Allylic amination and aziridination of olefins by aryl azides catalyzed by CoII(tpp): A synthetic and mechanistic study

Article abstract of DOI:10.1002/ejic.200800156

Co^II(tpp) catalyzes the reaction of aromatic azides (ArN ₃) with nonactivated olefins to yield allylic amines or aziridines in moderate-to-good yields. The chemoselectivity of the catalytic reaction is particularly high. Depending on the substrate employed, allylic amines or aziridines can be obtained. The reaction mechanism was investigated, and the reaction proceeds through reversible coordination of the aryl azide to the Co^II-porphyrin complex. The often postulated nitrene complex is not an intermediate in this reaction. The kinetics for the allylic amination is first order in azide, Co(tpp), and olefin. For the aziridination, the kinetics is again first order in azide and catalyst, but we observed a first-order dependence of the rate on α-methylstyrene only up to an olefin concentration of 6.9 M. An inhibiting role of the competitively formed 1-(4-nitrophenyl)-5-methyl-5-phenyl-1,2,3-triazoline was identified. The triazoline was shown to reversibly coordinate to Co(tpp), which blocks the free coordination site necessary for the catalytic reaction to proceed, and is it responsible for the catalyst deactivation in the aziridination reaction of α-methylstyrene by 4-nitrophenyl azide. Wiley-VCH Verlag GmbH & Co. KGaA, 2008.

Full text of DOI:10.1002/ejic.200800156

FULL PAPER
DOI: 10.1002/ejic.200800156
Allylic Amination and Aziridination of Olefins by Aryl Azides Catalyzed by
CoII(tpp): A Synthetic and Mechanistic Study
Alessandro Caselli,[a] Emma Gallo,[a] Simone Fantauzzi,[a] Simona Morlacchi,[a]
Fabio Ragaini,[a] and Sergio Cenini*[a]
Keywords: Cobalt / Porphyrinoids / Azides / Homogeneous catalysis / C–H activation
CoII(tpp) catalyzes the reaction of aromatic azides (ArN3)
with nonactivated olefins to yield allylic amines or aziridines
in moderate-to-good yields. The chemoselectivity of the cata-
lytic reaction is particularly high. Depending on the substrate
employed, allylic amines or aziridines can be obtained. The
reaction mechanism was investigated, and the reaction pro-
ceeds through reversible coordination of the aryl azide to the
CoII–porphyrin complex. The often postulated “nitrene”
complex is not an intermediate in this reaction. The kinetics
for the allylic amination is first order in azide, Co(tpp), and
olefin. For the aziridination, the kinetics is again first order
in azide and catalyst, but we observed a first-order depen-
dence of the rate on α-methylstyrene only up to an olefin
concentration of 6.9
M. An inhibiting role of the competitively
formed 1-(4-nitrophenyl)-5-methyl-5-phenyl-1,2,3-triazoline
was identified. The triazoline was shown to reversibly coordi-
nate to Co(tpp), which blocks the free coordination site nec-
essary for the catalytic reaction to proceed, and is it responsi-
ble for the catalyst deactivation in the aziridination reaction
of α-methylstyrene by 4-nitrophenyl azide.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2008)
fair to say that no truly general method for direct catalytic
C–N bond formation exists.
Introduction
Iminophenyliodinanes[4] have been extensively employed
as nitrogen sources in the last few years, but several prob-
lems are strictly related to their use.[5] The synthetic pro-
cedure to obtain these molecules does not allow the prepa-
ration of a large class of starting materials, the side product
of the reaction is iodobenzene, and, last but not least, the
sulfonyl group is always introduced in the aminated com-
pound. It is worth noting that iminoiodinane compounds
can be also prepared in situ from the corresponding sulfon-
ylamine and diacetoxyiodobenzene[6] or iodosylbenzene.[7]
Alternative nitrene sources such as chloramine-T[8] or bro-
mamine-T[9] have been employed to overcome some of the
limitations associated with the use of iminoiodinane com-
pounds. To get over the problem of the limited solubility of
these compounds in the commonly employed organic sol-
vents and to avoid the presence of trace amounts of water
that inevitably contaminates these last starting materials,
some years ago we considered an alkylammonium salt of N-
chloro-p-toluensulfonamide as an aminating agent.[10] Very
recently, us and Katsuki and coauthors published two re-
views on the use of azide compounds as a nitrogen source
for atom-efficient and ecologically benign nitrogen-atom-
Among the many different feasible catalytic transforma-
tions, C–C, C–O, and C–N bond-forming reactions have
received special attention, and consequently, they have had
a particularly high impact on both academic and industrial
research.[1] This is due to the versatility of these methodolo-
gies for the construction of important structural motifs. For
organic synthesis, the introduction of nitrogen functional
groups is as important as that of oxygen functional
groups.[2] Although reactions for introducing these func-
tional groups include isoelectronic activated species and
their reaction patterns are analogous (i.e., aziridination vs.
epoxidation and sulfimidation vs. sulfoxidation), nitrogen-
atom transfer reactions have been studied far-less fre-
quently.[3] This is partly due to the lack of appropriate nitro-
gen sources, as opposed to oxidants, which are available in
a wide variety of forms. Thus, it has been out of necessity
that the synthetic methods that have been developed for the
oxidative formation of the C–N bond have been of type
usually distinct than those suitable for oxidative C–O bond
formation. Catalytic aziridination and C–N bond-forming
reactions have been slow to develop, so much that it is still
[a] Dipartimento di Chimica Inorganica, Metallorganica e Analit- transfer reactions.[11,12] In particular, Katsuki obtained
ica “Lamberto Malatesta”, Università degli Studi di Milano,
interesting results with the use of tosylazide as a nitrene
and ISTM-CNR
precursor even in enantioselective reactions by using chiral
ruthenium Schiff base complexes as catalysts.
Organic azides are the precursors of the nitrene moiety
[RN] and the only side product of the decomposition is
Via Venezian 21, 20133 Milano, Italy
Fax: +39-02-50314405
E-mail: Sergio.Cenini@unimi.it
Supporting information for this article is available on the
WWW under http://www.eurjic.org or from the author.
Eur. J. Inorg. Chem. 2008, 3009–3019
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3009

A. Caselli, E. Gallo, S. Fantauzzi, S. Morlacchi, F. Ragaini, S. Cenini
FULL PAPER
molecular nitrogen. For this reason, they can be considered
atom-efficient nitrogen-atom-transfer agents. Organic
azides, depending on the experimental conditions and the
nature of the metal and the azide, can react with the metal
complex to give a metal–imido species or a metal–azide ad-
duct. The formation of a metal–nitrogen bond can occur
with coordination of the terminal nitrogen Nγ or Nα atom
or by oxidative addition of the azido group. To predict the
coordinating mode of the organic azide it is important to
remember that the nitrogen in the α position is more basic
than the terminal unsubstituted one. Among azides, aro-
matic azides present several advantages: (i) They can be eas-
ily synthesized from the corresponding substituted anilines
by a simple procedure. (ii) They allow an N-aryl group to
be introduced into the molecule when it is necessary. (iii)
Scheme 1. Benzylic amination reactions catalyzed by CoII porphy-
rin complexes.[24]
Results and Discussion
The reaction of 4-nitrophenyl azide with olefins in ben-
They are stable enough to be safely handled in the labora- zene at 75 °C can be efficiently catalyzed by Co(tpp) to give
tory. the corresponding allylic amines (A) or aziridines (B), de-
In our ongoing study on amination reactions catalyzed pending on the substrate used, in moderate-to-good yields
by transition metals,[13,14] we recently reported the use of (Scheme 2, Table 1). 4-Nitroaniline (C) was always observed
aromatic azides in the amination of unsaturated hydro- among the products. Whereas several selective catalysts for
carbons catalyzed by ruthenium[14–16] and cobalt[17,18] por- the aziridination reaction were recently reported, the selec-
phyrin complexes. The metal coordination mode of the por- tive insertion of nitrenoids into the allylic C–H bond has
phyrin complexes renders them a unique class of catalysts been explored only to a lower extent.[26] Much work is
for N-group transfer reactions,[19] and indeed, the first tran- needed to understand the mechanism of the alkene acti-
sition-metal complexes that showed good activity as cata- vation. In the present study, we chose four substrates with
lysts in aziridination reactions were metalloporphyrins.[20] allylic C–H bonds that can undergo allylic amination and
Cobalt porphyrin complexes have been recently shown to compared them with styrene and α-methylstyrene, which
be competent catalysts in the aziridination reaction of al- are commonly employed in aziridination reactions. All reac-
kenes by employing bromamine-T[21] or diphenylphos- tions reported in Table 1 were carried out in a 1:1 volume
phoryl azide[22] as the aminating agents. The same authors mixture of the olefin and benzene at 75 °C in order to have
also reported the use of cobalt–porphyrin complexes as cat- comparable results.
alysts for intramolecular C–H amination with arylsulfonyl
When cyclohexene was employed as the substrate
(Table 1, entry 1), allylic amine 1 was formed in good yields
azides.[23]
We reported that cobalt(II) porphyrin complexes are able (60% isolated yield) in 6 h. The only byproduct formed in
to activate aromatic azides for the amination of the C–H this reaction was 4-nitroaniline (39%) and no trace of azir-
bond of saturated organic compounds under mild condi- idine 2 was ever detected. Aziridine 2 is the only aminated
tions to give secondary amines and imines (Scheme 1).[24,25] olefin-based product that is obtained from the uncatalyzed
We report herein that cobalt(II) tetraphenylporphyrinate, reaction (see “Mechanistic Studies”). When the same cata-
Co(tpp) (tpp = dianion of tetraphenylporphyrin), can be lytic reaction was repeated in refluxing neat cyclohexene
used as a catalyst for the selective aziridination or amin- (b.p. 83 °C) we again isolated allylic amine 1 in 60% yield
ation of the allylic C–H bond of alkenes depending on the and in just 3 h.
nature of the substrate employed. Particular attention was
In contrast, in the catalytic reaction with cyclopentene,
given to the understanding of the mechanism of such reac- at complete conversion of the starting azide, allylic amine 3
tions with the aim to be able to direct the catalytic cycle was obtained in low yield (9%; Table 1, entry 2) and a ne-
towards the formation of the desired product.
arly equal amount of aziridine 4 (11%) was formed together
Scheme 2. Amination reactions catalyzed by Co(tpp).
3010
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 3009–3019

Co-Catalyzed Allylic Amination and Aziridination of Olefins by Aryl Azides
Table 1. Amination reactions of alkenes with 4-nitrophenyl azide fully characterized (Scheme 3). The rest of the starting azide
(ArN3) catalyzed by Co(tpp).[a]
was again recovered as 4-nitroaniline (see Supporting Infor-
mation). Triazoline 5 is not stable upon heating over 50 °C
and readily looses dinitrogen to yield the aziridine. This can
easily explain the loss of selectivity observed in the catalytic
reaction conducted in refluxing cyclopentene/benzene, 1:1.
Two competitive reactions occur at the same time: on the
one side, there is the Co(tpp)-catalyzed reaction yielding al-
lylic amine 3, whose rate is decreased by the fact that the
temperature is lower than 75 °C, and on the other side,
there is the uncatalyzed reaction giving triazoline 5, which
is readily converted into aziridine 4 by effect of the in-
creased temperature with respect to boiling cyclopentene.
[a] Experimental conditions: Co(tpp) (32 mg, 0.048 mmol) in ben-
zene (15 mL) and the olefin (15 mL) at 75 °C; Co(tpp)/ArN3, 4:50.
[b] Conversion of the starting azide. [c] Isolated yield. [d] Mixture
of two diastereoisomers in 55:45 ratio. [e] Mixture of four dia-
stereoisomers in 58:21:15:6 ratio. [f] Co(tpp): 8 mg (0.012 mmol)
was used; Co(tpp)/ArN3, 1:50.
Scheme 3. Reaction between cyclopentene and 4-nitrophenyl azide
in the presence or absence of Co(tpp).
To understand whether electrophilic attack of the azide
occurs at the double bond (with aziridine formation fol-
lowed by a β C–H insertion reaction[27] to yield the allylic
amine) or directly at the allylic C–H bond, two differently
substituted cyclohexenes were chosen as substrates for the
catalytic reactions, namely, 1-methylcyclohexene and 4-
methylcyclohexene. From the reaction with 1-methylcy-
clohexene one could expect to obtain two different sets of
diastereoisomers if the attack occurs at one or the other
with 4-nitroaniline (69%). It should be pointed out that at position (Scheme 4).
75 °C, the 1:1 mixture of cyclohexene and benzene is below
The catalytic reaction, upon complete conversion of the
the reflux temperature of both solvents, whereas the same starting azide, yielded 47% of allylic amines 6 and 7 in a
1
consideration does not hold with the use of cyclopentene, 45:55 ratio, as determined by H NMR spectroscopy. No
which has a boiling point of 44–46 °C. We know that the trace of allylic amine 8 was observed (Table 1, entry 3). It
temperature has a strong influence on the product distribu- should be pointed out that, similarly to what was observed
tion and on the catalytic activity of Co(tpp) in this reaction. for the radical reactions, the attack at the endocyclic CH2
In fact, when 4-nitrophenyl azide was treated in refluxing is faster than attack at the exocyclic CH3 and no trace of
cyclopentene in the absence of the catalyst, complete con- aminated products derived from CH3 activation were de-
version was observed in 5 h, and the only aminated olefin- tected. A similar result was obtained from the reaction with
based product obtained was triazoline 5 (41%), which was 4-methylcyclohexene (Table 1, entry 4). In this case, the re-
Scheme 4. Possible outcome of the Co(tpp)-catalyzed reaction between 1-methylcyclohexene with 4-nitrophenyl azide depending on the
attack site.
Eur. J. Inorg. Chem. 2008, 3009–3019
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3011

A. Caselli, E. Gallo, S. Fantauzzi, S. Morlacchi, F. Ragaini, S. Cenini
FULL PAPER
Figure 1. Allylic amines obtained from the catalyzed reaction between 4-methylcyclohexene and 4-nitrophenyl azide.
action yielded a mixture of four diastereoisomers in a ethylene. These results suggest that steric hindrance at the
global 47% yield in a 58:21:15:6 ratio. By comparison with olefinic substrate represents a major obstacle to the cata-
the 1H NMR spectra reported in the literature,[28] the major lytic reactivity. Different results were obtained with 1,2-di-
products were determined to be 9-trans and 9-cis (Figure 1); hydronaphthalene and 1,2-dihydro-4-phenylnaphthalene: in
the minor products were determined to be 10-trans and 10- these cases, the only aminated products obtained were the
cis derived from the other possible nitrene insertion into the benzylic amines derived from disproportionation reactions
more-hindered allylic C–H bond. In both cases, products of the starting olefins. These results will be the subject of a
derived from attack at the double bond were not observed, forthcoming paper.
and we can conclude that allylic amine formation occurs by
attack of the olefin to the cobalt-coordinated azide (see also
Mechanistic Studies).
Mechanistic Studies
When linear alkenes such as 1-octene were employed in
the present system, no reaction occurred. Conjugated
dienes also failed to react. In contrast, when styrene, which
has no allylic C–H bonds, was employed as a substrate, the
reaction product was the corresponding aziridine 11, albeit
in low yields (Table 1, entry 5). A similar result was ob-
tained with α-methylstyrene, even if in principle this olefin
could also give the allylic amine (Table 1, entry 6). In this
case, together with aziridine 12 (38%), 1-(4-nitrophenyl)-5-
methyl-5-phenyl-1,2,3-triazoline (13, 30%) was obtained
(Scheme 5). This particular triazoline is surprisingly stable,
as already reported by us some years ago.[15] Its formation
and the importance this compound has on the overall cata-
lytic cycle will be further discussed in the mechanistic stud-
ies reported below. Once formed, 13 is not converted into
12 by Co(tpp) under the reaction conditions. We can also
exclude the formation of 4-nitro-N-(2-phenylallyl)aniline
(14), which was synthesized by an independent route with
a methodology reported recently in our laboratory (see
Supporting Information).[29] This latter compound was
never detected among the products of the catalytic reactions
and remains unchanged under the reaction conditions.
Uncatalyzed Reaction
The uncatalyzed reaction between aromatic azides and
double bonds leading to the formation of 1,2,3-∆2-triazol-
ines was reported almost a century ago.[30] Nonactivated
double bonds give very slow reactions, and the triazoline in
most cases is not stable enough to be isolated. The aziridine
is obtained instead, although often in low yields. For exam-
ple, in our hands, the reaction of 4-nitrophenyl azide in re-
fluxing cyclohexene gave complete conversion of the start-
ing azide in 16 h to yield aziridine 2 with 70% selectivity
(the rest of the azide was transformed into 4-nitroaniline).
It should be pointed out that the aziridine is not an inter-
mediate in our catalytic reaction. No reaction was observed
if a catalytic amount of Co(tpp) was added to a stirred solu-
tion of aziridine 2, and if this last compound is added to a
catalytic run, it is recovered unaffected at the end of the
reaction (see Supporting Information).
Kinetic Studies
To clarify the mechanism of the amination reactions we
performed a kinetic study. The reaction between 4-ni-
trophenyl azide and cyclohexene catalyzed by Co(tpp) was
followed by IR spectroscopy by monitoring the intensity of
the 2120 cm–1 absorption of the azide. All reactions were
conducted at 75 °C and showed an excellent first-order de-
pendence of the rate on the azide concentration (R2 Ͼ 0.99
for all reactions, see Supporting Information). The kinetics
is also first order with respect to Co(tpp). We can rule out
any term independent from the catalyst in the reaction rate,
as the fitting for the plot of kobs. versus the concentration
of Co(tpp) can be forced to pass through the origin without
affecting the R2 value (R2 = 0.99; Figure S1, Supporting
Information). In other words, the uncatalyzed reaction
mentioned in the previous paragraph is so slow under the
conditions of the kinetic measurements not to be observable
at all. A first-order dependence was finally observed with
Scheme 5. Reaction between α-methylstyrene and 4-nitrophenyl az-
ide in the presence or absence of Co(tpp).
No reaction was observed with other styrenes such as β- respect to cyclohexene, but in this case, the straight line in
methylstyrene, cis-stilbene, trans-stilbene, and 1,1-diphenyl- the plot of kobs./[Co(tpp)] versus the olefin concentration
3012
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 3009–3019

Co-Catalyzed Allylic Amination and Aziridination of Olefins by Aryl Azides
showed an intercept on the axis of the reaction rate (Fig- cates that the presence of an electron-donating group on
ure 2). This indicates the presence of an olefin-independent the azide slows down the reaction. This data is consistent
reaction. Such a reaction between the aryl azide and the with electrophilic behavior of the coordinated azide in the
cobalt complex has indeed been independently observed, as transition state, as expected.
detailed in the following.
UV Studies
We next monitored the catalytic reactions by UV/Vis
spectroscopy in order to shed some light on the products
that are formed during the catalytic cycle. The UV/Vis spec-
tral changes for the reaction of 4-nitrophenyl azide with
cyclohexene catalyzed by Co(tpp) are shown in Figure 4.
Figure 2. Dependence of the rate of the allylic amination on the
concentration of cyclohexene.
Because the reaction products are different, a kinetic
study was also performed by employing 4-nitrophenyl azide
and α-methylstyrene as the substrate and Co(tpp) as the
catalyst. The kinetics is again first-order in azide (R2 Ͼ 0.98
Figure 4. UV/Vis spectral changes during the reaction of 4-ni-
for all reactions) and catalyst concentrations (see Support-
trophenyl azide (6.2ϫ10–1 mmol) with cyclohexene (30 mL) in the
ing Information). However, a first-order dependence of the
rate on the concentration of α-methylstyrene was observed
only up to an olefin concentration of 6.9  (90% v/v in
benzene). In neat α-methylstyrene, the reaction rate is lower
than it would be expected (Figure 3). We previously ob-
served an analogous outcome when studying the catalytic
activity of Ru(tpp)(CO) for this reaction.[15] In that case, an
inhibiting role of the competitively formed triazoline was
identified. The triazoline was shown to coordinate to ruthe-
nium, which blocks the free coordination site necessary for
the catalytic reaction to proceed. An analogous explanation
is likely to be involved here, and this possibility was investi-
gated by performing further spectroscopic studies.
presence of Co(tpp) (4.1ϫ10–2 mmol) at 84 °C. Scans were taken
at 30-min intervals. A = 4-NO2C6H4N3; B = 1.
The presence of the isosbestic point at 322 nm indicates
that the conversion of the starting aryl azide (A) into allylic
amine (B) 1 occurs without the accumulation of any long-
lived intermediate. We already showed that aziridines can-
not be intermediates in the synthesis of allylic amines. The
present result not only supports this view, but also indicates
that no other intermediate is observed along the reaction
pathway.
We next looked at the UV/Vis spectral changes in the
aziridination reaction. The spectral changes for the reaction
of 4-nitrophenyl azide (6.1ϫ10–1 mmol) with α-methylsty-
rene (90% v/v in benzene, 6.933 , 30 mL) catalyzed by
Co(tpp) (4ϫ10–4 ) at 75 °C are reported in Figure 5. Un-
Figure 3. Dependence of the aziridination rate on the concentration
of α-methylstyrene.
Figure 5. UV/Vis spectral changes during the reaction of 4-ni-
trophenyl azide (6.1ϫ10–1 mmol) with α-methylstyrene (90% v/v
in benzene, 30 mL) in the presence of Co(tpp) (1.2ϫ10–2 mmol) at
75 °C. Scans were taken at 1-h intervals. A = 4-NO2C6H4N3; D =
More information on the reactivity of the aryl azide com-
plexes could be gained by kinetic experiments run with a
different azide. When employing 4-methoxyphenyl azide, we
observed a sharp decrease in the reaction rate, which indi- 12 + 13.
Eur. J. Inorg. Chem. 2008, 3009–3019
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3013

A. Caselli, E. Gallo, S. Fantauzzi, S. Morlacchi, F. Ragaini, S. Cenini
FULL PAPER
der these conditions, we observed the disappearance of the tion between Co(tpp) and 4-nitrophenyl azide was con-
starting azide (A) (λ = 312 nm) with the simultaneous for- ducted at 60 °C in C6D6 new signals were observed in the
mation of aziridine 12 (λ = 341 nm) and triazoline 13 (λ = 1H NMR spectrum that cannot be attributed to any known
350 nm), which resulted in a broadened signal around compound, but again are identical to those of the product
345 nm (D). After 10 h, the conversion of the starting azide obtained by reaction of Co(tpp) with 4-(NO2)-
was quantitative and we could isolate aziridine 12 (43%) C6H4N=NC6H4-4-(NO2). All the above reported data, to-
and triazoline 13 (50%).
gether with the observation of only one set of resonances
During the catalytic reaction, the Soret band of Co(tpp) for the aryl groups of the diazene moiety of this complex,
(λ = 413 nm) remains unchanged. We can thus exclude the which indicates high symmetry, lead us to formulate the
formation of a long-lived CoIII intermediate, because in this identity of the final product as [Co(tpp)]2(ArN=NAr) [15;
case a considerable redshift of the Soret band would be ex- Ar = 4-(NO2)C6H4] (Scheme 6). A similar complex that has
pected (λ = 430 nm).[31] In contrast, the presence of a penta- an unsubstituted diazene (HN=NH) group bridging two ru-
coordinate Co(tpp) complex cannot be ruled out, as in this thenium–porphyrin complexes has been reported.[32] Unfor-
case no significant shift is observed. To be able to under- tunately, any attempt to grow single crystals suitable for X-
stand the changes at the metal center during the reaction, ray diffraction always gave the Co(tpp) complex without
we next looked at the stoichiometric reactions of Co(tpp) any coordinated ligand.
with both the olefin and the azide. No UV/Vis observable
shifts were observed with cyclohexene or α-methylstyrene.
When the stoichiometric reaction between Co(tpp) and 4-
nitrophenyl azide was monitored by UV/Vis we observed
the disappearance of the ArN3 signal at 312 nm and the
appearance of a new signal at 344 nm (Figure 6). The Soret
band of Co(tpp) at such a concentration goes out of range.
When the reaction with 4-nitrophenyl azide was repeated
employing substoichiometric amounts of Co(tpp) (mol ra-
tio = 10) the free diazene [4-(NO2)C6H4N=NC6H4-4-
(NO2)] could be detected (λ = 336 nm) as the major reac-
tion product (see Supporting Information).[24] Finally, ex-
amination of the reaction of Co(tpp) with 4-nitrophenyl di-
azene [4-(NO2)C6H4N=NC6H4-4-(NO2)] by UV/Vis spec-
troscopy indicates that the same final product is obtained
when either the azide or the diazene is employed as starting
Scheme 6. Proposed reaction product obtained from the reaction
of Co(tpp) with 4-nitrophenyl azide.
material.
The reaction between the azide and Co(tpp) was repeated
in C6D6 in the presence of the spin-trap compound tempo
(2,2,6,6-tetramethyl-1-piperidine-N-oxide), and the spectro-
scopic changes of the products were monitored by 1H
NMR spectroscopy. The signals due to paramagnetic Co-
II(tpp) gradually disappeared to give rise to a diamagnetic
cobalt(III) complex. Because there is no reaction between
the Co–porphyrin complex and tempo in the absence of the
azide, and because the addition of tempo does not have a
marked effect on the reaction of Co(tpp) with 4-(NO2)-
C6H4N=NC6H4-4-(NO2), this result suggests that the un-
paired spin density of the Co(tpp) complex is partly local-
ized on the azide nitrogen atoms in an intermediate com-
plex.
In order to investigate the role of complex 15 in the syn-
thesis of allylic amines, a series of catalytic reactions in
C6D6 at 60 °C with increasing amounts of cyclohexene were
Figure 6. UV/Vis spectra of 4-nitropenhyl azide (solid line, λmax
=
312 nm) and that of the product of the stoichiometric reaction be-
tween Co(tpp) (0.15 mmol) and 4-nitrophenyl azide (0.16 mmol) in
benzene (30 mL) at 75 °C after the complete consumption of the
starting azide (dashed line, λmax = 344 nm).
1
monitored by H NMR spectroscopy. First, we performed
the reaction between equimolar amounts of 4-nitrophenyl
azide and cyclohexene in the presence of stoichiometric
Co(tpp). Only complex 15 was formed and allylic amine 2
was not present in the final mixture. When the reaction was
NMR Spectroscopic Studies
If Co(tpp) is treated with an excess amount of olefin, repeated by employing a 1:10:170 ratio of Co(tpp)/azide/
such as cyclohexene or α-methylstyrene, in C6D6 no signifi- olefin, the azide signals disappeared in 9 h and allylic amine
1
cant shift in the H NMR signals is observed. In contrast, 2 was formed together with aniline (see Supporting Infor-
as already reported by us,[24] when the stoichiometric reac- mation). Complex 15, in contrast, was not formed during
3014
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 3009–3019

Co-Catalyzed Allylic Amination and Aziridination of Olefins by Aryl Azides
the catalytic cycle, because of the great excess of cyclohex- catalytic reaction. No reaction between 4-methoxyphenyl
ene.
azide and 13 was observed in the presence of Co(tpp). The
1
Similar results were obtained by following the H NMR coordination of 13 to the catalyst was observed even in the
spectral changes of a catalytic reaction of 4-nitrophenyl az- allylic amination of cyclohexene. The reaction rate mea-
ide with α-methylstyrene in the presence of Co(tpp). In this sured by monitoring the disappearance of the azide 1H
latter case, by employing CH2Cl2 as an internal standard in NMR signals of a catalytic reaction of 4-nitrophenyl azide
a sealed tube (see Experimental Section), a clean first-order with cyclohexene catalyzed by Co(tpp) [Co(tpp)/azide/ole-
dependence on the reaction rate with respect to the azide fin, 1:10:50; in C6D6 at 75 °C] in the presence of 13
was observed {Co(tpp)/azide/olefin, 1:10:50; in C6D6 at (5 equiv.; kobs./[Co] = 1.76ϫ10–2

–1 sec–1) is considerably
–1 sec–1;
75 °C, kobs./[Co] = 4.29ϫ10–2 –1 sec–1}. At the end of the slower than in its absence (kobs./[Co] = 4.83ϫ10–2


reaction, we measured a 49% yield of aziridine 12 together see Supporting Information).
with triazoline 13 (12%) and 4-nitroaniline (21%). The rest
Reaction Mechanism
of the starting azide was converted into an unidentified
product.
It is widely accepted that aryl azides can generate metal–
We next examined the effect of triazoline 13 on Co(tpp). nitrene (imido) intermediates in the presence of a suitable
A remarkable shift was observed in C6D6 at room tempera- metal ion. We already showed that cobalt(II) porphyrin
ture on the 1H NMR resonances of both the Co–porphyrin complexes are able to activate aromatic azides for the amin-
moiety and the organic molecule (Figure 7). Unfortunately ation of C–H bonds of saturated organic compounds under
any attempt to grow crystals suitable for X-ray structural mild conditions to give secondary amines and imines, with-
characterization of the coordination complex between out the intermediate formation of a metal–nitrene.[24] All
Co(tpp) and the triazoline met with failure, and again only the kinetic data reported above support, even in the present
Co(tpp) was recovered. This is probably due to a very weak case, the occurrence of a reversible interaction between
coordination. A similar Ru(tpp)(L)CO complex (L = 13) Co(tpp) and the arylazide, as proposed in Scheme 7.
was already reported by us.[15]
The equilibrium associated to k1 and k–1 in Scheme 7 is
shifted to the side of the reactants, in accord with the fact
that the intermediate complex between Co(tpp) and the aryl
azide is not observable. This complex can either react with
cyclohexene to yield the allylic amine or loose dinitrogen in
an unimolecular reaction to afford an intermediate, formu-
lated as a true imido complex (I, Scheme 7), as already sug-
gested for benzylic C–H activation.[24] We propose the
imido complex to be responsible not only for the formation
of diazene, but also for that of aniline for several reasons:
(i) Aniline formation occurs by hydrogen abstraction, very
likely from cyclohexene. Such hydrogen-abstraction reac-
tions are typically performed by radical species and odd-
electron species I is the most likely responsible for it. (ii)
Although diazene is formed in the absence of cyclohexene,
it was only obtained in trace amounts in the presence of the
latter. This suggests that both aniline and diazene are de-
rived from the same intermediate, which reacts with cyclo-
hexene at a much faster rate than that with an azide mole-
1
Figure 7. H NMR spectra of Co(tpp) (upper line) and Co(tpp) +
13 (lower line) in C6D6. Arrows evidence the shifts of the signal of
the protons of the porphyrin skeleton (Hβ = tpp pyrrolic protons;
Ho = tpp ortho aromatic protons; Hm+p = tpp meta and para aro-
matic protons).
Even if the coordination of the triazoline moiety to cule. (iii) If aniline is derived from reaction of I, then its
Co(tpp) appears to be weak, its effect on the catalytic azir- formation should be inhibited by an increase in the concen-
idination and allylic amination reactions is not negligible. tration of cyclohexene. In contrast, the selectivity should
In fact, when the catalytic reaction of 4-nitrophenyl azide not be altered if both the allylic amine and aniline originate
with α-methylstyrene in the presence of Co(tpp) was re- from the azido complex. Although a detailed study of the
peated in the presence of 13 [Co(tpp)/13/azide/olefin, effect on the selectivity of the cyclohexene concentration
1:5.8:10:50; in C6D6 at 75 °C] the reaction rate was slower was not performed, the allylic amine selectivity in neat cy-
than in its absence (kobs./[Co] = 3.36ϫ10–2

–1 sec–1 vs. clohexene is indeed higher than that in a 50:50 mixture of
kobs./[Co] = 4.29ϫ10–2 –1 sec–1). When the same reaction cyclohexene/benzene.[17] It should be considered that even

was repeated in the presence of a different triazoline, in the absence of cyclohexene, the decomposition of the az-
namely, 1-(4-methoxyphenyl)-5-methyl-5-phenyl-1,2,3-tri- ide in the presence of Co(tpp) leads to the formation of
azoline (16), no effect on the reaction rate was observed aniline as a byproduct, and we cannot exclude that its for-
(kobs./[Co] = 4.34ϫ10–2 –1 sec–1; see Supporting Infor- mation during the catalytic reaction may occur by other

mation). It should be noted that this last triazoline contain- routes.
ing an electron-donating substituent on the aryl ring is less
Note that any mechanism in which the aminating species
stable than 13, and it is probably transformed during the is the imido complex can be completely excluded by the
Eur. J. Inorg. Chem. 2008, 3009–3019
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3015

A. Caselli, E. Gallo, S. Fantauzzi, S. Morlacchi, F. Ragaini, S. Cenini
FULL PAPER
Scheme 7. Proposed reaction mechanism for the allylic amination and aziridination reactions.
available data. Indeed, the formation of such a complex deed what is observed indicates that this condition is met
cannot be an equilibrium, and so only two possibilities exist in the present system. Note that this was not the case for
for the kinetics of the reaction: either the rate of formation the related benzylic amination reactions. Equation (1) can
of the imido complex is slow relative to that of its subse- thus be simplified to give Equation (2):
quent reaction with the olefin, or it is fast. In the former
V = –d[ArN3]/dt = [Co][ArN3]k1 (k2[C6H10] + k3)/(k3 + k–1)
(2)
case, the kinetics should be zero order in the olefin concen-
tration. In the second, the kinetics should be zero order in
the azide, and the imido complex should be spectroscopi-
cally observable. The available data are clearly not compati-
ble with either alternative.
Application of the steady-state approximation to the in-
termediate arylazide–CoIIporphyrin complex leads, in the
case of allylic amination of cyclohexene, to Equation (1):
(iii) Equation (2) may be further simplified in case either
k2[C6H10]ϾϾk3 or k3 ϾϾk2[C6H10]. However, the experi-
mental results clearly show that neither case occurs. This is
clearly indicated by the positive slope and positive (and not
negligible) intercept in the rate versus [C6H10] plot (Fig-
ure 2).
A similar reaction mechanism can be proposed even for
the aziridination reactions. In this case, however, we should
consider that the kinetics of the triazoline formation (see
Supporting Information) indicate that at high olefin con-
V = –d[ArN3]/dt = [Co][ArN3]k1(k2[C6H10] + k3)/(k2[C6H10] + k3 +
k–1)
(1)
From this equation we can note the following: (i) The centration 13 is formed at a rate that is competitive with
term [Co] refers to the actual concentration of free Co(tpp). that of the catalytic reaction and that this last compound
However, because both the azido complex and the imido can reversibly coordinate to Co(tpp) (Scheme 7). The equi-
complex are present in very-low concentrations, in practice, librium constant, Keq. ≈ 3.6ϫ10–2 molL–1, for the complex
this term is coincident with the total cobalt concentration. formation was calculated by 1H NMR titration, which dem-
Note that this would not always be true in the case of the onstrates that even if the equilibrium is shifted toward the
aziridination reaction, where the concentration of the tria- reactants, the complex is formed even for low concentra-
zoline adduct is not negligible in some cases. (ii) If tions of 13 (see Supporting Information). Any attempt to
k
–1 ϾϾk2[C6H10], then the kinetics would become first or- isolate the complex in the solid state, however, failed due to
der in cyclohexene concentration. The fact that this is in- the low solubility of Co(tpp) relative to that of the newly
3016
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 3009–3019

Co-Catalyzed Allylic Amination and Aziridination of Olefins by Aryl Azides
(32 mg, 0.048 mmol) in a 1:1 volume mixture of the desired olefin
and benzene (30 mL). The resulting red solution was heated at
formed complex. The formation of this complex reduces the
active catalyst amount and consequently the reaction rate,
which thus explains the effect observed when the solution
contains more than 90% (v/v) of α-methylstyrene (see Fig-
ure 3).
In a manner similar to the intermediate coordination of
the azide proposed in Scheme 7, a mechanism that involves
N-bound tosylimidoioninane-adduct formation as the
active intermediate in the MnIII corrole catalyzed aziridin-
ation reaction, in route to the MnV imido product, was re-
cently unambiguously demonstrated in elegant work by
Abu-Omar and coworkers.[33] Structurally related iodo-
imine adducts have also been proposed on a kinetic ground
as the active intermediates in both the Mn/salen-catalyzed
sulfimidation of sulfides[34] and in the Mn/porphyrin-cata-
lyzed aziridination of olefins.[35]
75 °C until total consumption of the aryl azide (IR absorbance: ν
˜
= 2121 cm–1, Ͻ0.03) was observed. The final mixture was evapo-
rated to dryness in vacuo, and the products were separated by flash
chromatography (n-hexane/ethyl acetate, see Table 1 and Support-
ing Information). The reaction of 4-nitrophenyl azide (0.62 mmol)
in the presence of Co(tpp) (0.041 mmol) in cyclohexene (30 mL) at
84 °C was followed by UV/Vis spectroscopy. Spectra were taken at
30-min intervals. An isosbestic point at λ = 322 nm was observed.
Catalytic Aziridination of Olefins by Co(tpp): 4-Nitrophenyl azide
(101 mg, 0.62 mmol) was added to a solution of Co(tpp) (8.0 mg,
0.012 mmol) in a 1:1 mixture of the desired olefin and benzene
(30 mL). The resulting red solution was heated at 75 °C until total
consumption of the aryl azide (IR absorbance: ν = 2121 cm–1,
˜
Ͻ0.03) was observed. The final mixture was evaporated to dryness
in vacuo, and the products were isolated by flash chromatography
(n-hexane/CH2Cl2, see Table 1 and Supporting Information). The
reaction of 4-nitrophenyl azide (0.61 mmol) in the presence of
Co(tpp) (0.012 mmol) in α-methylstyrene (90% v/v in benzene,
30 mL) at 75 °C was followed by UV/Vis spectroscopy. Spectra
were taken at 1-h intervals.
Conclusions
In this paper we reported a new catalytic system for the
aziridination and allylic amination of nonactivated olefins.
The mechanistic aspects of the catalytic cycle of both reac-
tions were deeply investigated and the unusual aspects were
explained. Yields of the catalytic reactions are moderate to
good, and kinetic and mechanistic data allowed us to pro-
pose a reasonable mechanism. In particular, we showed that
in the present system, the often-postulated cobalt–imido
complexes are not the intermediates in the catalytic amin-
ation reactions. The selectivity of the present Co(tpp) sys-
tem is particularly high. Depending on the substrate em-
ployed, allylic amines or aziridines can be obtained.
1H NMR and UV/Vis Spectroscopic Analysis of [Co(tpp)]2(p-
NO2C6H4N=NC6H4NO2-p) (15)
1H NMR: Co(tpp) (10.3 mg, 0.015 mmol) and 4-nitrophenyl azide
(2.7 mg, 0.015 mmol) were dissolved in C6D6 (1 mL) at 60 °C. The
1
disappearance of the aryl azide was monitored by recording a H
NMR spectrum every 5 min. The absorptions due to Co(tpp) shift
during the reaction. The final spectrum shows signals at: 1H NMR
(300 MHz, C6D6, 60 °C): δ = 15.64 (br. s, 8 H), 12.92 (br. s, 8 H),
1
9.56 (m, 12 H), 7.82 (m, 4 H), 7.41 (m, 4 H) ppm. The H NMR
spectrum is identical to that of the product obtained by reaction of
Co(tpp) (17.3 mg, 0.026 mmol) with 4-nitrophenyldiazene (3.7 mg,
0.013 mmol) in C6D6 (1 mL) at 60 °C.
UV/Vis: The reaction mixture of Co(tpp) (100 mg, 0.15 mmol) and
4-nitrophenyl azide (25.7 mg, 0.16 mmol) in benzene (30 mL) was
heated at 80 °C until complete consumption of the aryl azide (UV/
Vis signal at 312 nm) was observed. The UV/Vis signal at 344 nm
attributed to complex 15 was observed.
Experimental Section
General Procedures: Unless otherwise stated, all operations were
carried out under an atmosphere of purified dinitrogen by using
modified Schlenk techniques. Solvents were dried and distilled be-
fore use by standard methods. Benzene, cyclohexene, cyclopentene,
α-methylstyrene, and 1-octene were distilled from sodium and
stored under dinitrogen. The synthesis of 4-nitrophenyl azide,[36]
4-nitrophenyldiazene,[37] 1-(4-nitrophenyl)-5-methyl-5-phenyl-1,2,3-
triazoline (13),[15] tetraphenylporphyrin (tppH2),[38] and Co(tpp)[39]
were carried out as reported in the literature. 1H and 13C{1H}
NMR spectra were obtained with Bruker Avance 300-DRX or
Avance 400-DRX spectrometers. Chemical shifts (ppm) are re-
ported relative to TMS. Infrared spectra were recorded with a BIO-
RAD FTS spectrophotometer. UV/Vis spectra were recorded with
an Agilent 8453E instrument. Elemental analyses and mass spectra
were recorded in the analytical laboratories of Milan University.
All other starting materials were commercial products and used as
received, unless otherwise reported.
Reaction of Co(tpp) with 4-Nitrophenyl Azide in the Presence of
tempo Followed by 1H NMR Spectroscopy: tempo (18.5 mg,
0.116 mmol) dissolved in C6D6 (0.1 mL) was added to a solution
of Co(tpp) (5.1 mg, 0.0076 mmol) in C6D6 (0.5 mL). The solution
was heated at 60 °C inside the NMR probe for 10 min [no changes
in the chemical shift of the signals of the paramagnetic CoII(tpp)
were observed], then 4-nitrophenyl azide (13.3 mg, 0.081 mmol)
dissolved in C6D6 (0.1 mL) was added, and the resulting red solu-
tion was heated at 60 °C inside the probe. A 1H NMR spectrum
was recorded every 5 min until complete consumption of the aryl
azide. The signals due to paramagnetic CoII(tpp) gradually disap-
peared to give rise to a new diamagnetic cobalt(III) complex. 1H
NMR (300 MHz, C6D6): δ = 9.17 (s, 8 H), 8.15 (m, 8 H), 7.63 (m,
12 H), 1.18 (s, ca. 12 H, CH3 tempo) ppm. Others signals due
to the tempo moiety are too broadened to be attributed without
uncertainty.
Caution: Like hydrogen azide, many other azides are also explosive
substances that decompose with the release of nitrogen through the
slightest input of external energy, for example, pressure, impact, or
heat, but this is not the case for the aryl azides employed in the
present study.[40]
Reaction of Co(tpp) with 13 Followed by 1H NMR Spectroscopy:
Compound 13 (2.0 mg, 0.0075 mmol) was added to a solution of
Co(tpp) (5.2 mg, 0.0076 mmol) in C6D6 (0.75 mL). A 1H NMR
spectrum was recorded after 5 min at room temperature. The 1H
Catalytic Allylic Amination of Olefins by Co(tpp): 4-Nitrophenyl NMR spectrum is reported in Figure 7. Among all the signals, the
1
azide (100 mg, 0.61 mmol) was added to a solution of Co(tpp)
proton resonances at: H NMR (300 MHz, C6D6): δ = 15.1 (br. s,
Eur. J. Inorg. Chem. 2008, 3009–3019
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3017

A. Caselli, E. Gallo, S. Fantauzzi, S. Morlacchi, F. Ragaini, S. Cenini
FULL PAPER
8 H), 11.7 (br. s, 8 H), 9.1 (m, 12 H) ppm can be attributed to the
Co(tpp) moiety.
by recording an IR spectrum every 15 min. The same reaction was
repeated six times with different ratios of the α-methylstyrene/ben-
zene solvent mixture (30:0, 28.5:1.5, 27:3, 22.5:7.5, 15:15,
7.5:22.5 mL).
Typical Procedures for the Determination of the Reaction Kinetics
with Respect to the Azide
Supporting Information (see footnote on the first page of this arti-
cle): Representative experimental procedures and characterization
data and kinetic graphics.
1H NMR: Co(tpp) (2.3 mg, 0.0034 mmol) was dissolved in C6D6
(0.8 mL) in an NMR tube. The substrates (see Supporting Infor-
mation for details) and CH2Cl2 (6 µL) were added, the tube was
sealed under N2, and then the sample was heated to 75 °C inside
1
the probe. A H NMR spectrum was recorded every 5 min.
Acknowledgments
IR: The rate constants of the reactions were measured by monitor-
ing the change in absorbance of the band at 2121 cm–1 of 4-nitro-
phenylazide. The first-order rate constants kobs were obtained by
linear fits of ln(A/Ao) versus time according to the equation ln(A/
Ao) = –kobs ϫt, where Ao and A are the initial absorbance and the
absorbance at the time t. For all reactions, the R2 value is larger
than 0.990.
We thank MiUR (Programmi di Ricerca Scientifica di Rilevante
Interesse Nazionale PRIN 2005035123) for financial support.
[1] H.-U. Blaser, A. Indolese, F. Naud, U. Nettekoven, A.
Schnyder, Adv. Synth. Catal. 2004, 346, 1583–1598.
[2] E. N. Jacobsen, A. Pfaltz, H. Yamamoto (Eds.), Comprehensive
Asymmetric Catalysis, Supplement: Volume 1. 2004, p. 238.
[3] R. S. Atkinson, Tetrahedron 1999, 55, 1519–1559.
[4] P. Dauban, R. H. Dodd, Synlett 2003, 1571–1586.
[5] J. A. Halfen, Curr. Org. Chem. 2005, 9, 657–669.
[6] J.-L. Liang, J.-S. Huang, X.-Q. Yu, N. Zhu, C.-M. Che, Chem.
Eur. J. 2002, 8, 1563–1572.
Allylic Amination Reaction of Cyclohexene
i) Determination of the Kinetics Order with Respect to 4-Nitrophenyl
Azide: Co(tpp) (13.5 mg, 0.020 mmol) was dissolved in cyclohexene
(15 mL) with 4-nitrophenyl azide (53.4 mg, 0.33 mmol), and the re-
sulting red solution was immersed in an oil bath preheated at 75 °C.
The consumption of the aryl azide was monitored by recording an
IR spectrum every 15 min.
[7] P. Dauban, L. Saniere, A. Tarrade, R. H. Dodd, J. Am. Chem.
Soc. 2001, 123, 7707–7708.
[8] D. Kano, S. Minakata, M. Komatsu, J. Chem. Soc. Perkin
Trans. 1 2001, 3186–3188.
ii) Determination of the Kinetics Order with Respect to Co(tpp):
Co(tpp) was dissolved in cyclohexene (30 mL) with 4-nitrophenyl
azide (100 mg, 0.61 mmol), and the resulting red solution was im-
mersed in an oil bath preheated at 75 °C following the consumption
of the aryl azide. An IR spectrum was recorded every 15 min. The
same reaction was repeated four times with four different amounts
of Co(tpp) (7.1 mg, 0.0102 mmol; 13.6 mg, 0.0201 mmol; 27.8 mg,
0.0414 mmol; 34.1 mg, 0.0508 mmol).
[9] B. M. Chanda, R. Vyas, A. V. Bedekar, J. Org. Chem. 2001, 66,
30–34.
[10] S. Cenini, A. Penoni, S. Tollari, J. Mol. Catal. A 1997, 124,
109–113.
[11] T. Katsuki, Chem. Lett. 2005, 34, 1304–1309.
[12] S. Cenini, E. Gallo, A. Caselli, F. Ragaini, S. Fantauzzi, C.
Piangiolino, Coord. Chem. Rev. 2006, 250, 1234–1253.
[13] A. Caselli, E. Gallo, F. Ragaini, A. Oppezzo, S. Cenini, J. Orga-
nometallic. Chem. 2005, 690, 2142–2148.
[14] C. Piangiolino, E. Gallo, A. Caselli, S. Fantauzzi, F. Ragaini,
S. Cenini, Eur. J. Org. Chem. 2007, 743–750.
[15] S. Fantauzzi, E. Gallo, A. Caselli, F. Ragaini, P. Macchi, N.
Casati, S. Cenini, Organometallics 2005, 24, 4710–4713.
[16] S. Fantauzzi, E. Gallo, A. Caselli, C. Piangiolino, F. Ragaini,
S. Cenini, Eur. J. Org. Chem. 2007, 6053–6059.
[17] S. Cenini, S. Tollari, A. Penoni, C. Cereda, J. Mol. Catal. A
1999, 137, 135–146.
[18] A. Caselli, E. Gallo, F. Ragaini, F. Ricatto, G. Abbiati, S. Cen-
ini, Inorg. Chim. Acta 2006, 359, 2924–2932.
[19] K. M. Kadish, K. M. Smith, R. Gouilard (Eds.), The Porphyrin
Handbook, Academic Press, San Diego, 2000–2003, vols. 1–20.
[20] J. T. Groves, T. Takahashi, J. Am. Chem. Soc. 1983, 105, 2073–
2074.
iii) Determination of the Kinetics Order with Respect to Cyclohex-
ene: Co(tpp) (27 mg, 0.040 mmol) was dissolved in the solvent mix-
ture (30 mL) with 4-nitrophenyl azide (100 mg, 0.61 mmol), and
the resulting red solution was immersed in an oil bath preheated
at 75 °C. The consumption of the aryl azide was monitored by
recording an IR spectrum every 15 min. The same reaction was
repeated six times with different ratios of the cyclohexene/benzene
solvent mixture (30:0, 22.5:7.5, 15:15, 8.6:21.4, 4.5:25.5, 0:30 mL).
Aziridination Reaction of α-Methylstyrene
i) Determination of the Kinetics Order with Respect to 4-Nitrophenyl
Azide: Co(tpp) (8.6 mg, 0.013 mmol) was dissolved in α-methylsty-
rene/benzene (1:3, 30 mL) with 4-nitrophenyl azide (101.4 mg,
0.618 mmol), and the resulting red solution was immersed in an oil
bath preheated at 75 °C. The consumption of the aryl azide was
monitored by recording an IR spectrum every 15 min.
[21] G. Y. Gao, J. D. Harden, X. P. Zhang, Org. Lett. 2005, 7, 3191–
3193.
[22] G. Y. Gao, J. E. Jones, R. Vyas, J. D. Harden, X. P. Zhang, J.
Org. Chem. 2006, 71, 6655–6658.
[23] J. V. Ruppel, R. M. Kamble, X. P. Zhang, Org. Lett. 2007, 9,
4889–4892.
[24] F. Ragaini, A. Penoni, E. Gallo, S. Tollari, C. Li Gotti, M.
Lapadula, E. Mangioni, S. Cenini, Chem. Eur. J. 2003, 9, 249–
259.
[25] S. Cenini, E. Gallo, A. Penoni, F. Ragaini, S. Tollari, Chem.
Commun. 2000, 2265–2266.
[26] H. M. L. Davies, M. S. Long, Angew. Chem. Int. Ed. 2005, 44,
3518–3520.
ii) Determination of the Kinetics Order with Respect to Co(tpp):
Co(tpp) was dissolved in α-methylstyrene/benzene (1:3, 30 mL)
with 4-nitrophenyl azide (100 mg, 0.61 mmol), and the resulting red
solution was immersed in an oil bath preheated at 75 °C following
the consumption of the aryl azide. An IR spectrum was recorded
every 15 min. The same reaction was repeated four times with four
different amount of Co(tpp) (8.6 mg, 0.013 mmol; 16.8 mg,
0.025 mmol; 24.5 mg, 0.037 mmol; 32.8 mg, 0.049 mmol).
[27] P. O’Brien, C. M. Rosser, D. Caine, Tetrahedron 2003, 59, 9779–
iii) Determination of the Kinetics Order with Respect to α-Methyl-
styrene: Co(tpp) (8.6 mg, 0.013 mmol) was dissolved in the solvent
mixture (30 mL) with 4-nitrophenyl azide (100 mg, 0.61 mmol),
and the resulting red solution was immersed in an oil bath pre-
heated at 75 °C. The consumption of the aryl azide was monitored
9791.
[28] M. Moreno-Manas, L. Morral, R. Pleixats, J. Org. Chem. 1998,
63, 6160–6166.
[29] F. Ragaini, S. Cenini, F. Turra, A. Caselli, Tetrahedron 2004,
60, 4989–4994.
3018
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 3009–3019

Co-Catalyzed Allylic Amination and Aziridination of Olefins by Aryl Azides
[30] L. Wolff, Justus Liebigs Ann. Chem. 1912, 394, 23–108.
[31] Y. Le Mest, M. L’Her, J. Y. Saillard, Inorg. Chim. Acta 1996,
248, 181–191.
[32] J. P. Collman, J. E. Hutchison, M. S. Ennis, M. A. Lopez, R.
Guilard, J. Am. Chem. Soc. 1992, 114, 8074–8080.
[33] M. J. Zdilla, M. M. Abu-Omar, J. Am. Chem. Soc. 2006, 128,
16971–16979.
[37] J. Dale, L. Vikersveen, Acta Chem. Scand. Ser. B 1988, 42, 354–
361.
[38] R. A. W. Johnstone, M. L. P. G. Nunes, M. M. Pereira, A. M.
d’A Rocha Gonsalves, A. C. Serra, Heterocycles 1996, 43,
1423–1437.
[39] A. D. Adler, F. R. Longo, V. Varadi, Inorg. Synth. 1976, 16,
213–220.
[34] C. Ohta, T. Katsuki, Tetrahedron Lett. 2001, 42, 3885–3888.
[35] T.-S. Lai, H.-L. Kwong, C.-M. Che, S.-M. Peng, Chem. Com-
mun. 1997, 2373–2374.
[36] M. Tanno, S. Sueyoshi, S. Kamiya, Chem. Pharm. Bull. 1982,
30, 3125–3132.
[40] P. A. S. Smith, Open-Chain Nitrogen Compounds, Benjamin,
New York, 1966, vol. 2, pp. 211–256.
Received: February 11, 2008
Published Online: May 27, 2008
Eur. J. Inorg. Chem. 2008, 3009–3019
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3019