AdsMind: A Physics-Grounded Multi-Agent System for Self-Correcting Discovery of Adsorption Configurations on Heterogeneous Catalyst Surfaces

2026-06-17

AI Summary

This work addresses the computational challenge of exploring vast configuration spaces in heterogeneous catalysis modeling, where existing methods lack physically informed self-correction capabilities. The authors propose AdsMind, a closed-loop multi-agent system that introduces, for the first time, a physics-driven self-correction mechanism. By integrating large language model agents, machine learning force fields (MLFF), and density functional theory (DFT) validation, AdsMind autonomously refines initial adsorption guesses through iterative MLFF-based structural relaxation feedback. Evaluated on the AA20 and OCD-GMAE62 benchmarks, AdsMind achieves success rates of 100% and 98.8%, respectively, requiring only 4–5 MLFF relaxations—reducing computational cost by approximately 14-fold compared to heuristic enumeration. DFT verification confirms that the predicted adsorption energies not only exhibit correct sign consistency but also superior quantitative accuracy.

Abstract

Identifying the lowest-energy surface-adsorbate configuration is critical for modeling heterogeneous catalysis, yet exhaustive exploration with ab initio calculations is computationally prohibitive. Machine-learning force fields (MLFFs) accelerate structural relaxation but leave the search over the vast configurational space a major bottleneck, and open-loop large language model (LLM) agents lack a physics-grounded feedback mechanism to correct erroneous initial guesses. We propose AdsMind (Adsorption configuration discovery with Machine intelligence and relaxation feedback), a closed-loop multi-agent framework that enables autonomous error correction through MLFF relaxation feedback. Across four LLM backends, AdsMind achieves consistently high search reliability, with success rates of 100% and 98.8% on the benchmarks AA20 and OCD-GMAE62. Relative to its single-pass (1-Shot) ablation it reduces cross-backend energy dispersion, and it uses only 4.11 and 4.67 MLFF relaxations per case, respectively -- an approximately 14-fold reduction over heuristic enumeration baselines. Density functional theory (DFT) validation using VASP/PBE on six representative AA20 systems shows that the reported open-loop Adsorb-Agent outputs exhibit qualitative adsorption-energy sign errors for molecular adsorbates, whereas AdsMind preserves the correct sign in all tested cases with closer quantitative agreement. AdsMind thus delivers reliability, self-reflection, and interpretability simultaneously, supporting more DFT-informed autonomous chemistry workflows.

Problem

• adsorption configuration
• heterogeneous catalysis
• configurational space search
• physics-grounded feedback
• machine-learning force fields

Innovation

• multi-agent system
• machine-learning force fields
• self-correction
• adsorption configuration
• heterogeneous catalysis