Synthesis of asymmetric iron-pybox complexes and their application to aziridine forming reactions

Article abstract of DOI:10.1016/j.tetlet.2004.10.047

The synthesis of a series of iron-pybox complexes and their employment in the catalytic asymmetric aziridine forming reaction is presented. When AgSbF₆ is used as an initiator, the i-pr- and t-bu-pybox complexes produce 47% of the cis-aziridine

Full text of DOI:10.1016/j.tetlet.2004.10.047

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 8987–8990
Synthesis of asymmetric iron–pybox complexes and
their application to aziridine forming reactions
Mark Redlich and M. Mahmun Hossain*
Department of Chemistry and Biochemistry, University of Wisconsin-Milwaukee, 3210N. Cramer Street,
Milwaukee, WI 53211-3029, USA
Received 22 September 2004; revised 4 October 2004; accepted 8 October 2004
Available online 22 October 2004
Abstract—The synthesis of a series of iron–pybox complexes and their employment in the catalytic asymmetric aziridine forming
reaction is presented. When AgSbF6 is used as an initiator, the i-pr- and t-bu-pybox complexes produce 47% of the cis-aziridine
in moderate eeÕs with the bulk of side products consisting of the trans-isomer and b-amino-a,b-unsaturated esters (AUEÕs).
Ó 2004 Elsevier Ltd. All rights reserved.
Methods for obtaining asymmetric aziridines by cataly-
sis generally proceed through one of two principal
routes: transfer of a nitrogen group to an olefin, or
transfer of a carbenoid to an imine. While the former
approach is more prevalent in the literature, increasing
attention has been given to the latter in recent years.1
Other recent work by Brookhart and co-workers,2 Gib-
son and co-workers,3 and Nomura and co-workers4 has
suggested that iron–pybox complexes of the type 1 (Fig.
1) might be effective candidates for the asymmetric azir-
idine forming reaction between imines and ethyl
diazoacetate.
O
O
N
1a, R = i-pr
1b, R = t-bu
1c, R = ph
N
N
Fe
R
R
Cl
Cl
Figure 1. Iron–pybox complexes 1.
all the solids were observed to dissolve (Scheme 2).7 In
the cases of the isopropyl and tert-butyl pybox com-
plexes (1a and 1b, respectively), the resultant solution
turned a dark blue-violet color instantly, and all solids
dissolved within a few minutes. In the complexation of
the phenyl pybox ligand, the solution was a deep magen-
ta color, and the magenta solid (1c) began to precipitate
as stirring continued.
The pybox ligands5 were synthesized from the commer-
cially available amino alcohol in three steps.6 2,6-Pyr-
idine dicarboxylic acid was refluxed with thionyl
chloride and isolated. The residue was treated with the
amino alcohol in chloroform at 0°C, followed by in situ
addition of thionyl chloride to yield the pybox ligand
(generally as a hydrochloride salt). Neutralization of
the salt was achieved by stirring a methanolic solution
of the salt with aqueous sodium hydroxide at room tem-
perature for 3days. Recrystallization from ethyl acetate
and pentane yielded the free pybox ligand as long, white
needles (Scheme 1).
The iron complexes exhibited different solubilities
throughout the series: both the isopropyl and tert-butyl
pybox complexes dissolved readily to form solutions in
dichloromethane, whereas the complex 1c did not unless
under high dilution (greater than 1000 times the amount
of solvent required to dissolve compared to the other
The iron–pybox complexes were prepared by stirring a
THF solution of pybox ligand and iron(II) chloride until
1
complexes). H NMR analysis of the crude 1a indicated
a paramagnetic complex, along with resonances corre-
sponding to the uncomplexed ligand. Repeated washing
of the crude solid with diethyl ether removed the uncom-
plexed ligand, and subsequently the NMR spectrum dis-
played only highly shifted resonances appearing as
Keywords: Iron–Pybox; Aziridines; Asymmetric catalysis; Lewis acid.
*
Corresponding author. Tel.: +1 4142296698; fax: +1 4142295530;
e-mail: mahmun@uwm.edu
0040-4039/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.10.047

8988
M. Redlich, M. M. Hossain / Tetrahedron Letters 45 (2004) 8987–8990
R
HO
NH2
SOCl2
reflux
O
O
O
O
N
N
a, R = ipr
b, R = tbu
c, R = Ph
OH
Cl
OH
Cl
O
O
1) SOCl2 , reflux
N
HN
OH
2) aq.NaOH, MeOH
HO
NH
R
R
O
a, R = ipr, 45%
b, R = tbu, 51%
c, R = Ph, 52%
O
N
N
N
R
R
Scheme 1. Synthesis of the pybox ligands.
O
O
O
O
N
N
THF
.
+
FeCl2 4H2O
N
reflux
N
N
N
Fe
R
R
R
R
Cl
Cl
a, R = ipr
b, R = tbu
c, R = Ph
1a, R = ipr, 92%
1b, R = tbu, 92%
1c, R = Ph, 90%
Scheme 2. Synthesis of the iron–pybox complexes.
broad singlets between 60 and 30ppm for the coordi-
nated pybox ligand. The other complexes displayed sim-
ilar behavior when analyzed by NMR. In an attempt to
achieve a higher incidence of ligand coordination, the
reaction was run again in refluxing THF, and allowed
ture revealed a 43% ee for the reaction. After workup,
the reaction yielded 41% of the cis-aziridine, with the
major byproducts being b-amino-a,b-unsaturated esters
(AUEÕs) as a 4:1 mixture (entry 2).
1
to stir for up to 48h. H NMR analysis of the reaction
Different amounts and types of initiator were attempted
(entries 3–6). In each case, NMR analysis of the crude
reaction mixture revealed small amounts of the trans-
isomer in a cis/trans ratio of approximately 85:15.
Increasing the amount of initiator increased the yield
of the reaction, but at the expense of the enantioselectiv-
ity. The best results were obtained with the use of 1equiv
of AgSbF6 (entry 5). The reaction was also attempted in
differing solvents (entries 7–10), and the results are listed
in Table 1. The best overall solvent for the reaction ap-
peared to be dichloromethane, although THF produced
a better yield (38% vs 32%), and nitromethane gave bet-
ter enantioselectivity (49% vs 42%). Results were poor in
both yields and ee categories when methanol was used as
the solvent. A catalyst loading of 5mol% was used for
the remainder of the iron–pybox catalyzed reactions.
There appears to be little correlation between the rela-
tive concentrations of the substrates and either the yield
or the enantioselectivity (entries 11–13). Nor was
enhancement in ee observed in any of the reactions with
the addition of extra equivalents of pybox ligand to the
progress showed no appreciable complexation occurring
after 5min. When the reaction was repeated at 0°C,
again, no change in the reaction profile was observed
by NMR.
The complex 1a was employed in an aziridine forming
reaction of N-benzylideneaniline and ethyl diazoacetate
(EDA), no consumption of the imine was observed by
TLC over 48h. The reaction was ceased, and only start-
ing materials were recovered (Table 1, entry 1). Another
reaction was attempted (Scheme 3); however, before the
iron–pybox complex was added to the imine, it was stir-
red in the presence of 1equiv of AgBF4 for 2h. The solu-
tion was then filtered under nitrogen into the flask
containing the imine, and EDA was added to the solu-
tion. After 48h, imine conversion was no longer detected
by TLC, and the reaction was halted.8 1H NMR analysis
of the crude reaction mixture revealed the cis-aziridine,9
plus small amounts of the trans-isomer in a cis/trans
ratio of 85:15. HPLC analysis of the crude reaction mix-

M. Redlich, M. M. Hossain / Tetrahedron Letters 45 (2004) 8987–8990
8989
Table 1. Iron–pybox catalyzed aziridine forming reactions of imines and ethyl diazoacetate (Scheme 4)
Entry
R
Catalyst
(mol%)
Ag+ Initiator
(mol%)
Equiv. EDAa
Solvent
Yieldb
cis-aziridine (%)
%eec
Combined yield
AUEÕs (%)b
1
Ph
Ph
1a (5)
1a (5)
1a (5)
1a (5)
1a (5)
1a (5)
1a (2)
1a (2)
1a (2)
1a (2)
1a (5)
1a (5)
1a (5)
1a (5)
1a (5)
1a (5)
1a (5)
1b (5)
1c (5)
1a (5)
1b (5)
1c (5)
—
1
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
NO2Me
THF
0
41
49
20
47
54
32
21
38
14
28d
34
42
40
37
30
28
47
15i
25i
39i
9i
—
43
30
32
45
31
42
49
35
20
40
28
31
40
32
15
40
49
5
—
31
2
AgBF4 (5)
AgBF4 (10)
AgOTf (5)
AgSbF6 (5)
AgSbF6 (10)
AgBF4 (2)
AgBF4 (2)
AgBF4 (2)
AgBF4 (2)
AgBF4 (5)
AgBF4 (5)
AgBF4 (5)
AgBF4 (5)
AgBF4 (5)
AgBF4 (5)
AgBF4 (5)
AgSbF6 (5)
AgSbF6 (5)
AgSbF6 (5)
AgSbF6 (5)
AgSbF6 (5)
1
3
Ph
Ph
1
Not detÕd
Not detÕd
Not detÕd
Not detÕd
Not detÕd
Not detÕd
Not detÕd
Not detÕd
Not detÕd
Not detÕd
Not detÕd
Not detÕd
Not detÕd
Not detÕd
Not detÕd
34
4
1
5
Ph
Ph
1
6
1
7
Ph
Ph
1
8
1
9
Ph
Ph
1
10
11
12
13
14e
15f
16g
17h
18
19
20
21
22
1
MeOH
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
Ph
Ph
0.85
1.2
1.5
1
Ph
Ph
Ph
Ph
1
1
Ph
Ph
1
1
Ph
1
8
CHPh2
CHPh2
CHPh2
1
20
28
0
6
1
9
>2
1
a Relative to imine.
b Isolated yield.
c Enantiomeric excess of the cis-aziridine was determined using a (S,S) Whelk-O column. The %ee of the trans-isomer was not determined.
d Yield based on EDA.
e An additional 5mol% pybox ligand added to substrate flask.
f An additional 10mol% pybox ligand added to substrate flask.
g An additional 25mol% pybox ligand added to substrate flask.
h Reaction run at 0°C.
i Incomplete consumption of the imine.
R
N
5mol% 1a
+
5mol% AgBF4
R
N
Ph
N
1, Ag+ initiator
H
H
Ph
+
N2CHCOOEt
N
solvent, rt, 48h
2h
H
H
CO2Et
+
N2CHCOOEt
CH2Cl2, rt, 48h
CO2Et
Scheme 4. Iron–pybox catalyzed aziridine-forming reaction.
Ph
Ph
H
H
Ph
Ph
H
H
N
N
O
N
O
H
CO2Et
+
pected, the formation of AUEÕs is reduced when the
more electron-rich benzhydryl imine is used;1b,c how-
ever, yield of cis-aziridine and ee are both reduced in
those reactions. In addition, unreacted benzhydryl imine
was recovered in all trials where it was the substrate, and
letting the reaction proceed for additional time (72h,
96h) did not improve percent conversion. The best over-
all results came when the tert-butyl pybox catalyst was
used, although results obtained with the isopropyl py-
box catalyst are very similar. It is interesting to note that
in both reaction types, the efficacy of the phenyl pybox
complex is markedly different than either of the other
two complexes. These results, coupled with the differing
physical properties of the phenyl pybox complex, war-
ranted further analysis of the pybox complexes, which
is underway.
H
OEt
OEt
Ph
>10%
31%, in a 4:1 ratio
Scheme 3. Reaction of complex 1a with N-benzylideneaniline and
EDA.
substrate solution (entries 14–16).6b,10 When the reac-
tion was attempted at 0°C, the yield was decreased to
28%, without improvement in the ee (entry 17).
Using the optimal set of conditions from the test reac-
tion, the entire series of pybox complexes were employed
in two aziridine forming reactions with either N-benzyl-
ideneaniline or N-benzylidene-N-(diphenyl-methyl)a-
mine (benzhydryl imine) as the substrate (Scheme 4
and Table 1, entries 18–22). In reactions where N-benzyl-
ideneaniline is the substrate, conversion of the imine
into products is essentially complete, with the AUEÕs
comprising the bulk of non-aziridine product. As ex-
A proposed catalytic cycle for the reaction is given in
Scheme 5. Initiation with 1equiv of Ag+ ion creates an
open site for coordination of the imine to the Lewis acid.

8990
M. Redlich, M. M. Hossain / Tetrahedron Letters 45 (2004) 8987–8990
N
Chem. Scand. 1998, 52, 1056; (h) Rasmussen, K. G.; Juhl,
K.; Hazell, R. G.; Jørgensen, K. A. J. Chem. Soc., Perkin
Trans. 2 1998, 1347; (i) Juhl, K.; Hazell, R. G.; Jørgensen,
K. A. J. Chem. Soc., Perkin Trans. 1 1999, 2293; (j)
Nagayama, S.; Kobayashi, S. Chem. Lett. 1998, 685.
2. Small, B. L.; Brookhart, M.; Bennett, A. M. A. J. Am.
Chem. Soc. 1998, 120, 4049.
3. Britovsek, G. J. P.; Bruce, M.; Gibson, V. C.; Kimberley,
B. S.; Maddox, P. J.; Mastroianni, S.; McTavish, S. J.;
Redshaw, C.; Solan, G. A.; Stro¨mberg, S.; White, A. J. P.;
Williams, D. J. J. Am. Chem. Soc. 1999, 121, 8728.
4. (a) Nomura, K.; Sidokmai, W.; Imanishi, Y. Bull. Chem.
Soc. Jpn. 2000, 73, 599; (b) Imanishi, Y.; Nomura, K. J.
Polym. Sci., Part A: Polym. Chem. 2000, 38, 4613.
5. For review on pybox ligands: see Desimoni, G.; Faita, G.;
Quadrelli, P. Chem. Rev. 2003, 103, 3119.
N
Fe
N
Cl
Cl
Ag+
AgCl
R
N
R
N
N
+
+
N2
H
Ph
H
N
Fe
N
Ph
CO2Et
Cl
N
N
N
R
Fe
N
N
H
6. (a) Nishiyama, H.; Sakaguchi, H.; Nakamura, T.; Hori-
hata, M.; Kondo, M.; Itoh, K. Organometallics 1989, 8,
846; (b) Nishiyama, H.; Kondo, M.; Nakamura, T.; Itoh,
K. Organometallics 1991, 10, 500.
N
R
Fe
N
N
Cl
Ph
Cl
+
CO2Et
Ph
H
N2
+
7. General procedure for the preparation of the iron–pybox
complexes: A 100mL side armed flask was charged with
the 1.5mmol pybox ligand, followed by 1.5mmol iron(II)
chloride tetrahydrate. The solids were dissolved in THF,
and stirred at room temperature for 1–2h. The solvent was
removed under reduced pressure, and the powder was
washed with cold ether to provide near-quantitative yields
of the iron–pybox complexes.
CO2Et
H
N2
O
O
N
N
N
N =
N
N
R
R
8. General procedure for the synthesis of aziridines: A side
armed flask was prepared and charged with 0.1mmol
iron–pybox Lewis acid. The solid was dissolved in 2mL
CH2Cl2. A separate side armed flask was charged with
silver salt initiator and suspended in 3mL CH2Cl2. T he
silver salt suspension was transferred over to the iron–
pybox solution via cannula and stirred in absence of light
for 2h. The solution was then stick-filtered through a
cotton plug to a side armed flask containing 2.0mmol
imine. The diazo compound was added after stirring for
10–15min, and the reaction was allowed to proceed for 1–
2days. The reaction solution was filtered through a plug of
silica and eluted with ether or EtOAc into a 100mL flask.
The solvent was removed by rotary evaporation, and the
crude product was analyzed by HPLC. After HPLC
analysis, the crude product was separated by column
chromatography (0–10% EtOAc/pentane) and the com-
Scheme 5. Proposed mechanism of monochelated iron–pybox com-
plexes.
Attack of the diazo compound occurs enantioselectively;
approach from one side is significantly more hindered by
the isopropyl or tert-butyl side chains. Backside ring clo-
sure and expulsion of a molecule of nitrogen creates the
aziridine ring, which disassociates to reform the active
catalyst.
Further analysis to determine the nature of the pybox
complexes and the mechanism of the origin of enantio-
selectivity in the reaction is currently in progress.
1
pounds were identified by matching their H NMR with
those of known compounds.1b,c,9
References and notes
9. Mayer, M. F.; Hossain, M. M. J. Org. Chem. 1998, 63,
6839–6844.
1. (a) Locaric, C.; Wulff, W. D. Org. Lett. 2001, 3, 3675; (b)
Antilla, J. C.; Wulff, W. D. Angew. Chem., Int. Ed. Engl.
2000, 39, 4518; (c) Antilla, J. C.; Wulff, W. D. J. Am.
Chem. Soc. 1999, 121, 5099; (d) Sengupta, S.; Mondal, S.
Tetrahedron Lett. 2000, 41, 6245; (e) Xie, W.; Fang, J.; Li,
J.; Wang, P. G. Tetrahedron 1999, 55, 12929; (f) Xie, W.;
Jin, Y.; Wang, P. G. Chem. Tech. 1999, 29(2), 23; (g)
Rasmussen, K. G.; Hazell, R. G.; Jørgensen, K. A. Acta
10. (a) Noyori, R.; Ohkuma, T.; Kitamura, M.; Takaya, H.;
Sayo, N.; Kumobayashi, H.; Akutagawa, S. J. Am. Chem.
Soc. 1987, 109, 5856; (b) Kitamura, M.; Ohluma, T.;
Inoue, S.; Sayo, N.; Kumobayashi, H.; Akutagawa, S.;
Ohta, T.; Takaya, H.; Noyori, R. J. Am. Chem. Soc. 1988,
110, 629; (c) Corey, E. J.; Bakshi, R. K.; Shibata, S.; Chen,
C.-P.; Singh, V. K. J. Am. Chem. Soc. 1987, 109, 7925.