Photodecomposition of some para-substituted 2-pyrazolylphenyl azides. Substituents affect the phenylnitrene S-T gap more than the barrier to ring expansion

Article abstract of DOI:10.1021/ja982585r

A series of para-substituted (H, Me, Cl, F, CF₃, OMe, NMe₂) phenyl azides bearing a dimethylpyrazolyl group in position 2 allowing intramolecular trapping of singlet nitrene have been photolyzed at both 295 and 90 K in ethanol. For t

Full text of DOI:10.1021/ja982585r

3104
J. Am. Chem. Soc. 1999, 121, 3104-3113
Photodecomposition of Some Para-Substituted 2-Pyrazolylphenyl
Azides. Substituents Affect the Phenylnitrene S-T Gap More Than
the Barrier to Ring Expansion
Angelo Albini, Gianfranco Bettinetti,* and Giovanna Minoli
Contribution from the Dipartimento di Chimica Organica, UniVersita` di PaVia,
Viale Taramelli 10, I-27100 PaVia, Italy
ReceiVed July 21, 1998
Abstract: A series of para-substituted (H, Me, Cl, F, CF3, OMe, NMe2) phenyl azides bearing a
dimethylpyrazolyl group in position 2 allowing intramolecular trapping of singlet nitrene have been photolyzed
at both 295 and 90 K in ethanol. For three significant models (H, CF3, NMe2), the reaction has been further
studied in the presence of diethylamine (DEA) and of oxygen. With all substituents but NMe2, singlet nitrene
(trapped intramolecularly to give pyrazolobenzotriazoles) and didehydroazepine (trapped with DEA to give
5H-azepines and then rearranging to 3H-azepines) are in equilibrium. With the NMe2 derivative, the
nonelectrophilic singlet is not trapped, while DEA adds to the benzoazirine, the precursor of the didehy-
droazepine. Thus, electronic effects do not hinder the equilibrium between singlet nitrene and its cyclic isomers,
while determining which of the above intermediates decays to a stable end product. The electron-donating
group NMe2 has a second important effect, causing a drastic enhancement of the triplet nitrene energy and
reduction of the S-T gap, so that triplet nitrene is also in equilibrium with the singlet and the benzoazirine.
As for triplet nitrenes, these have been characterized in matrix at 90 K, and the competition between dimerization
(to give azo compounds, as typical of such stabilized species) and hydrogen abstraction from the solvent
(involving a sizable barrier) has been studied. The energetic p-dimethylamino triplet undergoes hydrogen
abstraction exclusively. When present, oxygen adds efficiently to all of the nitrenes, giving a nitroso oxide,
likewise characterized in the matrix, which then converts to the nitroso and nitro derivatives in good yields.
Photochemical excitation of the triplet in matrix leads to intramolecular hydrogen abstraction.
Phenylnitrenes are conveniently generated under mild condi-
Scheme 1
tions by photolysis of the corresponding azides1 and are largely
used both in industrial polymer chemistry for cross-linking (e.g.,
for photoresists) and in biochemistry for photoaffinity labeling.2
The chemistry of these species is quite complex, and various
groups have devoted much work to the unraveling of the
mechanism. A limitation for mechanistic studies is that a poor
yield of isolable products is obtained from photolysis of phenyl
azides at room temperature, in contrast with the good yields
generally obtained from the seemingly related phenylcarbenes.3
This appears to be due to fast rearrangement of the primarily
formed phenylnitrene (11) to dehydroazepine (2).3e In turn, 2
polymerizes (mainly giving “tars”) under this condition.4 In the
presence of a nucleophile, adducts (azepines) have been obtained
3
ranges to the corresponding triplet 1 and this dimerizes to
azobenzene (Scheme 1).3,5 Recent advances in the field include
spectroscopic detection of 1 by picosecond flash photolysis6
1
and the determination of differential thermodynamic parameters
for singlet and triplet phenylnitrene in suitable models where
the singlet is trapped intramolecularly.7
Substitution on the aromatic ring is expected to affect the
reactivity of the nitrene, and choosing the right substituted
derivative is important for optimal results in the above applica-
tions in particular by favoring intermolecular rections, as
obtained, for example, with some poly(fluorophenyl azides).8,9
However, it has not been clarified as yet which step(s) of the
mechanism is (are) affected. Two main alternatives should be
1
and characterized.3 At low temperature, 1 preferentially rear-
* Corresponding author: (fax) 39-382-507323; (e-mail) albini@
chifis.unipv.it.
(1) Iddon, B.; Meth-Cohn, O.; O.; Scriven, E. F. V.; Suschitzky, H.;
Gallagher, P. T. Angew. Chem., Int. Ed. Engl. 1979, 18, 900; Angew. Chem.
1979, 91, 965. Scriven, E. F. V., Ed. Azides and Nitrenes: ReactiVity and
Utility; Academic Press: Orlando, 1984. Lwowski, W., Ed. Nitrenes;
Interscience: New York, 1970.
(2) Kotzyba-Hilbert, F.; Kapfer, I.; Goeldner, M. Angew. Chem., Int. Ed.
Engl. 1995, 34, 1296; Angew. Chem. 1995, 107, 1391.
(3) (a) Wentrup, C. ReactiVe Molecules; Wiley-Interscience: New York,
1984; Chapter 4. (b) Wentrup, C. AdV. Heterocycl. Chem. 1981, 28, 279.
(c) Platz, M. S.; Leyva, E.; Haider, K. Org. Photochem. 1991, 11, 367. (d)
Schuster, G. B.; Platz, M. S. AdV. Photochem. 1992, 17, 69. (e) Platz, M.
S. Acc. Chem. Res. 1995, 28, 487 and references cited.
(5) Leyva, E.; Platz, M.; Persy, G.; Wirz, J. J. Am. Chem. Soc. 1986,
108, 3873.
(6) Gritsan, N. P.; Yuzawa, T.; Platz, M. S. J. Am. Chem. Soc. 1997,
119, 5059. Singlet nitrene had been previously observed upon photolysing
p-dimethylaminophenyl azide: Kobayashi, T.; Ohtani, H.; Suzuki, K.;
Yamaoka, T. J. Phys. Chem. 1985, 89, 776.
(7) Albini, A.; Bettinetti, G.; Minoli, G. J. Am. Chem. Soc. 1997, 119,
7308.
(8) Karney, W. L.; Borden, W. T. J. Am. Chem. Soc. 1997, 119, 1378.
(4) Meijer, E. W.; Nijhuis, S.; Von Vroonhoven, F. C. B. M. J. Am.
Chem. Soc. 1988, 110, 7209.
10.1021/ja982585r CCC: $18.00 © 1999 American Chemical Society
Published on Web 03/19/1999

Photodecomposition of Para-Substituted 2-Pyrazolylphenyl Azides
J. Am. Chem. Soc., Vol. 121, No. 13, 1999 3105
Table 1. Products from the Photodecomposition of Azides 3a-g
in Ethanol
Scheme 2
productsa
azide
subst
H
T, K
4
5
trc
4
24
6
6
7 + 8b
3a
295
90d
295
90
295
90
295
90
295
90
295
90
295
90
72
1
32
41
10
3
18
3b
3c
3d
3e
3f
Me
36
7
45
Cl
64
tr
27
tr
66
11
15
28
15
1
17
3
10
33
29
19
1
7
F
64
CF3
OMe
NMe2
2
44
1
31
5
8
3e
4
3g
3
19
5
a Percent yield on the converted azide (conversion 50-75%). b The
total yield 7 + 8 is reported. Except for the case of azide 3g, which
gives only 7g, the dihydro derivatives 8 predominate in low-temperature
experiments. These are slowly oxidized to the corresponding 7 in
solution (∼1 month); attempts to accelerate the reoxidation caused
partial loss of the products. c Traces (<1%). d Irradiation in ethanol glass
at 90 K followed by “fast heating” (see text). e The azo compound is
obtained in good yield under different conditions, e.g., 62% by
irradiation of a 4 × 10-3 M solution in MeCN.10c
were carried out, and the products were determined by HPLC.
The products obtained by irradiation in carefully deoxygenated
ethanol at 295 K are shown in Table 1. The heteropentalenes 4
are the main products from the parent azide 3a as well as from
the 5-methyl (3b), chloro (3c), fluoro (3d), and trifluoromethyl
(3e) derivatives. With 3b and 3d an about equimolecular amount
of the corresponding amine was also formed, however. With
the 5-methoxy (3f) and 5-(dimethylamino) (3g) azides no
heteropentalene was formed, and the main products were the
corresponding amines 5 and azo compounds 6, with a minor
amount of the pyrazoloquinoxalines 7 and 8.
In representative cases, the irradiation was carried out in the
presence of 0.1 M diethylamine (DEA). Under such condition,
the yield of heteropentalene 4a from azide 3a decreased from
72 to 36% and a 22% yield of the 3H-2-diethylamino-3-
pyrazolylazepine 9a was obtained (Table 2, Scheme 3). The
structure of this adduct was suggested by NMR evidence and
finally proved through single-crystal X-ray analysis, thus
correcting a previous assignment as the isomeric 2-(diethyl-
amino)-7-pyrazolyl derivative7 (see Experimental Section).
With the trifluoromethyl azide 3e, the yield of heteropentalene
4e likewise decreased to less than half of the original value and
two aminoazepines were formed. One of these was by far the
main product immediately after irradiation. However, we did
not succeed in isolating it in the pure state because it converted
to the latter one (itself stable and easily crystallized) during
attempted purification. This compound was quantitatively rear-
ranged by stirring a benzene solution of the reaction products
in the presence of silica gel. Mass spectroscopy and NMR
analysis showed (see Experimental Section) that the original
main product was the 5H-2-(diethylamino)-3-pyrazolylazepine
10e and the rearranged product was the 3H tautomer 9e, the
structural analogue of 9a (Scheme 3).
confronted: (1) the substituent affects the barriers for the
interconversion of singlet and triplet nitrene and rearranged
isomers such as dehydroazepine 2 (see Scheme 1), and (2) the
intrinsic reactivity of some of these intermediates is changed.
In the hope of providing some experimental evidence for
clarifying the above issue, we made recourse to 1-(2-azidophen-
yl)-3,5-dimethylpyrazole (3a), which has been previously shown
to be a useful model because singlet and triplet nitrene give
well-characterized products through intramolecular trapping.7,10
We present here a systematic study of derivatives substituted
in position 5 of the phenyl ring (para to the azido functionality)
at different temperatures and in the presence of suitable
nucleophile and radical traps.
Results
Photochemistry of Azides 3 at Room Temperature. In a
previous preparative study, we showed that the products from
thermal and photochemical decomposition (in acetonitrile) of a
range of substituted 1-(2-azidophenyl)-3,5-dimethylpyrazoles 3
were the heteropentalenes 4, the amines 5, the azo derivatives
6, and the pyrazoloquinoxalines 7, the last ones via oxidation
of the corresponding 4,5-dihydro derivatives 8 (Scheme 2).10c
The product distribution depended strongly on the structure and
on the decomposition method used.
In the present work, we considered most of the above
derivatives and also the fluoro derivative. Small-scale photolyses
(9) (a) Poe, R.; Schnapp, K.; Young, M. J. T.; Grayzar, J.; Platz, M. S.
J. Am. Chem. Soc. 1992, 114, 5054. (b) Schnapp, K. A.; Poe, R.; Leyva,
E.; Soundararajan, N.; Platz, M. S. Bioconjugate Chem. 1993, 4, 172. (c)
Schnapp, K. A.; Platz, M. S. Bioconjugate Chem. 1993, 4, 178. (d) Gritsan,
N. P.; Zhai, H. B.; Yuzawa, T.; Karweik, D.; Brooke, J.; Platz, M. S. J.
Phys. Chem. A 1997, 101, 2833. (e) 4-Cyano- and 4-nitrophenyl azides
react with amines to give hydrazines, however: ref 9f, g. (f) Liang, J. M.;
Schuster, G. B. J. Am. Chem. Soc. 1987, 109, 7804. (g) Odum, R. A.;
Aaronson, A. M. J. Am. Chem. Soc. 1969, 91, 5680.
(10) (a) Lindley, J. M.; McRobbie, I. M.; Meth-Cohn, O.; Suscitzky, H.
J. Chem. Soc., Perkin Trans. 1 1980, 982. (b) Albini, A.; Bettinetti, G.;
Minoli, G. J. Am. Chem. Soc. 1991, 113, 6928. (c) Albini, A.; Bettinetti,
G.; Minoli, G. Heterocycles 1995, 40, 597.
The dimethylamino azide 3g also gave an adduct with DEA,
with corresponding lowering of the yield of the other products,
in particular of amine 5g. This had a different structure, however,
and spectroscopic evidence (see Experimental Section) showed
that it was the triaminophenylpyrazole 11 (Scheme 3).
Matrix Spectroscopy. Azides 3a-g were irradiated in glassy
ethanol at 90 K by using a 254-nm light source. In all cases,

3106 J. Am. Chem. Soc., Vol. 121, No. 13, 1999
Albini et al.
Table 2. Products from the Irradiation of Azides 3a, 3e, and 3g
under Various Conditions
Table 3. Absorption Spectra of Substituted Triplet Phenylnitrenes
16 and of the Nitroso Oxides 22 in Ethanol Matrix
productsa
main bands, λ, nm (log ꢀ)a
substrate
subst
azide subst T, K
condition
4
5
6
7 + 8b
other
16a
16b
16c
16d
16e
16f
16g
22a
22e
22g
H
Me
Cl
510 (3.3)
498 (2.8)
501 (2.9)
491 (2.8)
574 (3.2)
339 (3.6)
339 (3.6)
348 (3.6)
341 (3.8)
354 (3.3)
351 (3.7)
378 (3.8)
315 (3.8)
315 (3.8)
314 (4.1)
313 (3.8)
315 (3.7)
3a
H
295
160
72 trc
37
1
1
4
9
41
<10
10
4
90 fast heatingd
90 slow heatingd
F
43 23
CF3
OMe
NMe2
H
CF3
NMe2
90 double irradd 12
295 DEA 0.1 M 36
165 DEA 0.1 M 41
140 DEA 0.1 M 29
6
4
1
5
7
1
1
3
54
5
7
15
27
2
9 (22)
492 (3.6)
406 (3.9)
380 (3.9)
514 (3.5)
298 (3.8)
9 (10)
9 (4)
100 DEA 0.1 Me
90 O2
5
6
1
2
e
12 (49), 13 (13)
a Log ꢀ are approximate values. For comparison, triplet phenylnitrene
in EPA at 77 K shows the following bands: ∼500 (log ꢀ ≈ 2.5), 400
(3), 320 (3), and 240 (4) nm.
3e CF3
295
165
66
36
9
90 fast heatingd 11 17 44
8
15
77
90 slow heatingd
6
72
9
1
90 double irradd 10
295 DEA 0.1 M 30
165 DEA 0.1 M 23
K was complemented by product determination at low temper-
atures. The results obtained by photolysis in the matrix at 90 K
are reported in Table 1. Under this condition, the azo compounds
6 were the main products from all of the azides tested, including
the methoxy-substituted azide 3f, but not from the dimethyl-
amino derivative 3g from which amine 5g predominated.
9 (21), 10 (3)
9 (5), 10 (tr)
10 (5)
1
12
90 DEA 0.1 Me
6
3
12 39
1
6
e
90 O2
12 (40), 13 (18)
3g NMe2 295
33
31
29
36
22
14
13
17
14
20
5
4
5
1
3
3
165
90 fast heatingd
19
15
38
2
90 slow heatingd
90 double irradd
295 DEA 0.1 M
165 DEA 0.1 M
90 DEA 0.1 Me
90 DEA 0.1 Mf
The temperature effect was more extensively examined for
azides 3a, 3e, and 3g, for which it was also extended to
experiments in the presence of DEA (see Table 2). The
photolysis was carried out at 165 and at 90 K. In the latter case,
two follow-up procedures were used after photolysis. The first
one was the same used for the general case above; viz. the
temperature was allowed to increase quickly and the glass
melted in a few minutes. In the latter one, the temperature was
maintained at 90 K for 90 min and then kept at 100 (2 h) and
at 110 K (3 h). Spectra were taken at regular intervals; after
this time the sample was heated to room temperature and
analyzed. The main products obtained from azides 3a and 3e
by quick melting of the glass were the azo derivatives (>40%),
with none (from 3a) or little (from 3e) of the heteropentalene
obtained at room temperature. The results were similar in the
presence of DEA, with the azo as the main product and none
or only a little of the azepine. Slow heating of the glass, on the
other hand, mainly gave the amines 5a and 5e. Indeed, in the
case of 1e, the spectrum at 110 K after 3 h was virtually identical
to that of the amine (Figure 2a) and in fact no azo was formed
under this condition. Small amounts of the dihydropyrazolo-
quinoxalines 8a and 8e were also formed. The product distribu-
tion at 165 K was intermediate between the two previous
conditions.
11 (23)
11 (25)
3
10 11 (9)
9
11 (7)
e
90 O2
13 12 (10), 13 (14)
a Percent yield on the converted azide (conversion 50-75%). b See
note b, Table 1. c Traces (<1%). d See text. e Fast heating, see text.
f Slow heating, see text.
Scheme 3
In the case of the dimethylamino azide 3g, the yield of the
main product in ethanol, amine 5g, was only marginally affected
by the temperature, but the yield of the pyrazoloquinoxaline
7g increased considerably upon lowering the temperature; the
low amount of azo compound formed at room temperature was
maintained when the 90 K glass was rapidly melted, not upon
slow heating. In the presence of DEA, amine addition to give
11 was maintained, although with a lower yield than at 295 or
165 K.
virtually complete consumption of the azide was obtained after
a few minutes of irradiation, and a colored intermediate was
formed. The spectra observed by photolysis of azides 3a-e were
rather similar, with several well-separated bands extending out
to 550-600 nm. As it appears from Figure 1 and Table 3, there
were three main bands (or group of bands) well recognizable
along the series in the range 300-550 nm. The bands were
sensitive to substitution, as far as both the position and the
intensity are concerned. The longest wavelength band was not
detected with azide 3f. On the other hand, 3g gave rise to a
characteristic spectrum, little comparable to the previous ones
(see Figure 1). In all cases, the spectrum observed remained
unchanged for hours at 90 K, underwent minor modifications
by raising the temperature by 5-10 K, and faded upon melting
of the glass.
Double-Irradiation Experiments. The formation of a col-
ored intermediate at 90 K suggested the possibility of a double-
irradiation experiment. Thus, after the initial 254-nm photolysis
was completed, the samples were submitted to visible (λ > 455
nm) light irradiation. This led to bleaching of the visible band
and a change in the product distribution as determined after
melting the glass (see Table 2). Under this condition, the main
products from 3a and 3e were by far the dihydroquinoxalines
8, accompanied by minor amounts of the heteropentalenes and
Temperature-Dependent Photolysis. The above study at 295

Photodecomposition of Para-Substituted 2-Pyrazolylphenyl Azides
J. Am. Chem. Soc., Vol. 121, No. 13, 1999 3107
Figure 1. (a) Spectrum (A vs λ) obtained by irradiating (254 nm) 1 × 10-4 M azide 3a in EtOH (after degassing) at 90 K for 8 min. (b-g) Same
for azides 3b-3g.
amines. As shown in Figure 2b for the case of the trifluoro-
methyl derivative, the cyclized compound 8e was present before
melting of the glass, and the spectrum observed after double
irradiation was virtually identical to that of 8e. With the
dimethylamino derivative the quinoxaline 7g was the main
product, but the yield of the amine was decreased only slightly.
Matrix Photolysis in the Presence of Oxygen. Room-
temperature photolysis was carried out only in deoxygenated
solutions, since the efficient self-sensitized oxygenation of
heteropentalenes 4, the main products from most of these azides,
would otherwise complicate the result due to the intervening
of secondary photoreactions. Better suited for mechanistic
studies was irradiation in an ethanol matrix using an oxygen-
equilibrated solution. Under this condition, spectra different from
those observed after degassing were obtained (see Figure 3,
Table 3). Noteworthy, a portion of the same transient observed

3108 J. Am. Chem. Soc., Vol. 121, No. 13, 1999
Albini et al.
Figure 2. (a) Evolution of the spectrum (A vs λ) obtained by irradiating
1 × 10-4 M azide 3e in EtOH (after degassing) at 90 K upon heating.
The spectrum does not change for several hours at 90 K; bringing the
temperature a 100 K causes only a slight red shift of the long-
wavelength band. The spectra are taken at 30-min intervals at 100 K
(the first four lines) and then every 30 min at 110 K. Insets: (1) Final
spectrum; (2) spectrum of amine 5e in ethanol. (b) Evolution of the
spectrum (A vs λ) obtained by irradiating 1 × 10-4 M azide 3e in EtOH
(after degassing) at 90 K upon further irradiation at λ > 455 nm at 90
K. Initial spectrum, as observed after the 254-nm irradiation; following
spectra, after 4, 8, 12, and 24 min of 455-nm irradiation. Insets: (1)
Final spectrum after 24-min irradiation at 90 K; (2) spectrum of
dihydroquinoxaline 7e at room temperature; the band at ∼370 nm in
inset 1 is due to heteropentalene 4e formed as a minor product under
this condition (see Table 3).
Figure 3. (a) Evolution of the spectrum (A vs λ) obtained by irradiating
(254 nm) 1 × 10-4 M azide 3a in EtOH (oxygen equilibrated) at 90 K
for 8 min. The first spectrum is taken after complete consumption of
the azide, and following spectra are taken after 7, 17, 32, and 47 min,
respectively, at 90 K. The bands around 300 nm are due to residual
triplet nitrene. Inset: the spectrum after further 40 min at 115 K. (b)
Evolution of the spectrum (A vs λ) obtained by irradiating (254 nm) 1
× 10-4 M azide 3g in EtOH (oxygen equilibrated) at 90 K for 8 min.
The first spectrum is taken after complete consumption of the azide,
and following spectra are taken at 20-min intervals at 90 K. The band
at 492 nm is due to residual triplet nitrene. Inset: the spectrum after
80 min at 90 K.
accompanied by minor amounts of the corresponding nitro
compounds 13a and 13e (see Table 2). The corresponding
derivatives 12g and 13g were formed also in the case of 3g;
this occurred at the expense of quinoxaline 7g, while the yield
of the amine 5g did not decrease.
in oxygen-free glass could be detected immediately after
irradiation (see Figure 3) and was largely converted to the final
spectrum in minutes.
The matrix spectra both at 90 and at 115 K in the presence
of oxygen were stable for hours at these temperatures. Upon
further heating, the transient decomposed and the main products
from 3a and 3e were the nitroso derivatives 12a, and 12e
Discussion
The present azides can be divided into two groups as far as
their photochemical behavior at room temperature is concerned.
Thus, the parent compound 3a, the 5-methyl derivative 3b, and
all compounds containing an electron-withdrawing substituent
in position 5 (chloro, 3c; fluoro, 3d; trifluoromethyl, 3e) give
the heteropentalenes 4 by direct irradiation; compounds 4 are
by far the main product with 3a, 3c, and 3e and are formed in

Photodecomposition of Para-Substituted 2-Pyrazolylphenyl Azides
J. Am. Chem. Soc., Vol. 121, No. 13, 1999 3109
Scheme 4
Scheme 5
an amount equivalent to the amines 5 in the case of 3b and 3d
(see Table 1). We previously showed10c that acetophenone-
sensitized photolysis of the same azides gave the azo derivatives
6 as the main products (with a substantial amount of the amine
5e and the pyrazoloquinoxaline 7e from the trifluoromethyl
derivative 3e). The azides bearing an electron-donating sub-
stituent (methoxy or dimethylamino) form the second group.
No heteropentalene is formed from these substrates, and the
products are the corresponding azo and amino derivatives as
well as the pyrazoloquinoxalines. In this case, the same products
were obtained both from direct and from sensitized photolysis,
although in different yields.
short-lived and furthermore lacks the structural features that
make 2 easily identified by spectroscopy. Thus it is not
surprising that it has never been detected, although the analogue
in the naphthalene series has.15 Chemical trapping of 14 has
been reported in a single case up to now, with the isolation of
2-ethylthioaniline in the presence of EtSH;16 in the case of a
3-acetamido-4-trifluoroacetamidophenyl azide trapping of the
benzoazirine has been obtained also with amines.17 The above-
mentioned calculations also show that 2 is only marginally
stabilized with respect to the 1A2 nitrene (∼21 kJ mol-1). Thus,
all three species, 11, 14, and 2, are predicted to be in equilibrium
at room temperature.
This confirms the idea that led Suschitzky10a to originally
propose the present model. Thus, electrophilic singlet nitrene
is intramolecularly trapped by the pyrazole nitrogen to yield
the heteropentalene. Triplet nitrene, formed by sensitization,
abstracts a hydrogen atom from the neighboring methyl group
or from the solvent, yielding respectively the dihydropyrazolo-
quinoxaline or the amine, or alternatively dimerizes to the azo
compound; the last one turns out to be the most efficient process.
In the present series, the nitrene reactivity is modulated by
the para substituent. The selective intramolecular trapping of
both singlet and triplet nitrene, in connection with the use of
external nucleophilic (DEA) or radical (oxygen) traps, forms
the basis of the present study.
Singlet Nitrene, Benzoazirine, and Dehydroazepine. As
mentioned above, the failure of trapping singlet phenylnitrene
11 has been attributed to fast rearrangement to didehydroazepine
2.3e The intermediacy of 2 has been demonstrated both
spectroscopically5,11,12 and by reaction with amines (the rate
constants for the addition of DEA to para-substituted 2 is in
the range 106-108 M-1 s-1 in heptane).12 Recent computational
evidence by Karney and Borden13 supports the view14 that the
rearrangement is a two-step process via the benzoazirine 14
(Scheme 4) and shows that it proceeds from the lowest lying
singlet state (1A2), in which the p-π orbital is delocalized and
the C-N bond has double bond character (compare 15). The
With the present pyrazolylphenyl azides 3, the singlet is
1
trapped to give 4. Our previous study demonstrated that 16a
and didehydroazepine 17a (see Scheme 5) are in equilibrium.
The difference between the activation energies for 4a (intra-
molecular trapping of singlet nitrene) and for 9a (DEA addition
to the didehydroazepine, where the rate-determining step is
thought to be cyclization to benzoazirine 18a), is minimal (3
kJ mol-1). Intuitively, it is not surprising that a similar (and
low) barrier is encountered in these two rearrangements. In both
cases, the reaction involves bending of the nitrene C-N bond
to reach either cyclization onto the nN pyrazole orbital or onto
the phenyl π orbital (Scheme 5).
As for substituents, recent calculations showed that the
electronic effect on the cyclization barrier is small (e.g., for both
2-fluoro and 4-fluorophenylnitrene).8,18 On the contrary, steric
hindrance is determining (in the same way for both fluorine
and methyl).8 Indeed, experiments show that pyridine trapping
of unrearranged singlet nitrene is easier with 2,6-disubstituted
phenyl azides9 and that with 2-substituted phenyl azides
cyclization always occur away from the substituent, as shown
by the isolation of 3-(not 2-)substituted 3H-2-dialkyl-
aminoazepines in the presence of secondary amines.12
We explored the electronic effect of substituents by compar-
ing the photochemistry of the trifluoromethyl- and the di-
methylamino-substituted azides 3e and 3g. The first one behaved
similarly to 3a and gave 4e in neat ethanol and both 4e and
azepines 9e and 10e in the presence of DEA (with 0.1 M DEA
the heteropentalene/aminoazepine(s) ratio is 1.6 from unsub-
stituted 3a and 1.25 from 3e). Thus, an electron-withdrawing
substituent has little effect on the partitioning of singlet nitrene
between cyclization onto the pyrazole and onto the phenyl ring
(see Scheme 5).9
first step (11 f 14) is rate determining (∼25 kJ mol-1 barrier)
and is followed by a symmetry-allowed six-electron electrocyclic
ring opening.13 Benzoazirine 14 is thus expected to be quite
With a strong electron-donating substituent as in 3g, on the
other hand, cyclization to the heteropentalene is suppressed.
Donation from the substituent generates a negative charge on
(11) (a) Schrock, A. K.; Schuster, G. B. J. Am. Chem. Soc. 1984, 106,
5228. (b) Marciniek, A.; Leyva, E.; Whitt, D.; Platz, M. S. J. Am. Chem.
Soc. 1986, 108, 3783. (c) Chapman, O.; Le Roux, J. P. J. Am. Chem. Soc.
1978, 100, 282. (d) Hayes, J. C.; Sheridan, R. S. J. Am. Chem. Soc. 1990,
112, 5879.
(15) Dunkin, I. R.; Thompson, P. C. P. J. Chem. Soc., Chem. Commun.
1980, 499.
(12) Li, Y. Z.; Kirby, J. P.; George, M. W.; Poliakoff, M.; Schuster, G.
B. J. Am. Chem. Soc. 1988, 110, 8092.
(13) Karney, W. L.; Borden, W. T. J. Am. Chem. Soc. 1997, 119, 3347.
(14) Huisgen, R.; Vossius, D.; Appl, M. Chem. Ber. 1958, 91, 1. Huisgen,
R.; Appl, M. Chem. Ber. 1958, 91, 12.
(16) Carrol, S. E.; Nay, B.; Scriven, E. F. V.; Suscitzky, H.; Thomas,
D. R. Tetrahedron Lett. 1977, 3175.
(17) Younger, C. G.; Bell, R. A. J. Chem. Soc., Chem. Commun. 1992,
1359.
(18) Smith, B. A.; Cramer, C. J. J. Am. Chem. Soc. 1996, 118, 5490.

3110 J. Am. Chem. Soc., Vol. 121, No. 13, 1999
Albini et al.
Scheme 6
hydrogen shift rather than ring opening. Notice also that Schuster
did not detect a didehydroazepine intermediate from p-(di-
methylamino)phenyl azide.12 A way to rationalize this result is
to assume that the dimethylamino group stabilizes intermediate
21 through mesomeric structure 21′ and this provides the driving
force favoring hydrogen shift (reasonably a basesDEA-
catalyzed process, not a pericyclic reaction) over ring enlarge-
ment.
Nitrene Intersystem Crossing (ISC) and Temperature
Dependence of Azide Photoreactions. Calculation and experi-
ments show that triplet phenylnitrene is stabilized by ∼70 kJ
mol-1 with respect to the singlet state but ISC is slow (106-
107 s-1).3e Thus, the triplet nitrene product (azobenzene) is
obtained by photolysis of phenyl azide only at low temperature
(<160 K). With the pyrazolyl azide 3a, where there is no reason
to expect that the S-T gap is different, the differential activation
parameters for the singlet reaction and ISC were measured as
∆∆Hq ) -10 ( 2 kJ mol-1 and ∆∆Sq ) 34 ( 3 J mol-1 K-1
,
showing that it is the entropic term that disfavors ISC.7
Assuming that the latter process has negligible activation
enthalpy, the observed quantity (10 kJ mol-1) is directly the
activation enthalpy of intramolecular ring closure of singlet
1
nitrene 16 to give 4 (see Scheme 5). As shown above, the
barrier of the other singlet nitrene rearrangement, cyclization
to 18 and hence to 17, is only 3 kJ mol-1 larger, thus ∼13 kJ
mol-1.7 This is somewhat lower than recent experimental and
calculated data for the rearrangement of parent 11 to 14 (25 kJ
mol-1).13
the nitrene nitrogen (19), and this inhibits electrophilic attack
onto the pyrazole nitrogen. This does not hold for cyclization
The substituents affect the ISC to various degrees. The Cl
and CF3 groups have little effect, and as with the parent
compound, there are essentially no triplet products at room
temperature and only triplet products at 90 K. With the methyl
and fluoro derivatives, however, the proportion of triplet-derived
products (mainly the amines 5) is large already at 295 K (45
and 51%, respectively; the rest is the singlet derived hetero-
pentalene) and increases to 100% at 90 K. Thus, ISC is more
effective with these substrates.
onto the phenyl ring, however, possibly because the decrease
in the nitrene electrophilicity is compensated for by the
zwitterionic character of the nitrene (see 19). Indeed the isolated
yield of DEA addition products, whatever their structure,
remains the same (22-24% with 0.1 M DEA) with the three
azides examined (3a, 3e, 3g).
The case of the dimethylamino-substituted azide 3g has been
studied in more detail. In this case, direct intramolecular trapping
of singlet nitrene does not occur. However, the competition
between nucleophile trapping of the benzoazirine 18g and triplet-
derived products (5g, 7g) can be studied. Azide 3g shows the
smallest temperature effect, and nucleophile trapping (in the
presence of 0.1 M DEA) accounts for 59% of the total isolated
products (the rest being triplet-derived products) at 295 K and
moderately decreases to 25% in the glass at 90 K. This is the
lower limiting value for the benzoazirine, since bimolecular
The mode of nucleophile trapping of such intermediates
deserves some comment. Previous studies by Schuster showed
that para-substituted phenyl azides give a 3H-2-aminoazepine.12
We obtained the corresponding azepines 9a and 9e from azides
3a and 3e, but in the latter case the main product after photolysis
actually was isomeric 10e, which then tautomerized to 9e under
mild acid catalysis. This is informative, because the primary
product must be the 1H tautomer 20 (see Scheme 6), and this
is expected to undergo symmetry-allowed 1,5-suprafacial shift
to 10, while a further (forbidden) rearragement to 9 does not
occur concertedly but, for example, through acid catalysis. This
has now been demonstrated in the case of 10e, but presumably
occurs also with other azepines. As for the regiochemistry of
ring expansion, we confirm that, at least at room temperature,
nitrenes cyclize at position 6 and not position 2 in 2-substituted
phenyl azides independently from the electronic nature of the
substituent.13
3
trapping by DEA is hindered in the glass. Thus, triplet 16g
and azirine 18g compete over the entire range explored, and
the situation is really completely different from that of unsub-
stituted 3a, where the triplet products increase from 0 to 100%
in this interval. We see no explanation but to conclude that the
singlet triplet gap is dramatically reduced in this case, so that
singlet nitrene, benzoazirine, and triplet nitrene are all in
equilibrium.
This implies that the electronic structure of the nitrene must
be different in 16g: as mentioned above, the lowest lying singlet
In the case of the dimethylamino azide 3g, DEA gives the
benzenetriamine 11. This is the product expected from addition
to a benzoazirine, actually not the one directly formed by nitrene
cyclization (see 18, Scheme 6) but the rearranged intermediate
21. Thus, the initial nitrene rearrangement occurs as in the other
cases toward position 6, but the cyclization is followed by a
1
state in phenylnitrene is the open shell A2 where electron are
placed in pure p orbitals. With the dimethylamino group
donation from the ring to the nitrene nitrogen becomes important
(see 19) and this limits admixing of the N 2s orbital with the
nitrene lone pair. This removes the factor explaining the greater

Photodecomposition of Para-Substituted 2-Pyrazolylphenyl Azides
J. Am. Chem. Soc., Vol. 121, No. 13, 1999 3111
thermodynamic stability of nitrenes relative to carbenes3e,19 and
makes the structure of this nitrene more similar to that of
phenylcarbene. In the latter species, the singlet-triplet gap is
Scheme 7
small (calculated 8-16 kJ mol-1 19,20
and ISC is faster than
)
bimolecular reactions of both states, so that the triplet functions
as a reservoir for the faster reacting singlet.3e,21 With nitrene
16g the singlet is unreactive, but the triplet is a reservoir for
the azirine (in the presence of DEA, 11 is formed at the expense
of 5g; see Table 3). A support for the enhancement of triplet
energy comes from the observed chemistry: the main product
from 316g (both at 295 and 90 K) is the amine from
intermolecular hydrogen abstraction, rather than the azo com-
pound typical of “lazy” triplets such as 316a and most
phenylnitrenes (see further below).
Spectrum of Nitrene Triplets. Irradiation in an ethanol glass
at 90 K allows complete decomposition of the azides. With
parent 3a and with derivatives 3b-3e the spectra maintain the
same habitus and are closely comparable to those observed by
Platz and Wirz for phenyl azide (Figure 1, Table 3).5 These
spectra are attributed to triplet nitrene, and the observed thermal
and photochemistry of these transients (see below) is that
expected from a triplet. The longest-wavelength band is a πnx
transition.5,22 The systematic survey in Figure 1 shows that an
electron-withdrawing para substituent has little effect on this
band (see Figure 1a-d and Table 3, note a), while it affects
the complex band system in the 300-400-nm region. With the
trifluoromethyl derivative there is a marked red (and ipso-
chromic) shift of the low-energy band while the system at 300-
400 nm is split into two groups of bands (Figure 1e).
With electron-donating substituents, the spectra observed are
quite different. Apart from the introduction of strong bands at
270 nm (for the methoxy) and 300 nm (for the dimethylamino
derivative) related to the substituted π system, also the long-
wavelength part of the spectrum is also different. The structure
of the triplet is thus different in these cases, even though some
other transient may be present in the matrix and contribute to
the observed spectrum.
derivative, depends very little on conditions (the ratio is 55/
36/9 at 90 K, with fast heating, 68/28/<4 with slow heating,
and 89/11/0 at 295 K). This demonstrates that this nitrene
resembles more a good hydrogen abstractor such as the
phenylcarbene triplet than the stabilized phenylnitrene.
Oxygen Trapping. All of the present nitrenes add oxygen.
The matrix spectrum in the presence of oxygen is different from
that under degassed conditions (as an example, see Figure 3 vs
Figure 1), and the products obtained after melting are the nitroso
and nitro derivatives. The transient reasonably is the nitroso
oxide 22 (compare Figure 3b, where the residual absorption due
to triplet nitrene is observed immediately after irradiation and
gives way to the final spectrum in minutes, while oxygen
penetrates in the matrix). This is an expected process from
arylnitrenes, although it has been sparsely documented up to
now.23 Flash photolysis at room temperature has failed to reveal
any transient similar to that reported in Figure 3b,11a but oxygen-
transfer studies by Sawaki et al. are consistent with the expected
chemistry of nitroso oxides.24
Triplet Chemistry. Rapid heating allows nitrene migration
as soon as the glass softens and under such condition by far
3
3
the main process from 16a (90%) and from 16e (70%) is
dimerization to azobenzene, since hydrogen abstraction from
the solvent is too slow under this condition. On the other hand,
if the temperature is slowly raised, the matrix softens gradually
and hydrogen abstraction to give the anilines 5a (61%, azo
reduced to 33%) and 5e (83%, no azo) predominates. The fact
that it is sufficient to increase mobility through quick glass
softening to make coupling to the azo compound dominant and
that no intramolecular H abstraction from the pyrazole methyl
groups occurs demonstrate the poor radical character of triplet
phenylnitrenes (“lazy triplet” as dubbed by Suschitzky).10a
The CF3 group marginally increases hydrogen abstraction,
but this process is much more efficient with the dimethylamino-
Matrix studies reveal this transient and show a consistent blue
shift in the spectrum with electron-attracting substituents (see
Table 3; λmax 22g > 22a > 22e). Notice that oxygen addition
predominates (>90%) with “lazy” triplets (316a and 316e) while
3
H abstraction from the solvent remains competitive from 16g
(58-42%), once again demonstrating the high reactivity of this
species (see Scheme 7).
3
substituted nitrene 16g. In this case, the competition among
the three paths, viz. H abstraction from the solvent, intra-
molecular H abstraction (to give 8g), and coupling to the azo
Photochemistry of Nitrene Triplet. Selective excitation of
the triplet nitrenes by irradiation in the visible increases the
radical character and leads to intramolecular hydrogen abstrac-
(19) Kemnitz, C. R.; Karney, W. L.; Borden, W. T. J. Am. Chem. Soc.
1998, 120, 3499.
(23) (a) Phenyl azide and some of its substituted derivatives have been
shown to give 5-20% of the nitro derivative and 0-18% of the azoxy
compound upon photolysis in oxygen-flushed solution at 30 °C;23b 4-ni-
trophenyl azide gives a better yield23c and an oxygen adduct is obtained in
close to quantitative yield by photolysis of 2-phenazinyl azide.23d (b)
Abramovitch, R. A.; Challand, S. R. Chem. Commun. 1972, 964. (c) Liang,
T. Y.; Schuster, G. B. J. Am. Chem. Soc. 1987, 109, 7803. Bettinetti, G.;
Fasani, E.; Minoli, G.; Pietra, S. Gazz. Chim. Ital. 1988, 100, 175.
(24) Sawaki, Y.; Kishikawa, S.; Iwamura, H. J. Am. Chem. Soc. 1987,
109, 584.
(20) Matzinger, S.; Bally, T.; Patterson, E. V.; McMahon, R. J. J. Am.
Chem. Soc. 1996, 118, 1535. Wong, M. W.; Wentrup, C. J. Org. Chem.
1996, 61, 7030. Schreiner, P. R.; Karney, W. L.; Schleyer, P. v. R.; Borden,
W. T.; Hamilton, T. P.; Schaefer, H. F. J. Org. Chem. 1996, 61, 7030.
(21) Moss, R. A.; Dolly, U. H. J. Am. Chem. Soc. 1971, 93, 954. Baer,
T. A.; Gutsche, C. D. J. Am. Chem. Soc. 1971, 93, 5180. Savino, T. G.;
Kanakarajan, K.; Platz, M. S. J. Org. Chem. 1986, 51, 1305.
(22) Kim, S. J.; Hamilton, T. P.; Schaefer, H. F. J. Am. Chem. Soc. 1992,
114, 5349.

3112 J. Am. Chem. Soc., Vol. 121, No. 13, 1999
Albini et al.
tion (since the viscosity of the matrix does not allow for
intermolecular processes). The reaction is in accord with the π
f n character of this state, which favors a radical reactivity,
similarly to an iminyl radical (see 15). This is apparent with
316e, where intramolecular hydrogen abstraction occurs quite
product studies.25 This has been applied here to the substituent
effect. The main points are a negative result; viz. there is no
influence on singlet nitrene cyclization to the benzoazirine and
hence to the dehydroazepine, and a positive one; an electron-
donating substituent enhances the energy of triplet nitrene and
adds it to the pool of species in equilibrium.
3
cleanly (see Figure 2b and the high yield of 8e). With 16g,
intramolecular abstraction is the main process (64%), butsas
in the thermal reactionssintermolecular abstraction remains
important.
Experimental Section
General Information. NMR spectra were run on a Brucker 300
instrument and are reported in CDCl3 with TMS as the internal standard.
IR spectra were on a Perkin-Elmer Paragon 1000 spectrophotometer,
and mass spectra on a Finnigan LCQ instrument. 95% Ethanol was
spectroscopic grade solvent. Column chromatography was performed
with silica gel Merk HR 60. The azides 3a-c and 3e-f were prepared
and purified as previously reported.10a,c, 27
Substituent Effect. As pointed out by Platz,3e phenylnitrene
is a poor reagent in organic chemistry, giving a low yield of
characterized products, since ring enlargement to the didehy-
droazepine precludes the electrophilic intermolecular attack so
characteristic of phenylcarbene. This postulate is satisfactory,
and the study of the present pyrazolyl derivatives shows that
intramolecular electrophilic attack occurs efficiently in suitable
models and that singlet nitrene and the didehydroazepine are
in equilibrium (via the benzoazirine). Some effort has been
recently given to establish whether ring substitution may change
the situation, making intermolecular electrophilic reaction from
the singlet more competitive.8,13,18 This expectation is not borne
out by the present study, showing that the proportion of trappable
cyclized intermediates (the didehydroazepine or the azirine) is
only moderately affected by the substituent (at 295 K, 34% of
the isolated products from the parent azide, 44% from the
trifluoromethyl, and 59% from the dimethylamino derivative).
This is in accord with calculations demonstrating that electronic
factors exert a small effect on this rearrangement.8 Thus, singlet
trapping is effective only with ortho-disubstituted phenyl azides,
not on simple derivatives (unless, as in the present case, the
reaction is intramolecular) and only with unhindered n nucleo-
philes such as pyridine, not with π nucleophiles, which would
be much more interesting from the synthetic point of view. The
electron-donating NMe2 group does not affect the rearrangement
of the singlet; it only allows the trapping of the first intermediate,
the benzoazirine.
Synthesis of 1-[(2-Azido-5-(fluorophenyl)]-3,5-dimethylpyrazole
(3d). To a solution of acetylacetone (5 g, 50 mmol) in 150 mL of
ethanol containing 2 drops of concentrated HCl, 3.42 g (20 mmol) of
2-nitro-5-fluorophenylhydrazine28 was added and the mixture was
refluxed for 30 min. After cooling, a yellow precipitate formed was
discarded. Evaporation of the filtrate and recrystallization from cyclo-
hexane gave 1-(2-nitro-5-fluorophenyl)-3,5-dimethylpyrazole (13d, 2.5
g, 53% yield), light orange prisms, mp 93-94 °C. Anal., found C,
56.0; H, 4.1; N, 17.6. Calcd for C11H10FN3O2: C, 56.17; H, 4.25; N,
17.87. Catalytic hydrogenation of this product according to the usual
procedure gave the amine 5d as colorless crystals (65% yield), mp 69-
70 °C (petroleum ether). Anal., found C, 64.1; H, 5.9; N, 20.1. Calcd
for C11H12FN3: C, 64.39; H, 5.85; N, 20.49. IR (neat) 3442, 3320,
3207 cm-1; 1H NMR δ 2.2 (s, 3H), 2.3 (s, 3H), 3.9 (br, exch, 2H), 6.0
(s, 1H), 6.77 (dd, 1H, J ) 5, 9 Hz), 6.87 (dd, 1H, J ) 3, 9 Hz), 6.93
(ddd, 1H, J ) 3, 8, 9 Hz). Diazotisation of this material in AcOH-
HBF4 and reaction with sodium azide gave 3d as colorless needles,
mp 57-58 °C (petroleum ether, 79% yield). Anal., found: C, 56.8; H,
4.5; N, 30.0. Calcd for C11H10FN5: C, 57.14; H, 4.33; N, 30.30. IR
(KBr) 2127 cm-1; UV 286 nm (3.46), 239 (4.18); 1H NMR δ 2.15 (s,
3H), 2.3 (s, 3H), 6.0 (s, 1H), 7.1 (dt, 1H, J ) 2, 6 Hz), 7.18 (d, 1H, J
) 2, 6 Hz), 7.25 (s, 1H).
Preparation of 1-[2-nitroso-5-(trifluoromethyl)phenyl]-3,5-di-
methylpyrazole (12e). A solution of 1.0 g (3.9 mmol) of the
corresponding amine in 10 mL of chloroform was added with 2.6 g of
3-chloroperoxybenzoic acid (52% assay, 7.8 mmol) in 10 mL of
chloroform at room temperature, while stirring, in 15 min. The solution
turned yellow and after 2 h was extracted with 3 × 10 mL saturated
sodium carbonate. Evaporation of the organic solvent and chromatog-
raphy of the residue gave 0.6 g (55%) of a slightly yellow solid which
was recrystallized from ethanol, mp 122-23 °C. Anal., found: C, 53.6;
H, 3.9; N, 15.6. Calcd for C12H10F3N3O: C, 53.53; H, 3.74; N, 15.61.
1H NMR δ 2.1 (s, 2H), 2.38 (s, 3H), 6.15 (s, 1H), 6.35 (d, 1H, J ) 8
Hz), 7.63 (dd, 1H, J ) 2, 8 Hz), 8.18 (d, 1H, J ) 2 Hz). It should be
noted that in previous studies we showed that a nitroso derivative such
as 12a is better described as being in equilibrium with isomeric 1,3-
dimethylpyrazolo[1,2-a]benzotriazole 5-oxide.29
Substituents have a more important effect on ISC. The
rationalization of this effect is not straightforward, since groups
as different as methyl and fluoro both enhance the proportion
of triplet-derived products at room temperature. At least with
the dimethylamino derivative 16g, where the largest effect is
observed, the reason appears to be a substantial enhancement
of the phenylnitrene energy and a reduction of the singlet-
triplet gap. This phenylnitrene is thus closer to phenylcarbene
in the radical reactivity, but obviously not as far as electrophilic
reactions from the singlet are concerned, since these are
precluded by electron donation from the substituent. The higher
energy of the triplet in this case is reflected in the efficient
intermolecular hydrogen abstraction, as opposed to the usual
dimerization of phenylnitrenes. Triplet reactions may also be
exploited for applications. Reaction with oxygen is a limitation,
Preparative Irradiation. Preparative irradiation and product separa-
tion were carried out as in previous communications.7,10 The charac-
terization of several of the photoproducts was previously reported. The
main data about the new photoproducts are reported below.
3
but the triplets that are efficiently produced, such as 16g, are
also the most reactive ones, and abstract hydrogen also in the
presence of oxygen.
1,3-Dimethylpyrazolo-8-fluoro[1,2-a]benzotriazole (4d): lemon
yellow needles, mp 115-117 °C (cyclohexane). Anal., found: C, 64.8;
H, 5.0; N, 21.0. Calcd for C11H10FN3: C, 65.02; H, 4.92; N, 20.69. 1H
Thus, one may summarize that the introduction of substituents
does not cause electrophilic reactions of singlet phenylnitrene
to occur in a convenient way, while enhancement of the triplet
radical reactivity can be accomplished. This is of little signifi-
cance in synthesis, but may be important for applications such
as photochemical labeling and polymer cross-linking.
(25) (a) Other groups functioning as internal traps are phenyl25b,c and
methoxycarbonyl.26 (b) Swenton, J. S.; Ikeler, T. J.; Williams, B. H. J.
Am. Chem. Soc. 1970, 92, 3103. (c) Sundberg, R. J.; Heintzelman, R. W.
J. Org. Chem. 1979, 39, 2546.
(26) Tomioka, H.; Ichikawa, N.; Kamatsu, K. J. Am. Chem. Soc. 1993,
115, 8621.
(27) Albini, A.; Bettinetti, G.; Minoli, G. J. Org. Chem. 1983, 48, 1080.
(28) Menzel, K. H.; Puetter, R. Belg. Pat. 643 802; Chem. Abstr. 1965,
63, 4440f.
(29) Albini, A.; Bettinetti, G.; Minoli, G. J. Chem. Soc., Perkin Trans.1
1983, 581.
Conclusion. High-energy species such as arylnitrenes have
available a variety of accessible paths, not necessarily leading
to characterized end products. The use of “internal trapping”
models gives better yields allowing us to draw conclusions from

Photodecomposition of Para-Substituted 2-Pyrazolylphenyl Azides
J. Am. Chem. Soc., Vol. 121, No. 13, 1999 3113
NMR δ 2.55 (s, 1H), 2.7 (s, 1H), 6.3 8s, 1H), 7.08 (td, 1H, J ) 2.5,
9.5 Hz), 7.33 (dd, 1H, J ) 5, 9 Hz), 7.37 (dd, 1H, J ) 2.5, 8 Hz).
2,2′-Bis[1-(3,5-dimethylpyrazolyl)]-4,4′-difluoroazoben-
zene (6d): orange-yellow crystals, mp 184-185 °C. Anal., found: C,
64.7; H, 5.2; N, 20.5. Calcd for C22H20F2N6: C, 65.02; H, 4.92; N,
20.69. H NMR δ 2.02 (s, 3H), 2.32 (s, 3H), 6.05 (s, 1H), 7.1 (ddd,
1H, J ) 2.5, 7.5, 9 Hz), 7.3 (dd, 1H, J ) 2.5, 9), 7.5 (dd, J ) 5.5, 9
Hz).
6H), 1.7 (m, 2H), 2.12 (s, 3H), 2.3 (s, 3H), 2.8 (s, 6H), 2.95 (q, 4H),
3.8 (br, exch, 2H), 5.95 (s, 1H), 6.4 (d, 1H, J ) 4 Hz), 6.6 (d, 1H, J
) 4 Hz); 13C NMR δ 11.4 (Me), 12.6 (Me), 13.5 (Me), 42.0 (Me), 47.
5 (CH2), 105.1 (CH), 109.6 (CH), 110.6 (CH), 126.1, 133.4, 139.1,
140.6, 143.3, 148.9. A COLOC experiment on the δ 6.4 signal at 5 Hz
shows a coupling with the carbon signal at 126.1sa doublet attributed
at 1′-C; the signal at δ 133.4 is a triplet, showing that there is no vicinal
hydrogen for this amino-substituted carbon atom. NOE experiments:
irradiation of the δ 6.65 signal gives a 2.5 enhancement of the 2.8
singlet and of the 2.95 q; irradiation of the 2.95 q gives a 1%
enhancement of the 6.6 d. This demonstrates that the 4′-C is flanked
by dimethylamino and diethylamino groups.
Variable-Temperature Irradiation. A 1 × 10-4 M solution of one
of the azides 3 in EtOH (2 mL) in a 1-cm optical path quartz cell with
a quartz to glass graded seal was degassed by means of four freeze-
degas-thaw cycles and sealed. A series of experiments were carried
out by using oxygen-equilibrated solutions (see text). The cell was
inserted into an Oxford DN 1704 liquid nitrogen cryostat fitted with a
calibrated ITC4 temperature controller and placed in a UV-visible
Kontron Uvikon 941 spectrophotometer. The cell was irradiated from
the bottom by means of a bifilar low-pressure mercury arc inserted
below the cell (Helios Italquartz 15 W). Irradiation was discontinued
when the spectra were taken. In a typical experiment, the solution was
equilibrated for 30 min at the desired temperature and then irradiated
for 8 min. After this time, the temperature was either gradually raised
(see text) or allowed to quickly reach room temperature (within 30
min in the latter case). The solution was washed down in a round-
bottomed flask and evaporated in the dark, and the residue was
redissolved in 1 mL of MeCN and analyzed by HPLC. A Jasco PU
980 instrument with UV-975 detector was used, with a 25 cm × 4.6
mm Merck Purospher RP-18 LiChroCART 250-4 column (and a
Purospher RP-18 LiChroCART 4-4 precolumn). Various water-
acetonitrile mixtures were used as the eluent.
1
8-Fluoro-3-methyl-8-pyrazolo[1,2-a]quinoxaline (8d): light yellow
crystals, mp 132-133 °C (cyclohexane). Anal., found: C, 65.6; H,
1
4.1; N, 20.6. Calcd for C11H8FN3: C, 65.67; H, 3.98; N, 20.89. H
NMR δ 2.6 (s, 3H), 6.65 (s, 1H), 7.25 (s, 1H), 7.27 (dt, 1H, J ) 2.5,
9 Hz), 8.05 (dd, 1H, J ) 5.5, 9 Hz), 8.1 (dd, 1H, J ) 2.5, 9 Hz).
3H-2-(Diethylamino)-3-[1-(3,5-dimethylpyrazolyl)]azepine (9a):
erroneously reported as 2H-2-(3,5-dimethylpyrazolyl)-7-diethylami-
noazepine in ref 7; 1H NMR δ 0.85 (t, 6H), 2.12 (s, 3H), 2.25 (s, 3H),
2.78 (m, 2H), 3.15 (m, 2H), 4.02 (d, 1H, J ) 6 Hz, H-3), 5.9 (s, 1H),
5.9 (dd, 1H, J ) 6, 7.5 Hz, H-6), 6.2 (dd, 1H, J ) 6, 9 Hz, H-5), 6.45
(dd, 1H, J ) 6, 9 Hz, H-4), 7.23 (d, J ) 7.5 Hz, H-7), attribution on
the basis of double-irradiation experiments. NOE experiments: irradia-
tion of the 4.02 d enhances both CH2 m of the NEt2 group and the
azepine protons, in particular H-5 and H-7; irradiation of the 7.23 d
enhances the 5.9 d and the 4.02 d, but not the CH2 m of the NEt2
group; irradiation of either the 2.78 or the 3.15 m enhances the 4.02 d;
irradiation of the 2.25 s enhances the 4.02 d. The structure was finally
demonstrated by means of a single-crystal determination (data deposited
at the Cambridge Crystallographic Centre, no. 101778).
3H-2-(Diethyamino)-3-[1-(3,5-dimethylpyrazolyl)]-5-trifluoro-
methylazepine (9e): light yellow crystals, mp 75-78 °C (n-pentane).
Anal., found: C, 59.2; H, 6.5; N, 17.0. Calcd for C16H21F3N4: C, 58.89;
H, 6.44; N, 17.18. 1H NMR δ 0.9 (t, 6H), 2.15 (s, 3H), 2.3 (s, 3H), 2.8
(m, 2H), 3.22 (m, 2 H), 3.95 (d, 1H, J ) 6.5 Hz), 5.85 (s, 1H), 5.95 (d,
1H, J ) 8 Hz), 6.7 (d, 1H, J ) 6.5 Hz), 7.35 (d, 1H, J ) 8 Hz). NOE
experiments: irradiation of the 3.95 d enhances the CH2 m at 3.22 as
well as the azepine protons, similarly to the above isomer; 13C NMR
δ 10.7 (Me), 12.4 (Me), 13.4 (Me), 43.5 (CH2), 51.3 (CH), 103.5 (CH),
106.4 (CH), 111.8 (CH), 123.7 (CF3), 128.9 (C-CF3), 140.3, 142.3
(CH), 143.0, 148.1.
Double-Irradiation Experiments. Samples irradiated at 254 nm as
above were further irradiated by means of a focalized high-pressure
mercury arc (Osram 150 W) through a cutoff filter (>455 nm). The
beam reached the cell through a side opening perpendicular to the
analyzing beam. The sample was then rapidly heated at room temper-
ature and analyzed as above. Under this condition, dihydroquinoxalines
8 often were important products in the place of their rearomatized
counterpart (compounds 7) generally dominating upon photolysis at
room temperature (except than for the dimethylamino derivative, which
gave 7g under any condition). 8a was independently prepared and
characterized. As for the other derivatives, these rearomatized spontane-
ously in ethanol solution at room temperature in ∼1 month.7 In Tables
1 and 3, the total amount 7 + 8 is reported.
5H-2-(Diethyamino)-3-[1-(3,5-dimethylpyrazolyl)]-5-trifluoro-
methylazepine (10): obtained in a chromatographic fraction admixed
with some (20%) of the above isomer. 1H NMR δ 0.9 (t, 6H), 2.92 (m,
1H), 3.2 (m, 4H), 5.05 (dd, 1H, J ) 5, 7.5 Hz), 5.8 (s, 1H), 6.13 (d,
1H, J ) 7.5 Hz), 6.87 (d, 1H, J ) 5 Hz); 13C NMR δ 12.4 (Me), 13.2
(Me), 13.5 (Me), 40.3 (CH-CF3), 43.5 (CH2), 104.6 (CH), 106.4 (CH),
125.0 (CF3), 125.1 (CH), 128.9 (C-CF3), 140.5, 140.3 (CH), 149.3,
158.5.
1-[2-Amino-3-(diethylamino)-5-(dimethylamino)phenyl]-3,5-di-
methylpyrazole (11): yellow oil (purified by repeated chromatography).
Anal., found: C, 67.3; H, 9.3; N, 23.0. Calcd for C17H27N5: C, 67.73;
H, 9.03; N, 23.24. Mass spectrum, M+ 301 m/e; UV (EtOH) 330 nm
(log ꢀ 3.58), 238 (4.24); IR (neat) 3420, 3332 cm-1; 1H NMR δ 1.0 (t,
Acknowledgment. Partial support of this work by CNR,
Rome, and MURST, Rome, is gratefully aknowledged.
JA982585R