pith. sign in

arxiv: 2605.17207 · v1 · pith:UPJFNEIHnew · submitted 2026-05-17 · ❄️ cond-mat.mtrl-sci · physics.app-ph· physics.chem-ph

Structure of Molten FeCl2 and FeCl3

Fakhrul Hasan Bhuiyan , Jicheng Guo , Christopher James Benmore , Avery Blockmon , Denis Johnson , Alvaro Vazquez Mayagoitia This is my paper

Pith reviewed 2026-05-19 23:48 UTC · model grok-4.3

classification ❄️ cond-mat.mtrl-sci physics.app-phphysics.chem-ph
keywords molten iron chloridesstructure of meltsX-ray diffractionmolecular dynamicspolymerizationmachine learning potentialscoordination transition
0
0 comments X

The pith

Molten FeCl2 and FeCl3 both form extended chains of six or more iron centers linked by chlorine bridges rather than discrete Fe2Cl6 units.

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

The paper establishes the atomic structures of molten FeCl2 and FeCl3 by combining high-energy X-ray diffraction measurements with empirical potential structure refinement and molecular dynamics simulations that use machine-learned interatomic potentials. It identifies a change in iron coordination from octahedral to tetrahedral upon melting and uses polymer chain statistics on the simulation trajectories to show that both liquids consist predominantly of long chains containing six or more Fe centers. This finding revises earlier pictures of these melts as collections of small molecular units. The structural details matter for understanding ion transport and electrochemical performance in molten salts proposed for iron production and redox flow batteries.

Core claim

High-energy X-ray diffraction data for molten FeCl2 and FeCl3 are reproduced by EPSR refinement and by MACE-based MLIP molecular dynamics simulations. Both melts display a transition of Fe coordination from octahedral in the crystal to tetrahedral in the liquid. Analysis of the MD trajectories quantifies coordination numbers, bridging chlorine populations, and connectivity, showing that polymer chain statistics place the majority of iron centers in extended chains of six or more Fe atoms connected through Cl bridges, in contrast to prior reports of dominant Fe2Cl6 dimers.

What carries the argument

Polymer chain statistics applied to the connectivity patterns of Fe centers through bridging Cl atoms extracted from the molecular dynamics trajectories.

If this is right

  • The two melts exhibit distinct degrees of polymerization and local geometries that influence their transport properties.
  • Diffusion coefficients calculated from the MACE-MD trajectories provide direct comparisons between FeCl2 and FeCl3 melts.
  • The demonstrated reliability of MACE-based MLIPs offers a route to predictive modeling of other high-temperature molten salts.
  • The resulting structural benchmarks can guide optimization of electrochemical processes that use these liquids.

Where Pith is reading between the lines

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

  • Viscosity and ionic conductivity in these melts may be controlled by the average chain length rather than by isolated molecular species.
  • Analogous polymerization could appear in other molten transition-metal halides and alter their behavior in high-temperature applications.
  • Additional spectroscopic probes of chain length would offer a direct experimental test of the polymerization picture.

Load-bearing premise

The MACE machine-learning interatomic potentials and EPSR refinement together capture the true high-temperature coordination and chain connectivity without significant bias from model choice or fitting procedure.

What would settle it

An independent measurement, such as neutron diffraction or spectroscopy, that finds a majority of iron centers in short chains or isolated Fe2Cl6 units would contradict the extended-chain dominance.

Figures

Figures reproduced from arXiv: 2605.17207 by Alvaro Vazquez Mayagoitia, Avery Blockmon, Christopher James Benmore, Denis Johnson, Fakhrul Hasan Bhuiyan, Jicheng Guo.

Figure 1
Figure 1. Figure 1: FeCl2 crystal (left panel) and liquid (right panel) systems for MD simulations. Green atoms represent Cl and brown atoms represent Fe. Methods Experiments FeCl2 (Sigma-Alrich, ≥ 99.99 trace metal basis) and FeCl3 (Sigma-Aldrich,≥ 99.99 trace metal basis) were loaded in separated fused silica quartz tubes (VitroCom) in a UHP Ar glovebox (<0.1 ppm O2, < 1ppm H2O), which were then sealed under vacuum after be… view at source ↗
Figure 2
Figure 2. Figure 2: HEXRD-derived structure factors (left panel) and total RDFs (right panel) for molten FeCl2 (750°C) and FeCl3 (310°C). The molten structure of FeCl3 was previously studied using neutron diffraction, where the first and second peak positions were reported at 2.22 Å and approximately 3.7 Å, respectively.2,7 The first peak position observed in our HEXRD measurement (2.19 Å) matches closely with the neutron dif… view at source ↗
Figure 3
Figure 3. Figure 3: Comparison of experimental structure factor (left panel) and RDF (right panel) for FeCl2 with results from MD simulations with MACE foundation model (MACE-r 2SCAN) and finetuned model (ft-MACE-r 2SCAN), and EPSR modelling [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Comparison of experimental structure factor (left panel) and RDF (right panel) for FeCl3 with results from MD simulations with MACE foundation model (MACE-r 2SCAN) and finetuned model (ft-MACE-r 2SCAN), and EPSR modelling. The structure factor and RDF for FeCl3 obtained from MD simulations with the MACE models are shown in [PITH_FULL_IMAGE:figures/full_fig_p010_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Fe-Cl, Cl-Cl and Fe-Fe partial RDFs (left column) and coordination numbers (right column) for molten FeCl2 and FeCl3 from MD simulations with MACE foundation model (MACE￾r 2SCAN) and fine-tuned model (ft-MACE-r 2SCAN), and EPSR modelling [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Representative configurations from the MACE-MD simulations illustrating the octahedral coordination in the crystals and the tetrahedral coordination in the corresponding melts for FeCl2 and FeCl3. Each of the four panels shows the front and the side view of the same group of atoms randomly selected from the corresponding simulation trajectory. The resulting bridging‑Cl fraction, evaluated as a function of … view at source ↗
Figure 7
Figure 7. Figure 7: (a) The measured x-ray structure factor of molten FeCl3 at 310C (black circles) compared to our EPSR Fe2Cl6 molecular model (solid red line) and atomic FeCl3 model (blues dashed line). The insert shows the arrangement of molecules and atoms in the EPSR simulation box for the two models fixed at the experimental density. (b) the corresponding x-ray pair distribution function obtained from a Sine Fourier tra… view at source ↗
Figure 8
Figure 8. Figure 8: Analysis of FexCly polyhedral chain weights from MD simulations of liquid FeCl3 performed using the MACE-r 2SCAN foundation model. Polymerized chains are categorized by the number of Fe ions they contain, and the total mass of each category is normalized by the total simulation box mass to report weight percentages. Fe-Cl cutoff used was 3.0 Å for detecting the polyhedral chains [PITH_FULL_IMAGE:figures/f… view at source ↗
Figure 9
Figure 9. Figure 9: Variation of the fraction of bridging-Cl atoms, i.e., chlorides acting as bridges between iron atoms, as a function of the Fe-Cl cutoff distance in FeCl2 and FeCl3 melts in the MD simulations performed using the MACE-r 2SCAN foundation model and the fine-tuned model (ft￾MACE-r 2SCAN) [PITH_FULL_IMAGE:figures/full_fig_p018_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Distributions of Cl-Fe-Cl (intra-tetrahedral) and Fe-Cl-Fe (bridging) angles from MD simulations employing the MACE-r 2SCAN foundation model for FeCl2 and FeCl3. Fe-Cl bond [PITH_FULL_IMAGE:figures/full_fig_p018_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Arrhenius plot of the calculated diffusion coefficients for molten FeCl2 and FeCl3 from MD simulations employing the MACE-r 2SCAN foundation model. The total, Fe, and Cl diffusion coefficients are plotted on a logarithmic scale against inverse temperature (1000/T). Blue and orange markers denote the FeCl2 and FeCl3 systems, respectively. The straight lines represent linear least-squares fits to the Arrhen… view at source ↗
read the original abstract

Molten iron chlorides are central to emerging energy technologies, including electrochemical iron production and redox flow batteries. Optimizing their electrochemical performance and transport properties requires atomic-scale structural understanding, yet detailed data for molten FeCl2 and its differences from FeCl3 remain scarce. Here, we determined the structures of molten FeCl2 and FeCl3 using High Energy X-ray diffraction (HEXRD), Empirical Potential Structure Refinement (EPSR), and molecular dynamics (MD) simulations with machine learning interatomic potentials (MLIPs). HEXRD measurements provided structure factors and total radial distribution functions (RDFs), which were quantitatively reproduced through EPSR refinement directly constrained by experimental data. MD simulations using MACE foundation and fine-tuned models reproduced experimental structure factors as well as total and partial RDFs, capturing key structural differences between the melts. The models resolved the octahedral to tetrahedral coordination transition of Fe upon melting in FeCl3 and predicted a similar transition in FeCl2. Analysis of MD trajectories quantified coordination environments, bridging Cl populations, bond-angle distributions, and connectivity patterns, revealing distinct degrees of polymerization and local geometry. Polymer chain statistics further showed that, contrary to prior reports, both liquids predominantly consist of extended chains containing six or more Fe centers rather than discrete Fe2Cl6 units. Finally, diffusion coefficients of the two melts calculated from the MACE-MD simulations were compared. Together, these results establish atomic-scale structural benchmarks for molten FeCl2 and FeCl3 and demonstrate the reliability of MACE-based MLIPs for predictive modeling of high-temperature molten salts, while providing practical guidance for MLIP development in complex ionic liquids.

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 reports a combined HEXRD experimental, EPSR refinement, and MACE-MLIP MD study of the atomic structures of molten FeCl2 and FeCl3. It claims quantitative reproduction of experimental structure factors and RDFs by both EPSR and MD, an octahedral-to-tetrahedral coordination shift upon melting (for FeCl3 and predicted for FeCl2), and, from MD connectivity analysis, that both melts consist predominantly of extended polymer chains with six or more Fe centers rather than discrete Fe2Cl6 dimers. Diffusion coefficients are also compared.

Significance. If the structural conclusions are robust, the work supplies much-needed atomic-scale benchmarks for molten iron chlorides relevant to electrochemical iron production and redox flow batteries. The demonstration that MACE-based MLIPs can reproduce high-temperature molten-salt structure factors and RDFs is a positive contribution to the field of machine-learned potentials for ionic liquids.

major comments (2)
  1. [MD trajectory analysis and polymer chain statistics] The central claim that both melts are dominated by extended chains (≥6 Fe centers) rather than Fe2Cl6 units is extracted from MD connectivity analysis of bridging Cl populations. However, the reported agreement is limited to total S(Q) and total RDFs (see abstract and the MD results section). These observables are insensitive to the precise terminal-versus-bridging Cl partitioning; a 5–10 % shift in bridging fraction remains compatible with the stated residuals yet would change the dominant cluster size from chains of ≥6 to mostly dimers or trimers. Additional validation (partial structure factors, angle distributions, or sensitivity tests on bridge lifetimes) is required to support this load-bearing conclusion.
  2. [Results: comparison of HEXRD, EPSR and MACE-MD] The manuscript states that MACE foundation and fine-tuned models capture key structural differences and the coordination transition, but provides limited quantitative metrics (e.g., R-factors, χ² values, or cross-validation against held-out data) for the EPSR and MD fits. Without these, it is difficult to assess whether the models faithfully reproduce the experimental data or whether the reported chain-length statistics could be influenced by model bias.
minor comments (2)
  1. [Experimental Methods] Error bars on the experimental HEXRD structure factors, details of data exclusion or normalization procedures, and explicit validation metrics for the quantitative agreement between experiment and simulation are not fully described; adding these would improve reproducibility and reader confidence.
  2. [Structural Analysis] Notation for partial RDFs and coordination numbers should be clarified (e.g., consistent use of g_{Fe-Cl}(r) versus total RDF) to avoid ambiguity when comparing EPSR and MD results.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their thorough and insightful comments on our manuscript. We have carefully considered each point and provide detailed responses below. We believe the revisions will strengthen the presentation of our results on the structure of molten FeCl2 and FeCl3.

read point-by-point responses

  1. Referee: [MD trajectory analysis and polymer chain statistics] The central claim that both melts are dominated by extended chains (≥6 Fe centers) rather than Fe2Cl6 units is extracted from MD connectivity analysis of bridging Cl populations. However, the reported agreement is limited to total S(Q) and total RDFs (see abstract and the MD results section). These observables are insensitive to the precise terminal-versus-bridging Cl partitioning; a 5–10 % shift in bridging fraction remains compatible with the stated residuals yet would change the dominant cluster size from chains of ≥6 to mostly dimers or trimers. Additional validation (partial structure factors, angle distributions, or sensitivity tests on bridge lifetimes) is required to support this load-bearing conclusion.

    Authors: We appreciate the referee's careful scrutiny of the robustness of our polymer chain statistics. While the total S(Q) and RDFs are indeed the primary experimental observables, our MACE-MD simulations were validated against both total and partial RDFs, as stated in the manuscript. The partial RDFs, particularly Fe-Cl and Cl-Cl, are sensitive to the coordination and bridging character. Furthermore, the reported bond-angle distributions and coordination number analysis provide additional constraints on the local geometry that support the connectivity patterns. To directly address the concern regarding sensitivity to bridging fraction, we will add partial structure factors S_{αβ}(Q) and a sensitivity test varying the bridge lifetime cutoff in the revised manuscript. This will demonstrate that the chain-length distribution remains stable within the range consistent with the experimental data. revision: partial

  2. Referee: [Results: comparison of HEXRD, EPSR and MACE-MD] The manuscript states that MACE foundation and fine-tuned models capture key structural differences and the coordination transition, but provides limited quantitative metrics (e.g., R-factors, χ² values, or cross-validation against held-out data) for the EPSR and MD fits. Without these, it is difficult to assess whether the models faithfully reproduce the experimental data or whether the reported chain-length statistics could be influenced by model bias.

    Authors: We agree that explicit quantitative metrics enhance the assessment of model fidelity. In the original submission, we relied on visual comparison and residual plots for the structure factors and RDFs. For the revised version, we will compute and report R-factors and χ² values for the agreement between simulated and experimental total S(Q) and RDFs for both EPSR and the MACE-MD models. Where possible, we will also include details on the training/validation split for the fine-tuned MACE model to address potential bias concerns. revision: yes

Circularity Check

0 steps flagged

No significant circularity; results anchored in independent HEXRD data

full rationale

The paper's derivation begins with HEXRD experimental structure factors and RDFs as primary inputs. EPSR refinement is explicitly constrained by these data, and MD trajectories with MACE MLIPs are validated by quantitative reproduction of the same experimental quantities. Polymer chain statistics and connectivity analysis are downstream outputs from the validated trajectories rather than inputs redefined or fitted parameters relabeled as predictions. No self-definitional loops, load-bearing self-citations, or ansatzes smuggled via prior work appear in the provided text that would reduce the central claims (extended chains vs. dimers) to the inputs by construction. The structure is self-contained against external experimental benchmarks.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 0 invented entities

Abstract-only review limits visibility into specific parameters; EPSR refinement and MLIP fine-tuning likely introduce fitted quantities whose exact count and values are not stated.

free parameters (2)
  • EPSR empirical potential parameters
    Adjusted to match experimental structure factors for each melt.
  • MACE fine-tuning hyperparameters
    Used to improve agreement with HEXRD data for FeCl2 and FeCl3.
axioms (1)
  • domain assumption HEXRD structure factors accurately represent the pair correlations in the molten salts.
    Foundation for all subsequent EPSR and MD comparisons.

pith-pipeline@v0.9.0 · 5854 in / 1284 out tokens · 51488 ms · 2026-05-19T23:48:22.996558+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

49 extracted references · 49 canonical work pages

  1. [1]

    Shi, C. et al. Redox-structure dependence of molten iron oxides. Commun Mater 1, 80 (2020)

    work page 2020

  2. [2]

    Price, D. L. et al. Structure of molten iron chloride: Neutron scattering and modeling. Phys. Rev. B 57, 10496–10503 (1998)

    work page 1998

  3. [3]

    Mishra, M. et al. Recent advances in iron(III) chloride catalyzed synthesis of heterocycles. Tetrahedron Letters 60, 150925 (2019)

    work page 2019

  4. [4]

    R., Beltrán de Heredia, J., González, T

    Domínguez, J. R., Beltrán de Heredia, J., González, T. & Sanchez-Lavado, F. Evaluation of Ferric Chloride as a Coagulant for Cork Processing Wastewaters. Influence of the Operating Conditions on the Removal of Organic Matter and Settleability Parameters. Ind. Eng. Chem. Res. 44, 6539–6548 (2005)

    work page 2005

  5. [5]

    The Crystal Structure of Ferric Chloride FeCl3

    Wooster, N. The Crystal Structure of Ferric Chloride FeCl3. Zeitschrift für Kristallographie - Crystalline Materials 83, 35–41 (1932)

    work page 1932

  6. [6]

    Greenwood, N. N. & Earnshaw, A. Chemistry of the Elements. (Elsevier, 2012)

    work page 2012

  7. [7]

    S., Saboungi, M.-L., Price, D

    Badyal, Y. S., Saboungi, M.-L., Price, D. L., Haeffner, D. R. & Shastri, S. D. Atomic and electronic structure of liquid iron trichloride. EPL 39, 19 (1997)

    work page 1997

  8. [8]

    & Yelon, W

    Vettier, C. & Yelon, W. B. The structure of FeCl2 at high pressures. Journal of Physics and Chemistry of Solids 36, 401–405 (1975)

    work page 1975

  9. [9]

    A., Bjerrum, N

    Andreasen, H. A., Bjerrum, N. J. & Hansen, N. H. Densities of molten iron(III) chloride, potassium chloride-iron(III) chloride, and potassium chloride-aluminum chloride. J. Chem. Eng. Data 25, 236–239 (1980)

    work page 1980

  10. [10]

    & Moss, S

    Hashimoto, S., Forster, K. & Moss, S. C. Structure refinement of an FeCl3 crystal using a thin plate sample. J Appl Cryst 22, 173–180 (1989)

    work page 1989

  11. [12]

    Xu, J. et al. Molecular dynamics simulations of ionic transport, local structures, and physicochemical properties of multi-component NaCl-MgCl2-CaCl2-FeCl2/FeCl3 molten salt systems. Materials Today Communications 45, 112271 (2025)

    work page 2025

  12. [13]

    & Eriksson, G

    Robelin, C., Pelton, A., Chartrand, P. & Eriksson, G. Models for the Thermodynamic Properties, Density and Viscosity of Molten Salts. in Proceedings of the VIII International Conference on Molten Slags, Fluxes and Salts 673–684 (Gecamin, Santiago, Chile, 2009)

    work page 2009

  13. [14]

    Benmore, C. J. et al. Structure–property relations of binary ferrite melts. Journal of Applied Physics 137, 085903 (2025)

    work page 2025

  14. [15]

    Soper, A. K. Empirical potential Monte Carlo simulation of fluid structure. Chemical Physics 202, 295–306 (1996)

    work page 1996

  15. [16]

    Soper, A. K. Joint structure refinement of x-ray and neutron diffraction data on disordered materials: application to liquid water. J. Phys.: Condens. Matter 19, 335206 (2007)

    work page 2007

  16. [17]

    Roy, S. et al. Unraveling Local Structure of Molten Salts via X-ray Scattering, Raman Spectroscopy, and Ab Initio Molecular Dynamics. J. Phys. Chem. B 125, 5971–5982 (2021)

    work page 2021

  17. [18]

    & Costa Cabral, B

    Galamba, N. & Costa Cabral, B. J. First principles molecular dynamics of molten NaI: Structure, self-diffusion, polarization effects, and charge transfer. J. Chem. Phys. 127, 094506 (2007)

    work page 2007

  18. [19]

    S., Schreiter, E

    Sitze, M. S., Schreiter, E. R., Patterson, E. V. & Freeman, R. G. Ionic Liquids Based on FeCl3 and FeCl2. Raman Scattering and ab Initio Calculations. Inorg. Chem. 40, 2298–2304 (2001)

    work page 2001

  19. [20]

    Batzner, S. et al. E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials. Nat Commun 13, 2453 (2022)

    work page 2022

  20. [21]

    P., Payne, M

    Bartók, A. P., Payne, M. C., Kondor, R. & Csányi, G. Gaussian Approximation Potentials: The Accuracy of Quantum Mechanics, without the Electrons. Phys. Rev. Lett. 104, 136403 (2010)

    work page 2010

  21. [22]

    Klawohn, S. et al. Gaussian approximation potentials: Theory, software implementation and application examples. J. Chem. Phys. 159, 174108 (2023)

    work page 2023

  22. [23]

    & Parrinello, M

    Behler, J. & Parrinello, M. Generalized Neural-Network Representation of High- Dimensional Potential-Energy Surfaces. Phys. Rev. Lett. 98, 146401 (2007)

    work page 2007

  23. [24]

    Tovey, S. et al. DFT Accurate Interatomic Potential for Molten NaCl from Machine Learning. J. Phys. Chem. C 124, 25760–25768 (2020)

    work page 2020

  24. [25]

    Gibson, L., Chahal, R

    D. Gibson, L., Chahal, R. & S. Bryantsev, V. Computing chemical potentials with machine- learning-accelerated simulations to accurately predict thermodynamic properties of molten salts. Chemical Science 16, 3078–3091 (2025)

    work page 2025

  25. [26]

    Chahal, R. et al. Transferable Deep Learning Potential Reveals Intermediate-Range Ordering Effects in LiF–NaF–ZrF4 Molten Salt. JACS Au 2, 2693–2702 (2022)

    work page 2022

  26. [27]

    Shi, Y., T. Lam, S. & L. Beck, T. Deep neural network based quantum simulations and quasichemical theory for accurate modeling of molten salt thermodynamics. Chemical Science 13, 8265–8273 (2022)

    work page 2022

  27. [28]

    Yuan, E. C.-Y. et al. Foundation models for atomistic simulation of chemistry and materials. Nat Rev Chem 10, 212–230 (2026)

    work page 2026

  28. [29]

    & Jung, Y

    Choi, J., Nam, G., Choi, J. & Jung, Y. A Perspective on Foundation Models in Chemistry. JACS Au 5, 1499–1518 (2025)

    work page 2025

  29. [30]

    Batatia, I. et al. A foundation model for atomistic materials chemistry. J. Chem. Phys. 163, 184110 (2025)

    work page 2025

  30. [31]

    P., Simm, G

    Batatia, I., Kovacs, D. P., Simm, G. N. C., Ortner, C. & Csanyi, G. MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fields. in Advances in Neural Information Processing Systems 35 vol. 35 11423–11436 (2022)

    work page 2022

  31. [32]

    Kovács, D. P. et al. MACE-OFF: Short-Range Transferable Machine Learning Force Fields for Organic Molecules. J. Am. Chem. Soc. 147, 17598–17611 (2025)

    work page 2025

  32. [33]

    Shen, C. et al. SuperSalt: equivariant neural network force fields for multicomponent molten salts system. Nat Commun 16, 7280 (2025)

    work page 2025

  33. [34]

    Qiu, X., Thompson, J. W. & Billinge, S. J. L. PDFgetX2: a GUI-driven program to obtain the pair distribution function from X-ray powder diffraction data. J Appl Cryst 37, 678–678 (2004)

    work page 2004

  34. [35]

    Keen, D. A. A comparison of various commonly used correlation functions for describing total scattering. J Appl Cryst 34, 172–177 (2001)

    work page 2001

  35. [36]

    Thompson, A. P. et al. LAMMPS - a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales. Computer Physics Communications 271, 108171 (2022)

    work page 2022

  36. [37]

    Atomsk: A tool for manipulating and converting atomic data files

    Hirel, P. Atomsk: A tool for manipulating and converting atomic data files. Computer Physics Communications 197, 212–219 (2015)

    work page 2015

  37. [38]

    Martínez, L., Andrade, R., Birgin, E. G. & Martínez, J. M. PACKMOL: A package for building initial configurations for molecular dynamics simulations. Journal of Computational Chemistry 30, 2157–2164 (2009)

    work page 2009

  38. [39]

    Visualization and analysis of atomistic simulation data with OVITO–the Open Visualization Tool

    Stukowski, A. Visualization and analysis of atomistic simulation data with OVITO–the Open Visualization Tool. Modelling Simul. Mater. Sci. Eng. 18, 015012 (2009)

    work page 2009

  39. [40]

    Gao, M., Palmer, D. C. & Dove, M. T. A new approach to molecular and lattice simulations with CrystalMaker® 11. MRS Commun. 15, 1007–1016 (2025)

    work page 2025

  40. [41]

    P., Kondor, R

    Bartók, A. P., Kondor, R. & Csányi, G. On representing chemical environments. Phys. Rev. B 87, 184115 (2013)

    work page 2013

  41. [42]

    P., Csányi, G

    De, S., Bartók, A. P., Csányi, G. & Ceriotti, M. Comparing molecules and solids across structural and alchemical space. Phys. Chem. Chem. Phys. 18, 13754–13769 (2016)

    work page 2016

  42. [43]

    https://pubs.acs.org/doi/10.1021/acs.jpclett.0c02405

    Accurate and Numerically Efficient r2SCAN Meta-Generalized Gradient Approximation | The Journal of Physical Chemistry Letters. https://pubs.acs.org/doi/10.1021/acs.jpclett.0c02405

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

  43. [44]

    & Furthmüller, J

    Kresse, G. & Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169–11186 (1996)

    work page 1996

  44. [45]

    Z., Rambidi, N

    Zasorin, E. Z., Rambidi, N. G. & Akishin, P. A. An electron diffraction study of the structure of the ferric chloride molecule in vapors. J Struct Chem 4, 836–838 (1963)

    work page 1963

  45. [46]

    & Hargittai, I

    Hargittai, M., Tremmel, J. & Hargittai, I. Molecular structure of dimeric iron trichloride in the vapour phase as determined by electron diffraction. Journal of the Chemical Society, Dalton Transactions 0, 87–89 (1980)

    work page 1980

  46. [47]

    Papatheodorou, G. N. & Voyiatzis, G. A. Vibrational modes and structure of molten iron(III) chloride. Chemical Physics Letters 303, 151–156 (1999)

    work page 1999

  47. [48]

    ionic” to “molecular

    Hutchinson, F., Walters, M. K., Rowley, A. J. & Madden, P. A. The “ionic” to “molecular” transitions in AlCl3 and FeCl3 as predicted by an ionic interaction model. J. Chem. Phys. 110, 5821–5830 (1999)

    work page 1999

  48. [49]

    & Hargittai, I

    Hargittai, M., Tremmel, J. & Hargittai, I. Molecular structure of dimeric iron trichloride in the vapour phase as determined by electron diffraction. J. Chem. Soc., Dalton Trans. 87–89 (1980) doi:10.1039/DT9800000087

    work page doi:10.1039/dt9800000087 1980

  49. [50]

    Shannon, R. D. Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides. Acta Cryst A 32, 751–767 (1976). Supplementary Information Figure S1: Comparison of MACE foundation model performance for modelling molten iron salts. Structure factors are shown on the left panel and RDFs on the right panel. Three fo...

    work page 1976