Understanding the CO Oxidation on Pt Nanoparticles Supported on MOFs by Operando XPS

Abstract Metal‐organic frameworks (MOFs) are playing a key role in developing the next generation of heterogeneous catalysts. In this work, near ambient pressure X‐ray photoelectron spectroscopy (NAP‐XPS) is applied to study in operando the CO oxidation on Pt@MOFs (UiO‐67) and Pt@ZrO2 catalysts, revealing the same Pt surface dynamics under the stoichiometric CO/O2 ambient at 3 mbar. Upon the ignition at ca. 200 °C, the signature Pt binding energy (BE) shift towards the lower BE (from 71.8 to 71.2 eV) is observed for all catalysts, confirming metallic Pt nanoparticles (NPs) as the active phase. Additionally, the plug‐flow light‐off experiments show the superior activity of the Pt@MOFs catalyst in CO oxidation than the control Pt@ZrO2 catalyst with ca. 28 % drop in the T 50% light‐off temperature, as well as high stability, due to their sintering‐resistance feature. These results provide evidence that the uniqueness of MOFs as the catalyst supports lies in the structural confinement effect.

containing water (100 ml). The mixture was heated to 100 °C and stirred under reflux for 40 h.
After the synthesis, the resulting precipitate was filtered and washed using deionised water and then dried in an oven at 70°C. A PtCl2(bpydc) yield of ca. 75% (based on the initial amount of K2PtCl4) was achieved, which is close to the yield reported in the literature. [2a] 2 wt.% Pt supported on UiO-67 by the LD (LD-Pt@UiO-67): To a 100 ml Teflon-lined autoclave reactor containing 50 mL N,N-dimethylformamide was added ZrCl4 (1.28 mmol, 1 equiv), 4,4biphenyl-dicarboxylic acid (bpdc) (1.23 mmol), and benzoic acid (12.8, 10 equiv.). The resulting solution was sonicated for 10 min to dissolve the suspended particles until an opaque white solution was obtained. The reactor was heated at 95 °C (in an oil bath) for 4 days with the lid tightly sealed.
After the synthesis, the solution was allowed to cool to room temperature, and the liquid was decanted. The solid particles were exchanged for fresh ethanol several times and then filtered and dried under vacuum at 100 °C. LD-Pt@UiO-67 was reduced at 262 °C under hydrogen-argon flow (flowrate: 70 ml min −1 (5% H2), heating ramp: 5 °C min −1 ) to give LD-PtNPs@UiO-67.

Materials characterisation:
X-ray diffraction (XRD) of materials was carried out on a Rigaku Miniflex diffractometer using CuKα1 radiation (λ = 0.15406 nm, 30 kV, 15 mA). The measurement was perforce over a range of 4° < 2θ < 45° in 0.05 step size at a scanning rate of 1° min −1 . Scanning electron microscopy (SEM) was undertaken using a FEI Quanta 200 ESEM equipment using a work distance of 8-10 mm and an accelerating voltage of 20 kV. All samples were dispersed in ethanol and dropped onto SEM stubs, followed by the gold coating using an Emitech K550X sputter coater under vacuum (1×10 −4 mbar). Nitrogen (N2) sorption on materials at −196.15 °C was carried out using a Micromeritics ASAP 2020 analyser. Prior to the N2 adsorption, samples (~100 mg) were pretreated by degassing at 200°C under vacuum overnight. The surface area and total pore volume of the materials were calculated based on Brunauer-Emmett-Teller (BET) theory and at a relative pressure P/P 0 of 0.99, respectively. Thermogravimetric analysis (TGA) and was performed by use of a TG analyser (Beijing Boyuan Science and Technology Development Co., Ltd) from the room temperature to 700 °C in air (flowrate = 0.6 ml min −1 ) at a heating rate of 5 °C min −1 .
The Pt content of the synthesised catalysts was determined by inductively coupled plasma optical emission spectrometry (ICP-OES, Thermo iCAP 6000 SERIES). Samples were digested in the nitric acid solution overnight then solutions were analysed by ICP for the quantitative determination of the Pt contents. It was determined that about 2 wt.% Pt species were present in all catalysts.

Operando near ambient pressure XPS (NAP-XPS) study:
All XPS spectra were recorded with a SPECS NAP-XPS system employing a monochromatic Al Kα source (1486.6 eV). The catalyst powders were dispersed in ethanol and then drop-cast onto a silicon substrate. The drop-casting was repeated until a uniform layer of MOF covered the silicon pieces and no significant Si 2p signal was observable in XPS. The insulating nature of these MOFs presents a challenge to NAP-XPS analysis as conventional charge compensation mechanisms (electron flood sources) are not usable within the high-pressure environment of a NAP-XPS. In vacuum conditions, the samples charged by ~50 eV and substantial differential charging make the spectra unusable. However, it was found by admitting an appropriate amount of gas into the NAP cell and heating the sample, the sample could be fully charge compensated and usable spectra obtained. The spectra were recorded at a pass energy of 30 eV and charge corrected to the main component of the C 1s peak at 284.5 eV for aromatic carbon (the main constituent of the 4,4'biphenyl-dicarboxylic acid, bpdc, linker in UiO-67). The Zr 3d peak was used as a check to ensure that no differential charging was occurring. Its position was 182.2 ± 0.1 eV and did not significantly change throughout the experiment (see Fig 2d in main text), confirming that the shifts on the Pt 4f were due to real physical/chemical changes and not related to differential charging.
The NAP-XPS was performed during catalyst exposure to a mixture of CO:O2 (CO/O2 ratio = 2, total pressure = 3 mbar) at room temperature before it was heated in steps to 260 °C, with XPS spectra acquired at each temperature. The C 1s and O 1s core levels were also acquired, but information about C and O surface species on the Pt particles could not be obtained as the signal was dominated by the C and O in the bpdc linker. The acquisition time for the XPS spectra at each temperature was about 2 h in the temperature-programmed measurements (100-260 °C). XPS measurements were taken at 100 °C, 150 °C, 200 °C, 225 °C and 260 °C, respectively, to probe the chemical state of the Pt catalysts. Online gas analysis of the CO oxidation over the catalysts was monitored using a quadrupole mass spectrometer (QMS) (MKS EasyView 2) in the second differential stage of the NAP hemispherical analyser.

CO oxidation reaction the quartz plug-flow reactor:
The reaction was carried out in a quartz tubular reactor (9.0 mm ID) placed inside a programmable furnace at atmospheric pressure and with a flowrate of 100 ml min −1 (20 sccm CO, 10 sccm O2 and 70 sccm Ar). 50 mg of the pelletised catalysts (i.e. WI-PtNPs@UiO-67, LD-PtNPs@UiO-67 or PtNPs@ZrO2) were packed into the reactor and sandwiched by the quartz wool. Prior to the reaction, the catalyst was treated for 1 h in a reducing environment with the hydrogen-argon (Ar) flow (10 vol.% H2, the total flowrate of 100 ml min −1 , at 280 °C). The reaction temperature was ramped from the room temperature to 380 °C (for UiO MOFs catalysts) and to 580 °C (for PtNPs@ZrO2) at a heating rate of ca. 8 °C min −1 . The bed temperature was measured and recorded by a K type thermocouple adjacent to the catalyst bed. After each run, the furnace was turned off automatically to allow the reactor to cool down to the room temperature under Ar (at 100 ml min −1 ).
Mass spectrometry (MS) of the products was measured using an HPR20 QIC Hiden Analytical mass spectrometer. During the experiments, the spectrometer continuously monitored the ion currents at a mass-to-charge ratio (m/e) = 36, 18, 28, 32 and 44, corresponding to signals from Ar, H2O, CO, O2 and CO2, respectively. For cyclic catalyst deactivation tests, the same heating ramp and reactant gases were applied to the bed when the room temperature was achieved after each run.
where MPt is the molar mass of Pt (195 g mol −1 ), DPt is the Pt dispersion of the catalyst (-, estimated using the method below); and RPt is the Pt mass normalised reaction rate (mol s −1 gcatal. −1 ), defined as in Eq. (2): where of XCO is the measured CO conversion (-); CO n  is the molar flow rate of CO into the system (mol s −1 ); and mPt is the Pt mass present in the catalyst bed (g).

Pt dispersion estimation:
Calculation of Pt dispersion (DPt) for TOF estimation: Pt has a face-centred cubic (fcc) unit cell with the edge length of 0.392 nm (4 Pt atoms per unit cell). Therefore, Pt fcc unit cell volume is:

The effect of Pt precusors, catalysts preparation and reduction methods on the catalyst and acativty:
The method with Pt(acac)2 and acetone was used (for the preparation of WI-Pt@UiO-67 catalyst) due to the wettability issue with the K2PtCl4 and water system (which was used for the ZrO2 based catalyst) for impregnating UiO-67. In order to examine the effect of Pt precursors on the catalytic performance of wet impregnated UiO-67 catalysts, we prepared another reference WI-Pt@UiO-67catalyst using the method described below.
'UiO-67 support was synthesised by a microwave-assisted method proposed in our previous work. In the light-off experiment, the two WI-Pt@UiO-67 catalysts using (based on Pt(acac)2 (in acetone) and K2PtCl4 (in DMF) precursors) show comparable activity under the condition used, as shown in Figure S8. For the thermal decomposition of platinum(II) acetylacetonate, 200 °C is sufficient to decompose acetylacetonates thermally into acetone, as previously reported in the literature. [5] Additionally, we also performed TG analysis of platinum(II) acetylacetonate, showing the sharp drop in weight loss at around 200 °C ( Figure 2). Conversely, for the WI-PtNP@UiO-67 catalyst, there is no further weight loss at 200 °C ( Figure 2). It is worth noting that the weight loss from 250 °C to 350 °C is related to dehydroxylation of UiO-67 and removal of monocarboxylate ligands, [1] instead of the weight loss related to the decomposition of Pt acetylacetonate precursor. Therefore, we can conclude that the pre-treatment procedure used in this study is capable of removing the acetylacetonate groups from the catalyst. With regard to the Cl residues from Pt@ZrO2 prepared using K2PtCl4, we performed detailed ex situ X-ray absorption fine structure (XAFS) spectroscopy of the as synthesised and reduced samples to show the effectiveness of the reduction step to remove Cl residues ( Figures S11-12 and Table S1).
The XANES spectrum of PtNP@ZrO2 (reduced catalyst) compares well with the Pt 0 foil reference.
Fitting the EXAFS data of PtNP@ZrO2 using 1 st and 2 nd shell Pt-Pt scattering paths and a multiple   [b] amplitude = 0.85, 3.5< k <16.4, 1.15< R <3.0, number of independent points= 14.8 Figure S11. Magnitude of the k 2 weighted Fourier transform for the EXAFS data and fit for PtNP@ZrO2 and Pt@ZrO2. The imaginary part of the k 2 weighted FT of the data, fit and individual scattering paths is also shown for each. The χ(k) data of PtNP@ ZrO2 and Pt@ ZrO2 and the Pt foil are shown in Figure S12, showing that PtNP@ ZrO2 resembles the Pt metal reference closely. Thus, we can conclude that the pre-treatment procedure used in this study can remove the chloride and acetylacetonate groups from the catalyst surface, addressing the concern raised by the reviewer. The SI was also updated with relevant information to explain the choice of method for catalyst preparation and the effect of Pt precursors and catalyst pre-treatment methods on the resulting catalysts and activity.