Synthesis and characterization of heteromultinuclear Ni/M clusters (M = Fe, Ru, W) including a paramagnetic (NHC)Ni-WCp*(CO)₃ heterobinuclear complex

Abstract

A diverse range of heteromultinuclear Ni^I/[M_CO] clusters (M_CO = CpFe(CO)₂, CpRu(CO)₂, Cp*W(CO)₃) supported by a N-heterocyclic carbene ligand have been synthesized by reacting the Ni^I precursor, [IPrNi(μ-Cl)]₂, with [M_CO]⁻ reagents under various conditions. Clusters with Ni₂Fe₂, NiFe₂, Ni₂Ru, Ni₂Ru₂, NiRu₂, and Ni₂W, and NiW cores were all characterized using NMR and IR spectroscopies and X-ray crystallography. The Ni^I-containing paramagnetic heterobinuclear species, IPrNi-Wp* (7), was further characterized by EPR spectroscopy and DFT calculations. Notably, unlike previously studied (NHC)Cu^I-[M_CO] derivatives, complex 7 was found to coordinate Lewis bases like 3-chloropyridine to produce (IPr)(^3-Clpy)NiWp* (9). Complex 9 further underwent thermolytic C-Cl activation, proposed to involve NHC-free [(^3-Clpy)Ni(μ-Wp*)]₂ (10), to provide the C-arylated N-heterocyclic carbene product, [IPr(py-3-yl)]⁺[Cp*WCl₂(CO)₂]⁻ (11).

Graphical Abstract

INTRODUCTION

The development of new platforms for cooperative bond activation is a frontier research area in homogeneous catalysis.¹ One strategy for achieving cooperative bond activation is heterobinuclear cooperativity, which involves use of polar metal-metal interactions to activate inert small molecules for incorporation into organic substrates.^{1b,1c,1f,1i,2} In recent years, reported examples of organic transformations catalyzed by heterobinuclear metal complexes include C-H functionalization,³ carbonylation,⁴ hydrogenation,⁵ and hydrofunctionalization.⁶

Within this context, our group has been specifically interested in heterobimetallic systems consisting of a copper N-heterocyclic carbene (Cu-NHC) Lewis acid and a metal carbonyl anion ([M_CO]⁻ = e.g. FeCp(CO)₂⁻, Mn(CO)₅⁻, MoCp(CO)₃⁻, etc.) Lewis base.⁷ These complexes feature dative [M_CO]→Cu(NHC) bonds⁸ that are capable of C-X,⁹ H-H,¹⁰ H-B,¹¹ and H-Sn^6a activation. We have applied this platform to develop several classes of catalytic transformations,^{3b,4d,5b,6,12} and the modular nature of the catalyst design should allow for more applications to emerge as we expand our library of available heterobimetallic catalysts.

The catalytic mechanisms for the previously explored reactions catalyzed by (NHC)Cu-[M_CO] complexes involve redox cycling at the metal carbonyl unit, while the Cu(NHC) group remains redox neutral. New reactivity paradigms might be accessible if the d¹⁰ Cu^I center were replaced with a redox-active metal ion with an open valence shell. Generally, heterobimetallic complexes featuring [M_CO]-M′ bonds are limited to closed-shell systems where M′ is a d⁰ metal ion¹³ or a d¹⁰ metal ion.⁷ Exceptions include d⁸ Ni^II in square planar environments that have a strong preference for populating diamagnetic ground states and do not engage in Ni-based redox cycling.¹⁴ A recent survey of paramagnetic, metal-metal bonded, heterometallic complexes did not include any [M_CO]-M′ derivatives,¹⁵ and to our knowledge the only open shell examples are a series of ArM-FeCp(CO)₂ complexes reported by Power (M = Cr, Mn, Fe)¹⁶ and a (Ph₃P)₂Ni-WCp(CO)₃ complex,¹⁷ none of which have available reactivity data in the literature.

This paper reports our efforts at synthesizing (NHC)Ni-[M_CO] complexes in which Cu^I has been replaced by Ni^I, a redox-active d⁹ metal ion. Due to the possibility for redox chemistry, the synthesis of these heterobimetallic Ni^I complexes is not as straightforward as the Cu^I derivatives, and several heteromultinuclear Ni/Fe and Ni/Ru complexes were discovered serendipitously while targeting (NHC)Ni-MCp(CO)₂ molecules (M = Fe, Ru). In one case, a paramagnetic (NHC)Ni-WCp*(CO)₃ complex was successfully obtained. Reported here are this complex’s characterization by NMR and EPR spectroscopies, X-ray crystallography, DFT modeling, and preliminary reactivity data. While historically Ni^I has been considered a rare oxidation state,¹⁸ recently the diverse chemistry of this metal ion has emerged in many stoichiometric and catalytic transformations.¹⁹ We anticipate discovery of novel heterobinuclear catalysis based on the new heterobimetallic Ni^I complexes reported here.

RESULTS AND DISCUSSION

Synthesis and characterization of heteromultinuclear Ni/Fe and Ni/Ru clusters.

The most common method for synthesizing (NHC)Cu-[M_CO] complexes is to react (NHC)CuCl and [M_CO]⁻ precursors.²⁰ These salt metathesis reactions are typically quite clean and high-yielding, in part because there is little chance of competing electron transfer from [M_CO]⁻ to the d¹⁰ Cu^I center. Thus, we began our studies using [IPrNi(μ-Cl)]₂ as a Ni^I precursor²¹ in combination with NaFp and KRp (Fp = FeCp(CO)₂, Rp = RuCp(CO)₂). The products were fully characterized by ¹H NMR spectroscopy, IR spectroscopy, and X-ray crystallography. Although the targeted IPrNi-[M_CO] complexes were expected to be paramagnetic due to retention of d⁹ Ni^I (S = ½), these reactions with Fp⁻ and Rp⁻ produced exclusively diamagnetic products (Scheme 1), indicative of electron transfer between the [M_CO]⁻ anions and the Ni^I center.

Scheme 1.

Synthesis of heteromultinuclear (a) Ni/Fe and (b) Ni/Ru clusters (Fp = FeCp(CO)₂, Rp = RuCp(CO)₂)

The reaction between [IPrNi(μ-Cl)]₂ and NaFp in THF at room temperature produced a color change from yellow to dark brown immediately upon mixing, and ¹H NMR spectroscopy was used to monitor the reaction progress. In the first 5 minutes, two new resonances in the Cp region were evident at 3.88 and 4.28 ppm in a ratio of 2:1, while the 4.28-ppm resonance dominated after 6 h. These resonances were found to arise from the heterotetranuclear cluster, [IPrNi(μ-Fp)]₂ (1), and the cyclotrinuclear cluster, IPrNiFp₂ (2), respectively. The mixtures of 1 and 2 (ratio 2: 1) were found to quantitatively convert to 2 in C₆D₆ over 12 h at 60°C; presumably a IPrNi⁰ byproduct is lost during this transformation.

The Ni/Ru chemistry, while analogous to the Ni/Fe chemistry, was more complex. The reaction between [IPrNi(μ-Cl)]₂ and KRp in THF at room temperature gave mixtures of cyclotrinuclear IPrNiRp₂ (3) as the major product with two minor products, [IPrNi(μ-Rp)]₂ (4) and (IPrNi)₂(μ-Rp)(μ-CO)₃(μ₃-CO) (5). All of the assignments were confirmed by X-ray crystallography, as discussed below. Running the same reaction at 40°C instead of room temperature provided complex 3 exclusively. Reacting [IPrNi(μ-Cl)]₂ and KRp in toluene at −35°C gave yet another complex, (IPrNi)₂(μ-Rp)(μ-Cl) (6). Interestingly, complex 4 could be obtained by adding NaOtBu to 6 in toluene at 50 °C for 1 h. Adding additional KRp to 6 produced 3. When isolated complex 4 was heated to 80°C in C₆D₆ for 24 h, it quantitatively converted to 3 with loss of “IPrNi⁰” as with the Fe case. Thus, synthetic methods were identified for isolating complexes 3, 4, and 6. Although we accidently obtained a crystal of 5, we failed to identify conditions that allowed us to isolate 5, which we hypothesize is a decomposition product generated by loss of “CpRu” from 4.

Solid-state structures of complexes 1-6 are shown in Figure 1. The solid-state structure of 2 displays an isosceles NiFe₂ triangle with bent Ni-Fe-Fe (60.329(14) and 58.768(14)°) and Fe-Ni-Fe (60.903(14)°) angles and significantly bent C_NHC-Ni-Fe angles (139.33(8)° and 159.63(8)°). The Ni-Fe bond distances (2.4652(5) and 2.5050(5) Å) are in the range of Ni-Fe single bonds (typically 2.41–2.64 Å), and the Fe-Fe bond distance (2.5192(5) Å) is in the range of a Fe-Fe single bond (2.46–2.78 Å) based on literature precedents.²² The analogous NiRu₂ complex 3 is isostructural, with bent Ni-Ru-Ru (57.916(11) and 57.304(12)°) and Ru-Ni-Ru (64.780(12)°) angles along with significantly bent C_NHC-Ni-Ru angles (142.56(8) and 151.81(8)°). The Ni-Ru distances (2.5165(5) and 2.5337(4) Å are close to the Ni-Ru single bond distance in Tp^#Ni-RuCp(CO)₂ (2.512(1) Å; Tp^# = hydrotris(3,5-dimeth-yl-4-bromopyrazolyl)borato))²³, and the Ru-Ru bond distance (2.7053(5) Å) is in the range of a Ru-Ru single bond (2.59–2.75 Å) in some Ru cluster cases.²⁴ Previously, similar structures of related MoNi₂ clusters supported by NHCs have been reported by Chetcuti.²⁵ Calculation of asymmetry parameters²⁶ indicate that all CO ligands in 2 and 3 are in the semi-bridging regime. In each case, one CO unit can be considered as a semi-bridging μ₃-CO between Ni and two M (M = Fe or Ru), while the others are assigned as μ₂-CO ligands between Ni and M or M and M. The solid-state IR spectra of 2 and 3 have features at 1656 and 1665 cm⁻¹, respectively, assigned to the more activated semi-bridging μ₃-CO ligands, while the μ₂-CO ligands are assigned to features at 1841 and 1749 cm⁻¹, respectively.

Figure 1.

Solid-state structures of complexes 1-6 determined by X-ray crystallography. For clarity, hydrogen atoms and crystallized solvent molecules are omitted. NHC ligands and Cp groups are shown as wireframes. All other atoms are shown as 50% probability ellipsoids.

The solid-state structure of 1 displays a tetranuclear Ni-Fe-Fe-Ni core, featuring bent Ni-Fe-Fe (82.75(6)°) and C_NHC-Ni-Fe (158.60(19)°) angles. The Ni-Fe bond distance in 1 (2.3974(14) Å) is shorter than in 2, and the Fe-Fe bond distance is 2.525(2) Å, which is shorter than the Fe-Fe distance in Fp₂ (2.5389(3) Å)²⁷. The analogous Ni₂Ru₂ complex 4 is isostructural to 1, with bent Ni-Ru-Ru (81.454(4)°) and C_NHC-Ni-Ru (156.8(2)°) angles. The Ni-Ru bond distance in 4 (2.4785(11) Å) is shorter than in 3, and the Ru-Ru bond distance is 2.7334(15) Å, which is similar to that in Rp₂ (2.7412(4) Å)²⁸. According to the solid-state structures of 1 and 4, in each case two CO units can be considered semibridging μ₃-CO ligands between Ni and two M (M = Fe or Ru), while the others are assigned as semibridging μ₂-CO ligands between Ni and M. The elongated C-O bond distances in 1 (1.218(8) Å) and in 4 (1.220(9) Å) indicate the significant CO activation. The solid-state IR spectra of 1 and 4 show features at 1552 and 1554 cm⁻¹, respectively, assigned to semi-bridging μ₃-CO ligands, consistent with the longer C-O bond distances. The μ₂-CO ligands are assigned to features at 1757 and 1753 cm⁻¹, respectively.

Complex 6 has Ni-Ru bond distances of 2.5702(6) and 2.5580(6) Å and a Ni-Ni distance of 2.2974(7) Å. Interestingly, 6 displays an unusually short Ni-Ni distance compared to [IPrNi(μ-Cl)]₂ (2.5194(5) Å)², and slightly shorter than the Ni-Ni distances measured for [(IPr)Ni]₂(μ-Cl)(μ:η²-4-CH₃-C₆H₄) (2.3954(5) Å)^19e and [(IPr)Ni]₂(μ-CO)(μ-η²,η²-CO₂) (2.3374(4) Å)²⁹. Complex 6 has bent Ni-Ni-Ru (63.657(19) and 63.113(19)°) and Ni-Ru-Ni (53.231(17)°) angles and shows significantly bent C_NHC-Ni-Ni angles (171.69(10) and 172.72(11)°). The CO units can be considered as semi-bridging μ₂-CO ligands between Ni and Ru. The solid-state IR spectrum of the highly symmetric 6 displays a single feature at 1749 cm⁻¹.

Synthesis and characterization of a heterobinuclear Ni/W complex.

Compared to Fp⁻ (E° = −1.352 V vs. ferrocene) and Rp⁻ (E° = −1.057 V vs. ferrocene), Wp⁻ derivatives are known to be weaker reductants (Wp = WCp(CO)₃, E° = −0.491 V vs. ferrocene).³⁰ Furthermore, a (Ph₃P)₂Ni-Wp derivative was studied in the 1980s,¹⁷ indicating that Ni^I is compatible with Wp⁻. Thus, we hypothesized that electron transfer from [M_CO]⁻ to Ni^I could be avoided by use of tungsten carbonyl anions under appropriate reaction conditions. Accordingly, we were successful in preparing IPrNiWp* (7) from a 1:4 mixture of [IPrNi(μ-Cl)]₂ and LiWp* in THF at 40°C (Scheme 2, Wp* = WCp*(CO)₃). As expected, 7 displays a ¹H NMR spectrum with broadened and paramagnetically shifted resonances between 0 and 15 ppm. In addition, we successfully identified and isolated the intermediate, (IPrNi)₂(μ-Wp*)(μ-Cl) (8), from a 1:2 mixture of [IPrNi(μ-Cl)]₂ and LiWp* that had been allowed to react in toluene at room temperature for 2 h. Isolated complex 8 was converted to 7 with additional LiWp*. During investigation of the thermal stability of 7, we found that heating to 80°C for 24 h in C₆D₆ formed free IPr plus a decomposition product we are tentatively calling [Ni(μ-Wp*)]_n. In two cases, decomposition mixtures yielded X-ray quality crystals in which free IPr co-crystallized with either [Ni(μ-Wp*)]₃ or [Ni(μ-Wp*)]₄ that featured planar Ni_nW_n metallacycles with complex μ-CO networks. However, we have been unable to model these data sufficiently for publication.

Scheme 2.

Synthesis of a heteromultinuclear Ni/W complexes

The solid-state structures of 7 and 8 were determined by X-ray crystallography (Figure 2). Intermediate 8 has a structure closely related to that of 6, with Ni-W bond distances of 2.5910(7) and 2.6138(7) Å, a Ni-Ni distance of 2.4335(7) Å, and bent Ni-Ni-W (62.601(19) and 61.649(19)°), Ni-W-Ni (55.749(18)°), and C_NHC-Ni-Ni (166.03(6) and 161.08(7)°) angles. The three CO units can be considered as semibridging μ₂-CO ligands between Ni and W. The solid-state IR spectrum of 8 has features at 1763 and 1722 cm⁻¹ assigned to these semi-bridging μ₂-CO ligands.

Figure 2.

Solid-state structures of complexes 7-9 determined by X-ray crystallography. For clarity, hydrogen atoms and crystallized solvent molecules are omitted. NHC ligands and Cp groups are shown as wireframes. All other atoms are shown as 50% probability ellipsoids.

Heterobinuclear complex 7 features a Ni-W bond distance of 2.4977(10) Å, and a bent C_NHC-Ni-W angle (156.8(2)°). It has a significantly shorter Ni-W bond than the Ni-W single bond in (Ph₃P)₂Ni-WCp(CO)₃ (2.574(3) Å)¹² and is in the range of the Ni=W double bond (~2.466 Å) reported for the complex of Cp*Ni-W(η-C₅H₄Me)(CO)₃.³¹ According to the solid-state structure, two of the CO units are assigned as semi-bridging ligands according to their asymmetry parameters.⁶ The solid-state IR spectrum of 7 has two features at 1733 and 1760 cm⁻¹ assigned to the two semi-bridging CO ligands, and a feature at 1891 cm⁻¹ assigned to the terminal CO ligand. The closest Cu-W analogue to 7 is IPrCuWp (Wp = WCp(CO)₃),^1a which has been structurally characterized in two forms, one with two semi-bridging CO ligands and one with three. Selected metrics for the form with two semi-bridging CO ligands are presented in Table 1 for comparison to complex 7. The metal-metal bond distance in 7 is comparatively short despite the larger covalent radius of Ni compared to Cu. This could be due to Ni-W bond with more covalent character, or due to an enhanced W→Ni dative from the more electron rich [Wp*]⁻ compared to Wp⁻. Otherwise, the two complexes are roughly isostructural.

Table 1.

Comparison of Ni-W and Cu-W heterobinuclear complexes.

	(NHC)Ni-WCpR(CO)3	(NHC)Cu-WCp(CO)3c
d(M-W) (Å)a	2.4977(10)	2.5599(6)
d(M···CO) (Å)a	2.021(9), 2.072(9), 3.364(9)	2.294(7), 2.280(5), 3.858(6)
∠(CNHC-M-W) (°)a	156.8(2)	165.7(1)
νCO (cm−1)a	1733, 1760, 1891	1784, 1818, 1920
q(M)b	0.42	0.40
q(W)b	−0.89	−0.95
q(IMeM)a	0.57	0.69
q(WCpR(CO)3)b	−0.57	−0.69
M-W Wiberg bond indexb	0.31	0.29

^aSolid-state data for NHC = IPr and Cp^R = η⁵-C₅Me₅ (M = Ni) or Cp^R = η⁵-C₅H₅ (M = Cu).

^bCharges (q) and bond indices from NBO calculations for NHC = IMe and Cp^R = η⁵-C₅H₅.

^cFrom ref. 1a.

The X-band EPR spectrum of 7 in 2-methyltetrahydrofuran frozen solution is shown in Figure 3. The spectrum is rhombic;³² however, the g value pattern³³ is characteristic of a d⁹ metal ion with a singly-occupied molecular orbital (SOMO) of d_x2-y2 character, as typically seen for Cu^II, and of particular relevance here for Ni^I.³⁴ Particularly notable are two weak, satellite features associated only with g_mid. We assign these to a hyperfine-split doublet arising from coupling of the spin primarily on Ni^I to ¹⁸³W (I = 1/2, 14% abundant). This hyperfine coupling is very anisotropic. EPR simulations allow a very rough estimate as to the coupling at all g values: A(¹⁸³W = [30, 390, 120] MHz. The resulting A(¹⁸³W) tensor gives an isotropic coupling that corresponds to the SOMO having a maximum of ~3% W 6s character.^35,36 For comparison, no tungsten hyperfine was resolved in the EPR spectrum of (Ph₃P)₂Ni-Wp.¹² In general, previous examples of low-valent tungsten complexes characterized by EPR spectroscopy are somewhat rare and involve mononuclear complexes. The closest analogue to 7 is Bullock’s CpW(CO)₂(IMes) (IMes = N,N’-bis(2,4,6-trimethylphenyl)imidazol-2-ylidene).³⁷ Interestingly, despite Bullock’s complex being assigned as a W(I) metalloradical, no tungsten hyperfine features were resolved in the EPR spectra. In contast, Connelly characterized Tp′W(CO)₂(η²-RC≡CR) (Tp′ = hydridorotris(3,5-dimethylpyrazolyl)borate) complexes by EPR spectroscopy and measured an A_iso(¹⁸³W) value of 128 MHz (for R = Me; EPR was insensitive to R and gave resolved coupling at all g values: A(¹⁸³W) = –[37,38,53] MHz).^38,39 Additionally, Connelly’s complexes gave a different g value pattern: rhombic, with one value below g_e (2.00), g = [2.102, 2.015, 1.955]. In contrast, complex 7 gives all g values above g_e, and the maximum deviation from g_e is much larger. Assuming a limiting W(I) description for both systems, this difference could qualitatively be due to the monotungsten(I) having a SOMO of d_yz parentage but 7 having a SOMO of d_x2-y2 parentage. Thus, the comparison between Connelly’s monotungsten(I) complexes and heterobinuclear 7 is of limited utility but still indicates some measurable W(I) character of 7.

Figure 3.

X-band EPR spectrum of IPrNiWp* (7) in frozen Me-THF solution (black trace), with simulation (red trace). Experimental conditions: temperature, 77 K; microwave frequency, 9.255 GHz; microwave power, 6.3 mW (15 dB); field modulation amplitude, 0.1 mT; time constant, 1 s; scan time, 240 s. Simulation parameters: S = 1/2, g = [2.038, 2.125, 2.427], W = 25, 40, 60 MHz (gaussian, hwhm); a spectrum with a normalized intensity of 85.7 is added to a spectrum with a normalized intensity of 14.3 that includes hyperfine coupling to ¹⁸³W (I = 1/2) with A = [30, 390, 120] MHz (assumed collinear with g).

A truncated model of 7, IMeNiWp (7′, IMe = N,N’-dimethylimidazol-2-ylidene), was studied using density functional theory (DFT) calculations. The optimized structure of 7′ includes the presence of two semi-bridging CO ligands and a bent C_NHC-Ni-W angle of 164°, although the Ni-W distance is overestimated in the calculation as 2.57 Å. The calculated CO stretching frequencies are 1745, 1773, and 1870 cm⁻¹. In accordance with EPR analysis of 7, the calculated spin density for 7′ (Figure 4) resides mainly on Ni, with Mulliken spin densities of 0.92 on Ni, 0.10 on W, and −0.02 on C_NHC (sum = 1.0). Model complex 7′ was also analyzed with natural bond orbital (NBO) calculations, which allowed us to calculate natural charges and bond indices. These values are presented in Table 1 for comparison to previously calculated IMeCuWp.^1a Based on partial charges, the Ni-W bond is less polar than the corresponding Cu-W bond; for example, charge difference between [IMeM]⁺ and Wp⁻ fragments is 1.14e⁻ (M = Ni) vs. 1.38e⁻ (M = Cu). Similarly, the Ni-W bond is calculated to have slightly more covalent character than the corresponding Cu-W bond according to its Wiberg bond index. Collectively, the data are indicative of a dative W⁰→Ni^I interaction with slightly more covalent character than the corresponding W⁰→Cu^I derivative, thus giving rise to the minor W^I-Ni⁰ contribution indicated by EPR spectroscopy.

Figure 4.

Spin density plot (isovalue = 0001 au) for IMeNiWp (7′) determined by DFT (B3LYP/LANL2TZ(f) for Ni,W/6–31+G(d) for all other atoms).

Reactivity studies.

Recently, we reported the (IPr)CuFp-catalyzed 1,4-hydroboration of pyridines.^6b To expand our knowledge on Ni^I chemistry, we thus conducted preliminary reactivity studies between 7 and various substituted pyridines (^Rpy) to form complexes. These reactions were found to give products of the type (IPr)(^Rpy)Ni-WCp*(CO)₃. For example, addition of 3-chloropyridine to 7 immediately produced the mixtures of 3-chloropyridine adducts, with majority of (IPr)(^3-Clpy)NiWp* (9) along with minor [(^3-Clpy)Ni(μ-Wp*)]₂ (10). We propose that there is an equilibrium between 9 and 10 + IPr (Scheme 3), analogous to the IPr dissociation equilibrium proposed for 7. This behavior is notable, as all the (NHC)Cu-[M_CO] derivatives we studied previously were inert towards binding Lewis bases, presumably due to the preference of (NHC)CuX complex to adopt linear coordination geometries. The solid-state structure of 9 (Figure 2) shows a longer Ni-W bond distance (2.5793(8) Å) than in 7. This is a consequence of the increase in coordination number at Ni, and this Ni-W distance comparable to that in (Ph₃P)₂Ni-Wp (2.574(3) Å).¹⁰ The C_NHC-Ni-W angle is also more bent in 9 (136.02(16) °) than in 7. The identity of minor component 10 was also determined by X-ray crystallography.

Scheme 3.

Reactivity studies of 7 with 3-chloropyridine.

Heating complex 9 to 150°C for 0.5 h or to 100°C for 12 h resulted in ionic complex 11, in which the IPr ligand has been arylated at C_NHC upon C-Cl bond activation of the 3-chloropyridine substrate (Scheme 3). A trace amount of 3,3’-bipyridine was also detected in the product mixture. Monitoring formation of 11 by ¹H NMR in toluene-d₈ indicated the presence of an intermediate forming on the way to 11, which was determined to be [(^3-Clpy)Ni(μ-Wp*)]₂ (10). Given that Ni-catalyzed arylation of the C_NHC position of free N-heterocyclic carbenes has already been reported,⁴⁰ it is likely that 10 mediates the arylation of free IPr with concurrent release of [Cp*WCl₂(CO)₂]⁻. Collectively, the ease with which IPr dissociates from 7 and 9 should be noted for any future reactivity studies.

CONCLUSIONS

In conclusion, a new series of heteromutinuclear, Ni^I-containing complexes supported by a N-heterocyclic carbene ligand has been synthesized and characterized crystallographically. The transformations of complexes 1–8 indicate the versatile reactivity of redox-active d⁹ Ni^I compared to redox-inert d¹⁰ Cu^I metal sites. Moreover, the paramagnetic heterobimetallic complex 7 was successfully synthesized and further investigated by EPR spectroscopy and DFT calculations, indicating its Ni^I metalloradical character. While the Ni-W bond in 7 is slightly more covalent than its Cu-W analogue, stoichiometric reactivity data of 7 are consistent with a more labile Ni-NHC bond compared to the inert Cu-NHC bond. Additionally, unlike previously studied heterobinuclear Cu^I complexes,^7,8 complex 7 was found to coordinate pyridines at the Ni^I metal site, leading to heterobimetallic C-Cl activation of 3-chloropyridine. Further studies towards catalytic applications of 7 and its more detailed spectroscopic characterization are ongoing in our laboratory and will be informed by the results reported here.

EXPERIMENTAL SECTION

General considerations.

All reactions and manipulations were conducted under purified N₂ using standard Schlenk line techniques or in a glovebox. Reaction solvents (THF, toluene, diethyl ether, and pentane) were purified using a Glass Contour Solvent System built by Pure Process Technology, LLC. Deuterated solvents (C₆D₆, CD₂Cl₂, CDCl₃, and toluene-d₈) were purchased from Sigma-Aldrich and stored over activated 3-Å molecular sieves prior to use. ¹H and ¹³C NMR spectra were recorded using Bruker Avance 400-MHz or 500-MHz NMR spectrometers. NMR spectra were recorded at room temperature unless otherwise indicated, and chemical shifts were referenced to residual solvent peaks. FT-IR spectra were recorded in a glovebox on powder samples using a Bruker ALPHA spectrometer fitted with a diamond-ATR detection unit. Elemental analyses were performed by Midwest Microlab, LLC, in Indianapolis, IN. Literature methods were used to synthesize NaFp,^{20, 41} KRp,⁴² LiWp*,⁴³ and [IPrNi(μ-Cl)]_2.²¹ Although satisfactory combustion analysis data were not obtained after multiple attempts for complexes 4, 6, 7, 8, 9, and 11, NMR and IR data are included in the Supporting Information as indications of purity.

X-ray crystallography.

For each complex, a crystal was selected for data collection, encased in silicone oil, transferred to a nylon loop, and cooled to 100 K. Data collection was carried out using a Bruker D8 QUEST ECO diffractometer under its default manufacturer settings. Standard solution and refinement methods⁴⁴ within the SHELX package⁴⁵ were applied as indicated in the supporting CIF files.

EPR spectroscopy.

Frozen solution X-band EPR spectra were recorded on a modified Varian E-9 spectrometer using a liquid nitrogen finger dewar. Room temperature EPR spectra were attempted, but gave only very weak signals, so that no a_iso(¹⁸³W) value (i.e., the true isotropic value, as opposed to the average of the A matrix components in frozen solution) could be determined, in contrast to Connelly and co-workers who were able to record fluid solution (toluene) spectra.³⁸ EPR spectra were simulated using the program QPOW, originally by Belford and co-workers,⁴⁶ as modified by J. Telser.

Computational methods.

DFT calculations were performed with Gaussian09 Revision B.01⁴⁷ using the built-in hybrid functional, B3LYP.⁴⁸ The LANL2DZ(f) basis set was used for transition metals,⁴⁹ and the Gaussian09 internal 6–31+G(d) basis set was used for all other atoms. Input geometries were obtained from the crystallographic coordinates of 7. To reduce the computational cost, the 2,6-di-iso-propylphenyl groups in the structure were replaced by methyl groups. Vibrational frequencies were calculated to verify identification of an energy minimum (zero imaginary frequencies), and the optimized coordinates are provided as Supporting Information. Vibrational frequencies were corrected using standard scaling factors.⁵⁰

Preparation of [IPrNi(μ-Fp)]₂ (1).

In a nitrogen filled glovebox, [IPrNi(μ-Cl)]₂ (20.8 mg, 0.0215 mmol) and NaFp (9.0 mg, 0.046 mmol) were dissolved in tetrahydrofuran (THF, 4 mL) in a 20-mL scintillation vial and reacted at room temperature for 5 mins. After the reaction completed, the solution was dried in vacuo, resulting in a dark-brown solid residue. The solid was extracted with diethyl ether (Et₂O, 2 mL) and pipette-filtered through Celite. Crystallization was accomplished by leaving a concentrated solution in Et₂O (1 mL) with few drops of pentane at −35 °C to give dark brown crystals. ¹H NMR yield with tetraethoxysilane as internal standard: 41%. ¹H NMR (500 MHz, C₆D₆): δ 7.06 (s, 12H, Ar-H), 6.56 (s, 4H, NCH), 3.88 (s, 10H, Cp), 2.96 (sept., J = 7.0 Hz, 8H, CH(CH₃)₂), 1.50 (d, J = 7.0 Hz, 24H, CH(CH₃)₂), 1.06 (d, J = 7.0 Hz, 24H, CH(CH₃)₂). ¹³C{¹H} NMR (400 MHz, C₆D₆): δ 194.4 (NCNi), 146.7, 137.1, 130.1, 130.0, 124.3, 123.2, 85.3(Cp*), 28.8, 25.7, 23.4 (Note: no peak corresponding to bound CO was located). IR (solid, cm⁻¹): 1757 (ν_CO), 1552 (ν_CO). This mixture contained complexes 1 and 2 in the ratio of 2:1; spectral features from 2 are omitted in the peak lists.

Preparation of IPrNiFp₂ (2).

In a nitrogen filled glovebox, [IPrNi(μ-Cl)]₂ (20.5 mg, 0.0212 mmol) and NaFp (12.8 mg, 0.0640 mmol) were dissolved in THF (5 mL) in a 20-mL scintillation vial and reacted at room temperature for 6 h. The reaction turned dark-brown and then the solution was dried in vacuo, resulting in a dark-brown residue, which was extracted with toluene (2 mL) and pipette-filtered through Celite. Crystallization was accomplished by leaving a concentrated solution in toluene (1 mL) with few drops of pentane at −35 °C to give brown crystals. Yield: 22.4 mg, 0.0280 mmol, 66%. ¹H NMR (500 MHz, C₆D₆): δ 7.31 (t, J = 8.0 Hz, 2H, p-H), 7.20 (d, J = 8.0 Hz, 4H, m-H), 6.54 (s, 2H, NCH), 4.28 (s, 10H, Cp), 2.86 (sept., J = 6.5 Hz, 4H, CH(CH₃)₂), 1.44 (d, J = 7.0 Hz, 12H, CH(CH₃)₂), 1.03 (d, J = 7.0 Hz, 12H, CH(CH₃)₂). ¹³C{¹H} NMR (500 MHz, C₆D₆): δ 189.6 (NCNi), 146.3, 136.2, 130.0, 124.2, 124.1, 89.9 (Cp*), 28.8, 25.9, 22.9 (Note: no peak corresponding to bound CO was located). IR (solid, cm⁻¹): 1837 (ν_CO), 1797 (ν_CO), 1749 (ν_CO), 1656 (ν_CO). Anal. Calcd for C₄₁H₄₆Fe₂N₂NiO₄: C, 61.46; H, 5.79; N, 3.50. Found: C, 60.96; H, 6.35; N, 3.78.

Preparation of IPrNiRp₂ (3).

In a nitrogen filled glovebox, [IPrNi(μ-Cl)]₂ (16 mg, 0.0166 mmol) and KRp (3.0 equiv.) were dissolved in THF (4 mL) in a 20-mL scintillation vial and reacted at 40 °C for 4 h. After the reaction completed, the solution was dried in vacuo, resulting in a dark-brown solid residue. The solid was extracted with toluene (2 mL) and pipette-filtered through Celite. Crystallization was accomplished by leaving a concentrated solution in toluene (1 mL) with few drops of THF at −35 °C to give brown crystals. Yield: 21.3 mg, 0.0239 mmol, 72%. ¹H NMR (500 MHz, C₆D₆): δ 7.28 (t, J = 8.5 Hz, 2H, p-H), 7.20 (d, J = 7.5 Hz, 4H, m-H), 6.63 (s, 2H, NCH), 4.81 (s, 10H, Cp), 3.04 (sept., J = 7.0 Hz, 4H, CH(CH₃)₂), 1.49 (d, J = 7.0 Hz, 12H, CH(CH₃)₂), 1.06 (d, J = 7.0 Hz, 12H, CH(CH₃)₂).¹³C{¹H} NMR (500 MHz, C₆D₆): δ 235.6 (CO), 195.3(NCNi), 146.5, 136.2, 130.0, 124.2, 124.1, 90.2 (Cp*), 28.6, 26.3, 23.0. IR (solid, cm⁻¹): 1841 (ν_CO), 1801 (ν_CO), 1754 (ν_CO), 1665 (ν_CO). Anal. Calcd for C₄₁H₄₆Ru₂N₂NiO₄: C, 55.23; H, 5.20; N, 3.14. Found: C, 55.42; H, 5.94; N, 3.58.

Preparation of [IPrNi(μ-Rp)]₂ (4).

In a nitrogen filled glovebox, [(IPrNi)₂(μ-Rp)(μ-Cl)] (34.0 mg, 0.0295 mmol) and NaO^tBu (4.8 mg, 0.0499 mmol) were dissolved in THF (4 mL) in a 20-mL scintillation vial and reacted at 50 °C for 1 h. After the reaction completed, the solution was dried in vacuo, resulting in a red-brown solid residue that was washed with cold pentane (1 mL). The solid was extracted with toluene (2 mL) and pipette-filtered through Celite. Crystallization was accomplished by leaving a concentrated solution in toluene (1 mL) with few drops of pentane at −35 °C to give reddish brown crystals. Yield: 34.0 mg, 0.0252 mmol, 85%. ¹H NMR (500 MHz, C₆D₆): δ 7.14–7.07 (m, 12H), 6.54 (s, 4H, NCH), 4.35 (s, 10H, Cp), 2.98 (sept., J = 7.0 Hz, 8H, CH(CH₃)₂), 1.46 (d, J = 7.0 Hz, 24H, CH(CH₃)₂), 1.05 (d, J = 7.0 Hz, 24H, CH(CH₃)₂).¹³C{¹H} NMR (500 MHz, C₆D₆): δ 195.2 (NCNi), 146.8, 137.1, 130.1, 124.3, 123.1, 87.6 (Cp*), 28.7, 25.7, 23.5 (Note: no peak corresponding to bound CO was located). IR (solid, cm⁻¹): 1753 (ν_CO), 1554 (ν_CO).

Preparation of [(IPrNi)₂(μ-Rp)(μ-Cl)] (6).

In a nitrogen filled glovebox, [IPrNi(μ-Cl)]₂ (55.8 mg, 0.0578 mmol) and KRp (32.4 mg, 0.124 mmol) were dissolved at −35 °C in toluene (4 mL) in a 20-mL scintillation vial and reacted allowed to warm from −35 °C to room temperature for 1 h. After the reaction completed, the solution was pipette-filtered through Celite and dried in vacuo, resulting in a brown solid residue that was then washed with pentane (2 × 2 mL). Crystallization was accomplished by leaving a concentrated solution in toluene (1 mL) with few drops of pentane at −35 °C to give brown crystals. Yield: 61.0 mg, 0.0529 mmol, 91%. ¹H NMR (500 MHz, C₆D₆): δ 7.33 (t, J = 7.5 Hz, 4H, p-H), 7.25 (d, J = 7.5 Hz, 8H, m-H), 6.30 (s, 4H, NCH), 4.51 (s, 5H, Cp), 2.98 (sept., J = 7.0 Hz, 8H, CH(CH₃)₂), 1.52 (d, J = 7.0 Hz, 24H, CH(CH₃)₂), 1.00 (d, J = 7.0 Hz, 24H, CH(CH₃)₂).¹³C{¹H} NMR (500 MHz, C₆D₆): δ 240.9 (CO), 181.8 (NCNi), 146.0, 137.4, 129.7, 124.1, 124.0, 80.5(Cp*), 28.8, 25.4, 23.4. IR (solid, cm⁻¹): 1749 (ν_CO).

Preparation of IPrNiWp* (7).

In a nitrogen filled glovebox, [IPrNi(μ-Cl)]₂ (95.0 mg, 0.098 mmol) and LiWp* (162.1 mg, 0.394 mmol) were dissolved in THF (4 mL) in a 20-mL scintillation vial and reacted at 40 °C for 4 h. After the reaction completed, the solution was dried in vacuo, resulting in a dark greenish solid residue that was washed with pentane (2 × 2 mL). The solid was extracted with toluene (4 mL) and pipette-filtered through Celite. Crystallization was accomplished by leaving a concentrated solution in toluene (1 mL) with few drops of pentane at −35 °C to give green crystals. Yield: 141.2 mg, 0.166 mmol, 84%. ¹H NMR (500 MHz, C₆D₆): δ 14.58 (br,s), 7.99 (br,s), 7.58 (br,s), 6.61 (br,s), 3.71 (br,s,), 2.36 (br,s), 1.37 (br,s). IR (solid, cm⁻¹): 1891 (ν_CO), 1760 (ν_CO), 1733 (ν_CO).

Preparation of [(IPrNi)₂(μ-Wp*)(μ-Cl)] (8).

In a nitrogen filled glovebox, [IPrNi(μ-Cl)]₂ (20.7 mg, 0.0214 mmol) and LiWp* (18.0 mg, 0.043 mmol) were dissolved in toluene (4 mL) in a 20 ml scintillation vial and reacted at room temperature for 2 h. After the reaction completed, the solution was dried in vacuo, resulting in a dark brown-greenish solid residue that was washed with pentane (2 × 2 mL). The solid was extracted with toluene (2 mL) and pipette-filtered through Celite. Crystallization was accomplished by leaving a concentrated solution in toluene (1 mL) with few drops of pentane at −35 °C to give brown crystals. Yield: 24.5 mg, 0.0184 mmol, 86%. ¹H NMR (500 MHz, C₆D₆): δ 7.38 (t, J = 7.5 Hz, 4H, p-H), 7.32 (d, J = 7.5 Hz, 8H, m-H), 6.55 (s, 4H, NCH), 3.22 (sept., J = 6.5 Hz, 8H, CH(CH₃)₂), 1.97 (s, 15H, Cp*), 1.54 (d, J = 7.0 Hz, 24H, CH(CH₃)₂), 0.98 (d, J = 7.0 Hz, 24H, CH(CH₃)₂).¹³C{¹H} NMR (500 MHz, C₆D₆): δ 248.4 (CO), 146.0, 137.8, 129.4, 127.1, 124.1, 99.7(Cp*), 28.7, 26.2, 23.4, 11.6 (Note: no peak corresponding to NCNi was located). IR (solid, cm⁻¹): 1763 (ν_CO), 1722 (ν_CO).

Preparation of mixture containing (IPr)(^3-Clpy)NiWp* (9) and [(^3-Clpy)Ni(μ-Wp*)]₂ (10).

In a nitrogen filled glovebox, IPrNiWp* (31.0 mg, 0.0364 mmol) and 3-chloropyridine (4.3 mg, 0.0375 mmol) were dissolved in pentane (4 mL) and toluene (0.5 mL) in a 20-mL scintillation vial, resulting into brown solution immediately. After 5 min, the solution was pipette-filtered through Celite. Crystallization was accomplished by concentrating the filtrate (1.5 mL) and allowing to sit at −35 °C to give brown crystals. Complexes 9 and 10 were separated by sequential recrystallization: crystals of 10 were obtained in the 1^st recrystallization (but still contained some impurities); then, the mother liquor was filtered by Celite pipette and provided crystals of 9 in a 2^nd recrystallization. Yield of major complex 9: 17.6 mg, 0.0182 mmol, 50%. Complex 9, ¹H NMR (500 MHz, C₆D₆): δ 13.58 (br,s), 7.97 (br,s), 7.75 (br,s), 7.32 (br,s, overlap with d-benzene), 2.73 (br,s), 1.03 (br,s). IR (solid, cm⁻¹): 1852 (ν_CO), 1749 (ν_CO), 1716 (ν_CO).

Yield of minor complex 10: 3.3 mg, 0.0028 mmol, 14%. Complex 10, ¹H NMR (500 MHz, C₆D₆): δ 9.89 (m, 2H, 3-Cl-py), 8.04 (m, 2H, 3-Cl-py), 7.34 (m, 4H, 3-Cl-py), 2.36 (s, 30H, Cp*). IR (solid, cm⁻¹): 1872 (ν_CO), 1739 (ν_CO).

Preparation of [IPr(py-3-yl)]⁺[Cp*WCl₂(CO)₂]⁻ (11).

In a nitrogen filled glovebox, IPrNiWp* (29.8 mg, 0.035) and 3-chloropyridine (16.1 mg, 0.140 mmol) were dissolved in toluene (1 mL) in a 20-mL scintillation vial, resulting into brown solution that was immediately transferred to a J-Young NMR tube. The reaction was performed at 150 °C for 0.5 h or 100 °C for 12 h, after which the solution was pipette-filtered through Celite. Crystallization was accomplished by leaving a concentrated solution in toluene (0.5 mL) with few drops of pentane at −35 °C. Yield: 16.4 mg, 0.018 mmol, 51%. ¹H NMR (500 MHz, CDCl₃): δ 7.62 (br), 7.55 (br), 2.23 (br), 2.17 (br, s), 1.97 (br), 1.42 (br), 1.26 (br). IR (solid, cm⁻¹): 2009 (ν_CO), 1903 (ν_CO). Note: trace amount of 3,3’-bipyridine was found in the crude mixture, with ¹H NMR peaks matching reported values in the literature.⁵¹