Communication
red
E1/2 [IrIV/III]= +1.39 V vs. SCE in MeCN) was the most
either in the ground state or reduced form makes it unlikely
that excitation of the viologen co-catalyst plays a part in the
reaction mechanism.[17,18] α-Amino acids could also be readily
converted to their amide analogues in good efficiency (30–35,
45–85% yield), with both benzylic (30 and 31) and non-
benzylic (33–35) amino acids being well tolerated. The
transformation was amenable to a variety of ring sizes (33–
34), as well as acyclic variants (35). This methodology thus
provides a unique and facile means for the synthesis of amides
and lactams from amino acid precursors. Finally, we sought to
investigate the use of primary benzylic acids in this catalytic
protocol. We anticipated that primary acids would be challeng-
ing substrates due to the known tendency for aldehydes to
undergo aerobic oxidation to the corresponding carboxylic
acids under an atmosphere of oxygen. However, we were
pleased to find that under our optimized conditions arylacetic
acids could be converted to the corresponding substituted
benzaldehydes in excellent yields (22–24, 71–79% yield).
We then sought to explore whether other abundant func-
tional groups could also undergo oxidative dehomologation to
yield carbonyl products. Since the autoxidation of aldehydes
to their carboxylic acid analogues is a well-established
process,[19] we hypothesized that simple aldehydes would also
be amenable as substrates via a putative carboxylic acid
intermediate.
effective catalyst for this reaction (entry 4, 87% yield). This
phenomenon can potentially be attributed to more efficient
singlet oxygen sensitization by [Ir(dF(CF3)ppy)2(dtbbpy)]
PF6.[14] This process is detrimental to our reaction as the
excited *Ir(III) state of the catalyst is quenched without net
electron transfer, regenerating the Ir(III) ground state. To
facilitate generation of the more oxidizing Ir(IV) state, we
evaluated a range of oxidizing co-catalysts, and were pleased
to find that addition of 5 mol% ethyl viologen diperchlorate
(13) doubled the rate of the reaction (see Supporting
Information). Viologen catalysts are known to act as electron
shuttles in photochemical reactions,[15] and in addition their
reduced forms can readily reduce oxygen to the superoxide
anion.[16] We hypothesize that 13 can thus act as an electron
shuttle between the photocatalyst and molecular oxygen,
compensating for the low solubility of oxygen in our reaction
solvent.[10b] These combined changes to the reaction conditions
led to a 90% yield of acetophenone, with no 1-phenylethanol
observed, offering complete selectivity for ketone 10 over the
alcohol product (entry 5, 90% yield). Finally, control experi-
ments demonstrated that photocatalyst, visible light, and an
atmosphere of oxygen were all necessary for the reaction
(entries 6–8).
With the optimized conditions in hand, we investigated the
scope of the decarboxylative oxygenation (Table 2). A range
of secondary benzylic carboxylic acids were converted to their
ketone analogues in high efficiency (10, 14, 15 and 16, 75–
86% yield). The reaction can also generate fused bicyclic
fused bicyclic ketones, including tetrahydronaphthyl and
indanyl scaffolds (17 and 18, both 77% yield). Notably, the
drug molecules flurbiprofen, ketoprofen, and naproxen were
converted to the corresponding ketones in excellent yields (19,
20, and 21 respectively, 80–90% yield), highlighting the
application of this method to biologically-relevant molecules.
In addition to benzylic substrates, we were pleased to find that
aliphatic ketones were also generated in good levels of
efficiency (8, 25–29, 59–82% yield). The reaction was
amenable to acids containing a wide variety of aliphatic
structures, including acyclic (27 and 29), cyclic (8, 25, and
26), and bicyclic (28) scaffolds. In these cases, the viologen
co-catalyst proved vital for achieving high levels of efficiency
and selectivity for the ketone product over the reduced alcohol
byproduct. For example, in the case of product 8, we observed
a 40% yield of ketone 8 and 13% of the corresponding
secondary alcohol in the absence of the viologen co-catalyst.
However, upon addition of viologen co-catalyst 13, 67% of
the ketone and 5% of the alcohol byproduct were obtained
(see Supporting Information). As formation of the alcohol
product likely arises via photocatalyst-mediated reduction of
the hydroperoxide intermediate, we hypothesize that the
viologen co-catalyst helps favor the ketone product by
accelerating oxidative quenching of the *Ir(III) excited state.
The fact that no reaction is observed without photocatalyst
when viologen is present, as well as the fact that the viologen
salts absorb minimally in the blue region of the spectrum
Although minimal overoxidation of benzylic aldehydes
was previously observed, we hypothesized that longer reaction
times would enable the formation of the requisite acid
intermediate, which could then subsequently undergo decar-
boxylative oxygenation. We examined this oxidation-decar-
boxylation strategy in the context of both a benzylic and non-
benzylic aldehyde (36 and 37), and were pleased to observe
formation of ketones 10 and 8 in 85% and 17% yield,
respectively (Scheme 3, see Supporting Information).
Finally, we examined the possibility of using our decarbox-
ylative oxygenation strategy as a method for the oxidative
degradation of linear aliphatic chains. Such dehomologations
have previously been explored for the depolymerization of
lignin into simple aromatic molecules[20] and have also found
Scheme 3. Decarboxylative Oxygenation of Aldehydes.
©
Isr. J. Chem. 2020, 60, 1–7
2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
��
These are not the final page numbers!