Taking electrodecarboxylative etherification beyond Hofer–Moest using a radical C–O coupling strategy

Established electrodecarboxylative etherification protocols are based on Hofer-Moest-type reaction pathways. An oxidative decarboxylation gives rise to radicals, which are further oxidised to carbocations. This is possible only for benzylic or otherwise stabilised substrates. Here, we report the electrodecarboxylative radical-radical coupling of lithium alkylcarboxylates with 1-hydroxybenzotriazole at platinum electrodes in methanol/pyridine to afford alkyl benzotriazole ethers. The substrate scope of this electrochemical radical coupling extends to primary and secondary alkylcarboxylates. The benzotriazole products easily undergo reductive cleavage to the alcohols. They can also serve as synthetic hubs to access a wide variety of functional groups. This reaction prototype demonstrates that electrodecarboxylative C–O bond formation can be taken beyond the intrinsic substrate limitations of Hofer-Moest mechanisms.


Supplementary Note 1
General remarks: Solvents were purchased (puriss p.A.) from commercial suppliers and purified by standard procedures before use. 1 Commercial substrates were purchased from Merck ® (Sigma-Aldrich ® ), Acros ® , Alfa Aesar ® and Fluorochem ® , and were used without further purification.
Mass spectral data were acquired on a GC-MS Agilent ® 5977B MSD™. High-resolution mass spectrometry (HRMS) analyzes were obtained on a Waters ® GCT Premier CAB163™ with a TOF mass analyzer or on a Thermo Fischer Scientific ® LTQ Orbitrap XL™ with positive ion mode.
Melting points were measured on a Mettler Toledo ® One Click™-MP70.
Infrared experiments were carried out on a Bruker ® Vertex 70 Spectrometer™ with Universal ATR sampling Accessory.  In methanol: A 20 mL vessel equipped with two Pt-electrodes (2.0 × 1.0 cm), was charged with a solution of hydrocinnamic acid 1a (152 mg, 1.00 mmol) and NaOH (72 mg, 1.8 mmol, 1.8 eq.) in MeOH (12 mL). The reaction mixture was electrolyzed at a current of 500 mA for 30 min at room temperature. GC and GC-MS analysis of the resulting crude showed that styrene and 1,4-diphenylbutane (Kolbe dimer) were the major products, together with small amounts of non-selective derivatives of Hofer-Moest. In water: A 20 mL vessel equipped with two Pt-electrodes (2.0 × 1.0 cm), was charged with a solution of hydrocinnamic acid 1a (152 mg, 1.00 mmol) and NaOH (72 mg, 1.8 mmol, 1.8 eq.) in H2O (12 mL). The reaction mixture was electrolyzed at a current of 500 mA for 30 min at room temperature. 19 F NMRanalysis of the resulting crude showed traces of the expected trifluoroethanol (<2%), along with a mixture of products derived from elimination processes such as ethylbenzene (7a) and/or styrene (8a). In methanol: A 20 mL vessel equipped with two Pt-electrodes (2.0 × 1.0 cm), was charged with a solution of TFPA 1b (91 µL, 1.0 mmol) and NaOH (72 mg, 1.8 mmol, 1.8 eq.) in MeOH (12 mL). The reaction mixture was electrolyzed at a current of 500 mA for 30 min at room temperature. 19 F NMRanalysis of the resulting crude showed that 1,1,1,4,4,4-hexafluorobutane (Kolbe dimer) was the major product (42%) along with 10% of methyl ether 5b 2 and traces of 2,2,2-trifluoroethanol 5b 4 .

Supplementary
In water: A 20 mL vessel equipped with two Pt-electrodes (2.0 × 1.0 cm), was charged with a solution of TFPA 1b (91 µL, 1.0 mmol) and NaOH (72 mg, 1.8 mmol, 1.8 eq.) in H2O (12 mL). The reaction mixture was electrolyzed at a current of 500 mA for 30 min at room temperature. 19 F NMR-analysis of the resulting crude showed no trace of the expected trifluoroethanol, but a mixture of products derived from elimination or rearrangement processes.  (447 mg, 6.88 mmol, 1.71 eq.) and KNO3 (403 mg, 3.94 mmol, 0.996 eq.) in H2O (12 mL). A total charge of 27 C was passed through the cell using a current density of 900 mA/cm 2 (3.6 A) for 7.5 min at room temperature. The crude was extracted with Et2O (2 × 5.0 mL) and n-pentane (2 × 5.0 mL), the organic phases were collected, dried over Mg2SO4 and analysed by GC and GC-MS. A mixture of Kolbe and Hofer-Moest products was produced, with no trace of the corresponding nitrate ester. Following a procedure described by Fichter et al., 2 a 20 mL vessel equipped with two Pt-electrodes (2.0 × 1.0 cm), was charged with a solution of TFPA 1b (360 µL, 3.96 mmol), KOH (447 mg, 6.88 mmol, 1.71 eq.) and KNO3 (403 mg, 3.94 mmol, 0.996 eq.) in H2O (12 mL). A total charge of 27 C was passed through the cell using a current density of 900 mA/cm 2 (3.6 A) for 7.5 min at room temperature. The crude was extracted with Et2O (2 × 5.0 mL) and n-pentane (2 × 5.0 mL), the organic phases were collected, dried over Mg2SO4 and analysed by GC and GC-MS. 19 F NMR-analysis of the resulting crude showed no trace of the corresponding nitrate ester but a mixture of by-products that were presumably generated by elimination, over-oxidation, or rearrangement processes. Following a procedure described by Fichter et al., 2 a 20.0 mL vessel equipped with two Pt-electrodes (2.0 × 1.0 cm), was charged with a solution of hexanoic acid 1f (501 mg, 3.96 mmol), KOH (447 mg, 6.88 mmol, 1.71 eq.) and KNO3 (403 mg, 3.94 mmol, 0.996 eq.) in H2O (12 mL). A total charge of 27 C was passed through the cell using a current density of 900 mA/cm 2 (3.6 A) for 7.5 min at room temperature. The crude was extracted with Et2O (2 × 5.0 mL) and n-pentane (2 × 5.0 mL), the organic phases were collected, dried over Mg2SO4 and analysed by GC and GC-MS. As documented by Fichter et al., 8 although small amounts of pentyl nitrate were detected (~ 6%), the main products were those corresponding to Kolbe electrolysis in an aqueous medium.  Reaction conditions: 1a (mmol), 2a (eq.), base (eq.), MeOH/Py (ratio, 12 mL), current (mA), t (min), r.t.; Yields were determined by GC-analysis using n-undecane as internal standard. b Reaction conditions "B". c Reaction conditions "C".   Figure 13. Unsuccessful substrates in radical electrodecarboxylative etherification.

)-O cross-coupling
Primary carboxylic acids Note: due to its explosive nature, HOBt was used exclusively in the form of its stable hydrate.

Cyclic voltammetric studies
Cyclovoltammetry measurements were carried out using a Pt-disk working electrode (2 mm diameter, CH Instruments) and a Pt-wire counter electrode. A Haber-Luggin dual-reference electrode system as described by Speiser et al. was used. 31 This system consists of a Ag-wire, which is submerged in a solution of AgPF6 (0.01 M) and Bu4NPF6 (0.1 M) in MeCN. The silver wire is connected through a 0.01 µF capacitor to a Pt-wire, whose tip is placed in proximity to the working electrode. The reference solution is connected to the solution of analyte via a double junction and a Haber-Luggin capillary. The potential scale was calibrated against external ferrocene (Fc). The samples were measured with Bu4NPF6 (0.1 M) as supporting electrolyte. MeCN was dried by passing through a column of activated alumina using a MBRAUN Solvent Purification System. MeOH was dried by stirring with powdered activated molecular sieves (3Å) for 10 days and distilling the resulting mixture. All solvents were degassed by bubbling dry argon into the liquid for at least 15 minutes. The potentiostat was an EmStat³ from PalmSens BV. Unless stated otherwise, a scan rate of 100 mV/s was used, and for each sample, ten potential sweeps between −0.5 and +2.5 V vs. Ag/Ag + were performed; the first (red trace) and the tenth (blue trace) sweep are displayed. No significant changes are observed over the course of 10 cyclic voltammetry cycles, which demonstrates that the electrode surface is stable at these potentials.

 Measurements in MeOH:
If pure MeOH is used as the solvent, strong oxidation of the solvent is observed, which hinders the analysis of the oxidation of the substrates. The oxidation of phenylpropionic acid 1a is not visible due to overlap with the oxidation of MeOH. On the other hand, HOBt·H2O 2a clearly shows an oxidation peak (Epa=1.30 V). Interestingly, a small cathodic peak (Epc=0.12 V) is observed during the return sweep, possibly resulting from reduction of BtO • radicals.

 Measurements of hydrocinnamic acid 1a in MeOH/Py (4:1):
Addition of pyridine has several striking effects on the cyclic voltammograms. In the blank measurement, up to a potential of around 2 V, only relatively small currents are observed compared to pure MeOH. This demonstrates that, while pyridine slowly gets oxidized itself, it suppresses the oxidation of MeOH. The addition of Li2CO3 shifts the oxidation peak of the acid to lower potential (from 1.49 to 1.20 V). These measurements were performed using Bu4NBF4 as supporting electrolyte and AgBF4 as reference, instead of their respective PF6-salts. For the scans with varying scan rate, the potential sweep was performed between -0.5 and +1.0 V vs. Ag/Ag + , and only the first scans are displayed.

NMR Spectra
The chemical shifts of the used solvent signals observed for 1 H NMR, 13 C NMR and 19 F NMR spectra are listed in the following chart. The multiplicity is shown as "s" for a singlet, "d" for a doublet, etc.  13 C NMR (CDCl3, 75 MHz) 1 H NMR (CDCl3, 300 MHz) 13 C NMR (CDCl3, 75 MHz)