Organic Letters
Letter
light.7 This prerequisite for UV light is often associated with
challenges in the scalability of these transformations and
significant safety concerns due to possible UV radiation
exposure. Despite these drawbacks, complementary visible-
light-induced processes are significantly less advanced,
although the milder reaction conditions are expected to
significantly increase the functional group tolerance, scalability,
̀
and safety of these Paterno−Buchi reactions. Importantly, two
̈
successful examples of these [2 + 2] cycloadditions relying on
visible light irradiation have recently been reported by the
groups of Dell’Amico and Ouyang.8,9
Although an important advancement, these reactions are
inherently limited to starting materials that exhibit the required
photophysical features for direct excitation (Figure 1B). We
envisioned that a potentially more general synthetic approach
Figure 2. Evaluation of photocatalysts in visible-light-enabled
̀
Paterno−Buchi reactions. Reactions were performed with 12 (0.05
̈
mmol), alkene 13 (10 equiv), and photocatalyst (2.5 mol %) in
MeCN (0.1 M) under blue LED irradiation (465 nm) at ambient
temperature (fan cooling) for 0.5 h. aYield determined by quantitative
1H NMR analysis from the crude mixture using an internal standard.
̀
to visible-light-enabled Paterno−Buchi reactions could rely on
̈
utilizing a photocatalyst (10) to be the primary excited state
species in solution to ultimately generate the desired excited
state carbonyl (7) via triplet energy transfer (Figure 1C). In
related efforts, our lab has recently demonstrated that triplet
energy transfer can efficiently activate functionalized imines to
enable the synthesis of similarly strained four-membered
for additional details). The reaction was determined to
proceed in good yields of up to 75% relying on 1,2-
dichloroethane, methylene chloride, acetone, toluene, ethyl
acetate, methanol, and acetonitrile as reaction solvent. Catalyst
loadings as low as 0.5 mol % of 15 and equimolar amounts of
alkene 13 proved to be sufficient in forming oxetane 14 in up
to 70% yield.
̀
nitrogen-containing heterocycles in aza Paterno−Buchi re-
̈
actions.10 This transformation relies on a visible-light-
absorbing iridium photocatalyst to access the corresponding
excited state of the substrate via triplet energy transfer. We
envisioned a similar design principle to be applicable to
̀
carbonyl activation in a visible-light enabled Paterno−Buchi
̈
Next, we set out to investigate the scope of this visible-light-
reaction protocol.11 In this reaction platform, the photocatalyst
(10) is expected to absorb visible light, resulting in population
of the triplet excited state (11), followed by subsequent energy
transfer to the carbonyl (6), producing the reactive triplet
̀
enabled Paterno−Buchi reaction (Figure 3). Specifically, a
̈
variety of aryl glyoxylates and different alkene reaction partners
were investigated upon their ability to undergo the desired
transformation. Methyl, ethyl, adamantyl, and tert-butyl esters
all provided good to excellent yields of oxetanes 16−19 of up
to 99%. Interestingly, direct irradiation of ethyl benzoylformate
without photocatalyst or alkene showed complete decom-
position. We were pleased to see that, under our standard
conditions, rapid conversion to oxetane 17 was observed (see
electron-donating and -withdrawing groups on the aromatic
ring resulted in the formation of the desired oxetane products.
The anisole derivative 21 was obtained in 72% yield, while the
electron-withdrawing para-nitrile provided oxetane 20 in 66%
yield. Interestingly, an additional ester present on the aromatic
ring was tolerated under the optimal reaction conditions as
shown in the formation of oxetane 22 in 91% yield.
Furthermore, the para-bromo aromatic glyoxylate was
evaluated and proceeded in 77% yield to result in oxetane
23. Further efforts centered on the investigation of alkene
reaction partners showed that di-, tri-, and tetrasubstituted
alkenes undergo the desired transformation providing oxetanes
14 and 24−28 in yields ranging from 46% to 83%. Specifically,
cyclohexene as an alkene partner successfully formed the
bicyclic oxetane 25 in a good yield of 65%. Moreover, aromatic
alkenes like benzofuran and furan gave the cycloadducts 26
and 27 in 64% and 46% yield, respectively. Additionally, an
acyclic vinyl ether was able to provide oxetane 28 in 68% yield.
Mechanistic investigations were subsequently initiated to
gain additional insights into the controlling features of this
̀
energy state (7) allowing for efficient Paterno−Buchi reactions
̈
(Figure 1C). Importantly, this mode of activation negates the
need for visible-light-absorbing carbonyls, instead requiring
that their triplet energy (ET) lies close enough to that of the
photocatalyst. Based on their low triplet energy (approximately
60 kcal mol−1), we hypothesized that glyoxylate derivatives
could be a good starting point for the development of a visible-
̀
light-enabled Paterno−Buchi reaction relying on triplet energy
̈
transfer.12
Our initial investigations into the development of a mild
̀
protocol for visible-light-enabled Paterno−Buchi reactions was
̈
focused on methyl benzoylformate (12), due to its superior
stability compared to other glyoxylate substrates. Specifically,
Norrish type II reactivity is a known competing reaction path
for glyoxylate substrates under irradiation with UV light (see
alkene 13 were converted with catalytic amounts of [Ir(dF-
(Me)ppy)2(dtbbpy)]PF6 upon irradiation with 456 nm light in
acetonitrile for 0.5 h, the formation of oxetane 14 was observed
in 72% yield (entry 1, Figure 2). A selection of photocatalysts
varying in their triplet energies was subsequently evaluated
(entries 2−6, Figure 2). Ultimately, photocatalyst 15 was
identified as optimal due to its established reactivity and high
yield of oxetane 14 (entry 2, Figure 2), while photocatalysts of
similar triplet energies compared to 15 were found equally
sufficient in catalyzing the transformation. Importantly,
catalysts of lower triplet energy failed to provide the desired
oxetane product 14, which is consistent with a reaction
mechanism relying on an energy transfer mechanism (entries
4−6, Figure 2). Subsequent efforts focused on additional
reaction optimization, including varying the solvent, catalyst
̀
visible-light-enabled Paterno−Buchi reaction. Specifically, we
were interested in investigating the difference between a
possible direct excitation of the carbonyl and our proposed
triplet energy transfer process. When glyoxylate 12 and alkene
13 were irradiated with UV-A light, the formation of oxetane
̈
B
Org. Lett. XXXX, XXX, XXX−XXX