Pith. sign in

REVIEW 4 major objections 5 minor 88 references

Dynamic Vacancy Levels in CsPbCl3 Obey Equilibrium Defect Thermodynamics

T0 review · 4 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read In CsPbCl3, a chloride vacancy's optical level swings by more than 1 eV at 300 K, yet its thermodynamic charge transition level and capture barriers barely move, vindicating static 0 K defect theory.

desk verdict A well-executed MLFF study that makes a plausible but not yet airtight case that static 0 K defect thermodynamics holds in CsPbCl3, with the harmonic entropy term as the main soft spot. read the letter →

arxiv 2508.14513 v2 pith:WMXVZRNN submitted 2025-08-20 cond-mat.mtrl-sci

Irea Mosquera-Lois , Aron Walsh This is my paper · ORCID
classification cond-mat.mtrl-sci
keywords CsPbCl3halidevacancydefectthermodynamicschargetransitionlevelmachinelearningforcefieldperovskitesoftnessnon-radiativerecombinationfinite-temperaturetheory
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 asks whether the usual zero-temperature, static picture of semiconductor defects survives in a soft, dynamically disordered halide perovskite at room temperature. It follows the chloride vacancy in CsPbCl3 through 300 K molecular dynamics using a machine-learned force field that matches the accuracy of hybrid density-functional theory with spin-orbit coupling. The vacancy's optical transition level swings by more than 1 eV, but the thermodynamic charge transition level shifts only from 2.71 eV at 0 K to 2.66 eV at 300 K, and the hole-capture barrier stays at 2.53 eV. The authors conclude that thermal fluctuations do not change the vacancy's thermodynamic behavior: VCl is not the non-radiative recombination center it has been blamed for, and its performance impact comes from limiting voltage and promoting ion migration. The broader conclusion is that static 0 K defect theory remains a reliable basis for predicting defect thermodynamics in soft perovskites.

What carries the argument

The mechanism that carries the argument is the configurational-coordinate separation between two kinds of charge transition: the optical level, a vertical energy difference at fixed geometry that naturally jumps by more than 1 eV as the lattice breathes, and the thermodynamic level, a free-energy difference between relaxed charge states that sets capture barriers and equilibrium occupation. The tool that makes the 300 K calculation possible is a multi-task machine-learning force field: one shared feature representation, trained mostly on cheap DFT, with separate readouts for the expensive hybrid-functional-plus-spin-orbit reference, enabling constant-pressure, constant-temperature (NPT) traj

What would settle it

Measure the +/0 charge transition level of the chloride vacancy at cryogenic and room temperature, for example by deep-level transient spectroscopy or temperature-dependent photoluminescence; a shift larger than about 0.1 eV would contradict the claimed near-invariance. Alternatively, recompute the 300 K free-energy difference with anharmonic thermodynamic-integration sampling instead of harmonic entropy; if the -109 meV vibrational term changes by more than tens of meV, the cancellation that saves the 0 K picture is an artifact.

Watch

Extended reading notes

Core claim

The central claim is that the large dynamic oscillations of the chloride vacancy's optical level do not affect the quantities that actually control device behavior. The optical level is an instantaneous vertical transition and therefore samples the soft potential energy surface, but non-radiative capture and thermodynamic charge transition levels are adiabatic, equilibrium quantities governed by potential-energy-surface crossings and free-energy differences. Comparing a static 0 K calculation with constant-pressure, constant-temperature molecular dynamics at 300 K, the thermodynamic +/0 transition level moves from 2.71 eV above the valence band maximum to 2.66 eV, and the hole-capture barrie

Load-bearing premise

The key assumption is that the room-temperature vibrational-entropy term (-109 meV) is computed accurately by a harmonic approximation; if that term is off by more than tens of meV in this soft, anharmonic crystal, the 0 K and 300 K levels would no longer agree.

Editorial extensions

If this is right

  • Static 0 K calculations suffice to predict thermodynamic charge transition levels and capture barriers in halide perovskites, removing the need for expensive finite-temperature simulations for many defect-design questions.
  • VCl in CsPbCl3 should be treated as a reversible electron trap that limits open-circuit voltage and promotes ionic migration, not as a non-radiative recombination center.
  • Large breadth in an optical transition level cannot by itself be taken as evidence of non-radiative recombination activity; the relevant quantity is the adiabatic capture barrier.
  • The multi-task machine-learning strategy makes hybrid-functional-with-spin-orbit accuracy feasible for molecular dynamics of defects where direct simulation is prohibitive.
  • Passivation strategies aimed at removing VCl should focus on voltage and stability gains rather than on eliminating non-radiative recombination.

Reading between the lines

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

  • Because the cancellation between volume relaxation and vibrational entropy is what saves the 0 K result, the same test in even softer perovskites, or at higher temperatures, could break; anharmonic free-energy sampling would show where the static approximation fails.
  • This logic suggests that time-resolved measurements of optical-level fluctuations, such as linewidth broadening or transient absorption, should not be used to infer trap-assisted recombination; recombination assignments should instead be tied to capture-barrier measurements.
  • The same multi-task training recipe could be extended to other charged defects and interfaces in heavy-element or strongly correlated materials, where a cheap functional gives poor energetics but a hybrid reference is too costly for dynamics.
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 paper addresses whether static 0 K defect theory remains valid for halide perovskites, focusing on the chloride vacancy in orthorhombic CsPbCl3. The authors train a multi-task (multi-fidelity) MACE machine-learning force field that reproduces HSE+SOC energies, and use it to run NPT molecular dynamics at 300 K. They report that the optical transition level ϵopt(0→+1) fluctuates by more than 1 eV, yet its ensemble average (1.65 eV) is close to the 0 K value (1.58 eV). For the thermodynamic transition level, they obtain 2.71 eV at 0 K and 2.66 eV at 300 K after including spin and harmonic vibrational entropies, concluding that the static 0 K formalism remains accurate. They also compute carrier capture barriers and find a small electron capture barrier (0.05 eV) but a large hole capture barrier (2.53 eV), implying VCl is not an efficient non-radiative recombination center. The paper further argues that VCl limits device performance through Fermi-level pinning and ion migration, consistent with the beneficial effect of Cl-rich conditions.

Significance. If the central claims hold, the paper resolves an important controversy: despite large thermal fluctuations of defect levels, equilibrium defect thermodynamics in halide perovskites can be described by the conventional static 0 K framework. The methodology is also valuable: a multi-task MLFF that combines low-cost PBE data with a small set of HSE+SOC reference calculations is a practical route for defect simulations at high levels of theory. The authors use established open-source tools, report MLFF validation on energies/forces/stresses and on 0 K PES interpolation, and provide a transparent decomposition of finite-temperature contributions to the transition level. The strength of the paper is its clear separation of optical (vertical) and thermodynamic (adiabatic) quantities, and its direct test of the static approximation. The main weakness is that the central quantitative conclusion (50 meV agreement between 0 K and 300 K) is supported by a harmonic vibrational entropy term with no reported uncertainty, in a system known to be soft and anharmonic; this is the load-bearing approximation and it is not stress-tested.

major comments (4)
  1. [§II D, Eq. (2); Methods, Molecular dynamics] The central 0 K-vs-300 K agreement rests on the cancellation between +50 meV volume relaxation, -18 meV spin entropy, and -109 meV vibrational entropy. The -109 meV term is computed with phonopy in the harmonic approximation, but no details are given of the structure used (0 K relaxed volume, NPT average volume, or charge-state-specific volumes), the q-point sampling, or the convergence with supercell size. The MLFF is validated for energies/forces/stresses, but not for phonon frequencies. In a soft, anharmonic crystal at 300 K, harmonic entropy errors of tens of meV are plausible and would erase the 50 meV agreement. Please report the phonopy setup, the resulting entropy convergence, and an estimate of anharmonic contributions (e.g., two-phase thermodynamics, thermodynamic integration, or temperature-dependent phonon renormalization).
  2. [§II C and §III] The paper states that non-radiative capture barriers are 'not affected' by thermal dynamics, but the capture barriers (En=0.05 eV, Ep=2.53 eV) are computed from the static 0 K configurational-coordinate diagram using CarrierCapture on an 80-atom supercell. No finite-temperature free-energy barrier calculation is presented, and no direct test of whether the 300 K dynamics alter the crossing of the PES is given. Given that ϵopt fluctuates by >1 eV, the claim that capture kinetics are unaffected is not directly demonstrated and is load-bearing for the conclusion that VCl is not a non-radiative recombination center. A finite-temperature barrier estimate (e.g., umbrella sampling along the capture coordinate, or at least a justification from the 300 K thermodynamic level agreement) is needed.
  3. [§IV, Point defect calculations] The eFNV finite-size correction is computed for the static lowest-energy defect configuration and applied a posteriori to every MD frame, as the authors acknowledge. The correction is quoted as 60 meV for the 80-atom supercell, which is larger than the 50 meV agreement margin between 0 K and 300 K. Because the effective screening depends on the instantaneous configuration, the constant-correction approximation could shift the 300 K transition level by tens of meV. Please estimate the configuration dependence of the correction (e.g., by evaluating Ecorr for representative MD snapshots, or by testing a larger supercell for a subset of configurations) and include this in the uncertainty budget.
  4. [§II B] The agreement between the ensemble-averaged optical level at 300 K (1.65 eV) and the 0 K value (1.58 eV) is stated without statistical uncertainty. The trajectory is 200 ps (after 50 ps equilibration) and Figure A4 shows dominant low-frequency content near 0.12 THz, so the sampling may be marginal for fully converged averages. Please report the standard error of the mean and the equilibration check for the optical level and for the mean energies used in Eq. (2).
minor comments (5)
  1. [§II D, Eq. (2)] The notation ϵthermo(+1/0) is used interchangeably with ϵ0→+1 and ϵthermo; please define the charge-state ordering consistently. Also, Eq. (2) omits the temperature dependence of ϵVBM; this is stated in the text, but a brief note on its expected magnitude would help the reader judge the approximation.
  2. [§IV, Machine learning force fields] The description of the multi-task architecture is clear, but the term 'multi-task' is used for multi-fidelity learning within one charge state; later the authors say separate models are trained for VCl0 and VCl+1. Please clarify whether the model is a single multi-output model per charge state or two independent models, and how the HSE+SOC readout is selected during MD.
  3. [Figure 2 and Appendix A, Fig. A4] The Fourier transform of the optical level is shown only as a power spectrum with arbitrary units. Please specify the normalization and label the y-axis, and consider overlaying the phonon density of states to support the assignment of the low-frequency feature.
  4. [Introduction, ref. [6]] The phrase 'empirical defect tolerance' is used; this is a term of art in the field, but the reference [6] is a review. A brief definition or a pointer to the specific discussion in that review would improve accessibility.
  5. [Methods, DFT calculations] The HSE mixing parameter α=0.375 is set to reproduce the room-temperature band gap. Since the paper compares 0 K and 300 K properties, a brief comment on the consistency of using a 300 K-tuned α for 0 K defect levels would be useful.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the 300 K and 0 K defect levels are independent evaluations from the same DFT-trained MLFF; no target result is used as an input.

full rationale

The paper's central comparison—that the thermodynamic charge transition level of VCl in CsPbCl3 changes only from 2.71 eV (0 K static) to 2.66 eV (300 K with entropies)—is obtained by two independent evaluations using the same machine-learned force field. The MLFF is trained on reference total energies and forces from PBE and HSE+SOC DFT calculations, with validation on held-out configurations sampled from independent MD trajectories (Appendix A). The 0 K value comes from relaxed ground-state structures, while the 300 K value comes from NPT molecular dynamics averages plus spin and harmonic vibrational entropy corrections computed with phonopy (Eq. 2). These are not defined in terms of one another, and no parameter is fitted to make the two agree. The HSE mixing parameter α = 0.375 is set to reproduce the experimental band gap of CsPbCl3—an external benchmark, not the defect transition level or the capture barriers. The non-radiative capture barriers (En = 0.05 eV, Ep = 2.53 eV) are obtained from configurational-coordinate diagrams constructed from the same PES, but they are not fitted to the conclusion that VCl is not a recombination center; they are consistent with external experimental trapping data and with prior independent calculations for other halide vacancies. The paper's self-citations (e.g., ShakeNBreak, doped, prior MLFF defect methodology) are computational tools or reviews, not load-bearing arguments that reduce the central result to their own assertion. The acknowledged approximations—configuration-dependent eFNV corrections and the use of harmonic vibrational entropy in a soft anharmonic lattice—are accuracy/robustness concerns, not instances of circular reasoning. The derivation chain is therefore self-contained with respect to its inputs: the DFT reference data, the trained model, and the statistical-mechanical evaluation of free energies.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The central results rest on the adopted reference DFT level, the harmonic entropy approximation, the transferability of a static charge correction, the neglect of VBM(T), and the generalization accuracy of the MLFF. These are all stated or acknowledged, but only some are fully quantified.

free parameters (1)
  • HSE mixing parameter alpha and screening parameter omega = alpha=0.375, omega=0.1 A^-1
    Chosen to reproduce the 300 K bandgap of CsPbCl3, following Ref. 62; affects absolute defect-level positions but cancels in the 0K-vs-300K comparisons.
assumptions (5)
  • domain assumption HSE+SOC with alpha=0.375 accurately describes the potential energy surface and defect levels of CsPbCl3.
    The authors adopt this as the reference level of theory (Methods, DFT calculations); its accuracy is asserted from prior literature and band-gap fitting, not demonstrated here.
  • domain assumption Vibrational entropy is well approximated by the harmonic approximation at 300 K.
    Used to compute the -109 meV entropy term in Eq. 2 via phonopy; the material is soft and anharmonic, so this premise is fragile.
  • domain assumption The eFNV charge correction computed for the static defect geometry applies to all MD configurations.
    Acknowledged in Methods (Point defect calculations) as an approximation; the correction should depend on configuration but is small (60 meV for 80-atom supercell).
  • domain assumption The temperature dependence of the bulk band edges is negligible for the intrinsic defect comparison.
    Stated in Section II.D after Eq. 2; affects the absolute value of the 300 K transition level but not the comparison of intrinsic defect energetics.
  • domain assumption The machine-learned force fields generalize accurately beyond their training configurations.
    Validated on test-set MD frames and interpolation paths (Appendix A), but quantitative errors are in the unreleased SI.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dynamic Vacancy Levels in CsPbCl3 Obey Equilibrium Defect Thermodynamics." pith.science (2026). https://pith.science/paper/WMXVZRNN

@misc{pith2026250814513,
  author       = {Pith},
  title        = {Pith review of: Dynamic Vacancy Levels in CsPbCl3 Obey Equilibrium Defect Thermodynamics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WMXVZRNN}},
  note         = {Machine review of arXiv:2508.14513}
}
read the original abstract

Halide vacancies are the dominant point defects in perovskites with VCl identified as a detrimental trap for the optoelectronic performance of CsPbCl3, with applications ranging from photodetectors to solar cells. Understanding these defects under operating conditions is key since their electronic levels exhibit large thermal fluctuations that challenge the validity of static 0 K models. However, quantitative modelling of defect processes requires hybrid density functional theory with spin-orbit coupling, which is too expensive for direct molecular dynamic simulations. To address this, we train a multi-task machine learning force field to study VCl in orthorhombic CsPbCl3 at 300 K. While we observe strong oscillations in the optical transition level arising from the soft potential energy surface, neither the non-radiative capture barriers nor the thermodynamic charge transition levels are affected. Our results reveal that VCl is not responsible for the non-radiative losses previously assumed. Instead, its impact on performance arises from other mechanisms, such as limiting the open-circuit voltage and promoting ionic migration. Our findings demonstrate that, despite strong dynamical effects in halide perovskites, the conventional static formalism of defect theory remains valid for predicting thermodynamic behavior, providing a sound basis for the design of high-performance energy materials.

Figures

Figures reproduced from arXiv: 2508.14513 by the authors.

Figure 1
Figure 1. FIG. 1: a) Errors in the relative energies of [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2: a) Variations of the optical defect level [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3: a) Schematic energy-level diagram illustrating two types of charge transitions. The optical transition level [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

88 extracted references · 80 canonical work pages

  1. [1]

    Mosquera-Lois, S

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

    2023

  2. [2]

    D. A. Egger, A. Bera, D. Cahen, G. Hodes, T. Kirchartz, L. Kronik, R. Lovrincic, A. M. Rappe, D. R. Reichman and O. Yaffe, Adv. Mater., 2018, 30, 1800691

    2018

  3. [3]

    Zacharias, G

    M. Zacharias, G. Volonakis, F. Giustino and J. Even, npj Comp. Mater., 2023, 9, 153

    2023

  4. [4]

    Dubajic, J

    M. Dubajic, J. R. Neilson, J. Klarbring, X. Liang, S. A. Bird, K. C. Rule, J. E. Auckett, T. A. Selby, G. Tumen-Ulzii, Y. Lu et al., Nature Nanotech., 2025, 1–9

    2025

  5. [5]

    S. A. Seidl, X. Zhu, G. Reuveni, S. Aharon, C. Gehrmann, S. Caicedo-D´ avila, O. Yaffe and D. A. Egger, Phys. Rev. Mater., 2023, 7, L092401

    2023

  6. [6]

    Mosquera-Lois, Y.-T

    I. Mosquera-Lois, Y.-T. Huang, H. Lohan, J. Ye, A. Walsh and R. L. Z. Hoye, Nat Rev Chem, 2025, 9, 287–304

    2025

  7. [7]

    A. V. Cohen, D. A. Egger, A. M. Rappe and L. Kronik, J. Phys. Chem. Lett., 2019, 10, 4490–4498

    2019

  8. [8]

    J. Ran, B. Wang, Y. Wu, D. Liu, C. Mora Perez, A. S. Vasenko and O. V. Prezhdo, J. Phys. Chem. Lett., 2023, 14, 6028–6036

    work page 2023

Show all 88 references
  1. [9]

    L. Deng, J. Ran, B. Wang, A. Boziki, A. Tkatchenko, J. Jiang and O. V. Prezhdo, J. Phys. Chem. Lett., 2024, 15, 10465–10472

    2024

  2. [10]

    W. Chu, W. A. Saidi, J. Zhao and O. V. Prezhdo, Angew. Chem. Int. Ed., 2020, 59, 6435–6441

    2020

  3. [11]

    Liu, W.-H

    Y. Liu, W.-H. Fang and R. Long, J. Phys. Chem. Lett., 2024, 15, 3764–3771

    2024

  4. [12]

    Y. Shi, W. Chu, L. Zhang, B. Wang, W. A. Saidi, J. Zhao and O. V. Prezhdo, Chem. Mater., 2025, 37, 655–664

    2025

  5. [13]

    W. Chu, Q. Zheng, O. V. Prezhdo, J. Zhao and W. A. Saidi, Sci. Adv., 2020, 6, eaaw7453

    2020

  6. [14]

    Li, Y.-Y

    W. Li, Y.-Y. Sun, L. Li, Z. Zhou, J. Tang and O. V. Prezhdo, J. Am. Chem. Soc., 2018, 140, 15753–15763

    2018

  7. [15]

    B. Wang, W. Chu, Y. Wu, D. Casanova, W. A. Saidi and O. V. Prezhdo, J. Phys. Chem. Lett., 2022, 13, 5946–5952

    2022

  8. [16]

    Kang, Phys

    J. Kang, Phys. Rev. Mater., 2020, 4, 085405

    2020

  9. [17]

    C. Ming, H. Wang, D. West, S. Zhang and Y.-Y. Sun, J. Mater. Chem. A, 2022, 10, 3018–3024

    2022

  10. [18]

    Zhang, X

    J. Zhang, X. Zhang, M. E. Turiansky and C. G. Van de Walle, PRX Energy, 2023, 2, 013008

    2023

  11. [19]

    J. Kang, J. Li and S.-H. Wei, Appl. Phys. Rev., 2021, 8, 031302

    2021

  12. [20]

    Meggiolaro, S

    D. Meggiolaro, S. G. Motti, E. Mosconi, A. J. Barker, J. Ball, C. A. R. Perini, F. Deschler, A. Petrozza and F. D. Angelis, Energy Environ. Sci., 2018, 11, 702–713

    2018

  13. [21]

    Meggiolaro and F

    D. Meggiolaro and F. De Angelis, ACS Energy Lett., 2018, 3, 2206–2222

    2018

  14. [22]

    M.-H. Du, J. Phys. Chem. Lett., 2015, 6, 1461–1466

    2015

  15. [23]

    Zhang, M

    X. Zhang, M. E. Turiansky, J.-X. Shen and C. G. Van de Walle, J. Appl. Phys., 2022, 131, 090901

    2022

  16. [24]

    J. Heyd, G. E. Scuseria and M. Ernzerhof, J. Chem. Phys., 2003, 118, 8207–8215

    2003

  17. [25]

    Hirotsu, J

    S. Hirotsu, J. Phys. Soc. Jpn., 1971, 31, 552–560

    1971

  18. [26]

    M. I. Cohen, K. F. Young, T. Chang and W. S. Brower, Jr., J. Appl. Phys., 1971, 42, 5267–5272

    1971

  19. [27]

    Haeger, M

    T. Haeger, M. Ketterer, J. Bahr, N. Pourdavoud, M. Runkel, R. Heiderhoff and T. Riedl, J. Phys. Mater., 2020, 3, 024004

    2020

  20. [28]

    Y. He, C. C. Stoumpos, I. Hadar, Z. Luo, K. M. McCall, Z. Liu, D. Y. Chung, B. W. Wessels and M. G. Kanatzidis, J. Am. Chem. Soc., 2021, 143, 2068–2077

    2021

  21. [29]

    Dongre S, E

    S. Dongre S, E. E. Siddharthan, R. Thapa, S. Ramu and R. G. Balakrishna, ACS Appl. Nano Mater., 2023, 6, 13227–13237

    2023

  22. [30]

    D. Chen, J. Ding, Q. Tan, P. Yang, Y. Liu and Q. Wang, Phys. Scr., 2024, 99, 115911

    2024

  23. [31]

    D. E. Sommer, D. R. Gamelin and S. T. Dunham, Phys. Rev. Mater., 2022, 6, 025404

    2022

  24. [32]

    Ma, Y.-M

    J.-P. Ma, Y.-M. Chen, L.-M. Zhang, S.-Q. Guo, J.-D. Liu, H. Li, B.-J. Ye, Z.-Y. Li, Y. Zhou, B.-B. Zhang, O. M. Bakr, J.-Y. Zhang and H.-T. Sun, J. Mater. Chem. C, 2019, 7, 3037–3048

    2019

  25. [33]

    X. Zhu, L. Ge, Y. Wang, M. Li, R. Zhang, M. Xu, Z. Zhao, W. Lv and R. Chen, Adv. Opt. Mater., 2021, 9, 2100058

    2021

  26. [34]

    S. Seth, T. Ahmed, A. De and A. Samanta, ACS Energy Lett., 2019, 4, 1610–1618

    2019

  27. [35]

    Fiuza-Maneiro, J

    N. Fiuza-Maneiro, J. Ye, S. K. Sharma, S. Chakraborty, S. G´ omez-Gra˜ na, R. L. Z. Hoye and L. Polavarapu,ACS Energy Lett., 2025, 10, 1623–1632

    2025

  28. [36]

    Zheng, Y

    X. Zheng, Y. Hou, H.-T. Sun, O. F. Mohammed, E. H. Sargent and O. M. Bakr, J. Phys. Chem. Lett., 2019, 10, 2629–2640

    2019

  29. [37]

    L. Fei, L. Yang, P. Li and J. Ma, Adv. Opt. Mater., 2025, 13, 2402826

    2025

  30. [38]

    J. A. Peters, Z. Liu, M. C. De Siena, M. G. Kanatzidis and B. W. Wessels, J. Appl. Phys., 2022, 132, 035101

    2022

  31. [39]

    J. Ye, M. M. Byranvand, C. O. Mart ´ ınez, R. L. Z. Hoye, M. Saliba and L. Polavarapu,Angew. Chem. Int. Ed., 2021, 60, 21636–21660

    2021

  32. [40]

    Cheng, X

    S. Cheng, X. Zheng, Z. Hou, R. Hu, S. Jiang, S. Xi, G. Wen and X. Liu, J. Mater. Chem. C, 2022, 10, 5693–5706

    2022

  33. [41]

    Zhang, X

    Y. Zhang, X. Cheng, D. Tu, Z. Gong, R. Li, Y. Yang, W. Zheng, J. Xu, S. Deng and X. Chen, Angew. Chem. Int. Ed., 2021, 60, 9693–9698

    2021

  34. [42]

    G. H. Ahmed, J. K. El-Demellawi, J. Yin, J. Pan, D. B. Velusamy, M. N. Hedhili, E. Alarousu, O. M. Bakr, H. N. Alshareef and O. F. Mohammed, ACS Energy Lett., 2018, 3, 2301–2307

    2018

  35. [43]

    Yong, S.-Q

    Z.-J. Yong, S.-Q. Guo, J.-P. Ma, J.-Y. Zhang, Z.-Y. Li, Y.-M. Chen, B.-B. Zhang, Y. Zhou, J. Shu, J.-L. Gu, L.-R. Zheng, O. M. Bakr and H.-T. Sun, J. Am. Chem. Soc., 2018, 140, 9942–9951

    2018

  36. [44]

    Y. Li, Q. Liu, X. Liu, J. Feng, L. He, H. Li, C. Li and H. Zhang, J. Phys. Chem. Lett., 2021, 12, 10746–10752

    2021

  37. [45]

    Jiang, Z

    S. Jiang, Z. Hou, X. Zheng, Q. Wu, X. Yang, W. Cai, Y. Yi, S. Cheng, G. Wen and X. Liu, J. Mater. Chem. C, 2023, 11, 5667–5679

    2023

  38. [46]

    Brivio, K

    F. Brivio, K. T. Butler, A. Walsh and M. Van Schilfgaarde, Phys. Rev. B, 2014, 89, 155204

    2014

  39. [47]

    Umari, E

    P. Umari, E. Mosconi and F. De Angelis, Sci. Rep., 2014, 4, 4467. 11

    2014

  40. [48]

    Batatia, D

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

    2022

  41. [49]

    Wang, Y.-Q

    J. Wang, Y.-Q. Sun and D. Shi, Appl. Phys. Lett., 2023, 122, 041101

    2023

  42. [50]

    M. E. Turiansky, J. L. Lyons and N. Bernstein, Machine Learning Phonon Spectra for Fast and Accurate Optical Lineshapes of Defects, Preprint at https://arxiv.org/abs/2508.09113 (2025)

    2025 arXiv

  43. [51]

    W. Zia, M. Malekshahi Byranvand, T. Rudolph, M. Rai, M. Kot, C. Das, M. Kedia, M. Zohdi, W. Zuo, V. Yeddu, M. I. Saidaminov, J. I. Flege, T. Kirchartz and M. Saliba, ACS Energy Lett., 2024, 9, 1017–1024

    2024

  44. [52]

    Y. Zhao, C. Wang, T. Ma, L. Zhou, Z. Wu, H. Wang, C. Chen, Z. Yu, W. Sun, A. Wang, H. Huang, B. Zou, D. Zhao and X. Li, Energy Environ. Sci., 2023, 16, 2080–2089

    2023

  45. [53]

    Qiao, Y.-N

    S. Qiao, Y.-N. Wu, X. Yan, B. Monserrat, S.-H. Wei and B. Huang, Phys. Rev. B, 2022, 105, 115201

    2022

  46. [54]

    X. Wu, C. Ming, J. Shi, H. Wang, D. West, S. Zhang and Y.-Y. Sun, Chinese Phys. Lett., 2022, 39, 046101

    2022

  47. [55]

    Mosquera-Lois, J

    I. Mosquera-Lois, J. Klarbring and A. Walsh, Chem. Sci., 2025, 10.1039.D4SC08582E

    2025

  48. [56]

    S. R. Kavanagh, Journal of Physics: Energy, 2025, 7, 045002

    2025

  49. [57]

    Kresse and J

    G. Kresse and J. Furthm¨ uller,Comput. Mater. Sci., 1996, 6, 15–50

    1996

  50. [58]

    Kresse and J

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

    1993

  51. [59]

    Kresse and J

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

    1994

  52. [60]

    J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868

    1996

  53. [61]

    Steiner, S

    S. Steiner, S. Khmelevskyi, M. Marsmann and G. Kresse, Phys. Rev. B, 2016, 93, 224425

    2016

  54. [62]

    J. L. Lyons and M. W. Swift, J. Phys. Chem. C, 2023, 127, 12735–12740

    2023

  55. [63]

    Jinnouchi, J

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

    2019

  56. [64]

    $ #! ""!

    for Cs (5 s25p66s1), Pb (5 d106s26p2) and Cl (3 s23p5). We used two different exchange-correlation functionals to generate two separate datasets with PBE[ 60] and HSE+SOC[24, 61]. The mixing and screening parameters in the HSE functional were set to α = 0.375 and 0 .1 ˚A−1 in ...

  57. [65]

    Jinnouchi, F

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

    2019

  58. [66]

    Jinnouchi, K

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

    2020

  59. [67]

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

    2022

  60. [68]

    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

  61. [69]

    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

  62. [70]

    Kumagai and F

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

    2014

  63. [71]

    Mosquera-Lois and S

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

    2021

  64. [72]

    Mosquera-Lois, S

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

    2023

  65. [73]

    Mosquera-Lois, S

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

    2022

  66. [74]

    S. Kim, S. N. Hood, P. v. Gerwen, L. D. Whalley and A. Walsh, J. Open Source Softw., 2020, 5, 2102

    2020

  67. [75]

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

    2023

  68. [76]

    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

  69. [77]

    A. H. Larsen, J. J. Mortensen, J. Blomqvist, I. E. Castelli, R. Christensen, M. Du lak, 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

  70. [78]

    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

  71. [79]

    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

  72. [80]

    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

  73. [81]

    Himanen, M

    L. Himanen, M. O. J. J¨ ager, 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

  74. [82]

    Laakso, L

    J. Laakso, L. Himanen, H. Homm, E. V. Morooka, M. O. J¨ ager, M. Todorovi´ c and P. Rinke,J. Chem. Phys., 2023, 158, 234802

    2023

  75. [83]

    McInnes, J

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

    2018

  76. [84]

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

    2024

  77. [85]

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

    2007

  78. [86]

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

    2021

  79. [87]

    Stukowski, Model

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

    2010

  80. [88]

    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)

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.