Angewandte
Communications
Chemie
para) and a dimethyl-substituted phenyl ring as well as
These results indicate that copper(I) is essential as a single-
a polyarene naphthalene ring, were found to be suitable
substrates to afford 3G–3Q in 61–81% yields with 80–97% ee
(Table 2). Furthermore, many common functional groups,
such as ester (3R), amide (3S), nitrile (3T) and even nitro
(3U) ones, were all compatible with the reaction conditions.
Potentially reactive free aldehyde (3V) and alcohol (3W)
were well tolerated, without the need for protecting groups in
spite of the oxidative nature of this process. In addition,
substrates containing the other reactive double or triple bonds
afforded corresponding products 3X and 3Y with the addi-
tional double or triple bonds intact. More importantly,
substrate 1z without gem-disubstituents together with sub-
strates featuring a heteroaryl-substituted alkene moiety or
a 5-hexenol skeleton for the formation of a pyran ring
(Scheme 3S) all underwent the current reaction with unsat-
isfactory but promising enantioselectivity, which warrant
further condition optimization. These features indicate the
great functional group tolerance (halides, ester, amide, nitrile,
nitro, aldehyde, hydroxy, alkene and alkyne groups) of this
reaction with unique chemoselectivity, highlighting the gen-
erality of this transformation and offering opportunities for
further versatile modifications. The absolute configuration of
3V has been determined by X-ray structural analysis on its
hydrazone derivative.[11]
electron transfer catalyst to reduce Togniꢀs reagent in order
for generating CF3 radical and the activation of Togniꢀs
reagent could be facilitated by the phosphoric acid
[Eq. (1)].[13]
Then some experiments were conducted to ascertain the
role of pyridine in this reaction. In our reaction system,
pyridine derivatives could act either as a Lewis basic ligand on
copper metal[9] or as a Brønsted base (proton shuttle)[10c] to
facilitate deprotonation of alcohol. In support for the ligand
role, our initial high-resolution mass spectrometry analysis of
a reaction mixture identified a CuI-P1 complex formed from
CuBH4(PPh3)2 and P1 by ligand exchange (Figure S2). To
further provide support for pyridine as an ancillary ligand,[10]
we have surveyed a range of electronically differentiated
pyridines for this reaction (Table 1, and Figure S3a) and
found that a certain level of coordinating capability toward
copper but not Brønsted basicity is necessary for maintaining
high enantioselectivity. In particular, poorly coordinating 2,6-
di-tBu-pyridine with a pKa value (4.95 in water, Table S1)
within the optimal pKa window (3 to 5 in water) for high
enantioselectivity only delivered a comparable enantioselec-
tivity (50% ee) with that obtained in the absence of any
pyridine (45% ee). This latter fact excludes (or strongly
disfavors) a potential role of pyridine as a proton shuttle to
facilitate deprotonation of alcohol.[10c] The ligand role of
pyridine was further supported by its retarding effect on
reaction rate (Figure S3b, conducted at 108C), presumably by
stabilizing transient high-valent copper species.[9] Further-
more, the ee of products during reaction remained nearly
constant, supporting a uniform enantiodetermining transition
state along with the same reaction pathway for the reaction
(Figure S3c). Overall, the above results support that achiral
pyridine, at least primarily, acts as an ancillary ligand on
copper metal to greatly enhance the level of enantiocontrol.
On the basis of above mechanistic investigations and
previous studies,[4,5,12] a plausible catalytic cycle is tentatively
proposed (Scheme 2). At first, achiral pyridine is favorably
coordinated to CuI to form CuI-Py complex A. This complex
next reacts with CPA-activated Togniꢀs reagent by hydrogen
bonding[13] via single electron transfer, giving the crucial chiral
(LB)CuII phosphate complex B accompanied by the gener-
ation of CF3 radical. Subsequently, the addition of CF3 radical
to alkene gives a-CF3 alkyl radical C, which is trapped by B to
form a CuII species D. Subsequently a CuIII species E (path a)
may form.[4,5,9,14] In these two steps, the achiral pyridine-
coordinated copper complex and chiral phosphate counter-
anion work cooperatively to stabilize the reactive radical
intermediate and control the stereochemistry of this reaction.
Then, reductive elimination of E affords 3. However, the
other pathway via intramolecular single-electron oxidization
of intermediate D to the corresponding carbocation inter-
To gain some insight into the reaction mechanism, radical
trapping experiments were conducted by employing 2,2,6,6-
tetramethyl-1-piperidinyloxy and 1,4-benzoquinone, both of
which inhibited the reaction (Scheme S2a). Next, a radical
clock experiment with substrate 4 under the typical conditions
did not afford the expected product 5, while delivering 6 in
30% yield as a mixture of E/Z isomers, presumably via
a radical addition/cyclopropane ring opening/acid trapping
cascade process [Eq. (1)]. These observations, together with
previous studies,[4,5,12] suggest that CF3 radical is likely
generated in situ, which upon further addition to alkene
gives rise to a-CF3 alkyl radical C (Scheme 2). In addition,
only a trace amount of the desired product was obtained
either in the presence of a copper(II) salt or in the absence of
any copper catalysts (Scheme S2b). Furthermore, the reaction
did not work in the absence of phosphoric acid (Scheme S2b).
À
mediate F, which next undergoes C O bond formation to give
3, could not be excluded at the present stage (path b).
Alcohols are intrinsically less coordinative toward metal
than carboxylates and poorer hydrogen-bonding donors than
ureas, both of which render alcohols inapplicable in previ-
ously reported conditions.[4a,b,5] In this study, we have cap-
italized on a conceptually novel strategy with achiral pyridine
Scheme 2. Mechanistic proposal.
Angew. Chem. Int. Ed. 2017, 56, 1 – 5
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!