CO and CO 2 Co-Methanation on Ni / CeO 2 -ZrO 2 Soft-Templated Catalysts

: Supported nickel catalysts were synthesized, characterized, and employed in the carbon oxides co-methanation process. Five NiO / CeO 2 -ZrO 2 mixed oxides, with the same Ni content and di ﬀ erent Ce / Zr molar ratios, were prepared by the soft-template method. They were characterized through ICP-AES, N 2 adsorption, XRD, and TPR. Reduced Ni / CeO 2 -ZrO 2 catalysts were obtained by submitting the oxide systems to reduction treatment in H 2 at 400 ◦ C. They were characterized by XRD, H 2 -TPD, and CO 2 adsorption microcalorimetry and their catalytic performances in the carbon oxides co-methanation were investigated. Catalytic tests were performed in a ﬁxed-bed continuous-ﬂow microreactor at atmospheric pressure. The e ﬀ ect of experimental conditions (reaction temperature, space velocity, reactants molar ratio) was also studied. Almost complete CO conversion was obtained on any catalyst, whereas CO 2 conversion was much lower and increased with Ce content, at least up to Ce / Zr = 1. The beneﬁcial e ﬀ ect of the Ce content could be related to the increased NiO reducibility and to the higher ability to adsorb and activate CO 2 . However, at high Ce / Zr ratios, it is probably counterbalanced by an interplay of reactions involving CO and CO 2 .

Supported nickel-based catalysts are generally very active and selective in carbon oxides hydrogenation reactions; the additional advantage of being reasonably inexpensive makes them promising alternatives to noble metals-based catalysts (e.g., supported Ru or Rh). As for the co-methanation process, several oxidic systems have been used as a support for nickel in combination with another metal [45,47,51,52]. For supporting undoped nickel, besides Al 2 O 3 [47,51], CeO 2 -Al 2 O 3 [50], and Gd 2 O 3 -CeO 2 [52] systems, zirconia-pure [43,46] or in combination with other

Characterization of Fresh NiO-CeO 2 -ZrO 2 Mixed Oxides
NiO-CeO 2 -ZrO 2 mixed oxides with a nominal NiO content of 30 wt% and different Ce/Zr molar ratios were synthesized by means of the soft-template method and will hereafter be referred to as NiO/Ce x Zr 1−x O 2 , where x is the nominal CeO 2 molar fraction in the (CeO 2 ) x -(ZrO 2 ) 1−x support. Their chemical composition was determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES). Data reported in Table 1 show that the experimental composition is close to the nominal one for all the samples. Textural analysis was carried out by determining the nitrogen adsorption-desorption isotherms at −196 • C, which are shown in Figure 1. All the samples are characterized by a type IVa N 2 adsorption isotherm [54]; however, increasing Zr content results in a higher microporous contribution (at low p/p 0 ) and a more linear rise at high relative pressure (p/p 0 > 0.6). The pore size distribution curves, obtained by applying the BJH method to the isotherm adsorption branch [54,55], are displayed in the insets in Figure 1. They show a general decrease in pore size with increasing Zr content. In particular, at low Ce contents in the (CeO 2 ) x -(ZrO 2 ) 1−x support (x ≤ 0.25) the contribution of pores wider than 10 nm becomes negligible. Specific surface area calculated using the BET equation (S BET ) and specific pore volume (V p ) are reported in Table 1. The pure zirconia-supported sample is characterized by a high surface area, which generally decreases with increasing Ce content.  The X-ray diffraction (XRD) patterns of the fresh NiO-CeO 2 -ZrO 2 systems are shown in Figure 2. In all samples, NiO crystalline phase (PDF Card 441159) can be identified, and from the peak width (by means of the Scherrer equation [56]) an average crystallite size of ca. 4 nm can be calculated. In the XRD patterns of NiO/ZrO 2 and NiO/CeO 2 , signals ascribable to the corresponding pure oxide supports are also present. However, while the typical signals of nanocrystalline cubic CeO 2 phase (PDF Card 750151) can be identified, from which an average crystalline size of ca. 3 nm can be calculated, only one broad band ascribable to amorphous zirconia can be observed. In all the samples containing both Ce and Zr, besides the pure oxides (nanocrystalline ceria and amorphous zirconia), a ceria-zirconia solid solution is possibly present. In fact, signals corresponding to a cubic CeO 2 -like phase are visible. With increasing Zr content, they become wider and shift towards higher angles, suggesting the (at least partial) incorporation of the small zirconium ions into the ceria lattice, which would lead to the formation of Ce x Zr 1−x O 2 solid solutions characterized by smaller crystallite size and lattice parameter. The X-ray diffraction (XRD) patterns of the fresh NiO-CeO2-ZrO2 systems are shown in Figure  2. In all samples, NiO crystalline phase (PDF Card 441159) can be identified, and from the peak width (by means of the Scherrer equation [56]) an average crystallite size of ca. 4 nm can be calculated. In the XRD patterns of NiO/ZrO2 and NiO/CeO2, signals ascribable to the corresponding pure oxide supports are also present. However, while the typical signals of nanocrystalline cubic CeO2 phase (PDF Card 750151) can be identified, from which an average crystalline size of ca. 3 nm can be calculated, only one broad band ascribable to amorphous zirconia can be observed. In all the samples containing both Ce and Zr, besides the pure oxides (nanocrystalline ceria and amorphous zirconia), a ceria-zirconia solid solution is possibly present. In fact, signals corresponding to a cubic CeO2-like phase are visible. With increasing Zr content, they become wider and shift towards higher angles, suggesting the (at least partial) incorporation of the small zirconium ions into the ceria lattice, which would lead to the formation of CexZr1-xO2 solid solutions characterized by smaller crystallite size and lattice parameter.

Reducibility of NiO-CeO2-ZrO2 Mixed Oxides And Characterization of the Reduced Samples
Reducibility of NiO-CeO2-ZrO2 mixed oxides was investigated by means of temperature-programmed reduction (TPR) analyses, whose results are reported in Figure 3. The TPR profile of NiO/ZrO2 shows a single signal (maximum at ca. 590 °C) which -taking into account that soft-templated pure zirconia did not give rise to any significant hydrogen consumption [42] can be ascribed to the reduction of the NiO species present in the sample. However, when compared with the position of the maximum (ca. 500 °C) in the TPR profile of a soft-templated pure NiO sample [42], such signal appears shifted to higher temperature, thus indicating the presence of strong ZrO2-NiO interactions, which make the nickel oxide species more difficult to reduce. The TPR curves of the cerium-containing samples show a low-temperature (150-350 °C) small signal and a major one at higher temperatures. Both signals result from the overlapping of different contributions, which indicates the presence of different reducible species. Although also CeO2 species can be reduced, most hydrogen consumption is expected to be due to the reduction of NiO species [42]. The features of the NiO/CeO2 TPR profile can be described in the light of the literature on NiO-CeO2 systems [19,42,[57][58][59]. The low-temperature feature can be ascribed to the loss of oxygen species adsorbed on defective sites at the NiO-CeO2 interface, where Ce and Ni are in close interaction. The high-temperature signal clearly results from the partial overlapping of at least two

Reducibility of NiO-CeO 2 -ZrO 2 Mixed Oxides And Characterization of the Reduced Samples
Reducibility of NiO-CeO 2 -ZrO 2 mixed oxides was investigated by means of temperature-programmed reduction (TPR) analyses, whose results are reported in Figure 3. The TPR profile of NiO/ZrO 2 shows a single signal (maximum at ca. 590 • C) which-taking into account that soft-templated pure zirconia did not give rise to any significant hydrogen consumption [42]-can be ascribed to the reduction of the NiO species present in the sample. However, when compared with the position of the maximum (ca. 500 • C) in the TPR profile of a soft-templated pure NiO sample [42], such signal appears shifted to higher temperature, thus indicating the presence of strong ZrO 2 -NiO interactions, which make the nickel oxide species more difficult to reduce. The TPR curves of the cerium-containing samples show a low-temperature (150-350 • C) small signal and a major one at higher temperatures. Both signals result from the overlapping of different contributions, which indicates the presence of different reducible species. Although also CeO 2 species can be reduced, most hydrogen consumption is expected to be due to the reduction of NiO species [42]. The features of the NiO/CeO 2 TPR profile can be described in the light of the literature on NiO-CeO 2 systems [19,42,[57][58][59]. The low-temperature feature can be ascribed to the loss of oxygen species adsorbed on defective sites at the NiO-CeO 2 interface, where Ce and Ni are in close interaction. The high-temperature signal clearly results from the partial overlapping of at least two contributions, which can be assigned to the reduction of nickel oxide particles dispersed on the ceria surface and to NiO species strongly interacting with CeO 2 . Contributions similar to those distinguished for the pure ceria-supported sample are also present in the TPR curves of the systems containing both Ce and Zr, although with different relative areas. contributions, which can be assigned to the reduction of nickel oxide particles dispersed on the ceria surface and to NiO species strongly interacting with CeO2. Contributions similar to those distinguished for the pure ceria-supported sample are also present in the TPR curves of the systems containing both Ce and Zr, although with different relative areas. The synthetized systems were also characterized after being submitted to the same reduction pretreatment that they undergo prior to the catalytic tests (flowing H2 at 400 °C for 1 h). The reduced samples will hereafter be referred to as Ni/CexZr1-xO2. The XRD patterns of the hydrogen-treated systems are reported in Figure 4. As expected, the signals ascribable to the oxidic supports (amorphous zirconia, cubic ceria, and cubic solid solutions) are the same as those of the fresh NiO/CexZr1-xO2 samples (cf. Figure 2). For all the reduced systems, signals of Ni 0 phase (PDF Card 040850) are present. The metallic nickel average crystal size, calculated by means of the Scherrer equation, is ca. 6 nm for all the Ce-containing samples, and lower than 5 nm for Ni/ZrO2. Such small values indicate that the strong interaction with the support can hinder the sintering of the metal nanocrystals. The asymmetry of the main signals in the Ni/ZrO2 pattern is probably due to some unreduced NiO, whose presence also in the other samples cannot be excluded. The synthetized systems were also characterized after being submitted to the same reduction pretreatment that they undergo prior to the catalytic tests (flowing H 2 at 400 • C for 1 h). The reduced samples will hereafter be referred to as Ni/Ce x Zr 1−x O 2 . The XRD patterns of the hydrogen-treated systems are reported in Figure 4. As expected, the signals ascribable to the oxidic supports (amorphous zirconia, cubic ceria, and cubic solid solutions) are the same as those of the fresh NiO/Ce x Zr 1−x O 2 samples (cf. Figure 2). For all the reduced systems, signals of Ni 0 phase (PDF Card 040850) are present. The metallic nickel average crystal size, calculated by means of the Scherrer equation, is ca. 6 nm for all the Ce-containing samples, and lower than 5 nm for Ni/ZrO 2 . Such small values indicate that the strong interaction with the support can hinder the sintering of the metal nanocrystals. The asymmetry of the main signals in the Ni/ZrO 2 pattern is probably due to some unreduced NiO, whose presence also in the other samples cannot be excluded. The H2 adsorption properties where studied through hydrogen temperature-programmed desorption (H2-TPD) experiments. It resulted to be 55 mol g -1 for Ni/ZrO2 and to progressively increase with the Ce content up to 115 mol g -1 for Ni/CeO2. Probably the simple adsorption on the metallic Ni 0 surface is not the only cause for the total amount of hydrogen adsorbed, a contribution of spillover being expected. However, since the nickel content and the size of metal nanocrystals are similar for all the samples, the increasing trend of hydrogen adsorption with Ce content could mainly be explained by an increase in the NiO reduction degree at increasing Ce content. This would be in accordance with the TPR results and would imply that, even though no signals are clearly detectable in the XRD patterns, NiO phase is still present in the hydrogen-treated samples.
CO2 adsorption properties of the Ni/CexZr1-xO2 samples were studied through adsorption microcalorimetry. The differential heat of adsorption (Qdiff) is an indication of the strength of the adsorbent-adsorbate interaction, whereas CO2 uptake is a measure of the number of adsorbing sites. Therefore, reporting Qdiff as a function of CO2 uptake ( Figure 5) simultaneous information on the concentration of the adsorbing sites and on their strength distribution is obtained. Whatever the sample, strong sites (Qdiff  150 kJ mol -1 ) are present in low concentration. As CO2 uptake increases Qdiff decreases down to 50 kJ mol -1 (below which adsorption cannot be considered specific). For Ni/ZrO2 such decrease is very steep and the total amount of CO2 adsorbed is ca. 35 mol g -1 . On the other hand, Ni/CeO2 is able to adsorb a much higher amount of CO2 (ca. 230 mol g -1 ), most of which on medium-strength sites (75 kJ mol -1  Qdiff  150 kJ mol -1 ). These results, which are in agreement with those obtained on the pure oxidic phases [42], indicate the superior ability of CeO2 -in comparison with ZrO2 -in adsorbing and activating CO2. The systems containing both Ce and Zr show an intermediate behavior, in terms of both concentration of adsorbing sites and strength of interaction. The general enhancement in the adsorption of CO2 at increasing Ce content is not followed by Ni/Ce0.50Zr0.50O2 and Ni/Ce0.75Zr0.25O2, whose curves practically overlap. However, it has to be taken into account that the adsorption capacity also depends on the surface area, which for the reduced samples is expected to follow the same trend observed for the parent fresh oxide samples (Table 1). Compared to the sample with the same content of Ce and Zr, the Ce-richer sample is most probably characterized by a significantly lower surface area, which would counterbalance the higher capacity of the ceria component to adsorb CO2. The H 2 adsorption properties where studied through hydrogen temperature-programmed desorption (H 2 -TPD) experiments. It resulted to be 55 µmol g −1 for Ni/ZrO 2 and to progressively increase with the Ce content up to 115 µmol g −1 for Ni/CeO 2 . Probably the simple adsorption on the metallic Ni 0 surface is not the only cause for the total amount of hydrogen adsorbed, a contribution of spillover being expected. However, since the nickel content and the size of metal nanocrystals are similar for all the samples, the increasing trend of hydrogen adsorption with Ce content could mainly be explained by an increase in the NiO reduction degree at increasing Ce content. This would be in accordance with the TPR results and would imply that, even though no signals are clearly detectable in the XRD patterns, NiO phase is still present in the hydrogen-treated samples. CO 2 adsorption properties of the Ni/Ce x Zr 1−x O 2 samples were studied through adsorption microcalorimetry. The differential heat of adsorption (Q diff ) is an indication of the strength of the adsorbent-adsorbate interaction, whereas CO 2 uptake is a measure of the number of adsorbing sites. Therefore, reporting Q diff as a function of CO 2 uptake ( Figure 5) simultaneous information on the concentration of the adsorbing sites and on their strength distribution is obtained. Whatever the sample, strong sites (Q diff ≥ 150 kJ mol −1 ) are present in low concentration. As CO 2 uptake increases Q diff decreases down to 50 kJ mol −1 (below which adsorption cannot be considered specific). For Ni/ZrO 2 such decrease is very steep and the total amount of CO 2 adsorbed is ca. 35 µmol g −1 . On the other hand, Ni/CeO 2 is able to adsorb a much higher amount of CO 2 (ca. 230 µmol g −1 ), most of which on medium-strength sites (75 kJ mol −1 ≤ Q diff ≤ 150 kJ mol −1 ). These results, which are in agreement with those obtained on the pure oxidic phases [42], indicate the superior ability of CeO 2 -in comparison with ZrO 2 -in adsorbing and activating CO 2 . The systems containing both Ce and Zr show an intermediate behavior, in terms of both concentration of adsorbing sites and strength of interaction. The general enhancement in the adsorption of CO 2 at increasing Ce content is not followed by Ni/Ce 0.50 Zr 0.50 O 2 and Ni/Ce 0.75 Zr 0.25 O 2 , whose curves practically overlap. However, it has to be taken into account that the adsorption capacity also depends on the surface area, which for the reduced samples is expected to follow the same trend observed for the parent fresh oxide samples (Table 1). Compared to the sample with the same content of Ce and Zr, the Ce-richer sample is most probably characterized by a significantly lower surface area, which would counterbalance the higher capacity of the ceria component to adsorb CO 2 .

CO and CO2 Co-Methanation Catalytic Tests
Simultaneous CO and CO2 hydrogenation to SNG was carried out at 300 °C and atmospheric pressure on the Ni/CexZr1-xO2 samples, obtained in the reduced form by in situ pretreatment (flowing H2 at 400 °C for 1 h). A reactant gas mixture CO/CO2/H2 of molar composition of 1/1/5 was fed with a space velocity (SV) of 150,000 cm 3 h −1 gcat −1 . The results, in terms of CO and CO2 conversions and methane selectivity, are summarized in Table 2. Since, during the 6-h runs, catalytic activity was found to be stable with time-on-stream (t.o.s) for all the samples, average values are reported. Irrespective of the catalyst composition, a CH4 selectivity higher than 99 mol% is always observed. All the catalysts exhibit very high CO conversion, which seems to be slightly favored by high Ce contents. CO2 conversion is much lower and the beneficial effect of Ce is apparent up to an equimolar content of Ce and Zr, a further enrichment in Ce not producing any significant effect.
On Ni/Ce0.50Zr0.50O2, a long-time run at variable temperature was carried out, to study the catalytic stability under thermal stress. As shown in Figure 6, both CO and CO2 conversions decrease at increasing temperature, because of the exothermic character of both the CO and CO2 methanation reactions. However, when the reaction temperature is set back at 300 °C, the catalytic activity is

CO and CO 2 Co-Methanation Catalytic Tests
Simultaneous CO and CO 2 hydrogenation to SNG was carried out at 300 • C and atmospheric pressure on the Ni/Ce x Zr 1−x O 2 samples, obtained in the reduced form by in situ pretreatment (flowing H 2 at 400 • C for 1 h). A reactant gas mixture CO/CO 2 /H 2 of molar composition of 1/1/5 was fed with a space velocity (SV) of 150,000 cm 3 h −1 g cat −1 . The results, in terms of CO and CO 2 conversions and methane selectivity, are summarized in Table 2. Since, during the 6-h runs, catalytic activity was found to be stable with time-on-stream (t.o.s) for all the samples, average values are reported. Irrespective of the catalyst composition, a CH 4 selectivity higher than 99 mol% is always observed. All the catalysts exhibit very high CO conversion, which seems to be slightly favored by high Ce contents. CO 2 conversion is much lower and the beneficial effect of Ce is apparent up to an equimolar content of Ce and Zr, a further enrichment in Ce not producing any significant effect.
On Ni/Ce 0.50 Zr 0.50 O 2 , a long-time run at variable temperature was carried out, to study the catalytic stability under thermal stress. As shown in Figure 6, both CO and CO 2 conversions decrease at increasing temperature, because of the exothermic character of both the CO and CO 2 methanation reactions. However, when the reaction temperature is set back at 300 • C, the catalytic activity is restored, indicating a very good thermal stability of the catalyst. Methane selectivity (not reported) was always higher than 99 mol%.
On Ni/Ce 0.50 Zr 0.50 O 2 the influence of both feed composition and space velocity was also investigated, by performing co-methanation at 300 • C with two different feed compositions (CO/CO 2 /H 2 = 1/1/5 or 1/1/7) and three different space velocities (SV = 50,000, 150,000, or 450,000 cm 3 h −1 g cat −1 ). Results reported in Figure 7 show that when a higher H 2 concentration is used CO 2 conversion significantly increases, whereas only a very slight improvement is observed for the already very high CO conversion. Whatever the feed composition, both X CO and X CO2 decrease at increasing SV. However, whereas only small decreases in CO conversion are observed, the effect of SV on CO 2 conversion is particularly important, indicating that kinetic limitations are far more significant for CO 2 methanation. Noteworthy, for X CO2 the relative decrease is practically the same (ca. 37%) with either feed compositions. Also in these runs, methane selectivity (not reported) was always higher than 99 mol%.
Catalysts 2019, 9, x FOR PEER REVIEW 8 of 15 restored, indicating a very good thermal stability of the catalyst. Methane selectivity (not reported) was always higher than 99 mol%. On Ni/Ce0.50Zr0.50O2 the influence of both feed composition and space velocity was also investigated, by performing co-methanation at 300 °C with two different feed compositions (CO/CO2/H2 = 1/1/5 or 1/1/7) and three different space velocities (SV = 50,000, 150,000, or 450,000 cm 3 h −1 gcat −1 ). Results reported in Figure 7 show that when a higher H2 concentration is used CO2 conversion significantly increases, whereas only a very slight improvement is observed for the already very high CO conversion. Whatever the feed composition, both XCO and XCO2 decrease at increasing SV. However, whereas only small decreases in CO conversion are observed, the effect of SV on CO2 conversion is particularly important, indicating that kinetic limitations are far more significant for CO2 methanation. Noteworthy, for XCO2 the relative decrease is practically the same (ca. 37%) with either feed compositions. Also in these runs, methane selectivity (not reported) was always higher than 99 mol%.

Discussion
The NiO-CeO2-ZrO2 mixed oxides samples synthetized by the soft-template method are characterized by high surface areas and nanometric sizes of pores and crystallites. The strong interaction between the dispersed NiO phase and the ZrO2 support makes the NiO nanoparticles

Discussion
The NiO-CeO 2 -ZrO 2 mixed oxides samples synthetized by the soft-template method are characterized by high surface areas and nanometric sizes of pores and crystallites. The strong interaction between the dispersed NiO phase and the ZrO 2 support makes the NiO nanoparticles reduction more difficult than for the unsupported nickel oxide. When Ce is present different reducible species are formed and reducibility generally increases. XRD results indicate the reduction of NiO nanocrystals under the experimental pretreatment conditions. However, H 2 -TPD experiments suggest that such reduction is not complete, though favored by increasing Ce content.
In the carbon oxides co-methanation, a competition between CO and CO 2 hydrogenation reactions occurs. In the catalytic runs an almost complete CO conversion is obtained, whereas CO 2 is converted to a much smaller extent. The prevailing hydrogenation of CO rather than of CO 2 has already been reported [48,52] and is supported by thermodynamics [1]. Both reactions (Equations (1) and (2) in Scheme 1) are reversible and exothermic. Although both equilibrium constants decrease with temperature, at 300 • C still quite high values can be calculated, especially for CO hydrogenation [1]. Ni/Ce0.50Zr0.50O2 in the COx co-methanation reaction. T = 300 °C; red: CO/CO2/H2 = 1/1/5; blue: CO/CO2/H2 = 1/1/7.

Discussion
The NiO-CeO2-ZrO2 mixed oxides samples synthetized by the soft-template method are characterized by high surface areas and nanometric sizes of pores and crystallites. The strong interaction between the dispersed NiO phase and the ZrO2 support makes the NiO nanoparticles reduction more difficult than for the unsupported nickel oxide. When Ce is present different reducible species are formed and reducibility generally increases. XRD results indicate the reduction of NiO nanocrystals under the experimental pretreatment conditions. However, H2-TPD experiments suggest that such reduction is not complete, though favored by increasing Ce content.
In the carbon oxides co-methanation, a competition between CO and CO2 hydrogenation reactions occurs. In the catalytic runs an almost complete CO conversion is obtained, whereas CO2 is converted to a much smaller extent. The prevailing hydrogenation of CO rather than of CO2 has already been reported [48,52] and is supported by thermodynamics [1]. Both reactions (equations 1 and 2 in Scheme 1) are reversible and exothermic. Although both equilibrium constants decrease with temperature, at 300 °C still quite high values can be calculated, especially for CO hydrogenation [1]. On the present catalysts, CO hydrogenation is probably favored also from a kinetic point of view. In the co-methanation process, both the dispersed metallic nickel and the oxide support play a On the present catalysts, CO hydrogenation is probably favored also from a kinetic point of view. In the co-methanation process, both the dispersed metallic nickel and the oxide support play a role. H 2 and CO are adsorbed and activated on metallic Ni 0 species [8]. On the other hand, adsorption and activation of CO 2 occur on the oxide support, through the intermediate formation of either CO or carbonate and hydrogencarbonate species [27]. Since CO is activated on the same sites as H 2 , its hydrogenation is likely to be favored over that of CO 2 , which requires the migration of adsorbed hydrogen from the metallic to the oxide phase (hydrogen spillover).
It has also to be taken into account that several other reactions might take place in the carbon oxides co-methanation system [1]. For the present catalytic tests, calculation of methane selectivity and carbon balance allow excluding the occurrence of reactions leading to C 2+ hydrocarbons or to carbon deposits. However, CO and CO 2 are possibly involved also in reverse methane dry reforming and water-gas shift reactions (Equations (3) and (4) in Scheme 1). Although they are reversible reactions, CO conversion into CO 2 is thermodynamically favored; in particular, the reaction of reverse methane dry reforming is characterized by an equilibrium constant higher than those of the methanation reactions [1]. The combination of these reactions, or of the corresponding reverse reactions, with CO 2 or CO hydrogenation reactions would in the end lead to the production of methane. However, their occurrence needs to be considered when comparing the almost complete CO conversion and the low values obtained for CO 2 conversion. Most likely, not only CO 2 hydrogenation is repressed by the competitive CO hydrogenation, but CO 2 is also formed from CO. This is particularly true when the feed composition in the co-methanation runs is CO/CO 2 /H 2 = 1/1/5, i.e., hydrogen concentration is lower than the stoichiometric amount. When hydrogen is fed in higher concentration (CO/CO 2 /H 2 = 1/1/7; i.e., stoichiometric amount) CO 2 conversion significantly increases. However it does not reach the equilibrium value of 93 mol% [1], indicating the occurrence of kinetic limitations, as also suggested by the significant effect of SV.
The general beneficial effect of the Ce content on the catalytic activity of the present samples can be partly related to the increased reducibility of NiO species and the consequently higher amount of hydrogen adsorbed and activated. However, the role of the ceria component itself has to be taken into account. Besides favoring CO activation on metallic Ni 0 [5,20], ceria is the main responsible for CO 2 adsorption and activation, as confirmed by the microcalorimetric results. CO 2 would preferentially adsorb on the ceria component of the support, most probably forming first carbonate and hydrogencarbonate species, which would then be hydrogenated to formates and finally to methane by the hydrogen previously adsorbed and activated by the metallic Ni 0 nanoparticles [27,36,38,42]. Therefore, the beneficial effect of the Ce content in the hydrogenation of CO 2 is most likely due also to the superior ability of CeO 2 in activating CO 2 . However, a Ce molar fraction x > 0.50 in the support does not produce any further increase in X CO2 . In the light of the above description, this could be explained by the balance of different effects. As Ce content increases, increasing amounts of both CO 2 and H 2 can be adsorbed and activated. Therefore, increasing the Ce content up to Ce/Zr = 1 results in an increase in CO 2 conversion. A possible explanation for the observed behavior at Ce molar fraction x > 0.50 could be that the greater amount of activated hydrogen would favor the reverse methane dry reforming reaction. Thus, at high Ce contents the increased capacity in activating CO 2 would be counterbalanced by the increased amount of CO 2 produced through the reverse methane dry reforming reaction, resulting in a practically unchanged overall X CO2 . Further experiments should be performed in order to confirm such hypothesis.
Moreover, optimization of the co-methanation system, with the aim of obtaining higher CH 4 yields, could be achieved by analyzing the effect of the catalyst composition at different experimental conditions and better understanding the role of the catalytic active phases in the complex interplay of the different reactions involving carbon oxides.

Synthesis of Materials
NiO-CeO 2 -ZrO 2 mixed oxides were synthesized by means of the soft-template method [53,60]
Textural analysis was carried out with an ASAP 2020 system (Micromeritics), by determining the nitrogen adsorption-desorption isotherms at −196 • C. Before analysis, the sample was heated overnight under vacuum up to 250 • C (heating rate, 1 • C min −1 ). Surface area values were calculated by the BET equation. The pore size distribution profiles were determined by applying the BJH method to the isotherm adsorption branch [55,56].
Structural characteristics of the fresh and reduced samples were investigated by X-ray diffraction (XRD) using a X3000 diffractometer (Seifert) with θ-θ Bragg-Brentano geometry with Cu-Kα wavelength and a graphite monochromator before the detector. The average crystallite sizes were estimated by the Scherrer equation [56].
The studies on the reducibility of NiO-CeO 2 -ZrO 2 mixed oxides and on the hydrogen adsorption properties of the reduced samples were performed by means of temperature-programmed reduction (TPR) and hydrogen temperature-programmed desorption (H 2 -TPD) analyses, respectively, by using a TPD/R/O 1100 apparatus (ThermoQuest, Waltham, MA, USA) equipped with a thermal conductivity detector (TCD). Prior to the TPR experiment, the sample (typically 0.020 g) was pretreated in air (15 cm 3 min −1 ) at 450 • C for 1 h; reduction profiles were then recorded under flowing H 2 (5 vol% in N 2 ; flow rate, 30 cm 3 min −1 ) while heating (20 • C min −1 ) from 40 to 950 • C. As for H 2 -TPD analyses, the sample (ca. 0.100 g) was first reduced with pure H 2 (15 cm 3 min −1 ) at 400 • C for 1 h. Then, after purge with N 2 at 400 • C for 1 h, pulses of H 2 (5 vol% in N 2 ) were admitted in the reactor at 50 • C until the area of the peaks was observed constant; finally, the desorption profile was collected under flowing N 2 (20 cm 3 min −1 ) while heating (10 • C min −1 ) from 40 to 500 • C. CO 2 adsorption properties of the reduced samples were studied through adsorption microcalorimetry, using a Tian-Calvet heat flow calorimeter (Setaram, Caluire, France) connected to a volumetric vacuum line. Prior the analysis, each sample (ca. 0.150 g, 40 mesh) was reduced with pure H 2 at 400 • C for 1 h; then it was pretreated overnight at 250 • C under vacuum (10 −3 Pa) before adsorption. CO 2 adsorption was carried out at 80 • C by admitting successive doses of the probe gas; for each dose, the equilibrium pressure, the amount of gas adsorbed, and the corresponding thermal effect were recorded. The run was stopped at the final pressure of 133.3 Pa.

CO and CO 2 Co-Methanation Catalytic Tests
Simultaneous CO and CO 2 hydrogenation to SNG was carried out in a tubular (i.d. = 0.8 cm) quartz-glass fixed-bed continuous-flow microreactor at atmospheric pressure. Prior to the reaction, the appropriate amount of catalyst powder was placed inside the reactor and pretreated in situ: It was first treated in flowing air (30 cm 3 min −1 ) at 400 • C overnight; then, after purging in He (60 cm 3 min −1 ), it was reduced under H 2 flow (15 cm 3 min −1 ) at the same temperature for 1 h, and finally cooled to the reaction temperature in flowing He (60 cm 3 min −1 ). The whole series of Ni/Ce x Zr 1−x O 2 catalysts was tested for 6 h at 300 • C, with a reactant gas mixture CO/CO 2 /H 2 with molar composition of 1/1/5 (CO, 10 mol%; CO 2 , 10 mol%; H 2 , 50 mol%; balance N 2 , used as internal standard) and space velocity (SV) 150,000 cm 3 h −1 g cat −1 .
For each run, after the first hour on stream, on-line analysis of the reactor effluent was performed every hour (after removing water through an ice trap and a 3A molecular sieves trap, where no significant CO 2 adsorption proved to occur, placed between the reactor outlet and the injection valve) with a GC 6890 (Agilent), equipped with a HP Poraplot Q capillary column and a TCD. The results of the quantitative analysis of the carbon-containing components were used for checking the carbon mass balance and for calculating CO and CO 2 conversion and products selectivity.

Conclusions
NiO-CeO 2 -ZrO 2 mixed oxides with narrow mesopores and high specific surface areas were prepared by the soft-template method. NiO was dispersed in the form of small crystallites on the CeO 2 -ZrO 2 supports and its reducibility increased with the Ce content. Upon in situ reduction, active catalysts for the carbon oxides co-methanation were obtained. Whereas CO was almost completely converted, CO 2 conversion was much lower and resulted to depend on the catalyst composition. The increase in CO 2 conversion with Ce content up to Ce/Zr = 1 could be explained with the higher ability of the CeO 2 component of the support to adsorb and activate CO 2 and to promote NiO reduction, thus also favoring H 2 adsorption and activation. However, at high Ce/Zr ratios, such effects are probably counterbalanced by the competition among the reactions involving H 2 , CO, and CO 2 .