Hugging Face's logo

Spaces:
thesanogoeffect
/
grobid-papercheck
Sleeping

App Files Community
1
Jakub Werner commited on
Commit
83ff8fb
·
1 Parent(s): c6f9cb5

activating LFS and pushing

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/D08-1073.training.segmentation.tei.xml +64 -0
  2. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/Document_image_zone_classification_A_simple_high-p.training.segmentation.tei.xml +33 -0
  3. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/Duguet_12279.training.segmentation.tei.xml +33 -0
  4. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/E14-1007.training.segmentation.tei.xml +103 -0
  5. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/E14-1075.training.segmentation.tei.xml +84 -0
  6. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EJM_CavitationModelsComparison.training.segmentation.tei.xml +183 -0
  7. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/ENEURO.0140-18.2018.training.segmentation.tei.xml +119 -0
  8. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/ENEURO.0415-18.2019.training.segmentation.tei.xml +0 -0
  9. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EPJ0290369-CC.training.segmentation.tei.xml +36 -0
  10. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EPL0330153-CC.training.segmentation.tei.xml +38 -0
  11. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EPL0410207-CC.training.segmentation.tei.xml +32 -0
  12. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EPL0480073-CC.training.segmentation.tei.xml +38 -0
  13. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EPL0490086-CC.training.segmentation.tei.xml +21 -0
  14. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EPL0580589-CC.training.segmentation.tei.xml +58 -0
  15. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/Faraday-discuss-post-ref.training.segmentation.tei.xml +83 -0
  16. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/Fatnassi2015_Article_MagneticAndMagnetocaloricPrope.training.segmentation.tei.xml +75 -0
  17. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/FrClassTSNM.training.segmentation.tei.xml +48 -0
  18. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/French_in_Auvergne.training.segmentation.tei.xml +25 -0
  19. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/Gasiorowski_PLoS_Pathogens_2019-1_number_prefix.training.segmentation.tei.xml +210 -0
  20. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/HCII07-LongSteph.training.segmentation.tei.xml +29 -0
  21. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/Hadbi_11990.training.segmentation.tei.xml +16 -0
  22. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/IJPA-13-549.training.segmentation.tei.xml +95 -0
  23. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/IJPA-13-567.training.segmentation.tei.xml +89 -0
  24. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/IJPA-13-648.training.segmentation.tei.xml +65 -0
  25. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/J09-3003.training.segmentation.tei.xml +0 -0
  26. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/J11-1004.training.segmentation.tei.xml +0 -0
  27. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/J11-3009.training.segmentation.tei.xml +33 -0
  28. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/J90-2003.training.segmentation.tei.xml +0 -0
  29. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JAP0702234-CC.training.segmentation.tei.xml +60 -0
  30. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JAP0805169-CC.training.segmentation.tei.xml +62 -0
  31. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JAP0897669-CC.training.segmentation.tei.xml +42 -0
  32. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JCMM-23-1343.training.segmentation.tei.xml +85 -0
  33. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JP01103169-CC.training.segmentation.tei.xml +43 -0
  34. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPC0150519-CC.training.segmentation.tei.xml +28 -0
  35. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPC0150L17-CC.training.segmentation.tei.xml +30 -0
  36. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPS0710709-CC.training.segmentation.tei.xml +47 -0
  37. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPS0722453-CC.training.segmentation.tei.xml +34 -0
  38. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPS0723041-CC.training.segmentation.tei.xml +38 -0
  39. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPS0730819-CC.training.segmentation.tei.xml +38 -0
  40. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPS0731297-CC.training.segmentation.tei.xml +49 -0
  41. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPS0732912-CC.training.segmentation.tei.xml +39 -0
  42. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L087187001-CC.training.segmentation.tei.xml +55 -0
  43. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L089147002-CC.training.segmentation.tei.xml +29 -0
  44. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L089157004-CC.training.segmentation.tei.xml +31 -0
  45. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L090137001-CC.training.segmentation.tei.xml +31 -0
  46. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L090137002-CC.training.segmentation.tei.xml +21 -0
  47. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L091217001-CC.training.segmentation.tei.xml +57 -0
  48. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L092157004-CC.training.segmentation.tei.xml +33 -0
  49. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L092227003-CC.training.segmentation.tei.xml +27 -0
  50. grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L093156802-CC.training.segmentation.tei.xml +29 -0
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/D08-1073.training.segmentation.tei.xml ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei>
3
+ <teiHeader>
4
+ <fileDesc xml:id="__D08-1073"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <note place="footnote"> Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing, pages 698–706, <lb/>Honolulu, October 2008. c <lb/> 2008 Association for Computational Linguistics <lb/></note>
8
+
9
+ <front> Jointly Combining Implicit Constraints Improves Temporal Ordering <lb/> Nathanael Chambers and Dan Jurafsky <lb/> Department of Computer Science <lb/>Stanford University <lb/>Stanford, CA 94305 <lb/> {natec,jurafsky}@stanford.edu <lb/> Abstract <lb/> Previous work on ordering events in text has <lb/>typically focused on local pairwise decisions, <lb/>ignoring globally inconsistent labels. How-<lb/>ever, temporal ordering is the type of domain <lb/>in which global constraints should be rela-<lb/>tively easy to represent and reason over. This <lb/>paper presents a framework that informs lo-<lb/>cal decisions with two types of implicit global <lb/>constraints: transitivity (A before B and B be-<lb/>fore C implies A before C) and time expression <lb/>normalization (e.g. last month is before yes-<lb/>terday). We show how these constraints can <lb/>be used to create a more densely-connected <lb/>network of events, and how global consis-<lb/>tency can be enforced by incorporating these <lb/>constraints into an integer linear programming <lb/>framework. We present results on two event <lb/>ordering tasks, showing a 3.6% absolute in-<lb/>crease in the accuracy of before/after classifi-<lb/>cation over a pairwise model. <lb/></front>
10
+
11
+ <body>1 Introduction <lb/> Being able to temporally order events is a neces-<lb/>sary component for complete document understand-<lb/>ing. Interest in machine learning approaches for this <lb/>task has recently been encouraged through the cre-<lb/>ation of the Timebank Corpus (Pustejovsky et al., <lb/>2003). However, most work on event-event order-<lb/>ing has focused on improving classifiers for pair-<lb/>wise decisions, ignoring obvious contradictions in <lb/>the global space of events when misclassifications <lb/>occur. A global framework to repair these event or-<lb/>dering mistakes has not yet been explored. <lb/>This paper addresses three main factors involved <lb/>in a global framework: the global optimization al-<lb/>gorithm, the constraints that are relevant to the task, <lb/>and the level of connectedness across pairwise de-<lb/>cisions. We employ Integer Linear Programming to <lb/>address the first factor, drawing from related work <lb/>in paragraph ordering (Bramsen et al., 2006). After <lb/>finding minimal gain with the initial model, we ex-<lb/>plore reasons for and solutions to the remaining two <lb/>factors through temporal reasoning and transitivity <lb/>rule expansion. <lb/>We analyze the connectivity of the Timebank Cor-<lb/>pus and show how textual events can be indirectly <lb/>connected through a time normalization algorithm <lb/>that automatically creates new relations between <lb/>time expressions. We show how this increased con-<lb/>nectivity is essential for a global model to improve <lb/>performance. <lb/>We present three progressive evaluations of our <lb/>global model on the Timebank Corpus, showing a <lb/> 3.6% gain in accuracy over its original set of re-<lb/>lations, and an 81% increase in training data size <lb/>from previous work. In addition, we present the first <lb/>results on Timebank that include an unknown rela-<lb/>tion, establishing a benchmark for performance on <lb/>the full task of document ordering. <lb/> 2 Previous Work <lb/> Recent work on classifying temporal relations <lb/>within the Timebank Corpus built 6-way relation <lb/>classifiers over 6 of the corpus&apos; 13 relations (Mani et <lb/>al., 2006; Mani et al., 2007; Chambers et al., 2007). <lb/>A wide range of features are used, ranging from sur-<lb/>face indicators to semantic classes. Classifiers make <lb/>
12
+
13
+ <page>698 <lb/></page>
14
+
15
+ local pairwise decisions and do not consider global <lb/>implications between the relations. <lb/>The TempEval-07 (Verhagen et al., 2007) contest <lb/>recently used two relations, before and after, in a <lb/>semi-complete textual classification task with a new <lb/>third relation to distinguish relations that can be la-<lb/>beled with high confidence from those that are un-<lb/>certain, called vague. The task was a simplified clas-<lb/>sification task from Timebank in that only one verb, <lb/>the main verb, of each sentence was used. Thus, the <lb/>task can be viewed as ordering the main events in <lb/>pairwise sentences rather than the entire document. <lb/>This paper uses the core relations of TempEval <lb/>(before,after,vague) and applies them to a full docu-<lb/>ment ordering task that includes every labeled event <lb/>in Timebank. In addition, we extend the previous <lb/>work by including a temporal reasoning component <lb/>and embedding it within a global constraint model. <lb/> 3 The Timebank Corpus <lb/> The Timebank Corpus (Pustejovsky et al., 2003) is <lb/>a corpus of 186 newswire articles that are tagged <lb/>for events, time expressions, and relations between <lb/>the events and times. The individual events are fur-<lb/>ther tagged for temporal information such as tense, <lb/>modality and grammatical aspect. Time expressions <lb/>use the TimeML (Ingria and Pustejovsky, 2002) <lb/>markup language. There are 6 main relations and <lb/>their inverses in Timebank: before, ibefore, includes, <lb/>begins, ends and simultaneous. <lb/> This paper describes work that classifies the re-<lb/>lations between events, making use of relations be-<lb/>tween events and times, and between the times <lb/>themselves to help inform the decisions. <lb/> 4 The Global Model <lb/> Our initial model has two components: (1) a pair-<lb/>wise classifier between events, and (2) a global con-<lb/>straint satisfaction layer that maximizes the confi-<lb/>dence scores from the classifier. The first is based <lb/>on previous work (Mani et al., 2006; Chambers et <lb/>al., 2007) and the second is a novel contribution to <lb/>event-event classification. <lb/> 4.1 Pairwise Classification <lb/> Classifying the relation between two events is the <lb/>basis of our model. A soft classification with confi-<lb/>dence scores is important for the global maximiza-<lb/>tion step that is described in the next section. As <lb/>in Chambers et al. (2007), we build support vec-<lb/>tor machine (SVM) classifiers and use the probabili-<lb/>ties from pairwise SVM decisions as our confidence <lb/>scores. These scores are then used to choose an op-<lb/>timal global ordering. <lb/>Following our previous work, we use the set of <lb/>features summarized in figure 1. They vary from <lb/>POS tags and lexical features surrounding the event, <lb/>to syntactic dominance, to whether or not the events <lb/>share the same tense, grammatical aspect, or aspec-<lb/>tual class. These features are the highest performing <lb/>set on the basic 6-way classification of Timebank. <lb/> Feature <lb/>Description <lb/>Word* <lb/>The text of the event <lb/>Lemma* <lb/>The lemmatized head word <lb/>Synset* <lb/>The WordNet synset of head word <lb/>POS* <lb/>4 POS tags, 3 before, and 1 event <lb/>POS bigram* The POS bigram of the event and its <lb/>preceding tag <lb/>Prep* <lb/>Preposition lexeme, if in a preposi-<lb/>tional phrase <lb/>Tense* <lb/>The event&apos;s tense <lb/>Aspect* <lb/>The event&apos;s grammatical aspect <lb/>Modal* <lb/>The modality of the event <lb/>Polarity* <lb/>Positive or negative <lb/>Class* <lb/>The aspecual class of the event <lb/>Tense Pair <lb/>The two concatenated tenses <lb/>Aspect Pair <lb/>The two concatenated aspects <lb/>Class Pair <lb/>The two concatenated classes <lb/>POS Pair <lb/>The two concatenated POS tags <lb/>Tense Match <lb/>true if the events have the same tense <lb/>Aspect Match true if the events have the same as-<lb/>pect <lb/>Class Match <lb/>true if the events have the same class <lb/>Dominates <lb/>true if the first event syntactically <lb/>dominates the second <lb/>Text Order <lb/>true if the first event occurs first in <lb/>the document <lb/>Entity Match <lb/>true if they share an entity as an ar-<lb/>gument <lb/>Same Sent <lb/>true if both events are in the same <lb/>sentence <lb/>Figure 1: The features to learn temporal relations be-<lb/>tween two events. Asterisks (*) indicate features that are <lb/>duplicated, one for each of the two events. <lb/> We use Timebank&apos;s hand tagged attributes in the <lb/>feature values for the purposes of this comparative <lb/>
16
+
17
+ <page> 699 <lb/></page>
18
+
19
+ before after unknown <lb/>A r1 B <lb/>.5 <lb/>.3 <lb/>.2 <lb/>B r2 C <lb/>.4 <lb/>.3 <lb/>.3 <lb/>A r3 C <lb/>.4 <lb/>.5 <lb/>.1 <lb/>total <lb/> 1.3 <lb/> 1.1 <lb/>.6 <lb/>A r1 B <lb/>.5 <lb/>.3 <lb/>.2 <lb/>B r2 C <lb/>.4 <lb/>.3 <lb/>.3 <lb/>A r3 C <lb/>.2 <lb/>.7 <lb/>.1 <lb/>total <lb/>1.1 <lb/> 1.3 <lb/> .6 <lb/> Figure 2: Two sets of confidence scores. The first set <lb/>chooses before for all three labels, and the second chooses <lb/> after. Other lower-scoring valid relation sets also exist, <lb/>such as before, unknown, and before. <lb/> study of global constraints, described next. <lb/> 4.2 Global Constraints <lb/> Pairwise classifiers can make contradictory classifi-<lb/>cations due to their inability to consider other deci-<lb/>sions. For instance, the following three decisions are <lb/>in conflict: <lb/> A before B <lb/>B before C <lb/>A after C <lb/> Transitivity is not taken into account. In fact, there <lb/>are several ways to resolve the conflict in this exam-<lb/>ple. Given confidence scores (or probabilities) for <lb/>each possible relation between the three pairs, we <lb/>can compute an optimal label assignment. Differ-<lb/>ent scores can lead to different conflict resolutions. <lb/>Figure 2 shows two resolutions given different sets <lb/>of scores. The first chooses before for all three rela-<lb/>tions, while the second chooses after. <lb/> Bramsen et al. (2006) presented a variety of ap-<lb/>proaches to using transitivity constraints to help in-<lb/>form pairwise decisions. They found that Integer <lb/>Linear Programming (ILP) performed the best on a <lb/>paragraph ordering task, consistent with its property <lb/>of being able to find the optimal solution for a set <lb/>of constraints. Other approaches are variations on <lb/>a greedy strategy of adding pairs of events one at a <lb/>time, ordered by their confidence. These can lead to <lb/>suboptimal configurations, although they are guar-<lb/>anteed to find a solution. Mani et al. (2007) sub-<lb/>sequently proposed one of these greedy strategies as <lb/>well, but published results are not available. We also <lb/>implemented a greedy best-first strategy, but found <lb/>ILP outperformed it. <lb/>Our Integer Linear Programming framework uses <lb/>the following objective function: <lb/> max <lb/> i <lb/> j <lb/> p ij x ij <lb/> (1) <lb/>with added constraints: <lb/> ∀i∀j x ij ∈ {0, 1} <lb/> (2) <lb/> ∀i x i1 + x i2 + ... + x im = 1 <lb/> (3) <lb/>where x ij represents the ith pair of events classified <lb/>as the jth relation of m relations. Thus, each pair <lb/>of events generates m variables. Given n pairs of <lb/>events, there are n * m variables. p ij is the proba-<lb/>bility of classifying pair i with relation j. Equation <lb/>2 (the first constraint) simply says that each variable <lb/>must be 0 or 1. Equation 3 contains m variables for <lb/>a single pair of events i representing its m possible <lb/>relations. It states that one relation must be set to 1 <lb/> and the rest to 0. In other words, a pair of events <lb/>cannot have two relations at the same time. Finally, <lb/>a transitivity constraint is added for all connected <lb/>pairs i, j, k, for each transitivity condition that infers <lb/>relation c given a and b: <lb/>x ia + x jb − x kc &lt;= 1 <lb/> (4) <lb/>We generated the set of constraints for each doc-<lb/>ument and used lpsolve 1 to solve the ILP constraint <lb/>problem. <lb/>The transitivity constraints are only effective if <lb/>the available pairwise decisions constitute a con-<lb/>nected graph. If pairs of events are disconnected, <lb/>then transitivity makes little to no contribution be-<lb/>cause these constraints are only applicable to con-<lb/>nected chains of events. <lb/> 4.3 Transitive Closure <lb/> In order to connect the event graph, we draw on <lb/>work from (Mani et al., 2006) and apply transitive <lb/>closure to our documents. Transitive closure was <lb/>first proposed not to address the problem of con-<lb/>nected event graphs, but rather to expand the size <lb/>of training data for relations such as before. Time-<lb/>bank is a relatively small corpus with few examples <lb/>
20
+
21
+ <note place="footnote"> 1 http://sourceforge.net/projects/lpsolve <lb/></note>
22
+
23
+ <page> 700 <lb/></page>
24
+
25
+ Total Event-Event Relations After Closure <lb/> before after <lb/> Timebank <lb/> 592 <lb/>656 <lb/>+ closure <lb/>3919 3405 <lb/> Figure 3: The number of event-event relations after tran-<lb/>sitive closure. <lb/> of each relation. One way of expand the training <lb/>set is through transitive rules. A few rules are given <lb/>here: <lb/> A simultaneous B ∧ A bef ore C → B bef ore C <lb/>A includes B ∧ A ibef ore C → B bef ore C <lb/>A bef ore B ∧ A ends C → B af ter C <lb/> While the original motivation was to expand the <lb/>training size of tagged relations, this approach also <lb/>creates new connections in the graph, replacing pre-<lb/>viously unlabeled event pairs with their true rela-<lb/>tions. We adopted this approach and closed the orig-<lb/>inal set of 12 relations to help connect the global <lb/>constraint model. <lb/> 4.4 Initial Experiment <lb/> The first evaluation of our global temporal model <lb/>is on the Timebank Corpus over the labeled rela-<lb/>tions before and after. We merged ibefore and iafter <lb/> into these two relations as well, ignoring all oth-<lb/>ers. We use this task as a reduced evaluation to <lb/>study the specific contribution of global constraints. <lb/>We also chose this strict ordering task because it is <lb/>well defined from a human understanding perspec-<lb/>tive. Snow et al. (2008) shows that average inter-<lb/>net users can make before/after decisions with very <lb/>high confidence, although the distinction with an un-<lb/>known relation is not as clear. An evaluation includ-<lb/>ing unknown (or vague as in TempEval) is presented <lb/>later. <lb/>We expanded the corpus (prior to selecting the be-<lb/>fore/after relations) using transitive closure over all <lb/>12 relations as described above. Figure 3 shows the <lb/>increase in data size. The number of before and after <lb/> relations increase by a factor of six. <lb/>We trained and tested the system with 10-fold <lb/>cross validation and micro-averaged accuracies. The <lb/>folds were randomly generated to separate the 186 <lb/>files into 10 folds (18 or 19 files per fold). The same <lb/>10-way split is used for all the evaluations. We used <lb/> Comparative Results <lb/> Training Set <lb/>Accuracy <lb/>Timebank Pairwise <lb/> 66.8% <lb/> Global Model <lb/> 66.8% <lb/> Figure 4: Using the base Timebank annotated tags for <lb/>testing, accuracy on before/after tags in the two models. <lb/> libsvm 2 to implement our SVM classifiers. <lb/>Figure 4 shows the results from our ILP model <lb/>with transitivity constraints. The first row is the <lb/>baseline pairwise classification trained and tested on <lb/>the original Timebank relations. The second row <lb/>gives performance with ILP. The model shows no <lb/>improvement. The global ILP constraints did affect <lb/>local decisions, changing 175 of them (out of 7324), <lb/>but the changes cancelled out and had no affect on <lb/>overall accuracy. <lb/> 4.5 Loosely Connected Graph <lb/> Why didn&apos;t a global model help? The problem lies <lb/>in the graph structure of Timebank&apos;s annotated rela-<lb/>tions. The Timebank annotators were not required <lb/>to annotate relations between any particular pair of <lb/>events. Instead, they were instructed to annotate <lb/>what seemed appropriate due to the almost insur-<lb/>mountable task of annotating all pairs of events. A <lb/>modest-sized document of 30 events, for example, <lb/>would contain <lb/> 30 <lb/> 2 <lb/> = 435 possible pairs. Anno-<lb/>tators thus marked relations where they deemed fit, <lb/>most likely between obvious and critical relations to <lb/>the understanding of the article. The vast majority of <lb/>possible relations are untagged, thus leaving a large <lb/>set of unlabeled (and disconnected) unknown rela-<lb/>tions. <lb/>Figure 5 graphically shows all relations that are <lb/>annotated between events and time expressions in <lb/>one of the shorter Timebank documents. Nodes rep-<lb/>resent events and times (event nodes start with the <lb/>letter &apos;e&apos;, times with &apos;t&apos;), and edges represent tempo-<lb/>ral relations. Solid lines indicate hand annotations, <lb/>and dotted lines indicate new rules from transitive <lb/>closure (only one, from event e4 to time t14). As <lb/>can be seen, the graph is largely disconnected and <lb/>a global model contributes little information since <lb/>transitivity constraints cannot apply. <lb/>
26
+
27
+ <note place="footnote"> 2 http://www.csie.ntu.edu.tw/˜cjlin/libsvm <lb/></note>
28
+
29
+ <page> 701 <lb/></page>
30
+
31
+ Timebank Annotation of wsj 0551 <lb/> Figure 5: Annotated relations in document wsj 0551. <lb/> The large amount of unlabeled relations in the <lb/>corpus presents several problems. First, building a <lb/>classifier for these unknown relations is easily over-<lb/>whelmed by the huge training set. Second, many of <lb/>the untagged pairs have non-unknown ordering rela-<lb/>tions between them, but were missed by the annota-<lb/>tors. This point is critical because one cannot filter <lb/>this noise when training an unknown classifier. The <lb/>noise problem will appear later and will be discussed <lb/>in our final experiment. Finally, the space of an-<lb/>notated events is very loosely connected and global <lb/>constraints cannot assist local decisions if the graph <lb/>is not connected. The results of this first experiment <lb/> illustrate this latter problem. <lb/>Bethard et al. (2007) strengthen the claim that <lb/>many of Timebank&apos;s untagged relations should not <lb/>be left unlabeled. They performed an independent <lb/>annotation of 129 of Timebank&apos;s 186 documents, <lb/>tagging all events in verb-clause relationships. They <lb/>found over 600 valid before/after relations that are <lb/>untagged in Timebank, on average three per docu-<lb/>ment. One must assume that if these nearby verb-<lb/>clause event pairs were missed by the annotators, <lb/>the much larger number of pairs that cross sentence <lb/>boundaries were also missed. <lb/>The next model thus attempts to fill in some of the <lb/>gaps and further connect the event graph by using <lb/>two types of knowledge. The first is by integrating <lb/>Bethard&apos;s data, and the second is to perform tempo-<lb/>ral reasoning over the document&apos;s time expressions <lb/>(e.g. yesterday or january 1999). <lb/> 5 A Global Model With Time <lb/> Our initial model contained two components: (1) a <lb/>pairwise classifier between events, and (2) a global <lb/>constraint satisfaction layer. However, due to the <lb/>sparseness in the event graph, we now introduce <lb/>a third component addressing connectivity: (3) a <lb/>temporal reasoning component to inter-connect the <lb/>global graph and assist in training data expansion. <lb/>One important aspect of transitive closure in-<lb/>cludes the event-time and time-time relations during <lb/>closure, not just the event-event links. Starting with <lb/>5,947 different types of relations, transitive rules in-<lb/>crease the dataset to approximately 12,000. How-<lb/>ever, this increase wasn&apos;t enough to be effective in <lb/>global reasoning. To illustrate the sparsity that still <lb/>remains, if each document was a fully connected <lb/>graph of events, Timebank would contain close to <lb/>160,000 relations 3 , more than a 13-fold increase. <lb/>More data is needed to enrich the Timebank event <lb/>graph. Two types of information can help: (1) more <lb/>event-event relations, and (2) a separate type of in-<lb/>formation to indirectly connect the events: event-<lb/>X-event. We incorporate the new annotations from <lb/>Bethard et al. (2007) to address (1) and introduce <lb/>a new temporal reasoning procedure to address (2). <lb/>The following section describes this novel approach <lb/>to adding time expression information to further <lb/>connect the graph. <lb/> 5.1 Time-Time Information <lb/> As described above, we use event-time relations to <lb/>produce the transitive closure, as well as annotated <lb/>time-time relations. It is unclear if Mani et al. (2006) <lb/>used these latter relations in their work. <lb/>However, we also add new time-time links that <lb/>are deduced from the logical time intervals that they <lb/>describe. Time expressions can be resolved to time <lb/>intervals with some accuracy through simple rules. <lb/>New time-time relations can then be added to our <lb/>space of events through time stamp comparisons. <lb/>Take this newswire example: <lb/> The Financial Times 100-share index shed 47.3 points to <lb/>close at 2082.1, down 4.5% from the previous Friday, <lb/> and 6.8% from Oct. 13, when Wall Street&apos;s plunge helped <lb/>spark the current weakness in London. <lb/>
32
+
33
+ <note place="footnote"> 3 Sum over the # of events n d in each document d, <lb/> n d <lb/> 2 <lb/></note>
34
+
35
+ <page> 702 <lb/></page>
36
+
37
+ The first two expressions (&apos;previous Friday&apos; <lb/> and &apos;Oct. 13&apos;) are in a clear before relation-<lb/>ship that Timebank annotators captured. <lb/>The <lb/> &apos;current&apos; expression, is correctly tagged with the <lb/> PRESENT REF attribute to refer to the document&apos;s <lb/>timestamp. Both &apos;previous Friday&apos; and &apos;Oct. 13&apos; <lb/> should thus be tagged as being before this expres-<lb/>sion. However, the annotators did not tag either <lb/>of these two before relations, and so our timestamp <lb/>resolution procedure fills in these gaps. This is a <lb/>common example of two expressions that were not <lb/>tagged by the annotators, yet are in a clear temporal <lb/>relationship. <lb/>We use Timebank&apos;s gold standard TimeML an-<lb/>notations to extract the dates and times from the <lb/>time expressions. In addition, those marked as <lb/> PRESENT REF are resolved to the document times-<lb/>tamp. Time intervals that are strictly before or after <lb/>each other are thus labeled and added to our space <lb/>of events. We create new before relations based on <lb/>the following procedure: <lb/> if event1.year &lt; event2.year <lb/>return true <lb/>if event1.year == event2.year <lb/>if event1.month &lt; event2.month <lb/>return true <lb/>if event1.month == event2.month <lb/>if event1.day &lt; event2.day <lb/>return true <lb/>end <lb/>end <lb/>return false <lb/> All other time-time orderings not including the <lb/> before relation are ignored (i.e. includes is not cre-<lb/>ated, although could be with minor changes). <lb/>This new time-time knowledge is used in two sep-<lb/>arate stages of our model. The first is just prior to <lb/>transitive closure, enabling a larger expansion of our <lb/>tagged relations set and reduce the noise in the un-<lb/>known set. The second is in the constraint satisfac-<lb/>tion stage where we add our automatically computed <lb/>time-time relations (with the gold event-time rela-<lb/>tions) to the global graph to help correct local event-<lb/>event mistakes. <lb/> Total Event-Event Relations After Closure <lb/> before after <lb/> Timebank <lb/>3919 3405 <lb/>+ time-time <lb/>5604 5118 <lb/>+ time/bethard 7111 6170 <lb/> Figure 6: The number of event-event before and after re-<lb/>lations after transitive closure on each dataset. <lb/> Comparative Results with Closure <lb/> Training Set <lb/>Accuracy <lb/>Timebank Pairwise <lb/> 66.8% <lb/> Global Model <lb/> 66.8% <lb/> Global + time/bethard <lb/> 70.4% <lb/> Figure 7: Using the base Timebank annotated tags for <lb/>testing, the increase in accuracy on before/after tags. <lb/> 5.2 Temporal Reasoning Experiment <lb/> Our second evaluation continues the use of the two-<lb/>way classification task with before and after to ex-<lb/>plore the contribution of closure, time normaliza-<lb/>tion, and global constraints. <lb/>We augmented the corpus with the labeled rela-<lb/>tions from Bethard et al. (2007) and added the au-<lb/>tomatically created time-time relations as described <lb/>in section 5.1. We then expanded the corpus using <lb/>transitive closure. Figure 6 shows the progressive <lb/>data size increase as we incrementally add each to <lb/>the closure algorithm. <lb/>The time-time generation component automati-<lb/>cally added 2459 new before and after time-time re-<lb/>lations into the 186 Timebank documents. This is <lb/>in comparison to only 157 relations that the human <lb/>annotators tagged, less than 1 per document on av-<lb/>erage. The second row of figure 6 shows the dras-<lb/>tic effect that these time-time relations have on the <lb/>number of available event-event relations for train-<lb/>ing and testing. Adding both Bethard&apos;s data and <lb/>the time-time data increases our training set by 81% <lb/> over closure without it. <lb/>We again performed 10-fold cross validation with <lb/>micro-averaged accuracies, but each fold tested only <lb/>on the transitively closed Timebank data (the first <lb/>row of figure 6). The training set used all available <lb/>data (the third row of figure 6) including the Bethard <lb/>data as well as our new time-time links. <lb/>
38
+
39
+ <page> 703 <lb/></page>
40
+
41
+ Figure 7 shows the results from the new model. <lb/>The first row is the baseline pairwise classification <lb/>trained and tested on the original relations only. Our <lb/>model improves by 3.6% absolute. This improve-<lb/>ment is statistically significant (p &lt; 0.000001, Mc-<lb/>Nemar&apos;s test, 2-tailed). <lb/> 5.3 Discussion <lb/> To further illustrate why our model now improves <lb/>local decisions, we continue our previous graph ex-<lb/>ample. The actual text for the graph in figure 5 is <lb/>shown here: <lb/> docstamp: 10/30/89 (t14) <lb/> Trustcorp Inc. will become(e1) Society Bank &amp; Trust <lb/>when its merger(e3) is completed(e4) with Society Corp. <lb/>of Cleveland, the bank said(e5). Society Corp., which is <lb/>also a bank, agreed(e6) in June(t15) to buy(e8) Trustcorp <lb/>for 12.4 million shares of stock with a market value of <lb/>about $450 million. The transaction(e9) is expected(e10) <lb/>to close(e2) around year end(t17). <lb/> The automatic time normalizer computes and adds <lb/>three new time-time relations, two connecting t15 <lb/>and t17 with the document timestamp, and one con-<lb/>necting t15 and t17 together. These are not other-<lb/>wise tagged in the corpus. <lb/> Time-Time + Closure <lb/> Figure 8: Before and after time-time links with closure. <lb/> Figure 8 shows the augmented document. The <lb/>double-line arrows indicate the three new time-time <lb/>relations and the dotted edges are the new relations <lb/>added by our transitive closure procedure. Most crit-<lb/>ical to this paper, three of the new edges are event-<lb/>event relations that help to expand our training data. <lb/>If this document was used in testing (rather than <lb/>training), these new edges would help inform our <lb/>transitive rules during classification. <lb/>Even with this added information, disconnected <lb/>segments of the graph are still apparent. However, <lb/>the 3.6% performance gain encourages us to move <lb/>to the final full task. <lb/> 6 Final Experiment with Unknowns <lb/> Our final evaluation expands the set of relations to <lb/>include unlabeled relations and tests on the entire <lb/>dataset available to us. The following is now a clas-<lb/>sification task between the three relations: before, <lb/>after, and unknown. <lb/> We duplicated the previous evaluation by adding <lb/>the labeled relations from Bethard et al. (2007) and <lb/>our automatically created time-time relations. We <lb/>then expanded this dataset using transitive closure. <lb/>Unlike the previous evaluation, we also use this en-<lb/>tire dataset for testing, not just for training. Thus, all <lb/>event-event relations in Bethard as well as Timebank <lb/>are used to expand the dataset with transitive closure <lb/>and are used in training and testing. We wanted to <lb/>fully evaluate document performance on every pos-<lb/>sible event-event relation that logically follows from <lb/>the data. <lb/>As before, we converted IBefore and IAfter into <lb/> before and after respectively, while all other rela-<lb/>tions are reduced to unknown. This relation set co-<lb/>incides with TempEval-07&apos;s core three relations (al-<lb/>though they use vague instead of unknown). <lb/> Rather than include all unlabeled pairs in our un-<lb/>known set, we only include the unlabeled pairs that <lb/>span at most one sentence boundary. In other words, <lb/>events in adjacent sentences are included in the un-<lb/>known set if they were not tagged by the Timebank <lb/>annotators. The intuition is that annotators are more <lb/>likely to label nearby events, and so events in adja-<lb/>cent sentences are more likely to be actual unknown <lb/> relations if they are unlabeled. It is more likely that <lb/>distant events in the text were overlooked by con-<lb/>venience, not because they truly constituted an un-<lb/>known relationship. <lb/>The set of possible sentence-adjacent unknown re-<lb/>lations is very large (approximately 50000 unknown <lb/> compared to 7000 before), and so we randomly se-<lb/>lect a percentage of these relations for each evalu-<lb/>
42
+
43
+ <page> 704 <lb/></page>
44
+
45
+ Classification Accuracy <lb/> % unk base <lb/>global global+time <lb/>0 <lb/>72.0% 72.2% <lb/>74.0% <lb/>1 <lb/>69.4% 69.5% <lb/>71.3% <lb/>3 <lb/>65.5% 65.6% <lb/>67.1% <lb/>5 <lb/>63.7% 63.8% <lb/>65.3% <lb/>7 <lb/>61.2% 61.6% <lb/>62.8% <lb/>9 <lb/>59.3% 59.5% <lb/>60.6% <lb/>11 <lb/>58.1% 58.4% <lb/>59.4% <lb/>13 <lb/>57.1% 57.1% <lb/>58.1% <lb/> Figure 9: Overall accuracy when training with different <lb/>percentages of unknown relations included. 13% of un-<lb/>knowns is about equal to the number of befores. <lb/> ation. We used the same SVM approach with the <lb/>features described in section 4.1. <lb/> 6.1 Results <lb/> Results are presented in figure 9. The rows in the <lb/>table are different training/testing runs on varying <lb/>sizes of unknown training data. There are three <lb/>columns with accuracy results of increasing com-<lb/>plexity. The first, base, are results from pairwise <lb/>classification decisions over Timebank and Bethard <lb/>with no global model. The second, global, are re-<lb/>sults from the Integer Linear Programming global <lb/>constraints, using the pairwise confidence scores <lb/>from the base evaluation. Finally, the global+time <lb/> column shows the ILP results when all event-time, <lb/>time-time, and automatically induced time-time re-<lb/>lations are included in the global graph. <lb/>The ILP approach does not alone improve perfor-<lb/>mance on the event-event tagging task, but adding <lb/>the time expression relations greatly increases the <lb/>global constraint results. This is consistent with the <lb/>results from out first two experiments. The evalua-<lb/>tion with 1% of the unknown tags shows an almost <lb/> 2% improvement in accuracy. The gain becomes <lb/>smaller as the unknown set increases in size (1.0% <lb/>gain with 13% unknown). Unknown relations will <lb/>tend to be chosen as more weight is given to un-<lb/>knowns. When there is a constraint conflict in the <lb/>global model, unknown tends to be chosen because <lb/>it has no transitive implications. All improvements <lb/>from base to global+time are statistically significant <lb/>(p &lt; 0.000001, McNemar&apos;s test, 2-tailed). <lb/> Base Pairwise Classification <lb/> precision recall f1-score <lb/>before <lb/>61.4 <lb/>55.4 <lb/>58.2 <lb/>after <lb/>57.6 <lb/>53.1 <lb/>55.3 <lb/>unk <lb/>53.0 <lb/>62.8 <lb/>57.5 <lb/> Global+Time Classification <lb/> precision <lb/>recall <lb/>f1-score <lb/>before 63.7 (+2.3) 57.1 (+2.2) 60.2 (+2.0) <lb/> after <lb/>60.3 (+2.7) 54.3 (+2.9) 57.1 (+1.8) <lb/> unk <lb/>52.0 (-1.0) 62.9 (+0.1) 56.9 (-0.6) <lb/>Figure 10: Precision and Recall for the base pairwise de-<lb/>cisions and the global constraints with integrated time in-<lb/>formation. <lb/> The first row of figure 9 corresponds to the re-<lb/>sults in our second experiment in figure 7, but shows <lb/>higher accuracy. The reason is due to our different <lb/>test sets. This final experiment includes Bethard&apos;s <lb/>event-event relations in testing. The improved per-<lb/>formance suggests that the clausal event-event rela-<lb/>tions are easier to classify, agreeing with the higher <lb/>accuracies originally found by Bethard et al. (2007). <lb/>Figure 10 shows the precision, recall, and f-score <lb/>for the evaluation with 13% unknowns. This set was <lb/>chosen for comparison because it has a similar num-<lb/>ber of unknown labels as before labels. We see an <lb/>increase in precision in both the before and after de-<lb/>cisions by up to 2.7%, an increase in recall up to <lb/> 2.9%, and an fscore by as much as 2.0%. The un-<lb/>known relation shows mixed results, possibly due to <lb/>its noisy behavior as discussed throughout this pa-<lb/>per. <lb/> 6.2 Discussion <lb/> Our results on the two-way (before/after) task show <lb/>that adding additional implicit temporal constraints <lb/>and then performing global reasoning results in <lb/>significant improvements in temporal ordering of <lb/>events (3.6% absolute over simple pairwise deci-<lb/>sions). <lb/>Both before and after also showed increases in <lb/>precision and recall in the three-way evaluation. <lb/>However, unknown did not parallel this improve-<lb/>ment, nor are the increases as dramatic as in the two-<lb/>way evaluation. We believe this is consistent with <lb/>the noise that exists in the Timebank corpus for un-<lb/>labeled relations. Evidence from Bethard&apos;s indepen-<lb/>
46
+
47
+ <page> 705 <lb/></page>
48
+
49
+ dent annotations directly point to missing relations, <lb/>but the dramatic increase in the size of our closure <lb/>data (81%) from adding a small amount of time-time <lb/>relations suggests that the problem is widespread. <lb/>This noise in the unknown relation may be damp-<lb/>ening the gains that the two way task illustrates. <lb/>This work is also related to the task of event-time <lb/>classification. While not directly addressed in this <lb/>paper, the global methods described within clearly <lb/>apply to pairwise models of event-time ordering as <lb/>well. <lb/>Further progress in improving global constraints <lb/>will require new methods to more accurately iden-<lb/>tify unknown events, as well as new approaches to <lb/>create implicit constraints over the ordering. We ex-<lb/>pect such an improved ordering classifier to be used <lb/>to improve the performance of tasks such as summa-<lb/>rization and question answering about the temporal <lb/>nature of events. <lb/>
50
+
51
+ </body>
52
+
53
+ <back>
54
+
55
+ <div type="acknowledgement">Acknowledgments <lb/> This work is funded in part by DARPA through IBM <lb/>and by the DTO Phase III Program for AQUAINT. <lb/>We also thank our anonymous reviewers for many <lb/>helpful suggestions. <lb/></div>
56
+
57
+
58
+ <listBibl> References <lb/> Steven Bethard, James H. Martin, and Sara Klingenstein. <lb/>2007. Timelines from text: Identification of syntac-<lb/>tic temporal relations. In International Conference on <lb/>Semantic Computing. <lb/> Philip Bramsen, Pawan Deshpande, Yoong Keok Lee, <lb/>and Regina Barzilay. 2006. Inducing temporal graphs. <lb/>In Proceedings of EMNLP-06. <lb/> Nathanael Chambers, Shan Wang, and Dan Jurafsky. <lb/>2007. Classifying temporal relations between events. <lb/>In Proceedings of ACL-07, Prague, Czech Republic. <lb/>R Ingria and James Pustejovsky. 2002. TimeML specifi-<lb/>cation 1.0. In http://www.time2002.org. <lb/> Inderjeet Mani, Marc Verhagen, Ben Wellner, Chong Min <lb/>Lee, and James Pustejovsky. 2006. Machine learning <lb/>of temporal relations. In Proceedings of ACL-06, July. <lb/>Inderjeet Mani, Ben Wellner, Marc Verhagen, and James <lb/>Pustejovsky. 2007. Three approaches to learning <lb/>tlinks in timeml. Technical Report CS-07-268, Bran-<lb/>deis University. <lb/>James Pustejovsky, Patrick Hanks, Roser Sauri, Andrew <lb/>See, David Day, Lisa Ferro, Robert Gaizauskas, Mar-<lb/>cia Lazo, Andrea Setzer, and Beth Sundheim. 2003. <lb/>The timebank corpus. Corpus Linguistics, pages 647– <lb/>656. <lb/>Rion Snow, Brendan O&apos;Connor, Dan Jurafsky, and An-<lb/>drew Ng. 2008. Cheap and fast -but is it good? <lb/>evaluating non-expert annotations for natural language <lb/>tasks. In Proceedings of EMNLP-08, Waikiki, Hawaii, <lb/>USA. <lb/>Marc Verhagen, Robert Gaizauskas, Frank Schilder, <lb/>Mark Hepple, Graham Katz, and James Pustejovsky. <lb/>2007. Semeval-2007 task 15: Tempeval temporal re-<lb/>lation identification. In Workshop on Semantic Evalu-<lb/>ations. <lb/></listBibl>
59
+
60
+ <page> 706 </page>
61
+
62
+ </back>
63
+ </text>
64
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/Document_image_zone_classification_A_simple_high-p.training.segmentation.tei.xml ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <titlePage>See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/221415986 <lb/>Document image zone classification : A simple high-performance approach <lb/>Conference Paper • January 2007 <lb/>Source: DBLP <lb/>CITATIONS <lb/>57 <lb/>READS <lb/>446 <lb/>3 authors: <lb/>Some of the authors of this publication are also working on these related projects: <lb/>Religion in Native American literature View project <lb/>Dictionary Learning and Sparse Coding View project <lb/>Daniel Keysers <lb/>Google Inc. <lb/>139 PUBLICATIONS 5,893 CITATIONS <lb/>SEE PROFILE <lb/>Faisal Shafait <lb/>University of Western Australia <lb/>195 PUBLICATIONS 6,153 CITATIONS <lb/>SEE PROFILE <lb/>Thomas Breuel <lb/>Google Inc. <lb/>260 PUBLICATIONS 6,183 CITATIONS <lb/>SEE PROFILE <lb/>All content following this page was uploaded by Daniel Keysers on 12 July 2014. <lb/>The user has requested enhancement of the downloaded file. <lb/></titlePage>
8
+
9
+ <front>DOCUMENT IMAGE ZONE CLASSIFICATION <lb/> A Simple High-Performance Approach <lb/>Daniel Keysers, Faisal Shafait <lb/>German Research Center for Artificial Intelligence (DFKI) GmbH, Kaiserslautern, Germany <lb/>{daniel.keysers, faisal.shafait}@dfki.de <lb/>Thomas M. Breuel <lb/>Technical University of Kaiserslautern, Germany <lb/>[email protected] <lb/>Keywords: <lb/>Document Image Analysis, Zone Classification <lb/>Abstract: <lb/>We describe a simple, fast, and accurate system for document image zone classification -an important sub-<lb/>problem of document image analysis -that results from a detailed analysis of different features. Using <lb/>a novel combination of known algorithms, we achieve a very competitive error rate of 1.46% (n = 13811) <lb/>in comparison to (Wang et al., 2006) who report an error rate of 1.55% (n = 24177) using more complicated <lb/>techniques. The experiments were performed on zones extracted from the widely used UW-III database, which <lb/>is representative of images of scanned journal pages and contains ground-truthed real-world data. <lb/></front>
10
+
11
+ <body>1 INTRODUCTION <lb/>One important subtask of document image processing <lb/>is the classification of blocks detected by the physical <lb/>layout analysis system into one of a set of predefined <lb/>classes. For example, we may want to distinguish be-<lb/>tween text blocks and drawings to pass the former to <lb/>an OCR system and the latter to an image enhancer. <lb/>For a detailed discussion of the task and its relevance <lb/>please see e.g. (Wang et al., 2006). <lb/>During the design of our block classification sys-<lb/>tem we noticed that among the approaches we found <lb/>in the literature a detailed comparison of different fea-<lb/>tures was usually not performed, and in particular we <lb/>did not find a comparison that included features as <lb/>they are typically used in other image classification or <lb/>retrieval tasks. In this paper we address this shortcom-<lb/>ing by comparing a large set of commonly used fea-<lb/>tures for block classification and include in the com-<lb/>parison three features that are known to yield good <lb/>performance in content-based image retrieval (CBIR) <lb/>and are applicable to binary images (Deselaers et al., <lb/>2004). Interestingly, we found that the single feature <lb/>with the best performance is the Tamura texture his-<lb/>togram, which belongs to this latter class. Another re-<lb/>sult we transfer from experience in the area of CBIR <lb/>is that often a histogram is a more powerful feature <lb/>than using statistics of a distribution like mean and <lb/>variance only. We show that the use of histograms im-<lb/>proves the performance for block classification signif-<lb/>icantly in our experiments. By combining a number <lb/>of different features, we achieve a very competitive <lb/>error rate of less than 1.5% on a data set of blocks ex-<lb/>tracted from the well-known University of Washing-<lb/>ton III (UW-III) database. In addition to the data used <lb/>in prior work we include a class of &apos;speckles&apos; blocks <lb/>that often occur during photocopying and for which a <lb/>correct classification can facilitate further processing <lb/>of a document image. Figure 1 shows example block <lb/>images for each of the eight types distinguished in our <lb/>approach. We also present a very fast (but at 2.1% er-<lb/>ror slightly less accurate) classifier, using simple fea-<lb/>tures and only a fraction of a second to classify one <lb/>block on average on a standard PC. <lb/>
12
+ 2 RELATED WORK AND <lb/>CONTRIBUTION <lb/>We briefly discuss some related work in this section, <lb/>for a more detailed overview of related work in the <lb/>field of document zone classification please refer to <lb/>(Okun et al., 1999; Wang et al., 2006). Table 1 shows <lb/>an overview of related results in zone classification. <lb/>Inglis and Witten (Inglis and Witten, 1995) <lb/>math <lb/>logo <lb/>text <lb/>table <lb/>drawing <lb/>halftone <lb/>ruling <lb/>speckles <lb/>Figure 1: Examples of document image block types distinguished in our approach. <lb/>Table 1: Summary of UW zone classification error rates from the literature along with the number of pages, zones and block <lb/>types used. Note that an exact comparison between all error rates is not possible. <lb/>reference <lb/># pages # zones # types error [%] <lb/>(Inglis and Witten, 1995) <lb/>1001 <lb/>13831 <lb/>3 <lb/>6.7 <lb/>(Liang et al., 1996) <lb/>979 <lb/>13726 <lb/>8 <lb/>5.4 <lb/>(Sivaramakrishnan et al., 1995) <lb/>979 <lb/>13726 <lb/>9 <lb/>3.3 <lb/>(Wang et al., 2000) <lb/>1600 <lb/>24177 <lb/>9 <lb/>2.5 <lb/>(Wang et al., 2006) <lb/>1600 <lb/>24177 <lb/>9 <lb/>1.5 <lb/>this work <lb/>713 <lb/>13811 <lb/>8 <lb/>1.5 <lb/>present a study of the zone classification problem <lb/>as a machine learning problem. They use 13831 <lb/>zones from the UW database and distinguish the three <lb/>classes text, halftone, and drawing. Using seven <lb/>features based on connected components and run <lb/>lengths, the authors apply various machine learning <lb/>techniques to the problem, of which the C4.5 decision <lb/>tree performs best at 6.7% error rate. <lb/>The review paper by Okun et al. (Okun et al., <lb/>1999) succinctly summarizes the main approaches <lb/>used for document zone classification in the 1990s. <lb/>The predominant feature type is based on connected <lb/>components (see also for example (Liang et al., <lb/>1996)) and run-length statistics. Other features used <lb/>include the cross-correlation between scan-lines, ver-<lb/>tical projection profiles, wavelet coefficients, learned <lb/>masks, and the black pixel distribution. The most <lb/>common classifier used is a neural network. <lb/>The widespread use of features based on con-<lb/>nected components run-length statistics, combined <lb/>with the simplicity of implementation of such fea-<lb/>tures, led us to use these feature types in our exper-<lb/>iments as well, comparing them to the use of features <lb/>used in content-based image retrieval. Our CBIR fea-<lb/>tures are based on the open source image retrieval <lb/>system FIRE (Deselaers et al., 2004). We restrict <lb/>our analysis for zone classification to those features <lb/>that are promising for the analysis of binary images <lb/>as described in the following section. (The overall <lb/>most successful features in CBIR are usually based <lb/>on color information.) <lb/>The most recent and detailed overview of the <lb/>progress in document zone classification and a very <lb/>accurate system is presented in (Wang et al., 2006). <lb/>The authors use a decision tree classifier and model <lb/>contextual dependencies for some zones. In our work <lb/>we do not model zone context, although it is likely <lb/>that a context model (which can be integrated in a <lb/>similar way as presented by Wang et al.) would help <lb/>the overall classification performance. Wang et al. <lb/>use 24177 zones extracted from the UW-III database <lb/>to evaluate their approach. In our experiments we <lb/>use only 11804 labeled zones (plus 2007 additional <lb/>zones of type &apos;speckles&apos;) extracted from the UW-III <lb/>database because many zones occur in different ver-<lb/>sions in the database. In Section 5 we further illus-<lb/>trate this shortcoming and our approach to overcome <lb/>it. As the authors use 9-fold cross-validation to obtain <lb/>their results, it might be possible that the error rates <lb/>they present (the best result is an overall error rate of <lb/>1.5%) may be influenced positively by this fact, be-<lb/>cause it is likely that instances of blocks of the same <lb/>document occur in training and test set. In a simi-<lb/>lar direction, Wang et al. use one feature that &quot;uses <lb/>a statistical method to classify glyphs and was exten-<lb/>sively trained on the UWCDROM-III document im-<lb/>age database.&quot; It is not clear to us if this implies that <lb/>the glyphs that occur in testing have also been used in <lb/>the training of the glyph classifier. <lb/>
13
+ We expand on the work presented in (Wang et al., <lb/>2006) in the following ways: <lb/>• We include a detailed feature comparison includ-<lb/>ing a comparison with commonly used CBIR fea-<lb/>tures. It turns out that the single best feature is <lb/>the Tamura texture histogram which was not pre-<lb/>viously used for zone classification. <lb/>• We present results both for a simple nearest neigh-<lb/>bor classifier and for a very fast linear classifier <lb/>based on logistic regression and the maximum en-<lb/>tropy criterion. <lb/>• We introduce a new class of blocks containing <lb/>speckles that has not been labeled in the UW-III <lb/>database. This typical class of noise is important <lb/>to detect during the layout analysis especially for <lb/>images of photocopied documents. <lb/>• We present results for the part of the UW-III <lb/>database without using duplicates and achieve a <lb/>similar error rate of 1.5%. <lb/>• We introduce the use of histograms for the <lb/>measurements of connected components and run <lb/>lengths and show that this leads to a performance <lb/>increase. <lb/>3 FEATURE EXTRACTION <lb/>We extract the following features from each block, <lb/>where features 1-3 are chosen based on their perfor-<lb/>mance in CBIR (Deselaers et al., 2004) feature 4 was <lb/>expected to help distinguish between the types &apos;draw-<lb/>ing&apos; and &apos;text&apos; and features 5-9 were chosen based on <lb/>their common use in block classification (Okun et al., <lb/>1999; Wang et al., 2006). Due to space limitations we <lb/>refer the interested reader to the references for imple-<lb/>mentation details. <lb/>1. Tamura texture features histogram (TTFH) <lb/>2. Relational invariant feature histograms (RIFH) <lb/>3. Down-scaled images of size 32 × 32 (DSI) <lb/>4. The fill ratio, i.e. the ratio of the number of black <lb/>pixels in a horizontally smeared (Wong et al., <lb/>1982) image to the area of the image (FR) <lb/>5. Run-length histograms of black and white pixels <lb/>along horizontal, vertical, main diagonal, and side <lb/>diagonal directions; each histogram uses eight <lb/>bins, spaced apart as powers of 2, i.e. counting <lb/>runs of length ≤ 1, 3, 7, 15, 31, 63, 127 and ≥ 128 <lb/>(RL{B,W}{X,Y,M,S}H) <lb/>6. The vector formed by the total number, mean, <lb/>and variance of the runs of black and white pixels <lb/>along the horizontal, vertical, main diagonal, and <lb/>side diagonal directions as used in (Wang et al., <lb/>2006) (RL{B,W}{X,Y,M,S}V) <lb/>7. Histograms (as in 5) of the widths and heights of <lb/>connected components (CCXH, CCYH) <lb/>8. The joint distribution of the widths and heights of <lb/>connected components as a 2-dimensional 64-bin <lb/>histogram (CCXYH) <lb/>9. The histogram of the distances between a con-<lb/>nected component and its nearest neighbor com-<lb/>ponent (CCNNH) <lb/>4 CLASSIFICATION <lb/>To evaluate the various features, we use a simple near-<lb/>est neighbor classifier, that is, a test sample is clas-<lb/>sified into the class the closest training sample be-<lb/>longs to. The distance measures used are the Jensen-<lb/>Shannon divergence for histograms and the Euclidean <lb/>distance for all other features (Deselaers et al., 2004). <lb/>If different feature sets are combined, the overall dis-<lb/>tance is calculated as the weighted sum of the indi-<lb/>vidual normalized distances. The weights are pro-<lb/>portional to the inverse of the error rate of a particu-<lb/>lar feature. No tuning with respect to these weights <lb/>or with respect to the distance measures has been <lb/>performed. Although a k-nearest-neighbor approach <lb/>gives better results in many cases we only evaluated <lb/>the 1-nearest-neighbor classifier. The nearest neigh-<lb/>bor error rates are determined using leave-one-out <lb/>cross-validation. <lb/>The nearest neighbor classifier serves as a good <lb/>baseline classifier, although in many cases we can find <lb/>a more suitable classifier for a given task. As we con-<lb/>centrate on features in this paper, we did not test any <lb/>other classifiers. However, an important shortcoming <lb/>of the nearest neighbor classifier is its requirement on <lb/>computational resources. Both memory and run-time <lb/>can be prohibitive for some applications. To explore <lb/>a very fast approach with minimum requirements on <lb/>computational resources, we also trained a log-linear <lb/>classifier using the maximum entropy criterion (Key-<lb/>sers et al., 2002). The classification using this clas-<lb/>sifier can be obtained by computing a dot product of <lb/>the feature vector with a weight vector for each class <lb/>and choosing the maximum, and is thus very fast. <lb/>As only these weight vectors need to be stored, the <lb/>memory requirement is also minimal. Furthermore, <lb/>the maximum entropy approach yields a probabilistic <lb/>model, such that we obtain an estimate of the poste-<lb/>rior probability for each class. The maximum entropy <lb/>approach was evaluated on a regular 50/50 split of the <lb/>data into training and test set and thus only uses half <lb/>the amount of training data. The histograms were not <lb/>normalized for the maximum-entropy approach, but <lb/>the absolute numbers were used instead to allow the <lb/>classifier to utilize this additional information. <lb/>5 DATA SET <lb/>To evaluate our approach for document zone classifi-<lb/>cation, we use the University of Washington III (UW-<lb/>III) database (Guyon et al., 1997). The database con-<lb/>sists of 1600 English document images with bound-<lb/>ing boxes of 24177 homogeneous page segments or <lb/>blocks, which are manually labeled into different <lb/>classes depending on their contents, making the data <lb/>very suitable for evaluating a block classification sys-<lb/>tem, e.g. (Inglis and Witten, 1995; Wang et al., 2006). <lb/>The documents in the UW-III dataset are catego-<lb/>rized based on their degradation type as follows: <lb/>1. Direct scans of original English journals <lb/>2. Scans of first generation English journal photo-<lb/>copies <lb/>3. Scans of second or later generation English jour-<lb/>nal photocopies <lb/>Many of the documents in the dataset are dupli-<lb/>cated and differ sometimes only by the degradation <lb/>(a) E00D <lb/>(b) C000 <lb/>(c) E04A (photocopy) <lb/>(d) W033 (direct scan) <lb/>Figure 2: Example document pages from the UW-III database. Note that some documents, as shown on the right, occur in <lb/>different versions. For our experiments, we made sure that no such duplicates were used. <lb/>applied to them. This type of collection is useful <lb/>when one is evaluating a page segmentation algorithm <lb/>to see how well the algorithm performs when pho-<lb/>tocopy effect degradation is applied to a document. <lb/>However, the degradation introduced by photocopy-<lb/>ing a document does not affect the contents of a doc-<lb/>ument to a large extent. One such example can be <lb/>seen in Figure 2, where the same document is present <lb/>in the dataset four times (E04A, W033, S04A, W133, <lb/>two of them shown here). Although the photocopied <lb/>documents are darker than the corresponding direct <lb/>scans, the difference is not substantial. This dupli-<lb/>cation of documents tends to bias the evaluation re-<lb/>sults towards lower error rates when some of these <lb/>documents are used in training, while others are used <lb/>in testing. This effect seems to have been unnoticed <lb/>previously by some researchers who use the complete <lb/>dataset for the evaluation of their algorithms. <lb/>We decided to use a subset of the UW-III dataset <lb/>to avoid using duplicate documents. We chose doc-<lb/>uments in the scans from the first generation photo-<lb/>copies category because they were largest in number. <lb/>We use all the documents with prefixes A0, C0, D0, <lb/>IG, H0, J0, K0, E0, V0, I03, and I04. There are 713 <lb/>documents of this type. We extracted the ground-<lb/>truth zones and their labels from each of these 713 <lb/>documents. We observed that there were very few <lb/>examples of some of the zone types like &apos;seal&apos;, &apos;an-<lb/>nouncement&apos;, &apos;advertisement&apos;, etc. Therefore we se-<lb/>lected only those classes for evaluation that contained <lb/>at least ten example images. <lb/>One limitation of the UW-III ground-truth zones is <lb/>that they do not contain any example of noise regions, <lb/>i.e. regions that emerge from noise introduced dur-<lb/>ing the scanning or photocopy process. These regions <lb/>mostly consist of speckles and dots present along the <lb/>border of the document. Since such regions often ap-<lb/>pear in practice, it is important to detect such regions <lb/>as noise so that these can be removed from further <lb/>processing. The UW-III dataset images contain many <lb/>such regions but these are not labeled. In order to ex-<lb/>tract examples of such regions we used the page seg-<lb/>mentation algorithm from (Kise et al., 1998) to ex-<lb/>tract page segments. Then all the segments that did <lb/>not overlap with any of the ground-truth zones were <lb/>filtered out as examples of the noise zones. However <lb/>these contained both textual and non-textual noise. <lb/>Textual noise appears only along the left or the right <lb/>side of a document when the facing pages of a book <lb/>are scanned. Since these extraneous symbols cannot <lb/>be distinguished from the actual contents of the doc-<lb/>ument based on their appearance alone, we do not <lb/>consider examples of textual noise. Therefore we <lb/>only take examples of non-textual noise, i.e. speck-<lb/>les as noise class. The speckles heavily depend on the <lb/>degradation of the document and vary considerably <lb/>from the direct scan of a document to its first genera-<lb/>tion photocopy as can be seen in Figure 2. Therefore, <lb/>for the speckles class, we extracted examples from all <lb/>1600 documents of the UW-III database. The corre-<lb/>sponding number of examples used for each zone type <lb/>is included in Table 3. <lb/>6 EXPERIMENTAL RESULTS <lb/>Table 2 shows the error rates that the nearest neighbor <lb/>classifier achieves for each single feature along with <lb/>the dimensionality of the feature vectors and the av-<lb/>erage time used to compute the feature vector. (All <lb/>timing information is given for a standard PC with <lb/>1.8GHz AMD Athlon processor without special per-<lb/>formance tuning of the algorithms.) The last rows <lb/>show results for combined feature sets. <lb/>Table 2: Leave-one-out nearest neighbor error rates and ex-<lb/>traction run-times for each feature and for combinations. <lb/>feature <lb/># features extr.-time [s] error [%] <lb/>TTFH <lb/>512 <lb/>5.51 <lb/>3.4 <lb/>RIFH <lb/>512 <lb/>12.59 <lb/>7.8 <lb/>DSI <lb/>1024 <lb/>0.01 <lb/>8.1 <lb/>FR <lb/>1 <lb/>0.02 <lb/>27.3 <lb/>RLBXH <lb/>8 <lb/>0.01 <lb/>7.9 <lb/>RLWXH <lb/>8 <lb/>0.01 <lb/>5.1 <lb/>RLBYH <lb/>8 <lb/>0.01 <lb/>8.2 <lb/>RLWYH <lb/>8 <lb/>0.01 <lb/>5.6 <lb/>RLBMH <lb/>8 <lb/>0.01 <lb/>11.8 <lb/>RLWMH <lb/>8 <lb/>0.01 <lb/>6.6 <lb/>RLBSH <lb/>8 <lb/>0.01 <lb/>10.5 <lb/>RLWSH <lb/>8 <lb/>0.01 <lb/>6.2 <lb/>RLBXV <lb/>3 <lb/>0.01 <lb/>12.9 <lb/>RLWXV <lb/>3 <lb/>0.01 <lb/>9.7 <lb/>RLBYV <lb/>3 <lb/>0.01 <lb/>14.6 <lb/>RLWYV <lb/>3 <lb/>0.01 <lb/>12.1 <lb/>RLBMV <lb/>3 <lb/>0.01 <lb/>17.2 <lb/>RLWMV <lb/>3 <lb/>0.01 <lb/>12.6 <lb/>RLBSV <lb/>3 <lb/>0.01 <lb/>16.7 <lb/>RLWSV <lb/>3 <lb/>0.01 <lb/>12.2 <lb/>CCXH <lb/>8 <lb/>0.04 <lb/>14.5 <lb/>CCYH <lb/>8 <lb/>0.04 <lb/>14.9 <lb/>CCXYH <lb/>64 <lb/>0.04 <lb/>6.2 <lb/>CCNNH <lb/>8 <lb/>0.05 <lb/>19.0 <lb/>RL**V, constant weight <lb/>4.1 <lb/>RL**H, constant weight <lb/>1.8 <lb/>RL*, CC*, 1/error weight <lb/>1.5 <lb/>FR, RL*, CC*, 1/error weight <lb/>1.5 <lb/>TTFH, FR, RL*, CC*, 1/error weight <lb/>1.5 <lb/>RL*, CC*, logistic, 50/50 data split <lb/>2.1 <lb/>We can observe the following results: <lb/>• The Tamura texture feature is the single best fea-<lb/>ture but is more than 100 times slower to compute <lb/>than most other features. <lb/>• The use of as descriptors of the run-<lb/>lengths distribution leads to much lower error <lb/>rates than the use of number, mean, and variance. <lb/>The combination of these histograms alone leads <lb/>to a very good error rate of 1.8%. <lb/>• Interestingly, the use of the white (background) <lb/>runs for the computation of features consistently <lb/>leads to better results than the use of black (fore-<lb/>ground) runs. <lb/>• Among the run-lengths based features, those <lb/>based on the horizontal runs lead to the best er-<lb/>ror rates. <lb/>• The fill ratio as a single feature does not lead to <lb/>good results, which is not surprising as it consists <lb/>only of a single number. However, it is very use-<lb/>ful to distinguish drawings from text. This is how-<lb/>ever also achieved by using the distribution of the <lb/>white run lengths, such that the FR feature is not <lb/>part of the best observed feature set. <lb/>• By using a logistic classifier trained with the max-<lb/>imum entropy criterion (training time a few min-<lb/>utes, time for one classification in the order of a <lb/>few microseconds) on a feature set that is very <lb/>fast to extract, we can construct a zone type clas-<lb/>sifier that can classify more than five zones per <lb/>second even without performance tuning. At the <lb/>same time, the error rate is at 2.1% only slightly <lb/>higher than that of the best observed classifier. <lb/>Table 3 shows the frequency of misclassifications <lb/>between different classes of the best classifier. We can <lb/>observe that high recognition accuracy was achieved <lb/>for the text, ruling, speckles, math, halftone, and <lb/>drawing classes. However, our system failed to rec-<lb/>ognize logos correctly, and most of the logos were <lb/>misclassified as either text, or halftone/drawing. Note <lb/>that the accuracy rate for type &apos;logo&apos; in (Wang et al., <lb/>2006) is even lower at 0.0%. The reason for this ef-<lb/>fect is the very small number of samples for this class, <lb/>which on the other hand implies that it has only a very <lb/>small influence on the overall system error rate. Sim-<lb/>ilarly, the table detection accuracy was not high, and <lb/>about 21% of the tables were misclassified as text. <lb/>To visualize the errors made, we looked at the <lb/>nearest-neighbor images for each misclassified block. <lb/>Figure 3 shows some typical examples. It can be seen <lb/>that some of these images cannot be simply classified <lb/>correctly by using the block content alone, and even <lb/>humans are likely to make errors if they are asked to <lb/>classify these images. <lb/>7 CONCLUSION <lb/>From the analysis of the obtained results we can con-<lb/>clude that we can construct a very accurate classi-<lb/>fier based on run-lengths histograms alone. These <lb/>features are very easy to implement and fast to ex-<lb/>tract and thus should be part of any practical baseline <lb/>system. Interestingly, the distribution of the back-<lb/>ground runs is more important for document zone <lb/>classification than the distribution of the foreground <lb/>runs. Including a few more features based on run-<lb/>length and connected component measurements we <lb/>achieved a very competitive 1 error rate of below 1.5% <lb/>on zones extracted form the UW-III database without <lb/></body>
14
+
15
+ <note place="footnote">1 For a comparison to our results also note that at most <lb/>0.2% (53/24177) of the error rate Wang et al. present is <lb/>caused by their distinction between the text classes of dif-<lb/></note>
16
+
17
+ <body>Table 3: Contingency table showing the distribution of the classification of zones of a particular type in percent. (The total <lb/>number of errors equals 201 within 13811 tests.) The labels M, L, T, A, D, H, R, S correspond to the types math, logo, text, <lb/>table, drawing, halftone, ruling, and speckles, respectively. <lb/>M <lb/>L <lb/>T <lb/>A <lb/>D <lb/>H <lb/>R <lb/>S error [%] # samples <lb/>M 90.8 <lb/>0.0 <lb/>8.6 <lb/>0.0 <lb/>0.0 <lb/>0.6 <lb/>0.0 <lb/>0.0 <lb/>9.2 <lb/>476 <lb/>L <lb/>9.1 27.3 36.4 <lb/>0.0 <lb/>9.1 <lb/>9.1 <lb/>0.0 <lb/>9.1 <lb/>72.7 <lb/>11 <lb/>T <lb/>0.1 <lb/>0.0 99.8 <lb/>0.0 <lb/>0.0 <lb/>0.0 <lb/>0.0 <lb/>0.0 <lb/>0.2 <lb/>10450 <lb/>A <lb/>0.8 <lb/>0.0 20.7 68.6 <lb/>9.9 <lb/>0.8 <lb/>0.0 <lb/>0.0 <lb/>31.4 <lb/>121 <lb/>1.5 <lb/>0.3 <lb/>3.0 <lb/>5.5 86.0 <lb/>3.5 <lb/>0.0 <lb/>0.3 <lb/>14.0 <lb/>401 <lb/>H <lb/>0.0 <lb/>0.9 <lb/>0.0 <lb/>0.0 <lb/>9.7 86.7 <lb/>0.9 <lb/>1.8 <lb/>13.3 <lb/>113 <lb/>R <lb/>0.4 <lb/>0.0 <lb/>1.3 <lb/>0.0 <lb/>0.4 <lb/>0.0 96.1 <lb/>2.2 <lb/>3.9 <lb/>232 <lb/>S <lb/>0.1 <lb/>0.0 <lb/>0.5 <lb/>0.0 <lb/>0.1 <lb/>0.1 <lb/>0.0 99.4 <lb/>0.6 <lb/>2007 <lb/>the need for features based on glyphs or the Fourier <lb/>transform. By employing a fast logistic (log-linear) <lb/>classifier trained using the maximum entropy crite-<lb/>rion on these features, we arrived at a fast and ac-<lb/>curate, yet easy to implement overall classifier with <lb/>a slightly higher error rate of 2.1%. In our experi-<lb/>ments we did not use context information as done in <lb/>(Wang et al., 2006) and thus could keep the decision <lb/>rule very simple. However, context models are likely <lb/>to help in the overall classification and an inclusion <lb/>of our approach into Wang et al.&apos;s context model is <lb/>possible. Examining the errors made by the system <lb/>makes it seem likely that further improvements sig-<lb/>nificantly below the reached error rate may be difficult <lb/>to achieve without a significantly increased effort, for <lb/>example by using a dedicated sub-classifier to distin-<lb/>guish between text and table zones. <lb/></body>
18
+
19
+ <div type="acknowledgement">ACKNOWLEDGEMENTS <lb/>We wish to thank Oleg Nagaitsev for help with the im-<lb/>plementation and Thomas Deselaers for making avail-<lb/>able the open source image retrieval system FIRE, <lb/>which provided us with the implementation of some <lb/>of the features used. This work was partially funded <lb/>by the BMBF (German Federal Ministry of Education <lb/>and Research), project IPeT (01 IW D03). <lb/></div>
20
+
21
+ <listBibl>REFERENCES <lb/>Deselaers, T., Keysers, D., and Ney, H. (2004). Features for <lb/>image retrieval: A quantitative comparison. In DAGM <lb/>2004, Pattern Recognition, 26th DAGM Symposium, <lb/>volume 3175 of Lecture Notes in Computer Science, <lb/>pages 228-236, Tübingen, Germany. <lb/>Guyon, I., Haralick, R. M., Hull, J. J., and Phillips, I. T. <lb/>(1997). Data sets for OCR and document image un-<lb/></listBibl>
22
+
23
+ <note place="footnote">ferent font-sizes and the class &apos;other&apos; with the remaining <lb/>classes. On the other hand, we add a new class &apos;speckles&apos;, <lb/>which is related to 0.15% (21/13811) error. <lb/></note>
24
+
25
+ <listBibl>derstanding research. In Bunke, H. and Wang, P., <lb/>editors, Handbook of character recognition and doc-<lb/>ument image analysis, pages 779-799. World Scien-<lb/>tific, Singapore. <lb/>Inglis, S. and Witten, I. (1995). Document zone classifica-<lb/>tion using machine learning. In Proc Digital Image <lb/>Computing: Techniques and Applications, pages 631-<lb/>636, Brisbane, Australia. <lb/>Keysers, D., Och, F.-J., and Ney, H. (2002). Maximum en-<lb/>tropy and Gaussian models for image object recogni-<lb/>tion. In Pattern Recognition, 24th DAGM Symposium, <lb/>volume 2449 of Lecture Notes in Computer Science, <lb/>pages 498-506, Zürich, Switzerland. Springer. <lb/>Kise, K., Sato, A., and Iwata, M. (1998). Segmentation of <lb/>page images using the area Voronoi diagram. Com-<lb/>puter Vision and Image Understanding, 70(3):370-<lb/>382. <lb/>Liang, J., Phillips, I., Ha, J., and Haralick, R. (1996). Doc-<lb/>ument zone classification using the sizes of connected <lb/>components. In Proc. SPIE, volume 2660, Document <lb/>Recognition III, pages 150-157, San Jose, CA. <lb/>Okun, O., Doermann, D., and Pietikainen, M. (1999). Page <lb/>Segmentation and Zone Classification: The State of <lb/>the Art. Technical Report LAMP-TR-036, CAR-TR-<lb/>927, CS-TR-4079, University of Maryland, College <lb/>Park. <lb/>Sivaramakrishnan, R., Phillips, I. T., Ha, J., Subramanium, <lb/>S., and Haralick, R. M. (1995). Zone classification in <lb/>a document using the method of feature vector genera-<lb/>tion. In ICDAR &apos;95: Proceedings of the Third Interna-<lb/>tional Conference on Document Analysis and Recog-<lb/>nition (Volume 2), page 541ff. <lb/>Wang, Y., Haralick, R., and Phillips, I. (2000). Improve-<lb/>ment of zone content classification by using back-<lb/>ground analysis. In Fourth IAPR International Work-<lb/>shop on Document Analysis Systems (DAS2000). <lb/>Wang, Y., Phillips, I. T., and Haralick, R. M. (2006). Doc-<lb/>ument zone content classification and its performance <lb/>evaluation. Pattern Recognition, 39:57-73. <lb/>Wong, K. Y., Casey, R. G., and Wahl, F. M. (1982). Doc-<lb/>ument analysis system. IBM Journal of Research and <lb/>Development, 26(6):647-656. <lb/></listBibl>
26
+
27
+ <body>Missclassified <lb/>Nearest <lb/>Missclassified <lb/>Nearest <lb/>image <lb/>neighbor <lb/>image <lb/>neighbor <lb/>math <lb/>text <lb/>ruling <lb/>speckles <lb/>math <lb/>text <lb/>logo <lb/>speckles <lb/>logo <lb/>halftone <lb/>table <lb/>text <lb/>text <lb/>math <lb/>text <lb/>speckles <lb/>table <lb/>drawing <lb/>drawing <lb/>table <lb/>drawing <lb/>halftone <lb/>drawing <lb/>halftone <lb/>drawing <lb/>speckles <lb/>halftone <lb/>drawing <lb/>table <lb/>text <lb/>speckles <lb/>text <lb/>ruling <lb/>speckles <lb/>speckles <lb/>text <lb/>Figure 3: Examples of misclassifications showing the misclassified image and its nearest neighbor from a different class. <lb/></body>
28
+
29
+ View publication stats <lb/>View publication stats
30
+
31
+
32
+ </text>
33
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/Duguet_12279.training.segmentation.tei.xml ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei>
3
+ <teiHeader>
4
+ <fileDesc xml:id="__Duguet_12279"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front> Toward a computational and experimental model of a poly-epoxy <lb/>surface <lb/> Thomas Duguet * , Camille Bessaguet, Maëlenn Aufray, Jérôme Esvan, Cédric Charvillat, <lb/>Constantin Vahlas, Corinne Lacaze-Dufaure <lb/> CIRIMAT, CNRS – Université de Toulouse, 4, allée Emile Monso, BP-44362, 31030 Toulouse Cedex 4, France <lb/>Keywords: <lb/> Epoxy <lb/>Amine <lb/>Surface science <lb/>AFM <lb/>XPS <lb/>DFT <lb/> a b s t r a c t <lb/> A model poly-epoxy surface formed by the reaction of DGEBA and EDA is studied by the combination <lb/>of experiments and DFT calculations. A special synthesis protocol is presented leading to the formation <lb/>of a surface that is smooth (S a &lt; 1 nm), chemically homogeneous, and that presents a low-defect density <lb/>(0.21 񮽙m −2 ), as shown by AFM characterizations. Then, XPS is used for the determination of the elemental <lb/>and functional groups&apos; surface composition. DFT allows the identification and assignment of individual <lb/>bonds contributions to the experimental 1s core-level peaks. Overall, we demonstrate that such a model <lb/>sample is perfectly suitable for a use as a template for the study of poly-epoxy surface functionalization. <lb/> </front>
8
+
9
+ <body>1. Introduction <lb/> Poly-epoxy polymers are widely implemented in three families <lb/>of applications: adhesives, paints, and composite materials [1]. The <lb/>latters, such as epoxy/C fibers composites are increasingly found in <lb/>a wealth of devices and parts in the fields of leisure (skis, rackets, <lb/>boats, golf clubs, etc.), or transports, aeronautics and space (cars, <lb/>aircrafts, satellites, etc.), to name but a few. These composite mate-<lb/>rials possess stiffness and Young&apos;s modulus that compare well with <lb/>metallic alloys but with a much lower chemical reactivity and den-<lb/>sity. Therefore, they allow mass reduction and a large increase of <lb/>parts durability. <lb/>Replacement of metallic or ceramic parts by polymers often <lb/>requires surface functionalization in order to acquire optical, elec-<lb/>trical, magnetic, biomedical, esthetic, or chemical properties. The <lb/>main drawback when it comes to coat or to graft the surface of <lb/>polymer-based composites comes from the very low surface energy <lb/>of such materials once polymerized. This leads to a poor wett-<lb/>ability rendering painting or gluing difficult, and resulting in poor <lb/>adhesion. The surface energy of poly-ether ether ketone (PEEK) <lb/>or poly-epoxy is approximately 40–50 mJ/m 2 to be compared to <lb/> </body>
10
+
11
+ <front>* Corresponding author. Tel.: +33 05 34 32 34 39. <lb/> E-mail address: [email protected] (T. Duguet). <lb/></front>
12
+
13
+ <body>approximately 500 mJ/m 2 for aluminum. Moreover, the polar com-<lb/>ponent (due to H bonding) is as low as 6–7 mJ/m 2 which inhibits <lb/>the use of simple functionalization protocols [2–4]. Hence, a large <lb/>number of particular protocols has been described or patented, <lb/>where the increase of reactivity and roughness is sought. A selec-<lb/>tion amongst the wealth of publications can be found in Refs. <lb/>[5–16]. <lb/>Such protocols or methods that have been used until now <lb/>remain empirical despite the resulting improvement of the targeted <lb/>properties and/or the extension of the durability of the mate-<lb/>rial. Therefore, the need exists to access the basic mechanisms <lb/>which control the surface functionalization of polymers and to con-<lb/>trol them so as to achieve satisfactory functional properties and <lb/>adhesion. By subscribing in this perspective, our approach aims <lb/>at describing the nucleation and growth of metallic thin films on <lb/>polymer surfaces, by using an integrated method where all the <lb/>elementary mechanisms are taken into account. The first step in <lb/>this frame – object of the present study – is to obtain a model <lb/>of the polymer surface, both experimental and theoretical, at the <lb/>atomic/molecular level. Such a model will serve as a template <lb/>for further surface treatments, including pretreatments, molecu-<lb/>lar grafting, or application of films and coatings. It is worth noting <lb/>that, to the authors&apos; knowledge, no such a theoretical surface model <lb/>exists, most likely because of structural disorder and a lack of exper-<lb/>imental inputs. <lb/></body>
14
+
15
+ <front> http://dx.doi.org/10.1016/j.apsusc.2014.10.096 <lb/></front>
16
+
17
+ <body>Regarding our objectives, specifications of such an experimental <lb/>model polymer surface include: <lb/> • A 100% polymerization after curing to be comparable with calcu-<lb/>lations, where total polymerization is assumed. <lb/> • A low surface arithmetic roughness, namely R a &lt; 1 nm to make <lb/>sure that we can observe nano-islands or nano-clusters of a given <lb/>thin film. Otherwise, they would be hindered by roughness. <lb/> • A very low defect density to avoid heterogeneous nucleation at <lb/>defects. <lb/> • Chemical homogeneity to make sure that calculation models <lb/>where homogeneity is assumed are representative of the tracked <lb/>chemical reactivity. Also to make sure that chemical composition <lb/>is independent on the analyzed surface area corresponding to a <lb/>given probe size. <lb/>Our experimental approach is based on the method described <lb/>in [17,18] for forming model poly-epoxy surfaces. It consists <lb/>in the polymerization of the poly-epoxy in an Ar gloves box <lb/>at ambient temperature for at least 24 h, followed by a post-<lb/>curing at elevated temperature (polymer-dependent). Gu et al. [17] <lb/>synthesize samples from a stoichiometric mixture of DGEBA + 1,3-<lb/>di(aminomethyl)-cyclohexane, with a small amount of toluene for <lb/>decreasing viscosity and favoring an homogeneous stirring (7 min). <lb/>Samples are then stored for 24 h at ambient temperature, and post-<lb/>cured for 2 h at 130 • C in an air furnace. Characterizations of the <lb/>free surfaces are performed by atomic force microscopy (AFM) in <lb/>Tapping ® mode. Surface roughness and phase contrast are deter-<lb/>mined. It is shown that samples synthesized in an Ar glove box show <lb/>a lower surface roughness than those prepared in ambient condi-<lb/>tions, and that they are homogeneous in composition. Kansow et <lb/>al. [18] use a similar method with the aim of characterizing the for-<lb/>mation of Al, Cu, Ag, and Au films by physical vapour deposition. <lb/>DGEBA reacts with diethylene triamine in low excess at 55 • C under <lb/>controlled atmosphere, before it is left for 48 h at ambient temper-<lb/>ature. At this step, polymerization rate is about 75%. Completion is <lb/>achieved by post curing for 1 h at 120 • C. Surface roughness is about <lb/>1 nm. <lb/>Theoretically, our greatest challenge is to circumvent the <lb/>description of the disordered/amorphous structure and to limit <lb/>the number of atoms. To that end, we start with a small macro-<lb/>molecule made from the reaction of bisphenol A diglycidyl ether <lb/>(DGEBA) with ethylenediamine (EDA) (61 atoms). Even for this <lb/>moderately complex system, the analysis of the experimental <lb/>core-level XPS spectrum is not trivial and can lead to incorrect <lb/>conclusions. The help of accurate theoretical tools is thus needed <lb/>and density-functional theory (DFT) is usually used for computing <lb/>XPS core-level shifts in the case of small organic or inorganic sys-<lb/>tems. The application of this theoretical method to large systems, <lb/>e.g. polymers, is a challenge but it is established that experimental <lb/>spectra are directly related to the electronic states obtained from <lb/>calculations on smaller model molecules. For instance, Endo et al. <lb/>presented a comprehensive analysis of the XPS C 1s spectra for poly-<lb/>mers using the negative of the energy of molecular orbitals [19,20]. <lb/>More recently, they used the &apos;transition state&apos; theory [21] for the <lb/>calculation of the core electron binding energies [22,23]. Follow-<lb/>ing this work and in a first approach, we compute the molecular <lb/>orbitals energies on model molecules as preliminary input for the <lb/>assignment of experimental XPS spectra of the investigated poly-<lb/>mer. <lb/>We complement these results in the different DGEBA + EDA sys-<lb/>tem by implementing a more detailed description of surfaces by <lb/>AFM and XPS characterizations complemented by DFT calculations. <lb/>The paper is organized as follows. Experimental and computational <lb/>details are given in Section 2, followed by results in Section 3. Con-<lb/>clusions and perspectives are presented in Section 4. <lb/>
18
+
19
+ 2. Experimental and computational details <lb/> 2.1. Synthesis <lb/> We use a stoichiometric mixture of DGEBA (DER 332, Dow <lb/>Chemicals, n = 0.03) and EDA (analytical grade, purity &gt; 99.5%, <lb/>Sigma Aldrich). The mass of DGEBA (m DGEBA ) is fixed to 5 g. The <lb/>mass of EDA m EDA is thus determined following Eq. (1). <lb/> m DAE = <lb/> f DGEBA <lb/> f DAE <lb/> × <lb/> M DAE × m DGEBA <lb/> M DGEBA <lb/> = 0.43 g <lb/>(1) <lb/>where M DGEBA is the molar mass (348.52 g/mol) of this DGEBA <lb/>and f DGEBA is its functionality (2), and M EDA is the molar mass <lb/>(60.10 g/mol) and f DAE is the functionality (4) of the EDA. We assume <lb/>that no etherification occurs. <lb/>The mixture is then mechanically stirred (in an Ar glove box <lb/>when specified) for 7 min before it is poured into different molds <lb/>or deposited as a thin droplet on aluminum foil. Polymerization is <lb/>then allowed for 48 h at ambient temperature, followed by a post <lb/>curing of 2 h at 140 • C. For roughness comparison, we consider the <lb/>following poly-epoxy surfaces formed: <lb/>-At free surfaces, surfaces ref. either epoxy Air or epoxy Argon . <lb/>-At the interface with a 1 cm × 1 cm × 0.2 cm silicone mold, itself <lb/>molded on a Si wafer for transferring atomic flatness. Interfaces <lb/>ref. SiO Si /epoxy Air or SiO Si /epoxy Argon . <lb/>-At the interface with a 1 cm × 1 cm × 0.2 cm silicone mold, itself <lb/>molded on polystyrene (PS). Interfaces ref. SiO PS /epoxy Air or <lb/>SiO PS /epoxy Argon . <lb/>-By mechanical polishing up to a ¼ 񮽙m with diamond paste. Sur-<lb/>faces ref. polished Air . <lb/>Interfaces formed in the same molds but in air or Ar show <lb/>different roughnesses (shown hereafter). This is the reason <lb/>why SiO Si /epoxy Air and SiO Si /epoxy Argon , and SiO PS /epoxy Air and <lb/>SiO PS /epoxy Argon are differentiated. <lb/> 2.2. Bulk characterizations <lb/> Differential scanning calorimetry (DSC) is used for the determi-<lb/>nation of the glass transition temperature (T g ) of the poly-epoxy <lb/>under investigation. We use a DSC 204 Phoenix Series (NETZSCH) <lb/>coupled with a TASC 414/4 controller. The apparatus is calibrated <lb/>against melting temperatures of In, Hg, Sn, Bi, and Zn, applying a <lb/>+10 • /min temperature ramp. Samples are placed in aluminum cap-<lb/>sules. Mass is measured with an accuracy of ±0.1 mg. We choose to <lb/>report the onset T g-onset temperature. <lb/>Fourier transform infrared spectroscopy, FTIR (Frontier, <lb/>PerkinElmer equipped with a NIR TGS detector), is performed in <lb/>transmission in the 4000–8000 cm −1 range. 16 scans are collected <lb/>for each analysis with a resolution of 4 cm −1 . We monitor the <lb/>characteristic epoxy band (combination band of the –CH 2 of <lb/>the epoxy group) at 4530 cm −1 with increasing polymerization <lb/>time, and after post curing treatment. The reference band is the <lb/>combination band of C C with aromatic CH at 4623 cm −1 [24]. <lb/>Peak areas are then used for calculating the conversion rate (Xe NIR ) <lb/>of epoxy groups, following Eq. (2). <lb/> Xe NIR = 1 − <lb/> 񮽙 <lb/> A epoxy /A reference <lb/> 񮽙 <lb/> t=t <lb/> 񮽙 <lb/> A epoxy /A reference <lb/> 񮽙 <lb/> t=0 <lb/> (2) <lb/>where A epoxy and A reference are the peak areas of the epoxy and <lb/>reference groups, respectively. <lb/> Fig. 1. Model dimer (1 DGEBA + 1 EDA). <lb/> 2.3. Surface characterizations <lb/> Surface roughness and viscoelastic homogeneity are deter-<lb/>mined by AFM (Agilent Technologies model 5500) in ambient <lb/>conditions. The former is performed in contact mode with tips of <lb/>spring constant k approx. 0.292 N/m, whereas the latter is per-<lb/>formed in Tapping ® mode with tips of k = 25–75 N/m (AppNano). <lb/>Scanning rate is 2 񮽙m/s. Images are processed with the softwares <lb/>Gwyddion version 2.19 [25] and Pico Image (Agilent Technologies). <lb/>Surface roughness parameters follow the Geometric Product Spec-<lb/>ifications ISO 25178. S a is the arithmetic roughness, S q is the root <lb/>mean square roughness, and S z is the total roughness (maximum <lb/>peak-to-valley), determined by processing the AFM images. <lb/>XPS analysis is performed using a Thermoelectron Kalpha appa-<lb/>ratus. Photoemission spectra are recorded using Al-K񮽙 radiation <lb/>(h񮽙 = 1486.6 eV) from a monochromatized source. The X-ray spot <lb/>diameter on the sample surface is 400 񮽙m. The pass energy is <lb/>fixed at 30 eV for narrow scan and 170 eV for survey scans. The <lb/>spectrometer energy calibration was performed using the Au 4f 7/2 <lb/> (83.9 ± 0.1 eV) and Cu 2p 3/2 (932.8 ± 0.1 eV) photoelectron lines. <lb/>The background signal is removed using the Shirley method. Atomic <lb/>concentrations are determined from photoelectron peak areas <lb/>using the atomic sensitivity factors reported by Scofield [26] and <lb/>taking into account the transmission function of the analyzer. This <lb/>function was determined at different pass energies from Ag 3d and <lb/>Ag MNN peaks collected on a silver standard. Finally, photoelectron <lb/>peaks are analyzed and deconvoluted using a Lorentzian/Gaussian <lb/>(L/G = 30) peak fitting. <lb/> 2.4. Calculations <lb/> We used the model molecule shown in Fig. 1 that results of the <lb/>addition of one DGEBA and one EDA molecule. <lb/>The geometry of the model molecule was optimized at the <lb/>B3LYP/6-31G* level of theory using the Gaussian 03 software <lb/>package [27]. <lb/>
20
+
21
+ 3. Results <lb/>
22
+
23
+ Bulk characterizations are performed on samples polymerized <lb/>under ambient conditions. DSC is used for the determination of <lb/> T g-onset . Temperature ramps are doubled for each sample in order <lb/>to ensure that there is no physical aging and to verify that polymer-<lb/>ization is complete. For all samples T g-onset = 113 ± 1 • C. We assume <lb/>that T g-onset is not different after polymerization in the Ar glove box <lb/>(no bulk characterization for these samples). <lb/>We then monitor the polymerization rate with reaction duration <lb/>by following the gradual decrease of the epoxy peak area by FTIR, <lb/>and calculating the conversion rate using Eq. (2). Results are shown <lb/>in Fig. 2. <lb/>Experiments are performed from 15 min to 11520 min (8 days) <lb/>after mixing of the reactants. The conversion rate increases slowly <lb/>in the first hours and reaches an asymptote between 24 and <lb/>48 h. The maximum conversion at ambient temperature is 84% for <lb/> t ≥ 48 h. The only mean for achieving a complete polymerization <lb/>is to set the sample at a temperature above its glass transi-<lb/>tion. The post curing treatment (140 • C, 2 h) leads to a complete <lb/> Fig. 2. Epoxy group conversion rate as a function of polymerization over an 8-day <lb/>period of time. Dashed line indicates that polymerization is complete after post <lb/>curing at 140 • C for 2 h. <lb/> polymerization (&gt;98%, taking into account the FTIR spectrometer <lb/>sensitivity) illustrated by the dashed line in Fig. 2. <lb/>The different surfaces that we consider are then characterized by <lb/>AFM over 3 񮽙m × 3 񮽙m surface area images in order to determine <lb/>roughness parameters. Results are summarized in Fig. 3. <lb/>Roughness of the free surfaces is reduced by three orders of <lb/>magnitude when polymerization is performed in the Ar glove box. <lb/>Under Ar, S a and S q do not exceed 1.5 nm, except for sample <lb/>SiO Si /epoxy Argon , for which these two values are 4.9 nm and 6.8 nm, <lb/>respectively. The latter is not acceptable for the AFM observation <lb/>of metallic nanoislands or clusters that we target, in the range <lb/>of 1–20 nm in diameter [18]. In order to transfer atomic flatness <lb/>to the molds, and then to the SiO Si /epoxy surfaces, we mold sili-<lb/>cone molds against Si wafer or against PS. In these conditions, the <lb/>lowest roughness is again obtained when the surfaces are formed <lb/>under Ar atmosphere, and is similar between Si and PS processes. <lb/>Somehow, atmosphere also plays a role regarding roughness at the <lb/>substrate/polymer interface. However, a roughness as low as at that <lb/>of the free epoxy Argon surface is not achieved, indicating that mold-<lb/>ing in these conditions is not well suited for our purpose. Finally, the <lb/>roughness parameters of the polished surface are quite low but AFM <lb/>images show many scratches where nucleation may preferentially <lb/>occur. Since we want to avoid heterogeneous nucleation in order <lb/>to compare nucleation with adsorption energies at the molecular <lb/>level, polishing is abandoned. <lb/> Fig. 3. Roughness parameters determined by image processing on 3 񮽙m × 3 񮽙m sur-<lb/>faces characterized by AFM in contact mode. A polynomial of degree 2 is used in <lb/>order to correct image curvature. <lb/> Fig. 4. AFM images of the epoxy Air (a and b) and epoxy Argon (c and d) surfaces. Left column shows topographic images after a polynomial of degree 2 correction, and right <lb/>column shows deflection images (or phase contrast). <lb/> Fig. 4 shows a selection of AFM images of the epoxy Air (a and <lb/>b) and epoxy Argon (c and d) surfaces, obtained in Tapping ® mode. <lb/>Right column (Fig. 4a and c) corresponds to the surface topography <lb/>and left column (Fig. 4b and d) to the deflection of the cantilever, i.e. <lb/>to the phase contrast. Whereas Tapping ® mode leads to different <lb/>apparent values of roughness compared to contact mode, rough-<lb/>ness is again lower on the epoxy Argon surface, as can be noticed on <lb/>the contrast scale, on the right-hand side of the images. However, <lb/>both surfaces are quite flat and exhibit a very low phase contrast. <lb/>The measure of phase contrast probes the local viscoelastic prop-<lb/>erties that we assume to be an indication of chemical homogeneity <lb/>in the nanometer range. Finally, Fig. 4c and d is chosen on purpose <lb/>in order to illustrate the presence of defects, in the form of approx. <lb/>50 nm-in-diameter troughs. The density shown in Fig. 4 is not rep-<lb/>resentative (overestimated). A thorough count over a total 90 񮽙m 2 <lb/> surface area gives a defect density equal to 0.21 񮽙m −2 . <lb/>Epoxy Argon is selected as the best candidate for an experimen-<lb/>tal model surface of poly-epoxy. Thus, we investigate its surface <lb/>chemical composition by XPS and use the output of DFT calcula-<lb/>tions for peak identifications and binding energy assignments. A <lb/>first observation is made on free surfaces of samples synthesized in <lb/>silicone molds (i.e. that were not in contact with the mold). Survey <lb/>spectra show a strong Si 2p contribution at 101.8 ± 0.1 eV, which is <lb/>characteristic of siloxane groups [28]. It represents a large amount <lb/>of adsorbed silicone on the surface (approx. 8 at.%). Consequently, <lb/>epoxy Ar samples are now synthesized on Al foil (and silicone is <lb/>banished from the glove box). The significant thickness of the poly-<lb/>epoxy coupons (1 mm) ensures that Al does not diffuse up to the <lb/>free surface, since the measured interphases do not exceed 300 񮽙m <lb/> [29]. <lb/>The XPS survey spectrum of epoxy Argon surfaces polymerized <lb/>on aluminum foil show neither Si nor other elements than the one <lb/>expected in the polymer or from adsorbed molecules from the air. <lb/>Atomic composition of the surface is determined by 1s peaks fitting, <lb/>repeated at different x–y coordinates on the sample surface. We <lb/>determine the following surface composition: <lb/>81.5 at.% C, 1.8 at.% N, and 16.7 at.% O <lb/>
24
+
25
+ The result is slightly different from the bulk composition of <lb/>the poly-epoxy, where the basic motif is made of 2 DGEBA <lb/>(2 × 21 C + 2 × 4 O atoms) molecules for 1 EDA (2 C + 2 N atoms) <lb/>molecule, resulting in a bulk composition of:81.5 at . % C, 3.7 at . % N, <lb/>and 14.8 at . % O. Whereas the composition of the surface shows a <lb/>similar carbon content, it is richer in oxygen and poorer in nitro-<lb/>gen than the bulk. This is an indication of a mild surface oxidation <lb/>that may occur in the course of post-curing, when polymerization <lb/>is not yet complete (post-curing starts at 85% polymerization rate). <lb/>It is questioning though that the carbon content is apparently not <lb/>affected as well. <lb/>In order to further investigate the surface chemistry of the model <lb/>poly-epoxy surface, molecular orbitals extracted directly from DFT <lb/>results are studied. Table 1 shows the binding energies of 1s elec-<lb/>trons involved in the different bonds of the model dimer. The dimer <lb/>is made of 1 DGEBA and 1 EDA that virtually bonded through 1 <lb/>epoxy/1 amine proton reaction. Therefore, there are a few discrep-<lb/>ancies between the experimental fully-polymerized samples and <lb/>the model dimer. They are enlightened by the gray coloring of the <lb/>lines corresponding to secondary and primary amines (all should be <lb/>ternary) and to the epoxy group (no more epoxy rings in the 100% <lb/>polymerized sample). The binding energies shown are the nega-<lb/>tive value of the molecular orbitals energies. Therefore, absolute <lb/>values are not correct because (i) XPS binding energies correspond <lb/>to a multi-step process where photoelectrons interact with the cre-<lb/>ated holes, with the matrix and with their image before and after <lb/>extraction into vacuum, (ii) temperature is not considered, (iii) of <lb/>the limitation of Kohn–Sham orbital energies as reflecting initial <lb/>state effects [30]. Nevertheless, chemical shifts can be used if one <lb/>consider the latter processes constant in a given energy domain. <lb/>A minimum mean chemical shift of 0.2 eV is technically observ-<lb/>able with our XPS apparatus. Therefore, we discriminate phenyl <lb/>groups from CH 3 groups, and C OH &amp; part of the C O C bonds <lb/>from the other C O C bonds. Thanks to the support of DFT results, <lb/>we use 5 contributions to the C 1s peak deconvolution and 2 con-<lb/>tributions to the O 1s peak deconvolution. The fine fitting of the C <lb/>1s and O 1s spectra are shown in Fig. 5. N 1s spectrum is not shown <lb/>because it exhibits only one contribution for C N bonds centered <lb/> Table 1 <lb/> Molecular orbitals involving O, N, and C 1s atomic orbitals from DFT calculations on the model DGEBA–EDA dimer. Corresponding electronic binding energies ((−1) × orbital <lb/>energy), and mean chemical shifts for the given bond. Grayed cells do not have a counterpart in the experimental fully-reticulated poly-epoxy. <lb/>Molecular orbital <lb/>Binding energy (Hartree) <lb/>Binding energy (eV) <lb/>Mean chemical shift (±0.1 eV) <lb/>Bond <lb/>O 1s <lb/> −19.177 <lb/> 521.8 <lb/>+0.8 <lb/>C O C <lb/> −19.170 <lb/> 521.6 <lb/>C O C <lb/> −19.165 <lb/> 521.5 <lb/>+0.6 <lb/>Epoxy <lb/> −19.145 <lb/> 520.9 <lb/>Ref. <lb/>O H <lb/>N 1s <lb/> −14.324 <lb/> 389.8 <lb/>+0.3 <lb/>Secondary amine <lb/> −14.316 <lb/> 389.5 <lb/>Ref. <lb/>Primary amine <lb/>C 1s <lb/> −10.249 <lb/> 278.9 <lb/>+2.0 <lb/>C O C <lb/> −10.249 <lb/> 278.9 <lb/>C O C <lb/> −10.246 <lb/> 278.8 <lb/>C O C <lb/> −10.244 <lb/> 278.7 <lb/>C O C <lb/> −10.239 <lb/> 278.6 <lb/>+1.8 <lb/>C O C <lb/> −10.239 <lb/> 278.6 <lb/>C O C <lb/> −10.238 <lb/> 278.6 <lb/>C OH <lb/> −10.212 <lb/> 277.9 <lb/>+1.0 <lb/>C N <lb/> −10.209 <lb/> 277.8 <lb/>C N <lb/> −10.207 <lb/> 277.7 <lb/>C N <lb/> −10.205 <lb/> 277.7 <lb/>Quaternary C C <lb/> −10.186 <lb/> 277.2 <lb/>+0.2 <lb/>Phenyl <lb/> −10.185 <lb/> 277.1 <lb/>Phenyl <lb/> −10.185 <lb/> 277.1 <lb/>Phenyl <lb/> −10.183 <lb/> 277.1 <lb/>Phenyl <lb/> −10.182 <lb/> 277.1 <lb/>Phenyl <lb/> −10.181 <lb/> 277.0 <lb/>Phenyl <lb/> −10.181 <lb/> 277.0 <lb/>Phenyl <lb/> −10.181 <lb/> 277.0 <lb/>Phenyl <lb/> −10.180 <lb/> 277.0 <lb/>Phenyl <lb/> −10.176 <lb/> 276.9 <lb/>Phenyl <lb/> −10.173 <lb/> 276.8 <lb/>0.0 <lb/>CH3 <lb/> −10.173 <lb/> 276.8 <lb/>Ref. <lb/>CH3 <lb/> Fig. 5. XPS fine spectra of C 1s and O 1s. Spectra are fitted with contributions derived <lb/>from DFT calculations on the model dimer. <lb/> at 399.2 eV. Binding energy scale of the C 1s spectrum starts with <lb/>the –CH 3 contribution fixed at 284.4 eV. Then, mean chemical shifts <lb/>extracted from the energy difference between molecular orbitals of <lb/>DFT (see Table 1) are used for higher-binding–energy contributions <lb/>(284.4 + 0.2, +1.0, +1.8, +2.0 eV). <lb/>The filled area shows the envelope of the fitting curve. There is <lb/>an excellent matching with both O and C 1s experimental spec-<lb/>tra. Again, calculations ensure that contributions are real; even <lb/>C N, for instance, which is buried in the tails of neighboring con-<lb/>tributions. In order to consolidate these results, we now discuss <lb/>fitting with regards to the functional group composition shown in <lb/>Table 2. <lb/>Experimental atomic compositions in functional groups are con-<lb/>sistent. For instance, where one O 1s orbital of the C O C bonds <lb/>shows a composition of 15.5 at.%, two C 1s orbitals of the C O C <lb/>bonds show an approximately doubled composition of 30.5 at.% <lb/>(28.0 plus the contribution of C O C at 286.2 eV of about 3.7 <lb/>(C 1s C O C, C OH) – 1.2 (O 1s C OH) = 2.5 at.%). Similarly, N <lb/>1s and C 1s compare well in terms of composition in the C N <lb/>bonds (1.8 vs. 1.3 at.%). Finally, the last column of Table 2 shows <lb/>the expected composition in functional groups in a poly-epoxy <lb/>where the DGEBA:EDA ratio equals 2:1. For instance the number <lb/>of C 1s in phenyl groups is calculated as follows: 2 DGEBA × 2 <lb/>phenyls/DGEBA × 6 C atoms = 24 C 1s. Overall, one can find 4 C 1s <lb/>in –CH 3 , 24 C 1s in phenyls, 2 C 1s in C N, 4 C 1s in C OH, 4 C <lb/>1s in C O C 286.2 eV , 4 C 1s in C O C 286.4 eV , 2 N 1s in C N, 4 O <lb/>1s in C OH, and 4 O 1s in C O C. Therefore the total number of <lb/>considered 1s orbitals is 52. We observe large discrepancies con-<lb/>cerning the phenyl bonds concentration and the oxygenated bonds <lb/>C OH and C O C concentrations, a difference that was already <lb/>mentioned when considering the elemental atomic composition. <lb/>There are two possibilities for explaining these differences: either <lb/>the surface is oxidized and oxygenated bonds contribute to the C <lb/>1s and O 1s signals at neighboring binding energies, or the polymer <lb/>is oriented in such a way that C O C bonds emerge at the surface. <lb/> Table 2 <lb/> Results of the C, O, and N 1s deconvolutions; peak binding energy: BE, chemical shift imposed after DFT results; height in counts per second: CPS; full width at half-maximum: <lb/>FWHM; peak area; scofield relative sensitivity factor: RSF; atomic fraction: at.%; and the composition expected from the model polymer with a DGEBA:EDA ratio of 2:1. <lb/>Name <lb/>Peak BE (eV) <lb/>Chemical shift (eV) <lb/>Height (CPS) <lb/>FWHM (eV) <lb/>Area (CPS eV) <lb/>Scofield RSF <lb/>At.% <lb/>2 DGEBA:EDA motif (at.%) <lb/>C 1s CH3 <lb/>284.4 <lb/>0.0 <lb/>2249.74 <lb/>1.06 <lb/>2578.62 <lb/>1 <lb/>6.5 <lb/>7.7 <lb/>C 1s phenyl <lb/>284.6 <lb/>0.2 <lb/>11580.87 <lb/>1.3 <lb/>16279.01 <lb/>1 <lb/>40.8 <lb/>46.2 <lb/>C 1s C N <lb/>285.4 <lb/>1.0 <lb/>328.14 <lb/>1.41 <lb/>500.77 <lb/>1 <lb/>1.3 <lb/>3.8 <lb/>C 1s C O C, C OH <lb/>286.2 <lb/>1.8 <lb/>1009.93 <lb/>1.36 <lb/>1488.52 <lb/>1 <lb/>3.7 <lb/>C OH:7.7 + C O C:7.7 <lb/>C 1s C O C <lb/>286.4 <lb/>2.0 <lb/>7147.92 <lb/>1.44 <lb/>11172.06 <lb/>1 <lb/>28.0 <lb/>7.7 <lb/>C 1s shake up <lb/>291.2 <lb/>n/a <lb/>349.91 <lb/>1.32 <lb/>501.28 <lb/>1 <lb/>1.3 <lb/>n/a <lb/>N 1s C N <lb/>399.2 <lb/>n/a <lb/>808.92 <lb/>1.28 <lb/>1195.36 <lb/>1.8 <lb/>1.8 <lb/>3.8 <lb/>O 1s C OH <lb/>532.0 <lb/>0.0 <lb/>669.53 <lb/>1.73 <lb/>1255.76 <lb/>2.93 <lb/>1.2 <lb/>7.7 <lb/>O 1s C O C <lb/>532.9 <lb/>0.9 <lb/>9494.3 <lb/>1.53 <lb/>15728.75 <lb/>2.93 <lb/>15.5 <lb/>7.7 <lb/>
26
+
27
+ If we assume that a mild oxidation occurred in the course of sam-<lb/>ple preparation, it may be assigned to sub-stoichiometric groups, <lb/>such as amines (1.3–1.8 vs. 3.8 at.% expected) and phenyls (40.8 vs. <lb/>46.2 at.% expected). In that case deconvolution may be improved <lb/>by substituting or implementing additional contributions that we <lb/>are not able to identify now. <lb/> 4. Conclusions <lb/> We selected an epoxy-amine system which permits its use as <lb/>both an experimental and a computational template for further sur-<lb/>face treatments. DGEBA and EDA mixed in stoichiometric ratio and <lb/>slowly polymerized (48 h) in an Ar glovebox lead to the formation <lb/>of a poly-epoxy polymerized at a rate of 85%. Total polymeriza-<lb/>tion is achieved by post-curing at 120 • C for 2 h. Such a poly-epoxy <lb/>exhibits a glass transition temperature onset of 113 ± 1 • C. Dif-<lb/>ferent substrates and atmospheres were tested and compared <lb/>in terms of surface roughness. The lowest roughness (arithmetic <lb/>roughness = 0.2 nm, peak-to-valley = 1.5 nm) is obtained at the free <lb/>surface that polymerized under Ar atmosphere. AFM observations <lb/>reveal that, in addition to the high smoothness, the defect density of <lb/>the surface is low enough to avoid defect driven undesirable nuclea-<lb/>tion. Additionally, phase contrast is almost null which indicates that <lb/>the surface is chemically homogeneous. Atomic compositions from <lb/>XPS survey spectra at different positions confirm this result. Fine <lb/>XPS spectra over C, O, and N 1s core levels are analyzed in view <lb/>of the DFT calculations results. Theoretical binding energy chemi-<lb/>cal shifts allow an excellent fitting of the experimental 1s spectra. <lb/>A limitation has been emphasized concerning the compositions in <lb/>chemical groups: the main discrepancy concerning a much larger <lb/>composition in C O C than the one theoretically expected from <lb/>the perfect polymer model. In a near future, we will dedicate our <lb/>efforts to the improvement of (i) the poly-epoxy network model by <lb/>allowing a larger number of atoms and by using molecular dynam-<lb/>ics computations to freeze the structure at given temperatures, and <lb/>(ii) of the core-level binding energies calculations using the gener-<lb/>alized transition state method [21] that allows a better treatment of <lb/>the XPS photoemission process. Finally, the perspectives for exper-<lb/>imental work will be the formation of thin metallic films and the <lb/>mechanistic description of nucleation and growth. <lb/></body>
28
+
29
+ <listBibl> References <lb/> [1] J.-P. Pascault, H. Sautereau, J. Verdu, R.J.J. Williams, Thermosetting Polymers, <lb/>CRC Press, New York, 2002. <lb/>[2] J.W. Gooch, Encyclopedic Dictionary of Polymers, Springer, New York, 2010. <lb/>[3] E.M. Petrie, Epoxy Adhesives Formulation, McGraw-Hill, New York, 2006. <lb/>[4] E.M. Petrie, Handbook of Adhesives and Sealants, McGraw-Hill, New York, 2007. <lb/>[5] M. Charbonnier, M. Romand, Polymer pretreatments for enhanced adhesion <lb/>of metals deposited by the electroless process, Int. J. Adhes. Adhes. 23 (2003) <lb/>277–285. <lb/>[6] T.H. Baum, D.C. Miller, T.R. O&apos;Toole, Photoselective catalysis of electroless cop-<lb/>per solutions for the formation of adherent copper films onto polyimide, Chem. <lb/>Mater. 3 (1991) 714–720. <lb/>[7] S. Yoon, H.-J. Choi, J.-K. Yang, H.-H. Park, Comparative study between poly(p-<lb/>phenylenevinylene) (PPV) and PPV/SiO2 nano-composite for interface with <lb/>aluminum electrode, Appl. Surf. Sci. 237 (2004) 451–456. <lb/>[8] T.-G. Woo, I.-S. Park, K.-W. Seol, Effects of various metal seed layers on the <lb/>surface morphology and structural composition of the electroplated copper <lb/>layer, Met. Mater. Int. 15 (2009) 293–297. <lb/>[9] T. Duguet, F. Senocq, L. Laffont, C. Vahlas, Metallization of polymer composites <lb/>by metalorganic chemical vapor deposition of Cu: surface functionalization <lb/>driven films characteristics, Surf. Coat. Technol. 230 (2013) 254–259. <lb/>[10] J. Ge, M.P.K. Turunen, J.K. Kivilahti, Surface modification and characterization <lb/>of photodefinable epoxy/copper systems, Thin Solid Films 440 (2003) 198–207. <lb/>[11] C. Weiss, H. Muenstedt, Surface modification of polyether ether ketone (peek) <lb/>films for flexible printed circuit boards, J. Adhes. 78 (2002) 507–519. <lb/>[12] L. Di, B. Liu, J. Song, D. Shan, D.-A. Yang, Effect of chemical etching on the Cu/Ni <lb/>metallization of poly (ether ether ketone)/carbon fiber composites, Appl. Surf. <lb/>Sci. 257 (2011) 4272–4277. <lb/>[13] S. Siau, A. Vervaet, E. Schacht, A. Van Calster, Influence of chemical pre-<lb/>treatment of epoxy polymers on the adhesion strength of electrochemically <lb/>deposited cu for use in electronic interconnections, J. Electrochem. Soc. 151 <lb/>(2004) C133–C141. <lb/>[14] P.B. Messersmith, J. Dalsin, L. Lin, B.P. Lee, H. K., Polymeric compositions and <lb/>related method of use, WO 2005/118831 A2, Northwestern University (USA), <lb/>2005. <lb/>[15] M. Kemell, E. Färm, M. Ritala, M. Leskelä, Surface modification of thermoplastics <lb/>by atomic layer deposition of Al2O3 and TiO2 thin films, Eur. Polym. J. 44 (2008) <lb/>3564–3570. <lb/>[16] G. Cui, D. Wu, Y. Zhao, W. Liu, Z. Wu, Formation of conductive and reflective <lb/>silver nanolayers on plastic films via ion doping and solid–liquid interfacial <lb/>reduction at ambient temperature, Acta Mater. 61 (2013) 4080–4090. <lb/>[17] X. Gu, T. Nguyen, M. Oudina, D. Martin, B. Kidah, J. Jasmin, A. Rezig, L. Sung, <lb/>E. Byrd, J. Martin, D. Ho, Y.C. Jean, Microstructure and morphology of amine-<lb/>cured epoxy coatings before and after outdoor exposures—an AFM study, J. <lb/>Coat. Technol. Res. 2 (2005) 547–556. <lb/>[18] J. Kanzow, P.S. Horn, M. Kirschmann, V. Zaporojtchenko, K. Dolgner, F. Faupel, <lb/>C. Wehlack, W. Possart, Formation of a metal/epoxy resin interface, Appl. Surf. <lb/>Sci. 239 (2005) 227–236. <lb/>[19] K. Endo, S. Maeda, M. Aida, Simulation of C 1s spectra of C-and O-containing <lb/>polymers in XPS by ab initio MO calculations using model oligomers, Polym. J. <lb/>29 (1997) 171–181. <lb/>[20] K. Endo, S. Maeda, Y. Kaneda, Analysis of C 1s spectra of N-, O-, and X-containing <lb/>polymers in X-ray photoelectron spectroscopy by ab initio molecular orbital <lb/>calculations using model molecules, Polym. J. 29 (1997) 255–260. <lb/>[21] J.C. Slater, K.H. Johnson, Self-consistent-field X˛ cluster method for polyatomic <lb/>molecules and solids, Phys. Rev. B 5 (1972) 844–853. <lb/>[22] K. Takaoka, T. Otsuka, K. Naka, A. Niwa, T. Suzuki, C. Bureau, S. Maeda, K. Hyodo, <lb/>K. Endo, D.P. Chong, Analysis of X-ray photoelectron spectra of electrochemi-<lb/>cally prepared polyaniline by DFT calculations using model molecules, J. Mol. <lb/>Struct. 608 (2002) 175–182. <lb/>[23] T. Otsuka, K. Endo, M. Suhara, D.P. Chong, Theoretical X-ray photoelectron spec-<lb/>tra of polymers by deMon DFT calculations using the model dimers, J. Mol. <lb/>Struct. 522 (2000) 47–60. <lb/>[24] C.E. Miller, Near-infrared spectroscopy of synthetic polymers, Appl. Spectrosc. <lb/>Rev. 26 (1991) 277–339. <lb/>[25] D. Nečas, P. Klapetek, Gwyddion: an open-source software for SPM data anal-<lb/>ysis, Centr. Eur. J. Phys. 10 (2012) 181–188. <lb/>[26] J.H. Scofield, Hartree-Slater subshell photoionization cross-sections at 1254 <lb/>and 1487 eV, J. Electron Spectrosc. Relat. Phenom. 8 (1976) 129–137. <lb/>[27] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, <lb/>J.A. Montgomery, J.T. Vreven, K.N. Kudin, J.C. Burant, J.M. Millam, S.S. Iyengar, J. <lb/>Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A. Petersson, <lb/>H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. <lb/>Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox, H.P. Hratchian, <lb/>J.B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, J. <lb/>Austin, R. Cammi, C. Pomelli, J.W. Ochterski, P.Y. Ayala, K. Morokuma, G.A. <lb/>Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels, <lb/>M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, <lb/>J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A. <lb/>Liashenko, P. Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, <lb/>C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, <lb/> M.W. Wong, C. Gonzalez, J.A. Pople, Gaussian 03, Revision B05, Gaussian, Inc., <lb/>Pittsburg, PA, 2003. <lb/>[28] L.-A. O&apos;Hare, B. Parbhoo, S.R. Leadley, Development of a methodology for XPS <lb/>curve-fitting of the Si 2p core level of siloxane materials, Surf. Interface Anal. <lb/>36 (2004) 1427–1434. <lb/>[29] M. Aufray, A. André Roche, Epoxy–amine/metal interphases: influences <lb/>from sharp needle-like crystal formation, Int. J. Adhes. Adhes. 27 (2007) <lb/>387–393. <lb/>[30] P.S. Bagus, E.S. Ilton, C.J. Nelin, The interpretation of XPS spectra: insights into <lb/>materials properties, Surf. Sci. Rep. 68 (2013) 273–304. </listBibl>
30
+
31
+
32
+ </text>
33
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/E14-1007.training.segmentation.tei.xml ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei>
3
+ <teiHeader>
4
+ <fileDesc xml:id="__E14-1007"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, pages 58–67, <lb/>Gothenburg, Sweden, April 26-30 2014. c <lb/> 2014 Association for Computational Linguistics <lb/></front>
8
+
9
+ <front>Inducing Example-based Semantic Frames <lb/>from a Massive Amount of Verb Uses <lb/> Daisuke Kawahara † Daniel W. Peterson ‡ Octavian Popescu § Martha Palmer ‡ <lb/> † <lb/> Kyoto University, Kyoto, Japan <lb/> ‡ <lb/> University of Colorado at Boulder, Boulder, CO, USA <lb/> § <lb/> Fondazione Bruno Kessler, Trento, Italy <lb/> [email protected], {Daniel.W.Peterson, Martha.Palmer}@colorado.edu, [email protected] <lb/> Abstract <lb/> We present an unsupervised method for in-<lb/> ducing semantic frames from verb uses in <lb/>giga-word corpora. Our semantic frames <lb/>are verb-specific example-based frames <lb/>that are distinguished according to their <lb/>senses. We use the Chinese Restau-<lb/>rant Process to automatically induce these <lb/>frames from a massive amount of verb in-<lb/>stances. In our experiments, we acquire <lb/>broad-coverage semantic frames from two <lb/>giga-word corpora, the larger comprising <lb/>20 billion words. Our experimental results <lb/>indicate the effectiveness of our approach. <lb/></front>
10
+
11
+ <body> 1 Introduction <lb/> Semantic frames are indispensable knowledge for <lb/>semantic analysis or text understanding. In the <lb/>last decade, semantic frames, such as FrameNet <lb/>(Baker et al., 1998) and PropBank (Palmer et al., <lb/>2005), have been manually elaborated. These <lb/>resources are effectively exploited in many nat-<lb/>ural language processing (NLP) tasks, includ-<lb/>ing not only semantic parsing but also ma-<lb/>chine translation (Boas, 2002), information ex-<lb/>traction (Surdeanu et al., 2003), question answer-<lb/>ing (Narayanan and Harabagiu, 2004), paraphrase <lb/>acquisition (Ellsworth and Janin, 2007) and recog-<lb/>nition of textual entailment (Burchardt and Frank, <lb/>2006). <lb/>There have been many attempts to automati-<lb/>cally acquire frame knowledge from raw corpora <lb/>with the goal of either adding frequency informa-<lb/>tion to an existing resource or of inducing simi-<lb/>lar frames for other languages. Most of these ap-<lb/>proaches, however, focus on syntactic frames, i.e., <lb/>subcategorization frames (e.g., (Manning, 1993; <lb/>Briscoe and Carroll, 1997; Korhonen et al., 2006; <lb/>Lippincott et al., 2012; Reichart and Korhonen, <lb/>2013)). Since subcategorization frames represent <lb/>argument patterns of verbs and are purely syn-<lb/>tactic, expressions that have the same subcatego-<lb/>rization frame can have different meanings (e.g., <lb/>metaphors). Semantics-oriented NLP applications <lb/>based on frames, such as paraphrase acquisition <lb/>and machine translation, require consistency in the <lb/>meaning of each frame, and thus these subcatego-<lb/>rization frames are not suitable for these semantic <lb/>tasks. <lb/>Recently, there have been a few studies on au-<lb/>tomatically acquiring semantic frames (Materna, <lb/>2012; Materna, 2013). Materna induced seman-<lb/>tic frames (called LDA-Frames) from triples of <lb/>(subject, verb, object) in the British National <lb/>Corpus (BNC) based on Latent Dirichlet Allo-<lb/>cation (LDA) and the Dirichlet Process. LDA-<lb/>Frames capture limited linguistic phenomena of <lb/>these triples, and are defined across verbs based <lb/>on probabilistic topic distributions. <lb/>This paper presents a method for automati-<lb/>cally building verb-specific semantic frames from <lb/>a large raw corpus. Our semantic frames are verb-<lb/>specific like PropBank and semantically distin-<lb/>guished. A frame has several syntactic case slots, <lb/>each of which consists of words that are eligible to <lb/>fill the slot. For example, let us show three seman-<lb/>tic frames of the verb &quot; observe &quot; : 1 <lb/> observe:1 <lb/> nsubj:{we, author, ...} dobj:{effect, result, ...} <lb/>prep in:{study, case, ...} ... <lb/> observe:2 <lb/> nsubj:{teacher, we, ...} dobj:{child, student, ...} <lb/>prep in:{classroom, school, ...} ... <lb/> observe:3 <lb/> nsubj:{child, people, ...} dobj:{bird, animal, ...} <lb/>prep at:{range, time, ...} ... <lb/>
12
+
13
+ <note place="footnote">1 In this paper, we use the dependency relation names <lb/>of the Stanford collapsed dependencies (de Marneffe et al., <lb/>2006) as the notations of case slots. For instance, &quot; nsubj &quot; <lb/>means a nominal subject, &quot; dobj &quot; means a direct object, &quot; iboj &quot; <lb/>means an indirect object, &quot; ccomp &quot; means a clausal comple-<lb/>ment and &quot; prep * &quot; means a preposition. <lb/></note>
14
+
15
+ <page> 58 <lb/></page>
16
+
17
+ Frequencies, which are not shown in the above ex-<lb/>amples, are attached to each semantic frame, case <lb/>slot and word, and can be effectively exploited for <lb/>the applications of these semantic frames. The fre-<lb/>quencies of words in each case slot become good <lb/>sources of selectional preferences. <lb/>Our novel contributions are summarized as fol-<lb/>lows: <lb/> • induction of semantic frames based on the <lb/>Chinese Restaurant Process (Aldous, 1985) <lb/>from only automatic parses of a web-scale <lb/>corpus, <lb/> • exploitation of the assumption of one sense <lb/>per collocation (Yarowsky, 1993) to make the <lb/>computation feasible, <lb/> • providing broad-coverage knowledge for se-<lb/>lectional preferences, and <lb/> • evaluating induced semantic frames by us-<lb/>ing an existing annotated corpus with verb <lb/>classes. <lb/> 2 Related Work <lb/> The most closely related work to our semantic <lb/>frames are LDA-Frames, which are probabilistic <lb/>semantic frames automatically induced from a raw <lb/>corpus (Materna, 2012; Materna, 2013). He used a <lb/>model based on LDA and the Dirichlet Process to <lb/>cluster verb instances of a triple (subject, verb, ob-<lb/>ject) to produce semantic frames and slots. Both <lb/>of these are represented as a probabilistic distri-<lb/>bution of words across verbs. He applied this <lb/>method to the BNC and acquired 427 frames and <lb/>144 slots (Materna, 2013). These frames are over-<lb/>generalized across verbs and might be difficult <lb/>to provide with fine-grained selectional prefer-<lb/>ences. In addition, Grenager and Manning (2006) <lb/>proposed a method for inducing PropBank-style <lb/>frames from Stanford typed dependencies ex-<lb/>tracted from raw corpora. Although these frames <lb/>are based on typed dependencies and more seman-<lb/>tic than subcategorization frames, they are not dis-<lb/>tinguished in terms of the senses of words filling a <lb/>case slot. <lb/>There are hand-crafted semantic frames in the <lb/>lexicons of FrameNet (Baker et al., 1998) and <lb/>PropBank (Palmer et al., 2005). Corpus Pattern <lb/>Analysis (CPA) frames (Hanks, 2012) are another <lb/>manually created repository of patterns for verbs. <lb/>Each pattern represents a prototypical word usage <lb/>as extracted by lexicographers from the BNC. Cre-<lb/>ating CPA is time consuming, but our proposed <lb/>method may be employed to assist in the creation <lb/>of this type of resource, as shown in Section 4.4. <lb/>Our task can be regarded as clustering of verb <lb/>instances. In this respect, the models of Parisien <lb/>and Stevenson are related to our method (Parisien <lb/>and Stevenson, 2009; Parisien and Stevenson, <lb/>2010). Parisien and Stevenson (2009) proposed <lb/>a Dirichlet Process model for clustering usages <lb/>of the verb &quot; get. &quot; Later, Parisien and Stevenson <lb/>(2010) proposed a Hierarchical Dirichlet Process <lb/>model for jointly clustering argument structures <lb/>
18
+
19
+ (i.e., subcategorization frames) and verb classes. <lb/>However, their argument structures are not seman-<lb/>tic but syntactic, and also they did not evaluate the <lb/>resulting frames. There have also been related ap-<lb/>proaches to clustering verb types (Vlachos et al., <lb/> 2009; Sun and Korhonen, 2009; Falk et al., 2012; <lb/>Reichart and Korhonen, 2013). These methods in-<lb/>duce verb clusters in which multiple verbs partic-<lb/>ipate, and do not consider the polysemy of verbs. <lb/>Our objective is different from theirs. <lb/>Another line of related work is unsupervised <lb/>semantic parsing or semantic role labeling (Poon <lb/>and Domingos, 2009; Lang and Lapata, 2010; <lb/>Lang and Lapata, 2011a; Lang and Lapata, 2011b; <lb/>Titov and Klementiev, 2011; Titov and Klemen-<lb/>tiev, 2012). These approaches basically clus-<lb/>ter predicates and their arguments to distinguish <lb/>predicate senses and semantic roles of arguments. <lb/>Modi et al. (2012) extended the model of Titov and <lb/>Klementiev (2012) to jointly induce semantic roles <lb/>and frames using the Chinese Restaurant Process, <lb/>which is also used in our approach. However, <lb/>they did not aim at building a lexicon of semantic <lb/>frames, but at distinguishing verbs that have dif-<lb/>ferent senses in a relatively small annotated cor-<lb/>pus. Applying this method to a large corpus could <lb/>produce a frame lexicon, but its scalability would <lb/>be a big problem. <lb/>For other languages than English, Kawahara <lb/>and Kurohashi (2006a) proposed a method for au-<lb/>tomatically compiling Japanese semantic frames <lb/>from a large web corpus. They applied con-<lb/>ventional agglomerative clustering to predicate-<lb/>argument structures using word/frame similarity <lb/>based on a manually-crafted thesaurus. Since <lb/>Japanese is head-final and has case-marking post-<lb/>positions, it seems easier to build semantic frames <lb/>with it than with other languages such as English. <lb/>They also achieved an improvement in depen-<lb/>dency parsing and predicate-argument structure <lb/>
20
+
21
+ <page>59 <lb/> </page>
22
+
23
+ analysis by using their resulting frames (Kawahara <lb/>and Kurohashi, 2006b). <lb/> 3 Method for Inducing Semantic Frames <lb/> Our objective is to automatically induce verb-<lb/>specific example-based semantic frames. Each se-<lb/>mantic frame consists of a partial set of syntactic <lb/>slots: nsubj, dobj, iobj, ccomp and prep *. Each <lb/>slot consists of words with frequencies, which <lb/>could provide broad-coverage selectional prefer-<lb/>ences. <lb/>Frames for a verb should be semantically distin-<lb/>guished. That is to say, each frame should consist <lb/>of predicate-argument structures that have consis-<lb/>tent usages or meanings. <lb/>Our procedure to automatically generate seman-<lb/>tic frames from verb usages is as follows: <lb/>
24
+
25
+ 1. apply dependency parsing to a raw corpus <lb/>and extract predicate-argument structures for <lb/>each verb from the automatic parses, <lb/>2. merge the predicate-argument structures that <lb/>have presumably the same meaning based on <lb/>the assumption of one sense per collocation <lb/>to get a set of initial frames, and <lb/>3. apply clustering to the initial frames based <lb/>on the Chinese Restaurant Process to produce <lb/>the final semantic frames. <lb/>
26
+
27
+ Each of these steps is described in the following <lb/>sections in detail. <lb/> 3.1 Extracting Predicate-argument <lb/>Structures from a Raw Corpus <lb/> We first apply dependency parsing to a large raw <lb/>corpus. We use the Stanford parser with Stanford <lb/>dependencies (de Marneffe et al., 2006). 2 Col-<lb/>lapsed dependencies are adopted to directly extract <lb/>prepositional phrases. <lb/>Then, we extract predicate-argument structures <lb/>from the dependency parses. Dependents that have <lb/>the following dependency relations to a verb are <lb/>extracted as arguments: <lb/>nsubj, xsubj, dobj, iobj, ccomp, xcomp, <lb/>prep * <lb/> Here, we do not distinguish adjuncts from argu-<lb/>ments. All extracted dependents of a verb are han-<lb/>dled as arguments. This distinction is left for fu-<lb/>ture work, but this will be performed using slot <lb/>
28
+
29
+ <note place="footnote">2 http://nlp.stanford.edu/software/lex-parser.shtml <lb/></note>
30
+
31
+ Sentences: <lb/> They observed the effects of ... <lb/>This statistical ability to observe an effect ... <lb/>We did not observe a residual effect of ... <lb/>He could observe the results at the same time ... <lb/>My first opportunity to observe the results of ... <lb/>You can observe beautiful birds ... <lb/>Children may then observe birds ... <lb/>. . . <lb/> Predicate-argument structures: <lb/> nsubj:they observe dobj:effect <lb/>observe dobj:effect <lb/>nsubj:we observe dobj:effect <lb/>nsubj:he observe dobj:result prep at:time <lb/>observe dobj:result <lb/>nsubj:you observe dobj:bird <lb/>nsubj:child observe dobj:bird <lb/>. . . <lb/> Initial frames: <lb/> nsubj:{they, we, ...} observe dobj:{effect} <lb/>nsubj:{he, ...} observe dobj:{result} prep at:{time} <lb/>nsubj:{you, child, ...} observe dobj:{bird} <lb/>. . . <lb/>Figure 1: Examples of predicate-argument struc-<lb/>tures and initial frames for the verb &quot; observe. &quot; <lb/>frequencies in the applications of semantic frames <lb/>or the method proposed by Abend and Rappoport <lb/>(2010). <lb/>We apply the following processes to extracted <lb/>predicate-argument structures: <lb/> • A verb and an argument are lemmatized, and <lb/>only the head of an argument is preserved for <lb/>compound nouns. <lb/> • Phrasal verbs are also distinguished from <lb/>non-phrasal verbs. For example, &quot; look up &quot; <lb/>has independent frames from &quot; look. &quot; <lb/> • The passive voice of a verb is distinguished <lb/>from the active voice, and thus these have in-<lb/>dependent frames. Passive voice is detected <lb/>using the part-of-speech tag &quot; VBN &quot; (past <lb/>participle). The alignment between frames of <lb/>active and passive voices will be done after <lb/>the induction of frames using the model of <lb/>Sasano et al. (2013) in the future. <lb/> • &quot; xcomp &quot; (open clausal complement) is re-<lb/>named to &quot; ccomp &quot; (clausal complement) and <lb/> &quot; xsubj &quot; (controlling subject) is renamed to <lb/> &quot; nsubj &quot; (nominal subject). This is because <lb/>
32
+
33
+ <page>60 <lb/></page>
34
+
35
+ these usages as predicate-argument structures <lb/>are not different. <lb/> • A capitalized argument with the part-of <lb/>speech &quot; NNP &quot; (singular proper noun) or <lb/> &quot; NNPS &quot; (plural proper noun) is general-<lb/>ized to ⟨name⟩. Similarly, an argument of <lb/> &quot; ccomp &quot; is generalized to ⟨comp⟩ since the <lb/>content of a clausal complement is not impor-<lb/>tant. <lb/>Extracted predicate-argument structures are <lb/>collected for each verb and the subsequent pro-<lb/>cesses are applied to the predicate-argument struc-<lb/>tures of each verb. Figure 1 shows examples of <lb/>predicate-argument structures for &quot; observe. &quot; <lb/> 3.2 Constructing Initial Frames from <lb/>Predicate-argument Structures <lb/> A straightforward way to produce semantic frames <lb/>is to cluster the extracted predicate-argument <lb/>structures directly. Since our objective is to com-<lb/>pile broad-coverage semantic frames, a massive <lb/>amount of predicate-argument structures should <lb/>be fed into the clustering. It would take prohibitive <lb/>computational costs to conduct the sampling pro-<lb/>cedure, which is described in the next section. <lb/>To make the computation feasible, we merge the <lb/>predicate-argument structures that have the same <lb/>or similar meaning to get initial frames. These ini-<lb/>tial frames are the input of the subsequent cluster-<lb/>ing process. For this merge, we assume one sense <lb/>per collocation (Yarowsky, 1993) for predicate-<lb/>argument structures. <lb/>For each predicate-argument structure of a verb, <lb/>we couple the verb and an argument to make a unit <lb/>for sense disambiguation. We select an argument <lb/>in the following order by considering the degree of <lb/>effect on the verb sense: 3 <lb/> dobj, ccomp, nsubj, prep * , iobj. <lb/>This selection of a predominant argument order <lb/>above is justified by relative comparisons of the <lb/>discriminative power of the different slots for CPA <lb/>frames (Popescu, 2013). If a predicate-argument <lb/>structure does not have any of the above slots, it is <lb/>discarded. <lb/>Then, the predicate-argument structures that <lb/>have the same verb and argument pair (slot and <lb/>
36
+
37
+ <note place="footnote">3 If a predicate-argument structure has multiple preposi-<lb/>tional phrases, one of them is randomly selected. <lb/></note>
38
+
39
+ word, e.g., &quot; dobj:effect &quot; ) are merged into an ini-<lb/>tial frame (Figure 1). After this process, we dis-<lb/>card minor initial frames that occur fewer than 10 <lb/>times. <lb/>For example, we have 732,292 instances <lb/>(predicate-argument structures) for the verb &quot; ob-<lb/>serve &quot; in the web corpus that is used in our exper-<lb/>iment (its details are described in Section 4.1). As <lb/>the result of this merging process, we obtain 6,530 <lb/>initial frames, which become an input for the clus-<lb/>tering. This means that this process accelerates the <lb/>speed of clustering more than 100 times. <lb/>The precision of this process will be evaluated <lb/>in Section 4.3. <lb/> 3.3 Clustering using Chinese Restaurant <lb/>Process <lb/> We cluster initial frames for each verb to produce <lb/>final semantic frames using the Chinese Restau-<lb/>rant Process (Aldous, 1985). We regard each ini-<lb/>tial frame as an instance in the usual clustering of <lb/>the Chinese Restaurant Process. <lb/>We calculate the posterior probability of a se-<lb/>mantic frame f j given an initial frame v i as fol-<lb/>lows: <lb/> P (f j |v i ) ∝ <lb/> { n(f j ) <lb/> N +α · P (v i |f j ) f j ̸ = new <lb/> α <lb/>N +α · P (v i |f j ) f j = new, <lb/> (1) <lb/>where N is the number of initial frames for the <lb/>target verb and n(f j ) is the current number of ini-<lb/>tial frames assigned to the semantic frame f j . α <lb/> is a hyper-parameter that determines how likely <lb/>it is for a new semantic frame to be created. In <lb/>this equation, the first term is the Dirichlet process <lb/>prior and the second term is the likelihood of v i . <lb/> P (v i |f j ) is defined based on the Dirichlet-<lb/>Multinomial distribution as follows: <lb/> P (v i |f j ) = <lb/> ∏ <lb/> w∈V <lb/> P (w|f j ) count(v i ,w) , <lb/> (2) <lb/>where V is the vocabulary in all case slots cooc-<lb/>curring with the verb. It is distinguished by <lb/>the case slot, and thus consists of pairs of slots <lb/>and words, e.g., &quot; nsubj:child &quot; and &quot; dobj:bird. &quot; <lb/> count(v i , w) is the number of w in the initial <lb/>frame v i . <lb/> P (w|f j ) is defined as follows: <lb/> P (w|f j ) = <lb/> count(f j , w) + β <lb/> ∑ <lb/> t∈V count(f j , t) + |V | · β <lb/>, (3) <lb/>
40
+
41
+ <page>61 <lb/></page>
42
+
43
+ where count(f j , w) is the current number of w in <lb/>the frame f j , and β is a hyper-parameter of Dirich-<lb/>let distribution. For a new semantic frame, this <lb/>probability is uniform (1/|V |). <lb/> We use Gibbs sampling to realize this cluster-<lb/>ing. <lb/> 4 Experiments and Evaluations <lb/> 4.1 Experimental Settings <lb/> We use two kinds of large-scale corpora: a web <lb/>corpus and the English Gigaword corpus. <lb/>To prepare a web corpus, we first crawled the <lb/>web. We extracted sentences from each web <lb/>page that seems to be written in English based <lb/>on the encoding information. Then, we selected <lb/>sentences that consist of at most 40 words, and <lb/>removed duplicated sentences. From this pro-<lb/>cess, we obtained a corpus of one billion sen-<lb/>tences, totaling approximately 20 billion words. <lb/>We focused on verbs whose frequency was more <lb/>than 1,000. There were 19,649 verbs, includ-<lb/>ing phrasal verbs, and separating passive and ac-<lb/>tive constructions. We extracted 2,032,774,982 <lb/>predicate-argument structures. <lb/>We also used the English Gigaword corpus <lb/>(LDC2011T07; English Gigaword Fifth Edition) <lb/>to induce semantic frames. This corpus consists <lb/>of approximately 180 million sentences, which to-<lb/>taling four billion words. There were 7,356 verbs <lb/>after applying the same frequency threshold as the <lb/>web corpus. We extracted 423,778,278 predicate-<lb/>argument structures from this corpus. <lb/>We set the hyper-parameters α in (1) and β in <lb/>(3) to 1.0. The frame assignments for all the com-<lb/>ponents were initialized randomly. We took 100 <lb/>samples for each initial frame and selected the <lb/>frame assignment that has the highest probability. <lb/>These parameters were determined according to a <lb/>preliminary experiment to manually examine the <lb/>quality of resulting frames. <lb/> 4.2 Experimental Results <lb/> We executed the per-verb clustering tasks on a PC <lb/>cluster. It finished within a few hours for most <lb/>verbs, but it took a couple of days for very frequent <lb/>verbs, such as &quot; get &quot; and &quot; say. &quot; The clustering pro-<lb/>duced an average number of semantic frames per <lb/>verb of 15.2 for the web corpus and 18.5 for the <lb/>Gigaword corpus. Examples of induced semantic <lb/>frames from the web corpus are shown in Table 1. <lb/> slot <lb/>
44
+
45
+ instances <lb/>nsubj <lb/>i:5850, we:5201, he:3796, you:3669, ... <lb/>dobj <lb/>what:7091, people:2272, this:2262, ... <lb/>observe:1 prep in way:254, world:204, life:194, ... <lb/>. . . <lb/>nsubj <lb/>we:11135, you:1321, i:1317, ... <lb/>dobj <lb/>change:5091, difference:2719, ... <lb/>observe:2 prep in study:622, case:382, cell:362, ... <lb/>. . . <lb/>nsubj <lb/>student:3921, i:2240, we:2174, ... <lb/>dobj <lb/>child:2323, class:2184, student:2025, ... <lb/>observe:3 prep in classroom:555, action:509, ... <lb/>. . . <lb/>nsubj <lb/>we:44833, i:6873, order:4051, ... <lb/>dobj <lb/>card:28835, payment:22569, ... <lb/>accept:1 prep for payment:1166, convenience:1147, ... <lb/>. . . <lb/>nsubj <lb/>i:10568, we:9300, you:5106, ... <lb/>dobj <lb/>that:14180, this:12061, it:7756, ... <lb/>accept:2 prep as part:1879, fact:1085, truth:926, ... <lb/>. . . <lb/>nsubj <lb/>people:7459, he:6696, we:5515, ... <lb/>dobj <lb/>christ:13766, jesus:6528, it:5612, ... <lb/>accept:3 prep as savior:5591, lord:597, one:469, ... <lb/>. . . <lb/> Table 1: Examples of resulting frames for the verb <lb/> &quot; observe &quot; and &quot; accept &quot; induced from the web cor-<lb/>pus. The number following an instance word rep-<lb/>
46
+
47
+ resents its frequency. <lb/> 4.3 Evaluation of Induced Semantic Frames <lb/> We evaluate precision and coverage of induced se-<lb/>mantic frames. To measure the precision of in-<lb/>duced semantic frames, we adopt the purity met-<lb/>ric, which is usually used to evaluate clustering re-<lb/>sults. However, the problem is that it is impossible <lb/>to assign gold-standard classes to the huge num-<lb/>ber of instances. To automatically measure the <lb/>purity of the induced semantic frames, we make <lb/>use of the SemLink corpus (Loper et al., 2007), in <lb/>which VerbNet classes (Kipper-Schuler, 2005) and <lb/>PropBank/FrameNet frames are assigned to each <lb/>instance. We make a test set that contains 157 pol-<lb/>ysemous verbs that occur 10 or more times in the <lb/>SemLink corpus (sections 02-21 of the Wall Street <lb/>Journal). We first add these instances to the in-<lb/>stances from a raw corpus and apply clustering to <lb/>these merged instances. Then, we compare the in-<lb/>duced semantic frames of the SemLink instances <lb/>with their gold-standard classes. We adopt Verb-<lb/>Net classes and PropBank frames as gold-standard <lb/>classes. <lb/>For each group of verb-specific semantic <lb/>frames, we measure the purity of the frames as the <lb/>percentage of SemLink instances belonging to the <lb/>majority gold class in their respective cluster. Let <lb/>
48
+
49
+ <page> 62 <lb/></page>
50
+
51
+ PU <lb/> CO <lb/>F 1 <lb/> Mac <lb/>Mic <lb/>Mac <lb/>Mic <lb/>Mac <lb/>Mic <lb/>against <lb/>One frame <lb/>0.799 0.802 0.917 0.952 0.854 0.870 <lb/>VerbNet <lb/>Initial frames <lb/> 0.985 0.982 0.755 0.812 0.855 0.889 <lb/>Induced sem frames 0.900 0.901 0.886 0.928 0.893 0.914 <lb/> against <lb/>One frame <lb/>0.901 0.872 <lb/> ↑ <lb/> ↑ <lb/> 0.909 0.910 <lb/>PropBank Initial frames <lb/> 0.994 0.993 <lb/> ↑ <lb/>↑ <lb/> 0.858 0.893 <lb/>Induced sem frames 0.965 0.949 <lb/> ↑ <lb/>↑ <lb/> 0.924 0.939 <lb/> Table 2: Evaluation results of semantic frames from the web corpus against VerbNet classes and Prop-<lb/>Bank frames. &quot; Mac &quot; means a macro average and &quot; Mic &quot; means a micro average. <lb/>PU <lb/>CO <lb/>F 1 <lb/> Mac <lb/>Mic <lb/>Mac <lb/>Mic <lb/>Mac <lb/>Mic <lb/>against <lb/>One frame <lb/>0.799 0.804 0.855 0.920 0.826 0.858 <lb/>VerbNet <lb/>Initial frames <lb/> 0.985 0.981 0.666 0.758 0.795 0.855 <lb/>Induced sem frames 0.916 0.909 0.796 0.880 0.852 0.894 <lb/> against <lb/>One frame <lb/>0.901 0.874 <lb/> ↑ <lb/>↑ <lb/> 0.877 0.896 <lb/>PropBank Initial frames <lb/> 0.994 0.993 <lb/> ↑ <lb/>↑ <lb/> 0.798 0.859 <lb/>Induced sem frames 0.968 0.953 <lb/> ↑ <lb/>↑ <lb/> 0.874 0.915 <lb/> Table 3: Evaluation results of semantic frames from the Gigaword corpus against VerbNet classes and <lb/>PropBank frames. &quot; Mac &quot; means a macro average and &quot; Mic &quot; means a micro average. <lb/> N denote the total number of SemLink instances <lb/>of the target verb, G j the set of instances belong-<lb/>ing to the j-th gold class and F i the set of instances <lb/>belonging to the i-th frame. The purity (PU) can <lb/>then be written as follows: <lb/>PU = <lb/> 1 <lb/> N <lb/> ∑ <lb/> i <lb/> max <lb/> j <lb/>
52
+
53
+ |G j ∩ F i |. <lb/> (4) <lb/>For example, a frame of the verb &quot; observe &quot; con-<lb/>tains 11 SemLink instances, and eight out of them <lb/>belong to the class SAY-37.7, which is the ma-<lb/>jority class among these 11 instances. PU is cal-<lb/>culated by summing up such counts over all the <lb/>frames of this verb. <lb/>Usually, inverse purity or collocation is used <lb/>to measure the recall of normal clustering tasks. <lb/>However, these recall measures do not fit our task. <lb/>This is because it is not a real error to have similar <lb/>separate frames. Instead, we want to avoid hav-<lb/>ing so many frames that we cannot provide broad-<lb/>coverage selectional preferences due to sparsity. <lb/>To judge this aspect, we measure coverage. <lb/>The coverage (CO) measures to what extent <lb/>predicate-argument structures of the target verb in <lb/>a test set are included in one of frames of the verb. <lb/>We use the predicate-argument structures of the <lb/>above 157 verbs from the SemLink corpus, which <lb/>are the same ones used in the evaluation of PU. <lb/>We judge a predicate-argument structure as cor-<lb/>rect if all of its argument words (of the target slot <lb/>described in Section 3.1) are included in the corre-<lb/>sponding slot of a frame. If the clustering gets bet-<lb/>ter, the value of CO will get higher, because merg-<lb/>ing instances by clustering alleviates data sparsity. <lb/>These per-verb scores are aggregated into an <lb/>overall score by averaging over all verbs. We use <lb/>two ways of averaging: a macro average and a mi-<lb/>cro average. The macro average is a simple av-<lb/>erage of scores for individual verbs. The micro <lb/>average is obtained by weighting the scores for in-<lb/>dividual verbs proportional to the number of in-<lb/>stances for that verb. Finally, we use the harmonic <lb/>mean (F 1 ) of purity and coverage as a single mea-<lb/>sure of clustering quality. <lb/>For comparison, we adopt the following two <lb/>baseline methods: <lb/> One frame a frame into which all the instances <lb/>for a verb are merged <lb/> Initial frames the initial frames without cluster-<lb/>ing (described in Section 3.2) <lb/>Table 2 and Table 3 list evaluation results for <lb/>semantic frames induced from the web corpus and <lb/>the Gigaword corpus, respectively. 4 Note that CO <lb/>does not consider gold-standard classes, and thus <lb/>the values of CO are the same for the VerbNet <lb/>
54
+
55
+ <note place="footnote"> 4 We did not adopt inverse purity, but its values for the <lb/>induced semantic frames range from 0.42 to 0.49. <lb/></note>
56
+
57
+ <page> 63 <lb/></page>
58
+
59
+ and PropBank evaluations. The induced frames <lb/>outperformed the two baseline methods in terms <lb/>of F 1 in most cases. While the coverage of the <lb/>web frames was higher than that of the Giga-<lb/>word frames, as expected, the purity of the web <lb/>frames was slightly lower than that of the Giga-<lb/>word frames. This degradation might be caused <lb/>by the noise in the web corpus. <lb/>The purity of the initial frames was around <lb/>98%-99%, which means that there were few cases <lb/>that the one-sense-per-collocation assumption was <lb/>violated. <lb/>Modi et al. (2012) reported a purity of 77.9% <lb/>for the assignment of FrameNet frames to the <lb/>FrameNet corpus. We also conducted the above <lb/>purity evaluation against FrameNet frames for 140 <lb/>verbs.
60
+
61
+ 5 We obtained a macro average of 92.9% <lb/>and a micro average of 89.2% for the web frames, <lb/>and a macro average of 93.2% and a micro average <lb/>of 89.8% for the Gigaword frames. It is difficult <lb/>to directly compare these results with Modi et al. <lb/>(2012), but our frame assignments seem to have <lb/>higher accuracy. <lb/> 4.4 Evaluation against CPA Frames <lb/> Corpus Pattern Analysis (CPA) is a technique for <lb/>linking word usage to prototypical syntagmatic <lb/>patterns. 6 The resource was built manually by in-<lb/>vestigating examples in the BNC, and the set of <lb/>corpus examples used to induce each pattern is <lb/>given. For example, the following three patterns <lb/>describe the usage of the verb &quot; accommodate. &quot; <lb/>[Human 1] accommodate [Human 2] <lb/>[Building] accommodate [Eventuality] <lb/>[Human] accommodate [Self] to [Eventuality] <lb/>In this paper, we use CPA to evaluate the quality <lb/>of the automatically induced frames. By compar-<lb/>ing the induced frames to CPA patterns, we can <lb/>evaluate the correctness and relevance of this ap-<lb/>proach from a human point of view. To do that, <lb/>we associate semantic features to the set of words <lb/>in each slot in the frames, using SUMO (Niles <lb/>and Pease, 2001). For example, take the follow-<lb/>ing frame for the verb &quot; accomplish &quot; : <lb/> accomplish:1 <lb/> nsubj:{you, leader, employee, ...} <lb/>dobj:{developing, progress, objective, ...}. <lb/>
62
+
63
+ <note place="footnote">5 Since FrameNet frames are not assigned to all the verbs <lb/>of SemLink, the number of verbs is different from the evalu-<lb/>ations against VerbNet and PropBank. <lb/></note>
64
+
65
+ <note place="footnote">6 http://deb.fi.muni.cz/pdev/ <lb/></note>
66
+
67
+ all <lb/>K-means <lb/>Entropy (E) <lb/>0.790 <lb/>0.516 <lb/>Recovery Rate (RC) 0.347 <lb/>0.630 <lb/>Purity (P ) <lb/>0.462 <lb/>0.696 <lb/>Table 4: CPA Evaluation. <lb/>Using SUMO, we map this frame to the following: <lb/>nsubj: [Human] <lb/>dobj: [SubjectiveAssessmentAttribute], <lb/>which corresponds to pattern 3 for &quot; accomplish &quot; <lb/>in CPA. <lb/>We also associate SUMO attributes to the CPA <lb/>patterns with more than 10 examples (716 verbs). <lb/>There are many patterns of SUMO attributes for <lb/>any CPA frame or induced frame, since each <lb/>filler word in a particular slot can have more <lb/>than one SUMO attribute. We filter out the <lb/>non-discriminative SUMO attributes following the <lb/>technique described in Popescu (2013). Using <lb/>this, we obtain SUMO attributes for both CPA <lb/>clusters and induced frames, and we can use the <lb/>standard entropy-based measures to evaluate the <lb/>match between the two types of patterns: E — en-<lb/>tropy, RC — recovery rate, and P — purity (Li et <lb/>al., 2004): <lb/> E = <lb/> K <lb/> ∑ <lb/> j=1 <lb/> m j <lb/> m <lb/> · e j , RC = 1 − <lb/> K,L <lb/> ∑ <lb/> j,i=1 <lb/> p ij <lb/> m i <lb/> , <lb/> (5) <lb/> P = <lb/> K <lb/> ∑ <lb/> j=1 <lb/> m j <lb/> m <lb/> · p j , p j = max <lb/> i <lb/> p ij , <lb/> (6) <lb/> e j = <lb/> L <lb/> ∑ <lb/> i=1 <lb/> p ij log 2 p ij , p ij = <lb/> m ij <lb/> m i <lb/> , <lb/> (7) <lb/>where m j is the number of induced frames corre-<lb/>sponding to topic j, m ij is the number of induced <lb/>frames in cluster j and annotated with the CPA <lb/>pattern i, m is the total number of induced frames, <lb/> L is the number of CPA patterns, and K is the <lb/>number of induced frames. <lb/>We also consider a K-means clustering process, <lb/>with K set as 2 or 3 depending on the number of <lb/>SUMO-attributed patterns. The K-means evalu-<lb/>ation is carried out considering only the centroid <lb/>of the cluster, which corresponds to the prototypi-<lb/>cal induced semantic frame with SUMO attributes. <lb/>We compute E, RC and P using formulae (5) -<lb/>(7) for each verb and then compute the macro av-<lb/>erage, considering all the frames and only the K-<lb/>means centroids, respectively. The results for the <lb/>induced web frames are displayed in Table 4. <lb/>
68
+
69
+ <page> 64 <lb/></page>
70
+
71
+ The evaluation method presented here over-<lb/>comes some of the drawbacks of the previous ap-<lb/>proaches (Materna, 2012; Materna, 2013). First, <lb/>we did not limit the evaluation to the most frequent <lb/>patterns. Second, the mapping was carried out au-<lb/>tomatically and not by hand. The results above <lb/>compare favorably with the previous approaches, <lb/>especially considering that no filtering procedures <lb/>were applied to the induced frames. We anticipate <lb/>that the results based on the prototypical induced <lb/>frames with SUMO attributes would be competi-<lb/>tive. Our post-analysis revealed that the entropy <lb/>can be lowered further if an automatic filtering <lb/>based on frequencies is applied. <lb/> 4.5 Evaluation of the Quality of Selectional <lb/>Preferences <lb/> We also investigated the quality of selectional <lb/>preferences within the induced semantic frames. <lb/>The only publicly available test data for selectional <lb/>preferences, to our knowledge, is from Chambers <lb/>and Jurafsky (2010). This data consists of quadru-<lb/>ples (verb, relation, word, confounder) and does <lb/>not contain their context.
72
+
73
+ 7 <lb/> A typical way for using our semantic frames is <lb/>to select an appropriate frame for an input sen-<lb/>tence and judge the eligibility of the word uses <lb/>against the selected frame. However, due to the <lb/>lack of context for the above data, it is difficult to <lb/>select a corresponding semantic frame for a test <lb/>quadruple and thus the induced semantic frames <lb/>cannot be naturally applied to this data. To in-<lb/>vestigate the potential for selectional preferences <lb/>of the semantic frames, we approximately match <lb/>a quadruple with each of the semantic frames of <lb/>the verb and select the frame that has the highest <lb/>probability as follows: <lb/> P (w) = max <lb/> i <lb/> P (w|v, rel, f i ), <lb/> (8) <lb/>where w is the word or confounder, v is the verb, <lb/> rel is the relation and f i is a semantic frame. By <lb/>comparing the probabilities of the word and the <lb/>confounder, we select either of them according to <lb/>the higher probability. For tie breaking in the case <lb/>that no frames are found for the verb or both the <lb/>word and confounder are not found in the case slot, <lb/>we randomly select either of them in the same way <lb/>as Chambers and Jurafsky (2010). <lb/>We use the &quot; neighbor frequency &quot; set, which is <lb/>the most difficult among the three sets included <lb/>
74
+
75
+ <note place="footnote"> 7 A document ID of the English Gigaword corpus is avail-<lb/>able, but it is difficult to recover the context of each instance <lb/>from this information. <lb/></note>
76
+
77
+ in the data. It contains 6,767 quadruples and the <lb/>relations consist of three classes: subject, object <lb/>and preposition, which has no distinction of ac-<lb/>tual prepositions. To link these relations with our <lb/>case slots, we manually aligned the subject with <lb/>the nsubj (nominal subject) slot, the object with <lb/>the dobj (direct object) slot and the preposition <lb/>with prep * (all the prepositions) slots. For the <lb/>preposition relation, we choose the highest prob-<lb/>ability among all the preposition slots in a frame. <lb/>To match the generalized ⟨name⟩ with the word in <lb/>a quadruple, we change the word to ⟨name⟩ if it is <lb/>capitalized and not a capitalized personal pronoun. <lb/>Our semantic frames from the Gigaword corpus <lb/>achieved an accuracy of 81.7% 8 and those from <lb/>the web corpus achieved an accuracy of 80.2%. <lb/>This slight deterioration seems to come from the <lb/>noise in the web corpus. The best performance <lb/>in Chambers and Jurafsky (2010) is 81.7% on <lb/>this &quot; neighbor frequency &quot; set, which was achieved <lb/>by conditional probabilities with the Erk (2007)&apos;s <lb/>smoothing method calculated from the English Gi-<lb/>gaword corpus. Our approach for selectional pref-<lb/>erences does not use smoothing like Erk (2007), <lb/>but it achieved equivalent performance to the pre-<lb/>vious work. If we applied our semantic frames to a <lb/>verb instance with its context, a more precise judg-<lb/>ment of selectional preferences would be possible <lb/>with appropriate frame selection. <lb/> 5 Conclusion <lb/> This paper has described an unsupervised method <lb/>for inducing semantic frames from instances of <lb/>each verb in giga-word corpora. This method is <lb/>clustering based on the Chinese Restaurant Pro-<lb/>cess. The resulting frame data are open to the pub-<lb/>lic and also can be searched by inputting a verb via <lb/>our web interface. 9 <lb/> As applications of the resulting frames, we plan <lb/>to integrate them into syntactic parsing, semantic <lb/>role labeling and verb sense disambiguation. For <lb/>instance, Kawahara and Kurohashi (2006b) im-<lb/>proved accuracy of dependency parsing based on <lb/>Japanese semantic frames automatically induced <lb/>from a large raw corpus. It is valuable and promis-<lb/>ing to apply our semantic frames to these NLP <lb/>tasks. <lb/>
78
+
79
+ <note place="footnote">8 Since the dataset was created from the NYT 2001 portion <lb/>of the English Gigaword Corpus, we built semantic frames <lb/>again from the Gigaword corpus except this part. <lb/></note>
80
+
81
+ <note place="footnote">9 http://nlp.ist.i.kyoto-u.ac.jp/member/kawahara/cf/crp.en/ <lb/></note>
82
+
83
+ <page>65 <lb/></page>
84
+
85
+ </body>
86
+
87
+ <back>
88
+
89
+ <div type="acknowledgement">Acknowledgments <lb/> This work was supported by Kyoto University <lb/>John Mung Program and JST CREST. We grate-<lb/>fully acknowledge the support of the National Sci-<lb/>ence Foundation Grant NSF 1116782 -RI: Small: <lb/>A Bayesian Approach to Dynamic Lexical Re-<lb/>sources for Flexible Language Processing. Any <lb/>opinions, findings, and conclusions or recommen-<lb/>dations expressed in this material are those of the <lb/>authors and do not necessarily reflect the views of <lb/>the National Science Foundation. <lb/></div>
90
+
91
+ <listBibl>References <lb/>
92
+
93
+ Omri Abend and Ari Rappoport. 2010. Fully unsuper-<lb/>vised core-adjunct argument classification. In Pro-<lb/> ceedings of the 48th Annual Meeting of the Associa-<lb/>tion for Computational Linguistics, pages 226–236. <lb/>David Aldous. 1985. Exchangeability and related top-<lb/>ics. ´ <lb/>Ecole d&apos; ´ <lb/>Eté de Probabilités de Saint-Flour XIII <lb/> 1983, pages 1–198. <lb/>Collin Baker, Charles J. Fillmore, and John Lowe. <lb/>1998. The Berkeley FrameNet Project. In Pro-<lb/>ceedings of the 36th Annual Meeting of the Associ-<lb/>ation for Computational Linguistics and 17th Inter-<lb/>national Conference on Computational Linguistics, <lb/> pages 86–90. <lb/>Hans C. Boas. 2002. Bilingual framenet dictionaries <lb/>for machine translation. In Proceedings of the 3rd <lb/>International Conference on Language Resources <lb/>and Evaluation, pages 1364–1371. <lb/>Ted Briscoe and John Carroll. 1997. Automatic ex-<lb/>traction of subcategorization from corpora. In Pro-<lb/>ceedings of the 5th Conference on Applied Natural <lb/>Language Processing, pages 356–363. <lb/>Aljoscha Burchardt and Anette Frank. 2006. Approx-<lb/>imating textual entailment with LFG and FrameNet <lb/>frames. In Proceedings of the 2nd PASCAL Recog-<lb/>nizing Textual Entailment Workshop, pages 92–97. <lb/>Nathanael Chambers and Daniel Jurafsky. 2010. Im-<lb/>proving the use of pseudo-words for evaluating se-<lb/>lectional preferences. In Proceedings of the 48th <lb/>Annual Meeting of the Association for Computa-<lb/>tional Linguistics, pages 445–453. <lb/>Marie-Catherine de Marneffe, Bill MacCartney, and <lb/>Christopher D. Manning. 2006. Generating typed <lb/>dependency parses from phrase structure parses. In <lb/> Proceedings of the 5th International Conference on <lb/>Language Resources and Evaluation, pages 449– <lb/>454. <lb/>Michael Ellsworth and Adam Janin. 2007. Mu-<lb/>taphrase: Paraphrasing with framenet. In Proceed-<lb/>ings of the ACL-PASCAL Workshop on Textual En-<lb/>tailment and Paraphrasing, pages 143–150. <lb/>Katrin Erk. 2007. A simple, similarity-based model <lb/>for selectional preferences. In Proceedings of the <lb/>45th Annual Meeting of the Association of Compu-<lb/>tational Linguistics, pages 216–223. <lb/>Ingrid Falk, Claire Gardent, and Jean-Charles Lamirel. <lb/>2012. Classifying french verbs using french and en-<lb/>glish lexical resources. In Proceedings of the 50th <lb/>Annual Meeting of the Association for Computa-<lb/>tional Linguistics, pages 854–863. <lb/>Trond Grenager and Christopher D. Manning. 2006. <lb/>Unsupervised discovery of a statistical verb lexicon. <lb/>In Proceedings of the 2006 Conference on Empirical <lb/>Methods in Natural Language Processing, pages 1– <lb/>8. <lb/>Patrick Hanks. 2012. How people use words to make <lb/>meanings: Semantic types meet valencies. Input, <lb/>Process and Product: Developments in Teaching <lb/>and Language Corpora, pages 54–69. <lb/>Daisuke Kawahara and Sadao Kurohashi. 2006a. <lb/>Case frame compilation from the web using high-<lb/>performance computing. In Proceedings of the 5th <lb/>International Conference on Language Resources <lb/>and Evaluation, pages 1344–1347. <lb/>Daisuke Kawahara and Sadao Kurohashi. 2006b. A <lb/>fully-lexicalized probabilistic model for Japanese <lb/>syntactic and case structure analysis. In Proceedings <lb/>of the Human Language Technology Conference of <lb/>the NAACL, pages 176–183. <lb/>Karin Kipper-Schuler. 2005. VerbNet: A Broad-<lb/>Coverage, Comprehensive Verb Lexicon. Ph.D. the-<lb/>sis, University of Pennsylvania. <lb/>Anna Korhonen, Yuval Krymolowski, and Ted Briscoe. <lb/>2006. A large subcategorization lexicon for natural <lb/>language processing applications. In Proceedings of <lb/>the 5th International Conference on Language Re-<lb/>sources and Evaluation, pages 345–352. <lb/>Joel Lang and Mirella Lapata. 2010. Unsuper-<lb/>vised induction of semantic roles. In Human Lan-<lb/>guage Technologies: The 2010 Annual Conference <lb/>of the North American Chapter of the Association <lb/>for Computational Linguistics, pages 939–947. <lb/>Joel Lang and Mirella Lapata. 2011a. Unsupervised <lb/>semantic role induction via split-merge clustering. <lb/>In Proceedings of the 49th Annual Meeting of the <lb/>Association for Computational Linguistics: Human <lb/>Language Technologies, pages 1117–1126. <lb/>Joel Lang and Mirella Lapata. 2011b. Unsupervised <lb/>semantic role induction with graph partitioning. In <lb/> Proceedings of the 2011 Conference on Empirical <lb/>Methods in Natural Language Processing, pages <lb/>1320–1331. <lb/>Tao Li, Sheng Ma, and Mitsunori Ogihara. 2004. <lb/>Entropy-based criterion in categorical clustering. In <lb/> Proceedings of the 21st International Conference on <lb/>Machine Learning, volume 4, pages 536–543. <lb/>
94
+
95
+ 66 <lb/>
96
+
97
+ Thomas Lippincott, Anna Korhonen, and Diarmuid <lb/>O Séaghdha. 2012. Learning syntactic verb frames <lb/>using graphical models. In Proceedings of the 50th <lb/>Annual Meeting of the Association for Computa-<lb/>tional Linguistics, pages 420–429. <lb/>Edward Loper, Szu-Ting Yi, and Martha Palmer. 2007. <lb/>Combining lexical resources: mapping between <lb/>PropBank and VerbNet. In Proceedings of the 7th <lb/>International Workshop on Computational Linguis-<lb/>tics. <lb/> Christopher Manning. 1993. Automatic acquisition <lb/>of a large subcategorization dictionary from corpora. <lb/>In Proceedings of the 31st Annual Meeting of the As-<lb/>sociation for Computational Linguistics, pages 235– <lb/>242. <lb/>Jiří Materna. 2012. LDA-Frames: An unsupervised <lb/>approach to generating semantic frames. In Alexan-<lb/>der Gelbukh, editor, Proceedings of the 13th Inter-<lb/>national Conference CICLing 2012, Part I, volume <lb/>7181 of Lecture Notes in Computer Science, pages <lb/>376–387. Springer Berlin / Heidelberg. <lb/>Jiří Materna. 2013. Parameter estimation for LDA-<lb/>Frames. In Proceedings of the 2013 Conference of <lb/>the North American Chapter of the Association for <lb/>Computational Linguistics: Human Language Tech-<lb/>nologies, pages 482–486. <lb/>Ashutosh Modi, Ivan Titov, and Alexandre Klementiev. <lb/>2012. Unsupervised induction of frame-semantic <lb/>representations. In Proceedings of the NAACL-HLT <lb/>Workshop on the Induction of Linguistic Structure, <lb/> pages 1–7. <lb/>Srini Narayanan and Sanda Harabagiu. 2004. Ques-<lb/>tion answering based on semantic structures. In <lb/> Proceedings of the 20th International Conference on <lb/>Computational Linguistics, pages 693–701. <lb/>Ian Niles and Adam Pease. 2001. Towards a standard <lb/>upper ontology. In Proceedings of the International <lb/>Conference on Formal Ontology in Information Sys-<lb/>tems, pages 2–9. <lb/>Martha Palmer, Daniel Gildea, and Paul Kingsbury. <lb/>2005. The proposition bank: An annotated cor-<lb/>pus of semantic roles. Computational Linguistics, <lb/> 31(1):71–106. <lb/>Christopher Parisien and Suzanne Stevenson. 2009. <lb/>Modelling the acquisition of verb polysemy in chil-<lb/>dren. In Proceedings of the CogSci2009 Workshop <lb/>on Distributional Semantics beyond Concrete Con-<lb/>cepts, pages 17–22. <lb/>Christopher Parisien and Suzanne Stevenson. 2010. <lb/>Learning verb alternations in a usage-based <lb/>Bayesian model. In Proceedings of the 32nd annual <lb/>meeting of the Cognitive Science Society. <lb/> Hoifung Poon and Pedro Domingos. 2009. Unsuper-<lb/>vised semantic parsing. In Proceedings of the 2009 <lb/>Conference on Empirical Methods in Natural Lan-<lb/>guage Processing, pages 1–10. <lb/>Octavian Popescu. 2013. Learning corpus patterns us-<lb/>ing finite state automata. In Proceedings of the 10th <lb/>International Conference on Computational Seman-<lb/>tics, pages 191–203. <lb/>Roi Reichart and Anna Korhonen. 2013. Improved <lb/>lexical acquisition through DPP-based verb cluster-<lb/>ing. In Proceedings of the 51st Annual Meeting <lb/>of the Association for Computational Linguistics, <lb/> pages 862–872. <lb/>Ryohei Sasano, Daisuke Kawahara, Sadao Kurohashi, <lb/>and Manabu Okumura. 2013. Automatic knowl-<lb/>edge acquisition for case alternation between the <lb/>passive and active voices in Japanese. In Proceed-<lb/>ings of the 2013 Conference on Empirical Methods <lb/>in Natural Language Processing, pages 1213–1223. <lb/>Lin Sun and Anna Korhonen. 2009. Improving verb <lb/>clustering with automatically acquired selectional <lb/>preferences. In Proceedings of the 2009 Confer-<lb/>ence on Empirical Methods in Natural Language <lb/>Processing, pages 638–647. <lb/>Mihai Surdeanu, Sanda Harabagiu, John Williams, and <lb/>Paul Aarseth. 2003. Using predicate-argument <lb/>structures for information extraction. In Proceed-<lb/>ings of the 41st Annual Meeting of the Association <lb/>for Computational Linguistics, pages 8–15. <lb/>Ivan Titov and Alexandre Klementiev. 2011. A <lb/>Bayesian model for unsupervised semantic parsing. <lb/>In Proceedings of the 49th Annual Meeting of the <lb/>Association for Computational Linguistics: Human <lb/>Language Technologies, pages 1445–1455. <lb/>Ivan Titov and Alexandre Klementiev. 2012. A <lb/>Bayesian approach to unsupervised semantic role in-<lb/>duction. In Proceedings of the 13th Conference of <lb/>the European Chapter of the Association for Com-<lb/>putational Linguistics, pages 12–22. <lb/>Andreas Vlachos, Anna Korhonen, and Zoubin <lb/>Ghahramani. 2009. Unsupervised and constrained <lb/>dirichlet process mixture models for verb cluster-<lb/>ing. In Proceedings of the Workshop on Geomet-<lb/>rical Models of Natural Language Semantics, pages <lb/>74–82. <lb/>David Yarowsky. 1993. One sense per collocation. In <lb/> Proceedings of the Workshop on Human Language <lb/>Technology, pages 266–271. <lb/></listBibl>
98
+
99
+ <page> 67 </page>
100
+
101
+ </back>
102
+ </text>
103
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/E14-1075.training.segmentation.tei.xml ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei>
3
+ <teiHeader>
4
+ <fileDesc xml:id="__E14-1075"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, pages 712–721, <lb/>Gothenburg, Sweden, April 26-30 2014. c <lb/> 2014 Association for Computational Linguistics <lb/></front>
8
+
9
+ <front>Improving the Estimation of Word Importance for News Multi-Document <lb/>Summarization <lb/> Kai Hong <lb/> University of Pennsylvania <lb/>Philadelphia, PA, 19104 <lb/> [email protected] <lb/> Ani Nenkova <lb/> University of Pennsylvania <lb/>Philadelphia, PA, 19104 <lb/> [email protected] <lb/> Abstract <lb/> We introduce a supervised model <lb/>for predicting word importance that <lb/>incorporates a rich set of features. Our <lb/>model is superior to prior approaches <lb/>for identifying words used in human <lb/>summaries. <lb/>Moreover we show <lb/>that an extractive summarizer using <lb/>these estimates of word importance is <lb/>comparable in automatic evaluation with <lb/>the state-of-the-art. <lb/></front>
10
+
11
+ <body>1 Introduction <lb/> In automatic extractive summarization, sentence <lb/>importance is calculated by taking into account, <lb/>among possibly other features, the importance <lb/>of words that appear in the sentence. In this <lb/>paper, we describe experiments on identifying <lb/>words from the input that are also included in <lb/>human summaries; we call such words summary <lb/> keywords. <lb/> We review several unsupervised <lb/>approaches for summary keyword identification <lb/>and further combine these, along with features <lb/>including position, part-of-speech, subjectivity, <lb/>topic categories, context and intrinsic importance, <lb/>in a superior supervised model for predicting word <lb/>importance. <lb/>One of the novel features we develop aims <lb/>to determine the intrinsic importance of words. <lb/>To this end, we analyze abstract-article pairs in <lb/>the New York Times corpus (Sandhaus, 2008) <lb/>to identify words that tend to be preserved in <lb/>the abstracts. We demonstrate that judging word <lb/>importance just based on this criterion leads to <lb/>significantly higher performance than selecting <lb/>sentences at random. Identifying intrinsically <lb/>important words allows us to generate summaries <lb/>without doing any feature computation on the <lb/>input, equivalent in quality to the standard baseline <lb/>of extracting the first 100 words from the latest <lb/>article in the input. Finally, we integrate the <lb/>schemes for assignment of word importance into <lb/>a summarizer which greedily optimizes for the <lb/>presence of important words. We show that our <lb/>better estimation of word importance leads to <lb/>better extractive summaries. <lb/> 2 Prior work <lb/> The idea of identifying words that are descriptive <lb/>of the input can be dated back to Luhn&apos;s earliest <lb/>work in automatic summarization (Luhn, 1958). <lb/>There keywords were identified based on the <lb/>number of times they appeared in the input, <lb/>and words that appeared most and least often <lb/>were excluded. Then the sentences in which <lb/>keywords appeared near each other, presumably <lb/>better conveying the relationship between the <lb/>keywords, were selected to form a summary. <lb/>Many successful recent systems also estimate <lb/>word importance. The simplest but competitive <lb/>way to do this task is to estimate the word <lb/>probability from the input (Nenkova and <lb/>Vanderwende, 2005). Another powerful method <lb/>is log-likelihood ratio test (Lin and Hovy, 2000), <lb/>which identifies the set of words that appear in <lb/>the input more often than in a background corpus <lb/>(Conroy et al., 2006; Harabagiu and Lacatusu, <lb/>2005). <lb/>In contrast to selecting a set of keywords, <lb/>weights are assigned to all words in the input <lb/>in the majority of summarization methods. <lb/>Approaches based on (approximately) optimizing <lb/>the coverage of these words have become widely <lb/>popular. Earliest such work relied on TF*IDF <lb/>weights (Filatova and Hatzivassiloglou, 2004), <lb/>later approaches included heuristics to identify <lb/>summary-worthy bigrams (Riedhammer et al., <lb/>2010). Most optimization approaches, however, <lb/>use TF*IDF or word probability in the input as <lb/>word weights (McDonald, 2007; Shen and Li, <lb/>2010; Berg-Kirkpatrick et al., 2011). <lb/>
12
+
13
+ <page> 712 <lb/></page>
14
+
15
+ Word weights have also been estimated by <lb/>supervised approaches, with word probability and <lb/>location of occurrence as typical features (Yih et <lb/>al., 2007; Takamura and Okumura, 2009; Sipos et <lb/>al., 2012). <lb/>A handful of investigations have productively <lb/>explored the mutually reinforcing relationship <lb/>between word and sentence importance, iteratively <lb/>re-estimating each in either supervised or <lb/>unsupervised framework (Zha, 2002; Wan et <lb/>al., 2007; Wei et al., 2008; Liu et al., 2011). <lb/>Most existing work directly focuses on predicting <lb/>sentence importance, with emphasis on the <lb/>formalization of the problem (Kupiec et al., 1995; <lb/>Celikyilmaz and Hakkani-Tur, 2010; Litvak et al., <lb/>2010). There has been little work directly focused <lb/>on predicting keywords from the input that will <lb/>appear in human summaries. Also there has been <lb/>only a few investigations of suitable features <lb/>for estimating word importance and identifying <lb/>keywords in summaries; we address this issue by <lb/>exploring a range of possible indicators of word <lb/>importance in our model. <lb/> 3 Data and Planned Experiments <lb/> We carry out our experiments on two datasets from <lb/>the Document Understanding Conference (DUC) <lb/>(Over et al., 2007). DUC 2003 is used for training <lb/>and development, DUC 2004 is used for testing. <lb/>These are the last two years in which generic <lb/>summarization was evaluated at DUC workshops. <lb/>There are 30 multi-document clusters in DUC <lb/>2003 and 50 in DUC 2004, each with about 10 <lb/>news articles on a related topic. The task is <lb/>to produce a 100-word generic summary. Four <lb/>human abstractive summaries are available for <lb/>each cluster. <lb/>We compare different keyword extraction <lb/>methods by the F-measure 1 they achieve against <lb/>the gold-standard summary keywords. We do not <lb/>use stemming when calculating these scores. <lb/>In our work, keywords for an input are defined <lb/>as those words that appear in at least i of the <lb/>human abstracts, yielding four gold-standard sets <lb/>of keywords, denoted by G i . |G i | is thus the <lb/>cardinality of the set for the input. We only <lb/>consider the words in the summary that also <lb/>appear in the original input 2 , with stopwords <lb/>
16
+
17
+ <note place="footnote">1 2*precision*recall/(precision+recall) <lb/></note>
18
+
19
+ <note place="footnote">2 On average 26.3% (15.0% with stemming) of the words <lb/>in the four abstracts never appear in the input.<lb/></note>
20
+
21
+ excluded 3 . Table 1 shows the average number of <lb/>unique content words for the respective keyword <lb/>gold-standard. <lb/>i <lb/>1 <lb/>2 <lb/>3 4 <lb/>Mean |G i | 102 32 15 6 <lb/>Table 1: Average number of words in G i <lb/> For the summarization task, we compare results <lb/>using ROUGE (Lin, 2004). We report ROUGE-1, <lb/>-2, -4 recall, with stemming and without removing <lb/>stopwords. We consider ROUGE-2 recall as <lb/>the main metric for this comparison due to its <lb/>effectiveness in comparing machine summaries <lb/>(Owczarzak et al., 2012). All of the summaries <lb/>were truncated to the first 100 words by ROUGE 4 . <lb/>We use Wilcoxon signed-rank test to examine <lb/>the statistical significance as advocated by Rankel <lb/>et al. (2011) for both tasks, and consider <lb/>differences to be significant if the p-value is less <lb/>than 0.05. <lb/> 4 Unsupervised Word Weighting <lb/> In this section we describe three unsupervised <lb/>approaches of assigning importance weights to <lb/>words. <lb/>The first two are probability and <lb/>log-likelihood ratio, which have been extensively <lb/>used in prior work. We also apply a markov <lb/>random walk model for keyword ranking, similar <lb/>to Mihalcea and Tarau (2004). In the next <lb/>section we describe a summarizer that uses these <lb/>weights to form a summary and then describe <lb/>our regression approach to combine these and <lb/>other predictors in order to achieve more accurate <lb/>predictions for the word importance in Section 7. <lb/>The task is to assign a score to each word in the <lb/>input. The keywords extracted are thus the content <lb/>words with highest scores. <lb/> 4.1 Word Probability (Prob) <lb/> The frequency with which a word occurs in the <lb/>input is often considered as an indicator of its <lb/>importance. The weight for a word is computed <lb/>as p(w) = c(w) <lb/>N , where c(w) is the number of <lb/>times word w appears in the input and N is the <lb/>total number of word tokens in the input. <lb/>
22
+
23
+ <note place="footnote">3 We use the stopword list from the SMART system <lb/>(Salton, 1971), augmented with punctuation and symbols. <lb/></note>
24
+
25
+ <note place="footnote">4 ROUGE version 1.5.5 with parameters: -c 95 -r 1000 -n <lb/>4 -m -a -l 100 -x <lb/></note>
26
+
27
+ <page> 713 <lb/></page>
28
+
29
+ 4.2 Log-likelihood Ratio (LLR) <lb/> The log-likelihood ratio test (Lin and Hovy, 2000) <lb/>compares the distribution of a word in the input <lb/>with that in a large background corpus to identify <lb/>topic words. We use the Gigaword corpus (Graff et <lb/>al., 2007) for background counts. The test statistic <lb/>has a χ 2 distribution, so a desired confidence level <lb/>can be chosen to find a small set of topic words. <lb/> 4.3 Markov Random Walk Model (MRW) <lb/> Graph methods have been successfully applied to <lb/>weighting sentences for generic (Wan and Yang, <lb/>2008; Mihalcea and Tarau, 2004; Erkan and <lb/>Radev, 2004) and query-focused summarization <lb/>(Otterbacher et al., 2009). <lb/>Here instead of constructing a graph with <lb/>sentences as nodes and edges weighted by <lb/>sentence similarity, we treat the words as vertices, <lb/>similar to Mihalcea and Tarau (2004). The <lb/>difference in our approach is that the edges <lb/>between the words are defined by syntactic <lb/>dependencies rather than depending on the <lb/>co-occurrence of words within a window of k. We <lb/>use the Stanford dependency parser (Marneffe et <lb/>al., 2006). In our approach, we consider a word <lb/> w more likely to be included in a human summary <lb/>when it is syntactically related to other (important) <lb/>words, even if w itself is not mentioned often. <lb/>The edge weight between two vertices is equal to <lb/>the number of syntactic dependencies of any type <lb/>between two words within the same sentence in <lb/>the input. The weights are then normalized by <lb/>summing up the weights of edges linked to one <lb/>node. <lb/>We apply the Pagerank algorithm (Lawrence <lb/>et al., 1998) on the resulting graph. We set the <lb/>probability of performing random jump between <lb/>nodes λ=0.15. The algorithm terminates when <lb/>the change of node weight between iterations is <lb/>smaller than 10 −4 for all nodes. Word importance <lb/>is equal to the final weight of its corresponding <lb/>node in the graph. <lb/> 5 Summary Generation Process <lb/> In this section, we outline how summaries <lb/>are generated by a greedy optimization system <lb/>which selects the sentence with highest weight <lb/>iteratively. This is the main process we use in all <lb/>our summarization systems. For comparison we <lb/>also use a summarization algorithm based on KL <lb/>divergence. <lb/> 5.1 Greedy Optimization Approach <lb/> Our algorithm extracts sentences by weighting <lb/>them based on word importance. The approach is <lb/>similar to the standard word probability baseline <lb/>(Nenkova et al., 2006) but we explore a range <lb/>of possibilities for assigning weights to individual <lb/>words. For each sentence, we calculate the <lb/>sentence weight by summing up the weights of <lb/>all words, normalized by the number of words in <lb/>the sentence. We sort the sentences in descending <lb/>order of their scores into a queue. To create a <lb/>summary, we iteratively dequeue one sentence, <lb/>check if the sentence is more than 8 words (as <lb/>in Erkan and Radev (2004)), then append it to <lb/>the current summary if it is non-redundant. A <lb/>sentence is considered non-redundant if it is not <lb/>similar to any sentences already in the summary, <lb/>measured by cosine similarity on binary vector <lb/>representations with stopwords excluded. We use <lb/>the cut-off of 0.5 for cosine similarity. This value <lb/>was tuned on the DUC 2003 dataset, by testing the <lb/>impact of the cut-off value on the ROUGE scores <lb/>for the final summary. Possible values ranged <lb/>from 0.1 to 0.9 with step of 0.1. <lb/> 5.2 KL Divergence Summarizer <lb/> The KLSUM summarizer (Haghighi and <lb/>Vanderwende, 2009) aims at minimizing the KL <lb/>divergence between the probability distribution <lb/>over words estimated from the summary and <lb/>the input respectively. This summarizer is a <lb/>component of the popular topic model approaches <lb/>(Daumé and Marcu, 2006; Celikyilmaz and <lb/>Hakkani-Tür, 2011; Mason and Charniak, 2011) <lb/>and achieves competitive performance with <lb/>minimal differences compared to a full-blown <lb/>topic model system. <lb/> 6 Global Indicators from NYT <lb/> Some words evoke topics that are of intrinsic <lb/>interest to people. Here we search for global <lb/>indicators of word importance regardless of <lb/>particular input. <lb/> 6.1 Global Indicators of Word Importance <lb/> We analyze a large corpus of original documents <lb/>and corresponding summaries in order to identify <lb/>words that consistently get included in or excluded <lb/>from the summary. In the 2004-2007 NYT corpus, <lb/>many news articles have abstracts along with the <lb/>original article, which makes it an appropriate <lb/>
30
+
31
+ <page> 714 <lb/></page>
32
+
33
+ Metric <lb/> Top-30 words <lb/> KL(A ∥ G)(w) photo(s), pres, article, column, reviews, letter, York, Sen, NY, discusses, drawing, op-ed, holds, Bush <lb/>correction, editorial, dept, city, NJ, map, corp, graph, contends, Iraq, John, dies, sec, state, comments <lb/> KL(G ∥ A)(w) <lb/> Mr, Ms, p.m., lot, Tuesday, CA, Wednesday, Friday, told, Monday, time, a.m., added, thing, Sunday <lb/>things, asked, good, night, Saturday, nyt, back, senator, wanted, kind, Jr., Mrs, bit, looked, wrote <lb/> P rA(w) <lb/> photo, photos, article, York, column, letter, Bush, state, reviews, million, American <lb/>pres, percent, Iraq, year, people, government, John, years, company, correction <lb/>national, federal, officials, city, drawing, billion, public, world, administration <lb/> Table 2: Top 30 words by three metrics from NYT corpus <lb/> resource to do such analysis. We identified <lb/> 160, 001 abstract-original pairs in the corpus. <lb/>From these, we generate two language models, <lb/>one estimated from the text of all abstracts (LM A ), <lb/>the other estimated from the corpus of original <lb/>articles (LM G ). We use SRILM (Stolcke, 2002) <lb/>with Ney smoothing. <lb/>We denote the probability of word w in LM A as <lb/> P r A (w), the probability in LM G as P r G (w), and <lb/>calculate the difference P r A (w)−P r G (w) and the <lb/>ratio P r A (w)/P r G (w) to capture the change of <lb/>probability. In addition, we calculate KL-like <lb/>weighted scores for words which reflect both the <lb/>change of probabilities between the two samples <lb/>and the overall frequency of the word. Here <lb/>we calculate both KL(A ∥ G) and KL(G ∥ <lb/> A). Words with high values for the former score <lb/>are favored in the summaries because they have <lb/>higher probability in the abstracts than in the <lb/>originals and have relatively high probability in <lb/>the abstracts. The later score is high for words that <lb/>are often not included in summaries. <lb/> KL(A ∥ G)(w) = P r A (w) · ln <lb/> P r A (w) <lb/> P r G (w) <lb/> KL(G ∥ A)(w) = P r G (w) · ln <lb/> P r G (w) <lb/> P r A (w) <lb/> Table 2 shows examples of the global <lb/>information captured from the three types <lb/>of scores—KL(A ∥ G), KL(G ∥ A) and <lb/> P r A (w)—listing the 30 content words with <lb/>highest scores for each type. Words that tend to <lb/>be used in the summaries, characterized by high <lb/> KL(A ∥ G) scores, include locations (York, NJ, <lb/>Iraq), people&apos;s names and titles (Bush, Sen, John), <lb/> some abbreviations (pres, corp, dept) and verbs of <lb/>conflict (contends, dies). On the other hand, from <lb/> KL(G ∥ A), we can see that it is unlikely for <lb/>writers to include courtesy titles (Mr, Ms, Jr.) and <lb/>relative time reference in summaries. The words <lb/>with high P r A (w) scores overlaps with those <lb/>ranked highly by KL(A ∥ G) to some extent, <lb/>but also includes a number of generally frequent <lb/>words which appeared often both in the abstracts <lb/>and original texts, such as million and percent. <lb/> 6.2 Blind Sentence Extraction <lb/> In later sections we include the measures of <lb/>global word importance as a feature of our <lb/>regression model for predicting word weights for <lb/>summarization. Before turning to that, however, <lb/>we report the results of an experiment aimed to <lb/>confirm the usefulness of these features. We <lb/>present a system, BLIND, which uses only weights <lb/>assigned to words by KL(A ∥ G) from NYT, <lb/>without doing any analysis of the original input. <lb/>We rank all non-stopword words from the input <lb/>according to this score. The top k words are given <lb/>weight 1, while the others are given weight 0. <lb/>
34
+
35
+ The summaries are produced following the greedy <lb/>procedure described in Section 5.1. <lb/>Systems <lb/>R-1 <lb/>R-2 R-4 <lb/>RANDOM <lb/>30.32 4.42 0.36 <lb/>BLIND (80 keywords) 30.77 5.18 0.53 <lb/>BLIND (300 keywords) 32.91 5.94 0.61 <lb/>LASTESTLEAD <lb/>31.39 6.11 0.63 <lb/>FIRST-SENTENCE <lb/>34.26 7.22 1.21 <lb/>Table 3: <lb/>Blind sentence extraction system, <lb/>compared with three baseline systems (%) <lb/>Table 3 shows that the BLIND system has R-2 <lb/>recall of 0.0594 using the top 300 keywords, <lb/>significantly better than picking sentences from <lb/>the input randomly. It also achieves comparable <lb/>performance with the baseline in DUC 2004, <lb/>formed by selecting the first 100 words from <lb/>the latest article in the input (LASTESTLEAD). <lb/>However it is significantly worse than another <lb/>baseline of selecting the first sentences from the <lb/>input. Table 4 gives sample summaries generated <lb/>by these three approaches. These results confirm <lb/>that the information gleaned from the analysis <lb/>
36
+
37
+ <page> 715 <lb/></page>
38
+
39
+ Random Summary <lb/> It was sunny and about 14 degrees C(57 degrees F) in Tashkent on Sunday. The president is a strong person, and he has been <lb/>through far more difficult political situations, Mityukov said, according to Interfax. But Yeltsin&apos;s aides say his first term, <lb/>from 1991 to 1996, does not count because it began six months before the Soviet Union collapsed and before the current <lb/>constitution took effect. He must stay in bed like any other person, Yakushkin said. The issue was controversial earlier this <lb/>year when Yeltsin refused to spell out his intentions and his aides insisted he had the legal right to seek re-election. <lb/> NYT Summary from global keyword selection, KL(A ∥ G), k = 300 <lb/> Russia&apos;s constitutional court opened hearings Thursday on whether Boris Yeltsin can seek a third term. Yeltsin&apos;s growing <lb/>health problems would also seem to rule out another election campaign. The Russian constitution has a two-term limit for <lb/>presidents. Russian president Boris Yeltsin cut short a trip to Central Asia on Monday due to a respiratory infection that <lb/>revived questions about his overall health and ability to lead Russia through a sustained economic crisis. The upper house of <lb/>parliament was busy voting on a motion saying he should resign. The start of the meeting was shown on Russian television. <lb/> First Sentence Generated Summary <lb/> President Boris Yeltsin has suffered minor burns on his right hand, his press office said Thursday. President Boris Yeltsin&apos;s <lb/>doctors have pronounced his health more or less normal, his wife Naina said in an interview published Wednesday. President <lb/>Boris Yeltsin, on his first trip out of Russia since this spring, canceled a welcoming ceremony in Uzbekistan on Sunday <lb/>because he wasn&apos;t feeling well, his spokesman said. Doctors ordered Russian President Boris Yeltsin to cut short his Central <lb/>
40
+
41
+ Asian trip because of a respiratory infection and he agreed to return home Monday, a day earlier than planned, officials said. <lb/> Table 4: Summary comparison by Random, Blind Extraction and First Sentence systems <lb/>of NYT abstract-original pairs encodes highly <lb/>relevant information about important content <lb/>independent of the actual text of the input. <lb/> 7 Regression-Based Keyword Extraction <lb/> Here we introduce a logistic regression model <lb/>for assigning importance weights to words in the <lb/>input. Crucially, this model combines evidence <lb/>from multiple indicators of importance. We have <lb/>at our disposal abundant data for learning because <lb/>each content word in the input can be treated as <lb/>a labeled instance. There are in total 32, 052 <lb/> samples from the 30 inputs of DUC 2003 for <lb/>training, 54, 591 samples from the 50 inputs of <lb/>DUC 2004 for testing. For a word in the input, <lb/>we assign label 1 if the word appears in at least <lb/>one of the four human summaries for this input. <lb/>Otherwise we assign label 0. <lb/> In the rest of this section, we describe the rich <lb/>variety of features included in our system. We also <lb/>analyze and discuss the predictive power of those <lb/>features by performing Wilcoxon signed-rank test <lb/>on the DUC 2003 dataset. There are in total 9, 261 <lb/> features used, among them 1, 625 are significant <lb/>(p-value &lt; 0.05). We rank these features in <lb/>increasing p-values derived from Wilcoxon test. <lb/>Apart from the widely used features of word <lb/>frequency and positions, some other less explored <lb/>features are highly significant. <lb/> 7.1 Frequency Features <lb/> We use the Probability, LLR chi-square statistic <lb/>value and MRW scores as features. Since prior <lb/>work has demonstrated that for LLR weights in <lb/>particular, it is useful to identify a small set of <lb/>important words and ignore all other words in <lb/>summary selection (Gupta et al., 2007), we use <lb/>a number of keyword indicators as features. For <lb/>these indicators, the value of feature is 1 if the <lb/>word is ranked within top k i , 0 otherwise. Here k i <lb/> are preset cutoffs 5 . These cutoffs capture different <lb/>possibilities for defining the keywords in the input. <lb/>We also add the number of input documents that <lb/>contain the word as a feature. There are a total of <lb/> 100 features in this group, all of which are highly <lb/>significant, ranked among the top 200. <lb/> 7.2 Standard features <lb/> We now describe some standard features which <lb/>have been applied in prior work on summarization. <lb/> Word Locations: Especially in news articles, <lb/>sentences that occur at the beginning are often the <lb/>most important ones. In line with this observation, <lb/>we calculate several features related to the position <lb/>in which a word appears. We first compute <lb/>the relative positions for word tokens, where <lb/>the tokens are numbered sequentially in order of <lb/>appearance in each document in the input. The <lb/>relative position for one word token is therefore <lb/>its corresponding number, divided by total number <lb/>of tokens minus one in the document, e.g., 0 <lb/> for the first token, 1 for the last token. For <lb/>each word, we calculate its earliest first location, <lb/>latest last location, average location and average <lb/>first location for tokens of this word across all <lb/>documents in the input. In addition we have a <lb/>binary feature indicating if the word appears in the <lb/>
42
+
43
+ <note place="footnote">5 10, 15, 20, 30, 40, · · · , 190, 200, 220, 240, 260, 280, <lb/>300, 350, 400, 450, 500, 600, 700 (in total 33 values) <lb/></note>
44
+
45
+ <page>716 <lb/></page>
46
+
47
+ first sentence and the number of times it appears <lb/>in a first sentence among documents in one input. <lb/>There are 6 features in this group. All of them are <lb/>very significant, ranked within the top 100. <lb/> Word type: These features include Part of <lb/>Speech (POS) tags, Name Entity (NE) labels and <lb/>capitalization information. We use the Stanford <lb/>POS-Tagger (Toutanova et al., 2003) and Name <lb/>Entity Recognizer (Finkel et al., 2005). We have <lb/>one feature corresponding to each possible POS <lb/>and NE tag. The value of this feature is the <lb/>proportion of occurrences of the word with this <lb/>tag; in most cases only one feature gets a non-zero <lb/>value. We have two features which indicate if <lb/>one word has been capitalized and the ratio of its <lb/>capitalized occurrences. <lb/>Most of the NE features (6 out of 8) are <lb/>significant: there are more Organizations and <lb/> Locations but fewer Time and Date words in the <lb/>human summaries. Of the POS tags, 11 out of 41 <lb/> are significant: there are more nouns (NN, NNS, <lb/>NNPS); fewer verbs (VBG, VBP, VB) and fewer <lb/>cardinal numbers in the abstracts compared to the <lb/>input. Capitalized words also tend to be included <lb/>in human summaries. <lb/> KL: Prior work has shown that having estimates <lb/>of sentence importance can also help in estimating <lb/>word importance (Wan et al., 2007; Liu et al., <lb/>2011; Wei et al., 2008). The summarizer based <lb/>on KL-divergence assigns importance to sentences <lb/>directly, in a complex function according to the <lb/>word distribution in the sentence. Therefore, <lb/>we use these summaries as potential indicators <lb/>of word importance. We include two features <lb/>here, the first one indicates if the word appears <lb/>in a KLSUM summary of the input, as well as <lb/>a feature corresponding to the number of times <lb/>the word appeared in that summary. Both of the <lb/>features are highly significant, ranked within the <lb/>top 200. <lb/> 7.3 NYT-weights as Features <lb/> We include features from the relative rank of <lb/>a word according to KL(A ∥ G), KL(G ∥ <lb/> A), P r A (w)−P r G (w), P r A (w)/P r G (w) and <lb/> P r A (w), derived from the NYT as described in <lb/>Section 6. If the rank of a word is within top-k <lb/>or bottom-k by one metric, we would label it as <lb/> 1, where k is selected from a set of pre-defined <lb/>values 6 . We have in total 70 features in this <lb/>
48
+
49
+ <note place="footnote">6 100, 200, 500, 1000, 2000, 5000, 10000 in this case. <lb/></note>
50
+
51
+ category, of which 56 are significant, 47 having <lb/>a p-value less than 10 −7 . The predictive power of <lb/>those global indicators are only behind the features <lb/>which indicates frequency and word positions. <lb/> 7.4 Unigrams <lb/> This is a binary feature corresponding to each <lb/>of the words that appeared at least twice in the <lb/>training data. The idea is to learn which words <lb/>from the input tend to be mentioned in the human <lb/>summaries. There are in total 8, 691 unigrams, <lb/>among which 1, 290 are significant. Despite the <lb/>high number of significant unigram features, most <lb/>of them are not as significant as the more general <lb/>ones we described so far. It is interesting to <lb/>compare the significant unigrams identified in the <lb/>DUC abstract/input data with those derived from <lb/>the NYT corpus. Unigrams that tend to appear in <lb/>DUC summaries include president, government, <lb/>political. We also find the same unigrams among <lb/>the top words from NYT corpus according to <lb/> KL(A ∥ G) . As for words unlikely to appear in <lb/>summaries, we see Wednesday, added, thing, etc, <lb/>which again rank high according to KL(G ∥ A). <lb/> 7.5 Dictionary Features: MPQA and LIWC <lb/> Unigram features are notoriously sparse. To <lb/>mitigate the sparsity problem, we resort to <lb/>more general groupings to words according to <lb/>salient semantic and functional categories. We <lb/>employ two hand-crafted dictionaries, MPQA for <lb/>subjectivity analysis and LIWC for topic analysis. <lb/>The MPQA dictionary (Wiebe and Cardie, <lb/>2005) contains words with different polarities <lb/>(positive, neutral, negative) and intensities (strong, <lb/>weak). The combinations correspond to six <lb/>features. It turns out that words with strong <lb/>polarity, either positive or negative, are seldomly <lb/>included in the summaries. Most strikingly, <lb/>the p-value from significance test for the strong <lb/>negative words is less than 10 −4 —these words <lb/>are rarely included in summaries. There is no <lb/>significant difference on weak polarity categories. <lb/>Another dictionary we use is LIWC (Tausczik <lb/>and Pennebaker, 2007), which contains manually <lb/>constructed dictionaries for multiple categories <lb/>of words. The value of the feature is 1 for <lb/>one word if the word appears in the particular <lb/>dictionary for the category. 34 out of 64 LIWC <lb/>features are significant. Interesting categories <lb/>which appear at higher rate in summaries include <lb/>events about death, anger, achievements, money <lb/>
52
+
53
+ <page>717 <lb/></page>
54
+
55
+ and negative emotions. Those that appear at lower <lb/>rate in the summaries include auxiliary verbs, hear, <lb/>pronouns, negation, function words, social words, <lb/>swear, adverbs, words related to families, etc. <lb/> 7.6 Context Features <lb/> We use context features here, based on the <lb/>assumption that context importance around a word <lb/>affects the importance of this word. For context <lb/>we consider the words before and after the target <lb/>word. We extend our feature space by calculating <lb/>the weighted average of the feature values of the <lb/>context words. For word w, we denote L w as the <lb/>set of words before w, R w as the set of words <lb/>after w. We denote the feature for one word as <lb/> w.f i , the way of calculating the newly extended <lb/>word-before feature w.l f i could be written as: <lb/> w.l f i = <lb/> ∑ <lb/> i <lb/> p(w l ) · w l .f i , ∀w l ∈ L w <lb/> Here p(w l ) is the probability word w l appears <lb/>before w among all words in L w . <lb/>For context features, we calculate the weighted <lb/>average of the most widely used basic features, <lb/>including frequency, location and capitalization <lb/>for surrounding contexts. There are in total <lb/> 220 features of this kind, among which 117 are <lb/>significant, 74 having a p-value less than 10 −4 . <lb/> 8 Experiments <lb/> The performance of our logistic regression model <lb/>is evaluated on two tasks: keyword identification <lb/>and extractive summarization. We name our <lb/>system REGSUM. <lb/> 8.1 Regression for Keyword Identification <lb/> For each input, we define the set of keywords <lb/>as the top k words according to the scores <lb/>generated from different models. We compare <lb/>our regression system with three unsupervised <lb/>systems: PROB, LLR, MRW. To show the <lb/>effectiveness of new features, we compare our <lb/>results with a regression system trained only <lb/>on word frequency and location related features <lb/>described in Section 7. Those features are the <lb/>ones standardly used for ranking the importance <lb/>of words in recent summarization works (Yih et <lb/>al., 2007; Takamura and Okumura, 2009; Sipos et <lb/>al., 2012), and we name this system REGBASIC. <lb/>Figure 1 shows the performance of systems <lb/>when selecting the 100 words with highest weights <lb/>Figure 1: <lb/>Precision, Recall and F-score of <lb/>keyword identification, 100 words selected, G 1 as <lb/>gold-standard <lb/>as keywords. Each word from the input that <lb/>appeared in any of the four human summaries is <lb/>considered as a gold-standard keyword. Among <lb/>the unsupervised approaches, word probability <lb/>identifies keywords better than LLR and MRW <lb/>by at least 4% on F-score. REGBASIC does not <lb/>give better performance at keyword identification <lb/>compared with PROB, even though it includes <lb/>location information. Our system gets 2.2% <lb/> F-score improvement over PROB, 5.2% over <lb/>REGBASIC, and more improvement over the <lb/>other approaches. All of these improvements are <lb/>statistically significant by Wilcoxon test. <lb/>Table 5 shows the performance of keyword <lb/>identification for different G i and different <lb/>number of keywords selected. The regression <lb/>system has no advantage over PROB when <lb/>identifying keywords that appeared in all of the <lb/>four human summaries. However our system <lb/>achieves significant improvement for predicting <lb/>words that appeared in more than one or two <lb/>human summaries. 7 <lb/> 8.2 Regression for Summarization <lb/> We now show that the performance of extractive <lb/>summarization can be improved by better <lb/>estimation of word weights. We compare our <lb/>regression system with the four models introduced <lb/>in Section 8.1. We also include PEER-65, the best <lb/>system in DUC-2004, as well as KLSUM for <lb/>comparison. Apart from these, we compare our <lb/>model with two state-of-the-art systems, including <lb/>the submodular approach (SUBMOD) (Lin and <lb/>
56
+
57
+ <note place="footnote">7 We also apply a weighted keyword evaluation approach, <lb/>similar to the pyramid method for summarization. Still <lb/>our system shows significant improvement over the others. <lb/>See https://www.seas.upenn.edu/~hongkai1/regsum.html for <lb/>details. <lb/></note>
58
+
59
+ <page> 718 <lb/></page>
60
+
61
+ G i #words PROB LLR MRW REGBASIC REGSUM <lb/> G 1 <lb/> 80 <lb/>43.6 <lb/>37.9 <lb/>38.9 <lb/>39.9 <lb/>45.7 <lb/> G 1 <lb/> 100 <lb/>44.3 <lb/>38.7 <lb/>39.2 <lb/>41.0 <lb/>46.5 <lb/> G 1 <lb/> 120 <lb/>44.6 <lb/>38.5 <lb/>39.2 <lb/>40.9 <lb/>46.4 <lb/> G 2 <lb/> 30 <lb/>47.8 <lb/>44.0 <lb/>42.4 <lb/>47.4 <lb/>50.2 <lb/> G 2 <lb/> 35 <lb/>47.1 <lb/>43.3 <lb/>42.1 <lb/>47.0 <lb/>49.5 <lb/> G 2 <lb/> 40 <lb/>46.5 <lb/>42.4 <lb/>41.8 <lb/>46.4 <lb/>49.2 <lb/> G 3 <lb/> 10 <lb/>51.2 <lb/>46.2 <lb/>43.8 <lb/>46.9 <lb/>50.2 <lb/> G 3 <lb/> 15 <lb/>51.4 <lb/>47.5 <lb/>43.7 <lb/>49.8 <lb/>52.9 <lb/> G 3 <lb/> 20 <lb/>49.7 <lb/>47.6 <lb/>42.5 <lb/>49.3 <lb/>51.5 <lb/> G 4 <lb/> 5 <lb/>50.0 <lb/>48.8 <lb/>44.9 <lb/>43.6 <lb/>45.1 <lb/> G 4 <lb/> 6 <lb/>51.4 <lb/>46.9 <lb/>43.7 <lb/>45.2 <lb/>47.6 <lb/> G 4 <lb/> 7 <lb/>50.9 <lb/>48.2 <lb/>43.7 <lb/>45.8 <lb/>47.8 <lb/>Table 5: Keyword identification F-score (%) for different G i and different number of words selected. <lb/>Bilmes, 2012) and the determinantal point process <lb/>(DPP) summarizer (Kulesza and Taskar, 2012). <lb/>The summaries were kindly provided by the <lb/>authors of these systems (Hong et al., 2014). <lb/>As can been seen in Table 6, our system <lb/>outperforms PROB, LLR, MRW, PEER-65, <lb/>KLSUM and REGBASIC. These improvements <lb/>are significant on ROUGE-2 recall. Interestingly, <lb/>although the supervised system REGBASIC which <lb/>uses only frequency and positions achieve <lb/>low performance in keyword identification, the <lb/>summaries it generates are of high quality. The <lb/>inclusion of position features negatively affects the <lb/>performance in summary keyword identification <lb/>but boosts the weights for the words which appear <lb/>close to the beginning of the documents, which is <lb/>helpful for identifying informative sentences. By <lb/>including other features we greatly improve over <lb/>REGBASIC in keyword identification. Similarly <lb/>here the richer set of features results in better <lb/>quality summaries. <lb/>We also examined the ROUGE-1, -2, -4 <lb/>recall compared with the SUBMOD and DPP <lb/>summarizers 8 . There is no significant difference <lb/>on R-2 and R-4 recall compared with these <lb/>two state-of-the-art systems. DPP performed <lb/>significantly better than our system on R-1 recall, <lb/>but that system is optimizing on R-1 F-score in <lb/>training. Overall, our conceptually simple system <lb/>is on par with the state of the art summarizers and <lb/>points to the need for better models for estimating <lb/>word importance. <lb/>
62
+
63
+ <note place="footnote">8 The results are slightly different from the ones reported <lb/>in the original papers due to the fact that we truncated to 100 <lb/>words, while they truncated to 665 bytes. <lb/></note>
64
+
65
+ System <lb/>R-1 <lb/>R-2 R-4 <lb/>PROB <lb/>35.14 8.17 1.06 <lb/>LLR <lb/>34.60 7.56 0.83 <lb/>MRW <lb/>35.78 8.15 0.99 <lb/>REGBASIC 37.56 9.28 1.49 <lb/>KL <lb/>37.97 8.53 1.26 <lb/>PEER-65 <lb/>37.62 8.96 1.51 <lb/>SUBMOD <lb/>39.18 9.35 1.39 <lb/>DPP <lb/>39.79 9.62 1.57 <lb/> REGSUM 38.57 9.75 1.60 <lb/> Table 6: System performance comparison (%) <lb/> 9 Conclusion <lb/> We presented a series of experiments which <lb/>show that keyword identification can be improved <lb/>in a supervised framework which incorporates <lb/>a rich set of indicators of importance. We <lb/>also show that the better estimation of word <lb/>importance leads to better extractive summaries. <lb/>Our analysis of features related to global <lb/>importance, sentiment and topical categories <lb/>reveals rather unexpected results and confirms that <lb/>word importance estimation is a worthy research <lb/>direction. Success in the task is likely to improve <lb/>sophisticated summarization approaches too, as <lb/>well as sentence compression systems which use <lb/>only crude frequency related measures to decide <lb/>which words should be deleted from a sentence. 9 <lb/>
66
+ </body>
67
+
68
+ <back>
69
+
70
+ <div type="acknowledgement">9 The work is partially funded by NSF CAREER award <lb/> IIS 0953445. <lb/></div>
71
+
72
+ <page> 719 <lb/></page>
73
+
74
+ <listBibl>References <lb/> Taylor Berg-Kirkpatrick, Dan Gillick, and Dan Klein. <lb/>2011. Jointly learning to extract and compress. In <lb/> Proceedings of ACL-HLT, pages 481–490. <lb/>Asli Celikyilmaz and Dilek Hakkani-Tur. <lb/>2010. <lb/>A hybrid hierarchical model for multi-document <lb/>summarization. In Proceedings of ACL, pages <lb/>815–824. <lb/>Asli Celikyilmaz and Dilek Hakkani-Tür. <lb/> 2011. <lb/>Discovery of topically coherent sentences for <lb/>extractive summarization. <lb/>In Proceedings of <lb/>ACL-HLT, pages 491–499. <lb/>John M. Conroy, Judith D. Schlesinger, and Dianne P. <lb/>O&apos;Leary. 2006. Topic-focused multi-document <lb/>summarization using an approximate oracle score. <lb/>In Proceedings of COLING/ACL, pages 152–159. <lb/>Hal Daumé, III and Daniel Marcu. 2006. Bayesian <lb/>query-focused summarization. In Proceedings of <lb/>ACL, pages 305–312. <lb/>Gunes Erkan and Dragomir R. Radev. 2004. Lexrank: <lb/>graph-based lexical centrality as salience in text <lb/>summarization. Journal of Artificial Intelligence <lb/>Research, 22(1):457–479. <lb/>Elena Filatova and Vasileios Hatzivassiloglou. 2004. <lb/>A formal model for information selection in <lb/>multi-sentence text extraction. In Proceedings of <lb/>COLING. <lb/> Jenny Rose Finkel, Trond Grenager, and Christopher <lb/>Manning. <lb/>2005. <lb/>Incorporating non-local <lb/>information into information extraction systems by <lb/>gibbs sampling. In Proceedings of ACL, pages <lb/>363–370. <lb/>D. Graff, J. Kong, K. Chen, and K. Maeda. 2007. <lb/>English gigaword third edition. Linguistic Data <lb/>Consortium, Philadelphia, PA. <lb/> Surabhi Gupta, Ani Nenkova, and Dan Jurafsky. <lb/>2007. Measuring importance and query relevance <lb/>in topic-focused multi-document summarization. In <lb/> Proceedings of ACL, pages 193–196. <lb/>Aria Haghighi and Lucy Vanderwende. <lb/>2009. <lb/>Exploring content models for multi-document <lb/>summarization. In Proceedings of HLT-NAACL, <lb/> pages 362–370. <lb/>Sanda Harabagiu and Finley Lacatusu. 2005. Topic <lb/>themes for multi-document summarization. <lb/>In <lb/> Proceedings of SIGIR 2005, pages 202–209. <lb/>Kai Hong, John M. Conroy, Benoit Favre, Alex <lb/>Kulesza, Hui Lin, and Ani Nenkova. 2014. A <lb/>repositary of state of the art and competitive baseline <lb/>summaries for generic news summarization. In <lb/> Proceedings of LREC, May. <lb/>Alex Kulesza and Ben Taskar. 2012. Determinantal <lb/>point processes for machine learning. Foundations <lb/>and Trends in Machine Learning, 5(2–3). <lb/>Julian Kupiec, Jan Pedersen, and Francine Chen. 1995. <lb/>A trainable document summarizer. In Proceedings <lb/>of SIGIR, pages 68–73. <lb/>Page Lawrence, Brin Sergey, Rajeev Motwani, and <lb/>Terry Winograd. 1998. The pagerank citation <lb/>ranking: Bringing order to the web. Technical <lb/>report, Stanford University. <lb/>Hui Lin and Jeff Bilmes. 2012. Learning mixtures <lb/>of submodular shells with application to document <lb/>summarization. In UAI, pages 479–490. <lb/>Chin-Yew Lin and Eduard Hovy. <lb/>2000. <lb/>The <lb/>automated acquisition of topic signatures for text <lb/>summarization. In Proceedgins of COLING, pages <lb/>495–501. <lb/>Chin-Yew Lin. <lb/>2004. <lb/>Rouge: A package for <lb/>automatic evaluation of summaries. <lb/>In Text <lb/>Summarization Branches Out: Proceedings of the <lb/>ACL-04 Workshop, pages 74–81. <lb/>Marina Litvak, Mark Last, and Menahem Friedman. <lb/>2010. A new approach to improving multilingual <lb/>summarization using a genetic algorithm. <lb/>In <lb/> Proceedings of ACL, pages 927–936. <lb/>Fei Liu, Feifan Liu, and Yang Liu. 2011. A supervised <lb/>framework for keyword extraction from meeting <lb/>transcripts. Transactions on Audio Speech and <lb/>Language Processing, 19(3):538–548. <lb/>H. P. Luhn. <lb/>1958. <lb/>The automatic creation of <lb/>literature abstracts. IBM Journal of Research and <lb/>Development, 2(2):159–165, April. <lb/>M. Marneffe, B. Maccartney, and C. Manning. 2006. <lb/>Generating Typed Dependency Parses from Phrase <lb/>Structure Parses. In Proceedings of LREC-06, pages <lb/>449–454. <lb/>Rebecca Mason and Eugene Charniak. <lb/>2011. <lb/>Extractive multi-document summaries should <lb/>explicitly not contain document-specific content. <lb/>In Proceedings of the Workshop on Automatic <lb/>Summarization for Different Genres, Media, and <lb/>Languages, pages 49–54. <lb/>Ryan McDonald. 2007. A study of global inference <lb/>algorithms in multi-document summarization. In <lb/> Proceedings of ECIR, pages 557–564. <lb/>Rada Mihalcea and Paul Tarau. 2004. Textrank: <lb/>Bringing order into text. In Proceedings of EMNLP, <lb/> pages 404–411. <lb/>Ani Nenkova and Lucy Vanderwende. 2005. The <lb/>impact of frequency on summarization. Technical <lb/>report, Microsoft Research. <lb/>
75
+
76
+ <page>720 <lb/></page>
77
+
78
+ Ani Nenkova, Lucy Vanderwende, and Kathleen <lb/>McKeown. 2006. A compositional context sensitive <lb/>multi-document summarizer: exploring the factors <lb/>that influence summarization. In Proceedings of <lb/>SIGIR, pages 573–580. <lb/>Jahna Otterbacher, Günes Erkan, and Dragomir R. <lb/>Radev. <lb/>2009. <lb/>Biased lexrank: Passage <lb/>retrieval using random walks with question-based <lb/>priors. Information Processing and Management, <lb/> 45(1):42–54. <lb/>Paul Over, Hoa Dang, and Donna Harman. 2007. Duc <lb/>in context. Inf. Process. Manage., 43(6):1506–1520. <lb/>Karolina Owczarzak, John M. Conroy, Hoa Trang <lb/>Dang, and Ani Nenkova. 2012. An assessment <lb/>of the accuracy of automatic evaluation in <lb/>summarization. In NAACL-HLT 2012: Workshop <lb/>on Evaluation Metrics and System Comparison for <lb/>Automatic Summarization, pages 1–9. <lb/>Peter Rankel, John Conroy, Eric Slud, and Dianne <lb/>O&apos;Leary. 2011. Ranking human and machine <lb/>summarization systems. In Proceedings of EMNLP, <lb/> pages 467–473. <lb/>Korbinian Riedhammer, Benoˆ t Favre, and Dilek <lb/>Hakkani-Tür. <lb/>2010. <lb/>Long story short -<lb/>global unsupervised models for keyphrase based <lb/>meeting summarization. Speech Communication, <lb/> 52(10):801–815. <lb/>G. Salton. 1971. The SMART Retrieval System: <lb/>Experiments in Automatic Document Processing. <lb/> Prentice-Hall, Inc., Upper Saddle River, NJ, USA. <lb/>Evan Sandhaus. 2008. The new york times annotated <lb/>corpus. Linguistic Data Consortium, Philadelphia, <lb/>PA. <lb/> Chao Shen and Tao Li. 2010. Multi-document <lb/>summarization via the minimum dominating set. In <lb/> Proceedings of Coling, pages 984–992. <lb/>Ruben Sipos, Pannaga Shivaswamy, and Thorsten <lb/>Joachims. <lb/>2012. <lb/>Large-margin learning of <lb/>submodular summarization models. In Proceedings <lb/>of EACL, pages 224–233. <lb/>Andreas Stolcke. 2002. SRILM – an extensible <lb/>language modeling toolkit. <lb/>In Proceedings of <lb/>ICSLP, volume 2, pages 901–904. <lb/>Hiroya Takamura and Manabu Okumura. 2009. Text <lb/>summarization model based on maximum coverage <lb/>problem and its variant. In Proceedings of EACL, <lb/> pages 781–789. <lb/>Yla R Tausczik and James W Pennebaker. 2007. <lb/>The Psychological Meaning of Words: LIWC and <lb/>Computerized Text Analysis Methods. Journal of <lb/>Language and Social Psychology, 29:24–54. <lb/>Kristina Toutanova, Dan Klein, Christopher D. <lb/>Manning, and Yoram Singer. 2003. Feature-rich <lb/>part-of-speech tagging with a cyclic dependency <lb/>network. In Proceedings of the NAACL-HLT, pages <lb/>173–180. <lb/>Xiaojun Wan and Jianwu Yang. 2008. Multi-document <lb/>summarization using cluster-based link analysis. In <lb/> Proceedings of SIGIR, pages 299–306. <lb/>Xiaojun Wan, Jianwu Yang, and Jianguo Xiao. <lb/>2007. Towards an iterative reinforcement approach <lb/>for simultaneous document summarization and <lb/>keyword extraction. In Proceedings of ACL, pages <lb/>552–559. <lb/>Furu Wei, Wenjie Li, Qin Lu, and Yanxiang He. 2008. <lb/>Query-sensitive mutual reinforcement chain and <lb/>its application in query-oriented multi-document <lb/>summarization. In Proceedings of SIGIR, pages <lb/>283–290. <lb/>Janyce Wiebe and Claire Cardie. 2005. Annotating <lb/>expressions of opinions and emotions in language. <lb/>language resources and evaluation. In Language <lb/>Resources and Evaluation (formerly Computers and <lb/>the Humanities), page 1(2). <lb/>Wen-tau Yih, Joshua Goodman, Lucy Vanderwende, <lb/>and Hisami Suzuki. <lb/>2007. <lb/>Multi-document <lb/>summarization by maximizing informative <lb/>content-words. In Proceedings of IJCAI, pages <lb/>1776–1782. <lb/>Hongyuan Zha. 2002. Generic summarization and <lb/>keyphrase extraction using mutual reinforcement <lb/>principle and sentence clustering. In Proceedings <lb/>of SIGIR, pages 113–120. <lb/></listBibl>
79
+
80
+ <page> 721 </page>
81
+
82
+ </back>
83
+ </text>
84
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EJM_CavitationModelsComparison.training.segmentation.tei.xml ADDED
@@ -0,0 +1,183 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei>
3
+ <teiHeader>
4
+ <fileDesc xml:id="__EJM_CavitationModelsComparison"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front> A comparative study of cavitation models in a Venturi <lb/>flow <lb/> Boris Charì ere a , Jean Decaix b , Eric Goncalv es a, § <lb/> a <lb/> LEGI -University of Grenoble, 1025 rue de la Piscine, 38400 St Martin d&apos;Heres, France <lb/> b <lb/> University of Applied Sciences, Western Switzerland, CH-1950 Sion <lb/> Abstract <lb/> This paper presents a numerical study of an aperiodic cavitation pocket <lb/>developing in a Venturi flow. The mass transfer between phases is driven <lb/>by a void ratio transport equation model. A new free-parameter closure <lb/>relation is proposed and compared with other formulations. The re-entrant <lb/>jet development, void ratio profiles and pressure fluctuations are analyzed to <lb/>discern results accuracy. Comparisons with available experimental data are <lb/>done and good agreement is achieved. <lb/> Keywords: Cavitation, Mass transfer, Homogeneous model, RANS <lb/>simulation <lb/> Nomenclature <lb/> c <lb/> speed of sound <lb/> C p , C v <lb/> thermal capacities <lb/> E <lb/> total energy <lb/> e <lb/> internal energy <lb/> § <lb/> Corresponding author <lb/> Email address: [email protected] (Eric Goncalv es ) <lb/> Preprint submitted to Elsevier <lb/>January 8, 2014 <lb/></front>
8
+
9
+ <body> k <lb/> turbulent kinetic energy <lb/> P <lb/> static pressure <lb/> P vap <lb/> vapour pressure <lb/> P r , P rt <lb/> molecular and turbulent Prandtl numbers <lb/> Q <lb/> total heat flux <lb/> Re L <lb/> Reynolds number based on the length L <lb/> T <lb/> temperature <lb/> u, v <lb/> velocity components <lb/> V <lb/> vector velocity <lb/> w <lb/> conservative variables <lb/> Y <lb/> mass fraction of gas <lb/>˙ <lb/> m <lb/> mass transfer <lb/> AE <lb/> void fraction <lb/> ∞ <lb/> ratio of thermal capacities <lb/> &quot; <lb/> dissipation rate <lb/> ∏, ∏ t <lb/> molecular and turbulent thermal conductivity <lb/> µ, µ t <lb/> molecular and eddy viscosity <lb/> Ω <lb/> density <lb/> Ω I <lb/> interfacial density <lb/> ae <lb/> cavitation number <lb/> ø <lb/> total stress tensor <lb/> ! <lb/> specific dissipation <lb/>() l <lb/> liquid value <lb/>() v <lb/> vapour value <lb/>() v <lb/> viscous <lb/>
10
+
11
+ <page>2 <lb/></page>
12
+
13
+ () t <lb/> turbulent <lb/> 1. Introduction <lb/> Cavitation is the formation of vapour cavities within a given liquid due to <lb/>pressure drop. It may be observed in various engineering systems such as <lb/>hydraulic constructions, aeronautics, aerospace, power systems and turbo-<lb/>machinery. The cavitation development may be the origin of several nega-<lb/>tive eAEects, such as noise, vibrations, performance alterations, erosion and <lb/>structural damages. This makes cavitation an important issue in design and <lb/>operation, which should be controlled, or at least well understood. <lb/>Among the cavitation types that may develop, partial cavitation pockets <lb/>are often observed in hydraulic machines and is known to be responsible for <lb/>severe damage. Cavitating venturis are one of the simplest case to study <lb/>such cavities, both experimentally and numerically. This kind of cavitation <lb/>is characterized by a partial vapour cavity that detaches from the solid body <lb/>and extends downstream with the existence of a re-entrant jet. The dynamic <lb/>of these cavitation sheets, the interaction between cavitation and turbulence, <lb/>the behaviour of the turbulent boundary layer are not yet well known and <lb/>understood. <lb/>Although the numerical modelling of such cavitation has received a great <lb/>deal of attention, it is still a very di±cult and challenging task to simulate <lb/>such complex unsteady two-phase flow with an acceptable accuracy. Cavitat-<lb/>ing flows are challenging to model, since they are turbulent with a complex <lb/>
14
+
15
+ <page>3 <lb/></page>
16
+
17
+ interaction with two-phase structures, highly dynamic and involve non equi-<lb/>librium thermodynamic states. Several numerical models have been devel-<lb/>oped to investigate such cavitating flows, especially with one-fluid Reynolds-<lb/>Averaged Navier-Stokes (RANS) solvers. The homogeneous mixture model <lb/>treats the cavitating flows as a mixture of two fluids behaving as one. These <lb/>models are based on the assumption of local kinematic equilibrium between <lb/>phases (the local velocity is the same for both phases), local thermal and <lb/>mechanic equilibrium between the two components (local temperature and <lb/>pressure equality between phases). These models are composed by three <lb/>conservation laws for mixture quantities (mass, momentum and total en-<lb/>ergy). This model cannot reproduce strong thermodynamic non equilibrium <lb/>eAEects but, because of its simplicity, it is often used for numerical simulations <lb/>[1, 2, 3, 4, 5, 6, 7, 8]. <lb/>By assuming that one pure phase is on a metastable state, a supplementary <lb/>mass equation or void fraction equation is added. Various formulations of <lb/>four-equation model have been expressed. A very popular formulation has <lb/>been developed to simulate turbulent cavitating flows [9, 10, 11, 12, 13, 14]. <lb/>The main di±culty is related to the formulation of the source term and the <lb/>tunable parameters involved for the vaporization and condensation processes. <lb/>Moreover, these models are not thermodynamically well-posed [15]. Another <lb/>popular model devoted to ebullition problems uses a relaxation term (Ho-<lb/>mogeneous Relaxation Model). The source term involves a relaxation time <lb/>estimated from experimental data [16] or with an optimization problem on <lb/>the mixture entropy [17]. An original formulation was recently proposed for <lb/>the mass transfer between phases assuming its proportionality with the diver-<lb/>
18
+
19
+ <page>4 <lb/></page>
20
+
21
+ gence of the mixture velocity. This model was validated on various inviscid <lb/>and turbulent applications [18, 19]. <lb/>The present work is devoted to the numerical study of a partial cavitation <lb/>pocket appearing on a venturi geometry. For this test case, a transitional be-<lb/>haviour is observed between a stable pocket and a periodic cycles pocket. <lb/>A particular emphasis is placed on the comparison of various void ratio <lb/>transport equation models and their ability to capture the re-entrant jet <lb/>phenomenon. An in-house finite-volume code solving a four-equation RANS <lb/>compressible system was developed [19]. A new cavitation model is investi-<lb/>gated using a mixture of stiAEened gas equation of state (EOS). The formula-<lb/>tion does not involve any tunable parameter. Validation and comparisons are <lb/>done with experimental measurements (time-averaged void ratio and veloc-<lb/>ity profiles, RMS wall pressure fluctuations). A comparison is proposed with <lb/>OpenFOAM simulations in which the Kunz&apos;s void ratio transport equation <lb/>model is considered. The opensource software OpenFOAM was used and <lb/>validated in cavitating flows by various authors [20, 21, 22]. <lb/>In this paper, we will first review the theoretical formulation, including phys-<lb/>ical models, equation of state and elements of the numerical methods. This <lb/>is followed by sets of results on a Venturi geometry and discussions. <lb/> 2. The LEGI&apos;s numerical tool <lb/> The code is based on the solving of the one-fluid compressible RANS system <lb/>with transport-equation turbulence models. <lb/>
22
+
23
+ <page>5 <lb/></page>
24
+
25
+ 2.1. Reynolds-Averaged Navier-Stokes compressible equations <lb/> The compressible one-fluid RANS equations are used, coupled with a one-<lb/>or two-equation turbulence model. For low Mach number applications, an <lb/>inviscid preconditioner is introduced. These equations can be expressed as: <lb/> P °1 <lb/> c <lb/> @w <lb/> @t <lb/> + div (F c ° F v ) = S <lb/> (1) <lb/> w = <lb/> 0 <lb/> B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>@ <lb/> Ω <lb/>ΩV <lb/>ΩE <lb/>AE <lb/>Ωk <lb/>Ω™ <lb/> 1 <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>A <lb/> ; F c = <lb/> 0 <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>@ <lb/> ΩV <lb/>ΩV ≠ V + pI <lb/> (ΩE + p)V <lb/> AEV <lb/>ΩkV <lb/>Ω™V <lb/> 1 <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>A <lb/> ; F v = <lb/> 0 <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>B <lb/>@ <lb/> 0 <lb/> ø v + ø t <lb/> (ø v + ø t ).V ° Q v ° Q t <lb/> 0 <lb/>(µ + µ t /ae k ) grad k <lb/> (µ + µ t /ae ™ ) grad ™ <lb/> 1 <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>C <lb/>A <lb/> where w denotes the conservative variables and the void ratio, F c and F v <lb/> the convective and viscous flux densities and S the source terms, which con-<lb/>cern the void ratio equation and the turbulent transport equations. The <lb/>expression of the preconditioning matrix P c is given in [19]. k is the mixture <lb/>turbulent kinetic energy (TKE) and ™ is a mixture turbulent variable. In <lb/>multiphase flow, the divergence of the fluctuating phase velocity is not zero <lb/>[23]. Therefore, supplementary terms appear in the mixture TKE equation <lb/>(pressure-dilation term, dilatational dissipation rate), which are not taken <lb/>into account in the present paper. <lb/>The exact expression of the eddy-viscosity µ t and the source terms depends <lb/>on the turbulence model as well as constants ae k and ae ™ . <lb/>The total stress tensor ø is evaluated using the Stokes hypothesis, Newton&apos;s <lb/>
26
+
27
+ <page>6 <lb/></page>
28
+
29
+ law and the Boussinesq assumption. The total heat flux vector Q is obtained <lb/>from the Fourier law involving a turbulent thermal conductivity ∏ t with the <lb/>constant Prandtl number hypothesis. <lb/> ø = ø v + ø t = (µ + µ t ) <lb/> ∑ <lb/> ( grad V + ( grad V ) t ) ° <lb/> 2 <lb/>3 <lb/>( div V )I <lb/> ∏ <lb/> + <lb/>2 <lb/>3 <lb/> ΩkI <lb/> Q = Q v + Q t = ° (∏ + ∏ t ) grad T with ∏ t = <lb/> µ t C p <lb/> P rt <lb/> (2) <lb/>In pure phases, the viscosity is assumed to be constant. The mixture viscosity <lb/>is defined as the arithmetic mean of the liquid and vapour viscosities: <lb/> µ(AE) = AEµ V + (1 ° AE)µ L <lb/> (3) <lb/>The mixture thermal conductivity ∏ is also defined as the arithmetic mean <lb/>of the liquid and vapour values: <lb/> ∏(AE) = AE <lb/> µ V C p V <lb/> P r V <lb/> + (1 ° AE) <lb/> µ L C p L <lb/> P r L <lb/> (4) <lb/>The turbulent Prandtl number P rt is set to 1. <lb/>To close the system, an equation of state (EOS) is necessary to link the <lb/>pressure and the temperature to the internal energy and the density. For the <lb/>pure phases, we used the convex stiAEened gas EOS: <lb/> P (Ω, e) = (∞ ° 1)Ω(e ° q) ° ∞P 1 <lb/> (5) <lb/> P (Ω, T ) = Ω(∞ ° 1)C v T ° P 1 <lb/> (6) <lb/> T (Ω, h) = <lb/> h ° q <lb/>C p <lb/> (7) <lb/>where ∞ = C p /C v is the heat capacity ratio, C p and C v are thermal capacities, <lb/> q the energy of the fluid at a given reference state and P 1 is a constant <lb/>reference pressure. <lb/>
30
+
31
+ <page>7 <lb/></page>
32
+
33
+ 2.2. A void ratio transport equation <lb/> A void ratio equation can be expressed as [24]: <lb/> @AE <lb/> @t <lb/> + div (AEV ) = (K + AE) div V + <lb/>˙ <lb/> m <lb/> Ω I <lb/> (8) <lb/> K = <lb/> √ <lb/> Ω l c 2 <lb/> l ° Ω v c 2 <lb/> v <lb/> Ω l c 2 <lb/> l <lb/> 1°AE <lb/> + Ω v c 2 v <lb/> AE <lb/> ! <lb/> ; <lb/> Ω I = <lb/> √ c 2 <lb/> v <lb/> AE <lb/> + <lb/> c 2 <lb/> l <lb/> 1°AE <lb/> Ω l c 2 <lb/> l <lb/> 1°AE <lb/> + Ω v c 2 v <lb/> AE <lb/> ! <lb/> (9) <lb/>where ˙ <lb/> m is the mass transfer between phases and Ω I the interfacial density . <lb/>By assuming that the mass transfer is proportional to the divergence of the <lb/>velocity, it is possible to build a family of models in which the mass transfer <lb/>˙ <lb/> m is expressed as [18] <lb/>˙ <lb/> m = <lb/> Ω l Ω v <lb/> Ω l ° Ω v <lb/> µ <lb/> 1 ° <lb/> c 2 <lb/> c 2 <lb/> wallis <lb/> ∂ <lb/> div V <lb/> (10) <lb/>where c wallis is the propagation velocity of acoustic waves without mass trans-<lb/>fer [25]. This speed of sound is expressed as a weighted harmonic mean of <lb/>speeds of sound of each phase: <lb/>1 <lb/> Ωc 2 <lb/> wallis <lb/> = <lb/> AE <lb/>Ω v c 2 <lb/> v <lb/> + <lb/>1 ° AE <lb/>Ω l c 2 <lb/> l <lb/> (11) <lb/>A first model was built using the speed of sound associated with a sinu-<lb/>soidal barotropic EOS [18, 19]. In the following, this model will be named <lb/>4-equation barotropic model. It involves one tunable parameter c baro inter-<lb/>preted as the minimum value of the speed of sound in the mixture. For all <lb/>simulations the value was set to 0.5 m/s. <lb/> 2.3. A new cavitation model <lb/> The new model is based on a mixture of stiAEened gas EOS. By assuming <lb/>the pressure equilibrium between phases, an expression for the pressure can <lb/>
34
+
35
+ <page>8 <lb/></page>
36
+
37
+ be deduced, function of the void ratio AE and the vapour mass fraction Y = <lb/> AEΩ v /Ω: <lb/>P (Ω, e, AE, Y ) = (∞(AE) ° 1)Ω(e ° q(Y )) ° ∞(AE)P 1 (AE) <lb/>( 1 2 ) <lb/>1 <lb/> ∞(AE) ° 1 <lb/>= <lb/> AE <lb/>∞ v ° 1 <lb/>+ <lb/>1 ° AE <lb/>∞ l ° 1 <lb/>(13) <lb/> q(Y ) = Y q v + (1 ° Y )q l <lb/> (14) <lb/> P 1 (AE) = <lb/> ∞(AE) ° 1 <lb/> ∞(AE) <lb/> ∑ <lb/> AE <lb/>∞ v <lb/> ∞ v ° 1 <lb/> P v <lb/> 1 + (1 ° AE) <lb/>∞ l <lb/> ∞ l ° 1 <lb/> P l <lb/> 1 <lb/> ∏ <lb/> (15) <lb/>By assuming the thermal equilibrium between phases, the mixture tempera-<lb/>ture is expressed as: <lb/> T (Ω, h, Y ) = <lb/> h l ° q l <lb/> C p l <lb/> = <lb/> h v ° q v <lb/> C p v <lb/> = <lb/> h ° q(Y ) <lb/> C p (Y ) <lb/>(16) <lb/> C p (Y ) = Y C p v + (1 ° Y )C p l <lb/> (17) <lb/>The speed of sound in the mixture can be expressed as a function of the <lb/>enthalpy of each phase (see Appendix A): <lb/> Ωc 2 = <lb/>1 <lb/> ∞ ° 1 <lb/> ∑ Ω v Ω l <lb/> (Ω l ° Ω v ) <lb/>(h v ° h l ) <lb/> ∏ <lb/> (18) <lb/>Enthalpies of pure phase h l and h v are computed with the mixture temper-<lb/>ature T . <lb/>The mass transfer term is activated when the local pressure P is smaller than <lb/>the vapour pressure P vap . This model will be named 4-equation SG model. <lb/>It does not involve any tunable parameter. <lb/> 2.4. The turbulence model <lb/> Various turbulence models are considered: the Smith k ° ` model (KL) [26], <lb/>the one-equation Spalart-Allmaras model (SA) [27] and the Jones-Launder <lb/>
38
+
39
+ <page>9 <lb/></page>
40
+
41
+ k ° &quot; model (KE) [28]. For a correct simulation of the re-entrant jet, the <lb/>Reboud eddy-viscosity limiter is added [29, 30, 31]. For comparisons with <lb/>the OpenFOAM solver, the Menter k ° ! SST model [32] is used, assuming <lb/>the validity of the Bradshaw assumption in a two-phase turbulent boundary <lb/>layer. <lb/> 2.5. Wall functions <lb/> For the modelling of flow close to the wall, a two-layer wall law approach is <lb/>used: <lb/> u + = y + <lb/> if y + &lt; 11.13 <lb/> u + = <lb/>1 <lb/> ∑ <lb/> ln y + + 5.25 if y + &gt; 11.13 <lb/> u + = <lb/> u <lb/>U ø <lb/> ; <lb/> y + = <lb/> yU ø <lb/> ∫ w <lb/> ; U 2 <lb/> ø = <lb/> ø w <lb/> Ω w <lb/> (19) <lb/>where ∑ = 0.41 is the von Karman constant and the subscript &apos;w&apos; is used for <lb/>a wall value. <lb/>We assume that wall functions are similar in a two-phase flow and in a <lb/>single-phase flow. For unsteady flows, the existence of a wall law is assumed <lb/>to be valid at each instant. These assumptions have been studied in [33] and <lb/>comparisons were proposed with a thin boundary layer approach. <lb/> 2.6. Numerics <lb/> The numerical simulations are carried out using an implicit CFD code based <lb/>on a finite-volume discretization. For the mean flow, the convective flux <lb/>density vector on a cell face is computed with the Jameson-Schmidt-Turkel <lb/>scheme [34]. The artificial viscosity includes a second-order dissipation term <lb/>
42
+
43
+ <page>10 <lb/></page>
44
+
45
+ D 2 and a fourth-order dissipation term D 4 , which involve two tunable pa-<lb/>rameters k (2) and k (4) . <lb/>The viscous terms are discretized by a second-order space-centered scheme. <lb/>For the turbulence transport equations, the upwind Roe scheme [35] is used <lb/>to obtain a more robust method. The second-order accuracy is obtained by <lb/>introducing a flux-limited dissipation [36]. <lb/>Time integration is achieved using the dual time stepping approach and a <lb/>low-cost implicit method consisting in solving, at each time step, a system <lb/>of equations arising from the linearization of a fully implicit scheme. The <lb/>derivative with respect to the physical time is discretized by a second-order <lb/>formula. <lb/>The numerical treatment of boundary conditions is based on the use of the <lb/>preconditioned characteristic relationships. More details are given in [19]. <lb/> 3. The OpenFOAM code <lb/> The OpenFOAM code is an open source code distributed by ESI Group. It <lb/>is based on an orientated object framework [37]. It provides a large variety of <lb/>RANS turbulence models and cavitation models. For cavitation modelling, <lb/>two ways are available: either to use an equation of state for the mixture or <lb/>to use a transport equation for the volume fraction of liquid. The last one is <lb/>retained for the present simulation. <lb/> 3.1. Cavitation model <lb/> The cavitation model reads: <lb/> @AE l <lb/> @t <lb/> + u j <lb/> @AE l <lb/> @x j <lb/> = S <lb/> (20) <lb/>
46
+
47
+ <page>11 <lb/></page>
48
+
49
+ where AE l is the volume fraction of liquid and S is the mass source term. <lb/>Following Kunz development [10], the source term is expressed as the sum of <lb/>a vaporisation term m v and a condensation term m c : <lb/> S = m v + m c <lb/> (21) <lb/>with: <lb/> m c = <lb/> Ω <lb/>Ω l Ω v <lb/> C c <lb/> Ω v <lb/> t 1 <lb/> AE 2 <lb/> Llim <lb/> max (P ° P vap ; 0) <lb/>max (P ° P vap ; 0.01 P vap ) <lb/>(22) <lb/> m v = <lb/> Ω <lb/>Ω l Ω v <lb/> C v <lb/> Ω v <lb/> 1 <lb/>2 <lb/> Ω l U 2 <lb/> 1 t 1 <lb/> min (P ° P vap ; P 0 ) <lb/>( 2 3 ) <lb/> C c , C v , U 1 and t 1 are constant set by the user, whereas P 0 and AE Llim are <lb/>included to avoid non physical values. Usually U 1 is set to the freestream <lb/>value, and t 1 represents a relaxation time not well defined in the literature. <lb/>For the present computations, the following values are specified: <lb/> C c = 10 ; C v = 8000 ; U 1 = 10.8 m/s ; t 1 = 0.005 s <lb/> (24) <lb/>The model uncertainty should be analyzed using non-intrusive stochastic <lb/>methods as presented in [38]. <lb/> 3.2. The turbulence model <lb/> The k ° ! SST model proposed by Menter [32] is used to solve the turbulent <lb/>kinetic energy and the specific dissipation with the standard values of the <lb/>diAEerent parameters. <lb/> 3.3. Numerics <lb/> The time derivatives are computed with the backward second order scheme. <lb/>Excepted for the volume fraction of liquid, the convective flux are discretised <lb/>
50
+
51
+ <page>12 <lb/></page>
52
+
53
+ with the Total Variation Diminishing (TVD) scheme named &apos;limitedLinear&apos; <lb/>specific to OpenFOAM with the parameter set to 1. <lb/>Whereas the momentum equation and the Poisson equation are treated im-<lb/>plicitly, the equation for the volume fraction of liquid is treated explicitly and <lb/>separately. To maintain the boundedness of the liquid volume fraction, the <lb/>Multidimensional Universal Limiter for Explicit Solution (MULES) is used <lb/>and a counter-gradient is introduced to reduce the diAEusion of the interface <lb/>and enhanced the numerical stability. The convective flux of the transport <lb/>equation for the volume fraction of liquid is computed with the van Leer <lb/>scheme. <lb/>The set of equation is solved using a prediction-correction approach coupling <lb/>the SIMPLE and PISO algorithm. <lb/> 4. Experimental and numerical parameters <lb/> 4.1. Experimental conditions <lb/> The Venturi was tested in the cavitation tunnel of the CREMHyG (Centre <lb/>d&apos;Essais de Machines Hydrauliques de Grenoble). It is characterized by a <lb/>divergence angle of 4 ± , illustrated in Figure 1. The edge forming the throat <lb/>of the Venturi is used to fix the separation point of the cavitation cavity. <lb/>This geometry is equipped with five probing holes to allow various mea-<lb/>surements such as the local void ratio, instantaneous local speed and wall <lb/>pressure (Figure 1). The velocity is evaluated as the most probable value <lb/>and the void ratio is obtained from the signal of the double optical probe <lb/>using a post-processing algorithm. The relative uncertainty on the void ratio <lb/>measurement was estimated at roughly 15% [39]. <lb/>
54
+
55
+ <page>13 <lb/></page>
56
+
57
+ The selected operating point is characterized by the following physical pa-<lb/>rameters [39]: <lb/> U inlet = 10.8 m/s, the inlet velocity <lb/> ae inlet = <lb/> P inlet ° P vap <lb/> 0.5ΩU 2 <lb/> inlet <lb/> &apos; 0.55, the cavitation parameter in the inlet section <lb/> T ref &apos; 293K, the reference temperature <lb/> L ref =252 mm, the reference length <lb/> Re L ref = <lb/> U inlet L ref <lb/> ∫ <lb/> = 2.7 10 6 , the Reynolds number <lb/>With these parameters, a cavity length L ranging from 70 mm to 85 mm was <lb/>obtained. The experimental views for this geometry show a relatively stable <lb/>cavity behaviour (see Fig. 2). The attached cavity length corresponding to <lb/>the end of the re-entrant jet is around 30-35 mm. For this geometry, no <lb/>periodic cycles with large shedding were observed. <lb/> 4.2. Mesh and numerical parameters <lb/> The grid is a H-type topology. It contains 251 nodes in the flow direction <lb/>and 62 nodes in the orthogonal direction. A special contraction of the mesh <lb/>is applied in the main flow direction just after the throat to better simulate <lb/>the two-phase flow area. The y + values of the mesh, at the center of the first <lb/>cell, vary between 12 and 27 for a non cavitating computation. <lb/>Unsteady computations are performed with the dual time stepping method <lb/>and are started from the non cavitating numerical solution. The numerical <lb/>parameters are: <lb/>-the dimensionless time step, ¢t § = <lb/> ¢tU inlet <lb/> L ref <lb/> = 4.88 10 °3 <lb/> -sub-iterations of the dual time stepping method, 100 <lb/>
58
+
59
+ <page>14 <lb/></page>
60
+
61
+ - the CFL number, 0.2 <lb/>-Jacobi iterations for the implicit stage, 15 <lb/>-the two coe±cients of the artificial dissipation, k (2) = 1 and k (4) = 0.045. <lb/> 5. Computational results <lb/> Various computations were performed by varying the cavitation model and <lb/>the turbulence model, summarized in Table (1). The goal was to obtain a <lb/>sheet whose time-averaged attached and total length varied around 30-35 <lb/>mm and 75-85 mm, respectively. The time of simulation is between 2 and <lb/>3 s. <lb/> 5.1. Limitation of the eddy viscosity <lb/> A key point to compute the unsteadiness of the cavitation pocket is linked to <lb/>the over-production of eddy viscosity by standard turbulence models. Previ-<lb/>ous simulations based on a three-equation model illustrated the importance <lb/>of using an eddy-viscosity limiter to capture the re-entrant jet dynamics [30]. <lb/>In this study, the Reboud limiter is added to the turbulence model. The <lb/>eAEect of this limiter using the Spalart-Allmaras is showed in Figure 3 where <lb/>are plotted the contours of the density gradient modulus (Schlieren-like vi-<lb/>sualizations). When the turbulent viscosity is reduced by the correction, the <lb/>length of attached cavity reaches the experimental value around 0.35 m and <lb/>vapour clouds appear. <lb/>We observed the same eAEect for other turbulence models and results are not <lb/>presented. <lb/>
62
+
63
+ <page>15 <lb/></page>
64
+
65
+ 5.2. Turbulence models comparison <lb/> Computations are done using the four-equation barotropic model associated <lb/>to three turbulence models in which the Reboud limiter is added. All nu-<lb/>merical values are obtained by a time-averaged statistical treatment over a <lb/>simulation time of 2 s. <lb/>Fig. 4 presents the void ratio and velocity profiles from stations 3 to 5. At <lb/>stations 1 and 2, inside the attached cavity, all simulations provide the same <lb/>results in close agreement with the experimental data and are not presented. <lb/>At station 3, the re-entrant jet is observed on the velocity measurement. All <lb/>simulations indicate a recirculating behaviour with a re-entrant jet extend-<lb/>ing through half the sheet thickness. For the void ratio profiles, the three <lb/>simulations give very close results. <lb/>At station 4 and 5 the whole of simulation capture the re-entrant jet charac-<lb/>terized by negative velocitites close to the wall. As regard to the void ratio <lb/>profiles, we observe an over-estimation at station 4 using the k ° ¡ models <lb/>and a better estimation using the Spalart-Allmaras model. At station 5, the <lb/> k ° ` turbulence model overestimates a little the length of the cavity. <lb/>The dimensionless wall pressure distribution P °P vap <lb/> Pvap <lb/> is plotted in Figure 5 <lb/>versus the distance x ° x inlet . The first five data are located inside the cavity <lb/>(where the void ratio and velocity profiles are measured). All models provide <lb/>a pressure distribution similar to the experimental measurements upstream <lb/>of the re-compression. <lb/>
66
+
67
+ <page>16 <lb/></page>
68
+
69
+ The Root Mean Square (RMS) wall pressure fluctuations are plotted in Fig-<lb/>ure 6 versus the distance x ° x inlet . The pressure fluctuation is divided by <lb/>the time-averaged pressure P av. Experimental data indicate an augmenta-<lb/>tion of pressure fluctuations at the end of the sheet cavity. All simulations <lb/>predict a peak of pressure fluctuations located close to the experimental ab-<lb/>scissa. The magnitude of the peak is overestimated by the computations, <lb/>but with the lack of experimental values in this area the comparison cannot <lb/>be precise. Nevertheless, downstream the cavity, the pressure fluctuations <lb/>are over-predicted using the Spalart-Allmaras and k ° &quot; turbulence models. <lb/>This phenomenon has already been observed in a previous study and can be <lb/>corrected by modifying the source term in the void ratio equation [19]. <lb/>The dynamic of cavitation pocket is also studied with the iso-lines of the Q-<lb/>criterion. Positive values of the Q-criterion, defined as the second invariant <lb/>of the velocity gradient tensor @u i <lb/> @x j <lb/> [40], <lb/> Q = <lb/>1 <lb/>2 <lb/> &quot; µ @u i <lb/> @x i <lb/> ∂ 2 <lb/> ° <lb/> @u i <lb/> @x j <lb/> @u j <lb/> @x i <lb/> # <lb/> (25) <lb/>are used to identify vortices and local rotational areas. A dimensionless quan-<lb/>tity is built using the inlet velocity and the reference length. Iso-lines levels <lb/>vary between 0.005 and 0.1. The results are illustrated in Figure 7. For <lb/>all turbulence models, the shear layer creates vortical clouds of cavitation, <lb/>which are convected by the mean flow. No significant discrepancies on the <lb/>flow dynamic are noticeable between the three simulations. <lb/>To conclude, similar results have been obtained with the three turbulence <lb/>
70
+
71
+ <page>17 <lb/></page>
72
+
73
+ models associated to an eddy viscosity limiter. The re-entrant jet is well <lb/>estimated but vapour ratio tends to be overestimated especially in the re-<lb/>circulating area (station 3 and 4). The influence of the turbulence model is <lb/>weak for this partial cavitation pocket. <lb/>In the next part, a mixture of stiAEened gas EOS is proposed to replace the <lb/>barotropic EOS and to investigate the cavitation model influence. <lb/> 5.3. Cavitation models comparison <lb/> For this study, the Spalart-Allmaras turbulence model is chosen and the two <lb/>formulations of 4-equation model are compared. <lb/>The dynamic of the cavitation sheets is proposed in Figure 8 where the con-<lb/>tours of the density gradient modulus are presented at two diAEerent times. <lb/>Both simulations propose similar dynamics and the generation of vapour <lb/>cloud shedding are clearly illustrated. <lb/>A more precise comparison is led by studying void ratio and velocity profiles, <lb/>which are presented in Figure 9 from stations 3 to 5. At region of attached <lb/>cavity sheet (stations 1 and 2), both simulations provide a well estimation of <lb/>the cavity thickness and results are not presented. <lb/>DiAEerences appear at station 3 where the four-equation SG cavitation model <lb/>provides a better prediction of the void ratio profile characterized by a sig-<lb/>nificant decrease close to the wall. However, by comparison with the exper-<lb/>imental data and the four-equation barotropic velocity profiles, this model <lb/>underestimates the recirculating flow in this area. At stations 4 and 5 the <lb/>four-equation SG cavitation model also reproduces the re-entrant jet on the <lb/>
74
+
75
+ <page>18 <lb/></page>
76
+
77
+ half bottom of the cavity but it over-predicts the vapour ratio in the remain-<lb/>ing part. The wall value of the void ratio is better simulated with the new <lb/>model. <lb/>As it can be observed in Figure 10 both models give a time-averaged wall <lb/>pressure distribution in good agreement with the experimental data. The <lb/>Root Mean Square (RMS) wall pressure fluctuations are plotted in Figure 11. <lb/>Both models predict a similar peak intensity at the same location. It appears <lb/>that the four-equation SG cavitation model provides a better estimation of <lb/>the pressure fluctuations decrease in the re-compression area. <lb/>This study reveals that the four-equation SG cavitation is well adapted to <lb/>simulate the behaviour of the cavitation pocket especially in the closure part <lb/>of the cavity. Moreover, this model presents the large advantage to not <lb/>involve tunable parameters. In the next part, this model will be used to <lb/>compare the LEGI&apos;s and OpenFOAM softwares. <lb/> 5.4. Comparison with OpenFOAM simulations <lb/> Comparisons between the LEGI solver and OpenFOAM are proposed on a <lb/>similar mesh using the k ° ! SST turbulence model. The considered cavita-<lb/>tion models are the 4-equation SG model and the Kunz model, respectively. <lb/>The time of simulation is about 2 s. <lb/>Time-averaged void ratio and velocity profiles are presented in Figure 12 from <lb/>stations 1 to 5. Inside the attached cavity sheet, at stations 1 and 2, the both <lb/>solvers estimate a well cavity thickness and composition. At stations 3 and <lb/>
78
+
79
+ <page>19 <lb/></page>
80
+
81
+ 4, the LEGI solver provides a better prediction of the void ratio decrease <lb/>due to the presence of a mixture in this area. The OpenFOAM computation <lb/>over-predicts the vapour quantity inside the cavity (95% instead of 70% at <lb/>station 3 and 90% instead of 30% at station 3). Moreover, the thickness of <lb/>the cavity is under-estimated and a pure liquid phase is simulated close to <lb/>the wall. At station 5, both the cavity thickness and the re-entrant jet are <lb/>badly simulated by the OpenFOAM software. Although the re-entrant phe-<lb/>nomenon is well observed in experiments, the OpenFOAM simulation does <lb/>not reproduce the recirculation area. <lb/>The time-averaged wall pressure distributions are compared in Figure 13. <lb/>The two models present a diAEerent re-compression area respectively located <lb/>downstream the experimental measures with the LEGI solver and upstream <lb/>with the OpenFOAM solver. The rate of re-compression is also under-<lb/>estimated by the OpenFOAM solver. <lb/>A study of the Root Mean Square (RMS) wall pressure fluctuations is pro-<lb/>posed in Figure 14. The peak of fluctuations position is varying among the <lb/>case. The OpenFOAM solver provides a maximum of fluctuations located <lb/>upstream the closure area of the cavity sheet (x ° x i º 0.2 in experiments), <lb/>whereas it is predicted downstream by the LEGI solver. <lb/>To conclude, the topology of the cavitation pocket marks large discrepancies <lb/>between the two softwares, especially as regards the re-entrant jet develop-<lb/>ment. Maybe a better calibration of the production and evaporation con-<lb/>
82
+
83
+ <page>20 <lb/></page>
84
+
85
+ stants appearing in the mass transfer formulation should improve the results <lb/>for this test case. <lb/> 6. Conclusion <lb/> An aperiodic partial cavitation pocket has been studied in a 2D venturi con-<lb/>figuration by numerical one-fluid unsteady RANS simulations. Numerical <lb/>results have been compared with experimental data concerning the void ra-<lb/>tio, streamwise velocity, wall pressure and wall pressure fluctuations. <lb/>First, calculations have been carried out with an in-house code. The one-fluid <lb/>RANS equations have been successively coupled with the Smith k ° `, the <lb/>Spalart-Allmaras, the Jones-Launder k ° &quot; and the Menter k ° ! SST tur-<lb/>bulence models. Two cavitation models using an explicit formulation of the <lb/>mass transfer between phases have also been compared: the first one based <lb/>on a barotropic EOS and the second one based on a mixture of stiAEened gas <lb/>EOS. Comparisons revealed that similar results were obtained using diAEerent <lb/>turbulence models when an eddy viscosity limiter is introduced. For this test <lb/>case, the turbulence model influence is weak. Then, a cavitation model com-<lb/>parison was performed using the Spalart-Allmaras turbulence model. The <lb/>new four-equation model has improved the re-entrant jet simulation down-<lb/>stream the attached cavity by diminishing the void ratio values close to the <lb/>wall. A better estimation of the pressure fluctuations is also provided by this <lb/>model. A very positive point is the free-parameter formulation of the source <lb/>term, which avoids poisonous calibration problems. <lb/>Secondly, simulations has been carried out with the OpenFOAM software us-<lb/>
86
+
87
+ <page>21 <lb/></page>
88
+
89
+ ing another formulation of void ratio transport equation cavitation models. <lb/>Calculations were performed on a similar mesh using the Menter SST tur-<lb/>bulence model. Large discrepancies appeared between the two solvers and <lb/>the OpenFOAM simulations were not able to reproduce the re-entrant jet <lb/>phenomenon. Maybe it is due to a calibration problem of the mass transfer <lb/>model involving two tunable parameters. The new LEGI model is therefore <lb/>very attractive to simulate such partial cavitation pockets. <lb/>Additional works are in progress to pursue comparative analyses between <lb/>turbulence and cavitation models and to extend the formulation with ther-<lb/>modynamic eAEects. <lb/> Appendix <lb/> Appendix A: the speed of sound in a mixture of stiAEened gas <lb/> Starting from the usual thermodynamic relation <lb/> de = T ds + <lb/> P <lb/> Ω 2 dΩ <lb/> or <lb/> d(Ωe) = ΩT ds + hdΩ <lb/> (26) <lb/>And with the diAEerential of Ωe: <lb/> d(Ωe) = <lb/> µ @Ωe <lb/>@Ω <lb/> ∂ <lb/> P <lb/> dΩ + <lb/> µ @Ωe <lb/>@P <lb/> ∂ <lb/> Ω <lb/> dP <lb/> (27) <lb/>We can obtain the diAEerential of the pressure P : <lb/> µ @Ωe <lb/>@P <lb/> ∂ <lb/> Ω <lb/> dP = ΩT ds + <lb/> ∑ <lb/> h ° <lb/> µ @Ωe <lb/>@Ω <lb/> ∂ <lb/> P <lb/> ∏ <lb/> dΩ <lb/> (28) <lb/>We deduce an expression of the speed of sound: <lb/> c 2 = <lb/> µ @P <lb/>@Ω <lb/> ∂ <lb/> s <lb/> = <lb/> h ° <lb/> ≥ <lb/> @Ωe <lb/>@Ω <lb/> ¥ <lb/> P <lb/> ° @Ωe <lb/>@P <lb/> ¢ <lb/> Ω <lb/> (29) <lb/>
90
+
91
+ <page>22 <lb/></page>
92
+
93
+ With the stiAEened gas EOS, we have: <lb/> µ @Ωe <lb/> @P <lb/> ∂ <lb/> Ω <lb/> = AE <lb/> µ @Ω v e v <lb/> @P <lb/> ∂ <lb/> Ω <lb/> + (1 ° AE) <lb/> µ @Ω l e l <lb/> @P <lb/> ∂ <lb/> Ω <lb/> = <lb/>1 <lb/> ∞ ° 1 <lb/> µ @Ωe <lb/>@Ω <lb/> ∂ <lb/> P <lb/> = <lb/> @ <lb/>@Ω <lb/> ∑ <lb/> AE <lb/> µ P <lb/> ∞ v ° 1 <lb/>+ Ω v q v + <lb/> ∞ v P v <lb/> 1 <lb/> ∞ v ° 1 <lb/> ∂ <lb/> + (1 ° AE) <lb/> µ P <lb/> ∞ l ° 1 <lb/>+ Ω l q l + <lb/> ∞ l P l <lb/> 1 <lb/> ∞ l ° 1 <lb/> ∂∏ <lb/> = <lb/> Ω l h l ° Ω v h v <lb/> Ω v ° Ω v <lb/> Finally, the speed of sound is: <lb/> Ωc 2 = Ω <lb/> µ @P <lb/>@Ω <lb/> ∂ <lb/> s <lb/> = (∞ ° 1) <lb/> ∑ Ω v Ω l <lb/> (Ω l ° Ω v ) <lb/>(h v ° h l ) <lb/> ∏ <lb/> (30) <lb/></body>
94
+
95
+ <page>23 <lb/></page>
96
+
97
+ <listBibl> [1] D. Schmidt, C. Rutland, M. Corradini, A fully compressible, two-<lb/>dimensional model of small, high-speed, cavitating nozzles, Atomization <lb/>and Sprays 9 (1999) 255–276. <lb/>[2] T. Liu, B. Khoo, W. Xie, Isentropic one-fluid modelling of unsteady <lb/>cavitating flow, Journal of Computational Physics 201 (1) (2004) 80– <lb/>108. <lb/>[3] T. Barberon, P. Helluy, Finite volume simulation of cavitating flows, <lb/>Computers &amp; Fluids 34 (7) (2005) 832–858. <lb/>[4] E. Sinibaldi, F. Beux, M. Salvetti, A numerical method for 3D barotropic <lb/>flows in turbomachinery, Flow Turbulence Combustion 76 (2006) 371– <lb/>381. <lb/>[5] S. Schmidt, I. Sezal, G. Schnerr, Compressible simulation of high-speed <lb/>hydrodynamics with phase change, in: European Conference on Com-<lb/>putational Fluid Dynamics ECCOMAS 2006, Delft, The Netherlands, <lb/>2006. <lb/>[6] W. Xie, T. Liu, B. Khoo, Application of a one-fluid model for large <lb/>scale homogeneous unsteady cavitation: The modified Schmidt model, <lb/>Computers &amp; Fluids 35 (2006) 1177–1192. <lb/>[7] E. Goncalves, R. F. Patella, Numerical simulation of cavitating flows <lb/>with homogeneous models, Computers &amp; Fluids 38 (9) (2009) 1682– <lb/>1696. <lb/>[8] M. Bilanceri, F. Beux, M. Salvetti, An implicit low-diAEusive HLL scheme <lb/>
98
+
99
+ <page>24 <lb/></page>
100
+
101
+ with complete time linearization: application to cavitating barotropic <lb/>flows, Computers &amp; Fluids 39 (2010) 1990–2006. <lb/>[9] C. Merkle, J. Feng, P. Buelow, Computation modeling of the dynam-<lb/>ics of sheet cavitation, in: 3 rd International Symposium on Cavitation <lb/>CAV1998, Grenoble, France, 1998. <lb/>[10] R. Kunz, D. Boger, D. Stinebring, T. Chyczewski, J. Lindau, H. Gibel-<lb/>ing, S. Venkateswaran, T. Govindan, A preconditioned Navier-Stokes <lb/>method for two-phase flows with application to cavitation prediction, <lb/>Computers &amp; Fluids 29 (8) (2000) 849–875. <lb/>[11] G. Schnerr, J. Sauer, Physical and numerical modeling of unsteady cav-<lb/>itation dynamics, in: 4th Int. Conference on Multiphase Flow ICMF01, <lb/>New Orleans, USA, 2001. <lb/>[12] I. Senocak, W. Shyy, A pressure-based method for turbulent cavitating <lb/>flow computations, Journal of Computational Physics 176 (2) (2002) <lb/>363–383. <lb/>[13] A. Singhal, M. Athavale, H. Li, Y. Jiang, Mathematical basis and vali-<lb/>dation of the full cavitation model, Journal of Fluids Engineering 124 (3) <lb/>(2002) 617–624. <lb/>[14] Y. Utturkar, J. Wu, G. Wang, W. Shyy, Recent progress in modelling of <lb/>cryogenic cavitation for liquid rocket propulsion, Progress in Aerospace <lb/>Sciences 41 (2005) 558–608. <lb/>
102
+
103
+ <page>25 <lb/></page>
104
+
105
+ [15] E. Goncalves, R. F. Patella, Constraints on equation of state for cavitat-<lb/>ing flows with thermodynamic eAEects, Applied Math. and Computation <lb/>217 (2011) 5095–5102. <lb/>[16] P. Downar-Zapolski, Z. Bilicki, L. Bolle, J. Franco, The non-equilibrium <lb/>relaxation model for one-dimensional flashing liquid flow, Int. Journal <lb/>of Multiphase Flow 22 (3) (1996) 473–483. <lb/>[17] P. Helluy, N. Seguin, Relaxation models of phase transition flows, Math-<lb/>ematical Modelling and Numerical Analysis 40 (2) (2006) 331–352. <lb/>[18] E. Goncalves, Numerical study of expansion tube problems: Toward the <lb/>simulation of cavitation, Computers &amp; Fluids 72 (2013) 1–19. <lb/>[19] E. Goncalves, B. Charriere, Modelling for isothermal cavitation with a <lb/>four-equation model, International Journal of Multiphase Flow 59 (2014) <lb/>54–72. <lb/>[20] R. Bensow, G. Bark, Simulating cavitating flows with LES in openfoam, <lb/>in: V ECCOMAS CFD, Lisbon, Portugal, June 2010. <lb/>[21] S. Park, S. Rhee, Computational analysis of turbulent super-cavitating <lb/>flow around a two-dimensional wedge-shaped cavitator geometry, Com-<lb/>puters &amp; Fluids 70 (2012) 73–85. <lb/>[22] Z. Shang, Numerical investigations of supercavitation around blunt bod-<lb/>ies of submarine shape, Applied Mathematical Modelling 37 (2013) <lb/>8836–8845. <lb/>
106
+
107
+ <page>26 <lb/></page>
108
+
109
+ [23] J. Decaix, E. Goncalves, Compressible eAEects modelling in turbulent <lb/>cavitating flows, European J. of Mechanics B/Fluids 39 (2013) 11–31. <lb/>[24] R. Saurel, F. Petitpas, R. Abgrall, Modelling phase transition in <lb/>metastable liquids: application to cavitating and flashing flows, Journal <lb/>of Fluid Mechanics 607 (2008) 313–350. <lb/>[25] G. Wallis, One-dimensional two-phase flow, New York: McGraw-Hill <lb/>(1967). <lb/>[26] B. Smith, A near wall model for the k°l two equation turbulence model, <lb/>in: AIAA 94–2386, 25 sh Fluid Dynamics Conference – Colorado Springs, <lb/>Colorado, 1994. <lb/>[27] P. Spalart, S. Allmaras, A one-equation turbulence model for aerody-<lb/>namic flows, La Recherche Aérospatiale (1) (1994) 5–21. <lb/>[28] W. Jones, B. Launder, The prediction of laminarization with a two-<lb/>equation model of turbulence, Int. J. Heat and Mass Transfer 15 (1972) <lb/>301–314. <lb/>[29] J.-L. Reboud, B. Stutz, O. Coutier, Two-phase flow structure of cavita-<lb/>tion: experiment and modelling of unsteady eAEects, in: 3 rd International <lb/>Symposium on Cavitation CAV1998, Grenoble, France, 1998. <lb/>[30] E. Goncalves, Numerical study of unsteady turbulent cavitating flows, <lb/>European Journal of Mechanics B/Fluids 30 (1) (2011) 26–40. <lb/>[31] J. Decaix, E. Goncalves, Time-dependent simulation of cavitating flow <lb/>
110
+
111
+ <page>27 <lb/></page>
112
+
113
+ with k ° ` turbulence models, Int. Journal for Numerical Methods in <lb/>Fluids 68 (2012) 1053–1072. <lb/>[32] F. Menter, Two-equation eddy-viscosity turbulence models for engineer-<lb/>ing applications, AIAA Journal 32 (8) (1994) 1598–1605. <lb/>[33] E. Goncalves, J. Decaix, Wall model and mesh influence study for partial <lb/>cavities, European Journal of Mechanics B/Fluids 31 (1) (2012) 12–29. <lb/>[34] A. Jameson, W. Schmidt, E. Turkel, Numerical solution of the Euler <lb/>equations by finite volume methods using Runge-Kutta time stepping <lb/>schemes, in: AIAA Paper 81–1259, 1981. <lb/>[35] P. Roe, Approximate Riemann solvers, parameters vectors, and diAEer-<lb/>ence schemes, Journal of Computational Physics 43 (1981) 357–372. <lb/>[36] S. Tatsumi, L. Martinelli, A. Jameson, Flux-limited schemes for the <lb/>compressible Navier-Stokes equations, AIAA Journal 33 (2) (1995) 252– <lb/>261. <lb/>[37] H. Weller, G. Tabor, H. Jasak, C. Fureby, A tensorial approach to <lb/>computational continuum mechanics using objected-oriented techniques, <lb/>Computers in Physics 12 (1998) 620–631. <lb/>[38] M. Rodio, P. Congedo, Robust analysis of cavitating flows in Venturi <lb/>tube, European J. of Mechanics B/Fluids in press. <lb/>[39] S. Barre, J. Rolland, G. Boitel, E. Goncalves, R. F. Patella, Experiments <lb/>and modelling of cavitating flows in Venturi: attached sheet cavitation, <lb/>European Journal of Mechanics B/Fluids 28 (2009) 444–464. <lb/>
114
+
115
+ <page>28 <lb/></page>
116
+
117
+ [40] J. Hunt, C. Wray, P. Moin, Eddies, streams, and convergence zones in <lb/>turbulent flows, Tech. rep., Center for Turbulence Research, CTR-S88 <lb/>(1988). <lb/></listBibl>
118
+
119
+ <page>29 <lb/></page>
120
+
121
+ <body> Table 1: Unsteady computations, 4 ± Venturi. <lb/> cav. model turb. model <lb/> ae inlet attached and total sheet length (m) <lb/>4-eqt baro SA <lb/>0.59 <lb/>0.09 -0.11 <lb/>4-eqt baro SA + Reboud 0.60 <lb/>0.035 -0.076 <lb/>4-eqt baro KL + Reboud 0.61 <lb/>0.038 -0.085 <lb/>4-eqt baro KE + Reboud 0.61 <lb/>0.029 -0.078 <lb/>4-eqt SG <lb/>SA + Reboud 0.575 <lb/>0.038 -0.085 <lb/>4-eqt SG <lb/>KW SST <lb/>0.595 <lb/>0.030 -0.10 <lb/>4-eqt Foam KW SST <lb/>0.585 <lb/>0.070 <lb/>
122
+
123
+ <page>30 <lb/></page>
124
+
125
+ Figure 1: Schematic view of the 4 ± Venturi profile. <lb/>
126
+
127
+ <page> 31 <lb/></page>
128
+
129
+ Figure 2: Photograph of the cavitation pocket. <lb/>
130
+
131
+ <page> 32 <lb/></page>
132
+
133
+ Figure 3: Contours of the density gradient modulus, without limiter (top) and with limiter <lb/>(bottom). <lb/>
134
+
135
+ <page> 33 <lb/></page>
136
+
137
+ alpha <lb/> y (m) <lb/> 0 <lb/> 0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/> SA <lb/>KL <lb/>KE <lb/>EXPE <lb/> U (m/s) <lb/>y (m) <lb/> 0 <lb/> 5 <lb/>10 <lb/>15 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> SA <lb/>KL <lb/>KE <lb/>EXPE <lb/> alpha <lb/> y (m) <lb/> 0 <lb/> 0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> U (m/s) <lb/>y (m) <lb/> 0 <lb/> 5 <lb/>10 <lb/>15 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> alpha <lb/> y (m) <lb/> 0 <lb/> 0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> U (m/s) <lb/>y (m) <lb/> 0 <lb/> 5 <lb/>10 <lb/>15 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> Figure 4: Time-averaged velocity (right) and void ratio (left) profiles from station 3 to 5, <lb/>4-equation barotropic model. <lb/>
138
+
139
+ <page> 34 <lb/></page>
140
+
141
+ x-x i (m) <lb/>(P-Pvap)/Pvap <lb/> 0.1 <lb/> 0.15 <lb/>0.2 <lb/>0.25 <lb/>0.3 <lb/>0 <lb/>5 <lb/>10 <lb/>15 <lb/> EXPE <lb/>SA <lb/>KL <lb/>KE <lb/> Figure 5: Dimensionless time-averaged wall pressure evolution, 4-equation barotropic <lb/>model. <lb/>
142
+
143
+ <page> 35 <lb/></page>
144
+
145
+ x-x i (m) <lb/>P&apos; rms / Pav <lb/> 0.15 <lb/> 0.2 <lb/>0.25 <lb/>0.3 <lb/>0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>1.2 <lb/> EXPE <lb/>SA <lb/>KL <lb/>KE <lb/> Figure 6: RMS wall pressure fluctuations, 4-equation barotropic model. <lb/>
146
+
147
+ <page> 36 <lb/></page>
148
+
149
+ x(m) <lb/> y(m) <lb/> 0 <lb/> 0.02 <lb/>0.04 <lb/>0.06 <lb/>0.08 <lb/>0.1 <lb/>-0.005 <lb/>0 <lb/>0.005 <lb/>0.01 <lb/>0.015 <lb/>0.02 <lb/>0.025 <lb/>0.03 <lb/>0.035 <lb/>0.04 <lb/> SA model <lb/> x(m) <lb/> y(m) <lb/> 0 <lb/> 0.02 <lb/>0.04 <lb/>0.06 <lb/>0.08 <lb/>0.1 <lb/>-0.005 <lb/>0 <lb/>0.005 <lb/>0.01 <lb/>0.015 <lb/>0.02 <lb/>0.025 <lb/>0.03 <lb/>0.035 <lb/>0.04 <lb/> KL model <lb/>KE model <lb/> Figure 7: Isolines of the dimensionless Q-criterion. Turbulence models comparison, 4-<lb/>equation barotropic model. <lb/>
150
+
151
+ <page> 37 <lb/></page>
152
+
153
+ x(m) <lb/> y(m) <lb/> 0 <lb/> 0.02 <lb/>0.04 <lb/>0.06 <lb/>0.08 <lb/>0.1 <lb/>0.12 <lb/>0 <lb/>0.02 <lb/>0.04 <lb/>Grad-rho: 5000 21111.1 37222.2 53333.3 69444.4 85555.6 101667 117778 133889 150000 <lb/> x(m) <lb/> y(m) <lb/> 0 <lb/> 0.02 <lb/>0.04 <lb/>0.06 <lb/>0.08 <lb/>0.1 <lb/>0.12 <lb/>0 <lb/>0.02 <lb/>0.04 <lb/>Grad-rho: 5000 21111.1 37222.2 53333.3 69444.4 85555.6 101667 117778 133889 150000 <lb/> x(m) <lb/> y(m) <lb/> 0 <lb/> 0.02 <lb/>0.04 <lb/>0.06 <lb/>0.08 <lb/>0.1 <lb/>0.12 <lb/>0 <lb/>0.02 <lb/>0.04 <lb/>Grad-rho: 5000 21111.1 37222.2 53333.3 69444.4 85555.6 101667 117778 133889 150000 <lb/> x(m) <lb/> y(m) <lb/> 0 <lb/> 0.02 <lb/>0.04 <lb/>0.06 <lb/>0.08 <lb/>0.1 <lb/>0.12 <lb/>0 <lb/>0.02 <lb/>0.04 <lb/>Grad-rho: 5000 21111.1 37222.2 53333.3 69444.4 85555.6 101667 117778 133889 150000 <lb/> Figure 8: Density gradient modulus (kg.m °4 ) for 4-eqt SG model (left) and 4-eqt <lb/>barotropic model (right) <lb/>
154
+
155
+ <page> 38 <lb/></page>
156
+
157
+ alpha <lb/> y (m) <lb/> 0 <lb/> 0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/> 4-eqt baro <lb/>4-eqt SG <lb/>EXPE <lb/> U (m/s) <lb/>y (m) <lb/> 0 <lb/> 5 <lb/>10 <lb/>15 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> 4-eqt baro <lb/>4-eqt SG <lb/>EXPE <lb/> alpha <lb/> y (m) <lb/> 0 <lb/> 0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> U (m/s) <lb/>y (m) <lb/> 0 <lb/> 5 <lb/>10 <lb/>15 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> alpha <lb/> y (m) <lb/> 0 <lb/> 0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> U (m/s) <lb/>y (m) <lb/> 0 <lb/> 5 <lb/>10 <lb/>15 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> Figure 9: Time-averaged velocity (right) and void ratio (left) profiles from station 3 to 5, <lb/>4-equation models comparison. <lb/>
158
+
159
+ <page> 39 <lb/></page>
160
+
161
+ x-x i (m) <lb/>(P-Pvap)/Pvap <lb/> 0.1 <lb/> 0.15 <lb/>0.2 <lb/>0.25 <lb/>0.3 <lb/>0 <lb/>5 <lb/>10 <lb/>15 <lb/> EXPE <lb/>4-eqt baro <lb/>4-eqt SG <lb/> Figure 10: Dimensionless time-averaged wall pressure evolution, 4-equation models com-<lb/>parison. <lb/>
162
+
163
+ <page> 40 <lb/></page>
164
+
165
+ x-x i (m) <lb/>P&apos; rms / Pav <lb/> 0.15 <lb/> 0.2 <lb/>0.25 <lb/>0.3 <lb/>0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>1.2 <lb/> EXPE <lb/>4-eqt baro <lb/>4-eqt SG <lb/> Figure 11: RMS wall pressure fluctuations, 4-equation models comparison. <lb/>
166
+
167
+ <page> 41 <lb/></page>
168
+
169
+ alpha <lb/> y (m) <lb/> 0 <lb/> 0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>0.0005 <lb/>0.001 <lb/>0.0015 <lb/>0.002 <lb/> 4-eqt SG <lb/>OpenFOAM <lb/>EXPE <lb/> U (m/s) <lb/>y (m) <lb/> 0 <lb/> 5 <lb/>10 <lb/>15 <lb/>0 <lb/>0.0005 <lb/>0.001 <lb/>0.0015 <lb/>0.002 <lb/> 4-eqt SG <lb/>OpenFOAM <lb/>EXPE <lb/> alpha <lb/> y (m) <lb/> 0 <lb/> 0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/> U (m/s) <lb/>y (m) <lb/> 0 <lb/> 5 <lb/>10 <lb/>15 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/> alpha <lb/> y (m) <lb/> 0 <lb/> 0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/> U (m/s) <lb/>y (m) <lb/> 0 <lb/> 5 <lb/>10 <lb/>15 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> alpha <lb/> y (m) <lb/> 0 <lb/> 0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> U (m/s) <lb/>y (m) <lb/> 0 <lb/> 5 <lb/>10 <lb/>15 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> alpha <lb/> y (m) <lb/> 0 <lb/> 0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> U (m/s) <lb/>y (m) <lb/> 0 <lb/> 5 <lb/>10 <lb/>15 <lb/>0 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0.008 <lb/>0.01 <lb/> Figure 12: Time-averaged velocity (right) and void ratio (left) profiles from station 1 to <lb/>5, LEGI solver versus openFOAM. <lb/>
170
+
171
+ <page> 42 <lb/></page>
172
+
173
+ x-x i (m) <lb/>(P-Pvap)/Pvap <lb/> 0.1 <lb/> 0.15 <lb/>0.2 <lb/>0.25 <lb/>0.3 <lb/>0 <lb/>5 <lb/>10 <lb/>15 <lb/> EXPE <lb/>4-eqt SG <lb/>openFOAM <lb/> Figure 13: Dimensionless time-averaged wall pressure evolution, LEGI solver versus open-<lb/>FOAM. <lb/>
174
+
175
+ <page> 43 <lb/></page>
176
+
177
+ x-x i (m) <lb/>P&apos; rms / Pav <lb/> 0.15 <lb/> 0.2 <lb/>0.25 <lb/>0.3 <lb/>0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>1.2 <lb/> EXPE <lb/>4-eqt SG <lb/>openFOAM <lb/> Figure 14: RMS wall pressure fluctuations, LEGI solver versus openFOAM. <lb/></body>
178
+
179
+ <page> 44 </page>
180
+
181
+
182
+ </text>
183
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/ENEURO.0140-18.2018.training.segmentation.tei.xml ADDED
@@ -0,0 +1,119 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>Sensory and Motor Systems <lb/>Gas Anesthesia Impairs Peripheral Auditory <lb/>Sensitivity in Barn Owls (Tyto alba) <lb/>Nadine Thiele, 1 and Christine Köppl 1,2 <lb/>https://doi.org/10.1523/ENEURO.0140-18.2018 <lb/>1 Department of Neuroscience, School of Medicine and Health Sciences, Carl von Ossietzky University Oldenburg, <lb/>26129 Oldenburg, Germany and 2 Cluster of Excellence &quot;Hearing4all&quot; and Research Center Neurosensory Science, <lb/>Carl von Ossietzky University Oldenburg, 26129 Oldenburg, Germany <lb/>Visual Abstract <lb/>&lt; 1 .5 <lb/>k H z <lb/>1 .5 -3 <lb/>k H z <lb/>3 -4 .5 <lb/>k H z <lb/>4 .5 -6 <lb/>k H z <lb/>&gt; 6 k H z <lb/>Ket-terminal <lb/>Isofl-terminal <lb/>CF bin <lb/>Threshold re age-matched CAP (dB) <lb/>30 <lb/>20 <lb/>10 <lb/> 0 <lb/>-10 <lb/>-20 <lb/>-30 <lb/>-40 <lb/>-50 <lb/>Frequency (kHz) <lb/>) <lb/>L <lb/>P <lb/>S <lb/>B <lb/>d <lb/>( <lb/>d <lb/>l <lb/>o <lb/>h <lb/>s <lb/>e <lb/>r <lb/>h <lb/>t <lb/>R <lb/>B <lb/>A <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>6 <lb/>-2 <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>-1 <lb/>-2 <lb/>-3 <lb/>-4 <lb/>Time (ms) <lb/>) <lb/>V <lb/>µ <lb/>( <lb/>e <lb/>d <lb/>u <lb/>t <lb/>i <lb/>l <lb/>p <lb/>m <lb/>a <lb/>R <lb/>B <lb/>A <lb/>Ket-ABR <lb/>Isofl-ABR <lb/>Sevofl-ABR <lb/>Auditory nerve single-unit recordings were obtained from two groups of young barn owls (age, between <lb/>posthatching days 11 and 86) in terminal experiments under two different anesthetic regimes: ketamine (6 -11 <lb/>mg/kg) plus xylazine (ϳ2 mg/kg); or isoflurane (1-1.5%) in oxygen, delivered via artificial respiration. In a second <lb/>series of minimally invasive experiments, auditory brainstem responses (ABRs) were recorded in the same four <lb/>Significance Statement <lb/>Anesthesia and analgesia are necessary for most invasive experiments. Their effects are also a concern for <lb/>studying normal neural and sensory functions. We show a significant deterioration of hearing sensitivity of <lb/>the auditory nerve under gas anesthesia (isoflurane or sevoflurane), compared with injection anesthesia with <lb/>ketamine/xylazine, in barn owls. This generalizes similar findings across birds and mammals, and suggests <lb/>that while inhalants are widely recommended as safe and easy-to-use anesthetics in veterinary contexts, <lb/>they should only be used with great caution in auditory neurophysiology, even at the most peripheral level. <lb/>Future important questions are whether the deterioration of sensitivity at the periphery generalizes to other <lb/>senses and what the precise mechanisms are that determine the species-specific extent of sensitivity loss. <lb/> New Research <lb/>September/October 2018, 5(5) e0140-18.2018 1-14 <lb/>adult barn owls (Tyto alba; age, between 5 and 32 months) under three different anesthetic protocols: ketamine <lb/>(10 mg/kg) plus xylazine (3 mg/kg), isoflurane (1-1.5%), and sevoflurane (2-3%) in carbogen. Finally, the ABR <lb/>measurements on adult owls were repeated in terminal experiments including more invasive procedures such as <lb/>artificial respiration and higher isoflurane dosage. The main finding was a significant deterioration of auditory <lb/>sensitivity in barn owls under gas anesthesia, at the level of the auditory nerve (i.e., a very peripheral level of the <lb/>auditory system). The effect was drastic in the young animals that experienced threshold elevations in auditory <lb/>nerve single-unit responses of Ն20 dB. ABR thresholds assessed repeatedly in experiments on adult owls were <lb/>also significantly higher under isoflurane and sevoflurane, on average by 7 and 15 dB, compared with ketamine/ <lb/>xylazine. This difference already occurred with minimal dosages and was reversibly enlarged with increased <lb/>isoflurane concentration. Finally, there was evidence for confounding detrimental effects associated with artificial <lb/>respiration over many hours, which suggested oxygen toxicity. <lb/>Key words: auditory brainstem response; avian; bird; isoflurane; ketamine; physiology <lb/></front>
8
+
9
+ <body>Introduction <lb/>Anesthesia and analgesia are necessary components of <lb/>most invasive physiological experiments. Yet, their very <lb/>site of action, the nervous system, is a constant concern <lb/>for neuroscientists who strive to study normal neural and <lb/>sensory functions. Different anesthetic agents act on dif-<lb/>ferent neural targets and for some common anesthetic <lb/>agents these are known in considerable detail at the <lb/>molecular level (Lukasik and Gillies, 2003; Sonner et al., <lb/>2003; Rudolph and Antkowiak, 2004). However, due to <lb/>the highly interactive nature of the intact nervous system, <lb/>it remains difficult to predict their effect on a given neu-<lb/>ronal population under in vivo conditions (Antkowiak, <lb/>2001; Vahle-Hinz and Detsch, 2002; Windels, 2006). <lb/>In the present study, the anesthetic regime was <lb/>changed during an ongoing experimental series on the <lb/>development of auditory nerve responses in the barn owl <lb/>(Tyto alba). Previous auditory research in several labora-<lb/>tories had successfully used a combined ketamine/xyla-<lb/>zine or ketamine/diazepam injection anesthesia in both <lb/>adult and young owls (Cohen and Knudsen, 1995; Köppl, <lb/>1997; Keller and Takahashi, 2005; Bremen et al., 2007; <lb/>Köppl and Nickel, 2007). Nevertheless, a change to iso-<lb/>flurane inhalant anesthesia was recommended by the <lb/>consulting veterinarians, citing animal welfare concerns <lb/>(Varner et al., 2004). Isoflurane and related inhalants (ha-<lb/>logenated ethers, e.g., sevoflurane and desflurane; Cam-<lb/>pagna et al., 2003), are commonly recommended as the <lb/>first choice for veterinary procedures on a wide variety of <lb/>species, including birds (Korbel, 1998; Gunkel and Lafor-<lb/>tune, 2005; Lierz and Korbel, 2012; Raftery, 2013). Fre-<lb/>quently cited advantages are the rapid and easy control of <lb/>anesthetic depth, stable anesthetic state for lengthy pro-<lb/>cedures, and rapid recovery (Keegan, 2005). <lb/>Although no specific reports were available about the <lb/>effects of isoflurane on neural responses in birds, there <lb/>was no reason to expect a deterioration in auditory nerve <lb/>responses with gas anesthetic agents relative to ket-<lb/>amine/xylazine. In starlings and chickens, halothane, an-<lb/>other inhalant anesthetic, and ketamine/xylazine were <lb/>tested and compared for their effects on cochlear re-<lb/>sponses. Both types of anesthesia were found to act in an <lb/>equally depressive fashion on otoacoustic emissions (pro-<lb/>duced by the hair cells of the inner ear; Kettembeil et al., <lb/>1995). Similarly, the effects of ketamine or isoflurane on <lb/>auditory nerve responses in the Tokay gecko were mod-<lb/>erately depressive and differed little from each other <lb/>(Dodd and Capranica, 1992). In a bat, isoflurane had no <lb/>adverse effects on otoacoustic emissions (Drexl et al., <lb/>2004), suggesting that it could even be preferable to <lb/>ketamine, depending on the species. Unfortunately, it <lb/>became clear during the course of this study that the <lb/>change to isoflurane correlated with degraded hearing <lb/>sensitivity in the barn owl. Studies in several mammalian <lb/>species subsequently reported similar effects on cochlear <lb/>responses (Stronks et al., 2010; Cederholm et al., 2012; <lb/>Ruebhausen et al., 2012). In order to directly compare <lb/>different anesthetic agents and protocols, including pro-<lb/>longed anesthesia and artificial respiration, which are <lb/>common in invasive neurophysiology, a dedicated series <lb/>of experiments on adult barn owls was finally carried out. <lb/>In addition to ketamine/xylazine and isoflurane, the more <lb/>recently introduced inhalant sevoflurane was then also <lb/>included for testing. Sevoflurane shows a pharmacology <lb/>similar to isoflurane, but induction and recovery from <lb/>anesthesia are even more rapid and, as an additional <lb/>benefit, it is less of an irritant to the respiratory tract <lb/>(O&apos;Keeffe and Healy, 1999; Preckel and Bolten, 2005; <lb/>Flaherty, 2009; Burns, 2014). <lb/></body>
10
+
11
+ <front>Received April 9, 2018; accepted October 5, 2018; First published November <lb/>2, 2018. <lb/>The authors declare no competing financial interests. <lb/>Author contributions: C.K. designed research; N.T. and C.K. performed <lb/>research; N.T. and C.K. analyzed data; N.T. and C.K. wrote the paper. <lb/>This work was supported by Deutsche Forschungsgemeinschaft Grants KO <lb/>1143/11-1 and 11-2; and Bundesministerium für Bildung und Forschung Grant <lb/>01GQ1505B, as part of the German-USA Collaboration in Computational <lb/>Neuroscience &quot;Field Potentials in the Auditory System.&quot; <lb/>Acknowledgments: We thank Mark Konishi for the generous gift of the <lb/>software &quot;xdphys,&quot; custom written in his laboratory and used in the single-unit <lb/>recordings; and Rainer Beutelmann for the development and continuing sup-<lb/>port of custom-written software used for the ABR measurements. We also <lb/>thank STELS-OL (Scientific and Technical English Language Services, Olden-<lb/>burg, Germany) for English language editing. <lb/>Correspondence should be addressed to Christine Köppl, Cluster of Excel-<lb/>lence &quot;Hearing4all,&quot; Research Center Neurosensory Science, and Department <lb/>of Neuroscience, School of Medicine and Health Sciences, Carl von Ossietzky <lb/>University, 26129 Oldenburg, Germany. E-mail: [email protected]. <lb/>https://doi.org/10.1523/ENEURO.0140-18.2018 <lb/>Copyright © 2018 Thiele and Köppl <lb/>This is an open-access article distributed under the terms of the Creative <lb/>Commons Attribution 4.0 International license, which permits unrestricted use, <lb/>distribution and reproduction in any medium provided that the original work is <lb/>properly attributed. <lb/></front>
12
+
13
+ <note place="headnote">New Research <lb/></note>
14
+
15
+ <page>2 of 14 <lb/></page>
16
+
17
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org <lb/></note>
18
+
19
+ <body>Materials and Methods <lb/>Experiments were carried out over a time span of sev-<lb/>eral years, using two different laboratories and different <lb/>recording techniques. The two experimental series will be <lb/>referred to as young owls and adult owls. In young owls, <lb/>compound action potential (CAP) and auditory nerve <lb/>single-unit recordings were carried out. In adult owls, <lb/>auditory brainstem responses (ABRs) were recorded. <lb/>Within each group, different anesthetic protocols will be <lb/>abbreviated as follows: ketamine-terminal and isoflurane-<lb/>terminal (young owls); and ketamine-ABR, isoflurane-<lb/>ABR, sevoflurane-ABR, and ABR-terminal (adult owls). <lb/>Experimental animals <lb/>All animal procedures were performed in accordance <lb/>with the German Animal Welfare Act and were approved <lb/>by local authorities (permits AZ 209.1/211-2531-113/03 <lb/>and AZ 33.12 42502-13/1154). In the first experimental <lb/>series, 21 young barn owls of undetermined (either) sex, <lb/>18 T. alba and 3 Tyto furcata (formerly classified as T. alba <lb/>pratincola) were used. Their hatching dates were not al-<lb/>ways known to the specific day, therefore the develop-<lb/>mental stage was determined according to the study by <lb/>Köppl et al. (2005) and was expressed as the number of <lb/>days posthatching; the stage ranged from postnatal day <lb/>11 (P11) to P86. Barn owls are altricial and fledge from the <lb/>nest fully grown at about P65 (Köppl et al., 2005). The <lb/>ABR was recorded in four adult barn owls (T. alba), two <lb/>females and two males, between 5 and 32 months of age <lb/>and weighing between 310 and 390 g. <lb/>Anesthesia and homeostasis <lb/>All animals were deprived of food for ϳ12 h before the <lb/>initiation of anesthesia. Young owls received initial doses <lb/>of 10 mg/kg ketamine hydrochloride (Ketavet, Pharmacia) <lb/>and 3 mg/kg xylazine hydrochloride (Rompun, BayerVital), <lb/>injected intramuscularly. Young owls in the ketamine-<lb/>terminal group were maintained by supplementing ket-<lb/>amine and xylazine as needed, usually every 30 -40 min <lb/>during the surgical stage and every 40 -60 min during <lb/>electrophysiological recordings, at dosages of 6 -11 and <lb/>1.7-2.5 mg/kg, respectively. Young owls in the isoflurane-<lb/>terminal group were maintained on ketamine and xylazine <lb/>only for preliminary surgery, during which the trachea was <lb/>cut in the neck region and intubated, and the abdominal <lb/>air sac was exposed and opened. A one-way respiration <lb/>system was then connected (Burger and Lorenz, 1960; <lb/>Schwartzkopff and Brémond, 1963), delivering gases at a <lb/>constant pressure to the tracheal tube and providing an <lb/>outlet through a short tube inserted into the air sac. <lb/>Spontaneous breathing immediately ceased under artifi-<lb/>cial respiration in all cases. In most experiments, pure <lb/>oxygen was delivered with 1-1.5% isoflurane (Rhodia <lb/>Organique Fine or Essex Tierarznei) added by a vaporizer <lb/>(Vapor 19.3, Dräger), at a volume of 150 -400 ml/min, <lb/>depending on the size of the animal. In three experiments, <lb/>carbogen (95% oxygen and 5% carbon dioxide) was used <lb/>instead of pure oxygen. Respiratory gases were humidi-<lb/>fied via a wash bottle with distilled water before being <lb/>delivered to the animal. All young owls in the isoflurane-<lb/>terminal group also received analgesic injections of 20 -50 <lb/>mg/kg metamizole-sodium (Vetalgin, Intervet) at irregular <lb/>intervals of 2-8 h. <lb/>Each adult barn owl was tested under three different <lb/>anesthetic protocols, applied on separate days in a ran-<lb/>domized sequence, with 1 week of recovery in between. <lb/>Breathing was unaided in all cases. For the ketamine-ABR <lb/>condition, owls received an initial dose of 10 mg/kg ket-<lb/>amine hydrochloride (bela-pharm) and 3 mg/kg xylazine <lb/>hydrochloride (Medistar Arzneimittelvertrieb), i.m. Mainte-<lb/>nance doses of 1.6 -5 mg/kg ketamine and 0.6 -1.8 mg/kg <lb/>xylazine were given as needed, typically every 30 min. For <lb/>the isoflurane-ABR condition, anesthesia was both initi-<lb/>ated and maintained on isoflurane only. A concentration <lb/>of 0.5-1.5% isoflurane (CP-Pharma Handelsgesellschaft) <lb/>was added to carbogen (used to minimize the danger of <lb/>apnea) by a vaporizer (Fortec, Cyprane Kneighley) and <lb/>delivered via a custom-built respiration mask, at a volume <lb/>of 1 L/min. For the sevoflurane-ABR condition, 2-3% <lb/>sevoflurane (Ecuphar; vaporizer, Harvard Apparatus) in <lb/>carbogen was delivered. At the conclusion of each exper-<lb/>iment, the owl received a single dose of ϳ0.03 mg/kg <lb/>meloxicam (Boehringer Ingelheim Vetmedica), a non-<lb/>steroidal anti-inflammatory drug, for the recovery phase. <lb/>In a fourth and terminal experiment, ABR measurements <lb/>were repeated under different, sequentially applied pro-<lb/>tocols that included conditions closer to those of the <lb/>terminal experiments on young owls. The sequence al-<lb/>ways began with ketamine/xylazine injection anesthesia, <lb/>applied as before. The trachea was cut in the neck region <lb/>and intubated, in preparation for later artificial respiration. <lb/>However, breathing was still unaided for the first series of <lb/>measurements. After that, the abdominal air sac was <lb/>opened, and a one-way artificial respiration system with <lb/>pure oxygen (400 ml/min) was instigated, as in young <lb/>owls. After the completion of another series of ABR mea-<lb/>surements, ketamine/xylazine anesthesia was discontin-<lb/>ued and the anesthesia was switched to isoflurane, added <lb/>to the oxygen respiration at different concentrations (1%, <lb/>2%, and back to 1%) to investigate the effect of dosage <lb/>on the ABR. <lb/>All animals were killed by an overdose of sodium pen-<lb/>tobarbital (ϳ100 mg/kg) at the conclusion of the terminal <lb/>experiment. <lb/>The depth of anesthesia was constantly monitored via a <lb/>combined EKG and muscle potential recording between <lb/>needle electrodes inserted into muscles of one wing and <lb/>the contralateral leg. Body temperature was held constant <lb/>at 39°C by a feedback-controlled heating blanket (Har-<lb/>vard Systems) wrapped around the body of the animal, <lb/>with the probe inserted into the cloaca. The head temper-<lb/>ature of young barn owls was monitored separately by a <lb/>small thermoprobe placed in the throat. Barn owls be-<lb/>come homeothermic at ϳ3 weeks posthatching (Shaw-<lb/>yer, 1998), and individuals older than ϳP25 maintained a <lb/>constant head temperature at 37-38°C, under these con-<lb/>ditions. In younger animals, the unassisted head cooled <lb/>significantly relative to the body, and a heat lamp was <lb/>added to maintain head temperature at 35-38°C during <lb/>recordings. <lb/></body>
20
+
21
+ <note place="headnote">New Research <lb/></note>
22
+
23
+ <page>3 of 14 <lb/></page>
24
+
25
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org <lb/></note>
26
+
27
+ <body>Surgery <lb/>The heads of young barn owls were held firmly via an <lb/>individual head band modeled from plaster-of-Paris. For <lb/>CAP recordings, the round window of the inner ear on one <lb/>side was exposed. For single-unit recordings from the <lb/>auditory nerve, the brainstem was exposed by aspirating <lb/>part of the cerebellum. Note that in these experiments, the <lb/>surgical openings also vented the middle-ear space. This <lb/>avoids the buildup of negative middle-ear pressure under <lb/>anesthesia, which significantly reduces auditory sensitiv-<lb/>ity (for review, see Larsen et al., 2016). <lb/>For ABR measurements of adult owls, the beak of the <lb/>owl was fixed in a custom-built holder. To prevent the <lb/>buildup of negative middle-ear pressure under anesthesia <lb/>(Larsen et al., 2016), a sterile 27 gauge cannula was <lb/>inserted through the skull into the middle-ear cavity for <lb/>ventilation during ABR measurements. <lb/>Sound stimulation and electrophysiological <lb/>recordings <lb/>All CAP and single-unit recording measurements took <lb/>place in a custom-built double-walled sound-attenuating <lb/>chamber, and ABR measurements took place in a double-<lb/>walled chamber (model 1203A, Industrial Acoustics). In-<lb/>dividually calibrated acoustic stimuli were presented <lb/>through a custom-built miniature earphone and micro-<lb/>phone system sealed into the ear canal ipsilateral to the <lb/>recording electrode (ER-2 earphone, Etymotic Research; <lb/>FG-23329 microphone for ABR recordings, Knowles). <lb/>For CAP recordings in young owls, a silver wire elec-<lb/>trode, insulated except for a small bead melted at its tip, <lb/>was placed onto the round window membrane, and a <lb/>grounded reference electrode (Ag/AgCl pellet) was placed <lb/>under the skin near the incisions made in the head. Elec-<lb/>trode signals were amplified by a Tucker-Davis Technol-<lb/>ogies (TDT) DB4 amplifier, used at 10,000ϫ to 100,000ϫ <lb/>amplification, 0.1 or 0.2 kHz high-pass filtering, and 15 <lb/>kHz low-pass filtering (12 dB/octave Butterworth filters). <lb/>Signals were then fed to a TDT AD1 analog-to-digital <lb/>converter that was connected via an O1 optical interface <lb/>to an AP2 signal processor interface in a personal com-<lb/>puter. The same interface was used to synthesize the <lb/>acoustic stimuli, which were then antialiased (FT6-2, <lb/>TDT), variably attenuated (PA4, TDT), and fed to the ear-<lb/>phone. Stimuli were tone pips of 20 ms duration, including <lb/>1 ms cosine-shaped rise and fall times (2 ms at 500 Hz) <lb/>and were delivered at rates of 5/s. Stimuli had a fixed <lb/>starting phase and equal numbers of stimulus presenta-<lb/>tions with opposite starting phase were averaged (in total, <lb/>32 or 64). Stimulus generation and CAP recording were <lb/>conducted under the control of TDT software (SigGen and <lb/>BioSig). <lb/>For auditory nerve single unit recordings in young owls, <lb/>glass microelectrodes filled with 3 M KCL and with imped-<lb/>ances typically between 30 and 50 M��� were placed over <lb/>the brainstem under visual control and then remotely <lb/>advanced by a precision microdrive (inchworm 6000ULN, <lb/>Burleigh). A grounded reference electrode (Ag/AgCl pellet) <lb/>was placed under the skin near the incisions made on the <lb/>head. Signals were amplified and filtered (767 electrome-<lb/>ter, World Precision Instruments, TDT PC1 module), ac-<lb/>tion potentials were threshold discriminated (SD1, TDT), <lb/>and the resulting TTL pulses were fed to a TDT AP2 <lb/>interface board, via an event timer (model ET1, TDT) and <lb/>an analog-to-digital converter (model DD1 TDT). As <lb/>above, the same interface was used to synthesize the <lb/>acoustic stimuli, which were then antialiased (FT6-2, <lb/>TDT), variably attenuated (PA4, TDT), and fed to the ear-<lb/>phone. Stimulus generation and recording of the TTL <lb/>pulses was under the control of custom-written software. <lb/>ABR in adult owls was recorded between two subcu-<lb/>taneous platinum electrodes (Grass Technologies), one <lb/>placed on the vertex and one next to the left ear canal. <lb/>Signals were amplified 1000ϫ by an ISO 80 amplifier <lb/>(World Precision Instruments), band-pass filtered be-<lb/>tween 0.1 and 10 kHz, and digitized by a Hammerfall DSP <lb/>Multiface II Interface Card (RME Audio). Stimuli were tone <lb/>bursts with 10 ms duration and 1 ms rise/fall time, and <lb/>delivered at a rate of 7 bursts/s, generated by the same <lb/>interface and fed to the earphone via a TDT HB7 Head-<lb/>phone Buffer. ABR responses were averaged over 300 <lb/>stimulus repetitions. Stimulus generation and ABR re-<lb/>cording were conducted under the control of software <lb/>custom-written in MATLAB (MathWorks). <lb/>Data analysis <lb/>CAP responses in young owls were recorded to fre-<lb/>quencies of 500 Hz and 1 to 10 kHz in 1 kHz steps. At <lb/>each frequency, responses to a range of randomly pre-<lb/>sented levels were recorded, generally in 5 dB incre-<lb/>ments, and decreased to 3 dB near threshold. CAP <lb/>amplitude was defined as the difference between the first <lb/>negative peak N1 and the following most prominent pos-<lb/>itive peak. Thresholds were derived from linear regression <lb/>fits through the initial segment of the curve (four to six <lb/>data points collected at the lowest stimulus levels), as the <lb/>level eliciting a 5 V response. <lb/>For auditory nerve single units recorded in young owls, <lb/>the frequency-threshold curves were derived from re-<lb/>sponses to a matrix of tone bursts of 50 ms duration, <lb/>presented randomly at different frequencies and levels, <lb/>three times each, at a rate of five stimuli per second; the <lb/>threshold criterion was, on average, 20 spikes/s above <lb/>spontaneous rate. The spontaneous rate was estimated <lb/>from the same datasets, either by counting spikes in <lb/>the 50 ms window immediately before each stimulus <lb/>(ketamine-terminal group) or from randomly inserted silent <lb/>trials (isoflurane-terminal group). A new measure of rela-<lb/>tive sensitivity was defined that normalizes for the known <lb/>threshold changes that occur with age that have been <lb/>quantified for ketamine/xylazine anesthetized owls (Köppl <lb/>and Nickel, 2007). Age-typical CAP thresholds can be <lb/>derived for any desired age at 11 standard frequencies <lb/>between 0.5 and 10 kHz, from the published fits of the <lb/>CAP threshold as a function of posthatching age (Köppl <lb/>and Nickel, 2007, their Fig. 8, plus data for five frequen-<lb/>cies not shown). By linear interpolation between frequen-<lb/>cies, age-typical CAP thresholds were then calculated <lb/>for any desired frequency. The difference between the <lb/>threshold of a single unit at characteristic frequency (CF) <lb/></body>
28
+
29
+ <note place="headnote">New Research <lb/></note>
30
+
31
+ <page>4 of 14 <lb/></page>
32
+
33
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org <lb/></note>
34
+
35
+ <body>and the corresponding CAP threshold at that age and <lb/>frequency was taken as a measure of relative sensitivity. <lb/>In ABR recordings from adult owls, a standard set of six <lb/>frequencies was tested, at 1, 2, 4, 6, 8, and 10 kHz. At <lb/>each frequency, responses to a range of randomly pre-<lb/>sented levels were recorded, generally in 5 dB incre-<lb/>ments, and decreased to 3 dB near threshold (with few <lb/>exceptions). ABR thresholds were identified visually, and <lb/>peak-to-peak amplitude and peak latency were read out <lb/>with the use of a custom MATLAB script (Fig. 1, example). <lb/>To eliminate the audiogram threshold variation for graph-<lb/>ical summaries comparing the different anesthetic proto-<lb/>cols, ABR thresholds were normalized to the respective <lb/>individual value in the ketamine-ABR condition or the <lb/>ketamine condition in the ABR-terminal experiment. <lb/>Statistical analysis <lb/>Statistical analyses were carried out with the use of <lb/>PASW Statistics version 18.0.2 or SPSS Statistics ver-<lb/>sions 24 and 25 (IBM). Nonparametric measures and tests <lb/>were used throughout (Table 1). A p value of Յ0.01 was <lb/>the criterion for a significant difference. In case of multiple <lb/>post hoc comparisons, a Bonferroni adjustment of the <lb/>criterion p value was applied by dividing 0.01 by the <lb/>number of pairwise tests required. For example, if three <lb/>pairwise post hoc tests were performed, the Bonferroni-<lb/>corrected criterion p value was 0.003333. <lb/>Results <lb/>Single-unit auditory nerve thresholds were less <lb/>sensitive under isoflurane than under <lb/>ketamine/xylazine <lb/>A total of 57 auditory nerve fibers were recorded in the <lb/>ketamine-terminal condition, from owls aged P17 to P36, <lb/>and 360 auditory nerve fibers in the isoflurane-terminal <lb/>condition, from owls aged P11 to P86. As absolute sen-<lb/>sitivity is known to change within these age brackets <lb/>(Köppl and Nickel, 2007), the difference between the <lb/>threshold of a single unit and the CAP threshold typical for <lb/>that age and frequency was defined as a measure of relative <lb/>sensitivity (see Materials and Methods). This eliminated the <lb/>known maturational changes of auditory thresholds, allow-<lb/>ing for the identification of other factors influencing hearing <lb/>sensitivity. There was a significant difference in relative sen-<lb/>sitivity between auditory nerve fibers recorded under the two <lb/>anesthetic protocols (Table 1, References 1). Median values <lb/>were Ϫ2.4 dB for the isoflurane-terminal condition and <lb/>Ϫ26.5 dB for the ketamine-terminal condition (Fig. 2A). This <lb/>difference held across frequencies when tested separately <lb/>for different characteristic frequencies, binned into 1.5-kHz-<lb/>wide bands. Median relative thresholds for the isoflurane-<lb/>terminal condition were between 17 and 26 dB higher than <lb/>those recorded under the ketamine-terminal condition <lb/>(Table 1, References 2-5, Fig. 2B; note that for CFs Ͼ6 kHz, <lb/>the sample for the ketamine-terminal group was insuf-<lb/>ficient for a meaningful test). <lb/>CAP thresholds were less sensitive under isoflurane <lb/>than under ketamine/xylazine <lb/>CAP thresholds were obtained in only one young owl in <lb/>this study. Nevertheless, this case is included here as it <lb/>validated the principal assumption that CAP and single-<lb/>unit thresholds are tightly correlated. Single-unit data and <lb/>CAP recordings were obtained in the same individual, <lb/>aged P32, under isoflurane anesthesia. The median <lb/>single-unit sensitivity relative to the animal&apos;s own CAP <lb/>audiogram was Ϫ23 dB (n ϭ 7; CFs, 2.6 -4.4 kHz), while <lb/>it was ϩ10.2 dB relative to the age-matched CAP thresh-<lb/>old under ketamine/xylazine anesthesia (Köppl and <lb/>Nickel, 2007). This individual thus represented a drastic <lb/>case of threshold loss (Fig. 3). Furthermore, the direct <lb/>reference of single-unit thresholds to the animal&apos;s own <lb/>CAP thresholds supports the notion derived from the <lb/>population data that single-unit thresholds fall, on aver-<lb/>age, ϳ20 dB below the CAP thresholds obtained under <lb/>comparable conditions. This is typical for birds in general <lb/>(Köppl and Gleich, 2007). <lb/>Spontaneous discharge rates and frequency tuning <lb/>were much less affected <lb/>As a measure of discharge activity of auditory nerve <lb/>single units, spontaneous rates were evaluated for effects <lb/>of the anesthetic protocol. First, data were examined for <lb/>confounding age-related maturation of spontaneous rate. <lb/>There was evidence for lower spontaneous rates in very <lb/>young owls, aged P11 to P14, but no further changes in <lb/>animals older than that (Table 1, References 6 -16). To <lb/>minimize maturation effects, the comparison between the <lb/>anesthetic groups was therefore restricted to owls aged <lb/>P17 and older. In these groups, spontaneous rates were <lb/>significantly lower in the isoflurane-terminal group <lb/>(Table 1, Reference 17, Fig. 4A). Median values were 40 <lb/>spikes/s in the ketamine-terminal group and 33.3 spikes/s <lb/>in the isoflurane-terminal group. Since spontaneous rates <lb/>are, in addition, known to vary with CF (Köppl, 1997), the <lb/>data were further examined separately, for 1.5-kHz-wide <lb/>CF bands. Here, the difference between the anesthetic <lb/>Fig. 1. Example of a typical ABR recording from an adult owl. <lb/>Shown is an average response to 300 stimuli at 2 kHz, 51 dB <lb/>SPL. Only wave I was analyzed. Amplitude was defined as the <lb/>difference between the first positive (AmpMax) and following <lb/>negative (AmpMin) peak. Latency was defined as the latency of <lb/>the first positive peak (Tmax). <lb/></body>
36
+
37
+ <note place="headnote">New Research <lb/></note>
38
+
39
+ <page>5 of 14 <lb/></page>
40
+
41
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org <lb/></note>
42
+
43
+ <body>Table 1: List of statistical tests <lb/>Reference <lb/>number <lb/>Data structure <lb/>Parameter tested <lb/>Type of test <lb/>p value <lb/>Figure <lb/>Two independent samples: <lb/>ketamine-terminal (n ϭ 57) <lb/>isoflurane-terminal (n ϭ 351) <lb/>Single-unit threshold (relative to age-matched <lb/>CAP audiograms) <lb/>Mann-Whitney <lb/>&lt;0.001 <lb/>2A <lb/>Two independent samples: <lb/>ketamine-terminal (n ϭ 5) <lb/>isoflurane-terminal (n ϭ 49) <lb/>Single-unit threshold (relative to age-matched <lb/>CAP audiograms) <lb/>CF Ͻ1.5 kHz <lb/>Mann-Whitney <lb/>0.004 <lb/>2B <lb/>Two independent samples: <lb/>ketamine-terminal (n ϭ 14) <lb/>isoflurane-terminal (n ϭ 76) <lb/>Single-unit threshold (relative to age-matched <lb/>CAP audiograms) <lb/>CF 1.5-3 kHz <lb/>Mann-Whitney <lb/>&lt;0.001 <lb/>2B <lb/>Two independent samples: <lb/>ketamine-terminal (n ϭ 20) <lb/>isoflurane-terminal (n ϭ 89) <lb/>Single-unit threshold (relative to age-matched <lb/>CAP audiograms) <lb/>CF 3-4.5 kHz <lb/>Mann-Whitney <lb/>&lt;0.001 <lb/>2B <lb/>Two independent samples: <lb/>ketamine-terminal (n ϭ 17) <lb/>isoflurane-terminal (n ϭ 86) <lb/>Single-unit threshold (relative to age-matched <lb/>CAP audiograms) <lb/>CF 4.5-6 kHz <lb/>Mann-Whitney <lb/>&lt;0.001 <lb/>2B <lb/>Five independent samples (age groups): <lb/>P11 to P14 (n ϭ 99) <lb/>P17 (n ϭ 35) <lb/>P21 to 32 (n ϭ 87) <lb/>P35 to P40 (n ϭ 17) <lb/>P51 to P86 (n ϭ 38) <lb/>Single-unit spontaneous discharge rate <lb/>Kruskal-Wallis <lb/>&lt;0.001 <lb/>Two independent samples: <lb/>P11 to P14 (n ϭ 99) <lb/>P17 (n ϭ 35) <lb/>Single-unit spontaneous discharge rate <lb/>Mann-Whitney Bonferroni corrected <lb/>&lt;0.001 <lb/>Two independent samples: <lb/>P11to P14 (n ϭ 99) <lb/>P21 to P32 (n ϭ 87) <lb/>Single-unit spontaneous discharge rate <lb/>Mann-Whitney Bonferroni corrected <lb/>0.001 <lb/>Two independent samples: <lb/>P11 to P14 (n ϭ 99) <lb/>P35 to P40 (n ϭ 17) <lb/>Single-unit spontaneous discharge rate <lb/>Mann-Whitney Bonferroni corrected <lb/>0.003 <lb/>Two independent samples: <lb/>P11 to P14 (n ϭ 99) <lb/>P51 to P86 (n ϭ 38) <lb/>Single-unit spontaneous discharge rate <lb/>Mann-Whitney Bonferroni corrected <lb/>0.002 <lb/>Two independent samples: <lb/>P17 (n ϭ 35) <lb/>P21 to P32 (n ϭ 87) <lb/>Single-unit spontaneous discharge rate <lb/>Mann-Whitney Bonferroni corrected <lb/>0.346 <lb/>Two independent samples: <lb/>P17 (n ϭ 35) <lb/>P35 to P40 (n ϭ 17) <lb/>Single-unit spontaneous discharge rate <lb/>Mann-Whitney Bonferroni corrected <lb/>0.992 <lb/>Two independent samples: <lb/>P17 (n ϭ 35) <lb/>P51 to P86 (n ϭ 38) <lb/>Single-unit spontaneous discharge rate <lb/>Mann-Whitney Bonferroni corrected <lb/>0.614 <lb/>Two independent samples: <lb/>P21 to 32 (n ϭ 87) <lb/>P35 to P40 (n ϭ 17) <lb/>Single-unit spontaneous discharge rate <lb/>Mann-Whitney Bonferroni corrected <lb/>0.467 <lb/>Two independent samples: <lb/>P21 to P32 (n ϭ 87) <lb/>P51 to P86 (n ϭ 38) <lb/>Single-unit spontaneous discharge rate <lb/>Mann-Whitney Bonferroni corrected <lb/>0.718 <lb/>Two independent samples: <lb/>P35 to P40 (n ϭ 17) <lb/>P51 to P86 (n ϭ 38) <lb/>Single-unit spontaneous discharge rate <lb/>Mann-Whitney Bonferroni corrected <lb/>0.826 <lb/>Two independent samples: <lb/>ketamine-terminal (n ϭ 56) <lb/>isoflurane-terminal (n ϭ 260) <lb/>Single-unit spontaneous discharge rate, <lb/>all ages ՆP17 <lb/>Mann-Whitney <lb/>0.005 <lb/>4A <lb/>Two independent samples: <lb/>ketamine-terminal (n ϭ 5) <lb/>isoflurane-terminal (n ϭ 29) <lb/>Single-unit spontaneous discharge rate, <lb/>all ages ՆP17 and CF Ͻ1.5 kHz <lb/>Mann-Whitney <lb/>0.575 <lb/>4B <lb/>Two independent samples: <lb/>ketamine-terminal (n ϭ 14) <lb/>isoflurane-terminal (n ϭ 30) <lb/>Single-unit spontaneous discharge rate, <lb/>all ages ՆP17 and CF 1.5-3 kHz <lb/>Mann-Whitney <lb/>0.035 <lb/>4B <lb/>Two independent samples: <lb/>ketamine-terminal (n ϭ 19) <lb/>isoflurane-terminal (n ϭ 62) <lb/>Single-unit spontaneous discharge rate, <lb/>all ages ՆP17 and CF 3-4.5 kHz <lb/>Mann-Whitney <lb/>0.858 <lb/>4B <lb/>Two independent samples: <lb/>ketamine-terminal (n ϭ 17) <lb/>isoflurane-terminal (n ϭ 88) <lb/>Single-unit spontaneous discharge rate, <lb/>all ages ՆP17 and CF 4.5-6 kHz <lb/>Mann-Whitney <lb/>0.264 <lb/>4B <lb/>Two independent samples: <lb/>ketamine-terminal (n ϭ 53) <lb/>isoflurane-terminal (n ϭ 327) <lb/>Single-unit Q10 dB <lb/>Mann-Whitney <lb/>0.098 <lb/>(Continued) <lb/></body>
44
+
45
+ <note place="headnote">New Research <lb/></note>
46
+
47
+ <page>6 of 14 <lb/></page>
48
+
49
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org <lb/></note>
50
+
51
+ <body>regimes did not hold for any CF band (Fig. 4B, Table 1, <lb/>References 18 -21; note that for CFs Ͼ6 kHz, the sample <lb/>for the ketamine-terminal group was insufficient for a <lb/>meaningful test). Together, isoflurane thus appeared to <lb/>have a mildly depressive effect on spontaneous rates <lb/>when compared with data from ketamine/xylazine-ane-<lb/>sthetized juvenile owls. <lb/>The quality of frequency tuning, expressed as Q 10dB , <lb/>was not consistently affected by the anesthetic protocol. <lb/>An overall comparison of Q 10dB values between the ket-<lb/>amine-terminal and the isoflurane-terminal groups re-<lb/>vealed no significant difference (Table 1, Reference 22). <lb/>ABR thresholds were less sensitive with gas <lb/>anesthesia compared with ketamine/xylazine <lb/>Four adult owls were tested under three anesthetic <lb/>protocols each: ketamine-ABR, isoflurane-ABR, and sevo-<lb/>flurane-ABR. Importantly, the sequence of testing was <lb/>randomized and different for each owl. ABR audiograms <lb/>showed a similar overall shape for all conditions, suggest-<lb/>ing that the basic relationship between ABR threshold and <lb/>frequency was not affected (Fig. 5A). However, thresholds <lb/>differed significantly between the conditions (Table 1, <lb/>Reference 23). Specifically, thresholds in the ketamine-<lb/>ABR condition were significantly lower than thresholds <lb/>Table 1: Continued <lb/>Reference <lb/>number <lb/>Data structure <lb/>Parameter tested <lb/>Type of test <lb/>p value <lb/>Figure <lb/>23 <lb/>Three dependent samples (n ϭ 24): <lb/>ketamine-ABR <lb/>isoflurane-ABR <lb/>sevoflurane-ABR <lb/>ABR threshold <lb/>Friedman <lb/>&lt;0.001 <lb/>5 <lb/>24 <lb/>Two dependent samples (n ϭ 24): <lb/>ketamine-ABR <lb/>isoflurane-ABR <lb/>ABR threshold <lb/>Wilcoxon <lb/>Bonferroni corrected <lb/>&lt;0.001 <lb/>5 <lb/>25 <lb/>Two dependent samples (n ϭ 24): <lb/>ketamine-ABR <lb/>sevoflurane-ABR <lb/>ABR threshold <lb/>Wilcoxon <lb/>Bonferroni corrected <lb/>&lt;0.001 <lb/>5 <lb/>26 <lb/>Two dependent samples (n ϭ 24): <lb/>isoflurane-ABR <lb/>sevoflurane-ABR <lb/>ABR threshold <lb/>Wilcoxon <lb/>Bonferroni corrected <lb/>0.076 <lb/>5 <lb/>27 <lb/>Six independent samples: <lb/>1/2/4/6/8/10 kHz <lb/>(n ϭ 4 each) <lb/>ABR threshold difference: isoflurane-ABRϪ <lb/>ketamine-ABR condition <lb/>Kruskal-Wallis <lb/>0.406 <lb/>5B <lb/>28 <lb/>Six independent samples: <lb/>1/2/4/6/8/10 kHz <lb/>(n ϭ 4 each) <lb/>ABR threshold difference: sevoflurane-ABRϪ <lb/>ketamine-ABR condition <lb/>Kruskal-Wallis <lb/>0.472 <lb/>5B <lb/>29 <lb/>Three dependent samples (n ϭ 16): <lb/>ketamine-ABR <lb/>isoflurane-ABR <lb/>sevoflurane-ABR <lb/>ABR amplitudes 10 dB above threshold <lb/>Friedman <lb/>0.068 <lb/>30 <lb/>Three dependent samples (n ϭ 16): <lb/>ketamine-ABR <lb/>isoflurane-ABR <lb/>sevoflurane-ABR <lb/>ABR latencies 10 dB above threshold <lb/>Friedman <lb/>0.646 <lb/>31 <lb/>Two dependent samples (n ϭ 24): <lb/>ABR-terminal, ketamine <lb/>ABR-terminal, ketamine ϩ oxygen <lb/>ABR threshold <lb/>Wilcoxon <lb/>0.163 <lb/>32 <lb/>Three dependent samples (n ϭ 11): <lb/>ABR-terminal, 1% isoflurane <lb/>ABR-terminal, 2% isoflurane <lb/>ABR-terminal, 1% isoflurane repeat <lb/>ABR threshold <lb/>Friedman <lb/>&lt;0.001 <lb/>6A <lb/>33 <lb/>Two dependent samples (n ϭ 11): <lb/>ABR-terminal, 1% isoflurane <lb/>ABR-terminal, 2% isoflurane <lb/>ABR threshold <lb/>Wilcoxon <lb/>Bonferroni corrected <lb/>0.003 <lb/>6A <lb/>34 <lb/>Two dependent samples (n ϭ 11): <lb/>ABR-terminal, 2% isoflurane <lb/>ABR-terminal, 1% isoflurane repeat <lb/>ABR threshold <lb/>Wilcoxon <lb/>Bonferroni corrected <lb/>0.003 <lb/>6A <lb/>35 <lb/>Two dependent samples (n ϭ 11): <lb/>ABR-terminal, 1% isoflurane <lb/>ABR-terminal, 1% isoflurane repeat <lb/>ABR threshold <lb/>Wilcoxon <lb/>Bonferroni corrected <lb/>0.262 <lb/>6A <lb/>36 <lb/>Two dependent samples (n ϭ 24): <lb/>ABR-terminal, ketamine ϩ oxygen <lb/>ABR-terminal, 1% isoflurane <lb/>ABR threshold <lb/>Wilcoxon <lb/>0.004 <lb/>37 <lb/>Two dependent samples (n ϭ 24): <lb/>isoflurane-ABR (normalized <lb/>compared with ketamine) <lb/>ABR-terminal, 1% isoflurane <lb/>(normalized compared with ketamine) <lb/>ABR threshold <lb/>Wilcoxon <lb/>&lt;0.001 <lb/>Column 1 shows the serial number used to refer to specific tests throughout the article. Column 2 defines the samples, and column 3 the tested parameter. <lb/>Column 4 lists the specific nonparametric test used, and column 5 shows the resulting p value, which is highlighted in bold type if the null hypothesis was re-<lb/>jected. Note that the criterion p value was 0.01, or lower if a Bonferroni correction was applied, as indicated in Column 4. Finally, column 6 refers to the rele-<lb/>vant figure, if applicable. <lb/></body>
52
+
53
+ <note place="headnote">New Research <lb/></note>
54
+
55
+ <page>7 of 14 <lb/></page>
56
+
57
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org <lb/></note>
58
+
59
+ <body>for either the isoflurane-or sevoflurane-ABR condition <lb/>(Table 1, References 24 and 25). There was no significant <lb/>difference between ABR thresholds obtained with the two <lb/>anesthetic gases (Table 1, Reference 26). Threshold dif-<lb/>ferences to the respective threshold in the ketamine-ABR <lb/>condition showed no significant frequency dependence <lb/>(Table 1, References 27 and 28, Fig. 5B). Overall, thresh-<lb/>olds under isoflurane showed a median elevation of 7 dB, <lb/>thresholds under sevoflurane showed a median elevation <lb/>of 15 dB compared with the ketamine-ABR condition (Fig. <lb/>5C). <lb/>The amplitude and latency of ABR wave I were unaf-<lb/>fected by the anesthesia protocol. Of course, amplitude <lb/>increased and latency decreased with increasing sound <lb/>level. Therefore, this comparison was carried out at a <lb/>relative level of 10 dB above the respective ABR threshold <lb/>(Table 1, References 29 and 30). <lb/>In a final, terminal experiment, each adult owl was <lb/>tested with more invasive protocols to assess confound-<lb/>ing factors such as artificial respiration and variable inhal-<lb/>ant concentration. There was no significant change in <lb/>ABR thresholds after switching from unassisted breathing <lb/>through a tracheotomy to artificial respiration with oxygen, both <lb/>still under ketamine/xylazine anesthesia (Table 1, Reference <lb/>31). Next, the anesthetic protocol was switched to isoflurane <lb/>and stepped from 1% to 2% and back to 1%, with a minimum <lb/>equilibration time of 15 min before measurements were ob-<lb/>tained after a change in gas concentration. This revealed <lb/>a significant effect of isoflurane dosage (Table 1, Refer-<lb/>ence 32, Fig. 6A). Increasing isoflurane from 1% to 2% <lb/>resulted in a significant rise of ABR thresholds (Table 1, <lb/>Reference 33, Fig. 6A), which was reversible upon a return <lb/>to 1% (Table 1, References 34 and 35, Fig. 6A). Note that, <lb/>unfortunately, the full sequence of tests could be com-<lb/>pleted for only two owls. At the point of the initial switch <lb/>to 1% isoflurane, the full sample from all four owls could <lb/>still be obtained and showed a significant elevation of <lb/>thresholds relative to the ketamine condition with artificial <lb/>respiration tested immediately before (Table 1, Reference <lb/>36), thus confirming the principal effect observed in the <lb/>A <lb/>B <lb/>Fig. 2. Auditory nerve single-unit thresholds were severely ele-<lb/>vated under isoflurane. A, Box plot showing thresholds (normal-<lb/>ized to the age-matched CAP) for auditory nerve fibers recorded <lb/>under ketamine/xylazine and isoflurane, respectively. B, The <lb/>same data, separated into 1.5-kHz-wide CF bands, for the two <lb/>anesthetic conditions. Empty boxes represent data for the <lb/>ketamine-terminal conditions, and hatched boxes represent data <lb/>for the isoflurane-terminal condition. Note that thresholds under <lb/>isoflurane were significantly higher. Boxes and whiskers indicate <lb/>the interquartile ranges and 1.5 times the interquartile ranges, <lb/>respectively. Horizontal lines within boxes indicate medians, and <lb/>circular symbols indicate outliers that lie beyond 1.5 times the <lb/>interquartile range. <lb/>Fig. 3. CAP thresholds were severely elevated under isoflurane. <lb/>CAP threshold audiogram of an individual aged P32, under <lb/>isoflurane anesthesia (solid line). For comparison, the average <lb/>CAP audiogram for P32 owls under ketamine/xylazine anesthe-<lb/>sia is also shown (dashed line; after Köppl and Nickel, 2007). <lb/></body>
60
+
61
+ <note place="headnote">New Research <lb/></note>
62
+
63
+ <page>8 of 14 <lb/></page>
64
+
65
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org <lb/></note>
66
+
67
+ <body>previous ABR experiments with the same individuals. As <lb/>the terminal experiments progressed, however, an unex-<lb/>plained, gradual, and irreversible loss of sensitivity oc-<lb/>curred at different times for different owls that, in addition, <lb/>appeared to affect the higher frequencies more (Fig. 6B). <lb/>Some thresholds exceeded the maximum SPL of the <lb/>sound system, such that the sample gradually diminished <lb/>(Fig. 6B). The cause for this deterioration, which appeared <lb/>unrelated to the anesthetic protocol is unknown and will <lb/>be discussed below in the section about confounding <lb/>factors. Importantly, we observed no changes to the fre-<lb/>quency of heart beat or the shape of the EKG waveform <lb/>that might correlate with the loss of sensitivity. <lb/>Interaction with type of respiration <lb/>The above observations suggested significant factors <lb/>that influence auditory sensitivity, in addition to the anes-<lb/>thetic agent. This prompted us to compare thresholds <lb/>obtained with two types of respiration, both under isoflu-<lb/>rane anesthesia. Indeed, thresholds in the isoflurane-ABR <lb/>condition and the ABR-terminal-isoflurane 1% condition, <lb/>both normalized to the respective ketamine thresholds, <lb/>differed significantly (Table 1, Reference 37). The main <lb/>difference was the type of respiration: unassisted breath-<lb/>ing of 0.5-1.5% isoflurane delivered in carbogen versus <lb/>artificial respiration with 1% isoflurane delivered in pure <lb/>oxygen. In the condition we normalized to, owls always <lb/>breathed normal room air unassisted. While both isoflu-<lb/>rane conditions resulted in a threshold loss (as already <lb/>shown above), the loss was relatively greater when <lb/>breathing carbogen unaided, by a median of 1 dB. <lb/>Discussion <lb/>The main finding of the present study was a significant <lb/>deterioration of auditory sensitivity in barn owls under gas <lb/>anesthesia. The effect was drastic in young animals that, <lb/>compared with age-matched individuals that were anes-<lb/>thetized with a combination of ketamine and xylazine, <lb/>experienced threshold elevations in auditory nerve single-<lb/>unit responses of Ն20 dB. Consistent with this, ABR <lb/>thresholds assessed repeatedly in experiments on adult <lb/>owls were also significantly lower under ketamine/xyla-<lb/>zine anesthesia, compared with gas anesthesia with both <lb/>isoflurane and sevoflurane. Importantly, this difference <lb/>already occurred with minimal dosages and was revers-<lb/>ibly enlarged with increased isoflurane concentration. Fi-<lb/>nally, there was evidence for confounding detrimental <lb/>effects associated with the respiration mode. <lb/>Reports of anesthetic effects on peripheral auditory <lb/>responses across species <lb/>Evidence that inhalant anesthetics adversely affect co-<lb/>chlear sensitivity has accumulated in recent years for <lb/>mammals also. In the guinea pig, isoflurane was shown to <lb/>have a dose-dependent depressive effect on several au-<lb/>ditory evoked potentials [CAP, cochlear microphonic (CM) <lb/>potential, and ABR; Stronks et al., 2010). Thresholds, <lb/>amplitudes, and neural latencies were all negatively af-<lb/>fected. The effects were most pronounced at higher fre-<lb/>quencies Ն8 kHz, where CAP thresholds were elevated <lb/>by ϳ10-15 dB. Stronks et al. (2010) referenced the mea-<lb/>surements obtained under isoflurane to the awake condi-<lb/>tion, which does not necessarily suggest that isoflurane <lb/>acts worse than other anesthetics. In rat and mouse, ABR <lb/>thresholds were directly compared between anesthesia <lb/>with isoflurane and with ketamine/xylazine, and they were <lb/>found to be relatively elevated under isoflurane (Ceder-<lb/>holm et al., 2012; Ruebhausen et al., 2012). Furthermore, <lb/>in the gerbil, ABR thresholds under ketamine/xylazine <lb/>were not significantly different from those in the awake <lb/>condition (Smith and Mills, 1989). Together, these studies <lb/>A <lb/>B <lb/>Fig. 4. Auditory nerve spontaneous discharge rates were mildly <lb/>depressed by isoflurane. A, Box plot showing overall spontane-<lb/>ous discharge rates for auditory nerve fibers recorded under <lb/>ketamine/xylazine and isoflurane, respectively. The rates under <lb/>isoflurane were significantly lower. B, The same data, separated <lb/>into 1.5-kHz-wide CF bands, for the two anesthetic conditions. <lb/>Empty boxes represent data for the ketamine-terminal condi-<lb/>tions, and hatched boxes represent data for the isoflurane-<lb/>terminal condition. Boxes and whiskers indicate the interquartile <lb/>ranges and 1.5 times the interquartile ranges, respectively. Hor-<lb/>izontal lines within boxes indicate medians, and circular symbols <lb/>and stars indicate outliers that lie beyond 1.5 times (circles) or <lb/>beyond 3 times (stars) the interquartile range. <lb/></body>
68
+
69
+ <note place="headnote">New Research <lb/></note>
70
+
71
+ <page>9 of 14 <lb/></page>
72
+
73
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org <lb/></note>
74
+
75
+ <body>strongly support a differentially detrimental action of the <lb/>gas anesthetic, very similar to the present findings in the <lb/>barn owl. <lb/>However, there is clearly also species-specific variation <lb/>in the sensitivity to different anesthetics, and gas anes-<lb/>thetics do not fare universally worst. In the starling, a small <lb/>songbird, CAP amplitudes were depressed under halo-<lb/>thane anesthesia, at concentrations as low as 0.5%, com-<lb/>pared with the awake state (Kettembeil et al., 1995). The <lb/>same study, however, observed equally depressive ef-<lb/>fects of ketamine/xylazine and halothane anesthesia on <lb/>otoacoustic emissions, which reflect the responses of <lb/>sensory hair cells, in both starlings and chickens. Since <lb/>neural responses were not obtained under ketamine/xy-<lb/>lazine, it thus remained unclear whether they might be <lb/>differentially affected. In a systematic comparison of au-<lb/>ditory nerve single-unit responses under different anes-<lb/>thetic protocols in a lizard, the Tokay gecko, Dodd and <lb/>Capranica (1992) found significantly elevated thresholds <lb/>under both isoflurane and ketamine, compared with pen-<lb/>tobarbital anesthesia. Thus, in the gecko, too, ketamine <lb/>had similarly degrading effects compared with isoflurane, <lb/>albeit comparatively highly dosed at 440 mg/kg (Dodd <lb/>and Capranica, 1992). Furthermore, in the Tokay gecko, in <lb/>contrast to the present study in the owl, auditory nerve <lb/>discharge rates were also severely and differentially de-<lb/>pressed under the different anesthetic conditions: the <lb/>highest rates were observed under pentobarbital, fol-<lb/>lowed by ketamine, with isoflurane having the lowest rates <lb/>(Dodd and Capranica, 1992). <lb/>Finally, there is conflicting evidence regarding the ef-<lb/>fects of gas anesthetics on the responses of sensory hair <lb/>cells, specifically the outer hair cells of the mammalian <lb/>cochlea, measured as otoacoustic emissions. In humans, <lb/>several studies reported a selectively depressive effect of <lb/>gas anesthetics on evoked emissions (Ferber-Viart et al., <lb/>1998; Ropposch et al., 2014; Gungor et al., 2015). How-<lb/>ever, it is currently unclear whether this is a truly pharma-<lb/>cological effect on the cochlea or may be a secondary <lb/>consequence of changes in arterial blood pressure. In a <lb/>bat species, isoflurane was shown to have the opposite <lb/>effect (i.e., increased emission amplitudes; Drexl et al., <lb/>2004). It was suggested that this may reflect disinhibition <lb/>through inactivation of the olivocochlear efferent input <lb/>(see also next section). <lb/>Possible mechanisms of isoflurane and sevoflurane <lb/>action <lb/>The mechanisms that produce general anesthesia at <lb/>the systems level are still poorly understood (Rudolph and <lb/>Antkowiak, 2004; Ishizawa, 2007). The cellular sites of <lb/>action commonly involve ion channels and neurotransmit-<lb/>ter receptors that are widely expressed in the CNS and <lb/>should thus act at all levels. Nevertheless, as a general <lb/>rule, a gradual effect is observed, such that higher-level <lb/>cognitive functions are impaired at lower anesthetic con-<lb/>centrations than motor functions, early visual processing, <lb/>or basic homeostatic physiology (Campagna et al., 2003; <lb/>Rudolph and Antkowiak, 2004; Ishizawa, 2007). This sug-<lb/>gests that while the cellular sites of action may be similar, <lb/>higher centers tend to show the combined result of direct <lb/>anesthetic action and cumulative effects in neural net-<lb/>works. This also promotes the common assumption that <lb/>general anesthesia, when appropriately dosed, should not <lb/>significantly affect primary sensory processes. Therefore, <lb/>the pronounced effect of isoflurane and sevoflurane at the <lb/>most peripheral levels of the auditory system, the hair <lb/>cells and auditory nerve, is surprising. <lb/>Isoflurane and probably all inhalant anesthetics belong-<lb/>ing to the halogenated alkanes and ethers, such as halo-<lb/>thane and sevoflurane, have several known target sites of <lb/>action, all of which are predicted to suppress neural ac-<lb/>A <lb/>C <lb/>B <lb/>Fig. 5. ABR thresholds were elevated under gas anesthesia. A, Box plot showing ABR thresholds as a function of frequency, for the same <lb/>four adult individuals, tested with different anesthetic protocols in successive experiments. B, The same data, with thresholds now <lb/>normalized to the values at the respective frequency for the ketamine-ABR condition; as a visual reference, the dashed line indicates the <lb/>reference condition. Note that the minor variation across frequencies was not significant (Table 1, References 27 and 28). Therefore, C then <lb/>shows an overall comparison between anesthetic conditions. Thresholds for either the isoflurane-ABR or sevoflurane-ABR condition were <lb/>significantly higher than thresholds for the ketamine-ABR condition (Table 1, References 24 and 25). Thresholds for the ketamine-ABR <lb/>condition are shown as empty boxes, for the isoflurane-ABR condition as hatched boxes, and for the sevoflurane-ABR condition as gray <lb/>boxes. Boxes and whiskers indicate the interquartile ranges and 1.5 times the interquartile ranges, respectively. Horizontal lines within <lb/>boxes indicate medians. There were no outliers beyond 1.5 times the interquartile ranges. <lb/></body>
76
+
77
+ <note place="headnote">New Research <lb/></note>
78
+
79
+ <page>10 of 14 <lb/></page>
80
+
81
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org <lb/></note>
82
+
83
+ <body>tivity (for review, see Campagna et al., 2003; Rudolph and <lb/>Antkowiak, 2004; Burns, 2014). They suppress excitatory <lb/>transmission through the inhibition of glutamate recep-<lb/>tors, both the NMDA and AMPA subtypes, and through <lb/>inhibition of nicotinic acetylcholine (ACh) receptors. Con-<lb/>versely, they enhance inhibitory transmission through the <lb/>potentiation of GABA A and glycine receptors. In contrast, <lb/>the sites of ketamine and xylazine action are more re-<lb/>stricted. Ketamine predominantly inhibits NMDA-type glu-<lb/>tamate receptors and nicotinic ACh receptors, both <lb/>normally excitatory (Rudolph and Antkowiak, 2004). Xyla-<lb/>zine is a known agonist of the ␣ 2 -adrenergic receptor <lb/>(Lierz and Korbel, 2012), which is best known for mediat-<lb/>ing the inhibition of sympathetic activity of the autonomic <lb/>nervous system. <lb/>Considering these anesthetic profiles, where are the <lb/>potential sites of action at the cochlear level? Glutamate <lb/>receptors of all ionotropic subtypes are typically found on <lb/>the afferent terminals connecting to vertebrate hair cells <lb/>(Eatock and Lysakowski, 2006). In the mammalian co-<lb/>chlea, the functionally predominant receptors are those of <lb/>the AMPA subtype (Ruel et al., 2007; Glowatzki et al., <lb/>2008). Assuming the same for birds, isoflurane and sevoflu-<lb/>rane are indeed predicted to have a potentially larger direct <lb/>impact on auditory afferents than ketamine. However, in the <lb/>present study, the observed effect was curiously specific to <lb/>auditory thresholds and affected discharge rates only mildly. <lb/>This is not obviously compatible with a general suppressive <lb/>effect on the auditory afferents. <lb/>There is currently no evidence for the inhibitory neu-<lb/>rotransmitters GABA and glycine in the avian cochlea (for <lb/>review, see Köppl, 2011), so these are not likely to be <lb/>potential mediators of the observed threshold shifts in <lb/>birds. Remaining possible sites of action are the cholin-<lb/>ergic terminals of efferent fibers to the cochlear hair cells <lb/>(Köppl, 2011). Indeed, an inhibitory effect of isoflurane on <lb/>these has been indirectly suggested for mammals (Drexl <lb/>et al., 2004). Depending on the subtypes of efferents <lb/>activated, such an enhancement could conceivably me-<lb/>diate a suppressive effect on auditory nerve afferents. <lb/>However, the pharmacology of the hair cell ACh receptor <lb/>is of an unusual nicotinic subtype (Katz et al., 2011), and <lb/>there have been no direct tests for interactions with any <lb/>anesthetic agent. <lb/>In summary, selectively depressive effects of gas anes-<lb/>thetics relative to ketamine on cochlear neural responses <lb/>are consistent with their wider spectrum of known cellular <lb/>interaction sites, specifically their inhibitory action on <lb/>ionotropic glutamate receptors. However, the specific <lb/>pattern of a drastic threshold shift without a comparable <lb/>effect on discharge rates or frequency tuning, does not fit <lb/>any straightforward predictions. Importantly, it may point <lb/>to additional, confounding effects associated with pro-<lb/>longed and invasive protocols. <lb/>Confounding factors <lb/>The anesthetic agents were not the only difference <lb/>between the experimental groups in our initial study with <lb/>young barn owls. Young owls of the isoflurane-terminal <lb/>group were artificially respirated with oxygen or carbogen, <lb/>B <lb/>A <lb/>Fig. 6. A, Isoflurane dose dependence of ABR thresholds. Box <lb/>plot showing thresholds for different isoflurane concentrations <lb/>tested sequentially in ABR-terminal experiments. Thresholds un-<lb/>der 2% isoflurane were significantly elevated relative to those <lb/>under both 1% conditions (Table 1, References 32-35). B, Grad-<lb/>ual threshold deterioration with time in terminal ABR experi-<lb/>ments. Shown are ABR thresholds normalized to the values for <lb/>the initial ketamine/xylazine condition (animal breathing air un-<lb/>aided), separated according to frequency, for all conditions <lb/>tested sequentially. Note that the number of data points contrib-<lb/>uting to each box now varies; upward arrows indicate data that <lb/>dropped out because the threshold exceeded the limit of the <lb/>sound system and thus could not be determined. Note also that <lb/>high frequencies appear to be affected more. Thresholds ob-<lb/>tained under ketamine ϩ artificial oxygen respiration are shown <lb/>as empty boxes, for the 1% isoflurane condition as hatched blue <lb/>boxes, for the 2% isoflurane condition as hatched yellow boxes, <lb/>and for the repeated 1% isoflurane condition as hatched ma-<lb/>genta boxes. Boxes and whiskers indicate the interquartile <lb/>ranges and 1.5 times the interquartile ranges, respectively. Hor-<lb/>izontal lines within boxes indicate medians. There were no out-<lb/>liers beyond 1.5 times the interquartile ranges. <lb/></body>
84
+
85
+ <note place="headnote">New Research <lb/></note>
86
+
87
+ <page>11 of 14 <lb/></page>
88
+
89
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org <lb/></note>
90
+
91
+ <body>while the ketamine-terminal group breathed normal air <lb/>unaided. Therefore, in the follow-up study on adult owls, <lb/>each individual was tested repeatedly under otherwise <lb/>identical conditions. This confirmed the specific, detri-<lb/>mental effect of gas anesthesia. However, the threshold <lb/>difference relative to ketamine anesthesia was smaller for <lb/>the group of adult owls, opening several possibilities for <lb/>confounding effects. These were addressed in the termi-<lb/>nal experiments with adult owls. <lb/>Similar to previous studies (Kettembeil et al., 1995; <lb/>Stronks et al., 2010), it was shown that the inhalant an-<lb/>esthetic elevated peripheral auditory thresholds in a dose-<lb/>dependent manner. Thus, dosage is a likely confounding <lb/>factor in experiments using more invasive procedures, <lb/>such as single-unit recordings, which tend to require <lb/>higher anesthetic dosages. <lb/>Mode of respiration is another possible confounding <lb/>factor that was tested. Breathing carbogen unassisted <lb/>appeared to result in additional sensitivity losses, over <lb/>and above those related to isoflurane anesthesia, when <lb/>compared with artificial respiration with pure oxygen. We <lb/>had chosen carbogen (as opposed to pure oxygen) in the <lb/>experiments where the owls breathed the isoflurane mix-<lb/>ture unassisted, to minimize the danger of apnea. CO 2 is <lb/>known to be an important respiratory regulator, and low <lb/>partial pressure of CO 2 (p CO2 ) tends to depress respiration <lb/>(Powell, 2015). Our results, however, suggest that this <lb/>was misguided and instead resulted in a slight additional <lb/>loss of sensitivity. An obvious explanation for this effect is <lb/>lacking. We do not consider hypoxia very likely, since <lb/>carbogen still contains 95% oxygen (i.e., a concentration <lb/>far above normal air). Similarly, artificial respiration with <lb/>pure oxygen is unlikely to cause hypoxia but still appeared <lb/>to cause a decline in auditory sensitivity over time. While <lb/>no consistent, immediate deterioration associated with <lb/>the switch to artificial oxygen respiration was found, un-<lb/>explained drastic threshold losses occurred beyond 4 -5 h <lb/>in adult owls receiving artificial oxygen respiration. This is <lb/>reminiscent of the significant further deterioration within 1 <lb/>h reported for mice under isoflurane (but not ketamine) <lb/>anesthesia, breathing oxygen unaided (Cederholm et al., <lb/>2012). Such a time-dependent deterioration could also <lb/>have been a confounding factor for the auditory nerve <lb/>single-unit thresholds reported here, as these measure-<lb/>ments typically only began with a substantial delay after <lb/>anesthetic induction, due to prolonged surgery. Together, <lb/>these observations point to additional, detrimental changes <lb/>in the long term that would not be observed in short-term <lb/>experiments such as minimally invasive ABR measurements <lb/>or most veterinary procedures. Importantly, there were no <lb/>indications from our EKG monitor that the state of the ani-<lb/>mals may have been compromised. <lb/>Interestingly, a study (Jaensch et al., 2001a,b) exposing <lb/>awake budgerigars to pure oxygen over variable times, <lb/>from 3 h to 3 d, found indicators of oxygen stress by <lb/>reactive oxygen species from the shortest exposure and, <lb/>with longer exposure, additional evidence of pulmonary <lb/>inflammation due to oxygen toxicity. Although their birds <lb/>showed no outward signs, the authors concluded that &quot;in <lb/>a clinical setting, elevated inhalant oxygen tensions <lb/>should be provided to birds with caution, especially if <lb/>prolonged or repeated exposure is anticipated.&quot; This sug-<lb/>gests that there is a point where the delivery of pure <lb/>oxygen changes from being beneficial to damaging. This <lb/>point may well be reached earlier in healthy subjects, as <lb/>typically used in a neurophysiological research setting, <lb/>compared with veterinary patients. Data on respiratory or <lb/>other metabolic parameters, such as arterial partial pres-<lb/>sure of O 2 , p CO2 , or blood pressure and pH, would thus be <lb/>desirable but were not monitored in any of the auditory <lb/>studies. Unfortunately, they are known to be difficult to <lb/>obtain for small birds with body weights Ͻ400 g (Des-<lb/>marchelier et al., 2007; Lierz and Korbel, 2012). <lb/>General implications for invasive neurophysiology <lb/>The above findings suggest that despite their advan-<lb/>tages regarding animal welfare, isoflurane and related <lb/>inhalant anesthetics are not the first choice for experi-<lb/>ments in sensory and neurophysiology. The present arti-<lb/>cle focused on auditory physiology; however, there is <lb/>similar evidence, for example, for the visual cortex (Mi-<lb/>chelson and Kozai, 2018). Researchers should be aware <lb/>that sensory responses may already be depressed at the <lb/>most peripheral levels, as now shown clearly for the au-<lb/>ditory nerve. Effects on peripheral responses must also be <lb/>expected to propagate through brain nuclei. Indeed, re-<lb/>ports of reduced sensitivity and neural activity under iso-<lb/>flurane anesthesia in mammalian auditory cortex (Cheung <lb/>et al., 2001; Noda and Takahashi, 2015) are plausibly <lb/>explained by the extent of peripheral effects observed <lb/>here and in other studies. One should also be aware, <lb/>however, that the brain is not organized along one-<lb/>dimensional hierarchies. For example, the auditory sys-<lb/>tems of both birds and mammals show a multitude of <lb/>ascending and descending interconnections (Smith and <lb/>Spirou, 2002; Bolhuis et al., 2010). This makes it difficult <lb/>to predict to what extent a given peripheral impairment <lb/>will be evident in higher-order responses. Direct actions of <lb/>inhalants on higher-order neurons may add to any inher-<lb/>ited effects. <lb/>The degradation of auditory sensitivity was dose de-<lb/>pendent in the present and previous studies. In order to <lb/>minimize the detrimental effects of inhalant anesthetics, <lb/>their dosage should therefore be individually adjusted as <lb/>low as possible. In the veterinary literature, one recom-<lb/>mended way to reduce the dose further is to administer a <lb/>combination of isoflurane (or related inhalants) and nitrous <lb/>oxide (Korbel, 1998). However, regarding sensory re-<lb/>sponses, we caution that this may trade one evil for an <lb/>even worse one. Sloan et al. (2010) compared ABR, as <lb/>well as somatosensory and visual evoked responses in <lb/>the baboon, measured under anesthesia with different <lb/>proportionate mixtures of isoflurane and nitrous oxide. <lb/>They found evidence for a synergistic action of the two <lb/>agents (i.e., the combination produced more drastic ef-<lb/>fects on the sensitivity and latency of the responses than <lb/>predicted from a simple addition of the individual effects <lb/>of isoflurane and nitrous oxide). Consistent with that, <lb/>Anderson and Young (2004) found that adding nitrous <lb/>oxide in order to reduce the isoflurane necessary did not <lb/></body>
92
+
93
+ <note place="headnote">New Research <lb/></note>
94
+
95
+ <page>12 of 14 <lb/></page>
96
+
97
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org <lb/></note>
98
+
99
+ <body>avoid the depressive effects on sensitivity, discharge rate, <lb/>and extent of inhibition in cat dorsal cochlear nucleus <lb/>units. <lb/>Finally, there was clear evidence for additional detri-<lb/>mental effects on auditory sensitivity related to the modes <lb/>of respiration. While it may seem trivial that adequate <lb/>respiration needs to be provided, the observed effects <lb/>could not simply be traced to hypoxia. Instead, we ob-<lb/>tained tentative evidence for oxygen toxicity developing <lb/>over time and, furthermore, observed an unexplained de-<lb/>terioration when the owls breathed carbogen. Whether <lb/>these effects occur only in conjunction with isoflurane, or <lb/>may at least be exacerbated by it, remains to be shown. <lb/>Our results suggest that for prolonged experiments with <lb/>healthy experimental animals, normal air is the best op-<lb/>tion. <lb/></body>
100
+
101
+ <listBibl>References <lb/>Anderson MJ, Young ED (2004) Isoflurane/N2O anesthesia sup-<lb/>presses narrowband but not wideband inhibition in dorsal cochlear <lb/>nucleus. Hear Res 188:29 -41. CrossRef <lb/>Antkowiak B (2001) How do general anaesthetics work? Naturwis-<lb/>senschaften 88:201-213. Medline <lb/>Bolhuis JJ, Okanoya K, Scharff C (2010) Twitter evolution: converg-<lb/>ing mechanisms in birdsong and human speech. Nat Rev Neurosci <lb/>11:747-759. CrossRef Medline <lb/>Bremen P, Poganiatz I, Campenhausen M, Wagner H (2007) Sensi-<lb/>tivity to interaural time difference and representation of azimuth in <lb/>central nucleus of inferior colliculus in the barn owl. J Comp <lb/>Physiol A 193:99 -112. CrossRef Medline <lb/>Burger RE, Lorenz FW (1960) Artificial respiration in birds by unidi-<lb/>rectional air flow. Poultry Sci 39:236 -237. CrossRef <lb/>Burns P (2014) Isoflurane &amp; sevoflurane: mechanics &amp; effects. Clin <lb/>Brief 23-26. Available at https://www.cliniciansbrief.com/article/ <lb/>isoflurane-sevoflurane-mechanics-effects <lb/>Campagna JA, Miller KW, Forman SA (2003) Mechanisms of actions <lb/>of inhaled anesthetics. N Engl J Med 348:2110 -2124. CrossRef <lb/>Medline <lb/>Cederholm JM, Froud KE, Wong AC, Ko M, Ryan AF, Housley GD <lb/>(2012) Differential actions of isoflurane and ketamine-based an-<lb/>aesthetics on cochlear function in the mouse. Hear Res 292:71-<lb/>79. CrossRef Medline <lb/>Cheung SW, Nagarajan SS, Bedenbaugh PH, Schreiner CE, Wang X, <lb/>Wong A (2001) Auditory cortical neuron response differences un-<lb/>der isoflurane versus pentobarbital anesthesia. Hear Res 156:115-<lb/>127. CrossRef <lb/>Cohen YE, Knudsen EI (1995) Binaural tuning of auditory units in the <lb/>forebrain archistriatal gaze fields of the barn owl: local organization <lb/>but no space map. J Neurosci 15:5152-5168. Medline <lb/>Desmarchelier M, Rondenay Y, Fitzgerald G, Lair S (2007) Monitoring <lb/>of the ventilatory status of anesthetized birds of prey by using <lb/>end-tidal carbon dioxide measured with a microstream capnom-<lb/>eter. J Zoo Wildl Med 38:1-6. CrossRef Medline <lb/>Dodd F, Capranica RR (1992) A comparison of anesthetic agents and <lb/>their effects on the response properties of the peripheral auditory <lb/>system. Hear Res 62:173-180. CrossRef <lb/>Drexl M, Henke J, Kössl M (2004) Isoflurane increases amplitude and <lb/>incidence of evoked and spontaneous otoacoustic emissions. <lb/>Hear Res 194:135-142. CrossRef Medline <lb/>Eatock RA, Lysakowski A (2006) Mammalian vestibular hair cells. In: <lb/>Vertebrate hair cells (Eatock RA, Fay RR, Popper AN, eds), pp <lb/>348 -442. New York: Springer. <lb/>Ferber-Viart C, Preckel MP, Dubreuil C, Banssillon V, Duclaux R <lb/>(1998) Effect of anesthesia on transient evoked otoacoustic emis-<lb/>sions in humans: a comparison between propofol and isoflurane. <lb/>Hear Res 121:53-61. Medline <lb/>Flaherty D (2009) Anaesthetic drugs. In: Anaesthesia for veterinary <lb/>nurses (Welsh L, ed), pp 121-161. Chichester, UK: Wiley-<lb/>Blackwell. <lb/>Glowatzki E, Grant L, Fuchs P (2008) Hair cell afferent synapses. Curr <lb/>Opin Neurobiol 18:389 -395. CrossRef Medline <lb/>Gungor G, Bozkurt-Sutas P, Gedik O, Atas A, Babazade R, Yilmaz M <lb/>(2015) Effects of sevoflurane and desflurane on otoacoustic emis-<lb/>sions in humans. Eur Arch Otorhinolaryngol 272:2193-2199. <lb/>CrossRef Medline <lb/>Gunkel C, Lafortune M (2005) Current techniques in avian anesthe-<lb/>sia. Semin Avian Exotic Pet Med 14:263-276. CrossRef <lb/>Ishizawa Y (2007) Mechanisms of anesthetic actions and the brain. J <lb/>Anesth 21:187-199. CrossRef Medline <lb/>Jaensch S, Cullen L, Raidal SR (2001a) Normobaric hyperoxic stress <lb/>in budgerigars: enzymic antioxidants and lipid peroxidation. Comp <lb/>Biochem Physiol C Toxicol Pharmacol 128:173-180. CrossRef <lb/>Jaensch S, Cullen L, Morton L, Raidal SR (2001b) Normobaric hy-<lb/>peroxic stress in budgerigars: non-enzymic antioxidants. Comp <lb/>Biochem Physiol C Toxicol Pharmacol 128:181-187. CrossRef <lb/>Katz E, Elgoyhen AB, Fuchs PA (2011) Cholinergic inhibition of hair <lb/>cells. In: Auditory and vestibular efferents (Ryugo D, Fay RR, <lb/>Popper AN, eds), pp 103-133. New York: Springer. <lb/>Keegan RD (2005) Inhalant anesthetics: the basics. In: Recent ad-<lb/>vances in veterinary anesthesia and analgesia: companion animals <lb/>(Gleed RD, Ludders JW, eds). Ithaca, NY: International Veterinary <lb/>Information Service. <lb/>Keller CH, Takahashi TT (2005) Localization and identification of <lb/>concurrent sounds in the owl&apos;s auditory space map. J Neurosci <lb/>25:10446 -10461. CrossRef Medline <lb/>Kettembeil S, Manley GA, Siegl E (1995) Distortion-product otoa-<lb/>coustic emissions and their anaesthesia sensitivity in the European <lb/>Starling and the chicken. Hear Res 86:47-62. Medline <lb/>Köppl C (1997) Frequency tuning and spontaneous activity in the <lb/>auditory nerve and cochlear nucleus magnocellularis of the barn <lb/>owl, Tyto alba. J Neurophysiol 77:364 -377. CrossRef <lb/>Köppl C (2011) Evolution of the octavolateral efferent system. In: <lb/>Auditory and vestibular efferents (Ryugo D, Fay RR, Popper AN, <lb/>eds), pp 217-259. New York: Springer. <lb/>Köppl C, Gleich O (2007) Evoked cochlear potentials in the barn owl. <lb/>J Comp Physiol A Neuroethol Sens Neural Behav Physiol 193:601-<lb/>612. CrossRef <lb/>Köppl C, Nickel R (2007) Prolonged maturation of cochlear function <lb/>in the barn owl after hatching. J Comp Physiol A Neuroethol Sens <lb/>Neural Behav Physiol 193:613-624. CrossRef <lb/>Köppl C, Futterer E, Nieder B, Sistermann R, Wagner H (2005) <lb/>Embryonic and posthatching development of the barn owl (Tyto <lb/>alba): reference data for age determination. Dev Dyn 233:1248 -<lb/>1260. CrossRef Medline <lb/>Korbel R (1998) Vergleichende Untersuchungen zur Inhalationsanäs-<lb/>thesie mit Isofluran (Forene) und Sevofluran (SEVOrane) bei <lb/>Haustauben (Columba livia Gmel., 1789, var. domestica) und Vor-<lb/>stellung eines Referenz-Narkoseprotokolls für Vögel. Tierärztliche <lb/>Praxis 26:211-223. <lb/>Larsen ON, Christensen-Dalsgaard J, Jensen KK (2016) Role of <lb/>intracranial cavities in avian directional hearing. Biol Cybern 110: <lb/>319 -331. CrossRef Medline <lb/>Lierz M, Korbel R (2012) Anesthesia and analgesia in birds. J Exotic <lb/>Pet Med 21:44 -58. CrossRef <lb/>Lukasik VM, Gillies RJ (2003) Animal anaesthesia for in vivo magnetic <lb/>resonance. NMR Biomed 16:459 -467. CrossRef Medline <lb/>Michelson NJ, Kozai TDY (2018) Isoflurane and ketamine differen-<lb/>tially influence spontaneous and evoked laminar electrophysiology <lb/>in mouse V1. J Neurophyiol. Advance online publication. Retrieved <lb/>October 27, 2018. doi:10.1152/jn.00299.2018. <lb/>Noda T, Takahashi H (2015) Anesthetic effects of isoflurane on the <lb/>tonotopic map and neuronal population activity in the rat auditory <lb/>cortex. Eur J Neurosci 42:2298 -2311. CrossRef Medline <lb/>O&apos;Keeffe NJ, Healy TEJ (1999) The role of new anesthetic agents. <lb/>Pharmacol Ther 84:233-248. Medline <lb/></listBibl>
102
+
103
+ <note place="headnote">New Research <lb/></note>
104
+
105
+ <page>13 of 14 <lb/></page>
106
+
107
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org <lb/></note>
108
+
109
+ <listBibl>Powell FL (2015) Respiration. In: Sturkie&apos;s avian physiology, Vol6 <lb/>(Scanes CG, ed), pp 301-336. Amsterdam: Elsevier. <lb/>Preckel B, Bolten J (2005) Pharmacology of modern volatile anaes-<lb/>thetics. Best Pract Res Clin Anaesthesiol 19:331-348. Medline <lb/>Raftery A (2013) Avian anaesthesia. In Pract 35:272-278. CrossRef <lb/>Ropposch T, Walch C, Avian A, Mausser G, Spary M (2014) Effects <lb/>of the depth of anesthesia on distortion product otoacoustic emis-<lb/>sions. Eur Arch Otorhinolaryngol 271:2897-2904. CrossRef Med-<lb/>line <lb/>Rudolph U, Antkowiak B (2004) Molecular and neuronal substrates <lb/>for general anaesthetics. Nat Rev Neurosci 5:709 -720. CrossRef <lb/>Medline <lb/>Ruebhausen MR, Brozoski TJ, Bauer CA (2012) A comparison of the <lb/>effects of isoflurane and ketamine anesthesia on auditory brains-<lb/>tem response (ABR) thresholds in rats. Hear Res 287:25-29. <lb/>CrossRef Medline <lb/>Ruel J, Wang J, Rebillard G, Eybalin M, Lloyd R, Pujol R, Puel JL <lb/>(2007) Physiology, pharmacology and plasticity at the inner hair <lb/>cell synaptic complex. Hear Res 227:19 -27. CrossRef Medline <lb/>Schwartzkopff J, Brémond JC (1963) Méthode de dérivation des <lb/>potentiels cochléaires chez lЈoiseau. J Physiol Paris 55:495-518. <lb/>Shawyer CR (1998) The barn owl. Chelmsford, UK: Arlequin. <lb/>Sloan T, Sloan H, Rogers J (2010) Nitrous oxide and isoflurane are <lb/>synergistic with respect to amplitude and latency effects on sen-<lb/>sory evoked potentials. J Clin Monit Comput 24:113-123. Cross-<lb/>Ref Medline <lb/>Smith DI, Mills JH (1989) Anesthesia effects: auditory brain-stem <lb/>response. Electroencephal Clin Neurophysiol 72:422-428. Medline <lb/>Smith PH, Spirou GA (2002) From the cochlea to the cortex and <lb/>back. In: Integrative functions in the mammalian auditory pathway <lb/>(Oertel D, Fay RR, Popper AN, eds), pp 6 -71. New York: Springer <lb/>Verlag. <lb/>Sonner JM, Antognini JF, Dutton RC, Flood P, Gray AT, Harris RA, <lb/>Homanics GE, Kendig J, Orser B, Raines DE, Trudell J, Vissel B, <lb/>Eger EI (2003) Inhaled anesthetics and immobility: mechanisms, <lb/>mysteries, and minimum alveolar anesthetic concentration. Anesth <lb/>Analg 97:718 -740. Medline <lb/>Stronks HC, Aarts MCJ, Klis SFL (2010) Effects of isoflurane on <lb/>auditory evoked potentials in the cochlea and brainstem of guinea <lb/>pigs. Hear Res 260:20 -29. CrossRef Medline <lb/>Vahle-Hinz C, Detsch O (2002) What can in vivo electrophysiology in <lb/>animal models tell us about mechanisms of anaesthesia? Br J <lb/>Anaesth 89:123-142. Medline <lb/>Varner J, Clifton KR, Broderson R, Wyatt RD (2004) Lack of efficacy <lb/>of injectable ketamine with xylazine or diazepam for anesthesia in <lb/>chickens. Lab Anim 33:36 -39. CrossRef Medline <lb/>Windels F (2006) Neuronal activity: from in vitro preparation to be-<lb/>having animals. Mol Neurobiol 34:1-25. Medline <lb/></listBibl>
110
+
111
+ <note place="headnote">New Research <lb/></note>
112
+
113
+ <page>14 of 14 <lb/></page>
114
+
115
+ <note place="footnote">September/October 2018, 5(5) e0140-18.2018 <lb/>eNeuro.org </note>
116
+
117
+
118
+ </text>
119
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/ENEURO.0415-18.2019.training.segmentation.tei.xml ADDED
The diff for this file is too large to render. See raw diff
 
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EPJ0290369-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <front>High field magnetoresistivity of epitaxial La 2-x Sr x CuO 4 thin films <lb/>J. Vanacken 1,2 , L. Weckhuysen , T. Wambecq , P. Wagner , and V.V. Moshchalkov <lb/>Laboratorium voor Vaste-Stoffysica en Magnetisme, Katholieke Universiteit Leuven, <lb/>Celestijnenlaan 200 D, B-3001 Leuven, Belgium <lb/>Laboratoire National de Champs Magnétiques Pulsés, <lb/>143, Avenue de Rangueil BP 4245, F31432 Toulouse, France <lb/>3 Instituut voor Materiaalonderzoek, Limburgs Universitair Centrum, <lb/>Wetenschapspark 1, B-3590 Diepenbeek, Belgium <lb/>Abstract <lb/>A large positive magnetoresistivity (up to tens of percents) is observed in both underdoped and <lb/>overdoped superconducting La 2-x Sr x CuO 4 epitaxial thin films at temperatures far above the <lb/>superconducting critical temperature T c . For the underdoped samples, this magnetoresistance far <lb/>above T c cannot be described by the Kohler rule and we believe it is to be attributed to the influence of <lb/>superconducting fluctuations. In the underdoped regime, the large magnetoresistance is only present <lb/>when at low temperatures superconductivity occurs. T he strong magnetoresistivity, which persists <lb/>even at temperatures far above T c , can be related to the pairs forming eventually the superconducting <lb/>state below T c . Our observations support the idea of a close relation between the pseudogap and the <lb/>superconducting gap and provide new indications for the presence of pairs above T c . <lb/>PACS numbers: 74.25.-q, 74.25.Dw, 74.25.Fy, 74.40.+k <lb/></front>
9
+
10
+ <body>Introduction <lb/>One of the unusual features of high T c layered superconductors is the opening of a pseudo-gap in the <lb/>electronic energy spectrum at a temperature T * far above the critical temperature T c . Although the existence <lb/>of the pseudo-gap is commonly accepted and confirmed by several experimental techniques such as nuclear <lb/>magnetic resonance (NMR), tunneling spectroscopy, angle resolved photoemission spectroscopy (ARPES) <lb/>and electronic Raman scattering [1], the origin of the pseudo-gap is still not revealed. T he intriguing <lb/>question is whether the pseudogap and the superconducting gap have a common origin. If they are related, <lb/>the pseudogap might be associated with the presence of electronic pairs above T c . From this point of view, <lb/>superconductivity occurs when the phase of these pairs becomes coherent and not when they are first formed <lb/>in the phase incoherent state [2]. T he idea of a precursory pair formation at relatively high temperatures <lb/>T c &lt; T &lt; T * and its relevance for high T c superconductivity is supported by the experimental observation that <lb/>the pseudogap evolves into the superconducting gap at low temperatures, as clearly demonstrated by <lb/>scanning tunneling spectroscopy [3]. Moreover, the ARPES data [4,5] indicate that the pseudo-and the <lb/>superconducting gap both have d-wave symmetry. T he fact that the T *(p)-and the T c (p)-lines merge in the <lb/>overdoped regime [6], with p the hole density, may explain the difficulty to observe a pseudogap in the <lb/>strongly overdoped case. If preformed pairs exist, they should also influence the normal state transport <lb/>properties of high T c superconductors at temperatures T c &lt; T &lt; T *. Our paper is focussed on the in-plane <lb/>magnetoresistivity of the prototype system La 2-x Sr x CuO 4 that covers completely both the underdoped <lb/>(x&lt;0.15) and the overdoped (x&gt;0.15) regimes. T he results from earlier reports on the magnetoresistivity of <lb/>this system are rather contradictory: a negative magnetoresistivity has been obtained from pulsed field <lb/>transport measurements [7]; a positive [8,9,10] as well as a negative magnetoresistivity [11] has been <lb/>reported from DC field measurements. <lb/>In this article, we present magnetoresistivity data of La 2-x Sr x CuO 4 epitaxial thin films measured in <lb/>pulsed magnetic fields up to 50 T and in the temperature range from room temperature down to 4.2 K. It is <lb/>important to note that the magnetic field will not only destroy superconductivity but can affect, at the same <lb/>time, the scattering mechanisms in the normal state. Moreover, it is not known how the new sorts of <lb/>quasiparticles, introduced by theorists to explain the pseudogap phase, behave in an applied magnetic field. <lb/></body>
11
+
12
+ <body>T herefore, the study of the field-dependence of the resistivity is indispensable in that respect. By <lb/>systematically changing the hole concentration through the variation of the Sr-content, x, we have found that <lb/>superconductivity at low temperatures (T &lt; T c ) and a considerable positive magnetoresistivity at high <lb/>temperatures (T &gt;&gt; T c ) both appear at the same Sr-content, thus relating the large magnetoresistance with <lb/>superconductivity. We will present clear evidence for precursor effects from the high field transport data. <lb/>Expe rime ntal re sults <lb/>T he as-grown films were prepared by DC magnetron sputtering from stoichiometric targets [12,13]. The <lb/>magnetoresistivity measurements were carried out at the pulsed field facility of the Katholieke Universiteit <lb/>Leuven [14,15] by using a homemade flow-cryostat and 50 T coil. All data reported in this paper were <lb/>obtained on thin films (~150 nm), patterned (1000 x 50 µm strip) for four probe measurements in the <lb/>transverse geometry (µ H ⊥ I) with the magnetic field perpendicular to the film (µ 0 H // c) and the current <lb/>sent along the ab-plane (I // ab). <lb/></body>
13
+
14
+ <body>Figures 1 to 8 present the ρ ab (µ o H) curves measured at temperatures varying from T &gt;&gt; T c down to 4.2 K for <lb/>the La 2-x Sr x CuO 4 thin films with Sr content x = 0.045, 0.050, 0.055, 0.060, 0.100, 0.200, 0.250 and 0.270. <lb/>La 1.955 Sr 0.045 CuO 4 <lb/>T he La 1.955 Sr 0.045 CuO 4 sample (Figure 1) shows a very weak magnetoresistivity (less than 2 % at 45 T) in <lb/>the whole temperature range. T his is clearly demonstrated by the graphs A, B, C and D in the middle part of <lb/>Figure 1, which present the weak, in a first approximation, quadratic magnetoresistivity at the selected <lb/>temperatures 16 K, 20 K, 32 K and 176 K. No smoothing has been applied to the data, and both raising and <lb/>lowering field branches are shown. T he magnetoresistivity, indicated in the graphs is defined as <lb/>MR = (ρ ab (50 T ) -ρ(0 T ))/ρ(0 T ). <lb/>T he in-plane resistivity ρ ab (T) as a function of temperature at zero magnetic field is shown in Figure 1 <lb/>at the right side of the upper frame; it serves to better orientate the magnetoresistivity measurements. The open <lb/>circles denote the values of the resistivity at zero magnetic field, derived from pulsed field measurements. <lb/>Since La 1.955 Sr 0.045 CuO 4 exhibits, below T MI ~ 100 K, a resistivity that strongly diverges when lowering the <lb/>temperature (e.g. dρ/dT(4.2 K) ≈ -800 µΩcm/K), even minor heating effects in the pulsed field experiment can <lb/>artificially lead to negative magnetoresistivity effects at low temperatures. We judged that the <lb/>magnetoresistivity of La 1.955 Sr 0.045 CuO 4 at temperatures T &gt; 14 K could be adequately measured up to 50 T in <lb/>our setup. Indeed, no discrepancies between data taken during the rising and the lowering branch of the field <lb/>pulse could be found in this temperature range, a strong indication that heating effects do not influence the <lb/>results. Additional measurements in DC fields up to 8 T (not shown in this article), convincingly proved that <lb/>the resistivity of the sample is only slightly magnetic field dependent below 14 K as well. <lb/>La 1.95 Sr 0.05 CuO 4 <lb/>Upon approaching the insulator-superconductor transition in the (T,x)-phase diagram (x = 0.055), a <lb/>considerable positive magnetoresistivity appears (Figure 2). A magnetic field of 45 T causes an excess <lb/>resistivity of 10 % in La 1.95 Sr 0.05 CuO 4 at a temperature of 10 K. With increasing temperature, the <lb/>magnetoresistivity of the sample at 45 T goes down to a final decrease below 2 % around 40 K. Note that <lb/>La 1.95 Sr 0.05 CuO 4 does not show a sign of superconductivity at zero magnetic field down to 1.5 K, the lowest <lb/>temperature investigated. In contrast, the sample demonstrates an insulator-like behavior (dρ ab /dT &lt; 0) from <lb/>80.5 K (T MI ) down to the lowest temperature. For clarity, only the data taken during rising magnetic field are <lb/>shown in graphs A, B, C, and D of Figure 2. T he overview graph in the upper frame of Figure 2 depicts, at <lb/>the different temperatures, the data at zero field (open circles) and at 45 T (solid circles). <lb/>La 1.945 Sr 0.055 CuO 4 <lb/>Figure 3 illustrates that La 1.945 Sr 0.055 CuO 4 , situated at the border of the superconducting phase, <lb/>manifests strong magnetoresistivity effects. At 4.2 K, the magnetoresistivity at 45 T is 33 %; its value at <lb/>9.4 K is 18 %. Although situated very close to the insulator-superconductor transition, La 1.945 Sr 0.055 CuO 4 has <lb/>a robust insulator-like behavior from 72.7 K (T MI ) down to 1.5 K at zero magnetic field, seemingly not to be <lb/>correlated with the occurrence of superconductivity. T he graphs A, B, C, D, E and F in the lower part of <lb/>
15
+ Figure 3 give a clear presentation of the evolution of the resistivity with magnetic field for the <lb/>La 1.945 Sr 0.055 CuO 4 sample. At low temperatures (4.2 K), a saturating ρ(µ 0 H) behavior is observed. For <lb/>intermediate temperatures (20.7 K), a quadratic magnetoresistivity at low fields evolves into a behavior that <lb/>tends to saturate at higher fields. Only a quadratic behavior of the magnetoresistivity remains at sufficiently <lb/>high temperatures (47.7 K). Graphs A, B, C and D only show the data taken during the increasing branch of <lb/>the magnetic field pulse. <lb/>La 1.94 Sr 0.06 CuO 4 <lb/>In the La 1.94 Sr 0.06 CuO 4 compound (Figure 4), the insulating phase at low temperatures gives way to <lb/>superconductivity below T c = 2.4 K. T he rather low critical temperature T c = 2.4 K implies that the sample is <lb/>located very close to the insulator-superconductor transition. From figure 4 it is clear that the <lb/>magnetoresistivity effects become very pronounced upon increasing the charge carrier concentration through <lb/>the superconducting phase. At 4.2 K, which is nearly two times T c , a field of 45 T causes a magnetoresistivity <lb/>of 330 % in La 1.94 Sr 0.06 CuO 4 . Note that the resistivity is not even fully saturated at 45 T. Upon increasing the <lb/>temperature, the impact of the magnetic field on the resistivity diminishes, resulting in a crossing of the <lb/>ρ ab (µ o H) curves taken at temperatures below T MI = 63 K. T he fact that the ρ ab (µ o H) curves cross each other <lb/>reflects that the ground state at low temperatures, obscured below T c by the superconducting phase, has an <lb/>insulating character in La 1.94 Sr 0.06 CuO 4 . T his observation is in agreement with the results reported previous in <lb/>[7, 16] on underdoped superconducting La 2-x Sr x CuO 4 single crystals. T he temperature dependence of the <lb/>resistivity at 45 T is shown at the right side of the upper frame of Figure 4 by black circles. Below the metal-<lb/>to-insulator transition at T MI = 63 K, the resistivity at 45 T exhibits insulating properties (dρ ab /dT &lt; 0). The <lb/>graphs A, B, C, D, E and F show the functional dependence of the resistivity versus field in detail. T he <lb/>magnetoresistivity tends to saturate at low temperatures (4.2 K). Similar to the La 1.945 Sr 0.055 CuO compound <lb/>(Figure 3), this behavior gradually evolves into a quadratic dependence (50 K) upon increasing the temperature. <lb/>T o a lower extent, this behavior can be as well seen in the La 1.95 Sr 0.05 CuO 4 sample presented in Figure 2. <lb/>La 1.9 Sr 0.1 CuO 4 <lb/>
16
+ T he magnetoresistivity in La 1.9 Sr 0.1 CuO 4 (T c = 17.5 K) is shown in Figure 5. Again, the study in high magnetic <lb/>fields reveals an insulating ground state (dρ ab /dT &lt; 0) behind the superconducting phase, which is hidden in <lb/>zero magnetic field. T he use of high magnetic fields allows us to determine the metal to insulator transition <lb/>temperature T MI = 56 K. Below T MI = 56 K, the ρ ab (µ o H) curves cross each other. It looks like the curves <lb/>have a single intersection in the graph at the left side of the upper frame of Figure 5. However, an enlarged <lb/>view of this field region indicates that the crossing shifts systematically to higher fields when lowering the <lb/>temperature. Graphs A, B, C, D, E and F in Figure 5 provide a closer look upon the field dependence of the <lb/>resistivity in La 1.9 Sr 0.1 CuO 4 at the selected temperatures 4.2 K, 12 K, 18 K, 30 K, 71 K and 132.7 K. Below <lb/>T c = 17.5 K, the magnetoresistivity tends to saturate. Nevertheless, a complete saturation is still absent at all <lb/>temperatures. At the same time, the ρ ab (µ 0 H) curves do not exhibit a knee-shaped feature, marking the <lb/>position of the second critical field (H c2 ). Above T c = 17.5 K, the magnetoresistivity exhibits a familiar <lb/>behavior: it tends to saturate at low temperatures (18 K), a quadratic dependence at low fields bends down <lb/>with increasing field at intermediate temperatures (30 K) and a weak quadratic dependence remains at high <lb/>temperatures (71 K). T he magnetoresistivity at 45 T decreases from 30 % at 30 K down to below 2 % above <lb/>71 K. At high temperatures, for example 132.7 K, the magnetoresistivity for the superconducting <lb/>La 1.9 Sr 0.1 CuO 4 sample is comparable to that of the distinctly non-superconducting La 1.955 Sr 0.045 -CuO 4 <lb/>(Figure 1). <lb/>La 1.8 Sr 0.2 CuO 4 <lb/>In contrast to the underdoped samples, the overdoped La 1.8 Sr 0.2 CuO 4 (T c = 22.8 K) demonstrates <lb/>clear knee-shaped features in its field-dependent resistivity curves ρ ab (µ o H) at temperatures T &lt; T c (see <lb/>graphs A, B and C of Figure 6). As a consequence, the superconducting transitions and the second critical <lb/>field H c2 (T) can be determined for La 1.8 Sr 0.2 CuO 4 . Above H c2 (T), the resistivity still depends on the magnetic <lb/>field. At the same time, the overdoped La 1.8 Sr 0.2 CuO 4 shows magnetoresistivity, far above T c , dying out with <lb/>increasing temperature, similar to the superconducting underdoped samples La 1.94 Sr 0.06 CuO 4 and <lb/>La 1.9 Sr 0.1 CuO 4 . However, the functional dependence of the magnetoresistivity changed while crossing the <lb/>threshold of optimal doping. No tendency towards saturation has been observed in the magnetoresistivity of <lb/>La 1.8 Sr 0.2 CuO 4 , at temperatures well above the superconducting to normal transition. T his is illustrated by <lb/>
17
+ graphs D and E in Figure 6. A linear dependence of the resistivity with respect to the field is found at low <lb/>temperatures (30.6 K). At higher temperatures (79.6 K), a quadratic behavior is predominant. La 1.8 Sr 0.2 CuO 4 <lb/>exhibits a minimum in ρ ab (T) at T MI = 40 K and a crossover from metallic to insulator-like behavior upon a <lb/>temperature decrease, in a high magnetic field of 45 T . A low temperature insulating behavior persisting up <lb/>to optimal doping is reported both for La 2-x Sr x CuO 4 single crystals [16] and for the electron-doped <lb/>superconductor Pr 2-x Ce x CuO 4 [17]. In Bi 2 Sr 2-x -La x CuO 6+δ , it disappears at 1/8 hole doping, in the <lb/>underdoped regime [18]. Our results on thin films, on the other hand, show a metal to insulator-like <lb/>transition in La 1,8 Sr 0.2 CuO 4 , stretching well into the overdoped regime. <lb/>La 1.75 Sr 0.25 CuO 4 <lb/>T he resistivity data for the strongly overdoped La 1.75 Sr 0.25 CuO 4 compound (T c = 15.6 K) is shown in Figure 7 as a <lb/>function of the magnetic field. T he data for both branches of the field pulse are presented and no smoothing has <lb/>been applied. T he La 1.75 Sr 0.25 CuO 4 sample shows various kinds of superconducting transitions. While close to T c , the <lb/>ρ ab (µ o H) transition is narrow, it broadens significantly when lowering the temperature. This indicates that the <lb/>irreversibility line H irr (T) and the second critical field H c2 (T) gradually separate from each other. The graphs A, B <lb/>and C of Figure 7 illustrate that the resistivity of La 1.75 Sr 0.25 CuO 4 is linearly depending on the field above the <lb/>critical fields. Above T c (= 15.6 K), La 1.75 Sr 0.25 CuO 4 still demonstrates a considerable quadratic <lb/>magnetoresistivity but the ρ ab (µ o H)-curves lack any sign of saturation. Surprisingly, our high field studies <lb/>disclose a metal to insulator transition around 10 K (T MI ) in La 1.75 Sr 0.25 CuO 4 . Although this sample is <lb/>situated deeply in the overdoped regime, the resistivity values at 50 T demonstrate a distinct upturn when <lb/>lowering the temperature below T MI . T his insulator-like behavior at low temperatures could be related to a <lb/>weak pseudogap feature present even in this compound, or to disorder effects. <lb/>La 1.73 Sr 0.27 CuO 4 <lb/>Finally, Figure 8 contains the field dependent in-plane resistivity data of La 1.73 Sr 0.27 CuO 4 (T c = 9.2 K), our <lb/>strongest overdoped sample in this study. T he superconducting transitions are clear and sharp, just like in <lb/>La 1.75 Sr 0.25 CuO 4 . Since La 1.73 Sr 0.27 CuO 4 has, of all our samples, the lowest normal state resistivity, the <lb/>mechanical vibrations, caused by the high field-pulses, have a stronger impact on its resistivity data. The <lb/>effects of the vibrations, particularly present during the lowering branch of the pulse, are clearly visible in <lb/>the data taken at 12 K and 54 K, respectively shown in graphs B and D in Figure 8. T he other <lb/>magnetoresistivity curves, presented in Figure 8, contain only the data taken during increasing magnetic field <lb/>for clarity. In contrast with the previous samples, the ρ ab (µ o H)-curves of La 1.73 Sr 0.27 CuO 4 do not cross each <lb/>other at fields below 45 T . Above T c , the magnetic field dependences of the resistivity are essentially <lb/>quadratic, thus without tendency towards saturation. <lb/>In the field and temperature range used in our experiments, the magnetoresistivity is positive for all our <lb/>films, in agreement with the results of [19, 20, 21]. Our data on thin films differ from the results of <lb/>references [7, 16] on single crystals of La 2-x Sr x -CuO 4 with x = 0.08 and x = 0.13, where the <lb/>magnetoresistivity in the limit of high fields was found to be negative. <lb/>
18
+ Discussion <lb/>In the analysis of superconducting fluctuations the need to separate fluctuation and normal-state <lb/>conductivity contributions in the analysis is eminent. T his is a particularly difficult task in high-T c <lb/>superconductors, because both their normal state and their superconducting behavior are not yet understood. <lb/>In what follows, we present a possible analysis of the magnetoresistivity data of the La 2-x Sr x CuO 4 thin films. <lb/>Our observations give new and strong evidence for stripe formation in underdoped samples. <lb/>Absence of magnetoresistivvity in non-superconducting underdoped La 1.955 Sr 0.045 CuO 4 <lb/>Figure 1 (lower frame) presents the in-plane magnetoresistivity data for the non-superconducting <lb/>La 1.955 Sr 0.045 CuO 4 sample in a so-called Kohler-plot ((ρ-ρ 0 )/ρ 0 vs B 2 ). T he figure suggests that its resistivity <lb/>is proportional to H 2 at all temperatures. Kohler&apos;s rule [22] is however only valid when all the curves <lb/>coincide. We see that the curves at 87 K and 176 K nicely overlap but that the low temperature data deviate. <lb/>A violation of Kohler&apos;s rule at low temperatures can be expected for this compound since its low temperature <lb/>region is characterized by variable range hopping conductivity. It is most unlikely that one can describe the <lb/>charge transport in this temperature regime by a rule, based on a classical Boltzmann theory for metals. <lb/>Magnetoresistivity of La 2-x Sr x CuO 4 close to the insulator-superconductor transition <lb/>T he lower frames of Figures 2 and 3, show the Kohler plots for the La 2-x Sr x CuO 4 samples with x = 0.05 and 0.055, <lb/>situated very close to the insulator to superconductor transition. A considerable excess magnetoresistivity, which <lb/>depends not quadratically on the magnetic field but rather tends to saturate, appears at low temperatures. This <lb/>contribution becomes more pronounced when increasing the charge carrier concentration through the <lb/>superconducting phase, as evidenced by the Kohler plots for La 1.94 Sr 0.06 CuO 4 (T c = 2.4 K) and La 1.9 Sr 0.1 CuO <lb/>(T c = 17.5 K), presented in Figures 4 and 5, respectively. T his evolution strongly suggests that the non-quadratic <lb/>contribution to the magnetoresistivity can be attributed to superconducting fluctuations. <lb/>In zero magnetic field, La 1.95 Sr 0.05 CuO 4 and La 1.945 Sr 0.055 CuO 4 demonstrate an insulator-like behavior <lb/>(dρ ab /dT &gt; 0) at low temperatures down to at least 1.5 K. Nevertheless, the superconducting fluctuations <lb/>appear up to tens of Kelvin. For the superconducting samples La 1.94 Sr 0.06 CuO and La 1.9 Sr 0.1 CuO 4 , the <lb/>fluctuations extend over a temperature range, which exceeds several times T c . For example: the <lb/>La 1.94 Sr 0.06 CuO 4 compound shows at 25 K (seven times T c !) a magneto-resistivity of 8 % at T , which is <lb/>substantially higher than the ~ 1 % for the non-superconducting La 1.955 Sr 0.045 CuO 4 sample. The fluctuations <lb/>are moreover of an unusual strength: typical fields for a complete suppression of fluctuations in conventional <lb/>BCS bulk superconductors should not exceed the paramagnetic limiting field µ o H = 1.38 T c . However, the <lb/>La 1.94 Sr 0.06 CuO 4 system shows, at 4.2 K, a magnetoresistivity that is not even saturated at 50 T, a value which <lb/>is more than a decade higher than the conventional paramagnetic limit of µ o H = 4 T for this sample with T c ~ <lb/>2.4K. <lb/>Influence of the normal state magnetoresistivity for (strongly) overdoped La 2-x Sr x CuO 4. <lb/>T he lower frames of figures 6-8 present the Kohler plots for the overdoped samples La 1.8 Sr 0.2 CuO 4 <lb/>(T c = 22.8 K), La 1.75 Sr 0.25 CuO 4 (T c = 15.6 K) and La 1.73 Sr 0.27 CuO 4 (T c = 9.2 K). We see that the part of the <lb/>magnetoresistivity, which has no quadratic behaviour with respect to the applied field diminishes upon <lb/>increasing the charge carrier concentration and becomes finally undetectable in the La 1.73 Sr 0.27 CuO 4 sample. <lb/>
19
+ T his observation can be consistently interpreted in the context of our previous results that the one-<lb/>dimensional character of the charge transport (stripes) fades away upon doping [23]. As the dimensionality <lb/>of the electrical transport increases from 1D to 2D (or 3D), the fluctuations should indeed become less <lb/>pronounced. T he evidences for stripe formation from magnetoresistivity measurements are discussed further <lb/>below (see eq. (1)). <lb/>So far, we have ignored the part of the magnetoresistivity that appears in a Kohler plot as a straight line. In <lb/>Figure 8, we see that magnetic field dependencies of the La 1.73 Sr 0.27 CuO 4 -data are essentially H 2 up to 45 T. <lb/>Moreover, the straight lines coincide, which implies that H/ρ o scales the magnetoresistivity and hence that <lb/>the classical Kohler&apos;s rule is for valid in La 1.73 Sr 0.27 CuO 4 . Its magnetoresistivity can safely be attributed to <lb/>the normal state. T he data for the other strongly overdoped sample, La 1.75 Sr 0.25 CuO 4 , follows Kohler&apos;s rule <lb/>above 30 K. For the superconducting samples with a lower Sr content (Figures 2 to 6, lower frames), the <lb/>dependence with (H/ρ o ) 2 remains linear over a wide temperature range (far above T c ) but the slopes increase <lb/>when lowering the temperature. T he deviations are more pronounced in the underdoped samples. Many <lb/>authors observed this apparent violation of Kohler&apos;s rule in high-T c systems and speculated its origin [24, 19, <lb/>20, 25, 26, 27]. In early reports, it was assumed that the violation of Kohler&apos;s rule reflected the influence of <lb/>superconducting fluctuations [20, 25]. However, the observed temperature or field dependences could not be <lb/>reproduced in any fluctuation theory. <lb/>At a later stage, the effect has been explained as a normal state effect, caused by the presence of two separate <lb/>relaxation times in the normal state [19]. T his idea relies on the resonating valence bond model, in which <lb/>spin and charge are separated and are described by spinon and holon quasiparticles [28]. Within this concept <lb/>the relaxation times for carrier motion normal to the Fermi surface and parallel to it are different. T he <lb/>former, τ tr , is the usual transport relaxation time. It is related to the spinon-holon scattering, which leads to a <lb/>linear T dependence of the resistivity, i.e., τ tr <lb/>-1 ∝ T. T he latter, τ H , is the transverse (Hall) relaxation time. It <lb/>is the result of the spinon-spinon scattering that varies as T 2 like any other fermion-fermion interaction. <lb/>According to the theory, the magnetoresistivity should be proportional to the square of the Hall angle θ H . <lb/>However, the experiments of Balakirev [26] and Abe [27], respectively on La 2-x Sr x CuO 4 thin films and <lb/>La 1.905 Ba 0.095 CuO 4 single crystals, convincingly proved that the latter dependence does not hold. Both authors <lb/>considered the failure of Kohler&apos;s rule as an anomalous aspect of the normal-state transport in high-T c systems. <lb/>According to our data, a large positive magnetoresistivity that tends to saturate at high magnetic fields, <lb/>appears at low temperatures when crossing the Sr content through the insulator-superconductor transition. <lb/>T his tendency can be attributed to superconducting fluctuations. As follows from the data on La 1.73 Sr 0.27 CuO 4 , <lb/>the normal-state magnetoresistivity starts to play a considerable role at higher Sr contents. <lb/>One dimensional character of the superconducting fluctuations in underdoped La 2-x Sr x CuO 4 <lb/>We believe that an adequate theory to describe the magnetoresistivity in high-T c systems should <lb/>account for the inhomogeneous distribution of spin-rich and charge-rich areas, i.e. with the presence of <lb/>stripes (1-D charge areas) [30]. It is apparent from our data that the violation of Kohler&apos;s rule and the <lb/>superconducting fluctuations are much more pronounced in pseudogapped systems, where stripes are <lb/>formed. Although stripes are well established experimentally, they are, up to now, strongly neglected by <lb/>theoreticians working on magnetoresistivity effects. <lb/>Figure 9 shows the magnetoconductivity ∆σ = (σ(0 T ) -σ(50 T )) as a function of the temperature, both <lb/>presented in logarithmic coordinates for La 1.94 Sr 0.06 CuO 4 (T c = 2.4 K) and La 1.9 Sr 0.1 CuO 4 (T c = 17.5 K). Data <lb/>were taken in the temperature range 7.7 K -125 K and 20 K -132 K respectively. We observe a lowering of <lb/>
20
+ the magnetoresistivity with temperature following a similar power-law for both samples. Surprisingly, the <lb/>value of the power 1.54 is close to 3/2, which corresponds, according to the Aslamazov-Larkin (AL) [43] <lb/>expression (eq. 1) to superconducting fluctuations with a one-dimensional character! <lb/>3 <lb/>1 <lb/>2 <lb/>2 <lb/>1 <lb/>3~  <lb/> <lb/> <lb/>  <lb/> <lb/> <lb/>− <lb/>∆ <lb/>  <lb/> <lb/> <lb/>  <lb/> <lb/> <lb/>− <lb/>∆ <lb/>  <lb/> <lb/> <lb/>  <lb/> <lb/> <lb/>− <lb/>∆ <lb/>c <lb/>c <lb/>D <lb/>c <lb/>c <lb/>D <lb/>c <lb/>c <lb/>D <lb/>T <lb/>T <lb/>T <lb/>T <lb/>T <lb/>T <lb/>T <lb/>T <lb/>T <lb/>σ <lb/>σ <lb/>σ <lb/>(1) <lb/>T he nice correspondence with the 1D AL expression is an important result since both samples have a very <lb/>different critical temperature. T he dimensionality derived from the paraconductivity is moreover in good <lb/>agreement with stripe models. T o obtain the result of Figure 9 for La 1.94 Sr 0.06 CuO 4 and La 1.9 Sr 0.1 CuO 4 , we <lb/>neglected the influence of the normal state on the magnetoresistivity. Most probably, the normal-state contribution <lb/>is as small as the magnetoresistivity of the heavily underdoped La 1.955 Sr 0.045 CuO 4 compound: of the order of 1 %, <lb/>which justifies our procedure. <lb/>Since the contrast of the stripes with respect to their surrounding decreases upon doping, the dimensionality of the <lb/>charge-transport in the overdoped samples is not well defined. Although La 1.8 Sr 0.2 CuO 4 and La 1.75 Sr 0.25 CuO 4 reveal <lb/>distinct pseudogap features, they partially recover a 2D (or 3D) character. It is therefore not surprising that the <lb/>magnetoconductivity of the overdoped samples could not be fitted with a simple power law with respect to the <lb/>temperature. Secondly, the data for the overdoped samples are noisier because of their low resistivity, which <lb/>complicates an analysis like the one presented in Figure 9. Moreover, as follows from the data on <lb/>La 1.73 Sr 0.27 CuO 4, the normal-state contribution to the magnetoresistivity cannot be neglected in samples with a high <lb/>doping level. At the moment, there is however no theory available, allowing an accurate evaluation of this normal-<lb/>state background. It is even not &apos;a priori&apos; clear whether the violation of Kohler&apos;s rule should be attributed to <lb/>fluctuations or to the normal state (or to both). For example, if the mobile stripes bend in a magnetic field, they <lb/>may influence the normal state magnetoresistivity in an unconventional way. In this context, we would like to <lb/>mention the result of Ando and coworkers [29], who have reported a possible influence on the striped structure <lb/>in non-superconducting underdoped cuprates by a magnetic field, in the configuration where the magnetic field <lb/>is applied parallel to the ab-plane. Kimura et al. [20] found a strong suppression of the magnetoresistivity in <lb/>La 2-x Sr x CuO 4 around the hole concentration with x = 1/8, a concentration that is related to a more static nature <lb/>of the stripes [30]. T heir results are at least a manifestation of the importance of the striped structure in the <lb/>analysis of magnetoresistivity measurements. <lb/>Role of the pseudogap and pre-pair electronic states <lb/>T he pseudogap is emerging as an important indicator revealing the nature of the superconductivity as well as <lb/>the normal state in our high-T c samples. A possible scenario relates the pseudogap with the presence of <lb/>electronic pair states far above T c [30 -36]. T he idea is that Cooper pairs are formed at a temperature T * far <lb/>above T c, but bulk phase coherent superconductivity is only established when long-range phase coherence is <lb/>obtained below T c . T he models, which are based on this precursor superconductivity scenario, get growing <lb/>experimental support. Scanning tunneling spectroscopy measurements clearly demonstrate that the <lb/>pseudogap evolves into the superconducting gap at low temperatures [37]. Moreover, ARPES data indicate <lb/>that the pseudo-and the superconducting gap both have d-wave symmetry [4]. Our experimental observation <lb/>of a close relation between the pseudogap and the superconducting fluctuations (= precursor pairs) strongly <lb/>favor these models as well. Altshuler et al. [38] questioned the interpretation of the pseudogap as the <lb/>superconducting gap because a large fluctuation diamagnetism has not been observed between T c and T * . <lb/>
21
+ Emery et al. [34] stated however that the absence of dramatic diamagnetic effects is expected if the <lb/>superconducting fluctuations are one-dimensional, and if the Josephson coupling between stripes is small. In <lb/>this case, an applied magnetic field does not cause any significant orbital motion until full phase coherence <lb/>develops, close to T c . T o our knowledge, we are not aware of other publications, which unveiled the one-<lb/>dimensional character of the superconducting fluctuations experimentally (Figure 9). We found the one-<lb/>dimensional nature of the transport of precursor pairs thanks to the investigation of the magnetoresistivity <lb/>very close to the insulator-superconductor transition. <lb/>T he magnetoresistivity data for La 1.9 Sr 0.1 CuO 4 , presented in Figure 5, do show neither clearly <lb/>marked second critical fields H c2 (T) nor saturation at high fields. Fluctuating Cooper pairs seem to exist up <lb/>to very high fields, most probably above the field range accessible by our pulsed field setup. Following the <lb/>ideas outlined in [34, 37], T * is the mean-field critical temperature of the superconductor rather than T c . <lb/>When T * is used to obtain the paramagnetic limiting field for sample La 1.9 Sr 0.1 CuO 4 (T * ≈ 400 K, <lb/>T c = 17.5 K) instead of T c , a value of µ o H p ≈ 700 T is obtained, illustrating that a field of 50 T is indeed not <lb/>high enough to destroy completely the preformed pairs. T he ARPES research of Loeser et al. on the <lb/>pseudogap in Bi 2 Sr 2 CaCu 2 O 8+δ [4] revealed a binding energy of 75 meV in the precursor pairs. T hus a <lb/>magnetic field of about 130 T (µ o µ B H = k B T) would be needed to destroy them completely. If the idea of <lb/>precursor pairs is correct, the temperature seems to be a much more critical parameter for the existence of the <lb/>pairs than a magnetic field up to 50 T . T he &apos;resistive upper critical field&apos;, as defined by a line construction, is <lb/>certainly a questionable concept with respect to the underdoped high-T c compounds. It is possible that the <lb/>magnetoresistivity data of the samples, which manifest a pseudogap, just reflect the behavior of the precursor <lb/>pairs in a magnetic field, maybe even the localization of the pairs in a magnetic field. <lb/>Superconductivity in metals is the result of two distinct quantum phenomena, pairing and long-rang <lb/>phase coherence. T he influence of the stripes on superconductivity is therefore two-fold. First of all, the <lb/>one-dimensional character of the charge transport favors pair formation as follows from the similarities <lb/>between the pseudogap in high-T c superconductors and the spin-gap in ladder cuprates and from <lb/>experiments that demonstrate a connection between the superconducting-and the pseudogap. On the other <lb/>hand, the low dimensionality hinders the long-range phase coherence needed to establish bulk <lb/>superconductivity. It is a well-known fact that long-range phase coherence is impossible in a purely one-<lb/>dimensional system. T his is in agreement with the fact that (La,Sr,Ca) 14 Cu 24 O 41 , the only known <lb/>superconducting ladder compound, becomes superconducting under high pressure when the interactions <lb/>between the ladders are enhanced. <lb/>Like already stated before, the broadening of the superconducting transitions in underdoped cuprates, both in <lb/>field and temperature, is most probably due to sample inhomogeneities. However, the inhomogeneities do not <lb/>reflect a bad sample quality but rather an intrinsic property, related to a low charge carrier concentration and the <lb/>presence of stripes. <lb/>T he superconducting transitions for overdoped and underdoped La 2-x Sr x CuO 4 epitaxial thin films, showing <lb/>up in our transport measurements in pulsed magnetic fields up to 50 T , have a completely different nature. <lb/>While the transition in underdoped samples is smeared out over more than 40 T , the overdoped samples <lb/>
22
+ reveal well-defined second critical fields H c2 , where bulk superconductivity is suppressed. In these <lb/>overdoped samples, we found an upward curvature of H c2 (T) at low temperatures, in strong contrast with the <lb/>WHH model [39, 40], which predicts a saturation of the second critical field in this temperature range. A <lb/>similar anomalous behavior of H c2 with respect to the temperature has been reported in the literature for <lb/>several high-T c systems [41, 42, 29]. <lb/>Conclusion: Gerneric phase diagram including fluctuation area <lb/>T he pulsed field transport measurements at temperatures T &gt; T c , revealed a sudden appearance of a <lb/>large positive in-plane magneto-resistivity in La 2-x Sr x CuO 4 close to the insulator-superconductor transition at <lb/>x = 0.055. T his evolution suggests that the effect can be attributed to superconducting fluctuations. T he <lb/>fluctuations appear at temperatures, which exceed T c by several times. It is therefore reasonable to speak <lb/>about precursory pairing far above T c . By presenting the magnetoresistivity data in the form of classical <lb/>Kohler-plots, we found that the superconducting fluctuations are very pronounced in underdoped samples. <lb/>At the same time, the normal-state contribution to the magnetoresistivity dominates in overdoped samples. <lb/>T he region, where we observed superconducting fluctuations, is schematically shown in the phase diagram <lb/>of Figure 10 by the shaded area. In order to evaluate this region exactly, an adequate theory is needed, which <lb/>allows to separate fluctuations and normal-state contributions to conductivity. Unfortunately, such theory is <lb/>lacking at the moment. It is however clear from our data that there is a close link between the presence of <lb/>strong superconducting fluctuations and the pseudogap phase. Since fluctuations are expected to become <lb/>more pronounced in systems with a reduced dimensionality, this observation is in excellent agreement with <lb/>the idea that the pseudogap phase is characterized by 1D charge transport [23]. T he excess conductivity of <lb/>the La 1.94 Sr 0.06 CuO 4 (T c = 2.4 K) and La 1.9 Sr 0.1 CuO 4 (T c = 17.5 K) sample at 50 T indeed shows a simple <lb/>power-law behavior with respect to the temperature. T he experimentally found power 1.54 is close to 3/2, <lb/>which is characteristic for one-dimensional fluctuations, according to the basic theory of paraconductivity, <lb/>proposed by Aslamasov and Larkin [43]. Hence, our findings strongly favor stripe models [30-36] and are <lb/>consistent with the idea of a precursory behavior towards superconductivity far above T c . <lb/></body>
23
+
24
+ <div type="acknowledgement">Acknowledgements <lb/>T he Belgian IUAP, the Flemish GOA and FWO have supported this work. J.V. is a postdoctoral fellow of <lb/>the CNRS-France and the FWO -Vlaanderen.<lb/></div>
25
+
26
+ <listBibl>Re fe re nce s <lb/>[1] T . T imusk, B. Statt, Rep. Prog. Phys. 62, 61 (1999). <lb/>[2] V.J. Emery, S.A. Kivelson, Nature 374, 434 (1995). <lb/>[3] M. Kugler, Ø. Fischer, Ch. Renner, S. Ono, Y. Ando, Phys. Rev. Lett. 86, 4911 (2001). <lb/>[4] Loeser96A.G. Loeser, Z.-X. Shen, D.S. Dessau, D.S. Marshall, C.H. Park, P. Fournier, A. Kapitulnik, <lb/>Science 273, 325 (1996). <lb/>[5] H. Ding, T . Yokaya, J.C. Campuzano, T . T akahashi, M. Randeria, M.R. Norman, T . Mochiku, <lb/>K. Kadowaki, J. Giapinzakis, Nature 382, 51 (1996). <lb/>[6] P.J. White, Z.-X. Shen, C. Kim, J.M. Harris, A.G. Loeser, P. Fournier, A. Kapitulnik, Phys. Rev. B 54, <lb/>R15669 (1996). <lb/>[7] Ando95 Y. Ando, G.S. Boebinger, A. Passner, T . Kimura and K. Kishio, Phys. Rev. Lett. 75, 4662 <lb/>(1995). <lb/>[8] J.M. Harris, Y.F. Yan, P. Matl, N.P. Ong, P.W. Anderson, T . Kimura, K. Kitzawa, Phys. Rev. Lett. 75, <lb/>1391 (1995). <lb/>[9] A. Lacerda, J.P. Rodriguez, M.F. Hundley, Z. Fisk, P.C. Canfield, J.D. T hompson, S.W. Cheong, Phys. <lb/>Rev. B 49 13, 9097 (1994). <lb/>[10] T . Kimura, S. Miyasaka, H. T akagi, K. T amasaku, H. Eisaki, S. Uchida, K. Kitazawa, M. Hiroi, M. <lb/>Sera, N. Kobayashi, Phys. Rev. B 53, 8733 (1996). <lb/>[11] N.W. Preyer, M.A. Kastner, C.Y. Chen, R.J. Birgeneau, Y. Hidaka, Phys. Rev. B 44, 407 (1991). <lb/>[12] B. Wuyts, Z.X. Gao, S. Libbrecht, M. Maenhoudt, E. Osquiguil, Y. Bruynseraede, Physica C 203, 235 <lb/>(1992). <lb/>[13] P. Wagner, K.-Q. Ruan, I. Gordon, J. Vanacken, V.V. Moshchalkov, Y. Bruynseraede, Physica C 356, <lb/>107 (2001). <lb/>[14] F. Herlach, L. Van Bockstal, M. van de Burgt, G. Heremans, Physica B 155, 61 (1989). <lb/>[15] F. Herlach, Ch. Agosta, R. Bogaerts, W. Boon, I. Deckers, A. De Keyser, N. Harrison, A. Lagutin, L. <lb/>Li, L. T rappeniers, J. Vanacken, L. Van Bockstal, A. Van Esch, Physica B 216, 161 (1996). <lb/>[16] G.S. Boebinger, Y. Ando, A. Passner, T . Kimura, M. Okuya, J. Shimoyama, K. Kishio, K. T amasaku, <lb/>N. Ichikawa, S. Uchida, Phys. Rev. Lett. 77, 5417 (1996). <lb/>[17] P. Fournier, P. Mohanty, E. Maiser, S. Darzens, T . Venkateson, C.J. Lobb, G. Czjzek, R.A. Webb, <lb/>R.L. Greene, Phys. Rev. Lett. 81, 4720 (1998). <lb/>[18] S. Ono, Y. Ando, T . Murayama, F.F. Balakirev, J.B. Betts, G.S. Boebinger, Phys. Rev. Lett. 85, <lb/>(2000). <lb/>[19] J.M. Harris, Y.F. Yan, P. Matl, N.P. Ong, P.W. Anderson, T . Kimura, K. Kitazawa, Phys. Rev. Lett. 75, <lb/>1391 (1995). <lb/>[20] T . Kimura, S. Miyasaka, H. T akagi, K. T amasaku, H. Eisaki, S. Uchida, K. Kitazawa, M. Hiroi, <lb/>M. Sera, N. Kobayashi, Phys. Rev. B 53, 8733 (1996). <lb/>[21] A. Malinovski, M.Z. Cieplak, A.S. van Steenbergen, J.A. Perenboom, K. Karpinska, M. Nberkowski, S. <lb/>Guha, P. Lindenfeld, Phys. Rev. Lett. 79, 495 (1997). <lb/>[22] M. Kohler, Ann. Phys. 32, 211 (1938). <lb/>[23] J. Vanacken, L. T rappeniers, P. Wagner, L. Weckhuysen, V.V. Moshchalkov, Y. Bruynseraede, Phys. <lb/>Rev. B 64, 184425 (2001). <lb/>[24] A. Lacerda, J.P. Rodriguez, M.F. Hundley, Z. Fisk, P.C. Canfield, J.D. T hompson, S.W. Cheong, Phys. <lb/>Rev. B 49, 9097 (1994). <lb/>[25] K. Semba, A. Matsuda, Phys. Rev. B 55, (1997) and reference therein. <lb/>[26] F.F. Balakirev, I.E. T rofimov, S. Guha, M.Z. Cieplak, P. Lindenfeld, Phys. Rev. B 57, R8083 (1998). <lb/>[27] Y. Abe, Y. Ando, J. T akeya, H. T anabe, T . Watauchi, I. T anaka, H. Kojima, Phys. Rev. B 59, 14753 <lb/>(1999). <lb/>[28] P.W. Anderson, Science 235, 1196 (1987). <lb/>[29] Y. Ando, G.S. Boebinger, A. Passner, L.F. Schneemeyer, T . Kimura, M. Okuya, S. Watauchi, <lb/>J. Shimoyama, K. Kishio, K. T amasaku, N. Ichikama, S. Uchida, cond-mat/9908190 (1999). <lb/>[30] J. M. T ranquada, B.J. Sternlieb, J.D. Axe, Y. Nakamura, S. Uchida, Nature 375, 561 (1995) ; <lb/>J.M. T ranquada, J.D. Axe, N. Ichikawa, Y. Nakamura, S. Uchida, B. Nachumi, Phys. Rev. B 54, 7489 <lb/>(1996) ; J.M. T ranquada, Physica C 282-287, 166 (1997); J.M. T ranquada, Phys. Rev. Lett. 78, (1997) ; <lb/>J.M. T ranquada, Physica B 241-243, 745 (1998). <lb/>[31] M. Randeria, cond-mat/9710223 (1997). <lb/>[32] A. Bianconi, A. Valletta, A. Perali, N.L. Saini, Physica C 296, 269 (1998). <lb/>[33] J. Maly, J. Boldizar, K. Levin, cond-mat/9805018 (1998). <lb/>[34] V.J. Emery, S.A. Kivelson, Nature 374, 434 (1995); V.J. Emery, S.A. Kivelson, cond-mat/9902179 <lb/>(1999); V.J. Emery, S.A. Kivelson, J.M. T ranquada, cond-mat/9907228 (1999). <lb/>[35] J. Zaanen, Science 286, 251(1999). <lb/>[36] V.V. Moshchalkov, Sol. St. Comm. 86, 715 (1993); V.V. Moshchalkov, L. T rappeniers, J. Vanacken, <lb/>Europhys. Lett. 46, 75 (1999); V.V. Moshchalkov, J. Vanacken, L. T rappeniers, Phys. Rev. B. 64, 214504 <lb/>(2001). <lb/>[37] M. Kugler, Ø. Fischer, Ch. Renner, S. Ono, Y. Ando, Phys. Rev. Lett. 87, 4911 (2001). <lb/>[38] B.L. Altshuler, L.B. Ioffe, A.J. Millis, Phys. Rev. B 53, 415 (1996). <lb/>[39] N.R. Werthammer, E. Helfand, P.C. Hohenberg, Phys. Rev. 147, 295 (1966). <lb/>[40] E. Helfand, N.R. Werthamer, Phys. Rev. 147, 288 (1966). <lb/>[41] A.P. Mackenzie, S.R. Julian, A. Carrington, G.G. Lonzarich, D.J.C. Walker, J.R. Cooper, D.C. Sinclair, <lb/>Physica C 235-240, 233-236 (1994). <lb/>[42] M.S. Osofsky, R.J. Soulen (jr.), S.A. Wolf, J.M. Broto, H. Rakoto, J.C. Ousset, G. Coffe, S. Askenazy, <lb/>P. Pari, I. Bozovic, J.N. Eckstein, G.F. Virshup, Phys. Rev. Lett. 71, 2315 (1993). <lb/>[43] L.G. Aslamazov, A.I. Larkin, Sov. Solid State 10, 875 (1968). <lb/></listBibl>
27
+
28
+ <body>Figure captions <lb/>Figure 1. T he left side of the upper frame gives an overview of the field dependence of the in-plane <lb/>resistivity ρ ab (µ 0 H) of La 1.955 Sr 0.045 CuO 4 at different temperatures. T he right side of the upper frame depicts <lb/>the temperature dependence of the in-plane resistivity ρ ab (T ) at zero magnetic field (solid line). The open <lb/>circles mark the positions where the magnetoresistivity has been measured. For the positions labeled A, B, C <lb/>and D, the magnetoresistivity is shown in more detail in the middle part of the picture. T he <lb/>magnetoresistivity at 45 T (MR) is indicated in %. T he lower frame shows the Kohler plot at temperatures <lb/>T =18K, T =87K and T =176K for the 47 T pulsed field. <lb/>Figure 2. T he left side of the upper frame gives an overview of the field dependence of the in-plane <lb/>resistivity ρ ab (µ 0 H) of La 1.950 Sr 0.050 CuO 4 at different temperatures. T he right side of the upper frame depicts <lb/>the temperature dependence of the in-plane resistivity ρ ab (T ) at zero magnetic field (solid line). The open <lb/>circles denote the values of the resistivity in zero field, derived from pulsed field measurements; filled circles <lb/>mark the resistivity values at 45 T . For the positions labeled A, B, C, D, E and F, the magnetoresistivity is <lb/>shown in more detail in the middle part of the picture. T he magnetoresistivity at 45 T (MR) is indicated in <lb/>%. T he lower frame shows the Kohler plots for the La 1.950 Sr 0.050 CuO 4 sample at selected temperatures <lb/>14K &lt; T &lt; 72K for the 45 T pulsed field data <lb/>Figure 3. T he magnetoresistivity data of La 1.945 Sr 0.055 CuO 4 are presented in the same way as in Figure 1. <lb/>T he filled circles at the right side of the upper frame mark the resistivity values at 45 T . For the positions <lb/>labeled A, B, C, D, E and F, the magneto-resistivity is shown in more detail in the middle part of the picture. <lb/>T he magnetoresistivity at 45 T (MR) is indicated in %. T he lower frame shows the Kohler plots for the <lb/>La 1.945 Sr 0.055 CuO 4 sample at selected temperatures (4.2K &lt; T &lt; 97K) for the 47 T pulsed field data. <lb/>Figure 4. T he left side of the upper frame gives an overview of the field dependence of the in-plane <lb/>resistivity ρ ab (µ 0 H) of La 1.94 Sr 0.06 CuO 4 at different temperatures. T he right side of the upper frame depicts the <lb/>temperature dependence of the in-plane resistivity ρ ab (T ) at zero magnetic field (solid line). The open circles <lb/>denote the values of the resistivity in zero field, derived from pulsed field measurements; filled circles mark <lb/>the resistivity values at 45 T . For the positions labeled A, B, C, D, E and F, the magnetoresistivity is shown <lb/>in more detail in the lower part of the picture. T he magnetoresistivity at 45 T (MR) is indicated in %. The <lb/>lower frame shows the Kohler plots for the La 1.94 Sr 0.06 CuO 4 sample at selected temperatures used for the 47 <lb/>T pulsed field data. <lb/>Figure 5. T he magnetoresistivity data of La 1.9 Sr 0.1 CuO 4 are presented in the same way as in Figure 1. The <lb/>filled circles at the right side of the upper frame mark the resistivity values at 50 T . For the positions <lb/>labelled A, B, C, D, E and F, the magnetoresistivity is shown in more detail in the middle part of the picture. <lb/>T he magnetoresistivity at 50 T (MR) is indicated in %. T he lower frame shows the Kohler plots for the <lb/>La 1.9 Sr 0.1 CuO 4 sample at selected temperatures used for the 50 T pulsed field data. <lb/>
29
+ Figure 6. T he left side of the upper frame gives an overview of the field dependence of the in-plane <lb/>resistivity ρ ab (µ 0 H) of La 1.8 Sr 0.2 CuO 4 at different temperatures. T he right side of the upper frame depicts the <lb/>temperature dependence of the in-plane resistivity ρ ab (T ) at zero magnetic field (solid line). The open circles <lb/>denote the values of the resistivity at zero field, derived from pulsed field measurements; filled circles mark <lb/>the resistivity values at 45 T . For the positions labeled A, B, C, D, E and F, the magnetoresistivity is shown <lb/>in more detail in the middle part of the picture. T he magnetoresistivity at 45 T (MR) is indicated in % at <lb/>T &gt; T c . T he lower frame shows the Kohler plots for the La 1.8 Sr 0.2 CuO 4 sample at selected temperatures used <lb/>for the 48 T pulsed field data. <lb/>Figure 7. T he left side of the upper frame gives an overview of the field dependence of the in-plane <lb/>resistivity ρ ab (µ 0 H) of La 1.75 Sr 0.25 CuO 4 at different temperatures. T he right side of the upper frame depicts <lb/>the temperature dependence of the in-plane resistivity ρ ab (T ) at zero magnetic field (solid line). The open <lb/>circles denote the values of the resistivity at zero field, derived from pulsed field measurements; filled circles <lb/>mark the resistivity values at 45 T . For the positions labeled A, B, C, D, E and F, the magnetoresistivity is <lb/>shown in more detail in the lower part of the picture. T he magnetoresistivity at 50 T (MR) is indicated in % <lb/>at T &gt; T c . T he lower frame shows the Kohler plots for the La 1.75 Sr 0.25 CuO 4 sample at selected temperatures <lb/>used for the 50 T pulsed field data. <lb/>Figure 8. T he left side of the upper frame gives an overview of the field dependence of the in-plane <lb/>resistivity ρ ab (µ 0 H) of La 1.73 Sr 0.27 CuO 4 at different temperatures. T he right hand side of the upper frame <lb/>depicts the temperature dependence of the in-plane resistivity ρ ab (T ) at zero magnetic field (solid line). The <lb/>open circles denote the values of the resistivity at zero field, derived from pulsed field measurements; filled <lb/>circles mark the resistivity values at 45 T . For the positions labeled A, B, C and D, the magnetoresistivity is <lb/>shown in more detail in the middle part of the picture. T he lower frame shows the Kohler plots for the <lb/>La 1.73 Sr 0.27 CuO 4 sample at selected temperatures used for the 45 T pulsed field data (left side) and the 12 T <lb/>pulsed field data (right side). <lb/>Figure 9. T he logarithm of the magnetoconductivity (σ(0 T ) -σ(50 T )) as a function of the logarithm of the <lb/>temperature, which is rescaled with respect to T c . T he constant C = 1µΩcm accounts for the units along the <lb/>y-axis. T he linear fits follow the equations y = -5.5 + 1.55 x and y = -7.7 + 1.54 x for La 1.94 Sr 0.06 CuO 4 and <lb/>La 1.9 Sr 0.1 CuO 4 respectively. <lb/>Figure 10. (T ,x)-phase diagram of the La 2-x Sr x CuO 4 samples. T he antiferromagnetic (AF) and <lb/>superconducting (SC) regions are indicated. T he crossover temperature T * separates region I from region II, <lb/>where La 2-x Sr x CuO 4 has a pseudogap. T MI marks the metal to insulator transition and defines region III. The <lb/>dimensions of the electronic transport in region I and II, are labelled in the Figure. A shaded area is added, <lb/>which schematically shows the region, where we observed superconducting fluctuations.<lb/>
30
+
31
+ 0 <lb/>10 30 40 <lb/>µ o H(T) <lb/>ρ <lb/>ab (mΩ cm) <lb/>14 K <lb/>16 K <lb/>18K <lb/>20 K <lb/>25 K <lb/>32 K <lb/>176 K <lb/>56 K <lb/>87 K <lb/>0 <lb/>100 200 300 <lb/>T(K) <lb/>ρ <lb/>ab (mΩ cm) <lb/>1.0 <lb/>1.5 <lb/>2.0 <lb/>2.5 <lb/>1.0 <lb/>1.5 <lb/>2.0 <lb/>2.5 <lb/>A <lb/>B <lb/>C <lb/>D <lb/>0 <lb/>5 0 <lb/>2.46 <lb/>2.51 <lb/>0 <lb/>5 0 <lb/>2.13 <lb/>2.20 <lb/>0 <lb/>5 0 <lb/>1.67 <lb/>1.72 <lb/>0 <lb/>5 0 <lb/>1.68 <lb/>1.69 <lb/>16 K <lb/>20 K <lb/>32 K <lb/>176 K <lb/>M R: 0.8 % <lb/>M R: 1.1 % <lb/>M R: 0.9 % <lb/>M R: 0.3 % <lb/>A <lb/>B <lb/>C <lb/>D <lb/>ρ <lb/>ab (mΩ cm) <lb/>µ o H(T) <lb/>0 <lb/>5 <lb/>10 <lb/>-0.02 <lb/>-0.01 <lb/>0.00 <lb/>0.01 <lb/>0.02 <lb/>87K <lb/>176K <lb/>18K <lb/>(µ o H/ρ o ) 2 [10 -4 (T/µΩ cm) 2 ] <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>La 1.955 Sr 0.045 CuO 4 <lb/>Fig.01 <lb/>(µ o H/ρ o ) 2 [10 -4 (T/µΩ cm) 2 ] <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>0 <lb/>5 <lb/>15 <lb/>-0.025 <lb/>0.000 <lb/>0.025 <lb/>0.050 <lb/>0 <lb/>5 <lb/>1 0 <lb/>0.00 <lb/>0.05 <lb/>0.10 <lb/>14 K 16 K <lb/>18 K <lb/>20 K <lb/>25 K <lb/>30 K <lb/>40 K 53 K <lb/>72 K <lb/>0 <lb/>10 20 30 40 50 <lb/>16 K <lb/>20 K <lb/>74 K <lb/>10 K <lb/>12 K <lb/>K <lb/>25 K <lb/>30 K <lb/>116K <lb/>0 <lb/>100 200 300 <lb/>1.0 <lb/>1.5 <lb/>2.0 <lb/>2.5 <lb/>3.0 <lb/>1.0 <lb/>1.5 <lb/>2.0 <lb/>2.5 <lb/>3.0 <lb/>A <lb/>B <lb/>C <lb/>D <lb/>E <lb/>F <lb/>µ o H(T) <lb/>ρ <lb/>ab (mΩ cm) <lb/>T(K) <lb/>ρ <lb/>ab (mΩ cm) <lb/>ρ <lb/>ab (mΩ cm) <lb/>µ o H(T) <lb/>0 <lb/>5 0 <lb/>2.4 <lb/>2.7 A <lb/>M R: 10% <lb/>10 K <lb/>0 <lb/>5 0 <lb/>1.75 <lb/>1.85 B <lb/>18 K <lb/>M R: 6.3 % <lb/>0 <lb/>5 0 <lb/>1.65 <lb/>1.80 C <lb/>20 K <lb/>M R: 5.7 % <lb/>5 0 <lb/>1.5 <lb/>1.6 D <lb/>25 K <lb/>M R: 4.2 % <lb/>0 <lb/>5 0 <lb/>1.16 <lb/>1.20 E <lb/>M R: 0.8 % <lb/>5 0 <lb/>1.19 <lb/>1.23 F <lb/>116 K <lb/>M R: 0.2 % <lb/>K <lb/>La 1.950 Sr 0.050 CuO 4 <lb/>Fig.02 <lb/>ρ <lb/>ab (mΩ cm) <lb/>ρ <lb/>ab (mΩ cm) <lb/>ρ <lb/>ab (mΩ cm) <lb/>(µ o H/ρ o ) 2 [10 -4 (T/µΩ cm) 2 ] <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>0 <lb/>10 20 30 40 <lb/>1 <lb/>2 <lb/>3 <lb/>0 <lb/>100 200 <lb/>1 <lb/>2 <lb/>3 <lb/>4.2 K <lb/>5.6 K <lb/>9.4 K <lb/>14 K <lb/>200 K <lb/>20.7 K <lb/>97 K <lb/>47.7 K <lb/>A <lb/>B <lb/>C <lb/>D <lb/>E <lb/>F <lb/>µ o H(T) <lb/>T(K) <lb/>0 <lb/>5 0 <lb/>2.20 <lb/>3.20 <lb/>A <lb/>4.2 K <lb/>M R: 33 % <lb/>0 <lb/>5 0 <lb/>1.60 <lb/>1.90 B <lb/>9.4 K <lb/>M R: 18 % <lb/>0 <lb/>5 0 <lb/>1.30 <lb/>1.45 C <lb/>14 K <lb/>M R: 11.4 % <lb/>0 <lb/>5 0 <lb/>1.05 <lb/>1.10 <lb/>D <lb/>20.7 K <lb/>M R: 6.3 % <lb/>0 <lb/>5 0 <lb/>0.84 <lb/>0.88 E <lb/>0 <lb/>0 <lb/>0.84 <lb/>0.90 <lb/>F <lb/>47.7 K <lb/>95.5 K <lb/>M R: 1.3 % <lb/>M R: 0.8 % <lb/>0.00 <lb/>0.05 <lb/>0 <lb/>10 <lb/>20 <lb/>0 <lb/>5 <lb/>10 <lb/>15 <lb/>0.0 <lb/>0.1 <lb/>0.2 <lb/>0.3 <lb/>4.2 K <lb/>5.6 K <lb/>9.4 K <lb/>14 K <lb/>20.7 K <lb/>29.4 K <lb/>47.7 K <lb/>97 K <lb/>µ o H(T) <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>La 1.945 Sr 0.055 CuO 4 <lb/>Fig.03 <lb/>ρ <lb/>ab (mΩ cm) <lb/>ρ <lb/>ab (mΩ cm) <lb/>ρ <lb/>ab (µΩ cm) <lb/>(µ o H/ρ o ) 2 [10 -4 (T/µΩ cm) 2 ] <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>µ o H(T) <lb/>T(K) <lb/>µ o H(T) <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>0 <lb/>10 20 30 40 50 <lb/>0.0 <lb/>0.5 <lb/>1.0 <lb/>1.5 <lb/>4.2 K <lb/>287 K <lb/>7.6 K <lb/>14.6 K <lb/>17.25 K <lb/>22.6 K <lb/>124.9 K <lb/>63.2 K <lb/>100 200 300 <lb/>0.0 <lb/>0.5 <lb/>1.0 <lb/>1.5 <lb/>A <lb/>B <lb/>C <lb/>D E <lb/>F <lb/>0 <lb/>5 0 <lb/>0 <lb/>1500 <lb/>4.2 K <lb/>M R: 330 % <lb/>A <lb/>0 <lb/>5 0 <lb/>700 <lb/>850 <lb/>17.25 K <lb/>M R: 17 % <lb/>B <lb/>0 <lb/>5 0 <lb/>660 <lb/>720 <lb/>C <lb/>26.5 K <lb/>M R: 8.3 % <lb/>0 <lb/>5 0 <lb/>620 <lb/>650 D <lb/>33.3 K <lb/>M R: 5 % <lb/>5 0 <lb/>570 <lb/>580 <lb/>63.2 K <lb/>M R: 1.5 % <lb/>E <lb/>0 <lb/>5 0 <lb/>1465 <lb/>1485 <lb/>F <lb/>287 K <lb/>M R: 0.1 % <lb/>0.00 <lb/>0.01 <lb/>0.02 <lb/>0 <lb/>20 <lb/>40 <lb/>63 K <lb/>125 K <lb/>K <lb/>0 <lb/>20 <lb/>40 <lb/>60 <lb/>0.0 <lb/>0.2 <lb/>14.6 K <lb/>17.3 K <lb/>22.6 K <lb/>0 <lb/>200 <lb/>400 <lb/>0 <lb/>5 <lb/>4.2 K <lb/>7.7 K <lb/>26.5 K <lb/>33.3 K <lb/>Fig.04 <lb/> La 1.940 Sr 0.060 CuO <lb/> ρ <lb/>
32
+ ab (mΩ cm) <lb/>ρ <lb/>ab (mΩ cm) <lb/>ρ <lb/>ab (µΩ cm) <lb/>(µ o H/ρ o ) 2 [10 -4 (T/µΩ cm) 2 ] <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>µ o H(T) <lb/>T(K) <lb/>µ o H(T) <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>0 <lb/>20 30 50 <lb/>0.0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>4.2 K <lb/>6 K <lb/>10 K <lb/>14 K <lb/>18 K <lb/>25 K <lb/>30 K <lb/>71 K <lb/>132.7 K <lb/>0 <lb/>100 200 300 <lb/>0.2 <lb/>0.0 <lb/>0.4 <lb/>0.6 <lb/>A-B C <lb/>D <lb/>E F <lb/>0 <lb/>5 0 <lb/>0 <lb/>600 A <lb/>4.2 K <lb/>0 <lb/>5 0 <lb/>400 <lb/>12 K <lb/>B <lb/>0 <lb/>5 0 <lb/>0 <lb/>400 <lb/>18 K <lb/>C <lb/>0 <lb/>5 0 <lb/>250 <lb/>350 <lb/>5 0 <lb/>300 <lb/>320 <lb/>0 <lb/>5 0 <lb/>400 <lb/>420 <lb/>D <lb/>30 K <lb/>M R: 29 % <lb/>E 71 K <lb/>M R: 2.4 % <lb/>132.7 K <lb/>M R: 0.8 % <lb/>F <lb/>0.00 <lb/>0.05 <lb/>0.10 <lb/>0.0 <lb/>0.5 <lb/>1.0 <lb/>1.5 <lb/>22 K <lb/>25 K <lb/>30 K <lb/>0.00 <lb/>0.01 <lb/>0.02 <lb/>0.03 <lb/>0.00 <lb/>0.05 <lb/>71 K <lb/>La 1.900 Sr 0.100 CuO 4 <lb/>Fig.05 <lb/>ρ <lb/>ab (µΩ cm) <lb/>ρ <lb/>ab (µΩ cm) <lb/>ρ <lb/>ab (µΩ cm) <lb/>(µ o H/ρ o ) 2 [10 -4 (T/µΩ cm) 2 ] <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>µ o H(T) <lb/>T(K) <lb/>µ o H(T) <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>0 <lb/>100 <lb/>200 <lb/>0 <lb/>10 20 30 50 <lb/>134 K <lb/>79.2 K <lb/>4.2 K <lb/>7.0 K <lb/>10.6 K <lb/>14.6 K <lb/>22.0 K <lb/>30.6 K <lb/>0 50 100 150 <lb/>0 <lb/>100 <lb/>200 <lb/>A B C <lb/>D <lb/>E <lb/>F <lb/>0 <lb/>5 0 <lb/>0 <lb/>200 <lb/>A <lb/>4.2 K <lb/>0 <lb/>5 0 <lb/>0 <lb/>200 B <lb/>10.6 K <lb/>0 <lb/>5 0 <lb/>0 <lb/>150 <lb/>C <lb/>19.6 K <lb/>0 <lb/>5 <lb/>115 <lb/>130 D <lb/>30.6 K <lb/>M R: 12 % <lb/>0 <lb/>5 0 <lb/>154 <lb/>164 E <lb/>79.2 K <lb/>M R: 2 % <lb/>0 <lb/>5 0 <lb/>220 <lb/>230 F 134 K <lb/>M R: 0.5 % <lb/>0.00 0.05 0.10 0.15 <lb/>0.0 <lb/>0.1 <lb/>0.2 <lb/>0.00 <lb/>0.05 <lb/>0.10 <lb/>-0.02 <lb/>0.00 <lb/>0.02 <lb/>0.04 <lb/>26.6 K <lb/>30.6 K <lb/>34.7 K <lb/>52 K <lb/>79.2 K <lb/>101 K <lb/>116.6 K <lb/>134 K <lb/>La 1.800 Sr 0.200 CuO 4 <lb/>Fig.06 <lb/>ρ <lb/>ab (µΩ cm) <lb/>ρ <lb/>ab (µΩ cm) <lb/>ρ <lb/>ab (µΩ cm) <lb/>(µ o H/ρ o ) 2 [10 -4 (T/µΩ cm) 2 ] <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>µ o H(T) <lb/>T(K) <lb/>µ o H(T) <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>0 <lb/>50 <lb/>100 <lb/>10 20 30 40 50 <lb/>4.2 K, 6 K, 10 K, 14 <lb/>K, <lb/>18 K, 25 K, 30 K <lb/>K <lb/>156 K <lb/>0 50 100 150 <lb/>0 <lb/>50 <lb/>100 <lb/>A-C <lb/>D <lb/>E <lb/>F <lb/>0 <lb/>5 0 <lb/>0 <lb/>50 <lb/>4.2 K <lb/>A <lb/>0 <lb/>5 0 <lb/>0 <lb/>50 <lb/>8 K <lb/>B <lb/>0 <lb/>5 0 <lb/>0 <lb/>16 K <lb/>C <lb/>0 <lb/>5 <lb/>45 <lb/>55 D 30 K <lb/>M R: 10 % <lb/>0 <lb/>5 0 <lb/>50 <lb/>60 <lb/>E 43 K <lb/>M R: 8 % <lb/>0 <lb/>5 0 <lb/>110 <lb/>135 <lb/>F 156 K <lb/>M R: 0.7 % <lb/>0.0 <lb/>0.5 <lb/>1.0 <lb/>1.5 <lb/>0.0 <lb/>0.1 <lb/>0.2 <lb/>0.0 <lb/>0.1 <lb/>0.2 <lb/>0.0 <lb/>0.5 <lb/>1.0 <lb/>25 K <lb/>30 K <lb/>156 K <lb/>43 K <lb/>La 1.750 Sr 0.250 CuO 4 <lb/>Fig.07 <lb/>ρ <lb/>ab (µΩ cm) <lb/>ρ <lb/>ab (µΩ cm) <lb/>ρ <lb/>ab (µΩ cm) <lb/>(µ o H/ρ o ) 2 [10 -4 (T/µΩ cm) 2 ] <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>µ o H(T) <lb/>T(K) <lb/>µ o H(T) <lb/>(ρ-ρ <lb/>o )/ρ <lb/>o <lb/>0 <lb/>10 20 30 40 50 <lb/>K <lb/>54 K <lb/>4.2 K, 8 K, 12 K, <lb/>16 K, 22 K, 30 K <lb/>0 <lb/>50 <lb/>100 <lb/>0 <lb/>40 <lb/>60 <lb/>80 <lb/>0 <lb/>20 <lb/>40 <lb/>60 <lb/>80 <lb/>A B <lb/>C <lb/>D <lb/>0 <lb/>5 0 <lb/>0 <lb/>50 <lb/>8 K <lb/>A <lb/>0 <lb/>5 0 <lb/>0 <lb/>50 B <lb/>12 K <lb/>0 <lb/>3 0 <lb/>32 <lb/>40 <lb/>0 <lb/>3 0 <lb/>50 <lb/>55 <lb/>20 K <lb/>54 K <lb/>C <lb/>D <lb/>0.0 <lb/>0.5 <lb/>-0.2 <lb/>0.0 <lb/>0.2 <lb/>0.00 <lb/>0.05 <lb/>0.10 <lb/>-0.05 <lb/>0.00 <lb/>0.05 <lb/>99 K 54 K <lb/>30 K <lb/>99 K 54 K <lb/>30 K <lb/>Fig.08 <lb/>La 1.730 Sr 0.270 CuO <lb/>-4 <lb/>-2 <lb/>0 <lb/>2 <lb/>-12 <lb/>-10 <lb/>-8 <lb/>-6 <lb/>-4 <lb/>La 1.94 Sr 0.06 CuO 4 <lb/>La 1.9 Sr 0.1 CuO 4 <lb/>1.54 <lb/>ln(T c /(T-T c )) <lb/>ln(C(σ <lb/>0T -σ <lb/>50T )) <lb/>Fig.09 <lb/>antiferromagnetic insulator <lb/>0.0 <lb/>0.1 <lb/>0.2 <lb/>0.3 <lb/>0 <lb/>100 <lb/>200 <lb/>300 <lb/>I <lb/>II <lb/>SC <lb/>T * <lb/>T MI <lb/>T c <lb/>Sr-content x <lb/>T(K) <lb/>T N <lb/>metallic <lb/>2D <lb/>metallic <lb/>1D striped <lb/>Fig.10 </body>
33
+
34
+
35
+ </text>
36
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EPL0330153-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <front>arXiv:cond-mat/9504062v1 14 Apr 1995 <lb/>Linear Temperature Variation of the Penetration Depth in <lb/>YBa 2 Cu 3 O 7−δ Thin Films <lb/>L.A. de Vaulchier, J.P. Vieren, Y. Guldner, N. Bontemps <lb/>Laboratoire de Physique de la Matière Condensée, Ecole Normale Supérieure, 24 rue Lhomond, <lb/>Paris Cedex 05, France <lb/>R. Combescot <lb/>Laboratoire de Physique Statistique, Ecole Normale Supérieure, 24 rue Lhomond, 75231 Paris <lb/>Cedex 05, France <lb/>Y. Lemaître and J.C. Mage <lb/>Laboratoire Central de Recherches, Thomson-CSF,Domaine de Corbeville, 91404 Orsay Cedex, <lb/>France <lb/>(Received May 10, 2019) <lb/>Abstract <lb/>We have measured the penetration depth λ(T ) on YBa 2 Cu 3 O 7 thin films <lb/>from transmission at 120, 330 and 510 GHz, between 5 and 50 K. Our data <lb/>yield simultaneously the absolute value and the temperature dependence of <lb/>λ(T ). In high quality films λ(T ) exhibits the same linear temperature depen-<lb/>dence as single crystals, showing its intrinsic nature, and λ(0) = 1750Å. In a <lb/>lower quality one, the more usual T 2 dependence is found, and λ(0) = 3600Å. <lb/>This suggests that the T 2 variation is of extrinsic origin. Our results put the <lb/>d-wave like interpretation in a much better position. <lb/>1 <lb/>PACS numbers: 74.25.Gz, 74.25.Nf, 74.72.Bk, 74.76.Bz</front> <lb/>Typeset using REVT E X <lb/>
9
+
10
+ <body>The puzzling properties of high-T c superconductors have stimulated recently extensive <lb/>experimental and theoretical work in order to determine the actual symmetry of the order <lb/>parameter. While in the BCS weak coupling theory, the penetration depth λ(T ) varies as <lb/>exp(−∆/k B T ), a quadratic behavior λ(T ) ∼ T 2 was reported by many groups in YBCO <lb/>thin films [1-3]. Recently, in very high quality YBCO single crystals, a linear temperature <lb/>dependence was measured up to 40 K [4-6]. Such a variation, never observed so far in <lb/>YBCO thin films [7], is consistent with the occurrence of nodes in the gap and may suggest <lb/>a d-wave pairing mechanism [8]. However other possible interpretations have been raised, <lb/>both theoretically and experimentally, e.g. a possible proximity effect with a normal metal <lb/>layer [6,9], or the sensitivity of a conventional BCS type λ(T ) to the oxygenation of the <lb/>samples [10,11]. Several reasons for the discrepancy observed between thin films and single <lb/>crystals may be invoked: (i) within the framework of d-wave pairing, scattering due to im-<lb/>purities or defects may change the linear temperature dependence into a quadratic one in <lb/>thin films [12], where such defects would be more numerous. However T c should then also <lb/>be affected, which is not the case [4,13]. (ii) weak links, more likely to be present in thin <lb/>films, may yield an effective penetration depth (larger than the intrinsic one) with a different <lb/>temperature dependence [14,15]. The penetration depth λ has been essentially measured <lb/>in both single crystals and thin films by using surface impedance techniques at a single <lb/>frequency [2,4]. Such measurements actually measure the variation ∆λ(T ) = λ(T ) − λ(0), <lb/>but without the knowledge of λ(0) on each sample, comparing the temperature dependence <lb/>of various samples may not be significant. This paper reports for the first time, from high <lb/>frequency transmission data, both an unambiguous linear temperature dependence of λ(T ) <lb/>on high quality YBCO thin films, which is quantitatively the same as in single crystals [4], <lb/>and simultaneously λ(0) ∼ 1750Å. We observe a quadratic temperature dependence on a <lb/>lower quality sample, and λ(0) ∼ 3600Å. We suggest that this T 2 dependence in thin films <lb/>is not a consequence of strong scattering (in contrast with Zn or Ni doped materials [16]) <lb/>but is of extrinsic origin. The intrinsic dependence of λ(T ) at low temperature is linear. <lb/>This linear dependence points plainly toward the existence of low energy excitations, but our <lb/></body>
11
+
12
+ <page>3 <lb/></page>
13
+
14
+ <body>results alone do not allow to identify them. A natural and popular explanation is that they <lb/>are due to nodes in the gap arising from unconventional pairing of d-wave type [1,4,5], an <lb/>interpretation also conveyed by SQUID experiments [17,18]. Our results provide support to <lb/>d-wave pairing, because they show a linear behavior in films, but also because they eliminate <lb/>the uncomfortable need to explain the T 2 dependence by impurity scattering at the unitary <lb/>limit [12]. <lb/>The experimental transmission set-up uses carcinotron tubes as powerful, stable mi-<lb/>crowave sources and oversized waveguides in order to change easily the frequency (120 to <lb/>GHz) [19,20]. We took an extreme care to lower microwave leakages down to 60 dB. The <lb/>transmitted signal is detected by a helium cooled InSb bolometer. Measurements have been <lb/>performed by slowly varying the temperature between 5 and 110 K at fixed frequencies. The <lb/>transmission of the substrate was checked to vary by less than 1 % between 5 and 110 K. <lb/>We have screened the samples from stray magnetic field, so that the residual field is less <lb/>than 0.5 Gauss. For a film of thickness d (smaller than the skin depth in the normal state <lb/>or the penetration depth in the superconducting state) deposited onto a substrate of index <lb/>n, the transmission writes [21]: <lb/>T = <lb/>1 <lb/>1 + σdZ 0 <lb/>1+n <lb/>2 <lb/>(1) <lb/>
15
+ where σ = σ 1 − iσ 2 is the complex conductivity of the film, Z 0 the impedance of free space <lb/>and Z = 1/σd the impedance of the film. If in the energy and temperature range of interest <lb/>σ 2 (T ) ≫ σ 1 (T ) and σ 2 dZ 0 <lb/>1+n ≫ 1, we can write: <lb/>T <lb/>T 110 <lb/>= µ 2 <lb/>0 ω 2 λ 4 (T ) σ 2 <lb/>110 <lb/>(2) <lb/>where σ 110 is the normal 110 K conductivity of the film. We choose to normalize the trans-<lb/>mission at 110 K because the ratio thus obtained is experimentally more reliable than the <lb/>absolute value of the transmission. For simplicity we assumed σ 110 dZ 0 <lb/>1+n <lb/>≫ 1 in (2). Thus, <lb/>the measurement of T/T 110 yields an absolute value of λ(T ). The uncertainties on the λ(T ) <lb/></body>
16
+
17
+ <page>4 <lb/></page>
18
+
19
+ <body>value which arise when neglecting the interferences within the film and/or the substrate, and <lb/>the effect of the finite film thickness with respect to λ(T ) have been estimated: the trans-<lb/>mission is fairly insensitive to the interference effects at our frequencies as can be shown <lb/>by comparing a complete expression for T to (1). Moreover, no significant change of the <lb/>transmitted energy could be observed in the most sensitive range 440-550 GHz [20]. The <lb/>finite film thickness, yields an approximate 80Å overestimate of λ(T ) for d/λ(T ) ∼ 0.5. <lb/>σ 2 (ω, T ) ≫ σ 1 (ω, T ) holds at low temperature (T ≤ 20 K) for all frequencies, but should <lb/>break down at 40 K [5,22,23]. Indeed, σ 1 (35 GHz, 40 K) ∼ 2 × 10 7 Ω −1 m −1 , a value likely <lb/>larger than σ 2 (300 GHz). However we expect a strong decrease of σ 1 (T ) at 300 GHz [22-24]. <lb/>It is essential that we define the criteria we use in order to sort out the films that we <lb/>investigate. Our films are epitaxially grown either by laser ablation on MgO [25] or by <lb/>sputtering on LaAlO 3 [26]. They have a narrow transition (∆T c ≤ 1 K). T c is 86-92 K, <lb/>depending on the substrate. In previous papers, the correlation between the film quality in <lb/>terms of transmission, width of the rocking curve and surface resistance has been established <lb/>[19,20,27]. This led us to select the films either from their surface resistance: R S ≤ 0.5 mΩ <lb/>at 77 K and 10 GHz, or the width of their rocking curve ∆θ ≤ 0.5 • . The characteristics of <lb/>our samples are listed in Table I. Two types of films have been intentionally investigated. <lb/>The first film (A1) is of poor quality with respect to the above criteria. The two others <lb/>(B1 and B2) display either a low surface resistance or a narrow rocking curve. We show in <lb/>Fig. 1 and Fig. 2 the normalized transmission of the A1 and B2 films respectively. Strong <lb/>differences are observed, in particular for the residual transmission in the superconducting <lb/>state. At T ≤ 0.5 T c , the transmission increases as ω 2 in sample B2, as expected from (2) <lb/>and as shown in the inset of Fig. 2. In sample A1, the data can be analyzed by adding a <lb/>constant T to the ω 2 term [28]. We estimate T 0 from the 10 K data (see inset of Fig. 1). We <lb/>compute λ(T ) from (2) for the three frequencies, assuming that T 0 in the case of A1, does <lb/>not depend on temperature or frequency. The result of this analysis is shown on Fig. 3 and <lb/>Fig. 4, for the samples A1 and B2. The 120, 330 and 510 GHz curves collapse, up to 40 K <lb/>
20
+
21
+ for sample A1 and up to 55 K for sample B2. This confirms that the frequency dependent <lb/>part of the transmission varies as ω 2 , and that σ 2 (ω, T ) ≫ σ (ω, T ) in these temperature <lb/>ranges. λ(0) is found to be 1750 ± 160Å for B2 and 3600 ± 200Å for A1. The uncertainty <lb/>on λ(0) depends mostly on the accuracy on the σ 110 measurement, which is limited by the <lb/>uncertainty on the film thickness (±100Å) and by the Van der Pauw technique. Finally <lb/>the most striking result is the temperature dependence of λ(T ). For sample A1, we find <lb/>as shown in the inset, a clear T 2 dependence, a fairly common result for thin films [1-3], <lb/>associated with a very large value for λ(0). In contrast, for sample B2, we find a linear <lb/>temperature dependence up to 50 K along with a much shorter λ(0). Such a linear behavior <lb/>was similarly observed in film B1. <lb/>We now discuss the implications of these results. We believe that they demonstrate <lb/>clearly the extrinsic origin of the T 2 dependence found previously in thin films. Indeed they <lb/>agree with a phenomenological expression λ 2 (T ) = λ 2 <lb/>intr (T )+λ 2 <lb/>extr (T ) as proposed by Hylton <lb/>et al., with λ extr being for example the contribution of weak links (but our conclusions are <lb/>obviously independent of this specific expression). When λ extr is the dominating contribu-<lb/>tion as in our film A1 it may produce the T 2 dependence which is likely to change from <lb/>sample to sample. Indeed, the data of Porch et al. (see inset of Fig. 3) are not the same as <lb/>ours. On the other hand, for good enough films, λ extr gets negligible and we find the intrinsic <lb/>behavior for λ(T ) which is linear, and is expected to be unchanged from films to crystals. <lb/>The YBCO single crystals results of Hardy et al. are reported on Fig. 4, shifted to our λ(0) <lb/>value. The excellent agreement between the two sets of data confirms the intrinsic nature <lb/>of λ in film B2. We remark that the agreement between λ(T ) in two completely unrelated <lb/>samples is a very strong result since, except for a coincidence, it eliminates any extrinsic <lb/>explanation like poor surface effects or pair breaking due to magnetic impurities [10], for the <lb/>interpretations of the linear dependence. However the nature of the low energy excitations <lb/>is still unknown [29,30] and it is also not yet clear whether they are linked to the planes or to <lb/>the chains. Furthermore, the T 2 dependence appears in a film where the penetration depth <lb/></body>
22
+
23
+ <page>6 <lb/></page>
24
+
25
+ <body>is large. Our λ intr (0) is much shorter. It is larger though than λ ab (0) = 1450-1490Å derive <lb/>by µSR [31], and consistent with λ a (0) = 1600Å provided by infrared reflectance data [32]. <lb/>We believe that most techniques do not provide as a straightforward determination as ours. <lb/>Our results offer a solution to a previous uncomfortable situation arising in the d-wave <lb/>interpretation of the experimental results for λ(T ). Indeed in order to explain the T 2 behav-<lb/>ior in films together with the linear dependence in single crystals, theoretical calculations <lb/>had to call for a high impurity concentration in films. But this implies a significant differ-<lb/>ence between T c for films and crystals which is not observed. The proposed escape [33] was <lb/>to assume that the impurity concentration was rather low with a scattering very near the <lb/>unitary limit, a very specific hypothesis. In this case T c could be little changed by impurities <lb/>while the low T behavior of λ(T ) would be much more affected. Since our results show that <lb/>the T 2 behavior is extrinsic, e.g. due to weak links [20], all the results become coherent with <lb/>a reasonably low content of ordinary scatterers. <lb/>In order to interpret quantitatively our results, it is convenient to make use of the sim-<lb/>plified two-dimensional model of d-wave like pairing introduced by Xu et al. [12], where the <lb/>gap has a linear dependence near the nodes and is constant elsewhere: |∆(θ)| = µ∆ 0 θ for <lb/>0 ≤ θ ≤ 1/µ and |∆(θ)| = ∆ 0 for 1/µ ≤ θ ≤ π/4, θ being the angle with respect to the <lb/>node position on the Fermi surface and the rest of ∆(θ) being obtained by symmetry. The <lb/>density of states is taken as constant. This model contains the essential physical ingredients <lb/>to represent a general order parameter with d-wave symmetry, except for the possible effect <lb/>of Van Hove singularities. The slope of λ(T ) at T = 0 is given by 2 ln 2/µ∆ 0 , however when <lb/>∆ 0 is calculated with a weak coupling assumption, the result can only be made to agree with <lb/>experiment up to 20 K and it does not reproduce the remarkable linear behavior up to 55 K. <lb/>A simple way to account for this feature is to incorporate strong coupling effects. This is <lb/>justified on theoretical ground [8] and is also consistent with experiments since for example <lb/></body>
26
+
27
+ <page>7 <lb/></page>
28
+
29
+ <body>tunneling data give an effective gap larger than what is expected from BCS theory. If we <lb/>take 2∆ 0 /T c = 7 (a typical tunneling value), we obtain the solid curve on Fig. 4 which agrees <lb/>quite well with our experimental results. This gives further support to this interpretation. <lb/>In summary, although our results cannot be taken as a proof for d-wave pairing, they put <lb/>at least this interpretation in a much better position. <lb/>We are deeply indebted to M. Guilloux-Viry, C. Le Paven-Thivet and A. Perrin from <lb/>the Laboratoire de Chimie Inorganique et Moléculaire, Université de Rennes, for providing <lb/>us with well characterized thin films. We thank P. Monod for his critical reading of the <lb/>manuscript. This work has been supported by EEC Contract 93-2027.IL and by the New <lb/>Energy and Industrial Technology Development Organization (NEDO). Both laboratories <lb/>Physique de la Matière Condensée and Physique Statistique are associated to CNRS and <lb/>Universities Paris VI and Paris VII. <lb/></body>
30
+
31
+ <page>8 <lb/></page>
32
+
33
+ <listBibl>REFERENCES <lb/>[1] J. Annett, N. Goldenfeld and S.R. Renn, Phys. Rev. B 43, 2778 (1991). <lb/>[2] A. Porch et al., to appear in the Proceedings of the Applied Superconductivity Confer-<lb/>ence (1994), to be published in the IEEE Trans. on Applied Supercond. (1995). <lb/>[3] Z. Ma et al., Phys. Rev. Lett. 71, 781, (1993). <lb/>[4] W.N. Hardy et al., Phys. Rev. Lett. 70, 3999 (1993). <lb/>[5] Kuan Zhang et al., Phys. Rev. Lett. 73, 2484 (1994). <lb/>[6] J. Mao et al., submitted to Phys. Rev. B , (1994). <lb/>[7] M.R. Beasley, same issue as [2], reports a linear temperature dependence on one TlBa-<lb/>CaCuO thin film. <lb/>[8] P. Monthoux, A.V. Balatsky and D. Pines, Phys. Rev. Lett. 67, 3448 (1991). <lb/>[9] M.S. Pambianchi, J. Mao and S.M. Anlage, Phys. Rev. B 50, 13659 (1994). <lb/>[10] V.S. Kresin, to be published in J. Supercond. Proceedings of the Workshop on High-T c <lb/>Superconductors: Physical Properties and Mechanisms (5-11 Jan. 1995, Miami, USA). <lb/>[11] Y.Y. Xue et al., same issue as [10]. <lb/>[12] D. Xu, K. Yip and J.A. Sauls, to be published in Phys. Rev. B , (1994). <lb/>[13] M. Prohammer and J.P. Carbotte, Phys. Rev. B 43, 5370 (1991). <lb/>[14] T.L. Hylton et al., Appl. Phys. Lett. 53, 1343 (1988). <lb/>[15] J. Halbritter, J. Appl. Phys. 71, 339 (1992). <lb/>[16] D.A. Bonn et al., Phys. Rev. B 50, 4051 (1994). <lb/>[17] D.A. Wollman et al., Phys. Rev. Lett. bf 71, 2134 (1993). <lb/>[18] C.C. Tsuei et al., Phys. Rev. Lett. bf 73, 593 (1994). <lb/>9 <lb/>[19] L.A. de Vaulchier et al., Physica C 235-240, 1083 (1994). <lb/>[20] L.A. de Vaulchier et al., to be published in Phys. Rev. B , (1995). <lb/>[21] R.E. Glover and M. Tinkham, Phys. Rev. 108, 243 (1957). <lb/>[22] D.A. Bonn et al., Phys. Rev. B 47, (1993). <lb/>[23] R. Buhleier et al., Phys. Rev. B 50, 9672 (1994). <lb/>[24] U. Dähne et al., Physica C 235-240, 2066 (1994). <lb/>[25] M. Guilloux-Viry et al., Mat. Sc. Eng. B 18, 115 (1993). <lb/>[26] Y. Lemaitre et al., Physica C 235-240, 643 (1994). <lb/>[27] C. Thivet et al., submitted to Physica C, (1994). <lb/>[28] One possible origin for the extra transmission T could be holes in the film: indeed <lb/>observations with the optical microscope show in film A1 few holes of 100 µm typical <lb/>size which may give rise to excess transmission. <lb/>[29] A.M. Neminsky and P.N. Nikolaev, Physica C 212, 389 (1993). <lb/>[30] G.M. Eliashberg, G.V. Klimovitch and Rylyakov, J. Supercond. 4, 393 (1991). <lb/>[31] J.E. Sonier et al., Phys. Rev. Lett. 72, 744 (1994). <lb/>[32] D.N. Basov et al., Phys. Rev. Lett. 74, 598 (1995). <lb/>[33] P.J. Hirschfeld and N. Goldenfeld, Phys. Rev. B 48, 4219 (1993). <lb/></listBibl>
34
+
35
+ <body>FIGURES <lb/>FIG. 1. 110 K normalized transmission versus temperature for the YBCO A1 film. The inset <lb/>shows the ω 2 law of the normalized transmission at 10 K. <lb/>FIG. 2. 110 K normalized transmission versus temperature for the YBCO B2 film. The inset <lb/>shows the ω 2 law of the normalized transmission at 10 K. <lb/>FIG. 3. Temperature dependence of λ(T ) for the YBCO A1 film. The T 2 law (see inset) is <lb/>characteristic of a poor quality film. The bold squares represent the 8 GHz Porch et al. [2] data <lb/>points shifted to our λ(0) value. <lb/>FIG. 4. Temperature dependence of λ(T ) for the YBCO B2 film. The bold squares represent <lb/>the Hardy et al. [4] data points on a single crystal (shifted to our λ(0) value). The solid line is a <lb/>comparison with a d-wave like, strong coupling calculation, with 2∆ 0 = 7 k B T c (see text). <lb/>11 <lb/>TABLES <lb/>TABLE I. Characteristics of the YBa 2 Cu 3 O 7 thin films. d is their thickness and ∆θ the rocking <lb/>curve width. R S (77 K, 10 GHz) has been corrected for the film thickness. <lb/>Sample <lb/>d [Å] <lb/>T c [K] <lb/>∆θ [ • ] <lb/>R S [mΩ] <lb/>σ 110 [µΩcm] <lb/>A1 <lb/>1200 <lb/>86 <lb/>0.8 <lb/>∼ 10 <lb/>130 <lb/>B1 <lb/>1400 <lb/>88 <lb/>0.42 <lb/>0.45 <lb/>B2 <lb/>1000 <lb/>92 <lb/>0.27 <lb/>80 <lb/> 0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>20 <lb/>40 <lb/>60 <lb/>80 <lb/>100 <lb/>120 GHz <lb/>GHz <lb/>GHz <lb/>T <lb/>/ <lb/>T <lb/>6 [ K ] <lb/>0 <lb/>0.05 <lb/>0.1 <lb/>0.15 <lb/>0.2 <lb/>4 <lb/>8 <lb/>12 <lb/>M [ 10 &quot; rad s` ] <lb/>6 =10 K <lb/>T <lb/>/ <lb/>T <lb/>0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>20 <lb/>40 <lb/>60 <lb/>80 <lb/>100 <lb/>GHz <lb/>330 GHz <lb/>510 GHz <lb/>6 [ K ] <lb/>T <lb/>/ <lb/>T <lb/>0 <lb/>0.01 <lb/>0.02 <lb/>0.03 <lb/>0 <lb/>8 <lb/>12 <lb/>6 =10 K <lb/>T <lb/>/ <lb/>T <lb/>M [ 10 &quot; rad s` ] <lb/>3500 <lb/>4000 <lb/>4500 <lb/>5000 <lb/>5500 <lb/>6000 <lb/>25 <lb/>45 <lb/>65 <lb/>85 <lb/>120 GHz <lb/>330 GHz <lb/>510 GHz <lb/>Porch et al. <lb/>l <lb/>[ <lb/>Å <lb/>] <lb/>6 [ K ] <lb/>3600 <lb/>3700 <lb/>3800 <lb/>3900 <lb/>0 <lb/>400 <lb/>800 1200 1600 <lb/>l <lb/>[ <lb/>Å <lb/>] <lb/>6 [ K ] <lb/>% <lb/># <lb/>&apos; <lb/>!! <lb/>!% <lb/># <lb/># <lb/>&quot;# <lb/>$# <lb/>&amp;# <lb/>120 GHz <lb/>330 GHz <lb/>510 GHz <lb/>Hardy et al. <lb/>l <lb/>[ <lb/>Å <lb/>] <lb/>6 [ K ] <lb/>1750 <lb/>1800 <lb/>1850 <lb/>1900 <lb/>0 <lb/>20 <lb/>30 <lb/>40 <lb/>l <lb/>[ <lb/>Å <lb/>] <lb/>6 [ K ] </body>
36
+
37
+ </text>
38
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EPL0410207-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <front>arXiv:cond-mat/0103183v1 [cond-mat.supr-con] 8 Mar 2001 <lb/>Upper critical field in Ba 1−x K x BiO 3 : magnetotransport versus magnetotunneling <lb/>P. Samuely, 1 P. Szabó, 1,2 T. Klein, 3 A.G.M.Jansen, J. Marcus, 3 C.Escribe-Filippini, 3 and P. Wyder <lb/>1 Institute of Experimental Physics, Slovak Academy of Sciences, SK-04353 Košice, Slovakia. <lb/>2 Grenoble High Magnetic Field Laboratory, Max-Planck-Institut für Festkörperforschung and Centre National de la Recherche <lb/>Scientifique, B.P. 166, F-38042 Grenoble Cedex 9, France. <lb/>3 Laboratoire d&apos;Etudes des Propriétés Electroniques des Solides, Centre National de la Recherche Scientifique, B.P. 166, <lb/>F-38042 Grenoble Cedex 9, France. <lb/>(November 5, 2018) <lb/>Elastic tunneling is used as a powerful direct tool to determine the upper critical field Hc2(T ) in <lb/>the high-Tc oxide Ba1−xKxBiO3. The temperature dependence of Hc2 inferred from the tunneling <lb/>follows the Werthamer-Helfand-Hohenberg prediction for type-II superconductors. A comparison <lb/>will be made with resistively determined critical field data. <lb/>PACS numbers: 74.60.Ec, 74.25.Dw, 74.50.+r. <lb/></front>
9
+
10
+ <body>The upper critical field H c2 of high-T c superconduc-<lb/>tors remains a contradictory issue. In classical type-II <lb/>superconductors this quantity has been unequivocally de-<lb/>termined from the magnetotransport measurement and <lb/>its temperature dependence can, in most cases, be well <lb/>described by the Werthamer-Helfand-Hohenberg (WHH) <lb/>theory [1]. However, for the high-T c superconductors, <lb/>H c2 extracted from magnetotransport data reveals an un-<lb/>usual increase with decreasing temperatures without any <lb/>saturation down to low temperatures. This upward cur-<lb/>vature is observed in a pronounced way in the supercon-<lb/>ducting cuprates Sm 2−x Ce x CuO 4 [2], Tl 2 Ba 2 CuO 6+δ [3], <lb/>Bi 2 Sr 2 CuO y [4], and YBa (Cu 1−x Zn x )O 7−δ [5]. How-<lb/>ever, the effect is also found in the fully three dimensional <lb/>and non magnetic Ba 1−x K x BiO 3 [6]. Among others, a <lb/>bipolaron scenario [7], an unconventional normal state <lb/>[8], a strong electron-phonon coupling [9], the presence of <lb/>inhomogeneities and magnetic impurities [10] have been <lb/>put forward for an explanation of the anomalous H c2 (T ) <lb/>dependence. Because depinned vortices, either in the liq-<lb/>uid or solid state, cause a finite dissipative resistance be-<lb/>fore reaching the full transition to the normal state, the <lb/>complexity of the H − T phase diagram in high-T c &apos;s [11] <lb/>undermines any direct determination of the upper critical <lb/>field from magnetotransport data. There are indications <lb/>that, also in the fully 3D system of Ba 1−x K x BiO 3 , fluc-<lb/>tuations can lead to a melting of the vortex-glass state <lb/>[12]. This could be a reason complicating a determination <lb/>of H c2 from a dissipative measurement as magnetoresis-<lb/>tance. <lb/>Avoiding the dissipative mechanisms which could ob-<lb/>scure a determination of H c2 from transport measure-<lb/>ments, we show that tunneling measurements can be used <lb/>as an effective and direct method for the determination <lb/>of the upper critical field. The from tunneling obtained <lb/>temperature dependence of H c2 in Ba 1−x K x BiO 3 follows <lb/>the WHH model revealing a saturation at low tempera-<lb/>tures. At the tunneling H c2 values the resistance is very <lb/>close to the full resistive transition into the normal state <lb/>as measured on the same sample. This result was ob-<lb/>tained repeatedly on several samples. <lb/>The single-crystalline Ba 1−x K x BiO 3 samples grown by <lb/>electrochemical crystallisation were dark-blue crystals of <lb/>a cubic shape with a size of about 0.6 mm. The super-<lb/>conducting transition of our crystals was single stepped <lb/>and sharp with T c ≃ 23 K as determined by susceptibil-<lb/>ity measurements. The low temperature resistivity was <lb/>about 100 µΩcm with the metallic temperature depen-<lb/>dence. The tunnel junctions were prepared by painting <lb/>a silver spot of about 0.1 to 0.2 mm diameter on the sur-<lb/>face of the crystal. The interface between the silver and <lb/>Ba 1−x K x BiO 3 counter electrodes served as a natural bar-<lb/>rier forming a planar normal-insulator-superconductor <lb/>(N-I-S) tunnel junction. Low resistance electrical con-<lb/>tacts were prepared for the four-probe measurements of <lb/>the current-voltage (I-V ) and differential conductance <lb/>(dI/dV ) characteristics of the tunnel junction. The tun-<lb/>neling measurements were performed in magnetic fields <lb/>up to 30 T perpendicular to the planar junction enabling <lb/>the formation of the vortex state in the junction area. On <lb/>the same samples the magnetoresistance was measured <lb/>using a four-probe measurement at low frequencies. <lb/>Figure 1 shows the magnetoresistance R(H) of the <lb/>Ba 1−x K x BiO 3 single crystal up to 26 T at temperatures <lb/>from 1.5 K to T c ≃ 23 K. The resistive transitions are <lb/>shifted and broadened towards lower temperatures as the <lb/>magnetic field is increased, although the broadening is <lb/>much smaller than in the case of the cuprates. A sim-<lb/>ple evaluation of the transition field defined as H * (T ) <lb/>where R/R n equals, for instance, 0.1, 0.5, or 0.9 (R n is <lb/>the normal-state resistance) leads to a positive curvature <lb/>of H * (T ) down to the lowest temperatures. If H * (T ) is <lb/>defined for R even closer to R n , the curvature of H * (T ) <lb/>changes at the lowest temperatures which makes that the <lb/>dependence H * (T ) depends on the chosen criterion. <lb/>In the following the normalized tunneling conductances <lb/>of our junctions are presented, where the normal-state <lb/>conduction for the normalization is taken above H c2 (T ) <lb/>for the temperature T under investigation. Figure 2 <lb/>shows a quality certificate of our junction. The spectra <lb/></body>
11
+
12
+ <page>1 <lb/></page>
13
+
14
+ <body>0 <lb/>5 <lb/>10 <lb/>15 <lb/>20 <lb/>25 <lb/>30 <lb/>0.0 <lb/>0.5 <lb/>1.0 <lb/>18 K <lb/>14 K <lb/>10 K <lb/>6 K <lb/>22 K 20 K <lb/>16 K <lb/>12 K <lb/>8 K 4.2K 3K <lb/>1.5 K <lb/>MAGNETIC FIELD (T) <lb/>MAGNETORESISTANCE <lb/>FIG. 1. Magnetoresistance of the Ba1−xKxBiO3 single <lb/>crystal at different temperatures. Arrows -H t <lb/>c2 obtained from <lb/>tunneling. <lb/>can be perfectly described by the Dynes formula [13] <lb/>of the quasi-particle density of states ρ(ǫ) smeared by <lb/>the finite temperature at which the N-I-S junction has <lb/>been measured. ρ(ǫ) = Re[ǫ ′ /(ǫ ′2 − ∆ 2 ) 1/2 ] contains an <lb/>isotropic superconducting energy gap ∆ and a complex <lb/>energy ǫ ′ = ǫ − iΓ which takes account for some addi-<lb/>tional smearing Γ. The Γ smearing of the spectrum is <lb/>case dependent with a tendency Γ/∆ → in the best <lb/>junctions. As mentioned already in the original paper of <lb/>Dynes et al., such an &quot;intrinsic&quot; width of the spectrum <lb/>can be the consequence of anisotropy effects, noise, or <lb/>concentration fluctuations. The presence of microphases <lb/>due to fluctuations in the potassium and oxygen concen-<lb/>tration seems to be a general problem in a substitution <lb/>system like Ba 1−x K x BiO 3 [6]. The Dynes formula fits <lb/>our tunnel spectra at 1.5 K and zero magnetic field with <lb/>∆ = 3.9 ± 0.1 meV and Γ = 0.4 ± 0.1 meV, yielding <lb/>2∆/kT c = 3.9 ± 0.1 and indicating that Ba 1−x K x BiO 3 <lb/>is a BCS-like superconductor with a medium coupling <lb/>strength [14]. In the inset the temperature dependence <lb/>of the superconducting gap is shown for the data from <lb/>Fig.2 and also for data taken at B = 2 T. The broadening <lb/>parameter Γ is found to be independent of temperature. <lb/>At high magnetic fields the normalized tunneling con-<lb/>ductances of the Ba 1−x K x BiO 3 -Ag junction are displayed <lb/>in Fig.3 for different constant temperatures. With in-<lb/>creasing magnetic fields an increasing smearing of the <lb/>superconducting features in the tunneling spectra is ob-<lb/>served. At a certain field strength no structure from su-<lb/>perconductivity can be found anymore and the transi-<lb/>tion from a S-I-N to a N-I-N junction is accomplished. <lb/>Similar tunneling data have been obtained on a thin <lb/>Ba 1−x K x BiO 3 film in a parallel field up to 7 Tesla at <lb/>0.45 K [15]. Unlike a parallel-field configuration, our ex-<lb/>periment on a cubic single crystal involves the <lb/>FIG. 2. Differential conductance of the Ba1−xKxBiO3-Ag <lb/>tunnel junction measured at different temperatures. The inset <lb/>shows the temperature dependence of the superconducting <lb/>energy gap obtained from the tunneling conductances at zero <lb/>magnetic field and 2 T together with the BCS prediction (full <lb/>lines). <lb/>occurrence of a mixed state in a strong magnetic field. <lb/>In the mixed state the tunneling conductance will <lb/>probe an average of the local densities of states. For <lb/>an isolated vortex, the superconducting order parameter <lb/>is zero at the center, increases linearly up to a coherence-<lb/>length distance ξ away from the center where it saturates <lb/>to the zero-field value. The local quasi-particle density of <lb/>states (DOS) equals the normal-state DOS at the vortex <lb/>core, but is broadened near the vortex due to the pair-<lb/>breaking effect of the local magnetic field (as described by <lb/>the Abrikosov-Gor&apos;kov theory developed further by Maki, <lb/>de Gennes and others [16]). In the limit of moderate fields <lb/>(H &lt;&lt; H c2 ), Caroli, de Gennes and Matricon [17] have <lb/>shown that the main contribution to the density of states <lb/>at the Fermi energy comes from the low lying states local-<lb/>ized in a vortex core. Each isolated vortex gives a contri-<lb/>bution equivalent to a normal region of radius ξ yielding <lb/>for the density of states at the Fermi level ρ(0) ∝ ρ n (0)ξ 2 , <lb/>where ρ n (0) is the normal state DOS at the Fermi level. <lb/>Thus, the total averaged density of states at the Fermi <lb/>level is proportional to ρ n (0)ξ 2 per area (H c2 /H)ξ 2 oc-<lb/>cupied by one vortex giving ρ(0) ≃ ρ n (0)H/H c2 . How-<lb/>ever, of more relevance for critical field data, also close to <lb/>H c2 a linear field dependence of ρ(0) has been found by <lb/>solving the linearized Ginzburg-Landau equation in the <lb/>mixed state [18]. A very sensitive method to determine <lb/>the upper critical field from tunneling experiments is to <lb/>display the normalized zero-bias tunneling conductance <lb/>as a function of the field strength [19]. The observation <lb/>of a sharp transition is then taken as a proof of a good <lb/>homogeneity of the sample. <lb/>In Fig. 4 we present the zero-bias tunneling conduc-<lb/></body>
15
+
16
+ <page>2 <lb/></page>
17
+
18
+ <body>FIG. 3. Normalized tunneling conductances in magnetic <lb/>fields from zero up to 30 T in steps of 2 T (if not mentioned <lb/>else) at the indicated temperatures. <lb/>tance as a function of the applied magnetic field for dif-<lb/>ferent temperatures. A linear dependence of dI/dV (0) as <lb/>a function of applied field can be found in a limited field <lb/>range. At the highest fields a &quot;tailing&quot; of the zero-bias <lb/>conductance towards the normal-state value is observed, <lb/>and a finite value of the zero-bias conductance is found <lb/>already at zero magnetic field. The latter effect is ob-<lb/>viously due to the Γ broadening. Also the tailing effect <lb/>at the highest fields could be related to the same cause <lb/>as the Γ broadening, i.e. a certain inhomogeneity in the <lb/>sample. The observed tailing effect resembles the behav-<lb/>ior in the resistive transition close to the transition into <lb/>the normal state (see Fig.1). A linear extrapolation of the <lb/>zero-bias conductance to the field where the normalized <lb/>conductance equals unity, as shown by the full lines in <lb/>Fig.4, has been used to determine the upper critical field. <lb/>Figure 5 shows the obtained temperature dependence of <lb/>the upper critical field H t <lb/>c2 (T ). Also the points H * (T ) <lb/>obtained from 90 % of the magnetoresistance transition <lb/>(R/R n = 0.9) are indicated. The slope of H t <lb/>c2 (T ) near T c <lb/>is different from that of the transition field H * (T ) deter-<lb/>mined from resistance data. We note that the tunneling <lb/>critical fields H t <lb/>c2 (T ) are at fields where the bulk resis-<lb/>tivity is very close to the onset of superconductivity (as <lb/>indicated by the arrows for H t <lb/>c2 in Fig.1). <lb/>As a very significant result, H t <lb/>c2 (T ) shows a clear sat-<lb/>uration at the lowest temperatures as expected for the <lb/>WHH theory. From all dissipative measurements on dif-<lb/>ferent samples of Ba 1−x K x BiO 3 so far done [6,12,15], <lb/>FIG. 4. Zero-bias tunneling conductances as a function of <lb/>magnetic field for all measured temperatures with the linear <lb/>extrapolation to obtain H t <lb/>c2 . <lb/>only a linear increase of H * (T ) for decreasing tem-<lb/>peratures has been obtained. Also recent susceptibility <lb/>measurements reveal this effect in the temperature de-<lb/>pendence of the irreversibility field [20,21] down to the <lb/>lowest temperatures (0.4 K in [20]). <lb/>Besides the above mentioned tailing effect in the zero-<lb/>bias tunneling-conductance and the bulk resistive transi-<lb/>tion near the superconducting transition, the same tailing <lb/>can also be observed near T c in ∆(T ) (see Fig.2) and in <lb/>H t <lb/>c2 (T ) (see below in Fig. 5). We suppose that, despite <lb/>the quality of our sample, stochiometric inhomogeneity <lb/>could play a role in this phenomenon. However, as we will <lb/>discuss below, a more intrinsic cause related to supercon-<lb/>ducting fluctuations could also explain this broadening in <lb/>the superconducting transition. <lb/>Klein et al. [12] suggested that in Ba 1−x K x BiO 3 a <lb/>vortex-glass melting transition driven by fluctuations <lb/>can obscure the magnetotransport determination of H c2 . <lb/>Their measurements of the electric field versus current <lb/>density E − J show that a second order phase transi-<lb/>tion from a vortex glass to a vortex liquid state does <lb/>exist in this system. The presence of the liquid phase <lb/>can induce strong fluctuations below H c2 related to the <lb/>motion of the flux lines, but these fluctuations can be <lb/>quite small above H c2 . This can be the reason why H c2 <lb/>(see arrows in Fig.1) is quite close to the onset of the <lb/>resistive transitions . In this approach the foot of the <lb/>resistive transition is determined by the melting of the <lb/>vortex lattice giving a positive curvature in the temper-<lb/>ature dependence of the line H g (T ) for the liquid-solid <lb/>transition. The foots of the curves in Fig.1 could be in-<lb/>deed well fitted as R ∼ [H/H g (T ) − 1] β corresponding to <lb/>the vortex glass melting theory as introduced by Fisher et <lb/>al. [22], where H g (T ) is the magnetic field of the melting <lb/>transition. The resulting fitting parameter β = 4.1 ± 0.5 <lb/></body>
19
+
20
+ <page>3 <lb/></page>
21
+
22
+ <body>FIG. 5. H-T phase diagram of Ba1−xKxBiO3. Upper <lb/>closed circles -H t <lb/>c2 from tunneling, open triangles -H * from <lb/>magnetoresistance transition at R/Rn = 0.9, and closed <lb/>squares -the melting line Hg. <lb/>is in a perfect agreement with the value obtained on a <lb/>different Ba 1−x K x BiO 3 crystal with T c ≃ 31 K [12]. The <lb/>melting line H g (T ) reveals a positive curvature and can <lb/>be described by the power-law temperature dependence <lb/>H g = H g (0)(1 − T /T c ) 3/2 as shown in Fig. 5. This is <lb/>also in a good agreement with the recent measurement <lb/>of the irreversibility field on a similar sample [20] indi-<lb/>cating that the melting and irreversibility lines coincide <lb/>in Ba 1−x K x BiO 3 . <lb/>In the H − T phase-diagram of Fig.5 the initial slope <lb/>of the upper critical field (−dH c2 /dT ) Tc is about 1.7 ÷ <lb/>1.8 T/K. To emphasize more the fact of the saturation <lb/>of the upper critical field at the lowest temperatures, we <lb/>note the closeness of the zero-bias conductance data for <lb/>1.5, 3 and 4.2 K in comparison with the data taken at <lb/>other temperatures in Fig.4. In Fig.5 we also present <lb/>the WHH upper critical field line with an uncertainty <lb/>comparable to the experimental error bars. Taking into <lb/>account that in a system with important fluctuations the <lb/>H c2 boundary should not be very sharp a satisfactory <lb/>agreement is found. <lb/>We have presented here a direct non-dissipative de-<lb/>termination of the upper critical field in Ba 1−x K x BiO 3 <lb/>using the tunneling effect. H c2 (T ) can be satisfactorily <lb/>described by the Werthamer-Helfand-Hohenberg theory. <lb/>In the Cu-oxides, the existence of a resistive state within <lb/>a large part of the H − T diagram complicates an unam-<lb/>biguous determination of the critical field from transport <lb/>data. Therefore, it would be very interesting (and deci-<lb/>sive for certain proposed superconducting mechanisms) <lb/>to study the upper critical field in the cuprates with the <lb/>non-dissipative tunneling method. <lb/></body>
23
+
24
+ <div type="acknowledgement">ACKNOWLEDGMENTS <lb/>We acknowledge fruitful discussions with S.I. Vedeneev <lb/>and a support of the EU grant No.CIPA-CT93-0183 and <lb/>the Slovak VEGA contract No. 2/1357/94. <lb/></div>
25
+
26
+ <listBibl>[1] N.R. Werthamer, E. Helfand and P.C.Hohenberg, Phys. <lb/>Rev. 147, 295 (1966). <lb/>[2] I.W. Sumarlin et al., Phys. Rev. Lett. 68, 2228 (1992). <lb/>[3] A.P. Mackenzie et al., Phys. Rev. Lett. 71, 1238 (1993). <lb/>[4] M.S. Osofsky et al., Phys. Rev. Lett. 71, 2315 (1993). <lb/>[5] D.J.C. Walker et al., Phys. Rev. B 51, 9375 (1995). <lb/>[6] M. Affronte et al., Phys. Rev. B 49, 3502 (1994). <lb/>[7] A.S. Alexandrov, Phys. Rev. B 48, 10571 (1993). <lb/>[8] R.G. Dias and J.M. Wheatley, Phys. Rev. B 50, R13887 <lb/>(1994). <lb/>[9] F. Marsiglio and J.P. Carbotte, Phys. Rev. B 36, 3633 <lb/>(1987). <lb/>[10] Yu.N. Ovchinnikov and V.Z. Kresin, Phys. Rev. B 54, <lb/>1251 (1996). <lb/>[11] G. Blatter et al., Rev. Mod. Phys. 66, 1125 (1994). <lb/>[12] T. Klein et al., Phys. Rev. B 53, 9337 (1996). <lb/>[13] R.C. Dynes, V. Narayanamurti, and J.P. Garno, Phys. <lb/>Rev. Lett. 41, 1509 (1978). <lb/>[14] Q. Huang et al., Nature 347, 369 (1990). <lb/>[15] G. Roesler et al., IEEE Trans. Appl. Supercond. 3, 1280 <lb/>(1993). <lb/>[16] M. <lb/>Tinkham, <lb/>Introduction <lb/>to Superconductivity (McGraw-Hill, Kogakusha, Tokyo, <lb/>1975). <lb/>[17] C. Caroli, P.G. de Gennes, and J. Matricon, Phys. Lett. <lb/>9, 307 (1964). <lb/>[18] E.Guyon, A. Martinet, J. Matricon, and P. Pincus, Phys. <lb/>Rev. 138, A746 (1965). <lb/>[19] James L. Levine, Phys. Rev. 155, 373 (1967). <lb/>[20] G. Goll, A.G.M. Jansen, and J. Marcus, Czechoslovak J. <lb/>Phys. B 46, S2-849 (1996). <lb/>[21] V.F. Gantmakher et al., Phys. Rev. B 54, 6133 (1996). <lb/>[22] D. Fisher, M.P.A. Fisher, and D.A. Huse, Phys. Rev. B <lb/>43, 130 (1991). <lb/></listBibl>
27
+
28
+ <page>4 </page>
29
+
30
+
31
+ </text>
32
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EPL0480073-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>arXiv:cond-mat/9908383v1 [cond-mat.supr-con] 26 Aug 1999 <lb/>Universal relationship between the penetration depth and the <lb/>normal-state conductivity in YBaCuO. <lb/>A. Pimenov 1 , A. Loidl 1 , B. Schey 1 , B. Stritzker 1 , G. Jakob 2 , H. Adrian 2 , A. V. Pronin 3 , <lb/>and Yu. G. Goncharov <lb/>1 Institut für Physik, Universität Augsburg, 86135 Augsburg, Germany <lb/>Institut für Physik, Universität Mainz, 55099 Mainz, Germany <lb/>3 Institute of General Physics, Russian Acad. Sci., 117942 Moscow, Russia <lb/>Abstract <lb/>The absolute values of the conductivity in the normal state σ n and of <lb/>the low temperature penetration depths λ(0) were measured for a number <lb/>of different samples of the YBaCuO family. We found a striking correlation <lb/>between σ n and λ −2 regardless of doping, oxygen reduction or defects, thus <lb/>providing a simple method to predict the superconducting penetration depth <lb/>and to have an estimate of the sample quality by measuring the normal-state <lb/>conductivity. <lb/></front>
8
+
9
+ <body>The superconducting penetration depth, λ(0), is one of the most important electrody-<lb/>namic properties of the high-temperature superconductors and provides direct information <lb/>about the spectral weight of the superconducting condensate [1]. In the normal state the low <lb/>frequency electrodynamics can be characterized by the absolute value of the conductivity <lb/>σ n (ν, T ). At not too high frequencies, σ n (ν, T ) is in a good approximation a real function <lb/>of frequency. Because λ(0) and σ n (0, T ) both involve a Fermi-surface averaging and depend <lb/>on the same amplitude factors e 2 N(0)υ 2 <lb/>F [2], they may be expected to be connected to each <lb/>other. <lb/>In order to investigate possible correlations experimentally, we have carried out <lb/>submillimeter-wave (100 GHz ¡ ν ¡1100 GHz) transmission experiments on different thin <lb/>YBa 2 Cu 3 O 7−δ films. Optimally doped films on different substrates were obtained by laser <lb/>ablation and by magnetron sputtering. One film was oxygen reduced after the measurement <lb/>and then remeasured under identical conditions. Transmission experiments in the frequency <lb/>range 100-1100 GHz were performed utilizing a set of backward-wave oscillators. The mea-<lb/>surements were performed in a Mach-Zehnder interferometer arrangement [3] which allows <lb/>both, the measurements of transmission and phase shift. The properties of the substrates <lb/>were determined in separate experiments. Utilizing the Fresnel optical formulas for the <lb/>complex transmission coefficient for a double layer system, the complex conductivity was <lb/>determined from the observed spectra without any approximations. It is important to note, <lb/>that the absolute values of the conductivity could be deduced directly from the experiment. <lb/>The penetration depth was calculated from the conductivity data via λ = c/ωk, with k <lb/>being the imaginary part of the complex refractive index k = Im[i(σ 1 + iσ 2 )/ε ω] 1/2 . This <lb/>expression for λ gives the usual microwave result λ = (µ 0 ωσ 2 ) −1/2 in the limit σ 1 ≪ |σ 2 | and <lb/>reduces to the expression for the skin depth in the normal state (σ ≫ |σ 2 |). <lb/>The submillimeter properties of YBa Cu 3 O 7−δ films for a fixed frequency ν = 450 GHz <lb/>(15 cm −1 ) are summarized in the Table 1. In addition, some details of the film preparation, <lb/>the weight of the superconducting condensate (ω <lb/>p,s = c 2 /λ(0) 2 ), and the effective scattering <lb/>rate (Eq. 1) are given. The detailed submillimeter-wave data for the samples No.1,3,5 were <lb/>published previously [5-7]. The data in Tab.1 are represented in the order of decreasing <lb/>absolute values of the conductivities at 100 K. Comparing the conductivity column with the <lb/>penetration depth column, the correlation between these two quantities becomes obvious. <lb/>
10
+ With the exception of the sample No.4, the penetration depth increases with decreasing <lb/>conductivities. We note that the scattering of data for different samples may be connected <lb/>with experimental difficulties in measuring the absolute values of conductivity and penetra-<lb/>tion depth. In addition, in spite of substantial progress in prepairing good quality YBaCuO <lb/>films [4], the details of deposition process still may influence the experimental results. <lb/>To check the possible correlation between σ n and λ in more detail, the normal state con-<lb/>ductivity and the superconducting penetration depth were analyzed on the basis of published <lb/>data available up to now. The results of this analysis are summarized in Fig. 1. In order to <lb/>simplify the analysis only results on YBaCuO based materials were considered. To compare <lb/>the absolute values of the conductivity for different samples the characteristic temperature <lb/>of T =100 K was chosen. It is well known [8] that the resistivity even of oxygen reduced <lb/>or doped cuprates is to a reasonable approximation linear for temperatures between 100 K <lb/>and 300 K. A linear temperature scaling was therefore used, if the results were reported at <lb/>temperatures different from T =100 K. Taking into account the relatively strong scattering <lb/>of the data for different samples, this procedure certainly introduces no substantial errors. <lb/>Fig. 1 represents the data as measured by different experimental methods. In addition to <lb/>the thin film results by THz transmission technique [9-15], mutual inductance data [16,17] <lb/>and single crystal data [18-22] are shown. The remarkable feature of Fig. 1 is that the <lb/>majority of experimental points closely follows the dashed line, which is the best fit to our <lb/>data according to the expression λ(0) −2 /µ 0 = σ n /τ , with 1/τ = 22 THz. Fig. 1 thus suggest <lb/>that the absolute value of the conductivity in the normal state is approximately proportional <lb/>to the spectral weight of the superconducting condensate ω <lb/>p,s or to the inverse square of the <lb/>penetration depth λ(0) −2 . <lb/>In the following we would like to suggest rather simple arguments to understand this rela-<lb/>tion. Within the isotropic approximation the real part of the conductivity can be written as <lb/>σ n = ε 0 ω 2 <lb/>p,n τ , where τ −1 is the quasiparticle scattering rate and ε 0 ω 2 <lb/>p,n is the spectral weight <lb/>of the Drude peak. For a number of high temperature cuprates in the normal conducting <lb/>state the characteristic scattering rate has been shown to increase linear in temperature fol-<lb/>lowing h/τ ≈ 2k B T [23]. Assuming the value ∆ = 2k B T C for the energy gap [24], one obtains <lb/>the approximation for the ratio of mean free path to the coherence length, ℓ/ξ = πτ ∆/ℏ ≈ 3. <lb/>As the scattering rate decreases even stronger than linearly below T C [1], at low tempera-<lb/>tures the high-T C superconductors are in the clean limit and the spectral weight of a Drude <lb/>peak (ε 0 ω 2 <lb/>p,n = ne 2 /m * ) is expected to fully condense into the delta function at T =0K. <lb/>This conclusion is supported by the Drude-analysis of the frequency dependent conductivity <lb/>of YBa 2 Cu 3 O 7−δ [5] and by a comparison of the conductivity spectral weights at infrared <lb/>frequencies [25]. In the clean limit the following expression holds <lb/>1 <lb/>µ 0 λ(0) 2 = ε 0 ω 2 <lb/>p,s = ε 0 ω <lb/>p,n = σ n /τ <lb/>(1) <lb/>
11
+ The Eq.(1) thus suggests the possible explanation of the correlation in Fig. 1: obviously <lb/>defects (doping, sample quality) and oxygen concentration change only the density of states <lb/>at the Fermi-level, N(0), or the effective number of charge carriers, ω 2 <lb/>p,s and ω 2 <lb/>p,n , and do not <lb/>affect the effective scattering rate τ −1 . Theoretical calculations of the normal-state prop-<lb/>erties using spin-fluctuation scattering [26], Fermi-surface nesting [27] or phenomenological <lb/>marginal Fermi-liquid approach [28] result in the quasiparticle scattering rate, which is not <lb/>dependent on N(0), but is approximately proportional to temperature. Therefore, within a <lb/>reasonable approximation, τ −1 may be supposed to be unsensitve to the level of defects. As <lb/>long as the sample is in the clean limit, the complete Drude spectral weight ω 2 <lb/>p,n condenses <lb/>into the superconducting delta function retaining the correlation between the normal and <lb/>the superconducting state given by Eq. (1). <lb/>Interestingly, the Ni-and Zn-doped samples of Ulm et al. [17] and the Pr doped samples of <lb/>Brorson et al. [9] still follow the correlation of Fig.1, except for the samples with the highest <lb/>doping levels (6% Ni, 40% Pr, or 4% Zn doping). The different behavior of heavily doped <lb/>samples is probably due to the fact, that they are no more in the clean limit. Therefore a <lb/>substantial portion of the &quot;normal&quot; spectral weight is not condensed into the delta function <lb/>at low temperatures and the penetration depth λ(0) = c/ω p,s is substantially higher than <lb/>the value estimated from Eq. (1), λ(0) = c/ω p,n . In addition Fig. 1 shows that Zn doping <lb/>has the strongest effect on the electrodynamic properties of high-T C &apos;s compared to Ni or Pr. <lb/>Similar conclusion has been drawn on the basis of the microwave data by Bonn et al. [29]. <lb/>Although simple arguments, presented above, give an idea to understand the observed <lb/>correlation, we cannot exclude other possible explanations. E.g. even the application of <lb/>Fermi-liquid picture to high-temperature superconductors is still an intensively discussed <lb/>problem [2,30]. Further on and especially concerning optimally doped YBCO films, the <lb/>extrinsic effects may also play an important role in determining the conductivity values. The <lb/>influence of weak links on the surface impedance of granular superconductors was discussed <lb/>in detail by Halbritter [31]. <lb/>The scattering rate 1/τ = 22 THz, as deduced from the dashed line in Fig. 1, leads <lb/>to the following expression at T =100 K: h/τ = 1.7k B T . This relation is similar to h/τ ≈ <lb/>2k B T [23] cited above. On the basis of the above presented discussion we may conclude, <lb/>that the scattering rate h/τ = 1.7k B T is to a good approximation independent of doping, <lb/>oxygen depletion and defects. Hence, the proportionality between λ(0) −2 and σ n provides <lb/>a simple possibility to characterize the superconducting properties of the YBaCuO samples <lb/>by measuring the normal state conductivity by e.g. standard four point method. On the <lb/>basis of data, presented in Fig.1 the penetration depth of a certain sample is given by the <lb/>approximate expression <lb/>λ(0)[nm] = 190/ σ n (100K)[10 4 Ω −1 cm −1 ] = 19 • ρ n (100K)[µΩ • cm] <lb/>(2) <lb/>Eq. (2) describes 80% of all penetration depth data in Fig. 1 with deviations well <lb/>below 25%. Finally it should be noted, that most points of Fig.1 represent thin-film data. <lb/>The applicability of Eq. (2) for single crystals should be checked in more detail in further <lb/>investigations. <lb/>In conclusion, we have measured the conductivity and the penetration depth of a variety <lb/>of YBaCuO samples and compared our results to published data. We have plotted the <lb/>low temperature penetration depth as a function of the normal-state conductivity, where <lb/>the absolute values for both parameters were available on the same sample. In this plot <lb/>we observe a correlation λ −2 ∽ σ n . This observation allows the estimate of the penetration <lb/>depth of a given YBaCuO sample by measuring its normal state conductivity. In the normal <lb/>conducting state we find universal scattering rate h/τ = 1.7k B T independent of defects, <lb/>oxygen concentration and doping. <lb/></body>
12
+
13
+ <div type="acknowledgement">This research was partly supported by the BMBF under the contract number 13N6917/0 <lb/>(EKM). <lb/></div>
14
+
15
+ <page>5 <lb/></page>
16
+
17
+ <listBibl>REFERENCES <lb/>[1] Bonn D. A. and Hardy W. N., in Physical Properties of high Temperature Superconduc-<lb/>tors V, Ed. by D. M. Ginsberg (World Scientific, Singapore, 1996). p. 7. <lb/>[2] Hensen S., Müller G., Rieck C. T. and Scharnberg K., Phys. Rev. B56 (1997) 6237. <lb/>[3] Volkov A. A., Goncharov Yu. G., Kozlov G. V., Lebedev S. P. and Prochorov A. M., <lb/>Infrared Phys. 25 (1985) 369; G. V. Kozlov and A. A. Volkov in Millimeter and Sub-<lb/>millimeter Wave Spectroscopy of Solids, Ed. by G. Grüner (Springer, Berlin, 1998), <lb/>p.51. <lb/>[4] Lemberger T. R., in Physical Properties of high Temperature Superconductors III, Ed. <lb/>by D. M. Ginsberg (World Scientific, Singapore, 1992). p. 471. <lb/>[5] Pimenov A., Loidl A., Jakob G. and Adrian H., Phys. Rev. B59 (1999) 4390. <lb/>[6] Pimenov A., Pronin A. V., Schey B., Stritzker B. and Loidl A., Physica B244 (1998) <lb/>49. <lb/>[7] Pimenov A., Loidl A., Jakob G. and Adrian H., cond-mat/9901039 . <lb/>[8] Iye Y., in Physical Properties of high Temperature Superconductors III, Ed. by D. M. <lb/>Ginsberg (World Scientific, Singapore, 1992), p. 285. <lb/>[9] Brorson S. D., Buhleier R., Trofimov I. E., White J. O., Ludwig Ch., Balakirev F. F., <lb/>Habermeier H.-U. and Kuhl J., J. Opt. Soc. Am. B13 (1996) 1979 . <lb/>[10] Feenstra B. J., Klaassen F. C., van der Marel D., Barber Z. H., Pinaya R. P. and <lb/>Decroux M., Physica C278 (1997) 213 . <lb/>[11] Ludwig Ch., Jiang Q., Kuhl J. and Zegenhagen J., Physica C269 (1996) 249. <lb/>[12] Nagashima T., Harada S., Hangyo M. and Nakashima S., Physica C293 (1997) 283. <lb/>[13] Dähne U., Amrein T., Goncharov Y., Klein N., Kozlov G., Schultz L., Tellmann N. <lb/></listBibl>
18
+
19
+ <page>6 <lb/></page>
20
+
21
+ <listBibl>and Urban K., Physica C235-240 (1994) 2066; Dähne U., Goncharov Y., Klein N., <lb/>Tellmann N., Kozlov G. and Urban K., J. Supercond. 8 (1995) 129. <lb/>[14] Nuss M. C., Mankiewich P. M., O&apos;Malley M. L.,. Westerwick E. H and Littlewood P. <lb/>B., Phys. Rev. Lett. 66 (1991) 3305. <lb/>[15] de Vaulchier L. A., Vieren J. P., Guldner Y., Bontemps N., Combescot R., Lemaître <lb/>Y. and Mage J. C., Europhys. Lett. 33 (1996) 153; Djordjevic S., de Vaulchier L. A., <lb/>Bontemps N., Vieren J. P., Guldner Y., Moffat S., Preston J., Castel X., Guilloux-Viry <lb/>M. and Perrin A., Eur. Phys. J. B5 (1998) 847. <lb/>[16] Fiory A. T., Hebard A. F., Eick R. H., Mankiewich P. M., Howard R. E. and O&apos;Malley <lb/>M. L., Phys. Rev. Lett. 65 (1990) 3441. <lb/>[17] Ulm E. R., Kim J.-T., Lemberger T. R., Foltyn S. R. and Wu X., Phys. Rev. B51 <lb/>(1995) 9193; Kim J.-T., Lemberger T. R., Foltyn S. R. and Wu X., Phys. Rev. B49 <lb/>(1994) 15970; Sumner M. J., Kim J.-T. and Lemberger T. R., Phys. Rev. B47 (1993) <lb/>12248. <lb/>[18] Basov D. N., Liang R., Bonn D. A., Hardy W. N., Dabrowski B., Quijada M., Tanner <lb/>D. B., Rice J. P., Ginsberg D. M. and Timusk T., Phys. Rev. Lett. 74 (1995) 598. <lb/>[19] ZhangK., Bonn D. A., Kamal S., Liang R., Baar D. J., Hardy W. N., Basov D. and <lb/>Timusk T., Phys. Rev. Lett. 73 (1994) 2484. <lb/>[20] Kitano H., Shibauchi T., Uchinokura K., Maeda A., Asaoka H. and Takei H., Phys. <lb/>Rev. B51 (1995) 1401. <lb/>[21] Sonier J. E., Kiefl R. F., Brewer J. H.,. Bonn D. A, Carolan J. F., Chow K. H., Dosanjh <lb/>P., Hardy W. N., Liang R., MacFarlane W. A., Mendels P., Morris G. D., Riseman T. <lb/>M. and Schneider J. W., Phys. Rev. Lett. (1994) 744; Liang R., Dosanjh P., Bonn <lb/>D. A., Baar D. J., Carolan J. F. and Hardy W. N., Physica C195 (1992) 51. <lb/></listBibl>
22
+
23
+ <page>7 <lb/></page>
24
+
25
+ <listBibl>[22] Kamal S., Liang R., Hosseini A., Bonn D.A. and Hardy W. N., Phys. Rev. B58 (1998) <lb/>8933. <lb/>[23] Tanner D. B. and Timusk T. in Physical Properties of high Temperature Superconductors <lb/>III, Ed. by D. M. Ginsberg (World Scientific, Singapore, 1992), p. 363. <lb/>[24] Panagopoulos Ch. and Xiang T., Phys. Rev. Lett., 81 (1998) 2336. <lb/>[25] Tanner D. B., Liu H. L., Quijada M. A., Zibold A. M., Berger H., Kelley R. J., Onellion <lb/>M., Chou F. C., Johnston D. C., Rice J. P., Ginsberg D. M. and Markert J. T., Physica <lb/>B244 (1998) 1. <lb/>[26] Hirschfeld P. J., Putikka W. O., and Scalapino D. J., Phys. Rev. B50 (1994) 10250. <lb/>[27] Virosztek A. and Ruvalds J., Phys. Rev. B42 (1990) 4064. <lb/>[28] Littlewood P. B. and Varma C. M., J. Appl. Phys. (1991) 4979. <lb/>[29] Bonn D.A., Kamal S., Zhang K., Liang R., Baar D. J., Klein E. and Hardy W. N., <lb/>Phys. Rev. B50 (1994) 4051. <lb/>[30] Shulga D. V., Dolgov O. V., and Maksimov E. G., Physica C178 (1991) 266. <lb/>[31] Halbritter J., J. Supercond. 8 (1995) 691. <lb/></listBibl>
26
+
27
+ <page>8 <lb/></page>
28
+
29
+ <body>No. T C [K] Preparation <lb/>Substrate <lb/>d <lb/>[nm] <lb/>σ n <lb/>[10 4 Ω −1 cm −1 ] <lb/>λ <lb/>[nm] <lb/>ω 2 <lb/>p,s <lb/>[eV 2 ] <lb/>τ -1 <lb/>[THz] <lb/>Ref. <lb/>1 <lb/>89.5 magnetron sputt. NdGaO 3 81 <lb/>1.51 <lb/>1.71 22.8 [4] <lb/>2 <lb/>91.3 magnetron sputt. LaAlO 3 <lb/>1.0 <lb/>204 0.952 19.1 <lb/>3 <lb/>85.9 <lb/>laser ablation <lb/>MgO <lb/>85 <lb/>0.80 <lb/>243 0.671 16.8 [5] <lb/>4 <lb/>89 magnetron sputt. NdGaO 3 75 <lb/>0.64 <lb/>215 0.857 26.9 <lb/>5 <lb/>56.5 <lb/>No.1 reduced NdGaO 3 81 <lb/>0.60 <lb/>279 0.509 17.0 [6] <lb/>6 <lb/>91 <lb/>laser ablation NdGaO 3 70 <lb/>0.38 <lb/>292 0.464 24.6 <lb/>Table 1. Submillimeter-wave properties of YBa 2 Cu 3 O 7−δ films at ν = 450 GHz. The <lb/>normal state conductivity σ n and the superconducting penetration depth λ are given at <lb/>temperatures 100 K and 6 K respectively. The spectral weight of the superconducting <lb/>condensate ω 2 <lb/>p,s and the effective scattering rate 1/τ were calculated using Eq. (1). <lb/></body>
30
+
31
+ <page>9 <lb/></page>
32
+
33
+ <body>Figure caption. <lb/>Fig. 1. Low temperature (T K) penetration depth λ of different YBCO samples as <lb/>a function of the normal-state conductivity σ n at T =100 K. Dashed line is drawn according <lb/>to the expression λ(0) −2 /µ 0 = σ n /τ using 1/τ = 22THz. The characters in the symbols <lb/>correspond to: <lb/>Rhombs -single crystal data: <lb/>B -Basov et al. [18] and Zhang et al. [19]; K -Kitano et al. [20]; K1 -Kamal et al. [22]; <lb/>S -Sonier et al. [21]. <lb/>Circles -optimally doped YBa 2 Cu 3 O 7−δ films: <lb/>closed circles -this work, Table 1; B -Brorson et al. [9]; D -Dähne et al. [13]; F -Fiory <lb/>et al. [16]; L -Ludwig et al. [11]; N -Nagashima et al. [12] (λ is taken at T =40 K); N1 -<lb/>Nuss et al. [14]; V -de Vaulchier et al. [15]. <lb/>Squares -oxygen reduced YBa 2 Cu 3 O 7−δ films: <lb/>closed square -this work, Table 1; L -Ludwig et al. [11]. <lb/>Down triangles -doped YBa 2 Cu 3 O 7−δ films: <lb/>B -Brorson et al. [9] (20%, 30%, and 40% Pr -doped films); F -Feenstra et al. [10] <lb/>(DyBa 2 Cu 3 O 7−δ ); U -Ulm et al. [17] (2%, 4% and 6% Ni doped and 2% and 4% Zn doped <lb/>films). Dotted lines are guide to the eye. Arrows indicate the increasing doping direction.<lb/>
34
+ 0.1 <lb/>1000 <lb/>K <lb/>K <lb/>B <lb/>B <lb/>S <lb/>K1 <lb/>K1 <lb/>L <lb/>L <lb/>U <lb/>U <lb/>U <lb/>U <lb/>U <lb/>B <lb/>B <lb/>B <lb/>F <lb/>N <lb/>L <lb/>B <lb/>N1 <lb/>V <lb/>V <lb/>V <lb/>V <lb/>F <lb/>F <lb/>F <lb/>F <lb/>K1 <lb/>crystals <lb/>films opt. <lb/>films red. <lb/>films dop. <lb/>Fig. 1, Pimenov et al. <lb/>D <lb/>N i d o p e d <lb/>P r d o p e d <lb/>Z n d o p e d <lb/>F <lb/>V <lb/>τ -1 <lb/>= 2 2 T H z <lb/>YBaCuO <lb/>λ [nm] <lb/>σ 1 (100 K) [10 <lb/>4 Ω <lb/>-1 cm <lb/>-1 ] </body>
35
+
36
+
37
+ </text>
38
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EPL0490086-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <front>cond-mat/9903395 v2 24 Oct 1999 <lb/>The pseudogap in Bi 2 Sr 2 CaCu 2 O 8+x single crystals from <lb/>tunneling measurements: a possible evidence for the Cooper <lb/>pairs above T c . <lb/>A. Mourachkine <lb/>Université Libre de Bruxelles, Service de Physique des Solides, CP233, Blvd du Triomphe, B-1050 <lb/>Brussels, Belgium <lb/>(received 29 March 1999; accepted in final form ______ 1999) <lb/>PACS. 74.50.+r -Proximity effects, weak links, tunneling phenomena, and Josephson <lb/>effects. <lb/>PACS. 74.25.-q -General properties; correlation between physical properties in normal <lb/>and superconducting states. <lb/>PACS. 74.72.Hs -Bi-based cuprates. <lb/>Abstract. -We present electron-tunneling spectroscopy of slightly <lb/>overdoped Bi 2 Sr 2 CaCu 2 O 8-x (Bi2212) single crystals with T c = 87 -90 K <lb/>in a temperature range between 14 K and 290 K using a break-junction <lb/>technique. The pseudogap which has been detected above T c appears at T* <lb/>≈ 280 K. The analysis of the spectra shows that there is a contribution to <lb/>the pseudogap above T c , which disappears approximately at 110 -115 K. <lb/>We associate this contribution with the presence of incoherent Cooper <lb/>pairs. <lb/></front>
9
+
10
+ <body>The existence of a pseudogap in electronic excitation spectra of high-T c <lb/>superconductors (HTSC), which appears below a certain temperature T* &gt; T c , <lb/>is considered to be amongst the most important features of cuprates. Many <lb/>experiments have provided evidence (NMR [1], angle-resolved photoemission <lb/>(ARPES) [2], specific heat [3], electron-tunneling spectroscopy [4] and STM [5]) <lb/>for a gap-like structure in electronic excitation spectra. There is a consensus on <lb/>that the pseudogap is a characteristic feature of the underdoped regime of <lb/>copper-oxides, and the value of T* increases with the decrease of hole <lb/>concentration. For example, in tunneling measurements on underdoped Bi2212 <lb/>single crystals by STM, the pseudogap was detected even at room temperature <lb/>[5]. At the same time, tunneling measurements showed that the pseudogap is also <lb/>present in slightly overdoped samples as well [4]. An advantage of tunneling <lb/>spectroscopy is that it is particularly sensitive to the density of state (DOS) near <lb/>the Fermi level E F and, thus, is capable of detecting any gap in the quasiparticle <lb/>excitation spectrum at E F [6]. In addition to this, tunneling spectroscopy has a <lb/>high energy resolution [6]. <lb/>In general, the pseudogap in cuprates can originate from the charge-density-<lb/>wave (CDW) order, the spin-density-wave (SDW) order due to local <lb/>antiferromagnetic (AF) correlations, the presence of incoherent Cooper pairs, <lb/>or from their combination (so-called Ansatz) [7]. From earlier studies, the <lb/>pseudogap was associated with a SDW gap due to local AF correlations [1]. <lb/>Very recent measurements of high-frequency conductivity that track the phase-<lb/>correlation time in the normal state of underdoped Bi2212 showed that the <lb/>incoherent Cooper pairs remain up to T pair which is well above T c , however, <lb/>lower than T* [8]. Thus, between T c and T pair , there is a contribution of <lb/>incoherent Cooper pairs to the pseudogap. Indeed, there is a consensus on that, <lb/>to the contrary to conventional SCs, the formation of the Copper pairs and the <lb/>establishment of the phase coherence among them in cuprates occur at different <lb/>temperatures, T pair &gt; T c [9,8,5]. So, it is reasonable to expect that the incoherent <lb/>Cooper pairs have to be observed above T c in different types of measurements <lb/>including tunneling. The T pair scales with T*, thus, the difference between T pair <lb/>and T c is the highest in the underdoped regime and becomes smaller in the <lb/>overdoped regime [9]. <lb/>We present here direct measurements of the density of states by electron-<lb/>tunneling spectroscopy on slightly overdoped Bi2212 single crystals using a <lb/>break-junction technique. The pseudogap which has been detected above T c <lb/>appears at T* ≈ 280 K. The analysis of the spectra shows that there is a <lb/>contribution to the pseudogap above T c , which disappears between T c and T*. <lb/>We associate this contribution to the pseudogap with the presence of incoherent <lb/>Cooper pairs in Bi2212. To our knowledge, the data shown in Fig. 2 are the <lb/>most detailed tunneling data of the pseudogap presented in the literature. <lb/>The single crystals of Bi2212 were grown using a self-flux method and then <lb/>mechanically separated from the flux in Al 2 O 3 or ZrO 2 crucibles [10]. The <lb/>dimensions of the samples are typically 3×1×0.1 mm 3 . The chemical <lb/>composition of the Bi2212 phase corresponds to the formula <lb/>Bi 2 Sr 1.9 CaCu 1.8 O 8+x as measured by energy dispersive X-ray fluorescence <lb/>(EDAX). The crystallographic a, b, c values of the overdoped single crystals <lb/>are of 5.41 Å, 5.50 Å and 30.81 Å, respectively. The T c value was determined <lb/>by either dc-magnetization or by four-contacts method yielding T c = 87 -90 K <lb/>with the transition width ∆T c ~ 1 K. The single crystals were checked out to <lb/>assure that they are in the overdoped phase. <lb/>Experimental details of our break-junction technique can be found elsewhere <lb/>[11]. Shortly, many break-junctions were prepared by gluing a sample with <lb/>epoxy on a flexible insulating substrate and then were broken by bending the <lb/>substrate with a differential screw at 14 -18 K in a helium atmosphere. The <lb/>electrical contacts were made by attaching gold wires to a crystal with silver <lb/>paint. The sample resistance (with the contacts), R sample , at room temperature <lb/>varied from 5 Ω to about 1 kΩ, depending on the sample. The I(V) and <lb/>dI/dV(V) tunneling characteristics were determined by the four-terminal <lb/>method using a standard lock-in modulation technique. The tunneling spectra of <lb/>our break junctions on Bi2212 single crystals exhibit below T c the characteristic <lb/>features of typical tunneling spectra in Bi2212 [12,5,4]. <lb/>Figure 1 shows a set of normalized tunneling conductance curves measured <lb/>in superconductor-insulator-superconductor (SIS) between 14 K and 290 K on <lb/>an overdoped Bi2212 single crystal with T c = 88.5 K. All curves, except the last <lb/>one, show a gaplike structure. In the absence of generally accepted model for <lb/>the pseudogap, we consider the presence of the gap-like features in tunneling <lb/>spectra as a sign of the pseudogap (for more details, see Ref. 5). There is no <lb/>sign indicating at what temperature the superconducting gap was closed. Across <lb/>T c , the superconducting tunneling spectra evolve continuously into a normal-<lb/>state quasiparticle-gap structure which vanishes at 232 K &lt; T * &lt; 290 K but <lb/>remains almost unchanged with temperature up to 232 K [13]. Such evolution of <lb/>superconducting spectra into normal-state spectra across T c has been also <lb/>observed in superconductor-insulator-normal metal (SIN) junctions [5]. Thus, it <lb/>is clear that the pseudogap is related somehow to the superconductivity [5]. The <lb/>spectra in Fig. 1 show the change in conductance from very low to high <lb/>temperatures. It is difficult in such scale to see the differences between spectra <lb/>above T c . Therefore, Figure 2 shows tunneling spectra only above T c , measured <lb/>on another overdoped Bi2212 single crystal with T c = 88 K. In Fig. 2, one can <lb/>see that there is a gap-like structure just above T c which disappears at some <lb/>temperature T 0 much lower than T*. Most pseudogap data observed in our <lb/>study look similar to the data presented in Fig. 1, however, measurements on <lb/>the best samples [14] show some gap-like features in tunneling spectra above T c , <lb/>which look similar to the gap-like structure shown in Fig. 2. Therefore, we <lb/>believe that these features reflect intrinsic properties of overdoped Bi2212. <lb/>The spectra shown in Fig. 2 are asymmetrical about zero bias. Such behavior <lb/>is typical for tunneling spectra obtained on the nonmetallic surfaces [15]. In <lb/>fact, the contribution of the Cooper pairs to the pseudogap must not depend on <lb/>the direction of bias. Thus, it is more convenient to consider only an even part <lb/>of the conductivity, G e (V) ≡ [G(V) + G(-V)]/2 [16]. Figure 3 shows the even <lb/>part of the tunneling spectra presented in Fig. 2. In Fig. 3, one can clearly <lb/>observe two humps at different temperatures, one of them disappears at T 0 &gt; <lb/>100 K, and the second one is traceable up to high temperatures. It is difficult to <lb/>indicate exactly the origins of these humps, especially, the wide ones which are <lb/>traceable to high temperatures. However, we will try just to determine <lb/>characteristic temperatures from temperature dependencies of these humps, <lb/>which are shown in Fig. 4. We use straight (dash) line in Fig. 4 in order to <lb/>estimate the value of T 0 which is of the order of 110 -115 K. The origin of this <lb/>hump we discuss further. We concentrate now on the behavior of the curve with <lb/>dots in Fig. 4, which correspond to the temperature dependence of wide humps. <lb/>The fall of this curve with the increase of temperature between T c and T 0 <lb/>supports an idea that there are some changes in the system. The indication of <lb/>some changes which begin at ~ 210 K with the increase of temperature can be <lb/>explained by the disappearance of local AF correlations [17]. It seems that, <lb/>between T 0 and 210 K, there is no principal changes in the system. From Figs. <lb/>1, 2 and 3, the value of T * is of the order of 280 K. In Fig. 3, in addition to the <lb/>two humps which have been discussed above, one can observe also a hump at 70 <lb/>mV between 103 K and 122 K [18]. <lb/>We discuss now the origin of the contribution to the pseudogap, which <lb/>disappears at T 0 with the increase of temperature. This contribution can be <lb/>explained by the disappearance of the CDW or/and SDW order(s), or <lb/>incoherent Cooper pairs. From the common sense, the CDW or SDW order is <lb/>likely to be appear above T c and not to disappear. Consequently, it is most <lb/>likely that this contribution to the pseudogap is made by preformed pairs <lb/>[8,9,5]. Tunneling spectroscopy is sensitive exclusively to the magnitude of the <lb/>order parameter. Thus, if the humps originate from the presence of <lb/>preformed pairs, then the magnitude of the gap-like structure reflects the <lb/>average value of binding energy per particle [9]. However, our results alone <lb/>cannot prove that this contribution originates from preformed pairs. We can <lb/>argue that high-frequency conductivity measurements [8] showed that the <lb/>incoherent Copper pairs remain well above T c in Bi2212, and it was predicted <lb/>by the theory [17,19]. It is also in agreement with a MCS (Magnetic Coupling <lb/>between Stripes) model proposed very recently [20,21]. There is also an <lb/>alternative explanation for the presence of this contribution to the pseudogap, <lb/>it is possible that this contribution originates from the presence of the Bi2223 <lb/>phase with T c = 110 K [22]. Even in this case, our data obtained in SIS <lb/>junctions support in general the conclusions made on the basis of tunneling <lb/>data in SIN junctions, namely, that the pseudogap is directly related to <lb/>superconductivity [5]. <lb/>Recently, the pseudogap has been observed in Bi2212 with T c = 90 K <lb/>(optimum doping) and T c = 84 K (slightly overdoped) by SIN tunneling <lb/>technique [23]. The measured value of T* is equal to 300 K and 270 K, <lb/>respectively. Thus, the value of T* ≈ 280 K found in our measurements for <lb/>slightly overdoped single crystals is in an excellent agreement with the SIN <lb/>tunneling measurements [23]. In slightly overdoped Bi2212, the pseudogap has <lb/>been also observed in SIS tunneling measurements performed by break-<lb/>junction technique [24]. The value of T* is found to be of the order of 190 K. <lb/>However, the analysis of SIS spectra presented in Ref. 24 (see Fig. 4) shows <lb/>that, in fact, the T* value is of the order of 290 -300 K. In the literature, <lb/>there is a clear discrepancy in the definition of T*. For slightly overdoped <lb/>Bi2212, the temperature 270 -300 K corresponds most likely to the charge <lb/>ordering, T* charge ≈ 270 -300 K [17]. The temperature T* ≈ 190 K found in <lb/>Ref. 24 and the changes which were observed at ~ 210 K in our study (see Fig. <lb/>4) correspond most likely to the spin ordering, T* spin ≈ 190 -210 K [17]. <lb/>Thus, there is a very good agreement among tunneling data presented in Refs. <lb/>23, 24 and the present work. What is interesting, in SIS tunneling <lb/>measurements, the value of 116 K has been also discussed and assumed to be <lb/>the local onset of superconductivity [24]. <lb/>In summary, we presented direct measurements of the density-of-state by <lb/>tunneling spectroscopy on slightly overdoped Bi 2 Sr 2 CaCu 2 O 8-x single crystals <lb/>in a temperature range between 14 K and 290 K using the break-junction <lb/>technique. The pseudogap which has been detected above T c appears at T* ≈ <lb/>280 K. The fine analysis of the spectra shows that there is a contribution to the <lb/>pseudogap above T c . We associate this contribution with the presence of <lb/>incoherent Cooper pairs in Bi2212, which disappears approximately at 110 -<lb/>115 K. However, alternative explanations for the presence of this contribution <lb/>are also possible. In all cases, there is no doubts that the pseudogap observed <lb/>in our study is partially or entirely related to the superconductivity.<lb/></body>
11
+
12
+ * * * <lb/>
13
+
14
+ <div type="acknowledgement">I thank R. Deltour for discussion. This work is supported by PAI 4/10. <lb/></div>
15
+
16
+ <listBibl>REFERENCES <lb/>[1] Warren W. W. et al., Phys. Rev. Lett. 62 (1989) 1193. <lb/>[2] Norman M. R. et al., Nature 392 (1998) 157; Ding H. et al., Nature 382 <lb/>(1996) 51; Shen Z.-X. et al., Phys. Rev. Lett. 70 (1993) 1533. <lb/>[3] Loram J. et al, Phys. Rev. Lett., 71 (1993) 1740. <lb/>[4] Tao H. J., Lu Farun, and Wolf E. L., Physica C 282-287 (1997) 1507. <lb/>[5] Renner Ch. et al., Phys. Rev. Lett. 80 (1998) 149. <lb/>[6] Wolf E. L., Principles of Tunneling Spectroscopy (Oxford University <lb/>Press, New York, 1985). <lb/>[7] Markiewicz R. S., J. Phys. Chem. Sol. 58 (1997) 1179. <lb/>[8] Corson J. et al., Nature 398 (1999) 221. <lb/>[9] Deutscher G., Nature 397 (1999) 410. <lb/>[10] Davydov D. N. et al., Solid State Commun. 86 (1993) 267. <lb/>[11] Hancotte H. et al., Physica C 280 (1997) 71. <lb/>[12] DeWide Y. et al., Phys. Rev. Lett. 80 (1998) 153, and Miyakawa N. et al., <lb/>ibid. p. 157. <lb/>[13] In the absence of well established procedure of the determination of the <lb/>value of T* (or T o ), the T* (or T o ) value is determined when the gap-like <lb/>structure in tunneling spectra disappears. <lb/>[14] The samples with low value of R sample (~ 5 ÷ 50 Ω), we call the best. It <lb/>seems that there is a correlation between the value of R sample and how <lb/>easy is to detect a gaplike structure in I-V tunneling characteristics (i.e. <lb/>the SC DOS). The configuration of our break-junction technique does not <lb/>allow us to control the place for tunneling like the STM technique. We <lb/>control only the distance between two parts of broken crystal. In most <lb/>samples, it takes time to detect a gaplike structure by changing the <lb/>distance, going back and forth. As rule, in samples with low value of <lb/>R sample , it is easy to find a gaplike structure (i.e. the SC DOS), and <lb/>tunneling spectra above T c resemble the spectra shown in Fig. 2. It is not <lb/>easy to make any conclusions concerning the physical meaning of this <lb/>correlation which can be paraphrased as follows: tunneling spectra which <lb/>look similar to the spectra shown in Fig. 2 are observed as rule in more <lb/>&quot;metallic&quot; samples. <lb/>[15] Hasegawa T., Nantoh M., Ogino M., and Kitazawa K., in Bismuth-Based <lb/>High-Temperature Superconductors, edited by H. Maeda and K. Togano <lb/>(Marsel Dekker Inc., New York, 1996) p. 75 -92. <lb/>[16] Tsui, D. C., Dietz, R. E. &amp; Walker, L. R. Phys. Rev. Lett. 27 (1971) <lb/>1729. <lb/>[17] Emery V. J., Kivelson S. A., and Zachar O., Phys. Rev. B 56 (1997) <lb/>6120. <lb/>[18] It is difficult to indicate if the presence of this hump is meaningful or not. <lb/>However, it is clear that the amplitude of this hump is much weaker than <lb/>the amplitude of the hump which disappears at 110 -115 K. <lb/>[19] Alexandrov A. S., Kabanov V. V. &amp; Mott N. F., Phys. Rev. Lett. 77, <lb/>(1996) 4796. <lb/>[20] Mourachkine, A., to be published in J. Superconductivity (also cond-<lb/>mat/9902355). <lb/>[21] Mourachkine, A., proceedings of MOS-99 Conference (Stockholm 28 July <lb/>-2 August 1999), to be published in J. Low Temp. Physics (also cond-<lb/>mat/9908052). <lb/>[22] However, no presence of the Bi2223 phase has been detected in our <lb/>Bi2212 samples by EDAX measurements. <lb/>[23] A. Matsuda, S. Sugita, and T Watanabe, Phys. Rev. B 60, 1377 (1999). <lb/>[24] T. Ekino, Y. Sezaki, and H. Fujii, Phys. Rev. B 60, 6916 (1999). <lb/></listBibl>
17
+
18
+ <body>FIGURE CAPTIONS: <lb/>FIG. 1. Tunneling spectra measured in a SIS junction as a function of <lb/>temperature on an overdoped Bi2212 single crystal. The conductance scale <lb/>corresponds to the 290 K spectrum, the other spectra are offset vertically for <lb/>clarity. The curves have been normalized at -150 mV (or nearest point). <lb/>FIG. 2. Tunneling spectra measured as a function of temperature on an <lb/>overdoped Bi2212 single crystal. The conductance scale corresponds to the 290 <lb/>K spectrum, the other spectra are offset vertically for clarity. The curves have <lb/>been normalized at -200 mV (or nearest point). <lb/>FIG. 3. Even conductance of the spectra shown in Fig. 2 as a function of <lb/>temperature. <lb/>FIG. 4. Temperature dependence of quasiparticle DOS shown in Fig. 3. The <lb/>dashed line is a guide to the eye. <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>-200 <lb/>-100 <lb/>0 <lb/>100 <lb/>200 <lb/>Normalized Conductance <lb/>Bias (mV) <lb/>34 K <lb/>14 K <lb/>43 K <lb/>75 K <lb/>80 K <lb/>82 K <lb/>88 K <lb/>99 K <lb/>122 K <lb/>147 K <lb/>174 K <lb/>201 K <lb/>232 K <lb/>290 K <lb/>T = 88.5 K <lb/>c <lb/>overdoped <lb/>2∆ = 45 meV <lb/>FIG. 1 <lb/>-400 <lb/>-200 <lb/>0 <lb/>200 <lb/>400 <lb/>Normalized Conductance <lb/>Bias (mV) <lb/>86.8 K <lb/>T = 88 K <lb/>c <lb/>overdoped <lb/>88.1 K <lb/>91.5 K <lb/>98 K <lb/>99 K <lb/>103 K <lb/>109 K <lb/>122 K <lb/>134 K <lb/>150 K <lb/>160 K <lb/>180 K <lb/>200 K <lb/>220 K <lb/>290 K <lb/>280 K <lb/>270 K <lb/>230 K <lb/>240 K <lb/>250 K <lb/>1 <lb/>1.2 <lb/>1.4 <lb/>1.6 <lb/>1.8 <lb/>FIG. 2 <lb/>0 <lb/>100 <lb/>200 <lb/>300 <lb/>400 <lb/>Bias (mV) <lb/>Normalized Conductance, G <lb/>e <lb/>290 K <lb/>280 K <lb/>270 K <lb/>103 K <lb/>109 K <lb/>122 K <lb/>134 K <lb/>150 K <lb/>160 K <lb/>180 K <lb/>200 K <lb/>220 K <lb/>230 K <lb/>240 K <lb/>250 K <lb/>88.1 K <lb/>91.5 K <lb/>98 K <lb/>99 K <lb/>1 <lb/>1.2 <lb/>1.4 <lb/>1.6 <lb/>1.8 <lb/>FIG. 3 <lb/>0 <lb/>50 <lb/>100 <lb/>150 <lb/>200 <lb/>250 <lb/>300 <lb/>350 <lb/>100 <lb/>150 <lb/>200 <lb/>250 <lb/>300 <lb/>Hump position (mV) <lb/>Temperature (K) <lb/>T 0 <lb/>T c <lb/>T* <lb/>FIG. 4 </body>
19
+
20
+ </text>
21
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/EPL0580589-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <front>Europhys. Lett., 58 (4), pp. 589-595 (2002) <lb/>EUROPHYSICS LETTERS <lb/>15 May 2002 <lb/>Absence of pseudogap <lb/>in heavily overdoped Bi 2 Sr 2 CaCu 2 O 8+δ <lb/>from tunneling spectroscopy of break junctions <lb/>L. Ozyuzer 1,2 , J. F. Zasadzinski 1,3 , K. E. Gray , <lb/>C. Kendziora 4 and N. Miyakawa <lb/>1 Materials Science Division, Argonne National Laboratory -Argonne IL 60439, USA <lb/>Department of Physics, Izmir Institute of Technology -TR-35437 Izmir, Turkey <lb/>3 Physics Division, Illinois Institute of Technology -Chicago IL 60616, USA <lb/>4 Naval Research Laboratory -Washington DC 20375, USA <lb/>5 Department of Applied Physics, Science University of Tokyo -Tokyo, Japan <lb/>(received 29 September 2001; accepted in final form 21 February 2002) <lb/>PACS. 74.72.-h -High-Tc compounds. <lb/>PACS. 74.50.+r -Proximity effects, weak links, tunneling phenomena, and Josephson effects. <lb/>PACS. 74.62.Dh -Effects of crystal defects, doping and substitution. <lb/>Abstract. -We report tunneling spectroscopy of superconductor-insulator-superconductor <lb/>break junctions on heavily overdoped Bi2Sr2CaCu2O 8+δ with Tc = 56 K. At T <lb/>Tc, the <lb/>junction conductances display well-defined quasiparticle peaks at ±2∆ and in some cases a <lb/>Josephson current at zero bias. Gap values as small as ∆ = 10.5 meV have been observed <lb/>leading to 2∆/kTc near the BCS limit for d x 2 −y 2 pairing. Temperature dependence of the <lb/>gap magnitude, ∆(T ), follows the BCS relation and both the quasiparticle gap and Josephson <lb/>current vanish for T &gt; Tc. Above Tc, the tunneling conductance shows a flat background <lb/>without any indication of a pseudogap near the Fermi level. <lb/></front>
9
+
10
+ <body>High-temperature superconductors (HTSs) have emerged with properties that are very <lb/>different from those found in conventional superconductors. One such property of the under-<lb/>doped phase is the presence of a pseudogap observed well above T c , up to a characteristic <lb/>temperature T *, that has been detected by a number of experimental techniques, such as <lb/>in-plane resistivity, angle-resolved photoemission spectroscopy (ARPES), specific heat, and <lb/>NMR [1,2]. One reason why the presence of the pseudogap garners a lot of attention is that it <lb/>might be related to the phenomenon of superconductivity in HTSs. Numerous theories have <lb/>been proposed to explain the origin of the pseudogap in the underdoped phase, such as pairing <lb/>fluctuation theory, spin-charge separation scenario, phase fluctuations [3-6]. <lb/>While it is generally accepted that the pseudogap is a property of the underdoped phase, <lb/>it is still debatable whether it is also present in the overdoped region. Whereas T * can be <lb/>determined in various experiments (e.g., a change in slope of in-plane resistivity vs. T ) a <lb/>corresponding direct measurement of a pseudogap in the density of states is not always avail-<lb/>able. For overdoped HTSs, none of the experimental techniques listed earlier have detected a <lb/></body>
11
+
12
+ <note place="footnote">c EDP Sciences <lb/></note>
13
+
14
+ <page>590 <lb/></page>
15
+
16
+ <note place="headnote">EUROPHYSICS LETTERS <lb/></note>
17
+
18
+ <body>clear pseudogap [1]. Some results from scanning tunneling microscopy (STM) [7,8] and planar <lb/>tunnel [9] junctions have suggested that the pseudogap exists at all hole doping concentrations <lb/>in Bi 2 Sr 2 CaCu 2 O 8+δ (Bi-2212). In contrast, another STM study in the ab-plane of Bi-2212 <lb/>does not show any gap-like feature in the overdoped phase [10]. These conflicting observations <lb/>hinder our understanding of the pseudogap and demand further tunneling investigations of <lb/>the overdoped phase of Bi-2212. One reason for the conflict might be that the doping level of <lb/>Bi-2212 in the region of the junction is not the same as in the bulk. Miyakawa et al. [11] have <lb/>demonstrated a clear relationship between the energy gap (∆) and hole concentration and <lb/>therefore the measured gap provides important information on the local doping level. Here <lb/>we report measurements of heavily overdoped crystals (T c = 56 K) that exhibit a ∆ value as <lb/>low as 10.5 meV, much smaller than the optimal doped value of 38 meV and clearly in the <lb/>overdoped state. <lb/>The break junction method can be used to measure the superconductor-insulator-supercon-<lb/>ductor (SIS) tunneling conductance in conventional and HTSs. One advantage of this method <lb/>is that the junction can be fabricated at low temperatures so that it is unexposed to air. The <lb/>SIS tunneling spectroscopy is not only capable of measuring quasiparticle excitations but also <lb/>the Cooper pairs in the form of the Josephson current, I c , which is purely a superconducting <lb/>phenomenon. This capability provides an important advantage since I c (T ) → 0 defines the <lb/>junction T c . In other words, the phase coherence temperature can be identified by measuring <lb/>the temperature dependence of the critical current. <lb/>Single crystals of Bi-2212 were grown by a self-flux technique in a strong thermal gradient <lb/>to stabilize the direction of solidification. Overdoping was accomplished using stainless-steel <lb/>cells sealed with the sample immersed in liquid oxygen [12]. The critical temperature of the <lb/>samples was determined from the magnetization which showed sharp transitions. Tunneling <lb/>measurements were done with a point contact apparatus [13]. The Bi-2212 crystal is cleaved <lb/>along the ab-plane and mounted on a substrate so that the tip approaches along the c-axis. <lb/>A novel method is used to form the SIS break junctions. A differential micrometer-driven <lb/>Au tip approaches the sample at LHe temperature and superconductor-insulator-normal-<lb/>metal (SIN) junctions are first formed. These conductances exhibit quasiparticle peaks at <lb/>eV ∼ ±∆ and clearly show the standard dip and hump features above ∆ in the occupied part <lb/>of the DOS [7, 11]. Further increasing the force of the tip leads to an Ohmic contact (less <lb/>than 10 ohm) with the crystal and a mechanical bond. A Bi-2212 piece is easily dislodged by <lb/>cleaving along the double Bi-O layer. As a result of this process, an SIS junction forms between <lb/>the Bi-2212 piece and the rest of the crystal. The SIS conductances show quasiparticle peaks at <lb/>±2∆ that are consistent with the ∆ value obtained from the SIN junctions. Additionally, the <lb/>SIS conductances also show a symmetric dip, hump and background, and also the presence <lb/>of a Josephson current. The magnitude of this current generally decreases as the junction <lb/>resistance increases. <lb/>Figure 1(a) shows the temperature dependence of the tunneling conductance for junction <lb/># 1 on the heavily overdoped Bi-2212. The 4.2 K data show a tunneling conductance of <lb/>a Bi-2212 break junction with sharp quasiparticle peaks at ±2∆, a weak Josephson current <lb/>peak at zero bias, and the dip and hump features above ±2∆. One of the distinct features of <lb/>this junction is the relatively flat background. This is different than the typical SIS conduc-<lb/>tances found from optimally doped [14] and overdoped [15] Bi-2212 that show a decreasing <lb/>background up to ±350 meV. For SIS junctions at low temperatures, the thermal smearing <lb/>is minimal and the quasiparticle peaks can be used to estimate the energy gap size, which <lb/>is 14 meV for the 4.2 K spectra in fig. 1(a), giving a coupling ratio 2∆/kT c ∼ 5.8. A set <lb/>of temperature-dependent data for another SIS break junction (junction # 2) of Bi-2212 is <lb/>shown in fig. 1(b). Although the junction resistance in fig. 1(b) is approximately the same <lb/></body>
19
+
20
+ <note place="headnote">L. Ozyuzer et al.: Absence of pseudogap in heavily overdoped Bi-2212 <lb/></note>
21
+
22
+ <page>591 <lb/></page>
23
+
24
+ <body>0 <lb/>0.5 <lb/>1 <lb/>1.5 <lb/>2 <lb/>2.5 <lb/>3 <lb/>3.5 <lb/>-150-100 -50 0 50 100 150 <lb/>dI/dV(mS) <lb/>Voltage(mV) <lb/>T=4.2 K <lb/>T=15 K <lb/>T=30 K <lb/>T=39 K <lb/>T=48 K <lb/>T=50 K <lb/>T=55 K <lb/>T=59 K <lb/>T=67 K <lb/>(a) <lb/>0 <lb/>2 <lb/>4 <lb/>0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>1.2 <lb/>1.4 <lb/>1.6 <lb/>-150-100 -50 0 50 100 150 <lb/>dI/dV(mS) <lb/>dI/dV(mS) <lb/>T=41 K <lb/>T=44 K <lb/>T=48 K <lb/>T=51 K <lb/>T=61 K <lb/>T=4.2 K <lb/>Voltage (mV) <lb/>(b) <lb/>0.05 <lb/>0.1 <lb/>0.15 <lb/>0.2 <lb/>0.25 <lb/>0.3 <lb/>0.35 <lb/>0.4 <lb/>-300-200-100 0 100 200 <lb/>dI/dV (mS) <lb/>Voltage (mV) <lb/>(c) <lb/>T=10 K <lb/>T=31 K <lb/>T=50 K <lb/>T=63 K <lb/>T=74 K <lb/>T=78 K <lb/>T=81 K <lb/>T=72 K <lb/>Fig. 1 -The temperature evolution of the tunneling conductance for overdoped Bi-2212 with bulk <lb/>Tc = 56 K, (a) junction # 1 and (b) junction # 2; (c) underdoped Bi-2212 with bulk Tc = 77 K [11]. <lb/>The tunneling conductances (except for 4.2 K) are shifted upward for clarity. Note that the scale at <lb/>higher temperatures is different for (b). <lb/>as in fig. 1(a), we do not observe the Josephson current. This discrepancy may be due to a <lb/>slightly different junction orientation or possibly due to the smaller energy gap which is esti-<lb/>mated from the 4.2 K data to be 10.5 meV. To the best of our knowledge fig. 1(b) shows the <lb/>smallest gap ever found in the Bi-2212 system. The energy gap magnitude found from tunnel-<lb/>ing is consistent with that observed in recent ARPES measurements around (π,0) on similar <lb/>crystals [16]. Using the bulk T c , the coupling ratio 2∆/kT c is 4.35, which is very close to the <lb/>BCS mean field d x 2 −y 2 -wave (d-wave) prediction of 4.28. The decrease in gap size and T c with <lb/>increasing hole doping in the overdoped phase is consistent with recent experiments [11,14,17] <lb/>and the phase diagram of HTSs. <lb/>As the temperature increases towards T c , the tunneling conductances of fig. 1(a) and <lb/>fig. 1(b) show several notable changes. Most importantly, all traces of a superconducting gap, <lb/>or any other type of gap, have disappeared for temperatures above the bulk T c . Additionally, <lb/>the magnitude of the Josephson current peak in fig. 1(a) diminishes until it vanishes around T c . <lb/>This is further evidence that the junction T c is the same as the bulk value. Both figures show <lb/>no presence of any depression above T c such as is observed in underdoped Bi-2212 [7,11]. Note <lb/>the increased sensitivity of the vertical scale in fig. 1(b) for the higher-temperature data. For <lb/>comparison, we show some previously published data on underdoped Bi-2212 [11] in fig. 1(c). <lb/>The superconducting gap is washed out around the bulk T c = 77 K of the crystal but a weak <lb/>depression in the conductance remains which is consistent with a pseudogap. The absence <lb/>of any pseudogap in our present measurements is consistent with recent ARPES studies [16] <lb/>on the same type of samples. In that study, the midpoint of the spectral weight peak shifted <lb/>to the Fermi level at T c , showing no evidence of the pseudogap that is seen in optimally <lb/>doped or underdoped Bi-2212 [1]. Thus, there are now two measurements which show that <lb/>the pseudogap disappears in heavily overdoped Bi-2212. <lb/>Considering other studies, STM measurements of overdoped Bi-2212 with T c ∼ 74 K have <lb/></body>
25
+
26
+ <page>592 <lb/></page>
27
+
28
+ <note place="headnote">EUROPHYSICS LETTERS <lb/></note>
29
+
30
+ <body>0.85 <lb/>0.9 <lb/>0.95 <lb/>1 <lb/>1.05 <lb/>Normalized Conductance <lb/>T=48 K <lb/>D =9 meV <lb/>G=13 meV <lb/>(b) <lb/>0.95 <lb/>0.96 <lb/>0.97 <lb/>0.98 <lb/>0.99 <lb/>1 <lb/>1.01 <lb/>-100 <lb/>-50 <lb/>0 <lb/>50 <lb/>100 <lb/>D =6 meV <lb/>G=18 meV <lb/>D =5 meV <lb/>G=20 meV <lb/>D =4 meV <lb/>G=18 meV <lb/>D =3 meV <lb/>G=20 meV <lb/>T=55 K <lb/>Normalized Conductance <lb/>Voltage (mV) <lb/>(c) <lb/>0 <lb/>0.5 <lb/>1 <lb/>1.5 <lb/>2 <lb/>2.5 <lb/>3 <lb/>-150 -100 -50 <lb/>0 <lb/>50 100 150 <lb/>Normalized Conductance <lb/>Voltage (mV) <lb/>D=14 meV <lb/>G=0.2 meV <lb/>a=0.4 <lb/>d-wave SIS at 4.2 K <lb/>(a) <lb/>Fig. 2 -(a) Normalized conductance of SIS junction (dots), which is given in fig. 1(a). The data <lb/>corresponds to 4.2 K and normalized by a constant. The solid line is a SIS-generated d-wave fit <lb/>which includes the smearing factor Γ and a directionality function f (θ). (b) and (c) Normalized <lb/>conductance of SIS junctions (dots) for 48 K and 55 K, which is given in fig. 1(a). The solid line in <lb/>(b) is SIS-generated pure d-wave fit which only includes the smearing factor Γ. The solid lines in (c) <lb/>show different attemps to fit the data. <lb/>reported the observation of a pseudogap above T c [7]. Since STM is a local probe, there is a <lb/>possibility that it does not reflect the bulk T c of the sample and therefore it is important to <lb/>consider the energy gap magnitude. Both refs. [7] and [8] present energy gaps in the range <lb/>35-40 meV in addition to the pseudogap behavior. This gap magnitude is consistent with <lb/>optimally doped or slightly overdoped samples and thus the doping level at the junction is not <lb/>well established. In a more recent study, SIS break junctions of underdoped and overdoped <lb/>Bi-2212 with T c = 82 K were fabricated using an STM by M. Oda et al. [18]. Both sets of <lb/>data exhibited a pseudogap above T c ; however, the overdoped tunneling conductance showed a <lb/>smaller gap which fills in at a much lower temperature than underdoped and thus T * is closer <lb/>to T c . Taken together the tunneling studies have established that the pseudogap persists with <lb/>overdoping down to T c = 82 K, but is absent when T c = 56 K. <lb/>A natural question might be whether there is any change in the pairing symmetry as-<lb/>sociated with the loss of a pseudogap. One may assume that increasing hole doping makes <lb/>the Bi-2212 normal-state more Fermi-liquid-like, leading possibly to an s-wave BCS super-<lb/>conductor. There are ARPES observations that suggest s-wave BCS behavior in the heavily <lb/>overdoped phase [19]. However, the tunneling subgap conductance shapes at low tempera-<lb/>tures in figs. 1(a) and (b) are not consistent with the s-wave BCS theory prediction which <lb/>is supposed to be flat near zero bias. Here what we observed is that the d-wave-like sub-<lb/>gap structure persists at the T c = 56 K overdoped Bi-2212. Figure 2 shows the normalized <lb/>tunneling conductance of SIS break junction # 1 at 4.2 K (dots) which is given in fig. 1(a). <lb/>Since the tunneling conductance is flat above T c , the data have been normalized by a con-<lb/>stant. The fit is generated using a d-wave gap function, ∆ = ∆ cos(2θ), in the BCS DOS, <lb/>N s (E, θ) = (E − iΓ)/ (E − iΓ) 2 − ∆(θ) 2 . Here, Γ is a smearing parameter to account for <lb/></body>
31
+
32
+ <note place="headnote">L. Ozyuzer et al.: Absence of pseudogap in heavily overdoped Bi-2212 <lb/></note>
33
+
34
+ <page>593 <lb/></page>
35
+
36
+ <body>quasiparticle lifetime. The SIS conductance is calculated by <lb/>dI <lb/>dV <lb/>= c <lb/>d <lb/>dV <lb/>f (θ)N s (E, θ)N s (E + eV, θ)[F (E) − F (E + eV )]dEdθ , <lb/>where c is a proportionality constant. F (E) is the Fermi function. The formula also includes a <lb/>directionality function f (θ)=1+α cos(4θ) which corresponds to a preferred tunneling along the <lb/>d-wave lobes. Here α is the directionality strength, which is assumed temperature independent <lb/>(α = 0.4) [15]. The fit (solid line in fig. 2(a)) shows good agreement in the subgap region <lb/>with the experimental data. Thus there appears to be no change of symmetry away from <lb/>d-wave. However for |eV | &gt; 2∆ the data deviate from the BCS d-wave fit, exhibiting broader <lb/>peak widths and the dip feature. In general, these features are compensating which allows <lb/>conservation of states; however, the data in fig. 2(a) lead to an integrated area out to 150 meV <lb/>which is about 6% above the BCS d-wave fit. We attribute the discrepancy to the inability to <lb/>accurately know the background conductance at 4.2 K. The broad peaks and dip feature are <lb/>observed throughout the doping range and have been attributed to strong-coupling effects [20]. <lb/>The weak temperature dependence of quasiparticle peak position has been observed in <lb/>the bare STM tunneling conductance on underdoped, optimally doped and overdoped Bi-<lb/>[7]. Analysis of the underdoped data has been carried out by Franz and Millis [21] using <lb/>the classical phase fluctuation model. What they find in their model fit to the tunneling <lb/>conductance of underdoped Bi-2212 data is that the gap is decreasing until 3/4 of T c , and <lb/>sharply increases near T c . This unusual effect seems to be tied to the presence of a pseudogap. <lb/>In the present study, we show that the gap magnitude for heavily overdoped Bi-2212 shows <lb/>no anomalies near T c but continues to decrease and appears to close near the bulk T c value. <lb/>Figures 2(b) and (c) shows the d-wave fit for T c = 48 K and 55 K. The tunneling conductance <lb/>in the fig. 2(b) displays peak values near 30 meV, however, the best fit for fig. 2(b) corresponds <lb/>to ∆ = 9 meV and Γ = 13 meV, so the energy gap is smaller than obtained value at 4.2 K. For <lb/>K, we plot fits using different ∆ and Γ values in fig. 2(c) to prove that the gap is very small <lb/>around T c . The data are in good agreement with a fit using ∆ = 4 meV and Γ = 18 meV. <lb/>These results show the energy gap magnitude reduces with increasing temperature as indicated <lb/>in fig. 3 and appears to close at the bulk T c . <lb/>0 <lb/>4 <lb/>8 <lb/>12 <lb/>16 <lb/>0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>1.2 <lb/>1.4 <lb/>0 <lb/>1 0 <lb/>2 0 <lb/>3 0 <lb/>4 0 <lb/>5 0 <lb/>6 0 <lb/>∆(Τ), Γ(Τ) (meV) <lb/>Normalized Josephson Strength <lb/>Temperature (K) <lb/>Γ(T) <lb/>∆(T) <lb/>Fig. 3 -Graph obtained from analyzing fig. 1(a) temperature dependence of superconducting gap <lb/>∆(T ) (squares), quasiparticle scattering rate Γ(T ) (circles), and Josephson strength, IcRn (diamonds) <lb/>normalized by IcRn (4.2 K). Here, the Josephson current Ic is estimated from the peak in conductance <lb/>at zero bias. The full curve represents the BCS superconducting gap ∆(T ). <lb/></body>
37
+
38
+ <page>594 <lb/></page>
39
+
40
+ <note place="headnote">EUROPHYSICS LETTERS <lb/></note>
41
+
42
+ <body>Fig. 4 -The temperature and 2∆ vs. hole concentration for Bi-2212. The dashed area is the measured <lb/>T * by different experimental techniques (see ref. [18]). The circles are our data that were published <lb/>before [11]. The diamond is the present work. <lb/>Figure 3 presents the various parameters derived from fig. 1(a). The solid line corresponds <lb/>to the temperature dependence of s-wave BCS superconductor energy gap which is similar to <lb/>the d-wave BCS [22] behavior. Filled squares are energy gap magnitude, filled circles are Γ. <lb/>The T -dependence of the gap magnitude follows the BCS prediction, although the increased <lb/>smearing due to the scattering rate Γ prevents exact determination of the gap magnitude <lb/>around T c . <lb/>As discussed before, one of the advantages of SIS tunnel junctions is the possible Josephson <lb/>current which can be used to determine T c of the superconductor. Even if there is a pairing <lb/>above T c , such as preformed pairs [5], any Josephson current that might persist above T c , <lb/>would be extremely weak. Therefore, we use the tunneling conductance peak (a more sensi-<lb/>tive measurement) to obtain the Josephson strength I c R n . In fig. 3, diamonds correspond to <lb/>normalized Josephson strength which vanishes around bulk T c of the crystal, the same tem-<lb/>perature where the quasiparticle gap vanished. These two measurements point to the absence <lb/>of superconducting pairing and Cooper pairs above the bulk T c of heavily overdoped Bi-2212. <lb/>We now combine in fig. 4 these results along with our previously published data and <lb/>the phase diagram for Bi-2212 as suggested in ref. [18]. Figure 4 shows that our energy <lb/>gap values lie within the experimentally determined region of T * over a wide doping range. <lb/>Here T * is obtained from various experimental techniques including SIS break junctions using <lb/>STM [18]. Extrapolating T * to the overdoped phase suggests that T c and T * are very close <lb/>and might even merge in the overdoped region. This might be the reason that we do not <lb/>observe a pseudogap in our tunneling studies. The implication of the doping dependence of <lb/>superconducting energy gap and T * is that they are intimately related. The simplest notion <lb/>is that T * represents the mean-field temperature for pairing, whereas T c defines when long-<lb/>range phase coherence is established. However, if the two temperatures were distinct over the <lb/>entire doping range, then that would suggest that T * corresponds to physics distinct from <lb/></body>
43
+
44
+ <note place="headnote">L. Ozyuzer et al.: Absence of pseudogap in heavily overdoped Bi-2212 <lb/></note>
45
+
46
+ <page>595 <lb/></page>
47
+
48
+ <body>superconductivity. Our present result suggests that there is a particular doping value where <lb/>the two temperature scales merge and this provides further support for pseudogap models <lb/>based on superconducting fluctuations [4, 7]. <lb/>In summary, we have performed break junction tunneling spectroscopy on heavily over-<lb/>doped Bi-2212. The tunneling conductance displays a flat background above T c without any <lb/>indication of pseudogap near the Fermi level. In addition, the energy gap magnitude reaches <lb/>a value as low as 2∆/kT c = 4.35, very close to the BCS value for d-wave superconductors. <lb/>The quasiparticle gap appears to close at the same temperature that the Josephson current <lb/>disappears. Taken together, the results suggest a merging of the temperature scales for pairing <lb/>and phase coherence. This occurs at a region of hole concentration where the strong-coupling <lb/>ratio approaches the BCS limit. This gives further support for the ideas that the pseudogap <lb/>associated with T * is due to some type of precursor superconductivity.<lb/></body>
49
+
50
+ * * * <lb/>
51
+
52
+ <div type="acknowledgement">This work was partially supported by US Department of Energy, Division of Basic Energy <lb/>Sciences-Material Sciences under Contract No. W-31-109-ENG-38. <lb/></div>
53
+
54
+ <listBibl>REFERENCES <lb/>[1] Ding H. et al., Nature (London), 382 (1996) 51; Loeser A. G. et al., Science, 273 (1996) 325. <lb/>[2] Loram J. et al., Phys. Rev. Lett., 71 (1993) 1740; Warren W. W. et al., Phys. Rev. Lett., 62 <lb/>(1989) 1193. <lb/>[3] Chen Q., Kosztin I., Janko B. and Levin K., Phys. Rev. Lett., 81 (1998) 4708. <lb/>[4] Lee P. A and Wen X.-G., Phys. Rev. Lett., 78 (1997) 4111. <lb/>[5] Emery V. and Kivelson S. A., Nature (London), 374 (1995) 434. <lb/>[6] Timusk T. and Statt B., Rep. Prog. Phys., 62 (1999) 61. <lb/>[7] Renner Ch., Revaz B., Genoud J.-Y., Kadowaki K. and Fisher O., Phys. Rev. Lett., 80 <lb/>(1998) 149. <lb/>[8] Matsuda A., Sugita S. and Watanabe T., Phys. Rev. B, 60 (1999) 1377. <lb/>[9] Tao H. J., Lu F. and Wolf E. L., Physica C, 282-287 (1997) 1507. <lb/>[10] Gupta, A. K. and Ng K.-W., Phys. Rev. B, 58 (1998) R8901. <lb/>[11] Miyakawa N., Zasadzinski J. F., Ozyuzer L., Guptasarma P., Hinks D. G., Kendziora <lb/>C. and Gray K. E., Phys. Rev. Lett., 83 (1999) 1018. <lb/>[12] Kendziora C., Kelley R. J., Skelton E. and Onellion M., Physica C, 257 (1996) 74. <lb/>[13] Ozyuzer L., Zasadzinski J. F. and Gray K. E., Cryogenics, 38 (1998) 911. <lb/>[14] DeWilde Y., Miyakawa N., Guptasarma P., Iavarone M., Ozyuzer L., Zasadzinski J. <lb/>F., Romano P., Hinks D. G., Kendziora C., Crabtree G. W. and Gray K. E., Phys. <lb/>Rev. Lett., 80 (1998) 153. <lb/>[15] Ozyuzer L., Zasadzinski J. F., Kendziora C. and Gray K. E., Phys. Rev. B, 61 (2000) <lb/>3629. <lb/>[16] Yusof Z., Wells B. O., Valla T., Fedorov A. V., Johnson P. D., Li Q., Kendziora C., <lb/>Jian S. and Hinks D. G., cond-mat/0104367. <lb/>[17] Renner C., Revaz B., Genoud J.-Y. and Fisher O., J. Low Temp. Phys., 105 (1996) 1083. <lb/>[18] Oda M., Dipasupil R. M., Momono N. and Ido M., submitted to Phys. Rev. Lett. <lb/>[19] Kelley R. J., Quitmann C., Onellion M., Berger H., Almeras P., and Margaritondo <lb/>G., Science, 271 (1996) 1255. <lb/>[20] Zasadzinski J. F., Ozyuzer L., Miyakawa N., Gray K. E., Hinks D. G. and Kendziora <lb/>C., Phys. Rev. Lett., 87 (2001) 067005. <lb/>[21] Franz M. and Millis A. J., Phys. Rev. B, 58 (1998) 14572. <lb/>[22] Won H. and Maki K., Phys. Rev. B, 49 (1994) 1397. </listBibl>
55
+
56
+
57
+ </text>
58
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/Faraday-discuss-post-ref.training.segmentation.tei.xml ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <page>1 <lb/></page>
8
+
9
+ <front>On the accuracy of intermolecular interactions <lb/>and charge transfer: the case of TTF-CA <lb/>Pilar García a , Slimane Dahaoui a , Claudine Katan b , Mohamed <lb/>Souhassou a and Claude Lecomte a † <lb/>High-resolution X-ray diffraction experiments and state of the art density <lb/>functional theory calculations have been performed. The validity of the <lb/>atoms in molecules approach is tested for the neutral-ionic transition of <lb/>TTF-CA which involves a transfer of less than one electron between the <lb/>donor and acceptor molecules. Foremost, crystallographical data have been <lb/>reassessed along the temperature-induced neutral-ionic phase transition <lb/>undergone by this charge transfer complex. Based on accurate X-ray <lb/>structures at 105 and 15 K, topological analysis of both DFT and the <lb/>experimental multipolar electron densities allowed detailed characterization <lb/>of intra-and interstack intermolecular interactions. Direct quantification of <lb/>the intermolecular charge transfer and the dipole moment are discussed. <lb/></front>
10
+
11
+ <body>1. Introduction <lb/>Molecular and hybrid organic-inorganic materials will gradually replace the harder <lb/>inorganic materials. The efficiency of some of these materials is linked to the <lb/>cooperativity concept which is widely used among the molecular materials <lb/>community. However, cooperativity is still not understood on physical and chemical <lb/>grounds; it needs thorough investigations of intermolecular interactions (hydrogen <lb/>bonds, van der Waals contacts, hydrophobicity, charge transfer…) which is a very <lb/>big challenge for both experimental and theoretical methods. This paper will <lb/>concentrate on the characterization of intermolecular contacs and on the estimation <lb/>of the molecular and atomic charges. Intermolecular interactions are usually <lb/>experimentally characterized by crystallographic or Raman-IR methods: <lb/>conventional crystallography gives interatomic distances which should be compared <lb/>to the sum of the so called &quot;ionic&quot; or &quot;van der Waals&quot; radii which are defined from <lb/>statistical analysis of experimental interatomic distances. IR or Raman methods <lb/>correlate intermolecular interactions to the variation of typical phonon or molecular <lb/>group modes but the assumption are also empirical and only qualitative. Therefore <lb/>one has to go a step further by modelling accurately the charge density in the <lb/>intermolecular regions: this will give a direct estimation of the molecular charges. <lb/>This can be achieved either by theoretical methods (Ab initio, DFT) or/and accurate <lb/>high resolution crystallography. 1,2 Once a model of the continuous charge density is <lb/>obtained, then a method has to be chosen for analysing the interactions and <lb/></body>
12
+
13
+ <front>a Laboratoire de Cristallographie et Modélisation des Matériaux Minéraux et Biologiques LCM 3 B <lb/>CNRS UMR 7036.Université Henri Poincaré, Nancy 1, Faculté des Sciences, BP 239, 54506 <lb/>Vandoeuvre-lès-Nancy Cedex (France). Fax: +33-(0)383 406492; Tel: +33-(0383 684863; E-mail: <lb/>[email protected] <lb/>b Synthèse et ElectroSynthèse Organiques, CNRS UMR6510, Université de Rennes 1, Campus de <lb/>Beaulieu, Bât. 10A,Case 1003, 35402 Rennes Cedex, France. Fax : +33-(0)223 236955 ; Tel: +33-<lb/>(0)223 235682. <lb/> † Author to whom correspondece should be adressed. E-mail:[email protected] <lb/></front>
14
+
15
+ <page>2 <lb/></page>
16
+
17
+ <body>accurately estimate the atomic group or molecular charges. This method has to be <lb/>used on experimental and theoretical charge densities. <lb/>The charge concept of an atom is not based on physically measurable quantities <lb/>because it is subjected to atom partitioning. In fact, it is impossible to determine <lb/>unambiguously the atomic boundary inside which an atomic charge is located: <lb/>atomic charges depend on the form and the size of the integration volume and on the <lb/>atomic orbitals used (Ab Initio Methods). Hirshfeld has proposed in 1977 3 the <lb/>stockholder scheme using an atomic weighting function <lb/>( ) <lb/>( ) <lb/>( ) <lb/>r <lb/>r <lb/>r <lb/>w <lb/>sph <lb/>pro <lb/>sph <lb/>j <lb/>j <lb/>ρ <lb/>ρ <lb/>= <lb/>v <lb/>(1) <lb/>where ρ sph is the electron density of the independent atom ( <lb/>sph <lb/>j <lb/>ρ ) or of the <lb/>promolecule ( <lb/>sph <lb/>pro <lb/>ρ ). When multiplied by the modelled density ( ) <lb/>r <lb/>r <lb/>ρ <lb/>the function <lb/>gives an estimation of the charge density of the j th atom of the unit cell <lb/>( ) ( ) ( ) <lb/>r <lb/>w <lb/>r <lb/>r <lb/>j <lb/>j <lb/>r <lb/>r <lb/>r <lb/>. <lb/>ρ <lb/>ρ <lb/>= <lb/>(2) <lb/>The limiting atomic volume is therefore given by a threshold of w j . Such atoms have <lb/>therefore some overlap. Coppens 4 has also proposed a method based on the Wigner-<lb/>Seitz method which defines an atomic polyhedron which boundaries are defined by <lb/>( <lb/>) <lb/>( <lb/>) <lb/>B <lb/>AB <lb/>B <lb/>A <lb/>AB <lb/>A <lb/>R <lb/>R <lb/>r <lb/>r <lb/>R <lb/>R <lb/>r <lb/>r <lb/>r <lb/>r <lb/>r <lb/>r <lb/>r <lb/>r <lb/>− <lb/>= <lb/>− <lb/>(3) <lb/>R AB being the internuclear distance and R A and R B are &quot;a priori&quot; atomic radii. This <lb/>method which was used to estimate the charge transfer in the TTF-TCNQ material <lb/>relies on the van der Waals radii. In conclusion all these partitioning applied to both <lb/>experimental and theoretical models are based on &quot;a priori&quot; assumptions. On the <lb/>other hand, Mulliken partitioning 5 is only used with theoretical electron density <lb/>whereas the κ-Pv refinement method 6 is widely used by X-ray diffractionists only : it <lb/>assumes an expanded/contracted spherical pseudoatom density <lb/>( ) <lb/>) <lb/>( <lb/>) <lb/>( <lb/>3 <lb/>r <lb/>P <lb/>r <lb/>r <lb/>j <lb/>val <lb/>j <lb/>j <lb/>val <lb/>j <lb/>core <lb/>j <lb/>j <lb/>κ <lb/>ρ <lb/>κ <lb/>ρ <lb/>ρ <lb/>+ <lb/>= <lb/>(4) <lb/>where ρ core and ρ val are HF electron densities of the free atoms. Other methods to <lb/>estimate charges are potential fitted charges 7,8 . We have shown that these fitted <lb/>charges are very close to those obtained directly by κ-Pv refinement. <lb/>Bader 9 proposed a topological analysis of the total electron density based on <lb/>quantum mechanics: the whole molecular system in the crystal is represented by a <lb/>sum of non overlapping atomic basins: their boundary surface is limited by zero flux <lb/>surfaces <lb/>0 <lb/>) <lb/>( <lb/>). <lb/>( <lb/>= <lb/>∇ <lb/>r <lb/>n <lb/>r <lb/>r <lb/>r <lb/>ρ <lb/>(5) <lb/>Integration of charge, dipole and multipole moments inside the atomic basins can be <lb/>performed. This last method can be applied to both experimental and theoretical data <lb/></body>
18
+
19
+ <page>3 <lb/></page>
20
+
21
+ <body>and has the advantage to characterise intermolecular interactions 10 by bond paths <lb/>and charge density at the critical points ( <lb/>0 <lb/>) <lb/>( = <lb/>∇ r <lb/>r <lb/>r <lb/>ρ <lb/>). <lb/>The present discussion proposes to address the accuracy of such an approach on the <lb/>basis of up to date experimental and theoretical results obtained for a model <lb/>compound. The tetrathiafulvalene-p-chloranil (TTF-CA; C 6 S 4 H 4 ·C 6 Cl 4 O 2 ) was <lb/>chosen because it shows small charge transfer (less than 1e) via intermolecular <lb/>interactions that depends on various thermodynamic parameters (T, P, hν). The TTF-<lb/>CA complex is built from donor (D=TTF) and acceptor (A=CA) -D-A-D-A-D-A <lb/>stacks and undergoes a first-order neutral-ionic phase transition (NIT) under <lb/>temperature variation at a critical temperature of about 80K 11,12 as well as under <lb/>pressure variation at a critical pressure of about 11kbar. 12,13,14 N to I or I to N <lb/>conversion can also be induced by photoirradiation as was shown by Koshihara et <lb/>al. 15 allowing for time-resolved crystallography of such a transformation at the <lb/>picosecond time scale. 16 <lb/>According to vibrational spectroscopy and absorption spectra experiments 17 , the <lb/>charge transfer (CT) qCT has been estimated to increase from 0.3e in the N phase to <lb/>0.7 in the I phase. It has been shown that this NIT comes along with a structural <lb/>phase transition characterized by the loss of inversion symmetry (P21/n to Pn space <lb/>group ) and a dimerization along the stacking axis. 18 <lb/>Before getting onto the main purpose of this work, crystallographical data need to <lb/>be reassessed along the temperature-induced neutral-ionic phase transition <lb/>undergone by TTF-CA, to ensure appropriate starting point. Based on the <lb/>outcoming X-ray structures obtained at 105 and 15 K, intra-and intermolecular <lb/>interactions are discussed on the ground of the topology of both experimental and <lb/>theoretical total electron densities . Corresponding D (TTF) and A (CA) single <lb/>crystals have also been investigated to provide benchmarks. Quantification of the <lb/>intermolecular interaction, charge transfer and the dipole moments through the <lb/>topological analysis of the electron density will be discussed. <lb/>2. Experimental section <lb/>2.1 Experimental details <lb/>TTF and CA powders were purchased from Lancaster and purified. Single crystals of TTF were <lb/>obtained by sublimation method. Chloranil crystals were obtained in acetonitrile solution. After a <lb/>few days of slow evaporation, prismatic yellow single crystals appear in the solution. Single crystals <lb/>of the TTF-CA complex were grown by sublimation of the component materials in a vacuum-sealed <lb/>Pyrex glass tube, which was placed at 79 °C in an electric furnace for several days. Dark prismatic-<lb/>like crystals were obtained. <lb/>The temperature evolution of the unit cell parameters was studied using a home-made helium bath <lb/>cryostat 19 specially designed for a four-circle diffractometer equipped with a 2D detector. The <lb/>sample is mounted on a two-circle goniostat placed in the Helium tail of the cryostat. This goniostat <lb/>is magnetically coupled to a master magnet fit on the arm of the diffractometer which allows the <lb/>orientation of the sample inside the cryostat. The cell parameters were determined by the analysis of <lb/>a set of 84 and 42 images, when cooling and warming, respectively, repeated from 110K to 50K, <lb/>with an accuracy of 0.02K. <lb/></body>
22
+
23
+ <page>4 <lb/></page>
24
+
25
+ <body>For the structures and charge density determinations, three-dimensional diffraction data were <lb/>collected on Enraf-Nonius Kappa and Oxford-Xcalibur-Sapphire2 CCD-based diffractometers with <lb/>MoKα radiation and a nominal crystal-to-detector distance of 40 mm. The crystal was cooled from <lb/>room temperature to 105K with an Oxford Cryosystems N2 open flow cryostat 20 and to 15K with an <lb/>Oxford Diffraction Helijet open-flow He gas cryosystem. <lb/>Integration of frames and data reduction were performed with DENZO 21 and CrysAlis Red <lb/>programs. The multiple integrated reflections were averaged using SORTAV 23 adapted to area <lb/>detector data. Internal agreement factors for all data are given in Table 1. Only reflections having I ≥ <lb/>3σ(I) were used in the structure and electron density least-squares refinements. <lb/>At 105K, TTF-CA crystallizes in the monoclinic system, space group P21/n (systematic absences: <lb/>h0l, h+l ≠ 2n; 0k0, k ≠ 2n). At 15K, diffraction intensities of (0k0) with k=2n+1 do exist, while <lb/>reflections (h0l) with h+l=2n+1 remain systematically absent in agreement with the loss of the 21 <lb/>screw axis, the space group of the ionic phase is then Pn. In order to compare both structures, the <lb/>coordinates x and z of the Cl19 atom were fixed at high temperature (HT) structure values and only <lb/>the y coordinate was refined for the low temperature (LT) data. Sorted data were merged in the m <lb/>point group, and anomalous dispersion corrections were performed as described in Souhassou et al. 24 <lb/>Experimental details are reported in Table 1. <lb/>Molecular structures of TTF-CA were solved by direct methods and subsequent Fourier analysis, <lb/>then, refined, with a full-matrix least-squares refinement (independent atom model IAM, SHELXL-<lb/>97). 25 Hydrogen atoms were first constrained to positions contained from neutron diffraction <lb/>experiments. 18 The final refined parameters were imported into the MOLLY program to model the <lb/>aspherical electron density. 26 After convergence of the multipolar refinement, H atoms were allowed <lb/>to relax as their aspherical electron density features are taken into account in the multipolar <lb/>refinement, leading to unbiased H positions. Recently we have shown that this method gives best C-<lb/>H bond distances close within 0.03 A to distances usually found from neutron diffraction. 27 <lb/>It is very important to note that, although the molecules do not lie anymore on inversion centres in <lb/>the LT phase, molecular geometries are only slightly deformed, remaining almost centrosymmetric <lb/>as will be discussed later. This feature induces strong correlations between charge density <lb/>parameters of pseudo symmetry related atoms. In order to avoid these correlations, the following <lb/>approach was carried out: first all atoms positions and thermal parameters were refined together <lb/>without any constraints or restraints until convergence, and then the multipole parameters (all κ, Pv <lb/>and Plm) of atoms related by an inversion centre in the N phase were constrained to be equal. This <lb/>procedure leads to excellent statistical indexes as R factors are smaller than 1.3%. The maxima and <lb/>the minima in the residual density maps do not exceed 0.10e Ǻ -3 . The estimate of the average error <lb/>in the experimental electron density maps is 0.05 e.Ǻ -3 , as calculated by 28,29 <lb/>( <lb/>) <lb/>2 <lb/>/ <lb/>1 <lb/>2 <lb/>1 <lb/>2 <lb/>/ <lb/>1 <lb/>2 <lb/>2 <lb/> <lb/> <lb/> <lb/> <lb/> <lb/> <lb/>− <lb/>= <lb/>∑ − <lb/>H <lb/>m <lb/>obs <lb/>res <lb/>F <lb/>F <lb/>K <lb/>V <lb/>σ <lb/>(6) <lb/>This error is largely smaller in the intermolecular region. <lb/>Aspherical modelling of the experimental X-ray structure factors provides the total electron density <lb/>distribution ρ(r) which is the main observable in the &quot;Quantum Theory of Atoms in Molecules&quot; <lb/>developed by Bader. 30 The topology of the total charge density allows a unique partition of the <lb/>density into atomic contributions, by defining atomic basin as the volume limited by zero electron <lb/>density gradient flux. Topological properties at critical points (CP), as electron density, ρ(rCP); <lb/>Laplacian, ∇ 2 ρ(rCP); electron density curvatures, λ1 , λ2 , λ3 ; local kinetic G(rCP) and potential <lb/>V(rCP), energy densities can then be used to describe the molecular structure and to characterize <lb/>bonding and intermolecular interactions. Topological analysis of experimental electron density was <lb/>performed analytically using the NEWPROP program. 31 <lb/>2.2 Computational details <lb/></body>
26
+
27
+ <page>5 <lb/></page>
28
+
29
+ <body>Electronic structure calculations were performed in the framework of density functional theory <lb/>(DFT) using the local density approximation (LDA) parametrization by Perdew and Zunger 32 , <lb/>Becke&apos;s gradient correction to exchange energy 33 and Perdew&apos;s gradient correction to correlation <lb/>energy. 34 Calculations were carried out with the projector-augmented wave (PAW) method 35 which <lb/>uses augmented plane waves to describe the full wave functions and densities without shape <lb/>approximation. The core electrons were described within the frozen core approximation. The plane <lb/>wave cutoff for wave functions and densities were fixed respectively to 90 and 360 Ry to ensure <lb/>required precision for topological analysis. Atomic coordinates were obtained from experimental X-<lb/>ray data obtained after complete multipolar refinement. CA and TTF isolated molecule calculations <lb/>were conducted using tetragonal supercells (12.5 x 12.5 x 7.5 Å for TTF and 12 x 12 x 7 Å for CA) <lb/>including electrostatic decoupling from periodic images. 36 For TTF-CA, crystal calculations were <lb/>performed with three k-points along a* . 37 Topological analysis of theoretical densities was achieved <lb/>with the InteGriTy software package using a grid spacing smaller than 0.044 Å. 38 <lb/>Compared to indirect methods, previous DFT investigation of CT complexes found that charge <lb/>transfer (qCT) is overestimated in the HT phase 37, 39 , certainly due to the approximation used for the <lb/>exchange and correlation potential that suffers from the well-kown problem of the self-interaction <lb/>error and to a much lesser extent from the missing of the dynamical part of the electron-electron <lb/>interactions in any local or semi-local approaches. One of the major effects of the too rapid decay <lb/>of the LDA (or even GGA) exchange-correlation potential is the strong charge delocalization, that <lb/>leads to a systematic overestimation of qCT in the HT phase for instance. 40 <lb/>3. Results and discussion <lb/>3.1 Thermal behaviour of the cell parameters <lb/>The evolution of the cell parameters as a function of temperature gives a clear evidence of the phase <lb/>transition at Tc = 81K: a and c cell parameters display a slope anomaly, while b presents an abrupt <lb/>jump; in all cases the contraction rate decreases below Tc. 41 Fig. 1a shows that on cooling β <lb/>decreases from 99.06(2)° to 98.68(2)° between 82 and 81K. However, when warming the increase <lb/>of the β angle occurs between 82 and 83.5K. This hysteresis of about 1K is also observed on the b <lb/>parameter (Fig. 1b) and confirms the observed evolution of the intensity of the (0 3 0) reflection. 41 <lb/>Thus all the cell parameters are sensible to the phase transition, reflecting the 3D character of the <lb/>temperature induced transition which will be studied using AIM method. <lb/>3.2 Structural properties: Phase transition <lb/>The molecular structures of TTF-CA at 105 and 15K with atom labelling are reported in Fig. 2. In <lb/>the high temperature phase (HT, T&gt;Tc) the TTF-CA space group is P21/n (monoclinic), and the TTF <lb/>and CA molecules form mixed stacks along a axis (Fig. 2a). The TTF and CA molecules lie both on <lb/>inversion centres respectively at (0 1/2 0) and (1/2 1/2 0) and all atoms are in general <lb/>crystallographic positions. Since D and A lie on inversion centres, the asymmetric unit is made of <lb/>half a TTF molecule and half a CA molecule. Along the stack, the TTF axis, corresponding to the <lb/>central C═C bond, forms an angle of 105.9(5)° with the CA O···O axis. As mentioned by Mayerle et <lb/>al. 42 and in agreement with the general trend observed in TTF-XA (X=F, Cl, Br) 42,43 , the best <lb/>overlap conditions between the highest occupied molecular orbital (HOMO) of D and the lowest <lb/>unoccupied molecular orbital (LUMO) of A are fulfilled when the angle between the O···O axis of <lb/>the XA molecule is almost orthogonal to that of the TTF molecule. An edge-on view of the <lb/>molecules shows several features of interest regarding the molecular configurations and packing. <lb/>The distance between the TTF and CA molecular centres in the stack is about half the a parameter <lb/>(i.e. 3.615(2) Å). However, the mean planes of the TTF and CA molecules are not parallel. The <lb/>dihedral angle is 1.6(5)°. This angle induces a close contact along the stack between a sulphur atom <lb/>S4 and the C15-C17 bond, which has an important contribution to the valence band. 37 This short <lb/>contact (i.e. 3.325(2) Å) favours the intermolecular interactions needed for the CT. <lb/>In the low temperature phase (LT, T&lt; Tc), the crystal space group of TTF-CA becomes Pn. While <lb/>keeping the same lattice translational symmetry, the organic stack gets dimerized. The asymmetric <lb/>unit is made of one TTF and one CA molecules, whose centroids are shifted from their HT phase <lb/>positions, mainly along a axis, e.g., the stacking axis: -0.08 Ǻ and 0.1 Ǻ for TTF and CA <lb/></body>
30
+
31
+ <page>6 <lb/></page>
32
+
33
+ <body>respectively (Fig. 2b). TTF and CA form pairs of D and A characterized by different intra-(dintra) <lb/>and inter-(dinter) pair distances, defined as the distances between the molecular centres (dintra= <lb/>3.416(4) Ǻ and dinter=3.785(4) Ǻ). However, as discussed in the dimerization process of <lb/>tetrathiafulvalene-p-bromanil (TTF-BA; C6H4S4· C6Br4O2) 43 , it is more relevant to compare the <lb/>closest intermolecular contacts: the shortest atom to atom contact in the stack connects a carbon <lb/>atom of the central C=C bond of TTF to a CA carbon bonded to an oxygen. Due to the dimerization, <lb/>two different distances are observed: an intradimer C9···C21 distance (3.136(4) Ǻ) and an interdimer <lb/>C2···C15 (3.531 (4) Ǻ), while in the HT phase both distances are equivalent by crystal symmetry <lb/>(3.370 Ǻ). Other close contacts originate from the overlap between CA&apos;s π electrons and the pz <lb/>orbital of one of TTF&apos;s sulphur atoms, defining S···C-C contacts. Two non equivalent contacts also <lb/>appear: the intrapair S11···C23-C15 is reinforced and the corresponding distance reduced to 3.158(4) <lb/>Ǻ, while the interpair S4···C21-C17 distance increases to 3.503(4) Ǻ. However, relative orientation <lb/>of D and A along the stack are comparable to that of the HT phase, the central C=C bond of TTF <lb/>forms an angle of 105.7(1)° with the O···O axes of CA, and the dihedral angle between the mean <lb/>planes of D and A is 1.1(1)°. <lb/>Intramolecular geometries of TTF and CA slightly change when cooling below Tc. In this way, the <lb/>multipolar refinement shows an increase of the TTF C=C bond lengths. This effect is larger for the <lb/>central C=C bond, from 1.3678(7) Ǻ at 105K to 1.3862(7) Ǻ at 15K. An inverse effect at the limit of <lb/>the experimental accuracy occurs for the C-S bonds, with distances decreasing from 1.7494(4) and <lb/>1.7500(4) Ǻ at 105K to 1.7429(8), 1.7440(9), 1.7439(9) and 1.7499(9) Ǻ at 15K. This evolution of <lb/>the C=C and C-S bonds is related to the change of qCT in the complex. The HOMO of TTF is π-<lb/>antibonding with respect to the C-S bonds and π-bonding with respect to the C=C bonds, thus, when <lb/>qCT increases, C-S bond lengths decrease, while C=C increase. 44 <lb/>In the CA molecule all bond lengths are modified due to the phase transition, i.e. by the increase of <lb/>qCT. The single C-C bond lengths decrease, while C=C double bonds increase. However the effect is <lb/>larger for C=O and C-Cl bond lengths. The C=O double bonds increase from 1.2227(4) Ǻ at 105K <lb/>to 1.237(1) Ǻ for both C=O bonds at 15K. The C-Cl bond lengths also increase from 1.7062(4) and <lb/>1.7037(4) to 1.7234(9), 1.7190(9), 1.7176(9) and 1.7162(9) Ǻ. Hence, as the LUMO of CA is π <lb/>antibonding for both C=O and C-Cl, these bonds increase when qCT becomes higher. <lb/>Finally, the Uij anisotropic thermal parameters at 15K are roughly half than those at 105K. This <lb/>usual behaviour, also observed in TTF-BA 43 , largely contrasts with the trends observed in other CT <lb/>complexes, such as DMTTF-CA, where some anisotropic temperature factors along the stacking <lb/>axis increase below 65K, suggesting an order-disorder type of phase transition. 45 <lb/>We can conclude that the observed geometrical changes of TTF and CA are due to the increase of <lb/>the qCT between D and A in the complex. But, contrary to the neutron LT structure 18 , where the <lb/>bond lengths in half a TTF molecule are reported to be 1-2% longer than in the other half, this work <lb/>shows that the NIT does not induce such large differences. In fact, in the LT phase both TTF and <lb/>CA moieties remain almost centrosymmetric. Our results are in agreement with recent synchrotron <lb/>results 46 that show bond lengths differences of the order of the estimated errors (0.004 A). <lb/>Therefore, such small differences cannot characterize the phase transition. One reason for the <lb/>discrepancies between neutron and X-Ray structures arises from the small observation/parameter <lb/>ratio that allowed only for isotropic estimation of Debye Waller factors in the neutron case whereas <lb/>a full multipolar refinement was performed on X-ray data. <lb/>3.3 Charge transfer and volumes <lb/>CT is at the origin of the very interesting electronic, magnetic and optical properties of the CT <lb/>complexes, and therefore its accurate determination is essential to understand, classify and control <lb/>them. However, as discussed on the introductory paragraph, it is not a well-defined quantity and <lb/>there are several experimental and theoretical methods to calculate it. For TTF-CA, it has been <lb/>estimated from the CT absorption band 47 , from vibrational spectroscopy 17,48 and by empirical <lb/>relations between qCT and D or A geometry. 49 All these methods lead per se to indirect estimations <lb/>of qCT. In order to avoid both transferability problems and at the same time to allow for more direct <lb/>determinations of qCT, these determinations should be obtained from intrinsic properties, e.g., the <lb/>electronic structure of the CT complex. These has already been achieved for a couple of NIT <lb/></body>
34
+
35
+ <page>7 <lb/></page>
36
+
37
+ <body>compounds on the basis of DFT calculations 37,39 , in particular using &quot;Quantum Theory of Atoms in <lb/>Molecules&quot;. 30 For TTF-CA, integration over atomic basins leads to qCT of 0.48 and 0.64 e <lb/>respectively at 300 and 40K. Clearly, qCT is overestimated in the HT phase due to the approximation <lb/>used for the exchange and correlation (see section 2.2). <lb/>As discussed above and observed in several TTF-based donor molecules 50 , oxidation of donor <lb/>molecules produces variations in the central C=C and C-S bond lengths. First principle studies of <lb/>TTF 44 and CA 51 isolated molecules in different oxidation states have established the linear effects <lb/>of molecular ionicity on bond lengths and several vibrational frequencies. In particular, C=O and C-<lb/>Cl bonds of chloranil and central C=C and C-S bonds in TTF are shown to be particularly sensible to <lb/>molecular ionicity. These dependences have been related to the contributions of the bonds to the <lb/>LUMO of CA and HOMO of TTF. However, it has been stressed that these linear relationships may <lb/>fail due to intermolecular interactions. Moreover, concerning frequency, electron-intramolecular-<lb/>vibration interactions should not be disregarded. 52 Examination of the Cambridge Structural Data <lb/>Base has allowed Umland and co-workers 49 to establish linear relationships between qCT and several <lb/>functions (ratio and difference) of these C=C and C-S TTF bonds. This approach was used to <lb/>estimate qCT in TTF-CA 18 from experimental bond lengths: values of 0.4 and 0.8 e were deduced at <lb/>300 and 40K respectively. <lb/>Starting from our more accurate structural data, the estimates are qCT =0.40 e at 105K and qCT =0.65 <lb/>e at 15K. However, if we consider the values of the bond lengths of TTF and CA reported for <lb/>different oxidation states 44,51 , charge transfer variations (ΔqCT) from 105K to 15K of 0.4 and 0.5 e <lb/>are obtained using C=O and C-Cl bond lengths respectively. From the TTF C=C and C-S bonds, <lb/>ΔqCT of 0.5 and 0.15e are obtained respectively. When using bond lengths obtained for crystals of <lb/>pure TTF or CA as reference (qCT = 0) and the slopes determined by DFT calculations, 44,51 results <lb/>become worse. Indeed, C-Cl bonds lead to qCT ≈ -0.1 and qCT ≈ 0.4 e, C-S to qCT ≈ 0.8 and qCT ≈ 0.9 <lb/>e whereas the central C=C bond of TTF to qCT ≈ 0.7 and qCT ≈ 1.1 e, respectively at 105 and 15K. <lb/>Thus, using such approaches to determine qCT is clearly hazardous, and it gives only qualitative <lb/>trends for ΔqCT. From all these linear relationships ΔqCT ranges from 0.2 to 0.5e at the NIT. <lb/>Therefore a direct estimation is needed, based on intrinsic properties, like the electronic structure of <lb/>the complex. Experimentally, the charge transfer can be obtained from X-ray diffraction data in two <lb/>ways. First, a refinement in reciprocal space with a modified spherical valence shell (κ-Pv formalism <lb/>6 ), according to equation 4, allows an estimation of the net atomic charges. A similar method was <lb/>proposed by Espinosa et al. 53 to estimate the charge transfer in bis(thiodimethylene)-<lb/>tetrathiafulvalene tetracyanoquinodimethane (BTDMTTF-TCNQ, C10H8S6 + ·C12H4N4 -) using a Plm <lb/>multipolar refinement , the atomic net charges can be obtained as: <lb/>val <lb/>j <lb/>j <lb/>j <lb/>P <lb/>N <lb/>q <lb/>− <lb/>= <lb/>(7) <lb/>The obtained charge transfer for BTDMTTF-TCNQ (0.75e) has to be compared with the value <lb/>derived from the diffuse scattering experiments (0.5e). 54 <lb/>Both κ-Pv and Plm refinements have been performed in TTF-CA using electroneutrality constraint, <lb/>i.e. the total charge of the unit cell is zero. Thus, these two approaches lead respectively to charge <lb/>transfers of of 0.14±0.05 and 0.06±0.05e at 105K, while the values increase to 0.67±0.03 and <lb/>0.65±0.03 e at 15K. There are not significant differences between the results obtained which show <lb/>the consistency of the X-ray method. <lb/>Finally, the topology of the charge density provides another tool to determine the charge transfer, as <lb/>the sum of the atomic topological charges obtained by integration of the experimental/theoretical <lb/>charge density over the atomic basins of all atoms of each molecule leads to the molecular charge <lb/>and thus to qCT. These molecular charges and basin volumes obtained are listed in Table 2 and are <lb/>compared to those obtained for the CA and TTF single crystals. Within TTF-CA at 105K, the <lb/>experimental charges of CA and TTF are respectively -0.22(2) and 0.20(2) e, increasing to -0.71(2) <lb/>and +0.77(2) e at 15K. Consequently, the experimental qCT is 0.21 ± 0.02 at HT and 0.74 ± 0.02 at <lb/>LT, and ΔqCT =0.53 e. The same approach using the theoretical charge density leads to qCT ≈ 0.64e <lb/>for the LT phase. How can we gauge our approach? First, our values are in pretty good agreement <lb/>with the κ-Pv and Plm derived charges. Secondly, considering that the unit cell contains two TTF and <lb/>two CA molecules, a comparison between the volume filled by the atomic basins and the unit cell <lb/>volume allows for gauging the accuracy of the methods. Summation of the atomic volumes (Table <lb/>2) reproduces the cell volume within less than 0.3%, for both temperatures at experimental and <lb/>theoretical levels. This nicely demonstrates the ability of such approaches to tackle such small <lb/>changes, especially regarding atomic quantities. Correlatively, the net integrated charge of the unit <lb/>cell is also very close to zero. <lb/>As the temperature decreases from 105K to 15K, the TTF volume decreases from ≈ 1.5-2.5 %, while <lb/>the CA&apos;s volume remains almost constant. In fact, as CT increases, TTF looses electron density in <lb/>favour of CA and volume reduction follows. Moreover, from experiments on TTF-based CT <lb/>complexes 55 and ab-initio calculations 44 it is well known that the geometry of TTF is responsive to <lb/>environmental effects and phase transitions. On the other hand, the CA electron density increases <lb/>with CT while lattice parameters shrink, thus maintaining constant the CA molecular volume. <lb/>Additionally, CA peripheral atoms contain a large number of core electrons that prevents large <lb/>volume variations compared to TTF. This is particularly clear when volumes obtained within a TTF-<lb/>CA crystal are compared to those obtained on pure TTF and CA single crystals (Table 2). <lb/>3.4 Dipole moment <lb/>The calculation in the crystal axis system of the dipole moment for the dimer in the ionic phase can <lb/>be performed, considering both contributions: the net atomic charges and the atomic dipoles <lb/>obtained once the topological analysis has been completed. Obviously, the net charge contribution is <lb/>mainly along the stacking axis a: (-12.585, -0.086, 0.155) D. But the contribution of the atomic <lb/>dipole moments, (16.293, -7.027, 0.457) D, is large, almost lying in the (a,b) plane reflecting not <lb/>only the π-π interactions between D and A (along a), but also interstack interactions that are <lb/>dominated by hydrogen bonds as will be discussed in section 3.6.2. <lb/>Following dimerization, a ferroelectric dipole moment Δ µ= µdimer -µinterdimer appears. Values of 0.36 <lb/>D in the stack and 0.72 D in the unit cell are found. The corresponding dipolar energy, <lb/>Wd= ( Δ µ) 2 / 4πε0V, 56 where V is the volume occupied by the dipole moment, is 6.63x10 -21 J. This <lb/>small value compared to kBTN-I (Wd / kB=4.8K) shows that long-range dipolar interactions are not <lb/>the driving force for the NIT, in agreement with previous findings. 18 <lb/>3.5 Covalent bonds <lb/>A quantitative comparison of the covalent bonds strength can be derived from their topological <lb/>properties. For each type of covalent bond, similar features are found irrespective of molecular <lb/>environment, e.g. in TTF-CA, TTF or CA crystals or for calculated isolated molecules. <lb/>Accumulation of electron density is particularly visible in double bonds, the highest density being <lb/>observed within the C=O bond. The Laplacian of the density also reveals the oxygen lone pairs and <lb/>the non bonding valence shells of both chlorine and sulphur atoms. Interestingly, while other <lb/>covalent bonds show almost equal λ1 and λ2 curvatures at CPs, indicative of cylindrical symmetry <lb/>perpendicularly to the bonds, the λ1 and λ2 values obtained for all C=C bonds differ significantly <lb/>(about 30%). <lb/>The NIT affects clearly all bond critical points as exemplified by the C2=C9 bond, for which the <lb/>distance increases from 1.3678(7) to 1.3862(7) Å, the density decreases from 2.16 to 2.05 eǺ -3 and <lb/>∇ 2 ρ(rCP) changes from 21 eǺ -5 to 18 eǺ -5 when decreasing the temperature from 105 to 15K. <lb/>Again, this is directly related to the bonding character of the C=C bonds in the HOMO of TTF. Both <lb/>C-S and C-Cl bonds are less responsive to CT variation. For C-S, the average experimental <lb/>∇ 2 ρ(rCP) and ρ(rCP) vary from -7.2 eǺ -5 , 1.33 eǺ -3 at 105K to -6.2 eǺ -5 , 1.31 eǺ -3 at 15K. These <lb/>values are in good agreement with the literature. 57 On average, the C-Cl bond CPs lie at 0.76 Å and <lb/>0.96 Å from the C atom and the Cl atom respectively. This is directly related to their respective <lb/>atomic volumes. Moreover, the valence shell charge concentration of the Cl atom is strongly <lb/>polarized toward the nearby C atom (bonded charge concentration). In the direction of such a polar <lb/>bond, there is a lack of charge accumulation near the more electronegative atom. This behaviour is <lb/>also characteristic of C-O and C-F bonds. 58 During the NIT, C-Cl bonds lengths increase from <lb/>1.7045(4) to 1.7191(9) Å in average. ∇ 2 ρ(rCP) and ρ(rCP) change from -4.8 eǺ -5 , 1.40 eǺ -3 at HT to <lb/>-4.1 eǺ -5 , 1.32 eǺ -3 at LT. <lb/>The overall agreement between theoretical and experimental results is quite good for all bond CP <lb/>characteristics. Among the differences, theory gives systematically slightly higher ρ(rCP) than that <lb/>deduced from experiment and lower curvatures along the bond path (λ3) for all bonds except C=O. <lb/>The picture becomes more tricky for C=O bonds: calculated values of ∇ 2 ρ(rCP) show pronounced <lb/>deviation from experimental ones, especially for λ3: experimental/theoretical values of ∇ 2 ρ(rCP) are <lb/></body>
38
+
39
+ <page>9 <lb/></page>
40
+
41
+ <body>-28/3 eǺ -5 at 105K and -19/-1 in average at 15K. Such a discrepancy between experiment and theory <lb/>has already been reported in the literature, and explained as originated by the different nature of the <lb/>radial functions used in DFT and multipolar refinements. 59 It should also be underlined that, <lb/>because of the fast variation of the density along the C=O bond path, computations of λ3 and <lb/>∇ 2 ρ(rCP) are not straightforward, especially when working with densities on grids. 38 However, <lb/>experimental and theoretical ρ(rCP) are in excellent agreement: 2.91/2.71 at HT and 2.72/2.63 eǺ -3 at <lb/>LT. <lb/>3.6 Intermolecular Interactions <lb/>Up to now, in the TTF-CA family, crystallographic investigations have characterized the <lb/>intermolecular interactions only by distances between interacting atoms. Accurate electron density <lb/>studies may go a step further, using topological analysis to describe H bonds and van der Waals <lb/>interactions by energetical quantities at the intermolecular CPs associated with the interaction. For <lb/>example, Espinosa et al. 10, 60 have shown exponential relations between the charge density at CP and <lb/>the potential and kinetic energy densities at CP. Such analysis needs very accurate data and a <lb/>software able to calculate properties at the CP (where the density seldom reaches 0.1 eǺ -3 ) with a <lb/>good precision. Moreover, the corresponding gradient change is very small (flat density) makes the <lb/>determination of the bond path difficult. This is even more difficult when van der Waals interactions <lb/>are concerned. The experimental studies described here are accurate enough to characterize the <lb/>topological properties at CP within a relative error estimated to a maximum of 20%. The most <lb/>difficult problem is to locate accurately the CP. When the density is obtained from analytical model <lb/>like multipole refinement, an analytical calculation of the derivatives of ( ) <lb/>r <lb/>r <lb/>ρ <lb/>for estimating the <lb/>topological properties may be performed; 31 in that way the errors are limited only to the <lb/>experimental error on the charge density. The topological analysis of the DFT density has been done <lb/>on a grid 38 the positions and the properties of the CPS depend on the grid size and to low extended <lb/>on the interpolation procedure when the charge density is flat. <lb/>3.6.1. Intrastack Interactions <lb/>Starting from the published structures at 300 and 40 K 18 , Oison et al. 37 have conducted a thorough <lb/>ab-initio study of TTF-CA. Despite differences in molecular geometries discussed in sect. 3.2, our <lb/>experimental and theoretical findings on intermolecular interactions are consistent with their results. <lb/>In particular, topological analysis of the charge density of the HT phase shows the existence of (3,-<lb/>1) CPs lying between TTF and CA in a plane parallel to the molecules. According to their <lb/>Laplacian, charge density and potential energy density, only the most important contacts are <lb/>reported in Table 3 and discussed here. The strongest contact between D and A molecules is clearly <lb/>visible in Fig. 3a. At HT, a 0.06 eǺ -3 isocontour links the two molecules via the S4···C15-C17 <lb/>contact. This is directly related to the HOMO-LUMO overlap that shows up in the valence band. 37 <lb/>The positive Laplacian (0.59/0.56 eǺ -5 ) and the charge density (0.06/0.06 eǺ -3 ) at the CP are <lb/>characteristic of a closed-shell interaction (Fig. 2a). The agreement between experimental and <lb/>theoretical results is quite good, but discrepancies are observed in the CP locations and ellipticities <lb/>as discussed previously. Experimentally, the CP is closer to the C17 atom than to C15, whereas <lb/>theoretically C15 and C17 distances to the CP are similar. Both experimental and theoretical <lb/>ellipticities are quite large. This is an indication that the difference in CP location arises from the <lb/>flatness of the charge density distribution. In the same way, the central C=C bond of TTF is <lb/>connected to C15 in the experimental density and to C17 in the theoretical one, again with quite <lb/>large ellipticities. Moreover, both contacts involve the same carbon atoms of CA, namely C17 and <lb/>C15. Consequently, it becomes very risky to deduce whether these contacts link individual atoms or <lb/>bridge over bonds. Increased theoretical and experimental precision is required to come to a <lb/>conclusion for such low and flat charge density regions. <lb/>Other (3,-1) CPs are found between the S5 atom and both chlorine atoms of the asymmetric unit. <lb/>These Cl19···S5 and Cl20···S5 contacts exhibit similar characteristics with experimental positive <lb/>Laplacian of about 0.43 eǺ -5 , a small charge density of 0.04-0.05 eǺ -3 , and a potential energy <lb/>density of about -8.1 kJ/mol. CPs in Cl···S contacts are not very frequent in literature. Compared to <lb/>S···S contacts reported in L-cystine 57a , Laplacian, charge density and potential energy density are <lb/>stronger. <lb/>In the LT phase, the creation of pairs of DA dimers clearly shows up (Fig. 3b) with an increase of <lb/>intradimer charge density. Due to the dimerization process, different inter-and intra dimer CP are <lb/>observed (Fig. 2b, Table 3). The S···C-C intra dimer contact remains the dominant one, with a 81% <lb/>increase in its strength as measured by V(r). In the dimer pair the strength of the C···C-C and <lb/>S5···Cl20 increases, while that of S5···Cl19 is slightly reduced. All inter-dimer contacts are <lb/>systematically reduced and the observed reduction of CP characteristics, such as potential energy <lb/>density, demonstrates their ability to investigate that tiny structural changes. The overall agreement <lb/>between experimental and theoretical results is quite good. <lb/>3.6.2 Interchain Interactions <lb/>As c cell parameter is twice the length of the two other ones, the stacks are too far along c to allow <lb/>any atom to atom interactions. Consequently, contacts occur between stacks related by translation <lb/>along b axis, and by the n glide plane symmetry which remains in the ionic phase (P21/n to Pn). <lb/>Interestingly, (3,-1) CPs are found not only on C-H···O, Cl···H and Cl···S contacts between D and A <lb/>molecules, but also for S···H (D-D) and Cl···Cl (A-A) ones, as shown in Fig. 4a and in Fig. 4b; the <lb/>topological characteristics of the strongest interstack interactions for both HT and LT phases are <lb/>given in Table 4. <lb/>At first observed by the temperature-dependence of the thermal expansion tensor 61 , and as discussed <lb/>by structure analysis 18 and DFT calculations 37,62 , interchain hydrogen bonds are the dominant <lb/>interaction in TTF-CA for both inter and intra stack interactions. In both HT and LT phases each CA <lb/>oxygen atom is linked to two TTF molecules through two different hydrogen bonds. First, as shown <lb/>in Figures 4a and 4b, the oxygen atoms accept the strongest bond interaction: O18···H6-C1 and <lb/>O24···H13-C8 its symmetry equivalent in the HT phase, are almost linear in the [1, 2, 0] direction <lb/>and are reinforced in the LT phase, but asymmetrically, O18···H6 being the strongest. In the other <lb/>lone pair direction the O18···H14-C10 and O24···H7-C3 interactions are bent and therefore weaker. <lb/>They are also strengthened in the ionic phase as their potential energy density increases by more <lb/>than 40%. This is related to the contraction of the crystal which is the largest in the plane containing <lb/>this C-H···O network. 61 <lb/>The fact that the H bonds are the dominant contacts of all the intra-and interstack interactions <lb/>contrasts with the first idea of the quasi-one-dimensionality of TTF-CA. The CT was for a long time <lb/>considered to happen as a chain along a axis, with weak or no interchain coupling. In this way, the <lb/>electrostatic interaction between the stacks was assumed to be much smaller than the intrastack one. <lb/>However, as discussed above, the temperature evolution of the cell parameters does not agree with <lb/>this assumption. Furthermore, as reported by Kawamoto and co-workers 63 , if the charge distribution <lb/>on atoms is considered, the electrostatic energy between TTF and CA aligned along the [1, 2, 0] <lb/>direction is attractive and superior to the repulsive one between CA molecules in the b direction; It <lb/>is in clear agreement with the above analysis of the dipole moment. On the other hand, it is well <lb/>established that valence and conduction bands have a clear one-dimensional character. 37 Thus, the <lb/>dimensionality of TTF-CA depends strongly on the property of interest. <lb/>In TTF-BA, the estimation of qCT by the Umland empirical method 49 seems to show that <lb/>dimerization does not imply CT variation because this complex is already ionic at room <lb/>temperature. 43 Can we find a rational explanation by looking at the molecular packing and the <lb/>corresponding intermolecular interactions? In TTF-BA, (space group P-1 to P1), the D and A <lb/>molecules alternate along a and b crystallographic axes, the longest axes of the D and A molecules <lb/>of the two different stacks being quasiparallel. This geometry leads to bent H bonds, with distances <lb/>similar to those observed in TTF-CA and therefore of equivalent strength. The main difference <lb/>between the two CT complexes comes from the halogen-halogen (C1-X1···X2-C2) interactions. <lb/>Depending on the geometric parameters two types of halogen-halogen interactions have been <lb/>defined. 64,65 In the first one, called head-on contact, both C-X bonds form a nearly collinear <lb/>arrangement, with contacts angles θ1 (C1-X1···X2) and θ2 (X1···X2-C2) around 160±10°. These <lb/>contacts may arise when a particular crystallographic operation relates the two halogen atoms; the <lb/>interaction is repulsive as evidenced by the electrostatic potential. 66 In the second type, the <lb/>interaction is attractive and is called side-on or polarization-induced contact, one of the angles is <lb/>170±10° and the other 90±10°. 64,65 The preference to form one of these two types of halogen-<lb/>halogen contacts is related to the polarizability of the halogen atoms. In TTF-BA attractive type II <lb/>interactions occur along the [1 1 1] and [1, 1, -1] directions with Br···Br distances of 3.566(3) at <lb/></body>
42
+
43
+ <page>11 <lb/></page>
44
+
45
+ <body>100K and 3.540(3) Ǻ at 15K, much smaller than the sum of van der Waals radii (3.70Å). In TTF-<lb/>CA, Cl···Cl interactions are repulsive type I interactions with distances of about 3.470(3)Å close to <lb/>the sum of van der Waals radii (3.50Å). <lb/>In conclusion, attractive interactions lead to X···X distances smaller than van der Waals contacts, <lb/>whereas the repulsive Cl…Cl interactions do not allow such shortening. Topological parameters of <lb/>the Cl···Cl contacts do not change very much upon the phase transition, as shown in Table 4. <lb/>4. Conclusion and perspectives <lb/>Accurate X-ray measurement and electron density modelling is a necessary tool to directly <lb/>investigate intermolecular interactions and tiny charge transfers. For the TTF-CA crystal, the <lb/>topological analysis of the intermolecular interactions using theoretical DFT and experimental <lb/>electron densities show that C-H···O interstack contacts are the strongest interactions and are <lb/>reinforced in the LT phase, whereas the strength of the Cl···Cl interactions, which may be repulsive <lb/>in that case, do not change when crossing the transition. The role of the latter is still not clear and <lb/>needs further discussion as for example electrostatic potential calculations. <lb/>This study also demonstrates that agreement between theory and experiment is not fully satisfactory <lb/>and requires more accurate calculations and modelling: theory is not able to reproduce the charge <lb/>transfer occurring during the NIT and both experiment and theory disagree in the position of the <lb/>critical points of some important intermolecular interactions as Cl···Cl for example. More critical is <lb/>the disagreement for finding the C2=C2···C15 (or C17) interaction pathway: both methods find a (3,-<lb/>1) CP with similar characteristics but experiment would prefer a C2···C15 interaction when theory <lb/>finds a C2···C17 interaction. Therefore the present accuracy does not allow to discriminate if this <lb/>interaction links individual atoms (and which atoms) or bridges over bonds. <lb/>By comparison with previous work 43 , halogen... halogen interactions need a better description in <lb/>order to know when and why these interactions are attractive or repulsive and at which level they <lb/>compete with C-H···O interactions during the phase transition process. In order to rank these <lb/>interactions charge density and related electrostatic potential of TTF -XA (X = F, Br, I) are <lb/>underway using high resolution X-ray diffraction: as an example Fig. 5 shows the contribution of the <lb/>Cl atom to the electrostatic potential of CA in the ionic phase; clearly two negative lobes appear <lb/>which should be related with the attractive or repulsive behaviour of X···X interactions. Although <lb/>the potential distribution is the same than that observed in the vicinity of the terminal halide ligand <lb/>of CH3Cl 66 , where the electrostatic energy of the minima is -80 kJ/mol; in TTF-CA -25 and -148 <lb/>kJ/mol are respectively obtained at 105 and 15K Therefore, together with the TTF, CA and TTF-CA <lb/>charge densities, we will have a large panorama of the interactions as a function of the charges, <lb/>multipole moments, polarizabilities of the halogen atoms and of the geometry and environment of <lb/>the contacts. <lb/>It has been shown that the charge transfer is not an unambiguously defined quantity that can be <lb/>directly (charge density modelling) or indirectly estimated from molecular geometry or IR spectra. <lb/>Although D and A molecular geometries are sensible to the charge transfer, its estimation by <lb/>empirical models is only qualitative: examination of different bond lengths leads to different qCT <lb/>values. Similarly, the variation of qCT is supposed to be proportional to the frequency of typical <lb/>phonon modes (e.g., C=O stretching modes of CA). Moreover, these methods present inevitable <lb/>ambiguities: molecular geometries and frequencies of single TTF and CA molecules are used as <lb/>references, because those of TTF-CA with qCT=0 and qCT=1e are not known. <lb/>To our opinion the best way to determine the charges remains in the analysis of the experimental or <lb/>theoretical charge density. The case of TTF-CA complex has been chosen because the CT is less <lb/>than one electron shared by 26 atoms. We have shown that both κ-Pv and Plm models give similar <lb/>and realistic results for the charge transfer. However, they can not be compared with theoretical <lb/>results, except if we transform the charge density into the theoretical structure factors. A direct <lb/>comparison is possible with a topological analysis of the charge density as described by Bader that <lb/>allows for a direct accurate estimation of the charge transfer occurring during the NIT (in the <lb/>experimental case ,∆qCT=0.53e). <lb/></body>
46
+
47
+ <page>12 <lb/></page>
48
+
49
+ <body>In conclusion these results show that Kappa, multipole and topological methods applied to X-ray <lb/>data lead to statistically similar results respectively, ∆qC =0.53, 0.59 and 0.53e. This raises the <lb/>following question; why do we obtain such a good agreement as the definitions of the κ-Pv and <lb/>topological charges are clearly different: κ-Pv (or multipolar) charges are integrated from reciprocal <lb/>space calculations on spherical volumes obtained from expanded or contracted HF free infinite <lb/>atoms whereas the topological charges are integrated inside very non spherical atomic basins of <lb/>limited volumes. It evidently leads to very different atomic net charges as shown in Table 5 which <lb/>compares them in the LT phase. <lb/>An explanation may be the following: in TTF CA the charge which we consider is summed up over <lb/>the whole TTF or CA molecule and the different shapes of atoms therefore has no real importance <lb/>because the charge is integrated over the whole molecular volume. Therefore, as there is an <lb/>electroneutrality constraint in the refinement process and as we have shown that the total charge <lb/>summed over all atomic basins of the unit cell is only -0.02 and 0.06e at 105K and 15K, instead of <lb/>zero (Table 3), the small differences between the values obtained from the two methods comes only <lb/>from the periphery of the molecules which electron density is small. Therefore as both charge <lb/>partitioning agree it strengthens the X-ray approach but this conclusion drawn in this particular case <lb/>needs to be discussed and confirmed by other charge estimations on other selected charge transfer <lb/>materials or ionic species. <lb/></body>
50
+
51
+ <div type="acknowledgement">Acknowledgements <lb/>Calculations have been supported by the &quot;Centre Informatique National de l&apos;Enseignement <lb/>Supérieur&quot; (CINES-France). CK is greatful to P. E. Blöchl for his PAW code and to N. Audiffren <lb/>and Ph. Falandry (CINES-France) respectively for parallelization and graphical interfacing <lb/>(OpenDX) of InteGriTy. PG, SD and CL are especially indebted to A. Bouché and E. Wenger <lb/>(Service Commun de Diffraction X, UHP Nancy 1) and to Drs J. Angyán and Iann Gerber for useful <lb/>discussions. PG is grateful to the Ministère de l&apos;Education National et de la Recherche for a doctoral <lb/>fellowship. <lb/></div>
52
+
53
+ <listBibl>References <lb/>P. Coppens, X-Ray Charge Densities and Chemical Bonding, 1997, Oxford University Press UK <lb/>V. G. Tsirelson and R. P. Ozerov Electron Density and Bonding in Crystals, 1996, Inst. Of <lb/>Physics Publishing Philadelphia USA. <lb/>F.L. Hirshfeld, Theor. Chem. Acta, 1977, 44, 129. <lb/>P. Coppens, Phys. Rev. Lett., 1975, 35, 98. <lb/>R. S. Mulliken, J. Chem. Phys., 1955, 23, 1833. <lb/>P. Coppens, T. N. Guru Row, P. Leung, E. D. Stevens, P. J. Becker and Y. W. Yang, Acta Cryst., <lb/>1979, A35, 63. <lb/>N. Bouhmaida, N. E. Ghermani, C. Lecomte and A. Thalal, Acta Cryst., 1999, A55, 729. <lb/>(a) . G. Ángyán and C. Chipot, Int. J. Quantum Chem.,1994, 52, 17; (b) J. G. Ángyán, C. <lb/>Chipot, F. Dehez, C. Hättig, G. Jansen and C. Millot, J. Comp. Chem., 2003, 24, 997. <lb/>R. F. W. Bader, J. Chem. Phys., 1980, 73, 2871. <lb/>Espinosa, J. Molins and C. Lecomte Chem. Phys. Lett., 1998, 285, 170. <lb/>J. B. Torrance, A. Girlando, J. J. Mayerle, J. I. Crowley, V. Y. Lee, P. Batail and S. J. Laplaca, <lb/>Phys. Rev. Lett., 1981, 47, 1747. <lb/>Y. Tokura, T. Koda, T. Mitani and G. Saito, Solid State Communications, 1982, 43, 757. <lb/>J. B. Torrance, J. E. Vazquez, J. J. Mayerle and V. Y. Lee, Phys. Rev. Lett., 1981, 46, 253. <lb/>A. Girlando and A. Painelli, Phys. Rev., 1986, B34, 2131. <lb/>S. Koshihara, Y. Takahashi, H. Sakai, Y. Tokura and T. Luty, J. Phys. Chem., 1999, B103, 2592. <lb/>E. Collet, M. H. Lemée-Cailleau, M. Buron-LeCointe, H. Cailleau, M. Wulff, S. Koshihara, M. <lb/>Meyer, L. Toupet, P. Rabiller and S. Techert, Science, 2003, 300, 612. <lb/>A.Girlando, R. Bozio, C. Pecile and J. B. Torrance, Phys. Rev., 1982, B26, 2306. <lb/>M. L. LeCointe, M. H. Lemée-Cailleau, H. Cailleau, B. Toudic, L. Toupet, G. Heger, F. Moussa, <lb/>P. Scheiss, K. H. Kraft and N. Karl, Phys. Rev., 1995, B51, 3374. <lb/>R. Argoud, P. Fertey, P. Bordet and J. Reymann, Acta Cryst., 2000, A56, s221. <lb/></listBibl>
54
+
55
+ <page>13 <lb/></page>
56
+
57
+ <listBibl>J. Cosier and A. M. Glazer, J. Appl. Cryst. , 1986, 19, 105. <lb/>Z. Otwinowski, W. Minor, DENZO-SMN, Methods in Enzymology, Vol. 276, Macromolecular <lb/>Crystal-lography, Part A, edited by C. W. <lb/>CrysAlis, Oxford Diffraction Poland Sp., CCD Data Reduction GUI, version 1.171.27p5 beta <lb/>Blessing, R. H. Crystallogr. Rev. 1987, 1, 3. (New version 1998). <lb/>M. Souhassou, E. Espinosa, C. Lecomte, R. H. Blessing, Acta Cryst., 1995, B51, 661. <lb/>G. M. Sheldrick, SHELXL97. Programs for Crystal Structure Analysis (Release 97-2), <lb/>University of Göttingen, Germany (1997). <lb/>N. Hansen and P. Coppens, Acta Cryst., 1978, A34, 909. <lb/>C. Jelsch, V. Pichon-Pesme, B. Bartosz, B. Guillot and C. Lecomte, to be published. <lb/>D. W. J. Cruikshank, Acta Cryst., 1949, 2, 65. <lb/>B. Rees, Acta Cryst.,1976, A32, 483. <lb/>R. Bader, Atoms in Molecules: A Quantum Theory (Oxford University Press: New York, 1994). <lb/>M. Souhassou, R. H. Blessing, J. Appl. Cryst., 1999, 32, 210. <lb/>J. P. Perdew and A. Zunger, Phys. Rev., 1981, B23, 5048. <lb/>A. D. Becke, J. Chem. Phys. 1992, 96, 2155. <lb/>J. P. Perdew, Phys. Rev. B, 1986, 33, 8822. <lb/>P. E. Blöchl, Phys. Rev. B, 1994, 50, 17953. <lb/>P. E. Blöchl, J. Chem. Phys., 1995, 103, 7422. <lb/>V. Oison, C. Katan, P. Rabiller, M. Souhassou, C. Koening, Phys. Rev. B, 2003, 67, 035120. <lb/>C. Katan, P. Rabiller, C. Lecomte, M. Guezo, V. Oison, and M. Souhassou, J. Appl. Cryst. 2003, <lb/>36, 65. <lb/>V. Oison, P. Rabiller, C. Katan, J. Phys. Chem.., 2004, A108, 11049. <lb/>(a) R. Merkle, A. Savin and H. Preuss, J. Chem. Phys., 1992, 97, 9216; (b) E. Ruiz, D. R. <lb/>Salahub and A. Vela, J. Phys. Chem., 1996, 100, 12265; (c) Y. K. Zhang and W. T. Yang, J. <lb/>Chem. Phys., 1998, 109, 2604. <lb/>M. Buron-Le Cointe, M. H. Lemee-Cailleau, H. Cailleau, B. Toudic, A. Moreac, F. Moussa, C. <lb/>Ayache and N. Karl, Phys. Rev., 2003, B68, 064103. <lb/>J. J. Mayerle, J. B. Torrance and J. I. Crowley, Acta. Cryst., 1979, B35, 2988. <lb/>P. Garcia, S. Dahaoui, P. Fertey, E. Wenger and C. Lecomte, Phys. Rev., 2005, B72, 104115. <lb/>C. Katan; J. Phys. Chem., 1999, A103, 1407. <lb/>Y. Nogami, M. Taoda, K. Oshima, S. Aoki, T. Nakayama and A. Miura, Synth. Met., 1995, 70, <lb/>1219. <lb/>A. Tomita and S. Koshihara, 2006, in &quot; First France-Japan Advanced School on Chemistry and <lb/>Physics of Molecular Materials&quot;. <lb/>C. S. Jacobsen and J. B. Torrance, J. Chem. Phys., 1983, 78, 112. <lb/>A. Girlando, I. Zanon, R. Bozio and C. Pecile, J. Chem. Phys., 1978, 68, 22. <lb/>T. C. Umland, S. Allie, T. Kuhlmann and P. Coppens, J. Phys. Chem., 1988, 92, 6456. <lb/>P. Guionneau, C. J. Kepert, G. Bravic, D. Chasseau, M. R. Truter, M. Kurmoo and P. Day, <lb/>Synth. Meta., 1997, 86, 1973. <lb/>C. Katan, P. E. Blochl, P. Margl and C. Koeniq, Phys. Rev., 1996, B53, 12112. <lb/>C. Pecile, A. Painelli and A. Girlando, Mol. Cryst. Liq. Cryst., 1989, 171, 69. <lb/>E. Espinosa, E. Molins and C. Lecomte, Phys. Rev. B, 1997, 56, 1820. <lb/>C. Rovira, J. Tarrés, J. Llorca, E. Molins, J. Veciana, S. Yang, D. O. Cowan, C. Garrigou-<lb/>Lagrange, J. Amiell, P. Delhaes, E. Canadell and J. P. Pouget, Phys. Rev. , 1995, B52, 8747. <lb/>(a) J. L. Segura and N. Martin, Angew. Chem. Int. Ed., 2001, 40, 1372.; (b) M. R. Bryce, Adv. <lb/>Mater., 1999, 1, 11. <lb/>J. Lajzerowicz and J. Legrand, Phys. Rev. , 1978, B17, 1438. <lb/>(a) S. Dahaoui, V. Pichon-Pesme, J. A. K. Howard and C. Lecomte, J. Phys. Chem. A., 1999, <lb/>103, 6240; (b) S. Pillet, M. Souhassou, Y. Pontillon, A. Caneschi, D. Gatteschi and C. Lecomte, <lb/>New J. Chem., 2001, 1, 131; (c) F. Benabicha, V. Pichon-Pesme, C. Jelsch, C. Lecomte and A. <lb/>Khmou, Acta. Cryst., 2000, B56, 155; R. Guillot, N. Muzet, S. Dahaoui, C. Lecomte and C. <lb/>Jelsch, Acta. Cryst., 2001, B57, 567. <lb/>A. Bachn D. Lentz and P. Luger, J. Phys. Chem. , 2001, A105, 7405. <lb/>A. Volkov, and P. Coppens, Acta. Cryst., 2001, A57, 395. <lb/>E. Espinosa, M. Souhassou, H. Lachekar and C. Lecomte, Acta. Cryst. , 1999, B55, 563. <lb/>P. Batail, S. J. LaPlaca, J. J. Mayerle and J. B. Torrance, J. Am. Chem. Soc., 1981, 103, 951. <lb/>V. Oison, C. Katan and C. Koening, J. Phys. Chem., 2001, A105, 4300. <lb/>T. Kawamoto, T. Iizuka-Sakano, Y. Shimoi and S. Abe, Phys. Rev., 2001, B64, 205107. <lb/></listBibl>
58
+
59
+ <page>14 <lb/></page>
60
+
61
+ <listBibl>T. Sakurai, M. Sundaralingam, and G. A. Jeffrey, Acta Cryst., 1963, 16, 354. <lb/>N. Ramasubbu, R. Parthasarathy and P. Murray-Rust, J. Am. Chem. Soc., 1986, 108, 4308. <lb/>L. Brammer, E. A. Bruton, and P. Sherwood, Cryst. Growth Des., 2001, 1, 277. <lb/></listBibl>
62
+
63
+ <body>Figures <lb/>Fig. 1a: Temperature evolution of the beta angle (°). Cooling (circles) and warming (triangles). The <lb/>standard error estimated on statistical ground is 0.02°. <lb/>Fig. 1b: Temperature evolution of the b parameter (Ǻ). Cooling (circles) and warming (triangles). <lb/>The standard error estimated on statistical ground is 0.001 Ǻ. <lb/>Fig. 2a: Experimental critical points associated to the strongest intrastack contacts in TTF-CA at <lb/>105K and shortest distance (C2···C15). <lb/>Fig. 2b: Experimental critical points associated to the strongest intrastack contacts in TTF-CA at <lb/>15K and shortest intra and interdimer distances (C9···C21 and C2···C15 respectively). <lb/>Fig. 3a: Experimental total charge density at 105K. Isocontours of 0.065 e/Ǻ 3 . <lb/>Fig. 3b: Experimental total charge density at 15K. Isocontours of 0.065 e/Ǻ 3 : the dimer is clearly <lb/>identified. <lb/>Fig. 4a: Experimental critical points associated to the strongest interstack interactions at 15K. <lb/>Fig. 4b: Experimental critical points around CA molecule associated to the strongest interstack <lb/>interactions at 15K. <lb/>Fig. 5: Contribution of the chlorine atom to the electrostatic potential at 15K. Contours intervals are <lb/>0.02 e/Å. <lb/>Table 1 <lb/>Table 1 Experimental details <lb/></body>
64
+
65
+ <page>16 <lb/></page>
66
+
67
+ <body>T T F <lb/>C A <lb/>T T F -C A <lb/>T T F -C A <lb/>Formula <lb/>C6 S4H4 <lb/>C 6Cl4O2 <lb/>C 12S4Cl4O2H4 <lb/>C 12S4Cl4O2H4 <lb/>M <lb/>817.33 <lb/>491.72 <lb/>900.38 <lb/>900.38 <lb/>T(K) <lb/>100 <lb/>100 <lb/>105 <lb/>15 <lb/>Crystal system <lb/>Triclinic <lb/>Monoclinic <lb/>Monoclinic <lb/>Monoclinic <lb/>Space group <lb/>P -1 <lb/>P21/c <lb/>P21/n <lb/>Pn <lb/>Unit cell <lb/>dimensions <lb/>a= 8.3101(3) Ǻ <lb/>b=12.7800(3) Ǻ <lb/>c= 7.9582(1) Ǻ <lb/>α= 99.627(3)° <lb/>β = 95.906(3)° <lb/>γ= 101.083(3)° <lb/>a= 8.574(1) Ǻ <lb/>b= 5.690(1) Ǻ <lb/>c= 8.700(1) Ǻ <lb/>β=105.937(8) ° <lb/>a= 7.2297(5) Ǻ <lb/>b= 7.5933(3) Ǻ <lb/>c=14.4980(9) Ǻ <lb/>β=99.15(3)° <lb/>a= 7.1999(9) Ǻ <lb/>b= 7.5556(6) Ǻ <lb/>c=14.479(1) Ǻ <lb/>β=98.511(8)° <lb/>Volume (Ǻ 3 ) <lb/>817.33 <lb/>403.43 <lb/>786.2 (2) <lb/>779.02(13) <lb/>Z <lb/>4 <lb/>2 <lb/>2 <lb/>2 <lb/>Reflections <lb/>collected/ <lb/>independent <lb/>141108 / 16513 <lb/>85699 / 5445 <lb/>168470 / 9237 <lb/>75177 / 14634 <lb/>µ(Mo-Kα) (mm -1 ), <lb/>Resolution (Ǻ -1 ) <lb/>1.09, 1.10 <lb/>1.41, 1.19 <lb/>1.28, 1.14 <lb/>1.29, 1.16 <lb/>Rint <lb/>0 . 0 7 4 1 <lb/>0 . 0 2 6 9 <lb/>0 . 0 3 6 4 <lb/>0 . 0 2 6 6 <lb/>R (I&gt;3σ(I)) a <lb/>0 . 0 1 4 2 <lb/>0 . 0 1 6 7 <lb/>0 . 0 1 4 1 <lb/>0 . 0 1 3 8 <lb/>a At the end of multipolar refinement. <lb/>Table 2 <lb/>Table 2 Topological molecular volumes (Å 3 ) and charges (e), compared to those obtained with κ-Pv and <lb/>multipolar refinement. For TTF* and CA*, data are obtained from experiments on pure TTF and CA <lb/>single crystals. Other results are relative to molecules in single crystals of TTF-CA at either 105 or 15 <lb/>K. <lb/>T T F * <lb/>T T F 105K <lb/>TTF 15K <lb/>C A * <lb/>C A 105K <lb/>C A 15K <lb/>V exp <lb/>2 0 5 . 1 2 <lb/>1 9 1 . 2 4 <lb/>1 8 6 . 4 9 <lb/>2 0 4 . 9 2 <lb/>2 0 1 . 5 2 <lb/>2 0 1 . 6 9 <lb/>V theo <lb/>1 8 9 . 9 2 <lb/>1 8 7 . 0 7 <lb/>2 0 2 . 8 2 <lb/>2 0 2 . 2 7 <lb/>q exp <lb/>0 . 0 1 <lb/>0 . 2 0 <lb/>0 . 7 7 <lb/>-0 . 0 2 <lb/>-0 . 2 2 <lb/>-0 . 7 1 <lb/>q theo <lb/>0 . 5 4 <lb/>0 . 6 3 <lb/>-0 . 5 0 <lb/>-0 . 6 4 <lb/>q exp, κ-Pv <lb/>refinement <lb/>0 . 1 4 <lb/>0 . 6 7 <lb/>-0.14 <lb/>-0.67 <lb/>q exp, mult <lb/>refinement <lb/>0 . 0 6 <lb/>0 . 6 5 <lb/>-0.06 <lb/>-0.65 <lb/>Table 3 <lb/></body>
68
+
69
+ <page>17 <lb/></page>
70
+
71
+ <body>Table 3 Topological characteristics of the electron density at selected (3,-1) intrachain CPs. The λ1 , λ2 , <lb/>λ3 (e Å -5 ) are the curvatures of the three principal axes of the Hessian matrix, ∇ 2 ρ (e Å -5 ) is the <lb/>Laplacian of the electron density, ρ (e Å -3 ) is the electron density ε is the ellipticity value and V(r) <lb/>(kJ/mol) is the potential energy density. The values d1, d2 and d3 (Å) represent the distances from the CP <lb/>to the attractors 1, 2 and 3. First line: experimental electron density, second line: theoretical electron <lb/>density. <lb/>105K <lb/>d1 <lb/>d2 <lb/>d3 <lb/>∇ 2 ρ <lb/>ρ <lb/>ε <lb/>V ( r ) <lb/>λ3 <lb/>λ2 <lb/>λ1 <lb/>S4··· C17-C15 <lb/>1.816 <lb/>1.818 <lb/>1.628 <lb/>1.692 <lb/>1.802 <lb/>1.688 <lb/>0.59 <lb/>0.56 <lb/>0.06 <lb/>0.06 <lb/>1.53 <lb/>5.66 <lb/>-11.1 <lb/>-11.0 <lb/>0.77 <lb/>0.72 <lb/>-0.05 <lb/>-0.02 <lb/>-0.13 <lb/>-0.14 <lb/>C15··· C2-C2A <lb/>a <lb/>1.633 <lb/>1.742 <lb/>2.194 <lb/>0.45 <lb/>0.04 <lb/>6.68 <lb/>-7.0 <lb/>0.51 <lb/>-0.01 <lb/>-0.05 <lb/>C17··· C2-C2A <lb/>b <lb/>1.713 <lb/>1.809 <lb/>1.856 <lb/>0.43 <lb/>0.04 <lb/>7.52 <lb/>-6.8 <lb/>0.51 <lb/>-0.008 <lb/>-0.01 <lb/>Cl19 ···S5 <lb/>1.843 <lb/>1.817 <lb/>1.851 <lb/>1.875 <lb/>0 . 4 2 <lb/>0.45 <lb/>0.05 <lb/>0.04 <lb/>0.44 <lb/>0.24 <lb/>-8.1 <lb/>-7.2 <lb/>0.57 <lb/>0.61 <lb/>-0.06 <lb/>-0.06 <lb/>-0.09 <lb/>-0.08 <lb/>Cl20 ···S5 <lb/>1.797 <lb/>1.795 <lb/>1.840 <lb/>1.833 <lb/>0 . 4 4 <lb/>0.52 <lb/>0.05 <lb/>0.04 <lb/>1.22 <lb/>0.29 <lb/>-8.2 <lb/>-8.5 <lb/>0.58 <lb/>0.67 <lb/>-0.04 <lb/>-0.06 <lb/>-0.10 <lb/>-0.08 <lb/>15K <lb/>Intradimer <lb/>S11··· C23-C15 <lb/>1.730 <lb/>1.725 <lb/>1.518 <lb/>1.601 <lb/>1.811 <lb/>1.625 <lb/>0.74 <lb/>0.72 <lb/>0.10 <lb/>0.08 <lb/>0.42 <lb/>4.60 <lb/>-20.2 <lb/>-17.0 <lb/>1.19 <lb/>0.99 <lb/>-0.18 <lb/>-0.04 <lb/>-0.26 <lb/>-0.23 <lb/>C17··· C2-C9 <lb/>a <lb/>1.548 <lb/>1.710 <lb/>1.872 <lb/>0.69 <lb/>0.08 <lb/>2.23 <lb/>-15.6 <lb/>0.93 <lb/>-0.06 <lb/>-0.18 <lb/>C17-C21 ···C9 <lb/>b <lb/>1.617 <lb/>1.571 <lb/>1.846 <lb/>0.59 <lb/>0.06 <lb/>4.85 <lb/>-12.0 <lb/>0.75 <lb/>-0.02 <lb/>-0.14 <lb/>Cl20···S5 <lb/>1.718 <lb/>1.762 <lb/>1.845 <lb/>1.792 <lb/>0 . 5 2 <lb/>0.62 <lb/>0.06 <lb/>0.05 <lb/>1.61 <lb/>0.52 <lb/>-10.5 <lb/>-10.0 <lb/>0.69 <lb/>0.78 <lb/>-0.05 <lb/>-0.06 <lb/>-0.12 <lb/>-0.09 <lb/>Cl19···S5 <lb/>1.838 <lb/>1.801 <lb/>1.832 <lb/>1.848 <lb/>0 . 4 1 <lb/>0.50 <lb/>0.04 <lb/>0.04 <lb/>0.69 <lb/>0.51 <lb/>-6.7 <lb/>-8.1 <lb/>0.56 <lb/>0.66 <lb/>-0.06 <lb/>-0.06 <lb/>-0.10 <lb/>-0.09 <lb/>Interdimer <lb/>S4··· C21-C17 <lb/>1.855 <lb/>1.884 <lb/>1.738 <lb/>1.875 <lb/>1.967 <lb/>1.718 <lb/>0.38 <lb/>0.44 <lb/>0.04 <lb/>0.04 <lb/>1.85 <lb/>3.99 <lb/>-6.4 <lb/>-7.4 <lb/>0.50 <lb/>0.55 <lb/>-0.03 <lb/>-0.02 <lb/>-0.09 <lb/>-0.09 <lb/>Cl25···S12 <lb/>1.819 <lb/>1.828 <lb/>1.890 <lb/>1.879 <lb/>0 . 4 3 <lb/>0.43 <lb/>0.04 <lb/>0.04 <lb/>0.88 <lb/>0.08 <lb/>-6.8 <lb/>-6.8 <lb/>0.59 <lb/>0.59 <lb/>-0.06 <lb/>-0.08 <lb/>-0.11 <lb/>-0.08 <lb/>Cl26···S12 <lb/>1.820 <lb/>1.809 <lb/>1.839 <lb/>1.845 <lb/>0.44 <lb/>0.49 <lb/>0.04 <lb/>0.04 <lb/>2.12 <lb/>0.21 <lb/>-6.5 <lb/>-7.9 <lb/>0.59 <lb/>0.64 <lb/>-0.04 <lb/>-0.07 <lb/>-0.11 <lb/>-0.08 <lb/>a From experimental electron density. b From theoretical electron density. The differences in location of <lb/>the CP are discussed in the body of the paper. <lb/></body>
72
+
73
+ <page>18 <lb/></page>
74
+
75
+ <body>Table 4 <lb/>Table 4 Topological characteristics of the electron density at selected (3,-1) interchain CPs. Same <lb/>comments as Table 3. <lb/>105K <lb/>d1 <lb/>d2 <lb/>∇ 2 ρ <lb/>ρ <lb/>ε <lb/>V ( r ) <lb/>λ3 <lb/>λ2 <lb/>λ1 <lb/>H6···O18 <lb/>0.881 <lb/>0.878 <lb/>1.396 <lb/>1.404 <lb/>1.21 <lb/>1.08 <lb/>0.06 <lb/>0.09 <lb/>0.05 <lb/>0.02 <lb/>-16.7 <lb/>-20.2 <lb/>1.70 <lb/>1.75 <lb/>-0.24 <lb/>-0.33 <lb/>-0.25 <lb/>-0.33 <lb/>H7···O18 <lb/>1.000 <lb/>0.971 <lb/>1.443 <lb/>1.453 <lb/>0.90 <lb/>0.88 <lb/>0.06 <lb/>0.06 <lb/>0.14 <lb/>0.04 <lb/>-13.9 <lb/>-15.0 <lb/>1.26 <lb/>1.34 <lb/>-0.17 <lb/>-0.22 <lb/>-0.19 <lb/>-0.23 <lb/>Cl20···S4 <lb/>1.722 <lb/>1.757 <lb/>1.759 <lb/>1.723 <lb/>0.49 <lb/>0.54 <lb/>0.06 <lb/>0.04 <lb/>0.28 <lb/>0.08 <lb/>-10.2 <lb/>-8.5 <lb/>0.71 <lb/>0.72 <lb/>-0.10 <lb/>-0.08 <lb/>-0.12 <lb/>-0.09 <lb/>Cl20···Cl19 <lb/>1.732 <lb/>1.743 <lb/>1.734 <lb/>1.738 <lb/>0.50 <lb/>0.56 <lb/>0.05 <lb/>0.04 <lb/>0.10 <lb/>0.14 <lb/>-8.8 <lb/>-8.6 <lb/>0.70 <lb/>0.77 <lb/>-0.09 <lb/>-0.09 <lb/>-0.10 <lb/>-0.11 <lb/>15K <lb/>Short <lb/>H13···O24 <lb/>0.827 <lb/>0.813 <lb/>1.343 <lb/>1.346 <lb/>1.47 <lb/>1.36 <lb/>0.10 <lb/>0.11 <lb/>0.03 <lb/>0.01 <lb/>-26.8 <lb/>-29.0 <lb/>2.20 <lb/>2.32 <lb/>-0.36 <lb/>-0.48 <lb/>-0.37 <lb/>-0.48 <lb/>H14···O18 <lb/>0.901 <lb/>0.888 <lb/>1.387 <lb/>1.389 <lb/>1.28 <lb/>1.17 <lb/>0.09 <lb/>0.09 <lb/>0.07 <lb/>0.06 <lb/>-22.9 <lb/>-21.9 <lb/>1.89 <lb/>1.86 <lb/>-0.3 <lb/>-0.33 <lb/>-0.32 <lb/>-0.35 <lb/>Cl20···S11 <lb/>1.704 <lb/>1.721 <lb/>1.703 <lb/>1.683 <lb/>0.44 <lb/>0.62 <lb/>0.06 <lb/>0.05 <lb/>0.44 <lb/>0.06 <lb/>-9.7 <lb/>-10.0 <lb/>0.75 <lb/>0.85 <lb/>-0.12 <lb/>-0.11 <lb/>-0.18 <lb/>-0.12 <lb/>Cl26···Cl19 <lb/>1.770 <lb/>1.747 <lb/>1.727 <lb/>1.753 <lb/>0.51 <lb/>0.54 <lb/>0.05 <lb/>0.04 <lb/>0.14 <lb/>0.15 <lb/>-8.9 <lb/>-8.3 <lb/>0.73 <lb/>0.75 <lb/>-0.11 <lb/>-0.09 <lb/>-0.12 <lb/>-0.11 <lb/>Long <lb/>H6···O18 <lb/>0.858 <lb/>0.857 <lb/>1.373 <lb/>1.389 <lb/>1.28 <lb/>1.16 <lb/>0.09 <lb/>0.09 <lb/>0.02 <lb/>0.03 <lb/>-22.9 <lb/>-22.0 <lb/>1.90 <lb/>1.91 <lb/>-0.30 <lb/>-0.37 <lb/>-0.31 <lb/>-0.38 <lb/>H7···O24 <lb/>0.969 <lb/>0.935 <lb/>1.405 <lb/>1.425 <lb/>1.12 <lb/>1.02 <lb/>0.08 <lb/>0.07 <lb/>0.09 <lb/>0.08 <lb/>-19.5 <lb/>-16.7 <lb/>1.64 <lb/>1.57 <lb/>-0.25 <lb/>-0.26 <lb/>-0.27 <lb/>-0.28 <lb/>Cl26···S4 <lb/>1.725 <lb/>1.751 <lb/>1.740 <lb/>1.711 <lb/>0.41 <lb/>0.56 <lb/>0.06 <lb/>0.04 <lb/>0.23 <lb/>0.12 <lb/>-9.5 <lb/>-8.9 <lb/>0.71 <lb/>0.75 <lb/>-0.14 <lb/>-0.08 <lb/>-0.17 <lb/>-0.09 <lb/>Cl20···Cl25 <lb/>1.849 <lb/>1.745 <lb/>1.632 <lb/>1.733 <lb/>0.56 <lb/>0.56 <lb/>0.04 <lb/>0.04 <lb/>1.34 <lb/>0.14 <lb/>-8.0 <lb/>-8.7 <lb/>0.62 <lb/>0.78 <lb/>-0.02 <lb/>-0.10 <lb/>-0.04 <lb/>-0.11 <lb/></body>
76
+
77
+ <page>19 <lb/></page>
78
+
79
+ <body>Table 5 <lb/>Table 5 Comparison between κ-Pv, multipolar and topological net charges (e) of atoms of TTF-CA at <lb/>15K, statistical errors are 0.02e. <lb/>TTF Atom <lb/>C1/C8 <lb/>C2/C9 <lb/>C3/C10 <lb/>S4/S11 <lb/>S5/S12 <lb/>H6/H13 <lb/>H7/H14 <lb/>κ-Pv <lb/>-0.26 <lb/>0.13 <lb/>-0.24 <lb/>0.20 <lb/>0.30 <lb/>0.14 <lb/>0.06 <lb/>Multipolar <lb/>model <lb/>-0.59 <lb/>-0.23 <lb/>-0.50 <lb/>0.38 <lb/>0.55 <lb/>0.38 <lb/>0.34 <lb/>Topological <lb/>exp <lb/>-0.53 <lb/>-0.23 <lb/>-0.29 <lb/>0.38 <lb/>0.50 <lb/>0.31 <lb/>0.26 <lb/>Topological <lb/>theo <lb/>-0.14 <lb/>-0.30 <lb/>-0.13 <lb/>0.28 <lb/>0.33 <lb/>0.14 <lb/>0.13 <lb/>CA Atom <lb/>C15/C21 <lb/>C16/C22 <lb/>C17/C23 <lb/>O18/O24 <lb/>Cl19/Cl25 <lb/>Cl20/Cl26 <lb/>κ-Pv <lb/>0.22 <lb/>0.23 <lb/>0.14 <lb/>-0.42 <lb/>-0.27 <lb/>-0.23 <lb/>Multipolar <lb/>0.32 <lb/>-0.04 <lb/>0.20 <lb/>-0.44 <lb/>-0.11 <lb/>-0.25 <lb/>Topological <lb/>exp <lb/>0.92 <lb/>0.05 <lb/>0.30 <lb/>-1.09 <lb/>-0.19 <lb/>-0.34 <lb/>Topological <lb/>theo <lb/>0.95 <lb/>0.08 <lb/>0.08 <lb/>-1.11 <lb/>-0.16 <lb/>-0.16 </body>
80
+
81
+
82
+ </text>
83
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/Fatnassi2015_Article_MagneticAndMagnetocaloricPrope.training.segmentation.tei.xml ADDED
@@ -0,0 +1,75 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>ORIGINAL RESEARCH <lb/>Magnetic and magnetocaloric properties of nano-sized <lb/>La 0.8 Ca 0.2 Mn 12x Fe x O 3 manganites prepared by sol-gel method <lb/>D. Fatnassi 1,3 • Kheiria Sbissi 1 • E. K. Hlil 2 • M. Ellouze 1 • <lb/>J. L. Rehspringer 3 • F. Elhalouani 4 <lb/>Received: 28 May 2015 / Accepted: 8 July 2015 / Published online: 24 July 2015 <lb/>Ó The Author(s) 2015. This article is published with open access at Springerlink.com <lb/>Abstract We present an investigation on magnetic and <lb/>magnetocaloric properties of nano-sized La 0.8 Ca 0.2 Mn 1-x-<lb/>Fe x O 3 (x = 0, 0.01, 0.15, 0.2) manganites synthesized by <lb/>sol-gel process. X-ray diffraction and magnetization <lb/>measurements were performed to investigate both crystal-<lb/>lographic structure and magnetocaloric properties, respec-<lb/>tively. All samples show an orthorhombic structure with <lb/>Pnma space group. Ferromagnetic-paramagnetic transition <lb/>sensitive to iron doping is observed at Curie temperature <lb/>(T C ) ranging from 223 K (x = 0) to 70 K (x = 0.2). In <lb/>addition, a large magnetocaloric effect near T C is observed. <lb/>Under a magnetic field change of 5 T, a maximum of <lb/>magnetic entropy DS max <lb/>M <lb/>reaches 4.42, 4.32, 1.6, and <lb/>0.54 J kg -1 K -1 , for x = 0, x = 0.01, x = 0.15, and <lb/>x = 0.2, respectively. The relative cooling power (RCP) <lb/>values were computed as well. RCP values of 164 and <lb/>117 J kg -1 were found for La 0.8 Ca 0.2 MnO 3 (LCM) and <lb/>La 0.8 Ca 0.2 Mn .0.99 Fe 0.01 O 3 (LCMFe 0.01 ), respectively. The <lb/>large values of entropy changes and related RCP allow <lb/>concluding that our material could be a highly attractive <lb/>candidate for magnetic refrigeration. <lb/>Keywords Pechini sol-gel Á Manganites Á <lb/>Magnetocaloric effect Á Relative cooling power (RCP) <lb/></front>
8
+
9
+ <body>Introduction <lb/>The modern society is increasingly relying on refrigeration <lb/>technology. The vapor compression refrigerators have been <lb/>mainly used for cooling applications. However, the com-<lb/>pression and expanding processes in the refrigerators of a gas <lb/>are not sufficiently efficient. On the other hand, the use of <lb/>gases such as chlorofluorocarbons and hydro-chlorofluoro-<lb/>carbons is damaging to our living environment. For these <lb/>reasons, exploring a new type refrigeration technology that is <lb/>environmentally friendly and energy efficient becomes an <lb/>urgent necessity. Comparing to the conventional gas com-<lb/>pression (CGC), magnetic refrigeration (MR) based on the <lb/>magnetocaloric effect (MCE) [1] exhibits several advan-<lb/>tages [2, 3]. Indeed, the MR does not use global warming <lb/>gases and therefore is an environmentally friendly cooling <lb/>technology [4, 5]. So an ongoing research is necessary to find <lb/>appropriate materials with a large enough magnetic entropy <lb/>at moderate magnetic fields near room temperature [5, 6]. <lb/>The gadolinium (Gd) rare earth metal has been considered as <lb/>the most obvious material exhibiting a large MCE in room-<lb/>temperature magnetic refrigerators [2, 7, 8]. However, the <lb/>use of the Gd is limited due to its expensive cost price. <lb/>Nevertheless, some other candidates have been found <lb/>to exhibit large MCE, such as Gd 5 (Si x Ge 1-x ) 4 [9], La <lb/>(Fe 1-x Si x ) 13 [10], MnFeP 1-x As x [11], and Tb 1-x Gd x A l2 [12]. <lb/>In the last few years, manganites with a general formula <lb/>R 1-x A x MnO 3 (R = rare earth, A = alkali earth) have <lb/>attracted more attention as alternative candidates for MR <lb/>near room temperature. Compared to Gd, they show sev-<lb/>eral advantages such as higher chemical stability, higher <lb/></body>
10
+
11
+ <front>&amp; D. Fatnassi <lb/>[email protected] <lb/>1 <lb/> Faculty of Sciences of Sfax, Sfax University, <lb/>BP 1171-3000, Sfax, Tunisia <lb/> 2 <lb/> Institute Ne ´el, CNRS et Universite ´Joseph Fourier, BP 166, <lb/>38042 Grenoble Cedex 9, France <lb/>3 <lb/> Institute of Physics and Chemistry of Materials of Strasbourg, <lb/>UMR 7504 CNRS Universite ´de Strasbourg, BP 43, <lb/>67034 Strasbourg Cedex 2, France <lb/>4 <lb/>National Engineering School of Sfax, BP W, 3038 Sfax, <lb/>Tunisia <lb/> 123 <lb/> J Nanostruct Chem (2015) 5:375-382 <lb/>DOI 10.1007/s40097-015-0169-7 <lb/></front>
12
+
13
+ <body>resistivity, and lower cost. Their preparation can be <lb/>achieved without substantial difficulties. As other advan-<lb/>tages, they present the possibility to tune their magnetic <lb/>transition temperature by the substitution on both R-sites <lb/>and Mn-sites. The MCE of La 1-x A x MnO 3 (A = Ca, Sr, Ba) <lb/>manganites was first studied by Moreli et al. [13]. A large <lb/>MCE in La 1-x A x MnO 3 polycrystalline samples <lb/>(0.2 B x B 0. 33) is reported by Guo et al. [14, 15]. In fact, <lb/>for DH = 1.5 T, the DS M <lb/>j <lb/>j reaches a maximum of about <lb/>5.5 J/(kg K) at 230 K, 4.7 J/(kg K) at 224 K, and 4.3 J/ <lb/>(kg K) at 260 K for x = 0.2, 0.25, and 0.33, respectively <lb/>[14]. For the same magnetic field of 1.5 T, these values are <lb/>larger than that of Gd, DS M <lb/>j <lb/>j = 4.2 J/(kg K) [2]. As other <lb/>significant information, the magnitude of DS M <lb/>j <lb/>j was found <lb/>to be inversely proportional to the grain size [16]. This <lb/>paper is devoted to seeking for new perovskite manganites <lb/>with broad refrigerant capacity and large MCE demanding <lb/>only low applied magnetic fields close to the room tem-<lb/>perature. Precisely, we report the effect of Fe doping on the <lb/>magnetic and magnetocaloric properties of La 0.8 Ca 0.2-<lb/>Mn 1-x Fe x O 3 (x = 0, 0.01, 0.15, 0.2). They present large <lb/>magnetic entropy change values and high relative cooling <lb/>power (RCP) factors. <lb/>Experimental <lb/>The nano-sized La 0.8 Ca 0.2 Mn 1-x Fe x O 3 (x = 0, 0.01, 0.15, <lb/>0.2) manganites were synthesized using the sol-gel method. <lb/>The starting precursor MnO 2 was mixed in appropriate <lb/>proportion and dissolved in the concentrated nitric acid. <lb/>Suitable amounts of citric acid and ethylene glycol, as a <lb/>coordinate agent, were added. After the addition of the other <lb/>precursors La 2 O 3 , CaCO 3 , and Fe 2 O 3 , a clear black stained <lb/>solution was obtained. Then the solution is allowed to dry to <lb/>form a dried gel, followed by baking at 170°C to obtain black <lb/>precursor powder. Finally, the resulting powder was heated <lb/>in air at 950 °C for 24 h. The X-ray diffraction patterns at <lb/>room temperature were obtained using SIEMENS D8 X-ray <lb/>diffractometer with Cu Ka radiation. The FULLPROF pro-<lb/>gram based on the Rietveld method [17] was used for phase <lb/>analysis. The magnetic isotherms were recorded in the <lb/>magnetic field of up to 5 T and at the temperature ranging <lb/>from 4 to 400 K. The magnetocaloric effects (MCE) were <lb/>estimated via the Maxwell relation [2]. <lb/>Results and discussion <lb/>X-ray diffraction <lb/>Powder X-ray diffraction patterns (Fig. 1) show that the <lb/>samples show single phase and indexed in the <lb/>orthorhombic structure with Pnma group space (Fig. 2). <lb/>Refined cell parameters such as unit cell parameters, unit <lb/>cell volume, R factor, and the goodness-of-fit indicator (v 2 ) <lb/>are listed in Table 1. We can deduce that the substitution of <lb/>Mn 3? by Fe 3? ions induces an increase of the unit cell <lb/>volume. The linear increase is unexpected because we <lb/>substitute Mn 3? having 0.0645 nm as ionic radius by Fe 3? <lb/>with the same ionic radius (0.0645 nm). Consequently, no <lb/>change induced by this substitution is expected. Therefore, <lb/>the increase could be attributed to the lattice disorder <lb/>arising from the random occupancy of Fe and Mn ions on <lb/>the B-site. Indeed, in the pure perovskite La 0.8 Ca 0.2 MnO 3 <lb/>Fig. 1 Powder X-ray diffraction patterns for La 0.8 Ca 0.2 Mn 1-x Fe x O 3 <lb/>(x = 0, 0.01, 0.15, 0.2) <lb/>Fig. 2 Observed (solid symbols) and calculated (solid lines) X-ray <lb/>diffraction pattern for La 0.8 Ca 0.2 MnO 3 sample. Positions for the <lb/>Bragg reflections are markedly vertical bars. Differences between the <lb/>observed and the calculated intensities are shown at the bottom of the <lb/>figure <lb/></body>
14
+
15
+ <page>376 <lb/></page>
16
+
17
+ <note place="headnote">J Nanostruct Chem (2015) 5:375-382 <lb/></note>
18
+
19
+ 123 <lb/>
20
+
21
+ <body>system (LCMO), Mn shows a mixed valence with Mn 3? / <lb/>Mn 4? ratio close to 4 ([Mn 4? ] = 0.2 and [Mn 3? ] = 0.8) <lb/>with a valence of ?3 for La. The partial substitution of the <lb/>Mn ions by transition metal ions (Fe) in La 0.8 Ca 0.2 Mn 1-x-<lb/>Fe x O 3 manganites gives rise to changes in the Mn 3? /Mn 4? <lb/>rate, and some Mn 3? -O 2--Mn 4? networks are substituted <lb/>by Fe 3? -O 2--Mn 4? . This causes a disorder of the charge <lb/>transfer mechanism. Such disorder causes a change in the <lb/>Mn-O distances and Mn-O-Mn angles. Consequently, the <lb/>distortion of the ideal perovskite structure in which the <lb/>Mn-O-Mn angle is equal to 180°undergoes a modifica-<lb/>tion. These results are similar to those obtained by Othmani <lb/>et al. [18]. <lb/>Magnetic properties <lb/>To study the effect of substitution of iron in manganese <lb/>sites on the magnetic properties, we have analyzed the <lb/>magnetization variation versus temperature of La 0.8 Ca 0.2-<lb/>Mn 1-x Fe x O 3 (x = 0, 0.01, 0.15, 0.2) samples under an <lb/>applied magnetic field of H = 0.05 T (Fig. 3). The <lb/>M (T) curves reveal that all samples exhibit a ferromag-<lb/>netic (FM)-paramagnetic (PM) transition at Curie <lb/>temperature, T C = 223, 205, 114, and 70 K, for x = 0, <lb/>0.01, 0.15, 0.2, respectively. The Curie temperature T C , <lb/>defined as the peak of dM/dT in the M (T) curves, is <lb/>reported for all compositions in Table 3. This table gives <lb/>evidence that T C and the magnetization are sensitive to Fe <lb/>content. Indeed, the increase in Fe content causes an <lb/>increase in T C accompanied by a reduction of the magne-<lb/>tization. Probably, both changes are attributed to the <lb/>competition between the superexchange (Mn 4? -O-Mn 4? ) <lb/>and double-exchange (Mn 3? -O-Mn 4? ) interactions. The <lb/>Fe takes place at the Mn site as Fe 3? (replacement of some <lb/>Mn 3? -O-Mn 4? bonds by Mn 4? -O-Fe 3? bonds), giving <lb/>rise to an antiferromagnetic coupling between Mn and Fe <lb/>ions that favors the superexchange mechanism. The evo-<lb/>lution of magnetization (M) versus the applied magnetic <lb/>field (l 0 H) for x = 0, 0.01, 0.15, and 0.2 samples, obtained <lb/>at different temperatures and measured under applied <lb/>magnetic field ranging from 1 to 5 T, is shown in Fig. 4. <lb/>These curves show that, below the Curie temperature, the <lb/>magnetization greatly increases with the magnetic field and <lb/>the saturated M is reached at H = 1 T. For T [ T C , the <lb/>variation of M (T, l 0 H) does not reach the saturation and a <lb/>linear behavior appears. This result confirms that all sam-<lb/>ples present a typical ferromagnetic behavior. <lb/>Figure 5 presents the magnetization measurements per-<lb/>formed at 4 K under applied magnetic fields of up to 6 T, <lb/>for La 0.8 Ca 0.2 Mn 1-x Fe x O 3 (x = 0, 0.01, 0.15, 0.2) samples. <lb/>Table 2 lists the experimental and the calculated magnetic <lb/>moments per Mn ion, denoted by M Exp <lb/>Sat and M Theo <lb/>Sat , <lb/>respectively. The values of M Theo <lb/>Sat have been calculated by <lb/>considering that the spins of all Mn and Fe ions are aligned. <lb/>The magnetic moment of La 3þ <lb/>0:8 Ca 2þ <lb/>0:2 Mn 1Àx Fe x <lb/>ð <lb/>Þ 3þ <lb/>0:8 Mn 4þ <lb/>0:2 O 3 <lb/>is expressed as <lb/>M Theo <lb/>Sat ¼ 4 Â 0:8 À 0:8 Â x <lb/>ð <lb/>Þ À 5 Â 0:8 Â x <lb/>ð <lb/>Þþ3 Â 0:2 <lb/>ð Þ <lb/>½ <lb/>l B <lb/>¼ 3:8 À 7:2 Â x <lb/>ð <lb/>Þ l B : <lb/>ð1Þ <lb/>The magnetic moments of Mn 3? , Mn 4? , and Fe 3? ions <lb/>are l Mn <lb/>3? = 4l B , l Mn <lb/>4? = 3l B , and l Fe <lb/>3? = 5l B , respectively. <lb/>The x is the iron concentration and l B is the Bohr mag-<lb/>neton. We note that the magnetization saturation values <lb/>M Sat decrease with increasing Fe content, especially for <lb/>x = 0.2. It is worth noting that similar results are reported <lb/>in [19]. The difference between the measured and the <lb/>calculated values should be explained by the presence of a <lb/>magnetic inhomogeneity or by spin-canted state at low <lb/>temperature. <lb/>Arrott curve <lb/>To determine the nature of magnetic transition type (first or <lb/>second order), we have considered the experimental <lb/>Fig. 3 Magnetization versus temperature for La 0.8 Ca 0.2 Mn 1-x Fe x O 3 <lb/>(x = 0, 0.01, 0.15, 0.2) samples under 0.05 T <lb/>Table 1 Refined structural parameters of La 0.8 Ca 0.2 Mn 1-x Fe x O 3 <lb/>(x = 0, 0.01, 0.15, 0.2) compounds <lb/>Samples <lb/>x = 0 <lb/>x = 0.01 <lb/>x = 0.15 <lb/>x = 0.2 <lb/>a (A ˚) <lb/>5.46936 <lb/>5.46979 <lb/>5.47419 <lb/>5.47565 <lb/>b (A ˚) <lb/>7.7337 <lb/>7.73569 <lb/>7.74508 <lb/>7.74596 <lb/>c (A ˚) <lb/>5.4935 <lb/>5.49399 <lb/>5.49581 <lb/>5.49606 <lb/>V (A ˚3) <lb/>232.366 <lb/>232.465 <lb/>233.012 <lb/>233.119 <lb/>v 2 <lb/>1.47 <lb/>1.4 <lb/>1.14 <lb/>1.12 <lb/>Rf <lb/>5.73 <lb/>6.32 <lb/>11.5 <lb/>6.8 <lb/></body>
22
+
23
+ <note place="headnote">J Nanostruct Chem (2015) 5:375-382 <lb/></note>
24
+
25
+ <page>377 <lb/></page>
26
+
27
+ 123 <lb/>
28
+
29
+ <body>criterion given by Banerjee [20]. It consists in inspecting <lb/>the slope of isotherm plots of l 0 H/M versus M 2 . According <lb/>to this criterion, magnetic transition is of second order if all <lb/>the curves have positive slopes, while, if some of these <lb/>curves show a negative slope, the transition is first order. <lb/>Figure 6 shows the isotherm M 2 versus l 0 H/M above and <lb/>below T C for La 0.8 Ca 0.2 Mn 1-x Fe x O 3 (x = 0, 0.01, 0.15) <lb/>Fig. 4 Isothermal magnetization M (H) for La 0.8 Ca 0.2 Mn 1-x Fe x O 3 samples at different temperatures: a x = 0, b x = 0. 01, c x = 0.15, and <lb/>d x = 0.2 <lb/>Fig. 5 Magnetization versus applied magnetic field at 4 K for <lb/>La 0.8 Ca 0.2 Mn 1-x Fe x O 3 (x = 0, 0.01, 0.15, 0.2) samples <lb/>Table 2 Experimental and theoretical saturated magnetization <lb/>moment <lb/>Samples <lb/>M Exp <lb/>Sat ðl B Þ <lb/>M Theo <lb/>Sat ðl B Þ <lb/>x = 0 <lb/>3.75 <lb/>3.8 <lb/>x = 0.01 <lb/>3.6 <lb/>3.728 <lb/>x = 0.15 <lb/>2.3 <lb/>2.72 <lb/>x = 0.2 <lb/>0.36 <lb/>2.36 <lb/></body>
30
+
31
+ <page>378 <lb/></page>
32
+
33
+ <note place="headnote">J Nanostruct Chem (2015) 5:375-382 <lb/></note>
34
+
35
+ 123 <lb/>
36
+
37
+ <body>samples. Based on the of Banerjee&apos;s criterion, the LCM <lb/>and LCMF 0.01 systems exhibit a first-order ferromagnetic-<lb/>to-paramagnetic phase transition, whereas a second-order <lb/>transition is confirmed for LCMF 0.15 and LCMF 0.2 . <lb/>Magnetocaloric study <lb/>The MCE is defined as the heating or cooling of a magnetic <lb/>material due to the application or suppression of a magnetic <lb/>field, respectively. To estimate the magnetocaloric effect, <lb/>the change of magnetic entropy (DS M ) was calculated <lb/>numerically using the equation [21]: <lb/>ÀD S M <lb/>ð Þ ¼ <lb/>X ðM i À M iþ1 <lb/>T iþ1 À T i <lb/>ÞDH i : <lb/>ð2Þ <lb/>The M i and M i?1 are the experimental values of mag-<lb/>netization measured at temperatures T i and T i?1 , respec-<lb/>tively. The H i is the applied magnetic field. The magnetic <lb/>entropy change (DS M ) determined numerically using <lb/>Eq. (2) and the M (T, l 0 H) curves are shown in Fig. 7. The <lb/>(DS M ) value increases with temperature increase to reach a <lb/>maximum near T C and lowers above T c . To compare our <lb/>results with previously published data for other perovskite <lb/>manganites, we listed in Table 3 the data of several mag-<lb/>netic materials that could be used as magnetic refrigerants. <lb/>Also, the maximum magnetic entropy change of Fe-doped <lb/>manganites increases gradually with increasing applied <lb/>magnetic field for such materials. We noted that the max-<lb/>imum entropy change DS max <lb/>M <lb/>corresponding to a magnetic <lb/>field variation of 5 T for La 0.8 Ca 0.2 MnO 3 and La 0.8 Ca 0.2-<lb/>Mn 0.99 Fe 0.01 O 3 is about 4.42 and 4.32 J/(kg K), <lb/>respectively. <lb/>In Table 3, we compared our performances of MCE <lb/>with those of Gd [2]-based materials as well as rare earth <lb/>manganites. The highest value of the magnetic entropy <lb/>change for La 0.8 Ca 0.2 MnO 3 and La 0.8 Ca 0.2 Mn 0.99 Fe 0.01 O 3 <lb/>samples is observed with x = 0 content and is equal to 1.96 <lb/>and 4.42 J/(kg K) under magnetic fields of 1 and 5 T, <lb/>respectively. In addition, similar results were observed by <lb/>Shaobo Xi et al. for La 0.8 Ca 0.2 MnO 3 [25] and by S. <lb/>Fig. 6 Arrott curves M 2 versus l 0 H/M for La 0.8 Ca 0.2 Mn 1-x Fe x O 3 samples: a x = 0, b x = 0.01, c x = 0.15, and d x = 0.2 <lb/></body>
38
+
39
+ <note place="headnote">J Nanostruct Chem (2015) 5:375-382 <lb/></note>
40
+
41
+ <page>379 <lb/></page>
42
+
43
+ 123 <lb/>
44
+
45
+ <body>Ghodhbane et al. for Pr 0.8 Ba 0.2 MnO 3 compounds under <lb/>applied magnetic fields of 3 and 1 T, respectively. These <lb/>values are lower than that of pure Gd [2.8 J/(kg K)] in a <lb/>magnetic field change of 1 T [2] ) and Gd 5 (SixGe 1-x ) 4 <lb/>system [9] which have been considered as good magnetic <lb/>refrigerants. For LCMFe 0.15 and LCMFe 0.2 , the maximum <lb/>value of magnetic entropy change, DS max <lb/>M <lb/>; is 1.6 and <lb/>0.54 K/(kg K) under a magnetic field of 5 T, respectively. <lb/>Similar results have been reported for La 0.7 Ca 0.15 Sr 0.15-<lb/>Mn 0.9 Fe 0.1 O 3 [18] and La 0.63 Ca 0.33 Mn 0.8 Fe 0.2 O 3 [30]. <lb/>The temperature dependence of the DS M upon the <lb/>magnetic applied field changes of 5 T is shown in Fig. 8. <lb/>These curves reveal that the La 0.8 Ca 0.2 Mn 1-x Fe x O 3 (x = 0, <lb/>0.01) samples present large magnetic entropy change and <lb/>that DS M decreases when increasing the Fe content (x). <lb/>This behavior is understood as the reduction of the double-<lb/>exchange mechanism between Mn 3? and Mn 4? ions for <lb/>La 0.8 Ca 0.2 Mn 1-x Fe x O 3 samples when x increases. <lb/>Relative cooling power (RCP) <lb/>Another useful parameter which examines the efficiency of <lb/>a magnetocaloric material is the RCP or the refrigerant <lb/>capacity. It expresses the heat transfer between the hot and <lb/>the cold reservoirs during an ideal refrigeration cycle. This <lb/>is defined as the product of peak value of change in the <lb/>magnetic entropy and the full width at half maximum <lb/>(FWHM) of DS M versus T curve (Fig. 8, inset) [31]. <lb/>RCP ¼ ÀDS max <lb/>M Â dT FWHM <lb/>ð3Þ <lb/>We have represented in Fig. 9 the variation of the RCP <lb/>factor as a function of the applied magnetic field. The RCP <lb/>values exhibit a linear rise with increasing field for LCMO <lb/>and LCMO 0.01 samples. Under the influence of an applied <lb/>field of 5 T, the RCP values are found to be 146 and <lb/>116 J/kg for the samples x = 0 and 0.01, respectively. <lb/>Similar RCP values at 5 T (RCP = 140 J/kg) have been <lb/>Fig. 7 Magnetic entropy change versus temperature for La 0.8 Ca 0.2 Mn 1-x Fe x O 3 samples at several magnetic applied field changes: a x = 0, <lb/>b x = 0.01, c x = 0.15, and d x = 0.2 <lb/></body>
46
+
47
+ <page>380 <lb/></page>
48
+
49
+ <note place="headnote">J Nanostruct Chem (2015) 5:375-382 <lb/></note>
50
+
51
+ 123 <lb/>
52
+
53
+ <body>reported for La 0.8 Ca 0.2 MnO 3 [26] and La 0.7 Ca 0.15 Sr 0.15 <lb/>Mn 0.9 Fe 0.1 O 3 samples [24]. Another interesting feature in <lb/>the MCE plot is its asymmetric shape, especially under <lb/>high field. Similar behavior is observed in Fe-substituted <lb/>lanthanum calcium manganite [30]. For comparison, the <lb/>maximum magnetic entropy change, the Curie temperature, <lb/>and the relative magnetic cooling efficiency of several <lb/>manganese perovskites considered useful for room-tem-<lb/>perature magnetic refrigerators are summarized in Table 3. <lb/>Thus, due to the high DS M and RCP values, our compounds <lb/>with x = 0 and 0.01 could be considered as active mag-<lb/>netic refrigerants for near-room-temperature magnetic <lb/>refrigeration. <lb/>Fig. 8 Temperature dependence of the magnetic entropy change <lb/>under an applied magnetic field of 5 T for La 0.8 Ca 0.2 Mn 1-x Fe x O 3 <lb/>compounds <lb/>Fig. 9 Variation of the relative cooling power as a function of the <lb/>applied magnetic field for La 0.8 Ca 0.2 Mn 1-x Fe x O 3 (x = 0 and <lb/>x = 0.01) compound <lb/>Table 3 Maximum entropy change DS max <lb/>M <lb/>and relative cooling power (RCP) for La 0.8 Ca 0.2 Mn 1-x Fe x O 3 (x = 0, 0.01, 0.15, 0.2) samples, <lb/>occurring at the Curie temperature (T C ) under magnetic field variations, and compared to several materials considered for magnetic refrigeration <lb/>Composition <lb/>T C (K) <lb/>DS L (J/K kg) <lb/>l 0 H (T) <lb/>RCP (J/kg) <lb/>References <lb/>Gd <lb/>293 <lb/>9.5 <lb/>5 <lb/>410 <lb/>[2] <lb/>La 0.8 Ca 0.2 MnO 3 <lb/>223 <lb/>4.42 <lb/>5 <lb/>165 <lb/>Our work <lb/>La 0.8 Ca 0.2 Mn 0.99 Fe 0.01 O 3 <lb/>205 <lb/>4.32 <lb/>5 <lb/>116 <lb/>Our work <lb/>La 0.8 Ca 0.2 Mn 0.85 Fe 0.15 O 3 <lb/>114 <lb/>1.6 <lb/>5 <lb/>-<lb/>Our work <lb/>La 0.8 Ca 0.2 Mn 0.8 Fe 0.2 O 3 <lb/>70 <lb/>0.52 <lb/>5 <lb/>-<lb/>Our work <lb/>La 0.8 Ca 0.2 MnO 3 (annealed at 800 °C) <lb/>241 <lb/>8.1 <lb/>5 <lb/>[22] <lb/>La 0.8 Ca 0.2 MnO 3 (polycrystalline, annealed at 1200 °C) <lb/>183 <lb/>2.23 <lb/>2 <lb/>112.36 <lb/>[23] <lb/>La 0.8 Ca 0.2 MnO 3 (single crystal) <lb/>176 <lb/>3.67 <lb/>1.5 <lb/>99.09 <lb/>[24] <lb/>La 0.8 Ca 0.2 MnO 3 <lb/>[25] <lb/>D = 17 nm <lb/>0.6 <lb/>4.5 <lb/>140 <lb/>D = 28 nm <lb/>4.2 <lb/>4.5 <lb/>350 <lb/>D = 43 nm <lb/>8.63 <lb/>4.5 <lb/>225 <lb/>Pr 0.8 Ba 0.2 MnO 3 <lb/>295 <lb/>4.15 <lb/>5 <lb/>230 <lb/>[26] <lb/>La 0.8 Cd 0.2 MnO 3 <lb/>155 <lb/>1.01 <lb/>1.35 <lb/>32 <lb/>[27] <lb/>Pr 0.8 Pb 0.2 MnO 3 <lb/>175 <lb/>2.64 <lb/>1.35 <lb/>55 <lb/>[28] <lb/>La 0.8 Ca 0.2 Mn 0.95 Fe 0.05 O 3 <lb/>233 <lb/>3 <lb/>5 <lb/>238 <lb/>La0. 67 Ca 0.33 Mn 0.85 Fe 0.15 O 3 <lb/>147 <lb/>3.21 <lb/>5 <lb/>-<lb/>[29] <lb/>La 0.7 Ca 0.15 Sr 0.15 Mn 0.9 Fe 0.1 O 3 <lb/>225 <lb/>1.7 <lb/>5 <lb/>118 <lb/>[30] <lb/>La 0.63 Ca 33 Mn 0.8 Fe 0.2 O 3 <lb/>92 <lb/>0.3 <lb/>5 <lb/>-<lb/>[31] <lb/></body>
54
+
55
+ <note place="headnote">J Nanostruct Chem (2015) 5:375-382 <lb/></note>
56
+
57
+ <page>381 <lb/></page>
58
+
59
+ 123 <lb/>
60
+
61
+ <body>Conclusion <lb/>We have studied the structural, magnetic, and magne-<lb/>tocaloric properties of the Fe-doped manganite perovskite <lb/>La 0.8 Ca 0.2 Mn 1-x Fe x O 3 compounds with 0 B x B 0.2. The <lb/>results show that the samples crystallize in the orthorhombic <lb/>structure with Pnma space group. The magnetic properties <lb/>reveal that all samples exhibit a paramagnetic-ferromag-<lb/>netic transition when temperature decreases. From magne-<lb/>tocaloric study, the LCMO and LCMF 0.01 O samples have a <lb/>large magnetic entropy change. The maximum values of the <lb/>magnetic entropy changes decrease with the increase of Fe <lb/>concentration. In addition, due to the high DS M and RCP <lb/>values, Fe-doped manganite perovskite samples can be <lb/>considered a potential refrigerant for use in near-room-<lb/>temperature magnetic refrigeration. <lb/></body>
62
+
63
+ <div type="acknowledgement">Acknowledgments This study has been supported by the Tunisian <lb/>Ministry of Scientific Research and Technology and the Neel Institute <lb/>at Grenoble, France. <lb/>Open Access This article is distributed under the terms of the Crea-<lb/>tive Commons Attribution 4.0 International License (http://creative <lb/>commons.org/licenses/by/4.0/), which permits unrestricted use, <lb/>distribution, and reproduction in any medium, provided you give <lb/>appropriate credit to the original author(s) and the source, provide a link <lb/>to the Creative Commons license, and indicate if changes were made. <lb/></div>
64
+
65
+ <listBibl>References <lb/>1. Warburg, E.: Magnetische untersuchung. Ann. Phys. 13, 141-164 <lb/>(1881) <lb/>2. Gschneidner, K.A., Pecharsky, V.K., Tsokol, A.O.: Recent <lb/>developments in magnetocaloric materials. Rep. Prog. Phys. 68, <lb/>1479-1539 (2005) <lb/>3. Bruck, E.: Developments in magnetocaloric refrigeration. J. Phys. <lb/>D Appl. Phys. 38, R381-R391 (2005) <lb/>4. Vitalij, K., Pecharsky, V.K., Gschneidner Jr, K.A.: Magne-<lb/>tocaloric effect and magnetic refrigeration. J. Magn. Magn. <lb/>Mater. 200, 4456 (1999) <lb/>5. Phan, M.H., Yu, S.C.: Review of the magnetocaloric effect in <lb/>manganite materials. J. Magn. Magn. Mater. 308, 325-340 (2007) <lb/>6. Tishin, A.M., Spichkin, I.: The Magnetocaloric Effect and Its <lb/>Applications. Institute of Physics Publishing, Bristol (2003) <lb/>7. Gschneidner Jr, K.A., Pecharsky, V.K.: Advanced magne-<lb/>tocaloric materials: what does the future hold? Int. J. Refrig 29, <lb/>1239-1249 (2006) <lb/>8. Atalay, S., Gencer, H., Kolat, V.S.: Magnetic entropy change in <lb/>Fe 74-x Cr x Cu 1 Nb 3 Si 13 B 9 (x= 14 and 17) amorphous alloys. <lb/>J. Non-Cryst. Solids 351(30), 2373-2377 (2005) <lb/>9. Pecharsky, V.K., Gschneidner Jr, K.A.: Giant magnetocaloric <lb/>effect in Gd 5 (Si 2 Ge 2 ). Appl. Phys. Lett. 70, 3299 (1997) <lb/>10. Fujieda, S., Fujita, A., Fukamichi, K.: Large magnetocaloric <lb/>effect in La(Fe x Si 1-x ) 13 itinerant-electron metamagnetic com-<lb/>pounds. Appl. Phys. Lett. 81, 1276 (2002) <lb/>11. Tegus, Q., Bruck, E., Buschow, K.H.: Boer, F. R.: Transition-<lb/>metal-based magnetic refrigerants for room-temperature appli-<lb/>cations. Nature 415, 150 (2002) <lb/>12. Wang, F.W., Zhang, X.X., Hu, F.X.: Large magnetic entropy <lb/>change in TbAl 2 and (Tb 0.4 Gd 0.6 )Al 2 . Appl. Phys. Lett. 77, 1360 <lb/>(2000) <lb/>13. Morelli, D.T., Mance, A.M., Mantese, J.V., Micheli, A.L.: <lb/>Magnetocaloric properties of doped lanthanum manganite films. <lb/>J. Appl. Phys. 79, 373-375 (1996) <lb/>14. Guo, Z.B., Du, Y.W., Zhu, J.S., Huang, H., Ding, W.P., Feng, D.: <lb/>Large magnetic entropy change in perovskite-type manganese <lb/>oxides. Phys. Rev. Lett. 78, 1142 (1997) <lb/>15. Guo, Z.B., Zhang, J.R., Huang, H., Ding, W.P., Du, Y.W.: Lattice <lb/>effect in Pr doped La Sr Mn O perovskite. Solid State Commun. <lb/>100, 769-771 (1996) <lb/>16. Hueso, L.E., Sande, P., Miguens, D.R., Rivas, J., Rivadulla, F., <lb/>Lopez-Quintela, M.A.: Tuning of the magneto-caloric effect in <lb/>La 0.67 Ca 0.33 MnO 3-d nanoparticles synthesized by sol-gel tech-<lb/>niques. J. Appl. Phys. 91, 9943 (2002) <lb/>17. Rodrigez-Carjaval, J.: XVth congess of the international union of <lb/>crystallography. In: Proceedings of the satellite meeting on <lb/>powder diffraction, vol. 127.Toulouse (1990) <lb/>18. Othmani, S., Blel, R., Bejar, M., Sajieddine, M., Dhahri, E., Hlil, <lb/>E.K.: New complex magnetic materials for an application in <lb/>Ericsson refrigerator. J. Solid State Commun. 149, 969-972 <lb/>(2009) <lb/>19. Issaoui, F., Tlili, M.T., Bejar, M., Dhahri, E., Hlil, E.K.: Struc-<lb/>tural and magnetic studies of Ca 2-x Sm x MnO compounds <lb/>(x = 0-0.4). J. Supercond. Novel. Mag. 25(4), 1169-1175 (2012) <lb/>20. Banerjee, S.K.: On a generalised approach to first and second <lb/>order magnetic transitions. Phys. Lett. 12, 16-17 (1964) <lb/>21. Foldeaki, M., Chahine, R., Gopal, B.R., Bose, T.K.: Investigation <lb/>of the magnetic properties of the Gd 1-x Er x alloy system in <lb/>the x \ 0.62 composition range. J. Magn. Magn. Mater. 150(3), <lb/>421-429 (1995) <lb/>22. Nisha, P.: Pilla, i S.S., Varma, M.R., Surech, K.G.: Critical <lb/>behavior and magnetocaloric effect in La 0.67 Ca 0.33 Mn 1-x Cr x O 3 <lb/>(x = 0.1, 0.25). Solid State Sci. 14, 40-47 (2012) <lb/>23. Khlifi, M., Bejar, M., Sadek, O.E.L., Dhahri, E., Ahmed, M.A., <lb/>Hlil, E.K.: Structural, magnetic and magnetocaloric properties of <lb/>the lanthanum deficient in La 0.8 Ca 0.2-xÁx MnO 3 (x = 0-0.20) <lb/>manganites oxides. J. Alloy. Compd. 509, 7410-7415 (2011) <lb/>24. Phan, M.H., Phan, V.T., Yu, S.C., Rhee, J.R., Hur, N.H.: <lb/>Excellent magnetocaloric properties of La 0.7 Ca 0.3-x Sr x MnO 3 <lb/>(0.05-x-0.25) single crystals. Appl. Phys. Lett. 86, 072504 (2005) <lb/>25. Xi, S., Lu, W., Sun, Y.: Magnetic properties and magnetocaloric <lb/>effect of La0Á8Ca0Á2MnO 3 nanoparticles tuned by particle size. <lb/>J. Appl. Phys. (2012). doi:10.1063/1.3699037 <lb/>26. Ghodhbane, S., Dhahri, A., Dhahri, N., Hlil, E.K., Dhahri, J.: <lb/>Structural, magnetic and magnetocaloric properties of La 0.8-<lb/>Ba 0.2 Mn 1-x Fe x O 3 compounds with 0 O x O 0.1. J. Alloy. <lb/>Compd. 550, 358-364 (2013) <lb/>27. Gschneidner, K.A., Pecharsky, V.K.: Magnetocaloric materials. <lb/>Annu. Rev. Mater. Sci. 30, 387 (2000) <lb/>28. Phan, M.H., Peng, H.X., Yu, S.C., Hanh, D.T., Tho, N.D., Chau, <lb/>N.: Structure, magnetic, magnetocaloric and magnetoresistance <lb/>properties of Pr 1-x PbxMnO 3 perovskites. J. Appl. Phys. Q108, <lb/>9908 (2006) <lb/>29. Nisha, P., Pillai, S.S., Vacma, M.R., Surech, K.G.: Influence of <lb/>cobalt on the structural, magnetic and magnetocaloric properties of <lb/>La 0.67 Ca 0.33 MnO 3 . J. Magn. Magn. Mater. 327, 189-195 (2013) <lb/>30. Nisha, P., Pillai, S.S., Darbandi, A., Varma, M.R., Suresh, K.G., <lb/>Hahn, H.: Critical behavior and magnetocaloric effect in nano <lb/>crystalline La 0.67 Ca 0.33 Mn 1-x FexO 3 (x = 0.05, 0.2) synthesized <lb/>nebulized spray pyrolysis. Mater. Chem. Phys. 136(1), 66-74 (2012) <lb/>31. Kamilov, I.K., Gamzatov, A.G., Aliev, A.M., Batdalov, A.B., <lb/>Aliverdiev, A.A., Abdulvagidov, ShB, Melnikov, O.V., Gor-<lb/>benko, OYu., Kaul, A.R.: Magnetocaloric effect in La 1-x Ag y-<lb/>MnO 3 (y B x): direct and indirect measurements. J. Phys. D 40, <lb/>4413 (2007) <lb/></listBibl>
66
+
67
+ <page>382 <lb/></page>
68
+
69
+ <note place="headnote">J Nanostruct Chem (2015) 5:375-382 <lb/></note>
70
+
71
+ 123
72
+
73
+
74
+ </text>
75
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/FrClassTSNM.training.segmentation.tei.xml ADDED
@@ -0,0 +1,48 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei>
3
+ <teiHeader>
4
+ <fileDesc xml:id="__FrClassTSNM"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front> Manuscript submitted to <lb/>Website: http://AIMsciences.org <lb/>AIMS&apos; Journals <lb/>Volume X, Number 0X, XX 200X <lb/> pp. X–XX <lb/> AN ATTEMPT AT CLASSIFYING HOMOGENIZATION-BASED NUMERICAL <lb/>MATHODS <lb/> Emmanuel Frénod <lb/> Universit´é de Bretagne-Sud, UMR 6205, LMBA, F-56000 Vannes, France <lb/>AND <lb/>Projet INRIA Calvi, Université de Strasbourg, IRMA, <lb/>7 rue René Descartes, F-67084 Strasbourg Cedex, France <lb/> (Communicated by the associate editor name) <lb/> Abstract. In this note, a classification of Homogenization-Based Numerical Methods and (in <lb/>particular) of Numerical Methods that are based on the Two-Scale Convergence is done. In this <lb/>classification stand: Direct Homogenization-Based Numerical Methods; H-Measure-Based Nu-<lb/>merical Methods; Two-Scale Numerical Methods and TSAPS: Two-Scale Asymptotic Preserving <lb/>Schemes. <lb/> </front>
8
+
9
+ <body> 1. Introduction. A Homogenization-Based Numerical Method is a numerical method that incor-<lb/>porates in its conception concepts coming from Homogenization Theory. Doing this gives to the <lb/>built method the capability to tackle efficiently heterogeneities or oscillations. This approach can be <lb/>applied to problems occurring in a heterogeneous medium, that have oscillating boundary conditions <lb/>or that are constrained to oscillate by an external action (for instance a magnetic field on a charged <lb/>particle cloud). <lb/>This topic is currently active. The goal of this special issue is to emphasis recent advances in this <lb/>topic in a wide variety of application fields. <lb/>This introductory paper introduces a classification of Homogenization-Based Numerical Methods, <lb/>in which stand: Direct Homogenization-Based Numerical Methods; H-Measure-Based Numerical <lb/>Methods; Two-Scale Numerical Methods and TSAPS: Two-Scale Asymptotic Preserving Schemes. <lb/>2. Direct Homogenization-Based Numerical Methods. The context of Direct Homogeniza-<lb/>tion-Based Numerical Methods is depicted in the next diagram: <lb/> u ε solution to <lb/> O ε u ε = 0 <lb/> ε → 0 <lb/> / / <lb/> u solution to <lb/> O u = 0 <lb/> u∆z solution to <lb/> O ∆z u ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> ∆z → 0 <lb/> O O <lb/> ∆z → 0 <lb/> O O <lb/> ∆z → 0 <lb/> O O <lb/> (2.1) <lb/> </body>
10
+
11
+ <front>2000 Mathematics Subject Classification. Primary: 65L99, 65M99, 65N99. <lb/> Key words and phrases. Homogenization-Based Numerical Mathods; Homogenization; Asymptotic Analysis; As-<lb/>ymptotic Expansion; Numerical Simulation. <lb/></front>
12
+
13
+ <page> 1 <lb/></page>
14
+
15
+ <note place="headnote">2 <lb/> EMMANUEL FR <lb/>ENOD <lb/></note>
16
+
17
+ <body>It is when we face with an operator O ε that generates in solution u ε of equation O ε u ε = 0 oscillations <lb/>or heterogeneities of characteristic size ε -which is small -and when it is known that, in some sense, <lb/>for small ε, u ε (z) is close to u(z) for which is known a well-posed problem O u = 0. <lb/>In this context, it is possible, in place of building a numerical approximation of operator O ε , to <lb/>build a numerical operator O ∆z approximating O. Then solving O ∆z u ∆z gives a solution u ∆z (z) <lb/>which is close to u and consequently to u ε (z), when ε is small. This approach permits to obtain an <lb/>approximation of u ε (z) without resolving the oscillations the model to compute it contains. <lb/>In the case when a corrector result is known, i.e, if in association with u(z), a function u 1 (z), <lb/>solution to well-posed equation O 1 u 1 = 0, is such that u ε (z) is close to u(z) + εu 1 (z) for small ε, it <lb/>is possible build two numerical operators O ∆z and O 1 <lb/> ∆z that are discretizations of O and O 1 . Using <lb/>them, we can compute approximated solutions u ∆z (z) and u 1 <lb/> ∆z (z) of u(z) and u 1 (z) and obtain a <lb/>good approximation of u ε (z) computing u ∆z (z) + εu 1 <lb/> ∆z (z). Such a method is called order-1 Direct <lb/>Homogenization-Based Numerical Methods and is illutrated by the following diagram. <lb/> u ε solution to <lb/> O ε u ε = 0 <lb/> ε → 0 <lb/> / / <lb/> u, u1 solutions to <lb/> O u = 0 <lb/> O 1 u 1 = 0 <lb/> u∆z, u∆z solution to <lb/> O ∆z u ∆z = 0 <lb/> O 1 <lb/> ∆z u 1 <lb/> ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> ∆z → 0 <lb/> O O <lb/> ∆z → 0 <lb/> O O <lb/> ∆z → 0 <lb/> O O <lb/> (2.2) <lb/>The paper by Legoll &amp; Minvielle [10], by Laptev [9], by Bernard, Frénod &amp; Rousseau [3], and by Xu <lb/>&amp; Yue [13] of this special issue may enter this framework <lb/>3. H-Measure-Based Numerical Methods. The context of those kind of methods is when the <lb/>transition from a microscopic scale (of size ε) to a macroscopic one (of size 1) -with quantities of <lb/>interest at the microscopic scale that are not the same as the quantities of interest at the macroscopic <lb/>scale -needs to be described. This occurs for instance in the simulation of phenomena some parts <lb/>of which call upon quantum description or in the simulation of turbulence. This context can be <lb/>represented by the following diagram: <lb/> u ε solution to <lb/> O ε u ε = 0 <lb/> e ε = E(u ε ) <lb/> ε → 0 <lb/> / / <lb/> e solution to <lb/> M e = 0 <lb/> u ε <lb/> ∆z solution to <lb/> M ε <lb/> ∆z e ε <lb/> ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> ε → 0 <lb/> / / <lb/> u∆z solution to <lb/> M ∆z e ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> ∆z → 0 <lb/> O O <lb/> ∆z → 0 <lb/> O O <lb/> ∆z → 0 <lb/> O O <lb/> (3.1) <lb/>and explained as follows. The part in the top left of diagram (3.1) symbolizes a problem which is <lb/>set at the microscopic level. This problem writes O ε u ε = 0 and generates oscillations in its solution <lb/>
18
+
19
+ <note place="headnote"> CLASSIFYING HOMOGENIZATION-BASED NUMERICAL MATHODS <lb/></note>
20
+
21
+ <page>3 <lb/></page>
22
+
23
+ u ε . Besides, the quantity that makes sense at the macroscopic level is e ε ; it is related to u ε by a <lb/>non-linear relation e ε = E(u ε ) and it is, in some sense, close to e solution to M e = 0 (see the top <lb/>right of the diagram) which represents the model at the macroscopic level. <lb/>Then, the goal of a H-Measure-Based Numerical Method consists in building a numerical operator <lb/> M ε <lb/> ∆z , giving a numerical solution e ε <lb/> ∆z close to e ε , for any ε as soon as ∆z is small (see the bottom <lb/>left of the diagram), which behaves as a numerical approximation of M when ε is small (see the <lb/>bottom right of the diagram). <lb/>The paper by Tartar [12] of this special issue lays the foundation of the theory for those kinds of <lb/>methods. <lb/>4. Two-Scale Numerical Methods. The papers by Assyr, Bai &amp; Vilmar [1], Back &amp; Frénod [2], <lb/>Faye, Frénod &amp; Seck [5], Frénod, Hirtoaga &amp; Sonnendrücker [6], Lutz [11] and Henning &amp; Ohlberger <lb/>[7] of this special issue are related to this framework of Two-Scale Numerical Methods. <lb/>An order-0 Two-Scale Numerical Method may be explained using the following diagram: <lb/> u ε solution to <lb/> O ε u ε = 0 <lb/> ε → 0 <lb/> / / <lb/> ε → 0 , two-scale ) ) <lb/> R <lb/> R <lb/>R <lb/>R <lb/>R <lb/>R <lb/>R <lb/>R <lb/> u solution to <lb/> O u = 0 <lb/> U solution to <lb/> O U = 0 <lb/> 񮽙 <lb/> Z <lb/> dζ <lb/> 6 6 <lb/> n <lb/>n <lb/>n <lb/>n <lb/>n <lb/>n <lb/> u∆z solution to <lb/> O ∆z u ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> U∆z solution to <lb/> O ∆z U ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> 񮽙 Num <lb/> Z <lb/> dζ <lb/> 6 6 <lb/> n <lb/>n <lb/>n <lb/>n <lb/>n <lb/>n <lb/> (4.1) <lb/>The context includes the one of Direct Homogenization-Based Numerical Methods and diagram (4.1) <lb/>has to be regarded as a prism. Its deepest layer is nothing but diagram (2.1). Yet, if more is known <lb/>about the asymptotic behavior of u ε , i.e. if it is known that u ε (z) is close to U (z, z <lb/> ε ), with U (z, ζ) <lb/> periodic in ζ, when ε is small (which can be translated as u ε (z) Two-Scale Converges to U (z, ζ)) and <lb/>if it is known a well posed problem O U = 0 for U (see the middle of the diagram), that gives the <lb/>equation for u (see the top right of the diagram) when integrated with respect to periodic variable <lb/> ζ, it is possible to build a specific numerical method. <lb/>This method consists in building a numerical approximation O ∆z of operator O. Using this operator <lb/>can give a numerical solution U ∆z (see the bottom of the diagram) and U ∆z (z, z <lb/> ε ) is an approxima-<lb/>tion of u ε (z) for small ε. To be consistent with the continuous level, a numerical integration of the <lb/> O ∆z U ∆z = 0 needs to yield a numerical approximation of the equation for u (see the bottom right <lb/>of the diagram). <lb/>When a little more is known concerning the asymptotic behavior of u ε when ε is small, i.e. if u ε is <lb/>close to U (z, z <lb/> ε ) + εU 1 (z, z <lb/> ε ) with U 1 (z, ζ) also periodic in ζ and if a well-posed problem is known for <lb/> U 1 , we can enrich diagram (4.1) and obtain the following diagram of order-1 Two-Scale Numerical <lb/>
24
+
25
+ <page> 4 <lb/></page>
26
+
27
+ <note place="headnote">EMMANUEL FR <lb/>ENOD <lb/></note>
28
+
29
+ Methods: <lb/> u ε solution to <lb/> O ε u ε = 0 <lb/> ε → 0 <lb/> / / <lb/> ε → 0 , two-scale ( ( <lb/> Q <lb/> Q <lb/>Q <lb/>Q <lb/>Q <lb/>Q <lb/>Q <lb/>Q <lb/>Q <lb/>Q <lb/> u, u1 solutions to <lb/> O u = 0 <lb/> O 1 u 1 = 0 <lb/> U , U 1 solutions to <lb/> O U = 0 <lb/> O 1 U 1 = 0 <lb/> 񮽙 <lb/> Z <lb/> dζ <lb/> 6 6 <lb/> n <lb/>n <lb/>n <lb/>n <lb/>n <lb/>n <lb/> u∆z, u 1 <lb/> ∆z solutions to <lb/> O ∆z u ∆z = 0 <lb/> O 1 <lb/> ∆z u 1 <lb/> ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> U∆z, U 1 <lb/> ∆z solutions to <lb/> O ∆z U ∆z = 0 <lb/> O 1 <lb/> ∆z U 1 <lb/> ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> 񮽙 Num <lb/> Z <lb/> dζ <lb/> 6 6 <lb/> n <lb/>n <lb/>n <lb/>n <lb/>n <lb/>n <lb/> (4.2) <lb/>5. TSAPS: Two-Scale Asymptotic Preserving Schemes. To describe Two-Scale Asymptotic <lb/>Preserving Schemes, it is first needed to describe what is an Asymptotic Preserving Scheme (or <lb/>AP-Scheme in short). <lb/> u ε solution to <lb/> O ε u ε = 0 <lb/> ε → 0 <lb/> / / <lb/> u solution to <lb/> O u = 0 <lb/> u ε <lb/> ∆z solution to <lb/> O ε <lb/> ∆z u ε <lb/> ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> ε → 0 <lb/> / / <lb/> u∆z solution to <lb/> O ∆z u ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> ∆z → 0 <lb/> O O <lb/> ∆z → 0 <lb/> O O <lb/> ∆z → 0 <lb/> O O <lb/> (5.1) <lb/>For this, we comment on diagram (5.1). The context is when we are face-to-face with an operator <lb/> O ε which is approached, when ε is small, by another operator O which has not the same nature as <lb/> O ε . An Asymptotic Preserving Scheme to approximate problem O ε u ε = 0 (see the top left of the <lb/>diagram) is a numerical operator O ε <lb/> ∆z that gives, when solving O ε <lb/> ∆z u ε <lb/> ∆z = 0 (see the bottom right <lb/>of the diagram) a numerical solution u ε <lb/> ∆z which is close to u, with an accuracy depending on step <lb/> ∆z and not on ε. Besides, this operator needs to mimic the behavior of an numerical approximation <lb/>(see the bottom right of the diagram) of limit problem O u = 0 (see the top right of the diagram) <lb/>as ε is small. <lb/>For an introduction to this kind of method the reader is referred to Jin [8]. <lb/>
30
+
31
+ <note place="headnote"> CLASSIFYING HOMOGENIZATION-BASED NUMERICAL MATHODS <lb/></note>
32
+
33
+ <page>5 <lb/></page>
34
+
35
+ The explanation of TSAPS, will be based on the following diagram: <lb/> u ε solution to <lb/> O ε u ε = 0 <lb/> ε → 0 <lb/> / / <lb/> ε → 0 , two-scale <lb/> + + <lb/> W <lb/> W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/>W <lb/> u, u 1 solutions to <lb/> O u = 0, O 1 u 1 = 0 <lb/> U , U 1 solutions to <lb/> O U = 0 <lb/> O 1 U 1 = 0 <lb/> 񮽙 <lb/> Z <lb/> dζ <lb/> 6 6 <lb/> m <lb/>m <lb/>m <lb/>m <lb/>m <lb/>m <lb/>m <lb/>m <lb/>m <lb/> U ε solution to <lb/> O ε U ε = 0 <lb/> ζ = <lb/> z <lb/> ε <lb/> ^ ^ &gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/> ε → 0 <lb/> 7 7 <lb/> o <lb/>o <lb/>o <lb/>o <lb/>o <lb/>o <lb/>o <lb/> u ε <lb/> ∆z solution to <lb/> O ε <lb/> ∆z u ε <lb/> ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> ε → 0 <lb/> / / <lb/> u ∆z , u 1 <lb/> ∆z solutions to <lb/> O ∆z u ∆z = 0, O 1 <lb/> ∆z u 1 <lb/> ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> U ∆z , U 1 <lb/> ∆z solutions to <lb/> O ∆z U ∆z = 0 <lb/> O 1 <lb/> ∆z U 1 <lb/> ∆z = 0 <lb/> ∆z → 0 <lb/> O O <lb/> 񮽙 Num <lb/> Z <lb/> dζ <lb/> 6 6 <lb/> m <lb/>m <lb/>m <lb/>m <lb/>m <lb/>m <lb/>m <lb/>m <lb/> U ε <lb/> ∆z solution to <lb/> O ε <lb/> ∆z U ε <lb/> ∆z = 0 <lb/> ζ = <lb/> z <lb/> ε <lb/> ^ ^ &gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/>&gt; <lb/> ∆z → 0 <lb/> O O <lb/> ε → 0 <lb/> 7 7 <lb/> o <lb/>o <lb/>o <lb/>o <lb/>o <lb/>o <lb/> (5.2) <lb/>This diagram has to be regarded as a prism with three layers. The deepest one is the diagram of <lb/>the AP-schemes. At the top left of this layer is found the equation that generates in its solution <lb/>oscillations of size ε. At the top right stands the limit problem, as ε is small. (This limit problem <lb/>is assumed to be an order-1 problem, i.e. u ε ∼ u 0 + εu 1 for ε small and equations are known for u 0 <lb/> and u 1 .) At the bottom left stands the AP-Scheme that approximate equation O ε u ε = 0 for any ε <lb/> and that mimics an approximation of the limit problem when ε is small (see the bottom right of the <lb/>layer). <lb/>The middle layer is exactly the diagram of the order-1 Two-Scale Numerical Methods. <lb/>The top layer is the new part. at the bottom, stands the TSAPS. This is a numerical method that <lb/>gives a solution U ε <lb/> ∆z which depends on two variables z and ζ. When taken in ζ = z/ε, U ε <lb/> ∆z gives <lb/>a numerical approximation of the solution to the problem given at the top right of the diagram, <lb/>with an accuracy that only depends on the discretization step ∆z, and not on ε. Moreover, as ε is <lb/>small, the TSAPS O ε <lb/> ∆z needs to mimic the behavior of the order-1 Two-Scale Numerical Operator <lb/>(the couple (O, O 1 )). To builtd a TSAPS, a reformulation of problem O ε u ε = 0 calling upon a <lb/>Two-Scale Macro-Micro Decomposition (that reads O ε U ε = 0, see the middle of the diagram) is <lb/>used. A first step towards TSAPS is led in Crouseilles, Frénod, Hirstoaga &amp; Mouton [4]. <lb/></body>
36
+
37
+ <listBibl> REFERENCES <lb/> [1] A. Abdulle, Y. Bai, and G. Vilmart. Reduced basis finite element heterogeneous multiscale method for quasilin-<lb/>ear elliptic homogenization problems. Discrete and Continuous Dynamical Systems -Serie S. Special Issue on <lb/>Numerical Methods based on Homogenization and Two-Scale Convergence, In press. <lb/>[2] A. Back and E. Frénod. Geometric two-scale convergence on manifold and applications to the Vlasov equation. <lb/> Discrete and Continuous Dynamical Systems -Serie S. Special Issue on Numerical Methods based on Homoge-<lb/>nization and Two-Scale Convergence, In press. <lb/>
38
+
39
+ <page> 6 <lb/></page>
40
+
41
+ <note place="headnote"> EMMANUEL FR <lb/>ENOD <lb/></note>
42
+
43
+ [3] J.-P. Bernard, E. Frénod, and A. Rousseau. Paralic confinement computations in coastal environment with <lb/>interlocked areas. Discrete and Continuous Dynamical Systems -Serie S. Special Issue on Numerical Methods <lb/>based on Homogenization and Two-Scale Convergence, In press. <lb/>[4] N. Crouseilles, E. Frenod, S. Hirstoaga, and A. Mouton. Two-Scale Macro-Micro decomposition of the Vlasov <lb/>equation with a strong magnetic field. Mathematical Models and Methods in Applied Sciences, 23(08):1527–1559, <lb/>November 2012. <lb/> [5] I. Faye, E. Frénod, and D. Seck. Two-scale numerical simulation of sand transport problems. Discrete and <lb/>Continuous Dynamical Systems -Serie S. Special Issue on Numerical Methods based on Homogenization and <lb/>Two-Scale Convergence, In press. <lb/>[6] E. Frénod, S. Histoaga, and E. Sonnendrücker. An exponential integrator for a highly oscillatory Vlasov equa-<lb/>tion. Discrete and Continuous Dynamical Systems -Serie S. Special Issue on Numerical Methods based on <lb/>Homogenization and Two-Scale Convergence, In press. <lb/>[7] P. Henning and M. Ohlberger. Error control and adaptivity for heterogeneous multiscale approximations of <lb/>nonlinear monotone problems. Discrete and Continuous Dynamical Systems -Serie S. Special Issue on Numerical <lb/>Methods based on Homogenization and Two-Scale Convergence, In press. <lb/>[8] S Jin. Efficient asymptotic-preserving (ap) schemes for some multiscale kinetic equations. SIAM Journal of <lb/>Scientific Computing, 21:441–454, 1999. <lb/>[9] V. Laptev. Deterministic homogenization for media with barriers. Discrete and Continuous Dynamical Systems <lb/>-Serie S. Special Issue on Numerical Methods based on Homogenization and Two-Scale Convergence, In press. <lb/>[10] F. Legoll and W. Minvielle. Variance reduction using antithetic variables for a nonlinear convex stochastic homog-<lb/>enization problem. Discrete and Continuous Dynamical Systems -Serie S. Special Issue on Numerical Methods <lb/>based on Homogenization and Two-Scale Convergence, In press. <lb/>[11] M. Lutz. Application of Lie transform techniques for simulation of a charged particle beam. Discrete and Contin-<lb/>uous Dynamical Systems -Serie S. Special Issue on Numerical Methods based on Homogenization and Two-Scale <lb/>Convergence, In press. <lb/>[12] Tartar. Multi-scales h-measures. Discrete and Continuous Dynamical Systems -Serie S. Special Issue on Nu-<lb/>merical Methods based on Homogenization and Two-Scale Convergence, In press. <lb/>[13] X. Xu, S. Yue. Homogenization of thermal-hydro-mass transfer processes. Discrete and Continuous Dynamical <lb/>Systems -Serie S. Special Issue on Numerical Methods based on Homogenization and Two-Scale Convergence, <lb/> In press. <lb/></listBibl>
44
+
45
+ <note place="footnote">E-mail address: [email protected]</note>
46
+
47
+ </text>
48
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/French_in_Auvergne.training.segmentation.tei.xml ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei>
3
+ <teiHeader>
4
+ <fileDesc xml:id="__French_in_Auvergne"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <titlePage>French in Auvergne (Centre) : a speaker from <lb/>Clermont-Ferrand <lb/> Damien Chabanal, J. Durand, Corinne Ratier <lb/> To cite this version: <lb/> Damien Chabanal, J. Durand, Corinne Ratier. French in Auvergne (Centre) : a speaker from <lb/>Clermont-Ferrand. Oxford University Press. varieties of spoken French, Oxford University <lb/>Press, pp.5-15, 2015. &lt;hal-01062222&gt; <lb/> HAL Id: hal-01062222 <lb/>https://hal.archives-ouvertes.fr/hal-01062222 <lb/> Submitted on 10 Sep 2014 <lb/> HAL is a multi-disciplinary open access <lb/>archive for the deposit and dissemination of sci-<lb/>entific research documents, whether they are pub-<lb/>lished or not. The documents may come from <lb/>teaching and research institutions in France or <lb/>abroad, or from public or private research centers. <lb/>L&apos;archive ouverte pluridisciplinaire HAL, est <lb/>destinée au dépôt et a la diffusion de documents <lb/>scientifiques de niveau recherche, publiés ou non, <lb/>´ emanant des etablissements d&apos;enseignement et de <lb/>recherche français oú etrangers, des laboratoires <lb/>publics ou privés. <lb/></titlePage>
9
+
10
+ <front> French in Auvergne (Centre): <lb/>a speaker from Clermont-Ferrand <lb/> Damien Chabanal, Jacques Durand and Corinne Ratier <lb/> A paraître dans l&apos;ouvrage « varieties of spoken french », Oxford University Press <lb/></front>
11
+
12
+ <body>0. Introduction <lb/> This extract has been chosen to illustrate the variety of French in the Auvergne region, <lb/>more precisely in Clermont-Ferrand in the Puy de Dôme department, the town where our <lb/>selected speaker was born and lives. Linguistically, Clermont-Ferrand is part of North Occitan <lb/>(see Bec 1963: 7-8) but we are not aware of any published work on the variety of <lb/>contemporary spoken French in this part of France apart from the thumbnail phonological <lb/>sketch of a Clermont-Ferrand speaker in Walter (1982: 169). Clermont-Ferrand, with a <lb/>population of over 140 000 inhabitants and its famous Michelin factories, is usually <lb/>considered as the capital of Massif Central. It is an important university town since there are <lb/>37 000 students, according to its Mairie&apos;s official website. Indeed, our chosen speaker is a <lb/>student whose French is at first sight a good example of a levelled variety (see Armstrong and <lb/>Pooley 2010) but, which on closer inspection of the phonology, still shows a number of <lb/>regional features. <lb/> 1. Sociolinguistic profile and recording situation <lb/> The speaker we focus on is a young woman MM (PFC code: 63mm1), who was 22 years <lb/>old at the time of the recording. She was born in Clermont-Ferrand in 1988. She has always <lb/>lived in the area apart from two years spent in Paris for her studying. During her stay in Paris <lb/>she did her second year correspondence course in psychology while working at the same time. <lb/>Back in the Puy de Dôme she joined a school training students for a Diplôme d&apos;État <lb/>d&apos;Éducateur Spécialisé (DEES). <lb/>The recording was made at the very beginning of her first year course and the extract is <lb/>chosen from an informal conversation between MM and her mother which took place in <lb/>September 2010 in the family house. <lb/> 2. Content analysis and lexicon <lb/> MM tells her mother about her course which just started the day before the conversation, <lb/>her teachers and their expectations. The style, outside the precise references to education and <lb/>psychology, is informal. MM starts by describing her first teacher as un mec trop bizarre (l. <lb/>1, &apos;a very strange guy&apos;) where the use of the adverb trop to mean &apos;très&apos; is typical of the <lb/>younger generations in hexagonal French. The teacher&apos;s strangeness seems to derive from his <lb/>appearance (his long hair possibly seen as a sign of marginality in a professional context) but <lb/>also from his attitude and approach (the way he walks and prods students with repeated <lb/>questions). Every word used by a student has to be justified as to what lies behind it: Et <lb/>chaque mot que tu dis, il te dit : &apos;Mais tu sous-entends quoi ?&apos; (l. 7-8). The word marrant to <lb/>characterize the whole experience means &apos;funny&apos; and like its English equivalent is ambiguous <lb/>between the sense of &apos;peculiar&apos; and that of &apos;mirth provoking&apos;. Although the mother wonders if <lb/>this teacher is a psychologist, the daughter describes him as an éduc which is short for <lb/> éducateur (l. 12), a teacher of children with special needs. The strategy of truncating common <lb/>words is very frequent in hexagonal French and also illustrated within this extract by fac (l. <lb/>
13
+
14
+ 41) short for faculté (= university in this context) and prof (used by the mother, l. 52) short for <lb/> professeur (= teacher). <lb/>If this teacher&apos;s style is somewhat idiosyncratic, he is nevertheless said to be extremely <lb/>knowledgeable - il a un sacré niveau (l. 14, lit. &apos;he has an incredible level&apos;), where the use of <lb/> sacré as an intensificator is typical of informal speech. This remark allows MM to move on <lb/>to the course as such and its contents. MM stresses repeatedly how intensive the course is and <lb/>how much is expected of the students although they have no assignments as such (on a pas de <lb/>devoirs en soi, l. 19). <lb/>With a great deal of colloquial expressions, MM explains that she has to get stuck into <lb/>completing all her index cards (il faut que je me tape à faire toutes mes fiches, l. 23-24), read <lb/>lots of books (bouquins familiar for livres, l. 37) and attend &apos;millions&apos; of classes (une tonne <lb/>de cours, l. 31) on educational and psychological matters such as la séparation (l. 58, <lb/>&apos;divorce&apos;) and l&apos;hospitalisme (l. 58, &apos;hospitalism&apos; which can be defined as a type of mental <lb/>illness which affects children who have been prematurely separated from their mothers and <lb/>institutionalised). <lb/> Throughout the conversation, MM stresses the high level achieved by their teachers. <lb/>During the afternoon session, the students had a woman (une fille, lit. &apos;a girl&apos;, l. 39) who is <lb/> super sympa (l. 39, lit. super friendly) and who was initially a teacher of children with special <lb/>needs but then went to University and became a formatrice (an educationalist involved in the <lb/>training of future teachers). This woman is described as a militante, a term used to describe <lb/>activists involved in political or civil rights movements. In this context, it is likely that the <lb/>description applies to a person involved in the feminist movement but to establish the precise <lb/>reference of the phrase one would have to know more about the &apos;political&apos; or &apos;ideological&apos; <lb/>outlook of the speaker. <lb/>Many of the lexical choices are tightly integrated into the grammar used by MM and <lb/>would deserve more comments than is possible here. The relaxed register used by MM is <lb/>often typical of her generation but many of the lexical and grammatical features are typical of <lb/>colloquial speech. Although they are often branded &apos;très familier&apos;, and even &apos;populaire&apos; or <lb/>&apos;vulgaire&apos;, it is well known that all speakers use such words or constructions in informal <lb/>everyday conversation. Two obvious examples are the use of vachement (l. 57) to mean <lb/>&apos;beaucoup&apos; or &apos;très&apos;, a well-established feature of informal French, and the frequent use of <lb/>
15
+
16
+ machin (l. 32, 41, 66), a vague noun which like English &apos;thingummy&apos; or &apos;thingamabob&apos; <lb/>allows the speaker to refer to something or someone when you cannot remember or cannot be <lb/>bothered to find the proper word or name for them. The vagueness of the referent is <lb/>emphasized by et tout meaning &apos;and all&apos; (l. 32, 41). On the other hand, in l&apos;hospitalisme <lb/>machin (l. 66), machin appears to be used as a modifier with vague reference (&apos;any type of <lb/>hospitalism&apos;). The use of se taper is particularly interesting in this extract. <lb/>In colloquial usage, se taper is usually followed by a complement noun phrase and <lb/>means &apos;to get landed with&apos; (e.g. je me suis tapé toute la vaisselle = &apos;I got landed with (doing) <lb/>all the dishes&apos;). Here, however, the verb is followed by a verb phrase introduced by à: se <lb/>taper à faire toutes les fiches = &apos;to get landed with completing all the index cards&apos;. All these <lb/>examples illustrate the difficulty of defining registers in a rigid way. However we deal with <lb/>this question, it is essential not to look at all the features of spoken French present in the <lb/>extracts selected in this volume from a prescriptive point of view. Rather than seeing them as <lb/>illustrating a vulgar level of speech full of mistakes, we must learn to see them as typical of <lb/>spoken interaction between people who know each other well (see Blanche-Benveniste 2000: <lb/>35-63). <lb/> 3. Syntactic and discursive features <lb/> As emphasized above, this extract provides a striking illustration of informal speech. For <lb/>lack of space, we will not deal with features such as hesitations (there are 28 &apos;euh&apos; in the <lb/>extract), repetitions, reformulations, unfinished sentences or self-corrections which are typical <lb/>of spontaneous spoken interaction (see Rossi-Gensane, 2010, Ch. 6). Among features worth <lb/>examining, we will start with some comments on (i) ne in negative structures, (ii) the use of <lb/>personal pronouns, (iii) the use of tenses and moods, and then move on to some general <lb/>observations regarding the structure of this extract. <lb/>i) In one interpretation, ne is systematically omitted from all negative sentences in this <lb/>extract: cf. Non il est pas psychologue (l. 10-11), tu peux pas lire b/ dix bouquins (l. 37), vous <lb/>les lisez pas tous (l. 45), Si tu veux c&apos;est pas ça (l. 51). But it should be remembered that when <lb/>the pronoun on is used immediately a vowel initial word (e.g on a rien), it is impossible <lb/>according to most authorities to know whether the speaker used ne or not: on n&apos;a rien or on a <lb/>rien. In the relevant examples in this extract (cf. on a pas de devoirs en soi, on a rien de, <lb/> &lt;LM: Ouais.) on a rien à préparer (l. 19-20)), we have deliberately chosen not to include ne <lb/> in our transcriptions in line with the other clear examples where ne was indeed systematically <lb/>absent. It can&apos;t however be excluded that the speaker intended a ne which is not audible! <lb/>(ii) Unsurprisingly MM refers to herself as je (reinforced by moi, l. 60), and mother and <lb/>daughter use tu to refer to each other. It will be noticed that in reported speech (which there is <lb/>a great deal of in the extract), the teacher appears to use tu (or the oblique form toi) to address <lb/>the student which is not universal in higher education in France: e.g. Alors pour toi qu&apos;est-ce <lb/>que veut dire éducateur (l. 3-4), Oui mais tu sous-entends quoi par accompagner ? (l. 6-7). <lb/>The use of &apos;tu&apos; by a teacher when speaking to students probably depends on factors such as <lb/>age, &apos;political&apos; or &apos;ideological&apos; stance, outlook on life and the type of higher education <lb/>establishment one is dealing with. The reader will notice that many of the tu forms used by <lb/>MM are &apos;generic&apos; in value, often interpretable as on or nous but occasionally vaguer in <lb/>reference: e.g. Et chaque mot que tu dis (l. 7). It would appear that this &apos;generic&apos; use of tu is <lb/>currently replacing vous in hexagonal French (e.g. Et chaque mot que vous dites would have <lb/>been possible). The pronoun on is also very frequent and can either be replaced by nous or has <lb/>a generic value in line with its etymology (= homme). <lb/> (iii) As in many other extracts in this book, the tenses used are the indicative present, the <lb/>imperfect, the &apos;passé composé&apos;. Note that there are no futures but that MM uses aller + Vinf : <lb/>e.g. je vais continuer à lire mes fiches régulièrement (l. 63). The subjunctive, whose death has <lb/>frequently but rather imprudently been predicted for spoken French, is well-attested. It is <lb/>systematically triggered by the impersonal verb falloir: e.g. le problème c&apos;est qu&apos;il faut que ça <lb/>soit des notions qui soient acquises, c&apos;est-à-dire qu&apos;on f/faut qu&apos;on les apprenne ces notions <lb/> (l. 20-23). The absence of dummy subject il is in fact quite frequent before faut: cf. c&apos;est-à-<lb/>dire qu&apos;on f/faut qu&apos;on les apprenne ces notions (l. 22-23), Ah mais ce livre il est <lb/>indispensable &lt;LM: (XXXX)&gt; faut absolument que vous lisiez (l. 35-36), l e s notions <lb/>principales que tu vois en cours &lt;LM: Voilà.&gt; faut que tu les aies acquises (l. 55-56), nous <lb/> faut qu&apos;on sache ce que c&apos;est l&apos;hospitalisme (l. 66-67). In faut que tu les aies acquises (l. 56), <lb/>
17
+
18
+ the reader should note the agreement form acquises (feminine plural) triggered by a direct <lb/>object complement if it precedes the auxiliary verb avoir (here les which stands for les <lb/>notions principales). This type of agreement is getting rarer and rarer in speech, even among <lb/>cultivated speakers (see Rouayrenc 2010: 150). <lb/>This shows that registers are not watertight modules and that characteristics of colloquial <lb/>speech can coexist with more formal features. <lb/>At a more general level, word-order and sentence-structure also seem to be in line with <lb/>the conversational register. Most of the utterances are statements but the few questions used <lb/>either by the speaker or her mother deserve some brief comments. The question (Alors pour <lb/>toi,) qu&apos;est-ce que veut dire éducateur? (l. 3-4) corresponds to the more formal written form <lb/> que veut dire éducateur?. It shows the spread of Qu&apos;est-ce-que into QU-questions. The QU-<lb/>element stays in situ in the other reported speech question: Oui mais tu sous-entends quoi par <lb/>accompagner? (l. 6-7), instead of the normative form Oui mais qu&apos;est-ce que tu sous-entends <lb/>par accompagner ?. As for the YES-NO question used by the mother (Et tu as des devoirs?, l. <lb/>17-18), it uses the declarative order rather than an inverted form (Et as-tu des devoirs?) as is <lb/>generally the case in hexagonal spoken French. <lb/>Many of the statements are built around a nucleus made up of a pronoun-subject + verb + <lb/>optional complement (il arrive vers toi; il s&apos;avance, tu parles, etc.) with possible internal <lb/>expansions (other pronouns, negation, etc.). But it must be remembered that these basic <lb/>predicational nuclei (e.g. je le sais, l. 59) are announced by thematically salient phrases which <lb/>can themselves be placed after discursive markers, as illustrated by an utterance like: Donc <lb/>moi déjà, tout ce que j&apos;ai, toute la séparation, l&apos;hospitalisme tout ça je le sais (l. 56-59). <lb/> A connective like donc (or alors) can be argued to provide a temporal or causal frame to <lb/>the reporting of events but other discourse markers such as tu vois (l. 1), tu sais (l. 9), si tu <lb/>veux (l. 51, 52) show how the speaker involves the addressee (her mother) by maintaining the <lb/>channel of communication and allowing her, a least in theory, to disagree with what is being <lb/>said. Many modern specialists emphasize that these discourse markers are essential to the co-<lb/>construction of discourse. <lb/>4. Phonetic and phonological features <lb/> The pronunciation of our speaker (MM) is interesting in that its segmental system (in <lb/>particular, its oral vowel system, see 4.1) is typical of southern varieties of French; on the <lb/>other hand, as far as schwa is concerned the features which we have identified (see 4.2) are <lb/>typical of northern varieties of French. <lb/>4.1 Vowel system <lb/> Let us focus first of all on the oral vowels but leave aside the high vowels /i, y, u/ as there <lb/>are no striking differences between the system used by MM and that of most other varieties of <lb/>hexagonal French. On the other hand, the non-high vowels deserve our attention. First of all, <lb/>like typical southern French speakers MM has only one /a/ phoneme. She does not distinguish <lb/>a front and a back vowel even in the reading aloud of minimal pairs such as patte vs. pâte in <lb/>the word list. Secondly, the mid-vowels are reducible to three phonemes which we will <lb/>represent here as /E/, /Ø/ and /O/. The capital letters for the underlying phonemes are assumed <lb/>to range over two values in each case [e, o, ø] in open syllables and [ɛ, oe, ɔ] in closed <lb/>syllables. This is generally referred to as the &apos;loi de position&apos; (see Ch. 13 and Coquillon &amp; <lb/>Durand 2010). It will be noticed for instance that all –ais, -ait endings which are mid-low [ɛ] <lb/>in some northern varieties are pronounced with an [e]: e.g Mais il/tu sais c&apos;est marrant <lb/>comme il fait alors il, il marchait comme ça (l. 10), disait (l. 28), était (l. 42). The full <lb/>distribution of the mid-high vs. mid-low allophones of /E/, / Ø/ and /O/ is not attested in this <lb/>extract but the reader can verify that the &apos;loi de position&apos; is respected in the word-list and the <lb/>reading aloud of the text. Thus, chaude and chose in the text are both pronounced with an [ɔ]. <lb/>The nasal vowels used by MM do not have a typical southern realization (i.e. oral or <lb/>slightly nasalised vowel followed by a nasal appendix). Nevertheless, unlike modern speakers <lb/>of Parisian French, she appears to make an opposition between /ɛ̃ / and /oẽ /: compare the <lb/>pronunciation of un in the first line of this extract with that of bien: j&apos;ai eu euh, un mec trop <lb/>bizarre mais pff bien tu vois. It also seems to us that this true of the word-list and the text: <lb/>compare brun with brin which are differentiated although not in a minimal pair-context <lb/>triggering a more artificial pronunciation. If we are correct, this speaker has four nasal <lb/>vowels: /ɛ̃ , oẽ , ɑ̃ , ɔ̃ /. <lb/>4.2 Schwa <lb/>As far as schwa is concerned, the patterns of deletion are aligned on those typical in <lb/>northern varieties of French. Starting with polysyllabic words, the word-final position <lb/>corresponds to an absence of schwa: e.g. je vais quand mêm(e) fair(e) mes fich(e)s (l. 61). In <lb/>word-internal position, unless there is a preceding complex cluster, schwa is deleted: e.g. <lb/> malheureus(e)ment (l. 24), feuill(e)tez (l. 46), vach(e)ment (l. 57), régulièr(e)ment (l. 63). <lb/>Within word-initial syllables, the deletion of schwa exhibits variability. A consonant-cluster <lb/>at the end of the preceding word blocks schwa-deletion: on a pas d(e) devoir (l. 19), je vais <lb/>jus(te) relire mon cours (l. 60). A vowel-final preceding word seems to favour deletion, <lb/>
19
+
20
+ particularly with some frequent lexical items: les ch(e)veux (l. 2), la s(e)maine (l. 25). <lb/>However, note the presence of schwa in est repartie en fac (l. 41). There are two few <lb/>examples in this conversation to extract a full generalization. <lb/>Moving to monosyllables, this speaker seems to delete schwas as a matter of routine. <lb/>There are however some blocking contexts. The maintenance of a schwa may be triggered by <lb/>discourse planning factors: e.g. hesitation (remplies de, de dates) or a position at the end of an <lb/>unfinished sentence (vous essayez de. (l. 46)). Emphasis can also favour retention despite a <lb/>vowel on the left (e.g. mais ce livre l. 35). <lb/>A consonant at the end of the preceding word may appear to block deletion (cf. une <lb/>bibliographie donc de tous les livres, l. 43), which would traditionally be described in terms <lb/>of Grammont&apos;s famous &apos;loi des trois consonnes&apos; (i.e. avoid deleting a schwa if this results in a <lb/>cluster of three consonants or more ([kdt] in the above example). But note that the nature of <lb/> the consonant is highly relevant. Whereas &apos;donc&apos; ends in a plosive [k], a final [ʁ] does not <lb/>have the same effect and schwas are absent in the following examples despite the creation of <lb/>clusters of three consonants: prévoir l(e) soir ([ʁls], l. 26), histoir(e) d(e) l&apos;éducation ([ʁdl], l. <lb/>31), sur l(e) sujet ([ʁls], l. 44). In the same way, if a monosyllable is at the beginning of a <lb/>rhythmic group the schwa may be maintained (que les notions principales, l. 50) but note the <lb/>presence here of a plosive [k]. By contrast, in common clitics beginning with a fricative such <lb/>as je or ce, deletion seems to be automatic (C(e) matin (l. 1), J(e) sais pas comment (l. 48)). <lb/>4.3 Liaison <lb/> There are few contexts in the selected extract allowing for an extensive study of liaison. <lb/>Categorical liaisons as defined in the PFC project (see Ch. 28) are realized. Here the examples <lb/>are reduced to liaison between a clitic and the following item (e.g. on [n]a (l. 14), elle nous <lb/>[z]a données (l. 43)), Det + N (un [n]éduc, l. 12) and the set phrase tout-à-fait [tutafe] (l. 11). <lb/>There are a few contexts which are considered as variable in the recent literature and in all <lb/>such cases liaison is not realized: e.g. c&apos;est // un éduc (l. 12), c&apos;est // impressionnant (l. 49). <lb/>The reading aloud of the PFC text is therefore useful for a better understanding of this <lb/>speaker&apos;s system and of the possible role of register. Interestingly, the higher register <lb/>normally triggered by reading aloud is only detectable in the behaviour of the form est of the <lb/>verb être: est [t]en grand émoi, est [t]en revanche. In all other variable contexts, liaison is <lb/>absent: e.g. plural noun + adjective (e.g. circuits // habituels), avoir + past participle (ont // <lb/>eu), verb + complement (préparent // une journée chaude), and so forth. MM provides a good <lb/>example of the fact that adjective + noun is no longer an obligatory context of liaison: she <lb/>realized grand émoi without liaison but grand honneur, a more frequent combination, with <lb/>liaison. <lb/>This discrepancy has been much discussed in the PFC literature (e.g. Durand &amp; Lyche <lb/>2008, and Ch. 28). Overall, it can be said that MM realizes variable liaisons in a sparing <lb/>manner and limits herself to the frequent but restricted categorical contexts described in Ch. <lb/>28. <lb/> 4.4 Other features <lb/> The behaviour of the consonants in this extract does not require specific comments in an <lb/>overview such as this one. On the other hand, the reader should note quite a lot of reductions <lb/>and simplifications typical of fast informal speech. Thus, in the string si tu veux, the vowel /i/ <lb/>is dropped ([styvø]). The personal pronoun je can be reduced to [ʃ] through loss of the schwa <lb/>and devoicing of the [ʒ] by assimilation to a following voiceless consonant (a [t] in the <lb/>following example): faut que je trouve [fokʃtχuv] (l. 48). This example also illustrates the <lb/>fact that deletions of schwas can produce sequences of three consonants. Note that this [ʃ] can <lb/>absorb a following [s] as in je suis [ʃɥi] (l.11), je sais pas [ʃepa] (l. 48). Finally, the <lb/>pronunciation of expliquer as [ɛsplike] in sans les réexpliquer (l. 65) is often considered as a <lb/>feature of southern French but it is well attested in other northern varieties of French. <lb/></body>
21
+
22
+ <listBibl>REFERENCES <lb/> Armstrong, Nigel and Pooley, Tim (2010). Social and Linguistic Change in European <lb/>French. Basingstoke: Palgrave Macmillan. <lb/>Bec, Pierre (1963). La langue occitane. Paris: Presses Universitaires de France. <lb/>Blanche-Benveniste, Claire (2000) Approches de la langue parlée en français. Paris: <lb/>Ophrys. <lb/>Durand, Jacques &amp; Coquillon, Annelise (2010). Le français méridional : éléments de <lb/>synthèse. In: S. Detey, J. Durand, B. Laks &amp; C. Lyche (eds). Les variétés du français <lb/>parlé dans l&apos;espace francophone. Ressources pour l&apos;enseignement. Paris : Ophrys. pp. <lb/>185-197. <lb/>Durand, Jacques &amp; Lyche, Chantal (2008) French liaison in the light of corpus data. <lb/> Journal of French Language Studies. 18/1: 33-66. <lb/>Rossi-Gensane, Nathalie (2010). Oralité, syntaxe et discours. In S. Detey, J. Durand, <lb/>B. Laks &amp; Chantal Lyche (eds) Les variétés du français parlé dans l&apos;espace francophone. <lb/>Ressources pour l&apos;enseignement. Paris: Ophrys. <lb/>Rouayrenc, Catherine (2010). Le français oral. Vol. 1, Les composantes de la chaîne <lb/>parlée. Paris : Belin. <lb/>Walter, Henriette (1982). Enquête phonologique et variétés régionales du français. <lb/> Paris: PUF. </listBibl>
23
+
24
+ </text>
25
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/Gasiorowski_PLoS_Pathogens_2019-1_number_prefix.training.segmentation.tei.xml ADDED
@@ -0,0 +1,210 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>RESEARCH ARTICLE <lb/>HupA, the main undecaprenyl pyrophosphate <lb/>and phosphatidylglycerol phosphate <lb/>phosphatase in Helicobacter pylori is essential <lb/>for colonization of the stomach <lb/>Elise Gasiorowski 1,2,3 , Rodolphe Auger 4 , Xudong Tian 4 , Samia Hicham 1,2 , <lb/>Chantal Ecobichon 1,2 , Sophie Roure 4 , Martin V. Douglass 5,6,7 , M. Stephen Trent 5,6,7 , <lb/>Dominique Mengin-Lecreulx ID <lb/>4 <lb/>, Thierry Touze ´4*, Ivo Gomperts Boneca ID <lb/>1,2 <lb/>* <lb/>1 Institut Pasteur, Unite ´biologie et ge ´ne ´tique de la paroi bacte ´rienne, 28, rue du Docteur Roux, Paris, <lb/>France, 2 INSERM, Groupe Avenir, Paris, France, 3 Universite ´Paris Descartes, Sorbonne Paris Cite ´, Paris, <lb/>France, 4 Institute for Integrative Biology of the Cell (I2BC), CEA, CNRS, Univ Paris-Sud, Universite ´Paris-<lb/>Saclay, Gif-sur-Yvette, France, 5 Department of Infectious Diseases, College of Veterinary Medicine, <lb/>University of Georgia, Georgia, United States of America, 6 Department of Microbiology, Franklin College of <lb/>Arts and Sciences, University of Georgia, Georgia, United States of America, 7 Center for Vaccines and <lb/>Immunology, University of Georgia, Georgia, United States of America <lb/>* [email protected] (TT); [email protected] (IGB) <lb/>Abstract <lb/>The biogenesis of bacterial cell-envelope polysaccharides requires the translocation, across <lb/>the plasma membrane, of sugar sub-units that are produced inside the cytoplasm. To this <lb/>end, the hydrophilic sugars are anchored to a lipid phosphate carrier (undecaprenyl phos-<lb/>phate (C 55 -P)), yielding membrane intermediates which are translocated to the outer face of <lb/>the membrane. Finally, the glycan moiety is transferred to a nascent acceptor polymer, <lb/>releasing the carrier in the &quot;inactive&quot; undecaprenyl pyrophosphate (C 55 -PP) form. Thus, <lb/>C 55 -P is generated through the dephosphorylation of C 55 -PP, itself arising from either de <lb/>novo synthesis or recycling. Two types of integral membrane C 55 -PP phosphatases were <lb/>described: BacA enzymes and a sub-group of PAP2 enzymes (type 2 phosphatidic acid <lb/>phosphatases). The human pathogen Helicobacter pylori does not contain BacA homologue <lb/>but has four membrane PAP2 proteins: LpxE, LpxF, HP0350 and HP0851. Here, we report <lb/>the physiological role of HP0851, renamed HupA, via multiple and complementary <lb/>approaches ranging from a detailed biochemical characterization to the assessment of its <lb/>effect on cell envelope metabolism and microbe-host interactions. HupA displays a dual <lb/>function as being the main C 55 -PP pyrophosphatase (UppP) and phosphatidylglycerol phos-<lb/>phate phosphatase (PGPase). Although not essential in vitro, HupA was essential in vivo for <lb/>stomach colonization. In vitro, the remaining UppP activity was carried out by LpxE in addi-<lb/>tion to its lipid A 1-phosphate phosphatase activity. Both HupA and LpxE have crucial roles <lb/>in the biosynthesis of several cell wall polysaccharides and thus constitute potential targets <lb/>for new therapeutic strategies. <lb/>PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/>1 / 24 <lb/>a1111111111 <lb/>a1111111111 <lb/>a1111111111 <lb/>a1111111111 <lb/>a1111111111 <lb/>OPEN ACCESS <lb/>Citation: Gasiorowski E, Auger R, Tian X, Hicham <lb/>S, Ecobichon C, Roure S, et al. (2019) HupA, the <lb/>main undecaprenyl pyrophosphate and <lb/>phosphatidylglycerol phosphate phosphatase in <lb/>Helicobacter pylori is essential for colonization of <lb/>the stomach. PLoS Pathog 15(9): e1007972. <lb/>https://doi.org/10.1371/journal.ppat.1007972 <lb/>Editor: Mario F. Feldman, University of Alberta, <lb/>CANADA <lb/>Received: March 14, 2019 <lb/>Accepted: July 9, 2019 <lb/>Published: September 5, 2019 <lb/>Copyright: © 2019 Gasiorowski et al. This is an <lb/>open access article distributed under the terms of <lb/>the Creative Commons Attribution License, which <lb/>permits unrestricted use, distribution, and <lb/>reproduction in any medium, provided the original <lb/>author and source are credited. <lb/>Data Availability Statement: All relevant data are <lb/>within the manuscript and its Supporting <lb/>Information files. <lb/>Funding: Funding from the National Institutes of <lb/>Health (NIH; grants AI129940 and AI38576; https:// <lb/>www.nih.gov/) to M. Stephen Trent is gratefully <lb/>acknowledged. This study has received funding <lb/>from the French Government&apos;s Investissement <lb/>d&apos;Avenir program, Laboratoire d&apos;Excellence <lb/>&quot;Integrative Biology of Emerging Infectious <lb/>Author summary <lb/>Helicobacter pylori colonizes the human&apos;s gastric mucosa and infects around 50% of the <lb/>world&apos;s population. This pathogen is responsible for chronic gastritis, peptic ulcers and in <lb/>worst cases leads to gastric cancer. It has been classified as a class I carcinogen by the <lb/>World Health Organization in 1994. Here, we show that HP0851, renamed HupA, is the <lb/>major undecaprenyl pyrophosphate (C 55 -PP) phosphatase (UppP) and the major phos-<lb/>phatidylglycerol phosphate phosphatase (PGPase). This enzyme is also involved in cat-<lb/>ionic antimicrobial peptide (CAMP) resistance to which H. pylori hupA mutant shows an <lb/>increased sensitivity (4 fold). This mutant was unable to colonize the stomach in mouse <lb/>model of infection showing that even if hupA was not essential in vitro, this gene was <lb/>essential in vivo. Both HupA and LpxE have crucial roles in the biosynthesis of several cell <lb/>wall polysaccharides and thus constitute potential targets for new therapeutic strategies. <lb/></front>
8
+
9
+ <body>Introduction <lb/>The biogenesis of many bacterial cell-envelope polysaccharides (i.e., peptidoglycan (PGN), <lb/>lipopolysaccharides (LPS), teichoic acids, enterobacterial common antigen) requires the trans-<lb/>location, across the cytoplasmic membrane, of glycan units that are produced inside the cyto-<lb/>plasm [1]. Therefore, the hydrophilic sugars must be anchored to a lipid carrier (undecaprenyl <lb/>phosphate (C 55 -P)), yielding membrane intermediates which are translocated to the outer face <lb/>of the membrane [2]. In PGN biosynthesis, these intermediates are finally cross-linked by <lb/>transglycosylase and transpeptidase activities to a nascent acceptor polymer. These polymeri-<lb/>zation reactions release the lipid carrier in an &quot;inactive&quot; undecaprenyl pyrophosphate form <lb/>(C 55 -PP) which must be recycled to participate in new rounds of cell-envelope polysaccharides <lb/>biosynthesis. <lb/>C 55 -P originates from the dephosphorylation of C 55 -PP, itself arising from either (i) cyto-<lb/>plasmic de novo synthesis by condensation of eight isopentenyl pyrophosphate (C 5 -PP) mole-<lb/>cules with one farnesyl pyrophosphate (C 15 -PP) catalyzed by the essential C 55 -PP synthase <lb/>(UppS) [3] or (ii) recycling [4] when it is released at the periplasmic side of the membrane. <lb/>Two unrelated families of integral membrane proteins exhibiting C 55 -PP phosphatase (UppP) <lb/>activity were identified: BacA and members of the PAP2 (type 2 phosphatidic acid phospha-<lb/>tase) super-family. Escherichia coli cells possess four UppPs: BacA enzyme which accounts for <lb/>75% of UppP activity and three PAP2 enzymes (PgpB, YbjG and LpxT) which ensure the <lb/>remaining activity [5]. The plurality of UppPs as observed in E. coli and Bacillus subtilis [6] <lb/>seems to be shared by most of the bacteria as suggested by a search for homologues, raising the <lb/>question of the role of such a multiplicity. The simultaneous inactivation of bacA, ybjG and <lb/>pgpB is lethal in E. coli whereas any single or double deletions had no effect on bacterial <lb/>growth. Overexpression of BacA, PgpB or YbjG resulted in bacitracin resistance, and an <lb/>increase of the UppP activity contained in membrane extracts [7]. Bacitracin is an antibiotic <lb/>produced by Bacillus licheniformis [8] which strongly binds C 55 -PP, thereby inhibiting its <lb/>dephosphorylation and leading to an arrest of PGN biosynthesis. When overexpressed, the <lb/>UppP enzymes likely compete with the bacitracin for C 55 -PP binding, thus favoring its <lb/>dephosphorylation. LpxT was not able to sustain growth of the triple bacA-ybjG-pgpB mutant <lb/>and its overexpression did not lead to any bacitracin resistance suggesting that LpxT displays <lb/>another function. Nevertheless, LpxT was shown to catalyze the transfer of C 55 -PP distal phos-<lb/>phate group onto lipid A, the lipid moiety of LPS, yielding C 55 -P and a pyrophosphorylated <lb/>form of lipid A [9]. In addition to its UppP activity, PgpB is involved in phospholipids <lb/></body>
10
+
11
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
12
+
13
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
14
+
15
+ <page>2 / 24 <lb/></page>
16
+
17
+ <front>Diseases&quot; (grant n˚ANR-10-LABX-62-IBEID; http:// <lb/>www.agence-nationale-recherche.fr/ <lb/>investissements-d-avenir/). This work was <lb/>supported by the Agence National de la Recherche <lb/>(ANR; grant n˚ANR-11-BSV3-0002; http://www. <lb/>agence-nationale-recherche.fr/) and DIM Malinf <lb/>(grant n˚dim140053; http://www.dim-malinf.org/) <lb/>to Dominique Mengin-Lecreulx and Ivo G. Boneca. <lb/>Elise Gasiorowski was supported by a fellowship <lb/>from the Fondation pour la Recherche Me ´dicale <lb/>(FRM; fellowship FDT20170436808; https://www. <lb/>frm.org/en). The funders had no role in study <lb/>design, data collection and analysis, decision to <lb/>publish, or preparation of the manuscript. <lb/>Competing interests: The authors have declared <lb/>that no competing interests exist. <lb/></front>
18
+
19
+ <body>biosynthesis via the hydrolysis of phosphatidylglycerol phosphate (PGP) to form phosphatidyl-<lb/>glycerol (PG) [10]. In E. coli, PgpA and PgpC are two additional integral membrane enzymes <lb/>sharing the latter function. It has been shown that the co-inactivation of the three PGP phos-<lb/>phatases leads to a lethal phenotype [11]. Topology and structural analyses of PAP2 enzymes <lb/>from E. coli and B. subtilis showed that these enzymes exhibit their active site at the interface <lb/>between the plasma membrane and the periplasmic space, suggesting they are rather involved <lb/>in C 55 -PP recycling [6,12,13]. <lb/>More recently, the structure of BacA from E. coli was also resolved [14,15] showing here <lb/>again an enzyme with its active site residues accessible from the periplasmic side. Nevertheless, <lb/>the unique structure of BacA was reminiscent of that of transporters or channels raising the <lb/>possibility that BacA may have alternate active sites on either side of the membrane and/or <lb/>may function as a flippase allowing complete recycling of C 55 -P. <lb/>Helicobacter pylori is a microaerophilic, spiral-shaped, flagellated, Gram-negative bacte-<lb/>rium that colonizes the human&apos;s gastric mucosa [16]. This pathogen is responsible of chronic <lb/>gastritis and peptic ulcers [17] and is a risk factor for gastric cancer. It has been classified as a <lb/>class I carcinogen by the World Health Organization in 1994. In contrast to E. coli, H. pylori <lb/>does not contain a BacA protein but has four PAP2 enzymes: HP0021 (LpxE), HP0350, <lb/>HP0851 and HP1580 (LpxF). LpxE and LpxF have been shown to be involved in LPS modifica-<lb/>tions. The lipid A structure of H. pylori is unique and constitutively modified. LpxE is respon-<lb/>sible for the removal of the lipid A 1-phosphate group that is required for the further addition <lb/>of a phosphoethanolamine (PE) group at the same position [18]. LpxF is a lipid A 4&apos;-phosphate <lb/>phosphatase. Both modifications increase the net positive charge of the lipid A, which then <lb/>confers cationic antimicrobial peptide (CAMP) resistance and escape to the host innate <lb/>immune system, thereby allowing H. pylori to survive in the gastric mucosa [19]. The aim of <lb/>this study was to decipher the physiological role of the multiple PAP2 enzymes from H. pylori. <lb/>We describe that HP0851, renamed HupA, is the main UppP but is also involved in phospho-<lb/>lipid biosynthesis by catalyzing the dephosphorylation of PGP in PG. In addition, we show <lb/>that HupA has a role in CAMP resistance and is essential for colonization of the mouse stom-<lb/>ach. In a global context of increasing resistance to antibiotics, finding new potential therapeu-<lb/>tic targets is crucial especially against H. pylori, which colonizes half of the world&apos;s population <lb/>and is classified by the World Health Organization as a priority 2 pathogen regarding antibi-<lb/>otic resistance. A better understanding of essential metabolic pathways, and of the enzymes <lb/>involved in such processes, could lead to the development of new antibacterials. <lb/>Results <lb/>Purification of PAP2 proteins and determination of their UppP activities <lb/>To characterize the PAP2 enzymes from H. pylori (LpxE, HP0350, HP0851 and LpxF), the cor-<lb/>responding genes were cloned on the pTrcHis30 expression vector under the control of a <lb/>strong IPTG-inducible promoter (Table 1). The recombinant N-terminally His-tagged pro-<lb/>teins were overproduced in E. coli C43(DE3) cells. The integral membrane proteins were <lb/>extracted from membranes via their solubilization with n-dodecyl-β-D-maltoside (DDM) <lb/>detergent and the PAP2 proteins were purified by affinity chromatography using Ni 2+ -NTA <lb/>agarose beads (Materials and Methods). Due to the difficulty in getting high expression levels <lb/>of these membrane proteins in E. coli, they could not be purified to homogeneity. Nevertheless, <lb/>LpxE, HP0350 and HP0851 were enriched to a high level as judged from SDS-PAGE analysis <lb/>(Fig 1A), while LpxF could not be visualized by Coomassie blue staining. However, LpxF was <lb/>in the purified samples as confirmed by western blot analysis (Fig 1B). We then measured the <lb/>UppP activities present in these purified solutions. Considering the niche of H. pylori, which <lb/></body>
20
+
21
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
22
+
23
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
24
+
25
+ <page>3 / 24 <lb/></page>
26
+
27
+ <body>Table 1. Bacterial strains and plasmids. <lb/>Strains <lb/>Genotype <lb/>Resistance a Reference <lb/>E. coli <lb/>DH5α <lb/>F -endA1 glnV44 thi-1 recA1 relA1 gyrA96 deoR nupG purB20 φ80dlacZΔM15 Δ(lacZYA-argF)U169, <lb/>hsdR17(r K <lb/>-<lb/>m K <lb/>+ ), λ -<lb/>Life Science <lb/>Technologies <lb/>C43(DE3) <lb/>F -ompT gal dcm hsdS B (r B <lb/>-m B <lb/>-)(DE3) <lb/>Avidis <lb/>BWPGPTs <lb/>BW25113 ΔpgpA ΔpgpC ΔpgpB::Cm + pMAKkanpgpB <lb/>Cm, Kan <lb/>[6] <lb/>BWTsbacA <lb/>BW25113 ΔbacA ΔybjG ΔpgpB::Kan + pMAKbacA <lb/>Cm, Kan <lb/>[7] <lb/>H. pylori <lb/>HP-1 <lb/>N6 <lb/>[20] <lb/>HP-2 <lb/>N6 lpxE::Gm <lb/>Genta <lb/>This work <lb/> HP-3 <lb/>N6 hp0350::Km <lb/>Kan <lb/>This work <lb/>HP-4 <lb/>N6 lpxF::Km <lb/>Kan <lb/>This work <lb/>HP-13 <lb/>N6 hupA::Km <lb/>Kan <lb/>This work <lb/>HP-5 <lb/>N6 pILL2150 <lb/>Cm <lb/>This work <lb/>HP-7 <lb/>N6 pILL2150 bacA <lb/>Cm <lb/>This work <lb/>HP-8 <lb/>N6 pILL2150 lpxT <lb/>Cm <lb/>This work <lb/>HP-9 <lb/>N6 pILL2150 pgpB <lb/>Cm <lb/>This work <lb/>HP-10 <lb/>N6 pILL2150 ybjG <lb/>Cm <lb/>This work <lb/>HP-14 <lb/>N6 pILL2150 lpxE <lb/>Cm <lb/>This work <lb/>HP-15 <lb/>N6 pILL2150 hp0350 <lb/>Cm <lb/>This work <lb/>HP-16 <lb/>N6 pILL2150 lpxF <lb/>Cm <lb/>This work <lb/>N6 +pILL2157 bacA N6 pILL2157 bacA <lb/>Cm <lb/>This work <lb/>N6 +pILL2157 lpxT N6 pILL2157 lpxT <lb/>Cm <lb/>This work <lb/>N6 +pILL2157 pgpB N6 pILL2157 pgpB <lb/>Cm <lb/>This work <lb/>N6 +pILL2157 ybjG N6 pILL2157 ybjG <lb/>Cm <lb/>This work <lb/>HP-22 <lb/>X47-2AL <lb/>[21] <lb/>HP-50 <lb/>X47 hupA::Km <lb/>Kan <lb/>This work <lb/>HP-37 <lb/>N6 hupA::Km pILL2150 lpxE <lb/>Kan, Cm <lb/>This work <lb/>HP-38 <lb/>N6 hupA::Km pILL2150 hp0350 <lb/>Kan, Cm <lb/>This work <lb/>HP-39 <lb/>N6 hupA::Km pILL2150 hupA <lb/>Kan, Cm <lb/>This work <lb/>HP-40 <lb/>N6 hupA::Km pILL2150 lpxF <lb/>Kan, Cm <lb/>This work <lb/>HP-33 <lb/>N6 hupA::Km pILL2150 <lb/>Kan, Cm <lb/>This work <lb/>HP-49 <lb/>N6 pgpA::Km <lb/>Kan <lb/>This work <lb/>Plasmids <lb/>Topo TAΔlpxE:Gm <lb/>Genta <lb/>This work <lb/>Topo TAΔlpxE:Km <lb/>Kan <lb/>This work <lb/>Topo TAΔhp0350: <lb/>Km <lb/>Kan <lb/>This work <lb/>Topo TAΔhupA:Km <lb/>Kan <lb/>This work <lb/>Topo TAΔlpxF:Km <lb/>Kan <lb/>This work <lb/>pTrcHis30 lpxE <lb/>Amp <lb/>This work <lb/>pTrcHis30 hp0350 <lb/>Amp <lb/>This work <lb/>pTrcHis30 hupA <lb/>Amp <lb/>This work <lb/>pTrcHis30 lpxF <lb/>Amp <lb/>This work <lb/>pTrcHis30 pgpA <lb/>Amp <lb/>This work <lb/>pILL2150 lpxE <lb/>Cm <lb/>This work <lb/>pILL2150 hp0350 <lb/>Cm <lb/>This work <lb/>pILL2150 hupA <lb/>Cm <lb/>This work <lb/>pILL2150 lpxF <lb/>Cm <lb/>This work <lb/>(Continued ) <lb/></body>
28
+
29
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
30
+
31
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
32
+
33
+ <page>4 / 24 <lb/></page>
34
+
35
+ <body>ranges from the human stomach with an acidic pH to the epithelial interphase with a neutral <lb/>pH, we measured enzymatic activities on a large range of pHs (Fig 1C). <lb/>The uncharacterized HP0350 protein did not catalyze C 55 -PP dephosphorylation even <lb/>under prolonged incubation time and high protein concentration. LpxE displayed an UppP <lb/>activity of 900 nmol/min/mg at an optimal pH of 7.4, while HP0851 exhibited a 6.7-fold higher <lb/>UppP activity of 6039 nmol/min/mg at an optimal pH of 5. LpxE displayed UppP activity in <lb/>vitro in addition to its role as lipid A 1-phosphate phosphatase [19]. Considering the highest <lb/>activity of HP0851, this protein may account for a large part of C 55 -P (re)generation in vivo. <lb/>LpxF was shown to act as a lipid A 4&apos;-phosphate phosphatase [19]. In this study, we also <lb/>detected a low UppP activity for LpxF (Table 2 and Fig 1C). This activity is unlikely to arise <lb/>from contaminants since HP0350, which was purified using the same methodology, exhibited <lb/>no activity as compared to LpxF. <lb/>HP0851 accounts for more than 90% of the UppP activity <lb/>To further address the contribution of each PAP2 protein in the global UppP activity in H. <lb/>pylori, we generated the four corresponding single mutants in H. pylori N6. All four mutants <lb/>were readily obtained showing that none of these proteins is essential for survival of H. pylori <lb/>in vitro. We prepared DDM-solubilized membrane extracts from exponentially growing wild-<lb/>type and mutant cells and we measured the residual UppP activity present in these extracts <lb/>(Fig 2). The UppP activities present in membranes of hp0350, lpxE and lpxF single mutants <lb/>and wild-type strain were very similar. In contrast, the UppP activity decreased in hp0851 <lb/>membranes to about 10% of residual activity as compared to the wild-type. Thus, these data <lb/>confirmed the major contribution of HP0851 in C 55 -PP recycling in C 55 -P. <lb/>Only LpxE and HP0851 complement the E. coli conditionally UppP <lb/>deficient strain <lb/>The simultaneous inactivation of bacA, pgpB and ybjG is lethal in E. coli. The BWTsbacA strain <lb/>is a thermosensitive conditional triple mutant (ΔbacA, ΔybjG, ΔpgpB) containing an ectopic <lb/>copy of bacA on a plasmid whose replication is impaired at 42˚C [7]. This mutant accumulates <lb/>soluble PGN precursors and lyses after a shift from 30˚C to 42˚C, due to the depletion of the <lb/>pool of C 55 -P that arrests cell-wall synthesis. The ability of the PAP2 proteins from H. pylori to <lb/>restore the growth of BWTsbacA at the restrictive temperature was tested using the <lb/>pTrcHis30-based plasmids previously used for protein purification (Table 3). The hp0350 <lb/>gene was unable to complement BWTsbacA strain even in the presence of 1 mM IPTG. The <lb/>Table 1. (Continued) <lb/>Strains <lb/>Genotype <lb/>Resistance a Reference <lb/>pILL2150 bacA <lb/>Cm <lb/>This work <lb/>pILL2150 lpxT <lb/>Cm <lb/>This work <lb/>pILL2150 pgpB <lb/>Cm <lb/>This work <lb/>pILL2150 ybjG <lb/>Cm <lb/>This work <lb/>pILL2157 bacA <lb/>Cm <lb/>This work <lb/>pILL2157 lpxT <lb/>Cm <lb/>This work <lb/>pILL2157 pgpB <lb/>Cm <lb/>This work <lb/>pILL2157 ybjG <lb/>Cm <lb/>This work <lb/>(a) Genta or Gm, Kan or Km, Cm and Amp: gentamycin, kanamycin, chloramphenicol and ampicillin resistance cassettes, respectively. <lb/>https://doi.org/10.1371/journal.ppat.1007972.t001 <lb/></body>
36
+
37
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
38
+
39
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
40
+
41
+ <page>5 / 24 <lb/></page>
42
+
43
+ <body>lpxF gene was also unable to complement in the absence of inducer and was found to be toxic <lb/>in the presence of IPTG as no growth was observed either at 30˚C or at 42˚C. In contrast, lpxE <lb/>and hp0851 genes complemented BWTsbacA without the need of inducer, indicating that a <lb/>basal level of the corresponding proteins allows a supply of C 55 -P that is appropriate for opti-<lb/>mal growth of E. coli. The overproduction of LpxE and LpxF in E. coli, due to the addition of <lb/>IPTG, was lethal possibly due: 1) to interference with LPS biosynthesis, since the modifications <lb/>Fig 1. PAP2 proteins purification and effect of the pH on their UppP activity. The N-terminal His-tagged PAP2 proteins were purified on Ni 2+ -NTA-agarose beads <lb/>as detailed in Materials and Methods, and 10 μL aliquots of the elution fractions were analyzed by SDS-PAGE. The proteins were revealed either (A) by Coomassie blue <lb/>staining or (B) by Western blotting using anti�His tag antibody conjugated to the horseradish peroxidase (C) UppP phosphatase activity present in the purified solutions <lb/>was determined at different pH (from pH 3 to pH 11). LpxE (black), HP0350 (blue), HP0851 (HupA; brown) and LpxF (red) The values are also reported in Table 2. The <lb/>observed molecular weight of recombinant proteins are consistent with their calculated molecular weight: LpxE (21,2 kDa), HP0350 (24,5 kDa), HupA (25,8 kDa) and <lb/>LpxF (24,6 kDa). <lb/>https://doi.org/10.1371/journal.ppat.1007972.g001 <lb/></body>
44
+
45
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
46
+
47
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
48
+
49
+ <page>6 / 24 <lb/></page>
50
+
51
+ <body>they catalyze do not normally exist in E. coli and likely generate cytotoxicity, and/or 2) to accu-<lb/>mulation of large amounts of membrane proteins. These complementation assays perfectly <lb/>corroborate the previous biochemical data, further demonstrating that LpxE and HP0851 act <lb/>as the major and perhaps the sole C 55 -PP phosphatases in H. pylori. Hence, we renamed <lb/>HP0851 to HupA (Helicobacter UppP and PGPase A). <lb/>The simultaneous inactivation of lpxE and hupA is lethal <lb/>To address whether LpxE and HupA are the only C 55 -PP phosphatases in H. pylori, we <lb/>attempted to construct a lpxE/hupA double mutant. However, we failed to generate this strain <lb/>suggesting that co-inactivation of both genes is lethal. We then performed transformation effi-<lb/>ciency assays to test the capacity of each PAP2 from H. pylori to complement this apparent <lb/>lethality. We generated four strains deleted for hupA and carrying a copy of one PAP2 encod-<lb/>ing gene under the control of an IPTG-inducible promoter on the pILL2150 vector. We then <lb/>transformed these cells with the Topo TAΔlpxE plasmid in order to replace lpxE by a gentamy-<lb/>cin resistance cassette. The transformed H. pylori population was diluted and spread on nor-<lb/>mal (total number of bacteria) or selective (ΔlpxE recombinants) plates to measure the <lb/>transformation efficiency expressed in cfu/μg of Topo TAΔlpxE plasmid (Table 4). LpxF and <lb/>HP0350 were unable to complement the lpxE/hupA double mutant as no recombinant was <lb/>obtained even in the presence of IPTG. In contrast, LpxE and HupA complemented the double <lb/>mutant in the presence of IPTG. These data excluded LpxF and HP0350 as bona fide UppP <lb/>and confirm that only LpxE and HupA are major UppPs in H. pylori. <lb/>HupA is involved in resistance to CAMPs <lb/>H. pylori infects the stomach and persists in the mucosa during years due to the expression of <lb/>multiple virulence factors. Persistence also requires a resistance towards CAMPs, which are <lb/>secreted by the host [22]. To evaluate the involvement of PAP2 proteins in CAMPs resistance, <lb/>Table 2. UppP activity of H. pylori PAP2 enzymes. <lb/>UppP specific activity (nmol/min/mg) <lb/>pH <lb/>LpxE <lb/>HP0350 <lb/>HP0851 (HupA) <lb/>LpxF <lb/>3 <lb/>10 <lb/>ND <lb/>131 <lb/>30.8 <lb/>4 <lb/>13.8 <lb/>ND <lb/>1227 <lb/>97.2 <lb/>5 <lb/>21.3 <lb/>ND <lb/>6039 <lb/>130.3 <lb/>6 <lb/>291.3 <lb/>ND <lb/>4616 <lb/>97.5 <lb/>7 <lb/>671.3 <lb/>ND <lb/>3755 <lb/>91.8 <lb/>7 <lb/>735 <lb/>ND <lb/>4157 <lb/>84.8 <lb/>7.4 <lb/>900 <lb/>ND <lb/>3493 <lb/>71.5 <lb/>8 <lb/>492.5 <lb/>ND <lb/>2528 <lb/>48.8 <lb/>9 <lb/>285 <lb/>ND <lb/>1180 <lb/>28.7 <lb/>9 <lb/>76.3 <lb/>ND <lb/>974 <lb/>20.5 <lb/>10 <lb/>0 <lb/>ND <lb/>159 <lb/>9 <lb/>11 <lb/>8.8 <lb/>ND <lb/>19 <lb/>4.2 <lb/>The enzymatic activity was measured in the presence of 50 μM of [ 14 C]C 55 -PP substrate and an appropriate amount <lb/>of enzyme to obtain less than 30% of hydrolysis. Buffering of the reaction mixture was obtained with sodium acetate <lb/>(pH 3-7), Tris-HCl (pH 7-9) or sodium carbonate (pH 9-11). The C 55 -P product and the substrate were separated <lb/>by TLC and further quantified by radioactivity counting. Values represent the mean of at least three individual <lb/>experiments (the S.D. being within 15% of the presented values). ND, no detectable activity. <lb/>https://doi.org/10.1371/journal.ppat.1007972.t002 <lb/></body>
52
+
53
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
54
+
55
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
56
+
57
+ <page>7 / 24 <lb/></page>
58
+
59
+ <body>we used polymyxin B as a surrogate for CAMPs [23]. The modifications generated by LpxE <lb/>and LpxF through their lipid A 1-and 4&apos;-phosphate phosphatase activities promote resistance <lb/>to different CAMPs and they further confer host innate immune system evasion [19]. We con-<lb/>firmed the increased sensitivity already described in [19] for the lpxE and lpxF mutants (32-<lb/>and 128-fold, respectively, Table 5). The hp0350 mutant did not show any increase of poly-<lb/>myxin B sensitivity, while the sensitivity of hupA mutant increased by four-fold as compared <lb/>Fig 2. UppP activity in membranes of H. pylori. The UppP activity of the wild-type and the four single mutants of H. <lb/>pylori N6 strain was measured and normalized by the quantity of proteins in each membrane extracts. The relative <lb/>activities as compared to the wild-type strain membrane extracts are indicated. Each value is the mean of three <lb/>independent measurements. <lb/>https://doi.org/10.1371/journal.ppat.1007972.g002 <lb/>Table 3. Functional complementation of E. coli BWTsbacA conditional strain by PAP2 encoding genes from H. <lb/>pylori. <lb/>E. coli BWTsbacA <lb/>CFU/mL <lb/>30˚C <lb/>42˚C <lb/>30˚C <lb/>+IPTG 1mM <lb/>42˚C <lb/>+IPTG 1mM <lb/>-<lb/>782 <lb/>0 <lb/>876 <lb/>1 <lb/>+pTrc His30 lpxE <lb/>229 <lb/>222 <lb/>0 <lb/>0 <lb/>+pTrc His30 hp0350 <lb/>396 <lb/>2 <lb/>354 <lb/>0 <lb/>+pTrc His30 hupA <lb/>265 <lb/>219 <lb/>186 <lb/>172 <lb/>+pTrc His30 lpxF <lb/>180 <lb/>0 <lb/>0 <lb/>0 <lb/>The E. coli thermosensitive strain was transformed with the pTrcHis30-based plasmids and aliquots were plated onto <lb/>two ampicillin-containing 2YT agar plates incubated at either 30˚C or 42˚C. The CFU were counted after 24 h <lb/>incubation. <lb/>https://doi.org/10.1371/journal.ppat.1007972.t003 <lb/></body>
60
+
61
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
62
+
63
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
64
+
65
+ <page>8 / 24 <lb/></page>
66
+
67
+ <body>to the wild-type strain. The polymyxin B sensitivity of hupA mutant was fully restored upon <lb/>complementation with an ectopic copy of hupA gene. <lb/>HupA has no effect on lipid A structure <lb/>As already mentioned, LpxE and LpxF modify lipid A structure in H. pylori. We then investi-<lb/>gated whether the four-fold decrease in polymyxin B resistance of hupA mutant was a conse-<lb/>quence of an altered lipid A structure. LpxE exhibits UppP activity while also acting as a lipid <lb/>A 1-phosphate phosphatase. We hypothesized that, in the absence of HupA, LpxE has to cope <lb/>with this very low UppP activity. This hypothesis predicted the presence of a mixture of lipid A <lb/>species in hupA mutant with lipid A 1-phosphate together with the usual lipid A 1-PE. The <lb/>resulting rise in net negative charges at the outermost surface of the bacterium could then <lb/>account for the increased sensitivity towards polymyxin B. <lb/>To test this hypothesis, the lipid A from wild-type, hupA mutant and hupA mutant trans-<lb/>formed with pILL2150 hupA plasmid were extracted and analyzed by mass spectrometry. As <lb/>shown in Fig 3A-3C, the main form of lipid A present in wild-type strain was the tetra-acyl-<lb/>ated lipid A with a PE group at position 1 and no phosphate group at position 4&apos;, as described <lb/>[19]. Two additional minor species were observed, a penta-acylated lipid A with PE at position <lb/>1 and no phosphate at position 4&apos; and a hexa-acylated lipid A with PE at position 1 and with <lb/>phosphate at position 4&apos;. In hupA mutant and its complemented derivative, the lipid A profiles <lb/>were very similar to that of the wild-type strain. We then concluded that the moderate decrease <lb/>in polymyxin B resistance in hupA strain was independent of the lipid A structure. In parallel, <lb/>we confirmed the roles of LpxE and LpxF (Fig 3E and 3F) while we show that HP0350 is not <lb/>involved in lipid A modifications (Fig 3D). <lb/>Table 4. Complementation of lpxE/hupA double mutant with an ectopic copy of H. pylori PAP2 encoding genes. <lb/>Helicobacter pylori N6 <lb/>hupA::Km <lb/>Transformation rate <lb/>(transformants/cfu/μg TopoTAΔlpxE) <lb/>-ITPG <lb/>+IPTG (1mM) <lb/>+pILL2150 empty <lb/>0 <lb/>0 <lb/>+pILL2150 lpxE <lb/>0 <lb/>1.44E-05 <lb/>+pILL2150 hp0350 <lb/>0 <lb/>0 <lb/>+pILL2150 hupA <lb/>0 <lb/>1.16E-05 <lb/>+pILL2150 lpxF <lb/>0 <lb/>0 <lb/>Quantification of lpxE gene inactivation in ΔhupA single mutant transformed with plasmids expressing one PAP2 <lb/>from H. pylori. The transformation rates were measured in the presence and in the absence of IPTG. <lb/>https://doi.org/10.1371/journal.ppat.1007972.t004 <lb/>Table 5. Minimal Inhibitory Concentrations (MIC) of polymyxin B of H. pylori strains. <lb/>Helicobacter pylori N6 <lb/>Polymyxin B <lb/>Fold change vs <lb/>WT strain <lb/>WT <lb/>2048 <lb/>-<lb/>lpxE::Gm <lb/>64 <lb/>32 <lb/>lpxF::Km <lb/>16 <lb/>128 <lb/>hp0350::Km <lb/>2048 <lb/>1 <lb/>hupA::Km <lb/>512 <lb/>4 <lb/>hupA::Km + pILL2157 hupA <lb/>2048 <lb/>1 <lb/>MICs are reported as μg/ml and are the average of three experiments. <lb/>https://doi.org/10.1371/journal.ppat.1007972.t005 <lb/></body>
68
+
69
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
70
+
71
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
72
+
73
+ <page>9 / 24 <lb/></page>
74
+
75
+ <body>HupA is essential for stomach colonization <lb/>The maximal UppP activity of LpxE was reached at pH 7.4 (Fig 1) and all complementation <lb/>assays were performed at neutral pH. H. pylori resumes growth in acidic media only after <lb/>Fig 3. Mass spectrometry analysis of lipid A. Lipid A was isolated from wild-type (A), hupA (B), hp0350 (D), lpxE (E) and lpxF (F) mutant and complemented cells (C) <lb/>in strain N6 and analyzed by MALDI-TOF mass spectrometry in the negative-ion mode. The corresponding lipid A structure of the most abundant species is depicted <lb/>for each analyzed strain. <lb/>https://doi.org/10.1371/journal.ppat.1007972.g003 <lb/></body>
76
+
77
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
78
+
79
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
80
+
81
+ <page>10 / 24 <lb/></page>
82
+
83
+ <body>buffering the environment to a pH above 6 thanks to the urease enzyme [24], which precludes <lb/>any assays at acidic pH. However, in mouse colonization, the natural environment of H. pylori <lb/>is mainly acidic, ranging from 2 in the lumen to 5-6 in the mucus layer. But at low pH, we <lb/>noticed that the in vitro activity of LpxE drastically decreased by 43-fold (from 900 to 21 nmol/ <lb/>min/mg), while that of HupA increased by 1.7-fold (from 3493 to 6039 nmol/min/mg) <lb/>(Table 2). Therefore, LpxE may be unable to fully complement the reduced UppP activity in <lb/>the hupA mutant in an acidic environment. This issue was then addressed by mouse coloniza-<lb/>tion assays. The ΔhupA mutation was first introduced into the mouse-adapted strain H pylori <lb/>X47 and the kinetics of colonization of wild-type and hupA strains was followed by sacrificing <lb/>mice after 1, 4, 7, 15 and 32 days after bacterial inoculation. Seven mice per group were sacri-<lb/>ficed in two independent experiments and Fig 4 illustrates the CFU/g of stomach in a time <lb/>course. The hupA mutant was unable to colonize the stomach neither at early or later stages <lb/>after inoculation, showing therefore that HupA is absolutely essential for viability of H. pylori <lb/>in vivo, in sharp contrast to in vitro where neither of the four PAP2 mutants was affected in <lb/>growth (Fig 5). <lb/>HupA has major involvement in phospholipid biosynthesis <lb/>HupA is an orthologue of PgpB from E. coli, which is a bifunctional enzyme acting as a minor <lb/>UppP and being largely involved in phospholipid biosynthesis through the dephosphorylation <lb/>of PGP in PG. In E. coli, the latter activity is also provided by two other integral membrane <lb/>proteins named PgpA and PgpC, which are unrelated to the PAP2 superfamily [10,11,25]. One <lb/>putative gene encoding a PgpA homolog was identified in the genome of H. pylori, hp0737, <lb/>while no PgpC encoding gene was found. The hp0737 was cloned on a pTrcHis30 vector and <lb/>the N-terminal-His tagged protein was expressed in E. coli C43(DE3) cells and purified from <lb/>membranes to high homogeneity in DDM micelles (S1A Fig). In contrast to the PAP2 <lb/>enzymes, PgpA enzymes are Mg 2+ -dependent, which was confirmed for HP0737 that pre-<lb/>sented an optimal Mg 2+ concentration of 6 mM (S1B Fig). Its PGP phosphatase activity was <lb/>then fully confirmed in vitro with a specific activity of 1138 ± 174 nmol/min/mg. Noticeably, <lb/>HP0737 did not display any significant UppP activity and the plasmid carrying hp0737 gene <lb/>did not restore the growth of E. coli BWTsbacA strain at 42˚C. Considering the activity of <lb/>HP0737 both in vivo and in vitro, this protein was renamed PgpA. The pgpA null mutant was <lb/>readily generated by replacement with a resistance cassette (Table 1). These disagree with a <lb/>previous study in which pgpA was described as an essential gene [26]. Our data further sug-<lb/>gested the existence of one or several other PGP phosphatases. <lb/>We then assessed whether the PAP2 proteins from H. pylori are also involved in the synthe-<lb/>sis of PG. DDM-solubilized membrane extracts obtained from wild-type and mutant cells <lb/>(lpxE, hp0350, hupA, lpxF and pgpA) were assayed for their PGP phosphatase activity. This <lb/>activity was measured in the absence and presence of 6 mM Mg 2+ (Fig 6A). A drastic decrease <lb/>of the PGP phosphatase activity in the hupA mutant as compared to the wild-type strain was <lb/>observed, while the other mutants displayed similar activities as the control. Under these con-<lb/>ditions, HupA accounted for 98% of the PGP phosphatase activity in the membrane of H. <lb/>pylori. <lb/>To further confirm the involvement of HupA in the biosynthesis of PG, we also performed <lb/>complementation assays using the conditional BWPGPTs E. coli strain and the different <lb/>pTrcHis30-based plasmids. Like the BWTsbacA strain, the BWPGPTs is a thermosensitive tri-<lb/>ple mutant (ΔpgpA, ΔpgpB, ΔpgpC) containing an ectopic copy of pgpB on a plasmid whose <lb/>replication is impaired at 42˚C. LpxE, HupA and PgpA fully restored the growth at 42˚C of <lb/>this thermosensitive strain without IPTG, while LpxF and HP0350 were unable to complement <lb/></body>
84
+
85
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
86
+
87
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
88
+
89
+ <page>11 / 24 <lb/></page>
90
+
91
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
92
+
93
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
94
+
95
+ <page>12 / 24 <lb/></page>
96
+
97
+ <body>(Table 6). Thus, LpxE, HupA and PgpA carry enough of PGP phosphatase activity to support <lb/>growth of the BWPGPTs strain. <lb/>The apparent PGP phosphatase activity of HupA and LpxE in vivo could explain why pgpA <lb/>inactivation did not lead to a lethal phenotype. To further decipher the role of PAP2 enzymes <lb/>in the biosynthesis of PG in H. pylori, we tried to generate mutant strains inactivated for hupA <lb/>and pgpA but we failed, while a double mutant lpxE/pgpA was readily obtained. These results <lb/>suggest that HupA and PgpA are the only PGP phosphatases able to sustain growth of H. pylori <lb/>in vitro. <lb/>Broad substrate specificity of H. pylori PAP2 enzymes <lb/>Since the HupA and LpxE accept very dissimilar lipidic substrates in vivo, we further analyzed <lb/>the substrate specificity of PAP2 enzymes in vitro (Fig 6B). Again, HP0350 had no visible <lb/>phosphatase activity on any of the tested substrates. HupA was the most active phosphatase <lb/>using all substrates tested except phosphatidic acid and the highest activity was by far obtained <lb/>with the C 15 -PP substrate. Structural and topology analyses of integral membrane PAP2 <lb/>enzymes revealed that their active site residues are oriented towards the periplasm, which is <lb/>likely the same for HupA. Small molecules such as C 15 -PP are never exposed at the periplasmic <lb/>site, therefore the C 15 -PP should not be a natural substrate in vivo. Nevertheless, this analysis <lb/>demonstrates that HupA accepts a large range of pyro-and monophosphate substrates, in par-<lb/>ticular C 55 -PP and PGP. Noticeably, HupA and PgpA displayed very similar activities towards <lb/>Fig 4. In vivo colonization of the hp0851 mutant in strain X47 at days 1, 4, 7, 15 and 32 using OF1 mice. OF1 mice were infected orogastrically <lb/>with the indicated strains at 2 × 10 8 bacteria per mouse. Colonization rates were determined after 1 (A), 4 (B), 7 (C) 15 (D) and 32 (E) days by <lb/>enumeration of colony forming units per gram (CFU/g) of stomach. Circles and square represent individual mice while mean colonization levels <lb/>are illustrated by horizontal bars. Circles and square located on the x-axis represent mice with no colonization. The single hp0851 mutant showed <lb/>a statistically significant colonization defect at day 4, 7, 15 and 32 when compared to wild-type as indicated by a red asterisk ( � p&lt;0.05; <lb/>��� p&lt;0,001). Data from two independent cohorts of mice were combined to increase significance and robustness of our analysis. <lb/>https://doi.org/10.1371/journal.ppat.1007972.g004 <lb/>Fig 5. Growth curves of Helicobacter pylori N6. WT (black) and the four single mutants lpxE::Km (purple), hp0350::Km (blue), <lb/>hupA::Km (brown) and lpxF::Km (red) were grown with classic conditions previously described in Bacterial strains, plasmids and bacterial <lb/>growth conditions section. <lb/>https://doi.org/10.1371/journal.ppat.1007972.g005 <lb/></body>
98
+
99
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
100
+
101
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
102
+
103
+ <page>13 / 24 <lb/></page>
104
+
105
+ <body>PGP. LpxE had similar substrate specificity as HupA, except for C 5 -PP, and much lower activi-<lb/>ties were found. In contrast, LpxF only showed very low activity towards C 15 -PP and PGP. <lb/>Since HupA showed broad substrate specificity, we wondered whether the polymyxin B <lb/>sensitivity could be related to a general membrane permeability defect of the hupA mutant. <lb/>We measured the MIC of the wild-type strain and its four PAP2 mutants to four different <lb/>Fig 6. PGP phosphatase activities in membranes of H. pylori. The PGP phosphatase activity in presence or absence of MgCl 2 of <lb/>the wild-type and five single mutants of H. pylori N6 strain was measured and normalized by the quantity of proteins in membrane <lb/>extracts (A). Ratios were then normalized by the wild-type strain. (B) The phosphatase specific activity of the recombinant proteins <lb/>towards various substrates. <lb/>https://doi.org/10.1371/journal.ppat.1007972.g006 <lb/></body>
106
+
107
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
108
+
109
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
110
+
111
+ <page>14 / 24 <lb/></page>
112
+
113
+ <body>antimicrobial, one clinically relevant CAMP, colistin, and three large antibiotics, vancomycin, <lb/>teicoplanin and daptomycin that do not cross the outer membrane. Table 7 illustrates the <lb/>MICs to the four antimicrobials. <lb/>Absence of HupA, LpxE and LpxF affected specifically resistance to CAMPs without affect-<lb/>ing the resistance to large antimicrobials suggesting that these three PAP2 only affect overall <lb/>membrane charge and not permeability. Given the specific effect on CAMPs resistance and <lb/>the broad substrate specificity of HupA, we next analyzed the phospholipid composition of the <lb/>hupA mutant. As illustrated in S2 Fig, the wild-type N6 and its four PAP2 mutants showed a <lb/>similar phospholipid profile. <lb/>PgpB, YbjG and BacA from E. coli are functional in H. pylori <lb/>Since, HupA and LpxE are functional in E. coli, we tested if the PAP2/BacA enzymes from E. <lb/>coli are also functional in H. pylori. We performed complementation assays as previously <lb/>described by measuring the transformation rate of lpxE inactivation in hupA mutant contain-<lb/>ing different plasmids encoding PAP2/BacA from E. coli. We then generated eight strains <lb/>deleted for hupA and carrying a plasmid (pILL2150 or pILL2157 vector) bearing one of the <lb/>four E. coli genes (bacA, pbpB, ybjG and lpxT) placed under the control of an IPTG-inducible <lb/>promoter. The transgene is under control of a promoter from H. pylori in pILL2157, while it is <lb/>under control of an E. coli promoter in pILL2150. Consequently, the transgenes in pILL2157 <lb/>are likely more expressed in H. pylori as compared to those in pILL2150. Moreover, the <lb/>pILL2157 promoter was previously found to be leaky in contrast to that of pILL2150 vector <lb/>[27]. The transformation rates are summarized in Table 8. As expected, LpxT was unable to <lb/>complement the lack of UppP activity in H. pylori whatever the expression level. At a moderate <lb/>expression level (i.e. in pILL2150), only PgpB was able to complement the double lpxE/hupA <lb/>mutant (1.08 × 10 -5 transformants/cfu/μg DNA topoTAΔlpxE) in the presence of IPTG. In <lb/>contrast, at a higher level of expression (i.e. in pILL2157), BacA, PgpB and YbjG were able to <lb/>complement the double mutant both in the presence and in the absence of inducer. Therefore, <lb/>even though BacA does not belong to the PAP2 super-family, and H. pylori does not possess <lb/>UppP of the BacA type, this enzyme was functional in H. pylori. <lb/>Discussion <lb/>The recycling of C 55 -PP is an essential step for the biosynthesis of many polysaccharides such <lb/>as PGN, LPS O-antigen or teichoic acid [4]. This pathway was only studied in the Gram-<lb/>Table 6. Complementation of E. coli BWPGPTs conditional strain by PAP2 and PgpA encoding genes from H. <lb/>pylori. <lb/>E. coli BWPGPTs <lb/>CFU/mL <lb/>30˚C <lb/>42˚C <lb/>-<lb/>211 <lb/>1 <lb/>+pTrc His30 lpxE <lb/>97 <lb/>140 <lb/>+pTrc His30 hp0350 <lb/>152 <lb/>0 <lb/>+pTrc His30 hupA <lb/>289 <lb/>279 <lb/>+pTrc His30 lpxF <lb/>75 <lb/>0 <lb/>+pTrc His30 pgpA <lb/>652 <lb/>844 <lb/>The E. coli thermosensitive strain was transformed with the pTrcHis30-based plasmids and aliquots were plated onto <lb/>two ampicillin-containing 2YT agar plates incubated at either 30˚C or 42˚C. The CFU were counted after 24 h <lb/>incubation. <lb/>https://doi.org/10.1371/journal.ppat.1007972.t006 <lb/></body>
114
+
115
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
116
+
117
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
118
+
119
+ <page>15 / 24 <lb/></page>
120
+
121
+ <body>positive bacterium B. subtilis [6] and more intensively in the Gram-negative E. coli. Two <lb/>enzymes from H. pylori belonging to the PAP2 super-family, LpxE and LpxF, were previously <lb/>demonstrated to have critical functions through the dephosphorylation of the lipid A moiety <lb/>from LPS. The other two members of the PAP2 super-family, HupA and HP0350, were not <lb/>characterized. HupA and HP0350 are orthologues of PgpB from E. coli, which was described <lb/>as a dual functional enzyme exhibiting both C 55 -PP and PGP phosphatase activities, thus <lb/>involved in both cell-envelope polysaccharides and phospholipids biosynthesis [11]. <lb/>Here, we demonstrated, by in vitro and in vivo analyses, that HupA constitutes the major <lb/>C 55 -PP phosphatase in H. pylori, being responsible for 90% of the total UppP activity present <lb/>in the membranes. Interestingly, HupA presented an optimal pH of 5 for its enzymatic activity, <lb/>which is close to the pH that H. pylori cells face in the human stomach. To further confirm this <lb/>function, we showed that HupA was also fully functional in E. coli by complementing a PAP2/ <lb/>BacA conditionally deficient strain. The deletion of hupA gene was not lethal suggesting that <lb/>one or several other C 55 -PP phosphatases exist in H. pylori. Among the three other PAP2, we <lb/>showed that LpxE and LpxF exhibited UppP activity with an optimal pH of 7.4 and 5, respec-<lb/>tively. However, only LpxE was also capable of complementing the conditional E. coli strain. <lb/>The inactivation of both lpxE and hupA genes was found to be lethal, which was further <lb/>Table 7. Minimal Inhibitory Concentrations (MIC) of H. pylori strains. <lb/>Helicobacter pylori N6 <lb/>Colistin <lb/>Vancomycin <lb/>Teicoplanin <lb/>Daptomycin <lb/>WT <lb/>R <lb/>R <lb/>R <lb/>R <lb/>lpxE::Gm <lb/>12 <lb/>R <lb/>R <lb/>R <lb/>lpxF::Km <lb/>1 <lb/>R <lb/>R <lb/>R <lb/>hp0350::Km <lb/>R <lb/>R <lb/>R <lb/>R <lb/>hupA::Km <lb/>64 <lb/>R <lb/>R <lb/>R <lb/>MICs are reported as μg/ml or R: Resistant. R indicates above the highest concentration on the E-test, i.e. MIC &gt;256 μg/mL. <lb/>https://doi.org/10.1371/journal.ppat.1007972.t007 <lb/>Table 8. Complementation of lpxE/hupA double mutant with an ectopic copy of E. coli PAP2/BacA encoding <lb/>genes. <lb/>Helicobacter pylori N6 <lb/>hupA::Km <lb/>Transformation rate <lb/>(transformants/cfu/μg TopoTAΔlpxE) <lb/>-ITPG <lb/>+IPTG (1mM) <lb/>+pILL2150 empty <lb/>0 <lb/>0 <lb/>+pILL2150 bacA <lb/>0 <lb/>0 <lb/>+pILL2150 lpxT <lb/>0 <lb/>0 <lb/>+pILL2150 pgpB <lb/>0 <lb/>1.08E-05 <lb/>+pILL2150 ybjG <lb/>0 <lb/>0 <lb/>+pILL2157 bacA <lb/>3.85E-05 <lb/>5.25E-05 <lb/>+pILL2157 lpxT <lb/>+pILL2157 pgpB <lb/>1.51E-05 <lb/>2.07E-05 <lb/>+pILL21570 ybjG <lb/>5.86E-06 <lb/>1.06E-06 <lb/>Quantification of lpxE gene inactivation in ΔhupA single mutant transformed with plasmids expressing one PAP2/ <lb/>BacA from E. coli. Two types of expression vectors were used: pILL2150 and pILL2157. The pILL2150 vector <lb/>possesses an E. coli promoter and leads to low levels of expression in H. pylori, while pILL2157 displays a H. pylori <lb/>promoter yielding high levels of expression in H. pylori. The transformation rates were measured in the presence and <lb/>in the absence of IPTG. <lb/>https://doi.org/10.1371/journal.ppat.1007972.t008 <lb/></body>
122
+
123
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
124
+
125
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
126
+
127
+ <page>16 / 24 <lb/></page>
128
+
129
+ <body>confirmed by complementation assays of the double mutant by ectopic copies of lpxE or hupA. <lb/>All these data confirmed that HupA and LpxE are the only physiologically relevant UppP in H. <lb/>pylori. <lb/>In addition, we demonstrated that HupA plays critical roles in vivo. Indeed, HupA appeared <lb/>important for the resistance towards CAMPs by a yet uncharacterized process. In this study, <lb/>we excluded potential alteration of the structure of lipid A. More importantly, HupA was <lb/>shown to be essential for stomach colonization. Indeed, the PAP2 super-family members all <lb/>have their active site exposed to the periplasmic space where the pH is determined by the envi-<lb/>ronmental pH. We then hypothesized, based on in vitro data, that LpxE enzyme might not be <lb/>sufficiently active at acidic to sustain growth in the absence of HupA. Indeed, during colo-<lb/>nization, LpxE is unlikely to function properly since it was poorly active at pH 5 in vitro. In <lb/>fact, apart from the niche that is in contact with the epithelial cells where there exists a neutral <lb/>pH, H. pylori cells are always exposed to an acidic environment. In these conditions, LpxE will <lb/>likely be unable to provide the UppP activity required in absence of HupA enzyme. <lb/>The first step for stomach colonization by H. pylori is dependent on its motility and its ure-<lb/>ase enzyme to buffer its cytoplasm [24]. H. pylori escapes the lumen towards the mucus layer <lb/>to reach a more favorable environment with a higher pH. Therefore, we cannot exclude a rele-<lb/>vant physiological role of LpxE in C 55 -PP dephosphorylation after the first step of colonization, <lb/>once H. pylori has reached the mucus layer and the epithelial surface. Our previous work on <lb/>LpxE showed that LpxE is needed for long-term colonization. However, this is likely due to its <lb/>role as lipid A 1-phosphate phosphatase and escape to TLR4 signaling rather than to its UppP <lb/>activity since the mutant was able to colonize TLR4 KO mice [19]. <lb/>This study also highlighted the dual function of HupA, which has also a major involvement <lb/>in the biosynthesis of phospholipids through the synthesis of PG from PGP. By sequence <lb/>homology, only one additional putative PGP phosphatase was found in H. pylori, i.e. HP0737, <lb/>which is homologous to E. coli PgpA. A global transposon analysis performed on H. pylori <lb/>reported hp0737 as an essential gene [26]. However, we were able to delete hp0737 gene by <lb/>resistance cassette replacement. HP0737 could then account for the residual PGP phosphatase <lb/>activity in H. pylori hupA null mutant. In E. coli, three PGP phosphatases were described, <lb/>PgpA, PgpB and PgpC and only a triple mutant is lethal [11]. Even if the composition of phos-<lb/>pholipids differs between E. coli and H. pylori, i.e. PG represents 25% and 12.5% of the total <lb/>phospholipids, respectively [28,29], we can expect that the presence of PG is also essential in <lb/>H. pylori, especially because PG is the precursor of another important phospholipid, the cardi-<lb/>olipin. A pgpA/hupA double mutant is lethal and confirms the importance of PG in H. pylori. <lb/>Members of the PgpA-family have a predicted cytoplasmic active site while PAP2-like enzymes <lb/>have a periplasmic oriented active site. These results suggest that PGP is accessible to dephos-<lb/>phorylation on both sides of the cytoplasmic membrane similarly to E. coli. Thus, HupA and <lb/>PgpA are the only two physiologically relevant PGP phosphatases in H. pylori despite the weak <lb/>PGP phosphatase activity of LpxE. <lb/>The involvement of HupA in phospholipid biosynthesis may explain the decrease of resis-<lb/>tance to polymyxin B and colistin of the corresponding mutant. Indeed, these cationic peptides <lb/>form pores in the cytoplasmic membrane leading to cytoplasmic leakage. Therefore, if the <lb/>composition of the plasma membrane differs in the hupA mutant as compared to the wild-<lb/>type strain, particularly with a possible accumulation of PGP, the membrane net negative <lb/>charge might increase, favoring the binding of CAMPs, which could potentiate their capacity <lb/>to insert within the membrane. However, our analysis of phospholipid composition of the <lb/>wild-type N6 and its four PAP2 mutants showed similar phospholipid composition. Although <lb/>TLC analysis is mainly qualitative, these results suggest that changes in membrane charge are <lb/></body>
130
+
131
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
132
+
133
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
134
+
135
+ <page>17 / 24 <lb/></page>
136
+
137
+ <body>unlikely to be due to an accumulation of PGP and suggest that accumulation of C55-PP might <lb/>contribute to the mild increased sensitivity of the hupA mutant to CAMPs. <lb/>LpxE and HupA have similar broad substrate specificities although LpxE exhibits a much <lb/>weaker activity. In contrast, LpxF has a very narrow substrate specificity and, so far, only cata-<lb/>lyzes the dephosphorylation of the lipid A. It would be interesting to study the molecular basis <lb/>of such distinct substrate specificity. The low sequence conservation among the membrane <lb/>PAP2 proteins precludes in silico modelling studies using PgpB and BsPgpB. Insights into sub-<lb/>strate specificity will require 3D structures of HupA, LpxE and LpxF with their substrates. <lb/>In this report, we described HupA as the main C 55 -PP and PGP phosphatase in H. pylori. <lb/>This protein was also essential for colonization, likely due to its dual functionality and its cen-<lb/>tral role in several lipid biosynthetic pathways, suggesting that HupA is the sole enzyme capa-<lb/>ble to support C 55 -PP recycling and PG synthesis in its natural niche. Hence, HupA, as well as <lb/>LpxE and LpxF, constitute bona fide new targets for innovative therapeutic strategies against <lb/>H. pylori, which is becoming increasingly resistant to the existing antibiotic arsenal. <lb/>Materials and Methods <lb/>Ethics Statement <lb/>Animal experiments were done according to European (Directive 2010/63 EU) and French <lb/>regulation (De ´cret 2013-118) under the authorized protocol CETEA 2014-072 reviewed by <lb/>the Institut Pasteur Ethical Committee (registered as number 89 with the French Ministry of <lb/>Research). The experimental protocol was also approved by the French Ministry of Research <lb/>under the number APAFIS#11694-2017100510327765 v2. <lb/>Bacterial strains, plasmids and bacterial growth conditions <lb/>The bacterial strains and plasmids used in the study are summarized in Table 1. Precultures of <lb/>H. pylori were started from glycerol stocks routinely stored at -80˚C, plated onto 10% horse <lb/>blood agar medium supplemented with an antibiotic-antifungal mix [19] and incubated at <lb/>37˚C for 24 h in a microaerobic atmosphere (6% O 2 , 10% CO 2 , 84% N 2 ). The glycerol storage <lb/>media is composed of 25% glycerol, 38% Brain Heart Infusion liquid (BHI) (Oxoid) and 37% <lb/>sterilized water. Amplification of the preculture was performed in the same conditions as pre-<lb/>cultures in new plates at 37˚C for 24 h in a microaerobic atmosphere. Liquid precultures were <lb/>started from amplification plates and inoculated into BHI (Oxoid) supplemented with 10% <lb/>fetal bovine serum (FBS) incubated at 37˚C overnight in a microaerobic atmosphere. Liquid <lb/>cultures of H. pylori were started from overnight liquid culture and inoculated into BHI <lb/>(Oxoid) supplemented with 10% fetal bovine serum. In general, liquid cultures of H. pylori <lb/>were grown to an OD 600nm of ~1.0 at 37˚C under microaerobic conditions with shaking. Oth-<lb/>erwise indicated, E. coli was routinely grown at 37˚C in 2YT broth using standard conditions. <lb/>A non-polar kanamycin cassette was digested out of plasmid pUC18-Km2 [30] using <lb/>BamHI and KpnI and ligated into the PCR products of the 5&apos;and 3&apos; flanking regions of genes <lb/>lpxE, hp0350, hupA, lpxF and pgpA using standard cloning techniques. The oligonucleotides <lb/>used for the PCR amplification are described in S1 Table. The generated plasmids, TopoTA <lb/>ΔlpxE:kan, TopoTA Δhp0350:kan, TopoTA ΔhupA:kan, TopoTA ΔlpxF:kan and TopoTA <lb/>ΔpgpA:kan were used to create the corresponding null mutants in H. pylori N6 and/or X47 by <lb/>natural transformation. Similarly, a non-polar gentamycin cassette was digested out of plasmid <lb/>pUC18-Gm [31] using BamHI and KpnI to generate TopoTA ΔlpxE:Gm. <lb/>For the expression of PAP2, BacA and PgpA encoding genes from E. coli or H. pylori, the <lb/>genes were amplified by PCR using appropriate primers (S1 Table) and cloned into the <lb/></body>
138
+
139
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
140
+
141
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
142
+
143
+ <page>18 / 24 <lb/></page>
144
+
145
+ <body>pILL2150 or pILL2157 vectors (for expression in H. pylori) or in pTrcHis30 vector (for expres-<lb/>sion in E. coli). <lb/>Expression and purification of the PAP2 and PgpA proteins from H. pylori <lb/>E. coli C43(DE3) cells carrying pTrcHis30-based plasmids were used for the overproduction of <lb/>N-terminal His 6 -tagged proteins. Cells were grown in 1 liter 2YT-ampicillin at 37˚C until the <lb/>A 600nm reached 0.9, when the expression was induced by the addition of 1 mM IPTG and the <lb/>growth was continued for 3 h at 37˚C. Cells were harvested by centrifugation at 4˚C for 20 min <lb/>at 4000 × g, were washed and finally resuspended in buffer A (20 mM Tris-HCl pH 7.4, 400 <lb/>mM NaCl, 10% glycerol) before disruption by sonication with a Vibracell 72412 sonicator <lb/>(Bioblock). The suspension was centrifuged at 4˚C for 1 h at 100,000 × g to harvest the mem-<lb/>branes, which were washed three times in buffer A. The membranes were finally resuspended <lb/>in buffer A supplemented with 2% DDM for solubilization at 4˚C for 2 h with gentle agitation. <lb/>Solubilized membranes were loaded on nickel-nitrilotriacetate agarose (Ni 2+ -NTA-agarose, <lb/>Qiagen) equilibrated in buffer A supplemented with 0.2% DDM and 10 mM imidazole. The <lb/>column was washed with increasing imidazole concentration using the same buffer and the <lb/>elution was performed with 2 ml of buffer A supplemented with 0.2% DDM and 200 mM <lb/>imidazole. Desalting of samples was carried out using PD-10 desalting columns (GE Health-<lb/>care) and buffer A supplemented with 0.1% DDM. Protein concentrations were determined <lb/>with the QuantiProBCA assay kit (Sigma) or by densitometry analysis of the gel when <lb/>appropriate. <lb/>H. pylori membrane extracts preparation <lb/>H. pylori wild-type and single mutant strains were grown at 37˚C in BHI medium (200 ml cul-<lb/>ture) up to exponential phase. Cell free extracts, membrane free cytosol, and washed mem-<lb/>branes were prepared as previously described [22]. Cells were harvested by centrifugation, <lb/>washed and finally resuspended in 20 mM Tris-HCl, pH 7.4, 200 mM NaCl (buffer B). After <lb/>disruption by sonication, membranes were pelleted by centrifugation at 177 420 × g and resus-<lb/>pended in buffer B supplemented with 2% DDM for solubilization during 1.5 h in the cold. <lb/>The solubilized proteins were recovered by centrifugation at 177 420 × g and conserved at <lb/>-20˚C before activity measurement. <lb/>H. pylori membrane phospholipids extraction and staining <lb/>H. pylori wild-type and four single mutant strains were grown at 37˚C in BHI medium (50 ml <lb/>culture) up to exponential phase. Lipid extraction was performed using a new protocol (Nozeret <lb/>K et al. manuscript submitted). For thin-layer chromatography (TLC) analysis of phospholipids, <lb/>the dried lipid extracts and controls were dissolved respectively in 500μl and 350μl of chloro-<lb/>form. 10μl of the solution was spotted onto a TLC silica gel 60 plate. The TLC plate was devel-<lb/>oped in tanks equilibrated with dichloromethane-methanol-water (65:28:4 [vol/vol]). After <lb/>drying the plate, phospholipids were visualized with molybdenum blue reagent (Sigma). <lb/>C 55 -PP and PGP phosphatase assays <lb/>The C 55 -PP and PGP phosphatase assays were carried out in a 10 μl reaction mixture containing <lb/>20 mM Tris-HCl, pH 7.4, 10 mM β-mercaptoethanol, 150 mM NaCl, 0.2% DDM, 50 μM [ 14 C] <lb/>C 55 -PP or 50 μM [ 14 C]PGP (900 Bq) and enzyme. MgCl 2 was added at 6 mM in the reaction <lb/>mixture when PgpA activity was measured. Appropriate dilutions of purified phosphatases, or <lb/>of membrane extracts, were used to achieve less than 30% substrate hydrolysis. The reaction <lb/></body>
146
+
147
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
148
+
149
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
150
+
151
+ <page>19 / 24 <lb/></page>
152
+
153
+ <body>mixture was incubated at 37˚C for 10 to 30 min and the reaction was stopped by freezing in liq-<lb/>uid nitrogen. The substrates and products were then separated and quantified by thin layer <lb/>chromatography (TLC) analysis, as previously described for C 55 -PP [5] and PGP [11] hydroly-<lb/>sis. When the phosphatase activity was investigated at various pH values, buffering was achieved <lb/>in sodium acetate (pH 3-7), Tris-HCl (pH 7-9) or sodium carbonate buffer (pH 9-11). <lb/>The phosphatase activity towards other non-radiolabeled substrates: C 5 -PP, C 15 -PP, diacyl <lb/>(C8) glycerol-PP (DGPP) and phosphatidic acid (PA), was determined by measuring the <lb/>amount of released inorganic phosphate during catalysis. The reaction mixture was as <lb/>described above with 50 μM substrate in a final volume of 50 μl. After 10 min of incubation at <lb/>37˚C, the reaction was stopped by the addition of 100 μl of Malachite green solution (Biomol <lb/>green, Enzo Life Sciences), and the released phosphate was quantified by measurement of the <lb/>absorbance at 620 nm. <lb/>E. coli functional complementation <lb/>The E. coli thermosensitive strains BWTsbacA and BWPGPTs, carrying multiple chromo-<lb/>somal gene deletions and harboring a bacA-or pgpB-expressing plasmid, respectively, whose <lb/>replication is impaired at 42˚C, have been previously described [7]. These strains were trans-<lb/>formed by the pTrcHis30-based plasmids encoding the different H. pylori PAP2 and PgpA <lb/>encoding genes. Isolated transformants were subcultured at 30˚C in liquid 2YT medium sup-<lb/>plemented with 100 μg/ml ampicillin and, when the absorbance of the culture reached 0.5, the <lb/>cell suspension was diluted to 10 -5 in 2YT medium and 100 μl aliquots were plated onto two <lb/>ampicillin-containing 2YT agar plates which were incubated at either 30˚C or 42˚C for 24 h. <lb/>The colony forming units (CFU) were numerated on each plate and the functional comple-<lb/>mentation of the thermosensitive mutants was evaluated by the capacity of the transformants <lb/>to equally grow at both temperatures. <lb/>Transformation rate <lb/>Precultures of H. pylori were started from glycerol stocks routinely stored at -80˚C, plated onto <lb/>10% horse blood agar medium and incubated at 37˚C for 24 h in a microaerobic atmosphere <lb/>(6% O 2 , 10% CO 2 , 84% N 2 ). Amplification of the pre-culture was performed in the same con-<lb/>ditions such as pre-cultures in new plates at 37˚C for 24 h in a microaerobic atmosphere. 50 μl <lb/>of suspension in BHI at OD 600nm = 20 were made until the amplification plate and mixed with <lb/>10 ng of Topo TAΔlpxE plasmid. The whole mixture was put on a non-selective plate onto <lb/>10% horse blood agar medium and incubated at 37˚C for 24 h in a microaerobic atmosphere. <lb/>The transformation spot was then diluted (10 -1 to 10 -8 ). 50 μl of the non-diluted suspension <lb/>and 10 μl of the 10 -1 to 10 -4 dilutions were spread on 10% horse blood agar medium supple-<lb/>mented with chloramphenicol (4 μg/ml), kanamycin (20 μg/ml) and gentamycin (5 μg/ml) ± <lb/>IPTG 1mM to estimate the number of transformants. The 10 -5 to 10 -8 dilutions were spread <lb/>on non-selective plates containing 10% horse blood agar medium supplemented with chlor-<lb/>amphenicol (4 μg/ml), kanamycin (20 μg/ml) ± IPTG 1mM to estimate the total number of <lb/>bacteria. The CFU of all plates incubated at 37˚C for 5 days in a microaerobic atmosphere <lb/>were enumerated. The transformation rate was expressed by the number of transformants/ <lb/>total cfu/μg of DNA TopoTAΔlpxE. <lb/>Determination of Minimum Inhibitory Concentration (MIC) of polymyxin B <lb/>Strains were grown routinely on amplification plates. Two ml of suspension in 0.9% NaCl at <lb/>OD 600nm = 0.75 were made from the amplification plates and spread by inundation onto Muel-<lb/>ler Hinton medium (Difco) supplemented with 10% of FBS and 2,3,5-triphenyltetrazolium <lb/></body>
154
+
155
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
156
+
157
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
158
+
159
+ <page>20 / 24 <lb/></page>
160
+
161
+ <body>chloride (TTC, Sigma) (40 μg/ml). Polymyxin B was diluted by serial dilutions of two <lb/>(16384 μg/ml down to 0.5 μg/ml). Once the square plates with bacterial lawns were dried, 10 μl <lb/>of each dilution of polymyxin B was dropped on and allowed to dry. All plates were then incu-<lb/>bated at 37˚C for 3 days in a microaerobic atmosphere. TTC is a non-toxic dye which colors in <lb/>red the alive bacteria. MICs were determined as the lowest concentration of polymyxin B lead-<lb/>ing to a clear halo of inhibition. <lb/>Determination of Minimum Inhibitory Concentration (MIC) by Etest <lb/>Strains were grown routinely on amplification plates. Two ml of suspension in 0.9% NaCl at <lb/>OD 600nm = 0.75 were made from the amplification plates and spread by inundation onto 10% <lb/>horse blood agar medium. Once the square plates with bacterial lawns were dried, Etes, Bio-<lb/>me ´rieux, with different antibiotics targeting the cell wall (Vancomycin, Teicoplanin, Dapto-<lb/>mycin and Colistin) were put on plate and incubated at 37˚C for 48 h in a microaerobic <lb/>atmosphere. MICs were determined as the lowest concentration of antibiotics leading to a <lb/>clear halo of inhibition. <lb/>Colonization <lb/>OF1 female mice purchased from Charles River Laboratories aged 5 weeks were infected by <lb/>gavage with feeding needles with X47 strain (2 × 10 8 bacteria per mouse). Colonization rates <lb/>were determined after 1, 4, 7, 15 and 32 days by enumeration of CFU per gram of stomach. <lb/>Mice were euthanized with CO 2 and the stomachs were ground and homogenized in peptone <lb/>broth. The samples were then diluted and spread on blood agar plates supplemented with <lb/>10 μg/ml of nalidixic acid, to inhibit the growth of resident bacteria from the mouse foresto-<lb/>mach and 20 μg/ml of kanamycin for hupA mutant strain. The CFU were enumerated after 5 <lb/>days of incubation under microaerobic conditions. The results of two independent coloniza-<lb/>tion experiments (seven mice by cage) were pooled and a one tailed Mann-Whitney test was <lb/>used to determine statistical significance of observed differences (GraphPad Prism v5.0 Graph-<lb/>Pad Software, CA). <lb/>Isolation of lipid A for mass spectrometry analysis <lb/>For isolation of lipid A, H. pylori cultures were grown to an OD 600 of ~0.6. Lipid A chemical <lb/>extraction was carried out after mild acidic hydrolysis of LPS as previously described [32,33]. <lb/>For visualization of lipid A by mass spectrometry, lipids were analyzed using MALDI-TOF <lb/>(ABI 4700 Proteomic Analyzer) in the negative-ion linear mode similar to previously <lb/>described [34,35]. Briefly, lipid A samples were dissolved in a mixture of chloroform-methanol <lb/>(4:1, vol/vol), with 1 μL of sample mixed with 1 μL of matrix solution consisting of 5-chloro-<lb/>2meracaptobenzothiazole (CBMT) (20 mg/mL) resuspended in chloroform-methanol-water <lb/>(4:4:1, vol/vol/vol) mixed with saturated ammonium citrate (20:1, vol/vol), and 1 μL of sam-<lb/>ple-matrix mixture was loaded on to MALDI target plate. <lb/></body>
162
+
163
+ <div type="annex">Supporting information <lb/>S1 Fig. Purification and characterization of PgpA from H. pylori. (A) SDS-PAGE analysis <lb/>of His 6 -PgpA purification. Coloration was performed with Coomassie Blue R-250. Purified <lb/>PgpA protein is indicated with an asterisk in the elution fraction. (B) Mg 2+ -dependence of <lb/>PgpA PGPase activity. The results are expressed as the percentage of the optimal activity found <lb/>at a final concentration of 6 mM of MgCl 2 . The observed molecular weight of recombinant <lb/></div>
164
+
165
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
166
+
167
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
168
+
169
+ <page>21 / 24 <lb/></page>
170
+
171
+ <div type="annex">PgpA was lower than the calculated one (16,9 kDa). <lb/>(TIF) <lb/>S2 Fig. Membrane phospholipids of Helicobacter pylori. TLC analysis of total lipid extracts <lb/>from N6 (1) WT strain and the four single mutants (2) lpxE::Gm; (3) hp0350::Km; (4) <lb/>lpxF::Km; (5) hupA::Km grown to exponential phase in BHI medium. Left panel contains con-<lb/>trol phospholipids: CL: Cardiolipin; PC: Phosphotidylcholine; PE: Phosphotidylethanolamine; <lb/>PG: Phosphotidylglycerol; PS: Phosphotidylserine; Lip. EC: Lipid extracts from E. coli. <lb/>(TIF) <lb/>S1 Table. Oligonucleotides used in this study. Oligonucleotides which have an underlined <lb/>nucleotide sequence, the sequence corresponds to the restriction enzyme cutting site. <lb/>(DOCX) <lb/></div>
172
+
173
+ <div type="annex">Author Contributions <lb/>Conceptualization: Dominique Mengin-Lecreulx, Thierry Touze ´, Ivo Gomperts Boneca. <lb/>
174
+ Funding acquisition: M. Stephen Trent, Dominique Mengin-Lecreulx, Ivo Gomperts Boneca. <lb/>
175
+ Investigation: Elise Gasiorowski, Rodolphe Auger, Xudong Tian, Samia Hicham, Chantal <lb/>Ecobichon, Sophie Roure, Martin V. Douglass. <lb/>
176
+ Project administration: Thierry Touze <lb/>´. <lb/>
177
+ Resources: Chantal Ecobichon, Sophie Roure. <lb/>
178
+ Supervision: M. Stephen Trent, Thierry Touze ´, Ivo Gomperts Boneca. <lb/>
179
+ Validation: Rodolphe Auger, Samia Hicham, Ivo Gomperts Boneca. <lb/>
180
+ Visualization: Elise Gasiorowski, Rodolphe Auger, Martin V. Douglass. <lb/>
181
+ Writing -original draft: Elise Gasiorowski. <lb/>
182
+ Writing -review &amp; editing: Dominique Mengin-Lecreulx, Thierry Touze ´, Ivo Gomperts <lb/>Boneca. <lb/></div>
183
+
184
+ <listBibl>References <lb/>1. Barreteau H, Kovač A, Boniface A, Sova M, Gobec S, Blanot D. Cytoplasmic steps of peptidoglycan bio-<lb/>synthesis. FEMS Microbiol Rev. 2008; 32: 168-207. https://doi.org/10.1111/j.1574-6976.2008.00104.x <lb/>PMID: 18266853 <lb/>2. Bouhss A, Trunkfield AE, Bugg TDH, Mengin-Lecreulx D. The biosynthesis of peptidoglycan lipid-linked <lb/>intermediates. FEMS Microbiol Rev. 2008; 32: 208-233. https://doi.org/10.1111/j.1574-6976.2007. <lb/>00089.x PMID: 18081839 <lb/>3. Apfel CM, Taka ´cs B, Fountoulakis M, Stieger M, Keck W. Use of Genomics To Identify Bacterial Unde-<lb/>caprenyl Pyrophosphate Synthetase: Cloning, Expression, and Characterization of the Essential uppS <lb/>Gene. J Bacteriol. 1999; 181: 483-492. PMID: 9882662 <lb/>4. Manat G, Roure S, Auger R, Bouhss A, Barreteau H, Mengin-Lecreulx D, et al. Deciphering the Metabo-<lb/>lism of Undecaprenyl-Phosphate: The Bacterial Cell-Wall Unit Carrier at the Membrane Frontier. Microb <lb/>Drug Resist. 2014; 20: 199-214. https://doi.org/10.1089/mdr.2014.0035 PMID: 24799078 <lb/>5. Ghachi ME, Bouhss A, Blanot D, Mengin-Lecreulx D. The bacA Gene of Escherichia coli Encodes an <lb/>Undecaprenyl Pyrophosphate Phosphatase Activity. J Biol Chem. 2004; 279: 30106-30113. https://doi. <lb/>org/10.1074/jbc.M401701200 PMID: 15138271 <lb/>6. Ghachi ME, Howe N, Auger R, Lambion A, Guiseppi A, Delbrassine F, et al. Crystal structure and bio-<lb/>chemical characterization of the transmembrane PAP2 type phosphatidylglycerol phosphate phospha-<lb/>tase from Bacillus subtilis. Cell Mol Life Sci. 2017; 74: 2319-2332. https://doi.org/10.1007/s00018-017-<lb/>2464-6 PMID: 28168443 <lb/></listBibl>
185
+
186
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
187
+
188
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
189
+
190
+ <page>22 / 24 <lb/></page>
191
+
192
+ <listBibl>7. Ghachi ME, Derbise A, Bouhss A, Mengin-Lecreulx D. Identification of Multiple Genes Encoding Mem-<lb/>brane Proteins with Undecaprenyl Pyrophosphate Phosphatase (UppP) Activity in Escherichia coli. J <lb/>Biol Chem. 2005; 280: 18689-18695. https://doi.org/10.1074/jbc.M412277200 PMID: 15778224 <lb/>8. Azevedo EC, Rios EM, Fukushima K, Campos-Takaki GM. Bacitracin production by a new strain of <lb/>Bacillus subtilis. Extraction, purification, and characterization. Appl Biochem Biotechnol. 1993; 42: 1-7. <lb/>PMID: 8215347 <lb/>9. Touze ´T, Tran AX, Hankins JV, Mengin-Lecreulx D, Trent MS. Periplasmic phosphorylation of lipid A is <lb/>linked to the synthesis of undecaprenyl phosphate. Mol Microbiol. 2008; 67: 264-277. https://doi.org/ <lb/>10.1111/j.1365-2958.2007.06044.x PMID: 18047581 <lb/>10. Icho T, Raetz CR. Multiple genes for membrane-bound phosphatases in Escherichia coli and their <lb/>action on phospholipid precursors. J Bacteriol. 1983; 153: 722-730. PMID: 6296050 <lb/>11. Lu Y-H, Guan Z, Zhao J, Raetz CRH. Three Phosphatidylglycerol-phosphate Phosphatases in the Inner <lb/>Membrane of Escherichia coli. J Biol Chem. 2011; 286: 5506-5518. https://doi.org/10.1074/jbc.M110. <lb/>199265 PMID: 21148555 <lb/>12. Tatar LD, Marolda CL, Polischuk AN, van Leeuwen D, Valvano MA. An Escherichia coli undecaprenyl-<lb/>pyrophosphate phosphatase implicated in undecaprenyl phosphate recycling. Microbiology. 2007; 153: <lb/>2518-2529. https://doi.org/10.1099/mic.0.2007/006312-0 PMID: 17660416 <lb/>13. Touze ´T, Blanot D, Mengin-Lecreulx D. Substrate Specificity and Membrane Topology of Escherichia <lb/>coli PgpB, an Undecaprenyl Pyrophosphate Phosphatase. J Biol Chem. 2008; 283: 16573-16583. <lb/>https://doi.org/10.1074/jbc.M800394200 PMID: 18411271 <lb/>14. El Ghachi M, Howe N, Huang C-Y, Olieric V, Warshamanage R, Touze ´T, et al. Crystal structure of <lb/>undecaprenyl-pyrophosphate phosphatase and its role in peptidoglycan biosynthesis. Nat Commun. <lb/>2018; 9. <lb/>15. Workman SD, Worrall LJ, Strynadka NCJ. Crystal structure of an intramembranal phosphatase central <lb/>to bacterial cell-wall peptidoglycan biosynthesis and lipid recycling. Nat Commun. 2018;9. https://doi. <lb/>org/10.1038/s41467-017-01881-x <lb/>16. Marshall B, Warren JR. Unidentified curved bacilli in the stomach of patients with gastritis and peptic <lb/>ulceration. The Lancet. 1984; 323: 1311-1315. <lb/>17. Blaser MJ. Hypotheses on the pathogenesis and natural history of Helicobacter pylori-induced inflam-<lb/>mation. Gastroenterology. 1992; 102: 720-727. https://doi.org/10.1016/0016-5085(92)90126-j PMID: <lb/>1732141 <lb/>18. Tran AX, Karbarz MJ, Wang X, Raetz CRH, McGrath SC, Cotter RJ, et al. Periplasmic Cleavage and <lb/>Modification of the 1-Phosphate Group of Helicobacter pylori Lipid A. J Biol Chem. 2004; 279: 55780-<lb/>55791. https://doi.org/10.1074/jbc.M406480200 PMID: 15489235 <lb/>19. Cullen TW, Giles DK, Wolf LN, Ecobichon C, Boneca IG, Trent MS. Helicobacter pylori versus the Host: <lb/>Remodeling of the Bacterial Outer Membrane Is Required for Survival in the Gastric Mucosa. PLoS <lb/>Pathog. 2011;7. <lb/>20. Behrens W, Bo ¨nig T, Suerbaum S, Josenhans C. Genome sequence of Helicobacter pylori hpEurope <lb/>strain N6. J Bacteriol. 2012; 194: 3725-3726. https://doi.org/10.1128/JB.00386-12 PMID: 22740658 <lb/>21. Veyrier FJ, Ecobichon C, Boneca IG. Draft Genome Sequence of Strain X47-2AL, a Feline Helicobacter <lb/>pylori Isolate. Genome Announc. 2013; 1. <lb/>22. Stead CM, Zhao J, Raetz CRH, Trent MS. Removal of the outer Kdo from Helicobacter pylori lipopoly-<lb/>saccharide and its impact on the bacterial surface. Mol Microbiol. 2010; 78: 837-852. https://doi.org/10. <lb/>1111/j.1365-2958.2010.07304.x PMID: 20659292 <lb/>23. Gutsmann T, Hagge SO, David A, Roes S, Bo ¨hling A, Hammer MU, et al. Lipid-mediated resistance of <lb/>Gram-negative bacteria against various pore-forming antimicrobial peptides. J Endotoxin Res. 2005; <lb/>11: 167-173. https://doi.org/10.1179/096805105X37330 PMID: 15949145 <lb/>24. Smoot DT, Mobley HL, Chippendale GR, Lewison JF, Resau JH. Helicobacter pylori urease activity is <lb/>toxic to human gastric epithelial cells. Infect Immun. 1990; 58: 1992-1994. PMID: 2341188 <lb/>25. Icho T. Membrane-bound phosphatases in Escherichia coli: sequence of the pgpB gene and dual sub-<lb/>cellular localization of the pgpB product. J Bacteriol. 1988; 170: 5117-5124. https://doi.org/10.1128/jb. <lb/>170.11.5117-5124.1988 PMID: 2846511 <lb/>26. Salama NR, Shepherd B, Falkow S. Global Transposon Mutagenesis and Essential Gene Analysis of <lb/>Helicobacter pylori. J Bacteriol. 2004; 186: 7926-7935. https://doi.org/10.1128/JB.186.23.7926-7935. <lb/>2004 PMID: 15547264 <lb/>27. Boneca IG, Ecobichon C, Chaput C, Mathieu A, Guadagnini S, Pre ´vost M-C, et al. Development of <lb/>Inducible Systems To Engineer Conditional Mutants of Essential Genes of Helicobacter pylori. Appl <lb/>Environ Microbiol. 2008; 74: 2095-2102. https://doi.org/10.1128/AEM.01348-07 PMID: 18245237 <lb/></listBibl>
193
+
194
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
195
+
196
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
197
+
198
+ <page>23 / 24 <lb/></page>
199
+
200
+ <listBibl>28. Raetz CR. Enzymology, genetics, and regulation of membrane phospholipid synthesis in Escherichia <lb/>coli. Microbiol Rev. 1978; 42: 614-659. PMID: 362151 <lb/>29. Hirai Y, Haque M, Yoshida T, Yokota K, Yasuda T, Oguma K. Unique cholesteryl glucosides in Helico-<lb/>bacter pylori: composition and structural analysis. J Bacteriol. 1995; 177: 5327-5333. https://doi.org/10. <lb/>1128/jb.177.18.5327-5333.1995 PMID: 7665522 <lb/>30. Skouloubris S, Labigne A, Reuse HD. Identification and characterization of an aliphatic amidase in Heli-<lb/>cobacter pylori. Mol Microbiol. 1997; 25: 989-998. https://doi.org/10.1111/j.1365-2958.1997.mmi536.x <lb/>PMID: 9364923 <lb/>31. Bury-Mone ´S, Mendz GL, Ball GE, Thibonnier M, Stingl K, Ecobichon C, et al. Roles of alpha and beta <lb/>carbonic anhydrases of Helicobacter pylori in the urease-dependent response to acidity and in coloniza-<lb/>tion of the murine gastric mucosa. Infect Immun. 2008; 76: 497-509. https://doi.org/10.1128/IAI.00993-<lb/>07 PMID: 18025096 <lb/>32. Zhou Z, Lin S, Cotter RJ, Raetz CRH. Lipid A Modifications Characteristic of Salmonella typhimurium <lb/>Are Induced by NH4VO3 in Escherichia coli K12 detection of 4-amino-4-deoxy-l-arabinose, phos-<lb/>phoethanolamine and palmitate. J Biol Chem. 1999; 274: 18503-18514. https://doi.org/10.1074/jbc. <lb/>274.26.18503 PMID: 10373459 <lb/>33. Herrera CM, Crofts AA, Henderson JC, Pingali SC, Davies BW, Trent MS. The Vibrio cholerae VprA-<lb/>VprB two-component system controls virulence through endotoxin modification. mBio. 2014; 5. <lb/>34. Henderson JC, O&apos;Brien JP, Brodbelt JS, Trent MS. Isolation and chemical characterization of lipid A <lb/>from gram-negative bacteria. J Vis Exp JoVE. 2013; e50623. https://doi.org/10.3791/50623 PMID: <lb/>24084191 <lb/>35. Zhou P, Altman E, Perry MB, Li J. Study of matrix additives for sensitive analysis of lipid A by matrix-<lb/>assisted laser desorption ionization mass spectrometry. Appl Environ Microbiol. 2010; 76: 3437-3443. <lb/>https://doi.org/10.1128/AEM.03082-09 PMID: 20382818 <lb/></listBibl>
201
+
202
+ <note place="headnote">HupA, the main UppP and PGPase in H.pylori <lb/></note>
203
+
204
+ <note place="footnote">PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007972 September 5, 2019 <lb/></note>
205
+
206
+ <page>24 / 24 </page>
207
+
208
+
209
+ </text>
210
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/HCII07-LongSteph.training.segmentation.tei.xml ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei>
3
+ <teiHeader>
4
+ <fileDesc xml:id="__HCII07-LongSteph"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front> Computer-supported creativity: <lb/>Evaluation of a tabletop mind-map application <lb/> Stéphanie Buisine 1 , Guillaume Besacier 2 , Marianne Najm 1 , <lb/>Améziane Aoussat 1 , and Frédéric Vernier 2 <lb/> 1 <lb/> ENSAM-LCPI, 151 boulevard de l Hôpital, 75013 Paris, FRANCE <lb/>Corresponding author: [email protected] <lb/> 2 <lb/> LIMSI-CNRS, BP 133, 91403 Orsay Cedex, FRANCE <lb/> Abstract. The aim of this study is to investigate the usability and usefulness of <lb/>interactive tabletop technologies to support group creativity. We implemented a <lb/>tabletop interface enabling groups of 4 participants to build mind-maps (a tool <lb/>for associative thinking). With 24 users in a within-group design, we compared <lb/>its use to traditional paper-and-pencil mind-mapping sessions. The results <lb/>showed no difference in idea production, but the tabletop condition <lb/>significantly improved both subjective and collaborative dimensions, especially <lb/>by leading to better-balanced contributions from the group members. <lb/> Keywords: Creativity, Mind-map, Tabletop device. <lb/> </front>
8
+
9
+ <body>1 Creativity in industrial applications <lb/> Creativity is a high-level cognitive process which has given rise to researches in <lb/>various fields such as Psychology [4, 16], Engineering [7, 9] or Human-Computer <lb/>Interaction [3, 6, 14, 15]. Creativity applies to artistic work (e.g. fine arts, literature, <lb/>architecture, music), educative domain (e.g. early-learning and playing activities), <lb/>scientific skills (e.g. problem resolution, discoveries, epistemological breakthroughs), <lb/>and industrial applications (e.g. creation of product functions, stylistic design of <lb/>artifacts). <lb/>In this paper we consider creativity in industrial applications, for example when <lb/>some people design a product with new innovative functions (e.g. a mobile phone <lb/>including a positioning system) or search some applications to a new technology (e.g. <lb/>portable MP3 players). Understanding and supporting this kind of creativity is not <lb/>only an interesting research challenge: it also bears a strong potential for enhancing <lb/>industrial innovation and market opportunities. <lb/> 2 Enhancing creativity <lb/> To improve creativity, a wide-spread practice in companies is the group <lb/>brainstorming. Although creativity fundamentally remains an individual capacity, it <lb/> proves to be influenced by the subject s environment: in this respect, collective <lb/>creativity phenomena are often observed when group emulation improves the <lb/>expression of onee s own creative potential. This is especially true for industrial <lb/>creativity which can benefit from multiple, or even multidisciplinary viewpoints. <lb/>To further improve these collective creativity sessions, methodological toolkits [7, <lb/>9] have been formalized to structure the reflection and manage groups dynamics. <lb/>Consulting services specialized in creative problem solving also appeared to help <lb/>companies conduct creativity sessions and apply these methodologies. <lb/>Moreover, computer applications have been developed to support industrial <lb/>creativity 1 . According to Shneiderman [14], the existing software solutions can be <lb/>categorized into three approaches: inspirational tools (e.g. favoring visualization, free <lb/>association, or sources of inspiration), structural tools (e.g. databases, simulations, <lb/>methodical techniques of reasoning), and situational tools (e.g. based on the social <lb/>context, enabling peer-consultation, or dissemination). Lubart [8] adopted a <lb/>classification grounded on the role played by the computer in the creative process: <lb/>systems assisting the user in the management of creative projects (computer as <lb/>nanny), those supporting communication and collaboration within a team (computer <lb/>as pen-pal), systems implementing creativity enhancement techniques (computer as <lb/>coach) and those contributing to the idea production (computer as colleague). <lb/>In this context, our goal is to investigate the capacity of a tabletop computer (as a <lb/>physical device and as a digital interface) to support collaborative creativity related to <lb/>industrial issues. <lb/> 3 Tabletop systems <lb/> Fig. 1. Example of a tabletop system using MERL DiamondTouch device [5]. <lb/> Tabletop systems (see Fig. 1) are multi-user horizontal interfaces for interactive <lb/>shared displays. They implement around-the-table interaction metaphors allowing co-<lb/>located collaboration and face-to-face conversation in a social setting [12, 13]. <lb/>
10
+
11
+ <note place="footnote"> 1 <lb/> For example Goldfire Innovator (www.invention-machine.com) or ThoughtOffice <lb/>(www.ideacenter.com). <lb/></note>
12
+
13
+ Tabletop systems are used in various application contexts such as games, photo <lb/>browsing, map exploration, planning tasks, classification tasks, interactive exhibit <lb/>medium for museums, drawing, etc. [11]. Such systems being likely to favor <lb/>collaboration by providing around-the-table visualization facilities, they could be <lb/>thought of for supporting creativity sessions: in this respect they would fall into both <lb/>inspirational and situational creativity tools [14] or play pen-pal and coach roles [8] in <lb/>the creative process. Indeed, some tabletop systems were previously considered for <lb/>supporting creativity [17, 18] but their actual benefits were not experimentally <lb/>measured. To assess the usability and usefulness of tabletops in the context of <lb/>creativity sessions, we believe that it is necessary to compare their use with a control <lb/>condition relying on traditional paper-and-pencil tools. In the following section, we <lb/>introduce a creativity application we have developed for tabletop display in order to <lb/>conduct such an experiment. <lb/> 4 Our tabletop mind-map application <lb/> 4.1 Mind-maps as a collective creativity tool <lb/> Fig. 2. Example mind-map. <lb/> In general, creativity methodological framework [7] support a two-step process: first, <lb/>diverging by producing a vast number of ideas, then converging by selecting a few of <lb/>them to be further developed. Mind-maps [2] are used in the diverging step. The <lb/>mind-map principle is based on associative logics and is used for defining the <lb/>problem to address. The field to explore is written in a central box and the participants <lb/>express their free associations to this concept. Those ideas are written in new boxes <lb/>placed as a crown around the central concept (see Fig. 2). A second association level <lb/>is then built from the primary ideas, etc. Because the second level of association is not <lb/>directly related to the initial problem, new-original research directions can appear and <lb/>the realms of possibility grow. Mind-mapping can be performed individually or in a <lb/>group session. In the latter case, the session has to be managed by an animator whose <lb/> role is to coordinate speech turns and ensure that the group agrees on every idea. <lb/>Many software solutions 2 for desktop computers have been developed to support <lb/>mind-mapping, but none is adapted to tabletop interaction. This is why we <lb/>implemented our own tool. <lb/> 4.2 Implementation <lb/> Fig. 3. Example mind-map created with our tabletop application. <lb/> Implementation of Tabletop Mind-Maps (TMM, see Fig. 3) was conducted with the <lb/>DiamondSpin toolkit [13]. TMM was also based on our previous experience with a <lb/>hierarchy view, namely the Personal Digital Historian (PDH) application [12], which <lb/>is dedicated to organizing family pictures according to a hierarchy of people and <lb/>concepts present in the pictures. <lb/>A TMM session starts with a root label forced to be in center. The root displays the <lb/>field to explore, which is important to keep in mind, so we duplicated the label along <lb/>a symmetry axis to have it more readable from every point of view around the table. <lb/>The mind-map is then built top-down when users create new nodes with double-<lb/>tap-and-drop interaction. This concatenation of double-tap and drag-and-drop <lb/>appeared to be natural and easy to perform with direct manipulation. The double-tap <lb/>creates a new node in the sub-hierarchy of the tapped node, while the drag-and-drop <lb/>specifies the new node position. The background color of the node represents its level <lb/>in the hierarchy (green for 1 st level, blue for 2 nd level ). <lb/>
14
+
15
+ <note place="footnote">2 <lb/> A complete list of mind-mapping software is available at www.mind-mapping.org, the market <lb/>leader being MindManager (www.mindjet.com). <lb/></note>
16
+
17
+ TMM nodes are editable. The choice of a label being a collaborative activity in <lb/>mind-mapping, this aspect had to be reproduced in TMM: we chose to allow text <lb/>input only from a single source, i.e. a physical wireless keyboard with a particular <lb/>focus management. Indeed, in a tabletop system there can be more than one focused <lb/>or selected element, as the users interact seamlessly together or in parallel. We made <lb/>the keyboard focus persistent until the Enter key is pressed. While the text is being <lb/>keyed, users can create new nodes or select other ones (e.g. to check for possible <lb/>redundant nodee s name) without interfering with the edition. The font color of the <lb/>node represents the user who created it. <lb/>Nodes of the hierarchy are freely relocatable on the table. The nodes of a sub-<lb/>hierarchy will also follow their parent node when the latter is moved on. The <lb/>orientation of the nodes is adjusted online while they are being moved on so that the <lb/>text is always oriented outwards to be readable by the nearest user. Moreover, users <lb/>can rotate the whole display if they want to change the view without changing the <lb/>arrangement of the hierarchy. <lb/>Finally, we introduced a means of creating a temporary view of a sub-hierarchy. A <lb/>given node becomes the new central root, all the items outside of its sub-hierarchy <lb/>being temporarily hidden. <lb/> 5 Experimental study <lb/> This experiment was designed to evaluate the use of a tabletop interactive application <lb/>for mind-mapping by comparing it with a control paper-and-pencil condition. <lb/> 5.1 Method <lb/>Participants. 6 groups of 4 participants took part in the experiment. Each group <lb/>included students, professors and/or employees. We excluded groups only composed <lb/>of students in order to avoid too much familiarity among participants and simulate <lb/>more realistic conditions of creativity sessions. Overall, userss age ranged from 20 to <lb/>52 years old (mean = 28.7; SD = 7.9) and each group was composed of 2 male and 2 <lb/>female participants. <lb/> Materials. For the tabletop condition, we used MERL DiamondTouch [5]: the <lb/>participants were seated around the table with the experimenter sitting aside on a <lb/>highchair. The participants interacted on TMM display with finger-input to create, <lb/>edit or move the mind-map items. The experimenter typed down the content of the <lb/>nodes using the wireless keyboard. <lb/>In the control condition the participants were seated in front of a paperboard with <lb/>the experimenter standing beside it. The experimenter used a marker pen to build the <lb/>mind-map and write down its content according to the participants indications. <lb/> Procedure. Each group had to build 2 mind-maps on different topics: 1 in the <lb/>tabletop condition and 1 in the control condition. The topics were related to the <lb/> sectors of Media and Leisure : such topics simulate potential reflection for e.g. <lb/>companies trying to find a way to diversify, searching an application for a new <lb/>technology or trying to find new markets. These 2 topics were chosen so as to be <lb/>equivalent in level of abstraction and width of scope. The order of conditions and the <lb/>assignment of topics were counterbalanced across the whole sample (see Table 1). <lb/> Table 1. Counterbalancement of conditions: For each group (A to F), this table defines the <lb/>order of the 2 conditions (Tabletop and Control) and the topic addressed in each case (Media <lb/>and Leisure sectors). <lb/>Group ID <lb/>First mind-map <lb/>Second mind-map <lb/>A <lb/>Tabletop: Media <lb/> C o n t r o l : Leisure <lb/> B <lb/>Tabletop: Leisure <lb/> C o n t r o l : Media <lb/> C <lb/>Tabletop: Media <lb/> C o n t r o l : Leisure <lb/> D <lb/>Control: Media <lb/> T a b l e t o p : Leisure <lb/> E <lb/>Control: Leisure <lb/> Tabletop: Media <lb/> F <lb/>Control: Media <lb/> T a b l e t o p : Leisure <lb/> To conduct the session, the experimenter first asks the general question What <lb/>does leisure (resp. media) make you think of? The participants freely suggest some <lb/>ideas and concepts associated to the target sector, and the experimenter writes down <lb/>the ideas the group agrees on. Once the first level of the mind-map is completed, the <lb/>same process is repeated for the second level by focusing on first-level ideas one by <lb/>one (( What does xxx make you think of? ). In this experiment the mind-maps were <lb/>limited to 2 levels and the time to build them to 10 minutes. The differences between <lb/>tabletop and control conditions in building the mind-maps are summarized in Table 2. <lb/> Table 2. Differences between tabletop and control conditions in the process of mind-mapping. <lb/>Factor <lb/>Description <lb/>Spatial position of participants <lb/>Around the table vs. in front of the paperboard <lb/>Creation of new boxes <lb/>By the participants in the tabletop condition vs. by the <lb/>experimenter in the control condition <lb/>Modification / suppression of a <lb/>box <lb/>Allowed in tabletop but not in control condition <lb/>Spatial arrangement of items <lb/>Online modifications allowed in tabletop but not in control <lb/>condition <lb/>Rotation of the mind-map <lb/>Allowed in tabletop but not in control condition <lb/>Focus on a first-level idea <lb/>Explicit in tabletop (making the rest of the mind-map <lb/>disappear) vs. implicit in control condition (whole map <lb/>always displayed) <lb/> The tabletop condition was preceded by a familiarization phase for demonstrating <lb/>the tablee s functionalities to the participants. Both tabletop and control conditions <lb/>were then video-recorded. At the end of the experiment, users had to fill in a <lb/>questionnaire to assess the following dimensions: efficiency, usability, usefulness of <lb/>the tabletop system, satisfaction, and comparison with the control condition. Users <lb/>had to quantify their impressions on 7-point scales and were particularly prompted to <lb/> complete with free qualitative comments. The whole experiment lasted about 1 hour <lb/>for each group. <lb/> 5.2 Data analysis <lb/> Inferential analyses were performed by means of ANOVAs using SPSS software. <lb/>Three dimensions were investigated: the performance in mind-mapping, the <lb/>subjective experience of users, and the collaborative behaviors. <lb/> Performance. We chose to assess the performance dimension from the <lb/>exhaustiveness of the outcome. As we lack absolute standards to evaluate a mind-map <lb/>in itself, we decided to aggregate the mind-maps of the 6 groups for the same topic <lb/>and take this as a reference to be compared to each mind-map. We rated the <lb/>exhaustiveness of the mind-maps by considering both the total number of ideas and <lb/>the number of categories of ideas in comparison to the reference. <lb/> Subjective experience. This dimension was computed from the questionnaire ratings. <lb/>The analysis processed on these data also accounted for users gender and category <lb/>(student, professor or employee). <lb/> Collaboration. The participantss collaborative behaviors were annotated from the <lb/>video-recordings of the sessions. We collected the following behaviors: assertions <lb/>(e.g. giving an idea), information requests, action requests, answers to questions, <lb/>expression of opinions, communicative gestures related to the task, and off-task talks. <lb/>The communicative gestures variable includes e.g. pointing to the map, interrupting <lb/>s.o. or requesting the speech turn by a gesture, which can be observed in both <lb/>conditions. In the tabletop condition, it also includes gesture-inputs on the table, with <lb/>the exclusion of creation / edition / suppression actions which we did not consider as <lb/>communicative gestures. <lb/>We first analyzed the raw behavioral data for each participant, and then we <lb/>converted them into percentages in order to assess the respective contribution of each <lb/>participant in the group. Such an index finally enabled us to compute the difference <lb/>between the actual collaboration pattern of each group and a theoretical perfectly-<lb/>balanced pattern (each one of the 4 participants would contribute 25%). <lb/> 5.3 Results <lb/>Performance. No significant difference appeared between tabletop and control <lb/>conditions on our index of exhaustiveness of mind-maps (F(1/5) = 0.92, NS). <lb/> Subjective experience. There was no significant effect of the condition (tabletop vs. <lb/>control) on easiness (F(1/20) &lt; 0.1, NS) and efficiency (F(1/20) = 1.02, NS) of mind-<lb/>map building. However, the tabletop was rated as significantly more pleasant to use <lb/>(F(1/20) = 10.43, p = 0.004), enabling a more pleasant communication between <lb/> participants (F(1/20) = 5.01, p = 0.037), more efficient group work (F(1/20) = 3.56, <lb/>p = 0.074) and more pleasant group work (F(1/20) = 4.23, p = 0.053) see Fig. 4. <lb/>Users gender and category had no influence on any of the previous results. <lb/> 1 <lb/> 2 <lb/>3 <lb/>4 <lb/>5 <lb/>6 <lb/>7 <lb/> Pleasantness <lb/>Communication <lb/>pleasantness <lb/>Group efficiency <lb/>Group <lb/>pleasantness <lb/>Subhective ratings <lb/> Tabletop <lb/>Control <lb/> Fig. 4. Subjective ratings of participants for the tabletop and control conditions. <lb/> Collaboration. The variables expression of opinionn and off-task talkk comprised <lb/>too many missing values to be analyzed. The other raw behavioral data showed no <lb/>significant difference in the absolute number of any of the variables, except for the <lb/>communicative gestures category: tabletop led to more communicative gestures than <lb/>control condition (F(1/22) = 3.59, p = 0.071). <lb/> Tabletop condition <lb/>Control condition <lb/>Tabletop condition <lb/>Control condition <lb/> Fig. 5. Collaboration patterns in tabletop (left) and control (right) conditions: this graph <lb/>represents the average contribution of the 4 participants ranked on a leader / follower scale. <lb/>This figure illustrates that the contributions of the participants were significantly better <lb/>balanced in the tabletop condition (p = 0.013). <lb/> The analysis of collaboration patterns showed that participants verbal <lb/>contributions (sum of all behaviors without communicative gestures) were <lb/>significantly better balanced in tabletop than in control condition (F(1/22) = 7.35, <lb/>p = 0.013) i.e. they were significantly closer to the theoretical perfectly-balanced <lb/>pattern. Fig. 5 presents the average collaboration patterns in both conditions: to obtain <lb/> this figure, we ranked the participants of each group from the one who contributed the <lb/>most (the leader) to the one who contributed the least (the follower) and averaged the <lb/>data for the 6 groups. The same result applies for communicative gestures: the <lb/>gestural contributions were significantly better balanced in tabletop than in control <lb/>condition (F(1/22) = 8.94, p = 0.007). <lb/> 6 Conclusion and future work <lb/> The tabletop condition significantly improved both subjective and collaborative <lb/>dimensions of mind-mapping. First of all, the participants found that the tabletop <lb/>system was more pleasant to use, improved group communication and collaboration <lb/>efficiency. These effects on userss impressions could be explained e.g. by the spatial <lb/>position of participants favoring social interaction, the attraction of a new technology, <lb/>and/or the more active involvement of participants in this condition. <lb/>Moreover, the behavioral analysis showed that the tabletop system enabled a better <lb/>collaboration: while the control condition showed strong leaders and followers, in the <lb/>tabletop condition the participants collaborated in a better-balanced way. Some <lb/>benefits of a tabletop system compared to a wall display or a desktop computer were <lb/>previously observed by Rogers and Lindley [10] but their setting was noticeably <lb/>different from ours: their tabletop device supported only single-touch interaction <lb/>(with a pen) and a single viewpoint (so that the participants had to sit side by side and <lb/>not around the table). They observed more interaction and role changing (swapping <lb/>the possession of the input device) in the tabletop condition: it proved easier and more <lb/>natural to change roles because of the use of a direct input device (a pen has to be <lb/>placed directly on the display whereas a mouse controls the pointer from a distance) <lb/>and because of the physical proximity of the participants to this input device (higher <lb/>in the tabletop than in the wall display condition). In our experiment the collaborative <lb/>benefits cannot be explained by any of these reasons because all 4 participants had the <lb/>same role and interaction capacity. We could tentatively explain our results by the <lb/>spatial position of people around the table, which can facilitate idea exchange, or by <lb/>the attraction of a new technology, which could prompt the participants to interact <lb/>with the tabletop interface and thus to give new ideas. The second hypothesis is less <lb/>likely because it may have resulted in higher performance in idea generation. <lb/>Therefore we hypothesize that the collaborative benefits we observed come from the <lb/>around-the-table placement of people. This assumption will be tested with a new <lb/>control condition where the participants will have to build a paper-and-pencil mind-<lb/>map around a table. This new experiment will also complete the data about the <lb/>subjective preferences expressed in the present study. <lb/>Finally, despite all the advantages of our tabletop application (subjective <lb/>engagement, better collaboration, active involvement of users, focus on first-level <lb/>ideas, flexibility of the mind-map displayy ), the experimental results showed no <lb/>significant difference in the quality of outcomes between tabletop and control <lb/>conditions. In the next steps of the project, we intend to focus more deeply on the <lb/>performance dimension and search a way to improve it. We should develop a more <lb/>accurate analysis of mind-map process and outcome to better understand the idea <lb/> production mechanism. We also plan to test the influence of innovative interaction <lb/>styles (see e.g. the paper metaphor [1]) on idea production and organization. <lb/>The global experimental process followed in this study (comparison of tabletop <lb/>and traditional paper-and-pencil condition, variables collected ) is currently being <lb/>applied to other creativity tools such as brainstorming on sticky notes in order to <lb/>investigate whether the present results apply to other situations of group creativity. <lb/>
18
+
19
+ </body>
20
+
21
+ <back>
22
+
23
+ <div type="acknowledgement">Acknowledgements <lb/> This study was supported by the ANR-RNTL DigiTable project (www.digitable.fr). <lb/></div>
24
+
25
+ <listBibl> References <lb/> 1. Besacier, G., Rey, G., Najm, M., Buisine, S., Vernier, F.: Paper metaphor for tabletop <lb/>interaction design. HCII&apos;07 Human Computer Interaction International, Lecture Notes in <lb/>Computer Science, Springer (2007). <lb/>2. Buzan, T.: The Mind Map Book. Penguin Books (1991). <lb/>3. Candy, L., Hori, K.: The digital muse: HCI in support of creativity. Interactions, 10, (2003) <lb/>44-54. <lb/>4. Csikszentmihalyi, M.: Creativity: Flow and the psychology of discovery and invention. <lb/>Harper Perennial, New York (1996). <lb/>5. Dietz, P.H., Leigh, D.: DiamondTouch: A multi-user touch technology. UIST&apos;01 <lb/>International Conference on User Interface Software and Technology, ACM Press (2001) <lb/>219-226. <lb/>6. Farooq, U.: Eureka! Past, present, and future of creativity research in HCI. ACM <lb/>Crossroads, 12, (2005) 6-11. <lb/>7. Isaksen, S.G., Dorval, K.B., Treffinger, D.J.: Creative approaches to problem solving: A <lb/>framework for change. Kendall Hunt (2000). <lb/>8. Lubart, T.: How can computers be partners in the creative process. International Journal of <lb/>Human-Computer Studies, 63, (2005) 365-369. <lb/>9. Osborn, A.F.: Applied Imagination. Principles and procedures of creative problem-solving. <lb/>Charles Scribner&apos;s Sons (1953). <lb/>10. Rogers, Y., Lindley, S.: Collaborating around vertical and horizontal large interactive <lb/>displays: Which way is best? Interacting with Computers, 16, (2004) 1133-1152. <lb/>11. Scott, S.D., Carpendale, S. (Eds.): Interacting with digital tabletops. Special issue of IEEE <lb/>Computer Graphics and Applications, vol. 26 (2006). <lb/>12. Shen, C., Lesh, N., Vernier, F.: Personal Digital Historian: Story sharing around the table. <lb/>Interactions, 10, (2003) 15-22. <lb/>13. Shen, C., Vernier, F., Forlines, C., Ringel, M.: DiamondSpin: An extensible toolkit for <lb/>around-the-table interaction. CHI&apos;04 International Conference on Human Factors in <lb/>Computing Systems, ACM Press (2004) 167-174. <lb/>14. Shneiderman, B.: Creating creativity: User interfaces for supporting innovation. ACM <lb/>Transactions On Computer-Human Interaction (TOCHI), 7, (2000) 114 -138. <lb/>15. Shneiderman, B., Fischer, G., Czerwinski, M., Resnick, M., Myers, B.: Creativity support <lb/>tools: Report from a U.S. National Science Foundation sponsored workshop. International <lb/>Journal of Human-Computer Interaction, 20, (2006) 61-77. <lb/>16. Sternberg, R.J.: Handbook of Creativity. Cambridge University Press (1998). <lb/>17. Streitz, N.A., Geißler, J., Holmer, T., Konomi, S.: I-LAND, an interactive landscape for <lb/>creativity and innovation. CHI 99 International Conference on Human Factors in <lb/>Computing Systems, ACM Press (1999) 120-127. <lb/>18. Warr, A., OO Neill, E.: Public Social Private Design (PSPD). CHI 06 International <lb/>Conference on Human Factors in Computing Systems, ACM Press (2006) 1499-1504. </listBibl>
26
+
27
+ </back>
28
+ </text>
29
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/Hadbi_11990.training.segmentation.tei.xml ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei>
3
+ <teiHeader>
4
+ <fileDesc xml:id="__Hadbi_11990"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front> COMPARISON BETWEEN SYSTEM DESIGN OPTIMIZATION <lb/>STRATEGIES FOR MORE ELECTRIC AIRCRAFT NETWORKS <lb/> Djamel Hadbi * , ** , Nicolas Retière * , Frederic Wurtz * , Xavier Roboam ** , Bruno Sareni ** <lb/> * Univ. Grenoble Alpes, G2Elab, F-38000 Grenoble, France, CNRS, G2Elab, F-38000 Grenoble, France <lb/>[email protected] <lb/> ** Université de Toulouse, LAPLACE, UMR CNRS-INPT-UPS, 2 rue Camichel, 31071 Toulouse, France <lb/>[email protected] <lb/> Abstract. Nowadays, embedded aircraft system contains electrical devices which must cooperate in safe and light <lb/>weight network. For designing such systems, different local strategies have been developed but no global <lb/>optimization has been performed so far. In this paper, we present and compare three strategies applied to the sizing <lb/>of a whole network of more electric aircraft: a simplified case study with only two components is considered to <lb/>illustrate methodological issues. The quality of the solution found from each method is compared, with regards to <lb/>the &quot; cost of the collaborative approach &quot; and the volume of data generation. This comparison should provide system <lb/>designers an evaluation of the applicability of these methods according to the nature of the design problem. <lb/> Keywords: system design, integrated design, multilevel optimization, embedded electrical system. <lb/></front>
8
+
9
+ <body> INTRODUCTION <lb/> Modern engineering products are becoming increasingly complex, particularly in industries such as <lb/>railway, aerospace and automotive [1], [2]. Conventionally, expertise and classical analysis methods, especially <lb/>those based on simulations are used, aided by optimization methods in some part of the process. Each sub-<lb/>system is designed separately by his manufacturer using his own model and process: this method is called <lb/> &quot; mechanistic approach &quot; . Another approach called &quot; simultaneous design &quot; may be developed to integrate all the <lb/>components of a system into a single optimization. This extreme approach requires full cooperation for best <lb/>results [2]. Another diametrically opposite approach enables to design a system with a minimal collaboration. <lb/>This method called Extended Pareto Fronts was developed for the design of a railway application [3]. Industrial <lb/>design imposes other constraints, integration of a large scale multidisciplinary system, privacy issues and <lb/>decision level are fundamental criteria to elaborate feasible and efficient system design method [2]. <lb/> I. <lb/>DESIGN MODEL OF A SIMPLIFIED STUDIED CASE <lb/> Figure 1. Design of an actuator by Extended Pareto Front Method <lb/> Extended Pareto Front Method (EPFM) is based on the decomposition of a system in several devices, <lb/>each of them being previously optimized separately, only coupled by global variables (i. e. I, f, V in the figure <lb/>1). It intends to limit the communication of data between the suppliers. Only Extended Pareto Optimal Fronts <lb/>(space solution) are provided by suppliers to the system designer. First, a necessary condition has to be checked: <lb/>in a suitable decomposition, the system objective function has to be calculated by means of the objective <lb/>functions of sub-problems. If so, a classification of all problem variables allows identifying global variables that <lb/>describe couplings on the system. In the above example, there are N=3 couplings variables, i.e. current (I), <lb/>frequency (f) and voltage (V) of the bus, completed by O=2 objective functions for weight (W) and device <lb/>efficiency (K). So a N+O-dimension solution space is built for each component of the system. Gathering the <lb/>Extended Pareto Fronts of all amenities, the system manufacturer shall be able to make tradeoffs within the sets <lb/>of solutions in order to obtain the best solution at the system level. Of course, not any combination of equipment <lb/>is possible; the manufacturer must be careful of the consistency of coupling variables that ensure the <lb/>compatibility of the chosen amenities. <lb/> II. <lb/> DESIGN OF A SIMPLIFIED EMBEDDED ELECTRIC NETWORK <lb/> Figure 2. An example of a simplified embedded electric network <lb/> Embedded electric networks contain a high number of sources and loads connected to several buses: the <lb/>issue is then to analyze applicability of optimization methods to this class of complex system. In our study, <lb/>related to the development of strategies for system design, we initially relied on a &quot; simplified &quot; network, which is <lb/>voluntary limited to a single load in order to establish and compare optimization methods. It consists of a single <lb/>generating channel, connected with a unique non-linear load. This channel includes a generator, a rectifier and an <lb/>output filter. The load comprises an actuator, an inverter and an input filter. We investigate the design of this <lb/>channel (especially the filtering device sizing) using the three approaches previously presented: the optimization <lb/>goal is to minimize the whole network weight in compliance with quality standards; (current and voltage <lb/>harmonics are limited to a maximum threshold in a frequency band). Let us note that this case study consists of a <lb/>single objective optimization for which the previous method (EPFM) of course remains applicable. <lb/> III. <lb/>QUALITATIVE COMPARISON OF RESULTS <lb/> As shown in the figure 2 (left part), the values of the objective function obtained by two collaborative <lb/>methods, the simultaneous design approach and Extended Pareto Front Method are better than the value of the <lb/>objective function obtained with a conventional method ( &quot; mechanistic approach &quot; ) based on expertise and <lb/>classical analysis methods. <lb/>Collaboration level and calculation cost are conflicting characteristics as illustrated of the right part of <lb/>figure 2: higher is the collaboration level, lower the necessary computational cost to reach the perfect solution. <lb/>The mapping of the space solution of a sub-problem grows exponentially with the number of coupling variables <lb/>(if there are P values of M input global variable, P M optimizations are done). High level of collaboration means <lb/>that subsystem manufacturers need to share and communicate their models for the design process. So it&apos;s easy <lb/> for the system integrator to improve the whole system generating a small amount of data. On the other side, a <lb/>low level of collaboration means that each subsystem designer keeps its design secret, so that he must provide <lb/>more data to aircraft manufacturer to enable him finding the optimal combination of subsystems. <lb/>. <lb/> Figure 3. Comparison of optimization approaches in terms of objective function and costs (collaboration vs <lb/>computation cost) <lb/> CONCLUSION <lb/> Through our work, we managed to check out conditions are necessary to elaborate a design strategy for <lb/>complex system: collaboration and data generation. Thanks to this, we have proposed a multilevel collaborative <lb/>design strategy while limiting the data exchanged by the designer of system components which is a requirement <lb/>for industrial systems such as embedded networks. <lb/></body>
10
+
11
+ <listBibl> REFERENCES <lb/> [1] F. Moussouni-Messad, &quot; Multi-level and multi-objective design optimization tools for handling complex systems &quot; , these (PHD) de <lb/>l&apos;Ecole Centrale de Lille 2009. <lb/> [2] X. Roboam &amp; al, &quot; Integrated design by optimization of electrical energy systems&apos;&apos;, edited by ISTE Wiley, 2012, ISBN 978-1-84821-<lb/> 389-0. <lb/>[3] H. Nguyen-Huu , N. Retière, F. Wutz, &quot; A new approach for building the global Pareto border of an electromagnetic system by using <lb/> the Pareto borders of each sub-component of the system &quot; , IEEE-CEFC 2008 (International Conference on Electromagnetic Field <lb/>Computation), ID: OC1-4 in proceeding of the conference. <lb/></listBibl>
12
+
13
+ <body>0,6 <lb/> 0,8 <lb/>1 <lb/>1,2 <lb/> W <lb/> tot (kg) <lb/> Simultaneous desigh aproach <lb/>Extended Pareto Front Method <lb/>Mechanistic approach <lb/> ma comb atio of subsy <lb/> Collaboration level <lb/>Optimization cost <lb/> Simultaneous design approach <lb/>E. P. F. M. <lb/>Mechanistic <lb/>Approach <lb/>High <lb/>Medium <lb/>Minimal <lb/>1 2 <lb/>2 x P M</body>
14
+
15
+ </text>
16
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/IJPA-13-549.training.segmentation.tei.xml ADDED
@@ -0,0 +1,95 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.549-559 <lb/>549 <lb/>Available at: http://ijpa.tums.ac.ir <lb/>Original Article <lb/>Experimental Study on Plasmodium berghei, Anopheles <lb/>Stephensi, and BALB/c Mouse System: Implications for Malaria <lb/>Transmission Blocking Assays <lb/>Hossein DEHGHAN 1 , *Mohammad Ali OSHAGHI 1 , Seyed Hassan MOSA-KAZEMI 1 , <lb/>Mohammad Reza ABAI 1 , Fatemeh RAFIE 1 , Mehdi NATEGHPOUR 2 , Habib <lb/>MOHAMMADZADEH 3,4 , Leila FARIVAR 2 , Mulood MOHAMMADI BAVANI 5 <lb/>1. Dept. of Medical Entomology &amp; Vector Control, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran <lb/>2. Dept. of Parasitology and Medical Mycology, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran <lb/>3. Cellular and Molecular Research Center, Urmia University of Medical Sciences, Urmia, Iran <lb/>4. Dept. of Medical Parasitology and Mycology, School of Health, Urmia University of Medical Sciences, Urmia, Iran <lb/>5. Dept. of Medical Entomology, School of Health, Urmia University of Medical Sciences, Urmia, Iran <lb/>Received 22 Aug 2017 <lb/>Accepted 16 Jan 2018 <lb/>Abstract <lb/>Background: Plasmodium berghei is a rodent malaria parasite and has been very valuable <lb/>means in the progress of our understanding of the essential molecular and cellular biolo-<lb/>gy of the malaria parasites. Availability of hosts such as mice and vectors such as Anophe-<lb/>les stephensi has made this parasite a suitable system to study the parasite-host and vector-<lb/>parasite relationships. <lb/>Methods: This study was performed at Medical Entomology and Parasitology laborato-<lb/>ries of the School of Public Health, Tehran University of Medical Sciences, Iran in 2016. <lb/>The investigation was carried out to describe life cycle and parameters influencing <lb/>maintenance of the parasite within the mice or the mosquito. <lb/>Results: Results have revealed details and addressed some parameters and points influ-<lb/>ence maintenance of various life stages of the parasite including merozoites, <lb/>macrogametocytes, ookinetes, oocysts and sporozoites in the laboratory model P. berghei-<lb/>A. stephensi-BALB/c mouse. Injection of fresh infected blood results in higher <lb/>gametocytemia in the animals. The more injected parasites result in earlier and higher <lb/>parasitemia and exfelagellation centers in the mice blood. However, the highest number <lb/>of infected mosquitoes and oocysts formation were observed when the parasitemia and <lb/>exflagellation centers per microscopic field were 9% and 3.6 in the infected mice respec-<lb/>tively. The infected mosquitoes should be maintained on 8% (w/v) fructose, 0.05% <lb/>(w/v) PABA at 20±1 °C and 50%-80% relative humidity. <lb/>Conclusion: This study helps to understand the biology of vertebrate-parasite and mos-<lb/>quito-malaria interactions that may aid in the development of a new generation of <lb/>drug/vaccine and vector-based measures for malaria control. <lb/>Keywords: <lb/>Plasmodium berghei, <lb/>Anopheles stephensi, <lb/>BALB/c, <lb/>Malaria, <lb/>Lifecycle <lb/>* Correspondence <lb/>Email: <lb/>[email protected] <lb/>Iranian Society of Parasitology <lb/>http://isp.tums.ac.ir <lb/>Iran J Parasitol <lb/>Open access Journal at <lb/>http://ijpa.tums.ac.ir <lb/>Tehran University of Medical <lb/>Sciences Publication <lb/>http://tums.ac.ir <lb/></front>
8
+
9
+ <note place="headnote">Dehghan et al.: Experimental Study on Plasmodium berghei, Anopheles stephensi, … <lb/></note>
10
+
11
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
12
+
13
+ <page>550 <lb/></page>
14
+
15
+ <body>Introduction <lb/>alaria remains one of the most <lb/>prevalent tropical and infectious <lb/>diseases in the world, with an es-<lb/>timated more than 200 million clinical cases <lb/>every year (1). Attempts to restrict disease <lb/>transmission have focused on using imagicides <lb/>and adulticides to kill the larvae and adult <lb/>mosquito vectors accordingly, the develop-<lb/>ment of drugs to treat infected individuals, <lb/>and hindering vector-human contact using <lb/>physical barriers and repellents. The Plasmodi-<lb/>um life cycle requires two different organisms <lb/>to survive and develop: the vertebrate host <lb/>and the female of Anopheles mosquitoes (1, 2). <lb/>In recent years, transmission-blocking strat-<lb/>egies (TBS) are considered as a potential target <lb/>for malaria control focused on mosquito stag-<lb/>es of malaria parasites. The three main TBS <lb/>included gametocytocidal drugs, transmission-<lb/>blocking vaccines, and replacing wild mosqui-<lb/>toes with the ones harboring refractory traits. <lb/>The first strategy is focused on exploring the <lb/>gametocytocidal activity of commercially <lb/>available anti-malarial drugs. The second TBS <lb/>is based on immunological attack on sexual or <lb/>mosquito stages of the parasite life cycle using <lb/>different antigens including surface proteins of <lb/>gametocytes and gametes, zygotes and ooki-<lb/>netes, as well as, antigens in the mosquito <lb/>midgut surface such as calreticulin and alanyl <lb/>aminopeptidase (AnAPN1), and other mole-<lb/>cules such as chitinase (3-9). Some of these <lb/>vaccine candidates are currently in preclinical <lb/>phase and being considered for the first trials <lb/>in humans (10). The third TBS is founded on <lb/>changing the mosquito population towards a <lb/>Plasmodium-resistant vector phenotype using <lb/>an effector molecule triggering malaria refrac-<lb/>toriness. Some of effector molecules were <lb/>shown promising including phospholipase A2 <lb/>(PLA2) to inhibit ookinete invasion and SM1 <lb/>thought to block recognition sites for sporo-<lb/>zoites and ookinetes (11, 12). Introduction of <lb/>effector molecules can be achieved by releas-<lb/>ing genetically modified (GM) mosquitoes <lb/>(13), naturally refractory mosquitoes (11), arti-<lb/>ficial gene drive mechanisms (14, 15) and <lb/>third-party modified organisms (paratransgen-<lb/>esis). The third party could be a symbiotic bacte-<lb/>rium or fungus (16-20), and even viruses (21), <lb/>modified to express effector molecule/s inside <lb/>the mosquitoes. The midgut microbiome of <lb/>some Anopheles is characterized and could be <lb/>used as paratransgenesis means (22-27). <lb/>To assess the effectiveness of potential TBS <lb/>that are able to disrupt the life cycle of the <lb/>parasite in the mosquito vector, most studies <lb/>have focused on P. falciparum and P. berghei and <lb/>on the African mosquito, A. gambiae, and the <lb/>Asian mosquito, A. stephensi. However, P. <lb/>berghei-A. stephensi system is widely used when <lb/>studying mosquito-Plasmodium interactions <lb/>since P. berghei is a rodent malaria species, safe, <lb/>unable of infecting humans, its manageability <lb/>and exceptional robustness (28, 29). Numer-<lb/>ous studies have described establishment of P. <lb/>berghei life cycle and the rate of successful de-<lb/>velopment from one life stage to the next <lb/>within the mosquito and vertebrate hosts. <lb/>Since every developmental transition of the <lb/>parasite exhibits strong level of success, here <lb/>we tried to reveal essential requirements and <lb/>optimized situation for a P. berghei-A. stephensi-<lb/>BALB/c mouse laboratory model. Under-<lb/>standing the impact of variables influencing <lb/>on life stage of the parasite is important to <lb/>compare between potential TBS and their <lb/>evaluations. <lb/>Materials and Methods <lb/>The host and vector <lb/>All gametocytes were raised in BALB/c mice <lb/>and transmitted to A. stephensi strain Beech, <lb/>where maintained at 20±1 °C and 50%-80% <lb/>RH and fed on 8% fructose/0.05% para-amino-<lb/>benzoic acid (PABA) (28). The mice were used <lb/>to feed 24 h starved 3-5 d-old A. stephensi fe-<lb/>males for 30 min. In these experiments, more <lb/>than 50 fed mosquitoes were assayed per sample. <lb/>M <lb/></body>
16
+
17
+ <note place="headnote">Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.549-559 <lb/></note>
18
+
19
+ <page>551 <lb/></page>
20
+
21
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
22
+
23
+ <body>24-h after feeding, unfed mosquitoes were re-<lb/>moved by aspirator. Mosquitoes were then <lb/>maintained on fructose [8% (w/v) fructose, <lb/>0.05% (w/v) PABA] at 20±1 °C and 50%-80% <lb/>relative humidity. Day 9-10 post-feeding, mid-<lb/>guts of a subset of lived specimens were dissect-<lb/>ed in a drop 0.5% mercurochrome and their <lb/>midguts examined for oocysts by light micros-<lb/>copy. The remainder of each mosquito batch <lb/>was incubated a further 8-10 d before counting <lb/>the salivary gland sporozoites. <lb/>The parasite <lb/>We followed the protocols previously de-<lb/>scribed (30). To avoid the effect of parasite <lb/>genetic variability, we used clones of the ro-<lb/>dent malarial parasite P. berghei clone ANKA <lb/>2.34, a gift from Prof. Marcelo Jacob-Lorena <lb/>of Johns Hopkins Bloomberg School of Pub-<lb/>lic Health, Department of Molecular Microbi-<lb/>ology and Immunology, Malaria Research In-<lb/>stitute, Baltimore, USA. <lb/>All procedures were performed in accordance <lb/>with the terms of the Iran Animals (Scientific <lb/>Procedures) Act Project License and were ap-<lb/>proved by the Tehran University of Medical Sci-<lb/>ences Ethical Review Committee (No. 26231). <lb/>The parasites were maintained in 4-10 wk <lb/>old female BALB/c mice by serial mechanical <lb/>passages (up to a maximum of 8 passages). To <lb/>keep gametocyte infectivity to the mosquitoes, <lb/>hyper-reticulocytosis was induced 2-3 d be-<lb/>fore infection by treating mice with 100 μL <lb/>intraperitoneal (i.p.) phenylhydrazinium chlo-<lb/>ride (PH; 10 mg/mL in PBS) per 10 g mouse. <lb/>Mice infections were monitored on Giemsa-<lb/>stained tail-blood smears. Infections in mice <lb/>were done on d 2-5 when a low but rising <lb/>gametocytaemia succeeded (31). <lb/>Results <lb/>We only addressed the hints and fine details <lb/>normally not indicated exclusively in literature. <lb/>Impact of anesthetic materials on BALB/c <lb/>The impact of kind of anesthetized materials <lb/>on development of the parasite within mice, we <lb/>tested two chemicals: Acepromazine and mix <lb/>of Rompun® (Xylazine) and Vetalar® (Keta-<lb/>min) (1:Xylazine /2: Ketamin/3: PBS). We in-<lb/>jected intraperitoneal (i.p.) 50 μl (per mouse 15 <lb/>gr) of the chemicals which anesthetizes the ani-<lb/>mal for 30-45 min. Acepromazine caused con-<lb/>siderable reduction in the animal body tempera-<lb/>ture. This hindered the parasite development <lb/>within the animal and also the cold animal was <lb/>not attractive for mosquitoes to take blood meal. <lb/>Hyper-reticulocytosis in BALB/c mice <lb/>Hyper-reticulocytosis was induced 2-3 d be-<lb/>fore infection by treating the mice with 100 μl <lb/>(10 mg/mL in PBS) intraperitoneal (i.p.) <lb/>phenylhydrazinium chloride (PH). Treating the <lb/>BALB/c mice with 100 μL intraperitoneal (i.p.) <lb/>phenylhydrazinium chloride (PH; 10 mg/mL in <lb/>PBS) per 10 g animal weight was enough to <lb/>induce hyper-reticulocytosis in the animals. <lb/>Parasite storage <lb/>To store parasite-infected blood for further <lb/>experiments it is recommended to use cardiac <lb/>puncture method to harvest blood samples <lb/>from infected mice. The blood should be col-<lb/>lected into heparinized tubes and was immedi-<lb/>ately mixed with equal volume of PBS (30% <lb/>glycerin), transferred to freezer (-20 °C) for 30 <lb/>min, and then stored in -196 °C. It is possible <lb/>to harvest 1.2 -1.3 and up to 1.7 ml blood <lb/>from a BALB/c mouse weight 15-25 g. <lb/>Injection of fresh/frozen infectious blood <lb/>The frozen infectious blood is used to infect <lb/>noninfected mouse by injection of 200-300 <lb/>uli.p. infected blood to a new mouse (20-25 g). <lb/>It is recommended to thaw the frozen blood, <lb/>incubate at room temperature up to 37 °C for <lb/>a few minutes prior to injection. To make <lb/>proper infection using infected frozen blood, <lb/>it is recommended to inject the blood as soon <lb/>as possible without treating the mice with PH. <lb/>Appropriate parasitemia will appear in the <lb/>mice several days post frozen-blood injection <lb/>that is much later than the ones received fresh <lb/>infected blood (Fig. 1). <lb/></body>
24
+
25
+ <note place="headnote">Dehghan et al.: Experimental Study on Plasmodium berghei, Anopheles stephensi, … <lb/></note>
26
+
27
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
28
+
29
+ <page>552 <lb/></page>
30
+
31
+ <body>2 3 4 5 6 7 8 9 10 12 13 <lb/>0 <lb/>20 <lb/>40 <lb/>60 <lb/>80 <lb/>100 <lb/>10(4-5) Fresh parasites <lb/>10(9) Frozen parasites <lb/>Day <lb/>Parasitemia Rate <lb/>Fig. 1: Temporal and level of P. berghei (ANKA <lb/>clone 2.34) parasitemia in two groups of BALB/c <lb/>mice received either10 9 parasites per ml frozen <lb/>infectious blood or10 4-5 parasites per ml fresh in-<lb/>fectious blood. Bars indicate median and error <lb/>The number of parasites in the fresh blood <lb/>was 10 4 to 10 5 times less than the frozen blood. <lb/>Injection of fresh infected blood with more <lb/>than 50% parasitemia results in high <lb/>gametocytemia in the animals but may cause <lb/>early death in them. In two trails, one out of <lb/>five (20%) and three out of ten (30%) mice <lb/>died three days post-injection. In an inde-<lb/>pendent experiment, the mean death time for <lb/>the mice received 10 9 parasites/ml (p/ml) in-<lb/>fectious blood occurred at day 6 whereas the <lb/>mean death time occurred at day 10 (4 d later) <lb/>for the one received 10 4-5 p/ml infectious <lb/>blood (Fig. 2). <lb/>Infectious symptoms in the mice <lb/>The symptoms of infected mice included se-<lb/>cluding, temperature dysfunction, intense and <lb/>fast breathing and increased heart rate, hump-<lb/>ing, blistering body hair (ruffled fur) particu-<lb/>larly on back, eye weeping or closing, inactiva-<lb/>tion and reluctance to move in high parasitem-<lb/>ia, hard walking, losing skin, ear discoloration, <lb/>anemia and lightening the blood color from <lb/>dark red to light red/fawn, RBC reduction, <lb/>cerebral complications and death (Fig. 3). <lb/>10 4-5 <lb/>10 9 <lb/>0 <lb/>5 <lb/>10 <lb/>15 <lb/>Time of death (day) <lb/>Fig. 2: Time of death in two groups of BALB/c <lb/>mice received either 10 9 or 10 4-5 P. berghei (ANKA <lb/>clone 2.34) p/ml fresh infectious blood. Bars <lb/>show means and SDs <lb/>Obtaining P0 infected BALB/c mice <lb/>To obtain P0 infected mouse, the following <lb/>hints are useful: 1) Injection of PABA into the <lb/>non-infected mice 3 d prior to injection of in-<lb/>fected blood, 2) parasitemia in the donor in-<lb/>fected mouse should be more than 40%, 3) in-<lb/>jection of 230-250 ul of infectious blood to <lb/>non-infected BALB/c mouse is enough, 4) pri-<lb/>or to injection, the mouse should be main-<lb/>tained in 23±2 °C, 12:12 dark: light photoperi-<lb/>od regime, 5) two to three days post injection is <lb/>the best time for counting exflagellation centers. <lb/>Effect of number of parasites on para-<lb/>sitemia and exfelagellation centers <lb/>There is positive correlation between <lb/>number of injected parasites and level and or <lb/>temporal parasitemia in the BALB/c mice (Fig. <lb/>4). The more injected parasites result in earlier <lb/>and higher parasitemia and exfelagellation cen-<lb/>ters in the mice blood. <lb/>Oocyst formation in A. stephensi mosquitoes <lb/>Feeding the mosquitoes on the infectious <lb/>mice with high parasitemia or exflagellation <lb/>centers in each microscopic field result in high <lb/>oocyst formation, however, the number of <lb/>mosquitoes with no oocyst increased dramati-<lb/>cally (&gt;47-85%) (Fig.5). <lb/></body>
32
+
33
+ <note place="headnote">Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.549-559 <lb/></note>
34
+
35
+ <page>553 <lb/></page>
36
+
37
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
38
+
39
+ <body>Fig. 3: The symptoms of BALB/c mice infected by P. berghei ANKA clone 2.34 <lb/>Fig. 4: Effect of number of injected P. berghei (ANKA clone 2.34) parasites on the temporal and level of para-<lb/>sitemia (A) and exfelagellation centers (B) in BALB/c mice. Bars indicate medians and errors <lb/>High [7-20] exflagellation centers per micro-<lb/>scopic field or high parasitemia (&gt;17%) inten-<lb/>sify number of mosquitoes with no oocyst. <lb/>On the other hand, using infectious mice with <lb/>2.4-3.9 exflagellation centers in each micro-<lb/>scopic field and 9%-17% parasitemia results in <lb/>high frequency of infected mosquitoes and <lb/>oocysts formation (Fig. 6). The best results <lb/>were observed when the parasitemia and <lb/>exflagellation centers per microscopic field <lb/>was 9% and 3.6 in the infected mice respec-<lb/>tively (Fig. 6). <lb/>The best time for feeding mosquitoes on <lb/>infected mouse is when 3-4 exflagellation cen-<lb/>ters exist in each microscopic field. Based on <lb/>the number of parasites injected into the <lb/>mouse, this situation happens throughout days <lb/>3-4 and 5-6 post-injection when respectively <lb/>10 9 and 10 4-5 parasites injected (Fig. 3). <lb/>Overall, 4-7 d old mosquitoes should be <lb/>maintained on 8% (w/v) fructose, 0.05% <lb/>(w/v) PABA at 20±1 °C and 50%-80% rela-<lb/>tive humidity from the time of adult emer-<lb/>gence. The female mosquitoes were kept <lb/>starved for 12 h before blood meal. The hun-<lb/>ger females should be separated and trans-<lb/>ferred to a new cage. The infected mouse was <lb/>anesthetized by Ketamin/Xylasine and put on <lb/>top of the cage where the room temperature <lb/>was about 20±1 °C. <lb/></body>
40
+
41
+ <note place="headnote">Dehghan et al.: Experimental Study on Plasmodium berghei, Anopheles stephensi, … <lb/></note>
42
+
43
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
44
+
45
+ <page>554 <lb/></page>
46
+
47
+ <body>4 0 % <lb/>P , 7 <lb/>E <lb/>3 0 % <lb/>P , 2 0 <lb/>E <lb/>2 5 % <lb/>P , 1 8 <lb/>E <lb/>2 2 % <lb/>P , 1 9 <lb/>E <lb/>1 7 <lb/>% <lb/>P , 3 . 9 <lb/>E <lb/>9 % <lb/>P , 3 . 6 <lb/>E <lb/>1 0 % <lb/>P , 2 . 4 <lb/>E <lb/>0 <lb/>20 <lb/>40 <lb/>60 <lb/>80 <lb/>100 <lb/>Rate of mosquitoes with no oocyst <lb/>Fig. 5: Relationship between rate of parasitemia <lb/>(P) /number of exflagellation centers (E) of P. <lb/>berghei (ANKA clone 2.34) per microscopic field in <lb/>BALB/c blood and frequency of A. stephensi mos-<lb/>quito midgut with no oocyst <lb/>It is highly recommended to warm the ani-<lb/>mal body by putting a warm cotton pad on its <lb/>body and changing the pad every 5 min. This <lb/>helps to keep the animal body warm enough <lb/>and nullify cooling effect of the anesthetic <lb/>compounds. Totally, 20-30 min is enough for <lb/>most (&gt;90%) of females to take blood meal. <lb/>The fed mosquitoes should be maintained on <lb/>8% (w/v) fructose, 0.05% (w/v) PABA at <lb/>20±1 °C and 50%-80% relative humidity <lb/>through the experiments. The fructose-PABA <lb/>vial should be changed every two days. The <lb/>vial stocks should be kept at 20 °C before use. <lb/>Usually, about 60%-80% of fed females will <lb/>be alive 15-20 d post blood meal at optimized <lb/>condition. The mosquitoes lay eggs 4-5 d post <lb/>blood meal at 20±1 °C. <lb/>The parasite zygotes and ookinetes could be <lb/>seen accordingly 1 and 24 h after infected <lb/>blood intake. To support survival rate of the <lb/>female mosquitoes, it is recommended to of-<lb/>fer two more blood meals (from a non-<lb/>infected mouse) on day six and 12, following <lb/>laying eggs by the females. Day 9-10 and 20 <lb/>post blood meal is the best time for oocyst <lb/>and sporozoite observation accordingly. Oo-<lb/>cysts can be counted by phase contrast mi-<lb/>croscopy of freshly dissected midguts. Sporo-<lb/>zoites can be similarly counted on days 20 <lb/>onwards by hemocytometry of dissected sali-<lb/>vary glands of alive mosquitoes immediately <lb/>prior to dissection. <lb/>If the mosquitoes had sporozoites they are <lb/>ready for parasite transmission to a new host. <lb/>To ensure transferring the parasites from in-<lb/>fected mosquitoes to a new mouse, although <lb/>theoretically one mosquito is enough, it is rec-<lb/>ommended to use at least 10 infected females <lb/>per mouse because all of the mosquitoes may <lb/>not be infected and one mosquito may not <lb/>able to inoculate enough parasite to the host. <lb/>Younger mice with 10-15 g weight are pre-<lb/>ferred to older ones with more weight for par-<lb/>asite transmission. In our experiments, the P0 <lb/>parasites were observed on day six post mos-<lb/>quito bites in the mice (n=3) who received 13-<lb/>16 mosquito bites. Other mice group (n=3) <lb/>received fewer bites (7-8 and 9-10bites) the <lb/>parasites appeared on average day 15 and 10 <lb/>post mosquito bites accordingly (Fig.7). Fur-<lb/>ther analysis showed that rate of sporozoite <lb/>infection in the mosquitoes used in this exper-<lb/>iment was 80%. <lb/>Following mosquito bites, the sporozoites <lb/>enter the liver cells (hepatocytes) less than an <lb/>hour (30 min), and then after 48 h, the mero-<lb/>zoites form are released into the bloodstream <lb/>to attack red blood cells (RBC). In RBC, they <lb/>replicate asexually in 22-24-h cycle. When par-<lb/>asitemia of P0 reach to 20% in the mice blood, <lb/>the blood could be collected (10 ul blood + 90 <lb/>ul PBS/mouse 15g) and injected to new mice <lb/>to prepare P1 parasites. It takes almost one <lb/>week until the parasites (P1) appears in the <lb/>mice. When the parasitemia (P1) reach 30%-<lb/>50%, the blood can be taken and stored in <lb/>freezer for further investigation. For preparing <lb/>P2 parasites, the mouse blood containing P1 <lb/>parasites with 20%-30% parasitemia could be <lb/></body>
48
+
49
+ <note place="headnote">Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.549-559 <lb/></note>
50
+
51
+ <page>555 <lb/></page>
52
+
53
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
54
+
55
+ <body>injected into new mice. The new host (mouse) <lb/>should be treated with PH three days in ad-<lb/>vance of receiving the parasites. The P2 para-<lb/>sites will appear in the mice 3-4 d later. The <lb/>mice with P2-P7 parasites are suitable for <lb/>mosquito feeding for TBS assays. <lb/>0 <lb/>100 <lb/>200 <lb/>300 <lb/>400 <lb/>600 <lb/>650 <lb/>700 <lb/>750 <lb/>Parasitemia rate <lb/>Exflagellation centers <lb/>40 <lb/>7 <lb/>30 <lb/>20 <lb/>25 <lb/>18 <lb/>22 <lb/>19 <lb/>17 <lb/>3.9 <lb/>9 <lb/>3.6 <lb/>10 <lb/>2.4 <lb/>Mosquitoes (N) <lb/>Mean of oocyst/gut(+P) <lb/>Oocyst Range/gut(+P) <lb/>Mean/gut <lb/>Median/gut <lb/>Frequency of zero oocyst <lb/>(%) <lb/>20 <lb/>7.7 <lb/>1-19 <lb/>1.15 <lb/>0 <lb/>17 <lb/>(85) <lb/>27 <lb/>129.5 <lb/>11-350 <lb/>29.8 <lb/>0 <lb/>14 <lb/>(51.9) <lb/>97 <lb/>56.6 <lb/>1-700 <lb/>60 <lb/>0 <lb/>46 <lb/>(47.4) <lb/>18 <lb/>125.9 <lb/>1-350 <lb/>62.9 <lb/>0 <lb/>9 <lb/>(50) <lb/>17 <lb/>76.5 <lb/>3-267 <lb/>67.5 <lb/>40 <lb/>2 <lb/>(11.8) <lb/>16 <lb/>65.1 <lb/>0-249 <lb/>57 <lb/>40 <lb/>2 <lb/>(12.5) <lb/>27 <lb/>72.8 <lb/>1-250 <lb/>72.8 <lb/>21 <lb/>5 <lb/>(18.5) <lb/>No. Oocyts/gut <lb/>Fig. 6: Effect of parasitemia/number of exflagellation centers per microscopic field of P.berghei (ANKA clone <lb/>2.34) on oocyst formation in A.stephensi midgut. Bars show median. +p: mosquitoes with at least one oocyst <lb/>6 7 8 9 10 11 12 13 14 15 16 17 18 <lb/>6 <lb/>10 <lb/>15 <lb/>No of mosquito bites <lb/>Infection time (Day) <lb/>Fig. 7: Relationship between numbers of P. berghei infected A. stephensi bites and infection time in the BALB/c <lb/>blood. Prevalence (number of mosquitoes with at least one oocyst) of infected mosquitoes was 80%. The bars <lb/>indicate means with ranges <lb/></body>
56
+
57
+ <note place="headnote">Dehghan et al.: Experimental Study on Plasmodium berghei, Anopheles stephensi, … <lb/></note>
58
+
59
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
60
+
61
+ <page>556 <lb/></page>
62
+
63
+ <body>Discussion <lb/>We have revealed fine details for maintain-<lb/>ing P. berghei-A. stephensi-BALB/c mouse sys-<lb/>tem which is crucial and very helpful for stud-<lb/>ies involving malaria transmission blocking <lb/>assays. Although the system we explained here <lb/>was established for P. berghei-A. stephensi, it has <lb/>the potential applicability to other malaria vec-<lb/>tors. P. berghei is able to complete its life cycle <lb/>in different mosquito species (30, 32, 33). <lb/>However, when studying mosquito-malaria <lb/>interactions, one should consider the vector <lb/>intrinsic factors comprising Anopheles species <lb/>gametocyte activating factor(s), pH, tempera-<lb/>ture, and quality and quantity of blood inges-<lb/>tion which may influence on the mosquito <lb/>infectivity outcome (34, 35). <lb/>We used BALB/c mice as a vertebrate host <lb/>for P. berghei, however, other rodents including <lb/>Swiss Webster mice, Theiler&apos;s Original (TO-<lb/>outbred), and C57BL/6 have been used for <lb/>similar systems. The quality and quantity of <lb/>inhibitory factors present in the infected blood <lb/>of different mice such as antibodies, metabo-<lb/>lites, and cytokines (36, 37) may enhance or <lb/>diminish the mosquito infectivity outcome. <lb/>Further investigation needs to clarify the ef-<lb/>fect of the parasite/host combinations on the <lb/>maintenance of the parasite in vivo (38). <lb/>For the lab with less security level, this sys-<lb/>tem present here is preferred to the systems <lb/>use P. falciparum since P. berghei is a rodent ma-<lb/>laria species and is safe and unable of infecting <lb/>humans (29). However, the systems involved <lb/>human <lb/>malaria <lb/>species <lb/>such <lb/>as <lb/>A.stephensi/A.gambiae-P. falciparum might vary <lb/>from the rodent system we present here par-<lb/>ticularly in the gametocytemia and the mosqui-<lb/>to infectivity. <lb/>In our experiments, the number of infec-<lb/>tious parasites ingested by the mosquito from <lb/>the mouse has great effect on the success of <lb/>Plasmodium infection in the mosquitoes. Alt-<lb/>hough the number oocystes increased in the <lb/>mosquitoes fed on high parasitemic mouse, <lb/>number of infectious mosquitoes reduced in-<lb/>tensely which is in accordance with the finding <lb/>of other researchers (35, 39-44). Therefore it <lb/>is highly recommended for researchers who <lb/>use P. berghei-systems to use infected mouse or <lb/>its infected blood with low gametocyte densi-<lb/>ties (2-3 exflagellation centers per microscopic <lb/>field) when they evaluate TBS and calculate <lb/>value of interventions. Here we showed that <lb/>more inoculated sporozoites accelerate the <lb/>parasitemia in the host. In addition to the <lb/>number of mosquito bites, the number of in-<lb/>oculated sporozoites and the host-related fac-<lb/>tors may affect the quality and quantity of par-<lb/>asitemia in the host. This is in agreement with <lb/>previous studies indicating the positive rela-<lb/>tionship between parasitemia and number of <lb/>infected mosquito bites or sporozoites (31, 45, <lb/>46). <lb/>The results of this study showed that be-<lb/>tween the two parameters of parasitemia and <lb/>exflagellation centers, number of exflagellation <lb/>centers has more influence on the rate of in-<lb/>fected mosquitoes and proper number of oo-<lb/>cyst formation. However, this depends on <lb/>mosquito strain, amount of blood intake by <lb/>mosquito, number of active gametocytes in <lb/>the blood, and some other unknown factors in <lb/>the blood (8, 47). <lb/>Conclusion <lb/>The results of this study showed that various <lb/>factors influence on the maintenance of P. <lb/>berghei-A. stephensi-BALB/c mouse system. <lb/>This information helps to understand the bi-<lb/>ology of vertebrate-parasite and mosquito-<lb/>malaria interactions that may aid in the devel-<lb/>opment of a new generation of drug/vaccine <lb/>and vector-based measures for malaria con-<lb/>trol. <lb/></body>
64
+
65
+ <div type="acknowledgement">Acknowledgements <lb/>This study was supported by the Tehran <lb/>University of Medical Sciences, Iran Grant <lb/>number 26231. <lb/></div>
66
+
67
+ <note place="headnote">Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.549-559 <lb/></note>
68
+
69
+ <page>557 <lb/></page>
70
+
71
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
72
+
73
+ <div type="annex">Conflict of interest <lb/>The authors declare that there is no conflict <lb/>of interests. <lb/></div>
74
+
75
+ <listBibl>References <lb/>1. <lb/>WHO. Fact Sheet: World Malaria Report. <lb/>Geneva: World Health Organization, 2015. <lb/>http://www.who.int/malaria/media/world-<lb/>malaria-report-2015/en <lb/>2. <lb/>Vega-Rodríguez J, Ghosh AK, Kanzok SM <lb/>et al. Multiple pathways for Plasmodium <lb/>ookinete invasion of the mosquito midgut. <lb/>Proc Natl Acad Sci U S A. 2014; <lb/>111(4):E492-500. <lb/>3. <lb/>Carter R. Transmission blocking malaria <lb/>vaccines. Vaccine. 2001; 19: 2309-2314. <lb/>4. <lb/>Tomas AM, Margos G, Dimopoulos G et <lb/>al. P25 and P28 proteins of the malaria <lb/>ookinete surface have multiple and partially <lb/>redundant functions. EMBO J. 2001; <lb/>20(15):3975-83. <lb/>5. <lb/>Pradel G. Proteins of the malaria parasite <lb/>sexual stages: expression, function and po-<lb/>tential for transmission blocking strategies. <lb/>Parasitology. 2007; 134(Pt.14):1911-29. <lb/>6. <lb/>Saxena AK, Wu Y, Garboczi DN. Plasmodi-<lb/>um p25 and p28 surface proteins: potential <lb/>transmission-blocking vaccines. Eukaryot <lb/>Cell. 2007; 6(8):1260-5. <lb/>7. <lb/>Gholizadeh S, Basseri HR, Zakeri S et al. <lb/>Cloning, expression and transmission-<lb/>blocking activity of anti-PvWARP, malaria <lb/>vaccine candidate, in Anopheles stephensimy-<lb/>sorensis. Malar J. 2010; 9: 158. <lb/>8. <lb/>Bousema T, Drakeley C. Epidemiology and <lb/>infectivity of Plasmodium falciparum and Plas-<lb/>modium vivax gametocytes in relation to ma-<lb/>laria control and elimination. Clin Microbiol <lb/>Rev. 2011; 24(2):377-410. <lb/>9. <lb/>Borhani Dizaji N, Basseri HR, Naddaf SR <lb/>et al. Molecular characterization of calre-<lb/>ticulin from Anopheles stephensi midgut cells <lb/>and functional assay of the recombinant <lb/>calreticulin with Plasmodium berghei ooki-<lb/>netes. Gene. 2014; 550(2):245-52. <lb/>10. Nikolaeva D, Draper SJ, Biswas S. Toward <lb/>the development of effective transmission <lb/>blocking vaccines for malaria. Expert Rev <lb/>Vaccines. 2015; 14(5):653-80. <lb/>11. Zieler H, Keister DB, Dvorak JA. A snake <lb/>venom phospholipase A(2) blocks malaria <lb/>parasite development in the mosquito mid-<lb/>gut by inhibiting ookinete association with <lb/>the midgut surface. J Exp Biol. 2001;204(Pt <lb/>23):4157-67. <lb/>12. Ito J, Ghosh A, Moreira LA et al. Transgen-<lb/>ic anopheline mosquitoes impaired in <lb/>transmission of a malaria parasite. Nature. <lb/>2002; 417(6887):452-5. <lb/>13. Knols BG1, Bossin HC, Mukabana WR et <lb/>al. Transgenic mosquitoes and the fight <lb/>against malaria: managing technology push <lb/>in a turbulent GMO world. Am J Trop Med <lb/>Hyg. 2007; 77(6 Suppl):232-42. <lb/>14. Windbichler N, Menichelli M, Papathanos <lb/>PA et al. A synthetic homing endonuclease-<lb/>based gene drive system in the human ma-<lb/>laria <lb/>mosquito. <lb/>Nature. <lb/>2011; <lb/>473(7346):212-5. <lb/>15. Gantz VM, Jasinskiene N, Tatarenkova O <lb/>et al. Highly efficient Cas9-mediated gene <lb/>drive for population modification of the <lb/>malaria vector mosquito Anopheles stephensi. <lb/>Proc Natl Acad Sci U S A. 2015; <lb/>112(49):E6736-43. <lb/>16. Durvasula RV, Gumbs A, Panackal A et al. <lb/>Prevention of insect-borne disease: an ap-<lb/>proach using transgenic symbiotic bacteria. <lb/>Proc Natl Acad Sci U S A. 1997; <lb/>94(7):3274-8. <lb/>17. Riehle MA, Moreira CK, Lampe D et al. <lb/>Using bacteria to express and display anti-<lb/>Plasmodium molecules in the mosquito mid-<lb/>gut. Int J Parasitol. 2007; 37(6):595-603. <lb/>18. Wang S, Ghosh AK, Bongio N et al. <lb/>Fighting malaria with engineered symbiotic <lb/>bacteria from vector mosquitoes. Proc Natl <lb/>Acad Sci U S A. 2012; 109(31):12734-9. <lb/>19. Maleki-Ravasan N, Oshaghi MA, Afshar D <lb/>et al. Aerobic bacterial flora of biotic and <lb/>abiotic compartments of a hyperendemic <lb/>Zoonotic Cutaneous Leishmaniasis (ZCL) <lb/>focus. Parasit Vectors. 2015; 8: 63. <lb/>20. Dehghan H, Oshaghi MA, Moosa-Kazemi <lb/>SH et al. Dynamics of Transgenic Entero-<lb/>bacter cloacae Expressing Green Fluores-<lb/>cent Protein Defensin (GFP-D) in Anopheles <lb/></listBibl>
76
+
77
+ <note place="headnote">Dehghan et al.: Experimental Study on Plasmodium berghei, Anopheles stephensi, … <lb/></note>
78
+
79
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
80
+
81
+ <page>558 <lb/></page>
82
+
83
+ <listBibl>stephensi Under Laboratory Condition. J Ar-<lb/>thropod Borne Dis. 2017; 11(4):515-532. <lb/>21. Ren X, Hoiczyk E, Rasgon JL. Viral para-<lb/>transgenesis in the malaria vector Anopheles <lb/>gambiae. PLoS Pathog. 2008; 4(8):e1000135. <lb/>22. Straif SC, Mbogo CN, Toure AM et al. <lb/>Midgut bacteria in Anopheles gambiae and A. <lb/>funestus (Diptera: Culicidae) from Kenya and <lb/>Mali. J Med Entomol. 1998; 35(3):222-6. <lb/>23. Chavshin AR, Oshaghi MA, Vatandoost H <lb/>et al. Identification of bacterial microflora in <lb/>the midgut of the larvae and adult of wild <lb/>caught Anopheles stephensi: a step toward <lb/>finding suitable paratransgenesis candidates. <lb/>Acta Trop. 2012; 121(2):129-34. <lb/>24. Chavshin AR, Oshaghi MA, Vatandoost H <lb/>et al. Escherichia coli expressing a green fluo-<lb/>rescent protein (GFP) in Anopheles stephensi: <lb/>a preliminary model for paratransgenesis. <lb/>Symbiosis. 2013; 60: 17-24. <lb/>25. Chavshin AR, Oshaghi MA, Vatandoost H <lb/>et al. Isolation and identification of cultura-<lb/>ble bacteria from wild Anopheles culicifacies, a <lb/>first step in a paratransgenesis approach. <lb/>Parasit Vectors. 2014; 7: 419. <lb/>26. Chavshin AR, Oshaghi MA, Vatandoost H <lb/>et al. Malpighian tubules are important de-<lb/>terminants of Pseudomonas transstadial trans-<lb/>mission and longtime persistence in Anophe-<lb/>les stephensi. Parasit Vectors. 2015; 8: 36. <lb/>27. Wang S, Dos-Santos ALA, Huang W et al. <lb/>Driving mosquito refractoriness to Plasmo-<lb/>dium falciparum with engineered symbiotic <lb/>bacteria. Science. 2017; 357(6358):1399-<lb/>1402. <lb/>28. Blagborough AM, Delves MJ, Ramakrish-<lb/>nan C et al. Assessing transmission block-<lb/>ade in Plasmodium spp. Methods Mol Biol. <lb/>2013; 923:577-600. <lb/>29. Xu J, Hillyer JF, Coulibaly B et al. Wild <lb/>Anopheles funestus Mosquito Genotypes Are <lb/>Permissive for Infection with the Rodent <lb/>Malaria Parasite, Plasmodium berghei. PLoS <lb/>One. 2013; 8(4):e61181. <lb/>30. Sinden RE, Butcher GA, Beetsma AL. <lb/>Maintenance of the Plasmodium berghei life <lb/>cycle, In: Doolan DL(Ed.), Malaria Meth-<lb/>ods and Protocols. Humana Press Inc., To-<lb/>towa, NJ, 2002; pp. 25-40. <lb/>31. Dearsly AL, Sinden RE, Self IA. Sexual de-<lb/>velopment in malarial parasites: gametocyte <lb/>production, fertility and infectivity to the <lb/>mosquito vector. Parasitology. 1990; 100 Pt <lb/>3:359-68. <lb/>32. Vaughan JA, Narum D, Azad AF. Plasmodi-<lb/>um berghei ookinete densities in three <lb/>anopheline species. J Parasitol. 1991; <lb/>77(5):758-61. <lb/>33. Alavi Y, Arai M, Mendoza J et al. The dy-<lb/>namics of interactions between Plasmodium <lb/>and the mosquito: a study of the infectivity <lb/>of Plasmodium berghei and Plasmodium gallina-<lb/>ceum, and their transmission by Anophelesste-<lb/>phensi, Anopheles gambiae and Aedesaegypti. Int <lb/>J Parasitol. 2003; 33(9):933-43. <lb/>34. Billker O, Shaw MK, Margos G et al. The <lb/>roles of temperature, pH and mosquito fac-<lb/>tors as triggers of male and female gameto-<lb/>genesis of Plasmodium berghei in vitro. Para-<lb/>sitology. 1997; 115 ( Pt 1):1-7. <lb/>35. Arai M, Billker O, Morris HR et al. Both <lb/>mosquito-derived xanthurenic acid and a <lb/>host blood-derived factor regulate gameto-<lb/>genesis of Plasmodium in the midgut of the <lb/>mosquito. Mol Biochem Parasitol. 2001; <lb/>116(1):17-24. <lb/>36. Motard A, Landau I, Nussler A et al. The <lb/>role of reactive nitrogen intermediates in <lb/>modulation of gametocyte infectivity of ro-<lb/>dent malaria parasites. Parasite Immunol. <lb/>1993; 15(1):21-6. <lb/>37. Fleck SL, Butcher GA, Sinden RE. Plasmo-<lb/>dium berghei: serum-mediatedinhibition of in-<lb/>fectivity of infected mice to Anopheles ste-<lb/>phensi mosquitoes. Exp Parasitol. 1994; <lb/>78(1):20-7. <lb/>38. Sinden RE, Barker GC, Paton MJ et al. Fac-<lb/>tors regulating natural transmission of Plas-<lb/>modium berghei to the mosquito vector, and <lb/>the cloning of a transmission-blocking im-<lb/>munogen. <lb/>Parassitologia. <lb/>1993; <lb/>35 <lb/>Suppl:107-12. <lb/>39. Ponnudurai T, Lensen AH, Van Gemert GJ <lb/>et al. Infectivity of cultured Plasmodium falci-<lb/>parum gametocytes to mosquitoes. Parasit-<lb/>ology. 1989; 98 Pt 2:165-73. <lb/>40. Robert V, le Goff G, Gouagna LC et al. <lb/>Kinetics and efficiency of Plasmodium falcipa-<lb/>rum development in the midguts of Anophe-<lb/>les gambiae,A. funestus and A. nili. Ann Trop <lb/>Med Parasitol. 1998; 92(1):115-8. <lb/></listBibl>
84
+
85
+ <note place="headnote">Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.549-559 <lb/></note>
86
+
87
+ <page>559 <lb/></page>
88
+
89
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
90
+
91
+ <listBibl>41. Gouagna LC, Mulder B, Noubissi E et al. <lb/>The early sporogonic cycle of Plasmodium fal-<lb/>ciparum in laboratory-infected Anopheles gam-<lb/>biae: an estimation of parasite efficacy. Trop <lb/>Med Int Health. 1998; 3(1):21-8. <lb/>42. Drakeley CJ, Secka I, Correa S et al. Host <lb/>haematological factors influencing the <lb/>transmission of Plasmodium falciparum game-<lb/>tocytes to Anopheles gambiaes.s. mosquitoes. <lb/>Trop Med Int Health. 1999; 4(2):131-8. <lb/>43. Boudin C, Van Der Kolk M, Tchuinkam T <lb/>et al. Plasmodium falciparum transmission <lb/>blocking immunity under conditions of low <lb/>and high endemicity in Cameroon. Parasite <lb/>Immunol. 2004; 26(2):105-10. <lb/>44. Pollitt LC, Churcher TS, Dawes EJ et al. <lb/>Costs of crowding for the transmission of <lb/>malaria parasites. Evol Appl. 2013; 6(4):617-<lb/>29. <lb/>45. Mordmüller B, Supan C, Sim KL et al. Di-<lb/>rect venous inoculation of Plasmodium falci-<lb/>parum sporozoites for controlled human <lb/>malaria infection: a dose-finding trial in two <lb/>centres. Malar J. 2015; 14: 117. <lb/>46. Conteh S, Anderson C, Lambert L et al. <lb/>Grammomys surdaster, the Natural Host for <lb/>Plasmodium berghei Parasites, as a Model to <lb/>Study Whole-Organism Vaccines against <lb/>Malaria. Am J Trop Med Hyg. 2017; <lb/>96(4):835-841. <lb/>47. Lyimo IN, Keegan SP, Ranford-Cartwright <lb/>LC et al. The impact of uniform and mixed <lb/>species blood meals on the fitness of the <lb/>mosquito vector Anopheles gambiae s.s: does a <lb/>specialist pay for diversifying its host spe-<lb/>cies diet? J Evol Biol. 2012; 25(3):452-60. </listBibl>
92
+
93
+
94
+ </text>
95
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/IJPA-13-567.training.segmentation.tei.xml ADDED
@@ -0,0 +1,89 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.567-576 <lb/>567 <lb/>Available at: http://ijpa.tums.ac.ir <lb/>Original Article <lb/>Molecular Identification and Intra-species Variations among <lb/>Leishmania infantum Isolated from Human and Canine Visceral <lb/>Leishmaniasis in Iran <lb/>*Abdolhossein DALIMI 1 , Anita MOHAMMADIHA 1 , Mehdi MOHEBALI 2,3 , Asad <lb/>MIRZAEI 4 , Mohammadreza MAHMOUDI 5 <lb/>1. Dept. of Parasitology, Faculty of Medical Sciences, Tarbiat Modares University, Tehran, Iran <lb/>2. Dept. of Medical Parasitology and Mycology, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran <lb/>3. Center for Research of Endemic Parasites of Iran (CREPI), Tehran University of Medical Sciences, Tehran, Iran <lb/>4. Dept. of Parasitology, Faculty of Medicine, Ilam University of Medical Sciences, Ilam, Iran <lb/>5. Dept. of Microbiology and Parasitology, Faculty of Medicine, Guilan University of Medical Sciences, Rasht, Iran <lb/>Received 05 Aug 2017 <lb/>Accepted 09 Feb 2018 <lb/>Abstract <lb/>Background: In Iran, both forms of cutaneous (CL) and visceral leishmaniasis (VL) <lb/>have been re-ported; so the accurate species identification of the parasite(s) and the anal-<lb/>ysis of genetic diversity are necessary. <lb/>Methods: The investigation was conducted from 2014 to 2015 in the northwest and <lb/>south of Iran, where VL is endemic (7 provinces). Blood samples of patients and infect-<lb/>ed dogs were collected and sera separated for serologic examinations (DAT, rK39). <lb/>Spleen or bone marrow samples from infected dogs were also collected to confirm the <lb/>infection. DNAs of 70 samples amplified by targeting a partial sequence of ITS (18S <lb/>rRNA-ITS1-5.8S rRNA-ITS2) gene. All the amplicons were sequenced and analyzed <lb/>with restriction fragment length polymorphism (RFLP) using the TaqI enzyme. <lb/>Results: The cause of all 70 VL cases, were L. infantum, so, the dominant specie is L. <lb/>infantum. The sequencing results of all VL cases and RFLP analysis corroborate each <lb/>other. Discrimination of Iranian Leishmania isolates using ITS gene gives us this oppor-<lb/>tunity to detect, identify and construct the phylogenetic relationship of Iranian isolates. <lb/>In addition, detection and differentiation of Leishmania spp. DNA was confirmed by <lb/>amplification of variable area of the minicircle kDNA (conserved sequence blocks <lb/>(CSB)). <lb/>Conclusion: Low divergence and high likelihood were seen among L. infantum isolates <lb/>of human and dogs from Iran with a very slight divergence was seen between isolates <lb/>from northwest and south of Iran, thus grouped in a unique clad. No correlation was <lb/>observed between intraspecies divergence and geographic distribution of the isolates. <lb/>Keywords: <lb/>Phylogeny, <lb/>Leishmania infantum, <lb/>PCR-RFLP, <lb/>Sequencing, <lb/>Iran <lb/>* Correspondence <lb/>Email: <lb/>[email protected] <lb/>Iranian Society of Parasitology <lb/>http://isp.tums.ac.ir <lb/>Iran J Parasitol <lb/>Open access Journal at <lb/>http://ijpa.tums.ac.ir <lb/>Tehran University of Medical <lb/>Sciences Publication <lb/>http://tums.ac.ir <lb/></front>
8
+
9
+ <note place="headnote">Dalimi et al.: Molecular Identification and Intra-species Variations … <lb/></note>
10
+
11
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
12
+
13
+ <page>568 <lb/></page>
14
+
15
+ <body>Introduction <lb/>isceral Leishmaniasis (VL) is an in-<lb/>fectious systemic parasitic disease, <lb/>widespread in the Old and New <lb/>World. Leishmania parasites are obligatory in-<lb/>tracellular protozoans of the genus Leishmania, <lb/>which infects humans as well as domestic and <lb/>wild animals, transmitted via the infective <lb/>bites of Phlebotomine sand flies. Several clini-<lb/>cal syndromes with a wide clinical spectrum of <lb/>the severe disease ranges are exhibited in VL, <lb/>particularly in children and immunocompro-<lb/>mised patients (1). <lb/>VL spreads over the world in tropical and <lb/>subtropical regions reported an average annual <lb/>incidence of 500000 cases of the visceral form <lb/>(2). In six countries: India, Bangladesh, Sudan, <lb/>South Sudan, Ethiopia and Brazil, more than <lb/>90% of global VL cases occur (2, 3). <lb/>VL is caused mainly by L. infantum in our <lb/>country, follows by splenomegaly and hepa-<lb/>tomegaly, distributes in three foci (four prov-<lb/>inces) in northwest and south of Iran. In a few <lb/>cases, viscerotropic leishmaniasis can occur by <lb/>L. tropica, with no specific manifestation (4, 5); <lb/>and also, L. infantum can be a causative agent <lb/>of CL (6). Dogs (Canis familiaris) and some <lb/>wild canids such as wolves, jackals, and foxes <lb/>are the main reservoir hosts for VL (7). Fur-<lb/>thermore, VL has been reported in a sporadic <lb/>form in Iran, but reports from endemic areas <lb/>in northwestern and south of the country are <lb/>100-300 new cases annually (8). <lb/>In endemic areas as well as in Iran, where <lb/>two or more species are the causative agent of <lb/>VL, the accurate identification of the para-<lb/>site(s) and the analysis of genetic diversity are <lb/>necessary. Prognosis of the disease and choos-<lb/>ing and assessing a specific chemotherapeutic <lb/>regimen, effective control of the disease and <lb/>avoiding the disease transmission are the main <lb/>targets of these kind of the studies (9, 10). <lb/>The main molecular diagnostic technique in <lb/>discriminating of Leishmania parasites in any <lb/>kind of infected tissues is PCR-based. Different <lb/>genetic targets and various post-PCR tech-<lb/>niques provide a wide range of investigations <lb/>with various gains, all over the world (11-14). <lb/>In recent years, few studies have been car-<lb/>ried out to reveal the genetic diversity of Ira-<lb/>nian isolates by different methods, various <lb/>geographical regions for sampling and sample <lb/>sizes on the reservoir or final hosts (15-17). <lb/>The main aim of the present study was to as-<lb/>sess the similarities and proximity among spe-<lb/>cies responsible for the disease in the country. <lb/>This can lead to national policies in order to <lb/>narrow the protocols of treatments and pre-<lb/>vention of diseases in the country. <lb/>Therefore, we have used a conventional PCR <lb/>that amplifies the wide region of ITS for identi-<lb/>fication followed by sequencing and RFLP, <lb/>constructed the phylogenetic tree and analyzed <lb/>a number of visceral host-infecting Leishmania <lb/>isolates from different endemic areas of Iran. <lb/>Materials and Methods <lb/>Study Area <lb/>The study was conducted over a couple of <lb/>years (2014-2015), carried out in 7 provinces <lb/>of Iran. Seventy cases with clinical manifesta-<lb/>tion of VL were enrolled in the study: 39 pa-<lb/>tients and 31 isolates from infected dogs. Fars, <lb/>Boushehr, Ardabil, Alborz, East-Azerbaijan, <lb/>Qom and Tehran are the provinces involved <lb/>in this study (Fig. 1). <lb/>Human: The isolates obtained from HVL <lb/>were from Ardabil Province, northwest of <lb/>Iran, and the others from Tehran, Fars, Qom <lb/>and Bushehr provinces. <lb/>Dog: Thirty-one domestic dogs selected <lb/>from endemic areas from East-Azerbaijan, <lb/>Ardabil, and Alborz Provinces (Table 1). <lb/>Specimens <lb/>The characteristics and geographical origins <lb/>of the humans (39 isolates) and dogs (31 iso-<lb/>lates) included in this study are listed in Table <lb/>1. The sampling collection in our investigation <lb/>was performed passively. <lb/>V <lb/></body>
16
+
17
+ <note place="headnote">Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.567-576 <lb/></note>
18
+
19
+ <page>569 <lb/></page>
20
+
21
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
22
+
23
+ <body>Fig. 1: The samples were collected from 7 provinces; Fars, Boushehr, Ardabil, Alborz, East-Azerbaijan, Qom <lb/>and Tehran provinces which indicated with (L.i) in the map <lb/>Table 1: List of 70 Iranian strains of L.infantum characterized by ITS-PCR-RFLP. Isolates were from visceral <lb/>cases of leishmaniasis of Iran (2014-2015) <lb/>No. <lb/>Province <lb/>Sample No. <lb/>City <lb/>Disease <lb/>Source <lb/>Species <lb/>1 <lb/>Fars <lb/>9 <lb/>Kazeroun <lb/>HVL <lb/>Human <lb/>L.infantum <lb/>2 <lb/>Tehran <lb/>2 <lb/>Bumehen <lb/>HVL <lb/>Human <lb/>L.infantum <lb/>3 <lb/>Boushehr <lb/>3 <lb/>Borazjan <lb/>HVL <lb/>Human <lb/>L.infantum <lb/>4 <lb/>Arbors <lb/>7 <lb/>Karaj <lb/>CVL <lb/>Canine <lb/>L.infantum <lb/>5 <lb/>Ardabil <lb/>Qom <lb/>14 <lb/>Meshkin-shahr <lb/>Qom <lb/>HVL <lb/>Human <lb/>L.infantum <lb/>6 <lb/>12 <lb/>CVL <lb/>Canine <lb/>L.infantum <lb/>7 <lb/>East-Azerbaijan <lb/>2 <lb/>Ahar <lb/>HVL <lb/>Human <lb/>L.infantum <lb/>Human: For the passive survey, all suspect-<lb/>ed VL patients reported at least three clinical <lb/>signs, including abdominal distension, pale-<lb/>ness, and fever for at least two-week duration, <lb/>were referred to local health care centers and <lb/>were examined and confirmed by the physi-<lb/>cians, were enrolled in the study. At least 3 ml <lb/>blood samples were collected by trained health <lb/>care workers by venipuncture into 10 ml poly-<lb/>propylene tubes and processed 4-10 h after <lb/>collection. Approximately 2 ml of blood was <lb/>used for serum separation. Blood was centri-<lb/>fuged at 800 gr for 5-10 min and the serums <lb/>were stored at -20 °C. <lb/>Dogs: All the suspected CVL dogs which re-<lb/>ferred to health care centers by their owners; <lb/>were examined by a veterinarian. Based on clin-<lb/>ical evaluation (at least one clinical sign of CVL, <lb/>including loss of weight, lymphadenopathy, dry <lb/>exfoliative dermatitis, skin ulcers, periorbital <lb/>alopecia, diffuse alopecia, or onychogryphosis) <lb/>and DAT results, dogs were chosen. Spleen or <lb/>bone marrow aspirations were only carried out <lb/>from dogs after euthanizing or anesthesia with <lb/>acepromazine or Ketamine. <lb/>Sampling was conducted after gaining accu-<lb/>rate information about the place of infection <lb/>origin. Samples and data sheets were referred <lb/>to Parasitology and Entomology laboratory at <lb/>the Faculty of Medical Sciences of Tarbiat <lb/>Modares University. <lb/>Parasitological Examination <lb/>Spleen/bone marrow smears were collected <lb/>from dead dogs suspected of the disease from <lb/>different regions of Iran, then fixed with <lb/>methanol, stained with Giemsa, and examined <lb/>microscopically for the presence of Leishmania <lb/>amastigotes under high magnification (1000X) <lb/></body>
24
+
25
+ <note place="headnote">Dalimi et al.: Molecular Identification and Intra-species Variations … <lb/></note>
26
+
27
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
28
+
29
+ <page>570 <lb/></page>
30
+
31
+ <body>(18). Blood samples (3-5 ml) were taken from <lb/>cases (humans-dogs). <lb/>Negative and positive control: Five hu-<lb/>man samples exhibiting no Leishmania antibod-<lb/>ies (DAT ˗˗ ) and 5 dog DAT ˗˗ sera from non-<lb/>endemic areas with no history of VL consti-<lb/>tuted negative controls. Positive control and <lb/>human negative controls were prepared by the <lb/>Leishmaniasis laboratory at the School of <lb/>Public Health (SPH), TUMS. Canine negative <lb/>controls were prepared from the Veterinary <lb/>Faculty of Tehran University of Medical Sci-<lb/>ences, TUMS. Negative and positive control <lb/>and clinical samples were applied for PCR in <lb/>the same condition. <lb/>Serological Tests <lb/>All serum samples were examined by direct <lb/>agglutination test (DAT) and rK39 (recombi-<lb/>nant K39 antigen-based immunochromato-<lb/>graphic strip test) dipstick for detection of <lb/>Leishmaniaʼs antibodies. <lb/>DAT: The Iranian strain of L. infantum was <lb/>used for the preparation of DAT antigen in <lb/>the Leishmaniasis Laboratory, at the School of <lb/>Public Health and Institute of Public Health <lb/>Research, Tehran University of Medical Sci-<lb/>ences. The principal phases of the procedure <lb/>for making DAT antigen were mass produc-<lb/>tion of promastigotes of L. infantum <lb/>[MCAN/IR/07/Moheb-gh. (GenBank acces-<lb/>sion no. FJ555210)] (8). Peripheral blood <lb/>samples from all of 70 VL cases (human and <lb/>canine) were collected into tubes with sodium <lb/>citrate anticoagulant 4% (Merck, Germany) <lb/>for PCR testing and into tubes without antico-<lb/>agulant for DAT. All samples used for serolo-<lb/>gy and PCR were stored at −20 °C until use. <lb/>Based on many investigations carried out in <lb/>Iran, the cut-off values were determined anti-<lb/>Leishmania antibodies titers at ≥1:320 with <lb/>clinical signs considered as the disease of vis-<lb/>ceral leishmaniasis for the dogs and ≥1:3200 <lb/>with clinical signs considered as Leishmania <lb/>infection for human (8, 19). <lb/>rK39 RDT: For all the 70 subjects, rK39 <lb/>was performed from plasma samples accord-<lb/>ing to manufacturer&apos;s instruction provided as <lb/>product inserts (Cypress Diagnostic Company, <lb/>Belgium). The dipsticks were briefly placed <lb/>into 50µl of serum. After 1-4 min a red con-<lb/>trol line and, if positive, a second line ap-<lb/>peared on the test field. The test is based on a <lb/>combination of the protein-A colloidal gold <lb/>conjugate and rK39 Leishmania antigen to de-<lb/>tect anti-Leishmania antibody in serum or <lb/>plasma. Negative controls (10 samples) were <lb/>also negative by Dipstick rK39. <lb/>Culture of Reference Strains of Leishmania <lb/>Reference strains of Leishmania infantum was <lb/>MCAN/IR/96/LON49 that stored in liquid <lb/>nitrogen. Culture was carried out in biphasic <lb/>culture media (prepared from nutrient agar <lb/>containing 10% whole rabbit blood overlaid <lb/>with liver infusion tryptose broth containing <lb/>100-200 UI/ml penicillin G and 1 μg/ml <lb/>streptomycin with 10%-20% heat-inactivated <lb/>fetal bovine serum (Atlanta Biological, Atlanta, <lb/>CA). The inoculated cultures were incubated <lb/>at 21°C for up to six weeks and examined <lb/>weekly for the presence of promastigotes. <lb/>Meanwhile, Schneider Insect (HIMEDIA) and <lb/>RPMI1640 (GIBCO) media were used for <lb/>mass production of promastigotes. <lb/>DNA Isolation <lb/>Blood samples: Prior to DNA extraction <lb/>from the blood samples, 1 mL distilled water <lb/>was added to 300 µL samples followed by vor-<lb/>texing and centrifuging at 4000 gr for 5 min, to <lb/>completely remove interfering hemoglobin mol-<lb/>ecules. This step was repeated three times, and <lb/>finally, the pellets were washed with PBS (14). <lb/>A noteworthy result obtained in previous <lb/>studies to order to completely remove inter-<lb/>fering hemoglobin molecules from the sam-<lb/>ples prior to DNA extraction is; washing with <lb/>distilled water. <lb/>DNA extraction: DNA was extracted with <lb/>the DNG-plus Extraction Kit (Cinnagen, Iran) <lb/>according to the manufacturer&apos;s instructions. <lb/></body>
32
+
33
+ <note place="headnote">Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.567-576 <lb/></note>
34
+
35
+ <page>571 <lb/></page>
36
+
37
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
38
+
39
+ <body>The DNA pellet was dissolved in 50 μL of ster-<lb/>ile distilled water and incubated in a water bath <lb/>at 65 °C for 5 min. DNA concentration and <lb/>quality were determined using Nanodrop ND-<lb/>1000 Spectrophotometer (Nanodrop Technol-<lb/>ogies, Wilmington, DE, USA) at 260 and 280 <lb/>nm. DNA samples with A260/A280 ratios be-<lb/>tween 1.8 and 2 were selected and stored at -<lb/>20 °C for further analysis. <lb/>PCR Amplification by ITS-primers <lb/>For the first amplification primers were de-<lb/>signed based on the ITS region that identified, <lb/>including forwarding primer MOF (5&apos;-<lb/>GCAGCTGGATCATTTTCCGATG-3&apos;) and re-<lb/>verse <lb/>primer <lb/>MOR <lb/>(5&apos;-<lb/>GAATTCAACTTCGCGTTGGCC-3&apos;). The PCR <lb/>product size stays between 800 and 860 bp. <lb/>RFLP Analysis of Amplified ITS-gene. <lb/>Restriction fragment length polymorphism <lb/>(RFLP) analysis was performed on the ITS <lb/>amplicons which obtained from 70 blood <lb/>samples and 3 the reference strains, by using <lb/>the TaqI (1 μL) (Promega, USA) without prior <lb/>purification. After using the restriction en-<lb/>zyme obtained fragments were subjected to <lb/>electrophoresis in 2% agarose (Sigma-Aldrich, <lb/>St. Louis, MO) at 80V in 1x TAE buffer, <lb/>stained with safe stain (5 μL/100 mL), and <lb/>visualized and photographed using a UV <lb/>transilluminator. The obtained restricted <lb/>fragments were compared with the molecular <lb/>profiles of the WHO reference strains of L. <lb/>infantum (MCAN/IR/96/LON49). <lb/>Confirming of Identification of Leishma-<lb/>nia Species by kDNA-snPCR <lb/>A semi-nested PCR for detection of Leish-<lb/>mania spp. DNA was performed for amplifica-<lb/>tion of variable area of the minicircle kDNA <lb/>(with a slight modification) (20). The combi-<lb/>nation of primers LINR4 (forward), LIN17 <lb/>(reverse) and LIN19 (reverse) was used in a <lb/>seminested PCR technique. These primers <lb/>were designed within the conserved area of <lb/>the minicircle and contained conserved se-<lb/>quence blocks (CSB), CSB3, CSB2, and CSB1, <lb/>respectively (21). The mixture was incubated <lb/>at 94 ºC for 5 min followed by 30 cycles, each <lb/>consisting of 30 sec at 94 ºC, 30 sec at 52 ºC <lb/>and 1 min at 72 ºC. After the last cycle, the <lb/>extension was continued for a further 5 min. <lb/>For the second amplification first PCR prod-<lb/>uct was added to PCR-mix with 1μM, LIN19 <lb/>primer for 33 cycles under the conditions as <lb/>follows: 94 ºC for the 30 sec, 58 ºC for 30 sec <lb/>and 72 ºC for 1min) and the final extension at <lb/>72 ºC for 10 min. <lb/>Banding patterns of L. infantum, L. tropica <lb/>and L. major were 720, 760 and 560bp, respec-<lb/>tively; visualized on 2% agarose gel stained <lb/>with safe stain. <lb/>Sequencing and phylogenetic Analyses <lb/>The PCR products from visceral Leishmania <lb/>isolates (Table 1) were extracted from the gel <lb/>using a Vivantis Gel Purification kit (Vivantis, <lb/>Malaysia) according to the manufacturer&apos;s pro-<lb/>tocols, were sequenced using the same forward <lb/>and reverse primers used for amplification by an <lb/>ABI 3730 sequencer (Bioneer, Daejeon, South <lb/>Korea). The sequences were edited and manual-<lb/>ly checked with BioEdit Sequence Alignment <lb/>Editor (22), aligned (data not shown) and com-<lb/>pared with sequences from Critidia fasciculate, <lb/>Trypanosoma cruzi and Leishmania infantum <lb/>/Uzbekistan <lb/>by <lb/>ClustalX <lb/>2.12 <lb/>(23) <lb/>(http://www.clustal.org/clustal2/). The similari-<lb/>ties among our sequences were calculated (data <lb/>not shown) and phylogenetic tree (Fig. 1) was <lb/>constructed by Maximum Likelihood method in <lb/>Tamura 3 parameter option for DNA sequences <lb/>with the complete deletion procedure, by using <lb/>MEGA6 software (Molecular Evolutionary Ge-<lb/>netic Analysis Version 6) (24). The bootstrap <lb/>scores were calculated for 1000 replicates. <lb/>Ethical Approval <lb/>The trial was reviewed and approved by the <lb/>Ethics Committee of Tarbiat Modares Universi-<lb/>ty as well as the ethical committee of the Center <lb/>for Diseases Control of Iran in accordance with <lb/>the Helsinki Declaration and guidelines (Project <lb/></body>
40
+
41
+ <note place="headnote">Dalimi et al.: Molecular Identification and Intra-species Variations … <lb/></note>
42
+
43
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
44
+
45
+ <page>572 <lb/></page>
46
+
47
+ <body>No: 92013166). The patients and dogs&apos; owners <lb/>were aware that their samples (blood or spleen <lb/>and bone marrow) were needed for diagnosis of <lb/>the disease. Physicians obtained the written the <lb/>consents from the patients and dogs&apos; owners. <lb/>Results <lb/>Parasitological Results <lb/>Totally, 31 caught dogs and 3 references <lb/>were checked for the presence of Leishmaniaʼs <lb/>parasites in the spleen or bone marrow smears <lb/>by microscopy and all were positive. <lb/>Serological Test: DAT &amp; rK39 <lb/>DAT: Serological responses were obtained <lb/>for all 39 human sera (+ve ≥1:3200) and 31 <lb/>dog sera (+ve ≥1:320) and all 70 cases were <lb/>foun to be positive. All healthy patients and <lb/>dogs exhibited no antibody titer with DAT. <lb/>rK39: Anti-Leishmania antibody detected in all <lb/>70 serums. <lb/>Leishmania Identification by kDNA-<lb/>snPCR and ITS-PCR-RFLP Analysis <lb/>kDNA-snPCR: Leishmania spp. kDNA <lb/>were detected by seminested-PCR in 3 refer-<lb/>ences. All 70 samples showed the pattern of L. <lb/>infantum. <lb/>ITS-PCR-RFLP: Amplification of ITS-<lb/>rDNA from all 70 Leishmania isolates obtained <lb/>from VL cases and 3 references, were approx-<lb/>imately 800-860 bp. Digestion of amplicons <lb/>with the Taq1 enzyme produced banding pat-<lb/>terns, including the fragments of 326, 277, 142 <lb/>and 70 bp for L. infantum, 416, 296, 141 and <lb/>26 bp for L. major and 276, 193, 129, 115, 68 <lb/>and 28 bp for L. tropica. All samples showed <lb/>the pattern of L. infantum (Fig. 2). <lb/>Sequencing, Similarities and Phylogenetic <lb/>Tree <lb/>The numbers above the branches indicate <lb/>the percentage of bootstrap samplings. There <lb/>was no clear grouping among the 8 isolates <lb/>according to their geographical origin (Fig. 3). <lb/>Phylogenetic trees using Maximum Likelihood <lb/>(Fig. 3) showed intra-specific variations among <lb/>L. infantum isolates in this study and some other <lb/>mentioned parasites extracted from GenBank <lb/>(FJ001632/T.cruzi -FN398341/Uzbekistan). <lb/>Analysis of ITS sequence in our samples <lb/>showed the highest (100%) and lowest similari-<lb/>ty (97.03%) (Table 2). <lb/>Fig. 2: Agarose gel electrophoresis, showing PCR-RFLP results before (800-860bp) and after digestion with <lb/>the restriction enzyme TaqI on reference strains. From left to right: Lanes 1 and 2 L: molecular weight mark-<lb/>ers (100, 50 bp). Lanes 3 and 4: L. tropica; lane 5 and 6: L. major, Lanes 7 and 8: L. infantum. <lb/>After digestion by the restriction enzyme TaqI: L. major: 416, 296 and 141 bp.; L. tropica: 276, 193, 129, 115, 68 <lb/>and 28 bp.; L. infantum: 326, 277, 142, 70 and 33 bp <lb/></body>
48
+
49
+ <note place="headnote">Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.567-576 <lb/></note>
50
+
51
+ <page>573 <lb/></page>
52
+
53
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
54
+
55
+ . <lb/>
56
+
57
+ <body>Fig. 3: Phylogenetic tree of 8 Iranian isolates from visceral cases of leishmaniasis and 3 isolates selected from <lb/>GenBank, based on ITS-gene. The tree was constructed by using the Tamura3-parameter model in MEGA <lb/>software version 6. The evolutionary history was inferred using the Maximum Likelihood method, supported <lb/>by 1000 bootstrap replicates. The numbers above the branches indicate the percentage of bootstrap samplings <lb/>percentages. Samples isolated in the present study were compared to isolates selected from GenBank* (Red <lb/>diamonds). The bar at bottom of the tree shows a scale for magnitude of genetic distance between isolates <lb/>calculated by the software. <lb/>*(FJ001632/T.cruzi, C. fasciculata/Y00055 and Uzbakestan/FN398341/L. infantum) <lb/>Table 2: Levels of inter and intra-species mean similarity (one by one) among 8 Iranian of L.infantum Isolates <lb/>obtained from visceral cases of leishmaniasis and FJ001632/T. cruzi, C. fasciculata/Y00055 and Uzbeki-<lb/>stan/FN398341/L. infantum from GenBank, based on ITS-gene (2014-2016) <lb/>L. infantum Isolates <lb/>Levels of inter and intra-species (%) <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>6 <lb/>7 <lb/>8 <lb/>9 <lb/>10 <lb/>1. FJ001632/T.cruzi <lb/>100.00 <lb/>41.68 <lb/>42.31 <lb/>42.31 <lb/>42.31 <lb/>42.31 <lb/>42.31 <lb/>42.29 <lb/>42.29 <lb/>42.29 <lb/>2. C. fasciculata/Y00055 <lb/>41.68 <lb/>100.00 <lb/>57.01 <lb/>57.01 <lb/>57.01 <lb/>57.01 <lb/>57.01 <lb/>57.72 <lb/>57.72 <lb/>57.72 <lb/>3. Fars-Kazeroun/L.infantum <lb/>42.31 <lb/>57.01 <lb/>100.00 100.00 100.00 100.00 100.00 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>4. Fars-Borazjan/L.infantum <lb/>42.31 <lb/>57.01 <lb/>100.00 100.00 100.00 100.00 100.00 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>5. Albourz-Karaj/L.infantum <lb/>42.31 <lb/>57.01 <lb/>100.00 100.00 100.00 100.00 100.00 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>6. Qom-Qom/L.infantum <lb/>42.31 <lb/>57.01 <lb/>100.00 100.00 100.00 100.00 100.00 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>7. Uzbeki-<lb/>stan/FN398341/L.infantum <lb/>42.31 <lb/>57.01 <lb/>100.00 100.00 100.00 100.00 100.00 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>8. Ardabil-<lb/>MeshkinShahr/L.infantum <lb/>42.29 <lb/>57.72 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>100.00 <lb/>100.00 <lb/>100.00 <lb/>9.Tehran-Bumehen/L.infantum <lb/>42.29 <lb/>57.72 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>100.00 <lb/>100.00 <lb/>100.00 <lb/>10. East-Azerbaijan <lb/>Ahar/L.infantum <lb/>42.29 <lb/>57.72 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>97.28 <lb/>100.00 <lb/>100.00 <lb/>100.00 <lb/>Discussion <lb/>Both forms of VL and CL have been already <lb/>reported as important endemic diseases in <lb/>Iran. Annually about 100-300 cases of VL are <lb/>reported from different parts of Iran, but the <lb/>actual amount is several times higher (25). The <lb/>majority of the visceral cases reported from <lb/>the northwest (Ardabil and East-Azerbaijan) <lb/>and some of the southern provinces, including <lb/>Fars and Bushehr in Iran (5) and sporadic cas-<lb/>es have been reported from other regions (26, <lb/>27). <lb/></body>
58
+
59
+ <note place="headnote">Dalimi et al.: Molecular Identification and Intra-species Variations … <lb/></note>
60
+
61
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
62
+
63
+ <page>574 <lb/></page>
64
+
65
+ <body>In Iran, DAT is routinely performed for di-<lb/>agnosis and seroepidemiological studies of VL, <lb/>because it is simple and highly sensitive (92%-<lb/>100%) (7); thus in our study, we used DAT as <lb/>the standard serological test. Such passive in-<lb/>vestigations are usually encountered with the <lb/>dilemma of losing samples. Asymptomatic <lb/>dogs (7) and probably asymptomatic humans <lb/>(28) are sources of the parasite for phlebotom-<lb/>ine sandfly vectors and as it has been demon-<lb/>strated previously asymptomatic cases reserve <lb/>a high level of parasitemia in dogs and humans. <lb/>On the other hand, dogs support a larger res-<lb/>ervoir of the parasites than humans (14). In <lb/>this way, the cases which did not refer to <lb/>health centers were not enrolled, so we lost a <lb/>significant number of isolates. Furthermore, <lb/>patients who are in remission after treatment, <lb/>despite having antibody titers, but there are no <lb/>parasites in their bodies. <lb/>To detect Leishmania species in Iranian <lb/>leishmaniasis focuses, conventional and mo-<lb/>lecular methods have been employed (29, 30), <lb/>from different geographical areas of Iran, in-<lb/>cluding small sample sizes obtained from lim-<lb/>ited geographical regions (31, 32) as well as in <lb/>the large scale of sampling in a broad geo-<lb/>graphic area (16) by different molecular targets, <lb/>NAGT gene (16) and nuclear ITS-rDNA (32). <lb/>Semi-nested PCR of kDNA, a high sensitive <lb/>technique of PCR has been used formerly for <lb/>detection of Leishmania in the sandflies (20, <lb/>33) and reservoirs (11), is used in the present <lb/>study in order to confirm of species of Leish-<lb/>mania parasite which is in charge of VL in Iran. <lb/>We used ITS-RFLP and ITS-Sequencing <lb/>approaches for the investigation of genetic <lb/>diversity and population structure of three <lb/>strains of Leishmania spp. from different en-<lb/>demic areas for VL in Iran. RFLP of (TaqI <lb/>enzyme) ITS (18S rRNA-ITS1-5.8S rRNA-<lb/>ITS2)-rDNA gene was diagnostic for Leishma-<lb/>nia spp., and comparative for three Leishmania <lb/>species (L. tropica, L. major and L. infantum) <lb/>because of the size of the DNA fragment after <lb/>the enzyme digestion. The electrophoretic pat-<lb/>terns of 70 VL isolates compared with refer-<lb/>ence strains showed that all (70/70) isolates <lb/>belonged to L. infantum. In terms of molecular <lb/>epidemiology, our results are consonant with <lb/>previous epidemiologic studies performed in <lb/>Iran with introducing L. infantum as the main <lb/>causative agent of VL (34). <lb/>Based on our phylogenetic tree, low diver-<lb/>gence and high likelihood were seen among L. <lb/>infantum isolates of human and dogs from Iran <lb/>with a very slight divergence was seen between <lb/>isolates from northwest and south of Iran, <lb/>thus grouped in a unique clad. No correlation <lb/>was observed between intraspecies divergence <lb/>and geographic distribution of the isolates, as <lb/>all L. infantum isolates from different areas of <lb/>Iran causing VL with an isolate from Uzbeki-<lb/>stan. Despite the large variations in weather <lb/>conditions, geographical regions, vectors or <lb/>hosts and even in reservoirs, the similarity be-<lb/>tween isolates was 97.03%-100%, so there are <lb/>no noticeable differences in Iranian isolates. <lb/>This is in agreement with Waki (35), which <lb/>reported L. infantum as the less divergent <lb/>complexes and in consonant with Hajjaran <lb/>(16), employed ITS1 gene as DNA marker <lb/>and RAPD-PCR techniques, revealed no cor-<lb/>relation between isolates and geographical are-<lb/>as (36). <lb/>Conclusion <lb/>Low divergence was seen among L. infantum <lb/>isolates from Iran and no correlation was ob-<lb/>served between intraspecies divergence and <lb/>geographic distribution of the isolates. <lb/></body>
66
+
67
+ <div type="acknowledgement">Acknowledgements <lb/>This study was financially supported by <lb/>INSF (Project No: 92013166) and Tarbiat <lb/>Modares University. <lb/></div>
68
+
69
+ <div type="annex">Conflict of interest <lb/>The authors declare that there is no conflict <lb/>of interests. <lb/></div>
70
+
71
+ <note place="headnote">Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.567-576 <lb/></note>
72
+
73
+ <page>575 <lb/></page>
74
+
75
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
76
+
77
+ <listBibl>References <lb/>1. <lb/>World Health Organization. Leishmaniasis <lb/>and <lb/>Leishmania/HIV <lb/>coinfection, <lb/>WHO/CDC/CSR/ISR, Geneva. 2000. <lb/>2. <lb/>Desjeux P. Leishmaniasis: current situation <lb/>and new perspectives. Comp Immunol Mi-<lb/>crobiol Infect Dis. 2004; 27(5):305-18. <lb/>3. <lb/>Alvar J, Vélez ID, Bern C et al. Leishmania-<lb/>sis worldwide and global estimates of its in-<lb/>cidence. PLoS One. 2012; 7(5):e35671. <lb/>4. <lb/>Alborzi A, Rasouli M, Shamsizadeh A. <lb/>Leishmania tropica-isolated patient with vis-<lb/>ceral leishmaniasis in southern Iran. Am J <lb/>Trop Med Hyg. 2006; 74(2):306-7. <lb/>5. <lb/>Mohebali M. Epidemiological status of vis-<lb/>ceral leishmaniasis in Iran: experiences and <lb/>review of literature. J Clin Exp Pathol. 2012; <lb/>S3. <lb/>6. <lb/>Badirzadeh A, Mohebali M, Ghasemian M <lb/>et al. Cutaneous and post kala-azar dermal <lb/>leishmaniasis caused by Leishmania infantum <lb/>in endemic areas of visceral leishmaniasis, <lb/>Northwestern Iran 2002-2011: a case series. <lb/>Pathog Glob Health. 2013; 107(4):194-7. <lb/>7. <lb/>Mohebali M, Hajjaran H, Hamzavi Y et al. <lb/>Epidemiological aspects of canine visceral <lb/>leishmaniosis in the Islamic Republic of <lb/>Iran. Vet Parasitol. 2005; 129(3-4):243-51. <lb/>8. <lb/>Mohebali M, Edrissian Gh, Nadim A et al. <lb/>Application of direct agglutination test <lb/>(DAT) for the diagnosis and seroepidemio-<lb/>logical studies of visceral leishmaniasis in <lb/>Iran. Iran J Parasitol. 2006; 1(1):15-25. <lb/>9. <lb/>Leite RS, Ferreira Sde A, Ituassu LT et al. <lb/>PCR diagnosis of visceral leishmaniasis in <lb/>asymptomatic dogs using conjunctival swab <lb/>samples. Vet Parasitol. 2010; 170(3-4):201-6. <lb/>10. Van der Auwera G, Maes I, De Doncker S <lb/>et al. Heat-shock protein 70 gene sequenc-<lb/>ing for Leishmania species typing in Europe-<lb/>an tropical infectious disease clinics. Euro <lb/>Surveill. 2013; 18(30):20543. <lb/>11. Lachaud L, Marchergui-Hammami S, <lb/>Chabbert E et al. Comparison of six meth-<lb/>ods using peripheral blood for detection of <lb/>canine visceral leishmaniasis. J Clin Micro-<lb/>biol. 2002; 40(1):210-5. <lb/>12. Nasereddin A, Ereqat S, Azmi K et al. Sero-<lb/>logical survey with PCR validation for ca-<lb/>nine visceral leishmaniasis in northern Pal-<lb/>estine. J Parasitol. 2006; 92(1):178-83. <lb/>13. Mohammadiha A, Haghighi A, Mohebali M <lb/>et al. Canine visceral leishmaniasis: a com-<lb/>parative study of real-time PCR, conven-<lb/>tional PCR, and direct agglutination on sera <lb/>for the detection of Leishmania infantum in-<lb/>fection. Vet Parasitol. 2013; 192(1-3):83-90. <lb/>14. Mohammadiha A, Mohebali M, Haghighi A <lb/>et al. Comparison of real-time PCR and <lb/>conventional PCR with two DNA targets <lb/>for detection of Leishmania (Leishmania) in-<lb/>fantum infection in human and dog blood <lb/>samples. Exp Parasitol. 2013; 133(1):89-94. <lb/>15. Mahmoudzadeh-Niknam H, Ajdary S, Ria-<lb/>zi-Rad F et al. Molecular epidemiology of <lb/>cutaneous leishmaniasis and heterogeneity <lb/>of Leishmania major strains in Iran. Trop <lb/>Med Int Health. 2012; 17(11):1335-44. <lb/>16. Hajjaran H, Mohebali M, Teimouri A et al. <lb/>Identification and phylogenetic relationship <lb/>of Iranian strains of various Leishmania spe-<lb/>cies isolated from cutaneous and visceral <lb/>cases of leishmaniasis based on N-Acetyl <lb/>Glucosamine-1-phosphate Transferase gene. <lb/>Infect Genet Evol. 2014; 26: 203-12. <lb/>17. 17 Mirzaei A, Schweynoch C2, Rouhani S et <lb/>al. Diversity of Leishmania species and of <lb/>strains of Leishmania major isolated from de-<lb/>sert rodents in different foci of cutaneous <lb/>leishmaniasis in Iran. Trans R Soc Trop <lb/>Med Hyg. 2014; 108(8):502-12. <lb/>18. World Health Orzation. Basic laboratory <lb/>Methods in Medical Parasitology. First ed. <lb/>WHO, Geneva. 1991. <lb/>19. Edrissian GH, Hajjaran H, Mohebali M et <lb/>al. Application and evaluation of direct ag-<lb/>glutination test in sero-diagnosis of visceral <lb/>leishmaniasis in man and canine reservoirs <lb/>in Iran. Iran J Med Sci. 1996; 21, 119-124. <lb/>20. Aransay AM, Scoulica E, Tselentis Y. De-<lb/>tection and identification of Leishmania <lb/>DNA within naturally infected sandflies by <lb/>semi-nested PCR on minicircle kinetoplast <lb/>DNA. Appl Environ Microbiol. 2000; <lb/>66(5):1933-8. <lb/>21. Brewster S, Aslett M, Barker DC. Kineto-<lb/>plast DNA minicircle database. Parasitol <lb/>Today. 1998; 14(11):437-8. <lb/>22. Hall TA. BioEdit: a user-friendly biological <lb/>sequence alignment editor and analysis pro-<lb/></listBibl>
78
+
79
+ <note place="headnote">Dalimi et al.: Molecular Identification and Intra-species Variations … <lb/></note>
80
+
81
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
82
+
83
+ <page>576 <lb/></page>
84
+
85
+ <listBibl>gram for Windows 95/98/NT. Nucleic Ac-<lb/>ids Symposium Series, 1999; 41, 95-8. <lb/>23. Larkin MA, Blackshields G, Brown NP et al. <lb/>Clustal W and Clustal X version 2.0. Bioin-<lb/>formatics. 2007; 23(21):2947-8. <lb/>24. Tamura K, Stecher G, Peterson D et al. <lb/>MEGA6: molecular evolutionary genetics <lb/>analysis version 6.0. Mol Biol Evol. 2013; <lb/>30(12):2725-9. <lb/>25. Shirzadi MR. Epidemiological information <lb/>of cutaneous and visceral leishmaniasis in <lb/>I.R. Iran. In: Report of the constructive <lb/>meeting on cutaneous leishmaniasis. WHO. <lb/>Geneva. pp. 7-8. 2008. <lb/>26. Hosseininejad M, Mohebali M, Hosseini F <lb/>et al. Seroprevalence of canine visceral <lb/>leishmaniasis in asymptomatic dogs in Iran. <lb/>Iran J Vet Res. 2011; 13 (1): 38. <lb/>27. Malmasi A, Janitabar S, Mohebali M et al. <lb/>Seroepidemiology survey of canine visceral <lb/>leishmaniasis in Tehran and Alborz Prov-<lb/>inces of Iran. J Arthropod Borne Dis. 2014; <lb/>8(2):132-8. <lb/>28. Fakhar M, Motazedian MH, Hatam GR et <lb/>al. Asymptomatic human carriers of Leish-<lb/>mania infantum: possible reservoirs for Medi-<lb/>terranean visceral leishmaniasis in southern <lb/>Iran. Ann Trop Med Parasitol. 2008; <lb/>102(7):577-83. <lb/>29. Mirzaei A, Rouhani S, Taherkhani H et al. <lb/>Isolation and detection of Leishmania spe-<lb/>cies among naturally infected Rhombomys <lb/>opimus, a reservoir host of zoonotic cutane-<lb/>ous leishmaniasis in Turkemen Sahara, <lb/>North East of Iran. Exp Parasitol. 2011; <lb/>129(4):375-80. <lb/>30. Mirzaei A, Rouhani S, Kazerooni PA et al. <lb/>Molecular detection and identification of <lb/>Leishmania species in reservoir hosts of zo-<lb/>onotic cutaneous leishmaniasis in Fars <lb/>province, south of Iran. Iran J Parasitol. <lb/>2013; 8 (2):280-8. <lb/>31. Hajjaran H, Vasigheh F, Mohebali M et al. <lb/>Direct diagnosis of Leishmania species on <lb/>serosity materials punctured from cutane-<lb/>ous leishmaniasis patients using PCR-RFLP. <lb/>J Clin Lab Anal. 2011; 25(1):20-4. <lb/>32. Parvizi P, Ready PD. Nested PCRs of nu-<lb/>clear ITS-rDNA fragments detect three <lb/>Leishmania species of gerbils in sandflies <lb/>from Iranian foci of zoonotic cutaneous <lb/>leishmaniasis. Trop Med Int Health. 2008; <lb/>13(9):1159-71. <lb/>33. Rassi Y, Javadian E, Nadim A et al. <lb/>Phlebotomus (Larroussius) kandelakii the prin-<lb/>cipal and proven vector of visceral leish-<lb/>maniasis in north west of Iran. Pak J Biol <lb/>Sci. 2005; 8(12):1802-6. <lb/>34. Nadim A, Javadian E, Mohebali M, Zamen <lb/>Momeni A. Leishmania parasites and Leish-<lb/>maniases, in: Epidemiology of leishmaniasis <lb/>in Iran, Tehran University Press, Tehran, <lb/>2009. (Persian). <lb/>35. Waki K, Dutta S, Ray D et al. Transmem-<lb/>brane molecules for phylogenetic analyses <lb/>of pathogenic protists: Leishmania-specific <lb/>informative sites in hydrophilic loops of <lb/>transendoplasmic <lb/>reticulum <lb/>N-<lb/>acetylglucosamine-1-phosphate transferase. <lb/>Eukaryot Cell. 2007; 6(2):198-210. <lb/>36. Hajjaran H, Mohebali M, Mamishi S et al. <lb/>Molecular identification and polymorphism <lb/>determination of cutaneous and visceral <lb/>leishmaniasis agents isolated from human <lb/>and animal hosts in Iran. Biomed Res Int. <lb/>2013; 2013:789326. </listBibl>
86
+
87
+
88
+ </text>
89
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/IJPA-13-648.training.segmentation.tei.xml ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.648-654 <lb/> 648 <lb/> Available at: http://ijpa.tums.ac.ir <lb/> Short Communication <lb/>The Taxonomic Position and Phylogenetic Relationship between <lb/>Digramma interrupta and Ligula intestinalis Based on <lb/>Morphological and Molecular Diagnosis <lb/>Emad AHMADIARA 1 , *Seyed Hossein HOSSEINI 1 , Fatemeh JALOUSIAN 1 , Hossein Ali <lb/>EBRAHIMZADEH MOUSAVI 2 <lb/> 1. Dept. of Parasitology, School of Veterinary Medicine, University of Tehran, Tehran, Iran <lb/>2. Dept. of Aquatic Animal Health, School of Veterinary Medicine, University of Tehran, Tehran, Iran <lb/>Received <lb/>10 Oct 2017 <lb/>Accepted 04 Feb 2018 <lb/>Abstract <lb/>Background: The position of Digramma interrupta remains disputable as it was <lb/>raised by Cholodkovsky from Ligula alternans. This study aimed to survey the evolu-<lb/>tionary relationships and the taxonomic position of D. interrupta and L. intestinalis. It <lb/>also intended to support or reject the validity of D. interrupt as an independent ge-<lb/>nus and its correlation with L. intestinalis on the basis of their morphological charac-<lb/>teristics and a study on molecular data. <lb/>Methods: Overall, 1301 fish varieties, including 883 Alburnoides bipunctatus and 418 <lb/>Abramis brama, were collected from north and north-western parts of Iran. A. <lb/>bipunctatus samples were obtained from fresh water sources of the Maragheh dam <lb/>(northwest) and the Ramesar Lake (north). Moreover, samples of A. brama were <lb/>captured from the Aras Dam (northwest) and the Bandar-e-Anzali lagoon (north). <lb/>PCR was used to generate a fragment spanning two independent ITS-inclusive <lb/>parts: ITS1-5.8S and ITS2 with two pairs of primers. <lb/>Results: Nucleotide variation between L. intestinalis and D. interrupta samples <lb/>amounts to about 3% to 7%. Between samples of L. intestinalis and GenBank data, <lb/>and also between D. interrupta specimens and GenBank data, the diversity was seen <lb/>for about 1% to 3%. Moreover, about 1% to 4% nucleotide variation was seen only <lb/>in L. intestinalis samples caught from the same host, which could be supplementary <lb/>to the presence of a species and/or strains in this genus. <lb/>Conclusion: Maybe D. interrupta was just a rare diplogonadic form of the Ligula <lb/>species, not a different genus and not synonymous with the Ligula genus, but only <lb/>another species of the Ligula genus. <lb/>Keywords: <lb/> Ligula intestinalis, <lb/>Digramma interrupta, <lb/>Morphology, <lb/>Molecular diagnosis, <lb/>Iran <lb/> * Correspondence <lb/>Email: <lb/>[email protected] <lb/>Iranian Society of Parasitology <lb/>http://isp.tums.ac.ir <lb/>Iran J Parasitol <lb/>Open access Journal at <lb/>http://ijpa.tums.ac.ir <lb/>Tehran University of Medical <lb/>Sciences Publication <lb/>http://tums.ac.ir <lb/></front>
8
+
9
+ <note place="headnote">Ahmadiara et al.: The Taxonomic Position and Phylogenetic Relationship between … <lb/></note>
10
+
11
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
12
+
13
+ <page>649 <lb/></page>
14
+
15
+ <body>Introduction <lb/>embers of diphyllobothridae are <lb/>the most significant cestoda that <lb/>infect fish as the intermediated <lb/>host in the plerocercoid phase (1). This family <lb/>has significant genera like Ligula intestinalis, and <lb/>Digramma interrupta is distributed all around the <lb/>world. <lb/>The position of D. interrupta in <lb/>diphyllobothridae family remains disputable <lb/>since it was raised by Cholodkovsky from L. <lb/>alternans. Morphological differences, with one <lb/>set of reproductive organs in L. intestinalis and <lb/>two sets of this in D. interrupta, present the <lb/>most important differences between them. <lb/>However, the morphological features are not <lb/>completely reliable to distinguish genus from <lb/>each other (2). <lb/>One of the aims of this study was support or <lb/>reject the validity of D. interrupta and its corre-<lb/>lation with L. intestinalis on the basis of mor-<lb/>phological characteristics and a study of mo-<lb/>lecular data based on entire internal tran-<lb/>scribed spacer of the ribosomal DNA (ITS <lb/>rDNA) that contains ITS1-5.8S and ITS2. In <lb/>addition, this study intends to assess the ge-<lb/>netic diversity of L. intestinalis or D. interrupta <lb/>from different hosts and geographical regions. <lb/>Materials and Methods <lb/>Samples collection <lb/>Overall, 1301 fish varieties, including 883 <lb/>Alburnoides bipunctatus (A. bipunctatus) and 418 <lb/>Abramis brama (A. brama), were collected from <lb/>north and north-western parts of Iran. A. <lb/>bipunctatus samples were obtained from fresh <lb/>water sources of the Maragheh dam (north-<lb/>west) and the Ramesar Lake (north). Moreo-<lb/>ver, samples of A. brama were captured from <lb/>the Aras Dam (northwest) and the Bandar-e-<lb/>Anzali lagoon (north) (Fig. 1). The samples <lb/>were put into ice boxes and immediately trans-<lb/>ferred to the laboratory; they were isolated <lb/>from body cavities of infectious fish and pre-<lb/>served either by being placed in 96% ethanol <lb/>or being stored at -20 °C for morphological <lb/>and molecular analysis. <lb/>Morphological characterizations <lb/>The samples&apos; terminal segments were <lb/>stained with aceto-carmine and mounted with <lb/>the Canada balsam. The specimens were iden-<lb/>tified as L. intestinalis and D. interrupta by using <lb/>characters that are suitable for species identifi-<lb/>cation according to taxonomic keys (3). <lb/>Fig. 1: Localities in Iran where specimens of Ligula intestinalis and Digramma interrupta were collected <lb/>The morphological characterization of <lb/>plerocercoid was completed by observing <lb/>them under a light microscope equipped with <lb/>the camera lucida. Then, for having an accu-<lb/>M <lb/></body>
16
+
17
+ <note place="headnote">Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.648-654 <lb/></note>
18
+
19
+ <page>650 <lb/></page>
20
+
21
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
22
+
23
+ <body>rate survey, we draw the schematic character-<lb/>istics of the specimens, and transferred them <lb/>onto talk paper and scanned for more accurate <lb/>analysis (Figs. 2-5). <lb/>Fig. 2: Cross section of Digramma interrupta with two sets of reproductive organs <lb/>Fig. 3: Cross section of Ligula intestinalis with one set of reproductive organ <lb/>Fig. 4: Longitudinal sections of Digramma interrupta with two rows of reproductive organ <lb/>Fig. 5: Longitudinal sections of Ligula intestinalis with one row of reproductive organ <lb/>DNA extraction <lb/>Deoxyribonucleic acid (DNA) was extracted <lb/>by using a DNA isolation kit (Qiagen, Ger-<lb/>many) according to the manufacturer&apos;s in-<lb/>structions. <lb/>PCR amplification <lb/>PCR was used to generate a fragment span-<lb/>ning two ITS-inclusive independent parts, <lb/>namely ITS1-5.8S and ITS2 with two pairs of <lb/>specific primers. These primers were designed <lb/>by Vector NTI11. The first for the 891bp <lb/>fragment comprised ITS1-5.8S, and second <lb/>for the 421bp sequence-long inclusive ITS2 <lb/>locus (Table 1). The PCR product was puri-<lb/>fied by using the PCR purification kit (Qiagen, <lb/>Germany). The sequencing was performed <lb/>from both sites of each PCR products by <lb/>Kawsar Biotech Company in Iran on the basis <lb/>of the Sanger method (1977). The sequences <lb/>were analyzed by using the Chromas version <lb/>1.3 software and CLC Main Workbench 5, <lb/></body>
24
+
25
+ <note place="headnote">Ahmadiara et al.: The Taxonomic Position and Phylogenetic Relationship between … <lb/></note>
26
+
27
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
28
+
29
+ <page>651 <lb/></page>
30
+
31
+ <body>and they were compared with samples regis-<lb/>tered in GenBank by using the &apos;Basic Local <lb/>Alignment Search Tool&apos; (BLAST) program. <lb/>The phylogenetic tree was designed by MEGA <lb/>version 5.0 software. <lb/>Results <lb/>All isolates morphologically derived from A. <lb/>bipunctatus were distinguished from L. <lb/>intestinalis, and all specimens of A. brama were <lb/>determined to be D. interrupta. From 883 spec-<lb/>imens of A. bipunctatus and 418 collected sam-<lb/>ples of A. brama, 558 (63.19%) and 67 fishes <lb/>(16%) were infected, respectively (Table 2). <lb/>After blast, the data, according to the ITS <lb/>locus, 13 isolates were determined from D. <lb/>interrupta and 10 samples were determined L. <lb/>intestinalis. All the molecular results were coor-<lb/>dinated with the morphological outcome. <lb/>Table 1: Primer 1 (DigIr F 5&apos; &amp; DigIr R 5&apos;) for <lb/>891bp fragment consist of ITS1, 5.8S and ITS2; <lb/>Primer 2 (Ligir2F 5&apos;&amp; Ligir2R 5&apos;) for 421bp se-<lb/>quence long include ITS2 locus <lb/>Primer 1, product size: 890bp <lb/>DigIr F 5&apos; -CACGTTCCGTCTA TATGCGC-3&apos; <lb/>DigIr R 5&apos;-GGCAGCATCTCGCTTAAATG -3&apos; <lb/>Primer 2, product size:421bp <lb/>Ligir2F 5&apos; -TGGCGGGAAAACTCGGGCTT-3&apos; <lb/>Ligir2R 5&apos; -GCCGCCAACCACCAACAG -3&apos; <lb/>All 13 D. interrupta samples were registered in <lb/>the GenBank under accession numbers <lb/>KC900982-KC900994. Ten samples were re-<lb/>ferred to as L. intestinalis registered in the Gen-<lb/>Bank under accession numbers KC900972-<lb/>KC900981. Nucleotide variation between the <lb/>L. intestinalis and D. interrupta samples was <lb/>about 3% to 7%. <lb/>Table 2: Sampling localities and coordinates of Alburnoides bipunctatus and Abramis brama in this study <lb/>Sampled fish <lb/>Province <lb/>Locality <lb/>Geographical <lb/>position <lb/>Rate of <lb/>infection <lb/>Plerocercoid <lb/>detected <lb/>A. bipunctatus <lb/>East Azerbaijan <lb/>Maragheh Dam <lb/>Northwest <lb/>63.19% <lb/>L. intestinalis <lb/>A. brama <lb/>West Azerbaijan <lb/>Aras Dam <lb/>Northwest <lb/>16% <lb/>D. interrupta <lb/>A. bipunctatus <lb/>Guilan <lb/>Bandare-E-Anzali lagoon <lb/>North <lb/>-<lb/>-<lb/>A. brama <lb/>Mazandaran <lb/>Ramesar lake <lb/>North <lb/>-<lb/>-<lb/>Between L. intestinalis samples in this survey <lb/>and the GenBank samples, and also between <lb/>D. interrupta samples and the GenBank sam-<lb/>ples, nucleotide variation was about 1% to 3%. <lb/>Finally, L. intestinalis samples were caught in <lb/>Iran from the same host and nucleotide varia-<lb/>tions of about 1% to 4% were seen among <lb/>them wonderfully. <lb/>The results of the genealogy tree based on <lb/>the ITS2 locus display that two clades were <lb/>obvious. One containing L. intestinalis and D. <lb/>interrupta samples from present study with five <lb/>isolates of this other L. intestinalis and D. <lb/>interrupta that registered in GenBank, while <lb/>another clade contains genus Schistocephalus <lb/>solidus that is available in GenBank. All sam-<lb/>ples caught in this study, including 10 isolates <lb/>of L. intestinalis and 11 isolates of D. interrupta, <lb/>along with five isolates from GenBank were <lb/>located in the same clade and so we could call <lb/>them monophyletic. S. solidus is located in a <lb/>different clade (Table 3, Fig. 6). <lb/>Discussion <lb/>Studies on the distribution of D. interrupta <lb/>and L. intestinalis in Iran showed that most <lb/>such reports have been from the northwest <lb/>and western parts, while only a few reports <lb/>came from the northern area (4, 5). In this <lb/>study, none of the collected samples from the <lb/>northern part was infected, but, in contrast, <lb/>the infection rate was remarkably up in the <lb/>north-western part of country. Host specificity <lb/>and isolated lineages by geography were the <lb/>results of adjustment to local host fauna (6). <lb/>Moreover, L. intestinalis is remarkably host-<lb/></body>
32
+
33
+ <note place="headnote">Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.648-654 <lb/></note>
34
+
35
+ <page>652 <lb/></page>
36
+
37
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
38
+
39
+ <body>specific remarkably in Kenya (7). According to <lb/>the result of this study, infection with plero-<lb/>cercoid was highly correlated with habitat in <lb/>northwest Iran. <lb/>Table 3: Phylogenetic data collection to assess relationships between Ligula intestinalis and Digramma interrupta <lb/>No. Sample name <lb/>Locality <lb/>Host <lb/>Accession no. <lb/>1 <lb/>L. intestinalis 1 <lb/>Iran <lb/>A. bipunctatus <lb/>KC900972.1 <lb/>2 <lb/>L. intestinalis 2 <lb/>Iran <lb/>A. bipunctatus <lb/>KC900973.1 <lb/>3 <lb/>L. intestinalis 3 <lb/>Iran <lb/>A. bipunctatus <lb/>KC900974.1 <lb/>4 <lb/>L. intestinalis 4 <lb/>Iran <lb/>A. bipunctatus <lb/>KC900975.1 <lb/>5 <lb/>L. intestinalis 5 <lb/>Iran <lb/>A. bipunctatus <lb/>KC900976.1 <lb/>6 <lb/>L. intestinalis 6 <lb/>Iran <lb/>A. bipunctatus <lb/>KC900977.1 <lb/>7 <lb/>L. intestinalis 7 <lb/>Iran <lb/>A. bipunctatus <lb/>KC900978.1 <lb/>8 <lb/>L. intestinalis 8 <lb/>Iran <lb/>A. bipunctatus <lb/>KC900979.1 <lb/>9 <lb/>L. intestinalis 9 <lb/>Iran <lb/>A. bipunctatus <lb/>KC900980.1 <lb/>10 <lb/>L. intestinalis 10 <lb/>Iran <lb/>A. bipunctatus <lb/>KC900981.1 <lb/>11 <lb/>D. interrupta 1 <lb/>Iran <lb/>A. brama <lb/>KC900982.1 <lb/>12 <lb/>D. interrupta 2 <lb/>Iran <lb/>A. brama <lb/>KC900983.1 <lb/>13 <lb/>D. interrupta 3 <lb/>Iran <lb/>A. brama <lb/>KC900984.1 <lb/>14 <lb/>D. interrupta5 <lb/>Iran <lb/>A. brama <lb/>KC900986.1 <lb/>15 <lb/>D. interrupta 6 <lb/>Iran <lb/>A. brama <lb/>KC900987.1 <lb/>16 <lb/>D. interrupta 7 <lb/>Iran <lb/>A. brama <lb/>KC900988.1 <lb/>17 <lb/>D. interrupta 8 <lb/>Iran <lb/>A. brama <lb/>KC900989.1 <lb/>18 <lb/>D. interrupta 10 <lb/>Iran <lb/>A. brama <lb/>KC900991.1 <lb/>19 <lb/>D. interrupta 11 <lb/>Iran <lb/>A. brama <lb/>KC900992.1 <lb/>20 <lb/>D. interrupta 12 <lb/>Iran <lb/>A. brama <lb/>KC900993.1 <lb/>21 <lb/>D. interrupta 13 <lb/>Iran <lb/>A. brama <lb/>KC900994.1 <lb/>22 <lb/>S. solidus <lb/>Norway <lb/>Gasterosteus aculeatus <lb/>AY549508.1 <lb/>23 <lb/>L. intestinalis <lb/>Turkey <lb/>Silurus glanis <lb/>AY549516.1 <lb/>24 <lb/>L. intestinalis <lb/>Turkey <lb/>Chalcaburnus sp <lb/>AY549517.1 <lb/>25 <lb/>L. colymbi <lb/>Poland <lb/>Gavia stellata <lb/>EU241090.1 <lb/>26 <lb/>D. interrupta <lb/>Russia <lb/>Hemiculter lucidus <lb/>EU241114.1 <lb/>27 <lb/>D. interrupta <lb/>Russia <lb/>Hemiculter lucidus <lb/>EU241117.1 <lb/>Fig. 6: Maximum likelihood phylogenetic tree of Ligula intestinalis and Digramma interrupta based on ITS2 sequences. <lb/>Phylogenetic trees were obtained by comparing the ITS2 query sequences of L. intestinalis and D. interrupta with <lb/>those of other Cestoda species available in GenBank based on maximum parsimony. Similar topology was observed <lb/>among the trees obtained by distance-based (NJ) tree building methods in phylogenetic analysis using MEGA7 <lb/>software. The species included in the maximum parsimony analysis mainly clustered into two major clades and <lb/>Schistocephalus solidus (Pseudophyllidea: Diphyllobothriidae) rooted as an out-group and indicates its evolutionary <lb/>relationships <lb/></body>
40
+
41
+ <note place="headnote">Ahmadiara et al.: The Taxonomic Position and Phylogenetic Relationship between … <lb/></note>
42
+
43
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
44
+
45
+ <page>653 <lb/></page>
46
+
47
+ <body>Climatic conditions in the north-western <lb/>part of the country are more conducive to <lb/>these parasites, while another probable reason <lb/>could be the neighborhood of Turkey. Since <lb/>Turkey is a good source of pollution by Ligu-<lb/>lidae plerocercoid, maybe water imported <lb/>from Turkey can be attributed to infection of <lb/>water by plerocercoid in northwest Iran (8). <lb/>Another reason could be migratory birds that <lb/>could transmit the infection in nearby loca-<lb/>tions in two adjacent countries. <lb/>Host specificity probably exists in Iran be-<lb/>cause D. interrupta was detected only from A. <lb/>brama, in Russia (9). In addition, similar to a <lb/>survey (5), results from northwest Iran <lb/>showed detection of plerocercoid of L. <lb/>intestinalis from A. bipunctatus. However, the <lb/>morphological difference could be used to <lb/>detect the genus and species from each other, <lb/>but it is not enough and an accurate meter is <lb/>needed for distinguishing the species and line-<lb/>age from each other. Therefore, over the last <lb/>two decades, molecular studies have helped <lb/>scientists to improve and revise their infor-<lb/>mation about the taxonomical status of para-<lb/>sites and their phylogeny. <lb/>Despite various studies on the epidemiologi-<lb/>cal aspect of this Cestoda, any molecular char-<lb/>acterization is not found in Iran and also in <lb/>the entire Middle East zone. Thus, there is <lb/>lack of sufficient information about the mo-<lb/>lecular characteristics of this Pseudophyllidean <lb/>Cestoda in this zone. It confirms the necessity <lb/>of a study on molecular and morphological <lb/>features of this genus, and another genus of <lb/>this family, as well as their true taxonomical <lb/>station in this region. <lb/>Liao and Liang (10) watched in Digramma the <lb/>transitional reproductive organ structure. A <lb/>single set of reproductive organs was located <lb/>at the anterior end of the larva in the carassian <lb/>intermediate hosts and also two rows of this <lb/>were located at the posterior end. Sexual di-<lb/>morphism was reported in D. interrupta in A. <lb/>brama the reproductive organs found in <lb/>one row, whereas two rows occur in those <lb/>from carp (9). Maybe D. interrupta was just a <lb/>rare diplogonadic form of the Ligula species <lb/>(2). In L. intestinalis dimorphism with some of <lb/>them debating on the potential existence of <lb/>this species in the Ligula genus. For example, <lb/>based on ITS region and 28S rRNA, verified <lb/>that plerocercoid of L. intestinalis specimens <lb/>from Rutilus and Gobio may represent different <lb/>strains or species (11). Despite recognition of <lb/>species based on morphological features ap-<lb/>propriate for the discrimination between two <lb/>genera (3), the morphological specification of <lb/>them is not reliable and raises much confusion <lb/>concerning D. interrupta validity because of the <lb/>transitional shape of reproductive organs in <lb/>the proglottids of these two genera. <lb/>In the present study, D. interrupta exhibited <lb/>identical sequences with L. intestinalis with <lb/>about 1% heterology based on the ITS1-5.8S <lb/>regions analysis, whereas a high degree of ge-<lb/>netic diversity of about 3% to 7% has seen <lb/>based on the ITS2 locus. The present study <lb/>also suggests that ITS2, unlike ITS1-5.8S se-<lb/>quences, can act as a useful genetic marker. <lb/>There is not any yardstick for identifying <lb/>species or genus borders by using the result of <lb/>the DNA-sequence distinction, with a value of <lb/>divergence based on ITS2 sequences, D. <lb/>interrupta is almost different from L. intestinalis. <lb/>Therefore, with nucleotide variation levels be-<lb/>tween L. intestinalis and D. interrupta specimens, <lb/>besides morphological differences, it cannot <lb/>be verified whether D. interrupta is a synonym <lb/>of the genus L. intestinalis, against the theory of <lb/>some researchers (10, 11). <lb/>On the other hand, they are different genera <lb/>because of low levels of nucleotide variation <lb/>and unstable morphological difference be-<lb/>tween them. Maybe polymorphism seen in L. <lb/>intestinalis and D. interrupta is present only in <lb/>some global locations (10, 12). Besides low <lb/>levels of nucleotide variation (else two sam-<lb/>ples that have 6% and 7% diversity, the other <lb/>differences would amount to (3% to 4%), we <lb/>reach and verify the theory that maybe D. <lb/>interrupta was just a rare diplogonadic form of <lb/>the Ligula species (2) not a different genus and <lb/>not synonymous with the Ligula genus, but <lb/></body>
48
+
49
+ <note place="headnote">Iran J Parasitol: Vol. 13, No. 4, Oct-Dec 2018, pp.648-654 <lb/></note>
50
+
51
+ <page>654 <lb/></page>
52
+
53
+ <note place="footnote">Available at: http://ijpa.tums.ac.ir <lb/></note>
54
+
55
+ <body>only another species of the Ligula genus. Phy-<lb/>logenetic tree results have displayed differ-<lb/>ences between clade A and clade B, verifying <lb/>that the effect of geography in speciation and <lb/>its effect on molecular structures. L. intestinalis <lb/>and D. interrupta samples were different from <lb/>samples registered in GenBank. Eleven sam-<lb/>ples of D. interrupta and eight samples of L. <lb/>intestinalis were located in a common cluster so <lb/>that it may be support the analogy of this ge-<lb/>nus. Moreover, the discrepancy amount of 1% <lb/>to 4%, shown in the sequence of L. intestinalis <lb/>samples, were compared with each other in <lb/>the same host that could supplement the pres-<lb/>ence of a species and/or strains in this genus <lb/>much like results derived by other researchers <lb/>(12,13). <lb/>Conclusion <lb/>ITS2 may be one of the most useful markers <lb/>used to distinguish L. intestinalis from D. <lb/>interrupta. It can also be used to detect the var-<lb/>iations within L. intestinalis. Further, D. <lb/>interrupta may be another species of the Ligula <lb/>genus. On the other hand, they could not be <lb/>different genera because of low levels of nu-<lb/>cleotide variation arising from the possibility <lb/>of polymorphism and therefore D. interrupta is <lb/>another species of the Ligula genus. Its possi-<lb/>ble presence as a species and/or strains in the <lb/>Ligula genus is based on host specificity. <lb/></body>
56
+
57
+ <div type="acknowledgement">Acknowledgements <lb/>The authors thank Dr. Iraj Mobedi for help-<lb/>ful advice and his kind assistance in designing <lb/>the schematic. Moreover, we are grateful of <lb/>University of Tehran for supporting this re-<lb/>search. <lb/></div>
58
+
59
+ <div type="annex">Conflict of interest <lb/>The authors declare that there is no conflict <lb/>of interests. <lb/></div>
60
+
61
+ <listBibl>References <lb/>1. <lb/>Williams H, Jones A. Parasitic worms of fish <lb/>Taylor &amp; Francis, 1994. London. <lb/>2. <lb/>Luo HY, Nie P, Yao WJ, Wang GT, Gao Q. Is <lb/>the genus Digramma synonymous to the genus <lb/>Ligula (Cestoda: Pseudophyllidea)? Parasitol <lb/>Res. 2003; 89(5):419-21. <lb/>3. <lb/>Chubb JC, Pool DW, Veltkamp CJ. A key to <lb/>the species of cestodes (tapeworms) parasitic in <lb/>British and Irish freshwater fishes. J Fish Biol. <lb/>1987; 31:517-543. <lb/>4. <lb/>Jalali B, Barzegar M. Fish Parasites in Zarivar <lb/>Lake. J Agric Sci Technol. 2006; 8: 47-58. <lb/>5. <lb/>Hajirostamloo M. The occurrence and parasite <lb/>host of Ligula intestinalis in Sattarkhan Lake <lb/>(East Azerbaijan-Iran). J Anim Vet Adv. 2008; <lb/>7: 221-225. <lb/>6. <lb/>Bouzid W, Stefka J, Hypsa V, Lek S, Scholz T, <lb/>Legal L, Ben Hassine OK, Loot G. Geography <lb/>and host specificity: Two forces behind the ge-<lb/>netic structure of the freshwater fish parasite <lb/>Ligula intestinalis (Cestoda: Diphyllobothriidae). <lb/>Int J Parasitol. 2008; 38(12):1465-79. <lb/>7. <lb/>Britton JR, Jackson, MC, Harper, DM. Ligula <lb/>intestinalis (Cestoda: Diphyllobothriidae) in <lb/>Kenya: a field investigation into host specificity <lb/>and behavioral alterations. Parasitol. 2009; 136: <lb/>1367-73. <lb/>8. <lb/>Yavuzcan H, Korkmaz AS, Zencir O. The <lb/>infection of tench (Tincatinca) with Ligula <lb/>intestinalis plerocercoid in Lake Beysehir Turkey. <lb/>Bull Eur Assn Fish. 2003; 23: 223-227. <lb/>9. <lb/>Dubinina MN. Tapeworms (Cestoda, Liguli-<lb/>dae) of the fauna of the USSR Amerind, New <lb/>Delhi. 1980; 237-260. <lb/>10. Li J, Liao X. The taxonomic status of Di-<lb/>gramma (Pseudophyllidae: Ligulidae) inferred <lb/>from DNA sequences. J Parasitol. 2003; 89: <lb/>792-799. <lb/>11. Olson PD, Littlewood DT, Griffiths D, Ken-<lb/>nedy CR, Arme C. Evidence of the existence <lb/>of separate strains or species of Ligula in <lb/>Lough Neagh, Northern Ireland. J Helminthol. <lb/>2002; 76(2):171-4. <lb/>12. Arme C, Owen RW. Occurrence and patholo-<lb/>gy of Ligula intestinalis infections in British <lb/>fishes. J Parasitol. 1968; 54(2):272-80. <lb/>13. Arme C. Ligulosis in two cyprinid hosts: Ruti-<lb/>lus and Gobio. Helmintholgia. 1997; 34: 191-<lb/>196. </listBibl>
62
+
63
+
64
+ </text>
65
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/J09-3003.training.segmentation.tei.xml ADDED
The diff for this file is too large to render. See raw diff
 
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/J11-1004.training.segmentation.tei.xml ADDED
The diff for this file is too large to render. See raw diff
 
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/J11-3009.training.segmentation.tei.xml ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei>
3
+ <teiHeader>
4
+ <fileDesc xml:id="__J11-3009"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front> Book Reviews <lb/> Computational Modeling of Human Language Acquisition <lb/> Afra Alishahi <lb/> (University of the Saarland) <lb/>Morgan &amp; Claypool (Synthesis Lectures on Human Language Technologies, edited by <lb/>Graeme Hirst, volume 11), 2010, xiv+93 pp; paperbound, ISBN 978-1-60845-339-9, <lb/>$40.00; ebook, ISBN 978-1-60845-340-5, $30.00 or by subscription <lb/> Reviewed by <lb/>Sharon Goldwater <lb/>University of Edinburgh <lb/></front>
8
+
9
+ <body> For much of the last 25 years or more, researchers in natural language processing <lb/>(NLP) and those interested in human language acquisition have had little to say to <lb/>one another. NLP researchers were increasingly focusing on data-intensive supervised <lb/>learning methods, mostly using structured representations, while models of language <lb/>acquisition were typically based either on symbolic nativist accounts (Dresher and <lb/>Kaye 1990; Gibson and Wexler 1994) or on the unstructured distributed representations <lb/>of the connectionist approach (Rumelhart and McClelland 1986; Elman et al. 1996). <lb/>Moreover, language acquisition researchers have understandably been more interested <lb/>in unsupervised than supervised learning, and (perhaps due to the much more difficult <lb/>nature of this problem) have often focused on learning from toy data sets rather than <lb/>large naturalistic corpora. <lb/>Recent developments in both fields have led to a narrowing of the research gap, <lb/>however. NLP researchers have become increasingly interested in unsupervised and <lb/>minimally supervised methods, and the rise of probabilistic models of cognition (Chater <lb/>and Oaksford 1998; Griffiths, Kemp, and Tenenbaum 2008) means there is now a grow-<lb/>ing number of cognitive scientists who are well-versed in many of the same statistical <lb/>methods that are used in NLP. Thus, a short introductory text on computational mod-<lb/>eling of human language acquisition seems particularly apt at this time. Alishahi&apos;s slim <lb/>volume is not intended to be comprehensive, but rather to provide a brief overview <lb/>of the goals and methods of the field for researchers in related areas—either language <lb/>acquisition researchers with little computational experience or NLP researchers with-<lb/>out much knowledge of cognitive science. It aims for intuitive explanations rather <lb/>than highly technical ones, and includes a number of figures and diagrams, but no <lb/>equations. <lb/></body>
10
+
11
+ <front>© 2011 Association for Computational Linguistics <lb/></front>
12
+
13
+ <body>The book can be divided logically into two parts followed by a brief concluding <lb/>chapter. The first part (Chapters 1 and 2) provides an overview of the major research <lb/>questions and methodologies in the field. Chapter 1 begins by introducing some of the <lb/>main theoretical debates in the field of language acquisition—questions of modularity <lb/>and learnability. NLP researchers with some background in linguistics will probably <lb/>already be familiar with these debates at the level of detail presented here; these sections <lb/>will be more useful to those with a straight computer science background. Also included <lb/>in Chapter 1 is a section motivating the use of computational models as an alternative <lb/>to behavioral studies for studying language acquisition. This section will be useful to <lb/>anyone who is new to the idea of computational modeling of cognition, as will Chapter 2 <lb/>of the book. Chapter 2 discusses Marr&apos;s (1982) influential analysis of the different kinds <lb/> </body>
14
+
15
+ <front>Computational Linguistics <lb/>Volume 37, Number 3 <lb/></front>
16
+
17
+ <body>of explanations that models can provide, as well as the criteria for cognitive plausibility <lb/>against which models are judged, and the main frameworks for model development <lb/>(symbolic, connectionist, probabilistic). Finally, it describes the various ways models <lb/>can be evaluated, along with a list of corpus resources. <lb/>In the second part of the book (Chapters 3–5), Alishahi focuses on three areas of <lb/>language acquisition in particular: learning the meanings of words (Chapter 3), learning <lb/>morphology and syntax (Chapter 4), and learning relationships between syntax and <lb/>semantics, such as verb–argument structure and semantic roles (Chapter 5). Each chap-<lb/>ter is divided into sections focusing on more specific topics—for example, Chapter 4 <lb/>includes sections on morphology, syntactic categories, and syntactic structure. Each <lb/>section begins by reviewing the most salient empirical facts about children&apos;s acquisition <lb/>in that domain, along with relevant linguistic concepts and theories that have influenced <lb/>the modeling community (e.g., Mutual Exclusivity in word learning; Construction <lb/>Grammar and the Principles and Parameters theory in syntactic acquisition; theories <lb/>of selectional restrictions in the acquisition of verb–argument structure). The empirical <lb/>and theoretical background is followed by an overview of many of the models that have <lb/>been proposed in the area, and finally one or more &quot; case studies &quot; (more on these in a <lb/>moment). <lb/>In general the structure and content of this book is appropriate for researchers from <lb/>neighboring fields (including NLP) who want to get a quick taste of what computational <lb/>modeling of human language acquisition is all about. They can read the first couple of <lb/>chapters to get a general idea of the questions and methodologies, and then pick and <lb/>choose any topics from the remaining chapters that might be of interest. There are very <lb/>few dependencies between sections in the second part of the book, because most of the <lb/>relevant background for each section is provided in that section itself. Someone who is <lb/>interested in pursuing the field further (either a beginning graduate student or a more <lb/>advanced researcher moving into the field) will also find many useful references, at least <lb/>in the three areas that Alishahi focuses on. Many other active areas of modeling are not <lb/>covered at all (for example phonetic and phonological acquisition), although this choice <lb/>is understandable given the length of the book. <lb/>The main weakness of the book is in the execution of the case studies. Each case <lb/>study (with some exceptions, as noted subsequently) details a single model, with a half-<lb/>page to two-page description of the model&apos;s primary methods and assumptions, as well <lb/>as the input and results. It is an excellent idea to provide concrete examples showing <lb/>how models can be used to address important questions in language acquisition. But <lb/>in practice, most of the case studies fall short of this goal, as they are too light on <lb/>motivation and analysis. Alishahi is not always clear about why certain models, rather <lb/>than others, were chosen for case studies (are they ground-breaking in some way, or <lb/>merely a simple example of a particular theoretical idea put into practice?), nor how <lb/>each model&apos;s assumptions and results relate to the broader goals of modelling set out <lb/>in the first two chapters. In addition, three out of the four &quot; case studies &quot; in Chapter 4 <lb/>are really just additional review sections, covering models of the English past tense, <lb/>learning algorithms based on Principles and Parameters, and distributional models of <lb/>syntactic structure (all worthy topics, but not case studies). This leaves only one true <lb/>case study in Chapter 4, on the MOSAIC model of grammar induction (Jones, Gobet, <lb/>and Pine 2000). Finally, although there are case studies of symbolic, connectionist, and <lb/>probabilistic models, no Bayesian models are given a detailed look (though they are <lb/>included in the review sections). Bayesian models are, of course, a subset of probabilistic <lb/>models, but take a very different philosophical approach to most other models, includ-<lb/>ing the type of incremental probabilistic model that Alishahi discusses in more detail. <lb/>
18
+
19
+ <page>628 <lb/></page>
20
+
21
+ <note place="headnote">Book Reviews <lb/></note>
22
+
23
+ Bayesian modeling is now an important force in cognitive science generally and has <lb/>begun to make an impact in language acquisition specifically (Xu and Tenenbaum 2007; <lb/>Foraker et al. 2009; Goldwater, Griffiths, and Johnson 2009); as such it is worth taking a <lb/>bit more space to explain the ideas behind at least one of these models. <lb/>Despite these weaknesses in the case studies, there is enough useful material in <lb/>this book to make reading it worthwhile to any researcher who wants to get a quick <lb/>overview of the main goals and approaches in the field, along with some of the many <lb/>models that have been developed over the years. It will also be helpful to those who <lb/>are looking for a starting point for a more in-depth study of models in one of the three <lb/>areas of acquisition that Alishahi focuses on. Overall, it is a very accessible, if necessarily <lb/>selective, brief introduction to the field. <lb/></body>
24
+
25
+ <listBibl> References <lb/> Chater, Nicholas and Mike Oaksford, editors. <lb/>1998. Rational Models of Cognition. Oxford <lb/>University Press, Oxford. <lb/>Dresher, B. Elan and Jonathan Kaye. 1990. <lb/>A computational learning model for <lb/>metrical phonology. Cognition, <lb/> 34(2):137–195. <lb/>Elman, Jeffrey, Elizabeth Bates, Mark H. <lb/>Johnson, Anette Karmiloff-Smith, <lb/>Domenico Parisi, and Kim Plunkett. 1996. <lb/> Rethinking Innateness: A Connectionist <lb/>Perspective on Development. MIT <lb/>Press/Bradford Books, Cambridge, MA. <lb/>Foraker, Stephanie, Terry Regier, Naveen <lb/>Khetarpal, Amy Perfors, and Joshua B. <lb/>Tenenbaum. 2009. Indirect evidence and <lb/>the poverty of the stimulus: The case of <lb/>anaphoric one. Cognitive Science, <lb/> 33(2):287–300. <lb/>Gibson, Edward and Kenneth Wexler. <lb/>1994. Triggers. Linguistic Inquiry, <lb/> 25(3):407–454. <lb/>Goldwater, Sharon, Thomas L. Griffiths, <lb/>and Mark Johnson. 2009. A Bayesian <lb/>framework for word segmentation: <lb/>Exploring the effects of context. Cognition, <lb/> 112(1):21–54. <lb/>Griffiths, Thomas L., Charles Kemp, and <lb/>Joshua B. Tenenbaum. 2008. Bayesian <lb/>models of cognition. In Ron Sun, editor, <lb/> Cambridge Handbook of Computational <lb/>Cognitive Modeling. Cambridge University <lb/>Press, Cambridge, chapter 3. <lb/>Jones, Gary, Fernand Gobet, and Julian M. <lb/>Pine. 2000. A process model of children&apos;s <lb/>early verb use. In Proceedings of the 22nd <lb/>Meeting of the Cognitive Science Society, <lb/> pages 723–728, Philadelphia, PA. <lb/>Marr, David. 1982. Vision: A Computational <lb/>Approach. Freeman &amp; Co., San <lb/>Francisco, CA. <lb/>Rumelhart, David and James McClelland. <lb/>1986. On learning the past tenses of <lb/>English verbs. In David Rumelhart and <lb/>James McClelland, editors, Parallel <lb/>Distributed Processing: Explorations in the <lb/>Microstructure of Cognition. MIT Press, <lb/>Cambridge, MA, chapter 18. <lb/>Xu, Fei and Joshua B. Tenenbaum. 2007. <lb/>Word learning as Bayesian inference. <lb/> Psychological Review, 114(2):245–272. <lb/></listBibl>
26
+
27
+ <body>This book review was edited by Pierre Isabelle. <lb/>Sharon Goldwater is a Lecturer in the Institute for Language, Cognition and Computation, School <lb/>of Informatics, University of Edinburgh. Her research interests include unsupervised learning of <lb/>linguistic structure (including word segmentation, phonology, morphology, and syntax) and the <lb/>application of machine learning techniques, especially Bayesian methods, to the computational <lb/>study of human language acquisition. Goldwater&apos;s address is Informatics Forum, 10 Crichton <lb/>Street, Edinburgh EH8 9AB, United Kingdom; e-mail: [email protected]. <lb/></body>
28
+
29
+ <page> 629 </page>
30
+
31
+
32
+ </text>
33
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/J90-2003.training.segmentation.tei.xml ADDED
The diff for this file is too large to render. See raw diff
 
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JAP0702234-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <titlePage>High-pressure oxygen-induced bulk superconductivity in 1222 structure Tl-Pb-Sr-<lb/>Eu(Ce)-Cu-O <lb/>Z. Iqbal, A. P. B. Sinha, D. E. Morris, J. C. Barry, G. J. Auchterlonie, and B. L. Ramakrishna <lb/>Citation: Journal of Applied Physics 70, 2234 (1991); doi: 10.1063/1.349414 <lb/>View online: http://dx.doi.org/10.1063/1.349414 <lb/>View Table of Contents: http://scitation.aip.org/content/aip/journal/jap/70/4?ver=pdfcov <lb/>Published by the AIP Publishing <lb/>Articles you may be interested in <lb/>Oxygen-induced internal pressure in superconductors <lb/>Low Temp. Phys. 41, 879 (2015); 10.1063/1.4936225 <lb/>High-pressure structural and elastic properties of Tl2O3 <lb/>J. Appl. Phys. 116, 133521 (2014); 10.1063/1.4897241 <lb/>Superconductivity and microstructure of n-type Ln1.85Ce0.15CuO4−y (Ln=Pr, Sm, Eu) produced under <lb/>high-pressure sintering <lb/>Appl. Phys. Lett. 62, 3037 (1993); 10.1063/1.109131 <lb/>Bulk superconductivity with T c (zero) up to 95 K in a Tl0.5Pb0.5Ca0.9Ce0.1Sr2Cu2 oxide with an <lb/>Y1Ba2Cu3O y -like structure <lb/>Appl. Phys. Lett. 54, 2464 (1989); 10.1063/1.101536 <lb/>Crystal structure of the high T c (107 K) superconducting phase in Pb-Bi-Sr-Ca-Cu-O <lb/>Appl. Phys. Lett. 54, 2157 (1989); 10.1063/1.101516 <lb/>Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 130.102.82.20 On: Wed, 05 <lb/>Oct 2016 05:48:35 <lb/>I<lb/></titlePage>
9
+
10
+ <front>High-pressure oxygen-induced bulk superconductivity <lb/>in 1222 structure <lb/>I <lb/>TI-Pb-Sr-Eu(Ce)-Cu-0 <lb/>Z. lqbal <lb/>Allied-Signal Inc., Research and Technology, Morrfsfown, New Jersey 07962 <lb/>A. P. B. Sinha and D. E. Morris <lb/>Morris Research Inc., Berkeley, Califbrnid &apos;94704 <lb/>J. C. Barry and G. J. Auchterlonie <lb/>Electron Microscope Center, University of Queens&amp;d, Brisbane, Australia <lb/>B. L. Ramakrishna <lb/>Department of Chemistry, Arizona State University, Tempe. Arizona 85287 <lb/>(Received 4 March 1991; accepted for publication 4 May 1991) <lb/>A primarily single phase compound of composition (T1c.sPbc.s ) Srz ( Euz _ &amp;!e,) Cua09 (where <lb/>x ranges from (0.1 to 0.17) has been synthesized in the so-called 1222 structure. Bulk <lb/>superconductivity near 40 K is induced via oxygen doping in 100 bar of oxygen at 550 &quot;C!. The <lb/>material has been structurally characterized by diffraction and high-resolution lattice <lb/>imaging techniques. The lattice images reveal the presence of some intergrowths of a 1212 <lb/>structure nonsuperconducting (Tle,Pb&amp;Sr2EuCu207 <lb/>phase in the bulk 1222 <lb/>structure of the parent compound and indications of periodicity that suggest the possible <lb/>existence of a new, composite structure 2434 phase. <lb/></front>
11
+
12
+ <body>INTRODUCTION <lb/>Recently Bi-and Tl-based compounds of composition <lb/>(Bi or Tl),(Sr or Ba)z(Ln,L,Ce,)C.u2010 <lb/>(where Ln is <lb/>Sm, Eu, or Gd) have been successfully synthesized by <lb/>Tokura et al.&apos; In these so-called 2222~structure cuprates, <lb/>adjacent CuO, layers are separated by a Lnz _ ,Ce,o2 flu-<lb/>orite block and are alternately offset in the a direction. The <lb/>Bi-based compounds display bulk superconductivity after <lb/>annealing in 80 bar O2 at 550-600 &quot;C for 10 h. The transi-<lb/>tion temperature initially reported was about 25 K but this <lb/>has been recently improved to 40 K by optimizing &apos;the <lb/>synthesis and annealing conditions.2 Superconductivity has <lb/>so far not been found under these conditions in the Tl-s <lb/>based 2222 compound,&apos; and in the Tl-0 monolayer Tl-Ba-<lb/>Ln(Ce)-Cu-0 <lb/>1222 compound, which has been synthe-<lb/>sized by Martin et al3 <lb/>The discovery of superconductivity in the Tl-Sr-Ca-<lb/>Cu-0 system with T, at 20 K (Ref. 4) and its optimization <lb/>via Ce doping to a bulk T, of 65 K (Ref. 5), and the <lb/>discovery <lb/>of <lb/>superconductivity <lb/>in <lb/>isostructural <lb/>(Tlo,SPbO.s)Sr&amp;!aCuzO, with a Tc of 85 K (Ref. 6) [tihich <lb/>has now been raised to 105 K (Ref. 7) by Y doping at the <lb/>Ca sites], suggested that the synthesis of the related com-<lb/>pounds of composition ( Tle,Pbc.5) Sr2( Ln, _ &amp;e;)C!u209 <lb/>having the 1222 structure may be possible. Phases of for-<lb/>mula (Tli -,Pb,) (Sri -,LaJ2(Lnl <lb/>-yCe )2CuZ09 in the <lb/>1222 structure were indeed synthesized,&apos;but found to be <lb/>nonsuperconducting. In this paper we report the synthesis, <lb/>in essentially single phase form, of a new compound of <lb/>composition ( TlcsPbc.s) SrZ( Eu, _ ,Ce,) CuZ09 having the <lb/>1222 structure [abbreviated here as (Tl,Pb)-12221, and its <lb/>doping under high 0, pressure to give a bulk superconduc-<lb/>tor with a T, onset near 40 K. This i.s the, first report, to our <lb/>knowledge, of superconductivity <lb/>in the (Tl,Pb)-1222 <lb/>structure cuprates; superconductivity at 25 K has however <lb/>been reported by Maeda et al.&apos; in a (Pb,Cu)-1222 cuprate <lb/>of composition (Pb,Cu) ( Sr,Eu)Z( Eu,Ce)2Cu20Z. The gen-<lb/>eral features of the structure and microstructure of the <lb/>(Tl,Pb)-1222 compound have also been determined. by <lb/>powder x-ray dilIraction, selected area electron diffraction, <lb/>and high-resolution lattice imaging. <lb/>SYNTHESlS <lb/>The best specimens of (Tl,Pb)-1222 were obtained <lb/>with a-nominal content of Ce in the above-given formula of <lb/>x = 0.2: The samples were prepared in two steps. A matrix <lb/>of nominal composition Sr2( Euc.&amp;!ect ) &amp;uZO, was first <lb/>prepared from a well-ground precursor mixture of SrO, <lb/>EuzO3, Ce&amp;, and CuO, which was fired in air at 950 &quot;C! for <lb/>16 h. The matrix powder was then mixed with a stoichio-<lb/>metric amount of PbO and B 10% excess by weight of <lb/>T1203, followed by pelletization. The pellet was placedin-<lb/>sidea gold tube, which was closed by crimping on one side <lb/>and by spot welding on the other side. The gold tube was <lb/>pushed into a furnace at 925 &quot;C under flowing O2 and held <lb/>at this temperature for 6 h. At the end of this period the <lb/>furnace was opened and the sample allowed to cool rapidly <lb/>(down to 30 &quot;C in about 30 min) under flowing OF <lb/>RESULTS AND DISCUSSION <lb/>The, powder x-ray (Ct.&amp;a: radiation) diffraction pat-<lb/>tern (measured using a Rigaku diffractometer) of the as-<lb/>prepared material is shown in Fig. 1. Except for unreacted <lb/>CeOZ and a small amount of an impurity phase (see dis-<lb/>cussion below) indicated by crosses in Fig. 1, the diffrac-<lb/>tion lines can be indexed to the same structure (tetra-<lb/></body>
13
+
14
+ <page>2234 <lb/></page>
15
+
16
+ <note place="footnote">J. Appl. Phys. 70 (4), 15 August 1991 <lb/>0021-8979/91/042234-04$03.00 <lb/>@ 1991 American Institute of Physics<lb/></note>
17
+
18
+ <page>2234 <lb/></page>
19
+
20
+ <note place="footnote">Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 130.102.82.20 On: Wed, 05 <lb/>Oct 2016 05:48:35 <lb/></note>
21
+
22
+ <body>h <lb/>P 16161-<lb/>2 <lb/>s <lb/>s <lb/>12121-<lb/>% <lb/>.% <lb/>: <lb/>I <lb/>8080-<lb/>t <lb/>4040-<lb/>gonal, <lb/>space group <lb/>I4/mmm) <lb/>reported <lb/>for the <lb/>(Tlr -$&apos;b,) (Srt -..La,) (Lnt -,Ce,)@r20g <lb/>compounds by <lb/>Mochiku et aL8 <lb/>Selected area electron diffraction patterns yielded av-<lb/>erage lattice parameters a z b = 3.91 A and c = 29.98 A. <lb/>A representative high resolution lattice image obtained us-<lb/>ing a JEOL 4000 FX electron microscope is shown in Fig. <lb/>2 together with the corresponding electron diffraction pat-<lb/>FIG. 2. [X0] lattice image of a Tle5PbO&amp;Eu2 -xCexCuz09 (x(0.1 to <lb/>0.17) microcrystal. Inset on the left shows the corresponding selected <lb/>area diffraction pattern. Inset at the center shows a computer simulated <lb/>image. <lb/>FIG. 1. Powder x-ray diffraction (CIA&amp;) <lb/>of a typical as-synthesized sample of <lb/>TlasPbO.sSr,Euz _ .$e,Cu,O, (x(0.1 to <lb/>0.17). $ symbols refer to reflections due <lb/>to impurity. <lb/>tern and image simulation. The image is in the [loo] pro-<lb/>jection and shoivs the layer sequence of the structure. Us; <lb/>ing the structural parameters reported by Mochiku et al.,&apos; <lb/>a simulated image in good agreement with the structure is <lb/>obtained (inset, Fig. 2). The structure proposed for the <lb/>(Tl,Pb)-1222 compound is shown in Fig. 3. An important <lb/>feature of the structure is a MiO, fluorite block (where M&apos; <lb/>in our sample is Eu doped by Ce) which separates the <lb/>O(4) <lb/>P <lb/>TIO <lb/>c---------MO <lb/>c <lb/>I;-<lb/>--<lb/>TlO <lb/>b <lb/>a <lb/>FIG. 3. Proposed structure of the Tlo.sPbO.sSrzEuz _ $exCuz09 (x(0.1 to <lb/>0.17) compound. The Tl layer is composed of Tla,,Pb,s and M&apos; is Eu <lb/>doped with Ce. <lb/></body>
23
+
24
+ <page>2235 <lb/></page>
25
+
26
+ <note place="footnote">J. Appl. Phys., Vol. 70, No. 4, 15 August 1991 <lb/>lqbal et al.</note> <lb/>
27
+
28
+ <page>2235 <lb/></page>
29
+
30
+ <note place="footnote">Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 130.102.82.20 On: Wed, 05 <lb/>Oct 2016 05:48:35 <lb/></note>
31
+
32
+ <body>CuO? layers of the structure. As evident from the lattice <lb/>6 <lb/>image, the Cu atoms in consecutive Cu02 layers down the <lb/>c axis are offset in the a direction by (a + b)/2 (where a <lb/>and b are 2-dimensional unit cell vectors for the pseudo-<lb/>tetragonal Cu02 sheet). The charge reservoir layer is a <lb/>single Tl-0 plane in which half the Tl ions are replaced by <lb/>Pb ions. Because of the strongly oxidizing synthesis condi-<lb/>tions used, the oxidation states of Tl and Pb are likely to be <lb/>+ 3 and + 4, respectively. An oxidation state of + 4 has <lb/>been determined for Pb in the 77 K &quot;12 13&quot; superconduc-<lb/>tor of composition Pbo.sSrBal.zYo.,Cao.3CU307 + 6 (Ref. <lb/>10) prepared under similar conditions. In addition, the <lb/>lattice images from some of the microcrystals show inter-<lb/>growths of the single M&apos; layer (Tl,Pb)-1212 phase of com-<lb/>position (TlcSPbo.JSr,EuCuzO-/. In some of the crystal-<lb/>lites the (TI,Pb)-1212 intergrowths display some degree of <lb/>periodicity, indicating the possible existence of a new com-<lb/>posite phase; namely, a (TI,Pb)-2434 compound consisting <lb/>of an alternating fluorite block and single M&apos; layer primi-<lb/>tive unit cell. In the YBa,Cu,Os (Y-124) compound, for <lb/>example, similar periodic intergrowths of the YBa@@7 <lb/>(Y-123) <lb/>phase can occur to give a composite <lb/>Y2Ba4Cu701s (Y-247) phase.&apos; * Preliminary x-ray diffrac-<lb/>tion data indicate the formation of a (semiconducting) <lb/>(Tl,Pb)-2434 compound from a stoichiometric mixture of <lb/>the oxides under similar conditions as described above. <lb/>Results of doping experiments and detailed characteriza-<lb/>tion of this phase will be published elsewhere. In addition <lb/>to (Tl,Pb)-1212 intergrowths, some isolated crystallites of <lb/>the (Tl,Pb)-12 12 phase were also detected. X-ray reflec-<lb/>tions associated with an impurity phase in Fig. 1 are there-<lb/>fore likely to be due to this compound. This was also <lb/>confirmed <lb/>by <lb/>synthesizing <lb/>nearly <lb/>pure <lb/>(Tlc5Pb0.5)Sr2EuCuz07, which was found to be nonsuper-<lb/>conducting down to 4 K in the as-prepared form as well as <lb/>on 6-bar O2 annealing. <lb/>Analytical transmission electron microscopy using en-<lb/>ergy dispersive spectroscopy with an intrinsic germanium <lb/>high-angle x-ray detector was used to determine the atomic <lb/>compositions of a series of crystallites from various prep-<lb/>arations. Most of the crystallites corresponded to a <lb/>Tl(Pb)-Sr-Eu-Cu <lb/>1222 composition in which the Tl:Pb ra-<lb/>tio is 1:l. Ce was barely detectable in 80% of the crystal-<lb/>lites, consistent with the observation of unreacted crystal-<lb/>line Ce02 in the product. However, 20% of the grains did <lb/>contain Ce with x near 0.17. <lb/>Four-probe dc resistance measurements of the as-pre-<lb/>pared (Tl,Pb)-1222 product shows semiconducting behav-<lb/>ior (Fig. 4) consistent with the observations of Tokura et <lb/>al. in Tl-2222 and Mochiku et al8 in related compounds. <lb/>However, a sample annealed at 550 &quot;C for 24 h under 6 <lb/>bars of O2 in a high pressure bomb indicates reduced semi-<lb/>conducting behavior and a resistance drop with an onset <lb/>near 22 K, as evident from Fig. 41 Zero-field-cooled <lb/>SQUID magnetic susceptibility data shown in Fig. 5 rep-<lb/>resent results for the 6-bar 02-annealed sample and a 100-<lb/>bar 02-annealed sample processed for 5 h at 550 &quot;C! in a <lb/>commercial high-pressure O2 furnace.&quot; Magnetic suscep-<lb/>tibility data for an as-prepared sample shows no indication<lb/></body>
33
+
34
+ <page>2236 <lb/></page>
35
+
36
+ <note place="footnote">J. Appl. Phys., Vol. 70, No. 4, 15 August 1991<lb/></note>
37
+
38
+ <body>5 <lb/>4 <lb/>g3 <lb/>2 <lb/>1 <lb/>0 <lb/>0 <lb/>40 <lb/>60 <lb/>120 <lb/>160 <lb/>200 <lb/>240 <lb/>260 <lb/>T (K) <lb/>FIG. 4. Normalized four-probe dc resistivity (using I-mh measuring <lb/>current) as a function of temperature for an as-synthesized sample (curve <lb/>and <lb/>6-bars Orannealed <lb/>sample <lb/>f;)~,,Pbo,,Sr2J&amp; _xCexCuzOg (x(0.1 to 0.17). <lb/>(curve 6) <lb/>of <lb/>of diamagnetism consistent with the resistivity data. The <lb/>6-bar O,-annealed sample shows an approximately 3%-<lb/>volume diamagnetic fraction with diamagnetic onset near <lb/>20 K, consistent with the observed drop in resistance. The <lb/>lOO-bar.02-annealed sample clearly shows an enhancement <lb/>of the diamagnetic volume fraction to nearly 20% and an <lb/>increase of the diamagnetic onset T, to near 40 K. Field-<lb/>cooled Meissner fractions in this and similarly prepared <lb/>samples were found to be greater than 10% by volume, <lb/>indicating the occurrence of bulk superconductivity. How-<lb/>ever, on annealing the sample under 200 bars of O2 at <lb/>550 &quot;C! for 10 h, the diamagnetic fraction is sizably re-<lb/>duced. No changes are observed in the x-ray diffraction <lb/>patterns of the high 02-pressure annealed samples of <lb/>(Tl,Pb)-1222 indicating that doping affects only the oxy-<lb/>gen sublattice. Samples annealed at 100 bars of O2 show <lb/>Temperature <lb/>(K) <lb/>FIG. 5. Zero-field-cooled dc SQUID data at 10 Oe as a function of tem-<lb/>perature for: (A) 6-bars 02-annealed sample and (B) lCO-bar <lb/>Oz-annealed sample of Tle,Pbo.,Sr2Eu2 _ ,CexCu209 (x(0.1 to 0.17). <lb/></body>
39
+
40
+ <note place="headnote">lqbal et a/. <lb/></note>
41
+
42
+ <page>2236 <lb/></page>
43
+
44
+ <note place="footnote">Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 130.102.82.20 On: Wed, 05 <lb/>Oct 2016 05:48:35 <lb/></note>
45
+
46
+ <body>decreased resistance and sizable drops near T,, but zero <lb/>resistance down to -12 K (the limit of our resistance <lb/>data) was not observed, although large Meissner fractions <lb/>were found in these samples. In addition, the temperature <lb/>dependence of the resistance remained weakly semicon-<lb/>ducting, suggesting that the grain boundary regions are not <lb/>optimally doped in the ceramic samples. <lb/>The oxygen content of the Tl-based cuprates cannot be <lb/>determined either by iodometric titration or thermogravi-<lb/>metric methods. Also, neutron diffraction experiments are <lb/>complicated by strongly neutron absorbing rare earth ele-<lb/>ments in the (Tl,Pb)-1222 compounds. Nevertheless, the <lb/>oxygen content can be qualitatively estimated on the basis <lb/>of the expected oxidation state of the constituent ions; thus, <lb/>for (T10.sPbo.5)SrzCaCuzO~, the 0 content is estimated to <lb/>be 6.75, assuming that Pb is + 4 and Cu is + 2. In order <lb/>to achieve the oxygen level of 7 for the layered oxygen-<lb/>deficient perovskite 1212 structure, the compound has to <lb/>be oxygen doped to form Cu + 3, which leads to supercon-<lb/>ductivity when the doping reaches a value near the semi-<lb/>conductor-to-metal transition phase boundary. The results <lb/>of Barry et al6 for (T1,,5Pbo.5)Sr2CaCu207 have shown <lb/>that this doping level is achieved at l-bar O2 to give a T, <lb/>near 85 K. However, some samples showed a resistance <lb/>drop at 105 K, which, in the light of the more recent <lb/>results on electron doping via Y + 3 substitution at the <lb/>Ca + &apos; sites,7 sugg ests that the as-synthesized samples were <lb/>somewhat overdoped, and that an optimally doped sample <lb/>would have a T, near 105 K. In the case of <lb/>(Tlo.sPbo.s)SrzEuz _ XCe,Cu209 similar charge counting <lb/>(neglecting Ce) yields an 0 content of 8.75, whereas for <lb/>(TlesPbo,s)SrZEuCu207, the 0 content is 7.25. Both these <lb/>compounds are semiconducting in the as-prepared (l-bar-<lb/>0,) state-the (Tl,Pb)-1222 compound appears to remain <lb/>hole deficient under conditions of ambient pressure oxygen <lb/>synthesis, whereas ( Tlo.sPbo.s) Sr2EuCu207 is overdoped <lb/>under these circumstances. The hole deficiency is removed <lb/>in the (Tl,Pb) -1222 compound by high pressure (to about <lb/>loo-bar-02) oxygen annealing, but doping with oxygen at <lb/>interstitial sites occurs at pressures above 200-bar-O2 lead-<lb/>ing to a decrease in the superconducting volume fraction. <lb/>Note that the T, of 40 K for (Tl,Pb)-1222, Bi-2222, and <lb/>(Pb,Cu)-1222 corresponds to that of a typical single <lb/>CuOZ <lb/>layer <lb/>cuprate <lb/>such <lb/>as the <lb/>T-structure <lb/>La2 _ ,Sr,CuO$ compound. This suggests that in the 1222 <lb/>and 2222 cuprates the CuOZ sheets are effectively indepen-<lb/>dent since they are separated by Mao, fluorite blocks. <lb/></body>
47
+
48
+ <listBibl>&apos;Y. Tokura, T. Arima, H. Takagi, S. Uchida, T. Ishigaki, H. Asano, R. <lb/>Beyers, A. I. Nazzal, P. Lacorre, and J. B. Torrance, Nature (London) <lb/>342, 890 (1989). <lb/>&apos;T. Arima, Y. Tokura, H. Takagi, S. Uchida, R. Beyers, and J. B. Tor-<lb/>rance, Physica C 168, 79 ( 1990). <lb/>&apos;C. Martin, D. Bourgault, M. Hervieu, C. Michel, J. Provost, and B. <lb/>Raveau, Mod. Phys. Lett. B 3, 933 (1989). <lb/>4Z. Z. Sheng, A. M. Hermann, D. C. Vier, S. Schultz, S. B. Osero-ff, D. <lb/>J. George, and R. M. Hazen, Phys. Rev. B 38, 7074 (1988). <lb/>&apos;Z. Iqbal, B. L. Ramakrishna, and J. C. Barry, Physica C 169, 396 <lb/>(1990). <lb/>6J. C. Barry, Z. Iqbal, B. L. Ramakrishna, R. Sharma, H. Eckhardt, and <lb/>F. Reidinger, J. Appl. Phys. 65, 5207 (1989), and references therein. <lb/>&apos;R. S. Liu, J. M. L&amp;g, S. F. Wu, Y. T. Huang, P. T. Wu, and L. J. <lb/>Chen, Physica C 159, 385 (1989). <lb/>&quot;T. Mochiku, T. Nagashima, Y. Saito, M. Watahiki, H. Asano, and Y. <lb/>Fukai, Jpn. J. Appl. Phys. 29, L588 (1990). <lb/>&apos;T. Maeda, K. Sakuyama, S. Koriyama, A. Ichinose, H. Yamauchi, and <lb/>S. Tanaka, Physica C 169, 133 (1990). <lb/>*OX. X. Tang and D. E. Morris, Phys. Rev. B (in press). <lb/>&quot;D. E. Morris, N. G. Asmar, J. Y. T. Wei, J. H. Nickel, R. L. Sid, J. S. <lb/>Scott, and J. E. Post, Phys. Rev. B 40, 11406 (1989). <lb/>&apos;2Morris Research Inc., 1918 University Ave., Berkeley, CA 947041 <lb/></listBibl>
49
+
50
+ <page>2237 <lb/></page>
51
+
52
+ <note place="footnote">J. Appl. Phys., Vol. 70, No. 4, 15 August 1991 <lb/>lqbal et a/. <lb/></note>
53
+
54
+ <page>2237 <lb/></page>
55
+
56
+ <note place="footnote">Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 130.102.82.20 On: Wed, 05 <lb/>Oct 2016 05:48:35 </note>
57
+
58
+
59
+ </text>
60
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JAP0805169-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <titlePage>Thermoelectric power studies on Nd1.82−xSrxCe0.18CuOy:x≤0.18 <lb/>superconductors <lb/>Okram G. Singh, B. D. Padalia, Om Prakash, S. K. Agarwal, and A. V. Narlikar <lb/>Citation: J. Appl. Phys. 80, 5169 (1996); doi: 10.1063/1.363500 <lb/>View online: http://dx.doi.org/10.1063/1.363500 <lb/>View Table of Contents: http://jap.aip.org/resource/1/JAPIAU/v80/i9 <lb/>Published by the American Institute of Physics. <lb/>Related Articles <lb/>Chemical pressure and electron doping effects in SrPd2Ge2 single crystals <lb/>J. Appl. Phys. 111, 07E117 (2012) <lb/>Fluctuation-induced conductivity analyses of Be-doped (Bi0.25Cu0.25Li0.25Tl0.25)Ba2Ca2Cu3O10-δ <lb/>superconductors in the critical regime and beyond <lb/>J. Appl. Phys. 111, 033917 (2012) <lb/>Enhancement of the upper critical field in codoped iron-arsenic high-temperature superconductors <lb/>J. Appl. Phys. 110, 123906 (2011) <lb/>Superconductivity at 14.6 K in Fe(SeTe) single crystal and the role of excess Fe <lb/>J. Appl. Phys. 110, 033914 (2011) <lb/>Andreev reflection spectroscopy of the new Fe-based superconductor EuAsFeO0.85F0.15: Evidence of strong <lb/>anisotropy in the order parameter <lb/>Low Temp. Phys. 37, 280 (2011) <lb/>Additional information on J. Appl. Phys. <lb/>Journal Homepage: http://jap.aip.org/ <lb/>Journal Information: http://jap.aip.org/about/about_the_journal <lb/>Top downloads: http://jap.aip.org/features/most_downloaded <lb/>Information for Authors: http://jap.aip.org/authors <lb/>Downloaded 26 Feb 2012 to 14.139.97.73. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions<lb/></titlePage>
9
+
10
+ <front>Thermoelectric power studies on Nd 1.82؊x Sr x Ce 0.18 CuO y :xр0.18 <lb/>superconductors <lb/>Okram G. Singh a) and B. D. Padalia <lb/>Department of Physics, Indian Institute of Technology, Powai, Bombay 400 076, India <lb/>Om Prakash <lb/>Department of Metallurgical Engineering and Materials Science, Indian Institute of Technology, Powai, <lb/>Bombay 400 076, India <lb/>S. K. Agarwal and A. V. Narlikar b) <lb/>Superconductivity Group, National Physical Laboratory, New Delhi 110012, India <lb/>͑Received 21 March 1996; accepted for publication 15 July 1996͒ <lb/>Thermoelectric power (S) studies on a Nd 1.82Ϫx Sr x Ce 0.18 CuO y :xр0.18 superconducting system in <lb/>the temperature range 35-250 K are reported here. In the xϭ0.09 sample, synthesized in the <lb/>reduced environment, the small magnitude of S is highly metalliclike and its sign is negative, a <lb/>characteristic of electron conduction. The sign of S for the xϭ0.18 sample shows a crossover below <lb/>75 K from negative to positive, in apparent conflict with electronic conduction. Interestingly, after <lb/>oxygenation this sample exhibits a broadened but positive phonon draglike peak. This oxygenated <lb/>sample shows overcompensation of the carrier ͑electron͒ concentration. Critical analysis of the data <lb/>suggests that Sr doping seemingly causes a competition between electron-and holelike conduction. <lb/>The slope dS/dT is, in general, negative suggesting that the main contribution is coming from the <lb/>diffusive part. The observed thermopower features seem to fall in line with the theoretical curves of <lb/>Durczewski and Ausloos ͓Z. Phys. B 85, 59 ͑1991͒; Phys. Rev. B 53, 1762 ͑1996͔͒ based on the <lb/>inelastic scattering of quasifree electrons by phonons. © 1996 American Institute of Physics. <lb/>͓S0021-8979͑96͒06820-X͔ <lb/></front>
11
+
12
+ <body>I. INTRODUCTION <lb/>Thermoelectric power ͑TEP͒ studies are of considerable <lb/>interest in the understanding the nature of the charge carriers <lb/>and transport mechanism in cuprate superconductors. De-<lb/>spite several TEP studies on various cuprate superconducting <lb/>systems, 1-4 discrepancies continue to exist. Uher et al. 1 and <lb/>Kaiser 2 argue in favor of electron-phonon enhancement in <lb/>the temperature dependence of metallic diffusion ther-<lb/>mopower. The idea of the presence of two types of carriers <lb/>was suggested by Uher et al. 1 in the hole-doped <lb/>La 2Ϫx Sr x CuO 4 superconductor system which exhibits a posi-<lb/>tive Hall coefficient as well as negative diffusion ther-<lb/>mopower. Lopez-Morales et al., 3 however, do not favor such <lb/>a separation of carriers into specific negative and positive <lb/>entities. Trodahl 4 attempted to explain thermopower features <lb/>using the usual Fermi-liquid picture. Nevertheless, the tem-<lb/>perature dependence of the TEP still remains unresolved. <lb/>According to this scenario, the discovery of an electron-<lb/>doped R 2Ϫx Ce x CuO 4Ϫ␦ ; RϭNd, Sm, or Pr superconductor <lb/>system becomes significant. 6,7 These superconductors are <lb/>considered to be potential candidates for the crucial test in <lb/>formulating theories to understand the mechanism of super-<lb/>conductivity as the majority charge carriers in superconduct-<lb/>ing cuprates can be either holes or electrons. Electrons are <lb/>apparently doped into the Nd 2Ϫx Ce x CuO y ͑NCCO͒ system 6,8 <lb/>due primarily to ͑i͒ tetravalent Ce substitution for trivalent <lb/>Nd and ͑ii͒ creation of oxygen vacancy through reduction. <lb/>An increase in x leads to enhancement in conduction electron <lb/>density as Ce continues to be in mixed ͑3ϩ and 4ϩ͒ valence <lb/>states. 9,10 Eventually, for xϾ0.18, T c disappears. 6,11 <lb/>Interestingly, superconductivity gets revived in this sys-<lb/>tem when holes are nominally doped by the substitution of <lb/>divalent alkaline earth elements 12 for the trivalent Nd. This <lb/>result suggests that superconductivity is attained for certain <lb/>concentration of charge carriers and the majority of the car-<lb/>riers are electrons in both the parent NCCO and the revived <lb/>Nd 1.82Ϫx Sr x Ce 0.18 CuO y ͑NSCCO͒ superconductor systems. <lb/>Our preliminary study showed that the thermopower sign is <lb/>negative, consistent with electron doping. 12,13 However, it is <lb/>not always negative in its sign, thereby posing difficulty in <lb/>explaining the very nature of the transport properties in these <lb/>materials. In this article, we present the results of our inves-<lb/>tigations on the NSCCO system, synthesized in different en-<lb/>vironments. <lb/>II. EXPERIMENTS <lb/>The Nd 1.82Ϫx Sr x Ce 0.18 CuO y :р0.18 samples were syn-<lb/>thesized by the usual solid state reaction route. The synthesis <lb/>and characterization details are given elsewhere. An x-ray <lb/>diffraction ͑XRD͒ analysis revealed formation of the TЈ-type <lb/>structure and the presence of negligible impurities. The dc <lb/>four probe resistivity data were obtained in the temperature <lb/>range 5-300 K in a liquid helium bath cryostat using a car-<lb/>bon glass temperature sensor. The entire data acquisition sys-<lb/>tem, comprised of a ͑Keithley 181͒ nanovoltmeter, a ͑Kei-<lb/>thley 224͒ constant current source, a ͑Lakeshore model 82C͒ <lb/>temperature controller, and an indicator, was hooked to a HP <lb/>system controller. <lb/>a͒ <lb/>Present address: Nuclear Science Centre, Aruna Asaf Ali Road, New Delhi <lb/>110067, India. <lb/>b͒ Electronic mail: [email protected] <lb/></body>
13
+
14
+ <page>5169 <lb/></page>
15
+
16
+ <note place="footnote">J. Appl. Phys. 80 (9), 1 November 1996 <lb/>0021-8979/96/80(9)/5169/6/$10.00 <lb/>© 1996 American Institute of Physics <lb/>Downloaded 26 Feb 2012 to 14.139.97.73. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions <lb/></note>
17
+
18
+ <body>TEP was measured by the differential method in the <lb/>temperature range 35-250 K using a closed cycle refrigera-<lb/>tor ͑CCR͒. 14 The sample, with a uniform surface, was <lb/>clamped between two thin copper plates that served as ref-<lb/>erence samples as well as voltage leads. The voltages were <lb/>measured using a Keithley nanovoltmeter and the tempera-<lb/>tures were recorded using silicon diode sensors attached to <lb/>the plates. The measured thermopower was corrected for the <lb/>͑Cu͒ reference material. A Mylar piece wrapping each sensor <lb/>ensured good thermal contact and electrical insulation. One <lb/>of the copper plates was placed on the cold head of the CCR <lb/>and the other one was kept in contact with the clamp that lies <lb/>in the open space of the vacuum chamber. While the system <lb/>is cooling down, temperature of the cold-head-facing sample <lb/>surface is lower than that at the upper face thereby establish-<lb/>ing a temperature gradient ͑⌬T, typically ϳ2 K͒ across the <lb/>sample. The variation in the temperature gradient ⌬T with <lb/>the lowering of temperature does not seem to affect the re-<lb/>sulting data. The reason is that the TEP goes to zero at Rϭ0 <lb/>͑as expected͒ in all the measurements carried out in our labo-<lb/>ratory using the same setup on the higher T c superconductors <lb/>such as YBa 2 Cu 3 O 7Ϫ␦ and Bi 2 Sr 2 Ca 2 Cu 3 O y and, more im-<lb/>portant, its TEP features compare exactly with those already <lb/>published in the literature. 14 The mean of the temperatures <lb/>read at the two sensors was chosen as the sample tempera-<lb/>ture. For such a temperature, the voltage difference (⌬V) <lb/>between the two terminals was noted to find out the TEP <lb/>ϭ(⌬V/⌬T). In this manner, the data were collected at the <lb/>temperature interval of ϳ0.5 K but the plots were made us-<lb/>ing a few data points for the sake of clarity. The overall error <lb/>in the TEP measurements is ϳ10%. All these steps were <lb/>done automatically as in the dc four-probe method. The sign <lb/>of the TEP was assigned using a conventional criterion: if <lb/>the absolute TEP of the reference metal is zero, the higher <lb/>temperature terminal will be positive potential with respect <lb/>to the lower temperature one, provided the TEP of the <lb/>sample under consideration is positive. <lb/>III. RESULTS <lb/>Figure 1 shows the normalized resistance R vs T plot, <lb/>where RϭR(T)/R(300) for the Nd 1.82Ϫx Sr x Ce 0.18 CuO y : <lb/>xϭ0, 0.09, 0.18 ͑all reduced͒, and xϭ0.18 ͑oxygenated͒ <lb/>samples, designated as NCCO, Sr0.09, Sr0.18 and <lb/>Sr0.18͑O͒, respectively. The NCCO sample ͑curve a͒ shows <lb/>superconductivity onset only at 18 K and no Rϭ0 down to 5 <lb/>K. The normal state resistivity exhibits semiconductinglike <lb/>behavior. The Sr0.09 sample ͑curve b͒, on the other hand, <lb/>exhibits a T c onset at 20 K and Rϭ0 at 12 K. A more <lb/>interesting aspect of this sample is exhibition of a normal <lb/>state metalliclike behavior. The metallic behavior in the case <lb/>of electron-doped polycrystalline cuprates is rare, 15,16 indi-<lb/>cating the distinct feature of this ͑NSCCO͒ system. 12 For the <lb/>higher Sr concentration Sr0.18 sample ͑curve c͒ T c onset is <lb/>at 18 K and Rϭ0 at 8 K. The normal state behavior is <lb/>weakly semiconductinglike. The same sample when oxygen-<lb/>ated ͑curve d͒ shows only semiconductinglike resistivity be-<lb/>havior. <lb/>The thermopower S vs T of the NCCO, Sr0.09, Sr0.18, <lb/>and Sr0.18͑O͒ samples is shown in Fig. 2. All the reduced <lb/>samples reveal similar features in thermopower curves; the <lb/>magnitude of S is small and is comparable to that of noble <lb/>metals, electron-doped cuprates, and YBa 2 Cu 3 O 7 . 5,17,18 For <lb/>the NCCO and Sr0.09 samples, the sign of S is negative in <lb/>the entire temperature range of investigation, consistent with <lb/>electron doping. S suddenly increases below about 50 K sug-<lb/>gesting possible extrapolation to Sϭ0 at 12 K, ͓the observed <lb/>value of T c (Rϭ0)͔ and has a small dip at 55 K. The dip may <lb/>be compared with the phonon-drag peak observed in noble <lb/>metals and in some superconducting cuprates, often with <lb/>positive S values. 5,17 Surprisingly, the thermopower of the <lb/>NCCO sample with vanishing superconductivity ͑T c onset <lb/>ϳ18 K͒ shows an almost similar trend ͑cf. Refs. 11 and 12͒. <lb/>However, the magnitude of S is small, a typical characteristic <lb/>FIG. 1. Normalized resistance R vs T plots for the Nd 1.82Ϫx Sr x Ce 0.18 CuO y <lb/>with ͑a͒ xϭ0, ͑b͒ xϭ0.09, ͑c͒ xϭ0.18 ͑all reduced͒, and ͑d͒ xϭ0.18 ͑oxy-<lb/>genated͒ samples. <lb/>FIG. 2. S vs T plots for the Nd 1.82Ϫx Sr x Ce 0.18 CuO y with ͑a͒ xϭ0, ͑b͒ <lb/>xϭ0.09, ͑c͒ xϭ0.18 ͑all reduced͒, and ͑d͒ xϭ0.18 ͑oxygenated͒ samples. <lb/></body>
19
+
20
+ <page>5170 <lb/></page>
21
+
22
+ <note place="footnote">J. Appl. Phys., Vol. 80, No. 9, 1 November 1996 <lb/>Singh et al. <lb/></note>
23
+
24
+ <note place="footnote">Downloaded 26 Feb 2012 to 14.139.97.73. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions <lb/></note>
25
+
26
+ <body>of metals. This feature is in contrast to the semiconducting-<lb/>like behavior exhibited by the sample in the R vs T plot ͓Fig. <lb/>1, curve ͑c͔͒. <lb/>The Sr0.18 sample exhibits almost linear S(T) from 250 <lb/>down to 110 K. Below 110 K, S increases faster and offsets <lb/>from negative to positive value at 75 K. Whether the value of <lb/>S would turn back to zero at T c ϳ18 K is not clear from these <lb/>data ͑the lowest temperature T measured being about 35 K͒. <lb/>On the contrary, after the sample is oxygenated, ͑i͒ the mag-<lb/>nitude of S is 6-8 times larger, ͑ii͒ the phonon-drag peak is <lb/>broadened, with positive S value, and ͑iii͒ the two offsets, <lb/>negative to positive at 110 K and positive to negative at <lb/>͑extrapolated͒ 30 K, are exhibited. However, the enhanced <lb/>magnitude of S is of the same order as that of good quality <lb/>hole superconductors 5 albeit the resistivity data ͓Fig. 1, curve <lb/>͑d͔͒ indicate semiconductinglike behavior. <lb/>IV. DISCUSSION <lb/>For a metal, the sign of diffusion thermopower (S d ) is <lb/>considered to be indicative of the sign of the majority charge <lb/>carriers, and S d is given by 17 <lb/>S d ϭϪ 2 k 2 T/3e͕‫͓ץ‬ln ͑E ͔͒/‫ץ‬E͖ E F , <lb/>͑1͒ <lb/>where (E) is the conductivity at electron energy E, k is <lb/>Boltzmann&apos;s constant, e is the electronic charge, and E F is <lb/>the Fermi energy. Equation ͑1͒ assumes ͑i͒ a common relax-<lb/>ation time for scattering in both a temperature gradient and <lb/>an electric field, ͑ii͒ elastic scattering in pure metals when <lb/>temperature TϾ D and in alloys when the resistance is <lb/>dominated by impurities, and ͑iii͒ no restriction on the shape <lb/>of the Fermi surface provided the Fermi energy E F ӷkT. <lb/>S d is a linear function of temperature; its magnitude and <lb/>sign depend on how the conductivity (E) changes with <lb/>electron energy at the Fermi surface. Generally, conductivity <lb/>is expected to increase with increasing energy of the elec-<lb/>trons, thereby, as a rule, negative thermopowers are expected <lb/>unless other conditions are imposed. The detailed behavior <lb/>of the diffusion thermopower of all metals and alloys there-<lb/>fore depends on 17 <lb/>͑E ͒ϭe 2 /4 3 ប 2 <lb/>ϫ ͭ ‫ץ‬ lnͫ ͵ ͵͑ ‫ץ���E/‫ץ‬k͒ 2 ͑k͒dA/ٌ k Eͬ Ͳ ‫ץ‬E ͮ <lb/>E F <lb/>, <lb/>(2) <lb/>showing that S d is dependent on the Fermi surface through <lb/>(‫ץ‬E/‫ץ‬k) 2 and ٌ k E and on the scattering mechanism through <lb/>(k). Here k is the wave vector. Thus, interpretation of dif-<lb/>fusion thermopower is pretty difficult in the real physical <lb/>situation even for simple metals and alloys. As a result, lin-<lb/>earity of the thermopower with temperature is the theoretical <lb/>idealization. <lb/>The situation in the case of the Nd 1.82Ϫx Sr x Ce 0.18 CuO y <lb/>system, therefore, should obviously be more challenging <lb/>than in other high Temperature Superconductors ͑HTSCs͒ 5 <lb/>because the dopants or oxygen nonstoichiometry are lattice <lb/>imperfections in the crystallites of the polycrystalline <lb/>samples. 19 These, in addition to the unaccounted for but un-<lb/>avoidable impurities, are expected to collectively contribute <lb/>to the thermopower for such materials. <lb/>Figure 2 depicts the T dependence of S for the various <lb/>samples. In the Sr0.09 superconducting sample, there is mar-<lb/>ginal, although nonlinear, T dependence. The NCCO and <lb/>Sr0.18 samples, on the other hand, show negligible T depen-<lb/>dence. The Sr0.18͑O͒ sample also exhibits nonlinearity. The <lb/>temperature dependence of thermopower of the <lb/>Nd 1.82Ϫx Sr x Ce 0.18 CuO y materials, therefore, is not linear. <lb/>However, it may be mentioned that, in general, the S(T) <lb/>curves have dS/dTϽ0 ͑in the temperature range of 100-250 <lb/>K͒ in agreement with the earlier studies that diffusion is a <lb/>major contribution 13 to S. <lb/>We now consider contributions from the diffusion (S d ) <lb/>and phonon-drag (S g ) parts in the total thermopower S. This <lb/>can be expressed as 20 <lb/>SϭS d ϩS g ϭATϩBT 3 <lb/>or <lb/>S/TϭAϩBT 2 , <lb/>͑3͒ <lb/>where A and B are constants. Using this equation one can <lb/>extract the diffusive thermopower contribution S d through <lb/>the extrapolated intercept of the linear part of the S/T vs T 2 <lb/>plot. Such a plot also indicates the extent of S g contribution. <lb/>For this purpose, the S/T vs T 2 plots for the samples under <lb/>study are shown in Fig. 3. As can be seen, deviation from <lb/>linear behavior ͑i.e., the phonon-drag enhancement͒ begins <lb/>at temperatures ͑defined as T d ͒ 161, 179, 115, and 152 K for <lb/>the NCCO, Sr0.09, Sr0.18, and Sr0.18͑O͒ samples, respec-<lb/>tively. It is interesting to note that the extent of the phonon-<lb/>drag process is much larger ͑179 K͒ in the maximum T c <lb/>superconductor Sr0.09 than in that ͑115 K͒ of the Sr0.18 <lb/>sample. This is consistent with the general metallic and <lb/>semiconductinglike behavior of the materials. It is also <lb/>worth noting that, in the case of two semiconducting <lb/>samples, NCCO and Sr0.18͑O͒, T d is about 161 and 152 K, <lb/>respectively, which can be understood on the basis of their <lb/>composition and on synthesis conditions. The NCCO sample <lb/>synthesized under reducing conditions with maximum Ce <lb/>content possesses a high density electron concentration. On <lb/>the other hand, the Sr0.18͑O͒ sample prepared under an oxi-<lb/>dizing environment apparently overcompensates the electron <lb/>density generated through Ce incorporation, thereby leading <lb/>to a high density of hole concentration and thus behaving in <lb/>an almost opposite manner to the NCCO sample as far as the <lb/>phonon-drag part is concerned ͑Fig. 3͒. The small although <lb/>negative values of S encountered in the case of the Sr0.18͑O͒ <lb/>sample beyond the crossover at about 100 K indicates the <lb/>dominance of the electron contribution in the diffusive ther-<lb/>mopower region. <lb/>TEP, particularly that contributed from normal electrons <lb/>S that is also the case in the present study, is related to the <lb/>circulatory ( j ) and electronic ( e ) heat conductivities <lb/>j / e ϭ3eS⌬͑T͒/ 2 k 2 T, <lb/>͑4͒ <lb/>where ⌬(T) is the energy per normal carrier relative to the <lb/>superconducting carriers. Assuming that the normal carrier <lb/></body>
27
+
28
+ <note place="footnote">J. Appl. Phys., Vol. 80, No. 9, 1 November 1996 <lb/>Singh et al. <lb/></note>
29
+
30
+ <note place="footnote">Downloaded 26 Feb 2012 to 14.139.97.73. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions <lb/></note>
31
+
32
+ <body>TEP (S) below T c is similar to just below it, the ratio <lb/>j / e ϰS. Using the Wiedmann-Franz law and the electrical <lb/>resistivity, Jezowski and Klamut 22 obtained only a very <lb/>small electronic contribution ͑2.5% at ϳ300 K and 0.1% at <lb/>30 K͒ to the total thermal conductivity. The thermal conduc-<lb/>tivity is thus clearly dominated by phonons and compares <lb/>well with those found in a classical dielectric solid. In other <lb/>words, thermal conductivity associated with the sharp peak <lb/>in the Nd 2Ϫx Ce x CuO y system is extraordinarily large com-<lb/>pared to that in p-type cuprate superconductors. 22 This char-<lb/>acteristic may also be compared with the TEP dips ͑or peaks͒ <lb/>observed in the present samples as these features are found in <lb/>the temperature range where thermal conductivity also <lb/>peaks. However, the discrepancy lies in the peaks in <lb/>conductivity 22 and dips in the TEP of the reduced samples <lb/>that are exhibited. <lb/>Another focal point is the crossover of the TEP from <lb/>negative to positive value in the moderate temperature re-<lb/>gime where TEP diffusion dominates. In a recent theoretical <lb/>approach, model calculations were performed. In that <lb/>model, the simple free-electron scattering from the spherical <lb/>Fermi surface by acoustic phonons alone, without the usual <lb/>electron-phonon mass enhancement effect or Umklapp pro-<lb/>cess, was considered. Sign reversal in the TEP was shown to <lb/>have resulted from a competitive distribution of charge car-<lb/>riers of the different signs with the carriers being inelasti-<lb/>cally scattered. Further, the electron and phonon spectra are <lb/>considered to be on the same footing. Durczewski and <lb/>Ausloos 23 showed that the TEP bump ͑on the positive side͒ <lb/>grows with characteristic Debye energy or with decreasing <lb/>electron energy (⑀ s ), and is always negative for ⑀ s Ͼ48 K. <lb/>These characteristics can be compared with the TEP data of <lb/>the present work in Fig. 2, curves ͑c͒ and ͑d͒. The compari-<lb/>son is more striking in the case of the oxygenated sample <lb/>͓Fig. 2, curve ͑d͔͒. The TEP for the reduced sample ͓Fig. 2, <lb/>curve ͑c͔͒ is however not necessarily very compatible be-<lb/>cause it increases monotonically with decreasing temperature <lb/>down to 35 K. <lb/>We also note that the nonlinear S(T) is associated with <lb/>the decrease in dip as the Sr concentration (x) increases. The <lb/>dip decreases slowly with the introduction of superconduc-<lb/>tivity and may be correlated to the increased disorder due to <lb/>the increased x. 24-27 It is therefore tempting to say that the <lb/>thermopower of NSCCO is something more than the sum of <lb/>diffusion and phonon-drag components as it is in the <lb/>Nd 2Ϫx Ce x CuO y system. 28 <lb/>Presumably both electrons and holes take part in the <lb/>conductivity of ͑NSCCO͒ materials. 12 It, therefore, raises a <lb/>question as to how these two kinds of charge carriers are <lb/>going to respond to the overall transport and electronic prop-<lb/>erties? The observation of a negative sign of thermopower is <lb/>indicative of electrons as the majority carriers that in turn <lb/>corroborates the x-ray spectroscopic ͑XAS͒ data ͑cf. Fig. 2 <lb/>of the Ce L 3 edge of Ref. 10͒. However, in some cases, the <lb/>thermopower offsets to positive value thereby making it dif-<lb/>ficult to understand the transport mechanism. 1-4 Uher et al. 1 <lb/>suggested the presence of two types of carriers in the <lb/>La 2Ϫx Sr x CuO 4 superconductor. Wang et al. 29 made similar <lb/>conclusions in the Nd 2Ϫx Ce x CuO y system. But, Lopez-<lb/>Morales et al. 3 do not favor the idea of the separation of <lb/>carriers into specific negative and positive entities, in line <lb/>with the conclusions of Hu et al. 28 <lb/>For the NSCCO system there is, at present, no evidence <lb/>of the presence of two bands. 10 However, there is a system-<lb/>atic decrease in the unoccupied density of states with oxygen <lb/>depletion. 10 This tends to agree with better conductivity or <lb/>with superconductivity 12 which reflects the change in Fermi <lb/>level in the samples. Nevertheless, it should be pointed out <lb/>that two bands exist 30 in Bi 2 Sr 2 CaCu O y . This is evident <lb/>from the photoemission data that identify a band with the <lb/>Bi-O characteristic crossing the Fermi level and generating <lb/>an electronlike pocket to supplement the CuO plane holes. 30 <lb/>A similar model was also suggested for the Tl Ba 2 Ca 2 Cu 3 O y <lb/>system and was verified. 14,31 Precise understanding of the <lb/>thermopower of the NSCCO system is poor at this juncture <lb/>as it is in the other HTSCs. However, there are some ques-<lb/>tions to be addressed regarding the underlying physics that <lb/>may be hidden in material imperfections such as defects or <lb/>impurities. In Nd 2Ϫx Ce x CuO y and Nd 1.82Ϫx Sr x Ce 0.18 CuO y <lb/>systems, the local oxygen as well as cationic distributions <lb/>were investigated. 19,32 We report that, although the ͑cationic͒ <lb/>dopants are uniformly distributed, the lattice oxygen is non-<lb/>uniform locally which is believed to have far-reaching impli-<lb/>cations on the bulk physical properties of these materials. 19,32 <lb/>The oxygen distribution for the oxygenated samples ͑i.e., the <lb/>FIG. 3. S/T vs T 2 plots for the Nd 1.82Ϫx Sr x Ce 0.18 CuO y with ͑a͒ xϭ0, ͑b͒ <lb/>xϭ0.09, ͑c͒ xϭ0.18 ͑all reduced͒, and ͑d͒ xϭ0.18 ͑oxygenated͒ samples. <lb/>The straight line is drawn to indicate the dominance regime of the diffusion <lb/>thermopower. <lb/></body>
33
+
34
+ <page>5172 <lb/></page>
35
+
36
+ <note place="footnote">J. Appl. Phys., Vol. 80, No. 9, 1 November 1996 <lb/>Singh et al. <lb/></note>
37
+
38
+ <note place="footnote">Downloaded 26 Feb 2012 to 14.139.97.73. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions <lb/></note>
39
+
40
+ <body>nonsuperconductors͒ is uniform as anticipated. Such samples <lb/>do not form a Fermi surface as is evident from the observed <lb/>semiconductinglike behavior ͓Fig. 1, curve ͑d͔͒. The reduced <lb/>samples, on the other hand, form Fermi surfaces, as is <lb/>strongly suggested by the observation of better conductivity <lb/>or superconductivity. 6,16,32 <lb/>At this juncture it would be beneficial to see the thermal <lb/>properties in a wider perspective. 33-36 Hansel et al. 33 re-<lb/>ported that the transport entropy of magnetic flux lines in <lb/>Nd 1.85 Ce 0.15 CuO y at 18 K in 0.5 T is 1.6ϫ10 Ϫ14 J/km, in <lb/>agreement with the time prediction of Ginzburg-Landau <lb/>theory. On the other hand, the Nernst effect of the same thin <lb/>film composition in the mixed state was found to arise from <lb/>the motion of the vortices in response to thermal force, and it <lb/>is anomalously large. 34 This led Jiang et al. 34 to conclude <lb/>that both electrons and holes participate in the electrical <lb/>transport of the superconducting phase of Nd 2Ϫx Ce x CuO y . <lb/>This is, incidentally, coincidental with what Wang et al. <lb/>concluded. 29 The recent theoretical propositions for the ther-<lb/>mopower of semimetals and semiconductors that consider <lb/>the multiband structure and different types of carriers 37 ap-<lb/>parently describe the observed features of temperature de-<lb/>pendence of the TEP in the regime away from the usual low <lb/>temperature phonon-drag region of the high (xϭ0) and op-<lb/>timum (xϭ0.09) electron density samples ͓Fig. 2, curves ͑a͒ <lb/>and ͑b͔͒ where the majority carriers are doped electrons. <lb/>However, this does not seem to hold for the NSCCO system <lb/>as a whole. <lb/>The <lb/>reduced <lb/>samples <lb/>have <lb/>local <lb/>oxygen <lb/>nonuniformity 19,32 that suggests formation of different <lb/>shapes or sizes of Fermi surfaces locally. Therefore, the <lb/>Fermi surface, being the backbone in understanding the <lb/>mechanism of ͑super͒ conductivity or thermopower of such <lb/>materials delving into the Fermi surface topology, is ex-<lb/>pected to be quite revealing. However, in doing this, one <lb/>should consider the contrasting features of the semiconduct-<lb/>inglike behavior in the resistivity data and the metalliclike <lb/>small magnitude of thermopower in the oxygenated samples. <lb/>In this situation, one may visualize the resistivity and ther-<lb/>mopower arising from conduction by metallic and semicon-<lb/>ducting bands in parallel, 39 with the contribution of the semi-<lb/>conductor band to the total conductivity being small. Its <lb/>contribution to the total thermopower is, however, significant <lb/>owing to a very large intrinsic semiconductor thermopower. <lb/>V. CONCLUSIONS <lb/>In summary, we have measured the thermopower S of <lb/>the revived n-type Nd 1.82Ϫx Sr x Ce 0.18 CuO y superconductors <lb/>for different x values. In the Sr0.09 sample, the magnitude of <lb/>S is highly metalliclike and the sign is negative, consistent <lb/>with electron conduction. However, in the Sr0.18 sample, S <lb/>offsets to a positive value below 75 K, in apparent conflict <lb/>with the electron conduction. Similar behavior is observed in <lb/>the case of oxygenated sample. These data show that Sr <lb/>codoping is not a simple addition of holes. Based on the <lb/>estimation of the extent of diffusion thermopower, oxygen-<lb/>ation of the Sr0.18 sample seems to have overcompensated <lb/>the electron density in the material, thereby causing it to <lb/>behave in a mirror-imagelike fashion compared to the NCCO <lb/>sample. The thermopower behavior of the high and optimum <lb/>electron density samples seems to qualitatively match the <lb/>theoretical curves generated for the thermopower of semi-<lb/>metals and semiconductors by Durczewski and Ausloos. Fur-<lb/>ther, the thermopower sign reversal in the reduced and oxy-<lb/>genated (xϭ0.18) samples may be viewed in light of the <lb/>other recent theory by these authors. <lb/></body>
41
+
42
+ <div type="acknowledgement">ACKNOWLEDGMENTS <lb/>Three of the authors ͑O.G.S., B.D.P., and O.P.͒ acknowl-<lb/>edge the support of the Department of Science and Technol-<lb/>ogy, New Delhi ͑Project No. SBR 24͒ for this work. Thanks <lb/>are also due to Professor S. N. Bhatia for his comments and <lb/>Dr. R. Suba for her help. <lb/></div>
43
+
44
+ <listBibl>C. Uher, A. B. Kaiser, E. Gmelin, and L. Walz, Phys. Rev. B 36, 5676 <lb/>͑1987͒. <lb/>2 A. B. Kaiser, Phys. Rev. B 37, 5924 ͑1988͒. <lb/>3 M. E. Lopez-Morales, R. J. Savoy, and P. M. Grant, Solid State Commun. <lb/>71, 1079 ͑1989͒. <lb/>4 H. J. Trodahl, Phys. Rev. B 51, 6175 ͑1995͒. <lb/>5 A. B. Kaiser and C. Uher, Studies of High Temperature Superconductors, <lb/>edited by A. V. Narlikar ͑Nova Science, New York, 1990͒, Vol. 7, and <lb/>reference therein. <lb/>6 Y. Tokura, H. Takagi, and S. Uchida, Nature ͑London͒ 337, 345 ͑1989͒; <lb/>H. Takagi, S. Uchida and Y. Tokura, Phys. Rev. Lett. 62, 1197 ͑1989͒. <lb/>7 V. J. Emery, Nature ͑London͒ 337, 306 ͑1989͒. <lb/>8 Y. Hidaka and M. Suzuki, Nature ͑London͒ 338, 635 ͑1989͒. <lb/>9 J. M. Tranquada, S. M. Heald, A. R. Moodenbaug, G. Liang, and M. <lb/>Croft, Nature ͑London͒ 337, 720 ͑1989͒. <lb/>10 B. D. Padalia, Okram G. Singh, S. J. Gurman, K. Suba, and O. Prakash, <lb/>Physica B 208-209, 533 ͑1995͒. <lb/>11 O. G. Singh, B. D. Padalia, O. Prakash, K. Suba, A. V. Narlikar, and L. C. <lb/>Gupta, Physica C 219, 156 ͑1994͒. <lb/>12 O. G. Singh, O. Prakash, B. D. Padalia, and A. V. Narlikar, Phys. Rev. B <lb/>48, 13182 ͑1993͒. <lb/>13 O. G. Singh, B. D. Padalia, O. Prakash, V. N. Moorthy, and A. V. Narl-<lb/>ikar, J. Appl. Phys. 75, 6740 ͑1994͒. <lb/>14 V. P. S. Awana, V. N. Moorthy, and A. V. Narlikar, Phys. Rev. B 49, <lb/>6385 ͑1994͒; V. P. S. Awana, R. Lal, and A. V. Narlikar, J. Phys., Con-<lb/>dens. Matter. 7, L171 ͑1995͒. <lb/>15 M. E. Lopez-Morales, R. J. Savoy, and P. M. Grant, J. Mater. Res. 5, 2041 <lb/>͑1990͒. <lb/>16 O. G. Singh, O. Prakash, B. D. Padalia, D. Chandrasekharam, A. S. Tam-<lb/>hane, and L. C. Gupta, Semicond. Sci. Technol. 5, 561 ͑1992͒. <lb/>17 R. D. Barnard, Thermoelectricity in Metals and Alloys ͑Taylor and Fran-<lb/>cis, London, 1972͒. <lb/>18 H. Ishii, H. Sato, N. Kanazawa, H. Takagi, S. Uchida, K. Kitazawa, K. <lb/>Kishio, K. Fueki, and S. Tanaka, Physica B 148, 419 ͑1987͒. <lb/>19 O. G. Singh, C. S. Harendranath, O. Prakash, and B. D. Padalia, Solid <lb/>State Commun. ͑to be published͒. <lb/>20 F. J. Blatt, P. A. Schroeder, and C. L. Foiles, Thermoelectric Power of <lb/>Metals ͑Plenum, New York, 1976͒. <lb/>21 V. L. Ginzburg, J. Supercond. 2, 323 ͑1989͒. <lb/>22 A. Jezowski and P. W. Klamut, J. Less-Common Met. 169, L17 ͑1991͒. <lb/>23 K. Durczewski and M. Ausloos, Phys. Rev. B 53, 1762 ͑1996͒. <lb/>24 A. B. Kaiser, Phys. Rev. B 35, 4677 ͑1987͒. <lb/>25 H. Ma, G. Xiong, L. Wang, S. Wang, H. Zhang, L. Tong, S. Liang, and S. <lb/>Yan, Phys. Rev. B 40, 374 ͑1989͒. <lb/>26 V. Radhakrishnan, C. K. Subramaniam, V. Sankaranarayanan, G. V. <lb/>Subba Rao, and R. Srinivasan, Phys. Rev. B 40, 6850 ͑1989͒. <lb/>27 C. Uher and W.-N. Huang, Phys. Rev. B 40, 2694 ͑1989͒. <lb/>28 X.-Q. Hu, S. J. Hagen, W. Jiang, J. L. Peng, Z. Y. Li, and R. L. Greene, <lb/>Phys. Rev. B 45, 7356 ͑1992͒, and references therein. <lb/>Z. Z. Wang, T. R. Chien, N. P. Ong, J. M. Tarascon, and E. Wang, Phys. <lb/>Rev. B 43, 3020 ͑1991͒. <lb/>30 C. G. Olsen, R. Liu, D. W. Lynch, R. S. List, A. J. Arko, B. W. Veal, Y. <lb/>C. Chang, P. Z. Jiang, and A. P. Paulikas, Phys. Rev. B 42, 381 ͑1990͒; B. <lb/>O. Wells, Z.-X. Shen, D. S. Dessau, W. E. Spicer, C. G. Olsen, D. B. <lb/>Mitzi, A. Kapitulnik, R. S. List, and A. Arko, Phys. Rev. Lett. 65, 3056 <lb/>͑1994͒. <lb/></listBibl>
45
+
46
+ <page>5173 <lb/></page>
47
+
48
+ <note place="footnote">J. Appl. Phys., Vol. 80, No. 9, 1 November 1996 <lb/>Singh et al. <lb/></note>
49
+
50
+ <note place="footnote">Downloaded 26 Feb 2012 to 14.139.97.73. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions <lb/></note>
51
+
52
+ <listBibl>31 Y. Xin, W. Wong, C. X. Fan, Z. Z. Sheng, and F. T. Chan, Phys. Rev. <lb/>Lett. 48, 557 ͑1993͒. <lb/>32 O. G. Singh, Ph.D. thesis, Indian Institute of Technology, Bombay, 1994. <lb/>H. Hansel, F. Gollnik, A. Beck, and R. Gross, Physica C 235-240, 1367 <lb/>͑1994͒. <lb/>34 W. Jiang, S. N. Mao, X. X. Xi, X. Jiang, J. L. Peng, T. Venkatesan, C. J. <lb/>Lobb, and R. L. Greene, Phys. Rev. Lett. 73, 1291 ͑1994͒. <lb/>35 X. Jiang, W. Jiang, S. N. Mao, R. L. Greene, T. Venkatesan, and C. J. <lb/>Lobb, Physica C 254, 175 ͑1995͒. <lb/>36 C. Hohn, M. Gaffy, and A. Freimuth, Phys. Rev. B 50, 15875 ͑1994͒. <lb/>37 K. Durczewski and M. Ausloos, Z. Phys. B 85, 59 ͑1991͒. <lb/>38 Z.-X. Shen, W. E. Spicer, D. M. King, D. S. Dessau, and B. O. Wells, <lb/>Science 267, 343 ͑1995͒. <lb/>39 S. Yan, P. Lu, and Q. Li, Solid State Commun. 65, 355 ͑1988͒. <lb/></listBibl>
53
+
54
+ <page>5174 <lb/></page>
55
+
56
+ <note place="footnote">J. Appl. Phys., Vol. 80, No. 9, 1 November 1996 <lb/>Singh et al. <lb/></note>
57
+
58
+ <note place="footnote">Downloaded 26 Feb 2012 to 14.139.97.73. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions </note>
59
+
60
+
61
+ </text>
62
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JAP0897669-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <titlePage>University of Wollongong <lb/>Research Online <lb/>Faculty of Engineering -Papers (Archive) <lb/>Faculty of Engineering and Information Sciences <lb/>2001 <lb/>Superconductivity and flux pinning in Y and <lb/>heavily Pb codoped Bi-2212 single crystals <lb/>Xiaolin Wang <lb/>University of Wollongong, [email protected] <lb/>Hua-Kun Liu <lb/>University of Wollongong, [email protected] <lb/>S X. Dou <lb/>University of Wollongong, [email protected] <lb/>J. Horvat <lb/>University of Wollongong, [email protected] <lb/>D. Millikon <lb/>University of Wollongong <lb/>See next page for additional authors <lb/>http://ro.uow.edu.au/engpapers/96 <lb/>Research Online is the open access institutional repository for the University of Wollongong. For further information contact the UOW Library: <lb/>[email protected] <lb/>Publication Details <lb/>This article was originally published as: Wang, XL, Liu, HK, Dou, SX, Horvat, J, Millikon, D, Heine, G, Lang, W, Luo, HM &amp; Ding, SY, <lb/>Superconductivity and flux pinning in Y and heavily Pb codoped Bi-2212 single crystals, Journal of Applied Physics, 2001, 89(11), <lb/>7669-7671, and may be found here. Copyright 2001 American Institute of Physics. This article may be downloaded for personal use <lb/>only. Any other use requires prior permission of the author and the American Institute of Physics. <lb/> Authors <lb/>Xiaolin Wang, Hua-Kun Liu, S X. Dou, J. Horvat, D. Millikon, G. Heine, W. Lang, H. M. Luo, and Shichao <lb/>Ding <lb/>This journal article is available at Research Online: http://ro.uow.edu.au/engpapers/96 <lb/></titlePage>
9
+
10
+ <front>Superconductivity and flux pinning in Y and heavily Pb codoped Bi-2212 <lb/>single crystals <lb/>X. L. Wang, a) H. K. Liu, S. X. Dou, J. Horvat, and D. Millikon <lb/>Institute for Superconducting and Electronic Materials, University of Wollongong, NSW 2522, Australia <lb/>G. Heine and W. Lang <lb/>Institut fuer Materialphysik der Universitaet Wien and Ludwig Boltzman Institut fuer Festkoerperphysik, <lb/>Kopernikussgasse 15, A-1070 Wien, Austria <lb/>H. M. Luo and S. Y. Ding <lb/>National Laboratory of Solid State Microstructures and Department of Physics, Nanjing University, <lb/>Nanjing 210093, People&apos;s Republic of China <lb/> Studies of superconductivity and flux pinning were carried out on (Bi 1.64 Pb 0.36 ͒Sr 2 Ca 1Ϫx Y x Cu 2 O 8ϩy <lb/>͑xϭ0, 0.05, 0.11, 0.33͒ single crystals grown by the self-flux method. X-ray diffraction, transport, <lb/>and magnetic measurements were performed for purposes of characterization. X-ray analysis <lb/>revealed that the c lattice parameter systemically decreases as the Y doping level increases. The <lb/>superconducting transition temperature T c decreases from 80 to 30 K as x increases. A strong <lb/>annealing effect on T c and superconducting volume has been observed. Resistance measurements <lb/>show that xϭ0.33 samples are semiconductive over a wide temperature range between 4.2 and 300 <lb/>K for the as-grown state, but become metallic with T c of 65-70 K after air or oxygen annealing. <lb/>Flux pinning was studied by measuring the hysteresis loop at different temperatures and different <lb/>fields. A peak effect was observed in all the codoped samples. Results show that at low <lb/>temperatures, the peak field is smaller than in solely Pb doped crystals and decreases as x increases <lb/>(xϾ0.1). However, the peak field at high temperature for the xϭ0.05 sample is higher than in <lb/>heavily Pb doped Bi2212 crystals, indicative of a strong pinning due to the codoping. © 2001 <lb/>American Institute of Physics. ͓DOI: 10.1063/1.1356055͔<lb/></front>
11
+
12
+ <body>Due to the poor performance of Bi-Sr-Ca-Cu-O high <lb/>temperature superconducting materials under magnetic <lb/>fields, an improvement in the flux pinning capability through <lb/>effective doping is highly desirable. Taking into account the <lb/>Josephson coupling between CuO layers, Kim et al. 1 sug-<lb/>gested that the irreversibility field H irr is inversely propor-<lb/>tional to c , the resistivity along the c axis, and d s , the <lb/>distance between adjacent CuO planes, i.e., H irr ϳ1/( c <lb/>ϫd s ). This implies that reducing the c lattice parameter or <lb/>increasing the c-axis conductivity could enhance the flux <lb/>pinning of a material. It has been well established that all the <lb/>rare earth ͑RE͒ elements can substitute into the Ca site, 2 and <lb/>all 3d metal ions can substitute into the Cu site. 3 However, <lb/>most of the doped samples to date have been polycrystalline <lb/>bulk samples, and the investigators using RE dopants were <lb/>primary concerned with T c , the metal-insulator transition <lb/>and changes of normal state properties. 2,4,5 A very successful <lb/>enhancement of the intrinsic pinning, improving J c by 2 or-<lb/>ders of magnitude over the pure crystals, has already been <lb/>achieved by Pb doping into the Bi site in the BiO layer in <lb/>Bi2212 crystals. 6 A further enhanced flux pinning has been <lb/>reported in Pb and Cr ͑in the Cu site͒ codoped Bi-2212 <lb/>crystals. 7 No flux pinning investigations have been reported <lb/>in RE element doped Bi2212 crystals. <lb/>Y 3ϩ is the only ion among all the RE elements that is <lb/>smaller than Ca 2ϩ . It has been reported that Y doping causes <lb/>a reduction in the c lattice parameter. 8 According to the Kim <lb/>model, 1 this reduction is likely to be associated with an in-<lb/>crease in flux pinning. Unfortunately, no flux pinning inves-<lb/>tigations have been done on RE-doped Bi2212, except for Y <lb/>doped crystals which were reported to have strong flux pin-<lb/>ning at a 20% doping level. 9 These results have never been <lb/>replicated, and recent research on Y-doped polycrystalline <lb/>Bi2212 has indicated that Y doping does not improve pin-<lb/>ning, despite Y causing a reduction of the c lattice <lb/>parameter. 8 Since Pb doping makes the BiO 2 layer more con-<lb/>ductive and reduces the c lattice parameter, 10 Pb codoping <lb/>into the Bi site in the BiO 2 layer will be introduced into <lb/>Y-doped Bi2212 in order to further reduce the c lattice pa-<lb/>rameter and improve or increase the c-axis conductivity. It is <lb/>expected that Pb doping will reduce any increase in resistiv-<lb/>ity along the c axis that might be due to Y dopants. Here, we <lb/>show a detailed study of superconductivity and flux pinning <lb/>in both Y and heavily Pb codoped Bi-2212 crystals. <lb/>The Pb and Y codoped Bi2212 crystals used for the ex-<lb/>periment were grown using a self-flux method. High purity <lb/>Bi 2 O 3 , PbO, SrCO 3 , CaCO 3 Y 2 O 3 , and CuO were well <lb/>mixed according to the ratio Bi:Pb:Sr:Ca:Y:Cu <lb/>ϭ1.5:0.5:2:1Ϫx:x:2 (xр0.4) and put into Al 2 O 3 crucibles. <lb/>The crystal growth was carried out in a horizontal furnace <lb/>with a large temperature gradient. The sample was first <lb/>heated up to 1000°C and held there for about 2-4 h, then <lb/>fast cooled down to 950°C ͑200°C/h͒, then slowly cooled <lb/>down to 830°C at a rate of 5-20°C/h, and finally furnace <lb/>cooled down to room temperature. The real atomic compo-<lb/>sitions of the resulting crystals were determined by energy <lb/>dispersive analysis ͑EDA͒. Structure and lattice parameters <lb/>a͒ <lb/>Author to whom correspondence should be addressed; electronic mail: <lb/>[email protected] <lb/></body>
13
+
14
+ <note place="headnote">JOURNAL OF APPLIED PHYSICS <lb/>VOLUME 89, NUMBER 11 <lb/>1 JUNE 2001 <lb/></note>
15
+
16
+ <page>7669 <lb/></page>
17
+
18
+ <note place="footnote">0021-8979/2001/89(11)/7669/3/$18.00 <lb/>© 2001 American Institute of Physics <lb/>Downloaded 19 Jun 2006 to 130.130.37.6. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp <lb/></note>
19
+
20
+ <body>were determined using x-ray diffraction ͑XRD͒. The as-<lb/>grown crystals were annealed at different temperatures in air, <lb/>argon, and oxygen atmospheres. Superconductivity of the <lb/>crystals was characterized using standard four-probe trans-<lb/>port measurements and ac susceptibility. Flux pinning prop-<lb/>erties were investigated by the Physical Property Measure-<lb/>ment System, Quantum Design. <lb/>The single crystals obtained have dimensions of 1ϫ1-<lb/>4ϫ3 mm 2 in the ab plane depending on the amplitude of the <lb/>temperature gradient. The maximum sizes of the PbϩY <lb/>codoped crystals are bigger than those grown by the floating <lb/>traveling solvent zone technique. 6 The real atomic ratios <lb/>Bi:Pbϭ1.64:0.36 and Ca:Yϭ1.95:0.05, 0.89:0.11, 0.66:0.33 <lb/>were determined by EDA for three codoped samples used in <lb/>this work. XRD measurements showed that only ͑001͒ peaks <lb/>can be observed and no extra peaks from secondary phases <lb/>can be found, as shown in Fig. 1. X-ray analysis also re-<lb/>vealed that the ͑008͒ peak shifts to high angles with increas-<lb/>ing Y content ͑Fig. 2͒. This indicates that the c lattice pa-<lb/>rameters systemically decrease as the Y doping level <lb/>increases in agreement with the fact that the size of Y 3ϩ is <lb/>smaller than Ca 2ϩ . <lb/>The transition temperatures T c of the codoped crystals <lb/>are strongly dependent upon the Y doping level. T c deter-<lb/>mined by ac susceptibility decreases from 76 to 25 K as x <lb/>increases from 0.05 to 0.3 as shown in Fig. 3. This variation <lb/>in T c with increasing Y is similar to what is seen in solely Y <lb/>doped Bi2212 polycrystalline samples. 8 A strong annealing <lb/>effect on the T c was observed for xϭ0.11 and 0.33 samples <lb/>as shown in Fig. 3. It is clearly seen that the T c for the <lb/>as-grown crystal with xϭ0.11 shifts about 20 K higher after <lb/>air annealing, indicative of an oxygen underdoped state in <lb/>the as-grown crystals. It should be noted that the ac suscep-<lb/>tibility for the as-grown crystals with xϭ0.33 is very small, <lb/>but becomes larger after annealing in air. This implies that <lb/>the superconducting volume in this sample is enhanced after <lb/>annealing. Although the temperature where the major super-<lb/>conductivity occurs in the sample after air annealing is as <lb/>low as 25 K, the onset is as high as 60 K as shown in the <lb/>inset. Resistance measurements also show that the xϭ0.33 <lb/>sample is semiconductive over a wide temperature range be-<lb/>tween 4.2 and 300 K for the as-grown state, but becomes <lb/>metallic with a T c onset of 73 K after oxygen annealing, with <lb/>a T c0 of 30 K ͑Fig. 4͒, implying a small superconducting <lb/>volume above 30 K in agreement with ac susceptibility re-<lb/>sults. <lb/>The flux pinning in the Pb and Y codoped crystals was <lb/>investigated by measuring M -H over a wide temperature <lb/>range from 15 to 70 K. A peak effect can be observed in the <lb/>codoped samples. However, at low temperatures, the peak <lb/>field is smaller than in solely Pb doped crystals and de-<lb/>creases as x increases (xϾ0.05). The peak effect for x <lb/>ϭ0.05 starts at 20 K and can persist up to 70 K with the peak <lb/>fields around 2300 Oe at 20 K, close to what is seen in <lb/>heavily Pb doped Bi-2212 crystals. Figure 5 shows a series <lb/>of M -H loops measured at different temperatures for the x <lb/>ϭ0.05 sample. This indicated that the codoped crystals have <lb/>stronger flux pinning than in the solely Y doped Bi2212 <lb/>samples. 8 <lb/>Figure 6 shows peak fields as a function of T/T c for the <lb/>xϭ0.05 sample. As a comparison, data from a solely Pb <lb/>doped Bi2212 crystal with the same Pb content as in this <lb/>FIG. 1. XRD pattern of a Bi 1.64 Pb 0.36 Sr 2 Ca 0.66 Y 0.33 Cu 2 O 8ϩy single crystal. <lb/>FIG. 2. XRD pattern of ͑008͒ peaks for codoped crystals with different Y <lb/>content. <lb/>FIG. 3. Real part of ac susceptibility for different Y doped <lb/>Bi 1.64 Pb 0.36 Sr 2 Ca 1Ϫx Y x Cu 2 O 8ϩy crystals ͓xϭ0.33: as-grown ͑open squares͒, <lb/>air annealed ͑closed squares͒; xϭ0.11: as-grown ͑open circles͒, air annealed <lb/>͑closed circles͒; xϭ0.05: air annealed ͑solid line͔͒. <lb/></body>
21
+
22
+ <page>7670 <lb/></page>
23
+
24
+ <note place="headnote">J. Appl. Phys., Vol. 89, No. 11, 1 June 2001 <lb/>Wang et al. <lb/></note>
25
+
26
+ <note place="footnote">Downloaded 19 Jun 2006 to 130.130.37.6. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp <lb/></note>
27
+
28
+ <body>sample are also shown. It can be seen that the peak effect is <lb/>present over a wide temperature range 0.25рT/T c р0.85, <lb/>behaving the same as in heavily Pb doped Bi2212 crystals. It <lb/>should be noted that the peak field at high temperatures <lb/>above T/T c Ͼ0.5 for xϭ0.05 sample is higher than in solely <lb/>Pb doped crystals, indicative of a strong pinning due to the <lb/>codoping, even at relatively high fields. <lb/>Formation and decomposition of clusters of Bi 5ϩ and/or <lb/>Pb 4ϩ units in the Bi 3ϩ 2212 matrix, which can be controlled <lb/>by changing the oxygen or Pb content, have been proposed <lb/>as causes for the appearance of the peak effect in pure or Pb <lb/>doped Bi2212 crystals. 11 It is proposed that Bi 5ϩ and/or <lb/>Pb 4ϩ rich clusters of 2212 units, which are sensitive to the <lb/>annealing or oxygen content and distribution, exist in pure or <lb/>Pb doped Bi2212 crystals. On increasing oxygen or Pb con-<lb/>tent, the amount of Bi 5ϩ and/or Pb 4ϩ ͑which are smaller <lb/>than Bi 3ϩ ͒ increases, and the c-axis parameter decreases, re-<lb/>sulting in enhanced coupling along the ͑001͒ direction. This <lb/>is supported by the observed decrease of c and shift of H irr <lb/>for high quality Bi2212 crystals with different oxygen dop-<lb/>ing states ranging from overdoping and optimum to <lb/>underdoping. 11 The observed reduction in anisotropy in <lb/>heavily Pb doped Bi2212 crystals 10 also supports our sugges-<lb/>tion. By introducing Y into Ca sites, the c-lattice parameters <lb/>are further reduced as evidenced in Fig. 2. However, the <lb/>concentration of hole carriers is also depressed due to the <lb/>injection of electrons by Y 3ϩ , 12 which may increase the re-<lb/>sistivity and in turn reduce the flux pinning since H irr <lb/>ϳ1/( c ϫd s ). Therefore, there must be a competition be-<lb/>tween the carrier concentration and interlayer interaction that <lb/>are both induced by Y 3ϩ doping. It is most likely that the <lb/>contribution to flux pinning from the decrease in hole carrier <lb/>concentration caused by Y 3ϩ is dominant over the contribu-<lb/>tion from c-lattice parameter shrinking. However, it is pos-<lb/>sible that the smaller unit cells due to Y 3ϩ doping may be <lb/>more effective as pinning centers at high temperatures com-<lb/>pared with the situation in only heavily Pb doped Bi2212 <lb/>crystals. <lb/></body>
29
+
30
+ <listBibl>1 D. H. Kim, K. E. Gray, R. T. Kampwirth, J. C. Smith, D. S. Richeson, T. <lb/>J. Marks, J. H. Kang, J. Talvacchio, and M. Eddy, Physica C 177, 431 <lb/>͑1991͒. <lb/>2 Y. Gao, P. Pernambuco-Wise, J. E. Crow, J. O&apos;Reilly, N. Spencer, H. <lb/>Chen, and R. E. Salomon, Phys. Rev. B 45, 7436 ͑1992͒. <lb/>3 B. vom Hedt, W. Lisseck, K. Westerholt, and H. Bach, Phys. Rev. B 49, <lb/>9898 ͑1994͒. <lb/>4 P. P. Sumana, M. S. Rao, U. V. Ramachandra, G. V. Varadaraju, and G. <lb/>V. Subba Rao, Phys. Rev. B 50, 6929 ͑1994͒. <lb/>5 B. Beschoten, S. Sadewasser, G. Guntherodt, and C. Quitmann, Phys. <lb/>Rev. Lett. 77, 1837 ͑1996͒. <lb/>6 I. Chong, Z. Hiroi, M. Izumi, J. Shimoyama, Y. Nakayama, K. Kishio, T. <lb/>Terashima, Y. Bando, and M. Takano, Science 276, 770 ͑1997͒. <lb/>7 Y. P. Sun, W. H. Song, B. Zhao, J. J. Du, H. H. Wen, Z. X. Zhao, and H. <lb/>C. Ku, Appl. Phys. Lett. 76, 3795 ͑2000͒. <lb/>8 K. Kotitanta, T. Nakane, M. Karppinen, H. Yamauchi, and L. Niinistoe, J. <lb/>Low. Temp. Phys. ͑in press͒. <lb/>9 G. Villard, D. Pelloquin, A. Maignan, and A. Wahl, Appl. Phys. Lett. 69, <lb/>1480 ͑1996͒. <lb/>10 T. Motohashi, Y. Nakayama, T. Fujita, K. Kitazawa, J. Shinoyama, and K. <lb/>Kishio, Phys. Rev. B 59, 14080 ͑1999͒. <lb/>11 X. L. Wang, J. Horvat, H. K. Liu, S. X. Dou, G. Heine, and W. Lang, <lb/>Physica C 341-348, 651 ͑2000͒. <lb/>12 I. J. Hsu, R. S. Liu, J. M. Chen, R. G. Liu, L. Y. Jang, J. F. Lee, and K. <lb/>D. M. Harris, Chem. Mater. 12, 1115 ͑2000͒. <lb/></listBibl>
31
+
32
+ <body>FIG. 4. Resistance vs temperature for a Bi 1.64 Pb 0.36 Sr 2 Ca 0.66 Y 0.33 Cu 2 O 8ϩy <lb/>single crystal: ͑open squares͒ as-grown; ͑closed squares͒ annealed in O 2 for <lb/>24 h. <lb/>FIG. 5. M -H loops for xϭ0.05 crystal at different temperatures. <lb/>FIG. 6. Peak field vs T/T c for heavily Pb ͑closed circles͒ and both Pb and Y <lb/>codoped ͑open circles͒ crystals. <lb/></body>
33
+
34
+ <page>7671 <lb/></page>
35
+
36
+ <note place="headnote">J. Appl. Phys., Vol. 89, No. 11, 1 June 2001 <lb/>Wang et al. <lb/></note>
37
+
38
+ <note place="footnote">Downloaded 19 Jun 2006 to 130.130.37.6. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp </note>
39
+
40
+
41
+ </text>
42
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JCMM-23-1343.training.segmentation.tei.xml ADDED
@@ -0,0 +1,85 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>O R I G I N A L A R T I C L E <lb/>MiR-15a/16-1 deficiency induces IL-10-producing CD19 + <lb/>TIM-1 + cells in tumor microenvironment <lb/>Xiaoqin Jia 1,2 | Hao Liu 3 | Chong Xu 1 | Sen Han 1 | Yating Shen 1 | Xin Miao 1 | <lb/>Xiangyu Hu 1 | Zhijie Lin 1,2 | Li Qian 1,2 | Zhengbing Wang 3 | Weijuan Gong 1,2,4,5,6 <lb/>1 Department of Basic Medicine, Institute of <lb/>Translational Medicine, Medical College, <lb/>Yangzhou University,Yangzhou,P.R. China <lb/>2 Jiangsu Key Laboratory of Experimental &amp; <lb/>Translational Non-coding RNA Research, <lb/>Yangzhou, P.R. China <lb/>3 Department of General Surgery, Subei <lb/>People&apos;s Hospital of Jiangsu Province, <lb/>Yangzhou University, Yangzhou, P.R. China <lb/>4 Jiangsu Key Laboratory of Integrated <lb/>Traditional Chinese and Western Medicine <lb/>for Prevention and Treatment of Senile <lb/>Diseases, Yangzhou, P.R. China <lb/>5 Jiangsu Key Laboratory of Zoonosis, <lb/>Yangzhou, P.R. China <lb/>6 Jiangsu Co-innovation Center for <lb/>Prevention and Control of Important Animal <lb/>Infectious Diseases and Zoonoses, <lb/>Yangzhou, P.R. China <lb/>Correspondence <lb/>Weijuan Gong, Department of Basic <lb/>Medicine, Institute of Translational Medicine, <lb/>Medical College, Yangzhou University, <lb/>Yangzhou, P.R. China. <lb/>Email: [email protected] <lb/>and <lb/>Zhengbing Wang, Department of General <lb/>Surgery, Subei People&apos;s Hospital of Jiangsu <lb/>Province, Yangzhou University, Yangzhou, <lb/>P.R. China. <lb/>Email: [email protected] <lb/>Funding information <lb/>National Natural Science Foundation of <lb/>China, Grant/Award Number: 81273214, <lb/>81471547, 81671547, 81873867; <lb/>Innovation Project for Graduate Students in <lb/>Jiangsu Province, Grant/Award Number: <lb/>SJZZ16_0267; Natural Science Foundation <lb/>of Jiangsu Province, Grant/Award Number: <lb/>BK 20180925 <lb/>Abstract <lb/>IL-10-producing B cells (B10) are associated with autoimmune diseases, infection <lb/>and tumours. MiR-15a/16 as a tumour-suppressive gene is down-regulated in several <lb/>tumours, such as chronic lymphocytic leukaemia, pituitary adenomas and prostate <lb/>carcinoma. Here, increased frequency of IL-10-producing CD19 + Tim-1 + cells was <lb/>seen in both aged miR-15a/16 −/− mice (15-18 months) with the onset of B cell leu-<lb/>kaemia and young knockout mice (8-12 weeks) transplanted with hepatic cancer <lb/>cells. CD19 + Tim-1 + cells down-regulated the function of effector CD4 + CD25 low T <lb/>cells ex vivo dependent on IL-10 production, and adoptive transfer of CD19 + Tim-<lb/>1 + cells promoted tumour growth in mice. IL-10 production by CD19 + Tim-1 + cells <lb/>was involved with the STAT3 activation. Bioinformatics analysis shows that miR-16 <lb/>targets the 3′-untranslating region (3′-UTR) of STAT3 mRNA. Overexpression of <lb/>miR-16 in CD19 + Tim-1 + cells inhibited STAT3 transcription and its protein expres-<lb/>sion. Thus, the loss of miR-15a/16 promoted induction of regulatory CD19 + Tim-1 + <lb/>cells in tumour microenvironment. These results confirmed that miR-15a/16 could <lb/>be used in tumour therapy due to its inhibition of tumour and regulatory B cells. <lb/>K E Y W O R D S <lb/>CD19, IL-10, miR-15a/16, Tim-1, tumor <lb/>Xiaoqin Jia and Hao Liu contributed equally to this article. <lb/>----------------------------------------------------------------------------------------------------------------------------------------------------------------------<lb/>This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, <lb/>provided the original work is properly cited. <lb/>© 2018 The Authors. Journal of Cellular and Molecular Medicine published by John Wiley &amp; Sons Ltd and Foundation for Cellular and Molecular Medicine. <lb/>Received: 15 January 2018 | Revised: 30 August 2018 | Accepted: 26 October 2018 <lb/>DOI: 10.1111/jcmm.14037 <lb/>J Cell Mol Med. 2019;23:1343-1353. <lb/>wileyonlinelibrary.com/journal/jcmm | 1343 <lb/></front>
8
+
9
+ <body>1 | INTRODUCTION <lb/>The IL-10-secreting B cells (B10), as one of the important regulatory <lb/>B (Breg) cells, are associated with autoimmune diseases, infections <lb/>and tumours. B10 cells can directly inhibit macrophage activation, <lb/>down-regulate its secretion of NO and related inflammatory cytoki-<lb/>nes (TNF-α, IL-1, IL-6, IL-8, IL-12), and also reduce the phagocytosis <lb/>of macrophages. 1,2 Breg cells can inhibit antigen presentation and <lb/>expression of costimulatory molecules and cytokines of dendritic <lb/>cells (DCs). The Th1/Th17 effector cells are also regulated by Breg <lb/>cells to down-regulate the local inflammatory response. In addition, <lb/>Breg cells can also induce classical Treg and Tr1 cells to promote <lb/>immune-regulatory activity. 3,4 <lb/>So far, three subsets of IL-10-secreting B cells have been identi-<lb/>fied, namely, CD19 + CD5 + CD1d hi , 5,6 CD19 + Tim +7 and <lb/>CD19 + FcγIIb hi8 cells. After being stimulated with Tim-1 antibody, <lb/>45% of CD19 + CD5 + CD1d hi cells up-regulated Tim-1 expression and <lb/>secreted IL-10, whereas 8% of them could express Tim-1 and had <lb/>potential IL-10 secretion ability in non-CD5 + CD1d hi B cells, which <lb/>suggested that these two phenotypic Breg cells could not replace <lb/>each other. 7 A unique phenotype of CD19 hi FcγIIb hi B cells, which is <lb/>induced by regulatory DCs, also produces IL-10 to inhibit effector <lb/>CD4 + T cells. 8 The engagement of BCR is thought to play a role in <lb/>the early stage, whereas the activation of CD40 and Toll-like recep-<lb/>tors (TLRs) plays a role in the late stage of Breg cell differentiation. <lb/>Several signalling proteins, such as Erk, p38, glycogen synthase <lb/>kinase 3-β (GSK3-β), Smad, Signal transducers and activators of tran-<lb/>scription (STATs), mammalian target of rapamycin (mTOR) and <lb/>nuclear factor-kappa B (NF-κB), are involved in IL-10 production. 9-11 <lb/>The miR-15a/16 gene complex is located in the intron of the <lb/>DLEU2 gene in the human 13q14 region, which is recognized as a <lb/>tumour-suppressor gene. 12 The deletion of this gene region is closely <lb/>related not only with chronic B lymphocytic leukaemia, but also with <lb/>various solid tumours such as melanoma, colorectal cancer, prostate <lb/>cancer, breast cancer and bladder cancer. 13 The target genes of miR-<lb/>15a/16 include Bcl-2, WT-1, WNT3A, MCA1, MCL1 and CCDN1, <lb/>which are involved in tumour apoptosis. 14 MiR-15a/16 can also inhi-<lb/>bit vascular endothelial growth factor (VEGF) secretion and play <lb/>anti-angiogenesis activity. 15 Here, we showed that the deficiency of <lb/>miR-15a/16 was associated with induction of IL-10-producing Breg <lb/>cells under tumour microenvironment by up-regulating STAT3 <lb/>expression. <lb/>2 | MATERIALS AND METHODS <lb/>2.1 | Reagents and mice <lb/>MiR-15a/16 −/− mice (C57BL/6) were obtained from the Jackson Lab-<lb/>oratory (Sacramento, CA). All animal experiments were approved by <lb/>the Institutional Animal Care and Use Committee of Yangzhou <lb/>University (Yangzhou, China). The murine liver cancer cell line (H22) <lb/>was obtained from ATCC. Antibodies for flow cytometry were pur-<lb/>chased from either BioLegend (San Diego, CA) or eBioscience (San <lb/>Diego, CA): CD19 (1D3), B220 (RA3-6B2), CD5 (53-7.3), Tim-1 <lb/>(RMT1-4), FcγIIb (AT130-2), IL-10 (JES5-16E3), CD1d (1B1), CD4 <lb/>(RM4-5), CD69 (H1.2F3) and IFN-γ (XMG1.2). Antibodies for Wes-<lb/>tern blot were from Cell Signaling Technology (Boston, MA): STAT3 <lb/>(4904P), STAT3-pY705 (9145S) and STAT3-pS727 (94994T). <lb/>2.2 | ELISA <lb/>Splenic B cells (5 × 10 5 ) isolated either from miR-15a/16 −/− or wild-<lb/>type (WT) mice (C57BL/6) by magnetic sorting beads (Miltenyi) were <lb/>cultured in complete RPMI 1640 medium. For 24 or 48 hours, cell <lb/>supernatants were collected, and IL-10 concentrations were mea-<lb/>sured according to the manufacturer&apos;s protocol (BioLegend). Serum <lb/>IL-10 levels of miR-15a/16 knock-out (KO) or wild-type (WT) mice <lb/>were also determined by the same kit. <lb/>2.3 | Flow cytometric intracellular staining <lb/>Intracellular cytokine production was determined using a staining kit <lb/>(eBioscience). For IL-10 detection, B lymphocytes were cultured in <lb/>the presence of Brefeldin A (10 μg/mL) for 4 hours at 37°C. For <lb/>IFN-γ measurement, lymphocytes were stimulated with PMA (50 ng/ <lb/>mL)/ionomycin (5 μg/mL) in the presence of Brefeldin A for 4 hours <lb/>at 37°C. After lymphocytes were stained with surface markers, they <lb/>were fixed, permeabilized, stained with cytokine or isotype antibody <lb/>and analysed by flow cytometry. <lb/>2.4 | Coculture of B and T cells <lb/>CD19 + Tim-1 + cells were isolated from spleens of WT or KO mice which <lb/>were pre-transplanted with H22 cells by flow cytometry. CD4 + CD25 high <lb/>or CD4 + CD25 low cells from normal mice (C57BL/6) were also sorted out <lb/>by flow cytometry. Then, two lymphocyte subsets were mixed in a 1:1 <lb/>ratio and cultured overnight. The mixed lymphocytes were collected, <lb/>and CD69 expression and IFN-γ secretion were analysed in <lb/>CD4 + CD25 high or CD4 + CD25 low cells as described above. <lb/>2.5 | In vivo mouse tumour models <lb/>When H22 cells (2 × 10 6 ) were subcutaneously injected into mice <lb/>(n = 3), CD19 + Tim-1 + or CD19 + Tim-1 − cells (5 × 10 5 ) sorted form <lb/>KO mice were injected into the tail veins of tumour-bearing mice <lb/>every day. Tumour diameters were also documented every day. On <lb/>day 30, mice were sacrificed and tumour tissues were isolated. <lb/>2.6 | Western blot <lb/>Proteins of CD19 + Tim-1 + cells either from miR-15a/16 −/− or WT <lb/>mice were extracted after the lysis buffer (KeyGen, Nanjing, China) <lb/>was added into cells. After being separated on SDS-PAGE gels and <lb/>transferred onto polyvinylidene difluoride (PVDF) membranes, pro-<lb/>teins were stained with first and secondary antibodies sequentially. <lb/>The blotting signal was developed using an ECL kit (KeyGen) and <lb/>analysed with the Gel-Pro32 software. <lb/></body>
10
+
11
+ <page>1344 | <lb/></page>
12
+
13
+ <note place="headnote">JIA ET AL. <lb/></note>
14
+
15
+ <body>2.7 | Lentivirus infection <lb/>Lentivirus-expressing miR-16-1 (LV-miR-16) and a control lentivirus <lb/>(LV-control) were provided by GeneChem (Shanghai, China). <lb/>The transfection procedure was the same as in the previous <lb/>study. 16 <lb/>2.8 | Real-time polymerase chain reaction (PCR) <lb/>As CD19 + Tim-1 + cells from miR-15a/16 −/− mice were transfected <lb/>with the LV-miR-16 lentivirus for 72 hours, RNA was extracted by <lb/>the TRIzol reagent (Life Technologies; Carlsbad, CA), and cDNA was <lb/>generated by a QuantiTect ® reverse transcription kit (QIAGEN <lb/>GmbH; Hilden, Germany). The amplification of cDNA was conducted <lb/>using the QuantiNova ™ SYBR ® Green PCR Kit (QIAGEN) on ABI <lb/>7500 (PE Applied Biosystems, Carlsbad, CA, USA). Primer pairs for <lb/>STAT3 were 5′-CACCCAACAGCCGCCGTAGT and 5′-TCCATTCA-<lb/>GATCCTGCATGTCTCC. <lb/>2.9 | Statistical analysis <lb/>Differences between the two groups were analysed by Student&apos;s t <lb/>test. Data were evaluated by one-way ANOVA followed by Dun-<lb/>nett&apos;s test between control and various stimulation groups. Signifi-<lb/>cant differences were indicated when *P &lt; 0.05, **P &lt; 0.01, and <lb/>***P &lt; 0.001. <lb/>3 | RESULTS <lb/>3.1 | Increased IL-10-producing B cells in the aged <lb/>knockout (KO) mice with leukaemia <lb/>The miR-15a/16 −/− mice spontaneously develop B cell leukaemia at <lb/>the age of 15-18 months with a penetrance of 60%. 17 First, B cell <lb/>frequencies in spleens of KO or WT mice in the age of 15-<lb/>18 months or 8-12 weeks were analysed. The aged KO mice (15-<lb/>18 months) were verified to have B cell leukaemia as shown in the <lb/>Figure S1. As expected, B cell frequency and absolute number are <lb/>significantly enhanced in the aged KO mice (15-18 months) (Fig-<lb/>ure 1A,B). No changes of B cells were observed in young KO mice <lb/>(8-12 weeks), compared with WT mice. B cells from spleens of KO <lb/>or WT mice at different ages were isolated and cultured ex vivo for <lb/>24 or 48 hours. In aged mice, IL-10 concentrations of supernatants <lb/>from KO mice-derived B cells were significantly higher than those <lb/>from WT mice-derived B cells. We did not observe any significant <lb/>differences of IL-10 concentration of B cell supernatants from young <lb/>mice (Figure 1C). IL-10 production by B cells from aged KO mice <lb/>was confirmed by intracellular staining of flow cytometry, as shown <lb/>in Figure 1D. Multi-colour fluorescence labelling was used to analyse <lb/>surface markers of CD19 + IL-10 + cells. KO mice-derived CD19 + IL-<lb/>10 + cells displayed the increased expression of Tim-1 and the similar <lb/>expression of CD1d and FcγIIb, as compared with WT mice (Fig-<lb/>ure 1E). Serum IL-10 level of aged KO mice was also higher than <lb/>that of aged WT mice and young KO mice (Figure 1F). Thus, there <lb/>was a B cell population with IL-10-producing activity in the aged KO <lb/>mice. <lb/>3.2 | Increased CD19 + Tim-1 + cells in the aged KO <lb/>mice with leukaemia <lb/>Three Breg cell populations with IL-10 production (CD19 + Tim-1 + , <lb/>CD19 + CD5 + CD1d hi , CD19 + FcγIIb hi ) were detected in both aged and <lb/>young KO mice. In the aged KO mice (15-18 months), splenic <lb/>CD19 + Tim-1 + cell frequency was significantly increased compared <lb/>with WT mice at the same age. No significant changes of splenic <lb/>CD19 + Tim-1 + cell frequency were observed in young mice (Fig-<lb/>ure 2A). These CD19 + Tim-1 + cells were able to secrete IL-10 as <lb/>detected by flow cytometry (Figure 2B). The frequency of CD19 + <lb/>Tim-1 + cells was positively correlated with serum IL-10 level (Fig-<lb/>ure 2C). However, the CD19 + CD5 + CD1d hi cell frequency did not <lb/>vary obviously in aged or young KO mice (Figure 2D). CD19 + <lb/>FcγIIb hi cells also showed no significant changes between KO and <lb/>WT mice at different ages (Figure 2E). Therefore, CD19 + Tim-1 + <lb/>cells with IL-10 production were induced in aged KO mice bearing B <lb/>cell leukaemia. <lb/>3.3 | Increased CD19 + Tim-1 + cells in young KO <lb/>mice pre-transplanted with tumour cells <lb/>Considering IL-10-secreting B cells and leukaemic B cells in the aged <lb/>B cell-leukaemia bearing miR-15a/16 −/− mice (15-18 months) could <lb/>not be discriminated clearly, we determined whether CD19 + Tim-1 + <lb/>cells could be induced in young KO mice (8-12 weeks) bearing with <lb/>solid tumours. Hepatic cancer cells (H22) were subcutaneously trans-<lb/>planted into back of KO or WT mice. Tumour growth was up-regu-<lb/>lated in KO mice indicated by tumour size (Figure 3A). After 14 days <lb/>of H22 transplantation, Breg cells of KO or WT mice were analysed. <lb/>Compared with that in tumour-bearing WT mice, the splenic CD19 + <lb/>Tim-1 + cell frequency was significantly increased in tumour-bearing <lb/>KO mice (Figure 3B,C). Frequency of CD19 + IL-10 + Tim-1 + cells was <lb/>also increased in KO mice (Figure 3D). Similarly, no significant <lb/>changes of splenic CD19 + CD5 + CD1d hi and CD19 + FcγIIb hi cells were <lb/>observed in the two groups of mice (Figure 3E). The representative <lb/>results of CD19 + CD5 + CD1d hi and CD19 + FcγIIb hi cells detected by <lb/>flow cytometry were shown in Figure S2. Simultaneously, CD19 + <lb/>Tim-1 + cell frequencies in tumour tissues (Figure 3F) and peripheral <lb/>blood (Figure 3G) were significantly increased in KO mice. Because <lb/>sufficient CD19 + CD5 + cells infiltrated in tumour tissues can&apos;t be <lb/>acquired for analysis of distinct CD1d expression levels, we only <lb/>quantified the frequency of CD19 + CD5 + cells in tumour tissues and <lb/>peripheral blood (Figure S3). As shown in Figures 3F,G, there were <lb/>no significant differences of CD19 + CD5 + and CD19 + FcγIIb hi cell <lb/>frequencies in tumour tissues and peripheral blood between two <lb/>strains of mice. Therefore, CD19 + Tim-1 + cells with IL-10 production <lb/>could be also induced in young KO mice transplanted with tumour <lb/>cells. <lb/></body>
16
+
17
+ <note place="headnote">JIA ET AL. <lb/>| </note>
18
+
19
+ <page>1345 <lb/></page>
20
+
21
+ <body>F I G U R E 1 Enhanced IL-10-producing B cells in the aged (15-18 months) KO mice. (A, B) Detection of CD19 + B220 + cells of spleens from <lb/>both aged and young (8-12 weeks) mice by flow cytometry. (C) After CD19 + cells were isolated by magnetic beads and cultured ex vivo, cell <lb/>supernatants were collected, and IL-10 concentrations were measured with an ELISA kit. (D) Intracellular IL-10 of B cells from the aged mice <lb/>was detected by flow cytometry. (E) Expression of Tim-1, CD1d and FcγIIb on CD19 + IL-10 + cells detected by flow cytometry after multicolour <lb/>fluorescence labelling. (F) Serum IL-10 levels of aged and young KO or WT mice. **P &lt; 0.01; ***P &lt; 0.001; ns, no significance. <lb/></body>
22
+
23
+ <page>1346 | <lb/></page>
24
+
25
+ <note place="headnote">JIA ET AL. <lb/></note>
26
+
27
+ <body>F I G U R E 2 Analysis of regulatory B cell subsets in the aged (15-18 months) KO mice. CD19 + Tim-1 + (A), CD19 + CD5 + CD1d hi (D), and <lb/>CD19 + FcγIIb hi (E) cell frequencies in spleens were detected by flow cytometry. (B) IL-10 production by CD19 + Tim-1 + were determined by <lb/>intracellular staining of flow cytometry. (C) Correlation analysis of CD19 + Tim-1 + cell frequency with serum IL-10. **P &lt; 0.01; ns, no <lb/>significance <lb/></body>
28
+
29
+ <note place="headnote">JIA ET AL. <lb/>| </note>
30
+
31
+ <page>1347 <lb/></page>
32
+
33
+ <body>F I G U R E 3 Analysis of Breg cell subsets in young (8-12 weeks) KO mice bearing with transplanted tumours. (A) Tumour growth curve in KO <lb/>and WT mice after H-22 cells (5 × 10 6 ) were subcutaneously injected. (B, C) CD19 + Tim-1 + cell frequencies in spleens were detected by flow <lb/>cytometry. (D) Il-10 + Tim-1 + cells frequency gated on CD19 + cell was determined by intracellular staining of flow cytometry. (E) Splenic <lb/>CD19 + CD5 + CD1d hi and CD19 + FcγIIb hi cell frequencies of tumour-bearing mice. CD19 + Tim-1 + , CD19 + CD5 + and CD19 + FcγIIb hi cell <lb/>frequencies of tumour tissues (F) and peripheral blood (G). *P &lt; 0.05; ns, no significance <lb/></body>
34
+
35
+ <page>1348 | <lb/></page>
36
+
37
+ <note place="headnote">JIA ET AL. <lb/></note>
38
+
39
+ <body>F I G U R E 4 Immune-regulatory function mediated by CD19 + Tim-1 + cells via IL-10 production. (A, C) Variations of CD69 expression of IL-2-<lb/>stimulated CD4 + T cells from healthy C57BL/6 mice after in coculture with CD19 + Tim-1 + cells at a ratio of 1:1 (with/without IL-10 <lb/>neutralizing antibody, 10 μg/mL). (B, C) IFN-γ production by CD4 + T cells was detected by intracellular staining of flow cytometry after being <lb/>incubated with CD19 + Tim-1 + cells. (D) Growth curve of transplanted H22 cells after adoptive transfer of CD19 + Tim-1 + or CD19 + Tim-1 − <lb/>cells. (E) Stripped tumour tissues from mice. Each experiment was repeated at least thrice. *P &lt; 0.05; **P &lt; 0.01; ***P &lt; 0.001 <lb/></body>
40
+
41
+ <note place="headnote">JIA ET AL. <lb/>| </note>
42
+
43
+ <page>1349 <lb/></page>
44
+
45
+ <body>3.4 | Regulatory activity of CD19 + Tim-1 + cells <lb/>depends on IL-10 production <lb/>CD4 + CD25 high cells are generally regarded as regulatory T cells, and <lb/>CD4 + CD25 low cells as effector T cells. When CD19 + Tim-1 + cells <lb/>were incubated with IL-2-stimulating CD4 + cells, CD69 expression <lb/>was significantly down-regulated on CD4 + CD25 low cells, but no <lb/>changes on CD4 + CD25 high cells (Figure 4A,C). Moreover, IFN-γ <lb/>production was sharply inhibited in CD4 + CD25 low cells after these <lb/>T cells were cocultured with CD19 + Tim-1 + cells (Figure 4B,C). The <lb/>suppressive activity of Tim-1 + B-cells from KO mice was also higher <lb/>than that of WT mice (Figure 4B,C). Thus, CD19 + Tim-1 + cells <lb/>exerted immune-regulatory effects through inhibiting activity of <lb/>effector T cells. When IL-10 neutralizing antibodies were added into <lb/>the B-T cell coculture system, the inhibitory effect of CD69 and <lb/>IFN-γ expression was almost reversed (Figure 4A-C), demonstrating <lb/>that Breg cells mediated the regulatory effect via IL-10 secretion. <lb/>Next, CD19 + Tim-1 + or CD19 + Tim-1 − cells from KO mice were <lb/>adoptively transferred into mice pre-injected with H22 cells every <lb/>3 days and lasted for 30 days. As expected, the transfer of CD19 + <lb/>Tim-1 + cells significantly promoted tumour growth in vivo <lb/>(Figure 4D,E). <lb/>3.5 | STAT3 contributes to IL-10 production of <lb/>CD19 + Tim-1 + cells <lb/>STAT3 activation is involved in IL-10 production of immune cells. <lb/>STAT3 levels of CD19 + Tim-1 + cells were checked in tumour-bear-<lb/>ing KO and WT mice. STAT3 mRNA levels were higher in CD19 + <lb/>Tim-1 + cells of KO mice than those of WT mice (Figure 5A). Total <lb/>STAT3, STAT3-pY705 and STAT3-pS727 were all increased in <lb/>CD19 + Tim-1 + cells of tumour-bearing KO mice (Figure 5B,C). <lb/>When the STAT3 inhibitor (Stattic) was treated with KO mice-<lb/>derived CD19 + Tim-1 + cells, IL-10 production and Tim-1 expression <lb/>are both blocked (Figure 5D,E). The result indicated that the <lb/>increased STAT3 activity contributed to IL-10 production by <lb/>CD19 + Tim-1 + cells. <lb/>3.6 | MiR-16 over-expression in CD19 + Tim-1 + cells <lb/>inhibits STAT3 <lb/>Next, we determined whether miR-15a/16 directly down-regulated <lb/>the STAT3 mRNA level. By using bioinformatics analysis, miR-16, not <lb/>miR-15a, was found to directly bind with the 3′-untranslated region <lb/>(UTR) of STAT3 mRNA (Figure 6A). After KO mice-derived CD19 + <lb/>Tim-1 + cells were transfected with miR-16-containing lentivirus, miR-<lb/>16 expression levels were verified and shown in Figure 6B. At the <lb/>same time, the transcription of STAT3 was significantly inhibited <lb/>(Figure 6C). Western blot results also confirmed that STAT3 was <lb/>down-regulated in CD19 + Tim-1 + cells by being transfected with the <lb/>miR-16-containing lentivirus (Figure 6D). Therefore, overexpression <lb/>of miR-16 in KO mice-derived CD19 + Tim-1 + cells suppressed <lb/>STAT3 expression. <lb/>4 | DISCUSSION <lb/>The microRNA cluster of miR-15a/miR-16-1 (miR-15a/16) located <lb/>at 13q14.2 of the chromosome is regarded as a tumour-suppres-<lb/>sive gene. The loss of the gene cluster is involved in the develop-<lb/>ment of cancers such as chronic lymphocytic leukaemia (CLL), <lb/>pituitary adenomas, and prostate carcinoma. 13,14 Here, we deter-<lb/>mined that the loss of miR-15a/16 was associated with induction <lb/>of IL-10-producing B cells under tumour microenvironment. Breg <lb/>cells could down-regulate biologic activities of effector CD4 + T <lb/>cells and promote tumour growth with the characteristic pheno-<lb/>type of CD19 + Tim + . IL-10 production by B cells was dependent <lb/>on the STAT3 activation, and overexpression of miR-16 resulted <lb/>in inhibition of STAT3 and suppression of CD19 + Tim-1 + cells. We <lb/>described that Tim-1 + Breg cells with immune-suppressive activity <lb/>for tumour evasion was involved with the loss of the miR-15a/16 <lb/>gene cluster. <lb/>In humans, malignant cells from 90% of patients with CLL are <lb/>able to produce IL-10. These IL-10-competent CLL cells also express <lb/>cell surface phenotypes similar to nonmalignant B10 cells, indicating <lb/>a functional relationship between CLL and B10 cells. 18 We could not <lb/>discriminate B10 cells and leukemic B cells in the aged miR-15a/16 −/ <lb/>− mice (15-18 months) bearing B cell leukaemia. When the young <lb/>miR-15a/16 −/− mice (8-12 week) were transplanted with hepatic can-<lb/>cer cells, IL-10-producing CD19 + Tim-1 + cells were significantly <lb/>increased. In this case, as there is no B cell leukaemia in mice, IL-10-<lb/>producing CD19 + Tim-1 + cells were induced instead. No obvious <lb/>changes in IL-10-producing CD19 + Tim-1 + cells were observed in <lb/>healthy young miR-15a/16 −/− mice, suggesting that this Breg subset <lb/>was only induced in tumour microenvironment. Breg cells could be <lb/>recruited to the tumour and thereby attenuate anti-tumour immune <lb/>responses. <lb/>To date, the detailed molecular mechanisms of Breg development <lb/>in the tumour microenvironment remain unknown. Some tumour <lb/>cell-derived factors, such as leukotriene B4, 19 TNF-α, 20 placental <lb/>growth factor 21 and IL-21 secreted by local T cells, 22 have been rec-<lb/>ognized as Breg-induced factors. In addition, cell membrane mole-<lb/>cules (CD40L 23 or PD-1 24 ) of tumours are involved in Breg <lb/>development. The differentiation of Breg cells mainly depends on <lb/>the engagement of BCR and CD40. 25,26 In human B cells, STAT3 <lb/>and Erk activation induced by TLR controls IL-10 expression. 27 The <lb/>inhibition of STAT3 blocked IL-10 expression by CD19 + Tim-1 + cells, <lb/>suggesting that using STAT3 inhibitors in tumour patients also <lb/>retards B10 cell development. <lb/>Ectopic expression of miR-15a and miR-16-1 has been shown to <lb/>up-regulate 265 genes and down-regulate 3307 genes. 28 We found <lb/>that eight nucleotides of miR-16 are complementary to bases 295-<lb/>324 at the 3′-end of the STAT3 cDNA. The overexpression of miR-<lb/>16 led to the down-regulation of STAT3 mRNA and protein levels in <lb/>CD19 + Tim-1 + cells from KO mice. Whether miR-16 regulates <lb/>STAT3 expression through direct binding of its 3′-UTR needs further <lb/>study. It could be inferred that the STAT3 expression regulated by <lb/>miR-16 was not as strong as that by other miR-16 prominent target <lb/></body>
46
+
47
+ <page>1350 | <lb/></page>
48
+
49
+ <note place="headnote">JIA ET AL. <lb/></note>
50
+
51
+ <body>F I G U R E 5 STAT3 activity contributes to IL-10 production by CD19 + Tim-1 + cells. (A) STAT3 mRNAs of CD19 + Tim-1 + cells from both <lb/>mice were measured by reverse transcription and real-time PCR. (B) Total STAT3, STAT3-pY705, and STAT3-pS727 levels were determined by <lb/>Western blot. (C) Statistical analysis of STAT3 expression in CD19 + Tim-1 + cells from both mice. Variations of IL-10 (D) and Tim-1 (E) <lb/>expression in CD19 + Tim-1 + cells treated with the STAT3 inhibitor (Stattic). Each experiment was repeated at least thrice. *P &lt; 0.05; <lb/>***P &lt; 0.001 <lb/></body>
52
+
53
+ <note place="headnote">JIA ET AL. <lb/></note>
54
+
55
+ <page>| 1351 <lb/></page>
56
+
57
+ <body>genes (eg, BCL-2, MCL1, CCND1 and WNT3A). 13,14 In addition, <lb/>although we analysed the activity of miR-16 in B10 cell develop-<lb/>ment, the role of miR-15a (belonging to the same microRNA family) <lb/>could not be excluded. <lb/>In summary, deficiency of the microRNA cluster of miR-15a/16 <lb/>promoted the induction of IL-10-producing CD19 + Tim-1 + cells in <lb/>mice. The development of regulatory CD19 + Tim-1 + cells was <lb/>dependent on STAT3 activation. Overexpression of miR-16 inhibited <lb/>STAT3 expression. Considering that microRNAs target many genes, <lb/>this study confirmed that miR-15/16 could be used pharmaceutically <lb/>in tumour therapy. 29,30 <lb/></body>
58
+
59
+ <div type="acknowledgement">ACKNOWLEDGEMENTS <lb/>This work was supported by the National Natural Science Founda-<lb/>tion (No. 81671547, No. 81471547, No. 81273214, No. 81873867) <lb/>of China, Natural Science Foundation of Jiangsu Province (BK <lb/>20180925), Innovation Project for Graduate Students in Jiangsu Pro-<lb/>vince (SJZZ16_0267), the &quot;Six peaks&quot; Talent Project, and the &quot;333&quot; <lb/>Talent Project in Jiangsu Province. <lb/></div>
60
+
61
+ <div type="annex">CONF LICT OF I NTEREST <lb/>All authors have declared there are no financial conflicts of interest <lb/>with regard to this work. <lb/></div>
62
+
63
+ <front>O R C I D <lb/>Weijuan Gong <lb/>http://orcid.org/0000-0002-8543-1314 <lb/></front>
64
+
65
+ <listBibl>R E F E R E N C E S <lb/>1. Sarvaria A, Madrigal JA, Saudemont A. B cell regulation in cancer <lb/>and anti-tumor immunity. Cell Mol Immunol. 2017;14:662-674. <lb/>2. Mauri C, Menon M. Human regulatory B cells in health and disease: <lb/>therapeutic potential. J Clin Invest. 2017;127:772-779. <lb/>3. Gorosito Serrán M, Fiocca Vernengo F, Beccaria CG, et al. The <lb/>regulatory role of B cells in autoimmunity, infections and cancer: <lb/>perspectives beyond IL10 production. FEBS Lett. 2015;589:3362-<lb/>3369. <lb/>4. Lykken JM, Candando KM, Tedder TF. Regulatory B10 cell develop-<lb/>ment and function. Int Immunol. 2015;27:471-477. <lb/>5. Yanaba K, Bouaziz J-D, Haas KM, et al. A regulatory B cell subset <lb/>with a unique CD1d hi CD5 + phenotype controls T cell dependent <lb/>inflammatory responses. Immunity. 2008;28:639-650. <lb/>6. Xing C, Ma N, Xiao H, et al. Critical role for thymic <lb/>CD19 ± CD5 ± CD1d hi IL-10 ± regulatory B cells in immune homeostasis. <lb/>J Leukoc Biol. 2015;97:547-556. <lb/>7. Ding Q, Yeung M, Camirand G, et al. Regulatory B cells are identified <lb/>by expression of TIM-1 and can be induced through TIM-1 ligation <lb/>to promote tolerance in mice. J Clin Invest. 2011;121:3645-3656. <lb/>8. Qian L, Qian C, Chen Y, et al. Regulatory dendritic cells program B <lb/>cells to differentiate into CD19 hi FcγIIb hi regulatory B cells through <lb/>IFN-β and CD40L. Blood. 2012;120:581-591. <lb/>9. Bankó Z, Pozsgay J, Szili D, et al. Induction and differentiation of IL-<lb/>10-producing regulatory B cells from healthy blood donors and <lb/>rheumatoid arthritis patients. J Immunol. 2017;198:1512-1520. <lb/>10. Menon M, Blair PA, Isenberg DA, et al. A regulatory feedback <lb/>between plasmacytoid dendritic cells and regulatory B cells is aber-<lb/>rant in systemic lupus erythematosus. Immunity. 2016;44:683-697. <lb/>11. Yanaba K, Bouaziz JD, Matsushita T, et al. The development and <lb/>function of regulatory B cells expressing IL-10 (B10 cells) requires <lb/>antigen receptor diversity and TLR signals. J Immunol. <lb/>2009;182:7459-7472. <lb/>12. Pekarsky Y, Croce CM. Role of miR-15/16 in CLL. Cell Death Differ. <lb/>2015;22:6-11. <lb/></listBibl>
66
+
67
+ <body>F I G U R E 6 Overexpression of miR-16 <lb/>down-regulates STAT3. (A) Bioinformatics <lb/>analysis of complementary sequences <lb/>between miR16 and 3′-UTR of STAT3 <lb/>mRNA. (B) miR-16 expression of <lb/>CD19 + Tim-1 + cells determined by RT-PCR. <lb/>(C) After CD19 + Tim-1 + cells were <lb/>transfected by the lentivirus containing <lb/>miR-16, STAT3 mRNA was detected by <lb/>RT-PCR. (D) Variations of STAT3 protein <lb/>levels in CD19 + Tim-1 + cells detected by <lb/>Western blot after the miR-16 lentivirus <lb/>transfection. Each experiment was <lb/>repeated at least thrice. *P &lt; 0.05; <lb/>***P &lt; 0.001 <lb/></body>
68
+
69
+ <page>1352 | <lb/></page>
70
+
71
+ <note place="headnote">JIA ET AL. <lb/></note>
72
+
73
+ <listBibl>13. Aqeilan RI, Calin GA, Croce CM. MiR-15a and miR-16-1 in cancer: <lb/>discovery, function and future perspectives. Cell Death Differ. <lb/>2010;17:215-220. <lb/>14. Cui J. MiR-16 family as potential diagnostic biomarkers for cancer: a <lb/>systematic review and meta-analysis. Int J Clin Exp Med. <lb/>2015;8:1703-1714. <lb/>15. Sun CY, She XM, Qin Y, et al. miR-15a and miR-16 affect the angio-<lb/>genesis of multiple myeloma by targeting VEGF. Carcinogenesis. <lb/>2013;34:426-435. <lb/>16. Jia X, Li X, Shen Y, et al. MiR-16 regulates mouse peritoneal macro-<lb/>phage polarization and affects T-cell activation. J Cell Mol Med. <lb/>2016;20:1898-1990. <lb/>17. Klein U, Lia M, Crespo M, et al. The DLEU2/miR-15a/16-1 cluster <lb/>controls B cell proliferation and its deletion leads to chronic lympho-<lb/>cytic leukemia. Cancer Cell. 2010;17:28-40. <lb/>18. Dilillo DJ, Weinberg JB, Yoshizaki A, et al. Chronic lymphocytic leu-<lb/>kemia and regulatory B cells share IL-10 competence and immuno-<lb/>suppressive function. Leukemia. 2012;27:170-182. <lb/>19. Bichi R, Shinton SA, Martin ES, et al. Human chronic lymphocytic <lb/>leukemia modeled in mouse by targeted TCL1 expression. Proc Natl <lb/>Acad Sci USA. 2002;99:6955-6960. <lb/>20. Wejksza K, Lee-Chang C, Bodogai M, et al. Cancer-produced <lb/>metabolites of 5-lipoxygenase induce tumor-evoked regulatory B <lb/>cells via peroxisome proliferator-activated receptor α. J Immunol. <lb/>2013;190:2575-2584. <lb/>21. Schioppa T, Moore R, Thompson RG, et al. B regulatory cells and <lb/>the tumor-promoting actions of TNF-α during squamous carcinogen-<lb/>esis. Proc Natl Acad Sci USA. 2011;108:10662-10667. <lb/>22. Lindner S, Dahlke K, Sontheimer K, et al. Interleukin 21-induced <lb/>granzyme B-expressing B cells infiltrate tumors and regulate T cells. <lb/>Cancer Res. 2013;73:2468-2479. <lb/>23. Zhou X, Su YX, Lao XM, et al. CD19 + IL-10 + regulatory B cells affect <lb/>survival of tongue squamous cell carcinoma patients and induce rest-<lb/>ing CD4 + T cells to CD4 + Foxp3 + regulatory T cells. Oral Oncol. <lb/>2016;53:27-35. <lb/>24. Xiao X, Lao XM, Chen MM, et al. PD-1 hi identifies a novel regulatory <lb/>B-cell population in human hepatoma that promotes disease progres-<lb/>sion. Cancer Discov. 2016;6:546-559. <lb/>25. Watanabe R, Ishiura N, Nakashima H, et al. Regulatory B cells (B10 <lb/>cells) have a suppressive role in murine lupus: CD19 and B10 cell <lb/>defciency exacerbates systemic autoimmunity. J Immunol. <lb/>2010;184:4801-4809. <lb/>26. Tedder TF. B10 cells: a functionally defined regulatory B cell subset. <lb/>J Immunol. 2015;194:1395-1401. <lb/>27. Liu BS, Cao Y, Huizinga TW, et al. TLR-mediated STAT3 and ERK <lb/>activation controls IL-10 secretion by human B cells. Eur J Immunol. <lb/>2014;44:2121-2129. <lb/>28. Calin GA, Cimmino A, Fabbri M, et al. MiR-15a and miR-16-1 cluster <lb/>functions in human leukemia. Proc Natl Acad Sci USA. <lb/>2008;105:5166-5171. <lb/>29. Pekarsky Y, Balatti V, Croce CM. BCL2 and miR-15/16: from gene <lb/>discovery to treatment. Cell Death Differ. 2018;25:21-26. <lb/>30. Mohr A, Renaudineau Y, Bagacean C, et al. Regulatory B lymphocyte <lb/>functions should be considered in chronic lymphocytic leukemia. <lb/>Oncoimmunology. 2016;5:e1132977. <lb/></listBibl>
74
+
75
+ <div type="annex">S U P P O R T I N G I N F O R M A T I O N <lb/>Additional supporting information may be found online in the <lb/>Supporting Information section at the end of the article. <lb/></div>
76
+
77
+ <front>How to cite this article: Jia X, Liu H, Xu C, et al. MiR-15a/ <lb/>16-1 deficiency induces IL-10-producing CD19 + TIM-1 + cells <lb/>in tumor microenvironment. J Cell Mol Med. 2019;23:1343-<lb/>1353. https://doi.org/10.1111/jcmm.14037 <lb/></front>
78
+
79
+ <note place="headnote">JIA ET AL. <lb/>| </note>
80
+
81
+ <page>1353</page>
82
+
83
+
84
+ </text>
85
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JP01103169-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>arXiv:cond-mat/9903023v2 [cond-mat.supr-con] 24 Mar 1999 <lb/>Anisotropic resistivity of the antiferromagnetic <lb/>insulator Bi 2 Sr 2 ErCu 2 O 8 <lb/>T Kitajima, T Takayanagi, T Takemura and I Terasaki † <lb/>Department of Applied Physics, Waseda University, Tokyo 169-8555, Japan <lb/>Abstract. The anisotropic resistivities of Bi 2 Sr 2 Ca 1−x Er x Cu 2 O 8 single crystals were <lb/>measured and analyzed from 4.2 to 500 K with special interest taken in the parent <lb/>antiferromagnetic insulator with x=1.0. Although the resistivity is semiconducting <lb/>along both the in-plane and out-of-plane directions, the temperature dependences are <lb/>found to be significantly different. As a result, the resistivity ratio for x=1.0 exhibits <lb/>a broad maximum near room temperature. The electric conduction in the parent <lb/>antiferromagnetic insulators is different from that in other semiconductors, and is as <lb/>unconventional as that in high-temperature superconductors. <lb/></front>
8
+
9
+ <body>1. Introduction <lb/>Anisotropic transport properties in the normal state are one of the most striking features <lb/>of high-temperature superconductors (HTSC&apos;s) [1]. The metallic in-plane resistivity <lb/>(ρ ab ) accompanied by the non-metallic out-of-plane resistivity (ρ c ) enhances ρ c /ρ ab at <lb/>low temperature (T ) [2, 3], whereas ρ c /ρ ab is independent of T for conventional metals. <lb/>The enhancement of ρ c /ρ ab is often called &apos;confinement&apos; [4], and can be a key to the <lb/>elucidation of the mechanism of high-temperature superconductivity. We have studied <lb/>the anisotropic transport properties of slightly overdoped YBa 2 Cu 3 O y crystals [5, 6]. <lb/>Although their resistivities ρ c and ρ ab are both metallic, the anisotropy is difficult to <lb/>understand within the Fermi liquid theory. <lb/>The next question is that of whether ρ c /ρ ab is anomalous for a parent <lb/>antiferromagnetic (AF) insulator, whose resistivities ρ c and ρ ab are semiconducting. <lb/>To our knowledge, very little investigation has been done on ρ c /ρ ab . Thio et al. [7] <lb/>have found that ρ c /ρ ab for La 2 CuO 4 decreases with decreasing T below 200 K, which <lb/>is significantly incompatible with ρ c /ρ ab for HTSC&apos;s. Since it does not saturate near <lb/>200 K, a higher-temperature measurement is needed. <lb/></body>
10
+
11
+ <note place="footnote"> † To whom correspondence should be addressed. E-mail address: [email protected] <lb/></note>
12
+
13
+ <page>2 <lb/></page>
14
+
15
+ <body>Table 1. Characterization of Bi 2 Sr 2 Ca 1−x Er x Cu 2 O 8 single crystals. Note that the <lb/>actual composition ratio is represented by setting Cu = 2. <lb/>Nominal Actual Composition <lb/>Size <lb/>c-axis <lb/>x <lb/>Bi : Sr : Ca : Er : Cu <lb/>(mm 3 ) <lb/>( Å) <lb/>0 <lb/>1.9 : 1.9 : 1.2 : 0 : 2 <lb/>0.6×1×0.02 30.85 <lb/>0.1 <lb/>1.6 : 1.8 : 1.2 : 0.1 : 2 1×1.2×0.02 30.90 <lb/>0.5 <lb/>1.6 : 1.9 : 1.0 : 0.5 : 2 <lb/>1×1×0.004 <lb/>30.32 <lb/>1.0 <lb/>2.0 : 2.1 : 0 : 0.6 : 2 <lb/>1×1.2×0.004 30.33 <lb/>For studying ρ c /ρ ab over a wide temperature range, Bi 2 Sr 2 Ca 1−x R x Cu 2 O 8 (R: rare-<lb/>earth) is most suitable for the following reasons: (i) Oxygens for x=0 are chemically <lb/>stable up to 600 K in air where ρ ab remains &quot;T -linear&quot; [8]. (ii) When x reduces from 1 <lb/>to 0, the doping level varies from that of a parent AF insulator to that of a (slightly) <lb/>overdoped superconductor [9]. (iii) All of the Cu sites are equivalent, and only the <lb/>CuO 2 plane is responsible for the electric conduction. Here we report on measurements <lb/>and analyses of Bi 2 Sr 2 Ca 1−x Er x Cu 2 O 8 single crystals with x=0, 0.1, 0.5, and 1.0. We <lb/>have found that ρ c /ρ ab for a parent AF insulator (x=1.0) is quite unique; it increases <lb/>with T below 100 K, takes a broad maximum near 300 K, and decreases above room <lb/>temperature. This obviously indicates that a parent AF insulator exhibits a quite <lb/>different conduction mechanism from conventional semiconductors. <lb/>2. Experimental <lb/>Single crystals of Bi 2 Sr 2 Ca 1−x Er x Cu 2 O 8 were grown by a self-flux method [10]. Powders <lb/>of Bi 2 O 3 , SrCO 3 , CuO, CaCO 3 , and Er 2 O 3 of 99.9 % purity were mixed, well ground <lb/>in an Al 2 O 3 crucible, heated at 900 • C [1020 • C] for 10 h, and slowly cooled down to <lb/>760 • C [820 • C] by 2 • C/h for x=0 [x =0]. Since the single crystals were very thin along <lb/>the c axis, the thickness was measured with a scanning electron microscope (SEM). <lb/>The actual compositions were measured through inductively coupled plasma emission <lb/>spectroscopy. The x-ray diffraction pattern showed no trace of impurities, and the c-axis <lb/>lattice parameter for x=0 was evaluated to be 30.85 Å, which agrees with the value in <lb/>the literature [10, 11]. The measured compositions, sizes, and c-axis lattice parameters <lb/>are listed in table 1. We should note that crystals grown by a flux method are produced <lb/>with little stress, owing to the slow cooling rate near thermal equilibrium. In fact, we did <lb/>not observe any serious cracks in the SEM images of our samples. In order to examine <lb/>the influence of inhomogeneity and disorder on the resistivity, we made measurements <lb/>for more than 30 samples including ones grown from different batches. The measured <lb/>resistivities were reproducible enough to warrant the discussion in this paper. <lb/>The resistivity was measured with a dc current I in a four-probe configuration along <lb/></body>
16
+
17
+ <page>3 <lb/></page>
18
+
19
+ <body>10 -4 <lb/>10 -3 <lb/>10 -2 <lb/>10 -1 <lb/>10 0 <lb/>10 1 <lb/>10 2 <lb/>10 3 <lb/>ρ <lb/>ab (Ωcm) <lb/>Bi 2 Sr 2 Ca 1-x Er x Cu 2 O 8 <lb/>x=0 <lb/>x=0.1 <lb/>x=1 <lb/>x=0.5 <lb/>0 <lb/>100 <lb/>200 <lb/>300 <lb/>400 <lb/>500 <lb/>10 0 <lb/>10 1 <lb/>10 2 <lb/>10 3 <lb/>10 4 <lb/>10 5 <lb/>10 6 <lb/>Temperature (K) <lb/>ρ <lb/>c (Ωcm) <lb/>x=0 <lb/>x=0.1 <lb/>x=1 <lb/>x=0.5 <lb/>0 100 200 <lb/>10 0 <lb/>10 1 <lb/>10 2 <lb/>10 3 <lb/>T (K) <lb/>ρ/ρ <lb/>300 K <lb/>x=1.0 <lb/>ρ ab <lb/>ρ c <lb/>(a) <lb/>(b) <lb/>Figure 1. (a) The in-plane resistivity ρ ab and (b) out-of-plane resistivity ρ c of single <lb/>crystals of Bi 2 Sr 2 Ca 1−x Er x Cu 2 O 8 . ρ ab and ρ c for x=1.0 normalized at 300 K are <lb/>plotted as functions of temperature in the inset. <lb/>the in-plane direction (I ⊥ c), and in a ring configuration along the out-of-plane direction <lb/>(I c). We used two measurement systems below and above room temperature. From <lb/>4.2 to 300 K, the samples were slowly (100 K/h) cooled in a liquid-He cryostat, where T <lb/>was monitored through a cernox resistance thermometer. Above 300 K, the samples are <lb/>slowly (50-100 K/h) heated in air in a cylinder furnace with a Pt resistance thermometer. <lb/>3. Results and discussion <lb/>Figures 1(a) and 1(b) show ρ ab and ρ c of Bi 2 Sr 2 Ca 1−x Er x Cu 2 O 8 single crystals, <lb/>respectively. The magnitudes of ρ ab and ρ c increase with x, showing that the hole <lb/>concentration is reduced by the Er substitution. As is seen in the literature, ρ c is four or <lb/>five orders of magnitude lager than ρ ab for all the samples. For superconducting samples <lb/>(x=0 and 0.1), metallic ρ ab and semiconducting ρ c are observed above T c . Reflecting <lb/>the slightly overdoped nature of x=0, T c (∼84 K) for the x=0 sample is lower than T c <lb/>(∼87 K) for x=0.1. These results attest to the quality of our crystals. <lb/>Both ρ ab and ρ c for x=1.0 are semiconducting, but they exhibit different T <lb/>dependences. Above room temperature, where ρ ab decreases gradually in comparison <lb/></body>
20
+
21
+ <page>4 <lb/></page>
22
+
23
+ <body>0 <lb/>100 <lb/>200 <lb/>300 <lb/>400 <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>6 <lb/>Temperature (K) <lb/>ρ <lb/>c (T)/ρ <lb/>c (450 K) <lb/>Bi 2 Sr 2 Ca 1-x Er x Cu 2 O 8 <lb/>x=0 <lb/>x=0.1 <lb/>x=0.5 <lb/>x=1.0 <lb/>ρ <lb/>ab (T)/ρ <lb/>ab (450 K) <lb/>0 <lb/>1 <lb/>10 -4 <lb/>10 -2 <lb/>10 0 <lb/>10 2 <lb/>x <lb/>ρ <lb/>at 450K (Ωcm) <lb/>ρ c <lb/>ρ ab <lb/>Figure 2. The anisotropic resistivity ratios ρ c /ρ ab of Bi 2 Sr 2 Ca 1−x Er x Cu 2 O 8 single <lb/>crystals normalized at 450 K. Inset: The magnitudes of ρ ab and ρ c at 450 K are plotted <lb/>as functions of x. Note that each symbol represents the resistivity of a different sample <lb/>measured to check the reproducibility (see text). <lb/>with ρ c , ρ c /ρ ab decreases with increasing T . On the other hand, ρ ab becomes insulating <lb/>more rapidly than ρ c , as shown in the inset of figure 1 where the resistivities are <lb/>normalized at 300 K. Thus ρ c /ρ ab decreases with decreasing T below 300 K. These <lb/>results are not understandable on the basis of conventional theories. In the framework <lb/>of a band picture, anisotropy is mainly determined by effective masses, implying that <lb/>the T dependence of ρ is independent of the direction. In the case of a hole doped in <lb/>the AF background, the situations are nearly the same. In fact, a nearly T -independent <lb/>ρ c /ρ ab has been observed for La 2 NiO 4 [12] and Bi 2 M 3 Co 2 O 9 (M=Ca, Sr and Ba) [13]. <lb/>The magnitude of ρ c /ρ ab for a parent AF insulator is much more difficult to evaluate <lb/>than that for a superconductor, in that it is an exponentially varying quantity divided <lb/>by another exponentially varying quantity. Since we are interested in the T dependence <lb/>of ρ c /ρ ab , we normalize ρ c /ρ ab at 450 K in figure 2. As for the magnitude, we show ρ ab <lb/>and ρ c in the inset of figure 2, in which each symbol corresponds to a different sample. <lb/>From the inset one can see that the magnitude of ρ c /ρ ab at 450 K is nearly independent <lb/>of x. Accordingly the normalization at 450 K will not adversely affect the discussion <lb/>below. <lb/>We would like to point out three features in figure 2. First, ρ c /ρ ab changes smoothly <lb/>with x above room temperature; It increases with decreasing T , and the T dependence is <lb/>steeper for larger x (smaller hole concentration). If one looked at ρ c /ρ ab only above room <lb/>temperature, one could not distinguish a parent AF insulator from a superconductor. <lb/>Thus we may say that the holes are confined in a parent AF insulator as well as in <lb/>HTSC. In this context the former is as unconventional as the latter. Secondly, ρ c /ρ ab <lb/></body>
24
+
25
+ <page>5 <lb/></page>
26
+
27
+ <body>0 <lb/>0.05 <lb/>0.1 <lb/>0.15 <lb/>0 <lb/>100 <lb/>200 <lb/>300 <lb/>400 <lb/>Hole concentration per Cu <lb/>Temperature (K) <lb/>Bi 2 Sr 2 Ca 1-x R x Cu 2 O 8 <lb/>T max R=Er <lb/>T c R=Er <lb/>T N R=Y <lb/>T c R=Y <lb/>Fig. 3 Kitajima et al. <lb/>AF <lb/>SC <lb/>Figure 3. The phase diagram of Bi 2 Sr 2 Ca 1−x R x Cu 2 O 8 (R=Er and Y). AF and <lb/>SC represent the antiferromagnetic order and the superconducting phase, respectively. <lb/>T max (the temperature at which ρ c /ρ ab takes a maximum), T N (the Néel temperature), <lb/>and T c (the superconducting transition temperature) are plotted as a function of hole <lb/>concentration. T N and T c for R=Y are taken from Refs. [15] and [17]. The hole <lb/>concentration is estimated from the room-temperature thermopower, as is proposed in <lb/>Ref. [16]. The error bars in T max represent the variation from sample to sample. <lb/>for x=1.0 and 0.5 decreases with decreasing T below 100 K, which is consistent with <lb/>ρ c /ρ ab of La 2 CuO 4 [7]. The decrease of ρ c /ρ ab as T →0 could be understood from the <lb/>three-dimensional (3D) nature of the localization [14]. Thirdly, ρ c /ρ ab for x=1.0 and <lb/>0.5 takes a maximum at a certain temperature T max , which is very close to the Néel <lb/>temperature T N . (For x=0.5, a tiny fraction of a superconducting phase causes a small <lb/>drop of resistivity near 90 K, which obscures the position of T max .) <lb/>The localized f spins of Er 3+ make it difficult to measure T N of Cu 2+ for x=1.0 and <lb/>0.5. Instead, we will compare T N of Bi 2 Sr 2 Ca 1−x Y x Cu 2 O 8 [15], and we plot T c , T N and <lb/>T max in the electronic phase diagram of Bi 2 Sr 2 Ca 1−x R x Cu 2 O 8 in figure 3. We estimated <lb/>the hole concentration using an empirical relation to the thermopower [16], which we <lb/>measured with the same samples for R=Er (not shown here), and used Ref. [17] for <lb/>R=Y. T max is found to lie around the AF boundary. Since no structural transitions and <lb/>no phase separations are reported for Bi 2 Sr 2 Ca 1−x R x Cu 2 O 8 [9], it would be natural to <lb/>relate T max to the AF transition. <lb/>The confinement behavior above T N favors some theories based on the two-<lb/>dimensional (2D) spin fluctuation, which exists in parent AF insulators above T N [18] <lb/>and in HTSC&apos;s above T c as well [19]. We therefore propose that holes are confined in a <lb/></body>
28
+
29
+ <page>6 <lb/></page>
30
+
31
+ <body>CuO 2 plane in the presence of the 2D spin fluctuation, which occurs irrespective of doping <lb/>levels. As the 3D AF order grows below T N , the confinement becomes less effective. A <lb/>recent numerical analysis of the bilayer t − J model has also led to the assertion that <lb/>ρ c is substantially modified in the presence of the 2D spin fluctuation [20]. We further <lb/>note that a similar case is seen for a layered ferromagnet La 2−x Sr 1+x Mn 2 O 7 [21]. For <lb/>100 K &lt; T &lt; 250 K, this compound is in a 2D ferromagnetic state, and exhibits a <lb/>non-metallic ρ c together with a metallic ρ ab . Once the 3D ferromagnetic order appears <lb/>below 100 K, ρ c becomes metallic to behave in a 3D-like manner. We believe that the <lb/>out-of-plane conduction in parent AF insulators includes essentially the same physics <lb/>as for La 2−x Sr 1+x Mn 2 O 7 ; the only difference is as regards whether the material is an <lb/>antiferromagnetic insulator or a ferromagnetic metal. <lb/>4. Summary <lb/>We prepared Bi 2 Sr 2 Ca 1−x Er x Cu 2 O 8 , single crystals for x=0, 0.1, 0.5 and 1.0 and <lb/>measured the in-plane and out-of-plane resistivities (ρ ab and ρ c ) from 4.2 to 500 K. The <lb/>present study has revealed that ρ c /ρ ab for a parent antiferromagnetic insulator (x = 1.0) <lb/>strongly depends on temperature, and that the enhancement of ρ c /ρ ab with decreasing T <lb/>is observed above room temperature. In this sense, parent antiferromagnetic insulators <lb/>are as unconventional as high-temperature superconductors. Their ratios ρ c /ρ ab take <lb/>maxima at a certain temperature near the Néel temperature, and we propose that the <lb/>confinement in the CuO 2 plane is operative in the two-dimensional spin-fluctuation <lb/>regime regardless of the doping level. <lb/></body>
32
+
33
+ <div type="acknowledgement">Acknowledgments <lb/>The authors would like to thank T Itoh, T Kawata, K Takahata and Y Iguchi for <lb/>collaboration. They also wish to express their appreciation to S Kurihara and S Saito <lb/>for fruitful discussions and valuable comments. One of the authors (I T) is indebted <lb/>to S Tajima for the collaboration at a preliminary stage of this work. This work was <lb/>partially supported by Waseda University Grant for Special Research Projects (97A-565, <lb/>98A-618). <lb/></div>
34
+
35
+ <listBibl>References <lb/>[1] For a recent review of the anisotropic charge dynamics, see Cooper S L and Gray K E 1994 Physical <lb/>Properties of the High Temperature Superconductors IV, edited by Ginsberg D M (Singapore: <lb/>World Scientific) p. 61 <lb/>[2] Takenaka K, Mizuhashi K, Takagi H and Uchida S 1994 Phys. Rev. B50 6534 <lb/>[3] Nakamura Y and Uchida S 1993 Phys. Rev. B47 8369 <lb/></listBibl>
36
+
37
+ <page>7 <lb/></page>
38
+
39
+ <listBibl>[4] Anderson P W 1997 The Theory of Superconductivity in the High-T c Cuprates (Princeton: <lb/>Princeton University Press) <lb/>[5] Terasaki I, Sato Y, Miyamoto S, Tajima S and Tanaka S 1995 Phys. Rev. B52 16246 <lb/>[6] Terasaki I, Sato Y and Tajima S 1996 Phys. Rev. B55 15300 <lb/>[7] Thio T, Chen C Y, Freer B S, Gabbe D R, Jenssen H P, Kastner M A, Picone P J, Preyer N W <lb/>and Birgeneau R J 1990 Phys. Rev. B41 231 <lb/>[8] Martin S, Fiory A T, Fleming R M, Schneemeyer L F and Waszczak J V 1988 Phys. Rev. Lett. <lb/>60 2194 <lb/>[9] Quitmann C, Andrich D A, Jarchow C, Fleuster M, Beschoten B, Güntherodt G, Moshchalkov V <lb/>V, Mante G and Manzke R 1992 Phys. Rev. B46 11813 <lb/>[10] Kendziora C, Forro L, Mandrus D, Hartge J, Stephens P, Mihaly L, Reeder R, Moecher D, Rivers <lb/>M and Sutton S 1992 Phys. Rev. B45 13025 <lb/>[11] Ilyushin A S, Shi L, Leonyuk L I, Mustafa B M, Nikanorova I A, Red&apos;ko S V, Jia Y, Vetkin A G, <lb/>Zhou G, Zubov I V 1993 J. Mater. Res. 8 1791 <lb/>[12] Rao C N R, Buttrey D J, Otsuka N, Ganguly P, Harrison H R, Sandberg C J and Honig J M 1984 <lb/>J. Solid State. Chem. 51 266 <lb/>[13] Watanabe Y, Tui D C, Birminghem J T, Ong N P and Tarascon J M 1991 Phys. Rev. B43 3026 <lb/>[14] Shkolovskii B I 1977 Sov. Phys. Semicond. 11 1253 <lb/>[15] Nishida N, Okuma S, Miyatake H, Tamegai T, Iye Y, Yoshizaki R, Nishiyama K, Nagamine K, <lb/>Kadono R and Brewer J H 1990 Physica C168 23 <lb/>[16] Obertelli S D, Cooper J R, and Tallon J L 1992 Phys. Rev. B46 14928 <lb/>[17] Mandrus D, Forro L, Kendziora C and Mihaly L 1991 Phys. Rev. B44 2418 <lb/>[18] Shirane G, Endoh Y, Birgeneau R J, Kastner M A, Hidaka Y, Oda M, Suzuki M and Murakami <lb/>T 1987 Phys. Rev. Lett. 59 1613 <lb/>[19] Yamada K, Lee C H, Kurahashi K, Wada J, Wakimoto S, Ueki S, Kimura H, Endoh Y, Hosoya S, <lb/>Shirane G, Birgeneau R J, Greven M, Kastner M A and Kim Y J 1998 Phys. Rev. B57 6165 <lb/>[20] Eder R, Ohta Y and Maekawa S 1995 Phys. Rev. B51 3265 <lb/>[21] Kimura T, Tomioka Y, Kuwahara H, Asamitsu A, Tamura M and Tokura Y 1996 Science 274 <lb/>1698 </listBibl>
40
+
41
+
42
+ </text>
43
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPC0150519-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <front>arXiv:cond-mat/0305503v2 [cond-mat.supr-con] 26 May 2003 <lb/>Superconductivity in a layered cobalt oxyhydrate Na 0.31 CoO 2 •1.3H 2 O <lb/>Guanghan Cao, 1, * Chunmu Feng, Yi Xu, 1 Wen Lu, 1 Jingqin Shen, 1 Minghu Fang, 1 and Zhu&apos;an Xu 1 <lb/>1 Department of Physics, Zhejiang University, Hangzhou 310027, P. R. China <lb/>Test &amp; Analysis Center, Zhejiang University, Hangzhou 310027, P. R. China <lb/>(Dated: November 18, 2018) <lb/>We report the electrical, magnetic and thermal measurements on a layered cobalt oxyhydrate <lb/>Na0.31CoO2•1.3H2O. Bulk superconductivity at 4.3 K has been confirmed, however, the measured <lb/>superconducting fraction is relatively low probably due to the sample&apos;s intrinsic two-dimensional <lb/>characteristic. The compound exhibits weak-coupled and extreme type-II superconductivity with <lb/>the average energy gap ∆a(0) and the Ginzburg-Landau parameter κ of ∼ 0.50 meV and ∼ 140, <lb/>respectively. The normalized electronic specific heat data in the superconducting state well fit the <lb/>T dependence, suggesting point nodes for the superconducting gap structure. <lb/>PACS numbers: 74.70.-b, 74.25.-q, 74.20.Rp <lb/>Keywords: Cobalt oxyhydrate superconductor, Type-II superconductivity, Superconducting gap structure <lb/></front>
9
+
10
+ <body>The recent discovery of superconductivity in a two-<lb/>dimensional cobalt oxyhydrate [1] has been spurring new <lb/>round of intense interest in the field of superconductivity <lb/>research. It was mentioned [1, 2] that the cobalt oxy-<lb/>hydrate superconductor resembles the high-T c cuprates <lb/>in the two-dimensional (2D) MO (M=Co or Cu) lay-<lb/>ers and the existence of spin 1/2 for Co 4+ and Cu 2+ <lb/>ions, though their difference is obvious for the triangular <lb/>CoO 2 sheets in contrast with the nearly tetragonal CuO <lb/>planes. The fact that the superconductivity is derived <lb/>from the intercalation of H O into the host Na 0.35 CoO 2 , <lb/>which itself is not a superconductor, suggests that strong <lb/>two-dimensionality be important for the appearance of <lb/>superconductivity [1]. <lb/>The related theoretical work has been performed <lb/>quickly, though some basic physical property character-<lb/>izations of the new superconductors have not been re-<lb/>ported yet. By employing the t−J model on a planar tri-<lb/>angular lattice, different kinds of superconducting states, <lb/>such as time-reversal-symmetry-breaking d x 2 −y 2 + id xy <lb/>superconductivity [3, 4, 5], and spin triplet supercon-<lb/>ductivity [3, 6] have been proposed. Based on the den-<lb/>sity functional calculation, Singh [7] also speculates that <lb/>a triplet superconducting state may exist in this kind of <lb/>material. In a word, exotic superconductivity in the new <lb/>system seems to be a consensus for theorists. To verify <lb/>the theoretical result, therefore, the experimental inves-<lb/>tigations becomes very crucial on this topic. <lb/>Unfortunately, the development on the experimental <lb/>aspect goes relatively slowly. One of the major reasons <lb/>is that the preparation of samples is not optimized at <lb/>present. The other reason concerns about the chemical <lb/>instability of the oxyhydrate superconductor. It was re-<lb/>ported [8] that the material is exceptionally sensitive to <lb/>both temperature and humidity near ambient conditions, <lb/>which makes the experimental reproducibility rather dif-<lb/>ficult. Consequently, only a few experimental results, <lb/>such as the magnetic properties [9] and the hydrostatic <lb/>pressure effect on T c [10] have just been reported. Al-<lb/>though some unconventional magnetic properties were <lb/>revealed for the new superconductor [9], other basic prop-<lb/>erties such as the low-temperature specific heat have not <lb/>been reported yet for this newly-discovered supercon-<lb/>ductor. We recently succeeded in preparing the cobalt <lb/>oxyhydrate superconductor using a modified synthetic <lb/>route [11]. The problem of the sample&apos;s instability was <lb/>overcome to some extent by employing suitable experi-<lb/>mental procedure. In this Letter, we report some super-<lb/>conducting and normal-state properties of this intriguing <lb/>compound. <lb/>Our samples of Na 0.31 CoO 2 •1.3H 2 O were prepared in <lb/>four steps, briefly described as follows. First, single-<lb/>phase hexagonal Na 0.74 CoO 2 was prepared by a solid-<lb/>state reaction at 1083 K in flowing oxygen with Na 2 CO 3 <lb/>and Co 3 O 4 as the starting material. Second, partial <lb/>sodium in Na 0.74 CoO 2 was deintercalated by the exces-<lb/>sive bromine solved in acetonitrile, similar to the treat-<lb/>ment reported previously [1, 8]. Third, a hydration pro-<lb/>cess was carried out by the direct reaction with distilled <lb/>water. Last, the sample was slightly dehydrated and <lb/>then &quot;annealed&quot; under ambient condition. Powder X-<lb/>ray Diffraction (XRD) measurement indicates that the <lb/>final product is a hexagonal single phase with the cell <lb/>constants of a=2.820Å and c=19.65Å . The unit cell is <lb/>slightly stretched along the c-axis, compared with that <lb/>of the previous report [1]. This is probably due to the <lb/>difference in the Na content. By employing the Atomic <lb/>absorption spectroscopy, the atomic ratio of Na and Co <lb/>was determined as 0.31 for the final product. Thermo-<lb/>gravimetric analysis shows that the weight loss from 293 <lb/>K to 693 K is 19.8 %, indicating that the content of H O <lb/>is about 1.3 per formula. Therefore, the chemical formula <lb/>of the final product is expressed as Na 0.31 CoO 2 •1.3H 2 O . <lb/>Details of the sample&apos;s preparation and characterizations <lb/>will be given elsewhere [11]. <lb/>The physical property measurements were performed <lb/>at the temperature down to 1.8 K and under the field up <lb/>to 8 Tesla, on a Quantum Design PPMS system. While <lb/></body>
11
+
12
+ <page>2 <lb/></page>
13
+
14
+ <body>measured under &quot;zero field&quot;, there still exists a remanent <lb/>field of ∼ Oe. The precisions of ac magnetic suscep-<lb/>tibility (χ ac ) and dc susceptibility (χ dc ) are better than <lb/>∼ 10 −7 emu and ∼ −5 emu, respectively. The electri-<lb/>cal resistivity (ρ) was measured in a standard four-probe <lb/>configuration using a pressed sample bar. The heat ca-<lb/>pacity was measured using an automated relaxation tech-<lb/>nique with a square piece of ∼ 20 mg sample. The con-<lb/>tribution from the addenda has been subtracted. It is <lb/>noted that the handling of the sample and the detailed <lb/>measurement procedure sometimes affect the experimen-<lb/>tal result very much. So, we kept the same experimental <lb/>condition for the different measurements. <lb/>Figure 1(a) shows the temperature dependence of <lb/>magnetic susceptibility at low temperatures for the <lb/>Na 0.31 CoO 2 •1.3H 2 O sample. The real part of ac suscep-<lb/>tibility χ ′ shows the onset of diamagnetism at 4.3 K, <lb/>followed by a broad superconducting transition, similar <lb/>to the original report [1]. The diamagnetic screening sig-<lb/>nal at 1.9 K is 9.2 % of the full shielding when the ac <lb/>field amplitude (H ac ) is 2 Oe, suggesting relatively low <lb/>superconducting fraction. Considering that the χ ′ value <lb/>is not flat down to 1.9 K, the superconducting volume <lb/>fraction will be over 10 % under the remanent field of <lb/>∼ 1 Oe. The imaginary component of the ac susceptibil-<lb/>ity shows an incomplete dissipation peak, also suggesting <lb/>that the superconducting transition is not finished yet at <lb/>1.9 K. The dc susceptibility under 30 Oe shows even low <lb/>magnetic exclusion, which is primarily due to the very <lb/>low H c1 value as well as the magnetic penetration (see <lb/>the result below). An irreversible temperature can be <lb/>noticed, like that observed in the high T c cuprates [12]. <lb/>From the structural and chemical bonding points of <lb/>view, the present system should have very weak coupling <lb/>between the CoO layers, resulting in a strong 2D su-<lb/>perconductivity. It is proposed that the relatively low <lb/>superconducting fraction is mainly due to the sample&apos;s <lb/>intrinsic 2D characteristic. The following observations <lb/>are coincident with this point. First, the superconduct-<lb/>ing transition is broad. Second, zero resistance can never <lb/>be achieved in our experiments as well as the previous re-<lb/>port [1]. Third, the diamagnetic signal is enhanced when <lb/>decreasing H ac . Similar result was reported for a 2D or-<lb/>ganic superconductor (BEDT-TTF) 2 Cu(NCS) [13]. It <lb/>should be pointed out that the low superconducting frac-<lb/>tion is not mainly due to the sample&apos;s instability, because <lb/>our XRD experiment shows that the sample contains no <lb/>secondary phases before and after the magnetic suscep-<lb/>tibility measurements. <lb/>Figure 2(a) shows the magnetization loop at 1.9 K for <lb/>the Na 0.31 CoO 2 •1.3H 2 O sample. Narrow field hystere-<lb/>sis was observed, superposed with a paramagnetic back-<lb/>ground which can be described by the Brillouin func-<lb/>tion. The amplificatory plot using the upper-right co-<lb/>ordinates indicates the type-II superconductivity with <lb/>H c1 of about 10 Oe at 1.9 K. By the data fitting <lb/>0 <lb/>2 <lb/>4 <lb/>6 <lb/>8 <lb/>10 <lb/>-8 <lb/>-6 <lb/>-4 <lb/>-2 <lb/>0 <lb/>H ac =10 Oe <lb/>H ac =2 Oe <lb/>T (K) <lb/>χ&apos; <lb/>-10 χ&apos;&apos; <lb/>χ <lb/>ac (10 <lb/>-3 <lb/>emu/cm <lb/>3 <lb/>) <lb/>2 <lb/>4 <lb/>6 <lb/>8 <lb/>10 <lb/>-4 <lb/>-3 <lb/>-2 <lb/>-1 <lb/>0 <lb/>T irr =3.2 K <lb/>H=30 Oe <lb/>ZFC <lb/>FC <lb/>χ <lb/>dc (10 <lb/>-3 <lb/>emu/cm <lb/>) <lb/>T (K) <lb/>FIG. 1: Temperature dependence of ac magnetic suscepti-<lb/>bility at zero field for Na0.31CoO2•1.3H2O powdered sample. <lb/>The inset shows the dc magnetic susceptibility under the field <lb/>H=30 Oe. Hac, Tirr, FC and ZFC refer to the ac field am-<lb/>plitude, irreversible temperature, field cooling and zero-field <lb/>cooling, respectively. <lb/>of H c1 (T ) using the well-known equation: H c1 (T ) = <lb/>H c1 (0)[1 − (T /T c ) 2 ], H c1 (0) can be obtained as 13 Oe. <lb/>The H c2 value is difficult to be measured by the M − H <lb/>curve due to the very narrow hysteresis. Nevertheless, <lb/>by measuring the electrical resistance at fixed temper-<lb/>atures, one can basically obtain the H c2 (T ) data, as <lb/>shown in figure 2(b). H c2 (T ) is here determined as the <lb/>point where the resistance deviates from the linearity <lb/>in the R − H 2 curves [14]. The slope of H c2 at T c , <lb/>dH c2 /dT | Tc , is obtained as −34 kOe/K. H c2 (0) can <lb/>thus be estimated to be 1×10 5 Oe, using the WHH for-<lb/>mula [15]. Then, the average Ginzburg-Landau (GL) <lb/>coherent length ξ GL (0)=57Å can be calculated using <lb/>the formula ξ GL (0)=(Φ /2πH c2 (0)) 1/2 . On the other <lb/>hand, by solving the equation H c1 = Φ 0 ln(λ/ξ)/4πλ 2 , <lb/>the average penetration depths can also be obtained: <lb/>λ(0)=7900Å . Therefore, the Ginzburg-Landau param-<lb/>eter κ = λ/ξ GL is estimated as ∼ 140, indicating that <lb/>the cobalt oxyhydrate is an extreme type-II supercon-<lb/>ductor, like the high-T c cuprates. This conclusion has <lb/>also been drawn in a very recent report [9], in which dif-<lb/>ferent method was employed to determine the H c2 (T ). <lb/>It is worth while to note that, compared with the previ-<lb/>ous result, the values of H c1 (0) and H c2 (0) in the present <lb/>sample are remarkably smaller, which is possibly resulted <lb/>from the differences in the carrier-doping level and/or the <lb/>water content. <lb/>The result of low-temperature specific heat measure-<lb/>ment is shown in figure 3. At temperatures much below <lb/>the Debye temperature Θ D , and if neglected the possi-<lb/>ble magnetic contribution, the specific heat can be ex-<lb/>pressed as the sum of electron and phonon contributions: <lb/>C = γT + βT 3 , where the coefficient γ is generally called <lb/>
15
+
16
+ 0 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>6 <lb/>1.1 <lb/>1.2 <lb/>1.3 <lb/>1.4 <lb/>1.5 <lb/>(b) <lb/>R (Ω) <lb/>µ 0 H (T) <lb/>1.8 K <lb/>2.1 K <lb/>2.4 K <lb/>2.7 K <lb/>3.0 K <lb/>3.3 K <lb/>3.6 K <lb/>3.9 K <lb/>4.2 K <lb/>4.5 K <lb/>5.0 K <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>0 <lb/>20 <lb/>40 <lb/>60 <lb/>H <lb/>c2 (k Oe) <lb/>T (K) <lb/>-2 <lb/>-1 <lb/>1 <lb/>2 <lb/>-1.0 <lb/>-0.5 <lb/>0.0 <lb/>0.5 <lb/>1.0 <lb/>(a) <lb/>T=1.90 K <lb/>M (emu/g) <lb/>µ 0 H (T) <lb/>-600 <lb/>-400 <lb/>-200 <lb/>0 <lb/>200 <lb/>400 <lb/>600 <lb/>-0.06 <lb/>-0.04 <lb/>-0.02 <lb/>0.00 <lb/>0.02 <lb/>0.04 <lb/>0.06 <lb/>H (Oe) <lb/>FIG. 2: Magnetic field dependence of magnetization (a) and <lb/>electrical resistance (b) at certain temperatures. Note that <lb/>the upper-right axes are employed for the amplificatory plot <lb/>in (a). The inset of (b) shows the temperature dependence of <lb/>the upper critical field Hc2. <lb/>Sommerfeld parameter. The phonon contribution can be <lb/>separated by employing the T 2 vs C/T plot. It can be <lb/>seen that good linearity is satisfied in the temperature <lb/>range of 4.5 K&lt; T &lt;11 K. We thus obtained γ = 15.9 <lb/>mJ/K 2 •mol-f.u. (f.u. denotes formula unit) and β = <lb/>0.235 mJ/K 4 • mol-f.u. Θ D is then calculated to be 391 <lb/>K using the formula Θ D = ((12/5)N π 4 R/β) 1/3 , where <lb/>N =7.21 for Na 0.31 CoO 2 •1.3H 2 O and R=8.314 J/mol•K. <lb/>The γ value is significantly smaller than that of the par-<lb/>ent compound Na 0.5 CoO 2 (γ ∼ 40 mJ/K 2 •mol-Co [16]). <lb/>Since the Sommerfeld parameter γ is related to the den-<lb/>sity of state (DOS) at Fermi level, N (E F ), by the relation <lb/>γ = 1 <lb/>3 k 2 <lb/>B π 2 N (E F ) = 1 <lb/>3 k 2 <lb/>B π 2 N (0)(1 + λ ep ), where N (0) is <lb/>the bare, or band-structure electronic DOS at E F , λ ep an <lb/>electron-phonon interaction parameter [17], one can ob-<lb/>tain that N (E F )=6.7 states/eV•f.u. On the other hand, <lb/>λ ep can be calculated to be 0.57 using the formula <lb/>λ ep = <lb/>1.04 + µ * ln(Θ D /1.45T c ) <lb/>(1 − 0.62µ * )ln(Θ D /1.45T c ) − 1.04 <lb/>, <lb/>(1) <lb/>where Coulomb repulsion parameter µ * is assumed to be <lb/>0.13 empirically [17]. Therefore, N (0) is derived to be 4.3 <lb/>states/eV•f.u. We note that this value is almost identical <lb/>to the band calculation result (4.4 states/eV•Co) for the <lb/>parent compound Na 0.5 CoO 2 [18]. <lb/>It is noted that the sample&apos;s magnetic susceptibility (∼ <lb/>2.0×10 −3 emu/mol-f.u) is almost temperature indepen-<lb/>dent from 30 K to 300 K (not shown here). In order to <lb/>obtain the Pauli susceptibility χ P auli , the χ(T ) data were <lb/>fitted using the equation χ = χ 0 + AT 2 + C/(T − θ) [9]. <lb/>We obtained that χ 0 , A, C and θ are 0.0019 emu/mol, <lb/>2.6 ×10 −9 emu/mol•K 2 , 0.0024 emu•K/mol and 1.7 K, <lb/>respectively. The parameter C gives the small effective <lb/>magnetic moment of 0.14 µ B . The small positive θ value <lb/>suggests the existence of weak ferromagnetic correlations. <lb/>The unusually large χ 0 value should be dominantly con-<lb/>tributed by χ P auli , which is probably enhanced by the <lb/>Stoner-type ferromagnetic correlation. The Wilson ra-<lb/>tio, R W = π 2 k 2 <lb/>B χ P auli /3γµ 2 <lb/>B , is calculated to be 11, in <lb/>sharp contrast with the case of heavy fermion supercon-<lb/>ductor [19]. <lb/>At 4.3 K, specific heat anomalies can be seen, which <lb/>is ascribed to the superconducting transition. The spe-<lb/>cific heat jump at the T c under zero field, ∆C obs , is 6.9 <lb/>mJ/K•mol-f.u, further confirming the bulk superconduc-<lb/>tivity. When applying magnetic field, both the ∆C obs <lb/>and T c decrease as expected. It is noted that the T c (H) <lb/>values are basically consistent with the H c2 (T ) result de-<lb/>scribed above. <lb/>The specific jump at T c , ∆C, can be calculated us-<lb/>ing an approximate formula ∆C = H c (0) 2 /2πT c , where <lb/>H c (0) is the thermodynamic critical field. H c (0) is found <lb/>to be 505 Oe by using the formula H c (0) = H c2 (0)/ <lb/>√ <lb/>2κ, <lb/>where H c2 (0) and κ are 1×10 5 Oe and 140, respec-<lb/>tively. Then, ∆C should be 38.2 mJ/K•mol-f.u. There-<lb/>fore, the superconducting fraction is estimated to be <lb/>∆C obs /∆C=18.1 %, which is basically consistent with <lb/>the magnetic susceptibility measurement result. In addi-<lb/>tion, the average superconducting gap at zero tempera-<lb/>ture, ∆ a (0), can be obtained using the relation [20], <lb/>2∆ a (0) <lb/>k B T c <lb/>= <lb/>4π <lb/>√ <lb/>3 <lb/>[ <lb/>H c (0) 2 V m <lb/>8πγT 2 <lb/>c <lb/>] 1/2 . <lb/>(2) <lb/>∆ a (0) is then obtained to be 0.50 meV. The value of <lb/>2∆ a (0)/k B T c is found to be 2.71, suggesting that the <lb/>system belongs to the weak coupling limit. <lb/>A further data-analysis was carried out as follows. The <lb/>lattice specific-heat contribution, C L = βT 3 , was first <lb/>deducted, obtaining the electronic specific heat term: <lb/>C el = C − C L . If the superconducting fraction is η, the <lb/>electronic specific heat of the full superconductor can be <lb/>normalized as C es =[C el −(1−η)γT ]/η. Figure 3(b) shows <lb/>the result with η=18.1 %. The Sommerfeld-parameter <lb/>jump at the T c , ∆C/T c , becomes 9 mJ/K 2 •mol-f.u. So, <lb/>the dimensionless parameter ∆C/γT value is about 0.57, <lb/>which is remarkably lower than the expected value 1.43 <lb/>
17
+
18
+ 0 <lb/>20 <lb/>40 <lb/>60 <lb/>80 <lb/>100 <lb/>120 <lb/>0.00 <lb/>0.01 <lb/>0.02 <lb/>0.03 <lb/>0.04 <lb/>(a) <lb/>B=0 T <lb/>Linear Fit <lb/>C /T (J/mol K <lb/>) <lb/>T <lb/>2 (K <lb/>2 <lb/>) <lb/>0 <lb/>2 <lb/>4 <lb/>6 <lb/>8 <lb/>0.000 <lb/>0.005 <lb/>0.010 <lb/>0.015 <lb/>0.020 <lb/>0.025 <lb/>∆C/T <lb/>(b) <lb/>C es ~ T <lb/>2 <lb/>C es ~ T <lb/>3 <lb/>C es ~ exp (-1.5T c /T) <lb/>C <lb/>el /T (J/mol-K <lb/>) <lb/>T (K) <lb/>10 <lb/>20 <lb/>30 <lb/>40 <lb/>0.015 <lb/>0.020 <lb/>0.025 <lb/>B=8 T <lb/>B=2 T <lb/>B=0 T <lb/>FIG. 3: <lb/>Low-temperature specific heat result of the <lb/>Na0.31CoO2•1.3H2O superconductor. (a) plot of C/T vs T 2 . <lb/>The arrows in the inset point to the Tc under different field. <lb/>(b) temperature dependence of the normalized Sommerfeld <lb/>parameter. The electronic specific heat data in the super-<lb/>conducting state, Ces(T ), was fitted using different formula <lb/>containing just one fitting parameter (the coefficient). <lb/>for an isotropic gap. This suggests that the supercon-<lb/>ducting order parameter is basically not a s-wave. <lb/>As we know, the temperature dependence of C es may <lb/>give important information on the structure of the su-<lb/>perconducting gap. At the temperatures far below T c , <lb/>the temperature dependences of C es (T ) ∝ exp(−bT c /T ) <lb/>with b ∼ 1.5, C es (T ) ∝ T 3 and C es (T ) ∝ T 2 indicate an <lb/>isotropic BCS gap, point nodes and gap-zeroes along lines <lb/>in the superconducting gap structure, respectively [21]. <lb/>Though the extra-low temperature data is absent here <lb/>due to the experimental limitation, fitting on the present <lb/>data may give a preliminary clue. In figure 3(b), it can <lb/>be seen that the T 3 dependence best fits the C es (T ) data, <lb/>suggesting point nodes for the superconducting gap. It <lb/>should be mentioned that the T 3 dependence most favors <lb/>the data in the wide range of 13 % ≤ η ≤ 20 % (When <lb/>η ≤ 12 %, C es becomes a negative value at 1.8 K). <lb/>Based on symmetry and some preliminary experimen-<lb/>tal results, Tanaka and Hu [6] proposed spin triplet su-<lb/>perconductivity in the cobalt oxyhydrate. The p-wave <lb/>superconductivity was also suggested by Baskaran [3] for <lb/>the higher doping level. Owing to the ferromagnetic cor-<lb/>relation in the normal state, as stated above, spin-triplet <lb/>p-wave pairing is very probable. Considered the point <lb/>nodes for the superconducting gap, therefore, the gap <lb/>function will be ∆(k)=xk x +ŷk y , which shows the differ-<lb/>ence from the conclusion in the strontium ruthenate su-<lb/>perconductor [22]. Obviously, further experiments such <lb/>as NMR, neutron scattering, and µ SR will be needed to <lb/>make clearer picture for the symmetry of the supercon-<lb/>ducting order parameters. <lb/></body>
19
+
20
+ <div type="acknowledgement">The authors are indebted to H. H. Wen for the earlier <lb/>discussion. This work was supported by NSFC with the <lb/>Grant No. 10104012 and NSFC 10225417. Cao, Fang <lb/>and Xu also acknowledge the partial support (Project <lb/>No. nkbrsf-g1999064602) from the Ministry of Science <lb/>and Technology of China . <lb/></div>
21
+
22
+ <front> * Author to whom correspondence should be addressed; <lb/>Electronic address: [email protected] <lb/></front>
23
+
24
+ <listBibl>[1] K. Takada et al., Nature 422, 53 (2003). <lb/>[2] J. V. Badding, Nature Materials 2, 208 (2003). <lb/>[3] G. Baskaran, cond-mat/0303649. <lb/>[4] B. Kumar and B. S. Shastry, cond-mat/0304210. <lb/>[5] Q. H. Wang, D. H. Lee, and P. A. Lee, <lb/>cond-mat/0304377. <lb/>[6] A. Tanaka and X. Hu, cond-mat/0304409. <lb/>[7] D. J. Singh, cond-mat/0304532. <lb/>[8] M. L. Foo et al., cond-mat/0304464. <lb/>[9] H. Sakurai et al., cond-mat/0304503. <lb/>[10] B. Lorenz et al., cond-mat/0304537. <lb/>[11] G. Cao et al., unpublished. <lb/>[12] K. A. Müller et al., Phys. Rev. Lett. 58, 408 (1987). <lb/>[13] P. A. Mansky, P. M. Chaikin, and R. C. Haddon, Phys. <lb/>Rev. B 50, 15929 (1994). <lb/>[14] The transverse magneto-resistance in the normal state is <lb/>found to be proportional to the applied magnetic field, <lb/>i.e., ∆R = R(H) − R(0) ∝ H 2 . See: J. Callaway, Quan-<lb/>tum Theory of Solid State, (Academic Press, NewYork, <lb/>1991), p. 650. <lb/>[15] N. R. Werthamer, E. Helfand, and P. C. Hohenberg, <lb/>Phys. Rev. 147, 295 (1966)/R. R. Hake, Appl. Phys. <lb/>Lett. 10, 189 (1967)/T. P. Orlando et al., Phys. Rev. <lb/>B 19, 4545 (1979). <lb/>[16] Y. Ando et al., Phys. Rev. B 60, 10580 (1999). <lb/>[17] W. L. McMillan, Phys. Rev. 167, 331 (1968). <lb/>[18] D. J. Singh, Phys. Rev. B 61, 13397 (2000). <lb/>[19] G. R. Stewart, Rev. Mod. Phys. 56, 755 (1984). <lb/>[20] B. B. Goodman, C. R. Acad. Sci. 246, 3031 (1958). <lb/>[21] M. Sigrist and K. Ueda, Rev. Mod. Phys. 63, 239 (1991). <lb/>[22] Y. Maeno, T. M. Rice, and M. Sigrist, Physics Today 54, <lb/>42 (2001). </listBibl>
25
+
26
+
27
+ </text>
28
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPC0150L17-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <front>arXiv:cond-mat/0201165v1 [cond-mat.supr-con] 11 Jan 2002 <lb/>Oxygen-isotope effect on the in-plane penetration depth in underdoped <lb/>Y 1−x Pr x Ba 2 Cu 3 O 7−δ as revealed by muon-spin rotation <lb/>R. Khasanov (1,2) , A. Shengelaya (1) , K. Conder (3) , E. Morenzoni (2) , I. M. Savić (4) , and H. Keller (1) <lb/>(1) Physik-Institut der Universität Zürich, CH-8057 Zürich, Switzerland <lb/>(2) Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland <lb/>(3) Laboratory for Neutron Scattering, ETH Zürich and PSI Villigen, CH-5232 Villigen PSI, Switzerland <lb/>Faculty of Physics, University of Belgrade, 11001 Belgrade, Yugoslavia <lb/>The oxygen-isotope ( 16 O/ 18 O) effect (OIE) on the in-plane penetration depth λ ab (0) in under-<lb/>doped Y1−xPrxBa2Cu3O 7−δ was studied by muon-spin rotation. A pronounced OIE on λ −2 <lb/>ab (0) <lb/>was observed with a relative isotope shift of ∆λ −2 <lb/>ab /λ −2 <lb/>ab = −5(2)% for x = 0.3 and -9(2)% for <lb/>x = 0.4. It arises mainly from the oxygen-mass dependence of the in-plane effective mass m * <lb/>ab . <lb/>The OIE exponents of Tc and of λ −2 <lb/>ab (0) exhibit a relation that appears to be generic for cuprate <lb/>superconductors. <lb/>PACS numbers: 76.75.+i, 74.72.-h, 82.20.Tr, 71.38 <lb/></front>
9
+
10
+ <body>The pairing mechanism responsible for high-<lb/>temperature superconductivity remains elusive in <lb/>spite of the fact that many models have been pro-<lb/>posed since its discovery. A fundamental question is <lb/>whether lattice effects are relevant for the occurrence of <lb/>high-temperature superconductivity. In order to clarify <lb/>this point a large number of isotope-effect studies were <lb/>performed since 1987 [1]. The first oxygen-isotope effect <lb/>(OIE) studies on the transition temperature T c were <lb/>performed on optimally doped samples, showing no <lb/>significant isotope shift [2]. However, later experiments <lb/>revealed a small but finite dependence of T c on the <lb/>oxygen-isotope mass M O [3, 4, 5, 6], as well as on <lb/>the copper-isotope mass M Cu [7, 8]. Moreover, a <lb/>general trend in the dependence of the OIE exponent <lb/>α O = −d ln T c /d ln M O on the doping level was found <lb/>which appears to be generic for all cuprate supercon-<lb/>ductors [1, 5, 8, 9, 10]: In the underdoped region α O <lb/>is large, even exceeding the conventional BCS-value <lb/>α = 0.5 and becomes small in the optimally doped and <lb/>overdoped regime. <lb/>There is increasing evidence that a strong electron-<lb/>phonon coupling is present in cuprate superconductors, <lb/>which may lead to the formation of polarons (bare charge <lb/>carriers accompanied by local lattice distortions) [11, 12]. <lb/>One way to test this hypothesis is to demonstrate that <lb/>the effective mass of the supercarriers m * depends on <lb/>the mass M of the lattice atoms. This is in contrast to <lb/>conventional BCS superconductors, where only the &apos;bare&apos; <lb/>charge carriers condense into Cooper pairs, and m * is es-<lb/>sentially independent of M . For cuprate superconductors <lb/>(clean limit) the in-plane penetration depth λ ab is simply <lb/>given by λ −2 <lb/>ab (0) ∝ n s /m * <lb/>ab , where n s is the superconduct-<lb/>ing charge carrier density, and m * <lb/>ab is the in-plane effec-<lb/>tive mass of the superconducting charge carriers. This <lb/>implies that the OIE on λ ab is due to a shift in n s and/or <lb/>m * <lb/>ab : <lb/>∆λ −2 <lb/>ab (0)/λ −2 <lb/>ab (0) = ∆n s /n s − ∆m * <lb/>ab /m * <lb/>ab . <lb/>(1) <lb/>Therefore a possible mass dependence of m * <lb/>ab can be <lb/>tested by investigating the isotope effect on λ ab , provided <lb/>that the contribution of n s to the total isotope shift is <lb/>known. <lb/>Previous OIE studies of the penetration depth in <lb/>YBa 2 Cu 3 O 7−δ [13], La 2−x Sr x CuO 4 [10, 14, 15], and <lb/>Bi 1.6 Pb 0.4 Sr 2 Ca 2 Cu 3 O 10+δ [16] indeed showed a pro-<lb/>nounced oxygen-mass dependence on the supercarrier <lb/>mass. However, in all these experiments the penetration <lb/>depth was determined indirectly from the onset of magne-<lb/>tization [13, 16], from the Meissner fraction [10, 14], and <lb/>from magnetic torque measurements [15]. The muon-<lb/>spin rotation (µSR ) technique is a direct and accurate <lb/>method to determine the penetration depth λ in type II <lb/>superconductors. In this Letter, we report µSR measure-<lb/>ments of in-plane penetration depth λ ab in underdoped <lb/>Y 1−x Pr x Ba 2 Cu 3 O 7−δ (x = 0.3 and 0.4) with two differ-<lb/>ent oxygen isotopes ( 16 O and 18 O). A large OIE on λ ab <lb/>was observed which mainly arises from the oxygen-mass <lb/>dependence of m * <lb/>ab . <lb/>Polycrystalline samples of Y 1−x Pr x Ba 2 Cu 3 O 7−δ (x = <lb/>0.3 and x = 0.4) were prepared by standard solid state <lb/>reaction [17]. Oxygen isotope exchange was performed <lb/>during heating the samples in O 2 gas. In order to en-<lb/>sure the same thermal history of the substituted ( 18 O) <lb/>and not substituted ( O) sample, two experiments (in <lb/>O 2 and 18 O 2 ) were always performed simultaneously. <lb/>The exchange and back exchange processes were car-<lb/>ried out at 600 o C during 25 h, and then the samples <lb/>were slowly cooled (20 o C/h) in order to oxidize them <lb/>completely. The O content in the samples, as deter-<lb/>mined from a change of the sample weight after the iso-<lb/>tope exchange, was found to be 78(2)% for both sam-<lb/>ples. The total oxygen content of the samples was deter-<lb/></body>
11
+
12
+ <page>2 <lb/></page>
13
+
14
+ <body>mined using high-accuracy volumetric analysis [17]. To <lb/>examine the quality of the samples low-field (1mT, field-<lb/>cooled) SQUID magnetization measurements were per-<lb/>formed (see Fig. 1). For both concentrations the T c onset <lb/>for the 16 O samples was higher than for O with nearly <lb/>the same transition width. An oxygen back exchange of <lb/>the 18 O sample (x = 0.4) resulted within error in almost <lb/>the same magnetization curve as for the O sample, <lb/>confirming that the back exchange is almost complete. <lb/>The results of the OIE on T c are summarized in Table <lb/>I. Taking into account an isotope exchange of 78%, we <lb/>found α O = 0.22(4) for x = 0.3 and α O = 0.37(5) for <lb/>x = 0.4, in agreement with previous results [9, 18]. <lb/>50 <lb/>52 <lb/>54 <lb/>56 <lb/>58 <lb/>60 <lb/>62 <lb/>64 <lb/>-0.6 <lb/>-0.4 <lb/>-0.2 <lb/>0.0 <lb/>36 <lb/>38 <lb/>40 <lb/>42 <lb/>44 <lb/>46 <lb/>48 <lb/>50 <lb/>-0.6 <lb/>-0.4 <lb/>-0.2 <lb/>0.0 <lb/>DT <lb/>c <lb/>DT <lb/>c <lb/>= -1.3(1) K <lb/>x=0.3 <lb/>16 <lb/>O <lb/>18 <lb/>O <lb/>M <lb/>DT <lb/>c <lb/>= -1.7(1) K <lb/>x=0.4 <lb/>16 <lb/>O <lb/>O <lb/>18 <lb/>O <lb/>16 <lb/>O <lb/>M <lb/>T (K) <lb/>FIG. 1: Section near Tc of the low-field (1mT, field-cooled) <lb/>magnetization curves (normalized to the value at 10K) for <lb/>Y1−xPrxBa2Cu3O 7−δ (x = 0.3 and 0.4). <lb/>The µSR experiments were performed at the Paul <lb/>Scherrer Institute (PSI), Switzerland, using the πM3 µSR <lb/>facility. The samples consisted of sintered pellets ( 12 mm <lb/>in diameter, 3 mm thick) which were mounted on a <lb/>Fe 2 O 3 sample holder in order to reduce the background <lb/>from muons not stopping in the sample. The polycrys-<lb/>talline Y 1−x Pr x Ba 2 Cu 3 O 7−δ samples were cooled from <lb/>far above T c in a magnetic field of 200 mT perpendic-<lb/>ular to the sample disk. Time-differential µSR spec-<lb/>troscopy was employed, from which one can deduce the <lb/>probability distribution of the local magnetic field p(B) <lb/>of the vortex state by measuring the time evolution of <lb/>the muon-spin polarization [19]. In a powder sample the <lb/>magnetic penetration depth λ can be extracted from the <lb/>muon-spin depolarization rate σ(T ) ∝ 1/λ 2 (T ), which <lb/>probes the second moment ∆B 2 1/2 of p(B) in the mixed <lb/>state [19, 20]. For highly anisotropic layered supercon-<lb/>ductors (like the cuprate superconductors) λ is mainly <lb/>determinated by the in-plane penetration depth λ ab [20]: <lb/>σ(T ) ∝ 1/λ 2 <lb/>ab (T ) ∝ n s /m * <lb/>ab . <lb/>2.5 <lb/>2.0 <lb/>1.5 <lb/>1.0 <lb/>0.5 <lb/>0 <lb/>0 <lb/>20 <lb/>40 <lb/>T (K) <lb/>60 <lb/>80 <lb/>O <lb/>16 <lb/>O <lb/>18 <lb/>x = 0.3 <lb/>σ (µs ) <lb/>-1 <lb/>FIG. 2: Temperature dependence of the µSR depolarization <lb/>rate σ of Y1−xPrxBa2Cu3O 7−δ for x = 0.3, measured in a <lb/>field 200 mT (field-cooled). <lb/>The depolarization rate σ was extracted from the <lb/>µSR time spectra using a Gaussian relaxation func-<lb/>tion R(t) = exp[−σ 2 t 2 /2]. <lb/>Figure 2 shows the <lb/>temperature dependence of the measured σ for the <lb/>Y 1−x Pr x Ba 2 Cu 3 O 7−δ samples with x = 0.3. Similar re-<lb/>sults were obtained for the samples with x = 0.4. It is ev-<lb/>ident that the values of σ for 18 O are systematically lower <lb/>than those for O. As expected for a type II supercon-<lb/>ductor in the mixed state, σ continuously increases below <lb/>T c with decreasing temperature [20]. The sharp increase <lb/>of σ below ≃ 10 K is due to antiferromagnetic ordering <lb/>of the Cu(2) moments [21]. Above T c a small tempera-<lb/>ture independent depolarization rate σ nm ≃ 0.15 µs −1 is <lb/>seen, arising from the nuclear magnetic moments of Cu <lb/>and Pr. Therefore, the total σ is determined by three <lb/>contributions: a superconducting (σ sc ), an antiferromag-<lb/>netic (σ af m ), and a small nuclear magnetic dipole (σ nm ) <lb/>contribution. Because σ af m is only present at low tem-<lb/>peratures, data points below 10 K were not considered in <lb/>the analysis. The superconducting contribution σ sc was <lb/>then determined by subtracting σ nm measured above T c <lb/>from σ. In Fig. 3 we show the temperature dependence <lb/>of σ sc for the Y 1−x Pr x Ba 2 Cu 3 O 7−δ samples with x = 0.3 <lb/>and 0.4. It is evident that for both concentrations a re-<lb/>markable oxygen isotope shift on T c as well as on σ sc is <lb/>present. <lb/>The data in Fig. 3 were fitted to the power law <lb/>σ sc (T )/σ sc (0) = 1−(T /T c ) n [20] with σ sc (0) and n as free <lb/>parameters, and T c fixed. The values of T c were taken <lb/>from the magnetization measurements (see Table I). The <lb/></body>
15
+
16
+ <page>3 <lb/></page>
17
+
18
+ <body>1.5 <lb/>1.0 <lb/>0.5 <lb/>0 <lb/>1.00 <lb/>0.75 <lb/>0.50 <lb/>0.25 <lb/>0 <lb/>0 <lb/>10 <lb/>20 <lb/>30 <lb/>40 <lb/>50 <lb/>60 <lb/>T (K) <lb/>0.75 <lb/>0.50 <lb/>0.25 <lb/>0 <lb/>0.25 <lb/>0 <lb/>1.00 <lb/>1.00 <lb/>0.75 <lb/>0.50 <lb/>x = 0.3 <lb/>x = 0.4 <lb/>λ <lb/>ab (T) / <lb/>λ <lb/>ab (0) ( O) <lb/>-2 <lb/>-2 <lb/>18 <lb/>σ <lb/>sc (T) (µs ) <lb/>-1 <lb/>O <lb/>16 <lb/>18 O <lb/>18 O <lb/>O <lb/>16 <lb/>O <lb/>16 <lb/>18 O <lb/>FIG. 3: Temperature dependence of depolarization rate σsc <lb/>in Y1−xPrxBa2Cu3O 7−δ for x = 0.3 and 0.4 (200 mT, field-<lb/>cooled). On the right axis the normalized in-plane pene-<lb/>tration depth λ −2 <lb/>ab (T )/λ −2 <lb/>ab (0)( 18 O) is plotted for comparison <lb/>with Ref. [15]. The solid lines correspond to fits to the power <lb/>law σsc(T )/σsc(0) = 1 − (T /Tc) n . <lb/>values of σ sc (0) obtained from the fits are listed in Table <lb/>I and are in agreement with previous results [21]. The <lb/>exponent n was found to be n = 2.0(1) for x = 0.3 and <lb/>n = 1.5(1) for x = 0.4, which is typical for underdoped <lb/>YBCO [20]. Moreover, n is within error the same for <lb/>16 O and 18 O. This implies that σ sc has nearly the same <lb/>temperature dependence for the two isotopes (see Fig. 3). <lb/>In order to proof that the observed OIE on λ ab (0) are in-<lb/>trinsic, the 18 O sample with x = 0.4 was back exchanged <lb/>( 18 O → 16 O). As seen in Fig. 3, the data points of this <lb/>sample (cross symbols) indeed coincide with those of the <lb/>16 O sample. From the values of σ sc (0) listed in Table <lb/>I the relative isotope shift of the in-plane penetration <lb/>depth ∆λ −2 <lb/>ab (0)/λ −2 <lb/>ab (0) = [σ 18O <lb/>sc (0) − σ 16O <lb/>sc (0)]/σ 16O <lb/>sc (0) <lb/>was determined. Taking into account an isotope ex-<lb/>change of 78%, one finds ∆λ −2 <lb/>ab (0)/λ −2 <lb/>ab (0) = −5(2)% <lb/>and -9(2)% for x = 0.3 and 0.4, respectively (Table I). <lb/>For the OIE exponent β O = −d ln λ −2 <lb/>ab (0)/d ln M O , one <lb/>readily obtains β O = 0.38(12) for x = 0.3 and β O = <lb/>0.71(14) for x = 0.4 (Table I). This means that in un-<lb/>derdoped Y 1−x Pr x Ba 2 Cu 3 O 7−δ the OIE on λ −2 <lb/>ab as well <lb/>as on T c increase with increasing Pr doping x (decreas-<lb/>TABLE <lb/>I: <lb/>Summary <lb/>of <lb/>the <lb/>OIE <lb/>results <lb/>for <lb/>Y1−xPrxBa2Cu3O 7−δ extracted from the experimental <lb/>data (see text for an explanation). <lb/>O <lb/>18 O <lb/>x Tc <lb/>σsc(0) <lb/>Tc <lb/>σsc(0) <lb/>αO <lb/>∆λ −2 <lb/>ab (0) <lb/>λ −2 <lb/>ab (0) <lb/>βO <lb/>[K] <lb/>[µs −1 ] <lb/>[K] <lb/>[µs −1 ] <lb/>[%] <lb/>0.3 60.6(1) 1.63(2) 59.3(1) 1.57(2) 0.22(4) -5(2) 0.38(12) <lb/>0.4 45.3(1) 1.01(2) 43.6(1) 0.94(2) 0.37(5) -9(2) 0.71(14) <lb/>0.4 45.1(1) a 1.01(4) a <lb/>a results for the back-exchange ( 18 O→ 16 O) sample <lb/>ing T c ). This finding is in excellent agreement with <lb/>the recent magnetic torque measurements on underdoped <lb/>La 2−x Sr x CuO 4 [15]. <lb/>According to Eq. (1) the observed ∆λ −2 <lb/>ab (0)/λ −2 <lb/>ab (0) is <lb/>due to a shift of n s and/or m * <lb/>ab . For La 2−x Sr x CuO 4 <lb/>several independent experiments [10, 14, 15] have shown <lb/>that the change of n s during the exchange proce-<lb/>dure is negligibly small. <lb/>In the present work we <lb/>provide further evidence: (i) The fully oxygenated <lb/>Y 1−x Pr x Ba 2 Cu 3 O 7−δ samples (δ ≃ 0) were all prepared <lb/>under identical conditions, either in a 16 O 2 or 18 O 2 at-<lb/>mosphere [17], and the Pr content x did not change. <lb/>It is very unlikely that n s changes significantly upon <lb/>18 O substitution, and after the back-exchange ( 18 O→ <lb/>16 O) the same results are obtained (see Figs. 1, 3 and <lb/>Table I). (ii) According to a model [22] that describes <lb/>the suppression of T c in Y 1−x Pr x Ba 2 Cu 3 O 7−δ , the num-<lb/>ber of supercarriers decreases linearly with increasing <lb/>x in the range of 0.05 &lt; x &lt; 0.5, and consequently <lb/>∆n s /n s = −∆x/x. Moreover, for 0.1 &lt; x &lt; 0.5 the <lb/>transition temperature T c decreases linearly with x, with <lb/>∆T c /∆x ≃ −150 K/Pr atom [9]. Combining this two <lb/>relations one obtains: ∆T c ≃ −150 • x • ∆n s /n s . As-<lb/>suming that the observed OIE on λ −2 <lb/>ab is only due to a <lb/>change of n s (∆m * <lb/>ab /m * <lb/>ab ≃ 0), one can estimate the cor-<lb/>responding shift of T c . For x = 0.3 and x = 0.4 one finds <lb/>∆T c ≃ −1.8(4) K and −4.2(6) K, respectively. The ex-<lb/>perimental values, however, are much lower (see Fig. 1): <lb/>∆T c = −1.3(1) K (x = 0.3) and ∆T c = −1.7(1) K <lb/>(x = 0.4). We thus conclude that any change in n s dur-<lb/>ing the exchange procedure must be small, and that the <lb/>change of λ ab is mainly due to the OIE on the in-plane <lb/>effective mass m * <lb/>ab with ∆m * <lb/>ab /m * <lb/>ab ≃ 5(2) % and 9(2) % <lb/>for x = 0.3 and x = 0.4, respectively. This implies that <lb/>the effective supercarrier mass m * <lb/>ab in this cuprate su-<lb/>perconductor depends on the oxygen mass of the lattice <lb/>atoms, which is not expected for a conventional phonon-<lb/>mediated BSC superconductor. <lb/>In Fig. 4 the exponent β O versus the exponent α O for <lb/>Y 1−x Pr x Ba 2 Cu 3 O 7−δ is plotted. For comparison we <lb/>also included the recent magnetic torque results of un-<lb/>derdoped La 2−x Sr x CuO 4 [15]. It is evident that these <lb/></body>
19
+
20
+ <page>4 <lb/></page>
21
+
22
+ <body>exponents are linearly correlated: β O = A • α O + B. <lb/>A best fit yields A = 1.8(4) and B = −0.01(12), so <lb/>that β O ≃ A • α O . This empirical relation appears to <lb/>be generic for cuprate superconductors. Quantitatively <lb/>one can understand this behavior in terms of an empiri-<lb/>cal relation between T c and the µSR depolarization rate <lb/>σ sc (0) [23, 24]. It was shown [24] that for most fami-<lb/>lies of cuprate superconductors the simple parabolic re-<lb/>lation T c = 2σ(1 − σ/2) describes the experimental data <lb/>rather well (here T c = T c /T m <lb/>c , σ = σ sc (0)/σ m <lb/>sc (0), and <lb/>T m <lb/>c and σ m <lb/>sc (0) are the transition temperature and de-<lb/>polarization rate of the optimally doped system). Using <lb/>this parabolic Ansatz, one readily obtains the linear re-<lb/>lation between β O and α O : β O /α O = 1 + 1/2 [(1 − (1 − <lb/>T c ) 1/2 )/(1 − T c ) 1/2 ]. In the heavily underdoped regime <lb/>0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1.0 <lb/>0 <lb/>0.1 <lb/>0.2 <lb/>0.3 <lb/>0.4 <lb/>0.5 <lb/>α O <lb/>β <lb/>O <lb/>Y 1-x Pr x Ba 2 Cu 3 O 7 (µSR) <lb/>La 2-x Sr x CuO 4 (torque) <lb/>FIG. 4: Plot of the OIE exponents βO versus αO for <lb/>Y1−xPrxBa2Cu3O 7−δ (x = 0.3 and 0.4) and La2−xSrxCuO4 <lb/>(x = 0.080 and 0.086) from [15]. The dashed line represents <lb/>a best fit to the empirical relation βO = A • αO + B. <lb/>(T c → 0) β O /α O → 1. For the underdoped samples <lb/>shown in Fig. 4 the reduced critical temperature T c is <lb/>in the range 0.5 to 0.7, yielding β O /α O = 1.2 − 1.4, in <lb/>agreement with A = 1.8(4) obtained from the experi-<lb/>mental data. Very recently, it was pointed out [25] that <lb/>the unusual doping dependence of the OIE on T c and on <lb/>λ −2 <lb/>ab (0) naturally follows from the doping driven 3D-2D <lb/>crossover and the 2D quantum superconductor to insu-<lb/>lator transition in the underdoped limit. It is predicted <lb/>that in the underdoped regime β O /α O → 1, which is <lb/>consistent with the parabolic Ansatz. <lb/>In summary, we performed µSR measurements of <lb/>the in-plane penetration depth λ ab in underdoped <lb/>Y 1−x Pr x Ba 2 Cu 3 O 7−δ (x = 0.3, 0.4) for samples with two <lb/>different oxygen isotopes ( 16 O and 18 O). A pronounced <lb/>OIE on both the transition temperature T c and λ −2 <lb/>ab (0) <lb/>was observed, which increases with decreasing T c . The <lb/>isotope shift on λ −2 <lb/>ab (0) is attributed to a shift in the in-<lb/>plane effective mass m * <lb/>ab . For x = 0.3 and 0.4 we find <lb/>∆m * <lb/>ab /m * <lb/>ab = −5(2)% and -9(2)%, respectively. Further-<lb/>more, an empirical relation between the OIE exponents <lb/>β O and α O was found that appears to be generic for vari-<lb/>ous classes of cuprate superconductors. The OIE on m * <lb/>ab <lb/>implies that the superconducting carriers have polaronic <lb/>character, and that lattice effects play an essential role in <lb/>the occurrence of high-temperature superconductivity. <lb/></body>
23
+
24
+ <div type="acknowledgment">We are grateful to G.M. Zhao, T. Schneider, and <lb/>K.A. Müller for many fruitful discussions and to A. Am-<lb/>ato, U. Zimmermann, and D. Herlach from PSI for tech-<lb/>nical support during the µSR experiments. This work <lb/>was partly supported by the Swiss National Science Foun-<lb/>dation.</div> <lb/>
25
+
26
+ <listBibl>[1] For a review, see J.P. Franck, in Physical Properties High <lb/>Temperature Superconductors IV, edited by D.M Gins-<lb/>berg (World Scientific, Singapore, 1994), pp. 189-293. <lb/>[2] B. Batlogg et al., Phys. Rev. Lett. 58, 2333 (1987). <lb/>[3] B. Batlogg et al., Phys. Rev. Lett. 59, 912 (1987). <lb/>[4] M. Cardona et al., Solid State Commun. 67, 789 (1988). <lb/>[5] M.K. Crawford at al., Phys. Rev. B 41, 282 (1990). <lb/>[6] D. Zech et al., Nature (London) 371, 681 (1994). <lb/>[7] J.P. Franck, S. Harker, and J.H. Brewer, Phys. Rev. Lett. <lb/>71, 283 (1993). <lb/>[8] G.M. Zhao et al., Phys. Rev. B 54, 14956 (1996). <lb/>[9] J.P. Franck, J. Jung, and M. A.-K. Mohamed, Phys. Rev. <lb/>B 44, 5318 (1991). <lb/>[10] G.M. Zhao et al., J. Phys.: Condens Matter 10, 9055 <lb/>(1998). <lb/>[11] A.S. Alexandrov and N.F. Mott, Int. J. Mod. Phys. 8, <lb/>2075 (1994). <lb/>[12] K.A. Müller, Physica C 341-348, 11 (2000). <lb/>[13] G.M. Zhao and D.E. Morris, Phys. Rev. 51, 16487 <lb/>(1995). <lb/>[14] G.M. Zhao et al., Nature (London) 385, 236 (1997). <lb/>[15] J. Hofer et al., Phys. Rev. Lett. 84, 4192 (2000). <lb/>[16] G.M. Zhao et al., Phys. Rev. B 63, 220506-1 (2001). <lb/>[17] K. Conder, Mater. Sci. Eng. R32, 41-102 (2001). <lb/>[18] G.M. Zhao et al., Phys. Rev. B 54, 14982 (1996). <lb/>[19] S.L. Lee, in Muon Science, ed. by S.L. Lee, S.H. Kilcoyne, <lb/>and R. Cywinski, IOP Publishing, Bristol and Philadel-<lb/>phia, pp. 149-171 (1999), and references therein. <lb/>[20] P. Zimmerman et al., Phys. Rev. B 52, 541 (1995). <lb/>[21] C.L. Seaman et al., Phys. Rev. B 42, 6801 (1990). <lb/>[22] A. Liechtenstien and I. Mazin, Phys. Rev. Lett. 74, 1000 <lb/>(1995). <lb/>[23] Y. Uemura et al., Phys. Rev. Lett. 62, 2317 (1989). <lb/>[24] T. Schneider and H. Keller, Phys. Rev. Lett. 69, 3374 <lb/>(1992). <lb/>[25] T. Schneider and H. Keller, Phys. Rev. Lett. 86, 4899 <lb/>(2001) . </listBibl>
27
+
28
+
29
+ </text>
30
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPS0710709-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>arXiv:cond-mat/0112448v1 [cond-mat.supr-con] 25 Dec 2001 <lb/>Quasiparticle Excitation in the Superconducting Pyrochlore <lb/>Cd 2 Re 2 O 7 Probed by Muon Spin Rotation <lb/>Ryosuke Kadono * , Wataru Higemoto, Akihiro Koda, Yu Kawasaki 1 , Masashi <lb/>Hanawa 2 , and Zenji Hiroi 2 <lb/>Institute of Materials Structure Science, High Energy Accelerator Research Organization (KEK), <lb/>Tsukuba, Ibaraki 305-0801 <lb/>Graduate School of Engineering Science, Osaka University, Toyonaka, Osaka 560-8531 <lb/>2 Institute for Solid State Physics, University of Tokyo, Kashiwa, Chiba 277-8581 <lb/>(Received September 12, 2001) <lb/>Abstract <lb/>The quasiparticle excitations in the mixed state of Cd 2 Re 2 O 7 have been <lb/>studied by means of muon spin rotation/relaxation (µSR). The tempera-<lb/>ture dependence of the magnetic penetration depth (λ) is consistent with <lb/>a nearly isotropic superconducting order parameter, although a slight dis-<lb/>crepancy which is dependent on the details in the analysis may be present. <lb/>This is also supported by the relatively weak field dependence of λ. <lb/> * Also at School of Mathematical and Physical Science, The Graduate University for Advanced <lb/>Studies <lb/></front>
8
+
9
+ <page>1 <lb/></page>
10
+
11
+ <body>A class of metal oxides isostructural to mineral pyrochlore has been attracting consid-<lb/>erable attention because they exhibit a wide variety of interesting physical properties. [1] <lb/>The pyrochlore has a general formula of A 2 B 2 O 7 , consisting of BO 6 octahedra and eightfold <lb/>coordinated A cations, where A and B are transition metals and/or rare-earth elements. In <lb/>particular, the B sublattice can be viewed as a three-dimensional network of corner-sharing <lb/>tetrahedra, providing a testing ground for studying the role of geometrical frustration in sys-<lb/>tems which have local spins at B sites with antiferromagnetic (AFM) correlation. [2] Such <lb/>systems are known to remain frustrated even when the exchange interaction is ferromagnetic <lb/>(FM), provided that the spin correlation has local Ising anisotropy. [3] Recent studies have <lb/>revealed a rich variety of phenomena seemingly related to the geometrical frustration, such <lb/>as the occurrence of a spin-glass (SG) phase in R 2 Mo 2 O 7 with R=Y, Tb and Dy [4,5], the <lb/>unusual behavior of ordinary and anomalous Hall coefficients in the same compound with <lb/>R=Nd, Sm and Gd [6,7], and the &quot;spin-ice&quot; phase in R Ti 2 O 7 with R=Dy and Ho. [3,8] <lb/>Although metallic pyrochlores comprise a minority subgroup of the pyrochlore family, <lb/>they consist of distinct members such as Tl 2 Mn 2 O 7 , which exhibits colossal magnetoresis-<lb/>tance. [9,10] Moreover, the recently revealed superconductivity in Cd 2 Re 2 O 7 , a 5d transition <lb/>metal pyrochlore [11,12], demonstrates that the pyrochlores provide a fertile field for elec-<lb/>tronic correlation adjacent to the perovskite compounds. In this context, it is noteworthy <lb/>that LiV 2 O 4 , a cubic spinel compound in which the V sublattice is isostructural to the B <lb/>sublattice in pyrochlore, behaves similarly to a heavy fermion metal. [13] <lb/>It is reported that Cd 2 Re 2 O 7 falls into the bulk superconducting state below T c ≃ 1 ∼ <lb/>2 K, as confirmed by a large jump of specific heat ∆C c as well as large diamagnetism <lb/>due to the Meissner effect associated with the occurrence of zero resistivity. [11] The dc <lb/>magnetization curve indicates that the superconductivity is of type II with the upper critical <lb/>field close to 0.29 T at 0 K. [14] The ratio ∆C c /γT c (with γ being the Sommerfeld constant) <lb/>is reported to be 1.15, which is smaller than the predicted value of 1.43 for isotropic BCS <lb/>superconductors. Unfortunately, these measurements were performed only above 0.4 K and <lb/>are thus inconclusive in determining the detailed characteristics of superconductivity in <lb/>
12
+
13
+ Cd Re 2 O 7 . In this Letter, we report on the quasiparticle excitations in the mixed state of <lb/>Cd 2 Re 2 O 7 studied by muon spin rotation/relaxation (µSR). The magnetic penetration depth <lb/>λ, which reflects the population of normal electrons (&quot;quasiparticles&quot;) in the superconductive <lb/>state, is determined microscopically by measuring the muon spin relaxation due to the <lb/>spatial inhomogeneity of magnetic induction in the flux line lattice (FLL). We show that <lb/>the temperature dependence of λ is more or less consistent with the prediction based on the <lb/>BCS superconductors with an isotropic gap, although a slight discrepancy is suggested by <lb/>the detailed analysis. <lb/>The µSR experiments on both single-crystal and polycrystalline Cd 2 Re 2 O 7 were per-<lb/>formed on the M15 beamline at the TRIUMF muon facility which provides a beam of nearly <lb/>100% spin-polarized positive muons of momentum 28.6 MeV/c. The specimen was mounted <lb/>on the coldfinger of a He-4 He dilution refrigerator and cooled from a temperature above <lb/>T c after setting the magnetic field at every field point (i.e., field-cooling) to eliminate the <lb/>effect of flux pinning. The field and temperature scan data were obtained at T = 0.2 K and <lb/>at H = 0.1 T, respectively. Muons were implanted into the specimen (measuring about 10 <lb/>mm×10 mm and 1 mm thick) after being passed through a beam collimator. The initial <lb/>muon spin polarization was perpendicular to the magnetic field H and thus to the FLL in <lb/>the superconducting state. <lb/>Since the muons stop randomly along the length scale of the FLL, the muon spin pre-<lb/>cession signalP (t) provides a random sampling of the internal field distribution B(r), <lb/>P (t) ≡ P x (t) + iP y (t) = <lb/>∞ <lb/>−∞ <lb/>n(B) exp(iγ µ Bt)dB, <lb/>(0.1) <lb/>n(B) = <lb/>dr <lb/>dB <lb/>, <lb/>(0.2) <lb/>where γ µ is the muon gyromagnetic ratio (=2π×135.53 MHz/T), and n(B) is the spectral <lb/>density for muon precession determined by the local field distribution. These equations indi-<lb/>cate that the real amplitude of the Fourier-transformed muon precession signal corresponds <lb/>to the local field distribution n(B). The local field distribution can be approximated as the <lb/>sum of magnetic induction from isolated vortices in the London model to yield <lb/></body>
14
+
15
+ <page>3 <lb/></page>
16
+
17
+ <body>B(r) = B 0 <lb/>K <lb/>e −iK•r e −K 2 ξ 2 <lb/>v <lb/>1 + K 2 λ 2 + O(K 2 <lb/>x , K 2 <lb/>y ) <lb/>, <lb/>(0.3) <lb/>where K is a translation of the vortex reciprocal lattice, B 0 (≃ H) is the average internal <lb/>field, λ is the London penetration depth, and ξ v is the cutoff parameter. The term O(K 2 <lb/>x , K 2 <lb/>y ) <lb/>denotes the nonlocal effect in which the electromagnetic response kernel Q(K) generating <lb/>the supercurrent around the vortex depends on K. While this term is eliminated in the <lb/>conventional BCS superconductors with isotropic s-wave pairing, it becomes important for <lb/>the moe complex order parameters such as anisotropic s-wave or d-wave (e.g., d x −y 2 ). The <lb/>London penetration depth in the FLL state is related to the second moment ∆B 2 = <lb/>(B(r) − B 0 ) 2 of the field distribution reflected in the µSR line shape (where <lb/>means the <lb/>spatial average). In polycrystalline samples, a Gaussian distribution of local fields is a good <lb/>approximation, whereP <lb/>(t) ≃ exp(−σ 2 t 2 /2) exp(iγ µ Ht), <lb/>(0.4) <lb/>σ = γ µ ∆B 2 . <lb/>(0.5) <lb/>For the case of ideal triangular FLL with isotropic effective carrier mass m * and a cutoff <lb/>K ≈ 1.4/ξ v provided by the numerical solution of the Ginsburg-Landau theory, the London <lb/>penetration depth λ (with O(K 2 <lb/>x , K 2 <lb/>y ) = 0) can be deduced from σ using the following <lb/>relation [15-17], <lb/>σ [µs −1 ] = 4.83 × 10 4 (1 − h)λ −2 [nm], <lb/>(0.6) <lb/>where h = H/H c2 . While the above form is valid for h &lt; 0.25 or h &gt; 0.7, a more useful <lb/>approximation valid for an arbitrary field is [17] <lb/>σ [µs −1 ] = 4.83 × 10 4 (1 − h)[1 + 3.9(1 − h) 2 ] 1/2 λ −2 [nm]. <lb/>(0.7) <lb/>In both cases, λ is related to the superconducting carrier density n s as <lb/>λ 2 = <lb/>m * c 2 <lb/>4πn s e 2 , <lb/>(0.8) <lb/></body>
18
+
19
+ <page>4 <lb/></page>
20
+
21
+ <body>indicating that λ is enhanced upon the reduction of n s due to the quasiparticle excitations. <lb/>For simplicity, we adopt eq. (0.6) for the following analysis. <lb/>In a preliminary analysis, we found that the spin relaxation rate due to the FLL formation <lb/>is less than 0.1 µs −1 which is typically of the same order of magnitude as that due to <lb/>static random local fields from nuclear magnetic moments. This means that the additional <lb/>relaxation due to 111,113 Cd and 185,187 Re nuclear moments must be considered for the proper <lb/>estimation of λ in Cd 2 Re 2 O 7 . To this end, the following equation was used in the actual <lb/>fitting analysis of the time spectra, <lb/>P (t) = exp[−(σ 0 t) ν + σ 2 t 2 /2)] exp(iγ µ Ht), <lb/>(0.9) <lb/>where σ 0 is the relaxation rate due to the nuclear moments and ν is the power of relaxation. <lb/>The parameters σ 0 and ν were evaluated by fitting the time spectra above T c with σ set to <lb/>zero, yielding σ 0 ≃ 0.057(1) µs −1 and ν ≃ 1.0 in the polycrystalline specimen at H = 0.1 T. <lb/>Then, σ due to the formation of FLL below T c was deduced by analyzing data with σ 0 and <lb/>ν being fixed to the above values. A similar analysis was performed for the data obtained <lb/>for the single crystals. <lb/>Figure 1(a) shows the temperature dependence of σ in Cd 2 Re 2 O 7 at H = 0.1 T. Upon the <lb/>onset of FLL formation, σ exhibits a gradual increase with decreasing temperature just below <lb/>T c (0.1 T) ∼ 0.7 K at this field. According to the empirical two-fluid model approximately <lb/>valid for conventional BCS superconductors, we have <lb/>λ(t) = λ(0) <lb/>1 <lb/>√ <lb/>1 − t 4 , <lb/>(0.10) <lb/>which leads to <lb/>σ(t) = σ(0)(1 − t 4 ), <lb/>(0.11) <lb/>where t ≡ T /T c (0.1 T). The fitting analysis by the same formula with an arbitrary power, <lb/>σ(t) = σ(0)(1 − t β ), <lb/>(0.12) <lb/></body>
22
+
23
+ <page>5 <lb/></page>
24
+
25
+ <body>with T c as a free parameter yields β = 2.8(5) and T c = 0.65(3) K. We also found that <lb/>T c = 0.62(1) K when β = 4 is assumed, which is slightly lower than the value T c ≃ 0.7 K <lb/>estimated from the specific heat measurement for this field. Although the difference is not <lb/>obvious between these cases (solid curve for β = 2.8 and the dashed curve for β = 4) in <lb/>Fig. 1(a), we note that the reduced χ 2 for the former is almost two times smaller (better) than <lb/>the latter. Taking H c2 = 0.29 T, the penetration depth extrapolated to T = 0 (λ(0, 0.1 T)) <lb/>is 700(8) nm, leading to the Ginsburg-Landau parameter κ ≃ 21 with ξ ≃ 34 nm estimated <lb/>from H c2 (0). Thus, it is concluded that Cd 2 Re 2 O 7 is a typical type II superconductor with <lb/>large λ, which is consistent with the results of magnetization measurement. [11] <lb/>The finding that β ≃ 2.8 may indicate that the deviation ∆λ = λ(t)−λ(0) ∝ T β exhibits <lb/>a tendency predicted for the case of line nodes (d-wave pairing) with some disorder (i.e., <lb/>dirty limit), where ∆λ ∝ T 2 . [18] Such a temperature dependence has actually been observed <lb/>in high-T c (YBCO) cuprates. [19,20] Compared with the case of isotropic gap ∆k = ∆ 0 , the <lb/>quasiparticle excitations are enhanced along nodes (|∆k| = 0) to reduce average n s , leading <lb/>to the enhancement of λ. However, the result of the fitting analysis also suggests that the <lb/>discrepancy may be attributed to experimental uncertainty including the precise value of <lb/>T c . Thus, we are led to conclude that the order parameter in Cd 2 Re 2 O 7 is mostly isotropic <lb/>with a possibility of residual weak anisotropy as suggested by the slightly reduced value of <lb/>β. <lb/>As shown in Fig. 2(a), σ decreases with increasing external field, where the general <lb/>field dependence is determined by the increasing contribution of normal vortex cores and <lb/>the stronger overlap of field distribution around the cores which are described by the term <lb/>(1−h) in eq.(0.6). The fitting analysis by eq.(0.6) with H c2 and λ(H = 0) as free parameters <lb/>yields H c 2 = 0.37(5) T and λ(H = 0) = 796(12) nm at 0.2 K. Since the deduced value of H c2 <lb/>is consistent with that obtained from the magnetization measurement (≃ 0.29 T), we can <lb/>conclude that the observed field dependence of λ is mostly due to the vortex core/overlap <lb/>effect, except for the fields below ∼0.06 T where a slightly steeper reduction of σ is suggested. <lb/>The penetration depth deduced for each field using eq.(0.6) with H c2 = 0.37 T is plotted in <lb/>Fig. 2(b). In order to evaluate the relative strength of the pair-breaking effect, it is useful <lb/>to introduce a dimensionless parameter η to describe the field dependence of λ with the <lb/>following simple linear relation, <lb/>λ = λ(0.2K, h)[1 + ηh], <lb/>(0.13) <lb/>where h ≡ H/H c2 (0.2K) with H c2 (0.2K) = 0.37 T. From the analysis of data in Fig. 2(b) by <lb/>eq. (0.13), we find that η = 0.38(14) with λ(0.2K, 0) = 741(5) nm for 0 ≤ H ≤ 0.06 T. <lb/>In general, the field dependence of λ is enhanced by two different mechanisms, i.e., the <lb/>nonlinear effect in the semiclassical Doppler shift of the quasiparticle energy levels due to <lb/>the supercurrent around the vortex cores, and the nonlocal effect which further modifies <lb/>the quasiparticle excitation spectrum in the momentum space. In particular, the nonlocal <lb/>effect is important in the system with line nodes because the coherence length is inversely <lb/>proportional to the order parameter, such that ξ 0 (k) =hv F /π∆k. The divergence of ξ 0 along <lb/>the nodal directions |∆k| = 0 means that the response of quasiparticles near the nodes is <lb/>highly nonlocal. <lb/>It is predicted that η ≪ 1 for the isotropic s-wave pairing because the finite gap prevents <lb/>the shifted levels of quasiparticle excitations from being occupied at low temperatures. This <lb/>is supported, for example, by the recent observation in CeRu , in which η ≃ 0 over the <lb/>field region 0 ≤ h ≤ 0.5 where the system behaves more or less as a conventional BCS <lb/>superconductor with isotropic s-wave pairing. [21] On the other hand, stronger field depen-<lb/>dence is predicted for the case of d-wave or anisotropic s-wave pairing due to the excess <lb/>population of quasiparticles in the region where |∆k| is small or zero. Typical examples <lb/>for the d-wave pairing are those of high-T c cuprates in which η is reported to be 5∼6.6 for <lb/>YBCO. [22] Meanwhile, in the case of an s-wave superconductor YNi 2 B 2 C in which strong <lb/>anisotropy for ∆k is suggested experimentally [23], η ≃ 1 is reported from a µSR study. [24] <lb/>The comparison of these earlier results with our result suggests that the anisotropy of the <lb/>order parameter in Cd 2 Re 2 O 7 is considerably smaller than YNi 2 B 2 C. <lb/>Finally, we comment that the recent observation of a clear coherence peak below T c in <lb/></body>
26
+
27
+ <page>7 <lb/></page>
28
+
29
+ <body>the 187 Re NQR measurement [25] does not necessarily indicate the absence of anisotropy <lb/>in the order parameter. In the case of the anisotropic energy gap, the magnitude of the <lb/>coherence peak depends on the mean free path l. The coherence peak is enhanced in a <lb/>certain condition of l where quasiparticles probe only a limited region of the Fermi surface. <lb/>Thus, while the actual value of l is difficult to estimate because of the semimetallic character <lb/>of this compound, the NQR result does not completely rule out the presence of anisotropy <lb/>in general. <lb/>In summary, we have investigated the quasiparticle excitations in the mixed state of <lb/>Cd 2 Re 2 O 7 by µSR. The temperature and field dependence of the London penetration depth <lb/>indicates that the basic feature is consistent with the isotropic order parameter for BCS <lb/>s-wave pairing, although there remains a certain subtlety suggested by the small deviation <lb/>from the theoretical prediction which may be better understood by considering a weak <lb/>anisotropy. <lb/></body>
30
+
31
+ <div type="acknowledgment">We would like to thank the staff of TRIUMF for their technical support during the <lb/>experiment. This work was partially supported by a Grant-in-Aid for Scientific Research <lb/>on Priority Areas and a Grant-in-Aid for Creative Scientific Research from the Ministry of <lb/>Education, Culture, Sports, Science and Technology of Japan. <lb/></div>
32
+
33
+ <page>8 <lb/></page>
34
+
35
+ <listBibl>REFERENCES <lb/>[1] M. A. Subramanian, G. Aravamudan and G. V. Subba Rao: Prog. Solid St. Chem. <lb/>(1983) 55 . <lb/>[2] A. P. Ramirez: Annu. Rev. Mater. Sci. 24 (1994) 453. <lb/>[3] M. J. Harris, S. T. Bramwell, D. F. McMorrow, T. Zeiske and K. W. Godfrey: Phys. <lb/>Rev. Lett. 79 (1997) 2554. <lb/>[4] S. R. Dunsiger, R. F. Kiefl, K. H. Chow, B. D. Gaulin, M. J. P. Gingras, J. E. Greedan, <lb/>A. Keren, K. Kojima, G. M. Luke, W. A. MacFarlane, N. P. Raju, J. E. Sonier, Y. J. <lb/>Uemura and W. D. Wu: Phys. Rev. B 54 (1996) 9019. <lb/>[5] J. S. Gardner, B. D. Gaulin, S. H. Lee, C. Broholm, N. P. Raju and J. E. Greedan: <lb/>Phys. Rev. Lett. (1999) 211. <lb/>[6] Y. Taguchi and Y. Tokura: Phys. Rev. B 60 (1999) 10280. <lb/>[7] T. Katsufuji, H. Y. Hwang and S.-W. Cheong: Phys. Rev. Lett. 84 (2000) 1998. <lb/>[8] A. P. Ramirez, A. Hayashi, R. J. Cava, R. Siddharthan and B. S. Shastry: Nature <lb/>(1999) 333. <lb/>[9] Y. Shimakawa, Y. Kubo and T. Manako: Nature (1996) 53. <lb/>[10] Y. Shimakawa, Y. Kubo, T. Manako, Y. V. Sushko, D. N. Argyriou and J. D. Jorgensen: <lb/>Phys. Rev. B 55 (1999) 6399. <lb/>[11] M. Hanawa, Y. Muraoka, T. Tayama, T. Sakakibara, J. Yamaura and Z. Hiroi: Phys. <lb/>Rev. Lett. 87 (2001) 187001. <lb/>[12] R. Jin, J. He, S. McCall, C. S. Alexander, F. Drymiotis and D. Mandrus: Phys. Rev. B <lb/>64 (2001) 180503. <lb/>[13] S. Kondo, D. C. Johnston, C. A. Swenson, F. Borsa, A. V. Mahajan, L. L. Miller, T. <lb/>Gu, A. I. Goldman, M. B. Maple, D. A. Gajewski, E. J. Freeman, N. R. Dilley, R. P. <lb/>Dickey, J. Merrin, K. Kojima, G. M. Luke, Y. J. Uemura, O. Chmaissem and J. D. <lb/>Jorgensen: Phys. Rev. Lett.78 (1997) 3729. <lb/>[14] Z. Hiroi and M. Hanawa: to be published in J. Phys. Chem. Solid. <lb/>[15] P. Pincus, A.C. Gossard, V. Jaccarino and J.H. Wernick: Phys. Rev. Lett. 13 (1964) <lb/>21. <lb/>[16] G. Aeppli, R. J. Cava, E. J. Ansaldo, J. H. Brewer, S. R. Kreitzman, G. M. Luke, D. R. <lb/>Noakes and R. F. Kiefl: Phys. Rev. B 35 (1987) 7129. <lb/>[17] E. H. Brandt: Phys. Rev. B 37 (1988) 2349. <lb/>[18] P. J. Hirschfeld and N. Goldenfeld: Phys. Rev. B 48 (1993) 4219. <lb/>[19] D. A. Bonn, S. Kamal, K. Zhang, R. Liang, D. J. Baar, E. Klein and W. N. Hardy: <lb/>Phys. Rev. B 50 (1994) 4051. <lb/>[20] J. E. Sonier, J. H. Brewer, R. F. Kiefl, D. A. Bonn, S. R. Dunsiger, W. N. Hardy, R. <lb/>Liang, W. A. MacFarlane, R. I. Miller and T. M. Riseman: Phys. Rev. Lett. (1997) <lb/>2875. <lb/>[21] R. Kadono, W. Higemoto, A. Koda, K. Ohishi, T. Yokoo, J. Akimitsu, M. Hedo, Y. <lb/>Inada, Y. Onuki, E. Yamamoto and Y. Haga: Phys. Rev. B (2001) 224520. <lb/>[22] J. E. Sonier, R. F. Kiefl, J. H. Brewer, D. A. Bonn, S. R. Dunsiger, W. N. Hardy, R. <lb/>Liang, W. A. MacFarlane, T. M. Riseman, D. R. Noakes and C. E. Stronach: Phys. <lb/>Rev. B 55 (1997) 11789. <lb/>[23] T. Yokoya, T. Kiss, T. Watanabe, S. Shin, M. Nohara, H. Takagi and T. Oguchi: Phys. <lb/>Rev. Lett. 85 (2000) 4952. <lb/>[24] K. Ohishi, K. Kakuta, J. Akimitsu, W. Higemoto, R. Kadono, J. E. Sonier, A. N. Price, <lb/>R. I. Miller, R. F. Kiefl, M. Nohara, H. Suzuki and H. Takagi: submitted to Phys. Rev. <lb/></listBibl>
36
+
37
+ <page>10 <lb/></page>
38
+
39
+ <listBibl>B. <lb/>[25] O. Vyaselev, K. Kobayashi, K. Arai, J. Yamazaki, K. Kodama, M. Takigawa, M. Hanawa <lb/>and Z. Hiroi: to be published in J. Phys. Chem. Solid. <lb/></listBibl>
40
+
41
+ <body>FIGURES <lb/>FIG. 1. Temperature dependence of a) the muon spin relaxation rate σ due to flux line lattice <lb/>and b) the magnetic penetration depth λ, where solid circles are the data from the polycrystalline <lb/>specimen and open triangles are those from single crystals. Solid curves are results of fitting by a <lb/>relation σ ∝ 1/λ ∝ 1 − (T /T c ) β with β and T c being free parameters. Dashed curves are obtained <lb/>when β = 4 with T c as a free parameter. <lb/>FIG. 2. Magnetic field dependence of a) the muon spin relaxation rate σ due to flux line <lb/>lattice and b) the magnetic penetration depth λ. The solid line in a) is a result of fitting by <lb/>eq.(0.6) with H c2 being a free parameter. Dashed lines are obtained by fitting data with a relation <lb/>λ ∝ 1 + η(H/H c2 ) only for 0 ≤ H ≤ 0.06 T. <lb/></body>
42
+
43
+ <page>12 </page>
44
+
45
+
46
+ </text>
47
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPS0722453-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <front>59 Co-NMR Knight Shift of Superconducting Na x CoO 2 ⋅yH 2 O <lb/>Yoshiaki Kobayashi, Mai Yokoi and Masatoshi Sato <lb/>Department of Physics, Division of Material Science, Nagoya University, Furo-cho, Chikusa-ku <lb/>Nagoya, 464-8602 Japan <lb/>(Received June 12, 2003) <lb/>Layered Co oxide Na x CoO 2 ⋅yH 2 O with the superconducting transition temperature T c =4.5 K has <lb/>been studied by 59 Co-NMR. The Knight shift K estimated from the observed spectra for powder <lb/>sample exhibits almost temperature(T)-independent behavior above T c and decreases with decreasing T <lb/>below T c . This result and the existence of the coherence peak in the spin-lattice-relaxation-rate versus <lb/>T curve reported by the present authors indicate, naively speaking, that the singlet order parameter <lb/>with s-wave symmetry is realized in Na x CoO 2 ⋅yH 2 O. Differences of the observed behaviors between <lb/>the spectra of the non-aligned sample and the one aligned in epoxy adhesive by applying the external <lb/>magnetic field are discussed. <lb/>corresponding author: M. Sato (e-mail: [email protected]) <lb/>Keywords: Na x CoO 2 ⋅yH 2 O, superconductivity, NMR, Knight shift <lb/></front>
9
+
10
+ <page>2 <lb/></page>
11
+
12
+ <body>The recent discovery of superconductivity in the layered cobalt oxide Na x CoO 2 ⋅yH 2 O 1) has attracted <lb/>much attention, because it has the triangular lattice of Co atoms. The system can be obtained from <lb/>Na x CoO 2 by deintercalating Na + ions and simultaneous intercalating H 2 O molecules. The <lb/>superconducting transition temperature T c was reported to be ~5 K. Up to now, v arious kinds of <lb/>experimental and theoretical studies have been carried out to investigate the mechanism of the <lb/>superconductivity 1-16) . <lb/>In the previous report, the present authors ha ve shown by 59 Co-NQR studies of the system, that the <lb/>coherence peak of the 59 Co-nuclear spin relaxation rate 1/T 1 exists just below T c . It has been also <lb/>pointed out, based on detailed comparison of the relaxation rates 1/T 1 and the magnetic susceptibilities <lb/>χ between Na x CoO 2 ⋅yH 2 O and Na x CoO 2 that the former system is closer to the ferromagnetic phase <lb/>than the latter 7) . <lb/>In the present paper, the temperature (T) dependence of Knight shift K studied by 59 Co-NMR of <lb/>Na x CoO 2 ⋅yH 2 O is reported. Then, based on the observed behaviors of K and 1/T 1 , the paring symmetry <lb/>is argued. <lb/>The powder sample of Na x CoO 2 ⋅yH 2 O with T c ~4.5 K was prepared from Na 0.75 CoO 2 by the method <lb/>described in Ref. 1. The structural and magnetic properties of the sample were reported in the previous <lb/>paper 7) . The NMR experiments were carried out using a standard phase coherent type pulse <lb/>spectrometer. 59 Co NMR spectra were obtained by recording the spin-echo intensity with the applied <lb/>magnetic field changed step by step in a magnetic field range of 1-2 T. Measurements were carried out <lb/>on two kinds of samples of Na x CoO 2 ⋅yH 2 O, randomly oriented and aligned samples. The latter sample <lb/>was prepared by mixing the Na x CoO 2 ⋅yH 2 O powder with epoxy adhesive (Stycast 1266) and keeping <lb/>the mixture in a magnetic field of 11T at room temperature for ~12 h. Its X ray diffraction pattern <lb/>indicates that the ab plane of Na x CoO 2 ⋅yH 2 O crystallites align along the direction of the magnetic field <lb/>with the spreading of about (±6°). ( The lattice parameter c was found not to exhibit significant change <lb/>after the alignment processes, which indicates no drastic change of the structure.) The former sample <lb/>does not have epoxy adhesive. It was just put into a cylindrical holder made by cellophane tapes. The <lb/>obtained spectra are shown in Fig. 1(a) and 1(b) for randomly oriented powder and aligned samples, <lb/>respectively. <lb/>The positions of the peaks and shoulders of the 59 Co-NMR spectra observed for both kinds of <lb/>samples were well explained by considering the anisotropic Knight shifts K x ≠ K y ≠ K z ≠ K x and the <lb/>effects of electric quadrupole interaction up to the second order. The directions of the principal axes of <lb/>Knight shift are assumed to be same as those of the electric quadrupole interaction. In the present case, <lb/>the z-axis was defined along the direction, for which the component of the electric quadrupole tensor, <lb/>ν zz is larger than the other components. This z direction was found to be along the c-axis by fitting the <lb/>positions of calculated spectra to those of aligned sample. The fitted result for the data taken at 5 K in <lb/>the magnetic field H within the ab plane of Na x CoO 2 ⋅yH 2 O is shown in Fig. 1(b) and the parameters <lb/></body>
13
+
14
+ <page>3 <lb/></page>
15
+
16
+ <body>determined by the fitting are, K x = ~3.5±0.1 %, K y = ~3.1±0.05 %, the electric quadrupole resonance <lb/>frequency ν Q =3.6±0.05 MHz and the asymmetric parameter of the electric field gradient at nuclei <lb/>η=0.28±0.02. For the spectra of the randomly oriented sample, K x = ~3.8±0.1 %, K y = ~3.14±0.05 %, <lb/>K z = ~2.3±0.2 %, ν Q =4.05±0.05 MHz and η=0.24±0.01 were found to reproduce them well. The <lb/>comparison of the values of ν Q and η of the aligned sample with those of the randomly oriented one <lb/>justifies the present assignment of the z direction along the c axis. <lb/>Figure 2 shows the T dependence of the 59 Co-NMR spectra around the peak of central line for the <lb/>randomly oriented powder sample. The H value of the peak position is nearly T-independent above T c <lb/>(~4.5 K) and then increases with decreasing T below T c . This change of the peak position is due to the <lb/>T dependence of the Knight shift K y and not due to the second order effect of the electrical quadrupole <lb/>interaction, because ν Q and η are T-independent, as was confirmed by analyzing the spectra in the H <lb/>region between 0.8-2.4 T at 2.9 K, 4.7 K and 6.2 K. <lb/>It should be noted here following facts. For the aligned sample, the peak position of the central line <lb/>of the 59 Co-NMR spectra, which corresponds to K y , is nearly T-independent even below T c . It is <lb/>slightly puzzling, because the result indicates that the T dependence of K y of the aligned sample is <lb/>different from that of the powder sample. To explain this puzzle, we consider that the shift K y observed <lb/>in ref. 8 for a sample aligned in epoxy adhesive (Stycast) also exhibits the T-independent behavior. We <lb/>think that the specimens embedded in the Stycast have somewhat different properties from those of the <lb/>(non-aligned) samples. This idea is supported by the fact that we have observed different values of K x <lb/>between the two kinds of samples, which has already been stated above. Then, the data of the aligned <lb/>sample may not be intrinsic. <lb/>We have estimated the K y value of the (non-aligned) powder sample. The results are shown in Fig. 3. <lb/>Above T c , K y is nearly constant and decreases with decreasing T below T c . Because we have already <lb/>taken into account the effect of the electrical quadrupole interaction up to the second order in the <lb/>estimation of K y , the decrease of K y below T c with decreasing T should be due to the T dependence of <lb/>its spin component K spin,y and/or due to the effect of the shielding diamagnetism (The orbital <lb/>component K orb , y may be T-independent.) Here, in order to estimate the change of the magnetic field at <lb/>the Co nuclei by the shielding supercurrent, we consider the case where the H is applied along the y <lb/>direction, i.e., within the a-b plane of the crystallites, because we are observing K y . For the penetration <lb/>depth λ of 5000~8000 Å at low temperatures, 3,6) we expect that the reduction of the field is roughly <lb/>one order of magnitude smaller that the observed shift of the peak. It has been confirmed by the <lb/>following experimental observation. Even when the measuring frequency is changed from 16.09 MHz <lb/>to 30.044 MHz, the Knight shift does not change within the experimental error bars, indicating that the <lb/>spatial variation of the field or the shielding effect is negligible in the present experimental condition. <lb/>It is also confirmed by studying experimental observations for the similar two-dimensional system <lb/>YBa 2 Cu 3 O 7 with λ (~1400 Å at T&lt;&lt;T c ) larger than that of the present system does not have a serious <lb/></body>
17
+
18
+ <page>4 <lb/></page>
19
+
20
+ <body>effect of the shielding diamagnetism in the vortex state (at H=7.4 T). 17) Furthermore, it can be noted <lb/>that the shape of the peak corresponding to K y does not exhibit an appreciable change expected in the <lb/>case where the spatial variation of H induced by the shielding diamagnetism is a main cause of the <lb/>shift. <lb/>We have just tried to fit the K y -T curve below T c by the Yosida function 18) (see the solid curve in <lb/>Fig. 3) and obtained the superconducting energy gap ∆ of 8.1 K. For the value and T c =4.5 K, 2∆/k B T c <lb/>is estimated to be 3.6, which agrees well with the one expected from the BCS theory, though the rather <lb/>good fitting by the Yosida function may be accidental. To estimate K spin,y precisely, we have to know <lb/>the orbital component K orb , y , which is, at this moment, not easy. <lb/>Now, we have shown that the spin susceptibility χ spin of Na x CoO 2 ⋅yH 2 O studied by measuring the <lb/>Knight shift K y decreases with decreasing T in the superconducting state. The existence of the <lb/>coherence peak of the nuclear relaxation rate 1/T 1 just below T c has also been reported in the previous <lb/>paper by the present authors. 7) These results indicate, naively speaking, that the superconducting pairs <lb/>of Na x CoO 2 ⋅yH 2 O are in the singlet spin state with s-wave symmetry, though other possibilities may <lb/>not be completely ruled out: Strictly speaking, the (d 1 +id 2 ) state predicted in refs. 10-13 may have a <lb/>small coherence peak. Then, the precise calculation of the amplitude of the coherence peak is required <lb/>to exclude such the symmetry. Even for the p-wave state predicted in refs. 14 and 15, the decrease of <lb/>K spin ,y can be expected, if the direction of the triplet spins is pinned in the direction perpendicular to the <lb/>y-axis. At this moment, we do not know how strong the pinning force is. To answer this question, <lb/>further studies on the anisotropy of the Knight shift has to be carried out by using single crystal <lb/>specimens. <lb/>In summary, the T dependence of the Knight shift K y of Na x CoO 2 ⋅yH 2 O has been reported. For the <lb/>(non-aligned) powder sample, the T dependence of the spin component K spin,y roughly follows the <lb/>Yosida function in the superconducting state. This behavior and the existence of the coherence peak in <lb/>the 1/T 1 -T curve observed just below T c indicate that Na x CoO 2 ⋅yH 2 O has the spin singlet pairs with the <lb/>s-wave symmetry, though the observed results do not completely exclude the possibility of other kinds <lb/>of electron pair state. <lb/></body>
21
+
22
+ <div type="acknowledgement">Acknowledgments-The authors thank Prof. Y. Ono of Nagoya University for fruitful discussion. The work is <lb/>supported by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of <lb/>Science (JSPS) and by Grants-in-Aid on priority area from the Ministry of Education, Culture, Sports, <lb/>Science and Technology. <lb/></div>
23
+
24
+ <page>5 <lb/></page>
25
+
26
+ <listBibl>Reference <lb/>1. K. Takada, H. Sakurai, E. Takayama-Muromachi, F. Izumi, R.A. Dilanian and T. Sasaki: Nature 422 <lb/>(2003) 53. <lb/>2. M.L. Foo, R.E. Schaak, V.L. Miller, T. Klimczuk, N.S. Rogado, Yayu Wang, G.C. Lau, C. Craley, <lb/>H.W. Zandbergen, N.P. Ong and R.J. Cava: cond-mat/0304464. <lb/>3. H. Sakurai, K. Takada, S. Yoshii, T. Sasaki , K. Kindo and E. Takayama-Muromachi: <lb/>cond-mat/0304503. <lb/>4. B. Lorenz, J. Cmaidalka, R.L. Meng and C.W. Chu: cond-mat/0304537. <lb/>5. R.E. Schaak, T. Klimczuk, M.L. Foo and R.J. Cava: cond-mat/0305450 <lb/>6. G. Cao, C.Feng, Yi Xu, W. Lu, J. Shen, M. Fang and Z. Xu: cond-mat/0305503 <lb/>7. Y. Kobayashi, M. Yokoi and M. Sato: cond-mat/0305649 <lb/>8. T. Waki, C. Michioka, M. Kato, K. Yoshimura, K. Takada, H. Sakurai E. Takayama-Muromachi and <lb/>T. Sasaki: cond-mat/0306036 <lb/>9. R. Jin, B.C. Sales, P. Khalifah and D. Mandrus: cond-mat/0306066 <lb/>10. G. Baskaran: cond-mat/0303649. <lb/>11. Brijesh Kumar and B. Sriram Shastry: cond-mat/0304210. <lb/>12. Q.-H. Wang , D. -H. Lee and P. A. Lee: cond-mat/0304377. <lb/>13. M. Ogata: cond-mat/0304405. <lb/>14. A. Tanaka and X. Hu: cond-mat/0304409 <lb/>15. D. J. Singh: cond-mat/0304532 <lb/>16. K. Sano and Y. Ono: cond-mat/0304620. <lb/>17. M. Takigawa, P.C. Hammel, R.H. Heffner and Z. Fisk: Phys. Rev. B 39 (1989) 7371. <lb/>18. K. Yosida: Phys. Rev. 110 (1958) 769. <lb/></listBibl>
27
+
28
+ <page>6 <lb/></page>
29
+
30
+ <body>Figure captions <lb/>Fig. 1(a) 59 Co-NMR spectra of Na x CoO 2 ⋅yH 2 O taken at T =4.72 K and with the resonance frequency f <lb/>=16.09 MHz for the randomly oriented sample. Solid line is just for the guide to the eye. <lb/>Fig. 1 (b) 59 Co-NMR spectra of Na x CoO 2 ⋅yH 2 O taken at T =4.72 K and with the resonance frequency f <lb/>=16.09 MHz for the aligned sample, where the powder of Na x CoO 2 ⋅yH 2 O is embedded in the <lb/>epoxy adhesive (Stycast 1266). Solid line is just for the guide to the eye. Broken line shows <lb/>the spectra calculated with the parameters K x =~3.5 %, K y = ~3.1 %, ν Q =3.6 MHz and η=0.28. <lb/>Fig. 2 <lb/>Profiles of the sharp peak of the central line taken for the randomly oriented powder sample <lb/>at several fixed temperatures are shown. The peak positions correspond to the Knight shifts <lb/>K y . The vertical broken line shows the peak position above T c . <lb/>Fig. 3 <lb/>Knight shift K y is plotted against T. The solid line shows the Yosida function fitted to the <lb/>data. <lb/>0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>1.2 <lb/>0.8 <lb/>1.2 <lb/>1.6 <lb/>2 <lb/>2.4 <lb/>centeral peak <lb/>(H//CoO <lb/>2 <lb/>plane) <lb/>65 Cu line <lb/>63 Cu line <lb/>&amp; Na <lb/>H (T) <lb/>59 <lb/>Co-NMR spin-echo intensity (a. u.) <lb/>Fig. 1(a) <lb/>Y. Kobayashi et al. <lb/>Na <lb/>x <lb/>CoO <lb/>2 <lb/>⋅yH <lb/>2 <lb/>O <lb/>powder sample <lb/>T =4.72K <lb/>f =16.090MHz <lb/>0 <lb/>0.4 <lb/>0.8 <lb/>1.2 <lb/>1.6 <lb/>2 <lb/>2.4 <lb/>0.8 <lb/>1.2 <lb/>1.6 <lb/>2 <lb/>2.4 <lb/>observed data <lb/>calculated curve <lb/>T =4.72K <lb/>f =16.090MHz <lb/>H (T) <lb/>59 <lb/>Co-NMR spin-echo intensity (a. u.) <lb/>Na <lb/>x <lb/>CoO <lb/>2 <lb/>⋅yH <lb/>2 <lb/>O <lb/>aligned sample <lb/>Fig. 1(b) <lb/>Y. Kobayashi et al. <lb/>63 Cu <lb/>&amp; Na <lb/>65 Cu <lb/>centeral peak <lb/>(H//CoO <lb/>2 <lb/>plane) <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>1.2 <lb/>1.46 <lb/>1.47 <lb/>1.48 <lb/>1.49 <lb/>1.5 <lb/>1.51 <lb/>1.983K <lb/>2.356K <lb/>2.920K <lb/>3.149K <lb/>3.414K <lb/>3.897K <lb/>4.249K <lb/>4.727K <lb/>6.24K <lb/>7.926K <lb/>10.84K <lb/>Na <lb/>x <lb/>CoO <lb/>2 <lb/>⋅yH <lb/>2 <lb/>O <lb/>f =16.09MHz <lb/>59 <lb/>Co-NMR spin-echo intensity (a. u.) <lb/>H (T) <lb/>Fig. 2 <lb/>Y.Kobayashi et al. <lb/>2.6 <lb/>2.8 <lb/>3.2 <lb/>0 <lb/>2 <lb/>4 <lb/>6 <lb/>8 <lb/>1 0 <lb/>K (%) <lb/>T (K) <lb/>Na <lb/>x <lb/>CoO <lb/>2 <lb/>⋅yH <lb/>2 <lb/>O <lb/>Fig. 3 <lb/>Y. Kobayashi et al. </body>
31
+
32
+
33
+ </text>
34
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPS0723041-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <front>arXiv:cond-mat/0308506v1 [cond-mat.supr-con] 25 Aug 2003 <lb/> Typeset with jpsj2.cls &lt;ver.1.2&gt; <lb/>Letter <lb/>Unconventional Superconductivity and Nearly Ferromagnetic Spin Fluctuations in <lb/>Na x CoO 2 • yH 2 O <lb/>K. Ishida 1 * , Y. Ihara Y. Maeno 1 C. Michioka M. Kato 2 K. Yoshimura 2 K. Takada T. Sasaki H. <lb/>Sakurai and E. Takayama-Muromachi <lb/>1 Department of Physics, Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan, <lb/>Department of Chemistry, Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan <lb/>Advanced Materials Laboratory, National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki, 305-0044, Japan <lb/>4 Superconducting Materials Center, National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan. <lb/>Co NQR studies were performed in recently discovered superconductor NaxCoO2•yH2O to <lb/>investigate physical properties in the superconducting (SC) and normal states. Two samples <lb/>from the same NaxCoO2 were examined, SC bilayer-hydrate sample with Tc ∼ 4.7 K and <lb/>non-SC monolayer-hydrate sample. From the measurement of nuclear-spin lattice relaxation <lb/>rate 1/T1 in the SC sample, it was found that the coherence peak is absent just below Tc and <lb/>that 1/T1 is proportional to temperature far below Tc. These results, which are in qualitatively <lb/>agreement with the previous result by Fujimoto et al., suggest strongly that unconventional <lb/>superconductivity is realized in this compound. In the normal state, 1/T1T of the SC sample <lb/>shows gradual increase below 100K down to Tc, whereas 1/T1T of the non-SC sample shows the <lb/>Korringa behavior in this temperature range. From the comparison between 1/T1T and χ bulk in <lb/>the SC sample, the increase of 1/T1T is attributed to nearly ferromagnetic fluctuations. These <lb/>remarkable findings suggest that the SC sample possesses nearly ferromagnetic fluctuations, <lb/>which are possibly related with the unconventional superconductivity in this compound. The <lb/>implication of this finding is discussed. <lb/>KEYWORDS: superconductivity, hydrous sodium cobalt oxide, NQR, spin fluctuations <lb/></front>
9
+
10
+ <body>Superconductivity in Na x CoO 2 •yH 2 O (x ∼ 0.35, y ∼ <lb/>1.3) with the transition temperature T c ∼ 5 K was dis-<lb/>covered quite recently by Takada et al. 1) Although T c is <lb/>by one-order of magnitude smaller than that in cuprate <lb/>superconductors, much attention have been paid for its <lb/>unique crystal structure of the 2-dimensional (2-D) layer <lb/>where superconductivity occurs. The CoO forms a 2-D <lb/>hexagonal layered structure, which is in contrast with <lb/>the tetragonal structure of cuprates. Due to the vacancy <lb/>of the Na atom, 0.65 holes are doped into the band in-<lb/>sulating state of low-spin Co 3+ (3d 6 in t 2g orbits), which <lb/>is regarded alternatively as 0.35-electron doping state in <lb/>the triangular lattice consisting of S = 1/2 of Co 4+ . One <lb/>may expect the unconventional superconductivity with <lb/>magnetic frustrations originating from the triangular lat-<lb/>tice. <lb/>For understanding superconducting (SC) properties, <lb/>determination of the pairing symmetry is one of the most <lb/>important issues. For the purpose, nuclear magnetic reso-<lb/>nance (NMR) and nuclear quadrupole resonance (NQR) <lb/>measurements are suitable because they give crucial in-<lb/>formation about the orbital and spin states of the SC <lb/>pairs from a microscopic point of view. In particular, <lb/>NQR technique is powerful when some impurity phases <lb/>are included in a sample, because NQR can spectroscopi-<lb/>cally detect only a concerned phase from the difference of <lb/>the electric field gradient (EFG) in each phase. Further-<lb/>more, by using the NMR and NQR techniques, nuclear-<lb/>spin lattice relaxation rate 1/T 1 can be measured, which <lb/>is related with low-energy spin dynamics in compounds. <lb/></body>
11
+
12
+ <front>* E-mail address: [email protected] <lb/></front>
13
+
14
+ <body>Until now, there are three NMR and NQR reports on <lb/>Na x CoO 2 •yH 2 O. 2-4) Two groups report that 1/T 1 T is en-<lb/>hanced just below T c . 2, 3) On the other hand, Fujimoto <lb/>et al. show that 1/T 1 decreases suddenly just below T c <lb/>and follows the Korringa behavior (T T = const. behav-<lb/>ior) far below T c . 4) Their observation on 1/T 1 can be <lb/>interpreted by the unconventional SC model with line-<lb/>node gap. Thus the results reported so far contradict <lb/>each other. To settle the controversy over the NMR and <lb/>NQR results in the SC state, further 1/T 1 studies are <lb/>highly desired. <lb/>In this paper, we show our 1/T 1 results measured inde-<lb/>pendently. Our measurement was performed in the wider <lb/>temperature range between 65mK and 200 K and on two <lb/>samples with different character. One of the samples is <lb/>SC bilayer-hydrate Na x CoO 2 •yH 2 O with higher T c ∼ 4.7 <lb/>K than in the previous reports. 3, 4) The other is the non-<lb/>SC monolayer hydrate Na x CoO 2 •yH O due to partial <lb/>extraction of H O molecules between CoO layers. Our <lb/>1/T 1 result in the SC state is consistent with that by Fuji-<lb/>moto et al., i.e. absence of the coherence peak just below <lb/>T c and existence of the residual density of states far below <lb/>T c . 4) These are characteristic features of unconventional <lb/>superconductivity. In addition, we found the low-energy <lb/>spin-fluctuations in the SC sample, which is considered <lb/>as nearly ferromagnetic fluctuations. These results sug-<lb/>gest that unconventional superconductivity appears in <lb/>the metallic state with nearly ferromagnetic fluctuations, <lb/>which play an important role for the occurrence of super-<lb/>conductivity. <lb/>We used powder sample for our NQR measurements, <lb/>preparation of which was described in literatures. 1, 5) SC <lb/></body>
15
+
16
+ <note place="headnote">J. Phys. Soc. Jpn. <lb/>letter <lb/>Author Name <lb/></note>
17
+
18
+ <body>11.5 <lb/>12.0 <lb/>12.5 <lb/>13.0 <lb/>non -SC <lb/>monolayer <lb/>hydrate <lb/>Na x CoO 2 yH 2 O <lb/>Co -NQR Intensity ( arb. unit ) <lb/>Frequency ( MHz ) <lb/>Co NQR Spectrum <lb/>( 7 / 2 ⇔ 5 / 2 ) <lb/>SC T c ~ 4.7 K <lb/>bilayer hydrate <lb/>Fig. 1. Co-NQR spectra corresponding to ±5/2 ↔ ±7/2 tran-<lb/>sition in the SC bilayer-hydrate and non-SC monolayer-hydrate <lb/>samples. The spectra were obtained by frequency-swept method. <lb/>transition was confirmed by dc-susceptibility measure-<lb/>ment (χ bulk ). The non-SC sample was obtained by stor-<lb/>ing the SC sample in the vacuum space at room tem-<lb/>perature for three days. From X-ray measurement, the <lb/>c-axis lattice constant was found to 6.9Å, which cor-<lb/>responds to that in monolayer-hydrate sample. 5) The <lb/>bilayer-hydrate phase could not observed in the non-SC <lb/>sample at all. Since the non-SC sample was prepared <lb/>from the SC sample, Na content is equivalent in two sam-<lb/>ples. This is very important for considering the role of <lb/>H O layer, since the superconductivity appears in quite <lb/>narrow Na-concentration region. 6) <lb/>Figure 1 shows the Co-NQR spectra in the SC and non-<lb/>SC samples, which were obtained by frequency-swept <lb/>method. The Co-NQR spectra originate from Co (I <lb/>= 7/2) nuclear-level transition between ±5/2 ↔ ±7/2. <lb/>The SC sample shows a single peak at 12.35 MHz with <lb/>the linewidth of 0.4 MHz. The EFG frequency and the <lb/>asymmetric parameter are consistent with previous re-<lb/>ports. 3, 4) On the other hand, the non-SC sample shows <lb/>two sharp peaks at 12.1 and 12.35 MHz, together with <lb/>other small satellite peaks at lower and higher frequency <lb/>sides. It seems that all Co sites are crystallographycally <lb/>unique in the SC sample: the CoO 2 layer is sandwitched <lb/>by two hydrate layers. On the other hand, due to ex-<lb/>traction of one hydrate layer, some Co sites in the CoO <lb/>layer of the non-SC sample have a different local sym-<lb/>metry from those of the SC sample, e.g. the sharp peak <lb/>at 12.1 MHz arises from the Co site, in which one of the <lb/>hydrate molecules near Co is regularly replaced by Na, <lb/>10 <lb/>-1 <lb/>10 <lb/>0 <lb/>10 <lb/>1 <lb/>10 <lb/>2 <lb/>10 <lb/>-1 <lb/>10 <lb/>0 <lb/>10 <lb/>1 <lb/>10 <lb/>2 <lb/>10 <lb/>3~ <lb/>T <lb/>T ( K ) <lb/>Na x CoO 2 yH 2 O <lb/>SC bilayer hydrate <lb/>Co -NQR <lb/>T c ~ 4.7 K <lb/>T <lb/>3 <lb/>1 / T <lb/>1 ( sec <lb/>-1 <lb/>) <lb/>Fig. 2. Temperature dependence of 1/T 1 of the SC bilayer-<lb/>hydrate sample plotted in logarithmic scale. The dotted curve in <lb/>the SC state is the calculation using the 2-D line-node (∆(φ) = <lb/>cos(2φ) ) model with residual DOS induced by unitarity impuri-<lb/>ties. The appropriate fitting parameters are 2∆/k B Tc = 3.5 and <lb/>Nres/N 0 ∼ 0.32. <lb/>and the satellite peaks are from the Co sites, in which <lb/>two or three hydrate molecules are replaced by Na. For <lb/>identification of Co-NQR peaks in the non-SC sample, <lb/>further NQR studies are needed in different stoichiomet-<lb/>ric samples. <lb/>T 1 was measured at 12.35 MHz in the SC sample and <lb/>12.1 and 12.35 MHz in the non-SC sample, respectively. <lb/>At these peaks, the recovery of the nuclear magnetization <lb/>after saturation pulses can be fitted by the theoretical <lb/>curve 7) in whole temperature range except for very low <lb/>temperature below 300 mK. <lb/>First, we discuss temperature dependence of 1/T in <lb/>the SC state. Figure 2 shows temperature dependence <lb/>of 1/T 1 in logarithmic scale. 1/T 1 shows sharp decrease <lb/>below T c and crosses over to the Korringa behavior in low <lb/>temperatures, which is qualitatively in good agreement <lb/>with the result by Fujimoto et al. 4) <lb/>To see the behavior around T c in detail, we show in <lb/>Fig. 3 temperature dependence of 1/T 1 divided by T , <lb/>1/T 1 T of the SC sample below 10 K. For comparison, we <lb/>also plot bulk susceptibility showing SC transition. As <lb/>seen in the figure, 1/T T decreases just below T c deter-<lb/>mined by the susceptibility. It is obvious that the coher-<lb/>ence peak is absent below T c , which suggests that the <lb/>superconductivity should be classified to an unconven-<lb/>tional superconductor such as heavy-Fermion, cuprate, <lb/>ruthenate, and organic superconductors. <lb/>The whole temperature dependence in the SC state <lb/>can be understood by the 2-D line-node model with <lb/>∆(φ) = ∆ 0 cos(2φ) incorporated with residual density of <lb/></body>
19
+
20
+ <note place="headnote">J. Phys. Soc. Jpn. <lb/>Letter <lb/>Author Name <lb/></note>
21
+
22
+ <page>3 <lb/></page>
23
+
24
+ <body>2 <lb/>4 <lb/>6 <lb/>8 <lb/>10 <lb/>10 <lb/>20 <lb/>H = 200 Oe <lb/>T c ~ 4.7 K <lb/>T ( K ) <lb/>Na x CoO yH 2 O <lb/>SC bilayer Hydrate <lb/>Co -NQR <lb/>1 / T <lb/>1 <lb/>T ( sec <lb/>-1 <lb/>K <lb/>-1 <lb/>) <lb/>300 <lb/>400 <lb/>χ ( 10 <lb/>-9 <lb/>emu / mol ) <lb/>Fig. 3. Temperature dependence of 1/T T and χ bulk of the SC <lb/>bilayer-hydrate sample below 10 K. <lb/>states (N res ) ascribed to impurities and/or crystal imper-<lb/>fections. 8) The dotted lines in Fig. 2 is the calculation us-<lb/>ing the model with 2∆/k B T c = 3.5 and N res /N 0 ∼ 0.32, <lb/>where N 0 is the density of states at T = T c . We found <lb/>that N res /N 0 is smaller in the higher-T c samples than <lb/>those reported by Fujimoto et al. 4) (N res /N 0 ∼ 0.65 and <lb/>T c ∼ 3.9 K ). Similar tendency was already seen in uncon-<lb/>ventional superconductor Sr RuO 4 . 9) Based on the the-<lb/>oretical model by Hotta, 10) we can tentatively estimate <lb/>T c0 in a pure sample of this compound as 5.5 K. Due to <lb/>the presence of the residual density of states, we cannot <lb/>rule out the possibility of the isotropic gap ascribed to <lb/>D + iD state, but the most promising gap would be the <lb/>line-node gap with residual density of states induced by <lb/>unitarity impurities. 11) <lb/>Next we move on to the normal-state properties in <lb/>these compounds. Temperature dependence of 1/T 1 T of <lb/>SC and non-SC samples, which was measured at 12.35 <lb/>MHz, are shown in the main panel of Fig. 4, in which <lb/>the scale of T axis is in the logarithmic scale. 1/T 1 T <lb/>was also measured at 12.1MHz in the non-SC sample, <lb/>and the value of 1/T 1 T is 8.2 s −1 K −1 at 4.2 K, 10 % <lb/>smaller than 1/T 1 T at 12.35 MHz, and the temperature <lb/>dependence is the same as that at 12.35 MHz. As seen <lb/>in the figure, 1/T 1 T of the non-SC sample shows the <lb/>Korringa behavior from 100 K to 1.4 K, whereas 1/T T <lb/>of the SC sample increases with decreasing temperature <lb/>below 100K down to T c . The Korringa behavior is not <lb/>seen at all just above T c in the SC sample. It is obvious <lb/>that the low-energy spin-fluctuations are present in the <lb/>SC sample, which is not seen in the non-SC sample. <lb/>To understand the spin-fluctuation character, we com-<lb/>pare the behavior of 1/T 1 T with that of bulk susceptibil-<lb/>ity χ bulk . The inset of Fig. 4 shows the temperature de-<lb/>pendence of χ bulk in two samples. The gradual increase of <lb/>χ bulk in the SC sample is seen below K, where 1/T 1 T <lb/>also increases. On the other hand, χ bulk in the non-SC <lb/>sample shows a sharp increase below 50 K whereas 1/T 1 T <lb/>does not in this temperature range. It seems that the in-<lb/>crease of χ bulk in the SC and non-SC samples might be <lb/>different in origin, e.g. the sharp increase in the non-SC <lb/>0 <lb/>10 <lb/>1 <lb/>10 <lb/>2 <lb/>0 <lb/>10 <lb/>20 <lb/>30 <lb/>40 <lb/>non-SC monolayer hydrate <lb/>T c <lb/>SC bilayer hydrate <lb/>1 / T <lb/>1 <lb/>T ( sec <lb/>-1 <lb/>K <lb/>-1 <lb/>) <lb/>T ( K ) <lb/>0 <lb/>100 <lb/>200 <lb/>300 <lb/>non-SC monolayer <lb/>SC bilayer hydrate <lb/>10 <lb/>5 <lb/>5 <lb/>0 <lb/>0 <lb/>χ <lb/>bulk ( 10 <lb/>-4 <lb/>emu / mol ) <lb/>χ <lb/>bulk ( 10 <lb/>-4 <lb/>emu / mol ) <lb/>T ( K ) <lb/>Fig. 4. Temperature dependence of 1/T 1 T in SC bilayer-hydrate <lb/>and non-SC monolayer-hydrate samples. The inset shows tem-<lb/>perature dependence of χ bulk . In the inset, χ bulk is normalized, <lb/>so that the value of χ bulk at 300 K is identical in two samples. <lb/>sample originates from local moments in some impurity <lb/>phases, but the gradual increase in the SC sample is in-<lb/>trinsic behavior in the compound. This possibility is also <lb/>suggested by the comparison between 1/T T and χ bulk <lb/>in the SC sample as shown later. <lb/>In general, 1/T 1 T is related with low-energy part of <lb/>the q-dependent dynamical susceptibility in compounds. <lb/>From the comparison between 1/T T and χ bulk or Knight <lb/>shift, we can have an important information about spin-<lb/>fluctuation character, since the latter are related with <lb/>the static susceptibility at q = 0, χ(0). If AFM corre-<lb/>lations are dominant, which is the case in cuprate su-<lb/>perconductors, dynamical susceptibility has peaks at the <lb/>AFM wave vector Q apart from q = 0, therefore 1/T T is <lb/>mainly determined by staggered susceptibility, χ(Q). 12) <lb/>In most cuprate superconductors, 1/T 1 T and χ bulk show <lb/>different behavior, i.e., 1/T 1 T is enhanced whereas χ bulk <lb/>decreases with deceasing temperature due to the develop-<lb/>ment of AFM correlations. 13) Such information played a <lb/>crucial role in identifying the existence of AFM spin fluc-<lb/>tuations in underdoped cuprate superconductor. On the <lb/>contrary, when FM correlations are dominant, the dy-<lb/>namical susceptibility shows a peak at q = 0, thus 1/T 1 T <lb/>is dominant by χ(0) component. In the SCR theory which <lb/>describes the magnetic properties in weakly or nearly <lb/>FM metallic state successfully, 1/T 1 T was suggested to <lb/>be proportional to χ(0). 14) This relation was confirmed <lb/>experimentally in nearly FM metallic compounds such <lb/>as Pd 15) and YCo 2 . 16) The nearly FM fluctuations orig-<lb/>inate from the high density of states of d-electron at the <lb/>Fermi level. <lb/></body>
25
+
26
+ <page>4 <lb/></page>
27
+
28
+ <note place="headnote">J. Phys. Soc. Jpn. <lb/>letter <lb/>Author Name <lb/></note>
29
+
30
+ <body>0 <lb/>100 <lb/>150 <lb/>200 <lb/>0 <lb/>10 <lb/>20 <lb/>30 <lb/>1 / T 1 T <lb/>1 / T <lb/>1 <lb/>T ( sec <lb/>-1 <lb/>K <lb/>-1 <lb/>) <lb/>5 <lb/>0 <lb/>M <lb/>M <lb/>T ( K ) <lb/>χ bulk <lb/>χ <lb/>bulk ( 10 <lb/>-4 <lb/>emu / mol) <lb/>M <lb/>0 <lb/>5 <lb/>10 <lb/>15 <lb/>20 <lb/>1 / T <lb/>1 <lb/>T ( sec <lb/>-1 <lb/>K <lb/>-1 <lb/>) <lb/>0 <lb/>5 <lb/>χ bulk ( 10 <lb/>-4 emu / mol) <lb/>M <lb/>M <lb/>Fig. 5. Temperature dependence of 1/T T and χ bulk up to 200 <lb/>K. The inset shows the plot of 1/T 1 T against χ bulk . <lb/>Figure 5 shows temperature dependence of 1/T 1 T in <lb/>the normal state up to 200 K, together with bulk suscep-<lb/>tibility χ bulk for comparison. It was found that 1/T T <lb/>shows a good linear relation with χ bulk , as seen in the <lb/>inset of Fig. 5. 17) The good linear scaling between 1/T 1 T <lb/>and χ bulk strongly suggests that upturn of χ bulk in the <lb/>SC sample below 100 K is intrinsic behavior in the com-<lb/>pound, not from impurity phases. The linear relation in <lb/>the inset of Fig. 5 also suggests that the spin fluctuations <lb/>seen in the SC sample is nearly ferromagnetic ones, and <lb/>that superconductivity would appear in the nearly FM <lb/>metallic state, which is a quite contrast situation where <lb/>cuprate superconductivity appears. The nearly FM state <lb/>is also suggested by the large Wilson ratio (3.33), where <lb/>the values of χ bulk and γ of specific-heat measurement <lb/>just above T c <lb/>18) are adopted for the estimation. <lb/>Until now, it was found that most unconventional su-<lb/>perconductors appearing in the paramagnetic state have <lb/>strong AFM spin fluctuations. Even in spin-triplet super-<lb/>conductors reported so far, i.e., UPt 3 <lb/>19) and Sr 2 RuO 4 , 20) <lb/>dominant spin fluctuations are not FM but AFM like. Re-<lb/>cently unconventional superconductivity was discovered <lb/>far below FM transition, which is considered as spin-<lb/>triplet superconductors in the FM state. 21, 22) However, <lb/>although the existence of FM fluctuations above T c has <lb/>not been identified in the FM superconductors due to dif-<lb/>ficulty of experiments, it is likely that FM fluctuations <lb/>are mostly quenched deep into the FM ordered phase. <lb/>There has been no report about superconductivity ap-<lb/>pearing in the predominance of FM fluctuations. Thus, as <lb/>far as we know, Na x CoO 2 •yH 2 O reported here is the first <lb/>example that the superconductivity appears in the nearly <lb/>FM metallic state. In such a situation, spin-triplet super-<lb/>conductivity in analogy with superfluidity of He would <lb/>be the most promising state within possible SC pairing <lb/>states. To identify the spin state of SC pairs, Knight-<lb/>shift measurement in the SC state is most crucial. Quite <lb/>recently, Waki et al. reported that the Co-NMR spec-<lb/>trum in the aligned SC sample is unchanged on passing <lb/>through T c , 2) whereas Kobayashi et al. reported the de-<lb/>crease of the Knight shift in the non-aligned SC sam-<lb/>ple. 23) The Knight shift behavior in the SC state is now <lb/>controversial issue. The result of the former group sug-<lb/>gests a spin-triplet superconducting state which is con-<lb/>sistent with our NQR results. <lb/>In conclusion, we show by Co-NMR study that <lb/>unconventional superconductivity is realized in <lb/>Na x CoO 2 •yH 2 O. The promising SC-gap would be <lb/>line-node one with residual density of states induced <lb/>by unitarity impurities. We found the low-energy spin-<lb/>fluctuations present only in the SC sample, which are <lb/>considered as nearly ferromagnetic fluctuations from the <lb/>comparison between 1/T 1 T and χ bulk . We suggest that <lb/>Na x CoO 2 •yH 2 O is a new type of superconductivity, in <lb/>which nearly FM spin fluctuations play the primary role <lb/>for the mechanism of superconductivity. <lb/></body>
31
+
32
+ <div type="acknowledgement">We thank H.Yaguchi for experimental support, and H. <lb/>Ikeda, and K. Yamada for valuable discussions. One of <lb/>the authors (K.I.) thanks W. Higemoto, and R. Kadono <lb/>for valuable information. This work has been partially <lb/>supported by CREST of Japan Science and Technology <lb/>Corporation (JST) and the 21COE Research in Grant-<lb/>in-Aid for Scientific Research from the Ministry of Edu-<lb/>cation, Culture, Sport, Science and Technology of Japan <lb/>(MECSST), and by the Grants-in-Aid for Scientific Re-<lb/>search from Japan Society for Promotion of Science <lb/>(JSPS), and MECSST. <lb/></div>
33
+
34
+ <listBibl>1) K. Takada et al., Nature (London) 422, (2003) 53. <lb/>2) T. Waki et al., cond-mat/0306036. <lb/>3) Y. Kobayashi et al., cond-mat/0305649. <lb/>4) T. Fujimoto et al., cond-mat/0307 <lb/>5) K. Takada et al., to be published in J. Solid State Chem. <lb/>6) R. E. Schaak et al., Nature 424, (2003) 527 <lb/>7) D. E. MacLaughlin, J. D. Williamson and J. Butterworth, <lb/>Phys. Rev. B 4, (1971) 60. <lb/>8) K. Ishida et al., J. Phys. Soc. Jpn. 62, (1993) 2803. <lb/>9) K. Ishida et al., Phys. Rev. Lett. 84, (2000) 5387. <lb/>10) Y. Hotta, J. Phys. Soc. Jpn. 62, (1993) 274. <lb/>11) Y. Bang, M. J. Graf, and A. V. Balatsky, cond/mat 0307510. <lb/>12) For example, Y. Kitaoka et al., Physica C 170,(1990) 189. <lb/>13) M. Takigawa et al., Phys. Rev. B 43 43, (1991) 247. <lb/>14) T. Moriya, and K. Ueda, Solid State Commun. 15, (1974) 169. <lb/>15) M. Takigawa, and H. Yasuoka, J. Phys. Soc. Jpn. 51, 787 <lb/>(1982). <lb/>16) K. Yoshimura et al., J. Phys. Soc. Jpn. 53, 503 (1984). <lb/>17) The intercept of the x axis gives the orbital Van-Vleck sus-<lb/>ceptibility in the compound, which is estimated as 2 ×10 −5 <lb/>emu / mol. This suggests that the dominant contribution is <lb/>from the spin susceptibility of the Co-d electrons <lb/>18) B. G. Ueland et al., cond/mat 0307106. <lb/>19) G. R. Stewart et al., Phys. Rev. Lett. 52, (1984) 679. <lb/>20) Y. Maeno et al., Nature 372, (1996) 532. <lb/>21) S. S. Saxena et al., Nature 406, (2000) 587. <lb/>22) D. Aokiet al., Nature 413, (2001) 613. <lb/>23) Y. Kobayashi, M. Yokoi, and M. Sato, cond/mat 0306264 </listBibl>
35
+
36
+
37
+ </text>
38
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPS0730819-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+
8
+ <front>New Pyrochlore Oxide Superconductor RbOs 2 O 6 <lb/>Shigeki Yonezawa*, Yuji Muraoka, Yoshitaka Matsushita and Zenji Hiroi <lb/>Institute for Solid State Physics, University of Tokyo, Kashiwa, Chiba 277-8581 <lb/>Abstract <lb/>We report the discovery of a new pyrochlore oxide superconductor RbOs 2 O 6 . <lb/>The compound crystallizes in the same β-pyrochlore structure as the recently discovered <lb/>superconductor KOs 2 O 6 , where Os atoms form a corner-sharing tetrahedral network <lb/>called the pyrochlore lattice with Rb or K atoms in the cage. Resistivity, magnetic <lb/>susceptibility and specific heat measurements on polycrystalline samples evidence a <lb/>bulk superconductivity with T c = 6.3 K. <lb/>KEYWORDS: superconductivity, pyrochlore oxide <lb/>*E-mail address: [email protected] <lb/> </front>
9
+
10
+ <body>Pyrochlore oxides have the general chemical formula A 2 B 2 O 7 or A 2 B 2 O 6 O&apos;, where <lb/>A is a larger cation and B is a smaller transition metal (TM) cation. 1) The ideal <lb/>pyrochlore structure is composed of two types of cation-oxygen sublattices: one is a <lb/>corner-sharing tetrahedral network composed of A atoms with an O&apos; atom in the center <lb/>of each tetrahedron, and the other is another tetrahedral network of B atoms with each B <lb/>atom coordinated quasi-octahedrally by six O atoms. This type of tetrahedral network <lb/>is called the pyrochlore lattice, and has been studied extensively in order to elucidate the <lb/>effect of geometrical frustration on the properties of localized spin and itinerant electron <lb/>systems. <lb/>Recently, superconductivity was found in Cd 2 Re 2 O 7 (Re 5+ : 5d 2 ) at T c = K for the <lb/>first time in the family of pyrochlore oxides. 2-4) <lb/>The mechanism of the <lb/>superconductivity appears to be conventional, and may be understood in the framework <lb/>of the weak-coupling Bardeen-Cooper-Schrieffer (BCS) theory. 5) Very recently, we <lb/>have discovered a new pyrochlore oxide KOs 2 O 6 which exhibits superconductivity at <lb/>9.6 K. 6) Our preliminary structural analysis indicates that it crystallizes in a cubic <lb/>structure with space group <lb/>m <lb/>Fd3 , as in the ideal pyrochlore oxides, but with K atoms <lb/></body>
11
+
12
+ <page>1 <lb/></page>
13
+
14
+ <body>located at the O&apos; site of the ideal pyrochlore structure. It is known that the pyrochlore <lb/>structure sometimes tolerates vacancies at the A and O&apos; sites. 1) Such pyrochlore <lb/>oxides are generally called defect pyrochlores. By contrast, KOs 2 O 6 should not be <lb/>classified as defect pyrochlores, because of the apparent difference in the metal <lb/>occupations. Thus, we call this AB 2 O 6 type oxide a β-pyrochlore oxide to distinguish <lb/>it from ordinary defect pyrochlore oxides. In the following search for a new <lb/>superconductor, we have obtained a new ternary phase RbOs 2 O 6 with the β-pyrochlore <lb/>structure, which exhibits superconductivity at 6.3 K. <lb/>Polycrystalline samples were prepared from Rb 2 O and Os. The two powders <lb/>were mixed in the molar ratio of Rb 2 O:Os = 1:4, ground and pressed into a pellet in a <lb/>dry atmosphere. The pellet was heated in an evacuated silica tube at 773 K for 24 h. <lb/>It was necessary in the preparation process to avoid the formation of highly toxic OsO 4 . <lb/>In order to control the oxygen partial pressure in the silica tube, an appropriate amount <lb/>of AgO was added to the end of the silica tube: AgO decomposes into silver and oxygen <lb/>above 370 K, and thus generates an oxidizing atmosphere. The chemical composition of <lb/>the product examined by energy dispersive X-ray analysis in a scanning electron <lb/>microscope was Rb:Os ~ 1:2. <lb/>Figure 1 shows a powder X-ray diffraction (XRD) pattern taken at room <lb/>temperature. All the intense peaks can be indexed assuming a cubic unit cell with a <lb/>lattice constant a = 1.0114 nm. A few extra peaks from Os are also detected. <lb/>Extinctions observed in the XRD pattern are consistent with the space group of <lb/>m <lb/>Fd3 , <lb/>which is expected for the ideal pyrochlore structure. However, the relative peak <lb/>intensities are significantly different from those of typical pyrochlore oxides, and are <lb/>similar to those reported for KOs 2 O 6 . 6) Therefore, it is thought that RbOs 2 O 6 has the <lb/>same β-pyrochlore structure as KOs 2 O 6 . <lb/>Resistivity measurements were carried out down to 0.5 K by the standard <lb/>four-probe method in a Quantum Design PPMS equipped with a 3 He refrigerator. As <lb/>shown in Fig. 2(a), the temperature dependence of resistivity for RbOs 2 O 6 exhibits good <lb/>metallic behavior below room temperature, without any signs of phase transitions such <lb/>as observed in Cd 2 Os 2 O 7 <lb/>7) or Cd 2 Re 2 O 7 . 2) It is also significantly different from that <lb/>reported for KOs 2 O 6 : A clear T 2 temperature dependence is seen below 30 K for <lb/>RbOs 2 O 6 , which is absent for KOs 2 O 6 . 6) When a sample is further cooled, the <lb/>resistivity sharply drops below 6.5 K due to superconductivity. The resistivity below <lb/>the transition is nearly zero within our experimental resolution. The critical <lb/>temperature T c , defined as the midpoint temperature of the transition, is 6.3 K, and zero <lb/>resistivity is attained below 6.1 K. It is to be noted that the resistivity starts to decrease <lb/></body>
15
+
16
+ <page>2 <lb/></page>
17
+
18
+ <body>significantly at the high temperature of about 8 K, although the reason for this is not <lb/>clear. <lb/>In addition to the observation of the zero-resistivity transition, a large diamagnetic <lb/>signal associated with the Meissner effect has been observed below 6.3 K. Figure 2(b) <lb/>shows a temperature dependence of magnetic susceptibility measured on a powdered <lb/>sample in a Quantum Design MPMS. The measurements were carried out in a <lb/>magnetic field of 10 Oe on heating after zero-field cooling and then on cooling in the <lb/>field. A superconducting volume fraction estimated at 2 K from the zero-field cooling <lb/>experiment is almost 100 %, which is sufficiently large to constitute bulk <lb/>superconductivity. <lb/>The superconducting transition has also been detected by specific heat C <lb/>measurements. As shown in Fig. 2(c), the specific heat divided by temperature <lb/>suddenly increases below 6.3 K, and forms a broad maximum around 5.8 K. The <lb/>shape of this anomaly is unusual, and is different from what one expects for a <lb/>conventional superconductor. The details will be described elsewhere. <lb/>The superconductivity of RbOs 2 O 6 is robust against magnetic fields as shown in <lb/>the resistivity measurements under magnetic fields of Fig. 3(a). When the magnetic <lb/>field is increased, the transition curve systematically shifts to lower temperatures. The <lb/>superconductivity remains even at µ 0 H = 14 T at 0.5 K. The field dependence of T c , <lb/>which was determined as the midpoint of the transition, is plotted in Fig. 3(b). The <lb/>upper critical field at T = 0 may be around 17 T, which seems to be larger than Pauli&apos;s <lb/>limit, 12 T, for a weak-coupling BCS type superconductor in the absence of spin-orbit <lb/>interactions. However, as suggested by the previous band-structure calculations on <lb/>related compounds, 8) the spin-orbit interactions can be significantly large in the 5d TM <lb/>pyrochlore oxides, and thus the actual Pauli&apos;s limit can be larger than 12 T. <lb/>In conclusion, we found superconductivity with T c = 6.3 K in a new β-pyrochlore <lb/>oxide RbOs 2 O 6 . The nature of this superconductivity is to be clarified in a future study. <lb/>However, we believe that an interesting aspect of physics is involved in the <lb/>superconductivity of RbOs 2 O 6 , as in that of KOs 2 O 6 , on the basis of the itinerant <lb/>electrons on the pyrochlore lattice. <lb/></body>
19
+
20
+ <div type="acknowledgement">We thank F. Sakai for her help in the EDX analysis. This research was supported <lb/>by a Grant-in-Aid for Scientific Research on Priority Areas (A) provided by the <lb/>Ministry of Education, Culture, Sports, Science and Technology, Japan. <lb/></div>
21
+
22
+ <listBibl>References <lb/>1) M. A. Subramanian, G. Aravamudan and G. V. Subba Rao: Prog. Solid State Chem. <lb/>15 (1983) 55. <lb/>2) M. Hanawa, Y. Muraoka, T. Tayama, T. Sakakibara, J. Yamaura and Z. Hiroi: Phys. <lb/>Rev. Lett. 87 (2001) 187001. <lb/>3) H. Sakai, K. Yoshimura, H. Ohno, H. Kato, S. Kambe, R. E. Walstedt, T. D. <lb/>Matsuda, Y. Haga and Y. Onuki: J. Phys.: Condens. Matter 13 (2001) L785. <lb/>4) R. Jin, J. He, S. McCall, C. S. Alexander, F. Drymiotis and D. Mandrus: Phys. Rev. <lb/>B 64 (2001) 180503. <lb/>5) Z. Hiroi and M. Hanawa: J. Phys. Chem. Solids 63 (2002) 1021. <lb/>6) S. Yonezawa, Y. Muraoka, Y. Matsushita and Z. Hiroi: J. Phys.: Condens. Matter 16 <lb/>(2004) L9. <lb/>7) A. W. Sleight, J. L. Gillson, J. F. Weiher and W. Bindloss: Solid State Comm. 14 <lb/>(1974) 357. <lb/> 8) H. Harima: J. Phys. Chem. Solids 63 (2002) 1035. <lb/></listBibl>
23
+
24
+ <body>Figure captions <lb/>Fig. 1. Powder X-ray diffraction pattern of RbOs 2 O 6 . Peak index is given by <lb/>assuming a cubic unit cell with a lattice constant a = 1.0114 nm. Asterisks mark extra <lb/>peaks from Os. <lb/>Fig. 2. Temperature dependences of resistivity (a), magnetic susceptibility (b), and <lb/>specific heat divided by temperature (c). Insets in (a) and (c) show enlargements <lb/>around the superconducting transition. The magnetic susceptibility of (b) was <lb/>measured on a powdered sample of RbOs 2 O 6 in an applied field of 10 Oe. ZFC and <lb/>FC indicate zero-field cooling and field cooling curves, respectively. <lb/>Fig. 3. (a) Temperature dependence of resistivity as a function of magnetic fields. <lb/>(b) H-T phase diagram showing the temperature dependence of upper critical fields <lb/>determined from the resistivity data shown in (a). <lb/></body>
25
+
26
+ <page>4 <lb/></page>
27
+
28
+ <body>Fig.1 <lb/>Fig. 2 <lb/></body>
29
+
30
+ <page>6 <lb/></page>
31
+
32
+ <body>Fig. 3 <lb/></body>
33
+
34
+ <page>7 </page>
35
+
36
+
37
+ </text>
38
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPS0731297-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,49 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>Impurity Effects on the Superconducting Transition Temperature of <lb/>Na z CoO 2 ⋅yH 2 O <lb/>M. Yokoi, H. Watanabe, Y. Mori, T. Moyoshi, Y. Kobayashi and M. Sato <lb/>Department of Physics, Division of Material Science, Nagoya University, Furo-cho, <lb/>Chikusa-ku Nagoya, 464-8602 Japan <lb/>(Received <lb/>) <lb/>Superconducting transition temperature T c of Na z Co 1-x M x O 2 ⋅yH 2 O has been studied <lb/>for M=Ir and Ga, which can substitute for Co. It has been found that T c is <lb/>suppressed by the M-atom doping. The decreasing rates |dT c /dx| are of the order of <lb/>1 K/% and too small to be explained by the pair breaking mechanism for the <lb/>anisotropic order parameters by non-magnetic impurities. Brief arguments are given <lb/>on possible origins of the T c suppression. <lb/>corresponding author: M. Sato (e-mail: [email protected]) <lb/>Keywords: Na z CoO 2 ⋅yH 2 O, superconductivity, impurity effects <lb/></front>
8
+
9
+ <page>2 <lb/></page>
10
+
11
+ <body>Introduction <lb/>The occurrence of the superconductivity found in the layered cobalt oxide <lb/>Na z CoO 2 ⋅yH 2 O ( z~0.3 and y~1.3) by Takada et al. 1) has attracted much attention, <lb/>because it has the triangular lattice of Co atoms, in which the electrons are considered to <lb/>be strongly correlated and frustrated. The system can be obtained from Na z&apos; CoO 2 by <lb/>de-intercalating Na + ions and then, by intercalating H 2 O molecules. The transition <lb/>temperature T c is ~4.5 K. Up to now, various kinds of studies have been carried out to <lb/>investigate the mechanism of the superconductivity. 1-23) From the experimental side, <lb/>results of the specific heat ( C) measurements have been reported by several <lb/>groups, 6,10,11,13,15) and the nuclear spin-lattice relaxation rate (1/T 1 ) 7,12,16) and the NMR <lb/>Knight shift (K) 8) have also been reported as functions of temperature T, for example. <lb/>Among these works, the exotic nature of the superconductivity is presumed by several <lb/>groups. <lb/>From the theoretical side, a possible role of the interband excitations has been pointed <lb/>out in the Cooper pair formation in ref. 23. The resemblance of the electronic nature <lb/>with possible strong fluctuations of this frustrated lattice to that of high-T c oxides is <lb/>emphasized by several authors 17,20) , and a possibility of the triplet pairing is also <lb/>conceived by other groups. 18,19,21) In the previous report, 7) the present authors ha ve <lb/>shown by 59 Co-NQR studies of the system, that the coherence peak of the 59 Co-nuclear <lb/>spin relaxation rate 1/T 1 observed by NQR exists just below T c even though it is much <lb/>smaller than that expected for ordinary s-wave superconductors. We have also reported 8) <lb/>the NMR Knight shift observed for a peak of the NMR spectra, corresponding to the <lb/>crystallites oriented with their y-axes within the c-planes being parallel to the external <lb/>magnetic field H. Based on these results, we have pointed out that the superconducting <lb/>pairs can probably be considered to be in the singlet state, unless we do not consider a <lb/>situation where spins of triplet pairs are pinned along the direction perpendicular to the <lb/>y-axis. We have also pointed out that the symmetry may be s-like or fully gapped one <lb/>and added that other kinds of symmetry, ( d+id)-symmetry, for example, cannot be <lb/>excluded. Even if the symmetry is s-like, it does not necessarily indicate that the <lb/>superconductivity is not exotic, because the s-like symmetry or the fully gapped state <lb/>may be realized by the pairing mediated by the interband excitations of the electrons. 23) <lb/>Another possible way to study the pair symmetry is to examine an effect of the <lb/>impurity-doping on the superconducting transition temperature of the system: The pair <lb/>breaking effect of nonmagnetic impurities is negligible in ordinary s-wave <lb/>superconductors, while it can be large as in the case of magnetic impurities in usual <lb/>s-wave superconductors, if the symmetry of the superconducting order parameter has <lb/></body>
12
+
13
+ <page>3 <lb/></page>
14
+
15
+ <body>nonzero angular momentum l as the p-or d-wave one. For nonzero l, the absolute initial <lb/>slope |dT c /dx| is proportional to the scattering rate 1/τ of the electrons by the impurities. <lb/>(The relation is described by the pair breaking theory.) 24) Then, if impurities are in the <lb/>path of the conducting electrons, |dT c /dx| is expected to be very large. Because this <lb/>T c -suppression is expected essentially not to be observed, as stated above, for <lb/>non-magnetic impurities in s-wave superconductors, we may have a good way to <lb/>distinguish if the symmetry of the superconducting order parameter has non-zero l or <lb/>not. <lb/>Experiments <lb/>Polycrystalline samples of Na z&apos; Co 1-x M x O 2 (M=Ir and Ga) were prepared by solid <lb/>reactions at 750 ~850 °C for 24~36 hours, where the nominal value z&apos; was 0.7 or 0.75, <lb/>depending on series of samples. All samples of Na z&apos; Co 1-x M x O 2 (M=Ir and Ga) thus <lb/>obtained were found to have the CdI 2 structure (P6 3 /mmc) by X-ray powder diffraction <lb/>with FeKá radiation, where any impurity phases have not been observed for the small <lb/>values of the nominal concentration x studied here. The lattice parameter c determined <lb/>by the X-ray studies changed systematically with x for M=Ir, which guarantees that the <lb/>substitution is successfully carried out. For M=Ga, the lattice parameter c exhibits rather <lb/>large scattering as a function of x. However, because it shows, roughly speaking, a <lb/>decreasing tendency with increasing x and because the resistivity ρ and the transition <lb/>temperature T c of the system exhibit, as shown later, systematic x dependence, it can be <lb/>considered that the substitution is carried out. (In the actual experiments, besides M=Ir <lb/>and Ga, several atomic elements such as Cu and Al, for example, were tried to substitute <lb/>for Co. But, we could not find a firm experimental evidence that these elements really <lb/>substitute for Co: The lattice parameter did not show systematic dependence on the <lb/>nominal value x. The absolute value and the T-dependence of the resistivity ρ did not <lb/>change systematically with x, either. (The superconduting transition temperature was <lb/>almost independent of the nominal values of x.) Then, the obtained data of the doping <lb/>dependence of T c on x are, we think, reliable only for M=Ir and Ga. (For M=Rh, the data <lb/>may also be reliable.) and here these data are presented. <lb/>We employed a &quot;rapid heat-up&quot; technique to avoid the Na evaporation. 25) For each <lb/>sample series obtained through a single series of the processes, the c values were <lb/>determined by using 5-6 lines between the scattering angles of 20-90º, and the results <lb/>are shown in Figs. 1 and 5 for M=Ir and Fig. 9 for Ga. The c values at x=0 are found to <lb/>be different among these series probably due to the difference of the real Na <lb/>concentration, which depends on the preparation conditions such as the sintering <lb/></body>
16
+
17
+ <page>4 <lb/></page>
18
+
19
+ <body>temperature T a . (The lattice parameter c decreases with the Na concentration. 25) ) For <lb/>each series of samples, c increases with increasing x. The lattice parameter a is rather <lb/>insensitive to x. The resistivity ρ of Na z&apos; Co 1-x M x O 2 was measured by the four terminal <lb/>method. These samples were dried at 150~800 °C before the measurements. <lb/>Na z Co 1-x M x O 2 ⋅yH 2 O was prepared from the Na z&apos; Co 1-x M x O 2 as follows. Na atoms are <lb/>de-intercalated by dipping the Na z&apos; Co 1-x M x O 2 sample into acetonitrile, in which Br − ions <lb/>are dissolved. Then, the sample was washed by distilled water. In this process, H 2 O <lb/>molecules are intercalated. The magnetic susceptibilities χ of the samples were <lb/>measured by SQUID magnetometer with the magnetic field H of 5 G in the conditions <lb/>of the zero-field-cooling (ZFC). <lb/>Results and Discussion <lb/>Here, we present results of the studies on two series of Ir-doped system and a series <lb/>of Ga-doped system. For each series, experimental data of the lattice parameters c and <lb/>the electrical resistivities ρ of Na z&apos; Co 1-x M x O 2 , are shown. The magnetic susceptibilities <lb/>χ due to the shielding diamagnetism and the superconducting transition temperatures T c <lb/>of Na z Co 1-x M x O 2 ⋅yH 2 O derived from these series are also shown. <lb/>(1) Series A of Ir-doped system <lb/>For this series, the nominal value of z&apos; is 0.75. Figures 1 and 2 show the lattice <lb/>parameters c and the electrical resistivities ρ of Na z&apos; Co 1-x Ir x O 2 , respectively. As stated <lb/>above, the lattice parameter exhibits the systematic x-dependence, indicating that the <lb/>Ir-substitution for Co is really carried out. The electrical resistivity ρ also exhibits the <lb/>systematic x-dependence and it becomes insulating at x as small as 0.02. The <lb/>diamagnetic susceptibilities χ and the superconducting transition temperatures T c of <lb/>Na z Co 1-x Ir x O 2 ⋅yH 2 O obtained by intercalating H 2 O to the samples of Na z&apos; Co 1-x Ir x O 2 are <lb/>shown in Figs. 3 and 4, respectively. The values of T c are determined by using the <lb/>observed χ-T curves as the temperatures, at which the absolute values of χ exceeds a <lb/>certain value with decreasing T. (The parentheses in Fig. 4 indicate large ambiguities of <lb/>the data points.) The x-dependence seems to be linear and the slope |dT c /dx| is estimated <lb/>to be ∼1.0 K/%. <lb/>(2) Series B of Ir-doped system <lb/>For this series, the nominal value of z&apos; is 0.75, too. Figures 5-8 show the data taken <lb/>for the series B of Ir-doped system similar to those in Figs. 1-4, respectively. The <lb/>variations of c and ρ with x are, roughly speaking, systematic. For this series of samples, <lb/>the shielding diamagnetisms are smaller than those observed for the series A. The value <lb/>of T c at x=0 is larger than that of the series A. However, the initial slope |dT c /dx| is <lb/></body>
20
+
21
+ <page>5 <lb/></page>
22
+
23
+ <body>almost equal to the value observed for the series A. Then, we think that even though the <lb/>magnitudes of the shield ing diamagnetism are different between two series of A and B, <lb/>we are observing intrinsic behavior of the T c -x curve. <lb/>(3) Series of Ga-doped system <lb/>For this series, the nominal value of z&apos; is 0.7. Results are shown in Figs.9-12, as in <lb/>the cases of the Ir-doping. Although the increase of dρ/dT with increasing x found in <lb/>Fig.10 may not be explained by the simple Matthiessen&apos;s law, we have found almost <lb/>systematic x-dependences of ρ and T c . Then, the substitution of Ga for Co is, we think, <lb/>really carried out. In this system, the slope |dT c /dx| is ~2.0 K/%, which is about two <lb/>times larger than the values obtained for the two series of Na z Co 1-x Ir x O 2 ⋅yH 2 O. <lb/>We have so far presented the experimental data for three series of Na z&apos; Co 1-x M x O 2 <lb/>and Na z Co 1-x M x O 2 ⋅yH 2 O (M=Ir and Ga). Because T c has been found to be, roughly <lb/>speaking, linear in x in the region of small x, the pair breaking theory, which can be <lb/>described by <lb/>ln{T c (0)/T c (x)}=ψ{1/2+h/(2ηπk B T c (x)τ)} −ψ(1/2) <lb/>≡ ψ(1/2+α/2t) −ψ(1/2) <lb/>might be applicable to the present systems, where the pair breaking parameter α≡ <lb/>h/{ηπk B T c (0)τ}, t ≡T c (x)/T c (0), 1/τ is the scattering rate by the impurities and ψ(x) is the <lb/>digamma function. (By using the electron mean free path Λ~v F τ and the coherence <lb/>length ξ (=hv F /π∆ in the BCS theory), α is written as 1.76ξ/Λη, where v F and ∆{= <lb/>1.76k B T c (0)} are the Fermi velocity and the superconducting gap at x=T=0, <lb/>respectively.) The η value is equal to unity for the spin scattering by magnetic <lb/>impurities in s-wave superconductors and it is equal to 2 for both the potential and spin <lb/>scatterings by impurities when the order parameter has nonzero angular momentum l (If <lb/>l=0, α=0 for non-magnetic impurities and T c is not suppressed.). <lb/>Figure 13 shows the magnetic susceptibilities χ of the samples of Na z&apos; Co 1-x Ir x O 2 <lb/>(series B) with different x. Because the data do not indicate the increase of the number <lb/>of magnetic moments with increasing x, the impurities can be considered to be <lb/>non-magnetic, which is naturally expected, because non-magnetic Co atoms are <lb/>substituted with the isoelectronic Ir atoms or with Ga atoms which have only s and p <lb/>outer electrons. If the superconducting gap parameter has the s-wave symmetry, T c <lb/>should not, in an ideal case, be suppressed by the present doping. <lb/>It is interesting to compare the values of |dT c /dx| with those of other systems. In Fig. <lb/></body>
24
+
25
+ <page>6 <lb/></page>
26
+
27
+ <body>14, we summarize the T c -x curves obtained in the present experiments together with <lb/>those of well known systems, La 1.85 Sr 0.15 Cu 1-x Zn x O 4 <lb/>26) with the d-wave order parameter <lb/>and Mg(B 1-x C x ) 2 with the s-wave one. 27) The slopes |dT c /dx| observed for the present <lb/>sample series have the values similar to that of the Mg(B 1-x C x ) 2 system and much <lb/>smaller than that of La 1.85 Sr 0.15 Cu 1-x Zn x O 4 . It may be also useful to note that the values <lb/>of |dT c /dx| observed for the spinel system Cu(Co 1-x M x ) 2 S 4 are ~0.9 and ~0.6, for M=Rh <lb/>and Ir, respectively. 28) The system is considered to have the s-wave order parameter <lb/>from the existence of a significant coherence peak in the T-dependence of the nuclear <lb/>relaxation rate divided by T, 1/T 1 T. 29) It has the three dimensional network of the <lb/>corner-sharing (Co 1-x M x ) 4 and can be considered to be a three dimensional version of the <lb/>triangular lattice. For this system, |dT c /dx| is more than a half of that of <lb/>Na z Co 1-x Ir x O 2 ⋅yH 2 O. From these results, it is tempting to consider that the present <lb/>system has the s-wave order parameter and that the linear relationship between T c and x <lb/>is just caused by a mechanism different from the pair breaking due to the scattering by <lb/>non-magnetic impurities. <lb/>On the other hand, within the framework of the pair breaking theory, the relations <lb/>{T c (0)-T c (x)} ~πh/4ηk B τ~π 2 T c (0)α(x)/4~0.44π 2 (ξ/Λη)×T c (0) hold in the region of small <lb/>x (Note that ξ×T c (0) does not depend on T c (0). ). To explain the initial slope |dT c /dx| ~1 <lb/>K/%, ξ/(Λη)= hv F /{1.76πk B T c (0)}/( v F τ) should be of the order of ~1/20. If ξ is chosen <lb/>to be ~100 Å, 10) Λ has to be as large as 1000 Å at x=0.01, which seems to be too large as <lb/>compared with the average separation of neighboring impurities (~30 Å). The much <lb/>smaller values of |dT c /dx| for the present systems than that of La 1.85 Sr 0.15 Cu 1-x Zn x O 4 do <lb/>not seem to be easily understood only by the naive consideration of the pair breaking by <lb/>non-magnetic impurities, either <lb/>Then, what is the primary mechanism of the observed T c suppression? It is, roughly <lb/>speaking, to be linear in x, but the rate |dT c /dx| seems to be too small to understand by <lb/>the pair breaking mechanism. We may have to ask why Cu(Co 1-x M x ) 2 S 4 with the <lb/>s-symmetry of the order parameter exhibits the linear T c -decrease in x, 28) because the <lb/>similar origin of the T c suppression may be relevant to the present case. For example, <lb/>effects of the carrier localization on T c may be considered as one of possible origins, <lb/>because for M=Ir, the relatively small amount of the doping to the host material <lb/>Na z&apos; CoO 2 seems to induce the upturn of the resistivity with decreasing T (see Figs. 2 and <lb/>6). <lb/>Before summarizing the results of the present study, we also add followings. 30) The <lb/>Knight shifts K y studied as functions of T at several fixed H values (1.5 T ≤ H ≤ 4.5 T) <lb/>indicate that the spin susceptibility decreases with decreasing T below T c . The H-and <lb/></body>
28
+
29
+ <page>7 <lb/></page>
30
+
31
+ <body>T-dependences of K y cannot be explained by considering the superconducting <lb/>diamagnetism, as reported in our previous paper. 8) The magnetic field H c2 determined <lb/>from the T dependence of K y reproduces the H c2 -T curve for H applied within the c <lb/>plane reported in ref. 10. The H c2 value at T→0 can be understood by the effect of the <lb/>Zeeman splitting on singlet pairs, indicating again that the superconducting pairs are <lb/>definitely in the singlet state. It gives a restriction that the pairs have even parity. <lb/>In order to fully understand the observed values of |dT c /dx|, we have to examine <lb/>various quantities and study the electronic state of the present system in detail. In <lb/>summary, we have presented the results of the studies on the impurity-doping effects on <lb/>T c of Na z CoO 2 ⋅yH 2 O and shown that the values of |dT c /dx| are ~1 K/%, which is too <lb/>small to be explained by the pair breaking mechanism for the anisotropic (l≠0) order parameters <lb/>by non-magnetic impurities. <lb/></body>
32
+
33
+ <div type="acknowledgement">Acknowledgments-The work is supported by Grants-in-Aid for Scientific Research from <lb/>the Japan Society for the Promotion of Science (JSPS) and by Grants-in-Aid on priority <lb/>area from the Ministry of Education, Culture, Sports, Science and Technology. <lb/></div>
34
+
35
+ <page>8 <lb/></page>
36
+
37
+ <listBibl>References <lb/>1. K. Takada, H. Sakurai, E. Takayama-Muromachi, F. Izumi, R. A. Dilanian and T. Sasaki: <lb/>Nature 422 (2003) 53. <lb/>2. M. L. Foo, R. E. Schaak, V. L. Miller, T. Klimczuk, N. S. Rogado, Y. Wang, G. C. Lau, C. <lb/>Craley, H. W. Zandbergen, N. P. Ong and R. J. Cava: Solid State Commu. 127 (2003) 33. <lb/>3. H. Sakurai, K. Takada, S. Yoshii, T. Sasaki, K. Kindo and E. Takayama-Muromachi: Phys. <lb/>Rev. B 68 (2003) 132507. <lb/>4. B. Lorenz, J. Cmaidalka, R. L. Meng and C. W. Chu: Phys. Rev. B 68 (2003) 132504. <lb/>5. R. E. Schaak, T. Klimczuk, M. L. Foo and R. J. Cava: Nature 424 (2003) 527. <lb/>6. G. Cao, C. Feng, Y. Xu, W. Lu, J. Shen, M. Fang and Z. Xu: J. Phys.: Condens. Matter 15 <lb/>(2003) L519. <lb/>7. Y. Kobayashi, M. Yokoi and M. Sato: J. Phys. Soc. Jpn. 72 (2003) 2161. <lb/>8. Y. Kobayashi, M. Yokoi and M. Sato: J. Phys. Soc. Jpn. 72 (2003) 2453. <lb/>9. R. Jin, B.C. Sales, P. Khalifah and D. Mandrus: Phys. Rev. Lett. 91 (2003) 217001. <lb/>10. F. C. Chou, J. H. Cho, P. A. Lee, E . T. Abel, K. Matan and Y. S. Lee: <lb/>cond-mat/0306659. <lb/>11. B. G. Ueland, P. Schiffer, R. E. Schaak, M.L. Foo, V. L. Miller and R. J. Cava: <lb/>cond-mat/0307106. <lb/>12. T. Fujimoto, G. Zheng, Y. Kitaoka, R. L. Meng, J. Cmaidalka and C. W. Chu: <lb/>cond-mat/0307127. <lb/>13. H. D. Yang, J. -Y. Lin, C. P. Sun, Y. C. Kang, K. Takada, T. Sasaki, H. Sakurai and E. <lb/>Takayama-Muromachi: cond-mat/0308031. <lb/>14. J. Cmaidalka, A. Baikalov, Y. Y. Xue, R. L. Meng and J. Cmaidalka and C. W. Chu: <lb/>cond-mat/0308301. <lb/>15. B. Lorenz, J. Cmaidalka, R. L. Meng and C. W. Chu: cond-mat/0308143. <lb/>16. K. Ishida, Y. Ihara, Y. Maeno, C. Michioka, M. Kato, K. Yoshimura, T. Takada, T. <lb/>Sasaki, H. Sakurai and E. Takayama-Muromachi: J. Phys. Soc. Jpn. 72 (2003) 3041. <lb/>17. G. Baskaran: Phys. Rev. Lett. 91 (2003) 097003. <lb/>18. B. Kumar and B. S. Shastry: Phys. Rev. B 68 (2003) 104508. <lb/>19. Q.-H. Wang , D. -H. Lee and P. A. Lee: cond-mat/0304377. <lb/>20. M. Ogata: J. Phys. Soc. Jpn. 72 (2003) 1839. <lb/>21. A. Tanaka and X. Hu: cond-mat/0304409 <lb/>22. D. J. Singh: Phys. Rev. B 68 (2003) 020503. <lb/>23. K. Sano and Y. Ono: J. Phys. Soc. Jpn. 72 (2003) 1847. <lb/>24. A. A. Abrikosov and L. P. Gor&apos;kov: JETP 12 (1961) 1243. <lb/>25. T. Motohashi, E. Naujalis, R. Ueda, K. Isawa, M. Karppinen and H. Yamauchi: Appl. <lb/></listBibl>
38
+
39
+ <page>9 <lb/></page>
40
+
41
+ <listBibl>Phys. Lett. 79 (2001) 1480. <lb/>26. H. Harashina, T. Nishikawa, T. Kiyokura, S. Shamoto, M. Sato and K. Kakurai: Physica <lb/>C 212 (1993) 142. <lb/>27. T. Takenobu, T. Ito, Dam Hieu Chi, K. Prassides and Y. Iwasa: Phys. Rev. B 64 (2001) <lb/>134513. <lb/>28. N. Aito and M. Sato: in preparation. <lb/>29. H. Sugita, S. Wada, K. Miyatani, T. Tanaka and M. Ishikawa; Physica B 284-288 (2000) <lb/>473. <lb/>30. Y. Kobayashi, M. Yokoi and M. Sato: unpublished data. <lb/></listBibl>
42
+
43
+ <page>10 <lb/></page>
44
+
45
+ <body>Figure captions <lb/>Fig. 1 <lb/>Lattice parameters c of the sample series A of Na z&apos; Co 1-x Ir x O 2 at room temperature <lb/>are plotted against x. The nominal value of z&apos;=0.75. <lb/>Fig. 2 <lb/>Temperature dependence of the r esistivities ρ of the sample series A of <lb/>Na z&apos; Co 1-x Ir x O 2 is shown for various x. The nominal value of z&apos;=0.75. <lb/>Fig. 3 <lb/>Magnetic susceptibilities χ due to the shielding diamagnetism for the sample series <lb/>A of Na z Co 1-x Ir x O 2 ⋅yH 2 O are shown against x. <lb/>Fig. 4 <lb/>Superconducting transition temperatures T c of the sample series A of <lb/>Na z Co 1-x Ir x O 2 ⋅yH 2 O are shown against x. The parentheses indicate possible large <lb/>error bars. <lb/>Fig. 5 <lb/>Lattice parameters c of the sample series B of Na z&apos; Co 1-x Ir x O 2 at room temperature <lb/>are plotted against x. The nominal value of z&apos;=0.75. <lb/>Fig. 6 <lb/>Temperature dependence of the resistivities ρ of the sample series B of <lb/>Na z&apos; Co 1-x Ir x O 2 is shown for various x. The nominal value of z&apos;=0.75. <lb/>Fig. 7 <lb/>Magnetic susceptibilities χ due to the shielding diama gnetism for the sample series <lb/>B of Na z Co 1-x Ir x O 2 ⋅yH 2 O are shown against x. <lb/>Fig. 8 <lb/>Superconducting transition temperatures T c of the sample series B of <lb/>Na z Co 1-x Ir x O 2 ⋅yH 2 O are shown against x. The parentheses indicate possible large <lb/>error bars. <lb/>Fig. 9 <lb/>Lattice parameters c of the sample series of Na z&apos; Co 1-x Ga x O 2 at room temperature are <lb/>plotted against x. The nominal value of z&apos;=0.7. <lb/>Fig. 10 Temperature dependence of the resistivities ρ of the sample series of Na z&apos; Co 1-x Ga x O 2 <lb/>is shown for various x. The nominal value of z&apos;=0.7. <lb/>Fig. 11 Magnetic susceptibilities χ due to the shielding diamagnetism for the sample series <lb/>of Na z Co 1-x Ga x O 2 ⋅yH 2 O are shown against x. <lb/>Fig. 12 Superconducting transition temperatures T c of the sample series of <lb/>Na z Co 1-x Ga x O 2 ⋅yH 2 O are shown against x. The parentheses indicate possible large <lb/>error bars. <lb/>Fig. 13 The magnetic susceptibilities of Na z&apos; Co 1-x Ir x O 2 are shown against T for three Ir <lb/>concentration x. The nominal value of z&apos;=0.75. <lb/>Fig. 14 Superconducting transition temperatures T c of various superconducting systems are <lb/>plotted as functions of the doped atom concentrations. <lb/>10.88 <lb/>10.89 <lb/>10.9 <lb/>10.91 <lb/>10.92 <lb/>10.93 <lb/>0 0.005 0.01 0.015 0.02 0.025 0.03 <lb/>Na <lb/>0.75 <lb/>Co <lb/>1-x <lb/>Ir <lb/>x <lb/>O <lb/>2 <lb/>c <lb/>(Å) <lb/>x <lb/>(series A) <lb/>Fig.1 <lb/>Yokoi et al. <lb/>0 <lb/>10 <lb/>20 <lb/>30 <lb/>40 <lb/>50 <lb/>60 <lb/>0 <lb/>50 100 150 200 250 300 <lb/>Na <lb/>0.75 <lb/>Co <lb/>1-x <lb/>Ir <lb/>x <lb/>O <lb/>2 <lb/>ρ (mΩcm) <lb/>T (K) <lb/>x=0.025 <lb/>x=0.02 <lb/>x=0.015 <lb/>x=0.01 <lb/>x=0.0 <lb/>(series A) <lb/>Fig.2 <lb/>Yokoi et al. <lb/>-2.5 <lb/>-2 <lb/>-1.5 <lb/>-1 <lb/>-0.5 <lb/>0.5 <lb/>0 <lb/>2 <lb/>4 <lb/>6 <lb/>8 <lb/>10 <lb/>Na <lb/>z <lb/>Co <lb/>1-x <lb/>Ir <lb/>x <lb/>O <lb/>2 <lb/>⋅yH <lb/>2 <lb/>O <lb/>x=0.025 <lb/>x=0.02 <lb/>x=0.005 <lb/>x=0.001 <lb/>x=0.0 <lb/>χ (10 -3 <lb/>emu/g) <lb/>T (K) <lb/>H=5G ZFC <lb/>(series A) <lb/>Fig.3 <lb/>Yokoi et al. <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>0 0.005 0.01 0.015 0.02 0.025 0.03 <lb/>Na <lb/>z <lb/>Co <lb/>1-x <lb/>Ir <lb/>x <lb/>O <lb/>2 <lb/>⋅yH <lb/>2 <lb/>O <lb/>T <lb/>c <lb/>(K) <lb/>x <lb/>|dT <lb/>c <lb/>/dx|~1.0(K/%) <lb/>( ) <lb/>( ) <lb/>(series A) <lb/>Fig.4 <lb/>Yokoi et al. <lb/>10.92 <lb/>10.94 <lb/>10.96 <lb/>10.97 <lb/>10.99 <lb/>0 <lb/>0.01 0.02 0.03 0.04 0.05 0.06 <lb/>Na <lb/>0.75 <lb/>Co <lb/>1-x <lb/>Ir <lb/>x <lb/>O <lb/>2 <lb/>c <lb/>(Å) <lb/>x <lb/>(series B) <lb/>Fig.5 <lb/>Yokoi et al. <lb/>0 <lb/>10 <lb/>20 <lb/>30 <lb/>40 <lb/>50 <lb/>60 <lb/>0 <lb/>50 100 150 200 250 300 <lb/>Na <lb/>0.75 <lb/>Co <lb/>1-x <lb/>Ir <lb/>x <lb/>O <lb/>2 <lb/>ρ (mΩcm) <lb/>T (K) <lb/>x=0.05 <lb/>x=0.015 <lb/>x=0.025 <lb/>x=0.02 <lb/>x=0.005 <lb/>x=0.0 <lb/>(series B) <lb/>Fig.6 <lb/>Yokoi et al. <lb/>-7 <lb/>-6 <lb/>-5 <lb/>-4 <lb/>-3 <lb/>-2 <lb/>-1 <lb/>2 <lb/>6 <lb/>8 <lb/>10 <lb/>Na <lb/>z <lb/>Co <lb/>1-x <lb/>Ir <lb/>x <lb/>O <lb/>2 <lb/>⋅yH <lb/>2 <lb/>O <lb/>x=0.05 <lb/>x=0.025 <lb/>x=0.015 <lb/>x=0.01 <lb/>x=0.005 <lb/>x=0.0005 <lb/>x=0.0 <lb/>χ (10 <lb/>-3 <lb/>emu/g) <lb/>T (K) <lb/>H=5 G ZFC <lb/>(series B) <lb/>Fig.7 <lb/>Yokoi et al. <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>0 0.005 0.01 0.015 0.02 0.025 0.03 <lb/>Na <lb/>z <lb/>Co <lb/>1-x <lb/>Ir <lb/>x <lb/>O <lb/>2 <lb/>⋅yH <lb/>2 <lb/>O <lb/>T <lb/>c <lb/>(K) <lb/>x <lb/>|dT c /dx|~1.0(K/%) <lb/>( ) <lb/>( ) <lb/>(series B) <lb/>Fig.8 <lb/>Yokoi et al. <lb/>10.91 <lb/>10.92 <lb/>10.93 <lb/>10.94 <lb/>10.95 <lb/>0 0.005 0.01 0.015 0.02 0.025 0.03 <lb/>Na 0.7 Co 1-x Ga x O 2 <lb/>c <lb/>(Å) <lb/>x <lb/>Fig.9 <lb/>Yokoi et al. <lb/>0 <lb/>5 <lb/>10 <lb/>15 <lb/>0 <lb/>50 100 150 200 250 300 <lb/>Na 0.7 Co 1-x Ga x O 2 <lb/>ρ (mΩcm) <lb/>T (K) <lb/>x=0.05 <lb/>x=0.02 <lb/>x=0.01 <lb/>x=0.0 <lb/>x=0.005 <lb/>Fig.10 <lb/>Yokoi et al. <lb/>-10 <lb/>-8 <lb/>-6 <lb/>-4 <lb/>-2 <lb/>0 <lb/>2 <lb/>0 <lb/>2 <lb/>4 <lb/>6 <lb/>8 <lb/>10 <lb/>Na <lb/>z <lb/>Co <lb/>1-x <lb/>Ga <lb/>x <lb/>O <lb/>2 <lb/>⋅yH <lb/>2 <lb/>O <lb/>x=0.05 <lb/>x=0.005 <lb/>x=0.0025 <lb/>x=0.001 <lb/>x=0.0005 <lb/>x=0.0 <lb/>χ (10 -3 <lb/>emu/g) <lb/>T (K) <lb/>H=5 G ZFC <lb/>Fig.11 <lb/>Yokoi et al. <lb/>0 <lb/>1 <lb/>3 <lb/>4 <lb/>0 0.005 0.01 0.015 0.02 0.025 0.03 <lb/>Na <lb/>z <lb/>Co <lb/>1-x <lb/>Ga <lb/>x <lb/>O <lb/>2 <lb/>⋅yH <lb/>2 <lb/>O <lb/>T <lb/>c <lb/>(K) <lb/>x <lb/>|dT <lb/>c <lb/>/dx|~2.0(K/%) <lb/>( ) <lb/>( ) <lb/>Fig.12 <lb/>Yokoi et al. <lb/>0 <lb/>0.5 <lb/>1 <lb/>1.5 <lb/>2 <lb/>2.5 <lb/>3 <lb/>0 <lb/>50 100 150 200 250 300 <lb/>Na <lb/>0.75 <lb/>Co <lb/>1-x <lb/>Ir <lb/>x <lb/>O <lb/>2 <lb/>x=0.0 <lb/>x=0.025 <lb/>x=0.05 <lb/>χ (10 <lb/>-3 <lb/>emu/mol) <lb/>T (K) <lb/>H=1 T <lb/>(series B) <lb/>Fig.13 <lb/>Yokoi et al. <lb/>0 <lb/>10 <lb/>20 <lb/>30 <lb/>40 <lb/>0 0.005 0.01 0.015 0.02 0.025 0.03 <lb/>T <lb/>c <lb/>(K) <lb/>x <lb/>M=Ir <lb/>M=Ga <lb/>Na <lb/>0.3 <lb/>Co <lb/>1-x <lb/>M <lb/>x <lb/>O <lb/>2 <lb/>⋅yH <lb/>2 <lb/>O <lb/>La <lb/>1.85 <lb/>Sr <lb/>0.15 <lb/>Cu <lb/>1-x <lb/>Zn <lb/>x <lb/>O <lb/>4 <lb/>Mg(B <lb/>1-x <lb/>C <lb/>x <lb/>) <lb/>2 <lb/>Fig.14 <lb/>Yokoi et al. </body>
46
+
47
+
48
+ </text>
49
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/JPS0732912-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>High Pressure Effects on Superconductivity in the β-pyrochlore Oxides AOs 2 O 6 (A=K, Rb, Cs) <lb/>Takaki MURAMATSU, Shigeki YONEZAWA, Yuji MURAOKA and Zenji HIROI <lb/>Institute for Solid States Physics, University of Tokyo, Kashiwa, Chiba 277-8581 <lb/>KEYWORDS: superconductivity, β-pyrochlore oxide, pressure, magnetization <lb/>*E-mail address: [email protected] <lb/></front>
8
+
9
+ <body>Superconductivity in transition metal oxides (TMOs) has been one of exciting fields in <lb/>solid state physics due to its unconventional properties. Many superconductors in TMOs crystallize <lb/>in perovskite related structures. Recently, an exceptional TMO superconductor Cd 2 Re 2 O 7 was found <lb/>with T c = 1.0 K, which crystallizes in the pyrochlore structure. 1) The pyrochlore structure contains a <lb/>corner-sharing tetrahedral network made of transition metal cations called the pyrochlore lattice, and <lb/>is known as one of spin frustration systems in the case that the transition metal cation has a localized <lb/>moment with antiferromagnetic interactions between nearest neighbors. Very recently, three related <lb/>compounds AOs 2 O 6 (A = K, Rb, Cs) named the β-pyrochlore oxides were found, 2-4) which contain <lb/>the pyrochlore lattice made of Os atoms. They show superconductivity at T c = 9.6 K, 6.3 K and 3.3 <lb/>K, respectively. Particularly in KOs 2 O 6 the T c is almost one order higher than in Cd 2 Re 2 O 7 . The <lb/>mechanism of the superconductivity has been extensively studied and was suggested to be <lb/>unconventional. For example, a remarkably large upper critical field H c2 of 38 T exceeding Pauli&apos;s <lb/>limit was reported for KOs 2 O 6 . 5) Moreover, the µSR and NMR experiments suggested an anisotropic <lb/>order parameter. 6,7) On the other hand, there are a few reports which suggest conventional BCS <lb/>superconductivity for RbOs 2 O 6 . 8,9) <lb/>It is plausible to assume that the systematic change of T c in AOs 2 O 6 is due to the size <lb/>effect of alkaline metal ions. The lattice constant is almost proportional to the ionic radius of A ions, <lb/>and the T c increases with decreasing the lattice constant from Cs to K under a positive chemical <lb/>pressure. 4) This is in contrast to the case of conventional BCS superconductivity in a single-band <lb/>model, where the T c decreases under a positive pressure, because the density of state (DOS) <lb/>decreases with decreasing the lattice volume, as typically observed in alkali metal doped C 60 <lb/>superconductors A 3 C 60 . 10) This contrast may indicate the uniqueness of the superconductivity in <lb/>AOs 2 O 6 . Thus, it is interesting to examine physical pressure effects on T c . In this note, we report <lb/>diamagnetic measurements under pressures up to 1.2 GPa using a piston-cylinder type pressure cell <lb/>on the three compounds of AOs 2 O 6 . In the α-pyrochlore oxide superconductor Cd 2 Re 2 O 7 the high <lb/></body>
10
+
11
+ <page>1 <lb/></page>
12
+
13
+ <body>pressure study revealed that T c increases from 1.0 K to 3.0 K at 2 GPa and then decreases to vanish <lb/>above 3 GPa. 11,12) In the case of RbOs 2 O 6 , Khasanov et al. reported a high pressure experiment in <lb/>which the T c increases monotonously up to 1.0 GPa with the initial slope of 0.90 K/GPa. 8) <lb/>Polycrystalline samples were prepared as reported previously. 2-4) They were nearly <lb/>single-phase, but with a small amount of OsO 2 or Os metal. Magnetic susceptibility was measured in <lb/>a Quantum Design MPMS. Quasi-hydrostatic pressures up to 1.2 GPa were produced by a <lb/>piston-cylinder type pressure cell which is made of hardened CuBe alloy. Daphne oil 7373 was used <lb/>as a pressure transmitting medium. The actual pressure was determined in each experiment by <lb/>measuring the superconducting transition temperature of Pb or Sn which was put into the pressure <lb/>medium together with a sample and placed about 2 cm away from the sample. <lb/>A typical temperature dependence of magnetic susceptibility for KOs 2 O 6 is shown in Fig.1. <lb/>In a field of 10 Oe after zero-field cooling, a large diamagnetic signal due to the shielding effect was <lb/>observed at 4 K in the whole pressure range. On heating, it disappears rapidly above 9 K. We <lb/>defined T c in two ways. One is the onset temperature where the magnetic susceptibility starts to <lb/>deviate from a background signal at high temperature. This temperature corresponds to the T c <lb/>determined by specific heat measurements or resistivity measurements on bulk samples at ambient <lb/>pressure in a zero magnetic field. The other characteristic temperature T c &apos; is determined to be the <lb/>temperature where the diamagnetic signal becomes 5% of the maximum shielding signal. <lb/>The pressure dependence of T c as well as T c &apos; is plotted in Fig.2. In RbOs 2 O 6 and CsOs 2 O 6 , <lb/>the T c increases almost linearly with increasing pressure in the whole pressure range examined. The <lb/>slope for RbOs 2 O 6 is 0.78 K/GPa, which is slightly smaller than the value reported by Khasanov. 8) In <lb/>remarkable contrast, the T c of KOs 2 O 6 exhibits a downturn with a maximum value of 10.0 K at 0.56 <lb/>GPa, and then decreases to 9.5 K at 1.20 GPa. In order to compare the three systems, the pressure <lb/>dependence of T c (P) normalized by the T c at ambient pressure (AP) is shown in Fig. 2(b). The initial <lb/>slope of {T c (P)/T c (AP)}/P is 0.20, 0.14 and 0.13 GPa -1 for A = Cs, Rb and K, respectively; the <lb/>largest for CsOs 2 O 6 with the largest unit cell volume and the lowest T c at AP. The initial increase of <lb/>T c by applying physical pressure commonly observed for the three compounds is consistent with the <lb/>trend by chemical pressure. This means that the lattice volume is the key parameter to determine the <lb/>T c of AOs 2 O 6 . The origin of the saturation and the following downturn in T c for KOs 2 O 6 is not clear <lb/>at the moment but may indicate that a fluctuation relevant to the superconductivity is enhanced at <lb/>certain pressure and suppressed of higher pressure. Further experiments at higher pressures are <lb/>necessary to discuss this issure in more detail and are in progress. <lb/>In conclusion, we measured the pressure dependence of magnetization up to 1.2 GPa in <lb/>order to deduce the pressure effect of T c in the three β-pyrochlore oxides. It is found that the initial <lb/>pressure dependence of T c is positive for all the compounds. Only KOs 2 O 6 exhibits a saturation in T c <lb/>at 0.56 GPa and the downturn at higher pressure. <lb/></body>
14
+
15
+ <page>2 <lb/></page>
16
+
17
+ <div type="acknowledgement">This work is supported by a Grant-in-Aid for Scientific Research (14750549) given by the <lb/>Ministry of Education, Culture, Sports, Science and Technology. T. M. thanks for a financial support <lb/>by a Grant-in-Aid for Creative Scientific Research (13NP0201) <lb/></div>
18
+
19
+ <page>3 <lb/></page>
20
+
21
+ <body>Figure caption <lb/>Fig. 1 <lb/>Temperature dependence of magnetic susceptibility of KOs 2 O 6 measured at ambient pressure (AP) <lb/>and high pressures of P = 0.56 and 1.20 GPa. The measurements were carried out on heating after <lb/>zero-field cooling in a magnetic field of 10 Oe. <lb/>Fig. 2. <lb/>Pressure dependence of T c (a) and nomalized one by the value at AP (b) for AOs 2 O 6 (A=K, Rb, Cs). <lb/>The solid and open marks represent the onset T c and T c &apos;. <lb/></body>
22
+
23
+ <listBibl>ref) <lb/>1) M. Hanawa, Y. Muraoka, T. Tayama, T. Sakaibara, J. Yamaura and Z. Hiroi: Phys. Rev. Lett. 87 <lb/>(2001) 187001. <lb/>2) S. Yonezawa, Y. Muraoka, Y. Matsushita and Z. Hiroi: J. Phys.: Condes. Matter 16 (2004) L9. <lb/>3) S. Yonezawa, Y. Muraoka, Y. Matsushita and Z.Hiroi: J. Phys. Soc. Jpn. 73 (2004) 819. <lb/>4) S. Yonezawa, Y. Muraoka and Z. Hiroi: J. Phys. Soc. Jpn 73 (2004) 1655. <lb/>5) Z. Hiroi, S. Yonezawa and Y. Muraoka: J. Phys. Soc. Jpn 73 (2004) 1651. <lb/>6) A. Koda, W. Higemoto, K. Ohishi, S. R. Saha, R. Kadono, S. Yonezawa, Y. Muraoka and Z. <lb/>Hiroi: cond-mat/0402400. <lb/>7) K. Arai, J. Kikuchi, K. Kodama, M. Takigawa, S. Yonezawa, Y. Muraoka, Z. Hiroi:submitted to <lb/>Proceedings of SCES &apos;04. <lb/>8) R. Khasanov, D. G. Eshchenko, J. Karpinski, S. M. Kazakov, N. D. Zhigadlo, R. Brutsch, D. <lb/>Gavillet and H. Keller: cond-mat/0404542.; S. M. Kazakov, N. D. Zhigadlo, Bruhwiler, B. Batlogg, <lb/>J. Karpinski: cond-mat/0403588; M. Bruhwiler, S. M. Kazakov, N. D. Zhigadlo, J. Karpinski, B. <lb/>Batlogg:cond-mat/0403526. <lb/>9) R. Saniz, J. E. Medvedeva, Lin-Hui Ye, T. Shishidou and A. J. Freeman: cond-mat/0406538 <lb/>10) R. M. Fleming, A. P. Ramirez, J. J. Rosseinsky, D. W. Murphy, R. C. Haddon, S. M. Zahurak <lb/>and A. V. Makhija: Nature 353 (1991) 787. <lb/>11) Z. Hiroi, T. Yamauchi, T. Yamada, M. Hanawa, Y. Ohishi, O. Shimomura, M. Abliz, M. Hedo <lb/>and Y. Uwatoko: J. Phys. Soc. Jpn 71 (2002) 1553. <lb/>12) M. Abliz, M. Hedo, Z. Hiroi, T. Matsumoto, M. Hanawa and Y. Uwatoko: J. Phys. Soc. Jpn 72 <lb/>(2003) 3039. <lb/></listBibl>
24
+
25
+ <page>5 <lb/></page>
26
+
27
+ <note place="headnote">T. Muramatsu et al. <lb/></note>
28
+
29
+ <body>Fig. 1 <lb/>-1.0 <lb/>-0.5 <lb/>0.0 <lb/>4πM/H <lb/>11.0 <lb/>10.0 <lb/>9.0 <lb/>8.0 <lb/>T (K) <lb/>KOs 2 O 6 <lb/>H =10 Oe <lb/>0.56 GPa <lb/>AP <lb/>1.20 GPa <lb/>6 <lb/></body>
30
+
31
+ <note place="headnote">T. Muramatsu et al. <lb/></note>
32
+
33
+ <body>Fig. 2 <lb/>(a) <lb/>10 <lb/>5 <lb/>0 <lb/>T <lb/>c (K) <lb/>1.5 <lb/>1.0 <lb/>0.5 <lb/>0.0 <lb/>P (GPa) <lb/>AOs 2 O 6 <lb/>Cs <lb/>Rb <lb/>K <lb/>(b) <lb/>1.3 <lb/>1.2 <lb/>1.1 <lb/>1.0 <lb/>0.9 <lb/>T <lb/>c /T <lb/>c (AP) <lb/>1.5 <lb/>1.0 <lb/>0.5 <lb/>0.0 <lb/>P (GPa) <lb/>AOs 2 O 6 <lb/>Rb <lb/>Cs <lb/>K <lb/></body>
34
+
35
+ <page>7 </page>
36
+
37
+
38
+ </text>
39
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L087187001-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>arXiv:cond-mat/0106521v1 [cond-mat.str-el] 26 Jun 2001 <lb/>Superconductivity in a pyrochlore oxide Cd 2 Re 2 O 7 <lb/>Hironori Sakai 1,2 , Kazuyoshi Yoshimura 1 , Hiroyuki Ohno 1 , <lb/>Harukazu Kato , Shinsaku Kambe 2 , Russell E. Walstedt 2 , <lb/>Tatsuma D. Matsuda 2 , Yoshinori Haga 2 , <lb/>and YoshichikaŌnuki 2,3 <lb/>1 Department of Chemistry, Graduate School of Science, Kyoto University, Kyoto <lb/>606-8502, JAPAN <lb/>Advanced Science Research Center, Japan Atomic Energy Research Institute, <lb/>Tokai, Ibaraki 319-1195, JAPAN <lb/>3 Graduate School of Science, Osaka University, Toyonaka, Osaka 560-0043, JAPAN <lb/>Abstract. We make the first report that a metallic pyrochlore oxide Cd 2 Re 2 O 7 , <lb/>exhibits type II superconductivity at 1.1 K. The pyrochlore oxide is known to be <lb/>a geometrical frustrated system, which includes the tetrahedral network of magnetic <lb/>ions. A large number of compounds are classified in the family of pyrochlore oxides, and <lb/>these compounds exhibit a wide variety of physical properties ranging from insulator <lb/>through semiconductor and from bad metal to good metal. Until now, however, <lb/>no superconductivity has been reported for frustrated pyrochlore oxides. The bulk <lb/>superconductivity of this compound is confirmed by measurements of the resistivity and <lb/>the a. c. magnetic susceptibility. The H c2 , which is extrapolated to 0 K, is estimated as <lb/>about 0.8 T, using the resistivity measurements under aplied field. The plot of H c2 vs T <lb/>indicates that the Cooper pairs are composed of rather heavy quasiparticles. This fact <lb/>suggests that frustrated heavy electrons become superconducting in this compound. <lb/>PACS numbers: 74.10, 74.60, 74.70 <lb/></front>
8
+
9
+ <note place="headnote">Superconductivity in a pyrochlore oxide Cd 2 Re 2 O 7 <lb/></note>
10
+
11
+ <page>2 <lb/></page>
12
+
13
+ <body>Recently the subject of geometrical frustration in strongly correlated electron <lb/>systems has attracted considerable interest. The ground states of these systems are <lb/>expected to be highly degenerate. Such high degeneracies lead to thermodynamic <lb/>instability at low temperatures. Lifting these degeneracies makes possible the production <lb/>of exotic quantum ground states, such as spin-liquid, heavy fermion, and unconventional <lb/>superconductivity. <lb/>The pyrochlore oxide, which has a general formula A 2 B 2 O 6 O&apos;, contains a tetrahedral <lb/>network of A or B cations, leading to geometrical frustration. The A cations are eight-<lb/>fold coordinated with six O and two O&apos; anions and are located within distorted cubes, <lb/>while the smaller B cations are six-fold coordinated and are located within distorted <lb/>octahedra of which the six bond lengths from the central B cation to the corner O anion <lb/>are equal. The corner-sharing BO 6 octahedra compose a three-dimensional network as <lb/>shown in Fig. 1a. The sublattice of B cations composes a three-dimensional corner-<lb/>shared tetrahedral network, namely, the pyrochlore lattice as shown in Fig.1b. If <lb/>the A cations are non-magnetic and the B cations are magnetic, the B-spin magnetic <lb/>couplings are strongly frustrated under a nearest-neighbor antiferromagnetic exchange <lb/>interaction. For the case of localized electron spin moments, theoretical studies of the <lb/>Heisenberg model on the pyrochlore lattice have suggested that the ground state of such <lb/>an insulator would be long-range magnetic order[1], spin-freezing[2], or quantum spin <lb/>liquid[3]. The geometrical frustration plays a crucial role even in the case of itinerant <lb/>electrons in the pyrochlore lattice. Indeed, the metallic spinel oxide LiV 2 O 4 , which has <lb/>a pyrochlore lattice of vanadium, has been reported to show heavy fermion behavior at <lb/>low temperatures due to the geometrical frustration[4]. <lb/>A large number of compounds are classified in the family of pyrochlore oxides, and <lb/>these compounds exhibit a wide variety of physical properties ranging from insulator <lb/>through semiconductor and from bad metal to good metal[5]. Until now, however, <lb/>no superconductivity has been reported for frustrated pyrochlore oxides. Here, we <lb/>make the first report that a metallic cubic pyrochlore oxide Cd 2 Re 2 O 7 , exhibits type II <lb/>superconductivity at 1.1 K. <lb/>Polycrystalline samples of Cd 2 Re 2 O 7 were prepared by solid state reaction. A <lb/>stoichiometric mixture of CdO, ReO 3 , and Re metal was pelletized, and put into an <lb/>alumina Tammann tube. The pellet in the Tammann tube was further inserted into <lb/>an evacuated silica tube and preheated at 300˚C for several hours in order to avoid a <lb/>vaporization of the starting materials. Then, the pellet was heated at 1000˚C for several <lb/>hours. The powder XRD pattern measured at room temperature was identified as the <lb/>cubic pyrochlore structure with a lattice parameter a=10.221Å, which is consistent with <lb/>the previous report[6]. <lb/>The direct current (d. c.) electrical resistivity of the sintered sample of Cd 2 Re 2 O 7 <lb/>was measured using a standard four-probe technique in the temperature range of 0.3 to <lb/>300 K under applied field from 0 to 2 T. Low-temperature measurements below 1.9 K <lb/>were performed using a 3 He refrigerator. Figure 2 shows the temperature dependence <lb/>of the d.c. electrical resistivity of Cd 2 Re 2 O 7 under zero applied field. The electrical <lb/></body>
14
+
15
+ <note place="headnote">Superconductivity in a pyrochlore oxide Cd 2 Re 2 O 7 <lb/></note>
16
+
17
+ <page>3 <lb/></page>
18
+
19
+ <body>resistivity shows a steep descent below T * ∼ 200 K. This anomaly at T * was observed <lb/>in the d.c. magnetic susceptibility measurement as well. The origin of this anormaly <lb/>has not been identified. The electrical resistivity drops to zero sharply at the onset <lb/>superconducting temperature T c =1.1 K, and shows effectively zero resistivity below <lb/>1.05 K. Changes of driving electric current density produced slight differences in the <lb/>resistivity below 1.7 K, an effect which may be due to the superconductivity of a small <lb/>amount of impurity rhenium metal (T c =1.7 K). The superconductivity transition at 1.1 <lb/>K is not due to filamentary superconductivity of Re, since the observed critical field is <lb/>much larger than the critical field of Re (0.02 T), as will be described below. The large <lb/>residual resistivity of 4 × 10 −3 Ω•cm may suggest a low carrier density of this system. <lb/>From band calculations, it is pointed out that the system has a very small density of <lb/>states at the Fermi level, which is located within the 5d-band and in the valley between <lb/>the very flat bands of the rhenium 5d electrons[7]. However, the electronic specific heat <lb/>coefficient has been found to be large, i. e., γ=13.3 mJ/Re mol• K 2 [8], indicating that <lb/>heavy quasiparticles are formed due to the spin-frustration in this compound. <lb/>The alternating current (a. c.) magnetic susceptibility was measured by a mutual-<lb/>inductance method at a magnetic field of 2×10 −5 T in the temperature range of 0.3 to <lb/>1.8 K. As shown in Fig. 3, a strong diamagnetic signal (χ&apos;) has been observed below <lb/>1.06 K, which corresponds to the end-point transition temperature of the electrical <lb/>resistivity measurement. The dissipative component χ&quot; shows only a small peak around <lb/>T c , indicating no weak superconducting link between superconducting grains. The <lb/>superconducting volume fraction was estimated roughly as ∼ 50 % at 0.3 K. From this <lb/>experiment, a bulk superconducting state has been strongly confirmed in Cd 2 Re 2 O 7 <lb/>below 1.1 K. <lb/>To estimate the superconducting critical field, the magnetic field dependence <lb/>of T c has been determined by resistivity measurements. As the magnetic field was <lb/>applied, the sample maintained zero resistivity until the field reached H c2 . At H c2 <lb/>the superconductivity was quenched abruptly as shown in Fig.4. The value of the upper <lb/>critical field, which is extrapolated to zero temperature in the plot of H c2 vs T (Fig. 4, <lb/>inset) using the WHH model[9], is estimated as H c2 (0)=0.8 T. The superconducting <lb/>coherence length, ξ is expressed as <lb/>φ 0 <lb/>2πH c2 (T ) (where φ 0 is the fluxoid quantum). The <lb/>Pippard coherence length, ξ 0 =h v F <lb/>π∆(0) (where v F is the Fermi velocity, and ∆(0) is <lb/>the superconducting energy gap at T =0 K) was obtained as about 300Å using the <lb/>relation, ξ(T ) = 0.74ξ 0 <lb/>1− T <lb/>Tc <lb/>for a clean superconductor. This compound is considered to <lb/>be a clean superconductor, since the mean free path l is estimated as ∼ 4 × 10 −8 <lb/>m &gt; ξ 0 , using the Drude relation, ρ =h (3π 2 ) <lb/>1 <lb/>3 <lb/>e 2 l n − 2 <lb/>3 with resistivity ρ ∼ 4 × 10 −5 Ω• <lb/>m and carrier concentration n ∼ 1 × 10 24 m −3 which is a typical value for metallic <lb/>pyrochlore compounds. The observed large initial slope of H c2 , i.e., 1 <lb/>Tc <lb/>dH c2 <lb/>dT <lb/>T =Tc <lb/>∼ <lb/>1 T/K 2 , indicates that the Cooper pairs are composed of rather heavy quasiparticles, <lb/>since 1 <lb/>Tc <lb/>dH c2 <lb/>dT <lb/>T =Tc <lb/>is proportional to the square of effective mass m * 2 , in agreement with <lb/>the large γ value. This fact suggests, indeed, that frustrated heavy electrons become <lb/></body>
20
+
21
+ <note place="headnote">Superconductivity in a pyrochlore oxide Cd 2 Re 2 O 7 <lb/></note>
22
+
23
+ <page>4 <lb/></page>
24
+
25
+ <body>superconducting in this compound. <lb/>Generally, superconductivity due to a magnetic interaction competes with magnetic <lb/>frustration, since no particular magnetic excitation for a superconducting attractive <lb/>channel exists due to the frustration. The present study gives the first example of <lb/>coexistence of magnetic frustration and superconductivity. <lb/></body>
26
+
27
+ <listBibl>References <lb/>[1] Anderson P W 1956 Phys. Rev. 1008 <lb/>[2] Villain J 1979 Z. Phys. B 33 31 <lb/>[3] Reimers J N 1992 Phys. Rev. B 45 7287 <lb/>[4] Kondo S et al 1997 Phys. Rev. Lett. 3729 <lb/>[5] Subramanian M A, Aravamudan G and Subba Rao G V 1983 Prog. Solid State Chem. 15 55 <lb/>[6] Donohue P C, Longo J M, Rosenstein R D, and Katz L 1965 Inorg. Chem. 1153 <lb/>[7] Harima H et al private communication <lb/>[8] Blacklock K and White H W 1979 J.Chem.Phys 71 5287 <lb/>[9] Werthamer N R, Helfand E, and Hohenberg P C 1966 Phys. Rev. 295 <lb/></listBibl>
28
+
29
+ <note place="headnote">Superconductivity in a pyrochlore oxide Cd 2 Re 2 O 7 <lb/></note>
30
+
31
+ <page>5 <lb/></page>
32
+
33
+ <body>a <lb/>a <lb/>a <lb/>a <lb/>b <lb/>BO 6 octahedron <lb/>A cation <lb/>B cation <lb/>O&apos; anion <lb/>Fig. 1 H. Sakai et al. <lb/>Figure 1. Crystal structure of pyrochlore oxide. Figure 1a is drawn on a basis of the <lb/>network of BO 6 octahedra. Figure 1b shows the pyrochlore lattice of B cations. <lb/></body>
34
+
35
+ <note place="headnote">Superconductivity in a pyrochlore oxide Cd 2 Re 2 O 7 <lb/></note>
36
+
37
+ <page>6 <lb/></page>
38
+
39
+ <body>5000 <lb/>0 <lb/>ρ ( µΩ • cm ) <lb/>1.8 <lb/>1.6 <lb/>1.4 <lb/>1.2 <lb/>1.0 <lb/>0.8 <lb/>T ( K ) <lb/>J = 0.008 A/cm 2 <lb/>J = 0.23 A/cm 2 <lb/>J = 0.08 A/cm 2 <lb/>2.5x10 <lb/>4 <lb/>2.0 <lb/>1.5 <lb/>1.0 <lb/>0.5 <lb/>0.0 <lb/>ρ ( µΩ • cm ) <lb/>300 <lb/>200 <lb/>100 <lb/>0 <lb/>T ( K ) <lb/>Fig. 2 H. Sakai et al. <lb/>Figure 2. Temperature dependence of d. c. electrical resistivity for Cd 2 Re 2 O 7 in the <lb/>temperature range of 0.8 to 1.8 K. The inset shows the data in the range of 0.3 to 300 <lb/>K. <lb/></body>
40
+
41
+ <note place="headnote">Superconductivity in a pyrochlore oxide Cd 2 Re 2 O 7 <lb/></note>
42
+
43
+ <page>7 <lb/></page>
44
+
45
+ <body>0.5 <lb/>0.4 <lb/>0.3 <lb/>0.2 <lb/>0.1 <lb/>0.0 <lb/>-0.1 <lb/>χ&apos; <lb/>ac , <lb/>χ&apos;&apos; <lb/>ac ( arb. units ) <lb/>1.6 <lb/>1.2 <lb/>0.8 <lb/>0.4 <lb/>T ( K ) <lb/>χ&apos; ac <lb/>χ&apos;&apos; ac <lb/>Fig. 3 H. Sakai et al. <lb/>Figure 3. Temperature dependence of a.c. magnetic susceptibility for Cd 2 Re 2 O 7 . χ&apos; <lb/>and χ&quot; represent the real part and the imaginary part, respectively. <lb/></body>
46
+
47
+ <note place="headnote">Superconductivity in a pyrochlore oxide Cd 2 Re 2 O 7 <lb/></note>
48
+
49
+ <page>8 <lb/></page>
50
+
51
+ <body>5000 <lb/>0 <lb/>ρ ( µΩ • cm ) <lb/>2.0 <lb/>1.5 <lb/>1.0 <lb/>0.5 <lb/>0.0 <lb/>H ( T ) <lb/>T = 0.42 K <lb/>T = 0.49 K <lb/>T = 0.62 K <lb/>T = 0.69 K <lb/>T = 0.79 K <lb/>T = 0.90 K <lb/>T = 1.01 K <lb/>T = 1.10 K <lb/>T = 0.38 K <lb/>0.8 <lb/>0.6 <lb/>0.4 <lb/>0.2 <lb/>0.0 <lb/>H <lb/>C2 ( T ) <lb/>1.2 <lb/>0.8 <lb/>0.4 <lb/>0.0 <lb/>T ( K ) <lb/>Fig. 4 H. Sakai et al. <lb/>Figure 4. Magnetic field dependence of the resistivity for Cd 2 Re 2 O 7 . The inset shows <lb/>the plot of H c2 vs T . </body>
52
+
53
+
54
+ </text>
55
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L089147002-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>arXiv:cond-mat/0111187v1 [cond-mat.supr-con] 10 Nov 2001 <lb/>Temperature Dependence of the Magnetic Penetration Depth in the Vortex State of <lb/>the Pyrochlore Superconductor, Cd 2 Re 2 O 7 <lb/>M. D. Lumsden, 1 S. R. Dunsiger, J. E. Sonier, 3 R. I. Miller, 4 R. F. Kiefl, 4, 5, 9 <lb/>R. Jin, 1 J. He, 6 D. Mandrus, 1, 6 S. T. Bramwell, and J. S. Gardner <lb/>1 Solid State Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831 <lb/>2 Los Alamos National Laboratory, MST-10, MS K764 Los Alamos, NM 87545 <lb/>3 Department of Physics, Simon Fraser University, British Columbia Canada V5A 1S6 <lb/>4 Department of Physics and Astronomy, University of British Columbia, V6T 1Z1, Canada <lb/>TRIUMF, Vancouver, British Columbia V6T 2A3, Canada <lb/>6 Department of Physics and Astronomy, University of Tennessee, Knoxville, TN 37996 <lb/>7 Department of Chemistry, University College London, London WC1H OAJ, U.K. <lb/>National Research Council, NPMR, Chalk River Laboratories, Chalk River, Ontario, K0J 1J0, Canada <lb/>9 Canadian Institute for Advanced Research, 180 Dundas Street West, Toronto, Ontario, M5G 1Z8, Canada <lb/>(Dated: October 31, 2018) <lb/>We report transverse field and zero field muon spin rotation studies of the superconducting rhe-<lb/>nium oxide pyrochlore, Cd2Re2O7. Transverse field measurements (H=0.007 T) show line broaden-<lb/>ing below Tc, which is characteristic of a vortex state, demonstrating conclusively the type-II nature <lb/>of this superconductor. The penetration depth is seen to level off below about 400 mK (T /Tc ∼ 0.4), <lb/>with a rather large value of λ(T = 0) ∼ 7500Å. The temperature independent behavior below ∼ <lb/>mK is consistent with a nodeless superconducting energy gap. Zero-field measurements indicate no <lb/>static magnetic fields developing below the transition temperature. <lb/>PACS numbers: 74.70.Dd, 74.60.-w, 76.75.+i <lb/></front>
8
+
9
+ <body>The pyrochlore transition metal oxides, of general for-<lb/>mula A 2 B 2 O 7 , have been the topic of much interest in <lb/>recent years as they represent ideal systems for study-<lb/>ing the effects of geometrical frustration [1]. Both the A <lb/>and B sublattices form a network of corner-sharing tetra-<lb/>hedra such that it may not be possible to energetically <lb/>satisfy all the magnetic interactions simultaneously. The <lb/>resultant geometric frustration leads to the formation of <lb/>exotic ground states. Much of the recent work has con-<lb/>centrated on local moment systems where novel proper-<lb/>ties such as cooperative paramagnetism [2], partial, non-<lb/>collinear antiferromagnetic ordering [3, 4], spin-freezing <lb/>[5], and dipolar &quot;spin-ice&quot; behavior [6, 7] have been ob-<lb/>served. There has, however, been growing interest in the <lb/>interplay between itinerant and local moments in geo-<lb/>metrically frustrated systems. The metallic pyrochlore <lb/>Nd 2 Mo 2 O 7 exhibits a large anomalous Hall effect which <lb/>has been attributed to the Berry phase produced by spin <lb/>chirality on the pyrochlore lattice [8], while the spinel <lb/>compound, LiV 2 O 4 has been claimed to represent the <lb/>first known transition metal heavy-fermion system and <lb/>evidence exists that the unusual properties of this mate-<lb/>rial are related to geometrical frustration on the spinel <lb/>lattice. [9] <lb/>A vast body of work has been carried out on 3d and <lb/>4d transition metal pyrochlores. These are generally in-<lb/>sulators and possess either a spin-glass-like or long-range <lb/>ordered magnetic structure. In contrast, the 5d transition <lb/>metal pyrochlores are mainly metallic, resulting from the <lb/>extended nature of the 5d orbitals. The exception to this <lb/>is Cd 2 Os 2 O 7 [10] where the 5d 3 configuration of Os 5+ <lb/>results in a half-filled t 2g band and a metal-insulator <lb/>transition at 226 K. Despite the large number of transi-<lb/>tion metal compounds which crystallize in the pyrochlore <lb/>structure and the wide range of physical phenomena ob-<lb/>served in these materials, superconductivity had not been <lb/>observed until the recent discovery of bulk superconduc-<lb/>tivity in the 5d pyrochlore, Cd 2 Re 2 O 7 [11, 12]. <lb/>Cd 2 Re 2 O 7 crystallizes in the pyrochlore structure with <lb/>room temperature lattice constant a=10.219Å and an <lb/>oxygen positional parameter x=0.3089.[13] Recent inves-<lb/>tigations [11, 12, 14, 15, 16] have demonstrated the ex-<lb/>istence of two phase transitions in this compound. The <lb/>first, occurring at a temperature of about 200 K, is a <lb/>continuous structural transition which is accompanied <lb/>by drastic changes in resistivity and magnetic suscep-<lb/>tibility [15, 16]. On further lowering the temperature, <lb/>Cd Re 2 O 7 has been shown to exhibit bulk superconduc-<lb/>tivity below a sample dependent transition temperature <lb/>of about 1 K [11, 12, 14]. Preliminary measurements in <lb/>the superconducting state indicate that Cd 2 Re 2 O 7 is a <lb/>type-II superconductor with H c1 less than 0.002 T and <lb/>estimates of the upper critical field, H c2 , ranging from <lb/>0.2 T to 1 T [11, 12, 14]. None of the measurements <lb/>reported to date extend below 0.3 K (T /T c ∼0.3) and <lb/>hence, little can be concluded about the symmetry of the <lb/>order parameter in this system. An exponential form of <lb/>the specific heat as T approaches zero was speculated by <lb/>Hanawa et al. [12] but they point out that measurements <lb/>to lower temperatures are clearly needed. We report the <lb/>first measurements on Cd 2 Re 2 O 7 below 300 mK, tem-<lb/>peratures which are necessary (for T c ∼ 1 K) to extract <lb/></body>
10
+
11
+ <page>2 <lb/></page>
12
+
13
+ <body>information about the superconducting order parameter <lb/>symmetry. We have performed transverse field (TF) and <lb/>zero field (ZF) muon spin rotation (µSR) measurements <lb/>on single crystal samples of Cd 2 Re 2 O 7 . The ZF-µSR <lb/>measurements reveal very small internal magnetic fields <lb/>which are characteristic of nuclear dipoles, indicating no <lb/>significant electronic magnetism either above or below <lb/>T c . The TF-µSR results provide the first measurement <lb/>of the internal field distribution in the vortex state in this <lb/>material. In particular, temperature dependent studies <lb/>from 20 mK to 4 K indicate a penetration depth which <lb/>levels off as T → 0, suggestive of a fully gapped Fermi <lb/>surface with a rather large zero temperature value of the <lb/>penetration depth, λ(0)∼7500Å. <lb/>Muon spin rotation has proven to be a very effective <lb/>probe in the study of superconductivity [17]. In partic-<lb/>ular, TF-µSR provides a measure of the length scales <lb/>associated with type-II superconductors, the penetration <lb/>depth, λ and the vortex core radius r 0 [17]. In a TF-µSR <lb/>experiment, spin polarized muons, with polarization per-<lb/>pendicular to the applied magnetic field direction, are <lb/>implanted in a sample at a location which is random on <lb/>the length scale of the vortex lattice. The muon precesses <lb/>at a rate proportional to the local magnetic field provid-<lb/>ing a measure of the local field distribution, n(B). The <lb/>presence of the vortex lattice results in a spatially inho-<lb/>mogeneous field distribution and a resulting muon spin <lb/>depolarization. <lb/>Early TF-µSR measurements assumed a Gaussian dis-<lb/>tribution of magnetic fields and with this approximation, <lb/>the penetration depth can be directly obtained from the <lb/>Gaussian depolarization rate, σ ∼ 1/λ 2 . This approx-<lb/>imation has been shown to be reasonable for the case <lb/>of polycrystalline samples but is inadequate for the case <lb/>of single crystals [17]. In this case, a Ginzburg-Landau <lb/>(GL) model has been developed to model the magnetic <lb/>field distribution for a single crystal. In GL theory, the <lb/>size of the vortex core is determined by the applied mag-<lb/>netic field, H, and the GL coherence length normal to the <lb/>applied field, ξ GL , while the penetration depth provides <lb/>the length scale of the decay of magnetic field away from <lb/>the vortex core. The field distribution is calculated from <lb/>the spatial distribution of magnetic field [18], <lb/>B(r) = <lb/>Φ 0 <lb/>S <lb/>(1 − b 4 ) <lb/>G <lb/>e −iG•r uK 1 (u) <lb/>1 + λ 2 G 2 , <lb/>(1) <lb/>where u 2 = 2ξ 2 <lb/>GL G 2 (1 + b 4 )[1 − 2b(1 − b 2 )], K 1 (u) is a <lb/>modified Bessel function, G is a reciprocal lattice vector <lb/>of the vortex lattice, b = H/H c2 is the reduced field, Φ 0 <lb/>is the flux quantum and S is the area of the reduced unit <lb/>cell for a hexagonal vortex lattice. <lb/>Single crystals of Cd Re 2 O 7 were grown using vapor-<lb/>transport techniques as described elsewhere [14, 19]. <lb/>Three samples with an approximate surface area of 5×5 <lb/>mm 2 each were mounted using low temperature grease <lb/>FIG. 1: Typical µSR spectra in Cd2Re2O7 obtained in a <lb/>transverse magnetic field of 0.007 T at temperatures of (a) <lb/>T=1.5 K (above Tc) and (b) T=100 mK (below Tc). <lb/>such that the cubic (100) direction would be parallel to <lb/>the applied magnetic field direction. They were mounted <lb/>on intrinsic GaAs in order to eliminate any precession sig-<lb/>nal at the background frequency [20] from muons which <lb/>miss the sample and would otherwise land in the Ag sam-<lb/>ple holder. The samples were covered with 0.025 mm Ag <lb/>foil which was bolted to the sample holder to ensure tem-<lb/>perature uniformity. The TF and ZF-µSR measurements <lb/>were performed in an Oxford Instruments dilution refrig-<lb/>erator on the M15 beamline at TRIUMF at temperatures <lb/>from 20 mK up to 4 K. <lb/>Given the estimated critical field values, we selected a <lb/>field value of 0.007 T and the temperature dependence <lb/>was measured by cooling the sample in the presence of <lb/>this applied magnetic field to ensure a uniform flux line <lb/>lattice. Figs. 1(a) and (b) show typical µSR spectra in <lb/>a transverse field of 0.007 T, for temperatures above and <lb/>below T c respectively. Examination of this data clearly <lb/>shows an enhanced depolarization rate on entering the <lb/>superconducting state resulting from the inhomogeneous <lb/>field distribution associated with the flux line lattice. <lb/>This represents the first experimental observation of the <lb/>vortex lattice in Cd Re 2 O 7 and provides clear evidence <lb/></body>
14
+
15
+ <page>3 <lb/></page>
16
+
17
+ <body>FIG. 2: Linewidth parameter σ as a function of temperature <lb/>in a transverse magnetic field of 0.007 T applied along the <lb/>(100) direction. The solid line is a fit to Eq. 2. <lb/>that this material is a type-II superconductor. The ob-<lb/>served increase in the TF line broadening below T c can <lb/>be attributed entirely to the vortex lattice since the ZF <lb/>muon spin relaxation rate (not shown here) was small <lb/>and roughly temperature independent below 2 K. <lb/>The solid lines shown in Figs. 1(a) and (b) represent <lb/>fits of the individual time spectra to a sample signal con-<lb/>sisting of a Gaussian envelope with fixed asymmetry and <lb/>a background signal with fixed linewidth and asymme-<lb/>try. The background, from muons which miss the sample <lb/>and land in the heat shields, was obtained independently <lb/>by performing measurements with the sample removed. <lb/>The resulting sample linewidth, σ, is shown in Fig. 2 <lb/>as a function of temperature. As one can clearly see, <lb/>the magnitude of the depolarization rate as T → 0 is <lb/>very small, saturating at a value of about 0.1 µs −1 . The <lb/>residual linewidth from nuclear dipoles, taken from the <lb/>data above T c , is very small in Cd Re 2 O 7 (about 0.03 <lb/>µs −1 ), allowing for clear observation of the line broaden-<lb/>ing associated with the flux line lattice. Apparently the <lb/>muons stop in sites which are not close to the Re ions, <lb/>which have appreciable nuclear moments. <lb/>As can be seen from Eq. 1, the field distribution de-<lb/>pends on both the penetration depth and GL coherence <lb/>depth. The GL coherence length can be obtained from <lb/>the known value of the upper critical field using the ex-<lb/>pression ξ GL = (Φ 0 /2πH c2 ) 1/2 where Φ 0 is the flux quan-<lb/>tum. As mentioned above, a range of values for H c2 <lb/>have been reported and consequently, to provide a self-<lb/>consistent measurement of λ, the field dependence of the <lb/>linewidth was measured. To account for any possible <lb/>instrumental field-dependence in the linewidth, measure-<lb/>ments were made above the transition temperature (2 <lb/>K) at each field value after which the sample was field-<lb/>cooled to 100 mK. The measured linewidth in the nor-<lb/>mal state was subtracted in quadrature from that ob-<lb/>FIG. 3: Linewidth parameter σF C as a function of magnetic <lb/>field applied along the (100) direction. The normal state con-<lb/>tribution has been subtracted as described in the text. Mea-<lb/>surements were taken at a temperature of 100 mK. <lb/>served at 100 mK and the results are plotted in Fig. 3 <lb/>as a function of applied magnetic field. As can be clearly <lb/>seen, the linewidth decreases almost linearly with applied <lb/>field. This is attributed to the linear increase in the vol-<lb/>ume taken up by the vortices. The linewidth parameter, <lb/>σ F C , approaches zero at a field of 0.5 T which is our <lb/>estimate of H c2 (T→0) and is consistent with measure-<lb/>ments on other samples. This estimate of the critical <lb/>field corresponds to ξ GL ∼ 260Å. Using this value, we <lb/>obtain the penetration depth using the field distribution <lb/>shown in Eq. 1. The resulting temperature dependent <lb/>penetration depth is shown in Fig. 4. As expected from <lb/>the small values of linewidth, at the base temperature <lb/>we observe a rather large value of the penetration depth, <lb/>λ(0) ∼ 7500Å. We note that this value of penetration <lb/>depth is significantly larger than most oxide supercon-<lb/>ductors where values ranging from 1000-2000Å are typ-<lb/>ical [21, 22]. <lb/>As the penetration depth is related to the concentra-<lb/>tion of superconducting carriers, its temperature depen-<lb/>dence is a measure of the low-lying electronic excitations. <lb/>As such, the presence of a nodeless superconducting en-<lb/>ergy gap is indicated by a leveling off of the penetra-<lb/>tion depth as the temperature decreases below T c . As is <lb/>clearly seen in Figs. 2 and 4, the linewidth and penetra-<lb/>tion depth respectively become temperature independent <lb/>as the temperature decreases below about 0.4 K, consis-<lb/>tent with a fully gapped Fermi surface. Consequently, <lb/>we conclude that the superconducting order parameter in <lb/>Cd Re 2 O 7 is consistent with a nodeless energy gap sug-<lb/>gesting either s-wave symmetry or exotic pairing symme-<lb/>tries, such as p-wave which can also exhibit a fully gapped <lb/>Fermi surface. For comparison, the solid line in Figs. 2 <lb/></body>
18
+
19
+ <page>4 <lb/></page>
20
+
21
+ <body>FIG. 4: <lb/>Penetration depth as a function of temperature <lb/>in a magnetic field of 0.007 T applied parallel to the (100) <lb/>direction. The solid line is a fit to Eq. 2. <lb/>and represent fits to the two fluid approximation <lb/>σ(T ) <lb/>σ(0) <lb/>∼ <lb/>λ 2 (0) <lb/>λ 2 (T ) <lb/>= [1 − (T /T c ) 4 ], <lb/>(2) <lb/>while the dashed line in Fig. 4 is a fit to the BCS tem-<lb/>perature dependence <lb/>λ(T ) = λ(0) 1 + <lb/>π∆ 0 <lb/>2T <lb/>exp <lb/>−∆ 0 <lb/>T <lb/>(3) <lb/>where ∆ 0 = 1.74(11) K. <lb/>The London penetration depth, λ, provides a direct <lb/>measure of the ratio of superconducting carrier concen-<lb/>tration to effective mass, n s /m * , <lb/>1 <lb/>λ 2 = <lb/>4πn s e 2 <lb/>m * c 2 1 + <lb/>ξ 0 <lb/>l <lb/>−1 <lb/>, <lb/>(4) <lb/>where ξ 0 is the Pippard coherence length, and l is the <lb/>mean-free path. There is considerable uncertainty in <lb/>estimations of the mean-free path with reported values <lb/>ranging from 200-700Å[14, 23] and it is unclear whether <lb/>Cd 2 Re 2 O 7 is a superconductor in the clean or dirty limit. <lb/>If the material is in the clean limit, the present results <lb/>provide strong evidence for a fully gapped Fermi surface. <lb/>Under the assumption of a clean superconductor, such <lb/>that ξ 0 /l ≪ 1, a value of n s m e /m * of 5.0×10 25 m −3 <lb/>can be obtained using Eq. 4 and the measured penetra-<lb/>tion depth. On the other hand, if l ∼200Å (i.e. the dirty <lb/>limit) then we obtain ξ 0 ∼470Å and n s m e /m * ∼1.4×10 26 <lb/>m −3 . Clearly, precise determination of the mean free <lb/>path for Cd Re 2 O 7 is needed to allow accurate quantita-<lb/>tive information to be extracted. <lb/>In conclusion, we have performed µSR studies of the <lb/>superconducting state in the recently discovered py-<lb/>rochlore superconductor, Cd Re 2 O 7 . Zero-field measure-<lb/>ments indicate no significant magnetism in this super-<lb/>conductor, suggesting that magnetic frustration does not <lb/>play a direct role in the superconductivity. Transverse-<lb/>field measurements show that Cd Re 2 O 7 is a type-II su-<lb/>perconductor and indicate a superconducting order pa-<lb/>rameter consistent with a fully gapped Fermi surface <lb/>with a zero temperature value of penetration depth of <lb/>∼7500Å. However, considering that the superconductor <lb/>may be in the dirty limit, spectroscopic techniques which <lb/>directly measure the density of states would be required <lb/>to confirm this conclusion. <lb/></body>
22
+
23
+ <div type="acknowledgement">We would like to acknowledge valuable discussions <lb/>with M. Yethiraj, as well as the technical support of the <lb/>TRIUMF facility, in particular B. Hitti and M. Good. <lb/>Oak Ridge National Laboratory is managed by UT-<lb/>Battelle, LLC for the U.S. Department of Energy under <lb/>contract DE-AC05-00OR22725. Work at Los Alamos Na-<lb/>tional Laboratory was performed under the auspices of <lb/>the U.S. Department of Energy. <lb/></div>
24
+
25
+ <listBibl>[1] for recent reviews see Magnetic Systems with Com-<lb/>peting Interactions, edited by H. T. Diep (World Sci-<lb/>entific, Singapore, 1994); A. P. Ramirez, Annu. Rev. <lb/>Mater. Sci. 24, 453 (1994); P. Schiffer and A. P. Ramirez, <lb/>Comm. Cond. Mat. Phys. 18, 21, (1996). <lb/>[2] J. S. Gardner et al. , Phys. Rev. Lett. 82, 1012 (1999). <lb/>[3] N. P. Raju et al., Phys. Rev. B 59, 14489 (1999). <lb/>[4] J.D.M. Champion et al., Phys. Rev. B 64, 140407(R) <lb/>(2001). <lb/>[5] See, for instance N. P. Raju, E. Gmelin, R. K. Kremer, <lb/>Phys. Rev. B 46, 5405 (1992); M.J.P. Gingras et al., <lb/>Phys. Rev. Lett. 78, 947, (1997). <lb/>[6] M.J. Harris et al., Phys. Rev. Lett. 79, 2554 (1997); M.J. <lb/>Harris et al., Phys. Rev. Lett. 81, (1998) and S.T. <lb/>Bramwell and M.J. Harris, J. Phys.:Condens. Matter 10, <lb/>L215 (1998). <lb/>[7] A.P. Ramirez et al., Nature 399, (1999) and R. Sid-<lb/>dharthan et al., Phys. Rev. Lett. 83, 1854 (1999). <lb/>[8] Y. Taguchi et al., Science 291, 2573 (2001). <lb/>[9] S. Kondo et al., Phys. Rev. Lett. 78, 3729 (1997). C. <lb/>Urano et al., Phys. Rev. Lett. 85, 1052 (2000). <lb/>[10] D. Mandrus et al., Phys. Rev. B 63, 195104 (2001). <lb/>[11] Hironri Sakai et al., J. Phys. Condens. Matter 13, L785 <lb/>(2001). <lb/>[12] M. Hanawa et al., Phys. Rev. Lett. 87, 187001 (2001). <lb/>[13] P. Donohue et al., Inorg. Chem. 4, 1152 (1965). <lb/>[14] R. Jin et al., Phys. Rev. B 64, 180503(R) (2001). <lb/>[15] R. Jin et al., cond-mat/0108402 (2001). <lb/>[16] M. Hamawa et al., cond-mat/0109050 (2001). <lb/>[17] J.E. Sonier, J.H. Brewer and R.F. Kiefl, Rev. Mod. Phys. <lb/>72, 769 (2000) and references therein. <lb/>[18] A. Yaouanc et al., Phys. Rev. B 55, 11107 (1997). <lb/>[19] J. He et al., to be submitted. <lb/>[20] R. F. Kiefl et al., Phys. Rev. B 32, 530 (1985). <lb/>[21] Y.J. Uemura et al., Phys. Rev. Lett. 66, 2665 (1991). <lb/>[22] C.M. Aegerter et al., J. Phys. Condens. Matt. 10, 7445 <lb/>(1998). <lb/>[23] Z. Hiroi and M. Hanawa, cond-mat/0111126 (2001). </listBibl>
26
+
27
+
28
+ </text>
29
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L089157004-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>arXiv:cond-mat/0202251v2 [cond-mat.str-el] 7 Oct 2002 <lb/>Superconductivity and Quantum Criticality in CeCoIn <lb/>V. A. Sidorov, * M. Nicklas, P. G. Pagliuso, J. L. Sarrao, Y. Bang, A. V. Balatsky, and J. D. Thompson <lb/>Los Alamos National Laboratory, Los Alamos, NM 87545 <lb/>(Dated: October 29, 2018) <lb/>Electrical resistivity measurements on a single crystal of the heavy-fermion superconductor <lb/>CeCoIn5 at pressures to 4.2 GPa reveal a strong crossover in transport properties near P * ≈ 1.6 GPa, <lb/>where Tc is a maximum. The temperature-pressure phase diagram constructed from these data pro-<lb/>vides a natural connection to cuprate physics, including the possible existence of a pseudogap. <lb/>PACS numbers: 74.70.Tx, 74.62.Fj, 75.30.Mb, 75.40.-s <lb/></front>
8
+
9
+ <body>The relationship between unconventional superconduc-<lb/>tivity and quantum criticality is emerging as an im-<lb/>portant issue in strongly correlated materials, includ-<lb/>ing cuprates, organics and heavy-fermion intermetallics <lb/>[1, 2]. In each of these, either Nature or an externally <lb/>imposed parameter, such as pressure or chemical substi-<lb/>tutions, tunes some magnetic order to a T = 0 transi-<lb/>tion where quantum fluctuations introduce new excita-<lb/>tions that control thermodynamic and transport proper-<lb/>ties over a broad range of temperature-parameter phase <lb/>space. A frequent, but controversial [3], assumption is <lb/>that these same excitations may mediate Cooper pair-<lb/>ing in the anisotropic superconductivity that appears <lb/>in the vicinity of a quantum-critical point (QCP). In-<lb/>deed, the effective potential generated by proximity to a <lb/>quantum-critical spin-density wave [4] also can lead [5] <lb/>to anisotropic superconductivity. <lb/>Though the issue of quantum criticality and super-<lb/>conductivity came to prominence in the context of the <lb/>cuprates (see, eg [6]), there have been several recent ex-<lb/>amples of this interplay in heavy-fermion compounds [7]. <lb/>Like the cuprates, unconventional superconductivity in <lb/>these Ce-based heavy-fermion materials develops out of <lb/>a distinctly non-Fermi-liquid normal state that evolves <lb/>in proximity to a continuous T = 0 antiferromagnetic <lb/>transition; but unlike the cuprates, this state can be ac-<lb/>cessed cleanly by applied hydrostatic pressure and with-<lb/>out introducing extrinsic disorder associated with chemi-<lb/>cal substitutions. In addition to chemical inhomogeneity <lb/>in the cuprates, a further &apos;complication&apos; is the existence <lb/>of a pseudogap state above T c for dopings less than opti-<lb/>mal [8]. Whether evolution of the pseudogap with dop-<lb/>ing produces another QCP near the optimal T c in the <lb/>cuprates is an open question as is the origin of the pseu-<lb/>dogap [9]. Thus far, a pseudogap has not been detected in <lb/>canonical quantum-critical heavy-fermion systems, such <lb/>as CeIn 3 or CePd 2 Si 2 under pressure, but this may not <lb/>be too surprising. In these cases, superconductivity ap-<lb/>pears only at very low temperatures, less than 0.4 K, and <lb/>over a narrow window of pressures (≤ 0.8 GPa) centered <lb/>around the pressure where T N extrapolates to zero. In <lb/>analogy to the cuprates, we would expect a pseudogap in <lb/>these Ce compounds to exist only over a narrow temper-<lb/>ature window, 0.1 − 0.2 K above T c , and in a fraction of <lb/>the narrow pressure window, which certainly would make <lb/>detection of a pseudogap difficult. <lb/>CeCoIn 5 offers the possibility of making a connec-<lb/>tion between other Ce-based heavy-fermion materials <lb/>and the cuprates. Bulk, unconventional superconduc-<lb/>tivity is present in CeCoIn 5 at atmospheric pressure <lb/>[10, 11, 12, 13] and develops out of a heavy-fermion nor-<lb/>mal state in which the resistivity is approximately linear <lb/>in temperature. Application of a magnetic field suffi-<lb/>cient to quench superconductivity reveals a specific heat <lb/>diverging as −T ln T and associated low-temperature en-<lb/>tropy consistent with its huge zero-field specific heat <lb/>jump at T c (∆C/γT c = 4.5). Together, these proper-<lb/>ties suggest that CeCoIn 5 may be near an antiferromag-<lb/>netic quantum-critical point at P = 0. Its isostructural, <lb/>antiferromagnetic relative CeRhIn becomes a pressure-<lb/>induced unconventional superconductor near 1.6 GPa, <lb/>with a T c = 2.1 K close to that of CeCoIn 5 at P = 0 <lb/>[14, 15]. This and other similarities between CeCoIn at <lb/>P = 0 and CeRhIn 5 at P = 1.6 GPa reinforce specula-<lb/>tion [16, 17] that the nearby antiferromagnetic QCP in <lb/>CeCoIn may be at an inaccessible slightly negative pres-<lb/>sure. The 1.7% smaller cell volume of CeCoIn compared <lb/>to that of CeRhIn is consistent with this view. We have <lb/>studied the effect of pressure on the electrical resistivity <lb/>and superconductivity of CeCoIn 5 and find striking cor-<lb/>relations between them that are reminiscent of cuprate <lb/>behaviors. <lb/>Four-probe AC resistivity measurements, with current <lb/>flowing in the tetragonal basal plane, were made on a <lb/>single crystal of CeCoIn grown from excess In flux. Hy-<lb/>drostatic pressures to 5 GPa were generated in a toroidal <lb/>anvil cell [18] in which a boron-epoxy gasket surrounds <lb/>a teflon capsule filled with a glycerol-water mixture (3:2 <lb/>volume ratio) that served as the pressure transmitting <lb/>fluid. Pressure inside the capsule was determined at room <lb/>temperature and at low temperatures from the pressure-<lb/>dependent electrical resistivity and T c of Pb, respectively <lb/>[19]. The width of superconducting transition of Pb did <lb/>not exceed 15 mK, indicating good hydrostatic conditions <lb/>and providing an estimate of the pressure-measurement <lb/>uncertainty, ±0.04 GPa. <lb/>The response of ρ(T ) to pressure over a broad temper-<lb/>ature scale [16, 17] is typical of many Ce-based heavy-</body>
10
+
11
+ <page><lb/>2</page>
12
+
13
+ <body><lb/>3 <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>1.69 <lb/>Sidorov et al. <lb/>Figure 1 <lb/>CeCoIn <lb/>5 <lb/>4.2 <lb/>3.43 <lb/>3.15 <lb/>2.87 <lb/>2.16 <lb/>1.35 GPa <lb/>ρ (µΩ cm) <lb/>T(K) <lb/>1 <lb/>10 <lb/>0 <lb/>10 <lb/>3.43 <lb/>4.2 <lb/>3.15 <lb/>2.87 <lb/>2.16 <lb/>1.69 <lb/>ρ / <lb/>ρ(T <lb/>+ <lb/>c <lb/>) <lb/>T(K) <lb/>0 <lb/>1.35 <lb/>FIG. 1: Effect of pressure on the low-temperature resistivity <lb/>and superconducting Tc of CeCoIn5. The inset shows the <lb/>resistivity ρ normalized by ρ(T + <lb/>c ) up to 10 K. <lb/>fermion compounds and reflects a systematic increase in <lb/>the energy scale of spin fluctuations. Most interesting is <lb/>the low-temperature response (Fig. 1) where T c and the <lb/>resistivity just above T c (ρ(T + <lb/>c )) decrease rapidly with <lb/>pressures P &gt; 1.35 GPa. We have analyzed the low tem-<lb/>perature resistivity, plotted over a broader temperature <lb/>scale in the inset of Fig. 1, by either fitting to a form <lb/>ρ(T ) = ρ 0 + AT n , with ρ 0 , A and n fitting parameters, <lb/>or from plots of (ρ(T ) − ρ 0 )/T n . A drawback to the for-<lb/>mer approach is that parameters can be sensitive to the <lb/>temperature range chosen for fitting. The more sensitive <lb/>and revealing approach is the latter in which ρ 0 and n are <lb/>adjusted to give the best horizontal line. This procedure <lb/>requires one less parameter, A, which is provided directly <lb/>by the magnitude of (ρ(T ) − ρ 0 )/T n . Representative ex-<lb/>amples of this approach are plotted in Fig. 2. Parameters <lb/>obtained from these plots are well defined and agree, typ-<lb/>ically to 10% or better, with parameter values extracted <lb/>from straightforward fits. As indicated by arrows on the <lb/>lowest pressure curve in Fig. 2, (ρ(T ) − ρ 0 )/T n deviates <lb/>from the horizontal trend at a temperature T pg well above <lb/>the onset of superconductivity. This departure implies <lb/>either a decrease in the scattering rate or increase in car-<lb/>rier density. Resistivity measurements at P = 0 and in a <lb/>magnetic field (6 T) greater than H c2 (0) find [20] a sim-<lb/>ilar departure of (ρ(T ) − ρ 0 )/T n from a T -independent <lb/>curve at T pg =3 K. Thus, T pg does not originate from su-<lb/>perconducting fluctuations or from trace amounts of free <lb/>indium. We also note that this behavior was present in <lb/>previous measurements on other single crystals CeCoIn <lb/>but was missed in inspections of ρ versus T curves [16]. <lb/>We will return to a discussion of T pg later. <lb/>Values of ρ 0 , n and A obtained from these plots are <lb/>summarized in Figs. 3a and b. There is an unmis-<lb/>takable crossover in the pressure dependence of ρ and <lb/>12345 <lb/>0.02 <lb/>0.04 <lb/>0.06 <lb/>0.08 <lb/>1.00 <lb/>1.05 <lb/>T C <lb/>T C <lb/>T pg <lb/>1.35 GPa <lb/>CeCoIn 5 <lb/>T FL <lb/>3.43 <lb/>2.87 <lb/>( <lb/>ρ -<lb/>ρ <lb/>0 ) / T <lb/>n <lb/>( <lb/>µΩ cm K <lb/>-n <lb/>) <lb/>T(K) <lb/>FIG. 2: Representative plots of (ρ(T ) − ρ0)/T n vs T . At <lb/>low pressures, (ρ(T ) − ρ0)/T n is constant from Tpg ≈ (1.15 ± <lb/>0.05)Tc to about 10 K. At the higher pressures, there is a <lb/>well-defined range below TF L where (ρ − ρ0) ∝ T 2 is Fermi-<lb/>liquid like. The resistive onset of superconductivity, defined <lb/>by the intersection of linear extrapolations of ρ(T ) from above <lb/>and below Tc, is denoted by dashed vertical arrows. At the <lb/>highest pressures, the gradual rounding just above Tc is due <lb/>to a broadened transition (see Fig. 3) and not to Tpg <lb/>n at P * ≈ 1.6 GPa. Below P * , n = 1.0 ± 0.1 for <lb/>T pg ≤ T ≤ 10 K. This value of n is expected for a 2-<lb/>dimensional, antiferromagnetic quantum-critical system <lb/>[21] and is not too surprising given the layered crys-<lb/>tal structure [22] and anisotropy in electronic states of <lb/>CeCoIn [10, 23]. At pressures greater than P * , n rapidly <lb/>approaches the Fermi-liquid value of 2.0, which holds <lb/>from just above T c to T F L (see Fig. 2). Though not <lb/>shown in either Fig. 2 or 3a, at these higher pressures n <lb/>assumes a value of 1.5±0.1, characteristic of a 3-D antifer-<lb/>romagnetic QCP, in a temperature interval from slightly <lb/>greater than T F L to as high as ∼ 60 K. For P ≥ P * , ρ 0 <lb/>decreases reversibly by an order of magnitude to a very <lb/>small value of about 0.2 µΩcm. Clearly, pressure does not <lb/>remove impurities from the sample; the large decrease in <lb/>ρ 0 must be due to a pressure-induced change in inelas-<lb/>tic scattering processes. Theories of electronic transport <lb/>in quantum-critical systems show [24, 25] that impurity <lb/>scattering can be strongly enhanced by quantum-critical <lb/>fluctuations. This sensitivity provides a natural expla-<lb/>nation for a decrease in scattering with increasing pres-<lb/>sure, if there is a crossover near P * from quantum-critical <lb/>(P &lt; P * ) to a Fermi-liquid-like state for (P &gt; P * ). Such <lb/>a crossover is suggested by the pressure variation of n. <lb/>Experimental values of the T n -coefficient of resistiv-<lb/>ity, plotted in Fig. 3b, also decrease strongly with in-<lb/>creasing pressure for P &lt; <lb/>∼ P * . For comparison, we show <lb/></body>
14
+
15
+ <page>3 <lb/></page>
16
+
17
+ <body>0 <lb/>1 <lb/>3 <lb/>1.0 <lb/>1.5 <lb/>2.0 <lb/>0 <lb/>0.4 <lb/>0.8 <lb/>1.2 <lb/>1.6 <lb/>0 <lb/>100 <lb/>200 <lb/>300 <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>0.1 <lb/>CeCoIn 5 <lb/>ρ = ρ 0 + AT <lb/>n <lb/>a <lb/>ρ <lb/>0 ( <lb/>µΩ cm ) <lb/>n <lb/>b <lb/>A ( <lb/>µΩ cm K <lb/>-n <lb/>) <lb/>γ (mJ / mol Ce K <lb/>2 <lb/>) <lb/>c <lb/>∆T <lb/>C / T <lb/>C <lb/>P(GPa) <lb/>FIG. 3: (a) Values of ρ0 and n obtained from plots as shown in <lb/>Fig. 2. (b) Temperature coefficient of resistivity, A, associated <lb/>with values of n given in the top panel, and specific heat <lb/>coefficient γ measured directly (P &lt; 1.5 GPa) and, for P &gt; 2 <lb/>GPa , inferred from the T 2 coefficient of ρ(T ). See text for <lb/>details. (c) Superconducting transition width normalized by <lb/>Tc. ∆Tc and Tc are the half-width of the resistive transition <lb/>width and resistive midpoint, respectively. Solid lines in all <lb/>cases are guides to the eye. <lb/>the variation of the electronic specific heat divided by <lb/>temperature (C/T ≡ γ) at 3 K for pressures less than <lb/>1.5 GPa [26] and, at higher pressures, we plot values of γ <lb/>inferred from the empirical relationship A = 1 × 10 −5 γ 2 , <lb/>where A is the T 2 -coefficient of resistivity, followed by <lb/>several heavy-fermion compounds [27]. Though a T <lb/>coefficient proportional to γ 2 is expected for a Landau <lb/>Fermi-liquid [6], it is not obvious a priori why A(P ) ap-<lb/>proximately tracks the directly measured γ at low pres-<lb/>sures, but this seems to be the case. In any event, both <lb/>A(P ) and γ(P ) exhibit qualitatively different magnitudes <lb/>and functional dependencies above and below P * . The <lb/>dramatic differences in γ(P ) at low and high pressures <lb/>suggest that either the empirical relationship is invalid <lb/>and underestimates γ in the high-pressure regime by at <lb/>least two orders of magnitude or there is a very rapid <lb/>crossover in the density of low-energy excitations in the <lb/>vicinity of P * . In the Fermi-liquid-like state above P * , <lb/>we would expect the temperature scale T F L (see Fig. 2) <lb/>P* <lb/>P <lb/>0 <lb/>0 <lb/>c <lb/>b <lb/>PG <lb/>FL <lb/>NFL <lb/>AFM <lb/>SC <lb/>T <lb/>P <lb/>0 <lb/>1234 <lb/>5 <lb/>1 <lb/>2 <lb/>3 <lb/>T FL <lb/>T pg <lb/>a <lb/>CeCoIn 5 <lb/>SC <lb/>n = 1.5 ± 0.1 <lb/>n ≈ 1 <lb/>n = 2 <lb/>T(K) <lb/>P(GPa) <lb/>FIG. 4: (a) Temperature-pressure phase diagram for CeCoIn5 <lb/>constructed from data shown in Figs. 2 and 3. (b) Schematic <lb/>T −P phase diagram. AFM: Néel state; PG: pseudogap state; <lb/>SC: unconventional superconducting state; FL: Fermi liquid; <lb/>NFL: non-Fermi-liquid. See text for details. <lb/>to be proportional to A −0.5 instead of the approximate <lb/>T F L ∝ <lb/>√ <lb/>A found experimentally. We do not understand <lb/>this discrepancy. Direct measurements of C/T at pres-<lb/>sures above P * would be valuable. <lb/>Finally, we note that the resistive transition to the su-<lb/>perconducting state is equally sensitive to the crossover <lb/>at P * . As shown in Fig. 3c, the relative transition width <lb/>∆T c /T c passes through a pronounced minimum near P * . <lb/>To some extent this behavior is accentuated by the pres-<lb/>sure variation of T c , but ∆T c itself is a minimum near P * . <lb/>It is as if the optimally homogeneous state is singular in <lb/>the vicinity of P * . <lb/>From the data presented above, we construct a <lb/>temperature-pressure phase diagram for CeCoIn given <lb/>in Fig. 4a. This phase diagram is more like that of <lb/>the cuprates around optimal doping than the canonical <lb/>behavior of heavy-fermion systems exemplified, for ex-<lb/>ample, by CeIn or CePd 2 Si 2 [7]. There is no obvious <lb/>long-range ordered state in CeCoIn 5 at atmospheric or <lb/>higher pressures. Rather, in view of its relationship to <lb/>CeRhIn 5 , a reasonable speculation is that the antiferro-<lb/>magnetic QCP of CeCoIn 5 is at a slightly negative pres-<lb/>sure. The resistivity exponent n and specific heat (ref. <lb/></body>
18
+
19
+ <page>4 <lb/></page>
20
+
21
+ <body>[26]) are those of a non-Fermi-liquid up to P * where T c <lb/>is a maximum. Beyond P * , T c drops rapidly and param-<lb/>eters characterizing electronic transport are qualitatively <lb/>different, Fermi-liquid-like at temperatures just above T c <lb/>and those of a 3-D quantum-critical antiferromagnetic <lb/>state at higher temperatures. <lb/>The generic T −P phase diagram in Fig. 4b provides <lb/>a broader perspective of our experimental observations <lb/>and their relationship to CeRhIn 5 . In the spirit of this <lb/>diagram, the non-Fermi-liquid transport and thermody-<lb/>namic properties of CeCoIn 5 below P * are controlled <lb/>by an antiferromagnetic QCP at the inaccessible nega-<lb/>tive pressure denoted by P c in Fig. 4. We have chosen <lb/>P c to correspond approximately to the critical pressure <lb/>at which superconductivity appears in CeRhIn . Spin-<lb/>lattice relaxation measurements on CeRhIn identify [28] <lb/>a pseudogap whose signature first appears at pressures <lb/>slightly less than P c and at a temperature of ∼5 K and <lb/>then decreases toward T c with increasing pressure until <lb/>its signature is lost at P &gt; <lb/>∼ P c . If we associate the tem-<lb/>perature T pg (P ) in Fig. 4a with a resistive signature for a <lb/>pseudogap and extrapolate T pg (P ) in CeCoIn to slightly <lb/>negative pressures, there is a natural connection between <lb/>our observations and properties of CeRhIn . Neither <lb/>the 1/T 1 T measurements on CeRhIn 5 nor our transport <lb/>measurements on CeCoIn 5 permits definitive statements <lb/>about the origin or nature of the pseudogap [29]. We <lb/>note, however, that 4-fold anisotropy of the thermal con-<lb/>ductivity in the basal plane of CeCoIn persists [12] to <lb/>3.2 K ≈ T pg at atmospheric pressure. If this in-plane <lb/>modulation of the normal-state thermal conductivity is <lb/>due to the presence of a pseudogap, it implies d-wave <lb/>symmetry. <lb/>The schematic phase diagram, Fig. 4b, reflects the rela-<lb/>tionship among various phases of CeCoIn and CeRhIn 5 <lb/>and is similar to the T -doping phase diagram of the <lb/>cuprates. This diagram and experimental data from <lb/>which it is inferred indicate that the physics of heavy-<lb/>fermion systems may be more closely related to that of <lb/>the cuprates than previously appreciated. Aside from the <lb/>structural layering in CeCoIn 5 and CeRhIn 5 and their <lb/>relatively high T c &apos;s, they are similar in many respects <lb/>to other Ce-based heavy-fermion materials in which su-<lb/>perconductivity develops near a QCP. Further study <lb/>of CeCoIn , in parallel with the cuprates and other <lb/>pressure-induced heavy-fermion superconductors, holds <lb/>promise for bridging our understanding of the inter-<lb/>relationship between unconventional superconductivity <lb/>and quantum criticality across these interesting classes <lb/>of correlated matter. <lb/></body>
22
+
23
+ <div type="acknowledgement">We thank R. Movshovich for helpful discussions. VS <lb/>acknowledges S. M. Stishov for his support during the <lb/>initial stage of this research. Work at Los Alamos was <lb/>performed under the auspices of the US Department of <lb/>Energy. <lb/></div>
24
+
25
+ <front>* Permanent address: Institute for High Pressure Physics, <lb/>Russian Academy of Sciences, Troitsk, Russia. <lb/></front>
26
+
27
+ <listBibl>[1] J. Orenstein and A. J. Millis, Science 288, 468 (2000); S. <lb/>Sachdev, ibid. 288, 475 (2000). <lb/>[2] A. Y. Chubukov, D. Pines, and J. Schmalian, cond-<lb/>mat/0201140. <lb/>[3] P. W. Anderson, Physica B 318, 28 (2002). <lb/>[4] P. Coleman and C. Pepin, Physica B 312-313, 383 <lb/>(2002). <lb/>[5] P. Monthoux, A. V. Balatsky, and D. Pines, Phys. Rev. <lb/>Lett. 67, 3448 (1991). <lb/>[6] C. M. Varma, Z. Nussinov, and W. van Saarloos, Phys. <lb/>Rep. 361, 267 (2002). <lb/>[7] N. D. Mathur et al., Nature 394, 39 (1998). <lb/>[8] T. Timusk and B. Statt, Rep. Prog. Phys. 62, 61 (1999). <lb/>[9] See, for example, C. M. Varma, Phys. Rev. Lett. 83, 3538 <lb/>(1999); A. Sokol and D. Pines, Phys. Rev. Lett. 71, 2813 <lb/>(1993); S. Chakravarty et al., Phys. Rev. B 63, 094503 <lb/>(2001); Ar. Abanov, A.V. Chubukov, and J. Schmalian, <lb/>Europhys. Lett. 55, 369 (2001); and refernces therein. <lb/>[10] C. Petrovic et al., J. Phys.: Condens. Matter 13, L337 <lb/>(2001). <lb/>[11] R. Movshovich et al., Phys. Rev. Lett. 86, 5152 (2001). <lb/>[12] K. Izawa et al.,Phys. Rev. Lett. 87, 057002 (2001). <lb/>[13] Y. Kohori et al., Phys. Rev. B 64, 134526 (2001). <lb/>[14] H. Hegger et al., Phys. Rev. Lett. 84, 4986 (2000). <lb/>[15] R. A. Fisher et al., Phys. Rev. B 65, 224509 (2002). <lb/>[16] M. Nicklas et al., J. Phys.: Condens. Matter 13, L905 <lb/>(2001). <lb/>[17] H. Shishido et al., J. Phys. Soc. Jpn. 71, 162 (2002). <lb/>[18] L. G. Khvostantsev, V. A. Sidorov, and O. B. Tsiok, <lb/>in: Properties of Earth and Planetary Materials at High <lb/>Pressures and Temperatures, Ed. by M.H. Manghnani <lb/>and T. Yagi, Geophysical Monograph 101, American <lb/>Geophysical Union, 1998, p.89. <lb/>[19] A. Eiling and J. S. Schilling, J. Phys. F: Metal Phys. 11, <lb/>623 (1981). <lb/>[20] R. Movshovich, private communication. <lb/>[21] S. Sachdev, Quantum Phase Transitions (Cambridge <lb/>University Press, Cambridge, 1999). <lb/>[22] E. G. Moshopoulou et al., J. Solid State Chem. 158, 25 <lb/>(2001). <lb/>[23] R. Settai et al., J. Phys.: Condens. Matter 13, L627 <lb/>(2001). <lb/>[24] A. Rosch, Phys. Rev. Lett. 82, 4280 (1999). <lb/>[25] K. Miyake and O. Narikiyo, J. Phys. Soc. Jpn. 71, 867 <lb/>(2002). <lb/>[26] G. Sparn, et al. Physica B 312-313, 138 (2002); E. <lb/>Lengyel et al. High Pressure Res. 22, (2002). In these <lb/>references, for H ≥ Hc2(0) C/T = − ln T persists to <lb/>P = 1.48 GPa, but the rate of divergence decreases with <lb/>increasing P . <lb/>[27] K. Kadowaki and S. B. Woods, Solid State Comm. 58, <lb/>507 (1986). <lb/>[28] S. Kawasaki et al., Phys. Rev. B 65, 020504 (2001). <lb/>[29] There is no obvious evidence for a pseudogap at P = 0 in <lb/>specific heat or magnetic susceptibility data for CeCoIn5, <lb/>possibly because the signature is expected to be weak and <lb/>the temperature range between Tpg and Tc is limited to <lb/>no more than 1.1 K. </listBibl>
28
+
29
+
30
+ </text>
31
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L090137001-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>arXiv:cond-mat/0303278v1 [cond-mat.supr-con] 14 Mar 2003 <lb/>APS/123-QED <lb/>Superconductivity of the Sr 2 Ca 12 Cu 24 O 41 spin ladder system : Are the <lb/>superconducting pairing and the spin-gap formation of the same origin? <lb/>Naoki Fujiwara, * Nobuo Môri, † and Yoshiya Uwatoko <lb/>Institute for Solid State Physics, University of Tokyo, <lb/>5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8581,Japan <lb/>Takehiko Matsumoto <lb/>National Institute for Materials Science, Tsukuba 305-0047, Japan <lb/>Naoki Motoyama and Shinichi Uchida <lb/>Department of Superconductivity, University of Tokyo, <lb/>7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan <lb/>(Dated: July 17 2002) <lb/>Pressure-induced superconductivity in a spin-ladder cuprate Sr2Ca12Cu24O41 has not been studied <lb/>on a microscopic level so far although the superconductivity was already discovered in 1996. We have <lb/>improved high-pressure technique with using a large high-quality crystal, and succeeded in studying <lb/>the superconductivity using Cu nuclear magnetic resonance (NMR). We found that anomalous <lb/>metallic state reflecting the spin-ladder structure is realized and the superconductivity possesses <lb/>a s-wavelike character in the meaning that a finite gap exists in the quasi-particle excitation: At <lb/>pressure of 3.5GPa we observed two excitation modes in the normal state from the relaxation rate <lb/>T −1 <lb/>1 . One gives rise to an activation-type component in T −1 <lb/>1 , and the other T -linear component <lb/>linking directly with the superconductivity. This gapless mode likely arises from free motion of <lb/>holon-spinon bound states appearing by hole doping, and the pairing of them likely causes the <lb/>superconductivity. <lb/>PACS numbers: 74.25. Ha, 74.72. Jt,74.70. -b <lb/></front>
8
+
9
+ <body>Sr 14−x Ca x Cu 24 O 41 (x=11.5-13.5) is a spin-ladder sys-<lb/>tem in which superconducting state is realized by apply-<lb/>ing pressure [1]. The system possesses a structural unit <lb/>of the Cu 2 O 3 two-leg ladder [2], and holes are transferred <lb/>from the CuO chain unit by substituting Ca for Sr. The <lb/>dopant hole density can be controlled over a wide range <lb/>from 0.07 (x=0) to 0.24 (x=14) per ladder Cu [3]. It <lb/>is known from theoretical investigations that the ground <lb/>state of undoped system is a quantum spin liquid and <lb/>this state persists even in highly doped region [4-6]. The <lb/>spin gap has been observed by a number of works in <lb/>the present ladder system. The decrease of the spin gap <lb/>for the Ca substitution was observed in the NMR mea-<lb/>surements [7-9], although no change was observed in the <lb/>neutron scattering [10]. The relationship and interplay <lb/>between the spin gap and the superconductivity in the <lb/>hole-doped ladder system is of central concern [4-6, 11] <lb/>as is the case with high-T c cuprates in underdoped region <lb/>[12]. <lb/>The superconducting state is realized when high pres-<lb/>sure of 3-8GPa is applied for highly doped compounds <lb/>(x≥10) [3]. Pressure plays a role of stabilizing the metal-<lb/>lic state and suppressing anisotropy within the ladder <lb/>plane. In fact, the temperature (T ) dependence of the <lb/>resistivity in the direction perpendicular to the plane ρ c <lb/></body>
10
+
11
+ <front> * Email: [email protected] <lb/> † Department of Physics, Faculty of Science, Saitama University, <lb/>255 Simookubo, Saitama 338-8581, Japan <lb/></front>
12
+
13
+ <body>which is insulating at low pressures turns to be meallic <lb/>as pressure increases [13, 14]. The ratio of the resistiv-<lb/>ities along the rungs and the legs, ρ a /ρ c goes up to 80 <lb/>at ambient pressure, but is reduced below 30 at 3.5GPa <lb/>[13]. <lb/>The superconductivity has been studied on a macro-<lb/>scopic level by the resistivity and AC susceptibility mea-<lb/>surements because these measurements are performable <lb/>with a small amount of sample and cubic-anvil pressure <lb/>cell is available to reach high pressure [1, 13-15]. A <lb/>clamp-type pressure cell is much more convenient and <lb/>available to various methods in which a larger sample <lb/>volume is needed. However, usual clamp-type pressure <lb/>cell is made of CuBe alloy and the maximum pressure is <lb/>3GPa at most. Hence, microscopic study of the super-<lb/>conductivity has not been performed so far although the <lb/>superconductivity was discovered in 1996. Mayaffre et <lb/>al. made NMR measurements under high pressure up to <lb/>3GPa for the field (H) perpendicular to the plane, crystal <lb/>b axis and suggested that the spin gap is suppressed by <lb/>applying pressure [16, 17]. However, the measurements <lb/>were in fact done not in the superconducting state but in <lb/>the normal state above T c . <lb/>In the present work, we have used a clamp-type pres-<lb/>sure cell made of Ni alloy and performed NMR measure-<lb/>ment at pressures more than 3GPa by applying the field <lb/>parallel to the leg direction, crystal c axis. This enabled <lb/>us to study the superconducting state on a microscopic <lb/>level for the first time and to investigate pairing symme-<lb/>try as well as the relation between the spin gap and the <lb/></body>
14
+
15
+ <page>2 <lb/></page>
16
+
17
+ <body>61.5 <lb/>61.4 <lb/>61.3 <lb/>61.2 <lb/>61.1 <lb/>61.0 <lb/>60.9 <lb/>Frequency (MHz) <lb/>7.0 <lb/>6.0 <lb/>5.0 <lb/>4.0 <lb/>3.0 <lb/>2.0 <lb/>1.0 <lb/>T (K) <lb/>3.5Gpa <lb/>H // c axis <lb/>0T <lb/>5T <lb/>6T <lb/>7T <lb/>12 <lb/>8 <lb/>4 <lb/>0 <lb/>H(T) <lb/>4 <lb/>2 <lb/>0 <lb/>T(K) <lb/>FIG. 1: Resonance frequency of a NMR probe attached to a <lb/>pressure cell. The frequency changes at the superconducting <lb/>state. The onset corresponds to the upper critical field. Inset <lb/>shows Hc − Tc curve obtained from the onset of the frequency <lb/>at each field. Solid line is a guide for the eyes. <lb/>superconductivity. <lb/>Single crystal of Sr 2 Ca 12 Cu 24 O 41 with a volume of <lb/>4x2x1mm was prepared for the measurements. The <lb/>clamp-type pressure cell with an effective sample space <lb/>of φ4 x 20mm was used for the measurement. The ap-<lb/>pearance of the superconductivity was confirmed by mea-<lb/>suring resonance frequency of a NMR probe attached to <lb/>the pressure cell. The resonance frequency is roughly <lb/>given as f ∝ 1/ <lb/>√ <lb/>LC where L and C represent induc-<lb/>tance and variable capacitance of the NMR probe, re-<lb/>spectively. The sample is contained in a coil and the <lb/>onset of superconductivity is detected by the change of <lb/>the resonant frequency, i.e. the change of L value. This <lb/>method corresponds to AC susceptibility measurement. <lb/>The T dependence of the frequency at several fields is <lb/>shown in Fig. 1. If we probe the temperature at which <lb/>the resonance frequency starts to change against H, this <lb/>gives H c2 vs. T c characteristics as is shown in the inset. <lb/>For high-T c cuprates, the temperature has been identi-<lb/>fied as irreversibility temperature T irr [18]. T irr gives <lb/>a borderline between creeping and freezing of flux lines <lb/>(FL) which are formed through the planes for the H per-<lb/>pendicular to the plane. In the present case flux lines are <lb/>self-trapped between the planes since the field is applied <lb/>parallel to the plane, and thus such a creep is hardly <lb/>expected. <lb/>63 Cu-NMR signals for the ladder site can be separated <lb/>from those for the chain site since nuclei of these sites <lb/>possess different quadrupole coupling [8, 19]. We have <lb/>reported NMR spectra under pressure measured by us-<lb/>ing the present crystal in Ref. 20. The NMR spectra <lb/>at 3.5GPa were almost the same with those at ambient <lb/>pressure. One signal corresponding to the central transi-<lb/>10 <lb/>0 <lb/>10 <lb/>1 <lb/>10 <lb/>2 <lb/>10 <lb/>3 <lb/>10 <lb/>4 <lb/>1/T <lb/>1 (sec <lb/>-1 <lb/>) <lb/>2 <lb/>4 6 8 <lb/>10 <lb/>2 <lb/>4 6 8 <lb/>100 <lb/>2 <lb/>T (K) <lb/>T <lb/>6.2T Tc=2.8K <lb/>5.0T Tc=3.4K <lb/>3.5GPa <lb/>Sr 2 Ca 12 Cu 24 O 41 <lb/>H // c axis <lb/>10 <lb/>4 <lb/>5 <lb/>4 <lb/>3 <lb/>1 <lb/>FIG. 2: Nuclear ratice relaxation rate 1/T1 for 63 Cu nuclei. <lb/>A solid line shows Korringa relation expressed by Eq. (2). <lb/>1/T1 at low temperatures around Tc is expanded in the inset. <lb/>tion (I= -1/2⇔ 1/2) of 63 Cu nuclei was observed for the <lb/>ladder site and the relaxation rate (T −1 <lb/>1 ) was measured <lb/>at 70.0MHz which corresponds to 6.2T. T c at this field <lb/>is about 2.8K as is seen from the inset of Fig. 1. The T <lb/>dependence of T −1 <lb/>1 <lb/>is shown in Fig. 2. The spin gap is <lb/>observed even under high pressure as an activated T de-<lb/>pendence of T −1 <lb/>1 <lb/>at temperatures higher than 30K, i.e., <lb/>T −1 <lb/>1 <lb/>∝ exp(−∆ spin /T ). <lb/>(1) <lb/>The value of ∆ spin is estimated to be 173K. It should be <lb/>noted that the spin gap is seen in the state in which the <lb/>charge transport is metallic [13]. <lb/>By contrast, T −1 <lb/>1 <lb/>below 30K is dominated by a term <lb/>showing T -linear dependence followed by a peak devel-<lb/>oped below T c . The onset and the position of the peak <lb/>shift to higher temperature with decreasing the magnetic <lb/>field to 5.0T (open squares in Fig. 2). If the T -linear <lb/>component originates from extrinsic source, it should be <lb/>persistent in the superconducting state. However, T −1 <lb/>1 <lb/>measured at low temperatures obviously goes below the <lb/>T -linear line shown in the figure. Then, such a case is <lb/>excluded. As for the peak the possibility of vortex mo-<lb/>tion might be pointed out. The vortex motion has been <lb/>observed from T −1 <lb/>1 <lb/>of ligand sites in high-T c cuprates <lb/>such as YBa 2 Cu 4 O 8 [21] or HgBa 2 CuO 4+δ [22] when <lb/>the H is applied perpendicular to the plane. The FL <lb/>is self-trapped between the planes for the H parallel to <lb/>the plane, and thus the effect is hardly expected in the <lb/>present case. In fact, the effect disappears when the H is <lb/></body>
18
+
19
+ <page>3 <lb/></page>
20
+
21
+ <body>applied parallel to the planes in HBCO system [22]. Fur-<lb/>thermore, T −1 <lb/>1 /γ 2 <lb/>N (γ N : nuclear gyromagnetic ratio ) at <lb/>the peak is 80 or 25 times larger than those for YBCO <lb/>or HBCO, respectively. The value is too large to explain <lb/>the peak due to vortex motion. Hence, we conclude that <lb/>the peak of T −1 <lb/>1 <lb/>and the T -linear component have the <lb/>same origin in the electric state and/or spin fluctuations. <lb/>The peak can be assigned to a superconducting coherence <lb/>peak and the T -linear dependence of T −1 <lb/>1 <lb/>to Korringa-<lb/>type behavior, <lb/>T −1 <lb/>1 <lb/>∝ bT <lb/>(2) <lb/>where the value of b is about 6.1 (sec −1 K −1 ). The ob-<lb/>servation of the clear peak implies that a finite gap ex-<lb/>ists in the quasi-particle excitation at all wave vectors. <lb/>In the meaning that there exist no nodes in the pairing <lb/>symmetry, the superconductivity possesses a s-wavelike <lb/>character. <lb/>The spin gap is also observed in 63 Cu-NMR shift (K) <lb/>at relevant temperatures. The shift shows H −2 -linear <lb/>dependence because large quadrupole effect acts on 63 Cu <lb/>nuclei. The values free from the quadrupole effect are <lb/>obtained by plotting K vs. H −2 and extrapolating to <lb/>zero field [23]. The values at high temperatures are shown <lb/>in Fig. 3. The shift is given as a sum of two components, <lb/>the orbital and the spin parts (K = K orb + K spin ) and <lb/>the spin part is proportional to the spin susceptibility. <lb/>The T dependence of K at high temperatures fits well <lb/>the theoretical curve for a spin-ladder system [24], <lb/>K(T ) = K 0 + <lb/>K 1 <lb/>√ <lb/>T <lb/>exp(−∆ spin /T ). <lb/>(3) <lb/>The gap ∆ spin obtained from the fit is 217K and is com-<lb/>parable with that estimated from T −1 <lb/>1 . The value of <lb/>K 0 is estimated to be 0.25%. The main contribution <lb/>of K 0 comes from K orb comparable with that for high-<lb/>T c cuprates (K 0 of YBCO, for example, is 0.28% [25].) <lb/>The paramagnetic contribution corresponding to the Ko-<lb/>rringa term in T −1 <lb/>1 <lb/>(Eq. (2)) should be included in K 0 . <lb/>The shift in the low temperature region around T c is <lb/>shown in Fig. 4. The raw data at several fields are also <lb/>plotted in the figure since T c depends on the fields. As <lb/>seen from the figure, no appreciable change is seen at T c . <lb/>The paramagnetic contribution in K 0 should be consid-<lb/>erably small so that a change at T c might be difficult to <lb/>detect within the present experimental accuracy. <lb/>We have observed from T −1 <lb/>1 <lb/>and K unexpected fea-<lb/>tures in both normal and superconducting states. At <lb/>pressure of 3.5GP we observed two excitation modes in <lb/>the normal metallic state; one gives rise to the gapless T -<lb/>linear component in T −1 <lb/>1 ( T −1 <lb/>1 <lb/>∝ bT ) which links directly <lb/>with the superconductivity, and the other the activation-<lb/>type component expressed in Eq. (1). The persistence of <lb/>the spin gap at high pressures suggests that quasi-one-<lb/>dimensional spin-charge dynamics is preserved in the nor-<lb/>mal state although pressure increases coupling or hopping <lb/>between the ladders. In the case of conventional metals <lb/>1.0 <lb/>0.8 <lb/>0.6 <lb/>0.4 <lb/>0.2 <lb/>0.0 <lb/>Shift (%) <lb/>250 <lb/>150 <lb/>100 <lb/>T (K) <lb/>H // c axis <lb/>3.5GPa <lb/>Sr 2 Ca 12 Cu 24 O 41 <lb/>FIG. 3: NMR shift of Cu nuclei for the ladder sites at high <lb/>temperatures. Solid curve represents values calculated by Eq. <lb/>(3) in the text. <lb/>2.0 <lb/>1.5 <lb/>1.0 <lb/>0.5 <lb/>0.0 <lb/>Shift (%) <lb/>4.0 <lb/>3.5 <lb/>3.0 <lb/>2.5 <lb/>2.0 <lb/>1.5 <lb/>1.0 <lb/>T (K) <lb/>Tc <lb/>70MHz 6.2T <lb/>76MHz 6.7T <lb/>84MHz 7.4T <lb/>Extrapolation <lb/>(H=0T) <lb/>H // c axis <lb/>Sr 2 Ca 12 Cu 24 O 41 <lb/>FIG. 4: NMR shift of 63 Cu nuclei for the ladder sites at <lb/>low temperatures around Tc. The raw data which show the <lb/>H −2 -linear dependence due to the quadrupole effect are also <lb/>plotted in the figure. <lb/>the gapless T -linear component arises from paramagnetic <lb/>free electrons, however, such a term is hardly expected in <lb/>the present system unless peculiarity of the ladder struc-<lb/>ture is regarded, because majority of spins falls into the <lb/>spin-singlet state due to the existence of the large spin <lb/>gap ∆ spin at low temperatures. <lb/>Since the system is metallic under high pressure, the <lb/>system should be treated in k space. However, mi-<lb/>croscopic snapshot in real space saves to understand <lb/>the existence of the T -linear component as well as the <lb/>activation-type component. The snapshot in real space <lb/>
22
+
23
+ A <lb/>B <lb/>C <lb/>FIG. 5: (A) Illustration of spin and charge configuration at <lb/>high temperatures. Ellipses show spin dimmers on the rung. <lb/>Some of them are in the triplet states at high temperatures. <lb/>Rectangle implies the holon-spinon bound states. They move <lb/>independently in the ladder. (B) Illustration at intermediate <lb/>temperatures. A spin within the holon-spinon bound state is <lb/>free and paramagnetic. Spin dimmers in the ellipses are in <lb/>the singlet state at this temperature region. (C) Illustration <lb/>at the superconducting state. Two spins within the bound <lb/>state form the pairing. <lb/>can be described as follows: The ground state of the un-<lb/>doped ladders is understood as overlap of spin dimmers <lb/>on the rung [4]. Hole doping implies breaking of spin dim-<lb/>mers on the rung and gives rise to holon-spinon bound <lb/>state [26]. At high temperatures singlet-triplet spin ex-<lb/>citation in the spin dimmers away from the bound state <lb/>dominates, which causes the activated behavior of T −1 <lb/>1 <lb/>as is illustrated in Figs. 5A. At intermediate tempera-<lb/>tures majority of the spin dimmers falls into the singlet <lb/>ground state, but the spin in the bound states moves <lb/>rather freely and contributes to the gapless T -linear com-<lb/>ponent in T −1 <lb/>1 (Figs. 5B). The superconductivity would <lb/>be realized by the pairing of two bound states as illus-<lb/>trated in Figs. 5C. In this viewpoint, the spin-gap for-<lb/>mation observed from T −1 <lb/>1 <lb/>at high temperatures does not <lb/>contribute to the pairing formation of the superconduc-<lb/>tivity. The pairing force is expected to be magnetic since <lb/>the system is a strongly correlated electron system, how-<lb/>ever, we cannot exclude the possibility that conventional <lb/>phonon coupling plays some roles in the pairing. <lb/>Finally, it should be noted that the normal state in the <lb/>present system is free from large antiferromagnetic fluc-<lb/>tuation unlike high-T c cuprates. T −1 <lb/>1 <lb/>in typical high-T c <lb/>cuprates such as YBCO systems is expressed as a + bT <lb/>above T c where T -independent term a represents the an-<lb/>tiferromagnetic fluctuation (a ∼ 3x10 3 (sec −1 ) and b ∼ 6 <lb/>(sec −1 K −1 ) for YBCO) [27]. The antiferromagnetic fluc-<lb/>tuation is extremely suppressed due to the existence of <lb/>the spin gap in the present system, which might explain <lb/>why high T c is not realized in this system although the <lb/>structure is quite similiar to high-T c cuprates. <lb/>In conclusion, we have succeeded in obtaining mi-<lb/>croscopic information of the superconducting state in <lb/>Sr 2 Ca 12 Cu 24 O 41 by applying pressure up to 3.5GPa. In <lb/>this material, we observed two excitation modes in the <lb/>normal state. One gives rises to the activation-type com-<lb/>ponent in T −1 <lb/>1 , the other T -linear component linking di-<lb/>rectly with the superconductivity. The superconductiv-<lb/>ity possesses a s-wavelike character in the meaning that <lb/>a finite gap exists in the quasi-particle excitation. <lb/></body>
24
+
25
+ <div type="acknowledgement">Authors wish to thank Profs. H. Fukuyama, K. Ya-<lb/>mada, M. Imada and M. Takigawa and Drs. S. Fujimoto, <lb/>H. Kontani, and K. Kojima for fruitful discussion. The <lb/>present work was partially supported by Grant-in-Aid for <lb/>the Ministry of Education, Science and Culture, Japan, <lb/>and by grants of Ogasawara Foundation for the Promo-<lb/>tion of Science and Engineering, Japan and Simadzu Sci-<lb/>ence Foundation, Japan. <lb/></div>
26
+
27
+ <listBibl>[1] M. Uehara, et al., J. Phys. Soc. Jpn. 65, 2764 (1996) <lb/>[2] Z.Hiroi, et al., J. Solid. State. Chem., 95 (1991) <lb/>[3] T. Osafune, et al., Phys. Rev. Lett. 78 1980 (1997) <lb/>[4] E. Dagotto, J. Riera, and D. Scalapino, Phys. Rev. B45 <lb/>5644 (1992) <lb/>[5] T. M. Rice, S. Gopalan and M. Sigrist, Europhys. Lett. <lb/>23 445 (1993) <lb/>[6] M. Sigrist, T. M. Rice, F. C. Zhang, Phys. Rev. B49 <lb/>(1994) <lb/>[7] K. Kumagai, et al., Phys. Rev. Lett. 78 1992 (1997) <lb/>[8] K. Magishi, et al., Phys. Rev. B57 11533 (1998) <lb/>[9] T. Imai, et al., Phys. Rev. Lett. 81 220(1998) <lb/>[10] S. Katano, et al., Phys. Rev. Lett. 82 636 (1999) <lb/>[11] H. Kontani and K. Ueda, Phys. Rev. Lett. 5619 (1998) <lb/>[12] See for example, H. Yasuoka, et al., Strong Correlation <lb/>and Superconductivity, Springer, Berlin 254(1989): <lb/>Rossat-Mignod,et al., Dynamics of Magnetic fluctuations <lb/>in HighTc materials, Plemum Press, New York, 35 1990 <lb/>[13] T. Nagata, et al., Phys. Rev. Lett. 81 1091 (1998) <lb/>[14] N. Motoyama, et al., Phys. Rev. B55 R3386 (1997) <lb/>[15] T. Nakanishi, et al., Physica B281 282 (2000) <lb/>[16] H. Mayaffre,et al., Science 279 (1998) <lb/>[17] Y. Piskunov, et al. has recently extended the work of <lb/>Ref. [16]. They measured NMR under high pressure up to <lb/>3.6GPa for the field parallel to the b axis. Y. Piskunov,et <lb/>al., Eur. Phys. J. B13 417 (2000): Eur. Phys. J. B 24 443 <lb/>(2001) <lb/>[18] A.P.Melozemoff, et al., Phys. Rev. B 7203 (1988) <lb/>[19] M. Takigawa,et al., Phys. Rev. B57 1124 (1998) <lb/>[20] N. Fujiwara,et al., J. Phys. and Chem. Solids B63 1103 <lb/>(2002) <lb/>[21] M. Corti,et al., Phys. Rev. B 9469 (1996) <lb/>[22] B. J. Suh,et al., Phys. Rev. Lett. 1928 (1996) <lb/>[23] M. H. Kohen and F. Rief, Solid State Physics 321(1957) <lb/>[24] M. Troyer, et al., Phys. Rev. B50 13515 (1994) <lb/>[25] S. E. Barrett,et al., Phys. Rev. B41 6283 (1990) <lb/>[26] H. Tsunetsugu, et al., Phys. Rev. B49 R16078(1994) <lb/>[27] See for example, D. M. Ginsberg, Physical properties of <lb/>high temperature superconductors II, World Scientific 5 <lb/>269 (1990) </listBibl>
28
+
29
+
30
+ </text>
31
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L090137002-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>arXiv:cond-mat/0607671v1 [cond-mat.supr-con] 26 Jul 2006 <lb/>Role of pair-breaking and phase fluctuations in c-axis tunneling in <lb/>underdoped Bi 2 Sr 2 CaCu 2 O 8+δ <lb/>C.J. van der Beek, a P. Spathis, a S. Colson, a P. Gierlowski, b M. Gaifullin, c Yuji Matsuda, d <lb/>P.H. Kes e <lb/>a Laboratoire des Solides Irradiés, Ecole Polytechnique, 91128 Palaiseau, France <lb/>b Institute of Physics of the Polish Academy of Sciences, 32/46 Aleja Lotnikow, 02-668 Warsaw, Poland <lb/>c National Institute for Material Science, 1-2-1, Sengen, Ibaraki, Tsukuba, Japan <lb/>d Department of Physics, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan <lb/>e Kamerlingh Onnes Laboratorium, Rijksuniversiteit Leiden, P.O. Box 9506, 2300 RA Leiden, The Netherlands <lb/>Abstract <lb/>The Josephson Plasma Resonance is used to study the c-axis supercurrent in the superconducting state of under-<lb/>doped Bi2Sr2CaCu2O 8+δ with varying degrees of controlled point-like disorder, introduced by high-energy electron <lb/>irradiation. As disorder is increased, the Josephson Plasma frequency decreases proportionally to the critical tem-<lb/>perature. The temperature dependence of the plasma frequency does not depend on the irradiation dose, and is in <lb/>quantitative agreement with a model for quantum fluctuations of the superconducting phase in the CuO2 layers. <lb/>Key words: Disorder, Interlayer coupling, Josephson Plasma Resonance, Quantum fluctuations <lb/>PACS: 74.40.+k, 74.50.+r, 74.62.-c, 74.62.Dh <lb/></front>
8
+
9
+ <body>1. Introduction <lb/>From the d-wave symmetry of the order pa-<lb/>rameter of cuprate superconductors, one expects <lb/>an enhanced sensitivity of c-axis transport in the <lb/>superconducting state to disorder, due to the en-<lb/>hancement of the quasiparticle density of states <lb/>along the gap node directions, and due to impurity <lb/>assisted hopping [1]. Both mechanisms lead to as yet <lb/>unobserved T 2 -dependences of the c-axis superfluid <lb/>density ρ c <lb/>s at low T , with coefficients that strongly <lb/>depend on the scattering rate Γ. On the other hand, <lb/>underdoped cuprates are sufficiently disordered <lb/>for (quantum) fluctuations of the order parameter <lb/>phase to play a prominent role [2], leading to a T -<lb/>linear behavior of ρ c <lb/>s . Here, we report on the effect <lb/>of controlled point disorder on interlayer tunneling <lb/>of Cooper pairs in underdoped Bi Sr 2 CaCu 2 O 8+δ . <lb/>2. Experimental details <lb/>Single-crystalline <lb/>rods <lb/>of <lb/>underdoped <lb/>Bi 2 Sr 2 CaCu 2 O 8+δ were grown using the travel-<lb/>ling solvent floating zone technique under 25 mBar <lb/>oxygen partial pressure [3]. The samples used for <lb/>the study, with T c ≈ K, were cleaved from the <lb/>same crystalline piece. The crystals were then ir-<lb/>radiated with 2.5 MeV electrons using the van der <lb/>Graaf accelerator of the Laboratoire des Solides <lb/>Irradiés. The irradiation produces homogeneously <lb/>distributed Frenkel pairs, which have previously <lb/>been identified as strong scattering centers [4]. <lb/>The Josephson Plasma Resonance (JPR) fre-<lb/>quency f JP R of the crystals was then measured <lb/>using the cavity perturbation technique, exploiting <lb/>T M 01n harmonic modes to access different frequen-<lb/>cies [6], and the bolometric technique using a waveg-<lb/></body>
10
+
11
+ <note place="footnote">Preprint submitted to Elsevier Science <lb/>7 July 2018 <lb/></note>
12
+
13
+ <body>0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>× 10 <lb/>18 e <lb/>-cm <lb/>-2 <lb/>6 × 10 <lb/>19 e <lb/>-cm <lb/>-2 <lb/>7.7 × 10 <lb/>18 e <lb/>-cm <lb/>-2 <lb/>5.3 × 10 <lb/>17 e <lb/>-cm <lb/>-2 <lb/>unirradiated <lb/>f <lb/>JPR <lb/>(T)/f <lb/>JPR <lb/>(0) <lb/>T / T c <lb/>(b) <lb/>0 <lb/>1000 <lb/>2000 <lb/>3000 <lb/>4000 <lb/>0 <lb/>20 <lb/>40 <lb/>60 <lb/>80 <lb/>f <lb/>JPR <lb/>2 <lb/>(0) ( GHz 2 <lb/>) <lb/>T c ( K ) <lb/>(a) <lb/>Fig. <lb/>1. <lb/>JPR <lb/>frequency <lb/>of <lb/>the <lb/>e − -irradiated <lb/>Bi 2 Sr 2 CaCu 2 O 8+δ crystals, normalized to the JPR fre-<lb/>quency at T → 0, versus reduced temperature. <lb/>uide in the T E 01 travelling wave mode [5]. The <lb/>latter technique allows for swept-frequency mea-<lb/>surements, necessary to elucidate the weak f JP R (T ) <lb/>dependence at low T [5]. Note that f 2 <lb/>JP R is propor-<lb/>tional to the c-axis critical current j c <lb/>c and to the <lb/>c-axis superfluid density: f 2 <lb/>JP R = j c <lb/>c s/2πǫ 0 ǫ r Φ 0 = <lb/>c 2 /4π 2 ǫ r λ 2 <lb/>c ∝ ρ c <lb/>s , with ǫ r the low-frequency dielec-<lb/>tric constant and λ c is the c-axis penetration depth. <lb/>3. Results and Discussion <lb/>Sharp JPR resonant peaks were measured for all <lb/>samples under study. Both T c and f JP R (T → 0) <lb/>decrease with irradiation dose. The sensitivity of <lb/>f JP R to even weak additional disorder contradicts <lb/>the model of coherent interlayer Cooper pair tun-<lb/>neling [7]. Within the framework of a d-wave BCS <lb/>model, the measured proportionality between f <lb/>JP R <lb/>and T c can be understood as resulting either from <lb/>(i) the dependence of the interlayer Josephson cur-<lb/>rent j c <lb/>c ∝ ∆σ qp on the gap magnitude ∆ and the <lb/>quasiparticle conductivity σ qp [8], or (ii) from the <lb/>decrease of the in-plane superfluid density due to <lb/>strong phase fluctuations. <lb/>The origin of the temperature and disorder-<lb/>dependence of f JP R can be pinpointed using <lb/>Fig. 1(b). Normalising all results to f JP R (T → 0) <lb/>and plotting these versus reduced temperature, re-<lb/>veals a common T -dependence independent of disor-<lb/>der. This contradicts the prediction of Ref. [1] that <lb/>ρ s should follow different powers of T depending on <lb/>the ratio of T , ∆, and Γ. However, it agrees with a <lb/>dominant role of quantum phase fluctuations. Then, <lb/>0.01 <lb/>0.1 <lb/>1 <lb/>0.1 <lb/>1 <lb/>Eq. (1) <lb/>5.3×10 17 e -cm -2 <lb/>3×10 18 e -cm -2 <lb/>7.7×10 18 e -cm -2 <lb/>6×10 19 e -cm -2 <lb/>f <lb/>JPR <lb/>2 <lb/>(T) <lb/>λ <lb/>c <lb/>2 <lb/>(0) <lb/>f <lb/>JPR <lb/>2 <lb/>(0) <lb/>λ <lb/>c <lb/>2 <lb/>(T) <lb/>1 -<lb/>~ 1 -<lb/>T / T c <lb/>f JPR <lb/>2 <lb/>(T) <lb/>4α k <lb/>B <lb/>T <lb/>2 C 2 <lb/>σ k B <lb/>T <lb/>f JPR <lb/>2 (0) <lb/>σ ε 0 <lb/>s <lb/>3 <lb/>ε 0 <lb/>s <lb/>= 1 -( ) -<lb/>( ) 2 <lb/>ε 0 s/ k B = 400 K (Ref. [3]) <lb/>α = 0.5 [ Jacobs et al., PRL 75, 4516 (1995)] <lb/>σ = σs = 3.7 e 2 /h i.e. ρ = 1 mΩcm <lb/>C 2 = 10 <lb/>Fig. <lb/>2. <lb/>JPR <lb/>frequency <lb/>of <lb/>the <lb/>e − -irradiated <lb/>Bi 2 Sr 2 CaCu 2 O 8+δ crystals, plotted as 1−f 2 <lb/>J P R (T )/f 2 <lb/>J P R (0) <lb/>in order to bring out the low temperature T -dependence. <lb/>The line is a fit to Eq. 1 [9] with parameters as indicated. <lb/>f <lb/>JP R (T ) <lb/>f 2 <lb/>JP R (0) <lb/>≈ 1 − <lb/>4αk B T <lb/>ε 0 sσ <lb/>− C 2 <lb/>2σ <lb/>3 <lb/>k B T <lb/>ε 0 s <lb/>2 <lb/>(1) <lb/>where α = (∂ε 0 s/∂T ) T →0 , ε 0 = Φ 0 /4πµ 0 λ 2 <lb/>ab with <lb/>λ ab the in-plane penetration depth, s is the CuO 2 <lb/>layer spacing, and σ = σs is the CuO plane sheet <lb/>conductivity [9]. The temperature dependence of the <lb/>JPR frequency arises from the temperature of the in-<lb/>plane phase stiffness ε 0 . Figure 2 shows that Eq. (1) <lb/>describes the results very satisfactorily. <lb/>Summarizing, the temperature-and disorder de-<lb/>pendence of the c-axis Cooper pair tunnel current <lb/>in underdoped Bi Sr 2 CaCu 2 O 8+δ is well described <lb/>assuming a strong effect of quantum phase fluctua-<lb/>tions. A d-wave model without fluctuations cannot <lb/>account for the T -dependence of c-axis coupling. <lb/></body>
14
+
15
+ <listBibl>References <lb/>[1] T. Xiang and J.M. Wheatley, Phys. Rev. Lett. 77, 4632 <lb/>(1996). <lb/>[2] L.B. Ioffe and A. Millis, Science 285, 1241 (1999). <lb/>[3] Ming Li et al. Phys. Rev. B 66, 024502 (2002). <lb/>[4] F. Rullier-Albenque et al., Europhys. Lett. 50, 81 <lb/>(2000). <lb/>[5] M.B. Gaifullin et al., Phys. Rev. Lett. 83, 3928 (1999). <lb/>[6] S. Colson et al., Phys. Rev. Lett. 90, 137002 (2003). <lb/>[7] Yu. Latyshev et al., Phys. Rev. Lett. 82, 5345 (1999). <lb/>[8] Y. Sun and K. Maki, Phys. Rev. B 51, 6059 (1995). <lb/>[9] A. Paramekanti, Phys. Rev. B 65, 104521 (2002). <lb/></listBibl>
16
+
17
+ <page>3 </page>
18
+
19
+
20
+ </text>
21
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L091217001-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>arXiv:cond-mat/0408426v1 [cond-mat.supr-con] 19 Aug 2004 <lb/>CHINESE JOURNAL OF PHYSICS <lb/>VOL. unknown, NO. unknown <lb/>unknown <lb/>Unconventional superconductivity in Na x CoO 2 • yH 2 O <lb/>Hiroya Sakurai , Kazunori Takada , Takayoshi Sasaki 2 , and Eiji <lb/>Takayama-Muromachi 1 <lb/>1 Superconducting Materials Center, National Institute for Materials Science <lb/>(SMC/NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan. <lb/>2 Advanced Materials Laboratory, National Institute for Materials Science (AML/NIMS), <lb/>1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan/ CREST, Japan Science and Technology <lb/>Agency (JST). <lb/>(Received unknown) <lb/>We synthesized powder samples of Na x CoO 2 • yH 2 O changing the volume of the <lb/>water in the hydration process, then investigated their superconducting properties,. It <lb/>was proved that the volume of water is one of key parameters to obtain a single phase of <lb/>Na x CoO 2 • yH 2 O with good superconducting properties. The transition temperature, <lb/>T c , of the sample changed gradually while it was stored in the atmosphere of 70% <lb/>humidity. Superconducting behavior under high magnetic field was very sensitive to <lb/>T c . H c2 of a high quality sample with high T c seemed very high. <lb/>PACS. 70.70.-b -Superconducting Materials. <lb/></front>
8
+
9
+ <body>I. Introduction <lb/>The cobalt oxide superconductor Na x CoO 2 • yH 2 O [1] has characteristic features in <lb/>its structure and Co valence state[2]. The structure is composed of thick insulating layers <lb/>of Na atoms and H 2 O molecules, as well as conducting CoO layers. In the CoO layer, <lb/>CoO 6 octahedra are connected to each other by edge sharing with the Co sites forming a <lb/>triangular lattice. On the other hand, the valence of the Co ion is between 3+ (3d ) and 4+ <lb/>(3d 5 ) as in the case of Na 0.5 CoO 2 [3]. Co 3+ and Co 4+ ions often have low spin states with <lb/>S=0 and 1/2, respectively. At the early stage of research, this compound was considered <lb/>as a system of S=1/2 triangular lattice doped with electrons and the doped resonating-<lb/>valence-bond (RVB) model[4] was expected to be valid for explaining its superconductivity. <lb/>Now, it is widely recognized that the electron density of this compound is so large that the <lb/>doped RVB model can not be applied simply. Most of studies reported thus far suggest <lb/>strongly that the superconductivity of this compound is unconventional. <lb/>The symmetry of superconducting gap function is one of the most important param-<lb/>eters to elucidate the mechanism of superconductivity, and our understanding of super-<lb/>conductivity has become deeper with finding of various types of superconductivity with <lb/>different symmetries. Thus, many researchers have been greatly interested in the super-<lb/>conducting symmetry of Na x CoO 2 • yH 2 O since its discovery. <lb/>Theoretically, various possibilities have been proposed on the superconducting sym-<lb/>metry of the compound (or more generally, for the triangular lattice system), for example, <lb/></body>
10
+
11
+ <front>c 2000 THE PHYSICAL SOCIETY <lb/>OF THE REPUBLIC OF CHINA <lb/></front>
12
+
13
+ <page>2 <lb/></page>
14
+
15
+ <note place="headnote">Unconventional superconductivity in NaxCoO2 • yH2O <lb/>VOL. unknown <lb/></note>
16
+
17
+ <body>d-, chiral d-, chiral p-, and f -waves[5]. These states can be distinguished from each other <lb/>by microscopic experiments such as nuclear quadrupole resonance (NQR), muon spin ro-<lb/>tation/relaxation (µSR), and nuclear magnetic resonance (NMR). In the NQR measure-<lb/>ments, no Hebel-Schlichter peak (coherence peak) was observed in a 1/ 59 T 1 T -T curve (T 1 : <lb/>relaxation time, and T : temperature)[6, 7], which means that the superconducting gap <lb/>has nodes on the Fermi surface. Thus, the symmetry of s-, chiral d-, and chiral p-waves <lb/>are denied. It was pointed out previously for the case of chiral d-or chiral p-wave that if <lb/>superconductivity is disturbed by impurities, the coherence peak is reduced and may not <lb/>be observed[8]. However, an internal magnetic field should exist if chiral d or chiral p is the <lb/>case, and it has never observed by µSR[9]. These results are consistent with specific heat <lb/>data, which suggest the presence of line-nodes[10] in the superconducting gap. Therefore, <lb/>the superconducting symmetry is limited to d-or f -wave. The d-and f -wave symmetries <lb/>can be distinguished from T -dependence of Knight shift, K; sudden change of K below <lb/>T c is expected for spin singlet Cooper pairs with the d-wave symmetry, while invariable <lb/>K below T c along a certain direction for spin triplet pairing with the f -wave symmetry. <lb/>Unfortunately, inconsistent results have been reported thus far on K [11, 12]. The NMR <lb/>measurement usually needs an external magnetic field, and the magnetic field makes it <lb/>difficult to obtain the real (intrinsic) Knight shift, that is, K may apparently change below <lb/>T c caused by demagnetization effect resulting from the superconducting diamagnetism, or <lb/>on the contrary, K may be apparently invariant if the applied magnetic field is close to or <lb/>higher than the upper-critical field, H c2 . Thus, the estimation of H c2 is quite important <lb/>to interpret the NMR data. <lb/>Recently, many H c2 data have been reported. According to magnetic measurements, <lb/>we obtained an extremely high initial slope of dH c2 /dT c | H=0 = 19.3 T/K, by which H c2 <lb/>was calculated to be approximately 60 T if Werthamer-Helfand-Hohenberg (WHH) model <lb/>is applied[13]. This value of the initial slope is close to that estimated from the specific heat <lb/>data[10]. However, by transport measurements, much smaller H c2 has been obtained even <lb/>for the case that the magnetic field was applied perpendicular to c-axis[14]. In the present <lb/>study, we prepared samples carefully and elucidated influence of synthesis conditions on <lb/>the superconducting properties including H c2 . <lb/>II. Experimental <lb/>The powder samples of Na x CoO 2 • yH 2 O were synthesized from Na 0.7 CoO 2 essen-<lb/>tially in the same way as reported in the previous paper[1]. Synthesis of the precur-<lb/>sor of Na 0.7 CoO 2 was described elsewhere[15]. The duration times of the immersion in <lb/>Br 2 /CH 3 CN and water were both 5 days. Volume of water, V w , was varied to find an op-<lb/>timum condition; each sample of Na x CoO 2 (x ∼ 0.4)[2], which was made from Na 0.7 CoO 2 <lb/>with the mass of 1 g, was immersed in water with V w = 10, 50, 100, 300, 500, or 1000 ml. <lb/>The samples were filtered and stored in atmosphere of relative humidity of 70%. The sam-<lb/>ples were characterized by powder X-ray diffraction (XRD) and inductive-coupled plasma <lb/>atomic emission spectroscopy (ICP-AES). XRD measurements were carried out using a <lb/>Bragg-Brentano-type diffractometer (RINT2200HF, Rigaku) with Cu K α radiation. ICP-<lb/></body>
18
+
19
+ <note place="headnote">VOL. unknown <lb/>H. Sakurai et al. <lb/></note>
20
+
21
+ <page>3 <lb/></page>
22
+
23
+ <body>002 <lb/>004 <lb/>Intensity (arb. unit) <lb/>10 ml <lb/>50 ml <lb/>100 ml <lb/>300 ml <lb/>500 ml <lb/>1000 ml <lb/>20 <lb/>18 <lb/>16 <lb/>14 <lb/>12 <lb/>10 <lb/>8 <lb/>6 <lb/>Na <lb/>x CoO <lb/>2 <lb/>2θ ( ) <lb/>Figure 1: (a) XRD patterns of the samples immersed in water with V w . The peak at <lb/>2θ = 15.9˚is assignable to Na x CoO 2 . <lb/>AES measurements were done by dissolving a sample in hydrochloric acid to determine the <lb/>ratio of Na to Co, x. The magnetization, M , of the sample was measured using a com-<lb/>mercial magnetometer with a superconducting quantum interference device (MPMS-XL, <lb/>Quantum Design). Before each measurement under zero-field cooling (ZFC) condition, the <lb/>magnetic field, H, was reset to 0 Oe at T = 10 K or 300 K. <lb/>III. Results and Discussion <lb/>The XRD patterns of all the samples, which were measured just day after the <lb/>filtration, are shown in Fig. 1. As seen in the figure, all the samples immersed in water <lb/>with V w ≤ 500 ml included Na x CoO 2 as an impurity phase. The Na content decreased with <lb/>increasing V w as seen in Fig. 2(a) which means that the Na ions were partly deintercalated <lb/>in the hydration process[2]. The pH of the water was nearly independent of V w being <lb/>approximately 11 and this fact is consistent with the almost linear decrease of x with <lb/>increasing V w . The hydration proceeded completely only for x ≤ 0.35, and thus for <lb/>V w &gt; 500 ml, <lb/>The relative intensity of the impurity peak at 2θ = 15.9˚to the 002 peak of the <lb/>hydrated phase is shown in Fig. 2(b). After the sample was stored for 10 days under 70% <lb/>humidity, amount of the impurity of Na x CoO 2 decreased in every sample and the 500 ml <lb/>sample became single phase. This suggests that the intercalation and/or deintercalation of <lb/></body>
24
+
25
+ <page>4 <lb/></page>
26
+
27
+ <note place="headnote">Unconventional superconductivity in NaxCoO2 • yH2O <lb/>VOL. unknown <lb/></note>
28
+
29
+ <body>10 <lb/>6 <lb/>4 <lb/>2 <lb/>0 <lb/>800 <lb/>600 <lb/>400 <lb/>200 <lb/>0 <lb/>Relative intensity (%) <lb/>1 day after filtration <lb/>10 days after filtration <lb/>Vw (ml) <lb/>(b) <lb/>0.39 <lb/>0.38 <lb/>0.37 <lb/>0.36 <lb/>0.35 <lb/>0.34 <lb/>1000 <lb/>800 <lb/>600 <lb/>400 <lb/>200 <lb/>0 <lb/>1.45 <lb/>1.40 <lb/>1.35 <lb/>1.30 <lb/>1.25 <lb/>1.20 <lb/>Vw (ml) <lb/>(a) <lb/>Na content, x <lb/>Water content, y <lb/>Figure 2: (a) The Na and water contents of the samples immersed in water with various <lb/>V w . The analyses was performed 1 day after the filtration. The broken and dotted lines are <lb/>visual guides for x and y, respectively. (b) Relative intensities of the peak at 2θ = 15.9c <lb/>aused by Na x CoO 2 to the 002 peak of Na x CoO 2 • yH 2 O measured 1 day and 10 days after <lb/>filtration. The broken and dotted lines are visual guides. <lb/>water proceeded through the gas phase during the storage. Moreover, the Na atoms can <lb/>also be deintercalated forming NaOH on the surface of the grains. The 300 ml sample did <lb/>not become single phase even after it was stored for a month. <lb/>The magnetic susceptibility of each sample, which was measured within a few days <lb/>after filtration, is shown in Fig. 3. As seen in this figure, all the samples except the 1000 <lb/>ml one showed superconductivity. The variation of T c is shown as a function of V w in the <lb/>inset. Obviously, the samples immersed in the water with V w ≤ 300 ml showed better <lb/>superconducting properties, although they are not single phase. On the other hand, the <lb/>superconducting properties changed after the storage as seen in Fig. 4. The superconduct-<lb/>ing properties of the 500 ml and 1000 ml samples were improved drastically after they were <lb/>stored for 2 weeks, while those of the 300 ml sample became rather bad after the 2 month <lb/>storage. These facts are consistent with the idea that, during the storage, the intercalation <lb/>and/or deintercalation of water can proceed through the gas and the Na atoms can also <lb/>be deintercalated being extracted as NaOH. T c seems to depend on both the Na and the <lb/></body>
30
+
31
+ <note place="headnote">VOL. unknown <lb/>H. Sakurai et al. <lb/></note>
32
+
33
+ <page>5 <lb/></page>
34
+
35
+ <body>M/H (10 <lb/>-3 <lb/>emu/g) <lb/>T (K) <lb/>-12 <lb/>-10 <lb/>-8 <lb/>-6 <lb/>-4 <lb/>-2 <lb/>0 <lb/>6 <lb/>5 <lb/>4 <lb/>3 <lb/>2 <lb/>1000 ml <lb/>ml <lb/>ml <lb/>ml <lb/>50 ml <lb/>ml <lb/>T <lb/>C (K) <lb/>5.0 <lb/>4.5 <lb/>4.0 <lb/>3.5 <lb/>3.0 <lb/>2.5 <lb/>2.0 <lb/>1000 <lb/>800 <lb/>600 <lb/>400 <lb/>200 <lb/>0 <lb/>V W (ml) <lb/>Figure 3: M/H-T curves of the samples immersed in water with various V w measured a <lb/>few days after filtration. The inset shows T c of the samples. <lb/>Figure 4: Variations of M/H-T curves after the storage for the samples with V w = 500 <lb/>ml (a), 1000 ml (b), and 300 ml (c). <lb/>water content and there seems to be optimum values in them. According to the present <lb/>study, the best synthesis condition is the immersion to 500 ml water for 1 g of the starting <lb/>oxide followed by 2 weeks storage under 70% humidity. <lb/>The M/H-T curves measured under various magnetic fields are shown in Fig. 5. <lb/>The samples A and B are different from those mentioned above. Each sample was single <lb/>phases. The transition temperatures of samples A and B without magnetic field, T c (0), <lb/>are approximately 4.2 and 4.6 K, respectively. As shown in Fig. 5, T c under H ≥ 5 T is <lb/>hard to determine for the both samples, because the superconducting transitions are broad <lb/>[13]. However, it is notable that the superconductivity of sample A is suppressed almost <lb/>completely under 7 T while M/H of sample B shows clear downturn below, at least, 3.5 K <lb/>even under 7 T. Two explanations are possible for this result. First, H c2 depends strongly <lb/>on the sample quality and is improved greatly with slight increase of T c . Second, since 7 T <lb/>M/H-T curves of samples A and B both start to deviate from the linear lines near T c (0) <lb/>(see Fig. 5), their T c do not decrease significantly under 7 T but their superconducting <lb/>volume fractions are different under 7 T. In either case, H c2 seems to be quite high if it is <lb/>measured for a high quality of sample. <lb/>The important thing is that the superconductivity under high magnetic field depends <lb/>strongly on T c , and T c is influenced by the synthesis conditions such as the volume of water <lb/></body>
36
+
37
+ <page>6 <lb/></page>
38
+
39
+ <note place="headnote">Unconventional superconductivity in NaxCoO2 • yH2O <lb/>VOL. unknown <lb/></note>
40
+
41
+ <body>Figure 5: M/H-T curves of the samples A (a) and B (b) measured under various fields. <lb/>All the curves except those under 10 kOe are off-set to be distinguished. The dotted lines <lb/>are visual guides. <lb/>in the hydration process or the storage time as mentioned above. This seems to be the <lb/>reason for the inconsistent experimental results reported for the present system especially <lb/>in the estimations of H c2 and the NMR measurements. Our high quality samples with <lb/>high T c values have been investigated by µSR, NMR, specific heat measurements, and <lb/>magnetic measurements[9, 12, 13, 16] to give consistent results. In 1/ 59 T 1 T determined by <lb/>NMR, superconducting transition is detected clearly with T c at approximately 4 K even <lb/>under a high magnetic field of ∼7 T[17] consistent with the high H c2 estimated from the <lb/>specific heat and magnetic measurements. <lb/>IV. Summary <lb/>We synthesized powder samples of Na x CoO 2 • yH 2 O by changing the volume of water <lb/>in the hydration process, and investigated their superconducting properties. The single <lb/>phase of Na x CoO 2 • yH 2 O was obtained when the volume of water was 500 ml and 1000 <lb/>ml per 1 g of starting materials of Na 0.7 CoO 2 . T c of the sample changed gradually while <lb/>it was stored in the atmosphere of the humidity of 70%. The superconducting properties <lb/>under high magnetic field were very sensitive to T c , and thus, to the synthesis conditions. <lb/>These facts suggest that discrepancies in experimental results for the present compound <lb/>result from the difference in sample quality. H c2 is obviously higher than 7 T in a high <lb/>quality sample. <lb/></body>
42
+
43
+ <div type="acknowledgement">Acknowledge <lb/>Special thanks to S. Takenouchi (NIMS) for his chemical analyses. We would like <lb/>to thank H. D. Yang (National Sun Yat Sen University), J.-Y. Lin (National Chiao Tung <lb/>University), F. Izumi, R. A. Dilanian, A. Tanaka (NIMS), K. Yoshimura, M. Kato, C. <lb/>Michioka, T. Waki, K. Ishida (Kyoto University), W. Higemoto (JAERI), for useful dis-<lb/>cussion. This study was partially supported by Grants-in-Aid for Scientific Research (B) <lb/>from Japan Society for the Promotion of Science (16340111). One of the authors (H.S) is <lb/>research fellow of the Japan Society for the Promotion of Science. <lb/></div>
44
+
45
+ <listBibl>References <lb/>[1] K. Takada, H. Sakurai, E. Takayama-Muromachi, F. Izumi, R. A. Dilanian and T. <lb/>Sasaki, Nature (London) 422, 53 (2003). <lb/>[2] K. Takada, K. Fukuda, M. Osada, I. Nakai, F. Izumi, R. A. Dilanian, K. Kato, M. <lb/>Takata, H. Sakurai, E. Takayama-Muromachi and T. Sasaki, J. Mater. Chem. 14, <lb/>1448 (2004). <lb/></listBibl>
46
+
47
+ <note place="headnote">VOL. unknown <lb/>H. Sakurai et al. <lb/></note>
48
+
49
+ <page>7 <lb/></page>
50
+
51
+ <listBibl>[3] D. J. Singh, Phys. Rev. B 61, 13397 (2000). <lb/>[4] P. W. Anderson, Mat. Res. Bull. 153 (1973). <lb/>[5] G. Baskaran, Phys. Rev. Lett. 91, 097003 (2003); B. Kumar and S. Shastry, Phys. <lb/>Rev. B 68,104508 (2003); M. Ogata, J. Phys. Soc. Jpn. 72, 1839 (2003) ; A. Tanaka <lb/>and X. Hu, Phys. Rev. Lett. 91, 257006 (2003); C. Honerkamp, Phys. Rev. B 68, <lb/>104510 (2003); H. Ikeda, Y. Nisikawa and K. Yamada, J. Phys. Soc. Jpn. 73, 17 <lb/>(2004); Q.-H. Wang, D.-H. Lee and P. A. Lee, Phys. Rev. B 69, 092504 (2004); Y. <lb/>Tanaka, Y. Yanase and M. Ogata, J. Phys. Soc. Jpn. 73, 319 (2004); Y. Nisikawa, H. <lb/>Ikeda and K. Yamada, J. Phys. Soc. Jpn. 73, 1127 (2004); K. Kuroki, Y. Tanaka and <lb/>R. Arita, to appear in Phys. Rev. Lett. (cond-mat/0311619). <lb/>[6] K. Ishida, Y. Ihara, Y. Maeno, C. Michioka, M. Kato, K. Yoshimura, K. Takada, T. <lb/>Sasaki, H. Sakurai and E. Takayama-Muromachi, J. Phys. Soc. Jpn. 72, 3041 (2003). <lb/>[7] T. Fujimoto, G.-q. Zheng, Y. Kitaoka, R. L. Meng, J. Cmaidalka and C. W. Chu, <lb/>Phys. Rev. Lett. 92, 047004 (2004). <lb/>[8] Y. Bang, M. J. Graf and A. V. Balatsky, Phys. Rev. B 68, 212504 (2003); Q. Han <lb/>and Z. D. Wang, cond-mat/0308160. <lb/>[9] W. Higemoto, K. Ohishi, A. Koda, R. Kadono, K. Ishida, K. Takada, H. Sakurai, E. <lb/>Takayama-Muromachi and T. Sasaki, Phys. Rev. B, in print (cond-mat/0310324). <lb/>[10] H. D. Yang, J.-Y. Lin, C. P. Sun, Y. C. Kang, K. Takada, T. Sasaki, H. Sakurai and <lb/>E. Takayama-Muromachi, cond-mat/0308031. <lb/>[11] Y. Kobayashi, M. Yokoi and M. Sato, J. Phys. Soc. Jpn. 72, 2453 (2003). <lb/>[12] T. Waki, C. Michioka, M. Kato, K. Yoshimura, K. Takada, H. Sakurai, E. Takayama-<lb/>Muromachi and T. Sasaki, cond-mat/0306036. <lb/>[13] H. Sakurai, K. Takada, S. Yoshii, T. Sasaki, K. Kindo and E. Takayama-Muromachi, <lb/>Phys. Rev. B 68, 132507 (2003). <lb/>[14] T. Sasaki, P. Badica, N. Yoneyama, K. Yamada, K. Togano and N. Kobayashi, J. <lb/>Phys. Soc. Jpn. 73, 1131 (2004). <lb/>[15] H. Sakurai, S. Takenouchi, N. Tsujii and E. Takayama-Muromachi, J. Phys. Soc. Jpn. <lb/>73, No. 8 in print. <lb/>[16] C. Michioka, M. Kato, K. Yoshimura, K. Takada, H. Sakurai, E. Takayama-Muromachi <lb/>and T. Sasaki, cond-mat/0403293. <lb/>[17] M. Kato, C. Michioka, T. Waki, K. Yoshimura, K. Ishida, H. Sakurai, E. Takayama-<lb/>Muromachi, K. Takada and T. Sasaki, unpublished. <lb/></listBibl>
52
+
53
+ <body>This figure &quot;Fig4.gif&quot; is available in &quot;gif&quot; format from: <lb/>http://arxiv.org/ps/cond-mat/0408426v1 <lb/>This figure &quot;Fig5.gif&quot; is available in &quot;gif&quot; format from: <lb/>http://arxiv.org/ps/cond-mat/0408426v1 </body>
54
+
55
+
56
+ </text>
57
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L092157004-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>arXiv:cond-mat/0306659v2 [cond-mat.str-el] 17 Oct 2003 <lb/>Thermodynamic and Transport Measurements on Superconducting Na x CoO 2 • yH 2 O <lb/>Single Crystals Prepared by Electrochemical De-intercalation <lb/>F.C. Chou, 1 J.H. Cho, 2 † P.A. Lee, 1,2 E.T. Abel, 1,2 , K. Matan, 1,2 and Y.S. Lee 1,2 <lb/>1 Center for Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, <lb/>Department of Physics, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 <lb/>(Dated: November 3, 2018) <lb/>Superconducting single crystal samples of NaxCoO2 • yH2O have been produced using an elec-<lb/>trochemical technique which dispenses with the usual bromine chemical de-intercalation step and <lb/>permits much more precise control of the Na content. After careful hydration, nearly single-phase <lb/>crystals have been obtained in which over 90% of the sample&apos;s volume corresponds to the supercon-<lb/>ducting Na0.3CoO2 • 1.3H2O structure. Susceptibility and specific heat measurements confirm that <lb/>bulk superconductivity has been achieved. The extracted normal state density of states indicates <lb/>Fermi-liquid behavior with strong mass enhancement and a modest Wilson ratio. Measurements of <lb/>Hc2 for H c and H ab reveal significant anisotropy. The estimated value of Hc2 for H c yields <lb/>a coherence length of ∼ 100Å, consistent with an extremely narrow bandwidth. <lb/>PACS numbers: PACS numbers: 74.70.-b, 74.62.Bf, 74.25.Bt, 74.25.Fy <lb/></front>
8
+
9
+ <body>Sodium cobalt oxide, Na x CoO 2 (x ≈ 0.65 to 0.75), <lb/>has received considerable attention due to its unusual <lb/>thermal electric properties.[1] Recent studies have re-<lb/>vealed anomalous non-Fermi liquid behavior in trans-<lb/>port properties which point to the importance of strong <lb/>correlations.[2, 3] The cobalt ions form a hexagonal lay-<lb/>ered structure and the formally 1−x fraction is in the low <lb/>spin (S = 1 <lb/>2 ) Co 4+ state, while the x fraction is in the <lb/>S = 0 Co 3+ state. Interest in this material escalated with <lb/>the discovery of superconductivity by Takada et al.[4] in <lb/>Na x CoO 2 • yH 2 O when the sodium concentration is re-<lb/>duced to about 0.3 and water is intercalated between the <lb/>layers. The ability to control the sodium content is an <lb/>exciting development because, in principle, the limit of <lb/>x = 0 corresponds to a Mott insulator on a triangular <lb/>lattice with S = 1 <lb/>2 . Then the hydrated compound can <lb/>be viewed as electron doping of a Mott insulator with a <lb/>doping concentration x ≈ 0.3. As such, it is the second <lb/>known example of superconductivity arising from dop-<lb/>ing a Mott insulator after the high T c cuprates. The <lb/>S = 1 <lb/>antiferromagnet on a triangular lattice was in fact <lb/>the starting point of Anderson&apos;s resonating valence bond <lb/>(RVB) idea,[5] and the new superconductor has been in-<lb/>terpreted in this light.[6, 7, 8] <lb/>Until recently, the bulk properties of superconducting <lb/>Na 0.3 CoO 2 • 1.3H 2 O has been studied mostly using pow-<lb/>der samples.[9, 10, 11, 12, 13] Single crystal measure-<lb/>ments [13] are much less plentiful, though such measure-<lb/>ments are extremely important in order to understand <lb/>this anisotropic layered compound. In this paper we re-<lb/>port a new electrochemical method to extract Na from <lb/>Na x CoO 2 . This is an alternative to the chemical de-<lb/>intercalation of Na using Br ions introduced by Takada <lb/>et al.[4] Our electrochemical method permits precise con-<lb/>trol of the Na content and avoids the environmental haz-<lb/>ards associated with the use of high molar concentra-<lb/>tions of Br. In addition, we have succeeded in growing <lb/>large single crystals of Na x CoO 2 by the floating zone (FZ) <lb/>method. By applying the electrochemical procedure to <lb/>the FZ crystals, we have obtained superconducting single <lb/>crystal samples with which we have carried out a variety <lb/>of physical property measurements. <lb/>The initial Na 0.75 CoO 2 polycrystalline material was <lb/>prepared using Na 2 CO 3 and Co 3 O 4 with Na to Co ratio <lb/>at 0.75 to 1. The thoroughly mixed and ground powder <lb/>was calcined at 750C for 12 hours and then reacted at <lb/>850C for 24 hours with frequent grindings in between. <lb/>Na loss was minimized with a fast-heating method.[14] <lb/>A stoichiometric Na 0.75 CoO 2 feed rod was melted and <lb/>re-crystallized with an optical floating-zone furnace (4-<lb/>Xe lamp design, CSI Japan) under oxygen atmosphere. <lb/>Although Na vapor loss was observed from the white de-<lb/>posit on the walls of the quartz sample chamber, this loss <lb/>was minimized by choosing a fast pulling rate. A stable <lb/>molten zone can be maintained with various pulling rates <lb/>from 1.5 to 10 mm/hr. Electron probe microscopy anal-<lb/>ysis (EPMA) indicates the FZ crystal has Na content <lb/>between 0.71 and 0.74 for a crystal pulled with 2 mm/hr <lb/>rate. We have grown large single crystals of Na 0.75 CoO 2 <lb/>(5mm diameter × cm long) successfully with this FZ <lb/>method. Powder neutron diffraction confirms that the <lb/>Na level of the FZ crystal is essentially identical to that <lb/>in the feed rod.[15] <lb/>A three-electrode electrochemical cell was set up using <lb/>the Na x CoO 2 sample as a working electrode, platinum <lb/>foil as a counter electrode, Ag/AgCl as a reference elec-<lb/>trode (E 0 = +0.222 V vs. the standard hydrogen elec-<lb/>trode), and 1M NaOH as an electrolyte. The proposed <lb/>half reaction at the anodically polarized Na x CoO 2 elec-<lb/>trode responsible for Na extraction is <lb/>N a x CoO 2 + δ(OH) − → N a x−δ CoO 2 + δ(N aOH) + δe − , <lb/>where the generated NaOH becomes dissolved in the <lb/>electrolyte. Because Na x CoO 2 is metallic, the work-<lb/></body>
10
+
11
+ <page>2 <lb/></page>
12
+
13
+ <body>-0.2 <lb/>-0.1 <lb/>0 <lb/>0.1 <lb/>0.2 <lb/>0.3 <lb/>0.4 <lb/>0.5 <lb/>0.6 <lb/>0 <lb/>5 <lb/>10 15 20 25 30 35 40 <lb/>OCP (V vs Ag/AgCl Reference) <lb/>time (10 <lb/>3 sec) <lb/>Na <lb/>2/3 <lb/>Na <lb/>1/2 <lb/>Na <lb/>1/3 <lb/>FIG. 1: The quasi-open circuit potential (OCP) as a function <lb/>of time during charging of the EC cell. The arrows indicate <lb/>regions of special stability and are labelled with possible Na <lb/>concentrations associated with these regions. <lb/>ing electrode can be prepared directly using a com-<lb/>pressed powder or a single crystal sample. Hydrated <lb/>Na 0.3 CoO 2 • yH 2 O can be prepared by having the sam-<lb/>ple anodically polarized with a constant voltage (0.6-1.2 <lb/>V) until the decaying anodic current reaches a constant <lb/>value. Alternatively, a constant current (of 0.1-10 mA) <lb/>can be used to achieve the same charge level by wait-<lb/>ing until the final open circuit potential reaches ∼ 0.5 <lb/>V vs. Ag/AgCl. Samples obtained directly from the cell <lb/>are typically in a mixture of the partially hydrated [9] <lb/>(c ≃ 13.8Å) and fully hydrated (c ≃ 19.7Å) structures <lb/>and show only trace amounts of superconductivity. Fully <lb/>hydrated superconducting crystals can be achieved by <lb/>sealing the sample in a water vapor saturated container <lb/>at room temperature. We find that the rate of hydration <lb/>depends on grain size and several months are required <lb/>to fully hydrate mm-sized single crystals. The sodium <lb/>content in each sample was checked by electron probe <lb/>microanalysis. After electrochemical treatment, our as-<lb/>prepared floating-zone single crystals with x = 0.75 typ-<lb/>ically break apart into smaller crystals (∼5×5×2 mm) <lb/>with x ≃ 0.28 − 0.32. <lb/>The quasi-open circuit potential (OCP) was measured <lb/>as a function of time during a repeated sequence of hav-<lb/>ing a current density of 0.01 A/g turned on (for 100 sec-<lb/>onds) and off (for 100 seconds). Here, the quasi OCP <lb/>is the potential recorded during the time the current is <lb/>off. We can interpret the OCP measured in this manner <lb/>as the chemical potential of the surface layer of the sam-<lb/>ple. In Fig. 1 we plot the quasi-OCP vs. time where the <lb/>charging current density is 0.01 A/g. The time axis is <lb/>therefore proportional to the total charge supplied to the <lb/>sample surface. A plateau in the OCP corresponds to <lb/>two-phase coexistence. We can identify three prominent <lb/>plateaus in Fig. 1. The last plateau at 0.48 volts indi-<lb/>cates a saturation of the Na concentration, at least on <lb/>the surfaces of the grains of the crystal. Assuming that <lb/>this saturation level corresponds to Na 1/3 and knowing <lb/>0 <lb/>0.001 <lb/>0.002 <lb/>0 <lb/>50 <lb/>100 <lb/>150 <lb/>200 <lb/>250 <lb/>300 <lb/>M/H (cm <lb/>3 <lb/>/mole Co) <lb/>T (K) <lb/>H //c <lb/>H//ab <lb/>-0.08 <lb/>-0.06 <lb/>-0.04 <lb/>-0.02 <lb/>0 <lb/>0.02 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>AC <lb/>c&apos; &amp; <lb/>c&quot; (cm <lb/>3 <lb/>/mole Co ) <lb/>c&quot; <lb/>c&apos; <lb/>H//c <lb/>FIG. 2: The magnetic susceptibility measured using a SQUID <lb/>magnetometer with an applied field of 1 Tesla with H ab and <lb/>H c. Inset: The AC susceptibilities (in-phase χ ′ and out-of-<lb/>phase χ ′′ ) at low temperatures. <lb/>that the starting material is Na 0.75 , we have drawn ar-<lb/>rows in Fig. 1 to indicate the expected Na content with <lb/>the assumption that the Na extraction is proportional to <lb/>the charging. The close match of these arrows with the <lb/>plateaus suggests that stable intermediate phases exist <lb/>for Na 2/3 and Na 1/2 . We speculate that the special frac-<lb/>tions of x ≃ 1 <lb/>3 , 1 <lb/>2 , and 2 <lb/>3 may be indicative of at least <lb/>partial Na ordering relative to the hexagonal Co struc-<lb/>ture. The role of Na ordering on the physics of these <lb/>materials is an important subject for future studies. <lb/>Magnetic susceptibility, specific heat and resistivity <lb/>have been measured on our superconducting single crys-<lb/>tal samples. As shown in Fig. 2, the susceptibility <lb/>shows a large anisotropy, similar to that reported for <lb/>Na 0.68 CoO 2 .[2] However, the Curie-Weiss susceptibility <lb/>that is so prominent there is now absent. The small cusp <lb/>at 42 K is likely due to a Co 3 O 4 impurity phase, which <lb/>is known to have an antiferromagnetic transition in the <lb/>33 to 46 K range.[16] By comparing the size the magne-<lb/>tization cusp with that for bulk Co 3 O 4 , we estimate that <lb/>the Co 3 O 4 impurity fraction of our sample is small, at <lb/>the 1% level. However, we cannot detect the existence <lb/>of Co 3 O 4 within our x-ray diffraction sensitivity. Apart <lb/>from this anomaly and the low temperature Curie tail <lb/>(which arises from only 0.5% of the Co moments), the <lb/>susceptibility is nearly temperature independent. Note <lb/>the large magnitude and strong anisotropy. For compar-<lb/>ison, the magnitude is about 5 times that of lanthanum <lb/>strontium cuprate (La 2−x Sr x CuO 4 ) and 30 times that of <lb/>a simple metal like Na. The anisotropy is probably due <lb/>to a combination of anisotropy in the g factor and the <lb/>van Vleck term. AC susceptibility measurements were <lb/>taken in a field of 3 Gauss/100 Hz as shown in the inset <lb/>of Fig. 2. The diamagnetic signal is indicative of a su-<lb/>perconducting phase with an onset temperature of about <lb/>4.2 K, and the screening fraction is estimated to be about <lb/>120% (without geometric correction). <lb/>The specific heat of an array of co-aligned single crys-<lb/></body>
14
+
15
+ <page>3 <lb/></page>
16
+
17
+ <body>tals (combined mass of 4.7 mg) was measured using a <lb/>Physical Property Measurement System (Quantum De-<lb/>sign) in applied magnetic fields ranging from 0 T to 14 <lb/>T and temperatures ranging from 0.37 K to K as <lb/>shown in Fig. 3. The sample had been hydrated for three <lb/>months after electrochemical de-intercalation, and x-ray <lb/>diffraction indicates that 94% of the sample consists of <lb/>the fully hydrated superconducting structure. In zero <lb/>field, a pronounced peak is observed at 4.2 K, indicating <lb/>the transition to bulk superconductivity. A second broad <lb/>peak exists at lower temperature around 1.5 K. In a field <lb/>of 1 T, the superconducting anomaly at 4.2 K is strongly <lb/>suppressed; in contrast, the peak at 1.5 K is slightly en-<lb/>hanced. This suggests that the peak around 1.5 K is <lb/>not related to a second superconducting phase with a <lb/>lower T c . Such a peak may be compatible with a model <lb/>of weakly interacting localized Co moments, analogous to <lb/>the effects seen in impurity-doped semiconductors.[17] In <lb/>fields larger than 5 T, both peaks disappear and are re-<lb/>placed by a broad enhancement of C/T over a wide range <lb/>of temperatures. With increasing field, the enhancement <lb/>of C/T shifts to higher temperatures. <lb/>In the inset of Fig. 3, the top panel compares the sup-<lb/>pression of the superconducting anomaly for H c and <lb/>H ab. In order to achieve a comparable suppression, <lb/>an in-plane field with magnitude 5 times that of the the <lb/>out-of-plane field is required. This reflects the anisotropy <lb/>of H c2 in the different field orientations and is consistent <lb/>with our resistivity measurements of H c2 (discussed be-<lb/>low). We note that ∆C/T c in zero-field for our single <lb/>crystal sample is comparable to that reported by other <lb/>groups in powder samples.[11, 12] The data plotted in <lb/>the main part of Fig. 3 exhibit an upturn at the lowest <lb/>measured temperatures which increases with increasing <lb/>field. This is most likely due to a Schottky contribution <lb/>from the nuclear spins. The bottom panel of the inset <lb/>shows C/T data at high-fields in which a nuclear Schot-<lb/>tky contribution (6.7×10 −6 B 2 /T 2 ) has been subtracted. <lb/>These data demonstrate that in fields greater than ∼10 T <lb/>superconductivity is completely suppressed. Hence, 10 T <lb/>may be taken as an upper limit for H c2 with H ab. <lb/>On different single crystal sample which had a smaller <lb/>hydrated phase fraction, we had observed an anomaly in <lb/>the specific heat at T = 6 K. This feature is also notice-<lb/>able in the powder data of Jin et al.,[13]. We found that <lb/>this anomaly was insensitive to magnetic fields as large <lb/>as 14 T. The origin of this anomaly is not clear, and it <lb/>may be related to another order parameter in the vicinity <lb/>of the superconducting phase on the phase diagram. <lb/>From our specific heat data, we extract a normal state <lb/>γ value of 16.6 mJ/K Co-mole. This corresponds to <lb/>a free electron density of states (DOS) including both <lb/>spins of 7.09 states/eV. At first sight, this seems to com-<lb/>pare well with the LDA band structure result[18] of 4.4 <lb/>states/eV. However, this apparent agreement is mislead-<lb/>ing because the LDA band consists of 3 overlapping t 2g <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>6 <lb/>7 <lb/>0.01 <lb/>0.02 <lb/>0.03 <lb/>0.04 <lb/>0.05 <lb/>0.06 <lb/>0.07 <lb/>0.08 <lb/>1 <lb/>10 <lb/>0.01 <lb/>0.02 <lb/>0.03 <lb/>0.04 <lb/>0.05 <lb/>0.06 <lb/>0 <lb/>10 <lb/>20 <lb/>30 <lb/>40 <lb/>0.020 <lb/>0.025 <lb/>0.030 <lb/>0.035 <lb/>C <lb/>p <lb/>/T <lb/>(J/K <lb/>2 <lb/>mol) <lb/>T (K) <lb/>14 <lb/>10 <lb/>5 <lb/>1 <lb/>0.2 <lb/>0 <lb/>H (T) <lb/>H || c <lb/>T <lb/>2 (K <lb/>2 <lb/>) <lb/>10 T, H || c <lb/>14 T, H || c <lb/>10 T, H || ab <lb/>14 T, H || ab <lb/>T <lb/>0.2 T, H || c <lb/>T, H || ab <lb/>FIG. 3: The specific heat measured on a single crystal plot-<lb/>ted as cp/T versus T in various applied magnetic fields. In <lb/>the top panel of the inset, we plot cp/T versus T 2 showing <lb/>the suppression of the superconducting anomaly for H c and <lb/>H ab. In the bottom panel of the inset, we plot cp/T for fields <lb/>of 10 T and 14 T (a nuclear Schottky contribution has been <lb/>subtracted). <lb/>bands whereas the true quasiparticle is expected to form <lb/>a single band out of the A 1g orbital (symmetric combina-<lb/>tion of t 2g ), split off from the rest by correlation. A better <lb/>way is to extract a bandwidth which we estimate to be <lb/>1.4 eV. Even though the LDA calculation was done for <lb/>x = 0.5 and without hydration, the bandwidth should <lb/>be insensitive to these differences. On the other hand, <lb/>we can fit the observed DOS to that of the free electron <lb/>tight binding band on a triangular lattice with hopping <lb/>matrix element t ef f . With t ef f &lt; 0 and x = 0.35, the <lb/>tight binding DOS for both spins is 0.16/|t ef f |.[7] The <lb/>measured DOS then implies that |t ef f | ≈ meV, or a <lb/>full bandwidth of 9|t ef f | ≈ 0.2 eV, which is a factor of 7 <lb/>smaller than the LDA bandwith. We therefore conclude <lb/>that there is a mass enhancement of ∼ 7 compared with <lb/>band theory. <lb/>By assuming g = 2, we find that the measured χ c <lb/>of 3.5 × 10 −4 cm 3 /Co-mole corresponds to a free elec-<lb/>tron DOS of 10.86 states/eV. This gives a Wilson ra-<lb/>tio 4π 2 k 2 /3(gµ B ) 2 χ/γ of 1.53. In case a significant <lb/>portion of χ comes from the van Vleck term, the Wil-<lb/>son ratio will be even smaller. The combination of sus-<lb/>ceptibility and specific heat measurements indicate that <lb/>Na 0.3 CoO 2 •yH 2 O may be viewed as a Fermi liquid with <lb/>strong mass enhancement due to correlations. The occur-<lb/>rence of superconductivity in such a narrow band mate-<lb/></body>
18
+
19
+ <page>4 <lb/></page>
20
+
21
+ <body>0 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>6 <lb/>7 <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>5 <lb/>6 <lb/>7 <lb/>0 <lb/>50 <lb/>100 150 200 250 300 <lb/>r <lb/>ab <lb/>(mW cm) <lb/>r <lb/>c (W cm) <lb/>T(K) <lb/>r c (W cm) <lb/>r ab (mW cm) <lb/>0 <lb/>0.2 <lb/>0.4 <lb/>0.6 <lb/>0.8 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>r <lb/>c <lb/>(W cm) <lb/>FIG. 4: Resistivity in the ab plane and along the c axis. Low <lb/>temperature results for ρc are shown in an expanded scale in <lb/>the inset. <lb/>rial suggests an electronic rather than phononic mecha-<lb/>nism. <lb/>Four terminal resistivity measurements on a single <lb/>crystal are shown in Fig. 4. The zero resistance state <lb/>is achieved in ρ c and ρ ab at low temperatures below <lb/>T c ≃ 4.2 K (the inset shows an expanded view of ρ c (T ) <lb/>near the transition). The large room-to low-temperature <lb/>resistance ratio (∼40) indicates that the sample is a good <lb/>metal with large anisotropy between ρ c and ρ ab (grow-<lb/>ing to 10 4 at low temperatures). The conductance per <lb/>square at low T of each layer is 50 (e 2 /h), indicative of a <lb/>long mean free path. The current and voltage leads were <lb/>attached with silver paste which was allowed to dry in at-<lb/>mosphere for 12 hours. Even though partial dehydration <lb/>may have occurred during this process, it appears that <lb/>superconductivity in the single crystal sample, with its <lb/>smaller surface-to-volume ratio, is more robust than in <lb/>the powder samples where superconductivity is found to <lb/>degrade in a matter of minutes.[9] The large peak below <lb/>K reported by Jin et al.[13] is absent. <lb/>In order to make an estimate of H c2 , we have measured <lb/>the in-plane resistance R ab versus H at various tempera-<lb/>tures as shown in Fig. 5(a). There are two regimes (at low <lb/>fields and high fields) where the resistance is roughly pro-<lb/>portional to H. We define H c2 as the field at the crossing <lb/>point of the extrapolation of these regimes. In Fig. 5(b), <lb/>we plot H c2 vs. T for both field orientations. There is <lb/>about a factor of 5 difference in the slopes near T c , con-<lb/>sistent with the heat capacity data. The value of H c2 for <lb/>H ab is consistent with the Pauli paramagnetic limit <lb/>for pair breaking. From the H c2 vs. T curve for H c, <lb/>we determine a coherence length of ξ ≈ 100Å. This rela-<lb/>tively short coherence length is surprising for a supercon-<lb/>ductor with such a low T c , but is entirely consistent with <lb/>the narrow bandwidth. Assuming a parabolic band, the <lb/>BCS formula for ξ o can be expressed in terms of the DOS <lb/>ρ(ǫ F ) in the following way: ξ o = vF <lb/>π∆o = <lb/>√ <lb/>3kF a 2 <lb/>2π 2 ρ(ǫF )∆o . Us-<lb/>ing ρ(ǫ F ) = 7.09 eV −1 extracted from specific heat and <lb/>assuming 2∆ o = 3.52 kT c , we obtain ξ o ≈ 27a ≈ 76Å, <lb/>0 <lb/>5 <lb/>10 <lb/>0.000 <lb/>0.002 <lb/>0.004 <lb/>0.006 <lb/>0 <lb/>5 <lb/>10 <lb/>0.000 <lb/>0.005 <lb/>0.010 <lb/>0.015 <lb/>0.020 <lb/>0.025 <lb/>R <lb/>ab ( ) <lb/>H (T) <lb/>H || ab <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>0 <lb/>2 <lb/>4 <lb/>6 <lb/>8 <lb/>10 <lb/>12 <lb/>H <lb/>c2 <lb/>(T) <lb/>T (K) <lb/>H ||c, <lb/>ab <lb/>H || ab, <lb/>ab <lb/>H || c, <lb/>c <lb/>R <lb/>ab ( ) <lb/>H (T) <lb/>H || c <lb/>(a) <lb/>(b) <lb/>FIG. 5: (a) The in-plane resistance curves R ab (H) measured <lb/>for temperatures ranging from 2.0 K to 4.2 K. Field orienta-<lb/>tions with H c and H ab (inset) are measured. (b) Hc2 vs <lb/>T for both field orientations. <lb/>in reasonable agreement with the measured value. <lb/>In conclusion, by combining a novel electrochemical <lb/>method and floating-zone crystal growth we have suc-<lb/>ceeded in producing high quality single crystals of the <lb/>hydrated Na x CoO 2 • yH 2 O system which show bulk su-<lb/>perconductivity. Our measurements indicate that the <lb/>low temperature properties are consistent with those of a <lb/>Fermi liquid with strong mass enhancement. The avail-<lb/>ability of high quality single crystals opens the door to <lb/>many microscopic probes (such as x-ray and neutron <lb/>scattering studies) which should help achieve an under-<lb/>standing of this strongly correlated material. <lb/></body>
22
+
23
+ <div type="acknowledgement">We thank B. Khaykovich, R. Ott, and R. Lang for as-<lb/>sistance with the experimental measurements. This re-<lb/>search was supported by the National Science Foundation <lb/>under its MRSEC Program Award No. 02-13282 and also <lb/>by Grant No. DMR 0239377. J.H.C. was partially sup-<lb/>ported by Korea Research Foundation Grant.(KRF-2002-<lb/>005-C20001). <lb/></div>
24
+
25
+ <front> † Permanent address: RCDAMP and Department of <lb/>Physics, Pusan National University, Pusan 609-735, Ko-<lb/>rea <lb/></front>
26
+
27
+ <page>5 <lb/></page>
28
+
29
+ <listBibl>[1] I. Terasaki, et al., Phys. Rev. B56, R12685 (1997). <lb/>[2] Y. Wang, et al., Nature 423, 425 (2003). <lb/>[3] Y. Wang, et al., cond-mat/0305455. <lb/>[4] K. Takada et al., Nature 422, 53 (2003). <lb/>[5] P.W. Anderson, Science 235, 1196 (1987). <lb/>[6] G. Baskaran, Phys. Rev. Lett. 91, 097003 (2003). <lb/>[7] B. Kumar and B.S. Shastry, cond-mat/0304210. <lb/>[8] Q.-H. Wang, et al., cond-mat/0304377. <lb/>[9] M. Foo et al. cond-mat/0304464. <lb/>[10] G. Cao et al., cond-mat/0305503. <lb/>[11] B.G. Ueland, et al., cond-mat/0307106. <lb/>[12] H.D. Yang, et al., cond-mat/0308031. <lb/>[13] R. Jin, et al., cond-mat/0306066. <lb/>[14] T. Motohashi, et al., Appl. Phys. Lett. 79 (2001) 1480. <lb/>[15] Q. Huang, J.W. Lynn, Y.S. Lee, et al., unpublished. <lb/>[16] W.L. Roth, J. Phys. Chem. Solids 25, 1 (1964). <lb/>[17] M. Lakner, et al., Phys. Rev. B 50, 17064 (1994). <lb/>[18] D.J. Singh, Phys. Rev. B61, 13397 (2000). </listBibl>
30
+
31
+
32
+ </text>
33
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L092227003-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>arXiv:cond-mat/0404196v1 [cond-mat.supr-con] 8 Apr 2004 <lb/>&quot;Spin-Flop&quot; Transition and Anisotropic Magnetoresistance in Pr 1.3−x La 0.7 Ce x CuO 4 : <lb/>Unexpectedly Strong Spin-Charge Coupling in Electron-Doped Cuprates <lb/>A. N. Lavrov, 1 H. J. Kang, 2, 3 Y. Kurita, 1, * T. Suzuki, 1, * Seiki Komiya, <lb/>J. W. Lynn, S.-H. Lee, 4 Pengcheng Dai, 2, 3, † and Yoichi Ando 1, ‡ <lb/>1 Central Research Institute of Electric Power Industry, Komae, Tokyo 201-8511, Japan <lb/>2 Department of Physics and Astronomy, The University of Tennessee, Knoxville, Tennessee 37996-1200, USA <lb/>3 Condensed Matter Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6393, USA <lb/>4 NIST Center for Neutron Research, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, USA <lb/>(Dated: November 5, 2018) <lb/>We use transport and neutron-scattering measurements to show that a magnetic-field-induced <lb/>transition from noncollinear to collinear spin arrangement in adjacent CuO2 planes of lightly <lb/>electron-doped Pr1.3−xLa0.7CexCuO4 (x = 0.01) crystals affects significantly both the in-plane <lb/>and out-of-plane resistivity. In the high-field collinear state, the magnetoresistance (MR) does not <lb/>saturate, but exhibits an intriguing four-fold-symmetric angular dependence, oscillating from being <lb/>positive at B [100] to being negative at B [110]. The observed MR of more than 30% at low tem-<lb/>peratures induced by a modest modification of the spin structure indicates an unexpectedly strong <lb/>spin-charge coupling in electron-doped cuprates. <lb/>PACS numbers: 74.25.Fy, 75.25.+z, 74.20.Mn, 74.72.Jt <lb/></front>
8
+
9
+ <body>High-T c superconductivity (SC) in cuprates emerges <lb/>as the parent antiferromagnetic (AF) insulator is doped <lb/>with charge carriers, either holes or electrons. Despite <lb/>this apparent symmetry with respect to doping, it still <lb/>remains unclear whether the mechanism of SC in both <lb/>cases is the same. It is generally believed that in the <lb/>hole-doped cuprates, the SC pairing originates from an <lb/>interplay between the doped holes and AF spin correla-<lb/>tions. Indeed, many observations, including a fast sup-<lb/>pression of the Néel order by doped holes [1] which results <lb/>in the &quot;spin-glass&quot; state [1, 2, 3], and a strong tendency to <lb/>form spin-charge textures -&quot;stripes&quot; [4], point to a very <lb/>strong coupling between the charge and spin degrees of <lb/>freedom. <lb/>The behavior of doped electrons looks, however, much <lb/>different. Electron doping suppresses the AF order at vir-<lb/>tually the same slow rate as the substitution of magnetic <lb/>Cu 2+ ions with non-magnetic Zn 2+ [5, 6], and does not <lb/>induce any incommensurability in the spin correlations <lb/>[7]. This has been taken as evidence that the electrons <lb/>merely dilute the spin system [1, 5, 6]. Apparently, if the <lb/>charge transport and spin correlations are actually de-<lb/>coupled in the electron-doped cuprates, the SC pairing <lb/>should have a non-magnetic origin as well. A recent dis-<lb/>covery of the magnetic-field induced AF order in super-<lb/>conducting Nd 2−x Ce x CuO 4 [8, 9] has shown, however, <lb/>that antiferromagnetism and superconductivity may be <lb/>closely related in these compounds. <lb/>To probe the spin-charge coupling, one can determine <lb/>how the charge transport responds to such relatively <lb/>weak changes in the spin structure as spin-flop or meta-<lb/>magnetic transitions. In hole-doped La 2−x Sr x CuO 4 , <lb/>for instance, the conductivity changes by up to several <lb/>times [10, 11]. In this Letter, we use neutron scat-<lb/>tering and magnetoresistance (MR) measurements to <lb/>study the effect of magnetic field on the spin structure <lb/>and anisotropic conductivity of lightly electron-doped <lb/>Pr 1.3−x La 0.7 Ce x CuO 4 (PLCCO) single crystals. We find <lb/>that both the in-plane and out-of-plane resistivity (ρ ab <lb/>and ρ c ) are surprisingly sensitive to spin reorientation, <lb/>with ∆ρ ab /ρ ab exceeding 30% at low temperatures -the <lb/>same scale as in hole-doped La 2−x Sr x CuO 4 [11]. This re-<lb/>sult indicates that in electron-doped cuprates the charge <lb/>transport exhibits a similar degree of coupling to mag-<lb/>netism as in the hole-doped ones, and therefore the su-<lb/>perconductivity in both systems may have a universal <lb/>origin. <lb/>High-quality PLCCO single crystals with x = 0.01 <lb/>(mosaicity &lt; 1 • ) were grown by the traveling-solvent <lb/>floating-zone technique and annealed at ≈ 860 • C in <lb/>pure argon to remove excess oxygen. The partial sub-<lb/>stitution of Pr with La was used to stabilize the crys-<lb/>tal growth, without introducing significant lattice distor-<lb/>tions [12]. Neutron scattering measurements were per-<lb/>formed on the BT-2 and SPINS triple-axis spectrome-<lb/>ters at the NIST Center for Neutron Research. We la-<lb/>bel wavevectors Q = (q x , q y , q z ) inÅ −1 as (H, K, L) = <lb/>(q x a/2π, q y a/2π, q z c/2π) in the reciprocal lattice units <lb/>(r.l.u.) suitable for the tetragonal unit cell of PLCCO <lb/>(space group I4/mmm, a = 3.964 and c = 12.28Å <lb/>are in-plane and out-of-plane lattice paramters, respec-<lb/>tively). In this notation, [100]/[010] and [110]/[110] are <lb/>along the Cu-O-Cu bond direction and the diagonal Cu-<lb/>Cu direction, respectively. The experimental details are <lb/>described in Refs. [8, 9]. <lb/>Resistivity measurements were carried out by the ac <lb/>four-probe method on the same crystal used for neu-<lb/>tron measurements. It was cut and polished into suitable <lb/>shapes: 3.1 × 1 × 0.45 mm 3 for ρ ab and ≈ 1 × 1 × 1 mm 3 <lb/>for ρ c . The MR was measured by sweeping the magnetic <lb/>
10
+
11
+ FIG. 1: Field-induced transition from noncollinear to col-<lb/>linear spin arrangement in Pr2CuO4. (a) Zero-filed non-<lb/>collinear spin structure; only Cu spins are shown. (b) -(c) <lb/>Collinear spin-flop states induced by (b) a magnetic field ap-<lb/>plied along the Cu-Cu direction; (c) a magnetic field tilted <lb/>from [010]; and (d) B [010]. <lb/>field between ±14 T at fixed temperatures stabilized by <lb/>a capacitance sensor with an accuracy of ∼ 1 mK. <lb/>The peculiar spin structure of Pr 2 CuO 4 (PCO) is in-<lb/>teresting in its own right. While a strong intraplane ex-<lb/>change drives the AF spin ordering within CuO 2 planes, <lb/>all the isotropic exchange interactions between the planes <lb/>are perfectly canceled out due to the body-centered <lb/>tetragonal crystal symmetry. The three-dimensional or-<lb/>dering [Fig. 1(a)] that sets in below the Néel temperature <lb/>T N = 250 − 285 K [13, 14, 16] is governed by weak pseu-<lb/>dodipolar (PD) interactions, which favor a noncollinear <lb/>orientation of spins in adjacent planes (alternating along <lb/>the [100] and [010] directions) [13, 14, 15, 16]. A unique <lb/>feature of the interplane PD interaction is that its energy <lb/>does not change if the spin sublattices of adjacent CuO <lb/>50 100 150 200 250 300 <lb/>0.0 <lb/>0.1 <lb/>0.2 <lb/>0.3 <lb/>0.4 <lb/>0.5 <lb/>0.6 <lb/>0.7 <lb/>0.8 <lb/>100 150 200 250 300 <lb/>0 <lb/>50 <lb/>100 <lb/>150 <lb/>200 <lb/>Pr 1.3-x La 0.7 Ce x CuO 4 <lb/>x = 0.01 <lb/>(1/2, 1/2, 1) <lb/>(a) <lb/>M Cu <lb/>M Pr <lb/>Moment (arb. units) <lb/>T (K) <lb/>(b) <lb/>(1/2, 1/2, 3) <lb/>Intensity (counts/15 sec) <lb/>T (K) <lb/>FIG. 2: (a) Temperature dependence of the Cu 2+ and Pr 3+ <lb/>moments in PLCCO (x = 0.01). (b) Integrated intensity <lb/>of the ( 1 <lb/>2 , 1 <lb/>2 , 1) and ( 1 <lb/>2 , 1 <lb/>2 , 3) magnetic peaks. The ordered <lb/>moments are estimated by normalizing the magnetic intensity <lb/>to the weak (1,1,0) nuclear Bragg peak without considering <lb/>the absorption and extinction effects [9]. The solid lines are <lb/>power-law fits describing the contribution of Cu spins [13]. <lb/>0.96 0.98 1.00 1.02 1.04 <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>1 <lb/>2 <lb/>0 <lb/>20 <lb/>40 <lb/>60 <lb/>80 <lb/>1.92 1.96 2.00 2.04 2.08 <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>30 <lb/>40 <lb/>50 <lb/>60 <lb/>70 <lb/>80 <lb/>_ <lb/>T = 5 K <lb/>(1/2, 1/2, 1) <lb/>(a) <lb/>B = <lb/>0.5 T <lb/>1.0 T <lb/>1.5 T <lb/>2.0 T <lb/>2.5 T <lb/>3.0 T <lb/>4.0 T <lb/>5.0 T <lb/>7.0 T <lb/>B // [110] <lb/>Intensity (10 <lb/>3 <lb/>counts/10 sec) <lb/>H (r.l.u.) <lb/>(c) <lb/>T = <lb/>K <lb/>30 K <lb/>41 K <lb/>56 K <lb/>75 K <lb/>100 K <lb/>150 K <lb/>(1/2, 1/2, 1) <lb/>Intensity (counts/10 sec) <lb/>B (T) <lb/>(b) <lb/>(1/2, 1/2, 2) <lb/>H (r.l.u.) <lb/>(d) (1/2, 1/2, 2) <lb/>B (T) <lb/>FIG. 3: (a), (b) Effect of the B [110] field on ( <lb/>2 , 1 <lb/>2 , 1) and <lb/>( 1 <lb/>2 , 1 <lb/>2 , 2) magnetic peaks at 5 K. (c), (d) Field dependence <lb/>of the integrated intensity at various temperatures. We note <lb/>that the critical field for spin-flop transition in PLCCO is <lb/>lower than that of PCO [13]. <lb/>planes rotate in opposite directions [14, 15, 16]. Such a <lb/>continuous spin rotation can be induced by a magnetic <lb/>field parallel to Cu-Cu direction, which easily converts <lb/>the noncollinear structure of Fig. 1(a) into a collinear <lb/>one with spins along the [110] direction [Fig. 1(b)]. Note <lb/>that while these diagonal directions are hard spin axes in <lb/>the non-collinear phase, they become the easy axes in the <lb/>collinear one. A perfectly aligned field B [010] causes a <lb/>first-order transition directly to the spin-flop phase [Fig. <lb/>1(d)], while at intermediate field directions the magnetic <lb/>field first induces a transition into the collinear phase <lb/>[Fig. 1(c)], and then smoothly rotates the spins to align <lb/>them perpendicular to the field [16]. <lb/>The neutron diffraction measurements at zero field on <lb/>the (1/2, 1/2, L) magnetic Bragg peaks (L = 0, 1, 2, 3, 4) <lb/>show that in our PLCCO (x = 0.01) the Cu 2+ spins <lb/>order into the same non-collinear structure as in pure <lb/>PCO, albeit at a somewhat lower T N ≈ 229 K (Fig. 2). <lb/>The reduced T N is probably due to a partial substitution <lb/>of Pr 3+ with non-magnetic La 3+ , as well as to doped <lb/>electrons. Similar to PCO [13], the Pr 3+ ions in PLCCO <lb/>can be polarized by the ordered Cu 2+ moment. Upon <lb/>cooling below 100-150 K, the exchange field of the Cu 2+ <lb/>spins induces a small (up to ∼ 0.1 µ B ) ordered moment <lb/>
12
+
13
+ 0 50 100 150 200 250 300 <lb/>0 <lb/>20 <lb/>40 <lb/>60 <lb/>80 <lb/>100 <lb/>120 <lb/>140 <lb/>-15 -10 -5 <lb/>0 <lb/>10 <lb/>15 <lb/>0 <lb/>2 <lb/>4 <lb/>6 <lb/>8 <lb/>10 <lb/>12 <lb/>14 <lb/>16 <lb/>-15 -10 -5 <lb/>0 <lb/>5 <lb/>10 <lb/>15 <lb/>0 <lb/>2 <lb/>4 <lb/>6 <lb/>8 <lb/>10 <lb/>ρ <lb/>c (Ωcm) <lb/>T (K) <lb/>0 <lb/>10 <lb/>20 <lb/>30 <lb/>40 <lb/>50 <lb/>60 <lb/>70 <lb/>ρ ab <lb/>ρ c <lb/>(a) <lb/>ρ <lb/>ab (mΩcm) <lb/>(b) <lb/>10 K <lb/>20 K <lb/>T = 5 K <lb/>74 K <lb/>30 K <lb/>Pr 1.3-x La 0.7 Ce x CuO 4 <lb/>x = 0.01 <lb/>B // Cu-Cu <lb/>40 K <lb/>∆ρ <lb/>c <lb/>/ρ <lb/>c (%) <lb/>B (T) <lb/>55 K <lb/>(c) <lb/>K <lb/>55 K <lb/>40 K <lb/>30 K <lb/>20 K <lb/>10 K <lb/>T = 5 K <lb/>B// Cu-Cu <lb/>∆ρ <lb/>ab <lb/>/ρ <lb/>ab (%) <lb/>B (T) <lb/>40 K 55 K 74 K <lb/>FIG. 4: (a) In-plane and out-of-plane resistivity of PLCCO (x = 0.01) single crystals. The MR in ρc (b) and ρ ab (c) measured <lb/>for the in-plane magnetic field B [110]. <lb/>on the Pr 3+ ions (Fig. 2). <lb/>Figure 3 shows the effect of a B [110] field on the <lb/>(1/2, 1/2, 1) and (1/2, 1/2, 2) magnetic peaks at vari-<lb/>ous temperatures. Upon increasing the magnetic field, <lb/>the peak intensity changes, indicating a continuous non-<lb/>collinear to collinear phase transition. Indeed, for the <lb/>collinear spin arrangement [Fig. 1(b)], the magnetic in-<lb/>tensity vanishes at (1/2, 1/2, L) with L = 1, 3, 5. As can <lb/>be seen in Figs. 3(c) and 3(d), the critical field for the <lb/>non-collinear to collinear (&quot;spin-flop&quot;) transition, B c , in-<lb/>creases from less than 0.5 T at 150 K to ∼2 T at 5 <lb/>K. In comparison, the first-order spin-flop transition for <lb/>B [010] was reported to take place at several time larger <lb/>fields [16] and a c-axis aligned field does not change the <lb/>noncollinear spin structure [9]. <lb/>The transport properties of lightly electron-doped <lb/>PLCCO differ from those of its hole-doped analog LSCO <lb/>or YBa 2 Cu 3 O 6+x (YBCO). In contrast to hole-doped <lb/>cuprates [11, 17, 18], the doping of 1% of electrons into <lb/>the CuO 2 planes appears to be insufficient to induce <lb/>metallic in-plane conduction in PLCCO, and both ρ ab <lb/>and ρ c grow upon cooling below room temperature [Fig. <lb/>4(a)]. It is worth noting also that lightly doped PLCCO <lb/>turns out to be one of the most anisotropic cuprates with <lb/>ρ c /ρ ab ∼ 8000 at room temperature -an order of mag-<lb/>nitude larger than in LSCO and YBCO [11, 18]. <lb/>In further contrast to hole-doped cuprates [11, 18], no <lb/>anomaly is detected at the Néel transition in PLCCO ei-<lb/>ther in the in-plane or out-of-plane resistivity. At a first <lb/>glance, this supports the view that the charge motion in <lb/>electron-doped PLCCO is virtually decoupled from spin <lb/>correlations, and one therefore would expect the conduc-<lb/>tivity to ignore the spin reorientation sketched in Fig. <lb/>1. Surprisingly, the experiment shows that this is not <lb/>the case, and instead of being field-independent, both <lb/>ρ ab and ρ c exhibit a considerable increase upon transi-<lb/>tion into the collinear state [Figs. 4(b) and 4(c)]. We <lb/>have confirmed that this MR is of the spin origin and <lb/>contains no orbital terms, since no difference was ob-<lb/>served in ∆ρ ab /ρ ab for fields applied parallel or perpen-<lb/>dicular to the current. Moreover, ∆ρ ab /ρ ab and ∆ρ c /ρ c <lb/>demonstrate a remarkable similarity both in magnitude <lb/>and in field dependence, in spite of the huge resistivity <lb/>anisotropy. Finally, no MR anomaly is observed when a <lb/>c-axis aligned field is applied, consistent with the absence <lb/>of a spin-flop transition for such field orientation [9]. <lb/>The MR behavior in Fig. 4 is clearly reminiscent of <lb/>that in LSCO [11], though there are two important dif-<lb/>ferences. First is the sign of the anomalous MR, which is <lb/>always positive in PLCCO, but negative in LSCO. Sec-<lb/>ond, the MR features in LSCO and YBCO become dis-<lb/>cernible as soon as the AF order is established, but in <lb/>PLCCO they appear at temperatures much lower that <lb/>T N (at T &lt; 70 − 100 K), and quickly gain strength upon <lb/>decreasing temperature (Fig. 4). The latter indicates <lb/>that some other factors, such as magnetic moments of <lb/>Pr 3+ or a structural instability [19], that come into play <lb/>at low temperature, may be relevant to the observed MR. <lb/>A comparison of the neutron and resistivity data re-<lb/>veals one more interesting feature, namely, the transi-<lb/>tions observed by these two probes do not match each <lb/>other [inset to Fig. 5(a)]. One can see that the charge <lb/>transport ignores the initial spin rotation, and the steep-<lb/>est resistivity variation is observed at B c , where the <lb/>collinear structure is established. Although B c changes <lb/>substantially with temperature [Fig. 5(a)], the apparent <lb/>shift in the transitions holds consistently, with the peak <lb/>in dρ/dB roughly coinciding with the end of the transi-<lb/>tion observed by neutron scattering. <lb/>As the magnetic field deviates from the Cu-Cu direc-<lb/>tion [Fig. 1(c)], the spin-flop transition shifts towards <lb/>higher fields, reaching ultimately B c ∼ 12 T for B [010]; <lb/>the MR behavior for these two field orientations is com-<lb/>pared in Figs. 5(b) and 5(c) [20]. It becomes immediately <lb/>clear from these figures that the step-like increase of the <lb/>resistivity upon the transition to the collinear state does <lb/></body>
14
+
15
+ <page>4 <lb/></page>
16
+
17
+ <body>0 <lb/>20 <lb/>40 <lb/>60 <lb/>80 <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>-12 -8 <lb/>-4 <lb/>0 <lb/>4 <lb/>8 <lb/>12 <lb/>0 <lb/>2 <lb/>4 <lb/>6 <lb/>8 <lb/>10 <lb/>12 <lb/>14 <lb/>16 <lb/>18 <lb/>-12 -8 <lb/>-4 <lb/>0 <lb/>4 <lb/>12 <lb/>0 <lb/>5 <lb/>10 <lb/>15 <lb/>20 <lb/>25 <lb/>30 <lb/>1 <lb/>2 <lb/>0.0 <lb/>0.4 <lb/>0.8 <lb/>(a) <lb/>ρ ab <lb/>ρ c <lb/>B // Cu-Cu <lb/>B <lb/>c (T) <lb/>T (K) <lb/>* <lb/>B c <lb/>* <lb/>B c <lb/>(b) <lb/>B // Cu-O-Cu <lb/>T = 5 K <lb/>∆ρ <lb/>ab <lb/>/ρ <lb/>ab (%) <lb/>B (T) <lb/>* <lb/>B c <lb/>(c) <lb/>T=T <lb/>B // Cu-O-Cu <lb/>B // Cu-Cu <lb/>B // Cu-Cu <lb/>T = 5 K <lb/>∆ρ <lb/>c <lb/>/ρ <lb/>c (%) <lb/>B (T) <lb/>30 <lb/>60 <lb/>90 <lb/>120 <lb/>150 <lb/>180 <lb/>210 <lb/>240 270 <lb/>300 <lb/>330 <lb/>+ <lb/>-<lb/>+ <lb/>-<lb/>+ <lb/>-<lb/>+ <lb/>-<lb/>T = 30 K <lb/>∆ρ <lb/>c , <lb/>∆I <lb/>B (T) <lb/>FIG. 5: (a) The critical field Bc determined from peaks in dρ ab /dB and dρc/dB for B [110]. In the inset to (a), the <lb/>normalized field dependence of ρc (•) is compared with that of the ( 1 <lb/>2 , 1 <lb/>2 , 1)-peak intensity (•). ∆ρ ab /ρ ab (b) and ∆ρc/ρc (c) <lb/>for two directions of the in-plane magnetic field. The angular dependence of the high-field MR is sketched in the inset to (c). <lb/>not make a complete story. Regardless of the field direc-<lb/>tion within the ab plane, the resistivity exhibits roughly <lb/>the same increase at the spin-flop transition, but then (at <lb/>B &gt; B * <lb/>c ) it keeps changing without any sign of satura-<lb/>tion [Figs. 5(b) and 5(c)]. Even more surprising is that <lb/>this high-field MR changes its sign depending on the field <lb/>direction, as is schematically drawn in the inset to Fig. <lb/>5(c). One can conceive a spin structure upon rotating <lb/>the high magnetic field within the ab plane in the follow-<lb/>ing way: the spins always keep the collinear arrangement <lb/>and rotate as a whole, being almost perpendicular to the <lb/>magnetic field (Fig. 1). Our data show that the resistiv-<lb/>ity goes down as the spin direction approaches one of the <lb/>two equivalent spin easy axes (Cu-Cu directions) and in-<lb/>creases at the spin hard axes (Cu-O-Cu directions) [inset <lb/>to Fig. 5(c)]. Note that the resistivity changes are rather <lb/>large, ∆ρ ab /ρ ab reaches ≈ 18% at T = 5 K and exceeds <lb/>32% at 2.5 K, indicating that the magnetic field B [100] <lb/>can effectively localize the doped electrons. <lb/>Apparently, the fascinating MR oscillations in Fig. 5 <lb/>cannot originate from simple &quot;spin-valve&quot; effects, since <lb/>at high fields the spin structure always stays collinear, <lb/>and all that changes is the relative orientation of spins <lb/>with respect to the crystal axes. The MR may be related <lb/>to 2D spin fluctuations that were found to survive far <lb/>above B c , as manifested in the diffuse neutron scattering <lb/>[21], or to some unusual coupling of the charge trans-<lb/>port with low-energy spin dynamics. Though the exact <lb/>mechanism of the revealed MR features still remains to <lb/>be understood, what is certain is that the charge carriers <lb/>in electron-doped cuprates appear to have a remarkably <lb/>strong coupling with the spin order, which should play an <lb/>important role in determining their physical properties. <lb/>Upon preparing this paper, we became aware of similar <lb/>MR features observed for Pr 1.85 Ce 0.15 CuO 4 [22], which <lb/>gives evidence that the strong spin-charge coupling sur-<lb/>vives up to much higher electron-doping levels, that are <lb/>relevant for the superconducting state. <lb/></body>
18
+
19
+ <div type="acknowledgement">We thank K. Segawa and Shiliang Li for technical as-<lb/>sistance. This work was in part supported by the US <lb/>NSF DMR-0139882 and DOE under contract No. DE-<lb/>AC-00OR22725 with UT/Battelle, LLC. <lb/></div>
20
+
21
+ <front> * Also at Department of Physics, Tokyo University of Sci-<lb/>ence, Shinjuku-ku, Tokyo 162-8601, Japan. <lb/> † Electronic address: [email protected] <lb/> ‡ Electronic address: [email protected] <lb/></front>
22
+
23
+ <listBibl>[1] M. A. Kastner, R. J. Birgeneau, G. Shirane, and Y. En-<lb/>doh, Rev. Mod. Phys. 70, 897 (1998). <lb/>[2] Ch. Niedermayer et al., Phys. Rev. Lett. 80, 3843 (1998). <lb/>[3] A. N. Lavrov et al., Phys. Rev. Lett. 87, 017007 (2001). <lb/>[4] J. M. Tranquada et al., Nature 375, 561 (1995). <lb/>[5] B. Keimer et al., Phys. Rev. B 45, 7430 (1992). <lb/>[6] G. M. Luke et al., Phys. Rev. B 42, 7981 (1990). <lb/>[7] K. Yamada et al., Phys. Rev. Lett. 90, 137004 (2003). <lb/>[8] H. J. Kang et al., Nature 423, 522 (2003). <lb/>[9] M. Matsuura et al., Phys. Rev. B 68, 144503 (2003). <lb/>[10] T. Thio et al., Phys. Rev. B 38, 905 (1988). <lb/>[11] Y. Ando et al., Phys. Rev. Lett. 90, 247003 (2003). <lb/>[12] M. Fujita et al., Phys. Rev. B 67, 014514 (2003). <lb/>[13] I. W. Sumarlin et al., Phys. Rev. B 51, 5824 (1995). <lb/>[14] D. Petitgrand et al., Phys. Rev. B 59, 1079 (1999). <lb/>[15] R. Sachidanandam et al., Phys. Rev. B 56, 260 (1997). <lb/>[16] V. P. Plakhty et al., Europhys. Lett. 61, 534 (2003). <lb/>[17] Y. Ando et al., Phys. Rev. Lett. 87, 017001 (2001). <lb/>[18] A. N. Lavrov et al., Phys. Rev. Lett. 83 (1999) 1419. <lb/>[19] V. P. Plakhty et al., Solid State Commun. 103, 683 <lb/>(1997). <lb/>[20] For the case of B [010], the transition field is extremely <lb/>sensitive to the crystal alignment; just a few degrees tilt-<lb/>ing of the ρ ab sample in Fig. 5(b) notably shifted the tran-<lb/>sition to low fields, as compared with the better aligned <lb/>ρc sample in Fig. 5(c). <lb/>[21] D. Petigrand et al., App. Phys. A. 74, S853 (2002). <lb/>[22] P. Fournier et al., cond-mat/0309144. </listBibl>
24
+
25
+
26
+ </text>
27
+ </tei>
grobid-trainer/resources/dataset/segmentation/article/light/corpus/tei/L093156802-CC.training.segmentation.tei.xml ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0" ?>
2
+ <tei xml:space="preserve">
3
+ <teiHeader>
4
+ <fileDesc xml:id="_0"/>
5
+ </teiHeader>
6
+ <text xml:lang="en">
7
+ <front>arXiv:0708.0921v2 [cond-mat.supr-con] 8 May 2008 <lb/>Evidence for gap anisotropy in CaC 6 from directional point-contact spectroscopy <lb/>R.S. Gonnelli, 1 D. Daghero, 1 D. Delaude, 1 M. Tortello, 1 G.A. Ummarino, <lb/>V.A. Stepanov, 2 J.S. Kim, 3 R.K. Kremer, 3 A. Sanna, 4 G. Profeta, 5 and S. Massidda <lb/>1 Dipartimento di Fisica and CNISM, Politecnico di Torino, 10129 Torino, Italy <lb/>2 P.N. Lebedev Physical Institute, Russian Academy of Sciences, 119991 Moscow, Russia <lb/>3 Max-Planck-Institut für Festkörperforschung, D-70569 Stuttgart, Germany <lb/>4 Institut für Theoretische Physik, Freie Universität Berlin, Arnimallee 14, D-14195 Berlin, Germany <lb/>5 CNISM -Dipartimento di Fisica, Università degli studi dell&apos;Aquila, Italy <lb/>6 SLACS-INFM/CNR and Dipartimento di Fisica, Università degli Studi di Cagliari, Italy <lb/>We present the first results of directional point-contact spectroscopy in high quality CaC6 samples <lb/>both along the ab plane and in the c-axis direction. The superconducting order parameter ∆(0), <lb/>obtained by fitting the Andreev-reflection (AR) conductance curves at temperatures down to 400 <lb/>mK with the single-band 3D Blonder-Tinkham-Klapwijk model, presents two different distributions <lb/>in the two directions of the main current injection, peaked at 1.35 and 1.71 meV, respectively. By <lb/>ab-initio calculations of the AR conductance spectra, we show that the experimental results are in <lb/>good agreement with the recent predictions of gap anisotropy in CaC6. <lb/>PACS numbers: 74.50.+r, 74.45.+c, 74.70.Ad <lb/></front>
8
+
9
+ <body>The discovery of a relatively &quot;high-T c &quot; superconduc-<lb/>tivity in graphite intercalated with Ca [1, 2], Yb [1] and, <lb/>very recently, Sr [3, 4] has strongly revived the inter-<lb/>est in the Graphite Intercalated Compounds (GICs) and <lb/>their physics. The Ca-intercalated graphite, CaC 6 , with <lb/>its &quot;record&quot; T c of about 11.5 K, in particular, has been <lb/>the subject of various theoretical and experimental in-<lb/>vestigations in the past two years (for a short review of <lb/>the initial results see [5]). One of the most important <lb/>questions, however, is still not clear: what is the mag-<lb/>nitude and anisotropy of its superconducting gap? The <lb/>first experiments (STM, penetration depth, specific heat) <lb/>on CaC 6 have evidenced a single, apparently isotropic, s-<lb/>wave gap with a ratio 2∆/k B T c of the order of the BCS <lb/>value [6, 7, 8]. Recent tunnel spectroscopy results, on the <lb/>other hand, claimed the presence of an isotropic gap with <lb/>a magnitude more than 40% higher than that reported <lb/>earlier [9]. The spread of gap values measured up to now <lb/>range between 1.6 meV [6] and 2.3 meV [9]. The im-<lb/>portant point is that all these experiments have either <lb/>probed a bulk property [8] or a directional one along <lb/>the c-axis direction [6, 7, 9]. As pointed out in Ref.6, <lb/>the presence of anisotropic or two-gap superconductivity <lb/>in CaC 6 cannot be ruled out until tunneling or point-<lb/>contact measurements are performed also along the ab <lb/>direction. On the other hand, recent first-principles den-<lb/>sity functional calculations of the superconducting prop-<lb/>erties of CaC 6 have supported the presence of a moder-<lb/>ately anisotropic gap which varies between 1.1 and 2.3 <lb/>meV, depending on the k-point and the π or interlayer <lb/>(IL) sheet of the Fermi surface (FS) involved [10]. Such <lb/>an anisotropy can be revealed by directional spectroscopy <lb/>measurements performed along both c and ab direction. <lb/>In this paper we present the results of point-contact <lb/>Andreev-reflection (PCAR) spectroscopy performed on <lb/>high-quality bulk samples of CaC 6 [8]. By using a spe-<lb/>cial technique to realize the contacts, that proved very <lb/>successful and effective in the case of MgB 2 [11, 12], we <lb/>were able to perform directional PCAR spectroscopy at <lb/>very low temperature both along the ab plane and the <lb/>c-axis direction. Two different gap distributions in the <lb/>two directions can reproducibly be extracted from the ex-<lb/>perimental data. When compared to the results of new <lb/>first-principles calculations these findings unequivocally <lb/>prove the anisotropy of the superconducting gap in CaC 6 . <lb/>The high-quality CaC 6 bulk samples used for our mea-<lb/>surements were synthesized by reacting highly oriented <lb/>pyrolytic graphite (with a spread of the c axis orienta-<lb/>tion ≤ 0.4 • ) for several weeks at 350 • C with a molten <lb/>alloy of Li and Ca [8]. The resulting CaC 6 samples have <lb/>a shiny golden surface. They are very sensitive to air and <lb/>moisture which rapidly damage the sample surfaces. X-<lb/>ray analysis has shown mainly the CaC 6 reflections with <lb/>a small (&lt; 5%) contribution from impurity phases. Fur-<lb/>ther details on the characterization of the samples may be <lb/>found in Ref. 8. All samples used for PCAR spectroscopy <lb/>(size ≈ 1×1×0.2 mm 3 ) were selected to have a very sharp <lb/>superconducting transition (∆T c (10% − 90%) = 0.1 K) <lb/>with the onset at T c = 11.4 K. <lb/>The point contacts were made by using a non-<lb/>conventional technique we called &quot;soft&quot; PCAR spec-<lb/>troscopy [11, 12]. Instead of using the standard metallic <lb/>tip, a very small (∅ ≃ 50 µm) drop of Ag conductive <lb/>paint, put on the etched or freshly cleaved surfaces of <lb/>the sample is used as a counterelectrode. Such contacts <lb/>are particularly stable both in time and towards tem-<lb/>perature variations and they allow to inject the current <lb/>mainly perpendicular to the contact plane. A fine tuning <lb/>of the junction characteristics at low temperature can be <lb/>done by applying short voltage or current pulses. Fur-<lb/>ther details on the technique can be found in Refs. 11, 13. <lb/>Due to the mentioned high sensitivity of CaC 6 samples&apos; <lb/></body>
10
+
11
+ <page>2 <lb/></page>
12
+
13
+ <body>-10 <lb/>-5 <lb/>0 <lb/>5 <lb/>10 <lb/>0.95 <lb/>1.00 <lb/>1.05 <lb/>1.10 <lb/>¡ <lb/>¢ <lb/>£ <lb/>¤ <lb/>¥ <lb/>¦ <lb/> § <lb/>ab plane <lb/>(a) <lb/>-0.30 Ω <lb/>−1 <lb/>+ 0.41 Ω <lb/>−1 <lb/>+ 0.31 Ω <lb/>−1 <lb/>Conductance (Ω <lb/>-1 <lb/>) <lb/>Voltage (mV) <lb/>-10 <lb/>-5 <lb/>0 <lb/>5 <lb/>10 <lb/>0.50 <lb/>0.55 <lb/>0.60 <lb/>0.65 <lb/>0.70 <lb/>0.75 <lb/>0.80 <lb/>0.85© <lb/>c axis <lb/>(b) <lb/>-0.02 Ω <lb/>−1 <lb/>-0.03 Ω <lb/>−1 <lb/>-0.13 Ω <lb/>−1 <lb/>Conductance (Ω <lb/>-1 <lb/>) <lb/>Voltage (mV) <lb/>FIG. 1: (Color online) (a) Raw point-contact conductance <lb/>curves of various ab-plane contacts at 4.2 K. For clarity the <lb/>curves are vertically shifted of the amount shown close to each <lb/>curve. (b) The same as in (a) but for various contacts with <lb/>current injection mainly along the c axis. In each panel a <lb/>sketch of the contact geometry is also shown. <lb/>surface to air, the room-temperature preparation of the <lb/>contact was done inside a sealed glove bag filled with <lb/>pure He gas or in a glove box with Ar atmosphere. Af-<lb/>ter the contact was made the junction was very rapidly <lb/>transferred to the cryostat in a sealed container. Con-<lb/>tacts were made either on the flat ab-plane surface or on <lb/>the narrow lateral side of the samples. Referring to the <lb/>main direction of current injection, we call them c-axis <lb/>and ab-plane contacts, respectively (see insets of Fig. 1). <lb/>The conductance curves, dI/dV vs. V , were obtained <lb/>by numerical differentiation of the measured I −V curves <lb/>and subsequently normalized by dividing them by the <lb/>normal-state conductance measured at T ≥ T c . For this <lb/>reason, in all the contacts, we therefore carefully studied <lb/>the temperature dependence of the conductance in order <lb/>to determine the critical temperature of the junction, i.e. <lb/>the &apos;Andreev critical temperature&apos;, T A <lb/>c . In an overall of <lb/>different contacts, T A <lb/>c was found to be 11.3 ± 0.1 K, in <lb/>best agreement with the bulk T c &apos;s of the samples and in <lb/>contrast with a previous report [6]. This fact proves the <lb/>high quality of samples and surfaces in the contact re-<lb/>gion. For simplicity, we will therefore refer to the critical <lb/>temperatures of the contacts as T c in the following. <lb/>Fig. 1 shows several raw conductance curves as func-<lb/>tion of bias voltage measured both in ab-plane contacts <lb/>(a) and in c-axis ones (b) at 4.2 K. The curves show clear <lb/>Andreev-reflection (AR) features, an almost flat conduc-<lb/>tance (at V &gt; 8 − 10 meV) and no dips that usually <lb/>are a sign of the failure in reaching the conditions for <lb/>pure ballistic conduction in the contact [14, 15]. The <lb/>normal resistance R N of all the good contacts is between <lb/>0.75 and 6.4 Ω. By knowing the mean free paths and <lb/>the residual resistivities of CaC 6 along the ab plane and <lb/>in the c-axis direction, i.e. ℓ ab = 74 nm, ℓ c = 4.7 nm, <lb/>ρ 0,ab = 0.8 µΩ•cm and ρ 0,c = 24 µΩ•cm [16, 17] we can <lb/>apply the Sharvin formula for the contact resistance in <lb/>the ballistic regime in order to determine the contact <lb/>radius a = (4ρ 0 ℓ/3πR N ) 0.5 [15]. The condition for full <lb/>ballistic transport (a ≪ ℓ) is totally verified in ab-plane <lb/>contacts, where a ab ≈ 6 − 18 nm. In c-axis junctions, <lb/>where a c ≈ 14 − 24 nm but the conductance curves do <lb/>not show any sign of heating [13], the presence of at least <lb/>30 parallel contacts in the junction area is expected. <lb/>After normalization, the conductance curves were <lb/>fitted to the modified 3D Blonder-Tinkham-Klapwijk <lb/>(BTK) model [18, 19, 20]. In the single-band form it con-<lb/>tains three fitting parameters: The gap ∆, the barrier-<lb/>height parameter Z and the broadening Γ which accounts <lb/>for both intrinsic (quasiparticle lifetime) and extrinsic <lb/>phenomena that broaden the AR conductance [19]. <lb/>In order to increase the experimental resolution of our <lb/>measurements we decided to perform part of the PCAR <lb/>experiments at very low temperature in a Quantum De-<lb/>sign measurement system (PPMS) with 3 He insert. <lb/>Fig. 2 (a) shows the normalized conductance curves <lb/>(circles) of a typical ab-plane contact at various temper-<lb/>atures from 400 mK up to T c . At any temperature the <lb/>single-band 3D BTK model fits the data very well (solid <lb/>lines). At the lowest T , the values of the fitting param-<lb/>eters are: ∆ = 1.44 meV, Γ = 0.61 meV and Z = 0.75. <lb/>In panel (c) we display the order parameter ∆ obtained <lb/>from the data given in (a). Its temperature dependence <lb/>almost perfectly follows the BCS-like expression (solid <lb/>line) with 2∆(0)/k B T c = 2.98 which is sensibly smaller <lb/>than expected from BCS theory. <lb/>In Fig. 2 (b) and (d) we report the same data for a <lb/>c-axis contact. As for the ab-plane case, the curves are <lb/>well fitted by the single-band 3D BTK model which gives <lb/>at 400 mK: ∆ = 1.7 meV, Γ = 0.84 meV and Z = 0.97. <lb/>The temperature dependence of ∆ is very close to the <lb/>expected BCS one with a ratio 2∆(0)/k B T c = 3.48, in <lb/>best agreement with the weak-coupling BCS value. <lb/>It is worth noticing that the Z values observed in c-<lb/>axis contacts (between 0.74 and 1.01) are systematically <lb/>greater than those of ab-plane contacts (between 0.48 <lb/>and 0.75). According to the 3D BTK model [20], this <lb/>difference can be explained by the different Fermi veloc-<lb/>ities of CaC in the ab plane (v ab = 0.54 × 10 6 m/s) and <lb/>along c axis (v c = 0.29 × 10 6 m/s), thus confirming the <lb/>directionality of our point contacts. <lb/>The AR curves shown in Fig. 1 and 2 are rather <lb/>small in amplitude, as already observed in all the &quot;soft&quot; <lb/>PCAR measurements on MgB 2 and related compounds <lb/>[11, 12, 13], resulting in Γ values substantially greater <lb/>than those expected for the quasiparticle lifetime. As re-<lb/>cently observed in lithographically fabricated Cu-Pt-Pb <lb/>nanocontacts [21], this additional broadening can be ex-<lb/>plained by the presence of pair-breaking effects induced <lb/>by the scattering in a thin disordered layer present at the <lb/>NS interface. This is the case of our point contacts, due <lb/>to a disordered layer on the surface of Ag grains that also <lb/>makes the residual resistivity of the paint be five orders <lb/>of magnitude greater than in pure Ag. <lb/>The reproducibility of the PCAR data was very good. <lb/>Most of the contacts, obtained both in 4 He and in 3 He <lb/></body>
14
+
15
+ <page>3 <lb/></page>
16
+
17
+ <body>0 <lb/>2 <lb/>4 <lb/>6 <lb/>8 10 12 <lb/>0.0 <lb/>0.5 <lb/>1.0 <lb/>1.5 <lb/>2.0 <lb/>-10 <lb/>-5 <lb/>0 <lb/>5 <lb/>10 <lb/>0.8 <lb/>0.9 <lb/>1.0 <lb/>1.1 <lb/>1.2 (a) <lb/>Order parameter <lb/>∆ (meV) <lb/>Normalized conductance <lb/>T (K) <lb/>0.4 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>7 <lb/>8 <lb/>9 <lb/>10 <lb/>10.5 <lb/>11 <lb/>ab-plane <lb/>Normalized conductance <lb/>Voltage (mV) <lb/>0 <lb/>2 <lb/>4 <lb/>6 <lb/>8 10 12 <lb/>0.0 <lb/>0.5 <lb/>1.0 <lb/>1.5 <lb/>(c) <lb/>2∆/k B T C =2.98 <lb/>ab-plane <lb/>BCS fit <lb/>-10 <lb/>-5 <lb/>0 <lb/>5 <lb/>10 <lb/>0.8 <lb/>0.9 <lb/>1.0 <lb/>1.1 <lb/>1.2 <lb/>(b) <lb/>Temperature (K) <lb/>Temperature (K) <lb/>T (K) <lb/>0.4 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>6 <lb/>7 <lb/>8 <lb/>9 <lb/>10 <lb/>11 <lb/>c-axis <lb/>Voltage (mV) <lb/>(d) <lb/>2∆/k B T C =3.48 <lb/>c-axis <lb/>BCS fit <lb/>FIG. 2: (Color online) Normalized dI/dV vs. V curves at different temperatures down to 400 mK in an ab-plane contact (a) <lb/>and in a c-axis one (b) (open circles). Solid lines: best-fit curves given by the single-band 3D BTK model. Panels (c) and <lb/>(d) show the temperature dependency of the order parameter ∆ (full circles) in the ab-plane direction and in the c-axis one, <lb/>respectively, as determined from the BTK fits shown in (a) and (b). Solid lines are the BCS-like fits. <lb/>cryostat, show dI/dV curves and temperature dependen-<lb/>cies of quality similar to that presented in Fig. 2. In <lb/>15 ab-plane contacts the order parameter ∆(0) ranged <lb/>between 1.1 meV and 1.7 meV with the distribution <lb/>shown in Fig. (a). In 14 c-axis contacts ∆(0) ranged <lb/>between 1.3 meV and 1.94 meV with the distribution <lb/>shown in Fig. 3 (b). The figure also shows the Gaussian <lb/>curves that best fit the distributions. They are peaked at <lb/>∆ ab (0)=1.35 meV and ∆ c (0)=1.71 meV and show stan-<lb/>dard deviations s ab =0.14 meV and s c =0.08 meV, respec-<lb/>tively. The results in c-axis direction are in very good <lb/>agreement with the gap values previously reported in <lb/>Ref. 6, 7. A minority of contacts (3 in ab-plane and <lb/>3 in c-axis direction) have shown gap values between 2.1 <lb/>and 2.4 meV, similarly to the results of Ref. 9. <lb/>The complex microscopical nature of our point con-<lb/>tacts leaves some uncertainty about the true direction <lb/>of current injection, particularly in the case of contacts <lb/>on the side faces of the sample (i.e. ab-plane contacts) <lb/>where, due to the intrinsic inhomogeneity of the cleaved <lb/>surface, current injection along c axis is also possible. <lb/>1.0 <lb/>1.2 <lb/>1.4 <lb/>1.6 <lb/>1.8 <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>(a) <lb/>Frequency Counts <lb/>Order parameter ∆ (meV) <lb/>ab plane <lb/>1.3 <lb/>1.5 <lb/>1.7 <lb/>1.9 <lb/>0 <lb/>1 <lb/>2 <lb/>3 <lb/>4 <lb/>5 <lb/>6 <lb/>c axis <lb/>(b) <lb/>Frequency Counts <lb/>Order parameter ∆ (meV) <lb/>FIG. 3: (Color online) Distributions of the different ∆(0) val-<lb/>ues measured in ab-plane contacts (a) and in c-axis ones (b) <lb/>at 4.2 K (red) and at 400 mK (light red). Dash black lines <lb/>are the fits of the total distribution to a Gaussian curve. <lb/>However, the clear difference observed between the most <lb/>probable ∆(0) values in ab-plane and c-axis contacts pro-<lb/>vides strong evidence for a gap anisotropy in CaC 6 . <lb/>In order to compare our results with the theoretical <lb/>predictions of gap anisotropy in CaC 6 [10] we calcu-<lb/>lated the Andreev-reflection conductance curves by first-<lb/>principles methods. We have a SN junction where S = <lb/>CaC 6 and N = Ag. Let&apos;s label with the suffix i = 1,2,3 <lb/>the three sheets of the CaC 6 Fermi surface (FS) (π and <lb/>interlayer (IL) bands). If n is the unitary vector in the <lb/>direction of the injected current, v ik,n = v ik • n are the <lb/>corresponding components of the Fermi velocities in the <lb/>superconductor at wave vector k for band i-th. Tak-<lb/>ing into account that Ag has a quasi-spherical FS and <lb/>an almost constant Fermi velocity v N = v ik , the corre-<lb/>sponding quantity in the normal metal will be v N,n = v N . <lb/>Following Refs. 22, 23 we finally obtain the total AR con-<lb/>ductance as: <lb/>σ(E, n) = <lb/>i σ ikn (E) <lb/>v 2 <lb/>ik,n <lb/>v ik [v ik,n +vN] 2 FSi <lb/>i <lb/>v 2 <lb/>ik,n <lb/>v ik [v ik,n +vN] 2 FSi <lb/>(1) <lb/>where: FSi is the integral over the i-th FS, i.e. <lb/>v ik,n <lb/>2 <lb/>v ik [v ik,n +vN] 2 FSi = v ik,n &gt;0 <lb/>v 2 <lb/>ik,n <lb/>[v ik,n +vN] 2 δ(E ik )d 3 k. <lb/>σ ikn (E) is the BTK conductance of the i-th band ex-<lb/>pressed in terms of the quantities N q <lb/>ik (E) = E/ <lb/>p <lb/>E 2 <lb/>− ∆ 2 <lb/>ik <lb/>and N p <lb/>ik (E) = ∆ ik / <lb/>p <lb/>E 2 <lb/>− ∆ 2 <lb/>ik (whose real parts are the <lb/>quasiparticle and the pair density of states in the same <lb/>band, respectively) and of Z n values. ∆ ik is the gap value <lb/>for band i-th at point k over the FS, recently calculated <lb/>from first principles [10]. The values of Z n used in the <lb/>calculation are taken similar to those of the curves shown <lb/></body>
18
+
19
+ <page>4 <lb/></page>
20
+
21
+ <body>-6 <lb/>-4 <lb/>-2 <lb/>0 <lb/>2 <lb/>4 <lb/>6 <lb/>0.8 <lb/>1.0 <lb/>1.2 <lb/>1.4 <lb/>1.6 <lb/>1.8 <lb/>2.0 <lb/>2.2 <lb/>2.4 <lb/>T = 0.4 K <lb/>T = 0 K <lb/>T = 0 K <lb/>a axis <lb/>(a) <lb/>exp <lb/>Γ = 0.6 meV <lb/>Γ = 0 <lb/>Normalized conductance <lb/>Voltage (mV) <lb/>BTK fit <lb/>-6 <lb/>-4 <lb/>-2 <lb/>0 <lb/>2 <lb/>4 <lb/>6 <lb/>0.8 <lb/>1.0 <lb/>1.2 <lb/>1.4 <lb/>1.6 <lb/>1.8 <lb/>2.0 <lb/>2.2 <lb/>2.4 <lb/>Γ = 0 <lb/>Γ = 0.8 meV <lb/>exp <lb/>T = 0.4 K <lb/>T = 0 K <lb/>T = 0 K <lb/>c axis <lb/>(b) <lb/>Normalized conductance <lb/>Voltage (mV) <lb/>BTK fit <lb/>FIG. 4: (Color online) Theoretical AR conductances calcu-<lb/>lated at T = 0 by Eq. (1). (a) Current injected along the <lb/>a axis with Z = 0.75 and Γ = 0 (black) and Γ = 0.6 (red); <lb/>(b) current injected along the c axis with Z = 1 and Γ = 0 <lb/>(black) and Γ = 0.8 (red). Experimental curves at 400 mK <lb/>are shown for comparison (blue circles). <lb/>in Fig. 2, i.e. Z ab = 0.75 and Z c = 1. The explicit ex-<lb/>pression of σ ikn (E) can be found in Ref. 20. <lb/>The results of the calculations given by Eq. (1) are <lb/>shown at T = 0 K and for a-and c-axis directions (top <lb/>curves in Fig. 4 (a) and (b)). The conductance calcu-<lb/>lated along the b direction is almost identical to the one <lb/>in a direction. At T = 0 the topology of the CaC 6 FS <lb/>and the calculated anisotropy of the π and IL gaps re-<lb/>sult in a sizeable anisotropy of the AR conductance. In <lb/>ab direction, it exhibits a sharp peak (related to the π <lb/>gap) at about 1.38 meV and a broad shoulder (mainly <lb/>related to the IL gap) at about 1.9 meV. In c direction, <lb/>as expected from the shape of the FS, the role of the IL <lb/>(Ca) gap becomes more important and the conductance <lb/>shows two distinct peaks with almost the same height. <lb/>However, for Γ = 0 even at T = 0 K these anisotropic <lb/>features are rapidly smeared out. The middle curves of <lb/>Fig. 4(a) and (b) show the effect on the theoretical con-<lb/>ductances of a broadening similar to that observed at <lb/>very low T in the experimental curves of Fig. 2. The con-<lb/>ductances become similar to single-gap ones and can be <lb/>perfectly fitted by single-gap 3D BTK curves, as shown <lb/>in Fig. 4 (black dash lines). The use of more complex <lb/>fitting models (anisotropic or two-band BTK) that we <lb/>tested on our data does not improve the fit substantially, <lb/>as already pointed out in Ref. 6, 9. In Fig. 4 the ex-<lb/>perimental ab-plane and c-axis conductances measured <lb/>at 400 mK are included too (circles) in order to show <lb/>the remarkable agreement with the theoretical curves for <lb/>the same level of broadening. Although this broadening <lb/>washes out the fine anisotropic structures of the conduc-<lb/>tance, a clear sign of the underlying gap anisotropy is <lb/>still present since the 3D BTK fit of the theoretical con-<lb/>ductances gives different order parameters in the two di-<lb/>rections, ∆ = 1.5 meV (Γ = 0.65 meV, Z = 0.765) for <lb/>a-axis direction and ∆ = 1.7 meV (Γ = 0.92 meV, Z = <lb/>1.015) for c-axis one. The c-axis value is in perfect agree-<lb/>ment with the experimental results (both single curves at <lb/>400 mK and the peak of the distribution of the 14 dif-<lb/>ferent contacts). In the ab-plane case, the experimental <lb/>∆ values from the curves at 400 mK and from the peak <lb/>of the distribution of Fig. 3 (a) (ranging from 1.3 meV <lb/>to 1.44 meV) are smaller than the value obtained from <lb/>the fit of the theoretical conductance. This discrepancy <lb/>could be ascribed to a possible slight overestimation of <lb/>the small π gap (and, maybe, an underestimation of the <lb/>large IL gap associated with Ca FS) in the theoretical <lb/>calculations. This fact appears reasonable if one consid-<lb/>ers that the first-principle calculations of Ref. [10] led to <lb/>an underestimation of T c of about 17 %. <lb/>In conclusion, the first directional PCAR measure-<lb/>ments in CaC 6 carried out also at T = mK both <lb/>along the ab-plane and the c-axis direction give strong <lb/>and reproducible evidence of the predicted anisotropic <lb/>nature of the superconducting gap in this GIC. New first-<lb/>principles calculations of the expected anisotropy in the <lb/>AR conductance curves fully support this conclusion and <lb/>indicate that the actual gap anisotropy in CaC 6 could be <lb/>even slightly greater than theoretically predicted. <lb/></body>
22
+
23
+ <div type="acknowledgement">We thank Lilia Boeri, O.V. Dolgov, E.K.U. Gross and <lb/>I.I. Mazin for useful discussions. This work was done <lb/>within the projects: PRIN (No. 2006021741) and Cy-<lb/>bersar (cofunded by MUR under PON). V.A.S. acknowl-<lb/>edges the support of RFBR (Proj. No. 06-02-16490). <lb/></div>
24
+
25
+ <listBibl>[1] T. E. Weller et al., Nature Phys. 1, 39 (2005). <lb/>[2] N. Emery et al., Phys. Rev. Lett. 95, 087003 (2005). <lb/>[3] J. S. Kim et al., Phys. Rev. Lett. 99, 027001 (2007). <lb/>[4] M. Calandra and F. Mauri, Phys. Rev. B 74, 094507 <lb/>(2006). <lb/>[5] I. I. Mazin et al., Physica C 460-462, 116 (2007). <lb/>[6] N. Bergeal et al., Phys. Rev. Lett. 97, 077003 (2006). <lb/>[7] G. Lamura et al., Phys. Rev. Lett. 96, 107008 (2006). <lb/>[8] J. S. Kim et al., Phys. Rev. Lett. 96, 217002 (2006). <lb/>[9] C. Kurter et al., Phys. Rev. B 76, 220502 (2007). <lb/>[10] A. Sanna et al., Phys. Rev. B 75, 020511(R) (2007). <lb/>[11] R.S. Gonnelli et al., Phys. Rev. Lett. 89, 247004 (2002). <lb/>[12] R.S. Gonnelli et al., Phys. Rev. Lett. 97, 037001 (2006). <lb/>[13] D. Daghero et al., Phys. Rev. B 74, 174519 (2006). <lb/>[14] Goutam Sheet, S. Mukhopadhyay, and P. Raychaudhuri, <lb/>Phys. Rev. B 69, 134507 (2004). <lb/>[15] Y.G. Naidyuk, I.K. Yanson, Point-Contact Spectroscopy, <lb/>Springer Series in Solid-State Sciences, Vol. 145, 2004. <lb/>[16] J. S. Kim, unpublished results. <lb/>[17] A. Gauzzi et al., Phys. Rev. Lett. 98, 067002 (2007). <lb/>[18] G.E. Blonder, M. Tinkham and T.M. Klapwijk, Phys. <lb/>Rev. B 25, 4515 (1982). <lb/>[19] A. Plecenik et al., Phys. Rev. B 49, 10016 (1994). <lb/>[20] S. Kashiwaya et al., Phys. Rev. B 53, 2667 (1996). <lb/>[21] P. Chalsani et al., Phys. Rev. B 75, 094417 (2007). <lb/>[22] I.I Mazin, Phys. Rev. Lett. 83, 1427 (1999). <lb/>[23] A. Brinkman et al., Phys. Rev. B 65, 180517(R) (2002). </listBibl>
26
+
27
+
28
+ </text>
29
+ </tei>