Molybdenum and tungsten complexes with carbon dioxide and ethylene ligands

Abstract

The first examples of stable metal complexes with coordinated ethylene and carbon dioxide ligands are reported. Reaction of tris(ethylene) complexes mer-M(C₂H₄)₃(PNP) (M = Mo and W; PNP = 2,6-bis(diphenylphosphinomethyl)pyridine) with CO₂ yields the corresponding, mixed cis-M(C₂H₄)₂(CO₂)(PNP) derivatives. X-ray studies reveal six-coordinate structures exhibiting η²-ethylene and κ²-C,O carbon dioxide coordination. Remarkably, the formation of the molybdenum CO₂ adduct occurs also in the solid state at room temperature, under 4 bar of CO₂, in a nearly quantitative manner.

---

Introduction

Since the twentieth century, organometallic chemistry has provided countless examples of molecules exhibiting two C-donor ligands possessing the potential to undergo metal-induced C–C coupling reactions.¹ Yet to be discovered are, however, stable transition metal compounds with coordinated ethylene and carbon dioxide, which could be relevant in the context of acrylate formation² by C–C bond formation. Notwithstanding decades of interest and intense research, catalytic conversion of C₂H₄ and CO₂ into acrylate derivatives with the aim of replacing current non-environmentally friendly methods is still an elusive goal.³ In 1983, Hoberg⁴ achieved nickelalactone formation by means of a Ni(0)-mediated coupling of C₂H₄ and CO₂ (Scheme 1a). This seminal discovery paved the way for the development of Ni- and Pd-catalyzed reactions, although productivity values were well below industrial needs.^5,6 Only two years after Hoberg's findings, our parent laboratory disclosed that tertiary phosphine stabilized bis(ethylene) complexes of Mo(0) and W(0) promoted stoichiometric acrylate formation in the presence of CO₂ under mild conditions (Scheme 1b).⁷ Though neither of the purported M(C₂H₄)(CO₂) adduct and metallalactone intermediates could be detected, metal–hydrido-acrylate end-products derived from the latter by β–H elimination were isolated and characterized. Lately, Bernskoetter and coworkers described⁸ a modified system alike in reactivity but bearing a tridentate pincer PPP ligand (Scheme 1c). Solution studies suggested the intermediacy of a five-coordinate, sixteen-electron species with cis C₂H₄ and CO₂ ligands, which could not be isolated.⁹ Although for the Mo and W systems CO₂ and C₂H₄ coordination appears to be a prerequisite for acrylate formation,¹⁰ this proposal is at doubt for Ni.¹¹ Related metallacycles have been reported for titanium,¹² zirconium,¹³ vanadium,¹⁴ rhodium,¹⁵ cobalt¹⁶ and iron¹⁷ but corresponding species with the metal coordinated to both C₂H₄ and CO₂ have escaped isolation. Chirik and coworkers have identified recently an iron-mediated C₂H₄–CO₂ coupling to carboxylates of variable chain length up to 21 carbon atoms, and although they demonstrated control of the reaction outcome by the metallacyclic intermediates, evidence for the Fe(C₂H₄)(CO₂) precursors was not provided.¹⁸

Scheme 1 Transition metal complexes with ethylene and carbon dioxide as co-ligands.

Inspired by the work¹⁹ of Nishibayashi, Beller or Kichner, among others, about the capacity of Mo(PNP) linkages to promote different stoichiometric or catalytic transformations with low reactive molecules, we envisioned replacing our original ligand scaffold based on monodentate phosphines⁷ with the simple, –PPh₂ bearing, pyridine-based pincer ligand 2,6-bis(diphenylphosphinomethyl)pyridine (from now on PNP in shorthand notation; see Scheme 1, bottom). Besides providing kinetic stability to M(0) complexes, the tridentate PNP unit would additionally thwart the mer-to-fac coordination change documented for the PPP tridentate phosphine ligand in Scheme 1c, and favor M(C₂H₄)(CO₂) adducts. Herein, we report that the said tridentate ligand allows the synthesis of tris(ethylene) complexes, mer-M(C₂H₄)₃(PNP) (M = Mo, 1a; W, 1b), as well as their conversion into the target mixed adducts, M(C₂H₄)₂(CO₂)(PNP) (M = Mo, 3a; W, 3b). Notably, formation of CO₂ adducts, 3, proceeds readily in solution and the solid state for Mo in a quantitative manner. In spite of the presence of C₂H₄ and CO₂cis ligands in the molecules of 3, these complexes are stable toward the formation of metallalactone and acrylate products. Computational support for the above observations is also reported.

Results and discussion

Synthesis of tris(ethylene) Mo(0) and W(0) complexes 1a and 1b

We found that sodium amalgam reduction of MCl₃(PNP) under ethylene, at room temperature, yielded the tris(ethylene) complexes mer-M(C₂H₄)₃(PNP) (M = Mo, 1a; W, 1b), which were isolated as dark red crystalline solids in good yields (Scheme 2). Attempts to prepare analogous complexes of PNP ligands with other –PR₂ termini (R = i-Pr, Cy and t-Bu) proved fruitless. Spectroscopic data are in good agreement with the proposed structure (see the ESI†). In particular, the C₂H₄ ligands of complexes 1 give rise to two ¹³C{¹H} NMR resonances at about 40 (two trans C₂H₄) and 49 ppm, both showing ²J_C–P values close to 5 Hz. Comparison with the δ value found for free C₂H₄ of 123 ppm reveals Δδ shifts of 83 and 74 ppm to lower frequencies, consistent with increased electronic density due to strong back-donation.²⁰ To our knowledge, complexes 1 are the first isolated tris(ethylene) complexes of molybdenum and tungsten, though Bernskoetter and co-workers detected in solution a species of this type when monitoring the reaction of a Mo(H)₄(PPP) complex with C₂H₄.⁹

Scheme 2 Synthesis of the tris(ethylene) complexes 1a and 1b.

X-ray crystallography²¹ led to unequivocal structure authentication (Fig. 1) and confirmed the meridional arrangement of the ethylene molecules. The two in trans are staggered with respect to one another, whereas the two carbon atoms of the third, which occupy the coordination site trans to the pyridine N atom, are essentially co-planar to the PNP donor atoms. No significant difference exists among the Mo–C_alkene distances to all C₂H₄ ligands, despite their diverse geometrical distribution. Thus, for 1a the Mo–C distances have values of 2.26–2.27 Å. Likewise, the three coordinated C=C bonds feature similar lengths, for instance 1.416(2) (C22–C23) and 1.425(2) Å (C24–C25). These metrical parameters, and those disclosed for 1b (see the ESI†), are significantly longer than the 1.34 Å distance corresponding to free C₂H₄,²² denoting the existence of strong electronic interactions between the electron-rich M(0) centers of complexes 1 and the π–C₂H₄ ligands.

Fig. 1 Solid-state structures of tris(ethylene) complexes 1a (top) and 1b (bottom). Hydrogen atoms have been omitted for clarity.

Reactivity of complexes 1a an 1b toward carbon monoxide

In contrast to results reported for trans-M(C₂H₄)₂(PMe₃)₄ complexes (M = Mo and W),²³ which are known to react with CO to give the mono-carbonyls trans,mer-M(C₂H₄)₂(CO)(PMe₃)₃, exposure of complexes 1 to an atmosphere of CO originated the trans-M(C₂H₄)(CO)₂(PNP) dicarbonyls (2a, Mo; 2b, W) as depicted in Scheme 3. The complexes were fully characterized by microanalysis, IR data [ν_asym(C–O) is 1807 and 1781 cm⁻¹ for 2a and 2b, respectively], NMR spectroscopy and X-ray studies²¹ (see Scheme 3). The exchange of the axial ethylene ligands instead of the equatorial one provides a singular feature of compounds 1 when compared to the related, previously described bis(ethylene) complexes of Mo and W. This difference in behavior must be a consequence of the use of the PNP ancillary ligand instead of the monodentate PMe₃ ligands, a fact that should also affect the reactivity toward carbon dioxide (see below).

Scheme 3 Synthesis of the dicarbonyls 2a and 2b and the molecular structure of 2a. See the ESI† for the structure of the tungsten analogue 2b. Hydrogens are omitted for clarity.

The mixed complexes M(C₂H₄)₂(CO₂)(PNP) (3a–3b)

When complexes 1 were submitted to a carbon dioxide atmosphere only one of the trans C₂H₄ ligands underwent ligand substitution upon exposure to CO₂ (1 bar, toluene solution). For molybdenum, CO₂ coordination took place at room temperature, but formation of the tungsten analogue needed heating at 80 °C for 4 h (Scheme 4). The new compounds were isolated as brown and yellow-brown crystalline solids, respectively, in 68% and 48% yields. With reference to the molybdenum complex 3a an IR absorption (nujol mull) centered at ca. 1700 cm⁻¹ can be taken as an indication of κ²-C,O coordination of the molecule of carbon dioxide.²⁴ Additional support for this proposal comes from the observation of a triplet ¹³C{¹H} resonance with δ 200.8 ppm and ²J_C–P = 14 Hz (Scheme 4), the two parameters resembling closely those reported for trans-Mo(CO₂)₂(PMe₃)₄.²⁵ The two ethylene ligands seem to undergo fast rotation around the respective Mo–C₂H₄ bond axis at room temperature, so that only two ¹³C{¹H} resonances are recorded (43.8, br s, and 50.6 ppm, t, ²J_C–P = 4 Hz). Although the two ³¹P nuclei of these complexes are equivalent and resonate at 46.9 ppm in the ³¹P{¹H} NMR (datum for 3a), the reduced symmetry brought about by the C₂H₄-by-CO₂ replacement becomes apparent from the observation of a ¹H AB spin system (δ_A = 4.75, δ_B = 4.65 ppm; ²J_AB = 18 Hz) for the now diastereotopic methylene protons of the PNP ligand. Signature ¹³C{¹H} resonances for the tungsten complex 3b were recorded at 210.2 (W–CO₂, t, ²J_C–P = 13 Hz), 45.4 (W–C₂H₄, t, ²J_C–P = 4 Hz) and 44.0 ppm (W–C₂H₄, br s), and are also comparable to literature data for related complexes (see the ESI† for full description).^23,26

Scheme 4 Synthetic route to M(C₂H₄)₂(CO₂)(PNP) (M = Mo, 3a; M = W, 3b) and region of their ¹³C{¹H} NMR spectra containing the CO₂ resonances.

The molecular complexity proposed for complexes 3 on the basis of microanalytical and spectroscopic data was confirmed by X-ray crystallography.²¹ These studies provided the solid-state molecular geometries presented in Fig. 2, which possess the two C₂H₄ ligands in a cis distribution, hence a cis,mer-M(C₂H₄)₂(CO₂)(PNP) formulation. The molecule of carbon dioxide is bound to the metal in a κ² fashion through one of the C=O bonds, in a staggered orientation with respect to the trans ethylene ligand. The coordinated C=O bond eclipses the M–N bond, while the two C=C bonds eclipse the trans P–M–P vector. The four Mo–C₂H₄ distances in 3a have similar magnitude, close to 2.25 Å. The Mo1–C1 bond within the Mo–CO₂ linkage is nonetheless significantly shorter at 2.107(4) Å, and it is also shorter than the associated Mo1–O2 separation of 2.215(3) Å. These metrics (see the ESI† for corresponding X-ray data for 3b) along with the elongation of the coordinated C=O bond relative to the non-coordinated one (C1–O2 = 1.261(4) and C1–O1 = 1.213(4) Å; 1.243(12) and 1.248(13) Å in the tungsten analog, 3b) hint at the existence of strong M–CO₂ electronic interactions in these complexes.

Fig. 2 The structures in the solid state of M(C₂H₄)₂(CO₂)(PNP) (M = Mo, 3a, left; M = W, 3b, right). Hydrogen atoms have been omitted for clarity.

Solid state ethylene replacement with carbon dioxide

Most notably, the reaction of crystalline samples of 1a with CO₂ occurred also in the solid state at room temperature. Loss of crystallinity and a color change to brown (Fig. 3) were observed after 24 hours under 4 bar of carbon dioxide. NMR analysis of the resulting powdery brown solid revealed clean conversion into complex 3a, though minor amounts (<2%) of unreacted 1a were still present. For the tungsten case, such mild conditions were not enough to promote the exchange. However, heating solid 1b under 4 bar of CO₂ for 24 h at 80 °C led to the formation of 3b in 20% yield, with 70% of 1b yet unreacted and 10% of free PNP ligand, formed upon decomposition, being also observed in the ³¹P{¹H} NMR spectrum (see the ESI†).

Fig. 3 Left: solid, crystalline sample of 1a before being submitted to 4 bar of CO₂. Right: the same sample after 12 h, containing >98% of complex 3a.

The observation of the formation of these compounds in the solid state constitutes an uncommon observation linking transition metal carbon dioxide chemistry and solid-state molecular organometallic chemistry. The latter is a highly important, though still underdeveloped field of research.²⁷

Stability of the complexes M(C₂H₄)₂(CO₂)(PNP) (3a,b)

The presence of both ethylene and carbon dioxide ligands in the coordination sphere of the M(C₂H₄)₂(CO₂)(PNP) complexes resembles the intermediate previously proposed (Scheme 1b) en route to their coupling, although in that case, which was also verified by DFT studies,¹⁰ the two ethylene ligands were in mutually trans disposition at variance with the cis geometry shown in 3a,b. Notwithstanding the presence of cis C₂H₄ and CO₂ ligands in the coordination polyhedron of complexes 3, the putative C–C coupling reaction leading to metallalactone or acrylate derivatives could not be ascertained. Heating toluene solutions of compounds 3 under argon, in the absence or in the presence of C₂H₄ and CO₂, with added base or with Lewis acids, resulted only in gradual decomposition with slow liberation of the PNP ligand, and no observation of acrylate formation.

Computational studies

To gain a deeper understanding of all the findings reported in this contribution, computational studies were undertaken. Calculations revealed that conversion of the tris(ethylene) molybdenum complex into the carbon dioxide adduct 3a is a facile process, in agreement with the already described experimental findings. Furthermore, two intermediate isomeric rotamers, i3a and i4a (Fig. 4), were identified and were shown to interconvert by means of a concerted synchronous rotation²⁸ of the trans CO₂ and C₂H₄ ligands. Not unexpectedly, the 1a-to-3a conversion requires the dissociation of one of the trans ethylene molecules of the 18-electron complex 1, and subsequent coordination of CO₂. The initially formed κ¹-O adduct isomerizes to the κ²-C,O rotamer i3a, which is only 2.2 kcal mol⁻¹ above the tris(ethylene) complex 1a, through TS2-3a, with energy just below 20 kcal mol⁻¹ above 1a (see Fig. 4). Rotamers i3a and i4a have similar energies but the experimentally observed rotamer 3a is about 5 kcal mol⁻¹ more stable, making the overall reaction exergonic by −3.1 kcal mol⁻¹ (toluene solvent). Similar ΔG values have been computed for the tungsten system (Scheme S2 and Fig. S2 in the ESI†), with the principal difference with respect to the molybdenum analogue being a significantly higher transition state TS2-3b at 27.1 kcal mol⁻¹ (also in toluene) relative to complex 1b. This is in excellent agreement with the slower formation of 3b, which requires heating at 80 °C for 4 h.

Fig. 4 Comparative ΔG profiles for the formation of CO₂-complex 3a and lactone-complex i6a in toluene.

In accordance with expectations, metallalactone product formation was computed to be endergonic (3.2 and 2.1 kcal mol⁻¹, in toluene, for the Mo and W systems, respectively). As shown in Fig. 4, two routes were calculated for the formation of the metallalactone species i6a starting from rotamers i3a and i4a. They are closely related to those described by the group of Papai for the M(C₂H₄)₂(PMe₃)₄-plus-CO₂ reactions¹⁰ (see the ESI†). A transition state for the purported formation of a hydrido-acrylate product from i6a could not be found, probably due to the rigidity of the Mo–PNP^19c and Mo–metallated lactone structures. The putative hydrido acrylate structure has been computed as lying 23.6 kcal mol⁻¹ above i6a (25.7 kcal mol⁻¹ for i6b).

Conclusions

In conclusion, a suitable choice of ancillary ligands has allowed for the isolation of the first stable complexes bearing both C₂H₄ and CO₂ ligands in the metal (Mo and W) coordination sphere. We have also shown that metallalactone formation by C–C coupling of the coordinated ligands is slightly endergonic, while the subsequent rearrangement to a hydrido-acrylate isomeric structure is strongly hindered kinetically by the rigidity of the M–PNP chelating structure, besides being thermodynamically uphill. We believe these results will be useful to design suitable catalytic systems for the conversion of C₂H₄ and CO₂ into acrylate derivatives.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We thank MINECO for support with Grants CTQ2017-82893-C2-1-R and CTQ2017-82893-C2-2-R and a FPI fellowship (MA). A. G. thanks the Centro de Servicios de Informática y Redes de Comunicaciones (CSIRC), Universidad de Granada, for providing the computing time. We thank Francisco Molina for X-ray structure determination.

Notes and references

1. J. F. Hartwig, Organotransition Metal Chemistry: From Bonding to Catalysis, University Science books, 2009.
2. (a) M. Limbach, Adv. Organomet. Chem., 2015, 63, 175–202; (b) X. Wang, H. Wang and Y. Sun, Chem, 2017, 3, 211–228.
3. H. J. Arpe, Industrial Organic Chemistry, Wiley-VCH, 5th edn, 2010.
4. (a) H. Hoberg and D. Schaefer, J. Organomet. Chem., 1983, 251, C51–C53; (b) H. Hoberg, Y. Peres, C. Krüger and Y.-H. Tsay, Angew. Chem., Int. Ed., 1987, 26, 771–773.
5. (a) N. Huguet, I. Jevtovikj, A. Gordillo, M. L. Lejkowski, R. Lindner, M. Bru, A. Y. Khalimon, F. Rominger, S. A. Schunk, P. Hofmann and M. Limbach, Chem.–Eur. J., 2014, 20, 16858–16862; (b) C. Hendriksen, E. A. Pidko, G. Yang, B. Schäffner and D. Vogt, Chem.–Eur. J., 2014, 20, 12037–12040.
6. (a) S. Mazini, A. Cadu, A.-C. Schimdt, N. Huguet, O. Trapp, R. Paciello and R. Schaub, ChemCatChem, 2017, 9, 2269–2274; (b) S. C. E. Stieber, N. Huguet, T. Kageyama, I. Jevtovikj, P. Ariyananda, A. Gordillo, S. A. Schunk, F. Rominger, P. Hofmann and M. Limbach, Chem. Commun., 2015, 51, 10907–10909.
7. (a) R. Álvarez, E. Carmona, D. J. Cole-Hamilton, A. Galindo, E. Gutiérrez-Puebla, A. Monge, M. L. Poveda and C. Ruíz, J. Am. Chem. Soc., 1985, 107, 5529–5531; (b) R. Álvarez, E. Carmona, A. Galindo, E. Gutiérrez-Puebla, J. M. Marín, A. Monge, M. L. Poveda, C. Ruíz and J. M. Savariault, Organometallics, 1989, 8, 2430–2439; (c) A. Galindo, A. Pastor, P. J. Pérez and E. Carmona, Organometallics, 1993, 12, 4443–4451; (d) C. Collazo, M. M. Conejo, A. Pastor and A. Galindo, Inorg. Chim. Acta, 1998, 272, 125–130.
8. W. H. Bernskoetter and B. T. Tyler, Organometallics, 2011, 30, 520–527.
9. Y. Zhang, B. S. Hanna, A. Dineen, P. G. Williard and W. H. Bernskoetter, Organometallics, 2013, 32, 3969–3979.
10. G. Schubert and I. Papai, J. Am. Chem. Soc., 2003, 125, 14847–14858.
11. M. Al-Ghamdi, S. V. C. Vummaleti, L. Falivene, F. A. Pasha and D. J. Beetstra, Organometallics, 2017, 36, 1107–1112.
12. S. A. Cohen and J. E. Bercaw, Organometallics, 1985, 4, 1006–1014.
13. H. G. Alt and C. E. Denner, J. Organomet. Chem., 1990, 390, 53–60.
14. B. Hessen, A. Meetsma, F. van Bolhuis, J. H. Teuben, G. Helgesson and S. Jagner, Organometallics, 1990, 9, 1925–1936.
15. M. Aresta and E. Quaranta, J. Organomet. Chem., 1993, 463, 215–221.
16. I. Knopf, M.-A. Courtemanche and C. C. Cummins, Organometallics, 2017, 36, 4834–4843.
17. H. Hoberg, K. Jenni, K. Angermund and C. Krüger, Angew. Chem., Int. Ed., 1987, 26, 153–155.
18. S. M. Rummelt, H. Zhong, I. Korobkov and P. J. Chirik, J. Am. Chem. Soc., 2018, 140, 11589–11593.
19. (a) K. Arashiba, Y. Miyake and Y. Nishibayashi, Nat. Chem., 2011, 3, 120–125; (b) T. Leischner, A. Spannenberg, K. Junge and M. Beller, Organometallics, 2018, 37, 4402–4408; (c) O. Öztopcu, C. Holzhacker, M. Puchberger, M. Weil, K. Mereiter, L. F. Veiros and K. Kirchner, Organometallics, 2013, 32, 3042–3052.
20. L. Cavallo, A. Macchioni, C. Zuccaccia, D. Zuccaccia, I. Irabona and F. Ruffo, Organometallics, 2004, 23, 2137–2145.
21. CCDC 1911725 (1a), 1911727 (1b), 1911730 (2a), 1911733 (2b), 1911735 (3a), and 1911736 (3b) contain the supplementary crystallographic data for this paper.†.
22. H. J. Bernstein, Trans. Faraday Soc., 1961, 57, 1649–1656.
23. (a) E. Carmona, J. M. Marín, M. L. Poveda, J. L. Atwood and R. D. Rogers, J. Am. Chem. Soc., 1983, 105, 3014–3022; (b) E. Carmona, A. Galindo, M. L. Poveda and R. D. Rogers, Inorg. Chem., 1985, 24, 4033–4039.
24. A. Pastor, F. Montilla and A. Galindo, Adv. Organomet. Chem., 2017, 68, 1–91.
25. (a) R. Álvarez, E. Carmona, M. L. Poveda and R. Sánchez-Delgado, J. Am. Chem. Soc., 1984, 106, 2731–2732; (b) R. Álvarez, E. Carmona, E. Gutiérrez-Puebla, J. M. Marín, A. Monge and M. L. Poveda, J. Chem. Soc., Chem. Commun., 1984, 1326–1327.
26. Related concerted synchronous rotation of CO₂ groups was demonstrated in trans,mer-Mo(CO₂)₂(PMe₃)₃(CNR) and related compounds. See: E. Carmona, A. K. Hughes, M. A. Muñoz, D. M. O'Hare, P. J. Perez and M. L. Poveda, J. Am. Chem. Soc., 1991, 113, 9210–9218.
27. S. D. Pike and A. S. Weller, Philos. Trans. R. Soc., A, 2015, 373, 20140187.
28. Computed barriers for ethylene rotation in 3a are lower than 15.4 kcal mol⁻¹.

---

Footnote

† Electronic supplementary information (ESI) available: Synthesis and characterization of new compounds, crystallographic data and DFT calculations. CCDC 1911725, 1911727, 1911730, 1911733, 1911735 and 1911736. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c9sc03225h

---