Coordination-Resolved Surrogate Models for Thermodynamic Stability, Band Gaps, and Magnetic Moments of Spinel Oxides, Sulfides, and Selenides
arXiv:2607.11996
The authors curated a dataset of 320 cubic spinel oxides, sulfides, and selenides and trained tree‑ensemble surrogate models to predict formation energy, hull distance, band gap, magnetic moment, and metallicity, using coordination‑based grouping of cations. Their models achieve low mean absolute errors and are evaluated with rigorous grouped cross‑validation and interpretability analyses.
Abstract
We curated 320 cubic () spinel entries from the Materials Project-nitrides, oxides, sulfides, and selenides, including single-cation mixed-valence compounds-and trained tree-ensemble surrogates for the formation energy, energy above the convex hull, band gap, total magnetization, and metallicity. Cations were assigned to tetrahedral-like and octahedral-like groups from CrystalNN coordination numbers rather than from element identity, and the evaluation was group-aware throughout: splits were grouped by reduced formula, every transform was fit on training folds only, and champion models were selected on cross-validated scores before the holdout was examined. Over twenty repeated grouped holdouts (single-seed refits that reuse the tuned hyperparameters, and are therefore mildly optimistic) the champions reach mean absolute errors of eV/atom for the formation energy, eV/atom for the hull distance, and ~\muBfu{} for the magnetization, with a metallicity accuracy of . Band-gap regression does not beat a trivial baseline on the 19-member non-metal holdout under paired bootstrap testing, we report this negative result and trace it to sample scarcity and to the semi-local DFT labels. On the identical grouped split, the tabular champion is more accurate than an untuned MEGNet trained from scratch (0.087 versus 0.209 eV/atom formation-energy MAE on the primary holdout), a comparison that bounds, rather than settles, the descriptor-versus-graph question at this data scale. SHAP attribution ties the magnetization model to octahedral -occupancy and the formation-energy and band-gap models to electronegativity descriptors, and grouped conformal intervals, permutation nulls, and leave-one-chemistry-out tests bound a domain of applicability that is uneven across anions and cations.