Multimodal Imaging of Semiconductor Nanocrystal Deformation Trajectories at Atomic Resolution with Liquid-Phase Transmission Electron Microscopy

Semiconductor nanocrystals are inherently less stable than bulk materials due to their high surface-to-volume ratio, making them susceptible to decomposition under external stimuli. Understanding the degradation mechanisms of semiconductor nanocrystals, particularly the pathways of structural transformation during deformation, is key to developing nanocrystals and devices with enhanced stability and robust structures. Here, we combine the real-space liquid-phase transmission electron microscopy (TEM) imaging and reciprocal-space 4D-STEM to look at the morphological and structural evolution of II-VI and III-V based two-dimensional (2D) semiconductor nanoplates. The in-situ observation shows sequential removal of atomic layers following layer-by-layer etching at atomic resolution during the etching of wurtzite InP (w-InP). By controlling etching rate, morphologies with different surface roughness could be obtained, which is rationalized by Monte Carlo (MC) simulations. Meanwhile the customized machine-learning (ML) based data-mining of the 4D-STEM datasets indicates the formation of an amorphous layer during the etching process. In addition, by controlling the etching rate, we captured the evolution of strain in CdS nanoplates forming structural-dependent patterns, demonstrating the strain engineering in semiconductor nanocrystals by selective etching. Our observations suggest the structure and morphology of QDs or semiconductor nanocrystals could be controlled via external etching, shedding light on defect control and strain engineering in semiconductor nanocrystals and quantum dots for optoelectronic properties and devices.