Promising Catalytic Systems for CO 2 Hydrogenation into CH 4 : A Review of Recent Studies

: The increasing utilization of renewable sources for electricity production turns CO 2 methanation into a key process in the future energy context, as this reaction allows storing the temporary renewable electricity surplus in the natural gas network (Power-to-Gas). This kind of chemical reaction requires the use of a catalyst and thus it has gained the attention of many researchers thriving to achieve active, selective and stable materials in a remarkable number of studies. The existing papers published in literature in the past few years about CO 2 methanation tackled the catalysts composition and their related performances and mechanisms, which served as a basis for researchers to further extend their in-depth investigations in the reported systems. In summary, the focus was mainly in the enhancement of the synthesized materials that involved the active metal phase (i.e., boosting its dispersion), the di ﬀ erent types of solid supports, and the frequent addition of a second metal oxide (usually behaving as a promoter). The current manuscript aims in recapping a huge number of trials and is divided based on the support nature: SiO 2 , Al 2 O 3 , CeO 2 , ZrO 2 , MgO, hydrotalcites, carbons and zeolites, and proposes the main properties to be kept for obtaining highly e ﬃ cient carbon dioxide methanation catalysts.


Introduction
The promotion of electricity production using renewable sources is in the origin of a general interest for electric energy storage systems able to deal with their well-known intermittency [1][2][3]. Among all, the use of the temporary overproduction of renewable electricity for the synthesis of fuels (energy vectors) has attracted attention [1,[3][4][5][6]. CO 2 methanation is foreseen to play a key role in the future energy context [4,5], especially because CH 4 is the main component of natural gas, which allows its injection in the existing and well-established network (Power-to-Gas) [7].
In addition, important efforts must be carried out in order to constrain CO 2 emissions in the next years [8][9][10], being the development of suitable technologies for its capture and conversion to Besides green hydrogen production and CO2 capture, CO2 methanation efficiency depends on the used catalysts. Therefore, finding active, selective and stable materials is considered as the core of the process [17]. Taking into account the significant number of publications dealing with this topic, the present work elaborates a review and compares recent studies regarding heterogeneous supported catalysts for conventional thermal CO2 hydrogenation to CH4. The results will be presented and organized based on the support's nature. This work aims to complement excellent reviews published in the area of carbon dioxide methanation in the last few years [17][18][19][20][21][22][23][24][25][26][27][28][29]. The main findings reported will be summarized and the best performances (T: reaction temperature, XCO2: CO2 conversion, SCH4: selectivity to CH4) will be presented along with the main preparation (active metal(s) loading-wt%, given in italic in all the tables-and incorporation method) and operation (Tred: reduction temperature, H2/CO2 ratio, QT/W: flowrate per mass of catalyst) conditions applied. Additionally, and taking into account that CO2 methanation mechanism is strongly influenced by the catalyst's composition, the main findings regarding mechanistic approaches will be also included.
Despite the fact that only a few studies have dealt with the use of unsupported catalysts for CO2 methanation, their results must be taken into account. Indeed, Choe et al. [30] and Ren et al. [31] carried out theoretical mechanistic studies using Ni (111). Both proposed that CO2 methanation proceeds with CO as intermediate, suggesting a mechanism based on several elementary steps consisting in two different processes: carbon formation (from CO2 dissociation into CO and O, and the subsequent CO dissociation into C and O) and C hydrogenation to CH4, being concluded that CO dissociation was the rate-determining step of the process. Zhao et al. [32] also developed selfsupported cobalt nanoparticles composed of Cox(CoO)1-x with different ratios of x = Co/(Co + CoO) via reduction of precursor Co3O4. The Co0.2(CoO)0.8 sample exhibited the highest catalytic performances and was active even at low temperatures (160-180 °C), which was attributed to the increased amount and moderate binding of adsorbed CO2 on CoO sites. However, these results were obtained with a significantly high contact time. Besides green hydrogen production and CO 2 capture, CO 2 methanation efficiency depends on the used catalysts. Therefore, finding active, selective and stable materials is considered as the core of the process [17]. Taking into account the significant number of publications dealing with this topic, the present work elaborates a review and compares recent studies regarding heterogeneous supported catalysts for conventional thermal CO 2 hydrogenation to CH 4 . The results will be presented and organized based on the support's nature. This work aims to complement excellent reviews published in the area of carbon dioxide methanation in the last few years [17][18][19][20][21][22][23][24][25][26][27][28][29]. The main findings reported will be summarized and the best performances (T: reaction temperature, X CO2 : CO 2 conversion, S CH4 : selectivity to CH 4 ) will be presented along with the main preparation (active metal(s) loading-wt%, given in italic in all the tables-and incorporation method) and operation (T red : reduction temperature, H 2 /CO 2 ratio, Q T /W: flowrate per mass of catalyst) conditions applied. Additionally, and taking into account that CO 2 methanation mechanism is strongly influenced by the catalyst's composition, the main findings regarding mechanistic approaches will be also included.
Despite the fact that only a few studies have dealt with the use of unsupported catalysts for CO 2 methanation, their results must be taken into account. Indeed, Choe et al. [30] and Ren et al. [31] carried out theoretical mechanistic studies using Ni (111). Both proposed that CO 2 methanation proceeds with CO as intermediate, suggesting a mechanism based on several elementary steps consisting in two different processes: carbon formation (from CO 2 dissociation into CO and O, and the subsequent CO dissociation into C and O) and C hydrogenation to CH 4 , being concluded that CO dissociation was the rate-determining step of the process. Zhao et al. [32] also developed self-supported cobalt nanoparticles composed of Co x (CoO) 1-x with different ratios of x = Co/(Co + CoO) via reduction of precursor Co 3 O 4 . The Co 0.2 (CoO) 0.8 sample exhibited the highest catalytic performances and was active even at low temperatures (160-180 • C), which was attributed to the increased amount and moderate binding of adsorbed CO 2 on CoO sites. However, these results were obtained with a significantly high contact time.
Going back to the main purpose of this review, the following paragraphs will be divided based on the support nature applied in the CO 2 methanation reaction, as shown in Figure 2. Inside each Figure 2. Structure of the present review article in terms of chapters and corresponding topics.

SiO 2 -Based Catalysts
Regarding the use of supported catalysts, the firstly analyzed material is SiO 2 . The main catalytic systems found in the literature and the type of metal dispersed over the silica surface can be viewed in Table 1. with the most remarkable findings summarized below.

Mesoporous Silica-Based Catalysts
In the last years, mesoporous-based samples and especially silicas such as KIT-6, SBA-15, MCM-41 or KCC-1 have gained a great attention due to their positive feature of managing in confining nickel particles preventing sintering. The main reported findings on such type of supports are discussed below and summarized in Table 2.  [62] a Surface area determined by Brunauer-Emmett-Teller (BET) method. Values in brackets in this column correspond to pore diameters (nm); b Hydrothermal synthesis; c Direct synthesis method. •

Mesoporous silica nanoparticle (MSN)-based catalysts
Mesoporous silica nanoparticles (MSNs) present unique characteristics such as an ordered and nanosize porous structure with well-defined and tunable pore sizes (1.5-10 nm) and high textural properties (e.g., pore volume, surface area) [63]. Aziz et al. [43] studied metal-promoted mesostructured silica nanoparticles (MSN, prepared by sol-gel method) using 5 wt% of Rh, Ru, Ni, Fe, Ir, Cu, Zn, V, Cr, Mn, Al or Zr incorporated by impregnation. Authors found that Rh/MSN catalyst was the most active while Ir/MSN presented the poorest activity. Additionally, no catalytic activity was revealed by the Zn/MSN, V/MSN, Cr/MSN, Mn/MSN, Al/MSN and Zr/MSN in the studied temperature range. The same authors [63] prepared 5 wt% Ni-based catalysts supported on MSN, MCM-41, HY zeolite, SiO 2 and Al 2 O 3 . The activity was found to follow the order: Ni/MSN > Ni/MCM-41 > Ni/HY > Ni/SiO 2 > Ni/Al 2 O 3 . The better results of the MSN supported catalyst were assigned to the presence of both intra-and inter-particle porosity, which led to a higher concentration of basic sites in this sample. Researchers who worked on these materials also verified the relevance of the basicity in the catalytic performances and highlighted the fact that CO 2 methanation activity increased with the concentration of basic sites. Again, same authors [44] studied the influence of Ni loading and the presence of steam in the reactor feed flow. Regarding the metal content, the catalytic results of the prepared materials followed the order: 10%Ni/MSN ≈ 5%Ni/MSN > 3%Ni/MSN > 1%Ni/MSN. However, higher Ni loadings decreased catalysts crystallinity, surface area and basic sites. Consequently, the order of the obtained performances was due to a balance between basic sites and available metallic Ni sites. Additionally, authors emphasized that the presence of steam in the feed stream induced a negative effect on the activity (30% of activity loss after 5 h) and attributed this behavior to the formation of  [45] reported a highly active 50 wt% Ni-MSN catalyst. MSN support was prepared by sol-gel method while Ni was incorporated by impregnation using urea together with nickel nitrate, which was responsible for an enhancement in the reducibility and dispersion of Ni species and favored CO 2 adsorption capacity. They also optimized the calcination time (2, 3 or 4 h) and verified the best results when using the intermediate value (3 h). They finally included 1% of CO in the reaction flow, increasing the overall yield by its effective conversion to CH 4 .

•
KIT-6-based catalysts KIT-6 mesoporous silica presents a 3-D channel network with thick pore walls, hydrothermal stability, high surface area and large pore volume [64,65]. Its interconnected 3-D structure promotes the location of active species inside the porous system, favoring metallic dispersion and diffusional processes and hindering pores blockage [64,65]. Zhou et al. [46,47] prepared mesoporous Co/KIT-6 and Co/meso-SiO 2 catalysts with the same Co content (wt% not specified) via hydrogen reduction. Co was better dispersed on KIT-6 than on SiO 2 , explaining the higher activity revealed by Co/KIT-6 sample. In addition, the highly ordered and mesoporous structure of the Co/KIT-6 catalyst improved the selectivity to CH 4 . Liu et al. [48] studied the effect of Co content on the performances of Co/KIT-6 catalysts, concluding that the best Co loading was 25 wt%. However, for the 15 to 20 wt% Co loading based samples there was a limitation in the number of active sites. In addition, 30 wt% Co catalyst presented larger and less dispersed Co particles with lower specific surface area and a disordered pore structure that hindered the diffusion and migration of CO 2 and H 2 molecules towards Co active sites. The same authors [49] investigated the influence of the reduction temperature in Co/KIT-6 catalysts activity towards CO 2 methanation. They obtained the best results by reducing the sample at 400 • C, due to the larger specific surface area formed, the higher number of Co species reduced and the increased CO 2 adsorption and activation capacity. Additionally, Merkache et al. [50] prepared Fe-KIT-6 materials with Si/Fe molar ratios of 10, 30 and 60. The best performances were reported for the sample with higher Fe content (n Si /n Fe = 10) and were attributed to the higher concentration of metallic active sites on the KIT-6 structure. Cao et al. [51] reported recently Ni-V/KIT-6 catalysts, with metal loadings varying from 5 to 40 wt% Ni and 0.1 to 2 wt% V. They obtained the best results when using 20 wt% Ni and 0.5 wt% V, achieving a high metal dispersion. The surface basicity of the materials was further boosted by a synergistic effect between Ni and V (present as V 2 O 5 in the catalysts).
• SBA-15-based catalysts SBA-15 presents dual porosity composed by mesoporous and intrawall micropores or secondary mesoporous channels. It presents large surface areas, thick pore walls (3-6 nm), uniformly distributed cylindrical channels (5-10 nm) and high hydrothermal stability [66]. Liu et al. [52] prepared Ni-based SBA-15 catalysts by one-pot hydrothermal method with a 3-D network structure and 15 wt% of Ni. For comparison purposes, they also prepared an impregnated sample containing 15 wt% Ni over SBA-15 support. Ni/SBA-15 One-pot catalyst presented higher surface area, larger pore volume, better dispersion, higher catalytic activity and anti-sintering properties for Ni 0 particles. In addition, Lu et al. [53] prepared Ni/SBA-15 catalysts by grafting method. A chemical bond was formed between silicon and Ni atoms via an oxygen atom (-O-Ni-O-Si-O-), by using Ni ammonia complex ions with NH 3 /Ni molar ratios of 2 to 4 (equivalent to 4 to 10 wt% Ni). No bulk nickel oxides were depicted in the Ni-grafted SBA-15 sample and the CO 2 conversion and methane selectivity were higher than those obtained for equivalent Ni/SBA-15 catalysts prepared by impregnation. Authors also indicated that the increase of the Ni content enhanced the activity and selectivity. Bacariza et al. [54] prepared Ni and Ce/Ni impregnated SBA-15 catalysts with 15 wt% Ni and 15 wt% Ce and with SBA-15 prepared by a conventional and a microwaves-assisted hydrothermal treatment. Despite the synthesis method of the support, relevant performances were detected, with the microwave-method sample presenting slightly higher conversions, especially between 250 and 350 • C. Additionally, an enhancement of the performances was observed by adding Ce, being this attributed to the role of CeO 2 species as active sites for CO 2 activation. Finally, Li et al. [55] investigated bimetallic Ni-Pd/SBA-15 alloy catalyst for selective hydrogenation of CO 2 to methane by changing the atomic ratio of Ni/Pd (Ni 0. 25  The excellent catalytic activity of Ni 0.75 Pd 0.25 /SBA-15 was related to the synergistic effect between Ni and Pd species. In this way, the charge transfer from Pd to Ni atoms provided a negative charged surface, which promoted the activation of CO 2 and the dissociation of intermediates into CH 4 , crucial for improving the catalytic performance. • MCM-41-based catalysts MCM-41 presents a purely mesoporous 2-D hexagonal structure formed by unidirectional and non-interconnecting cylindrical pores (2-10 nm of diameter) [67,68]. Du et al. [56] reported 1 to 3 wt% Ni containing MCM-41 samples with Ni ions incorporated into the MCM-41 structure. Well-dispersed Ni 0 species were formed in the samples, especially after reduction at high temperature (700 • C), leading to superior performances. Additionally, Bacariza et al. [54] prepared Ni and Ce/Ni-based MCM-41 catalysts with 15 wt% Ni and 15 wt% Ce. These samples presented greater activity than equivalent SBA-15 catalytic systems. MCM-41 was found as capable of promoting the metal-support interactions, favoring the CO 2 and H 2 activation. Additionally, the lower affinity of MCM-41 to water served as another advantage in encountering the water produced as a product in the reaction (according to the literature [44,69,70]). Finally, and as mentioned previously for other type of mesoporous silica supports, the incorporation of Ce to the Ni/MCM-41 sample triggered a boost in its performance during the methanation reaction. To better emphasize on this point, Wang et al. [57] also prepared Ni/MCM-41 promoted with CeO 2 , containing different CeO 2 loadings (0, 10, 20, 30 wt%). The addition of CeO 2 improved the dispersion of Ni on the support, providing a larger specific metallic surface area and a higher reducibility. Besides, the interaction between Ni active sites, CeO 2 and the support facilitated the activation of adsorbed CO 2 species at low temperatures. Thus, within a certain composition range, the incorporation of CeO 2 was able to improve the catalytic performances, with 20 wt% CeO 2 leading to the best catalytic activity and stability. Recently, Taherian et al. [58] reported the beneficial effects of incorporating Y and Mg to Ni-based MCM-41 catalysts synthesized by direct synthesis method. The incorporation of these two metals to the MCM-41 mesoporous system decreased the average Ni 0 size by improving Ni species reducibility and creating new active sites for CO 2 activation.
• KCC-1-based catalysts KCC-1 is formed by a fibrous surface morphology arranged in a 3-D structure forming nanospheres. Its fibrous structure facilitates the accessibility to the available surface area and improves the hydrothermal and mechanical stability [71]. Hamid et al. [59] prepared a KCC-1 material by microemulsion and used it as catalyst for CO 2 methanation in absence of active metals sites. They compared the synthesized sample to MCM-41 and SiO 2 materials. As a result, KCC-1, which presented higher basicity and oxygen vacancies than MCM-41 and SiO 2 , led to the best performances. It was proposed that both CO 2 and H 2 were activated and dissociated in the KCC-1 support oxygen vacancies, where the CO 2 interaction with the material is related to the basicity. The same authors [60] reported promising results when loading KCC-1 mesoporous support with 5 wt% of Ni, Co or Zn, with superior behavior found for Ni/KCC-1. Recently, Lv et al. [61] evaluated the effects of impregnating increasing Ni loadings (5 to 25 wt%) over a fibrous KCC-1 nanospheres synthesized by microemulsion hydrothermal method. 20 wt% Ni sample exhibited better results than those found for equivalent Ni/SiO 2 and Ni/MCM-41 catalysts. The formation of dendrimer mesoporous channels not only improved the metallic dispersion due to the confinement effect but also hindered the occurrence of sintering processes. Furthermore, the appearance of intermediate carbonate species was faster when using Ni/KCC-1 instead of Ni/SiO 2 or Ni/MCM-41. • FDU-12-based catalysts FDU-12 is a 3-D ordered mesoporous silica with a face-centered cubic structure. It presents a window of~9 nm and pore cages with sizes ranging from 12 to 42 nm. Its structure presents advantages when comparing to SBA-15 or MCM-41, such as higher resistance towards pores blockage due to its cage-like network [62]. Recently, Liu and Dong [62] reported, for the first time, Ni-based FDU-12 catalysts for CO 2 methanation. Authors prepared a series of 10 wt% Ni-containing FDU-12 and evaluated the influence of the impregnation solvent (water or ethylene glycol) and the effects of doping the catalysts with trehalose and Ce (3 wt%). A remarkable reduction of the average Ni 0 particle size was observed when using ethylene glycol (5.9 nm) instead of water (10 nm). These values were further improved (from 5.9 to 2.6-3.3 nm) by incorporating increasing amounts of trehalose in the impregnation solution, while Ce did not influence this parameter. The enhancement of the metallic dispersion was responsible for an improvement of the catalytic performances, being a CH 4 yield >75% obtained at 425 • C for the 10Ni/FDU-12 catalyst prepared using ethylene glycol and a fraction of trehalose. Complementarily, the already reported positive effect of Ce in the reaction was proved, being the activity boosted at lower reaction temperatures (<350 • C). Finally, authors also verified a reduction of the activation energy with the optimization of the impregnation solvent and Ce incorporation (from 116.23 to 77.49 kJ mol −1 ).

Comparative analysis
Taking into account the various experimental conditions used in the works reported in the literature, a proper comparison among the different types of mesoporous silica supports is difficult to establish. In this way, in an attempt to compare monometallic catalysts prepared by impregnation method, CH 4 formation rates as a function of the mass of catalyst (mol CH 4 s −1 g cat −1 ) and metal (mol CH 4 s −1 g metal −1 ) were determined, far from thermodynamic equilibrium, for a series of Ni and Co-supported SiO 2 [34] (reference), MSN [45], KIT-6 [46], SBA-15 [54], MCM-41 [54], KCC-1 [61] and FDU-12 [62] (Figure 3). As seen, the use of Ni instead of Co leads generally to better performances, which insists on the suitability of Ni as CO 2 methanation active metal. Additionally, SBA-15, MCM-41 and FDU-12 are the materials leading to the most outstanding results, both per g catalyst and g metal . The catalytic performances obtained for the different mesoporous silicas will be likely influenced by several factors such as the metallic dispersion, the activation induced by the support (resulting from the metallic-support interactions), the location of the metal, the ability to activate carbon dioxide (adsorption sites, pores acting as CO 2 reservoir) or even the hydrophobicity. Even with well-designed, systematic and specific studies reported in the literature, no definitive conclusions regarding the superior performances revealed by SBA-15, MCM-41 or FDU-12 catalysts can be drawn (e.g., it is not possible to quantify certainly the contribution of each individual factor to the activity).
several factors such as the metallic dispersion, the activation induced by the support (resulting from the metallic-support interactions), the location of the metal, the ability to activate carbon dioxide (adsorption sites, pores acting as CO2 reservoir) or even the hydrophobicity. Even with welldesigned, systematic and specific studies reported in the literature, no definitive conclusions regarding the superior performances revealed by SBA-15, MCM-41 or FDU-12 catalysts can be drawn (e.g., it is not possible to quantify certainly the contribution of each individual factor to the activity).

Mechanistic Aspects
Wu et al. [33] reported that Ni loading affects the selectivity to methane and the mechanism pathways over Ni/SiO2 catalysts (Figure 4-Left). Indeed, they prepared two 0.5 and 10 wt% Ni/SiO2 samples and proposed that formate species in a monodentate configuration are involved in CO2 hydrogenation on both cases. On one hand, the consecutive pathway, which was favored on small Ni particles (0.5%Ni/SiO2 catalyst, Ni 0 average size < 1 nm), was attributed to low H2 coverage on the Ni surface, leading to dissociation of formate intermediates resulting in CO formation and high CO selectivity. On the other hand, the reaction on large Ni particles (10%Ni/SiO2 catalyst, Ni 0 average size of 9 nm, stronger H2 adsorption and enhanced H2 coverage), was proposed to be controlled by mixed consecutive and parallel routes, promoting formate species hydrogenation to CO or CH4 as part of a parallel reaction pathway. The sites corresponding to kink, corner or step positions on the Ni/SiO2 surface were proposed as the primary active sites for CO2 hydrogenation. In addition, Park et al. [36,37] proposed a bifunctional mechanism for Pd-MgO/SiO2 catalysts (Figure 4-Right). They considered CO2 adsorption over MgO as a surface carbonate that is sequentially hydrogenated to form methane. Consequently, in this mechanism an active metal (Pd in this case) is required for the dissociation of H2 and the supply of H atoms to the CO2 derived species activated on Mg sites.

Mechanistic Aspects
Wu et al. [33] reported that Ni loading affects the selectivity to methane and the mechanism pathways over Ni/SiO 2 catalysts (Figure 4-Left). Indeed, they prepared two 0.5 and 10 wt% Ni/SiO 2 samples and proposed that formate species in a monodentate configuration are involved in CO 2 hydrogenation on both cases. On one hand, the consecutive pathway, which was favored on small Ni particles (0.5%Ni/SiO 2 catalyst, Ni 0 average size < 1 nm), was attributed to low H 2 coverage on the Ni surface, leading to dissociation of formate intermediates resulting in CO formation and high CO selectivity. On the other hand, the reaction on large Ni particles (10%Ni/SiO 2 catalyst, Ni 0 average size of 9 nm, stronger H 2 adsorption and enhanced H 2 coverage), was proposed to be controlled by mixed consecutive and parallel routes, promoting formate species hydrogenation to CO or CH 4 as part of a parallel reaction pathway. The sites corresponding to kink, corner or step positions on the Ni/SiO 2 surface were proposed as the primary active sites for CO 2 hydrogenation. In addition, Park et al. [36,37] proposed a bifunctional mechanism for Pd-MgO/SiO 2 catalysts (Figure 4-Right). They considered CO 2 adsorption over MgO as a surface carbonate that is sequentially hydrogenated to form methane. Consequently, in this mechanism an active metal (Pd in this case) is required for the dissociation of H 2 and the supply of H atoms to the CO 2 derived species activated on Mg sites.
Regarding mesoporous silica-based catalysts, Aziz et al. [43] considered CO as a reaction intermediate in CO 2 methanation over Ni/MSN catalysts. Indeed, they proposed the dissociation of CO 2 into CO and O on the Ni surface and the progressive hydrogenation of CO into methane. Due et al. [56] also suggested that CO 2 methanation passes through CO as intermediate over Ni/MCM-41 catalysts. Finally, Hamid et al. [59] suggested a mechanism for CO 2 methanation over KCC-1 ( Figure 5), where CO 2 was activated on oxygen vacancies forming adsorbed carbonyl species. They demonstrated that linear carbonyl species were the precursors for CH 4 formation. Additionally, they implied that hydrogen was activated and dissociated in E' centres (defined as dangling bonds, ≡Si*, formed on the SiO x material for x < 2) acting as active sites and giving rise to atomic H species.  [59] suggested a mechanism for CO2 methanation over KCC-1 ( Figure  5), where CO2 was activated on oxygen vacancies forming adsorbed carbonyl species. They demonstrated that linear carbonyl species were the precursors for CH4 formation. Additionally, they implied that hydrogen was activated and dissociated in E' centres (defined as dangling bonds, ≡Si*, formed on the SiOx material for x < 2) acting as active sites and giving rise to atomic H species.

Al2O3-Based Catalysts
Alumina-based catalysts are, without doubt, the most commonly used in carbon dioxide methanation reaction. In summary, the main reported catalytic systems are disclosed below in Table  3, while the corresponding conclusions of the published works will be deeply discussed in the upcoming paragraphs.  Regarding mesoporous silica-based catalysts, Aziz et al. [43] considered CO as a reaction intermediate in CO2 methanation over Ni/MSN catalysts. Indeed, they proposed the dissociation of CO2 into CO and O on the Ni surface and the progressive hydrogenation of CO into methane. Due et al. [56] also suggested that CO2 methanation passes through CO as intermediate over Ni/MCM-41 catalysts. Finally, Hamid et al. [59] suggested a mechanism for CO2 methanation over KCC-1 ( Figure  5), where CO2 was activated on oxygen vacancies forming adsorbed carbonyl species. They demonstrated that linear carbonyl species were the precursors for CH4 formation. Additionally, they implied that hydrogen was activated and dissociated in E' centres (defined as dangling bonds, ≡Si*, formed on the SiOx material for x < 2) acting as active sites and giving rise to atomic H species.

Al2O3-Based Catalysts
Alumina-based catalysts are, without doubt, the most commonly used in carbon dioxide methanation reaction. In summary, the main reported catalytic systems are disclosed below in Table  3, while the corresponding conclusions of the published works will be deeply discussed in the upcoming paragraphs.

Al 2 O 3 -Based Catalysts
Alumina-based catalysts are, without doubt, the most commonly used in carbon dioxide methanation reaction. In summary, the main reported catalytic systems are disclosed below in Table 3, while the corresponding conclusions of the published works will be deeply discussed in the upcoming paragraphs.

Monometallic Al 2 O 3 -Supported Catalysts
Garbarino et al. [73] studied Ni/Al 2 O 3 catalysts prepared by impregnation and with 16, 39 and 125 wt% of Ni by wt% of support. Authors performed catalytic tests without a pre-reduction treatment and obtained very low activities, confirming that Ni 0 species are indeed required for obtaining significant CO 2 conversion degrees. They demonstrated also that the higher the Ni content, the larger the particles but also the greater the number of active sites. Consequently, these two parameters affecting in opposite ways the performances contributed to the following activity results order: 125%Ni/Al 2 O 3 > 16%Ni/Al 2 O 3 > 39%Ni/Al 2 O 3 . The same authors [72] synthesized Ni nanoparticles (NPs) with <8 nm of size and compared their performances with the ones obtained for 125%Ni/Al 2 O 3 catalyst from their previous work. Better outcome was obtained for 125%Ni/Al 2 O 3 (with Ni particles of~36 nm), while Ni NPs resulted in poorly active metallic sites. They came up with a hypothesis that the support could play a role in the activation of CO 2 while Ni nanoparticles could be mainly responsible for the H 2 dissociation, which explains the better results obtained for the supported Ni catalyst. Lately, they also reported a study comparing two 3 wt% Ru and 20 wt% Ni based Al 2 O 3 catalysts [75] revealing that, despite the higher metal content of the Ni sample, Ru/Al 2 O 3 led to higher performances due to the favored dispersion of Ru species. Quindimil et al. [74] investigated the influence of the metal nature (Ni, Ru) and loading (4 to 20 wt% for Ni and 1 to 5 wt% for Ru) on CO 2 methanation catalysts supported on Al 2 O 3 . Authors clearly observed that, in both cases, higher metal loading enhanced the catalytic performances, which could be due to the increase of active metal sites available for H 2 dissociation. However, for Ni/Al 2 O 3 series , metallic dispersion decreased with metal loading due to the formation of larger particles, while the dispersion of Ru/Al 2 O 3 was not significantly influenced. Moreover, it was suggested that the reducibility of the catalysts plays a key factor in the catalytic performances. Indeed, the complete reduction of ruthenium species was achieved at 250 • C, whereas for nickel-based catalysts a temperature of 900 • C was required, reducing the available Ni 0 sites for H 2 dissociation. Besides, at low temperature (T < 300 • C), turnover frequency (TOF) values for Ru/Al 2 O 3 were considerably higher than those found for Ni/Al 2 O 3 , which indicates that Ru is more effective than Ni in H 2 dissociation. Taking this into account and considering the saturation effect on CO 2 conversion with metal loading, the best performances were obtained by 12%Ni/Al 2 O 3 and 4%Ru/Al 2 O 3 formulations, with Ru catalyst leading to higher conversion and CH 4 selectivity. As a final remark, no works studying the influence of chloride and/or chlorine doping over Ru/Al 2 O 3 catalysts were found in the literature so far. This strategy could be interesting for Sabatier reaction, as promising results were reported for CO methanation [99][100][101].

Bimetallic Al 2 O 3 -Supported Catalysts
Hwang et al. [76] studied 20%Ni-5%Fe-Al 2 O 3 catalysts prepared by co-precipitation using different precipitation agents ((NH 4 ) 2 CO 3 , Na 2 CO 3 , NH 4 OH or NaOH). They verified that Ni 0 dispersion was enhanced by using ammonium carbonate, favoring both the CO 2 conversion and the CH 4 selectivity. Serrer et al. [77] also studied the effect of Fe in Ni-Fe-γ-Al 2 O 3 catalysts synthesized via homogeneous precipitation with urea, a method optimized in previous publications from the same authors [102,103]. During catalyst activation, authors observed a synergistic effect between nickel (13 wt%) and iron (4 wt%) that led to higher fractions of reduced nickel compared to the monometallic 17 wt% Ni/γ-Al 2 O 3 reference. Ni remained in its reduced metal state under CO 2 methanation conditions. On the contrary, the oxidation of Fe 0 was immediately observed. This indicated that Fe exhibits a highly dynamic behavior through a Fe 0 Fe 2+ Fe 3+ redox cycle, likely located at the interface between the FeO x clusters and the surface of the metal particles, which promoted CO 2 dissociation during the reaction. Burger et al. [79] reported also highly active and thermally stable Mnand Fe-promoted NiAlO x catalysts prepared by co-precipitation and tested under CO 2 methanation conditions using a pressure of 8 bar. They concluded that while Mn improves the CO 2 adsorption capacity (especially in terms of medium basic sites) without significantly modifying the available nickel surface area, Fe promotes thermal stability through the formation of a Ni-Fe alloy and slightly increases the CO 2 uptake. These authors [104] also reported an improvement of NiFeAlO x catalysts deactivation resistance at elevated temperature and pressure in thermodynamic equilibrium. This phenomenon was attributed to the partial segregation of (γFe, Ni) nanoparticles formed during the activation step giving rise to Fe 2+ , offering an alternative reaction pathway through CO 2 activation on disordered Fe y O. They also studied the effects of incorporating Fe by surface redox reaction (SRR) method on Ni-based alumina catalysts [78]. They prepared a 11 wt% Ni/Al 2 O 3 catalyst by deposition-precipitation and a 48 wt% NiAlO x material by co-precipitation method. In the case of Ni/Al 2 O 3 -derived catalysts, higher Fe loadings (0.5 to 1.8 wt%) induced better performances (tests done at 8 bar), which was attributed to the enhancement of the electronic properties due to the formation of Ni-Fe alloys. In addition, increasing Fe loadings (3.6 to 8.6 wt%) increased the apparent thermal stability of NiAlO x -derived catalysts, which was ascribed to the partial segregation of the alloyed Ni-Fe particles, in accordance with their previous studies.
Xu et al. [80] investigated a series of Co-Ni doped ordered mesoporous Al 2 O 3 (OMA) oxides with different Co/(Co+Ni) molar ratios fabricated by the one-pot evaporation-induced self-assembly (EISA) method. This synthesis method produced homogeneous incorporation of metals among the Al 2 O 3 framework achieving as well a high Co and Ni dispersion. Bimetallic samples exhibited higher catalytic performances than those obtained for the monometallic Ni and Co OMA catalysts, especially at lower temperatures. This trend was related to the synergetic effect between Co and Ni ( Figure 6), since Co presents high activity towards CH 4 formation at low temperature, owing to remarkable CO 2 activation capacity and allowing the decrease of the apparent activation energy. In addition, they observed that only an appropriate Co/(Co + Ni) molar percentage (20%) could maximally enhance the catalytic performances. Besides, catalysts showed an outstanding thermal stability and anti-sintering properties. Alrafei et al. [81] also optimized Ni (5 to 25 wt%) and Ni-Co catalysts supported on γ-Al 2 O 3 . Authors prepared the alumina support in the form of extrudates and incorporated Ni and Co by impregnation (co-impregnation for bimetallic samples). They observed, in accordance with other studies already discussed, that Ni loading strongly influenced the catalytic activity, with values above 15 wt% leading to no remarkable differences. In terms of Co effect, they observed that this metal presented only a positive impact in the performances when Ni loading was 10 wt%. They attributed this to an enhancement of nickel species reducibility and dispersion.
observed that only an appropriate Co/(Co + Ni) molar percentage (20%) could maximally enhance the catalytic performances. Besides, catalysts showed an outstanding thermal stability and antisintering properties. Alrafei et al. [81] also optimized Ni (5 to 25 wt%) and Ni-Co catalysts supported on γ-Al2O3. Authors prepared the alumina support in the form of extrudates and incorporated Ni and Co by impregnation (co-impregnation for bimetallic samples). They observed, in accordance with other studies already discussed, that Ni loading strongly influenced the catalytic activity, with values above 15 wt% leading to no remarkable differences. In terms of Co effect, they observed that this metal presented only a positive impact in the performances when Ni loading was 10 wt%. They attributed this to an enhancement of nickel species reducibility and dispersion. Several studies have dealt with the beneficial incorporation of lanthanides and metal oxides to CO2 methanation catalysts due to their appropriate electronic or acidic/basic properties, among all [105,106]. Indeed, Rahmani et al. [82] reported Ni-based catalysts supported on mesoporous γ-Al2O3 promoted with Ce, Mn, Zr and La oxides and prepared by impregnation, all samples containing 20 wt% Ni and 2 wt% of promoter. As authors identified Ce as the promoter with the most effective results, they studied the effect of Ce loading (2, 4 and 6 wt%) being observed that, despite the small differences between samples, 2 wt% Ce seems to be the most suitable content. Cerium positive effect was explained, as seen in SiO2 based materials section, by the presence of oxygen vacancies in the cerium containing samples, enhancing CO2 activation. In the Mn, Zr and La samples, the presence of the promoters had only an effect in the coverage of Ni sites, reducing the accessibility and causing blockage, so that no enhancement of the performances was found. Liu et al. [83] also reported the effect of adding CeO2 to 15%Ni/Al2O3 catalysts. The catalytic performances were strongly dependent on the CeO2 content, being the sample with 2 wt% Ce the one leading to the best results, as also found by Rahmani et al. [82]. The positive effect of Ce was attributed to the improved reducibility of Ni species (e.g., weakening of Ni-Al2O3 interactions) and enhanced activation of CO2 on Ce sites. Guo et al. [84] developed an innovative catalyst by synthesizing Ni-Al2O3 LDHs (layered double support) and using CeO2 as promoter. Authors obtained the best performances when using ~50 wt% CeO2 due to the small and highly dispersed Ni 0 particles formed and due to the existence of oxygen vacancies that directly increased the active sites and promoted CO2 adsorption. Tada et al. [85] verified also the Several studies have dealt with the beneficial incorporation of lanthanides and metal oxides to CO 2 methanation catalysts due to their appropriate electronic or acidic/basic properties, among all [105,106]. Indeed, Rahmani et al. [82] reported Ni-based catalysts supported on mesoporous γ-Al 2 O 3 promoted with Ce, Mn, Zr and La oxides and prepared by impregnation, all samples containing 20 wt% Ni and 2 wt% of promoter. As authors identified Ce as the promoter with the most effective results, they studied the effect of Ce loading (2, 4 and 6 wt%) being observed that, despite the small differences between samples, 2 wt% Ce seems to be the most suitable content. Cerium positive effect was explained, as seen in SiO 2 based materials section, by the presence of oxygen vacancies in the cerium containing samples, enhancing CO 2 activation. In the Mn, Zr and La samples, the presence of the promoters had only an effect in the coverage of Ni sites, reducing the accessibility and causing blockage, so that no enhancement of the performances was found. Liu et al. [83] also reported the effect of adding CeO 2 to 15%Ni/Al 2 O 3 catalysts. The catalytic performances were strongly dependent on the CeO 2 content, being the sample with 2 wt% Ce the one leading to the best results, as also found by Rahmani et al. [82]. The positive effect of Ce was attributed to the improved reducibility of Ni species (e.g., weakening of Ni-Al 2 O 3 interactions) and enhanced activation of CO 2 on Ce sites. Guo et al. [84] developed an innovative catalyst by synthesizing Ni-Al 2 O 3 LDHs (layered double support) and using CeO 2 as promoter. Authors obtained the best performances when using~50 wt% CeO 2 due to the small and highly dispersed Ni 0 particles formed and due to the existence of oxygen vacancies that directly increased the active sites and promoted CO 2 adsorption. Tada et al. [85] verified also the effect of CeO 2 loading on the performances of Ru/CeO 2 /Al 2 O 3 catalysts. They observed that the performances followed the trend: Ru/30%CeO 2 /Al 2 O 3 > Ru/60%CeO 2 /Al 2 O 3 > Ru/CeO 2 > Ru/Al 2 O 3 . With a lower CeO 2 content (30 wt%), a larger surface area was achieved due to the smallest average size of CeO 2 crystallites in this sample when comparing to the rest of the tested catalysts. Additionally, authors suggested, according to the results of performed mechanistic studies, that samples without CeO 2 presented further difficulties in the dissociation and decomposition of intermediate formate species. Along with the previously discussed studies, Ahmad et al. [86] tested the effect of La, Ce, Pr, Eu and Gd as promoters of Ni/γ-Al 2 O 3 catalysts with 5 wt% of the promoter and 12 wt% of Ni added by impregnation. In this case, Pr was pointed out as the best promoter, even if all lanthanides led to an increase of Ni dispersion without remarkable effects on the textural properties when comparing to the 12%Ni/γ-Al 2 O 3 catalyst. Finally, Karam et al. [87] reported Ni-Mg-Al 2 O 3 catalysts prepared by EISA one-pot and optimized the materials in terms of Mg and Ni loadings. The best performances were achieved with 7 wt% Mg and 15 wt% Ni, being higher than those revealed for other Al 2 O 3 samples in the literature. This outcome was due to the important improvement of the metallic dispersion and surface basicity.

Trimetallic Al 2 O 3 -Supported Catalysts and Composites
Wang-Xin et al. [88] prepared highly dispersed Ni-Ce-Zr/γ-Al 2 O 3 catalysts by citric acid assisted impregnation. The addition of citric acid in the preparation procedure promoted the dispersion of the Ni-Ce-Zr oxide species over the γ-Al 2 O 3 surface and improved metal-support interactions. Consequently, Ni-Ce-Zr/γ-Al 2 O 3 catalysts were highly active and the higher the citric acid content used in the synthesis method the better the results. Toemen et al. [89] studied Ru/Mn/Ce/Al 2 O 3 catalysts prepared by sequential impregnation with the aim of optimizing the Ce loading as well as the calcination temperature. Firstly, calcination at 1000 • C was found as the optimum temperature when comparing with the results obtained for the samples calcined at 400, 600 and 800 • C due to the modification of the phases formed during this thermal treatment. Additionally, Ce loadings of 55 to 85 wt% were tested, being the best results obtained for the 65 wt% Ce containing sample, which was attributed to higher dispersion and the favorable morphology of this catalyst.
Zamani et al. [107] carried out a deep study regarding the use of M*/Mn/Cu-Al 2 O 3 (M* = Pd, Rh and Ru) catalysts prepared by impregnation. After comparing three catalysts containing the same Pd, Rh and Ru content over Mn/Cu-Al 2 O 3 , the best material was the one containing Ru. Consequently, as a second step, the Ru loading (1.6 to 18.4 wt%) and the calcination temperature (831 to 1168 • C) were optimized by using Design-Expert ® software, where 10.9 wt% Ru and a calcination temperature of 1035 • C were described as the best conditions. Bakar et al. [90] studied Ru and Pd supported Mn/NiO/Al 2 O 3 catalysts and reported Ru as a more favorable noble metal for this reaction. They also optimized the Ru/Mn/Ni ratio and the calcination temperature with better performances obtained when the parameters were 5:35:60 and 1000 • C, respectively. Authors defined this optimum catalyst as spherical nanoparticles with aggregated and agglomerated mixtures of metal species on the surface. It must be highlighted that the authors did not report any kind of pre-reduction treatment. However, they considered that unreduced NiO species could be active sites for CO 2 methanation. To further confirm their hypothesis, they synthesized a NiO based Al 2 O 3 catalyst and reported~15 % of CO 2 conversion without carrying out a pre-reduction treatment. They asserted the potential role of NiO as an active site for the reaction since, even after the tests, no reduced nickel species were observed in the samples. Franken et al. [108] prepared CoAl 2-x Mn x O 4 (x = 0, 0.1, 0.5, 1) solid solutions by co-precipitation with the aim of understanding the effects of Mn on catalysts properties and performances. CoMn 0.5 Al 1.5 O 4 led to the best performances while the introduction of Mn into the spinel support was responsible for a reduction in the basicity and an improvement of the hydrogen spillover from Co active sites to the support surface. These modifications led to a reaction mechanism with formate species as intermediates (no CO) accompanied by a remarkable decrease of the activation energy.
Do et al. [91] studied the effect of Ca incorporation (0, 1.0, 3.0, 5.0, 7.0 and 10.0 %mol) to perovskite NiTiO 3 (containing~11 wt% Ni) loaded by impregnation on γ-Al 2 O 3 . Ca incorporation favored the nickel dispersion, boosted the CO 2 adsorption capacity and improved the textural properties, which resulted not only in a remarkable increase of the performances but also in a significant stability under reaction conditions for 10 days.
Several authors also reported the suitability of Al 2 O 3 -based composites and mixed oxides as supports for CO 2 methanation catalysts. Indeed, Cai et al. [92] studied 12%Ni/(ZrO 2 -Al 2 O 3 ) catalysts checking not only the promoter effect of ZrO 2 but also the effects of the ZrO 2 -Al 2 O 3 composite preparation method (co-precipitation, impregnation of Zr over commercial γ-Al 2 O 3 or drying and impregnation followed by precipitation). In all cases, Ni was incorporated by impregnation. Regarding the preparation method of ZrO 2 -Al 2 O 3 , impregnation followed by precipitation led to the best catalytic performances with high stability due to the enhanced Ni species reducibility and dispersion over this support. Additionally, this sample also presented the highest basicity, which was considered by the authors as a key parameter, also favoring the methanation performances. Regarding the effect of the ZrO 2 loading (0, 3, 9 and 15 wt%), 3 wt% ZrO 2 enhanced both the activity and the stability. Lin et al. [93] also studied mesoporous Al 2 O 3 -ZrO 2 modified Ni catalysts prepared via a single-step epoxide-driven sol-gel method and varying the Al/Zr ratio and Ni loading. They observed that the incorporation of ZrO 2 into the Ni/Al 2 O 3 structure decreased the interaction between Ni species and Al 2 O 3 , promoting nickel reducibility and dispersion. Thus, increasing ZrO 2 loading favored the formation of active metallic Ni sites and surface oxygen vacancies, which represent two key factors for improving the catalytic activity at low temperatures. However, increasing both Zr and Ni loading was found as beneficial only for a certain range, with the catalyst containing 20% wt% Ni and an Al/Zr ratio of 1.0 presenting the best results.
Furthermore, Yang et al. [95] prepared CaO-Al 2 O 3 composites by co-precipitation and incorporating Ni by impregnation. The best reported sample contained 15 wt% of Ni and a CaO:Al 2 O 3 ratio of 20:80. Authors concluded that CaO was responsible for restraining the growth of NiO nanoparticles, improving its dispersion and creating a moderate interaction between NiO and Al 2 O 3 . Liu et al. [96] also reported ordered mesoporous Ni-Ru-doped CaO-Al 2 O 3 composites synthesized by one-pot evaporation-induced self-assembly method with 10, 1 and 2 wt% of NiO, RuO 2 and CaO, respectively. This material was compared to reference samples such as Ni-Al 2 O 3 , Ni-Ru-Al 2 O 3 and Ru-Al 2 O 3 . Monometallic Ru or Ni catalysts presented considerably similar CO 2 conversions due to the close average size values of the metal species in both catalysts (~10 nm). However, the selectivity to methane was quite poor in the Ni sample. The better results obtained for the Ni-Ru bimetallic sample were attributed by the authors to the synergistic effect between Ni and Ru and to the favored H 2 activation developed by Ru species. Finally, the addition of CaO enhanced the CO 2 adsorption capacity providing Ni-Ru-Ca sample a finer performance.
Xu et al. [94] studied Ru/TiO 2 -Al 2 O 3 catalysts prepared by impregnation. When comparing the performances of two 5 wt% Ru catalysts supported on Al 2 O 3 and TiO 2 -Al 2 O 3 , the second catalytic system exhibited a much higher activity. This was due to the more dispersed particles found in the Ti-containing sample as a result of establishing a strong interaction between Ru and TiO 2 . The effect of TiO 2 content (5, 10 and 15 wt%) and the calcination temperature after TiO 2 -Al 2 O 3 synthesis were also evaluated. While TiO 2 loading did not influence the results, CO 2 conversions increased with higher calcination temperatures (from 600 to 950 • C) and decreased for 1100 • C, being 950 • C the optimum.
Abate et al. [97] prepared γ-Al 2 O 3 -ZrO 2 -TiO 2 -CeO 2 composites as supports for Ni catalysts. The best results were obtained for the sample presenting the composite composition of γ-Al 2 O 3 :ZrO 2 :TiO 2 :CeO 2 = 55:15:15:15, prepared by impregnation-precipitation method using commercial γ-Al 2 O 3 powder as a host, and 20 wt% Ni. The incorporation of Ni to the composites instead of γ-Al 2 O 3 provoked enhanced Ni 0 dispersion and reducibility, due to the modification of the metal-support interactions. Mebrahtu et al. [98] also studied Ni catalysts supported on ternary and quaternary alumina-zirconia-titania-ceria mixed oxides. Results showed that, for ternary oxide supported Ni, the catalytic activity depends on textural properties improvement while, for quaternary oxide supported Ni, it depends mainly on reducibility and metallic dispersion. Moreover, TiO 2 and ZrO 2 incorporation was found to modify mainly textural properties while CeO 2 addition promoted Ni dispersion and reducibility. Besides, 15 wt% of CeO 2 strengthened the catalyst stability, decreasing the deactivation rate of about one order of magnitude for the best ternary system (20%Ni/90%γ-Al 2 O 3 -5%ZrO 2 -5%TiO 2 ).
Finally, some authors focused on finding new methods to prepare Ni/Al 2 O 3 catalysts. This is the case of Daroughegi et al. [109], who used ultrasound-assisted co-precipitation; Xu et al. [110], who used cold plasma; Song et al. [111], who used microwaves; Schubert et al. [112], who used double flame spray pyrolysis; Aljishi et al. [113] and Xu et al. [114], who used evaporation induced self-assembly technique (EISA) or Shang et al. [115], who used partial hydrolysis of aqueous solutions containing Al(NO 3 ) 3 and Ni(NO 3 ) 2 with (NH 4 ) 2 CO 3 , without templates or organic surfactants. Additionally, Le et al. [116] studied the effect of the type of Al 2 O 3 crystalline phase in the performances of Ni/Al 2 O 3 catalysts.

Mechanistic Aspects
Beuls et al. [117] proposed some insights for Rh/γ-Al 2 O 3 catalysts mechanism at 50-150 • C and 2 atm. They suggested that CO 2 is adsorbed over the material by dissociation, forming CO (ads) and O (ads) and inducing Rh species oxidation. Indeed, Rh species oxidation state played a crucial role in the distribution of the adsorbed species on the catalysts, with Rh-dicarbonyls being identified as those leading to the formation of CH 4 . In short, authors consider CO as the true reaction intermediate in their catalytic system. Garbarino et al. [73] suggested that CO 2 methanation over Ni/Al 2 O 3 catalysts occurs at the expense of CO intermediate on the corners of Ni nanoparticles interacting with alumina, likely due to the formation of formate or other oxygenate species and their further hydrogenation into CH 4 . Zhang et al. [118] studied the mechanistic implications of different Ni loadings on Ni/ -Al 2 O 3 catalysts by diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) studies. They observed that, while low Ni contents favor the occurrence of reverse water gas shift (RWGS) reaction promoting CO formation, increasing loadings hinder the formation of stable and non-reactive carbonate species and promote formates conversion to CH 4 , boosting the activity towards CO 2 methanation rather than RWGS. Finally, Cárdenas-Arenas et al. [119] proposed that CO 2 methanation proceeds over Ni/Al 2 O 3 through the adsorption and consecutive chemisorption of CO 2 over hydroxyl groups created due to the H 2 reduction of the NiO-Al 2 O 3 interface (Figure 7). They reported that chemisorbed CO 2 hampered carbon dioxide molecules chemisorption/dissociation and led, alternatively, to the formation of water and formates species. Even if formate species could be successively hydrogenated to methane, part of them gave rise to CO, decreasing CH 4 selectivity.

Mechanistic Aspects
Beuls et al. [117] proposed some insights for Rh/γ-Al2O3 catalysts mechanism at 50-150 °C and 2 atm. They suggested that CO2 is adsorbed over the material by dissociation, forming CO(ads) and O(ads) and inducing Rh species oxidation. Indeed, Rh species oxidation state played a crucial role in the distribution of the adsorbed species on the catalysts, with Rh-dicarbonyls being identified as those leading to the formation of CH4. In short, authors consider CO as the true reaction intermediate in their catalytic system. Garbarino et al. [73] suggested that CO2 methanation over Ni/Al2O3 catalysts occurs at the expense of CO intermediate on the corners of Ni nanoparticles interacting with alumina, likely due to the formation of formate or other oxygenate species and their further hydrogenation into CH4. Zhang et al. [118] studied the mechanistic implications of different Ni loadings on Ni/ɣ-Al2O3 catalysts by diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) studies. They observed that, while low Ni contents favor the occurrence of reverse water gas shift (RWGS) reaction promoting CO formation, increasing loadings hinder the formation of stable and non-reactive carbonate species and promote formates conversion to CH4, boosting the activity towards CO2 methanation rather than RWGS. Finally, Cárdenas-Arenas et al. [119] proposed that CO2 methanation proceeds over Ni/Al2O3 through the adsorption and consecutive chemisorption of CO2 over hydroxyl groups created due to the H2 reduction of the NiO-Al2O3 interface (Figure 7). They reported that chemisorbed CO2 hampered carbon dioxide molecules chemisorption/dissociation and led, alternatively, to the formation of water and formates species. Even if formate species could be successively hydrogenated to methane, part of them gave rise to CO, decreasing CH4 selectivity.

CeO 2 -Based Catalysts
Even if cerium oxide has been typically used as a promoter for CO 2 methanation catalysts, several authors also reported interesting results when using it as support. Indeed, the most notable results can be seen in Table 4. and their main findings are discussed in the following sections.

Monometallic CeO 2 -Supported Catalysts
Tada et al. [85,120] studied 10 wt% Ni-based catalysts supported on CeO 2 , γ-Al 2 O 3 , TiO 2 and MgO and identified CeO 2 as the best choice. The surface coverage by CO 2 -derived species on CeO 2 surface as well as the partial reduction of CeO 2 surface was reported to be responsible for the great performances of the 10%Ni/CeO 2 catalyst. Atzori et al. [121] also studied NiO-CeO 2 mixed oxides highly active for CO 2 methanation and prepared by soft-template and impregnation methods. Soft-template produced Ni 0 particles of~7 nm while impregnation contributed to the formation of larger particles (>30 nm) after the pre-reduction treatment. Despite the synthesis method and the differences in the Ni 0 size, both samples presented high catalytic performances. Consequently, authors suggested that H 2 activation on Ni 0 is not the critical step for a high catalytic activity at such conditions, pointing out on the important role of ceria support in the reaction pathway. The proposed explanation was that only the highly uncoordinated Ni atoms at the metal-support interface, whose number strongly depends on the average particle size, remain available to hydrogenate CO 2 activated on the nearby ceria sites. Ratchahat et al. [122] also prepared structured Ni/CeO 2 catalysts with different configurations: plain, segmented, stacked and multi-stacked. When compared to the plain configuration, the stacked and segmented ones provided an enhancement of convective heat and mass transfer due to the random flow-channels and the addition of a gap distance in the case of the stacked, which contributed to the improvement of the methanation performances at low temperature and increased the contact time. Moreover, the multi-stacked catalyst was tested under industrial-like conditions (high feed rate and pure feed gas component), being the moderate hot spot observed responsible for a re-boosting of the conversion. Furthermore, this catalyst presented high stability (<0.6% CO 2 conversion loss after 76 h under reaction conditions), being its shape and surface morphology maintained. Cárdenas-Arenas et al. [123] also reported Ni/CeO 2 catalysts prepared by different Ni-incorporation method in order to modify the Ni-Ce interaction. Authors synthesized three dimensionally ordered macroporous (3DOM) structures by co-impregnation, successive impregnation of Ni and Ce precursors or Ni impregnation after CeO 2 3DOM synthesis. Besides, materials with uncontrolled structures were also prepared by co-precipitation and Ni impregnation. Authors identified two types of active sites: NiO-CeO 2 interface and reduced Ni 0 particles efficient for CO 2 and H 2 dissociation, respectively. The optimal sites proportion required to obtain the maximum conversion was 25% Ni 0 for H 2 dissociation and 75% NiO-CeO 2 for CO 2 dissociation, achieved by the Ni-uncontrolled structured catalyst prepared by impregnation. Moreover, pulse experiments with isotopic CO 2 confirmed that this catalyst was more effective in hydrogenating the surface carbon species.
Sharma et al. [124] studied Ni, Co, Pd and Ru doped CeO 2 catalysts prepared by combustion method. Ce x Ru 1-x O 2 presented the best catalytic performances while Ni, Co and Pd induced low CH 4 selectivity, with Ni and Co catalyzing both methanation and RWGS reaction and Pd producing only CO. Furthermore, Vita et al. [125] studied Ni/GDC (gadolinium-doped ceria) catalysts both in powder and monolith forms. For the powder systems, the activity increased with Ni loading, which was explained by the more favourable metal-support interaction, the increased number of moderate-basic sites and the presence of surface oxygen vacancies. Powdered catalysts presented also lower CH 4 productivity than a high loaded monolith (0.5 g cm −3 ), due to the high surface-to-volume ratio, good interphase mass transfer and low pressure drops of the monolithic system.
Ocampo et al. [126,127] prepared Ce x Zr 1-x O 2 (CZ) mixed oxides as supports for 5-15 wt% Ni-based catalysts by sol gel method. CZ fluorite structure was maintained after Ni incorporation but the replacement of some Zr 4+ by Ni 2+ cations led to a reduction in the lattice parameter of the structure, especially in the 10 and 15 wt% Ni samples. Even if all samples presented great performances and stability, 10% Ni-CZ catalyst reported the best results due to the high oxygen storage capacity of CZ and its ability to effectively disperse Ni particles. Additionally, the incorporation of nickel cations into the CZ structure and the higher dispersion of NiO at its surface were found responsible for improving the redox properties of the materials, able to hinder sintering processes. A similar strategy was followed by Ashok et al. [128], who prepared a series of nickel catalysts supported on Ce x Zr 1-x O 2 (CZ) by ammonia evaporation, impregnation and deposition-precipitation methods. Among all, the Ni/CZ catalyst prepared via ammonium evaporation method led to the best performances and stability. These positive results were explained by the fact that this catalyst was activated at low temperatures, and the incorporation of some Ni cations in the ceria structure led to the creation of oxygen vacancies and to better reducibility of Ni species in this sample.

Bi-and Trimetallic CeO 2 -Supported Catalysts
Le et al. [129] studied Ni-Na/SiO 2 and Ni-Na/CeO 2 catalysts prepared by impregnation in order to compare the effect of the support nature and evaluate Na loading effects. In both cases, increasing Na contents led to lower performances which is ascribed to a reduction in the textural properties and in the CO 2 adsorption capacity while higher performances were obtained by CeO 2 -supported materials. In addition, Sun et al. [131] developed a novel ICCU (integrated carbon capture and utilization) process for carbon capture and methane production, using dual function materials composed by a mechanical mixture of MgO and Ru/CeO 2 catalysts with increasing Ru loadings. This ICCU process allows the simultaneous regeneration of the sorbent and the conversion of the captured CO 2 to CH 4 in a single reactor at intermediate temperature (300 • C). Authors found that MgO+10Ru/CeO 2 showed the best performance in the 1st cycle of ICCU due to a higher number of oxygen vacancies. However, after 10 cycles, MgO+5Ru/CeO 2 exhibited the highest catalytic activity because of the more favorable metal-support interactions established in this sample. Finally, Pastor-Pérez et al. [130] prepared Ni-Co/CeO 2 -ZrO 2 and Ni-Mn/CeO 2 -ZrO 2 catalysts (~15 wt% Ni and 3.5 wt% Co/Mn) and evaluated the impact of methane incorporation in the feed. Authors verified a negative effect of Mn on the activity, which attributed to a stronger than desired CO 2 adsorption over these sites, limiting the subsequent hydrogenation of the reactant towards methane. On the contrary, Co favored the performances through the improvement of the catalyst's electronic properties and the inhibition of Ni sintering. They also verified that incorporating methane in the feed led to a slight enhancement of the catalytic performances, which they ascribed to low temperature reforming reactions taking place simultaneously with methanation.

Mechanistic Aspects
In terms of mechanisms, Konishcheva et al. [132] proposed that CO 2 methanation proceeds over Ni/CeO 2 catalysts through the activation of CO 2 on CeO 2 surface and the stepwise hydrogenation to methane through hydrocarbonate and formate intermediates by the hydrogen spilled over from Ni. Authors considered the route where CO acts as a reaction intermediate negligible.
Cárdenas-Arenas et al. [119] also reported the presence of two independent active sites for CO 2 methanation over Ni/CeO 2 ( Figure 8): NiO-CeO 2 interface for CO 2 activation and Ni 0 particles for H 2 dissociation. Even if they identified water desorption as the slowest step of the mechanism, the co-existence of two types of actives sites and the high oxygen mobility of Ni/CeO 2 catalysts hindered water inhibitory effect.

Mechanistic Aspects
In terms of mechanisms, Konishcheva et al. [132] proposed that CO2 methanation proceeds over Ni/CeO2 catalysts through the activation of CO2 on CeO2 surface and the stepwise hydrogenation to methane through hydrocarbonate and formate intermediates by the hydrogen spilled over from Ni. Authors considered the route where CO acts as a reaction intermediate negligible. Cárdenas-Arenas et al. [119] also reported the presence of two independent active sites for CO2 methanation over Ni/CeO2 (Figure 8): NiO-CeO2 interface for CO2 activation and Ni 0 particles for H2 dissociation. Even if they identified water desorption as the slowest step of the mechanism, the co-existence of two types of actives sites and the high oxygen mobility of Ni/CeO2 catalysts hindered water inhibitory effect. Aldana et al. [126] studied Ni-CexZr1-xO2 catalysts and proposed a bifunctional mechanism without CO as intermediate species (Figure 9a). Actually, they observed that H2 molecules were activated and dissociated into H on Ni 0 sites while CO2 was adsorbed on the Ce-Zr support oxygen vacancies forming monodentate carbonates. These monodentate carbonates were then sequentially hydrogenated producing formates, formaldehydes, methoxy species and, finally, methane. Aldana et al. [126] studied Ni-Ce x Zr 1-x O 2 catalysts and proposed a bifunctional mechanism without CO as intermediate species (Figure 9a). Actually, they observed that H 2 molecules were activated and dissociated into H on Ni 0 sites while CO 2 was adsorbed on the Ce-Zr support oxygen vacancies forming monodentate carbonates. These monodentate carbonates were then sequentially hydrogenated producing formates, formaldehydes, methoxy species and, finally, methane. Additionally, authors proposed a pathway for CO formation and verified its inhibitory role in the reaction (Figure 9b). Aldana et al. [126] studied Ni-CexZr1-xO2 catalysts and proposed a bifunctional mechanism without CO as intermediate species (Figure 9a). Actually, they observed that H2 molecules were activated and dissociated into H on Ni 0 sites while CO2 was adsorbed on the Ce-Zr support oxygen vacancies forming monodentate carbonates. These monodentate carbonates were then sequentially hydrogenated producing formates, formaldehydes, methoxy species and, finally, methane. Additionally, authors proposed a pathway for CO formation and verified its inhibitory role in the reaction (Figure 9b).

ZrO 2 -Based Catalysts
In terms of ZrO 2 supported catalysts, the main results found in the literature can be seen in Table 5 while the main conclusions of the published works will be discussed below.

Monometallic ZrO 2 -Supported Catalysts
Da Silva et al. [133] studied Ni/ZrO 2 catalysts and compared them to Ni/SiO 2 samples, both prepared by impregnation and containing 10 wt% Ni. The sample supported on zirconia was more active, which authors explained by the considerably weaker interaction of CO 2 with the SiO 2 sample and the higher Ni dispersion found in the ZrO 2 catalyst. In addition, Lu et al. [134] prepared mesoporous zirconia-modified clays as supports for Ni catalysts by one-pot method (hydrothermal treatment of the mixture of the clay suspension and the ZrO(NO 3 ) 2 solution). The bimodal pore structure of the zirconia-modified clays was reported as beneficial for the dispersion of Ni. As concluded by previous authors, ZrO 2 favored the dispersion of the Ni species, preventing sintering during the reaction and reducing carbon deposition. The catalyst with 20 wt% of zirconia led to the best performances while the most stable sample was the 15%Ni/ZrO 2 . Jia et al. [135] studied also 9 wt% Ni-supported ZrO 2 catalysts and reported the beneficial effect of carrying out the decomposition of the precursor salts after impregnation (calcination) using dielectric barrier discharge (DBD) plasma instead of a conventional thermal treatment at 500 • C (from 30 to 70% of CO 2 conversion at 300 • C when performing thermal and DBD plasma calcination, respectively). The positive effect of plasma was related to the remarkable enhancement of the metallic dispersion when using this non-conventional treatment. Additionally, Zhao et al. [136] prepared 10, 15 and 25 wt% Ni/ZrO 2 catalysts by combustion method using several combustion mediums such as urea, glycerol, glycol, ethanol and n-propanol. The highest performances were reported for the sample synthesized using urea, being the results attributed to the favored reducibility of the Ni species, the better dispersion and the greater CO 2 adsorption capacity.
Furthermore, Li et al. [137] reported Co-based catalysts supported on ZrO 2 , Al 2 O 3 , SiO 2 , SiC, TiO 2 and activated carbon (AC), all containing 10 wt% of Co. Their catalytic tests were performed at 30 atm of pressure, contrary to what was done in the majority of works reported for CO 2 methanation. ZrO 2 was the best performing Co-based support, with the order as follows: 10%Ni/ZrO 2 > 10%Ni/SiO 2 > 10%Ni/Al 2 O 3 ≈ 10%Ni/SiC > 10%Ni/AC > 10%Ni/TiO 2 . The zirconia support contributed to the better reducibility and dispersion of Ni species. Besides, the presence of oxygen vacancies on the ZrO 2 structure prevented the formation of carbon deposits during the tests. The same authors [138] studied later the effect of the impregnation solvent for preparing 2 wt% Co catalysts supported over ZrO 2 . Among all solvents, citric acid-assisted impregnation led to the best catalytic results. The beneficial effect of citric acid arises from the enhancement of the Co particles dispersion and its interaction with ZrO 2 due to the formation of cobalt citrate complex as precursor. This intensified interaction also provided oxygen vacancies and enhanced nickel reducibility, improving CO 2 adsorption capacity and, consequently, the obtained performances.
Recently, Nagase et al. [139] reported~3 wt% Ru/ZrO 2 catalysts (using amorphous and crystalline ZrO 2 ) prepared by selective deposition method and compared them with an equivalent Ru/SiO 2 sample. Ruthenium particles with average sizes below 10 nm were obtained in all cases with amorphous ZrO 2 promoting the activity and selectivity. The effects of the preparation method for Ru/amorphous ZrO 2 was studied, by comparing the use of a NaOH or NH 3 in the selective deposition method with a conventionally impregnated sample. Impregnated and NH 3 samples led to the similar results, but lower than those exhibited by the catalyst prepared using NaOH. The higher activity of the later was related not only to the promoted dispersion, also found when using NH 3 , but especially to the presence of residual Na species in the catalyst able to activate the intermediate formate species during the reaction.

Bimetallic ZrO 2 -Supported Catalysts
Takano et al. [140] prepared 50 wt% Ni/Sm-ZrO 2 by mechanical mixture followed by calcination. The best results were obtained for the Zr/Sm ratio of 5 and after calcination at 650 or 800 • C. Sm 3+ ions were substituted in the ZrO 2 lattice, what promoted the presence of oxygen vacancies, favored the interaction of CO 2 weakening the C-O bonds and, consequently, enhanced the CO 2 activation.
The same authors [141] prepared Ni supported Yttrium-Zirconia catalysts by co-impregnation and with several Y 3+ concentrations. Ni/Y-ZrO 2 catalysts showed higher catalytic activity than the Ni/ZrO 2 sample, being the performances dependent upon the Y 3+ concentration, with the best results obtained for the catalyst with a Y/(Zr+Y) molar ratio of 0.333. Kesavan et al. [142] also studied Ni-based Y-stabilized ZrO 2 (YSZ) catalysts with 10 wt% Ni prepared by different methods: wetness impregnation, wetness impregnation with EDTA, electroless plating, mechanical mixing with commercial NiO nano-powder and mechanical mixing with nano-powder transformed in micro-powder NiO. The most efficient and stable catalyst was the one prepared by wet impregnation with Ni(EDTA) 2− complex due to the smaller and highly dispersed Ni 0 particles formed. Additionally, Kosaka et al. [143] studied Ni-based tubular catalysts again supported over Y-stabilized ZrO 2 with different Ni loadings. Even under high gas hourly space velocity (GHSV) conditions, catalysts with higher Ni content showed better catalytic activity, which was also favored by their tubular structure that provided a higher CH 4 yield when compared to the same powder catalysts. Authors studied also the effect of Ni loading on the temperature profiles and realized that the high CO 2 methanation performances arise from the interaction between heat generation and acceleration of the reaction rate in the catalysts. Through a numerical analysis based on reaction kinetics, heat transfer and fluid dynamics, authors deduced that both high CH 4 yield and prevention of hot spots formation can be achieved by properly arranging catalysts with different activities in the reactor.
Ren et al. [144] studied Ni/ZrO 2 catalysts impregnated with a second metal (Fe, Co and Cu). Fe was the metal leading to the best performances, which resulted from the improved dispersion and reducibility of Ni species as well as to the partial reduction of ZrO 2 in the Ni-Fe/ZrO 2 sample. Lu et al. [145] also studied the effects of adding Fe, Co, Ce and La to Ni based ZrO 2 modified clays by impregnation. They observed that by adding 1 wt% of promoter to the 15%Ni/ZrO 2 catalysts, Ni species dispersion and reducibility and catalysts' thermal stability were enhanced. Even if the performances were quite similar for all the promoted samples, Fe led to the best results attributed to synergistic effects between Fe, ZrO 2 and Ni species. Dumrongbunditkul et al. [146] prepared Co-Cu-ZrO 2 catalysts by co-precipitation method and tested them at 30 atm. They also prepared monometallic samples with better results obtained for the Co-ZrO 2 than for the Cu-ZrO 2 samples. Additionally, the catalyst with Co:Cu:Zr = 20:40:40 molar ratio led to the highest performances due to its high surface area and the more homogeneous dispersion of the mixed metal oxides.
Tan et al. [147] prepared a series of MgO-modified Ni/ZrO 2 catalysts with 6 wt% Ni and Mg/Ni molar ratios ranging from 1/6 to 1/1 using citrate-complexing impregnation method. Among all the synthesized materials, the catalyst with Mg/Ni molar ratio of 1/4 showed the best catalytic performances. In accordance with Li et al. [138], the use of citric acid was considered as responsible for a reduction in the Ni particles size. Furthermore, the addition of MgO stabilizes the highly dispersed Ni nanoparticles and has a confinement effect, limiting the growth of Ni and improving the resistance towards sintering and carbon deposition. Takano et al. [148] also studied tetragonal ZrO 2 doped with Ca 2+ and Ni 2+ ions. The incorporation of Ca 2+ and Ni 2+ ions to the zirconia led to the formation of oxygen vacancies in the ZrO 2 lattice and favored the performances.
In order to compare the different promoters reported for Ni-based ZrO 2 catalysts prepared by impregnation and calcined under thermal conditions, CH 4 production rates were determined for a Ni/ZrO 2 catalyst [135] and for bimetallic materials containing Y [141], Fe [144], Co [144], Cu [144], Mg [147] and Ca [148] (Figure 10). As seen, the main effect of the promoters is verified at lower reaction temperatures while Mg, Co or Fe are suitable candidates for developing further optimization studies.
ZrO2 doped with Ca and Ni ions. The incorporation of Ca and Ni ions to the zirconia led to the formation of oxygen vacancies in the ZrO2 lattice and favored the performances.
In order to compare the different promoters reported for Ni-based ZrO2 catalysts prepared by impregnation and calcined under thermal conditions, CH4 production rates were determined for a Ni/ZrO2 catalyst [135] and for bimetallic materials containing Y [141], Fe [144], Co [144], Cu [144], Mg [147] and Ca [148] (Figure 10). As seen, the main effect of the promoters is verified at lower reaction temperatures while Mg, Co or Fe are suitable candidates for developing further optimization studies. Figure 10. CH4 production rates per gram of catalyst determined, far from thermodynamic equilibrium, for a series of Ni-based ZrO2 materials with without promoters all prepared by impregnation method as a function of the reaction temperatures. Figure 10. CH 4 production rates per gram of catalyst determined, far from thermodynamic equilibrium, for a series of Ni-based ZrO 2 materials with without promoters all prepared by impregnation method as a function of the reaction temperatures.

Mechanistic Aspects
Takano et al. [140,141,148] proposed a mechanism for Ni/ZrO 2 catalysts where CO 2 is adsorbed as carbonate species on zirconia oxygen vacancy sites rather than CO while hydrogen is activated on Ni particles. Then, H atoms must be supplied from Ni to ZrO 2 sites to reduce the adsorbed carbonate species progressively into formates, formaldehydes, methoxy species and, finally, methane.

MgO and Hydrotalcite-Based Catalysts
Besides all the catalysts already presented in this work and containing Mg as a promoter, in this chapter catalysts supported on MgO and hydrotalcite(HT)-derived materials will be summarized, with their main results shown in Table 6

Monometallic MgO and Hydrotalcite-Derived Catalysts
Loder et al. [149] prepared a bifunctional Ni/MgO catalyst for CO 2 methanation. They tested the influence of Ni loading (11,17,20 and 27 wt% of Ni) and the matrix composition (MgO pure and MgO/CaO). Results confirmed that MgO is an active compound and the CaO did not participate actively in the reaction, since the thermal stability of CaCO 3 is much higher than MgCO 3 . In this way, Ni/MgO catalyst showed better results than those obtained for Ni/MgO-CaO. The catalytic activity increased linearly with the Ni loading since a remarkable number of active sites were found even at high Ni loadings (27 wt%). Based on the reaction kinetics, authors also modelled a Langmuir-Hinshelwood based rate law taking the bifunctional catalytic action of the catalyst into account.
Wierzbicki et al. [150] studied the effect of the Ni loading (10 to 43 wt%) on Ni based hydrotalcite-derived materials with constant M(II)/M(III) (valence II to valence III metals) molar ratio, prepared by pH controlled co-precipitation and the decomposition of the precursor hydrotalcites carried out by thermal treatment. After this treatment, periclase-like structured (homogeneous mixed oxide) materials were formed being Ni present as NiO. Increasing Ni contents weakened the interaction between Ni and the hydrotalcite matrix, improving its reducibility and increasing the number of active sites. This resulted in an enhancement of the performances. Additionally, Bette et al. [151] prepared ã 59 wt% Ni catalyst supported on a (Mg,Al)O x mixed oxide derived from a (Ni,Mg,Al)-hydrotalcite-like (HT) precursor. The precursor HT structure was synthesized by pH-controlled co-precipitation and the reduction at 900 • C led to metallic Ni particles supported on a (Mg,Al)O x matrix. Authors reported that the hydrotalcite-like derived material provided anti-sintering and stability properties to the catalysts, what justified the good performances revealed by the supported materials. Finally, Abate et al. [152] prepared Ni-Al-hydrotalcite materials and compared them to a Ni/Al 2 O 3 commercial sample with 80 wt% Ni. Ni content was fixed with an Al/(Al + Ni) = 0.25-0.27 corresponding to a theoretical loading of 75-80 wt% NiO. Better results were obtained for the HT derived sample due to the favored catalysts reducibility and the higher metal surface area and metal dispersion.

Bi-and Trimetallic MgO and Hydrotalcite-Derived Catalysts
Yan et al. [153] prepared Ni-Mg mixed oxides promoted with Co, Mo, Mn, Fe, Cu and W. Authors also prepared MgO, NiO, Ni a MgO x (a = 0.5, 0.8, 1.0, 1.2) and Ni/MgO Impregnated samples. NiO and MgO catalysts revealed quite poor activity, Ni-Mg mixed oxides led to considerably higher levels of CO 2 conversion and CH 4 selectivity and the Ni/MgO Impregnated sample presented intermediate results between the NiO and the mixed oxides. Being Ni 1.0 MgO x the best mixed oxide, samples containing transition metals were synthesized using this optimized composition. The performances varied as follows: W > Co > No promoter > Mo > Mn > Fe > Cu. Then, as the W-promoted sample presented the best results, authors optimized the W/Ni molar ratio with the best performances obtained for the 1/1 value. They considered that the remarkable results reported by the NiWMgO x catalyst were related to the great anti-CO-poison ability, strong resistance against coke formation and sintering and the higher amount of CO 2 adsorption sites due to the contribution of W species. Varun et al. [154] studied NiO-MgO and M/NiO-MgO (M = 2 wt% Co, Cu or Fe) catalysts and evaluated the composites preparation method and the promoter nature. They reported three NiO-MgO materials synthesized by solution combustion, sonochemical and co-precipitation methods and obtained the best results when using the second due to the highly dispersed particles formed. Regarding the promoter nature (Co, Cu or Fe; incorporated by impregnation), Co led to the best performances, explained by the lowest activation energy determined for this catalyst due to Co reducing nature.
Ho et al. [155] reported hydrotalcite-derived~60 wt% Ni catalysts promoted with 9.5 wt% La, 6.3 wt% Y or 9.5 wt% Ce. The best catalytic performances were exhibited by the Ni-La sample (Ni-La > Ni-Ce ≈ Ni-Y > Ni) and were attributed to La 2 O 3 species ability to simultaneously promote basicity and Ni 0 dispersion. Authors concluded that an equilibrium between Ni 0 particle size and basicity is fundamental for obtaining the best CO 2 methanation performances. Furthermore, Wierzbicki et al. [156] studied the promoter effect of La on the performances of Ni-HT derived materials, keeping the M(III)/(M(III)+M(II)) molar ratio equal to 0.25. Authors verified that the addition of lanthanum resulted in the formation of a separate phase, not integrated in the hydrotalcite-derived structure. Indeed, the incorporation of 2 wt% of lanthanum favored the catalytic performances due to the enhanced basicity, reducibility and dispersion of Ni species. Lately, the same authors [157] studied different preparation methods for La introduction (e.g., co-precipitation, impregnation and ion exchange using La-EDTA) in order to investigate its influence on the catalytic performances. La affected catalysts CO 2 adsorption capacity in terms of medium strength basic sites, which were strongly influenced by the used preparation method. Among all, ion exchange was the most efficient method to increase the amount of medium strength basic sites, leading to the best catalytic performances. Zhang et al. [158] also studied La-promoted Ni/Mg-Al catalysts with different La contents (0 to 8 wt%) and prepared by two different methods: urea hydrolysis and co-precipitation. La-promoted catalysts showed an enhanced catalytic activity, with the one containing 5 wt% of La presenting the best Ni dispersion. Besides, La increased significantly the amount of moderate basic sites that enhanced CO 2 adsorption capacity. Furthermore, urea hydrolysis was a more efficient preparation method than co-precipitation, as it promoted Ni dispersion and CO 2 adsorption capacity, leading to better catalytic performances.
Wang et al. [159] studied a Ni-Al-hydrotalcite-derived catalyst modified with Fe or Mg and prepared by co-precipitation. Authors prepared samples with different (Fe or Mg)/Al molar ratios (Fe/Al = 0.05 and 0.25, Mg/Al = 0.1 and 1). The performances of Fe catalysts were the same for both ratios while, in Mg samples, the lower the Mg content, the better the results. They suggested that both Fe and Mg favored Ni dispersion while Mg enhanced also the basicity, helping in CO 2 adsorption and activation, and Fe improved the reducibility and changed the pore distribution. Mebrathu et al. [162] also studied Ni-Fe bimetallic catalysts derived from hydrotalcites and prepared by co-precipitation, using variable Fe/Ni ratios (0.1 to 1.5, corresponding to 12 wt% Ni and 1.2 to 18 wt% Fe). Bimetallic catalysts showed higher activities and stabilities than monometallic ones, with the sample presenting the lowest Fe content (1.2 wt% Fe, Fe/Ni = 0.1) being the most outstanding. The incorporation of Fe led to weaker Ni-support interactions, promoting nickel reducibility. Additionally, an improvement of the metallic dispersion (with lower Fe loadings) and basicity was associated with Fe incorporation. In accordance with other studies already discussed in this work, the basicity of the support and the metal particle size were found as key catalysts properties. Later, the same authors [160] studied the deactivation mechanism for the best catalyst reported in the previous work (Fe/Ni = 0.1) and for the monometallic Ni sample. They observed the formation of nickel hydroxide by the reaction of Ni species with the water from methanation during the long-term tests (30-40 h). The formation of these species decreased the number of available Ni 0 sizes for the reaction, favored sintering processes and facilitated the formation of hardly reducible nickel aluminates. Authors concluded that water partial pressure was the main factor influencing the deactivation rate of the Ni catalyst, while the higher resistance of Ni-Fe material was due to Fe ability to limit the formation of nickel hydroxide.
Finally, He et al. [161] prepared Ni-Al 2 O 3 -HT catalysts as well as reference samples with Ni incorporated by impregnation. Ni-Al 2 O 3 -HT catalyst presented highly dispersed Ni particles along with strong basic sites, what led to considerably better results than the impregnated samples. Additionally, one sample containing K (K-Ni-Al 2 O 3 -HT) was synthesized contributing to results slightly higher than those of the un-promoted catalyst, due to the extra strong basic sites provided by K. Therefore, authors proposed that the combination of small metallic Ni particles with strong basic sites over Al 2 O 3 could effectively promote the conversion of CO 2 into CH 4 . This study contrasts with other works where medium strength basic sites are considered as the most favorable for CO 2 methanation reaction, as the strongest ones lead to inactive carbonate phases.

Carbon-Based Catalysts
Few studies have focused on the use of carbon-based catalysts for this reaction. In summary, their main results (Table 7) and findings can be found in the following sections.

Monometallic Carbon-Supported Catalysts
Jiménez et al. [163] prepared Ru impregnated carbon nanofibers (CNFs) with different natures (orientation of graphite planes): platelet, fishbone and ribbon. The catalytic performances were not affected by the nature of the carbon nanofibers used as support. In addition, Li et al. [164] studied Co-based porous carbon (PC) catalysts with controlled crystal morphology and size by using ZIF-67 metal organic framework (MOF) as template and with the aid of surfactants. After carbonization, the obtained samples inherited the original morphology and size of ZIF-67 crystals with a distorted surface. Catalytic tests were carried out at 30 atm and Co nanoparticles inside the carbon matrix ranged between 7 and 20 nm were separated by the graphite-like carbon effectively avoiding metal sintering. Consequently, interesting performances were reported for these samples, especially when compared with similar catalysts supported on commercial activated carbon. Finally, Gödde et al. [165] prepared Ni nanoparticles supported on nitrogen functionalized carbon nanotubes (CNTs) with different Ni loadings (10 to 50 wt%). The optimum loading of 30-40 wt% led to high surface area and small Ni particle sizes, ascribed to the efficient anchoring on the N-doped CNTs. Besides, high stability of the catalysts was demonstrated for 100 h time-on-stream.

Bi-and Trimetallic Carbon-Supported Catalysts
Romero-Sáez et al. [166] reported Ni-Zr-supported CNTs catalysts by both sequential and co-impregnation methods (metal loadings: 5 wt% Ni and 20 wt% Zr). Authors obtained better results after sequential impregnations, since co-impregnation method resulted in the formation of a core-shell structure where NiO particles were surrounded by ZrO 2 , limiting the access of the reactants to Ni active sites during the catalytic tests. On the contrary, sequential impregnation favored Ni reducibility and CO 2 adsorption capacity. Wang et al. [167] prepared Ni-based catalysts supported on multi-walled CNTs and promoted with Ce by ultrasonic-assisted co-impregnation (12 wt% of Ni and Ce contents of 0, 1.5, 3, 4.5 and 6 wt%). Additionally, authors prepared samples supported on γ-Al 2 O 3 for comparison purposes. They verified, in accordance with other studies previously discussed in this work, that Ce enhanced the Ni 0 dispersion, promoted the reduction of metal oxides and favored the CO 2 activation. Meanwhile, the confinement effect of CNTs and the promotion effect of cerium could efficiently prevent active species migration and sintering, as well as restrict carbon deposition. The best results were reported for the 12%Ni-4.5%Ce/CNTs. Le et al. [168] prepared Ni based activated carbons (AC) and verified the effect of adding Ce-Zr oxides as promoters. 7%Ni/AC reference sample was prepared by impregnation while the promoted 7%Ni/Ce 0.2 Zr 0.8 O 2 /AC sample was prepared by depositing the Ce 0.2 Zr 0.8 O 2 mixed oxide prepared by hydrothermal method over the activated carbon and later impregnating 7 wt% Ni over the Ce 0.2 Zr 0.8 O 2 /AC support. Authors verified that the ceria-zirconia solid solution phase could effectively disperse and stabilize nickel species as well as favor activated carbon CO 2 adsorption capacity. Consequently, 7%Ni/Ce 0.2 Zr 0.8 O 2 /AC catalyst exhibited higher activity than 7%Ni/AC sample. Finally, Gaidai et al. [169] synthesized carbon fiber (CF) supported Fe-Co catalysts by uniform infiltration of a metal nitrate solution through the carrier (CF) surface, followed by drying. Authors, who performed several treatments in the carbon fibers, found out that oxygenation increased the amount of oxygen in the surface layer, decreasing active metals reducibility. Carbon fibers pre-reduction increased the amount of oxygen containing groups on the surface through pulling them from the bulk, which was detrimental for the catalytic performances.

Mechanistic Aspects
Lapidus et al. [170] studied CO 2 methanation mechanism over copper and nickel-based carbon catalysts (Ni/C and Cu/C). They only found CO and water as products on Cu/C, suggesting that CO 2 was so strongly adsorbed on the catalyst's surface that there was not enough adsorbed hydrogen to convert the formed CO into CH 4 . In the case of the Ni/C sample, authors observed that CH 4 , CO and water were produced, indicating the presence of a sufficient amount of adsorbed hydrogen on the Ni catalyst surface. However, they observed that a longer period of time was needed for detecting CH 4 in the reactor effluent, whereas CO appeared from the beginning of the reaction. Thus, they proposed that CO was an indispensable reaction intermediate. Wang et al. [167] studied Ni-Ce/CNTs and proposed a mechanism where Ce, in conjunction with the CNTs support, donated electrons and facilitated their transfer from the support to the metallic nickel and the absorbed CO 2 , accelerating the activation of CO 2 molecules and the formation of atomic hydrogen. Then, the activated CO 2 dissociated on the surface of Ni formed carbonate species further hydrogenated to methane. They also pointed out that CO could be formed as by-product during this process, without playing a role in CH 4 formation.

Zeolite-Based Catalysts
Several works have been reported in the last few years dealing with the utilization of zeolite-based materials as catalysts for CO 2 methanation, with a recent review recently published on this topic [24]. Consequently, only the main results found in the literature recently are shown in Table 8 in order to allow a proper comparison with the rest of the catalysts presented in the current work.
Starting by an interesting study recently reported by Hussain et al. [171], fibrous silica-mordenite (without any additional active metal) was synthesized by microemulsion and tested under CO 2 methanation conditions, comparing its properties and performances with those obtained by a commercial MOR zeolite. Both samples presented CH 4 yields below 50% in all the temperature range studied (250-500 • C), but fibrous SiO 2 -MOR exhibited better performances (conversion, selectivity and stability for 50 h under reaction conditions), especially at higher temperatures. This behavior was explained by the higher basicity, textural properties and oxygen vacancies found in the sample, able to promote the activity towards CO 2 methanation and suppress coke deposition, boosting stability with time-on-stream.

Monometallic Zeolite-Supported Catalysts
Franken et al. [172] recently reported Fe/13X (intact and collapsed zeolite structure) catalysts with increasing metal loadings (1, 5 and 10 wt%) and tested under atmospheric and elevated pressures (5-15 bar). Authors verified that the preservation of the zeolite structure allows obtaining higher metallic dispersion and, hence, more favorable performances. In addition, better performances were obtained for the sample presenting the lowest Fe loading, where highly dispersed metallic particles were formed. Generally, authors verified that higher pressures favored the catalytic performances as expected taking into account the characteristics of the Sabatier reaction. However, the 1 wt% Fe/13X catalyst exhibited lower reaction rates at 15 than at 10 bar, which was ascribed to limitations in terms of intermediate species desorption or modifications on Fe species nature at elevated pressure.
Kitamura Bando et al. [173] studied Rh-Y zeolite and Rh/SiO 2 catalysts and performed tests at 30 atm. Authors reported better performances for the Rh-Y but it possessed an intense deactivation after 100 min due to the accumulation of H 2 O (arising from the reaction) inside the cavities of the zeolite. Later, Graça et al. [184] reported Ni/USY catalysts and evaluated the effects of the Ni incorporation method (ion exchange or impregnation) and Ni loading, with 15 wt% found to be the best compromise. The same authors [188,189] reported that the chosen calcination and reduction temperatures as well as the impregnation solvent can be crucial in these materials. Bacariza et al. evaluated the effects of the compensating cation nature [190], Si/Al ratio [174] and zeolite framework type [175] in the performances of Ni-based zeolites. Among all, larger cations (e.g., Cs + ), higher Si/Al ratios and USY zeolite led to the best results. In these studies, an additional relevant property of the support was proposed: the hydrophobicity. In fact, several zeolite samples series presenting a systematic and gradual enhancement of hydrophobicity exhibited improved methane yields, which was attributed to a reduced negative effect of the competitive adsorption of produced water over the same sites adsorbing CO 2 [118].
Goodarzi et al. [176] prepared 5 wt% Ni catalysts supported over silicalite-1 (S1) and desilicated silicalite-1 (d-S1) zeolites. Authors observed an improvement of the catalytic performances when using as support d-S1 support, mainly due to the confinement effect induced in Ni 0 particles. Chen et al. [178] reported recently a novel Ni-ZSM-5 catalyst synthesized by hydrothermal method and using as Si source a Ni/SiO 2 catalyst prepared by impregnation. They compared the properties and performances of this material with conventional Ni/SiO 2 and Ni/ZSM-5 catalysts (all with~10 wt%). Ni-ZSM-5 was strongly stable exhibiting an excellent stability over 40 h time on stream, attributed to an embedment structure able to hinder sintering, while the others reflected severe deactivation after 5 h. Guo et al. [177] compared a Ni/ZSM-5 catalyst with equivalent Ni samples supported over SBA-15, MCM-41, Al 2 O 3 and SiO 2 . Among them, the Ni/ZSM-5 sample presented the best performances. As already found by Goodarzi et al. [176], desilication provided an encapsulation effect for Ni particles, being CH 4 yields significantly improved.
Czuma et al. [180] synthetized an X zeolite from waste fly ashes to be used as support in the preparation of Ni catalysts for CO 2 methanation. Despite the partial micropores blockage due to Ni incorporation, the catalyst still presented an adequate surface area, showing a predominant presence of strong basic sites promoting the formation of very stable CO 2 species and, consequently, low methane yields. In spite of the CO 2 conversions obtained (~50%) being slightly lower than those obtained when using commercial zeolites, the use of fly ash zeolites offers the opportunity to use waste from energy sector, which is a positive economic and ecological aspect. Wei et al. [181] prepared 5 wt% Ni supported 5A and 13X zeolites and studied the effects of the Ni precursor salt (nitrate, citrate or acetate) and calcination temperature (300, 350, 400 and 450 • C) on the properties and performances. While no remarkable enhancements were found in the 5A-based samples, the activity, selectivity and stability were highly improved when using nickel citrate as salt and 350-400 • C as calcination temperature for 5Ni/13X (320 • C: 79% of CO 2 conversion, 100% of CH 4 selectivity and stability after 200 h of time-on-stream). They attributed the higher performances of 13X samples to the location of Ni species in the zeolite channels, allowing a confinement effect, while the use of citrate and a mild calcination temperature improved Ni species reducibility and metallic dispersion. Recently, Da Costa-Serra et al. [182] reported a series of ITQ-2 (prepared by delamination using ITQ-1 and MCM-22 as starting materials) and ZSM-5 zeolites impregnated with 5 wt% Ni. They evaluated the effect of the zeolite framework type and the Si/Al ratio and elaborated an enhancement of the performances when decreasing the Al content for both structures due to the improvement of the materials' hydrophobic properties. The best results were revealed by the sample supported over the ITQ-2 with Si/Al = ∞, a purely silica zeolite that led the formation of highly dispersed Ni 0 nanoparticles (<3 nm).
Monometallic zeolites have been also applied in sorption-enhanced methanation. Indeed, Delmelle et al. [191] prepared 5Ni/13X and 5Ni/5A catalysts by impregnation obtaining similar CH 4 yields. Borgschulte et al. [192] studied a Ni-5A catalyst prepared by ion exchange with <6 wt% Ni, obtaining a CH 4 yield of~100% when using a H 2 :CO 2 ratio of 8:1. Walspurger et al. [193] mixed a Ni-based commercial catalyst with a hydrophilic 4A zeolite and, as suggested by the previous studies from Delmelle et al. [191] and Borgschulte et al. [192], attributed the beneficial effects of using an hydrophilic zeolite to the displacement of the reaction equilibrium to the formation of CH 4 . In this context, Isah et al. [178] reported recently a beneficial effect when adding a small fraction (2 wt%) of zeolite over a 10 wt% Ni/Al 2 O 3 catalyst, that improved the textural properties without affecting the metallic surface area available. As shown, many works have dealt with the preparation of Ni-supported zeolites. Based on a comparison of the zeolite type (USY [174], MOR [175], ZSM-5 [175], BEA [179], 13X [181], 5A [181] or ITQ-2 [182]- Figure 11), one can observe that USY, MOR, ZSM-5 and ITQ-2 present generally better performances, while 5A is the zeolite leading to the lowest performances. Monometallic zeolites have been also applied in sorption-enhanced methanation. Indeed, Delmelle et al. [191] prepared 5Ni/13X and 5Ni/5A catalysts by impregnation obtaining similar CH4 yields. Borgschulte et al. [192] studied a Ni-5A catalyst prepared by ion exchange with <6 wt% Ni, obtaining a CH4 yield of ~100% when using a H2:CO2 ratio of 8:1. Walspurger et al. [193] mixed a Ni- Figure 11. CH 4 production rates per mass of catalyst (left) or Ni (right) determined, far from thermodynamic equilibrium, for a series of Ni/zeolites prepared by impregnation method as a function of the reaction temperatures. Italic numbers correspond to the wt% of Ni in the catalysts.

Bimetallic Zeolite-Supported Catalysts
Regarding bimetallic catalysts, several studies reported the beneficial effects of adding promoters such as Mg (Bacariza et al. [183]), Ce (Graça et al. [184], Bacariza et al. [185]) or La (Quindimil et al. [179]) over Ni/Zeolite catalysts due to the general enhancement of the CO 2 adsorption capacity and Ni 0 dispersion. In addition, Boix et al. [186] reported interesting results when using Pt-Co/MOR catalysts due to the formation of PtCo x O y active species. Recently, Wei et al. [187] reported bimetallic Ni-Ru catalysts supported over 13X and 5A zeolites. Authors kept constant the total metal content (5 wt%) and varied Ru and Ni loadings. They also prepared two monometallic reference samples containing 5 and 2.5 wt% Ni and Ru, respectively. Despite the effects of Ru on the selectivity to CH 4 , no remarkable enhancement of the catalytic performances was generally achieved when preparing bimetallic Ni-Ru/Zeolite catalysts, being 5Ni/13X and 5Ni/5A the catalysts exhibiting the highest CO 2 conversions. In terms of mechanism, Westermann et al. proposed some mechanism insights for Ni/USY [194] and Ce/Ni/USY [195] zeolites prepared by impregnation. In both cases formate species were reported as key intermediates.

Other Types of Supported Catalysts
Finally, some catalysts not included in the previous chapters, found in the literature, are presented in Table 9 and discussed below.
In this way, Lu et al. [196] studied the promotion effect of VO x (3, 5 and 8 wt% of V 2 O 5 ) on 20 wt% Ni catalysts supported on raw (RB) and modified bentonite (B) and prepared by impregnation method. The presence of more dispersed Ni 0 particles in the modified bentonite explained the better results obtained for this sample. After adding VO x , the catalytic activity was further improved due to the higher H 2 uptakes, the increased Ni 0 dispersion and the superior anti-coking and anti-sintering properties of the promoted samples. Additionally, Le et al. [197] reported Ni-based catalysts using SiC and SiO 2 as supports prepared by wet impregnation and deposition-precipitation (DP) methods. Authors observed an enhancement of the metal-support interactions and metallic dispersion when using DP method, which induced a significant effect in the improvement of the CO/CO 2 methanation performances over the SiC supported catalyst. SiC turned out to be the most behaving support due to its superior thermal conductivity.
Additionally, authors reported the beneficial effect of including Mn in the formulation, due to the favored metallic dispersion and CO 2 activation capacity. In addition, Zhi et al. [198] studied Ni and 15%Ni-5%La catalysts supported on SiC and prepared by impregnation. La 2 O 3 promoted the dispersion of Ni species on the SiC support simultaneously increasing the number of active sites on the catalyst surface and leading to better performances than the one reported for the Ni/SiC catalyst.
Kirchner et al. [199] studied different Fe 2 O 3 catalysts and reported the best activity for nanosized γ-Fe 2 O 3 . Authors associated the efficiency of this catalyst with reactive surface carbon species. Baysal and Kureti [200] also reported the use of commercial α-Fe 2 O 3 promoted with alkali (Li, Na, K, Rb and Cs), alkali earth (Mg, Ca and Ba), transition (Al, Mn, Cu and Mo) or rare earth (La, Ce and Sm) metals for CO 2 methanation at variable pressures. Among all, Mg (2 wt%) was identified as the most promising promoter. Indeed, MgO species, highly dispersed over iron oxide, were responsible for favorable changes in the iron phase composition and in the intermediate carbon species (e.g., Fe carbides).
Wang et al. [201] prepared hierarchically porous network-like Ni/3D draped Co 3 O 4 catalysts, where Ni clusters and nanodeposits were uniformly dispersed across the high surface area of the Co 3 O 4 support. The pre-reduction treatment was found to induce oxygen vacancies formation, improving Ni-Co interaction, enhancing the catalytic performances and lowering the activation energy. Wang et al. [202] studied two perovskite LaCoO 3 based catalysts: PdO-LaCoO 3 with an encapsulated structure and PdO/LaCoO 3 with a surface dispersing structure. The best catalytic performances were achieved by the encapsulated structure, which was attributed to the strong interaction of PdO NPs and LaCoO 3 and also to the H 2 spillover effect of Pd. Thus, the decomposition of perovskite structure was significantly promoted, benefiting the extraction of metallic Co NPs from LaCoO 3 to prepare a well-dispersed catalyst. Besides, the spent catalysts revealed that well-dispersed catalytically active Co 2 C was in situ formed accompanied with perovskite structure decomposition during the catalytic reaction. LaCoO 3 and PdO/LaCoO 3 catalysts showed a weak interaction and the perovskite structure remained with few Co 2 C, explaining the significant difference of catalytic performance.
Branco and Ferreira [210][211][212][213] reported a series of Co and Ni-based bimetallic materials containing f block elements such as Th, U, La, Ce, Sm, Dy or Yb. Starting by Ni-Th and Ni-U catalysts [210], authors studied two preparation methods: controlled oxidation under dry air, using intermetallic binary compounds as precursors, and a modified sol-gel. The intermetallic route induced more favorable interactions between Ni and f block elements, resulting in better performances. Additionally, Th led to better results than U. Furthermore, authors used electrospinning followed by a controlled oxidation treatment for synthesizing Ni-La, Ni-Ce, Ni-Sm, Ni-Dy and Ni-Yb catalysts in the form of nanofibers and nanoparticles [211]. While the morphology of the catalysts did not affect the performances, results were found to increase along the lanthanide series, with Dy and Ce leading to the best performances. Authors also compared Ni-La, Fe-Dy and Co-Sm catalytic systems prepared by electrospinning followed by oxidation [212], being the first catalyst the most significantly active towards CO 2 methanation (Fe-Dy and Co-Sm presented >95% selectivity towards CO). Finally, Branco et al. [213] evaluated the effect of the calcination temperature on Co-Lanthanide (La, Ce, Sm, Gd, Dy and Yb) aerogels prepared by epoxide addition method. They observed that higher temperatures promote the formation of perovskite structures (except for Ce and Yb), enhancing catalysts basicity and activity, but presenting lower selectivity towards methane. Among all, the use of Ce as promoter and a calcination temperature of 900 • C led to the best results.
Zhen et al. [203] prepared a series of Ni/MOF-5 with increasing Ni loadings and obtained the highest catalytic activity with 10Ni/MOF-5 catalyst. This material presented high specific surface area and large pore volume that provided highly uniform dispersed Ni particles in the framework. Besides, this catalyst showed high stability and almost no deactivation during a long-term stability test (100 h).
Zhou et al. [204] studied Ni/TiO 2 catalysts prepared by impregnation followed by DBD plasma decomposition. Authors found out the structure-sensitivity of the reaction over Ni/TiO 2 catalysts and reported the beneficial effect of Ni (111) in the activity achieved. They suggested that, while the existence of multiple Ni facets induces a reaction mechanism passing through formate species as intermediates and with Ni only participating in H 2 dissociation, Ni (111) promotes also CO 2 dissociation into CO, leading to a mechanism where carbon monoxide acts as an intermediate for the final production of CH 4 . Kim et al. [214] studied Ru/TiO 2 catalysts and the effect of mixing anatase and rutile TiO 2 phases. Mixtures were done by using several ratios and at different stages: before RuO 2 deposition and before or after annealing. A synergetic effect was verified when the mixing was done before the annealing, leading to a higher catalytic activity. In this case, RuO 2 nanoparticles migrated towards the TiO 2 phase to be stabilized and lately reduced. Petala et al. [206] studied the effects of alkali (Li, Na, K, Cs) incorporation on Ru/TiO 2 based catalysts (0.5 and 5 wt% of Ru). The effect of the alkali promoters in the observed performances was stronger over the low Ru content catalyst (0.5%Ru/TiO 2 ) while, in the 5%Ru/TiO 2 sample no significant effects were verified. Catalytic activity was strongly improved with the addition of small alkali contents (0.2 wt%), following the order: TiO 2 (unpromoted) < Lĩ K < Cs < Na, with Na-promoted sample being 3 times more active that the unpromoted catalyst. Zuzeng et al. [207] evaluated the effects of doping TiO 2 with Y (0, 1.0, 2.0 and 3.0 wt%) to obtain Co/Y-TiO 2 catalysts (20 wt% Co). The best catalytic activity was exhibited by the 20%Co/2%Y-TiO 2 and the incorporation of Y increased the specific surface area, enhanced the reducibility, generated oxygen vacancies and improved CO 2 adsorption capacity through the creation of medium basic sites. Finally, Marwood et al. [215] used 2%Ru/TiO 2 to perform CO 2 methanation mechanism studies and observed the formation of CO as a reaction intermediate. However, they suggested that CO was not formed as a result of the dissociative adsorption of CO 2 but as a product of formate species.
Cerdá-Moreno et al. [208] reported Ni-Sepiolite catalysts, prepared by precipitation and impregnation, and a Ni-Todorokite synthesized by sequential method and with 5 wt% Ni. Authors studied not only the properties and performances but also proposed the mechanism over these two materials based on Operando DRIFTS studies. Overall, todorokite led to higher performances, even at lower reaction temperatures, mainly due to its ability to activate CO 2 forming intermediate species (carbonates) further hydrogenated to formates and, later, CH 4 (associative mechanism). Oppositely, sepiolite did not adsorb CO 2 , being CO the main reaction intermediate (dissociative mechanism). Regarding the preparation method followed for Ni/Sepiolite catalysts, precipitation improved the reducibility of the Ni species formed and the metallic dispersion.
Liang et al. [209] reported Ni/Attapulgite catalysts prepared by impregnation with varied Ni loading (5, 15 and 25 wt%). Higher Ni content reduced the textural properties of the material, the metallic-support interactions and the number of basic sites, while the average Ni 0 size increased. However, and in line with several works already discussed, increasing the metal loading improved the activity and, mainly, the selectivity to CH 4 as a result of the CO formation suppression.
In order to compare among the different types of new or not widely explored supports and as previously done, CH 4 production rates were determined far from the equilibrium for Ni-based bentonite [196], SiC [197], Co 3 O 4 [201], MOF-5 [203], sepiolite [208], todorokite [208] and attapulgite [209] (Figure 12). Based on the plotted data, the use and optimization of Co 2 O 3 , todorokite and sepiolite-based catalysts in CO 2 methanation reaction constitute an interesting topic for further research in this area.

Final Remarks and Conclusions
As concluded from the results presented in previous chapters, supported nickel catalysts are the most studied for CO2 methanation reaction. However, their main drawback arises from their sintering processes and the deactivation at low temperatures, which is due to the interaction of nickel metal particles with CO and the formation of mobile nickel sub-carbonyls. Consequently, nickel catalysts are typically modified by adding promoters and stabilizers (e.g., Ce, Zr, La, Ti, Mg) in order to avoid deactivation processes [216]. Additionally, ruthenium could be considered as the alternative metal used on CO2 methanation catalysts since, although more expensive, is highly active and stable for this reaction. Even if several studies have also reported interesting results when using alternative metals (e.g., iron, cobalt), they typically required pressures above the atmospheric, as their activity is generally lower than that exhibited by Ni catalysts.
Regarding the supports, it can be concluded that their nature is crucial for the achievement of significant catalytic performances. Indeed, the dispersion and reducibility of the active metals, the activation of CO2 and H2, and the reaction mechanisms are deeply influenced by the choice of the support. As seen, SiO2 and Al2O3 have been widely and extensively studied due to their good initial activities and relatively cheap prices. However, SiO2 and Al2O3 based catalysts often suffer from sintering and serious carbon deposition at high temperatures. Additionally, CeO2 and ZrO2 have been revealed responsible for CO2 activation due to the presence of oxygen vacancies in their structure as well as promoting the reducibility and dispersion of other metals. Mesoporous SiO2 materials allow the favorable dispersion of Ni particles due to their porous structure, as occurs also in some carbon and zeolite-based supports, with the last being easily tunable in terms of hydrophobicity and basicity. MgO and hydrotalcite-derived materials present beneficial effects in terms of interaction with CO2 due to the enhanced basicity of these materials even if the reducibility of the metal species is sometimes compromised due to the strong metal-support interactions. Finally, some other materials were found promising for CO2 methanation, such as Fe and Co oxides, SiC, MOFs, sepiolite, todorokite or TiO2. Figure 12. CH 4 production rates per mass of catalyst (left) or Ni (right) determined, far from thermodynamic equilibrium, for a series of Ni-supported catalysts prepared using non-conventional supports a function of the reaction temperatures. Italic numbers correspond to the wt% of Ni in the catalysts.

Final Remarks and Conclusions
As concluded from the results presented in previous chapters, supported nickel catalysts are the most studied for CO 2 methanation reaction. However, their main drawback arises from their sintering processes and the deactivation at low temperatures, which is due to the interaction of nickel metal particles with CO and the formation of mobile nickel sub-carbonyls. Consequently, nickel catalysts are typically modified by adding promoters and stabilizers (e.g., Ce, Zr , La, Ti, Mg) in order to avoid deactivation processes [216]. Additionally, ruthenium could be considered as the alternative metal used on CO 2 methanation catalysts since, although more expensive, is highly active and stable for this reaction. Even if several studies have also reported interesting results when using alternative metals (e.g., iron, cobalt), they typically required pressures above the atmospheric, as their activity is generally lower than that exhibited by Ni catalysts.
Regarding the supports, it can be concluded that their nature is crucial for the achievement of significant catalytic performances. Indeed, the dispersion and reducibility of the active metals, the activation of CO 2 and H 2 , and the reaction mechanisms are deeply influenced by the choice of the support. As seen, SiO 2 and Al 2 O 3 have been widely and extensively studied due to their good initial activities and relatively cheap prices. However, SiO 2 and Al 2 O 3 based catalysts often suffer from sintering and serious carbon deposition at high temperatures. Additionally, CeO 2 and ZrO 2 have been revealed responsible for CO 2 activation due to the presence of oxygen vacancies in their structure as well as promoting the reducibility and dispersion of other metals. Mesoporous SiO 2 materials allow the favorable dispersion of Ni particles due to their porous structure, as occurs also in some carbon and zeolite-based supports, with the last being easily tunable in terms of hydrophobicity and basicity. MgO and hydrotalcite-derived materials present beneficial effects in terms of interaction with CO 2 due to the enhanced basicity of these materials even if the reducibility of the metal species is sometimes compromised due to the strong metal-support interactions. Finally, some other materials were found promising for CO 2 methanation, such as Fe and Co oxides, SiC, MOFs, sepiolite, todorokite or TiO 2 .
To conclude, this work aims to become a starting point or a roadmap for researchers intending to enhance heterogeneous catalysts properties and performances towards CO 2 methanation. Indeed, a remarkable number of publications regarding supported catalysts for thermal Sabatier reaction were summarized not only in terms of main conclusions but also in terms of achieved catalytic performances and preparation conditions used so that researchers could have an idea about the main properties to guarantee and the main drawbacks of the chosen supports.
Based on the previous findings, we claim that the main properties to be considered for CO 2 methanation catalysts must be the ones summarized in Figure 13 and described below.

•
Metallic dispersion. Based on the analyzed studies, catalysts presenting higher metallic dispersion and smaller particles present typically better performances due to the favored H 2 dissociation capacity. These properties can be enhanced by tuning the preparation conditions (e.g., method, solvent, calcination and reduction temperatures), adding promoters (e.g., CeO 2 , MgO, ZrO 2 , La 2 O 3 ) or, among all, through encapsulation strategies; • Basicity. CO 2 can be, based on the previously summarized works, adsorbed and activated on basic sites. Generally, authors have identified medium strength basic sites as the most promising for CO 2 methanation, since stronger sites lead to the formation of inactive carbonate species. These types of sites can be obtained by adding a basic promoter, such as MgO, or using a support able to interact with CO 2 (e.g., CeO 2 , zeolites, Al 2 O 3 ); • Oxygen vacancies. Several authors referred to the presence of oxygen vacancies and its responsibility in the enhancement of CO 2 activation. This can be achieved by using pure or modified CeO 2 , ZrO 2 or TiO 2 oxides or even KCC-1 mesoporous material; • Metal-support interactions. Typically, this property is related to the metal dispersion and average particle sizes. Indeed, strong metal-support interactions (SMSI) can promote the formation of smaller and well-dispersed metallic particles, typically resisting sintering and carbon formation. However, the use of higher reduction temperatures could be required. SMSI can be obtained by tuning the preparation conditions. Indeed, while impregnation typically leads to weaker interactions, the insertion of the active metals into the support framework (e.g., ZrO 2 , CeO 2 , Al 2 O 3 , TiO 2 ) represents a promising alternative; • Reducibility. Even if few works suggested a possible role of metal oxides in the reaction (e.g., NiO), metallic states of transition or noble metals are typically considered as the active phases for the CO 2 methanation reaction. This property, hardly dependent on the metal-support interactions, could be responsible for a reduction in the number of available active sites. It is important to deem in mind that increasing the reduction temperature to maximize the amount of reduced species can lead to severe sintering processes, which can negatively influence the observed performances. Authors found improvements in reducibility by adding promoters (e.g., CeO 2 , MgO, ZrO 2 , La 2 O 3 ); • Hydrophobicity. Water inhibitory effect in the CO 2 methanation reaction was proved by several research studies. From the literature about zeolite supports, it was pointed out that the lower the affinity of the support for water adsorption (higher hydrophobicity), the higher the methane yields produced. The adsorption of water on the same sites for CO 2 adsorption may comprehensively induce a negative kinetic effect on the methanation process. Consequently, it is important to use supports and/or promoters presenting low and weak affinity with water; • Textural properties. Mesoporous materials have gained attention in the last years for this reaction as they can be responsible for encapsulation effects able to reduce metallic particle sizes and strengthen metal-support interactions, resulting in higher catalytic performances. Apart from the use of conventional mesoporous materials, carbons or zeolites as supports, strategies for obtaining ordered mesoporous structures such as Al 2 O 3 represent promising strategies.  Finally, further research on the topics presented below could be useful towards a better understanding of CO2 methanation reaction and its application at industrial scale:


Supports. The optimization and further development of new catalysts based on promising but not deeply explored materials such as MOFs or mesoporous silicas would be a promising pathway [217][218][219]. In addition, the utilization of waste materials (e.g., fly ash, rice husk) as support precursors constitutes an interesting route towards higher cost-efficient catalysts [220];  Active metals. The systematic analysis of alternative transition (e.g., Co, Fe) or noble (e.g., Ru, Rh) active metals for CO2 methanation is a topic of high interest. Complementary, further efforts towards the identification and optimization of synergistic effects in bimetallic systems (e.g., Ni-Co, Ni-Fe) through advanced characterization techniques would be highly valuable;  Mechanistic approaches. Modelling and in situ/Operando spectroscopy studies carried out using density-functional theory (DFT), DRIFTS or Operando FTIR under conventional and more realistic methanation conditions (e.g., incorporation of pollutants/minor compounds present in flue gases, biogas or hydrogen streams in the feed) will be advantageous. In addition, further studies dealing with the elucidation of the deactivation mechanisms over different types of catalysts, using as basis long-term and aging experiments, will be helpful;  Catalytic testing. Although few works dealt with the effects of incorporating CO, CH4 or even steam in the reactor feed, evaluating CO2 methanation performances under realistic conditions will be key for identifying strategies to obtain catalysts with high resistance to, among all, oxygen, steam or H2S. In addition, further research on the preparation of scale-up catalysts (e.g., monoliths) will be important.   Finally, further research on the topics presented below could be useful towards a better understanding of CO 2 methanation reaction and its application at industrial scale:

•
Supports. The optimization and further development of new catalysts based on promising but not deeply explored materials such as MOFs or mesoporous silicas would be a promising pathway [217][218][219]. In addition, the utilization of waste materials (e.g., fly ash, rice husk) as support precursors constitutes an interesting route towards higher cost-efficient catalysts [220]; • Active metals. The systematic analysis of alternative transition (e.g., Co, Fe) or noble (e.g., Ru, Rh) active metals for CO 2 methanation is a topic of high interest. Complementary, further efforts towards the identification and optimization of synergistic effects in bimetallic systems (e.g., Ni-Co, Ni-Fe) through advanced characterization techniques would be highly valuable; • Mechanistic approaches. Modelling and in situ/Operando spectroscopy studies carried out using density-functional theory (DFT), DRIFTS or Operando FTIR under conventional and more realistic methanation conditions (e.g., incorporation of pollutants/minor compounds present in flue gases, biogas or hydrogen streams in the feed) will be advantageous. In addition, further studies dealing with the elucidation of the deactivation mechanisms over different types of catalysts, using as basis long-term and aging experiments, will be helpful; • Catalytic testing. Although few works dealt with the effects of incorporating CO, CH 4 or even steam in the reactor feed, evaluating CO 2 methanation performances under realistic conditions will be key for identifying strategies to obtain catalysts with high resistance to, among all, oxygen, steam or H 2 S. In addition, further research on the preparation of scale-up catalysts (e.g., monoliths) will be important.