Journal of the American Chemical Society
Article
NMR signals of HPPhMes and (R)-BINAP appeared at δ
−77.12 and δ −15.22 ppm, respectively (Figure 1A and B). A
certain complex formed from Cu(CH3CN)4PF6 and HPPhMes
gave a broad peak at δ −68.25 ppm (Figure 1C). The signal of
Cu(I)-(R)-BINAP was recorded as δ −1.31 ppm (Figure 1D).
When 1 equiv of HPPhMes was added to the solution of
Cu(I)-(R)-BINAP in THF, one new sharp peak (δ 1.36 ppm)
and one new broad peak (δ −63.89 ppm) were observed,
indicating the formation of the Cu(I)-(R)-BINAP-HPPhMes
complex (Figure 1E). When 5 equiv of HPPhMes was added,
the signal at δ −15.26 ppm indicated that some amount of (R)-
BINAP was liberated from the Cu(I) complex (Figure 1F).
However, a substantial amount of Cu(I)-(R)-BINAP complex
remained, which was indicated by the signal at δ 1.47 ppm.
When 20 equiv of HPPhMes was added, although a significant
amount of free (R)-BINAP was detected, a certain Cu(I)-(R)-
BINAP complex still remained (δ 1.14 ppm) (Figure 1G).
These 31P NMR experiments suggested that HPPhMes has a
much weaker competitive coordinating ability toward the
Cu(I)-(R)-BINAP complex than HPPh2 possibly due to its
larger steric hindrance. Previously, it was realized that the
Cu(I)-(R,RP)-TANIAPHOS complex was relatively stable in
the presence of stoichiometric HPPh2.44 Thus, it was
conjectured that the Cu(I)-(R,RP)-TANIAPHOS complex
would be more stable in the presence of stoichiometric
HPPhMes. The high enantioselectivity in the present catalytic
asymmetric alkylation could be attributed to the high
performance of the Cu(I)-(R,RP)-TANIAPHOS complex in
asymmetric induction. Moreover, the 31P NMR experiment of
the mixture of Cu(CH3CN)4PF6 (1 equiv), (R)-BINAP (1
equiv), HPPhMes (1 equiv), and Barton’s base (0.1 equiv) was
performed, which gave the same signals as the mixture without
Barton’s base (0.1 equiv) (Figure 1H vs E). When 1 equiv of
Barton’s base was added, the mixture gave unidentified
multiple signals, suggesting the formation of many copper(I)
complexes (Figure 1I). Definitely, a certain catalytically active
Cu(I)-PPhMes-(R)-BINAP complex existed. However, its 31P
NMR signal was not unambiguously identified. Furthermore, it
should be pointed out that Barton’s base not only worked as an
organic base but also served as a ligand to copper(I), which
was revealed by the presence of a significant amount of free
(R)-BINAP (δ −15.37 ppm).
HPPhMes itself might serve as a ligand to Cu(I). In the
reaction with (R)-TOL-BINAP, the yield in 24 h increased to
73%, indicating that the reaction catalyzed by Cu(I)-(R)-TOL-
BINAP complex was not accelerated significantly. However,
the reactions with (R,R)-Ph-BPE and (R,RP)-TANIAPHOS
were much faster than the background reaction. Evidently, the
moderate enantioselectivity in the reaction with (R,R)-Ph-BPE
was largely due to its unsatisfactory performance in asymmetric
induction. Such an interesting ligand-accelerating effect was
the key to achieve high enantioselectivity.
Except for allyl bromide 2a, a series of other electrophiles
with an allyl group (8−15) was studied as described in Scheme
3. Obviously, allyl chloride 8 was a much less efficient
Scheme 3. Trials with Various Allyl Electrophiles (1H NMR
Yields Are Given)
electrophile under the present reaction conditions, as 3aa was
generated in 24% yield with 85% ee. As for allyl sulfonate, less
bulky allyl mesylate 9 was transformed to 3aa in 22% yield
with 60% ee. Bulky allyl tosylate (10) was an appropriate
substrate, and 3aa was produced in quantitative yield with 85%
ee. Unfortunately, the reaction with much bulkier allyl 2,4,6-
trimethylbenzenesulfonate (11) provided 3aa in slightly
decreased enantioselectivity (80% ee). Subsequently, allyl
methyl carbonate (12), allyl 2,2,2-trichloroethyl carbonate
(13), and allyl acetate (14) were tried, which were widely used
in palladium- or iridium-catalyzed asymmetric allylic alkyla-
tion.73 However, no reaction occurred. Furthermore, allyl
diethyl phosphate (15) frequently used in copper-catalyzed
asymmetric allylic alkylation73−78 was employed. Nevertheless,
only trace 3aa was observed. Therefore, we believe that the
present alkylation proceeds through a simple SN2 substitution
mechanism.
Moreover, the ligand-accelerating effect72 in this reaction
was roughly investigated with HPPhMes and allyl bromide as
shown in Figure 2. Without an external ligand, the copper(I)-
catalyzed nonenantioselective reaction proceeded slowly and
provided the product in 40% yield in 24 h. In this reaction,
With the assistance of the above control experiments and
our previous realization toward the [Cu]*-PPh2 species,43,44
a
plausible mechanism for the copper(I)-catalyzed asymmetric
alkylation is proposed in Scheme 4. As previously reported, the
coordination of HPAr1Ar2 to Cu(I)-(R,RP)-TANIAPHOS
complex (U) results in complex V and thus activates
HPAr1Ar2, which enables its facile deprotonation to afford
two diastereoisomers, Cu(I)-PAr1Ar2-(R,RP)-TANIAPHOS
complexes W and W′. Then in the presence of RBr, the
asymmetric SN2 substitution of W occurs fast to afford the
desired chiral RPAr1Ar2 as the product and the complex U is
regenerated as the catalyst. The asymmetric SN2 substitution of
W′ is very slow, which finally leads to high asymmetric
induction in the present reaction. It should be pointed out that
the bromide anion was finally transformed to CsBr in the
presence of stoichiometric Cs2CO3.
Figure 2. Ligand-accelerating effect in the catalytic asymmetric
allylation of HPPhMes with allyl bromide (1H NMR yields are given).
9917
J. Am. Chem. Soc. 2021, 143, 9912−9921