pith. sign in

arxiv: 2605.16091 · v1 · pith:X5DVB4ZEnew · submitted 2026-05-15 · ❄️ cond-mat.dis-nn · cond-mat.mtrl-sci

Machine learning potential as a guide for eutectic in ultra-refractory multicomponent ceramics

V.E. Valiulin , A.V. Mikheyenkov , N.M. Chtchelkatchev , E.A. Levashov This is my paper

Pith reviewed 2026-05-19 17:13 UTC · model grok-4.3

classification ❄️ cond-mat.dis-nn cond-mat.mtrl-sci
keywords machine learning potentialeutectic pointultra-refractory ceramicsTi-B-C systemliquid phase simulationneural network interatomic potentialmulticomponent alloyshigh-temperature materials
0
0 comments X

The pith

A neural-network interatomic potential locates eutectic compositions in ultra-refractory alloys by simulating only the liquid phase.

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

The paper shows that a machine-learning potential can predict eutectic concentrations in alloys whose melting points exceed 3000 K, where direct melting experiments are impractical. It trains the potential on the Ti-B-C system and uses it to derive a concentration criterion that matches known behavior without any input from the solid crystal structure. If the method holds, researchers can screen candidate ultra-refractory multicomponent ceramics computationally first and then target only the most promising compositions for costly high-temperature tests. This shifts the bottleneck from exhaustive trial-and-error melting runs to validation of a few guided points. The approach therefore supplies a practical route to explore new high-temperature materials whose phase diagrams would otherwise remain unknown.

Core claim

The central claim is that a neural-network machine-learning interatomic potential, trained to reproduce ab initio accuracy, can be run entirely in the liquid state to extract a transferable criterion for the eutectic composition in ultra-refractory multicomponent ceramics. Verification on the well-studied Ti-B-C system demonstrates that the liquid-phase simulation alone is sufficient to identify the eutectic point, removing any requirement for prior knowledge of the solid alloy's crystalline structure.

What carries the argument

A neural-network interatomic potential that computes liquid-phase thermodynamics and supplies a concentration criterion for the eutectic point.

If this is right

  • High-temperature experiments can be limited to a narrow composition window around the predicted eutectic instead of broad screening.
  • Phase-diagram exploration becomes feasible for alloys with melting points above 3000 K that cannot be melted in conventional furnaces.
  • The liquid-only workflow can be repeated for additional refractory elements without first solving their solid-state crystal structures.
  • Multicomponent ultra-refractory ceramics can be screened computationally before any synthesis attempt.

Where Pith is reading between the lines

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

  • If the liquid-phase criterion proves robust, it could be embedded in automated composition optimizers that propose new ceramic candidates for extreme environments.
  • The same liquid simulation framework might later be used to estimate other liquidus features such as congruent melting points or metastable extensions in refractory systems.
  • Rapid computational guidance could shorten the discovery cycle for thermal-protection or cutting-tool materials by directing experimental effort to only a few compositions.

Load-bearing premise

The potential trained on the Ti-B-C system accurately reproduces the liquid thermodynamics that determine the true eutectic composition, and the derived criterion carries over to other ultra-refractory multicomponent systems.

What would settle it

Measure the experimental eutectic temperature and composition in the Ti-B-C system (or a second ultra-refractory ternary) and compare it directly with the composition predicted by the liquid-phase simulation; a mismatch larger than the claimed ab initio accuracy would refute the criterion.

Figures

Figures reproduced from arXiv: 2605.16091 by A.V. Mikheyenkov, E.A. Levashov, N.M. Chtchelkatchev, V.E. Valiulin.

Figure 1
Figure 1. Figure 1: FIG. 1. (a) Pseudo-binary sections under consideration are colored as follows: orange for Line 1 and green for Line 2; (b) [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Validation of the machine-learning interatomic po [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Accuracy assessment of the DeepMD-se-a2 MLIP via radial distribution function (RDF) benchmark. Total and partial [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Accuracy assessment of the DeepMD-se-a2 MLIP via velocity autocorrelation function (VACF) benchmark. The [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. The derivative of density by concentration as the [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. The derivative of the total energy by concentration as [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7. Dependence on concentration in Line 1. The experimentally estimated eutectic region [42] is shaded (57 [PITH_FULL_IMAGE:figures/full_fig_p008_7.png] view at source ↗
Figure 9
Figure 9. Figure 9: FIG. 9. The derivative of the total energy by concentration as [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8. The derivative of density by concentration as the [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 10
Figure 10. Figure 10: FIG. 10. Dependence on concentration in Line 2. The experimentally estimated eutectic region [42] is shaded (32 [PITH_FULL_IMAGE:figures/full_fig_p009_10.png] view at source ↗
read the original abstract

The experimental determination of eutectic points is a long-established and widely used technique, but it is generally only practical for systems with relatively low melting points. Many modern, promising materials, however, are ultra-refractory, with melting points exceeding 3000 K. For these systems, conventional melting experiments become prohibitively expensive and technically challenging. Advanced AI modeling can serve as a powerful precursor to guide successful experimentation in such cases. This work proposes a novel criterion for determining the eutectic point concentration in ultra-refractory alloys. The approach is verified using the Ti-B-C system - the most thoroughly studied three-component refractory system to date. The core of the algorithm is a machine-learning interatomic potential, based on a neural network, which achieves accuracy comparable to ab initio methods. Crucially, the algorithm operates effectively in the liquid phase, eliminating the need for information about the solid alloy's crystalline structure to estimate eutectic points.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes a novel criterion for determining eutectic compositions in ultra-refractory multicomponent ceramics. The method centers on a neural-network machine-learning interatomic potential that achieves accuracy comparable to ab initio calculations and operates exclusively in the liquid phase, thereby avoiding any requirement for solid-phase crystalline structure information. The approach is verified on the Ti-B-C system, described as the most thoroughly studied three-component refractory system.

Significance. If the liquid-phase thermodynamics are shown to be accurate, the work would provide a practical computational guide for locating eutectics in materials with melting points above 3000 K, where direct experiments are extremely difficult. The liquid-only formulation is a clear strength that could extend to other multicomponent ultra-refractory systems lacking complete solid-phase data.

major comments (2)
  1. [Abstract] Abstract: the claim that the neural-network potential 'achieves accuracy comparable to ab initio methods' and 'operates effectively in the liquid phase' is load-bearing for the eutectic criterion, yet no validation metrics (e.g., MAE on energies/forces or direct comparison of liquid mixing enthalpies/chemical potentials against AIMD at ~3000 K) are reported for the Ti-B-C melt. Without these, the reliability of the derived eutectic location cannot be assessed.
  2. [Verification on Ti-B-C] Verification on Ti-B-C: the abstract states that the algorithm is verified on this system as an external check, but provides neither the predicted eutectic composition, its deviation from known experimental values, nor the explicit form of the novel criterion (e.g., an equation based on liquid free-energy surfaces or activity coefficients). This prevents evaluation of whether the liquid-only approach reproduces the true eutectic without solid-phase input.
minor comments (2)
  1. The abstract would be strengthened by including at least one quantitative result, such as the computed eutectic composition for Ti-B-C and its agreement with literature.
  2. Consider adding a schematic or equation in the methods section that defines the new eutectic criterion in terms of liquid-phase quantities to make the central advance immediately clear.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments. We address each major comment below and outline the revisions we will implement.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim that the neural-network potential 'achieves accuracy comparable to ab initio methods' and 'operates effectively in the liquid phase' is load-bearing for the eutectic criterion, yet no validation metrics (e.g., MAE on energies/forces or direct comparison of liquid mixing enthalpies/chemical potentials against AIMD at ~3000 K) are reported for the Ti-B-C melt. Without these, the reliability of the derived eutectic location cannot be assessed.

    Authors: We agree that explicit validation metrics strengthen the claim and should be more visible. The full manuscript reports mean absolute errors for energies and forces from the neural-network potential against DFT reference data for the Ti-B-C system, along with direct comparisons of liquid mixing enthalpies obtained from the ML potential versus AIMD trajectories at temperatures near 3000 K. To address the referee's concern, we will revise the abstract to include the key MAE values and a concise statement on the liquid-phase agreement with AIMD. revision: yes

  2. Referee: [Verification on Ti-B-C] Verification on Ti-B-C: the abstract states that the algorithm is verified on this system as an external check, but provides neither the predicted eutectic composition, its deviation from known experimental values, nor the explicit form of the novel criterion (e.g., an equation based on liquid free-energy surfaces or activity coefficients). This prevents evaluation of whether the liquid-only approach reproduces the true eutectic without solid-phase input.

    Authors: We concur that the abstract would benefit from concrete verification details. The novel criterion identifies the eutectic composition as the point of minimum in the liquid chemical-potential surface (or equivalently where the activity coefficients satisfy the two-phase equilibrium condition) computed exclusively from liquid-phase simulations. The results section reports the predicted eutectic composition for Ti-B-C and its deviation from the established experimental value. We will update the abstract to state the predicted composition, the magnitude of its deviation from experiment, and a brief equation or description of the liquid-only criterion. revision: yes

Circularity Check

0 steps flagged

No significant circularity in the ML potential derivation for eutectic prediction

full rationale

The paper trains a neural-network interatomic potential on the Ti-B-C system and applies it to locate eutectic composition via liquid-phase thermodynamics without solid-structure input. Verification on Ti-B-C is presented as an external benchmark rather than a self-referential fit. No equations, criteria, or steps reduce the eutectic prediction to a parameter defined from the same data by construction, nor do they rely on load-bearing self-citations or imported uniqueness theorems. The central claim remains independent of its inputs and is self-contained against the described ab initio-comparable accuracy benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Based on abstract only; no explicit free parameters, axioms, or invented entities are described in the provided text.

pith-pipeline@v0.9.0 · 5718 in / 1091 out tokens · 47941 ms · 2026-05-19T17:13:12.588292+00:00 · methodology

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

97 extracted references · 97 canonical work pages

  1. [1]

    This approach enabled systematic sampling across the entire concentration space

    Active learning To generate initial configurations at each DPGEN iter- ation (Deep Potential GENerator [71]), we employed the Dirichlet-Rescale (DRS) algorithm [72, 73], which pro- duces uniformly distributed random vectors constrained to unit summation. This approach enabled systematic sampling across the entire concentration space. For each DPGEN step, ...

    work page

  2. [2]

    Kurchatov In- stitute

    MLIP verification To validate the accuracy and reliability of our newly developed Deep Potential Molecular Dynamics (DeePMD) potential for molten Ti–B–C systems, we conducted a series of comprehensive benchmark tests. 6 0 1 002 003 004 000.00.20.40.60.81.00 1 02 03 04 05 00123450 1 02 03 04 05 00.00.51.01.52.02.53.03.50 1 02 03 04 05 00.00.51.01.52.02.53....

    work page 2000

  3. [3]

    White, Takashi Goto, and Eliz- abeth C

    Wei-Ting Chen, Ryan M. White, Takashi Goto, and Eliz- abeth C. Dickey. Directionally solidified boride and car- bide eutectic ceramics.J. Am. Ceram. Soc., 99(6):1837– 1851, 2016

    work page 2016

  4. [4]

    Pollock, Dierk Raabe, Marc Andr´ e Meyers, Assel Aitkaliyeva, Kerri-Lee Chintersingh, Zachary C

    Suhas Eswarappa Prameela, Tresa M. Pollock, Dierk Raabe, Marc Andr´ e Meyers, Assel Aitkaliyeva, Kerri-Lee Chintersingh, Zachary C. Cordero, and Lori Graham- Brady. Materials for extreme environments.Nat. Rev. Mater., 8(2):81–88, February 2023

    work page 2023

  5. [5]

    Schreiber, Ruth Schwaiger, Martin Heilmaier, and Scott J

    Daniel K. Schreiber, Ruth Schwaiger, Martin Heilmaier, and Scott J. McCormack. Materials properties charac- terization in the most extreme environments.Mrs Bull., 47(11):1128–1142, November 2022

    work page 2022

  6. [6]

    Fahrenholtz and Greg E

    William G. Fahrenholtz and Greg E. Hilmas. Ultra-high temperature ceramics: Materials for extreme environ- ments.Scripta Mater., 129:94–99, March 2017

    work page 2017

  7. [7]

    Otani, T

    S. Otani, T. Aizawa, and N. Kieda. Solid solution ranges of zirconium diboride with other refractory diborides: HfB2, TiB 2, TaB 2, NbB 2, VB 2 and CrB 2.J. Alloy. Compd., 475(1):273–275, May 2009

    work page 2009

  8. [8]

    V. V. Kurbatkina, E. I. Patsera, D. V. Smirnov, E. A. Levashov, S. Vorotilo, and A. N. Timofeev. Part 2. Structure, mechanical and thermophysical properties of consolidated ceramics based on (Hf,Ta)B 2.Ceram. Int., 45(3):4076–4083, February 2019

    work page 2019

  9. [9]

    Vorotilo, K

    S. Vorotilo, K. Sidnov, V. V. Kurbatkina, P. A. Loginov, E. I. Patsera, T. A. Sviridova, T. A. Lobova, E. A. Lev- ashov, and V. V. Klechkovskaya. Super-hardening and localized plastic deformation behaviors in ZrB 2-TaB2 ce- ramics.J. Alloy. Compd., 901:163368, April 2022

    work page 2022

  10. [10]

    Neuman, Matthew Thompson, William G

    Eric W. Neuman, Matthew Thompson, William G. Fahrenholtz, and Gregory E. Hilmas. Thermal proper- ties of ZrB 2-TiB2 solid solutions.J. Eur. Ceram. Soc., 41(15):7434–7441, December 2021

    work page 2021

  11. [11]

    Synthesis of fine ZrB 2 powders by solid solution of TaB 2 and their densification and mechanical properties.Ceram

    Wei-Ming Guo, Da-Wang Tan, Ling-Yong Zeng, Hao- Li, Hua-Tay Lin, and Cheng-Yong Wang. Synthesis of fine ZrB 2 powders by solid solution of TaB 2 and their densification and mechanical properties.Ceram. Int., 44(4):4473–4477, March 2018

    work page 2018

  12. [12]

    V. I. Ivashchenko, P. E. A. Turchi, V. I. Shevchenko, N. R. Mediukh, Leonid Gorb, and Jerzy Leszczynski. Phase diagram, electronic, mechanical and thermody- namic properties of TiB 2–ZrB2 solid solutions: A first- principles study.Mater. Chem. Phys., 263:124340, April 2021. 11

    work page 2021

  13. [13]

    Solid solutioning in ZrB 2 with HfB 2: Effect on densification and oxidation resistance.Int

    Rubia Hassan, Shobit Omar, and Kantesh Balani. Solid solutioning in ZrB 2 with HfB 2: Effect on densification and oxidation resistance.Int. J. Refract. Met. Hard Mater, 84:105041, November 2019. [12]Ultra-refractory materials, in particular ultra-high tem- perature ceramics (UHTC) are commonly defined as those with melting points higher than 3000 oC (see,...

    work page 2019

  14. [14]

    E. A. Levashov, A. S. Mukasyan, A. S. Rogachev, and D. V. Shtansky. Self-propagating high-temperature syn- thesis of advanced materials and coatings.Int. Mater. Rev., 62(4):203–239, 2017

    work page 2017

  15. [15]

    Levashov Yuri M

    Evgeny A. Levashov Yuri M. Maksimov Alexander S. Mukasyan Alexander S. Rogachev Inna P. Borovinskaya, Alexander A. Gromov.Concise Encyclopedia of Self- Propagating High-Temperature Synthesis. Elsevier, Am- sterdam, 2017

    work page 2017

  16. [16]

    K. Morsi. The diversity of combustion synthesis process- ing: a review.J. Mater. Sci., 47(1):68–92, January 2012

    work page 2012

  17. [17]

    Aruna and Alexander S

    Singanahally T. Aruna and Alexander S. Mukasyan. Combustion synthesis and nanomaterials.Curr. Opin. Solid State Mater. Sci., 12(3):44–50, June 2008

    work page 2008

  18. [18]

    Levashov S

    E.A. Levashov S. Vorotilo, A.Yu. Potanin. SHS of refrac- tory carbides, borides, related solid solutions and com- posite.Ceramics in Modern Technologies, 1(1):25–43, 2019

    work page 2019

  19. [19]

    Vorotilo, K

    S. Vorotilo, K. Sidnov, I. Yu. Mosyagin, A. V. Khvan, E. A. Levashov, E. I. Patsera, and I. A. Abrikosov.Ab- initiomodeling and experimental investigation of proper- ties of ultra-high temperature solid solutions TaxZr1-xC. J. Alloy. Compd., 778:480–486, March 2019

    work page 2019

  20. [20]

    V. V. Kurbatkina, E. I. Patsera, P. A. Loginov, T. A. Sviridova, V. V. Klechkovskaya, and E. A. Levashov. Structure and properties of (Ta1-xZrx)C and (Ta1- xHfx)C solid solutions produced by SHS and hot press- ing.Ceram. Int., 47(18):26205–26214, September 2021

    work page 2021

  21. [21]

    V. V. Kurbatkina, E. I. Patsera, E. A. Levashov, and A. N. Timofeev. Self-propagating high-temperature syn- thesis of single-phase binary tantalum-hafnium carbide (Ta,Hf)C and its consolidation by hot pressing and spark plasma sintering.Ceram. Int., 44(4):4320–4329, March 2018

    work page 2018

  22. [22]

    A. A. Zaitsev, A. Yu. Potanin, Yu. S. Pogozhev, I. O. Filonenko, and E. A. Levashov. Mechanical Activa- tion Assisted Self-Propagating High-Temperature Syn- thesis of HfB2–HfC Composites.Int. J. Self Propag. High Temp. Synth., 32(2):157–168, June 2023

    work page 2023

  23. [23]

    Potanin A.N

    A.Yu. Potanin A.N. Astapov I.O. Vakhrusheva V.V. Korolev S.I. Rupasov E.A. Levashov A.A. Zaitsev, Yu.S. Pogozhev. The structure and properties of the promising ultra-high-temperature hfb2–hfc–sic ceramics obtained from heterophase shs powders.Int. J. Self Propag. High Temp. Synth., 33(2):122–137, 2024

    work page 2024

  24. [24]

    I. V. Iatsyuk, Yu. S. Pogozhev, E. A. Levashov, A. V. Novikov, N. A. Kochetov, and D. Yu. Kovalev. Combus- tion synthesis of high-temperature ZrB2-SiC ceramics.J. Eur. Ceram. Soc., 38(7):2792–2801, July 2018

    work page 2018

  25. [25]

    S. S. Ordan’yan, A. I. Dmitriev, E. K. Stepanenko, N. Yu. Aulova, and N. E. Semenov. SiC-TiB 2 system — A base of high-hardness wear-resistant materials.Sov. Powder Metall. Met. Ceram., 26(5):375–377, May 1987

    work page 1987

  26. [26]

    A study on in-situ syn- thesis of TiB 2–SiC ceramic composites by reactive hot pressing.Ceram

    Guolong Zhao, Chuanzhen Huang, Hanlian Liu, Bin Zou, Hongtao Zhu, and Jun Wang. A study on in-situ syn- thesis of TiB 2–SiC ceramic composites by reactive hot pressing.Ceram. Int., 40(1, Part B):2305–2313, January 2014

    work page 2014

  27. [27]

    Fahrenholtz, and Gregory E

    Alireza Rezaie, William G. Fahrenholtz, and Gregory E. Hilmas. Effect of hot pressing time and tempera- ture on the microstructure and mechanical properties of ZrB2–SiC.J. Mater. Sci., 42(8):2735–2744, April 2007

    work page 2007

  28. [28]

    Zhang, Greg E

    Shi C. Zhang, Greg E. Hilmas, and William G. Fahren- holtz. Mechanical properties of sintered ZrB2–SiC ceram- ics.J. Eur. Ceram. Soc., 31(5):893–901, May 2011

    work page 2011

  29. [29]

    Preparation of WC-SiC Whisker Composites by Hot Pressing and Their Mechanical Properties.Mater

    Shigeaki Sugiyama, Daiki Kudo, and Hitoshi Taimatsu. Preparation of WC-SiC Whisker Composites by Hot Pressing and Their Mechanical Properties.Mater. Trans., 49(7):1644–1649, 2008

    work page 2008

  30. [30]

    Microstructure and Mechanical Properties of WC-SiC Composites

    Akihiro Nino, Yoshinao Nakaibayashi, Shigeaki Sugiyama, and Hitoshi Taimatsu. Microstructure and Mechanical Properties of WC-SiC Composites. Mater. Trans., 52(8):1641–1645, 2011

    work page 2011

  31. [31]

    Wentao Zhou, Jia Song, Lve Lin, Huilong Yang, Shao- qiang Guo, Guang Ran, and Yafei Wang. Amorphiza- tion evolution study of crcofenimn high entropy alloy for mechanical performance optimization by deep potential molecular dynamics.npj Computational Materials, 11(1), March 2025

    work page 2025

  32. [32]

    Hongyang Liu, Bo Chen, Rong Chen, Jingzhi He, Dong- dong Kang, and Jiayu Dai. Computational simulation of short-range order structures in high-entropy alloys: a re- view on formation patterns, multiscale characterization, and performance modulation mechanisms.Advances in Physics: X, 10(1), July 2025

    work page 2025

  33. [33]

    R. L. Ashbrook. Directionally solidified ceramic eutectics. J. Am. Ceram. Soc., 60(9-10):428–435, 1977

    work page 1977

  34. [34]

    Javier Llorca and Victor M. Orera. Directionally solidi- fied eutectic ceramic oxides.Prog. Mater. Sci., 51(6):711– 809, August 2006

    work page 2006

  35. [35]

    L. N. Brewer, M. U. Guruz, and V. P. Dravid. Interfacial fracture mechanisms in solid solution directionally solidi- fied eutectic oxide composites.Acta Mater., 52(13):3781– 3791, August 2004

    work page 2004

  36. [36]

    L. N. Brewer, R. A. Peascoe, C. R. Hubbard, and V. P. Dravid. Residual stress distributions in the solid solution eutectic, Co1-xNixO/ZrO 2(CaO).J. Am. Ceram. Soc., 86(12):2188–2194, 2003

    work page 2003

  37. [37]

    Yu. B. Paderno, V. N. Paderno, V. B. Filippov, Yu. V. Mil’man, and A. N. Martynenko. Structure features of the eutectic alloys of borides with the d- and f-transition metals.Sov. Powder Metall. Met. Ceram., 31(8):700–706, August 1992

    work page 1992

  38. [38]

    Identification of the Eutectic Points in the Multicomponent Systems with the High-Throughput CALPHAD Approach.J

    Jize Zhang, Rui Wang, and Yu Zhong. Identification of the Eutectic Points in the Multicomponent Systems with the High-Throughput CALPHAD Approach.J. Phase Equilib. Diff., 43(6):844–857, December 2022

    work page 2022

  39. [39]

    Sopata, M

    M. Sopata, M. Sadej, and J. Jakubowicz. High temper- ature resistance of novel tantalum-based nanocrystalline refractory compounds.J. Alloy. Compd., 788:476–484, June 2019

    work page 2019

  40. [40]

    Melt/solid interaction and melt super-saturation effect on the microstructure asset of ultra-refractory compos- ites prepared by reactive melting infiltration.J

    Laura Silvestroni, Antonio Vinci, Nicola Gilli, Luca Zoli, Diletta Sciti, Dietmar Koch, and Marius K¨ utemeyer. Melt/solid interaction and melt super-saturation effect on the microstructure asset of ultra-refractory compos- ites prepared by reactive melting infiltration.J. Alloy. 12 Compd., 947:169521, June 2023

    work page 2023

  41. [41]

    Shaikh, B

    Sufyan M. Shaikh, B. S. Murty, and Satyesh K. Yadav. Designing a thermodynamically stable and intrinsically ductile refractory alloy.J. Alloy. Compd., 939:168597, April 2023

    work page 2023

  42. [42]

    Ultrahigh-Temperature Ceramics (UHTCs) Systems

    Zhenkun Huang and Laner Wu. Ultrahigh-Temperature Ceramics (UHTCs) Systems. In Zhenkun Huang and Laner Wu, editors,Phase Equilibria Diagrams of High Temperature Non-oxide Ceramics, pages 103–162. Springer, Singapore, 2018

    work page 2018

  43. [43]

    Dipankar Banerjee and J. C. Williams. Perspectives on Titanium Science and Technology.Acta Mater., 61(3):844–879, February 2013

    work page 2013

  44. [44]

    Momma and F

    K. Momma and F. Izumi. VESTA 3 for three-dimensional visualization of crystal, volumetric and morphology data. J. Appl. Crystallogr., 44(6):1272–1276, December 2011. Publisher: International Union of Crystallography

    work page 2011

  45. [45]

    Molecular dynamics simulation on glass formation, microstructural evolution, and transport properties of cu54hf46 alloy during rapid solidification.J

    Murat Celtek. Molecular dynamics simulation on glass formation, microstructural evolution, and transport properties of cu54hf46 alloy during rapid solidification.J. Chem. Phys., 163(14):144505, 10 2025

    work page 2025

  46. [46]

    In- teraction strength of carbon dioxide on graphene from periodic quantum diffusion monte carlo.J

    Flaviano Della Pia, Giaan Kler-Young, Andrea Zen, Fabian Berger, Dario Alfe, and Angelos Michaelides. In- teraction strength of carbon dioxide on graphene from periodic quantum diffusion monte carlo.J. Chem. Phys., 163(7):071101, 08 2025

    work page 2025

  47. [47]

    Non-collinear ferro- electric h2o ice viii.J

    Hiroshi Fukui and Toshiaki Iitaka. Non-collinear ferro- electric h2o ice viii.J. Chem. Phys., 163(8):084511, 08 2025

    work page 2025

  48. [48]

    Kinetic pathways of coesite densification from metadynamics.J

    David Vrba and Roman Martoˇ n´ ak. Kinetic pathways of coesite densification from metadynamics.J. Chem. Phys., 163(11):114503, 09 2025

    work page 2025

  49. [49]

    Atomsk: A tool for manipulating and converting atomic data files.Comput

    Pierre Hirel. Atomsk: A tool for manipulating and converting atomic data files.Comput. Phys. Commun., 197:212–219, December 2015

    work page 2015

  50. [50]

    Kairi Masuda and Andrew M. Rappe. Phase field model- ing for atoms.Phys. Rev. B, 110(10):104107, September 2024

    work page 2024

  51. [51]

    Basavaraj, and Aditi Ray

    Anuradha Singla, K. Basavaraj, and Aditi Ray. Shock compression behavior and dynamic strength properties of preheated tantalum: Role of crystal anisotropy.Phys. Rev. B, 110(18):184111, November 2024

    work page 2024

  52. [52]

    Highly ac- curate and efficient potential for bcc iron assisted by ar- tificial neural networks.Phys

    Meng Zhang, Koki Hibi, and Junya Inoue. Highly ac- curate and efficient potential for bcc iron assisted by ar- tificial neural networks.Phys. Rev. B, 110(5):054110, August 2024

    work page 2024

  53. [53]

    Deringer, Miguel A

    Volker L. Deringer, Miguel A. Caro, and Gabor Csanyi. Machine learning interatomic potentials as emerging tools for materials science.Adv. Mater., 31(46):1902765, 2019

    work page 2019

  54. [54]

    N. M. Chtchelkatchev, R. E. Ryltsev, M. V. Magnit- skaya, S. M. Gorbunov, K. A. Cherednichenko, V. L. Solozhenko, and V. V. Brazhkin. Local structure, ther- modynamics, and melting of boron phosphide at high pressures by deep learning-driven ab initio simulations. J. Chem. Phys., 159(6):064507, 08 2023

    work page 2023

  55. [55]

    E. O. Khazieva, N. M. Chtchelkatchev, and R. E. Ryltsev. Transfer learning for accurate description of atomic transport in al–cu melts.J. Chem. Phys., 161(17):174101, 11 2024

    work page 2024

  56. [56]

    Ceperley

    Shubhang Goswami, Scott Jensen, Yubo Yang, Markus Holzmann, Carlo Pierleoni, and David M. Ceperley. High temperature melting of dense molecular hydrogen from machine-learning interatomic potentials trained on quan- tum monte carlo.J. Chem. Phys., 162(5):054118, 02 2025

    work page 2025

  57. [57]

    Optimizing ma- chine learning interatomic potentials for hydroxide trans- port: Surprising efficiency of single-concentration train- ing.J

    Jonas H¨ anseroth and Christian Dreßler. Optimizing ma- chine learning interatomic potentials for hydroxide trans- port: Surprising efficiency of single-concentration train- ing.J. Chem. Phys., 163(8):084118, 08 2025

    work page 2025

  58. [58]

    Developing a neural network machine learning interatomic potential for molecular dynamics simulations of la–si–p systems.J

    Ling Tang, Weiyi Xia, Gayatri Viswanathan, Ernesto Soto, Kirill Kovnir, and Cai-Zhuang Wang. Developing a neural network machine learning interatomic potential for molecular dynamics simulations of la–si–p systems.J. Chem. Phys., 163(8):084109, 08 2025

    work page 2025

  59. [59]

    Comparison of intermediate-range order in geo2 glass: Molecular dynamics using machine-learning inter- atomic potential vs reverse monte carlo fitting to exper- imental data.J

    Kenta Matsutani, Shusuke Kasamatsu, and Takeshi Usuki. Comparison of intermediate-range order in geo2 glass: Molecular dynamics using machine-learning inter- atomic potential vs reverse monte carlo fitting to exper- imental data.J. Chem. Phys., 161(20):204103, 11 2024

    work page 2024

  60. [60]

    Deep-learning interatomic potential for iron at extreme conditions.Phys

    Zhi Li and Sandro Scandolo. Deep-learning interatomic potential for iron at extreme conditions.Phys. Rev. B, 109(18):184108, May 2024

    work page 2024

  61. [61]

    Environment- adaptive machine learning potentials.Phys

    Ngoc Cuong Nguyen and Dionysios Sema. Environment- adaptive machine learning potentials.Phys. Rev. B, 110(6):064101, August 2024

    work page 2024

  62. [62]

    Deep learning in- teratomic potential for thermal and defect behaviour of aluminum nitride with quantum accuracy.Comp

    Tao Li, Qing Hou, Jie-chao Cui, Jia-hui Yang, Ben Xu, Min Li, Jun Wang, and Bao-qin Fu. Deep learning in- teratomic potential for thermal and defect behaviour of aluminum nitride with quantum accuracy.Comp. Mater. Sci., 232:112656, January 2024

    work page 2024

  63. [63]

    Machine Learning Interatomic Po- tentials: Keys to First-Principles Multiscale Modeling

    Bohayra Mortazavi. Machine Learning Interatomic Po- tentials: Keys to First-Principles Multiscale Modeling. In Timon Rabczuk and Klaus-J¨ urgen Bathe, editors,Ma- chine Learning in Modeling and Simulation: Methods and Applications, pages 427–451. Springer International Pub- lishing, Cham, 2023

    work page 2023

  64. [64]

    DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics

    Han Wang, Linfeng Zhang, Jiequn Han, and Weinan E. DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics. Comput. Phys. Commun., 228:178–184, July 2018

    work page 2018

  65. [65]

    Lucien Dupuy and Neepa T. Maitra. Exciting deepmd: Learning excited-state energies, forces, and non-adiabatic couplings.J. Chem. Phys., 161(13):134103, 10 2024

    work page 2024

  66. [66]

    Machine learning potential-accelerated multiscale dynamical simulations of nanodiamond structural reconstruction.J

    Rui-Hong He and Jing-Shuang Dang. Machine learning potential-accelerated multiscale dynamical simulations of nanodiamond structural reconstruction.J. Chem. Phys., 163(16):164310, 10 2025

    work page 2025

  67. [67]

    Alou An- gulo, Mathieu Bertin, Jean-Hugues Fillion, Alejandro Rivero Santamar´ ıa, and Maurice Monnerville

    Maxime Infuso, Samuel Del Fre, Gilberto A. Alou An- gulo, Mathieu Bertin, Jean-Hugues Fillion, Alejandro Rivero Santamar´ ıa, and Maurice Monnerville. Deep potential-driven molecular dynamics of co ice analogs: Investigating desorption following vibrational excitation. J. Chem. Phys., 163(8):084303, 08 2025

    work page 2025

  68. [68]

    Understanding the heterogeneous nucleation of ice on silver iodide using deep potential molecular dynamics.J

    Yaochen Yu and Haiyang Niu. Understanding the heterogeneous nucleation of ice on silver iodide using deep potential molecular dynamics.J. Chem. Phys., 163(13):134704, 10 2025

    work page 2025

  69. [69]

    Hafner and G

    J. Hafner and G. Kresse. The Vienna AB-Initio Simula- tion Program VASP: An Efficient and Versatile Tool for Studying the Structural, Dynamic, and Electronic Prop- erties of Materials. In Antonios Gonis, Annemarie Meike, and Patrice E. A. Turchi, editors,Properties of Complex Inorganic Solids, pages 69–82. Springer US, Boston, MA, 1997

    work page 1997

  70. [70]

    End-to-end symmetry preserving inter-atomic potential energy model for fi- 13 nite and extended systems

    Linfeng Zhang, Jiequn Han, Han Wang, Wissam Saidi, Roberto Car, and Weinan E. End-to-end symmetry preserving inter-atomic potential energy model for fi- 13 nite and extended systems. In S. Bengio, H. Wal- lach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, and R. Garnett, editors,Advances in Neural Information Processing Systems, volume 31. Curran Associate...

    work page 2018

  71. [71]

    Dp-gen: A concurrent learning platform for the generation of reliable deep learning based potential energy models.Computer Physics Communications, 253:107206, August 2020

    Yuzhi Zhang, Haidi Wang, Weijie Chen, Jinzhe Zeng, Linfeng Zhang, Han Wang, and Weinan E. Dp-gen: A concurrent learning platform for the generation of reliable deep learning based potential energy models.Computer Physics Communications, 253:107206, August 2020

    work page 2020

  72. [72]

    David Griffin, Iain Bate, and Robert I. Davis. Gener- ating utilization vectors for the systematic evaluation of schedulability tests. In2020 IEEE Real-Time Systems Symposium (RTSS), pages 76–88, 2020

    work page 2020

  73. [73]

    Work-in-progress: Evaluation framework for self-suspending schedulability tests

    Mario G¨ unzel, Harun Teper, Kuan-Hsun Chen, Georg von der Br¨ uggen, and Jian-Jia Chen. Work-in-progress: Evaluation framework for self-suspending schedulability tests. In2021 IEEE Real-Time Systems Symposium (RTSS), pages 532–535, 2021

    work page 2021

  74. [74]

    Jinzhe Zeng, Duo Zhang, Denghui Lu, Pinghui Mo, Zeyu Li, Yixiao Chen, Marian Rynik, Liang Huang, Ziyao Li, Shaochen Shi, Yingze Wang, Haotian Ye, Ping Tuo, Ji- abin Yang, Ye Ding, Yifan Li, Davide Tisi, Qiyu Zeng, Han Bao, Yu Xia, Jiameng Huang, Koki Muraoka, Yibo Wang, Junhan Chang, Fengbo Yuan, Sigbjorn Loland Bore, Chun Cai, Yinnian Lin, Bo Wang, Jiaya...

    work page 2023

  75. [75]

    Giese, Weile Jia, Ben Xu, Darrin M

    Jinzhe Zeng, Duo Zhang, Anyang Peng, Xiangyu Zhang, Sensen He, Yan Wang, Xinzijian Liu, Hangrui Bi, Yi- fan Li, Chun Cai, Chengqian Zhang, Yiming Du, Jia- Xin Zhu, Pinghui Mo, Zhengtao Huang, Qiyu Zeng, Shaochen Shi, Xuejian Qin, Zhaoxi Yu, Chenxing Luo, Ye Ding, Yun-Pei Liu, Ruosong Shi, Zhenyu Wang, Sigb- jorn Loland Bore, Junhan Chang, Zhe Deng, Zhaoha...

    work page 2025

  76. [76]

    Kresse and D

    G. Kresse and D. Joubert. From ultrasoft pseudopoten- tials to the projector augmented-wave method.Physical Review B, 59(3):1758–1775, January 1999

    work page 1999

  77. [77]

    McGill, and Tim Mueller

    Pandu Wisesa, Kyle A. McGill, and Tim Mueller. Ef- ficient generation of generalized monkhorst-pack grids through the use of informatics.Physical Review B, 93(15):155109, April 2016

    work page 2016

  78. [78]

    Rapid generation of optimal generalized monkhorst-pack grids.Computa- tional Materials Science, 187:110100, February 2021

    Yunzhe Wang, Pandu Wisesa, Adarsh Balasubramanian, Shyam Dwaraknath, and Tim Mueller. Rapid generation of optimal generalized monkhorst-pack grids.Computa- tional Materials Science, 187:110100, February 2021

    work page 2021

  79. [79]

    Anubhav Jain, Shyue Ping Ong, Geoffroy Hautier, Wei Chen, William Davidson Richards, Stephen Dacek, Shreyas Cholia, Dan Gunter, David Skinner, Gerbrand Ceder, and Kristin A. Persson. Commentary: The ma- terials project: A materials genome approach to accel- erating materials innovation.APL Materials, 1(1), July 2013

    work page 2013

  80. [80]

    arXiv e-prints , keywords =

    Duo Zhang, Anyang Peng, Chun Cai, Wentao Li, Yuan- chang Zhou, Jinzhe Zeng, Mingyu Guo, Chengqian Zhang, Bowen Li, Hong Jiang, Tong Zhu, Weile Jia, Lin- feng Zhang, and Han Wang. A graph neural network for the era of large atomistic models.arXiv:2506.01686, 2025

    work page arXiv 2025

Showing first 80 references.