Scanning Tunnelling Microscopes (STM) are capable of obtaining images of surfaces with atomic scale resolution. They accomplish this by scanning an atomically sharp probe across the surface of a sample while monitoring an electric current. The quality of STM images depends greatly on the exact geometry and composition at the apex of the scanning probe. Blunt tips result in blurry images, while contaminated tips can interact with the sample during a scan to produce noisy images.
To improve image quality, the probe can be conditioned via a process called ‘tip shaping’, which involves poking the apex of the probe into a clean metallic substrate to reshape it. This is a somewhat random process that is repeated until a tip capable of acquiring images of sufficient quality is achieved. Along with surveying samples to locate regions of interest, tip shaping is a necessary, time-consuming task performed by a trained human operator. The simple and iterative-nature of these tasks makes them ideal to control via simple algorithmic and machine-learning protocols. This project, dubbed ‘Scanbot’, automates probe conditioning and sample surveying, along with several other STM data acquisition techniques.
About the presenter
Julian Ceddia is a PhD student at Monash University with CI Agustin Schiffrin. His project aims to design electronically functional solid-state devices based on 2D organic materials, as part of FLEET’s Research Theme 1, Topological Materials.