Unveiling the Role and Stabilization Mechanism of Cu+ into Defective Ce-MOF Clusters during CO Oxidation

Copper single-site catalysts supported on Zr-based metal–organic frameworks (MOFs) are well-known systems in which the nature of the active sites has been deeply investigated. Conversely, the redox chemistry of the Ce-counterparts is more limited, because of the often-unclear Cu2+/Cu+ and Ce4+/Ce3+ pairs behavior. Herein, we studied a novel Cu2+ single-site catalyst supported on a defective Ce-MOF, Cu/UiO-67(Ce), as a catalyst for the CO oxidation reaction. Based on a combination of in situ DRIFT and operando XAS spectroscopies, we established that Cu+ sites generated during catalysis play a pivotal role. Moreover, the oxygen vacancies associated with Ce3+ sites and presented in the defective Cu/UiO-67(Ce) material are able to activate the O2 molecules, closing the catalytic cycle. The results presented in this work open a new route for the design of active and stable single-site catalysts supported on defective Ce-MOFs.

The synthesis has been carried out following a scaled-up previously reported recipe 1 : H2BPDC (Biphenyl-4,4´-dicarboxylic acid) (1.42 g, 5.86 mmol) and 33 mL of N,N-dimethylformamide (DMF) were placed in a round-bottomed flask.Then, an aqueous solution of cerium ammonium nitrate (11 mL, 0.53 M) was added to the mixture.The flask with the resulting solution was sealed and heated under magnetic stirring for 15 min at 100°C.The as-obtained yellow precipitate was decanted by centrifugation from the mother solution and washed with dimethyl sulfoxide (DMSO), DMF and acetone.To avoid the presence of unreacted organic ligand on the MOF pores, the material was treated overnight with DMF at 100ºC.Finally, the UiO-67(Ce) was decanted and washed with DMF and acetone, and subsequently dried in an oven at 100ºC for 1 hour.

Synthesis of Cu/UiO-67(Ce).
The Cu-supported UiO-67(Ce) material was prepared following a previous synthetic method with some modifications. 2 Firstly, Cu(OAc)2•H2O (70 mg, 0.35 mmol) was dissolved in 67 mL of DMF.The resulting solution was added to a Teflon-lined autoclave containing 1.25 g of UiO-67(Ce).The suspension was heated at 100°C in an oven during 24 h.The as-synthesized Cu/UiO-67(Ce) was decanted and washed with DMF and acetone.Finally, the material was dried in an oven at 100ºC for 1 hour.

Sample Characterization
The PXRD measurements were carried out using the Bragg-Brentano geometry with a PANalytical PW3050/60 X'Pert PRO MPD diffractometer with a Cu anode (Kα = 1.5418Å) and an X'Celerator detector.
The morphology of the samples, placed on a carbon tape, was studied by field emission scanning electron microscopy (FESEM) using a FEG-SEM TESCAN S9000G microscope, equipped with a FEG, Scottky type of source.Energy dispersive X-ray (EDX) spectrometry analyses were performed at the same time to obtain the metallic distribution on the MOF particles.DR UV-vis spectra were measured with a Varian Cary5000 spectrophotometer, equipped with a diffuse reflectance sphere, where the samples were placed in powder form.The spectra were collected in a reflectance mode and successively converted as Kubelka-Munk F(R) function.
Isothermal N2 physisorption measurements at liquid nitrogen temperature (LNT) were performed on a Micromeritics 3Flex.Prior to the measurement, the powders were degassed 4h at 110°C.Specific surface areas using the Brunauer-Emmett-Teller (BET) model were calculated following the Rouquerol´s criterion.Pore size distribution was obtained by applying the N2-Cylindrical Pores-Oxide Surface DFT model.
Thermogravimetric analysis (TGA) data were recorded with a TA Instruments Q600 thermobalance in air flow (100 mL/min) with a ramp of 5°C/min from room temperature (RT) to 500°C working with ~5 mg of sample in an alumina crucible.
ICP analyses were carried out in a Varian 715-ES ICP-Optical Emission spectrometer after solid dissolution in H2SO4/H2O2 aqueous solution.
Fourier Transform Infrared (FT-IR) spectroscopy in transmission mode was employed to characterize surface properties of the materials by following the adsorption/desorption of CO as probe molecule.Absorption/transmission FTIR spectra were collected using a Bruker Vertex 70 spectrophotometer equipped with a Mercury Cadmium Telluride (MCT) cryo-detector in the 4000-600 cm -1 range with 2 cm -1 resolution.Powders were pressed in self-supporting discs (~4 mg/cm 2 ) and placed in quartz IR cells suitable for thermal treatments in controlled atmosphere and for spectra recording even at LNT.Before IR measurements, catalysts were activated from RT to 110°C at 5°C/min holding at 110°C until the pressure reached 5.10-4 mbar.The pre-reduced Cu/UiO-67(Ce) was treated as follow: 1) The sample was degassed at 110°C (5°C/min) until dynamic vacuum.2) Then, H2 (100 mbar) was dosed into the cell and it was heated up to 200°C during 30 min.3) Finally, the cell was cooled down and evacuated until the pressure reached 5.10-4 mbar.Spectra were treated using Bruker OPUS spectroscopy software.All the reported spectra were normalized for the pellet weight and area.
In situ IR experiments were conducted with a Bruker Invenio spectrophotometer equipped with a 'Sandwich type' IR cell. 3The catalyst wafer was inserted in the round-shaped homemade sample holder.Bottles of gases were connected to Bronkhorst EL-flow mass flowmeters.Background spectra was collected after flushing the empty cell for 30 minutes under He flow (30 mL/min).Spectra were recorded with 60 kHz and every spectrum was the results of the average of 32 scans (ca.60 s/scan).
The measurement protocol consisted in: 1) catalyst activation at 200°C (3°C/min) under He flow (30 mL/min) for 30 min.2) Sample reduction at 200ºC under a pure H2 flow (30 mL/min) during 30 min to be then removed with pure He during 10 min.
3) This process was then repeated but O2 was used instead H2. 4) After the last cleaning step in He, the pellet was flushed with CO (30 mL/min, 3% CO and 27% He) at 200ºC during 30 min.
DRIFT experiments were conducted with a Bruker Invenio spectrophotometer equipped with a commercial cell (PIKE TECHNOLOGIES, DIFFUSIR™).
Measurements were performed under similar reaction conditions followed in the catalytic tests: the Cu/UiO-67(Ce) was activated at 200 o C during 30 minutes passing a flow of He (45 mL/min).Then, the reaction mixture was flowed (45 mL/min, containing 2.22% CO and 1.11% O2) at atmospheric pressure at 200 o C for 12 h.Spectra were collected (every 10 min) and reported in Kubelka-Munk units.
Background subtraction of the spectra was performed using a spectrum recorded initially in a flow of He (45 mL/min) at 200ºC.X-Ray Absorption Spectra (XAS) were measured at BM23 beamline of the European Synchrotron Radiation Facility (ESRF) during in-house beamtimes IH-CH-1712 and IH-CH-1726. 4 The storage ring was operating in 16 bunch mode with maximum ring current of 75 mA for Cu K-edge experiment while Ce L3-edge were collected in 4 bunch mode with maximum ring current of 32 mA.Cu K-edge and Ce L3-edge XAS spectra were collected in transmission mode on mass optimized pellets of ≈0.2 cm 2 and 6.8 and 1.8 mg, respectively.Two ion chambers (IC) were employed to measure incoming beam (IC0) and beam transmitted from the sample (IC1).A third IC (IC2) was located after IC1 to measure transmitted beam from a reference sample employed for energy alignment purposes.Cu K-edge spectra were collected in the 8.8 -10.2 keV energy range in continuous mode with 0.3 eV/point energy resolution and 0.04 s/point integration time for a total of 3.5 minutes/scan.Cu metal foil was used for energy calibration and alignment.Ce L3-edge spectra were collected in the 5.55 -6.15 keV energy range with 0.25 eV/point energy resolution and 0.04 s/point integration time for a total of 2 minutes/scan.Pellet of CeO2 was used for energy alignment.Spectra of steady states resulted from the average of 10 spectra.Spectra were energy aligned, background subtracted and edge jump normalized with Larch based python script. 5All the reported FT-EXAFS spectra were extracted with Hanning window in the 2.2-11.4Å -1 k-range.
The sample was placed in a home-made reaction cell suitable for thermal treatments under gas flows (Transmission Pellet Cell, TPC).Inlet gas flows were provided by Bronkhorst mass flow controllers while reactants consumption/products formation were determined with an online Mass Spectrometer connected to the cell outlet.The operando XAS experiment followed a protocol similar to the one employed for DRIFT experiment i.e.: the sample was heated to 200⁰C (5⁰C/min) under He flow (45 mL/min) and kept at this temperature for 30 minutes.Then, the reaction mixture was flowed (45 mL/min, containing 4.44% CO and 2.22% O2) at atmospheric pressure at 200 o C for 20 h.The cell was then purged under He (50 mL/min) for 30 minutes and cooled to RT (5 o C/min).The outlet gas composition was monitored with a mass spectrometer.Spectra of CuO, Cu2O, CeO2 and Ce(NO3)3 references were measured on mass optimized pellets in transmission mode with the same parameters.Cu(OAc)2 spectra was recorded in water solution.

CO oxidation reaction.
In a typical experiment, 80 mg of Cu/UiO-67(Ce) were loaded in a conventional tubular plug-flow reactor (ID = 4 mm).The reactor was heated at 200ºC during 30 minutes flowing He through the catalyst (15 mL/min) to eliminate the physisorbed molecules.Then, the reaction mixture was flowed at atmospheric pressure at different temperatures from 100 to 200ºC every 25ºC: 15 mL/min (GHSV = 11250 mL/gcat•h), containing 6.67% CO and 3.33% O2.The downstream reaction effluents were analyzed continuously by gas chromatography.During thermal activation Cu K-edge XANES white-line (Figure S15a) and FT-EXAFS Cu-O shell (Figure S15b) intensities decrease in line with Cu dehydration.
Coordination number of Cu-O shell during activation was not evaluated due to the presence of thermal Debye Waller contribution as observable from the dampening of EXAFS spectra intensity (Figure S15c).Formation of Cu + was observed during thermal activation from the ΔXANES (Figure 3b and Figure S15a inset).ΔXANES was calculated as the difference between each spectrum and the first spectrum collected at RT under He.where Reff is the distance from the provided atomic model while doff and α are parameters refined during the fit describing the initial offset from Reff and the bond expansion, respectively.ΔE and CN were fit to the same value for all the scans.The scattering amplitude factor was fixed to 0.89 after being evaluated from fit of reference Cu metal FT-EXAFS.As can be observed from the fit results in Figure S16 and Table S4 all the experimental spectra were well described by the employed model.The obtained α parameter (Table S4) indicated that no expansion occurred during cooling.However, the evaluated CN (≈3) is suspiciously low indicating as the single Cu-O scattering path might not properly describe Cu local structure.The EXAFS spectra of the sample collected at RT-He was fit using three Cu-O single scattering paths calculated by Feff6 using the simulated structural model reported in Figure S13.The scattering amplitude factor was fixed to 0.89 (value obtained from fit of Cu metal foil).As reported in Figure S17c, the three Cu-O scattering paths presented imaginary components partially in antiphase, inducing a high correlation between their radial distances and DW factors.For this reason, we have constrained their DW factors to 0.0042 Å 2 .This value was obtained from the fit during cooling reported above and it reasonably describes the DW factor value at RT.For the same reason coordination numbers were constrained to those calculated by FEFF6.With this approach the only free variable were the E0 and the scattering paths radial distances.As reported in Table S5, all the four structures converged to the same results.For sake of clarity, the fit results are reported in Table S5 while only the best fit curves obtained using the structure Cu-2 are reported in Figure S17.Table S5.Results of the fit of the EXAFS spectra collected at RT under He after the reaction protocol.The results obtained using all the structures calculated by DFT are reported for clarity.The FT-EXAFS was extracted in the 2.2-11.4Å -1 k-space whit Hanning window while fit was performed in the 1.4-2.4Å range.*These parameters were fixed and not refined.

Operando Ce L3-edge
Ce 3+ /Ce 4+ content was evaluated through Linear Combination Fit of the experimental spectra considering Ce(NO3)3 as reference for Ce 3+ .As previously described, 22,23 instead of using CeO2 as Ce 4+ reference, we have employed the first spectrum of UiO-67(Ce), hypothesizing as its hydrated state contains only Ce 4+ .

2. 6 .
Figure S6.IR spectra of the UiO-67(Ce) (black line) and Cu/UiO-67(Ce) (red line).ν(OH) and ν(CH3) regions are reported in panel a and b, respectively, while panel c shows the full spectra.The presence of isolated acetates could be discarded since ν(CH3)sym vibrations at ~2940 cm -1 in the IR spectrum of Cu/UiO-67(Ce) reported in panel b were not detected.

Figure S7 .
Figure S7.Difference IR spectra of CO desorption at LNT on the pre-reduced Cu/UiO-67(Ce) sample.

Figure
Figure S8.IR spectra of Cu/UiO-67(Ce), focused in the ν(OH) range, collected a) during He activation (from RT to 200⁰C with a 3⁰C/min ramp) followed by b) H2 treatment, c) O2 treatment and d) CO treatment (10% in He) at 200⁰C (time evolution from red to orange line).Inset in panel a describes the desorption of the DMF molecules trapped into the MOF pores while inset in panel d shows the difference IR spectra during CO adsorption focused on ν(CO) region.PXRD patterns of the fresh Cu/UiO-67(Ce) (red line) and after be treated with H2 (orange line) are reported in the inset of panel b.

3. 1 .
Cu effect on the catalytic performance of the UiO-67(Ce) material

Figure S10 .
Figure S10.N2 adsorption-desorption (filled and empty circles, respectively) isotherms of the fresh Cu/UiO-67(Ce) (red circles) and after be tested in the CO oxidation reaction (orange circles).

Figure S11 .
Figure S11.PXRD patterns of the fresh Cu/UiO-67(Ce) (red line) and after be used in the CO oxidation reaction (orange line).

Figure S12 .
Figure S12.Difference IR spectra of CO desorption at LNT on the a) fresh Cu/UiO-67(Ce) (red line) and b) after the catalytic test (orange line).

Figure 5 .
Figure S13.DFT-optimized structures of Cu 2+ units bounded to a defective Ce-based node in Cu/UiO-67(Ce).Electronic energies with respect to Cu-1 are shown in kcal.mol - .

Figure S16 .
Figure S16.Experimental (colored lines) and best fit (orange line) of a) k 2 -weighted EXAFS, b) phase uncorrected magnitude and c) imaginary components of k 2 -weighted FT-EXAFS spectra collected during cooling from 200⁰C to RT under He.For clarity, only spectra measured every 50⁰C are reported.D) Cu-O DW factors (black squares) evaluated at each temperature using the Einstein model.

Figure S17 .
Figure S17.Experimental (red line) and best fit (orange line) of Cu/UiO-67(Ce) a) k2weighted EXAFS and FT-EXAFS b) magnitude and c) imaginary components measured at RT under He.FT-EXAFS imaginary components of Cu-OI, Cu-OII and Cu-OIII are reported in light blue, green and dark red line in panel c), respectively.

Figure S18 .
Figure S18.Cu K-edge a) XANES and k 2 -weighted FT-EXAFS spectra b) magnitude and c) imaginary components of Cu/UiO-67(Ce) at 200⁰C before (red line) and after exposure to CO/O2 mixture for 24h (orange line).References CuO (dark red line), Cu2O (green line) and Cu acetate solution (light blue line) are reported for clarity.Detail of Cu + 1s→4p transition is reported in the inset in panel a. d) CO and CO2 MS signals collected during CO oxidation reaction.

Figure
Figure S19.a) In situ Ce L3-XANES spectra collected during sample activation (from RT-He to 200⁰C-He, from dark red to red line).Spectrum of Cu/UiO-67(Ce) after 24h of reaction is reported with orange line while references Ce(NO3)3 and CeO2 are reported with light blue and green lines, respectively.b) Ce 4+ /Ce 3+ ratio estimated from LCF. c) CO and CO2 MS signals collected during 24h reaction.

Table S2 .
3.2.Comparison of the catalytic performance for the Cu/UiO-67(Ce) with other stateof-the-art Cu supported on different Ce-based materials Catalytic performance of Cu-catalysts supported on different Ce-based materials.

Table S4 .
Results of the fit of the FT-EXAFS collected during cooling under He.For all the spectra FT-EXAFS was extracted in the 2.2-11.4Å -1 k-space with Hanning window while fit was performed in the 1.4-2.4Å range.Error bars are not reported since Larch does not evaluate errors when correlations between variables are <0.1.