Plasma-assisted CO2 methanation: effects on the low-temperature activity of an Ni–Ce catalyst and reaction performance

Ni–Ce three-dimensional material with macropore diameter of 146.6 ± 8.4 nm was synthesized and used as a methanation catalyst. Firstly, H2 reduction of the catalyst was conducted in the thermal fixed bed and plasma reactor, respectively, then X-ray diffraction (XRD) and CO2 temperature programmed desorption experiments on the two reduced samples were carried out to reveal the plasma effect on the catalyst's physico-chemical properties. It was found that plasma reduction created more abundant basic sites for CO2 adsorption, in particular the medium basic sites were even doubled compared with the thermal-reduced catalysts. The plasma-reduced catalyst exhibited excellent low-temperature activity, ca 50–60°C lower than the thermal catalyst (the maximum CO2 conversion point). Based on the optimum reduced catalyst, plasma effect in the reactor level was further investigated under high gas hour space velocity of approximately 50 000 h−1. The plasma reactor showed higher CO2 conversion capacity and efficiency than the thermal reactor.

few of the methanation technologies reached the commercial level [1]. Due to the increasing demand to store the surplus renewable power, the concept of CO 2 methanation coupled with water electrolysis, known as power to gas (PtG), was proposed by German in 2009. This concept revived the ancient technology and brought new attention to it [1,3]: Apparently, reaction (1) is a highly exothermic and volume contrast reaction, which means pressure and temperature significantly influence the reaction equilibrium. From the thermodynamic viewpoint, it is desirable to operate the reaction at low temperature to achieve high CO 2 conversion. However, low temperature also means slow reaction kinetics and when the temperature is below 250°C, the conversion is negligible. At temperature above 450°C, the CH 4 yield decreases significantly with the CO rapid formation as a by-product through the reversed water gas shift reaction route [4][5][6]. Also, high temperature usually leads to the sinter of the catalyst.
In recent years, numerous studies have been conducted on the catalysts of CO 2 methanation. Among all the catalysts, Ni-based catalysts are the most widely investigated material catalysing the hydrogenation of CO 2 to methane due to their relatively high activity and low cost. These studies focused on improving the activity, stability and the resistance to carbon deposition of the catalysts by adding promoters, loading on different supports and modifying catalyst preparation methods [1][2][3][7][8][9][10][11][12][13][14]. Ceria is often employed as a promoter to improve Ni nanoparticles dispersion and decrease the size; besides, it can also work as a single support. Oxygen vacancies in ceria caused by the existence of Ce 3+ /Ce 4+ ion pairs lead to efficient activation of CO 2 [7,8,12]. Nie et al. [8] found that the addition of CeO 2 could enhance the catalyst activity by lowering Ni particle size and generating oxygen vacancies. The equilibrium conversion can be achieved over Ni-Ce supported γ-Al 2 O 3 catalyst at 300°C with optimal 3wt% CeO 2 addition.
Plasma known as a partially ionized state consists of many active species such as electrons, ions and radicals. Non-thermal plasma (NTP) can generate high energetic electrons (1-10 eV), while the bulk temperature can keep as low as room temperature [15][16][17]. Among the NTP technologies, the dielectric barrier discharge (DBD) plasma is convenient to couple with the catalyst bed. The combination of catalyst with DBD plasma could initiate chemical reactions such as ionization, excitation and dissociation at a lower temperature and may also have an additional effect on the physico-chemical properties of the catalyst. Nizio et al. [6,18] reported methanation with 80% CO 2 conversion and nearly 100% CH 4 selectivity under the assistance of plasma. Wang et al. [19] prepared a highly dispersed Co nanoparticle catalyst and created more active sites using a plasma-assisted method. Besides, the quick on-off switch property of DBD methanation equipment is suitable for the PtG system.
In this work, the 3DOM Ni-Ce catalyst was successfully synthesized and a comparative study for the catalytic performance of the 3DOM catalyst was evaluated both in the DBD plasma reactor and thermal fixed-bed reactor. This well-fabricated catalyst exhibited an excellent catalytic activity and high selectivity for methane under high gas hour space velocity (GHSV) with the assistance of plasma.

Catalyst preparation
All reagents purchased from Sinopharm Chemical Reagent Co., Ltd. are analytical-grade, apart from the methyl methacrylate which is chemical-grade, and they were used as received without further purification. The 3DOM Ni-Ce catalyst was synthesized through Ni and Ce precursor solutions coimpregnating on the poly-methyl methacrylate (PMMA) colloidal crystal template.

Synthesis of PMMA colloidal crystal template
Monodisperse PMMA spheres were prepared via a soap-free emulsion polymerization method. Typically, the polymerization took place in a four-neck round-bottom flask equipped with a thermometer, an inlet for N 2 feeding, a water-cooling condenser and a mechanical stirring paddle. First, 580 ml deionized water and royalsocietypublishing.org/journal/rsos R. Soc. open sci. 6: 190750 24.48 g methyl methacrylate were added into the flask. The solution was heated to 70°C by a hot water bath under constant stirring at 350 r.p.m. and N 2 bubbling. In a separate 50 ml beaker, 0.20 g potassium persulfate (KPS) was dissolved in 20 ml deionized water and then also heated to 70°C. After N 2 bubbling to remove oxygen for at least 30 min, the preheated KPS solution was added to the flask, and then the polymerization was allowed to run at 70°C for 65 min under constant stirring and N 2 bubbling. The resulting monodisperse PMMA microsphere emulsion was cooled to room temperature and then centrifuged at 2200 r.p.m. for 6 h. After abandoning the supernatant liquor, the solid matter was dried at 40°C, and the highly ordered PMMA colloidal crystal template was obtained [20,22,27,28].

Synthesis of 3DOM Ni-Ce catalyst
Prior to the impregnation, the PMMA template was heated to 100°C for 5 min in the oven to increase the mechanical strength and eliminate the remaining MMA in the template. Typically, 5.04 g Ce(NO 3 ) 3 ·6H 2 O, and 3.87 g Ni(NO 3 ) 2 ·6H 2 O (molar ratio 1.15 : 1) were dissolved in 25.0 ml 95% ethanol solution with the addition of equimolar citric acid (5.30 g) as a complexing agent. PMMA template was impregnated with a continuous droplet of the Ni-Ce precursors on a Buchner funnel under vacuum condition. Then, the infiltrated template was removed from the funnel and transferred to a vacuum oven. It was dried at 40°f or 4 h and then dried at 60°C for another 4 h. These impregnation and drying steps were repeated four times to fill the template uniformly and completely. The hybrids were heated to 300°C at a ramp rate of 1°C min −1 , and held at 300°C for 1 h. Then, the precursor was heated again at 1°C min −1 to 550°C and calcined for 6 h. After cooling down, the 3DOM Ni-Ce catalyst was obtained [21,22,25,30].

Catalyst test and analysis
2.2.1. Experimental set-up and procedure CO 2 methanation over a 3DOM Ni-Ce catalyst were evaluated in two reactors: a DBD plasma reactor and a thermal fixed-bed reactor. The schematic diagram of the DBD plasma reactor is shown in figure 1. The DBD plasma reactor consists of a quartz tube as the dielectric material, stainless rod as the inner electrode linked to the cover and ground, stainless net outside the tube as the high-voltage electrode. The plasma discharges in the tube gap between the electrodes. The 0.375 g catalyst (30-60 mesh) diluted with 2.00 g quartz sand of the same size were loading in the plasma zone. The voltage and current applied to the plasma system is 55-60 V and 1.55-1.60 A during methanation reaction, respectively, and the output voltage is ca 20 kV.
The thermal catalytic methanation of CO 2 was performed in a vertical continuous flow fixed-bed quartz tube surrounded by a temperature-controlled electrical oven. The same amount of catalysts and  royalsocietypublishing.org/journal/rsos R. Soc. open sci. 6: 190750 quartz sand were placed in the middle of the tube and furnace. The reaction temperature was measured by a K-type thermocouple penetrated into the catalyst bed.
The gas injection was the same for the two reactors controlled by the calibrated mass flow controller (MFC). A gas mixture of H 2 and CO 2 with a molar ratio of 4 : 1 was fed into the reactor for methanation reaction. The effluent gases were cooled down in a condenser at −5.0°C and passed through a drier to remove the water. Then, the dried gases were analysed by an online gas chromatograph (GC) equipped with TCD and the flow rate was also monitored by a bubble flowmeter. The influence factor for each component was calibrated by two standard gas mixtures through the external standard method. The overall mass balance on carbon basis was more than 98%. Thus, the conversion of CO 2 (X CO2 ) and the selectivity of CH 4 (S CH4 ) and CO (S CO ) were calculated as follows: where F i were the molar fraction of i component (i = CH 4 , CO and CO 2 ) in the effluent gases.

Catalyst reduction method
Prior to the catalytic methanation, the catalysts need to be reduced to the active phases under H 2 atmosphere. Two reduction methods were investigated for comparison: plasma reduction and thermal reduction. The plasma-assisted reduction was conducted in the DBD reactor under 100 ml min −1 hydrogen and plasma atmosphere with 70 V and 1.70 A electrical input for 30 min. The thermal reduction was performed in the thermal fixed-bed reactor under 100 ml min −1 H 2 , and three temperatures (350, 400 and 450°C) were chosen for comparison with a retention time of 30-60 min.

Characterization
Scanning electron microscope (SEM) analyses of 3DOM catalyst were performed on a HITACHI SU8010 operating at 10.0 kV to investigate the morphology and structure of the catalyst. N 2 isothermal adsorption-desorption analysis was conducted on a Micromeritics ASAP 2020 Sorptometer at −196°C. Prior to the analysis, the sample was degassed at 200°C for 6 h with a ramping rate of 10°C min −1 . The specific surface area was calculated by the Brunauer-Emmett-Teller (BET) method. The sample had a macroporous structure and the Barrett-Joyner-Halenda (BJH) method is only appropriate for mesoporous materials. The density functional theory (DFT) method was employed to obtain the pore size distribution.
Powder X-ray diffraction (XRD) patterns of catalysts were measured with a Bruker D8 Advance diffractometer using Cu-K α radiation at 40 kV and 40 mA at a scanning rate of 4°min −1 over the 2θ range of 5-80°with a step size of 0.02°. The crystallite sizes of metallic Ni, NiO (200) and CeO 2 were calculated by the Scherrer equation based on Ni (111) and CeO 2 (111) peaks, respectively. CO 2 temperature programmed desorption (TPD) was performed on a Micromeritics Chemisorption Analyzer 2920TR equipped with a thermal conductivity detector (TCD). The catalyst sample was first treated in He flow at 300°C for 120 min with a ramping rate of 10°C min −1 . After cooling down to 100°C, the sample was exposed to a 10% CO 2 /He gas mixture with a flow rate of 20 ml min −1 for 90 min. Then the sample was swept by 20 ml min −1 He for 80 min to remove the physically absorbed CO 2 until the baseline remained stable. Finally, the signal of TCD was recorded from 100 to 650°C with a heating rate of 10°C min −1 .  From figure 2c,d and f, it can be seen that the NiO-CeO 2 material is an inverse replica of the PMMA template, possessing open and ordered 3D macropores with an inner diameter of 146.6 ± 8.4 nm. Compared with the PMMA colloidal template, the shrinkage of the pores is about 39%, which was caused by the PMMA sphere shrinkage during the calcination. And this is consistent with the 32% contraction ratio of 3DOM Al 2 O 3 derived from the PMMA colloidal template [22] and 60% contraction ratio of 3DOM CeO 2 derived from the polystyrene template [21].

Results
The N 2 isothermal adsorption and desorption and the pore size distribution are shown in figure 3. The physisorption isotherm exhibits a type V with an H3 hysteresis loop which indicates the existence of macropores. The BET surface is 32.19 m 2 g −1 in consistence with 30-40 m 2 g −1 for 3DOM CeO 2 [21]. The DFT method for pore size distribution reveals that macropores larger than 100 nm contribute a lot to the pore volume and there also exists a certain number of micropores of around 3 nm, which may be caused by the calcination of citric acid.   [31] in H 2 /Ar plasma-assisted reduction of Ni/ Al 2 O 3 catalyst. And Zhang et al. [32] found that the H 2 -TPR result of Ni/SiO 2 catalyst after N 2 and H 2 plasma treatment demonstrated the existence of unreduced NiO phase. During the plasma-assisted reduction process, the temperature was less than 250°C, and the incomplete reduction of NiO may be caused by the mild temperature condition. There are only irregular and small drum-bags at the  The chemical adsorption of CO 2 on the catalyst is the first step in the methanation reaction and plays a significant role in the activation process [6,12]. The CO 2 adsorption properties can be evaluated by CO 2 -TPD experiment. The CO 2 -TPD results of catalyst samples reduced in plasma and thermal conditions, respectively, are illustrated in figure 5. Apparently, there are two CO 2 desorption peaks over the temperature range investigated: one is around 175°C and another is around 450°C. The first peak can be ascribed to the existence of weak basic sites such as OH and O groups, and the second peak is due to medium basic sites such as O 2− [2]. Obviously, the order of CO 2 desorption peak area is: 350°C reduction ≈ 400°C reduction < 450°C reduction << plasma reduction. The normalized peak area of weak basic sites is 12.4 for the plasma-reduced catalyst, while it is 9.1 for the 450°C thermal-reduced one; the peak area of medium basic sites in the plasma-reduced catalyst is more than double that of the 450°C thermal-reduced catalyst (33.6 versus 14.7).
Compared to Ni-Ce catalysts reported in the literature [6], the present Ni-Ce catalyst reduced under plasma has more abundant medium basic sites, due to the second peak area being much larger than that of the first peak. According to the literature [33], during the methanation, CO 2 chemisorption on basic sites forms carbonate and it is then hydrogenated to methane via an intermediate-formate species. CO 2 adsorption on the weak basic sites such as OH and O groups tends to form hydrogen carbonate that becomes bidentate carbonate during further hydrogenation. It is prone to form monodentate carbonate and is then hydrogenated to monodentate formate, if the CO 2 adsorption occurs on the medium basic sites such as the O 2− group [33]. The further hydrogenation of monodentate formate takes place faster than that of bidentate carbonate. The medium basic sites play a more significant role than the weak sites for this kind of catalyst in enhancing the catalytic activity during CO 2 methanation [33]. Thus, plasma reduction could produce more abundant basic sites, especially the medium basic sites in the catalyst, which may favour higher CO 2 conversion capacity and efficiency.

Effect of plasma reduction on catalytic activity
Methanation reaction is highly exothermic, low temperature is thermodynamically favourable for high CO 2 conversion and the development of catalysts possessing low-temperature activity has attracted much attention. In §3.2, it was found that plasma reduction brought more medium basic sites and mild plasma reduction may lead to smaller crystal size representing higher activity. Therefore, lowtemperature activity of different catalyst samples reduced in the plasma and thermal reactor was tested in the thermal reactor by regulating the furnace temperature gradually, as shown in figure 6. Due to its exothermal character, once the trigger-point of methanation reaction is reached, a temperature-rise in bed temperature will be detected by the thermocouple inserted in the middle of the catalyst bed. Therefore, CO 2 conversion is plotted versus furnace temperature and bed temperature, respectively, in figure 6 to give a comprehensive view of the activity test. Obviously, the catalyst reduced under plasma conditions showed a higher low-temperature activity and CO 2 conversion than those reduced by the thermal method. As shown in figure 6a, the trigger temperature of the plasma-reduced catalyst is in the range of 190-200°C, and when the furnace temperature reached 268°C, the CO 2 conversion reached the maximum ca 87%. On the other hand, the trigger temperature of the thermalreduced catalyst is 15-20°C higher than that of the plasma catalyst, and the maximum CO 2 conversion was only ca 80% until the furnace temperature reached ca 320°C. Figure 6b also shows a similar temperature difference in the catalyst bed. The maximum CO 2 conversion over plasma-reduced catalyst corresponds to a bed temperature of ca 303°C, and it is ca 370°C for the case of thermal-reduced catalysts. The CH 4 selectivity is almost the same for the two cases around 99.4%. For the thermal reduction process, it is clear that 450°C is an optimum condition, which yields clear Ni crystal peaks in the XRD spectrum. Furthermore, it is found that the order of the catalyst's activity is in accordance with the capacity of CO 2 adsorption (shown in figure 5); in particular, it is more related to the quantity of the medium basic sites.

Effect of plasma on conversion capacity
Besides high activity, more efficient conversion of CO 2 in per unit reactor or catalyst is also an important factor. In order to investigate the feasibility and capacity of the plasma reactor, high GHSV experiments were carried out both in the thermal and plasma reactor with the same catalyst samples for comparison. Figure 7 shows the results of the GHSV test in the thermal fixed-bed reactor loading with optimal Ni-Ce catalyst ( plasma reduced). As shown, when the GHSV increased from 8600 to 40 000 h −1 , although the  royalsocietypublishing.org/journal/rsos R. Soc. open sci. 6: 190750 CH 4 selectivity kept almost constant, the CO 2 conversion drastically and linearly decreased from 87.8 to 71.3%. At the same time, it was found that the catalyst bed temperature ramped up from 430 to 525°C with the increase in GHSV, due to reaction heat. Figure 8 shows the contrast experiment results from the DBD plasma reactor. Apparently, the CO 2 conversion curve in figure 8 is flatter than the one in figure 7, and the plasma reactor showed higher conversion capacity and efficiency for methanation reaction. In detail, the CO 2 conversion remained constant between 88 and 89% in the GHSV range of 4000-20 000 h −1 and then a slow descent occurred at GHSV of 25 000 h −1 ; but the downward trend gradually disappeared, and the conversion maintained at ca 82% from 40 000 to 50 000 h −1 GHSV. The high conversion capacity of the DBD plasma reactor should be due to the unique method of molecular activation through high energetic electrons (1-10 ev). The electrons can activate molecules of CO 2 and H 2 in a more directed and efficient manner. The low temperature of plasma atmosphere will also lead to a low adiabatic reaction temperature that thermodynamically favours CO 2 conversion.

Conclusion
Firstly, the 3DOM Ni-Ce catalyst with pore diameter of 146.6 ± 8.4 nm was successfully synthesized through inverse replica of PMMA templates, and it was combined with plasma technology to enhance the efficiency and capacity of the CO 2 methanation reaction. In the catalyst reduction process, plasma demonstrated a positive impact on the material's physico-chemical properties. CO 2 -TPD results revealed that the plasma-reduced catalyst possessed more abundant weak and medium basic sites than the thermal-reduced catalysts, favouring CO 2 adsorption and activation. A methanation reaction test in the thermal fixed-bed reactor showed that plasma reduction generated higher low-temperature activity than the thermal method. Furthermore, the plasma reaction mode was compared with the thermal mode. The CO 2 conversion decreased almost linearly with the increase in GHSV when used in the thermal reactor. Instead, the plasma reactor gave a relatively stable and higher CO 2 conversion capacity; even GHSV reached 50 000 h −1 , CO 2 conversion was still 82%. This work contributes to a further understanding of the plasma's effect on the catalyst's characteristics and reaction capacity.
Ethics. Our investigation was carried out in full accordance with the ethical guidelines of our research institution and in compliance with Chinese legislation.
Data accessibility. Data supporting this study are available from the Dryad Digital Repository: https://doi.org/10.5061/ dryad.kg76k5b [34]. SARI, CAS) and the National Key R&D Program of China (grant no. 2017YFE0105500). D.H. would like to thank the Start-up Foundation for Advanced Talents of Qingdao University of Science and Technology (grant no. 010022919).