Autonomous heterogeneous catalyst discovery with a self-evolving multi-agent digital twin

2026-06-03

AI Summary

This work proposes CatDT, a self-evolving multi-agent digital twin system for heterogeneous catalysis that overcomes the limitations of traditional simulation methods, which often deviate from experimental conditions and are restricted to specific material systems. Requiring only crystal structures and natural-language reaction descriptions as input, CatDT autonomously predicts stable facets, reconstructs working surfaces, enumerates reaction pathways, and computes kinetics. Key innovations include the UniMech algorithm, which identifies dominant reaction paths at three orders of magnitude lower computational cost; a memory-augmented reinforcement learning loop that boosts transition state search success rates from 41% to 84%; and an integrated architecture combining eight specialized agent types, 27 scientific tools, and large language model coordination. Validated across seven gas–solid catalytic systems, CatDT achieves predictions within 0.5–2× of experimental values and autonomously discovers a highly active non-precious Ni@ZrO₂ catalyst for propane dehydrogenation (TOF = 1.63 s⁻¹, selectivity ≈ 100%).

Abstract

Theoretical heterogeneous catalysis promises rapid catalyst discovery, yet computational and machine-learning predictions often deviate from experiment and stay confined to narrow material families, for want of a faithful, condition-aware catalytic simulator. We present CatDT (Catalysis Digital Twin), a self-evolving multi-agent system that builds an autonomous digital twin of a working catalyst, unifying gas-solid and liquid-solid modeling. From only a bulk crystal and a natural-language reaction description, eight specialized agents and 27 scientific tools predict stable facets, reconstruct working surfaces, enumerate and rank reaction pathways, locate transition states, and compute kinetics in 5-30 min on a single GPU. Two innovations address the hardest steps: UniMech finds dominant pathways for novel materials at over $10^3\times$ lower cost than exhaustive enumeration by fusing agent-guided proposals with energy-cached graph search, and a memory-augmented reinforcement loop raises barrier-calculation success from 41\% to 84\% across 600 catalytic surfaces. Across seven gas-solid benchmarks -- stepped metals, single-atom catalysts, ordered intermetallics, vacancy-rich 2D sulfides and carbides, and a strong-metal--support-interaction (SMSI) interface -- every CatDT prediction lies within 0.5-2 times experiment over four orders of magnitude. For propane dehydrogenation, CatDT independently discovers non-precious candidates rivaling the Pt-based industrial benchmark, with a proposed Ni@ZrO$_2$ SMSI overlayer reaching a simulated TOF of $1.63~\text{s}^{-1}$ at $\sim$100\% selectivity. More broadly, the decisive factor for a faithful catalyst digital twin -- or any multi-stage scientific simulator -- is not raw LLM capability but the engineered harness around it: deterministic tools, persistent memory, and verified self-improvement that compound across models, tools, and runs.

Problem

• heterogeneous catalysis
• catalyst discovery
• digital twin
• reaction pathways
• condition-aware simulation

Innovation

• digital twin
• multi-agent system
• heterogeneous catalysis
• reaction pathway prediction
• self-evolving simulation