Organic Letters
Letter
A more promising way to achieve a catalytic [2 + 2]
photocycloaddition was discovered when studying the reaction
of 2-benzimidazolyl styryl sulfones. In contrast to aryl sulfones
3t−3v which had nicely provided the respective cyclobutanes
4t−4v (Scheme 2), the attempted reaction of the N-benzyl
protected 2-benzimidazolyl sulfone 7a with 2,3-dimethyl-2-
butene took an unexpected course. Under the optimized
conditions employed for the other sulfones (Scheme 2),
cleavage of the C−S bond was observed19 and the C2-alkylated
product 8 was isolated in 56% yield (Scheme 3). It was shown
Figure 3. Enantiomerically enriched cyclobutane products 9 obtained
by irradiation of the respective sulfones 7, excess (30 equiv) 2,3-
dimethyl-2-butene and catalytic quantities (2 mol %) of chiral
rhodium catalyst (+)-Λ-10.
Scheme 3. Photochemical Reactions of Sulfone 7a and 2,3-
Dimethyl-2-butene: Fragmentation/C−C Bond Formation
upon Direct Irradiation and Lewis Acid Catalyzed
Cyclobutane Formation
For the N-methyl analogue 7b irradiation at λ = 437 nm
(LED) with an excess of 2,3-dimethyl-2-butene delivered
product (+)-9b in 67% yield and 54% ee. At a longer
wavelength (λ = 457 nm) the enantioselectivity was much
higher (87% ee) but the reaction became impractically slow
(15% conversion after 24 h). The absolute configuration
assigned to products 9 is tentative and is based on a chelate-
type association of sulfones 7 to rhodium complex (+)-Λ-10
and an s-trans conformation around the S−styryl bond.
Preliminary DFT calculations (see the SI for details) on the
1:1 complex between 7b and (+)-Λ-10 corroborate an
enantioface differentiation that leads preferably to a cyclo-
butane with the depicted structure (+)-9b.
In summary, a straightforward access to sulfonyl-substituted
cyclobutanes has been established which relies on the direct
irradiation of the respective sulfones with UV light. The
method complements the existing synthetic arsenal of [2 + 2]
photocycloaddition chemistry and displays some potential for
enantioselective catalysis. The combination of Lewis acids with
triplet sensitizers has not yet been explored22 and might be a
viable route to enantiopure compounds. Other chelating Lewis
acids beyond complex (+)-Λ-10 might provide even better
results in combination with a sulfonyl substituent that allows
for chelation (e.g., 2-imidazolyl, ortho-hydroxyphenyl).
that this cleavage is not restricted to styryl sulfones but occurs
also with a methyl sulfone (see the SI for details) and the
reaction likely proceeds via the respective 2-benzimidazolyl
radical. Remarkably, the C−S bond scission was completely
suppressed when the reaction was performed with the same
substrates at λ = 368 nm and at −78 °C in the presence of
catalytic amounts of AlBr3 as a Lewis acid. Under these
conditions, the desired cyclobutane 9a was isolated in 76%
yield. Control reactions indicated that both the irradiation
source (no reaction in the absence of light) and the Lewis acid
were required for the reaction to occur.
We hypothesized that a chelating chiral Lewis acid would
potentially bind to sulfone 7a via the sulfonyl oxygen atom and
the nitrogen atom N3 of the benzimidazole. Despite the fact
that free rotation around the S−styryl bond remains possible
within 7a even in a putative chelate complex, we performed a
preliminary screening on potential chiral Lewis acids. From
these experiments, chiral-at-metal rhodium complex (+)-Λ-10
emerged as the most promising catalyst. In a seminal study,
Meggers and co-workers had demonstrated that this catalyst
can be successfully employed for enantioselective [2 + 2]
photocycloaddition reactions of α,β-unsaturated carbonyl
compounds20 and its preparation had been described in
detail.21 Still, in the specific [2 + 2] photocycloaddition of stryl
sulfones the catalyst turned out to be capricious. Most
experiments were performed with sulfone 7a and 2,3-
dimethyl-2-butene at different wavelengths (2 mol % catalyst).
There was an inverse correlation between yield and
enantiomeric excess (ee). If the yield was low, the ee was
high and vice versa. The low yields do not result from retro
[2 + 2] photocycloaddition reactions but rather from low
reaction rates. The best compromise (41% yield, 77% ee) was
found for product (−)-9a to be an irradiation with a light
emittiing diode (LED) that exhibits a narrow emission band
centered at λ = 420 nm (Figure 3). Emission spectra of all light
ASSOCIATED CONTENT
* Supporting Information
■
sı
The Supporting Information is available free of charge at
Materials and methods, optimization of reaction
conditions, synthetic procedures and full character-
ization of new compounds (3f, 3l, 4a−4v, 5a−5d, 7a,
7b, 8, 9a, 9b), spectroscopic data, NMR spectra, DFT
calculations, HPLC traces, emission spectra of the light
FAIR data, including the primary NMR FID files, for
compounds 3f, 3l, 4a−4v, 5a−5d, 7a, 7b, 8, 9a, 9b
AUTHOR INFORMATION
Corresponding Author
■
Thorsten Bach − Department Chemie and Catalysis Research
Center (CRC), Technische Universität Munchen, 85747
̈
5676
Org. Lett. 2021, 23, 5674−5678