Synthesis and Characterization of Di- and Tetracarbene Iron(II) Complexes with Chelating N-Heterocyclic Carbene Ligands and Their Application in Aryl Grignard−Alkyl Halide Cross-Coupling

A series of new and known bis(imidazolium) chloride and bromide salts bridged by either a methylene group (1−8, 10a,b) or an ethylene group (9a,b) and bearing different N substituents (Me, Et, Bn, tBu, Mes) have been reacted with [Fe{N(SiMe3)2}2]2 to yield the four-coordinate iron(II) complexes [LFeX2] (11−20; X = Cl, Br; L = chelating bis(imidazolylidene) ligand). Molecular structures of six of these complexes have been characterized by X-ray crystallography, and selected examples have been characterized by H NMR and UV−vis spectroscopy, cyclic voltammetry, Mo ̈ssbauer spectroscopy, and SQUID magnetometry. In all cases the iron(II) is found in a distorted-tetrahedral environment; it is in the high-spin state and shows large quadrupole splittings in the range 3.67−4.03 mm·s−1 (δ = 0.73−0.81 mm·s−1). Subtleties of the metric parameters depend on the bridging unit between the two imidazolylidene groups, the peripheral N substituents, and the coligand (Cl or Br). In case of rather small (Me, Et) or flexible (Bn) N substituents the dicarbene species [LFeX2] are formed together with ferrous tetracarbene complexes [L2FeX2] (21−23), which are difficult to separate and could not be isolated in pure form. When the latter are dissolved in MeCN in the presence of residual [FeBr2(solv)x], however, they transform into the ionic complexes [L2Fe(MeCN)2][FeBr4] (24−26), which have been characterized by single-crystal X-ray diffraction. They feature lowspin iron(II) (Mössbauer parameters δ ≈ 0.15 mm s−1, ΔEQ ≈ 1.36 mm s−1) and distorted-octahedral structures with the two MeCN ligands in a cis configuration. Selected examples of the new dicarbene complexes [LFeX2] have been tested as catalysts for the standard cross-coupling reaction between p-tolylmagnesium bromide and bromoor chlorocyclohexane. They show moderate activity that appears to be generally lower than for related complexes with two monodentate NHC ligands, but the activities clearly depend on the peripheral N substituents and the linker between the two imidazolylidene groups; the best results are obtained for complex 19, which features a long ethylene bridge and bulky Mes substituents, and hence the most shielded metal center. ■ INTRODUCTION N-heterocyclic carbenes (NHCs) have gained great popularity as ligands in organometallic chemistry and are finding widespread use in homogeneous catalysis. They are usually viewed as strongly σ-donating ligands with little or negligible π back-bonding. Chemical and substitutional inertness is often mentioned among their particular advantages, and this is expected to be further amplified by incorporating the NHC moiety into a chelating scaffold. Mainly in the past decade a variety of iron complexes with monodentate NHC ligands and several NHC ligated iron clusters have been published. Some of the mononuclear iron NHC complexes as well as species generated in situ from imidazolium salts and suitable iron precursors showed high activities in homogeneous catalysis. However, only a few iron complexes with chelating bisor tris(imidazolylidene) ligands are known so far. These comprise iron complexes with tridentate 2,6-bis(imidazolylidene)pyridine, tris(imidazolylidene)borate, and tris(imidazolylidene)triethylamine as well as methylenebridged bis(imidazolylidene) ligands. The latter have also been employed in model systems for [FeFe]-hydrogenases. More generally, bidentate methyleneor ethylene-bridged bis(imidazolylidene) ligands are quite popular scaffolds for transition-metal catalysts, especially in palladium chemistry. Among the reasons are the simple preparation of the respective bis(imidazolium) salts, the precursors of NHC ligands, and the relatively high stability of the bis(imidazolylidene) species. The synthesis of NHC complexes is usually performed in situ by deprotonation of the imidazolium salts and subsequent treatment with metal complexes (Scheme 1; route A), by direct reaction of the imidazolium salt with metal compounds containing basic ligands (route B), or by cyclization of metal-bound isocyanides (route C). Almost all iron NHC complexes reported so far have been prepared via route A, while palladium NHC complexes are often prepared according to route B by the use of Pd(OAc)2. 16 In contrast, Fe(OAc)2 has not yet been successfully used for NHC complex synthesis, and our attempts have also failed. Received: September 15, 2011 Published: November 17, 2011 Article pubs.acs.org/Organometallics © 2011 American Chemical Society 6692 dx.doi.org/10.1021/om200870w |Organometallics 2011, 30, 6692−6702 Some iron NHC complex syntheses have been reported using [Fe{N(SiMe3)2}2]2 or [Cp*FeN(SiMe3)2] for imidazolylidene formation and complexation. Since the only side product, hexamethyldisilazane, is rather unreactive and can be removed by evaporation, and since [Fe{N(SiMe3)2}2]2 is readily accessible, 17 we have chosen route B for the present study. The bis-chelate NHC iron(II) dihalide complexes presented herein are related to monodentate NHC iron(II) dihalide complexes initially reported by Grubbs et al. and recently investigated by Deng et al. (Scheme 2), which in both cases were synthesized via route A. We report the synthesis and characterization of several new iron(II) bis-chelate NHC complexes that are accessible via in situ carbene generation using the respective bis(imidazolium) salts and [Fe{N(SiMe3)2}2]2. Depending on the ligand substituents, either dicarbene iron complexes or mixtures of dicarbene and tetracarbene iron complexes are formed. For each type of complexes several representative examples have been analyzed by X-ray diffraction. The catalytic activity of selected dicarbene iron complexes is examined in a benchmark aryl Grignard−alkyl halide cross-coupling reaction. During the final stages of the preparation of this paper, similar dicarbene complexes were reported by Zlatogorsky et al. ■ RESULTS AND DISCUSSION Synthesis of Bis(imidazolium) Salts 1−10. The preparation of bis(imidazolium) salts was performed in an ACS glass autoclave by stirring the reactants for 1−3 days at 110 °C (DCM) or 130 °C (THF) in analogy to the procedure reported by Scherg et al., although yields of the known salts 1−3, 5, 6a, 7a and 8a−10a were in most cases lower via this route (Table 1). Reaction conditions and workup of the chloride salts 6b−10b as well as the phenyl-substituted bromide salt 4 have been slightly modified compared to the literature procedure (see the Experimental Section). Prior to further use the bis(imidazolium) salts were evaporated to absolute dryness in high vacuum (10−3 mbar) at 100−150 °C (i.e., below the melting point). The absence of residual H2O was checked by H NMR spectroscopy. This procedure is necessary, because most bis(imidazolium) salts are hygroscopic and [Fe{N(SiMe3)2}2]2, which is used in the following step, is highly water and oxygen sensitive. Bis(imidazolium) salts 4, 6b, 7b and 10b have not been reported before and have been fully characterized. 4 was analyzed by X-ray diffraction, since no molecular structure of any CHPh-bridged bis(imidazolium) salt has been published yet. Colorless single crystals were obtained by slow diffusion of Et2O into a MeCN solution. The molecular structure of 4, which crystallizes in the monoclinic space group P21/n, is shown in Figure S3 (Supporting Information). The angle formed by the two imidazolium rings is 72.9°; the rings are bridged by C4 with an N1−C4−N3 angle of 106.9°. The C1/5−N distances of the imidazolium rings lie within a range of 1.32−1.34 Å with N−C−N angles of 107.7° (N3−C5−N4) and 108.4° (N1−C1−N2). The protons H1, H4, and H5 form rather short C−H···Br hydrogen bonds (2.57−2.82 Å) to the counteranions; these distances are in the common range and are in accordance with values reported for compounds with Scheme 1. Typical Routes for the Synthesis of TransitionMetal NHC Complexes Scheme 2. Dicarbene Iron(II) Dihalide Complexes Synthesized by Grubbs et al. (X = Br, Cl; R = iPr) and Deng et al. (X = Cl; R = Me, Et, iPr) Table 1. Yields of Bis(imidazolium) Salts 1−10 (Left) and Iron(II) NHC Complexes 11−20 (Right) salt R Y X found (lit.) yield/% complex R Y X yield/% 1 Me CH2 Br 41 (83) 19 11 Me CH2 Br 2 Me (CH2)2 Br 67 (90) 19 12 Me (CH2)2 Br 84 3 Me (CH2)3 Br 60 (70) 20 13 Me (CH2)3 Br 80 4 Me CHPh Br 42 14 Me CHPh Br 5 Et CH2 Br 78 (80) 21 15 Et CH2 Br 65 6a Bn CH2 Br 64 (78) 22 16a Bn CH2 Br 79 6b Bn CH2 Cl 59 16b Bn CH2 Cl 44 7a Bu CH2 Br 44 (79) 19 17a Bu CH2 Br 83 7b Bu CH2 Cl 35 17b Bu CH2 Cl 90 8a Mes CH2 Br 84 (68) 19 18a Mes CH2 Br 82 8b Mes CH2 Cl 53 (76) 23 18b Mes CH2 Cl 86 9 Mes (CH2)2 Br 68 (45) 24 19 Mes (CH2)2 Br 68 10a Ad CH2 Br 61 (67) 19 20a Ad CH2 Br 85 10b Ad CH2 Cl 59 20b Ad CH2 Cl 63 Organometallics Article dx.doi.org/10.1021/om200870w |Organometallics 2011, 30, 6692−6702 6693 similar C−H···Br interactions and with related bis(imidazolium) salt derivatives. Synthesis of Dicarbene Iron(II) Complexes 11−20. A suspension of the respective bis(imidazolium) salt in Et2O or THF is treated with [Fe{N(SiMe3)2}2]2, and the ferrous NHC complex precipitates during the reaction (Scheme 3). Filtration and rinsing affords the desired product as a white to brownish powder, usually in around 80% yield. The yields of 11, 14, 15, and 16b are derogated by tetracarbene complex byproduct formation (see below); in the case of 11 no pure material could be isolated. In some cases further purification was hampered by the low solubility and limited stability in solution. The solid complexes 12−15 are quite air sensitive, and complexes 16 and 17 decompose within a couple of minutes up to 1 h in air, while complexes 18−20 with sterically demanding substituents (Mes and Ad) are air stable for several hours. In solution all these NHC complexes quickly decompose in the presence of air. The solubility of the new dicarbene iron(II) dihalide complexes in nonpolar solvents is generally poor, while polar and particularly protic solvents induce decomposition. They are insoluble in alkanes and toluene, sparingly soluble in DCM and THF, and soluble but slowly decompose in MeCN, DMF, and DMSO. Complexes with bis(imidazolylidene) ligands bearing mesityl and benzyl substituents (16a,b, 18a,b, and 19) represent exceptions, since they show better solubility in DCM. Bulk crystallization from DCM, CHCl3, THF, and dioxane in most cases was not successful; only 19 recrystallizes from DCM. Appending a phenyl group at the backbone methylene bridge (14) slightly increases the solubility. A similar effect was also observed for tetracarbene complexes 21 (insoluble in DCM) and 22 (phenyl substituted, soluble in DCM; see below). As a consequence, the purification and characterization of almost all dicarbene iron(II) complexes 11− 20 is limited by the poor solubility or stability in solution. However, the purity of most complexes after filtration, rinsing, and evaporation of the remaining solvents was quite good, as indicated by elemental analysis (EA) and Mo ̈ssbauer spectroscopy (MB, see below). Spectroscopic, Electrochemical, and Magnetic Characterization of Complexes 12−20. From the series of new dicarbene iron(II) dihalide complexes the molecular structures of 14, 15, 16b, 17a, 18a, and 19 were determined by X-ray diffraction. All complexes were characterized by EI-MS, MB, and EA (except for 11, 14, and 17b, which contained significant amounts of impurities). Those complexes soluble in DCM (17a, 18a,b, and 19) were additionally characterized by H NMR and UV−vis spectroscopy. H NMR spectra of 17a in DMSO-d6 and 18a,b and 19 in CD2Cl2 showed broad signals in the range of −5 to 60 ppm. Except for the H NMR spectrum of 18a, the number of signals and their integrals did not match expectations, likely due to paramagnetic broadening beyond detection for some signals. Paramagnetic signal broadening also prevented reasonable C NMR measurements; hence, C−H correlation experiments were not carried out. Cyclic voltammetry was performed for 17a, 18a,b, and 19 in MeCN or DCM (Table 2). All these complexes show irreversible processes in both oxidation and reduction. The first oxidation (anodic peak potential Ep ) is observed at around +1.1 V vs NHE for several [LFeBr2] complexes (the cyclic voltammogram of 17a is shown in Figure S1 (Supporting Information) as an example), but at +1.48 V for chloride complex 18b. Apparently the corresponding ferric complexes are unstable or major structural changes occur upon oxidation of the ferrous complexes. We also tried to synthesize the corresponding dicarbene iron(III) species directly when using 8a and [Fe{N(SiMe3)2}3] instead of [Fe{N(SiMe3)2}2]2, but it was not possible to identify or isolate any dicarbene iron(III) complex. Instead, reduction occurred and iron(II) NHC complex 18a crystallized from the reaction mixture at −30 °C. These observations suggest that the potential dicarbene iron(III) halide species are not accessible, in accordance with electrochemical findings. Mössbauer spectroscopy of complexes 12, 13, and 15−20 at 80 K showed isomer shifts of 0.73−0.81 mm s−1 and large quadrupole splittings in the range 3.67−4.03 mm s−1 (Table 3), which is at the upper end of the 3−4 mm s−1 range typical for high-spin iron(II) in a tetrahedral environment (S = 2). This comparison refers to iron complexes with thiourea or thiolate ligands, since Mössbauer data of tetrahedral iron NHC complexes are scarce. The high-spin state (S = 2) was further corroborated by a SQUID measurement of 18a, giving μ eff = 5.1 μ B (Figure 1). Mössbauer spectra of all iron complexes with bromide ligands show rough backgrounds due to a strong Scheme 3. Preparation of Dicarbene Iron(II) Dihalide Complexes from Bis(imidazolium) Salts Table 2. Electrochemical Data for Selected Complexes complex Ep a Ep c solvent 17a +1.11 −1.74 MeCN