Hydrogenation of CO₂ over Biochar-Supported Catalysts

Abstract

The study investigates hydrogenation of CO₂ over mono- and bimetallic catalysts supported on biochar. In this reaction, bimetallic iron–cobalt catalysts were shown to surpass monometallic iron and cobalt catalysts in terms of catalytic performance. The optimal combination of performance parameters was reached at an iron to cobalt ratio of 3 : 1. The composition and genesis of the active phase in the bimetallic Fe–Co catalyst were identified, and the CO₂ hydrogenation mechanism was suggested for an iron-dominated bimetallic catalyst. Using biochar as a support was found to provide an active phase composition favorable for CO₂ hydrogenation.

Process solutions that reduce the climate impact of greenhouse gas emissions, including those of carbon dioxide, are of paramount importance today.

Although the production of hydrocarbons (HCs) from CO₂ has been researched since the 1970s, for a long time this process attracted very limited interest due to its poor cost effectiveness. The situation dramatically changed in the mid-2000s, when the development of process techniques for chemical recycling of carbon dioxide by converting it into value-added chemicals became a widespread trend [1–3]. For instance, increasing attention has been paid to hydrogenation of CO₂ into synthetic liquid hydrocarbons (SLH).

To date, research in this area has mostly been focused on one-step synthesis of SLH over bifunctional catalysts. Two main mechanisms for CO₂ hydrogenation have been described [4–6]. The first mechanism involves the conversion of carbon dioxide into carbon monoxide in the presence of hydrogen (the reverse water–gas shift reaction, RWGS) followed by the hydrogenation of CO into HCs:

$${\rm{C}}{{\rm{O}}_2} + {{\rm{H}}_2} \to {\rm{CO}} + {{\rm{H}}_2}{\rm{O}},$$

((1))

$$n{\rm{CO}} + (2n + 1){{\rm{H}}_2} \to {{\rm{C}}_n}{{\rm{H}}_{2n + 2}} + n{{\rm{H}}_2}{\rm{O}}.$$

((2))

The second potential mechanism involves the formation of methanol from CO₂ followed by conversion to HCs, primarily alkenes:

$${\rm{C}}{{\rm{O}}_2} + {\rm{CO}} + 5{{\rm{H}}_2} \to 2{\rm{C}}{{\rm{H}}_3}{\rm{OH}} + {{\rm{H}}_2}{\rm{O}},$$

((3))

$$2{\rm{C}}{{\rm{H}}_3}{\rm{OH}} \to {\rm{C}}{{\rm{H}}_3}{\rm{OC}}{{\rm{H}}_3} + {{\rm{H}}_2}{\rm{O}},$$

((4))

$$n{\rm{C}}{{\rm{H}}_3}{\rm{OC}}{{\rm{H}}_3} + m{\rm{C}}{{\rm{H}}_3}{\rm{OH}} \to {{\rm{C}}_{n + 2m}}{{\rm{H}}_{2n + 4m}} + (n + m){{\rm{H}}_2}{\rm{O}}.$$

((5))

Yang et al. [4] demonstrated that the first mechanism prevails in the presence of bimetallic catalysts, and that the conversion of CO₂ to HCs via methanation becomes preferential for zeolite-supported catalysts.

He et al. [5] found, for hydrogenation of CO₂ over Co–Mn catalysts, that CO was hardly detectable in the reaction products. They further identified, by a labeling test and in situ Fourier transform infrared characterization, that these catalysts promoted direct hydrogenation of CO₂ into HCs without intermediate CO formation. According to the suggested reaction mechanism, the CO₂ adsorbed on the catalyst surface was gradually reduced to CH₂/CH₃ moieties via CO₂^δ−, HCOO−, −CH₂OH, and/or CH₃O− intermediates.

Cu, Mn, alkali metals, and/or noble metals can be added as modifiers to both cobalt and iron catalysts; for example, the addition of Cu and K to an iron catalyst facilitates the formation of a metallic and/or carbide active phase, on which the chain grows [7].

Iron and cobalt catalysts represent the two major existing approaches to catalytic hydrogenation of carbon oxides because they lead to different types of active sites generated on the surface. Guo et al. [8] were the first to use bimetallic iron–cobalt catalysts, i.e., to combine Fe and Co active sites. These bimetallic catalysts were synthesized by coprecipitation or impregnation. The synergy between Fe and Co contributed to an enhancement of the CO₂ hydrogenation selectivity (up to 87% for C₂₊). The researchers believed that incorporating a metallic Co increased the CO₂ adsorption and promoted the formation of active iron carbides, thus facilitating the generation of C–C bonds on the iron carbide phase. In conventional catalysts for the hydrogenation of carbon oxides, alumina, ceria, silica, or various zeolites have generally functioned as a support and structure-directing agent [7].

Of late, carbon materials such as biochar have been increasingly used as a support in this type of catalysts. Most biochar-supported catalysts are prepared by loading a metallic phase on an activated carbon material via impregnation with corresponding metal salts. This allows the metal ions to be bound with the support surface. The active phase, i.e. metal oxides and carbides, is formed under thermal treatment of the catalyst precursor [9, 10].

Biochar-supported Co and Fe catalysts have proven to be efficient in CO hydrogenation, a reaction similar to CO₂ hydrogenation in chemistry. Kuz’min et al. [11] investigated various nanosized iron catalysts, with the active phase being represented by iron carbides (like in the case of CO₂ hydrogenation). The tested catalysts exhibited high activity and selectivity in the synthesis of C₅₊.

The purpose of the present study was to investigate bimetallic Fe–Co catalysts supported on biochar in hydrogenation of CO₂. Similar monometallic iron and cobalt catalysts were used as comparative samples.

EXPERIMENTAL

Synthesis of catalysts. The catalyst support, specifically biochar, was prepared by hydrothermal carbonization of cellulose in a 0.5 L autoclave-type steel reactor equipped with a mechanical stirrer under isothermal conditions (at 190°C) for 24 h. The carbonizate was filtered off and dried at 105°C for 24 h. The material obtained after drying was calcined in a muffle furnace at 400°C for 1 h. The active components were loaded from an aqueous alcoholic solution of their precursors, followed by thermal treatment in an inert atmosphere (400°C, 1 h). In all cases, the total amount of active components was 20 wt % based on metal.

Catalytic test. The catalytic tests were run in a flow-type catalytic setup with a fixed catalyst bed. The inlet and outlet zones of the reactor were backfilled with quartz beads to ensure efficient feed distribution, heat transfer, and mass transfer. The synthesis was carried out in a continuous mode at 2.0 MPa and a feed gas hourly space velocity (GHSV) of 500 h^–1 (CO₂/H₂ = 1 : 3 mol/mol) and in the temperature range of 240 to 320 (340)°C. The temperature was increased stepwise (by 20°C every 12 h), with relevant gaseous and liquid products being presampled for analysis.

Prior to the catalytic test, the samples were activated at 450°C, 2.0 MPa, and a H₂ GHSV of 1000 h^–1 for 3 h.

Analysis of reactants and products. The initial CO and gaseous products were analyzed using a Crystal Lux 4000M chromatograph (Meta-Chrom, Russia) equipped with a thermal conductivity detector and two columns, with helium as a carrier gas. A 3 m×3 mm column filled with CaA molecular sieve was used to separate CO, CH₄, and N₂ (internal standard) under isothermal conditions (80°C).

A 3 m×3 mm Haye Sep R packed column was used to separate CO₂ and C₂–C₄ HCs. The temperature was programmed for heating from 50 to 200°C at a rate of 8°C/min.

Liquid HCs were identified on a Crystal Lux 4000M gas chromatograph (GC) equipped with a flame ionization detector and a 50 m×0.32 mm OV-351 capillary column. The temperature was programmed as follows: 50°C (2 min); 50–260°C, 6°C/min; 260–270°C, 5°C/min; 270°C (10 min). The weight content of components (e.g., n-paraffins, iso-paraffins, and olefins) and the fractional composition (e.g., gasoline fuel, diesel fuel, and paraffins) were measured by GC.

Oxygenates in the aqueous phase were also GC-analyzed on a Crystal Lux 4000M chromatograph equipped with a flame ionization detector and a 50 m×0.32 mm capillary column filled with HP-FFAP (nitroterephthalic modified polyethylene glycol). The temperature was programmed as follows: 70°C (8 min); 70–110°C, 10°C/min; 110–220°C, 15°C/min; 220°C (10 min). Isobutanol was used as an internal standard.

To test the catalytic activity, the following parameters were evaluated: CO₂ conversion, product yield (the amount of the product obtained by passing 1 m³ of the gas mixture through the catalyst, this amount being normalized to standard conditions, g), selectivity (the percentage of CO₂ expended to produce a specific reaction product per the total amount of CO₂ entered into the reaction). The measurement error was obtained by estimating the errors for each individual component using an indirect error estimation procedure.

Physicochemical characterization of catalysts. The physicochemical properties of the catalysts were examined by X-ray diffraction (XRD) using a Rigaku Rotaflex D/MAX-RC diffractometer equipped with a rotating copper anode and a secondary graphite monochromator (with a CuK_α wavelength of 0.1542 nm) as an X-ray source (continuous θ–2θ scanning in the 2θ range of 10°–90° with a scanning speed of 2°/min and a step of 0.04°). The XRD patterns were processed using the MDI Jade 6.5 software package; the phase composition of the catalyst samples was identified using the ICDD PDF-2 diffraction database.

The X-ray beam was focused in Bragg–Brentano focusing geometry with two Soller slits. The interplanar distance was derived from the Wolf–Bragg equation:

$$2d\sin \theta = n\lambda ,$$

where d is the interplanar distance (nm); θ is the diffraction angle; n is the reflection order; and λ is the wavelength of X-rays (nm).

RESULTS AND DISCUSSION

In the first part of the study, we compared the catalytic activity of the monometallic iron and cobalt catalysts and the bimetallic Fe–Co catalyst (Fe : Co = 1 : 1). The comparative data are summarized in Tables 1 and 2.

Table 1. Yield of CO₂ hydrogenation products vs. process temperature for catalysts with different active components

Table 2. Selectivity towards CO₂ hydrogenation products vs. process temperature for catalysts with different active components

The CO₂ conversion as a function of process temperature was plotted for monometallic and bimetallic catalysts (Fig. 1). Over the entire temperature range, the highest CO₂ conversion was achieved in the presence of the cobalt catalyst; the CO₂ conversion for the iron catalyst was markedly lower. At the lower temperature end, the bimetallic equilibrium catalyst exhibited CO₂ conversion comparable to that over the iron catalyst. Under heating, however, it provided the most rapid conversion boost, to approximate the performance of the Co-based sample at 320°C (58.0% vs. 60.6%, respectively). It is further worth noting that the CO₂ conversion curve for the Co catalyst passed through a maximum (64.9% at 300°C), whereas the curves for the iron and bimetallic catalysts displayed a steady growth over the entire temperature range.

Fig. 1.

CO₂ conversion as a function of process temperature for different catalyst types.

The yields of target products over the different catalyst types largely depended on process temperature (Fig. 2). The bimetallic catalyst proved to be markedly superior to its monometallic counterparts over the entire temperature range. In the presence of the bimetallic sample, the yield of C₅₊ steadily increased with temperature, to reach 36.6 g/m³ at 320°C (Fig. 2a). However, the total yield of C₅₊ and oxygenates passed through a maximum of 49.3 g/m³ at 280°C followed by a decrease with further heating (Fig. 2b). In the case of the monometallic cobalt catalyst, no formation of oxygenates was observed; the yield of C₅₊ did not exceed 28.8 g/m³ at 280°C and declined with further temperature elevation. The monometallic iron catalyst exhibited the overall lowest activity; in addition, in its presence both the yield of C₅₊ and the total C₅₊/oxygenate yield displayed only a negligible variation over the wide range of 260–320°C.

Fig. 2.

Yield of C₅₊ (a) and total yield of C₅₊ and oxygenates (b) as functions of process temperature for different catalyst types.

The yield of methane also heavily depended on the catalyst type (Fig. 3). The most significant methane formation was observed in the presence of the Co catalyst; the Fe catalyst exhibited on average a three-fold lower methane yield at equal process temperatures. The bimetallic catalyst provided the lowest methane yield over the temperature range of 240–300°C.

Fig. 3.

Yield of methane as a function of temperature for different catalyst types.

Another important characteristic of catalyst performance in CO₂ hydrogenation is the content of CO in the reaction products (Table 1). In the case of the Co catalyst, no CO was detected in the reaction products, obviously due to its effective hydrogenation into methane. The monometallic Fe catalyst and bimetallic Fe–Co catalyst exhibited the opposite temperature effects on CO yield, specifically a direct correlation in the first case and an inverse correlation in the second.

Thus, in terms of catalytic performance the bimetallic iron–cobalt catalyst surpassed its monometallic Fe-based and Co-based counterparts. The second part of the study was intended to determine the contributions of iron and cobalt as active components to the catalytic activity in CO₂ hydrogenation. Three samples varying in the iron to cobalt molar ratio were tested: 1 : 3, 1 : 1 (equilibrium), and 3 : 1. The test data are summarized in Tables 3 and 4.

Table 3. Yields of CO₂ hydrogenation products vs. process temperature for catalysts with different ratios of active components

Table 4. Selectivity towards CO₂ hydrogenation products vs. process temperature for catalysts with different ratios of active components

These data clearly show that the molar ratio of active components in the bimetallic catalyst had no significant effect on the CO₂ conversion as a function of temperature: for all the tested samples, the CO₂ conversion increased from 23–26% at 240°C to 55–58% at 320°C (Fig. 4).

Fig. 4.

CO₂ conversion as a function of temperature for catalysts with different ratios of active components.

In contrast, the yields of C₅₊ and oxygenates were largely dependent on the ratio of active components in the catalyst. The highest yield of C₅₊ and the highest total yield of C₅₊/oxygenates were achieved in the presence of the Fe : Co 3 : 1 sample (Fig. 5). The highest yield of C₅₊ for this sample (50.4 g/m³) was obtained at 320°C (Fig. 5a). Intriguingly, the C₅₊ yield curve for this catalyst, unlike those for the samples with smaller Fe : Co ratios, displays considerable growth (by 11.2 g/m³) over the temperature range of 300–320°C. The highest total yield of C₅₊ and oxygenates for the Fe : Co 3 : 1 sample (53 g/m³) was obtained over the range of 280–300°C; further heating up to 320°C led this parameter to drop (Fig. 5b). A similar correlation between the total C₅₊/oxygenate yield and temperature was observed for the equilibrium iron–cobalt sample. The catalyst with the prevalence of cobalt (Fe : Co 1 : 3) exhibited a steady increase in the total C₅₊/oxygenate yield over the entire temperature range, reaching the highest point (47.7 g/m³) at 320°C.

Fig. 5.

Yield of C₅₊ (a) and total yield of C₅₊ and oxygenates (b) as functions of temperature for catalysts with different ratios of active components.

The ratio of active components also had a significant effect on the yield of methane (Fig. 6). Although the Fe : Co 1 : 3 sample achieved the lowest yield of methane at low temperatures (240–260°C), it provided the highest methane yield at high temperatures (300–320°C). Above 260°C, i.e. within the temperature range that ensured effective CO₂ hydrogenation, the lowest methane yield was obtained for the Fe : Co 3 : 1 sample.

Fig. 6.

Yield of methane in CO₂ hydrogenation as a function of temperature for catalysts with different ratios of active components.

Thus, the best combination of CO₂ conversion and yields of target products (C₅₊ and oxygenates) was achieved for the Fe : Co 3 : 1 bimetallic catalyst.

The composition of the active phase was examined by XRD. See the XRD patterns in Figs. 7 and 8.

Fig. 7.

XRD patterns: (1) Fe : Co 1 : 3; and (2) Fe : Co 3 : 1.

Fig. 8.

XRD patterns of Fe : Co 1 : 1 sample: (1) fresh; (2) activated; and (3) catalytically tested.

Figure 7 illustrates the phase compositions of the samples with Fe : Co ratios of 1 : 3 and 3 : 1 after CO₂ hydrogenation. The patterns clearly indicate that both samples contained the phases of magnetite, iron carbide, cobalt carbide, and an iron–cobalt alloy. The predominant phases in the synthesized catalysts were governed by their compositions: the iron-dominated sample (Fe : Co = 3 : 1) had markedly higher concentrations of iron carbide and magnetite, and markedly lower concentrations of cobalt carbide, than the cobalt-dominated sample (Fe : Co = 1 : 3).

The genesis of the active phase in the bimetallic catalyst was determined by successive examination of the fresh, activated, and catalytically tested (spent) samples. The examination results for the equilibrium iron–cobalt sample (Fe : Co = 1 : 1) are provided in Fig. 8.

These data clearly show that the fresh sample was characterized by low content of well-crystallized phases and the presence of magnetite and mixed Fe/Co oxide. The activated sample exhibited the predominance of the Fe–Co alloy, with a Fe to Co ratio of about 60 : 40 (judging from the lattice spacing in the alloy). The surface of the tested sample consisted of iron carbide, cobalt carbide, and iron oxide. The active phase had average crystallite sizes of about 10 nm for Fe₃O₄ and about 20 nm for Fe₅C₂ and Co₂C.

Thus, it was found that, in an iron–cobalt catalyst for CO₂ hydrogenation, the active phase forms immediately in the initial process step. Given the predominance of the Fe–Co alloy in the activated sample, and the significant decrease in the concentration of this phase in the spent sample with a simultaneous increase in the oxide and carbide phases active in the catalytic hydrogenation of CO₂, it is fair to state that the active phase was generated from the metal atoms contained in the alloy. In this case, the synergy between the iron and cobalt active phases manifested itself: the crystallites of the iron and cobalt active phases were not isolated but rather generated from the common clusters of the Fe–Co alloy. The ratio of active phases depends on the metal proportions in the bimetallic composition and is critical for the catalytic activity of the sample. The interaction between the cobalt and iron phases prevented the generation of methanation sites typical of monometallic cobalt catalysts. At the same time, the cobalt phase facilitated the hydrogenation and shifted the catalytic activity towards lower temperatures.

The test data suggest the following CO₂ hydrogenation mechanism for a bimetallic catalyst with Fe prevailing over Co:

—At low temperatures (240–260°C), CO₂ hydrogenation occurs mostly through CO formation followed by Fischer–Tropsch synthesis (reactions (1) and (2)). Under these conditions, hydrocarbons predominate in the products, and the gas phase has a high concentration of unreacted CO, with oxygenates being formed in negligible amounts.

—As the process temperature rises to 280–300°C, the oxygenate formation reactions (reactions (3) and (4)) make a markedly increasing contribution. This dramatically boosts the content of oxygenates in the products and noticeably reduces the CO content in the gas phase.

—At 320° the oxygenates interact to produce HCs (reaction (5)). This abruptly reduces the oxygenate concentration in the products (down to their complete disappearance) and considerably enhances the yield of C₅₊.

CONCLUSIONS

An investigation of the genesis of bimetallic catalysts for hydrogenation of CO₂ demonstrated that using biochar as a support facilitates the synergy between iron and cobalt active sites. In contrast to oxide supports, loading iron and cobalt compounds on biochar does not promote the formation of a massive mixed oxide phase. This oxide phase, usually spinel-like, is typical of the well-explored catalysts for Fischer–Tropsch synthesis. This type of active phase is not favorable for the generation of iron and cobalt carbides active in CO₂ hydrogenation, but rather intensifies methanation even at 220–240°C. In contrast, a carbon support when activated ensures that an essentially different active phase is generated, specifically an iron–cobalt alloy, which transforms into catalytically active carbides of iron and cobalt when CO₂ hydrogenation is launched. This mechanism precludes the formation of metallic cobalt highly active in methanation. Moreover, the distinctive features of biochar, such as the lack of developed porosity and the tendency to bind metal atoms to the surface through hydroxyl groups, prevent the active components from being agglomerated, as is typically the case with oxide supports. In contrast, biochar favors the formation of Fe–Co alloy crystallites about 10–20 nm in size, which subsequently transform into proportionately sized carbide clusters. Due to the formation of carbides from the alloy, iron and cobalt atoms are closely spaced, thus ensuring their synergistic catalytic activity. Varying the ratio of active components incorporated into the catalyst makes it possible to adjust the catalytic performance in terms of the yield and composition of the reaction products. This approach holds promise for further research to optimize the properties both of the biochar support and of the active phase generated on its surface.