Probing Active Sites in Cu_xPd_y Cluster Catalysts by Machine-Learning-Assisted X-ray Absorption Spectroscopy

Abstract

Size-selected clusters are important model catalysts because of their narrow size and compositional distributions, as well as enhanced activity and selectivity in many reactions. Still, their structure–activity relationships are, in general, elusive. The main reason is the difficulty in identifying and quantitatively characterizing the catalytic active site in the clusters when it is confined within subnanometric dimensions and under the continuous structural changes the clusters can undergo in reaction conditions. Using machine learning approaches for analysis of the operando X-ray absorption near-edge structure spectra, we obtained accurate speciation of the Cu x Pd y cluster types during the propane oxidation reaction and the structural information about each type. As a result, we elucidated the information about active species and relative roles of Cu and Pd in the clusters.

Solving the structure of “single-atom” catalysts using machine learning – assisted XANES analysis

“…The first challenge could be solved if spectra that correspond to unique structures could be isolated. To solve the first part, several approaches have been proposed, ranging from forward modelling 38,39 to spectral deconvolution methods, such as principal component analysis (PCA), 40 linear combination analysis, 41 and multivariate curve resolution-alternating least square (MCR-ALS). 40 The latter approaches are particularly attractive if a series of spectra are available, such as when the sample is exposed to reaction conditions, where the speciation of different structures of catalytically active metal complexes changes with the reaction.…”

“…To solve the first part, several approaches have been proposed, ranging from forward modelling 38,39 to spectral deconvolution methods, such as principal component analysis (PCA), 40 linear combination analysis, 41 and multivariate curve resolution-alternating least square (MCR-ALS). 40 The latter approaches are particularly attractive if a series of spectra are available, such as when the sample is exposed to reaction conditions, where the speciation of different structures of catalytically active metal complexes changes with the reaction. In that case, a sequence of XANES spectra collected during the reaction can be interpreted as a combination of spectra coming from unique species with unique local geometry.…”

“…43–46 After training the model with theoretical spectra, the NN can be used to make predictions on the experimental data to extract unknown structural descriptors. 47,48 ML-assisted XANES analysis has been applied to mono- and bimetallic nanoparticles, 34,43,49 size-selected clusters, 50 metal clusters, 40 and metal oxide clusters, 51 and SACs. 52 ML-based extended X-ray absorption fine structure (EXAFS) analysis of SACs was reported recently.…”

Interpreting the Operando X-ray Absorption Near-Edge Structure of Supported Cu and CuPd Clusters in Conditions of Oxidative Dehydrogenation of Propane: Dynamic Changes in Composition and Size

“…41 The experimental spectra referenced have been previously published. 28 LCF was performed for energies between -20 and 20 eV around the white line peak.…”

“…The spectra for the Cu 3 Pd/Al 2 O 3 and Cu 4 /Al 2 O 3 clusters have been published previously, 15,28 and are reproduced with permission in Figure 1b. Our approach towards interpreting these spectra differs from that of Ref 29, whose principle component analysis approach for determining the primary components of the experimental spectra involves no inherent chemical meaning.…”