Fast reconstruction of atomic-scale STEM-EELS images from sparse sampling.

Area of Science:

• Materials Science
• Microscopy
• Spectroscopy

Background:

• Image reconstruction is crucial for accelerating data acquisition in scanning transmission electron microscopy (STEM).
• Existing methods for 3D spectral image reconstruction in STEM electron energy loss spectroscopy (EELS) are often either inaccurate or computationally intensive.
• There is a need for reconstruction methods that balance speed and accuracy for atomic-scale EELS data.

Purpose of the Study:

• To develop and evaluate a novel, fast, and accurate reconstruction method for partially sampled spectrum-images in STEM-EELS.
• To address the limitations of existing reconstruction techniques in terms of accuracy and computational complexity.
• To compare the proposed method with established techniques like beta process factor analysis (BPFA) for STEM-EELS data.

Main Methods:

• A new reconstruction algorithm tailored for atomic-scale EELS data was developed.
• The proposed method was applied to partially sampled spectrum-images.
• Performance was evaluated against existing methods, including beta process factor analysis (BPFA), on both real and synthetic datasets.

Main Results:

• The proposed method demonstrates a favorable balance between reconstruction accuracy and computational efficiency.
• Comparison with BPFA shows competitive or superior performance for STEM-EELS data reconstruction.
• The method is suitable for accelerating the acquisition of atomic-scale EELS data without significant loss of information.

Conclusions:

• The developed reconstruction method offers a significant improvement for processing partially sampled STEM-EELS data.
• This approach enables faster data acquisition while maintaining high accuracy, crucial for atomic-scale analysis.
• The study validates the proposed method's effectiveness and potential for widespread adoption in EELS microscopy.