Dual reaction pathways in the magnesium-mediated synthesis of aziridines from benzal halides and imines

Article abstract of DOI:10.1016/S0040-4039(01)00271-4

Reaction between benzal dihalides, benzaldehyde imines, and magnesium in ether affords aziridines in modest yields. Two mechanistic pathways for aziridine formation are discerned. One path involves nucleophilic attack by an alpha-halo Grignard species on

Full text of DOI:10.1016/S0040-4039(01)00271-4

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 2759–2762
Dual reaction pathways in the magnesium-mediated synthesis of
aziridines from benzal halides and imines
Mark R. Biscoe and Albert J. Fry*
Chemistry Department, Wesleyan University, Middletown, CT 06459, USA
Received 16 January 2001; revised 12 February 2001; accepted 13 February 2001
Abstract—Reaction between benzal dihalides, benzaldehyde imines, and magnesium in ether affords aziridines in modest yields.
Two mechanistic pathways for aziridine formation are discerned. One path involves nucleophilic attack by an alpha-halo Grignard
species on the imine; the other involves electrophilic attack on imine by phenylcarbene to afford an azomethine ylide. © 2001
Elsevier Science Ltd. All rights reserved.
A variety of methods for the preparation of aziridines
have been reported.1–4 Some, such as the thermolysis of
b-haloamines,5 involve transformation of a pre-existing
C–C–N chain. Others can be characterized as of [1+2]
type; they involve (a) addition of a nitrogen species to
an alkene or (b) a carbon species to an imino (CꢀN)
linkage. Type (a) processes include the copper-catalyzed
addition of chloramine to alkenes6 and the addition of
carboethoxynitrene to activated esters.7 Type (b) pro-
cesses include the addition of diazo compounds or
sulfur ylides to imines.8–11 Another [1+2] sequence
could be envisioned in which a benzal dihalide (1) first
reacts with magnesium to form an alpha-halo Grignard
reagent (2), the latter adds to an imine component of
the solution, and the resulting adduct (4) then cyclizes
to the aziridine (Scheme 1). We report successful imple-
mentation of this synthetic route and the surprising
discovery that there are in fact two mechanistic paths to
aziridine using this protocol.
benzaldehyde n-propyl imine (3a) and magnesium
under ultrasonic irradiation. The products isolated
(flash chromatography) were a mixture of aziridine 5,
reductive dimers of the starting imine (6), the imidazo-
lidine 7, and a mixture of cis- and trans-stilbenes (8)
(Scheme 2, Table 1). Initial experiments in THF
afforded low yields of aziridines, but ether was found to
afford better yields and was used for the experiments in
Table 1. A few experiments were carried out with
substituted imines 3b,c (runs 9–12). Authentic samples
of cis- and trans-5 were prepared by reaction between
the appropriate stilbene oxide and propylamine to
afford the appropriate b-aminoethanol (9), followed by
cyclization of 9 with triphenylphosphine/diethyl azodi-
carboxylate.12,13 Dimers 6 are produced by reaction of 3
with magnesium. They were formed in significant
amounts only in those reactions in which the reaction
of 1 with magnesium appeared to be slow. Only stilbe-
nes were produced if imine 3 was omitted from the
solution. Imidazolines 7 hydrolyzed to a mixture of
benzaldehyde and 6 during chromatography and were
characterized only by mass spectrometry.14
Benzal chlorobromide (1b) and benzal bromide (1a)
were allowed to react with varying molar ratios of
Scheme 1.
* Corresponding author. Tel.: 860-685-2622; fax: 860-685-2211; e-mail: afry@wesleyan.edu
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)00271-4

2760
M. R. Biscoe, A. J. Fry / Tetrahedron Letters 42 (2001) 2759–2762
Scheme 2.
Table 1. Magnesium-promoted reaction between benzal dihalides with imines
Run
Halide
Imine
Ratio of 3 to 1
5 (%)
cis:trans Ratio
7 (%)
1
2
3
4
5
6
7
8
9
1a
1a
1a
1b
1b
1b
1a
1a
1a
1a
3a
3a
3a
3a
3a
3a
3b
3b
3c
3c
1:1
3:1
6:1
1:1
3:1
6:1
1:1
3:1
1:1
3:1
24
39
60
10
19
21
23
35
21
32
72:28
67:33
61:39
29:71
19:81
17:83
75:25
65:35
58:42
49:51
9
12
17
3
4
4
12
18
8
10
11
The dibromide (1a) and chlorobromide (1b) were found
to differ in several significant respects in this reaction.
First and most striking, the cis-aziridine (cis-5) is the
major isomer produced with 1a, while trans-5 predomi-
nates in the reactions carried out with 1b. Secondly,
substantial quantities of imidazolidine 7 are produced
when halide 1a is used, whereas the amount of 7
produced is very low with 1b as reactant. Finally, the
yields of 5 and 7 are substantially dependent upon the
concentration of imine with dibromide 1a but much less
so with chlorobromide 1b. We conclude from these
results that there are two paths to aziridine 5 in this
system. Path A produces all or mostly trans-5, is less
sensitive to the concentration of imine, and does not
produce imidazolidine 7. Path B affords cis-5 as the
major or exclusive product, is quite sensitive to the
concentration of imine, and can lead to 7 as a side
product. We propose that path A involves the mecha-
nism of Scheme 1, i.e. nucleophilic attack of the ini-
tially-generated a-halomagnesium species 2 upon 3 to
form a b-chloro nitrogen species (4), which then cyclizes
to 5. The preferred stereochemistry of addition of 2 to
3 should be such that the two phenyl rings are as far
apart as possible in the transition state. This suggests
that the major diastereomer of 4 produced should be the
erythro diastereomer, which would then be expected to
cyclize to trans-5 (Scheme 3). A minor pathway would
involve formation of threo-4 and ultimately cis-5.
We suggest that path B involves loss of magnesium
halide from 2 to afford phenylcarbene (10) (Scheme 4).
Nucleophilic attack on 10 by imine 3 would produce
the azomethine ylide 11. Molecular mechanics
calculations15 on 11 predict that the s-cis/s-trans struc-
Scheme 3.

M. R. Biscoe, A. J. Fry / Tetrahedron Letters 42 (2001) 2759–2762
2761
Scheme 4.
(1)
Scheme 5.
ture shown is lower energy than the all s-trans or all
s-cis stereoisomers. Orbital symmetry-controlled
cyclization of this stereoisomer should proceed in con-
rotatory fashion to afford cis-3, and 7 would be formed
by [2+3] cycloaddition of 11a to starting imine 3.16,17
This path would be favored for 1a because of the better
leaving group ability of bromide, though it is unlikely
that either dihalide reacts exclusively by just one path.
1. For this reason we were surprised to find the substi-
tuted stilbenes 12 (from 3b) and 13 (from both 3b and
3c) among the reaction products in runs 7–10 (total
yields of 12 and 13 ca. 5–10%).
We believe that this constitutes evidence that azome-
thine ylide formation is reversible (Scheme 5). Because
carbenes are electrophilic, both methyl and fluorine
should stabilize the substituted carbene (10b,c) resulting
from decomposition of mixed ylide 11b,c.19 Trapping of
10b,c by MgBrX or 2a would then lead on to 12 and
13. As Scheme 5 would predict, runs 7–10 also pro-
duced the unsubstituted imine 3a as well as aziridines
and imidazolidines containing varying degrees of
substitution.
A final unexpected result must be mentioned here. As
mentioned above, cis- and trans-stilbenes 8 are pro-
duced as side products in these reactions. These are
presumably formed by reaction of carbene 10 with 2,
followed by loss of magnesium halide (Eq. (1)).18 Both
aryl groups therefore originate from the geminal halide

2762
M. R. Biscoe, A. J. Fry / Tetrahedron Letters 42 (2001) 2759–2762
1H NMR (300 MHz): l 0.95 (t, 3H), 1.6 (m, 2H), 2.05
Acknowledgements
(m, 1H), 2.5 (m, 1H), 3.1 (br s, 1H), 3.35 (br s, 1H),
7.3–7.6 (m, 10H); MS (EI): 237 (26), 236 (61), 194 (100),
165 (29); exact mass (CI) calcd for M+1: 238.1596, found:
238.1594; (b) cis-5a: 1H NMR: l 1.05 (t, 3H), 1.75 (m,
2H), 2.65 (t, 2H), 2.85 (s, 2H), 7.1–7.4 (m, 10H); MS (EI):
237 (20), 236 (53), 194 (100), 165 (27); exact mass (CI)
calcd for M+1: 238.1596, found: 238.1590; (c) erythro-9:
1H NMR: l 0.9 (t, 3H), 1.35 (s, 1H), 1.5 (m, 2H), 2.5 (m,
2H), 3.25 (br s, 1H), 3.9 (d, 1H), 4.85 (d, 1H), 7.15 (m,
4H), 7.25–7.35 (m, 6H); 13C NMR: 141, 140, 128.4, 128.3,
128.1, 127.8, 127.7, 127, 77, 69, 49, 23, 12; MS (EI): 254
(0.3), 237 (0.2), 236 (0.3), 208 (0.4), 194 (0.4), 148 (100),
106 (23); (d) threo-9: 1H NMR: l 0.9 (t, 3H), 1.35 (s, 1H),
1.5 (m, 2H), 2.5 (m, 2H), 3.3 (br s, 1H), 3.85 (d, 2H), 4.85
(d, 2H), 7.3–7.6 (m, 10H); MS (EI): 254 (0.3), 237 (0.2),
236 (0.3), 208 (0.4), 194 (0.4), 148 (100), 106 (23); (e) 3b:
1H NMR: l 0.95 (t, 3H), 1.72 (m, 2H), 3.56 (t, 2H), 7.1
(dt, 2H), 7.75 (m, 2H), 8.24 (s, 1H); 13C NMR: 164.3 (d,
J=249 Hz), 159.4, 132.79 (doublet, J=3 Hz), 129.95 (d,
J=8 Hz), 115.65 (d, J=23 Hz), 63.5, 24.2, 11.9; MS (EI):
165 (5), 164 (17), 136 (57), 109 (100); (f) trans-5b: 1H
NMR: l 0.85 (t, 3H), 1.6 (m, 2H), 2.45 (m, 2H), 3.0 (br
s, 1H), 3.25 (br s, 1H), 7–7.8 (m, 9H) MS (EI): 254 (20),
212 (100), 165 (15); (g) cis-5b: MS (EI): 254 (21), 212
Financial support by the National Science Foundation
(Grant No. CHE-97-13306 to A.J.F.) is gratefully
acknowledged. Mark Biscoe was the recipient of two
Hughes Foundation summer research grants.
References
1. Deyrup, J. A. In The Chemistry of Heterocyclic Com-
pounds; Hassner, A., Ed.; Wiley: New York, 1983; Vol.
42, pp. 11–83.
2. Kametani, T.; Honda, T. Adv. Heterocycl. Chem. 1986,
39, 181.
3. Padwa, A.; Woolhouse, A. D. In Comprehensive Hetero-
cyclic Chemistry; Lwowski, W., Ed.; Pergamon: Oxford,
1984; Vol. 7, pp. 80–93.
4. Tanner, D. Angew. Chem., Int. Ed. Engl. 1994, 33, 599.
5. Bartnik, R.; Mloston, G.; Lesniak, S. Pol. J. Chem. 1979,
53, 537.
6. Ando, T.; Minakata, S.; Ryu, I.; Komatsu, M. Tetra-
hedron Lett. 1998, 39, 309.
7. Carducci, M.; Fioravanti, S.; Loreto, M. A.; Pellacani,
L.; Tardella, P. A. Tetrahedron Lett. 1996, 37, 3777.
8. Gunnoe, T. B.; White, P. S.; Templeton, J. L.; Cassarru-
bios, L. J. Am. Chem. Soc. 1997, 119, 3171.
9. Gyenes, F.; Bergmann, K. E.; Welch, J. T. J. Org. Chem.
1998, 63, 2824.
1
(100), 165 (15): H NMR: l 0.95 (t, 3H), 1.75 (m, 2H),
3.55 (t, 2H), 7.10 (m, 2H), 7.7 (m, 2H), 8.24 (s, 1H); 13C
NMR: 166, 162.5, 159.5, 132.7 (doublet), 130, 116, 62, 24,
1
12; (h) 6b: H NMR: l 0.8 (t, 3H), 1.35 (m, 2H), 1.55 (br
s, 2H; disappears in the presence of D2O), 2.2 (t, 2H), 3.7
(s, 2H), 6.85 (m, 4H), 7.3 (m, 6H); MS (EI): 274 (0.5), 166
(M/2) (100), 148 (13), 124 (13); (i) 7b: MS (EI): 325
(M−C6H4F) (25), 273 (85), 255 (100), 230 (85%); (j) 7a:
MS (EI): 383 (M−1) (31), 355 (5), 341 (4), 307 (79), 237
(100), 236 (67), 194 (58).
10. Mohan, J. M.; Uphade, B. S.; Choudhary, V. R.; Ravin-
dranathan, T.; Sudalai, A. Chem. Commun. 1997, 1429.
11. Wang, D.-K.; Dai, L.-X.; Hou, X.-L. Chem. Commun.
1997, 1231.
12. Carlock, J. T.; Mack, M. P. Tetrahedron Lett. 1978, 5153.
13. Stereochemical inversion at nitrogen is fairly slow at
room temperature for 5; comparison of experimental
NMR spectra with computed spectra (Quantum Chem-
istry Program Exchange program DNMR5) indicated a
rate of ꢀ35 s−1 at 25°C.
15. PCMODEL7, Serena Software: Bloomington, IN.
16. Hall, J.; Huisgen, R. J. Chem. Soc. D 1971, 1187.
17. 1,3-Dipolar Cycloaddition Chemistry; Padwa, A., Ed.;
Wiley: New York, 1984.
18. Kirmse, W. Carbene Chemistry; Academic: New York,
1964.
14. Spectroscopic properties for new substances (1H COSY
NMR spectra were carried out on all new substances and
were consistent with the assigned structures): (a) trans-5a:
19. Hine, J. Divalent Carbon; Ronald Press: New York, 1964.
.