skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Attention:

The NSF Public Access Repository (PAR) system and access will be unavailable from 11:00 PM ET on Friday, May 16 until 2:00 AM ET on Saturday, May 17 due to maintenance. We apologize for the inconvenience.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


  • NSF Public Access
  • Search Results
  • Unsupervised machine learning for unbiased chemical classification in X-ray absorption spectroscopy and X-ray emission spectroscopy
Title: Unsupervised machine learning for unbiased chemical classification in X-ray absorption spectroscopy and X-ray emission spectroscopy
We report a comprehensive computational study of unsupervised machine learning for extraction of chemically relevant information in X-ray absorption near edge structure (XANES) and in valence-to-core X-ray emission spectra (VtC-XES) for classification of a broad ensemble of sulphorganic molecules. By progressively decreasing the constraining assumptions of the unsupervised machine learning algorithm, moving from principal component analysis (PCA) to a variational autoencoder (VAE) to t-distributed stochastic neighbour embedding (t-SNE), we find improved sensitivity to steadily more refined chemical information. Surprisingly, when embedding the ensemble of spectra in merely two dimensions, t-SNE distinguishes not just oxidation state and general sulphur bonding environment but also the aromaticity of the bonding radical group with 87% accuracy as well as identifying even finer details in electronic structure within aromatic or aliphatic sub-classes. We find that the chemical information in XANES and VtC-XES is very similar in character and content, although they unexpectedly have different sensitivity within a given molecular class. We also discuss likely benefits from further effort with unsupervised machine learning and from the interplay between supervised and unsupervised machine learning for X-ray spectroscopies. Our overall results, i.e. , the ability to reliably classify without user bias and to discover unexpected chemical signatures for XANES and VtC-XES, likely generalize to other systems as well as to other one-dimensional chemical spectroscopies.  more » « less
Award ID(s):
1633216 1904437
PAR ID:
10353651
Author(s) / Creator(s):
; ;
Date Published:
Journal Name:
Physical Chemistry Chemical Physics
Volume:
23
Issue:
41
ISSN:
1463-9076
Page Range / eLocation ID:
23586 to 23601
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; (, ECS Meeting Abstracts)
    Tracking the change in electronic structure of target elements is crucial to investigate the nature of redox reactions occurring in battery electrodes. X-ray emission spectroscopy (XES) and x-ray absorption fine structure (XAFS) perform this role well with high sensitivity and throughput, but the requisite of synchrotron facilities often limits those availability for material characterization. Using a lab-scale x-ray emission/absorption spectrometer, we investigated the changes in the local structure and chemistry around the 3d transition metal elements of LiMO 2 cathodes for Li-ion batteries as a function of the battery state of charge (SoC). Ex situ measurement was prepared from the electrode samples with discrete difference in SoC. Coupled with ex situ measurement, operando measurement was performed using pouch cells with LiMO 2 cathode, which enabled a real-time monitoring of chemical shift in an element-specific manner resulted from changing electrode potential. Through the XES mode of the bench-top spectrometer, fluorescence emissions from the LiMO 2 cathode, or the cell containing it, was monochromatized by a spherically bent crystal analyzer (SBCA). The Kβ emissions of 3d transition metal elements such as cobalt display position/shape difference of spectrum with varying SoC. The trend of chemical shift and change in spectral features provided the information on the electronic structure variation, such as oxidation state change of 3d transition metals in LiMO 2 during charge and discharge (i.e., delithiation and lithiation). Furthermore, valence-to-core (VtC) emission signals helped enable in-depth analysis such as spin structure characterization. In addition to the XES analysis, we could measure K-edge XAFS for the same 3d transition metals in LiMO 2 as well. In the XAFS mode of the spectrometer, SBCA monochromatized bremsstrahlung x-ray generated from a high-power x-ray tube is used to make an incident source energy-dispersive. While Kβ XES probed occupied levels, K-edge XAFS examined unoccupied levels providing comprehensive understanding on the change in electronic structure of 3d transition metals in LiMO 2 . Figure 1 
    more » « less
  2. ; ; ; ; (, Physical Chemistry Chemical Physics)
    Cation site occupation is an important determinant of materials properties, especially in a complex system with multiple cations such as in ternary spinels. Many methods for extracting the cation site information have been explored in the past, including analysis of spectra obtained through K-edge X-ray absorption spectroscopy (XAS). In this work, we measure the effectiveness of X-ray emission spectroscopy (XES) for determining the cation site occupation. As a test system we use spinel phase Co x Mn 3−x O 4 nanoparticles contaminated with CoO phases because Co and Mn can occupy all cation sites and the impurity simulates typical products of oxide syntheses. We take advantage of the spin and oxidation state sensitive Kβ 1,3 peak obtained using XES and demonstrate that XES is a powerful and reliable technique for determining site occupation in ternary spinel systems. Comparison between the extended X-ray absorption fine structure (EXAFS) and XES techniques reveals that XES provides not only the site occupation information as EXAFS, but also additional information on the oxidation states of the cations at each site. We show that the error for EXAFS can be as high as 35% which makes the results obtained ambiguous for certain stoichiometries, whereas for XES, the error determined is consistently smaller than 10%. Thus, we conclude that XES is a superior and a far more accurate method than XAS in extracting cation site occupation in spinel crystal structures. 
    more » « less
  3. ; ; (, X-Ray Spectrometry)
    Spherically‐bent crystal analyzers (SBCAs) see considerable use in very high‐resolution hard X‐ray wavelength dispersive X‐ray fluorescence spectroscopy, often called X‐ray emission spectroscopy (XES). While Si and Ge are the most frequently used diffractive components of SBCAs, we consider here the somewhat classical choice of muscovite mica as the dispersing element. We find that the various harmonics of a highest‐quality mica‐based SBCA show ~5–~40% of the integral reflectivity per unit solid angle of a typical Si or Ge SBCA in the hard X‐ray range, and that the mica SBCA have comparable energy resolution to the traditional SBCAs. Interestingly, the choice of mica comes with a practical benefit: the primary (0,0,2) reflection has sufficiently strong harmonics that are fairly tightly spaced in energy so that they span the complete energy range from ~4 to ~11 keV when used at convenient Bragg angles in a Rowland circle spectrometer. Hence, a single mica SBCA can be used for every K‐shell emission line of three dimensional transition metals and every L‐shell emission line of the lanthanide elements simply by selecting the correct mica (0,0,2) harmonic with a final energy‐dispersive solid state detector. The loss in efficiency is counteracted by an operational efficiency, i.e., the “universal” application of a single analyzer over a very large range of elements. This performance suggests future application of mica SBCAs in both laboratory‐based XES and synchrotron‐based photon‐in, photon‐out spectroscopies in the hard X‐ray range. 
    more » « less
  4. ; ; ; ; ; ; ; (, Physical Chemistry Chemical Physics)
    “Single-atom” catalysts (SACs) have demonstrated excellent activity and selectivity in challenging chemical transformations such as photocatalytic CO 2 reduction. For heterogeneous photocatalytic SAC systems, it is essential to obtain sufficient information of their structure at the atomic level in order to understand reaction mechanisms. In this work, a SAC was prepared by grafting a molecular cobalt catalyst on a light-absorbing carbon nitride surface. Due to the sensitivity of the X-ray absorption near edge structure (XANES) spectra to subtle variances in the Co SAC structure in reaction conditions, different machine learning (ML) methods, including principal component analysis, K-means clustering, and neural network (NN), were utilized for in situ Co XANES data analysis. As a result, we obtained quantitative structural information of the SAC nearest atomic environment, thereby extending the NN-XANES approach previously demonstrated for nanoparticles and size-selective clusters. 
    more » « less
  5. ; ; ; ; ; ; ; ; ; ; et al (, Scientific Data)
    Abstract The L-edge X-ray Absorption Near Edge Structure (XANES) is widely used in the characterization of transition metal compounds. Here, we report the development of a database of computed L-edge XANES using the multiple scattering theory-based FEFF9 code. The initial release of the database contains more than 140,000 L-edge spectra for more than 22,000 structures generated using a high-throughput computational workflow. The data is disseminated through the Materials Project and addresses a critical need for L-edge XANES spectra among the research community. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email