An Artificial Intelligence (AI) workflow for catalyst design and optimization

In the pursuit of novel catalyst development to address pressing environmental concerns and energy demand, conventional design and optimization methods often fall short due to the complexity and vastness of the catalyst parameter space. The advent of Machine Learning (ML) has ushered in a new era in the field of catalyst optimization, offering potential solutions to the shortcomings of traditional techniques. However, existing methods fail to effectively harness the wealth of information contained within the burgeoning body of scientific literature on catalyst synthesis. To address this gap, this study proposes an innovative Artificial Intelligence (AI) workflow that integrates Large Language Models (LLMs), Bayesian optimization, and an active learning loop to expedite and enhance catalyst optimization. Our methodology combines advanced language understanding with robust optimization strategies, effectively translating knowledge extracted from diverse literature into actionable parameters for practical experimentation and optimization. In this article, we demonstrate the application of this AI workflow in the optimization of catalyst synthesis for ammonia production. The results underscore the workflow's ability to streamline the catalyst development process, offering a swift, resource-efficient, and highprecision alternative to conventional methods. Keywords: Catalysts; Large Language Models; Active Learning; Bayesian Optimization; Ammonia Synthesis 1. Introduction The development of novel catalysts to address increasing energy demand and consumption has become an urgent task in the realm of renewable energy This surge is driven not only by escalating demands from applications in process optimization, yield improvement, and energy saving but also by a heightened awareness and concern for environmental issues, particularly the increase in carbon dioxide emissions. Several optimization strategies are conventionally employed to identify the optimal set of condition parameters, thereby enhancing the performance of the catalyst. The'One Factor At a Time' (OFAT) method is frequently employed as an alternative technique for chemical process optimization and comprehension While these conventional optimization methods and their advancements have undeniably made significant contributions to the field, certain gaps persist that limit their full potential in optimizing catalyst synthesis. The predominant reliance on the empirical knowledge and intuition of seasoned chemists, while invaluable, is not systematically scalable and transferable. Techniques like OFAT and DoE, though statistically rigorous, are often unable to keep pace with the sheer complexity and vastness of the catalyst parameter space, leaving much of it unexplored and underutilized.