Bayesian tomographic reconstruction for high angle annular dark field (HAADF) scanning transmission electron microscopy (STEM)

HAADF-STEM data is increasingly being used in the physical sciences to study materials in 3D because it is free from the diffraction effects seen in Bright Field STEM data and satisfies the projection requirement for tomography. Typically, reconstruction is performed using Filtered Back Projection (FBP) or the SIRT algorithm. In this paper, we develop a Bayesian reconstruction algorithm for HAADF-STEM tomography which models the image formation, the noise characteristics of the measurement, and the inherent smoothness in the object. Reconstructions of polystyrene functionalized Titanium dioxide nano particle assemblies show results that are qualitatively superior to FBP and SIRT reconstructions, significantly suppressing artifacts and enhancing contrast.