A Transformer-based Model for Rapid Microstructure Inference from Four-Dimensional Scanning Transmission Electron Microscopy Data

TL;DR

This paper introduces a transformer-based machine learning framework that rapidly infers microstructural features from 4D-STEM data, significantly speeding up the characterization process for crystalline materials and aiding materials design.

Contribution

The novel framework combines transformers with 4D-STEM data to achieve fast, high-resolution microstructure inference, outperforming traditional template-matching methods.

Findings

Inferred crystallographic orientations up to 100 times faster.

Enabled high-throughput microstructure characterization.

Facilitated structure-property relationship studies.

Abstract

Properties of crystalline materials are closely linked to microstructure arising from the spatial arrangement, orientation, and phase of nanocrystals. Rapid characterization of crystalline microstructure can accelerate the identification of these links and the development of materials with desired properties. Here, we combine a machine learning framework with four-dimensional scanning transmission electron microscopy (4D-STEM) to enable fast inference of crystalline microstructure over large fields of view. The framework employs a transformer-based architecture to predict crystallographic orientations and phases from 4D-STEM diffraction patterns, yielding spatially resolved maps of microstructural features at the nanoscale. With this framework, crystallographic orientations are inferred up to two orders of magnitude faster than widely used correlative template-matching approaches. This capability enables high-throughput characterization of complex crystalline materials and facilitates the establishment of structure-property relationships central to materials design and optimization.