Synthesis of CeO2 nanoshapesCeO2 nanoparticles with rodlike, cubic and octahedral morphologies were synthesized by a hydrothermal method that has been reported in a previous study22. For CeO2 nanorods and nanocubes, 4.80 g sodium hydroxide (NaOH, Sigma-Aldrich) was dissolved in 18 mL deionized (DI) water. After stirring for 30 min, 2 mL solution containing 0.434 g cerium nitrate hexahydrate (Ce(NO3)3·6H2O, Sigma-Aldrich) was added dropwise into the stirring NaOH solution. The resulting slurry was transferred into a 50 mL Teflon-lined stainless-steel autoclave after an additional 30 min stirring. After hydrothermal treatment at 90 and 180 °C for 24 h, respectively, CeO2 nanorods and nanocubes are obtained. For the synthesis of CeO2 octahedra, 0.008 g NaOH and 0.434 g Ce(NO3)3·6H2O were used, while the amount of DI water remained the same as for the synthesis of nanorods and nanocubes. The hydrothermal treatment temperature and time were 180 °C and 24 h. The resulting precipitates were collected by centrifugation, washed with DI water, and dried in a vacuum oven at 80 °C overnight. Finally, the dried powder samples were calcined in a tube furnace at 400 °C for 4 h with a 20 mL min−1 flow of air and ramping rate of 1 °C min−1.Synthesis of Pt/CeO2 catalystsTo obtain Pt/CeO2 catalysts, a chemical reduction method was used. Briefly, 0.10 mL chloroplatinic acid hexahydrate solution (H2PtCl6·6H2O, Sigma-Aldrich, 0.05 g/mL) was diluted by 20 mL DI water. After stirring for 30 min, 0.2 g CeO2 nanorods powder was added into the solution, giving a Pt loading of ~1 wt%. After stirring for another 2 h, 20 mL solution containing 0.03 mL hydrazine monohydrate (NH2NH2·H2O, Sigma-Aldrich) was slowly added into the mixture at room temperature. The resulting Pt/CeO2 nanorods catalyst was collected by centrifuge after stirring for 12 h, washed thoroughly with DI water, and dried at room temperature for 24 h. For the Pt/CeO2 nanocubes and octahedra, the synthetic procedure was the same as Pt/CeO2 nanorods. The obtained samples are labeled as Pt/CeO2-rod, Pt/CeO2-cube, and Pt/CeO2-oct, referring to the different supports.CharacterizationWide-angle X-ray diffraction (XRD) on a Rigaku SmartLab Universal Diffractometer at the Center for Functional Nanomaterials at Brookhaven National Laboratory was used to investigate the structure of Pt/CeO2 catalysts.Ambient-pressure X-ray photoelectron spectroscopy (AP-XPS) was performed at the 23-ID-2 beamline (IOS) of the National Synchrotron Light Source II (NSLS-II)47. Photon energies of 1100 and 320 eV was used. The main chamber (base pressure 2.0 × 10−9 Torr) of the end-station was equipped with a differentially pumped hemispherical analyzer (Specs Phoibos 150 NAP). The Ce 3d photoemission line with the strongest Ce4+ feature at 916.9 eV was used for energy calibration of the AP-XPS signals. For sample pretreatment, 0.40 Torr oxygen (O2) was introduced, and the sample was heated to 400 °C for 30 min to remove any surface-bound carbon species. After O2 was pumped out, 0.40 Torr hydrogen (H2) was introduced to pre-reduce the sample at 300 °C for 30 min when the pressure reached ~5.0 × 10−8 Torr. For the WGSR, 0.25 Torr CO was introduced after the pressure of AP-XPS chamber reached ~5.0 × 10−8 Torr, and the sample was heated to 300 °C for 60 min, followed by introducing 0.50 Torr H2O for another 60 min. During all the experiments, a mass spectrometer was used to monitor the gas compositions in the main chamber. Peak fitting of XPS data was performed using CasaXPS peak fitting software, and fitting parameters are summarized in the Supplementary Tables 5–7.In situ transmission infrared spectroscopy (TIR) was performed on a Bruker Vertex 80 V spectrometer. Powder sample was pressed on to a tungsten mesh (75% transparency, GoodFellow) to form a thin film. The samples were in situ pretreated at 400 °C under 0.40 Torr O2 for 30 min, followed by pretreatment at 300 °C under 0.40 Torr H2 for 30 min. The system was pumped down to vacuum condition (<1 × 10−5 Torr) before introducing gas. After pretreatment, a background spectrum of sample at 300 °C was collected by averaging 512 scans at 4 cm−1 resolution between 4000 and 800 cm−1. Then, a set of spectra as function of time were collected after 0.05 Torr CO was introduced into the chamber, followed by 0.10 Torr H2O was added into the system. The sample temperature was maintained at 300 °C during the data collection. To track the changes of IR spectra, the peak areas of different species were integrated for comparison. The peak area integration performed at 3703–3630, 2450–2274, 2244–2045, and 2044–1846 cm−1 correspond to OH group, gas phase CO2, gas phase CO and CO adsorption on oxidized Pt species (Pt – O), and CO adsorption on metallic Pt species (Pt0), respectively. The normalization of peak areas is obtained by comparing the peak area different species obtained at different time with the peak area obtained at 3 min. Therefore, for all the formed new species in the IR spectra, the normalized peak area is 1 at 3 min and it increases with time. For comparison, the consumed OH group in the presence of CO and during the WGSR, the negative value was used.In situ environmental transmission electron microscopy (ETEM) was conducted on a FEI Titan 80-300 environmental transmission electron microscope. The samples were prepared by dispersing Pt/CeO2 powder in water, followed by deposition onto a nano-chip with through hole windows, which was then loaded into s DENSsolution Wildfire heating holder. After loading the sample into the ETEM, 0.40 Torr O2 was introduced into the chamber and the catalyst was pretreated in situ at 400 °C for 30 min. The catalyst was subsequently cooled to 300 °C and the O2 gas was pumped out. To minimize electron beam effects, the catalyst was imaged after pretreatment. Then, 0.40 Torr H2 was introduced into the chamber and the catalyst was reduced in situ at 300 °C for 30 min. After pumping out gas, the catalyst was imaged at 300 °C under vacuum conditions. To produce a water-gas-shift reaction conditions, 0.10 Torr CO and 0.20 Torr H2O were introduced to the chamber. The evolution of structural reconstruction was recorded and imaged under 300 °C in the presence of reactants.Catalytic activity testAll the catalysts were evaluated in a flow reactor. 50 mg of sample was loaded into a quartz tube with an inner diameter of 4.0 mm. A stream of 50 mL min−1 O2 gas was passed over the sample, which was then heated to 400 °C. After 60 min, a stream of 15 vol% H2/Ar was introduced into the reactor with a flow rate of 100 mL min−1. The catalyst was reduced by H2 at 300 °C for 60 min. After pretreatment, the catalyst was cooled down to room temperature under 50 mL min-1 Ar flow. Then, a stream of 1 vol% CO and 2 vol% H2O balanced with Ar were introduced into the reactor with a total flow rate of 157 mL min−1, given a space velocity of 188,400 ml gcat−1 h−1. The temperature was gradually increased to 300 °C. At each temperature, the system was stabilized for 30 min to inject the gas into an Agilent 7890B gas-chromatography coupled with mass spectrometer (GC-MS) system equipped with a thermal conductivity detector. The CO conversion was calculated based on the concentration of CO and CO2 in the gas stream:$$CO\, ( \% )=\frac{formation\,of\,C{O}_{2}(mol\,{h}^{-1})}{Initial\,amount\,of\,CO(mol\,{h}^{-1})}\times 100 \%$$
(1)
In situ ETEM indicates that the Pt species are highly dispersed on the CeO2 supports after O2 and H2 pretreatment. The catalysts show interface reconstruction under reaction conditions. The size of Pt nanoparticles changes with time. In addition, AP-XPS indicates that the concentration and dispersion of Pt on three different CeO2 supports are different. Therefore, the formation rates of CO2 and H2 on three different catalysts was calculated by comparing the concentration of formed H2 and CO2 with the total amount of Pt (1 wt%; 50 mg catalyst):$$formation\,rate\, (mol\,{g}^{-1}\,{h}^{-1})=\frac{formation\,of\,C{O}_{2}/{H}_{2}(mol\,{h}^{-1})}{total\,amount\,of\,Pt(5\times {10}^{-3}g)}$$
(2)
Since the AP-XPS results indicates that the concentration/dispersion of surface Pt on the CeO2 octahedra is much higher than that on the CeO2 nanorods and nanocubes. Therefore, the activity per surface sites (Pt-CeO2) of Pt/CeO2-oct may be much lower than that of Pt/CeO2-rod, as shown in Supplementary Fig. 4c. However, due to the complexity of Pt/CeO2 interfacial structure as well as the structure reconstruction during the WGSR, it’s challenging to calculate the activity per site for three different catalysts. But the rank of activity per site on CeO2 supported Pt catalysts remains unchanged: Pt/CeO2-rod > Pt/CeO2-cube > Pt/CeO2-oct.In order to investigate the thermodynamic equilibrium of the WGSR, the reaction was simulated in ASPEN Plus V10 (Gibbs reactor). A feed stream of CO and H2O with a ratio of 1: 2 was applied. Mole fractions of each compound in the feed and product with respect to the temperature of the reactor are shown in the supplementary information.
Tracking the dynamics of catalytic Pt/CeO2 active sites during water-gas-shift reaction
