Pith. sign in

REVIEW 2 major objections 5 minor 55 references

Fundamental Factors Governing Stabilization of Janus 2D-Bulk Heterostructures with Machine Learning

T0 review · 2 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 Bulk properties, not the 2D layer, set heterostructure stability.

desk verdict A genuinely useful dataset and model, but the bulk-dominance conclusion is oversold beyond the sampled chemical space. read the letter →

arxiv 2502.04603 v1 pith:PDSWWCCM submitted 2025-02-07 cond-mat.mtrl-sci cond-mat.mes-hall

Tara M. Boland (Technical University of Denmark) , Rachel Gorelik (Arizona State University) , Arunima K. Singh (Arizona State University) This is my paper · ORCID
classification cond-mat.mtrl-scicond-mat.mes-hall
keywords Janus2Dmaterialsheterostructurestabilityrandomforestregressionbindingenergyz-separationdistancebulksurfaceelectronegativityhigh-throughputDFT
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

The paper asks which partner controls the stability of a three-atom-thick Janus 2D material placed on a metal surface. Based on 1147 density-functional-theory heterostructures and random-forest models, the answer is the bulk support: bulk electronegativity is the leading predictor of binding energy, and bulk surface energy is the leading predictor of the interfacial z-separation. The models reproduce the DFT data with root-mean-square errors of 0.047 eV/atom and 0.139 Å, and the calculated trends align with experimental observations of which metal substrates grow transition-metal dichalcogenides successfully. If this holds, substrate selection becomes the primary design lever for stabilizing and tuning 2D-on-bulk interfaces.

What carries the argument

The machinery is a pair of random-forest regression models trained on the lowest-energy DFT configuration for each 2D-bulk pair, using 155 candidate features from the 2D material, the bulk slab, and the relaxed interface. Feature reduction (low-variance, high-correlation, recursive elimination) leaves 15 features for the binding-energy model and 10 for the z-separation model. A deliberately restricted z-separation model uses only features available from databases and elemental properties—bulk surface energy, atomic number, packing efficiency, molar volume—and still predicts separation with R2 of 0.838. The physical argument that bulk properties should dominate is that the bulk surface is created by breaking strong metallic bonds while the 2D material's exfoliation breaks weak van der Waals bonds, so the bulk side sets the interaction energy scale.

What would settle it

Extend the study to non-Janus 2D materials with electronegativities below 1.5 or above 3.0 on the same 19 substrates, recompute binding energies, and retrain the random forest; if 2D electronegativity or the 2D–bulk difference enters the top features, the bulk-dominance claim fails for the broader class.

Watch

Extended reading notes

Core claim

The central claim is that in Janus 2D-bulk heterostructures, the bulk material's intrinsic properties dominate the interface energetics and geometry. For binding energy, bulk electronegativity has the highest feature importance (29%), ahead of the 2D–bulk electronegativity difference (13%), and removing the difference from the model slightly improves accuracy. For z-separation, bulk surface energy is the top feature, and higher surface energy correlates with smaller separation, consistent with more dangling bonds at the surface. The random-forest models, trained on the most stable configuration of each of 438 2D-bulk pairs, achieve R2 of 0.940 for binding energy and 0.838 for z-separation when the z-model is restricted to features that require no DFT calculation. The authors interpret the bulk dominance physically: creating the bulk surface breaks strong metallic bonds, whereas exfoliating the 2D layer breaks weak van der Waals bonds, so the bulk side carries the energy scale.

Load-bearing premise

The claim that bulk properties dominate stability rests on a dataset in which the 2D materials vary little (electronegativity 1.98–2.59) while the bulk set varies widely (0.95–2.54), so the bulk features have more variance to explain; a wider range of 2D materials could change the ranking.

Editorial extensions

If this is right

  • Substrate choice, not the 2D layer, becomes the primary design knob for stabilizing metastable Janus monolayers on metals.
  • The random-forest models can screen 2D-bulk pairs at negligible cost, with the z-separation model needing only database-available properties.
  • The published database of over 1200 heterostructures gives other researchers a benchmark and training set for interface models.
  • The computed stability ordering rationalizes experimental reports that some metal substrates (for example, Cu) grow many TMDCs while others (for example, Ni) grow few.
  • Because binding energy and z-separation correlate, a single bulk descriptor such as surface energy can guide targeted tuning of interface geometry.

Reading between the lines

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

  • The bulk-dominated feature ranking may partly reflect the narrow electronegativity range of the 2D set (1.98–2.59) compared with the bulk set (0.95–2.54); including 2D materials with more extreme electronegativities could shift importance toward the 2D side.
  • The 19 elemental cubic (111) substrates are all metals, so the conclusion that bulk properties dominate may not transfer to polar, compound, or differently oriented surfaces without retesting.
  • Stability here is thermodynamic (negative adsorption formation energy); kinetic barriers to nucleation and growth are not addressed, so a predicted stable pair might still fail in synthesis.
  • The no-DFT z-separation model's modest accuracy drop (R2 0.886 to 0.838) suggests it is best used for coarse screening, with DFT reserved for finalists.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 5 minor

Summary. The manuscript presents a high-throughput DFT study of 1147 Janus 2D-bulk heterostructures formed from 51 Janus 2D materials and 19 metallic cubic (111) substrates, together with random-forest regression models that predict the binding energy and interfacial z-separation of the most stable configuration for each of 438 2D-bulk pairs. The reported models achieve RMSEs of 0.047 eV/atom and 0.139 Å, respectively. The central physical claim is that bulk-material properties, specifically bulk electronegativity for binding energy and bulk surface energy for z-separation, are the leading fundamental factors governing heterostructure stabilization, implying that substrate choice is the primary design lever. The computed data are released in the aiHD database.

Significance. If the central claim holds, the work is significant: it provides a large, openly available database of ab initio 2D-bulk interface data, demonstrates that random-forest models can capture trends in binding energy and interfacial separation across this chemical space, and offers a physically plausible case study of charge redistribution and hybridization at the SbTeI/Sr interface. Strengths include the consistent vdW-corrected DFT workflow, the large dataset relative to prior 2D-bulk studies, the explicit cross-validation of the ML models, and the authors' own acknowledgment of the narrow 2D electronegativity range in the feature-importance discussion. However, the generalization of the feature-importance conclusion is not yet fully supported because the dataset is deliberately narrow on the 2D side and the feature-importance rankings are sensitive to that sampling.

major comments (2)
  1. [Machine Learning Insights into the Fundamental Factors Governing Janus 2D Heterostructure Stability and Figure 7(b)] The headline claim that bulk electronegativity and bulk surface energy are the 'leading fundamental factor' is not established by the reported feature-importance analysis. Random-forest feature importance depends on the range and variance of features in the training set. Here the Janus 2D electronegativities lie between 1.98 and 2.59, while the bulk values span 0.95 to 2.54, and all 19 substrates are metallic cubic (111) surfaces. Under this sampling, the bulk electronegativity has far more variance to explain ΔEb, so its high importance is partly a sampling artifact. The authors acknowledge this limitation for electronegativity ('Thus, for this subset of materials, the bulk electronegativity plays a much larger role than the difference'), yet the abstract and conclusion state the dominance as a general finding without that caveat. A concrete control is needed: for example, retrain the models on a set with a broader 2D electronegativity range (e.g., oxides, fluorides, or alkali-based 2D compounds) or use permutation importance on variance-balanced resamples. If bulk dominance persists under such controls, the claim is strengthened; otherwise, it should be restricted to the studied chemical subspace.
  2. [Table 1 and Section 'Energetic Stability of Janus 2D Materials'] The term 'thermodynamically stable' overstates what is computed. The criterion ΔEf_ads < 0 compares the heterostructure with the freestanding 2D material and its 3D bulk counterpart; it does not establish stability against all competing phases and includes no temperature or pressure dependence. The 828 'stable' configurations are better described as 'bulk-stabilized relative to the freestanding 2D material.' Because the number 828 is cited in the abstract as a central result, the wording should be corrected to match the actual definition.
minor comments (5)
  1. [Table 1 and Figure 7] The reported RMSE and MAE values are cross-validated on the same dataset, but no uncertainty intervals are given. Reporting standard deviations across cross-validation folds would help assess whether the differences between model variants (e.g., RFRz−sep and RFRz−sep*) are meaningful.
  2. [Machine Learning Methodology] The feature-reduction process is described only at a high level, and the final lists of 15 and 10 features are not given in the main text. Since the feature set is central to the interpretation, the full feature lists and their importances should be included or more completely summarized in the main text.
  3. [Abstract] The abstract says 'nearly 1000 heterostructures' but the paper reports 1147 heterostructure configurations; please make the count consistent.
  4. [DFT Methods] There is a typo in 'V ASP' (should be VASP) and in 'M atminer's database' (should be Matminer) in the text near Figure 3.
  5. [Energetic Stability of Janus 2D Materials] The phrase 'the lower the ΔEb the lower the stability' is confusing: since ΔEb is defined as (E2D + ES − E2D+S)/N2D, a lower (more negative) ΔEb means stronger binding and thus higher stability. Please clarify the sign convention.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the ML models are cross-validated fits, the feature-importance claims are explicitly confined to the sampled chemical space, and the self-citations are methodological rather than load-bearing.

full rationale

The paper's derivation chain is: DFT-computed dataset → random forest regression → feature importances → conclusion that bulk electronegativity and bulk surface energy dominate. None of these steps is circular by the standards of the review. The reported RMSE values (0.047 eV/atom and 0.139 Å) are cross-validated performance metrics on the same dataset; this is standard supervised machine learning and does not constitute 'a fitted parameter renamed as prediction.' The models are empirical fits, not first-principles derivations, and the paper does not claim otherwise. The central feature-importance conclusion is also presented with an explicit sampling caveat: 'The Janus 2D materials in our study all had electronegativities falling between 1.98-2.59, which is a much narrower range than that of the study's bulk materials (0.95-2.54). Thus, for this subset of materials, the bulk electronegativity plays a much larger role than the difference in determining heterostructure stability (i.e. ∆ Eb ).' This admission locates the bulk-dominance finding within the studied chemical subspace rather than deriving it from the definition of any feature or target. The concern that a wider 2D electronegativity range or different substrate classes could change the ranking is an external-validity or sampling-risk criticism, not a circularity. Self-citations to Ref. 27 are used as methodological references for the Hetero2d package, DFT parameters, and the set of 19 bulk materials; they are not invoked as an unverified theorem to force a conclusion. No equation is recycled as its own output, and no known result is merely renamed. Therefore the paper is self-contained against external benchmarks in the sense relevant to circularity, and the honest finding is no significant circularity.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central claims rest on four classes of unverified inputs: DFT exchange-correlation accuracy, the statistical interpretation of feature importance, the representativeness of the materials sample, and the stability criterion. No new physical entities are introduced, and the only fitted quantities are ML hyperparameters and screening cutoffs.

free parameters (2)
  • Machine learning hyperparameters = 150 or 500 trees; K=8 or 5; correlation threshold 65%; RFECV feature counts 15 and 10
    Tuned with GridSearchCV on the training data; they affect reported RMSE and feature importance, so they are choices fitted to the dataset.
  • Lattice-matching screening thresholds = maximum surface area 80 A^2; maximum 2D strain 3%
    Hand-chosen cutoffs in Hetero2d screening; they determine which 438 pairs enter the dataset and therefore shape the ML feature-importance conclusion.
assumptions (5)
  • domain assumption vdW-DF/optB88 accurately captures 2D-bulk van der Waals interactions and binding energies.
    All DFT energies and z-separations depend on this exchange-correlation approximation; no verification against experiment or higher-level theory is provided.
  • domain assumption Random forest feature importance reflects the physical factors governing stabilization.
    The paper interprets feature importance as 'fundamental factors,' but feature importance is a correlation measure and does not establish causation.
  • domain assumption The 51 Janus 2D and 19 bulk materials sampled are representative of 2D-bulk heterostructures generally.
    The conclusions about bulk dominance are drawn from this specific set; the paper itself notes the narrow 2D electronegativity range.
  • domain assumption Negative adsorption formation energy identifies thermodynamic stability.
    Stability is judged only by Delta_E_f_ads relative to the freestanding 2D material and slab; entropic, temperature, and competing-phase effects are not included.
  • domain assumption Lattice-matched low-strain configurations are representative of synthesizable interfaces.
    Only pairs with gamma_SA < 80 A^2 and sigma_2D < 3% on (111) metal surfaces are considered; other orientations and lattice mismatches are excluded by design.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Fundamental Factors Governing Stabilization of Janus 2D-Bulk Heterostructures with Machine Learning." pith.science (2026). https://pith.science/paper/PDSWWCCM

@misc{pith2026250204603,
  author       = {Pith},
  title        = {Pith review of: Fundamental Factors Governing Stabilization of Janus 2D-Bulk Heterostructures with Machine Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PDSWWCCM}},
  note         = {Machine review of arXiv:2502.04603}
}
abstract

The more-than-6000 2D materials predicted thus far provide a huge combinatorial space for forming functional heterostructures with bulk materials, with potential applications in nanoelectronics, sensing, and energy conversion. In this work, we investigate nearly 1000 heterostructures, the largest number of heterostructures thus far, of 2D Janus and bulk materials' surfaces using ab initio simulations and machine learning (ML) to deduce the structure-property relationships of the complex interfaces in such heterostructures. We first perform van der Waals-corrected density functional theory simulations using a high-throughput computational framework on 51 Janus 2D materials and 19 metallic, cubic phase, elemental bulk materials that exhibit low lattice mismatches and low coincident site lattices. The formation energy of the resultant 1147 Janus 2D-bulk heterostructures were analyzed and 828 were found to be thermodynamically stable. ML models were trained on the computed data, and we found that they could predict the binding energy and $z$-separation of 2D-bulk heterostructures with root mean squared errors (RMSE) of 0.05 eV/atom and 0.14 angstroms, respectively. The feature importance of the models reveals that the properties of the bulk materials dominate the heterostructures' energies and interfacial structures heavily. These findings are in-line with experimentally observed behavior of several well-known 2D materials-bulk systems. The data used within this paper is freely available in the Ab Initio 2D-Bulk Heterostructure Database (aiHD). The fundamental insights on 2D-bulk heterostructures and the predictive ML models developed in this work could accelerate the application of thousands of 2D-bulk heterostructures, thus stimulating research within a wide range of electronic, quantum computing, sensing, and energy applications.

Figures

Figures reproduced from arXiv: 2502.04603 by the authors.

Figure 1
Figure 1. 1H-Janus 2D-bulk heterostructure model showing the orientation of the 2D ma￾terial on top of the bulk surface. The Y-M-X represents the atomic layers of the Janus 2D material. In this report, we introduce the Ab Initio 2D-Bulk Heterostructure Database (aiHD), the first freely available database of ab initio-calculated data on 2D-bulk heterostructures and their post-adsorbed properties. To construct, optimize, and an… view at source ↗
Figure 2
Figure 2. (a) Schematic illustrating the criteria for creating heterostructures, (b) the 2D [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Schematic diagram of the steps taken to prepare the data set and features to [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: (a) Adsorption formation energy, ∆E f ads, of 1T- (triangle markers) and 1H-SbTeI (circle markers) on the lattice- and symmetry-matched bulk surfaces. The bulk materials are noted on the x-axis, and the ∆E f ads for the lowest energy heterostructure configuration is sh…
Figure 5
Figure 5. Figure 5: Heatmap of the ∆Eb for the (a) 1T and (b) 1H phase of the Janus 2D materials with the 2D materials listed on the y-axis and the bulk materials on the x-axis and the color map represents the (∆Eb) binding energy. The materials are sorted based on their electronegativiti…
Figure 6
Figure 6. Figure 6: Heatmap of the z-separation for the (a) 1T and (b) 1H phases of the Janus 2D materials, sorted based on the electronegativity of the 2D materials (y-axis) and bulk’s surface energy (x-axis). The color map represents the z-separation. The clear correlation between ∆Eb a…
Figure 7
Figure 7. Figure 7: (a) Scatter plot compares the DFT-computed and machine learning-predicted [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 8
Figure 8. Figure 8: (a) Scatter plot comparing the DFT-computed and machine learning-predicted [PITH_FULL_IMAGE:figures/full_fig_p018_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 54 canonical work pages

  1. [1]

    S.; Geim, A

    Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Zhang, Y.; Dubonos, S. V.; Grigorieva, I. V.; Firsov, A. A. Electric Field Effect in Atomically Thin Carbon Films. Science 2004, 306, 666--669

    work page 2004

  2. [2]

    K.; Hennig, R

    Singh, A. K.; Hennig, R. G.; Davydov, A. V.; Tavazza, F. Al _2 O _3 as a suitable substrate and a dielectric layer for n-layer MoS _2 . Applied Physics Letters 2015, 107, 053106

    work page 2015

  3. [3]

    K.; Zhuang, H

    Singh, A. K.; Zhuang, H. L.; Hennig, R. G. Ab initio synthesis of single-layer III-V materials . Physical Review B - Condensed Matter and Materials Physics 2014, 89, 1--10

    work page 2014

  4. [4]

    Zhuang, A

    Houlong L. Zhuang, A. K. S. R. G. H., Michelle D. Johannes Doping-controlled phase transitions in single-layer MoS2. Physical Review B 2017, 96, 165305

    work page 2017

  5. [5]

    K.; Huang, Z.-F.; Elder, K

    Smirman, M.; Taha, D.; Singh, A. K.; Huang, Z.-F.; Elder, K. R. Influence of misorientation on graphene Moir \' e patterns . Physical Review B 2017, 95, 1--7

    work page 2017

  6. [6]

    T.; Singh, A

    Paul, J. T.; Singh, A. K.; Dong, Z.; Zhuang, H.; Revard, B. C.; Rijal, B.; Ashton, M.; Linscheid, A.; Blonsky, M.; Gluhovic, D.; Guo, J. Computational methods for 2D materials: discovery, property characterization, and application design . Journal of Physics: Condensed Matter 2017, 29, 473001

    work page 2017

  7. [7]

    K.; Hennig, R

    Singh, A. K.; Hennig, R. G. Computational prediction of two-dimensional group-IV mono-chalcogenides . Applied Physics Letters 2014, 105, 042103

    work page 2014

  8. [8]

    K.; Mathew, K.; Zhuang, H

    Singh, A. K.; Mathew, K.; Zhuang, H. L.; Hennig, R. G. Computational screening of 2D materials for photocatalysis. Journal of Physical Chemistry Letters 2015, 6, 1087--1098

    work page 2015

Show all 55 references
  1. [9]

    K.; Mathew, K.; Davydov, A

    Singh, A. K.; Mathew, K.; Davydov, A. V.; Hennig, R. G.; Tavazza, F. High throughput screening of substrates for synthesis and functionalization of 2D materials. SPIE 2015, 9553

    2015

  2. [10]

    Pakdel, S.; Rasmussen, A.; Taghizadeh, A.; Kruse, M.; Olsen, T.; Thygesen, K. S. High-throughput computational stacking reveals emergent properties in natural van der Waals bilayers. Nature Communications 2024, 15, 932

    2024

  3. [11]

    K.; Zhao, W

    Zhang, Q. K.; Zhao, W. H.; Zhou, Z. P.; Cao, L. M.; Yin, W. J.; Wei, X. L.; Tang, Z. K.; Zhang, H. 2D Janus MoSSe/MoGeSiN4 vdW heterostructures for photovoltaic and photocatalysis applications. Journal of Alloys and Compounds 2023, 938, 168708

    2023

  4. [12]

    H.; Ribeiro-Palau, R.; Hone, J

    Rhodes, D.; Chae, S. H.; Ribeiro-Palau, R.; Hone, J. Disorder in van der Waals heterostructures of 2D materials . Nature materials 2019, 18, 541--549

    2019

  5. [13]

    Computational discovery of stable phases of graphene and h-BN van der Waals heterostructures composed of group III-V binary compounds

    Akiyama, T.; Kawamura, T.; Ito, T. Computational discovery of stable phases of graphene and h-BN van der Waals heterostructures composed of group III-V binary compounds . Applied Physics Letters 2021, 118

    2021

  6. [14]

    O.; Azadani, J

    \"Oz c c elik, V. O.; Azadani, J. G.; Yang, C.; Koester, S. J.; Low, T. Band alignment of two-dimensional semiconductors for designing heterostructures with momentum space matching. Phys. Rev. B 2016, 94, 035125

    2016

  7. [15]

    K.; Grigorieva, I

    Geim, A. K.; Grigorieva, I. V. Van der Waals heterostructures . Nature 2013, 499, 419--425

    2013

  8. [16]

    Predicting Van der Waals Heterostructures by a Combined Machine Learning and Density Functional Theory Approach

    Willhelm, D.; Wilson, N.; Arroyave, R.; Qian, X.; Cagin, T.; Pachter, R.; Qian, X. Predicting Van der Waals Heterostructures by a Combined Machine Learning and Density Functional Theory Approach . ACS Applied Materials & Interfaces 2022, 14, 25907--25919

    2022

  9. [17]

    Novel optoelectronic devices: transition-metal-dichalcogenide-based 2D heterostructures

    Zeng, Q.; Liu, Z. Novel optoelectronic devices: transition-metal-dichalcogenide-based 2D heterostructures. Advanced Electronic Materials 2018, 4, 1700335

    2018

  10. [18]

    S.; Hinsche, N

    Haastrup, S.; Strange, M.; Pandey, M.; Deilmann, T.; Schmidt, P. S.; Hinsche, N. F.; Gjerding, M. N.; Torelli, D.; Larsen, P. M.; Riis-Jensen, A. C.; Gath, J.; Jacobsen, K. W.; Mortensen, J. J.; Olsen, T.; Thygesen, K. S. The Computational 2D Materials Database: High-Throughpu...

    2018

  11. [19]

    B.; Hennig, R

    Ashton, M.; Paul, J.; Sinnott, S. B.; Hennig, R. G. Topology-scaling identification of layered solids and stable exfoliated 2D materials . Physical Review Letters 2017, 118, 106101

    2017

  12. [20]

    High-throughput Identification and Characterization of Two-dimensional Materials using Density functional theory

    Choudhary, K.; Kalish, I.; Beams, R.; Tavazza, F. High-throughput Identification and Characterization of Two-dimensional Materials using Density functional theory. Scientific Reports 2017 7:1 2017, 7, 1--16

    2017

  13. [21]

    High-throughput computational screening of layered and two-dimensional materials

    Zhang, X.; Chen, A.; Zhou, Z. High-throughput computational screening of layered and two-dimensional materials. Wiley Interdisciplinary Reviews: Computational Molecular Science 2019, 9, e1385

    2019

  14. [22]

    The Materials Cloud 2D database (MC2D

    Campi, D.; Mounet, N.; Gibertini, M.; Pizzi, G.; Marzari, N. The Materials Cloud 2D database (MC2D. Materials Cloud Archive 2022,

    2022

  15. [23]

    L.; Jahnatek, M.; Chepulskii, R

    Curtarolo, S.; Setyawan, W.; Hart, G. L.; Jahnatek, M.; Chepulskii, R. V.; Taylor, R. H.; Wang, S.; Xue, J.; Yang, K.; Levy, O.; Mehl, M. J.; Stokes, H. T.; Demchenko, D. O.; Morgan, D. AFLOW: An automatic framework for high-throughput materials discovery. Computational Materi...

    2012

  16. [24]

    P.; Hautier, G.; Chen, W.; Richards, W

    Jain, A.; Ong, S. P.; Hautier, G.; Chen, W.; Richards, W. D.; Dacek, S.; Cholia, S.; Gunter, D.; Skinner, D.; Ceder, G.; Persson, K. A. The Materials Project: A materials genome approach to accelerating materials innovation . APL Materials 2013, 1, 011002

    2013

  17. [25]

    E.; Meredig, B.; Thompson, A.; Doak, J

    Kirklin, S.; Saal, J. E.; Meredig, B.; Thompson, A.; Doak, J. W.; Aykol, M.; R\" u hl, S.; Wolverton, C. The Open Quantum Materials Database (OQMD): assessing the accuracy of DFT formation energies. npj Computational Materials 2015, 1, 15010

    2015

  18. [26]

    I.; Technology NIST Inorganic Crystal Structure Database

    of Standards, N. I.; Technology NIST Inorganic Crystal Structure Database. NIST Standard Reference Database Number 3

  19. [27]

    M.; Singh, A

    Boland, T. M.; Singh, A. K. Computational synthesis of 2D materials: A high-throughput approach to materials design. Computational Materials Science 2022, 207, 111238

    2022

  20. [28]

    C.; Deilmann, T.; Olsen, T.; Thygesen, K

    Riis-Jensen, A. C.; Deilmann, T.; Olsen, T.; Thygesen, K. S. Classifying the Electronic and Optical Properties of Janus Monolayers . ACS Nano 2019, 13, 13354--13364

    2019

  21. [29]

    Recent advances in emerging Janus two-dimensional materials: From fundamental physics to device applications

    Zhang, L.; Yang, Z.; Gong, T.; Pan, R.; Wang, H.; Guo, Z.; Zhang, H.; Fu, X. Recent advances in emerging Janus two-dimensional materials: From fundamental physics to device applications . Journal of Materials Chemistry A 2020, 8, 8813--8830

    2020

  22. [30]

    Multifunctional two-dimensional van der Waals Janus magnet Cr-based dichalcogenide halides

    Hou, Y.; Xue, F.; Qiu, L.; Wang, Z.; Wu, R. Multifunctional two-dimensional van der Waals Janus magnet Cr-based dichalcogenide halides . npj Computational Materials 2022, 8, 1--6

    2022

  23. [31]

    P.; Richards, W

    Ong, S. P.; Richards, W. D.; Jain, A.; Hautier, G.; Kocher, M.; Cholia, S.; Gunter, D.; Chevrier, V. L.; Persson, K. A.; Ceder, G. Python Materials Genomics (pymatgen): A robust, open-source python library for materials analysis . Computational Materials Science 2013, 68, 314--319

    2013

  24. [32]

    Zur, A.; McGill, T. C. Lattice match: An application to heteroepitaxy . Journal of Applied Physics 1984, 55, 378--386

    1984

  25. [33]

    K.; Gabriel, J

    Mathew, K.; Singh, A. K.; Gabriel, J. J.; Choudhary, K.; Sinnott, S. B.; Davydov, A. V.; Tavazza, F.; Hennig, R. G. MPInterfaces: A Materials Project based Python tool for high-throughput computational screening of interfacial systems . Computational Materials Science 2016, 12...

    2016

  26. [34]

    Ab initio molecular dynamics for liquid metals

    Kresse, G.; Hafner, J. Ab initio molecular dynamics for liquid metals . Phys. Rev. B 1993, 47, 558

    1993

  27. [35]

    Ab initio molecular-dynamics simulation of the liquid-metal-amorphous-semiconductor transition germanium

    Kresse, G.; Hafner, J. Ab initio molecular-dynamics simulation of the liquid-metal-amorphous-semiconductor transition germanium . Phys. Rev. B 1994, 49, 14251

    1994

  28. [36]

    Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave set

    Kresse, G.; Furthm\" u ller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave set . Comput. Mat. Sci. 1996, 6, 15

    1996

  29. [37]

    Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set

    Kresse, G.; Furthm \" u ller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set . Physical Review B 1996, 54, 11169--11186

    1996

  30. [38]

    From ultrasoft pseudopotentials to the projector augmented-wave method

    Kresse, G.; Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wave method . Physical Review B 1999, 59, 1758

    1999

  31. [39]

    I.; Langreth, D

    Rydberg, H.; Dion, M.; Jacobson, N.; Schr\" o der, E.; Hyldgaard, P.; Simak, S. I.; Langreth, D. C.; Lundqvist, B. I. Phys. Rev. Lett. 2003, 91

    2003

  32. [40]

    Bowler, J. K. D. R.; Michaelides, A. Van der Waals density functionals applied to solids . Physical Review B 2011, 83, 195131

    2011

  33. [41]

    A grid-based Bader analysis algorithm without lattice bias

    Tang, W.; Sanville, E.; Henkelman, G. A grid-based Bader analysis algorithm without lattice bias . Journal of Physics Condensed Matter 2009, 21

    2009

  34. [42]

    A fast and robust algorithm for Bader decomposition of charge density

    Henkelman, G.; Arnaldsson, A.; J nsson, H. A fast and robust algorithm for Bader decomposition of charge density. Computational Materials Science 2006, 36

    2006

  35. [43]

    Scikit-learn: Machine Learning in Python

    Pedregosa, F.; Varoquaux, G.; Gramfort, A.; Thirion, B.; Grisel, O.; Prettenhofer, P.; Weiss, R.; Dubourg, V.; Vanderplas, J.; Passos, A.; Cournapeau, D.; Brucher, M.; Perrot, M. Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research 2011, 12, 2825--2830

    2011

  36. [44]

    API design for machine learning software: experiences from the scikit-learn project

    Buitinck, L.; Louppe, G.; Blondel, M.; Pedregosa, F.; Mueller, A.; Grisel, O.; Niculae, V.; Prettenhofer, P.; Gramfort, A.; Grobler, J.; Layton, R.; VanderPlas, J.; Joly, A.; Holt, B.; Varoquaux, G. API design for machine learning software: experiences from the scikit-learn pr...

    2013

  37. [45]

    Machine learning in materials research: developments over the last decade and challenges for the future

    Jain, A. Machine learning in materials research: developments over the last decade and challenges for the future. 2024,

    2024

  38. [46]

    Machine learning in materials science: From explainable predictions to autonomous design

    Pilania, G. Machine learning in materials science: From explainable predictions to autonomous design. Computational Materials Science 2021, 193, 110360

    2021

  39. [47]

    Biswas, T.; Singh, A. K. Incorporating quasiparticle and excitonic properties into material discovery. arXiv preprint arXiv:2401.17831 2024,

    2024 arXiv

  40. [48]

    Ward, L. et al. Matminer: An open source toolkit for materials data mining. Computational Materials Science 2018, 152, 60--69

    2018

  41. [49]

    K.; Hennig, R

    Singh, A. K.; Hennig, R. G. Computational synthesis of single-layer GaN on refractory materials . Applied Physics Letters 2014, 105

    2014

  42. [50]

    Adhesion Energy of MoS _2 Thin Films on Silicon-Based Substrates Determined via the Attributes of a Single MoS _2 Wrinkle

    Deng, S.; Gao, E.; Xu, Z.; Berry, V. Adhesion Energy of MoS _2 Thin Films on Silicon-Based Substrates Determined via the Attributes of a Single MoS _2 Wrinkle . ACS Publications 2017, 9, 7812--7818

    2017

  43. [51]

    Geng, D.; Yang, H. Y. Recent advances in growth of novel 2D materials: beyond graphene and transition metal dichalcogenides. Advanced Materials 2018, 30, 1800865

    2018

  44. [52]

    Growth of multi-layer hBN on Ni (111) substrates via MOCVD

    Siegel, G.; Gryzbowcki, G.; Hilton, A.; Muratore, C.; Snure, M. Growth of multi-layer hBN on Ni (111) substrates via MOCVD. Crystals 2019, 9, 339

    2019

  45. [53]

    A.; Chua, D

    Loh, T. A.; Chua, D. H. Growth mechanism of pulsed laser fabricated few-layer MoS2 on metal substrates. ACS applied materials & interfaces 2014, 6, 15966--15971

    2014

  46. [54]

    T.-W.; Chang, C.-S.; Li, L.-J.; others Synthesis of large-area MoS2 atomic layers with chemical vapor deposition

    Lee, Y.-H.; Zhang, X.-Q.; Zhang, W.; Chang, M.-T.; Lin, C.-T.; Chang, K.-D.; Yu, Y.-C.; Wang, J. T.-W.; Chang, C.-S.; Li, L.-J.; others Synthesis of large-area MoS2 atomic layers with chemical vapor deposition. arXiv preprint arXiv:1202.5458 2012,

    2012 arXiv

  47. [55]

    IDATxq G7魽U Pr # LG : XT &#0 V V

    Novoselov, K. S.; Colombo, L.; Gellert, P.; Schwab, M.; Kim, K.; others A roadmap for graphene. nature 2012, 490, 192--200 mcitethebibliography Main-Figs/0000775000000000000000000000000014751257573011350 5ustar rootrootMain-Figs/Fig2_iface_criteria.png0000664000000000000000000...

    1972

Pith tools

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