pith. sign in

arxiv: 2605.31303 · v1 · pith:DXFRLK2Vnew · submitted 2026-05-29 · ❄️ cond-mat.mtrl-sci

A Practical Guide for Diagnosing Imaginary Phonon Modes in Metal--Organic Frameworks: The Case of MOF-5

Julia Santana-Andreo , Caterina Cocchi This is my paper

Pith reviewed 2026-06-28 21:48 UTC · model grok-4.3

classification ❄️ cond-mat.mtrl-sci
keywords MOF-5phonon calculationsdynamical stabilitymetal-organic frameworksimaginary modesfinite-displacement methodnumerical convergence
0
0 comments X

The pith

Numerical settings like force thresholds and grid resolution can create spurious imaginary phonon modes in MOF-5 calculations.

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

The paper demonstrates that phonon calculations on metal-organic frameworks such as MOF-5 are highly sensitive to numerical choices, with force convergence thresholds, real-space grid resolutions, symmetry standardization, and unit-cell representations each capable of introducing or removing imaginary modes that have no physical basis. A finite-displacement workflow is laid out to vary these settings systematically until numerical noise is ruled out. Only the remaining imaginary modes are then examined through mode mapping and stochastic Monte Carlo symmetry-breaking to search for lower-energy local minima. This approach matters for distinguishing which computationally proposed frameworks are likely to be dynamically stable enough for experimental realization.

Core claim

Imaginary phonon modes in MOF-5 can be produced by insufficient force convergence, coarse real-space grids, inconsistent symmetry handling, or alternative unit-cell choices. A finite-displacement workflow isolates these artifacts by controlled variation of the settings; once they are excluded, any leftover imaginary modes are investigated with mode mapping or Monte Carlo symmetry-breaking distortions to identify lower-energy structures and thereby assess genuine dynamical stability.

What carries the argument

finite-displacement workflow that systematically varies numerical parameters to separate calculation artifacts from true lattice instabilities

If this is right

  • Many reported imaginary modes in MOF-5 disappear once force convergence and grid resolution are improved.
  • Symmetry-standardization choices and alternative unit-cell descriptions can reverse a stability prediction.
  • Mode mapping after numerical cleanup reveals whether an imaginary mode points to a nearby stable structure.
  • Stochastic Monte Carlo symmetry breaking locates lower-energy minima once numerical noise is removed.

Where Pith is reading between the lines

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

  • The same workflow could be tested on other large flexible frameworks where phonon results have been reported as unstable.
  • Some previously dismissed MOF structures might regain candidate status if re-run with tighter numerical controls.
  • Standardized numerical protocols might be needed before computational screening of MOFs can be considered reliable.

Load-bearing premise

Imaginary modes that survive after all numerical settings have been tightened still indicate real dynamical instability rather than some other hidden computational problem.

What would settle it

A set of MOF-5 phonon calculations that progressively tightens force thresholds and grid spacing until all imaginary modes disappear, or that finds a lower-energy minimum via Monte Carlo whose displacement pattern matches the original imaginary mode.

Figures

Figures reproduced from arXiv: 2605.31303 by Caterina Cocchi, Julia Santana-Andreo.

Figure 1
Figure 1. Figure 1: Conventional unit cell representation of MOF-5 and zoom-in of the metal–organic linker environment. 1 arXiv:2605.31303v1 [cond-mat.mtrl-sci] 29 May 2026 [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Workflow for diagnosing imaginary phonon modes in MOFs. potential energy surface. For the lattice dynamics calculations, the harmonic interatomic force constants (IFCs) are computed with phonopy using the finite-displacement approach [25, 17]. Following geometry optimization under tight convergence thresholds to minimize residual stress, the MOF-5 framework is standardized according to its space-group symm… view at source ↗
Figure 3
Figure 3. Figure 3: Phonon dispersions of MOF-5 calculated (a) using the r2SCAN functional in a primitive 2 × 2 × 2 supercell under maximum-force (MF) thresholds of 10−5 , 10−6 , and 10−7 Ha/bohr, and (b) with PBE on a conventional unit cell, with CUTOFF values of 2200 Ry, 2700 Ry, and 3200 Ry at fixed REL CUTOFF of 600 Ry. typically causes an artificial softening of the interatomic force-constant matrix, leading to spurious … view at source ↗
Figure 4
Figure 4. Figure 4: Influence of unit-cell standardization and grid parameters on the structural integrity and lattice dynamics of MOF-5 in (a) its primitive cell, (b) conventional cell, and (c) fragmented primitive cell. Panels (d–f) illustrate the corresponding 2 × 2 × 2 supercell expansions, and (g–i) display the resulting harmonic phonon dispersion relations. converged real-space grid and an overly loose symmetry toleranc… view at source ↗
Figure 5
Figure 5. Figure 5: Conventional cubic cell reference used for the lattice dynamics of MOF-5. (a) Symmetry-standardized conventional structure and (b) its corresponding harmonic phonon dispersion relation [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Effect of unit-cell and supercell representation on the harmonic phonon dispersion of MOF-5. Results of calculations performed for (a) the conventional unit cell, (b) the primitive unit cell, and (c) the primitive 2 × 2 × 2 supercell are shown. Sec. 4.2, the CUTOFF required to stabilize the softest phonon branches is highly sensitive to the spatial representation of the lattice. As illustrated in Fig. 6b, … view at source ↗
Figure 7
Figure 7. Figure 7: Mode mapping of the unstable A2g phonon mode in MOF-5. (a) One-dimensional potential-energy scan along the unstable normal coordinate Q. (b) Phonon dispersion and atom-projected phonon density of states for the high-symmetry Fm¯3m reference structure (blue), alongside the phonon dispersion of the symmetry-lowered Fm¯3 phase (red). 4.5 Mode mapping of soft phonon instabilities Once numerical artifacts have … view at source ↗
read the original abstract

Assessing the dynamical stability of computationally predicted metal--organic frameworks (MOFs) is essential to distinguish synthetically feasible structures from dynamically unstable ones. However, reliable first-principles phonon calculations on these systems remain challenging: their large, flexible unit cells and soft collective modes make the vibrational spectrum highly sensitive to the numerical settings. Using MOF-5 as a representative case study, we establish a finite-displacement workflow to identify and isolate the origins of imaginary phonon modes. We demonstrate how numerical force convergence thresholds, real-space grid resolutions, symmetry-standardization protocols, and alternative unit-cell representations can qualitatively and spuriously alter the predicted lattice stability. Once numerical noise is confidently excluded, the remaining imaginary modes can be analyzed through mode mapping or stochastic Monte Carlo symmetry-breaking distortions to locate lower-energy local minima. This protocol provides a robust, transferable strategy for the reliable assessment of dynamical stability and lattice vibrations in flexible porous frameworks.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The manuscript presents a finite-displacement workflow for diagnosing imaginary phonon modes in metal-organic frameworks, using MOF-5 as a case study. It claims that numerical settings including force convergence thresholds, real-space grid resolutions, symmetry-standardization protocols, and unit-cell representations can qualitatively induce spurious imaginary modes, and that once these are excluded, remaining modes indicate genuine dynamical instability that can be resolved by mode mapping or stochastic Monte Carlo symmetry-breaking distortions to locate lower-energy, dynamically stable minima. The abstract frames this as a robust, transferable protocol but supplies no quantitative results, convergence tests, or validation spectra.

Significance. A validated, practical guide for handling numerical artifacts in phonon calculations on large, flexible MOFs would address a recurring practical barrier in the field and improve reliability of dynamical stability assessments. The manuscript correctly identifies that standard methods are sensitive to settings in these systems, but the absence of any demonstrated before/after phonon spectra, energy comparisons, or confirmation that mapped/MC structures are free of imaginary modes under the same converged settings means the central claim of a reliable resolution protocol cannot yet be evaluated.

major comments (2)
  1. [Abstract] Abstract, final paragraph: the assertion that 'remaining imaginary modes can be analyzed through mode mapping or stochastic Monte Carlo symmetry-breaking distortions to locate lower-energy local minima' is load-bearing for the claimed protocol, yet no explicit validation is supplied (e.g., phonon spectra of the resulting structures showing absence of imaginary modes, or energy lowering relative to the original cell under identical converged settings).
  2. [Abstract] Abstract: the central claim that the workflow provides a 'robust, transferable strategy' rests on demonstrations that numerical settings 'can qualitatively and spuriously alter the predicted lattice stability,' but the text contains no quantitative data, convergence thresholds tested, or specific examples of how modes change with grid resolution, force tolerance, or cell choice.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. We address the major comments point-by-point below, clarifying the content of the full manuscript while agreeing to strengthen the abstract and add explicit quantitative summaries where helpful.

read point-by-point responses

  1. Referee: [Abstract] Abstract, final paragraph: the assertion that 'remaining imaginary modes can be analyzed through mode mapping or stochastic Monte Carlo symmetry-breaking distortions to locate lower-energy local minima' is load-bearing for the claimed protocol, yet no explicit validation is supplied (e.g., phonon spectra of the resulting structures showing absence of imaginary modes, or energy lowering relative to the original cell under identical converged settings).

    Authors: The full manuscript (Sections 3–4 and associated figures) already presents the phonon spectra of the mapped and Monte Carlo-distorted structures, confirming the absence of imaginary modes under the same converged settings, together with the computed energy lowerings relative to the original cell. The abstract is necessarily concise and does not repeat these details; we will revise the final paragraph to explicitly reference these validations. revision: yes

  2. Referee: [Abstract] Abstract: the central claim that the workflow provides a 'robust, transferable strategy' rests on demonstrations that numerical settings 'can qualitatively and spuriously alter the predicted lattice stability,' but the text contains no quantitative data, convergence thresholds tested, or specific examples of how modes change with grid resolution, force tolerance, or cell choice.

    Authors: The body of the manuscript supplies the requested demonstrations through explicit examples and figures showing the appearance or disappearance of imaginary modes as a function of force convergence threshold, real-space grid spacing, symmetry standardization, and choice of unit-cell representation. To make these quantitative aspects more immediately visible, we will add a concise summary table of the tested settings and their effects on the phonon spectrum. revision: yes

Circularity Check

0 steps flagged

No circularity: workflow applies standard finite-displacement phonon methods with external numerical benchmarks

full rationale

The paper describes a finite-displacement workflow to diagnose imaginary modes in MOF-5 phonons by varying force convergence, grid resolution, symmetry protocols, and unit-cell choices. These are presented as empirical demonstrations on external DFT calculations rather than any derivation, fitted parameter, or self-citation that reduces the central claim to its own inputs by construction. No equations or steps equate a 'prediction' to a fitted input, and the resolution suggestions (mode mapping, MC distortions) are framed as practical next steps without load-bearing self-referential definitions.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The paper relies on standard assumptions of density functional theory and finite-displacement phonon methods without introducing new free parameters or postulated entities.

axioms (2)
  • domain assumption Density functional theory provides accurate forces for finite-displacement phonon calculations when numerical parameters are converged
    Invoked throughout the workflow description for MOF-5.
  • domain assumption Symmetry standardization and unit-cell choice do not alter the underlying physical stability when properly handled
    Central to the claim that certain imaginary modes are spurious.

pith-pipeline@v0.9.1-grok · 5695 in / 1178 out tokens · 32049 ms · 2026-06-28T21:48:18.797728+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

47 extracted references · 44 canonical work pages

  1. [1]

    Yaghi O M, O’Keeffe M, Ockwig N W, Chae H K, Eddaoudi M and Kim J 2003Nature423 705–714 URLhttps://doi.org/10.1038/nature01650

    work page doi:10.1038/nature01650

  2. [2]

    Adv.4eaat9180 URL https://doi.org/10.1126/sciadv.aat9180

    Kalmutzki M J, Hanikel N and Yaghi O M 2018Sci. Adv.4eaat9180 URL https://doi.org/10.1126/sciadv.aat9180

    work page doi:10.1126/sciadv.aat9180

  3. [3]

    Chen Z, Kirlikovali K O, Li P and Farha O K 2022Acc. Chem. Res.55579–591 URL https://doi.org/10.1021/acs.accounts.1c00707

    work page doi:10.1021/acs.accounts.1c00707

  4. [4]

    Rev.112724–781 URLhttps://doi.org/10.1021/cr2003272

    Sumida K, Rogow D L, Mason J A, McDonald T M, Bloch E D, Herm Z R, Bae T H and Long J R 2012Chem. Rev.112724–781 URLhttps://doi.org/10.1021/cr2003272

    work page doi:10.1021/cr2003272

  5. [5]

    Nanotechnol.17911–923 URL https://doi.org/10.1038/s41565-022-01168-3

    Knebel A and Caro J 2022Nat. Nanotechnol.17911–923 URL https://doi.org/10.1038/s41565-022-01168-3

    work page doi:10.1038/s41565-022-01168-3

  6. [6]

    Li H, Eddaoudi M, O’Keeffe M and Yaghi O M 1999Nature402276–279 ISSN 1476-4687 URLhttps://doi.org/10.1038/46248

    work page doi:10.1038/46248

  7. [7]

    Eddaoudi M, Li H and Yaghi O M 2000J. Am. Chem. Soc.1221391–1397 URL https://doi.org/10.1021/ja9933386

    work page doi:10.1021/ja9933386

  8. [8]

    Lock N, Christensen M, Wu Y, Peterson V K, Thomsen M K, Piltz R O, Ramirez-Cuesta A J, McIntyre G J, Nor´ en K, Kutteh R, Kepert C J, Kearley G J and Iversen B B 2013 Dalton Trans.421996–2007 URLhttps://doi.org/10.1039/C2DT31491F

    work page doi:10.1039/c2dt31491f 2013

  9. [9]

    Rimmer L H N, Dove M T, Goodwin A L and Palmer D C 2014Phys. Chem. Chem. Phys.16 21144–21152 URLhttps://doi.org/10.1039/C4CP01701C

    work page doi:10.1039/c4cp01701c

  10. [10]

    Theory Simul.21900093 URL https://doi.org/10.1002/adts.201900093

    Ryder M R, Maul J, Civalleri B and Erba A 2019Adv. Theory Simul.21900093 URL https://doi.org/10.1002/adts.201900093

    work page doi:10.1002/adts.201900093

  11. [11]

    Edzards J, Santana-Andreo J, Saßnick H D and Cocchi C 2025J. Chem. Theory. Comput.21 7062–7074 pMID: 40653644 URLhttps://doi.org/10.1021/acs.jctc.5c00399

    work page doi:10.1021/acs.jctc.5c00399

  12. [12]

    Andreeva A B, Le K N, Chen L, Kellman M E, Hendon C H and Brozek C K 2020 J. Am. Chem. Soc.14219291–19299 URLhttps://doi.org/10.1021/jacs.0c09499

    work page doi:10.1021/jacs.0c09499 2020

  13. [13]

    Dubbeldam D, Walton K S, Ellis D E and Snurr R Q 2007Angew. Chem. Int. Ed.46 4496–4499 URLhttps://doi.org/10.1002/anie.200700218

    work page doi:10.1002/anie.200700218

  14. [14]

    Baroni S, de Gironcoli S, Dal Corso A and Giannozzi P 2001Rev. Mod. Phys.73515–562 URLhttps://doi.org/10.1103/RevModPhys.73.515

    work page doi:10.1103/revmodphys.73.515

  15. [15]

    Struct.4033002 URL https://doi.org/10.1088/2516-1075/ac78b3

    Pallikara I, Kayastha P, Skelton J M and Whalley L D 2022Electron. Struct.4033002 URL https://doi.org/10.1088/2516-1075/ac78b3

    work page doi:10.1088/2516-1075/ac78b3

  16. [16]

    Kamath P D and Persson K A 2026arXiv preprint arXiv:2602.07295URL https://arxiv.org/abs/2602.07295

    Pith/arXiv arXiv

  17. [17]

    Togo A 2023J. Phys. Soc. Jpn.92012001 URLhttps://doi.org/10.7566/JPSJ.92.012001 12 IOP PublishingJournalvv(yyyy) aaaaaa Authoret al

    work page doi:10.7566/jpsj.92.012001

  18. [18]

    Mater.10258 URLhttps://doi.org/10.1038/s41524-024-01437-w

    Zhu Z, Park J, Sahasrabuddhe H, Ganose A M, Chang R, Lawson J W and Jain A 2024 npj Comp. Mater.10258 URLhttps://doi.org/10.1038/s41524-024-01437-w

    work page doi:10.1038/s41524-024-01437-w 2024

  19. [19]

    Inhomogeneous electron gas,

    Hohenberg P and Kohn W 1964Phys. Rev.136(3B) B864–B871 URL https://link.aps.org/doi/10.1103/PhysRev.136.B864

    work page doi:10.1103/physrev.136.b864

  20. [20]

    Self-consistent equations including exchange and correlation effects,

    Kohn W and Sham L J 1965Phys. Rev.140(4A) A1133–A1138 URL https://link.aps.org/doi/10.1103/PhysRev.140.A1133

    work page doi:10.1103/physrev.140.a1133

  21. [21]

    Perdew J P, Burke K and Ernzerhof M 1996Phys. Rev. Lett.77(18) 3865–3868 URL https://doi.org/10.1103/PhysRevLett.77.3865

    work page doi:10.1103/physrevlett.77.3865

  22. [22]

    Furness J W, Kaplan A D, Ning J, Perdew J P and Sun J 2020J. Phys. Chem. Lett.11 8208–8215 URLhttps://doi.org/10.1021/acs.jpclett.0c02405

    work page doi:10.1021/acs.jpclett.0c02405

  23. [23]

    Grimme S, Antony J, Ehrlich S and Krieg H 2010J. Chem. Phys.132154104 ISSN 0021-9606 URLhttps://doi.org/10.1063/1.3382344

    work page doi:10.1063/1.3382344

  24. [24]

    Born M and Huang K 1954Dynamical Theory of Crystal Lattices(Oxford: Clarendon Press) URLhttps://doi.org/10.1093/oso/9780192670083.001.0001

    work page doi:10.1093/oso/9780192670083.001.0001

  25. [25]

    Togo A and Tanaka I 2015Scripta Mater.1081–5 URL https://doi.org/10.1016/j.scriptamat.2015.07.021

    work page doi:10.1016/j.scriptamat.2015.07.021 2015

  26. [26]

    Parlinski K, Li Z Q and Kawazoe Y 1997Phys. Rev. Lett.784063–4066 URL https://doi.org/10.1103/PhysRevLett.78.4063

    work page doi:10.1103/physrevlett.78.4063

  27. [27]

    Hoffman A E J, Senkovska I, Wieme J, Krylov A, Kaskel S and Van Speybroeck V 2022 J. Mater. Chem. A1017254–17266 URLhttps://doi.org/10.1039/D2TA01678H

    work page doi:10.1039/d2ta01678h 2022

  28. [28]

    Phys.234513–527 URL https://doi.org/10.1515/zkri-2018-2152

    Kuchta B, Formalik F, Rogacka J, Neimark A V and Firlej L 2019Z. Phys.234513–527 URL https://doi.org/10.1515/zkri-2018-2152

    work page doi:10.1515/zkri-2018-2152 2018

  29. [29]

    K¨ uhne T D, Iannuzzi M, Del Ben M, Rybkin V V, Hutter Jet al.2020J. Chem. Phys.152 194103 URLhttps://doi.org/10.1063/5.0007045

    work page doi:10.1063/5.0007045

  30. [30]

    Goedecker S, Teter M and Hutter J 1996Phys. Rev. B54(3) 1703–1710 URL https://link.aps.org/doi/10.1103/PhysRevB.54.1703

    work page doi:10.1103/physrevb.54.1703

  31. [31]

    VandeVondele J and Hutter J 2007J. Chem. Phys.127114105 ISSN 0021-9606 (Preprint https://pubs.aip.org/aip/jcp/article-pdf/doi/10.1063/1.2770708/14787836/ 114105_1_online.pdf) URLhttps://doi.org/10.1063/1.2770708

    work page doi:10.1063/1.2770708/14787836/

  32. [32]

    Lippert, J

    Lippert G, Hutter J and Parrinello M 1997Mol. Phys.92477–488 URL https://doi.org/10.1080/002689797170220

    work page doi:10.1080/002689797170220

  33. [33]

    VandeVondele J, Krack M, Mohamed F, Parrinello M, Chassaing T and Hutter J 2005 Comput. Phys. Commun.167103–128 URLhttps://doi.org/10.1016/j.cpc.2004.12.014

    work page doi:10.1016/j.cpc.2004.12.014 2005

  34. [34]

    Skelton J M 2017 Modemaphttps://github.com/JMSkelton/ModeMapaccessed: 2026-04-26 URLhttps://github.com/JMSkelton/ModeMap

    2017

  35. [35]

    Setyawan W and Curtarolo S 2010Comp. Mater. Sci.49299–312 URL https://doi.org/10.1016/j.commatsci.2010.05.010

    work page doi:10.1016/j.commatsci.2010.05.010 2010

  36. [36]

    Saßnick H D, Edzards J, Reents T and Cocchi C 2026arXiv preprint arXiv:2604.26551URL https://arxiv.org/abs/2604.26551

    Pith/arXiv arXiv

  37. [37]

    Adamo C and Barone V 1999J. Chem. Phys.1106158–6170 ISSN 0021-9606 (Preprint https: //pubs.aip.org/aip/jcp/article-pdf/110/13/6158/19068890/6158_1_online.pdf) URL https://doi.org/10.1063/1.478522

    work page doi:10.1063/1.478522

  38. [38]

    Sun J, Ruzsinszky A and Perdew J P 2015Phys. Rev. Lett.115036402 URL https://doi.org/10.1103/PhysRevLett.115.036402 13 IOP PublishingJournalvv(yyyy) aaaaaa Authoret al

    work page doi:10.1103/physrevlett.115.036402

  39. [39]

    Ning J, Kothakonda M, Furness J W, Kaplan A D, Ehlert S, Brandenburg J G, Perdew J P and Sun J 2022Phys. Rev. B106(7) 075422 URL https://link.aps.org/doi/10.1103/PhysRevB.106.075422

    work page doi:10.1103/physrevb.106.075422

  40. [40]

    Tan J C and Cheetham A K 2011Chem. Soc. Rev.401059–1080 URL https://doi.org/10.1039/C0CS00163E

    work page doi:10.1039/c0cs00163e

  41. [41]

    Sci.1010666–10679 URL https://doi.org/10.1039/C9SC04249K

    Redfern L R and Farha O K 2019Chem. Sci.1010666–10679 URL https://doi.org/10.1039/C9SC04249K

    work page doi:10.1039/c9sc04249k

  42. [42]

    Krause S, Hosono N and Kitagawa S 2020Angew. Chem. Int. Ed.5915325–15341 URL https://doi.org/10.1002/anie.202004535

    work page doi:10.1002/anie.202004535

  43. [43]

    Togo A, Shinohara K and Tanaka I 2024Sci. Technol. Adv. Mater. Meth.42384822–2384836 URLhttps://doi.org/10.1080/27660400.2024.2384822

    work page doi:10.1080/27660400.2024.2384822 2024

  44. [44]

    Skelton J M, Burton L A, Parker S C, Walsh A, Kim C E, Soon A, Buckeridge J, Sokol A A, Catlow C R A, Togo A and Tanaka I 2016Phys. Rev. Lett.117075502 URL https://doi.org/10.1103/PhysRevLett.117.075502

    work page doi:10.1103/physrevlett.117.075502

  45. [45]

    Whalley L D, Skelton J M, Frost J M and Walsh A 2016Phys. Rev. B94220301 URL https://doi.org/10.1103/PhysRevB.94.220301

    work page doi:10.1103/physrevb.94.220301

  46. [46]

    Theory Simul.21800184 URL https://doi.org/10.1002/adts.201800184

    Eriksson F, Fransson E and Erhart P 2019Adv. Theory Simul.21800184 URL https://doi.org/10.1002/adts.201800184

    work page doi:10.1002/adts.201800184

  47. [47]

    Mater.11125 URLhttps://doi.org/10.1038/s41524-025-01611-8 14

    Elena A M, Kamath P D, Jaffrelot Inizan T, Rosen A S, Zanca Fet al.2025 npj Comp. Mater.11125 URLhttps://doi.org/10.1038/s41524-025-01611-8 14

    work page doi:10.1038/s41524-025-01611-8 2025