Publications

This study utilizes 4D-STEM to reveal strain-dependent domain structures and unique stacking orders in 2D vdW ferroelectric SnSe.

This study utilizes STEM and STM to demonstrate robust supermoiré structures in large-angle single-twist bilayers, overcoming lattice disorder to realize tunable flat bands and strong correlations.

Using 4D-STEM and molecular dynamics, this study reveals that sharp corners in core-shell nanoparticles stabilize high surface strain by reducing shear stress, leading to superior durability and performance.

By combining statistical 4D-STEM analysis with in situ heating, this work elucidates the precise strain distributions in fivefold-twinned nanoparticles and captures the dynamic transition to a strain-relieved state via partial dislocation formation.

By employing machine learning-assisted operando 4D-STEM, this study resolves the complex grain boundary networks formed as copper nanowires evolve into active metallic nanograins.

This paper reports a semi-automated, unsupervised machine learning method for discovering complex material deformations in 4D-STEM data without requiring prior structural assumptions.

By integrating human-in-the-loop feedback with a multi-agent architecture, this work introduces an LLM-powered system that automates code generation and parameter tuning to achieve high-quality ptychographic imaging.