The effect of direct steric interaction between substrate substituents and ligand substituents on enantioselectivities in asymmetric addition of diethylzinc to aldehydes catalyzed by sterically congested ferrocenyl aziridino alcohols

Article abstract of DOI:10.1016/j.tetasy.2006.07.028

The direct strong steric interaction between substrate substituents and ligand substituents was first discovered in asymmetric addition of diethylzinc to aldehydes catalyzed by sterically congested ferrocenyl aziridino alcohol derivatives. In addition, th

Full text of DOI:10.1016/j.tetasy.2006.07.028

Tetrahedron: Asymmetry 17 (2006) 2126–2132
The effect of direct steric interaction between substrate
substituents and ligand substituents on enantioselectivities
in asymmetric addition of diethylzinc to aldehydes catalyzed
by sterically congested ferrocenyl aziridino alcohols
Min-Can Wang, Xue-Hui Hou, Chao-Xian Chi and Ming-Sheng Tang*
*
Department of Chemistry, Zhengzhou University, Zhengzhou, Henan 450052, China
Received 23 June 2006; accepted 20 July 2006
Abstract—The direct strong steric interaction between substrate substituents and ligand substituents was first discovered in asymmetric
addition of diethylzinc to aldehydes catalyzed by sterically congested ferrocenyl aziridino alcohol derivatives. In addition, this non-
bonded steric repulsion influenced significantly enantioselectivities of this reaction, and even led to inversion of the absolute configura-
tion. This fact was further confirmed by the theoretical calculations and the design of a new chiral ferrocenyl aziridino alcohol ligand.
A plausible mechanism for this asymmetric reaction was also proposed.
ꢀ
2006 Elsevier Ltd. All rights reserved.
1. Introduction
H2
N
H2
N
H2
N
R
R
R
Zn
Zn
Zn
H
H
H
Since the initial report of Oguni and Omi on the reaction of
diethylzinc with benzaldehyde in the presence of a catalytic
O
O
O
O
O
R
H
O
Zn
Zn
H
amount of (S)-leucinol with moderate enantioselectivity
Zn
R
R
R
1
(
49% ee) in 1984, many chiral ligands have been synthe-
H
anti
syn
inversion
R
R
sized to induce asymmetry in this reaction, the majority
2
of which have been b-amino alcohols. Several types of
Figure 1. Transition states characterized by Yamakawa and Noyori.
transition states for this transformation have been pro-
2
a,3
posed in the literature.
Noyori and co-workers have
studied the reaction mechanism extensively, both experi-
give inversion of the migrating alkyl group. These results
have been reproduced by Houk and Goldfuss in the PM3
and ONIOM (EHF/LanL2DZ:UFF) TS models.
4
5
mentally and theoretically. In 1995, based on the mole-
cular orbital calculations at the restricted Hartree–Fock
6
5
a
(
4
RHF) level, Yamakawa and Noyori established the 5/
/4-fused tricyclic transition states (TS)—syn- and anti-ori-
According to these theoretical studies, the tricyclic transi-
entation of the terminal rings and one bicyclic TS (Fig. 1).
Among the three possible stereoisomeric TS, the tricyclic
anti-configuration is the most favored, being 12–13 kJ/
mol more stable than tricyclic syn-configuration, and
tion state model of the anti-type correctly predicts the abso-
2a,5
lute configuration of products.
As illustrated in Figure
2
, diethylzinc reacts firstly with the ligand b-amino alcohols
to afford the corresponding zinc aminoalkoxide 1 and then
is converted to the zinc monoalkoxide–diethylzinc complex
2
9 kJ/mol more stable than the bicyclic TS. In the tricyclic
anti-TS, alkyl migration takes place with retention of con-
figuration, whereas the high-energy bicyclic pathway would
2
. Benzaldehyde coordinates at the less hindered face of the
five-membered ring chelate 1. The ethyl group transfers
from the diethylzinc to the aldehyde both from the Si-face
and Re-face results in the anti-5/4/4-fused tricyclic transi-
tion states 3 and 4, respectively. In the transition state 3,
a nonbonded repulsion between Et and Ph is absent. As
*
Corresponding authors. Tel./fax: +86 371 67769024 (M.-C.W.); e-mail:
wangmincan@zzu.edu.cn
0957-4166/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2006.07.028

M.-C. Wang et al. / Tetrahedron: Asymmetry 17 (2006) 2126–2132
2127
a result, this transition state is the favored structure and
2. Results and discussion
leads to (S)-1-phenyl-1-propanol. On the other hand, the
significant steric interaction between the Et and Ph groups
disfavors transition state 4, which leads to (R)-1-phenyl-1-
propanol.
More recently, a variety of chiral ferrocenyl aziridino alco-
hol ligands such as 5 and 6 were designed and synthe-
7
,8
sized
because they were b-amino alcohols possessing
unique structural characteristics as follows: (1) the aziridine
ring has the smallest ring area, but most rigid three-mem-
bered ring backbone, as compared with the corresponding
azetidine and pyrrolidine rings. Therefore, we call chiral
ferrocenyl aziridino alcohols ‘rigid and sterically congested
ligands’. (2) The introduction of bulky and rigid ferrocenyl
groups on the nitrogen atom makes this structure become
more rigid. In addition, the crystal structures of some chi-
ral ferrocenyl aziridino alcohols reveal that the nitrogen
The tricyclic transition state model of the anti-type qualita-
tively predicts the degree of enantioselectivity as well. As
2
a
is seen in Figure 2, two bulky cis-disubstituted groups (R
L
0
and R ) in the five-membered chelate complexes 1 are the
L
most desirable, because the directing effects of the vicinal
substituents cooperate in keeping a single chiral integrity
3
, which leads to excellent enantioselectivities. If two bulky
0
groups (R and R ) in the Zn chelate 1 are positioned anti
L
L
7
a,8b
to each other, the directing effects partly offset each other,
but the a stereogenic center is more influential than the b
center. As a result, a decrease in enantioselectivity is gener-
atom on the aziridine ring becomes also a stereocenter.
One possible reason is that the strong repulsive interaction
between the bulky and rigid ferrocenyl group on the nitro-
gen atom of the aziridine ring and the bulky diphenylhydr-
oxymethyl group on the three-membered ring leads to this
tetrahedral structure that is a stable conformation.
Another reason is due to formation of an intramole-
cular hydrogen bond, which further prevents nitrogen
pyramidal inversion.
ated compared to two cis-disubstituted groups (R and
L
0
R ).
L
R2
R2
N
N
Et
R'L
RL
R'S
RS
Zn
Ph
H
R'L
RL
Zn Et
ZnEt2
O
O
7
In our previous papers, we reported the synthesis of chiral
O
PhCHO
R'S
ferrocenyl aziridino alcohol ligands 5 possessing only one
single stereogenic center and application in the enantiose-
lective addition of diethylzinc to aldehydes with excellent
enantioselectivities of up to 98.8% ee (substrate: m-
Zn Et
RS
1
2
Et
Si-face
Et
Re-face
R2
N
R2
N
Et
ClC H CHO). When we investigated the effect of newly
R'L
R'L
RL
R'S
6
4
Zn
Zn
H
Ph
H
introduced methyl group (R = Me) on the aziridine ring
and newly formed stereogenic carbon on the reactivity
and enantioselectivity of reaction in the presence of chiral
ferrocenyl aziridino alcohol ligands 6 bearing two stereo-
RL
R'S
O
O
O
O
Ph
Zn
Zn
Et
RS
RS
Me
Me
Et
8a
3
4
genic centers, we found that introduction of a methyl
group, which was positioned syn to the dialkylhydroxy-
methyl group on the aziridine ring, sharply decreased the
enantioselectivity from 92.6% ee to 8% ee in the addition
of diethylzinc to benzaldehyde. The results suggested that
the steric hindrance of the methyl group, compared to
hydrogen atom, led to the remarkable drop in ee values.
(
S)-1-phenyl-1-propanol
(R)-1-phenyl-1-propanol
Figure 2. Transition structures derived from b-amino alcohol.
Noyori and Yamakawa,5b based on transition state model
and theoretical calculations, believed that steric repulsions
between substituents are absent in the tricyclic transition
state model of the anti-type 3, which is the most favored
transition structure possessing an electronically favored
anti-5/4/4 tricyclic alignment, which contains little steric
constraint, suitably accommodating the substituents in
space. So, direct steric interaction between carbonyl sub-
stituents and a- or b-substituents of the chiral auxiliaries
is also unimportant. To the best of our knowledge, stud-
ies on the effect of nonbonded repulsion between carbonyl
substituents and chiral ligand substituents on enantioselec-
tivities have never been reported. Interestingly, in the pres-
ence of rigid and sterically congested ferrocenyl aziridino
alcohol ligands, the remarkable effect of strong direct steric
interaction between substrate substituents and ligand sub-
stituents on enantioselectivities was firstly observed in
asymmetric addition of diethylzinc to aldehydes. In this
paper, we report our findings on the effect of strong direct
steric interaction between substrate substituents and
ligand substituents on enantioselectivities and theoretical
calculations.
Ph
HO
Ph
HO
H
Ph
R
Ph
N
N
Fc
: 92.6% ee
Fc
6 (R = Me): 8% ee
5
2
a
According to Noyori and Yamakawa’s reaction mecha-
5
nism (Fig. 3), the steric hindrance of the methyl group
of 6 can strengthen the directing effects in the five-mem-
bered chelate complexes 7, followed by forming the favored
anti-5/4/4-fused tricyclic transition state 8, which gives (R)-
1-phenyl-1-propanol, by the reaction of 7 with diethylzinc
and benzaldehyde. Based on transition state model 8, it is
very difficult to explain our experimental results. How does
the steric hindrance of the methyl group work?
For many catalytic processes, the use of different ligands
belonging to the same family does not provoke any mech-
anistic change. Based on a great number of previous theo-

2128
M.-C. Wang et al. / Tetrahedron: Asymmetry 17 (2006) 2126–2132
steric interaction between the R and Ph groups is. When
Ph
Zn
H
R = H (the ligand 5), this steric nonbonded repulsion
between H and Ph is not obvious. Consequently, the tran-
sition state 12 is the favored structure and leads to (S)-1-
phenyl-1-propanol with outstanding enantioselectivity of
up to 92.6% ee. When R = Me (the ligand 6), this direct
steric interaction between Me and Ph increases and leads
to (S)-1-phenyl-1-propanol with poor enantioselectivity of
8% ee. The results suggested that the direct nonbonded
repulsion between carbonyl substituents and substituents
of chiral ligands exists in the asymmetric addition of dieth-
ylzinc to aldehydes catalyzed by aziridino alcohol deriva-
tives. In addition, this interaction is very strong and has
a significant impact on enantioselectivities.
Et
O
R'
R'
O
O
Et
R'
R'
Me Zn
Me
(R)-1-phenyl-1-propanol
Zn
N
N
Fc
7
8
Fc
Figure 3. Transition structures derived from ferrocenyl aziridino alcohols.
2
a,3–6
retical studies on the mechanism of this reaction
and
7
,8
our results, a plausible mechanism for the enantioselec-
tive addition of diethylzinc to benzaldehydes catalyzed by
ligands 5 and 6 is proposed (Fig. 4).
In order to verify this fact, the activation energies of struc-
tures 12 and 13 were calculated for formation of the R and
S enantiomers at HF/6-31G level. Due to calculation diffi-
culty, a ferrocenyl group of ligands 5 and 6 is replaced by a
phenyl group. The results are summarized in Table 1. As
can be seen from Table 1, the favored (S)-generating struc-
ture, anti-Si-12, is 16.60 kJ/mol more stable than the form-
ing structure anti-Re-13 when R group is hydrogen atom,
while the favored (S)-generating structure, anti-Si-12, is
only 9.78 kJ/mol more stable than the forming structure
anti-Re-13 when R is methyl group. The smaller difference
H
Ph
Ph
O
Et
O
Ph
Ph
R
Me
R
Zn
N
H Zn
N
9
10
Fc
Fc
Et
Et
Ph
H
Zn
O
O
Ph
Ph
R
Zn Et
N
(DE = 9.78 kJ/mol) between these activation energies
a
(R = Me), as compared to 16.60 kJ/mol (R = H), appears
to arise mainly from nonbonded repulsion between the
phenyl group of benzaldehyde and methyl group of chiral
ligands.
Si-face
H
Re-face
Fc
1
1
Ph
H
Ph
Zn
Zn
In order to further verify this fact, we designed and synthe-
sized a new chiral ferrocenyl aziridino alcohol ligand 18
from starting material allo-L-threonine according to our
O
Ph
Ph
O
Ph
Ph
O
R
O
Et
R
Zn
N
Et
Zn
N
8
previous reported procedure. The preparation of aziridino
alcohol 18 is shown in Scheme 1. First, ferrocene-carboxal-
Fc
dehyde 14 was reacted in CH OH in the presence of Et N
3 3
1
3
Fc
12
with methyl allo-L-threonine ester hydrochloride giving the
corresponding imine 15. Then, reduction of 15, carried out
on the crude reaction mixtures, with sodium borohydride
in methanol provided the compound 16 in 60% overall
yield after work-up. Next, 16 was cyclized to the aziridino
(
S)-1-phenyl-1-propanol
(R)-1-phenyl-1-propanol
Figure 4. Transition structures derived from ferrocenyl aziridino alcohols.
ester 17 in 95% yield by reaction with Ph P in the presence
3
The reaction of diethylzinc with the ligand 5 and 6 first gen-
erates the corresponding zinc aminoalkoxide 9, which acts
of CCl4 in CH CN. Treatment of 17 with excess of
PhMgBr, afforded the ligand 18 (99%).
3
5
a
as a bifunctional catalyst. In this sterically congested
structure 9, the presence of nonbonded repulsions between
R and Et groups makes Me group of the tricoordinate Zn
atom far away from R group, as envisaged from 10 shown
in Figure 4. The lone pair electrons of the oxygen atom of
benzaldehyde coordinate the Lewis acidic Zn atom at the
less hindered face of the five-membered ring chelate 10,
and then the adjacent basic oxygen accepts diethylzinc at
Zn to form the product-forming, mixed-ligand complex
With the new chiral ligand 18 in hand, under conditions
identical to those for 6, the reaction of diethylzinc with
benzaldehyde afforded (S)-1-phenyl-1-propanol in excellent
yield (99%) with outstanding enantioselectivity of up to
96% ee (Scheme 2). As expected, the absolute configuration
change at 3-position on the aziridine ring led to a substan-
tial increase in enantioselectivity from 8% to 96% ee.
1
1. The ethyl group transfer from the diethylzinc to the
Comparing the structure of 18 with that of 6, the differ-
ences of the compounds 18 and 6 are that methyl group
on the aziridine ring was positioned anti to the diphenyl-
hydroxymethyl group in 18; whereas the methyl group
was syn to the diphenylhydroxymethyl group in 6. Based
on our proposed mechanism, the reaction of the chiral
ligand 18 with diethylzinc and benzaldehyde generated
aldehyde both from the Si-face and Re-face results in the
anti-5/4/4-fused tricyclic transition states 12 and 13, respec-
tively. In the transition state 12, a nonbonded repulsion be-
tween Et and Ph is absent, but steric repulsive interaction
between the R and Ph groups disfavors the transition state
1
2. The more bulky the R group is, the stronger the direct

M.-C. Wang et al. / Tetrahedron: Asymmetry 17 (2006) 2126–2132
2129
a
Table 1. Calculated activation energies of structures 12 and 13
b
TS
R
H
E (Hartree)
TS (Hartree)
E
a
(kJ/mol)
DE
6.60
.78
a
(kJ/mol)
anti-Si-12
anti-Re-13
anti-Si-12
anti-Re-13
ꢀ1679.869555
ꢀ1718.884552
ꢀ1679.838344
ꢀ1679.832022
ꢀ1718.849591
ꢀ1718.845868
81.94
98.54
91.79
1
CH
3
9
101.57
a
All calculations were performed using GAUSSIAN 03 program. All geometries of transition states as well as the reactants were optimized at the HF level
using the 6-31G basis set for C, H, N, and O, LanL2DZ basis set for Zn. The frequency analysis was carried out on the basis of the RHF calculations.
The single-point energy was determined at the same level.
b
Energies of reactants.
OH
HO
HO
CO CH
CH
3
CH3
CO CH
2
3
H
O
H C
N
3
N
NH HCl
CO CH
3
NaBH4
2
3
2
2
Fe
Fe
Fe
NEt , CH OH
3
3
1
4
15
16: 60%
CH3
CH3
Ph
PPh , CCl
3
4
N
C H MgX
N
Fe
6 5
Et N, MeCN
CO CH
3
Fe
3
2
HO
Ph
1
7: 95%
18: 99%
Scheme 1. Synthesis of chiral ligand 18.
O
OH
According to the above analyses, we predict that an inver-
sion of absolute configuration in product will take place
when the steric hindrance of R group on the aziridine ring
of chiral ligands 6 is big enough.
ZnEt , 5% mol 18
2
Toluene, 0-5 ºC, rt, 48 h
H
Ph
Ph
up to 96% ee
Scheme 2. Addition of diethylzinc to benzaldehyde.
Our prediction was supported by Tanner and Andersson’s
experimental results. For instance, the replacement of a
9
methyl group with a bulky phenyl group led to the inver-
sion from (S)-configuration to the (R)-configuration, com-
paring compounds 20 and 21 with 22 and 23, respectively.
However, these phenomena could not be rationalized by
Tanner and Andersson.
the transition state 19, which was the favored structure
because a direct strong steric interaction between methyl
and Ph was absent. Therefore, (S)-1-phenyl-1-propanol
with high enantioselectivity of up to 96% ee was obtained.
Me
Me
Me
Me
HO
HO
HO
HO
Me
Me
Ph
Ph
H
Ph
Zn
Ph
HO
O
Et
Ph
Ph
N
N
N
N
H
Ph
H
O
Ph
0: S, 38% ee
Ph
21: S, 30% ee
Ph
22: R, 87% ee
Ph
23: R, 42% ee
Zn
N
2
N
Me
Me
Fc
Fc
To the best of our knowledge, this is the first discovery that
the direct strong nonbonded interaction between substrate
substituents and ligand substituents leads to remarkable
differences of enantioselectivities in asymmetric addition
of diethylzinc to aldehydes, and even inversion of the con-
figuration. We believe that this direct strong nonbonded
repulsion between substrate substituents and ligand sub-
stituents exists when rigid and sterically congested aziridino
alcohol ligands are used as catalysts for asymmetric addi-
tion of diethylzinc to aldehydes. As a supplement to Noy-
ori and Yamakawa’s transition state model, we expect the
present work to be of interest to all scientists dealing with
the design of chiral ligands and reaction mechanism and
asymmetric catalysis.
1
8
19
Based on our proposed mechanism and transition state
model, the direct nonbonded repulsion between R and phe-
nyl groups will further be strengthened when the steric hin-
drance of R group on the aziridine ring of chiral ligands 6
increases further. Therefore, the significant steric interac-
tion between the R and Ph groups makes the transition
state 12 become the disfavored structure, which gives (S)-
-phenyl-1-propanol. On the other hand, the absent steric
interaction between the R and Ph groups favors the transi-
tion state 13, which affords (R)-1-phenyl-1-propanol.
1

2130
M.-C. Wang et al. / Tetrahedron: Asymmetry 17 (2006) 2126–2132
a
Table 2. Asymmetric addition of diethylzinc to aldehydes using ligands 7
measured with Perkin ELmer, model 341 Polarimeter at
0 ꢁC in CHCl . The ee value was determined by HPLC
2
O
OH
3
ZnEt2 , 5% mol 18
using a chiral column with hexane–propan-2-ol (ratio as
indicated) as the eluent. The chromatographic system con-
sisted of a JASCO model PU-1580 intelligent HPLC pump
and a JASCO model UV-1575 intelligent UV/vis detector
(254 nm). The injection loop had a 20 lL capacity. The col-
umn used was a Chiralcel OD (250 · 4.6 mm) from Daicel
Chemical Ind., Ltd (Japan). The column was operated at
o
Toluene, 0-5 C, rt, 48 h Ar
Ar
H
b
c
d
Entry
ArCHO
CHO
p-MeOC
m-MeOC
o-MeOC
Heliotripine
m-ClC CHO
m-BrC CHO
Yield (%)
Ee (%)
Confign.
1
2
3
4
5
6
7
8
9
C
6
H
5
99
96
85
90
83
92
99.8
91
92
81
S
S
S
S
S
S
S
S
S
6
H
4
CHO
CHO
CHO
81
88
96
92
82
91
89
92
6
H
4
H
4
1
13
6
ambient temperature. NMR Spectra ( H and C) were
performed on a Bruker DPX 400 (400 MHz) spectrometer
using solutions in CDCl (referenced internally to Me Si); J
6
H
4
3
4
6
H
4
values are given in hertz. IR Spectra were determined on a
Therme Nicolet IR 200 spectrophotometer. TLC was per-
formed on dry silica gel plates developed with hexane/ethyl
acetate. Mass spectra were obtained using a Waters a-Tof
Ferronenyl aldehyde
PhCH@CHCHO
a
b
c
The mole ratio Et
Isolated yields.
2
Zn/aldehyde was 2/1.
TM
micro instrument with an electrospray ionization source
Determined by HPLC using a chiral OD column.
d
(ESIMS). All the ESIMS spectra were performed using
MeOH as the solvent.
Absolute configuration assigned by comparison with the sign of specific
rotation of a known compound and the known elution order from a
Chiralcel OD column.
4.2. Reagents and solvents
The above exciting results prompted us to probe whether
this improvement existed in other aldehydes. Thus, a vari-
ety of aldehydes were used subsequently with chiral ligand
Methyl allo-L-threonine ester hydrochloride was prepared
1
0
according to a published procedure. Ether was distilled
from sodium benzophenone ketyl. Tetrahydrofuran
(THF) was pre-dried over calcium chloride, and then dis-
1
8. The results are summarized in Table 2. The results re-
veal that the new chiral ligand 18 is a highly efficient cata-
lyst for the reaction of diethylzinc with arylaldehydes to
afford optically active secondary alcohols in high yield
tilled from LiAiH . Et N was dried with KOH pellets, then
4 3
refluxed for 2 h with p-toluenesulfonyl chloride and dis-
tilled. p-Toluenesulfonyl chloride was purified by dissolv-
(
81–99%) and with excellent enantioselectivities in the
range 83–99.8% ee. The ligand 18 was also tested with an
a,b-unsaturated aliphatic aldehyde. It was found that this
catalyst is also very effective in the enantioselective addi-
tion of diethylzinc to a,b-unsaturated aliphatic aldehyde
ing (10 g) in the minimum volume of CHCl (ca. 25 mL)
3
filtered, and diluted with five volumes (i.e., 125 mL) of
pet ether (30–60 ꢁC) to precipitate impurities. The solution
was filtered, clarified with charcoal, concentrated to 15 mL
by evaporation and gave 7 g of white crystals. All other
reagents are commercially available and were used as
received.
(
Table 2, entry 9).
3. Conclusion
4.3. Synthesis of methyl (2S,3S)-N-ferrocenylmethyl-allo-L-
threomine ester 16
In conclusion, we have reported a new discovery of the
presence of the strong direct steric interaction between sub-
strate substituents and ligand substituents in the asymmet-
ric addition of diethylzinc to aldehydes catalyzed by
ferrocenyl aziridino alcohol derivatives and the significant
effect of this nonbonded repulsion on enantioselectivities.
In addition, this fact was verified by theoretical calcula-
tions and the design of a new chiral ferrocenyl aziridino
alcohol ligand 18. A plausible mechanism for this asym-
metric reaction is proposed. Based on our transition state
model, a prediction of the inversion of configuration was
made, and then was supported by Tanner and Andersson’s
experimental results. Our efforts to develop novel ferroce-
nyl aziridino alcohols as chiral ligands for asymmetric syn-
thesis are underway in our laboratory and will be reported
in due time.
Methyl allo-L-threonine ester hydrochloride (1.55 g) was
dissolved in 8 mL of anhydrous methanol and cooled to
0 ꢁC. Triethylamine (1.4 mL) was added, and the reaction
was stirred for 10 min. Ferrocene carboxaldehyde 14
(1.95 g) was added, and the reaction was monitored by
TLC. After the reaction mixture was stirred for 4 h, sodium
borohydride (0.4 g) was added portionwise to the reaction
mixture over a period of 0.5 h. After stirring for 5 h, meth-
anol was evaporated under reduced pressure at 40 ꢁC. The
resulting residue was carefully neutralized with 3% HCl to
pH = 7–8, and extracted three times with 3 · 20 mL por-
tion of EtOAc. The combined ether extract was washed
with brine, dried over Na SO , and evaporated under the
2
4
reduced pressure. The resulting residue was purified by
the preparative TLC with petroleum (60–90 ꢁC)/Et O
2
(
1:1) as developing solvent to give 16 in 60% yield (1.8 g);
20
1
4
. Experimental
mp 87.8–88.8 ꢁC; ½aꢁ ¼ ꢀ40:8 (c 1.00, CHCl ). H NMR
D
3
(400 MHz, CDCl ): d 1.09 (d, J = 6.4 Hz, 3H, CHCH ),
3
3
4
.1. General
2.1–2.8 (br s, 2 H, OH, NH), 3.46 (d, J = 7.6 Hz, 1H,
COCH), 3.42, 3.58 (dd, J = 12.4 Hz, each 1H, NHCH2),
3.76 (s, 3H, OCH ), 4.01–4.09 (m, 1H, CH CH), 4.12–
Melting points were determined using YRT-3 melting point
apparatus, and were uncorrected. Optical rotations were
3
3
1
3
4.24 (m, 9H, FcH). C NMR (100 MHz, CDCl ): d
3

M.-C. Wang et al. / Tetrahedron: Asymmetry 17 (2006) 2126–2132
2131
1
8
2
1
8.62, 47.83, 51.88, 65.34, 66.85, 67.72, 67.78, 68.34, 68.45,
5.75, 173.25. IR (KBr): 3322, 3171, 2982, 2943, 2907,
851, 1726, 1479, 1377, 1318, 1264, 1206, 1171, 1128,
126.50, 127.01, 128.06, 128.11. IR (KBr): 3314, 3085,
3055, 3024, 2994, 2854, 1667, 1597, 1489, 1447, 1387,
1319, 1265, 1180, 1114, 1084, 1029, 997, 901, 818, 750,
ꢀ1
+
093, 1030, 1008, 987, 917, 818 cm . HRMS (ESI): m/z
698, 488. HRMS (ESI): m/z (M+H) calcd for C H Fe-
2
7
27
+
(
M+H)
calcd for C H FeNO 331.0871; found:
NO: 438.1520; found: 438.1508.
1
6
21
3
3
31.0875.
4.6. General procedure for the enantioselective addition of
Et Zn to arylaldehydes
4.4. Synthesis of methyl (2S,3R)-1-ferrocenylmethyl-3-
methylaziridine-2-carboxylate 17
2
A solution of diethylzinc (1 M in n-hexane, 1.1 mL) was
added to a solution of a chiral catalyst (0.025 mmol,
5 mol %) in dry toluene under a nitrogen atmosphere.
The mixture was cooled to 0 ꢁC, and stirred for 30 min.
Freshly distilled benzaldehyde (0.05 mL, 0.5 mmol) was
added to the mixture. The resulting mixture was stirred
for 10 h in 0–5 ꢁC and was allowed to warm to room tem-
perature, and kept stirring for another 38 h at the same
temperature. The reaction was quenched by the addition
To a stirred solution of triphenylphosphine (3.5 g,
1
3.3 mmol) in acetonitrile (40 mL) was added carbon tetra-
chloride (20 mL). The solution turned yellow over a period
of 0.5 h, at which time 16 (1.5 g, 4.5 mmol) was added
dropwise in a solution of Et N (2 mL, 14.4 mmol) and ace-
3
tonitrile (15 mL). The reaction mixture was stirred for 2 h
at 0 ꢁC, 14 h at room temperature. Removal of the solvent,
followed by the preparative TLC of the residue with petro-
leum (60–90 ꢁC)/Et O (1:1) as developing solvent afforded
of saturated aqueous NH Cl (4 mL). The mixture was ex-
2
4
1
7: 1.34 g (95%). 17 was obtained as a 1.7:1 mixture of
tracted with Et O (3 · 8 mL). The combined organic layers
2
20
invertomers at nitrogen: ½aꢁ ¼ ꢀ86:6 (c 0.77, CHCl ).
were washed with brine, dried over anhydrous Na SO ,
D
3
2
4
1
H NMR (400 MHz, CDCl ): d major invertomer: 1.20
and evaporated under reduced pressure. Purification of
the residue by the preparative silica gel TLC plate (hex-
ane/EtOAc = 4/1) afforded the (S)-1-phenyl-1-propanol.
The ee was determined by HPLC analyses using a chiral
column.
3
(
d, J = 5.6 Hz, 3H, CHCH ); 2.24–2.28 (m, 1H, CH CH);
3 3
2
1
4
6
2
.34 (d, J = 2.8 Hz, 1H, COCH); 3.63, 4.07 (dd, J =
0
2.8 Hz, each 1H, NCHH ); 3.75 (s, 3H, OCH ); 4.10–
3
.26 (m, 9H, FcH). Minor invertomer: 1.39 (d, J =
.0 Hz, 3H, CHCH ); 1.99 (d, J = 2.8 Hz, 1H, COCH);
3
.48–2.51 (m, 1H, CH CH); 3.34, 3.57 (dd, J = 12.8 Hz,
3
0
Acknowledgments
each 1H, NCHH ); 3.70 (s, 3H, OCH ); 4.10–4.26 (m,
3
1
3
9
H, FcH). C NMR (100 MHz, CDCl ): d major inver-
3
We are grateful to the National Natural Sciences Founda-
tion of China (NNSFC: 20172047), the Ministry of Educa-
tion of China, and the Education Commission of Henan
Province for the financial supports.
tomer: 17.91, 40.99, 42.75, 50.63, 67.83, 67.89, 68.04,
6
1
6
2
1
8.48, 68.61, 68.74, 85.12, 170.28. Minor invertomer:
0.92, 40.03, 44.15, 51.99, 67.83, 67.89, 68.04, 68.48,
8.61, 68.74, 84.89, 171.64. IR (KBr): 3100, 2994, 2952,
925, 1729, 1438, 1400, 1341, 1202, 1179, 1126, 1105,
ꢀ
1
054, 1023, 1001, 925, 818 cm . HRMS (ESI): m/z (M+
References
+
Na) calcd for C H FeNO : 336.0663; found: 336.0668.
1
6
19
2
1
2
. Oguni, N.; Omi, T. Tetrahedron Lett. 1984, 25, 2823–2824.
. Reviews on enantioselective organozinc additions to alde-
hydes: (a) Noyori, R.; Kitamura, M. Angew. Chem., Int. Ed.
Engl. 1991, 30, 49–69; (b) Soai, K.; Niwa, S. Chem. Rev. 1992,
92, 833–856; (c) Pu, L.; Yu, H.-B. Chem. Rev. 2001, 101, 757–
824.
4.5. Synthesis of (2S,3R)-1-ferrocenylmethyl-3-methylaziri-
din-2-yl(diphenyl)methanol 18
A Grignard reagent was prepared in the usual way from
7
3
6.8 mg (3.2 mmol) of magnesium and bromobenzene
.2 mmol in THF (5 mL). The solution was cooled to
3
. (a) Itsuno, S.; Fr e´ chet, J. M. J. J. Org. Chem. 1987, 52, 4140–
4
142; (b) Evans, D. A. Science 1988, 240, 420–426; (c) Corey,
E. J.; Yuen, P.-W.; Hanon, F. J.; Wierda, D. A. J. Org. Chem.
990, 55, 784–786; (d) Rasmussen, T.; Norrby, P.-O. J. Am.
ꢀ
20 ꢁC before addition of a solution of 17 (125 mg,
0
.4 mmol) in THF (2 mL). The mixture was allowed to
1
reach room temperature. After stirring for 24 h, the reac-
Chem. Soc. 2001, 123, 2464–2465; (e) Vidal-Ferran, A.;
Moyano, A.; Peric a´ s, M. A.; Riera, A. Tetrahedron Lett.
tion was quenched with saturated aqueous NH Cl
4
(
10 mL) at 0 ꢁC. The phases were separated and the aque-
1
997, 38, 8773–8776; (f) V a´ zquez, J.; Peric a´ s, M. A.; Maseras,
ous phase was extracted with Et O (3 · 5 mL). The com-
bined organic phases were washed with brine (15 mL),
dried over Na SO , and after filtration, the solvent was re-
2
F.; Lled o´ s, A. J. Org. Chem. 2000, 65, 7307–7309; (g)
Kozlowski, M. C.; Dixon, S. L.; Panda, M.; Lauri, G. J. Am.
Chem. Soc. 2003, 125, 6614–6615.
2
4
moved under reduced pressure. The resulting residue was
purified by preparative TLC with hexane/EtOAc (4:1) as
developing solvent to give 18 (173 mg, 99%); mp 144–
4. (a) Kitamura, M.; Okada, S.; Suga, S.; Noyori, R. J. Am.
Chem. Soc. 1989, 111, 4028–4036; (b) Kitamura, M.; Oka, H.;
Noyori, R. Tetrahedron 1999, 55, 3605–3614.
20
1
5. (a) Yamakawa, M.; Noyori, R. J. Am. Chem. Soc. 1995, 117,
1
45 ꢁC; ½aꢁ ¼ ꢀ10:0 (c 0.46, CHCl3).
H
NMR
D
6
327–6335; (b) Yamakawa, M.; Noyori, R. Organometallics
(
400 MHz, CDCl ): d 1.39 (d, J = 5.8 Hz, 3H, CH );
3
3
1
999, 18, 128–133.
2
.12–2.16 (m, 1H, CH CH); 2.20 (d, J = 3.2 Hz, 1H,
3
6
. (a) Goldfuss, B.; Houk, K. N. J. Org. Chem. 1998, 63, 8998–
9006; (b) Goldfuss, B.; Steigelmann, M.; Khan, S. I.; Houk,
K. N. J. Org. Chem. 2000, 65, 77–82; (c) Goldfuss, B.;
Steigelmann, M.; Rominger, F. Eur. J. Org. Chem. 2000,
1785–1792.
CH CHCH); 3.20, 3.64 (dd, J = 13.2 Hz, each 1H,
3
0
CHH N); 3.89–4.09 (m, 9H, FcH); 7.23–7.38 (m, 10H,
1
3
PhH). C NMR (100 MHz, CDCl ): d 11.15, 35.06,
3
4
9.51, 53.05, 65.90, 67.93, 68.48, 68.92, 73.97, 126.12,

2132
M.-C. Wang et al. / Tetrahedron: Asymmetry 17 (2006) 2126–2132
7
. (a) Wang, M.-C.; Wang, D.-K.; Zhu, Y.; Liu, L.-T.; Guo, Y.- Xu, C.-L.; Liu, L.-T.; Li, G.-L.; Wang, D.-K. Synthesis 2005,
F. Tetrahedron: Asymmetry 2004, 15, 1289–1294; (b) Wang,
M.-C.; Liu, L.-T.; Zhang, J.-S.; Shi, Y.-Y.; Wang, D.-K.
Tetrahedron: Asymmetry 2004, 15, 3853–3859.
. (a) Wang, M.-C.; Wang, D.-K.; Lou, J.-P.; Hua, Y.-Z. Chin.
J. Chem. 2004, 22, 512–514; (b) Wang, M.-C.; Hou, X.-H.;
3620–3626.
9. Tanner, D.; Kornø, H. T.; Guijarro, D.; Andersson, P. G.
Tetrahedron 1998, 54, 14213–14232.
10. Mayers, A. I.; Schmidt, W.; Mcdennon, M. J. Synthesis 1993,
250–362.
8