Interpretable Inverse Design of Metal-Organic Frameworks with Large Language Model Agents
Abstract
Inverse design of metal-organic frameworks (MOFs) requires navigating combinatorial spaces with costly property labels and opaque machine-learning models. We introduce LLM4MOF, a closed-loop multi-agent framework that converts a natural-language target into chemical hypotheses, constraints, diagnostic tests, and feedback. One agent proposes interpretable hypotheses over metal nodes, linkers, pore geometry, and functionality. Another converts them into constraints selecting MOFs defined by a node, linker, and topology. The Matchmaker forms four beams to attribute gains to geometry, chemistry, or metal choice: full hypothesis, chemistry, metal only, and random baseline. Blind to database landscapes, LLM4MOF enriches top performers across six adsorption, separation, and electronic-structure tasks within 400 evaluations. It also designs and live-simulates de novo MOFs spanning H2 storage, SF6 capture, and C2H6/C2H4 separation, deriving a distinct design rule for each objective. Under an identical nominal evaluation budget it consistently outperforms random search, Bayesian optimization, and genetic algorithms, and the outcome is insensitive to the language-model backend. All of this uses not a single property-labeled training structure, whereas generative alternatives train on thousands to hundreds of thousands. A new objective requires only a natural-language request: interpretable inverse design at a fraction of the data cost of existing approaches.
Related
- Phenomenon-first problem formulation with language models in metal-organic frameworks
- Chiroptical Nonlinearity Found in Covalent Organic Frameworks
- Ultrasound-Assisted Facile Nanoconfinement of Luminescent Guests in Zirconium Metal-Organic Frameworks
- Molecular Curvature Enables Tubular π-Stacked Pathways in Three-Dimensional Conductive Metal-Organic Frameworks
- Designing Geared Molecular Rotors in Flexible Metal-Organic Frameworks
BibTeX
@misc{nam2026interpretable,
title = {{Interpretable Inverse Design of Metal-Organic Frameworks with Large Language Model Agents}},
author = {Kyungmin Nam and Seunghee Han and Jihan Kim},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2606.29459},
eprint = {2606.29459},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2606.29459},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - Interpretable Inverse Design of Metal-Organic Frameworks with Large Language Model Agents
AU - Kyungmin Nam
AU - Seunghee Han
AU - Jihan Kim
PY - 2026
DA - 2026/09/13
PB - arXiv
DO - 10.48550/arxiv.2606.29459
UR - https://arxiv.org/abs/2606.29459
AB - Inverse design of metal-organic frameworks (MOFs) requires navigating combinatorial spaces with costly property labels and opaque machine-learning models. We introduce LLM4MOF, a closed-loop multi-agent framework that converts a natural-language target into chemical hypotheses, constraints, diagnostic tests, and feedback. One agent proposes interpretable hypotheses over metal nodes, linkers, pore geometry, and functionality. Another converts them into constraints selecting MOFs defined by a node, linker, and topology. The Matchmaker forms four beams to attribute gains to geometry, chemistry, or metal choice: full hypothesis, chemistry, metal only, and random baseline. Blind to database landscapes, LLM4MOF enriches top performers across six adsorption, separation, and electronic-structure tasks within 400 evaluations. It also designs and live-simulates de novo MOFs spanning H2 storage, SF6 capture, and C2H6/C2H4 separation, deriving a distinct design rule for each objective. Under an identical nominal evaluation budget it consistently outperforms random search, Bayesian optimization, and genetic algorithms, and the outcome is insensitive to the language-model backend. All of this uses not a single property-labeled training structure, whereas generative alternatives train on thousands to hundreds of thousands. A new objective requires only a natural-language request: interpretable inverse design at a fraction of the data cost of existing approaches.
ER -