Organic Letters
Letter
benzo[h]quinolone afforded only C2 product 3i. Five
membered fused heteroarenes are amenable to the alkylation
as well (3j−3k). Further extension to challenging pyridine
derivatives gave monoalkylated products 3l−3n in regiose-
lective manner. To prove the robustness of this protocol, we
applied this protocol to late-stage functionalization of agro-
chemical and pharmaceutical agents. Hydroquinine, cincho-
nine and fasudil underwent this CDC alkylation process
smoothly, providing 3o, 3p and 3q in 41%, 37% and 29%
yields, respectively. Furthermore, a gram-scaled reaction by
reacting 1.0 g of 1a (7 mmol) with cyclohexane afforded 3a
(1.08 g) in 69% yield with 27% recovery of 1a.
2:1. Alkylation of linear alkanes only occurred at the methylene
group, providing products 4m′ and 4m′′ in a ratio of 5:4.
To gain insight into the reaction mechanism, we conducted
a series of experiments. First, a light on/off experiment
demonstrated that the transformation needed continuous
irradiation of visible light (see the SI, Scheme S1). The radical
scavenger, 2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO),
inhibited the alkylation process and allowed more than 95%
recovery of the starting materials, implying a single electron
transfer (SET) process (Scheme 4). Radical capture reaction
with diphenylethene revealed the generation of cyclohexyl
radical during the process (Scheme 4).
Next, diverse C(sp3)−H sources were subjected to the
standard conditions (Scheme 3). Cycloalkanes underwent the
Scheme 4. Cyclohexyl Radical Capture Experiments
a b
,
Scheme 3. Substrate Scope of Diverse C(sp3)−H Sources
Furthermore, the reaction vanished in the absence of
oxidants even with larger loadings of photocatalyst (see the
SI, Table S2, entry 2). However, the Minisci reaction of 1a and
2a resulted in successful alkylation of 1a (Table 1, entry 8).
These results together with the inhibition test by anthracene
(Table 1, entry 2) imply that 4-CzIPN is an energy transfer
catalyst rather than a photoredox catalyst for the redox process
and hydrogen abstraction. This hypothesis is further confirmed
by the redox potential of B and 4-CzIPN. As the reaction
undergoes acid-free conditions, the intermediates and plausible
pathway would be distinct from the previous reports.24b24c In a
previous reported mechanism, B is normally protonated for a
further redox process owing to an acid additive. However, in
this acid-free conditions, in either the photoredox catalysis
pathway or energy transfer routine, B is in its neutral form. The
oxidation potential of B can be estimated using literature data
(Ep/2red = −1.19 V vs SCE).29 The lower reduction potential of
a
Reaction conditions: 1 (0.1 mmol), 2a (2 mL.), 4-CzIPN (0.01 mol
%), (NH4)2S2O8 (2 equiv), and DMSO (1.8 mL) under blue LED
irradiation under argon atmosphere for 15 h. Isolated yield of the
product.
b
red
Minisci reaction smoothly with 1a, providing 4a and 4b. Given
that the rigid caged structure of adamantane leads to an
increased 3 °C−H bond dissociation energy (BDE) of 99 kcal/
mol over the 2° C−H BDE of 96 kcal/mol,26 it is unexpected
to observed that the CDC process occurred at 3° C−H bond
of both adamantane and adamantanyl ketone in a regiose-
lective manner (4c and 4d). Compared to the recent works,27
it demonstrates that our visible light catalytic system would be
an alternative inexpensive strategy for selective functionaliza-
tion of adamantane derivatives, the unique structure and
chemical properties of which have led to many applications in
nanoscale frameworks, optical materials, and clinically
approved drugs (e.g., memantine, antidementia).28 Further
extension of this protocol to ethers also gave desired products
4e−4l in moderate to excellent yields. Notably, s-trioxane
failed to react with 1a under acid conditions (Table 1, entries
15−16) owing to its instability in the presence of acids. In the
case of 1, 2-dimethoxy ethane the arylation takes place at two
positions, viz., methylene C−H and methoxy C−H, to give
isomers 4k′ and 4k′′ were in a total yield of 63%, in a ratio of
4-CzIPN (Ep/2 = −1.22 V vs SCE) over the complex B
suggests that the oxidation of radical B to produce 3a with 4-
CzIPN is thermodynamically unfavorable. Additionally, given
the common O−O bond energy of 45 kcal/mol,25 the energy
of the triplet state of *4-CzIPN (ET > 59 kcal/mol)21 is
sufficient for homolytic cleavage of the O−O bond in
persulfate via energy transfer, and the resulting sulfate radical
would be the hydrogen abstractor for alkanes [O3SO−H− BDE
= 105.0 kcal mol−1 versus C(sp3)−H of cyclohexane BDE =
100.0 kcal mol−1].30
On the other hand, we conducted a series electron
paramagnetic resonance experiments between persulfate and
4-CzIPN including (NH4)2S2O8/5,5-dimethyl-1-pyrroline N-
oxide (DMPO) under irradiation (I), (NH4)2S2O8/DMPO
without irr-adiation (II) and (NH4)2S2O8/DMPO/4-CzIPN
under irradiation (III) (Scheme 5). DMPO trapped sulfate
•−
radical signals (DMPO−SO4
with hyperfine splitting
constants of αN = 13.2 G, αH = 9.6 G, αH = 1.48 G, and αH
= 0.78 G)31 appeared in both I and III, while the signal
vanished without irradiation (II). These results suggest the
C
Org. Lett. XXXX, XXX, XXX−XXX