On the incidence of non-covalent intramolecular interligand interactions on the conformation of carbene complexes: a case study

Abstract

The prevalence of a counterintuitive carbene conformation in a series of piano-stool Mn(I) alkylalkoxycarbene complexes was rationalized by means of DFT calculations and high-resolution XRD analysis in terms of stabilizing non-covalent C–H⋯C≡O interligand interactions.

---

Seminal MO calculations by Hoffmann et al.¹ have predicted that carbene ligands in piano-stool complexes of the type Cp(CO)₂M = C(R¹)R² should adopt the so-called vertical coordination mode (A or B, Fig. 1), rather than the horizontal one (C, Fig. 1),² so as to maximize the overlap between the frontier orbitals of the metal fragment and of the carbene ligand. This situation is actually found in the solid-state structure of all Mn(I) piano-stool alkoxycarbene complexes reported so far^3–5 except one.⁶ Early theoretical (SCF)⁷ and structural “normal resolution” XRD^3b studies of the Cp(CO)₂Mn=C(OR)Ph (R = Me or Et) complexes have, however, revealed some discrepancies between the predicted and observed rotamers.

Fig. 1 Vertical vs. horizontal coordination mode of carbene ligand in piano-stool complexes Cp(CO)₂M = C(R¹)R².

In this context and in line with our previous investigations on the recurrent incidence of non-covalent interligand interactions on the conformation of carbene ligands in their complexes,⁸ we were prompted to reinvestigate the structure of Mn(I) alkylalkoxycarbene complexes by means of DFT calculations and high-resolution XRD analysis.

Initial DFT calculations using the dispersion-corrected M06L functional (M06L/def2-TZVPP level)^† on the model complex Cp(CO)₂Mn=C(OMe)Me revealed five possible minima on the potential energy surface: 1-A_anti, 1-A_syn, and 1-B_syn where the carbene ligand adopts a vertical coordination mode, 1-C_anti where it sits in a horizontal coordination mode, and one minimum, we shall call eclipsed, 1-D_anti, very close in energy to 1-C_anti, in which the carbene moiety plane and one CO ligand appear to be coplanar (Fig. 2). Gratifyingly, the most stable conformer was found to be a vertical one, in agreement with the above-mentioned seminal studies,¹ and appeared to be the one in which the methoxy group is proximal to the carbonyl ligands and points toward them in an anti conformation, 1-A_anti.⁹ Such an orientation might be considered as counterintuitive for steric reasons as it brings two hydrogen atoms of the methoxy substituent in the vicinity of the carbonyl ligands, i.e. the computed H1_OMe⋯C1_CO and H2_OMe⋯C2_CO distances being relatively short (2.423 Å and 2.542 Å, respectively), far below the sum of the VdW radii of the corresponding atoms (2.9 Å). Yet, our calculations suggest that this conformation is the result of attractive interactions. Indeed, application of the second-order perturbation theory (SOPT) of the Natural Bond Orbital (NBO)^† method revealed the occurrence of stabilizing two-electron delocalizations from both σ-(C–H_OMe) molecular orbitals (MO) of the methoxy group into the π*-(C≡O) MO of the closest carbonyl ligand, coupled with a reverse donation from the π-(C≡O) to the vacant σ*-(C–H_OMe) (Fig. 3). The computed associated SOPT energies (ΔE⁽²⁾ of −0.95, −0.82, and −0.68 kcal mol⁻¹ for the σ(C–H1) → π*-(C1≡O), σ(C–H2) → π*-(C2≡O), and π-(C1≡O) → σ*(C–H1) delocalizations, respectively)¹⁰ indicate that these interactions are weak yet significant and in part responsible for the anti conformational preference.

Fig. 2 M06L/def2-TZVPP optimized geometries for the possible conformers of complex 1 (bond lengths in Å and relative energies (E_rel) in kcal mol⁻¹).

Fig. 3 NBO-molecular orbitals responsible for the C–H⋯C≡O stabilization stabilizing interactions in conformer 1-A_antiantianti.

Remarkably, a literature search revealed that nine³ over a total of eleven^3,5,6 piano-stool Mn(I) alkoxycarbene complexes possessing methyl or methylene hydrogen atoms disposed to develop a similar interaction exhibit an A_anti conformation, whereas those lacking such hydrogen atoms invariably display a B_syn conformation.⁴ This suggests that the possibility of building C–H⋯C≡O interligand interactions may play a key role regarding the conformation of the former complexes in the solid-state.

Being aware by experience that the archetypal Mn(I) alkoxycarbene complex MeCp(CO)₂Mn=C(OEt)Me (2), displaying an A_anti conformation, could be obtained as high quality single crystals, and considering that the DFT calculations did predict in 2 the occurrence of similar stabilizing C–H⋯CO interligand interactions,¹¹ we seized the opportunity to visualize them experimentally by examining the topology of the electron density (ED) distribution ρ(r) established by high-resolution XRD in the framework of Bader's “Atoms In Molecules” (AIM) theory.¹²

An accurate high-resolution XRD data set was thus collected on 2 at 100 K up to sin θ/λ < 1.1 Å⁻¹ (MoK_α).^†,‡ The experimental ED distribution ρ(r) was derived from the XRD data using the MoPro multipolar refinement suite,¹³ whereas its subsequent topological analysis was achieved using VMoPro¹³ and WinXPro.¹⁴ For the sake of comparison, the topological analysis of the theoretical ED based on the DFT gas-phase optimized geometry of 2 was also carried out using AIMAll.¹⁵

The carbene ligand in 2 effectively exhibits the expected A_anti conformation ({cnt–Mn1–C3–O3} = 172.4°; Mn1–C3–O3–C5 = −3.4°), which, as in the model complex 1, brings the methylene hydrogen atoms of the ethoxy substituent in the vicinity of the carbonyl ligands (Fig. 4 (left)). Noticeably, the multipolar refinement leads to short and dissymmetrical H5a⋯C1 and H5b⋯C2 distances, which are well reproduced in the gas-phase DFT optimized structure (Table 1), thus suggesting that crystal packing has little influence, if any, on the position of the CH₂ group relative to the CO ligands.

Fig. 4 Perspective view (left; 50% probability level) and experimental molecular graph (right; blue lines are bp's, red sphere are bcp's, yellow cubes are rcp's, and purple diamonds are ccp's) for complex 2.

Table 1 Values of the experimental and theoretical^a (italic) orthodox parameters^b at the intramolecular interligand bond critical points

Contact	d	ρ(r)	∇^2ρ(r)	G	V	E int
a ED computation (M06L/def2-TZVPP) based on the gas-phase optimized geometry (M06L/def2-TZVPP). b d (Å), ρ(r) (e Å^−3), ∇^2ρ(r) (e Å^−5), G (hartree Å^−3), V (hartree Å^−3), and Eint (kcal mol^−1) are the interatomic distance, the ED, the Laplacian of the ED, kinetic, potential, electron energy densities, and interaction energy, respectively.
H5a⋯C1	2.41	0.08	0.97	0.06	−0.05	2.1
	2.422	0.08	0.95	0.06	−0.05	2.2
H5b⋯C2	2.53	0.09	1.07	0.07	−0.05	2.5
	2.515	0.08	0.99	0.06	−0.05	2.2

---

The molecular graph resulting from the topological analysis of the experimental ED in 2 (Fig. 4 (right)) reveals that all intraligand and metal-to-ligand bond critical points (bcp's) and bond paths (bp's) that can be anticipated on the basis of chemical intuition are present,^† except for the (η⁵-MeCp)Mn unit, which displays a [3, 3, 1 + 1] topology – using Farrugia's terminology¹⁶ – instead of the expected [5, 5, 1 + 1] one.¹⁷ Remarkably, the molecular graph also reveals two pairs of extra intramolecular interligand bcp's located approximately at mid distance between H5a and C1, and H5b and C2, and between H4a, and H12 and H13. The topological analysis shows that these bcp's are associated with H5a⋯C1, H5b⋯C2, H4a⋯H13, and C4⋯H12 bp's. This brings experimental support to the occurrence of stabilizing interactions between the corresponding atom pairs or triad, and in particular, between the methylene hydrogen atoms and the proximal carbonyl carbon atoms, in agreement with the results of the DFT/NBO approach. The orthodox parameters associated with the interligand bcp's are gathered in Table 1. The relatively low values of ρ(r), and positive values of the Laplacian ∇²ρ(r) and the electron energy density H(ρ) at the bcp's, denote weak interactions of the closed-shell type.¹⁸ Their interaction energies, E_int, were thus estimated using the Espinosa, Molins and Lecomte correlation scheme based on the semi-quantitative relationship between E_int and the value of the potential energy density V(ρ) at the bcp's.^19,20 The E_int was found to be 2.1 kcal·mol⁻¹ and 2.5 kcal·mol⁻¹, for the CH₂/CO interactions, respectively (Table 1), and roughly half of those values, 1.2 kcal·mol⁻¹ and 1.1 kcal·mol⁻¹, for MeCp/Me ones.^† The topological analysis of the experimental ED also revealed several intermolecular bcp's associated with H⋯O, and H⋯H bcp's whose associated interaction energies range from 0.2 to 1.0 kcal·mol⁻¹.^† Naturally, these intermolecular interactions not only participate in the cohesion of the crystal, but they also necessarily act to some extent on the conformation of the molecules constitutive of the crystal cell, in a delicate balance with the intramolecular interactions. Yet, among all the weak interactions detected, the intramolecular ones between the methylene group and the carbonyl ligands clearly remain the strongest ones, which reinforces the idea that they do play a determining structuring role.

Farrugia et al. have recently demonstrated that bcp's and bp's associated with very weak interaction may actually be the result of artefacts inherent to the Hansen–Coppens multipole model,²¹ which is used to determine the experimental ED distribution. In this respect, it is worth noting that the theoretical molecular graph for 2 globally matches the experimental one^† (in spite of two notable differences²²) providing evidence for the existence of two bcp's between H5a and C1, and H5b and C2, respectively, while the topological parameters at these points are in excellent agreement with the experimental ones (Table 1).

In conclusion, the case study presented here supports growing evidence that, beyond the observation of molecular structures obtained from normal resolution XRD, a conjunction of matching high-resolution XRD (experimental) analyses and DFT (theoretical) calculations is very helpful to rationalize the occurrence of weak interactions,²³ including intramolecular interligand interactions,⁸ having possible effects on the conformation of organometallic complexes bearing flexible ligands like carbenes.

Acknowledgements

The authors thank financial support from the PICS grant 4873 (CNRS)/09-03-91060 (RFBR), Spanish MICINN and CAM (Grants CTQ2010-20714-CO2-01/BQU, Consolider-Ingenio 2010, CSD2007-00006, S2009/PPQ-1634, and Ramón y Cajal contract to I.F.). CALMIP (France) is acknowledged for HPC resources (Grant p1006). R.B. thanks the French MESR for a PhD fellowship.

Notes and references

1. B. E. R. Schilling, R. Hoffmann and D. L. Lichtenberger, J. Am. Chem. Soc., 1979, 101, 585.
2. K. G. Caulton, Coord. Chem. Rev., 1981, 38, 1.
3. (a) S. Fontana, U. Schubert and E. O. Fischer, J. Organomet. Chem., 1978, 146, 39; (b) U. Schubert, Organometallics, 1982, 1, 1085; (c) C. S. Yi, G. L. Geoffroy, C. A. White and R. L. Rheingold, J. Am. Chem. Soc., 1993, 115, 3806; (d) Y. Yong, C. Jiabi, W. Xiaoying, W. Qiangjin and L. Qiutan, J. Organomet. Chem., 1996, 516, 81; (e) C. Mongin, N. Lugan and R. Mathieu, Organometallics, 1997, 16, 3873; (f) R. Streubel, M. Hobbold, J. Jeske and P. G. Jones, J. Organomet. Chem., 2000, 595, 12; (g) A. Rabier, N. Lugan and R. Mathieu, J. Organomet. Chem., 2001, 617–618, 681.
4. (a) C. Mongin, K. Gruet, N. Lugan and R. Mathieu, Tetrahedron Lett., 2000, 41, 7341; (b) E. O. Fischer, J. Chen and U. Schubert, Z. Naturforsch., B: Anorg. Chem. Org. Chem., 1982, 37, 1284; (c) D. A. Valyaev, M. G. Peterleitner, L. I. Leont'eva, L. N. Novikova, O. V. Semeikin, V. N. Khrustalev, M. Yu. Antipin, N. A. Ustynyuk, B. W. Skelton and A. H. White, Organometallics, 2003, 22, 5491.
5. H. Fischer, K. Weissenbach, C. Karl and A. Geyer, Eur. J. Inorg. Chem., 1998, 339.
6. For the only exception of a horizontal conformation mode see: M. Landman, W. Barnard, P. H. Van Rooyen and D. C. Liles, J. Mol. Struct., 2012, 1021, 76.
7. N. M. Kostić and R. F. Fenske, J. Am. Chem. Soc., 1982, 104, 3879.
8. (a) D. A. Valyaev, R. Brousses, N. Lugan, I. Fernández and M. A. Sierra, Chem.–Eur. J., 2011, 17, 6602; (b) I. Fernández, N. Lugan and G. Lavigne, Organometallics, 2012, 31, 1155.
9. (a) I. Fernández, F. P. Cossío, A. Arrieta, B. Lecea, M. J. Mancheño and M. A. Sierra, Organometallics, 2004, 23, 1065; (b) D. M. Andrada, M. E. Zoloff Michoff, I. Fernández, A. M. Granados and M. A. Sierra, Organometallics, 2007, 26, 5854.
10. The corresponding π-(C2≡O) → σ*(C–H2) delocalization was not evidenced by the SOPT of the NBO method (ΔE⁽²⁾ < −0.5 kcal mol⁻¹).
11. Computed associated SOPT energies ΔE⁽²⁾ of −0.95, −0.87, and −0.71 kcal mol⁻¹ for the σ(C–H5a) → π*-(C1≡O), σ(C–H5b) → π*-(C2≡O), and π-(C1≡O) → σ*(C–H5a) delocalizations.
12. R. F. W. Bader, Atoms in Molecules: A Quantum Theory, Clarendon Press, Oxford, 1990.
13. C. Jelsch, B. Guillot, A. Lagoutte and C. Lecomte, J. Appl. Crystallogr., 2004, 38, 38.
14. A. Stash and V. J. Tsirelson, J. Appl. Crystallogr., 2002, 35, 371.
15. T. A. Keith, AIMAll, aim.tkgristmill.com, 2012.
16. L. Farrugia, C. Evans, D. Lentz and M. Roemert, J. Am. Chem. Soc., 2009, 131, 1251.
17. (a) The topology of π-carbocyclic/transition metal complexes arrangement has been studied in detail and is out of the focus of the present communication. See ref. 16 and references therein, and ref. 17b; ; (b) A. O. Borissova, M. Y. Antipin and K. A. Lyssenko, J. Phys. Chem. A, 2009, 113, 10845.
18. U. Koch and P. L. A. Popelier, J. Phys. Chem., 1995, 99, 9747.
19. (a) E. Espinosa, E. Molins and C. Lecomte, Chem. Phys. Lett., 1998, 285, 170; (b) E. Espinosa, I. Alkorta, I. Rozas, J. Elguero and E. Molins, Chem. Phys. Lett., 2001, 336, 457.
20. (a) E _int = −313.754 V, where E_int is the interaction energy (in kcal mol⁻¹); ; (b) Yu. V. Nelyubina, M. Yu Antipin and K. A. Lyssenko, Russ. Chem. Rev., 2010, 79, 167.
21. L. J. Farrugia and H. M. Senn, J. Phys. Chem. A, 2012, 116, 738.
22. First, the (η⁵-MeCp)Mn unit here shows the expected [5, 5, 1 + 1] topology and second, no bcp's could be evidenced in the median Me/MeCp region, probably in correlation with the slightly different orientation of the methyl group in the gas-phase-optimized molecule (Table 1).
23. (a) P. M. Dominiak, E. Espinosa and J. G. Ángyán, Intermolecular interaction energies from experimental charge density studies, in Modern Charge-Density Analysis, ed. C. Gatti and P. Macchi, 2012, Springer, London, pp. 387–433; (b) For a recent perspective review see: K. A. Lyssenko, Mendeleev Commun., 2012, 22, 1 and references therein.

---

Footnotes

† Electronic supplementary information (ESI) available: DFT computational details, details on the multipolar refinement for 2. CCDC 907309. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c2dt32554c

‡ Crystal data for 2: C₁₂H₁₅MnO₃, M = 262.175; monoclinic, P2₁/c, a = 11.1515(1) Å, b = 7.8616(1) Å, c = 14.1583(2) Å, β = 109.389(7)°, V = 1170.84(6) ų, T = 100(1) K, Z = 4, F₀₀₀ = 544; D_c = 1.486 g cm⁻³; μ = 1.114 mm⁻¹. 168 793 Bragg reflections were used up to sin θ/λ = 1.1 Å⁻¹. Final multipolar refinement of 488 parameters against 11 120 observed reflections (I > 3σ(I)) converged to R = 0.0130 and R_w = 0.0176, GOF = 1.13. CCDC 907309.

---