Stereoselective isomerisation of N-Allyl aziridines into geometrically stable Z enamines by using rhodium hydride catalysis

Article abstract of DOI:10.1002/chem.200701322

In the presence of rhodium(I) hydride catalysts, tertiary N-allylamines are known to isomerise into E enamines. In contrast, we have recently found that N-allylaziridines isomerise to form Z enamines. On the basis of literature data, the most likely mecha

Full text of DOI:10.1002/chem.200701322

DOI: 10.1002/chem.200701322
Stereoselective Isomerisation of N-Allyl Aziridines into Geometrically Stable
Z Enamines by Using Rhodium Hydride Catalysis
Derek S. Tsang, Sharon Yang, France-AimØe Alphonse, and Andrei K. Yudin*[a]
Abstract: In the presence of rhodi-
um(I) hydride catalysts, tertiary N-al-
lylamines are known to isomerise into
E enamines. In contrast, we have re-
cently found that N-allylaziridines iso-
merise to form Z enamines. On the
basis of literature data, the most likely
mechanism of isomerisation would in-
volve a rhodium hydride addition/b-hy-
dride elimination sequence. We show
that the observed selectivity cannot be
adequately explained by this pathway
and is more consistent with initial CH-
activation followed by rearrangement
to form a five-membered cyclometal-
lated rhodium intermediate. This inter-
mediate subsequently undergoes reduc-
ꢀ
tive elimination to form a C H bond.
The resulting geometrically stable Z
enamines are useful building blocks for
stereoselective synthesis.
Keywords: aziridines · enamines ·
isomerisation
rhodium
·
metallacycles
·
Introduction
Enols, metal enolates and enamines are amongst the most
important carbon nucleophiles in both biological and chemi-
cal synthesis.[1] Many useful transformations owe their effi-
ciency to selective formation of E or Z metal enolates. In
contrast, the chemistry of enamines is restricted to reactions
of their E isomers because the classical condensation be-
tween amine and carbonyl functional groups is an equilibri-
um process that affords thermodynamically favoured E
products. There are no useful, kinetically controlled reac-
tions that result in Z enamine formation. Although strong
base-promoted isomerisation of N-allylamines into Z enam-
ines was reported,[2] attempts to isolate the Z products result
in rapid Z-to-E isomerisation or hydrolysis (Scheme 1).
Even in the metal-catalysed cross-coupling of geometrically
pure (Z)-vinyl bromides with amine nucleophiles, formation
of E enamines is observed due to in-situ isomerisation of
the kinetic Z product into the E isomer.[3]
Scheme 1. Z enamines are difficult to prepare, often isomerising or hy-
drolysing during isolation.
Noyoriꢀs BINAP/rhodium(I) catalyst has been used in one
of the most acclaimed asymmetric processes practiced in in-
dustry, synthesis of menthol.[5] In a generally accepted mech-
anistic description of the reaction depicted in Scheme 2,
CH-activation at the allylic position leads to an organome-
tallic species A, which is stabilised through iminium com-
plex B. Subsequent hydride migration to the terminal olefin-
ic carbon atom affords intermediate C which undergoes de-
complexation, releasing the E enamine product. Interesting-
Cationic rhodium(I) diphosphine complexes are known to
induce selective isomerisation of N-allylamines into enam-
ines.[4] Once again, the E geometry is strongly favoured.
[a] D. S. Tsang, S. Yang, Dr. F.-A. Alphonse, Prof. A. K. Yudin
Davenport Research Laboratories
Department of Chemistry
University of Toronto
80 St. George St., Toronto
Ontario, M5S 3H6 (Canada)
Fax : (+1)416-946-7676
Scheme 2. Accepted mechanism of isomerisation of allyl amines with cat-
E-mail: ayudin@chem.utoronto.ca
ionic rhodium(I) diphosphine complexes.
886
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 886 – 894

FULL PAPER
ly, failure of the cationic rhodium(I) catalysts to induce iso-
merisation of allyl amines derived from ethylene imine has
been used as an indirect piece of evidence in favour of imi-
nium complex formation.[6]
selectivity in transition-metal-catalysed isomerisations, al-
though the levels of selectivity are generally quite low.
These mechanisms often involve some form of coordination
between the transition-metal catalyst and the substrate,
forming a metallocycle, such as in the case of N-allyla-
mides.[18] What is unique about N-allylaziridines are the high
Z selectivities observed.
In this study, we put forth mechanistic evidence in support
of a cyclometallated rhodium intermediate, which operates
concurrently with a non-productive, hydride addition–elimi-
nation mechanism that has traditionally been suggested as
the sole mechanism of catalytic olefin isomerisation with
late transition-metal hydrides.
Rhodium(I) hydride catalysts are also known to induce
isomerisation of N-allylamines.[7,8] The generally accepted
mechanism of isomerisation in this case involves metal hy-
dride addition across the carbon–carbon double bond fol-
lowed by b-hydride elimination, resulting in double-bond
migration. This transformation generally leads to E enamine
geometry in both the kinetic and thermodynamic product.
Previous work by our group revealed that NH-aziridines
behave differently from other secondary amines with respect
to palladium-catalysed allylic amination.[9,10a] With common
secondary amine nucleophiles, thermodynamically favoured
linear products are formed due to in-situ palladium-cata-
lysed isomerisation. On the other hand, with NH-aziridine
nucleophiles, the kinetically favoured branched products do
not undergo further isomerisation into the linear allyl
amines. This mechanistic insight has allowed us to find con-
ditions under which any primary or secondary amine can un-
dergo palladium-catalysed allylic amination with high
branched selectivity.[10b] With a reliable route to allyl aziri-
dines in hand, we opted for examine transition-metal-pro-
moted isomerisation of N-allylaziridines into enamines. As a
result, we uncovered yet another instructive aberration in
aziridine chemistry: N-allylaziridines behave differently
from N-allylamines with respect to metal-catalysed isomeri-
sation. Whereas N-allylamines form E isomers upon expo-
sure to rhodium(I) hydride catalysts, it was found that iso-
merisation of N-allylaziridines results in unusual selectivity
for the thermodynamically disfavoured (Z)-N-propenylaziri-
dine (Scheme 3).[11] As aziridine rings are stepping stones to
more complex amines by means of diverse ring-opening pro-
cesses, this chemistry is expected to find useful synthetic ap-
plications.
Results
With our recently disclosed approach to allyl aziridines,[9,10a]
we investigated the isomerisation of N-allylcyclohexene
imine 1 with 1.5 mol% of [Rh
A
ACHTREUNG
of rhodium(I) prepared in situ from [Rh
AHCTREUNG
BINAP.[11] In contrast to previous reports with allylated eth-
ylene imine, we did obtain the N-(1-propenyl)aziridine 2,
albeit with a low 25% conversion (Scheme 4). The rear-
Scheme 4. Initial isomerisation studies. a) [Pd
A
PPh3, K2CO3, THF, RT, 16 h, 76%; b) [Rh(cod)2]OTf, rac-BINAP, THF,
AHCTREUNG
reflux, 608C, 16 h, 25% conv. cod=1,5-cyclooctadiene; BINAP=
(2R,3S)-2,2’-bis(diphenylphosphino)-1,1’-binapthyl.
rangement of 1 (t1/2 =164.0 min) was significantly slower
than the rearrangement of N-allylmorpholine under similar
conditions (t1/2 =1.3 min). Interestingly, unexpected selectivi-
ty for the Z isomer of 2 was observed. There was no evi-
dence for isomerisation into the E isomer in the course of
isolation.
Similar to cationic rhodium(I) catalysts equipped with bi-
dentate ligands, such as BINAP, rhodium(I) hydride cata-
lysts are known to induce allyl amine isomerisation. Reports
in the literature have previously indicated that amines such
as N-allylpiperidine (3) can be isomerised with [Rh(CO)H-
Scheme 3. Isomerisation of N-allylaziridines and N-allylamines leads to
(PPh3)3], leading exclusively to the E isomer 4 (Scheme 5).[8]
In our hands, treatment of N-allylmorpholine with
[Rh(CO)H(PPh3)3] also showed exclusive preference for the
different geometries. cat.=<5 mol%; RhH=[Rh(CO)H(PPh3)3].
A
AHCTREUNG
It should be noted that unexpected Z geometry has been
observed with other substrates, including simple elimination
reactions with base,[12] isomerisation of olefins varying in
chain length from 1-butene to 1-heptene,[13,14,15,16] allyl vinyl
ethers,[17] and N-allylamides.[18] The classical hydride mecha-
nism is often mentioned in relation to both isomerisa-
E enamine, whereas treatment of N-allylcyclohexene imine
under similar conditions gave the Z product. Notably, the
tion[8,14,15,18] and hydrogenation[7,19] with [Rh(CO)H
(PPh3)3],
A
Scheme 5. Isomerisation of 1-allylpiperidine.[8] a) NaH, THF, RT, 16 h,
54% conv.; b) RhH (1.5 mol%), THF, RT, 16 h, 31%.
but this mechanism is expected to produce E products. Al-
ternate mechanisms have been suggested to explain the Z
Chem. Eur. J. 2008, 14, 886 – 894
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
887

A. K. Yudin et al.
Table 1. Selected substrate scope for isomerisation of N-allylaziridines to
rearrangement of N-allylmorpholine (t1/2 =330.0 min) was
significantly slower than the rearrangement of N-allylaziri-
dine under similar conditions (t1/2 =29.0 min).
N-vinylaziridines with [Rh(CO)H
(PPh3)3] catalyst.[a]
A
Entry
Substrate
Conv. [%][b]
100
Z/E selectivity [%][b]
The Z stereoselectivity in the course of allyl aziridine iso-
1
95:5
merisation, observed with [Rh(CO)H(PPh3)3] as the catalyst,
G
was increased to 95:5 when the reaction was performed at
ꢀ788C. The same selectivity was achieved in both toluene
and dichloromethane. The double-bond migration consis-
tently afforded high Z selectivity with full conversion at
room temperature (Scheme 6a). GC analysis of N-allylcyclo-
hexene imine (1) isomerisation, performed with 1.5 mol%
2[c]
3[d]
4[e]
100
100
100
82:18
76:24
91:9
5
100
80
96:4
6[c]
75:25
7[f]
trace
trace
–
–
8[f]
[a] [Substrate]=0.7m, [substrate]/[Rh]=66 at ꢀ788C in THF under N2.
[b] Determined by GC analysis. [c] [Substrate]/[Rh]=10 at 608C, 72 h.
[d] [Substrate]/[Rh]=20 at 608C, 24 h. [e] [Substrate]/[Rh]=66 at 208C,
24 h. [f] Only starting material was recovered.
of hydrogen.[7] The mechanism of hydrogenation passes
through the same rhodium–carbon intermediates as those
formed in the course of double-bond migration. To detect
and trap these rhodium–carbon intermediates, isomerisa-
tions were carried out in the presence of hydrogen gas. Hy-
drogenation of 1 resulted in a mixture of hydrogenated and
isomerised products in a Z/E ratio consistent with the ratios
observed in the absence of H2 gas (Scheme 6d). Further-
more, hydrogenation of 5 was extremely slow and resulted
in low yields (Scheme 6f), demonstrating the inhibitory
effect of having a nitrogen substituent proximal to the site
of hydrogenation.
Substitution at the allylic position was found to impede
isomerisation.[7] Furthermore, the addition of substituents to
the N-allyl chain creates a tri- (7) or di-substituted (8) olefin
that does not isomerise because of the thermodynamic sta-
bility of the starting material (Scheme 7).
Scheme 6. Isomerisation and hydrogenation experiments. a) RhH
(1.5 mol%), C6H6 or THF, RT, 1 d; b) RhH (1 equiv), THF, RT, 5 d;
c) hn (lmax =254 nm), quartz, THF, RT, 6 h; d) H2 (1 atm), RhH (cat.),
THF, RT, 16 h, 91%; e) Cu
A
(1 atm), RhH (1 mol%), C6H6, RT, 1 week, 32%.
of [Rh(CO)H(PPh3)3] in THF at room temperature, indicat-
U
ed that the E and Z aziridine isomers do not interconvert
under the reaction conditions or during isolation. This iso-
merisation protocol was extended to a number of different
allyl aziridine substrates (Table 1). All of the terminal ole-
fins were reactive, consistent with the known selectivity of
[Rh(CO)H(PPh3)3] towards unhindered terminal double
A
bonds.[7] Moreover, all of the substrates showed high selec-
tivity for the Z enamine product. A conversion to a thermo-
dynamic mixture of geometric isomers was effected upon
extended treatment of 1 with a stoichiometric amount of
rhodium(I) hydride catalyst, which caused slow isomerisa-
tion to a 54:46 Z/E isomeric mixture over five days
(Scheme 6b). A similar thermodynamic ratio was photo-
chemically induced upon UV irradiation (lmax =254 nm;
Scheme 6c).
Deuterium-labelling studies were undertaken to elucidate
the mechanism of isomerisation. Scheme 8a shows a stoi-
chiometric isomerisation of 1 with [Rh(CO)D(PPh3)3] which
A
resulted in deuterium incorporation at the methyl (C3) and
central vinylic (C2) positions. Extended treatment with stoi-
chiometric [Rh(CO)D(PPh3)3] over five days resulted in
G
scrambling of deuterium across all three positions, as well as
formation of a thermodynamic mixture of Z/E isomers
(Scheme 8b). Experiments to trap the rhodium–carbon in-
termediate A (Scheme 2) by using the 3,3-dimethyl sub-
strate, which did not undergo isomerisation, showed no deu-
In the presence of H2, [Rh(CO)H(PPh3)3] acts as a mild
G
and efficient hydrogenation catalyst, able to reduce terminal
carbon–carbon double bonds under an atmospheric pressure
888
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 886 – 894

Isomerisation of N-Allyl Aziridines
FULL PAPER
Scheme 7. Substituted N-allylaziridines do not isomerise under RhH cat-
alysis.
Scheme 9. Further isomerisation studies. a) RhH (15 mol%), C6D6, in-
situ experiment (utilising NMR spectroscopy); b) RhH (cat.), THF, RT,
16 h, 94% conv.
Discussion
This study was designed to elucidate differences between
the isomerisation of N-allylaziridines and allyl amines, as
the same rhodium(I) hydride catalyst leads to two different
enamine geometries from these substrates. We previously
postulated a mechanism of methylene CH-activation, fol-
lowed by the formation of a cyclic intermediate and reduc-
tive CH-bond elimination of the Z enamine (Scheme 10).[11]
However, isomerisation of 1 into the Z enamine product
by using stoichiometric rhodium(I) deuteride resulted in
deuterium incorporation at the C2- and C3-positions of iso-
lated product, suggesting that a hydride-based mechanism
may be involved. This pathway would proceed by means of
hydride addition across the double bond and subsequent b-
Scheme 8. Deuterium labelling and crossover experiments. a) RhD
(1 equiv), THF, RT, 30 min; b) RhD (1 equiv), THF, RT, 5 d; c) RhH
(3 mol%), THF, RT, 12 h. RhH=[Rh(CO)H
(PPh3)3].
A
ACHTREUNG
ꢀ
terium incorporation at position 1 (Scheme 7). Finally, a
crossover experiment allowed us to observe deuterium
transfer from one aziridine substrate to another, indicating
that deuterium from C1 of the 1,1-dideuterioallylaziridine is
transferred to the rhodium(I) complex, which in turn trans-
fers deuterium onto the protio-aziridine substrate
(Scheme 8c).
hydride elimination to re-form the C2 C3 double bond in a
cycle that we have termed the “hydride loop” (Scheme 11).
The intermediate rhodium–carbon species D was trapped
when we isolated n-propyl cyclohexene imine under hydro-
genation conditions.
If b-hydride elimination from intermediate D were also
responsible for forming the C1 C2 double bond of the en-
ꢀ
To further probe the mechanism, the isomerisations were
investigated on simple allylamine and allylcyclopropanes.
The use of 1-allylamine as a substrate revealed transient for-
mation of the unstable 1-aminopropene in a Z/E ratio of
80:20 with a half-life (t1/2) of 142.2 min (Scheme 9a). Isomer-
isation of 10, the cyclopropane equivalent of 1, resulted in
poor selectivity, with a Z/E ratio of 64:36 (Scheme 9b). This
suggests that the presence of a three-membered aziridine
ring played an important role
amine product, it should result in E enamine geometry be-
cause the activation barrier to form the Z isomer is known
to be significantly higher.[20] However, our experiments dem-
onstrate the kinetic nature of the isolated Z product. The
Z/E ratio remained constant throughout the reaction. This
suggests that the enamine product is relatively unreactive to
further isomerisation via hydride addition and subsequent b-
hydride elimination.[7] However, isomerisation can be forced
in determining high Z selectiv-
ity.
It is evident that allylaziri-
dines differ from other allyla-
mine substrates, including ter-
tiary amines, in that they show
very strong selectivity for the
Z enamine product. Why is it
that the aziridine nitrogen in-
duces such unique stereoselec-
tivity? We propose a mecha-
nism that explains these re-
sults.
Scheme 10. Proposed mechanism of rhodium hydride catalysed CH-activation/isomerisation (L=PPh3).
Chem. Eur. J. 2008, 14, 886 – 894
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
889

A. K. Yudin et al.
bond, releasing the Z enamine
product. Crossover of deuteri-
um has been observed between
the deuterated and protio-azir-
idine (Scheme 8), suggesting
that
[Rh(CO)D
A
is
formed in situ following CH-
activation of the deuterated
substrate and reductive elimi-
nation from the ensuing cyclo-
metallated rhodium intermedi-
ate. RhD then enters the CH-
activation path with the protio-
Scheme 11. Hydride addition and b-hydride elimination (the “hydride loop”). The rhodium-substrate inter-
mediates may be trapped by hydrogenation.
by prolonged exposure to stoichiometric [Rh(CO)H-
(PPh3)3], resulting in a thermodynamic mixture of Z and E
aziridine substrate to effect crossover of deuterium
(Scheme 12).
ACHTREUNG
enamines. Importantly, deuterium scrambling and partial
isotope incorporation at the C1-position was also observed
upon extended treatment with the RhD catalyst. Such ex-
tended treatment causes scrambling of deuterium through-
out the three-carbon substituent by means of slow hydride
addition followed by b-hydride elimination. The E product
was formed after such prolonged exposure: accordingly, E
enamine formation must proceed by means of b-hydride
elimination. By the principle of microscopic reversibility, the
reverse reaction—addition to the enamine double bond—is
slow, which is reflected in the slow rate of Z-to-E conversion
as well as the fact that neither E nor Z aziridine enamine,
nor the unsubstituted vinyl aziridine 5, undergo hydrogena-
Scheme 12. Mechanism of deuterium crossover. 1,1-Dideuterioallylaziri-
dine transfers deuterium to rhodium(I), forming [Rh(CO)D(PPh3)3] in
A
situ.
tion at a reasonable rate when using the [Rh(CO)H
catalyst (Scheme 6).
G
Why is it that b-hydride elimination cannot be responsible
for Z enamine formation? We attribute this intriguing find-
ing to the special properties of the aziridine nitrogen, includ-
ing its high degree of pyramidalisation and its high barrier
to inversion. In contrast to other amines, two substituents
on aziridine rings are locked in a three-membered heterocy-
cle and thus are swept back from the nitrogen (Figure 1).
The conventional hydride mechanism alone cannot ex-
plain Z enamine formation in the case of allyl aziridines. We
conclude that a lower-barrier Z-selective pathway must op-
erate. Concurrently with the non-productive addition across
the C2–C3 double bond (“hydride loop”, Scheme 11), the
RhH catalyst can enter into a methylene CH-bond activa-
tion manifold (Scheme 10). The resulting rhodium inter-
mediate cannot be stabilised by the iminium ion B due to
high strain. Thus, the formation of any E product by means
of the tertiary allyl amine isomerisation mechanism is not
expected along this pathway. On the other hand, formation
of an isomeric five-membered cyclometallated rhodium
complex C can be expected to take place, via a s-allyl inter-
mediate. It is important to mention that at this point we do
not have experimental evidence ruling out direct hydride
migration from the p-allyl intermediate which should also
result in Z enamine formation. The rhodium complex C has
not been directly observed by NMR spectroscopy, but sever-
al arguments speak in its favour. The wA/p interaction is
known to stabilize the bisected geometry in stereoelectroni-
cally related vinylcyclopropanes.[21] This interaction, coupled
with Rh–N coordination, is expected to contribute towards
the formation of C. Furthermore, five-membered ring organ-
ometallic intermediates have been postulated to account for
Z selectivity in the case of alkali base-induced rearrange-
ments.[22] Following the formation of C, facile reductive
Figure 1. Representations of (Z)-N-vinylaziridine 2 and (Z)-N-vinylpiper-
idine 4.
This relieves steric strain in the cyclometallated intermedi-
ate and the resulting Z enamines that are formed. Further-
more, coordination of rhodium(I) to the sterically unencum-
bered lone pair of nitrogen in intermediate E (Scheme 13)
would prevent formation of a syn-periplanar rotamer, which
is a prerequisite for b-hydride elimination. Interestingly, a
similar coordinative arrangement between palladium(II)
and hydroxyl group was reported to affect the regioselectivi-
ty of b-hydride elimination in the Heck reaction.[23] Tertiary
amines, which lead to the formation of E enamines, do not
appear to exhibit this behaviour.[8] This explains why N-allyl-
morpholine and N-allylpiperidine isomerisations occur with
E selectivity—although they have a nitrogen lone pair that
elimination is expected to take place to form a new C H
890
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 886 – 894

Isomerisation of N-Allyl Aziridines
FULL PAPER
drate was used as purchased from Strem Chemicals without further pu-
rification. NMR spectra were referenced to TMS and run at 258C. CDCl3
was stored over anhydrous K2CO3. Sodium deuteride (98 atom% D) and
EtOD (99 atom% D) were purchased from Aldrich and were used as re-
ceived. Isotopic purity of metal catalysts was determined by 1H NMR
spectroscopy. Deuterium-labelled products were detected by 2H NMR
spectroscopy. GC runs were carried out on a Hewlett Packard 6890 gas
chromatograph with an HP-5 column (crosslinked 5% phenyl methyl si-
loxane, 30 m”0.32 mm”0.25 mm film thickness), and were configured to
start at 508C, run 1.0 min at 508C, ramp 108Cminꢀ1 to 1508C; hold
2.0 min, then ramp 158Cminꢀ1 to 2508C, and hold 6.0 min.
Cyclohexene imine:[9,10a] In a 100 mL round-bottomed flask equipped
with stirrer bar and condenser, NaN3 (5.08 g, 77.5 mmol) was dissolved in
1:1 H2O/acetone (35 mL). Cyclohexene oxide was then added to this so-
lution (3.1 mL, 30.7 mmol). The reaction was refluxed for 16 h and
cooled, at which point a biphasic system was observed. Without separat-
ing the layers, the acetone was evaporated under reduced pressure and
then the layers were extracted with diethyl ether (3”20 mL). Organic ex-
tracts were combined, washed with brine, dried with Na2SO4, then evapo-
rated in vacuo. Crude trans-2-azidocyclohexanol (4.98 g) was recovered
as an orange oil, all of which was dissolved in dry THF (26 mL). PPh3
(9.3 g, 35.5 mmol) was added with stirring to this solution. Bubbling of N2
gas was observed as the solution was refluxed for 16 h. The solvent was
removed gently under reduced pressure. Pentanes were poured in and tri-
phenylphosphine oxide precipitate was repeatedly filtered. Pentanes in
the filtrate were removed under reduced pressure and the oil was dis-
tilled in vacuo to yield 1 g of cyclohexene imine (10.3 mol, 34%, GC
time: 4.3 min) as a clear, colourless liquid that solidified upon standing at
ꢀ208C. 1H NMR (300 MHz, CDCl3): d=2.10 (2H, m), 1.80 (4H, m),
1.30 ppm (4H, m); 13C NMR (75 MHz, CDCl3): d=20.3, 24.8, 29.5 ppm.
Scheme 13. Aziridine nitrogen may coordinate to rhodium(I), forming a
transient four-membered ring that precludes syn-periplanar arrangement
and b-hydride elimination.
appears capable of forming a cyclometallated intermediate,
such a ring would be subject to severe steric constraint. The
only pathway open to tertiary amine isomerisation is b-hy-
dride elimination. Further work was done with the unsubsti-
tuted parent allyl amine, which was subjected to the isomeri-
sation protocol. This reaction produced a kinetic preference
for the Z isomer (80:20), suggesting that unencumbered
Rh–N coordination is a critical factor that contributes to re-
tarding the rate of b-hydride elimination.
Conclusion
We have shown that the special properties of the aziridine
nitrogen allow allyl aziridines to undergo isomerisation into
Z enamines through a CH-activation manifold that is not ac-
cessible to common tertiary amines. Our mechanistic investi-
gations are consistent with two concurrent reaction path-
ways operating during the isomerisation. The hydride addi-
tion and b-hydride elimination (“hydride loop”) are non-
productive with respect to enamine formation, while the
principal reaction pathway proceeds by means of CH-activa-
tion, rearrangement to a five-membered intermediate, and
subsequent reductive elimination producing (Z)-vinylaziri-
dine. The involvement of highly strained p intermediates
has not been ruled out. This catalytic method provides facile
access to a heretofore underutilised class of substrates—geo-
metrically stable Z enamines. These molecules can be used
as carbon nucleophiles in stereoselective reactions and
should facilitate explorations of Z enamine transforma-
tions.[11] Current efforts are aimed at further understanding
the relative reactivity of s and p intermediates involved in
this reaction.
7-Allyl-7-azabicyclo
flask equipped with stirrer bar, [Pd
ACHTREUNG
AHCTREUNG
1.5 mol%), PPh3 (70 mg, 0.27 mmol, 4 mol%), and K2CO3 (1.7 g,
17.0 mmol) were suspended in THF (11 mL). Allyl acetate (0.6 mL,
6.1 mmol) and cyclohexene imine (0.5 g, 5.2 mmol) were then added to
this solution. The contents were stirred for 16 h under N2, then poured
into pentanes (25 mL) and filtered through a pad of Celite. The filtrate
was concentrated in vacuo, then distilled under reduced pressure to yield
537 mg of 1 in 75% yield. H NMR (400 MHz, CDCl3): d=5.90 (1H, m),
5.20 (1H, d, J=17 Hz), 5.10 (1H, d, J=8.6 Hz), 2.80 (2H, d, J=5.3 Hz),
1.75 (4H, m), 1.50 (2H, m), 1.35 (2H, m), 1.18 ppm (2H, m).
1
Hydrogenation of 1 to a mixture of saturatedandisomerisedproducts
50 mL flame-dried Schlenk flask equipped with stirrer bar was charged
with [Rh(CO)H(PPh3)3] (5 mg, 0.005 mol, 5 mol%) under inert atmos-
: A
AHCTREUNG
phere. The flask was evacuated, then filled with H2 at 1 atm. Dry THF
(1.5 mL) was added by syringe, and the yellow solution was stirred for
5 min. The flask was equipped with a H2-filled balloon, then 1 (15 mg,
0.11 mmol) was added by syringe. After stirring the reaction overnight,
GC analysis revealed a 91% conversion, with 54% consisting of hydro-
genated product, 34% (Z)-N-vinylaziridine, 3% (E)-N-vinylaziridine and
9% 7-allyl-7-azabicyclo[4.1.0]heptane (1, starting material). The reaction
solution was reduced to 0.2 mL in vacuo, then chromatographed on silica
gel (9:1 hexanes/EtOAc, I2 stain) to yield 6.5 mg (0.05 mmol, 43% yield)
G
of crude hydrogenated product 7-propyl-7-azabicyclo[4.1.0]heptane (Rf =
N
0.55). The starting material, N-vinylaziridines and product all eluted to-
gether upon column chromatography and were not separable. 1H NMR
(400 MHz, CDCl3): d=2.16 (2H, t, J=7.3 Hz), 1.75 (m), 1.56 (2H, q, J=
7.4 Hz), 0.9 ppm (m); ESI-MS: m/z (%): 140.1
(100) [M+1+] , 141.1 (10)
A
Experimental Section
[M+2+].
N-Allylpiperidine (3): In a 100 mL round-bottomed flask equipped with
stirrer bar, 60% NaH in oil (1.83 g, 45.8 mmol) was suspended in THF
(25 mL). Piperidine (2.5 mL, 2.1 g, 24.7 mmol) was added followed by
allyl bromide (2.3 mL, 3.3 g, 27.3 mmol). After stirring overnight, water
was poured in and the solution was extracted with EtOAc, dried with an-
hydrous MgSO4, and reduced in vacuo. The resulting oil was distilled
under reduced pressure to yield 1.9 g (54% isolated yield) of 3 (GC
time: 4.4 min). 1H NMR (200 MHz, CDCl3): d=5.90 (1H, m), 5.10 (2H,
m), 2.90 (2H, d), 2.40 (4H, brs), 1.60 ppm (m, 4H), 1.45 ppm (m, 2H).
General procedures: THF was freshly distilled from sodium benzophe-
none ketyl prior to use. Anhydrous acetone was stored over 4 Š molecu-
lar sieves. All procedures involving [Rh(CO)H(PPh3)3] were carried
A
under Ar/N2 in flame-dried glassware. n-Butyllithium was titrated with
N-benzylbenzamide prior to use.[24] N-Benzylbenzamide was synthesised
and purified according to literature procedures.[24] Triphenylphosphine
(PPh3) was recrystallised from hot ethanol prior to use to remove triphe-
nylphosphine oxide (Ph3PO). Other solvents and reagents were pur-
chased from Aldrich and used as received. Rhodium(III) chloride trihy-
U
Chem. Eur. J. 2008, 14, 886 – 894
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
891

A. K. Yudin et al.
2,4,6-Trivinylcyclotriboroxane–pyridine complex:[25] THF (75 mL) with
trimethylborate (10 mL, 9.32 g, 90 mmol) was cooled under N2 to ꢀ788C.
Vinylmagnesium bromide (60 mL, 1.0m in THF, 60 mmol) was added
dropwise by means of an addition funnel over 35 min. The reaction was
stirred for 1 h, at which point HCl (1m, 25 mL) was added. The reaction
was warmed to room temperature, then brine (20 mL) was added. The
slurry was extracted with ether (4”50 mL). The organic extracts were
washed with water (50 mL), brine (50 mL), dried with Na2SO4, reduced
to 25 mL in vacuo, and then pyridine (10 mL, 9.83 g, 0.12 mol) was added
under N2. This solution was stirred for 4 h. The solvent was removed
under high vacuum, then the remaining oil was distilled at 2008C under
reduced pressure. The complex (2.5 g, 10 mmol, 52%), which freezes to
form oily white crystals when cooled to ꢀ208C, was isolated.
acryloyl chloride (6.68 mL, 74.7 mmol) was added dropwise while the
temperature was controlled at ꢀ10 ꢁ 28C. The mixture was stirred at
ꢀ108C for 3 h, after which time, water (2.6 mL) was added slowly, fol-
lowed by 15% aqueous NaOH (2.6 mL) and more water (2.6 mL). The
resulting slurry was stirred for 1 h and then filtered. The filtrate was con-
centrated at 608C and the resulting crude oil was used in the next step
without further purification. 1H NMR (400 MHz, CDCl3): d=6.01 (1H,
dd), 5.20 ppm (2H, m).
1,1-Dideuterioallyl tosyl ester: A 250 mL Schlenk flask was charged with
1,1-dideuterioallyl alcohol (5 g, 82.2 mmol, crude product from previous
step), tosyl chloride (15.8 g, 83.2 mmol) and anhydrous ether (80 mL).
The mixture was cooled to 08C and powdered NaOH (9.15 g, 0.228 mol)
was added in portions under N2. The reaction was then warmed to room
temperature and stirred for 12 h. The precipitate was filtered and the fil-
trate concentrated in vacuo. The resulting oil was subjected to column
chromatography (silica gel, 90:10 hexane/EtOAc) to the yield pure prod-
uct (16.5 g, 95%). 1H NMR (300 MHz, CDCl3): d=7.80 (2H, d), 7.33
(2H, d), 5.81 (1H, dd), 5.28 (1H, d, 3J=17.4 Hz, 2J2 =1.6 Hz), 5.16 (1H,
7-Vinyl-7-azabicyclo[4.1.0]heptane (5): Copper(II) acetate (0.92 g,
G
4.6 mmol, >1 equiv) and 4 Š molecular sieves (0.3 g) were added to a
flame-dried 50 mL Schlenk flask. 2,4,6-Trivinylcyclotriboroxane–pyridine
complex (0.33 g, 2.0 mmol, 6.0 mmol vinyl equivalents), pyridine (0.9 mL,
11.2 mmol) and cyclohexene imine (0.35 mL, 3.6 mmol) were added in
succession, and the system was connected to a drying tube with oxygen
flow-through. The reaction was stirred for 10 h, poured into pentanes
(75 mL) and then filtered through a 1 cm pad of silica gel. The filtrate
was reduced to 3–4 mL in vacuo, and the resulting liquid was distilled
under reduced pressure to yield 235 mg of 7-vinyl-7-azabicyclo-
2
d, 3J=10.2 Hz, J2 =1.6 Hz), 2.46 ppm (3H, s).
N-1-(1,1-Dideuterioallyl)-2-methyl-3-phenylaziridine (9): A 25 mL flask
was charged with 2-methyl-3-phenylaziridine[11] (220 mg, 1.66 mmol),
K2CO3 (382 mg, 2.76 mmol) and DMF (4 mL). A solution of 1,1-dideuter-
ioallyl tosyl ester (296 mg, 1.38 mmol) in DMF (2 mL) was added to the
reaction mixture and stirred at room temperature for 12 h. When TLC
showed no starting material remained, water (15 mL) was added and the
solution was extracted with EtOAc (2”10 mL). The combined organic
layers were washed with brine (15 mL) then dried over Na2SO4. T he
drying agent was filtered off and filtrate was concentrated in vacuo. The
resulting oil was subjected to column chromatography (silica gel, 80:20
hexane/EtOAc) to yield pure 9 (240 mg, 80%). 1H NMR (400 MHz,
CDCl3): d=7.32 (4H, m), 7.22 (1H, m), 5.99 (1H, dd), 5.25 (1H, d, 3J1 =
17.2, 2J2 =2 Hz), 5.11 (1H, dd, 3J1 =10.4, 2J2 =2 Hz), 2.53 (1H, d, J=
6.8 Hz), 1.80 (1H, m), 0.94 ppm (d, J=5.6 Hz).
A
164.5 mg of product or 27%, GC time: 4.8 min). 1H NMR (300 MHz,
CDCl3): d=6.33 (1H, dd, J1 =15.2, J2 =7.8 Hz), 4.46 (1H, d, J=15.2 Hz),
4.35 (1H, d, J=7.9 Hz), 1.70–1.95 (m, 6H), 1.40 (m, 2H), 1.25 ppm (m,
2H).
7-Ethyl-7-azabicyclo
A
(6):
[Rh(CO)H
A
(6 mg,
0.0065 mmol, 1 mol%) was dissolved in THF (6 mL) in a dry 25 mL
Schlenk flask with stirrer bar. The flask was flushed twice with H2, filled
with H2 (1 atm), then 7-vinyl-7-azabicyclo[4.1.0]heptane 5 (80 mL,
G
0.65 mmol) was added by syringe. The reaction was followed by GC anal-
ysis and observed to have slowed after two days, thus 6 mg of additional
catalyst were added every two days. After six days (and two 6 mg supple-
ments of catalyst), 32% conversion was achieved. The solvent was evapo-
rated in vacuo, and chromatographed on silica gel (7:3 hexanes/EtOAc, I2
7,7-Dibromobicyclo
[4.1.0]heptane:[26] Benzyltriethylammonium chloride
G
(293 mg, 1.29 mmol, 2 mol%) and cyclohexane (6.08 mL, 4.93 g,
60.0 mmol) were mixed in a flask under N2. The suspension was cooled
to 08C, then dichloromethane (6 mL), absolute EtOH (0.25 mL), bromo-
form (7.9 mL, 22.8 g, 90.3 mmol) and 50% aqueous NaOH (30 mL) were
added in succession. The solution turned light brown, and was stirred for
16 h as the bath slowly warmed to room temperature. Water was added
and the slurry was extracted with hexanes (3”100 mL), then dried with
anhydrous MgSO4. Solvent was evaporated under reduced pressure, then
the oil was chromatographed on silica gel (hexanes) to yield 15.8 g of a
clear, colourless liquid (Rf =0.9, GC time: 10.5 min) of crude 7,7-
1
stain) to yield trace amounts of 6 (Rf =0.15, GC time: 4.4 min). H NMR
(400 MHz, CDCl3): d=2.13 (q), 1.80 (m), 1.30 (m), 1.15 (t); ESI-MS: m/z
(%): 126.1 (75) [M+1+], 127.1 ppm (10) [M+2+].
7-(3-Methylbut-2-enyl)-7-azabicyclo a 50 mL
[4.1.0]heptane (7):[9] In
G
Schlenk flask, K2CO3 (1.35 g, 13.4 mmol) was suspended in anhydrous
acetone (30 mL). Cyclohexene imine (500 mL, 4.9 mmol) was added by
syringe, followed by the dropwise addition of prenyl bromide (575 mL,
4.9 mmol). The reaction was stirred for 1.5 h under N2, filtered, and ace-
tone was removed under reduced pressure. The resulting yellow oil was
chromatographed on silica gel (95:5!90:10 hexanes/EtOAc) to isolate
667 mg of pure 7 in 79% yield (Rf =0.25 under 95:5 EtOAc/hexanes, GC
time: 9.3 min). The branched product (7-(2-methylbut-3-en-2-yl)-7-
dibromobicyclo
[4.1.0]heptane (1258C =2 gmLꢀ1) contaminated with bro-
R
1
moform. This compound was used without further purification. H NMR
(400 MHz, CDCl3): d=2.00 (2H, m), 1.83 (2H, m), 1.58 (2H, m), 1.35
(2H, m), 1.18 ppm (2H, m); 13C NMR (125 MHz, CDCl3): d=41.1, 27.3,
20.9, 20.4 ppm.
[4.1.0]heptane:[26] 7,7-Dibromobicyclo
[4.1.0]heptane
A
azabicyclo[4.1.0]heptane) was found at Rf =0.5 but was not isolated.
C
trans-7-Bromobicyclo
(1.5 mL, 3.0 g, 11.9 mmol) in THF (20 mL) was cooled to ꢀ958C under
N2, at which point, n-butyllithium (8.2 mL, 1.5m in hexanes, 12.3 mmol)
was added. The solution was stirred for 10 min, quenched with absolute
EtOH (4 mL) and stirred for 1 h while allowed to warm to room temper-
ature. Water (15 mL) was poured into the solution, and the layers were
extracted with hexanes (4”30 mL). The extracts were combined, dried
with MgSO4, then reduced in vacuo. The oil was chromatographed on
silica gel (hexanes, Rf =0.95, GC time: 6.9 min) to yield 1.9 g crude trans-
1.71 (3H, m), 1.58 (3H, s), 1.50–1.80 (8H, m), 1.15 ppm (2H, m).
7-(Cyclohex-2-enyl)-7-azabicyclo[4.1.0]heptane (8): In a 100 mL round-
G
bottomed flask equipped with septum and stirrer bar, K2CO3 (945 mg,
6.8 mmol) was suspended in anhydrous acetone (30 mL). Cyclohexene
imine (350 mL, 3.4 mmol) was added by syringe, followed by 3-bromocy-
clohexene (394 mL, 549 mg, 3.4 mmol). The reaction was left stirring for
22 h, filtered, and the acetone was removed under reduced pressure. The
resulting oil was chromatographed on silica gel (95:5!90:10 hexanes/
EtOAc) to yield 264 mg of a clear, colourless oil 8 in 44% isolated yield
(GC time: 11.2 min). 1H NMR (400 MHz, CDCl3): d=5.75 (1H, m), 5.65
(1H, m), 2.10 (1H, m), 1.95 (1H, m), 1.80 (m), 1.55 (m), 1.35 (2H, m),
1.16 ppm (2H, m); ESI-MS: m/z (%): 178.1 (100) [M+1+], 179.1 (13)
[M+2+].
7-bromobicyclo[4.1.0]heptane, contaminated with hexanes. This com-
G
pound was used without further purification. 1H NMR (300 MHz,
CDCl3): d=2.60 (1H, t, J=3.5 Hz), 1.80 (2H, m), 1.70 (2H, m), 1.38
(2H, m), 1.10–1.30 ppm (4H, m); 13C NMR (100 MHz, CDCl3): d=26.0,
22.7, 21.5, 21.1 ppm.
7-Allylbicyclo
[4.1.0]heptane (10):[27] Copper(I) iodide (0.75 g, 3.94 mmol)
H
1,1-Dideuterioallyl alcohol: A 250 mL flame-dried three-necked flask
fitted with magnetic stirrer bar, thermometer, addition funnel and nitro-
gen inlet was charged with LiAlD4 (2.2 g, 47.6 mmol) and anhydrous
ether (100 mL). The mixture was cooled to ꢀ108C and a solution of
was suspended in THF (14 mL) under N2 at ꢀ488C. To this solution, n-
butyllithium (5.7 mL, 1.4m in hexanes, 7.98 mmol) was added, followed
by trans-7-bromobicyclo
A
30 min, allyl bromide (1 mL, 1.40 g, 11.6 mmol) was added and the reac-
892
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 886 – 894

Isomerisation of N-Allyl Aziridines
FULL PAPER
tion was stirred for a further 30 min. The reaction was quenched with
MeOH (4 mL), then water. The copper salts were filtered through a pad
of Celite and then the filtrate was extracted with hexanes. Hexane ex-
tracts were dried with MgSO4 and then chromatographed on silica gel
(hexanes, Rf =0.95, GC time: 6.05 min) to yield 167 mg of crude 9, conta-
minated with hexanes. 1H NMR (400 MHz, CDCl3): d=5.88 (1H, m),
5.05 (1H, d, J=17.2 Hz), 4.95 (1H, d, J=10.2 Hz), 1.94 (2H, m), 1.85
(2H, m), 1.62 (2H, m), 1.10–1.40 (4H, m), 0.60 (2H, m), 0.43 ppm (1H,
m).
reaction (t1/2 =142.2 min). The ratio of Z/E remained constant at 80:20
but the peaks decreased in intensity, likely due to degradation of the
highly unstable enamine product. The product propenyl-1-amine was not
isolated, and only one vinylic proton from the product was clearly dis-
cernible by NMR spectroscopy. 1H NMR (400 MHz, C6D6): d=4.49
3
(0.2H, dq, 3Jtrans =13.3 Hz; CH3CH=CHNH2), 4.31 (0.8H, dq, Jcis
=
8.4 Hz; CH3CH=CHNH2).
Chlorocarbonylbis(triphenylphosphine)rhodium(I):[28]
Triphenylphos-
phine (1.1 g, 4.2 mmol) was dissolved in absolute EtOH (43 mL) in a
250 mL three-necked flask equipped with condenser and stirrer bar
under N2. The solution was heated to reflux, then RhCl3·3H2O (284 mg,
1.1 mmol) suspended in hot EtOH (10 mL) was added by syringe. To the
red solution, 40% aqueous formaldehyde (5 mL) was quickly added. The
suspension was refluxed for 5 min, cooled, filtered, and dried in vacuo to
General protocol for the isomerisation of N-allylaziridines to N-vinylazir-
idines:[11] In a flame-dried 25 mL Schlenk flask equipped with stirrer bar,
[Rh(CO)H(PPh3)3] (30 mg, 0.03 mmol, 1.5 mol%) was dissolved in THF
G
or C6H6 (8 mL). N-Allylaziridine 1 (2.0 mmol) was added via syringe and
the solution was stirred overnight. Progress of the reaction may be moni-
tored by GC analysis. Reactions were worked up by one of two methods:
recover 699 mg [Rh(CO)Cl(PPh3)2] as pale-yellow crystals in 94% yield.
H
M.p.: 195–1998C decomp., lit. 195–1978C; 1H NMR (300 MHz, CDCl3):
d=7.70 (12H, m), 7.40 ppm (18H, m); 31P NMR (120 MHz, CDCl3): d=
30.0 ppm (d, J=127 Hz).
Method A, large scale: THF was evaporated under reduced pressure to
1 mL of solution, then the residue was distilled under reduced pressure
to produce clear, colourless N-vinylaziridine 2 (1.2 mmol, 60%).
Deuteriocarbonyltris(triphenylphosphine)rhodium(I):[19,29] In
a 250 mL
Method B, microscale: The reaction solution was poured into pentanes
(20–40 mL) and filtered on a pad of Celite. The filtrate was reduced in
vacuo and the residue washed with pentanes and filtered again. This can
be repeated once more to isolate a clear, yellow oil of N-vinylaziridine 2,
contaminated with trace amounts of rhodium salts and triphenylphos-
phine oxide.
three-necked flask equipped with condenser and stirrer bar, chlorocarbo-
nylbis(triphenylphosphine)rhodium(I) (599 mg, 0.89 mmol) and PPh3
(0.9 g, 3.4 mmol) were suspended in EtOD (20 mL) and heated to reflux.
NaBD4 (300 mg, 7.2 mmol) in EtOD (20 mL) was added over one
minute. The resulting solution was refluxed for 20 min, cooled, and fil-
tered. The collected product was washed once with EtOH and once with
7-(Prop-1-enyl)-7-azabicyclo[4.1.0]heptane (2): Procedure as above, fol-
U
ether, then dried in vacuo to yield 770 mg [Rh(CO)D(PPh3)3] in 97%
C
lowing workup method A. Conversion to product: 97%; isolated yield:
57% (GC time: Z isomer: 6.5 min, 90%; E isomer: 6.8 min, 10%).
1H NMR (300 MHz, CDCl3): d=5.95 (d, 0.1H, J=13.5 Hz), 5.78 (1H, d,
J=7.9 Hz), 5.02 (0.1H, m), 4.78 (1H, m), 1.67 (3H, dd), 1.56 (0.3H, dd),
1.10–2.00 ppm (m).
yield (97 atom D%). 1H NMR (400 MHz, C6D6): d=7.40 (18H, m),
6.90 ppm (27H, m); 31P NMR (120 MHz, C6D6): d=40.5 ppm (d, J=
155 Hz); 2H NMR (60 MHz, C6H6): d=ꢀ9.3 ppm (brs).
1-[(E)-Prop-1-enyl]piperidine (4):[8] Procedure as above, following
workup method A. Isolated yield: 31% (GC time: 5.8 min). 1H NMR
(400 MHz, CDCl3): d=5.83 (1H, d, J=13.9 Hz), 4.40 (1H, m), 2.73 (4H,
m), 1.50–1.70 ppm (6H, m).
Acknowledgements
7-(Prop-1-enyl)bicyclo[4.1.0]heptane (11): Procedure as above, following
A
workup method B. 94% conversion was achieved, as determined by GC
analysis (GC time: Z isomer: 6.8 min; E isomer: 7.0 min). 1H NMR
(400 MHz, CDCl3): d=5.43 (0.6H, dq, J1 =15, J2 =7 Hz), 5.31 (1H, dq,
J1 =11.5, J2 =7 Hz), 4.95 (0.6H, m), 4.80 (1H, m), 1.72 (3H, dd), 1.64
(1.8H, dd), 0.85–1.90 ppm (10H, m).
We gratefully acknowledge financial support from the NSERC. D.T.
thanks the NSERC for an undergraduate summer research award
(USRA). F.-A.A. is grateful to MCMM-MRC Global Health for a post-
doctoral fellowship.
Deuterium crossover study: Procedure as above with 1 equiv of each azir-
idine and [Rh(CO)H(PPh3)3] (3 mol%) in THF for 12 h at room temper-
G
[1] a) G. Stork, R. Terrell, J. Szmuszkovicz, J. Am. Chem. Soc. 1954, 76,
2029–2030; b) The Chemistry of Enamines (Parts 1 & 2) (Ed.: Z.
Rappoport), Wiley, New York, 1994; c) R. E. Gawley, J. Aubꢂ in
Principles of Asymmetric Synthesis (Eds.: J. Baldwin, R. M. Wil-
liams, J.-E. Bꢃckvall), Tetrahedron Organic Chemistry Series 14, Per-
gamon, Oxford, 1996.
[2] S. Escoubet, S. Gastaldi, M. Bertrand, Eur. J. Org. Chem. 2005,
3855–3873.
[3] J. Barluenga, C. Valdes, Chem. Commun. 2005, 4891–4901.
[4] S. Inoue, H. Takaya, K. Tani, S. Otsuka, T. Sato, R. Noyori,J. Am.
Chem. Soc. 1990, 112, 4897–4905.
[5] R. Noyori, Angew. Chem. 2002, 114, 2108–2123; Angew. Chem. Int.
Ed. 2002, 41, 2008–2022, and references therein.
[6] S. Otsuka, K. Tani, Synthesis 1991, 9, 665–680.
[7] F. H. Jardine, Polyhedron 1982, 1, 569–605.
[8] S. Krompiec, M. Pigulla, M. Krompiec, B. Marciniec, D. Chadyniak,
J. Mol. Catal. A 2005, 237, 17–25.
[9] I. D. G. Watson, S. A. Styler, A. K. Yudin, J. Am. Chem. Soc. 2004,
126, 5086–5087.
ature. The reaction was not worked up; rather, an aliquot of solution was
run under ESI-MS in which the [M+2+] (137+2=139) peak was detect-
ed, indicating the presence of deuterated product beyond normal isotopic
distributions.
Preparative photochemical synthesis of (E)-N-vinylaziridine 2: An ali-
quot of 2 (50 mL), as prepared by rhodium(I)-catalysed isomerisation of
1, was added by syringe to dry THF (10 mL) purged with N2 for 10 min.
The sample was irradiated for 6 h through quartz (lmax =254 nm), then
solvent was evaporated in vacuo to leave an oil (0.5 mL). GC analysis in-
dicated isomerisation to an equilibrium mixture of (Z)- and (E)-N-vinyl-
aziridine (Z/E 54:46).
Compounds listed in Table 1 without bold numerical labels were reported
in reference [11], a previous communication from our group.
General protocol for NMR spectroscopic experiments: A dry NMR tube
was loaded with [Rh(CO)H(PPh3)3] (20.1 mg, 0.022 mmol, 1 equiv) and
G
C6D6 (0.75 mL). The aziridine (0.022 mmol) was then dispensed into the
tube by syringe. The tube was sealed with a NMR cap, shaken and then
spectra were taken at ten minute intervals on a 400 MHz spectrometer at
258C. Product ratios for kinetics, if applicable, were measured by
1H NMR spectroscopic integration.
[10] a) I. D. G. Watson, A. K. Yudin, J. Am. Chem. Soc. 2005, 127,
17516–17529; b) I. Dubovyk, I. D. G. Watson, A. K. Yudin, J. Am.
Chem. Soc. 2007, 127, in press.
Propenyl-1-amine: Procedure as above, except 15 mol% [Rh(CO)H-
A
[11] F.-A. Alphonse, A. K. Yudin, J. Am. Chem. Soc. 2006, 128, 11754–
11755.
A
[12] T. Mecozzi, M. Petrini, Synlett, 2000, 73–74.
[13] R. Cramer, J. Am. Chem. Soc. 1966, 88, 2272–2282.
Chem. Eur. J. 2008, 14, 886 – 894
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
893

A. K. Yudin et al.
[14] a) D. Bingham, D. E. Webster, P. B. Wells, J. Chem. Soc. Dalton
Trans. 1974, 1514–1518; b) ibid, J. Chem. Soc. Dalton Trans. 1974,
1519–1521.
[15] M. Yagupsky, G. Wilkinson, J. Chem. Soc. A 1970, 941–944.
[16] W. Strohmeier, R. Fleischmann, W. Rehder-Stirnweiss, J. Organo-
met. Chem. 1973, 47, C37–40.
[17] P. Golbon, F. Scheinmann, J. Chem. Soc. Perkin Trans. 1 1973, 2870–
2875.
[18] J. K. Stille, Y. Becker, J. Org. Chem. 1980, 45, 2139–2145.
[19] C. OꢀConnor, G. Wilkinson, J. Chem. Soc. A 1968, 2665–2671.
[20] R. H. Crabtree, The Organometallic Chemistry of the Transition
Metals, 3rd ed, Wiley, New York, 2001.
[23] S.-K. Kang, H.-W. Lee, S.-B. Jang, T.-H. Kim, S.-J. Pyun, J. Org.
Chem. 1996, 61, 2604–2605.
[24] A. F. Burchat, J. M. Chong, N. Nielsen, J. Organomet. Chem. 1997,
542, 281–283.
[25] F. Kerins, D. F. OꢀShea, J. Org. Chem. 2002, 67, 4968–4971.
[26] A. Inoue, J. Kondo, H. Shinokubo, K. Oshima, Chem. Eur. J. 2002,
8, 1730–1740.
[27] K. Kitatani, T. Hiyama, H. Nozaki, Bull. Chem. Soc. Jpn. 1977, 50,
3288–3294.
[28] D. Evans, J. A. Osborn, G. Wilkinson, Inorg. Synth. 1969, 11, 99.
[29] C. Krug, J. F. Hartwig, J. Am. Chem. Soc. 2002, 124, 1674–1679.
[21] P. Rademacher, G. Irsch, W. Sicking, E.-U. Wꢄrthwein, J. Mol.
Struct. 1989, 197, 291–305.
Received: August 24, 2007
[22] A. G. M. Barrett, M. A. Seefeld, Tetrahedron 1993, 49, 7857–7870.
Published online: November 8, 2007
894
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 886 – 894