Photochemical Reactions of Mesityl Azide with Tetracyanoethylene: Competitive Trapping of Singlet Nitrene and Didehydroazepine

Article abstract of DOI:10.1021/jo9622901

Irradiation of the title azide 4 in the presence of TCNE gives a mixture of two stable adducts. One of them is identified as the azomethine ylide 5, the structure of which is strictly determined by X-ray crystallography. The other is spectroscopically assigned to the spiroazepine 6. The effect of wavelength of the light employed in the photolysis reveals that the TCNE-4 charge-transfer complex (λ_max 454 and 550 nm in dichloromethane) does not participate in the adduct formation. The ratio of the adducts obtained in the photolysis is dependent linearly upon the initial concentration of TCNE, which strongly suggests that the adducts 5 and 6 are produced by competitive trapping of singlet mesitylnitrene (8S) and trimethyldidehydroazepine (9), respectively. The rate constant for the reaction of 8S with TCNE is estimated to be on the order of 10⁹ M^-1 s^-1 or greater. The PM3 calculation indicates that the azomethine ylide 5 is thermodynamically more stable than the aziridine 7, which is thought to be initially formed by the reaction of 8S with TCNE. Thus, we propose that these findings make the first example of competitive trapping of singlet arylnitrene and its ring-expanded isomer with an alkene, which definitely reveals the intervention of singlet nitrene in the photolysis of an aryl azide.

Full text of DOI:10.1021/jo9622901

J . Org. Chem. 1997, 62, 3055-3061
3055
P h otoch em ica l Rea ction s of Mesityl Azid e w ith
Tetr a cya n oeth ylen e: Com p etitive Tr a p p in g of Sin glet Nitr en e a n d
Did eh yd r oa zep in e
Shigeru Murata,* Shizue Abe,† and Hideo Tomioka*,†
Department of Basic Sciences, Graduate School of Arts and Sciences, The University of Tokyo, Meguro,
Tokyo 153, J apan, and Chemistry Department for Materials, Faculty of Engineering, Mie University, Tsu,
Mie 514, J apan
Received December 9, 1996X
Irradiation of the title azide 4 in the presence of TCNE gives a mixture of two stable adducts. One
of them is identified as the azomethine ylide 5, the structure of which is strictly determined by
X-ray crystallography. The other is spectroscopically assigned to the spiroazepine 6. The effect of
wavelength of the light employed in the photolysis reveals that the TCNE-4 charge-transfer complex
(λmax 454 and 550 nm in dichloromethane) does not participate in the adduct formation. The ratio
of the adducts obtained in the photolysis is dependent linearly upon the initial concentration of
TCNE, which strongly suggests that the adducts 5 and 6 are produced by competitive trapping of
singlet mesitylnitrene (8S) and trimethyldidehydroazepine (9), respectively. The rate constant for
the reaction of 8S with TCNE is estimated to be on the order of 109 M-1 s-1 or greater. The PM3
calculation indicates that the azomethine ylide 5 is thermodynamically more stable than the
aziridine 7, which is thought to be initially formed by the reaction of 8S with TCNE. Thus, we
propose that these findings make the first example of competitive trapping of singlet arylnitrene
and its ring-expanded isomer with an alkene, which definitely reveals the intervention of singlet
nitrene in the photolysis of an aryl azide.
Sch em e 1
Studies of the photodecomposition processes of aryl
azides have advanced in recent years.1 Platz and his co-
workers established the scheme of the photochemical
decomposition of phenyl azide (1) by the use of spectro-
scopic techniques and quantitative product analyses,
where singlet phenylnitrene (2S) was identified as the
temperature-dependent branching point (Scheme 1).2 At
room temperature the ring expansion from 2S to dide-
hydroazepine (3) is favored over intersystem crossing to
triplet phenylnitrene (2T), while at low temperature 2T
is responsible for the product formation since the rate of
ring expansion decreases considerably.
Thus, singlet phenylnitrene (2S) is postulated to be a
primary reactive intermediate formed upon the photolysis
of phenyl azide (1). However, although triplet phenylni-
trene (2T), as well as didehydroazepine (3), has been
directly observed by means of IR and UV spectroscopy
in matrices3 and in solutions,4 2S has not yet been
detected by spectroscopy. Odum and Wolf reported that
the singlet nitrene and the didehydroazepine formed
photolytically from p-cyanophenyl azide were competi-
tively trapped with dimethylamine to give the hydrazine
and the 3H-azepine derivative, respectively.5 It was also
reported that highly electrophilic singlet nitrenes formed
from aryl azides substituted with electron-withdrawing
groups were captured by alkanes, alkenes, aromatics, and
amines.6 However, attempts to trap singlet phenylni-
trene (2S), as well as arylnitrenes having no electron-
withdrawing groups, with these external reagents have
been unsuccessful. Very recently, it was reported that
singlet nitrenes from the irradiation of phenyl, 4-biphen-
ylyl, and 2-fluorenyl azide could be trapped by protona-
tion in aqueous solutions to give nitrenium ions.7
† Mie University.
X Abstract published in Advance ACS Abstracts, April 15, 1997.
(1) The photochemistry of aryl azides has reviewed; see, for ex-
ample: (a) Iddon, B.; Meth-Cohn, O.; Scriven, E. F. V.; Suschitzky,
H.; Gallagher, P. T. Angew. Chem., Int. Ed. Engl. 1979, 18, 900. (b)
Scriven, E. F. V. In Reactive Intermediates; Abramovitch, R. A., Ed.;
Plenum Press: New York, 1982. (c) Azides and Nitrenes; Scriven, E.
F. V., Ed.; Academic Press: New York, 1984. (d) Wentrup, C. Reactive
Molecules; Wiley: New York, 1984; Chapter 4.
In the course of our studies on trapping of singlet
arylnitrenes with electron-deficient alkenes, we found
(5) Odum, R. A.; Wolf, G. J . Chem. Soc., Chem. Commun. 1973, 360.
(6) (a) Banks, R. E.; Sparkes, G. R. J . Chem. Soc., Perkin Trans.1
1972, 2964. (b) Banks, R. E.; Prakash, A. J . Chem. Soc., Perkin Trans.
1 1974, 1365. (c) Abramovitch, R. A.; Challand, S. R.; Yamada, Y. J .
Org. Chem. 1975, 40, 1541. (d) Leyva, E.; Young, M. J . T.; Platz, M. S.
J . Am. Chem. Soc. 1986, 108, 8307. (e) Young, M. J . T.; Platz, M. S.
Tetrahedron Lett. 1989, 30, 2199. (f) Keana, J . F. W.; Cai, S. X. J . Org.
Chem. 1990, 55, 3640. (g) Poe, R.; Schnapp, K.; Young, M. J . T.;
Grayzar, J .; Platz, M. S. J . Am. Chem. Soc. 1992, 114, 5054.
(7) (a) McClelland, R. A.; Davidse, P. A.; Hadzialilc, G. J . Am. Chem.
Soc. 1995, 117, 4173. (b) McClelland, R. A.; Kahley, M. J .; Davidse, P.
A.; Hadzialic, G. J . Am. Chem. Soc. 1996, 118, 4794.
(2) Leyva, E.; Platz, M. S.; Persy, G.; Wirz, J . J . Am. Chem. Soc.
1986, 108, 3783.
(3) (a) Chapman, O. L.; Le Roux, J .-P. J . Am. Chem. Soc. 1978, 100,
282. (b) Donnelly, T.; Dunkin, I. R.; Norwood, D. S. D.; Prentice, A.;
Shields, C. J .; Thompson, P. C. P. J . Chem. Soc., Perkin Trans. 2 1985,
307. (c) Hayes, J . C.; Sheridan, R. S. J . Am. Chem. Soc. 1990, 112,
5879.
(4) (a) Shields, C. J .; Chrisope, D. R.; Schuster, G. B.; Dixon, A. J .;
Poliakoff, M.; Turner, J . J . J . Am. Chem. Soc. 1987, 109, 4723. (b) Li,
Y.-Z.; Kirby, J . P.; George, M. W.; Poliakoff, M.; Schuster, G. B. J . Am.
Chem. Soc. 1988, 110, 8092.
S0022-3263(96)02290-6 CCC: $14.00 © 1997 American Chemical Society

3056 J . Org. Chem., Vol. 62, No. 10, 1997
Murata et al.
F igu r e 1. UV-vis spectrum of 4 in the absence (- - -) and in the presence (s) of TCNE in acetonitrile: [4] ) 5.1 × 10-2 M,
[TCNE] ) 1.0 × 10-2 M.
that the singlet nitrene, as well as the didehydroazepine,
derived photolytically from mesityl azide (4) was trapped
with tetracyanoethylene (TCNE) to give a stable adduct
in an appreciable yield. In this paper, we wish to report
the first example of competitive trapping of singlet
arylnitrene and didehydroazepine with an alkene, which
reveals the intervention of singlet nitrene in the pho-
tolysis of an aryl azide having no electron-withdrawing
groups.
Resu lts
Absor p tion Sp ectr a of Mesityl Azid e (4) in th e
P r esen ce of TCNE. Mesityl azide (4) is a light yellow
liquid, the absorption spectrum of which shows an
intense transition with maximum at 204 nm and a
F igu r e 2. Concentration dependence of the absorbance for
maximum absorption of the TCNE-4 charge-transfer complex
in dichloromethane.
shoulder at around 240 nm (log ꢀ ) 4.45 and 3.82 in
acetonitrile, respectively). The spectrum also exhibits a
weak tailing to ca. 400 nm, so that the azide 4 can be
photolyzed efficiently with the light of a high-pressure
mercury lamp filtered by Pyrex or even by the optical
glass filter cutting wavelength shorter than 350 nm (log
ꢀ ) 2.05 and 1.06 at 313 and 366 nm in acetonitrile,
respectively).
by the method of Merrifield and Phillips.8 If the complex
is assumed to be 1:1, the absorbance A for maximum
absorption of the charge-transfer complex is correlated
with the association constant K and the molar absorp-
tivity ꢀ by eq 1,
T0l
A
T0
When TCNE was added to a solution of mesityl azide
(4) in dichloromethane, the solution took on a dark red
color. The UV-vis spectrum exhibited a transition with
maximum at 454 nm and a shoulder at around 550 nm
(Figure 1), both of which were not present in the
spectrum of either 4 or TCNE. In acetonitrile, the new
visible band appeared as shoulders around 415 and 510
nm. The colored species could be identified as a charge-
transfer complex between mesityl azide (4) and TCNE
on the basis of its similarity to well-known complexes
between TCNE and a variety of other aromatics.8 The
two charge-transfer absorption bands observed in the
complex between 4 and TCNE originate from the splitting
of the degeneracy of highest occupied orbitals in benzene
by ring substituents. The similar spectral splitting has
been observed in the TCNE-p-xylene or -durene charge-
transfer complex.8
1
B0
1
ꢀK
1
ꢀ
)
+
+
(1)
[
]
ꢀ
where T0 and B0 are the initial concentrations of TCNE
and mesityl azide (4), respectively, and l is the cell length
(1 cm). Figure 2 displays the plot of T0l/A against 1/B0,
which gives a straight line (correlation coefficient 0.999).
By the least-squares analysis of the plot, we obtain K and
ꢀ as 0.917 and 1090, respectively. The association
constant K for the formation of the TCNE-4 charge-
transfer complex is ca. 20 times smaller than that for
the TCNE-mesitylene complex reported under the same
conditions.8 This observation is possibly explained in
terms of steric and electronic repulsion between TCNE
and an azide moiety of 4, which prevents the close
approach between two components and, therefore, the
formation of a stable complex.9
In order to estimate the association constant for
complex formation K, the concentration dependence of
the absorption spectra of the TCNE-4 charge-transfer
complex was examined in dichloromethane and analyzed
(9) The electronic influence of the azido group on an aromatic ring
should be also considered. At the present time, however, little is known
about the electronic character of the azido group because of its high
reactivity. Smith and his co-workers reported that the azido group
strengthened the acidity of benzoic acid, while electrophilic aromatic
substitution was markedly activated by the azido substitution: Smith,
P. A. S.; Hall, J . H.; Kan, R. O. J . Am. Chem. Soc. 1962, 84, 485.
(8) Merrifield, R. E.; Phillips, W. D. J . Am. Chem. Soc. 1958, 80,
2778.

Trapping of Singlet Nitrene and Didehydroazepine
J . Org. Chem., Vol. 62, No. 10, 1997 3057
deviations of atoms from the mean plane defined by these
11 atoms are within 0.24 Å. Moreover, the mean plane
of the azomethine ylide unit is completely perpendicular
to the plane defined by carbon atoms comprising the
mesitylene unit, so that 5 has a molecular framework of
C2v symmetry.
The second adduct obtained in 73% yield was identified
as the spiroazepine 6. This adduct is the analogue of the
TCNE adduct of 2-nitrenobiphenyl, for which the struc-
ture has been determined by X-ray crystallographic
1
analysis.10 In the H NMR spectrum of 6, the protons
on the azepine ring appeared in a higher field (δ 6.48
and 6.74) compared with the corresponding protons of
the aromatic adduct 5 (δ 7.13). The 13C NMR spectrum
showed a signal assigned to the imino carbon in a
considerably low field (δ 175.2). Three signals assigned
to the carbons of cyclopropane ring and four cyano
carbons appeared at δ 24.8, 29.7, and 66.1 and δ 106.6,
107.8, 108.7, and 108.8, respectively. An inequivalence
of the two dicyanomethylene groups is attributed to the
slow inversion of the spiroazepine ring in solution. All
of these spectroscopic features are in agreement with
those of the TCNE adduct of 2-nitrenobiphenyl.
F igu r e 3. ORTEP diagram of 5. Carbon and nitrogen atoms
are shown as 50% probability surfaces, and hydrogen atoms
are displayed as arbitrary spheres.
P h otop r od u cts of Mesityl Azid e (4) in th e P r es-
en ce of TCNE. A solution of 4 (10 mM) containing 2
equiv of TCNE in acetonitrile was irradiated with the
Pyrex-filtered light of a high-pressure mercury lamp.
After separation by the use of gel permeation liquid
chromatography, two adducts 5 and 6 were isolated. The
No other products were detected in the 1H NMR
spectrum of the photoreaction mixture. The ratio of the
adducts formed during the early stages of the photore-
action remained constant. However, continued irradia-
tion of the solution containing both adducts with a Pyrex-
filtered light resulted in an increase in the product ratio
[5]:[6], which was due to the photochemical decomposi-
tion of the spiroazepine 6. Both adducts 5 and 6 were
obtained in dichloromethane and benzene in the analo-
gous ratios, though the yields were slightly lower com-
pared with in acetonitrile. Finally, we confirmed that
neither 5 nor 6 was formed in the dark.
Both adducts were fairly stable at room temperature,
but they decomposed gradually on heating in benzene at
130 °C to give a complex mixture from which nothing
could be identified. Neither photochemical nor thermal
interconversion between 5 and 6 could be observed.
Thus, two adducts 5 and 6 are isolated in the photolysis
of mesityl azide (4) in the presence of TCNE, the
structures of which formally correspond to the TCNE
adducts of mesitylnitrene (8) and trimethyldidehy-
droazepine (9), respectively. The mechanism for the
formation of both adducts is discussed in the following
sections.
first adduct was obtained in 22% yield as yellow granules.
In its 1H NMR spectrum, the methyl signals appeared
at δ 2.35 and 2.39 with an intensity ratio of 2:1, and the
aromatic protons appeared at δ 7.13 as a singlet. The
13C NMR spectrum of the adduct showed two signals
assigned to the cyano carbons at δ 108.6 and 109.8.
These spectroscopic data indicated the presence of a plane
of symmetry in this adduct, which was not inconsistent
with the aziridine structure 7. There were no precedents,
Wa velen gth Effect on th e P h otolysis. The effect
of wavelength on the photolysis of 4 in the presence of
TCNE was examined to elucidate whether the charge-
transfer complex between 4 and TCNE was responsible
for the formation of the adducts 5 and 6. However, on
irradiation of the charge-transfer absorption band in
acetonitrile with a xenon arc lamp through the glass filter
cutting wavelength shorter than 520 nm for 20 h, neither
5 nor 6 was obtained, and more than 90% of the starting
however, for the isolation of aziridines in the reaction of
an arylnitrene having no electron-withdrawing group
with an alkene, though pentafluorophenylnitrene was
reported to add to tetramethylethylene to give an aziri-
dine derivative.6c,g In order to get the unambiguous
molecular structure of the adduct, X-ray crystallographic
study was carried out. To our surprise, the adduct is not
the aziridine 7, but the azomethine ylide 5, where the
original carbon-carbon double bond of TCNE is cleaved.
The ORTEP diagram is given in Figure 3. The azome-
thine ylide unit N[C(CN)2]2 in 5 is almost planar, where
(10) Murata, S.; Sugawara, T.; Iwamura, H. J . Chem. Soc., Chem.
Commun. 1984, 1198.

3058 J . Org. Chem., Vol. 62, No. 10, 1997
Murata et al.
Sch em e 2
materials was recovered. It was reported that irradiation
of the charge-transfer absorption bands of TCNE-
toluene systems gave 3-phenylpropane-1,1,2,2-tetracar-
bonitrile.11 We could not detect this type of addition
product in the excitation of the charge-transfer absorp-
tion bands of the TCNE-4 complex, either, which is
probably due to the instability of the complex as indicated
by the small association constant K. On the other hand,
irradiation of 4 in the presence of TCNE (2 equiv) with
the light filtered by a 30 nm band-pass filter centered at
360 nm for 6 h resulted in the formation of the adducts
5 and 6 in 18% and 71% yields, respectively, in which
the conversion of the starting azide 4 was 60%. Note that
the ratio of the adducts 5 and 6 obtained in the photolysis
with the light through the band-pass filter (360 ( 15 nm),
where the charge-transfer absorption bands were not
excited, was nearly identical to that obtained with the
Pyrex-filtered light (>300 nm). The same wavelength
effect on the photolysis of 4 in the presence of TCNE was
observed in dichloromethane.
These observations strongly suggest that both adducts
5 and 6 are produced by trapping of reactive intermedi-
ates formed from free rather than complexed azide and
that the TCNE-4 charge-transfer complex does not
participate in the adduct formation. Moreover, the
mechanism involving the radical cation of the azide 4 and
the TCNE radical anion, which would be formed through
a single electron transfer from the excited azide to TCNE,
can be ruled out for the adduct formation because both
adducts 5 and 6 were obtained in analogous yields in
irradiation of the azide 4 and TCNE in the oxygen-
saturated solution.12
F igu r e 4. Dependence of the product ratio upon the TCNE
concentration in the photolysis of 4.
where [T] is the concentration of TCNE, which is as-
sumed to remain constant throughout the reaction, and
the rate constants k’s are defined as shown in Scheme 2.
We examined the dependence of the product ratio [5]:[6]
on the TCNE concentration under the defined conditions
([4] ) 5.0 mM, [T] ) 10-60 mM, in acetonitrile, λ > 300
nm, 1 h), where the conversion of the azide 4 was 20-
30%. Under these conditions, the photodecomposition of
the adducts could be neglected, and the concentration of
TCNE could be considered unchanged during the photo-
reaction. Figure 4 illustrates a plot of [5]/[6] versus [T],
which gives a straight line (correlation coefficient 0.999).
Thus, the dependence of the product ratio [5]:[6] upon
the TCNE concentration [T] is found to obey eq 2, which
appears to justify the mechanism for the adduct forma-
tion shown in Scheme 2.
Com m en ts on P h otoch em ica l Rea ction s of Ar yl
Azid es w ith Electr on -Deficien t Alk en es. Finally, we
have to refer to the photochemical reaction of aryl azides
other than mesityl azide (4) with electron-deficient alk-
enes. As reported previously, irradiation of 2-azidobi-
phenyl in the presence of TCNE gives a stable spiro-
azepine-type adduct, the structure of which has been
unambiguously determined by X-ray crystallographic
analysis.10 When phenyl, 2-methylphenyl, or 2,4-dimeth-
ylphenyl azide is irradiated in acetonitrile containing
TCNE, the 1H NMR spectrum of the photoreaction
mixture indicates the formation of TCNE adducts of the
corresponding nitrene. However, all attempts to isolate
the adducts from the reaction mixture have been unsuc-
cessful because of their instability. One of the adducts
is characterized by a proton signal at δ 8.2-8.5. This
signal gives reliable evidence of the formation of a
spiroazepine-type adduct because the isolated adduct of
2-nitrenobiphenyl shows a doublet assigned to the imino
proton at δ 8.25 in its 1H NMR spectrum.10 Thus, the
didehydroazepine formed photolytically from phenyl,
2-methylphenyl, or 2,4-dimethylphenyl azide is undoubt-
edly trapped with TCNE to give the corresponding
spiroazepine-type adduct. On the other hand, we have
no evidence for the formation of an azomethine ylide-type
TCNE adduct in the photolysis of aryl azides other than
Thus, we propose that both adducts 5 and 6 originate
primarily from the reaction of the nitrene generated
through the direct excitation of the azide 4 with TCNE.
Dep en d en ce of P r od u ct Ra tio u p on TCNE Con -
cen tr a tion . We found that the ratio of the two adducts
5 and 6 obtained in the photolysis of the azide 4 in the
presence of TCNE was dependent upon the initial
concentration of TCNE; the yield of the azomethine ylide
5 increased with increasing TCNE concentration, and the
spiro-azepine 6 decreased. If the two adducts 5 and 6
are produced by trapping of singlet mesitylnitrene (8S)
and trimethyldidehydroazepine (9) with TCNE, respec-
tively, and these two reactive intermediates are related
as shown in Scheme 2, then the ratio of the adducts [5]:[6]
is given by eq 2,
k1k-i
k2ki
k
[5]
[6]
)
+
1[T]
ki
(2)
(11) Ohashi, M.; Suwa, S.; Osawa, Y.; Tsujimoto, K. J . Chem. Soc.,
Perkin Trans. 1 1979, 2219.
(12) It has been reported that the radical anion intermediates are
effectively quenched by O2: (a) Spada, L. T.; Foote, C. S. J . Am. Chem.
Soc. 1980, 102, 391. (b) Akaba, R.; Kamata, M.; Sakuragi, H.;
Tokumaru, K. Tetrahedron Lett. 1992, 33, 8105. (c) Murata, S.;
Nakatsuji, R.; Tomioka, H. J . Chem. Soc., Perkin Trans. 2 1995, 793.

Trapping of Singlet Nitrene and Didehydroazepine
J . Org. Chem., Vol. 62, No. 10, 1997 3059
1
4 owing to the lack of characteristic H NMR signals in
s-1 estimated for 2S by Platz and his co-workers.2,14 The
reason for this discrepancy is not elucidated yet, though
it would be possible that three methyl groups on the
benzene ring retard the isomerization from the singlet
nitrene to the didehydroazepine or that Platz and his co-
workers underestimated the activation energy for the
ring expansion process from 2S to 3. In any event, it
should be emphasized that the rate constant for the
reaction of singlet mesitylnitrene (8S) with TCNE is
predicted to be on the order of 109 M-1 s-1 or greater.
An alternative candidate for the reactive intermediate
participating in the formation of the azomethine ylide 5
is triplet mesitylnitrene (8T). In analogy with the
carbene chemistry,15 the spin state of the nitrene involved
in an addition reaction to alkenes would be determined
by whether the stereochemistry of the alkenes is retained
or not. Unfortunately, these stereochemical studies could
be inapplicable to the formation of 5 because of the
cleavage of the initial carbon-carbon bond of the alkene.
Thus, though we cannot strictly exclude the mechanism
involving 8T, we should point out that the triplet mech-
anism is incompatible with our observations as follows.
It has been established that the ring expansion from
singlet phenylnitrene (2S) is favored over intersystem
crossing to its triplet state 2T at room temperature and
that this is also the case for singlet 2,6-dimethylphe-
nylnitrene.2 Thus, it is reasonable to think that the rate
of isomerization of the singlet nitrene 8S to the didehy-
droazepine 9 is much greater than that of intersystem
crossing to 8T under our photoreaction conditions. If 5
is produced by trapping of 8T with TCNE, then 9 is the
bifurcation point for the formation of the azomethine
ylide 5 and the spiroazepine 6 because 8T is formed
through intersystem crossing of 8S, which is regenerated
by 9. Consequently, an increase in the initial concentra-
tion of TCNE should result in the acceleration of trapping
of 9 with TCNE, which would lead to a decrease in the
product ratio [5]:[6]. It is not the case as demonstrated
in Figure 4.
We wish to give a discussion on the pathway to the
azomethine ylide 5 from singlet mesitylnitrene (8S) and
TCNE. Taking into account the reactions of a singlet
carbene with alkenes,15 it seems reasonable to think that
8S reacts concertedly with TCNE to give the aziridine 7.
However, we could not detect 7 in the photoreaction
mixture. In order to gain further information about the
TCNE-adducts derived from 8S, theoretical calculations
using the PM3 method were carried out. We optimized
the geometries of the azomethine ylide 5 and the aziri-
dine 7 and calculated their heat of formation. Some
selected bond lengths and angles of the optimized struc-
tures, as well as the calculated heat of formation, are
given in Table 1. The PM3-optimized geometry of 5 is
in fair agreement with the molecular structure deter-
mined by X-ray crystallography. As shown in Table 1,
the heat of formation of 7 is much greater than that of
5, which is probably due to a steric strain in 7, so that it
is revealed that the cleavage of the C(CN)2-C(CN)2 bond
of 7 to give 5 is a thermodynamically favored process.
Thus, we propose that the aziridine 7 would be initially
this type of adduct. Thus, at the present stage we cannot
determine whether an azomethine ylide-type adduct is
also produced but decomposes in the course of isolation
or the formation of the azomethine ylide 5 is due to a
higher nucleophilicity of singlet mesitylnitrene (8S)
compared with the nitrenes generated by photolysis of
phenyl, 2-methylphenyl, or 2,4-dimethylphenyl azide. In
any event, it is thought that the success in competitive
trapping of singlet nitrene and didehydroazepine with
TCNE in the photolysis of mesityl azide (4) is primarily
attributed to the high stability of the both adducts 5 and
6, which arises from steric and electronic effects of three
methyl groups.
We also examined the photochemical reaction of mesi-
tyl azide (4) with electron-deficient alkenes other than
TCNE. However, when 4 was irradiated in the presence
of maleic anhydride, dimethyl fumarate, or fumaronitrile,
no corresponding adducts could be detected in the pho-
toreaction mixture, but azomesitylene and 2,4,6-tri-
methylaniline were isolated as photoproducts. Azoben-
zenes and anilines are known to be produced through the
dimerization and the hydrogen abstraction of triplet
arylnitrenes, respectively.1 Thus, it is found that in the
case of alkene having a lower electrophilicity compared
with TCNE, the rate of reaction of 8S and 9 with the
alkene is considerably reduced, so that intersystem
crossing from 8S to 8T becomes a predominant path-
way.
Discu ssion
On the basis of our observations, we propose that
singlet mesitylnitrene (8S) and trimethyldidehydroazepine
(9) are responsible for the formation of the azomethine
ylide 5 and the spiroazepine 6, respectively, as shown in
Scheme 2. In this section, the reactivities of these
intermediates with TCNE, as well as other possible
mechanisms for the formation of the adducts, are dis-
cussed.
F or m a tion of th e Azom eth in e Ylid e 5. The azome-
thine ylide 5 seems to be produced by trapping of singlet
mesitylnitrene (8S) with TCNE. This provides the first
example of the capture of a singlet arylnitrene having
no electron-withdrawing groups with an alkene. The
least-squares analysis of the plot shown in Figure 4 gives
us kinetic information on the reactivity of 8S. From the
slope of the plot, the ratio of the rate constant for the
reaction of 8S with TCNE to that for the ring expansion
of 8S to 9, k1/ki, is obtained as 25.2 M-1. Platz and his
co-workers reported the absorption spectrum of didehy-
droazepine (3) by laser flash photolysis of phenyl azide
(1),2 and more recently, Schuster and his co-workers
directly observed 3 and its derivatives by the use of time-
resolved IR spectroscopy.4 Both groups detected 3 im-
mediately after the photolysis of 1 with the 20 ns laser
pulse, which leads to a value of >5 × 107 s-1 for the rate
constant for the ring expansion of 2S to 3. If we assume
ki > 5 × 107 s-1, the rate constant for the reaction of 8S
with TCNE, k1, is calculated to be >1.3 × 109 M-1 s-1
.
However, if we assume that the reaction of 8S with
TCNE is diffusion controlled (in acetonitrile, k1 ) 1.9 ×
1010 s-1),13 a maximum rate for ki is calculated to be 7.5
× 108 s-1, which is not consistent with the value of 1011(1
(14) McClelland and his co-workers estimated the rate constant for
the ring expansion of 2S as (2-4) × 1010 s-1, which was based on the
assumption that the rate constant for the protonation of 2S was equal
to that of alkylamines: ref 7b.
(15) For reviews on carbene chemistry, see, for example: (a)
Carbenes; Moss, R. A., J ones, M., J r., Eds.; Wiley: New York, 1973,
1975; Vols. 1 and 2. (b) Kirmse, W. Carbene Chemistry; Academic
Press: New York, 1971. See also ref 1d.
(13) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photo-
chemistry, 2nd ed.; Marcel Dekker: New York, 1993.

3060 J . Org. Chem., Vol. 62, No. 10, 1997
Murata et al.
Ta ble 1. Selected Bon d Len gth s (Å), Bon d An gles (Deg),
Tor sion An gles (Deg), a n d Ca lcu la ted Hea t of F or m a tion
(k ca l m ol-1) of 5 a n d 7a
distinguished from trapping of 9 with TCNE. Finally,
the formation of 6 can be interpreted by an electrophilic
reaction of TCNE with 9, which has a canonical structure
9a .3a
5
7
PM3
X-ray
PM3
N1-C1
1.470
1.363
2.435b
1.479(7)
1.347(4)
2.429(5)b
1.439
1.495
1.531
61.6
177.7
180.0
111.4
53.2
N1-C10
C10-C11
C10-N1-C11
C10-C12-N2
C10-C13-N3
C12-C10-C13
C2-C1-N1-C10
126.6
128.8(4)
175.0
177.9
114.9
89.8
173.5(5)
177.4(4)
117.1(3)
88.7(2)
Thus, in this paper, we assign the reactive intermedi-
ate involved in the formation of 6 tentatively as trimeth-
yldidehydroazepine (9). In order to validate the inter-
mediacy of the carbene 10, further experimental, as well
as theoretical, studies are required.
heat of formation
188.855
200.233
a
Atom numbers of 7 correspond to those of 5, which are
b
designated in Figure 3. Interatomic distances.
produced by the concerted reaction of 8S with TCNE and
then isomerize immediately to the azomethine ylide 5.
To validate this assumption, however, estimation of the
activation energy for the isomerization of 7 to 5 is
required.
Con clu sion s
Irradiation of mesityl azide (4) in the presence of TCNE
gave a mixture of two stable adducts, which were
identified as the azomethine ylide 5 and the spiroazepine
6. On the basis of the effects of wavelength of the light
employed in the photolysis and the TCNE concentration
on the product ratio, it is proposed that the adducts 5
and 6 are produced by trapping of singlet mesitylnitrene
(8S) and trimethyldidehydroazepine (9) with TCNE,
respectively. To our knowledge, this provides the first
example of competitive trapping of singlet arylnitrene
and its ring-expanded isomer with an alkene. Moreover,
the intermediacy of singlet nitrene in the photochemistry
of an aryl azide having no electron-withdrawing groups
is established for the first time by these findings.
F or m a tion of th e Sp ir oa zep in e 6. Kinetic informa-
tion on the reactivity of trimethyldidehydroazepine (9)
with TCNE can be obtained from the intercept of the plot
of [5]/[6] against [T], which corresponds to k1k-i/k2ki.
Unfortunately, as shown in Figure 4, the intercept of the
plot is zero within the experimental error. Assuming
kik-i/k2ki < 0.01, we obtain the ratio of the rate constant
for the regeneration of 8S from 9 to that for the reaction
of 8S with TCNE, k-i/k2, as <4 × 10-4 M. Thus, it is
found that the isomerization of 8S to 9 is practically
irreversible in the presence of TCNE (10-60 mM). The
intrinsic lifetime for didehydroazepine (3) has been
estimated to be 4.8 ms in heptane by the extrapolation
of the plot of the decay rate constant of 3 against the
phenyl azide concentration to zero.4b If we employ this
value to estimate k-i, the rate constant for the reaction
of 9 with TCNE, k2, is calculated to be >5.2 × 105 M-1
Exp er im en ta l Section
Gen er a l Meth od s. 1H and 13C NMR spectra were recorded
at 270 and 67.8 MHz, respectively. Gel permeation liquid
chromatography (GPC) was performed on a J ASCO HLC-01
high-pressure liquid chromatograph equipped with a Shodex
GPC H-2001 column. Column chromatography was done on
Fuji Davison silica gel BW-127ZH. TCNE was synthesized
according the literature18 and purified by sublimation.
Mesityl Azid e (4). To a solution of 910 mg (0.73 mmol) of
2,4,6-trimethylaniline (Tokyo Kasei Kogyo Co., Ltd.) in 12 mL
of THF was added a mixture of concentrated hydrochloric acid
(3 mL) and water (24 mL). The reaction mixture was stirred
for 15 min at 0 °C, and a solution of 925 mg (13.4 mmol) of
NaNO2 in 10 mL of water was added dropwise to the solution,
keeping the temperature of the reaction mixture below 5 °C.
The reaction mixture was stirred for 15 min at 0 °C, and excess
nitrous acid was removed by addition of urea. To the cooled
reaction mixture was added dropwise a solution of 1.09 g (16.7
mmol) of NaN3 in 11 mL of water. After the addition, the
reaction mixture was stirred for 1 h at room temperature. The
organic material was extracted with ether, and the extract was
washed with dilute hydrochloric acid (0.1 M) and dried over
Na2SO4. The solvent was removed under reduced pressure,
and the residue was developed on a silica gel column with
hexane to give 748 mg (69%) of 4 as a light yellow liquid. The
s-1
.
The spiroazepine 6 could be considered to be obtained
by trapping of 2-aza-3,5,7-trimethylcycloheptatrienylidene
(10) with TCNE. However, at the present stage we
cannot insist that 10 is the reactive intermediate respon-
sible for the formation of 6 for the following three reasons.
First, though didehydroazepine (3) is established as a
reactive intermediate in the photochemistry of phenyl
azide (1),1 the intermediacy of 2-azacycloheptatrienyl-
idene (11) is controversial at present. There have been
no reports for a direct observation of 11 or an isolation
of the products derived from 11. In 1974, Wentrup
estimated that 11 was 25 kcal mol-1 thermochemically
less stable than phenylnitrene (2).16 The semiempirical
calculations carried out by Shillady and Trindle also
suggested that 11 was less stable than 2, but the energy
gap was much dependent on the method.17 However,
recent MNDO calculations of Schuster and his co-workers
showed that 11 in a triplet state was an accessible
species, the energy of which was only 3 kcal mol-1 above
that of didehydroazepine (3).4b Second, in the analysis
of the plot of [5]/[6] against the TCNE concentration, the
intermediacy of 10 in the formation of 6 is not necessarily
required, since trapping with TCNE of 10 equilibrated
with trimethyldidehydroazepine (9) cannot be kinetically
1
identity and purity of 4 were established by H NMR and IR
spectra: 1H NMR (CDCl3) δ 2.16 (3H, s), 2.23 (6H, s), 6.73
(2H, s); IR (NaCl) 2080, 1480, 1270, 850 cm-1; UV (acetonitrile)
λmax (log ꢀ) 204 (4.45), 242sh (3.82) nm.
Con cen tr a tion Dep en d en ce of th e Absor p tion Sp ectr a
of th e TCNE-4 Ch a r ge-Tr a n sfer Com p lex. Five solutions
ranging in concentration from 2.00 × 10-2 to 9.99 × 10-2 M of
4 in dichloromethane containing TCNE (1.01 × 10-2 M) were
(16) Wentrup, C. Tetrahedron 1974, 30, 1301.
(17) Shillady, D. D.; Trindle, C. Theor. Chim. Acta 1976, 43, 137.
(18) Carboni, R. A. Organic Synsesis; Wiley: New York, 1963;
Collect. Vol. IV, p 877.

Trapping of Singlet Nitrene and Didehydroazepine
J . Org. Chem., Vol. 62, No. 10, 1997 3061
prepared, and the absorbance for maximum absorption (454
nm) of the TCNE-4 charge-transfer complex formed in the
solutions was measured by means of a UV-vis spectropho-
tometer. The results are analyzed according to eq 1 and are
shown in Figure 2.
concentration dependence of the product ratio, six solutions
ranging in TCNE concentration from 1.02 × 10-2 to 6.02 ×
10-2 M were irradiated with a 300-W high-pressure mercury
lamp. The results are analyzed according to eq 2 and are
shown in Figure 4.
Ir r a d ia tion of 4 for P r ep a r a tive Exp er im en ts. The
azide 4 (32 mg, 0.20 mmol) was dissolved in 20 mL of
acetonitrile containing TCNE (51 mg, 0.40 mmol). The solu-
tion was placed in a Pyrex tube, purged with N2 for 10 min,
and irradiated for 8 h with a 300-W high-pressure mercury
lamp at room temperature. After evaporation of the solvent
under reduced pressure, the residue was separated by GPC
with chloroform eluent to give two TCNE-mesitylnitrene
adducts 5 and 6. The identity and purity of the new adducts
X-r a y Cr ysta llogr a p h y of 5. Crystal data: C15H11N5, FW
) 261.29, orthorhombic, space group Fdd2, a ) 13.4088(6) Å,
b ) 17.6432(9) Å, c ) 11.8141(8) Å, V ) 2794.9(3) Å3, Z ) 8,
Dc ) 1.242 g cm-3, µ ) 6.01 cm-1. A crystal suitable for the
crystallography was grown from dichloromethane-hexane. A
yellow crystal (0.8 × 0.8 × 0.5 mm) was mounted on a Rigaku
AFC-5R automated four-circle diffractometer, and the intensity
data were collected using Cu KR radiation (λ ) 1.541 78 Å).
The ω - 2θ scan technique was employed at a scan rate of
16.0° min-1 in ω, and the scan range was calculated by 1.37°
+ 0.30° tan θ. A total of 598 independent reflections with in
the range 2θ < 120.2° were measured, and 541 reflections with
I > 3.0σI were used for the structure determination and
refinement. The structure was solved by direct methods
(MITHRIL) and refined by a full-matrix least-squares tech-
nique with anisotropic thermal parameters for non-hydrogen
atoms. All of the hydrogen atoms were located on a difference
map, except two hydrogens of the methyl group at the
4-position, the positional parameters of which were obtained
by calculation. The positional and isotropic thermal param-
eters of hydrogens were not refined but were included in the
calculation of structure factors. The weighting scheme w )
4Fo2/σ2(Fo2) was employed. The final R and Rw values were
0.056 and 0.061, respectively. All calculations were carried
out using the TEXSAN crystallographic software package of
Molecular Structure Corporation.
1
were established by H and 13C NMR spectra. Moreover, the
structure of 5 was strictly determined by X-ray crystal-
lography. The yield of the adducts mentioned in the text was
determined by the integration of 1H NMR in the crude reaction
mixture on the basis of the reacted material. [N-Mesityl-N-
(dicyanomethyl)imino]dicyanomethanide (5): yellow granules;
1
mp 213 °C dec; H NMR (CDCl3) δ 2.35 (6H, s), 2.39 (3H, s),
7.13 (2H, s); 13C NMR (CDCl3) δ 16.4, 21.4, 29.7, 108.6, 109.8,
131.3, 132.3, 136.1, 144.6. 4-Aza-1,1,2,2-tetracyano-5,7,9-
trimethylspiro[2.6]nona-4,6,8-triene (6): yellow granules; mp
155 °C dec; 1H NMR (CDCl3) δ 2.22 (3H, s), 2.30 (3H, s), 2.36
(3H, s), 6.48 (1H, s), 6.74 (1H, s); 13C NMR (CDCl3) δ 18.9,
24.7 (overlapping of two carbons), 24.8, 29.7, 66.1, 106.7, 107.8,
108.7, 108.8, 126.7, 130.7, 134.1, 149.7, 175.2.
Ir r a d ia tion of 4 for An a lytica l Exp er im en ts. In a
typical run, a solution (5 mL) of 4 (4.1 mg, 0.025 mmol) and
TCNE (6.4 mg, 0.05 mmol) in acetonitrile was placed in a
Pyrex tube, purged with N2 for 10 min, and irradiated for 1 h
at room temperature. The consumption of the material (20-
30%) and the yield of the adducts based on the reacted material
Su p p or t in g In for m a t ion Ava ila b le: 1H and 13C NMR
spectra of new compounds 4-6 (5 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
ordered from the ACS; see any current masthead page for
ordering information.
1
were determined by the integration of H NMR in the crude
reaction mixture. In the experiment for the effect of wave-
length employed in the photolysis, irradiation was carried out
with a 300-W high-pressure mercury lamp through band-pass
filter (360 ( 15 nm) or with a 500-W xenon arc lamp through
cutoff glass filter (>520 nm). In the experiment for the TCNE-
J O9622901