Rational electronic tuning of CBS catalyst for highly enantioselective borane reduction of trifluoroacetophenone

Article abstract of DOI:10.1039/c0cc03706k

α,α,α-Trifluoroacetophenone (2), which is susceptible to noncatalytic reduction by BH₃, could be reduced to chiral alcohol up to 90% ee by using electronically tuned-CBS catalyst (1) with BH₃. The enantioselectivities highly correlated with the differential orbital energies between 1-BH₃ adduct and 2, which were calculated by DFT method.

Full text of DOI:10.1039/c0cc03706k

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Rational electronic tuning of CBS catalyst for highly enantioselective
borane reduction of trifluoroacetophenonew
Toshinobu Korenaga,* Kenji Nomura, Kazutaka Onoue and Takashi Sakai*
Received 7th September 2010, Accepted 24th September 2010
DOI: 10.1039/c0cc03706k
a,a,a-Trifluoroacetophenone (2), which is susceptible to non-
catalytic reduction by BH3, could be reduced to chiral alcohol up
to 90% ee by using electronically tuned-CBS catalyst (1)
with BH3. The enantioselectivities highly correlated with the
differential orbital energies between 1–BH3 adduct and 2, which
were calculated by DFT method.
In asymmetric catalysis, the enantioselectivity of the reaction
product is generally controlled by steric repulsion between
asymmetric catalyst and substrate. The electronic effect of a
ligand substituent in the catalyst1 or substrate2 can play an
important role in this phenomenon. Among asymmetric
catalysts, the Corey–Bakshi–Shibata (CBS) catalyst (1) and
Fig. 1 Comparison of HOMO and LUMO between 2 and acetophenone.
other oxazaborolidine (OAB) catalysts are useful for achieving
the asymmetric reduction of ketones.3 In asymmetric BH3
reduction of 2 with high enantioselectivity by means of
an electron-tuned CBS catalyst. The electronic effect was
reductions using these catalysts, the enantioselectivity of the
investigated by Density Functional Theory (DFT) calculations.
product can be affected by introducing electron-donating
or -withdrawing substituents into the catalyst or substrate.4
To control the Lewis acidity of the CBS catalyst, we
modified its B-aryl group by introducing EWGs (Fig. 2).
Reduction of ketonic substrates bearing electron-withdrawing
The steric effect of functional groups on B-aryl seems to
groups (EWGs) tends to afford products with relatively low
enantioselectivity.5 In particular, OAB-catalyzed asymmetric
not greatly influence the transition state of the asymmetric
reduction.10 Therefore, introducing varying numbers of
BH3 reduction of a,a,a-trifluoroacetophenone (2) gives
fluoro-functional groups into B-aryl allows fine electronic
a-(trifluoromethyl)benzyl alcohol (3) with poor enantio-
selectivity;6,7 when the OAB catalyst is the popular CBS
tuning of the catalyst without significantly affecting the
transition state. We easily prepared CBS catalysts 1b–1h from
catalyst B-Me–(S)-CBS (1a), our preliminary study shows that
(S)-a,a-diphenyl-2-pyrrolidinemethanol and the corresponding
reaction gives 3 with only 2% ee. This low value is attributed
to the anomalous electronic nature of 28 rather than to its
undiscriminating nature in the transition state. The p*CQO
arylboronic acid by the conventional method, and prepared
1i by an alternative method.11 Asymmetric BH3 reduction of 2
was performed by catalysis with 10% B-R–(S)-CBS (1) at
orbital energy of 2 (LUMO = À2.46 eV) is much lower than
30 1C for 1 h (Table 1). In all cases, the product (S)-3 was
that of acetophenone (LUMO = À1.69 eV) and similar to that
obtained quantitatively. The enantioselectivity of 3 was greatly
of activated acetophenone formed by catalysis with CBS
(LUMO = À2.52 eV) (Fig. 1).9 These results indicate
improved by changing the substituent on the boron atom of 1
from Me to Ar (entry 1 vs. 2–9). The highest enantioselectivity
that 2 is a highly reactive substrate for BH3 reduction
without catalyst. Furthermore, the nO orbital energy of 2
was obtained with 1h (entry 8). Interestingly, the enantio-sense
of 3 was opposite to that of the product of the same reaction
catalyzed by B-nBu–(S)-CBS with catecholborane as the reducing
acetophenone (HOMO = À6.94 eV) (Fig. 1),9 indicating the
(HOMO À 2 = À7.85 eV) is much lower than that of
agent.7 In our catalytic system, asymmetric introduction to
difficulty of coordinating 2 to the CBS catalyst. As a result,
ketone 2 takes place by the same rule as for acetophenone.
noncatalytic reduction of 2 with BH3 occurs before coordination
We performed reactions using the catalyst (S)-1h under
to the catalyst, giving racemic 3. Therefore, we speculated that
varying reaction conditions of both CBS catalyst and
highly enantioselective BH3 reduction of 2 requires a CBS
catalyst with high Lewis acidity to form a catalyst–2 adduct.
In this communication, we attained the asymmetric BH3
Division of Chemistry and Biochemistry,
Graduate School of Natural Science and Technology,
Okayama University, Okayama 700-8530, Japan.
E-mail: korenaga@cc.okayama-u.ac.jp; Fax: +81 86-251-8092;
Tel: +81 86-251-8092
w Electronic supplementary information (ESI) available: Experimental
procedures, cartesian coordinates and schematic presentation of
orbitals for calculations. See DOI: 10.1039/c0cc03706k
Fig. 2 Electronic tuning of 1 for asymmetric reduction of 2.
c
8624 Chem. Commun., 2010, 46, 8624–8626
This journal is The Royal Society of Chemistry 2010

View Article Online
Table 1 Asymmetric BH3 reduction of 2a,b
Entry
Oxazaborolidine (R)
Taft’s s* of R
% ee of 3
1
2
3
4
5
6
7
8
9
1a (Me)
1b (4-Me–C6H4)
1c (Ph)
1d (4-F–C6H4)
1e (4-CF3–C6H4)
1f (3,5-F2–C6H3)
1g (3,4,5-F3–C6H2)
1h (2,4,6-F3–C6H2)
1i (C6F5)
0
0.46
0.6
0.63
0.96
1.01
1.04
1.08
1.5
1.7 (S)
76.2 (S)
79.3 (S)
79.2 (S)
71.8 (S)
80.2 (S)
80.8 (S)
84.9 (S)
54.6 (S)
a
The reaction was performed at 30 1C for 1 h. 1 M of ketone solution
was added at 20 mL minÀ1 by syringe pump. The conversions were
100% in all reactions, which were confirmed by H NMR.
b
Fig. 3 Typical catalytic cycle of CBS reduction.
1
Table 2 Optimization of asymmetric BH3 reduction of 2a,b
CBS 1h/
mol%
BH3ÁSMe2/
Temp./
1C
% ee
of 3
Entry
mol%
Solvent
1
2
3
4
5
6
7
8
9
10
10
10
10
10
10
10
10
10
8
60
60
60
60
100
120
100
100
100
100
THF
THF
THF
THF
THF
THF
Toluene
CH2Cl2
THF
0
30
40
50
40
40
40
40
40
40
5 (S)
85 (S)
87 (S)
86 (S)
90 (S)
88 (S)
90 (S)
88 (S)
90 (S)
87 (S)
5
THF
Fig. 4 Transition states 5h at the hydride transfer step.
1 M of ketone solution was added at 20 mL minÀ1 by syringe
pump. The conversions were 100% in all reactions, which were
a
b
Both aryl groups on boron in 5h oriented their aromatic plane
toward the ketonic substrate, suggesting that the steric effect
of the substituent on B-Ar had negligible influence on the
enantioselectivity of the product of reduction of 2. Therefore,
we focused again on electronic effects. Because ketone 2
coordinates to 1–BH3 adduct 4 before hydride transfer
(Fig. 3, step B), we calculated the p* orbital energy of boron
(p*B) in 4. The corresponding orbitals were found in the
LUMO + 4 of 4a and the LUMO of 4b–4i. DE values
were calculated by subtracting the nO orbital energy of 2
(HOMO À 2 = À7.85 eV) from the p*B orbital energy of 4,
indicating the coordination capability between 2 and 4. The
DE values correlated highly with the values of enantio-
selectivity of 3 (ln[(S)-product/(R)-product]) (Fig. 5), and the
correlation divided into two systems at catalyst 1h (4h). The
1
confirmed by H NMR.
BH3ÁSMe2 concentration, solvent, and temperature (Table 2).
In the temperature range of 30–50 1C, enantioselectivity
slightly changed (entries 2–4), but at lower reaction temperature
it decreased significantly (entry 1); this trend is commonly
observed in OAB-catalyzed asymmetric BH3 reductions of
ketones.12 At 1 equiv. of BH3ÁSMe2, enantioselectivity peaked
at 90% ee (entry 5 vs. the surrounding 3 and 6). Enantio-
selectivity was less affected by solvent (entries 5, 7, 8). Regarding
CBS concentration, catalyst loading could be reduced from
10% to 8% without loss of enantioselectivity, but further
reduction decreased enantioselectivity (entry 9 vs. 10). Thus,
the choice of optimal conditions permitted highly enantio-
selective CBS-catalyzed BH3 reduction of 2.
Next, we considered the influence of B-Ar on enantio-
selectivity. Although an OAB catalyst with EWGs tended to
give products with high enantioselectivity, the data showed no
correlation with Taft’s s* values for the aryl group,13 causing us
to suspect that the steric effect of the fluoro-functional groups of
aryl affects enantioselectivity. The mechanism of OAB-catalyzed
asymmetric BH3 reduction is shown in Fig. 3.3c Asymmetric
induction to 2 occurs at the transition state of hydride transfer
(Fig. 3, step C). In our case, Ph was recognized as larger group
than CF3 (Ph 4 CF3), similar to the case of reduction of
acetophenone (Ph 4 CH3)3c and contrary to the case of
reduction of 2 with catecholborane (Ph o CF3) as mentioned
above.3c,7,14 Calculations of the transition states 5h by
two-layered ONIOM (MP2/6-31G(d,p):B3LYP/6-31G(d))15
demonstrated a preference for the formation of (S)-3 (Fig. 4).
Fig. 5 Correlation between enantioselectivities of 3 and DE.
Chem. Commun., 2010, 46, 8624–8626 8625
c
This journal is The Royal Society of Chemistry 2010

View Article Online
In conclusion, we accomplished highly enantioselective
CBS-catalyzed asymmetric BH3 reduction of 2 by electronic
tuning of the catalyst. The enantioselectivity of reaction
product (S)-3 is influenced by the coordination capability of
4. We are now able to predict roughly the enantioselectivities
of the products of reduction of 6 from calculated DE values,
although at present these predictions can be applied exclusively
to trifluoroacetophenone analogues. Such electronic control of
CBS catalysis should be effective for other highly reactive
substrates if the relationship between catalyst and substrate is
well designed.
Fig. 6 DE values 6–4h for prediction of % ee of BH3 reduction.
Notes and references
high correlation indicated that the enantioselectivity of the
product formed by reduction of 2 was mainly influenced by
coordination between 2 and catalyst. When the substituent
was B-Me (4a), coordination between 2 and 4a was very weak
due to the high p*B orbital energy of 4a (that is, large DE),
indicating that noncatalytic reduction of 2 with BH3 occurred
dominantly (Table 1, entry 1). When the substituent was
changed to B-Ar, coordination between 2 and 4 drastically
improved (Table 1, entries 2–9), owing to the lower p*B orbital
energy (that is, smaller DE) of 4 relative to that of 4a.
Although large numbers of fluoro-functional groups tended
to decrease DE, the order turned back in the case of 4e, 4f, 4g
and 4h. This disorder is because of the mesomeric effect of
fluorine’s lone pair to the boron atom, particularly at the 2, 4,
and 6 positions. The enantioselectivity of product 3 increased
with decreasing DE values for 4 until 4h (OAB 1h) (Fig. 5).
However, further decrease in DE resulted in a loss of enantio-
selectivity for 3, suggesting the generation of an inhibition
process. Although one likely reason for the inhibition is
dimerization of OAB,4b dimer suspect signals could not be
observed in 19F NMR analyses of the catalysts 1e, 1f, and 1h in
THF at 30 1C. Therefore, we hypothesized that sluggish
elimination of the product owing to the strong Lewis acidity
of the catalyst inhibited catalyst reproduction (Fig. 3, reaction
step D) in the cases of 1e, 1f, 1g, and 1i. Although the reasons
for this phenomenon are not clear, a subtle inhibition process
would decrease the enantioselectivity of (S)-3 considerably
because of the ease of noncatalytic reduction of 2.
1 Review: S. P. Flanagan and P. J. Guiry, J. Organomet. Chem.,
2006, 691, 2125.
2 Recent example: (a) W. Kashikura, J. Itoh, K. Mori and
T. Akiyama, Chem.–Asian J., 2010, 5, 470; (b) L. Yang, Q. Zhu,
S. Guo, B. Qian, C. Xia and H. Huang, Chem.–Eur. J., 2010, 16,
1638; (c) S. O’Neill, S. O’Keeffe, F. Harrington and A. R. Maguire,
Synlett, 2009, 2312.
3 Reviews: (a)
G. Paul, in Name Reactions for Functional
Group Transformations, ed. J. J. Li and E. J. Corey, Wiley, 2007,
pp. 2–21; (b) B. T. Cho, Tetrahedron, 2006, 62, 7621; (c) E. J. Corey
and C. J. Helal, Angew. Chem., Int. Ed., 1998, 37, 1986;
(d) V. A. Glushkov and A. G. Tolstikov, Russ. Chem. Rev.
(Engl. Transl.), 2004, 73, 581.
4 (a) W. Xu, H. Guo, J. Zhang, Q. Zhu and X. Hu, J. Mol. Catal. A:
Chem., 2009, 300, 25; (b) J. Du, Z. Li, D.-M. Du and J. Xu, J. Mol.
Catal. A: Chem., 2008, 284, 40; (c) H. Liu and J. Xu, J. Mol. Catal. A:
Chem., 2006, 244, 68.
5 (a) E. J. Corey and C. J. Helal, Tetrahedron Lett., 1995, 36, 9153;
(b) J. Xu, T. Wei and Q. Zhang, J. Org. Chem., 2004, 69, 6860;
(c) E. Fuglseth, E. Sundby, P. Bruheim and B. H. Hoff,
Tetrahedron: Asymmetry, 2008, 19, 1941.
6 (a) S. Goushi, K. Funabiki, M. Ohta, K. Hatano and M. Matsui,
Tetrahedron, 2007, 63, 4061; (b) M. Zhao and F. Xiao, Huaxue
Yanjiu Yu Yingyong, 1999, 11, 282.
7 When catecholborane was used instead of BH3 as a reducing agent,
(R)-3 was obtained with 90% ee in the presence of B-nBu-(S)-CBS
at À78 1C: E. J. Corey, J. O. Link and R. K. Bakshi, Tetrahedron
Lett., 1992, 33, 7107.
8 (a) N. Asao, T. Asano and Y. Yamamoto, Angew. Chem., Int. Ed.,
2001, 40, 3206; (b) A. Vargas, T. Burgi, M. von Arx, R. Hess and
¨
A. Baiker, J. Catal., 2002, 209, 489.
9 All calculations were performed at the B3LYP/6-311G(2d,p)//
B3LYP/6-311G(2d,p) by using Gaussian 03 (see ESIw).
10 In ref. 4b, for example, the asymmetric reduction of
p-nitroacetophenone by using OAB bearing several B-Ar groups
gave similar enantioselectivity at 110 1C in toluene.
We obtained the DE values between each of the other
trifluoroacetophenone analogues 6a–6c and 4h by DFT
calculations. Each DE value between 6a and 4h and between
6b and 4h was similar to that between 2 and 4h within the
difference of Æ0.1 eV; however, the difference of those from
the value between 6c and 4h was 0.33 eV (Fig. 6). These values
indicated that, compared to the product of reduction of 2, the
products of reduction of 6a and 6b should have similar
enantioselectivity and the product of reduction of 6c should
have lower enantioselectivity.16 In fact, the products of BH3
reduction of 6a and 6b had similar enantioselectivities (86%
and 90% ee, respectively) to the product of reduction of 2
(90% ee) when the catalyst (S)-1h was used under the same
condition, and the product of reduction of 6c had low
enantioselectivity (54% ee) owing to weak coordination
between 6c and 4h, as indicated by the DE value (Fig. 6).
11 T. Korenaga, F. Kobayashi, K. Nomura, S. Nagao and T. Sakai,
J. Fluorine Chem., 2007, 128, 1153.
12 J. Xu, T. Wei and Q. Zhang, J. Org. Chem., 2003, 68, 10146 and
references therein.
13 T. Korenaga, K. Kadowaki, T. Ema and T. Sakai, J. Org. Chem.,
2004, 69, 7340.
14 One likely reason for the different results between our catalytic
system and CBS with catecholborane is due to asymmetric
reduction with another catecholborane in the presence of
CBS–catecholborane adduct: Y.-Y. Yeung, R.-J. Chein and
E. J. Corey, J. Am. Chem. Soc., 2007, 129, 10346.
15 Definition of two layers was shown in ESIw (Fig. S2). Harmonic
vibrational frequencies were computed for all stationary points in
order to characterize them as saddlepoints. The energies were
corrected using the zero-point energy. IRC calculations were
performed to confirm the formation of product.
16 The correlations in Fig. 5 are not applied to the results of reduction
of 6 as it is, because EWG in phenyl group accelerates
non-catalytic BH3 reduction, particularly, in 6c. See ref. 8b.
c
8626 Chem. Commun., 2010, 46, 8624–8626
This journal is The Royal Society of Chemistry 2010