Angewandte
Chemie
the corresponding adducts 4 in excellent yields. The reaction
of bromo malonate (2b) was successfully implemented on
a gram scale, and the product 4b was isolated in 99% yield
HBr formation. We confirmed that trace amounts of acid
were also generated during the ATRA reaction, which greatly
[
14]
affected the reactivity, thus explaining the need for a base
[15]
(
3.3 g) while the catalyst 1a was not consumed and almost
such as 2,6-lutidine (entry 3 in Table 1). Detection of such
[16]
fully recovered (> 90%). a-Bromo ester derivatives were
smoothly converted into the corresponding products 4c–e.
The method allowed direct introduction of functional groups
other than an ester, as the alcohol 4 f was isolated in excellent
yield and the nitrile 4g was prepared from bromoacetonitrile.
In addition, the use of carbon tetrachloride and perfluor-
ohexyl iodide granted access to the synthetically valuable
polyhalogenated products 4h and 4i, respectively.
photolysis products
suggests that 1a can be capable of
sensitizing the alkyl halides 2 to the excited triplet state upon
*
illumination, thus inducing an n!s transition which would
result in a rapid homolytic dissociation of the CꢀX bond in
[
17]
2. The photochemistry of 1a is dominated by reactions of
[13]
the lowest electronically excited triplet state, which pos-
[
18]
sesses a relatively long lifetime along with a triplet energy
ꢀ1
(E ) of 300 kJmol . This energetic value is in the range of the
T
We next examined the reactivity of different olefins with
b (Figure 2b). Both terminal olefins and the inherently less
carbon–halogen bond dissociation energies of the alkyl
ꢀ1
2
halides used in this ATRA methodology (260–300 kJmol ,
see Table S1 in the Supporting Information), which is
congruent with a homolytic cleavage of 2 induced by an
energy-transfer mechanism. It is of note that the alkyl halides,
reactive internal cyclic alkenes, including 2-norbornene,
cyclohexene, and cyclooctene, were competent substrates,
thus enabling the construction of complex frameworks 5a–g
from simple precursors. As a testament to the very mild
nature of the protocol, terminal olefins adorned with a wide
array of sensitive functionalities were efficiently transformed
into the corresponding products 5h–o. The ATRA process
was also possible with a terminal disubstituted olefin and an
alkyne (Figure 2c). Electron-rich silyl enol ethers allowed the
preparation of functionalized carbonyl compounds (Fig-
ure 2d). In addition, the naturally occurring compounds
[
19]
by means of the external heavy atom effect, can facilitate
the intersystem crossing which drives the singlet–triplet state
conversion of the excited 1a.
Consistent with the triplet sensitization mechanism
depicted in Figure 3a, the model photo-organocatalytic
ATRA reaction was completely inhibited by the presence of
oxygen, an efficient triplet state quencher. The process was
largely insensitive to solvent polarity (e.g. DMSO, DMF,
CH Cl , benzene, and n-hexane offered 83, 95, 85, 95, and
(
R)-limonene and (ꢀ)-b-pinene readily participated in the
2
2
ATRA process to give 5t and 5u, respectively (Figure 2e).
The opening of the cyclobutane ring leading to 5u further
highlighted the radical nature of the process. Similarly, the
diene 2v gave the product 5v as a result of a 5-exo-trig radical
cyclization (Figure 2 f). The selectivity for all the ATRA
reactions was excellent, since dimers or dehalogenated
products were not detected. Control experiments, performed
for all the reactions presented here, confirmed that the
absence of the aldehyde 1a or of light completely suppressed
the process.
86% yield, respectively, of 4a after 18 h). This observation
further supports an energy-transfer pathway, since charged
radical ion intermediates, possibly generated through elec-
tron-transfer manifolds, would be strongly destabilized in
nonpolar media. In addition, the rate of the reaction was
greatly increased when conducted in solvents with a high
viscosity, such as cyclohexanol and benzyl alcohol (viscosity at
258C = 5.47 and 57.5 cP, respectively). We found that, in
glycerol (934 cP), the model reaction reached completion
(95% yield of 4a) after 7 hours when using only 5 mol% of
1a. This reactivity profile is consonant with the notion that the
rate of exergonic triplet-energy transfers generally increases
to the limit of diffusion control when conducted in viscous
We then sought to elucidate the mode of catalysis of 1a.
We determined the light frequency required for an efficient
ATRA process between 2a and 3a catalyzed by 1a (the
model reaction depicted in Table 1). The use of a 300 W
Xenon lamp equipped with a cut-off filter at l = 385 nm
resulted in no reaction. Complete reactivity was restored
when using a band-pass filter at l = 360 nm. These experi-
ments indicated that the near UV part of the CFL emission
[20]
media.
We also performed the model reaction in the presence of
a triplet quencher (2,5-dimethylhexa-2,4-diene; entry 1 in
Figure 3b) and an additive with a lower triplet-state energy
than 1a (entry 2), which effectively quenched the catalyst,
thus ultimately lowering the reaction rate. Finally, we
reasoned that, on the basis of the proposed mechanism,
other organo-sensitizers could be as efficient as 1a if provided
with adequate excitation and triplet-state energies. When
performing the ATRA reaction under 23 W CFL irradiation,
the use of benzophenone (1b), 4,4’-dimethoxy benzophenone
(1e), and carbazole (1 f) inferred only a moderate reactivity
(Figure 3c). However, much higher conversions were
obtained using a 15 W black light CFL bulb (lmax = 360),
which showed a broader band in the near-UV region (for the
emission spectra of the bulbs used in these experiments, see
Figures S1 and S2 in the Supporting Information).
[12]
spectrum was required to bring the catalyst 1a to an excited
*
state through the established symmetry forbidden n!p
[
13]
transition manifold (lmax = 344 nm for 1a).
In contrast,
the organic halides 2, the olefins 3, and 2,6-lutidine do not
absorb within this frequency region. We then performed
investigations to correlate the photochemical activity of the
catalyst 1a with the substrates. We did not detect any ground-
state association between the reaction components by spec-
troscopic and absorption measurements. The possibility for
the olefins 3 to undergo a photochemical pathway when
mixed with 1a under irradiation was excluded, since Paternꢀ–
[
13]
Bꢁchi cycloadducts were not detected. In contrast, illumi-
nating a mixture of 1a and halide 2a provided trace amounts
of the corresponding dehalogenated malonate derivative
along with acidification of the medium, likely arising from
In summary, we have shown that the photochemical
[22]
activity of a simple organic molecule enables the intermo-
lecular atom-transfer radical addition of haloalkanes onto
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!
Arceo, Elena
