Pith. sign in

REVIEW 4 major objections 5 minor 54 references

High-Entropy Solid Electrolytes Discovery: A Dual-Stage Machine Learning Framework Bridging Atomic Configurations and Ionic Transport Properties

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A two-stage machine-learning pipeline can screen thousands of high-entropy solid electrolytes and points to a candidate with 4.53 mS/cm room-temperature ionic conductivity.

desk verdict A useful screening workflow and a new dataset, but the banner 4.53 mS/cm conductivity claim rests on an undocumented Arrhenius extrapolation. read the letter →

arxiv 2505.18571 v1 pith:HWXPRQPU submitted 2025-05-24 cond-mat.mtrl-sci

Xiao Fu , Jing Xu , Qifan Yang , Xuhe Gong , Jingchen Lian , Liqi Wang , Zibin Wang , Ruijuan Xiao
show 1 more author
Hong Li
This is my paper · ORCID
classification cond-mat.mtrl-sci
keywords high-entropysolidelectrolytesmachinelearninginteratomicpotentialsCHGNetionicconductivityNASICONLi3Zr2Si2PO12structure-propertypredictionhigh-throughputscreening
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

This paper tries to show that a two-stage machine-learning pipeline can replace trial-and-error in the search for high-entropy solid electrolytes. The first stage fine-tunes a pretrained neural-network interatomic potential on DFT data for the NASICON electrolyte Li$_3$Zr$_2$Si$_2$PO$_{12}$ (LZSP), so that atomic relaxations and molecular dynamics run at near-DFT accuracy but at a fraction of the cost. The second stage maps simple relaxed-structure features, lithium content, cell parameters, and oxygen-coordination distortions, onto lithium mean squared displacement, producing a fast predictor for ionic transport. Applied to 4,575 quinary high-entropy LZSP compositions, the pipeline singles out Li$_{2.625}$Zr$_{0.25}$Hf$_{0.1875}$Sn$_{0.1875}$Ti$_{0.1875}$Nb$_{0.1875}$Si$_2$PO$_{12}$, whose extrapolated room-temperature conductivity is 4.53 mS/cm, about three orders of magnitude above pristine LZSP, with a migration barrier reduced from 0.46 to 0.24 eV. If true, this would cut the evaluation of one candidate composition from days to minutes and give a transferable recipe for other solid-electrolyte families.

What carries the argument

The load-bearing machinery is a dual-stage pipeline. Stage one is a fine-tuned CHGNet interatomic potential: the pretrained universal potential is refined first for structural relaxation and then for molecular dynamics on DFT-generated images of LZSP-family structures, including images chosen specifically for lithium transition-state geometry. Stage two is the SF-MSD model, a regression on 17 structural descriptors of the relaxed cell that outputs the lithium mean squared displacement, replacing direct molecular-dynamics screening. The bridge between stages is the assumption that relaxed-structure descriptors encode the bottleneck geometry and percolation network that control diffusion. The final conductivity figure comes from an Arrhenius fit of CHGNet molecular dynamics at elevated temperatures down to 300 K.

What would settle it

Synthesize LZHSTNSP and measure its ionic conductivity at 300 K, or run the same CHGNet molecular dynamics at 400, 600, and 800 K and check whether the Arrhenius plot remains linear with the same 0.24 eV slope; a bend in the plot or a measured conductivity far from 4.53 mS/cm would falsify the extrapolation.

Watch

Extended reading notes

Core claim

The central claim is that ionic transport in high-entropy LZSP can be predicted from the relaxed structure alone, without running expensive molecular dynamics for every candidate. By fine-tuning CHGNet on a purpose-built dataset of relaxation paths and lithium transition-state images, the authors obtain a potential that reproduces DFT energies, forces, cell volumes within about 1%, and molecular-dynamics mean-squared-displacement curves for doped and high-entropy variants. On top of that, the SF-MSD model uses structural descriptors such as lithium content, unit-cell parameters, bond lengths, and continuous symmetry measures of the oxygen polyhedra to predict the 100-ps, 1000 K mean squared displacement, with gradient-boosting and random-forest variants outperforming a compressed-sensing descriptor baseline on out-of-sample quaternary and multinary tests. The framework's payoff is the identification of LZHSTNSP in the 4,575-composition quinary space, with a predicted migration barrier of 0.24 eV and extrapolated 300 K conductivity of 4.53 mS/cm. The paper also uses feature-importance attribution to argue that the conductivity gain comes from a favorable lithium vacancy concentration together with uniform Zr-site octahedral distortions, not from any single dopant.

Load-bearing premise

The headline room-temperature conductivity assumes that lithium diffusion follows the same unchanged mechanism from the 1000 K simulation temperature down to 300 K, without any phase transition or ordering effect in between.

Editorial extensions

If this is right

  • Evaluation time per candidate drops from days to minutes, so a 4,575-composition space becomes screenable in one pass.
  • The fine-tuned potential and SF-MSD predictor are released as a queryable dataset, letting researchers search high-entropy LZSP compositions by element, proportion, and formula.
  • The identified composition LZHSTNSP is a concrete candidate for experimental synthesis and impedance testing.
  • The same two-stage recipe is claimed to transfer to other high-entropy solid electrolytes such as garnets and argyrodites, and to other properties of high-entropy cathode materials.
  • The feature-importance analysis identifies lithium vacancy concentration and octahedral distortion uniformity as design levers, rather than composition alone.

Reading between the lines

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

  • The paper reports the 4.53 mS/cm value as an extrapolation; a natural next step the authors do not take is to run CHGNet molecular dynamics at intermediate temperatures such as 400, 600, and 800 K to check whether the Arrhenius line stays straight and whether the same diffusion mechanism persists.
  • Because the SF-MSD model is trained on 1000 K mean-squared-displacement values, applying it to room-temperature design implicitly assumes that ranking by high-temperature MSD equals ranking by room-temperature conductivity, an assumption that could be tested directly by comparing predicted and measured conductivities on a handful of compositions.
  • A similar descriptor-based shortcut could be built for other framework families, but the 17 selected descriptors would need re-derivation because bottleneck geometry and coordination chemistry differ.
  • The framework does not include thermodynamic stability or synthesis feasibility, so composition screening would need to be paired with phase-stability checks before experimental follow-up.
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

4 major / 5 minor

Summary. The manuscript proposes a dual-stage machine-learning framework for discovering high-entropy solid electrolytes in the Li3Zr2Si2PO12 (LZSP) family. In the first stage, a pretrained CHGNet potential is fine-tuned on DFT relaxation and MD data to provide fast structural relaxation and molecular dynamics. In the second stage, a structure-feature-to-mean-squared-displacement (SF-MSD) surrogate model, comparing SISSO, gradient boosting, and random forest, is trained on CHGNet MD results and used to screen 4575 quinary HE-LZSP compositions. The authors identify Li2.625Zr0.25Hf0.1875Sn0.1875Ti0.1875Nb0.1875Si2PO12 (LZHSTNSP) and report an extrapolated room-temperature ionic conductivity of 4.53 mS/cm, three orders of magnitude higher than pristine LZSP, with a migration barrier reduced from 0.46 to 0.24 eV. The paper also presents SHAP analyses linking transport to features such as lithium vacancy concentration and octahedral distortion.

Significance. If the screening results are reliable, this work would provide a practical template for accelerating discovery in large composition spaces of high-entropy solid electrolytes, where exhaustive DFT MD is intractable. The framework is appealing in combining a fine-tuned universal potential with a lightweight interpretable surrogate, and the authors report several useful validation checks: CHGNet relaxation errors are compared against DFT for selected compositions, an AIMD comparison is provided for a quaternary structure, and model selection among SISSO, GB, and RF is documented with test-set RMSE values. The manuscript also makes a substantial computational dataset available (about 500 MD-simulated HE-LZSP compositions and over 5000 relaxed structures with predicted MSDs), which is a clear strength. However, the central quantitative discovery claim rests on an Arrhenius extrapolation whose details and uncertainties are not reported, and the validation chain is largely circular because the surrogate is trained and evaluated on the same MLIP-MD methodology used for the final candidate. These issues must be resolved before the quantitative conductivity claim can be accepted.

major comments (4)
  1. [§2.3, Fig. 6(c); §4.2] The central claim that LZHSTNSP has a room-temperature ionic conductivity of 4.53 mS/cm and a migration barrier of 0.24 eV is obtained by an Arrhenius extrapolation, but the manuscript does not report the number of MD temperatures used for the fit, the temperature range, the fitted prefactor, the goodness of fit, or any uncertainty on the extracted barrier and intercept. Section 4.2 only states that 'all simulations were set at 1000 K with the exception of the final MD simulation that was used to fit the Arrhenius curve.' This omission is load-bearing: the conductivity depends exponentially on the barrier, so a plausible ±0.05 eV uncertainty in the fitted slope changes the 300 K extrapolation by a factor of roughly exp(0.05/k_B·300 K) ≈ 7. Without reporting the fit details and statistical uncertainties, the 'three orders of magnitude' improvement over pristine LZSP cannot be critically assessed.
  2. [§2.3; Supplementary Fig. S10] The only AIMD consistency check for the LZSP framework is performed on pristine LZSP at 1000 K and 1100 K (Fig. S10), not on the candidate LZHSTNSP or at lower temperatures. All CHGNet MD data in the dataset are generated at 1000 K (except the final Arrhenius simulation, whose temperature is not specified). Extrapolating over roughly 700 K assumes that the diffusion mechanism, attempt frequency, and correlation factor are temperature-independent and that no phase transition or order–disorder change occurs. None of these assumptions is tested. To support the quantitative room-temperature claim, the authors should provide at least one independent check for the candidate, such as AIMD or DFT NEB barrier calculations, or explicitly reframe the reported value as a high-temperature screening metric rather than an established 300 K conductivity.
  3. [§2.2, §2.3; Table 2; Fig. 6(a,b)] The validation of the SF-MSD model is circular with respect to the transport property being predicted. The model is trained on MSD values obtained from fine-tuned CHGNet MD, tested on further fine-tuned CHGNet MD, and the final candidate is selected by the surrogate and then 'verified' by the same fine-tuned CHGNet MD potential. Consequently, systematic errors in CHGNet's prediction of ionic transport are not probed. Moreover, Table 2 reports multinary test-set RMSEs of 20.3 Ų (GB) and 20.1 Ų (RF), with quaternary test-set RMSEs of 23.9–24.8 Ų, and Fig. 6(a) states that the largest absolute errors occur in the high-MSD region, which is exactly where the candidate lies. The authors should report the per-region errors, the actual MD MSD for LZHSTNSP, and an independent validation (e.g., AIMD or DFT NEB) for the candidate to break the circularity.
  4. [§2.3] The screening procedure for the 4575 quinary compositions is not fully specified. The manuscript categorizes predicted MSD values into high, medium, and low regions (MSD>120, 40–120, <40 Ų) and selects 10 random structures per region for MD validation, but it does not state how the final candidate LZHSTNSP was chosen from the high-MSD region, nor does it quantify the screening's false-discovery rate. Given that only 30 of 4575 candidates are MD-verified, the paper should report the selection rule and show how many of the 30 validation samples were correctly classified, so the reader can assess the reliability of the high-throughput screening claim.
minor comments (5)
  1. [Fig. 4 caption] The caption contains a typo ('Comparation') and should define the notation 'Lix', 'csm_Xavg', 'csm_Xsd', and 'X-O' explicitly.
  2. [§2.2] The sentence 'This dataset was partitioned, allocating a small subset as the training set for the SF-MSD model and the majority as the predicted targets' is unclear; it should specify the number of training, validation, and test compositions and clarify the role of the 'predicted targets.'
  3. [§4.2] The MSD analysis protocol is described for the 100 ps dataset runs (averaging the 20–80 ps window), but it is not stated whether the same protocol was used for the final Arrhenius MD simulations. This should be clarified for reproducibility.
  4. [Supplementary Information] The figure numbering in the supplementary material is disordered (Fig. S10 appears after Fig. S11 in the supplied text), and the caption for Fig. S10 should be placed with the figure. Please renumber all supplementary figures consistently.
  5. [§2.3] The sentence 'The results confirm that the current model maintains strong predictive performance, as effective dataset generation strikes an optimal balance between computational cost and model accuracy' is vague; the cross-validation RMSEs (28.7 Ų for GB, 26.1 Ų for RF) are not compared statistically with the main model's RMSEs, so the claim of 'optimal balance' is not supported by a quantitative comparison.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the screening surrogate is validated on held-out CHGNet MD data and the final conductivity claim comes from direct CHGNet MD Arrhenius analysis, not from the surrogate itself.

full rationale

The paper's derivation chain is not circular under the rules used here. Stage 1 fine-tunes the pretrained CHGNet MLIP on DFT relaxation and MD data, with generalization checks against DFT (Table 1) and AIMD (Fig. 3, Fig. S10); this is an external anchor for the interatomic potential. Stage 2 trains the SF-MSD model on CHGNet-derived MSD labels and evaluates it on held-out multinary and quaternary sets (Table 2) plus 30 randomly selected quinary structures (Fig. 6a,b). This is a standard supervised-learning validation: the surrogate is a learned approximation of CHGNet MD, not an algebraic identity, and the held-out errors are empirical rather than forced. The central discovery claim for LZHSTNSP does not rest on the surrogate's fitted value: the 4.53 mS/cm room-temperature conductivity is stated to be 'extrapolated from the Arrhenius curve' obtained from MD simulations, and the migration barrier is the slope of that same Arrhenius fit. While the Arrhenius fit's temperature range, number of points, and uncertainties are not reported, and the candidate itself is not cross-checked against AIMD, those are validation and reporting limitations rather than circularity. No load-bearing self-citation, imported uniqueness theorem, or ansatz-by-citation was found. The self-citations ([11], [12]) are background references only. Consequently, no step reduces by construction to its own input.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The computational results rest entirely on modeled physics: CHGNet as a stand-in for DFT, a reduced set of structural descriptors as complete for transport, and Arrhenius extrapolation across a roughly 700 K gap. These are domain assumptions, not derived facts.

free parameters (3)
  • Arrhenius pre-exponential factor and activation energy for LZHSTNSP = Ea = 0.24 eV; pre-factor not reported
    Fitted to CHGNet MD MSD at elevated temperatures; the extrapolated 4.53 mS/cm depends on them (Section 2.3).
  • SF-MSD model hyperparameters (gradient boosting / random forest) = Not reported (chosen via scikit-learn default or randomized search)
    The surrogate's predictive performance, and therefore the screening results, depend on these fitted model parameters (Section 4.4).
  • MSD classification thresholds for high/medium/low regions = 120 and 40 Angstrom^2
    Hand-chosen cutoffs (Section 2.3) determine which structures receive MD validation and thus influence reported validation statistics.
assumptions (5)
  • domain assumption CHGNet pretrained on Materials Project, after fine-tuning on a small LZSP DFT dataset, is an accurate surrogate for DFT across the HE-LZSP composition space.
    Invoked in Sections 2.1 and 4.3. Accuracy is demonstrated on a small test set and on two AIMD comparisons, but generalization to arbitrary HE compositions is assumed.
  • domain assumption The 17 structural features (Li content, cell volume, X-O bond lengths, CSM values) are sufficient to determine the lithium MSD in HE-LZSP.
    The SF-MSD model in Sections 2.2 and 2.3 uses only these features; no proof that omitted features (e.g., exact cation ordering) are irrelevant.
  • domain assumption Ionic conduction follows the Arrhenius relation over the entire temperature range from simulation temperatures to 300 K, with a single activation energy.
    The room-temperature conductivity of LZHSTNSP in Section 2.3 is extrapolated from high-temperature MD via an Arrhenius fit; a change in diffusion mechanism would invalidate the value.
  • domain assumption The MSD averaged over the 20 to 80 ps window of a 100 ps NVT MD at 1000 K is a converged estimate of the ionic diffusion coefficient.
    Section 4.2 states the averaging protocol; the representativeness for all 500 simulated compositions is assumed.
  • domain assumption The ideal configuration entropy Sconfig is a meaningful feature for predicting transport in high-entropy systems.
    Sconfig is added as a feature in the multinary model (Section 2.3); its physical relevance is posited, not derived.

how reviews work

0 comments
Cite this review

Pith. "Pith review of High-Entropy Solid Electrolytes Discovery: A Dual-Stage Machine Learning Framework Bridging Atomic Configurations and Ionic Transport Properties." pith.science (2026). https://pith.science/paper/HWXPRQPU

@misc{pith2026250518571,
  author       = {Pith},
  title        = {Pith review of: High-Entropy Solid Electrolytes Discovery: A Dual-Stage Machine Learning Framework Bridging Atomic Configurations and Ionic Transport Properties},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HWXPRQPU}},
  note         = {Machine review of arXiv:2505.18571}
}
abstract

The rapid development of computational materials science powered by machine learning (ML) is gradually leading to solutions to several previously intractable scientific problems. One of the most prominent is machine learning interatomic potentials (MLIPs), which expedites the study of dynamical methods for large-scale systems. However, a promising field, high-entropy (HE) solid-state electrolytes (SEs) remain constrained by trial-and-error paradigms, lacking systematic computational strategies to address their huge and high-dimensional composition space. In this work, we establish a dual-stage ML framework that combines fine-tuned MLIPs with interpretable feature-property mapping to accelerate the high-entropy SEs discovery. Using Li$_3$Zr$_2$Si$_2$PO$_{12}$ (LZSP) as a prototype, the fine-tuned CHGNet-based relaxation provides atomic structure for each configuration, the structure features - mean squared displacement (SF-MSD) model predicts the ionic transport properties and identifies critical descriptors. The theoretical studies indicate that the framework can satisfy the multiple requirements including computational efficiency, generalization reliability and prediction accuracy. One of the most promising element combinations in the quinary HE-LZSP space containing 4575 compositions is identified with a high ionic conductivity of 4.53 mS/cm as an application example. The framework contains generalizability and extensibility to other SE families.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

54 extracted references · 27 canonical work pages

  1. [1]

    B. Dunn, H. Kamath, J.M. Tarascon, Electrical Energy Storage for the Grid: A Battery of Choices, Science 334 (2011) 928–935. https://doi.org/10.1126/science.1212741

    doi:10.1126/science.1212741 2011

  2. [2]

    Q. Zhao, S. Stalin, C. -Z. Zhao, L.A. Archer, Designing solid -state electrolytes for safe, energy-dense batteries, Nat. Rev. Mater. 5 (2020) 229 –252. https://doi.org/10.1038/s41578-019-0165-5

    work page doi:10.1038/s41578-019-0165-5 2020

  3. [3]

    Gao, A.M

    Y . Gao, A.M. Nolan, P. Du, Y . Wu, C. Yang, Q. Chen, Y . Mo, S.-H. Bo, Classical and Emerging Characterization Techniques for Investigation of Ion Transport Mechanisms in Crystalline Fast Ionic Conductors, Chem. Rev. 120 (2020) 5954 –6008. https://doi.org/10.1021/acs.chemrev.9b00747

    doi:10.1021/acs.chemrev.9b00747 2020

  4. [4]

    H. Wang, T. Fu, Y . Du, W. Gao, K. Huang, Z. Liu, P. Chandak, S. Liu, P. V an Katwyk, A. Deac, A. Anandkumar, K. Bergen, C.P. Gomes, S. Ho, P. Kohli, J. Lasenby, J. Leskovec, T.-Y . Liu, A. Manrai, D. Marks, B. Ramsundar, L. Song, J. Sun, J. Tang, P. Veličković, M. Welling, L. Zhang, C.W. Coley, Y . Bengio, M. Zitnik, Scientific discovery in the age of ar...

    doi:10.1038/s41586 2023

  5. [5]

    M. Liu, Z. Rong, R. Malik, P. Canepa, A. Jain, G. Ceder, K.A. Persson, Spinel compounds as multivalent battery cathodes: a systematic evaluation based on ab initio calculations, Energy Environ. Sci. 8 (2015) 964–974. https://doi.org/10.1039/C4EE03389B

    work page doi:10.1039/c4ee03389b 2015

  6. [6]

    Ong, V .L

    S.P. Ong, V .L. Chevrier, G. Hautier, A. Jain, C. Moore, S. Kim, X. Ma, G. Ceder, V oltage, stability and diffusion barrier differences between sodium -ion and lithium -ion intercalation materials, Energy Environ. Sci. 4 (2011) 3680. https://doi.org/10.1039/c1ee01782a

    work page doi:10.1039/c1ee01782a 2011

  7. [7]

    K.J. Kim, J. Wortman, S. -Y . Kim, Y . Qi, Atomistic Simulation Derived Insight on the Irreversible Structural Changes of Si Electrode during Fast and Slow Delithiation, Nano Lett. 17 (2017) 4330–4338. https://doi.org/10.1021/acs.nanolett.7b01389

    work page doi:10.1021/acs.nanolett.7b01389 2017

  8. [8]

    Banerjee, X

    A. Banerjee, X. Wang, C. Fang, E.A. Wu, Y .S. Meng, Interfaces and Interphases in All- Solid-State Batteries with Inorganic Solid Electrolytes, Chem. Rev. 120 (2020) 6878 –

    work page 2020

Show all 54 references
  1. [9]

    Y . Li, K. Leung, Y . Qi, Computational Exploration of the Li -Electrode|Electrolyte Interface in the Presence of a Nanometer Thick Solid-Electrolyte Interphase Layer, Acc. Chem. Res. 49 (2016) 2363–2370. https://doi.org/10.1021/acs.accounts.6b00363

    2016 doi

  2. [10]

    Cheng, B.V

    T. Cheng, B.V . Merinov, S. Morozov, W.A.I. Goddard, Quantum Mechanics Reactive Dynamics Study of Solid Li-Electrode/Li6PS5Cl-Electrolyte Interface, ACS Energy Lett. 2 (2017) 1454–1459. https://doi.org/10.1021/acsenergylett.7b00319

    2017 doi

  3. [11]

    Y . Wang, S. Wu, W. Shao, X. Sun, Q. Wang, R. Xiao, H. Li, Accelerated strategy for fast ion conductor materials screening and optimal doping scheme exploration, J. Materiomics 8 (2022) 1038–1047. https://doi.org/10.1016/j.jmat.2022.02.010

    2022 doi

  4. [12]

    X. Fu, Y . Wang, J. Xu, Q. Yang, H. Mao, R. Xiao, H. Li, First-principles study on a new chloride solid lithium -ion conductor material with high ionic conductivity, J. Mater. Chem.A 12 (2024) 10562–10570. https://doi.org/10.1039/D3TA07943K

    2024 doi

  5. [13]

    Jun, Y .Z

    K.J. Jun, Y .Z. Sun, Y . Xiao, Y . Zeng, R.H. Kim, H.Y . Kim, L.J. Miara, D.M. Im, Y . Wang, G. Ceder, Lithium superionic conductors with corner-sharing frameworks, Nat. Mater. 21 (2022) 924–931. https://doi.org/10.1038/s41563-022-01222-4

    2022 doi

  6. [14]

    Ouyang, Y

    B. Ouyang, Y . Zeng, The rise of high-entropy battery materials, Nat. Commun. 15 (2024)

    2024

  7. [15]

    Y . Feng, L. Yang, Z. Yan, D. Zuo, Z. Zhu, L. Zeng, Y . Zhu, J. Wan, Discovery of high entropy garnet solid -state electrolytes via ultrafast synthesis, Energy Storage Mater. 63 (2023) 103053. https://doi.org/10.1016/j.ensm.2023.103053

    2023

  8. [17]

    S. Wang, Y . Liu, Y . Mo, Frustration in super-ionic conductors unraveled by the density of atomistic states, Angew. Chem. Int. Ed. 62 (2023) e202215544. https://doi.org/10.1002/anie.202215544

    2023 doi

  9. [18]

    H. Wang, L. Zhang, J. Han, W. E, DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics, Comput. Phys. Commun. 228 (2018) 178–184. https://doi.org/10.1016/j.cpc.2018.03.016

    2018 doi

  10. [19]

    Batzner, A

    S. Batzner, A. Musaelian, L. Sun, M. Geiger, J.P. Mailoa, M. Kornbluth, N. Molinari, T.E. Smidt, B. Kozinsky, E(3) -equivariant graph neural networks for data -efficient and accurate interatomic potentials, Nat. Commun. 13 (2022) 2453. https://doi.org/10.1038/s41467-022-29939-5

    2022 doi

  11. [20]

    B. Deng, P. Zhong, K. Jun, J. Riebesell, K. Han, C.J. Bartel, G. Ceder, CHGNet as a pretrained universal neural network potential for charge -informed atomistic modelling, Nat. Mach. Intell. 5 (2023) 1031–1041. https://doi.org/10.1038/s42256-023-00716-3

    2023 doi

  12. [21]

    Musaelian, S

    A. Musaelian, S. Batzner, A. Johansson, L. Sun, C.J. Owen, M. Kornbluth, B. Kozinsky, Learning local equivariant representations for large -scale atomistic dynamics, Nat. Commun. 14 (2023) 579. https://doi.org/10.1038/s41467-023-36329-y

    2023 doi

  13. [22]

    Szczuka, B

    C. Szczuka, B. Karasulu, M.F. Groh, F.N. Sayed, T.J. Sherman, J.D. Bocarsly, S. Vema, S. Menkin, S.P. Emge, A.J. Morris, C.P. Grey, Forced Disorder in the Solid Solution Li 3 P–Li 2 S: A New Class of Fully Reduced Solid Electrolytes for Lithium Metal Anodes, J. Am. Chem. Soc. ...

    2022 doi

  14. [23]

    Huang, L

    J. Huang, L. Zhang, H. Wang, J. Zhao, J. Cheng, W. E, Deep potential generation scheme and simulation protocol for the Li 10 GeP 2 S 12 -type superionic conductors, J. Chem. Phys. 154 (2021) 094703. https://doi.org/10.1063/5.0041849

    2021 doi

  15. [24]

    W. Ye, H. Zheng, C. Chen, S.P. Ong, A Universal Machine Learning Model for Elemental Grain Boundary Energies, Scr. Mater. 218 (2022) 114803. https://doi.org/10.1016/j.scriptamat.2022.114803

    2022

  16. [25]

    Chang, I

    Y . Chang, I. Benlolo, Y . Bai, C. Reimer, D. Zhou, H. Zhang, H. Matsumura, H. Choubisa, X.-Y . Li, W. Chen, P. Ou, I. Tamblyn, E.H. Sargent, High-entropy alloy electrocatalysts screened using machine learning informed by quantum -inspired similarity analysis , Matter 7 (2024)...

    2024 doi

  17. [26]

    Pablo-García, S

    S. Pablo-García, S. Morandi, R.A. Vargas-Hernández, K. Jorner, Ž. Ivković, N. López, A. Aspuru-Guzik, Fast evaluation of the adsorption energy of organic molecules on metals via graph neural networks, Nat. Comput. Sci. 3 (2023) 433 –442. https://doi.org/10.1038/s43588-023-00437-y

    2023 doi

  18. [27]

    Ouyang, S

    R. Ouyang, S. Curtarolo, E. Ahmetcik, M. Scheffler, L.M. Ghiringhelli, SISSO: A compressed-sensing method for identifying the best low -dimensional descriptor in an immensity of offered candidates, Phys. Rev. Materials 2 (2018) 083802. https://doi.org/10.1103/PhysRevMaterials.2.083802

    2018 doi

  19. [28]

    H. Wang, R. Ouyang, W. Chen, A. Pasquarello, High-Quality Data Enabling Universality of Band Gap Descriptor and Discovery of Photovoltaic Perovskites, J. Am. Chem. Soc. 146 (2024) 17636–17645. https://doi.org/10.1021/jacs.4c03507

    2024 doi

  20. [29]

    Vazquez, P

    G. Vazquez, P. Singh, D. Sauceda, R. Couperthwaite, N. Britt, K. Youssef, D.D. Johnson, R. Arróyave, Efficient machine-learning model for fast assessment of elastic properties of high -entropy alloys, Acta Mater. 232 (2022) 117924. https://doi.org/10.1016/j.actamat.2022.117924

    2022

  21. [30]

    Bartel, C

    C.J. Bartel, C. Sutton, B.R. Goldsmith, R. Ouyang, C.B. Musgrave, L.M. Ghiringhelli, M. Scheffler, New tolerance factor to predict the stability of perovskite oxides and halides, Sci. Adv. 5 (2019) eaav0693

    2019

  22. [32]

    B. Deng, Y . Choi, P. Zhong, J. Riebesell, S. Anand, Z. Li, K. Jun, K.A. Persson, G. Ceder, Systematic softening in universal machine learning interatomic potentials, Npj Comput. Mater. 11 (2025) 1–9. https://doi.org/10.1038/s41524-024-01500-6

    2025 doi

  23. [33]

    He, Y .Z

    X.F. He, Y .Z. Zhu, A. Epstein, Y .F. Mo, Statistical variances of diffusional properties from ab initio molecular dynamics simulations, Npj Comput. Mater. 4 (2018) 1–9

    2018

  24. [34]

    Behler, Atom-centered symmetry functions for constructing high -dimensional neural network potentials, J

    J. Behler, Atom-centered symmetry functions for constructing high -dimensional neural network potentials, J. Chem. Phys. 134 (2011) 074106. https://doi.org/10.1063/1.3553717

    2011 doi

  25. [35]

    Bartók, R

    A.P. Bartók, R. Kondor, G. Csányi, On representing chemical environments, Phys. Rev. B 87 (2013) 184115. https://doi.org/10.1103/PhysRevB.87.184115

    2013 doi

  26. [36]

    Pinsky, D

    M. Pinsky, D. Avnir, Continuous symmetry measures. 5. The classical polyhedra, Inorg. Chem. 37 (1998) 5575–5582. https://doi.org/10.1021/ic9804925

    1998 doi

  27. [37]

    Kresse, J

    G. Kresse, J. Furthmüller, Efficient iterative schemes for ab initio total -energy calculations using a plane -wave basis set, Phys. Rev. B 54 (1996) 11169 –11186. https://doi.org/10.1103/PhysRevB.54.11169

    1996 doi

  28. [38]

    Blöchl, Projector augmented-wave method, Phys

    P.E. Blöchl, Projector augmented-wave method, Phys. Rev. B 50 (1994) 17953 –17979. https://doi.org/10.1103/PhysRevB.50.17953

    1994 doi

  29. [39]

    Perdew, K

    J.P. Perdew, K. Burke, M. Ernzerhof, Generalized Gradient Approximation Made Simple, Phys. Rev. Lett. 77 (1996) 3865–3868. https://doi.org/10.1103/PhysRevLett.77.3865

    1996 doi

  30. [40]

    Perdew, M

    J.P. Perdew, M. Ernzerhof, K. Burke, Rationale for mixing exact exchange with density functional approximations, J. Chem. Phys. 105 (1996) 9982 –9985. https://doi.org/10.1063/1.472933

    1996 doi

  31. [41]

    Jain, S.P

    A. Jain, S.P. Ong, G. Hautier, W. Chen, W.D. Richards, S. Dacek, S. Cholia, D. Gunter, D. Skinner, G. Ceder, K.A. Persson, Commentary: The Materials Project: A materials genome approach to accelerating materials innovation, APL Mater. 1 (2013) 011002. https://doi.org/10.1063/1.4812323

    2013 doi

  32. [42]

    Ong, W.D

    S.P. Ong, W.D. Richards, A. Jain, G. Hautier, M. Kocher, S. Cholia, D. Gunter, V .L. Chevrier, K.A. Persson, G. Ceder, Python Materials Genomics (pymatgen): A robust, open-source python library for materials analysis, Comput. Mater. Sci. 68 (2013) 314 –

    2013

  33. [43]

    Larsen, J.J

    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. Maxson, T...

    2017

  34. [44]

    Pedregosa, G

    F. Pedregosa, G. Varoquaux, G. Varoquaux, N. Org, A. Gramfort, A. Gramfort, V . Michel, V . Michel, L. Fr, B. Thirion, B. Thirion, O. Grisel, O. Grisel, M. Blondel, P. Prettenhofer, P. Prettenhofer, R. Weiss, V . Dubourg, V . Dubourg, J. Vanderplas, A. Passos, A. Tp, D. Courna...

    2011

  35. [45]

    Lundberg, S

    S.M. Lundberg, S. -I. Lee, A Unified Approach to Interpreting Model Predictions, in: Advances in Neural Information Processing Systems 30, Curran Associates, Inc., 2017: pp. 4765 –4774. https://proceedings.neurips.cc/paper_files/paper/2017/hash/8a20a8621978632d76c43df d28b6776...

    2017

  36. [49]

    Elements are selected by referring research [[1]] and [[2]], and the pristine LZSP structure is adopted from reference[[3]]

  37. [50]

    This approach produces a dataset with controlled size that better accommodates the large structural changes during the relaxation calculations

    For structural relaxation dataset, 50 images for each composition consist of the first 10 consecutive high-energy images and follow-by 40 random images from the DFT relaxation ionic steps. This approach produces a dataset with controlled size that better accommodates the large...

  38. [51]

    For MD dataset, it is important to find the structure of Li in the transition state from a range of structures, and this information is valuable for subsequent simulations of the ion ic migration process. In order to find the desired structures, the closest distance from each ...

  39. [52]

    If structures of excessively high energies are entered, this can lead to poor model performance or unstable iterations while fine-tuning CHGNet

    It is important to check whether high-energy images are abnormal structures in both of two datasets. If structures of excessively high energies are entered, this can lead to poor model performance or unstable iterations while fine-tuning CHGNet. Supplementary text 2 Discussion...

  40. [53]

    Y . Zeng, B. Ouyang, J. Liu, Y .-W. Byeon, Z. Cai, L.J. Miara, Y . Wang, G. Ceder, High- entropy mechanism to boost ionic conductivity, Science 378 (2022) 1320 –1324. https://doi.org/10.1126/science.abq1346

    2022 doi

  41. [54]

    Rao, K.K

    Y .B. Rao, K.K. Bharathi, L.N. Patro, Review on the synthesis and doping strategies in enhancing the Na ion conductivity of Na3Zr2Si2PO12 (NASICON) based solid electrolytes, Solid State Ionics 366 –367 (2021) 115671. https://doi.org/10.1016/j.ssi.2021.115671

    2021

  42. [55]

    L. Zhu, Y . Wang, J. Chen, W. Li, T. Wang, J. Wu, S. Han, Y . Xia, Y . Wu, M. Wu, F. Wang, Y . Zheng, L. Peng, J. Liu, L. Chen, W. Tang, Enhancing ionic conductivity in solid electrolyte by relocating diffusion ions to under -coordination sites, Sci. Adv. 8 (2 022) eabj7698. h...

    doi

  43. [56]

    Zheng, G

    W. Zheng, G. Liang, Q. Liu, J. Li, J.A. Yuwono, S. Zhang, V .K. Peterson, Z. Guo, The promise of high-entropy materials for high-performance rechargeable Li-ion and Na-ion batteries, Joule 7 (2023) 2732–2748. https://doi.org/10.1016/j.joule.2023.10.016. Fig S12. Waterfall plot...

    2023 doi

  44. [319]

    https://doi.org/10.1016/j.commatsci.2012.10.028

    2012 doi

  45. [973]

    https://doi.org/10.1038/s41467-024-45309-9

    doi

  46. [6933]

    https://doi.org/10.1021/acs.chemrev.0c00101

    doi

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.