The Dynamic Structure of Au38(SR)24 Nanoclusters Supported on CeO2 upon Pretreatment and CO Oxidation

Atomically precise thiolate protected Au nanoclusters Au38(SC2H4Ph)24 on CeO2 were used for in-situ (operando) extended X-ray absorption fine structure/diffuse reflectance infrared fourier transform spectroscopy and ex situ scanning transmission electron microscopy–high-angle annular dark-field imaging/X-ray photoelectron spectroscopy studies monitoring cluster structure changes induced by activation (ligand removal) and CO oxidation. Oxidative pretreatment at 150 °C “collapsed” the clusters’ ligand shell, oxidizing the hydrocarbon backbone, but the S remaining on Au acted as poison. Oxidation at 250 °C produced bare Au surfaces by removing S which migrated to the support (forming Au+-S), leading to highest activity. During reaction, structural changes occurred via CO-induced Au and O-induced S migration to the support. The results reveal the dynamics of nanocluster catalysts and the underlying cluster chemistry.


Synthesis of Au38(SC2H4Ph)24
Au38(SC2H4Ph)24/CeO2 was prepared as described in previous reports. 1 Au38(SC2H4Ph)24, here referred to as Au38(SR)24, was obtained by the following protocol: 1g of HAuCl4・3H2O and 3.12g of L-Glutathione were dissolved in 100ml Acetone and stirred vigorously for 30min at 0°C. The obtained yellow suspension was reduced with 960mg of NaBH4 dissolved in 30ml H2O (milli-Q grade, T=0°C). The liquid was decanted, and the black precipitant was dried at room temperature. 30ml of H2O, 1.6ml of ethanol, 10ml toluene and 10ml of 2-Phenylethanethiol were added. The reaction mixture was heated to 80°C and kept at this temperature for 4h. The product was cleaned multiple times with methanol. Further purification was achieved by Size Exclusion Chromatography (SEC). The purity of the clusters was determined by UV/Vis spectroscopy (dissolved in DCM) and by MALDI-MS. Figure S1. UV/Vis spectrum of Au38(SR)24 Figure S2. MALDI-MS spectrum of Au38(SR)24

Synthesis of CeO2
Ceria was synthesized by combustion method. 11.48g of cerium nitrate and 0.2g of Pluronic were dissolved in the minimum quantity of ethanol. The reaction mixture was heated up to 300°C with a 10°C/min rate and kept for 4h. The product was a voluminous flaky yellow powder.

Catalyst preparation
To prepare the catalyst, the theoretical amount of 2wt% of cluster was dissolved in toluene and added dropwise to a toluene ceria suspension. The suspension was stirred overnight. The remains of toluene were removed by centrifugation and decantation. The catalyst was washed with methanol and dried under vacuum. All experiments were performed with the same batch of catalyst.

EXAFS
XAS measurements were performed at the Claess Beamline at Alba Synchrotron in fluorescence mode (Au-L3 edge) in the beamline's solid-gas reactor "Multipurpose cell". The catalysts were pressed into pellets. The samples were pretreated inside the multipurpose cell (pre-treatment: 21vol% O2 in He, total flow: 20ml/min, ramp: 10°C/min) at 150°C or 250°C for 30min.
After cooling down (100vol% He), the gas mix was changed to reaction conditions (reaction: 3.3vol% CO, 7vol% O2 in He, total flow: 60ml/min). The samples were heated up to 150°C in 5°C/min steps. Once the maximum temperature was reached, the reaction chamber was cooled (100vol% He). EXAFS spectra were taken at 40°C in He at the beginning, after pretreatment and after reaction for each sample, without opening the reaction chamber in between.
The Artemis package 2 that uses the FEFF8 code 3 was applied for EXAFS data treatment, building a cluster model based on the known crystal structure of Au38(SC2H4Ph)24. The disorder factors σ0 2 of Au 0 core-S and Au + staple-S were set equal in the fitting model ( Figure S3).  Figure S3. EXAFS fitting in R space S4

KINETICS
Kinetic measurements were performed both in the operando EXAFS and DRIFTS cells, keeping the same flow conditions as mentioned above, yielding the same catalytic performance.

STEM-HAADF
Microstructural characterization by STEM-HAADF was performed at 200kV with a Tecnai G 2 F20 S-TWIN microscope equipped with a field emission electron source. c) d) Figure S6. STEM-HAADF images of a) as prepared sample, b) as prepared after CO oxidation, c) pret150 sample after CO oxidation and d) pret250 sample after CO oxidation S5 XPS X-ray photoelectron spectroscopy (XPS) measurements were performed within a UHV system (base pressure: 5x10 -10 mbar), equipped with a Phoibos 100 hemispherical analyzer and XR 50 X-ray source (SPECS GmbH). The catalyst powder samples were placed on transferrable sample holders using UHV-compatible conductive carbon tape. Spectra were measured at room temperature using Al-Kα radiation (1486.61eV) and an electron emission angle of 0°, with the analyzer operated in "large area" transmission mode. Using CasaXPS, all spectra were referenced to the C 1s signal (C-C, 284.6eV).
Subsequently, peaks were fitted after Shirley background subtraction utilizing Gauss-Lorentz sum functions, with peak positions and full width half maxima (FWHM) left unconstrained. For Au 4f, a doublet separation of 3.7eV (according to NIST XPS database) and a peak area ratio of 4 f5/2:4 f7/2 = 3:4 were assumed.

DRIFTS
DRIFTS studies were carried out on a Bruker Vertex 70 spectrometer with a liquid N2-cooled MCT detector and with 4cm −1 resolution. The stainless-steel flow cell (Pike) has a CaF2 window and an oven. The inlet of the cell was connected to a gas manifold system with calibrated mass flow controllers to adjust the gas mixtures (pre-treatment: 21vol% O2 in He, total flow: 20ml/min, reaction: 3.3vol% CO, 7vol% O2 in He, total flow: 60ml/min) and a mass spectrometer for kinetic measurements.
Each sample was placed into a small ceramic cup and the exact weight was taken for normalization (~30mg). The sample was pretreated at 250°C with a temperature ramp of 10°C/min and kept at the maximum temperature for 30min (21vol% O2 in He). The sample was cooled to room temperature (100vol% He) and then the gases were changed to reaction conditions without removing the sample in between. The reaction temperature was increased by 1°C/min steps and kept at the maximum temperature of 150°C (reached after 125min) for 4h. DRIFTS spectra were taken over the course of the whole experiment by averaging 256 scans to achieve good signal to noise ratio. For background removal CeO2 pretreated at the same conditions was used.   Figure S12. MS spectrum accompanying the DRIFTS surface plot. CO2 Ion current during CO oxidation on 2wt% Au38(SR)24 on CeO2 (flow: 3,3vol% CO, 7vol% O2, 89,7vol% He, total flow: 60ml/min, ramp: 1°C/min, dwell: 4h at 150°C): CO2 MS trace normalized to inert gas.