A Hybrid Approach to Calibrate the Affine Transformation Between Scan-Camera Coordinates for 4D-STEM Datasets

Accurate geometrical calibration between scan-camera coordinates is critical in four-dimensional scanning transmission electron microscopy (4D-STEM) [1] for both quantitative imaging and ptychographic reconstructions. Such calibration is, however, not a trivial task since ( i ) 4D-STEM is more sensitive to the vortical nature of electron trajectories caused by the electromagnetic lenses compared to other STEM imaging modes that use detectors with circular symmetries; ( ii ) Most 4D-STEM experiments still need relatively long dwell time ( e.g., > 100μs per pixel) making them highly susceptible to instrumental instabilities and environmental disturbances. We argue that existing calibration methods are not completely satisfactory. For instance, imaging registration methods [2-7] that require multiple 4D-STEM frames could still be practically challenging for most instruments. A popular method by J. Hatchel et al. , [8,9] which utilizes the electric-field information from the 4D-STEM