pith. sign in

arxiv: 2605.18057 · v1 · pith:FKFJCDD3new · submitted 2026-05-18 · ❄️ cond-mat.mtrl-sci · cond-mat.mes-hall

Chemo-mechanical coupling stabilizes mixed Ag_(x)Cu_(1-x)GaSe₂ solar-cell absorbers: Insights from Monte-Carlo simulations assisted by ab initio informed machine-learning potentials

Vasilios Karanikolas , Delwin Perera , Linus Erhard , Jochen Rohrer , Karsten Albe This is my paper

Pith reviewed 2026-05-20 09:42 UTC · model grok-4.3

classification ❄️ cond-mat.mtrl-sci cond-mat.mes-hall
keywords Ag-Cu miscibilitycoherency strainmachine learning potentialMonte Carlo simulationsolar cell absorbersCuGaSe2chemo-mechanical couplingphase stability
0
0 comments X

The pith

Incorporating coherency strains leads to complete Ag-Cu miscibility in mixed gallium selenide solar absorbers.

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

The paper investigates the stability of Ag-alloyed CuGaSe2 for solar cells, where ab initio calculations predict a miscibility gap but experiments show stable high-efficiency devices. Using off-lattice Monte Carlo simulations with a machine-learning interatomic potential, the authors find that coherency strains from a coherent interface eliminate the phase separation tendency, resulting in full miscibility. Without coherency, as in an incoherent setup with misfit dislocations, the expected separation occurs. This chemo-mechanical coupling explains the observed stability.

Core claim

Off-lattice Monte-Carlo simulations assisted by ab initio informed machine-learning potentials demonstrate that elastic energy contributions in a coherent setup result in complete Ag-Cu miscibility in Ag_x Cu_{1-x} GaSe_2, whereas phase separation occurs in the absence of coherency strains with respect to the end boundary phases mimicked by an incoherent interface.

What carries the argument

Machine-learning interatomic potential trained on ab initio data, used in off-lattice Monte-Carlo simulations to compute the free energy including both chemical and elastic contributions under coherent and incoherent boundary conditions.

If this is right

  • Complete miscibility implies that Ag-alloyed absorbers can maintain uniform composition without decomposing at operating temperatures.
  • The approach resolves the controversy between theory and experiment regarding long-term stability of record-efficiency solar cells.
  • The ML-MC framework offers a general method to study thermodynamic stability in systems where strain and chemistry compete.
  • Similar stabilization effects may be expected in other alloyed chalcopyrite materials for photovoltaics.

Where Pith is reading between the lines

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

  • Interface design to maintain coherency could be a strategy to enhance stability in thin-film solar cells.
  • Applying this method to other compositions or temperatures could predict optimal alloying ranges.
  • Validation would require comparing phase behavior in epitaxial coherent films versus polycrystalline or relaxed samples.

Load-bearing premise

The machine-learning potential accurately reproduces the elastic and chemical energies that determine phase stability when coherency strain is imposed.

What would settle it

An experiment demonstrating phase separation in a coherently strained thin film of Ag_x Cu_{1-x} GaSe_2 at ambient temperature would falsify the stabilization claim.

Figures

Figures reproduced from arXiv: 2605.18057 by Delwin Perera, Jochen Rohrer, Karsten Albe, Linus Erhard, Vasilios Karanikolas.

Figure 1
Figure 1. Figure 1: Upper panel: Ag0.5Cu0.5)GaSe2 structure with Ag-Cu atoms homo￾geneously distributed in sublattice-(I). Lower panel: AgGaSe2 and CuGaSe2 structures with a− and c− lattice parameters, which are investigated in Sec. III. the validation of structural and elastic properties (III A), the construction of the composition–temperature phase diagram (III B), and relax–tfMC/MC simulations of coherent and inco￾herent l… view at source ↗
Figure 2
Figure 2. Figure 2: We present the workflow used to develop the machine-learning (ML) potential. Molecular-dynamics (MD) simulations are performed following a cook [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: (a) shows that the static a-lattice parameter increases with increasing Ag content and follows the data obtained from DFT calculations. There is also good agreement with the ex￾perimental values from Ref. [24], which are, however, taken at room temperature. Also, the non-linear dependence of the c-lattice parameter on the Ag content is captured by the ML potential and DFT training data, even though the qua… view at source ↗
Figure 4
Figure 4. Figure 4: (a) The components of the stiffness tensor Ci j and the bulk modulus B are calculated using the ML potential, for varying Ag to Cu ratios. (b) The elas￾tic constants Ci j are used to calculate the effective moduli for a polycrystalline sample using the Voigt, Reuss and Hill measures. All calculation are done with 4 × 4 × 2 supercells and fully relaxed structures with different Ag to Cu ratios. we observe i… view at source ↗
Figure 5
Figure 5. Figure 5: (a) Mixing enthalpy per formula unit for varying Ag to Cu ratios; the circular points give the DFT values and the star-shaped points give the predicted [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Equilibration of a coherent and incoherent layered interfaces using the relax-tfMC [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Panels (a-c) use as a reference the (Ag0.5Cu0.5GaSe2) coherent lattice parameters, whereas panels (d-f) use the end-point references corresponding to AgGaSe2 and CuGaSe2. Panels (a,d) show the induced coherency strain, panels (b,e) show ∆Hmix, Hel, and Htotal, and panels (c,f) show the Gibbs free energy of mixing at different temperatures. All quantities are reported as a function of composition (Ag to Cu … view at source ↗
read the original abstract

Alloying Ag into Cu(In,Ga)Se$_2$ has enabled record solar-cell efficiencies ($\sim23.6\%$), yet their long-term stability remains in question because initio calculations predict a Ag-Cu miscibility gap near ambient temperature. By off-lattice Monte-Carlo simulations using a newly developed machine learning (ML) interatomic potential we show that the presence of coherency strain is resolving the controversy between experimental observations and the predicted phase stability. Incorporating elastic energy contributions present in a coherent setup results in complete Ag-Cu miscibility, whereas the expected phase separation occurs in the absence of coherency strains with respect to the end boundary phases, which are mimicked by an incoherent interface with misfit dislocations. The developed ML-MC framework provides a novel approach for resolving discrepancies in thermodynamic stability for systems where mechanical and chemical effects compete.

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 / 2 minor

Summary. The manuscript uses off-lattice Monte Carlo simulations driven by a machine-learning interatomic potential trained on ab initio data to examine Ag-Cu mixing in Ag_x Cu_{1-x} GaSe_2. It reports that imposing coherent boundary conditions (with coherency strains) produces complete miscibility, while incoherent boundaries that permit misfit dislocations recover the expected phase separation, thereby resolving the discrepancy between ab initio miscibility-gap predictions and experimental stability observations in solar-cell absorbers.

Significance. If the central result is robust, the work is significant because it isolates the role of chemo-mechanical coupling in suppressing phase separation in a technologically relevant photovoltaic alloy. The ML-MC framework offers a scalable route to treat competing chemical and elastic energies in large cells, which could be applied to other alloy systems where direct ab initio sampling is prohibitive.

major comments (2)
  1. [Methods] Methods (ML potential training and validation): the manuscript does not provide quantitative benchmarks showing that the surrogate potential reproduces the long-range elastic strain fields that penalize phase separation under coherent boundary conditions. Because ab initio training sets are typically limited to small supercells, it remains possible that the potential underestimates coherency-strain energy relative to the chemical mixing term; this directly affects the load-bearing claim that coherency produces complete miscibility.
  2. [Results] Results (coherent vs. incoherent comparison): the quantitative partitioning of total energy into chemical and elastic contributions is not shown for the two boundary-condition cases. Without this decomposition (e.g., via separate runs with and without the elastic term), it is difficult to confirm that the observed stabilization arises from coherency strain rather than from an imbalance in the fitted potential.
minor comments (2)
  1. [Abstract] The abstract states that the ML potential is 'newly developed' but the main text should include a short comparison of its accuracy on formation energies and elastic constants against existing empirical or other ML potentials for chalcopyrites.
  2. [Figures] Figure captions for the Monte Carlo snapshots should explicitly label the coherent and incoherent boundary conditions and note the system size used in each panel.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments and for recognizing the potential significance of the chemo-mechanical coupling mechanism in resolving the stability discrepancy for Ag-Cu alloying in solar-cell absorbers. We address each major comment in detail below.

read point-by-point responses

  1. Referee: [Methods] Methods (ML potential training and validation): the manuscript does not provide quantitative benchmarks showing that the surrogate potential reproduces the long-range elastic strain fields that penalize phase separation under coherent boundary conditions. Because ab initio training sets are typically limited to small supercells, it remains possible that the potential underestimates coherency-strain energy relative to the chemical mixing term; this directly affects the load-bearing claim that coherency produces complete miscibility.

    Authors: We agree that explicit validation of the ML potential's reproduction of long-range elastic strain fields is essential to support the central claim. The training set incorporated strained and defect-containing configurations from ab initio calculations to capture elastic contributions, but we acknowledge that direct quantitative benchmarks against ab initio strain energies in larger coherent cells were not presented. In the revised manuscript we will add such benchmarks, including direct comparisons of coherency-strain energies for supercells up to the largest sizes feasible with ab initio methods, to demonstrate that the potential does not systematically underestimate the elastic penalty relative to the chemical term. revision: yes

  2. Referee: [Results] Results (coherent vs. incoherent comparison): the quantitative partitioning of total energy into chemical and elastic contributions is not shown for the two boundary-condition cases. Without this decomposition (e.g., via separate runs with and without the elastic term), it is difficult to confirm that the observed stabilization arises from coherency strain rather than from an imbalance in the fitted potential.

    Authors: We concur that an explicit energy decomposition would strengthen the attribution of stabilization to coherency strain. In the revised manuscript we will include a quantitative partitioning of the total energy into chemical (local bonding) and elastic (long-range strain) components for both the coherent and incoherent boundary-condition simulations. This will be obtained by performing auxiliary runs in which atomic relaxations are constrained to isolate the elastic contribution, allowing direct comparison that confirms the miscibility arises from the coherency-strain term rather than any fitting artifact. revision: yes

Circularity Check

0 steps flagged

No significant circularity in the derivation chain

full rationale

The paper's central result—that coherency strains stabilize complete Ag-Cu miscibility—is obtained from explicit off-lattice Monte Carlo sampling in large cells with boundary conditions that enforce coherency versus an incoherent interface. The ML interatomic potential is trained on independent ab initio calculations, and the elastic versus chemical energy balance is evaluated numerically rather than imposed by definition or by a fitted parameter that is then relabeled as a prediction. No step in the reported workflow reduces the final miscibility outcome to an internal fit or to a self-citation whose content is itself unverified.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the transferability of the ML potential from ab-initio training data to finite-temperature Monte-Carlo runs and on the modeling choice that coherent interfaces carry only elastic strain while incoherent ones allow misfit dislocations. No new particles or forces are postulated.

axioms (1)
  • domain assumption The machine-learning potential trained on ab-initio calculations faithfully reproduces both chemical mixing energies and elastic moduli at the relevant compositions and temperatures.
    Invoked when the authors state that the ML potential enables off-lattice Monte-Carlo that captures chemo-mechanical coupling.

pith-pipeline@v0.9.0 · 5719 in / 1305 out tokens · 32557 ms · 2026-05-20T09:42:19.573372+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

70 extracted references · 70 canonical work pages

  1. [1]

    Oliveira, J.P

    A.J.N. Oliveira, J.P. Teixeira, D. Ramos, P.A. Fernan- des, P.M.P. Salomé, Exploiting the Optical Limits of Thin-Film Solar Cells: A Review on Light Management Strategies in Cu(In,Ga)Se2, Adv. Photonics Res. 3 (2022) 2100190. https://doi.org/10.1002/adpr.202100190

    work page doi:10.1002/adpr.202100190 2022

  2. [2]

    Mufti, T

    N. Mufti, T. Amrillah, A. Taufiq, Sunaryono, Aripri- harta, M. Diantoro, Zulhadjri, H. Nur, Review of CIGS-based solar cells manufacturing by structural engineering, Sol. Energy 207 (2020) 1146–1157. https://doi.org/10.1016/j.solener.2020.07.065

    work page doi:10.1016/j.solener.2020.07.065 2020

  3. [4]

    Keller, K

    J. Keller, K. Kiselman, O. Donzel-Gargand, N.M. Martin, M. Babucci, O. Lundberg, E. Wallin, L. Stolt, M. Edoff, High-concentration silver alloying and steep back-contact gallium grading enabling copper indium gallium selenide solar cell with 23.6% efficiency, Nat. Energy 9 (2024) 467–478. https://doi.org/10.1038/s41560-024-01472-3

    work page doi:10.1038/s41560-024-01472-3 2024

  4. [5]

    Erslev, J

    P.T. Erslev, J. Lee, G.M. Hanket, W.N. Shafarman, J.D. Cohen, The electronic structure of Cu(In 1−xGax)Se2 al- loyed with silver, Thin Solid Films 519 (2011) 7296–

    work page 2011

  5. [6]

    https://doi.org/10.1016/j.tsf.2011.01.368

    work page doi:10.1016/j.tsf.2011.01.368 2011

  6. [7]

    S.-C. Yang, J. Sastre, M. Krause, X. Sun, R. Hertwig, M. Ochoa, A.N. Tiwari, R. Carron, Silver-Promoted High- Performance (Ag,Cu)(In,Ga)Se 2 Thin-Film Solar Cells Grown at Very Low Temperature, Sol. RRL 5 (2021) 2100108. https://doi.org/10.1002/solr.202100108

    work page doi:10.1002/solr.202100108 2021

  7. [8]

    Abou-Ras, N

    D. Abou-Ras, N. Schafer, T. Rissom, M.N. Kelly, J. Haarstrich, C. Ronning, G.S. Rohrer, A.D. Rol- lett, Grain-boundary character distribution and corre- lations with electrical and optoelectronic properties of CuInSe2 thin films, Acta Mater. 118 (2016) 244–252. https://doi.org/10.1016/j.actamat.2016.07.042

    work page doi:10.1016/j.actamat.2016.07.042 2016

  8. [9]

    Valdes, J.W

    N. Valdes, J.W. Lee, W. Shafarman, Comparison of Ag and Ga alloying in low bandgap CuInSe 2-based solar cells, Sol. Energy Mater. Sol. Cells 195 (2019) 155–159. https://doi.org/10.1016/j.solmat.2019.02.022. 11

    work page doi:10.1016/j.solmat.2019.02.022 2019

  9. [10]

    Kim, J.W

    K. Kim, J.W. Park, J.S. Yoo, J. Sik Cho, H.-D. Lee, J.H. Yun, Ag incorporation in low-temperature grown Cu(In,Ga)Se2 solar cells using Ag precursor layers, Sol. Energy Mater. Sol. Cells 146 (2016) 114–120. https://doi.org/10.1016/j.solmat.2015.11.028

    work page doi:10.1016/j.solmat.2015.11.028 2016

  10. [11]

    Y . Gu, C. Zhou, W. Chen, Y . Zhang, Y . Yao, Z. Zhou, Y . Sun, W. Liu, Silver (Ag) substitution in Cu(In,Ga)Se2 solar cell: insights into current challenges and future prospects, Appl. Phys. A 130 (2024) 636. https://doi.org/10.1007/s00339-024-07796-x

    work page doi:10.1007/s00339-024-07796-x 2024

  11. [12]

    Kim, S.K

    K. Kim, S.K. Ahn, J.H. Choi, J. Yoo, Y .J. Eo, J.S. Cho, A. Cho, J. Gwak, S. Song, D.H. Cho, Y .D. Chung, J.H. Yun, Highly Efficient Ag-Alloyed Cu(In,Ga)Se2 Solar Cells with Wide Bandgaps and Their Application to Chalcopyrite-Based Tan- dem Solar Cells, Nano Energy 48 (2018) 345–352. https://doi.org/10.1016/j.nanoen.2018.03.052

    work page doi:10.1016/j.nanoen.2018.03.052 2018

  12. [13]

    Zhang, Z

    Y . Zhang, Z. Hu, S. Lin, S. Cheng, Z. He, C. Wang, Z. Zhou, Y . Sun, W. Liu, Facile Silver-Incorporated Method of Tuning the Back Gradient of Cu(In,Ga)Se 2 Films, ACS Appl. Energy Mater. 3 (2020) 9963–9971. https://doi.org/10.1021/acsaem.0c01644

    work page doi:10.1021/acsaem.0c01644 2020

  13. [14]

    Krause, S.C

    M. Krause, S.C. Yang, S. Moser, S. Nishiwaki, A.N. Ti- wari, R. Carron, Silver-Alloyed Low-Bandgap CuInSe 2 Solar Cells for Tandem Applications, Sol. RRL 7 (2023) 2201122. https://doi.org/10.1002/solr.202201122

    work page doi:10.1002/solr.202201122 2023

  14. [15]

    Siebentritt, T.P

    S. Siebentritt, T.P. Weiss, Chalcopyrite solar cells— state-of-the-art and options for improvement, Sci. China Phys. Mech. Astron. 66 (2023) 217301. https://doi.org/10.1007/s11433-022-2001-4

    work page doi:10.1007/s11433-022-2001-4 2023

  15. [16]

    Chen, X.G

    S. Chen, X.G. Gong, S.-H. Wei, Band-structure anomalies of the chalcopyrite semiconductors CuGaX2 versus AgGaX 2 (X=S and Se) and their alloys, Phys. Rev. B 75 (2007) 205209. https://doi.org/10.1103/PhysRevB.75.205209

    work page doi:10.1103/physrevb.75.205209 2007

  16. [17]

    Sopiha, J.K

    K.V . Sopiha, J.K. Larsen, O. Donzel-Gargand, F. Khavari, J. Keller, M. Edoff, C. Platzer-Björkman, C. Pers- son, J.J.S. Scragg, Thermodynamic stability, phase sep- aration and Ag grading in (Ag,Cu)(In,Ga)Se 2 solar absorbers, J. Mater. Chem. A 8 (2020) 8740–8751. https://doi.org/10.1039/D0TA00363H

    work page doi:10.1039/d0ta00363h 2020

  17. [18]

    Robbins, J.C

    M. Robbins, J.C. Phillips, V .G. Lambrecht Jr., Solid solu- tion formation in the systems CuMIIIX2–AgMIIIX2 where MIII =Al, Ga, In and X=S, Se, J. Phys. Chem. Solids 34 (1973) 1205–1209. https://doi.org/10.1016/S0022- 3697(73)80210-0

    work page doi:10.1016/s0022- 1973

  18. [19]

    J.E. Avon, K. Yoodee, J.C. Woolley, Solid solution, lattice parameter values, and effects of composition in AgxCu1−xInSe2 chalcopyrite, J. Appl. Phys. 55 (1984) 524–535. https://doi.org/10.1063/1.333058

    work page doi:10.1063/1.333058 1984

  19. [20]

    Albornoz, R.M

    J.G. Albornoz, R.M. Rojas L., J.M. Merino, M. Leon, Structural, thermal and electrical properties of the semi- conductor system Ag 1−xCuxInSe2, J. Phys. Chem. Solids 75 (2014) 1–7. https://doi.org/10.1016/j.jpcs.2013.08.003

    work page doi:10.1016/j.jpcs.2013.08.003 2014

  20. [21]

    Hanket, J.H

    G.M. Hanket, J.H. Boyle, W.N. Shafarman, Characterization and Device Performance of (AgCu)(InGa)Se2 Absorber Layers, in: Proc. 34th IEEE Photovoltaic Specialists Conf. (PVSC), Philadelphia, PA, USA, 2009, pp. 001240–001245. https://doi.org/10.1109/PVSC.2009.5411241

    work page doi:10.1109/pvsc.2009.5411241 2009

  21. [22]

    Hanket, J.H

    G.M. Hanket, J.H. Boyle, W.N. Shafarman, G. Teeter, Wide-bandgap (AgCu)(InGa)Se 2 absorber layers de- posited by three-stage co-evaporation, in: Proc. 35th IEEE Photovoltaic Specialists Conf. (PVSC), Honolulu, HI, USA, 2010, pp. 003425–003429. https://doi.org/10.1109/PVSC.2010.5614576

    work page doi:10.1109/pvsc.2010.5614576 2010

  22. [23]

    Boyle, B.E

    J.H. Boyle, B.E. McCandless, G.M. Hanket, W.N. Shafarman, Structural characterization of the (AgCu)(InGa)Se2 thin film alloy system by x-ray diffraction, Thin Solid Films 519 (2011) 7292–7295. https://doi.org/10.1016/j.tsf.2011.01.138

    work page doi:10.1016/j.tsf.2011.01.138 2011

  23. [24]

    Chen, J.W

    L. Chen, J.W. Lee, W.N. Shafarman, The Com- parison of (Ag,Cu)(In,Ga)Se 2 and Cu(In,Ga)Se 2 Thin Films Deposited by Three-Stage Coevap- oration, IEEE J. Photovolt. 4 (2014) 447–451. https://doi.org/10.1109/JPHOTOV .2013.2280471

    work page doi:10.1109/jphotov 2014

  24. [25]

    Boyle, B.E

    J.H. Boyle, B.E. McCandless, W.N. Shafarman, R.W. Birkmire, Structural and optical proper- ties of (Ag,Cu)(In,Ga)Se 2 polycrystalline thin film alloys, J. Appl. Phys. 115 (2014) 223504. https://doi.org/10.1063/1.4880243

    work page doi:10.1063/1.4880243 2014

  25. [26]

    Erhard, J

    L.C. Erhard, J. Rohrer, K. Albe, V .L. Deringer, A machine-learned interatomic potential for silica and its re- lation to empirical models, npj Comput. Mater. 8 (2022)

    work page 2022

  26. [27]

    https://doi.org/10.1038/s41524-022-00768-w

    work page doi:10.1038/s41524-022-00768-w

  27. [28]

    Mortazavi, X

    B. Mortazavi, X. Zhuang, T. Rabczuk, A.V . Shapeev, Atomistic modeling of the mechanical proper- ties: the rise of machine learning interatomic potentials, Mater. Horiz. 10 (2023) 1956–1968. https://doi.org/10.1039/D3MH00125C

    work page doi:10.1039/d3mh00125c 2023

  28. [29]

    Erhard, J

    L.C. Erhard, J. Rohrer, K. Albe, V .L. Deringer, Modelling atomic and nanoscale structure in the silicon–oxygen sys- tem through active machine learning, Nat. Commun. 15 (2024) 1927. https://doi.org/10.1038/s41467-024-45840- 9

    work page doi:10.1038/s41467-024-45840- 2024

  29. [30]

    Jacobs, D

    R. Jacobs, D. Morgan, S. Attarian, J. Meng, C. Shen, Z. Wu, C.Y . Xie, J.H. Yang, N. Artrith, B. Blaiszik, G. Ceder, K. Choudhary, G. Csanyi, E.D. Cubuk, B. Deng, R. Drautz, X. Fu, J. Godwin, V . Honavar, O. Isayev, A. Johansson, B. Kozinsky, S. Martiniani, S.P. Ong, I. Poltavsky, K. Schmidt, S. Takamoto, A.P. Thompson, 12 J. Westermayr, B.M. Wood, A prac...

    work page doi:10.1016/j.cossms.2025.101214 2025

  30. [31]

    Baskaran, C

    A. Baskaran, C. Ratsch, P. Smereka, Modeling the elas- tic energy of alloys: Potential pitfalls of Fourier-based multilattice methods, Phys. Rev. E 92 (2015) 062406. https://doi.org/10.1103/PhysRevE.92.062406

    work page doi:10.1103/physreve.92.062406 2015

  31. [32]

    Dereli, Stillinger-Weber type potentials in Monte Carlo simulations of amorphous silicon, Mol

    G. Dereli, Stillinger-Weber type potentials in Monte Carlo simulations of amorphous silicon, Mol. Simul. 8 (1992) 351-360. https://doi.org/10.1080/08927029208022490

    work page doi:10.1080/08927029208022490 1992

  32. [33]

    Timonova, J

    M. Timonova, J. Groenewegen, B.J. Thijsse, Modeling diffusion and phase transitions by a uniform-acceptance force-bias Monte Carlo method, Phys. Rev. B 81 (2010) 144107. https://doi.org/10.1103/PhysRevB.81.144107

    work page doi:10.1103/physrevb.81.144107 2010

  33. [34]

    M.J. Mees, G. Pourtois, E.C. Neyts, B.J. Thi- jsse, A. Stesmans, Uniform-acceptance force-bias Monte Carlo method with time scale to study solid- state diffusion, Phys. Rev. B 85 (2012) 134301. https://doi.org/10.1103/PhysRevB.85.134301

    work page doi:10.1103/physrevb.85.134301 2012

  34. [35]

    Bal, E.C

    K.M. Bal, E.C. Neyts, On the time scale associated with Monte Carlo simulations, J. Chem. Phys. 141 (2014) 204104. https://doi.org/10.1063/1.4902136

    work page doi:10.1063/1.4902136 2014

  35. [36]

    Eshelby, The determination of the elastic field of an ellipsoidal inclusion, and related prob- lems, Proc

    J.D. Eshelby, The determination of the elastic field of an ellipsoidal inclusion, and related prob- lems, Proc. R. Soc. Lond. A 241 (1957) 376–396. https://doi.org/10.1098/rspa.1957.0133

    work page doi:10.1098/rspa.1957.0133 1957

  36. [37]

    and Spaepen, F

    F. Larche, J.W. Cahn, A linear theory of thermo- chemical equilibrium of solids under stress, Acta Met- all. 21 (1973) 1051–1063. https://doi.org/10.1016/0001- 6160(73)90021-7

    work page doi:10.1016/0001- 1973

  37. [38]

    Larche, J.W

    F.C. Larche, J.W. Cahn, The effect of self-stress on dif- fusion in solids, Acta Metall. 30 (1982) 1835–1845. https://doi.org/10.1016/0001-6160(82)90023-2

    work page doi:10.1016/0001-6160(82)90023-2 1982

  38. [39]

    Larche, J.W

    F.C. Larche, J.W. Cahn, The Interactions of Composition and Stress in Crystalline Solids, J. Res. Natl. Bur. Stand. 89 (1984) 467–500. https://doi.org/10.6028/jres.089.026

    work page doi:10.6028/jres.089.026 1984

  39. [40]

    Fratzl, O

    P. Fratzl, O. Penrose, J.L. Lebowitz, Modeling of phase separation in alloys with coherent elas- tic misfit, J. Stat. Phys. 95 (1999) 1429–1503. https://doi.org/10.1023/A:1004587425006

    work page doi:10.1023/a:1004587425006 1999

  40. [41]

    Kamachali, L

    R.D. Kamachali, L. Wang, Elastic energy of multi-component solid solutions and strain origins of phase stability in high-entropy alloys, Scr. Mater. 206 (2022) 114226. https://doi.org/10.1016/j.scriptamat.2021.114226

    work page doi:10.1016/j.scriptamat.2021.114226 2022

  41. [42]

    Weissmuller, S

    J. Weissmuller, S. Shi, Giant compliance and spontaneous buckling of beams containing mo- bile solute atoms, Acta Mater. 227 (2022) 117696. https://doi.org/10.1016/j.actamat.2022.117696

    work page doi:10.1016/j.actamat.2022.117696 2022

  42. [43]

    Robbins, V .G

    M. Robbins, V .G. Lambrecht Jr., Solid solution forma- tion in chalcopyrite systems of the type AgInX2–AgGaX2, X=S, Se, Te, J. Solid State Chem. 15 (1975) 167–170. https://doi.org/10.1016/0022-4596(75)90240-6

    work page doi:10.1016/0022-4596(75)90240-6 1975

  43. [44]

    Drautz, Atomic cluster expansion for accurate and transferable interatomic po- tentials, Phys

    R. Drautz, Atomic cluster expansion for accurate and transferable interatomic po- tentials, Phys. Rev. B 99 (2019) 014104. https://doi.org/10.1103/PhysRevB.99.014104

    work page doi:10.1103/physrevb.99.014104 2019

  44. [45]

    Lysogorskiy, C

    Y . Lysogorskiy, C. van der Oord, A. Bochkarev, S. Menon, M. Rinaldi, T. Hammerschmidt, M. Mrovec, A.P. Thomp- son, G. Csanyi, C. Ortner, R. Drautz, Performant imple- mentation of the atomic cluster expansion (PACE) and application to copper and silicon, npj Comput. Mater. 7 (2021) 97. https://doi.org/10.1038/s41524-021-00559-9

    work page doi:10.1038/s41524-021-00559-9 2021

  45. [46]

    Leimeroth, L.C

    N. Leimeroth, L.C. Erhard, K. Albe, J. Rohrer, Machine- learning interatomic potentials from a users perspec- tive: a comparison of accuracy, speed and data effi- ciency, Modell. Simul. Mater. Sci. Eng. 33 (2025) 065012. https://doi.org/10.1088/1361-651X/adf56d

    work page doi:10.1088/1361-651x/adf56d 2025

  46. [47]

    Qamar, M

    M. Qamar, M. Mrovec, Y . Lysogorskiy, A. Bochkarev, R. Drautz, Atomic cluster expansion for quantum- accurate large-scale simulations of carbon, J. Chem. Theory Comput. 19 (2023) 5151–5167. https://doi.org/10.1021/acs.jctc.2c01149

    work page doi:10.1021/acs.jctc.2c01149 2023

  47. [48]

    Podryabinkin, A.V

    E.V . Podryabinkin, A.V . Shapeev, Active learn- ing of linearly parametrized interatomic poten- tials, Comput. Mater. Sci. 140 (2017) 171–180. https://doi.org/10.1016/j.commatsci.2017.08.031

    work page doi:10.1016/j.commatsci.2017.08.031 2017

  48. [49]

    Novikov, K

    I.S. Novikov, K. Gubaev, E.V . Podryabinkin, A.V . Shapeev, The MLIP package: moment tensor poten- tials with MPI and active learning, Mach. Learn.: Sci. Technol. 2 (2021) 025002. https://doi.org/10.1088/2632- 2153/abc9fe

    work page doi:10.1088/2632- 2021

  49. [50]

    Lysogorskiy, A

    Y . Lysogorskiy, A. Bochkarev, M. Mrovec, R. Drautz, Active learning strategies for atomic cluster expan- sion models, Phys. Rev. Mater. 7 (2023) 043801. https://doi.org/10.1103/PhysRevMaterials.7.043801

    work page doi:10.1103/physrevmaterials.7.043801 2023

  50. [51]

    Bochkarev, Y

    A. Bochkarev, Y . Lysogorskiy, S. Menon, M. Qamar, M. Mrovec, R. Drautz, Efficient parametrization of the atomic cluster expansion, Phys. Rev. Mater. 6 (2022) 013804. https://doi.org/10.1103/PhysRevMaterials.6.013804

    work page doi:10.1103/physrevmaterials.6.013804 2022

  51. [52]

    Kresse, J

    G. Kresse, J. Furthmüller, J. Hafner, Theory of the crystal structures of selenium and tel- lurium, Phys. Rev. B 50 (1994) 13181–13185. https://doi.org/10.1103/PhysRevB.50.13181

    work page doi:10.1103/physrevb.50.13181 1994

  52. [53]

    Kresse, J

    G. Kresse, J. Furthmüller, Efficiency of ab-initio total en- ergy calculations for metals and semiconductors using a plane-wave basis set, Comput. Mater. Sci. 6 (1996) 15–

    work page 1996

  53. [54]

    https://doi.org/10.1016/0927-0256(96)00008-0. 13

    work page doi:10.1016/0927-0256(96)00008-0

  54. [55]

    Perdew, K

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

    work page doi:10.1103/physrevlett.77.3865 1996

  55. [56]

    (36) Blöchl, P

    G. Kresse, D. Joubert, From ultrasoft pseu- dopotentials to the projector augmented-wave method, Phys. Rev. B 59 (1999) 1758–1775. https://doi.org/10.1103/PhysRevB.59.1758

    work page doi:10.1103/physrevb.59.1758 1999

  56. [57]

    Thompson, H.M

    A.P. Thompson, H.M. Aktulga, R. Berger, D.S. Bolin- tineanu, 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, S.J. Plimpton, LAMMPS – a flexible simulation tool for particle-based materials modeling at the atomic, meso, and contin- uum scales, Comp. Phys. Comm. 271 (2022) 1081...

    work page doi:10.1016/j.cpc.2021.108171 2022

  57. [58]

    Sadigh, P

    B. Sadigh, P. Erhart, A. Stukowski, A. Caro, E. Martinez, L. Zepeda-Ruiz, Scalable parallel Monte Carlo algorithm for atomistic simulations of precip- itation in alloys, Phys. Rev. B 85 (2012) 184203. https://doi.org/10.1103/PhysRevB.85.184203

    work page doi:10.1103/physrevb.85.184203 2012

  58. [59]

    Neyts, A

    E.C. Neyts, A. Bogaerts, Combining molecular dy- namics with Monte Carlo simulations: implementations and applications, Theor. Chem. Acc. 132 (2013) 1320. https://doi.org/10.1007/s00214-012-1320-x

    work page doi:10.1007/s00214-012-1320-x 2013

  59. [60]

    https://doi.org/xxx

    In the Supplemental Materials we provide validation plots for the ML potential, energy-volume curves for different AgxCu1−xGaSe2 phases, the temperature dependence of the lattice parameters, the anion displacement, present the relax-tfMC/MD protocol and its application to a spherical inclusion structure. https://doi.org/xxx

    work page

  60. [61]

    Grimsditch, G.D

    M.H. Grimsditch, G.D. Holah, Brillouin scat- tering and elastic moduli of silver thiogal- late (AgGaS 2), Phys. Rev. B 12 (1975) 4377. https://doi.org/10.1103/PhysRevB.12.4377

    work page doi:10.1103/physrevb.12.4377 1975

  61. [62]

    Neumann, Bulk modulus-volume rela- tionship in ternary chalcopyrite compounds, Phys

    H. Neumann, Bulk modulus-volume rela- tionship in ternary chalcopyrite compounds, Phys. Status Solidi A 96 (1986) K121–K125. https://doi.org/10.1002/pssa.2210960245

    work page doi:10.1002/pssa.2210960245 1986

  62. [63]

    Neumann, Simple theoretical estimate of surface en- ergy, bulk modulus, and atomization energy of A IBIIICVI 2 compounds, Cryst

    H. Neumann, Simple theoretical estimate of surface en- ergy, bulk modulus, and atomization energy of A IBIIICVI 2 compounds, Cryst. Res. Technol. 18 (1983) 665–670. https://doi.org/10.1002/crat.2170180517

    work page doi:10.1002/crat.2170180517 1983

  63. [64]

    Fernandez, S.M

    B. Fernandez, S.M. Wasim, Sound Veloc- ities and Elastic Moduli in CuInTe 2 and CuInSe2, Phys. Status Solidi A 122 (1990) 235. https://doi.org/10.1002/pssa.2211220122

    work page doi:10.1002/pssa.2211220122 1990

  64. [65]

    Garbato, F

    L. Garbato, F. Ledda, R. Rucci, Structural distor- tions and polymorphic behaviour in ABC 2 chalcopy- rites, Prog. Cryst. Growth Charact. 15 (1987) 1–41. https://doi.org/10.1016/0146-3535(87)90008-6

    work page doi:10.1016/0146-3535(87)90008-6 1987

  65. [66]

    H.H. Falk, S. Eckner, M. Seifert, K. Ritter, S. Lev- cenko, T. Pfeiffelmann, E. Welter, S. Botti, W.N. Shafarman, C.S. Schnohr, Peculiar bond length de- pendence in (Ag,Cu)GaSe 2 alloys and its impact on the bandgap bowing, APL Mater. 11 (2023) 111105. https://doi.org/10.1063/5.0164407

    work page doi:10.1063/5.0164407 2023

  66. [67]

    Shay, J.H

    J.L. Shay, J.H. Wernick, Ternary Chalcopyrite Semicon- ductors: Growth, Electronic Properties and Applications, Pergamon Press, Oxford, 1975

    work page 1975

  67. [68]

    Verma, S

    A.S. Verma, S. Sharma, R. Bhandari, B.K. Sarkar, V .K. Jindal, Elastic properties of chalcopyrite struc- tured solids, Mater. Chem. Phys. 132 (2012) 416–420. https://doi.org/10.1016/j.matchemphys.2011.11.047

    work page doi:10.1016/j.matchemphys.2011.11.047 2012

  68. [69]

    Kushwaha, C.-G

    A.K. Kushwaha, C.-G. Ma, M.G. Brik, S. Bin Omran, R. Khenata, Zone-center phonons and elastic properties of ternary chalcopyrite ABSe 2 (A=Cu and Ag; B= Al, Ga and In), Mater. Chem. Phys. 227 (2019) 324–331. https://doi.org/10.1016/j.matchemphys.2019.02.024

    work page doi:10.1016/j.matchemphys.2019.02.024 2019

  69. [70]

    Stukowski, Visualization and analysis of atomistic simulation data with OVITO—the Open Visualization Tool, Modell

    A. Stukowski, Visualization and analysis of atomistic simulation data with OVITO—the Open Visualization Tool, Modell. Simul. Mater. Sci. Eng. 18 (2010) 015012. https://doi.org/10.1088/0965-0393/18/1/015012

    work page doi:10.1088/0965-0393/18/1/015012 2010

  70. [71]

    Aboulfadl, K.V

    H. Aboulfadl, K.V . Sopiha, J. Keller, J.K. Larsen, J.J.S. Scragg, C. Persson, M. Thuvander, M. Ed- off, Alkali Dispersion in (Ag,Cu)(In,Ga)Se 2 Thin Film Solar Cells—Insight from Theory and Experiment, ACS Appl. Mater. Interfaces 13 (2021) 7188–7199. https://doi.org/10.1021/acsami.0c20539. 14

    work page doi:10.1021/acsami.0c20539 2021