Reactivity of singlet and triplet arylnitrenes: Temperature-dependent photodecomposition of 1-(2-azidophenyl)-3,5-dimethylpyrazole

Article abstract of DOI:10.1021/ja970588i

At >200 K photolysis of 1-(2-azidophenyl)-3,5-dimethylpyrazole (5) gives 1,3-dimethylpyrazolobenzotriazole (6, by electrophilic cyclization of singlet nitrene 4) or, in the presence of diethylamine, aminoazepine 8 (by addition of the nucleophile and rearr

Full text of DOI:10.1021/ja970588i

7308
J. Am. Chem. Soc. 1997, 119, 7308-7315
Reactivity of Singlet and Triplet Arylnitrenes:
Temperature-Dependent Photodecomposition of
1-(2-Azidophenyl)-3,5-dimethylpyrazole
Angelo Albini, Gianfranco Bettinetti,*,† and Giovanna Minoli
Contribution from the Department of Organic Chemistry, The UniVersity of PaVia,
Viale Taramelli 10, 27100 PaVia, Italy
ReceiVed February 24, 1997X
Abstract: At >200 K photolysis of 1-(2-azidophenyl)-3,5-dimethylpyrazole (5) gives 1,3-dimethylpyrazolobenzo-
triazole (6, by electrophilic cyclization of singlet nitrene 14) or, in the presence of diethylamine, aminoazepine 8 (by
addition of the nucleophile and rearrangement). At lower temperatures, the yield of these products decreases and
3
the azo derivative 9 (from the dimerization of triplet nitrene 4) as well as products from intramolecular radical
cyclization (again via 34) is obtained, to become the only products at <100 K. Differential thermodynamic parameters
1
3
for the reactions of 4 and 4 are determined by analysis of the temperature dependence of products distribution
(∆∆Hq ) -10 kJ mol-1 and ∆∆Sq ) 34 J mol-1 K-1 in ethanol). Addition of the amine may require previous ring
enlargement to give the dehydroazepine 12; there is no indication that this is an intermediate of any stability, however,
3
and if formed, this is in equilibrium with 14. Triplet nitrene 4 is characterized in matrix by UV spectroscopy, and
its photoreactions (to give mainly intramolecular hydrogen abstraction) are separately studied.
There is now current agreement that the photolysis of phenyl
Scheme 1
azide (1) in solution generates singlet nitrene (12) and this
partitions between intersystem crossing to the triplet 32 and ring
enlargement to the dehydroazepine 3 (Scheme 1).1-5 Irradiation
of a solution of 1 at room temperature gives some azobenzene,
in a yield depending on the initial phenyl azide concentration,
and untractable “tars” arising via a polymerization process
thought to be initiated by a reaction of 3. The latter intermediate
can be trapped by nucleophiles, typically amines, to give isolable
azepines in a moderate yield. However, lowering the temper-
ature slows down the reaction with amines and below 160 K
the main product formed is azobenzene, arising from triplet
nitrene. A brief irradiation in a glassy matrix at 77 K yields
3
been estimated,4 it may be disappointing that the data obtained
are based on a reaction of the dehydroazepine and not on a
direct reaction of singlet nitrene, which is the primary photo-
product and the temperature-dependent branching point. On
other hand, no direct reaction of singlet phenylnitrene has been
reported, except for nucleophilic addition by amines to give a
hydrazine in the case of 4-nitro and 4-cyanophenyl azide.9
We reasoned that a further improvement of the model could
be obtained by resorting to intramolecular trapping, and
particularly studying the (nitrenophenyl)dimethylpyrazole 4.
This species was originally proposed by Suschitzky and
colleagues as a probe for testing the nucleophilic/radical
character of phenylnitrene through addition to the pyrazole
nitrogen and hydrogen abstraction from the methyl group,
respectively (Scheme 2).10 We found that, with some modifica-
tions, this model is satisfactory. Noteworthy, photolysis of azide
5 yields the heteropentalene 6 (see below), the product from
intramolecular trapping of singlet nitrene, in a good yield at
room temperature.11,12 This is an advantage with respect to
parent 1 and its simple derivatives and gives confidence that a
temperature-dependence study may offer new mechanistic
predominantly 2, recognized by its UV and EPR spectra, but
continuous photolysis leads to 3.4,6 IR detection has been
particularly valuable, and after some initial conflict it is now
clear that 2 is the main product also in Ar matrix4,5,7,8 and 3
3
has been revealed at room temperature in solution.2c
While the mechanism shown in Scheme 1 can be regarded
as fully satisfactory and the pertinent kinetic parameters have
† Fax: 39-382-507323. E-mail: minoli@chifis.unipv.it.
X Abstract published in AdVance ACS Abstracts, July 1, 1997.
(1) (a) Schuster, G. B.; Platz, M. S. AdV. Photochem. 1992, 17, 69. (b)
Platz, M. S.; Leyva, E.; Haider, K. Org. Photochem. 1991, 11, 367. (c)
Platz, M. S. Acc. Chem. Res. 1995, 28, 487 and references cited.
(2) (a) Li, Y. Z.; Kirby, J. P.; George, M. W.; Poliakoff, M.; Schuster,
G. B. J. Am. Chem. Soc. 1988, 110, 8092. (b) Schrock, A. K.; Schuster, G.
B. J. Am. Chem. Soc. 1984, 106, 5228. (c) Li, Y. Z.; Kirby, J. P.; George,
M. W.; Poliakoff, M.; Schuster, G. B. J. Am. Chem. Soc. 1988, 110, 8092.
(3) Meijer, E. W. Nijhuis, S.; Von Vroonhoven, F. C. B. M. J. Am. Chem.
Soc. 1988, 110, 7209.
(4) Leyva, E.; Platz, M. S.; Persy, G.; Wirz, J. J. Am. Chem. Soc. 1986,
108, 3783.
(5) Marcinek, A.; Leyva, E.; Whitt, D.; Platz, M. S. J. Am. Chem. Soc.
1993, 115, 8609.
(6) Reiser, A.; Bowes, G.; Horne, R. J. Trans. Faraday Soc. 1966, 62,
3162.
(7) Hayes, J. C.; Sheridan, R. S. J. Am. Chem. Soc. 1990, 112, 5879.
(8) (a) Chapman, O. L.; Le Roux, J. P. J. Am. Chem. Soc. 1978, 100,
282. (b) Chapman, O. L. Sheridan, R. S. J. Am. Chem. Soc. 1979, 101,
3690. (c) Dunkin,I. R.; Donelly, T.; Lockhart, T. S. Tetrahedron Lett. 1985,
359. (d) 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,
2147.
(9) (a) Liang, T. Y.; Schuster, G. B. J. Am. Chem. Soc. 1987, 109, 7804.
(b) Odum, R. A.; Aaronson, A. M. J. Am. Chem. Soc. 1969, 91, 5680.
(10) Lindley, J. M.; McRobbie, I. M.; Meth-Cohn, O.; Suschitzky, H. J.
Chem. Soc., Perkin Trans. 1 1980, 982.
(11) Albini, A.; Bettinetti, G.; Minoli, G. Chem. Lett. 1981, 301.
(12) Albini, A.; Bettinetti, G.; Minoli, G. Heterocycl. 1995, 597.
S0002-7863(97)00588-X CCC: $14.00 © 1997 American Chemical Society

ReactiVity of Singlet and Triplet Arylnitrenes
J. Am. Chem. Soc., Vol. 119, No. 31, 1997 7309
Table 2. Products from the Variable Temperature Irradiation of
Azide 5 in Ethanol (HPLC Monitoring)a
Scheme 2
in neat solvent
with 0.1 M DEA
T/K
6
7
9
10 + 11
6
7
8
9
10 + 11
297 72 trc
270 85 tr
230 82 tr
180 60 tr
140 27 tr
36
48
64
65
29
5
4
4
3
3
5
7
22
23
23
18
4
b
b
b
5
15
27
2
1
b
b
b
6
27
42
100
2
4
13
6
Scheme 3
a 8 min irradiations; conversion is ca. 90%. See Figure 1 for data at
more temperatures. b Low amount, difficult to quantify due to peak
superimposition in the HPLC trace. c tr ) trace.
Table 1. Products from the Preparative Irradiation of Azide 5
(Isolated Yields from Column Chromatography)
Figure 1. Temperature dependence of the photoproducts yield by
irradiation of azide 5 in EtOH, (b) product 6, (2) product 9, as well
as in EtOH containing 0.1 M DEA, (O) product 6, (0) product 8.
solvent (additive)
products (% yield)
EtOHa
6 (66), 7 (9)
6 (58)
MeCNa
a
C6H12
6 (53), 7 (1)
Table 3. Products from the Variable Temperature Irradiation of
Azide 5 in MPa
C6H12, Et2NH 0.06 M
6 (7), 7 (tr),b 8 (51)
b
a See ref 17. tr ) trace.
in neat solvent
with 0.1 M DEA
T/K
6
7
9
10 + 11
6
7
8
9
10 + 11
insight, since a satisfactory material balance is expected under
all conditions. We found that this is indeed the case, as reported
in the following.
295
56
trf
tr
2
8
11
3
46
tr
295b 54
240
200
160
90
59
33
5
4
3
15
13
3
11
11
7
c
c
c
42
39
19
2
Results
9d tr
c
11
19
21e
4
5
tr
Room Temperature Photolysis. Irradiation of azide 5 at
room temperature gave the heteropentalene 6 in good yield along
with a minor amount of the amine 7 in a variety of solvents
(see Scheme 3; the scheme gathers the products formed under
various conditions, see below). Typical results from preparative
irradiation and chromatographic separation are reported in Table
1. As it appears from the table, the reaction is little affected
by the medium. In the presence of amines, the yield of 6 was
decreased and the aminoazepine 8 was obtained as a major
product.
Temperature-Dependent Photolysis. Experiments in EtOH.
A set of small-scale experiments was carried out at different
temperatures. 2-Methylpentane (MP) and ethanol were chosen
as suitable solvents for this study. The product distribution was
determined by standardized HPLC measurements. Compound
6 is sensitive to irradiation in the presence of oxygen,13 and
thus the experiments were routinely carried out in degassed
solutions. The amount of azide decomposed did not signifi-
cantly change on decreasing the temperature, but the yield of
the products was deeply affected. Thus, in EtOH the azo
derivative 9, not significantly formed at room temperature, was
detected below 230 K, was found in equal amounts to 6 at 140
77
7
14
10
77b
a 8 min irradiations; conversion is ca. 90%. See Figure 2 for data at
more temperatures. b In oxygen-equilibrated solution. c Low amount,
difficult to quantify due to peak superimposition in the HPLC trace.
d Further products (M+ + 1, 371) revealed by HPLC; see text. e 3,5-
Dimethyl-1-(2′-nitrosophenyl)pyrazole and the corresponding nitro
derivative also present. f tr ) trace.
K, and became the main product at lower temperatures when 6
was present only in trace amounts (Table 2, Figure 1). Apart
from a small amount of the amine 7, further products formed
at low temperature were 2-methylpyrazolo[1,2-a]quinoxaline
(10) and its dihydro derivative 11 (see Experimental Section
for the characterization). In the presence of 0.1 M diethylamine
(DEA) the azepine 8 was formed competitively with hetero-
pentalene 6 at 150 K, but not at lower temperatures, where the
azo compound 9 remained the main product, with qualitatively
the same product distribution as in neat EtOH.
Experiments in MP. The behavior in MP was roughly
similar, but more complex. In this case, the yield of product 6
decreased in the 220-150 K range, more abruptly than in the
EtOH case (Table 3, Figure 2), in a temperature range where
compound 9 was still unimportant. The formation of further
peaks was chromatographically revealed in this range, and
(13) Albini, A.; Bettinetti, G.; Minoli, G. J. Org. Chem. 1983, 48, 1080.
Albini, A.; Bettinetti, G.; Minoli, G. J. Chem. Soc., Perkin Trans. 1 1983,
581.

7310 J. Am. Chem. Soc., Vol. 119, No. 31, 1997
Albini et al.
Discussion
The availability of intramolecular reaction paths for (singlet
and triplet) nitrene 4 results in a better chemical yield than that
from parent 2. Analysis of product distribution and of the
temperature effect on it, as well as spectroscopic evidence, allow
a detailed discussion of the mechanism.
Identification of the Intermediate in a Glassy Matrix and
Chemistry of Triplet Nitrene. It is convenient to start the
discussion from the experiments in glassy matrix. The spectra
shown in Figures 3 and 4 are closely related to those previously
reported for triplet phenylnitrene,4,6 and accordingly the present
species is identified as nitrene 34. The spectrum of the nitrene
has a minimum in the 250-280 nm region, where the azide
has a maximum (see Figure 3b). Due to the efficient photo-
decomposition of the latter, irradiation for a few minutes at 254
nm leads to its full consumption with negligible further reaction
of the primarily formed triplet nitrene (vide infra). Therefore,
the spectrum in Figure 3a does not require deduction for
unreacted azide and can be regarded as that of the nitrene, free
also from other contaminants (see the chemical evidence below).
It shows several well-resolved bands. In particular, the lowest
energy absorption is further shifted to the red with respect to
parent 32 (λoo 560 nm in the present case, ca. 500 experimental,4
540 nm calculated14 in the latter one). The entire spectrum is
similar but more intense in the present case, λmax 510 nm (log
ꢀ 3.3), 339 (3.6), 315 (3.8), 305 (3.9), 231 (4.3) in EtOH, 542
(3.3), 345 (3.72), 307 (4), 236 (4.5) in MP. Actually, the
spectrum of 34 is nearly identical to that of triplet (2,6-
dimethylphenyl)nitrene.4 Thus, the pyrazole ring, expected to
be noncoplanar with the phenyl ring due to steric hindrance,15a
exerts only a moderate effect and the electronic structure of the
Figure 2. Temperature dependence of the photoproducts yield by
irradiation of azide 5 in MP, (b) product 6, (2) product 9, as well as
in MP containing 0.1 M DEA, (O) product 6, (0) product 8.
although we were unable to isolate them, HPLC/MS analysis
demonstrated that the molecular weight was m/e 370 (twice that
of the nitrene 4). In turn, the yield of these products decreased
by further lowering the temperature, and below 130 K the main
product was, as in the EtOH case, the azo derivative 9 along
with compounds 7, 10, and 11. Notice finally that below 100
K the last two derivatives increased at the expense of 9.
In the presence of 0.1 M DEA the azepine 8 was formed
competitively with 6, and indeed in a higher proportion than in
EtOH. Lowering the temperature at 220 K and below decreased
the yields of both 6 and 8, but this change could be explored
only down to 160 K, since then the amine separated out.
Irradiation at 77 K (apparently involving an inhomogeneous
glass) formed no azepine and gave products 7, 9, 10, and 11 in
yields similar to those obtained in a neat MP glass.
3
present considered nitrene is quite similar to that of parent 2.
The spectrum undergoes a modest solvent effect, with a slight
red shift in MP with respect to EtOH (compare Figures 3a and
4).
The effect of oxygen was tested in a few cases. At room
temperature the effect was small, while a 77 K in MP 1-(2-
nitrosophenyl)- and 1-(2-nitrophenyl)-3,5-dimethylpyrazole were
formed (see Table 3).
The persistence of the spectrum in the glass and its bleaching
upon melting are consistent with what previously observed with
3
parent 2. Heating of the EtOH glass from 90 to 105 K (i.e.,
below the melting point) induces some modification of both
the visible band and of the groups of bands in the 290-400
nm region (see Figure 3c). This is reasonably due either to a
temperature-dependent complexation with the nucleophilic
solvent (see below) or to a conformational equilibrium.
The chemistry occurring upon melting the glass (the glassy
solution softens at ca. 95 K for MP and 120 K for EtOH) is
Irradiation in a Glassy Matrix. Irradiation at 90 K in glassy
EtOH by using a 254 nm light source led to the formation of a
colored intermediate (Figure 3a). After a few minutes of
irradiation, virtually complete consumption of the azide was
obtained (compare initial and final spectra in Figure 3b). The
spectrum did not change for hours at 90 K. Heating up to 105
K caused some change in the spectrum (see Figure3c), and at
a higher temperature the spectrum bleached upon melting the
glass. A similar spectrum was obtained in MP at 77 K (Figure
4). At ca. 90 K the MP glass became less rigid and the
intermediate accumulated to a lesser degree. On the other hand,
when degassing was omitted, a different spectrum was observed
(see Figure 5).
Secondary Photolysis. Prolonged irradiation at 254 nm led
to disappearance of the above spectrum. More expediently, a
clean bleaching of that spectrum could be obtained by irradiation
into the visible band. Taking advantage of this property,
solutions of 5 were first irradiated at 254 nm until a complete
conversion of the azide was reached and then at λ > 455 nm
until the visible absorption was bleached. Chemical analysis
after this double irradiation showed that compounds 10 and 11
were the main products in MP as well as in EtOH, where,
however, a sizeable amount of heteropentalene 6 was also
obtained (Table 4). The product distribution was little affected
by the temperature (in the range 110-90 K) in EtOH and
somewhat more in MP (in the range 100-77 K). The presence
of DEA did not greatly change the result, in particular did not
gave the azepine 8, except in a 2% yield in MP.
3
fully compatible with 4 being the only species present under
these conditions, since all products arise from triplet nitrene,
with no significant amount of those arising from singlet nitrene
or the deydroazepine 12 (see below). The main reactions are
dimerization to yield an azo compound (9) as well as intramo-
lecular hydrogen abstraction, here monitored by the formation
of products 10 and 11, the latter arising from oxidation of the
former one, and to a minor degree by intermolecular hydrogen
abstraction to give the amine 7. Previous experiments had
shown that these products are formed upon triplet-sensitised
photodecomposition of azide 5.11,12 At 90-100 K in EtOH
hydrogen abstraction is largely the minor process, in accordance
with the poor radical character of triplet arylnitrenes previously
documented for the present nitrene12 as well as for the general
case,1 although the ratio (10 + 11)/9 is larger in the experiments
in MP.
Temperature Dependence of the Reaction. Singlet-
Triplet Nitrene Competition. In a fluid solvent, triplet nitrene
obviously does not accumulate upon steady state irradiation.
(14) Kim, S. J.; Hamilton, T. P.; Schaefer, H. F. J. Am. Chem. Soc. 1992,
114, 5349.

ReactiVity of Singlet and Triplet Arylnitrenes
J. Am. Chem. Soc., Vol. 119, No. 31, 1997 7311
Figure 4. Spectrum (A vs λ, nm) obtained by irradiating 6.15 × 10-5
M azide 5 in MP at 77 K for 8 min.
Figure 5. Spectrum (A vs λ, nm) obtained by irradiating an air
equilibrated 1 × 10-4 M azide 5 in EtOH at 100 K for 8 min.
Table 4. Products from Double Irradiation Experimentsa
solvent
EtOH
T/K
6
7
8
9
trb
tr
10 + 11
110
100
90
100
100
90
22
20
12
17
6
3
5
6
2
10
5
41
48
54
20
15
20
64
58
EtOH/0.1 M DEA
MP
tr
2
tr
20
19
tr
tr
77
77
2
7
MP/0.1 M DEA
2
2
a Glassy solutions of azide 5 were first irradiated for 8 min at 254
nm and then for 10 min at λ > 455 nm. b tr ) trace.
However, the intermediacy of triplet nitrene is recognized from
product distribution15b and is significant when a solution of the
azide 5 is irradiated at a sufficiently low temperature. Indeed,
enhancing the temperature induces a regular decrease of the
(15) (a) Crystallographic analysis on azide 5 shows that the pyrazole is
tilted out of the phenyl ring plane; two independent molecules in the cell,
torsion angles 65.77 and 83.30°, respectively (B. Bovio, Pavia, personal
communication). Molecular model calculation reproduce a similar result
both from the azide and the nitrene. (b) The glassy solution softens at a
relatively low temperature (95 K for MP and 120 K for EtOH). This makes
the observed product distribution significant for the irradiation temperature,
since the intermediates are directly converted to the final products in the
experiments at these temperatures or immediately upon softening of the
glass for the few experiments below them.
Figure 3. (a) Spectrum (A vs λ, nm) obtained by irradiating 1 × 10-4
M azide 5 in EtOH at 90 K for 8 min. (b) Evolution of the absorption
spectrum by irradiation of 1 × 10-4 M azide 5 in EtOH at 100 K.
Spectra were registered at 2 min intervals (a, initial spectrum, e spectrum
after 8 min). (c) Spectrum obtained by irradiating 1 × 10-4 M azide 5
in EtOH at 90 K for 8 min (a) and spectra observed by heating the
solution at 95 (curve b) and 105 K (curve c).

7312 J. Am. Chem. Soc., Vol. 119, No. 31, 1997
Albini et al.
Scheme 4
Figure 6. Arrhenius plot for the ratio of the singlet nitrene derived
product 6 vs the triplet derived products 7, 9, 10, and 11, (O) in EtOH,
(b) in MP.
triplet-derived products (7, 9-11, see Tables 2 and 3, Figures
1 and 2) over the whole range explored. In EtOH, this is
accompanied by a regular increase of the singlet nitrene derived
product, heterocycle 6, and the overall yield of identified
products remains high and close to constant throughout the range
considered, from the glass to room temperature. In MP, on the
other hand, the yield of 6 increases more slowly with temper-
ature, and in the 150-230 K range some (two main) further
products are formed. These have a molecular weight corre-
sponding to a nitrene dimer. There is previous literature
evidence for photoproducts with molecular weights twice that
of the nitrene. These are the ring-cleaved products (unsatured
iminonitriles) obtained from the photolysis of quinoxalyl16a and
phenazinyl azides.16b These have been rationalized as arising
from nitrene-azide addition. Analogy suggests that products
of structure 13 should be the products formed under this
condition (see the mass spectra in the Experimental Section).
Amine Trapping and Temperature Effect on It. In the
presence of DEA the aminoazepine 8 is formed, and the
combined yields of 6 + 8 are close to the yield of 6 alone in
the absence of DEA (see Figure 1). The temperature depen-
dence of the yield of azepine 8 follows a pattern similar to that
seen above for 6. At a very low temperature (<150 K in EtOH)
addition of the amine has little influence on the yield of triplet
nitrene derived products; in MP the examination cannot be
extended below 160 K due to phase separation. At relatively
high temperatures the azepine 8 is formed and is obtained in a
larger amount in MP than in EtOH. The yield of the azepine
in EtOH decreases more rapidly with decreasing temperature
than that of heteropentalene 6, but the effect is relatively
small: the ratio 8/6 decreases from 0.6 to 0.14 from 297 to
140 K. The trend is qualitatively similar in the MP/DEA
experiments above 160 K, except that, as mentioned above, the
8/6 ratio is here much higher (decreasing from 4.2 to 2.7 in the
295-160 K range) and that examination at a lower temperature
is precluded by nonmiscibility (see Figure 2).
Figure 7. Eyring plot for the ratio of the singlet nitrene derived product
6 vs the triplet derived products 7, 9, 10, and 11 in EtOH.
supported by the demonstration that the same type of chemistry
is observed by generating the nitrene both from the azide and
from other precursors, as reported both for the present azide
(5)17 and for parent 1.5 14 is a branching point in the process
and has available two irreversible reactions: intramolecular
electrophilic cyclization to give 6 and intersystem crossing to
give 34. The former process predominates at a high temperature,
the latter one below ca. 140 K. A reversible rearrangement
between 14 and dehydroazepine is also indicated, justifying the
fact that, in the presence of DEA, azepine 8 is formed and at
all temperatures explored the sum 6 + 8 remains close to the
yield of 6 alone in the absence of DEA.
Differential Thermodynamic Parameters. At least in
ethanol, a reasonable material balance is obtained over all the
temperature domain explored, 297-90 K. Furthermore, viscos-
ity changes do not affect the result, since unimolecular processes
are compared. Thus, differential thermodynamic parameters for
the reactions of 14 can be obtained. Arrhenius analysis applied
to the ratio of the singlet nitrene derived product 6 vs the
summation of the triplet-derived products (7 + 9 + 10 + 11)
gives a linear plot (Figure 6). This yields ∆Ea ) -10.5 ( 2
kJ mol-1 for the above competing process. Similarly, Eyring
analysis gives ∆∆Hq ) -10 ( 2 kJ mol-1 and ∆∆Sq ) 34 (
3 J mol-1 K-1 (Figure 7). Assuming that intersystem crossing
has negligible activation enthalpy, the observed quantity (10
kJ mol-1) is directly the activation enthalpy of intramolecular
Mechanistic Scheme. The previous evidence is rationalized
through the mechanism depicted in Scheme 4. This is an
adapted version of Scheme 1, with the difference that intramo-
lecular trapping of singlet nitrene is also taken into account.
As with related phenyl azides, photodecomposition proceeds
from singlet excited azide (sensitised experiments give largely
different results)11,12 and its quantum yield does not change with
1
temperature. There is no doubt that singlet nitrene 4 is the
first-formed intermediate. This had been previously also
(16) (6) Hayes, R.; Schofield, J. M.; Smalley, R. C.; Scopes, D. I. C.
Tetrahedron 1990, 46, 2089. (b) Albini, A.; Bettinetti, G.; Minoli, G. J.
Org. Chem. 1991, 56, 6993.
(17) Albini, A.; Bettinetti, G.; Minoli, G. J. Am. Chem. Soc. 1991, 113,
6928.

ReactiVity of Singlet and Triplet Arylnitrenes
J. Am. Chem. Soc., Vol. 119, No. 31, 1997 7313
Scheme 5
and in view of the similar structure is expected to be very similar
to that of parent 2.
In the present case, two intramolecular processes are com-
3
pared, cyclization to 6 and ISC to 4. Previous studies had
compared ISC with amine addition, supposed to occur with
negligible Ea after rate-determining cyclization (again an in-
tramolecular electrophilic attack) to a dehydroazepine1,4,5 (in
our case 12). The equivalence of the combined 6 + 8 yield to
that of 6 alone in the absence of the amine shows that if 12 is
an intermediate, it is in equilibrium with 14. The thermodynamic
parameters for reactions attributed to the two species are very
Figure 8. Arrhenius plot for the ratio of the azepine 8 vs the singlet
nitrene derived product 6, (O) in EtOH), (b) in MP.
ring closure to give 6. As expected, this is small. It compares
with the previously estimated activation energy for the singlet
phenylnitrene-dehydroazepine rearrangement, 12 ( 4 kJ mol-1
(also evaluated as a differential parameter in comparison with
intersystem crossing to the triplet), but it refers in this case to
a chemical reaction directly from singlet nitrene. This enables
determination of the differential entropy, which is favorable to
the singlet cyclization. Analysis of the data in MP is less
straightforward, since the material balance of isolated products
is low in the intermediate temperature range. A rough estimate
can be obtained by using the high- and low-temperature points
(Figure 6). The slope is somewhat smaller than that in EtOH,
1
similar. If cyclization of 4 to 12 is the rate-determining step
for amine addition, then such a process has Ea ) 3 kJ mol-1
which is a really low value.
,
Alternatively, it may be held that the dehydroazepine is not
formed as a kinetic distinct intermediate and amine addition
occurs concertedly with ring enlargement (see Scheme 5),
although there is no precedent for such a mechanism. Notice
also that the molecular vibration involved in the conversion of
14 to both 6 and 12 (or directly to 8) is the same, in plane
bending of the C-N bond (Scheme 5).
with ∆Ea ≈ -8 kJ mol-1
.
As seen above, the aminoazepine vs heteropentalene ratio
depends on the solvent but changes little with the temperature.
Figure 8 shows that the two processes encounter very similar
barriers, ∆Ea ) -3 kJ mol-1 in EtOH and -1 kJ mol-1 in MP.
Thus, intermolecular reaction with a nucleophile encounters a
barrier only slightly higher than intramolecular cyclization onto
a nucleophilic center.
1
The activation energy measured4 for amine addition to 2 is
1
only slightly larger than that for 4, thus the same mechanism
should hold also for the parent phenylnitrene, and also in that
case the dehydroazepine is formed reversibly, if at all. Such
intermediates are well characterized in a matrix at a low
temperature, where they are formed by secondary photoreaction
of the primarily formed triplet nitrene. Furthermore, they have
been detected at room temperature by flash photolysis with IR
detection.2c However, the present data suggest that such
transients are not necessarily an intermediate on the path to the
azepine, or at any rate that no significant minimum is associated
to the dehydroazepine configuration in fluid solution.
Detailed Mechanism. The mechanism can now be discussed
in more detail, taking also in account the difference between
the two solvents explored. Ethanol, as the EPA mixture and
methyltetrahydrofuran used in previous studies,4,5 offers some
stabilization to the electrophilic nitrene, whereas this is not
possible with MP. As seen above, in EtOH there is a smooth
transition in the 140-200 K range between triplet and singlet
nitrene chemistry, and a uniformly good material balance is
maintained near the turning interval as well as in all the ranges
studied. Adding the good nucleophile DEA simply involves
partial substitution of the azepine 8 for the singlet nitrene product
6.
When generated in MP, singlet nitrene enjoys no solvent
stabilization. Such a stabilization is small, however, and no
major change intervenes in the mechanism. In particular, the
activation energy for cyclization to 6 does not increase (indeed,
from the limited data available, its difference from that of ISC
to 34 rather decreases). However, intermolecular reactions are
here more important (see below), showing that the advantage
for intramolecular attack observed in EtOH is in part due to a
weak bond with the slightly nucleophilic solvent which main-
tains the nitrene in the correct configuration for cyclization but
does not compete for nucleophilic addition. In accordance with
this, 6 is quantitatively formed at relatively high temperatures
(above ca. 220 K). However, the above evidence suggests that
below that point even a weak nucleophile such as azide 5 adds
(giving 13). In the presence of DEA the ratio 8/6 is much higher
than that in EtOH, and trapping continues also below 220 K,
where the intramolecular reaction is too slow, until the point is
reached where, as in EtOH, ISC to 34 takes over. Again Figure
6 shows that ∆Ea for formation of 6 and 8 is minimal, and thus
there is no evidence for 12 being a kinetically distinguished
intermediate in this temperature range. As for the chemistry
Thus, intersystem crossing from 14 to the strongly stabilized
triplet state 34 (the energy difference between parent singlet and
triplet 2 has been recently estimated as ca. 70 kJ mol-1 14,18a,b
;
previous estimates have given a much lower value, 17 kJ
mol-1 18c,d
involves probably negligible activation enthalpy, but
)
it is kinetically disfavored, with a 34 J mol-1 K-1 difference in
the entropic term. There is previous evidence that ISC from 12
3
to 2 is relatively slow (106-107 s-1) due to the large energy
gap, spatial differentiation of the singly occupied MOs, and poor
contribution from spin-orbit coupling.1c The small entropic
barrier we measure for 4 is reasonably explained on this basis
(18) (a) Hrovat, D. A.; Waali, E. E.; Borden, W. T. J. Am. Chem. Soc.
1992, 114, 8698. (b) Travers, M. J.; Cowles, D. C.; Clifford, E. P.; Ellison,
G. B. J. Am. Chem. Soc. 1992, 114, 8699. (c) Drzaic, P. S.; Brauman, J. I.
J. Phys. Chem. 1984, 88, 5285. (d) Drzaic, P. S.; Brauman, J. I. J. Am.
Chem. Soc. 1984, 106, 3443.
3
of 4, this is little affected by the solvent, in accordance with

7314 J. Am. Chem. Soc., Vol. 119, No. 31, 1997
Albini et al.
Scheme 6
the well-known poor radical character of arylnitrene triplet. Thus
hydrogen absorption from the solvent is ineffective over all the
temperature ranges where the triplet is the key intermediate.
Reactions of Triplet Nitrene with Oxygen. As pointed out
3
above, triplet nitrene 4 is virtually the only product obtained
by irradiation of the azide 5 in a glassy matrix. Noteworthy,
when degassing is omitted the transient spectrum observed is
completely different (compare Figure 5 with Figures 3 and 4).
This is reasonably attributed to oxygen addition to the nitrene
to yield a diradical or an oxygen-centered radical (see, e.g.,
formula 14, see Scheme 6). Similar intermediates have been
previously suggested for explaining the formation of nitro,
nitroso, and azoxy derivatives from the decomposition of some
aromatic and heterocyclic azides in the presence of oxygen.2c,19,20
Indeed, irradiation in oxygen-equilibrated MP gave some nitroso
derivative (15) with a minor amount of nitro compound 16,
along with products 10 and 11, while only 6 is formed at room
temperature (see Table 2). This shows that singlet nitrene is
unaffected by oxygen, while the triplet adds to it, and the nitroso
derivative is formed under this condition in place of the azo
compound. On the other hand, a complete temperature-de-
pendence study in the presence of oxygen was discouraged by
the previously observed sensitivity of compound 6 to radicals
and various oxidative conditions,13 which would have made it
difficult to obtain clear-cut conclusions.
Scheme 7
dependent and, although intramolecular radical attack remains
predominant, 12-22% cyclization to 6 occurs. On the other
hand, very little (e2%) azepine 8 is obtained by irradiation in
the glass in the presence of DEA in both solvents, although the
yield of 10 + 11 decreases. In contrast, some azepine has been
obtained from triplet phenyl azide.4,7 Thus, photoconversion
3
of 4 to 12 is unimportant, or at least is slower than the other
reactions of 34* under the present conditions. Obviously,
intramolecular radical attack (to give 10 and 11) is possible only
Photochemistry of the Triplet Nitrene. As mentioned
3
above, the spectrum of 4 (see Figures 3 and 4) is rather close
3
3
to that of parent 2, and the electronic structures of the two
with the present nitrene, not with 2, but this reaction at least
species are similar, since the noncoplanar pyrazole ring has only
a marginal effect. In particular, we observe a low-intensity,
long-wavelength band above 500 nm. A similar (somewhat blue
shows that ring enlargement occurs with a relatively low rate
from the excited nitrene triplet and that the iminyl radical
character of this species is revealed when a suitable model is
chosen.
3
shifted) band of 2 has been observed by Leyva et al. and
attributed to an nπ* or πn transition.4 Detailed calculations by
Kim et al. showed that the T1 state results from the promotion
of an electron from the highest lying π orbital to the nx orbital
(see Scheme 7), with a predicted adiabatic energy of 18 600
cm-1 (539 nm).14 This state is considerably puckered.
Conclusion. The discussion above has shown that the
structure of the nitrene 4 is similar to that of parent 2. Therefore
the conclusion presented here should be significant for phen-
ylnitrene and its simple derivatives. With respect to what was
previously known in the literature, this work documents directly
the competition between an electrophilic reaction of singlet
nitrene and intersystem crossing and shows that the former has
a differential activation enthalpy 10 kJ mol-1 lower than the
latter one, but is favored from the entropic point of view. It is
further recognized that a dehydroazepine is either not a
kinetically distinguished isomer or is in equilibrium with singlet
nitrene. It has been previously suggested that the lack of
intermolecular electrophilic reactions of singlet phenylnitrene
(as compared, for example, with the carbene) is due to the fast
(10-100 ps at room temperature) ring enlargement.1c It is now
shown that intramolecular electrophilic reaction is fast when
there is a convenient nucleophilic group, and on the other hand,
amine addition seems not to necessarily require previous
rearrangement (and indeed it appears difficult to prove participa-
tion of a dehydroazepine in a fluid solution, since the involved
reactions are too fast). It may be that azepines arise via attack
to the C atom R to the nitrene function by a nucleophile. Thus
the difference between singlet nitrene and carbene may be due
to an intrinsic difference of the rates of reaction with various
nucleophiles, which are significant with the latter species only
with polar nucleophiles. As for triplet nitrene, this is strongly
Visible irradiation of triplet nitrene in MP at 77 K caused
conversion of the main part of the nitrene to products 10 and
11, rather than to the azo compound obtained by thermal
dimerization of the nitrene (see Table 4). However, at a higher
temperature (90 or 100 K) more of the azo and, in the latter
case, also some of the cyclization product 6 were obtained. As
3
for the EtOH glass, irradiation of 4 in the 90-110 K range
gave none of the azo compound and a ca. 1 to 2 proportion of
compounds 6 and 10 + 11. Electronic excitation as shown in
Scheme 7 is expected to increase the radical character of the
nitrene which would now be more similar to an iminyl radical.
In the noncomplexing MP intramolecular radical hydrogen
abstraction to give 10 and 11 is the exclusive process at 77 K,
while at a higher temperature, where the glass is less rigid,
nitrene dimerization to give 9 and, to a small extent, electrocyclic
cyclization to 6 are also observed. In EtOH, some interaction
with the solvent occurs; the reaction is not very temperature-
(19) Abramovitch, R. A.; Challand, S. R. J. Chem. Soc., Chem. Commun.
1972, 964.
(20) Bettinetti, G. F.; Fasani, E.; Minoli, G.; Pietra, S. Gazz. Chim. Ital.
1979, 109, 175.

ReactiVity of Singlet and Triplet Arylnitrenes
J. Am. Chem. Soc., Vol. 119, No. 31, 1997 7315
ITC4 temperature controller and placed in a UV-vis Kontron Uvikon
941 spectrophotometer. The cell was irradiated from the bottom by
means of a bifilar low-pressury mercury arc inserted below the cell
(Helios Italquartz 15 W). Irradiation was discontinued when taking
the spectra. 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 solution was allowed to reach room temperature (within
30 min). 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 Supelcosil LC-
ABZ column (and a Supelguard LC-ABZ precolumn). The eluant was
a MeCN-water 1:1 mixture and the flux was 0.5 mL/min. The
analyzing wavelength was 290 nm, and the spectra of every peak were
routinely measured in the range 200-400 nm. All of the main peaks
corresponded to compounds isolated form preparative irradiations, and
the analysis was based on a comparison with authentic samples of the
photoproducts prepared as above (5, tR 14.8 min; 6, tR 11.7 min; 7, tR
10.4 min; 8, tR 18.4 min; 9, tR 23.4 min; 10, tR 12.8 min; 11, tR 12.3
min). The results are shown in Table 2 and Figure 1.
stabilized with respect to the carbene and shows a poor radical
reactivity, even in the intramolecular case.
A further part of this study takes advantage of the fact that
triplet nitrene is obtained in a clean way in a glassy solution.
This allows the study of the secondary photochemistry of this
species and reveals that its excited state behaves as an
electrophilic radical.
Experimental Section
General. 1H NMR spectra were run on a Brucker 300 instrument,
IR spectra on a Perkin-Elmer Paragon 1000 spectrophotometer, UV-
visible spectra on a Kontron Uvikon 941 spectrophotometer, and mass
spectra on a Finnigan LCQ instrument. Methylpentane and 95% ethanol
were spectroscopic grade solvents and used without further purification.
Column chromatography was performed with silica gel Merk HR 60.
Compounds 5, 6, 7, 9, and 10 were prepared as previously reported.17
2-(3,5-Dimethylpyrazolyl)-7-(diethylamino)-2H-azepine (8). An-
hydrous cyclohexane (300 mL) in an immersion well irradiation
apparatus was refluxed and cooled under argon. Azide 5 (300 mg,
4.66 × 10-3 M) and diethylamine (1.9 mL , 6 × 10-2 M) were added
and after a further 15 min of Ar flushing the solution was irradiated
by means of a 15 W low-pressure mercury arc while maintaining an
Ar stream. After 30 min the solvent was evaporated under reduced
pressure and the residue chromatographed on a silica gel column eluting
with a 7:3 cyclohexane-ethyl acetate mixture to give 135 mg (51%)
yield of the title compound, mp 55-56 °C (n-hexane); IR (Nujol) 1560,
It is important to note that in this instrumental setting not all the
contents of the cell was illuminated. Therefore, analysis at the end of
the irradiation time showed the presence of some residual azide (ca.
10%) even if the UV spectrum of the irradiated sample showed complete
consumption of this compound.
The experiments in MP were analogously carried out using 6.15 ×
10-5 M solutions (Table 3, Figure 2).
1
1508 cm-1; UV (MeCN) λ 301 nm (log ꢀ 3.89), 268 (sh, 3.75); H
The softening and melting temperatures of the solutions were
determined in parallel cryoscopic experiments. In this way it was
checked that the glass softens at ca. 95 K for the MP frozen solution
and at ca. 120 K for the EtOH frozen solution.
NMR δ 0.87 (t, 6H, J ) 7 Hz), 2.15 (s, 3H), 2.25 (s, 3H), 2.87 (m,
2H), 3.20 (m, 2H), 4.02 (dd, 1H, J ) 6.5, 1 Hz, H-2), 5.87 (s, 1H),
5.88 (ddt, 1H, J ) 7.5, 5.5, 1 Hz, H-5), 6.27 (dd, 1H, J ) 9, 6.5 Hz,
H-3), 6.46 (ddt, 1H, J ) 9, 5.5, 1 Hz, H-4), 7.25 (dt, 1H, J ) 7.5, 1
Hz, H-6).
Double Irradiation Experiments. Samples irradiated as above were
further irradiated by means of a focalized high-pressure mercury arc
(Osram 150 W) through a cutoff filter (λtr > 455 nm). The beam
reached the cell through a side opening perpendicular to the analyzing
beam.
2-Methyl-4,5-dihydropyrazolo[5,4-a]quinoxaline (11). A solution
of compound 10 (37 mg, 0.2 mmol) in 10 mL of ethanol containing 1
mg of Pd/C was hydrogenated at room temperature and pressure until
1 equiv of hydrogen was adsorbed. The mixture was separated on a
silica gel column eluting with a benzene-ethyl acetate 9:1 mixture
under argon to afford the title compound (15 mg) as well as some amine
7 and unreacted 10. 11: oil which solidifies on treating with n-hexane,
mp 69-70 °C; IR (Nujol) 3235 cm-1; UV (MeCN) λ 317 (log ꢀ 3.68),
267 (3.66); 1H NMR δ 2.35 (s, 3H), 3.9 (s, 1H, NH), 4.5 (s, 2H), 5.92
(s, 1H), 6.7 (dd, 1H, J ) 1.5, 7 Hz, H-6), 6.85 (dt, 1H, J ) 1.5, 7 Hz,
H-8), 6.98 (dt, 1H, J ) 1.5, 7 Hz, H-7), 7.27 (s, 1H, H-3), 7.76 (dd,
1H, J ) 1.5, 7 Hz, H-9).
HPLC MS Measurements. HPLC examination of the photolysate
in MP at a temperature in the range 230-150 K range showed the
presence of some peaks different form those reported above. HPLC
MS analysis for those peaks (tR 7.4 min, tR 8.6 min) showed an identical
mass fragmentation pattern [chemical ionization, m/z 371 (M+ + 1),
275, 198] which was consistent with the structure of isomeric unsatured
imino nitriles 13 (loss of pyrazole, 96, and of the unsaturated chain,
78).
Variable Temperature Irradiation. A 1 × 10-4 M solution of
azide 5 in EtOH (2 mL) in a 1 cm optical path spectrophotometric
quartz cell with a quartz to glass graded seal was degassed by means
of four freeze-degas-thaw cycles and sealed. This was inserted into
an Oxford DN 1704 liquid nitrogen cryostat fitted with a calibrated
Acknowledgment. Partial support of this work by MURST,
Rome, and CNR, Rome, is gratefully acknowleged.
JA970588I