4D STEM is a start-of-the-art method for extracting structural information from complex samples such as novel semi-conductor devices, battery materials, carbon capture devices, and viruses. In 4D STEM, a detector captures a diffraction pattern at each probe position in the scan. As a result, the acquisition speed is limited by how quickly the detector can capture each frame, which is typically on the order of 10x to 100x slower than typical STEM imaging. This leads to smaller fields of view, increased beam damage potential, stage drift, and vast datasets.
In this tutorial, we will demonstrate a subsampled approach to 4D STEM from SenseAI, allowing real-time feedback during survey mode, as well as the results of typical 4D STEM analysis such as ptychography. By acquiring data in a subsampled regime, SenseAI can increase effective detector readout speeds by up to ten times, making dynamical 4D STEM a possibility for those even without the latest hardware.
