Electrochemical promotion of catalysis over Pd nanoparticles for CO2 reduction† †Electronic supplementary information (ESI) available. See DOI: 10.1039/c6sc04966d Click here for additional data file.

Electrochemical promotion of catalysis was observed over Pd nanoparticles with a significant rate enhancement ratio (ρ) for catalyzing CO2 reduction to produce formate in 1 M KHCO3 solution at ambient temperature.


Introduction
Electrochemical promotion of catalysis (EPOC), discovered by M. Stoukides and C. Vayenas in the 1980s, 1,2 has been widely investigated in more than 100 heterogeneous catalytic reactions 3-6 on either metal or metal oxide surfaces, which are interfaced with a solid 7 or an aqueous electrolyte solution. [8][9][10][11][12][13][14] By applying an electrical current between the working electrode, coated with catalyst, and the counter electrode, the electronic properties of the supported catalyst can be tuned, accompanied by the alteration in the adsorption strength of the reactants, and in some cases, a signicant enhancement in the catalytic performance can be observed. To date, only a few reports have demonstrated the EPOC effect in an aqueous electrolyte solution at ambient temperature for reactions such as the H 2 oxidation, 8,9 hydrocarbon isomerization, 10,11 CO oxidation, 12,13 and hydrazine oxidation. 14 Herein, we report that the EPOC effect can also be observed for electrochemical reduction reactions such as the reduction of CO 2 in an aqueous electrolyte solution at ambient temperature.
The reduction of carbon dioxide to produce formic acid is an attractive route to store renewable electricity and an important strategy for the utilization of carbon cycle. 7,[15][16][17][18][19] However, CO 2 is thermodynamically stable and notoriously unreactive; therefore, high reaction temperatures 7,20-23 and high overpotentials [24][25][26][27] are usually essential to activate and transform CO 2 during thermocatalysis and electrocatalysis. Herein, we report a signicant EPOC effect for the CO 2 reduction to produce formate over Pd nanoparticles (NPs) in a 1 M KHCO 3 aqueous solution at ambient temperature. Thermocatalytic and electrocatalytic reduction of CO 2 over Pd nanoparticles (NPs) occur simultaneously and compete with each other, which are promoted by the applied negative potential and H 2 in the feeds, respectively. The shared reaction intermediate, namely HCOO*, is formed over the Pd NPs and is proposed as the origin of the EPOC effect during the thermocatalytic and electrocatalytic reduction of CO 2 . Inspired by the EPOC effect, a Pd/C-Pt/C composite electrode was constructed for the CO 2 reduction, such that the addition of H 2 to the feeds could be avoided. H 2 generated from the electrolysis of water over Pt NPs effectively promotes formate production over Pd NPs.

Results and discussion
Carbon-supported Pd NPs were prepared using sodium citrate as a stabilizing agent and NaBH 4 as a reducing agent. 28 Each sample was named on the basis of the average size of Pd NPs present in it, such as 3.7 nm Pd indicates that Pd NPs in the Pd/C catalyst have an average particle size of 3.7 nm (Fig. S1 †). The catalyst ink, containing Pd/C and Naon ionomer, was deposited on a piece of carbon paper (Toray TGP-H-060) with a microporous layer and dried to serve as a porous electrode for CO 2 reduction in a 20% H 2 /CO 2 -saturated 1 M KHCO 3 solution. To quantify the EPOC effect, the rate enhancement ratio was dened as follows: 3 where r and r 0 are the rates of the promoted (by applying negative potentials) and unpromoted (at open-circuit voltage, OCV) reactions, respectively. In this study, all r values measured under different atmospheres or over different electrodes are calculated by the following equation: r ¼ total amount of formate generated in one hour ðin molÞ total amount of Pd in electrode ðin milligramÞ Â 1 h CO 2 reduction experiments were conducted in an H-cell, as shown in Fig. 1a. The porous electrode coated with Pd/C catalyst was immersed in a 1 M KHCO 3 aqueous solution, and 20% H 2 /CO 2 was fed into the cathode chamber for the CO 2 reduction reaction at OCV and different negative potentials. The maximum temperature increment of the electrolyte solution was 0.8 C during the constant-potential electrolysis for 1 h, and the effect of temperature on the formate production rate at OCV and different negative potentials can be ignored. Fig. 1b shows the rate enhancement ratio for the formate production at different negative potentials compared to that at OCV over differently-sized Pd NPs. There are volcano-like curves for the value of r over differently-sized Pd NPs within the studied potential range. The value of r is 54 over 2.4 nm Pd at À0.1 V and reaches the maximum value of 143 at À0.2 V. Further negatively shiing the potential to À0.3 V and À0.4 V would decrease the rate enhancement ratio to 95 and 39, respectively. The ratio over 3.7 nm Pd increases from 58 to 119 when the potential is shied from À0.1 V to À0.2 V and drops to 100 and 17 at À0.3 V and À0.4 V, respectively. The ratio over 7.8 nm Pd is obviously smaller than that over 2.4 nm Pd and 3.7 nm Pd, and the maximum ratio over 7.8 nm Pd is 23 at À0.2 V. The reaction of adsorbed hydrogen on Pd hydride surface with CO 2 to form adsorbed HCOO* is considered as the rate-determining step for CO 2 reduction. 16 Upon negatively shiing the potential from À0.1 V to À0.4 V, the hydrogen adsorption strength on the Pd hydride surface is weakened due to a favored hydrogen evolution reaction. Therefore, the optimum adsorption strength for surface-adsorbed hydrogen to react with CO 2 occurs at À0.2 V, resulting in the highest r value. Since small-sized Pd NPs prefer to adsorb more hydrogen over coordinatively unsaturated sites, 29 the r value is much higher over 2.4 nm Pd as compared to that over 3.7 nm Pd and 7.8 nm Pd at À0.4 V.
To investigate the origin of the EPOC effect during CO 2 reduction, an isotope-labeling experiment was conducted by replacing 20% H 2 /CO 2 with 20% D 2 /CO 2 , and the products were analyzed by nuclear magnetic resonance (NMR) spectroscopy. PdD x could also be generated under D 2 atmosphere, and the properties of PdD x were close to those of PdH x ; thus, it was suggested that the electrochemical measurements under 20% D 2 /CO 2 atmosphere were comparable to those under 20% H 2 /CO 2 atmosphere. 30-32 HCOO À and DCOO À were quantied by 1 H-NMR and 2 H-NMR spectra, respectively. The amount of D 2 O and HDO in the electrolyte solution aer constant-potential electrolysis at À0.2 V for 1 h was also quantied by 2 H-NMR spectra, which is 0.028%, whereas the natural abundance of D is about 0.015%. 33 Therefore, non-electrochemical exchange between adsorbed D and H + could be ignored, 34 and DCOO À and HCOO À were considered to be produced from the CO 2 + D 2 thermocatalytic reaction and CO 2 electrocatalytic reaction, respectively. Fig. 2 shows the percentage of HCOO À and DCOO À formed over 3.7 nm Pd at different negative potentials. The percentage of DCOO À is higher than that of HCOO À at À0.1 V and À0.2 V, which sharply decreases at À0.3 V and reaches below the detection limit at À0.4 V. Thus, the thermocatalytic and electrocatalytic reduction of CO 2 occur simultaneously and compete with each other when applying negative potentials over Pd NPs.
We also measured the electrocatalytic reduction of CO 2 without adding H 2 to the feeds. As shown in Fig. 3, S2 and S3, † the current density becomes unstable at negatively shied potentials, which is caused by poisoning from trace CO, a minor side product from the CO 2 electroreduction. 16,[35][36][37] With the addition of H 2 , the stability of current density increases, indicating that the electrocatalytic reduction of CO 2 is stabilized. X-ray diffraction (XRD) patterns, of 3.7 nm Pd under different atmosphere were obtained to investigate the active phase of Pd  NPs, as shown in Fig. S4. † XRD pattern of 3.7 nm Pd under 20% N 2 /CO 2 atmosphere matches well with that of Pd (JCPDS 46-1043), and the diffraction peak of the Pd (111) plane is located at 39.9 . When the atmosphere is switched to 20% H 2 /CO 2 , the diffraction peak of the Pd (111) plane quickly shis to 38.8 , accompanied by a shi in all the other peaks. The pattern is consistent with that of PdH 0.706 (JCPDS 18-0951), which is facilely generated under H 2 atmosphere. Since the state and structure of the supported Pd NPs are not affected by water, 38 the PdH x active phase is expected to be stable in 20% H 2 /CO 2 -saturated 1 M KHCO 3 solution. Fig. 3c shows the electric charge ratios over 3.7 nm Pd at various potentials, calculated from the I-t plots, in Fig. 3a and b. At À0.1 V, the electric charge does not change over Pd NPs because surface PdH x is stable and the rate of proton reduction can meet the requirement for the CO 2 electroreduction. At À0.2 V, CO 2 electroreduction is accelerated, and the rate of proton reduction cannot match the rate of CO 2 electroreduction, resulting in a decrease of current density. The enhancement of the electric charge ratio is the highest at À0.3 V since the addition of H 2 effectively stabilizes surface PdH x that tends to decompose through hydrogen evolution. This enhancement of electric charge ratio indicates that the electrocatalysis is to some extent also promoted by H 2 in the feeds, which has not been reported earlier.
The rate of formate production was enhanced when the reaction atmosphere was changed from 20% N 2 /CO 2 to 20% H 2 /CO 2 . Fig. 3c shows the enhancement ratio of r H 2 /r N 2 over 3.7 nm Pd at different negative potentials, where r H 2 and r N 2 represent the formate production rates in 20% H 2 /CO 2 and 20% N 2 /CO 2 -saturated electrolyte solutions, respectively. The enhancement ratio of formate production rates is higher than that of electric charge accumulated in 1 h. For instance, r H 2 at À0.2 V reaches about 1.9 mol formate mg Pd À1 h À1 , and r H 2 /r N 2 is more than twice that of Q H 2 /Q N 2 . The additional improvement is considered as a contribution from the thermocatalytic reaction, 39 also identied in the isotope labeling experiment. The signicant potential dependence of the r H 2 /r N 2 and Q H 2 /Q N 2 values over 3.7 nm Pd is attributed to the instability of Pd hydride and CO poisoning at more negative potentials during the electrocatalytic reduction of CO 2 . Similar enhancement effects over 2.4 nm and 7.8 nm Pd are shown in Fig. S2 and S3. † Since the strong hydrogen adsorption on small-sized Pd NPs could stabilize the surface Pd hydride, the current densities over 2.4 nm Pd were more stable than those over 3.7 nm Pd ( Fig. 3 and S2 †). Therefore, the r H 2 /r N 2 and Q H 2 /Q N 2 values over 2.4 nm Pd are smaller than those over 3.7 nm Pd. The lack of potential dependence of r H 2 /r N 2 and Q H 2 /Q N 2 over 7.8 nm Pd indicates that surface Pd hydride is difficult to form over large-sized Pd NPs, as also conrmed by the unstable current densities shown in Fig. S3. † Based on the abovementioned results, thermocatalytic and electrocatalytic reduction of CO 2 occur simultaneously as follows: [40][41][42] (a) Thermocatalytic reduction reaction Pd + xH 2 (g) 4 H* (3) HCOO* + H 4 HCOOH* HCOOH* 4 HCOOH(l) + * (b) Electrocatalytic reduction reaction CO 2 (g) + * + e À + H + (aq.) 4 HCOO* HCOO* + e À + H + (aq.) 4 HCOOH* + H 2 O(l) HCOOH* 4 HCOOH(l) + * where * and H represent the palladium hydride active phase and the free H atom adsorbed on the catalyst surface, respectively. HCOO* and HCOOH* represent the corresponding intermediate species. It is clear from eqn (4) and (7) that the thermocatalytic and electrocatalytic reduction of CO 2 share the same intermediate species HCOO*. Recent research on hydrazine electrooxidation indicates that the mechanistic origin of the EPOC effect lies in structurally similar activated transition states and/or adsorbed surface intermediates arising from hydrazine oxidation and decomposition. 14 Therefore, the negative electrode potentials could affect and promote the heterogeneous catalytic reduction of CO 2 . Fig. 4 shows the schematic of CO 2 + D 2 reduction over Pd NPs. D 2 is split into D atoms on the surface of Pd NPs, which subsequently diffuse into the Pd lattice to form the PdD x phase. The atomic D on the PdD x surface, derived from D 2 dissociation, reacts with CO 2 to form DCOO À via the thermocatalytic pathway, whereas H + in water reacts with CO 2 to form HCOO À via the electrocatalytic pathway. At À0.1 V and À0.2 V, the electrocatalytic reduction rate is constrained due to the low overpotentials and the percentage of HCOO À is lower than that of DCOO À . Upon increasing the overpotential, the electrocatalytic reduction rate is accelerated, exceeding the thermocatalytic rate, and thus the electrocatalytic reduction pathway dominates for CO 2 reduction at À0.4 V (Fig. 2). H 2 generated from the electrolysis of water is promising for practical applications in CO 2 reduction. Within the potential range for the CO 2 reduction via CO 2 + H 2 reaction, the hydrogen evolution reaction can occur over Pt NPs, which would provide H 2 by the electrolysis of water at the same negative potentials. 43,44 We designed a composite electrode in which 3.7 nm Pd was deposited on the microporous layer, whereas commercial Pt/C catalyst was deposited on the other side of the carbon paper. The cross-sectional scanning electron microscopy (SEM) image and corresponding energy-dispersive X-ray spectroscopy mapping image of the electrode are shown in Fig. 5a and S5. † As illustrated in Fig. 5b, the Pt/C and Pd/C catalyst layers are separated by the carbon paper with a microporous layer, and H 2 generated over Pt NPs diffuses across the carbon paper or the electrolyte solution towards the Pd NPs, which promotes the thermocatalytic and electrocatalytic reduction of CO 2 over Pd NPs. The current density is clearly improved and stabilized aer depositing Pt/C catalyst on the opposite side of the Pd/C catalyst layer ( Fig. 3a and 5c). The activity and selectivity of the Pt/C electrode, used as a control in this study, were also measured, and only H 2 was produced (Fig. S6 †). The enhancement ratio for formate production with the Pd/C-Pt/C composite electrode is shown in Fig. 5d. The maximum value reaches 8.2 at À0.4 V as a result of the combination of thermocatalytic and electrocatalytic reduction of CO 2 .

Conclusions
In summary, a signicant EPOC effect was observed over Pd NPs during CO 2 reduction to generate formate in 1 M KHCO 3 solution at ambient temperature. Both thermocatalytic and electrocatalytic reduction of CO 2 over Pd NPs were promoted by applying negative potentials and by adding H 2 to the feeds, respectively. The shared reaction intermediate HCOO* over Pd NPs was proposed as the origin of the EPOC effect during the thermocatalytic and electrocatalytic reduction of CO 2 . Based on the abovementioned understanding, the Pd/C-Pt/C composite electrode was constructed for CO 2 reduction without the direct addition of H 2 to the feeds. H 2 generated through water electrolysis over Pt NPs effectively promoted the formate production over Pd NPs. The significant rate enhancement ratio for CO 2 reduction not only reveals a new example of EPOC in a low-temperature aqueous electrochemical reaction, but also provides an alternative strategy to promote the electrocatalytic reduction of CO 2 .