Fine registration of SEM and AFM images using Monte Carlo simulations

For the characterization of objects on the micro-and nanoscale scanning electron microscopy (SEM) and atomic force microscopy (AFM) are common imaging techniques. Due to the different technical approaches of both methods they provide different information about the sample so a fusion of AFM and SEM images into a single image is beneficial. As SEM and AFM images show a number of individual effects like edge effects in SEM and tip convolution in AFM existing algorithms for multi-modal image registration can only yield a coarse transformation between an AFM and an SEM image of the same scene. Fine registration describes the process of finding an optimal transformation between images which therefore is a difficult task for AFM and SEM images. This paper shows a technique for enhancing AFM/SEM fine registration results using Monte Carlo simulations of SEM images. It is shown that based on an AFM image representing the height profile of a virtual material surface the output of an SEM simulation can be fine registered with the original SEM image to gain a clearly enhanced transformation between the AFM and SEM image.