CoChat
Sign Up
Cited by 36Metal-Organic Frameworks: Synthesis and ApplicationsMachine Learning in Materials ScienceNanomaterials for catalytic reactions

Conversational Large-Language-Model Artificial Intelligence Agent for Accelerated Synthesis of Metal-Organic Frameworks Catalysts in Olefin Hydrogenation

Lin, Jing, Zhao, Danfeng, Lu, Shaopeng, Li, Rushuo, Xu, Xinmeng, Wang, Zhaokun, Li, Wenqing, Ji, Yujing +5 more·2025·ACS Nano
DOI 10.1021/acsnano.5c04880
Open this paper in CoChat — free account

Abstract

Metal–organic frameworks (MOFs) attract significant attention for their structural diversity and design flexibility, making them ideal candidates for catalytic applications. However, the traditional trial-and-error approach for optimizing MOF synthesis remains inefficient. In this study, we introduce the MOFsyn agent, an AI-driven framework that harnesses large language models (LLMs) for MOF synthesis optimization. This system integrates data automatic analysis, material mechanism analysis, and experimental protocol navigation by employing retrieval-augmented generation (RAG) to refine synthetic strategies based on natural language inputs. Using Ni@UiO-66(Ce) for olefin hydrogenation as a case study, the MOFsyn agent analyzed the relationship between synthesis conditions, structural characteristics, and catalytic performance, with a particular focus on the electronic structure of nickel. Through adaptive optimization, a novel stepwise reduction strategy was proposed that outperformed conventional one-pot reduction. The optimized Ni@UiO-66(Ce)-R2T1, synthesized under MOFsyn agent’s guidance, exhibited nearly twice the Ni 0 /Ni total ratio compared to the best-performing sample from an initial experimental set and achieved 100% conversion and selectivity for dicyclopentadiene hydrogenation under mild conditions (70 °C, 2 MPa). These results validate the accuracy and efficiency of the MOFsyn agent. This study provides an efficient tool for intelligent material synthesis, enabling researchers without programming expertise to accelerate material development.

Integrity signals

Registry

Crossref · 10.1021/acsnano.5c04880

Cites

47 of 47 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Claims the paper makes

This paper’s full text isn’t openly available, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Catalysis and adsorption tasks formed the largest application block, represented by six studies on electrocatalyst design, adsorption-configuration search, biomass-valorization catalysis, MOF catalyst optimization, and single-atom-alloy discovery.

    “Catalysis and adsorption tasks formed the largest application block, represented by six studies on electrocatalyst design, adsorption-configuration search, biomass-valorization catalysis, MOF catalyst optimization, and single-atom-alloy discovery [19,20,27,28,30,34].”

    From From Scientific Copilots to Tool-Grounded Autonomy: AI Agents in Simulation-Driven Materials Discovery (2026) · cites this paper as [20]

  • A second group of studies used hybrid human-in-the-loop designs, where the agentic layer guided design or analysis but final judgment remained partly manual.

    “A second group of studies used hybrid human-in-the-loop designs, where the agentic layer guided design or analysis but final judgment remained partly manual [20,22,27,30,31].”

    From From Scientific Copilots to Tool-Grounded Autonomy: AI Agents in Simulation-Driven Materials Discovery (2026) · cites this paper as [20]

  • In several papers, the strongest reproducibility evidence came from public code, tutorials, datasets, or supplementary traces.

    “In several papers, the strongest reproducibility evidence came from public code, tutorials, datasets, or supplementary traces [18-21,23,25,28,30,31,35-42].”

    From From Scientific Copilots to Tool-Grounded Autonomy: AI Agents in Simulation-Driven Materials Discovery (2026) · cites this paper as [20]

  • Other studies act more as discovery managers around retrieval, ranking, surrogate models, or protocol navigation.

    “Others act more as discovery managers around retrieval, ranking, surrogate models, or protocol navigation [19,20,26,30,41].”

    From From Scientific Copilots to Tool-Grounded Autonomy: AI Agents in Simulation-Driven Materials Discovery (2026) · cites this paper as [20]

  • Hybrid human-in-the-loop systems accounted for four of the six experimentally validated papers, which suggests that human oversight appears to be retained more often where the downstream consequences of an incorrect decision are more costly or irreversible, although the small sample precludes strong inference.

    “Hybrid human-in-the-loop systems accounted for four of the six experimentally validated papers, which suggests that human oversight appears to be retained more often where the downstream consequences of an incorrect decision are more costly or irreversible, although the small sample precludes strong inference [20,22,27,30,31].”

    From From Scientific Copilots to Tool-Grounded Autonomy: AI Agents in Simulation-Driven Materials Discovery (2026) · cites this paper as [20]

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Lin, J., Zhao, D., Lu, S., Li, R., Xu, X., Wang, Z., Li, W., Ji, Y., Zhang, C., Shi, L., Jin, X., Gao, H., & Wang, G.
Affiliations
  • University of Science and Technology Beijing (CN)
  • Research Institute of Petroleum Exploration and Development (CN)
  • Beihang University (CN)
  • Research Institute of Petroleum Exploration & Development
Corresponding author
  • Gao, Hongyi
  • Wang, Ge
ORCID
Venue
ACS Nano
Publisher
American Chemical Society (ACS)
Published
2025-06-24
Volume
19
Issue
26
Pages
23840-23858
ISSN
  • 1936-0851 (ISSN-L)
  • 1936-086X
DOI
10.1021/acsnano.5c04880
OpenAlex
W4411572762
Cited by
36
References
47
Citations by year
2026: 28 · 2025: 8
FWCI
7.95
Type
journal article
Language
en
Open access
closed
Licence
Locations
  • ACS Nano (publishedVersion)
  • PubMed (publishedVersion)
Funders
  • National Key Research and Development Program of China (2021YFB3500700)
  • State Key Laboratory of Virtual Reality Technology and Systems (VRLAB2025C11)
  • Natural Science Foundation of Beijing Municipality (L233011)
  • National Natural Science Foundation of China (52373261)
Topics
  • Physical Sciences › Chemistry › Inorganic Chemistry › Metal-Organic Frameworks: Synthesis and Applications
  • Physical Sciences › Materials Science › Materials Chemistry › Machine Learning in Materials Science
  • Physical Sciences › Chemistry › Organic Chemistry › Nanomaterials for catalytic reactions
Keywords
metal-organic frameworks, olefin hydrogenation, nickel catalysts, UiO-66(Ce), retrieval-augmented generation, large language models, conversational AI agents

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review