Multi-modal and multi-scale non-local means method to analyze spectroscopic datasets.

Area of Science:

• Materials Science
• Spectroscopy
• Microscopy

Background:

• Low signal-to-noise ratio (SNR) in transmission electron microscopy (TEM) datasets hinders atomic-scale analysis.
• Existing denoising techniques often compromise spatial resolution or fail to leverage multi-modal data.
• Simultaneous dark-field and spectroscopic data offer complementary information for improved analysis.

Purpose of the Study:

• To develop and validate a novel denoising method for enhancing atomically resolved spectroscopic maps from low-SNR TEM datasets.
• To improve the signal-to-noise ratio (SNR) of spectroscopic datasets without sacrificing spatial resolution.
• To demonstrate the method's capability in analyzing atomic-scale diffusion and surface oxidation states.

Main Methods:

• Proposed a multi-modal and multi-scale non-local means (M3S-NLM) algorithm.
• Integrated simultaneous dark-field signals with spectroscopic data (electron energy dispersive X-ray and electron-energy-loss spectroscopy - EELS).
• Applied the M3S-NLM method to Al_1-xIn_xN alloy and perovskite nanocatalyst datasets.

Main Results:

• Successfully extracted atomically resolved spectroscopic maps from low-SNR datasets.
• Demonstrated improved SNR in electron-energy-loss spectroscopy (EELS) datasets.
• Retrieved atomic-scale diffusion in Al_1-xIn_xN alloys and surface oxidation states of perovskite nanocatalysts.
• Obtained atomically resolved oxidation maps of LaMnO₃ nanoparticles using fine structure absorption edge analysis.

Conclusions:

• The M3S-NLM method effectively enhances spectroscopic datasets from TEM, enabling atomic-scale material characterization.
• The technique allows for the study of dynamic processes like diffusion and surface chemistry at the atomic level.
• This advancement provides a powerful tool for analyzing complex materials and nanomaterials with improved accuracy and resolution.