Scanning Tunneling Microscope Control: A Self-Tuning PI Controller Based on Online Local Barrier Height Estimation

We identify the dynamics of a scanning tunneling microscope (STM) in closed loop and show that the plant dc gain is proportional to the square root of local barrier height (LBH), a quantum mechanical property of the sample and/or tip that affects the tunneling current. We demonstrate that during a scan, the LBH may undergo significant variations and this can adversely affect the closed-loop stability if the controller parameters remain fixed. Feedback instabilities increase the risk of tip-sample crash in STMs. In order to improve the closed-loop performance, we estimate the LBH, on the fly, and use that to adaptively tune the proportional-integral (PI) controller parameters. Experimental results obtained with the self-tuning PI controller confirm the improved STM performance compared to the conventional fixed-gain PI controller. Additional experiments confirm effectiveness of the proposed method in extending the tip lifetime by lowering the chance of a tip/sample crash. ©2018 IEEE