Pith. sign in

REVIEW 4 major objections 5 minor 37 references

Human-AI collaboration for modeling heat conduction in nanostructures

T0 review · 4 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper shows that three human-identified stacking-order parameters, surfaced by an entropy-based sampler and embedded in a symbolic-regression formula, predict thermal conductivity in graphene-WS2 heterostructures.

desk verdict The workflow is appealing, but Eq. 4 cannot produce the low-κ values the paper is after, so the central prediction claim fails on the paper's own numbers. read the letter →

arxiv 2502.05576 v1 pith:OU7X4ACR submitted 2025-02-08 physics.comp-ph

Wenyang Ding , Jiang Guo , Meng An , Koji Tsuda , Junichiro Shiomi This is my paper · ORCID
classification physics.comp-ph
keywords thermalconductivitygraphene-WS2heterostructuresphonontransmissionhuman-AIcollaborationsymbolicregressionentropicpopulationannealingatomisticGreen'sfunctioninterpretablemachinelearning
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Human involvement in materials machine learning is usually limited to setting up and monitoring the algorithm; this paper asks what happens when human expertise sits in the loop, interpreting what the AI proposes before a second AI builds the final model. Using heat conduction in 14-layer graphene-WS2 heterostructures as the test bed, the authors show that a sampling method called SLEPA can produce a small dataset that reproduces the full distribution of thermal conductivities, a human then extracts three stacking-order parameters Pa, Pb, and Pc from the extremes of that distribution, and symbolic regression compresses the parameters into a compact formula. The paper's central claim is that this human-AI sequence yields an interpretable predictive model—κ = 0.178/(9.89 + Pa + Pb + 0.143Pc) + 0.0109—that captures the dependence of thermal conductivity on stacking order and, through mode-resolved atomistic Green's function analysis, identifies the specific frequencies and incidence angles at which each parameter suppresses phonon transmission. If true, it would mean that expert intuition can be injected into materials discovery in a data-efficient way, producing design rules rather than black-box predictions.

What carries the argument

The carrying mechanism is an AI-human-AI pipeline whose load-bearing components are three human-identified stacking parameters computed from the binary layer sequence of the heterostructure (0 = graphene, 1 = WS2). Pa = (n0+1)/(sum0+1) measures how strongly graphene is buried away from the outer WS2 layers; Pb = (n>00+1)/(sum0+1) measures the share of graphene that appears in runs of length two or more; Pc = (n1 n11 +1)/(sum1+1) measures the product of single- and double-layer WS2 runs. SLEPA, a sampling method combining entropic sampling with a surrogate Gaussian-process model, supplies a small dataset that reproduces the full thermal-conductivity distribution, and the human step converts that distribution into these three descriptors. Symbolic regression then fits κ = 0.178/(9.89 + Pa + Pb + 0.143Pc) + 0.0109. The descriptors carry the argument because they are discrete-structure statistics that both correlate with conductivity and map onto specific phonon-suppression windows in frequency-incidence space.

What would settle it

Recompute thermal conductivities of the same 14-layer stacking sequences with an ab initio or experimentally benchmarked phonon method and check whether Pa, Pb, and Pc still separate low- from high-conductivity structures; even a single sequence with large Pa, Pb, and Pc that conducts better than Eq. (4) predicts would break the claimed design rule.

Watch

Extended reading notes

Core claim

On its own terms, the paper discovers three physically interpretable descriptors of stacking order in graphene-WS2 heterostructures and shows that they control the thermal conductivity through distinct phonon-suppression channels. Pa captures whether graphene layers are concentrated between outer WS2 layers, Pb captures whether graphene forms runs of two or more consecutive layers, and Pc captures the product of single- and double-layer WS2 runs. The final model, κ = 0.178/(9.89 + Pa + Pb + 0.143Pc) + 0.0109, predicts thermal conductivity from these three numbers, and the mode-resolved AGF analysis attributes each descriptor to a specific suppression regime: Pa suppresses normally incident low-frequency phonons, Pb suppresses normally incident mid-frequency phonons, and Pc suppresses both normally incident mid-frequency and obliquely incident high-frequency phonons. The paper argues that this mechanism-resolved, closed-form model can guide nanostructure design directly.

Load-bearing premise

The ground truth for every prediction is the thermal conductivity computed by atomistic Green's function with empirical interatomic potentials (Tersoff for graphene, Stillinger-Weber for WS2, and Lennard-Jones for van der Waals contacts); if those potentials misrepresent phonon transmission at the interface, the three parameters and the fitted formula would be artifacts of the force field rather than physical rules.

Editorial extensions

If this is right

  • The three stacking parameters can rank any 14-layer graphene-WS2 heterostructure by thermal conductivity without running a full phonon calculation.
  • The design rules translate into fabrication guidance: terminate with WS2 layers, keep graphene in multi-layer blocks, limit WS2 to one or two contiguous layers, and balance the two materials.
  • Because the model is an explicit formula, it can be inverted to search for stacking sequences that achieve a target thermal conductivity.
  • The frequency and incidence-angle map in Table 1 identifies which phonon populations to engineer, such as tuning Pa to suppress low-frequency normal-incidence phonons.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Outside the paper's own claims, the same SLEPA-to-human-features-to-symbolic-regression loop could be applied to other interface-controlled properties, with the human step identifying local structural motifs rather than global stack order.
  • Because Pa, Pb, and Pc are simple run-length statistics, they may generalize to other layered heterostructures and to longer layer counts, though the model may need an explicit layer-number dependence.
  • The model is trained on zero-temperature AGF conductances; an extension to finite-temperature or anharmonic effects would test whether the same parameters preserve their ranking of structures.
  • The near-minimum structures found by SLEPA and by the symbolic-regression formula are not identical, so an exhaustive check over all 16,384 candidates would directly quantify how much the human-selected descriptors miss.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes a human-AI collaboration workflow for 14-layer graphene-WS2 heterostructures: SLEPA combined with mode-resolved atomistic Green's function (AGF) generates a small thermal-conductivity dataset; human inspection defines three sequence descriptors Pa, Pb, and Pc; random forest and symbolic regression produce a compact formula κ = 0.178/(9.89 + Pa + Pb + 0.143Pc) + 0.0109 (Eq. 4); and mode-resolved AGF attributes each descriptor to specific frequency and incidence-angle phonon-suppression mechanisms. The paper claims that this interpretable model predicts thermal conductivity across the full space of 16,384 stacking sequences and yields actionable design rules for ultralow thermal conductivity.

Significance. If the central quantitative claim were sound, the work would provide a useful template for combining active learning, human feature construction, and transparent symbolic regression in phonon engineering. The full N=10 enumeration (1024 AGF calculations), the use of mode-resolved transmission data to interpret each descriptor, and the explicit design rules are valuable contributions. However, the central predictive claim is currently undermined by an inconsistency between Eq. 4 and the reported AGF values in the low-κ regime, and the reported accuracy metrics lack the protocol needed to support generalizability. As a computational study, all conclusions are also conditional on the empirical interatomic potentials used for the AGF ground truth, which the paper does not discuss critically.

major comments (4)
  1. [Construction of a predictive model, Eq. (4) and Fig. 8(c)] The model's stated theoretical minimum of 0.0256 W/m-K is inconsistent with the reported AGF value of 0.0180 W/m-K for the same structure '10000100110011'. Because Eq. 4's denominator is bounded for the defined descriptors, the model output cannot reach the low-κ values that SLEPA identifies and that the paper explicitly targets. This is not a small extrapolation error but a systematic failure in the region of interest. The authors should refit the symbolic expression, add a residual analysis for the low-κ tail, or restrict the predictive claims to the range actually supported by Eq. 4.
  2. [Construction of a predictive model, Fig. 8] The reported R2 = 0.70 for the random forest and the '64% accuracy' for symbolic regression are not accompanied by a train/test split, cross-validation, error bars, or residual plots. Since the descriptors Pa, Pb, and Pc were selected by human inspection of the same SLEPA-generated dataset that is later used to fit the models, these metrics are at risk of being in-sample and cannot establish out-of-sample predictive performance. A clear data-splitting or resampling protocol and a residual-vs-κ plot are needed, with particular attention to the low-κ regime.
  3. [Validation of SLEPA, Figs. 2–3] The claim that SLEPA 'mimics the original large dataset' is supported only by visual comparison of histograms. No quantitative distribution-distance metric (e.g., Kolmogorov-Smirnov or Hellinger distance) is reported, and no repeated-run statistics are provided to show that the 100-case SLEPA outcome is robust. Adding such metrics for the 10%, 20%, 30%, and 40% sample sizes would make the validation conclusion load-bearing rather than qualitative.
  4. [Identification of meaningful features and Construction of a predictive model] The feature-construction loop uses the SLEPA dataset both to discover Pa, Pb, and Pc and to fit the RF/SR models; this creates an in-sample selection effect that the paper does not address. The independent mode-resolved AGF mechanism analysis in Fig. 9 provides useful external grounding for the physical interpretation, but it does not validate the numerical accuracy of Eq. 4. The authors should clarify the chronology of descriptor selection and model fitting and evaluate the final model on held-out structures outside the SLEPA training pool.
minor comments (5)
  1. [Figure 3 caption] The caption lists '(b) SLEPA, (b) Bayesian optimization', duplicating the label for two different panels; the second panel should be labeled (c).
  2. [Table 1] The heading contains the typo 'frequences'; it should read 'frequencies'.
  3. [Results, SLEPA for 14-layer heterostructures] The text says the SLEPA-optimal structure '11000000101101' has a thermal conductivity of 0.018 W/m-K, while the SR-optimal structure is later said to be 0.0180 W/m-K and 'only slightly higher'. The rounding and the comparison should be made consistent so the reader can see whether these are the same value.
  4. [Methods, Mode-resolved AGF] The sentence 'the left and right leads consist of two layers of graphene or graphite' is ambiguous; it should specify whether the leads are graphene, graphite, or both depending on the terminal layers of the central heterostructure.
  5. [Data availability] The data availability statement only offers data 'from the corresponding author on reasonable request'; given the reproducibility emphasis of the study, a persistent repository for the 1024 and 1300 AGF datasets would strengthen the paper.

Circularity Check

1 steps flagged · score 4.0 of 10

Descriptor discovery and RF/SR model evaluation both use the same SLEPA dataset, so the reported R2 and 64% accuracy are in-sample fits; SLEPA self-citation is not load-bearing because SLEPA is validated against full 10-layer ground truth.

  1. fitted input called prediction [Results, 'Identification of meaningful features' and 'Construction of a predictive model' (Eq. 4, Fig. 8)]
    "During this process, the small dataset obtained for the thermal conductivity distribution was used as the basis for extracting and testing features, and the optimized heterostructure was used as inspiration to determine possible features. ... An RF model was trained using the three parameters Pa, Pb, and Pc. ... The actual thermal conductivities were compared to the values predicted by the trained RF model ... with an R2 value of 0.70. ... The actual thermal conductivities were compared to the predicted values obtained from the trained SR model ..."

    The descriptors Pa, Pb, and Pc were created by human inspection of the same SLEPA-generated dataset ('basis for extracting and testing features'); the RF and SR models were then trained on that dataset, and their 'predicted' values were compared to the very same actual values in Fig. 8. No train/test split or held-out set is described, so the reported R2=0.70 and 64% accuracy measure in-sample fit rather than predictive skill. Eq. 4 is a regression fitted to those points; labeling the resulting values 'predictions' and using them to claim an accurate predictive model is the fitted-input-called-prediction pattern. The independent modal AGF analysis (Fig. 9) and the N=10 SLEPA validation are separate evidence, but they do not convert the Fig. 8 comparison into an out-of-sample prediction.

full rationale

The paper's main derivation chain is otherwise self-contained. SLEPA is cited from the authors' prior work (Ref. [26]), but that citation is not load-bearing circularity: the paper independently validates SLEPA against full ground truth for all 1024 ten-layer structures (Figs. 2-3), so SLEPA's ability to produce a representative small dataset is externally checked. The AGF transmission calculations use standard literature formalisms (Ong, Landauer) and established empirical potentials; the mode-resolved mechanism analysis in Fig. 9 compares independently selected structures with matched control parameters and is not an output of the fitted model. The main circular concern is the model-evaluation loop: human-chosen features, model fitting, and the claimed R2/accuracy are all anchored to the same SLEPA thermal-conductivity values, with no described train/test separation. A separate correctness inconsistency exists (Eq. 4's theoretical minimum 0.0256 W/m-K exceeds the paper's own reported AGF value 0.0180 W/m-K for the SLEPA optimum), but that inconsistency is not circularity; it is a quantitative failure of the fitted model. Overall, the central mechanistic content retains independent support, so the circularity score is moderate rather than severe.

Assumptions & free parameters 6 free parameters · 5 assumptions · 3 invented entities

The central model rests on several fitted constants (Eq. 4) and on the assumed validity of empirical potentials and AGF. The human-selected descriptors are ad hoc but are independently tested via mode-resolved AGF, providing some external grounding.

free parameters (6)
  • SR model coefficient A = 0.178 W/m-K
    Fitted constant in Eq. (4), adjusts magnitude of kappa.
  • SR model coefficient B = 9.89 (dimensionless)
    Fitted denominator constant in Eq. (4).
  • SR model coefficient C = 0.143 (dimensionless)
    Weight for Pc in Eq. (4).
  • SR model offset D = 0.0109 W/m-K
    Fitted offset in Eq. (4).
  • SLEPA initial sample size = 100
    Initial M particles used to train the first Gaussian process model; chosen by the authors, affects exploration.
  • SLEPA rounds = 13 (about 10% of 16384)
    Stopping criterion; more rounds change the dataset seen by human experts.
assumptions (5)
  • domain assumption Mode-resolved atomistic Green's function (AGF) formalism is valid for coherent phonon transport in these heterostructures.
    Equations (6)-(8) are taken from Ref. [35] and assumed to describe transmission in the device region.
  • standard math Landauer formula for thermal conductance applies with the Bose-Einstein distribution.
    Eq. (8) uses the Landauer formalism, a standard result; no derivation given.
  • domain assumption Empirical interatomic potentials (optimized Tersoff for graphene, Stillinger-Weber for WS2, Lennard-Jones for vdW interactions) yield accurate force constants.
    Methods section specifies these potentials; their accuracy for thermal transport is not benchmarked against experiments in this paper.
  • domain assumption SLEPA produces an unbiased representative sample of the full configuration space.
    The claim that SLEPA mimics the true distribution (Figure 2) is based on visual comparison, not a quantitative statistical test; relies on Ref. [26].
  • domain assumption Binary layer sequence fully determines the heterostructure properties.
    The descriptor space is all 2^14 sequences; in-plane structure, defects, and temperature effects are ignored.
invented entities (3)
  • Pa independent evidence
    purpose: Descriptor for fraction of graphene layers located between the outermost WS2 layers
    Defined in Eq. (1); its phonon suppression role is independently tested with AGF in Fig. 9(a-d).
  • Pb independent evidence
    purpose: Descriptor for fraction of graphene layers in subsequences of length at least 2
    Defined in Eq. (2); mechanism tested in Fig. 9(e-h).
  • Pc independent evidence
    purpose: Descriptor for product of counts of single and double WS2 subsequences
    Defined in Eq. (3); mechanism tested in Fig. 9(i-l).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Human-AI collaboration for modeling heat conduction in nanostructures." pith.science (2026). https://pith.science/paper/OU7X4ACR

@misc{pith2026250205576,
  author       = {Pith},
  title        = {Pith review of: Human-AI collaboration for modeling heat conduction in nanostructures},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OU7X4ACR}},
  note         = {Machine review of arXiv:2502.05576}
}
read the original abstract

In recent years, materials informatics, which combines data science and artificial intelligence (AI), has garnered significant attention owing to its ability to accelerate material development, reduce costs, and enhance product design. However, despite the widespread use of AI, human involvement is often limited to the initiation and oversight of machine learning processes and rarely includes more substantial roles that capitalize on human intuition or domain expertise. Consequently, true human-AI collaborations, where integrated insights can be maximized, are scarce. This study considers the problem of heat conduction in a two-dimensional nanostructure as a case study. An integrated human-AI collaboration framework is designed and used to construct a model to predict the thermal conductivity. This approach is used to determine the parameters that govern phonon transmission over the full range of frequencies and incidence angles. During operation, the self-learning entropic population annealing technique, which combines entropic sampling with a surrogate machine learning model, generates a small dataset that can be interpreted by a human. Therefore, data-efficient and global modeling is achieved, and parameters with physical interpretations are developed, which can guide nanostructural design to produce materials with specific properties. The proposed framework can leverage the complementary strengths of humans and AI, thereby enhancing the understanding and control of materials.

Figures

Figures reproduced from arXiv: 2502.05576 by the authors.

Figure 7
Figure 7. Relationship between thermal conductivity and subsequence length of WS2. (a) Pa = 0.8, Pb = 0.6 and Pa= 0.8, Pb = 0.8. Here, n is subsequence length and n = 1, 2 represents subsequences “010” and “0110,” and n ≥ 2 represents subsequences “0110,” “01110,” etc. The x-axis, titled “number of WS2,” denotes the summation of the WS2 subsequence length across all the subsequences. (b) Optimal heterostructure identified by … view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 37 canonical work pages

  1. [1]

    Silver, T

    D. Silver, T. Hubert, J. Schrittwieser, I. Antonoglou, M. Lai, A. Guez, M. Lanctot, L. Sifre, D. Kumaran, T. Graepel, T. Lillicrap, K. Simonyan, D. Hassabis, Science 362, 1140-1144 (2018)

    work page 2018

  2. [2]

    Brown, T

    N. Brown, T. Sandholm, Science 365, 885-890 (2019)

    work page 2019

  3. [3]

    K. He, X. Zhang, S. Ren, J. Sun, IEEE International Conference on Computer Vision (ICCV 2015) 1502, (2015)

    work page 2015

  4. [4]

    Gulshan, R.P

    V. Gulshan, R.P. Rajan, K. Widner, D.J. Wu, P. Wubbels, T. Rhodes, K. Whitehouse, M. Coram, G.S. Corrado, K. Ramasamy, R. Raman, L.H. Peng, D.R. Webster, JAMA ophthalmology DOI (2019)

    work page 2019

  5. [5]

    Hitsuwari, Y

    J. Hitsuwari, Y. Ueda, W. Yun, M. Nomura, Computers in Human Behavior 139, 107502 (2023)

    work page 2023

  6. [6]

    Sharma, I.W

    A. Sharma, I.W. Lin, A.S. Miner, D.C. Atkins, T. Althoff, Nature Machine Intelligence 5, 46-57 (2023)

    work page 2023

  7. [7]

    Petrescu, A.S

    M. Petrescu, A.S. Krishen, Journal of Marketing Analytics 11, 263-274 (2023)

    work page 2023

  8. [8]

    R. Chen, Y. Tian, J. Cao, W. Ren, S. Hu, C. Zeng, Journal of Applied Physics 135, (2024)

    work page 2024

Show all 37 references
  1. [9]

    S. Hu, S. Ju, C. Shao, J. Guo, B. Xu, M. Ohnishi, J. Shiomi, Mater. Today Phys. 16, 100324 (2021)

    2021

  2. [10]

    S. Ju, T. Shiga, L. Feng, Z. Hou, K. Tsuda, J. Shiomi, Phys. Rev. X 7, 021024 (2017)

    2017

  3. [11]

    Spiece, S

    J. Spiece, S. Sangtarash, M. Mucientes, A.J. Moli na-Mendoza, K. Lulla, T. Mueller, O. Kolosov, H. Sadeghi, C. Evangeli, Nanoscale 14, 2593-2598 (2022)

    2022

  4. [12]

    P.-Z. Jia, Y. -J. Zeng, D. Wu, H. Pan, X. -H. Cao, W. -X. Zhou, Z. -X. Xie, J. -X. Zhang, K.-Q. Chen, J. Phys. Condens. Matter 32, 055302 (2019)

    2019

  5. [13]

    Geim, I.V

    A.K. Geim, I.V. Grigorieva, Nature 499, 419-425 (2013)

    2013

  6. [14]

    Zhang, Y

    L. Zhang, Y. Zhong, X. Qian, Q. Song, J. Zhou, L. Li, L. Guo, G. Chen, E.N. Wang, ACS Applied Materials & Interfaces 13, 46055-46064 (2021)

    2021

  7. [15]

    Vaziri, E

    S. Vaziri, E. Yalon, M. Muñ oz Rojo, S.V. Suryavanshi, H. Zhang, C.J. McClellan, C.S. Bailey, K.K.H. Smithe, A.J. Gabourie, V. Chen, S. Deshmukh, L. Bendersky, A.V. Davydov, E. Pop, Sci. Adv. 5, eaax1325 (2019)

    2019

  8. [16]

    Guo, Y.-D

    R. Guo, Y.-D. Jho, A.J. Minnich, Nanoscale 10, 14432-14440 (2018)

    2018

  9. [17]

    Z.-S. Wu, W. Ren, L. Gao, J. Zhao, Z. Chen, B. Liu, D. Tang, B. Yu, C. Jiang, H.- M. Cheng, ACS Nano 3, 411-417 (2009)

    2009

  10. [18]

    X. Li, J. Yu, S. Wageh, A.A. Al-Ghamdi, J. Xie, Small 12, 6640-6696 (2016)

    2016

  11. [19]

    Sorkin, H

    V. Sorkin, H. Pan, H. Shi, S.Y. Quek, Y.W. Zhang, Crit. Rev. Solid State 39, 319- 367 (2014)

    2014

  12. [20]

    Solanki, D.N

    G.K. Solanki, D.N. Gujarathi, M.P. Deshpande, D. Lakshminarayana, M.K. Agarwal, Cryst. Res. Technol. 43, 179-185 (2008)

    2008

  13. [21]

    C. Lee, J. Hong, M.-H. Whangbo, J.H. Shim, Chem. Mater. 25, 3745-3752 (2013)

    2013

  14. [22]

    Jiang, X

    P. Jiang, X. Qian, X. Gu, R. Yang, Adv. Mater. 29, 1701068 (2017)

    2017

  15. [23]

    Lindroth, P

    D.O. Lindroth, P. Erhart, Phys. Rev. B 94, 115205 (2016)

    2016

  16. [24]

    Seeber, E

    I. Seeber, E. Bittner, R.O. Briggs, T. de Vreede, G. -J. de Vreede, A. Elkins, R. Maier, A.B. Merz, S. Oeste-Reiß , N. Randrup, G. Schwabe, M. Sö llner, Information & Management 57, 103174 (2020)

    2020

  17. [25]

    Shrestha, S

    Y. Shrestha, S. Ben -Menahem, G. Krogh, California Management Review 61, 000812561986225 (2019)

    2019

  18. [26]

    J. Li, J. Zhang, R. Tamura, K. Tsuda, Digital Discovery 1, 295-302 (2022)

    2022

  19. [27]

    Cohen, Empirical Methods for Artificial Intelligence, 1995

    P. Cohen, Empirical Methods for Artificial Intelligence, 1995

    1995

  20. [28]

    Breiman, Mach

    L. Breiman, Mach. Learn. 45, 5-32 (2001)

    2001

  21. [29]

    Cranmer, A

    M. Cranmer, A. Sanchez Gonzalez, P. Battaglia, R. Xu, K. Cranmer, D. Spergel, S. Ho, Adv. Neural Inf. Process. Syst. 33, 17429-17442 (2020)

    2020

  22. [30]

    Zhang, T.S

    W. Zhang, T.S. Fisher, N. Mingo, J. Heat Transfer. 129, 483-491 (2006)

    2006

  23. [31]

    Zhang, T.S

    W. Zhang, T.S. Fisher, N. Mingo, Numerical Heat Transfer, Part B: Fundamentals 51, 333-349 (2007)

    2007

  24. [32]

    Plimpton, J

    S. Plimpton, J. Comput. Phys. 117, 1-19 (1995)

    1995

  25. [33]

    Lindsay, D.A

    L. Lindsay, D.A. Broido, Phys. Rev. B 81, 205441 (2010)

    2010

  26. [34]

    Mobaraki, A

    A. Mobaraki, A. Kandemir, H. Yapicioglu, O. Gü lseren, C. Sevik, Comput. Mater. Sci. 144, 92-98 (2018)

    2018

  27. [35]

    Ong, Phys

    Z.-Y. Ong, Phys. Rev. B 98, 195301 (2018)

    2018

  28. [36]

    L. Yang, B. Latour, A.J. Minnich, Phys. Rev. B 97, 205306 (2018)

    2018

  29. [37]

    A. Dhar, D. Roy, J. Stat. Phys. 125, 801-820 (2006)

    2006

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.