Convolutional Neural Network for Graphene Detection using Automated Optical Microscope

Graphene has been used extensively as a platform for quantum physics research and for quantum-based electronic and optical applications. Graphene has unique electronic and optical properties which are attractive to researchers in various fields. However, graphene production is incredibly expensive, slow, and labor-intensive. It involves researchers searching for graphene flakes with the correct thickness under an opt microscope and finding the best geometry. In this talk, we present the design and implementation of a quicker cost-effective way of detecting graphene samples. Our approach to identifying samples is to apply a Convolutional neural networks algorithm to rapidly detect monolayer graphene. Currently, we have made great progress in automating the microscope and camera with stepper motors and creating a graphical interface to communicate with the system. Future progress entails training the convolutional neural network algorithm, with data from motors, and camera. This project aims to reduce the production costs and time of graphene by efficiently detecting usable samples.