High-throughput determination of sample visibility in liquid-phase SEM via Monte Carlo simulations

Abstract

We developed a Monte Carlo simulation workflow to investigate the effects of sample thickness, membrane thickness, sample composition, and incident electron energy on the visibility of samples in liquid-phase scanning electron microscopy based on electron-transparent Si₃N₄ membranes. By using a thin wedge as the sample geometry and non-uniform spacing of the scan points, we avoided the need to generate numerous geometries for each configuration and reduced the computation time by up to 2 orders of magnitude. Quantitative analyses of the threshold current for visible contrast and the spatial resolution revealed that secondary electrons may visualize thin samples more effectively than BSEs at incident electron energies down to 3 keV when aided by efficient in-lens detectors and Si₃N₄ membranes with a thickness of 20 nm or lower. The simulations also supported the trends of decreasing spatial resolution with the thickness of both the membrane and the sample material.