Chiral C2-boron-bis(oxazolines) in asymmetric catalysis - A theoretical study of the catalyzed enantioselective reduction of ketones promoted by catecholborane

Article abstract of DOI:10.1002/ejoc.200600133

C₂-Symmetrical boron complexes, prepared by the reactions of 2,2′-methylenebis(oxazolines) (BOXs) with catecholborane (CATBH), can be used as catalysts (5-10 mol-%) in the enantioselective reduction of prochiral ketones (ee 72-86 %), giving the desired alcohols in satisfactory yields. We have theoretically investigated the mechanism of the reduction of chloroacetophenone at the DFT level and the computational results have provided a complete mechanistic picture, which explains the stereochemical outcome of the reaction. The B-BOXate complex binds both the reducing agent CATBH and the carbonyl compound, activating the former as a hydride donor and enhancing the electrophilicity of the latter. Moreover, the structure of two boron-BOX (BOXate) complexes has been confirmed by means of X-ray diffraction techniques. Wiley-VCH Verlag GmbH & Co. KGaA, 2006.

Full text of DOI:10.1002/ejoc.200600133

FULL PAPER
DOI: 10.1002/ejoc.200600133
Chiral C -Boron-Bis(oxazolines) in Asymmetric Catalysis – A Theoretical
2
Study of the Catalyzed Enantioselective Reduction of Ketones Promoted by
Catecholborane
[
a]
[a]
[a]
[a]
Marco Bandini, Andrea Bottoni, Pier Giorgio Cozzi, Gian Pietro Miscione,*
[a]
[a]
[a]
Magda Monari, Rossana Pierciaccante, and Achille Umani-Ronchi
Keywords: Ab initio calculations / Reaction mechanism / Asymmetric catalysis / Reduction / Boron-bis(oxazolines)
C
2
-Symmetrical boron complexes, prepared by the reactions
BOXate complex binds both the reducing agent CATBH and
the carbonyl compound, activating the former as a hydride
donor and enhancing the electrophilicity of the latter. More-
over, the structure of two boron–BOX (BOXate) complexes
has been confirmed by means of X-ray diffraction tech-
niques.
of 2,2Ј-methylenebis(oxazolines) (BOXs) with catecholbor-
ane (CATBH), can be used as catalysts (5–10 mol-%) in the
enantioselective reduction of prochiral ketones (ee 72–86%),
giving the desired alcohols in satisfactory yields. We have
theoretically investigated the mechanism of the reduction of
chloroacetophenone at the DFT level and the computational
results have provided a complete mechanistic picture, which
explains the stereochemical outcome of the reaction. The B–
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
Introduction
Chiral bis(oxazoline) (BOX) species are considered “priv-
ileged ligands”[ and have been used in many catalyzed
stereo-controlled reactions.[ Chiral Lewis acids are usually
1]
2]
obtained from bis(oxazolines) and an appropriate metal Scheme 1.
with free coordination sites. The role of the metal is to acti-
The main BOX skeleton is identical to that of β-diketim-
vate electrophiles towards the attack of nucleophiles or
[
3]
inate (Figure 1), which has been known for a long time and
was initially employed in the spectroscopic studies of coor-
masked enolates. C -Bis(oxazolines) bearing a methylene
2
bridge between the two dihydrooxazole rings can be easily
deprotonated and several research groups have been able to
functionalize this type of ligand (Scheme 1) by introducing
carbon chains with different functions and to anchor the
[
11]
dination compounds. In addition to many applications of
[12]
monoanionic β-diketiminates in coordination chemistry,
[
13]
Smith III and co-workers have recently reported an inter-
esting study of the structure and properties of β-diketimin-
ate boron complexes. The similarity between BOX and β-
diketiminate framework, the structural features of β-diket-
iminate boron compounds and their stability, has prompted
us to consider the possibility of designing boron–BOX com-
plexes to be used in catalytic enantioselective reactions. As
a result of our research, we have discovered that achiral and
chiral BOX compounds react with catecholborane affording
a stable boron-bis(oxazolinate) adduct and that chiral bis-
[
4]
[5]
BOX species to solid supports or to soluble polymers.
Furthermore, the organometallic chemistry of metallated
[
6]
bis(oxazolines) has been described in several papers.
Hoffmann, for instance, has reported the preparation of
magnesium–BOX compounds[ (Scheme 1, M = Mg), while
7]
[
8]
Zn–BOX complexes have been synthesized by Nakamura
[
9]
and Singh. Quite recently our group has described the
preparation of Ti–BOX complexes and their use in catalytic
enantioselective reduction of ketones. In this case the whole
catalytic cycle has also been theoretically investigated at the
(
oxazolines) are able, in the presence of CATBH, to reduce
[
14,15]
[
10]
pro-chiral ketones with good enantioselectivity.
DFT level.
In this paper, in order to elucidate the mechanism of this
reaction and to understand the reasons for the aforemen-
tioned stereoselectivity, we have carried out a theoretical
investigation of the whole catalytic cycle at the Density
Functional Theory (DFT) level. To this purpose, we have
used a model system consisting of one chiral boron–BOX
complex, one molecule emulating the CATBH and a chlo-
[
a] Dipartimento di Chimica “G. Ciamician”, Università di
Bologna,
Via Selmi 2, 40126 Bologna, Italy
Fax: +39-051-2099456
E-mail: gianpietro.miscione@unibo.it
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
4596
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4596–4608

2
Chiral C -Boron-Bis(oxazolines) in Asymmetric Catalysis
FULL PAPER
terized[17] by 11B NMR, the characteristic signals for the
central atom were observed at δ = 11.6 and 11.7 ppm,
respectively.
The X-ray diffraction study carried out on 4 and 6 is
consistent with the NMR interpretation. The ORTEP plots
with the two structures are depicted in Figure 2 and Fig-
ure 3 whereas the relative values obtained for selected bond
lengths and angles are collated in Table 1. The two mole-
Figure 1. BOX and β-diketiminate. The two species have similar
skeletons.
roacetophenone molecule as the substrate. Although several
important computational studies have been carried out on
similar catalytic systems,[ as far as we know, no theoreti-
cal investigations on this specific subject can be found in
the literature. We also report details and information about
the structure of the chiral boron–BOX complex obtained
from X-ray diffraction techniques.
cules conform to an idealized C -symmetry with the boron
2
atom exhibiting a slightly distorted tetrahedral geometry
upon coordination of a chelating bis(oxazolinate) ligand
and a chelating catecholate unit. The BOX and catecholate
ligands lie on almost perpendicular planes [the dihedral an-
gle between the N1–B1–N2 and O3–B1–O4 planes is 89.69
16]
(8)° and 87.7 (1)° in 4 and 6, respectively]. The boron atom
is coplanar with the catecholate ring in 4 but “tipped” out
of the plane by ca. 0.15 Å in compound 6, presumably be-
cause of crystal packing forces. The six-membered rings,
formed by coordination of the two nitrogen atoms of the
bis(oxazolinate) ligand to the boron, show some deviation
Results and Discussion
A. Experimental Results
To prepare coordinatively saturated boron-bis(oxazolin- from planarity. The negative charge of the bis(oxazolinate)
ate) complexes, the reaction of achiral BOX 1 (Scheme 2) ligand is delocalized, as clearly indicated by the C–C and
with catecholborane was carried out in CH Cl for 24 h at C–N bond interactions [C4–C5 1.370 Å and C4–C3
2
2
room temperature. A white solid, stable to moisture and 1.375(2) Å; N1–C3 1.316 Å and N2–C5 1.1.319(2) Å in 4;
1
oxygen, was isolated after removal of the solvent. The H, C11–C10 1.382 Å and C11–C27 1.371(3) Å; N1–C10 1.332
1
3
11
C and B NMR spectroscopic investigations supported Å and N2–C27 1.319(2) Å in 6]. Interestingly, the N–B–N
the hypothesis for the formation of a single complex derived bite angle [105.8° and 105.2(1)° in 4 and 6, respectively] is
from the reaction of CATBH with BOX 1 (boron–BOXate). rather large when compared to the bite angles in related
The reaction is quite general and the synthesis of other bo- tetrahedral metal complexes containing the chelating bis-
ron–BOXate species was also readily accomplished from (oxazolinate) ligand. For example, the N–Metal–N bite an-
chiral bis(oxazolines). In this way, the boron–BOXate com- gle of the metallacycle is 93.13(7)° in (BOX-Me )AlMe and
2
2
[
18]
plexes 4–6 (from 1, 2 and 3, respectively) were prepared 97.35(8)° in (BOX-Me )AlCl2 and as narrow as 85.2(1)°
2
[
19]
and isolated in good yields, monitoring the deprotonation in the [(4S)iBuBox]Lu[CH(TMS) ] complex. This rather
2
2
process at the methylene bridge by NMR spectroscopy. The wide range of values demonstrates the high flexibility of the
diagnostic NMR signals for the boron–BOXate complexes bis(oxazolinate) ligand.
derived from the achiral bis(oxazoline) 1 are the singlet for
Achiral and chiral boron complexes 4–6 were able to re-
the proton bridge at δ = 4.45 ppm and the corresponding duce pro-chiral ketones with good enantioselectivity, em-
1
3
C NMR signal at δ = 57.8 ppm, both characteristic of ploying CATBH as the reducing agent, as shown in
1
1
[14]
methine carbons. On the other hand, the B NMR signal Scheme 3. The corresponding results (yields and enantio-
at δ = 8.87 ppm represents a typical value for a tetrahedral meric excesses) are reported in Table 2.
coordination at the boron atom. Recently, borates derived
On the basis of the experimental evidence, it turned out
by the reaction between the reaction of (1R,2S)-norephed- to be very difficult to propose a plausible reaction mecha-
rine and diphenylprolinol with catecholborane were charac- nism: the saturated boron coordination spheres in these B–
Scheme 2.
Eur. J. Org. Chem. 2006, 4596–4608
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4597

G. P. Miscione et al.
FULL PAPER
Scheme 3.
Table 2. Yields and the enantiomeric excesses obtained with various
ketones.
Figure 2. ORTEP drawing for compound 4. The hydrogen atoms
are omitted for clarity.
Entry
Ketone
Yield (%)
ee (%)
1
2
3
4
5
6
7
8
9
7a
7b
7c
7d
7e
7f
7g
7h
7i
80
65
57
85
54
78
72
65
67
55
42
76 (R)
76 (R)
74(R)
84 (S)
86 (S)
72 (R)
74 (R)
74 (R)
81 (S)
57 (R)
58 (R)
1
1
0
1
7j
7k
BOXate complexes do not imply trivial mechanistic expla-
nations. Thus, in order to obtain detailed information about
the process of the reaction, with particular regard to the
stereochemical aspects and to understand the role played
by the B–BOXate, we have examined the mechanism of the
process at a theoretical level, using a DFT approach. The
results are reported in the following section.
Figure 3. ORTEP drawing for compound 6. The hydrogen atoms
are omitted for clarity.
Table 1. Selected bond lengths [Å] and angles (°) for compounds 4
and 6 obtained in the X-ray diffraction study. The values in the B. Computational Results: The Choice of the Model
brackets refer to the error in the measurement. It can be noted
that the data are quite accurate. For comparison, the geometrical
parameters obtained computationally for the isolated BOX-ate
The model system, represented in Figure 4, was chosen
as it closely resembles the real molecules involved in the
reaction. Also, the two basis sets used in the computations
(which is more similar to 6, see Figure 5) have been included.
Compound 4
Compound 6 Computational are indicated in the Figure: the DZVP basis for the atoms
results
directly involved in the reaction and the 3-21G* basis set
B1–N1
B1–N2
B1–O3
B1–O4
N1–C3
N2–C5
C3–C4
C4–C5
1.541(2)
1.548(2)
1.481(2)
1.479(2)
1.316(2)
1.319(2)
1.375(2)
1.370(2)
–
1.525(2)
1.544(2)
1.472(2)
1.484(2)
–
1.548
1.551
1.482
1.478
for the remaining atoms (see the Computational Details
section). This system includes a chloroacetophenone mole-
cule (substrate, corresponding to entry 7d in Scheme 3), one
molecule emulating the catecholborane (CATBH) and a
chiral catalytic B–BOXate species, which mimics compound
6 (obtained from 3) of Scheme 2. In the real complex, which
is characterized by an inversion center, a benzene ring is
bound to each of the two five-membered cycles, pointing
above and under the BOX plane. To avoid heavy computa-
tions, we have reduced the size of the B–BOXate and we
have considered only one of these cycles. Furthermore, we
have replaced the benzene ring fused with this cycle with a
C=C double bond and we have used the same approxi-
mation to simplify the catecholborane molecule. Thus, a
penta-atomic unsaturated ring, approximately orthogonal
to the BOX plane, characterizes the resulting B–BOXate
complex.
–
–
–
N1–C10
N2–C27
1.332(2)
1.319(2)
1.382(3)
1.371(3)
104.8(1)
105.2(1)
–
115.6(2)
111.8(2)
112.7(2)
113.9(2)
108.6(2)
1.335
1.334
1.391
1.390
104.9
104.1
–
–
–
C10–C11
C11–C27
O3–B1–O4
N1–B1–N2
C5–C4–C3
C10–C11–C27
N1–B1–O3
N2–B1–O3
N1–B1–O4
N2–B1–O4
104.5(1)
105.8(1)
115.7(2)
–
111.3(1)
111.9(1)
111.8(1)
111.6(1)
115.7
111.7
111.8
112.8
111.7
4598
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4596–4608

2
Chiral C -Boron-Bis(oxazolines) in Asymmetric Catalysis
FULL PAPER
C. Computational Results: The Reaction Mechanism
In this section we discuss in details the singlet potential
energy surface associated with the various steps of the reac-
tion. The corresponding mechanistic scheme is illustrated
in Scheme 4. A representation of the energy profile is given
in Figure 6, while the structures of the various critical
points located along it are shown in Figure 7, Figure 8, Fig-
ure 9, Figure 10 and Figure 11. In Figure 6, the energy val-
ues obtained with single-point calculations with the DZVP
basis set on all atoms are also reported in square brackets.
The reaction sequence reported at the bottom of Figure 6
corresponds to the mechanistic steps of Scheme 4.
Figure 4. The model system used to investigate the mechanism of
the catalyzed ketone reduction. The different basis sets used in the
various molecular regions are indicated.
A preliminary encounter complex (long-range complex
M1 in Figure 7), involving the B–BOXate, the CATBH
molecule and the substrate, is found in the first reaction
step. Formation of I does not require any activation barrier.
In this complex, which is 7.0 kcalmol–1 more stable than
the reactants (asymptotic limit), the carbonyl bond of the
substrate interacts with the B–BOXate system through a hy-
drogen bond with the bridging CH group of the central B–
BOXate ring (O5–H distance 2.476 Å). A second interac-
tion that contributes to stabilize the complex involves the
catecholborane boron atom B2 and the B–BOXate oxygen
O4, the B2–O4 distance being 2.803 Å.
The next step of the catalytic cycle (IǞII) is the ketone
reduction and represents the rate-determining step of the
process. Several transformations take place to allow the hy-
dride transfer. The O4–B1 bond is replaced by the O4–B2
bond, which determines the opening of the catecholborane
ring of the B–BOXate system. This makes the boron atom
This B–BOXate complex, although being different from
the real system, is chiral and can be successfully used to
obtain information on the stereochemical outcome of the
reaction. This is possible in a theoretical study, by making
the reaction take place in the desired position: in the present
case this has been achieved by placing the reducing agent
and the substrate under the B–BOXate plane, as shown in
Figure 5, in such a way as to let them interact with the
unsaturated cycle. Of course, under the real experimental
conditions, the reaction can take place on both sides of the
molecular plane. Thus, by forcing the reaction to occur on
one side of the plane only, where the chirality elements are
placed, we can use the present system as a reliable model
of the real one, without adding the second five-membered
ring.
(electronically saturated and sterically hindered in M1)
available for the interaction with the ketone acting as a
Lewis base. In this new arrangement, the boron atom B1
adopts a trigonal coordination state. Furthermore, the con-
jugation of the O4 lone pair with the LUMO of B2 shifts
the electron density along the B1–O3–C12–C13–O4–B2–H
chain and has the effect of making B1 more electrophilic
and the hydride more nucleophilic. In spite of extensive se-
arch it was not possible to locate any critical point preced-
ing the reduction and corresponding to the previously de-
scribed transformations (breaking of the O4–B1 and forma-
tion of O4–B2 bonds). Our investigation has shown that
this region of the potential energy surface is very flat in-
deed.
During the ketone reduction a hydride ion is transferred
from the boron atom B2 to the prochiral carbon C14 lead-
ing to a new chiral center. It is evident that, because of the
presence of the five-membered unsaturated cycle in the B–
BOXate catalytic complex, the probabilities for the attack
Figure 5. A schematic representation of the direction of attack of
the substrate and CATBH with respect to the B–BOXate molecular
plane. The values of selected geometrical parameters (Å and °) are
reported.
It is interesting to compare the values of some selected of the hydride ion on the two faces of the carbonyl group
geometrical parameters computed at the DFT level for the are not the same. The reduction can occur along two dif-
model B–BOXate complex to the corresponding values ob- ferent channels involving two different diastereotopic tran-
tained in the X-ray diffraction study for compounds 4 and sition states: TS1(S) and TS1(R) (Figure 8 and Figure 9,
6
(see Table 1). The agreement between the two sets of val- respectively). TS1(S) describes the transfer of the hydride
ues is very satisfactory. This result indicates that the compu- onto the Re face of the ketone. In this structure, the new
tational level used here is adequate to provide a reliable de- forming C14–H bond is 2.071 Å, while the breaking B2–H
scription of the system.
bond is 1.256 Å. This transition state is 22.7 kcalmol–1
Eur. J. Org. Chem. 2006, 4596–4608
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4599

G. P. Miscione et al.
FULL PAPER
Scheme 4. The reaction mechanism.
above the reactants and leads to configuration S at the new TS1(R)]. This interaction certainly makes the C14 carbon
chiral center. TS1(R) (transfer of the hydride onto the Si more electrophilic in TS1(S) than in TS1(R) and makes the
face; C14–H 1.542 Å and B2–H 1.335 Å), leading to config- nucleophilic attack of the hydride ion easier.
uration R, is significantly higher in energy (by
The key factors that prevent the ketone carbonyl group
–
1
4.3 kcalmol ). Since this is the key step for determining the from getting close enough to the B1 atom in TS1(R) are
stereochemical outcome of the reaction and is also clearly the steric interactions between the ketone substituents and
irreversible, the formation of the stereoisomer with configu- the B–BOXate system and the repulsion between the O5
ration S at C14 is highly favored in agreement with the ex- and O4 oxygen atoms. This repulsion is certainly lower in
perimental evidence. The lower energy of TS1(S) with re- TS1(S) than in TS1(R) since the O5–O4 distance is 2.940 Å
spect to TS1(R) is partly due to the stabilizing B1–O5 inter- in the former case and 2.893 Å in the latter. In TS1(S), the
action (acid–base interaction), which is considerably carbonyl group of the ketone can get closer to B1 and in-
stronger in the former case than in the latter [the B1–O5 crease the stabilizing acid–base interaction (without making
distance is 1.901 Å in TS1(S) and becomes 2.392 Å in the O5–O4 interaction too strong), by decreasing the B1–
4600
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4596–4608

2
Chiral C -Boron-Bis(oxazolines) in Asymmetric Catalysis
FULL PAPER
Figure 6. Total energy profile for the whole catalytic cycle corresponding to the reduction of chloroacetophenone. The energy values
include the ZPE corrections. In square brackets, energy values obtained with single-point calculations with DZVP basis on all atoms.
These distances are 2.537 and 3.028 Å, respectively. It is
interesting to point out that a rotation around the C–C14
bond could not help to reduce the repulsion, since a de-
crease of one H···H distance after rotation, is necessarily
accompanied by an increase of the other H···H distance.
Things are different in TS1(S), where the planar phenyl
ring, by rotating around the C(phenyl)–C(carbonyl) bond,
can easily find the best orientation to minimize the repul-
sion that involves one H···H pair of atoms (see Figure 8)
only. In the end, the position of the ketone (and thus the
magnitude of the stabilizing O5–B1 interaction and the dif-
ferent stability of the two transition states) is the result of
a complex interplay between the O5–O4 repulsion and the
repulsion involving the ketone substituents and the unsatu-
rated five-membered B–BOXate cycle.
Figure 7. Schematic representation of the structure of the critical
The free-energy difference between TS1(R) and TS1(S) is
–1
point M1. The total energy value (kcalmol ) relative to reactants
B–BOXate + CATBH + chloroacetophenone) and some selected
geometrical parameters (Å) are reported. The energy values include
the ZPE corrections. In square brackets, energy values obtained excess above 99%, and thus, does not exactly predict the
–1
4
.4 kcalmol and thus, almost identical to the total energy
(
difference. This value would correspond to an enantiomeric
with single-point calculations with DZVP basis on all atoms.
experimental outcome (corresponding to an enantiomeric
excess of about 80%). However, it provides the correct indi-
cation of the stereochemical preference of the reaction, in
agreement with the experimental evidence.
It can be argued that the relative position of the two reac-
tants (catecholborane including the reducing hydride and
ketone) may be interchanged with respect to the five-mem-
bered cycle of the BOX system (i.e the ketone may be placed
far from it, on the right side, according to the orientation
of the figures). The transition states for this alternative ar-
rangement have been calculated. Actually, these points
O5–C14 angle that becomes 147.2°. More precisely, as the
ketone approaches the catalyst, it can slightly bend with
respect to an ideal straight line of attack (B1–O5–C14 180°)
to minimize the repulsive steric interactions of its substitu-
ent with the catalyst. A similar distortion of the B1, O5,
C14 system is not possible in TS1(R) (here the B1–O5–C14
angle is larger i.e. 154.8°) since in this case it would firstly
cause an enhancement of the steric repulsion between the
unsaturated ring of the B–BOXate and the CH Cl group.
2
[TS1(S)Ј and TS1(R)Ј reported in the Supporting Infor-
In Figure 9 we have reported the distances between the
two hydrogen atoms of the methylene group and the closest
hydrogen of the B–BOXate unsaturated ring in TS1(R).
mation] are almost degenerate to TS1(S) and TS1(R),
respectively, indicating that this different approach is ener-
Eur. J. Org. Chem. 2006, 4596–4608
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4601

G. P. Miscione et al.
FULL PAPER
Figure 8. Schematic representation of the structure of the critical
Figure 9. Schematic representation of the structure of the critical
points TS1(R) and M2(R). The total energy value (kcalmol ) rela-
tive to reactants (B–BOXate + CATBH + chloroacetophenone)
and some selected geometrical parameters (Å) are reported. The
energy values include the ZPE corrections. In square brackets, en-
ergy values obtained with single-point calculations with DZVP ba-
sis on all atoms.
–
1
–1
points TS1(S) and M2(S). The total energy value (kcalmol ) rela-
tive to reactants (B–BOXate + CATBH + chloroacetophenone)
and some selected geometrical parameters (Å) are reported. The
energy values include the ZPE corrections. In square brackets, en-
ergy values obtained with single-point calculations with DZVP ba-
sis on all atoms.
getically equivalent to that originally considered. Thus, even M2(S) [the O5–C(ring) distance is 3.290 and 3.136 Å in
in this case a preference for the path leading to the stereo- M2(S) and M2(R), respectively]. Furthermore, in M2(R)
isomer with S configuration at C14 is still observed [the the O3–C12–C13–O4–B2 chain is oriented toward the ring
total energy difference and the free-energy difference be- and this causes an increase of the hydrogen C12–hydro-
tween TS1(S)Ј and TS1(R)Ј are 2.8 kcalmol–1 and gen(ring) interaction. These structural features have the ef-
–
1
3
.2 kcalmol , respectively].
fect of making M2(S) more stable than M2(R).
The two transition states TS1(S) and TS1(R) lead to two
Since the results of the calculations indicate that M2(S)
diastereomeric reduced species [M2(S) and M2(R), respec- is the most likely reaction product, in agreement with the
tively], both corresponding to an oxaborane bound to the experimental outcomes (the stereoisomer with S configura-
[
14]
B–BOXate system and differing only for the configuration tion at C14 is formed in 84% ee ), from now on, only the
at C14. The two complexes (see Figure 8 and Figure 9), remaining path related to this isomer will be discussed. This
where B1 has a tetrahedral coordination, feature the “un- path, corresponding to the closure of the catalytic cycle (re-
folding” of the two chains bound to B1, one above and the generation of the catalyst and release of the oxaborane), is
other under the N–B1–N plane. The groups bound to C14 accomplished in two steps: IIǞIII [corresponding to the
(
new chiral center) are oriented in such a way that the M2(S)ǞTS2ǞM3 transformation] and IIIǞIV (corre-
phenyl substituent approximately points in the opposite di- sponding to the M3ǞTS3ǞM4 transformation).
rection with respect to the five-membered ring. Both species In TS2 (see Figure 10), the O3–C12–C13–O4–B2 chain
are very stable: M2(S) and M2(R) are 35.9 and is folded down to form the new B2–O5 bond. The transition
–
1
3
3.9 kcalmol lower than the asymptotic limit, respectively. state has a strong product-like character with the boron-
This indicates that the transfer of the hydride is a highly oxygen bond almost completely formed (the B2–O5 dis-
exothermic process. Because of the different configuration tance is 1.811 Å). This structural change, which entails a
at the chiral center C14, the oxygen atom O5 in M2(R) is change of the B2 coordination state (from tri-coordinated
closer to the unsaturated five-membered cycle than in to tetra-coordinated), results in a lengthening of the B1–O5
4602
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4596–4608

2
Chiral C -Boron-Bis(oxazolines) in Asymmetric Catalysis
FULL PAPER
Figure 10. Schematic representation of the structure of the critical
Figure 11. Schematic representation of the structure of the critical
points TS3 and M4. The total energy value (kcalmol ) relative to
reactants (B–BOXate + CATBH + chloroacetophenone) and some
selected geometrical parameters (Å) are reported. The energy val-
ues include the ZPE corrections. In square brackets, energy values
obtained with single-point calculations with DZVP basis on all
atoms.
–1
–1
points TS2 and M3. The total energy value (kcalmol ) relative to
reactants (B–BOXate + CATBH + chloroacetophenone) and some
selected geometrical parameters (Å) are reported. The energy val-
ues include the ZPE corrections. In square brackets, energy values
obtained with single-point calculations with DZVP basis on all
atoms.
and B2–O4 bonds that become 1.547 and 1.411 Å, respec- being 32.9 kcalmol–1 lower than reactants. In M4 the B1–
tively. The barrier required by this transformation is O3–C13–C12–O4 cycle of the regenerated catalyst is com-
–
1
1
1.1 kcalmol and probably, is mainly caused by the B1– pleted with the two B1–O3 and B1–O4 bonds almost equiv-
O5 and B2–O4 lengthening and by the steric effect of the alent (1.483 and 1.472 Å, respectively). The product mole-
penta-atomic cycle bonded to O4. The resulting intermedi- cule is still slightly interacting with the catalyst, the B2–O4
ate M3 (Figure 10) is characterized by a seven-member cy- distance being 3.754 Å.
cle that includes the O5 oxygen bonded to the chiral C14
Further calculations have been carried on a more realistic
–
1
carbon. M3 is 10.7 kcalmol higher than M2(S) and model system that includes the phenyl group attached to
–
1
2
5.2 kcalmol below the asymptotic limit. This energy in- the five-membered cycle. In these calculations we have con-
crease is certainly caused by the proximity of the CATBH sidered either the two transition states where the ketone is
ring to the unsaturated penta-atomic cycle. close to the five-membered ring and the phenyl group [TS1-
To close the catalytic cycle the breaking of two boron- phen(S) and TS1-phen(R), see Figure 12] and those corre-
oxygen bonds (B2–O4 and B1–O5) and the formation of sponding to the opposite arrangement [TS1-phen(S)Ј and
a new boron-oxygen bond (B1–O4) must take place. This TS1-phen(R)Ј, see Supporting Information]. The results
transformation can be accomplished in one step (step have confirmed the following important aspects: (i) the al-
IIIǞIV): release of the oxaborane and regeneration of the ready observed arrangement of the two reactants in TS1(S)
–
1
catalyst by overcoming a barrier of 22.6 kcalmol . In the and TS1(R), i.e. the ketone near to the five-membered ring,
corresponding transition state TS3 (see Figure 11), the two is favored [TS1-phen(S) is 1.8 kcalmol–1 more stable than
–
1
breaking bonds B2–O4 and B1–O5 are 2.613 and 3.796 Å, TS1-phen(S)Ј, whereas TS1-phen(R) is 4.4 kcalmol below
respectively. The B1–O3–C13–C12–O4 chain is simulta- TS1-phen(R)Ј]. The corresponding free-energy values are
neously folded down to form the new bond B1–O4, which 1.2 kcalmol–1 and 2.6 kcalmol , respectively: (ii) a signifi-
is still far from being completed (B1–O4 2.961 Å). The reac- cant preference for the path leading to the S stereoisomer
tion is highly exothermic, the final complex M4 (Figure 11) is once again observed. The total energy and free-energy
–1
Eur. J. Org. Chem. 2006, 4596–4608
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4603

G. P. Miscione et al.
FULL PAPER
differences between TS1-phen(S) and TS1-phen(R) are to those discussed in the previous section. The stabilizing
–
1
–1
1
.1 kcalmol and 0.9 kcalmol , respectively. It is interest- B1–O5 interaction is stronger in the former case than in the
ing to underline that this free-energy difference would latter (the B1–O5 distance is 1.643 Å in TS(R) prop and
correspond to an enantiomeric excess of about 80%, in ex- 1.664 Å in TS(S) prop), while the O5–O4 repulsion has the
cellent agreement with the experimental outcome.
opposite trend as indicated by the O5–O4 distance (2.940
and 2.879 Å in TS(R) prop and TS(S) prop, respectively).
Once again the planar phenyl ring plays a key role in al-
lowing a shorter B1–O5 distance in TS(R) prop without in-
creasing the O5–O4 repulsion too much and the repulsion
between the phenyl ring itself and the unsaturated five-
membered cycle of the B–BOXate moiety.
Figure 12. Schematic representation of the structure of the critical
points TS1-phen(S) and TS1-phen(R), which include the phenyl
ring in the model system.
D. Computational Results: the Case of Propyl Phenyl
Ketone as a Substrate
To check the reliability of our model in predicting the Figure 13. Schematic representation of the structure of the critical
points TS(R) and TS(S) in the case of propyl phenyl ketone as
substrate. Some selected geometrical parameters (Å) are reported.
stereoselectivity of the reaction, we have re-computed the
two transition states TS1(R) and TS1(S) using a model sys-
tem where the substrate is a propyl phenyl ketone molecule
(entry 7b in Scheme 3). In this case an excess of the stereo-
isomer with configuration R at C14 was experimentally ob-
served (ee of 76%).
E. Computational Results: The Basis Set Effect
The two diastereotopic transition states (TS(R) prop and
To validate the results obtained with the mixed DZVP/
TS(S) prop) are schematically represented in Figure 13. 3-21G* basis set, we have carried out single-point computa-
TS(R) prop, i.e. the transition state leading to the product tions on the optimized structures using the DZVP basis set
with configuration R at C14, is 4.9 kcalmol–1 more stable on all atoms. These energy values are reported in square
than TS(S) prop leading to the opposite configuration. brackets in Figure 6. The mechanistic scenario, which stems
Thus, the prediction of our model, indicating that for the from these data is identical to that previously discussed. For
product an excess of the stereoisomer with configuration R instance, the barriers for the two transition states TS1(S)
at C14, is again in agreement with the experimental obser- and TS1(R) of the rate-determining step change from 22.7
vation. The key factors that determine the lower activation and 27.0 kcalmol–1 to 25.1 and 31.1 kcalmol , respectively.
energy of TS(R) prop with respect to TS(S) prop are similar Thus, again the pathway leading to the product with config-
–1
4604
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4596–4608

2
Chiral C -Boron-Bis(oxazolines) in Asymmetric Catalysis
FULL PAPER
uration S at C14 is favored. Also, the intermediate M2(S) the reaction, is due to complex interplay between the stabi-
and the product M4 are now 36.0 and 32.0 kcalmol–1 below lizing B1–O5 interaction (acid–base interaction) involving
the asymptotic limit, respectively (35.9 and 32.9 kcalmol–1 the B–BOXate boron and the carbonyl oxygen, the repul-
with the less accurate basis). Finally, the barrier for TS3 sion between the ketone carbonyl oxygen and one CATBH
–
1
varies from 22.6 to 23.3 kcalmol .
oxygen and the repulsion between the side-unsaturated B–
BOXate ring and the ketone substituents. This last interac-
3
tion is particularly important. When the sp carbon (the
carbon in an alpha position to the carbonyl in the chloro-
Conclusions
In this paper the reaction of achiral and chiral bis(oxa- acetophenone and propyl phenyl ketone) interacts with the
zolines) (BOX) with catecholborane (CATBH) have been B–BOXate ring, the system cannot easily arrange to de-
investigated. This reaction provides boron–BOXate com- crease the repulsive interactions and to increase the boron–
plexes that can be used as catalysts in the enantioselective oxygen stabilizing effect. This becomes much easier when
reduction of ketones. The X-ray analysis of two boron– the configuration at the new chiral centre changes and a
3
BOXate complexes is in agreement with NMR spectro- phenyl group replaces the sp carbon. In this case, a rota-
scopic data and has shown that the chelating bis(oxazoline) tion of the phenyl ring around the carbon–carbon bond can
ligand forms a six-membered heterocycle when it coordi- decrease the repulsion and allow a more stabilizing boron–
nates the CATBH boron atom. Furthermore, the BOX and oxygen interaction.
the catecholborane skeletons lie on two orthogonal planes
(v) Two possible arrangements of the two reactants (cate-
with the boron characterized by a tetrahedral coordination cholborane and ketone) with respect to the chiral element
geometry. It has been demonstrated that it is possible to on the catalyst have been considered. In the preferred ar-
achieve good enantiomeric excesses in the reduction of aro- rangement the ketone is close to the chiral element. This
matic ketones (ee = 72–86%) using a commercially available minimizes the steric repulsions and the strong acid–base in-
bis(oxazoline) in the presence of catecholborane.
teraction between the B–BOXate boron and the carbonyl
A DFT investigation has been carried out to shed light oxygen.
on the mechanistic details of the catalytic reduction of aro-
(vi) The computation of the free-energy difference be-
matic ketones. The most interesting results can be summa- tween the two transition states TS1(S) and TS1(R)
–1
rized as follows:
(4.4 kcalmol ) corresponds to an enantiomeric excess
(i) The structural features of the B–BOXate complex ob- above 99%, thus much larger than the experimental value
tained at the computational level are in good agreement (84%). However, when a more realistic model system in-
with the experimental outcomes provided by the X-ray cluding a benzene ring, is used, the free-energy difference
analysis.
between the two diastereotopic transition states becomes
–1
(ii) The B–BOXate catalyst simultaneously plays different 0.9 kcalmol . This value predicts an enantiomeric excess
roles. Firstly, it binds the reducing agent CATBH (interac- of about 80%, in very good agreement with the experimen-
tion between the B–BOXate oxygen and the CATBH tal result.
boron) and causes the weakening of the bond between the
(vii) The role played by the ketone phenyl ring also ex-
boron and the hydrogen that must be transferred as a hy- plains why the reaction is much easier with aromatic than
dride ion. Secondly, as expected, the boron of the B–BOX- aliphatic ketones. Our analysis indicates that, in the case of
ate unit interacts with the carbonyl oxygen and makes the aliphatic ketones, it is much more difficult to minimize the
prochiral carbon atom more electrophilic. Finally, the re- repulsions and to increase the stabilizing boron–oxygen in-
pulsions between the side group of the B–BOXate unit (the teraction and this should make the energy barrier of the
side unsaturated five-membered ring in our model) and the rate-determining step larger.
ketone substituents determine the enantioselectivity of the
reaction.
(
iii) The transfer of the hydride ion from the boron atom
Experimental Section and Computational Details
of CATBH to the prochiral carbonyl is the rate-determining
step of the catalytic reaction. This transfer can occur on two
different pathways involving two diastereotopic transition General Remarks: H NMR spectra were recorded using a Varian
states, corresponding to the attack of the hydride either 200 MHz or a Varian 300 MHz spectrometer. Chemical shifts are
onto the Re or onto the Si face of the ketone. The transition reported in ppm relative to tetramethylsilane with the solvent reso-
A) Experimental Section
1
nance as the internal standard (deuteriochloroform: δ = 7.27 ppm).
Data are reported as follows: chemical shifts, multiplicity (s = sing-
state TS1(S) leading to configuration S at the new chiral
centre (attack onto the Re face) in the case of chloroaceto-
–
1
let, d = doublet, t = triplet, q = quartet, br = broad, m = multiplet),
phenone has an activation barrier 4.3 kcalmol lower than
the barrier required by the attack onto the Si face [transi-
tion state TS1(R)]. Thus, our model correctly predicts an
excess of the stereoisomer with configuration S at the new
chiral centre in agreement with the experimental evidence.
1
3
coupling constants (Hz). C NMR spectra were recorded using a
Varian 50 MHz or a Varian 75 MHz spectrometer with complete
proton decoupling. Chemical shifts are reported in ppm relative to
tetramethylsilane with the solvent as the internal standard (deuter-
iochloroform: δ = 77.0 ppm). Mass spectra were performed at an
(iv) The energy difference between the two diastereotopic ionizing voltage of 70 eV. Chromatographic purification was car-
transition states, which determines the enantioselectivity of ried out using 240–400 mesh silica gel. Analytical gas chromatog-
Eur. J. Org. Chem. 2006, 4596–4608
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4605

G. P. Miscione et al.
FULL PAPER
raphy (GC) was performed using a Hewlett–Packard HP 6890 gas
chromatograph with a flame ionization detector and split mode
capillary injection system, using a Crosslinked 5% PH ME Silox-
ane (30 m) column or a Megadex5 chiral (25 m) column. Analytical
high performance liquid chromatograph (HPLC) was performed
on a HP 1090 liquid chromatograph equipped with a variable wave-
1.01, CH
mal) t
phenylethanol: 15.77 min.
2
Cl
2
). GC analysis: column temperature, 115 °C (isother-
r
r
(major) (R)-1-phenylethanol: 12.61 min. t (minor) (S)-1-
(
(
9
R)-1-Phenylpropanol (7b): ee = 76%, 44 mg, yield 65%. [α]
D
= +28
= +21.5 (χ =
.3, diethyl ether). GC analysis: column temperature, 110 °C (iso-
thermal) t (major) (R)-1-phenylpropanol: 25.3 min t (minor) (S)-
-phenylpropanol: 25.7 min.
χ = 0.8, CHCl
3 D
). Ref.[ (R)-1-phenylpropanol [α]
21]
length UV detector (deuterium lamp 190–600 nm), using a Daicel
r
r
TM
Chiralcel
OD column (0.46 cm I.D.× 25 cm) (Daicel Inc.).
1
HPLC grade 2-propanol and hexane were used as the eluting sol-
vents. Elemental analyses were carried out using a EACE 1110
CHNOS analyzer. All the reactions were carried out in a nitrogen
atmosphere in flame-dried glassware using standard inert tech-
niques for introducing reagents and solvents. All ketones were dis-
tilled prior to use. All the other commercially obtained reagents
were used as received. Neat commercially available catecholborane
(R)-3-Methyl-1-phenylbutan-1-ol (7c): ee = 74%, 47 mg, yield 57%.
D 3 D
= +23 (χ = 1.2, CHCl
). Ref.[ (R)-1-phenylbutan-1-ol [α]
22]
[α]
= –39.9 (χ = 0.05, heptane). GC analysis: column temperature,
120 °C (isothermal) t (minor) (S)-3-methyl-1-phenylbutan-1-ol:
(major) (R)-3-methyl-1-phenylbutan-1-ol: 14.1 min.
r
13.8 min. t
r
(
S)-2-Chloro-1-phenylethanol (7d): ee = 84%, 66 mg, yield 85%.
(
Aldrich) stored 5 °C was used for the catalytic reductions. Anhy-
drous CH Cl , THF, Et O, toluene and CH CN were purchased
from the Fluka Co.
[23]
[α]
D
3 D
= +42 (χ = 2.4, CHCl ). Ref. [α] = +49.6 (χ = 2.92, cyclo-
2
2
2
3
hexane). GC analysis: column temperature, 120 °C 20 min, 120 to
30 °C 2 °C/min (program). t (major) (S)-1-(2Ј-chloro-1-phenyl)-
(minor) (R)-1-(2Ј-chloro-1-phenyl)ethanol:
2
r
Preparation of B-Boxate 4: In a oven dried flask anhydrous CH
2
Cl
2
ethanol: 25.1 min t
r
(2 mL) was introduced, then achiral BOX 1 ligand (0.05 mmol) and
27.6 min.
catecholborane(0.75 mmol) were added. The resulting solution was
stirred at 25 °C for 3 h then the solvent was evaporated under vac-
uum. The white solid obtained was washed with anhydrous diethyl
ether and collected by filtration (13 mg, yield 86%). 1H NMR
(
S)-2-Bromo-1-phenylethanol (7e): ee = 86%, 54 mg, yield 54%.
[24]
[
α]
CH
minor) (R)-1-(2Ј-bromo-1-phenyl)ethanol: 18.5 min t
-(2Ј-bromo-1-phenyl)ethanol: 19.1 min.
D
= +42 (χ = 1.0, CHCl
3
). Ref.
D
[α] = +45.2 (χ = 1.01,
2
Cl
2
). GC analysis: column temperature, 110 °C (isothermal) t
r
(
1
r
(major) (S)-
(
1
1
=
C
3
200 MHz, CDCl ): δ = 6.70 (s, 4 H), 4.40 (s, 1 H), 4.16 (s, 4 H),
13
.30 (s, 12 H) ppm. C NMR (75 MHz, CDCl
3
): δ = 150.7, 119.1,
): δ
09.3, 91.6, 76.4, 62.8, 57.6, 26.3 ppm. 11B (96 MHz, CDCl
(R)-1-Naphthylethanol (7f): ee = 72%, 67 mg, yield 78%. [α]
+38.5 (χ = 1.06, CHCl ). Ref.
=
=
D
3
[25]
–
1
(S)-1-naphthylethanol [α]D
8.87. IR (Nujol): ν˜ = 2725, 1558, 1233, 1057, 910, 897 cm .
(296.17): calcd. C 66.22, H 6.45, N 3.29; found: C
3
–
(
(
54.8 (χ = 3.34, CHCl
isothermal) t (minor) (S)-1-naphthylethanol: 27.9 min t
R)-1-naphthylethanol: 29.15 min.
3
). GC analysis: column temperature, 150 °C
17
2 2
H21BN O
r
r
(major)
6
6.19, H 6.44, N 3.33.
B-Boxate 5: (25 mg, yield 86%). M. p. 168–171 °C. 1H NMR
300 MHz, CDCl ): δ = 6.99 (m, 24 H), 6.01 (d, 2 H, J = 10.5 Hz),
.67(d, 2 H, J = 10.5 Hz), 3.91 (s, 1 H) ppm. 13C NMR (75 MHz,
(
5
3
(R)-1-(4Ј-Chlorophenyl)ethanol (7g): ee = 74%, 56 mg, yield 72%.
[26]
[α]
anol [α]
ture, t (major) (R)-1-(4Ј-chlorophenyl)ethanol: 19.9 min t
(S)-1-(4Ј-chlorophenyl)ethanol: 25.1 min.
D
= +38.6 (χ = 1.01, CHCl
3
). Ref. (S)-1-(4Ј-chlorophenyl)eth-
O). GC analysis: column tempera-
(minor)
CDCl
3
): δ = 169.5, 150.0, 143.9, 135.5, 133.7, 128.0, 128.9, 127.8,
D
= –49.6 (χ = 1.8, Et
2
1
6
27.5, 127.4, 126.1, 120.9, 129.1, 118.4, 115.6, 109.4, 108.2, 87.9,
5.2, 57.6 ppm. [α] = –165 (χ = 0.69, CHCl ) ppm. C37
r
r
D
3
2 4
H29BN O
(
4
576.22): calcd. C77.09, H 5.07, N 4.86; found: C 77.30, H 5.17, N
.93.
B-Boxate 6: (17 mg, yield 75%). M. p. 158–160 °C. 1H NMR nol [α]
300 MHz, C ): δ = 7.94 (d, J = 7.5 Hz, 1 H), 7.60–7.40 (m, 6
115 °C (isothermal) t
H), 7.08–705 (m, 1 H), 6.82 (d, J = 7.5 Hz, 1 H), 6.70 (m, 3 H),
(
R)-1-(4Ј-Methylphenyl)ethanol (7h): ee = 74%, 44 mg, yield 65%.
[29]
[α]
D
= +28 (χ = 1.0, CHCl
= +35 (χ = 1, CHCl
(major) (R)-1-(4Ј-methylphenyl)ethanol:
(minor) (S)-1-(4Ј-methylphenyl)ethanol: 15.9 min.
3
). Ref. (R)-1-(4Ј-methylphenyl)etha-
D
3
). GC analysis: column temperature,
(
6
D
6
r
11.8 min t
r
5
.29 (s, 1 H), 5.05 (d, 1 H, J = 6.9 Hz), 4.52 (t, J = 6.6 Hz, 1 H),
(
R)-1-(4Ј-Fluorophenyl)ethanol (7i): ee = 81%, 47 mg, yield 67%.
1
3
2
.86 (d, J = 9.9 Hz, 1 H), 2.52 (dd, J = 6.6, 9.9 Hz, 1 H) ppm.
): δ = 169.3, 152.3, 144.5, 139.7, 139.5, 129.1,
26.1, 125.6, 121.1, 120.0, 115.7, 109.70, 87.3, 66.8, 57.8, 37.5 ppm.
α] = –330 (χ = 0.65, CHCl ). C27 (448.16): calcd. C
2.34, H 4.72, N 6.25; found: C 72.47, H 4.81, N 6.28.
C
[27]
[α]
D
= +34 (χ = 1.16, CHCl
3
) Ref.
(S)-1-(4Ј-fluorophenyl)etha-
6 6
NMR (75 MHz, C D
1
[
7
nol [α]
ture, 120 °C (isothermal) t
.2 min t (minor) (S)-1-(4Ј-fluorophenyl)ethanol: 5.5 min.
D
= –37.2 (χ = 0.93, MeOH). GC analysis: column tempera-
r
(major) (R)-1-(4Ј-fluorophenyl)ethanol:
D
3
2 4
H21BN O
4
r
(
R)-1-(2Ј-Bromophenyl)ethanol (7j): ee = 57%, 55 mg, yield 55%
General Procedure for the Catalytic Reduction: In an oven dried
flask anhydrous CH Cl (2 mL) was introduced, then BOX ligand
0.05 mmol) and catecholborane (1 mmol, neat) were added. The
).) Ref.[ [α]
28]
= +38.5 (χ = 1.8,
[
α]
CHCl
30 °C 2 °C/min (program). t
nol: 15.2 min t (minor) (R)-1-(4Ј-bromophenyl)ethanol: 16.5 min.
D
= +27 (χ = 1.08, CHCl
). GC analysis: column temperature, 120 °C 20 min, 120 to
(major) (S)-1-(4Ј-bromophenyl)etha-
3 D
2
2
3
(
2
r
resulting solution was stirred at 25 °C for 3 h and then was cooled
to 0 °C. The freshly distilled ketone (0.5 mol) was added, then the
clear pale yellow solution was stirred at the same temperature for
r
(S)-2-Methoxy-1-phenylethanol (7k): ee = 58%, 32 mg, yield 42%.
[α]
perature, 110 °C (isothermal) t
thanol: 28.0 min
30.0 min.
). Ref.[ GC analysis: column tem-
(major) (S)-2-methoxy-1-phenyle-
(minor) (R)-2-methoxy-1-phenylethanol:
29]
48 h. The reaction mixture was diluted with Et
2
O (5 mL), then
D 3
= +23 (χ = 1.08, CHCl
NaOH 2  (5 mL) was carefully added (attention, gas evolution).
The two phases were stirred 20–30 min and during this period the
aqueous phase turned dark brown. The phases were separated and
r
t
r
the aqueous phase extracted with Et
phase was collected and dried with Na
chromatography (hexane: Et O, 9:1–7:3).
2
O (2×3 mL). The organic
X-ray Crystallographic Study of 4 and 6: The diffraction experi-
ments for 4 and 6 were carried out at room temperature using a
Bruker AXS SMART 2000 CCD based diffractometer using graph-
2
SO , then purified by
4
2
(
R)-1-Phenylethanol (7a): ee = 76%, 49 mg, yield 80%. [α]
D
= +36.4
). Ref.[ (R)-1-phenylethanol [α]
= +45.2 (χ =
3 D
ite-monochromated Mo-K
data were measured over full diffraction spheres using 0.3° wide
α
radiation (λ = 0.71073 Å). Intensity
20]
(χ = 2.4, CHCl
4606
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4596–4608

Chiral C
2
-Boron-Bis(oxazolines) in Asymmetric Catalysis
FULL PAPER
ω
scans, crystal-to-detector distance 5.0 cm. The software
S. W. Tregay, J. Am. Chem. Soc. 1999, 121, 686–699; l) D. A.
Evans, K. A. Scheidt, J. N. Johnston, M. C. Willis, J. Am.
Chem. Soc. 2001, 123, 4480–4491; m) D. A. Evans, D. M.
Barnes, J. S. Johnson, T. Lectka, P. von Matt, S. J. Miller, J. A.
Murry, R. D. Norcross, E. A. Shaughnessy, K. R. Campos, J.
Am. Chem. Soc. 1999, 121, 7582–7594; n) D. A. Evans, S. J.
Miller, T. Lectka, P. von Matt, J. Am. Chem. Soc. 1999, 121,
[
30a]
SMART
was used for collecting frames of data, indexing reflec-
tions and determination of lattice parameters. The collected frames
were then processed for integration by software SAINT
[30a]
and an
[
30b]
empirical absorption correction was applied with SADABS.
The structures were solved by direct methods (SIR 97)[
30c]
and sub-
sequent Fourier syntheses and refined by full-matrix least-squares
calculations on F (SHELXTL)
parameters to the non-hydrogen atoms. The aromatic hydrogen
atoms were placed in calculated positions and refined with ideal-
ized geometry [C(sp )–H 0.93 Å] whereas the other H atoms were
located in the Fourier map and refined isotropically. Crystal data
and details of the data collection for 4 and 6 are reported in
Table 1.
7559–7573; o) D. A. Evans, T. Rovis, M. C. Kozlowski, J. S.
2
[31]
attributing anisotropic thermal
Tedrow, J. Am. Chem. Soc. 1999, 121, 1994–1995.
[4] M. I. Burguete, J. M. Fraile, J. I. Garcia, E. Garcia-Verdugo,
C. I. Herrerias, J. Luis, S. V. A. Mayoral, J. Org. Chem. 2001,
66, 8893–8901 and references cited therein.
2
[
5] R. Annunziata, M. Benaglia, M. Cinquini, F. Cozzi, M. Pitillo,
J. Org. Chem. 2001, 66, 3160–3166.
[
6] a) H. W. Görlitzer, M. Spiegler, R. Arwander, J. Chem. Soc.,
Dalton Trans. 1999, 4287; b) R. Schumacher, F. Dammast,
R. H. U. Reißig, Chem. Eur. J. 1997, 3, 614–619; c) J. M.
Brown, P. J. Guiry, D. W. Price, M. B. Hursthouse, K. Karulov,
Tetrahedron: Asymmetry 1994, 5, 561–564; d) D. Müller, G.
Umbricht, B. Weber, A. Pfaltz, Helv. Chim. Acta 1991, 74, 232–
CCDC-163357 and -278924 contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
240.
B) Computational Details: All the DFT computations reported here
have been performed with the Gaussian 98[ series of programs
using the non-local hybrid Becke’s three-parameter exchange func-
tional (denoted as B3LYP in the Gaussian formalism) and locally
[7] V. Schultze, R. W. Hoffmann, Chem. Eur. J. 1999, 5, 337–344.
[8] M. Nakamura, A. Hirai, M. Sogi, E. Nakamura, J. Am. Chem.
Soc. 1998, 120, 5846–5847.
[9] R. P. Singh, Bull. Soc. Chim. Fr. 1997, 134, 765–768.
[10] a) M. Bandini, P. G. Cozzi, L. Negro, A. Umani-Ronchi,
Chem. Commun. 1999, 39–40; b) M. Bandini, F. Bernardi, A.
Bottoni, P. G. Cozzi, G. P. Miscione, A. Umani-Ronchi, Eur. J.
Org. Chem. 2003, 2972–2984.
32]
dense basis sets (LDBS).[33] According to the LDBS approach, in
all geometry optimizations the system has been partitioned into
two different regions, which were assigned basis sets of different
accuracy. One region contains the atoms directly involved in the
reaction or in the formation of hydrogen bonds. In this case we
have used the DZVP basis, which is a Local Spin Density (LSD)-
[
[
11] J. E. Parks, R. H. Holm, Inorg. Chem. 1968, 7, 1408–1416.
12] C. E. Radzewich, M. P. Coles, R. F. J. Jordan, J. Am. Chem.
Soc. 1998, 120, 9384–9385.
[34]
optimized basis set of double-zeta quality. This basis, which in-
cludes polarization functions, is suitable to describe weak hydrogen
interactions such as those occurring in the system investigated here.
The other region includes all the remaining atoms. For these atoms,
contained within the contour line in Figure 4, the 3-21G* basis
[
13] B. Qian, S. W. Baek, M. R. Smith III, Polyhedron 1999, 18,
2405–2414.
[14] M. Bandini, P. G. Cozzi, M. Monari, R. Pierciaccante, S. Selva,
A. Umani-Ronchi, Chem. Commun. 2001, 1318–1319.
[15] For enantioselective reduction of ketones with boranes see: a)
J. Xu, T. Wei, Q. Zhang, J. Org. Chem. 2004, 69, 6860–6866;
b) J. Xu, T. Wei, Q. Zhang, J. Org. Chem. 2003, 68, 10146–
[
32]
set has been employed. To obtain more accurate energy values,
single-point calculations have been performed using the DZVP ba-
sis set on all atoms. The geometry of the various critical points on
the reaction surface has been fully optimized with the gradient
method available in Gaussian 98. A computation of the harmonic
vibrational frequencies has been carried out to determine the na-
ture of each critical point and the ZPE correction has been added
to all the energy values.
1
0151; c) N. J. Gilmore, S. Jones, Tetrahedron: Asymmetry
2
003, 14, 2115–2118; d) S. M. Nettles, K. Matos, E. R. Burkh-
ardt, D. R. Rouda, J. A. Corella, J. Org. Chem. 2002, 67, 2970–
2
976; e) H. S. Wilkinson, G. J. Tanoury, S. A. Wald, C. H. Sen-
anayake, Org. Process Res. Dev. 2002, 6, 146–148; f) C.
Puigjaner, A. Vidal-Ferran, A. Moyano, M. A. Pericas, A. Ri-
era, J. Org. Chem. 1999, 64, 7902–7911. For review: see, g) E. J.
Corey, C. J. Helal, Angew. Chem. Int. Ed. 1998, 37, 1986–2012;
h) V. A. Glushkov, A. G. Tolstikov, Russian Chem. Rev. 2004,
Supporting Information (see also the footnote on the first page of
this article): Calculated transition states for TS1(S)Ј and TS1(R)Ј.
73, 581–608; i) M. Wills, J. Hannedouche, Curr. Opin. Drug
Discovery Dev. 2002, 5, 881–891; j) J. M. Brunel, G. Buono,
Top. Curr. Chem. 2002, 220 (New Aspects, in: Phosphorus
Chemistry I), 79–105; k) M. Wills, M. Gamble, M. Palmer, A.
Smith, J. Studley, J. Kenny, J. Mol. Catal. A 1999, 146, 139–
[
[
[
1] T. P. Yoon, E. N. Jacobsen, Science 2003, 299, 1691–1693.
2] A. Pfaltz, Synlett 1999, 835–843.
3] For recent examples of C
as catalysts, see: a) J. S. Johnson, D. A. Evans, Acc. Chem. Res.
000, 33, 325–335; b) N. Gathergood, W. Zhuang, K. A.
2
-bis(oxazoline) Lewis acid complexes
148.
[
16] a) G. Alagona, C. Ghio, M. Persico, S. Tomasi, J. Am. Chem.
Soc. 2003, 125, 10027–10039 and references cited therein; b) W.
Harb, M. F. Ruiz-López, F. Coutrot, C. Grison, P. Coutrot,
J. Am. Chem. Soc. 2003, 126, 6996–7008 and references cited
therein.
17] V. Stepanenko, M. Ortiz-Marciales, W. Correa, M. De Jesùs, S.
Espinosa, L. Ortiz, Tetrahedron: Asymmetry 2006, 17, 112–115.
18] S. Dagorne, S. Bellemin-Laponnaz, R. Welter, Organometallics
2
Jørgensen, J. Am. Chem. Soc. 2000, 122, 12517–12522; c) K. B.
Jensen, J. Thorhauge, R. G. Hazell, K. A. Jørgensen, Angew.
Chem. Int. Ed. 2001, 40, 160–163; d) H. Audrain, K. A.
Jørgensen, J. Am. Chem. Soc. 2000, 122, 11543–11544; e) K.
Juhl, N. Gathergood, K. A. Jørgensen, Angew. Chem. Int. Ed.
[
[
[
2
001, 40, 2995–2997; f) K. R. Knudsen, T. Risgaard, N. Nishi-
waki, K. V. Gothelf, K. A. Jørgensen, J. Am. Chem. Soc. 2001,
23, 5843–5844; g) N. Nishiwaki, K. R. Knudsen, K. V. Go-
2004, 23, 3053–3061.
1
19] S. Hong, S. Tian, M. V. Metz, T. J. Marks, J. Am. Chem. Soc.
2003, 125, 14768–14783.
[20] A. Hashiguchi, K. J. Fujii, K. Haack, T. Matsumura, R. Ika-
riya, R. Noyori, Angew. Chem. Int. Ed. Engl. 1997, 36, 288–
290.
thelf, K. A. Jørgensen, Angew. Chem. Int. Ed. 2001, 40, 2992–
2
995; h) K. Juhl, K. A. Jørgensen, J. Am. Chem. Soc. 2002,
124, 2420–2421; i) J. Thorhauge, M. Roberson, R. G. Hazell,
K. A. Jørgensen, Chem. Eur. J. 2002, 8, 1888–1898; j) D. A.
Evans, M. C. Kozlowski, J. A. Murry, C. S. Burgey, K. R.
Campos, B. T. Connell, R. J. Staples, J. Am. Chem. Soc. 1999,
[21] J. B. Morrison, E. Grandbois, S. I. Howard, J. Org. Chem.
121, 669–685; k) D. A. Evans, C. S. Burgey, M. C. Kozlowski,
1980, 45, 4229–4231.
Eur. J. Org. Chem. 2006, 4596–4608
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4607

G. P. Miscione et al.
FULL PAPER
[
[
[
[
[
22] M. B. Carter, B. Schiott, A. Gutierrez, S. L. Buchwald, J. Am.
[32] M. J. Frisch, G. W. Trucks, H. B. Schlegel, E.G. Scuseria, M. A.
Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery,
R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam,
A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C.
Adamo, S. Clifford, J. Ochterski, G. A. Petersson, Q. Cui, K.
Morokuma, D. K. Malik, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B. Stefanov, G.
Liu, A. Liashenko, P. Piskorz, I. Kamaromi, R. Gomperts,
R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill,
B. G. Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonza-
lez, M. Head-Gordon, E. S. Replogle, J. A. Pople, Gaussian 98,
Revision A.6; Gaussian, Inc.: Pittsburgh, PA, 1998.
Chem. Soc. 1994, 116, 11667–11668.
23] E. J. Corey, S. Shibata, R. K. Bakshi, J. Org. Chem. 1988, 53,
2861–2863.
24] M. Locatelli, P. G. Cozzi, Angew. Chem. Int. Ed. 2003, 42,
928–4930.
25] A. Kohda, K. Nagayoshi, K. Maemoto, T. Sato, J. Org. Chem.
983, 48, 425–432.
26] Y. Kita, Y. Takebe, K. Murata, T. Naka, S. Akai, J. Org. Chem.
000, 65, 83–88.
4
1
2
[
[
27] K. Nakamura, T. Matsuda, J. Org. Chem. 1998, 63, 8957–8964.
28] J. M. Cherng Fang, T. J. Lu, J. Org. Chem. 1999, 64, 3207–
3212.
[
29] M. Bandini, P. G. Cozzi, M. De Angelis, A. Umani-Ronchi,
Tetrahedron Lett. 2000, 41, 1601–1605.
30] a) SMART & SAINT Software Reference Manuals, Version
[33] a) J. S. Wright, G. A. Di Labio, D. A. Pratt, Chem. Phys. Lett.
1998, 297, 181–186; b) J. S. Wright, E. R. Johnson, G. A.
Di Labio, J. Am. Chem. Soc. 2001, 123, 1173–1183.
[
5.051 (Windows NT Version), Bruker Analytical X-ray Instru-
ments Inc.: Madison, Wi, 1998; b) G. M. Sheldrick, SADABS,
program for empirical absorption correction, University of
Göttingen, Germany, 1996; c) A. Altomare, M. C. Burla, M.
Camalli, G. L. Cascarano, C. Giacovazzo, A. Guagliardi,
A. G. G. Moliterni, G. Polidori, R. Spagna, J. Appl. Crys-
tallogr. 1999, 32, 115–119.
[34] a) N. Godbout, D. R. Salahub, J. Andzelm, E. Wimmer, Can.
J. Chem. 1992, 70, 560–571; b) UniChem. DGauss, Version
2.3.1, 1994, Cray Research, Inc..
Received: February 15, 2006
Published Online: August 10, 2006
[
31] G. M. Sheldrick, SHELXTLplus Version 5.1 (Windows NT ver-
sion) Structure Determination Package; Bruker Analytical X-
ray Instruments Inc.: Madison, WI, 1998.
4608
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4596–4608