Low-temperature catalytic oxidative coupling of methane in an electric field over a Ce–W–O catalyst system

We examined oxidative coupling of methane (OCM) over various Ce–W–O catalysts at 423 K in an electric field. Ce2(WO4)3/CeO2 catalyst showed high OCM activity. In a periodic operation test over Ce2(WO4)3/CeO2 catalyst, C2 selectivity exceeded 60% during three redox cycles. However, Ce2(WO4)3/CeO2 catalyst without the electric field showed low activity, even at 1073 K: CH4 Conv., 6.0%; C2 Sel., 2.1%. A synergetic effect between the Ce2(WO4)3 structure and electric field created the reactive oxygen species for selective oxidation of methane. Results of XAFS, in-situ Raman and periodic operation tests demonstrated that OCM occurred as the lattice oxygen in Ce2(WO4)3 (short W–O bonds in distorted WO4 unit) was consumed. The consumed oxygen was reproduced by a redox mechanism in the electric field.

Because of its stable tetrahedral structure, methane activation requires temperatures higher than 973 K. Furthermore, the reactivity of ethylene is higher than that of methane. Consequently, C 2 selectivity decreases because gas-phase non-selective and sequential oxidation with oxygen to form CO and CO 2 is unavoidable at such high temperatures. Therefore, it is extremely difficult to obtain high C 2 yield with OCM.
To resolve the difficulties described above, we adopted a non-conventional catalytic system, a catalytic reaction in an electric field, in anticipation of methane activation at low temperatures. Results show that various low-temperature catalytic reactions such as methane steam reforming [6][7][8][9][10] can proceed in the electric field. We also reported that OCM proceeded at a low temperature (423 K) in the electric field over Sr-La 2 O 3 (Sr/La = 1/20) catalyst 11,12 . Although the Sr-La 2 O 3 catalyst showed high C 2 selectivity (49.0%), CH 4 conversion remained low (8.9%) in the electric field with imposition of 3.0 mA of current. Further catalyst development is necessary for OCM in an electric field at low temperatures.
Based on the discussion presented above, we investigated application of the electric field to Ce-W-O system catalysts including HPAs for low-temperature methane conversion. Important findings included the following: 1) 40 wt%TBA-PW 12 O 40 /CeO 2 catalyst showed high OCM activity only in the electric field. The structure was changed to Ce 2 (WO 4 ) 3 /CeO 2 . 2) The Ce 2 (WO 4 ) 3 structure, which functioned as an active site structure for selective oxidation of methane, worked only in an electric field. 3) Short W-O bonds of the distorted WO 4 unit in Ce 2 (WO 4 ) 3 was an active site in the electric field; OCM occurred by a redox mechanism.
Next, we evaluated the effects of an electric field on OCM activity over various catalysts including TBA-PW 12 O 40 /CeO 2 catalyst (Supplementary Information Tables S2 and S3). TBA-PW 12 O 40 /CeO 2 catalyst showed high OCM activity (CH 4 Conv., 52.8%; C 2 Sel., 32.0%; C 2 Yield, 16.9% at imposed current 7.0 mA) in the electric field. Bare CeO 2 catalyst showed high CH 4 conversion (28.2%), but CeO 2 contributed to complete oxidation of methane to produce CO 2 with 98% selectivity in the electric field. Results show that the OCM activity of TBA-PW 12 O 40 /CeO 2 was derived from the supported TBA-PW 12 O 40 . However, bare TBA-PW 12 O 40 was unsuitable for imposing the electric field because of its electric conductivity.
As-made and after-reaction samples were characterized with Raman, XRD, and FT-IR (Supplementary Information Figs S1-S3). As-made samples were attributed to TBA-PW 12 O 40 and CeO 2 . After reaction, the peaks corresponding to TBA-PW 12 O 40 disappeared. New peaks, which were attributable to WO 3 and Ce 2 (WO 4 ) 3 , appeared. Therefore, it is likely that these oxides contribute to OCM activity.
To confirm the OCM activity for these oxides without an electric field, OCM over TBA-PW 12 O 40 /CeO 2 catalyst at 573-1073 K without an electric field was conducted after reaction with an electric field (423 K, 3.0 mA) to form the oxides described above ( Supplementary Information Fig. S4). With an electric field, TBA-PW 12 O 40 /CeO 2 catalyst showed the following: CH 4 Conv., 14.1%; C 2 Sel., 44.4%. After turning off the electric field, CH 4 and O 2 conversion and C 2 selectivity decreased dramatically at 1073 K: CH 4 Conv., 2.1%; C 2 Sel., 5.0%. Results show that the formed oxides can produce reactive oxygen species and activate methane only when an electric field is applied.
Clarification of the active site structure. As described in the section above, the supported TBA-PW 12 O 40 on CeO 2 was converted to Ce 2 (WO 4 ) 3 and WO 3 during OCM with an electric field. The formed oxides might play an important role for OCM with an electric field. Therefore, OCM activities over Ce 2 (WO 4 ) 3 /CeO 2 , WO 3 / CeO 2 , Ce 2 (WO 4 ) 3 , and WO 3 catalysts were investigated in the electric field. Structures of the supported and the unsupported oxide catalysts were confirmed using Raman and XRD measurements ( Fig. 1(a) and Supplementary  Information Fig. S5(a)). Results of activity tests over these oxides at 423 K with an electric field are presented in Table 1, which shows that Ce 2 (WO 4 ) 3 /CeO 2 , WO 3 /CeO 2 and Ce 2 (WO 4 ) 3 showed OCM activity, although WO 3 showed no OCM activity. Actually, Ce 2 (WO 4 ) 3 /CeO 2 showed the following for T tc of 649 K: CH 4 Conv., 13.6%; C 2 Sel., 39.0%. For T tc of 634 K, WO 3 /CeO 2 showed the following: CH 4 Conv., 14.3%; C 2 Sel., 32.4%. Also, Ce 2 (WO 4 ) 3 showed the following for T tc of 659 K: CH 4 Conv., 9.7%; C 2 Sel.: 41.2%. Additionally, WO 3 catalyst showed no OCM activity even at almost the same temperature (around 650 K) or at almost the same input power (around 2.5 W) as the other catalysts (Supplementary Information Tables S4 and S5). Ce 2 (WO 4 ) 3 /CeO 2 catalyst showed the higher OCM activity than TBA-PW 12 /CeO 2 catalyst at almost the same input power (2.6-2.7 W) (Supplementary Information Table S4).
The structures of Ce 2 (WO 4 ) 3 /CeO 2 , WO 3 /CeO 2 , Ce 2 (WO 4 ) 3 , and WO 3 catalysts before and after reaction with electric field were analyzed using Raman spectroscopy and XRD ( Fig. 1(b) and Supplementary Information Fig. S5(b)). As shown in Fig. 1(b), the structures of Ce 2 (WO 4 ) 3 /CeO 2 , Ce 2 (WO 4 ) 3 , and WO 3 after reaction with the electric field were not markedly different from as-made. However, the spectrum of WO 3 /CeO 2 after reaction with an electric field differed considerably from as-made. The Raman spectrum showed that W species in WO 3 /CeO 2 changed to Ce 2 (WO 4 ) 3 from WO 3 after reaction in the electric field. Supplementary Information Fig. S5 (XRD patterns) presents similar results. It is conceivable that the formation of Ce 2 (WO 4 ) 3 proceeded as a solid-solid reaction between CeO 2 and WO 3 in the reaction with an electric field 32 . In light of the activity test over WO 3 /CeO 2 , one can infer that the Ce 2 (WO 4 ) 3 formed on WO 3 /CeO 2 activated methane and that it showed OCM activity. Therefore, activity tests and characterizations demonstrated that the active site structure for OCM with the electric field was Ce 2 (WO 4 ) 3 . Moreover, the active site structure, Ce 2 (WO 4 ) 3 , showed stable OCM activity for at least one hour ( Supplementary Information Fig. S6).

Contribution of the active site to OCM activity.
To elucidate the contribution of the active site to OCM activity, OCM at 573-1073 K over Ce 2 (WO 4 ) 3 /CeO 2 catalyst was conducted without an electric field. Results of activity tests (573-1073 K) are presented in Table 2. In the reaction at 1073 K without the electric field, the catalyst showed results (CH 4 Conv., 6.0%; C 2 Sel., 2.1%) that were much lower than those in the reaction with the electric field at external temperature of 423 K (T tc measured catalyst bed temperature by thermocouple: 649 K): CH 4 Conv., 13.6%; C 2 Sel., 39.0%. These results demonstrate that Ce 2 (WO 4 ) 3 /CeO 2 catalyst can produce reactive oxygen species and activate methane only when an electric field is applied.
In the reaction with an electric field, the catalyst bed temperature increased by Joule heating. Therefore, to elucidate the influence of Joule heating on OCM in the gas phase, the temperature dependency of OCM in the electric field was evaluated. Table 3 presents the activities of OCM over Ce 2 (WO 4 ) 3 /CeO 2 at 423, 673, and 873 K with the electric field (3.0 mA). C 2 selectivity and C 2 yield decreased in association with increasing furnace temperature. This reduction caused by combustion of C 2 species with O 2 in gas phase because O 2 conversion increased in proportion to increasing temperature. Therefore, OCM in the gas phase is not promoted by Joule heating from the electric field. The effect of Joule heating by an electric field is unimportant in the system. In other words, because the catalyst was able to activate methane at a low gas-phase temperature, C 2 selectivity was high in the OCM with the electric field at low temperature. The same trend was obtained in the case of electric power fixing to normalize the electric factor (Supplementary Information Table S6).   Next, to clarify the reaction mechanism, the influence of contact time (W/F CH4 ) on OCM activity over Ce 2 (WO 4 ) 3 /CeO 2 in the electric field was investigated. As presented in Fig. 2, CH 4 conversion and O 2 conversion increased and C 2 selectivity decreased concomitantly with increasing contact time. The decrease of C 2 selectivity resulted from combustion of C 2 hydrocarbons with O 2 in gas phase. Figure 3 shows the relation between CH 4 conversion and C 2 H 4 , C 2 H 6 , and C 2 H 2 selectivity over Ce 2 (WO 4 ) 3 /CeO 2 in the electric field (423 K, 3.0 mA). As shown in Fig. 3, as CH 4 conversion approaches 0%, C 2 H 4 and C 2 H 6 selectivity increase and C 2 H 2 selectivity decreases. In the range of very low CH 4 conversion, production of C 2 H 4 and C 2 H 6 were the main reactions. One can infer that methyl radical and carbene were produced on the catalyst surface in the electric field. Accordingly, Ce 2 (WO 4 ) 3 can extract one or two H atoms from CH 4 and produced methyl radical and carbene. However, C 2 H 2 selectivity increased concomitantly with increased CH 4 conversion up to 15% and then decreased. These results suggest that C 2 H 2 was generated through oxidative dehydrogenation of C 2 H 4 or through coupling of CH species, which was formed from CH 4 with electric energy 33 ; then it was oxidized to CO and CO 2 .
To confirm that reactive oxygen species are formed on the catalyst surface in the electric field, the periodic operation test at 473 K over Ce 2 (WO 4 ) 3 /CeO 2 catalyst was conducted with the electric field. Results of activity tests are presented in Table 4. The results were the activities at 2 min after from methane+ Ar supply. Table 4 shows that the C 2 selectivity was higher than 60% and that it was maintained during three cycles. When increasing the CH 4 conversion by increasing the contact time, C 2 selectivity was higher than 65% and was maintained during three cycles. However, in the periodic operation test without an electric field at 1073 K, C 2 selectivity was very low; CO and CO 2 were mainly formed ( Supplementary Information Table S7). These results demonstrate that reactive oxygen species suitable for OCM were formed on the catalyst surface in the electric field. The synergic effect of Ce 2 (WO 4 ) 3 and electric field created the reactive oxygen species for selective oxidation of methane and activated methane because the active site structure for OCM with the electric field was Ce 2 (WO 4 ) 3 .

In-situ Raman over Ce 2 (WO 4 ) 3 /CeO 2 with and without electric field. Many researchers have
reported that Na 2 WO 4 -Mn 2 O 3 /SiO 2 catalyst has high OCM activity 2,34 . A short W-O bond in the distorted WO 4 unit is proposed as the active site of Na 2 WO 4 -Mn 2 O 3 /SiO 2 . This W-O bond is observed at 927 cm −1 in Raman spectrum 35 . Specific examination of the Ce 2 (WO 4 ) 3 structure reveals that it has WO 4 units of two kinds. Both are physically distorted. It is likely that W-O bonds in Ce 2 (WO 4 ) 3 /CeO 2 work as an active site for OCM in the electric field. Comparison of Ce 2 (WO 4 ) 3 with Na 2 WO 4 reveals that Ce 2 (WO 4 ) 3 has a stable structure and short W-O bonds in the distorted WO 4 unit, which are observed at 949 cm −1 and 931 cm −1 32 . We conducted in-situ Raman measurements over Ce 2 (WO 4 ) 3 /CeO 2 in the electric field to clarify the W-O bond behavior in the electric field. Figure 4 portrays Raman spectra with and without the electric field. The peak positions of the short W-O bonds   in distorted WO 4 unit are summarized in Supplementary Information Table S8. Figure 4 shows that the spectra of Ce 2 (WO 4 ) 3 /CeO 2 with an electric field (Fig. 4c-e) differed from inert (Fig. 4a) and the peak of W-O bonds shifted to a lower wavenumber and broadened. From Supplementary Information Table S8, the peak shift of W-O peaks in the electric field was about 7-10 cm −1 . The shift and broadening of the W-O peaks were attributed to the electric field because the spectrum without electric field after imposing electric field in CH 4 + O 2 flow (Fig. 4f) was almost identical to that of the spectrum without an electric field under inert (Fig. 4a). These results suggest that the Ce 2 (WO 4 ) 3 /CeO 2 structure was distorted and that the W-O bonds were weakened by the electric field. It acted as the reactive oxygen suitable for selective oxidation of methane to C 2 hydrocarbons. For comparison, the Raman spectrum over Ce 2 (WO 4 ) 3 /CeO 2 at 603-703 K without an electric field (Fig. 4b) was measured because the catalyst bed temperature was increased by Joule heating in the electric field. The shift of W-O peaks was about 2-3 cm −1 , indicating a small effect of Joule heat from the electric field on W-O bond activation. These results imply that W-O bond activation is attributable mainly to the electric field and that the W-O bond activated by the electric field functioned as an active site for OCM in the electric field. Next, XAFS spectra were recorded. The results are presented in Supplementary Information Figs S7 and S8 and Table S9. Regarding the coordination number of W-O bonds in Ce 2 (WO 4 ) 3 from XAFS measurements, the catalyst after 1 cycle in periodic operation test in the electric field showed a smaller coordination number than in the as-made material and after O 2 supply in the periodic operation test in the electric field. This result supports our inference that surface oxygen of Ce 2 (WO 4 ) 3 activated by the electric field was consumed by methane. Therefore, results of XAFS, in-situ Raman, and periodic operation tests demonstrated that OCM which occurred as lattice oxygen in Ce 2 (WO 4 ) 3 (short W-O bonds in distorted WO 4 unit) was consumed and reproduced by the redox mechanism. A possible reaction mechanism is described in Supplementary Information Fig. S9. In addition, redox reaction of Ce cation (Ce 3+ ↔ Ce 4+ + e -) might be also responsible for the OCM reaction.

Conclusion
Oxidative coupling of methane (OCM) over 40 wt%TBA-PW 12 O 40 /CeO 2 in the electric field was conducted. The catalyst showed high OCM activity, although the catalyst showed extremely low OCM activity, even at the high temperature of 1073 K without an electric field. After reaction with the electric field, Raman spectra confirmed that the structure of TBA-PW 12 O 40 /CeO 2 catalyst was changed to Ce 2 (WO 4 ) 3 /CeO 2 . Then, OCM activities over Ce 2 (WO 4 ) 3 /CeO 2 , WO 3 /CeO 2 , Ce 2 (WO 4 ) 3 , and WO 3 catalysts in an electric field (3.0 mA) were investigated to clarify the structure of the active site. Ce 2 (WO 4 ) 3 /CeO 2 , WO 3 /CeO 2 and Ce 2 (WO 4 ) 3 catalysts showed OCM activity; WO 3 showed no OCM activity. After reaction with the electric field, Raman spectra showed that W species in WO 3 /CeO 2 changed to Ce 2 (WO 4 ) 3 from WO 3 . Therefore, the active site structure for OCM with the electric field was Ce 2 (WO 4 ) 3 . Also, Ce 2 (WO 4 ) 3 /CeO 2 catalyst showed extremely low OCM activity at 1073 K. In the periodic operation test over Ce 2 (WO 4 ) 3 /CeO 2 catalyst, C 2 selectivity was higher than 60% and was maintained during three cycles. Therefore, synergic effects of Ce 2 (WO 4 ) 3 and the electric field created the reactive oxygen species suitable for selective oxidation of methane, activated methane, and progressed OCM only when an electric field was applied.
Results of XAFS, in-situ Raman and periodic operation test demonstrated that OCM occurred using the lattice oxygen of Ce 2 (WO 4 ) 3 (short W-O bonds in distorted WO 4 unit), which were consumed and reproduced by the redox mechanism in the electric field.
Keggin-type TBA-HPAs supported on CeO 2 (JRC-CEO-1) catalysts were prepared by impregnation with acetone as the impregnation solvent. The loading amount of TBA-HPAs was 40 wt%. First, acetone (30 mL) and CeO 2 (0.6 g) were added to a 300 mL eggplant flask and were stirred for 2 h using a rotary evaporator. Subsequently, TBA-HPAs (0.4 g) dissolved into acetone (10 mL) were added to the flask and were stirred for 2 h again. The resulting suspension was dried. Then the resulting solid was dried overnight at 393 K. WO 3 /CeO 2 catalyst containing 11.9 wt% W was prepared using impregnation with water as the impregnation solvent, as described in a previous report 32 . An ammonium metatungstate hydrate ((NH 4 ) 6 H 2 W 12 O 40 ·H 2 O) was used as a precursor. After impregnation, the resulting suspension was dried. Then the resulting solid was dried at 393 K overnight, followed by calcination for 3 h in air at 773 K under a ramping rate of 0.5 K min −1 .
Ce 2 (WO 4 ) 3 /CeO 2 catalyst containing 11.9 wt% W was prepared by impregnating CeO 2 with an aqueous solution of ammonium metatungstate hydrate using a similar method to that for WO 3 /CeO 2 , except that the calcination temperature was 1173 K 32 . As a reference, unsupported Ce 2 (WO 4 ) 3 was prepared using a complex method combining ethylenediamine tetraacetic acid and citrate ions, as described in previous reports 40,41 . Activity test. Catalytic activity tests were conducted with a fixed bed flow-type reactor equipped with a quartz tube (4.0 mm i.d.). A schematic image of the reaction system is presented in Supplementary Information  Fig. S10. The catalyst was sieved into 355-500 μ m. Then 100 mg of it was charged in the reactor. The reactant feed gases were methane, oxygen, and Ar (CH 4 : O 2 : Ar = 25: 15: 60, total flow rate: 100 SCCM). The effect of contact time (W/F CH4 ) was investigated by changing the total flow rate. The standard W/F CH4 was 1.6 g cat h mol −1 . For the reaction in the electric field, two stainless steel electrodes (2.0 mm o.d.) were inserted contiguously into the catalyst bed in the reactor. The electric field was controlled using a constant current (3, 5, or 7 mA) with a DC power supply. The imposed voltage depended on the electric properties of the catalyst. Current and voltage profiles were measured using an oscilloscope (TDS 3052B; Tektronix Inc.). The reactor temperature was set to 423 K to avoid the condensation of water produced by the reactions, except for reactions that used no electric field. Product gases after passing a cold trap were analyzed using GC-FID (GC-14B; Shimadzu Corp.) with a Porapak N packed column and methanizer (Ru/Al 2 O 3 catalyst), and using a GC-TCD (GC-2014; Shimadzu Corp.) with a molecular sieve 5A packed column. The respective calculation formulae for conversion, C 2 yield, C 2 selectivity, and Faradaic number in this study are shown below (eqs 2-6). A periodic operation test was conducted to elucidate surface active species on the catalyst in the following steps. In the first step, oxygen and Ar were supplied to the reactor with an electric field for 10 min for oxidation of the catalyst surface. For the second step, residual oxygen in the gas phase of the reactor was removed with Ar purge for 5 min. For the third step, methane and Ar were supplied to the reactor with an electric field for 12 min to evaluate the oxidation catalysis of the surface oxygen species on the catalyst. As the final step, Ar purge was conducted for 20 min to remove all residual gases. The steps described above were repeated for three cycles. Product gases were analyzed at 5 min after oxygen+ Ar supply, and at 2 and 12 min after methane + Ar supply (CO x and desorbed CH 4 were detected at 5 min after from oxygen+ Ar supply and no products were detected at 12 min after Scientific RepoRts | 6:25154 | DOI: 10.1038/srep25154 from methane + Ar supply). Gas flow was O 2 : Ar = 1: 10, total 55 mL min −1 (for oxidation of the catalyst surface) and CH 4 : Ar = 1: 10 or 1: 2, total 55 or 75 mL min −1 (for oxidation of supplied methane by surface oxygen species). The reactor temperature was fixed at 473 K. The imposed current was set at 3.0 or 7.0 mA.

Characterization of catalyst.
FT-IR spectra were recorded on a spectrometer (FT-IR/6200; Jasco Corp.) using a KBr pelletizing method. Raman spectra were recorded on a Raman spectrometer (excitation line λ = 532 nm, NRS-1000; Jasco Corp.). The crystalline structure was characterized using powder X-ray diffraction (XRD, RINT-Ultima III; Rigaku Corp.) operating at 40 kV and 40 mA with Cu-Kα radiation. The specific surface area of the catalyst was measured using N 2 adsorption using the BET method (Gemini VII; Micromeritics Instrument Corp.) after pre-treatment at 473 K in N 2 atmosphere for 2 h. Results of BET measurements are presented in Supplementary Information Table S10. W L 3 -edge X-ray absorption fine structure (XAFS) spectra were recorded on BL14B2 in SPring-8 (Hyogo, Japan). Catalysts treated in the reaction condition were ground into powder and were pressed into pellets. Then, pellets were packed into gas-barrier bags. The pellets were diluted with BN to adjust for XAFS measurement. EXAFS analysis and curve fitting were performed using software (Athena ver. 0.8.056; Artemis ver. 0.8.012).
In-situ Raman. In-situ Raman measurements in the electric field were conducted using a Raman spectrometer with a hand-made glass reactor and gold wire electrodes (see Supplementary Information Fig. S11). The reactant feed gases were supplied with canned standard gases (CH 4 (0.995%) + Ar and/or O 2 (99.9%)). The electric field was imposed using a constant current at 6.0 mA.
Raman spectra for catalyst heated at reaction temperature (603-703 K) without an electric field were also observed to elucidate the effect of Joule heating on the catalyst structure. A Ni-Cr wire was inserted into the sample for heating by resistance heating.