Pith. sign in

REVIEW 2 major objections 4 minor 143 references

Point defect formation at finite temperatures with machine learning force fields

T0 review · 2 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Thermal entropy, not just the 0 K internal energy, sets the formation free energy of the tellurium interstitial in CdTe, raising its predicted equilibrium concentration by roughly two orders of magnitude.

desk verdict Very solid benchmark of finite-temperature defect free energies; the main thermal-effect claim survives even if the TI migration-entropy handling needs a fix. read the letter →

arxiv 2412.16741 v1 pith:FSISAG7P submitted 2024-12-21 cond-mat.mtrl-sci physics.chem-ph

Irea Mosquera-Lois , Johan Klarbring , Aron Walsh This is my paper · ORCID
classification cond-mat.mtrl-sciphysics.chem-ph
keywords pointdefectsdefectformationfreeenergymachinelearningforcefieldsfinite-temperaturethermodynamicsthermodynamicintegrationCdTetelluriuminterstitialconfigurationalentropy
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper argues that the usual shortcut for predicting point-defect concentrations—using the 0 K internal energy of the defect's ground-state structure as a stand-in for the formation free energy—can fail badly for defects that are dynamic at operating temperatures. Using CdTe as a test case, it trains a machine-learned force field to follow the tellurium interstitial $\mathrm{Te_i^{+1}}$, which flips between two near-degenerate configurations, migrates between sites, and rotates its Te–Te bond on the nanosecond timescale at 300 K. Combining harmonic, quasiharmonic, and fully anharmonic thermodynamic-integration free energies with analytical estimates of spin, electronic, orientational, and configurational entropy, the paper finds that the formation free energy $g_f$ at the 840 K annealing temperature is about 0.5 eV below the 0 K value $u_f(0\,\mathrm{K})$, raising the predicted interstitial concentration by roughly two orders of magnitude. By contrast, the tellurium vacancy $V_{\mathrm{Te}}^{+2}$, which has no accessible metastable states and no symmetry-breaking, shows a negligible thermal correction. The message is that finite-temperature entropy dominates defect thermodynamics precisely when a defect reconstructs its local structure and has low-energy competing configurations.

What carries the argument

The central machinery is a machine-learned force field trained on DFT data for each species in the defect reaction, used as a surrogate to run nanosecond NPT molecular dynamics and thermodynamic integration. The free-energy calculation proceeds by non-equilibrium thermodynamic integration—a reversible-switching method that starts from an Einstein crystal of independent harmonic oscillators, integrates to the anharmonic crystal at 100 K, and then sweeps temperature up to 840 K—with the electronic and spin entropies added analytically; separately, the harmonic and quasiharmonic vibrational free energies are obtained from phonon calculations. The defect's active degrees of freedom (configurational, orientational, and migrational) are the mechanism that generates the entropy, and the paper's decomposition shows the vibrational contribution dominates.

What would settle it

Repeat the fully anharmonic thermodynamic-integration calculation for $\mathrm{Te_i^{+1}}$ at 840 K using direct DFT energies on the same 65-atom supercell and compare the formation free energy with the machine-learned result; a disagreement larger than about $0.02$ eV per supercell (the reported TI error) would undermine the claimed 0.5 eV thermal shift and the roughly 500-fold concentration increase. An independent check would be a temperature-dependent experimental measurement of interstitial concentration in CdTe, if the predicted slope of $\ln[c]$ versus $1/T$ can be resolved.

Watch

Extended reading notes

Core claim

For a defect whose local geometry is bistable—$\mathrm{Te_i^{+1}}$ in CdTe, with a $\mathrm{C}_{2v}$ split-interstitial ground state and a $C_s$ metastable configuration only 18 meV higher—the thermal formation free energy is not the 0 K internal energy. Molecular dynamics with a machine-learned force field shows the interstitial changes configuration, hops between lattice sites, and reorients within nanoseconds at 300 K, with barriers of 28–100 meV. Summing the entropic contributions, the vibrational term dominates ($4.2\,k_{\mathrm{B}}$ at 840 K), followed by structural and spin terms; the net effect lowers $g_f$ by 0.5 eV relative to $u_f(0\,\mathrm{K})$ and multiplies the predicted equilibrium concentration by a factor ~500. The anharmonic thermodynamic-integration result agrees with the harmonic approximation to within the computed error, because anharmonic contributions largely cancel between the defective and pristine supercells. $V_{\mathrm{Te}}^{+2}$ behaves classically: it stays in its $T_d$ ground state and its formation free energy shifts by only 0.08 eV, showing the contrast is defect-specific, not a general failure of the 0 K picture.

Load-bearing premise

The calculation rests on the machine-learned force field reproducing the interstitial's energy landscape accurately enough to resolve free-energy differences of a few hundredths of an electron-volt, and on the assumption that defect migration during thermodynamic integration contributes little to the free energy.

Editorial extensions

If this is right

  • The predicted equilibrium concentration of the tellurium interstitial in CdTe at annealing temperatures is roughly two orders of magnitude higher than the static 0 K estimate, so relative defect populations and derived carrier concentrations shift accordingly.
  • The 0 K description remains adequate for defects like the tellurium vacancy that keep a single high-symmetry ground state and lack low-energy metastable configurations.
  • For dynamic defects, the harmonic approximation to the vibrational formation entropy matches the fully anharmonic thermodynamic-integration result at 840 K, validating cheaper phonon-based studies in similar cases.
  • High-temperature applications—thermoelectrics, catalysts, fuel cells—are where thermal corrections to defect formation are large enough to matter, and charge-transition levels may also move with temperature.
  • Machine-learned force fields need training sets that include both pristine and defective supercells if they are to be used for absolute defect formation energies in larger cells.

Reading between the lines

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

  • This result suggests that defect-population models which solve for the Fermi level self-consistently will see temperature-dependent shifts in charge-state transitions for any defect with an accessible metastable manifold, not just in CdTe.
  • A concrete extension would be to run the same protocol on a reconstructive defect in a soft lattice such as a halide perovskite, where anharmonic cancellation between bulk and defect is less likely and the harmonic approximation may break down.
  • Because the electronic entropy term was evaluated with a functional that underestimates the band gap, the small electronic contribution is the least certain part of the entropy budget; training the force field on hybrid-functional data would test whether that term changes the total formation free energy by more than the reported thermodynamic-integration error.
  • The paper's recommendation to train on both pristine and defective supercells has a practical consequence: defect-focused machine-learned force fields trained only on defective cells may be unsuitable for computing absolute formation energies in larger supercells.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 4 minor

Summary. The paper trains MACE machine learning force fields on DFT data for CdTe and two charged defects, Te_i^+1 and V_Te^+2, and uses them to compute defect formation free energies at finite temperature. The authors compare harmonic, quasiharmonic, and fully anharmonic (thermodynamic integration) treatments and add electronic, spin, orientational, and structural entropy contributions. They report that Te_i^+1 is dynamically active at 300 K (configurational, orientational, and migrational motions) and that its formation free energy at 840 K differs by about 0.5 eV from the static 0 K internal-energy estimate, increasing the predicted concentration by a factor of about 500. For V_Te^+2, entropic effects are found to be negligible. The central claim is that finite-temperature entropy dominates the defect formation thermodynamics for defects that undergo structural reconstructions and have low-energy metastable configurations.

Significance. If the quantitative claim is correct, the paper provides a strong case against the standard static 0 K approximation for defect concentrations and demonstrates a practical MLFF-based route to anharmonic defect free energies. The study is carefully executed in many respects: the force fields are validated against independent DFT test sets, the potential-energy surface along the configurational path matches DFT, thermodynamic-integration convergence is checked with low standard errors, and the computed Te melting point (704 K) is close to experiment (722 K). The authors also make transparent use of open-source tools (doped, ShakeNBreak, calphy, phonopy). However, the central factor-of-500 concentration enhancement rests on the relation between the unconstrained thermodynamic-integration free energy and the fixed-site formation free energy used in Eq. (4). If the TI free energy includes the site-to-site migration entropy, the reported enhancement may be largely an artifact of double-counting the site multiplicity. This issue is load-bearing for the paper's main conclusion and must be resolved before the quantitative claims can be accepted.

major comments (2)
  1. [Methods, Thermodynamic integration; Eq. (4)] The thermodynamic integration is performed on an unconstrained 65-atom supercell containing one Te_i^+1 interstitial, and the Methods explicitly note that defect diffusion occurs during the temperature-scaling runs. For a defect that can hop among the M=32 primitive cells of the 2x2x2 conventional supercell and N_sites=12 symmetry-equivalent interstitial sites per primitive cell, the unconstrained partition function contains a factor M*N_sites=384 relative to a defect pinned at a single site. This contributes a site entropy of k_B T ln(384) ≈ 0.43 eV at 840 K that is included in g_TI_f. Equation (4) then multiplies by N_sites/V, counting the site degeneracy a second time. The statement that the migration contribution is small because the defect spends most time near local minima is not a valid statistical-mechanical argument: the free-energy weight of additional basins is determined by their number and phase-space volume, not by the residence time in transition states. Since the double-counting factor is of the same order as the reported factor-of-500 concentration enhancement, the central claim of the paper is not established by the present calculations.
  2. [Fig. 3b and Eq. (3)] The agreement between the harmonic and anharmonic free energies in Fig. 3b is cited as validation of the harmonic approximation and of the decoupling assumption. However, if the anharmonic TI free energy includes the site-to-site migration entropy discussed above, it cannot agree with the harmonic free energy, which is computed for a single fixed interstitial site with phonopy. The authors need to reconcile this contradiction: either the TI path does not actually sample the full site multiplicity (in which case the 'fully anharmonic' label and the diffusion statement in Methods are misleading), or the harmonic calculation inadvertently contains the same multiplicity. A direct test would be to compute the TI free energy with the interstitial restrained to one lattice site (e.g., a harmonic umbrella restraint) and compare it with the unrestrained value; the difference quantifies the migration entropy and should be removed from g_TI_f before applying Eq. (4).
minor comments (4)
  1. [Abstract and Section II.C] The abstract states that thermal effects increase the predicted concentration by 'two orders of magnitude', while the main text reports a factor of 500; please use a single, consistent quantitative statement.
  2. [Eq. (4)] The text defines N_sites as the number of symmetry-equivalent sites per primitive cell but calls V the 'crystallographic unit cell volume'. For CdTe the conventional cell is four times the primitive cell, so the concentration would be off by a factor of four if the conventional cell volume were used. Please specify explicitly which volume is used and keep the definitions consistent.
  3. [Methods, Structural entropy] The analytical structural entropy gives s_struc(840 K) = 0.6 k_B (reported as 0.7 k_B in Fig. 3a), while the inherent-structures method gives 1.05 k_B. The text says these are in the same order, but the discrepancy is nearly a factor of two; a brief justification for using the analytical value and an estimate of its uncertainty would strengthen the analysis.
  4. [Methods, Electronic entropy] The electronic entropy calculation relies on a mid-gap Fermi level for the defect supercell and an assumed excess electron concentration n = 10^15 cm^-3. Since the electronic entropy contribution is small (0.1 k_B), this does not affect the conclusions, but a sensitivity test or justification for these choices would be useful.

Circularity Check

1 steps flagged · score 7.0 of 10

Central factor-500 concentration increase reduces to a site-degeneracy double count between the diffusion-containing thermodynamic-integration free energy and the Nsites prefactor in Eq. (4).

  1. other [Methods, Thermodynamic integration (migration note) and Eq. (4) in Results II C; site multiplicities in Methods, Defect calculations]
    "We note that during the temperature scaling runs of the interstitial, defect diffusion occurs within the simulation timescale. ... We expect that their contribution is small as the defects spend more time around their local minima configurations. ... We calculate the equilibrium defect concentration with [c] = Nsites/V exp(-gf/kBT) ... The site multiplicities per primitive cell are 12 and 24 for the C2v and Cs configurations, respectively."

    The TI path is run on a 64-atom (M=32 primitive-cell) supercell in which the interstitial hops between sites, so its partition function already includes a factor M×Nsites (32×12=384 for C2v): g_TI_f ≈ g_f^site − kBT ln(384) ≈ g_f^site − 0.43 eV at 840 K. Eq. (4) then multiplies by Nsites/V, re-adding the site degeneracy the TI already counted: c ∝ M Nsites^2 exp(−g_f^site/kBT). The spurious factor exp(0.43 eV/kBT) ≈ 400 is the same order as the reported 500× enhancement, so the headline number is built into the ensemble plus prefactor rather than derived from the physics. The paper's claim that the migration contribution is small refers to time spent in barriers, not to the ln(M Nsites) multiplicity of minima, and the harmonic–anharmonic agreement in Fig.

full rationale

The paper's quantitative headline is not self-contained: the anharmonic formation free energy is obtained from thermodynamic-integration runs in which the Te_i+1 interstitial demonstrably migrates between lattice sites, so the computed free energy includes the site-selection entropy of the supercell (≈kBT ln(M Nsites)); Eq. (4) then multiplies by Nsites as if g_f were the fixed-site value. This double counting injects ≈0.43 eV at 840 K, i.e., a factor ≈400 in concentration, which accounts for most of the claimed 500× increase. The remaining machinery is genuinely non-circular: the MACE force fields are trained on DFT energies/forces and tested on independent DFT configurations; phonon and TI free energies are actual calculations; spin, electronic, and orientational entropy estimates follow standard external formalisms; and the structural entropy is cross-checked by three methods. The self-citations (doped, ShakeNBreak, Ref. 2 formalism, Kavanagh et al.'s Te_i+1 characterization) are tool/background citations, not the source of the final free-energy values, and do not by themselves raise the score. Because the central factor-500 claim reduces largely to the definitional double count of the same site degeneracy, the circularity score is 7 rather than a lower value; the harmonic-anharmonic agreement cited as validation needs re-examination in light of the extra site entropy in the TI path.

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

The central claim rests on the accuracy of the PBEsol DFT reference and the trained MACE potentials, plus standard free-energy approximations (harmonic and quasiharmonic phonons, fixed-DOS electronic entropy, additive entropy decomposition). The dominant vibrational term is computed both harmonically and with thermodynamic integration, with good agreement. No target-result fitting or invented entities are used; the main free choices are the Fermi-level placement and electron concentration in the small electronic-entropy term.

free parameters (2)
  • Fermi level EF for Te_i+1 electronic entropy = 0.37 eV above VBM
    Set from the defect's empty state 0.75 eV above VBM; affects electronic entropy, which is a small contribution (0.1 kB at 840 K). Not fitted to the target concentration.
  • Excess electron concentration n = 1e15 cm^-3
    Assumed dilute-limit value in computing the entropy of the added conduction-band electron (Methods, Electronic entropy). Small effect on final free energy.
assumptions (5)
  • domain assumption PBEsol DFT accurately describes the relevant defect potential energy surfaces and vibrational properties of CdTe.
    All MLFF training and validation reference data use PBEsol; hybrid HSE06 is used only for electronic entropy. Reference 12 reported the same defect configurations with HSE06.
  • domain assumption MACE force fields faithfully represent the DFT potential energy surface over the sampled temperature range (100-900 K).
    Supported by test-set MAE/RMSE and potential energy surface path agreement, but unsampled transition states and rare configurations are not proven error-free.
  • domain assumption Ionic, electronic, spin, orientational, and structural degrees of freedom are separable.
    Equation (3) decomposes the free energy; the anharmonic TI includes ionic contributions and the authors compare with the decoupled harmonic result to justify the approximation.
  • domain assumption Electronic entropy can be evaluated with a temperature-independent density of states and assumed Fermi levels.
    Methods, Electronic entropy: EF is set mid-gap or from the defect level, and the excess electron concentration is assumed to be 1e15 cm^-3.
  • domain assumption The Te reservoir model including liquid Te is accurate for the chemical potential in the defect reaction.
    The Te model covers solid and liquid phases and reproduces the melting point at 704 K (experiment 722 K).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Point defect formation at finite temperatures with machine learning force fields." pith.science (2026). https://pith.science/paper/FSISAG7P

@misc{pith2026241216741,
  author       = {Pith},
  title        = {Pith review of: Point defect formation at finite temperatures with machine learning force fields},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FSISAG7P}},
  note         = {Machine review of arXiv:2412.16741}
}
abstract

Point defects dictate the properties of many functional materials. The standard approach to modelling the thermodynamics of defects relies on a static description, where the change in Gibbs free energy is approximated by the internal energy. This approach has a low computational cost, but ignores contributions from atomic vibrations and structural configurations that can be accessed at finite temperatures. We train a machine learning force field (MLFF) to explore dynamic defect behaviour using $\mathrm{Te_i^{+1}}$ and $\textit{V}{\mathrm{_{Te}^{+2}}}$ in CdTe as exemplars. We consider the different entropic contributions (e.g., electronic, spin, vibrational, orientational, and configurational) and compare methods to compute the defect free energies, ranging from a harmonic treatment to a fully anharmonic approach based on thermodynamic integration. We find that metastable configurations are populated at room temperature and thermal effects increase the predicted concentration of $\mathrm{Te_i^{+1}}$ by two orders of magnitude -- and can thus significantly affect the predicted properties. Overall, our study underscores the importance of finite-temperature effects and the potential of MLFFs to model defect dynamics at both synthesis and device operating temperatures.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

143 extracted references · 67 canonical work pages

  1. [1]

    A. M. Stoneham, Theory of defects in solids, Oxford University Press, 1975

    1975

  2. [2]

    Mosquera-Lois, S

    I. Mosquera-Lois, S. R. Kavanagh, J. Klarbring, K. Tolborg and A. Walsh, Chem. Soc. Rev., 2023, 52, 5812--5826

    2023

  3. [3]

    Freysoldt, B

    C. Freysoldt, B. Grabowski, T. Hickel, J. Neugebauer, G. Kresse, A. Janotti and C. G. Van De Walle, Rev. Mod. Phys., 2014, 86, 253--305

    2014

  4. [4]

    Mosquera-Lois, S

    I. Mosquera-Lois, S. R. Kavanagh, A. Walsh and D. O. Scanlon, npj Comput. Mater., 2023, 9, 1--11

    2023

  5. [5]

    Mosquera-Lois and S

    I. Mosquera-Lois and S. R. Kavanagh, Matter, 2021, 4, 2602--2605

    2021

  6. [6]

    X. Wang, S. R. Kavanagh, D. O. Scanlon and A. Walsh, Phys. Rev. B, 2023, 108, 134102

    2023

  7. [7]

    X. Wang, S. R. Kavanagh and A. Walsh, ACS Energy Lett., 0, 0, 161--167

  8. [8]

    X. Wang, S. R. Kavanagh, D. O. Scanlon and A. Walsh, Joule, 2024, 2105--2122

    2024

Show all 143 references
  1. [9]

    A. G. Squires, L. Ganeshkumar, C. N. Savory, S. R. Kavanagh and D. O. Scanlon, ACS Energy Lett., 2024, 9, 4180--4187

    2024

  2. [10]

    J.-H. Yang, L. Shi, L.-W. Wang and S.-H. Wei, Sci Rep, 2016, 6, 21712

    2016

  3. [11]

    Alkauskas, C

    A. Alkauskas, C. E. Dreyer, J. L. Lyons and C. G. Van de Walle, Phys. Rev. B, 2016, 93, 201304

    2016

  4. [12]

    S. R. Kavanagh, D. O. Scanlon, A. Walsh and C. Freysoldt, Faraday Discuss., 2022, 239, 339--356

    2022

  5. [13]

    S. R. Kavanagh, A. Walsh and D. O. Scanlon, ACS Energy Lett., 2021, 1392--1398

    2021

  6. [14]

    S. R. Kavanagh, R. S. Nielsen, J. L. Hansen, R. S. Davidsen, O. Hansen, A. E. Samli, P. C. Vesborg, D. O. Scanlon and A. Walsh, Intrinsic point defect tolerance in selenium for indoor and tandem photovoltaics, Preprint at 10.26434/chemrxiv-2024-91h02 (2024)

    2024 doi

  7. [15]

    D. Guo, C. Qiu, K. Yang and H.-X. Deng, Phys. Rev. Appl., 2021, 15, 064025

    2021

  8. [16]

    W. B. Fowler, M. Stavola, A. Venzie and A. Portoff, J. Appl. Phys., 2024, 135, 170901

    2024

  9. [17]

    A. S. Bochkarev, A. van Roekeghem, S. Mossa and N. Mingo, Phys. Rev. Materials, 2019, 3, 093803

    2019

  10. [18]

    Cheng and M

    B. Cheng and M. Ceriotti, Phys. Rev. B, 2018, 97, 054102

    2018

  11. [19]

    Chiesa, P

    S. Chiesa, P. M. Derlet and S. L. Dudarev, Phys. Rev. B, 2009, 79, 214109

    2009

  12. [20]

    de Koning, C

    M. de Koning, C. R. Miranda and A. Antonelli, Phys. Rev. B, 2002, 66, 104110

    2002

  13. [21]

    De Koning, S

    M. De Koning, S. Ramos De Debiaggi and A. Monti, DDF, 2003, 224-225, 59--74

    2003

  14. [22]

    Glensk, B

    A. Glensk, B. Grabowski, T. Hickel and J. Neugebauer, Phys. Rev. X, 2014, 4, 011018

    2014

  15. [23]

    Grabowski, L

    B. Grabowski, L. Ismer, T. Hickel and J. Neugebauer, Phys. Rev. B, 2009, 79, 134106

    2009

  16. [24]

    Lucas and R

    G. Lucas and R. Schäublin, Nucl. Instrum. Methods Phys. Res., 2009, 267, 3009--3012

    2009

  17. [25]

    J. Luo, C. Zhou, Y. Cheng, Q. Li, L. Liu, J. F. Douglas and T. Sinno, Phys. Rev. Mater., 2022, 6, 064603

    2022

  18. [26]

    T. A. Mellan, A. I. Duff, B. Grabowski and M. W. Finnis, Phys. Rev. B, 2019, 100, 024303

    2019

  19. [27]

    E. V. Safonova, Y. P. Mitrofanov, R. A. Konchakov, A. Y. Vinogradov, N. P. Kobelev and V. A. Khonik, J. Phys.: Condens. Matter, 2016, 28, 215401

    2016

  20. [28]

    Satta, F

    A. Satta, F. Willaime and S. de Gironcoli, Phys. Rev. B, 1998, 57, 11184--11192

    1998

  21. [29]

    G. S. Smirnov and V. V. Stegailov, J. Phys.: Condens. Matter, 2019, 31, 235704

    2019

  22. [30]

    Shin and C

    D. Shin and C. Wolverton, Acta Mater., 2012, 60, 5135--5142

    2012

  23. [31]

    Y. Gong, B. Grabowski, A. Glensk, F. Körmann, J. Neugebauer and R. C. Reed, Phys. Rev. B, 2018, 97, 214106

    2018

  24. [32]

    Smirnova, S

    D. Smirnova, S. Starikov, G. D. Leines, Y. Liang, N. Wang, M. N. Popov, I. A. Abrikosov, D. G. Sangiovanni, R. Drautz and M. Mrovec, Phys. Rev. Mater., 2020, 4, 013605

    2020

  25. [33]

    Zhang, B

    X. Zhang, B. Grabowski, T. Hickel and J. Neugebauer, Comput. Mater. Sci., 2018, 148, 249--259

    2018

  26. [34]

    Mathes, T

    L. Mathes, T. Gigl, M. Leitner and C. Hugenschmidt, Phys. Rev. B, 2020, 101, 134105

    2020

  27. [35]

    Sinno, Z

    T. Sinno, Z. K. Jiang and R. A. Brown, Appl. Phys. Lett., 1996, 68, 3028--3030

    1996

  28. [36]

    O. K. Al-Mushadani and R. J. Needs, Phys. Rev. B, 2003, 68, 235205

    2003

  29. [37]

    Rauls and T

    E. Rauls and T. Frauenheim, Phys. Rev. B, 2004, 69, 155213

    2004

  30. [38]

    P. E. Blöchl, E. Smargiassi, R. Car, D. B. Laks, W. Andreoni and S. T. Pantelides, Phys. Rev. Lett., 1993, 70, 2435--2438

    1993

  31. [39]

    Maroudas and R

    D. Maroudas and R. A. Brown, Phys. Rev. B, 1993, 47, 15562--15577

    1993

  32. [40]

    M. I. Mendelev and Y. Mishin, Phys. Rev. B, 2009, 80, 144111

    2009

  33. [41]

    P. J. Ungar, T. Halicioglu and W. A. Tiller, Phys. Rev. B, 1994, 50, 7344--7357

    1994

  34. [42]

    Wynblatt, J

    P. Wynblatt, J. Phys. Chem. Solids, 1969, 30, 2201--2211

    1969

  35. [43]

    J. H. Harding, Physica B+C, 1985, 131, 13--26

    1985

  36. [44]

    J. H. Harding and A. M. Stoneham, Philos. Mag. B, 1981, 43, 705--713

    1981

  37. [45]

    J. H. Harding, Phys. Rev. B, 1985, 32, 6861

    1985

  38. [46]

    J. Luo, C. Zhou, Q. Li and L. Liu, J. Chem. Phys., 2022, 156, 214113

    2022

  39. [47]

    J. Luo, C. Zhou, Q. Li and L. Liu, Materials, 2022, 15, 4026

    2022

  40. [48]

    Mishin, M

    Y. Mishin, M. R. Sorensen and A. F. Voter, Philos. Mag. A, 2001, 81, 2591--2612

    2001

  41. [49]

    Wynblatt, Phys

    P. Wynblatt, Phys. Status Solidi B, 1969, 36, 797--808

    1969

  42. [50]

    Lapointe, T

    C. Lapointe, T. D. Swinburne, L. Proville, C. S. Becquart, N. Mousseau and M.-C. Marinica, Phys. Rev. Mater., 2022, 6, 113803

    2022

  43. [51]

    P. W. M. Jacobs, J. Chem. Soc., Faraday Trans., 1990, 86, 1197--1201

    1990

  44. [52]

    Nam and R

    J. Nam and R. Gómez-Bombarelli, Interpolation and differentiation of alchemical degrees of freedom in machine learning interatomic potentials, Preprint at http://arxiv.org/abs/2404.10746 (2024)

    2024 arXiv

  45. [53]

    Ramos de Debiaggi, M

    S. Ramos de Debiaggi, M. de Koning and A. M. Monti, Phys. Rev. B, 2006, 73, 104103

    2006

  46. [54]

    K. M. Carling, G. Wahnström, T. R. Mattsson, N. Sandberg and G. Grimvall, Phys. Rev. B, 2003, 67, 054101

    2003

  47. [55]

    S. M. Foiles, Phys. Rev. B, 1994, 49, 14930--14938

    1994

  48. [56]

    S. K. Estreicher, M. Sanati, D. West and F. Ruymgaart, Phys. Rev. B, 2004, 70, 125209

    2004

  49. [57]

    Sanati and S

    M. Sanati and S. K. Estreicher, Phys. B: Condens. Matter, 2003, 340-342, 630--636

    2003

  50. [58]

    Sanati and S

    M. Sanati and S. K. Estreicher, Solid State Commun., 2003, 128, 181--185

    2003

  51. [59]

    Sanati and S

    M. Sanati and S. K. Estreicher, Phys. Rev. B, 2005, 72, 165206

    2005

  52. [60]

    Smargiassi and R

    E. Smargiassi and R. Car, Phys. Rev. B, 1996, 53, 9760--9763

    1996

  53. [61]

    Catlow, J

    C. Catlow, J. Corish, P. Jacobs and A. Lidiard, J. Phys. C: Sol. Stat. Phys., 1981, 14, L121

    1981

  54. [62]

    Ágoston and K

    P. Ágoston and K. Albe, Phys. Chem. Chem. Phys., 2009, 11, 3226--3232

    2009

  55. [63]

    Zacherle, P

    T. Zacherle, P. C. Schmidt and M. Martin, Phys. Rev. B, 2013, 87, 235206

    2013

  56. [64]

    Smith, S

    T. Smith, S. Moxon, J. S. Tse, J. M. Skelton, D. J. Cooke, L. J. Gillie, E. L. d. Silva, R. M. Harker, M. T. Storr, S. C. Parker and M. Molinari, J. Phys. Energy, 2023, 5, 025004

    2023

  57. [65]

    Zacherle, A

    T. Zacherle, A. Schriever, R. A. De Souza and M. Martin, Phys. Rev. B, 2013, 87, 134104

    2013

  58. [66]

    Walsh, A

    A. Walsh, A. A. Sokol and C. R. A. Catlow, Phys. Rev. B Condens. Matter, 2011, 83, 224105

    2011

  59. [67]

    Baldassarri, J

    B. Baldassarri, J. He and C. Wolverton, Phys. Rev. Mater., 2024, 8, 055407

    2024

  60. [68]

    S. L. Millican, J. M. Clary, C. B. Musgrave and S. Lany, Chem. Mater., 2022, 34, 519--528

    2022

  61. [69]

    Moxon, J

    S. Moxon, J. Skelton, J. S. Tse, J. Flitcroft, A. Togo, D. J. Cooke, E. L. d. Silva, R. M. Harker, M. T. Storr, S. C. Parker and M. Molinari, J. Mater. Chem. A, 2022, 10, 1861--1875

    2022

  62. [70]

    Y. Sun, T. Liu, Q. Chang and C. Ma, J. Phys. Chem. Solids, 2018, 115, 228--232

    2018

  63. [71]

    Miceli and A

    G. Miceli and A. Pasquarello, Phys. Rev. B, 2016, 93, 165207

    2016

  64. [72]

    Grieshammer, T

    S. Grieshammer, T. Zacherle and M. Martin, Phys. Chem. Chem. Phys., 2013, 15, 15935--15942

    2013

  65. [73]

    Cazorla, Phys

    C. Cazorla, Phys. Rev. Appl., 2017, 7, 044025

    2017

  66. [74]

    J. M. Wynn, R. J. Needs and A. J. Morris, Structures and finite-temperature abundances of defects in In _2 O _3 - II from first-principles calculations , Preprint at https://arxiv.org/abs/1609.04760 (2016)

    2016 arXiv

  67. [75]

    Youssef and B

    M. Youssef and B. Yildiz, Phys. Rev. B, 2012, 86, 144109

    2012

  68. [76]

    L. N. Holtzman, P. A. Vargas, R. G. Hennig and K. Barmak, J. Chem. Phys., 2024, 161, 144105

    2024

  69. [77]

    Gorfer, R

    A. Gorfer, R. Abart and C. Dellago, Phys. Rev. Mater., 2024, 8, 073602

    2024

  70. [78]

    R. J. Tarento and J. H. Harding, J. Phys. C: Solid State Phys., 1987, 20, L677

    1987

  71. [79]

    Zhang, F

    C. Zhang, F. Gygi and G. Galli, Phys. Rev. Mater., 2024, 8, 046201

    2024

  72. [80]

    T. S. Bjørheim, M. Arrigoni, D. Gryaznov, E. Kotomin and J. Maier, Phys Chem Chem Phys, 2015, 17, 20765--20774

    2015

  73. [81]

    Menon, Y

    S. Menon, Y. Lysogorskiy, J. Rogal and R. Drautz, Phys. Rev. Mater., 2021, 5, 103801

    2021

  74. [82]

    Kapil, E

    V. Kapil, E. Engel, M. Rossi and M. Ceriotti, J. Chem. Theory Comput., 2019, 15, 5845--5857

    2019

  75. [83]

    Batatia, D

    I. Batatia, D. P. Kovacs, G. Simm, C. Ortner and G. Csanyi, Adv Neural Inf Process Syst, 2022, 35, 11423--11436

    2022

  76. [84]

    W. K. Metzger, S. Grover, D. Lu, E. Colegrove, J. Moseley, C. L. Perkins, X. Li, R. Mallick, W. Zhang, R. Malik, J. Kephart, C.-S. Jiang, D. Kuciauskas, D. S. Albin, M. M. Al-Jassim, G. Xiong and M. Gloeckler, Nat Energy, 2019, 4, 837--845

    2019

  77. [85]

    Mosquera-Lois, S

    I. Mosquera-Lois, S. R. Kavanagh, A. Walsh and D. O. Scanlon, J. Open Source Softw., 2022, 7, 4817

    2022

  78. [86]

    A. Togo, L. Chaput, T. Tadano and I. Tanaka, J. Phys. Condens. Matter, 2023, 35, 353001

    2023

  79. [87]

    Knoop, T

    F. Knoop, T. A. R. Purcell, M. Scheffler and C. Carbogno, Phys. Rev. Mater., 2020, 4, 083809

    2020

  80. [88]

    M. Choi, F. Oba, Y. Kumagai and I. Tanaka, Adv. Mater., 2013, 25, 86--90

    2013

  81. [89]

    J. Park, B. Xu, J. Pan, D. Zhang, S. Lany, X. Liu, J. Luo and Y. Qi, npj Comput Mater, 2023, 9, 1--13

    2023

  82. [90]

    Mosquera-Lois, S

    I. Mosquera-Lois, S. R. Kavanagh, A. M. Ganose and A. Walsh, npj Comput Mater, 2024, 10, 1--9

    2024

  83. [91]

    L. D. Whalley, J. Phys. Chem. C., 2023, 127, 15738--15746

    2023

  84. [92]

    S. R. Kavanagh, A. G. Squires, A. Nicolson, I. Mosquera-Lois, A. M. Ganose, B. Zhu, K. Brlec, A. Walsh and D. O. Scanlon, J. Open Source Softw., 2024, 9, 6433

    2024

  85. [93]

    S. S. Naghavi, A. A. Emery, H. A. Hansen, F. Zhou, V. Ozolins and C. Wolverton, Nat Commun, 2017, 8, 285

    2017

  86. [94]

    Sikstrom and V

    D. Sikstrom and V. Thangadurai, Ionics, 2024, 1, 1

    2024

  87. [95]

    J. Gao, G. Tian, A. Sorniotti, A. E. Karci and R. Di Palo, Appl. Therm. Eng., 2019, 147, 177--187

    2019

  88. [96]

    Zhong, F

    X. Zhong, F. Höfling and T. John, Hydrogen diffusion in garnet: insights from atomistic simulations, Preprint at http://arxiv.org/abs/2410.07724 (2024)

    2024 arXiv

  89. [97]

    J. P. Perdew, A. Ruzsinszky, G. I. Csonka, O. A. Vydrov, G. E. Scuseria, L. A. Constantin, X. Zhou and K. Burke, Phys. Rev. Lett., 2008, 100, 136406

    2008

  90. [98]

    Kresse and J

    G. Kresse and J. Furthmüller, Comput. Mater. Sci., 1996, 6, 15--50

    1996

  91. [99]

    Kresse and J

    G. Kresse and J. Hafner, Phys. Rev. B, 1993, 47, 558--561

    1993

  92. [100]

    Kresse and J

    G. Kresse and J. Hafner, Phys. Rev. B, 1994, 49, 14251--14269

    1994

  93. [101]

    J. M. Skelton, D. Tiana, S. C. Parker, A. Togo, I. Tanaka and A. Walsh, J. Chem. Phys., 2015, 143, 064710

    2015

  94. [102]

    Jinnouchi, J

    R. Jinnouchi, J. Lahnsteiner, F. Karsai, G. Kresse and M. Bokdam, Phys. Rev. Lett., 2019, 122, 225701

    2019

  95. [103]

    Jinnouchi, F

    R. Jinnouchi, F. Karsai and G. Kresse, Phys. Rev. B, 2019, 100, 014105

    2019

  96. [104]

    Jinnouchi, K

    R. Jinnouchi, K. Miwa, F. Karsai, G. Kresse and R. Asahi, J. Phys. Chem. Lett., 2020, 11, 6946--6955

    2020

  97. [105]

    P. Liu, C. Verdi, F. Karsai and G. Kresse, Phys. Rev. B, 2022, 105, L060102

    2022

  98. [106]

    A. Jain, S. P. Ong, G. Hautier, W. Chen, W. D. Richards, S. Dacek, S. Cholia, D. Gunter, D. Skinner, G. Ceder and K. A. Persson, APL Materials, 2013, 1, 011002

    2013

  99. [107]

    Martínez, R

    L. Martínez, R. Andrade, E. G. Birgin and J. M. Martínez, J Comput Chem, 2009, 30, 2157--2164

    2009

  100. [108]

    Akola, R

    J. Akola, R. O. Jones, S. Kohara, T. Usuki and E. Bychkov, Phys. Rev. B, 2010, 81, 094202

    2010

  101. [109]

    J. F. Ziegler and J. P. Biersack, in The Stopping and Range of Ions in Matter, ed. D. A. Bromley, Springer US, Boston, MA, 1985, pp. 93--129

    1985

  102. [110]

    J. D. Morrow, J. L. A. Gardner and V. L. Deringer, J. Chem. Phys., 2023, 158, 121501

    2023

  103. [111]

    Kumagai and F

    Y. Kumagai and F. Oba, Phys. Rev. B Condens. Matter, 2014, 89, 195205

    2014

  104. [112]

    Eriksson, J

    O. Eriksson, J. M. Wills and D. Wallace, Phys. Rev. B, 1992, 46, 5221--5228

    1992

  105. [113]

    Zhang, B

    X. Zhang, B. Grabowski, F. Körmann, C. Freysoldt and J. Neugebauer, Phys. Rev. B, 2017, 95, 165126

    2017

  106. [114]

    Metsue, A

    A. Metsue, A. Oudriss, J. Bouhattate and X. Feaugas, J. Chem. Phys., 2014, 140, 104705

    2014

  107. [115]

    Willaime, A

    F. Willaime, A. Satta, M. Nastar and O. Le Bacq, Int. J. Quantum Chem., 2000, 77, 927--939

    2000

  108. [116]

    A. G. Squires, D. O. Scanlon and B. J. Morgan, J. Open Source Softw., 2023, 8, 4962

    2023

  109. [117]

    Buckeridge, Comput

    J. Buckeridge, Comput. Phys. Commun., 2019, 244, 329--342

    2019

  110. [118]

    A. H. M. Smets, K. Jäger, O. Isabella, R. A. v. Swaaij and M. Zeman, Solar energy : the physics and engineering of photovoltaic conversion, technologies and systems, UIT Cambridge Ltd., Cambridge, England, 2016

    2016

  111. [119]

    S. Lany, J. Chem. Phys., 2018, 148, 071101

    2018

  112. [120]

    Togo and I

    A. Togo and I. Tanaka, Scr. Mater., 2015, 108, 1--5

    2015

  113. [121]

    A. P. Thompson, H. M. Aktulga, R. Berger, D. S. Bolintineanu, W. M. Brown, P. S. Crozier, P. J. in 't Veld, A. Kohlmeyer, S. G. Moore, T. D. Nguyen, R. Shan, M. J. Stevens, J. Tranchida, C. Trott and S. J. Plimpton, Comp. Phys. Comm., 2022, 271, 108171

    2022

  114. [122]

    Freitas, M

    R. Freitas, M. Asta and M. de Koning, Comput. Mater. Sci., 2016, 112, 333--341

    2016

  115. [123]

    F. C. Kracek, J. Am. Chem. Soc., 1941, 63, 1989--1990

    1941

  116. [124]

    A. H. Larsen, J. J. Mortensen, J. Blomqvist, I. E. Castelli, R. Christensen, M. Dułak, J. Friis, M. N. Groves, B. Hammer, C. Hargus, E. D. Hermes, P. C. Jennings, P. B. Jensen, J. Kermode, J. R. Kitchin, E. L. Kolsbjerg, J. Kubal, K. Kaasbjerg, S. Lysgaard, J. B. Maronsson, T....

    2017

  117. [125]

    Shen and J

    J.-X. Shen and J. Varley, J. Open Source Softw., 2024, 9, 5941

    2024

  118. [126]

    J.-X. Shen, L. F. Voss and J. B. Varley, J. Appl. Phys., 2024, 135, 145102

    2024

  119. [127]

    S. P. Ong, W. D. Richards, A. Jain, G. Hautier, M. Kocher, S. Cholia, D. Gunter, V. L. Chevrier, K. A. Persson and G. Ceder, Comput. Mater. Sci., 2013, 68, 314--319

    2013

  120. [128]

    A. Jain, S. P. Ong, G. Hautier, W. Chen, W. D. Richards, S. Dacek, S. Cholia, D. Gunter, D. Skinner, G. Ceder and K. A. Persson, APL Materials , 2013, 1, 011002

    2013

  121. [129]

    S. P. Ong, S. Cholia, A. Jain, M. Brafman, D. Gunter, G. Ceder and K. A. Persson, Comput. Mater. Sci., 2015, 97, 209--215

    2015

  122. [130]

    Himanen, M

    L. Himanen, M. O. J. J \"a ger, E. V. Morooka, F. Federici Canova, Y. S. Ranawat, D. Z. Gao, P. Rinke and A. S. Foster, Comput. Phys. Commun., 2020, 247, 106949

    2020

  123. [131]

    Laakso, L

    J. Laakso, L. Himanen, H. Homm, E. V. Morooka, M. O. J \"a ger, M. Todorovi \'c and P. Rinke, J. Chem. Phys., 2023, 158, 234802

    2023

  124. [132]

    McInnes, J

    L. McInnes, J. Healy, N. Saul and L. Großberger, J. Open Source Softw., 2018, 3, 861

    2018

  125. [133]

    J. Qi, T. W. Ko, B. C. Wood, T. A. Pham and S. P. Ong, Npj Comput. Mater., 2024, 10, 43

    2024

  126. [134]

    J. D. Hunter, Comput. Sci. Eng., 2007, 9, 90--95

    2007

  127. [135]

    M. L. Waskom, J. Open Source Softw., 2021, 6, 3021

    2021

  128. [136]

    Stukowski, Model

    A. Stukowski, Model. Simul. Mater. Sci. Eng. , 2010 , 18 , 015012

    2010

  129. [137]

    CrystalMaker Software Ltd, Oxford, England (www.crystalmaker.com)

    CrystalMaker®: a crystal and molecular structures program for Mac and Windows. CrystalMaker Software Ltd, Oxford, England (www.crystalmaker.com)

  130. [138]

    Henkelman, B

    G. Henkelman, B. P. Uberuaga and H. Jónsson, J. Chem. Phys., 2000, 113, 9901--9904

    2000

  131. [139]

    Henkelman and H

    G. Henkelman and H. Jónsson, J. Chem. Phys., 2000, 113, 9978--9985

    2000

  132. [140]

    Knoop, T

    F. Knoop, T. A. R. Purcell, M. Scheffler and C. Carbogno, J. Open Source Softw., 2020, 5, 2671

    2020

  133. [141]

    o ping University, SE-581 83, Link\

    C. Chen, Y. Zuo, W. Ye, Q. Ji and S. P. Ong, Maml - materials machine learning package , https://github.com/materialsvirtuallab/maml, 2020 mcitethebibliography main.tex0000664000000000000000000015730614731611437011247 0ustar rootroot [ preprint, superscriptaddress, nofootinbib...

    2020

  134. [142]

    C. Chen, Y. Zuo, W. Ye, Q. Ji and S. P. Ong, Maml - materials machine learning package , https://github.com/materialsvirtuallab/maml, 2020

    2020

  135. [143]

    F. C. Kracek, J. Am. Chem. Soc., 1941, 63, 1989--1990 mcitethebibliography SI_for_main.tex0000664000000000000000000004162314731611437012502 0ustar rootroot center Supplementary Information for `Point defect formation at finite temperatures with machine learning force fields' c...

    1941

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.