Exfoliation of large high-quality graphene coupled with physically informed automated identification

First isolated in 2004 with a piece of scotch-tape, graphene monolayers display unique properties and promising technological potential in next generation electronics, optoelectronics, and energy storage. This simple yet effective mechanical exfoliation technique has been applied to analogous materials to discover more two-dimensional (2D) atomic crystals which demonstrate distinct physical properties from their bulk counterpart, opening the new era of materials research. However, the difficulty in fabricating large flakes of high purity and the impractical manual inspection of optical images to identify 2D flakes, hinders practical commercial applications and fundamental research of these thin materials. Furthermore, despite the advancements brought by coupling deep learning algorithms with optical microscopy for automated flake identification, their high computational complexities, large dataset requirements, and more importantly, opaque decision-making processes limit their accessibilities. Therefore, as an alternative we have developed physically informed, transparent tree-based machine learning (ML) algorithms for the automated identification of exfoliated 2D atomic crystals under different optical settings. We then couple these successful ML methods with mechanical exfoliation followed by vacuum annealing of graphene to promote scalable fabrication of large flakes. We evaluate the purity of the flakes with Raman and atomic force microscopy.