generated in situ from metal salts and N,N’-dioxide L1 were
evaluated. Pleasingly, Sc(OTf)3 produced 3aa in 99% yield
with 84% enantiomeric excess (ee) and 99:1 diastereomeric
ratio (d.r.) (Table 1, entry 4 vs. entries 1–3). Based upon this
result, various ligands of N,N’-dioxides were screened to be
tron-donating substituents at the meta- or para-position of
the aryl group allowed the maintenance of the high enantio-
selectivity (up to 99% ee) and complete syn-diastereoselec-
tivity (99:1 d.r.), although the yields of the products were
dependent on the electronic effect of the substituents
(Table 2, entries 2–12). Generally, 3,3’-dihalostilbene oxides
reacted more slowly than 4,4’-dihalostilbene oxides, and
longer reaction times were needed to achieve satisfactory
yields (Table 2, entries 2 and 3 vs. entries 7 and 8). Compa-
ratively, epoxides 1i and 1j with electron-donating substitu-
ents at the meta-position showed higher reactivity than ep-
oxides 1g and 1h with electron-withdrawing groups
(Table 2, entries 9 and 10 vs. entries 7 and 8).
For 4,4’-ditrifluoromethylstilbene oxide 1e, a moderate
yield with 97% ee was obtained in the presence of 5 mol%
of L6–ScACHTNUGTRNE(NUG OTf)3 catalyst with a longer reaction time
(Table 2, entry 5). However, the current catalyst system was
not efficient for reactions with cycloalkene oxides. After a
slight modification of the ligand structure, ligand L7 coordi-
AHCTUNGTRENNUNG
attached to ScACHTUNGTRENNUNG(OTf)3. Modification of the chiral backbone
and the amide of the ligand gave an interesting insight into
the structural elements that were crucial for the enantiose-
lectivity of the process. The l-ramipril derivative L3 was su-
perior to l-proline-derived L1 and l-pipecolic acid derived
L2 in both stereoselectivity and reactivity (Table 1, entry 6
vs. entries 4 and 5). The steric hindrance of the amide subu-
nits of l-ramipril-derived N,N’-dioxides had a crucial influ-
ence on both the yield and the stereoselectivity of the reac-
tion. The use of ligand L4 containing 2,6-dimethyl aniline
gave the product in only 41% yield, 94:6 d.r., and 16% ee,
whereas ligand L6 containing 3,5-dimethyl aniline provided
excellent enantioselectivity (98% ee) and diastereoselectiv-
ity (99:1 d.r.) as well as high reactivity (Table 1, entry 9 vs.
entry 7). Thus, the combination of L6 and Sc
G
nated with ScACTHUNGTERNNU(G OTf)3 gave a moderate ee value for cyclohex-
adopted as the optimal catalyst system. Pleasingly, when the
catalyst loading was reduced from 10 to 5 mol%, the enan-
tioselectivity and yield of the desired product, 3aa, was
maintained. Carrying out the reaction with an even lower
catalyst loading (1 mol%) in the presence of 3 ꢀ molecular
sieves maintained the striking results (Table 1, entry 12 vs.
entry 11). Other conditions, such as the solvent and reaction
temperature, were also investigated, but no superior results
were obtained.
ene oxide 1m (Table 2, entry 13).
Given the remarkable performance of the present catalyst
system, its further applicability was also examined with
more pyrazole derivatives (Table 3). To our delight, the cor-
responding syn-b-pyrazole-substituted alcohols were ob-
tained in good to excellent yields and high diastereo- and
enantioselectivities at 5 mol% catalyst loading (Table 3, en-
tries 1–10). Dimethyl-substituted 1H-pyrazole 2c provided
the corresponding products with higher enantioselectivity
than 1H-pyrazole 2b (Table 3, entries 2, 7, and 9 vs. entry 1).
A 4-nitro substituent on 1H-pyrazole had a slight influence
on the stereoselectivity but reduced the reactivity sharply
(Table 3, entries 1 and 3). 4-Bromo-3,5-dimethyl-1H-pyra-
zole 2 f gave comparable stereoselectivities with either 4,4’-
difluorostilbene oxide 1b or 3,3’-dimethoxylstilbene oxide
1j (Table 3, entries 5, 8, and 10). Remarkably, 1H-indazole
gave the desired 2-indazolyl-1,2-diphenylethanol with up to
99% ee (Table 3, entry 6). Additionally, almost complete
conversions of epoxides 1b and 1j could be achieved with
pyrazole 2c in 92% and 97% ee, respectively (Table 3, en-
tries 7 and 9).
With the optimized conditions in hand, various cis-stil-
bene oxide derivatives were examined. As summarized in
Table 2, meso-epoxides with electron-withdrawing or elec-
Table 2. Scope of epoxides for the reaction[a]
Entry
1
R
t
Yield
[%][b]
d.r.[c]
ee
[h]
[%][d]
1
2
1a
1b
1c
1d
1e
1 f
1g
1h
1i
Ph
4-FC6H4
12
12
12
18
96
27
72
48
14
14
27
18
24
99
81
85
99
47
89
90
99
82
89
69
99
57
99:1
99:1
99:1
99:1
99:1
99:1
99:1
99:1
99:1
99:1
99:1
99:1
99:1
98
99
96
95
97
99
93
97
98
98
97
98
45
It is worth pointing out that trans-stilbene oxide reacted
sluggishly in current reaction conditions (Table 4, entry 1).
When cis- and trans-stilbene oxides were mixed in a ratio of
1:1 or 3:1 as the starting materials, the syn-b-pyrazole substi-
tuted alcohol was observed in high diastereo- and enantiose-
lectivity. These results indicated that the catalytic system is
capable of distinguishing between cis- and trans-stilbene
oxides. It provided an efficient way to get the syn-product
even if a mixture of cis- and trans-stilbene oxides was used.
The absolute configuration of the b-pyrazole-substituted
alcohol 3aa was unambiguously determined to be (1R,2R)
by single-crystal X-ray diffraction analysis of the corre-
3[e]
4
4-ClC6H4
4-BrC6H4
4-CF3C6H4
4-PhC6H4
3-FC6H4
3-ClC6H4
3-MeC6H4
3-MeOC6H4
3-PhOC6H4
2-naphthyl
-(CH2)4-
5[e]
6
7
8[e]
9
10
11
12
13[f]
1j
1k
1l
1m
[a] Unless otherwise noted, reactions were carried out with 1 (0.1 mmol),
1.5 equiv of 2a and 3 ꢀ MS (20 mg) in CH2Cl2 (0.3 mL) at 358C.
[b] Yield of isolated product. [c] Determined by chiral HPLC analysis
and NMR spectroscopy. [d] Determined by chiral HPLC analysis.
sponding
methanesulfonyl-protected
derivative
4
(Scheme 2).[16]
A mechanism to explain the large difference in the reac-
[e] 5 mol% of catalyst was used. [f] 10 mol% of L7–ScACTHNUGTRENUNG(OTf)3 catalyst
was used.
tivity of the ring-opening reaction in the case of cis- and
3474
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 3473 – 3477