Development and Implementation of the Optical Cataloger for Automatic 2D Crystallite Identification

Van der Waals crystals are a large class of materials characterized by strong in-plane bonding and weak out-of-plane bonding. This characteristic allows them to easily exfoliate from their bulk form to individual atomic layers. The resulting layers have surfaces free of dangling bonds, allowing for arbitrary re-stacking of any combination of materials and at any twist angles one desires. This presents a huge opportunity to design material stacks, called heterostructures, with tailored optical, electronic, and magnetic properties absent in the stack's individual layers. These heterostructures are already being investigated for potential applications such as next-generation field-effect transistors and single photon light sources for quantum information technologies, as well as physical quantum simulators of the Hubbard model. A major hurdle to investigating the potential applications and new physics found in these heterostructures, is the slow and laborious process by which they are constructed. One must first exfoliate all the materials necessary for the desired heterostructure, then optically identify crystallites of interest with a brightfield microscope, and finally the crystallites must be stacked on top of each other in the desired configuration. It is the second step in this process, the optical identification of crystallites, which is the most time consuming. In this work, we introduce an automated scanning brightfield microscope with integrated AI-based image segmentation algorithms and point spectroscopy capabilities, for the automatic identification of 2D crystallites based on the brightfield optical images and the confirmation of the layer number of the crystallites with point photoluminescence and Raman spectroscopy. Additionally, this instrument is designed to be integrated into a fully automated production pipeline for heterostructures, where two other robotic instruments handle the exfoliation and the stacking. This production pipeline will be housed in a glovebox with an inert gas atmosphere, allowing the processing of van der Waals materials that are air sensitive. The entire system has the potential to greatly accelerate the production of heterostructures, allowing researchers to spend more time characterizing and developing them into viable technologies.