Effect of nickel on combustion synthesized copper/fumed‐SiO2 catalyst for selective reduction of CO2 to CO

In this study, we explore the effect of nickel incorporation in Cu/fumed‐SiO2 catalyst for CO2 reduction reaction. Two catalysts, Cu and CuNi supported on fumed silica were synthesized using a novel surface restricted combustion synthesis technique, where the combustion reaction takes place on the surface of the inert fumed‐SiO2 support. An active solution consisting of a known amount of metal nitrate precursors and urea (fuel) was impregnated on fumed silica. The catalyst loading was limited to 1 wt% to ensure localized combustions on the surface of fumed‐SiO2 by restricting the combustion energy density. The synthesized catalysts were tested for CO2 hydrogenation reaction using a tubular packed bed reactor between temperature 50°C and 650°C, where Cu/SiO2 showed high CO2 conversion to carbon monoxide, and the addition of Ni further improved the catalytic performance and showed some tendency for methane formation along with CO. Moreover, both the catalysts were highly stable under the reaction conditions and did not show any sign of deactivation for ~42 hours time on stream (TOS). The catalysts were characterized using X‐ray diffractometer (XRD), scanning electron microscope/energy dispersive X‐ray spectrometer (SEM/EDX), transmission electron microscope (TEM), and the Brunauer‐Emmet‐Teller (BET) surface area measurement technique to understand their structural properties and to assess the effect of CO2 conversion reaction. In situ DRIFTS was also used to investigate the reaction pathway followed on the surface of the catalysts.


| INTRODUCTION
The global energy demand has surged over the past decades and is expected to continuously grow in the near future to fulfill the demand generated by continuous improvement in human life quality. So far, most of the energy is produced from fossil fuel-based sources that emit carbon dioxide (CO 2 ) as a byproduct during energy generation process, which is known to adversely affect the global climate. 1,2 A high concentration of CO 2 in the atmosphere is directly linked to global temperature rise. 2 Being a thermodynamically stable molecule CO 2 can be captured and stored underground for a long time, or it can be utilized for producing commodity chemicals as a way of recycling CO 2 and reducing the overall global impact of their release into the atmosphere. While many groups are working on CO 2 capture and sequestration aspects, leading to industrial applications such as enhanced oil recovery etc. 3,4 ; more efforts are required in converting CO 2 into other carbon-containing products. Utilizing CO 2 for the production of chemicals and fuels seems a sustainable approach, though it requires considerable energy at this stage on account of the relatively inert nature of the CO 2 molecule. However, it is anticipated that the renewable energy will become cheaply available in the future, which will make the cost of CO 2 conversion to fuels and chemicals drop drastically. Catalysts play a critical role in lowering the energy barrier for CO 2 conversion reactions. A suitable catalyst can bring the energy requirement to a cost-competitive stage while selectively producing the desired products. Efforts in the direction of photo-catalysis, 5,6 electro-catalysis, [7][8][9][10][11][12][13] and thermal catalysis [14][15][16][17][18] are underway, each having its advantages and disadvantages. Nonetheless, the thermal catalysis route seems more favorable for hydrocarbon reforming industries as high-pressure CO 2 streams can be easily processed within the reforming facilities while simultaneously utilizing the excess heat for an overall energy-efficient conversion process. The reduction of CO 2 in presence of H 2 can follow reverse water-gas shift (RWGS) 19 or further proceed to methanation as shown below 20 The choice of catalyst plays a crucial part in the conversion of CO 2 as there are clear evidences that the reaction pathway may follow methane formation via intermediate CO generation, or in some cases, CO generation can be completely avoided. 14,19,23 It is also clear that single-atom sites behave differently than the clusters of the same atoms by promoting CO formation in presence of hydrogen. 14 Herein, we investigate the promotional effect of nickel on fumed-silica supported copper catalysts for CO 2 reduction in presence of H 2 . The choice of catalytic sites (Cu, and Ni) is based on their effectiveness in converting CO 2 to chemicals as evidenced in numerous reported literature. 10,19,[24][25][26][27][28] Nickel is a well-known reforming catalyst that has been used in CO 2 methanation reaction, and recent studies indicate how single atom nickel active sites prefer CO formation as compared to C H bond formation on nickel nanoparticles. 14 Copper on the other hand is known to have potential for a variety of products formation from CO 2 hydrogenation reaction, which includes alcohols, aldehydes, acids along with hydrocarbon and CO. 10,19,29 Various synthesis methods can be adopted to alloy copper with other elements to enhance a targeted product selectivity in the thermal, electrocatalytic, and photocatalytic routes. [10][11][12][13]19,30 Here, we chose to use a localized surface combustion synthesis technique as a novel method to synthesize Cu/fumed-SiO 2 and CuNi/ fumed-SiO 2 catalysts with 1 wt% loading of the metal. A small loading of the active metal allows the combustion to be localized with low energy density combustion system distributed over the inert silica support. Combustion synthesis has gained a considerable research interest for catalysts preparation over the past decade due to the simplicity of the technique that requires only a few synthesis steps for preparing reasonably large amount of catalysts. [31][32][33][34][35][36][37] Primarily combustion-based techniques have been used to synthesize unsupported materials, however, more recently a shift in the direction of supported catalyst is clearly seen. [38][39][40][41][42] The choice of fumed-SiO 2 is based on some recent reports highlighting its role in achieving a high dispersion of active sites on the surface compared to other supports. 18 Fumed SiO 2 is widely used in industries, however, studies employing it as a catalyst support are limited, though the availability of fibrous morphology resulting in large surface area has demonstrated its potential for application in catalysis. 15,18 Here we utilize fumed-SiO 2 as an inert oxide surface to perform localized combustion synthesis of Cu and CuNi active sites for CO 2 thermal reduction in a flow reactor.

| EXPERIMENTAL TECHNIQUES
A catalyst consisting 1 wt% of Cu and 1 wt% of CuNi (Cu = 0.5 wt%, Ni = 0.5 wt%) over fumed silica support was synthesized using surface combustion synthesis technique. The catalysts were synthesized on the basis of a 3 g batch, which is sufficient to perform catalytic tests and structural characterizations. A known amount of metal nitrate precursors (Cu(NO 3 ) 2 Á6H 2 O and Ni(NO 3 ) 2 Á6H 2 O both from BDH Chemicals Ltd., $98% pure) were dissolved in deionized water (DIW, 20 mL per g of SiO 2 ). Then an equivalent amount of urea was added to the solution corresponding to a fuel to oxidizer ratio (φ, defined elsewhere 34,43,44 ) of 1, and thoroughly mixed. Thereafter a pre-calculated amount of fumed-silica (from Alfa Aesar, amorphous fumed Silicon [IV] oxide) is added to the solution and the entire mixture, consisting of fumed silica, metal nitrates, and urea, was placed in a crucible and inside a tubular furnace for combustion in an inert atmosphere. The tubular furnace was set at 400 C and a constant flow of 30 sccm of N 2 was maintained for 1 hour to complete the combustion resulting in the synthesis of metal supported on fumed-SiO 2 . Thereafter the as-synthesized powders were collected, hand-ground and pelletized using a hydraulicpress to obtain uniform pellets with 0.6-1 mm size for assessing the catalytic properties. The Weisz-Prater criteria were used to confirm that the reaction is not diffusion-limited. [45][46][47][48] The details of calculations are provided in the supporting information. The supported catalysts were used directly for catalytic properties evaluation, without any activation/reduction in hydrogen as the synthesis conditions were maintained to get metallic/bimetallic form of copper and nickel.
The carbon dioxide conversion reaction was completed in a tubular flow reactor, BTRS-JR reactor system from Autoclave Engineers with the inner diameter of the reactor being 0.8 cm. A total of 1 g of pelletized supported catalyst was loaded inside the reactor, and a 70 sccm (CO 2 = 5 sccm, H 2 = 20 sccm, He = 45 sccm) flow rate of the reaction mixture, CO 2 :H 2 of 1:4, was maintained while raising the temperature from 50 C to 650 C and maintaining a total pressure of 1 atm. These conditions result in a catalytic bed height of 3.95 cm. The reactor effluent is passed through a condenser unit, where water and other condensable products are removed, and the gases are sent to a gas chromatograph (GC, Agilent Technologies 7890A) for further analysis. The GC consisted of a Carboxen 1010 PLOT column for the separation of expected gases (such as CO, CO 2 , H 2 , CH 4 ) and a thermal conductivity detector (TCD) for a quantitative analysis of the gaseous stream. The stability of the catalysts was investigated at 650 C for $42 hours at ambient pressure. The following definitions were used in our calculation of CO 2 conversion and CH 4 and CO yield (or conversion of CO 2 to CH 4 and CO) respectively: CO yield : The activation energy calculations were performed at low conversion (<15%) for both catalysts. The reaction rate data were obtained at three different temperature values while maintaining the same flow rate and concentration of the reactants. The reaction rates were calculated by multiplying the CO 2 molar inflow by CO 2 conversion divided by the moles of the metal content in the catalyst to get the reaction rate in [moles of CO 2 converted]/h/[moles of metal]. Arrhenius plot was obtained by plotting the logarithmic value of reaction rate with the inverse of temperature (in Kelvin) and fitting a straight line. The slope of the straight line was multiplied by the universal gas constant to get the activation energy values, as in Equation (6).
where R represents the reaction rate, E act is the activation energy, R 0 is the universal gas constant, k 0 is the preexponential factor and f(C) represents the concentrationdependent term in the rate law. The fresh catalyst after synthesis, as well as the used catalysts after the stability tests was characterized for their structural and physio-chemical properties. The details of experimental conditions used during the characterization are discussed elsewhere, [23][24][25][26][27][28][29]49 and only a brief summary is provided here. PANalytical EMPYREAN X-ray diffractometer (XRD) was utilized to perform the XRD analysis of the samples within a 2θ range of 10 to 90 as well as to perform the low angle XRD within 2θ < 5 . AimSizer machine (model AS-3012) was employed to measure the total BET surface area and distribution of pore size using liquid nitrogen to provide isothermal conditions. The morphological analysis was conducted using a field emission scanning electron microscope (FESEM, Philips, XL-30) equipped with an energy dispersive X-ray spectrometer (EDX). The composition of the as-synthesized catalysts was investigated using inductively coupled plasma mass spectroscopy (ICP-MS, using NexION 300D, PerkinElmer, USA), whereas a transmission electron microscope (TEM, FEI Tecnai G2 TF20) was used for further morphological analysis.
The catalysts were assessed for their reducibility by temperature-programmed reduction (TPR) technique with H 2 (50 sccm at STP of 10% H 2 in Ar with temperature ramp of 10 C/min, approx. 0.043 g catalyst) using Micromeritics AutoChem 2950 system. Prior to the TPR experiments, the catalysts were treated at 750 C in the open air for 3 hours. The experimentally obtained TPR data were further treated using OriginLab software to correct the baseline and get the quantitative information. The hydrogen uptake in TPR profile is calibration using a standard Ag 2 O reference sample. In addition, in situ FTIR (in DRIFTS mode) studies were conducted over the catalysts surface in presence of the reactive gases using Thermo-Nicolet 6700 FTIR between 50 C and 450 C. A praying-mantis optical assembly and a DRIFTS reaction cell (both accessories from Harrick-Scientific) were used to focus the IR beam and conduct the catalytic reaction, respectively. For DRIFTS analysis, the supported catalysts were hand-ground and screened to get smaller particles with size <106 μm, which were placed in the catalyst holder and gently pressed to obtain a visually flat surface. The reactor was heated in presence of N 2 to collect background spectra at different temperatures and then cooled down to 50 C to start the DRIFTS experiments. The reaction gas mixture (CO 2 + H 2 ) was introduced and the temperature was gradually raised from 50 C to 450 C and the reaction spectra were recorded with temperature to formulate a reaction mechanism explaining the observed products.

| RESULTS AND DISCUSSION
The catalysts were synthesized with a fuel/oxidizer ratio of φ = 1 inside a tubular furnace in an inert atmosphere to ensure the synthesis of metallic phases rather than oxides. The synthesized catalysts as well as the used catalysts after the reaction were analyzed using XRD to identify the synthesized phases as shown in Figure 1. In case of Cu/fumed-SiO 2 synthesis, the crystalline peaks are not clearly visible except for a broad peak associated with fumed silica at 21.6 . This could be because of low loading of just 1 wt% of Cu on the support, however, zooming in around 2θ value of 40 to 70 indicates a small peak around 43.2 corresponding to the (111) crystal plane of Cu (JCPDS# 04-0836) in both the fresh and used catalysts. There is no apparent change in the XRD structure of the fresh and used Cu/fumed-SiO 2 catalyst which indicates the robustness of the catalyst and stable performance for CO 2 conversion reaction, though the low loading of 1 wt% makes it difficult to notice any structural changes if they exist. The XRD profile for CuNi/fumed-SiO 2 ( Figure 1B) is much clear with evident peaks at $43.6 and $50.9 , closely match with the (111) and (200) crystal planes of CuNi bimetallic phase (JCPDS# 65-9048). A broad peak at 21.6 corresponding to fumed-SiO 2 is also present as in case of Figure 1A for Cu/fumed-SiO 2 samples. There are no visible changes in the crystal structure of the used catalyst, which points to the stable nature of CuNi/fumed-SiO 2 catalyst for the CO 2 reduction reaction. Small angle XRD (SAXRD) results shown in Figure S1 also indicate a very similar profile for both the samples before and after the reaction. Slight deflections are noticeable at 2θ values of 0.9 and 1.8 aligning with the (100) and (200) planes of hexagonal SiO 2 structure. [50][51][52] The composition of the catalyst after synthesis was further identified using ICP-MS analysis. The Cu/SiO 2 catalyst indicated 0.014 g of copper per gram of sample, whereas CuNi/SiO 2 indicated 0.007 g Cu and 0.007 g Ni per gram of the sample, respectively. So the total amount of metal content in both the samples is approx. 1.4 wt% as compared to the targeted 1 wt%. This change perhaps could be due to the loss of some SiO 2 during combustion process, as it is anticipated that some of the particles will fall out of the crucible during the sudden exothermic combustion reaction that releases a significant amount of gaseous products.
The SEM images of as-synthesized as well as the used samples are provided in Figure 2. Looking at the fresh catalysts in Figure 2 (A, Cu/fumed-SiO 2 and C, CuNi/ fumed-SiO 2 ), it is clear that both the catalysts have similar and highly porous structures. This could be due to the dominating effect of SiO 2 which constitutes 99 wt% of the total sample. Fumed SiO 2 is known for its fibrous and porous characteristics that make it suitable for various industrial applications. As it is synthesized using flame combustion of silicon tetrachloride, 53 a finely dispersed and fluffy structure emerges as compared to the coprecipitated silica, which has been utilized recently in synthesizing highly dispersed active sites for different catalytic applications. 15,18 The SEM images of used catalysts after the reaction show no visible changes to the structure as shown in Figure 2B,D for Cu/fumed SiO 2 and CuNi/fumed-SiO 2 respectively. The highly porous structure is still maintained as in the fresh catalysts. TEM analysis was conducted on the synthesized samples to further analyze the structure. As shown in Figure 2E,F, both the samples show highly porous and similar structures, presumably due to high concentration of SiO 2 phase. These observations are in coherence with the XRD results discussed earlier in Figure 1, where no apparent F I G U R E 1 XRD pattern of fresh and spent catalysts; (A) Cu/fumed-SiO 2 and (B) CuNi/fumed-SiO 2 change to the crystalline structure is observed. XRD and SEM analysis together indicate that the reaction environment with temperature up to 650 C does not create any adverse impact on the catalyst, which becomes evident in our analysis of the CO 2 conversion reaction. The EDX elemental analysis of the samples was also conducted before and after the reaction as presented in Table S1. As shown, there is not much change in the composition after the reaction. No carbon was identified in any of the samples before and after the reaction.
The two catalysts were evaluated for their performance for CO 2 conversion reaction in a tubular flow reactor. The details of reaction conditions are provided in the experimental section. Figure 3A shows the conversion of carbon dioxide as a function of temperature over a range of 50 C to 650 C. Both the catalysts show activity for CO 2 conversion reaction with CuNi/fumed-SiO 2 being more active than Cu/fumed-SiO 2 , throughout the temperature range. Both the catalysts start to show an exponential increase in conversion from temperature around 250 C, with 50% conversion temperature, T 50 , being 550 C and 607 C for CuNi/fumed-SiO 2 and Cu/fumed-SiO 2 respectively, with the former showing approx. 57 C lower temperature to achieve 50% conversion of CO 2 . Their conversion seems to merge at higher temperature of around 650 C with values being 61% and 60% for CuNi/fumed-SiO 2 and Cu/fumed-SiO 2 respectively, close to the thermodynamic equilibrium conversion. 54 Figure 3B shows the product yield for the two catalysts, with Cu/fumed-SiO 2 being exclusively selective for CO and no trace of CH 4 is seen up to 650 C temperature. CuNi/fumed-SiO 2 on the other hand shows a good selectivity for CO, however between 400 C and 650 C methane formation is also observed, nonetheless CO yield is always higher than CH 4 throughout the entire temperature range of this study. Methane formation starts around 400 C and reaches a maximum yield of 8.5% at 500 C and thereafter gradually decreases to zero at 650 C. The addition of Ni helps in methanation process by promoting the C H bond formation as it is a well-known catalyst for CO and CO 2 methanation reaction. 15,25,55,56 Thermodynamically, a higher conversion of CO 2 at low temperatures (50 C) is possible, with selectivity towards more stable hydrocarbons (eg, CH 4 ) rather than CO, which is selective at higher temperatures only, 54,57 however, in presence of catalysts we do not see any methane formation at a temperature below 400 C and only CO as a product is observed. It should be noted that the thermodynamic calculations are performed under simplified assumptions such as homogeneous phases, and gaseous compounds following ideal gas laws, etc. The catalytic reaction environment could be significantly different involving multiple phases and interaction among nonideal gases, leading to product distribution different from projected thermodynamic calculations. The catalysts were further evaluated for their stability towards CO 2 conversion reaction at 650 C for approx. 42 hours TOS. As shown in Figure 3C, both Cu/fumed-SiO 2 and CuNi/ fumed-SiO 2 showed stable performance without any deactivation. The XRD and SEM images provided in Figures 1B and 2B,D belong to the used catalysts after the stability analysis; and the structure remains intact without showing any sign of agglomeration or deformation in the porous morphology.
The rate of reaction and activation energies were calculated at low conversion conditions (<15%) for the two catalysts. The parameters used for the calculation of activation energy (eg, temperature, conversion, and corresponding reaction rates) are provided in Table 1. As shown in Figure 4, the addition of Ni helps in decreasing the activation barrier and thereby accelerates the reaction at a relatively lower temperature. The activation energy for Cu/fumed-SiO 2 was calculated to be 54.1 kJ/mol, whereas for CuNi/fumed-SiO 2 , the value was found to be only 33.8 kJ/mol. From the reaction rate calculations at low CO 2 conversion values, it is clear that the addition of Ni helps almost double the rate of reaction. As shown in Table 1, the reaction rate values are 4:4 mol CO 2 =h=mol metal and 10:7 mol CO 2 =h=mol metal at 300 C and 350 C for CuNi/fumed-SiO 2 as compared to 1:98 mol CO 2 =h=mol metal and 5:91 mol CO 2 =h=mol metal for Cu/fumed-SiO 2 for the same temperature values.
TPR studies were used to obtain the reducibility profiles of the catalysts, and to understand the impact of Ni addition on the reducibility characteristics of Cu/fumed-SiO 2 catalyst. As the results indicate in Figure 5, Cu/fumed-SiO 2 shows a broad peak at 249 C for the reduction of copper oxides to copper. The peak was further deconvoluted resulting in a combination of two peaks, which correspond to a high-intensity peak at 249 C and a low-intensity peak at 322 C. The lower temperature peak is related to finely dispersed copper oxide phase, whereas the hightemperature peak is normally linked to the bulk nanoparticle phases or copper oxides confined in micro-pores of silica that are difficult to reduce. [58][59][60] As the intensity trend indicates, Cu/fumed-SiO 2 primarily consists of highly dispersed copper phases that form the active sites responsible for catalytic conversion of CO 2 to CO.  Figure 5B shows the TPR result of CuNi/fumed-SiO 2 to illustrate the effect of nickel incorporation in the catalyst. A deconvolution of the TPR profile shows the presence of three different peaks at 249 C, 287 C, and 313 C, with gradually decreasing intensity. The H 2 -consumption corresponding to 249 C has clearly increased significantly in presence of Ni, which indicates a better dispersion of the copper sites. According to the literature, the lowtemperature peaks indicate well dispersed active sites that could be composed of oxides of copper and nickel, whereas the high-temperature peak represents sites that are difficult to reduce. [61][62][63] It is well known that nickel improves the reducibility and dispersion of copper, 62,63 and interfacing copper with silica is also reported to enhance catalytic properties of copper. 64 These two factors are combined here to generate a highly dispersed active phase. Addition of nickel results in a 4-fold increase in H 2 update with values of 0.89 cc/g ($40 μmol/g) and 3.516 cc/g ($157 μmol/g) for Cu/fumed-SiO 2 and CuNi/fumed-SiO 2 respectively. Assuming 1 mole hydrogen uptake to be equivalent to 1 mole of the metal (Cu or Ni), the dispersion of metals can be calculated to be 25.2% and 95.8% respectively for the two catalysts, indicating approximately four times increase in the metal dispersion in presence of nickel. The higher dispersion and lower activation energy justify the improvement in the catalytic activity of CuNi/fumed-SiO 2 as observed in Figure 3. In addition to high metal dispersion, the CuNi/fumed-SiO 2 catalyst also displayed a slightly higher BET surface area compared to the Cu/fumed-SiO 2 catalyst, with values being 574.9 m 2 /g and 499.2 m 2 /g for the two catalysts respectively. The pore volumes on the other hand were very similar with 3.45 cc/g and 3.21 cc/g for CuNi/fumed-SiO 2 and Cu/fumed-SiO 2 catalysts. Nature of the active sites, high metal dispersion as well as improved BET area could have all contributed positively in improving the CO 2 conversion activity of CuNi/fumed-SiO 2 catalyst. The CO 2 activity of our catalysts are compared with selected Cu and Ni based catalysts as presented in Table 2. Figure 6 shows the results of DRIFT studies on Cu/fumed-SO 2 and CuNi/fumed-SiO 2 catalysts. A schematic of the reaction pathway on the two catalysts is also provided in Figure 7. The two sharp peaks around 2360 cm −1 , present in both the catalysts at all the investigated temperature values belong to the CO 2 molecule adsorbed on the active sites. The peak around 3740 to 3750 cm −1 is stronger at lower temperature and gradually decreases with an increase in temperature, correspond to O H bond. The presence of peaks between 2850 and 2970 cm −1 confirms the formation of formate species, with peaks related to C H stretching and bending modes. [65][66][67] As these peaks are present in both Cu/fumed-SiO 2 and CuNi/fumed-SiO 2 , even though at low intensities, the catalyst surface is active at 50 C itself. The peak around 1650 cm −1 is also associated with the CO stretching in formates. 15 For temperature ≤100 C, both the catalysts show the presence of carbonate species (between 1000 and 1670 cm −1 ) attached to the surface that disappear as the temperature is increased. 68 the CuNi/fumed-SiO 2 surface, a weak peak around 3020 cm −1 emerges at 450 C, which provides support for the formation of CH 4 , possibly by successive associative reaction between adsorbed CO and adsorbed H, that is followed by the release of H 2 O molecule as shown in Figure 7. Similar reaction pathways have been reported by other authors on Cu and Ni supported catalysts. 28,71 The formation of formate/carbonate intermediates are reported to precede CO formation, however, a direct splitting of the adsorbed CO 2 * to adsorbed CO and adsorbed O cannot be completely ruled out. 28 The formate and carbonate intermediates have been reported to temperatures as low as 30 C to 150 C on Cu/CeO 2 and Cu/SiO 2 catalysts. 71

| CONCLUSION
The effect of nickel incorporation on catalytic properties of Cu/fumed-SiO 2 catalyst for CO 2 hydrogenation reaction was investigated. The catalysts were synthesized (1 wt% Cu/fumed-SiO 2 and 1 wt% CuNi/fumed-SiO 2 [Cu = 0.5 wt%, Ni = 0.5 wt%]) using a localized surface combustion synthesis technique, where a low concentration of metal precursors in presence of an active fuel ensures the combustion to take place on the surface of fumed-SiO 2 support. Both the catalysts were active and stable for CO 2 hydrogenation reaction, with CuNi/ fumed-SiO 2 showing more activity than Cu/fumed-SiO 2 .
Both the catalysts showed selectivity for CO at lower temperature, and some CH 4 production was also observed in presence of nickel. The TPR results indicate a higher dispersion of the metals in presence of nickel. The metal dispersion increased almost four times when nickel was introduced leading to the availability of a large number