Inﬂuence of Nickel Loading on Reduced Graphene Oxide-Based Nickel Catalysts for the Hydrogenation of Carbon Dioxide to Methane

: In this study, a series of novel nickel catalysts supported on reduced graphene oxide nanosheets (Ni / rGO) with Ni loadings of 10, 15 and 20 wt% were successfully synthesized via the incipient wetness impregnation method. The physicochemical properties of the catalysts and rGO support were thoroughly characterized by thermogravimetric analyser, X-ray di ﬀ raction, fourier-transform infrared spectroscopy, Raman spectroscopy, N 2 adsorption-desorption, temperature programmed reduction, temperature programmed CO 2 desorption and ﬁeld emission scanning electron microscopy with energy dispersive X-ray spectroscopy. The properties of the catalysts are correlated to its catalytic activity for CO 2 methanation which were investigated using three-phase slurry reactor at low temperature and pressure of 240 ◦ C and 10 bar, respectively. Among the three catalysts of di ﬀ erent Ni loading, Ni 15 / rGO shows the highest activity of 51% conversion of CO 2 with total selectivity towards CH 4 . N 2 -physisorption and CO 2 -TPD analysis suggest that high catalytic performance of Ni 15 / rGO is attributed to the high surface area, strong basic sites and special support e ﬀ ect of rGO in anchoring the active metal.


Introduction
The concentration of carbon dioxide (CO 2 ), a major component of greenhouse gases, is rapidly increasing in the atmosphere [1]. With over 30 gigaton (Gt) emissions a year, high atmospheric CO 2 is alarming as it will create critical risks for the Earth's climate system [2]. Due to the global concern about the climate change, CO 2 conversion into valuable low-carbon fuels such as methane (CH 4 ) is seen as a viable solution to reduce anthropogenic CO 2 emissions [3]. CO 2 methanation, as in Equation (1), was initially reported by Sabatier and Senderens in 1902 [4]. Ideally in a power-to-gas (P2G) technology, CH 4 will be produced from Sabatier reaction by reacting CO 2 recovered from industrial process with hydrogen (H 2 ) from water splitting.

Physicochemical Properties of rGO and Ni/rGO
A thermogram of the rGO support showing the DTG curve is shown in Figure 1. The TGA profile of rGO implies that rGO has excellent thermal stability because heating up to 900 • C only resulted in a 15.14% weight loss. The first weight loss step is observed between 100 • C to 150 • C which is caused by the loss of physisorbed water from the sample (2.28% loss). The presence of physisorbed water on the rGO surface is due to the porous characteristics and oxygen functional groups on the edge of graphene [20,25,26]. Further heating to the calcination temperature of 400 • C only resulted in another 4% degradation; hence, rGO does not degrade in the Ni/rGO synthesis step. High thermal stability of the rGO is important to avoid decomposition which could result in metal sintering during CO 2 methanation; hence, reducing the conversion efficiency. At elevated reaction temperatures, rGO can anchor the Ni active phase without being decomposed during CO 2 methanation reaction. From Raman spectrum in Figure 2, three fundamental vibrations are observed in the range of 1300 to 3000 cm −1 . D band, the breathing mode of κ-point phonons, is observed at 1352 cm −1 with relatively low intensity which indicates the presence of defects in rGO. Even though the presence of defects can disrupt the structure of carbon atoms at the graphitic edges [27,28] but it can also provide active sites for chemical reactions [29]. The G vibration band, assigned to the E2g phonon of sp 2 C atoms in graphite single crystals [30], appears at 1572 cm −1 . As compared to rGO synthesized according to the literature [26,31], this commercial rGO has a sharper G band which indicates a higher sp 2 domain and lower degree of disorder [28]. The 2D band of rGO is observed at 2704 cm −1 . The I2D/IG ratio can give insights on the graphene layers [27,32]. The result shows that I2D/IG <1; hence, the rGO is multilayer. Upon impregnation of Ni onto rGO, the Raman spectrum is blue-shifted. The frequency shift of the Raman spectra of graphene was explained by Dervishi et al. who indicated that the increase in Raman shift of D and G bands indicates an increase in defects size [33]. Apart from that, the ID/IG ratio was found to increase as the defect density increases. The intensity ratio of rGO and Ni15/rGO is almost similar, From Raman spectrum in Figure 2, three fundamental vibrations are observed in the range of 1300 to 3000 cm −1 . D band, the breathing mode of κ-point phonons, is observed at 1352 cm −1 with relatively low intensity which indicates the presence of defects in rGO. Even though the presence of defects can disrupt the structure of carbon atoms at the graphitic edges [27,28] but it can also provide active sites for chemical reactions [29]. From Raman spectrum in Figure 2, three fundamental vibrations are observed in the range of 1300 to 3000 cm −1 . D band, the breathing mode of κ-point phonons, is observed at 1352 cm −1 with relatively low intensity which indicates the presence of defects in rGO. Even though the presence of defects can disrupt the structure of carbon atoms at the graphitic edges [27,28] but it can also provide active sites for chemical reactions [29]. The G vibration band, assigned to the E2g phonon of sp 2 C atoms in graphite single crystals [30], appears at 1572 cm −1 . As compared to rGO synthesized according to the literature [26,31], this commercial rGO has a sharper G band which indicates a higher sp 2 domain and lower degree of disorder [28]. The 2D band of rGO is observed at 2704 cm −1 . The I2D/IG ratio can give insights on the graphene layers [27,32]. The result shows that I2D/IG <1; hence, the rGO is multilayer. Upon impregnation of Ni onto rGO, the Raman spectrum is blue-shifted. The frequency shift of the Raman spectra of graphene was explained by Dervishi et al. who indicated that the increase in Raman shift of D and G bands indicates an increase in defects size [33]. Apart from that, the ID/IG ratio was found to increase as the defect density increases. The intensity ratio of rGO and Ni15/rGO is almost similar, The G vibration band, assigned to the E 2g phonon of sp 2 C atoms in graphite single crystals [30], appears at 1572 cm −1 . As compared to rGO synthesized according to the literature [26,31], this commercial rGO has a sharper G band which indicates a higher sp 2 domain and lower degree of disorder [28]. The 2D band of rGO is observed at 2704 cm −1 . The I 2D /I G ratio can give insights on the graphene layers [27,32]. The result shows that I 2D /I G <1; hence, the rGO is multilayer. Upon impregnation of Ni onto rGO, the Raman spectrum is blue-shifted. The frequency shift of the Raman spectra of graphene was explained by Dervishi et al. who indicated that the increase in Raman shift of D and G bands indicates an increase in defects size [33]. Apart from that, the I D /I G ratio was found to increase as the defect density increases. The intensity ratio of rGO and Ni 15 /rGO is almost similar, Catalysts 2020, 10, 471 4 of 15 possibly suggesting the defect density does not increase due to Ni impregnation but only the defect size increases in the graphene layer. Figure 3 depicts the X-ray diffractograms of rGO and Ni/rGO catalysts. XRD pattern of rGO support shows good agreement with the hexagonal phase of the graphene crystal lattice. A major peak at 26.52 • is indexed to the (002) facet of graphite crystals [34]. This peak describes the layers of rGO that are stacked at an interlayer spacing (d) of 0.336 nm [35]. The calculated crystallite size (L c ) in the stacking direction and width (L a ) of each rGO layer are 12.27 nm and 25.37 nm, respectively. Therefore, there are about 36 graphitic layers in each rGO crystal. The sharp and intense peak reflects the high crystallinity of rGO due to the structure that is only edge-oxidized, so its stacking is more comparable to well-ordered graphite. XRD patterns of synthesized rGO from other studies show broader (002) peaks due to poor ordering of the sheets along the stacking direction [25,26,36]. Another peak at 43.80 • is indexed to the (100) plane of rGO. The commercial rGO used in this study is a multiple layer carbon allotrope where each layer consists of carbon atoms arranged in a honeycomb fashion with defects and oxygen functional groups at the edge [31]. After Ni is impregnated onto rGO, the (002) peak of rGO is observed with same intensity and position as in pristine rGO. This observation suggests that there was no disruption of the rGO crystallinity. All Ni/rGO catalysts reflect NiO peaks that are consistent with those of cubic phase which shows the viability of the wetness impregnation method to support Ni onto rGO. Diffraction peaks at 37.  Table 1. possibly suggesting the defect density does not increase due to Ni impregnation but only the defect size increases in the graphene layer. Figure 3 depicts the X-ray diffractograms of rGO and Ni/rGO catalysts. XRD pattern of rGO support shows good agreement with the hexagonal phase of the graphene crystal lattice. A major peak at 26.52° is indexed to the (002) facet of graphite crystals [34]. This peak describes the layers of rGO that are stacked at an interlayer spacing (d) of 0.336 nm [35]. The calculated crystallite size (Lc) in the stacking direction and width (La) of each rGO layer are 12.27 nm and 25.37 nm, respectively. Therefore, there are about 36 graphitic layers in each rGO crystal. The sharp and intense peak reflects the high crystallinity of rGO due to the structure that is only edge-oxidized, so its stacking is more comparable to well-ordered graphite. XRD patterns of synthesized rGO from other studies show broader (002) peaks due to poor ordering of the sheets along the stacking direction [25,26,36]. Another peak at 43.80° is indexed to the (100) plane of rGO. The commercial rGO used in this study is a multiple layer carbon allotrope where each layer consists of carbon atoms arranged in a honeycomb fashion with defects and oxygen functional groups at the edge [31]. After Ni is impregnated onto rGO, the (002) peak of rGO is observed with same intensity and position as in pristine rGO. This observation suggests that there was no disruption of the rGO crystallinity. All Ni/rGO catalysts reflect NiO peaks that are consistent with those of cubic phase which shows the viability of the wetness impregnation method to support Ni onto rGO. Diffraction peaks at 37.07°, 43.27°, 62.72°, 75.16° and 79.13° are assigned to the (111), (200), (220), (311) and (222) lattice planes, respectively. The intensity of the NiO peaks increases with increasing Ni content from 10 to 20 wt.% due to the formation of larger NiO crystals on the rGO [37]. The crystallite size (D) of NiO on Ni10/rGO, Ni15/rGO and Ni20/rGO catalysts are calculated from Scherrer equation at peak (111) and are shown in Table 1.    Isotherms of rGO and Ni/rGO catalysts are presented in Figure 4, while comparative values obtained from surface area and porosity analysis are shown in Table 1. rGO exhibits a Type IV isotherm with H3 hysteresis loop based on the IUPAC classification [38]. rGO is a mesoporous material (its pores range from 2-50 nm) with aggregates of plate-like particles so the pores have a slit shape [13,39]. The BET surface area and BJH pore volume of rGO are 258 m 2 g −1 and 0.39 cm 3 g −1 , respectively. A sharp increase at P/P o is observed due to the presence of micropores (with pore sizes of less than 2 nm). rGO has a micropore surface area of 43 m 2 g −1 with a pore volume of 0.019 cm 3 g −1 . The porous structure and high surface area features of rGO can serve as sites for Ni dispersion and active sites for chemical reactions. The isotherms and hysteresis loop patterns of the catalysts are like those of the rGO support. Furthermore, the pore volume of rGO decreases when Ni is loaded. Therefore, these results suggest that Ni nanoparticles are distributed and dispersed on the rGO surface and pores since the Ni crystallite size is smaller than the rGO pore diameter [40]. The amount of N 2 adsorbed decreased due to loading of Ni which has much higher density than rGO [41]. The surface area increases as the metal content is increased from 10% to 15% and decreases with further increment of the metal loading of 20%. This trend observed for the surface area is similar to that mentioned in the literature [42][43][44]. Ni 15 /rGO shows the highest surface area, possibly because of a better dispersion and smaller particle size of the Ni on rGO. Uniform dispersion of the nickel metal on rGO was achieved with 15% loading, which is indicated from the FESEM/EDX-mapping. It is proposed that the 15% Ni loading led to small Ni particle size formation and lower metal agglomeration. The pore diameter of Ni 15 /rGO is proven to be the smallest among the different Ni/rGO catalysts, suggesting a high concentration of smaller pores and hence reflecting the highest surface area [45]. A further increase of Ni loading to 20 wt% resulted in a lower surface area of the catalyst. Based on a previous study by Jang et al. [46], this could be attributed to the agglomeration of nickel particles on the rGO. Isotherms of rGO and Ni/rGO catalysts are presented in Figure 4, while comparative values obtained from surface area and porosity analysis are shown in Table 1. rGO exhibits a Type IV isotherm with H3 hysteresis loop based on the IUPAC classification [38]. rGO is a mesoporous material (its pores range from 2-50 nm) with aggregates of plate-like particles so the pores have a slit shape [13,39]. The BET surface area and BJH pore volume of rGO are 258 m 2 g −1 and 0.39 cm 3 g −1 , respectively. A sharp increase at P/Pο is observed due to the presence of micropores (with pore sizes of less than 2 nm). rGO has a micropore surface area of 43 m 2 g −1 with a pore volume of 0.019 cm 3 g −1 .
The porous structure and high surface area features of rGO can serve as sites for Ni dispersion and active sites for chemical reactions. The isotherms and hysteresis loop patterns of the catalysts are like those of the rGO support. Furthermore, the pore volume of rGO decreases when Ni is loaded. Therefore, these results suggest that Ni nanoparticles are distributed and dispersed on the rGO surface and pores since the Ni crystallite size is smaller than the rGO pore diameter [40]. The amount of N2 adsorbed decreased due to loading of Ni which has much higher density than rGO [41]. The surface area increases as the metal content is increased from 10% to 15% and decreases with further increment of the metal loading of 20%. This trend observed for the surface area is similar to that mentioned in the literature [42][43][44]. Ni15/rGO shows the highest surface area, possibly because of a better dispersion and smaller particle size of the Ni on rGO. Uniform dispersion of the nickel metal on rGO was achieved with 15% loading, which is indicated from the FESEM/EDX-mapping. It is proposed that the 15% Ni loading led to small Ni particle size formation and lower metal agglomeration. The pore diameter of Ni15/rGO is proven to be the smallest among the different Ni/rGO catalysts, suggesting a high concentration of smaller pores and hence reflecting the highest surface area [45]. A further increase of Ni loading to 20 wt% resulted in a lower surface area of the catalyst. Based on a previous study by Jang et al. [46], this could be attributed to the agglomeration of nickel particles on the rGO.   The FTIR spectra of rGO and Ni 15 /rGO catalyst are presented in Figure 5, while the assignments of each absorption peak are listed in Table 2. A strong broad absorption at 3420 cm −1 is attributed to the O-H stretching vibration, very likely coming from oxygen-containing functional groups at the edge of the graphene sheets [47]. The peak at 1725 cm −1 reflects the vibration band of carbonyl functional groups (C=O, COOH) and the medium broad peak at 1087 cm −1 is assigned to the stretching mode of C-O and C-O-C. The presence of skeletal C=C in rGO is observed from the stretching and bending of C=C at 1574 cm −1 and 800 cm −1 , respectively. The spectrum of Ni 15 /rGO shows no formation of new peaks after metal loading, so the FTIR spectrum of Ni/rGO is similar to that of rGO in terms of peak position of each functional group, possibly suggesting that the oxygen functional groups in rGO were not degraded in the synthesis step, in line with the TGA result that indicated a high stability of rGO.
Catalysts 2020, 10, 471 6 of 15 The FTIR spectra of rGO and Ni15/rGO catalyst are presented in Figure 5, while the assignments of each absorption peak are listed in Table 2. A strong broad absorption at 3420 cm −1 is attributed to the O-H stretching vibration, very likely coming from oxygen-containing functional groups at the edge of the graphene sheets [47]. The peak at 1725 cm −1 reflects the vibration band of carbonyl functional groups (C=O, COOH) and the medium broad peak at 1087 cm −1 is assigned to the stretching mode of C-O and C-O-C. The presence of skeletal C=C in rGO is observed from the stretching and bending of C=C at 1574 cm −1 and 800 cm −1 , respectively. The spectrum of Ni15/rGO shows no formation of new peaks after metal loading, so the FTIR spectrum of Ni/rGO is similar to that of rGO in terms of peak position of each functional group, possibly suggesting that the oxygen functional groups in rGO were not degraded in the synthesis step, in line with the TGA result that indicated a high stability of rGO.  Temperature programmed reduction (TPR) analysis using H2 was performed to study the relation between reducibility and metal-support interaction. The TPR profiles of the Ni/rGO catalysts are presented in Figure 6. The profile provided the amount of H2 gas consumed by the catalysts at different temperatures caused by redox reaction between H2 and NiO according to the following equation: NiO(s) + H2(g) → Ni°(s) + H2O(g) [13]. From Figure 6, the three catalysts show strong signals at temperatures around 400-600 °C caused by the reduction of Ni 2+ species that strongly interact with the graphene support [48,49]. Ni15/rGO shows the highest reduction temperature owing to the strongest metal-support interaction. This can be explained as due to the fact the energy needed to overcome the interactions of NiO and rGO increases with the increasing strength of such interactions. Furthermore, the strongest interactions in Ni15/rGO also relates to the high dispersion of the NiO on the rGO support which is proven by EDX mapping and highest BET surface area [42]. In addition, a  Temperature programmed reduction (TPR) analysis using H 2 was performed to study the relation between reducibility and metal-support interaction. The TPR profiles of the Ni/rGO catalysts are presented in Figure 6. The profile provided the amount of H 2 gas consumed by the catalysts at different temperatures caused by redox reaction between H 2 and NiO according to the following equation: [13]. From Figure 6, the three catalysts show strong signals at temperatures around 400-600 • C caused by the reduction of Ni 2+ species that strongly interact with the graphene support [48,49]. Ni 15 /rGO shows the highest reduction temperature owing to the strongest metal-support interaction. This can be explained as due to the fact the energy needed to overcome the interactions of NiO and rGO increases with the increasing strength of such interactions. Furthermore, the strongest interactions in Ni 15 /rGO also relates to the high dispersion of the NiO on the rGO support which is proven by EDX mapping and highest BET surface area [42]. In addition, a shoulder peak is observed at around 300-350 • C for all catalysts. This peak is assigned to the surface species of bulk Ni that are weakly contacting with the support, called 'free' nickel species [40,50]. The intensity of this peak increases with increasing Ni loading and it became apparent in Ni 20 /rGO catalyst possibly because of the agglomeration of larger NiO crystallites on the surface, in agreement with the XRD and BET results. Based on Table 3, H 2 consumption reflects the order of Ni 10 /rGO < Ni 15 /rGO < Ni 20 /rGO which follows the amount of Ni concentrations in the samples. In comparison to some literature values, the reduction temperature of Ni/rGO is relatively lower than those of Ni/Al 2 O 3 [12,40,51], Ni/SiO 2 [13] and Ni/ZrO 2 [15], which reflects the higher reducibility of Ni 2+ in our Ni/rGO catalysts.
Catalysts 2020, 10, 471 7 of 15 shoulder peak is observed at around 300-350 °C for all catalysts. This peak is assigned to the surface species of bulk Ni that are weakly contacting with the support, called 'free' nickel species [40,50]. The intensity of this peak increases with increasing Ni loading and it became apparent in Ni20/rGO catalyst possibly because of the agglomeration of larger NiO crystallites on the surface, in agreement with the XRD and BET results. Based on Table 3, H2 consumption reflects the order of Ni10/rGO < Ni15/rGO < Ni20/rGO which follows the amount of Ni concentrations in the samples. In comparison to some literature values, the reduction temperature of Ni/rGO is relatively lower than those of Ni/Al2O3 [12,40,51], Ni/SiO2 [13] and Ni/ZrO2 [15], which reflects the higher reducibility of Ni 2+ in our Ni/rGO catalysts.   Figure 7 shows the CO2-TPD spectra of the Ni/rGO catalysts and rGO support. NiO is also evaluated for control purposes. CO2-TPD evaluates the CO2 adsorption and activation capability of the catalyst for methanation. rGO does not show any significant desorption, indicating that it contains only a small number of basic sites. On the other hand, NiO shows a broad signal for the γ peak at 583 °C. When Ni is supported on rGO, a sharp and intense desorption peak at ~550 °C appears. This γ peak corresponds to the strong basic sites [49,52], responsible for chemical adsorption of CO2 [2]. On closer inspection, there is relatively small peak at around 115 °C for all Ni/rGO catalysts, indicating   Figure 7 shows the CO 2 -TPD spectra of the Ni/rGO catalysts and rGO support. NiO is also evaluated for control purposes. CO 2 -TPD evaluates the CO 2 adsorption and activation capability of the catalyst for methanation. rGO does not show any significant desorption, indicating that it contains only a small number of basic sites. On the other hand, NiO shows a broad signal for the γ peak at 583 • C. When Ni is supported on rGO, a sharp and intense desorption peak at~550 • C appears. This γ peak corresponds to the strong basic sites [49,52], responsible for chemical adsorption of CO 2 [2]. On closer inspection, there is relatively small peak at around 115 • C for all Ni/rGO catalysts, indicating the weak basic sites. From this result, it is therefore deduced that the basic sites of the catalyst contribute to the presence of Ni active metal which is basic in nature. The low signal for NiO in the chromatogram is most probably due to the agglomeration of the Ni particles [3]. Once supported, the dispersion of Ni is enhanced, and new interactions based on Lewis acid-base interaction of electron transfer between Ni and support were created which led to higher number of basic sites. The desorption temperature of CO 2 from the catalyst is tabulated in Table 3. As shown in Table 3, the desorption temperature of Ni 15 /rGO is the highest since it has strongest basic sites and metal-support interaction.
Catalysts 2020, 10, 471 8 of 15 the weak basic sites. From this result, it is therefore deduced that the basic sites of the catalyst contribute to the presence of Ni active metal which is basic in nature. The low signal for NiO in the chromatogram is most probably due to the agglomeration of the Ni particles [3]. Once supported, the dispersion of Ni is enhanced, and new interactions based on Lewis acid-base interaction of electron transfer between Ni and support were created which led to higher number of basic sites. The desorption temperature of CO2 from the catalyst is tabulated in Table 3. As shown in Table 3, the desorption temperature of Ni15/rGO is the highest since it has strongest basic sites and metal-support interaction. The FESEM-EDX mapping images of Ni/rGO catalysts are displayed in Figure 8. The multilayer nanosheets of graphene can be clearly observed in each FESEM image of the catalysts. From the EDX mapping, it can be observed that the Ni is dispersed well on the surface of graphene. This is due to the high surface area and porous structure of graphene. The higher density of Ni on the graphene is due to higher Ni content loaded on the graphene. As the Ni content increase, there could be agglomeration occurred for Ni20/rGO catalyst as observed in the EDX mapping. Agglomeration is also proven from the lowest surface area and high Ni crystallite size of Ni20/rGO among other Ni/rGO catalysts. The FESEM-EDX mapping images of Ni/rGO catalysts are displayed in Figure 8. The multilayer nanosheets of graphene can be clearly observed in each FESEM image of the catalysts. From the EDX mapping, it can be observed that the Ni is dispersed well on the surface of graphene. This is due to the high surface area and porous structure of graphene. The higher density of Ni on the graphene is due to higher Ni content loaded on the graphene. As the Ni content increase, there could be agglomeration occurred for Ni 20 /rGO catalyst as observed in the EDX mapping. Agglomeration is also proven from the lowest surface area and high Ni crystallite size of Ni 20 /rGO among other Ni/rGO catalysts.

Catalytic Activity
The catalytic activity of Ni/rGO catalyst is presented in Table 3, outlining the CO2 conversion and CH4 selectivity. For comparison purposes, Raney nickel was also studied for the catalytic conversion of CO2 into CH4. NiO has low conversion of CO2 which is only 5.7%, suggesting the necessity of a catalyst support in the CO2 methanation. The low conversion of CO2 on NiO nanoparticles is related to the lower number of basic sites and hydrogen consumption of the catalysts due to agglomeration of the particles. After being supported on rGO, the conversion of CO2 and product selectivity to methane was notably increased. This is possibly due to the presence of rGO which improve the dispersion of Ni and reduces agglomeration. Consequently, the specific activity per Ni atom increases for CO2 methanation. The trend for the activities of the catalysts is directly correlated with their respective SBET, where Ni15/rGO shows the highest conversion. The higher the surface area, the higher the conversion because the number of active sites increases due to the presence of more pores. The trend is also related to the basic sites which act as the active sites for reaction and the strength of the metal-support interactions. As shown in the CO2-TPD and H2-TPR result, Ni15/rGO has the strongest basic sites due to the high temperature for CO2 desorption and high reduction temperature. This indicates that CO2 first dissociates on the basic sites of the catalyst prior to its conversion into methane. Previous studies found that Ni metal became the active site for hydrogen dissociation which was then added to C form methane [53]. A graphical representation of the conversion of CO2 into methane on the Ni/rGO catalyst from this study is displayed in Figure 9. Overall, both physicochemical and activity profiles of catalysts were significantly affected by the variation of Ni content. Physicochemical investigations suggested 15 wt% of Ni as an optimum amount for the rGO-supported Ni catalyst. Activity data confirmed the Ni15/rGO catalyst as the most

Catalytic Activity
The catalytic activity of Ni/rGO catalyst is presented in Table 3, outlining the CO 2 conversion and CH 4 selectivity. For comparison purposes, Raney nickel was also studied for the catalytic conversion of CO 2 into CH 4 . NiO has low conversion of CO 2 which is only 5.7%, suggesting the necessity of a catalyst support in the CO 2 methanation. The low conversion of CO 2 on NiO nanoparticles is related to the lower number of basic sites and hydrogen consumption of the catalysts due to agglomeration of the particles. After being supported on rGO, the conversion of CO 2 and product selectivity to methane was notably increased. This is possibly due to the presence of rGO which improve the dispersion of Ni and reduces agglomeration. Consequently, the specific activity per Ni atom increases for CO 2 methanation. The trend for the activities of the catalysts is directly correlated with their respective S BET , where Ni 15 /rGO shows the highest conversion. The higher the surface area, the higher the conversion because the number of active sites increases due to the presence of more pores. The trend is also related to the basic sites which act as the active sites for reaction and the strength of the metal-support interactions. As shown in the CO 2 -TPD and H 2 -TPR result, Ni 15 /rGO has the strongest basic sites due to the high temperature for CO 2 desorption and high reduction temperature. This indicates that CO 2 first dissociates on the basic sites of the catalyst prior to its conversion into methane. Previous studies found that Ni metal became the active site for hydrogen dissociation which was then added to C form methane [53]. A graphical representation of the conversion of CO 2 into methane on the Ni/rGO catalyst from this study is displayed in Figure 9. Overall, both physicochemical and activity profiles of catalysts were significantly affected by the variation of Ni content. Physicochemical investigations suggested 15 wt% of Ni as an optimum amount for the rGO-supported Ni catalyst. Activity data confirmed the Ni 15 /rGO catalyst as the most efficient catalyst in terms of methane productivity (24.9 g/kg·cat·h), followed by Ni 10 /rGO which has high TOF. Ni 10 /rGO has highest TOF as it has the lowest number of basic sites.
Catalysts 2020, 10, 471 10 of 15 efficient catalyst in terms of methane productivity (24.9 g/kg·cat·h), followed by Ni10/rGO which has high TOF. Ni10/rGO has highest TOF as it has the lowest number of basic sites. Figure 9. CO2 conversion to CH4 on Ni/rGO catalyst.
Apart from that, a comparative study of the current catalyst with the recently reported data [16,41,42,48] for CO2 hydrogenation methanation is presented in Table 4. As depicted from the tabulated data, the synthesized catalyst showed good activity for methane synthesis and good CO2 conversion as compared to the catalysts supported on other conventional supports such as Al2O3, SiO2, alumina, MCM-41, SBA-15 and MSU-F. In this work, Ni/rGO exhibited high selectivity which is due to its high CO2 desorption temperature and CO2 uptake of the catalysts. The high selectivity is possibly due to the low temperature used in this reaction because the thermodynamic analysis of CO2 methanation revealed that a lower temperature of reaction suppresses the production of byproducts especially CO [8,54,55]. Table 4. Activity comparison of the synthesized Ni-based catalyst with recent reported data.

Materials
Commercial reduced graphene oxide (rGO) from Sigma Aldrich (MO, USA cat. 796034) was used as received and was characterized thoroughly to understand its properties as outlined in Section 2.3. Nickel nitrate hexahydrate (Ni(NO3)2·6H2O), was obtained from Merck (Darmstadt, Germany). All the solutions were prepared using distilled water.

Synthesis of Ni-Based Catalyst
NiO catalyst was synthesized based on precipitation method [56] in which 3.96 g (0.012 mol) of Ni(NO3)2·6H2O was added to 20 mL distilled water. The solution was constantly stirred with the Apart from that, a comparative study of the current catalyst with the recently reported data [16,41,42,48] for CO 2 hydrogenation methanation is presented in Table 4. As depicted from the tabulated data, the synthesized catalyst showed good activity for methane synthesis and good CO 2 conversion as compared to the catalysts supported on other conventional supports such as Al 2 O 3 , SiO 2 , alumina, MCM-41, SBA-15 and MSU-F. In this work, Ni/rGO exhibited high selectivity which is due to its high CO 2 desorption temperature and CO 2 uptake of the catalysts. The high selectivity is possibly due to the low temperature used in this reaction because the thermodynamic analysis of CO 2 methanation revealed that a lower temperature of reaction suppresses the production of byproducts especially CO [8,54,55].

Materials
Commercial reduced graphene oxide (rGO) from Sigma Aldrich (St. Louis, MO, USA cat. 796034) was used as received and was characterized thoroughly to understand its properties as outlined in Section 3.3. Nickel nitrate hexahydrate (Ni(NO 3 ) 2 ·6H 2 O), was obtained from Merck (Darmstadt, Germany). All the solutions were prepared using distilled water.

Synthesis of Ni-Based Catalyst
NiO catalyst was synthesized based on precipitation method [56] in which 3.96 g (0.012 mol) of Ni(NO 3 ) 2 ·6H 2 O was added to 20 mL distilled water. The solution was constantly stirred with the simultaneous addition of 20 mL of 15% NaOH solution. After 15 min, a green precipitate of Ni(OH) 2 was obtained that was separated using vacuum filtration and dried at 100 • C. Finally, the sample was calcined at 400 • C for 2 h, producing NiO black powder.
Ni/rGO catalysts were synthesized via the incipient wetness impregnation method. Initially, 10, 15 and 20 wt.% (relative to the weight of rGO) of Ni precursors were added dropwise to commercial rGO support using an aqueous solution prepared by dissolving required amount of Ni(NO 3 ) 2 ·3H 2 O. The volume of impregnating solution was adjusted to the volume of the support. The impregnated sample was left for 12 h. Then, it was dried at 80 • C for 12 h in oven before calcined at temperature 400 • C for 2 h. Each catalyst was characterized as described in Section 3.3.

Material Characterization
Thermogravimetric analysis (TGA) of rGO support was conducted using a TGA/DSC STARe thermogravimetric analyser (Mettler Toledo, Columbus, OH, USA) equipped with a GC10 gas controller over the temperature range of 25 • C to 900 • C (heating ramp: 20 • C/min) in a N 2 atmosphere. The purpose of TGA was to understand the thermal properties and analyse weight loss profile of rGO support which is used for determination of catalyst calcination temperature.
The Raman spectrum of rGO was obtained using a HR800 spectrometer Horiba Jobin Yvon (Kyoto, Japan) with a 514 nm green laser as excitation source in the Raman shift of 400 to 4000 cm −1 range to identify the structure and degree of disorder in carbon-based material of the rGO support.
X-ray diffractograms for rGO and Ni/rGO catalysts were recorded on a powder diffractometer (X'Pert 3 Powder & Empyrean, PANalytical, Malver, UK) with Cu Kα radiation source between 2θ of 5 • to 90 • (scanning step of 0.01 • /step) on continuous scanning. Interlayer spacing (d), crystal stack height (L c ), crystallite size (L a ) and estimated number of layers (N GP ) were obtained from Equations (2) to (5) respectively: λ = 2d sinθ (2) where λ is wavelength of Cu Kα radiation (1.5418 Å), K is shape factor (0.89), K 2 is Warren Form Factor (1.84) and B is line broadening at half maximum of peak. Surface area and porosity (SAP) analysis of rGO, and Ni/rGO catalysts were performed using an ASAP 2020 analyser (Micromeritics, Norcross, GA, USA) with N 2 as adsorbate. The surface area was determined from the BET model, pore size and volume were determined using the BJH method and micropore analysis was carried out using t-plot analysis.
A Fourier-transform infrared (FTIR) spectrophotometer (Spectrum One, Perkin Elmer, Waltham, MA, USA) was used to study the functional groups of the rGO support and Ni 15 /rGO catalyst in the range of 400-4000 cm −1 with a resolution of 4 cm −1 . Sample pellets were prepared by grinding with KBr.
Hydrogen temperature-programmed reduction (H 2 -TPR) of the catalysts was carried out using a TPD/R/O 110 MS (Thermo scientific, Waltham, MA, USA) equipped with a thermal conductivity detector (TCD) to evaluate the reducibility and metal-support interaction of the catalysts. The analysis was performed in temperature range of 30 • C to 800 • C at heating rate of 10 • C/min in the gas flow of 5 vol.% H 2 /N 2 ratio with flow rate of 20 mL/min. CO 2 temperature-programmed desorption (CO 2 -TPD) of rGO, Ni/rGO and NiO was carried out on a TPD/R/O (1100CE) instrument equipped with TCD. The sample was pretreated at 100 • C under helium flow (20 mL/min) and cooled down to 75 • C. Sorption study was carried out using CO 2 gas (10 mL/min) for 30 min. Then, temperature was decreased to 40 • C. Then, desorption was performed using He (20 mL/min) in the temperature range of 40-800 • C at a heating rate of 10 • C/min.
The surface morphologies of Ni/rGO catalysts were observed by field emission scanning electron microscopy equipped with energy dispersive X-ray spectroscopy (FESEM-EDX). The analysis was performed using a Supra55 Variable Pressure (VP) system (Zeiss, Jena, Germany) under 5 kV accelerating voltage, magnification range of 5 to 100 k× and 4 nm working distance.

Catalytic Activity of Ni/rGO
The scheme of the methanation setup is presented in Figure 10. Each catalyst was reduced beforehand using H 2 -temperature programmed reduction (H 2 -TPR) under 5% H 2 in argon at the total rate of 20 mL/min from temperature 50 to 500 • C with a heating ramp of 10 • C/min. CO 2 methanation was performed for 2 h in a 100 mL autoclave slurry reactor (Parr 4593 with a regular Parr 4848, Parr, Moline, IL, USA) at 10 bar with 240 • C reaction temperature and 500 rpm agitator speed. Feed gas was supplied to the reactor in 1:4 molar ratio of CO 2 :H 2 . Catalyst (0.5 g) was placed in 40 mL dodecane liquid phase to form a 1.25% (w/v) slurry of catalyst in solvent. A natural gas analyzer (Perkin Elmer Clarus 580 GC) was used for the analysis of reagents and product using ASTMD 1945 standard method. The reaction condition was chosen based on previous literature [22,57] that used slurry reactor with some modification at lower temperature. CO 2 conversion, CH 4 selectivity, space time yield (STY) and turnover frequency (TOF) are calculated from the respective equations below: S CH 4 (%) = n CH 4,out n product out × 100% (7) STY = Quantity of methane (g) Weight of catalyst (kg.cat)×time (h) (8) TOF = Mole of CO 2 consumed Number of basic sites (mmol·g cat)×mass of catalyst (g)×time (s) (9) Catalysts 2020, 10, 471 12 of 15 The surface morphologies of Ni/rGO catalysts were observed by field emission scanning electron microscopy equipped with energy dispersive X-ray spectroscopy (FESEM-EDX). The analysis was performed using a Supra55 Variable Pressure (VP) system (Zeiss,Jena, Germany) under 5 kV accelerating voltage, magnification range of 5 to 100 k× and 4 nm working distance.

Catalytic Activity of Ni/rGO
The scheme of the methanation setup is presented in Figure 10. Each catalyst was reduced beforehand using H2-temperature programmed reduction (H2-TPR) under 5% H2 in argon at the total rate of 20 mL/min from temperature 50 to 500 °C with a heating ramp of 10 °C/min. CO2 methanation was performed for 2 h in a 100 mL autoclave slurry reactor (Parr 4593 with a regular Parr 4848, Parr, IL, USA) at 10 bar with 240 °C reaction temperature and 500 rpm agitator speed. Feed gas was supplied to the reactor in 1:4 molar ratio of CO2:H2. Catalyst (0.5 g) was placed in 40 mL dodecane liquid phase to form a 1.25% (w/v) slurry of catalyst in solvent. A natural gas analyzer (Perkin Elmer Clarus 580 GC) was used for the analysis of reagents and product using ASTMD 1945 standard method. The reaction condition was chosen based on previous literature [22,57] that used slurry reactor with some modification at lower temperature. CO2 conversion, CH4 selectivity, space time yield (STY) and turnover frequency (TOF) are calculated from the respective equations below: Initial experiments were repeated two times to check for reproducibility. Gas chromatograph analysis are, in general, reproducible within a maximum of 6% but mostly within a few percent.

Conclusions
Ni/rGO catalysts with three different Ni loadings were successfully synthesized via the incipient wetness impregnation method and their catalytic activity towards CO2 methanation was assessed. It Initial experiments were repeated two times to check for reproducibility. Gas chromatograph analysis are, in general, reproducible within a maximum of 6% but mostly within a few percent.

Conclusions
Ni/rGO catalysts with three different Ni loadings were successfully synthesized via the incipient wetness impregnation method and their catalytic activity towards CO 2 methanation was assessed. It was found that the variation of Ni loading affected both the physicochemical properties and catalytic activity. A Ni loading of 15 wt% relative to the weight of rGO has the highest surface area and strongest basic sites. Due to these properties, Ni 15 /rGO shows highest conversion of CO 2 . The high surface area of the catalysts provide active sites for the chemical reaction whereas strong basic sites of catalysts serve as sites for CO 2 dissociation.