Surface Protonics Promotes Catalysis

Catalytic steam reforming of methane for hydrogen production proceeds even at 473 K over 1 wt% Pd/CeO2 catalyst in an electric field, thanks to the surface protonics. Kinetic analyses demonstrated the synergetic effect between catalytic reaction and electric field, revealing strengthened water pressure dependence of the reaction rate when applying an electric field, with one-third the apparent activation energy at the lower reaction temperature range. Operando–IR measurements revealed that proton conduction via adsorbed water on the catalyst surface occurred during electric field application. Methane was activated by proton collision at the Pd–CeO2 interface, based on the inverse kinetic isotope effect. Proton conduction on the catalyst surface plays an important role in methane activation at low temperature. This report is the first describing promotion of the catalytic reaction by surface protonics.

same trends as those reported from earlier studies, showing that the order for methane partial pressure is almost 1 and that for water is 0. From these results, we inferred that the mechanism of SR with our prepared Pd catalyst showed the same mechanism as that reported, indicating that the rate-determining step is the methane dissociative adsorption step.

Calculation of rate constant k for SR and ER
To evaluate the apparent activation energy (Ea), we calculated rate constant (k). The reaction rate equations are estimated as equations (1)-(3) from our kinetic analyses, as shown in Tables S1 and   Table S3. For SR, we ascertained the reaction rate equation as equation (2) because our measurements for the dependence of partial pressure were almost identical around 623-723 K. The orders for methane and water pressure were average values. However, for ER, we extracted the increased reaction rate (rER) with equation (1). The apparent activation energy was calculated using kER.
(1) rSR = kSR PCH4 0.27 + PH2O 0. 26 (2) Our kinetic analyses for rER revealed that the methane pressure dependence of (α') rER was approximately 0, and that the water pressure dependence (β') changed with the reaction temperature, as presented in Figure S4. Therefore, we used β' at each reaction temperature from measured values and estimated values. The estimated values were obtained as an average from measured values. The obtained orders for β' are presented in Table S4.

Operando-DRIFTS spectra and analyses for products via SR and ER under various conditions
Operando-DRIFTS spectra before subtracting after ER spectrum at 473 K with Pd/CeO2 or CeO2 catalyst are shown in Figure S5. The assignments are presented in Table S5. Furthermore, operando-DRIFTS spectra over Pd/CeO2 catalyst with H2O/D2O or without water (only CH4) are shown in Figure S6. The operando-DRIFTS spectra over Pd/CeO2 catalyst with CH4 and H2O/D2O at various temperature are presented in Figures S7 and S8. Results of analyses of products are presented in Tables   S6 and S7.

Calculation of O-H bond energy from wavenumbers of IR spectrum
As shown in Figure 2(A) and Figure 2(C), the red-shift of O-H stretching peak was observed by the application of an electric field of 3699 cm -1 to 3675 cm -1 or from 3649 cm -1 to 3627 cm -1 . Using these wavenumbers and following equations (4) and (5)

Inverse Kinetic Isotope Effect (Inverse KIE)
KIE is defined as the ratio of kD/kH, where kD is the rate constant of the reaction with D2O, and where kH is the rate constant of the reaction with H2O. With this conversion, KIEs greater than unity are called "inverse." Those less than unity are called "normal" (33). As Figure 3 shows, the decrease of gas phase methane was greater with D2O, than with H2O. Assuming that kD/kH represents the ratio of the gas phase methane decrease, our results show that inverse KIE was observed around 473-573 K during ER, presented in Figure S9. These results strongly support our assumption that methane is activated by proton collision derived from the Grotthuss mechanism.

TOF determined by Pd specific surface area (TOF-s) and Pd perimeter (TOF-p)
To evaluate the influence of the number of Pd active site or the length of Pd interface on the activities for ER at 473 K and SR at 673 K, we prepared different amounts of Pd-loaded catalyst and conducted operando-DRIFTS measurements with this catalyst. Operando-DRIFTS spectra are presented in Figures S10 and S11, and the analyses for products are presented in Table S8. Using these results and the result of CO pulse, presented in Table S9, TOF-s and TOF-p were calculated. To evaluate the particle size of Pd on CeO2 correctly, the influence of CO adsorption onto CeO2 was examined first. Figure S12 shows the results of CO pulse for CeO2 and 1.0wt%Pd/CeO2 catalyst. As Figure S12 shows, the GC intensity of CO for CeO2 at first time and the saturated GC intensity of CO for 1.0wt%Pd/CeO2 were almost the same. Figure S13 shows IR spectra recorded 1 h after supplying nearly 40% CO flow with Ar at 323 K for CeO2 and 1.0wt%Pd/CeO2 catalyst. After purged with Ar for 1 h, the adsorbed CO remained only on Pd loaded catalyst. From these results, we concluded that the influence of CO adsorption onto CeO2 was negligible when using CO pulse dosing.
As shown in Table S9, the particle size of Pd was almost the same for 0.5wt% and 1.0wt% Pd loaded catalysts, about 1.0 nm. And the particle size became larger than 1.0 nm for more than 1.0wt% AC Impedance measurements for evaluating proton conduction via adsorbed water on surface of CeO2 To evaluate the electrical properties of CeO2, we conducted AC impedance measurements. The results of characterization (XRD and SEM) are shown in Figure S14. The phase of CeO2 was cubic, and the average particle size of bulk was 110 nm. The relative density of CeO2 disc was 61%, so it has a kind of pores, as shown in SEM images. Figure S15 presents the examples of Nyquist plots and fitting results. At temperatures from 423 to 673 K, all Nyquist plots obtained under wet condition (PH2O = 0.026 atm) showed smaller arcs rather than those under dry condition. Therefore we considered the parallel equivalent circuit, as shown in Figure S16, where R stands for the resistance, CPE represents the capacitance with constant phase element, and b, gb, surf. b and surf.gb respectively correspond to a bulk, a grain boundary, a surface of bulk, and a surface of grain boundary. The disc of CeO2 has some pores in itself, the surface conduction via adsorbed water on bulk and grain boundary could be observed. First, the data under dry condition were fitted, and we obtained R, C (Capacitance), and CPE values respectively. Then these parameters under dry condition were kept fixed, and the data under wet condition were fitted to obtain surface R, C and CPE values respectively. Also, we can exclude the component of electrodes with Nyquist plots and equivalent circuit in case of necessity. Figure S17 shows the temperature dependency of conductivity under dry and wet conditions. The results under dry condition at high temperatures (573 < T < 773 K) showed the typical temperature dependency of CeO2 on conductivity with mixed (ionic and electronic) conduction (38)(39)(40). The apparent activation energies under high temperature region were 1.19 eV for grain boundary and 1.30 eV for bulk. However the apparent activation energies decreased to 0.36 eV for grain boundary and 0.27 eV for bulk respectively even under dry condition at lower temperatures (423 < T < 573 K). These results are considered to reflect the mobility of lattice oxygen. It is reported that the lattice oxygen of CeO2 starts to move around over 600 K (41). Therefore, at low temperatures, the electron conduction is dominant and the barrier for electron hopping is considered to be relatively low (40). The results under wet condition (surf. b and  surf.gb) showed higher conductivity compared to those (b and gb) under dry. These two conductivity is considered to present proton conductivity, especially via adsorbed water onto the bulk and grain boundary surface, because surf. b and  surf.gb became larger with lowering temperature at lower temperatures (T < 573 K for surf. b and T < 448 K for  surf.gb). These phenomena was studied, for example with TiO2 (42), and the reason conductivity increases under lower temperatures is considered to be related with increasing the amount of adsorbed water on the surface of bulk and grain boundary according to Grotthuss mechanism: proton hopping. From these results, it is revealed that the surface protoics could be occurred via adsorbed water onto CeO2. .

In-situ XAFS measurements for evaluating the electronic state and structure of Pd/CeO2 catalyst during ER
To evaluate the change of electronic state and structure for Pd/CeO2 catalyst, we conducted in-situ XAFS measurements at SPring-8, Hyogo in Japan using an in-situ cell as shown in figure S18. Figures S19 and S20 show the obtained XAFS spectra with an electric field for the Pd-K edge and Ce-K edge at 473 K. As presented in Figure S19(a), Pd was slightly oxidized with the supply of raw materials: CH4 and H2O. Pd was then reduced with application of an electric field because the reaction proceeded and hydrogen was produced. However, no drastic change of the electronic state for Pd was confirmed, suggesting that the main effect of the electric field is not the change for work function of loaded metal (Pd) on catalyst. The EXAFS spectra shown in Figure S19 Figure S12 The results of CO pulse for CeO2 and 1.0wt% Pd/CeO2 catalyst at 323 K.  Figure S17 Temperature dependency of conductivity under dry and wet (PH2O = 0.026 atm) conditions.  Intensity / arb. unit R / ang.