Dynamic structure of active sites in ceria-supported Pt catalysts for the water gas shift reaction

Oxide-supported noble metal catalysts have been extensively studied for decades for the water gas shift (WGS) reaction, a catalytic transformation central to a host of large volume processes that variously utilize or produce hydrogen. There remains considerable uncertainty as to how the specific features of the active metal-support interfacial bonding—perhaps most importantly the temporal dynamic changes occurring therein—serve to enable high activity and selectivity. Here we report the dynamic characteristics of a Pt/CeO2 system at the atomic level for the WGS reaction and specifically reveal the synergistic effects of metal-support bonding at the perimeter region. We find that the perimeter Pt0 − O vacancy−Ce3+ sites are formed in the active structure, transformed at working temperatures and their appearance regulates the adsorbate behaviors. We find that the dynamic nature of this site is a key mechanistic step for the WGS reaction.


Supplementary Figures
Supplementary Figure 1. Comparison of Pt L3 edge (a) normalized XANES, (b) k 2 weighted χ(k), and (c) Fourier transformed k 2 χ(k) spectra of as-prepared Pt/CeO2 (this work) and ceria supported Pt single atoms. 1 Supplementary Figure 2. The histogram of particle size measured using scanning transmission electron microscopyannular dark field (STEM-ADF) images, for observed clusters (250 counts) in the reacted Pt/CeO2 catalyst. The average particle size is about 1.7 nm. The standard deviation is 0.3 nm. Figure 3. The temperaturedependent CO -DRIFTS spectra of Pt supported on ceria. The comparison of CO -DRIFTS spectra collected in the ramp-up and ramp-down process at 26 °C, 120 °C, 180 °C, and 240 °C (left to right). The spectroscopic features (peak position and shape) for species above 180 °C are almost identical between heat-up and cool-down. Similarly, samples at 120 °C and 26 °C during cool down share similar, narrow lineshapes but spectra during ramp-up at 26°C and 120°C are quite different. The as-prepared Pt/CeO2 at 26°C has a narrow peak centered at about 2095 cm -1 , assigned to CO adsorbed on single ionic Pt 2+ sites. 2 The presence of the sites is confirmed by STEM and XAS (see main text). After heating and cooling, the adsorbed species at 26 °C has a redshifted peak centered at 2072 cm -1 , implying CO is adsorbed to a Pt nanocluster. 3 Upon reaching 120 °C during the ramp-up process, the DRIFTS spectrum broadens with several features appearing between 2200-1900 cm -1 , suggesting the coexistence of ionic and metallic Pt species on the surface. This mixture of states is not observed at temperatures above 120 °C nor at 120 °C during ramp-down. Instead, a single, narrow cooling peak with the maximum at about 2069 cm -1 and with the tail at the low wavenumber side is observed at 120 °C during cooling. The as-prepared, supported Pt single atoms are not stable upon heating up to 180 °C and they aggregate into nanoclusters irreversibly at elevated temperatures. As deduced from the adsorbate vibrational modes, the following valence specific associations can be made. The as-prepared catalysts are exclusively ionic Pt single atoms at room temperature. Upon heating to 120 °C, a mixture of ionic and metallic Pt species coexists, while only metallic Pt species persist at T ≥ 180 °C. Metallic Pt remains the only predominant species present upon ramp-down to room temperature. These observations corroborate those from the activity test, STEM, and XAS evaluation ( Supplementary Fig. 4).  Figures (a2 -a5) and (b2 -b5). Figures (a6/b6) show the time-averaged image over the entire 6 second movie for each condition. The same Pt nanoparticle, which is different from the particle shown in the main text, is shown in all frames. The behavior shown in both figures is representative of many other particles imaged under the same conditions. corresponding fitting spectra (red) for the ceria supported Pt cluster under the WGS reaction conditions at different temperatures. For Pt 4f XPS spectra, the data were fitted by Pt 2+ (blue) and Pt 0 (green) components. For Pt 2+ , the experimentally observed binding energy of the Pt-4f7/2 core level is 72.8 eV and for Pt 0 , is 71.5 eV. For Ce 3d XPS spectra, the data were fitted by Ce 4+ (blue) and Ce 3+ (green) components. For Ce 4+ , the experimentally observed binding energy of the Ce-3d5/2 core level is 882.6 eV and of the Ce-3d3/2 core level is 901.2 eV. For Ce 3+ , the experimentally observed binding energy of the Ce-3d5/2 core level is 880.6 eV and of the Ce-3d3/2 core level is 899.2 eV. Figure 8. The percentage of interfacial and peripheral boundary atoms in the hemispherical cuboctahedron with the size less than 12 nm. For regular polyhedra, the number of interfacial and perimeter atoms could be analytically expressed as the function of the cluster order L. The cluster order L is defined as the number of spacings between adjacent atoms along the edge of the cluster. 4 Assuming the hemispherical cuboctahedron geometry of supported clusters, the number of interfacial atoms in the cluster is Ni=3L 2 +3L+1, the number of perimeter atoms of the interface is Np=6L, and the total number of atoms in the cluster is N=5L 3 /3+4L 2 +10L/3+1. For the hemispherical cuboctahedron geometry, the particle size could be estimated by using D=2Lr (D is the diameter of the cluster, r is metal-metal bond distance and for Pt-Pt distance, is ≈2.7Å). Then the relationship between the percentage of interfacial/perimeter sites in the particle and the particle site could be made via the cluster order L. Assuming a hemispherical cuboctahedron geometry for a supported nanoparticle, in a 1.7 nm nanocluster, there are approximately 40% atoms at the interface, accounting for the fraction of Pt 2+ detected by XPS at and below 200 °C. Figure 9. The comparison between fitted and experimental spectrum for the XAS data collected at 200°C, 250°C, 300°C and their corresponding post room temperatures under WGS condition. EXAFS data analysis was performed using IFEFFIT package. 5 The amplitude reduction factor S 0 2 was obtained by fitting the spectrum of Pt foil measured at the same beamline.

Supplementary
The value of S 0 2 (0.83) was used for all subsequent fitting the spectra of Pt/CeO2. In fitting the spectra of Pt/CeO2, two nearest-neighboring photoelectron paths chosen in the fitting model are Pt-O and Pt-Pt. For all data, the fitting k range is 2.8 Å -1 to 14.0 Å -1 and the fitting R range is 1.6 Å to 3.2 Å. The best fitting results are summarized in Supplementary Table 2.
Supplementary Figure 10. The DRIFTS spectra at different temperatures (ramp-down process) under WGS and CO conditions for Pt/CeO2 catalyst. The assignments to the peaks are based on the literature. [6][7][8][9] We note here that the existence of carbon-containing intermediates (formate, carbonate, carboxylic acid) suggests that the associative mechanism may also play role. However, to identify or discriminate active/spectator species in the WGS reaction, advanced transient isotopic experiments are needed.
Supplementary Figure 11. Activity test using the DRIFTS reactor: the H2 and CO2 productions at different temperatures in the ramp up and down processes. Similar trend was observed using the Clausen cell ( Fig. 1(a)).

Supplementary Tables
Supplementary Table 1