Review of Iron-Based Catalysts for Carbon Dioxide Fischer–Tropsch Synthesis

Abstract

Capturing and utilizing CO₂ from the production process is the key to solving the excessive CO₂ emission problem. CO₂ hydrogenation with green hydrogen to produce olefins is an effective and promising way to utilize CO₂ and produce valuable chemicals. The olefins can be produced by CO₂ hydrogenation through two routes, i.e., CO₂-FTS (carbon dioxide Fischer–Tropsch synthesis) and MeOH (methanol-mediated), among which CO₂-FTS has significant advantages over MeOH in practical applications due to its relatively high CO₂ conversion and low energy consumption potentials. However, the CO₂-FTS faces challenges of difficult CO₂ activation and low olefins selectivity. Iron-based catalysts are promising for CO₂-FTS due to their dual functionality of catalyzing RWGS and CO-FTS reactions. This review summarizes the recent progress on iron-based catalysts for CO₂ hydrogenation via the FTS route and analyzes the catalyst optimization from the perspectives of additives, active sites, and reaction mechanisms. Furthermore, we also outline principles and challenges for rational design of high-performance CO₂-FTS catalysts.

Introduction

In recent years, carbon neutrality has gained significant attention, and global efforts have been devoted to mitigating CO₂ emissions. Due to its abundance and nontoxicity, the conversion of CO₂ into value-added products is an attractive approach to addressing the issue of CO₂ storage and sequestration and reducing dependence on fossil fuels by using CO₂ as feedstock for various processes [1,2,3]. Because CO₂ is very stable, its conversion into high-value-added products is challenging [4, 5]. Among various methods for CO₂ hydrogenation, thermal catalysis, photocatalysis, and electrocatalysis are the most promising approaches [6,7,8]. Particularly, thermal catalytic CO₂ hydrogenation for olefin production has recently made significant strides [9]. As important raw materials, olefins are traditionally produced by the petrochemical industry, and it was predicted that in 2022, the total global ethylene production capacity alone would reach 218 million tons, which would lead to more than 20 billion tons of CO₂ emissions [10, 11]. Therefore, direct synthesis of light and heavy olefins from CO₂ is a promising approach to achieve CO₂ neutralization.

CO₂ hydrogenation to olefins proceeds through two routes, namely the methanol route and the CO₂-FTS route (Fig. 1). The methanol route is through the conversion of CO₂ into methanol using the carbon-to-methanol process and then methanol into olefins using the methanol-to-olefins (MTO) technology. The catalysts used for the methanol route are typically bifunctional catalysts that comprise a metal/oxide component for methanol synthesis and a zeolite component for the MTO process, which can directly convert CO₂ into olefins [5]. Meanwhile, the CO₂-FTS route combines the reverse water–gas shift (RWGS) reaction and the Fischer–Tropsch synthesis (FTS) reaction. Similarly, the catalysts for the CO₂-FTS route also have two active sites, one for the RWGS reaction and one for the FTS reaction [1, 4, 5, 12, 13]. For CO₂ to olefins through the CO₂-FTS route, the RWGS reaction is endothermic, while the FTS reaction is exothermic, and thus, the CO₂-FTS route is thermodynamically more favorable by combining these two reactions. Furthermore, the FTS reaction consumes CO, which shifts the equilibrium of the RWGS reaction to the right and enhances CO₂ conversion [12, 13]. In addition, it has been shown that by regulating the active sites of catalysts, the CO₂-FTS route could possibly deviate from the Anderson–Schulz–Flory distribution, which opens an opportunity to achieve higher olefin selectivity. Therefore, the CO₂-FTS route is more favorable for practical applications due to its high efficiency in CO₂ conversion, selectivity, and energy utilization.

Fig. 1

Two pathways of CO₂ hydrogenation to olefin

CO₂ activation remains a major challenge in CO₂-FTS compared with conventional CO-FTS [14, 15]. CO₂-FTS catalysts typically exhibit a higher surface H/C ratio than CO-FTS catalysts, which can reduce the chain propagation ability and increase methane formation. Moreover, the low partial pressure of CO can enhance the secondary reactions of olefins, leading to the formation of more saturated hydrocarbons. Therefore, efficient CO production from RWGS is crucial for the production of olefins through FTS [16,17,18]. To overcome these challenges, there is a need for improved adsorption and activation of CO₂, as well as enhanced C–C coupling, to obtain olefins with higher carbon numbers. In addition, there is a need to increase the selectivity toward olefins while suppressing the formation of C₁ species and alkanes and improving the stability of the catalyst to prevent catalyst sintering and carbon deposition [10, 11, 16,17,18,19,20]. Therefore, a rational design of catalysts is essential to achieving high-efficiency CO₂ conversion to olefins.

Transition metal-based catalysts, such as Fe, Co, and Ru, are commonly used in CO₂-FTS. Co-based catalysts are highly active and preferentially produce linear hydrocarbons but suffer from low RWGS activity. Ru-based catalysts exhibit good activity at low temperatures but have low RWGS activity and are expensive [14]. Meanwhile, Fe-based catalysts are cheap, versatile, and highly efficient in catalyzing the RWGS reaction, which draws significant attention for their use in CO₂ conversion through the Fischer–Tropsch synthesis pathway.

Recently, various promoters such as N, Co, and Cu have been examined to improve the catalytic efficiency of Fe-based catalysts [21,22,23]. The active phases of Fe-based catalysts for RWGS and FTS are reported to be iron oxide (Fe₃O₄) and iron carbide (Fe₅C₂) [24]. For example, the introduction of transition metals such as Cu, Zn, and Co can enhance the carbon chain growth, likely due to the enhanced formation and distribution of Fe₅C₂ through metal–metal interaction [22], which improves the C–C coupling ability [21, 22, 25]. The introduction of alkali metals such as Na and K can increase the olefin selectivity, likely due to the increased surface alkalinity caused by the alkali metals, which enhances CO₂ adsorption and suppresses H₂ activation, leading to an increased surface C/H ratio [1,2,3, 26,27,28,29,30]. Moreover, the introduction of alkaline earth metals such as Sr can promote the dispersion of Fe active sites and facilitate the formation and stabilization of Fe₅C₂ phases. Moreover, Sr can enhance the electron interaction between Na and Fe species, leading to a synergistic effect that improves the C–O dissociation adsorption and the subsequent C–C coupling [30, 31].

CO₂ conversion has been reviewed extensively. Wang et al. [11] discussed the reaction mechanisms of CO₂ hydrogenation through CO₂-FTS and MeOH routes and the optimization of reaction conditions. Meanwhile, the main focus of Sun et al. [32] was on spinel ferrite-based catalysts for CO₂ hydrogenation to fuel-related chemicals. In contrast to previous reports, this review mainly focuses on Fe-based catalysts and their role in CO₂ conversion through the CO₂-FTS route (Fig. 2). The effects of various promoters, nature of active sites, and reaction mechanisms are included and discussed comprehensively, with the goal of providing insights into the rational design of high-efficiency CO₂-FTS catalysts [33].

Fig. 2

Scope and contents of the review on iron-based catalysts for CO₂-FTS

Promoters for Iron-Based Catalysts

Alkali Metal Promoters

The selectivity of olefins can be significantly enhanced by introducing alkali metals into catalysts. Sodium and potassium are common choices, as they can donate electrons and create an alkaline surface environment. This would facilitate the adsorption of CO₂, raise the surface C/H ratio, and inhibit the adsorption of olefins [9, 34,35,36]. Moreover, the addition of electrons to Fe can strengthen the Fe–C bond, contributing to the formation of iron carbide active sites [1,2,3, 9, 12, 26,27,28,29, 34, 36]. In addition, both Na and K can cause the isomerization and hydrogenation of alpha-olefins, leading to more branched-chain alkanes (Fig. 3). Compared to Na, K-doped samples had a higher linear/branched hydrocarbon ratio, suggesting that K had lower isomerization activity [1].

Fig. 3

Reproduced with permission from Ref. [44]. Copyright 2015, American Chemical Society

Secondary reaction pathways of Na- and K-iron-based catalysts.

The use of Na- and K-doped zinc-iron catalysts has been shown to improve the apparent reaction rate and yield of C₂₊ linear olefins for both CO and CO₂ under the same conditions [1]. The Na/Fe–Zn catalyst exhibited a higher apparent reaction rate and yield of C₂₊ linear olefins compared to the K/Fe–Zn catalyst. This suggests that Na can create a suitable balance between oxides and carbides, which is beneficial for producing long-chain olefins, especially C₅₊ olefins. This balance enhances the conversion of CO and CO₂ and increases the number of active sites on the surface. Moreover, Tu et al. [3] showed that Na promoters could enhance the stability of Fe₅C₂ by preventing phase oxidation during the reaction process (Fig. 4). Wei et al. [29] found that Na additives can also influence the particle size of Fe₂O₃ and Fe₃O₄, which can reduce the particle size of Fe₅C₂ during the CO reduction process.

Fig. 4

Copyright 2021, Elsevier

a Possible mechanism of Na on CO₂ hydrogenation of Fe₅C₂; b DRIFTS spectrum of the CO reduced (10 kPa CO, 523 K, 0.83 cm³/s) Fe₅C₂–ZnO (black line) and Fe₅C₂–ZnO–3Na (red line) catalysts after exposing to 10 kPa H₂ at 593 K for 1 h. (IR Spectrum were collected at 323 K under flowing Ar, 0.5 cm³/s); c product distribution at 593 K and 1.5 MPa H₂/CO₂ = 3 10 000 cm³/( g_cat·h) Reproduced with permission from Ref. [3].

It was reported that K also provides more electrons than Na during the reaction and promotes more carbide formation [1]. However, K also has higher hydrophilicity than Na, which causes strong physical adsorption of H₂O on K₂O. This may interfere with the adsorption of CO on iron carbides and affect the conversion rate. Guo et al. [2] showed that K tends to accumulate on the surface during the reaction. These surface-accumulated K species may be close to the active sites of iron carbide, providing active sites for activating C = O and coupling C–C bonds, and the surface-accumulated K was found more effective in promoting carbide formation than dispersed K [36]. In the study conducted by Kim et al. [9], they found that in mesoporous CuAl₂O₄-supported Fe catalysts, a small amount of K does not significantly improve CO₂ conversion and C₅₊ selectivity. A noticeable effect can be seen only with a larger loading amount (3 wt%). This is due to changes in the amphoteric properties of the catalyst surface, which affect the binding of CO₂ and change the activation energy for electron transfer between CO₂ and the catalyst surface [37]. According to Xu et al. [28], Na affects Fe–Mn catalysts by interfering with the contact between MnO_x and Fe₂O₃ in the oxide precursors of FeMnNa catalysts, thus weakening the inhibition effect of MnO_x on the reduction of Fe₂O₃. However, excessive alkali metals such as Na and K can significantly reduce the activity of Fe-based catalysts, which in turn results in excessive reduction of Fe due to their electron-donating ability, resulting in carbon deposition and the coverage of active sites on the catalyst surface [1, 9]. Moreover, excessive alkali metals enhance the surface alkalinity of the catalyst, making it difficult for olefins to adsorb, hampering carbon chain elongation, and affecting selectivity [27, 37].

Transition Metals and Oxides

Copper and Its Oxides

Copper doping can help reduce the catalysts and form active phases like Fe₃O₄ and Fe₅C₂, which are active sites for CO₂ hydrogenation through the CO₂-FTS route [13, 21, 22]. The interaction between Cu and Fe was reported to speed up the Fe reduction and carburization to form more active sites [9, 13]. Besides, Cu was also reported to participate in the RWGS reaction, which increases the CO levels, contributing to olefin formation processes [9, 13, 22, 38,39,40].

Generally, Cu promotes CO₂ conversion and the formation of more long-chain olefins. In the study of Liu et al. [39], they compared catalysts prepared using co-impregnation and sequential impregnation methods for FeCuK/Al₂O₃. They discovered that C₅₊ selectivity increased monotonically with increasing Cu content until the mass fraction of Cu reached 3%. Using Fe and Cu separately could result in the relatively complete growth of the hematite phase. However, the interaction between Fe and Cu in the co-impregnation catalysts was successfully regulated, leading to stronger metal interactions compared to the sequentially prepared catalysts. Similarly, Kim et al. [9] prepared a mesoporous Fe catalyst supported on CuAl₂O₄ and compared it with physically mixed 22Fe3K/SiO₂ and CuAl₂O₄. Both experimental data and characterization results demonstrated that the close chemical interaction between the Fe and Cu catalytic components is crucial for enhancing the CO₂ conversion rate, reducing C₁ product selectivity, and promoting C–C coupling reactions to improve C₅₊ selectivity.

Using a hydrothermal method, Zeng et al. [38] made CuFeO₂ catalysts that achieved high C₄₊⁼ selectivity (66.9%) and CO₂ conversion (27.3%) at atmospheric pressure (Fig. 5). They found that CO adsorbs without breaking on the Cu–Fe interface sites, and this CO^* adsorption is needed for CO insertion. The CO insertion at the Cu–Fe interface and the carbide mechanism on the carbide iron both lead to high C₄₊⁼ selectivity. They also studied the deactivation and regeneration of the catalysts and showed that the separation and re-diffusion of Cu and Fe elements cause deactivation, which could be inhibited at high-pressure reaction conditions. Choi et al. [41] investigated the impact of synthesis methods on the performance of CuFeO₂ catalysts. They discovered that the duration of the hydrothermal reaction significantly influences the crystal structure of CuFeO₂. As the hydrothermal duration extends from 6 to 24 h, there is a gradual decrease in the intensity of impurity phases such as Fe₂O₃ and Cu₂O. The nature of the precursor plays a pivotal role during the reduction process. In the spinel CuFe₂O₄, both metals exist in a fully oxidized state as Cu²⁺ and Fe³⁺. However, in CuFeO₂, Cu is in an intermediate oxidation state of Cu⁺, which may be less thermodynamically stable to reduction. In contrast, within the FeAlK system, the composition of Fe carbides increases with an increase in Cu content. Furthermore, Cu has proven effective in producing oxygen-containing substances [21].

Fig. 5

Copyright 2022, Springer Nature

a CO₂ conversion and product selectivity of activated Cu–Fe binary oxides with different Cu:Fe ratios. Reaction conditions: 320 °C, 0.1 MPa, H₂/CO₂ = 3:1, and 2 400 mL/( g_cat·h). Time on stream = 4 h. b SEM image of activated CuFeO₂; c schematic illustration of C–C coupling on the surface of activated CuFeO₂ Reproduced with permission from Ref. [38].

However, different observations for the roles of copper were also reported. Cai et al. [22] synthesized CuFe₂O₄-K catalysts using Prussian Blue Analogue (PBA). They found that, although with the highest selectivity toward C₅₊ products (85.50%), the incorporation of Cu was found to significantly reduce CO₂ conversion. Characterization results suggested that severe sintering of the Fe–Cu particles could be a contributing factor to the low CO₂ conversion rate. Yang et al. [13] prepared FeCu–Na catalysts in which Fe and Cu exhibited significant phase separation. The distribution of Cu was more concentrated, inhibiting C–C coupling and decreasing the selectivity toward long-chain hydrocarbons. Moreover, Cu was reported to act as an electron promoter, enhancing the surface basicity of the catalyst. And, the inclusion of Cu significantly improved the activation of H₂ and enhanced the secondary olefin hydrogenation capability, resulting in an O/P ratio noticeably lower than other catalysts.

Zinc and Its Oxides

The introduction of Zn improves the performance of the catalyst in several ways. First, it facilitates CO₂ adsorption, activation, CO adsorption, and hydrogen dissociation [22, 23, 42, 43]. Second, it creates a strong interaction between Fe and Zn to prevent the aggregation of iron species, stabilize carbide iron, and enhance C–C coupling capability. The Zn also acts as electron-donating groups to create a surface alkaline environment and suppress secondary hydrogenation of olefins [13, 22, 23, 25, 44].

Zn could be introduced in Fe-based catalysts by a simple co-precipitation method [13]. The introduction of Zn was found to improve the dispersion of Fe species and reduce the iron particle size, which facilitates the formation of active phases such as Fe₃O₄ and Fe₅C₂ (Fig. 6) [45]. Moreover, Zn forms a ZnFe₂O₄ spinel structure with Fe, which strengthens the interactions between Fe and Zn and prevents the aggregation of iron carbides and iron oxides during the initial structural evolution. Zn also enhances the activation of H₂, which suppresses the oxidation of Fe₅C₂. These factors result in the remarkable stability of FeZn-based catalysts.

Fig. 6

Copyright 2023, Elsevier

a TEM; b HRTEM image; c EDX element map of waste FeZn–Na catalyst; d effect of space velocity on the catalytic performance over the FeZn–Na catalyst. Reaction conditions: 320 °C, 3 MPa, H₂/CO₂ = 3 Reproduced with permission from Ref. [13].

In the study of Zhang et al. [46], they prepared a Zn-doped Fe-based catalyst, which transformed into Fe₅C₂ and ZnO species after reduction. The ZnO species enhanced the adsorption of CO₂ and CO, facilitated the generation of new active sites on Fe₅C₂ for CO₂ activation, and enhanced the chemical adsorption and dissociation of CO. The strong adsorption of CO inhibited the adsorption of H₂, delaying the hydrogenation of surface intermediates and favoring the formation of olefins and C–C coupling intermediates (–CH₂ groups). The electron donation from Zn species also made olefins more easily desorbed on Fe–Zn catalysts, resulting in a significant decrease in their secondary hydrogenation ability. Xu et al. [25] prepared Fe/Zn/Al–Na catalysts through co-precipitation (Fig. 7). The catalyst promoted by Zn and Al exhibited better stability. Zn existed in the form of spinel in Fe₆Zn₁Al₁, while in Fe₃Zn₁, it existed in the form of ZnO. Fe₃Zn₁ had a higher proportion of Fe_xC due to the strong stabilizing ability of Zn toward Fe_xC. Zn and Al exhibited synergistic effects by retaining the positive effect of Al while removing the Fe–Al spinel layer. This removal of spatial hindrance on the Fe₅C₂ surface by the Fe–Zn–Al spinel, which did not encapsulate Fe₅C₂ particles, improved the selectivity to α-olefins.

Fig. 7

Copyright 2021, ACS Publications

a Reaction mechanism; b, c Catalyst performance. Reaction condition (330 °C, 1.5 MPa, H₂/CO₂ = 3:1, and 15,000 mL/(g_cat·h)) Reproduced with permission from Ref. [25].

When Zn is introduced into FeAlK using the impregnation method [21], the reduced catalysts consist of ferrites, regardless of the Zn content (Fig. 8). Ferrites account for over 70% of the total composition and possess alkaline sites that greatly promote carbon–carbon coupling. Zn increases the adsorption strength and enhances the carbon–carbon coupling rate, favoring the production of long-chain hydrocarbons. Thus, Zn–FeAlK tends to generate more long-chain hydrocarbons than FeAlK. The doping of Zn in the zinc-iron spinel catalyst prepared from PBAs (Prussian Blue analogs) increases the difficulty of catalyst reduction but also promotes the dispersion of catalyst particles, significantly reducing their size. The strong interaction between iron species and Zn contributes to the excellent performance of Zn–Fe catalysts (Fig. 9) [22].

Fig. 8

Copyright 2022, Royal Soc Chemistry

a CO₂-TPD profiles of the catalyst samples; b H₂-TPR profiles of the catalyst samples; c Zn–FeAlK catalyst performance Reproduced with permission from Ref. [21].

Fig. 9

Copyright 2022, Royal Soc Chemistry

a CO₂ hydrogenation over Zn–Fe at different temperatures. Reaction conditions: 3 MPa, H₂/CO₂ = 3:1, 8 h, and 12 000 mL/(g_cat·h); b stability of the bimetallic Zn–Fe catalysts. Reaction conditions: 320 °C, 3 MPa, H₂/CO₂ = 3:1, and 12 000 mL/(g_cat·h). c Structural evolution and mechanism of C₅₊ hydrocarbon synthesis by CO₂ hydrogenation of MFe₂O₄ (M = zinc, copper, nickel, cobalt) Reproduced with permission from Ref. [22].

The effects of calcination temperature on the performance of Na_0.2/Fe₁Zn_1.2O_x catalyst were investigated by Yang et al. [44]. They found that catalysts calcinated at 400 °C exhibited smaller crystal sizes and lower total surface basicity. However, this catalyst also tended to convert α-olefin selectivity into branched alkanes. Wu et al. [43] prepared a layered K–Fe–Zn–Ti catalyst using high-temperature solid-phase reaction. They found that the interaction between Zn and Ti is weak. After Zn introduction, a stable ZnFe₂O₄ phase was formed, and the surface K content was significantly reduced, which facilitates the adsorption of CO₂, contributing to higher CO₂ conversion.

Manganese and Its Oxides

Manganese also has a strong interaction with iron, which inhibits the formation of iron carbide and enhances the reducibility of iron oxide. Therefore, manganese can improve the activity and selectivity of iron catalysts for various reactions. In addition, manganese also reduces the adsorption capacity of H₂ on iron surfaces. This can prevent hydrogen poisoning and increase the stability of iron catalysts [13, 28, 47,48,49].

For the FeMn–Na catalyst prepared by the co-precipitation method [13, 28, 47], it was found that the introduction of Mn significantly enhances the dispersion of Fe species, reduces the average grain size, and improves the reducibility of iron oxide. Mn possesses stronger reducibility compared to iron oxide and can assist in the removal of adsorbed oxygen species on the Fe surface through the overflow of oxygen vacancies in manganese oxide. The reduced catalyst contains FeMnO_x, indicating a close interaction between Fe and Mn. However, the strong interaction between Mn and Fe inhibits the interaction between Fe and CO, hindering the carburization of Fe and impeding the further conversion of CO intermediates. Therefore, the presence of single Mn suppresses the activity of Fe catalysts. To overcome this problem, some studies have introduced other promoters to modify the interaction between Fe and Mn and enhance the catalytic performance. For example, Praewpilin et al. [48] prepared a K/Mn/Fe/NCNT catalyst loaded on nitrogen-doped carbon nanotubes. The manganese and potassium promoters were coated on carbon nanotubes (MnK-CNTs) to avoid the formation of amorphous MnO_x phases that could block the iron phase. In contrast, Liang et al. [49] prepared Mn-modified Na/Fe catalysts, where the addition of Mn to the Na/Fe catalysts contributes to the formation of active Fe₅C₂ species. There is a strong interaction between Mn promoter and Fe species, which lowers the quantity and strength of CO adsorption, thereby helping to weaken the chain growth reaction and leading to high selectivity for light olefins.

Cobalt and Its Oxides

Cobalt (Co) is a traditional active metal in Fischer–Tropsch synthesis (FTS), a process that converts carbon monoxide (CO) and hydrogen (H₂) into liquid hydrocarbons. Co was also found to be available in CO₂-FTS[50]. In recent years, Fe–Co catalysts with excellent performance have been reported for CO₂-FTS. The introduction of Co can significantly decrease the particle size of the catalyst [2, 21, 42, 51], improving the reduction and adsorption of CO₂ molecules [52]. Furthermore, Co facilitates the formation of electron-rich Fe₅C₂, which enhances the electron density of carbides and strengthens the Fe–C bond while weakening the C–O bond. This can effectively enhance the catalytic performance of Fe–Co catalysts for CO₂-FTS [42, 52]. Guo et al. [2] prepared a supported bimetallic catalyst using iron and cobalt on Y zeolite modified by potassium ion exchange and found that the introduction of Co prompts the formation of more carbides, contributing to the improvement in selectivity toward heavy olefins. In the FeAlK system [21], introducing Co through co-precipitation would lead to the formation of heterogeneous Co metal in the iron lattice, resulting in defects and a reduction of iron crystal size.

Moreover, the presence of Co also enhances the adsorption of H₂ due to its strong H bonding energy [53]. However, it also hinders the migration of surface H to neighboring Fe species. In contrast, the CoFe₂O₄-K catalyst prepared using a Prussian blue analog (PBA) as a precursor exhibits good selectivity toward C_2–4 olefins (Fig. 9) [22]. Guo et al. [42] employed co-precipitation to introduce both zinc (Zn) and Co into Fe-based catalysts. A spinel structure of K-Zn(FeCo)₂O₄ was synthesized, which exhibited remarkably high activity (60.4%) and low CO selectivity (4.5%). The doping of Co metal enhanced the utilization of carbon elements and resulted in higher olefin yields. The KZFe-Co catalyst was further synthesized using a carbon template method [52]; with the introduction of Co, the cascade reaction between the reverse water–gas shift (RWGS) and chain propagation was significantly enhanced, and the CO₂ adsorption strength increased notably. Co and Fe were found to form an alloy phase (Co₃Fe₇) during the reaction, which improves the CO₂ adsorption and promotes the formation of oxygen-containing species (CO^*, HCOO^*, CO₃^2*, and HCO₃^*), which can further support chain propagation through the oxygenation reaction mechanism.

However, there is competition between methane formation and the generation of oxygen-containing functional groups, and an excess amount of Co can hinder the production of long-chain products [52]. When Co is used alone in CO₂-FTS, it exhibits a high tendency for methane formation. Therefore, when Fe and Co are used together, they may enhance the selectivity toward paraffins and suppress the production of olefins, attributing to the properties of Co promoting the hydrogenation of surface CH species, which lowers the probability of chain growth [54]. The conversion of CO₂ was found to increase with increasing Co content in iron-based catalysts. However, the intrinsic methane formation activity of Co can lead to a decrease in selectivity toward long-chain hydrocarbons (C₅₊) while increasing the selectivity toward light hydrocarbons (C_1–4). Therefore, the optimal amount of Co addition is crucial for achieving high performance. When a small amount of Co is added (5Co-FeAlK), Co atoms can be well integrated into the Fe structure, and the methane formation activity is not significant [55, 56]. On the other hand, when a larger amount of Co is added (12Co- and 28Co-FeAlK), the Co atoms in the FeCo alloy are likely to form clusters, accelerating the hydrogenation of surface CH_x species and leading to methane production. Co-based catalysts perform poorly for CO₂ hydrogenation: they display extremely high methane selectivity [22, 53, 56, 57]. They limit the growth of carbon chains due to the presence of a single cobalt carbide phase and decrease the conversion of CO₂–CO. The interaction between Fe and Co facilitates the transformation of Fe(III) to Fe(II) and further to carbide species to improve the adsorption and activation of CO₂ and CO intermediates [2, 21, 55, 56, 58]. Co can form a Co₃Fe₇ alloy phase with Fe, which also serves as an active site for FTS.

Nitrogen and Sulfur Doping

Carbon-based support materials, such as carbon nanotubes (CNTs) and activated carbon (AC), can be modified by nitrogen doping, a common method that also enables nitrogen-doped carbon materials, such as g-C₃N₄, to serve as catalyst supports. Generally, nitrogen doping could enhance the surface basicity of carbon materials, which would probably contribute to improving CO₂ adsorption and suppressing the absorption and dissociation of H₂, thereby inhibiting the secondary hydrogenation of light olefins [59, 60].

Nitrogen atom doping has been reported to facilitate the reduction of iron oxides and promote the formation of active carbon phases while inhibiting crystal aggregation [23, 48, 59, 61, 62]. Chew et al. [49] used a dry impregnation method to prepare iron catalysts supported on N-functionalized multi-walled carbon nanotubes (CNTs). Their experimental results showed that iron oxide nanoparticles supported on nitrogen-doped carbon nanotubes were more easily reduced than those supported on oxygenated carbon nanotubes. Using N-functionalized carbon nanotubes (NCNTs) as support, Kangvansura et al. [48] prepared an iron-based catalyst that exhibited high reducibility due to the high dispersion of iron within the NCNTs. The NCNTs were modified with 10 mol/L nitric acid, and the nitrogen in them induced certain distortions in the carbon-iron (π-d) interaction on the curved surface of the carbon nanotubes [63]. Zhang et al. [59] synthesized Fe₃O₄–FeC_x heterogeneous catalysts with active sites confined within N-doped graphene shells on the surface of N-doped ordered mesoporous carbon (N-OMC). The nitrogen doping promoted the formation of FeC_x and inhibited crystal aggregation (Fig. 10). By modifying the surface electron density with alkali metals and nitrogen atoms, they increased the surface alkalinity, suppressed the absorption and dissociation of H₂, enhanced CO₂ adsorption, and inhibited the excessive hydrogenation of CH_x, thereby improving the yield of light olefins.

Fig. 10

Copyright 2021, Elsevier

Surface analysis of catalysts with different Fe loading. a XPS spectra of 0.2Fe@N-OMC, 0.4Fe@N-OMC, 0.6Fe@N-OMC, and 0.8Fe@N-OMC; b CO₂ conversion and CO selectivity of catalysts with various Fe loading; c schematic of the preparation process of Fe₃O₄–FeC_x@N-OMC Reproduced with permission from Ref. [59].

Furthermore, nitrogen doping creates a rich electron environment that weakens the adsorption strength of hydrocarbon intermediates and promotes the formation of long-chain products such as C₅₊ [55, 61]. Nitrogen atoms in pyridine-like structures have higher electron density than those in pyrrole-like structures, making them more favorable for stabilizing Fe–C bonds [62]. The improved CO₂ hydrogenation performance depends on several factors influenced by the type of nitrogen doping, such as the catalyst’s specific surface area, the carbonization degree of the iron precursor, the number of defect sites, and the content of pyridine-like nitrogen structures [64]. Liu et al. [23] prepared a zinc oxide and nitrogen-doped carbon (NC)-coated iron-based catalyst, Fe@NC (Fig. 11), which showed a 25% increase in the reaction rate and a 24-fold enhancement in the O/P ratio compared to the benchmark Fe₃O₄ catalyst. This enhancement was attributed to ZnO, which is beneficial for CO₂ adsorption and hydrogen dissociation. The introduction of NC and alkaline accelerators improved the selectivity and O/P of light olefins.

Fig. 11

Copyright 2019, ACS Publications

a XRD patterns of Fe@NC; b FTY and product distribution over Fe₃O₄ and Fe@NC catalysts Reproduced with permission from Ref. [23].

Contrary to the conventional view of sulfur as a toxic component, sulfur was reported to act as a promoter for iron-based catalysts in FTS when doped at low concentrations [65]. In combination with alkali metal Na, sulfur was reported to lower methane selectivity, increase chain growth probability, enhance olefin selectivity, and facilitate the reduction and carburization of iron phases [66]. Using sulfur and sodium as promoters, Galvis et al. [65, 67] prepared iron nanoparticles uniformly dispersed on alumina or carbon nanofiber supports. The presence of sulfur was found to reduce methane selectivity and increase olefin selectivity by blocking the active sites for hydrogenation reactions. However, the sulfur-promoted catalysts also showed a higher degree of coking than the non-promoted samples.

Reaction Mechanism and Kinetics

Nature of Active Sites

The CO₂-FTS reaction involves two steps: the RWGS reaction, which converts CO₂ to CO on the iron oxide phase (mainly Fe₃O₄), and the subsequent CO hydrogenation FTS process, which produces hydrocarbons on the iron carbide phase (mainly χ-Fe₅C₂) [68, 69]. The composition, structure, and bonding of the iron oxide and iron carbide phases significantly affect CO₂-FTS. Before the reaction, the precursor is pretreated with H₂ or CO. The CO₂ in the reaction atmosphere and the water produced during the reaction are strong oxidizing agents, while H₂ is a reducing agent. Therefore, the catalyst undergoes drastic structural changes during reduction and reaction, making it complex to fabricate mixed iron oxide and iron carbide phases as working catalysts [70].

The dynamic equilibrium between the carbide and oxide phases is hard to control, because the final composition and structure of different phases involving iron and promoter elements depend on both the initial composition of the precursor and the activation/reaction processes [25, 29]. For most catalysts that require hydrogen activation, the oxidized phase is reduced to Fe after activation, and Fe interacts with CO₂ to form Fe₃O₄ in the early stages of the reaction. Over time, the promoter elements facilitate the in situ formation of Fe₅C₂ [21, 22, 25]. The in situ formation of Fe₅C₂ ensures the uniform distribution of the carbide phase, enabling better cooperation between the Fe₅C₂ and Fe₃O₄ active sites. Wei et al. [24] synthesized Na-Fe₃O₄ and bound it to the HZSM-5 molecular sieve to provide three types of active sites (Fe₃O₄, Fe₅C₂ and acidic sites) for the reaction, showing excellent performance (Fig. 12). Xu et al. [68] observed a significant amount of FeO_x overlayer on the χ-Fe₅C₂ main phase during the CO₂ hydrogenation process. The total RWGS rate was positively correlated with the surface content of FeO_x when the CO₂ conversion was less than 10% but negatively correlated when the CO₂ conversion was above 10%. In the latter case, the CO₂ conversion was limited by the subsequent removal of CO in FTS, which depended on the supply of surface FeC_x species. The inherent activity of RWGS is much higher than FTS. Therefore, stabilizing FeC_x through effective surface modification, either kinetically or thermodynamically, and preventing the oxidation of the carbide phase are crucial for realizing the full potential of iron-based catalysts in carbon-hydrogen compound production.

Fig. 12

Copyright 2017, Springer Nature

a CO₂ conversion and product selectivity over different Na–Fe₃O₄/Zeolite catalysts; b catalytic performances over composite catalysts as a function of the mass ratio of Na–Fe₃O₄/HZSM-5 in the composite catalysts. reaction conditions: H₂/CO₂ = 3320 °C, 3 MPa and 4000 mL/( h·g_cat); c reaction scheme for CO₂ hydrogenation to gasoline-range hydrocarbons Reproduced with permission from Ref. [24].

The in situ formation of carbide iron is significantly affected by the type of promoter elements and the bonding modes with iron elements. Both spinel structures [13, 21, 22, 25, 42] and perovskite structures [31] have shown good performance. Hou et al. [31] synthesized an ABO₃-type perovskite catalyst, Sr_1-xK_xFeO₃, with good thermal stability and redox activity. The introduction of K accelerated the oxygen release of SrFeO₃ and promoted the synchronous formation of Fe₃O₄ and Fe₅C₂. The reversibility of the perovskite catalyst ensured the high dispersion of active phases Fe₃O₄ and Fe₅C₂ within the SrCO₃ phase. Zhang et al. [59] synthesized a Fe₃O₄–FeC_x heterostructure and demonstrated that CO spillover on the FeO_x-FeC_x heterojunction was favored over diffusion between independent FeO_x and FeC_x species. This reduced the selectivity toward CO and improved the conversion of CO₂ and the yield of C₂₊ products. These phase interfaces also exist in Cu- or Zn-promoted Fe-based catalysts, which exhibit excellent carbon chain elongation ability under higher pressure through the CO insertion mechanism, leading to the formation of oxygen-containing products in CO₂-FTS. However, Zeng et al. [38] observed that as the reaction proceeded, the distance between Fe and M metals increased, and the metal phase interface was disrupted, resulting in a significant decrease in reaction activity.

In addition, the χ-Fe₅C₂ surfaces were also reported to be capable of activating CO₂. As revealed by Nie et al. [71] by DFT calculation, they found that the χ-Fe₅C₂(510) surface exhibits higher activity for the direct dissociation of CO₂ into CO^* and O^* with a barrier of 0.24 eV, while the χ-Fe₅C₂(111) surface is more favorable for the hydrogenation of CO₂ into HCO^* intermediate with a barrier of 0.25 eV. Despite the different reaction pathways, both the (510) and (111) surfaces are good candidate surfaces for the initial activation and transformation of CO₂ (Fig. 13).

Fig. 13

Copyright 2022, Elsevier

a Comparison of Gibbs free energy barriers (673 K) for CO₂* direct dissociation; b CO₂* hydrogenation to HCOO* and COOH* intermediates on different surfaces of χ-Fe₅C₂ Reproduced with permission from Ref. [71].

Kinetic Features of CO₂-FTS

As stated above, the hydrogenation of CO₂ over iron-based catalysts proceeds through the CO₂-FTS route, which combines the RWGS reaction and CO-FTS. The CO₂-FTS shows higher selectivity toward methane than the CO-FTS process, and the product distribution was also found to deviate from the Anderson–Schulz–Flory (ASF) distribution. As proposed by Visconti et al. [57, 72], two main reasons would account for the changes in product distributions. Firstly, the different adsorption strengths of CO and CO₂ on the catalyst surface led to different H/C ratios, which affected the chain growth ability. The relatively weak adsorption of CO₂ results in a higher H/C ratio on the catalyst surface, which promotes the hydrogenation of surface intermediates to CH₄ and reduces chain growth reactions. Secondly, as CO originates from the RWGS reaction, the surface coverage of which is relatively lower than that for the CO-FTS process, which influences the secondary reactions of olefins, leading to more saturated hydrocarbons. Moreover, due to the involvement of two-step reactions and two types of active sites, the bonding modes of these two sites significantly influence the diffusion and adsorption of intermediates, thus affecting the overall reaction [59, 68]. Therefore, to produce olefins with high selectivity in CO₂-FTS, it is necessary to enhance the activation of CO₂ and the synergistic combination of the two reactions, in addition to traditional measures to enhance CO-FTS.

Using DFT calculations, Wang et al. [73] studied the energetically favorable pathways and the selectivity factors of hydrocarbon production from CO₂ hydrogenation. They found that the favorable pathway for CO₂ hydrogenation involves HCOO^* intermediate and crucial CH^* species, which lead to the formation of CH₄ and C₂H₄, respectively. The CH^* species, passing through the HCO intermediate, is the key C₁ intermediate over χ-Fe₅C₂(510). The Microkinetic simulation results showed that CO₂ hydrogenation has higher selectivity toward CH₄ than C₂H₄, while CO hydrogenation shows the opposite trend. The main difference between CO₂ and CO hydrogenation is the different surface coverages of key species such as H^*, CH_x^*, and O^*. The higher surface coverage of O^* from CO₂ conversion occupies crucial active sites and impedes the coupling of C–C and C₂ species on χ-Fe₅C₂(510). Therefore, except for enhancing the adsorption and activation of CO₂, introducing additives such as alkaline metals and secondary transitional metals is crucial to trigger the formation and distribution of the active phases and optimization of the electronic environment of active sites to enhance the adsorption and transformation of critical intermediates [23, 59, 73].

Carbon Chain Extension Mechanism

There are two pathways for CO₂ dissociation: one is through hydrogenation of CO₂ to formate (HCOO*) intermediate, which further hydrogenates to generate methanol, and the other is through the formyl (^*HOCO) intermediate, which further dissociates into OH^*and CO^*. The latter dominates in Fe-based catalysts [3, 11, 16]. After the formation of CO through RWGS, the CO serves as the initial surface species for further carbon chain extension reactions. Mainly two mechanism-the carbide mechanism (direct and hydrogen-assisted CO dissociation mechanism) and the CO insertion mechanism-have been reported for the activation of CO to long-chain hydrocarbons [52, 58, 74, 75].

In the CO direct dissociation mechanism, the C–O bond is directly activated to form carbides as intermediates on the surface of catalysts forms during the FTS process [76]. These carbide species undergo hydrogenation to form CH_x groups, which serve as the building blocks for alkyl chain growth [77, 78]. In this mechanism, the C between Fe and M metals increased O cleavage precedes the C–C coupling, and the coverage of CH_x groups must be high enough to favor chain growth over chain termination through hydrogenation [78]. This requires a high rate of CO dissociation. Experimental and theoretical calculations have shown that direct CO dissociation has a high energy barrier and is more difficult [79, 80]. Therefore, the hydrogen-assisted CO activation mechanism has been proposed as an alternative.

In the hydrogen-assisted CO activation mechanism, CO^* is hydrogenated to form formyl (HCO) and hydroxymethyl (HCOH) intermediates before C–O cleavage [75]. During this process, CO first reacts with H to form the formyl intermediate (^*HCO). Then, the O atom of the HCO intermediate is further hydrogenated to form the hydroxymethyl (HCOH) intermediate. Finally, the HCOH dissociates to generate the CH species, which serves as the monomer and initiator for chain growth [76, 81, 82]. It has been reported that the activation of CO through direct CO dissociation has higher activation barriers on the surfaces of Fe, Co, and Ru catalysts than that through hydrogen-assisted ones [76]. The activation mechanism of CO to form long-chain hydrocarbons depends on the exposed surface during the reaction; CO tends to be directly activated on terraced-like χ-Fe₅C₂(510) surface using the hydrogen-assisted dissociation mechanism [79], but vice versa tendency was observed over χ-Fe₅C₂(010) [80], χ-Fe₅C₂(001) [83] and χ-Fe₅C₂(100) surfaces [84].

The hydrogenation of CO^* is highly endothermic (120 kJ/mol), and the C–O activation energy (178 kJ/mol) in the hydrogen-assisted pathway is still too high to match the observed reaction kinetics [75]. Therefore, the CO insertion mechanism has been proposed [85]. In this mechanism, the RCH_x^* group couples with CO^* before C–O cleavage and directly inserts into intermediates, elongating the carbon chain. As shown in Fig. 14, this process occurs at the metal interface such as Cu–Fe [38, 41] and Co–Fe [52, 58]. Because the CO insertion process relies on non-dissociative adsorption of CO, more oxygen-containing compounds would be formed in products [38]. In general, the carbide mechanism occurs simultaneously with the CO insertion mechanism [38, 52, 86]. Which pathway is more dominant appears to depend on the nature of active sites [38, 41, 52, 58]. It was found that direct CO dissociation is more favorable than the CO insertion pathway over these ferric carbide surfaces [87]. Moreover, on the Cu–Fe or Zn–Fe interfaces, the CO insertion mechanism is more advantageous [38].

Fig. 14

Copyright 2022, ACS Publications

a, c Reaction pathway for the formation of long-chain hydrocarbons through the carbide mechanism; b, d pathway of the oxygenated compound mechanism for the formation of long-chain hydrocarbons; e CO₂ hydrogenation stability over the KZFe–5.0Co catalyst at 320 °C, 2.0 MPa, 6000 mL/(g_cat·h) Reproduced with permission from Ref. [52].

Conclusions and Perspectives

Using CO₂ as a feedstock for olefin production is an effective way to utilize CO₂ and reduce dependence on petroleum and natural gas. The main challenge is to achieve high selectivity toward olefins and suppress C₁ by-products (CO and CH₄). This article summarizes the recent progress on iron-based catalysts for CO₂ hydrogenation to olefins (CO₂-FTS). CO₂-FTS has made significant advancements in recent years, especially in synthesizing long-chain olefins. However, the nature of active sites, the interactions between promoters and supports, and the reaction mechanism are still debated. Improved catalysts that combine RWGS activity with chain growth are needed. Catalysts can be tailored to achieve better performance by understanding the fundamental structure–composition–activity relationship of these catalytic systems.

The performances of iron-based catalysts can be enhanced by doping transition elements such as Cu, Zn, Mn, and Co. The interaction between transition metal elements and Fe affects the formation, dispersion, and stability of active phases, especially carbides in iron-based catalysts. Enough carbide content and proper distribution can enhance C–C coupling and increase the proportion of high-carbon number products. Therefore, improving the content and distribution of carbides through interactions between Fe and second metals may be a strategy to enhance product carbon number. Alkali metals can improve the selectivity toward olefins by increasing surface electron enrichment, enhancing the surface basicity of the catalyst, and facilitating olefin desorption. Moreover, increased surface basicity favors CO₂ adsorption, leading to reduced surface H/C ratio, suppression of intermediate hydrogenation, promotion of C–C coupling, and increased olefin yield. Therefore, improving the surface basicity conditions of catalysts and increasing the number of moderately strong basic sites for CO₂ adsorption are effective ways to enhance the proportion of olefin products (Table 1).

Table 1 Performances of iron-based catalysts for CO₂ hydrogenation through the CO₂-FTS route

Besides alkali metals, surface-basic nitride carbon supports can also be incorporated. The combination modes of promoter elements with Fe are also crucial for product selectivity. A well-designed structure can coordinate the carbide and oxide phases of iron better. Novel carbon materials such as carbon nanotubes and nitride carbons with new structures or special functional groups can also be explored as catalyst supports. Supports with specific spatial structures can limit the growth of active phase particle size caused by sintering and improve catalyst stability. Understanding key steps and intermediates in the reaction process through DFT calculations and stabilizing critical intermediates or lowering the energy barriers of key steps are meaningful research directions. Furthermore, it is important to consider the influence of the support and active phase environment and design an optimal CO₂-FTS catalyst with suitable metal phases, particle sizes, and surface structures, considering the impact of support and active phase surroundings.