Highly sensitive 2D X-ray absorption spectroscopy via physics informed machine learning

Zeyuan Li,Thomas Flynn,Tongchao Liu,Sizhan Liu,Wah-Keat Lee,Ming Tang,Mingyuan Ge

DOI: https://doi.org/10.1038/s41524-024-01313-7

IF: 12.256

2024-06-19

npj Computational Materials

Abstract:Improving the spatial and spectral resolution of 2D X-ray near-edge absorption structure (XANES) has been a decade-long pursuit to probe local chemical reactions at the nanoscale. However, the poor signal-to-noise ratio in the measured images poses significant challenges in quantitative analysis, especially when the element of interest is at a low concentration. In this work, we developed a post-imaging processing method using deep neural network to reliably improve the signal-to-noise ratio in the XANES images. The proposed neural network model could be trained to adapt to new datasets by incorporating the physical features inherent in the latent space of the XANES images and self-supervised to detect new features in the images and achieve self-consistency. Two examples are presented in this work to illustrate the model's robustness in determining the valence states of Ni and Co in the LiNi x Mn y Co 1-x-y O 2 systems with high confidence.

materials science, multidisciplinary,chemistry, physical

What problem does this paper attempt to address?

The paper primarily aims to address the challenges encountered in two-dimensional X-ray absorption spectroscopy imaging (2D XANES) technology when probing localized chemical reactions at the nanoscale, particularly the quantitative analysis challenges posed by poor image signal-to-noise ratio when the concentration of the element of interest is low. Specifically, the research team developed a post-image processing method based on deep neural networks to reliably enhance the signal-to-noise ratio of 2D XANES images. The proposed neural network model can be trained by incorporating the physical characteristics in the latent space of XANES images and self-supervised to detect new features and achieve self-consistency. This method can significantly improve the accuracy of valence state determination for low-concentration elements such as nickel (Ni) and cobalt (Co). The paper provides two examples to demonstrate the model's ability to determine the valence states of Ni and Co in the LiNiₓMnyCo₁₋ₓ₋yO₂ system with high confidence. Additionally, the research shows how self-supervised learning can be utilized to improve the model's generalization capability, especially when dealing with datasets with specific background patterns. In summary, the paper aims to enhance the performance of 2D XANES imaging technology through deep learning techniques, particularly for the analysis of low-concentration elements, thereby improving the spatial resolution and spectral resolution in materials science research.

Highly sensitive 2D X-ray absorption spectroscopy via physics informed machine learning

Towards Large-Scale and Spatiotemporally Resolved Diagnosis of Electronic Density of States by Deep Learning

Leveraging Machine Learning for Advanced Nanoscale X-ray Analysis: Unmixing Multicomponent Signals and Enhancing Chemical Quantification

Robust Machine Learning Inference from X-ray Absorption Near Edge Spectra through Featurization

Classification of Local Chemical Environments from X-ray Absorption Spectra using Supervised Machine Learning

Machine-Learning X-Ray Absorption Spectra to Quantitative Accuracy

Machine learning spectral indicators of topology

High-sensitivity extreme-ultraviolet transient absorption spectroscopy enabled by machine learning

Deep neural network for x-ray photoelectron spectroscopy data analysis

Machine Learning Automated Approach for Enormous Synchrotron X-Ray Diffraction Data Interpretation

A graph neural network and transformer based approach to XANES data analyis

Determining electronic properties from L-edge X-ray absorption spectra of transition metal compounds with artificial neural networks

Identification of chemical compositions from “featureless” optical absorption spectra: Machine learning predictions and experimental validations

Decoding Structure-Spectrum Relationships with Physically Organized Latent Spaces

Deep Learning of Atomically Resolved Scanning Transmission Electron Microscopy Images: Chemical Identification and Tracking Local Transformations

Subspace modeling for fast and high-sensitivity X-ray chemical imaging

Analysis of extended X-ray absorption fine structure (EXAFS) data using artificial intelligence techniques

AI-NERD: Elucidation of relaxation dynamics beyond equilibrium through AI-informed X-ray photon correlation spectroscopy

A machine learning model for textured X-ray scattering and diffraction image denoising

Machine Learning Automated Analysis of Enormous Synchrotron X-ray Diffraction Datasets