Synthesis and carbene-transfer reactivity of dimeric nickel carbene cations supported by N-heterocyclic carbene ligands

Abstract

Dimeric nickel bridging-carbene complex salts [{(IPr)Ni}₂(μ-Cl)(μ-CR¹R²)][B(Ar^F)₄](6, R¹ = R² = Ph; 7, R¹ = H, R² = SiMe₃) are synthesized by reaction of {(IPr)Ni(μ-Cl)}₂ (3) with NaB(Ar^F)₄ and N₂CPh₂ or N₂CHSiMe₃, with elimination of N₂ and NaCl. The solid-state structure of 6 features an unsymmetric μ,η³-bonding motif for the bridging diphenylcarbene ligand in which one Ni center binds to the CPh₂ ligand in a π-fashion involving the ipso- and one ortho-carbon of a phenyl ring along with the carbene carbon. The solution structure (NMR) of 6 indicates a C_2v-symmetric structure even at −80 °C. The solid-state structure of 7 shows a (trimethylsilyl)carbene unit symmetrically disposed between the two nickel centers and bound through only the carbene-carbon atom. Diphenylcarbene-group transfer from 6 to carbon monoxide (3 equiv) gives O=C=CPh₂ and the dimeric Ni(I)-Ni(I) carbonyl complex [{(IPr)Ni(CO)}₂(μ-Cl)][B(Ar^F)₄] (8). Excess pivaloisocyanide reacts with 6 to afford ^tBuN=C=CPh₂, [(IPr)Ni(CN^tBu)₃][B(Ar^F)₄] (9), and (IPr)NiCl(CN^tBu) (10). Mesitylazide reacts with 6 to give the ketimine MesN=CPh₂ and the bridging mesitylimido dimer [{(IPr)Ni}₂(μ-Cl)(μ-NMes)][B(Ar^F)₄] (5). A secondary reaction of diphenyldiazomethane with 6 to give Ph₂C=CPh₂ and Ph₂C=N–N=CPh₂ prevents catalytic CPh₂-group transfer reactions from being realized in these systems.

---

Introduction

Transition metal carbene complexes are important reagents and intermediates in many synthetically useful reactions, including olefin metathesis and cyclopropanation.¹ We have reported the isolation and characterization of a monomeric nickel carbene, (dtbpe)Ni=CPh₂ (1, dtbpe = 1,2-bis(di-tert-butylphosphino)ethane), from (dtbpe)Ni(cod) and diphenyl–diazomethane (N₂CPh₂) (cod = 1,5-cyclooctadiene; Scheme 1).²Group transfer reactions from 1 to form cyclopropane, cyclopropene, ketimine, ketene, and ketone products have been demonstrated.^3,4 However, carbene-transfer catalysis with the dtbpe system is very limited (a few turnovers for the cyclopropanation of 1-hexene) due to azo-coupling between 1 and diphenyldiazomethane to form tetraphenylazine, Ph₂C=N–N=CPh₂.^3b In related chemistry, imido complexes (dtbpe)Ni=NR (2, Scheme 1) can be prepared from the appropriate organoazide (N₃R) and {(dtbpe)Ni}₂(C₆H₆), and undergo similar stoichiometric “nitrene” transfer reactions with appropriate substrates to give aziridine, carbodiimide, isocyanate, and diazene derivatives.^3a,4Catalytic nitrene transfer from 2 is also limited by the reaction of 2 with N₃R to give diazenes (RN=NR) and dinitrogen at faster rates than for nitrene-transfer to other substrates.⁴

Scheme 1 Synthesis of 1 from N₂CPh₂ and (dtbpe)Ni(cod) with N₂ elimination. Related imido complexes (2) can be prepared from organoazides. Mes = 2,4,6-trimethylphenyl, 1-Ad = 1-adamantyl.

The undesirable reaction between nickel-imido moieties and N₃R noted above can be obviated by using bimetallic frameworks containing N-heterocyclic carbene ancillary ligands.⁵ We have shown that the Ni(I)-Ni(I) dimer {(IPr)Ni(μ-Cl)}₂ (3, IPr = 1,3-(2,6-ⁱPr₂C₆H₃)₂imidazolin-2-ylidene) reacts with excess N₃Mes to form the Ni(II)-Ni(II) bridging imido complex {(IPr)NiCl}₂(μ-NMes) (4).⁵ Removal of a chloride ligand from 4 gives the complex salt [{(IPr)Ni}₂(μ-Cl)(μ-NMes)][B(Ar^F)₄] (5; B(Ar^F)₄ = B{3,5-(CF₃)₂C₆H₃)}₄; Scheme 2). Importantly, neither 4 nor 5 react further with N₃Mes, and are competent catalysts for the formation of carbodiimides (MesN=C=NR) and mesityl isocyanate (MesN=C=O) from N₃Mes and isocyanides or carbon monoxide, respectively (illustrated for carbodiimide synthesis in Scheme 2).

Scheme 2 Catalytic carbodiimide synthesis with a NHC-supported dimer 4. Catalytic mesityl isocyanate formation can be accomplished with 5. A⁻ = B(Ar^F)₄⁻ = B{3,5-(CF₃)₂C₆H₃)}₄⁻. (See ref. 5)

We were intrigued by the possibility that a similar synthetic strategy using 3 and diazoalkanes might afford access to dimeric complexes with bridging carbene ligands that could participate in group-transfer reactions. Herein we describe the synthesis and characterization of two bridging dinickel carbene complexes from 3 with diazoalkanes serving as the carbene synthons. The reactivity of the new carbene complexes and their potential catalytic efficacy in group-transfer reactions are evaluated and compared with the structurally analogous bridging-imido complexes.

Results and discussion

Carbene syntheses and structures

In an attempt to access a carbene analogue of the bridging imido complex 4, we investigated the reaction of the Ni(I) chloride dimer 3 with diphenyldiazomethane. Stoichiometric (1 : 1) addition of N₂CPh₂ to diethyl ether solutions of 3 resulted in vigorous evolution of N₂ and darkening of the solution. A ¹H-NMR spectrum of the crude reaction mixture showed formation of multiple products (along with unreacted 3; see discussion below), and while resonances consistent with a product of the formulation “{(IPr)Ni(Cl)}₂(μ-CPh₂)” were observed, attempts to isolate the neutral carbene dimer by selective crystallization were unsuccessful. In contrast, addition of one equivalent of N₂CPh₂ to a stirred diethyl ether solution containing equimolar amounts of 3 and NaB(Ar^F)₄ resulted in N₂ evolution and clean formation of the diphenylcarbene complex salt [{(IPr)Ni}₂(μ-Cl)(μ-CPh₂)][B(Ar^F)₄] (6) as forest-green crystals in 89% isolated yield (Scheme 3). Chloride abstraction with NaB(Ar^F)₄ facilitated removal of neutral contaminants in a simple manner.

Scheme 3 Synthesis of bridging, cationic carbene complexes 6 and 7 from diazoalkanes.

The diphenylcarbene complex 6 is diamagnetic and exhibits a ¹H-NMR spectrum consistent with C_2v symmetry at both 195 K and room temperature. The bridging carbene-carbon of 6 resonates at δ 274 in the ¹³C-NMR spectrum (cf., δ 222 for the terminal carbene ligand in 1),² and significantly downfield from the NHC-carbon attached to Ni (δ 177). The solid-state structure of 3, however, shows an unsymmetrical dimer in which the carbene functionality adopts η³-coordination involving the ipso- and one ortho-carbon of a phenyl ring along with the carbene carbon (see Fig. 1, Table 1). It is noteworthy that the Ni(2)–C(50) bond length is close to the value expected for a Ni=C double bond (cf., 1.836(2) Å in 1) while the Ni(1)–C(50) bond is significantly longer at 1.958(3) Å, and is closer to distances found for Ni–C single bonds in cationic nickel alkyls like [(dtbpe)Ni(CH₂CMe₂Ph)][PF₆] with Ni–C = 1.954(3) Å.⁶

Fig. 1 Solid-state structures of the complex cations of 6 (top) and 7 (bottom) drawn with 50% thermal ellipsoids. H-atoms (except on the carbene carbon), the B(Ar^F)₄ anions, IPr iso-propyl groups (for 6), and co-crystallized solvent have been omitted for clarity.

Table 1 Relevant bond distances for 6 and 7

6			7
Atom 1	Atom 2	Dist./Å	Atom 1	Atom 2	Dist./Å
Ni(1)	Ni(2)	2.4312(10)	Ni(1)	Ni(2)	2.4385(9)
Ni(1)	C(50)	1.958(3)	Ni(1)	C(50)	1.849(3)
Ni(2)	C(50)	1.831(5)	Ni(2)	C(50)	1.858(3)
Ni(1)	C(1)	1.975(5)	C(50)	Si(1)	1.882(3)
Ni(1)	C(51)	2.134(5)
Ni(1)	C(52)	2.368(5)
C(50)	C(511)	1.474(7)

---

Reaction of N₂CHSiMe₃ with 3 in the presence of NaB(Ar^F)₄ affords purple, crystalline [{(IPr)Ni}₂(μ-Cl)(μ-CHSiMe₃)][B(Ar^F)₄] (7) in 91% isolated yield (Scheme 3). The highly deshielded carbene carbon of 7 resonates at δ 294 in the ¹³C-NMR spectrum, and the unique proton at δ 14.27 in the ¹H-NMR spectrum. Carmona has reported several bridging Ni carbenes with “Ni₂(μ-CHR)” cores that are formed via α–H elimination from a cyclometallated neophyl ligand in (PMe₃)₂Ni(CH₂CMe₂-o-C₆H₄), including examples with phosphine and cyclopentadienyl ancillary ligands.⁷ For comparison, Carmona's complexes exhibit bridging carbene-ligand resonances at δ 160–167 and δ 3.0–4.6 in their ¹³C- and ¹H-NMR spectra, respectively.

A single-crystal X-ray study of 7 reveals a symmetrically bridging trimethylsilylcarbene ligand with Ni–C bond distances of 1.849(3) and 1.858(3) Å. The Ni–Ni distance of 7 is nearly identical to that in 6 despite the difference in carbene coordination mode (2.4312(10) vs. 2.4389(9) Å, respectively). Ni(1), Ni(2), Cl, and the carbene carbon are coplanar.

Polynuclear Ni complexes containing unsaturated carbyl fragments are relatively rare, primarily represented by bridging carbynes (L_nNi_xCR),⁸ in addition to the aforementioned Carmona carbenes. The majority of these carbynes are trimeric species supported by CpNi subunits, exemplified by (CpNi)₃(μ₃-CCH₃), a product of the reaction of excess LiCH₃ with Cp₂Ni.^8e Heteronuclear trimetallic complexes have also been prepared by the reaction of Cp*W(CO)₂(CR) with {(Cp)₂Ni(μ-CO)}₂, giving (CpNi)₂Cp*W(μ₃-CR)(CO)₄.^8c,d Kubiak has reported that N-protonation or alkylation of bridging isocyanide ligands affords [Ni₂(μ-CNMeH)(CNMe)₂(dppm)₂]⁺ (dppm = bis(diphenylphosphino)methane) featuring bridging aminocarbyne moieties.⁹

Reactivity and group-transfer studies

In contrast to the tolerance of the imido dimers 4 and 5 toward excess mesitylazide (see above), both 3 and 6 readily convert excess N₂CPh₂ to a ∼ 1 : 1 mixture of Ph₂C=CPh₂ and Ph₂C–N=N–CPh₂ with elimination of dinitrogen (Scheme 4).^10e Previous reports of both free and transition-metal-stabilized carbene and diazoalkane decomposition note the corresponding azine (in this case, Ph₂C=N–N=CPh₂) as the major product with the corresponding olefin (here, Ph₂C=CPh₂) often constituting a minor component.¹⁰

Scheme 4 Reaction of N₂CPh₂ with 3 or 6 forms tetraphenylethylene and benzophenonediazine in a ∼ 1 : 1 product ratio.

Reactions of 6 with mesityl azide, carbon monoxide, and t-butyl isocyanide were explored to elucidate its carbene group-transfer capabilities (see Schemes 4–6). It is noteworthy that Warren has reported related bridging dicopper carbenes displaying similar metric parameters which effect facile cyclopropenation and addition to suitable nucleophiles.¹¹ As shown in Scheme 4, reaction of 6 with N₃Mes results in its conversion to 5 with elimination of the ketimine Ph₂C=NMes, isolated in 80% yield. The reaction is slow, occurring over a period of three weeks at room temperature in the presence of excess N₃Mes (CD₂Cl₂). Attempts to increase the reaction rate by heating to 60 °C resulted in rapid decomposition of 6. ¹H-NMR studies of the room temperature reaction of 6 and N₃Mes show the only nickel-containing product to be 5. We have reported on carbene–imide and carbene–oxo coupling with the nickel diphenylcarbene complex 1.⁴ Computational and kinetic studies indicate the formation of a 1,3-dipolar cycloaddition intermediate prior to N₂ release. The slow reaction rate of 6 is possibly a consequence of steric inhibition of a related cycloaddition intermediate. Alternatively, the poor nucleophilicity of N₃Mes may inhibit coordination to the sterically hindered Ni centers and slow formation of bridging μ-NMes or μ-N₃Mes intermediates.

Scheme 5 Combination of excess CO and 6 leads to formation of diphenylketene and the Ni(I)-Ni(I) dicarbonyl salt 8.

Scheme 6 Addition of excess CN^tBu to 6 forms Ph₂C=C=N^tBu with dissociation of the dimeric metal complex to Ni(I) monomers.

Diphenylketene is readily formed by carbene-group transfer in the reaction of 6 with carbon monoxide. Exposure of 6 to an excess of CO results in clean conversion to free Ph₂C=C=O and the diamagnetic, bridging Ni(I)-Ni(I) carbonyl dimer [{(IPr)Ni(CO)}₂(μ-Cl)][B(Ar^F)₄] (8, Scheme 5) as a purple salt. If the reaction is carried out in a 1 : 1 stoichiometry (6:CO), diphenylketene is formed in only ∼30% yield along with 8 and unreacted 6. Complex salt 8 is stable in the presence of 1 atm CO, and disproportionation to Ni(0) and Ni(II) is not observed. This is in contrast to the reaction of excess CO with 3 that results in disproportionation to the Ni(0) carbonyl complex (IPr)Ni(CO)₃ and the Ni(II) dimer {(IPr)Ni(Cl)}₂(μ-Cl)₂.⁵ The solid-state structure of 8 (Fig. 2, Table 2) shows a symmetrical dimer with a carbonyl ligand binding in a κ¹ fashion to each Ni(I) center. The Ni(1)–Ni(2) distance in 8 at 2.4967(5) Å is only slightly longer than the corresponding values found in 6 and 7. The carbonyl ligands bind to nickel in linear manners (Ni–C–O = 179.5(3) and 176.6(3)°) with C–O bond distances of 1.134(3) and 1.145(3) Å that indicate limited back-bonding. The IR data (ν_CO = 2063, 2023 cm⁻¹) corroborate minimal activation of the carbonyl ligands in this cationic complex. For comparison, previously isolated examples of dimeric Ni(I) carbonyls exhibit ν_CO typically ranging from 1775–2020 cm⁻¹.¹²

Fig. 2 Solid-state structure of the complex cation of 8 drawn with 50% thermal ellipsoids. H-atoms, co-crystallized solvent, and the B(Ar^F)₄ counterion have been omitted for clarity.

Table 2 Relevant bond distances for 8 and 9

8			9
Atom 1	Atom 2	Dist./Å	Atom 1	Atom 2	Dist/Å
Ni(1)	Ni(2)	2.4967(5)	Ni(1)	C(41)	1.916(3)
Ni(1)	C(1)	1.943(3)	Ni(1)	C(42)	1.906(3)
Ni(2)	C(2)	1.958(3)	Ni(1)	C(43)	1.919(3)
Ni(1)	C(51)	1.762(3)	C(41)	N(41)	1.152(4)
Ni(2)	C(52)	1.754(3)	C(42)	N(42)	1.164(4)
C(51)	O(51)	1.134(3)	C(43)	N(43)	1.160(4)
C(52)	O(52)	1.145(3)	Ni(1)	C(1)	1.988(3)

---

Addition of an excess of tert-butyl isocyanide to a solution of 6 gives the carbene-coupling product Ph₂C=C=N^tBu in 76% isolated yield (Scheme 6). Although alkyl isocyanides often give metal products isostructural with the CO analogues, the fate of the Ni(I) fragment in this reaction differs from 8, undergoing dimer dissociation to give the mononuclear Ni(I) salt [(IPr)Ni(CN^tBu)₃][B(Ar^F)₄] (9) and the known 3-coordinate Ni(I) complex (IPr)Ni(CN^tBu)(Cl) (10).⁵ Complex 9 is a paramagnetic pale-yellow salt with μ_eff = 1.9 μ_B (Evans). The X-ray structure of 9 (Fig. 3, Table 2) shows a distorted tetrahedral Ni(I) center with three linear isocyanide ligands (Ni(1)-C(42)-N(42) = 169.4(3), Ni(1)-C(43)-N(43) = 176.1(3), Ni(1)-C(41)-N(41) = 169.2(3)°). When three equivalents of CN^tBu are used, a mixture of 6, 9, and 10 is obtained (and not the dimeric pivaloisocyanide analogue of 8). The difference in the rate of reaction between CO, CN^tBu, and N₃Mes may be due to the ability of CO and CN^tBu to readily coordinate to the Ni centers.

Fig. 3 Solid-state structure of the complex cation of 9 drawn with 50% thermal ellipsoids. H-atoms, co-crystallized solvent, and the B(Ar^F)₄ counterion have been omitted for clarity.

Conclusions

Cationic, bridging dinickel carbene salts of the type [{(IPr)Ni}₂(μ-Cl)(μ-CR¹R²)][B(Ar^F)₄] (6, R¹ = R² = Ph; 7, R¹ = H, R² = SiMe₃) can be isolated in good yields from reactions of the Ni(I)-Ni(I) dimer {(IPr)Ni(μ-Cl)}₂ (3) with diazoalkanes in the presence of NaB(Ar^F)₄. Carbene transfer from 6 to the nitrene fragment “NMes” (from N₃Mes), CO, and CN^tBu was found to occur stoichiometrically to form Ph₂C=NMes, Ph₂C=C=O, and Ph₂C=C=N^tBu, respectively. Exposure of 6 to excess carbon monoxide leads to the formation a Ni(I)-Ni(I) carbonyl dimer [{(IPr)Ni(CO)}₂(μ-Cl)][B(Ar^F)₄] (8) that is stable with respect to dimer dissociation. The related substrate CN^tBu, however, effects dimer dissociation in its reaction with 6 with formation of the monomeric Ni(I) complexes [(IPr)Ni(CN^tBu)₃][B(Ar^F)₄] (9) and (IPr)Ni(Cl)(CN^tBu) (10).

Attempts to isolate the neutral carbene variant, {(IPr)Ni(Cl)}₂(μ-CR¹R²), were stymied by the competitive, facile decompostion of the diazoalkane substrate initiated by either 3 or the product carbene. This catalyticdiazoalkane decomposition (to the azine) prevents catalytic turnover for the stoichiometric carbene-transfer reactions we have explored since the substrate is readily consumed before productive carbene-transfer occurs.

Acknowledgements

This work was generously supported by the US National Science Foundation through Grant CHE-0957816 to G.L.H.

Notes and references

1. (a) E. O. Fisher and A. Maasböl, Angew. Chem., Int. Ed. Engl., 1964, 3, 580; (b) R. R. Schrock, Reactions of Coordinated Ligands; Braterman, P. S., Ed.; Plenum Press: New York, 1986; (c) K. H. Dötz, Transition Metal Carbene Complexes; Verlag Chemie: Weinheim, Germany, 1983; (d) R. R. Schrock, Angew. Chem., Int. Ed., 2006, 45, 3748; (e) R. H. Grubbs, Angew. Chem., Int. Ed., 2006, 45, 3760.
2. D. J. Mindiola and G. L. Hillhouse, J. Am. Chem. Soc., 2002, 124, 9976.
3. (a) D. J. Mindiola and G. L. Hillhouse, Chem. Commun., 2002, 1840; (b) R. Waterman and G. L. Hillhouse, J. Am. Chem. Soc., 2003, 125, 13350.
4. N. D. Harrold, R. Waterman and G. L. Hillhouse, J. Am. Chem. Soc., 2009, 131, 12872.
5. C. A. Laskowski and G. L. Hillhouse, Organometallics, 2009, 28, 6114.
6. K. D. Kitiachvili, D. J. Mindiola and G. L. Hillhouse, J. Am. Chem. Soc., 2004, 126, 10554.
7. T. R. Balderrain, E. Gutierrez, A. Monge, M. C. Nicasio, M. Paneque, M. L. Poveda and E. Carmona, Organometallics, 1993, 12, 4431.
8. Representative examples: (a) H. Beurich, R. Blumhofer and H. Vahrenkamp, Chem. Ber., 1982, 115, 2409; (b) R. Blumhofer, K. Fischer and H. Vahrenkamp, Chem. Ber., 1986, 119, 194; (c) G. P. Elliot, J. A. Howard, T. Mise, I. Moore, C. M. Nunn and F. G. A. Stone, J. Chem. Soc., Dalton Trans., 1986, 2091; (d) E. Delgado, J. Hein, J. C. Jeffery, A. L. Ratermann, F. G. A. Stone and L. J. Farrugia, J. Chem. Soc., Dalton Trans., 1987, 1191; (e) H. Lehmkuhl, C. Krüger, S. Pasynkiewicz and J. Poplawska, Organometallics, 1988, 7, 2038.
9. D. L. DeLaet, P. E. Fanwick and C. P. Kubiak, Organometallics, 1986, 5, 1807.
10. (a) W. E. Parham and W. R. Hasek, J. Am. Chem. Soc., 1954, 76, 935; (b) C. Ruchardt and G. N. Schrauzer, Chem. Ber., 1960, 93, 1840; (c) K. Ishiguro, Y. Sawaki and H. Iwamura, Chem. Lett., 1987, 9, 1853; (d) D. Griller, M. Majewski, W. G. McGimpsey, A. S. Nazran and J. C. Scaiano, J. Org. Chem., 1988, 53, 1550; (e) P. Hofmann, I. V. Shishkov and F. Rominger, Inorg. Chem., 2008, 47, 11755.
11. Y. M. Badiei and T. M. Warren, J. Organomet. Chem., 2005, 690, 5989.
12. (a) R. A. Jones, A. L. Stuart, J. L. Atwood and W. E. Hunter, Organometallics, 1983, 2, 874; (b) Z. Zhang, H. Wang, H. Wang, R. Wang, W. Zhao and L. Yang, J. Organomet. Chem., 1988, 347, 269; (c) D. J. Elliot, D. G. Holah, A. N. Hughes, H. A. Mirza and E. Zawada, J. Chem. Soc., Chem. Commun., 1990, 32.

---

Footnote

† Electronic supplementary information (ESI) available: All experimental protocols and complete crystallographic details for 6, 7, 8, and 9 (CIF) and data regarding the crystallographic refinements (PDF). CCDC reference numbers 791319–791322. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c0sc00464b

---