Selective electrochemical generation of benzylic radicals enabled by ferrocene-based electron-transfer mediators

The use of ferrocene mediators offers significant advantages over direct electrolysis in the generation and functionalization of radicals from benzylboronates.


Instruments and techniques:
1 H, 13 C, and 19 F NMR spectra were recorded on Bruker 400 and 500 MHz spectrometers. Chemical shifts are given in parts per million (ppm) relative to residual solvent peaks in the 1 H and 13 C NMRs and relative to CFCl 3 in the 19 F NMR. High-resolution mass spectra were obtained using a Thermo Q ExactiveTM Plus by the mass spectrometry facility at the University of Wisconsin. Chromatographic purification of products was accomplished by chromatography on Silicycle P60 silica gel (particle size 40-63 µm, 230-400 mesh) using Teledyne Isco Combiflash R f or Biotage Isolera One flash chromatography systems. Thin-layer chromatography (TLC) was performed on Silicycle silica gel UV254 pre-coated plates (0.25 mm). Visualization of the developed chromatogram was performed by using UV lamps and KMnO 4 stain.
All cyclic voltammetric, chronoamperometric and chronopotentiometric measurements were performed at room temperature using a Pine WaveNow PGstat. The CV experiments were carried out in a three-electrode cell configuration with a glassy carbon (GC) working electrode (3 mm diameter, unless otherwise stated) and a platinum wire counter electrode. The potentials were measured versus an Ag/AgNO 3 (0.01 M) reference electrode (all electrodes from BASi) and recorded against a ferrocene/ferrocenium reference. The GC working electrode was polished with alumina before each experiment. Bulk electrolysis experiments were performed in a custom-built inert atmosphere divided cell with a Schlenk tap attached. Reticulated vitreous carbon (RVC) was used for the working electrode material, a platinum wire for the counter electrode, and Ag/AgNO 3 (0.01 M) for the reference electrode.

S2
Electrochemical cells: Figure S1: Electrolysis cell used for anaerobic electrolyses. Cell is equipped with a gas/vacuum line inlet, cathodic cell separated from the anodic solution by a porous glass frit.

2.
Oxidation potential study of boronates    Oxidation potential study of other boronates

CV studies of electrode fouling
These CV studies were performed by cycling the potential and measuring the current. Following each complete cycle, the solution was stirred and allowed to settle again before initiating another cycle. The electrode was not polished in between each cycle. The rate of the electrode fouling was found to be dependent on the concentration of boronate. The current could only be restored upon polishing and not after washing with solvent.

SEM analysis of electrode fouling
Electrode preparation for SEM before fouling A 1 mm diameter glassy carbon (GC) disk electrode was polished, washed with MeCN and dried. A blank CV of only TBAP (0.1 M) in MeCN (5 mL) was run (100 mV/s, -0.5-1.5 V (vs Ag/Ag + )) followed by a chronoamperometry (CA) experiment (1 V (vs Ag/Ag + ), 60 s) and again a CV (100 mV/s, -0.5-1.5 V (vs Ag/Ag + )). The surface of the GC electrode was rinsed with fresh MeCN, air-dried, and analysed by SEM (see images below).

Electrode fouling and preparation for SEM
The electrode was polished again, washed with MeCN and dried. The electrode was immersed into a solution of 1b (50 mM), TBAP (0.1 M) in MeCN (5 mL) and a CV was run (100 mV/s, -0.5-1.5 V (vs Ag/Ag + )) followed by a chronoamperometry (CA) experiment (1 V (vs Ag/Ag + ), 60 s) and a CV (100 mV/s, -0.5-1.5 V (vs Ag/Ag + )). These CVs are displayed in Figure S14 and clearly show the fouling effect that has occurred. The surface of the GC electrode was rinsed with fresh MeCN, air-dried, and analysed by SEM (see images below).

SEM images after fouling
Figure S17: SE2 images of the GC electrode surface after fouling. Amorphous organic deposits found on the surface were only found to cover <10% of the surface, which otherwise appeared relatively clean and free of the microparticles observed before fouling. This suggests a much thinner, molecular covering of the electrode is responsible for the electrical insulation observed by CV and bulk electrolysis.  Figure S18: The amorphous deposits on the electrode surface after fouling were found to contain fluoride originating from 1b.

Bulk electrolyses
General procedure To the anodic compartment of an oven-dried divided electrochemical cell ( Figure S1) was added a magnetic stir bar, Bu 4 NClO 4 (171 mg) and an appropriate, catalyst (0.01 mmol). To the cathodic compartment was added Bu 4 NClO 4 (171 mg). A reticulated vitreous carbon (RVC) working electrode, Ag/AgNO 3 reference electrode, and glassy carbon disk electrode were inserted to the anodic compartment. A platinum wire counter electrode was inserted in the cathodic compartment. The cell was further dried under vacuum and purged with N 2 . Under a stream of N 2 , TEMPO was added to the anodic compartment. To the anodic and cathodic cells were added 5 mL dry and degassed MeCN.
The benzylic boronic ester 2a (0.1 mmol) and base (0.2 mmol) were added to the anodic chamber. When NaOH was employed, it was pre-stirred with 2a in a Teflon-capped 1.5-dram vial charged with a magnetic stir bar that was then purged with N 2 gas. Dry and degassed 0.1-0.2 mL MeCN was added and the solution was stirred vigorously for 30 s. The solution was then transferred to the anodic compartment of the electrochemical cell.
A CV was obtained of the solution using the glassy carbon disk electrode. If running a constant potential experiment, the potential chosen to be applied was 50-100 mV lower than the relevant oxidation peak. If running a constant current experiment, a 20 s chronoamperogram was collected with a potential applied at 30 mV lower than the relevant peak potential. Bulk electrolysis was then carried out at half the current measured with this chronoamperogram. The reaction was stopped after 1.5 F/mol had been applied or after the measured potential of the solution increased to the oxidation potential of excess TEMPO. A sample from the reaction was then removed and analysed by 19 F NMR. Entries 1 and 2 demonstrate the direct oxidation of 2a with two different bases and trapping with a large excess of TEMPO. Introducing a catalyst to the system (Entry 3) did not improve the yield when TBAF was used as a base due to deactivating interactions between ferrocenium and fluoride, see reference 26 in main text. Employing the strong base, sodium tert-butoxide provided improved yields and mass balance (Entry 4). The butoxide anion did not show deactivating interactions with the electrochemically generated ferrocenium derivative. NaOH, which is sufficiently insoluble in MeCN, provided an acceptable mass balance without catalyst deactivation (Entry 5). An increased loading of NaOH (Entry 6) in the bulk solution increased the NMR yield of product, presumably by helping to retain formation of the anionic boronate in solution. A variety of proton sources were tested to facilitate proton reduction at the counter electrode. Strong acids are commonly added 4-9 as sacrificial oxidants, but we found that water led to good increases in yield (Entry 7). Running the reaction with constant current led to an identical reaction outcome (Entry 8). Analysis of the applied potential under constant current conditions, was found to remain more consistent over the duration of the reaction, if a proton source (sacrificial oxidant) was added. Acetic acid as sacrificial oxidant was also found to facilitate catalysis and yielded the highest yield and mass balance observed (Entry 10). Finally, running the reaction with a platinum working electrode in place of the RVC one, resulted in a very poor yield (Entry 9). 7. Reactions with stoichiometric oxidant 2a (0.05 mmol) and NaOH (0.075 mmol) were added to a N 2 purged vial and stirred with MeCN (1 M) for 30 seconds. This was then diluted into a solution containing TEMPO (0.1 mmol) and 4,4'difluorobiphenyl (0.01 mmol) in MeCN (0.5 mL). To this was added dropwise a solution containing FcMe 8 BF 4 (0.024 mmol) in MeCN (0.5 mL) and stirred for 10 min before being analysed by 19 F NMR.
A solution containing FcBr 2 BF 4 (0.024 mmol) in THF (0.5 mL) was added dropwise to a solution of 2b (0.02 mmol), 4,4'-difluorobiphenyl (0.005 mmol) and oxo-TEMPO (0.04 mmol) in THF (0.5 mL) with stirring for 10 min before being analysed by 19 F NMR. Oxo-TEMPO was used in this reaction for its enhanced oxidative stability compared to TEMPO, which is necessary with the more highly oxidising ferrocenium oxidant. The benzylic boronic ester (0.1 mmol) and powdered NaOH (7 mg, 0.2 mmol) were added to a Teflon-capped 1.5-dram vial charged with a magnetic stir bar that was then purged with N 2 gas. Dry and degassed 0.2 mL MeCN was added to the vial and the solution was stirred vigorously for 30 s. The solution was then transferred to the anodic compartment of the electrochemical cell. Methyl benzoate was added to the anodic compartment as an internal standard. A CV was obtained of the solution using the glassy carbon electrode, and the peak potential of Fc(Me) 8 was established. Bulk electrolysis was carried out at 0.4 mA. The electrolysis was stopped after 1.5 F/mol had been applied or after the measured potential of the solution began to increase, whichever came first. After the electrolysis, an aliquot was taken up from the anodic compartment under N 2 for subsequent NMR analysis.