Catalytic asymmetric aziridination with borate catalysts derived from VANOL and VAPOL ligands: Scope and mechanistic studies

Article abstract of DOI:10.1002/chem.200701558

An extended study of the scope and mechanism of the catalytic asymmetric aziridination of imines with ethyl diazoacetate mediated by catalysts prepared from the VANOL and VAPOL ligands and triphenylborate is described. Nonlinear studies with scalemic VANOL and VAPOL reveal an essentially linear relationship between the optical purity of the ligand and the product suggesting that the catalyst incorporates a single molecule of the ligand. Two species are present in the catalyst prepared from B(OPh)₃ and either VANOL or VAPOL as revealed by ¹H NMR studies. Mass spectral analysis of the catalyst mixture suggests that one of the species involves one ligand molecule and one boron atom (B1) and the other involves one ligand and two boron atoms (B2). The latter can be formulated as either a linear or cyclic pyroborate and the ¹¹B NMR spectrum is most consistent with the linear pyroborate structure. Several new protocols for catalyst preparation are developed which allow for the generation of mixtures of the B1 and B2 catalysts in ratios that range from 10:1 to 1:20. Studies with catalysts enriched in the B1 and B2 species reveal that the B2 catalyst is the active catalyst in the VAPOL catalyzed asymmetric aziridination reaction giving significantly higher asymmetric inductions and rates than the B1 catalyst. The difference is not as pronounced in the VANOL series. A series of 12 different imines were surveyed with the optimal catalyst preparation procedure with the finding that the asymmetric inductions are in the low to mid 90s for aromatic imines and in the mid 80s to low 90s for aliphatic imines for both VANOL and VAPOL catalysts. Nonetheless, the crystallinity of the N-benzhydryl aziridines is such that nearly all of the 12 aziridine products screened can be brought to >99% ee with a single recrystallization.

Full text of DOI:10.1002/chem.200701558

FULL PAPER
DOI: 10.1002/chem.200701558
Catalytic Asymmetric Aziridination with Borate Catalysts Derived from
VANOL and VAPOL Ligands: Scope and Mechanistic Studies
Yu Zhang, Aman Desai, Zhenjie Lu, Gang Hu, Zhensheng Ding, and
William D. Wulff*[a]
Abstract: An extended study of the
scope and mechanism of the catalytic
asymmetric aziridination of imines with
ethyl diazoacetate mediated by cata-
lysts prepared from the VANOL and
VAPOL ligands and triphenylborate is
described. Nonlinear studies with scale-
mic VANOL and VAPOL reveal an es-
sentially linear relationship between
the optical purity of the ligand and the
product suggesting that the catalyst in-
corporates a single molecule of the
ligand. Two species are present in the
(B1) and the other involves one ligand
and two boron atoms (B2). The latter
can be formulated as either a linear or
cyclic pyroborate and the 11B NMR
spectrum is most consistent with the
linear pyroborate structure. Several
new protocols for catalyst preparation
are developed which allow for the gen-
eration of mixtures of the B1 and B2
catalysts in ratios that range from 10:1
to 1:20. Studies with catalysts enriched
in the B1 and B2 species reveal that
the B2 catalyst is the active catalyst in
the VAPOL catalyzed asymmetric azir-
idination reaction giving significantly
higher asymmetric inductions and rates
than the B1 catalyst. The difference is
not as pronounced in the VANOL
series. A series of 12 different imines
were surveyed with the optimal catalyst
preparation procedure with the finding
that the asymmetric inductions are in
the low to mid 90s for aromatic imines
and in the mid 80s to low 90s for ali-
phatic imines for both VANOL and
VAPOL catalysts. Nonetheless, the
crystallinity of the N-benzhydryl aziri-
dines is such that nearly all of the 12
aziridine products screened can be
brought to >99% ee with a single re-
crystallization.
catalyst prepared from B(OPh)3 and
G
either VANOL or VAPOL as revealed
by 1H NMR studies. Mass spectral
analysis of the catalyst mixture suggests
that one of the species involves one
ligand molecule and one boron atom
Keywords: asymmetric catalysis
·
asymmetric synthesis · aziridines ·
chiral ligands
Introduction
the formation of the cis-substituted aziridine 3. The process
is thought to involve coordination of the Lewis acid to the
À
The formation of aziridines from the Lewis acid mediated
reaction of imines with diazo compounds has been known
for quite some time dating to the early observations involv-
ing boron trifluoride etherate and zinc iodide.[1–3] The mech-
anism of these reactions (Scheme 1) is distinctively different
from those involving copper or rhodium catalysts where a
metal carbene or metal carbenoid is involved.[4] The scope
of the Lewis acid mediated reactions was not appreciated
until the reports by Brookhart and Templeton in 1996[5] and
Jorgensen in 1997.[6] The reaction of ethyl diazoacetate with
imines mediated by a Lewis acid is normally selective for
imine and then carbon carbon bond formation resulting
from the attack of the diazo carbon in 1 on the activated
imine 6. From steric considerations, the approach of 1 and 6
may be expected to most favorably generate the zwitterion
7, which after bond rotation to give 8 and then backside dis-
placement of nitrogen would give the cis-isomer of 3. The
reaction is also known to give varying amounts of the iso-
meric enamines 4 and 5 which are likely the result of nitro-
gen loss concomitant with a 1,2-carbon or 1,2-hydrogen shift
in the zwitterionic intermediate 8, respectively. Consistent
with this view are the increased amounts of enamine prod-
ucts observed in more polar solvents in which the zwitterion
8 may be expected to have a longer lifetime. Interestingly,
Brookhart and Templeton reported that aziridine 3 could
only be observed with substoichiometric amounts of Lewis
acids.[5] In the presence of one equivalent of BF3·Et2O, no
aziridine is observed even though complete conversion is
[a] Y. Zhang, A. Desai, Z. Lu, G. Hu, Z. Ding, Prof. Dr. W. D. Wulff
Department of Chemistry, Michigan State University
East Lansing, MI 48824 (USA)
Fax : (+1)517-353-1793
E-mail: wulff@chemistry.msu.edu
Chem. Eur. J. 2008, 14, 3785 – 3803
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3785

presumably achieved. Apparently, a stoichiometric complex
of the aziridine 3 with BF3·Et2O is not stable relative to ring
opening processes.
Scheme 1. Mechanism of the Lewis acid mediated aziridination reaction.
Taking the lead from the pioneering studies of Brookhart
and Templeton[5] and of Jorgensen,[6] we developed a catalyt-
ic asymmetric version of this reaction[7,8] with chiral boron
derived Lewis acids of the vaulted biaryl ligands[7g] VANOL
and VAPOL in 1999[7a] from BH3·THF and the following
year with similar catalysts prepared from triphenylborate.[7b]
The latter catalysts were quite general giving high asymmet-
ric inductions and high cis-selectivities for the aziridines 10
from benzhydryl imines 9 prepared from both aliphatic and
aromatic aldehydes. One of the more remarkable observa-
tions is that the VANOL and VAPOL ligands give very
closely related profiles of asymmetric induction over the
entire scope of substrates that were examined. We have
never seen this before for any other reaction that we have
studied with these ligands[9] and it is certainly not under-
stood at this point. In all cases that have been determined,
catalysts derived from the R enantiomers of VANOL and
VAPOL give Re-face addition to the benzhydryl imines and
those from the S enantiomers give Si-face addition. The pro-
cedure for generation of the catalyst is shown in Scheme 2
(Procedure A) and involves heating either ligand with three
Scheme 2. Lewis acid catalyzed asymmetric aziridination using VANOL
and VAPOL as chiral ligands.
for this reaction of 95Æ2% ee. When this discrepancy was
first discovered it was assumed that the purity of one of the
reagents had either been compromised or an impurity was
making an appearance in one of the reagents. Possible sour-
ces of this change in asymmetric induction could include: 1)
the purity of the imine, 2) the purity of the commercially
available triphenylborate, 3) the purity of the commercially
available ethyl diazoacetate, 4) the optical purity or chemi-
cal purity of the ligand, and 5) experimental techniques in-
cluding the introduction or exclusion of water. All of these
variables were thoroughly investigated over a period of sev-
eral months to no avail. This included the independent syn-
thesis of triphenylborate and ethyl diazoacetate by at least
two different published procedures. We were left to con-
clude that whatever the source of this difference, with rigor-
ously purified reagents, the reaction of imine 9b with ethyl
diazoacetate gives the aziridine 10b in 89Æ2% ee. A reeval-
uation of a few other imines are shown in Table 1 and as
can be seen, the asymmetric inductions are 0–8% ee lower
with aryl imines and 8–16% lower with alkyl imines. Thus,
the purpose of the present investigation is to optimize this
catalytic asymmetric aziridination reaction. This will include
efforts to identify the structure of the active catalyst and the
optimization of catalyst formation as well as evaluation of
the reactions conditions. Finally, the new optimized protocol
will be employed in a study of a thorough reevaluation of
the scope of this catalytic asymmetric aziridination which
will include 12 different imines.
equivalents of B(OPh)3 in methylene chloride at 558C for
A
1 h and then removing the volatiles under vacuum (0.1 mm
Hg) at 558C for 30 min providing the catalyst as a white to
pale-yellow semi-solid. Part of the purpose of the present
work is to shed light on possible structures for this catalyst
and another is to optimize and examine the scope of the cat-
alytic asymmetric aziridination (AZ) reaction.
In the course of pursuing various applications of this cata-
lytic asymmetric aziridination reactions over the last several
years, we began to find that some of the reactions we had
published in 2000[7b] were not reproducible. For example,
the reaction of the benzhydryl imine of benzaldehyde 9b
was originally reported to give 95% ee but more recently
we have found that this reaction gives 89% ee. We first no-
ticed this difference somewhere near the end of 2002. Prior
to 2003at least six members of our research group had re-
peated this reaction and measured an asymmetric induction
3786
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3785 – 3803

Catalytic Asymmetric Aziridination
FULL PAPER
Table 1. Reinvestigation of aziridinations with catalyst prepared by Procedure A in Scheme 2.[a]
Catalyst loading and reaction
scale-up: The standard catalyst
loading that was employed in
our
original
report
was
10 mol%.[7b] In an effort to
define the turnover efficiency
of this catalyst, a catalyst load-
ing study was carried out with
the reaction of imine 9b and
ethyl diazoacetate with the
VAPOL catalyst 13 generated
as shown in Scheme 2. This
study was carried out in both
methylene chloride and carbon
tetrachloride as solvent and the
results are presented in Table 3.
The reaction of 9b in carbon
tetrachloride under the stan-
Entry
Series
R
Ligand
Yield 10 [%][b]
(ref. [7b])
ee 10 [%][c]
(ref. [7b])
Yield 10 [%][b]
(this work)
ee 10 [%][c]
(this work)
N
G
1
2
3
4
5
6
7
8
9
a
1-naphthyl
Ph
(S)-VAPOL
(S)-VAPOL
(S)-VANOL
(S)-VAPOL
(S)-VANOL
(S)-VAPOL
(S)-VANOL
(S)-VAPOL
(S)-VAPOL
(S)-VANOL
87
77[e]
85
92[d]
95
96
94
91
98
98
94
91
97
87
8389
81
93
b
88
88
83
90
90
78
69
68
c
o-MeC6H4
65
69
91[f]
85
85[f]
86
f
p-BrC6H4
cyclohexyl
tert-butyl
k
l
74
74
78
8383
10
77[g]
93[g]
83
[a] Unless otherwise specified, all reactions were run in CH2Cl2 containing 0.5 m imine at 258C for 24 h with
1.1 equiv of ethyl diazoacetate and 10 mol% of the catalyst which was prepared by Procedure A in Scheme 2.
The reactions with (R)-VANOL give ent-10. [b] Isolated yield after purification by chromatography on silica
gel. [c] Determined by HPLC on a Chiralcel OD-H column. [d] The retention time for the minor enantiomer
was mis-assigned in the original report. [e] Reaction with 2.5 mol% catalyst. [f] Solvent is toluene/CH2Cl2 1:1.
[g] Reactions conditions: toluene, 08C for 4 h and then 258C for 20 h.
dard
catalyst
loading
of
10 mol% was found to give
higher asymmetric induction
than the reaction in methylene
chloride; 93% ee for the former
Results and Discussion
and 89% ee for the latter (Table 3, entries 1 vs 5). All of the
reactions in Table 3were carried out with 0.1 mmol of
VAPOL and as the catalyst loading was decreased the
amount of substrate was increased. As a result of the in-
creased scale of the reactions at low loading, the product
from all of the reactions in Table 3carried out at less than
10 mol% catalyst was first isolated by recrystallization. In
order to monitor the effect of catalyst loading on the total
reaction yield and on the total asymmetric induction, in
each reaction after two crops of the product were taken, the
remaining product was isolated from the mother liquor by
column chromatography on silica gel. The total yield for the
Background reaction and Lewis acid evaluation: Triphenyl-
borate is not a particularly strong Lewis acid and would not
be expected to promote the reaction of ethyl diazoacetate
and benzhydryl imines as readily as other stronger Lewis
acids.[10] This expectation was borne out by the data in
Table 2 for the reaction of the 1-naphthyl imine 9a with
ethyl diazoacetate. Nonetheless, triphenylborate will pro-
mote the reaction to some extent but not as well as
BF3·etherate or Yb
A
G
to 20% conversion with 10 mol% catalyst in 24 h at room
temperature in methylene chlo-
ride. Interestingly, the best cata-
Table 2. Lewis acid mediated aziridination of benzhydryl imine 9a.[a]
lyst for this reaction was that
generated from either VANOL
or VAPOL and triphenylborate.
They were so efficient in fact,
that it was found that the best
method for the preparation of
the racemates to be used as
standards in the measurement
of asymmetric inductions of the
various aziridines examined in
the present work was to employ
catalysts prepared from racemic
VAPOL. The VANOL and
VAPOL catalysts gave lower
amounts of the enamine side-
products 14 and 15 as well as
higher yields of the aziridine
product 10 and higher cis/trans
selectivities for the aziridine 10.
Entry Lewis acid
Conv. [%][b] Yield 10a [%][c] cis/trans 10a[b] ee 10a [%][d] Ratio
10a/14a/15a[b]
1
2
3Yb
4
5
6
7
BF3·OEt2
Yb(OTf)3
(OTf)3
(OPh)3
rac-VAPOL catalyst 13
(S)-VAPOL catalyst 13
(S)-VANOL catalyst 13
100
85
4325:1
39
37
–
5:0.021.0:0.3
G
25:1
10:1
30:1
–
–
–
1.0:0.30:0.05
1.0:–[f]:0.04
1.0:0.17:0.21
1.0:0.03:0.01
1.0:0.03:0.06
1.0:0.01:0.05
G
70[e]
20
B
G
15
100
100
100
88
87
85
33:1
>30:1
>30:1
–
92
88
[a] Unless otherwise specified, all reactions were carried out at 0.5 m in 9a in CH2Cl2 at 258C for 24 h with
10 mol% catalyst and 1.1–1.2 equiv of ethyl diazoacetate. The VAPOL and VANOL catalysts were prepared
as indicated in Scheme 2. [b] Determined by 1H NMR on the crude reaction mixture. [c] Isolated yield after
chromatography on silica gel. [d] Determined by HPLC on a Chiralcel OD-H column. [e] Reaction in hexane.
[f] Not determined due to absorptions for 14a being obscured by those from unidentified compounds.
Chem. Eur. J. 2008, 14, 3785 – 3803
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3787

W. D. Wulff et al.
Table 4. Effect of solvent on the aziridination of imine 9b.[a]
reaction does not significantly change as the catalyst loading
is reduced and the total induction only drops slightly. The
reaction does not fail to go to completion until the catalyst
loading is reduced to 0.25 mol% in methylene chloride
(44% completion, entry 4) and 0.125 mol% in carbon tetra-
chloride (68% completion, entry 9). Thus, the reaction of
9b with the VAPOL catalyst 13 will turnover 176 times in
methylene chloride and 544 times in carbon tetrachloride.
Although the asymmetric induction of the reaction drops to
86% ee in carbon tetrachloride at 0.25 mol% catalyst, the
reaction goes to completion and the product can be ob-
tained in 64% yield and 98% ee with a single recrystalliza-
tion. In this reaction 54 mg of VAPOL was used to give
8.78 g of 10b of 98% ee.
Given the finding of the increased asymmetric induction
in carbon tetrachloride, a screen of solvents for the asym-
metric aziridination reaction was carried out on the benzhy-
dryl imine 9b and ethyl diazoacetate (Table 4). Given its
Lewis basic nature, it was not a surprise that the reaction in
acetonitrile was sluggish and only went to 77% completion
in 24 h with 10 mol% of the VAPOL derived catalyst 13.
What was surprising is that the reaction in THF did go to
completion and gave essentially the same yield and asym-
metric induction as methylene chloride. Toluene and ben-
zene both gave slightly better inductions than methylene
chloride and gave about the same yield. A variety of other
halogenated solvents were also examined but none gave a
higher asymmetric induction than carbon tetrachloride
(93% ee). The differences in the asymmetric inductions be-
tween methylene chloride, toluene and carbon tetrachloride
are small but reproducible; each reaction was carried out a
minimum of 4–6 times with an induction of 89Æ2, 91Æ2,
and 93Æ2% ee, respectively.
Entry
Solvent
Yield 10b [%][b]
ee 10b [%][c]
1
2
3CF
4
5
CH3CN
CH2Cl2
3C6H5
CHCl3
THF
60[d]
8389
82
81
79
26
90
90
90
6
7
8
9
CH3C6H5
CS2
C6H6
8391
7391
8392
84
CCl4
93
[a] Unless otherwise specified, all reactions were carried out at 0.5m in
imine at 258C with 1.1–1.2 equiv of ethyl diazoacetate and 10 mol% cat-
alyst which was prepared by Procedure A in Scheme 2, and all reactions
went to completion. [b] Isolated yield after chromatography on silica gel.
[c] Determined by HPLC on a Chiralcel OD-H column. [d] 77% conver-
sion.
The rate of the reaction of imine 9b and ethyl diazoace-
tate was examined in the solvents methylene chloride,
carbon tetrachloride and toluene. This was done by lowering
the catalyst loading to 2 mol% and then stopping the reac-
tion after various time intervals. The results in Table 5
reveal that the reaction is slowest in methylene chloride and
fastest in carbon tetrachloride with toluene close behind.
Thus, the solvent of choice is clearly toluene when the fac-
tors of rates, asymmetric induction, expense and environ-
mental concerns are taken into consideration.
Table 3. Catalyst loading study with imine 9b with the VAPOL catalyst.[a]
Catalyst structure and nonlinear
effects: It has been well estab-
lished that catalysts prepared
from triphenylborate and the
BINOL ligand can have two
Entry 9b
[mmol]
Solvent Loading
[mol%]
1st Crop
2nd Crop
ee mother
Overall
equivalents of BINOL per
boron.[12] Given the increased
size of the VANOL and
VAPOL ligands compared with
BINOL it might be anticipated
that the formation of catalysts
composed of two equivalents of
VANOL or VAPOL per boron
would not be favorable. How-
ever, an examination of CPK
models reveals that this could
be possible. A popular method
for probing the stoichiometric
composition of a catalyst is to
test for a nonlinear relationship
between the optical purity of
the ligand and that of the prod-
uct. This method and its short-
liquor [%][c]
yield
ee
yield
ee
yield
ee
[%][b]
[%][c] [%][b]
[%][c]
[%][d]
[%][c,e]
1
1
CH2Cl2 10
nd
62
5398
–
nd
98
22
–
nd
16
nd
60
66
–
nd
59
8389
83
86
–
84
86
88
88
–
2
10
CH2Cl2
1
89
320
4[f]
5
CH
2Cl2 0.5
CH2Cl2 0.25
59
79
40
1
–
–
–
CCl4
CCl4
CCl4
CCl4
CCl4
10
1
0.5
0.25
0.125
nd
59
58
64
–
nd
99
98
98
–
nd
24
25
19
–
nd
77
65
50
–
nd
62
65
66
–
93
91
86
86
–
6
7
10
20
40
80
8
9[g]
[a] Unless otherwise specified, all reactions were carried out at 0.5 m in 9b with 1.1 equiv of ethyl diazoacetate
at 258C for 24 h and reached 100% conversion. nd=not determined. The catalyst was prepared by Procedure
A in Scheme 2. [b] Yield from 9b. recrystallization of the crude product was performed with a mixture of
CH2Cl2 and hexanes and after two crops were taken, the mother liquor was purified by silica gel chromatogra-
phy to give the remainder of 10b. [c] Determined by HPLC on a Chiralcel OD-H column. [d] Determined by
the weight of the combined 10b from the first and second crops and from the mother liquor. [e] Determined
after the first and second crops were combined with the 10b recovered from the mother liquor. [f] Reaction
went to 44% conversion. [g] Reaction performed with 1.0m in imine and went to 68% conversion.
3788
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3785 – 3803

Catalytic Asymmetric Aziridination
FULL PAPER
Table 5. Relative rate study in toluene, CH2Cl2 and CCl4.[a]
Entry
Solvent
t [min]
Conversion [%][b]
1
2
CH2Cl2
toluene
15
15
15
38
40
56
0
3CCl
4
4
CH2Cl2
3
50
66
5
6
toluene
CCl4
30
3
55
0
7
8
9
CH2Cl2
toluene
CCl4
CH2Cl2
toluene
CCl4
2Cl2
toluene
CCl4
60
60
60
200
200
200
20[c]
20[c]
20[c]
59
75
79
90
99
95
84
98
98
Figure 1. Plot of optical purity of ligand versus product in the reaction of
&
^
imine 9b; : VAPOL as ligand, : VANOL as ligand.
10
11
12
13CH
14
15
mulas C46H29BO3 and C52H34B2O5. Catalyst structures consis-
tent with these formulas are the borate 18a for the former
and either of the pyroborates 16a or 17a for the latter. The
assignment of the pyroborate as the major catalyst species
was made possible by a catalyst mixture prepared from p-
tolylborate. The major species in this mixture was deter-
[a] Unless otherwise specified, all reactions were carried out at 0.5m in
imine at 258C with 1.1 equiv of ethyl diazoacetate and 2 mol% catalyst
which was prepared by Procedure A in Scheme 2 with the exception that
the catalyst was prepared in toluene at 808C. [b] Determined by
1H NMR on the crude reaction mixture by the relative integration of 10b
and 9b. [c] Reaction with 10 mol% catalyst.
1
mined by integration of the methyl signals in the H NMR
to have two p-methylphenol units per VAPOL ligand and
the minor species only had one p-methylphenol per
VAPOL. The high resolution mass spectrum of the mixture
of the two species generated from p-tolylborate had ions for
the molecular formulas C47H31BO3 and C54H38B2O5 consis-
tent with the structures 18b and 16b/17b, respectively.
The next question to be addressed is what is the active
catalyst in the asymmetric catalytic aziridination reaction,
comings were first clearly defined by Kagan.[13] This tech-
nique was applied to the reaction of the imine 9b with ethyl
diazoacetate in carbon tetrachloride at 258C with 10 mol%
catalyst. The catalyst was prepared by a variation of Proce-
dure A (Scheme 2) which involves heating three equivalents
of triphenylborate with either the VANOL or VAPOL
ligand in carbon tetrachloride at 808C for 1 h and then re-
moving the volatiles under high vacuum at 808C for 30 min.
As shown by the plot in Figure 1, the relationship between
the optical purity of each ligand and the product is essential-
ly linear. Although, an observed linearity does not disprove
the association of the boron with two or more ligands, none-
theless, the results of these experiments suggest that it is
likely that only one molecule of the ligand is involved in the
active catalyst.
Spectroscopic analysis of the catalyst prepared by the
original Procedure A shown in Scheme 2 revealed the pres-
ence of two species. The bay region proton in the VAPOL
ligand (Hb in 12, Scheme 3) is a convenient spectroscopic
handle for probing the number of catalyst species that are
generated because this proton (d=9.77 ppm) is significantly
deshielded relative to the rest of the aromatic protons. The
catalyst prepared from VAPOL and B(OPh)3 by Procedure
A
A gives two species in a ratio that ranged from 3:1 to 5:1
(over 5 runs) with the bay proton doublet for the minor spe-
cies at d=9.51 ppm and a doublet for the major species at
d=9.22 ppm (see Figure 2 for spectrum of catalyst generat-
ed by Procedure E). The high resolution mass spectrum of
the catalyst mixture gives two ions with the molecular for-
Scheme 3. The B1 and B2 catalysts generated from VAPOL and triaryl
borates.
Chem. Eur. J. 2008, 14, 3785 – 3803
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3789

W. D. Wulff et al.
the borate 18 (B1 catalyst), the pyroborate 16/17 (B2 cata-
lyst) or both? In the preparation of the catalyst by Proce-
dure A we had seen variation in the ratio of B2 to B1 from
3:1 to 5:1 but these different ratios gave the same asymmet-
ric induction in the aziridination reaction. Thus procedures
were sought that could selectively produce either the B1 or
B2 catalyst or at least produce substantially enriched sam-
ples of each. The data in Table 6 summarize the asymmetric
aziridination of imine 9b with catalysts prepared from the
VANOL and VAPOL ligands with five different procedures
which give B2/B1 ratios ranging from 1:10 to 20:1. At a B2/
B1 ratio of 1:10 the aziridine 10b is obtained in 50% ee
(entry 1) whereas, with a B2:B1 ratio of 20:1, 10b is formed
in 91% ee (entry 6). These results clearly reveal that the py-
roborate species B2 is responsible for the high asymmetric
induction observed in the aziridination reaction and that the
B1 species is not the active catalyst as might have been an-
ticipated. Interestingly, the differences in induction between
the B2 and B1 forms of VANOL are not large (Table 6,
entry 7 vs 8 and 9). Furthermore, the ratio of VANOL de-
rived B2 to B1 catalyst could not be driven past 2:1 even
with conditions (Procedure F) that gave a 20:1 selectivity
for the VAPOL ligand. The B2 and B1 forms of VANOL
formed in oven and flame-dried Schlenk flasks and employs
freshly distilled and dried solvents. Thus, it was considered
most likely that the H2O comes from the commercially
available B
and most samples of commercially available B
partially hydrolyzed and contain free phenol. The purity of
(OPh)3 obtained from a variety of suppliers was never
A
ACHTREUNG
AHCTREUNG
B
ACHTREUNG
higher than 85%. Thus, perhaps it would be possible to in-
crease the proportion of the B2 catalyst if H2O were added
during catalyst preparation. The results of such an investiga-
tion are summarized in Table 7. First, it can be seen that
without H2O, catalyst formation can be driven to a 10.6:1
ratio in favor of B2 by increasing the amount of B(OPh)3 to
H
5 equiv (entries 1–3). As expected, the addition of H2O
leads to an increase in the B2/B1 ratio (entries 4 and 5 vs
entry 2) but the addition of too much water (1.5 equiv) leads
to substantial amounts of unreacted VAPOL (entry 6). With
the addition of a given quantity of H2O, the conversion of
VAPOL is higher when the catalyst is prepared at 808C
rather than at 558C (entries 4 and 5 vs 7 and 8). After much
experimentation, the optimal conditions for effecting high
conversion and high selectivity for the B2 catalyst includes
the use of 4 equiv of B(OPh)3 and 1 equiv of H2O and cata-
C
1
were also tentatively assigned by H NMR and high resolu-
tion mass spectrum.
The formation of a pyroborate from VAPOL and B-
lyst preparation at 808C (entry 11). This is Procedure F and
as can be seen from entry 6 in Table 6 the catalyst prepared
with this procedure gives the highest induction observed for
aziridine 10b in CH2Cl2.
A
tion requires an equivalent of H2O. The original procedure
for the preparation of the aziridination catalyst (Procedure
A, Scheme 2) involves moisture-free conditions and is per-
Procedure F gives a 19.6:1 mixture of B2 to B1 catalysts
and gives the highest induction (91% ee) of any of the other
procedures compared in Table 6, however, since this proce-
dure calls for the use of 4 equiv
of
B(OPh)3 it is not clear
A
whether a 19.6:1 mixture of B2
to B1 in the absence of any
Table 6. Effect of B2:B1 ratio on the induction in the aziridination of imine 9b.[a]
extra B(OPh)3 might actually
U
give higher than 91% ee. To
probe for the effects of excess
B(OPh)3, a series of aziridina-
T
tions were performed with the
imine 9 f in which the catalyst
was prepared with various
Entry Ligand
Cat prep[b] B2/B1[c] Yield 10b [%][d] ee 10b [%][e] cis/trans 10b[f] Yield 14/15[f]
1
2
3
4
5
6
7
8
9
(S)-VAPOL
B
D
A
C
E
1:10
1:4
4.5:1
8:1
11:1
20:1
1:8
1.7:1
1.8:1
2.1:1
47
66
8389
80
75
67
81
81
82
77
50
72
nd
16:1
>30:1
nd
>50:1
ꢀ33:1
nd
50:1
nd
100:1
nd
21
3
nd
4
2
nd
13
amounts of excess B(OPh)3 up
G
(S)-VAPOL
(S)-VAPOL
(S)-VAPOL
(S)-VAPOL
(S)-VAPOL
(S)-VANOL
(R)-VANOL
(S)-VANOL
(R)-VANOL
to 30 equiv and the results are
summarized in Table 8. The
control reaction reveals that
89
91
91
84
88
93nd
91
3mol% B (OPh)3 will catalyze
G
F
B
this reaction in toluene to the
tune of 26% conversion in 24 h
at room temperature (entry 8).
However, the reaction with a
A[g]
C
10
F[g]
5
[a] Unless otherwise specified, all reactions were carried out at 0.5 m in 9b with 1.1–1.2 equiv of ethyl diazo-
acetate at 258C for 24 h. nd=not determined. [b] Procedure A is shown in Scheme 2. Procedure B involves
catalyst preparation from 1 equiv of BH3·Me2S and 1 equiv of phenol in toluene at 1008C. Procedure C in-
volves catalyst preparation from 2 equiv BH3·Me2S, 3equiv of phenol and 1 equiv of H 2O in toluene at 1008C.
catalyst prepared from B(OPh)3
E
and VAPOL does not show any
drop off in asymmetric induc-
tion until a 10-fold excess of B-
Procedure D involves syringe pump addition of 1.5 equiv B
ACHTREUNG
1008C. Procedure E involves syringe pump addition of VAPOL to a solution of 5 equiv B
ACHTREUNG
A
808C. Procedure F: See entry 11 in Table 7. [c] Determined by 1H NMR. [d] Isolated yield after chromatogra-
phy on silica gel. [e] Determined by HPLC on a Chiralcel OD-H column. [f] Determined by 1H NMR on the
crude reaction mixture. [g] 5 mol% catalyst is used. The product is ent-10b.
preparation (the error in the %
ee measurements is Æ2%).
3790
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3785 – 3803

Catalytic Asymmetric Aziridination
FULL PAPER
Table 7. Optimization of the preparation of the catalyst.[a]
Table 8. Effect of excess triphenylborate on the asymmetric induction.[a]
Entry
T [8C]
B
N
H2O
[equiv]
B2/B1/VAPOL[b]
A
ACHTREUNG
1
2
355
4
5
6
7
8
9
10
11
12[c]
13[d]
14[e]
15
55
55
2
30
0
0
4.5:1:1.2
.0 7.5:1:3
10.6:1:0.1
5
55
55
55
80
80
80
80
80
80
80
80
80
30.5 .8:1:7.5 13
31.0
31.5
30.5 .9:1:0.8 13
31.0
4
4
4
4
4
Entry
B
(OPh)3 [mol%] Conv. [%][b] Yield 10 f [%][c] ee 10 f [%][d]
A
.8 12.4:1:3
9.9:1:19.1
1
2
1.5
2
26
98
25
82
83
85
12.5:1:0.4
32.5
4
5
97
82
88
0
13.2:1:0.0
13.4:1:0.0
19.6:1:<0.1
11.8:1:0.03
11.8:1:0.02
12.8:1:0.0
14.6:1:0.06
3100
5
85
–
88
–
0.5
1.0
1.0
1.0
1.0
1.0
100
100
100
85
85
84
89
87
83
6
10
30
7
8[e]
326
4
5
[a] The catalyst is prepared by heating (S)-VAPOL (1 mol%) with the
requisite number of equivalents of B(OPh)3 in toluene at 808C for 1 h,
AHCTREUNG
[a] The catalyst is prepared by heating the VAPOL ligand with the indi-
cated amount of B(OPh)3 and H2O in CH2Cl2 at 558C (or in toluene at
and then removal of the volatiles under vacuum (0.1 mm Hg) at 808C for
30 min. Unless otherwise specified, all reactions were carried out at 0.5 m
of imine in toluene at 258C with 1.1 equiv of ethyl diazoacetate for 24 h.
[b] Determined by 1H NMR on the crude reaction mixture and the rela-
tive integration of 10 f and 9 f. [c] Isolated yield after chromatography on
ACHTREUNG
808C) for 1 h and then exposure to high vacuum (0.1 mm Hg) for 0.5 h at
558C (or 808C). Unless otherwise specified, a newly opened bottle of B-
(OPh)3 was used. [b] Determined by 1H NMR. [c] A one year old bottle
of B
ACHTREUNG
silica gel. [d] Determined by HPLC on
[e] VAPOL was not used. The catalyst consisted only of 3mol%
(OPh)3.
a Chiralcel OD-H column.
year old bottle of B
ACHTREUNG
tor. [e] A three year old bottle of B
on the bench.
A
B
ACHTREUNG
Thus, it is concluded that Procedure F represents the opti-
mal procedure both in terms of B2/B1 ratio and in terms of
induction in the aziridination reaction.
reacted VAPOL (d=9.77) and the presence of the species
B1 (d=9.51) and the B2 species (d=9.22). In principle it
should be possible to distinguish between the linear pyrobo-
rate 16a and the cyclic pyroborate 17a on the basis of the
11B NMR since 16a should have two different boron signals
whereas, 17a is symmetrical and should give a single boron
signal. However, the boron spectrum from the catalyst pre-
pared by Procedure F (entry b in Figure 2) is expected to be
complicated by the fact that Procedure F utilizes excess B-
In an effort to further support the structures of the B1
species 18a and the B2 species 16a or 17a and in an effort
to distinguish between the latter, the 11B NMR spectrum of
the catalyst was taken and this is shown in Figure 2. Entry a
1
is the H NMR spectrum of the catalyst prepared by Proce-
dure F and entry b is the corresponding 11B NMR spectrum.
The 1H NMR spectrum reveals the presence of some un-
A
U
Figure 2. a) 1H NMR (CDCl3, 500 MHz) spectrum of VAPOL catalyst prepared from 4 equiv B
(OPh)3 and 1 equiv H2O in toluene at 808C (Procedure
G
F). b) 11B NMR spectrum corresponding to 1H NMR spectrum a. c) 1H NMR spectrum of VAPOL catalyst prepared from 2 equiv BH3·Me2S, 3equiv
1
PhOH and 1 equiv H2O in toluene at 808C (Procedure C). d) 11B NMR spectrum corresponding to H NMR spectrum c.
Chem. Eur. J. 2008, 14, 3785 – 3803
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3791

W. D. Wulff et al.
to be removed under high vacuum. Therefore, a different
procedure for the generation of the catalyst was employed
(Procedure C, Table 6) which utilizes a combination of
BH3·Me2S and phenol where any unreacted boron can be re-
matography, a 46% recovery of (S)-VAPOL with >99% ee
along with a 49% yield of the EDA adduct 21. The same re-
action with 1.2 equivalents of EDA gives only the EDA
adduct 21 in 98% yield. The EDA adduct 21 can be recycled
to optically pure (S)-VAPOL via a Curtius rearrangement[16]
or by samarium diiodide reduction (Scheme 5).[17] After hy-
1
moved after the catalyst is prepared. The H NMR of the
catalyst prepared from this method is shown in entry c in
Figure 2 and is essentially the
same as that for Procedure F
(entry a, Figure 2). However,
the 11B NMR spectrum (entry
d, Figure 2) is quite different
and shows two partially re-
solved peaks that appear to be
of roughly equal intensity. On
this basis we tentatively as-
signed the major species pres-
ent in the mixture as the linear
pyroborate 16a.
At this point it is not known
whether the pyroborate catalyst
16 can function as a mono- or
bidentate Lewis acid. Many ex-
amples of pyroborates are
known, however, only one ex-
ample is known in which both
of the borons form Lewis acid–
Lewis base complexes
(Scheme 4). The bis-catechol
Scheme 5. Recovery of VAPOL ligand.
pyroborate 19 has been characterized as its complex with
acetate ion by X-ray analysis in the solid state.[14] An at-
tempt to prepare a related acetate complex with a pyrobo-
rate derived from a chiral diol failed.[14] The only pyroborate
that we are aware of that contains a chiral diol is compound
20 which had been reported from the reaction of BINOL
with boric acid.[15] The pyroborate 20 was used in the resolu-
tion of BINOL but its properties as a chiral Lewis acid cata-
lyst have not been explored.
drolysis of 21, carboxylic acid 22 is treated with diphenyl-
phosphoryl azide (DPPA) and triethylamine and the result-
ing acyl azide is rearranged to an isocyanate. Trapping the
isocyanate with H2O gives a carbamate that decarboxylates
to give a hemiaminal that hydrolyzes to (S)-VAPOL. How-
ever, some of the acyl azide is trapped intramolecularly by
the phenol to give lactone 23. The overall result is a mixture
of free (S)-VAPOL (56%) and lactone 23 (34%). Although
lactone 23 can be recycled to ethyl ester 21, a more efficient
method for the liberation of VAPOL is the direct reduction
of 21 with samarium diiodide[17] which gives (S)-VAPOL in
91% yield and 99.8% ee.
The data outlined in Table 6 reveals that, at least for the
reaction of imine 9b with ethyl diazoacetate, the optimal
catalyst is that prepared by heating the VAPOL or VANOL
ligand with 4 equiv of B(OPh)3 and 1 equiv of H2O at 808C
G
(Procedure F). Although the present study finds that tolu-
ene is the solvent of choice for this reaction, the scope was
explored in CH2Cl2 as well as toluene since CH2Cl2 is the
solvent in which these aziridination reactions were originally
examined.[7b,c] The scope of the catalytic asymmetric aziridi-
nation reaction with catalysts generated from both the
VANOL and VAPOL ligands with Procedure F was exam-
ined in CH2Cl2 with 12 different imines and the results are
shown in Table 9. The general finding in CH2Cl2 is that
imines from aromatic aldehydes typically give higher asym-
metric inductions than those from aliphatic aldehydes. In ad-
dition the catalyst from VANOL generally gives higher
asymmetric induction than that from VAPOL. Averaged
Scheme 4. Known examples of pyroborates.
The VAPOL can be recovered from the reaction in high
optical purity, however, usually part or all of the VAPOL is
recovered as the ethyl diazoacetate (EDA) adduct 21.[7f] The
ratio of VAPOL 12 to the VAPOL-EDA adduct 21 that is
recovered at the end of the reaction depends on the amount
of excess EDA that is used in the reaction. For example,
with 1.1 equivalents of EDA, the reaction performed with
the catalyst prepared by Procedure F (Table 7, entry 11,
5 mol% catalyst) gave, after purification by silica gel chro-
3792
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3785 – 3803

Catalytic Asymmetric Aziridination
FULL PAPER
Table 9. Aziridination of imine 9 in methylene chloride with both VANOL and VAPOL catalysts from Proce-
trast to the VANOL catalyst
which only gives on average a
0.4% ee increase in toluene
versus CH2Cl2. Interestingly,
since as discussed above it was
found that the VANOL catalyst
gave on average 8% ee higher
inductions than VAPOL in
CH2Cl2, the increases for
VAPOL in toluene cancels out
the advantage that VANOL
had in CH2Cl2 such that in tolu-
ene, the differences in asym-
dure F.[a]
Entry Series
R
Ligand
Yield 10 [%][b] ee 10 [%][c] cis/trans[d]
Yield
14+15 [%][d]
1
a
2
3
b
4
5
c
6
7
d
8
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
76
732913:1
67
77
56
57
80
82
40
89
8
26:1
5
1-naphthyl
Ph
91
91
85
88
88
93
75
85
84
92
61
86
77
88
ꢀ33:1
100:1
10:1
2
5
10
14
8
o-MeC6H4
p-MeC6H4
11:1
ꢀ50:1
>100:1
2.0:1
2.2:1
20:1
34:1
13:1
34:1
5:1
2
9
10
13
13
2
12
11
5
2
5
7
10
[e]
e
o-BrC6H4
41
71
81
11
12
f
p-BrC6H4
metric
induction
between
13
14
61[f]
76[g]
42[h]
51[i]
8386
88
VANOL and VAPOL is only
1.2% ee averaged over the 12
substrates in favor of VANOL
and this is less than the error
for these measurements (Æ2%
ee). Although the VANOL cat-
alyst has nearly equal asymmet-
ric inductions on average in tol-
uene and CH2Cl2, a significant
difference is seen in the yields
for the VANOL catalyst with
an average of 4% higher yields
in toluene versus CH2Cl2. Fur-
thermore, the VANOL catalyst
gives on average 6% higher
yields than the VAPOL catalyst
in toluene. Similar amounts of
the side products 14 and 15 are
observed with both the VAPOL
g
p-NO2C6H4
p-MeOC6H4
3,4-(OAc)2C6H3
n-propyl
15
16
h
6:1
>100:1
>100:1
17
18
i
91
738:1
81
76
83
19
20
24
55
76
15
j
14:1
17
<1
4
5
6
21
22
ꢀ50:1
ꢀ50:1
ꢀ16:1
>100:1
k
cyclohexyl
tert-butyl
80
23
24
66[j]
85
73
84
l
[a] Unless otherwise specified, all reactions were carried out at 0.5m in imine in methylene chloride with
1.2 equiv of ethyl diazoacetate and 5 mol% catalyst at 258C for 24 h and went to 100% conversion. The cata-
lyst was prepared from either (S)-VAPOL or (R)-VANOL and 4 equiv of B(OPh)3 and 1 equiv of H2O in tolu-
U
ene at 808C for 1 h and then all volatiles were removed at 808C under high vacuum (0.1 mm Hg) for 0.5 h.
The reaction with (R)-VANOL gives the enantiomer of 10 as drawn in Scheme 2. [b] Isolated yield after chro-
matography on silica gel. [c] Determined by HPLC with either a Chiralcel OD-H column or Chiralcel OD
1
column. [d] Determined by H NMR on the crude reaction mixture. [e] Reaction time is 48 h. [f] 81% conver-
sion. [g] 97% conversion. [h] 70% conversion. [i] 81% conversion. [j] 87% conversion.
over all 12 substrates, VANOL gives 8% ee higher induction
than VAPOL which is significantly higher than the Æ2%
error for these measurements. In addition, the VANOL cat-
alyst gives an 8% higher yield than the VAPOL catalyst
when the yield differences are averaged over all 12 sub-
strates. This figure is curious since the VANOL catalyst
gives on average a 3% higher yield of the side-products 14
and 15 than the VAPOL catalyst. Since the reactions in
Tables 9–11 go to completion (except where indicated),
there must obviously be other side-products formed that do
not elute from silica gel under conditions in which the aziri-
dines are mobile. There is not a significant difference be-
tween VANOL and VAPOL on the cis/trans selectivity.
A survey of the catalytic asymmetric aziridination reac-
tion with VANOL and VAPOL catalysts prepared by Proce-
dure F with the same 12 substrates in toluene is summarized
in Table 10. A common feature of toluene and CH2Cl2 is the
difference between imines from aliphatic versus aromatic al-
dehydes. In both solvents, the inductions are generally in the
90s for aromatic substrates and in the 80s for aliphatic sub-
strates. There are, however, some significant differences be-
tween the two solvents. The first is that asymmetric induc-
tions for the VAPOL catalyst are higher in toluene than in
CH2Cl2. Averaged over the 12 substrates, the VAPOL cata-
lyst gives 7% ee higher inductions in toluene. This is in con-
and VANOL catalyst in toluene. As in CH2Cl2, there does
not seem to be a difference between VANOL and VAPOL
on the cis/trans ratios in toluene.
The optimization of the asymmetric induction for a few of
the substrates was undertaken by examining their aziridina-
tion reactions in toluene at 08C and the results are present-
ed in Table 11 along with the data at room temperature for
comparison. The substrate that had the greatest response to
lowering the temperature from ambient to 08C was the p-ni-
trophenyl imine 9g. The asymmetric induction increased
from 79 to 95% ee with the VAPOL catalyst, but interest-
ingly, only a small increase was noted for the VANOL cata-
lyst. The imines from cyclohexane carboxaldehyde and piva-
ladehyde gave small increases in asymmetric induction for
the VAPOL catalyst but not for the VANOL catalyst. Final-
ly, for the imine of n-butyraldehyde, small increases in in-
duction were seen for both the VAPOL and VANOL cata-
lysts and significant increases in the yield were also seen.
As revealed by the data in Tables 10–11, excellent asym-
metric inductions can be generally achieved for the aziridi-
nations of benzhydryl imines in toluene with catalysts gener-
ated from both the VAPOL and VANOL ligands. The asym-
metric inductions are generally in the low to mid 90s for ar-
omatic imines and from the mid 80s to the low 90s for ali-
Chem. Eur. J. 2008, 14, 3785 – 3803
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3793

W. D. Wulff et al.
Table 10. Aziridination of imine 9 in toluene with both VANOL and VAPOL catalysts from Procedure F.[a]
phatic imines. All 12 of the
benzhydryl aziridines examined
in Tables 9–11 are solids and
can be very readily recrystal-
lized. The optical purity of all
12 of the aziridines 10a–10l
could be significantly enhanced
Entry Series
R
Ligand
Yield 10 [%][b] ee 10 [%][c] cis/trans[d] Yield 14+15 [%][d]
1
a
2
3
b
4
5
c
6
7
d
8
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
76
80
82
87
6391
67
4:1 933
9351:1
<1
1-naphthyl
Ph
2
94
93100:1
10:1
90
92
94
82
1.9:1
90
94
79
89
86
87
89
93
81
77
ꢀ50:1
14
12:1
ꢀ50:1
ꢀ50:1
1.6:1
<1
2
o-MeC6H4
p-MeC6H4
11
<1
2
by
a
single recrystallization
80
79[e]
37
from either hexanes or from an
ethyl acetate/hexanes mixture
and the data is shown in
Table 12. Remarkably, the opti-
cal purity of all but two of the
aziridines could be increased to
ꢀ99% ee with a single recrys-
tallization and with good recov-
ery. The two exceptions were
the optical purities of o-bromo-
phenyl substituted aziridine 10e
which could be enhanced from
85 to 98.6% ee and the n-
propyl aziridine 10j which
could be enhanced from 86 to
96.6% ee with one recrystalliza-
tion.
9
10
10
[f]
e
o-BrC6H4
4382
78[e]
86
24
11
12
20:1
ꢀ20:1
15:1
100:1
6:1
<1
14
<1
<1
23
<1
6
<1
7
19
<1
6
f
p-BrC6H4
13
14
79[g]
86
g
p-NO2C6H4
p-MeOC6H4
3,4-(OAc)2C6H3
n-propyl
15
16
51[e,h]
61
h
34:1
17
18
87
84
40
54
100:1
ꢀ100:1
14:1
i
19
20
j
14:1
21
22
7381
79
ꢀ50:1
ꢀ50:1
100:1
ꢀ100:1
k
cyclohexyl
tert-butyl
82
87
85
23
24
72[i]
89
<1
4
l
[a] Unless otherwise specified, all reactions were carried out at 0.5m in imine in toluene with 1.2 equiv of ethyl
diazoacetate and 5 mol% catalyst at 258C for 24 h and went to 100% conversion. The catalyst was prepared
from either (S)-VAPOL or (R)-VANOL and 4 equiv of B(OPh)3 and 1 equiv of H2O in toluene at 808C for
N
1 h and then all volatiles were removed at 808C under high vacuum (0.1 mm Hg) for 0.5 h. The reaction with
(R)-VANOL gives the enantiomer of 10 as drawn in Scheme 2. [b] Isolated yield after chromatography on
silica gel. [c] Determined by HPLC with either a Chiralcel OD-H column or Chiralcel OD column. [d] Deter-
mined by 1H NMR on the crude reaction mixture. [e] Solvent is 4:1 toluene:CH2Cl2. [f] Reaction time is 48 h.
[g] 95% conversion. [h] 73% conversion. [i] 93% conversion.
Conclusion
In conclusion, the catalytic
asymmetric aziridination (AZ)
reaction
of
N-benzhydryl
Table 11. Effect of temperature on the aziridination of imine 9 in toluene with catalysts prepared by Proce-
dure F.[a]
imines with ethyl diazoacetate
is a broadly applicable method
for the asymmetric preparation
of aziridines. The scope of the
reaction includes electron rich
and electron poor aromatic
benzhydryl imines and primary,
secondary and tertiary aliphatic
benzhydryl imines. A set of 12
different imines were screened
with catalysts prepared from B-
Entry
Series
R
Ligand
T
[8C]
Yield
ee
cis/
Yield
10 [%][b]
10 [%][c]
trans[d]
14+15 [%][d]
1
2
3
4
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
25
25
0
79[e]
86
90
79
89
95
15:1
100:1
33:1
<1
<1
<1
<1
g
p-NO2C6H4
0
9391300:1
5
6
7
8
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
25
25
0
40
54
81
77
86
8433
14:1
14:1
25:1
7
19
13
j
n-propyl
cyclohexyl
tert-butyl
54[f]
6:10[f]
0
3
AHCTREUNG
9
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
25
25
0
7381
79
70
ꢀ50:1
ꢀ50:1
33:1
<1
6
13
5
10
11
12
82
85
82
k
l
VANOL and VAPOL catalysts
gave essentially the same pro-
file of asymmetric inductions
across the entire screen with %
ee values ranging from the mid
80s to the mid 90s for all sub-
strates. The utility is further en-
hanced by the fact that the opti-
cal purity of nearly all of the 12
aziridine products could taken
to greater than 99% ee with a
0
81
100:1
13
14
15
16
(S)-VAPOL
(R)-VANOL
(S)-VAPOL
(R)-VANOL
25
25
0
72[g]
89
87
85
100:1
ꢀ100:1
<1
4
75[f]
58[f]
4:1 933
83100:1
3
0
<2
[a] Unless otherwise specified, all reactions were carried out at 0.5m in imine in toluene with 1.2 equiv of ethyl
diazoacetate and 5 mol% catalyst at 258C for 24 h and went to 100% conversion. The catalyst was prepared
from either (S)-VAPOL or (R)-VANOL and 4 equiv of B(OPh)3 and 1 equiv of H2O in toluene at 808C for
G
1 h and then all volatiles were removed at 808C under high vacuum (0.1 mm Hg) for 0.5 h. The reaction with
(R)-VANOL gives the enantiomer of 10 as drawn in Scheme 2. [b] Isolated yield after chromatography on
silica gel. [c] Determined by HPLC with either a Chiralcel OD-H column or Chiralcel OD column. [d] Deter-
mined by 1H NMR on the crude reaction mixture. [e] 95% conversion. [f] 10 mol% catalyst; reaction time is single recrystallization.
48 h. [g] 93% conversion.
3794
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3785 – 3803

Catalytic Asymmetric Aziridination
FULL PAPER
Table 12. Enhancement of the optical purity of aziridines 10 by recrystal-
Melting points were determined on a Thomas Hoover capillary melting
point apparatus. IR spectra were taken on a Nicolet IR/42 spectrometer.
1H NMR and 13C NMR were recorded on a Varian 300 MHz or VXR-
500 MHz instrument in CDCl3 unless otherwise noted. CDCl3 was used
as the internal standard for both 1H NMR (d=7.24) and 13C NMR (d=
77.0). Low-resolution mass spectra and elemental analysis were per-
formed in the Department of Chemistry at Michigan State University.
Analytical thin-layer chromatography (TLC) was performed on Silicycle
silica gel plates with F-254 indicator. Visualization was by short wave
(254 nm) and long wave (365 nm) ultraviolet light, or by staining with
phosphomolybdic acid in ethanol or with potassium permanganate.
Column chromatography was performed with silica gel 60 (230–450
mesh).
lization.[a]
Entry Series
R
ee 10 [%] ee 10 [%] Yield 10 [%][b]
before
recryst
1st Crop
1st Crop
1
2
3
4
5
6
7
8
a
b
c
d
e
f
g
h
i
1-naphthyl
Ph
89
94
91
94
85
94
94.5
87
92.5
86
8399.1
87
99.9
99.4
99.374
99.2
98.6
99.4
99.7
99.9
99.0
96.6
55
62
o-MeC6H4
p-MeC6H4
o-BrC6H4
p-BrC6H4
p-NO2C6H4
p-MeOC6H4
3,4-(OAc)2C6H3
n-propyl
80
65
76
74
81
67
40
HPLC analyses were carried out using a Varian Prostar 210 Solvent De-
livery Module with a Prostar 330 PDA Detector and a Prostar Worksta-
tion. Chiral HPLC data for the aziridines were obtained using a Chiralcel
OD-H column, others were obtained using a Chiralcel OD column. Opti-
cal rotations were obtained on a Perkin–Elmer 341 polarimeter at a
wavelength of 589 nm (sodium D line) using a 1.0 decimeter cell with a
total volume of 1.0 mL. Specific rotations are reported in degrees per
decimeter at 208C and the concentrations are given in gram per 100 mL
in CH2Cl2 unless otherwise noted.
9
10
11
12
j
k
l
cyclohexyl
tert-butyl
80
99.7
76
[a] Recrystallization from boiling hexanes or hexanes/ethyl acetate.
[b] Yield based on scalemic aziridine. The cis/trans ratio was ꢀ50:1 in
each case.
General procedure for the preparation of the aldimines 9: All liquid al-
dehydes were distilled before use and all the solid aldehydes were used
as purchased from Aldrich. All imines 9 could be purified by recrystalli-
zation except imine 9j which was a liquid at room temperature and was
used in the aziridination reaction without further purification. The reac-
tion time for the formation of imine 9h was 24 h.
The catalyst generated from B(OPh)3 and both the
A
VANOL and VAPOL ligands contains two species and
1H NMR and mass spectral evidence suggests that the major
species contains one molecule of the ligand and two boron
atoms (B2 catalyst) and the minor species contains one mol-
ecule of the ligand and one boron atom (B1 catalyst). A
series of new protocols for catalyst generation were devel-
oped which allows for the selective generation of the B2
and B1 catalysts in ratios that range from 20:1 to 1:10. With
the aid of these enriched samples of each catalyst, it is re-
vealed that the B2 catalyst is the active catalyst in the AZ
reaction of the VAPOL series as the B2 catalyst gives much
higher asymmetric induction and higher rates than the B1
catalyst. In the VANOL series the difference in the two cat-
alysts is greatly reduced. The structure of the B2 catalyst is
tentatively assigned as a pyroborate in which the VAPOL
ligand forms a borate ester with one of the boron atoms and
the other boron remains as an acyclic borate ester with two
phenol groups. Pyroborate esters of chiral diols are not
common and at this point it is not clear whether an unsym-
metrical chiral pyroborate will function as a monodentate or
bidentate Lewis acid or both.
N-Benzylidene-1,1-diphenylmethanamine (9b): MgSO4 (4 g, 33.3 mmol)
and dried CH2Cl2 (40 mL) were added to a flame-dried 100 mL round-
bottom flask filled with argon. This was followed by the addition of di-
phenylmethanamine (3.46 g, 18.9 mmol). After stirring for 5 min, benzal-
dehyde (2 g, 18.9 mmol) was added. The reaction mixture was stirred for
15 h at room temperature. Thereafter, the reaction mixture was filtered
through Celite; the Celite bed was washed with CH2Cl2 (15 mLꢁ3). The
filtrate was then concentrated by rotary evaporation and placed under
high vacuum (0.1 mm Hg) for 5 min to give crude imine 9b as an off-
white solid. Recrystallization (ethyl acetate/hexanes 1:9) afforded 9b
(3.9 g, 14.5 mmol, 77%) as a white solid. M.p. 99–1018C (lit.[19] 98–
1008C); 1H NMR (CDCl3, 300 MHz): d=8.46 (s, 1H), 7.20–7.90 (m,
15H), 5.64 ppm (s, 1H); 13C NMR (CDCl3, 75 Hz): d=160.48, 143.64,
136.07, 130.47, 128.24, 128.19, 128.15, 127.40, 126.69, 77.62 ppm.
N-(1-Naphthylidene)-1,1-diphenylmethanamine (9a):[7b] Recrystallization
(ethyl acetate/hexanes 1:5) afforded 9a in 85% isolated yield as a white
solid. M.p. 1058C; 1H NMR (CDCl3, 300 MHz): d=9.06 (d, J=7 Hz,
1H), 9.00 (s, 1H), 7.84–7.91 (m, 3H), 7.18–7.55 (m, 12H), 5.62 ppm (s,
1H); 13C NMR (CDCl3, 75 MHz): d=160.91, 144.02, 131.24, 129.83,
128.57, 128.48, 127.66, 127.22, 126.97, 126.02, 125.15, 124.71, 79.33 ppm.
N-(2-Methylbenzylidene)-1,1-diphenylmethanamine (9c):[7a] Recrystalli-
zation (ethyl acetate/hexanes 1:5) afforded 9c in 92% isolated yield as
white crystals. M.p. 99–1008C; 1H NMR (CDCl3, 300 MHz): d=8.67 (s,
1H), 7.93(d, J=7 Hz, 2H), 7.10–7.40 (m, 12H), 5.52 (s, 1H), 2.48 ppm
(s, 3H); 13C NMR (CDCl3, 75 MHz): d=159.65, 144.09, 137.87, 134.18,
130.81, 130.22, 128.39, 127.60, 126.88, 126.03, 78.74, 19.64 ppm.
Experimental Section
N-(4-Methylbenzylidene)-1,1-diphenylmethanamine (9d):[20] Recrystalli-
zation (ethyl acetate/hexanes 1:5) afforded 9d in 79% isolated yield as
white crystals. M.p. 73–748C; 1H NMR (CDCl3, 300 MHz): d=8.45 (s,
1H), 7.80 (d, J=8 Hz, 2H), 7.26–7.48 (m, 12H), 5.64 (s, 1H), 2.44 ppm
(s, 3H); 13C NMR (CDCl3, 75 MHz): d=160.67, 143.98, 141.01, 133.88,
129.21, 128.41, 128.38, 127.66, 126.89, 76.57, 21.50 ppm; IR (thin film): n˜
= 3026s, 2853s, 1639vs, 1599s, 1452s, 1383s, 1030s, 700scmÀ1; MS: m/z
(%): 285 (14) [M]+, 168 (16), 167 (100), 152 (27), 76 (9); elemental analy-
sis calcd (%) for C21H19N: C 88.38, H 6.71, N 4.91; found: C 88.23, H
6.88, N 4.82.
N-(2-Bromobenzylidene)-1,1-diphenylmethanamine (9e):[21] Recrystalli-
zation (ethyl acetate/hexanes 1:5) afforded 9e in 81% isolated yield as a
white solid. M.p. 113–1148C; 1H NMR (CDCl3, 300 MHz): d=8.86 (s,
1H), 8.27 (dd, J=8, 2 Hz, 1H), 7.27–7.61 (m, 13H), 5.71 ppm (s, 1H);
General: All experiments were performed under an argon atmosphere.
Flasks were flame-dried and cooled under argon before use. Dichlorome-
thane and carbon tetrachloride were distilled from calcium hydride under
nitrogen. Toluene, diethyl ether and THF were distilled from sodium
under nitrogen. Hexanes and ethyl acetate were ACS grade and used as
purchased. Reagents were purified by simple distillation or recrystalliza-
tion with simple solvents. Ethyl diazoacetate and triphenylborate were
used as purchased from Aldrich. VAPOL and VANOL were purified by
column chromatography with 3:1 hexanes/dichloromethane. All aldi-
mines were synthesized by a known procedure[18] and purified by recrys-
tallization from hexanes or from a mixture of hexanes/ethyl acetate. Azir-
idines were purified by column chromatography with hexanes/ethyl ace-
tate and further purified by recrystallization from hexanes/ethyl acetate
if needed.
Chem. Eur. J. 2008, 14, 3785 – 3803
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3795

W. D. Wulff et al.
13C NMR (CDCl3, 75 MHz): d=159.80, 143.58, 132.96, 131.89, 129.23,
128.47, 127.61, 127.53, 127.06, 78.06 ppm; IR (thin film): n˜ = 3061m,
3026m, 1631s, 1493s, 1028s, 756scmÀ1; MS: m/z (%): 351 (4, 81Br) [M]+,
349 (5, 79Br) [M]+, 165 (100), 152 (53), 151 (84), 88 (52); elemental analy-
sis calcd (%) for C20H16BrN: C 68.58, H 4.60, N 4.00; found: C 68.39, H
4.73, N 3.93.
N-(4-Bromobenzylidene)-1,1-diphenylmethanamine (9 f):[22] Recrystalli-
zation (ethyl acetate/hexanes 1:5) afforded 9 f in 70% isolated yield as a
white solid. M.p. 96–978C; 1H NMR (CDCl3, 300 MHz): d=8.28 (s, 1H),
7.64 (d, J=7 Hz, 2H), 7.47 (d, J=7 Hz, 2H), 7.15–7.35 (m, 10H),
5.23ppm (s, 1H); 13C NMR (CDCl3, 75 MHz): d=159.51, 143.62, 131.75,
129.84, 128.46, 127.60, 127.05, 77.85 ppm.
N-(4-Nitrobenzylidene)-1,1-diphenylmethanamine (9g):[22] Recrystalliza-
tion (ethyl acetate/hexanes 1:1) afforded 9g in 80% isolated yield as an
off-white solid. M.p. 132–1348C (lit.[18] 134–1358C; 1H NMR (CDCl3,
300 MHz): d=8.52 (s, 1H), 8.31 (d, J=8 Hz, 2H), 8.08 (d, J=8 Hz, 2H),
7.30–7.40 (m, 10H), 5.76 ppm (s, 1H); 13C NMR (CDCl3, 75 MHz): d=
158.51, 143.14, 129.10, 128.57, 127.55, 127.28, 123.80, 78.09 ppm.
solve the two reagents and this was followed by the addition of water
(0.9 mL, 0.05 mmol). The Teflon valve was closed and the flask was
heated at 808C for 1 h. The threaded Teflon valve was opened to gradu-
ally apply high vacuum (0.1 mm Hg) and to remove the solvent. The
vacuum is maintained for a period of 30 min at a temperature of 808C.
The flask was then filled with argon and the catalyst mixture was allowed
to cool to room temperature.
To the flask containing the catalyst was first added aldimine 9b (271 mg,
1 mmol) and then dry toluene (2 mL). Upon addition of the imine and
solvent the reaction mixture turned a yellow color. Ethyl diazoacetate
(124 mL, 1.2 mmol) was added via syringe and the Teflon valve was
closed and the reaction mixture was stirred at room temperature for
24 h. Immediately upon addition of ethyl diazoacetate the reaction mix-
ture became an intense yellow and the formation of bubbles from the re-
lease of nitrogen was noted. The mixture was then diluted with hexanes
(15 mL) and transferred to a 100 mL round-bottom flask. The reaction
flask was rinsed twice with dichloromethane (5 mL) and the rinse was
added to the round-bottom flask. Rotary evaporation of the solvent fol-
lowed by exposure to high vacuum (0.1 mm Hg) for 5 min gave the crude
aziridine as an off-white solid. The conversion was determined from the
1H NMR spectrum of the crude reaction mixture by integration of the
aziridine ring methine protons relative to either the imine methine
proton or the proton on the imine carbon. The cis/trans ratio was deter-
mined on the crude reaction mixture to be ꢀ100:1 by 1H NMR integra-
tion of the ring methine protons for each aziridine. The cis (J=7–8 Hz)
and trans (J=2–3Hz) coupling constants were used to differentiate the
two isomers. The yields of the acyclic enamine products (14b, 15b) were
N-(4-Methoxybenzylidene)-1,1-diphenylmethanamine (9h): Recrystalliza-
tion (ethyl acetate/hexanes 1:5) afforded 9h in 85% isolated yield as
white crystals. M.p. 108–1098C (lit.[18] 108–1098C); H NMR (CDCl3,
300 MHz): d=8.34 (s, 1H), 7.78 (d, J=8.8 Hz, 2H), 7.28–7.40 (m, 10H),
6.91 (d, J=8.8 Hz, 2H), 5.55 (s, 1H), 3.82 ppm (s, 3H); 13C NMR
(CDCl3, 75 MHz): d=160.01, 144.11, 129.99, 128.37, 127.67, 126.85,
113.88, 77.77, 55.32 ppm; IR (thin film): n˜
= 2849m, 1632s, 1493m,
1028m, 756scmÀ1; MS: m/z (%): 301 (16) [M]+, 168 (10), 167 (100), 164
(41), 152 (22), 76 (11); elemental analysis calcd (%) for C21H19NO: C
83.69, H 6.35, N 4.65; found: C 83.60, H 6.35, N 4.52.
1
determined from the H NMR spectrum of the crude reaction mixture by
integration of the N-H proton of the enamine relative to the aziridine
ring methine protons with the aid of the isolated yield of the cis-aziri-
dine: <1% yield of 14b and 2% yield of 15b. Purification of the crude
aziridine by chromatography (35 mmꢁ400 mm column) on silica gel with
an eluent mixture of ethyl acetate/hexanes 1:9 gave the pure aziridine
10b in 87% isolated yield (310.2 mg, 0.87 mmol). The optical purity of
10b was determined to be 93% ee by HPLC analysis (Chiralcel OD-H
column, hexanes/2-propanol 90:10, 222 nm, flow rate 0.7 mLminÀ1). Re-
tention times: tR =4.44 min (major enantiomer) and tR =8.18 min (minor
enantiomer). Spectral data for (2S,3S)-10b: Rf =0.3(ethyl acetate/hex-
anes 1:9); 1H NMR (CDCl3, 500 MHz): d=7.69 (d, J=7 Hz, 2H), 7.57
(d, J=7 Hz, 2H), 7.49 (d, J=7 Hz, 2H), 7.41 (t, J=7 Hz, 2H), 7.33 (m,
5H), 7.25 (m, 2H), 4.08 (s, 1H), 4.00 (m, 2H), 3.30 (d, J=7 Hz, 1H),
2.76 (d, J=7 Hz, 1H), 1.03ppm (t, J=7 Hz, 3H); 13C NMR (CDCl3,
125 Hz): d=167.65, 142.48, 142.35, 135.00, 128.43, 127.74, 127.71, 127.49,
127.35, 127.27, 127.17, 77.64, 60.48, 47.98, 46.34, 13.88 ppm; IR (thin
film): n˜ = 3030m, 2981m, 1737s, 1600s, 1200s, 1097scmÀ1; MS: m/z (%):
357 (<1) [M]+, 190 (100), 167 (60), 117 (34); elemental analysis calcd
(%) for C24H23NO2: C 80.84, H 6.48, N 3.92; found: C 80.92, H 6.70, N
3.88; [a]2D3 = À41.0 (c = 1.0, CH2Cl2) on 99.4% ee material (HPLC);
white solid: m.p. 128–1298C on 99.4% ee material.
4-((Benzhydrylimino)methyl)-1,2-phenylene diacetate (9i):[7b] Recrystalli-
zation (ethyl acetate/hexanes 1:5) afforded 9i in 66% isolated yield as
white crystals. M.p. 138–1398C; 1H NMR (CDCl3, 300 MHz): d=8.37 (s,
1H), 7.77 (d, J=2 Hz, 1H), 7.68 (dd, J=8, 2 Hz, 1H), 7.38 (d, J=8 Hz,
4H), 7.33 (t, J=8 MHz, 4H), 5.62 (s, 1H), 7.24 (m, 3H), 2.30 (s, 3H),
2.29 ppm (s, 3H); 13C NMR (CDCl3, 75 MHz): d=168.22, 168.02, 158.85,
144.07, 143.59, 142.44, 135.16, 128.50, 127.68, 127.11, 126.99, 123.60,
122.88, 77.62, 20.70, 20.64 ppm; IR (thin film): n˜ = 1775s, 1640scmÀ1
;
MS: m/z (%): 387 (10) [M]+, 167 (100); elemental analysis calcd (%) for
C24H21NO4: C 74.46, H 5.47, N 3.62; found: C 74.17, H 5.66, N 3.58.
N-Butylidene-1,1-diphenylmethanamine (9j):[7a] Crude product obtained
as a light yellow oil in 74% yield. 1H NMR (CDCl3, 500 MHz): d=7.84
(t, J=5 Hz, 1H), 7.1–7.4 (m, 10H), 5.35 (s, 1H), 2.33 (dt, J=7.5, 5 Hz,
2H), 1.60 (q, J=7.5 Hz, 2H), 0.95 ppm (t, J=7.5 Hz, 3H).
N-(Cyclohexylmethylene)-1,1-diphenylmethanamine (9k):[19] Recrystalli-
zation (ethyl acetate/hexanes 1:5) afforded 9k in 74% isolated yield as
an off-white solid. M.p. 49–518C (lit.[2] 48–498C; 1H NMR (CDCl3,
300 MHz): d=7.59 (d, J=5.5 Hz, 1H), 7.00–7.60 (m, 10H), 5.21 (s, 1H),
2.20 (brs, 1H), 1.10–1.90 ppm (m, 10H); 13C NMR (CDCl3, 75 MHz): d=
169.51, 144.41, 128.73, 127.97, 127.20, 78.35, 43.91, 30.13, 26.41,
25.82 ppm.
N-(2,2-Dimethylpropylidene)-1,1-diphenylmethanamine (9l):[7b] Recrys-
tallization (ethyl acetate/hexanes 1:9) afforded 9l in 35% isolated yield
as white crystals. M.p. 51–51.58C; 1H NMR (CDCl3, 300 MHz): d=7.85
(s, 1H), 7.49 (d, J=7 Hz, 4H), 7.44 (t, J=7 Hz, 4H), 7.34 (t, J=7 Hz,
2H), 5.50 (s, 1H), 1.27 ppm (s, 9H); 13C NMR (CDCl3, 75 Hz): d=
171.48, 144.23, 128.25, 127.44, 126.68, 77.36, 36.38, 26.94 ppm; IR (thin
film): n˜ = 1666scmÀ1; MS: m/z (%): 251 (<1) [M]+, 167 (100); elemen-
tal analysis calcd (%) for C18H21N: C 86.08, H 8.43, N 5.58; found: C
85.82, H 8.58, N 5.53.
Each aldimine 9a–l was subjected to the catalytic asymmetric aziridina-
tion reaction with the procedure described above (catalyst preparation
Procedure F) in four different variations: with catalysts derived from
(R)-VANOL and (S)-VAPOL ligands and with the solvents toluene and
CH2Cl2. The results for all these reactions can be found in Tables 9 and
10. Aldimines 9g, 9j, 9k and 9l were also subjected to the catalytic asym-
metric aziridination reaction at a reaction temperature of 08C and these
results are presented in Table 11.
Optical purity enhancement by recrystallization: The chemically pure
aziridine (2S,3S)-10b (261 mg, 0.73mmol, 94% ee) obtained from
column chromatography was placed in a 100 mL round-bottom flask. An
air condenser with an argon balloon was attached to the round-bottom
flask. A small amount of a 1:9 mixture of EtOAc/hexane (3–5 mL) was
added via syringe and the solvents brought to boil with a heat gun as the
flask was swirled. Additional solvent mixture was added and mixture was
returned to a boil. This process was continued until a clear solution was
obtained (10–30 mL). The flask was then kept in an insulated place un-
touched for 10–15 h, upon which aziridine 10b crystallized out. The first
crop was collected (162 mg, 0.45 mmol, 62% recovery) and determined
General procedure for the catalytic asymmetric aziridination with aldi-
mines: Catalyst preparation Procedure F
A
catalyst was prepared by the following method (Procedure F). A magnet-
ic stir bar was added to a 25 mL pear-shaped flask that had its 14/20 joint
replaced by a high vacuum threaded T-shaped Teflon valve and then the
flask was flame-dried and cooled under argon. To the flask was added
(R)-VANOL (21.9 mg, 0.05 mmol) and triphenylborate (58 mg,
0.2 mmol). Under an argon flow, dry toluene (2 mL) was added to dis-
3796
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3785 – 3803

Catalytic Asymmetric Aziridination
FULL PAPER
to be 99.4% ee by HPLC (see conditions above). A summary of the re-
crystallization of aziridines 10a–10l is presented in Table 12.
7.2 Hz, 2H), 3.93 (s, 1H), 3.17 (d, J=6 Hz, 1H), 2.64 (d, J=6.9 Hz, 1H),
2.28 (s, 3H), 1.00 ppm (t, J=7.1 Hz, 3H); 13C NMR (CDCl3, 75 Hz): d=
167.75, 142.51, 142.41, 136.84, 131.94, 128.41, 127.63, 127.47, 127.31,
127.19, 127.13, 76.57, 60.47, 47.98, 46.29, 21.09, 13.94 ppm; IR (thin film):
n˜ = 3030m, 2980m, 1739s, 1454m, 1197s, 1178s, 1066 mcmÀ1; MS: m/z
(%): 371 (<1) [M]+, 204 (83), 203 (58), 167 (40), 164 (46), 131 (58), 130
(100), 129 (58), 77 (26); elemental analysis calcd (%) for C25H25NO2: C
80.83, H 6.78, N 3.77; found: C 80.67, H 6.50, N 3.66; [a]2D3 =À27.8 (c =
1.0, CH2Cl2) on 99.2% ee material; white solid: m.p. 164–1658C on
99.2% ee material.
Variation of Procedure F: The above procedure for the preparation of
the aziridine 10b was also repeated with a slight modification of the pro-
cedure in which the catalyst solution was transferred to a solution of the
imine and identical results were obtained.
A
1-benzhydryl-3-(naphthalen-1-yl)aziridine-2-carboxylate
(10a):[7b] Imine 9a (321 mg, 1 mmol) was reacted according to the general
Procedure F described above with (R)-VANOL as ligand. Purification by
column chromatography on silica gel (ethyl acetate/hexanes 1:9) gave the
pure aziridine (2S,3S)-10a in 80% isolated yield (325 mg, 0.80 mmol);
cis/trans 51:1. Enamine side products: <1% yield of 14a and 2% yield
of 15a. The optical purity of 10a was determined to be 93% ee by HPLC
analysis (Chiralcel OD-H column, hexanes/2-propanol 99:1, 222 nm, flow
rate 0.7 mLminÀ1). Retention times: tR =32.89 min (major enantiomer)
and tR =25.62 min (minor enantiomer). A single recrystallization (ethyl
acetate/hexanes 1:15) of 89% ee material gave 10a with 55% recovery
and 99.9% ee. Rf =0.25 (ethyl acetate/hexanes 1:9); 1H NMR (CDCl3,
300 MHz): d=8.12 (d, J=7 Hz, 1H), 7.81 (d, J=7 Hz, 1H), 7.70 (m,
4H), 7.58 (d, J=7 Hz, 2H), 7.48 (m, 2H), 7.38 (m, 3H), 7.30 (m, 3H),
7.22 (m, 1H), 4.10 (s, 1H), 3.77 (d, J=7 Hz, 1H), 3.75 (m, 2H), 2.94 (d,
J=7 Hz, 1H), 0.65 ppm (t, J=7 Hz, 3H); 13C NMR (CDCl3, 75 MHz):
d=167.75, 142.45, 142.22, 133.01, 131.38, 130.48, 128.54, 128.48, 127.85,
127.58, 127.14, 127.10, 126.51, 125.82, 125.40, 125.29, 122.93, 77.91, 60.35,
46.36, 45.98, 13.55 ppm; IR (thin film): n˜ = 3030w, 2980w, 1737s, 1598m,
1191scmÀ1; MS: m/z (%): 407 (5) [M]+, 240 (59), 167 (100), 139 (9); ele-
mental analysis calcd (%) for C28H25NO2: C 82.59, H 6.19, N 3.44; found:
C 81.86, H 6.37, N 3.26; [a]2D3 = +16.0 (c = 1.0, CH2Cl2) on 99.9% ee
material; white solid: m.p. 128–1298C on 99.9% ee material.
A
1-benzhydryl-3-(2-bromophenyl)aziridine-2-carboxylate
(10e): Imine 9e (349 mg, 1 mmol) was reacted according to the general
Procedure F described above with (R)-VANOL as ligand, the only differ-
ence being the reaction time which was 48 h for this reaction. Purification
by column chromatography (ethyl acetate/hexanes 1:9) gave the pure
aziridine 10e in 43% isolated yield (188 mg, 0.43mmol); cis/trans
ꢀ100:1. Enamine side products: 11% yield of 14e and 13% yield of 15e.
The optical purity of 10e was determined to be 82% ee by HPLC analy-
sis (Chiralcel OD-H column, 98:2 hexanes/2-propanol, 222 nm, flow rate
0.7 mLminÀ1). Retention times: tR =6.06 min (major enantiomer) and
tR =7.91 min (minor enantiomer). A single recrystallization (1:19 ethyl
acetate) of 85% ee material gave 10e with 65% recovery and 98.6% ee.
Rf =0.33 (ethyl acetate/hexanes 1:9); 1H NMR (CDCl3, 300 MHz): d=
7.64 (d, J=7.8 Hz, 3H), 7.56 (d, J=7.1 Hz, 2H), 7.48 (d, J=8.0 Hz, 1H),
7.22–7.44 (m, 7H), 7.12 (t, J=7.6 Hz, 1H), 3.98 (s, 1H), 3.92 (q, J=
7.2 Hz, 2H), 3.32 (d, J=6.8 Hz, 1H), 2.77 (d, J=7.0 Hz, 1H), 0.94 ppm
(t, J=7.1 Hz, 3H); 13C NMR (CDCl3, 75 MHz): d=167.54, 142.34,
142.11, 134.40, 131.54, 130.78, 128.77, 128.54, 128.49, 127.65, 127.57,
127.18, 126.98, 126.71, 123.22, 76.57, 60.58, 48.77, 45.86, 13.85 ppm; IR
(thin film): n˜ = 1738s, 1199s, 1028m, 74mcmÀ1; MS: m/z (%):437 (<1,
81Br) [M]+, 43 5 (<1, 79Br) [M]+, 270 (22), 268 (31), 167 (100), 165 (50);
elemental analysis calcd (%) for C24H22BrNO2: C 66.06, H 5.08, N 3.21;
found: C 66.01, H 4.98, N 3.06; [a]2D3 =À26.0 (c = 1.0, CH2Cl2) on 98.6%
ee material (HPLC); white solid: m.p. 147–1488C on 98.6% ee material.
A
1-benzhydryl-3-o-tolylaziridine-2-carboxylate
(10c):[7a]
Imine 9c (285 mg, 1 mmol) was reacted according to the general Proce-
dure F described above with (R)-VANOL as ligand. Purification by
column chromatography on silica gel (ethyl acetate/hexanes 1:9) gave
pure aziridine 10c in 67% isolated yield (250 mg, 0.67 mmol); cis/trans
12:1. Enamine side products: 2% yield of 14c and 9% yield of 15c. The
optical purity of 10c was determined to be 90% ee by HPLC analysis
(Chiralcel OD-H column, hexanes/2-propanol 99:1, 222 nm, flow rate
1 mLminÀ1). Retention times: tR =6.02 min (major enantiomer) and tR =
7.47 min (minor enantiomer). A single recrystallization (ethyl acetate/
hexanes 1:19) of 91% ee material gave 10c with 74% recovery and
99.3% ee. Rf =0.33 (ethyl acetate/hexanes 1:9); 1H NMR (CDCl3,
500 MHz): d=7.75 (d, J=7 Hz, 2H), 7.68 (d, J=7 Hz, 1H), 7.65 (d, J=
7 Hz, 2H), 7.45 (t, J=7 Hz, 2H), 7.38 (m, 3H), 7.28 (m, 1H), 7.22 (m,
2H), 7.15 (d, J=7 Hz, 1H), 4.07 (s, 1H), 4.00 (q, J=7 Hz, 2H), 3.34 (d,
J=7 Hz, 1H), 2.86 (d, J=7 Hz, 1H), 2.43(s, 3H), 1.00 ppm (t, J=7 Hz,
3H); 13C NMR (CDCl3, 125 Hz): d=167.80, 142.48, 142.33, 135.90,
133.05, 129.01, 128.41, 128.39, 127.63, 127.43, 127.06, 127.04, 125.26, 77.76,
60.33, 46.81, 45.53, 18.70, 13.73 ppm; IR (thin film): n˜ = 3054m, 2982m,
1740s, 1600m, 1184scmÀ1; MS: m/z (%): 371 (<1) [M]+, 204 (100), 167
(43), 131 (41); elemental analysis calcd (%) for C25H25NO2: C 80.83, H
6.78, N 3.37; found: C 80.84, H 6.94, N 3.64; [a]2D3 = À42.6 (c = 1.0,
CH2Cl2) on 99.3% ee material; white solid: m.p. 164–1658C on 99.3% ee
material.
A
1-benzhydryl-3-(4-bromophenyl)aziridine-2-carboxylate
(10 f):[7a] Imine 9 f (349 mg, 1 mmol) was reacted according to the general
Procedure F described above with (R)-VANOL as ligand. Purification by
column chromatography on silica gel (ethyl acetate/hexanes 1:9) gave
pure aziridine 10 f in 86% isolated yield (373 mg, 0.86 mmol); cis/trans
ꢀ20:1. Enamine side products: 5% yield of 14 f and 9% yield of 15 f.
The optical purity of 10 f was determined to be 94% ee by HPLC analy-
sis (Chiralcel OD-H column, hexanes/2-propanol 98:2, 222 nm, flow rate
1 mLminÀ1). Retention times: tR =5.37 min (major enantiomer) and tR =
13.48 min (minor enantiomer). A single recrystallization (ethyl acetate/
hexanes 1:19) of 94% ee material gave 10 f with 76% recovery and
99.4% ee. Rf =0.33 (ethyl acetate/hexanes 1:9); 1H NMR (CDCl3,
500 MHz): d=7.65 (d, J=7 Hz, 2H), 7.50 (d, J=7 Hz, 2H), 7,29–7.45 (m,
9H), 7.23(t, J=7 Hz, 1H), 4.01 (s, 1H), 4.00 (q, J=7 Hz, 2H), 3.19 (d,
J=7 Hz, 1H), 2.74 (d, J=7 Hz, 1H), 1.07 ppm (t, J=7 Hz, 3H);
13C NMR (CDCl3, 125 MHz): d=167.37, 142.29, 142.12, 134.06, 130.86,
129.52, 128.49, 127.46, 127.40, 127.25, 127.11, 121.31, 77.55, 60.67, 47.31,
46.44, 13.96 ppm; IR (thin film): n˜ = 1734s, 1201s, 1067 mcmÀ1; MS: m/z
(%): 437 (<1, 81Br) [M]+, 43 5 <( 1, 79Br) [M]+, 270 (42, 81Br), 268 (43,
79Br), 167 (100, 81Br), 165 (19, 79Br); elemental analysis calcd (%) for
C24H22BrNO2: C 66.06, H 5.27, N 3.09; found: C 66.06, H 5.08, N 3.21;
[a]2D3 =À12.5 (c = 1.0, CH2Cl2) on 99.4% ee material; white solid: m.p.
155–1578C on 99.4% ee material.
(2S,3S)-Ethyl 1-benzhydryl-3-p-tolylaziridine-2-carboxylate (10d): Imine
9d (285 mg, 1 mmol) was reacted according to the general Procedure F
described above with (R)-VANOL as ligand. The only difference in the
procedure was that a 4:1 toluene/CH2Cl2 (2 mL) solvent system was used
for the reaction. Purification by column chromatography on silica gel
(ethyl acetate/hexanes 1:9) gave the pure aziridine 10d in 79% isolated
yield (293mg, 0.79 mmol); cis/trans ꢀ50:1. Enamine side products: <1%
yield of 14d and 2% yield of 15d. The optical purity of 10d was deter-
mined to be 94% ee by HPLC analysis (Chiralcel OD-H column, hex-
anes/2-propanol 90:10, 222 nm, flow rate 0.7 mLminÀ1). tR =4.29 min
(major enantiomer) and tR =7.60 min (minor enantiomer). A single re-
crystallization (ethyl acetate/hexanes 1:19) of 94% ee material gave 10d
with 80% recovery and 99.2% ee (HPLC). Rf =0.30 (ethyl acetate/hex-
anes 1:9); 1H NMR (CDCl3, 300 MHz): d=7.60 (d, J=7.3Hz, 2H), 7.48
(d, J=7.2 Hz, 2H), 7.13–7.36 (m, 8H), 7.05 (d, J=8 Hz, 2H), 3.95 (q, J=
A
1-benzhydryl-3-(4-nitrophenyl)aziridine-2-carboxylate
(10g):[7a] Imine 9g (316 mg, 1 mmol) was reacted according to the general
Procedure F described above with (R)-VANOL as ligand. The only dif-
ference was that the reaction was carried out at 08C. Purification by
column chromatography on silica gel (ethyl acetate/hexanes 1:5) gave
pure aziridine 10g (371 mg, 0.92 mmol, 93%); cis/trans:100:1. Enamine
side products: <1% yield of 14g and <1% yield of 15g. The optical
purity of 10g was determined to be 93% ee by HPLC analysis (Chiralcel
OD-H column, hexanes/2-propanol 90:10, 222 nm, flow rate
0.7 mLminÀ1). Retention times: tR =8.70 min (major enantiomer) and
tR =11.37 min (minor enantiomer). A single recrystallization (ethyl ace-
Chem. Eur. J. 2008, 14, 3785 – 3803
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3797

W. D. Wulff et al.
tate/hexanes 1:15) of 94.5% ee material gave 10g with 74% recovery and
99.7% ee. Rf =0.3(ethyl acetate/hexanes 1:5);
1H NMR (CDCl3,
was used and the reaction time was 48 h. Purification by column chroma-
tography (ethyl acetate/hexanes 1:19) gave pure aziridine 10j in 60% iso-
lated yield (194 mg, 0.60 mmol); cis/trans 33:1. Enamine side products:
<1% yield of 14j and 4% yield of 15j. The optical purity of 10j was de-
termined to be 83% ee by HPLC analysis (Chiralcel OD-H column, hex-
anes/2-propanol 99:1, 222 nm, flow rate 1 mLminÀ1). Retention times:
tR =3.51 min (major enantiomer) and tR =7.44 min (minor enantiomer).
A single recrystallization (hexanes) of 86% ee material gave 10j with
40% recovery and 96.6% ee. Rf =0.33 (ethyl acetate/hexanes 1:9);
1H NMR (CDCl3, 500 MHz): d=7.49 (d, J=7 Hz, 2H), 7.39 (d, J=7 Hz,
2H), 7.27 (m, 2H), 7.33 (m, 4H), 4.17 (m, 2H), 3.66 (s, 1H), 2.28 (d, J=
7 Hz, 1H), 2.05 (q, J=7 Hz, 1H), 1.52 (m, 1H), 1.45 (m, 1H), 1.25 (t, J=
7 Hz, 3H), 1.10 (m, 1H), 1.05 (m, 1H), 0.74 ppm (t, J=7 Hz, 3H);
13C NMR (CDCl3, 125 MHz): d=169.46, 142.77, 142.42, 128.29, 128.27,
127.82, 127.30, 127.10, 126.94, 77.88, 60.62, 46.62, 43.32, 29.85, 20.26,
500 MHz): d=8.15 (d, J=8 Hz, 2H), 7.63(m, 4H), 7.55 (d, J=8 Hz,
2H), 7.38 (t, J=7 Hz, 2H), 7.29 (m, 3H), 7.23 (t, J=7 Hz, 1H), 4.04 (s,
1H), 3.98 (q, J=7 Hz, 2H), 3.30 (d, J=7 Hz, 1H), 2.84 (d, J=7 Hz, 1H),
1.06 ppm (t, J=7 Hz, 3H); 13C NMR (CDCl3, 125 MHz): d=166.92,
142.49, 142.03, 141.09, 128.74, 128.60, 128.57, 127.64, 127.40, 127.34,
127.02, 123.00, 60.89, 47.02, 46.88, 29.64, 13.96 ppm; IR (thin film): n˜ =
2980w, 1742s, 1605s, 1520s, 1346s, 1340s, 1202scmÀ1; MS: m/z (%): 402 (<
1) [M]+, 167 (100), 165 (12), 152 (8), 89 (3); elemental analysis calcd (%)
for C24H22N2O4: C 71,63, H 5.51, N 6.96; found: C 71.58, H 5.71, N 6.82;
[a]2D3 =+11.2 (c 1.0, CH2Cl2) on 99.7% ee material. White solid: m.p.
139–1408C on 99.7% ee material.
A
(10h): Imine 9h (301 mg, 1 mmol) was reacted according to the general
Procedure F described above with (R)-VANOL as ligand. The silica gel
for column chromatography was pre-conditioned by preparing a slurry in
a 1:9 mixture of Et3N/CH2Cl2 which was loaded into a column, the sol-
vent was drained and then the silica gel was dried by flushing with nitro-
gen for 1 h. The silica gel column was then saturated with a 1:9 mixture
of ethyl acetate/hexane, the crude aziridine was loaded onto the column
and then elution with the same solvent mixture gave the pure aziridine
10h in 61% isolated yield (236 mg, 0.61 mmol); cis/trans 34:1. Enamine
side products: <1% yield of 14h and <1% yield of 15h. The optical
purity of 10h was determined to be 87% ee by HPLC analysis (Chiralcel
OD-H column, hexanes/2-propanol 95:5, 222 nm, flow rate 0.7 mLminÀ1).
Retention times: tR =6.35 min (major enantiomer) and tR =15.00 min
(minor enantiomer). A single recrystallization (ethyl acetate/hexanes
1:25) of 87% ee material gave 10h with 81% recovery and 99.9% ee.
14.21, 13.57 ppm; IR (thin film): n˜ = 3040m, 2959m, 1732s, 1194scmÀ1
;
MS: m/z (%): 323 (2) [M]+, 167 (100), 156 (91), 152 (15), 128 (23), 82
(17); elemental analysis calcd (%) for C21H25NO2: C 77.98, H 7.79, N
4.33; found: C 78.06, H 7.94, N 4.21; [a]2D3 =À112.2 (c = 1.0, CH2Cl2) on
96.6% ee material; ehite solid: m.p. 93–958C on 96.6% ee material.
(2S,3S)-Ethyl 1-benzhydryl-3-cyclohexylaziridine-2-carboxylate (10k):[7a]
Imine 9k (277 mg, 1 mmol) was reacted according to the general Proce-
dure F described above with (R)-VANOL as ligand. The only difference
was that the reaction was carried out at 08C. Purification by column
chromatography on silica gel (ethyl acetate/hexanes 1:15) gave pure aziri-
dine 10k in 81% isolated yield (295 mg, 0.81 mmol); cis/trans 100:1. En-
amine side products: 5% yield of 14k and <1% yield of 15k. The opti-
cal purity of 10k was determined to be 82% ee by HPLC analysis (Chir-
alcel OD-H column, hexanes/2-propanol 99:1, 222 nm, flow 1 mLminÀ1).
Retention times: tR = 3.45 min (major enantiomer) and tR = 6.99 min
(minor enantiomer). A single recrystallization (ethyl acetate/hexanes
1:19) of 83% ee material gave 10k with 80% recovery and 99.1% ee.
Rf =0.2 (ethyl acetate/hexanes 1:15); 1H NMR (CDCl3, 500 MHz): d=
7.45 (d, J=7 Hz, 2H), 7.37 (m, 2H), 7.31 (m, 4H), 7.24 (m, 2H), 4.25 (m,
2H), 3.63 (s, 1H), 2.29 (d, J=7 Hz, 1H), 1.83(dd, J=7, 3Hz, 1H), 1.28
(t, J=7 Hz, 3H), 0.95–1.66 (m, 10H), 0.52 ppm (dq, J=10, 3Hz, 1H);
13C NMR (CDCl3, 125 MHz): d=169.63, 142.72, 142.33, 128.35, 128.30,
128.26, 127.49, 127.06, 126.82, 126.80, 78.18, 60.67, 52.12, 43.39, 36.27,
30.71, 30.11, 25.53, 25.34, 14.27 ppm; IR (thin film): n˜ = 2927 m, 2917 m,
2850 m, 1731s, 1190s, 1180s; MS: m/z (%): 363 (1) [M]+, 196 (100), 167
(64), 102 (18), 95 (29); elemental analysis calcd (%) for C24H29NO2: C
79.44, H 8.07, N 3.64; found: C 79.30, H 8.04, N 3.85; [a]2D3 =À145.2 (c
1.0, CH2Cl2) on 99.1% ee material; white solid: m.p. 165–1668C on
99.1% ee material.
1
Rf =0.2 (ethyl acetate/hexanes 1:9); H NMR (CDCl3, 300 MHz): d=7.63
(d, J=7.3Hz, 2H), 7.51 (d, J=7.3Hz, 2H), 7.15–7.39 (m, 8H), 6.82 (d,
J=8.8 Hz, 2H), 3.97 (q, J=7.2 Hz, 2H), 3.96 (s, 1H), 3.74 (s, 3H), 3.19
(d, J=6.7 Hz, 1H), 2.66 (d, J=6.8 Hz, 1H), 1.03ppm (t, J=7.0 Hz, 3H);
13C NMR (CDCl3, 75 MHz): d=167.80, 158.84, 142.53, 142.38, 128.82,
128.40, 127.46, 127.31, 127.15, 127.05, 113.16, 76.57, 60.45, 55.06, 47.67,
46.26, 13.94 ppm; IR (thin film): n˜ = 3030w, 2934w, 1738s, 1614m, 1516s,
1250s, 1033scmÀ1; MS: m/z (%): 388 (0.9) [M+1]+, 315 (10), 222 (12),
221 (100), 167 (21), 166 (20), 147 (25), 146 (19), 91 (19); elemental analy-
sis calcd (%) for C25H25NO3: C 77.49, H 6.50, N 3.61; found: C 77.67, H
6.63, N 3.58; [a]2D3 =À27.6 (c = 1.0, CH2Cl2) on 99.9% ee material; white
solid: m.p. 136–1378C on 99.9% ee material.
4-((2S,3S)-1-Benzhydryl-3-(ethoxycarbonyl)aziridin-2-yl)-1,2-phenylene
diacetate (10i):[7b] Imine 10i (387 mg, 1 mmol) was reacted according to
the general Procedure F described above with (R)-VANOL as ligand. Pu-
rification by column chromatography on silica gel (ethyl acetate/hexanes
1:2) gave pure aziridine 10i in 84% isolated yield (214 mg, 0.45 mmol);
cis/trans ꢀ100:1. Enamine side products: <1% yield of 14i and <1%
yield of 15i. The optical purity of 10i was determined to be 93% ee by
HPLC analysis (Chiralcel OD column, hexanes/2-propanol 85:15,
222 nm, flow rate 0.7 mLminÀ1). Retention times: tR =28.62 min (major
enantiomer) and tR =25.38 min (minor enantiomer). A single recrystalli-
zation (ethyl acetate/hexanes 1:5) of 92.5% ee material gave 10i with
67% recovery and 99% ee. Rf =0.28 (ethyl acetate/hexanes 1:2);
1H NMR (CDCl3, 300 MHz): d=7.81 (d, J=7 Hz, 2H), 7.45 (d, J=7 Hz,
2H), 7.28 (m, 7H), 7.19 (m, 1H), 7.07 (d, J=9 Hz, 1H), 3.95 (m, 2H),
3.95 (s, 1H), 3.18 (d, J=7 Hz, 1H), 2.68 (d, J=7 Hz, 1H), 2.25 (s, 3H),
2.24 (s, 3H), 0.99 ppm (t, J=7 Hz, 3H); 13C NMR (CDCl3, 75 MHz): d=
168.24, 168.07, 167.45, 142.21, 141.57, 141.35, 133.97, 128.65, 128.55,
127.61, 127.45, 127.30, 127.18, 126.05, 122.78, 122.75, 77.49, 60.89, 47.03,
46.57, 20.64, 13.84 ppm; IR (thin film): n˜ = 3030w, 2980w, 1770w, 1731s,
1600 mcmÀ1; MS: m/z (%): 474 (21) [M+1]+, 306 (12), 195 (10), 167
(100); elemental analysis calcd (%) for C28H27NO6: C 71.02, H 5.75, N
2.96; found: C 71.23, H 5.88, N 2.94; [a]2D3 =À19.7 (c = 1.0, CH2Cl2) on
99% ee material; white solid: m.p. 141–1438C on 99% ee material.
(2S,3S)-Ethyl 1-benzhydryl-3-tert-butylaziridine-2-carboxylate (10l):[7b]
Imine 9l (251 mg, 1 mmol) was reacted according to the general Proce-
dure F described above with (R)-VANOL as ligand. Purification by
column chromatography on silica gel (1:9 ethyl acetate:hexanes) gave
pure aziridine 10l in 89% isolated yield (300 mg, 0.89 mmol); cis/trans
ꢀ100:1. Enamine side products: 4% yield of 14l and <1% yield of 15l.
The optical purity of 10l was determined to be 85% ee by HPLC analysis
(Chiralcel OD-H, hexanes/2-propanol 99:1, 222 nm, flow rate
1 mLminÀ1). Retention times: tR =3.60 min (major enantiomer) and tR =
9.76 min (minor enantiomer). A single recrystallization (ethyl acetate/
hexanes 1:19) of 87% ee material gave 10l with 76% recovery and
99.7% ee. Rf =0.33 (ethyl acetate/hexanes 1:9); 1H NMR (CDCl3,
300 MHz): d=7.67 (d, J=7 Hz, 2H), 7.40 (d, J=7 Hz, 2H), 7.28 (m,
4H), 7.20 (m, 2H), 4.24 (m, 1H), 4.09 (m, 1H), 3.59 (s, 1H), 2.16 (d, J=
7 Hz, 1H), 1.76 (d, J=7 Hz, 1H), 1.29 (t, J=7 Hz, 3H), 0.70 ppm (s,
9H); 13C NMR (CDCl3, 75 MHz): d=169.72, 143.43, 142.07, 128.26,
128.19, 128.17, 127.36, 127.24, 126.83, 79.19, 60.58, 56.07, 43.37, 31.59,
27.39, 14.09 ppm; MS: m/z (%): 338 (14) [M+1]+, 195 (15), 167 (100); el-
emental analysis calcd (%) for C22H27NO2: C 78.30, H 8.06, N 4.15;
found: C 78.27, H 8.27, N 4.13; [a]2D3 =À149.4 (c 1.0, CH2Cl2) on 99.7%
ee material; white solid: m.p. 150–1528C on 99.7% ee material.
A
1-benzhydryl-3-propylaziridine-2-carboxylate
(10j):[7a]
Imine 9j (237 mg, 1 mmol) was reacted according to the general Proce-
dure F described above with (R)-VANOL as ligand. The only differences
were that the reaction was carried out at 08C, 10 mol% catalyst loading
General procedure for the preparation of racemic aziridines, illustrated
for the reaction of imine 9a catalyzed by triphenylborate
3798
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3785 – 3803

Catalytic Asymmetric Aziridination
FULL PAPER
rac-cis-Ethyl
1-benzhydryl-3-(naphthalen-1-yl)aziridine-2-carboxylate
Table 13. Study on nonlinear effects with VAPOL/VANOL-borate cata-
lyst.[a]
(10a): A 25 mL round-bottom flask, flame-dried and cooled under argon,
was fitted with a rubber septum and an argon balloon. To the flask was
added triphenyl borate (29 mg, 0.1 mmol) and imine 9a (321 mg,
1 mmol). Next was added dry CH2Cl2 (2 mL) followed by the addition of
ethyl diazoacetate (119 mL, 1.15 mmol) via syringe. The reaction mixture
was allowed to stir for 24 h and then was diluted with hexanes (15 mL).
Rotary evaporation of the solvent followed by applying high vacuum
(0.1 mm Hg) for 5 min afforded crude aziridine 10a as an off-white solid.
The cis/trans ratio, conversion and enamine side-product amounts were
calculated from the crude 1H NMR spectrum as described in the general
Procedure F above. Purification of this crude aziridine by column chro-
matography (35 mmꢁ400 mm column) on silica gel with ethyl acetate/
hexanes 1:9 gave pure aziridine 10a in 15% isolated yield (61 mg,
0.15 mmol) as a white solid. M.p. 152–1548C; cis/trans 30:1. Enamine
side products: 1.7% yield of 14a and 2.1% yield of 15a. Conversion:
20%. Spectral data for 10a matches that given above. Other Lewis acids
have been examined for the preparation of racemic samples of aziridines
9a–9l utilizing this procedure and the results are presented in Table 2.
Entry Ligand
ee ligand [%] Yield 10b [%][b] ee 10b [%][c]
1
2
3
4
5
6
7
8
9
(R)-VAPOL
20
20
40
84
86
84
23
19
43
(R)-VANOL
(R)-VAPOL
(R)-VANOL
(R)-VAPOL
(R)-VANOL
(R)-VAPOL
(R)-VANOL
(R)-VAPOL
(R)-VANOL
40
60
4
833
82
58
53
76
60
80
87
85
80
>99.5
>99.5
8372
78
88
92
92
10
[a] Reactions were performed with the ligand generated from the proper
mixture of racemic and enantiomerically pure ligand. The catalysts were
prepared from the ligand with 3equiv of B (OPh)3 at 808C according to
Procedure A. [b] Isolated yield after chromatography on silica gel.
[c] Determined by HPLC on a Chiralcel OD-H column.
C
Large-scale catalytic asymmetric aziridinations with low catalyst loading,
a general procedure illustrated for entry 8 in Table 3: The catalyst was
prepared by the following method (Procedure A). A magnetic stir bar
was added to a 25 mL pear-shaped flask that had its 14/20 joint replaced
by a high vacuum threaded T-shaped Teflon valve and then the flask was
flame-dried and cooled under argon. To the flask was added (S)-VAPOL
(54 mg, 0.1 mmol) and triphenylborate (87 mg, 0.3mmol). Under an
argon flow, dry CCl4 (2 mL) was added to dissolve the two reagents. The
Teflon valve was closed and the flask was heated at 558C for 1 h. The
threaded Teflon valve was opened to gradually apply high vacuum
(0.1 mm Hg) to remove the solvent. The vacuum is maintained for a
period of 30 min at a temperature of 558C. The flask was then filled with
argon and the catalyst mixture was allowed to cool to room temperature.
and 80% ee were prepared by mixing the proper amounts of racemic and
the R enantiomer of each ligand. The reaction were worked up and ana-
lyzed according to the procedure described above for the preparation of
aziridine 10b prepared from a catalyst prepared by general Procedure F.
The results are presented in Table 13and plotted in Figure 1.
Characterization of borate and pyroborate species present in the
VAPOL/B(OPh)3 catalyst (Scheme 6)
G
Catalyst preparation Procedure A: A 25 mL pear-shaped flask that had
its 14/20 joint replaced by a high vacuum threaded T-shaped Teflon valve
and which was flame-dried and cooled under argon to room temperature
was equipped with a stir bar and charged with (S)-VAPOL (54 mg,
A 250 mL round-bottom flask was flame-dried, cooled under argon was
charged with imine 9b (10.854 g, 40 mmol) and CCl4 (76 mL). The cata-
lyst prepared above was dissolved in CCl4 (2ꢁ2 mL) and added to the re-
action flask via syringe. Stirring for 5 min gave a light yellow solution. To
this solution was added ethyl diazoacetate 1 (4.6 mL, 46 mmol) via sy-
ringe. Vigorous bubbling was observed after the addition. The solution
was stirred at room temperature for 24 h, and then quenched by the addi-
tion of hexane (80 mL). The mixture was stirred open to air for 15 min,
and then filtered through Celite to remove insoluble solids. The solids
were washed with CH2Cl2 and then the combined filtrate and washings
were evaporated to dryness and the excess EDA was removed under
high vacuum (0.1 mm Hg) to afford an off-white solid. The crude product
was recrystallized from a boiling mixture of hexane and CH2Cl2 98:2
(360 mL). Collection of the first crop afforded aziridine 10b (9.127 g,
64%) with 98.2% ee. The mother liquor was evaporated to dryness and
the residue was recrystallized from a mixture of hexane and CH2Cl2 98:2
(65 mL) to give additional aziridine 10b (2.694 g, 19%) with 50% ee.
The remaining mother liquor was stripped of solvent and the residue pu-
rified by chromatography on silica gel (EtOAc/hexane 1:19 to 1:15) to
give 0.728 g aziridine 10b with 66% ee. The combined yield for aziridine
10b was 12.55 g (35.2 mmol, 88% yield) and the overall ee was deter-
mined to be 86% ee by chiral HPLC analysis on the combined product.
HPLC conditions and spectral data for 10b are as given above under
Procedure F.
0.1 mmol), B
A
purity) and dry CCl4 (2 mL). The Teflon valve was closed to seal the
flask under argon and the mixture was stirred at 808C for 1 h and then a
vacuum (0.1 mm Hg) was carefully applied by partially opening the
Teflon valve to remove the solvent and volatile materials and the result-
ing residue was heated at 808C under high vacuum for 30 min. The cata-
lyst was dissolved in an appropriate deuterated solvent and examined by
1H NMR which indicated that the ratio of the two catalyst species (16a/
18a) prepared by Procedure A ranges from 3–5:1. The bay protons for
these two species were chosen as diagnostic peaks. In [D6]benzene at
300 MHz the bay protons appear at d=9.80 (d, J=8.4 Hz) for 18a and at
d=9.59 (d, J=8.7 Hz) for 16a (Figure 3). In CDCl3 at 300 MHz the bay
protons appear at d=9.59 (d, J=8.4 Hz) for 18a and at d=9.29 (multip-
let) for 16a (Figure 4). In CDCl3 at 500 MHz the bay protons appear at
d=9.51 (d, J=8.4 Hz) for 18a and at d=9.22 (d, J=8.0 Hz) for 16a (see
Figure 2 in text). Mass spectrum of catalyst mixture: low resolution in
FAB, NPOE matrix, relative intensity of 18a is 4.6% and relative intensi-
ty of 16a is 12%. High resolution mass spectrum of 18a: m/z: calcd for
C46H29BO3: 640.2210, found: 640.2214. High resolution mass spectrum of
16a: m/z: calcd for C52H34B2O5: 760.2592, found: 760.2596. The use of
this catalyst (10 mol%) in the aziridination of imine 9b in CH2Cl2 as
shown in Table 6 gives a 83% yield of aziridine 10b with 89% ee and a
Nonlinear studies on the catalysts gen-
erated from VANOL and VAPOL
(Table 13): The reaction of imine 9b
and ethyl diazoacetate 1 was carried
out utilizing a catalyst derived from
either the VANOL or VAPOL ligand
with the catalyst preparation Proce-
dure A described above for the large
scale preparation of aziridine 10b. The
reactions were carried out in carbon
tetrachloride at room temperature for
24 h. The scalemic ligands of 20, 40, 60
Scheme 6. Characterization of borate and pyroborate species present in the VAPOL–B(OPh)3 catalyst.
A
Chem. Eur. J. 2008, 14, 3785 – 3803
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3799

W. D. Wulff et al.
0.1 mmol), phenol (28 mg, 0.3mmol), toluene (2 mL) and BH 3Me2S
(100 mL, 2.0m in toluene, 0.2 mmol). To this solution was added 1.8 mL
H2O (0.1 mmol). The Teflon valve was closed to seal the flask and the
mixture was then heated to 808C for 1 h. The volatiles were removed by
carefully opening the Teflon valve and applying a high vacuum. When
dry, the flask was heated at 808C for 0.5 h under high vacuum (0.1 mm
Hg). After cooling, the flask with the catalyst was moved into a glove
box and the catalyst was dissolved into approximately 0.75 mL CDCl3
and analyzed by NMR. The catalyst prepared by this method typically
gave a ratio of VAPOL/18a/16a 0.1:1.0:8.2 as determined by 1H NMR.
The spectrum shown in Figure 2 has
a small amount of unreacted
VAPOL. The 11B NMR was taken in CDCl3 in a quartz NMR tube and
as indicated in Figure 2 shows two broad peaks at 18.25 and 16.20 ppm
(baseline separation not achieved) relative to BF3·Et2O. The use of this
catalyst (10 mol%) in the aziridination of imine 9b in CH2Cl2 as shown
in Table 6 gives a 80% yield of aziridine 10b with 89% ee.
Catalyst preparation Procedure D: To
a flame-dried Schlenk flask
Figure 3. 1H NMR spectrum (C6D6, 300 MHz) of the VAPOL catalyst
prepared by Procedure A.
cooled under argon was added (S)-VAPOL (54 mg, 0.1 mmol) and dry
toluene (9 mL). The flask was stirred at 1008C under argon protection. A
solution of triphenyl borate (0.15 mmol, 1.5 mL of 1m solution in tolu-
ene) was added via syringe pump to the VAPOL solution over 30 min.
After the addition, the solution was stirred for 40 min, then a vacuum
(0.1 mm Hg) was applied to remove the solvent and volatile materials
and the residue was heated at 1008C under high vacuum for 30 min. The
prepared catalyst was dissolved in CDCl3 and examined by 1H NMR
which indicated that the conversion of VAPOL was > 95% and the
ratio of the two catalyst species (18a/16a) is 4.3:1. The use of this catalyst
(10 mol%) in the aziridination of imine 9b in CH2Cl2 as shown in
Table 6 gives a 66% yield of aziridine 10b with 72% ee and a 16:1 cis/
trans ratio. A 21% yield of 14b and 15b is observed as determined by
1
the crude H NMR spectrum.
Catalyst preparation Procedure E: To a flame-dried Schlenk flask cooled
under argon was added triphenylborate (145 mg, 0.5 mmol) and dry tolu-
ene (9 mL). The flask was stirred at 808C under argon protection. A so-
lution of VAPOL (54 mg, 0.1 mmol) in toluene (1 mL) was added via sy-
ringe pump to the solution of triphenylborate over 30 min. After the ad-
dition, the solution was stirred for 40 min, then a vacuum (0.1 mm Hg)
was applied to remove the solvent and volatile materials and the residue
was heated at 808C under high vacuum for 30 min. The prepared catalyst
was dissolved in CDCl3 and examined by 1H NMR which indicated that
the conversion of VAPOL was 100% and the ratio of the two catalyst
species (18a/16a) is 1:11. The use of this catalyst (10 mol%) in the aziri-
dination of imine 9b in CH2Cl2 as shown in Table 6 gives a 75% yield of
aziridine 10b with 91% ee and a >50:1 cis/trans ratio. A 4% yield of
14b and 15b is observed as determined by the crude 1H NMR spectrum.
Figure 4. 1H NMR spectrum (CDCl3, 300 MHz) of the VAPOL catalyst
prepared by Procedure A.
>30:1 cis/trans ratio. A 3% yield of 14b and 15b is observed as deter-
1
mined by the crude H NMR spectrum.
Catalyst preparation Procedure B: A 25 mL pear-shaped flask that had
its 14/20 joint replaced by a high vacuum threaded T-shaped Teflon valve
and which was flame-dried and cooled under argon to room temperature
was equipped with a stir bar and charged with (S)-VAPOL (54 mg,
0.1 mmol), phenol (9.4 mg, 0.1 mmol), toluene (2 mL) and BH3Me2S
(50 mL, 2.0m in toluene, 0.1 mmol). The Teflon valve was closed and the
sealed flask was heated to 808C for 1 h. The threaded Teflon valve was
slightly opened under high vacuum such that the volatiles were gradually
removed. Once all volatiles were removed, the Teflon valve was opened
all the way and the contents of the flask were exposed to high vacuum
(0.1 mm Hg) at 808C for 0.5 h. After cooling, the flask with the catalyst
was moved into a glove box. The catalyst was dissolved into approxi-
mately 0.75 mL CDCl3 and analyzed by NMR. The catalyst prepared by
this method typically afforded a ratio of VAPOL/18a/16a 1.0:9.9:1.0 as
determined by 1H NMR. 1H NMR (CDCl3, 500 MHz): VAPOL d=9.77
(d, J=8.5 Hz), 18a d=9.51 (d, J=8.5 Hz), 16 d=9.22 (J=8.0 Hz). The
11B NMR was taken in a Norell quartz NMR tube and shows a broad
peak at 19.83ppm relative to BF 3·Et2O in CDCl3. The use of this catalyst
(10 mol%) in the aziridination of imine 9b in CH2Cl2 as shown in
Table 6 (in text) gives a 47% yield of aziridine 10b with 50% ee.
Catalyst preparation Procedure F: To a 25 mL pear-shaped flask that
had its 14/20 joint replaced by a high vacuum threaded T-shaped Teflon
valve and which was oven-dried and cooled under argon to room temper-
ature, was added (S)-VAPOL (54 mg, 0.1 mmol) and triphenyl borate
(116 mg, 0.4 mmol). The mixture was dissolved in dry toluene (2 mL),
followed by the addition of H2O (1.8 mL, 0.1 mmol). All operations were
done under the protection of argon. The flask was sealed under argon by
closing the Teflon valve and the solution was stirred at 808C for 1 h and
then a vacuum (0.1 mm Hg) was applied carefully. Upon removal of sol-
vent, the vacuum was maintained for 30 min while the flask was heated
at 808C. The flask was then filled with argon and allowed to cool down
to room temperature. To this prepared catalyst was added freshly neu-
tralized dry CDCl3 (1 mL, the CDCl3 was passed through a short pipette
column packed with activated basic Al2O3 prior to each NMR experi-
ment. The basic Al2O3 was activated at 1008C under vacuum for 16 h,
then cooled to room temperature and stored under N2). The NMR tube
was flame-dried and cooled to room temperature under the protection of
argon; the catalyst solution was then transferred to the NMR tube via sy-
ringe.The ratio of 16a/18a by 1H NMR is 19.6:1 with no detectable
amount of VAPOL remaining. The use of this catalyst (10 mol%) in the
aziridination of imine 9b in CH2Cl2 as shown in Table 6 gives a 67%
yield of aziridine 10b with 91% ee and ꢀ33:1 cis/trans. A 2% yield of
14b and 15b is observed as determined by the crude 1H NMR spectrum.
Catalyst preparation Procedure C: A 25 mL pear-shaped flask that had
its 14/20 joint replaced by a high vacuum threaded T-shaped Teflon valve
and which was flame-dried and cooled under argon to room temperature
was equipped with a stir bar and charged with (S)-VAPOL (54 mg,
3800
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3785 – 3803

Catalytic Asymmetric Aziridination
FULL PAPER
Characterization of borate and pyroborate species present in the
24) is 3.0:1 by integration of the multiplet at d=7.88 and the doublet at
d=7.79, respectively. The spectrum of the aromatic region is shown in
Figure 5. In a different run the catalyst was obtained with a ratio of 25/24
1.7:1. The use of this catalyst (5 mol%) in the aziridination of imine 9b
in CH2Cl2 as shown in Table 6 gives an 81% yield of aziridine 10b with
88% ee and ꢀ50:1 cis/trans. A 13% yield of 14b and 15b is observed as
VAPOL–B(p-tolyl)3 catalyst
N
Preparation of p-tris-tolylborate: A 250 mL round-bottom flask fitted
with a Dean-Stark trap was charged with boric acid (3.09 g, 50 mmol), p-
cresol (18.9 g, 175 mmol) and 50 mL toluene. The solution was refluxed
under argon for 48 h, and the water produced in the reaction was re-
moved until complete conversion was indicated by the cessation of water
generation. The solvent was removed in vacuo and the crude product was
distilled under high vacuum. The product boiled at 1818C (0.1 mm Hg),
and the viscous oil that was collected solidified upon standing. The prod-
uct was isolated as white solid (13.54 g, 41 mmol, 82%). The 1H NMR of
the product indicated that the isolated p-tris-tolylborate[23] was greater
than 95% chemical purity.
1
determined by the crude H NMR spectrum.
Catalyst preparation Procedure A: A 25 mL pear-shaped flask that had
its 14/20 joint replaced by a high vacuum threaded T-shaped Teflon valve
and which was flame-dried and cooled under argon to room temperature
was equipped with a stir bar and charged with (S)-VAPOL (54 mg,
0.1 mmol), p-tris-tolylborate (133 mg, 0.4 mmol) and dry CH2Cl2 (2 mL).
The Teflon valve was closed to seal the flask under argon and the mix-
ture was stirred at 558C for 1 h. A vacuum (0.1 mm Hg) was gradually
applied by slightly opening the Teflon valve under a high vacuum to
remove the solvent and volatile materials. The residue was heated at
558C under a high vacuum for 30 min. The catalyst was dissolved in
CDCl3 and examined by 1H NMR which indicated that the ratio of the
two catalyst species (16b/18b) is 1.3:1 by integration of the bay protons
and 1.2:1 by integration of the p-tolyl protons. VAPOL was completely
consumed as indicated by the absence of a doublet at d=9.77 ppm.
1H NMR (CDCl3, 300 MHz) for 18b: d=9.63(d, J=8.4 Hz), 2.50 (s), the
relative integration is 1:1.8, respectively. 1H NMR (CDCl3, 300 MHz) for
16b: d=9.35 (multiplet), 2.24 (s), the relative integration is 1:3.5, respec-
tively. Mass spectrum of catalyst mixture: low resolution in FAB, NPOE
matrix, relative intensity of 18b is 11% and the relative intensity of 16b
is 14%. High resolution mass spectrum of 18b: m/z: calcd for
C47H31BO3: 654.2366, found: 654.2376. High resolution mass spectrum of
16b: m/z: calcd for C54H38B2O5: 788.2905, found: 788.2918.
Figure 5. 1H NMR spectrum (CDCl3, 300 MHz) of the VANOL catalyst
prepared by Procedure A.
Catalyst preparation Procedure F: A 25 mL pear-shaped flask that had
its 14/20 joint replaced by a high vacuum threaded T-shaped Teflon valve
and which was flame-dried and cooled under argon to room temperature
was equipped with a stir bar and charged with (S)-VAPOL (54 mg,
0.1 mmol), p-tris-tolylborate (133 mg, 0.4 mmol), 1.8 mL of water
(0.1 mmol) and dry toluene (2 mL). The Teflon valve was closed to seal
the flask under argon and the mixture was stirred at 808C for 1 h. A
vacuum (0.1 mm Hg) was gradually applied by slightly opening the
Teflon valve to remove the solvent and volatile materials and the residue
was heated at 808C under high vacuum for 30 min. The catalyst was dis-
solved in CDCl3 and examined by 1H NMR which indicated that the
ratio of the two catalyst species (16b/18b) is 8.4:1 by integration of the
bay protons and 8.2:1 by integration of the p-tolyl protons. VAPOL was
completely consumed as indicated by the absence of a doublet at d=
9.77 ppm. 1H NMR (CDCl3, 300 MHz) for 18b: d=9.63(d, J=8.4 Hz),
2.50 (s), the relative integration is 1:1.8, respectively. 1H NMR (CDCl3,
300 MHz) for 16b: d=9.35 (multiplet), 2.24 (s), the relative integration is
1:3.4, respectively.
Scheme 7. Characterization of borate and pyroborate species present in
the VANOL–B(OPh)3 catalyst.
G
Catalyst preparation Procedure B: The protocol described for the prepa-
ration of the VAPOL borate complex 18a and VAPOL pyroborate com-
plex 16a with the catalyst preparation Procedure B was followed exactly
except that (S)-VANOL (44 mg, 0.1 mmol) was added in place of
VAPOL. The catalyst made by this method typically afforded a ratio of
VANOL/24/25 0.2:8.0:1.0 as determined by 1H NMR. The aromatic hy-
drogen on the naphthalene ring peri to the oxygen was chosen as the di-
agnostic peak: VANOL d=8.43(m), 24 d=8.07 (d, J=8.0 Hz), 25 d=
8.16 (m). The use of this catalyst (10 mol%) in the aziridination of imine
9b in CH2Cl2 as shown in Table 6 gives a 81% yield of aziridine 10b with
84% ee.
Characterization of borate and pyroborate species present in the
VANOL–B(OPh)3 catalyst (Scheme 7)
A
Catalyst preparation Procedure A: A 25 mL pear-shaped flask that had
its 14/20 joint replaced by a high vacuum threaded T-shaped Teflon valve
and which was flame-dried and cooled under argon to room temperature
was equipped with a stir bar and charged with (S)-VANOL (44 mg,
Catalyst preparation Procedure C: The protocol described for the prepa-
ration of the VAPOL borate complex 18a and VAPOL pyroborate com-
plex 16a with the catalyst preparation Procedure C was followed except
(S)-VANOL (44 mg, 0.1 mmol) was added in place of VAPOL. The cata-
lyst made by this method typically afforded a ratio of VANOL/24/25
<0.1:1.0:1.8. Mass spectrum (EI+, direct probe) m/z (%): 660 (13), 540
(100); HRMS: m/z: calcd for C38H25BO3: 540.1897, found: 540.1896.
HRMS: m/z: calcd for C44H30B2O5: 660.2279, found: 660.2289. The use of
this catalyst (10 mol%) in the aziridination of imine 9b in CH2Cl2 as
shown in Table 6 gives a 82% yield of aziridine 10b with 93% ee.
0.1 mmol), B
A
purity) and dry CCl4 (2 mL). The Teflon valve was closed to seal the
flask under argon and the mixture was stirred at 808C for 1 h. A vacuum
(0.1 mm Hg) was gradually applied by slightly opening the Teflon valve
to remove the volatile materials and the residue was heated at 808C
under high vacuum for 30 min. The catalyst was dissolved in CDCl3 and
examined by 1H NMR which indicated that the ratio of the two catalyst
species (25/24) is 2.8:1 by integration of the multiplet at d=8.16 and the
doublet at d=8.07, respectively. The ratio of the two catalyst species (25/
Catalyst preparation Procedure D: A 25 mL flame-dried Schlenk flask
was cooled under argon to room temperature, equipped with a stir bar
Chem. Eur. J. 2008, 14, 3785 – 3803
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3801

W. D. Wulff et al.
and charged with (S)-VANOL (44 mg, 0.1 mmol) and dry toluene
(9 mL). The flask was stirred at 1008C under argon protection. A solu-
tion of triphenyl borate (0.15 mmol, 1.5 mL of 1m solution in toluene)
was added via syringe pump to the VANOL solution over 30 min. After
the addition, the solution was stirred for 40 min and then a vacuum
(0.1 mm Hg) was gradually applied to remove the solvent and volatile
materials and then the residue was heated at 1008C under high vacuum
for 30 min. The prepared catalyst was dissolved in CDCl3 and examined
by 1H NMR which indicated that the conversion of VANOL was 100%
(d=8.43, doublet) and the ratio of the two catalyst species (25/24) is
1:4.8 by integration of the multiplet at d=8.14 and the doublet at d=
8.04, respectively. The ratio of the two catalyst species (25:24) is 1:5.0 by
integration of the multiplet at d=7.86 and the doublet at d=7.77, respec-
tively. The spectrum of the aromatic region is shown in Figure 6.
flask with a rubber septum and an argon balloon, ethanol (reagent grade,
17 mL, 0.294 mmol) and hexamethylphosphoramide (HMPA, 153 mL,
0.882 mmol) were added via syringe. The SmI2/THF solution
(0.392 mmol, 2.6 mL) was then transferred via syringe to the solution of
21. The reaction mixture was stirred at room temperature for 1 h, during
which time the reaction went to completion (TLC, ethyl acetate/hexanes
1:9). To the reaction flask was then added saturated NaHCO3 solution
(20 mL) and the mixture extracted with ethyl acetate (4ꢁ20 mL). The or-
ganic layers were combined, washed with brine, dried over MgSO4 and
the solvent removed by rotary evaporation to afford the crude VAPOL
ligand 12. The ligand was then purified by column chromatography on
silica gel with a mixture of ethyl acetate/hexanes 1:19, which afforded the
pure (S)-VAPOL product 12 (48 mg, 0.089 mmol, 91%). The optical
purity of the recovered VAPOL was determined to be 99.8% ee by
chiral HPLC analysis (Regis Pirkle Covalent D-Phenylglycine column,
hexanes/2-propanol 75:25, 260 nm, flow rate 2 mLminÀ1). Retention
times: (S)-VAPOL=18.54 min, (R)-VAPOL=12.50 min.
Liberation of the VAPOL ligand via Curtius rearrangement[16]
Hydrolysis of the EDA adduct to acid 22: A 100 mL round-bottom flask
was flame dried and cooled under argon and then the VAPOL–EDA
adduct 21 (232 mg, 0.372 mmol) was introduced into the flask. The
adduct was dissolved in ethanol (20 mL) and then 20% (w/v) aqueous so-
lution of NaOH (20 mL) was added. This resulted in an instant color
change from colorless to intense yellowish green. The reaction mixture
was stirred at room temperature for 1 h. Thereafter, 1n HCl (140 mL)
was added to adjust the pH of the mixture to pH 1, upon which the prod-
uct carboxylic acid 22 precipitated from the reaction mixture. The prod-
uct was isolated by vacuum filtration and then dissolved in ethyl acetate.
The filtrate was extracted once with ethyl acetate (30 mL) and the organ-
ic layers combined, washed with brine (2ꢁ30 mL), dried over MgSO4
and then the solvent was removed via rotary evaporation to afford crude
carboxylic acid product 22 as a yellow solid (217.2 mg, 0.36 mmol, 98%).
1H NMR (CDCl3, 500 MHz): d=9.71 (d, J=8.7 Hz, 1H), 9.31 (d, J=
8.1 Hz, 1H), 7.92 (d, J=7.7 Hz, 1H), 7.82 (d, J=8.7 Hz, 1H), 7.71–7.64
(m, 3H), 7.47–7.64 (m, 7H), 7.43 (s, 1H), 7.04–7.13 (m, 2H), 6.95–7.04
(m, 4H), 6.78–6.86 (m, 4H), 6.51 (brs, 1H), 4.39 (d, J=15.7 Hz, 1H),
4.29 ppm (d, J=15.7 Hz, 1H); 13C NMR (CDCl3, 125 MHz): d=171.64,
154.63, 152.07, 142.32, 140.41, 139.76, 139.19, 135.15, 134.55, 133.10,
132.84, 130.40, 129.61, 129.23, 129.07, 129.02, 128.90, 128.86, 128.77,
128.53, 128.19, 128.08, 127.73, 127.71, 127.62, 127.30, 127.12, 127.05,
126.99, 126.84, 126.73, 126.63, 126.09, 126.03, 123.13, 123.01, 120.74,
119.92, 118.70, 115.25, 67.72 ppm (one sp2 carbon not located).
Figure 6. 1H NMR spectrum (CDCl3, 300 MHz) of the VANOL catalyst
prepared by Procedure D.
General procedure for the recovery and recycling of VAPOL from the
catalytic asymmetric aziridination reaction
Recovery of VAPOL–EDA adduct 21: The aziridination reaction of the
imine 9b (542 mg, 2 mmol) with ethyl diazoacetate 1 (250 mL, 2.4 mmol)
was carried out in toluene, for 24 h at room temperature and with
5 mol% of the (S)-VAPOL–borate catalyst generated according to the
general Procedure F described above. Thus for the preparation of the
Curtius rearrangement[16] of acid 22: A 25 mL round-bottom flask was
flame dried and cooled under argon and then charged with crude carbox-
ylic acid 22 (41.2 mg, 0.069 mmol). The solid was then dissolved by the
addition of toluene (3mL) and DMF (1 mL). This was followed by the
addition of triethyl amine (11 mL, 0.079 mmol) and diphenylphosphoryl
azide (DPPA, 15.7 mL, 0.072 mmol). The flask was then fitted with a
water condenser and an argon balloon and the reaction mixture was re-
fluxed for 3h. After cooling down to room temperature, water (3mL)
was added via syringe and the reaction mixture was refluxed again for
2 h. After cooling to room temperature, 2n HCl (5 mL) and ethyl acetate
(10 mL) were added and the layers separated. The aqueous layer was ex-
tracted with ethyl acetate (2ꢁ10 mL), the organic layers combined,
washed twice with brine, dried over MgSO4 and the solvent evaporated
by rotary evaporation to afford the crude reaction product. This crude
product was then subjected to column chromatography on silica gel with
ethyl acetate/hexanes 1:9 to afford (S)-VAPOL 12 (20.8 mg, 0.039 mmol,
56%, >99% ee) and lactone 23 (13.6 mg, 0.024 mmol, 34%). Spectral
catalyst, (S)-VAPOL (54 mg, 0.1 mmol), B(OPh)3 (116 mg, 0.4 mmol),
G
H2O (1.8 mL, 0.1 mmol) and toluene (2 mL) were heated at 808C for 1 h,
then a vacuum (0.1 mm Hg) was applied carefully. Upon removal of sol-
vent, the vacuum was kept for 30 min with continual heating at 808C.
After the aziridination reaction, the crude reaction mixture obtained was
subjected to separation by column chromatography with
a mixture
EtOAc/hexanes 1:9, which afforded pure aziridine 10b in 73% yield
(522 mg, 1.46 mmol) as well as the VAPOL-EDA adduct 21 in 98% yield
(61 mg, 0.098 mmol). No VAPOL was detected under these reactions
conditions. The Rf values for VAPOL, the aziridine 10b and the
VAPOL–EDA adduct 21 with a mixture of EtOAc/hexanes 1:9 are 0.34,
0.30 and 0.25, respectively. The characterization data for 21 was identical
to that previously reported by our group.[7f] The amount of the EDA 21
that is formed is variable and depends on the amount of excess ethyl di-
azoacetate that is used. For example, if 1.1 equivalents of ethyl diazoace-
tate is used then the adduct 21 is isolated in 49% yield along with a 46%
recovery of unreacted VAPOL that is >99% ee.
Samarium diiodide[17] reduction of EDA adduct 21: A 25 mL round-
bottom flask was flame dried and cooled under argon and charged with
samarium metal (128 mg, 0.85 mmol) and dry THF (5.2 mL). The flask
was then fitted with a rubber septum and an argon balloon. Freshly dis-
tilled diiodomethane (63 mL, 0.784 mmol) was then added via syringe.
The reaction mixture was stirred for 2 h at room temperature to give a
dark blue slurry. To another 25 mL round-bottom flask which had been
flame dried and cooled under argon was added the VAPOL–EDA
adduct 21 (61 mg, 0.098 mmol) and dry THF (1 mL). After fitting the
1
data for 23: H NMR (CDCl3, 500 MHz): d=9.29–9.34 (m, 2H), 7.95–8.0
(m, 2H), 7.79–7.85 (m, 2H), 7.64–7.76 (m, 7H), 7.60 (s, 1H), 7.04–7.10
(m, 2H), 6.89–6.98 (m, 4H), 6.67 (d, J=7.1 Hz, 2H), 6.54 (d, J=7.1 Hz,
2H), 5.25 (d, J=13.5 Hz, 1H), 4.92 ppm (d, J=13.5 Hz, 1H); 13C NMR
(CDCl3, 125 MHz): d=166.31, 154.25, 147.74, 140.45, 140.06, 139.24,
139.14, 135.03, 134.20, 133.47, 133.21, 129.21, 129.18, 129.12, 128.99,
128.94, 128.93, 128.84, 128.67, 128.37, 128.28, 127.83, 127.55, 127.49,
127.40, 127.25, 127.10, 127.01, 126.93, 126.82, 126.74, 122.01, 121.14,
71.29 ppm (four sp2 carbons not located); IR (thin film): n˜ = 3055w,
3802
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3785 – 3803

Catalytic Asymmetric Aziridination
FULL PAPER
2920w, 1761 m, 1641 mcmÀ1; MS: m/z (%): 578 (32) [M]+, 295 (17), 294
(100), 221 (25).
ganometallics 2002, 21, 1540; e) M. F. Mayer, M. M. Hossain, J. Or-
ganomet. Chem. 2002, 654, 202; f) M. Redlich, M. M. Hossain, Tetra-
hedron Lett. 2004, 45, 8987; g) P. Wipf, M. A. Lyon, ARKIVOC
2007, 91.
Ethanolysis of lactone 23: A 25 mL round-bottom flask was flame dried
and cooled under argon and then the crude reaction mixture (12+23)
from the Curtius rearrangement reaction (77.8 mg, 0.0975 mmol (scale
determined from the amount of the original starting material: the carbox-
ylic acid 22)) was added which was dissolved in EtOH (4 mL) and THF
(1.2 mL) to obtain a clear yellow solution. To this mixture was added
K2CO3 (134.7 mg, 0.975 mmol) and the reaction mixture stirred for 6 h to
obtain a brownish green slurry at which point the TLC indicated com-
plete disappearance of the lactone 23. To this solution was then added
2n HCl (5 mL) and diethyl ether (10 mL) and the layers separated. The
aqueous layer was extracted with diethyl ether (2ꢁ10 mL), the organic
layers combined and washed twice with brine, dried over MgSO4 and the
solvent removed by rotary evaporation to give the crude reaction mix-
ture. This was then subjected to purification by column chromatography
on silica gel with ethyl acetate/hexanes 1:19 to afford (S)-VAPOL 12
(26.1 mg, 0.042 mmol, 38%, >99% ee) and the VAPOL-EDA adduct 21
(26.7 mg, 0.05 mmol, 32%).
[9] For other reactions employing chiral catalysts generated from the
VANOL and VAPOL ligands, see: a) J. Bao, W. D. Wulff, A. L.
Rheingold, J. Am. Chem. Soc. 1993, 115, 3814; b) J. Bao, W. D.
Wulff, Tetrahedron Lett. 1995, 36, 3321; c) D. P. Heller, D. R. Gold-
berg, W. D. Wulff, J. Am. Chem. Soc. 1997, 119, 10551; d) S. Xue, S.
Yu, Y. Deng, W. D. Wulff, Angew. Chem. 2001, 113, 23 3 1A; ngew.
Chem. Int. Ed. 2001, 40, 2271; e) C. Bolm, J.-C. Frison, Y. Zhang,
W. D. Wulff, Synlett 2004, 1619; f) G. B. Rowland, H. Zhang, E. B.
Rowland, S. Chennamadhavuni, Y. Wang, J. C. Antilla, J. Am.
Chem. Soc. 2005, 127, 15696; g) D. P. Heller, D. R. Goldberg, H.
Wu, W. D. Wulff, Can. J. Chem. 2006, 84, 1487; h) G. Li, Y. Liang,
J. C. Antilla, J. Am. Chem. Soc. 2007, 129, 5830; i) C. A. Newman,
J. C. Antilla, P. Chen, A. V. Predeus, L. Fielding, W. D. Wulff, J. Am.
Chem. Soc. 2007, 129, 7216; J. E. B. Rowland, G. B. Rowland, E.
Rivera-Otero, J. C. Antilla, J. Am. Chem. Soc. 2007, 129, 12084.
[10] M. Yasuda, S. Yoshioka, S. Yamasaki, T. Somyo, K. Chiba, A. Baba,
Org. Lett. 2006, 8, 761.
[11] S. Nagayama, S. Kobayashi, Chem. Lett. 1998, 685.
[12] a) K. Ishihara, M. Miyata, K. Hattori, T. Tada, H. Yamamoto, J.
Am. Chem. Soc. 1994, 116, 10520; b) M. Periasamy, L. Venkatraman,
S. Sivakumar, N. Sampathkumar, C. R. Ramanathan, J. Org. Chem.
1999, 64, 7643; c) J. P. Cros, Y. Perez-Fuertes, M. J. Thatcher, S. Ari-
mori, S. D. Bull, T. D. James, Tetrahedron: Asymmetry 2003, 14,
1965.
Acknowledgement
This work was supported by a grant from the National Institutes of
Health (GM 63019).
[13] a) C. Girard, H. B. Kagan, Angew. Chem. 1998, 110, 3088; Angew.
Chem. Int. Ed. 1998, 37, 2922; b) H. B. Kagan, T. O. Luukas, in
Comprehensive Asymmetric Catalysis, Vol. 1 (Eds.: E. N. Jacobsen;
A. Pfaltz; H. Yamamoto), Springer, 1999, pp. 101; c) H. B. Kagan,
Synlett 2001, 888.
[14] N. D. Coombs, S. Aldridge, G. Wiltshire, D. L. Kays (nee Coombs),
B. Bresner, L. Ooi, J. Organomet. Chem. 2005, 690, 2725.
[15] Z. Shan, Y. Xiong, W. Li, D. Zhao, Tetrahedron: Asymmetry 1998, 9,
3985.
[1] a) J. C. Sheehan, K. Nakajima, E. Chacko, Heterocycles 1979, 13,
227; b) V. J. Jephcote, D. I. John, D. J. Williams, J. Chem. Soc. Perkin
Trans. 1 1986, 2195.
[2] R. Bartnik, G. Mloston, Synthesis 1983, 924.
[3] For a comprehensive list of citations on the Lewis acid mediated
synthesis of aziridines from imines and diazo compounds, see: A. P.
Patwardhan, V. R. Pulgam, Y. Zhang, W. D. Wulff, Angew. Chem.
2005, 117, 6325; Angew. Chem. Int. Ed. 2005, 44, 6169, and referen-
ces therein.
[16] D. Mirk, S. R. Waldevogel, Tetrahedron Lett. 2004, 45, 7911.
[17] a) K. Kyeuda, J. Inanaga, M. Yamaguchi, Tetrahedron Lett. 1989, 30,
2945; b) T. Hanamoto, N. Shimomoto, T. Kikukawa, J. Inanaga, Tet-
rahedron: Asymmetry 1999, 10, 2951.
[18] D. Armesto, M. J. Ortiz, R. Perez-Ossorio, J. Chem. Soc. Perkin
Trans. 1 1986, 2021.
[19] G. Cainelli, D. Giacomini, A. Trere, P. P. Boyle, J. Org. Chem. 1996,
61, 5134.
[20] S. C. Joshi, P. K. Tikuu, K. N. Mehrutra, Indian J. Chem. Sect B.
1980, 19, 1009.
[21] C. A. Kruter, K. W. Kuntz, C. O. Ozierba, W. A. Wirschun, J. D. Si-
canyi, M. L. Snapper, A. H. Hoveyda, J. Am. Chem. Soc. 1999, 121,
4284.
[22] D. Green, G. Patel, S. Elgendy, J. A. Baban, G. Claeson, V. V.
Kakkar, J. Deadman, Tetrahedron 1994, 50, 5099.
[4] For a review of different catalytic methods for the synthesis of aziri-
dines, see: P. Muller, C. Fruit, Chem. Rev. 2003, 103, 2905.
[5] L. Casarrubios, J. A. Perez, M. Brookhart, J. L. Templeton, J. Org.
Chem. 1996, 61, 8358.
[6] K. G. Rasmussen, K. A. Jorgensen, J. Chem. Soc. Perkin Trans. 1
1997, 1287.
[7] a) J. C. Antilla, W. D. Wulff, J. Am. Chem. Soc. 1999, 121, 5099;
b) J. C. Antilla, W. D. Wulff, Angew. Chem. 2000, 112, 4692; Angew.
Chem. Int. Ed. 2000, 39, 4518; c) C. Loncaric, W. D. Wulff, Org. Lett.
2001, 3, 3675; d) ref. [3]; e) Y. Deng, Y. R. Lee, C. A. Newman,
W. D. Wulff, Eur. J. Org. Chem. 2007, 2068; f) Z. Lu, Y. Zhang,
W. D. Wulff, J. Am. Chem. Soc. 2007, 129, 7185; g) the VANOL and
VAPOL ligands are commmercially available from Aldrich Chemi-
cal Co. and from Strem Chemicals, Inc.
[8] For other examples of catalytic asymmetric aziridination of imines
with diazo compounds with chiral Lewis acids, see: a) ref. [6];
b) K. G. Rasmussen, R. G. Hazall, K. A. Jorgensen, Acta. Chem.
Scand. 1998, 52, 1056; c) K. Juhl, R. G. Hazall, K. A. Jorgensen, J.
Chem. Soc. Perkin Trans. 1 1999, 2293; d) D. Morales, J. Perez, L.
Riera, V. Riera, R. Corzo-Suarez, S. Garcia-Granda, D. Miguel, Or-
[23] T. Colclough, W. Gerrard, M. F. Lappert, J. Chem. Soc. 1955, 907.
Received: October 3, 2007
Published online: February 27, 2008
Chem. Eur. J. 2008, 14, 3785 – 3803
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3803