Skip to content

eReaxFF force field development for BaZr0.8Y0.2O3-delta solid oxide electrolysis cells applications

Summary

This corpus PDF is an uncorrected proof galley of the npj Computational Materials article on eReaxFF for BZY20 solid-oxide electrochemistry (same DOI as the version-of-record file paired as 2024hossain-npj-computat-ereaxff-force). Scientific content matches that publication: eReaxFF training to DFT/C-DFT for vacancies, water reactions, and explicit-electron effects, plus large-scale AMS/LAMMPS-style simulations of steam on BZY20 surfaces.

Galley PDFs sometimes differ in pagination, line breaks, and figure placement; operators should cite locator-independent quantities (reaction labels, barrier heights, supercell stoichiometries) from whichever file is designated canonical after manifest review.

Methods

QM fitting targets mirror the VOR article: bulk BZY20 Y-doping preference (Type 3), oxygen-vacancy energies in bulk and on (100)/(110) slabs (Ba-O vs Zr-Y-O terminations; surface vacancies up to 25% in supercells), EOS, H2O adsorption/splitting, H2 release energetics, and C-DFT excess-electron localization and migration barriers on cluster models. eReaxFF balances reproduction of relative energies and structural metrics; absolute vacancy formation energies are described as qualitatively tracked.

Atomistic simulations use the fitted eReaxFF with ACKS2 charges and optional explicit electrons; bond-biased scans estimate adsorption, water-splitting, and H2 formation barriers. Production MD (e.g. ~3000-atom, 1000 K, NVT, repeated H2O dosing) probes vacancy filling, proton hopping into the bulk, and bias-assisted H2 pathways; galley pagination differs, but equations and numerical examples align with the journal PDF.

Training prioritizes reproducing the ordering of surface elementary steps and proton-electron coupling trends needed for electrolysis modeling rather than perfect absolute formation energies for every oxygen vacancy configuration.

1 — MD / 2 — Force-field details align with the VOR article’s Methods; this proof PDF is not the preferred citation for page numbers. RMD examples use AMS/eReaxFF molecular dynamics as in the VOR: 3D PBC metallic-oxide slabs, NVT kinetics of H2O at ~1000 K, ~1.4 ns-scale cumulative steam exposure, 0 GPa external hydrostatic target (N/A to NPT Parrinello barostats in the same excerpt). For timestep, Langevin/Nose thermostat, and per-segment ps/ns duration, use 2024hossain-npj-computat-ereaxff-force. 3 — Static-only QM as standalone primary claim: N/A—DFT is in service of eReaxFF training, as in the sibling page.

Findings

The proof reports the same core validation outcomes as the article: Type 3 doping stability, EOS agreement, semiquantitative Ov trends, surface reaction-energy ordering (Zr-Y-O vs Ba-O terminations), high thermal barriers for H2 evolution without bias, lowering of barriers when excess electrons or H-H bond biasing stand in for applied voltage, and long MD showing subsurface proton transfer without H2 in zero-bias steam simulations. Use 2024hossain-npj-computat-ereaxff-force for page-cited reading when possible.

Readers should interpret explicit-electron and biased-bond workflows as mechanistic analogues of electrochemical driving force, not literal constant-voltage continuum boundary conditions. For kcal/mol and reaction labels, prefer the version-of-record wiki page and PDF (2024hossain-npj-computat-ereaxff-force).

Limitations

Proof PDFs can diverge slightly from the final layout; any figure numbering referenced externally should be checked against the journal version-of-record PDF linked by DOI.

Relevance to group

Co-authored eReaxFF development for proton-conducting perovskite electrolytes with Adri C. T. van Duin, supporting solid-oxide electrochemistry and interface simulation threads in the KB.

Citations and evidence anchors