LLM-Driven Discovery of High-Entropy Catalysts via Retrieval-Augmented Generation

AI Scientists; Xinyi Lin; Danqing Yin; Ying Guo

cond-mat.mtrl-sciMar 16, 2026

LLM-Driven Discovery of High-Entropy Catalysts via Retrieval-Augmented Generation

Authors: AI Scientists, Xinyi Lin, Danqing Yin, Ying Guo

Organizations: School of Biomedical Sciences, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Pokfulam, Hong Kong SAR, 999077, China · Laboratory of Data Discovery for Health Limited (D24H), Pak Shek Kok, Hong Kong SAR, 999077, China · Room 312, Lau Chung Him Building, 8 Castle Peak Road, Tuen Mun, Hong Kong

Abstract

CO2 reduction requires efficient catalysts, yet materials discovery remains bottlenecked by 10-20 year development cycles requiring deep domain expertise. This paper demonstrates how large language models can assist the catalyst discovery process by helping researchers explore chemical spaces and interpret results when augmented with retrieval-based grounding. We introduce a retrieval-augmented generation framework that enables GPT-4 to navigate chemical space by accessing a database of 50,000+ known materials, adapting general-purpose language understanding for high-throughput materials design. Our approach generated over 250 catalyst candidates with an 82% thermodynamic stability rate while addressing multi-objective constraints: 68% achieved < $100/kg cost with metallic conductivity (band gap<0.1eV) and mechanical stability (B/G>1.75). The best-performing Fe0.2Co0.2Ni0.2Ir0.1Ru0.3 achieves 0.285V limiting potential (25% improvement over IrO2), while Cr0.2Fe0.2Co0.3Ni0.2Mo0.1 optimally balances performance-cost trade-offs at$ 18/kg. Volcano plot analysis confirms that 78% of LLM-generated catalysts cluster near the theoretical activity optimum, while our system achieves 200x computational efficiency compared to traditional high-throughput screening. By demonstrating that retrieval-augmented generation can ground AI creativity in physical constraints without sacrificing exploration, this work demonstrates an approach where natural language interfaces can streamline materials discovery workflows, enabling researchers to explore chemical spaces more efficiently while the LLM assists in result interpretation and hypothesis generation.

Figures & tables

Figure 1: LLM-driven catalyst discovery pipeline: RAG retrieval → LLM generation → DFT validation.

Catalyst Composition	$\eta_{OER}$	$E_{hull}$	Band Gap	B/G	Cost	Score ∗
	(V)	(meV/atom)	(eV)	Ratio	($/kg)
Fe 0.2 Co 0.2 Ni 0.2 Ir 0.1 Ru 0.3	0.285	32	0.0	2.1	27,000	0.72
Mn 0.15 Fe 0.25 Co 0.25 Ni 0.2 Pt 0.15	0.298	28	0.0	1.9	4,500	0.85
Cr 0.2 Fe 0.2 Co 0.3 Ni 0.2 Mo 0.1	0.312	41	0.0	2.3	18	0.91
V 0.1 Cr 0.2 Mn 0.2 Fe 0.25 Co 0.25	0.325	37	0.08	1.8	15	0.88
Ti 0.1 Fe 0.3 Co 0.3 Ni 0.2 Cu 0.1	0.334	45	0.0	2.0	19	0.89

Table 1: Multi-objective performance comparison of top LLM-generated catalysts. Beyond catalytic activity (

\eta_{OER}

), we evaluate conductivity (band gap), mechanical stability (Pugh’s ratio B/G), and material cost. Statistical significance assessed using Wilcoxon signed-rank test with Bonferroni correction (

\alpha

=0.0002 for 250 comparisons).

Figure 2: Comprehensive comparison of material properties between known catalysts and LLM-generated catalysts (HEA: High-Entropy Alloy, DA: Doped Alloy). The visualization maps catalysts by mixing enthalpy and d-band center, with LLM-HEAs occupying the favorable lower-left quadrant. Property distributions show LLM-HEAs exhibit more negative mixing enthalpies (mean -0.794 eV/atom) indicating higher stability, and more negative d-band centers (mean -2.891 eV) correlating with enhanced catalytic activity.

Figure 3: Volcano plot analysis showing the relationship between oxygen binding energy ( $\Delta E_{*O}$ ) and theoretical overpotential for LLM-generated catalysts (blue circles) compared to known catalysts (red triangles). The optimal region near the volcano peak is highlighted, where most LLM candidates cluster, explaining their superior performance. Error bars represent standard deviations from ensemble DFT calculations.

Figure 4: Performance ranking of all validated catalysts showing the distribution of limiting potentials. LLM-generated HEAs (blue) consistently outperform both traditional catalysts (red) and randomly generated compositions (gray). The top quartile is dominated by LLM discoveries, with 18 of the best 25 catalysts originating from our approach.

Figure 5: Activity landscape and optimization paths showing the iterative refinement process. The contour map represents limiting potential as a function of $\Delta E_{NOH}$ and mixing enthalpy, with the best catalyst (red star) identified through systematic exploration. Red paths trace the convergence trajectory from initial candidates to the optimal composition, demonstrating efficient navigation of the 2D property space.

Appendix figures & tables11 assets

Supplementary material from the paper’s appendix.

Appendix

Figure 6: Detailed ablation results showing RAG impact on thermodynamic stability (3.6× improvement), comparison of different prompt engineering strategies, and iterative refinement convergence over 5 cycles demonstrating plateau at cycle 4.

Configuration	Stability (%)	$\eta_{OER}$ (V)	Diversity	Time (h)
Full System	$82.4\pm 1.8$	$0.362\pm 0.015$	3.2	24
No RAG	$23.1\pm 4.2$	$0.521\pm 0.043$	4.1	18
No Iteration	$64.3\pm 3.1$	$0.412\pm 0.021$	3.0	5
Constraint Only	$68.2\pm 2.7$	$0.395\pm 0.018$	1.8	22
Analogy Only	$41.3\pm 3.9$	$0.438\pm 0.027$	3.5	21
Random Baseline	$3.2\pm 1.1$	$0.612\pm 0.071$	4.5	20

Appendix

Table 2: Full ablation study examining all component combinations. Each configuration tested with 200 generated candidates over 5 independent runs.

Parameter	Range Tested	Optimal	Impact
Temperature	0.1-1.0	0.7	Critical
Top-p	0.5-1.0	0.95	Moderate
k (retrieval)	5-50	20	High
Similarity threshold	0.7-0.95	0.85	Low
Beam width	1-10	5	Moderate
Iterations	1-10	5	High

Appendix

Table 3: Extended hyperparameter sensitivity analysis

Comparison	Cohen’s d	Interpretation
LLM vs IrO 2 baseline	2.31	Very large
LLM vs known catalysts	1.87	Large
With RAG vs without	3.42	Very large
Combined vs constraint-only prompts	1.42	Large
Combined vs analogy-only prompts	2.18	Very large

Appendix

Table 10

Source	Entries	Material Types	Avg. Elements
Materials Project	25,000	Bulk alloys, oxides	3.8
NOMAD	12,000	Alloys, surfaces	4.2
OC20	13,000	Surface+adsorbates	2.5
Total	50,000	—	3.6

Appendix

Table 4: Distribution of database entries across sources and material types

Figure 7: Complete correlation matrix showing relationships between all catalyst properties including overpotential, stability metrics, d-band center, and compositional features for the full set of LLM-generated catalysts.

Figure 8: 3D activity landscape of HEA catalysts showing the relationship between NOH adsorption energy ( $\Delta E_{NOH}$ ), mixing enthalpy, and limiting potential. The surface color represents catalytic activity, with dark purple regions indicating optimal performance. Black circles mark individual catalyst compositions, demonstrating clustering in the favorable low-potential region.

Figure 9: Statistical comparison of key properties across catalyst types. Box plots show mixing enthalpy distribution with LLM-HEAs exhibiting most negative values (median -0.8 eV/atom) indicating superior stability, and d-band center distribution with LLM-HEAs centered at -2.8 eV correlating with enhanced activity.

Figure 10: Property distributions for HEA catalysts showing mixing enthalpy right-skewed distribution (mean -0.593 eV/atom), multimodal d-band center distribution (mean -2.425 eV), broad $\Delta E_{NOH}$ distribution (mean 0.774 eV), and left-skewed limiting potential distribution with exceptional catalysts in the tail. Vertical lines indicate mean (red) and median (green) values.

Feature	$\eta_{OER}$	Stability	d-band	EN	Size
$\eta_{OER}$	1.00
Stability	-0.42**	1.00
d-band center	-0.73***	0.31*	1.00
Avg. EN	0.28*	-0.19	-0.35**	1.00
Size mismatch	0.15	-0.52***	-0.08	0.21	1.00
Fe content	-0.38**	0.27*	0.41**	-0.15	-0.03

Appendix

Table 5: Pearson correlations. *p<0.05, **p<0.01, ***p<0.001 after Bonferroni correction

Composition	Method	Conditions
Fe 0.2 Co 0.2 Ni 0.2 Ir 0.1 Ru 0.3	Arc melting	1800 ∘ C, Ar
Mn 0.15 Fe 0.25 Co 0.25 Ni 0.2 Pt 0.15	Sputtering	400 ∘ C, 5 mTorr
Cr 0.2 Fe 0.2 Co 0.3 Ni 0.2 Mo 0.1	Ball milling	500 rpm, 20h
V 0.1 Cr 0.2 Mn 0.2 Fe 0.25 Co 0.25	Carbothermal	2000 ∘ C flash

Appendix

Table 17

Explore similar work

Feb 25, 2026

Reasoning-Driven Design of Single Atom Catalysts via a Multi-Agent Large Language Model Framework

Large language models (LLMs) are becoming increasingly applied beyond natural language processing, demonstrating strong capabilities in complex scientific tasks that traditionally require human expertise. This progress has extended into materials discovery, where LLMs introduce a new paradigm by leveraging reasoning and in-context learning, capabilities absent from conventional machine learning approaches. Here, we present a Multi-Agent-based Electrocatalyst Search Through Reasoning and Optimization (MAESTRO) framework in which multiple LLMs with specialized roles collaboratively discover high-performance single atom catalysts for the oxygen reduction reaction. Within an autonomous design loop, agents iteratively reason, propose modifications, reflect on results and accumulate design history. Through in-context learning enabled by this iterative process, MAESTRO identified design principles not explicitly encoded in the LLMs' background knowledge and successfully discovered catalysts that break conventional scaling relations between reaction intermediates. These results highlight the potential of multi-agent LLM frameworks as a powerful strategy to generate chemical insight and discover promising catalysts.

Dong Hyeon Mok, Seoin Back, Victor Fung +1

Department of Chemical and Biomolecular Engineering, Institute of Emergent Materials, Sogang University, Seoul 04107, Republic of Korea · KU-KIST Graduate School of Converging Science and Technology, Korea University, Seoul 02841, Republic of Korea · Department of Integrative Energy Engineering, Korea University, Seoul 02841, Republic of Korea +4

Materials Science Large Language Model-Guided Scientific Discovery Multi-Agent LLMs

Ask PDF

Mar 1, 2026

Catalyst-Agent: Autonomous heterogeneous catalyst screening with an LLM Agent

The discovery of catalysts for electrochemical applications such as the oxygen reduction reaction (ORR), nitrogen reduction reaction (NRR), and CO2 reduction reaction (CO2RR) remains a central challenge in chemistry and materials science. Machine-learning interatomic potentials (MLIPs) and graph neural network models now accelerate individual adsorption-energy calculations by orders of magnitude relative to density functional theory. However, true large-scale screening is still blocked by human decisions: selecting candidates, constructing slabs, enumerating adsorption sites, interpreting descriptor failures, and choosing follow-up modifications. Here, we introduce Catalyst-Agent, a Model Context Protocol (MCP) server-based, LLM-powered agent that autonomously coordinates closed-loop catalyst screening. Catalyst-Agent searches materials databases through OPTIMADE, constructs slabs, computes adsorption energies using Meta FAIRchem's UMA MLIP within AdsorbML, evaluates reaction-specific descriptors, and applies structural modifications to refine near-miss candidates. In ORR, NRR, and CO2RR campaigns, Catalyst-Agent demonstrates high performance and converges in 1.40-3.41 trials per successful material on average. It identified Sn3Sc, Sn3Y, Tl3La, Pb3Y and In3Y as CO2RR candidates for further validation that were not previously reported in the literature. DFT single-point checks confirmed screening outcomes for representative NRR and CO2RR candidates. Ablations show these gains arise from chemically informed candidate selection and feedback-directed modification rather than brute-force evaluation: fully randomized screening dropped to 13.3%, 16.7%, and 0% success for ORR, NRR, and CO2RR, respectively. These results show that tool-grounded LLM agents can shift catalyst screening from manual trial-and-error toward more autonomous, reproducible and adaptive workflows.

Achuth Chandrasekhar, Janghoon Ock, Amir Barati Farimani

Mechanical Engineering, Carnegie Mellon University, Pittsburgh, PA 15213, USA · Department of Chemical and Biomolecular Engineering, University of Nebraska–Lincoln, Lincoln, NE 68588, USA · Biomedical Engineering, Carnegie Mellon University, Pittsburgh, PA 15213, USA +3

LLM Agent Orchestration Materials Science Large Language Model-Guided Scientific Discovery Autonomous Research Agents

Ask PDF

Jul 13, 2026

CatRetriever: Contrastive Representation Learning for Slab-to-Bulk Retrieval in Generative Catalyst Discovery

Inverse design is an emerging data-driven paradigm for efficiently navigating vast chemical spaces to discover new materials with targeted properties, and in the context of heterogeneous catalysis, surface generative models have recently advanced this goal by directly generating catalyst surface-adsorbate structures. However, these models typically operate at the slab level and do not provide the corresponding parent bulk structure, making it difficult to assess bulk-dependent properties such as formation energy, surface energy, crystallographic symmetry, and synthesizability. Here, we address this missing slab-to-bulk connection as a retrieval problem and introduce CatRetriever, a contrastive representation learning model that aligns slab and bulk crystal representations in a shared latent space. From a slab query, CatRetriever accurately retrieves plausible parent bulk candidates with R@1 > 91% and R@3 > 98% on both the in-distribution and holdout evaluation sets. We further extend the CatRetriever framework into an adsorption energy targeted bulk discovery pipeline that combines bulk retrieval, generative search space expansion, and adsorption energy distribution analysis. This workflow evaluates candidates by both structural compatibility with the query slab and their ability to access the target adsorption energy range across diverse surface environments. CatRetriever therefore provides a scalable route for connecting catalyst generative models with physically plausible and adsorption energy compatible bulk catalyst discovery.

Jungho Oh, Woosung Kim, Dong Hyeon Mok +2

KU-KIST Graduate School of Converging Science and Technology, Korea University, Seoul 02841, Republic of Korea · Department of Materials Science and Engineering, Korea University, Seoul 02841, Republic of Korea · Department of Chemical and Biomolecular Engineering, Sogang University, Seoul 04107, Republic of Korea +5

Contrastive Learning Materials Science Computational Materials Science

Ask PDF