Spatial and spectral dynamics in STEM hyperspectral imaging using random scan patterns

Area of Science:

• Materials Science
• Electron Microscopy
• Spectroscopy

Background:

• Scanning transmission electron microscopy (STEM) traditionally uses raster scan modes.
• Acquisition speed and electron dose effects are key limitations in STEM.
• Arbitrary scan pathways are being explored to overcome these limitations.

Purpose of the Study:

• To implement and demonstrate a hardware-level random scan operating mode in STEM.
• To evaluate the benefits of random scanning for image acquisition speed and dose reduction.
• To showcase the application of random scanning in spectro-microscopy for enhanced data analysis.

Main Methods:

• Implementation of a custom scan control module for hardware-level random scanning in STEM.
• Utilizing a pre-defined, shuffled raster pattern to sample regions of interest.
• Application of image reconstruction techniques to subsampled random sparse images.
• Demonstration using electron energy loss spectroscopy (EELS) and cathodoluminescence (CL) spectrum imaging.

Main Results:

• Random scan mode effectively reduces electron dose accumulation compared to conventional raster scanning.
• Decoupling of spatial and temporal information in hyperspectral images is achieved.
• Atomically-resolved elemental maps and nanoscale CL spectrum images were successfully generated.
• Precise tracking and correction of sample instabilities and spectral diffusion were demonstrated.

Conclusions:

• The random scan operating mode in STEM offers a flexible approach to enhance acquisition speed and minimize dose effects.
• This method enables advanced analysis of hyperspectral data, including tracking sample dynamics.
• Random scanning provides a powerful tool for spectro-microscopy applications, yielding high-resolution elemental and spectral information.