of Selective C-H bond electro-oxidation of benzylic acetates and alcohols to benzaldehydes

A chemical oxidant-free and mediator-free, direct electro-oxidation of both benzylic alcohols and benzylic esters are reported. The scope of the reaction is explored as a function of both steric and electronic effects. Expansion of the scope to non-benzylic and heteroaryl substrates are investigated. Functionalisation of esters and alcohols selectively to the aldehyde oxidation level using a traceless electron approach is reported.


Introduction
The controlled and chemo-selective oxidation of primary alcohols to aldehydes and secondary alcohols to the corresponding ketones are fundamental reactions in organic synthesis. 1 In turn, these aldehydes and ketones serve as precursors for a variety of complexity generating reactions. 2 However, the controlled oxidation of primary alcohols to aldehydes can be problematic due to over-oxidation to the carboxylic acid oxidation state. 3 There are a range of versatile chemical oxidants available to the academic and industrial chemist that enable this transformation to be performed on demand. 4 However, the majority of these oxidation reactions are stoichiometric in nature and therefore suffer from the generation of quantities of chemical waste. 5 Recently, the field of electrosynthesis has undergone a renaissance 6 and has found application in a variety of organic synthetic transformations, such as: C-H bond activation, 7 total synthesis, 8 Diels-Alder reaction 9 amongst others. 10 A fundamental advantage of the electrosynthesis approach is the replacement of the need to use stoichiometric oxidants and instead the oxidation reaction is performed on the electrode surface via quantum mechanical tunelling 11 or through a mediator in solution. 12 To address the challenge of identifying a cleaner oxidation, we explored the use of electrosynthesis to replace the need for both a chemical oxidant and mediator in these REDOX transformations. The simultaneous removal of oxidant and mediator would minimise chemical waste associated with the reaction. We have recently investigated the Shono-type oxidation of C-H bonds adjacent to a tertiary amide 13 and our initial foray into this area began with attempting to expand the scope of the amide oxidation to esters (Figure 1). It is known that the Shono electro-oxidation of amides proceeds through an N-acyl iminium species (A). 14 It was therefore postulated that a transient O-acyloxonium species (B) 15 could form under similar electro-oxidative conditions in esters bearing an ɑmethylene or methide group. This unstable O-acyloxonium species (B) would in turn react further with adventious water to form the aldehyde product and a carboxylic acid as a by-product via intermediate (C). Alternatively, an ester and aldehyde product could potentially form when conducted in an alcohol. To the best of our knowledge there are only limited reports of electro-oxidative cleavage of an ester group.

Results and discussion
To probe whether this methodological leap was indeed feasible, a collection of benzylic acetates were prepared including the parent benzyl acetate 1a, a mild electron donating example 1b and a mild electron withdrawing example 1c (structures shown in Table 2). Cyclic voltammetry was recorded for 1a ( Figure 2).
Using a sensitive measurement of 10 mVs -1 scan rate it was observed in both electrolyte systems (LiClO 4 and Bu 4 NClO 4 ) 17 that oxidation waves for all three substrates were observed with +1.0 V (1a), +1.2 V (1b) and +0.9 V (1c) oxidation potentiasl (relative to Ag/AgCl) and a slightly improved peak current (Ip) measurement in Bu 4 NClO 4 . On the basis of this positive oxidation result, screening of potential conditions to enable viable electro-oxidation of model benzylic acetate 1a was attempted (Table 1).
In the first instance, potentiostatic conditions were screened with 1a (entries 1-8) using both Bu 4 NClO 4 and LiClO 4 as the electrolyte. The applied voltage was varied around the observed oxidation potential for 1a (cyclic voltammetry measurement, +1.0 V) with up to an extra 300 mV applied to compensate for expected iR drop across the electrode surface. 18 In all cases, the reaction was performed until Fmol -1 equivalent to 4 electrons per mole of substrate was passed or starting material consumption was observed. Near the oxidation potential of 1a trace conversion to the desired aldehyde 2a was observed in both electrolyte systems (entries 3-4 and 6-7, respectively). However, at higher applied voltages degradation products were observed (entries 5 and 8, Bu4NClO4 MeCN/MeOH -20 27  Please do not adjust margins Please do not adjust margins approach was respectively), coupled with excessive time required for sufficient charge to be passed (24 h to 5 days), a controlled voltage ruled out in this system early on. Switching to a galvanostatic approach, using the same electrolyte systems and reticulated vitreous carbon (RVC) electrodes produced more promising results (entries 9-15).
In particular, it was observed that a single solvent system gave higher yields than the previously optimized solvent system for the Shono oxidation, 13 acetonitrile-methanol (entries 9 vs 10 and 12 vs 13, respectively). Lowering the current density across the electrode surface from an applied 20 mA to 10 mA, led to a doubling in reaction time (approx. 5-6 h) and a concomitant improvement in reaction yield in both electrolyte systems (entries 11 vs 10 and 14 vs 13, respectively). Lowering the current rate further led to modest improvements in conversion but unacceptable lengthening of the reaction time. It became clear that lithium perchlorate was superior to tetrabutylammonium perchlorate (entry 14 vs 11) plus coupled with its ease of separation from organic products was selected as the electrolyte. To probe whether additional water improved the reaction yield (entry 15) based on a postulated mechanism led to a reduced yield versus entry 14. Furthermore, passing no electrical current led to no reaction (entry 16). To test the hypothesis that C-H bond Table 3. Results of direct alcohol electro-oxidation to aldehydes (c.e. = current efficiency).
oxidation adjacent to an ester was possible, a collection of benzylic esters were prepared. Our initial results using the optimised procedure are detailed in Table 2. It was found using the optimised conditions, appreciable amounts of the desired aldehyde (2a-2c) were obtained from the benzylic acetates (1a-1c). To address, the mechanism issue identified in Figure 1, a simple aqueous base wash removed the acid by-product. However, there were still limitations to this approach for example electro-oxidation of cyclic benzylic ester 1d or homologated ester 1e, did not afford the aldehyde 2d and 2e, respectively. These limitations coupled with the use of an acyl ancillary group still did not meet our green chemistry standards due to the additional manipulation step required to prepare the acetate. We therefore considered whether a stabilising group on the heteroatom adjacent to the C-H bond was indeed essential to electro-oxidation. Promisingly, cyclic voltammetry measurements on benzyl alcohol 3a showed an oxidation wave at +1.2 V relative to Ag/AgCl. 19 Using our previously optimised conditions for the ester electro-oxidation we explored the direct mediator-free oxidation on a range of commercially available alcohols (Table 3). It was possible to cleanly convert benzyl alcohol (3a) to benzaldehyde (2a) under the mild conditions of lithium perchlorate in methanol in near quantitative yield and in improved yield compared to the chemical manipulation of forming the benzyl acetate (99% vs 80%, respectively). 20 The scope of this reaction was further investigated via exploration of the effects of electron withdrawing and donating group around the ring system and steric effects around the reacting centre. The use of a chlorine atom as an electron withdrawing group in the ortho-(2f, 81%), meta-(2g, 55%) and para-(2h, 86%) positions was well tolerated. The use of a methyl group as a mild electron donating group in the ortho-(2i, 66%), meta-(2j, 99%) and para-(2b, 83%) positions also afforded good to excellent yields of the aldehyde.
Intriguingly, a strong electron donating group para to the reacting centre (from para-methoxy benzyl alcohol, 3i (not shown)) resulted in a greatly reduced yield of aldehyde 2k (27%). The relatively low yield of 2k compared with other benzaldehydes is likely to be due to the para-methoxy group stabilizing the carbon-centred benzylic radical leading to further unproductive reaction pathways and oxidative decomposition of 2k.
The result for a strongly electron withdrawing group in the 4-position of the substrate 3j afforded only trace conversion to the desired product as a result of the highly electron withdrawing nature of the para-nitro group. Based on the experimental results above and EPR literature 21 of related systems a tentative mechanism can be proposed (Scheme 1).
Electro-oxidation of the benzyl alcohol reveals a benzyl radical, subsequent oxidation of the hydroxyl group will deliver an aldehyde after deprotonation. Similarly, the benzylic acetate system could proceed via analogous initial steps to an Oacyloxonium species that could be trapped with water or methanol to deliver a hemiacetal or acetal species, respectively. In turn, the hemi-acetal or acetal intermediate would collapse upon work-up to deliver the aldehyde (Scheme 1B). We have shown that mild electron withdrawing and donating groups are tolerated in both systems but those that strongly donate or withdraw electrons were less compatible with direct electro-oxidation. Nitrogen containing heterocycle (3n) and non-benzylic substrates (3k-3m) gave trace conversion (<5%) to the aldehyde. However, changing the benzene ring to a thiophene was successful, affording aldehyde 2h in quantitative yield, demonstrating the utility of the approach in other classes of heterocyclic systems.

Conclusion
The use of mediator-free, galvanostatic, direct electrooxidation of benzylic alcohols and acetates shows scope on a range of substrates and offers an alternative and complementary approach to the preparation of valuable aldehydes. Mechanistic investigations to determine the sequence of electro-oxidation steps are now underway.

Experimental section General methods
Reactions were carried out under nitrogen. Organic solutions were dried over MgSO 4 . Starting materials were purchased from commercial suppliers and were used without further purification. Solvents were dried over molecular sieves (3-4Å) Flash silica chromatography was performed using Sigma-Aldrich high-purity grade, pore size 60 Å, 200-400 mesh particle size silica gel. 1 H and 13 C NMR spectra were recorded on a JEOL ECS 400 NMR Spectrometer at 400 MHz or Bruker AVIII 300 or 400 MHz spectrometers. Chemical shifts (δ) are reported relative to TMS (δ=0) and/or referenced to the solvent in which they were measured. All chemical shifts (δ) are reported in parts per million, and coupling constants (J) are reported in Hertz. Low and High-resolution mass spectrometry analysis were obtained using an Agilent 6450 LC-MS/MS system. Infrared (IR) spectra were recorded on a ThermoScientific Nicolet Impact-380 ATR-FTIR spectrometer.

Conflicts of interest
There are no conflicts to declare.