Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
To evaluate whether general-purpose foundation machine learning interatomic potentials (MLIPs) generalize to surface chemistry, the authors constructed HetCat26 and benchmarked fifteen pre-trained models across surface energetics, metal-metal oxide interfaces, adsorption, and reaction networks. Interfacial validation is critical because standard materials science benchmarks correlate only weakly with performance in catalytic regimes. The evaluation reveals that while top models like eSEN-30M-OAM and MACE-MH-1-OMAT reliably predict surface energies and reaction barriers, current foundation MLIPs broadly fail to reproduce adsorption energetics and DFT site preferences, issues exacerbated by mixing PBE and PBE+U functionals in training corpora.
Top materials foundation models can accurately predict complex catalytic reaction barriers, yet routinely fail at basic adsorption site preferences—a failure mode that general benchmarks leave entirely hidden.
Foundation machine learning interatomic potentials (MLIPs) promise near-density functional theory (DFT) accuracy across broad areas of chemistry and materials science. However, their performance in describing systems and processes relevant to heterogeneous catalysis remains underexplored. Here, we introduce HetCat26, a collection of benchmark tests designed to assess pre-trained MLIPs across key aspects of catalytic modeling, including surface energetics, metal-metal oxide interactions, adsorption, and catalytic reaction networks. Evaluating fifteen foundation models, we find that performance on existing general materials benchmarks is only weakly predictive of performance on HetCat26; transferability to heterogeneous catalysis cannot be inferred from these general benchmarks. Across the benchmark tests, current models describe surface energetics and, perhaps surprisingly, reaction barriers well, whereas larger errors are observed for adsorption, and DFT site preferences are often not reproduced. Two models, eSEN-30M-OAM and MACE-MH-1-OMAT, nevertheless achieve high accuracy across the properties evaluated. Beyond model benchmarking, HetCat26 highlights the importance of training data consistency: PBE and PBE+U calculations should not be mixed within a training set. Overall, we identify challenges limiting the transferability of current foundation MLIPs to heterogeneous catalysis and, more broadly, to chemical reactions at interfaces, providing guidance for the development of the next generation of foundation models.