Bromide-Mediated Silane Oxidation: A Practical Counter-Electrode Process for Nonaqueous Deep Reductive Electrosynthesis

The counter-electrode process of an organic electrochemical reaction is integral for the success and sustainability of the process. Unlike for oxidation reactions, counter-electrode processes for reduction reactions remain limited, especially for deep reductions that apply very negative potentials. Herein, we report the development of a bromide-mediated silane oxidation counter-electrode process for nonaqueous electrochemical reduction reactions in undivided cells. The system is found to be suitable for replacing either sacrificial anodes or a divided cell in several reported reactions. The conditions are metal-free, use inexpensive reagents and a graphite anode, are scalable, and the byproducts are reductively stable and readily removed. We showcase the translation of a previously reported divided cell reaction to a >100 g scale in continuous flow.


Electrochemical analytical techniques
All cyclic voltammetric (CV) experiments were performed at room temperature using a MultiPalmsens 4. CV experiments were carried out with a working electrode (GC = glassy carbon, Pt, Au, Ni = 1-3 mm diameter), a counter electrode (platinum wire) and a 0.01 M Ag/AgNO3 reference electrode.All working electrodes were polished before each experiment.Before each CV, the solution was stirred for approximately 10 seconds, whilst being degassed by a stream of N2.The CV cell was maintained under an atmosphere of N2 during analysis.After analysis, ferrocene was added to reference the data.

Batch electrochemical setups Electrode materials and suppliers
Zinc plates, magnesium plates and graphite plates (8 x 52.5 x 2 mm) were purchased from IKA.
Pt wire was purchased from Advent research materials: diameter 0.4 mm, 99.99% temper annealed, part number PT5441.Platinum electrodes were made as described in our previous report 1 by wrapping platinum wire around PTFE tubing to create a surface area approximately ~1 cm 2 .The platinum wire was then fed through PTFE tubing by creating a small hole on the side of the tubing.The wire was then spot-welded to a copper wire.
Stainless Steel grating was purchased from GERAWOO via Amazon: 304 Stainless Steel Mesh Sheet,20 Mesh.

Electrode cleaning
Ni foil was discarded after each use.Other electrodes were sonicated in MeOH or MeCN then rinsed with acetone.Zn, Gr, Mg, and Al electrodes were sanded before us.Zn, Mg, and Al were scoured with 20% aq.HCl between polishing if polishing did not afford a smooth finish.

Electrochemical equipment Schlenk tube
Electrode wire were fed through a 3D printed polypropylene stopper, attaching to the electrode material either directly or via a crocodile clip (purchased Amazon, sold for spray-painting or DIY decorative arts).Blue tack was used to ensure airtightness around the wires or crocodile clips.

ElectraSyn vials
Electrasyn vial and lids were purchased from IKA and run on either an ElectraSyn 2.0, a MultiPalmsens 4 or a PLH250 Aim-tti PSU.The electrovortex has a 1.25 mm gap size and a reaction volume of 10.8 mL, using a graphite rod (inner electrode) with a stainless-steel outer (outer electrode) and the electrodes are attached to a 720 w power supply unit (PSU).Rotation of the graphite rotor is supplied by a custom-built motor controller attached to a brushless motor.Both the inlet and outlet pumps are peristaltic pumps using PTFE pump heads with 1/8 th OD PFA tubing.The chiller uses a mixture of IPA and H2O (1:1).

Flow electrolysis setup
Online analysis is obtained using a FT-IR attached with a 6.3 mm AgX DiComp probe inserted into a PEEK IR cell, spectra were collected in 16 scan averages every 15 seconds.

Optimisation table Optimisation of the undivided hydrodefluorination
Preliminary studies were performed using a Ni gauze cathode, in which many different co-reductants, solvents, and electrochemical parameters were tested.However, these studies did not lead to easily interpreted results, possibly due to the difficulty in cleaning of the gauze.Ni foil gave results that were more easily interpreted and therefore we show these.
Table S1: Optimisation of the undivided hydrodefluorination of trifluoromethylarenes. 19b] Pt wire as anode [a] Reaction run at 0.2 M, as opposed to 0.1 M and only 3 eq.TMSCl.

Impact of the charge on the product distribution and selectivity:
The entry 15 of Table S1 was run for two different amounts of charge in order to assess the selectivity against charge passed (Table S2).Increasing the charge from 3F to 4F led to a 29% point increase in conversion leads only to 13% point increase in "CF2" species.This can be explained by an increase of overreduction of the "CF2" species to "CF1" species by 13% point.As a result, the choice of the amount charge to pass is a balancing act between increasing conversion and limiting overreduction.
Table S2: Impact of the charge in the double concentration experiment 19 F NMR yields relative to internal C6F6 standard."CF2" refers to all monodefluorinated species detected, "CF1" refers to all didefluorinated species detected.The selectivity at 3F is 14:1, while at 4F the selectivity has been reduced to 5:1 despite a higher yield.For comparison, the reported selectivity in the original paper was 25:1. [2]As a result, the undivided system may have a modest but deleterious impact on selectivity.

Entry Charge Conversion
Direct trapping of the by-products of the TES oxidation by the substrate is likely not a significant factor as this accounts for less than 1% of the product distribution.,which must be due to the low amount of [TES + ] species compared to TMSCl, and the bulkiness of it.
In any case, the fact that Gen2 conditions offer higher selectivity and a yield matching the original divided cell conditions [2] seems to indicate that the selectivity drop is not intrinsic to the undivided system.

Optimisation of the "Generation 2 conditions"
Procedure for the optimisation of "Generation 2 conditions" Under air, a 40 mL dried Schlenk tube equipped with a PTFE stir bar was charged with Bu4NBr (241 mg, 0.75 mmol, 1.5 eq.) and other electrolytes (if indicated).MeCN (non-anhydrous grade, 2.5 ml or 5.0 mL) was added, followed by the silanes, additives and 4-fluorotrifluoromethylbenzene 1a (0.500 mmol, 63.0 μL, 1.00 eq.)(if liquid) via micropipette.The tube was fitted with a polypropylene stopper with two crocodile clips holding a Stainless Steel grating (3x2 cm) cathode and a graphite plate anode.
Electrolysis was performed for 25 or 50 mA for the indicating charge with a stir rate of 200 rpm.After electrolysis, the electrodes were rinsed using MeCN, TIPSCl (109 μL, 0.500 mmol, 1.00 eq.) was added as a fluoride scavenger, and C6F6 (20 μL) was added for 19 F NMR yield assay.
b] Calculated by measuring the quantity of TIPSF detected (quant. 19d] Formation of adducts as byproducts.
Yield 2a b] Calculated by measuring the quantity of TIPSF detected (quant. 19d] Formation of adducts as byproducts.
Under a positive pressure of N2, the suba-seal was swapped for polypropylene stopper cap with two corocodile clips holding two graphite plate electropdes.
Electrolysis was performed for 30 mA for 6400 s (2F) with a stirring rate of 500 rpm.After electrolysis C6F6 (20.0 μL) was added for 18 F NMR yield assay.
a] ] Determined by 19 F NMR using C6F6 as an internal standard Entry Base Conversion [%] [a] Yield 9a [%] [a] Mass balance [%] None of the bases tested significantly improved the reaction profile (Table S5).LiOAc (entry 4) lead to a suppression of the reaction, possibly due competitive Li + reduction.BuN4OAc (entry 5) lead to a complex mixture, which may be attributed to the acetate nucleophilicity.

Impact of the amount of Bu4NBr of the product distribution and Faradaic efficiency in the hydrodefluorination
Under a N2 flow, a 10 mL wide neck oven-dried Schlenk tube equipped with a PTFE stir bar was charged with Bu4NPF6 (580 mg, 1.5 mmol, 3.0 eq.) and Bu4NBr .Anhydrous, degassed MeCN (5 mL) was inserted via syringe, and Et3SiH (0.24 mL, 1.5 mmol, 3 eq.)and the trifluoromethylarene substrate (0.500 mmol, 1.00 eq., 63 μL) were added via micropipette over an N2 flow.TMSCl (0.38 mL, 3.0 mmol, 6.0 eq.) was then added via needle.The septum was swapped for a polypropylene stopper with two crocodile clips holding a Ni foil (3x2 cm) cathode and a graphite plate anode.
Electrolysis was performed for 20 mA for 7200 s (3F) (unless otherwise specified) with a stir rate of 200 rpm.After electrolysis, the electrodes were rinsed using MeCN and C6F6 (20 μL) was added for NMR yield assay.Increasing the amount of Bu4NBr increases the Faradaic efficiency of the reaction, with a maximum at 1.5 eq.This supports the hypothesis of 1 eq. of bromide being able to mediate the generation of 2 eq. of electrons at the anode.

Formation of Et3SiF with Bu4NF
In an NMR tube was inserted MeCN (1 mL), followed by C6F6 (1.0 eq.) and either Et3SiH (0.1 mmol, 1.0 eq.) or Et3SiCl (1.0 mmol, 1.0 eq.) by micropipette.(If Bu4NBBr3 was added, it was added as a solid (48 mg, 1.0 eq.)).The tube was agitated and left standing for 30 min.Subsequently, Bu4NF (1 M in THF, 0.1 mL, 1.0 eq.) was added via syringe.The tube was agitated and analysed by quantitative 19 F NMR after 2 h.Analysis of the spectra (Figure S3) shows clear formation of the TESF adduct at 175.7 ppm (with the characteristic multiplicity) in the reaction of TBAF and Et3SiCl (bottom spectra,85 %) as well as in the reaction of TBAF and Et3SiH and TBABr3 (middle spectra, 75 %).In contrast, the reaction without TBABr3 (top spectra) show no Et3SiF signal.

Cyclic voltammetry experiments
Each CV was run in 2.5 mL anhydrous MeCN with 0.6 M Bu4PF6, using a GC disk electrode at a scan rate of 100 mV/s, unless otherwise specified (see general considerations for full details).As shown in Figure S4, the addition of water has a drastic impact on the current response of bromide oxidation above 0.5 V, but only in the presence of silane.This observation supports the hypothesis of a more efficient regeneration of bromide due to hydrolysis of the formed Et3SiBr in the presence of water.This is consistent with the need for only a catalytic amount of Bu4NBr in aqueous media (see the pinacol coupling results).
As shown in Figure S5, increasing the amount of silane has a small but noticeable effect on the current response of the first feature of the oxidation of bromide to tribromide.This supports the hypothesis of regeneration of bromide from "oxidised bromide" by the silane, albeit slowly at CV time scales.
Figure S6 shows that once normalised according to the Randles-Sevcik equation (by dividing by the square root of the scan rate), the current responses at different scan rates are not equal.The lower scan rates show a greater normalised current response than the faster ones.This is consistent with the hypothesis of the regeneration of bromide on the CV timescale: at slower scan rates more bromide is regenerated over the course of the experiment leading to higher current responses.
As the current response for the second oxidation peak is much greater in the presence of silane, Figure S4, the regeneration of bromide from bromine is significantly faster.CV experiment showed that TMSCl only displays a reduction feature at very negative potentials (peak at −3.3 V -−3.4 V vs Fc/Fc + , Figure S8), very close to the edge of the solvent windows.
Addition of TMSBr increased the current response in that range, indicating a very close reduction potential of these two species.Addition of water reduced the intensity of this peak, indicating that this feature is intrinsic to TMSBr and TMSCl, and is not due to hydrolysis product reduction.
As a result, generation of TESBr should not affect the reduction process.

Oxidative stability to Br3 -with or without TES:
To assess the tolerance of functional groups to the new process and in particular oxidatively sensitive groups, we conducted a set of experiments in which different compounds were subjected to Bu4NBr3 in the presence and absence of Et3SiH (TES).

General procedure:
Two solutions of each substrate (see list in the table below) were prepared: the substrate 0.50 mmol (1 eq., 1.0 M) and C6F6 (9.6 μL, 0.083 mmol, 0.17 eq., internal standard) were dissolved in a mixture of MeCN (0.50 mL) and THF (0.20 mL).For each substrate, Et3SiH (160 μL, 1.0 mmol, 1.0 eq.) was added to one of the solutions.For solubility reasons, 0.50 mL of THF was added to each solution of PPh3, and 0.20 mL of water was added to each solution of hydroquinone.
A solution of Bu4NBr3 (9.6 g, 20 mmol) and CH2Br2 (352 μL, 5.00 mmol, as 1 H NMR standard) in MeCN (15.0 mL) was prepared.After sonication, the total volume was 22.3 mL.0.75 mL of this solution was added over 1 min to each of the substrate solutions via syringe (corresponding to 1.3 eq.Bu4NBr3 and 0.34 eq.CH2Br2).The resulting solutions were shaken and sonicated for 5 minutes, then left to stand for 18 h.100 μL of each solution was transferred to an NMR tube (if solid were present, only the supernatant was pipetted) and diluted with 0.5 mL of CDCl3 before being analysed by 1 H NMR and 19 F NMR spectrometry.Quant.

%
For entries 1-4 (electron-rich rings, alkenes and phosphine) TES protected the substrate against bromination.For entry 5, the thiophenyl ether was not affected under either conditions.For entry 6, TES partly protected the silyl enol ether against bromination.For entry 7, reduction of the aldehyde by TES was observed.S8, Faradaic efficiency of the hydrodefluorination is lower in the silane/Br system.However, for the acylation, disylation and pinacol, the FE is better for the silane/Br system
Electrolysis was performed for 20 mA for 7200 s (3F) (unless otherwise specified) with a stir rate of 50 rpm.After electrolysis, the electrodes were rinsed using MeCN and C6F6 (20 μL) was added for NMR yield assay.Compound 2c was further purified while 2a, 2b, 2e, 2f, 3a were not.

4-Fluoro-(fluoromethyl)benzene, 3a
3a was synthesised from 4-Fluoro-(difluoromethyl)benzene (2a) using General Procedure 1. Bu4NF (1 M in THF, 5.0 mL, 5.0 mmol, 10 eq.) was added to the crude reaction mixture after electrolysis.Because of its volatility, it was not isolated: the 19 F NMR yield was determined to be 48%.The following diagnostic NMR signals were observed in the crude reaction mixture: (Difluoromethyl)benzene, 2b 2b was synthesised from benzotrifluoride using General Procedure 1 with 4F.Because of its volatility, it was not isolated: the 19 F NMR yield was determined to be 73%.The following diagnostic NMR signals were observed in the crude reaction mixture: Ethyl 4-(difluoromethyl)benzoate, 2c 2c was synthesised from 4-(trifluoromethyl)benzoate using General Procedure 1.After concentration under reduced pressure, Bu4NF (1M in THF, 1 mL, 1 mmol, 2 eq.) was added to the crude reaction mixture.The mixture was agitated for a few minutes, loaded unto silica by concentrating it under reduced pressure and purified by FCC to yield a colourless oil (53 mg, 53% (9:1 2c: over-reduced 3c), which represents a yield of 47% of 2c when correcting for 10% contamination with 3c, as also observed in the divided cell conditions).
Electrolysis was performed for 20 mA for 9600 s (4F) with a stir rate of 50 rpm.After electrolysis, the electrodes were rinsed using MeCN and C6F6 (20 μL) was added for NMR yield assay ( 19 F NMR = 64%).Evaporation of the volatile under reduced pressure was followed by FCC (5-15% EtOAc in pentane) to afford a white solid (55%).
Electrolysis was performed for 20 mA for 6000 s (2.5F) with a stir rate of 200 rpm.After electrolysis, the electrodes were rinsed using MeCN.The reaction mixture was then concentrated under reduced pressure.The crude was handled differently depending on the desired product: For 7c, the solids were triturated and filtered off with diethyl ether (2 x 25 mL) then EtOAc (25 mL).
The residue was concentrated and dissolved in a minimum amount of EtOAc and purified by FCC (pure pentane, liquid injection in EtOAc).
For 7a, 7b, and 7d, the solids were dissolved in THF (50 mL) and cooled to 0°C.Bu4NF (2.0 mL, 1 M in THF, 2.0 mmol, 4.0 eq.) was added and the mixture was stirred for 1.5 h.Loading onto silica by concentration under reduced pressure and FCC afforded the desired compounds.

Metal-free disylation of styrenes in a Schlenk tube (General procedure 3)
A 40 mL oven-dried Schlenk tube was taken into a nitrogen-filled glovebox and charged with Bu4NPF6 (700 mg, 1.8 mmol, 1.8 eq.) and Bu4NBr (970 mg, 3.0 mmol, 3.0 eq.).Anhydrous THF (9 mL), the substrate (1.00 mmol, 1.00 eq.), Et3SiH (0.48 mL, 3.0 mmol, 3.0 eq.) and the trialkylsilane (3.0 mmol, 3.0 eq.) were added.The tube was closed with a polypropylene stopper with two crocodile clips holding graphite plates.Electrolysis was performed for 30 mA for 8000 s (2.5F).After electrolysis, the stopper was swapped for a septum and the reaction tube was taken out of the glovebox and connected to a N2 Schlenk line.Ethanolamine (0.30 mL, 5.0 mmol, 5.0 eq.) was added via syringe and the mixture was stirred under N2 for an additional hour.
The reaction mixture was then filtered through a 25 mL silica plug and eluted with additional solvent, concentrated under reduced pressure, loaded unto silica gel, and purified by FCC to afford the title compounds.Following General Procedure 3, after a plug with 5% Et2O/pentane (2 x 100 mL), concentration under reduced pressure and FCC (pure pentane), a clear oil was obtained (70%).

Metal-free disilylation of 4-fluorostyrene using the ElectraSyn
An oven-dried 10 mL Electrasyn vial was taken into a nitrogen-filled glovebox and charged with Bu4NClO4 (270 mg, 1.1 mmol, 1.1 eq.) and Bu4NBr (970 mg, 3.0 mmol, 3.0 eq.).The thread of the vial was wrapped in PTFE tape, and the vial was sealed using an ElectraSyn lid fitted with two graphite plates.The vial was taken out and placed under a positive pressure of N2 via a needle connected to a Schlenk line.
Electrolysis was performed for 30 mA for 6240 s (2.5F).After electrolysis, ethanolamine (0.30 mL, 5.0 mmol, 5.0 eq.) was added via syringe and the mixture was stirred under N2 for an additional hour.Afterwards, the electrodes were rinsed using EtOAc and C6F6 (20.0 μL) was added for NMR yield assay ( 19 F NMR = 78%).The reaction mixture was then filtered through a 25 mL silica plug using pentane (50 mL), filtered over cotton, concentrated under reduced pressure, loaded unto silica gel, and purified by FCC (pure pentane) to afford 9a as clear oil (72%).
Electrolysis was performed for 20 mA for 6240 s (1.3F).After electrolysis, the electrodes were rinsed using MeCN.The reaction mixture was then concentrated under reduced pressure, loaded onto silica gel and purified by FCC to afford the title compounds.

Cathode
Anode Stir rate Current Duration Charge graphite plate graphite plate 200 rpm 20 mA 6240s 1.3F
Major: Data in agreement with the literature, assuming that the major isomer is dl and the minor isomer is meso. 6
Major:   Data in agreement with the literature, assuming that the major isomer is dl and the minor isomer is meso. 6
Major: Data in agreement with the literature, assuming that the major isomer is dl and the minor isomer is meso. 6
Major: Data in agreement with the literature, assuming that the major isomer is dl and the minor isomer is meso. 6
An oven-dried wide neck 10 mL Schlenk tube was taken into a nitrogen-filled glovebox and charged with Bu4NClO4 (200 mg, 0.6 mmol, 3.0 eq.), Bu4NBr (260 mg, 0.8 mmol, 4.0 eq.).The tube was sealed, taken out of the glovebox, and placed under a positive pressure via a Schlenk line.
Electrolysis was performed for 30 mA for 2560 s (4F) (Et3SiH or 1,1,3,3-tetramethlydisiloxane) or 5 mA for 11520 s (3F) (diphenylsilane, triethoxysilane or dimethylphenylsilane).After electrolysis, the electrodes were rinsed using MeCN and the reaction mixture was concentrated under reduced pressure.After the addition of a known amount of CH2Br2, 1 HNMR analysis showed that, despite the conversion of diphenylchloromethane (<50% remaining), no significant amount of product was formed.A complex mixture was observed with multiple signals in the aromatic area (δ 6.5 -7.5 ppm).Diphenylmethane (characteristic singlet at δ 3.9) was observed but was not the main species.
It is suspected that the silane may interact unfavourably with the formed diphenylmethyl radical.
Electrolysis was performed for 10 mA or 30 mA for 3-6F. 1 H NMR of the reaction mixture did not detect any product, with mostly starting material remaining.

Flow electrochemical experiments using the Electrovortex Reactor General procedure 5 for running the ElectroVortex reactor
The chiller is turned on and set to 10 °C (IPA/MeOH, 1:1).A flask containing MeCN is attached to the inlet line and both peristaltic pumps are set to 5 mlmin -1 , the rotor is then turned to the desired rotation and the reactor is allowed to flush for 10 minutes.A flask containing the reaction solution is connected to inlet line and the inlet pump is set to the desired flow rate.The solution is then left pumping until a steady state concentration is achieved within the system, this is monitored by inline FTIR.The power supply is then turned on and set to the desired current using the constant current mode.A sample is then collected for analysis after steady state.has been observed, this takes approximately 5 reactor volumes (54 mL) of solution.After all experimentation is complete the inlet pump is opened and the reactor is flushed with solvent (MeCN) and the power supply unit is turned off.The ElectroVortex is then detached and dismantled before being cleaned with MeOH and a brush, both the inner and outer electrodes are cleaned.The vortex reactor is then reassembled and reconnected to the system and the reactor is ready for operation.

Optimisation of the electrochemical synthesis of (difluoromethyl)benzene, "Generation 1 conditions"
Under air, a solution of trifluorotoluene (45.0 mmol, 5.50 mL) in MeCN (150 mL) was prepared and triethylsilane (135 mmol, 21.5 mL), Bu4NBr (67.5 mmol, 21.8 g), TMSCl (see table S9) and Bu4NPF6 (30.0 mmol, 11.6 g) were added.The reaction solution is then connected to inlet line 2 and the desired flow rates and rotor speed are set, general procedure 5 is then followed.Each run allows for 2 reaction conditions to be investigated, with each sample being analysed using quantitative 19 F NMR with fluorobenzene as an internal standard.

Multigram synthesis of (difluoromethyl)benzene in continuous flow, "Generation 1 conditions"
Under air, a solution of trifluorotoluene (0.36 mol, 44.2 mL) in MeCN (1200 mL) was prepared and triethylsilane (1.08 mol, 171.6 mL), Bu4NBr (0.54 mol, 174.1 g), TMSCl (1.08 mol, 135.6 mL) and Bu4NPF6 (0.24 mol, 92.9 g) were added to the solution sequentially.The flow rate of the inlet pump was the set to 5 mlmin -1 and the rotor turned to 2000 rpm, the reaction solution was then added in inlet line 2 and general procedure 5 was followed.The PSU current was set to 7.05 A (3.8F) and the reaction left to achieve S.S., the reaction was then left collecting in a glass bottle for 4 hours.Hourly timepoint were collected and the conversion and yield were measured by quantitative 19 F NMR.

S41
Late-stage functionalisation of fluoxetine (Prozac) using the electrovortex reactor, "Generation 1 conditions" Under air, a solution of fluoxetine (30.0 mmol, 9.30 g) in MeCN (100 mL) was prepared and triethylsilane (90.0 mmol, 14.3 mL), Bu4NBr (45.0 mmol, 14.50 g), TMSCl (120 mmol, 15.1 mL) and Bu4NPF6 (20.0 mmol, 7.70 g) were added to the solution sequentially.The flow rate of the inlet pump was the set to 2 mlmin -1 and the rotor turned to 2000 rpm, the reaction solution was then added in inlet line 2 and general procedure 5 was followed.2 reaction conditions were investigated (Table S11), with each sample being analysed using quantitative 19 F NMR with fluorobenzene as an internal standard.Electrochemical synthesis of (difluoromethyl)benzene, using "Generation 2 conditions" Under air, in a 100 mL volumetric flask trifluorotoluene (20 mmol, 2.45 mL) was dissolved in MeCN (50 mL).Followed by the stepwise addition of Bu4NBr (30 mmol, 9.67 g, 1.5 eq.), tetramethyldisiloxane (60 mmol, 10.60 mL, 3.0 eq.) and hexamethyldisiloxane (120 mmol, 25.50 mL, 6 eq.).The volumetric flask was then filled up to 100 mL using MeCN and poured into a round bottom flask.As the starting solution is biphasic a magnetic stirrer and plate were added to the reaction set-up and the solution was stirred throughout the reaction.General procedure 5 was then followed, and the conditions were altered as desired (Table S12).Quantitative 19 F NMR analysis was used for the calculations of yields from a S.S. sample with fluorobenzene as an internal standard.

S43
PAT analysis in continuous flow

IR Data
An inline FTIR was used to collect all data using a probe connected to a probe holder, with the reaction mixture being flowed directly through it.Spectra were obtained at 4 cm -1 resolution and with 16 scans.Spectra were first averaged together (4), a baseline was then applied to the obtained spectra before normalising to the MeCN peak and subtracting the solvent spectrum.An MCR model was constructed to track the concentration of reagents and product as the reaction progressed.This was done using PLS toolbox from Eigenvector and involved using calibration solutions of each reagent in a range of concentrations (0 to 1 M).

Figure S1 :
Figure S1: Example electrochemical setup with crocodile clips/polypropylene lid, Ni foil and Gr.Plate.

Figure S4 :
Figure S4: Impact of water and Et3SiH on the current response of Bu4NBr oxidation.TES = Et3SiH; TBAB = Bu4NBr.10 mg of each analyte was used.

Figure S5 :
Figure S5: Impact of the amount of Et3SiH on the current response of Bu4NBr oxidation.TES = Et3SiH; TBAB = Bu4NBr.10 mg of Bu4NBr and water used.

Figure S6 :
Figure S6: Impact of the scan rate on the normalised current response of Bu4NBr oxidation.The two 0.1 V/s scans are repeats to ensure there was no significant change or drifting due to concentration change or solvent evaporation; the blue trace was acquired before all the other scan rates, and the orange afterwards.

Figure S8 :
Figure S8: CV experiment of TMSCl, TMSBr, and water.Each CV was run in 3mL anhydrous MeCN with 0.4 M Bu4PF6, using a Ni disk electrode at a scan rate of 100 mV/s.Only forward scan shown.

Figure S10 :
Figure S10: FTIR spectra of reagents and product in the reaction at approximately operating concentrations.

Figure S11 :
Figure S11: Left: MCR spectral loads showing starting reagents (orange) and product (blue).Right: MCR scores of each spectral load showing the reagents flow in and then converted to products when the current is turned on.This provides a qualitative indication how the reaction progresses.

Figure S12 :
Figure S12: FTIR spectra of the hydrodefluorination of Prozac in continuous flow.Blue is the starting materials flowing in, and green to yellow is the consumption of Prozac as the current is applied.

Table S6 :
Impact of the amount of Bu4NBr on the product distribution and the Faradaic efficiency.

Table S7 :
Stability of various electron-rich or easily oxidisable substrates to Br3 -in the presence or absence of Et3SiH.Recovery was determined by19F NMR or 1 H NMR.

Table S9 :
Continuous flow optimisation of the synthesis of (difluoromethyl)benzene.All yields calculated versus fluorobenzene.*Reaction performed at 0.1 M.

Table S10 :
Consumption and yields over a 4h continuous flow electrolysis.All yields calculated versus fluorobenzene Figure S 9: Product distribution over a 4h continuous flow electrolysis, determined by quantitative 19 F NMR.

Table S11 :
Continuous flow optimisation of the synthesis of the defluorination of fluoxetine (Prozac).All yields quoted versus fluorobenzene.

Table S12 :
Brief investigation of the generation 2 conditions in flow.All yields quoted versus fluorobenzene.