Pith. sign in

REVIEW 3 major objections 4 minor 51 references

A machine-learning phonon screen of 1,620 metal–sandwich combinations finds 1,208 dynamically stable monolayers, with MoSe2 the most effective casing.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review

2026-08-04 21:00 UTC pith:X6NLZ45Y

load-bearing objection A useful high-throughput UMLIP screening of metallene sandwiches with a defensible screening hypothesis, but the 1208-structure count rests on a single MLIP validated against DFT for only two systems — treat the rankings as hypotheses until spot-checked. the 3 major comments →

arxiv 2608.01779 v1 pith:X6NLZ45Y submitted 2026-08-03 cond-mat.mtrl-sci physics.comp-ph

Machine-Learning-Accelerated Metallene Stabilization from High-Throughput Sandwich Modeling

Mohammad Bagheri , Pekka Koskinen This is my paper
classification cond-mat.mtrl-sci physics.comp-ph
keywords metallenesmetallene sandwich heterostructuresvan der Waals heterostructuresuniversal machine-learning interatomic potentialsphonon dynamical stabilitytransition-metal dichalcogenidestensile-strain stabilizationhigh-throughput screening
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

This paper tries to show that fragile one-atom-thick metal sheets, called metallenes, can be kept from collapsing by squeezing them between two inert layered materials, and that the recipe can be mapped systematically. Using a universal machine-learning interatomic potential instead of expensive quantum-mechanical calculations, it builds 1,620 sandwiches from six casing layers and 45 metals in six different lattices, performs phonon calculations, and finds 1,208 dynamically stable combinations. It identifies MoSe2 as the best casing, buckled hexagonal lattices as the most stable metal geometries, and argues that stabilization works through a few percent of tensile strain plus sufficiently strong interlayer forces. If the screen is right, experimentalists get a quantitative menu of casing/metal/lattice combinations to try for making real metallenes.

Core claim

On its own terms, the paper establishes that van der Waals sandwiching is a general, tunable route to stable two-dimensional elemental metals. It constructs 1,620 metallene sandwich heterostructures—six casing layers (graphene, h-BN, MoS2, MoSe2, WS2, WSe2) around monolayers of 45 metals in six lattices (hexagonal, square, honeycomb, and their buckled forms)—and, using MatterSim's machine-learning phonons, finds 1,208 dynamically stable combinations. The most effective casing is MoSe2 (235 stable of 270), followed by WSe2 and MoS2, while h-BN and graphene stabilize far fewer. Buckled hexagonal metallenes are the most robust, and the paper connects this to two factors: the casing imposes a te

What carries the argument

The load-bearing machinery is a high-throughput phonon screen built on a universal machine-learning interatomic potential (MatterSim). Supercells are matched to below 3% lattice mismatch, relaxed to a strict force tolerance, and classified by imaginary phonon frequencies, with a structure counted unstable if any imaginary frequency exceeds 0.2 THz. The explanatory mechanism is a pair of quantities derived from the relaxed geometries: sandwich-induced biaxial strain and the maximum interlayer force constant between metal and casing atoms, computed with the FCDimen method. Stable sandwiches cluster at tensile strains below about 5% and at force constants of 1.48–3.42 eV/Ų, which together act

Load-bearing premise

The stability census stands on MatterSim's transferability to metal–TMD interfaces; it was benchmarked against first-principles phonons for only two of the 1,620 sandwiches, and the paper concedes larger phonon-band shifts for the MoS2|Mg case.

What would settle it

Run density-functional perturbation theory phonons for all 45 MoSe2-sandwiched buckled-hexagonal metallenes and count mismatches in imaginary-frequency classification; if MatterSim mislabels more than a handful, the 1,208 total and the MoSe2 ranking are not reliable.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • The experimentally demonstrated MoS2-squeezing route is not an isolated case: the same stabilization works across many casings, and the paper's list of 1,208 combinations is a direct candidate pool for synthesis attempts.
  • For late and post-transition metals (Bi, Pb, In, Ga, Sn, and noble metals), the casing choice decides success or failure, so experiments with these metals should prioritize TMD casings such as MoSe2 over h-BN or graphene.
  • The stability criteria translate into design targets: keep the metal under mild tensile strain below about 5% and ensure a maximum metal–casing force constant near 1.5–3.4 eV/Ų.
  • Buckled hexagonal lattices are the most stabilizable geometry and buckled honeycomb the least, meaning the casing can select which monolayer phase of a given metal actually forms.
  • The 3% lattice-mismatch construction rule provides a practical bound for designing future sandwich experiments.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • This screen is a filter on static and short-time dynamical stability, not a growth prediction: a metal listed as stable still needs a kinetic pathway into the sandwich, so the 1,208 count is better read as an upper bound on synthesizable metallenes.
  • If the strain/force-constant window holds for alloys, the same two-number design rule could be used to screen bimetallic or doped monolayer membranes instead of elemental metals.
  • Because the machine-learning potential was validated against DFT phonons for only two of the 1,620 sandwiches, targeted DFPT spot-checks on late- and post-transition-metal cases would harden the MoSe2 ranking and the overall stability count.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The manuscript proposes a high-throughput computational workflow for predicting dynamically stable metallene sandwich heterostructures (MSHs). Using the universal machine-learning interatomic potential MatterSim, the authors construct 1620 MSHs from 6 sandwich layers (graphene, h-BN, MoS2, MoSe2, WS2, WSe2), 45 metals, and 6 monolayer lattices (hex, sq, hc and their buckled forms). After relaxation and phonon calculations, they report 1208 dynamically stable MSHs, identify MoSe2 as the most effective sandwich layer and buckled hexagonal as the most stable lattice, validate two systems against DFT phonons, test six selected MSHs with DFT molecular dynamics, and analyze stabilization in terms of sandwich-induced tensile strain and interlayer force constants. The conclusion is that TMD sandwiches, especially MoSe2, stabilize metallenes over a wide composition range and that the stability landscape can guide synthesis.

Significance. If the central result is correct, this is a useful screening map for a large and experimentally relevant materials space: the 1208-count inventory and the MoSe2/buckled-hexagonal rankings are concrete, falsifiable predictions that could guide future synthesis. The paper has clear strengths: the workflow is transparent, the data are deposited, the geometry and phonon comparison for two representative MSHs is a sensible first check, and the strain/force-constant mechanism is physically plausible and consistent with prior work. The main limitation is that the quantitative screening claim rests on the transferability of MatterSim to metal/TMD van der Waals interfaces, and the paper's own DFT validation covers only two hex-lattice systems. Because the central claims are screening counts and rankings, the adequacy of that validation is the load-bearing issue.

major comments (3)
  1. [Results and discussion, Fig. 2 and the 'Dynamical stability analysis' paragraph] The central claim — 1208 stable MSHs and the layer/lattice rankings — depends entirely on MatterSim's ability to classify dynamical stability across 45 metals, 6 lattices, and 6 encapsulants. The only DFT phonon checks are BN|Cu(hex)|BN and MoS2|Mg(hex)|MoS2. Both are planar hexagonal lattices, and no check covers the buckled lattices (bhex, bsq, bhc) or late/post-transition metals (Ru, Rh, Re, Ir, Pt, Bi, Pb, etc.) that dominate the failure cases in Fig. 3. The sentence 'the UMLIP prediction still serves our purpose of distinguishing dynamically unstable MSHs' is asserted but is not backed by a false-positive/false-negative rate or by any DFT-derived stability classification. Please provide a representative DFT validation set (e.g., 10–20 systems spanning stable/unstable, several lattices, and several metal families) and a confusion-matrix-style comparison using the same 0.2 THz thresho
  2. [Results, 'Dynamical stability analysis' and Fig. 3] The 0.2 THz imaginary-frequency cutoff is an unbenchmarked free parameter. In MoS2|Mg(hex)|MoS2, the paper states that DFPT gives small imaginary frequencies 'comparable' to UMLIP but that phonon-band shifts are larger. If UMLIP's error on imaginary modes is of order 0.1–0.5 THz for weakly bound TMD/metal interfaces, many structures near the threshold — likely numerous among post-transition metals and bhc/bsq lattices — would flip between stable and unstable. This would materially change the reported count of 1208 and the MoSe2/bhex rankings. Please report a cutoff-sensitivity analysis (e.g., counts and rankings for thresholds of 0.1, 0.2, and 0.3 THz) and, ideally, validate the threshold against DFT classification for the representative set requested above.
  3. [Stabilization mechanism, Fig. 5 and Eqs. (1)–(4)] The strain and interlayer force-constant analysis is read off the same UMLIP-relaxed geometries and UMLIP force constants that produce the stable/unstable labels. The statement that stable materials have maximum interface force constants in the range 1.48–3.42 eV/Å^2 while unstable cases have near-zero force constants is therefore partly an internal consistency check, not an independent mechanistic test. The ordering of MoSe2 versus graphene in Fig. 5(c) could be an artifact of the same potential bias that drives the stability classification. Please either validate the strain and FC distributions for a subset with DFT (e.g., DFT-relaxed geometries and DFPT force constants) or explicitly state that the mechanism analysis is UMLIP-only and carries the same transferability uncertainty as the screening.
minor comments (4)
  1. [Abstract and main text] There are numerous formatting inconsistencies, including missing spaces in '1620', '45 metals', and 'MoSe 2 asthemosteffectivesandwichlayerforstabilizingmetallenes.' A careful copyedit is needed.
  2. [Fig. 4 caption and text] The caption refers to 'Free energy (eV)' from molecular dynamics simulations. In a finite-temperature MD run, the quantity plotted is total or potential energy, not the Helmholtz free energy. Please clarify which energy is shown and whether the simulations are NVT or NPT.
  3. [Methods/SI] The manuscript repeatedly refers to Supporting Information Figures S1–S8 for atom counts, stability details, and compressive-strain distributions. The arXiv version does not contain the SI. Please ensure the SI is included with the submission and that the main text states the MatterSim model version/checkpoint used, since this is needed for reproducibility.
  4. [Notation] The notation S|M(L)|S is used to define MSHs, but 'L' is introduced as the metallene's lattice and then appears both as a superscript and in textual references. Please define the notation explicitly and use it consistently, including in the figure captions.

Circularity Check

0 steps flagged

No circularity: the central screening is an output of an externally trained MLIP, not a fitted input or self-citation reduction.

full rationale

The paper's central claim—1208 dynamically stable MSHs, MoSe2 as the most effective sandwich layer, and buckled hexagonal lattices as the most stable—is produced by applying the externally trained MatterSim universal MLIP to 1620 constructed heterostructures under a fixed stability criterion (removing structures with imaginary frequencies > 0.2 THz). No parameter is fitted to the stability outcomes, and the DFT/DFPT cross-checks on BN|Cu(hex)|BN and MoS2|Mg(hex)|MoS2 are external validation steps rather than fitting steps. The authors' explicit concession that phonon-band shifts are larger in the MoS2 case, while the UMLIP prediction 'still serves our purpose of distinguishing dynamically unstable MSHs,' is a transferability limitation, not a logical circularity. The strain and interlayer force-constant mechanism is read off the same UMLIP-relaxed geometries and forces, which creates a mild internal correlation with the stability labels, but it does not reduce by construction to those labels: the paper compares stable versus unstable strain distributions and computes sandwich-layer-dependent FC differences using a method fully defined in Eqs. (1)-(4). Self-citations (e.g., refs. 13, 38, and UMLIP-application papers) are supportive, contextual, or methodological; none carries the load of the stability claim, and no uniqueness theorem or ansatz is imported from the authors' prior work as a substitute for evidence. Thus no prediction is an input renamed, and the derivation chain is self-contained with respect to circularity; the principal risk is external validity of MatterSim, which is a correctness concern, not a consistency flaw.

Axiom & Free-Parameter Ledger

2 free parameters · 4 axioms · 0 invented entities

No new particles, forces, or conserved quantities are postulated; the sandwich heterostructures are constructed from existing materials. The main hidden input is the universal MLIP's transferability, which is assumed rather than demonstrated for the full chemical space.

free parameters (2)
  • Imaginary-frequency stability cutoff = 0.2 THz
    Hand-chosen threshold; structures with imaginary phonon frequencies above 0.2 THz are classified unstable, so the 1208 count is sensitive to this value.
  • Maximum lattice mismatch for supercell construction = 3%
    Expanding supercells until lattice mismatch falls below 3% selects commensurate configurations; different tolerances would change the candidate pool and strain distribution.
axioms (4)
  • domain assumption Harmonic phonon imaginary frequencies computed from the MLIP potential determine dynamical stability.
    Central screening methodology; stability classification relies on this standard but approximate criterion (Fig. 1, 'Phonon calculations').
  • domain assumption MatterSim reproduces first-principles forces and phonons for all 45 metals and 6 sandwich layers in the heterostructure library.
    Load-bearing transferability assumption; validated only for two systems (Fig. 2), so accuracy across the full library is assumed.
  • ad hoc to paper An imaginary frequency threshold of 0.2 THz cleanly separates stable from unstable metallene phases.
    Threshold is chosen by the authors; no sensitivity analysis is provided, so weaker instabilities might be mislabeled stable.
  • domain assumption Density-functional theory is an adequate reference for validating the MLIP and for room-temperature MD.
    DFT/DFPT and DFT-MD are standard reference methods; the validation covers only Cu and Mg systems.

reviewed 2026-08-04 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Machine-Learning-Accelerated Metallene Stabilization from High-Throughput Sandwich Modeling." pith.science (2026). https://pith.science/paper/X6NLZ45Y

@misc{pith2026260801779,
  author       = {Pith},
  title        = {Pith review of: Machine-Learning-Accelerated Metallene Stabilization from High-Throughput Sandwich Modeling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/X6NLZ45Y}},
  note         = {Machine review of arXiv:2608.01779}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Metallenes have appealing properties, but stabilizing them in a monolayer phase poses challenges for their synthesis. A recent experiment showed that the van der Waals squeezing method can stabilize certain metallenes in a MoS2 sandwich. This pioneering work motivates systematic studies, but such studies are experimentally impractical, while first-principles modeling remains prohibitive. Here, armed with universal machine-learning interatomic potentials, we constructed 1620 metallene sandwich heterostructures containing 6 different sandwich layers and 45 metals. We performed phonon calculations, which revealed 1208 dynamically stable structures. We found that transition-metal dichalcogenides, particularly MoSe2, are highly effective in stabilizing metallenes. Specifically, buckled hexagonal and honeycomb crystal lattices exhibit the greatest stability. We further evaluated the thermal stability of selected heterostructures with density-functional theory molecular dynamics simulations at room temperature. By uncovering the physical and chemical factors governing the stabilization of metallenes, our results provide systematic insights to guide and accelerate synthesis for future applications.

Figures

Figures reproduced from arXiv: 2608.01779 by Mohammad Bagheri, Pekka Koskinen.

Figure 1
Figure 1. Figure 1: High-throughput calculation workflow for metallene sandwich heterostructures (MSHs). Sandwich layers consist of graphene (C), hexagonal boron nitride (h-BN), and TMDs with the chemical formula of MX2, where M: Mo, W, and X: S, Se). The atomic species colors are indicated in insets. Metallene with three planar lattices: hexagonal (hex), square (sq), and honeycomb (hc), and their buckled forms (bhex, bsq, an… view at source ↗
Figure 2
Figure 2. Figure 2: Comparison of first principles and UMLIP calculations for selected MSHs. Phonon dispersion curves of (a) BN|Cu(hex)|BN supercell, and (b) MoS2|Mg(hex)|MoS2 supercell. The DFPT bands are shown as solid blue lines, and the UMLIP bands as dashed red lines. The chemical structures of sandwich heterostructures with average bond lengths and interlayer distances from DFT (red) and UMLIP (blue) are indicated. crit… view at source ↗
Figure 3
Figure 3. Figure 3: Dynamical stability analysis. (a) Heatmap of the total number of dynamically stable (no imaginary frequency > 0.2 THz) metallene sandwich heterostructures. (b) Number of dynamically stable sandwich heterostructures vs sandwich layer type grouped by metallenes’ lattice. stable Bi metallene with two hc and one hex lattice (Figure S7). We found the hc is the most stable lattice for Ga and Sn (appearing on all… view at source ↗
Figure 4
Figure 4. Figure 4: Finite-temperature structural stability. Density-functional theory molecular dynamics of selected stable MSHs at room temperature for 10 ps. The insets show the side views of the initial (0 ps) and final (10 ps) structures. Using these scalar pairs, the peak interaction strength experienced by any individual atom i is isolated by finding its largest force constant with any distinct neighboring atom Φ max i… view at source ↗
Figure 5
Figure 5. Figure 5: Stabilization mechanism of metallene sandwich heterostructures. (a) Distribution of sandwich-induced strain on the metallene in both stable (blue) and unstable (red) heterostructures. (b) Number of stable structures as a function of Maximum force constants between metal and supporting atom from sandwich layers. (c) Comparison of the maximum force constants between metallene’s (hc) atoms and the supporting … view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

51 extracted references · 43 canonical work pages

  1. [1]

    Realization of 2

    Zhao, Jiaojiao and Li, Lu and Li, Peixuan and Dai, Liyan and Dong, Jingwei and Zhou, Lanying and Wang, Yizhe and Zhang, Peihang and Ji, Kunshan and Zhang, Yangkun and Yu, Hua and Wei, Zheng and Li, Jiawei and Li, Xiuzhen and Huang, Zhiheng and Wang, Boxin and Liu, Jieying and Chen, Yutong and Zhang, Xingchao and Wang, Shuopei and Li, Na and Yang, Wei and ...

    doi:10.1038/s41

  2. [2]

    Nano Letters , volume =

    Zhang, Qian and Yang, Lizhou and Zeng, Chuanjing and Lv, Jirui and Yang, Jinlong and Hu, Wei , title =. Nano Letters , volume =. 2025 , doi =

    work page 2025

  3. [3]

    and Greczynski, Grzegorz and Magnuson, Martin and Andersson, Mike and Rosen, Johanna and Hultman, Lars , journal =

    Kashiwaya, Shun and Shi, Yuchen and Lu, Jun and Sangiovanni, Davide G. and Greczynski, Grzegorz and Magnuson, Martin and Andersson, Mike and Rosen, Johanna and Hultman, Lars , journal =. 2024 , title =. doi:10.1038/s44160-024-00518-4 , pages =

    work page doi:10.1038/s44160-024-00518-4 2024

  4. [4]

    Metallene: Ångstr\"

    Ren, Fengzhu and Han, Zhaoyang and Zhu, Lingfeng and Lei, Zhihao and Shi, Guozheng and Li, Zhixuan and Lin, Chun‐Ho and Hu, Long and Li, Hui and Guan, Xinwei and Jia, Baohua and Kumar, Prashant and Ma, Tianyi , year =. Metallene: Ångstr\". Advanced Materials , publisher =. doi:10.1002/adma.202512683 , number =

    work page doi:10.1002/adma.202512683

  5. [5]

    General-purpose machine-learned potential for 16 elemental metals and their alloys , volume =

    Song, Keke and Zhao, Rui and Liu, Jiahui and Wang, Yanzhou and Lindgren, Eric and Wang, Yong and Chen, Shunda and Xu, Ke and Liang, Ting and Ying, Penghua and Xu, Nan and Zhao, Zhiqiang and Shi, Jiuyang and Wang, Junjie and Lyu, Shuang and Zeng, Zezhu and Liang, Shirong and Dong, Haikuan and Sun, Ligang and Chen, Yue and Zhang, Zhuhua and Guo, Wanlin and ...

  6. [6]

    Journal of Applied Physics , volume =

    Cao, Shuo and Tang, Benrui and Zeng, Zezhu and Fan, Zheyong and Su, Ye and Yan, Yu , title =. Journal of Applied Physics , volume =. 2026 , doi =

    work page 2026

  7. [7]

    Loew, Antoine and Sun, Dewen and Wang, Hai-Chen and Botti, Silvana and Marques, Miguel A. L. , year =. Universal Machine Learning Interatomic Potentials Are Ready for Phonons , volume =. npj Computational Matererials , publisher =. doi:10.1038/s41524-025-01650-1 , pages =

    doi:10.1038/s41524-025-01650-1

  8. [8]

    arXiv , url=

    MatterSim: A Deep Learning Atomistic Model Across Elements, Temperatures and Pressures , author=. arXiv , url=. 2024 , eprint=

    work page 2024

  9. [9]

    Lateral graphene-metallene interfaces at the nanoscale

    Bagheri, Mohammad and Koskinen, Pekka. Lateral graphene-metallene interfaces at the nanoscale. Nanoscale. 2026. doi:10.1039/D5NR02770E

    work page doi:10.1039/d5nr02770e 2026

  10. [10]

    Chemistry of Materials , volume =

    Bagheri, Mohammad and Berger, Ethan and Komsa, Hannu-Pekka and Koskinen, Pekka , title =. Chemistry of Materials , volume =. 2026 , doi =

    work page 2026

  11. [11]

    doi:10.1088/3050-287X/ae1208 , year =

    Universal Machine Learning Potential for Systems with Reduced Dimensionality , author=. doi:10.1088/3050-287X/ae1208 , year =

    work page doi:10.1088/3050-287x/ae1208

  12. [12]

    Dispersion-Corrected Machine Learning Potentials for 2

    Sauer, Mikkel Ohm and Lyngby, Peder Meisner and Thygesen, Kristian Sommer , journal =. Dispersion-Corrected Machine Learning Potentials for 2. 2025 , publisher =

    work page 2025

  13. [13]

    Small , volume =

    Berger, Ethan and Bagheri, Mohammad and Komsa, Hannu-Pekka , title =. Small , volume =. 2025 , doi =

    work page 2025

  14. [14]

    Physical Review B , volume =

    Dynamical stability of two-dimensional metals in the periodic table , author =. Physical Review B , volume =. 2020 , publisher =

    work page 2020

  15. [15]

    Nanoscale , year =

    Abidi, Kameyab Raza and Koskinen, Pekka , title =. Nanoscale , year =

    work page

  16. [16]

    Journal of Materials Chemistry C , year =

    Ren, Yinti and Hu, Liang and Shao, Yangfan and Hu, Yijian and Huang, Li and Shi, Xingqiang , title =. Journal of Materials Chemistry C , year =

    work page

  17. [17]

    doi:10.1088/2632-2153/adfa68 , year =

    Han, Bowen and Cheng, Yongqiang , title =. doi:10.1088/2632-2153/adfa68 , year =

    doi:10.1088/2632-2153/adfa68

  18. [18]

    Journal of Physics: Condensed Matter , volume=

    Ask Hjorth Larsen and Jens Jørgen Mortensen and Jakob Blomqvist and Ivano E Castelli and Rune Christensen and Marcin Dułak and Jesper Friis and Michael N Groves and Bjørk Hammer and Cory Hargus and Eric D Hermes and Paul C Jennings and Peter Bjerre Jensen and James Kermode and John R Kitchin and Esben Leonhard Kolsbjerg and Joseph Kubal and Kristen Kaasbj...

  19. [19]

    Physical Review Letters , volume =

    Structural Relaxation Made Simple , author =. Physical Review Letters , volume =. 2006 , publisher =

    work page 2006

  20. [20]

    Journal of Physics: Condensed Matter , volume =

    Togo, Atsushi and Chaput, Laurent and Tadano, Terumasa and Tanaka, Isao , title =. Journal of Physics: Condensed Matter , volume =. 2023 , doi =

    work page 2023

  21. [21]

    Journal of the Physical Society of Japan , volume =

    Togo, Atsushi , title =. Journal of the Physical Society of Japan , volume =. 2023 , doi =

    work page 2023

  22. [22]

    and Furthm

    Kresse, G. and Furthm. Efficiency of Ab-Initio Total Energy Calculations for Metals and Semiconductors using a Plane-Wave Basis Set , volume =. doi:10.1016/0927-0256(96)00008-0 , journal =

    doi:10.1016/0927-0256(96)00008-0

  23. [23]

    Physical Review B , volume =

    Efficient Iterative Schemes for Ab Initio Total-Energy Calculations using a Plane-Wave Basis Set , author =. Physical Review B , volume =. 1996 , publisher =

    work page 1996

  24. [24]

    Physical Review B , volume =

    Projector Augmented-Wave Method , author =. Physical Review B , volume =. 1994 , doi =

    work page 1994

  25. [25]

    Physical Review B , volume=

    Atlas for the properties of elemental two-dimensional metals , author=. Physical Review B , volume=. 2018 , publisher=

    work page 2018

  26. [26]

    Geim, A. K. and Grigorieva, I. V. , year =. Van der Waals heterostructures , volume =. Nature , publisher =. doi:10.1038/nature12385 , number =

    doi:10.1038/nature12385

  27. [27]

    The Journal of Physical Chemistry Letters , volume =

    Cui, Ziyue and Tang, Qian and Lv, Mingyue and Yu, Guangtao and Chen, Wei , title =. The Journal of Physical Chemistry Letters , volume =. 2026 , doi =

    work page 2026

  28. [28]

    and Majumdar, Arun and Dionne, Jennifer A

    Cui, Yi and Xu, Kun and Ren, Peiwen and Yuan, Lin and Czaja, Pawel and Barnum, Andrew and Sarkar, Prasanna and Altman, Aaron and Bustillo, Karen and Kundu, Sudipta and Ramdas, Akash and Wang, Xueji and Wan, Gang and Wang, Yu and Wang, Jiayue and Song, Chengyu and Lim, Cedric and Zheng, Qi and Yao, Helen and Heinz, Tony and Hwang, Harold Y. and Majumdar, A...

    work page 2026

  29. [29]

    Physical Review Letters , volume =

    Crystal Structure and Pair Potentials: A Molecular-Dynamics Study , author =. Physical Review Letters , volume =. 1980 , publisher =

    work page 1980

  30. [30]

    and Rahman, A

    Parrinello, M. and Rahman, A. , title =. Journal of Applied Physics , volume =. 1981 , issn =

    work page 1981

  31. [31]

    Chemical Society Reviews , year =

    Lu, Chenxin and Li, Ruoyao and Miao, Zhaohua and Wang, Fei and Zha, Zhengbao , title =. Chemical Society Reviews , year =

    work page

  32. [32]

    Journal of Energy Chemistry , year=

    Recent progress in two-dimensional metallenes and their potential application as electrocatalyst , author=. Journal of Energy Chemistry , year=. doi:10.1016/j.jechem.2024.02.068 , publisher=

    work page doi:10.1016/j.jechem.2024.02.068 2024

  33. [33]

    Materials Horizons , year =

    Xie, Minghao and Tang, Sishuang and Zhang, Bowen and Yu, Guihua , title =. Materials Horizons , year =

    work page

  34. [34]

    ACS Materials Letters , volume =

    Liu, Yaoda and Dinh, Khang Ngoc and Dai, Zhengfei and Yan, Qingyu , title =. ACS Materials Letters , volume =. 2020 , doi =

    work page 2020

  35. [35]

    doi:10.1088/2053-1583/adc470 , year =

    Kashiwaya, Shun and Shi, Yuchen and Rosen, Johanna and Hultman, Lars , title =. doi:10.1088/2053-1583/adc470 , year =

    work page doi:10.1088/2053-1583/adc470 2053

  36. [36]

    Nanoscale , year =

    Nevalaita, Janne and Koskinen, Pekka , title =. Nanoscale , year =

    work page

  37. [37]

    and Ta, Huy Quang and Gemming, Thomas and van Gog, Heleen and van Huis, Marijn A

    Mendes, Rafael G. and Ta, Huy Quang and Gemming, Thomas and van Gog, Heleen and van Huis, Marijn A. and Bachmatiuk, Alicja and Rümmeli, Mark H. , title =. Advanced Functional Materials , volume =. doi:10.1002/adfm.202412889 , year =

    work page doi:10.1002/adfm.202412889

  38. [38]

    , title =

    Zhao, Xiaoxu and Dan, Jiadong and Chen, Jianyi and Ding, Zijing and Zhou, Wu and Loh, Kian Ping and Pennycook, Stephen J. , title =. Advanced Materials , volume =. doi:10.1002/adma.201707281 , year =

    work page doi:10.1002/adma.201707281

  39. [39]

    and Mendes, Rafael G

    Zhao, Liang and Ta, Huy Q. and Mendes, Rafael G. and Bachmatiuk, Alicja and Rummeli, Mark H. , title =. Advanced Materials Interfaces , volume =. doi:10.1002/admi.202000436 , year =

    work page doi:10.1002/admi.202000436

  40. [40]

    ACS Nano , volume =

    Zhu, Qi and Hong, Youran and Cao, Guang and Zhang, Yin and Zhang, Xiaohan and Du, Kui and Zhang, Ze and Zhu, Ting and Wang, Jiangwei , title =. ACS Nano , volume =. 2020 , doi =

    work page 2020

  41. [41]

    and Mendes, Rafael G

    Ta, Huy Quang and Yang, Qin Xiao and Liu, Shuyuan and Bachmatiuk, A. and Mendes, Rafael G. and Gemming, Thomas and Liu, Yu and Liu, Lijun and Tokarska, Klaudia and Patel, Rajen B. and Choi, Jin-Ho and R. In Situ Formation of Free-Standing Single-Atom-Thick Antiferromagnetic Chromium Membranes , journal =. 2020 , doi =

    work page 2020

  42. [42]

    Nano Letters , volume =

    Li, Wenbo and Jia, Yanyan and Gao, Mengyue and Huang, Weiping and Wang, Kuan-Wen and Dai, Sheng , title =. Nano Letters , volume =. 2026 , doi =

    work page 2026

  43. [43]

    Nano Letters , volume =

    Jiang, Wenxuan and Wu, Yu and Su, Rina and Xu, Weiqing and Yang, Wenhong and Qiu, Yiwei and Cai, Yujia and Wang, Canglong and Hu, Liuyong and Gu, Wenling and Zhu, Chengzhou , title =. Nano Letters , volume =. 2024 , doi =

    work page 2024

  44. [44]

    Nano Letters , volume =

    Guo, Hongyu and Shi, Jia and Li, Lu and Han, Xiaocang and Shang, Changshuai and Luo, Heng and Cao, Xiaoqing and Tao, Lu and Tan, Hao and Gu, Yu and Qian, Zhengyi and Zhang, Wenyu and Luo, Mingchuan and Zhao, Xiaoxu and Guo, Shaojun , title =. Nano Letters , volume =. 2024 , doi =

    work page 2024

  45. [45]

    The Journal of Physical Chemistry Letters , volume =

    Bagheri, Mohammad and Berger, Ethan and Komsa, Hannu-Pekka , title =. The Journal of Physical Chemistry Letters , volume =. 2023 , doi =

    work page 2023

  46. [46]

    doi:10.1088/2632-2153/adea2d , year =

    Shuang, Fei and Wei, Zixiong and Liu, Kai and Gao, Wei and Dey, Poulumi , title =. doi:10.1088/2632-2153/adea2d , year =

    doi:10.1088/2632-2153/adea2d

  47. [47]

    A Comprehensive Assessment and Benchmark Study of Large Atomistic Foundation Models for Phonons , doi =

    Anam, Md Zaibul and Aghoghovbia, Ogheneyoma and Al‐Fahdi, Mohammed and Kong, Lingyu and Fung, Victor and Hu, Ming , year =. A Comprehensive Assessment and Benchmark Study of Large Atomistic Foundation Models for Phonons , doi =. Advanced Intelligent Discovery , publisher =

    work page

  48. [48]

    Small , volume =

    Kabiraz, Mrinal Kanti and Wahidah, Hafidatul and Hong, Jong Wook and Choi, Sang-Il , title =. Small , volume =. doi:https://doi.org/10.1002/smll.202500858 , year =

    work page doi:10.1002/smll.202500858

  49. [49]

    Harriet , title =

    Joudi, Wael and Ghaderzadeh, Sadegh and Trentino, Alberto and Mizohata, Kenichiro and Mustonen, Kimmo and Besley, Elena and Kotakoski, Jani and Åhlgren, E. Harriet , title =. ACS Nano , volume =. 2025 , doi =

    work page 2025

  50. [50]

    ACS Nano , volume =

    Jiang, Bo and Guo, Yanna and Sun, Fengyu and Wang, Shengyao and Kang, Yunqing and Xu, Xingtao and Zhao, Jingjing and You, Jungmok and Eguchi, Miharu and Yamauchi, Yusuke and Li, Hexing , title =. ACS Nano , volume =. 2023 , doi =

    work page 2023

  51. [51]

    Physical Review B , volume =

    Metallenes: Emergent superconductivity in atomically thin metal layers , author =. Physical Review B , volume =. 2026 , publisher =

    work page 2026

This paper was first reviewed by deepseek-v4-flash on August 4, 2026.