catalysts Well-Dispersed MgAl 2 O 4 Supported Ni Catalyst with Enhanced Catalytic Performance and the Reason of Its Deactivation for Long-Term Dry Methanation Reaction

: Dry methanation of syngas is a promising route for synthetic natural gas production because of its water and cost saving characteristics, as we reported previously. Here, we report a simple soaking process for the preparation of well-dispersed Ni/MgAl 2 O 4 -E catalyst with an average Ni size of 6.4 nm. The catalytic test results showed that the Ni/MgAl 2 O 4 -E catalyst exhibited considerably higher activity and better stability than Ni/MgAl 2 O 4 -W catalyst prepared by conventional incipient wetness impregnation method in dry methanation reaction. The long-term stability test result of 335 h has demonstrated that the deactivation of the Ni/MgAl 2 O 4 -E catalyst is inevitable. With multiple characterization techniques including ICP, EDS, XRD, STEM, TEM, SEM and TG, we reveal that the graphitic carbon encapsulating Ni nanoparticles are the major reasons responsible for catalyst deactivation, and the rate of carbon deposition decreases with reaction time. 2 O 4 -W catalyst synthesized by conventional incipient wetness impregnation method. The catalyst structure evolution and deactivation behavior during long-term stability test of 335 h were probed by using ICP, EDS, XRD, STEM, TEM, SEM and TG. We disclosed that the deactivation of Ni/MgAl 2 O 4 catalyst is attributed to the encapsulation of Ni nanoparticles by graphitic carbon. TEM and STEM obtained on a transmission a high-angle annular dark ﬁeld scanning transmission microscopy Energy dispersive X-ray spectroscopy (EDS) analysis was performed on an adjacent ISIS/INCA energy dispersive X-ray spectrometer Oxford, UK) equipped with an ultrathin window (UTW) detector. The amount of carbon deposits on the spent catalysts were measured using a thermogravimetric analyzer SDT by temperature programmed oxidation method with a of ◦ C/min in an air ﬂow


Introduction
The process for direct combustion of coal usually releases many pollutants such as SO x , NO x , CO x , particulate matter (PM), etc., which has resulted in severe environmental pollution [1,2]. Replacing coal with natural gas for combustion can improve the combustion efficiency and significantly minimize pollution because of the high calorific value of natural gas and low emission of pollutants for its combustion. Natural gas import has risen in China in recent years in spite of the emergence of shale gas and combustible ice [3,4]. Therefore, the process of coal to synthetic natural gas has attracted considerable attention in the past [5][6][7]. Methanation of syngas is a key process for the production of natural gas from coal. Dry methanation of syngas with H 2 /CO ratio of 1 has the advantages of water and cost saving compared with conventional methanation routes due to not needing to adjust the H 2 /CO ratio of raw syngas through water-gas shift (Equation (1)) process before the reaction [8,9]. However, most of the literature focuses on the conventional methanation of syngas with H 2 /CO ratio of 3 (Equation (2)). Dry methanation of syngas with H 2 /CO ratio of 1 has rarely been investigated thus far (Equation (3)).
CO + H 2 O = CO 2 + H 2 ∆H 298K = −41.2 kJ/mol (1) CO + 3H 2 = CH 4 + H 2 O ∆H 298K = −206.1 kJ/mol (2) 2CO + 2H 2 = CH 4 + CO 2 ∆H 298K = −247.3 kJ/mol Nickel based catalysts are most widely used in conventional methanation of syngas with H 2 /CO ratio of 3 or above due to its high activity and low cost [10][11][12][13]. Similar to this The basic physicochemical characteristics of Ni/MgAl 2 O 4 -W and Ni/MgAl 2 O 4 -E samples are listed in Table 1. The specific surface areas for the Ni/MgAl 2 O 4 -W and Ni/MgAl 2 O 4 -E samples are 242.7 and 220.3 m 2 /g, respectively. It is obvious that the S BET values of these two samples have no significant difference. The corresponding pore sizes are 3.5 and 8.6 nm, and the corresponding pore volumes are 0.40 and 0.57 cm 3 /g. The pore sizes were calculated from the desorption branch of the N 2 adsorption-desorption isotherms using the Barrett-Joyner-Halenda (BJH) method. Ni/MgAl 2 O 4 -E has larger pore sizes and pore volumes than Ni/MgAl 2 O 4 -W. It is obvious that the pore size of the Ni/MgAl 2 O 4 -E is the same to that of MgAl 2 O 4 support, while for Ni/MgAl 2 O 4 -W, it is lower than that of MgAl 2 O 4 support. The Ni contents were measured at 5.  impregnation method. The catalyst structure evolution and deactivation behavior during long-term stability test of 335 h were probed by using ICP, EDS, XRD, STEM, TEM, SEM and TG. We disclosed that the deactivation of Ni/MgAl2O4 catalyst is attributed to the encapsulation of Ni nanoparticles by graphitic carbon.

Catalyst Characterization
The basic physicochemical characteristics of Ni/MgAl2O4-W and Ni/MgAl2O4-E samples are listed in Table 1. The specific surface areas for the Ni/MgAl2O4-W and Ni/MgAl2O4-E samples are 242.7 and 220.3 m 2 /g, respectively. It is obvious that the SBET values of these two samples have no significant difference. The corresponding pore sizes are 3.5 and 8.6 nm, and the corresponding pore volumes are 0.40 and 0.57 cm³/g. The pore sizes were calculated from the desorption branch of the N2 adsorption-desorption isotherms using the Barrett-Joyner-Halenda (BJH) method. Ni/MgAl2O4-E has larger pore sizes and pore volumes than Ni/MgAl2O4-W. It is obvious that the pore size of the Ni/MgAl2O4-E is the same to that of MgAl2O4 support, while for Ni/MgAl2O4-W, it is lower than that of MgAl2O4 support. The Ni contents were measured at 5.2 and 4.9 wt.% for the Ni/MgAl2O4-W and Ni/MgAl2O4-E samples , respectively, by ICP-OES analysis.
The XRD patterns of Ni/MgAl2O4-W and Ni/MgAl2O4-E samples are presented in Figure 1. These two samples show the characteristic diffraction peak of MgAl2O4 spinel (JCPDS 21-1152). Ni/MgAl2O4-W displays a distinct diffraction peak at 2θ of 51.8°, which is attributed to the (200) plane of metallic Ni (JCPDS 04-0850), while for Ni/MgAl2O4-E, a broadly weak diffraction peak of metallic Ni at 51.8° is observed. The average crystallite sizes of Ni, which were calculated by using Scherrer equation based on the diffraction of Ni (200), are 9.6 (±2.0) and 5.6 (±1.4) nm for Ni/MgAl2O4-W and Ni/MgAl2O4-E samples, respectively. This result suggests that Ni possesses smaller average size and higher dispersion on MgAl2O4-E than MgAl2O4-W.  Figure 2 shows the STEM images of the freshly reduced Ni/MgAl2O4-W and Ni/MgAl2O4-E samples. The histograms of Ni particle size distribution for these two samples were obtained by statistical analysis of counting more than 300 Ni particles. Ni/MgAl2O4-E shows the average Ni nanoparticle size of about 6.4 nm with a narrow size distribution in the range of 3-11 nm, whereas the Ni/MgAl2O4-W shows the average Ni nanoparticle size of about 10.1 nm with a wide size distribution in the range of 5-19 nm. The Ni nanoparticle sizes obtained from STEM are well consistent with the ones calculated from XRD data. It indicates that Ni/MgAl2O4-E has better dispersion of Ni nanoparticles on MgAl2O4 support than Ni/MgAl2O4-W.

Catalytic Activity and Stability
We assessed the catalytic performance of the Ni/MgAl2O4-W and Ni/MgAl2O4-E catalysts for dry methanation at 0.1 MPa and GHSV of 10,000 mL h −1 gcat. −1 , as shown in Figure  3. The CO conversion (XCO) increases from 51.5% to 91.6% when the temperature increases from 350 to 400 °C over Ni/MgAl2O4-E catalyst. It reaches the equilibrium CO conversion calculated, assuming no carbon formation occurs at 450 °C, and then surpasses them at higher temperatures, while remaining far below the ones calculated, assuming carbon formation occurs, suggesting that carbon formation occurred in the reaction system. Above 450 °C, the XCO and XH2 decreases with increasing reaction temperature due to the limitation of chemical equilibrium. The selectivity of CH4 (SCH4) decreases from 61.9% to 49.6% and then stays at 50-48%, while the selectivity of CO2 (SCO2) increases from 33.8% to 46.6% and maintains at 47-45% when the temperature increases from 350 to 400 °C and above. The carbon balances are above 95% at all tested temperatures as shown in Table S1. For Ni/MgAl2O4-W catalyst, when the reaction temperature increases from 350 to 400 °C and above, the XCO increases from 8.0% to 46.8% and then approaches to the equilibrium conversion calculated, assuming no carbon formation occurs. Meanwhile, the SCH4 decreases from 74.6% to 52.7% and then holds at 48.1-48.5%, and the SCO2 increases from 19.0% to 40.5% and then maintains at 47-46%. As a result, CH4 yields about 45.4% (theoretical yield is 48.2% at 400 °C) and is obtained at 400 °C over Ni/MgAl2O4-E catalyst. All these above results suggest that Ni/MgAl2O4-E catalyst shows better catalytic performance for dry methanation reaction than that of Ni/MgAl2O4-W catalyst.

Catalytic Activity and Stability
We assessed the catalytic performance of the Ni/MgAl 2 O 4 -W and Ni/MgAl 2 O 4 -E catalysts for dry methanation at 0.1 MPa and GHSV of 10,000 mL h −1 g cat.
−1 , as shown in Figure 3. The CO conversion (X CO ) increases from 51.5% to 91.6% when the temperature increases from 350 to 400 • C over Ni/MgAl 2 O 4 -E catalyst. It reaches the equilibrium CO conversion calculated, assuming no carbon formation occurs at 450 • C, and then surpasses them at higher temperatures, while remaining far below the ones calculated, assuming carbon formation occurs, suggesting that carbon formation occurred in the reaction system. Above 450 • C, the X CO and X H 2 decreases with increasing reaction temperature due to the limitation of chemical equilibrium. The selectivity of CH 4 (S CH 4 ) decreases from 61.9% to 49.6% and then stays at 50-48%, while the selectivity of CO 2 (S CO 2 ) increases from 33.8% to 46.6% and maintains at 47-45% when the temperature increases from 350 to 400 • C and above. The carbon balances are above 95% at all tested temperatures as shown in Table S1. For Ni/MgAl 2 O 4 -W catalyst, when the reaction temperature increases from 350 to 400 • C and above, the X CO increases from 8.0% to 46.8% and then approaches to the equilibrium conversion calculated, assuming no carbon formation occurs. Meanwhile, the S CH 4 decreases from 74.6% to 52.7% and then holds at 48.1-48.5%, and the S CO 2 increases from 19.0% to 40.5% and then maintains at 47-46%. As a result, CH 4 yields about 45.4% (theoretical yield is 48.2% at 400 • C) and is obtained at 400 • C over Ni/MgAl 2 O 4 -E catalyst. All these above results suggest that Ni/MgAl 2 O 4 -E catalyst shows better catalytic performance for dry methanation reaction than that of Ni/MgAl 2 O 4 -W catalyst.
The effect of gas hourly space velocity (GHSV) on the Ni/MgAl 2 O 4 -E catalyst performance for dry methanation reaction at 450 • C was studied, as shown in Figure 4. With increasing the GHSV, the X CO starts decreasing from the CO equilibrium conversion (calculated with assuming no carbon formation occurs) when the GHSV is larger than 20,000 mL h −1 g cat.
−1 , indicating that mitigation of mass transport limitation begins. Meanwhile, the S CH 4 increases and S CO 2 decreases, consistent with the increase in the difference between X H 2 and X CO , which reflects that methanation of syngas with H 2 /CO molar ratio above 1, which is pronounced at high GHSV. This requires low GHSV to obtain high X CO with dry methanation selectivity. The effect of gas hourly space velocity (GHSV) on the Ni/MgAl2O4-E catalyst performance for dry methanation reaction at 450 °C was studied, as shown in Figure 4. With increasing the GHSV, the XCO starts decreasing from the CO equilibrium conversion (calculated with assuming no carbon formation occurs) when the GHSV is larger than 20,000 mL h −1 gcat. −1 , indicating that mitigation of mass transport limitation begins. Meanwhile, the SCH 4 increases and SCO 2 decreases, consistent with the increase in the difference between XH 2 and XCO, which reflects that methanation of syngas with H2/CO molar ratio above 1, which is pronounced at high GHSV. This requires low GHSV to obtain high XCO with dry methanation selectivity.    The effect of gas hourly space velocity (GHSV) on the Ni/MgAl2O4-E catalyst performance for dry methanation reaction at 450 °C was studied, as shown in Figure 4. With increasing the GHSV, the XCO starts decreasing from the CO equilibrium conversion (calculated with assuming no carbon formation occurs) when the GHSV is larger than 20,000 mL h −1 gcat. −1 , indicating that mitigation of mass transport limitation begins. Meanwhile, the SCH 4 increases and SCO 2 decreases, consistent with the increase in the difference between XH 2 and XCO, which reflects that methanation of syngas with H2/CO molar ratio above 1, which is pronounced at high GHSV. This requires low GHSV to obtain high XCO with dry methanation selectivity.    −1 . It is obvious that the Ni/MgAl 2 O 4 -E catalyst displays a relatively stable X CO of 88-90% during the total test time of 24 h, while the X CO over Ni/MgAl 2 O 4 -W decreases from 90% to 78.2% in the last 16 h of reaction. Both the S CH 4 and S CO 2 approach 50% and remain almost unchanged during the total test time because of the high X CO , which is also confirmed by the above results (Figures 3 and 4). The above results indicate that Ni/MgAl 2 O 4 -E shows better catalytic stability than Ni/MgAl 2 O 4 -W catalyst.
The catalytic stability of the Ni/MgAl 2 O 4 -E catalyst in dry methanation reaction was further investigated by the long-term test at 450 • C and GHSV of 10,000 mL h −1 g cat. −1 . As shown in Figure 6, the X CO continuously decreases from 91.1% to 18.4% during 300 h time on stream (TOS). It decreases by 6.5%, 10.8%, 12.1% and 17.5% in the first four consecutive 50 h periods, respectively, displaying increasing deactivation rates. The S CH 4 increases from 50.6% to 60.5% and S CO 2 decreases from 47.9% to 26.1% along with the decrease of X CO in similar ways to that presented in Figure 4. After reaction, significant coke in the catalyst bed is visible. After burning coke with air and reducing the spent catalyst with H 2 at 650 • C, the regenerated catalyst exhibits initial X CO of 86.0% and S CH 4 and S CO 2 of 50.6% and 47.2%, respectively. However, the X CO decreases dramatically to 43.8% within 24 h TOS and the S CH 4 and S CO 2 turn to 60.4% and 41.2%, indicating that irreversible deterioration of the catalyst occurs probably due to coke combustion.
the above results (Figures 3 and 4). The above results indicate that Ni/MgAl2O4-E shows better catalytic stability than Ni/MgAl2O4-W catalyst.
The catalytic stability of the Ni/MgAl2O4-E catalyst in dry methanation reaction was further investigated by the long-term test at 450 °C and GHSV of 10,000 mL h −1 gcat. −1 . As shown in Figure 6, the XCO continuously decreases from 91.1% to 18.4% during 300 h time on stream (TOS). It decreases by 6.5%, 10.8%, 12.1% and 17.5% in the first four consecutive 50 h periods, respectively, displaying increasing deactivation rates. The SCH4 increases from 50.6% to 60.5% and SCO2 decreases from 47.9% to 26.1% along with the decrease of XCO in similar ways to that presented in Figure 4. After reaction, significant coke in the catalyst bed is visible. After burning coke with air and reducing the spent catalyst with H2 at 650 °C, the regenerated catalyst exhibits initial XCO of 86.0% and SCH4 and SCO2 of 50.6% and 47.2%, respectively. However, the XCO decreases dramatically to 43.8% within 24 h TOS and the SCH4 and SCO2 turn to 60.4% and 41.2%, indicating that irreversible deterioration of the catalyst occurs probably due to coke combustion.   the above results (Figures 3 and 4). The above results indicate that Ni/MgAl2O4-E shows better catalytic stability than Ni/MgAl2O4-W catalyst. The catalytic stability of the Ni/MgAl2O4-E catalyst in dry methanation reaction was further investigated by the long-term test at 450 °C and GHSV of 10,000 mL h −1 gcat. −1 . As shown in Figure 6, the XCO continuously decreases from 91.1% to 18.4% during 300 h time on stream (TOS). It decreases by 6.5%, 10.8%, 12.1% and 17.5% in the first four consecutive 50 h periods, respectively, displaying increasing deactivation rates. The SCH4 increases from 50.6% to 60.5% and SCO2 decreases from 47.9% to 26.1% along with the decrease of XCO in similar ways to that presented in Figure 4. After reaction, significant coke in the catalyst bed is visible. After burning coke with air and reducing the spent catalyst with H2 at 650 °C, the regenerated catalyst exhibits initial XCO of 86.0% and SCH4 and SCO2 of 50.6% and 47.2%, respectively. However, the XCO decreases dramatically to 43.8% within 24 h TOS and the SCH4 and SCO2 turn to 60.4% and 41.2%, indicating that irreversible deterioration of the catalyst occurs probably due to coke combustion.   A series of characterization techniques were used to probe the structure evolution of the Ni/MgAl 2 O 4 -E catalyst in the long-term stability test. Figure 7 shows the STEM images for the Ni/MgAl 2 O 4 -E catalyst after different reaction times at reaction conditions of 450 • C and 10,000 mL h −1 g cat. −1 . The average Ni nanoparticle sizes of the Ni/MgAl 2 O 4 -E after reaction for 5 and 150 h are 6.9 and 6.4 nm, respectively (Figure 7a,b), which show similar sizes as that of the fresh one (Figure 2b). The average Ni nanoparticle size for the regenerative catalyst of Ni/MgAl 2 O 4 -E increases to 11.5 nm, and the Ni nanoparticles with large sizes of 25-35 nm are clearly observed because of presumed deterioration of the catalyst (the insert image of Figure 7c). These results are in good accord with the sizes calculated by Scheer equation using the XRD data (Table S2 and Figure S1). The obvious increase for the Ni nanoparticle size of the regenerative catalyst may explain the phenomenon of its fast deactivation in dry methanation reaction.
for the Ni/MgAl2O4-E catalyst after different reaction times at reaction conditions of 450 °C and 10,000 mL h −1 gcat. −1 . The average Ni nanoparticle sizes of the Ni/MgAl2O4-E after reaction for 5 and 150 h are 6.9 and 6.4 nm, respectively (Figure 7a,b), which show similar sizes as that of the fresh one (Figure 2b). The average Ni nanoparticle size for the regenerative catalyst of Ni/MgAl2O4-E increases to 11.5 nm, and the Ni nanoparticles with large sizes of 25-35 nm are clearly observed because of presumed deterioration of the catalyst (the insert image of Figure 7c). These results are in good accord with the sizes calculated by Scheer equation using the XRD data (Table S2 and Figure S1). The obvious increase for the Ni nanoparticle size of the regenerative catalyst may explain the phenomenon of its fast deactivation in dry methanation reaction. TEM images reveal the massive deposits of carbon for the spent catalyst, which has a filamentous structure with a Ni nanoparticle attached to the end (Figure 8b-d). HRTEM images display that the graphitic carbon encapsulated the Ni nanoparticles as shown in a representative sample of S150 (Figure 8c). The measured d-spacing values of the graphitic carbon is 0.34 nm, which is the well-crystallized graphitic carbon (JCPDS 75-1621) and is also verified by XRD crystallography ( Figure S1). SEM images of the spent catalysts also show the formation of filamentous carbon deposits with Ni nanoparticles at their tips (Figure S2a-c). The deposits of carbon over Ni/MgAl2O4-S150 were completely removed by regeneration operation (Figure S2d). The amount of carbon deposits over spent catalysts was measured by thermogravimetric (TG) analysis during temperature-programmed oxidation. The weight losses of the spent catalysts below 350 °C are attributed to the desorption of the physiosorbed water, as we previously reported, for the similar types of catalysts. The weight losses of Ni/MgAl2O4-S5, Ni/MgAl2O4-S10, Ni/MgAl2O4-S50, Ni/MgAl2O4-S150 and Ni/MgAl2O4-S335* are 4.5%, 5.0%, 13.6%, 28.3% and 11.4%, respectively, in the temperature range of 350 and 650 °C, which is associated with the oxidation of carbon deposits during temperature programmed processes (Table S2 and Figure S3) [9]. The amount of carbon deposits with reaction time over spent catalysts obtained from TG data are presented in Figure 9a. It is obvious that the longer reaction times give rise to more deposits of carbon. Notably, the amount of carbon deposits over Ni/MgAl2O4-S335 were obtained by recording the weight of reactor before and after reaction. Figure 9b shows the relationship between the average accumulation rate of carbon deposits over Ni/MgAl2O4-E and the reaction time. It demonstrates clearly that the accumulation rate of carbon deposits decreases with the increase of reaction time.
In order to check whether the amount of Ni over Ni/MgAl2O4-E is loss under reaction condition, the Ni content of the Ni/MgAl2O4-E catalyst after different reaction time was determined by ICP and EDS analysis. The Ni content of Ni/MgAl2O4-S5, Ni/MgAl2O4-S10, Ni/MgAl2O4-S50, Ni/MgAl2O4-S150 and Ni/MgAl2O4-S335* determined by ICP is 4.9, 5.0, 5.0, 5.2, and 5.1 wt.%, respectively (Table S2). The EDS for elemental analysis further re- TEM images reveal the massive deposits of carbon for the spent catalyst, which has a filamentous structure with a Ni nanoparticle attached to the end (Figure 8b-d). HRTEM images display that the graphitic carbon encapsulated the Ni nanoparticles as shown in a representative sample of S150 (Figure 8c). The measured d-spacing values of the graphitic carbon is 0.34 nm, which is the well-crystallized graphitic carbon (JCPDS 75-1621) and is also verified by XRD crystallography ( Figure S1). SEM images of the spent catalysts also show the formation of filamentous carbon deposits with Ni nanoparticles at their tips ( Figure S2a-c). The deposits of carbon over Ni/MgAl 2 O 4 -S150 were completely removed by regeneration operation (Figure S2d). The amount of carbon deposits over spent catalysts was measured by thermogravimetric (TG) analysis during temperatureprogrammed oxidation. The weight losses of the spent catalysts below 350 • C are attributed to the desorption of the physiosorbed water, as we previously reported, for the similar types of catalysts. The weight losses of Ni/MgAl 2 O 4 -S5, Ni/MgAl 2 O 4 -S10, Ni/MgAl 2 O 4 -S50, Ni/MgAl 2 O 4 -S150 and Ni/MgAl 2 O 4 -S335* are 4.5%, 5.0%, 13.6%, 28.3% and 11.4%, respectively, in the temperature range of 350 and 650 • C, which is associated with the oxidation of carbon deposits during temperature programmed processes (Table S2 and Figure S3) [9]. The amount of carbon deposits with reaction time over spent catalysts obtained from TG data are presented in Figure 9a. It is obvious that the longer reaction times give rise to more deposits of carbon. Notably, the amount of carbon deposits over Ni/MgAl 2 O 4 -S335 were obtained by recording the weight of reactor before and after reaction. Figure 9b shows the relationship between the average accumulation rate of carbon deposits over Ni/MgAl 2 O 4 -E and the reaction time. It demonstrates clearly that the accumulation rate of carbon deposits decreases with the increase of reaction time.
In order to check whether the amount of Ni over Ni/MgAl 2 O 4 -E is loss under reaction condition, the Ni content of the Ni/MgAl 2 O 4 -E catalyst after different reaction time was determined by ICP and EDS analysis. The Ni content of Ni/MgAl 2 O 4 -S5, Ni/MgAl 2 O 4 -S10, Ni/MgAl 2 O 4 -S50, Ni/MgAl 2 O 4 -S150 and Ni/MgAl 2 O 4 -S335* determined by ICP is 4.9, 5.0, 5.0, 5.2, and 5.1 wt.%, respectively (Table S2). The EDS for elemental analysis further reveals that the Ni content of Ni/MgAl 2 O 4 -S335* is 4.9 wt.% ( Figure S4), which is well consistent with the results obtained from ICP test. These results suggest that the Ni content of Ni/MgAl 2 O 4 -E catalyst hardly changes during the long-term stability test under reaction condition. veals that the Ni content of Ni/MgAl2O4-S335* is 4.9 wt.% ( Figure S4), which is well consistent with the results obtained from ICP test. These results suggest that the Ni content of Ni/MgAl2O4-E catalyst hardly changes during the long-term stability test under reaction condition.

Discussion
Ni/MgAl2O4-E catalyst shows higher activity and better stability than that of Ni/MgAl2O4-W catalyst in dry methanation reaction at 450 °C and GHSV of 10,000 mL h −1 gcat. −1 (Figures 3 and 5). According to the ICP results, these two catalysts have similar Ni content of ~5 wt.% (Table 1). XRD crystallography and STEM images reveal that the Ni/MgAl2O4-E catalyst shows higher dispersion of Ni with the average particle size of ~6 nm, while the Ni/MgAl2O4-W catalyst exhibits lower dispersion of Ni with the average particle size of ~10 nm (Table 1, Figures 1 and 2). According to the literature, the amount of deposited coke increases with the size of Ni nanoparticles increasing from ~6 to ~10 nm  veals that the Ni content of Ni/MgAl2O4-S335* is 4.9 wt.% ( Figure S4), which is well consistent with the results obtained from ICP test. These results suggest that the Ni content of Ni/MgAl2O4-E catalyst hardly changes during the long-term stability test under reaction condition.

Discussion
Ni/MgAl2O4-E catalyst shows higher activity and better stability than that of Ni/MgAl2O4-W catalyst in dry methanation reaction at 450 °C and GHSV of 10,000 mL h −1 gcat. −1 (Figures 3 and 5). According to the ICP results, these two catalysts have similar Ni content of ~5 wt.% (Table 1). XRD crystallography and STEM images reveal that the Ni/MgAl2O4-E catalyst shows higher dispersion of Ni with the average particle size of ~6 nm, while the Ni/MgAl2O4-W catalyst exhibits lower dispersion of Ni with the average particle size of ~10 nm ( Table 1, Figures 1 and 2). According to the literature, the amount of deposited coke increases with the size of Ni nanoparticles increasing from ~6 to ~10 nm

Discussion
Ni/MgAl 2 O 4 -E catalyst shows higher activity and better stability than that of Ni/MgAl 2 O 4 -W catalyst in dry methanation reaction at 450 • C and GHSV of 10,000 mL h −1 g cat. −1 (Figures 3 and 5). According to the ICP results, these two catalysts have similar Ni content of~5 wt.% (Table 1). XRD crystallography and STEM images reveal that the Ni/MgAl 2 O 4 -E catalyst shows higher dispersion of Ni with the average particle size of~6 nm, while the Ni/MgAl 2 O 4 -W catalyst exhibits lower dispersion of Ni with the average particle size of 10 nm (Table 1, Figures 1 and 2). According to the literature, the amount of deposited coke increases with the size of Ni nanoparticles increasing from~6 to~10 nm [39]. These are the possible reasons for the large difference in catalytic performance for dry methanation reaction over Ni/MgAl 2 O 4 -W and Ni/MgAl 2 O 4 -E catalysts.
However, Ni/MgAl 2 O 4 -E catalyst deactivation is still inevitable during the 335 h long-term stability test in dry methanation at 450 • C and GHSV of 10,000 mL h −1 g cat. stability test for different time was observed according to the XRD and STEM results (Figures 1 and 7). From the TEM images and SEM images, obvious filamentous carbon with a nickel crystallite attached to the end formed over Ni/MgAl 2 O 4 -E catalyst after the stability test for different times (Figure 8 and Figure S2). HRTEM images reveal that the Ni nanoparticles were encapsulated by the graphitic carbon. Moreover, considerably more carbon deposits formed with longer stability test time, as evidenced by TG results (Table S2 and Figures 9a and S3). In addition, the Ni content of the Ni/MgAl 2 O 4 -E after the stability test for different time does not (or hardly) change, including under the reaction atmosphere of high CO concentration, as determined by ICP and EDS analysis (Table S2 and Figure S4). The catalytic activity of the deactivated Ni/MgAl 2 O 4 -E cannot be fully recovered by regeneration, which is probably due to the increased Ni particle size during regeneration process (Table S2 and Figures 7c and 8d). All these above results indicate that carbon deposition is the main reason for the deactivation of the Ni/MgAl 2 O 4 -E catalyst in dry methanation reaction.

Catalyst Preparation
MgAl 2 O 4 support was synthesized by a solvothermal method, as we previously reported [9,40]

Catalyst Characterization
The specific surface areas, pore size and pore volume of the samples were measured at −196 • C on a Micromeritics ASAP 2460 instrument (Norcross, GA, USA). All samples were degassed at 350 • C for 5 h under vacuum prior to the adsorption measurements. The Ni loading of the samples was determined by inductively coupled plasma optical emission spectroscopy (ICP-OES, Waltham, MA, USA). X-ray diffraction (XRD) patterns of the samples were recorded on an X-ray diffractometer (PANalytical PW 3040/60 X'Pert PRO, Almelo, The Netherlands) with a Cu Kα (λ = 0.154 nm) radiation source at 40 kV and 40 mA. The micro-morphologies of the samples were characterized on a scanning electron microscopy (JEOL JSM-7800F, Tokyo, Japan). All the samples were grinded into powder and coated onto a conductive tape before characterization. TEM and STEM images were obtained on a transmission electronic microscopy (JEM-2100F, Tokyo, Japan) with a highangle annular dark field scanning transmission electron microscopy (HAADF-STEM, Tokyo, Japan) detector. Energy dispersive X-ray spectroscopy (EDS) analysis was performed on an adjacent ISIS/INCA energy dispersive X-ray spectrometer (Oxford Instruments, Oxford, UK) equipped with an ultrathin window (UTW) detector. The amount of carbon deposits on the spent catalysts were measured using a thermogravimetric analyzer (TA instrument, SDT Q600, New Castle, DE, USA) by temperature programmed oxidation method with a heating rate of 10 • C/min in an air flow of 100 mL/min.

Catalytic Performance Tests
Dry methanation reaction tests were performed in a fixed-bed quartz tubular reactor (i.d. = 10 mm) at atmosphere pressure. For each test, all the catalysts powders were diluted with 2 g of quartz sand and then loaded into a U-shaped reactor. The reactant gases consisted of 48 vol.% CO, 48 vol.% H 2 and 4 vol.% N 2 (internal standard). The flow rates of reactant gases were controlled by mass flow controllers and the reaction temperature was controlled by a programmable temperature controller with a K-type thermocouple. The products and unconverted reactants at the reactor outlet were analyzed using an online gas chromatography (Agilent 7890B, Santa Clara, CA, USA) equipped with a thermal conductivity detector (TCD, Santa Clara, CA, USA) and two packed columns (Parapak N and 5A molecular sieve). The CO and H 2 conversion were calculated as X i = (n i,in − n i,out )/n i,in , where i is CO and H 2 . The CH 4 and CO 2 selectivity were calculated as S j = n j,out /(n CO,in − n CO,out ), where j is CH 4 and CO 2 . The carbon balance was calculated as the moles of carbon at the reactor outlet divided by the moles of carbon at the reactor inlet.

Conclusions
In this work, we report that the well-dispersed Ni/MgAl 2 O 4 -E catalyst with smaller average size of Ni nanoparticles shows better catalytic performance than the Ni/MgAl 2 O 4 -W catalyst in dry methanation reaction under the studied reaction conditions. High and stable catalytic performance was obtained at 400-450 • C over Ni/MgAl 2 O 4 -E catalyst. However, Ni/MgAl 2 O 4 -E catalyst deactivation is inevitable during the long-term stability test at 450 • C due to the encapsulation of Ni nanoparticles by the graphitic carbon. The regeneration of catalyst did not recover the performance of Ni/MgAl 2 O 4 -E catalyst, most likely owing to the increase of Ni nanoparticles size in the process of burning coke. We will focus on the research of exploring suitable regeneration operations condition and designing anti-coking catalysts in the future.
Author Contributions: Investigation, F.W., X.Y. and J.Z.; supervision, F.W. and X.Y.; data curation, F.W. and X.Y.; writing-original draft preparation, F.W. and X.Y.; writing-review and editing, F.W. and J.Z; funding acquisition, F.W. and X.Y. All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by the Anhui Provincial Natural Science Foundation (2108085QB49), the Scientific Research Foundation for the Introduction of Talent, Anhui University of Science and Technology (13200001) and the University-level General Projects of Anhui University of Science and Technology (xjyb2020-07).