Versatile Ni-Ru catalysts for gas phase CO 2 conversion: Bringing closer dry reforming, reverse water gas shift and methanation to enable end-products flexibility

Advanced catalytic materials able to catalyse more than one reaction efficiently are needed within the CO 2 utilisation schemes to benefit from end-products flexibility. In this study, the combination of Ni and Ru (15 and 1 wt%, respectively) was tested in three reactions, i.e. dry reforming of methane (DRM), reverse water – gas shift (RWGS) and CO 2 methanation. A stability experiment with one cycle of CO 2 methanation-RWGS-DRM was carried out. Outstanding stability was revealed for the CO 2 hydrogenation reactions and as regards the DRM, coke formation started after 10 h on stream. Overall, this research showcases that a multicomponent Ni-Ru/ CeO 2 -Al 2 O 3 catalyst is an unprecedent versatile system for gas phase CO 2 recycling. Beyond its excellent performance, our switchable catalyst allows a fine control of end-products selectivity.


Introduction
Climate change is a major challenge faced by the entire global community.Greenhouse gas (GHG) emissions continue to rise due to anthropogenic activities.In particular, carbon dioxide (CO 2 ) emissions are the most alarming ones due to the global dependency on fossil fuels [1].In 2019, the power sector itself accounted for 36% of the global CO 2 emissions [2].It is estimated that CO 2 is responsible for 60% of the global warming effects [3].Its average monthly recorded concentration in September 2021 was 413 ppm, the highest ever [4].One solution is carbon capture and storage (CCS) [1,5], but CO 2 utilisation is another attractive solution as it addresses the issue of resource depletion by using CO 2 as a carbon pool, eventually producing valuable chemicals and fuels [6][7][8].
DRM produces syngas using the two most abundant greenhouse gases, CO 2 and CH 4 .The H 2 /CO ratio of the produced syngas is lower than the stoichiometric 1:1, because of the simultaneous occurrence of the RWGS reaction.The low ratio from DRM is more suitable for long chain hydrocarbon production via Fischer-Tropsch (FT) synthesis [10,17,18] and as a result, DRM has attracted attention recently [10,17,[19][20][21][22][23][24].RWGS is of interest, especially for fixing the syngas ratio in various applications and when combined with other processes, such as the FT synthesis and CAMERE process, which produce hydrocarbons and methanol, respectively [12,25,26].More specifically, the combination of the RWGS and the methanol synthesis in the so-called CAMERE process leads to higher methanol production [27][28][29][30].CO 2 hydrogenation to methane, which is also referred to as synthetic natural gas (SNG), using H 2 from renewable energy driven water electrolysis is known as Power-to-Gas technology.Power-to-Gas schemes are appealing as a large scale storage option for renewable electricity produced in excess of demand [14,[31][32][33].
As high temperatures are needed for the endothermic reactions, catalysts are prone to deactivation via sintering [21,34].Even for the exothermic CO 2 methanation reaction, the development of hot spots may occur thus deactivating the catalysts through thermal sintering [35,36].Carbon deposition is another major culprit while these reactions are carried out because the active sites of the catalyst are covered [18].Coking of the catalysts occurs because of some side reactions, i.e.Boudouard, CO and CO 2 reduction, and CH 4 decomposition (4-7), that take place simultaneously and influence the catalytic performance [16,18,34,37].
Among the different active phases, Ni is the most frequently studied metal due its low cost and high activity in the DRM [18,38,39], in the RWGS [40][41][42] and in the CO 2 methanation [43][44][45].Ni is prone to deactivation due to the formation of coke and sintering and alternatives are needed to boost the catalytic performance [46].Noble metals, such as ruthenium (Ru), rhodium (Rh), palladium (Pd), and platinum (Pt), have shown high activity and stability as well as resistance to sintering and carbon formation in the DRM [18,47,48], in the RWGS [49,50] and in the CO 2 methanation [51,52].Ru especially is the most active CO 2 methanation catalyst [53] and it is the cheapest noble metal (Ru = 22.2€/g, Rh = 530.7€/g,Pd = 69.8€/g,Pt = 29.6€/g on 1 Sept. 2021) [54].A synergistic effect has been achieved in prior reports upon combining Ni and Ru, which resulted in increased Ni dispersion, enhanced performance and lower Ru loading [31,55,56].
In terms of supports and promoters, cerium oxide (CeO 2 ) has been extensively studied as a support in reforming [19,38,57] and hydrogenation reactions [49,58,59], due to its excellent oxygen storage capacity.CeO 2 forms oxygen vacancies, helping the oxidation of carbon [57,60,61].CeO 2 -Ni interfaces in particular have been shown to facilitate activation of CO 2 [62].CeO 2 has a low specific surface area which can be improved by combination with aluminium oxide (Al 2 O 3 ) [57], due to the latter's high specific surface area which can improve metal dispersion [22,63].However, its increased acidity enhances coke formation and sintering [22].Consequently, the combination of CeO 2 and Al 2 O 3 is a promising support since the result is a high specific surface area basic material.The addition of CeO 2 into an Al 2 O 3 -supported catalyst used in the DRM, the RWGS and the CO 2 methanation reactions separately, enhances its activity and stability, promotes the oxidation of carbon deposits and increases the active metal dispersion, while extending the catalyst's lifetime due to its excellent redox properties [19,38,42,57,60,64].Ce 0.5 Zr 0.5 O 2 has reported resistance to sintering at elevated temperatures, and a remarkable oxygen accessibility and mobility on the surface [21,31,44,61].Characteristically, this support has oxygen bulk diffusion two times higher than ceria itself hence leading to an enhanced oxygen storage and an increased efficiency in the redox reactions of ceria [61].
While these reactions are often treated separately, CO 2 activation forms a common starting point to develop high performance catalysts with robust activity for all three CO 2 utilisation reactions.This flexibility in operation would then enable switchable process designs, that adjust to the changing demands in natural gas and chemicals production from CO 2 .Processes using switchable catalysts can operate in "Power-to-gas" mode to store renewable energy at times of excess production and can otherwise be switched to the chemicals production mode to supply CO or syngas as carbon building blocks to the chemical industry.Hence, switchable catalysts offer us means to reconcile a continuous supply of captured CO 2 with a variable demand for products of CO 2 utilisation in power-to-gas schemes.
No research has been done into combining these three reactions: DRM, RWGS, and CO 2 methanation.In this study, we show that a single catalyst can be used for all three of them and would enable flexibility needed for a combined power, gas, chemicals and liquid fuels grid, based on the recycling of greenhouse gases.This switchable catalyst should be both active and stable for all of them by altering its temperature and reactants.Our team demonstrated that a combination of 15 wt% Ni and 1 wt% Ru displayed promising activity for the RWGS and CO 2 methanation reactions, resulting in an exceptional performance for CO 2 hydrogenation [31].The RWGS and CO 2 methanation reactions compete against each other, but due to their different thermodynamic natures, the RWGS reaction is favoured at high temperatures and CO 2 methanation at lower ones.Nevertheless, careful catalyst design is required in the temperature range where the two reactions overlap in order to achieve the desired product selectivity [65,66].Ni catalysts, promoted with noble metals like Ru have also been used previously for DRM, [18,56] prompting us to explore Ni-Ru bimetallic catalysts as switchable catalysts that can operate with renewable hydrogen or in its absence to recycle greenhouse gases.Consequently, a bimetallic Ni-Ru mixture was chosen to be the active site of the catalysts along with two supports.The first one was CeO 2 -Al 2 O 3 and the second one Ce 0.5 Zr 0.5 O 2 , which are both promising supports.

Catalyst synthesis
The two catalysts were prepared by co-impregnation.Firstly, Ni (NO 3 ) 2 ⋅6H 2 O (Acros Organics) and Ru(NO)(NO 3 ) 3 solution (1.5 w/v Ru, Alfa Aesar) were mixed in 40 mL of deionised water and added to Ce 0.5 Zr 0.5 O 2 (Daiichi Kigenso Kagaku Kogyo Co.) in order to obtain a 15 wt% of Ni and 1 wt% of Ru catalyst.The same procedure was followed for the second catalyst, but the support used was a commercial CeO 2 -Al 2 O 3 support (SCFa-160 Ce20 Puralox, Sasol), containing 20 wt% of CeO 2 and 80 wt% of Al 2 O 3 .
The two resulting suspensions were initially stirred at room temperature and then concentrated in a rotary evaporator under reduced pressure.After that, the materials obtained were dried at 110 • C overnight and calcined at 500 • C (5 • C/min ramp) for 3 h.For the sake of simplicity, the two catalysts will be referred to as NiRu/CeZr and NiRu/ CeAl and the supports as CeZr and CeAl.

Catalyst characterisation
The specific surface area and the pore volume of each catalyst were calculated by using the Brunauer-Emmett-Teller (BET) equation and the Barett-Joyner-Halenda (BJH) method, respectively.The samples were initially degassed at 250 • C in vacuum for 4 h.Then, nitrogen adsorption-desorption measurements were carried out at liquid nitrogen temperature (-195 • C) in a Micrometrics 3Flex apparatus in order to obtain the textural properties of each catalyst and their supports.
The apparatus used for the X-ray diffraction (XRD) analysis was the X'Pert Powder from PANalytical.The 2θ o angle recorded was over the range of 10-90 • with a step size of 0.05 • every 240 secs.Additionally, by using a Cu Kα radiation (λ = 0.154 nm), the diffraction patterns were obtained at 30 mA and 40 kV.XRD analysis was carried out on fresh, reduced at 850 • C for 1 h, post-temperature screening and post-stability experiment samples.The average crystallite size was estimated using the Scherrer equation.
The H 2 -Temperature Programmed Reduction (TPR) was conducted in a continuous fixed bed quartz reactor.A mass spectrometer (Omni-Star GSD 320) measured the H 2 consumption while the data was logged by using the Quadera software package.50 mg of each catalyst were initially dried at 150 • C with argon (Ar) for 30 mins.After they had cooled down, a flow of 50 mL/min of a 10% H 2 /Ar mixture was used while the temperature was increased to 1000 • C with a 10 • C/min rate.The temperature programmed oxidation (TPO) was carried out on the spent catalyst in the same equipment as H 2 -TPR and the data were obtained accordingly.The samples were heated up from room temperature to 900 • C with a 50 mL/min flow of 3% O 2 /Ar and a ramp rate of 10 • C/min.Thermogravimetric analysis (TGA) was conducted on the spent catalyst after the stability test in an SDT Q600 V8.3 Instrument from TA Instruments.A flow of 100 mL/min of air was used, while the temperature was increased from room temperature to 900 • C at 10 • C/min.
X-ray photoelectron spectroscopy (XPS) measurements were carried out in a VGMicrotechMultilab 3000 spectrometer.The XPS spectra were obtained with a hemispherical electron analyser and a Mg Kα (h = 1253.6eV; 1 eV = 1.6302 × 10-19 J) 300-W X-ray source with a major axis length of 400 μm.The pass energy used was 50 eV.The samples were maintained in the analysis chamber until a residual pressure of ca. 5 × 10 -7 N/m 2 was achieved.CasaXPS software was used to carry out the XPS data analysis and the experimental data were fitted in a combination of Lorentzian (30%) and Gaussian (70%) lines.The C1s, whose binding energy (BE) peak was located at 284.6 eV, was used as a reference binding energy.Before the experiment, the samples were reduced ex situ at 850 • C and conserved in octane.

Catalytic activity
All the reactions were carried out in a tubular fixed bed quartz reactor with an outside diameter of 0.5 in, which was placed vertically in a tubular furnace.The catalysts were located on top of a quartz wool bed inside that reactor.A thermocouple was used to enable the monitoring of the temperature throughout the experiments and the deviation from the temperature setpoints if it had occured.An ABB AO2020 online gas analyser was used to detect the amount, i.e. the percentage, of CO 2 , CH 4 , CO, and H 2 in the product stream after water was separated in a chiller.The initial composition of the reactants was identified using the gas analyser and the total volumetric flow rate was measured using a bubble meter.The bubble meter was used because of the difference in the inlet and exit flow rates.The total outlet flow rate and the composition of the reactants were monitored frequently; they were measured/detected at every temperature in the temperature screening experiments and usually every 30 mins in the stability experiments.In the stability experiment, the trend of the flow rate was found based on the regular measurements, thus enabling the accurate calculation of the various parameters.
Prior to all experiments, an in situ reduction was carried out at 850 • C with a 50 mL/min flow of a 10% H 2 /N 2 mixture for 1 h.This reduction temperature was selected because of it being the highest operating temperature in our work.The catalysts powder particles used in this project were 100-200 µm.The mass of the catalyst in each temperature screening (TS) experiment was 0.125 g and the weight hourly space velocity (WHSV) 24 L/g h.Two TS experiments were carried out for each catalyst: one for the DRM reaction and one for the RWGS and CO 2 methanation reactions.The inlet stream in the DRM TS experiment had a mole composition of 20% CO 2 , 20% CH 4 , and 60% N 2 , whereas the one in RWGS and CO 2 methanation reactions had a composition of 10% CO 2 , 40% H 2 , and 50% N 2 .All the TS experiments started at the highest temperature and performed by decreasing the temperature in 50 • C increments of 30 mins.The DRM reaction was performed in the 550-850 • C temperature range while the RWGS-CO 2 methanation was conducted across the 200-850 • C range.
In the stability experiment conducted, the reaction temperature was initially 350 • C (CO 2 methanation), increased to 700 • C (RWGS) and finally the reactants were changed at the same temperature, allowing DRM to take place.It was reported in our previous work that the conversion of the NiRu/CeZr catalyst is slightly affected when the stability experiments start with the RWGS instead of the CO 2 methanation reaction [31].Therefore, the stability experiment in this research started with the CO 2 methanation reaction at 350 • C. The 700 • C temperature was selected so as to observe the deactivation of the catalyst, if any, without having a CO 2 conversion close to 100%.The total flow was mL/min with 10% CO 2 , 40% H 2 , and 50% N 2 for the RWGS and CO methanation, and 25% CO 2 , 25% CH 4 , and 50% N 2 for the DRM.The test was performed using 0.1 g of catalyst reaching a WHSV of 120 L/g h.Each reaction was carried out for 20 h, yielding a total time on stream of 60 h.
The equations used for measuring the catalytic activity are presented below (8)(9)(10)(11)(12)(13)(14).The F indicates the flow rates of the respective gas and the subscripts in and out refer to the inlet and outlet streams.

Thermodynamic simulations
Chemstations' ChemCad software was used so as to acquire the thermodynamic equilibrium conversions for all the reactants of the three reactions over a range of temperatures.The equation of state used was the Soave− Redlich− Kwong.The reactor was simulated as a Gibbs reactor in order to calculate thermodynamic equilibrium limits.The total inlet flow rate and the percentages of the reactants used in the ChemCad simulation were the same as those of both the temperature screening and the stability experiments.

Textural properties
The textural properties of the catalysts are presented in Table 1.The NiRu/CeAl catalyst had a higher surface area than the NiRu/CeZr, which was expected as alumina is widely used as support due to its large surface area [18,67].The isotherms of all the catalysts and supports are presented in Figure S1.The type of isotherms generated were Type IV with a characteristic H1 hysteresis loop, according to the IUPAC classification.This suggests the existence of well-developed cylindrical mesopores [68].This can be confirmed from the average pore diameter in Table 1.

Crystalline structure
The XRD patterns for the fresh and reduced NiRu/CeZr and NiRu/ CeAl samples are shown in Fig. 1a and Fig. 1b respectively.Both samples presented the characteristic peaks of γ-Al 2 O 3 and Ce 0.5 Zr 0.5 O 2 phases respectively (JCPDS 00-004-0880 & JCPDS 00-038-1436).While CeO 2 was detected in the fresh NiRu/CeAl sample (JCPDS 03-065-5923), it was not detected in the reduced sample.The cerium aluminate phase (CeAlO 3 ) phase was not detected here however its presence was confirmed by XPS.According to previous data found in literature, the CeAlO 3 phase should exist in the reduced and in the post-stability samples, but as highly dispersed or incorporated into, or over the other materials [19].The CeAlO 3 phase was in fact detected in the XRD pattern of the sample after H 2 -TPR (Figure S2) confirming the restructuration of ceria and alumina.NiO was detected in the fresh samples as there were diffraction peaks at 2θ o = 37.2 • , 43.3 • , 62.9 • and 75.4 • (JCPDS 00-047-1049).Nickel oxide was reduced to metallic Ni 0 in the reduced samples, which was considered as the active species of the reactions (JCPDS 01-070-1849).However, the existence of NiO in the reduced NiRu/CeZr sample suggested that the sample was not fully reduced after the in situ reduction (carried out at 850 • C for 1 h).This might have been because of the large nickel particle size maybe in conjunction with mass transfer limitation issues, thus hindering the reduction of internal nickel oxide [67].NiAl 2 O 4 spinels were not present in any of the samples.As for Ru crystalline phases, traces of RuO 2 appeared in the fresh samples (JCPDS 01-070-2662) and of metallic Ru 0 , in the reduced NiRu/CeAl sample (JCPDS 01-070-0274).The estimate metallic Ni particle size using the Scherrer equation at 2θ o = 44.48• were 12.4 and 34.1 nm for the NiRu/CeAl and NiRu/CeZr respectively.

Reducibility: H 2 -TPR
The redox properties and the interactions between the support and the metallic species are observed in the H 2 temperature programmed reduction.The H 2 -TPR profiles of the CeAl, CeZr, NiRu/CeAl, and the NiRu/CeZr are shown in Fig. 2. Overall, it was observed that Ni and Ru shifted the reduction of the supports at lower temperatures pointing out an intimate metal-support interaction.
As regards the CeAl support, a wide reduction process was observed.Ceria surface reduction took place between 150 • C and 700 • C, while the peak at 850 • C was attributed to the reduction of the bulk ceria crystallites.When Ni and Ru were added, the reduction of ceria was facilitated, showing a close Ni-CeO 2 interaction [38,69].The peak at 380 • C was mainly attributed to the reduction of the medium-sized NiO x species interacting with the CeO 2 particles [31,69,70].In the CeZr sample, a peak at 570 • C was evident where the reduction of Ce 4+ to Ce 3+ occurred.Concerning the NiRu/CeZr, the main peak at 320 • C was due to the surface oxygen reduction of CeZr and the Ni 2+ species interacting with the support.It appeared that Ni-Ru facilitated the reduction of the surface oxygen of the CeZr support, thus shifting the main reduction peak towards lower temperatures, showing the interaction between the Ni and Ru species and the support.In the NiRu/CeAl and NiRu/CeZr samples, the peaks at 130 • C and 170 • C respectively corresponded to the reduction of Ru 4+ to metallic Ru and their difference in the reduction peak position was associated with different forms of interaction with the support [31,55].In general, a strong interaction of the Ni species with the support is associated with a high catalytic activity [71,72].We will discuss this potential correlation within the studied CO 2 conversion reactions in the next sections.Figure 3a shows the Ni2p 3/2 spectra of the two reduced catalysts.In both catalysts, a mixture of metallic Ni (851-853 eV) and Ni 2+ (854-857 eV) was observed [31].Therefore, the Ni 2+ species were not fully reduced upon the reduction treatment employed prior to reaction for any of the catalysts.However, there was a higher amount of metallic nickel in the NiRu/CeZr sample compared to that of the NiRu/CeAl, as reported in the Ni 0 /Ni Total ratio in Table 2.This indicated that there was a strong metal-support interaction that facilitated the reduction of nickel oxide to metallic nickel in that sample.Regarding the NiRu/CeAl catalyst, it appeared that there was a higher amount of oxidised nickel, which was also evident in the H 2 -TPR results (Fig. 2), showing a slow rate of reduction after 450 • C.Moreover, the Ni/support ratio observed in Table 2 was higher in the NiRu/CeZr sample, indicating that a significantly higher Ni dispersion at the surface was achieved.In contrast with the XRD results, which suggested that only the Ni in the NiRu/CeZr hadn't been fully reduced, the XPS results illustrated that there was a higher amount of NiO in the NiRu/CeAl sample.The latter suggest a slightly different surface (XPS) and bulk (XRD) compositions which is common in this kind of Ni-based systems [31].

XPS was performed on reduced catalysts to obtain information about the chemical state of the elements and coordination environments of Ni
The binding energies of Ru3p 3/2 are presented in Table 2 and its spectra in Fig. 3b.Similarly to nickel, both metallic Ru and Ru 4+ existed in the two samples.However, the ratio of Ru 0 /Ru Total indicated that there was a higher amount of metallic Ru in the NiRu/CeAl sample, showing a higher reducibility of the Ru species in good agreement with the H 2 -TPR results.Additionally, a higher value was obtained for the Ru/support ratio in the case of NiRu/CeZr.This suggests that a higher Ru dispersion at the surface was achieved in that sample, which was in good agreement with its XRD data and with the absence of Ru 0 in its reduced form.
The Ce3d region is presented in Fig. 3c.A mixture of Ce 3+ and Ce 4+ was present in the NiRu/CeAl sample, corresponding to CeAlO 3 and CeO 2 , respectively.This was particularly evident in the significantly diminished peak at ~ 917 eV, suggesting a considerably lower amount of Ce 4+ [19,73].The CeAlO 3 phase was also detected by XRD on this catalyst after H 2 -TPR (Figure S2).A mixture of Ce 3+ and Ce 4+ was also observed in the NiRu/CeZr sample, thus both Ce 2 O 3 and CeO 2 existed.

Catalytic behaviour 3.5.1. Temperature screening experiments
The catalysts' performance as regards the CO 2 conversion in DRM is presented in Fig. 4a, while the H 2 /CO ratio in Fig. 4b.It can be seen that the two catalysts exhibited very similar performance.At higher temperatures, i.e. above 750 • C, the CO 2 conversions reached thermodynamic equilibrium.As the DRM is an endothermic reaction, higher temperatures are needed in order to achieve higher conversions.At lower temperatures, the NiRu/CeAl catalyst performed slightly better than the NiRu/CeZr.Nevertheless, the two catalysts appeared to have very similar CO 2 conversion.
As far as the CH 4 conversion was concerned, which can be seen in Figure S3, both catalysts were below the equilibrium, even at high temperatures.The NiRu/CeAl sample performed better overall, with a difference in CH 4 conversion between the catalysts being as high as 13% at 700℃, i. e. 62% for the NiRu/CeAl and 49% for the NiRu/CeZr.In general, CH 4 conversion is smaller than that of CO 2 due to two phenomena.Firstly, the energy barrier of methane activation, which is considered to be the rate-determining step, is larger than that of carbon dioxide activation.Secondly, the presence of carbon dioxide and hydrogen drives the RWGS, which is a side reaction of the DRM.Therefore, an amount of CO 2 reacts with H 2 instead of CH 4 and it is eventually converted into CO [34,74].
Just as in the CO 2 conversion, the H 2 /CO ratio did not vary significantly for the two catalysts.As aforementioned, this ratio should be 1 for the DRM.However, it was evident that above 650 • C, this ratio was above 1 as a result of two events.Firstly, since the RWGS is a side reaction of the DRM, H 2 O is present in the reactor.Therefore, the bireforming reaction, which has a syngas ratio of 2, occurs simultaneously [38,75].Secondly, methane decomposition (7) also takes place.Consequently, an excess amount of H 2 is produced, increasing the ratio [34].The methane decomposition was evident in the NiRu/CeZr sample at 850 • C.This increased value in Fig. 4b is in agreement with the CH 4 conversion of that catalyst at the same temperature (Figure S3).Additionally, the H 2 /CO ratio was below 1 for temperatures under 650 • C due to the RWGS [34].
Although the two catalysts showed similar performance in DRM, different behaviours were obtained for CO 2 hydrogenation.The CO conversion and the CO and CH 4 selectivity of the two catalysts in the RWGS and CO 2 methanation reactions are presented in Fig. 5a and Fig. 5b, respectively.The CO 2 conversion for both catalysts reached the equilibrium values at higher temperatures.It might have been due to the fact that CeO 2 also catalyses the RWGS, due to the existence of nearby oxygen vacancies at the nickel-ceria interface being considered as efficient active site for CO 2 hydrogenation [58].Concerning the methanation reaction, it was clear that the NiRu/CeAl catalyst had a superior activity; it achieved equilibrium conversion at a temperature as low as 350 • C. As this is reaction is structure sensitive [53,76], it was suggested that their difference in the CO 2 conversion was due to the smaller particle size of Ni in the NiRu/CeAl catalyst in good agreement with our XRD data.In addition to that, it is known that the CO 2 conversion is improved when the Ce 3+ is stabilised in the form of CeAlO 3 , which acts as a driving force for CO 2 activation [77], thus further boosting the   L.-P.Merkouri et al. activity of this catalyst.
The CO selectivity is increased with increasing temperature, while the opposite is true for the CH 4 selectivity.More specifically, RWGS was favoured at temperatures higher than 550-600 • C, and CO 2 methanation at lower ones.Fig. 5b shows that above 750 • C, both catalysts showed 100% CO selectivity due to the endothermic nature of RWGS.Below 700 • C, the NiRu/CeZr showed a higher selectivity towards CO, and the NiRu/CeAl towards CH 4 .At 400 • C, the latter achieved a 100% CH 4 selectivity, demonstrating its excellent performance.The NiRu/CeZr catalyst reached 100% CH 4 selectivity at 200 • C, but at such low temperatures, CO 2 conversion was less than 5%.The difference in selectivity of the two catalysts was attributed to smaller particle size of the NiRu/ CeAl catalysts, which favoured the CH 4 formation instead of CO [76] and to the CeAlO 3 phase, which enhanced the CO 2 activation in the CO 2 methanation reaction.
Certainly, the CO 2 activation offered us a starting point in order to search for switchable catalysts.The results of the catalytic activity experiments showed that the Ni-Ru catalysts are active for the dry reforming of methane, the reverse water-gas shift and the CO 2 methanation.Especially NiRu/CeAl is a very active catalyst for all the CO 2 utilisation reactions, which can provide us with the opportunity to provide chemicals and fuel in various power-to-gas schemes.

Stability experiment
The stability experiments are an essential part of catalytic research as they showcase whether the catalyst is durable and how much it is affected by coking and/or sintering over time.Additionally, as our vision of switchable catalysts is not only to be active in all the CO 2 utilisation reaction, but also to be stable in all of them.The switchable catalyst should be stable in these reactions while the modes are switched from the "Power-to-gas" mode" to store renewable energy at times of excess production to the chemicals-production mode to supply CO or syngas as carbon building blocks to the chemical industry and vice versa.As NiRu/CeAl showed overall a better activity compared to NiRu/CeZr, it was selected for the stability experiments.
In the stability experiment, the NiRu/CeAl catalyst manifested an impressive performance and stability in the CO 2 methanation; the CO 2 conversion remained constant at 70% after 20 h, while the methane yield was 69% throughout the entire duration.The NiRu/CeAl was also very stable in the RWGS reaction, having a 73% conversion, which is that of the equilibrium, and a 56% CO yield.These results indicated the excellent stability of the bi-metallic Ni-Ru catalysts in the CO 2 hydrogenation reactions in line with our previous findings [31].It should be noted that the methanation performance of NiRu/CeAl was different in the stability experiment compared to the temperature screening experiment because of the higher space velocity used in the former.As far as the stability of the catalyst in the DRM was concerned, Fig. 6 shows that the catalyst maintained a CO 2 conversion of 71% after 10 h on stream.After 20 h on stream, that conversion decreased by 10% mainly due to the catalyst's coke formation as shown by post reaction characterisation (Section 3.6).It is worth noting that the space velocity used was higher than the 10,000 h − 1 operational space velocities for static reformers in industrial applications [38,78].In any case, the excellent behaviour of our switchable catalyst under very high space velocity is commendable.Beyond the product flexibility, its performance at a high space velocity involves a significant reduction of the overall reactor volume (and hence its capital costs) thus allowing the design of compact CO 2 conversion units.
Undoubtedly, the impressive stability of NiRu/CeAl makes it an ideal candidate catalyst for our vision to use switchable catalysts in the various CO 2 utilisation scenarios.This catalyst showed a stable CO 2 conversion, even when the conditions changed, which is an important characteristic that a switchable catalyst should have.Obviously, NiRu/ CeAl catalyst has manifested a remarkable stability performance in the CO 2 hydrogenation reactions, but a catalyst improvement is needed to further enhance its performance in the DRM reaction and accomplish a stable cyclic operation between these three CO 2 utilisation reactions.

1 X-Ray diffraction (XRD)
The XRD patterns for the post temperature screening NiRu/CeZr and NiRu/CeAl samples are shown in Fig. 7a and Fig. 7b respectively.In the NiRu/CeZr XRD pattern, there was no NiO in contrast with the reduced sample (Fig. 1a) meaning that the remaining NiO was converted into metallic Ni during the reaction.A closer observation shows that there was a small carbon peak after the DRM temperature screening experiment (TS), suggesting that the catalyst had started to deactivate.This is in line with the methane decomposition observed during the TS experiment.However, there was no crystalline carbon peak in the NiRu/CeAl sample.By applying the Scherrer equation to find the average particle size of Ni, the extent of sintering was observed.The average Ni particle size was estimated to be 12.5 nm both post RWGS-CO 2 methanation and DRM TS for NiRu/CeAl and 33.5 nm post RWGS-CO 2 methanation TS and 34.7 nm post DRM TS for NiRu/CeZr (compared to the initial 12.4 and 34.1 nm for the NiRu/CeAl and NiRu/CeZr respectively).The minor differences on Ni particle size found in the reduced NiRu/CeZr sample was considered to be within the error limits during the calculations.Negligible sintering was observed in the NiRu/CeAl samples after the TS experiments.Concerning NiRu/CeZr, it was observed that sintering affected that catalyst more during DRM than RWGS and CO 2 methanation, suggesting that sintering is dependent on the reaction mixture.This might have been because of the formation of coke, which is generally known to enhance sintering, and/or the slow sintering rate in a H 2 atmosphere, thus in the CO 2 hydrogenation reactions [79].
The XRD pattern of the spent catalysts after the stability experiment are shown in Fig. 8.A carbon peak was observed at 2θ o = 26 • , indicating the crystalline coke formation following the stability experiment.Additionally, Fig. 8 reveals the extent of the sintering phenomenon, which resulted in a slight catalyst deactivation during DRM, as Ni particle size was calculated to be 14.3 nm.Consequently, there was a small degree of metallic sintering.In general, an increase in the Ni particle size was expected because the reactions temperatures of the RWGS and DRM were higher than the Tammann temperature of Ni [64].Therefore, the catalyst deactivation observed in the DRM stability was attributed to a slight degree of sintering, but mainly to carbon deposition.L.-P.Merkouri et al.

Carbonaceous deposits
As indicated in the XRD, crystalline carbonaceous deposits were detected on the samples.Fig. 9a shows the thermogravimetric analysis (TGA) performed on the post stability sample, while Fig. 9b shows its TPO profile.
The TGA plot shows the amount of coke deposition formed after 60 h on stream.It was observed that there was a carbon formation of 0.29 g c / g cat in the spent sample.The TPO profile shows two distinct peaks, meaning that multiple carbon types were present in the sample.The peaks at lower temperatures indicated the existence of softer carbon with a poorer degree of crystallinity, while the peaks at higher temperatures indicated the existence of harder carbon with greater degree of crystallinity.Harder carbon requires higher temperature to oxidise, while softer carbon can be regenerated at mild conditions [38].Therefore, the formation of softer carbon is more desirable compared to harder.Hence, in the TPO profile it was observed that there was a bigger amount of softer carbon, which is more advantageous if regeneration is to be performed.Overall, there was a high amount of carbon being formed indicated that further improvement of the catalyst is needed in the future.In any case, as positive outcome of the carbon deposits study, our TPO data indicate that no matter the reaction conditions the carbon deposits can be burned at relatively mild temperatures (700 • C) opening excellent opportunities for catalysts regeneration.

Conclusions
We report for the first time a switchable Ni-Ru catalyst able to effectively catalyse three CO 2 utilisation reactions in the gas phase, i.e. dry reforming of methane, reverse water-gas shift, and carbon dioxide    This study specifically examined the impact of the supports on the performance of a Ni-Ru catalyst.The physiochemical properties of the different catalysts were studied while their activities were tested over a range of temperatures and times on stream for all the reactions.The positive effects of a Ce-Al support were observed.The increased specific surface area of the NiRu/CeAl catalyst, attributed to the alumina content, and its Ni small particle size result are advantageous.Indeed, the Ce-Al catalyst showed remarkable activity and stability levels in the CO 2 methanation and RWGS reactions while maintaining a high yield to desired products CH 4 and CO respectively.Especially in the CO 2 methanation reaction, this catalyst was very active, mainly due to its small particle size.Similar successful results were also obtained for the DRM reaction, during which some degree of deactivation occurred.Deactivation due to carbon deposition was observed in the DRM stability experiment and motivates further development of switchable catalysts with increased resistance to coking in order to establish a cyclic operation between these three CO 2 utilisation reactions.Our TPO results encourage a potential catalyst regeneration at temperatures around 700 • C.
Overall, this paper outlines a novel direction for combating CO 2 emissions by developing catalysts for not individual CO 2 utilisation schemes but for a switchable operation scenario allowing for flexibility in CO 2 recycling based on renewable electricity availability and changing product demands.The NiRu/CeAl catalyst offers this opportunity to control products of CO 2 utilisation using reactant composition and temperature, and it is a major step towards our vision of adaptable CO 2 recycling.
L.-P.Merkouri et al.   methanation.Such switchable catalyst allows flexibility in synthesis of chemicals from CO 2 .

Table 1
Textural properties of the two catalysts.

Table 2
Binding Energies of the Ni 2p3/2 and Ru 3p3/2 levels for the reduced catalysts and Ni/support, Ni 0 /Ni Total and Ru 0 /Ru Total atomic ratios.