Superior Energy Storage Performance in High-Entropy Dielectric Ceramics Discovered by AI Materials Scientist

Keywords: AI Materials Scientist; Agent; Energy storage ceramics; Autonomous experimentation platform

Abstract: The design of advanced energy storage materials is hindered by vast and complex compositional spaces that are intractable for traditional trial-and-error methodologies. High-entropy ceramics (HECs) represent a promising class of dielectrics, but their multi-element nature exponentially expands this search space. To address this challenge, we deployed an "AI Materials Scientist"---an autonomous research agent---to navigate the high-dimensional landscape of lead-free perovskite HECs. The AI agent successfully identified a novel, non-intuitive five-component composition: 0.36 BaTiO3 - 0.32 BiFeO3 - 0.09 Bi0.5Na0.5TiO3 - 0.19 CaZrO3 - 0.04 Sr0.7La0.2TiO3. Experimental synthesis and characterization validated the AI's prediction, revealing a phase-pure ceramic with a dense, fine-grained microstructure. This material exhibits a breakthrough recoverable energy density (Wrec) of 10 J/cm3 and a high energy efficiency (η) of 80% at a breakdown strength of approximately 850 kV/cm, outperforming most existing lead-free dielectric ceramics. This work not only introduces a state-of-the-art energy storage material but also demonstrates the transformative potential of AI-driven autonomous systems to accelerate the discovery of complex, high-performance materials.

Submission Number: 312