Benzene hydrogenation utilizing organometallic early transition metal precursors

Abstract

Arene hydrogenation is an important catalytic reaction for the production of intermediates used in the synthesis of nylon, purification of gasoline fuels, and development of liquid organic hydrogen carriers. Herein is described the general utilization of organometallic complexes of early transition metals as pre-catalysts for the hydrogenation of benzene to cyclohexane. The metal precursors investigated for group 4 were [M(CH₂tBu)₄] (M = Ti, Zr, and Hf) and [V(CH₂SiMe₃)₃(THF)] and [M₂(μ₂-CSiMe₃)₂(CH₂SiMe₃)₄] (M = Nb and Ta) were chosen for group 5. These pre-catalysts were tested under semi-batch conditions at 120 °C and 400 psi of H₂, with the highest initial activity from group 4 being [Hf(CH₂tBu)₄] (TOF = 1155 mol C₆H₆ mol M⁻¹ h⁻¹) and for group 5 [Nb₂(μ₂-CSiMe₃)₂(CH₂SiMe₃)₄] (TOF = 1055 mol C₆H₆ mol M⁻¹ h⁻¹). These activities surpass 5 wt% Pd/C (TOF = 393 mol C₆H₆ mol M⁻¹ h⁻¹) and RANEY® nickel (TOF = 72 mol C₆H₆ mol M⁻¹ h⁻¹) tested under similar conditions. These catalytic performances offer new avenues in developing base metal catalysts for hydrocarbon conversion.

---

Introduction

The hydrogenation of aromatic substrates is an important transformation in modern chemical industries. It plays a key role in the production of bulk and fine chemicals, petrochemical refining, and in the development of liquid organic hydrogen carriers.^1,2 The hydrogenation of benzene accounts for nearly all industrial production of cyclohexane, which is most prevalently used as an intermediate in the production of nylon. The first reported catalytic hydrogenation of benzene was in 1901 by Sabatier and Senderens utilizing a Ni-based heterogeneous catalyst.³ Nickel is still utilized, in the form of RANEY® nickel, for the industrial hydrogenation of benzene; heterogeneous catalysts based on other group 10 metals have also been developed. The precise catalyst composition employed is highly dependent on specific reaction conditions in a process and the purity of the benzene source.^4,5

Despite heterogenous catalysts being widely used for industrial arene hydrogenation, homogenous systems active in catalytic arene hydrogenation have also been reported. These typically offer distinct advantages over heterogeneous catalysts in that ligand frameworks are utilized to promote chemo-, regio-, and stereoselectivity during arene hydrogenation. Homogeneous systems based on late transition metals have received significant interest in these directions due to a combination of catalyst activity and tunability.^2,6–8 Regarding homogeneous catalysts based on early transition metals, pioneering work by Rothwell and co-workers reported a series of Nb and Ta-aryloxide complexes that are capable of catalytic intra- and intermolecular hydrogenation of a variety of aromatic substrates.⁹ These systems invoke metal hydrides, typically prepared through hydrogen treatment of organometallic fragments, as the active species. This pioneering work has spurred other efforts in developing homogeneous catalysts based on early transition metals for the complete hydrogenation of arene substrates.^10–15

In an effort to generate heterogeneous catalysts based on homogeneous organometallic precursors, Rothwell and co-workers impregnated silica with solutions of organometallic complexes.¹⁶ While these materials were reported to be active in benzene hydrogenation at 120 °C and 1400 psi of H₂, a molecular description of the active sites for these systems has not been reported. However, other well-defined heterogeneous catalysts prepared by surface organometallic chemistry have provided insight into the nature of the active sites for arene hydrogenation.^17–21 Several organometallic Zr complexes ([CpZr(R)₃], R = –CH₃, and –Ph, –CH₂Ph) immobilized on sulfated oxide supports have shown activity in arene hydrogenation with catalysis occurring at room temperature and moderate pressures. In these studies, the organometallic precursors used in preparing the heterogeneous catalysts were reported to be ineffective catalysts for benzene hydrogenation.^17,18 We reasoned that under appropriate reaction conditions, organometallic precursors of early transition metals should be active in benzene hydrogenation without the need of immobilization on a support.

Herein, we report the general use of early transition metal organometallic complexes as pre-catalysts for arene hydrogenation reactions. We demonstrated that under relatively mild conditions (120 °C and 400 psi of H₂), transition metals of groups 4 and 5 are active in the catalytic hydrogenation of benzene to cyclohexane. The most active catalysts for groups 4 and 5 are Hf- and Nb-based displaying initial turnover frequencies (TOFs) of 1155 and 1055 mol C₆H₆ mol M⁻¹ h⁻¹, respectively. These values surpass those of a typical precious metal catalyst, namely 5 wt% Pd/C, and RANEY® Ni under similar reaction conditions.

Results and discussion

The organometallic precursors investigated and their catalytic performance in the hydrogenation of benzene to cyclohexane are summarized in Fig. 1 and Table 1. The range of molecular precursors encompasses group 4 and 5 transition metals. For group 4, the homoleptic –CH₂tBu complexes, [M(CH₂tBu)₄] (M = Ti, Zr, and Hf), were selected while –CH₂SiMe₃ complexes of group 5, [V(CH₂SiMe₃)₃(THF)] and [M₂(μ₂-CSiMe₃)₂(CH₂SiMe₃)₄] (M = Nb and Ta), were chosen. This selection of organometallic precursors provides a set of molecular complexes to evaluate general reactivity. Hydrogenation reactions were performed utilizing a benzene solution of the molecular pre-catalyst (ca. 1500 equiv. of C₆H₆ M⁻¹) where benzene was the solvent and substrate. The catalytic experiments were performed in an H.E.L. semi-batch reactor maintaining a constant pressure of H₂ (400 psi) and rate of stirring (1000 rpm) throughout the duration of the reaction. Hydrogen up-take was determined using HPCS-WinISO software and product analysis was performed by ¹H NMR spectroscopy of the filtered reaction mixtures after reaction times monitored up to 4 hours.

Fig. 1 a) General reaction conditions for benzene hydrogenation. b) Organometallic precursors investigated. c) Representative molar H₂ uptake for the hydrogenation of benzene to cyclohexane at 120 °C and 400 psi of H₂. H₂ uptake is shown up to 240 min and is normalized per mole of the metal in the reaction. Plateaus observed in the H₂ uptake for Nb, Hf, and Zr are indicative of full conversion.

Table 1 Catalytic hydrogenation of benzene to cyclohexane using organometallic precursors of groups 4 and 5^a

Pre-catalyst	Time^b	% Conv^c^,^d	TOF^c^,^e
a 4mL benzene, 120 °C, 400 psi of H2, 0.025–0.030 mmol M. b Reaction times less than 4 h indicate that full conversion was achieved at the listed time. c Reactions were performed in a minimum of triplicate and the reported values are an average of runs. d % Conv = (molinitial C6H6 – molfinal C6H6)/molinitial C6H6. e TOF = mol C6H6 M^−1 h^−1, determined over t = 0.25 h. f TOF was determined after an induction period, t = 0.67 h. g 4 mL benzene, 120 °C, 400 psi of H2, 0.150–0.165 mmol Ni.
[Ti(CH2tBu)4]	4 h	23 ± 5	173 ± 39
[Zr(CH2tBu)4]	3.2 h	97 ± 2	691 ± 46
[Hf(CH2tBu)4]	2.4 h	96 ± 4	1155 ± 24
[V(CH2SiMe3)3(THF)]	4 h	0	0
[Nb2(μ2-CSiMe3)2(CH2SiMe3)4]	2.1 h	98 ± 2	1055 ± 12
[Ta2(μ2-CSiMe3)2(CH2SiMe3)4]^f	4 h	44 ± 5	177 ± 11
5% Pd/C	4 h	56 ± 8	393 ± 31
RANEY® Ni^g	4 h	68 ± 12	72 ± 8

---

The overall catalytic performance in the hydrogenation of benzene to cyclohexane performed at 120 °C and 400 psi of H₂ was determined from a combination of H₂ uptake measurements and solution ¹H NMR spectroscopy of the products. Fig. 1 shows a comparison of H₂ uptake measurements from which TOFs are calculated for all investigated pre-catalysts at the initial times of the reaction. In all cases, cyclohexane was the only product detected by ¹H NMR spectroscopy of the reaction mixtures after catalysis. Using ferrocene as an internal standard, residual benzene in each spectrum was quantified and used to calculate conversions. No products of partial hydrogenation, hydrogenolysis, or ring contraction were detected, highlighting the high selectivity of the organometallic precursors investigated.

The most active pre-catalysts for benzene hydrogenation are [Nb₂(μ₂-CSiMe₃)₂(CH₂SiMe₃)₄], displaying an initial TOF of 1055 mol C₆H₆ mol M⁻¹ h⁻¹ with complete conversion being reached after 2.1 h, and [Hf(CH₂tBu)₄] with an initial TOF of 1155 mol C₆H₆ mol M⁻¹ h⁻¹ and full conversion occurring at 2.4 h. [Zr(CH₂tBu)₄] is slightly less active, when compared to Nb and Hf, with a TOF of 691 mol C₆H₆ mol M⁻¹ h⁻¹ and requiring 3.2 h to reach full conversion. Isostructural to Nb, the [Ta₂(μ₂-CSiMe₃)₂(CH₂SiMe₃)₄] pre-catalyst displays a lower initial TOF (177 mol C₆H₆ mol M⁻¹ h⁻¹) while reaching 44% conversion after 4 hours. For Ta, a short induction period is observed when performing the reaction at 120 °C. Higher reaction temperatures significantly reduce the induction period, while lower temperatures drastically reduce H₂ uptake over the duration of the reaction (see the ESI for details). In contrast to the pre-catalysts mentioned above, [Ti(CH₂tBu)₄] and [V(CH₂SiMe₃)₃(THF)] exhibited comparably low (173 mol C₆H₆ mol M⁻¹ h⁻¹) or no catalytic activity, respectively. The diminished activity is attributed to the thermal instability of the organometallic compounds.^22,23 To test this, we performed benzene hydrogenation reactions over the temperature range of 80 to 160 °C (Fig. 2). For [Ti(CH₂tBu)₄], an improvement in the performance is observed at lower temperatures, with increased TOFs from 173 mol C₆H₆ mol M⁻¹ h⁻¹ to 226 mol C₆H₆ mol M⁻¹ h⁻¹ when performed at 80 °C. In the case of V, lower temperatures did not result in an active catalyst. Optimal performance was observed between 120 and 140 °C for other 4d and 5d metals. When investigating temperatures greater than 140 °C, a decrease in catalytic performance is apparent. These findings support our claim that under the appropriate reaction conditions, organometallic precursors of early transition metals are active for benzene hydrogenation.

Fig. 2 Temperature dependence for the hydrogenation of benzene performed over a temperature range of 80–160 °C using a) group 4 and b) group 5 organometallic pre-catalysts. Reported values are an average of a minimum of three runs.

Comparison of the investigated organometallic pre-catalysts with homogeneous and heterogeneous systems based on group 4 and 5 metals is summarized in Table S3.† Homogeneous systems reported in the literature encompass Ti, Zr, and Nb, all being tested under different reaction conditions.^9,10,12,18 Of these, the ^R[P₂N₂]-Nb(CH₂SiMe₃) (R = Ph and Cy; [P₂N₂] = [RP(CH₂SiMe₂NSiMe₂CH₂)₂PR]) pre-catalyst reported by Fryzuk and co-workers has the highest reported TON (225 mol C₆H₆ mol M⁻¹) over the duration of 20 h.¹⁰ Under optimal reaction temperatures and heightened pressures of H₂ (1200 psi), our molecular Nb pre-catalyst has a TON of 3285 mol C₆H₆ mol M⁻¹ occurring over 4 h, a remarkable increase when compared to the homogeneous-Nb system. When compared to the heterogeneous catalysts of groups 4 and 5, a range of the TONs spanning ca. 100–4000 mol C₆H₆ mol M⁻¹ are reported. Catalysts with the highest TON were prepared via impregnation of organometallic precursors of Zr, Hf, Nb, and Ta on the surface of silica.¹⁶ These values were obtained over 5–24 h at 120 °C and 1200–1400 psi of H₂. The comparable TONs reported here highlight that support immobilization is not necessary for developing active arene hydrogenation catalysts.

The activities of the organometallic precursors were also compared to 5 wt% Pd/C and RANEY® Ni under similar reaction conditions in our reactor setup. Pd/C yields an initial TOF of 393 mol C₆H₆ mol M⁻¹ h⁻¹ and 56% conversion after 4 hours while the initial TOF for RANEY® Ni was 72 mol C₆H₆ mol M⁻¹ h⁻¹ with 68% conversion after 4 hours. Due to the high wt% of Ni in the latter, this was performed with a lower substrate to metal ratio (ca. 300 equiv. of C₆H₆ M⁻¹) in comparison to the other catalytic experiments. This is consistent with the higher conversion despite exhibiting significantly lower initial activity. Comparison revealed that all pre-catalysts showing activity exhibit a higher initial TOF when compared to RANEY® Ni. When compared to Pd/C, the Hf and Nb organometallic precursors show more than double initial activities (Table 1). Under the tested reaction conditions, this clearly demonstrates that base metals can compete with the catalytic performance of precious metals.

Post-analysis of the reaction mixtures via¹H NMR spectroscopy reveals the presence of neopentane and tetramethylsilane for group 4 and 5 complexes, respectively, with no residual organometallic precursor remaining in solution. In all cases, the post-catalysis reaction mixtures are colorless with the concomitant formation of a dark colored precipitate, which we are tentatively assigning to insoluble but active reduced forms of the metal precursors. Both early and late transition metal homogeneous systems have taken advantage of reducing conditions to generate active catalysts for hydrogenation that are likely heterogeneous.^24–27 For the most active group 5 pre-catalyst, [Nb₂(μ₂-CSiMe₃)₂(CH₂SiMe₃)₄], we performed reusability studies demonstrating that the isolated insoluble material post-catalysis remains active for arene hydrogenation in subsequent catalytic runs, albeit with slightly lower activity. This deactivation is attributed to the apparent deviation from zero-order kinetics with respect to the arene substrate (see the ESI† for details).

Homogeneous Nb-aryloxide catalysts have displayed high regioselectivity and stereoselectivity in the hydrogenation of aromatic substrates.^9,28,29 Hence, we probed benzene hydrogenation using the Nb pre-catalyst investigated in this study while implementing C₆D₆ as the substrate under established reaction conditions (120 °C, 400 psi of H₂, 4 h). Analysis of the products by GC-MS indicates a distribution of isotopologues spanning from cyclohexane-d₀ to cyclohexane-d₁₂ with the major isotopologue being cyclohexane-d₆ (Fig. 3). This suggests that hydrogenation and C–H activation pathways are competitive during the reaction. Expanding on this, the same reaction was performed at reduced reaction times to avoid full conversion of the substrate. These experiments revealed that even at ca. 80% conversion, a distribution of cyclohexane-d₀ to cyclohexane-d₁₂ products is observed, accompanied by the presence of unreacted C₆D₆. To unequivocally demonstrate that H/D exchange occurs with cyclohexane, we also performed the reaction of cyclohexane-d₀ with D₂ (120 °C, 200 psi of D₂, 4 h), revealing deuterium incorporation via GC-MS and ²H NMR spectroscopy. These experiments indicate that the C–H activation of hydrocarbon substrates and hydrogenation reactivity are feasible with the Nb pre-catalyst.

Fig. 3 a) Hydrogenation of C₆D₆ resulting in cyclohexane-d_0–12 isotopologues. b) H/D exchange of C₆H₁₂ with D₂ resulting in cyclohexane-d_1–4 isotopologues.

To examine the generality of these metals as catalysts for arene hydrogenation, homoleptic metal–amide complexes, [M(NMe₂)_n] (M = Zr and Hf, n = 4; M = Nb and Ta, n = 5) were also utilized. These precursors showed no detectable activity in the hydrogenation of benzene under similar reaction conditions (120 °C, 400 psi of H₂). We postulate that the in situ generation of HNMe₂ under conditions for benzene hydrogenation when using [M(NMe₂)_n] precursors could inhibit catalysis, highlighting the unique catalytic properties of the investigated organometallic complexes presented in this study.

Conclusions

In summary, we have demonstrated a general reactivity pattern of organometallic early transition metal complexes serving as pre-catalysts for the hydrogenation of benzene to cyclohexane under mild reaction conditions. Interestingly, discrete homoleptic arene complexes of both group 4 and 5 metals have been previously reported under reducing conditions.³⁰ This presents intriguing possibilities for the in situ generation of homogeneous single atom catalysts of base metals displaying exceptional catalytic performance. We are currently examining the exact nature of the active sites in the reported catalysts as well as investigating the activity and selectivity amongst other substrates. This will be explored through a variety of catalyst design strategies.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The University of Florida is thanked for partial financial support of this work. This work was also supported by the donors of the ACS Petroleum Research Fund under the Doctoral New Investigator Grant 65805-DNI5. K.S. served as the principal investigator on ACS PRF 65805-DNI5 that provided support for P.H. We also thank the Mass Spectrometry Research and Education Center and the funding source NIH S10 OD021758-01A1.

Notes and references

1. P. Preuster, C. Papp and P. Wasserscheid, Acc. Chem. Res., 2017, 50, 74–85.
2. M. P. Wiesenfeldt, Z. Nairoukh, T. Dalton and F. Glorius, Angew. Chem., Int. Ed., 2019, 58, 10460–10476.
3. P. Sabatier and J.-B. Senderens, C. R. Hebd. Seances Acad. Sci., 1901, 132, 210–212.
4. M. L. Campbell, Cyclohexane, in Ullmann's Encyclopedia of Industrial Chemistry, John Wiley & Sons, Ltd, 2011,  DOI:10.1002/14356007.a08_209.pub2.
5. K. Weissermel and H. Arpe, Ind. Org. Chem., 2003, pp. 337–385.
6. F. Glorious, L. Lückemeier and M. Pierau, Chem. Soc. Rev., 2023, 52, 4996–5012.
7. D.-S. Wang, Q.-A. Chen, S.-M. Lu and Y.-G. Zhou, Chem. Rev., 2012, 112, 2557–2590.
8. A. N. Kim and B. M. Stolz, ACS Catal., 2020, 10, 13834–13851.
9. I. P. Rothwell, Chem. Commun., 1997, 1331–1338.
10. M. D. Fryzuk, C. M. Kozak, M. R. Bowdridge and B. O. Patrick, Organometallics, 2002, 21, 5047–5054.
11. G. Zhu, K. Pang and G. Parkin, J. Am. Chem. Soc., 2008, 130, 1564–1565.
12. A. Gómez-Torres, J. R. Aguilar-Calderón, A. M. Encerrado-Manriquez, M. Pink, A. J. Metta-Magaña, W. Y. Lee and S. Fortier, Chem. – Eur. J., 2020, 26, 2803–2807.
13. P. Viereck, G. Hierlmeier, P. Tosatti, T. P. Pabst, K. Puentener and P. J. Chirik, J. Am. Chem. Soc., 2022, 144, 11203–11214.
14. M. V. Joannou, M. J. Bezdek and P. J. Chirik, ACS Catal., 2018, 8, 5276–5285.
15. G. Hierlmeier, P. Tosatti, K. Puentener and P. J. Chirik, J. Am. Chem. Soc., 2023, 145, 21027–21039.
16. R. D. Profilet, A. P. Rothwell and I. P. Rothwell, J. Chem. Soc., Chem. Commun., 1993, 42–44.
17. L. A. Williams, N. Guo, A. Motta, M. Delferro, I. L. Fragalà, J. T. Miller and T. J. Marks, Proc. Natl. Acad. Sci. U. S. A., 2013, 110, 413–418.
18. W. Gu, M. M. Stalzer, C. P. Nicholas, A. Bhattacharyya, A. Motta, J. R. Gallagher, G. Zhang, J. T. Miller, T. Kobayashi, M. Pruski, M. Delferro and T. J. Marks, J. Am. Chem. Soc., 2015, 137, 6770–6780.
19. M. M. Stalzer, C. P. Nicholas, A. Bhattacharyya, A. Motta, M. Delferro and T. J. Marks, Am. Ethnol., 2016, 128, 5349–5353.
20. C. P. Nicholas, H. Ahn and T. J. Marks, J. Am. Chem. Soc., 2003, 125, 4325–4331.
21. H. Ahn, C. P. Nicholas and T. J. Marks, Organometallics, 2002, 21, 1788–1806.
22. G. S. Girolami, J. A. Jensen, D. M. Pollina, C. M. Allocca, A. E. Kaloyeros and W. S. Williams, J. Am. Chem. Soc., 1987, 109, 1579–1580.
23. D. S. Belov, D. A. Fenoll, I. Chakraborty, X. Solans-Monfort and K. V. Bukhryakov, Organometallics, 2021, 40, 2939–2944.
24. P. J. Dyson, Dalton Trans., 2003, 2964–2974.
25. J. A. Widegren and R. G. Finke, J. Mol. Catal. A: Chem., 2003, 198, 317–341.
26. B. L. Tran, J. L. Fulton, J. C. Linehan, J. A. Lercher and R. M. Bullock, ACS Catal., 2018, 8, 8441–8449.
27. E. L. Muetterties, H. Schaeffer, R. I. Mink, M. Y. Darensbourg, M. Millar, T. Groshens and K. J. Klabunde, Inorg. Chem., 1979, 18, 883–884.
28. J. S. Yu, B. C. Ankianiec, I. P. Rothwell and M. T. Nguyen, J. Am. Chem. Soc., 1992, 114, 1927–1929.
29. M. C. Potyen and I. P. Rothwell, J. Chem. Soc., Chem. Commun., 1995, 849–851.
30. J. E. Ellis, Dalton Trans., 2019, 48, 9538–9563.

---

Footnotes

† Electronic supplementary information (ESI) available: Experimental details, calculations, spectra, temperature studies, catalyst reusability data, concentration/pressure dependence studies, isotopic labeling experiments, and GC-MS data. See DOI: https://doi.org/10.1039/d4cy01275e

‡ These authors contributed equally.

---