Towards Real-Time Imaging of Atomic Vibrations with a Pixelated Detector

Aberration-corrected scanning transmission electron microscopy (STEM) is a potent technique for the direct observation of atomic structures and local material chemistry. Electrons scattered at high angles are primarily governed by thermal diffuse scattering, which depends on atomic displacements, particularly those due to atomic vibrations [1]. Therefore, quantitative measurement of atomic vibrational properties is achievable by analyzing detailed distributions of thermal diffuse scattering. Recently developed multi-dimensional detectors, such as segmented and pixelated detectors, can detect changes in scattering distribution attributable to atomic vibrations. Utilizing the segmented detector, we have acquired anisotropy of thermal diffuse scattering and observed anisotropic atomic vibrations of the clathrate compound Ba₈Ga₁₆Ge₃₀ [2]. In contrast, pixelated detectors, which can acquire more detailed distributions of thermal diffuse scattering, are anticipated to be powerful tools for analyzing physical properties related to atomic vibrations in more detail, such as localized phonons.<br/>Recent advancements in pixel detector speeds have underscored the critical need for in situ processing of live 4D data and the real-time rendering of processed images. In this study, we developed real-time imaging applications to process 4D data captured by the JEOL JEM-ARM300F, equipped with a DECTRIS ARINA detector [3], at speeds reaching up to 100,000 fps. Our application rapidly transfers the acquired raw data to high-performance GPUs, enabling live display of results through highly parallelized processing using CUDA. We explored the potential for the live visualization of atomic vibrations through the anisotropic scattering imaging using this application. The presentation will report the outcomes of adapting this application to atomic vibration observation and other live observation scenarios.