NiFe(CoFe)/silica and NiFe(CoFe)/alumina nanocomposites for the catalytic hydrogenation of CO2

The fumed SiO2 and Al2O3 oxides with a specific surface area of about 80 m2 g–1 were used for the synthesis of Ni(80)Fe(20)/SiO2, Co(93)Fe(7)/SiO2, Ni(80)Fe(20)/Al2O3 and Co(93)Fe(7)/Al2O3 nanocomposites, and numbers between brackets indicate the metal content in wt%, being 10 wt% of the mass of catalysts. Catalytically active bimetallic compositions (NiFe and CoFe) that modified the fumed oxides’ surface were prepared using the solvate-stimulated method with subsequent thermal decomposition and reduction of the metal oxides to corresponding metals with hydrogen. The catalysts were characterized using the TGA in dynamic hydrogen, nitrogen physisorption, and PXRD methods. The complete conversion of carbon dioxide is observed in the temperature range of 350–425 °C at the maximum methane yield of 72–84%. The long-time catalytic test demonstrates the high stability of the catalyst during 5 weeks of exposure to the reaction mixture. The yield of methane was decreased by 3–14% after 1–2 months of long-time testing.


Introduction
The environment's temperature rises and other accompanying atmospheric processes are stimulated by the increased anthropogenic greenhouse gas (GHG) emissions (Ronsch et al. 2016).Climate change caused by global warming is becoming intensive, resulting in sea-level rise, coastal lowland inundation, glacier melting, extreme weather, species extinction, and other environmental problems (Ma et al. 2021).To prevent global warming, GHG emissions should be reduced (Mohsin et al. 2019).
It could be done, for example, by reducing the number of fossil fuels (for example, oil and coal) in operation and using the null-carbon and other greener and cleaner technologies for mitigating the consequences of climate change to allow adaptation (Embong et al. 2021).Global climate change harms the environment and ecological systems.Today's agricultural production has declined, and water resources showed increasing shortages (Nema et al. 2012).Typically, the heating of the atmosphere on a couple of grads contributes to ice melting, causes insufficient harvesting, lack of drinking water, and indirectly impacts human health (Meehl 2005).According to Olivier et al. (2020), carbon dioxide (CO 2 ) remains the main component of GHGs available in the atmosphere.Therefore, the main attention of engineers and researchers is paid to finding ways to reduce its emissions.To date, there are two main approaches to emission reduction to capture CO 2 in geological formations to reduce its amount or convert CO 2 into low-carbon fuels (Aziz et al. 2015;Devarajan et al. 2022).The "energy to gas" (E-G) and "energy to liquid" (E-L) technologies allow chemical storage of excess energy from regenerative sources by the reaction of renewable hydrogen (H 2 ) with carbon monoxide (CO) to produce methane (CH 4 ) and other energetic gas products (E-G) or liquid fuels (E-L) (Chen et al. 2009;Weitzel and Glock 2018;Saeidi et al. 2014).In focused interest is the concept of recovered CO 2 emission and fuel synthesis with the Fischer-Tropsch reaction, and such an approach can significantly reduce carbon emissions (Arakawa et al. 2001).The CO 2 methanation reaction has a number of advantages over other chemical processes, as CH 4 gas can be directly piped by existing natural gas pipelines.Besides, it can be a fuel or feedstock for the production of miscellaneous chemicals.On the other hand, this reaction produces the target product under atmospheric pressure.Thus, the formation of CH 4 from CO 2 at low temperatures became an important task to converse the CO 2 captured from industrial exhausts, although the conversion was very low at first (Beuls et al. 2012).
The CO 2 methanation with hydrogen (H 2 ) gas produced from renewable sources (CO 2 + 4H 2 ↔ CH 4 + 2H 2 O; ΔH 298 K = −252.9kJ mol -1 ) is of fundamental scientific interest.This reaction is thermodynamically favorable and is much faster than other CO 2 reduction reactions producing hydrocarbons or alcohols (Gao et al. 2012).Catalytic CO 2 methanation is an important process with potential commercial applications (Miao et al. 2016;Frontera et al. 2017).Nickel-based metal catalysts are typical low-cost metal catalysts producing CH 4 efficiently (Delmelle et al. 2016;Muroyama et al. 2016;Ashok et al. 2020).These catalysts are frequently engaged to remove CO 2 impurities within the industrial cycle of ammonia synthesis (Jürgensen et al. 2015).Not only Ni, but other transition metals are catalytically active in the CO 2 methanation reaction (Zhludenko et al. 2018;Ghadamgahi and Rahmani 2020;Ischenko et al. 2020Ischenko et al. , 2021)).The unsupported transition metal catalysts show an increase in the CO 2 conversion as Ru > Ir > Rh > Ni, while the selectivity order varies as Ni > Co > Fe > Ru (Roy and Peter 2020).Increasing the thermal stability of Ni and Co catalysts can be achieved by introducing the Fe component (Hwang et al. 2012;2013).This component supplies high thermal stability to composite, contrasting to the effect of Ni and Co metals (Ren et al. 2015).The carrier is an important fraction in the catalyst composition, which can participate in the CO and CO 2 methanation reactions directly or indirectly through interaction with Ni or Co (Wang et al. 2018;Ferreira et al. 2019;Dyachenko et al. 2021).This loading prevents a rapid deactivation of the catalyst and significantly reduces the catalyst's acting mass (Kulyk et al. 2010;Linnik et al. 2013;Lesnyak et al. 2007).
Nanocomposites (NCs) based on silica or alumina with NiFe or CoFe bimetallic nanoparticles (NPs) are prominent CO 2 hydrogenation catalysts (Dyachenko et al. 2022a).But, we observed that high specific surface areas of the used silica or alumina, at the level of 200-300 m 2 g −1 , have a negative impact on the catalytic performance of NiFe or CoFe bimetallic NPs.High specific surface area could be a reason for an uneven surface distribution of active mass components.Some of the distributed components cannot contribute to the formation of catalytically active sites and the high conversion rate of CO 2 to CH 4 cannot be reached (Dyachenko et al. 2022a, b).Therefore, the aim of the present study is to examine the CO 2 methanation performance over NiFe(CoFe)/Silica and NiFe(CoFe)/Alumina NCs with a lower specific surface area of ~ 70-80 m 2 g -1 .

Synthesis of nanocomposites
Fumed oxides, SiO 2 (Si60) and Al 2 O 3 (pilot plant of Chuiko Institute of Surface Chemistry, Kalush, Ukraine) with the specific surface area (S BET ) of 80 and 72 m 2 g -1 , respectively, were used as sorbents to adsorb transition metal salts for the preparation of bimetallic nanocomposites (NCs).High water-soluble metal salts, Ni(NO 3 ) 2 •6H 2 O, Co(NO 3 ) 2 •6H 2 O, and Fe(CHO 2 ) 3 •2H 2 O, were purchased from Khimlabor Reactive Ltd. (Brovary, Ukraine) and used as modifying agents in the process of sorbent surface treatment.
The method of obtaining bimetallic nanocomposite catalysts based on silica and alumina involved of three stages.In the first step, the method of solvate-stimulated modification was applied for the uniform distribution of appropriate metals ions according to the routine described by Goncharuk et al. (2019).The metals ratios in the active masses were Ni(80)Fe(20) and Co(93)Fe(07), where numbers between brackets are the metal content in wt% (Zhludenko et al. 2018;Meshkini-Far et al. 2018a).To obtain NCs with 10 wt% metal contents, the required mass of salts was weighted, and then the salts were dissolved in 10 ml of water.When 20 g of silica or alumina was placed in a ceramic ball mill, the water solution of salts was gradually added, and after that the resulting mixture was homogenized for 90 min.After achieving homogeneity, the resulting mixture was air-dried for 24 h.In the second stage, the dried powder was calcined in a muffle furnace at 600 °C for 1 h.This treatment leads to the thermal decomposition of adsorbed salts and their transformation into transition metal oxides.The last stage is the reduction of the NCs being in oxide form in a dynamic equivolume hydrogen-helium (H 2 /He) gas mixture for 2 h at atmospheric pressure.The heating was performed at the optimal reduction temperature determined from the data of thermogravimetric analysis (TGA) in dynamic hydrogen gas.The obtained reduction products are metallic Ni, Co, and Fe NPs, which are uniformly distributed on the surface of fumed silica or alumina.

Catalytic tests
The catalytic test of prepared NCs was provided in a flowbed quartz microreactor (inner diameter: 8 mm) under atmospheric pressure, as described by Dyachenko et al. (2022a).The volume of the catalyst package was 1 cm 3 or 0.6-0.7 g.During the chromatographic analysis, the reaction gas mixture of CO 2 (2 vol%) and H 2 (55 vol%) in He (43 vol%) was used at a gas flow rate of 100 ml min -1 .The gas mixture from the reactor outlet was analyzed for the content of CO 2 , CO, and CH 4 using a gas chromatograph "Shimadzu 2014" equipped with a thermal conductivity detector and a separation column filled with 5A molecular sieves.
Catalytic activity was studied using two consecutive temperature rises from 20 to 450 °C.In the process of the first raise, the catalyst pre-treatment occurs, involving the preliminary formation of catalytically active sites on the surface (Yatsimirskii et al. 2005a).The conversion of CO 2 and the amount of the formed product was registered during the second temperature rise.The reaction temperature inside the catalyst bed was monitored with covered thermocouples (Yatsimirskii et al. 2005b;Veselovskyi et al. 2012).The efficiency of the catalyst was evaluated by the temperature at the maximum conversion of CO 2 into CH 4 .The CO 2 conversion ( X CO 2 ) and CH 4 yield ( Y CH 4 ) were determined by means of the following equations (Meshkini Far et al. 2018b;Beierlein et al. 2019): where CO 2 out , and CH 4 out are the volume concentrations of CO 2 and CH 4 gases, respectively, in the gas mixtures sampled at the reactor outlet.
The catalytic performance of the prepared catalysts was also studied over time.The experiment for each catalyst was carried out for 30-50 days in a flow reactor that operated in the following mode: 8 h of work in the CO 2 methanation process in the temperature range of 350-450 °C, then the catalyst was kept in the reaction mixture at room temperature for 1 week.The cycle was repeated 3-4 times according to the procedure that is described above. (1)

Methods
Thermogravimetry analysis (TGA ) in dynamic hydrogen gas was used to study the weight changes during the reduction of transition metal oxides in the oxide NCs to their metallic state.For this purpose, the studied samples were heated in a dynamic flow of equivolume argon-hydrogen gas mixture.The heating was carried out in the temperature range of 30-550 °C at a rate of 10 °C/min.With an optimal reduction temperature found from the TGA experiments, one can avoid thermal decomposition and sintering of the active surface centers during the preparation of NC catalysts at the reduction stage.

Powder X-ray diffraction (PXRD) analysis
The reduced samples and post-reaction NC catalysts were investigated by the PXRD method using a DRON-UM1 diffractometer with monochromatic CuК α (λ = 1.5418Å) and CoК α (λ = 1.7902Å) radiations in the 2θ range from 10 to 90°.The Match! analytical software (Crystal Impact, Germany) was used for the phase identification procedure.
The half-widths of the fitted PXRD patterns were taken for the calculation of the crystallite size (Julian R.H.Ross 2019).
Nitrogen adsorption-desorption isotherms were measured on an automatic surface area and porosity analyzer ASAP 2020 at liquid nitrogen temperature (Micromeritics Instrument Corp., USA).Before adsorption measurements, the samples were outgassed at 110 °C for 2 h in a vacuum chamber.The standard Brunauer-Emmett-Teller (BET) method (Rouquerol et al. 2013) was used to calculate the specific surface area (S BET ).The total pore volume V p was evaluated by converting the volume of adsorbed nitrogen measured over the relative equilibrium adsorption pressure (p/p 0 ) to the volume of liquid nitrogen per gram of adsorbent.The calculation was carried out taking the ratio p/p 0 = 0.98-0.99,where (p/p 0 ) is the nitrogen gas vapor pressure in the system, and p 0 is the saturation pressure of nitrogen at 77.4 K.The procedure of a self-consistent regularization (SCR) under non-negativity condition (f ≥ 0 at any pore radius R) at a fixed regularization parameter α = 0.01 with voids (V) between spherical nonporous nanoparticles packed in random aggregates (V/SCR model) (Gun'ko 2014) were applied to estimate the pore size distributions (PSDS, differential f V ~ dV p /dR and f S ~ dS/dR) using the experimental nitrogen desorption data.

Results and discussion
TGA in dynamic hydrogen flow was conducted to determine the optimal reduction temperature for the preparation of the active metal phase of catalysts.Figure 1  For the SiO 2 sorbent, the first weight loss is observed in the temperature range of 100-200 °C (Fig. 1a).This effect corresponds to the physisorbed water release.The second weight loss effect at and above 320 °C can be assigned to the dehydroxylation of the sorbent surface.For the Ni(80) Fe(20)/SiO 2 (Fig. 1c) and Co(93)Fe(07)/SiO 2 (Fig. 1e) NCs, the weight loss measured in the temperature range of 50-250 °C corresponds to water thermodesorption.The intensive reduction of the oxide phase for these samples is observed between 300 and 500 °C.Aiming to prevent possible sintering of metal particles, the samples were reduced with hydrogen at 440 °C.
For the Al 2 O 3 sorbent, intensive weight loss is observed between 50 and 250 °C (Fig. 1b).We referred the weight loss at and above 250 °C to the surface dehydroxylation.For the Ni(80)Fe(20)/Al 2 O 3 (Fig. 1d) and Co( 93)Fe(07)/ Al 2 O 3 (Fig. 1f) NCs, the weight loss assigned to water release, is registered between 30 and 200 °C in a hydrogen atmosphere; an intense weight loss in the temperature range of 250-500 °C can be assigned to the conversion of metal oxides to metals or their alloys.An intense weight loss ranging from 300 to 375 °C (Fig. 1d, f) may also indicate the thermal decomposition of some functional surface groups.The observed weight loss is due to the release of the gaseous products of thermal decomposition.To prepare the NC catalysts, the reduction of Ni( 80)Fe(20)/Al 2 O 3 and

Nitrogen adsorption-desorption
The specific surface area and pore size parameters are important features affecting catalytic activity.The results of N 2 adsorption-desorption characterization of the synthesized NCs, alumina and silica sorbents are shown in Fig. 2 and are presented in Table 1.According to the IUPAC classification (Gregg and Sing 1982;Thommes et al. 2015), the obtained nitrogen adsorption-desorption isotherms (Fig. 2a, c, e, g) are assigned to the IUPAC type II with a hysteresis loop of H3-type indicating the textural porosity of aggregated nonporous NPs.These results present the prevailing contribution of mesopores.
The textural characteristics (Table 1) of the studied NCs and the alumina and silica show a slight decrease in the S BET value in the sequence sorbent → oxide NC → reduced NC.This decrease is mainly due to the contribution of mesopores, which prevail for the alumina and silica sorbent and for the synthesized NCs in the oxide and reduced forms.The IPSD data shown in Fig. 2b, d, f, h demonstrate an insignificant contribution of micropores to the total porosity.The latter remains unchanged after modifying the alumina or silica sorbents by transition metal oxides and their subsequent transformation into metals.
Decreasing the contribution of mesopores in the total porosity is 4-6% for the SiO 2 -based NCs and 10-12% for the Al 2 O 3 -based NCs in the oxide forms, considering the porosity of sorbent.The reduction process of NCs taken in the oxide form to the corresponding metal forms causes a further decrease in the mesoporosity by 4-7% for the NiFe/ SiO 2 , CoFe/SiO 2 , and NiFe/Al 2 O 3 NCs.The exception is the CoFe/Al 2 O 3 NC.This sorbent shows a slight increase in the values of S BET , V p , R p , and V.But only those values related to S meso and V meso .These changes in texture characteristics after sorbent modifying and forming oxide NC and further to the metal NC catalyst may be due to compacting of the sorbent surface.We believe that the formed metallic particles (NiFe and CoFe) are localized at the boundary of some pseudo-pores.The latter are voids between close spherical particles (Khavryuchenko et al. 2011).The specific surface area and pore radius are reducing, and this observation is a sign of such an arrangement.
Usually, after the loading of metal oxide nanoparticles on the surface of the carrier, an increase in the total pore volume is observed for all studied NCs.For the SiO 2 -based NCs, the increase in V p is due to the contribution of both V meso and V macro .For the Al 2 O 3 -based NCs, the total volume V p increases only due to V meso , while macropores do not undergo significant transformation.The pore volume for all composites remains practically unchanged after subsequent treatment of the nanoxide composites in a reducing atmosphere and their operation in the catalytic process.The effect of increasing the pore volume as a result of fixing the guest oxide phase on the surface of the carrier matrix can have several explanations.First, in the process of solvate-stimulated modification, there is a change of packing between the aggregated particles of the carrier itself, which leads to the redistribution of pores.Secondly, the oxide phase may have a mesoporous structure, which contributes to the volume of mesopores.
For the NiFe/Si60 NC (Fig. 3a) and for the NiFe/Al 2 O 3 NC in its reduced form (Fig. 3c), the diffraction reflections of metallic Ni and Fe have the same positions (44.6° and 51.5°).Most likely, a solid solution is formed, where some Ni atoms in the crystal lattice are replaced by Fe atoms.Diffraction of alumina complicates the interpretation of PXRD patterns and the determination of individual crystalline phases in the NC compositions.However, the most intense diffraction reflections corresponded to metal crystallites.Thus, for the CoFe/Al 2 O 3 sample (Fig. 3d), the diffraction reflections at 51.7° and 60.1° correspond to the metallic Co phase, while the diffraction reflection at 52.5° corresponds to the Fe phase.Table 2 shows  Figure 4 shows the CO 2 conversion and yield of CH 4 against temperature over the bimetallic NC catalysts with NiFe and CoFe active mass localized on the surface of highly dispersed oxides.
In general, all NC catalysts activate the CO 2 methanation reaction starting from a temperature of 200 °C, after which the fraction of converted CO 2 increases sharply and reaches a maximum value in the range of 350-450 °C (Fig. 4a).The increase in the yield of CH 4 is more gradual.In the presence of NC catalysts, one can see a steep increase in the yield of CH 4 with a temperature rise reaching the saturation at about 300-325 °C. Figure 4  The high efficiency of the catalysis passage with respect to the CH 4 product is shown in the presence of the NiFe/ SiO 2 NC catalyst (Fig. 4a, b).The CO 2 conversion starts at 225 °C and gradually increases to a maximum at 375 °C.For the CoFe/SiO 2 NC catalyst, the near total conversion of CO 2 is achieved at 400 °C (Y(CH 4 ) = 70%), and the maximum  temperature at the highest CO 2 conversion level is 450 °C (Y(CH 4 ) = 72%), see Fig. 4b.For the NiFe/Al 2 O 3 NC catalyst, a sharp increase in the CO 2 conversion is observed at 225 °C (Fig. 4).The CO 2 conversion curve has a characteristic sharp rise and saturation at about 300 °C.The near total conversion of CO 2 can be achieved at 350 °C.At the same time, the Y(CH 4 ) in the presence of the NiFe/Al 2 O 3 catalyst is 76% at 350 °C and gradually increases to 84% when the temperature rises to 450 °C (Fig. 4b).The near total conversion of CO 2 in the presence of the CoFe/Al 2 O 3 NC catalyst is achieved at 425 °C.
To compare the results of the studies presented in this work, Table 3 shows the catalytic activity of NC catalysts synthesized on the basis of highly dispersed oxides with a specific surface area of 270 m 2 g -1 .For these NC catalysts, the near total CO 2 conversion is achieved already at 325-350 °C, which is certainly an important factor for determining the activity of the catalyst.However, a more valuable indicator of catalyst efficiency is the Y(CH 4 ) rate.For the NCs based on SiO 2 with S BET = 80 m 2 g -1 , the Y(CH 4 ) value increases from 76 to 83% in the temperature range of the near total CO 2 conversion.The same trend is observed for alumina-based NCs studied in the presented research (Table 3).For higher dispersed NCs, the increase in the Y(CH 4 ) value is insignificant.Notably, for some NC catalysts, the Y(CH 4 ) increases with increasing temperature at a steady CO 2 conversion, which is an interesting effect.
We studied catalytic performances over time to evaluate the operation stability of the NC catalysts.The diagrams presented in Fig. 5 show the change in the catalytic behavior of the studied NC catalysts during long-term exposure in the reaction mixture at the maximum CO 2 conversion temperature of 350-450 °C.The data show thatthe maximal Y(CH 4 ) value, which was 84% for the NiFe/SiO 2 NC (Fig. 5a), 72% for CoFe/SiO 2 NC (Fig. 5b), 83% for NiFe/ Al 2 O 3 NC (Fig. 5c), and 77% for CoFe/Al 2 O 3 NC (Fig. 5d) in the first days of the experiment, decrease over time.This observation indicates time-dependent deterioration of the catalyst action.Typically, the bimetallic NiFe/Al 2 O 3 and CoFe/Al 2 O 3 NC catalysts show a decrease in the Y(CH 4 ) value after 1 or 2 months of catalytic tests.The same effect Fig. 5 Diagrams showed the change in the catalytic CO 2 performance at the temperature at the near total conversion of CO 2 over time was noted for the NiFe/SiO 2 and CoFe/SiO 2 NC catalysts.This decrease in the Y(CH 4 ) value is varied from 3 to 14%.
According to the results of catalytic activity, the studied NC catalysts can be arranged by the Y(CH 4 ) changes in descending order as NiFe/Al 2 O 3 (84%) > NiFe/Si60 (83%) > CoFe/Al 2 O 3 (77%) > CoFe/Si60 (72%).In sum, we proved that the specific surface areas of alumina and silica have a crucial effect on the catalytic performance of the studied NC catalysts in the CO 2 methanation.

Conclusions
The study of the CO 2 methanation reaction over the bimetallic NiFe/SiO 2 and NiFe/Al 2 O 3 NCs, and CoFe/ SiO 2 and CoFe/Al 2 O 3 NC catalysts showed the total CO 2 conversion in the temperature range of 350-420 °C while the maximum CH 4 yield is 72-84%.The CoFe/SiO 2 and CoFe/Al 2 O 3 NC catalysts showed lower efficiency than highly active NiFe/Si60 and NiFe/Al 2 O 3 NC catalysts.This experimental fact can be explained by the low Fe content, which, due to dispersion on the surface of the alumina or silica, prevents the rapid reaction passage on catalytically active centers.Therefore, in further research, this problem is planned to be solved by increasing the Fe content in the ratio between the components within the active catalytic mass or by increasing the amount of active mass in the NC while maintaining the already studied mass ratio of Co( 93)Fe( 7).However, a long-term catalytic test showed that the low Fe content in the CoFe/SiO 2 and CoFe/Al 2 O 3 NC catalysts did not affect the long-term performance of the catalyst in the CO 2 methanation reaction.
According to the obtained PXRD data, the crystalline lattice of alumina contributes to the formation of smaller crystallites for both oxide and metal NCs.The CoFe/SiO 2 and CoFe/Al 2 O 3 NCs are characterized by larger crystallites of the transition metal oxides and respective transition metals than those in the NiFe/SiO 2 and NiFe/Al 2 O 3 NCs.In the reduced NiFe/SiO 2 and NiFe/Al 2 O 3 NCs, the formation of the NiFe solid solutions of replacement type is observed.Separating the bimetallic CoFe compositions into individual metal particle phases does not affect the CH 4 production positively.On the other hand, the NiFe composition forms the solid solution that is more stable under exposure to the reaction mixture at the catalysis temperatures and, consequently, more active in catalytic CH 4 synthesis.

Fig. 3
Fig. 3 PXRD patterns of a NiFe/SiO 2 , b CoFe/SiO 2 , c NiFe/Al 2 O 3 , and d CoFe/Al 2 O 3 NCs that nanocrystals of active mass of smaller sizes are formed for the NC based on Al 2 O 3 .The CoFe/SiO 2 and CoFe/Al 2 O 3 NCs are characterized by large crystallites of metal oxides and metals compared with those found in the NiFe/SiO 2 and NiFe/Al 2 O 3 NCs.Obviously, the Ni and Fe can form a solid solution, while Co and Fe in the respective NCs can decompose into separate metallic Co and metallic Fe phases.The calculated crystallite size in the studied NCs before and after their use in the catalytic process shows the sizing of crystallites.The high-temperature treatment of transition metal oxides with hydrogen transforms bigger crystallites of nickel and cobalt oxides into smaller crystallites of respective metals.
illustrates the clear advantage of the NiFe/Al 2 O 3 and NiFe/SiO 2 NCs in catalytic methane formation compared to the CoFe/Al 2 O 3 and CoFe/SiO 2 NCs, in pairs, correspondingly.

Fig. 4 a
Fig. 4 a CO 2 conversions and b CH 4 yield over NCs versus the reactor temperature

Table 3
The parameters of the CO 2 methanation process over the bimetallic NCs catalysts