Large language model agents accelerate inverse design of metal-organic frameworks for gas separation

2026-07-12

AI Summary

This work addresses the formidable challenge of inverse design for metal–organic frameworks (MOFs), which must simultaneously satisfy chemical validity, separation performance, and structural diversity within an astronomically large combinatorial space. To this end, the authors propose LEMO Agent, a novel framework that, for the first time, integrates large language model (LLM) agents into MOF inverse design. Operating in the standardized MOFid space, LEMO Agent enables closed-loop optimization by synergistically combining LLM-based generation, explicit validity checking, Transformer-based performance prediction, structured memory, and multi-island parallel exploration. Evaluated on CH₄/N₂ and CO₂/N₂ separation tasks, the approach substantially increases the proportion of high-performance candidates and prediction accuracy, leading to the successful synthesis and experimental validation of new MOFs. This paradigm offers an interpretable, scalable, and efficient design strategy that effectively balances constraints with diversity.

Abstract

Metal-organic frameworks (MOFs) offer a highly modular platform for adsorptive gas separation, yet their vast reticular design space makes inverse design difficult under simultaneous constraints of chemical validity, separation performance, and structural diversity. Here, we present LEMO Agent, a large-language-model agent framework for closed-loop inverse design of gas-separation MOFs in MOFid space. LEMO Agent couples language-based candidate generation with MOFid standardization, explicit validity checking, Transformer-based property prediction, structured design memory, and multi-island exploration. Through iterative generate--validate--evaluate--remember cycles, the agent uses feedback from both successful and failed candidates to guide chemically constrained search across linker, metal, and topology choices. We evaluate LEMO Agent on CH$_4$/N$_2$ and CO$_2$/N$_2$ separation tasks. Compared with representative generative, optimization, and agentic baselines, LEMO Agent enriches high-performing candidates, improves predicted separation performance, and maintains broad chemical and topological diversity. Selected candidates are further reconstructed, evaluated by GCMC simulations, and passed through an experimental down-selection workflow based on chemical feasibility and ligand purchasability, leading to initial wet-lab synthesis and SEM characterization. These results demonstrate that large language model agents can serve as interpretable and scalable design engines for accelerating MOF discovery beyond conventional fixed-library screening.

Problem

• inverse design
• metal-organic frameworks
• gas separation
• chemical validity
• structural diversity

Innovation

• large language model agent
• inverse design
• metal-organic frameworks
• MOFid
• closed-loop discovery