Fluorine-Substituted Arylphosphine for an NHC-Ni(I) System, Air-Stable in a Solid State but Catalytically Active in Solution

Monovalent NHC-nickel complexes bearing triarylphosphine, in which fluorine is incorporated onto the aryl groups, have been synthesized. Tris(3,5-di(trifluoromethyl)-phenyl)phosphine efficiently gave a monovalent nickel bromide complex, whose structure was determined by X-ray diffraction analysis for the first time. In the solid state, the Ni(I) complex was less susceptible to oxidation in air than the triphenylphosphine complex, indicating greatly improved solid-state stability. In contrast, the Ni(I) complex in solution can easily liberate the phosphine, high catalytic activity toward the Kumada–Tamao–Corriu coupling of aryl bromides.


Introduction
Recent developments in the field of nickel catalysis have opened up possibilities of new catalytic processes directly involving monovalent nickel complexes in organometallic chemistry [1][2][3][4][5][6] along with those involving conventional zerovalent nickel catalysts. Although limited to the structures of specific ligands and complexes, monovalent nickel complexes are thermally stable and isolable [7]. Several studies have used well-defined nickel(I) complexes as catalyst precursors [8][9][10][11][12][13][14], and reactions where nickel(I) complexes are the key intermediates have been proposed based on theoretical calculations [15][16][17][18][19][20]. Recently, experimental results have also supported the catalyst chemistry of nickel(I) [21][22][23]. One of the current bottlenecks in nickel(I) catalysis research is developing catalytically active, air-stable nickel(I) catalyst precursors. It is safe to say that a breakthrough here would go a long way in furthering our knowledge about nickel catalyst chemistry.
In the course of our research, we have found a desirable ligand (fluorine-substituted monodentate phosphine) to stabilize three-coordinated nickel(I) complex. Intriguingly, when using P(Ar m -CF3 ) 3 , the obtained monovalent nickel complex in the crystalline state did not suffer air oxidation for some minutes, while simultaneously maintaining the catalytic activity in the Kumada-Tamao-Corriu cross-coupling reaction in solution under an inert-gas atmosphere. Here, we have reported the synthetic procedure, structures, and properties of the nickel(I) complexes and results in a Kumada-Tamao-Corriu cross-coupling reaction. This is the first example of a well-defined nickel complex bearing the m-CF 3 -substituted triarylphosphine.
The 1 H-NMR spectral measurement for the crystals of 2a-2c dissolved in C 6 D 6 revealed a partial regeneration of the starting complex 1 accompanied by free triarylphosphine from these products. A similar phenomenon is observed in a solution of the previous Ni(I) chloride analogues, suggesting the existence of coordination-elimination equilibrium of weakly coordinating triarylphosphine in solution [34]. It is notable that the ratio of the regenerated complex 1 from the mononuclear complex 2c was only slightly smaller than the ratio from 2a.

X-ray Crystallography and Theoretical Studies
The structures of complexes 2a, 2b, and 2c were confirmed by X-ray crystallography by using single crystals after recrystallization. Figure 1 shows representative examples of the structures of 2a and 2c. The result from the crystal of 2b was insufficient to discuss the detailed structure, unfortunately. There were no significant differences between the 2a and 2c structures, as far as the bond lengths and angles of nickel-bromine bonds are concerned. (Table 1). On the other hand, some differences were observed in the bond lengths between the nickel-phosphorus and the nickel-carbene carbon atoms: Ni(1)-P(1), 2.201(1) (2a) and 2.182(1) (2c); Ni(1)-C(1), 1.918(5) (2a), and 1.936(3) Å (2c). Similar differences between the metal-phosphorus atom in platinum(II) complexes bearing PPh 3 and P(Ar m -CF3 ) 3 were reported as ca. 0.02 Å [36]. The extension of the nickel-carbon bond in 2c appears to be in balance with the strength of the back-donation to the nickel-phosphorus bond.
106.44 (4) 109.29 (4) A space-filling model of 2c was depicted in Figure 1c,d to show the coordination sphere of nickel. The substituents of P(Ar m -CF3 ) 3 largely occupied the remaining spaces other than IPr and bromine around nickel. In particular, one CF 3 on the aryl group is sandwiched between the two isopropyl groups contained in the two wingtip groups of IPr, which hampers the free rotation of the Ni-P axis. If so, the 19 F signals from the CF 3 groups could not be equivalently observed. The averaged equivalent 19 F signal of 2c may be provided by the fast equilibrium of elimination and the coordination of phosphine. Because this complex has only one phosphine ligand, intramolecular CF···π interaction between phosphine ligands did not exist in the crystal structure.
The distribution of the single electron-occupied molecular orbital (SOMO) of 2a and 2c was investigated by performing single-point density functional theory (DFT) calculations (B3LYP/6-31G(d,p) level) implemented in the Gaussian 16 program package [37] at the fixed geometries given by the crystallographic coordinates. Similar electron distributions were given in both complexes as shown in Figure 2a,b. The unpaired electron in SOMO is distributed mainly in one d orbital of nickel, resulting in the formation of two σ*(d-n σ ) orbitals with σ-type non-bonding orbitals (n σ ) of the phosphorus atom and the NHC unit, and the π*(d-p) orbital between nickel and bromine. Because of the electron-withdrawing property of the CF 3 groups, the acceptor ability of the arylphosphine may be enhanced in 2c. The ratio of the unpaired electron in the nickel d-orbital was 34.9% in 2c, lower than that of 2a (41.4%), whereas that in the phosphorus atom was 12.4% and 10.6% in 2c and 2a, respectively. Interestingly, in the case of bromine, the figure was much larger in 2c (28.3%) than that in 2a (21.7%). The reason for this may be the fact that the energy level of NHC-Ni-PPh 3 unit in 2c providing the main component of SOMO is lower-shifted upon the substitution of the CF 3 group into arylphosphine to become closer to the 4p orbital of bromine. This can lead to a greater occupation of the 4p orbital in the Ni-Br π*-orbital component of SOMO as compared to that of 2a.

Oxidation Potentials
It was found that the stability of 2c in air was improved more than that of 2a. In fact, in the solid state, it had not been degraded by oxidation for a period of time: the surface crystal color of 2c began to change gradually after several minutes upon oxidation in air, whereas 2a was oxidized as soon as it was exposed to air. On the other hand, the resistance to oxidation could be evaluated by cyclic voltammogram (CV), which was measured with THF solutions of 2a or 2c, an excess amount of free phosphine (2 equiv), and Bu 4 NPF 6 . The free phosphine (PPh 3 for 2a and P(Ar m -CF3 ) 3 for 2c) was added in order to suppress the generation of the Ni(I) dimer 1 with the liberation of phosphine in equilibrium. As shown in Figure 3, in the CV measured using Ag/AgCl as a reference electrode, the one-electron oxidation potential of 2a was 1.0 V, and that of 2c was higher, 1.3 V. These were not oxidation potentials of the free phosphines. Therefore, it is believed that the oxidation resistance of the nickel center is shown to be improved by the fluorine-substituted phosphine, which may also be reflected in the stability in the solid state, along with the fluorine-derived inter-or intramolecular interactions.

Catalytic Performance for Cross-Coupling Reaction
The Kumada-Tamao-Corriu coupling of aryl bromides using complexes 2a, 2b, and 2c were representatively conducted under the same reaction conditions. Because the equilibrium ratio of the regenerated dimer 1 observed in the 1 H-NMR spectrum was only slightly different between 2a and 2c, it is expected that the catalytic activity of these complexes will not change. As shown in Scheme 3, 0.5 mol% of 2a-2c were added to the reaction media in the presence of 5 equiv of the corresponding triarylphosphine, and the mixture was stirred under an inert gas atmosphere at room temperature for 18 h. The addition of the excess amount of triarylphosphine can shift the ligand-elimination equilibrium from the dimer 1 toward the mononuclear complexes 2. In the reaction of 4-bromotoluene with phenylmagnesium bromide, the product 1-methyl-4-phenylbenzene was successfully obtained in excellent yields (86, 94, and 89% using 2a, 2b, and 2c, respectively). More inactive substrate (4-bromoanisole) also produced 1-methoxy-4-phenylbenzene in high yields (76, 79, and 91% with 2a, 2b, and 2c, respectively). These results indicated that the reactions proceeded efficiently under these conditions, regardless of the nature of triarylphosphines, as expected.

Discussion
As noted above, apparently 2c is more stable in air than 2a in the crystalline state. It is believed that the stability in the crystalline state is derived from several factors. One of them is a steric effect. X-ray crystallography demonstrated that the bulky meta-CF 3 substituted phosphine buried the coordination sphere around the nickel center other than those of the NHC and bromine, kinetically blocking the facile access of small molecules such as dioxygen. As is supportde by the discussion above, more hindered, ortho-fluorine substituted P(C 6 F 5 ) 3 was difficult to coordinate to nickel. Additionally, the increase in the oxidation potential from Ni(I) to Ni(II) for 2c should lead to the resistance of the nickel center to oxidation in air, compared with 2a. However, in solution, the stability of the complex derived from the phosphine should be greatly reduced by its facile elimination to form the Ni(I) dimer 1, enabling a similar catalyst performance of 2c to that of 2a in Kumada-Tamao-Corriu coupling of aryl bromides via the dinickel reaction pathway [9].
As indicated in the literature, the presence of intermolecular interactions in the crystals formed with fluorine atoms may not be negligible [29]. This hypothesis for the specific intermolecular interactions was strongly supported by the results of X-ray crystallography. It should be noted that there are characteristic, multiple short contacts involving fluorine atoms between 2c molecules, such as the F···H interaction observed with hydrogens of aryl para-C-H in phosphine and isopropyl CH 3 groups in IPr, as well as many F···F interactions (Figure 4b and ESI). Such F···F interactions have been visualized as a strong dispersed interaction (i.e. van der Waals interaction) in perfluoropolymers [38] and perfluoroalkanes [39]. Interestingly, such short contacts were mostly distributed in layers, suggesting that these interactions could be a driving force to induce and subsequently stabilize the crystal packing. In this packing structure, the CF···π interaction was not observed, in contrast to the palladium chemistry [29]. On the other hand, in a crystal of 2a, the fewer intermolecular short contacts were observed only between C-H groups (Figure 4a and ESI). It was difficult to evaluate the strength of the intermolecular interactions experimentally, because thermal measurements such as differential scanning calorimetry (DSC) or thermo gravimeter (TG) may break the metal-phosphorus and metal-carbon bonds more easily. However, it is believed that the higher oxidation potential and these steric and electronic factors specifically derived from meta-CF 3 substituted phosphine contribute to the air-stable nature of 2c in the solid state compared with that of 2a, at present.

Materials
Super-dehydrated grade THF, toluene, and hexane were used as solvents as purchased from WAKO Pure Chemical Industries, Ltd., Tokyo, Japan. Benzene-d 6 was distilled from sodium benzophenone ketyl and stored under a nitrogen atmosphere. Organic reagents used for coupling reactions were distilled just before use or used as purchased. N-Heterocyclic carbene (IPr) [40] and nickel dimer, [Ni(IPr)(µ-Br)] 2 (1) [35], were prepared according to the literature methods.

Methods
All experiments were carried out under an inert gas atmosphere using standard Schlenk techniques and a glove box (MBraun UniLab, München, Germany) unless otherwise noted. Column chromatography of organic products was carried out using silica gel (Kanto Kagaku, silica gel 60 N (spherical, neutral), Tokyo, Japan). The 1 H-NMR spectra were taken with a Bruker Avance-III400 Y plus 400 MHz spectrometer (Bruker BioSpin, Billerica, MA, USA) at room temperature. Chemical shifts (δ) were recorded in ppm from the solvent signal. The magnetic properties of the materials were investigated using a Quantum Design MPMS-5S superconducting quantum interference device (SQUID) magnetometer (Quantum Design Inc., San Diego, CA, USA). The elemental analysis was carried out with J-Science CHN Corder JM-11 (J-Science Lab Co. Ltd., Kyoto, Japan), equipped with AUTO-SAMPLER, using tin foil, where the samples were held in a glove box. The X-band EPR measurements were collected with a Bruker EMX Plus spectrometer (Bruker BioSpin, Billerica, MA, USA) equipped with a continuous flow N 2 cryostat.

X-ray Crystallography
Single crystals of 2a, 2b, and 2c for X-ray diffraction were grown at −30 • C from toluene/hexane (2a) and THF/hexane (2b and 2c) solutions. All the data were obtained at 125 (2a), 151 (2b), and 180 K (2c) using a Rigaku Saturn CCD diffractometer with a confocal mirror and graphite-monochromated Mo Kα radiation (λ = 0.71070 Å). Data reduction of the measured reflections was performed using the software package CrystalStructure [41]. The structures were solved by direct methods (SHELXT-2014) [42] and refined by full-matrix least-squares fitting based on F 2 , using the program SHELXL-2014 [43]. All non-hydrogen atoms were refined with anisotropic displacement parameters. All hydrogen atoms were located at ideal positions and included in the refinement but were restricted to riding on the atom to which they were bonded. Unfortunately, the refinement for the crystal 2b cannot be completed, and only the preliminary structure was shown in the supporting information. CCDC 1947417-1947418 contains the supplementary crystallographic data of 2a and 2c for this paper. A copy of the data can be obtained free of charge via http://www.ccdc.cam.ac.uk/cgi-bin/catreq.cgi.

Theoretical Details
All the DFT calculations were performed utilizing the GAUSSIAN 16 Rev. A.03 program package (Gaussian Inc., Wallingford, CT, USA) [37]. The B3LYP functional was employed with a standard split valence-type basis set, 6-31G(d,p). The single-point calculations to obtain SOMOs were carried out using crystallographic coordinates without geometry optimization, and done with a tight self-consistent field (SCF) convergence criterion. All the computation was carried out using the computer facilities at Research Institute for Information Technology, Kyushu University, Fukuoka, Japan.

Electrochemistry
The cyclic voltammogram of 2a and 2c were recorded on an ALS/chi electrochemical analyzer Model/610A with a platinum working electrode, a silver wire reference electrode, and a platinum wire counter electrode, with a scan rate of 50 mV·s −1 . The analyte solutions of these complexes were prepared with a 0.1 M solution of tetra-n-butyl ammonium perchlorate in acetonitrile. Ferrocene was used as an internal standard and the potential reported here is referenced to the ferrocene/ferricinium couple.

Conclusions
In summary, the Ni(I)-IPr complex bearing meta-CF 3 substituted triarylphosphine ligand was successfully isolated. Although the "superstable" Pd(0) catalyst has been derived using palladium and phosphine, such a triarylphosphine complex using nickel has been determined here for the first time. In the solid state, steric bulk, appropriate electron-withdrawing properties, and the presence of intermolecular interactions make nickel less susceptible to oxidation in air where the triarylphosphine is coordinated. Particularly, intermolecular multiple interactions including those between F and H, and F and F were found to be gathered in layers in the crystalline state. On the other hand, it is possible to maintain the same catalytic activity as the analogous PPh 3 complex, since the phosphine can be easily eliminated in solution, resulting in the generation of active species. In the future, we plan to work on the development of other useful catalytic reactions and detailed mechanistic research using this Ni(I) complex.
Supplementary Materials: The supplementary materials are available online. Figure S1-S8: NMR spectra for the complexes, Figure S9: χ mol T vs T plot, Figure S10: ESR spectrum, Figure S11-S12: Packing views of the complexes (expanded views), details of the Kumada coupling reactions including NMR spectra of products.
Author Contributions: T.F. and T.I. conceived and designed the experiments; R.H. performed the experiments; Y.K. analyzed the data; Y.Y. contributed theoretical calculations; K.M. managed this research and wrote the paper.