Nickel-catalyzed coupling reaction of alkyl halides with aryl Grignard reagents in the presence of 1,3-butadiene: mechanistic studies of four-component coupling and competing cross-coupling reactions

The detailed reaction mechanism of anionic Ni complex-promoted C–C bond forming reactions was clarified by experimental and theoretical methods.


Reaction condition screening
Using the optimized conditions for ortho-unsubstituted aryl Grignard reagents (entry 4 in Table 1), we examined various parameters to improve the selectivity for the four-component coupling product 3aa. Selected examples are summarized in Table S1. The use of 20 mol% of Ni catalyst and PPh3 did not change the selectivity, but slightly affected the reaction efficiency (entry 1 vs. 2).
1,1'-Bis(diphenylphosphino)ferrocene (dppf) depressed the selectivity to 0.9 (entry 3. Tetramethylethylenediamine (TMEDA) improved neither the total yield nor the selectivity (entry 4). The addition of LiI largely affected the selectivity, yielding 4aa as the major product (entry 5). The addition of MgBr2 afforded a 1:1 mixture of 3aa and 4aa in 61% total yield (entry 6). When the reaction was conducted in a mixed solvent of toluene and THF in 1:1 ratio, the yields of both 3aa and 4aa decreased to 24% and 31%, respectively, whereas the selectivity (3aa/4aa) dropped to 0.8 (entry 8). A similar tendency was observed in the mixed solvent system of hexane and THF (entry 9). On the other hand, the use of polar solvents such as dioxane and HMPA as co-solvents affected both four-component and cross-coupling reactions, resulting in a low yield (25% total yield) and no reaction, respectively.

3-2. Four-component coupling reaction using the nickelate complex 7 (eqn 4) S1
An oven-dried test tube was charged with the formed nickelate complex 7 (14.1 mg, 5 mol%) and a stirring bar and closed with a septum cap in a glove box. The test tube was brought out of the glove box and into it were added 1-fluorooctane (66.6 mg, 0.5 mmol), THF (0.6 mL), 2,6-dimethylphenylmagnesium bromide (in THF, 0.83 M, 0.9 mL, 0.75 mmol), and 1,3-butadiene (34 mL as gas, 1.5 mmol) by syringe at -78 ºC. The mixture was then stirred at 40 ºC. After 10 h, the reaction mixture was diluted with Et2O, carefully quenched by 1N HCl aq., and analyzed by GC using dodecane as an internal standard to determine the yield of the desired product 3am (88%). S10

3-3. The reaction of complex 11 with alkyl bromides
When complex 11 (119 mg, 0.20 mmol) was treated with n-OctBr (52 L, 0.30 mmol) in THF at rt for 1 h, 0.23 mmol of n-OctBr was consumed to give the cross-coupling product, n-Oct-Ph, in 37% yield along with biphenyl (0.06 mmol) (eqn S1). This result shows that alkyl bromides could react with anionic Ni complex having Li as the cation in sharp contrast with the case of alkyl fluorides (Schemes 6 and 7).

5-2. Activation parameters (Fig. 4)
The reactions using 2j and 2m were conducted between 30 °C and 50 °C, and time-course of the reactions was traced as shown in Figs. 4a and 4b. Under the conditions with a constant amount of Ni catalyst, the reaction rate obeys pseudo-first-order kinetics for 1a with v = k[1a][NiBr2(dme)] = kobs [1a].

S18
From the slope of these plots, kobs at each reaction temperatures were determined, and the k (= kobs[NiBr2(dme)]) are summarized in Table S8. Eyring plot of these data are shown in Fig. 4c.

5-3. Hammett plot regarding RDS (Section 3.6)
The reaction of n-OctF (1a) with aryl Grignard reagents having a methyl group at the ortho-position and various substituents at the para-position was conducted, and the obtained results are shown in Table   S9.  shows plots of ln([1a]/[1a]0) against reaction time. From the slope of the plots, kobs was obtained. The rate parameters and Yukawa-Tsuno's  0 are summarized in Table S10. Hammett plot of these data is shown in Fig. 5.

5-4. Competitive reactions (Section 3.8)
The competitive reactions shown in Table 7 were conducted, and the amounts of the product formed are plotted against reaction time as shown in Fig. S14, where the competitive reaction of o-TolMgBr (2j) and 4-fluoro-2-methylphenylmagnesium bromide (2o) was difficult to analyze by GC. Therefore, the relative reactivity of 2o against 2j was estimated by the competitive reaction with 2q (Fig. S14b). The ratio of reaction constants, kX/kH, was directly calculated from the slope of the plots, and the obtained data are summarized in Table 7. Among substrate constants tested, p + showed the best linearity.
Hammett plot of the data against p + is shown in Fig. 9.

6-2. Optimized structure and dihedral angle of INT1 with different aryl groups
The structures of the anionic part of INT1 and their dihedral angles between the aryl and the coordination plane of the Ni are shown in Fig. S16. In the case of INT1-2,6-dimethylphenyl, the 2,6-dimethylphenyl group is orthogonally bound to Ni, in good agreement with the X-ray structure. In the case of INT1-Ph, the dihedral angle of the optimized structure is 118.2°, in sharp contrast to the crystal structure. S7 NMR studies of the nickelate complexes bearing Ph and 2,6-diemthylphenyl groups at -20 °C show symmetric Ph group and unsymmetric 2,6-dimethylphenyl group (vide supra). These results indicate that the phenyl group easily rotates even at low temperatures. However, when methyl groups are introduced into the ortho-positions, the rotation becomes slow or restrained due to the steric interaction.

6-3. Rotation of aryl group in anionic Ni complexes
As mentioned above, 1 H NMR of isolated anionic Ni complexes 11 (Ph) and 8 (2,6-dimethylphenyl) showed symmetric Ph and unsymmetric 2,6-dimethylphenyl groups at -20 °C, respectively. These observations clearly indicate that the rotation of aryl ring is restrained by ortho-substituent(s). To evaluate rotation barrier of the anionic Ni complex having 2-methylphenyl group, we conducted theoretical calculation with the dihedral angle restriction. The initial structure with the dihedral angle of 98.6° is obtained directly from INT1 structure by removing MgBr·4THF. Plot of relative energy against dihedral angle is shown in Fig. S17, revealing rotation barrier to be ca. 55 kJ/mol. To confirm the rotation barrier, we further investigated by VT NMR. The anionic Ni complex having 2-methylphenyl group was synthesized by the reaction of Ni(cod)2 with 2-methylphenyllithium reagent in the same way as the case of complexes 7 and 11. 1 H NMR spectrum of the isolated complex showed two signals arising from ortho-methyl group at 4.99 and 4.47 ppm at -20 °C in THF-d8.
Rotation rate was estimated by line-shape analysis of the 1 H NMR spectra recorded at -60 to -20 °C using gNMR simulation software S8 . The obtained rate constants are summarized in Table S11. Eyring S25 plot of the obtained data is shown in Fig. S18. From the slope and intercept, H ‡ and S ‡ are determined to be 24.2 ± 2.8 kJ/mol and -104 ± 12 J/Kmol, respectively. Therefore, the rotation barrier G ‡ 273 is 52.7 ± 6.0 kJ/mol, the value which is consistent with the calculated value of 55 kJ/mol.

6-4. Geometry of Ni complexes bearing 2-methylphenyl group
Because of the orthogonal orientation of the aryl group, Ni complex bearing 2-methylphenyl group has two possible geometries, in which the -allyl group and methyl group on the 2-methylphenyl group point in either the same (up) or opposite (down) directions. We thus calculated TS3 and TS10 with n-Pr-F for both geometries at M06/6-31G(d,p) level of theory, and the results are summarized in Fig.   S19. In TS3, TS3-2-methyl-up is more favorable compared to TS3-2-methyl-down. In contrast, in TS10, TS10-2-methyl-down is favored rather than TS10-2-methyl-down due to the steric repulsion between ortho-methyl group and n-Pr moiety. It should be noted that TS10-2-methyl-down is the most favorable in energy, being inconsistent with the experimental results. In both TS10-2-methyl-up and  Table S12 summarizes important bond distances in TS10 with various Grignard reagents and alkyl fluorides, e.g. in TS10 with Ph and MeF, the C-F bond is 1.28 times longer than MeF (C-F = 1.379 Å), and the C-Ni bond is 1.28 times longer than INT11 (C-Ni = 1.914 Å). The C-F and C-Ni distances in TS10 with Ph and n-OctF are longer than those in TS10 with MeF, where both bonds are 1.36 times and 1.35 times elongated, respectively. Although this step is the nucleophilic substitution of alkyl fluorides with anionic Ni center by the aid of the Mg cation, the details of the reaction mechanism of this step differs according to alkyl fluorides. In the case of MeF, the reaction proceeds through pure SN2 mechanism. On the other hand, the process of n-OctF proceeds through SN2 mechanism with a somewhat SN1 character. This difference may arise from the stability of the carbocation, where methyl cation is unfavorable than n-Oct cation. Therefore, the C-F and C-Ni distances of TS10 with MeF are shorter than that of TS10 with n-OctF. When methyl group(s) was introduced into the ortho-positions, both C-F and C-Ni bonds are elongated probably due to the steric repulsion, and the elongation of these bond distances is more significant in the case of n-OctF. For instance, the change of the sum of C-F and C-Ni bonds is 0.112 Å for n-Oct but only 0.051 Å for MeF by changing Ph to 2,6-dimethylphenyl. This result could be explained by the stability of the carbocation as mentioned above. In the case of MeF, the corresponding Me cation is relatively unstable and therefore elongation of C-F and C-Ni bond in TS10 is impermissible, resulting in the increase of the energy barrier by steric repulsions. On the other hand, the steric repulsion is loosened by elongation of these bonds in the case of TS10 with n-OctF. Therefore, these TSs with n-OctF are not sensitive to steric hindrances compared to TSs with MeF.  Table S13. Among key atoms in the bond cleavage and formation, the NBO charge of the reacting carbon is drastically altered by changing MeF to n-OctF (from -0.137 to +0.101). This large difference in NBO charge clearly supports the above mentioned difference in mechanism and the SN1 character of TS10 with n-OctF. The relatively S29 small difference of Ni in NBO charge is due to the -allyl group in trans-position (vide infra).  Table S14. It should be noted that electronic effects of the substituents affect the NBO charge of the -carbon in the -allyl group. However, that of Ni atom is constant in all cases. This is probably due to the electronic transfer between Ni and the -allyl moiety at the trans-position. Indeed, the total NBO charge of Ni and the -allyl depends on the substituents. were purchased from Kanto Chemical Company and purified by SPS S9 prior to use. Dehydrated toluene and pentane were purchased from Wako Pure Chemical Industries and used as received. THF-d8 and 1,2-dimethoxyethane (DME) were distilled from sodium benzophenone ketyl. 12-Crown-4 was distilled from sodium metal. Grignard reagents 2f, 2k, 2l, and 2m were prepared by a standard procedure. Other THF solutions of Grignard reagents were purchased from Aldrich or TCI. A cyclohexane/ether solution of phenyllithium was purchased from Kanto. Ether solutions of 2,6-dimethylphenyllithium and 2-methylphneyllithium were prepared by the reaction of the corresponding bromide with lithium under an Ar atmosphere. These organomagnesium and lithium reagents were used after titration using I2.
The resulting mixture was refluxed for 2 days. Since reaction was not completed, solvent was changed to DMF (5 mL) and reaction was continued at 100 ºC for 40 h. The reaction mixture was cooled to ambient temperature, diluted with Et2O and washed with water for three times and brine. The organic layer was dried over Na2SO4 and concentrated. Purification by silica gel column chromatography A mixture of the chloride (2.022 g, 9.86 mmol) and n-Bu4N•F(H2O)5 (6.94 g, 19.7 mmol) was stirred at 80 ºC for 8 h. After addition of H2O (10 mL), products were extracted by EtOAc for three times. The combined organic layer was dried over Na2SO4, through a short pad of silica gel, and concentrated to give an orange oil. The fluorinated product was purified silica gel column chromatography (hexane/EtOAc = 100/0 to 90/10) to give the title compound as a colorless oil (1.138 g, 6.03 mmol,