pith. sign in

arxiv: 2601.10174 · v2 · pith:LJMDRMC7new · submitted 2026-01-15 · ❄️ cond-mat.mtrl-sci

A Neuroevolution Potential for Gallium Oxide: Accurate and Efficient Modeling of Polymorphism and Swift Heavy-Ion Irradiation

Yaohui Gu , Binbo Li , Lingyang Jiang , Yuhui Hu , Wenqiang Liu , Lijun Xu , Pengfei Zhai , Jie Liu
show 1 more author
Jinglai Duan
This is my paper

Pith reviewed 2026-05-21 15:22 UTC · model grok-4.3

classification ❄️ cond-mat.mtrl-sci
keywords gallium oxideneuroevolution potentialmachine learning interatomic potentialswift heavy-ion irradiationphase transformationpolymorphismbeta-Ga2O3ion track
0
0 comments X

The pith

Neuroevolution potential for Ga2O3 matches experiments on phase transformations in ion tracks.

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

The paper develops a machine-learning interatomic potential for gallium oxide using the neuroevolution potential framework to address challenges in modeling its polymorphic behavior under nonequilibrium conditions. It combines this with an energy-dependent weighting strategy and a physically process-oriented sampling method to augment training data for targeted phenomena. The resulting potential outperforms the tabGAP approach in accuracy and speed, and when applied to swift heavy-ion irradiation of beta-Ga2O3, the simulations achieve quantitative agreement with experiments while explaining reported discrepancies in phase transformations.

Core claim

A robust NEP potential for Ga2O3 is constructed that shows clear advantages over tabGAP in accuracy and computational efficiency; a dedicated version for SHI irradiation simulations of beta-Ga2O3 produces results in quantitative agreement with experimental observations and supplies a consistent physical explanation for discrepancies in reported phase transformations within the ion track.

What carries the argument

Neuroevolution potential (NEP) framework with energy-dependent weighting strategy and physically process-oriented sampling to augment the training dataset for irradiation-specific configurations.

If this is right

  • The potential supports large-scale simulations of nonequilibrium phase changes in Ga2O3 that were previously inaccessible.
  • It offers a unified explanation that reconciles conflicting experimental reports on ion-track phase behavior.
  • The approach enables efficient modeling of polymorphism and radiation effects for Ga2O3 device applications.
  • Direct corollaries include improved predictions of structural evolution under varying irradiation conditions.

Where Pith is reading between the lines

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

  • The sampling strategy may transfer to modeling radiation damage in other wide-bandgap oxides with complex polymorphs.
  • Higher efficiency could permit simulations over longer timescales to capture post-irradiation annealing.
  • Quantitative match suggests the potential can forecast outcomes for irradiation parameters not yet tested experimentally.

Load-bearing premise

The physically process-oriented sampling strategy sufficiently covers the atomic configurations encountered during swift heavy-ion irradiation without requiring significant extrapolation beyond the fitted data.

What would settle it

Experimental data on phase transformation sequences or final phases inside the ion track of beta-Ga2O3 that deviate from the simulated trajectories would falsify the claimed quantitative agreement.

Figures

Figures reproduced from arXiv: 2601.10174 by Binbo Li, Jie Liu, Jinglai Duan, Lijun Xu, Lingyang Jiang, Pengfei Zhai, Wenqiang Liu, Yaohui Gu, Yuhui Hu.

Figure 1
Figure 1. Figure 1: Detailed comparison of the prediction accuracy [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: Comparison of the MAE of (a) energy, (b) force, and (c) virial predictions by NEP and tabGAP for different config types in the GAP dataset. correctly captures their relative energetics and thermo￾dynamic similarity. In contrast, the accurate description of the δ phase but the substantial overprediction of the γ phase by tabGAP suggest that tabGAP may fail to reproduce the thermodynamic proximity between th… view at source ↗
Figure 6
Figure 6. Figure 6: Energy predictions for configurations sampled from [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 4
Figure 4. Figure 4: (a) Illustration of the energy-dependent weighting strategy, where the black curve represents the rescaling fac￾tor as a function of the average potential energy. All samples in the GAP dataset are shown as dots colored according to their configuration types; their positions along the horizontal axis indicate their average potential energies. (b) Comparison of the energy-prediction performance of tabGAP, N… view at source ↗
Figure 5
Figure 5. Figure 5: Energy-volume curves for the α, β, δ, γ, and κ phases predicted by NEP (solid circles) and tabGAP (solid squares), compared with DFT reference data (open circles). Different phases are color-coded as indicated [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: Analysis of a representative ion track in [PITH_FULL_IMAGE:figures/full_fig_p008_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Transverse and longitudinal cross-sectional views of [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Comparison of the track diameters predicted in this [PITH_FULL_IMAGE:figures/full_fig_p009_9.png] view at source ↗
read the original abstract

Gallium oxide (Ga2O3) is a wide-bandgap semiconductor with promising applications in high-power and high-frequency electronics. However, its complex polymorphic nature poses substantial challenges for fundamental studies, particularly in understanding phase-transformation behaviors under nonequilibrium conditions. Here, we develop a robust, accurate, and computationally efficient machine-learning interatomic potential (MLIP) for Ga2O3 based on the neuroevolution potential (NEP) framework combined with an energy-dependent weighting strategy. The resulting NEP potential demonstrates clear advantages over the state-of-the-art tabGAP potential with respect to both accuracy and computational efficiency. Furthermore, we introduce a physically process-oriented sampling strategy to systematically augment the training dataset, thereby enhancing the MLIP performance for targeted physical phenomena. As a representative application, a dedicated NEP potential is constructed for swift heavy-ion (SHI) irradiation simulations of \b{eta}-Ga2O3. The simulated results are in quantitative agreement with experimental observations and provide a consistent physical explanation for the reported experimental discrepancies regarding phase transformations in the ion track of \b{eta}-Ga2O3.

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 develops a neuroevolution potential (NEP) machine-learning interatomic potential for Ga2O3 that incorporates an energy-dependent weighting strategy and a physically process-oriented sampling approach to augment the training set. It reports superior accuracy and efficiency relative to the tabGAP potential, and applies a dedicated NEP to swift heavy-ion irradiation simulations of β-Ga2O3, claiming that the results are in quantitative agreement with experiments and provide a physical explanation for reported discrepancies in phase transformations within the ion track.

Significance. If the central claims are substantiated, the work supplies an efficient, accurate MLIP for large-scale simulations of polymorphism and nonequilibrium radiation damage in a technologically relevant wide-bandgap oxide. The targeted, process-oriented sampling strategy represents a methodological advance that could generalize to other materials under extreme conditions. Reproducible code and parameter-free aspects of the NEP framework, if provided, would further strengthen the contribution to computational materials science.

major comments (2)
  1. [SHI irradiation simulations] Section describing the SHI irradiation application and validation: the central claim of quantitative agreement with experimental phase-transformation outcomes in the β-Ga2O3 ion track lacks reported force or energy RMSE on hold-out snapshots extracted from the production irradiation trajectories. Without such metrics, it is impossible to verify that the potential does not extrapolate in the high-energy, disordered configurations inside the track, which directly undermines the strength of the agreement.
  2. [Training dataset and sampling strategy] Training dataset augmentation subsection: the physically process-oriented sampling strategy is presented as key to performance, yet no quantitative test (e.g., distribution overlap, maximum force deviation, or uncertainty quantification on ion-track-like configurations) is shown to confirm coverage of transient melt or defect-cascade environments. This is load-bearing for the assertion that the NEP accurately models the observed phase transformations without significant extrapolation.
minor comments (2)
  1. [Results figures] Figure captions for the irradiation results could explicitly state the simulation cell size, timestep, and electronic stopping model used, to allow direct comparison with experimental track radii.
  2. [Abstract and Results] The abstract states 'quantitative agreement' but the main text would benefit from a dedicated table comparing simulated and experimental phase fractions or track dimensions with uncertainties.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the positive evaluation of our work and the constructive major comments. We address each point below and describe the revisions we will implement to strengthen the validation of the NEP potential.

read point-by-point responses

  1. Referee: [SHI irradiation simulations] Section describing the SHI irradiation application and validation: the central claim of quantitative agreement with experimental phase-transformation outcomes in the β-Ga2O3 ion track lacks reported force or energy RMSE on hold-out snapshots extracted from the production irradiation trajectories. Without such metrics, it is impossible to verify that the potential does not extrapolate in the high-energy, disordered configurations inside the track, which directly undermines the strength of the agreement.

    Authors: We agree that explicit error metrics on hold-out snapshots drawn directly from the production irradiation trajectories would provide a more direct assessment of extrapolation risk in the high-energy, disordered regimes. The original manuscript validated the dedicated NEP through reproduction of experimental phase-transformation outcomes and through the inclusion of high-energy configurations generated via process-oriented sampling. To address the concern, we will extract representative snapshots from the irradiation trajectories, compute energy and force RMSE values on these hold-out sets, and report the results together with the existing experimental comparisons in the revised manuscript. revision: yes

  2. Referee: [Training dataset and sampling strategy] Training dataset augmentation subsection: the physically process-oriented sampling strategy is presented as key to performance, yet no quantitative test (e.g., distribution overlap, maximum force deviation, or uncertainty quantification on ion-track-like configurations) is shown to confirm coverage of transient melt or defect-cascade environments. This is load-bearing for the assertion that the NEP accurately models the observed phase transformations without significant extrapolation.

    Authors: The process-oriented sampling was constructed by augmenting the dataset with configurations obtained from preliminary high-temperature and high-energy molecular-dynamics runs that target the transient melt and defect-cascade states encountered during swift heavy-ion irradiation. While the submitted manuscript emphasized the resulting improvement in predictive accuracy for the target phenomena, it did not include explicit quantitative diagnostics such as force-distribution overlap or maximum deviations on ion-track-like configurations. We will add these quantitative tests, including comparisons of force histograms and reported maximum deviations for configurations representative of the ion-track interior, to the revised manuscript to substantiate the coverage achieved by the sampling strategy. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained

full rationale

The paper constructs an NEP MLIP by fitting to reference calculations (with energy-dependent weighting and process-oriented sampling augmentation), then deploys the resulting potential in independent SHI irradiation simulations of β-Ga2O3. The reported phase-transformation outcomes and quantitative agreement with experiment are forward applications rather than quantities that reduce by construction to the training inputs or to any self-citation. No load-bearing step equates a prediction to a fitted parameter, imports uniqueness via author citation, or renames a known result; the central claim therefore rests on external validation against experimental data.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the assumption that DFT reference data plus the chosen sampling strategy produce a transferable potential; many internal parameters of the NEP model are fitted to that data.

free parameters (1)
  • NEP model parameters
    Hyperparameters and coefficients in the neuroevolution potential are fitted to reference calculations.
axioms (1)
  • domain assumption Density functional theory calculations supply sufficiently accurate reference energies and forces for training.
    Standard premise in machine-learning interatomic potential development.

pith-pipeline@v0.9.0 · 5765 in / 1185 out tokens · 122114 ms · 2026-05-21T15:22:14.327821+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

59 extracted references · 59 canonical work pages

  1. [1]

    As shown in Fig

    and added to the augmented training dataset. As shown in Fig. 6(c), the NEP model trained on the aug- mented dataset exhibits significantly improved perfor- mance across the entire energy range. The overall MAE is reduced to 8.8 meV/atom, which is approximately half of that obtained with models trained on the original GAP dataset. Although the softening pr...

    work page

  2. [2]

    Based on the transverse cross-sectional view [Fig

    05 keV/nm in TREKIS), with a total simulation time of 200 ps. Based on the transverse cross-sectional view [Fig. 7(a)], the atomic configuration is manually divided into four concentric regions from the center to the pe- riphery: the amorphous core (region I), the γ phase (re- gion II), a deformed β phase (region III), and the pris- tine β phase (region IV...

    work page

  3. [3]

    The f-sum rule: 2 π Ω 2 p ∫ ∞ Ip Im( − 1 ϵ(ω, 0) )shellω dω = Ne,shell (A4) where Ω 2 p = 4 πe 2nat/m e is the plasma frequency, Ne,shell is the number of electrons in selected shell, and Ip is ionization potential

    work page

  4. [4]

    The KK-sum rule: 2 π ∫ ∞ 0 Im( − 1 ϵ(ω, 0) ) dω ω = 1 (A5) The free-electron approximation is used to calculate the dispersion relation for the oscillator energy Ei and trans- ferred momentum q: Ei(q) = Ei(0) + ℏ2q2 2me (A6) The lower and upper limits of the transferred energy dur- ing an inelastic scattering event are as follows: W− = Ip (A7) W+ = 4Em1m2...

    work page

  5. [5]

    S. J. Pearton, J. Yang, P. H. Cary, F. Ren, J. Kim, M. J. Tadjer, and M. A. Mastro, A review of Ga 2O3 materi- als, processing, and devices, Appl. Phys. Rev. 5, 011301 (2017)

    work page 2017

  6. [6]

    M. J. Tadjer, Toward gallium oxide power electronics, Science 378, 724 (2022)

    work page 2022

  7. [7]

    A. S. Pratiyush, S. Krishnamoorthy, R. Muralidharan, S. Rajan, and D. N. Nath, Advances in Ga 2O3 solar-blind uv photodetectors, in Gallium Oxide (Elsevier, 2019) pp. 369–399

    work page 2019

  8. [8]

    J. Zhu, Z. Xu, S. Ha, D. Li, K. Zhang, H. Zhang, and J. Feng, Gallium Oxide for Gas Sensor Applications: A Comprehensive Review, Materials 15, 7339 (2022)

    work page 2022

  9. [9]

    Yoshioka, H

    S. Yoshioka, H. Hayashi, A. Kuwabara, F. Oba, K. Mat- sunaga, and I. Tanaka, Structures and energetics of Ga 2 O3 polymorphs, J. Phys.: Condens. Matter 19, 346211 (2007)

    work page 2007

  10. [10]

    H. He, R. Orlando, M. A. Blanco, R. Pandey, E. Amza- llag, I. Baraille, and M. R´ erat, First-principles study of the structural, electronic, and optical properties of Ga2O3 in its monoclinic and hexagonal phases, Phys. Rev. B 74, 195123 (2006)

    work page 2006

  11. [11]

    J. Xu, J. Zhai, and X. Yao, Structure and dielectric properties of barium titanate thin films grown by sol- gel-hydrothermal process, Appl. Phys. Lett. 89, 252902 (2007)

    work page 2007

  12. [12]

    J. B. Varley, J. R. Weber, A. Janotti, and C. G. Van De Walle, Oxygen vacancies and donor impurities in β - Ga2O3, Appl. Phys. Lett. 97, 142106 (2010)

    work page 2010

  13. [13]

    Yan and S

    Z. Yan and S. Kumar, Phonon mode contributions to thermal conductivity of pristine and defective β -Ga2O3, Phys. Chem. Chem. Phys. 20, 29236 (2018)

    work page 2018

  14. [14]

    J. Yang, Y. Xu, X. Wang, X. Zhang, Y. He, and H. Sun, Lattice thermal conductivity of β -, α - and κ-Ga2O3 : a first-principles computational study, Appl. Phys. Express 17, 011001 (2024)

    work page 2024

  15. [15]

    Bermudez, The structure of low-index surfaces of β - Ga2O3, Chem

    V. Bermudez, The structure of low-index surfaces of β - Ga2O3, Chem. Phys. 323, 193 (2006)

    work page 2006

  16. [16]

    Q. Fan, R. Zhao, W. Zhang, Y. Song, M. Sun, and U. Schwingenschl¨ ogl, Low-energy Ga 2O3 polymorphs with low electron effective masses, Phys. Chem. Chem. Phys. 24, 7045 (2022)

    work page 2022

  17. [17]

    Azarov, C

    A. Azarov, C. Bazioti, V. Venkatachalapathy, P. Vajee- ston, E. Monakhov, and A. Kuznetsov, Disorder-Induced Ordering in Gallium Oxide Polymorphs, Phys. Rev. Lett. 128, 015704 (2021)

    work page 2021

  18. [18]

    X. Han, Y. Liu, Y. Li, M. L. Crespillo, E. Zarkadoula, W. Mu, and P. Liu, Unraveling the Atomic Mechanism of the Crystalline Phase-Dependent Structural Features and Special Spectral Design of α -, β -, and ϵ-Ga2O3, Adv. Sci. 12, e08207 (2025)

    work page 2025

  19. [19]

    M. A. Blanco, M. B. Sahariah, H. Jiang, A. Costales, and R. Pandey, Energetics and migration of point defects in Ga2O3, Phys. Rev. B 72, 184103 (2005)

    work page 2005

  20. [20]

    E. O. Pyzer-Knapp, J. W. Pitera, P. W. J. Staar, S. Takeda, T. Laino, D. P. Sanders, J. Sexton, J. R. Smith, and A. Curioni, Accelerating materials discovery using artificial intelligence, high performance computing and robotics, npj Comput. Mater. 8, 84 (2022)

    work page 2022

  21. [21]

    T. W. Ko and S. P. Ong, Recent advances and outstand- ing challenges for machine learning interatomic poten- tials, Nat. Comput. Sci. 3, 998 (2023)

    work page 2023

  22. [22]

    R. Li, Z. Liu, A. Rohskopf, K. Gordiz, A. Henry, E. Lee, and T. Luo, A deep neural network interatomic poten- tial for studying thermal conductivity of β -Ga2O3, Appl. Phys. Lett. 117, 152102 (2020)

    work page 2020

  23. [23]

    Liu, J.-Y

    Y.-B. Liu, J.-Y. Yang, G.-M. Xin, L.-H. Liu, G. Cs´ anyi, and B.-Y. Cao, Machine learning interatomic potential developed for molecular simulations on thermal proper- ties of β -Ga2O3, J. Chem. Phys. 153, 144501 (2020)

    work page 2020

  24. [24]

    J. Zhao, J. Byggm¨ astar, Z. Zhang, F. Djurabekova, K. Nordlund, and M. Hua, Phase transition of two- dimensional ferroelectric and paraelectric Ga 2O3 mono- layers: A density functional theory and machine learning study, Phys. Rev. B 104, 054107 (2021)

    work page 2021

  25. [25]

    Z. Sun, Z. Qi, K. Liang, X. Sun, Z. Zhang, L. Li, Q. Wang, G. Zhang, G. Wu, and W. Shen, A neuroevo- lution potential for predicting the thermal conductivity of α -Ga2O3, β -Ga2O3, and ε-Ga2O3, Appl. Phys. Lett. 123, 192202 (2023)

    work page 2023

  26. [26]

    X. Wang, J. Yang, P. Ying, Z. Fan, J. Zhang, and H. Sun, Dissimilar thermal transport properties in κ- Ga2O3 and β -Ga2O3 revealed by homogeneous nonequi- librium molecular dynamics simulations using machine- learned potentials, J. Appl. Phys. 135, 065104 (2024)

    work page 2024

  27. [27]

    Y. Liu, H. Liang, L. Yang, G. Yang, H. Yang, S. Song, Z. Mei, G. Cs´ anyi, and B. Cao, Unraveling Ther- mal Transport Correlated with Atomistic Structures in Amorphous Gallium Oxide via Machine Learning Combined with Experiments, Adv. Mater. 35, 2210873 (2023)

    work page 2023

  28. [28]

    J. Zhao, J. Byggm¨ astar, H. He, K. Nordlund, F. Djurabekova, and M. Hua, Complex Ga 2O3 poly- morphs explored by accurate and general-purpose machine-learning interatomic potentials, npj Comput. Mater. 9, 159 (2023)

    work page 2023

  29. [29]

    H. He, J. Zhao, J. Byggm¨ astar, R. He, K. Nord- lund, C. He, and F. Djurabekova, Threshold displace- ment energy map of Frenkel pair generation in β -Ga2O3 from machine-learning-driven molecular dynamics simu- lations, Acta Mater. 276, 120087 (2024)

    work page 2024

  30. [30]

    R. He, J. Zhao, J. Byggm¨ astar, H. He, and F. Djurabekova, Ultrahigh stability of oxygen sublattice in β -Ga2O3, Phys. Rev. Materials 8, 084601 (2024)

    work page 2024

  31. [31]

    T. Liu, Z. Li, J. Zhao, X. Fei, J. Feng, Y. Zuo, M. Hua, Y. Guo, S. Liu, and Z. Zhang, Orientation-dependent sur- face radiation damage in β -Ga2O3 explored by atomistic simulations, Acta Mater. 300, 121484 (2025)

    work page 2025

  32. [32]

    Azarov, C

    A. Azarov, C. Radu, A. Galeckas, I. F. Mercioniu, A. Cernescu, V. Venkatachalapathy, E. Monakhov, F. Djurabekova, C. Ghica, J. Zhao, and A. Kuznetsov, Self-Assembling of Multilayered Polymorphs with Ion Beams, Nano Lett. 25, 1637 (2025)

    work page 2025

  33. [33]

    Zhang, J

    J. Zhang, J. Zhao, J. Chen, and M. Hua, Orientation- dependent atomic-scale mechanism and defect evolution in β -Ga2O3 thin film epitaxial growth, Appl. Phys. Lett. 124, 022102 (2023)

    work page 2023

  34. [34]

    Q. Li, J. Zhao, N. Lin, X. Cheng, X. Zhao, Z. Liu, Z. Jia, and M. Hua, Edge-Dependent Step-Flow Growth Mech- anism in β -Ga2O3 (100) Facet at the Atomic Level, J. Phys. Chem. Lett. 16, 5101 (2025). 13

    work page 2025

  35. [35]

    Azarov, J

    A. Azarov, J. G. Fern´ andez, J. Zhao, F. Djurabekova, H. He, R. He, Ø. Prytz, L. Vines, U. Bek- tas, P. Chekhonin, N. Klingner, G. Hlawacek, and A. Kuznetsov, Universal radiation tolerant semiconduc- tor, Nat. Commun. 14, 4855 (2023)

    work page 2023

  36. [36]

    J. Zhao, J. G. Fern´ andez, A. Azarov, R. He, Ø. Prytz, K. Nordlund, M. Hua, F. Djurabekova, and A. Kuznetsov, Crystallization Instead of Amorphization in Collision Cascades in Gallium Oxide, Phys. Rev. Lett. 134, 126101 (2025)

    work page 2025

  37. [37]

    X. Han, Y. Li, M. L. Crespillo, E. Zarkadoula, Y. Liu, W. Mu, S. Zhao, and P. Liu, Electronic Excitation- Driven β -Ga2 O3 Metastability Transformation and Self-Organization Mechanism: β → κ/γ/δ Phases, Adv. Mater. , e19259 (2025)

    work page 2025

  38. [38]

    Abdullaev, J

    A. Abdullaev, J. G. Fernandez, C. Nozais, J. O’Connell, R. Tlegenov, K. Sekerbayev, A. Azarov, A. Leino, T. F. Bouvier, J. Zhao, A. A. Pena, N. Medvedev, Z. Utegulov, O. Prytz, F. Djurabekova, and A. Kuznetsov, Ions leav- ing no tracks (2025), version Number: 2

    work page 2025

  39. [39]

    K. Xu, H. Bu, S. Pan, E. Lindgren, Y. Wu, Y. Wang, J. Liu, K. Song, B. Xu, Y. Li, T. Hainer, L. Svens- son, J. Wiktor, R. Zhao, H. Huang, C. Qian, S. Zhang, Z. Zeng, B. Zhang, B. Tang, Y. Xiao, Z. Yan, J. Shi, Z. Liang, J. Wang, T. Liang, S. Cao, Y. Wang, P. Ying, N. Xu, C. Chen, Y. Zhang, Z. Chen, X. Wu, W. Jiang, E. Berger, Y. Li, S. Chen, A. J. Gabouri...

    work page 2025

  40. [40]

    Z. Fan, Z. Zeng, C. Zhang, Y. Wang, K. Song, H. Dong, Y. Chen, and T. Ala-Nissila, Neuroevolution machine learning potentials: Combining high accuracy and low cost in atomistic simulations and application to heat transport, Phys. Rev. B 104, 104309 (2021)

    work page 2021

  41. [41]

    Z. Fan, Y. Wang, P. Ying, K. Song, J. Wang, Y. Wang, Z. Zeng, K. Xu, E. Lindgren, J. M. Rahm, A. J. Gabourie, J. Liu, H. Dong, J. Wu, Y. Chen, Z. Zhong, J. Sun, P. Erhart, Y. Su, and T. Ala-Nissila, GPUMD: A package for constructing accurate machine-learned po- tentials and performing highly efficient atomistic simula- tions, J. Chem. Phys. 157, 114801 (2022)

    work page 2022

  42. [42]

    K. Song, R. Zhao, J. Liu, Y. Wang, E. Lindgren, Y. Wang, S. Chen, K. Xu, T. Liang, P. Ying, N. Xu, Z. Zhao, J. Shi, J. Wang, S. Lyu, Z. Zeng, S. Liang, H. Dong, L. Sun, Y. Chen, Z. Zhang, W. Guo, P. Qian, J. Sun, P. Erhart, T. Ala-Nissila, Y. Su, and Z. Fan, General-purpose machine-learned potential for 16 ele- mental metals and their alloys, Nat. Commun....

    work page 2024

  43. [43]

    Liang, K

    T. Liang, K. Xu, E. Lindgren, Z. Chen, R. Zhao, J. Liu, E. Berger, B. Tang, B. Zhang, Y. Wang, K. Song, P. Ying, N. Xu, H. Dong, S. Chen, P. Erhart, Z. Fan, T. Ala- Nissila, and J. Xu, NEP89: Universal neuroevolution po- tential for inorganic and organic materials across 89 ele- ments (2025), arXiv:2504.21286 [cond-mat]

    work page arXiv 2025

  44. [44]

    R. A. Rymzhanov, N. Medvedev, and A. E. Volkov, Dam- age threshold and structure of swift heavy ion tracks in Al2O3, J. Phys. D: Appl. Phys. 50, 475301 (2017)

    work page 2017

  45. [45]

    N. A. Medvedev, R. A. Rymzhanov, and A. E. Volkov, Time-resolved electron kinetics in swift heavy ion irradi- ated solids, J. Phys. D: Appl. Phys. 48, 355303 (2015)

    work page 2015

  46. [46]

    Rymzhanov, N

    R. Rymzhanov, N. Medvedev, and A. Volkov, Effects of model approximations for electron, hole, and photon transport in swift heavy ion tracks, Nucl. Instrum. Meth- ods Phys. Res., Sect. B 388, 41 (2016)

    work page 2016

  47. [47]

    B. Deng, Y. Choi, P. Zhong, J. Riebesell, S. Anand, Z. Li, K. Jun, K. A. Persson, and G. Ceder, Systematic soften- ing in universal machine learning interatomic potentials, npj Comput. Mater. 11, 9 (2024)

    work page 2024

  48. [48]

    C. Chen, Y. Li, R. Zhao, Z. Liu, Z. Fan, G. Tang, and Z. Wang, NepTrain and NepTrainKit: Automated Active Learning and Visualization Toolkit for Neuroevolution Potentials (2025), arXiv:2506.01868 [cs]

    work page arXiv 2025

  49. [49]

    L. Xu, P. Zhai, P. Hu, S. Zhang, J. Zeng, Z. Li, W. Li, M. E. Toimil-Molares, C. Trautmann, and J. Liu, Ther- mal stability of swift heavy ion tracks in β -Ga2O3 an- nealed at high temperatures, Appl. Phys. Lett. 127, 151903 (2025)

    work page 2025

  50. [50]

    W. Ai, L. Xu, S. Nan, P. Zhai, W. Li, Z. Li, P. Hu, J. Zeng, S. Zhang, L. Liu, Y. Sun, and J. Liu, Radiation damage in β -Ga2O3 induced by swift heavy ions, Jpn. J. Appl. Phys. 58, 120914 (2019)

    work page 2019

  51. [51]

    L. E. Ratcliff, T. Oshima, F. Nippert, B. M. Janzen, E. Kluth, R. Goldhahn, M. Feneberg, P. Mazzolini, O. Bierwagen, C. Wouters, M. Nofal, M. Albrecht, J. E. N. Swallow, L. A. H. Jones, P. K. Thakur, T. Lee, C. Kalha, C. Schlueter, T. D. Veal, J. B. Varley, M. R. Wagner, and A. Regoutz, Tackling Disorder in γ-Ga2O3, Adv. Mater. 34, 2204217 (2022)

    work page 2022

  52. [52]

    Huang, C.-N

    Q.-S. Huang, C.-N. Li, M.-S. Hao, H.-P. Liang, X. Cai, Y. Yue, A. Kuznetsov, X. Zhang, and S.-H. Wei, Nature of Disordering in γ-Ga2O3, Phys. Rev. Lett. 133, 226101 (2024)

    work page 2024

  53. [53]

    H. Y. Playford, A. C. Hannon, E. R. Barney, and R. I. Walton, Structures of Uncharacterised Polymorphs of Gallium Oxide from Total Neutron Diffraction, Chem. - Eur. J. 19, 2803 (2013)

    work page 2013

  54. [54]

    K. Xu, G. Wang, T. Liang, Y. Xiao, D. Ding, H. Guo, X. Gao, L. Tong, X. Wan, G. Zhang, and J. Xu, Device- Scale Atomistic Simulations of Heat Transport in Ad- vanced Field-Effect Transistors (2025), arXiv:2511.18915 [cond-mat]

    work page arXiv 2025

  55. [55]

    Gervais and S

    B. Gervais and S. Bouffard, Simulation of the primary stage of the interaction of swift heavy ions with con- densed matter, Nucl. Instrum. Methods Phys. Res., Sect. B 88, 355 (1994)

    work page 1994

  56. [56]

    T. M. Jenkins, W. R. Nelson, and A. Rindi, Monte Carlo transport of electrons and photons , Vol. 38 (Springer Sci- ence & Business Media, 2012)

    work page 2012

  57. [57]

    Wright, E

    D. Wright, E. Guzman, M. S. H. Bijoy, R. B. Wil- son, D. H. Mudiyanselage, H. Fu, F. Kargar, and A. A. Balandin, Acoustic phonon characteristics of (001) and (

    work page

  58. [58]

    β -Ga2O3 single crystals investigated with brillouin–mandelstam light scattering spectroscopy (2025)

    work page 2025

  59. [59]

    A Neuroevolution Potential for Gallium Oxide: Accurate and Efficient Modeling of Polymorphism and Swift Heavy-Ion Irradia tion

    Y. Zhang, M. Liu, D. Jena, and G. Khalsa, Tight-binding band structure of β - and α -phase Ga 2O3 and Al 2O3, J. Appl. Phys. 131, 175702 (2022). Supplemental Material for “A Neuroevolution Potential for Gallium Oxide: Accurate and Efficient Modeling of Polymorphism and Swift Heavy-Ion Irradia tion” Yaohui Gu, 1, 2 Binbo Li, 1, 2, ∗ Linyang Jiang, 1, 2 Yuhui...

    work page 2022