Organic Letters
Letter
of CuI/CuII and MnII/MnIII, taking also place above 2.5 V.
According to the known electrochemical oxidation behavior of
acetonitrile at platinum electrodes in the presence of water,
solvent oxidation appears as a logical, and in this case
synergistic, candidate.26
Scheme 1. Standard and Electrochemical Wacker−Tsuji
Oxidation
Once the optimization process was completed, different
combinations of manganese and copper compounds were
tested as potential catalytic systems (Table S2). It is interesting
to note that combinations of chlorides and bromides of Mn(II)
and Cu(II) are almost equally suitable catalysts for the process
(entries S1, S5, and S15), with the combined use of both
dibromides (entry S1) leading to the highest (86%) yield. On
the contrary, the combination of both chlorides (entry S12)
results in a mediocre catalyst. Thus, the presence of bromide
ions is of primordial importance for catalytic activity. The
reaction did not produce any 2a when Cu(acac)2, involving a
highly chelated Cu(II) species, was used as a co-catalyst (entry
S6). It is also worth mentioning that polymerization of styrene
was observed when Cu(OTf)2 replaced CuCl2 as the Cu(II)
source (entry S4).27 As already mentioned, no acetophenone
was detected in the absence of a Mn(II) source (entries S16−
S20).
Once the optimal reaction conditions had been fully
established, the applicability of the Mn/Cu co-catalyzed
electrochemical oxidation was studied on a representative
series of substrates 1a−z containing in their structures an aryl
group conjugated with a carbon−carbon double bond
(Scheme 2). In general, the reaction tolerates alkyl groups
and medium polarity substituents, such as halogen or
trifluoromethyl groups and even groups with strong with-
drawing character (2n). With respect to regiochemistry, para-
substituted substrates are those leading to higher yields.
Also in this respect, the reaction appears to be sensitive to
steric effects, since ortho-substituted substrates, like 1d, 1f, and
1h, afford the corresponding oxidation products in lower yield
than the corresponding para-isomers and heavily ortho-
substituted substrates, like 1u and 1x, fail to react.
Interestingly, substrates containing 1,2-disubstituted double
bonds are efficiently oxidized, irrespective of their cyclic (1q,
1s) or acyclic nature (1p, 1r). In the case of 1p, no appreciable
bias exists with respect to the stereochemistry of the double
bond in the substrate. 2H-Chromene (1t), a substrate
belonging to an important class of natural substances, was
successfully oxidized to chromanone 2t (47% yield). However,
the analogue 2,2-dimethyl-2H-chromene 1y failed to react,
thus indicating that heavy substitution on the double bond is
deleterious to the reaction. On the other hand, when extension
of the electrochemical oxidation to commercially available
dihydroquinoline 1z was attempted, fast deprotection of the
carbamate moiety took place, but oxidation did not proceed. It
is also worth mentioning that allylbenzene, a regioisomer of 1o
of nonstyrene nature, completely failed to provide the
corresponding oxidation product.28 Finally, p-methoxystyrene
(1t) and m-nitrobenzaldehyde (1u) failed to provide the
corresponding acetophenone products 2t and 2u for
completely different reasons. While 1t underwent a very fast
reaction, but led to ill-defined products of oligomeric nature,
1u was reluctant to electrochemical Wacker−-Tsuji oxidation.
This behavior can be rationalized through the tentative
mechanistic proposal shown in Figure 1.
of this mechanism under electrooxidative conditions. As shown
in entries S1 and S2, no coupling took place between primary
alcohols used as cosolvents and styrene 1a, while acetophe-
none 2a was detected in low yield when 1:1 2-propanol/water
was used as a solvent in the presence of 5 mol % MnBr2 as
catalyst at 60 °C (entry S3). Binary mixtures of polar aprotic
solvents and water also afforded poor results (entries S4 and
S5), but the use of 4:1 vol/vol acetonitrile/water was
somewhat promising, especially when performing the reaction
at 60 °C (entries S7 and S8). The use of NiCl2 (entry S9),
FeCl3 (entry S10), or even PdCl2 (entry S11) instead of
MnBr2 was deleterious. On the other hand, the combined use
of MnBr2 and CuCl2 (5 mol % each) led to a very significant
yield increase (67%, entry S12). Alternatively, the exclusive use
of CuCl2 did not produce any detectable amount of 2a (entry
S13).
Once we had established the determining roles in the
reaction of MnBr2 and CuCl2, we proceeded to optimize the
initial concentration of 1a (Figure S4) and of the support
electrolyte LiClO4 (Figure S5) were also optimized.
Interestingly, the reaction gave a similar yield in the presence
or in the absence of oxygen in the reaction cell. Besides the
operative advantage in reaction practicality, this observation
provides a clear indication that oxygen reduction is not
involved in the observed reaction and that the oxygen atom in
the final product arises from water. We accordingly studied the
effect of the proportion of water in the solvent system on the
efficiency of the process (Figure S3). As anticipated, the
presence of water in the solvent is a requisite for the reaction
to take place, and the optimal yield (77%) is achieved with a
volumetric composition of 80% MeCN and 20% water. Further
increases in the amount of water lead to a rather sharp decrease
in yield, probably because of the insolubility of styrene in those
solvent mixtures.
As a final parameter, we studied the effect of the applied
potential on the reaction yield while working in constant
voltage mode (Figure S6). The onset voltage for the reaction
to proceed was shown to be 1.7 V. The yield slightly increased
with the applied potential, and at ca. 2.5 V, a sudden increase
in the slope of the yield vs voltage graph occurs, with a
maximum 85% yield being achieved at 2.8 V.25 These
experiments were later repeated in a single-compartment,
three-electrode cell including a AgCl reference electrode (see
Figure S7). Interestingly, the results between 2.4 and 3.0 V are
exactly duplicated, with the highest (85%) yield being achieved
again at 2.8 V. The observed yield vs applied voltage behavior
is strongly indicative of some additional redox process, on top
As can be seen, we propose that three parallel oxidative
events could take place at the carbon felt anode: the standard
oxidation of Mn(II) to Mn(III) [E° = 1.56 V29 and peak at ca.
B
Org. Lett. XXXX, XXX, XXX−XXX