Lithium-promoted hydrogenation of carbon dioxide to formates by heterobimetallic hydridozinc alkoxide clusters

Abstract

The remarkably distinct reactivity of hydridozinc heterobimetallic cubanes [(HZnO^tBu)_4−n(thf·LiO^tBu)_n] 1a–1d towards CO₂ is reported—the hydride transfer from Zn–H to CO₂ is drastically accelerated in the presence of Li ions in 1b–1d which led to the respective metal formate hydrates; the systems are inspiring models for the selective conversion of water gas into formates on lithium-promoted ZnO supports.

---

Carbon dioxide (CO₂) is an abundant yet low-value carbon source of enormous impact in Nature. However, the extraordinarily high stability of CO₂ has hampered its utilization as efficient C₁-building block for large-scale industrial syntheses of useful organic compounds. Thus, current efforts in developing efficient catalytic processes that exploit CO₂ as a source for valuable organic products belong to one of the most ambiguous challenges in industrial chemistry.¹ Of particular interest is the reductive transformation of CO₂ in the presence of H₂ into renewable sources such as methanol (MeOH).² The world demand for MeOH is currently enormously increasing because of its role as a precursor for many useful organic chemicals (e.g., formaldehyde, acetic acid) and as a substitute for fuels.³ This can only be managed by an efficient large-scale industrial process. In fact, MeOH synthesis is very efficiently achieved by the heterogeneously catalyzed conversion of syngas (CO, H₂) and water gas (CO₂, H₂), respectively, on heterometal-promoted ZnO carriers. Accordingly, the most commonly commercialized unit for the production of MeOH is the low-temperature ICI process, which converts a high-pressure gas mixture of CO, CO₂ and H₂ into MeOH at 250–300 °C, using Cu-promoted ZnO which is dispersed on alumina.⁴ As expected, the mechanisms of the heterogeneously catalyzed reduction of carbon oxides depends sensitively on the nature of the support material (i.e., composition, particle size, structure, defects) and the presence of promoters (e.g., Cu). Notably, also Cu-free pure ZnO model systems show considerable catalytic activities in methanol synthesis under particular circumstances. The catalytic activity of pure ZnO supports does not increase linearly with increasing BET surface area but requires the presence of polar ZnO facets⁵ and oxygen-vacancies as particularly important active sites.⁶ Recently, it has also been shown that the catalytic activity of pure ZnO supports for using feed gas mixtures containing CO and H₂ correlate with the amount of oxygen-vacancies, whereas CO₂ has a poisoning effect presumably because it quenches oxygen defect sites.⁷ Apparently, the exothermic conversion of CO₂ with H₂ (water gas) to methanol on ZnO supports, according to eqn (1), requires a different mechanism.

CO₂ + 3H₂→ H₃COH + H₂O Δ_RH = −40.9 kJ mol⁻¹ (1)

In line with that, five reactions on ZnO surfaces (eqn (2)–(6)) have been proposed in the literature, which play a crucial role for the conversion of water gas to methanol.⁸

H₂(g) → 2 H (ad) (2)

CO₂(g) → CO₂ (ad) (3)

CO₂(ad) + H(ad) → HCO₂(ad) (4)

HCO₂(ad) + 4H(ad) → H₃CO(ad) + H₂O (5)

H₃CO(ad) + H(ad) → H₃COH(ad) → H₃COH(g) (6)

The initial step of the catalytic process is the hydrogenation of ZnO, leading to surface-terminated ZnH and OH sites⁹ which indicate the heterolytic fission of dihydrogen. It has also been shown for a few cases that alkali metals can promote the catalytic performance of ZnO.¹⁰ On the molecular level, however, little is known about the consecutive mechanism of chemisorption and reduction steps. This lack of knowledge could be partially overcome by using hydridozinc alkoxides as molecular models which resemble some electronic features of hydrogenated ZnO. Additionally, the promoting influence of heterometals could be mimicked by using heterobimetallic hydridozinc clusters (e.g., mixed lithium hydridozinc alkoxides). Recently, we reported the synthesis and structure of a series of well-defined homo- and heterobimetallic hydridozinc tert-butoxide clusters of the formula [(HZnOtBu)_4−n(LiOtBu)_n] 1a–1d¹¹ (Scheme 1) which could be useful to study Zn–H assisted and heterometal promoted hydrogenation of CO₂.

Scheme 1 The homometallic cubane 1avs. heterobimetallic lithium zinc tert-butoxide cubanes 1a–1d.

Here we report the remarkably different reactivity of the lithium-free vs. lithium-containing hydridozinc heterocubane-like clusters 1avs.1b–1d. The latter show the pivotal role of lithium ions for an accelerated reduction of CO₂ at Zn–H sites at ambient temperature and atmospheric pressure to give selectively zinc formate.

The reaction progress for the consumption of CO₂ has been monitored by in situ IR measurements. In contrast to ZnH₂ which remains unchanged even after one day in a pure CO₂ atmosphere, the powdered hydridozinc heterocubane cluster 1a reacts slowly with CO₂ to give additional vibrational bands in the region of 1330, 1600–1650, 1798–1813 and 2700–2800 cm⁻¹ in the IR spectrum which can be assigned to the formate ion in different coordination modes.¹² The consumption of 1a is complete after ca. three days. The formation of formate is similar to the process of insertion of CO₂ into the Zn–H bond of a hydridozinc tris(pyrazole)borate described by Parkin et al. which enabled the isolation of an η¹-formate zinc complex.¹³ Interestingly, the insertion of CO₂ into the Zn–H bond is drastically accelerated in the presence of Li ions in the heterocubane framework: thus, powders of 1b react immediately with CO₂ as shown by IR measurements (Fig. 1).

Fig. 1 Insitu IR spectra of the gradual conversion of 1b with CO₂. (a) Vibrational bands between 1900 and 1500 cm⁻¹. (b) Vibrational modes between 1450 and 1250 cm⁻¹.

The characteristic Zn–H vibrational mode of 1b at 1769 cm⁻¹ decreases gradually during the reaction progress. Concomitantly, an additional band at 1805 cm⁻¹ is growing in and additional new bands appear at 1336, 1630–1674, and 2700–2800 cm⁻¹. Furthermore, the appearance of a broad OH stretching vibration in the region of 3200–3500 cm⁻¹ suggests the formation of water which is coordinated to zinc and/or lithium formate. In fact, the vibrational modes can be unequivocally assigned to zinc formate–hydrate species based on characteristic reference data¹³ and confirmed by quantum chemical calculations.¹⁴ From the literature it is known that the formate ligand can adopt the η¹- and η²-coordination mode, respectively. We considered in our DFT calculations three different coordination modes for the formate ligand: the η¹-, symmetric η²- and asymmetric η²-coordination modes, respectively. As expected, the η²-coordination modes are favoured with a slight preference for the asymmetric one. Our calculations suggest that both coordination modes contribute to the vibrational spectra of the formate species obtained by the reaction of 1b and CO₂. This is in accordance with the MAS ¹³C-NMR spectrum of the product which shows two resonances for the HCO₂ moiety at δ = 168 and 169 ppm. Furthermore, powder-XRD studies confirm the presence of Zn(HCO₂)₂ and Li(HCO₂) hydrates as microcrystalline components of the reaction mixture. Other hydrogenated products of CO₂ such as formaldehyde and methanol or CO could not be detected (IR, NMR, MS). The experiments have been performed under rigorous anhydrous and anaerobic conditions. Accordingly, using the deuterated isotopomer of 1b, [(DZnO^tBu)₃(thf·LiO^tBu)], leads to [Zn(DCO₂)₂·2D₂O] as proven by IR. Remarkably, contamination of CO₂ with water vapour (0.05–0.1%) prevents hydrogenation of CO₂ and leads merely to partial hydrolysis of Zn–H bonds in 1b and sorption of CO₂ to give carbonates exclusively. The water molecules of [Zn(HCO₂)₂·2H₂O] are presumably formed by secondary reduction of formate to give as yet unidentified reduction products. Apparently, the formation of water necessitates the presence of lithium ions, since the Li-free cluster 1a leads exclusively to anhydrous zinc formate. The formation of water during the hydrogenation of CO₂ on ZnO supports has previously been discussed by Bailey et al..¹⁵ In the latter case, however, formate species are also formed as initial products at elevated temperature via hydride transfer apart from other hydrogenation products (e.g., methanolate) and water. Interestingly, the formation of Zn(HCO₂)₂ and Li(HCO₂) hydrates occurs also in solution. Thus, clear solutions of 1b in thf react with CO₂ leading immediately to precipitation of [Zn(HCO₂)₂·2H₂O] which has been characterized by IR, ¹H- and ¹³C-NMR spectroscopy. Crystals of the latter have been characterized by a single-crystal X-ray diffraction analysis (Fig. 2). The crystal structure¹⁶ consists of a coordination polymer with two octahedrally coordinated Zn ions linked by a formate anion as a bridging bidentate ligand.¹⁷

Fig. 2 Structural unit of polymeric [Zn(HCO₂)₂·2H₂O] in the crystal. The Zn2-atom coordinates four water molecules.

The two Zn ions are in different environments and lie on independent inversion centres: while Zn1 is coordinated to six formate ligands, Zn2 is surrounded by two formate ligands and four water molecules. As proven by ¹H-NMR spectroscopy of the filtrate, compound 1b has been completely consumed and neither another Zn–H compound nor formate species remain in solution. Instead, resonance signals of as yet unknown metal tert-butoxide aggregates and uncharacterized organic side-products can be observed.

The heterobimetallic cubanes 1c (ratio Li : Zn = 2 : 2) and 1d (ratio Li : Zn = 3 : 1) react also with CO₂ which has been monitored by in situ IR spectroscopy. Fig. 3 shows the changes of selected characteristic vibrational modes for the gradual conversion of 1c which are practically identical with those of 1d. However, reaction progress is significantly slower than that for the monolithium cluster 1b. While the conversion of 1b with CO₂ is complete after ca. 5 min, it takes ca. 30 min to consume the same molar quantity of 1c and 1d, respectively, affording [Zn(HCO₂)₂·2H₂O] and Li(HCO₂) hydrates as major products.

Fig. 3 In situ IR spectra of the gradual conversion of 1c with CO₂. (a) Vibrational bands between 1250 and 1450 cm⁻¹. (b) Vibrational modes between 1500 and 1900 cm⁻¹.

The distinct reactivity of 1bvs. 1c and 1d suggests that hydride transfer from the Zn–H bond to CO₂ is significantly reduced by increasing the molar ratio of Li : Zn. In line with that, the relatively low reactivity of 1a indicates that the presence of at least one Li ion as a Lewis-acidic centre in proximity to the Zn–H moiety fosters the hydride transfer to CO₂.

On the other hand, increasing the Li : Zn ratio reduces the basicity of the Zn–H moiety due to a stronger O → Li vs. O → Zn coordination. In conclusion, our model systems demonstrate the pivotal role of Li ions for an accelerated reduction of CO₂ at Zn–H sites. Although, the mechanism for the accelerated reduction of CO₂ through the presence of Li ions is still unknown, our preliminary results on the model systems 1a–1d suggest that the selective conversion of water gas (hydrogenation of CO₂) into formic acid derivatives (e.g., formic acid methylester) could be strongly favoured by using lithium-promoted ZnO supports. In line with our previous results on synthesizing nanoscaled zinc oxide materials through the organometallic precursor approach,¹⁸1b–1d are promising molecular single-source precursors for the synthesis of Li-promoted, nanoscaled ZnO materials. Respective investigations on the synthesis and catalytic performance of Li-promoted ZnO nanoparticles for the selective catalytic conversion of water gas to formic acid derivatives are currently underway.

Notes and references

1. Review: H. Arakawa, M. Aresta, J. N. Armor, M. A. Barteau, E. J. Beckman, A. T. Bell, J. E. Bercaw, C. Creutz, E. Dinjus, D. A. Dixon, K. Domen, D. L. DuBois, J. Eckert, E. Fujita, D. H. Gibson, W. A. Goddard, D. W. Goodman, J. Keller, G. J. Kubas, H. H. Kung, J. E. Lyons, L. E. Manzer, T. J. Marks, K. Morokuma, K. M. Nicholas, R. Periana, L. Que, J. Rostrup-Nielson, W. M. H. Sachtler, L. D. Schmidt, A. Sen, G. A. Somorjai, P. C. Stair, B. R. Stults and W. Tumas, Chem. Rev., 2001, 101, 953–996 and references therein.
2. (a) G. C. Chinchen, K. C. Waugh and D. A. Whan, Appl. Catal., 1986, 25, 101; (b) M. Kurtz, H. Wilmer, T. Genger, O. Hinrichsen and M. Muhler, Catal. Lett., 2003, 86, 77.
3. (a) M. S. Spencer, Catal. Lett., 1989, 50, 37–40; (b) G. A. Olah, Angew. Chem., 2005, 117, 2692 (Angew. Chem., Int. Ed., 2005, 44, 2636).
4. Review: L. Lloyd, D. E. Ridler and M. V. Twigg, in Catalyst Handbook, ed. M. V. Twigg, Wolfe, London, 1989.
5. H. Wilmer, M. Kurtz, K. V. Klementiev, O. P. Tkatschenko, W. Grünert, O. Hinrichsen, A. Birkner, S. Rabe, K. Merz, M. Driess, C. Wöll and M. Muhler, Phys. Chem. Chem. Phys., 2003, 5, 4736.
6. (a) S. A. French, A. A. Sokol, S. T. Bromley, C. R. A. Catlow, S. C. Rogers, F. King and P. Sherwood, Angew. Chem., Int. Ed., 2001, 40, 4437; (b) S. A. French, A. A. Sokol, S. T. Bromley, C. R. A. Catlow and P. Sherwood, Top. Catal., 2003, 24, 161.
7. S. Polarz, J. Strunk, V. Ischenko, M. W. E. Van den Berg, O. Hinrichsen, M. Muhler and M. Driess, Angew. Chem., Int. Ed., 2006, 45, 2965–2969.
8. I. A. Fisher and A. T. Bell, J. Catal., 1997, 172, 222.
9. (a) G. Ghiotti, A. Chioino and F. Boccuzzi, Surf. Sci., 1993, 297, 228–234; (b) J. C. Lavalley, S. Saussey and T. Rais, J. Mol. Catal., 1982, 17, 289–298.
10. (a) D. M. Minahan, W. S. Epling and G. B. Hoflundy, J. Catal., 1998, 179, 241–257; (b) J. G. Nunan, C. E. Bogdan, K. Klier, K. J. Smith, C. W. Young and R. G. Herman, J. Catal., 1989, 116(1), 195–221; (c) P. Winiarek and J. Kijenski, J. Chem. Soc., Faraday Trans., 1998, 94(1), 167–172; (d) A. Szabo, Prog. Catal., 2000, 9(1–2), 65–72.
11. W. Marchiniak, K. Merz, M. Moreno and M. Driess, Organometallics, 2006, 25, 4931–4933.
12. D. G. Rethwisch and J. A. Dumesic, Langmuir, 1986, 2, 73–79.
13. R. Han, I. B. Gorell, A. G. Looney and G. Parkin, J. Chem. Soc., Chem. Commun., 1991, 717.
14. R. Ahlrichs, M. Bär, M. Häser, H. Horn and C. Kölmel, Chem. Phys. Lett., 1989, 162, 165–169.
15. S. Bailey, G. F. Froment, J. W. Snoeck and K. C. Waugh, Catal. Lett., 1994, 30, 99.
16. Crystal data for [Zn(HCO₂)₂·2H₂O] see ESI.† Crystal data for C₂H₆O₆Zn, M_r = 191.44, monoclinic, space group P2₁/c (No. 14), a = 8.6803(9), b = 7.1241(10), c = 9.3060(12) Å, β = 97.664(3)°, V = 570.3(1) ų, ρ_calcd = 2.229 g cm⁻³, μ = 4.265 mm⁻¹, Z = 4, λ = 0.71073 Å, T = 213 K, 209 reflections collected (±h, ±k, ±l), [(Θ range: 3.71 to 25.02], 926 independent (R_int=0.019) and 734 observed reflections [I > 2σ(I)], 110 refined parameters, R = 0.022, wR₂ = 0.062. CCDC 650772. For crystallographic data in CIF or other electronic format see DOI: 10.1039/b714806b.
17. A. S. Lipton, M. D. Smith, R. D. Adams and P. D. Ellis, J. Am. Chem. Soc., 2002, 124, 410–414.
18. (a) K. Merz, R. Schoenen and M. Driess, J. Phys. IV, 2001, 11, 467; (b) M. Driess, K. Merz, R. Schoenen, R. Rabe, F. E. Kruis, A. Roy and A. Birkner, C. R. Chim., 2003, 6(3), 273; (c) J. Hambrock, S. Rabe, K. Merz, A. Wohlfarth, A. Birkner, R. A. Fischer and M. Driess, J. Mater. Chem., 2003, 13, 1731; (d) M. Kurtz, N. Bauer, C. Buescher, H. Wilmer, O. Hinrichsen, R. Becker, S. Rabe, K. Merz, M. Driess, R. A. Fischer and M. Muhler, Catal. Lett., 2004, 92, 49; (e) S. Polarz, A. Roy, M. Merz, S. Halm, D. Schröder, L. Schneider, G. Bacher, F. E. Kruis and M. Driess, Small, 2005, 1, 2; (f) V. Ischenko, S. Polarz, D. Grote, V. Stavarache, K. Fink and M. Driess, Adv. Funct. Mat., 2005, 15, 1945; (g) D. Schröder, H. Schwarz, S. Polarz and M. Driess, Phys. Chem. Chem. Phys., 2005, 7, 1049.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental procedure, characterisation of [Zn(HCO₂)₂·2H₂O] and DFT calculations of IR vibrations. See DOI: 10.1039/b714806b

---