Improving Image Quality and Reducing Drift Problems via Automated Data Acquisition and Averaging in Cs-corrected TEM

Image acquisition with a CCD camera is a single-press-button activity: after selecting exposure time and adjusting illumination, a button is pressed and the acquired image is perceived as the final, unmodified proof of what was seen in the microscope. Thus it is generally assumed that the image processing steps of e.g., 'dark-current correction' and 'gain normalization' do not alter the information content of the image, but rather eliminate unwanted artifacts. Image quality therefore is, among a long list of other parameters, defined by the dynamic range of the CCD camera as well as the maximum allowable exposure time depending on sample drift (ignoring sample damage). Despite the fact that most microscopists are satisfied with present, standard image quality we found that it is a relatively easy to improve on existing routines in at least two aspects: (1) Suppression of lateral image drift during acquisition by using significantly shorter exposure times with a plurality of exposures (3D-data set); and (2) Improvement in the Signal/Noise ratio by averaging over a given data set by exceeding the dynamic range of the camera.