Towards AI-driven autonomous growth of 2D materials based on a graphene case study

The scalable synthesis of two-dimensional (2D) materials remains a key challenge for their integration into solid-state technology. While exfoliation techniques have driven much of the scientific progress, they are impractical for large-scale applications. Advances in artificial intelligence (AI) now offer new strategies for materials synthesis. This study explores the use of an artificial neural network (ANN) trained via evolutionary methods to optimize graphene growth. The ANN autonomously refines a time-dependent synthesis protocol without prior knowledge of effective recipes. The evaluation is based on Raman spectroscopy, where outcomes resembling monolayer graphene receive higher scores. This feedback mechanism enables iterative improvements in synthesis conditions, progressively enhancing sample quality. By integrating AI-driven optimization into material synthesis, this work contributes to the development of scalable approaches for 2D materials, demonstrating the potential of machine learning in guiding experimental processes. The authors propose an artificial neural network enabled to autonomously learn and refine time-dependent protocols for graphene growth, optimizing synthesis quality through iterative feedback based on Raman spectral analysis. This AI-driven methodology advances the development of large-scale and high-quality 2D-Materials.