Additional data for "How accurate are foundational machine learning interatomic potentials for heterogeneous catalysis?"

Additional ASE Atoms objects required for the analyses in the paper "How accurate are foundational machine learning interatomic potentials for heterogeneous catalysis?", but not provided in the data publicly available for the cited publications.

These objects are downloaded whenever necessary by the respective data set implementations in the MLIP Cats package, and are hosted here so that the package does not need to contain any data files. This data is not required for analysis of the data from the paper cited above; for this, we refer to the main data set for the paper.

This data was originally generated in the following publications:

nanoclusters.zip: Kempen, L.H.E., Andersen, M. Inverse catalysts: tuning the composition and structure of oxide clusters through the metal support. npj Computational Materials 11, 8 (2025)

nanoclusters-formate.zip: Nielsen, M.J., et al. Interpretable machine learned predictions of adsorption energies at the metal–oxide interface. The Journal of Chemical Physics 163, 4, 044708 (2025)

single-atom-alloys.zip: Cheula, R., Andersen, M. Transition States Energies from Machine Learning: An Application to Reverse Water–Gas Shift on Single-Atom Alloys. ACS Catalysis 15, 13, 11377–11388 (2025)