Carbon Dioxide Reforming of Methane over Ni Supported SiO 2 : Inﬂuence of the Preparation Method on the Resulting Structural Properties and Catalytic Activity

: Ni-C / SiO 2 and Ni-G / SiO 2 catalysts were prepared by a complexed-impregnation method using citric acid and glycine as complexing agents, respectively. Ni / SiO 2 was also prepared by the conventional incipient impregnation method. All the catalysts were comparatively tested for carbon dioxide reforming of methane (CDR) at P = 1.0 atm, T = 750 ◦ C, CO 2 / CH 4 = 1.0, and GHSV = 60,000 mL · g − 1 · h − 1 . The results showed that Ni-C / SiO 2 and Ni-G / SiO 2 exhibited better CDR performance, especially regarding stability, than Ni / SiO 2 . The conversions of CH 4 and CO 2 were kept constant above 82% and 87% after 20 h of reaction over Ni-C / SiO 2 and Ni-G / SiO 2 while they were decreased from 81% and 88% to 56% and 59%, respectively, over the Ni / SiO 2 . The characterization results of the catalysts before and after the reaction showed that the particle size and the distribution of Ni, as well as the interactions between Ni and the support were signiﬁcantly inﬂuenced by the preparation method. As a result, an excellent resistance to the coking deposition and the anti-sintering of Ni was obtained over the Ni-C / SiO 2 and Ni-G / SiO 2 , leading to a highly active and stable CDR performance.


Introduction
With increasing environmental awareness, the capture and use of CO 2 and CH 4 , which are the two main greenhouse gases, have received increasing attention [1,2]. The carbon dioxide reforming of methane (CDR), which simultaneously converts CO 2 and CH 4 into syngas, has been extensively reported [3][4][5]. In addition to the reduction in greenhouse gas emissions, the syngas produced is H 2 /CO ≤ 1 for CDR, which is the raw material for synthesizing high-value chemicals, such as methanol, higher oxygenates, and long-chain hydrocarbons from Fischer-Tropsch (FT) synthesis [6]. More importantly, compared with other CO 2 conversion and utilization technologies, CDR can be directly applied to the reforming reaction of CH 4 and CO 2 in flue gas [7,8]. Therefore, accelerating the industrialization process of the CDR reaction plays an important role in achieving CO 2 emission reduction and efficient utilization.
Although precious metal catalysts such as Pt and Rh show excellent catalytic activity and resistance to coking for CDR, the cost of precious metal-based catalysts is too high, which limits their application

Structural Properties of Catalyst
The XRD characterization of the calcined and reduced Ni-based catalysts, shown in Figure 1, was used to analyze the structural properties. As shown in Figure 1A, a broad diffraction peak at approximately 23 • (2θ) was observed, which is the characteristic diffraction peak of amorphous silica species [35]. Moreover, the characteristic diffraction peaks of (111), (200), (220), (311), and (222) of the cubic NiO plane were obtained at 2θ of approximately 37, 43, 63, 75, and 79 • on the calcined catalyst [36,37], implying that only cubic NiO was formed on the calcined catalysts. However, the intensity of the NiO diffraction peaks obviously varied for all the catalysts. Ni-C/Q-10 and Ni-G/Q-10 showed weaker diffraction peaks of NiO than Ni/Q-10, and the NiO crystal size was calculated, as shown in Table 1. From the table, Ni/Q-10 exhibited the largest NiO grain size of 16.54 nm, while Ni-C/Q-10 and Ni-G/Q-10 showed smaller NiO particle sizes of 11.58 and 10.01 nm, respectively. This can be explained by the formation of complexes between Ni 2+ and the complex agents, which prevents the agglomeration of Ni particles [38,39]. The XRD pattern for the reduced catalysts was showed in Figure 1B, in addition to the broad peak at 2θ of 23 • for the amorphous silica species, the diffraction peaks at 2θ of 44, 52, and 76 • were assigned to the (111), (200), and (220) diffractions of metallic Ni [36,37,40]. The Ni particle size was also calculated and is shown in Table 1. Unexpectedly, the Ni particle sizes were significantly larger than those of the NiO for the corresponding calcined catalysts, implying that the Ni particles were sintered under the reduction conditions. Moreover, Ni particle sizes of approximately 12 nm were obtained for Ni-C/Q-10 and Ni-G/Q-10, which was obviously smaller than that of Ni/SiO 2 (25.34 nm), and the dispersion of Ni was calculated, as shown in Table 1 [41]. The dispersion of Ni on Ni-C/Q-10 and Ni-G/Q-10 was larger than that on Ni/Q-10, which is in agreement with the reference reports [40]. These results are consistent with the TEM results ( Figure 2). The distributions of Ni particles obtained for Ni-C/Q-10 and Ni-G/Q-10 were narrower than that for Ni/SiO 2 , as shown in Figure 2. These results can be ascribed to complexation with the complexing agents (glycine and citric acid) [40,42].
Catalysts 2020, 10, x FOR PEER REVIEW 3 of 11 consistent with the TEM results ( Figure 2). The distributions of Ni particles obtained for Ni-C/Q-10 and Ni-G/Q-10 were narrower than that for Ni/SiO2, as shown in Figure 2. These results can be ascribed to complexation with the complexing agents (glycine and citric acid) [40,42].

Textural Properties of the Catalysts
The BET surface area, pore volume, and pore diameter of the Ni-based catalysts calculated from the N2 adsorption-desorption isotherms are listed in Table 1. After impregnation with Ni, the BET surface area, pore volume, and pore diameter all obviously decreased. This occurred because the nickel precursor was loaded on the surface of SiO2 and the pore structure, which caused the pores to be blocked, leading to a reduction in the BET specific surface area, pore volume, and pore diameter. However, compared with that of Ni/Q-10, the BET surface areas of Ni-C/Q-10 and Ni-G/Q-10 were larger, implying that the complexing agents formed new pores by decomposing the frame of the complex between Ni 2+ and the complexing agents [43,44]. Moreover, the pore volume and pore diameter of Ni-C/Q-10 and Ni-G/Q-10 were close to those of Ni/Q-10, implying that the complexing agents were unimportant.   [41].
Catalysts 2020, 10, x FOR PEER REVIEW 3 of 11 consistent with the TEM results ( Figure 2). The distributions of Ni particles obtained for Ni-C/Q-10 and Ni-G/Q-10 were narrower than that for Ni/SiO2, as shown in Figure 2. These results can be ascribed to complexation with the complexing agents (glycine and citric acid) [40,42].

Textural Properties of the Catalysts
The BET surface area, pore volume, and pore diameter of the Ni-based catalysts calculated from the N2 adsorption-desorption isotherms are listed in Table 1. After impregnation with Ni, the BET surface area, pore volume, and pore diameter all obviously decreased. This occurred because the nickel precursor was loaded on the surface of SiO2 and the pore structure, which caused the pores to be blocked, leading to a reduction in the BET specific surface area, pore volume, and pore diameter. However, compared with that of Ni/Q-10, the BET surface areas of Ni-C/Q-10 and Ni-G/Q-10 were larger, implying that the complexing agents formed new pores by decomposing the frame of the complex between Ni 2+ and the complexing agents [43,44]. Moreover, the pore volume and pore diameter of Ni-C/Q-10 and Ni-G/Q-10 were close to those of Ni/Q-10, implying that the complexing agents were unimportant.

Textural Properties of the Catalysts
The BET surface area, pore volume, and pore diameter of the Ni-based catalysts calculated from the N 2 adsorption-desorption isotherms are listed in Table 1. After impregnation with Ni, the BET surface area, pore volume, and pore diameter all obviously decreased. This occurred because the nickel precursor was loaded on the surface of SiO 2 and the pore structure, which caused the pores to be blocked, leading to a reduction in the BET specific surface area, pore volume, and pore diameter. However, compared with that of Ni/Q-10, the BET surface areas of Ni-C/Q-10 and Ni-G/Q-10 were larger, implying that the complexing agents formed new pores by decomposing the frame of the complex between Ni 2+ and the complexing agents [43,44]. Moreover, the pore volume and pore diameter of Ni-C/Q-10 and Ni-G/Q-10 were close to those of Ni/Q-10, implying that the complexing agents were unimportant. Figure 3 shows the reduction behavior of the Ni-based catalysts. According to the literature reports, the reduction peak at low temperature (<400 • C) is the reduction of large particles of NiO, and the reduction peak at high temperature can be attributed to the reduction of small particles of NiO and NiO with strong interaction and the support [39,45]. Only one reduction peak at~370 • C was obtained for Ni/Q-10, while in addition to the reduction peak at approximately 370 • C, a significant reduction peak at approximately 570 • C was obtained for Ni-C/Q-10 and Ni-G/Q-10. This can be explained by the complexation of the complexing agent with metallic Ni, which can significantly reduce the particle size of Ni on the support, resulting in a higher reduction peak [46]. The results are consistent with the previous XRD and TEM results.

Reduction Behavior of the Catalysts
Catalysts 2020, 10, x FOR PEER REVIEW 4 of 11 Figure 3 shows the reduction behavior of the Ni-based catalysts. According to the literature reports, the reduction peak at low temperature (<400 °C) is the reduction of large particles of NiO, and the reduction peak at high temperature can be attributed to the reduction of small particles of NiO and NiO with strong interaction and the support [39,45]. Only one reduction peak at ~370 °C was obtained for Ni/Q-10, while in addition to the reduction peak at approximately 370 °C, a significant reduction peak at approximately 570 °C was obtained for Ni-C/Q-10 and Ni-G/Q-10. This can be explained by the complexation of the complexing agent with metallic Ni, which can significantly reduce the particle size of Ni on the support, resulting in a higher reduction peak [46]. The results are consistent with the previous XRD and TEM results.  Figure 4 shows the CDR performance of the Ni-based catalysts under the conditions of P = 1.0 atm, T = 750 °C, CH4/CO2 = 1.0, and GHSV = 60,000 mL·g −1 ·h −1 . As shown in Figure 4A, all the Ni-based catalysts showed higher initial CDR activity, which is close to the equilibrium CH4 conversion (83%), and the CH4 conversion at 0.5 h was above 81%. This result is in agreement with the reference reports [47]. However, there were large differences in their stability. The CH4 conversion of Ni/Q-10 decreased rapidly to 56% after 20 h of reaction. Compared with that of Ni/Q-10, the CDR performance of Ni-C/Q-10 and Ni-G/Q-10 was significantly improved, and an unobservable decrease in CH4 was obtained. Therefore, the preparation method has a great influence on the stability of the catalyst for CDR. This can be explained by the smaller size and more uniform distribution of the catalyst prepared by the coordination-impregnation method, which reduces the carbon deposit of the CDR [38][39][40]42]. This is also completely consistent with the results of the previous discussion.

Catalytic Performance
The CO2 conversion, H2/CO and H2 yield over all the catalysts exhibited a similar trend as the CH4 conversion ( Figure 4B-D). When the conversions of CH4 and CO2 ( Figure 4A,B) were carefully compared, all the catalysts showed the higher CO2 conversion than corresponding to CH4 conversion. This can be ascribed to the reverse water gas shift reaction, which is also explained the lower H2/CO than the equilibrium value shown in Figure 4C. Moreover, a higher H2 yield was obtained over the Ni-C/Q-10 and Ni-G/Q-10 shown in Figure 4D.  Figure 4 shows the CDR performance of the Ni-based catalysts under the conditions of P = 1.0 atm, T = 750 • C, CH 4 /CO 2 = 1.0, and GHSV = 60,000 mL·g −1 ·h −1 . As shown in Figure 4A, all the Ni-based catalysts showed higher initial CDR activity, which is close to the equilibrium CH 4 conversion (83%), and the CH 4 conversion at 0.5 h was above 81%. This result is in agreement with the reference reports [47]. However, there were large differences in their stability. The CH 4 conversion of Ni/Q-10 decreased rapidly to 56% after 20 h of reaction. Compared with that of Ni/Q-10, the CDR performance of Ni-C/Q-10 and Ni-G/Q-10 was significantly improved, and an unobservable decrease in CH 4 was obtained. Therefore, the preparation method has a great influence on the stability of the catalyst for CDR. This can be explained by the smaller size and more uniform distribution of the catalyst prepared by the coordination-impregnation method, which reduces the carbon deposit of the CDR [38][39][40]42]. This is also completely consistent with the results of the previous discussion.

Catalytic Performance
The CO 2 conversion, H 2 /CO and H 2 yield over all the catalysts exhibited a similar trend as the CH 4 conversion ( Figure 4B-D). When the conversions of CH 4 and CO 2 ( Figure 4A,B) were carefully compared, all the catalysts showed the higher CO 2 conversion than corresponding to CH 4 conversion. This can be ascribed to the reverse water gas shift reaction, which is also explained the lower H 2 /CO than the equilibrium value shown in Figure 4C. Moreover, a higher H 2 yield was obtained over the Ni-C/Q-10 and Ni-G/Q-10 shown in Figure 4D.

Characterization of Used Catalysts
The used catalysts after CDR for TOS = 20 h were subjected to XRD characterization, and the results are shown in Figure 5. Compared with the reduced catalysts (Figure 1), except for the clear shoulder peak at 2θ of approximately 26° for the used Ni/Q-10, similar diffraction peaks were obtained for the used catalysts, implying that they had the same Ni crystal structure after CDR. The diffraction peak at 2θ of approximately 26° was a diffraction peak of graphitic carbon, suggesting that less coke was deposited on Ni-C/Q-10 and Ni-G/Q-10 than on Ni/Q-10, which was attributed to the smaller Ni particle size [8,37]. When the intensities of the Ni diffraction peaks were carefully compared, the peak intensities of the used catalysts were greater than those of the corresponding reduced catalysts, implying Ni sintering after CDR, which agrees with previous reports [37,39]. To clearly express the degree of Ni sintering, the Ni crystal sizes of the used catalyst were calculated and are listed in Table 1. From Table 1, the Ni crystal sizes of the used Ni/Q-10 significantly increased from 25.34 nm to 32.75 nm, while the used Ni-C/Q-10 and Ni-G/Q-10 displayed slight increases, suggesting superior anti-sintering on Ni-C/Q-10 and Ni-G/Q-10. These results are also consistent with the results of a previous report stating that good CDR performance was achieved with catalysts with superior anti-sintering characteristics [37,39,48], which is also supported by the results of TPR (Figure 3), showing a stronger interaction between the Ni and supports [37,39,45,49].

Characterization of Used Catalysts
The used catalysts after CDR for TOS = 20 h were subjected to XRD characterization, and the results are shown in Figure 5. Compared with the reduced catalysts (Figure 1), except for the clear shoulder peak at 2θ of approximately 26 • for the used Ni/Q-10, similar diffraction peaks were obtained for the used catalysts, implying that they had the same Ni crystal structure after CDR. The diffraction peak at 2θ of approximately 26 • was a diffraction peak of graphitic carbon, suggesting that less coke was deposited on Ni-C/Q-10 and Ni-G/Q-10 than on Ni/Q-10, which was attributed to the smaller Ni particle size [8,37]. When the intensities of the Ni diffraction peaks were carefully compared, the peak intensities of the used catalysts were greater than those of the corresponding reduced catalysts, implying Ni sintering after CDR, which agrees with previous reports [37,39]. To clearly express the degree of Ni sintering, the Ni crystal sizes of the used catalyst were calculated and are listed in Table 1. From Table 1, the Ni crystal sizes of the used Ni/Q-10 significantly increased from 25.34 nm to 32.75 nm, while the used Ni-C/Q-10 and Ni-G/Q-10 displayed slight increases, suggesting superior anti-sintering on Ni-C/Q-10 and Ni-G/Q-10. These results are also consistent with the results of a previous report stating that good CDR performance was achieved with catalysts with superior anti-sintering characteristics [37,39,48], which is also supported by the results of TPR (Figure 3), showing a stronger interaction between the Ni and supports [37,39,45,49]. Moreover, the used catalysts after the CDR reaction were characterized by TG, as shown in Figure 6, to confirm the coke deposition. An observable weight increase occurred at temperatures from approximately 200 °C to 400 °C for all the used catalysts due to the oxidation of Ni to NiO [50], while the weight loss starting from 400 °C was assigned to the gasification of the coke deposited after the CDR reaction. As shown in Figure 6, a slight weight loss about ~1.4-1.7% was obtained for the used Ni-C/Q-10 and Ni-G/Q-10 after 20 h CDR testing, and it was significantly lower than that of the used Ni/Q-10 (3.1%), which is significantly lower than the weight loss of the reference reports [51,52]. These results showed that the coke deposition was significantly reduced by adding complexing agent, which was attributed to its smaller Ni particle sizes ( Figure 2 and Table 1) leading to a high stability of CDR ( Figure 4).  Moreover, the used catalysts after the CDR reaction were characterized by TG, as shown in Figure 6, to confirm the coke deposition. An observable weight increase occurred at temperatures from approximately 200 • C to 400 • C for all the used catalysts due to the oxidation of Ni to NiO [50], while the weight loss starting from 400 • C was assigned to the gasification of the coke deposited after the CDR reaction. As shown in Figure 6, a slight weight loss about~1.4-1.7% was obtained for the used Ni-C/Q-10 and Ni-G/Q-10 after 20 h CDR testing, and it was significantly lower than that of the used Ni/Q-10 (3.1%), which is significantly lower than the weight loss of the reference reports [51,52]. These results showed that the coke deposition was significantly reduced by adding complexing agent, which was attributed to its smaller Ni particle sizes ( Figure 2 and Table 1) leading to a high stability of CDR ( Figure 4). Moreover, the used catalysts after the CDR reaction were characterized by TG, as shown in Figure 6, to confirm the coke deposition. An observable weight increase occurred at temperatures from approximately 200 °C to 400 °C for all the used catalysts due to the oxidation of Ni to NiO [50], while the weight loss starting from 400 °C was assigned to the gasification of the coke deposited after the CDR reaction. As shown in Figure 6, a slight weight loss about ~1.4-1.7% was obtained for the used Ni-C/Q-10 and Ni-G/Q-10 after 20 h CDR testing, and it was significantly lower than that of the used Ni/Q-10 (3.1%), which is significantly lower than the weight loss of the reference reports [51,52]. These results showed that the coke deposition was significantly reduced by adding complexing agent, which was attributed to its smaller Ni particle sizes ( Figure 2 and Table 1) leading to a high stability of CDR ( Figure 4).

Preparation of Catalysts
The Ni-supported SiO 2 catalysts were prepared by the complexed-impregnation method. Using citric acid (C) and glycine (G) as the complexing agents, commercial SiO 2 (Q-10 was purchased from Fujisilicia) was used as the support, and Ni(NO 3 ) 2 ·6H 2 O was used as the precursor of Ni. The nickel content in the catalyst was set to 10 wt.%. Before the impregnation, Q-10 was calcined at 750 • C for 4 h. The Ni(NO 3 ) 2 ·6H 2 O and complexing agent were dissolved in distilled water containing the same volume of Q-10. The molar ratio of the complexing agent to Ni was 0.5. Then, the Q-10 was added to the mixed aqueous solution, and it was continuously oscillated for 30 min until evenly mixed. After the sample was stored at room temperature overnight, it was dried at 80 • C for 12 h and calcined at 500 • C for 4 h. The catalysts were designated Ni-x/Q-10, where x is the complexing agent used. For comparison, Ni/Q-10 was prepared by the same procedure with the same preparation parameters but without the addition of a complexing agent.

Characterization Techniques Used on the Catalysts
XRD was performed on the samples by an X-ray powder diffractometer (Bruker, Germany, D8 Advance) with Cu Kα radiation (40 kV, 40 mA). The grain sizes of the Ni and NiO were determined by the Scherrer equation based on the (111) and (200) planes, respectively.
The textural properties were measured by a BelSorp-Max physisorption analyzer under liquid nitrogen temperature. The sample was pretreated under vacuum (10 −2 kPa) at 300 • C for 10 h. The specific surface area of the sample was calculated by the BET method, and the pore volume was tested from the nitrogen adsorption data at the relative pressure (P/P 0 ) = 0.99.
TEM images were taken on a JEOL JEM-2100 transmission electron microscope. The sample was dispersed in an ethanol solution by ultrasonic sonication for 2 min, dropped on copper grids and dried in air.
The coke deposition of the used catalyst was measured using a Q1000DSC+LNCS+FACS Q600SDT type thermogravimetry-differential thermal analyzer under air atmosphere with a total flow of 100 mL·min −1 .

Activity Evaluation of the Catalysts
The CDR was carried out in a fixed bed reactor with a quartz reaction tube (the inner diameter was 8 mm). Then, 0.10 g catalyst (40-60 mesh) was loaded and reduced by 50 mL·min −1 10% H 2 /N 2 , and the temperature was raised to 700 • C for 150 min. After the reduction, it was purged with 40 mL·min −1 N 2 for 30 min, the temperature was raised to 750 • C, and the flow was switched to the feed gas (CH 4 /CO 2 = 1.0). The CDR reaction was performed under the conditions of T = 750 • C, P = 1.0 atm, and gas hourly space velocity (GHSV) = 60,000 mL·g −1 ·h −1 . The product gases, which were condensed in an ice water bath to remove the water, were separated by Molecular Sieve 5A and Porapak Q capillary columns and analyzed by online gas chromatography (GC9720II, Zhejiang Fuli chromatographic analysis Co., Ltd., Zhejiang, China).

Conclusions
Using citric acid and glycine as complexing agents, Ni-C/SiO 2 and Ni-G/SiO 2 catalysts were prepared by the complexing-impregnation method. Compared with Ni/SiO 2 prepared with the traditional incipient impregnation method, Ni-C/SiO 2 and Ni-G/SiO 2 show smaller particle sizes and better dispersion of Ni and stronger interactions between Ni and the support SiO 2 . The CDR performance of the Ni-based catalysts was comparatively studied. Both the catalyst preparation of the coordination-impregnation method and the traditional incipient impregnation method showed