Pd-incorporated polyoxometalate catalysts for electrochemical CO2 reduction

Polyoxometalates (POMs), representing anionic metal–oxo clusters, display diverse properties depending on their structures, constituent elements, and countercations. These characteristics position them as promising catalysts or catalyst precursors for electrochemical carbon dioxide reduction reaction (CO2RR). This study synthesized various salts—TBA+ (tetra-n-butylammonium), Cs+, Sr2+, and Ba2+—of a dipalladium-incorporated POM (Pd2, [γ-H2SiW10O36Pd2(OAc)2]4−) immobilized on a carbon support (Pd2/C). The synthesized catalysts—TBAPd2/C, CsPd2/C, SrPd2/C, and BaPd2/C—were deposited on a gas-diffusion carbon electrode, and the CO2RR performance was subsequently evaluated using a gas-diffusion flow electrolysis cell. Among the catalysts tested, BaPd2/C exhibited high selectivity toward carbon monoxide (CO) production (ca. 90%), while TBAPd2/C produced CO and hydrogen (H2) with moderate selectivity (ca. 40% for CO and ca. 60% for H2). Moreover, BaPd2/C exhibited high selectivity toward CO production over 12 h, while palladium acetate, a precursor of Pd2, showed a significant decline in CO selectivity during the CO2RR. Although both BaPd2/C and TBAPd2/C transformed into Pd nanoparticles and WOx nanospecies during the CO2RR, the influence of countercations on their product selectivity was significant. These results highlight that POMs and their countercations can effectively modulate the catalytic performance of POM-based electrocatalysts in CO2RR.


Experimental Section
Instruments GC analyses were carried out by Nexis GC-2030 gas chromatography (Shimadzu Corporation) equipped with a barrier ionization discharge (BID) detector and a ShinCarbon ST Micropacked column (2.0 m × 1.0 mm I.D., Shinwa Chemical Industries Ltd.).Electrochemical measurements were conducted by VSP-300 multichannel potentiostat (BioLogic). 1 H NMR (500.16MHz) measurements were performed a by JNM ECA-500 spectrometer (JEOL Ltd.) using 5-mm outer-diameter tubes. 1 H NMR chemical shifts were referenced to dimethyl sulfoxide signal (2.6 ppm).S1 IR measurements were carried out by FT/IR-4100 (JASCO Corporation) using KBr disks.ICP-AES measurements were conducted by ICP-8100 (Shimadzu Corporation) and iCAP PRO XP ICP-OES Duo (Thermo Fisher Scientific Inc.) at the Analytical Chemistry Center of the School of Engineering, The University of Tokyo.AAS measurement was performed by ZA3000 (Hitachi High-Tech Corporation) at the Analytical Chemistry Center of the School of Engineering, The University of Tokyo.CHN analyses were carried out by CE-440F Elemental Analyzer (Exeter Analytical Inc.) at the Analytical Chemistry Center of the School of Engineering, The University of Tokyo.TEM observations were conducted by JEM-2010F (JEOL Ltd.).HAADF-STEM observations and STEM-EDS mappings were performed by JEM-ARM200F Thermal FE (JEOL Ltd.).

Synthesis of strontium trifluoromethanesulfonate
Strontium trifluoromethanesulfonate (Sr(OTf)2) was prepared by modifying the reported procedures.S3,S4 To a solution containing water (10 mL) and HOTf (5 mL, ca.55 mmol), excessive amount of Sr(OH)2•8H2O (9.21 g, 35 mmol) was added.Then, the mixture was refluxed at 130℃ for 1 h.The white precipitate was removed by membrane filtration and the resulting filtrate was evaporated to obtain white crude product.This crude product was dissolved into acetone (20 mL) and insoluble materials were filtered out by syringe filtration.Finally, the resulting filtrate was evaporated to obtain Sr(OTf)2.

Preparation of TBAPd2/C
Immobilization of TBAPd2 on a carbon support (Vulcan XC 72R) was performed by modifying the reported procedure about immobilization of Keggin-type polyoxometalates (POMs) on single-walled carbon nanotubes.S5 A carbon support (Vulcan XC 72R, 20 mg) was dispersed in ethyl acetate (25 mL) with the aid of ultrasonication.While vigorously stirring the suspension, an acetone solution (2.5 mL) containing TBAPd2 (20 mg, 5.3 μmol) was quickly added to the suspension.This mixture was stirred for 1 h and the resultant suspension was stand still for additional 1 h.The black precipitates were collected by membrane filtration, washed with ethyl acetate, and dried under suction to obtain TBAPd2/C.Preparation of CsPd2/C, SrPd2/C, and BaPd2/C CsPd2/C was prepared by modifying the reported procedure about immobilization of vanadium-incorporated POMs on oxide supports.S3 TBAPd2/C (32 mg) was added to an acetone solution (16 mL) containing CsOTf (4.8 mg, 17 μmol, the cation exchange step), followed by vigorous stirring for 2 h.The black precipitates were collected by membrane filtration and washed with an acetone solution (16 mL) containing CsOTf (2.4 mg, 8.5 μmol, the washing step).Then, the black precipitates were washed with acetone (16 mL) three times and dried under suction to obtain CsPd2/C.SrPd2/C and BaPd2/C were prepared by a similar procedure to CsPd2/C, using Sr(OTf)2 (3.3 mg, 8.5 μmol) and Ba(OTf)2 (3.7 mg, 8.5 μmol), respectively, for the cation exchange step and Sr(OTf)2 (1.6 mg, 4.3 μmol) and Ba(OTf)2 (1.9 mg, 4.3 μmol), respectively, for the washing step.
For the preparation of the catalyst ink containing Pd(OAc)2, the mixture of a carbon support (10 mg), Pd(OAc)2 (1.12 mg, 4.98 µmol), neutralized Nafion solution (200 μL), and acetonitrile (4 mL) was ultrasonicated for 1 h.The catalyst ink (800 μL) was dropcast onto a carbon electrode (Sigracet 39 BB or AvCarb P75T, 2.5×2.5 cm 2 ) and the electrode was dried at 313 K for 3 h.As a carbon electrode, Sigracet 39 BB was used for a 1 h reaction and AvCarb P75T was used for a 12 h reaction.

Electrochemical measurements
Constant potential electrolysis was carried out in a gas-diffusion flow electrolysis cell (Fig. S3) with three-electrode system.S6,S7 This gas-diffusion flow electrolysis cell is composed of gas chamber for gas delivery and gaseous CO2RR products collection, catholyte chamber for catholyte circulation and liquid CO2RR products collection, and anolyte chamber for anolyte circulation.Gas chamber and catholyte chamber were separated by working electrode (WE, catalyst-modified electrode with catalytic area of 1.4×1.4cm 2 ) while catholyte chamber and anolyte chamber were separated by pre-activated AEM.Reference electrode (RE, Ag/AgCl electrode filled with a 3 M NaCl aqueous solution, BAS Inc.) and counter electrode (CE, Pt mesh electrode) were placed in catholyte chamber and anolyte chamber, respectively.All potentials were recorded against reversible hydrogen electrode (RHE) using the following equation: where pH is a pH of the catholyte after the reaction measured by HM-30G (DKK-TOA CORPORATION), I [A] is a current, and R [Ω] is a solution resistance determined by potentiostatic electrochemical impedance spectroscopy.
Working electrode, reference electrode, and counter electrode were connected to VSP-300 electrochemical station (BioLogic) to perform electrochemical measurements.A 1 M KHCO3 aqueous solution (30 mL) was used as the catholyte and the anolyte, and they were circulated at 5 mL/min by using a peristaltic pump (TOKYO RIKAKIKAI CO., LTD.).The inlet gas flow rate was controlled at 80 mL/min by 8700MC mass flow controller (KOFLOC Corp.) for CO2 and 8500MC mass flow controller (KOFLOC Corp.) for Ar.The outlet gas flow rate was measured for each measurement by GFM-2000 flow meter (Shimadzu GLC Ltd.) to quantify gaseous products.Every experiment was conducted under ambient pressure (1 atm) and temperature (298 K).Two independent experiments in each condition were performed.
Carbon monoxide (CO) stripping voltammetry was performed by using the abovementioned gas-diffusion flow electrolysis cell (Fig. S3).Prior to the experiment, the constant potential electrolysis around −0.8 VRHE under CO2 atmosphere was performed for 10 min in a 1 M KHCO3 aqueous solution.Then, the applied potential was changed to 0.1 VRHE and CO was introduced to the gas chamber for 5 min.Subsequently, while maintaining the applied potential to 0.1 VRHE, the gas chamber and the catholyte were purged with Ar for more than 30 min to remove excess CO.Finally, the potential was linearly swept between 0 VRHE and 1.5 VRHE at a scan rate of 10 mV/s.
The Faradaic efficiency of product i (FEi) was calculated according to below equation: where ni is the number of electrons used to produce one molecule of i, F is the Faraday constant (96485 C/mol), Ci [%] is the concentration of i in the outlet gas, v [m 3 /s] is the outlet gas flow rate, P is the atmospheric pressure (1.013×10 5 Pa), R is the universal gas constant (8.314Pa•m 3 /(mol•K)), T is the room temperature (298 K), and is a current.

Liquid CO2RR products analysis
The liquid CO2RR products, formate (HCOO − ) and methanol (CH3OH), were quantified by referring to the previous literature.S8 The catholyte after the reaction (400 μL) was first mixed with D2O (200 μL) and internal standard solution (50 μL) containing 20 mM phenol and 10 mM dimethyl sulfoxide.The 1 H NMR measurement of this mixture was performed at 303 K using a water suppression method.The concentration of product i in the catholyte (Ci) was determined by using 5-point calibration curve.
The Faradaic efficiency of i (FEi) was calculated according to below equation: where ni is the number of electrons used to produce one molecule of i, F is the Faraday constant (96485 C/mol), Ci [mol/L] is the concentration of i in the catholyte, V [L] is the volume of the catholyte (0.030 L), and Q [C] is a total charge passed through the working electrode. ´

XAFS analysis
XAFS measurements were carried out at the BL14B2 beamline of SPring-8.The X-ray beam was monochromatized using a single pair of Si(311) crystal monochromators for the Pd K-edge and W L3-edge XAFS.X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) data were analyzed using Athena and Artemis software (Demeter, ver.0.9.025; Bruce Ravel).The data reduction procedure for EXAFS consisted briefly of the following steps: pre-edge subtraction, background determination, normalization, and spectra averaging.The edge position is defined to be the first inflection point on the leading absorption peak.The energy was calibrated by Pd foil for the Pd K-edge XAFS and W foil for the W L3-edge XAFS.The background in the EXAFS region was approximated using a cubic spline routine and optimized according to the criteria described by Cook and Sayers.S9 Then, the spectra were normalized by the edge-step at 50 eV after the absorption edge.The k 3 -weighted EXAFS functions were Fourier-transformed, filtered, and fitted in R-space in the range of 3-12 Å −1 for Pd and 3-13 Å −1 for W. Fourier filtering was used to isolate the contributions of specific shells and to eliminate low frequency background and high frequency noise.Fourier filtering was done by choosing a window in the Fourier-transformed spectrum and calculating the inverse Fourier transform of the selected R-range.The interatomic distance (R), the coordination number (C.N.), the difference of the Debye-Waller factor from the reference (σ 2 ), and the correction of the threshold energy (ΔEj0) were treated as free parameters during the fitting unless otherwise specified in Table S1.
The quality of the fit was estimated from R-factor.R-factor represents the residuals between the observed and calculated spectrum in the fitted range.Low values of R-factor indicate a good agreement between the data and model.

Supplementary Figures
Fig. S2 IR spectra of TBAPd2 and BaPd2.
Fig. S3 The configuration of a gas-diffusion flow electrolysis cell for CO2RR.

Supplementary Tables
Table S1 Fitting parameters of EXAFS spectra of TBAPd2, TBAPd2/C, BaPd2, and

Table S2
The results of ICP-AES and AAS measurements of TBAPd2/C, CsPd2/C, SrPd2/C, and BaPd2/C.

Table S5
The results of control experiments for 1 h in 1 M KHCO3 using a carbon support (C) and BaPd2/C.

Table S8 (
continued) The summary of POM-based electrocatalysts for CO2RR in aqueous electrolytes.