Unsupervised machine learning for unbiased chemical classification in X-ray absorption spectroscopy and X-ray emission spectroscopy

Abstract

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)...

Informed Chemical Classification of Organophosphorus Compounds via Unsupervised Machine Learning of X-ray Absorption Spectroscopy and X-ray Emission Spectroscopy

“…The sulforganics study of Holden et al for the experimental VtC-XES and calculations showed excellent agreement, as did additional calculations and comparison to XANES in Tetef et al Here, in Figures S1 and S2, we more modestly validate the performance of against several VtC-XES taken with the same instrument and methodology as Holden et al, and also validate the performance against several XANES spectra from Persson et al…”

“…In our prior work on sulforganics and in the present above work on the more complex case of organophosphorus compounds, we have demonstrated a convincing utility of advanced, nonlinear unsupervised ML tools for evaluating the chemically relevant information in VtC-XES and XANES spectra. We now return to our hypothesis presented in the introduction and illustrated in Figure , where we propose that such an unsupervised ML method can productively inform the use of supervised ML tasks.…”

“…Additionally, the Stuttgart RLC ECP was substituted for atoms heavier than phosphorus. As in Tetef et al, a post-processing energy-dependent linear broadening scheme was applied to XANES transitions, starting with a full width half-maximum (FWHM) Lorentz broadening of 0.5 eV at the whiteline, and then linearly increasing to 4.0 eV FWHM at 20 eV past the whiteline. An energy shift of 50 eV was applied to all XANES transitions to align with experimental data …”

Computational approaches for XANES, VtC-XES, and RIXS using linear-response time-dependent density functional theory based methods

“…135,136 This procedure corrects for many-body and multichannel effects as it introduces orbital relaxation in the final valence-ionized states. This approach has been successfully explored by Govind and co-workers 84,102,[136][137][138][139] and recently in machine learning models to chemically classify sulphorganic 140 and organophosphorus compounds.…”

How Much Structural Information Could Be Extracted from XANES Spectra for Palladium Hydride and Carbide Nanoparticles

“…We investigate a possible dimensions reduction by the principal component analysis (PCA) and link the pure components to the structural descriptors. In the work of Tetef et al, 20 unsupervised ML was applied to a database of calculated molecules to investigate better the performance of Variational AutoEncoder, which is alternative to the PCA. However, this method involves a more complicate procedure of understanding parametric space and its correlation with structural descriptors.…”