Automating Convergence of STM Controls using Bayesian Optimization

Scanning Tunneling Microscopy (STM) is a widely used tool for atomic-scale imaging of novel materials. However, the tip optimization is a tedious process due to the extremely complex nature of the tip-surface interaction, and thus limits the throughput efficiency. Here we demonstrate a Machine Learning (ML) based framework to realize the automated optimization of the scan controls for high quality STM imaging of graphene. We deploy a Bayesian Optimization (BO) method on the STM controls in real-time to enhance the imaging quality, given by the peak intensity in the Fourier space. The BO prediction is dynamically incorporated into the microscopy controls, i.e., the current setpoint and the tip bias, to rapidly optimize the imaging conditions. We present strategies to either selectively explore or exploit across the parameter space. As a result, suitable policies are developed for autonomous convergence of the control-parameters. The ML-based framework serves as a general workflow methodology across a wide range of materials.