Exploring the use of large language models for EIS: A feasibility study on organic coatings for corrosion protection

Large Language Models (LLMs) were applied to automate the interpretation of electrochemical impedance spectroscopy (EIS) data, enabling classification and parameter estimation without the need for a task specific machine learning training. The approach achieved classification accuracies up to 96% and produced fitting results comparable to those obtained with specifically trained neural networks. The methodology reduces reliance on labelled data and manual intervention. While demonstrated in the context of organic coatings, the framework provides a scalable AI-based workflow that could, in principle, be extended to conceptually similar tasks in materials and corrosion research, subject to dedicated validation. • A structured methodology is proposed integrating Large Language Models (LLMs) with algorithmic logic to automate the interpretation of Electrochemical Impedance Spectroscopy, without the requirement of training dataset, increasing the accessibility and scalability. • The methodology enables the identification of the impedance response from mathematically different equivalent circuits using prompt-based and reference image comparisons and allows the estimation of electrical equivalent circuit components (fitting). • The proposed approach has been validated with simulated and experimental from organic coatings • The proposed approach has been benchmarked with traditional ML models used with the same datasets presented in this study. • This study demonstrates the practical potential of LLMs for fast, reliable, and expert-guided corrosion assessment.