| Abstract Scope |
We present an integrated 4D scanning transmission electron microscopy (4D-STEM) and machine learning framework for quantitative mapping of nanoscale structure and dynamics in emerging electronic materials. The approach leverages nanodiffraction datasets to extract local symmetry, medium-range order, and time-dependent correlations, while neural network models enable classification and reconstruction of atomic motifs without reliance on prior crystallographic assumptions. This methodology is applied across a range of systems, including ultra wide bandgap semiconductors, two dimensional materials, amorphous thin films, and hybrid interfaces relevant to microelectronics. Representative problems include defect and dislocation characterization, interface driven charge noise in quantum devices, and structure–property correlations in ALD grown materials. The combined framework provides statistically robust descriptors that connect local structural heterogeneity to functional performance, offering a generalizable route for characterization and discovery in next generation electronic nanomaterials. |