Pith. sign in

REVIEW 3 major objections 5 minor 52 references

Statistical Mechanics of Thermal Diffusion in Rough Energy Landscapes

T0 review · 3 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read A new identity splits the temperature dependence of diffusion barriers into three computable fluctuations.

desk verdict A clean, parameter-free decomposition of dQ/dβ into thermodynamic and kinetic fluctuation terms, with a good exact example and a weaker alloy validation; the T-independence assumption is the real limit on 'broad applicability.' read the letter →

arxiv 2607.22742 v1 pith:NPEP5OD3 submitted 2026-07-22 cond-mat.stat-mech cond-mat.mtrl-sci

Soham Chattopadhyay , Blas P. Uberuaga This is my paper · ORCID
classification cond-mat.stat-mechcond-mat.mtrl-sci PACS 66.30.-h05.20.-y
keywords diffusiondisorderedmaterialsvariationalprinciplemachinelearningnon-Arrheniusbehavioractivationbarrierkineticdistributiontransitionstatetheory
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 claims that non-Arrhenius diffusion in rough energy landscapes is governed by three microscopic fluctuation terms: the variance of state energies, the variance of transition-state energies under a kinetics-controlled distribution, and a kinetic-correlation term. It derives an exact identity for the first temperature derivative of the diffusion activation barrier and shows that approximately Arrhenius behavior emerges when these fluctuations compete or cancel. Calculations on oxygen-vacancy diffusion in a bixbyite oxide and vacancy diffusion in a Ni-Cu solid solution demonstrate that the identity can be evaluated from local energetics and machine-learned jump statistics. This matters because it replaces empirical fits with a quantitative, microscopic accounting of why diffusion plots curve.

What carries the argument

The variational principle for mass transport expresses the self-diffusion coefficient as an infimum over trial functions; its minimizer, the relaxation vector η^a_n, encodes correlations between successive jumps and satisfies a discrete Poisson equation. From this, the activation barrier Q^a becomes an average of transition-state energies under the kinetic distribution Pκ, and differentiating that average yields Eq. 5. The relaxation-vector derivative dβη is shown to obey its own Poisson equation and variational principle, making the new kinetic term accessible to machine-learned approximations from single-step kinetic Monte Carlo data.

What would settle it

On a lattice model with temperature-independent barriers, compute D(T) by kinetic Monte Carlo, numerically differentiate ln D to obtain dQ/dβ, and compare with the right-hand side of Eq. 5 evaluated from the same jump statistics; agreement at all temperatures confirms the identity. Repeating with barriers that depend on temperature, for example through vibrational entropy, should produce a discrepancy proportional to those derivatives.

Watch

Extended reading notes

Core claim

The central claim is Eq. 5 of the main text: dQ^a/dβ = ΔE_n − Δ_{κ,a} E_{TS} + Δ_{dβη}. The first-order temperature dependence of the per-defect diffusion activation barrier is exactly three statistical fluctuations: the variance of state energies under the equilibrium Boltzmann distribution, the variance of transition-state energies under the kinetic distribution Pκ (which weights each jump by its squared contribution to net diffusion), and the averaged fluctuations of the temperature derivative of the relaxation vectors across the diffusion network. If this identity holds, non-Arrhenius behavior is not a free fit parameter but a consequence of calculable thermodynamic and kinetic fluctuati

Load-bearing premise

State energies, transition-state energies, entropic prefactors, and jump displacement vectors are assumed temperature-independent, and defect formation and annihilation kinetics are set aside; if thermal expansion, vibrational entropy, or temperature-dependent barriers matter, Eq. 5 misses extra derivative terms.

Editorial extensions

If this is right

  • Deviations from Arrhenius behavior in complex solids can be computed from local energies, barriers, and jump statistics rather than inferred from fits.
  • Approximately Arrhenius diffusion at high temperature in solid solutions follows from configurational entropy and cancellation between the thermodynamic and kinetic variance terms; non-Arrhenius behavior appears as short-range ordering sets in.
  • Kinetic traps, such as repeated back-and-forth jumps, raise the effective activation barrier at low temperature, while competing variance in transition-state energies lowers it; the competition limits the curvature of the Arrhenius plot.
  • The relaxation-vector derivatives satisfy their own variational principle, so the new kinetic term can be approximated with physics-conforming neural networks and linear bias corrections trained on kinetic Monte Carlo data.
  • The identity generalizes superbasin transition-state theory and shows that transitions with transition-state energies below the thermal average can contribute negative terms to the activation barrier.

Reading between the lines

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

  • If Eq. 5 survives scrutiny, the sign of an Arrhenius plot's curvature becomes a diagnostic: positive curvature would indicate thermodynamic energy fluctuations dominating, negative curvature would indicate kinetic barrier variance, and near-flat behavior would indicate cancellation.
  • The supplement's general treatment of derivatives with respect to macroscopic variables suggests analogous identities for activation volumes and elastodiffusion tensors, where the same fluctuation decomposition might hold under strain or external fields.
  • Existing kinetic Monte Carlo datasets could be post-processed to compute all three terms, offering a testable extension to high-entropy alloys where 'sluggish diffusion' would be attributed to specific fluctuations rather than assumed.
  • The role of configurational entropy near a miscibility gap suggests that the same identity could be used to detect the onset of chemical ordering from diffusion data alone.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper derives an exact closed-form expression for the temperature derivative of the per-defect diffusion activation barrier Q^a within harmonic transition state theory, starting from Trinkle's variational principle for mass transport. Defining a kinetic distribution Pκ,a over transition states, the authors obtain dQ^a/dβ = Var(E_n) - Var_κ(E_TS) + Δ_{dβη}, i.e., a decomposition into the equilibrium variance of state energies, the kinetic variance of transition-state energies, and a correlation term involving the temperature derivative of the relaxation vector. The theory is illustrated with two examples: an exact Green's-function calculation for oxygen-vacancy diffusion in Sm2O3, and a machine-learning-based (NN+SRBC) study of vacancy diffusion in Ni-20%Cu. The paper concludes that approximate Arrhenius behavior at high temperature arises from cancellation of thermodynamic and kinetic fluctuations, and that non-Arrhenius behavior appears when ordering and kinetic trapping become important.

Significance. If the central derivation is correct, Eq. (5) is a valuable result: it converts a formal derivative of the diffusion coefficient into three separately interpretable and computable statistical quantities. The Sm2O3 example is a genuinely exact benchmark with parameter-free inputs from published barriers, and it cleanly demonstrates how a kinetic trap can produce a nearly Arrhenius but non-trivial effective barrier. The Ni-Cu example shows how the formalism can be exported to a realistic disordered alloy using machine-learned surrogate models. The main limitations are the explicitly assumed temperature-independence of the energy landscape (energies, entropies, and jump vectors) and the restriction to per-defect diffusion; both limit the 'broadly applicable' claim made in the abstract. The alloy validation is also weaker than it appears because the extrapolations and reference data derive from the same surrogate.

major comments (3)
  1. [Sec. 1 / Eq. (5)] The derivation of Eq. (5) is sound only under the assumption stated on p. 2 that ρ_n,n', E_TS, ρ_n, E_n, and δx are temperature-independent. When these quantities depend on T—thermal expansion, vibrational entropy, phonon renormalization—the derivative d(P0W)/dβ in Eq. S10 acquires additional terms, and Eq. (5) is not the complete dQ^a/dβ. The paper acknowledges this in Sec. 1, but the abstract and conclusions claim a 'broadly applicable statistical framework' for 'complex solids'. This is a load-bearing scope mismatch. Please either restrict the claims to the T-independent landscape model, or use the formalism of Sec. S3 (which already treats λ-dependent free energies) or an explicit quasiharmonic estimate to quantify the omitted derivative terms.
  2. [Sec. 3 / Fig. 3] The Ni-Cu demonstration does not validate Eq. (5) or the 'nearly Arrhenius at high temperature' conclusion. The first- and second-order extrapolations in Fig. 3 are constructed from Q_v and dQ_v/dβ computed with the NN+SRBC surrogate, and the reference points being extrapolated are the same NN+SRBC estimates. Agreement thus only demonstrates that the truncated Taylor expansion is internally consistent with the surrogate; it is circular with respect to the physics. The subset-resampling uncertainties in Fig. 2(a) are statistical only and do not bound surrogate error. I request either a direct brute-force kMC benchmark for at least one pair of temperatures, or a clear statement that the alloy results are surrogate-model predictions rather than independent validation.
  3. [Sec. 1 (last paragraph) and Sec. 3] The paper explicitly sets aside defect formation and annihilation. Consequently, all equations and examples describe per-defect mobility, not the total self-diffusion coefficient that tracer experiments measure. The comparison to the experimental tracer activation barrier of CoNiCrFeMn in Sec. 3 is indirect: an equilibrium vacancy concentration contributes its own Arrhenius temperature dependence. This restriction should appear in the abstract and conclusion; as written, 'thermal diffusion in complex solids' and 'vacancy diffusion in solid solutions' overstate the demonstrated scope.
minor comments (5)
  1. [Sec. 1, after Eq. (4)] The statement that Pκ,a is 'independent of E_n' is too strong: η^a depends on the rates W and hence on the state energies E_n. Recommend rephrasing as 'has no explicit dependence on the equilibrium state probabilities P0_n'.
  2. [Eq. (5)] The symbol Δdβη^a is easy to misread as a product. Consider writing it as Δ_{d_β η} or adding a one-line definition of the subscript.
  3. [Sec. 2] In the sentence 'kinetic correlations induced by fast, low-barrier mechanisms that that can hinder diffusion', the duplicated 'that' should be removed.
  4. [Sec. S1 D] Equation (46) for dΓ/dβ is stated without derivation; a one-line derivation would help the reader verify the sign.
  5. [Sec. S2 A] Please clarify whether the vacancy participates in the Metropolis swap moves, and whether the reported state energies include the vacancy. The fixed vacancy concentration is stated, but the MC protocol is otherwise ambiguous.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Eq. 5 is derived self-contained from the variational principle, and the alloy validation is an out-of-sample check rather than a fitted prediction.

full rationale

I tracked the derivation chain for the paper's central result, Eq. 5. The derivation is analytically self-contained: S1 A differentiates P0_n W_{n→n'} under the explicitly stated temperature-independence of E_n, E_TS, ρ, and δx, yielding d(P0_n W)/dβ = −P0_n W (E_TS − ⟨E_n⟩0). S1 B derives a variational Poisson equation for dη/dβ from the original relaxation-vector equation. S1 C then combines these pieces to obtain dQ^a/dβ = ΔE_n − Δ_{κ,a} E_TS + Δ_{dβη} exactly as an algebraic identity. No fitted parameter is renamed as a prediction, and no input quantity is defined in terms of the output. The analytical Sm2O3 example is computed exactly from the given barriers with the Green-function method, so it is not circular. The alloy example uses NN+SRBC as an approximation method, but the central claims do not reduce to that method; they are illustrations of Eq. 5. Figure 3 extrapolates ln D from a single temperature using Q and dQ/dβ computed at that temperature and compares against NN+SRBC values at other temperatures; those comparison values are generated from equilibrium states and barriers at the other temperatures, so the agreement is an out-of-sample check rather than a self-consistency loop. Self-citations [26] and [33] are methodological or supportive, not load-bearing uniqueness arguments, and [26] has publicly available source code. The acknowledged T-independence and defect-formation exclusions are scope limitations, not circularity. No specific reduction of a claimed prediction to its inputs by construction was found.

Assumptions & free parameters 3 free parameters · 7 assumptions · 0 invented entities

The central derivation adds no free physical parameters: it follows from differentiation of the variational principle under the stated assumptions. The only fitted objects are the neural-network/SRBC surrogate coefficients used in the Ni-Cu example and an assumed attempt frequency for absolute D_v values. The clean derivation is therefore the main contribution; the computational demonstration pulls its model inputs (EAM potential, Sm2O3 barriers) from prior literature.

free parameters (3)
  • NN weights and SRBC coefficients θ*_SRBC = not reported numerically; optimized per temperature
    All Ni-Cu alloy results for D_v, Q_v and dQ_v/dβ depend on the machine-learned surrogate trained to minimize Eq. 1 on 10^4 single-step KMC samples at each temperature. No independent validation of the surrogate's accuracy on this specific system is provided.
  • Vacancy-atom exchange attempt frequency = 10 THz
    Assumed for absolute D_v values in Fig. 3. It cancels in Q_v and dQ_v/dβ but sets the plotted diffusivity scale.
  • Sm2O3 migration barrier inputs = 0.84 eV, 1.07 eV, 1.42 eV
    Taken from ref [35] as literature inputs, not fitted here. The Sm2O3 demonstration's quantitative results depend on these values.
assumptions (7)
  • standard math Variational principle for mass transport (Eq. 1, ref [25])
    The entire derivation starts from this theorem; correctness of D as an infimum and the Poisson equation for η are assumed.
  • domain assumption Harmonic transition state theory rates and detailed balance
    Transition rates are assumed to share forward/backward transition states with entropic prefactors; used in every subsequent formula.
  • ad hoc to paper Temperature independence of E_n, E_TS_n,n', ρ_n, ρ_n,n', δx
    Explicitly stated in Section 1. If violated, Eq. 5 misses extra derivative terms and the decomposition is incomplete.
  • domain assumption Equilibrium Boltzmann state distribution and no phase transformations
    P0_n and the expansion of Q assume near-equilibrium thermodynamics and a fixed state space; the authors note 'in the absence of phase transformations'.
  • ad hoc to paper Per-defect diffusion; defect formation/annihilation set aside
    The main text says examples focus on per-defect diffusion properties; total diffusivity in real solids requires formation kinetics.
  • domain assumption EAM potential for Ni-Cu (Foiles 1985) and NEB transition states
    All alloy energetics and barriers come from this potential; an inaccurate potential changes Q_v and dQ_v/dβ.
  • domain assumption NN+SRBC surrogate converges to true relaxation vectors
    Alloy Q_v and D_v are computed with the approximation from [26]; no convergence certificate against exact kMC is supplied.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Statistical Mechanics of Thermal Diffusion in Rough Energy Landscapes." pith.science (2026). https://pith.science/paper/NPEP5OD3

@misc{pith2026260722742,
  author       = {Pith},
  title        = {Pith review of: Statistical Mechanics of Thermal Diffusion in Rough Energy Landscapes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NPEP5OD3}},
  note         = {Machine review of arXiv:2607.22742}
}
read the original abstract

Starting from a variational principle for mass transport, we present a broadly applicable statistical framework describing thermal diffusion in complex solids. We show that microscopic thermodynamic and kinetic fluctuations govern non-Arrhenius diffusion, with kinetic terms described by machine-learnable relative diffusion contributions. For vacancy diffusion in solid solutions, deviation from Arrheniusness may occur when ordering effects start to become important but at higher temperatures, approximately Arrhenius behavior can occur, aided by configurational entropy and competing energy fluctuations.

Figures

Figures reproduced from arXiv: 2607.22742 by the authors.

Figure 1
Figure 1. FIG. 1. Di [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Energetics of isolated vacancy di [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Extrapolations of [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 10 canonical work pages

  1. [1]

    Tunable interstitial and vacancy diffusivity by chemical ordering control in CrCoNi medium-entropy alloy,

    Yangen Li, Jun-Ping Du, Shuhei Shinzato, and Shigenobu Ogata, “Tunable interstitial and vacancy diffusivity by chemical ordering control in CrCoNi medium-entropy alloy,” npj Computational Mate- rials10(2024), 10.1038/s41524-024-01322-6

    work page doi:10.1038/s41524-024-01322-6 2024

  2. [2]

    Enhancing radiation tolerance by controlling defect mobility and migration pathways in multicom- ponent single-phase alloys,

    Chenyang Lu, Liangliang Niu, Nanjun Chen, Ke Jin, Taini Yang, Pengyuan Xiu, Yanwen Zhang, Fei Gao, Hongbin Bei, Shi Shi, Mo-Rigen He, Ian M. Robertson, William J. Weber, and Lumin Wang, “Enhancing radiation tolerance by controlling defect mobility and migration pathways in multicom- ponent single-phase alloys,” Nature Communications7(2016), 10.1038/ncomms13564

    work page doi:10.1038/ncomms13564 2016

  3. [3]

    Cation-disordered rocksalt-type high-entropy cathodes for Li-ion batteries,

    Zhengyan Lun, Bin Ouyang, Deok-Hwang Kwon, Yang Ha, Emily E. Foley, Tzu-Yang Huang, Zi- jian Cai, Hyunchul Kim, Mahalingam Balasubramanian, Yingzhi Sun, Jianping Huang, Yaosen Tian, Haegyeom Kim, Bryan D. McCloskey, Wanli Yang, Rapha¨ele J. Cl´ement, Huiwen Ji, and Gerbrand Ceder, “Cation-disordered rocksalt-type high-entropy cathodes for Li-ion batteries...

    2020

  4. [4]

    High-entropy mechanism to boost ionic conductivity,

    Yan Zeng, Bin Ouyang, Jue Liu, Young-Woon Byeon, Zijian Cai, Lincoln J. Miara, Yan Wang, and Gerbrand Ceder, “High-entropy mechanism to boost ionic conductivity,” Science378, 1320–1324 (2022)

    2022

  5. [5]

    High-entropy electrolytes in boosting battery performance,

    Jijian Xu, “High-entropy electrolytes in boosting battery performance,” Materials Futures2, 047501 (2023)

    2023

  6. [6]

    High-entropy strategy flattening lithium ion migration energy landscape to enhance the conductivity of garnet-type solid-state electrolytes,

    Shuhan Wang, Xiaojuan Wen, Zhenweican Huang, Haoyang Xu, Fengxia Fan, Xinxiang Wang, Guilei Tian, Sheng Liu, Pengfei Liu, Chuan Wang, Chenrui Zeng, Chaozhu Shu, and Zhenx- ing Liang, “High-entropy strategy flattening lithium ion migration energy landscape to enhance the conductivity of garnet-type solid-state electrolytes,” Advanced Functional Materials35...

    work page doi:10.1002/adfm.202416389 2024

  7. [7]

    Configurational entropy-driven oxide-ion transport in rare-earth high-entropy fluorite electrolytes for low-temperature fuel cells,

    Asma Noor, Chunyu Guo, Muhammad Yousaf, Li Xiuxiu, Bushra Bibi, Qaisar Hayat, and Yuzheng Lu, “Configurational entropy-driven oxide-ion transport in rare-earth high-entropy fluorite electrolytes for low-temperature fuel cells,” ACS Sustainable Chemistry & Engineering14, 8259–8270 (2026)

    2026

  8. [8]

    Correlation factors for diffusion in nondilute alloys,

    John R. Manning, “Correlation factors for diffusion in nondilute alloys,” Physical Review B4, 1111–1121 (1971)

    1971

Show all 52 references
  1. [9]

    Diffusion, correlation, and percolation in a random alloy,

    G. E. Murch and S. J. Rothman, “Diffusion, correlation, and percolation in a random alloy,” Philo- sophical Magazine A43, 229–238 (1981)

    1981

  2. [10]

    Diffusion in a rough potential

    R Zwanzig, “Diffusion in a rough potential.” Proceedings of the National Academy of Sciences85, 2029–2030 (1988). 11

  3. [11]

    A self-consistent theory of matter transport in a random lattice gas and some simulation results,

    L. K. Moleko, A. R. Allnatt, and E. L. Allnatt, “A self-consistent theory of matter transport in a random lattice gas and some simulation results,” Philosophical Magazine A59, 141–160 (1989)

    1989

  4. [12]

    Collective and tracer diffusion kinetics in the ternary random alloy,

    I V Belova, A R Allnatt, and G E Murch, “Collective and tracer diffusion kinetics in the ternary random alloy,” Journal of Physics: Condensed Matter14, 6897–6907 (2002)

    2002

  5. [13]

    A self-consistent mean field theory for diffusion in alloys,

    Maylise Nastar and Vincent Barbe, “A self-consistent mean field theory for diffusion in alloys,” Fara- day Discuss.134, 331–342 (2007)

    2007

  6. [14]

    Variational approach to the diffusion on inhomogeneous lattices,

    Magdalena A. Załuska-Kotur, “Variational approach to the diffusion on inhomogeneous lattices,” Ap- plied Surface Science304, 122–126 (2014)

    2014

  7. [15]

    Anomalous dimensionality dependence of diffusion in a rugged energy landscape: How pathological is one dimension?

    Kazuhiko Seki, Kaushik Bagchi, and Biman Bagchi, “Anomalous dimensionality dependence of diffusion in a rugged energy landscape: How pathological is one dimension?” The Journal of Chemical Physics144(2016), 10.1063/1.4948936

    2016 doi

  8. [16]

    A high accuracy diffusion kinetics formalism for random multicomponent alloys: application to high entropy alloys,

    A. R. Allnatt, T. R. Paul, I. V . Belova, and G. E. Murch, “A high accuracy diffusion kinetics formalism for random multicomponent alloys: application to high entropy alloys,” Philosophical Magazine96, 2969–2985 (2016)

    2016

  9. [17]

    Vacancy diffusion in multi-principal element alloys: The role of chemical disorder in the ordered lattice,

    Spencer L. Thomas and Srikanth Patala, “Vacancy diffusion in multi-principal element alloys: The role of chemical disorder in the ordered lattice,” Acta Materialia196, 144–153 (2020)

    2020

  10. [18]

    Kineclue: A kinetic cluster expansion code to compute transport coefficients beyond the dilute limit,

    Thomas Schuler, Luca Messina, and Maylise Nastar, “Kineclue: A kinetic cluster expansion code to compute transport coefficients beyond the dilute limit,” Computational Materials Science172, 109191 (2020)

    2020

  11. [19]

    Estimating linear mass transport coefficients in solid solutions via correlation splitting and a law of total diffusion,

    Manuel Ath `enes, Gilles Adjanor, and J ´erˆome Creuze, “Estimating linear mass transport coefficients in solid solutions via correlation splitting and a law of total diffusion,” Physical Review Materials6 (2022), 10.1103/physrevmaterials.6.013805

    2022 doi

  12. [20]

    On the existence and origin of sluggish diffusion in chemically disordered concentrated alloys,

    Yuri N. Osetsky, Laurent K. B ´eland, Alexander V . Barashev, and Yanwen Zhang, “On the existence and origin of sluggish diffusion in chemically disordered concentrated alloys,” Current Opinion in Solid State and Materials Science22, 65–74 (2018)

    2018

  13. [21]

    Collective diffusion in the binary random alloy,

    I. V . Belova and G. E. Murch, “Collective diffusion in the binary random alloy,” Philosophical Maga- zine A80, 599–607 (2000)

    2000

  14. [22]

    Revealing the crucial role of rough energy landscape on self-diffusion in high-entropy alloys based on machine learning and kinetic Monte Carlo,

    Biao Xu, Jun Zhang, Shihua Ma, Yaoxu Xiong, Shasha Huang, J.J. Kai, and Shijun Zhao, “Revealing the crucial role of rough energy landscape on self-diffusion in high-entropy alloys based on machine learning and kinetic Monte Carlo,” Acta Materialia234, 118051 (2022)

    2022

  15. [23]

    Influence of defect thermodynamics on self-diffusion in com- 12 plex concentrated alloys with chemical ordering,

    Anus Manzoor and Yongfeng Zhang, “Influence of defect thermodynamics on self-diffusion in com- 12 plex concentrated alloys with chemical ordering,” JOM74, 4107–4120 (2022)

    2022

  16. [24]

    Prediction of vacancy defect diffusion paths in high entropy alloys via machine learning on molecular dynamics data,

    C. Reimer, P. Saidi, C. Casert, C. Beeler, C. G. Tetsassi Feugmo, S. Whitelam, E. Mansouri, A. Mar- tinez, L. Beland, and I. Tamblyn, “Prediction of vacancy defect diffusion paths in high entropy alloys via machine learning on molecular dynamics data,” Journal of Applied Physi...

    2025 doi

  17. [29]

    Atomic movements in solids – phenomenological equations,

    A. R. Allnatt and A. B. Lidiard, “Atomic movements in solids – phenomenological equations,” in Atomic Transport in Solids(Cambridge University Press, 1993) p. 1–49

    1993

  18. [30]

    Diffusion and transformation kinetics of small helium clusters in bulk tungsten,

    Danny Perez, Thomas V ogel, and Blas P. Uberuaga, “Diffusion and transformation kinetics of small helium clusters in bulk tungsten,” Physical Review B90(2014), 10.1103/physrevb.90.014102

    2014 doi

  19. [31]

    S3, we show that similar variational relations also exist for relaxation deriva- tives with respect to macroscopic variables

    In supplementary Sec. S3, we show that similar variational relations also exist for relaxation deriva- tives with respect to macroscopic variables

  20. [32]

    Understanding the physical metallurgy of the CoCrFeMnNi high-entropy alloy: an atomistic simulation study,

    Won-Mi Choi, Yong Hee Jo, Seok Su Sohn, Sunghak Lee, and Byeong-Joo Lee, “Understanding the physical metallurgy of the CoCrFeMnNi high-entropy alloy: an atomistic simulation study,” npj Computational Materials4(2018), 10.1038/s41524-017-0060-9

    2018 doi

  21. [33]

    Larger than uncorrelated vacancy diffusion contribu- tions in chemically disordered crystalline materials,

    Soham Chattopadhyay and Blas P. Uberuaga, “Larger than uncorrelated vacancy diffusion contribu- tions in chemically disordered crystalline materials,” Scripta Materialia270, 116928 (2026)

    2026

  22. [34]

    Interpreting oxygen vacancy migration mechanisms in oxides using the layered structure motif,

    Blas Pedro Uberuaga and Kurt E. Sickafus, “Interpreting oxygen vacancy migration mechanisms in oxides using the layered structure motif,” Computational Materials Science103, 216–223 (2015)

    2015

  23. [35]

    The impact of chemistry on anion migration in bixbyite-structured lanthanide oxides,

    Blas Pedro Uberuaga and Soham Chattopadhyay, “The impact of chemistry on anion migration in bixbyite-structured lanthanide oxides,” Solid State Ionics444, 117264 (2026)

    2026

  24. [36]

    Clustering in Cu-Ni alloys: A diffuse neutron-scattering study,

    J. Vrijen and S. Radelaar, “Clustering in Cu-Ni alloys: A diffuse neutron-scattering study,” Physical Review B17, 409–421 (1978)

    1978

  25. [37]

    On the 13 miscibility gap of Cu-Ni system,

    Yusuke Iguchi, G ´abor L. Katona, Csaba Cserh ´ati, G ´abor A. Langer, and Zolt ´an Erd ´elyi, “On the 13 miscibility gap of Cu-Ni system,” Acta Materialia148, 49–54 (2018)

    2018

  26. [38]

    Phase equilibria and thermodynamics of binary copper systems with 3d-metals. vi. copper-nickel system,

    M. A. Turchanin, P. G. Agraval, and A. R. Abdulov, “Phase equilibria and thermodynamics of binary copper systems with 3d-metals. vi. copper-nickel system,” Powder Metallurgy and Metal Ceramics 46, 467–477 (2007)

    2007

  27. [40]

    A climbing image nudged elastic band method for finding saddle points and minimum energy paths,

    Graeme Henkelman, Blas P. Uberuaga, and Hannes J ´onsson, “A climbing image nudged elastic band method for finding saddle points and minimum energy paths,” The Journal of Chemical Physics113, 9901–9904 (2000)

    2000

  28. [42]

    Group equivariant convolutional networks,

    Taco Cohen and Max Welling, “Group equivariant convolutional networks,” inProceedings of The 33rd International Conference on Machine Learning, Proceedings of Machine Learning Research, V ol. 48, edited by Maria Florina Balcan and Kilian Q. Weinberger (PMLR, New York, New York...

    2016

  29. [44]

    Tracer diffusion in single crystalline CoCrFeNi and CoCrFeMnNi high-entropy alloys: Kinetic hints towards a low-temperature phase instability of the solid-solution?

    Daniel Gaertner, Josua Kottke, Yury Chumlyakov, Fabian Hergem¨oller, Gerhard Wilde, and Sergiy V . Divinski, “Tracer diffusion in single crystalline CoCrFeNi and CoCrFeMnNi high-entropy alloys: Kinetic hints towards a low-temperature phase instability of the solid-solution?” S...

    2020

  30. [45]

    ABC- FIRE

    and the Onsager package [3] were used to construct supercells, write LAMMPS data and for symmetry analysis. A. Generating Initial States MC simulations to construct the datasets were done using the LAMMPS software for energy optimization [8]. All energy relaxations were carrie...

  31. [46]

    direct” approximation is obtained as described in Sec. S1 D, while the “variational

    The neural networks (NNs) are able to substantially capture the correlation effects in the system, with subsequent SRBC correction providing more accurate results. All samples (both training and validation sets) were considered in these plots as they were found to be in reason...

  32. [47]

    Variational principle for mass transport,

    Dallas R. Trinkle, “Variational principle for mass transport,” Physical Review Letters121(2018), 10.1103/physrevlett.121.235901

    2018 doi

  33. [48]

    Frequency factors and isotope effects in solid state rate processes,

    George H. Vineyard, “Frequency factors and isotope effects in solid state rate processes,” Journal of Physics and Chemistry of Solids3, 121–127 (1957)

    1957

  34. [49]

    Diffusivity and derivatives for interstitial solutes: activation energy, volume, and elastodiffusion tensors,

    Dallas R. Trinkle, “Diffusivity and derivatives for interstitial solutes: activation energy, volume, and elastodiffusion tensors,” Philosophical Magazine96, 2714–2735 (2016)

    2016

  35. [50]

    Contributions to diffusion in complex materials quan- tified with machine learning,

    Soham Chattopadhyay and Dallas R. Trinkle, “Contributions to diffusion in complex materials quan- tified with machine learning,” Physical Review Letters132(2024), 10.1103/physrevlett.132.186301

    2024 doi

  36. [51]

    S4 of [4], due to a typographical error, a negative sign is missing on the right-hand side

    In Eq. S4 of [4], due to a typographical error, a negative sign is missing on the right-hand side. Eq. 41 is the correct equation for solving the optimal linear expansion coefficients and also correctly reflects all code implementation and results in both [4] and the current work

  37. [52]

    Calculation of the surface segregation of Ni-Cu alloys with the use of the embedded- atom method,

    S. M. Foiles, “Calculation of the surface segregation of Ni-Cu alloys with the use of the embedded- atom method,” Physical Review B32, 7685–7693 (1985)

    1985

  38. [53]

    The atomic simulation environment—a python library for working with atoms,

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

    2017

  39. [54]

    LAMMPS - a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales,

    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, “LAMMPS - a flexible simulation tool for particle-based mate...

    2022

  40. [55]

    Abc-fire: Accelerated bias-corrected fast inertial relaxation engine,

    Sebasti ´an Echeverri Restrepo and Predrag Andric, “Abc-fire: Accelerated bias-corrected fast inertial relaxation engine,” Computational Materials Science218, 111978 (2023)

    2023

  41. [56]

    Group equivariant convolutional networks,

    Taco Cohen and Max Welling, “Group equivariant convolutional networks,” inProceedings of The 33rd International Conference on Machine Learning, Proceedings of Machine Learning Research, 23 V ol. 48, edited by Maria Florina Balcan and Kilian Q. Weinberger (PMLR, New York, New Y...

    2016

  42. [57]

    Pytorch: An imperative style, high-performance deep learning library,

    A. Paszke, S. Gross, F. Massa, ..., and S. Chintala, “Pytorch: An imperative style, high-performance deep learning library,” inAdvances in Neural Information Processing Systems 32(2019) pp. 8024– 8035

    2019

  43. [58]

    Adam: A method for stochastic optimization,

    Diederik P. Kingma and Jimmy Ba, “Adam: A method for stochastic optimization,” (2017), arXiv:1412.6980 [cs.LG]

    2017 arXiv

  44. [59]

    A multivariate faa di bruno formula with applications,

    G. Constantine and T. Savits, “A multivariate faa di bruno formula with applications,” Transactions of the American Mathematical Society348, 503–520 (1996). 24

    1996

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.