Sodium-promoted bimetallic M–CoO_x catalysts (M = In, Ga, Mo, Mn, and V) for the hydrogenation of CO₂ to C₂₊ hydrocarbons

Abstract

In CO₂ hydrogenation, Na-V-modified cobalt nanoparticles redirect the product distribution from methane to C₂₊ hydrocarbons. Sodium facilitates C–C chain propagation and suppresses the methanation reaction, while vanadium enhances site dispersion and regulates the Co⁰/Co^δ+ ratio, collectively driving the coupled reverse water–gas shift and Fischer–Tropsch synthesis reactions.

---

Driven by green hydrogen technology, the utilization of CO₂ as a carbon feedstock offers a sustainable route to carbon-based fuels and chemicals.^1–3 Co-based catalysts are widely applied in Fischer–Tropsch (FT) synthesis due to their high chain-growth capability.^4,5 However, in CO₂ hydrogenation, Co-based catalysts often catalyze the formation of light alkanes, especially CH₄.^6,7 Studies on CO and CO₂ hydrogenation over Co-based catalysts indicate that Co⁰ favours H₂ dissociation and C–C coupling, whereas Co^δ+ sites (0 ≤ δ ≤ 3) enhance CO₂ adsorption and facilitate the production of CO by reverse water–gas shift (RWGS) reaction.^8–11 Therefore, tuning the chemical environment of Co could potentially regulate the product selectivity towards more valuable products in CO₂ hydrogenation, such as C₂₊ hydrocarbons (C₂₊H).

Introducing a second metal offers an effective way to tailor the geometric and electronic structures of Co sites, thereby adjusting the product distribution of CO₂ hydrogenation. Wang et al. found an in situ produced Cu⁰–Co₂C interface that facilitated CO insertion and exhibited a C₂₊H selectivity of 60% at 300 °C.¹² Bibi et al. developed a Co⁰ core and ZrO₂ shell architecture that stabilized Co nanoparticles (NPs) and offered a C₂₊H selectivity of 57% at CO₂ conversion of 70%.¹³ Jo et al. reported a Na and Mn co-promoted core–shell Co@CoO_x/Co₂C catalyst, in which the Mn-induced CoO_x shell suppressed excessive methanation, enabling a C₅₊H yield of 21%.¹⁴ The addition of alkali promoters has also been demonstrated to adjust surface C/H coverage on Co sites, thus tuning the hydrocarbon selectivities.^15,16

Although several bimetallic Co-based catalysts have shown promise for CO₂ to C₂₊H conversion, the catalyst library remains narrow, and the factors governing C₂₊H selectivity are still underexplored. Herein, we investigate the synergistic mechanism and structural modulation of Na–M–CoO_x catalysts (M = In, Ga, Mo, Mn, and V, Na = 1.0 wt%, M = 1.0 wt%) for CO₂ hydrogenation to C₂₊H. Under 250 °C and 3 MPa, the optimum Na-1V-CoO_x catalyst achieves a C₂₊H selectivity of 63% at a CO₂ conversion of 30%, close to the performance of benchmark catalysts reported in this area (Table S1, SI). The structure–activity relationship has been proposed based on a series of characterizations.

Pristine CoO_x catalyst, gave a CO₂ conversion of 42% and a CH₄ selectivity of 93% (Fig. 1a). The introduction of Na with only 1.0 wt% amount significantly influenced the catalytic performance, with the CO₂ conversion decreased to 15%. However, the CH₄ selectivity was suppressed to 39%, and the C₂₊H and C₅₊H selectivities were increased to 61% and 34%, respectively. The chain growth probability (α) of Na–CoO_x was estimated to be 0.71 (Fig. S1, SI). The CO selectivity was only 6.5%. Generally, the combination of Na–CoO_x with In, Ga, Mo, Mn, and V didn’t change the product distribution significantly, but the CO₂ conversion. Among them, Na–In–CoO_x exhibited a CO₂ conversion of 17% with the C₂₊H selectivity of 54%. Na–Ga–CoO_x achieved a CO₂ conversion of 18% with the C₂₊H selectivity of 62%. Na–Mo–CoO_x delivered a CO₂ conversion of 24% with the C₂₊H selectivity of 66%, but the CO selectivity increased to 11%. Na–Mn–CoO_x exhibited a CO₂ conversion of 28% with the C₂₊H selectivity of 61%.

Fig. 1 (a) Catalytic performance of CO₂ hydrogenation over CoO_x and Na–M–CoO_x (M = In, Ga, Mo, Mn, and V, Na = 1.0 wt%, M = 1.0 wt%) catalysts. (b) Stability of Na-1V-CoO_x. Reaction conditions: 0.1 g, 250 °C, H₂/CO₂ = 3, 3 MPa, 6000 mL g_cat.⁻¹ h⁻¹. C⁰_2–4 and C⁼_2–4 denote C₂–C₄ alkanes and C₂–C₄ olefins, respectively.

Na-V-CoO_x afforded the optimum promotional effect with CO₂ conversion of 30% and C₂₊H selectivity of 63%. The C₂₊H space time yield (STY) was 12 mmol g_cat.⁻¹ h⁻¹. The selectivity of liquid hydrocarbons (C₅₊H) was 37%. Besides, CO selectivity was only 1.8%, significantly lower than that over Fe-based and Ru-based catalysts in CO₂ hydrogenation.^17,18 Overall, V serves as an effective promoter for Na–Co-based catalysts in CO₂ hydrogenation to enhance the formation of desired C₂₊H product. We further investigated the V loading and found that a loading of 1.0 wt% was optimum for the formation of C₂₊H (Fig. S2, SI). With the V loading increased to 5.0 wt%, the CO₂ conversion decreased to 25%, while the selectivities for CO and C₂₊H remained nearly constant. For Na-10V-CoO_x, although the C₂₊H selectivity increased to 74%, the CO₂ conversion dropped sharply to 6.5%. Besides, the CO selectivity increased to 35%. The activation energies (E_a) calculated on C₂₊H STY at varying reaction temperatures (Fig. S3, SI) suggest that the introduction of 1.0 wt% V markedly reduced the E_a from 93 to 62 kJ mol⁻¹. However, if the V loading was increased to 10 wt%, the E_a was 76 kJ mol⁻¹. This trend generally agrees well with the catalyst reactivity in CO₂ hydrogenation (Fig. S2, SI). We evaluated the stability of both Na-1V-CoO_x and Na-CoO_x catalysts (Fig. 1b and Fig. S4, SI). During 70 h on stream, Na-1V-CoO_x exhibited significantly better stability than Na-CoO_x, especially in product distribution. The C₂₊H STY of Na-1V-CoO_x decreased slightly from 12 to 9.5 mmol g_cat.⁻¹ h⁻¹, whereas Na-CoO_x deactivated more severely, with C₂₊H STY dropping from 7.4 to 3.8 mmol g_cat.⁻¹ h⁻¹.

The reaction kinetics have been investigated over the optimal Na-1V-CoO_x catalyst. Raising the reaction temperature from 210 to 290 °C increased CO₂ conversion from 5.8% to 66% and decreased CO selectivity from 7.9% to 0.5% (Fig. 2a), indicating that a higher temperature is favourable for CO₂ activation and CO consumption. However, higher temperatures also shifted product selectivity from C₂₊H to CH₄. Catalytic performance is strongly dependent on reaction pressures (Fig. 2b). With the pressure increased from 1 to 4 MPa, the CO₂ conversion rose from 14% to 34%, while the CO selectivity decreased from 11% to 1.1%. These trends indicate that the consumption of CO intermediates likely follows the FT synthesis mechanism. Meanwhile, the C₂₊H selectivity remained nearly constant at 63–65% in the pressure range of 1–3 MPa, then gradually decreased to 56% at 4 MPa, suggesting that high pressure favors the deep hydrogenation of CO-derived intermediates.

Fig. 2 Catalytic performance of Na-1V-CoO_x catalyst. (a) Effect of reaction temperature. Reaction condition: 0.1 g, 210–290 °C, H₂/CO₂ = 3, 3 MPa, 6000 mL g_cat.⁻¹ h⁻¹. (b) Effect of pressures. Reaction conditions: 0.1 g, H₂/CO₂ = 3, 250 °C, 1–4 MPa, 6000 mL g_cat.⁻¹ h⁻¹. (c) Effect of WHSV. Reaction condition: 0.1 g, 250 °C, H₂/CO₂ = 3, 3 MPa, 3000–12 000 mL g_cat.⁻¹ h⁻¹. (d) Effect of reduction temperature. Reaction conditions: 0.1 g, H₂/CO₂ = 3, 250 °C, 3 MPa, 6000 mL g_cat.⁻¹ h⁻¹.

To probe the reaction pathway, we evaluated Na-1V-CoO_x at different weight hourly space velocities (WHSV) (Fig. 2c). As the WHSV increased from 3000 to 12 000 mL g_cat.⁻¹ h⁻¹, the CO₂ conversion decreased from 42% to 10%, while CO selectivity rose from 0 to 6.2%, indicating that shorter residence time limits the conversion of CO. C₂₊H selectivity was not affected, but C₅₊H selectivity decreased from 46% to 32% as the WHSV increased from 3000 to 12 000 mL g_cat.⁻¹ h⁻¹, and the olefin/paraffin (o/p) ratio decreased from 0.6 to 0.3 (Table S2, SI). This implies a CO₂ → CO → C₂₊H → C₅₊H pathway. Since the redox state of Co may influence the hydrogenation behavior, we then assessed the impact of reduction conditions on the CO₂ hydrogenation over Na-1V-CoO_x (Fig. 2d). At the reduction temperature of 250 °C, the CO₂ conversion was only 7.0% with a high CO selectivity of 80%. This implies that Co-based catalysts in an oxidized state preferentially catalyze the RWGS reaction, and the formed CO cannot be efficiently consumed for C–C chain growth. As the reduction temperature increased to 300 °C, the reactivity of CO₂ hydrogenation increased significantly. However, further increasing the reduction temperature to 350 °C lowered the CO₂ conversion to 25%, along with the increase of CH₄ and C⁰_2–4 selectivities. Generally, C₂₊H selectivity decreases as the CO₂ conversion increases, while CH₄ selectivity follows an inverse trend according to statistics of all catalytic results (Fig. S5 and Table S2, SI). Compared with reported Co-based and Fe-based catalysts, which were typically evaluated in a high temperature range of 270–350 °C to sustain high C₂₊H STY, our catalyst achieves a comparable C₂₊H yield and STY under relatively mild conditions (Table S1, SI).

The ICP-OES analysis evidences that the Na-promoted catalysts have similar Na loadings of 0.9 wt%, and V-promoted catalysts have V loadings of about 0.9 and 9.5 wt%, respectively (Table S3, SI), which are close to their dosages in catalyst preparation. XRD analysis shows that all samples exhibit diffraction peaks corresponding to Co₃O₄ (Fig. S6, SI). Notably, no discernible diffraction peaks for crystalline VO_x or Na species were observed, suggesting that V and Na are highly dispersed. The V incorporation can reduce the cobalt crystallite size (Table S3, SI). N₂ physisorption reveals a type IV isotherm with hysteresis loops for all the catalysts, indicating mesoporous structures arising from interparticle space (Fig. S7 and Table S3, SI). With the increase of V loading, the surface area and pore volume rise from 11 m² g⁻¹/0.1 cm³ g⁻¹ (Na-CoO_x) to 18 m² g⁻¹/0.1 cm³ g⁻¹ (Na-1V-CoO_x) and 32 m² g⁻¹/0.2 cm³ g⁻¹ (Na-10V-CoO_x). SEM reveals that Na-CoO_x shows a layered morphology with densely packed, relatively large NPs and limited interparticle space (Fig. 3a–c). Introducing V partially disrupts the layered framework and yields smaller particles in the form of bundle-like aggregates, which is consistent with the increased BET surface area and pore volume.

Fig. 3 SEM images of (a) Na-CoO_x, (b) Na-1V-CoO_x, and (c) Na-10V-CoO_x catalysts. (d) TEM image of Na-1V-CoO_x. (e) HAADF-STEM image and EDS mapping of Na-1V-CoO_x catalyst.

Ultraviolet–visible spectroscopy (UV-vis) shows that the three catalysts exhibit similar spectral features, with a strong 200–400 nm Co–O charge-transfer band and a broad 600–800 nm Co²⁺ d–d band (Fig. S8, SI).¹⁹ Compared with Na–CoO_x, V-containing samples show a red shift of the shoulder, suggesting that dispersed VO_x modulates the local Co electronic structure.²⁰ To analyse the crystalline phases and elemental distributions, we performed HRTEM and STEM-EDS mapping on the three catalysts. For Na–CoO_x, the average size of Co NPs is 37 nm, and the lattice spacing of 0.25 nm corresponds to the (311) plane of Co₃O₄ (Fig. S9a–c, SI). After introducing 1.0 wt% V, the particle size decreases to 25 nm (Fig. 3d and Table S3, SI). Meanwhile, EDS mapping reveals a homogeneous distribution of V and Na (Fig. 3e). When the V loading is further increased to 10 wt%, the particle size decreases to 21 nm, and Co₃O₄ lattice planes such as (311) and (220) can be observed, whereas V-related species remain highly dispersed (Fig. S9d–f, SI).

H₂ temperature-programmed reduction (H₂-TPR) reveals the role of V loading in tuning the reduction behavior of Co species (Fig. S10, SI). Na-CoO_x exhibits the typical reduction peaks at 343 °C (Co³⁺ → Co²⁺) and 408 °C (Co²⁺ → Co⁰).²¹ Both Na-1V-CoO_x and Na-10V-CoO_x show broadened and prolonged reduction peaks, and the peak area is smaller than that of Na–CoO_x, indicating that V doping suppresses the overall reducibility of Co species. The reduction degrees of Na-CoO_x, Na-1V-CoO_x, and Na-10V-CoO_x, as estimated from H₂ consumption at 300 °C, are 74%, 53%, and 33%, respectively (Fig. S11 and Table S3, SI). This indicates that the Co exists in a mixed-valence state of Co⁰ and Co^δ+ after reduction. This is also the reason that the cobalt catalyst is named as CoO_x in this work.

XRD analysis for reduced catalysts suggests that Na–CoO_x contains both CoO and Co⁰, indicating partial reduction of Co₃O₄ (Fig. S12, SI). With 1.0 wt% V added, CoO and Co⁰ still coexist, but the Co⁰ peaks weaken, indicating V suppresses the reduction of Co oxides. The absence of V-related diffraction peaks implies that V remains highly dispersed after reduction. For Na-10V-CoO_x, the CoO peaks become broader and weaker, and the Co⁰ peak is almost absent. XRD patterns of the used catalysts reveal their structural evolution and stability under reaction conditions (Fig. S13, SI). After reaction, Na-CoO_x is dominated by Co⁰. Na-1V-CoO_x exhibits much weaker and broadened characteristic diffraction of Co⁰, with the 42–45° peaks shifting to higher 2θ values than Na-CoO_x, indicating lattice contraction in Co domains likely caused by lattice distortion.²² For used Na-10V-CoO_x, CoO and weak metallic Co⁰ reflections are observed, indicating that high V loading suppresses the reduction of Co species.

To elucidate the surface chemistry of used catalysts, X-ray photoelectron spectroscopy (XPS) was performed (Fig. 4a and Table S4, SI). Deconvolution of Co 2p reveals the presence of Co⁰ (778.1 eV), Co^δ+ (779.9 eV), and Co²⁺ (Co₃O₄) (781.2 eV) in all samples.^13,14,23 According to the literature, oxidized Co species (Co^δ+) are generally regarded as active for RWGS, whereas Co⁰ provides for C–C coupling in FT synthesis.^5,6 On this basis, we then examined how the Co⁰/Co^δ+ ratio correlates with catalytic performance. A volcano-shaped relationship is observed when the XPS-derived Co⁰/Co^δ+ peak-area ratio is plotted against the C₂₊H STY. At low Co⁰/Co^δ+ ratios, the formation of CO by RWGS reaction is favoured, while at high Co⁰/Co^δ+ ratios, methanation reaction dominates the reaction process (Fig. 4b and Table S5, SI).

Fig. 4 (a) Co 2p XPS of the used yV-CoO_x and Na-yV-CoO_x catalysts (y = 0, 1, and 10 wt%) after reaction for 18 h. (b) Correlation of C₂₊H formation rates and surface Co⁰/Co^δ+ ratios.

In situ DRIFTS was performed to probe surface intermediates and reaction pathways during CO₂ hydrogenation (Fig. S14, SI). Upon introducing the CO₂/H₂, a series of carbon–oxygen species were observed, including carbonate (*CO₃²⁻, 1400–1460 cm⁻¹), bicarbonate (*HCO₃⁻, 1500/1650 cm⁻¹), carboxyl (*COOH, 1700–1760 cm⁻¹), CO (2100–2200 cm⁻¹), and CH_x (C–H stretching at 3000 cm⁻¹) species.^24,25 Compared with Na-CoO_x and Na-1V-CoO_x, Na-10V-CoO_x exhibits much stronger carbonate, bicarbonate, and carboxyl signals, indicating its weaker hydrogenation ability, which is consistent with its low reactivity in CO₂ hydrogenation (Fig. S1, SI). Combined with reaction kinetic data (Fig. 2), we speculate that CO₂ is first hydrogenated to CO via the bicarbonate/carboxyl route on Co^δ+ sites. The generated CO is immediately consumed for CH_x formation and C–C coupling (FT synthesis) on the adjacent Co⁰ sites. Therefore, the balance between Co⁰/Co^δ+ is critical to determining the density of surface intermediates and final products.

In summary, we obtained an optimized Na-1V-CoO_x catalyst, offering CO₂ conversion of 30% with C₂₊H selectivity of 63% at 250 °C, and 3 MPa. The CO selectivity was only 1.8%. The introduction of V has been evidenced to suppress particle growth, leading to smaller Co domains and increased surface area. Besides, Na and V electronically modulate the electron density of Co, stabilizing an appropriate surface Co⁰/Co^δ+ ratio that balances the RWGS and FT synthesis steps. This balanced interplay enables efficient C–C coupling and significantly enhances the production of value-added C₂₊ hydrocarbons.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article are included in this file and in the supplementary information (SI). Supplementary information: experimental procedures, additional characterization, catalysis, and analysis data. See DOI: https://doi.org/10.1039/d6cc03041f.

Acknowledgements

This work was supported by the National Key Research and Development Program of Ministry of Science and Technology (No. 2022YFA1504600), the National Natural Science Foundation of China (No. 22121001, U25B6005, and 22572150), the Fujian Provincial Natural Science Foundation of China (2026J011001), the Fundamental Research Funds for the Central Universities (20720250081), and the Open Project of State Key Laboratory of Physical Chemistry of Solid Surfaces (No. 202102).

References

1. R. Ye, J. Ding, T. R. Reina, M. S. Duyar, H. Li, W. Luo, R. Zhang, M. Fan, G. Feng, J. Sun and J. Liu, Nat. Synth., 2025, 4, 288–302.
2. K. Cheng, Y. Li, J. Kang, Q. Zhang and Y. Wang, Acc. Chem. Res., 2024, 57, 714–725.
3. J. Ye, N. Dimitratos, L. M. Rossi, N. Thonemann, A. M. Beale and R. Wojcieszak, Science, 2025, 387, eadn9388.
4. K. Cheng, J. Kang, D. L. King, V. Subramanian, C. Zhou, Q. Zhang and Y. Wang, Adv. Catal., 2017, 60, 125–208.
5. A. Y. Khodakov, W. Chu and P. Fongarland, Chem. Rev., 2007, 107, 1692–1744.
6. M. Wang, G. Zhang, J. Zhu, W. Li, J. Wang, K. Bian, Y. Liu, F. Ding, C. Song and X. Guo, Chem. Eng. J., 2022, 446, 137217.
7. G. Melaet, W. T. Ralston, C.-S. Li, S. Alayoglu, K. An, N. Musselwhite, B. Kalkan and G. A. Somorjai, J. Am. Chem. Soc., 2014, 136, 2260–2263.
8. X. Zhou, G. A. Price, G. J. Sunley and C. Copéret, Angew. Chem., Int. Ed., 2023, 135, e202314274.
9. K. Li, X. Li, L. Li, X. Chang, S. Wu, C. Yang, X. Song, Z.-J. Zhao and J. Gong, JACS Au, 2023, 3, 508–515.
10. A. Parastaev, V. Muravev, E. H. Osta, T. F. Kimpel, J. F. M. Simons, A. J. F. van Hoof, E. Uslamin, L. Zhang, J. J. C. Struijs, D. B. Burueva, E. V. Pokochueva, K. V. Kovtunov, I. V. Koptyug, I. J. Villar-Garcia, C. Escudero, T. Altantzis, P. Liu, A. Béché, S. Bals, N. Kosinov and E. J. M. Hensen, Nat. Catal., 2022, 5, 1051–1060.
11. I. C. t Have, J. J. G. Kromwijk, M. Monai, D. Ferri, E. B. Sterk, F. Meirer and B. M. Weckhuysen, Nat. Commun., 2022, 13, 324.
12. M. Wang, S. Guo, S. Yan, Z. Wang, G. Zhang, H. Gao, M. Zhang, K. Bian, J. Liu, X. Nie, J. Zeng, C. Song and X. Guo, J. Am. Chem. Soc., 2025, 147, 42051–42060.
13. S. S. Bibi, H. Jo, J. R. Sugiarto, S. Ahmed, M. Irshad, W. Yoon and J. Kim, J. Energy Chem., 2025, 108, 769–784.
14. H. Jo, H.-J. Chun, J. R. Sugiarto, M. K. Khan, M. Irshad, W. Yoon, S. K. Kim and J. Kim, Appl. Catal., B, 2024, 359, 124457.
15. Z. Shi, H. Yang, P. Gao, X. Li, L. Zhong, H. Wang, H. Liu, W. Wei and Y. Sun, Catal. Today, 2018, 311, 65–73.
16. Z. Shi, H. Yang, P. Gao, X. Chen, H. Liu, L. Zhong, H. Wang, W. Wei and Y. Sun, Chin. J. Catal., 2018, 39, 1294–1302.
17. Z. Liu, F. Guo, L. Ge, S. Jiang, X. Lang, A. Dutta Chowdhury, Y. Zheng, K. Cheng and Y. Wang, Energy Fuels, 2025, 39, 20957–20965.
18. Q. Yang and E. V. Kondratenko, Acc. Mater. Res., 2024, 5, 1314–1328.
19. A. B. Vennela, D. Mangalaraj, N. Muthukumarasamy, S. Agilan and K. V. Hemalatha, Int. J. Electrochem. Sci., 2019, 14, 3535–3552.
20. S. Farhadi, J. Safabakhsh and P. Zaringhadam, J. Nanostruct. Chem., 2013, 3, 69.
21. K. Cheng, V. Subramanian, A. Carvalho, V. V. Ordomsky, Y. Wang and A. Y. Khodakov, J. Catal., 2016, 337, 260–271.
22. R. D. Shannon and C. T. Prewitt, Acta Crystallogr., Sect. B, 1969, 25, 925–946.
23. S. Zhang, K. An, J. Xin, Y. Jiang, M. Niu, P. Song, C. Wu, H. Wang and Y. Liu, Chem. Eng. J., 2024, 486, 150334.
24. W. Yu, M. Zhu, Z. Yang and Y.-F. Han, ACS Catal., 2025, 15, 3428–3441.
25. M. Wang, L.-C. Wang, B. Liu, Y. Ding, Y. Yang and D. Ding, ACS Catal., 2025, 15, 5954–5967.

---