A Zinc Catalyzed C(sp3)−C(sp2) Suzuki–Miyaura Cross‐Coupling Reaction Mediated by Aryl‐Zincates

Abstract The Suzuki–Miyaura (SM) reaction is one of the most important methods for C−C bond formation in chemical synthesis. In this communication, we show for the first time that the low toxicity, inexpensive element zinc is able to catalyze SM reactions. The cross‐coupling of benzyl bromides with aryl borates is catalyzed by ZnBr2, in a process that is free from added ligand, and is compatible with a range of functionalized benzyl bromides and arylboronic acid pinacol esters. Initial mechanistic investigations indicate that the selective in situ formation of triaryl zincates is crucial to promote selective cross‐coupling reactivity, which is facilitated by employing an arylborate of optimal nucleophilicity.


General Remarks
Unless otherwise stated all manipulations were carried out using standard Schlenk techniques under argon or in an MBraun UniLab glovebox, under an atmosphere of argon. THF, 2MeTHF, dioxane and cyclopentyl methyl ether (CPME) were dried over and distilled from potassium and stored over activated 3 Å molecular sieves. Hexane was dried and distilled from either calcium hydride or NaK alloy and stored over a potassium mirror. All other reagents were purchased from commercial chemical suppliers and used as received. NMR spectra were recorded on Bruker AvanceIII-400, Bruker AvanceII-500 or Bruker Ascend-400 spectrometers. Chemical shifts are reported as dimensionless values and are frequency referenced relative to residual protio-impurities in the NMR solvents for 1 H and 13 C{ 1 H} respectively, while 11 B{ 1 H}, 19 F{ 1 H}, 7 Li and 31 P shifts are referenced relative to external BF3-etherate, hexafluorobenzene, LiCl, and H3PO4 respectively. Coupling constants J are given in Hertz (Hz) as positive values regardless of their real individual signs. The multiplicity of the signals are indicated as "s", "d", or "q" for singlet, doublet, or quartet respectively. GC-MS analysis was performed on either of two instruments. An Agilent Technologies 7890A GC system equipped with an Agilent Technologies 5975C inert XL EI/CI MSD with triple axis detector, fitted with a HP-5Ms column, with dimensions 30 m length; 0.250 mm internal diameter; and 0.25 μm film. Or an Agilent Technologies 6890N GC equipped with an Agilent Technologies 5973N EI MSD, fitted with a HP-5MS column, with dimensions 30 m length; 0.250 mm internal diameter; and 0.25 μm film.
The relative response factors for GCMS analysis of the heterocoupled and homocoupled products derived from the fluorinated electrophile, 2b, were calculated using values from 19 F{ 1 H} NMR spectra (with a delay time of 35s to allow full spin-lattice relaxation) where their integrals could be measured accurately. When these resonances were overlapped in the 19 F { 1 H} NMR spectra (which occurred in a number of solvents), GCMS analysis was used to calculate their ratio, and yields calculated by using the overall integral of the overlapped peak with this ratio applied (accounting for the 2 equivalents of electrophile involved in the homocoupled product). The relative response factors used in this calculation for GCMS analysis were calculated from the results where 19 F resonances could be accurately integrated. Yields are based on the electrophile as the limiting reagent and for homocoupling impurities the 19F integrals are scaled by 0.5 to give a molar ratio vs. heterocoupling (i.e. A 1:1 hetero:homocoupled product ratio at full conversion, would be reported as 33 % : 33 %, due to the additional molecule of starting material required in the production of the homocoupled product). In a number of cases the 13 C resonance for the carbon atom directly bonded to boron was not observed due to the effect of quadrupolar relaxation.

Synthesis of borate nucleophiles General Procedure
The borates were synthesised according to a modified literature procedure 1 . In an oven dried Schlenk flask the appropriate arylboronic acid pinacol ester (1-1.05 eq.) was dissolved in anhydrous hexane and cooled to -78 °C before dropwise addition of tert-butyllithium (1.7M in pentane, 1 eq.). The reaction was allowed to warm to room temperature and stirred overnight at room temperature, over which period a precipitate formed. The borate was isolated by filtration, washed with anhydrous hexane and residual solvent removed under reduced pressure.

Synthesis of [Li][(O t Bu)(Ph)B(Pin)] (6)
An oven dried Schlenk tube was charged with phenyl boronic acid pinacol ester (500 mg, 2.5 mmol) and anhydrous hexane (9 ml) and a solution of lithium tert-butoxide (200 mg, 2.5 mmol) in hexane (10 ml) was added slowly. The homogenous mixture was stirred and of 2.5 ml THF was added, before stirring for a further 4 hrs. During this time, a white precipitate had formed, which was isolated by filtration and washed with anhydrous hexane (2 x 4ml). Residual solvent was removed under reduced pressure, giving a free flowing white powder in 258 mg. The solvent components were combined and stirred overnight leading to more precipitation which was isolated above to give a second crop of 133mgs. Combined yield of both crops = 56 %.

Preliminary catalysis investigations
In an oven dried J Young's ampoule [Li][( t Bu)(Ph)B(Pin)] 1a (94 mg, 0.35 mmol) and ZnPh2(5 mg, 0.02 mmol) and 3-methoxybenzyl bromide (32 µl, 0.23 mmol) were combined and dissolved in the appropriate solvent (2 ml). The reaction was heated to the desired temperature for 17 hours prior to addition of mesitylene (32 µl, 0.23 mmol) as an internal standard and transfer to an NMR tube under ambient conditions. Subsequently the mixture was diluted with DCM, filtered through a silica plug and analysed by GCMS.

Solvent optimisation reactions General Procedure
In an oven dried J Young's ampoule [Li][( t Bu)(Ph)B(Pin)] 1a (94 mg, 0.35 mmol) and ZnBr2 (5 mg, 0.02 mmol) were dissolved in the appropriate solvent (2 ml). Then 4-fluorobenzyl bromide (29 µl, 0.23 mmol) was added. The reaction was heated to the appropriate temperature for 18 hours before quenching with ethanol followed by addition of fluorobenzene (22 µl, 0.23 mmol) and mesitylene (32 µl, 0.23 mmol) as internal standards. The mixture was directly analysed by 19 F NMR spectroscopy before dilution with DCM, filtration through a silica plug and analysis by GCMS. The fraction at 9.65 min retention times has a m/z of 202.1 thus is not heterocoupling or biphenyl. Currently it is an unidentified by-product from the reaction S13

Kumada coupling with ZnBr2 and with FeBr2
ZnBr2 (5mg, 10 mol%) was added to an ampoule and dissolved in 1.5 ml 2-MeTHF. 4fluorobenzylbromide was then added (29l, 0.233 mmol, 1 equiv). PhMgBr (480l, 0.725 M solution in 2-MeTHF, 0.35 mmol 1.5 eq.) was added slowly and the reaction was heated at 60 o C for 18 h. The mixture was quenched by addition of 0.2 ml EtOH followed by addition of fluorobenzene (22 µl, 0.23 mmol) and mesitylene (32 µl, 0.23 mmol) as internal standards. The mixture was directly analysed by 19 F{ 1 H} NMR spectroscopy before dilution with DCM, filtration through a silica plug and analysis by GCMS.
An identical procedure was used replacing ZnBr2 with FeBr2.

Attempted synthesis of cycloheptyl benzene
In an oven dried ampoule 1a (141 mg 0.525 mmol, 1.5eq) was dissolved in 2.25ml 2MeTHF. To this was added 750 µl of a 0.047M stock solution of ZnBr2 in 2MeTHF (0.035mmol, 0.1 eq.), immediately followed by cycloheptyl bromide (48 l, 0.35 mmol, 1 eq.). The reaction was heated at 60 °C for 24 hours before quenching with EtOH (~2ml), followed by extraction into DCM (3 x ~10 ml) and removal of solvent under reduced pressure. An aliquot was then taken for analysis by GC-MS. The desired product was produced in only trace quantities, with the major product being cycloheptene. This is in contrast to work by Bedford et al using iron catalysts which efficiently couple cycloheptyl bromide with 1a 4 .

S15
An analogous reaction was run using octylbromide in place of cyclohepthylbromide under otherwise identical conditions. Analysis by GC-Ms again showed minimal coupling with the major species being the starting electrophile along with triphenylboroxine.

Procedure for reaction using NaBPh4
In an oven dried ampoule NaBPh4 (180 mg, 0.525 mmol, 1.5eq) was dissolved/suspended in 2.25ml 2MeTHF. To this 750 µl of a 0.047M stock solution of ZnBr2 in 2MeTHF (0.035mmol, 0.1 eq) was added, immediately followed by 4-fluorobenzyl bromide (44 µl, 0.35 mmol). The reaction was heated at 60 °C for 24 hours before quenching with ethanol (~0.2ml), followed by extraction into DCM (3 x ~10 ml) and removal of solvent under reduced pressure. The solid was dissolved in CDCl3 followed by addition of fluorobenzene (22 µl, 0.23 mmol) and mesitylene (32 µl, 0.23 mmol) as internal standards. The mixture was directly analysed by 19 F{ 1 H} NMR spectroscopy and analysis by GCMS. A J. Youngs NMR tube was loaded with 4-Br-C6H4-BPin (84.9 mg, 0.3 mmol) and dissolved in a 0.5 ml mixture of C6D6/C6H6 and then ZnEt2 was added (0.3 mL of a 1 M hexanes solution) to furnish a colourless homogenous reaction mixture. After rotating for 30 minutes multinuclear NMR spectroscopy revealed that minimal transmetallation had occurred (30 minutes was chosen to be comparable with the transmetallation with borate 1a on page S9). At this stage one equivalent of the electrophile, 3methoxybenzylbromide) was added (0.3 mmol, 42 L) which resulted in no observable change (visibly or by 11 B NMR spectroscopy) even after 1 h at 20 o C. Subsequent heating overnight led to minimal Csp 2 -Csp 3 coupling (as indicated by the resonance at 30.3 ppm in the 11 B NMR spectrum dominating which is consistent with the starting arylBPin reagent). Benzene was chosen as reaction solvent in this case to maximise the transmetallation and subsequent coupling (coordinating solvents would hinder both steps by binding to the zinc Lewis acids).

Substrate scope screening reactions and experimental data General Procedure
In an oven dried ampoule the appropriate borate salt (0.525 mmol, 1.5eq) was dissolved in 2.25 ml 2MeTHF. To this was added 750 µl of a 0.047M stock solution of ZnBr2 in 2MeTHF (0.035 mmol, 0.1 eq), immediately followed by the alkyl halide (0.35 mmol, 1 eq.). The reaction was heated at 60 °C for 24 hours before quenching with 1M aqueous HCl (~2ml), followed by extraction into DCM (3 x ~10 ml) and removal of solvent under reduced pressure. Triphenylmethane (85.5 mg, 0.35 mmol, 1 eq.) or mesitylene was added as an internal standard for NMR yield calculations and the mixture was dissolved in CDCl3 and analysed by NMR spectroscopy. An aliquot was then taken for analysis by GC-MS.

1-methyl-4-(4-(trifluoromethyl)benzyl)benzene (3k)
Synthesised according to the above general procedure from 4-(trifluoromethyl)benzyl bromide (54 l) and [Li][( t Bu)(p-tol)B(Pin)] (148 mg). 1 H NMR spectroscopic data is consistent with previously reported values 9 . Yield 75% using triphenylmethane as an internal standard. Purification of the crude mixture by silica gel column chromatography was attempted (using PET ether as eluent), although NMR analysis showed the product was present (25 mg 75 % yield) in 87% purity (therefore yield of the desired C2 functionalised = 65%), due to the presence of 13 % of the 3-isomer from Friedel Crafts functionalisation of the beta thiophene position, related chemistry has been observed previously using anisole-zinc Lewis acid reagents 10 . .

H} (C6D6) NMR spectrum of the products from 3l
The major isomer is assigned as the 2-(4-fluorobenzyl)-5-methyl isomer, while the minor isomer is assigned as the 3-(4-fluorobenzyl)-5-methyl isomer, based on two observations. Firstly, 1 H NMR data for the analogous 2-phenyl-5-methyl-isomer has been previously published 11 , and this shows strong similarity to the major isomer of 3l, particularly the CH2 benzyl resonances (4.07 vs 4.05 ppm). Additionally, HMBC shows that for the minor isomer, where each thienyl resonance is distinct, there is coupling between the benzyl protons and both thienyl carbon (containing a C-H) resonance, which would be much more likely to occur in the 3-(4-fluorobenzyl)-5-methyl isomer, as both of these positions are only separated by 3 bonds, whereas for the 2-(4-fluorobenzyl) isomer, one is separated by 4 bonds. However, since the thienyl resonances of the major isomer are coincident, completely unambiguous assignment is not possible. Finally, in the work reported herein using anisole derivatives very little Freidel Crafts substitution products are observed in contrast to previous work, 10 indicating the organometallic coupling is the preferred process in this work.  , with heating to 60°C for 72 hours. 1 H NMR spectroscopic data is consistent with previously reported values 12 . Yield 59% using triphenylmethane as an internal standard.  . The reaction was quenched with ethanol and the conversion was measured by 19 F NMR spectroscopy. Conversion : 30% after 24 hours.

Reaction of ZnBr2 with [Li][ t BuPhBPin] (1a)
An oven dried J Young's NMR tube equipped with a DMSO-d6 capillary insert was loaded with [Li][ t BuPhBPin] (19 mg, 0.07 mmol) 2-MeTHF (0.7ml). The sample was analysed by 1 H NMR spectroscopy prior to addition of zinc bromide (8 mg, 0.035 mmol) and sonication for 30s. The sample was analysed by 11 B NMR spectroscopy showing complete consumption of the borate starting material and formation of t BuBPin.

Reaction of ZnPh2 with [Li][ t BuPhBPin] (1a)
An oven dried J Young's NMR tube equipped with a DMSO-d6 capillary insert was loaded with [Li][ t BuPhBPin] (86 mg, 0.32 mmol, 2 eq.), mesitylene (22.25 µl, 0.16 mmol as internal standard) and protio-2MeTHF (0.7ml). The sample was analysed by 1 H NMR spectroscopy and diphenyl zinc (35 mg, 0.16 mmol, 2 eq.) was added and the sample heated to 60 °C for 90 minutes. The sample was analysed by 1 H and 11 B NMR spectroscopy showing transfer of 1 phenyl equivalent. The sample was then heated at 60 °C for a further 16 hours. Analysis by NMR spectroscopy showed that 1 equivalent of the neutral t BuBPin had been formed by integration vs. the mesitylene standard, suggesting formation of LiZnPh3 but not Li2ZnPh4.

Reaction of NaBPh4 with ZnPh2
An oven dried J Young's NMR tube equipped with a DMSO-d6 capillary insert was loaded with diphenyl zinc (35 mg, 0.16 mmol, 1 eq.), and sodium tetraphenylborate (55 mg, 0.16 mmol, 1 eq.) and 2-MeTHF (600 µl). The mixture was sonicated for 1 minute to aid dissolution of the solids, after standing at room temperature for 90 minutes, the mixture was analysed by 11 B NMR spectroscopy revealing no phenyl group transfer. The mixture was heated to 60°C and analysed by 11 B NMR spectroscopy after 3 hours and 41 hours, which showed the presence of only unreacted NaBPh4 (no (2-MeTHF)-BPh3 was observed).