Materials Optimization and Physics Discovery via Automated Scanning Probe Microscopy of Combinatorial Libraries

For over three decades, scanning probe microscopy (SPM) has been a pivotal method for exploring material structures and functionalities at nanometer and often atomic scales in various environments, including ambient, liquid, and vacuum. Historically, SPM applications have predominantly been downstream, with images serving as illustrations or sources of fundamental physical knowledge. However, the rapidly growing developments in rapid materials synthesis via self-driving labs and established applications such as combinatorial spread libraries are poised to change this paradigm. Rapid synthesis demands matching capabilities to probe materials on small scales and with speed, characteristics inherent to SPM. However, many SPM methods are not intrinsically quantitative, and all require constant monitoring and optimization by human operator.<br/><br/>In this presentation, I will illustrate the development of fully automated SPMs for probing combinatorial libraries of functional materials including photovoltage in hybrid perovskites, ferroelectricity in classical ferroelectrics, and topography in multicomponent alloys. I discuss the optimization of the full SPM discovery cycle, and introduce the machine learning methods necessary for the exploration of ternary and higher-dimensional systems. These include automated optimization of classical topographic imaging, Kelvin Probe Microscopy, and Piezoresponse Force Microscopy. I will further illustrate existing ML strategies for automated exploration of combinatorial libraries via automated large stage and probing structure-property relationships within the image. The corresponding software libraries are fully open. Finally, I will overview the coming challenges in the field. Overall, SPM will play a crucial role in closing the loop from materials prediction and synthesis to characterization.