PVmatAgent: A Large Language Model (LLM) Agent for Perovskite Photovoltaic Material Design and Analysis

Abstract

The rapid advancement of large language models (LLMs) has opened new opportunities for materials informatics. However, LLMs fall short in photovoltaic (PV) material design due to their lack of domain grounding, unreliable outputs, and inability to perform integrated computational tasks. To address these issues, this work presents PVmatAgent, an autonomous LLM‐based computational agent designed specifically for PV materials design and analysis. The system integrates 11 domain‐specific tools (organized into 9 functional modules), including the machine learning force field CHGNet for geometric relaxation and formation energy calculation, the graph neural network MEGNet for bandgap prediction, the Goldschmidt tolerance factor and octahedral factor for perovskite stability screening, the photovoltaic performance evaluator including the Shockley–Queisser limit, the SLME model and tandem current matching module, the materials project agent for structure retrieval, and the retrieval‐augmented generation (RAG) knowledge base. A hallucination truncation mechanism prevents the model from fabricating numerical outputs. The system was validated on four representative scenarios. Results demonstrate that PVmatAgent effectively executes computational tasks and provides corrective recommendations grounded in literature evidence for photovoltaic material design. This work offers a practical paradigm for LLM‐based autonomous agents in AI‐driven optoelectronic materials discovery.