*3.1. Promoter Effect*

Fe-based catalysts have been widely studied in CO2 hydrogenation, and usually show unsatisfactory selectivity toward lower olefins. The addition of suitable promotors to increase the yield of light olefins and the stability of the catalysts by controlling the electronic and structural properties have been extensively studied. Alkali metals such as K and Na have been broadly used as promotors to control the electronic properties. Mn, Ce, Ca metals have been used as structural promotors. Transition metals such as Zn, Co, Cu, V, Zr, etc., have been used as both electronic and structural promotors. Some representative catalysts on the promoter effect for CO2 hydrogenation to light olefins with improved catalyst stability are presented in Table 1.

**Table 1.** Some representative catalysts on promoter effect for CO2 hydrogenation to light olefins.


<sup>a</sup> refers to high valued olefins (HVO); <sup>b</sup> includes C2-C4 and C5+ hydrocarbons; <sup>c</sup> refer to C2–C7 <sup>=</sup> olefins.

Adding alkali metals (i.e., Na, K) could increase the selectivity towards light olefins due to the enhanced CO2 adsorption on the more electron-rich Fe phases and suppressed H2 chemisorption, which inhibits olefin re-adsorption. Numpilai et al. reported on the effect of varying the content of the K promoter on the Fe-Co/K-Al2O3 catalysts via the CO2–FTS reaction pathway. Unpromoted catalysts evidenced low-light olefin yields when compared to K-promoted ones with an ascending K/Fe ratio from 0 to 2.5. The maximum light olefin (C2 <sup>=</sup>−C4 =) distribution of 46.7% and O/P ratio of 7.6 were achieved over the catalyst promoted with a K/Fe atomic ratio of 2.5. The positive effect of K's addition is attributed to the strong interaction of H adsorbed with the catalyst surface caused by the electron donor from K to Fe species. This notion is also rationalized by the fact that the K promoter enhances the bond strength of absorbed CO2 and H2, retarding the hydrogenation of olefins to paraffins. In the same operating conditions, the catalyst promoted with a K/Fe atomic ratio of 0.5 provides the maximum light olefin (C2 <sup>=</sup>−C4 =) yield of 16.4%, which is significantly higher than that of 2.5 KFe catalysts (13.4%). This is explained by the K enriched surface of 2.5 KFe catalysts significantly reducing the BET surface area and generating a hydrogen-lean environment, ultimately lessening the catalytic activity [101].

A different promoter source plays an important role to affect catalytic CO2 hydrogenation. Han et al. demonstrated that as the series of K-promoters changes from K2CO3, CH3COOK, KHCO3, and KOH, the electron transfer from potassium to iron species is facilitated, which forms a more active and distinct χ-Fe5C2-K2CO3 interface during CO2 hydrogenation. This results in a higher selectivity to light olefins (75%) and a higher CO2 conversion (32%). In contrast, the non-carbonaceous K-promoters do not facilitate iron species to form iron carbides, which causes an undesirable catalytic performance (Figure 7a). Additionally, the close proximity between carbonaceous K-promoters and Fe/C catalyst components produced high olefin yields and catalytic stability (Figure 7b) [94]. Guo et al. reported that K derived from biological rather than inorganic precursors showed a stronger migration ability during the CO2 hydrogenation to light olefins. These surfaceenriched K ions extracted from corncobs could promote the carburization of iron species to form more Fe5C2, promoting both the reverse water–gas shift reaction and subsequent C–C coupling [97].

**Figure 7.** (**a**) Distribution of iron species content over different spent catalysts. (**b**) CO2 hydrogenation over Fe/C−K2CO3 catalysts with varying proximity. Adapted with permission from ref. [94]. Copyright 2020 American Chemical Society.

Metal organic frameworks as precursors for the preparation of heterogeneous catalysts have been used recently [99,106,107]. Ramirez et al. used a metal organic framework as a catalyst precursor to synthesize a highly active, selective, and stable catalyst, as shown in Figure 8a–f for the hydrogenation of CO2 to light olefins. Comparing the addition of Cu, Mo, Li, Na, K, Mg, Ca, Zn, Ni, Co, Mn, Fe, Pt, and Rh to an Fe/C composite, only K is able to enhance olefin selectivity, as shown in Figure 8c. The presence of K promoted the formation of Fe5C2 and Fe7C3 carbides, as confirmed by XRD (Figure 8e). K helped keep a good balance between the iron oxide for RWGS and iron carbide for FTS. The results presented in Figure 8f indicated a trend in which methane formation decreased and olefin selectivity increased as the K loading increased. The catalyst Fe/C+K(0.75) exhibited good stability (Figure 8d) and outstanding C2−C4 olefin space time yields of 33.6 mmol·gcat−1·h−<sup>1</sup> at XCO2 = 40%, 593K, 30 bar, H2/CO2 = 3, and 24,000 mL·g−1·h−<sup>1</sup> [99].

Some work may shed light on the ways in which the alkali promoters affect the behavior of iron catalysts. By the precisely controlled addition of promoters to fine tune the catalytic performance for the hydrogenation of CO2 to olefins, Yang et al. investigated how a zinc ferrite catalyst system could be affected by the addition of sodium and potassium promoters, specifically on the conversion of CO and CO2 to olefins. It was found that the catalyst's composition of iron oxides and iron carbides was altered in the presence of the promoters, which affected the CO and CO2 conversion. The production of C2+ olefins was greatly facilitated by the Na- and K-promoted catalysts. The Na/Fe-Zn catalyst was found to possess the optimal olefin productivity, and inhibited the competitive methanation reaction. It showed a total carbon conversion of 34.0%, which decreased by only 12.2% over 200 h [96]. Similarly, Wei et al. unraveled the effect of the Na promoter on the evolution of iron and carbon species, as well as the consequent tuning effect on the hydrogenation of CO2 to olefins. With the contents of the Na promoter increasing from 0 wt% to 0.5 wt%, the ratio of olefins to paraffins (C2+) rose markedly, from 0.70 to 5.67. The in situ XRD and temperature programed surface reaction (TPSR) confirmed that the introduction of the Na promoter decreased the particle size of Fe5C2 and regulated the distribution of surface carbon species. Furthermore, the in situ XRD and Raman demonstrated that the interaction between the Na promoter and the catalysts inhibited the hydrogenation of Fe5C2 and surface graphitic carbon species, consequently improving the stability of the Fe5C2 and enhancing the formation of olefins by inhibiting the hydrogenation of the intermediate carbon species [92]. Using a similar approach, Liang et al. modified the xNa/Fe-based catalysts with tunable amounts of sodium promoter for CO2 hydrogenation to alkenes, with CO2 conversion at 36.8% and a light olefin selectivity of 64.3%. It was found that the addition of the Na promoter into Fe-based catalysts boosted the adsorption of CO2, facilitated the formation and stability of the active Fe5C2 phase, and inhibited the secondary hydrogenation of alkenes under the CO2 hydrogenation reaction conditions (Figure 9a–c). The content of Fe5C2 correlated with the amount of Na is shown in Figure 9d [98].

Wei et al. synthesized a series of Fe3O4-based nanocatalysts with varying sodium contents. The residual sodium markedly influenced the textural properties of the Fe3O4-based catalysts, and faintly hampered the reduction of the catalysts. However, it discernibly promoted the surface basicity and prominently improved the carburization degrees of the iron catalysts, which is favored for olefin production. Compared with the sodium-free Fe3O4 catalysts, the sodium-promoted Fe3O4 catalysts displayed higher activity and selectivity for C2–C4 olefins. The FeNa catalyst (1.18) (Na/Fe weight ratio of 1.18/100) exhibited a high degree of catalytic activity with a high olefin/paraffin ratio (6.2) and selectivity to C2–C4 olefins (46.6%), and fairly low CO and CH4 production at a CO2 conversion of 40.5%. This catalyst also exhibited superb stability during the 100 h test at 593K. Comparing the scanning electron microscopy (SEM) image after reduction, there was no apparent indication of particle size growth after catalytic reaction for 100 h, further revealing the improved reaction stability of these iron nanoparticles [31]. Zhang et al. fabricated a Na- and Znpromoted iron catalyst by a sol-gel method, and demonstrated its high activity, selectivity and stability towards the formation of C2+ olefins in the hydrogenation of CO2 into C2+ olefins. The selectivity of the C2+ olefins reached 78%, and the space–time yield of olefins was as high as 3.4 g gcat−<sup>1</sup> h−1. The catalyst was composed of ZnO and χ-Fe5C2 phases with Na<sup>+</sup> dispersed on both ZnO and χ-Fe5C2. Zhang et al. found that ZnO functions for the RWGS reaction of CO2 to CO, while χ-Fe5C2 is responsible for CO hydrogenation to olefins. The presence of Na<sup>+</sup> enhanced the selectivity of C2+ olefins by regulating the

hydrogenation ability and facilitating the desorption of olefins (Figure 10a). The presence of ZnO not only efficiently catalyzes the RWGS reaction but also improves the activity and stability of CO2 hydrogenation by controlling the size of χ-Fe5C2 (Figure 10c,d). It was further discovered that the close proximity between ZnO and χ-Fe5C2 is beneficial for the conversion of CO2 to olefins (Figure 10b). The larger interface could facilitate the diffusion and transfer of intermediate CO from ZnO to χ-Fe5C2, favoring CO2 adsorption and subsequent CO hydrogenation to C2+ olefins [90]. Malhi et al. also investigated the effect of Na and Zn on iron-based catalysts, and found that the modified Fe-based catalyst exhibited a good performance for CO2 hydrogenation to olefins, with a CO2 conversion of 43%, a selectivity of 54.1% to C2+<sup>=</sup> olefins, and a high olefins-to-paraffins ratio of 7.3 [93].

**Figure 8.** (**a**) Illustrated synthesis of Fe/C catalysts. (**b**) TEM image of Fe/C catalysts. (**c**) Catalytic performance over promoted and unpromoted Fe catalysts. (**d**) CO2 conversion after 50 h of TOS. (**e**) XRD of promoted and unpromoted Fe catalysts. (**f**) Effect of K loading on the selectivity and CO2 conversion after 50 h of TOS. Testing conditions: 593 K, 30 bar, H2/CO2 molar ratio = 3, and GHSV = 24,000 mL·g−1·h−1. Adapted with permission from ref. [99]. Copyright 2018 American Chemical Society.

Chaipraditgul et al. investigated the effect of transition metals (Cu, Co, Zn, Mn or V) on the Fe/K-Al2O3 catalyst and found that the inclusion of the transition metal remarkably affected the interaction between the catalysts' surface and the adsorptive CO2 and H2.

The Fe/K-Al2O3 promoted with Cu, Co or Zn showed a lower the olefin to paraffin ratio, owning to a markedly increased number of weakly adsorbed H atoms resulting from the enhanced hydrogenation ability of the promoted catalysts. On the contrary, the addition of a Mn promoter to Fe/K-Al2O3 reduced the number of weakly adsorbed H atoms, lowering the hydrogenation ability to result in a high olefin to paraffin ratio of 7.4. The presence of either Mn or V inhibited the CO hydrogenation to hydrocarbon, leading to the low CO2 conversion, while the CO2 conversion was enhanced by incorporating either Co or Cu onto the Fe/K-Al2O3 catalyst [33]. Gong et al. investigated the promoting effect that Cu had on Fe-Mn-based catalysts in the production of light olefins via the CO2–FTS process. The Cu promoter was found to facilitate the reduction process and enhance CO dissociative adsorption by altering the interactions between Fe, Mn and the SiO2 binder, which led to increased activity. The addition of Cu weakened the surface basicity, which in turn decreased the chain growth probability and yielded a higher selectivity of light olefins [108].

**Figure 9.** (**a**) Illustrated scheme of CO2 hydrogenation. (**b**) Conversion and selectivity of CO2 hydrogenation over an xNa/Fe catalyst (testing conditions: H2/CO2 molar ratio = 3; P = 3 MPa; T = 593 K; GHSV = 2040 mL h−<sup>1</sup> gcat<sup>−</sup>1; TOS = 10 h). (**c**) Mössbauer spectra of the spent Na-free/Fe and 1Na/Fe catalysts. (**d**) Fe5C2 content of the spent xNa/Fe catalyst vs. the Na content. Adapted with permission from ref. [98]. Copyright 2019 American Chemical Society.

Jiang et al. reported the synthesis of Mn-modified Fe3O4 microsphere catalysts. These catalysts demonstrated excellent catalytic performance, with a 44.7% CO2 conversion, 46.2% light olefin selectivity, and 18.7% light olefin yield over the 10 Mn−Fe3O4 catalyst. The O/P ratio increased from 3.7 for the unpromoted Fe3O4 catalyst to 6.5 for the Mn-promoted catalyst. An even distribution of manganese was found over the surface of the Fe3O4 microsphere. Such homogeneous dispersion allows for an increase in the basicity of the catalyst, which prevents the further hydrogenation of olefins into paraffins. It was noted that the synergistic effects between Fe and Mn improve the dissociation and conversion of CO2 to hydrocarbons. The addition of Mn was found to promote the production of Fe carbides and enhance the active phases of CO2 hydrogenation and the FTS reaction, as well as preventing the hydrogenation of light olefins into paraffins and chain growth into

longer hydrocarbons [88]. A similar effect of the addition of Mn to Na/Fe catalysts was also observed by Liang et al. [95].

**Figure 10.** (**a**) Illustrated reaction mechanism for CO2 hydrogenation over the catalyst Na-Zn-Fe. (**b**) Effect of the proximity between ZnO and Na+Fe5C2 on catalytic behaviors for CO2 hydrogenation. Catalyst stability in CO2 hydrogenation over (**c**) an Na+Fe5C2 catalyst and (**d**) a Na-Zn-Fe catalyst. Testing conditions: H2/CO2 molar ratio = 3, P = 2.5 MPa, W = 0.10 g, F = 25 mL min−1, T = 613 K. Adapted with permission from ref. [90]. Copyright 2021 Elsevier.

Zhang et al. synthesized uniform microspheres of Fe-Zr-Ce-K catalysts by microwaveassisted homogeneous precipitation, and found that the reducibility, surface basicity and surface atom composition of the catalysts were greatly affected by varying the Ce content. CeO2, as the structural promoter, restrained the growth of Fe2O3 crystallite, weakening the interaction between Fe species and zirconia, and enabling the easier reduction of Fe2O3. The best performance was obtained on a 35Fe-7Zr-1Ce-K catalyst at 593 K and 2 MPa, with a CO2 conversion of 57.34%, a C2–C4 olefin selectivity of 55.67%, and a ratio of olefin/paraffin of 7 [100].

Extensive research efforts have been exerted on the development of bi-metallic catalysts for the conversion of CO2 to light olefins. Yuan et al. demonstrated the influence of Na, Co and intimacy between Fe and Co on the catalytic performance of Fe-Co bimetallic catalysts for CO2 hydrogenation that offers an olefin to paraffin ratio of 6 at a CO2 conversion rate of 41%. With the introduction of Co into the Fe catalyst, the CO2 conversion is significantly enhanced. The intimate contact between the Fe and Co sites favored the production of C2–C4 =. When Na was added to the system, the surface of the catalyst became carbon-rich and hydrogen-poor, allowing C–C coupling to form light olefins and suppress the methane formation. Moreover, the addition of a Na promoter facilitated the generation of χ-(Fe1-xCox)5C2 under the CO2 hydrogenation reaction conditions, and thus further improved the catalytic performances. A superb stability over 100 h was observed (Figure 11) [91]. Witoon et al. investigated the effect of Zn addition to Fe-Co/K-Al2O3 catalysts. The addition of Zn resulted in the improved dispersion and reducibility of iron oxides. For example, the 0.58 wt% Zn-promoted Fe-Co/K-Al2O3 catalyst afforded a large number of active sites for the adsorption of CO and H2 due to higher dispersion and an

eased reducibility (Figure 12a). The catalyst exhibited superior activity for light olefin formation with yield of 19.9% under the optimum testing conditions of 613 K, 25 bar, 9000 mL gcat−<sup>1</sup> h−<sup>1</sup> and a H2/CO2 ratio of 4. Figure 12b also shows a gradual decrease in the olefin to paraffin ratio, with an almost constant CO2 conversion as a function of the time-on-stream (TOS). The X-ray photoelectron spectroscopy (XPS) analysis of the spent catalyst showed the continuous growth of iron carbide with the time-on-stream, indicating that iron carbide may be the active component resulting in paraffin production (Figure 12c). XRD confirmed the formation of Fe-C phases over the spent 0.58 wt% Zn-promoted Fe-Co/K-Al2O3 catalyst at the time-on-stream (Figure 12d) [89].

**Figure 11.** Catalytic performance of the CO2 hydrogenation over the Na-CoFe2O4 catalyst at TOS (reaction conditions: H2/CO2 molar ratio = 3, T= 593 K, P = 3 MPa, GHSV = 7200 mL h−<sup>1</sup> gcat−1, TOS = 100 h). Adapted with permission from ref. [91]. Copyright 2021 Elsevier.

**Figure 12.** (**a**) Illustrated reaction mechanism. (**b**) Catalytic performance of the CO2 hydrogenation over a Zn-promoted Fe-Co/K-Al2O3 catalyst. (**c**) XPS spectra (Fe 2p region) of the 0.58 wt% Zn-promoted Fe-Co/K-Al2O3 catalysts. (**d**) XRD pattern of the 0.58 wt% Zn-promoted Fe-Co/K-Al2O3 catalyst at varying TOS. Testing conditions: T = 613 K, P = 25 bar, GHSV = 9000 mL gcat−<sup>1</sup> h−<sup>1</sup> and H2/CO2 molar ratio = 4. Adapted with permission from ref. [89]. Copyright 2021 Elsevier.

Wang et al. reported the synthesis of γ-alumina supported Fe-Cu bimetallic catalysts, and found a strong bimetallic promotion for selective CO2 conversion to olefin-rich C2+ hydrocarbons resulting from the combination of Fe and Cu at a specific composition. The suppression of the undesired CH4 formation was achieved by the addition of Cu to Fe while

simultaneously enhancing the C–C coupling for C2+ hydrocarbon formation. The formation of the Fe-Cu alloy in the Fe-Cu(0.17)/Al2O3 catalyst is suggested by the XRD results. Furthermore, the addition of K into the Fe-Cu considerably enhanced the production of C2 =–C4 <sup>=</sup> light olefins and the O/P ratio over Fe-Cu bimetallic catalysts. The Fe-Cu/K catalysts exhibited the superior selectivity of C2+ hydrocarbons compared to Fe-Co/K catalysts under the same reaction conditions [102]. Kim et al. synthesized monodisperse nanoparticles (NPs) of CoFe2O4 by the thermal decomposition of metal−oleate complexes, as shown in Scheme 4. The prepared NPs were supported on carbon nanotubes (CNTs), and Na was added to investigate the promoter and support effects on the catalyst for CO2 hydrogenation to light olefins. The resulting Na-CoFe2O4/CNT exhibited a superior CO2 conversion of 34% and a light olefin selectivity of 39% compared to other reported Fe-based catalysts under similar reaction conditions. The superb performance of Na-CoFe2O4/CNT was attributed to the formation of a bimetallic alloy carbide, (Fe1−xCox)5C2. Higher CO2 conversion and better light olefin selectivity were found in comparison with conventional Fe-only catalysts which possess χ- Fe5C2 active sites and drastically improved the C2+ hydrocarbon formation in comparison with Co-only catalysts which contain Co2C sites [103].

**Scheme 4.** Schematic demonstration of CO2 hydrogenation over CNT supported bi-metallic catalyst CoFe2O4. Adapted with permission from ref. [103]. Copyright 2020 American Chemical Society.

Song et al. investigated titania-supported monometallic and bimetallic Fe-based catalysts for CO2 conversion, and found that the mono-metallic catalyst (Fe-, Co-, Cu-) performed poorly for C–C coupling reactions. However, adding a small amount of a second metal (Co and Cu) to Fe revealed the synergetic promotion on the CO2 conversion and the space–time yields (STY) of hydrocarbon products. The inclusion of K and La as promoters further improved the activity, giving a higher hydrocarbon selectivity and O/P ratio, indicating that the promotor facilitated the CO2 activation and C–C couplings over bi-metallic catalysts [106]. Zhang et al. investigated Fe-Zn bimetallic catalysts for CO2 hydrogenation to C2+ olefins. A high C2+ olefin selectivity of 57.8% after 200 h of time-onstream at a CO2 conversion of 35.0% was obtained over an Fe2Zn1 catalyst. In bimetallic Fe5C2-ZnO catalysts, the ZnO plays a crucial role in improving the performance by altering the structure of the Fe components. Without ZnO, the chief deactivation mechanism was attributed to a phase transition from FeCx to FeO*<sup>x</sup>* over Fe2O3. However, with the addition of Zn to Fe2O3, the phase transformation and the carbon deposits over Fe2Zn1 were greatly diminished. Furthermore, the addition of Na inhibited the oxidation of χ-Fe5C2 active species for Fe-Zn bimetallic catalysts. During activation, both Zn and Na were shown to migrate onto the catalysts' surfaces. The oxidation of FeC*<sup>x</sup>* by H2O and CO2 was shown to be diminished by the interaction between Zn and Na [28].

Xu et al. investigated the roles of Fe-Co interactions over ternary spinel-type ZnCo*x*Fe*2-x*O4 catalysts for CO2 hydrogenation to produce light olefins. As shown in Figure 13, a high light olefin selectivity of 36.1%, a low CO selectivity of 5.8% at a high CO2 conversion of 49.6%, and an excellent catalyst stability were obtained over the ZnCo0.5Fe1.5O4 via the RWGS–FTS reaction pathway. It was shown that during the CO2 hydrogenation over ternary ZnCo0.5Fe1.5O4 catalysts, the formation of electron-rich Fe<sup>0</sup> atoms in the CoFe alloy phase significantly boosted the generation of the active χ-Fe5C2, Co2C, and θ-Fe3C phases, in which the χ-Fe5C2 phase facilitated the C–C coupling, the Co2C species suppressed the formation of CH4, and the formation of the θ-Fe3C phase with lower hydrogenation activity inhibited the second hydrogenation of light olefins [105].

**Figure 13.** (**a**) Schematic illustration of the structural transformations of as-formed ZnCoxFe2-xO4 catalysts during the reduction and reaction steps. (**b**) CO2 conversion and product distributions over K-containing ZnCoxFe2-xO4 catalysts with various Fe/Co molar ratios. (**c**) The stability of the Kcontaining ZnCo0.5Fe1.5O4 catalyst in CO2 hydrogenation (testing conditions: T = 583 K, P = 2.5 MPa, GHSV = 4800 mLh−1gcat−1, CO2/H2 molar ratio = 1:3). Adapted with permission from ref. [105]. Copyright 2021 Elsevier.

In summary, the use of the appropriate K or Na promoter, the inclusion of Cu, Co, Zn, Mn or Ce in the Fe phase, and the bi-metallic formation played important roles for enhanced catalytic performance and stability.

#### *3.2. Support Effect*

Catalyst support plays an important role in the overall activity and selectivity due to the interactions between the active metal components and the support during CO2–FTS. Some representative catalysts of the support effect for CO2 hydrogenation to light olefins with improved catalyst stability are presented in Table 2.


**Table 2.** Some representative catalysts of the support effect for CO2 hydrogenation to light olefins.

Owen et al. investigated the effect of Co-Na-Mo on various supports (SiO2, CeO2, ZrO2, γ-Al2O3, TiO2, ZSM-5 (NH4 +) and MgO) for CO2 hydrogenation. It was found that the surface area of the support and the metal–support interaction played a key role in the determination of the cobalt crystallite size, which strongly affected the catalytic activity. Cobalt particles with sizes < 2 nm supported on MgO showed low RWGS conversion with negligible FT activity, which is in agreement with the work of de Jong et al. [51]. When the cobalt particle size increased to 15 nm supported on SiO2 and ZSM-5, both the CO2 conversion and C2+ hydrocarbon selectivity increased markedly. When the cobalt particle size further increased to 25–30 nm, a lower CO2 conversion but higher C2+ light olefin selectivity was obtained. The authors reported that the higher the metal–support interaction, the higher the growth chain probability of the hydrocarbons. By altering the TiO2/SiO2 ratio in the support, the CO2 conversion and C2+ light olefin selectivity could be tuned [115]. Li et al. evaluated cobalt catalysts supported on TiO2 with different crystal forms of anatase (a-TiO2) and rutile (r-TiO2), and it was found that the addition of Zr, K, and Cs improved the CO, CO2, and H2 adsorption in both the capacity and strength over a-TiO2- and r-TiO2-supported catalysts. The surface C/H ratio increased drastically in the presence of promoters, leading to a high C2+ selectivity of 17% with 70% CO2 conversion over a K-Zr-Co/a-TiO2 catalyst. As a result, the product distribution could be tuned by adjusting the metal–support interaction and surface C/H ratio through Zr, K, and Cs modification over Co-based catalysts for CO2 hydrogenation, as shown in Scheme 5 [10].

**Scheme 5.** Schematic illustration of CO2 hydrogenation over unpromoted and Zr- and K-promoted cobalt catalysts supported on a-TiO2 and r-TiO2. Adapted with permission from ref. [10]. Copyright 2013 American Chemical Society.

Da Silva et al. found the Fe-Cr catalyst, promoted with K and supported on niobium oxide, was more active (CO2 conversion = 20%) and selective to light olefins (25%) compared to the same composition supported on silica (CO2 conversion = 11%, light olefin selectivity = 18%) under the same testing conditions. Alkali metal promotion increased the selectivity of olefins, probably due to electron-donor effects and the basicity of niobium oxide. A niobium oxide-supported Fe-Cr catalyst presented higher activity and selectivity to olefins, which is probably due to strong metal–support interactions when compared with traditional SiO2 [4]. Very recently, Huang et al. revealed the dynamic evolution of the active Fe and carbon species over different phases of zirconia (m-ZrO2 and t-ZrO2) on CO2 hydrogenation to light olefins, as shown in Scheme 6. Fe-K/m-ZrO2 catalysts performed better than the corresponding Fe-K/t-ZrO2 catalysts under the optimal reaction conditions. Among them, the 15Fe-K/m-ZrO2 catalyst showed remarkable catalytic activity, with a CO2 conversion of 38.8% and a C2–C4 <sup>=</sup> selectivity of 42.8%. More active species (Fe3O4 and χ-Fe5C2) with smaller particle sizes were obtained for the Fe-K/m-ZrO2 catalysts. The larger specific surface area facilitated the highly dispersed Fe species on the surface of the m-ZrO2 support when compared to the t-ZrO2 support. In addition, the

monoclinic phase m-ZrO2 support provided more strong basic sites, effectively decreasing the deposited carbon species and coke generation. Moreover, the electron-donating ability of iron elements and more oxygen vacancies (Ov) improved the charge transfer between ZrO2 and Fe. The synergy effect between K2O and ZrO2 fostered the generation of active carbide species. The formation of more χ-Fe5C2 species contributed to the high yield of light olefins [112]. Similarly, Gu et al. investigated Fe-K supported on ZrO2 with different crystal phases, revealing 40.5% CO2 conversion, 15.0% light olefin selectivity, and excellent stability (Figure 14) over 10Fe-1K/m-ZrO2 (10 wt% Fe and 1 wt% K) at 2.0 MPa and 613 K. The CO2 conversion was almost 200% higher than that of 10Fe-1K/t-ZrO2 [114].

**Scheme 6.** CO pre-reduction and CO2 hydrogenation process on (**a**) m-ZrO2- and (**b**) t-ZrO2 supported Fe-Zr catalysts. Adapted with permission from ref. [112]. Copyright 2021 American Chemical Society.

**Figure 14.** The stability of 10Fe1K/m-ZrO2 and 10Fe1K/t-ZrO2 for CO2 conversion, and the light olefin selectivity at a TOS of 100 h at 613K. Adapted with permission from ref. [114]. Copyright 2019 Elsevier.

Torrente-Murciano demonstrated that iron-based catalysts could be improved not only through the inclusion of promoters but also by the judicious control of the morphology of the ceria support (nanoparticle, nanorods, nanocubes) for CO2 hydrogenation to light olefins. For example, 20 wt% Fe/CeO2 cubes provided better catalytic performance (CO2 conversion = 15.2%, C2–C4 <sup>=</sup> selectivity = 20.2%) when compared with nanorods and their nanoparticle counterparts. TPR showed that the ceria reducibility decreased in the order of rods > particles > cubes, suggesting that the catalytic effect had a direct dependence on the reducibility of the different nanostructured ceria supports and their interaction with the iron particles [113]. By the physical mixing of Fe5C2 and K-modified Al2O3, Liu et al. discovered that Fe5C2-10K/a-Al2O3 exhibited a CO2 conversion of 40.9% and C2+ selectivity of 73.5%, containing 37.3% C2–C4 <sup>=</sup> and 31.1% C5+ (Figure 15). The superior catalytic performance was due to the potassium which migrated into the Fe5C2 during the reaction, and the intimate contact between the Fe5C2 and K/a-Al2O3. Among the various supports tested, as shown in Figure 15, alkaline Al2O3 is the best support for the high selectivity of value-added hydrocarbons [15].

**Figure 15.** (**a**) Catalytic performance of CO2 hydrogenation over Fe5C2-based catalysts on various supports. (**b**) Catalytic performance and stability over an Fe5C2-10K/a-Al2O3 catalyst (testing conditions: T = 593 K, P = 3.0 MPa, GHSV = 3600 mLg<sup>−</sup>1h−1, H2/CO2 molar ratio = 3). Adapted with permission from ref. [15]. Copyright 2018 American Chemical Society.

Dai et al. synthesized hierarchical porous carbon monoliths (HPCMs) by an adaptable strategy employing a one-step desilication process for a coke-deposited spent zeolite catalyst. This hierarchical porous carbon was shown to be a better support for the reduction of the nanoparticle size and heightening the synergism of the Fe–K catalyst for CO2 hydrogenation, with a CO2 conversion of 33.4% and a C2 =–C4 <sup>=</sup> selectivity of 18.0% [109].

Metal organic frameworks (MOFs) as novel porous materials had a considerable effect on the activity and selectivity of Fe-based catalysts. Hu et al. synthesized a type of hydrothermally stable MOF, zeolitic imidazolate frameworks (ZIF-8) with different sizes and morphologies, which were used as supports for CO2 hydrogenation. The acidity, internal diffusion process and crystal size enabled the ZIF-8 supports to show different levels of substantial light olefin selectivity [110]. Raghav et al. developed a simple method for the synthesis of hierarchical molybdenum carbide (β-Mo2C). The β-Mo2C phase exhibited the strongest metallic and some ionic character, and it behaved as both a support and co-catalyst for CO2 hydrogenation to light olefins. The Fe(0.5)-Mo2C catalyst exhibited a conversion of CO2 of 7.3% and a C2 <sup>=</sup> olefin selectivity of 79.4% at 300 ◦C and 4.0 mPa. The XRD patterns of the fresh and used Fe(0.5)-Mo2C catalyst did not show a noticeable difference, indicating the stability of the catalysts to achieve high olefin selectivity [111].

In summary, various supports (SiO2, CeO2, m-ZrO2, γ-Al2O3, TiO2, ZSM-5, MgO, NbO HPCMs, MOFs, β-Mo2C) have been used for the dispersal of active species. The surface area, basicity, reducibility, oxygen vacancies, and morphology of the support played important roles, in most cases with the presence of promoters (K, Zr, Cs), in affecting the amount and particle size of the active carbide species; the synergy effect; the metal–support interaction; the strength and capacity of the CO, CO2, and H2 adsorption on support; and the surface C/H ratio for CO2 hydrogenation. By tuning the above-mentioned characteristics properly, the physically deposited carbon species, coke generation and metal sintering could be mitigated as reported.

#### *3.3. Bifunctional Composite Catalyst Effect*

The zeolite–methanol composite catalyst can also be improved by compositional modifications. The composite catalyst is composed of two functional components: one is the target for methanol synthesis, mainly Cu, Zn, and In metal oxide catalysts; the other one is for the MTO process, mainly zeolite catalysts. Here, in this section, the recent progress on composite catalysts for improved catalytic performance and stability are described accordingly. Some representative catalysts for the bifunctional composite catalyst effect for CO2 hydrogenation to light olefins with improved catalyst stability are presented in Table 3.

**Table 3.** Some representative catalysts for the bifunctional composite catalyst effect for CO2 hydrogenation to light olefins.


<sup>a</sup> CO is not considered when calculating selectivity.

Wang et al. prepared kaolin-supported CuO-ZnO/SAPO-34 catalysts using kaolin as the support and raw material to prepare SAPO-34 molecular sieves. It was found that the resultant SAPO-34 molecular sieves showed a lamellar structure, relatively high crystallinity, and a larger specific surface area, which enabled the good dispersion of CuO-ZnO on the surface of the kaolin, and exposed more active sites for CO2 conversion. The confinement effect of (CuO-ZnO)-kaolin/SAPO-34 catalysts could prevent methanol dissipation, and provided an increased driving force for the conversion of CO2. Furthermore, the lamellar structure of SAPO-34 molecular sieves shortened the diffusion path of the intermediate product, and therefore enhanced the catalytic lifetime [116].

Gao et al. shown a selective hydrogenation process to directly convert CO2 to light olefins via a bifunctional catalyst composed of a methanol synthesis catalyst (In2O3-ZrO2) and a MTO catalyst (SAPO-34) by simple physical mixing. This bifunctional process exhibited an outstanding light olefin (C2–C3 =) selectivity of 80–90% with a CO2 conversion of ~20% and superior catalyst stability, running 50 h without obvious deactivation. The excellent catalytic performance was ascribed to the hybrid catalyst that suppressed the usually uncontrollable surface polymerization of CH*x* in conventional CO2–FTS. This was the highest selectivity reported to date, which dramatically surpassed the value obtained from traditional Fe or Co CO2–FTS catalysts (typically less than 50%) [43].

Similarly, Tan et al. evaluated CO2 conversion to light olefins over an In2O3-ZrO2/ SAPO-34 hybrid catalyst. This hybrid catalyst combined a In2O3-ZrO2 component, which would provide the benefit of oxygen vacancy to foster CO2 activation for hydrogenation into methanol, and a SAPO-34 component, to provide sites for the dehydration of the formed methanol into light olefins (Figure 16a). The light olefin selectivity reached 77.6% with less than 5% CO formation, which was ascribed to the strong adsorption of CO2 to defects in the In2O3 and ZrO2 components, creating a large energy barrier that suppressed CO2 dissociation into CO. The weaker acidity from In2O3-ZrO2 suppressed the further hydrogenation of the generated light olefins to paraffins. The catalyst displayed excellent stability, running for 100 h without obvious deactivation (Figure 16b) [44].

Furthermore, Gao et al. discovered that a bifunctional catalyst with an appropriate proximity containing In−Zr oxide, which was responsible for the CO2 activation, and SAPO-34, which was responsible for the selective C−C coupling, could greatly improve the CO2 hydrogenation to lower olefins with excellent selectivity (80%) and high activity (35% CO2 conversion) (Figure 17a). They showed that the incorporation of zirconium significantly improved the catalytic stability by preventing the sintering of the oxide

nanoparticles caused by the increase in surface oxygen vacancies. No obvious deactivation was observed over 150 h (Figure 17b) [19].

**Figure 16.** (**a**) Illustrated reaction mechanism over the bifunctional composite catalyst In2O3-ZrO2/SAPO-34, and (**b**) the stability of the In2O3-ZrO2/SAPO-34 composite catalyst for CO2 hydrogenation to light olefins (testing conditions: P = 2.0 MPa, T = 573 K, GHSV = 2160 cm3h−1gcat−1). Adapted with permission from ref. [44]. Copyright 2019 Elsevier.

**Figure 17.** (**a**) Effect of the proximity of the active components on the CO2 conversion and product selectivity, and (**b**) the catalytic stability of the composite catalyst In-Zr/SAPO-34 (testing conditions: T = 673 K, P = 3.0 MPa, GHSV = 9000 mL gcat−<sup>1</sup> h<sup>−</sup>1, molar ratio of H2/CO2/N2 = 73/24/3, and mass ratio of oxide/zeolite = 2). Adapted with permission from ref. [19]. Copyright 2018 American Chemical Society.

Wang et al. developed a new catalyst system composed of a Zn0.5Ce0.2Zr1.8O4 solid solution and H-RUB-13 zeolite. This composite exhibited a remarkable C2 =–C4 <sup>=</sup> yield as high as 16.1%, with a CO selectivity of only 26.5% due to the hindering of the RWGS reaction. It was demonstrated that methanol was first generated on the Zn0.5Ce0.2Zr1.8O4 solid solution via the formate–methoxyl intermediate mechanism, and was then converted into light olefins on H-RUB-13. By adjusting the H-RUB-13 acidity, the light olefin distribution can be effectively regulated, with propene and butene accounting for 90% of the light olefins [117].

Li et al. proposed a new synthetic strategy to prepare the bifunctional catalysts ZnZrOx/bio-ZSM-5. Hierarchically porous structured bio-ZSM-5 was prepared by using a natural rice husk as a template, which was then integrated with the ZnZrO*x* solid solution nanoparticles by physical mixing. The derived bifunctional catalysts ZnZrO*x* and bio-ZSM-5 exhibited superior light olefin selectivity and stability due to their unique pore structure, which was advantageous for mass transport and coke formation inhibition. \*CH*x*O was identified to be the key intermediate formed on the ZnZrO*x* surface, and was transferred to the Brønsted acid sites in the bio-ZSM-5 for the subsequent conversion to light olefins. The addition of a Si promoter to the ZnZrO*x*/bio-ZSM-5 catalyst prominently enhanced the

light olefin selectivity. The ZnZrO*x*/bio-ZSM-5−Si catalyst exhibited an outstanding light olefin selectivity of 64.4%, with a CO2 conversion of 10% and an excellent stability without noticeable deactivation during 60 h on stream (Figure 18a). In addition, the proximity of the catalyst components plays a key role in light olefin selectivity. As seen in Figure 18b, increasing the proximity resulted in a greater olefin selectivity [118]. By incorporating proper amounts of Ce or Cr ions into indium oxides, the methanol selectivity is increased, along with a reduction in the CH4 amount, as shown in Figure 19. Upon complexing with SAPO-34, a CO2 conversion of 33.6% and a C2 =–C4 <sup>=</sup> selectivity of 75.0% were achieved over InCrO*x*(0.13)/SAPO-34, which was about 1.5–2.0 times those obtained on In2O3/SAPO-34 and In–Zr/SAPO-34. This is because the incorporation of Ce or Cr ions into In2O3 lattice sites promoted the generation of more surface oxygen vacancies, as shown in Figure 19a, and enhanced the electronic interaction of HCOO\* with InCeO*x*(0.13) and InCrO*x*(0.13) surfaces, which decreased the free energy barrier and enthalpy barrier for the formation of HCOO\* and CH3OH. The composite catalysts also displayed excellent stability after 120 h on stream (Figure 19b) [119].

**Figure 18.** (**a**) Catalytic performance over the bifunctional composite catalysts ZnZrOx/bio-ZSM-5−Si at TOS (testing conditions: mass of catalyst = 0.6 g, T = 653 K, P = 3 MPa, gas flow rate = 20 mL min<sup>−</sup>1). (**b**) Effect of the proximity of the active components of ZnZrO*x*/bio-ZSM-5-Si on the catalytic performance. Adapted with permission from ref. [118]. Copyright 2021 American Chemical Society.

**Figure 19.** (**a**) The content of the surface oxygen vacancies (Ov) from O (1s) XPS spectra for the catalysts In2O3, InCeO*x*(0.13), and InCrO*x*(0.13). (**b**) The catalytic stability of the bifunctional composite catalysts InCrO*x*(0.13) and SAPO-34 for CO2 hydrogenation (testing conditions: H2/CO2 molar ratio = 3/1, T = 623 K, P = 3.5 MPa, and GHSV = 1140 mLgcat<sup>−</sup>1h−1). Adapted with permission from ref. [119]. Copyright 2020 Elsevier.

Similarly, Li et al. developed a bifunctional composite catalyst ZnZrO/SAPO-34 containing a ZnOZrO2 component to activate CO2 and H2 to form methanol, and a SAPO-34 component to perform C–C bond formation for the conversion of the produced methanol to light olefins. The derived dual function tandem catalyst exhibited an outstanding light olefin selectivity of 80% with good stability, and a CO2 conversion of 12.6% (Figure 20a,b). The kinetic and thermodynamic coupling between the tandem reactions enabled the highly efficient conversion of CO2 to lower olefins through the transfer and migration of CH*x*O intermediate species [13].

**Figure 20.** (**a**) CO2 hydrogenation over the bifunctional composite catalyst ZnZrO/SAPO-34, with the effect of the proximity of the active components of ZnZrO and SAPO-34 on the catalytic performance. (**b**) The catalytic stability of the catalyst ZnZrO/SAPO-34 (testing conditions: T = 653K,P=2 MPa, and GHSV = 3600 mL gcat−<sup>1</sup> h<sup>−</sup>1). Adapted with permission from ref. [13]. Copyright 2017 American Chemical Society.

Dang et al. advanced a series of dual function tandem catalysts containing In2O3- ZnZrO*x* oxides and various SAPO-34 zeolites with varying crystal sizes (0.4–1.5 mm) and pore structures. It was found that decreasing the crystal size of SAPO-34 could shorten the diffusion path from the surface to the acid sites inside the zeolite pores, thus favoring the mass transfer of intermediate species for efficient C–C coupling to produce lower olefins and enhance the selectivity of C2 =–C4 =. Interestingly, further HNO3 post-treatment caused the formation of the SAPO-34 zeolites with a hierarchical structure comprised of micro- /meso-/macropores, and reduced the amount of the Brønsted acid sites, both of which led to a significant increase in the catalytic performance, with the C2 =–C4 <sup>=</sup> selectivity reaching as high as 85% among all of the hydrocarbons (Figure 21a), a very low CH4 selectivity of only 1%, and an O/P ratio of 7.7 at a CO2 conversion of 17%. The C2 =–C4 <sup>=</sup> selectivity is much higher than the maximum predicted by the Anderson–Schulz–Flory distribution over modified FTS catalysts. The composite catalysts also exhibited excellent stability after 90 h on stream (Figure 21b) [42].

Liu et al. synthesized bifunctional composite catalysts composed of a spinel binary metal oxide ZnAl2O4/ZnGa2O4 and SAPO-34, with the selectivity of C2–C4 olefins reaching 87% at CO2 conversions of 15%. This study revealed that the oxygen vacancy site on metal oxides played a crucial role in the adsorption and activation of CO2, while the -Zn-Odomain accounted for H2 activation. It was demonstrated that the methanol reaction intermediates formed on the metal oxide, then converted to lower olefins at the Brønsted acid sites in SAPO-34 zeolite [121]. Tong et al. developed a dual-function composite catalyst, 13%ZnO-ZrO2/Mn0.1SAPO-34, and attained a high CO2 conversion of 21.3% with a light olefin selectivity of 61.7%, and suppressed the selectivity of CO below 43% and the CH4 selectivity below 4%. The fine-tuned acidity of zeolite by the addition of Mn and the

granule stacking arrangement contributed to the excellent catalytic performance. Mn was embedded into the zeolite ionic structure to tune the acidity of the molecular sieve and limit secondary hydrogenation reactions. The granule stacking arrangement facilitated the tandem catalysis [122]. Dang et al. presented a series of bifunctional catalysts containing In-Zr composite oxides with different In/Zr atomic ratios and SAPO-34 zeolite for CO2 conversion to light olefins. It was demonstrated that the inclusion of a certain amount of ZrO2 could provide more oxygen vacancy sites (Figure 22a), stabilize the intermediates in the CO2 hydrogenation, and prevent the sintering of the active nanoparticles. This, in turn, would lead to significantly enhanced catalytic activity, selectivity of hydrocarbons and stability for direct CO2 hydrogenation to lower olefins at the relatively high reaction temperature of 653K. A light olefin selectivity as high as 80% at a CO2 conversion rate of 27% and less than 2.5% methane selectivity was obtained over the optimized indiumzirconium/SAPO-34 bifunctional catalyst. The catalyst exhibited excellent stability for over 140 h without showing obvious deactivation (Figure 22b) [18].

**Figure 21.** (**a**) Catalytic performance for CO2 hydrogenation over a In2O3-ZnZrO*<sup>x</sup>* catalyst with different types of SAPO-34. (**b**) The stability of the bifunctional composite catalysts In2O3-ZnZrO*x*/SAPO-34-H-a (testing conditions: T = 653 K, P = 3.0 MPa, GHSV = 9000 mLgcat−1h−1, molar ratio of H2/CO2/N2 = 73:24:3, mass ratio of oxide/zeolite = 0.5). Adapted with permission from ref. [42]. Copyright 2019 Wiley-VCH.

**Figure 22.** (**a**) XPS spectra (O1s) of various oxides and the content of surface oxygen vacancies (Ov). (**b**) The stability tests of the bifunctional composite catalysts In-Zr(4:1)/SAPO-34 (testing conditions: T = 653 K, P = 3.0 MPa, GHSV = 9000 mL gcat−<sup>1</sup> h−1, molar ratio of H2/CO2/N2 = 73/24/3, and mass ratio of oxide/zeolite = 0.5). Adapted with permission from ref. [18]. Copyright 2018 Elsevier.

In summary, the majority of the catalysts tested for CO2 hydrogenation to light olefins via the MeOH-mediated route involve two active components (metal oxides and zeolite), which are so-called bifunctional composite catalysts. In this section, multiple variations (acidity, particle size, proximity, oxygen vacancy) in the combination of methanol synthesis catalysts (Cu, Zn, In, Ce, Zr, etc. metal oxides) with various zeolites (SAPO-34 and ZSM5) have been reported to give improved olefin selectivity and catalyst stability by mitigating coke formation, reducing the particle size growth of active carbide species, and inhibiting inactive species formation for CO2 hydrogenation to light olefins.
