Revisiting the thermal decomposition of five ortho-substituted phenyl azides by calorimetric techniques

Article abstract of DOI:10.1007/s10973-009-0572-8

The thermal decomposition (TD) of 2-azidophenylmethanol (1), 2-azidobenzenecarbaldehyde (2), 1-(2-azidophenyl)-1-ethanone (3), (2-azidophenyl)(phenyl)methanone (4) and 1-azido-2-nitrobenzene (5) was analysed by DSC, TG and C80 calorimetric techniques under both oxidative and non-oxidative conditions. The TD of these azides in solution is well known to give the corresponding benzoxazoles, generally in good yields, with the exception of azide 1. When both the outcomes from the solid phase and in 'solution phase' TD reactions combined with the results from EI-MS experiments were considered, sufficient information was available to estimate the azides intrinsic molecular reactivity (MIR).

Full text of DOI:10.1007/s10973-009-0572-8

J Therm Anal Calorim (2010) 100:191–198
DOI 10.1007/s10973-009-0572-8
Revisiting the thermal decomposition of five ortho-substituted
phenyl azides by calorimetric techniques
•
•
•
Paolo Cardillo Lucia Gigante Angelo Lunghi
Paolo Zanirato
Received: 11 March 2009 / Accepted: 16 October 2009 / Published online: 6 November 2009
´
´
ꢀ Akademiai Kiado, Budapest, Hungary 2009
Abstract The thermal decomposition (TD) of 2-azid-
ophenylmethanol (1), 2-azidobenzenecarbaldehyde (2),
1-(2-azidophenyl)-1-ethanone (3), (2-azidophenyl)(phenyl)
methanone (4) and 1-azido-2-nitrobenzene (5) was analysed
by DSC, TG and C80 calorimetric techniques under both
oxidative and non-oxidative conditions. The TD of these
azides in solution is well known to give the corresponding
benzoxazoles, generally in good yields, with the exception
of azide 1. When both the outcomes from the solid phase
and in ‘solution phase’ TD reactions combined with the
results from EI-MS experiments were considered, sufficient
information was available to estimate the azides intrinsic
molecular reactivity (MIR).
and nitrogen-containing heterocycles (azirines, aziridines,
triazoles, triazolines and azoles). Azides have several
defence and industrial applications, including use as pro-
pellants, explosives, polymer cross linkers, rubber vulca-
nisation agents, reactive dyes and blowing agents, and are
also found in biologically and pharmaceutically active
compounds [1–3]. Although most chemists recognise that
organic azides have extremely useful applications in
modern organic chemistry, very few reports mention their
potentially dangerous properties [4]. However, an over-
whelming interest in the explosive properties of metal
azides [5] and poly-azide organic compounds [6] has
developed during the last decade.
Compounds containing one (or more) azido groups,
especially in conjunction with other ‘explosophores’, are
known for their treacherous sensitivity to heat and/or shock
[4, 6–8]. Despite the apparent lack of shock sensitivity
warnings on its MSDS, the 4-azidobenzaldehyde is too
shock sensitive to be isolated in other than very small
quantities [9]. It is often prudent to perform thermal anal-
yses of such compounds using a differential scanning cal-
orimeter (DSC) and/or a thermal gravimeter (TG), as well
as C80 calorimetric analysis. Papers discussing the most
likely thermal decomposition (TD) hazards of aryl [10] and
heteroaryl [11, 12] azides have recently been published.
In contrast, the TD of simple organic azides under
‘controlled’ conditions is a synthetically fruitful reaction
that gives rise to a great variety of products. The outcomes
of TD reactions are generally dependent upon the medium,
as well as the substrate and its substituent(s). These factors
may direct the reaction’s mechanism toward a concerted
extrusion of nitrogen or toward the formation of a ‘free’
nitrene [13]. The intermediate nitrene can exist in either a
singlet or triplet (S or T) electronic ground state, where
the former is normally favoured by conjugated electron
Keywords DSC Á TGA-FTIR Á C80 techniques Á
Heat flux calorimetry Á Kinetic and thermodynamic
parameters Á MS (EI) spectroscopy Á Pericyclic
reactions and mechanism Á Phenyl azides Á Thermal
decomposition ion in ‘solution phase’ and ‘solid phase’
Introduction
Organic azides are a well-studied class of compounds that
are easily prepared and transformed into a variety of
functional groups (amines, triazenes, aza-ylides and iso-
cyanates), reactive intermediates (nitrenes, nitrenium ions)
P. Cardillo (&) Á L. Gigante Á A. Lunghi
Stazione Sperimentale per i Combustibili SSC,
Viale A. De Gasperi 3, 20097 San Donato Milanese, MI, Italy
e-mail: cardillo@ssc.it
P. Zanirato
Dip.to di Chimica Organica ‘A. Mangini’, Viale Risorgimento 4,
40136 Bologna, Italy
123

192
P. Cardillo et al.
The isolated solid organic azides 1–5 were handled using
standard techniques, while their purification and character-
isation were carried out by standard methods [25–29]. The
IR, 1H-NMR and mass spectroscopic data obtained for
azides 1–5 are consistent with their assigned chemical
structures. Chromatographic filtration was carried out on
‘Florisil’ BDH, 60-100 mesh, using petroleum ether (distil-
lation range 30–60 ꢁC) as the eluent: (1, mp 52–53 ꢁC [25];
2, mp 34–36 ꢁC [26]; 3, mp 22–23 ꢁC (bp. 121 ꢁC/17 mm)
[27]; 4, mp 36–37 ꢁC [28]; 5, mp 51–53 ꢁC [29].
NO2
N3
N3
R
N3
CH2OH
2
3
4
; R = COH
; R = COMe
; R = COPh
1
5
Scheme 1 o-Phenyl azides used in this study
withdrawing groups and the latter is typically favoured by
electron-donating groups on the phenyl azide [14–18].
The aim of this study is to disclose the hazardous nature,
if any [6–9], of 2-azidophenylmethanol (1) and the
o-oxophenylazides: 2-azidobenzenecarbaldehyde (2), 1-(2-
azidophenyl)-1-ethanone (3), (2-azidophenyl)(phenyl)meth-
anone (4) and 1-azido-2-nitrobenzene (5) (Scheme 1). The
kinetic and thermodynamic studies of the ‘controlled’ TD
reactions of the compounds 2–5 in solution have been pre-
viously reported [19–21]. These compounds were also
selected for study because re-examining their TD reactions
should shed further light on their relevance in the long
debated electrocyclic mechanism of benzoxazole formation.
These studies were performed using advanced calorimetric
and thermo-analytical methods (DSC, TG and C80) under
both oxidative and non-oxidative conditions, which were
coupled to analyses of the gasses emitted by TGA- and C80-
FTIR. In contrast, azide 1 was selected as a comparative
probe to assess the energy required for (presumably) non-
concerted TD processes.
Physical measurements and experimental conditions
Spectral characterisation of azides 1–5 [30]
The IR spectra of azides 1–5 were measured as neat films
using a Perkin-Elmer Spectrum 2000 FT-IR spectrometer.
1H-NMR and 13C-NMR spectra for all five compounds
were recorded on a Varian Gemini 300 and agree with the
data reported in the literature [31–35].
Mass spectra were recorded on VG7070E instruments
using an electron impact ionisation method (70 eV) at the
same pressure (10-7 atm) and temperature (27 ꢁC). The
melting points of azides 1–5 were devised from the endo-
thermic melting peak in the DSC’s thermal plot and were in
agreement with the previously reported values [24–29].
Thermal performances were measured relative to a ref-
erence sample in the DSC 820 and 823e Mettler Toledo
instruments with the following accuracy standards: (i)
5 ꢁC min-1 for the evaluation of decomposition enthalpy
under both oxidative and closed steel vessel non-oxidative
conditions from 30 to 300 ꢁC; (ii) 2, 5, 10 and 15 ꢁC min-1
for the evaluation of kinetic data from dynamic heating
under a nitrogen atmosphere in a closed glass vessel from
30 to 300 ꢁC, unless otherwise stated [36]. The best mea-
surements of self-heating rates were obtained using a C80
Setaram calorimeter from 30 to 300 ꢁC, to a scan rate (U)
of 0.3 ꢁC min-1 unless otherwise stated.
In order to predict and detect the early stage processes
involved in the TDs of azides 1–5, both theoretical calcu-
lations using the CHETAH (CHEmical Thermodynamic and
Hazard Evaluation) software package and electron impact
mass spectroscopy (EI-MS, 70 eV) were performed.
Experimental section
Materials and chemicals
The phenyl azides (1–5) used in this study were prepared as
follows: 2-Azidophenylmethanol (1), 1-(2-azidophenyl)-
1-ethanone (3), (2-azidophenyl)(phenyl)methanone (4) and
1-azido-2-nitrobenzene (5) were prepared according to the
standard procedure of Noelting and Michel [22] with some
modifications to improve safety [23]. The desired azides
were prepared by diazotization of the corresponding amines,
obtained from Sigma–Aldrich of Italy, and subsequent
treatment of the resulting diazonium salt solutions with a
sodium acetate-buffered aqueous solution of sodium azide.
Alternatively, 2-azidobenzenecarbaldehyde (2) was pre-
pared by the oxidation of azide 1 with pyridinium chloro-
chromate according to the previously described protocol
(mp 36–37 ꢁC) [24].
The thermogravimetric analysis, which measures weight
loss under both oxidative and non-oxidative heating con-
ditions and which takes place in an open vessel, allows the
IR and mass spectra (MS) of the released gases to be
recorded. The instruments used in the apparatus were a TG/
DSC 1 Star system from Mettler Toledo, a Nexus TGA
FTIR with a Nicolet interface and a Pfeifer Vacuum
Thermostar MS spectrometer.
Results and discussion
Thermodynamic values for azides 1–5 (heats of formation/
DfH, combustion/DcH, and decomposition/DdH) were
123

Revisiting the thermal decomposition
193
Table 2 DSCa,b and C80c determined heats of decomposition/J g-1
and temperature/ ꢁC for the thermal decompositions (TD) azides 1–5
under various conditions
calculated using the CHETAH software package (ASTM
Computer Program for Chemical Thermodynamic and
Energy Release Evaluation, CHETAH, ver. 8.0; 2005) and
are listed in Table 1. The heats of decomposition were
predicted to be from -5.21 to -3.30 kJ g-1. For reference,
when a compound’s DdH value is over the threshold of
-2.93 kJ g-1, it is considered to be dangerous.
DdHaair
DdHb
DdHc
Tid
Tpe
Tif
Tpg
1 -1928.53 -1563.42 -1809.13 135.21 194.71 116.6 162.8
2
3
4
5
a
-708.85 -709.03 -680.9 113.06 159.49 91.97 127.86
-462.43 -464.35 -394.54 82.16 130.83 58.57 101.69
The data predicted by the upgraded version of CHE-
TAH, although approximate [37], are very useful [38] for
the characterisation and comparison of substances
belonging to the same class [39]. In this case, the combined
software calculated total risk value, called ERP (Energy
Release Potential), appears to be high (Fig. 1).
-374.02 -377.53 -381.72 89.57 138.7
77.72 120.55
-527.13 -576.9 -471.5 71.34 119.81 119.81 92.6
DSC data refers to different heating scan rate U = 5 ꢁC min-1 in
static air
b
Under nitrogen
c
C80 data refers to a scan rate U = 0.3 ꢁC min-1 in a closed-vessel
under nitrogen
The use of DSC, TG-DTA and Setaram C80 FTIR calo-
rimetry techniques enabled us to characterise the (in)sta-
bility of azides 1–5 by determining their physical parameters
(pressure/P, temperature/Tn, decomposition rate/k and
energetic parameters (activation energy/Ea, enthalpy/DH
and entropy/DS of decomposition). The experimental heats
of decomposition/kJ g-1 obtained by DSC analysis, as well
as the related initial/Ti and peak/Tp temperature/ꢁC, are
shown in Table 2.
d
Initial decomposition temperature in static air
e
Peak temperature in static air
f
Initial decomposition temperature under nitrogen
g
Peak temperature under nitrogen
nitrogen (DdH) were comparable, or slightly larger in the
case of the more oxygenated azide 5, with those measured
in static air (DdHair), which confirms a concerted pathway.
In contrast, under nitrogen and in a closed vessel only azide
1 exhibited a significant decrease in (DdH) such that
(DdHair) ꢀ (DdH). As expected the individual TD C80
FTIR profile for compound 1, measured at various heating
rates, displays an endothermic melting peak at 53 ꢁC,
corresponding to the solid to liquid transition, which is
followed by a single exothermic peak starting at a Ti of
116.6 ꢁC, and characterised by a Tp of 162.8 ꢁC with a
decomposition enthalpy (DdH) of -1809.13 J g-1 and
pressure increase (DP) of 8 bar (Fig. 2).
The DSC data for the TD of oxoazides 2–5 show similar
behaviour. They were insensitive to atmospheric condi-
tions, while azide 1’s behaviour was strongly affected by
the closed- or open-vessel conditions. The absolute values
of heat capacity measured for compounds 2–5 under
Table 1 Heats of formation/DfH, combustion/DcH, and decompo-
sition/DdH for azides 1–5 calculated with CHETAH
1
2
3
4
5
DfH/kJ mol-1
DcH/kJ g-1
DdH/max/kJ g-1
202.84
266.32
-24.64 -26.13 -29.42 -19.59
-3.83 -3.42 -3.3 -5.21
215.95
364.03
370.91
Interestingly, azides 2–5 show DSC and C80 FTIR
profiles containing a second exothermic peak that corre-
sponds to the decomposition of the initial cyclisation
products 2a–5a. In the instance of the azide 4, a third peak,
-25.5
-3.61
9.0
45
40
35
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
30
25
20
15
10
5
0
–5
–10
–20
40 60 80 100 120 140 160 180 200 220 240 260
Temperature/°C
25 50 75 100 125 150 175 200 225 250 275
Temperature/°C
Fig. 1 DSC profiles of the TD of the azide 1 performed at heating
scan rate U = 5 ꢁC min-1 under nitrogen (upper) and in static air
(lower)
Fig. 2 C80 plot of the TD in glass vessel at scan rate
U = 0.3 ꢁC min-1 under nitrogen of the azide 1
123

194
P. Cardillo et al.
4b, was also detected (see Figs. 3, 4 for the TD of com-
pound 4).
O
N
O
N+
According to the previously reported results from
‘controlled’ TDs in inert solvent, we expected that the
second (and third) exothermic peaks would correspond to
the cyclised compounds namely: benzo[c]isoxazole 2a
[19, 20], 3-methylbenzo[c]isoxazole 3a [19, 20], 3-phen-
ylbenzo[c]isoxazole 4a, acridin-9(10H)-one 4b [19–21]
and benzo[c][1,2,5]oxadiazole 1-oxide 5a [19, 20] (see
Scheme 2). The formation of these cyclisation products
accounts for the lower decomposition enthalpy (DdH)
observed for azides 2–5, relative to azide 1, since part of
the heat is devolved to the cyclisation process.
O
O
-O
O
N
N
N
H
4b
N
5a
2a
3a
4a
Scheme 2 Main products detected from TD of the o-phenyl azides
2–5
13C-NMR spectra of the final residues resulting from azides
3 and 4.
Table 3 reports the experimental values for the heat of
decomposition/J g-1 and the associated temperatures, ini-
tial/Ti and peak/Tp, for the cyclisation products.
The structures of 2a–5a and 4b were confirmed by cou-
pled TG-SDTA-FTIR-MS analysis of the gases evolved
during the TD reactions, and by analysis of 1H- and
The failure of the o-carbinol group to assist the peri-
cyclic mechanism is significant and can be seen by com-
paring the Arrhenius activation energy (Ea) for azides 1, 3
and 4. The Ea parameter can be used to calculate enthalpies
[Ea = DHà ? RT] and entropies (DSa) of activation. The
DSC profiles (see Fig. 5 for the DSC profile of azide 1)
were measured at different heating rates (four DSC runs at
2, 5, 10 and 15 ꢁC min-1) and were analysed using the
standard ASTM E698 method for reactions whose behav-
iour is well described by the Arrhenius equation. Figure 5
shows a plot of log b (b = heating rate in K/min) versus
1/K, which allows the apparent activation energy (Ea), the
pre-exponential factor (A0) and the kinetic constant k at
each absolute temperature to be calculated.
The TD curves at various temperatures ranging from 90
to 200 ꢁC for azides 1, 3 and 4 were fit to first-order
kinetics as plots of ln(k/min-1) versus T/K and displayed a
linear trend with correlation coefficients of greater than
0.999. The activation energies (Ea) of these azides were
calculated, along with the corresponding frequency factor
ln(A0/min-1) values. Assuming unimolecular processes, we
then calculated the activation enthalpies (DaHà) and
40 60 80 100 120 140 160 180 200 220 240 260 280
Temperature/°C
Fig. 3 DSC profiles of the TD of the azide 3 performed at heating
scan rate U = 5 ꢁC min-1 under nitrogen (upper) and in static air
(lower)
Table 3 DSCa,b and C80c determined heats of decomposition/J g-1
and thermal decomposition temperature/ ꢁC for the products formed
by intramolecular cyclisation 2a–5a and 4b from 2–5
50
80
45
DdHaair
DdHb
DdHc
Tid
Tpe
40
35
70
60
50
2a
3a
4a
4b
5a
a
-932.21
-1086.69
-60.45
-953.99
-904.9
-1054.59
-1137.66
-54.29
152.26
164.24
143.26
190.76
126.95
189.36
220.96
149.06
245.66
193.42
30
25
20
15
10
5
-41.24
40
30
20
10
-488.3
-612.01
-1835.67
-825.83
-1953.91
-1846.64
DSC data refers to different heating scan rate U = 5 ꢁC min-1 in
0
static air
b
Under nitrogen
-5
25 50 75 100 125 150 175 200 225 250
Temperature/°C
c
C80 data refers to heating scan rate U = 0.3 ꢁC min-1 in a closed-
vessel under nitrogen
d
Initial decomposition temperature under nitrogen
Fig. 4 C80 plot of the TD in glass-vessel at heating scan rate
U = 0.3 ꢁC min-1 under nitrogen of the azide 3
e
Peak decomposition temperature under nitrogen
123

Revisiting the thermal decomposition
195
100
4 under ‘controlled’ thermal conditions were lower than
observed for azide 1 and were comparable at 107.9 and
102.7 kJ mol-1, respectively [19, 21]. This implies con-
siderable anchimeric assistance via the phenyl ring, instead
of an intermediate free nitrene, in the quick extrusion of
nitrogen from 3 and 4 in route to the formation of the
corresponding benzoisoxazoles 3a and 4a. This peculiar
trend is also confirmed by the negative activation entropies
(DaS) calculated for the TD of aryl azides 3 and 4 under
both conditions (see Tables 4, 5).
T[433]K
90
T[428]K
T[423]K
T[418]K
80
T[413]K
70
60
50
40
30
20
10
0
T[408]K
The C80 analysis of azide 4 (see Fig. 6) detected at least
three exothermic processes, with Tp = 120.45, 149.06
and 245.66 ꢁC, and with DdH = -381.72, -54.29 and
-825.83 J g-1, respectively.
250
0
50 100 150 200
300 350 400 450 500 550 500
Time/min
Fig. 5 C80, TD curves of the azide 1; mass loss/% versus times/min
at various temperatures/K
The C80 analysis of azide 4s TD is characterised by an
overall pressure increase of about 6 bars with the change
during the first decomposition, which presumably corre-
sponds to the extrusion of nitrogen. A second pressure
increase, which can be ascribed to the loss of carbon
monoxide from benzoisoxazole 4a, rather than the loss of a
residual nitrogen, suggests the possible initial formation of
a bridged bicyclic compound resulting from an intramo-
lecular 1,3-dipolar addition [21], but can be ruled out in
this case. It is well known that at high temperatures 4a
isomerises to acridinone 4b [28], and the TD behaviour of
o-(2-azidophenyl)(phenyl)methanone (4) should not be
considered fundamentally different. In light of the reported
thermodynamic studies [40] of acridinone 4b, the exo-
thermic enthalpy peak of -825.8 J g-1 appears to result
from the conversion of 4a to 4b, which can be detected by
1H-NMR, along with small amounts of various other
products (N,N-diphenylamine, 2-aminobenzaldehyde and
carbazole) that are detected by TG-FTIR.
entropies (Sa) from the data (see Table 4), which we
compared to those reported for ‘controlled’ decompositions
(Table 5).
The ASTM calculated half-life (t1/2) for the TD of azide
1 at 25 ꢁC (Table 4) was about thousand times slower than
that of azides 3 and 4 (the latter two being comparable),
and the corresponding activation energies (Ea) were 134.3
versus 105.3 and 105.3 kJ mol-1, respectively. Unfortu-
nately, no kinetic or energetic data concerning azide 1
under ‘controlled’ thermal conditions has been reported,
but the activation energies (Ea) calculated for azides 3 and
Table 4 Kinetic parameters and activation energies for the TD of
azides 1, 3 and 4 calculated from thermo-analytic data obtained using
the ASTM E698 method
ka/min-1
ln Aꢁ Ea/
kJ mol-1
DrHà/
Saa/
kJ mol-1
J K-1 mol-1
1 2.19 9 10-11 29.65 134.3
3 2.12 9 10-8 24.83 105.3
4 2.12 9 10-8 24.84 105.3
131.8
102.8
102.8
-0.7
-40.3
-40.4
The thermal behaviours of azides 1–5 were analysed by
EI-MS (70 eV) under the ion chamber conditions at 27 ꢁC
and a pressure of 10-7 atm. In the case of azide 4, the
DSC data refer to four runs at heating scan rate U = 2, 5, 10 and
15 ꢁC min-1 in static air
7.0
6.0
a
Calculated at 298 K
25
20
15
10
5
Table 5 Kinetic parameters and activation energies calculated from
data involving ‘controlled’ TDs of azides 1, 3 and 4
5.0
4.0
3.0
2.0
k/min-1
ln Aꢁ Ea/
kJ mol-1
DrHà/
Sa/
J K-1 mol-1
kJ mol-1
1a,b 4.5 9 10-4
–
–
–
–
0
3a 8.20 9 10-7 27.1 107.9
4c 9.25 9 10-7 28.8 102.7
105.4
99.6
-35.6
-57.6
–5
25 50 75 100 125 150 175 200 225 250 275
Temperature/°C
Obtained in decalin solutions
a
From [19]
b
Measured at 161.6 ꢁC
Data from [21]
Fig. 6 C80 plot of the TD in glass vessel at heating scan rate
U = 0.3 ꢁC min-1 under nitrogen of the azide 4
c
123

196
P. Cardillo et al.
characteristic initial fragmentation (M-28) of the molecular
ion at m/z 223 (M?•, \3%) leaves a mass spectrum iden-
tical to that reported for benzoisoxazole 4a [41], but which
is different than that reported for acridinone 4b. However,
4b does display some peaks in the MS that are common to
4a. The authors suggest a possible thermal partial con-
version of 4a to 4b via a nitrene intermediate under these
conditions may account for these results.
100
0
40 60 80 100 120 140 160 180 200 220 240 260 280
Similarly, the mass spectra of azides 2 and 3 show
molecular ions at m/z 147 and 161 (M?•, 7 and 3%),
respectively, and after the initial fragmentation (M-28)
follow overall patterns identical to those reported for the
corresponding benzoisoxazoles 2a and 3a, respectively
[42]. In contrast, the mass spectrum of azide 1 shows a
remarkably different profile. Its mass spectrum is typical
for a molecule undergoing multiple fragmentations. It has a
40 60 80 100 120 140 160 180 200 220 240 260 280
Temperature/°C
Fig. 7 TG curve (upper) of weight loss, its first derivate and the heat
flux measured by SDTA (differential thermal analysis) (lower) for
azide 5 undergoing TD between 30 and 300 ꢁC with U = 5 ꢁC min-1
under a 60.0 mL min-1 stream of nitrogen
relatively intense molecular ion peak at m/z 149 (M?•
,
[37%). Some of the other fragments can be attributed to
red-ox processes involving the phenyl ring (aniline, 2-
aminobenzaldehyde and benzaldehyde), or the extrusion of
small molecules (proprionitrile, cyclopropylamine, allyl
amine), whose formation greatly increases the exothermic
peak’s magnitude.
The thermal behaviour of 5 was also investigated by
DSC tests (U = 5 ꢁC min-1), carried out under both oxi-
dative and inert conditions that consisted of TDs performed
in two steps as illustrated in Fig. 8.
The DSC profile was characterised by an initial exo-
thermic peak at Tp = 121.26 ꢁC, corresponding to the
decomposition of azide 5. When the vessel was cooled to
allow the crystallisation of the resulting product, subsequent
heating of the vessel to 300 ꢁC showed an endothermic
melting peak at 71 ꢁC for 5a (lit. mp 71–73 ꢁC [20]) and an
exothermic decomposition starting at Ti = 180.61 with
Tp = 192.87. It is noteworthy that the enthalpy of decom-
position for azide 5 (DdH = -471.5 J g-1) measured by the
C80 under nitrogen, resulted in a value four times smaller
than that of the benzo[c][1,2,5]oxadiazole 1-oxide (5a)
(DdH = -1953.9 J g-1). Moreover, the C80 investigation
reveals a significant pressure increase of about 20 bars for
It is well known that the controlled decomposition of
azide 5 in boiling toluene gives the corresponding
benzo[c][1,2,5]oxadiazole 1-oxide 5a in 88% yield [29,
43]. The reaction has been deeply investigated from both
the experimental and theoretical [44] points of view, since
the reaction produces heterocyclic compounds important in
biochemical and pharmacological applications. The MS-EI
spectrum of azide 5 at 27 ꢁC again shows the initial
molecular ion at m/z 164 with a very low intensity (M?•
,
\5%), which is followed by fragmentation (M-28), and
then by an overall pattern identical to that reported for the
corresponding benzoxadiazole 5a [45].
Among the azides under investigation, azide 5 contains
the most favourable oxygen balance, which has a clear
influence on the heat capacity of its TD when carried out in
a closed or open vessel (see Table 2). TG-SDTA-FTIR-MS
techniques allow the spectroscopic (IR and MS) charac-
terisation of the thermal process’s products and a qualita-
tive understanding of the flux of heat, which is associated
with the sample’s first weight loss in the DSC profile. As
shown in Fig. 7, the sample loses 30% of its weight during
the evolution of the first exothermic peak. The C80-FTIR
spectrum shows no sign of any IR active gasses, and the
mass spectrum indicates the exclusive presence of nitrogen.
During the second weight loss at temperatures over
160 ꢁC, an IR spectrum identical to that of the benzo[c]
[1,2,5]oxadiazole 1-oxide 5a was observed [46]. The
identity of this compound was also confirmed using the
mass spectrum instrument interfaced directly with the TD
vessel.
Temperature/°C
50 100
10
100
50
100 150 200 250
60 70 80 90
0
20
30
40
50
Time/min
Fig. 8 DSC curve of the probe (U = 5 ꢁC min-1) for the product 5a
from the TD of the azide 5
123

Revisiting the thermal decomposition
197
the entire process. This is associated with 5a’s highly exo-
thermic enthalpies of decomposition, especially under oxi-
dative conditions, which have been reported to be mainly
due to the low dissociation enthalpy of the N–O bond [47,
48]. For most classes of compounds, including organic
azides, the primary cleavage represents a qualitative indi-
cator of the stability of a compound and correlates with the
ground state energy content, which can be used to compare
the relative stabilities of compounds. The investigation of
TD mechanisms by computational modelling is important
both to gain an appreciation for how energy is stored and
released and to gain information useful for the design of
more stable molecules. Using density functional theory
(DFT) at the B3LYP/6-31G* level [the data were obtained
using the standard program Spartan ‘06 Windows, released
by Wavefunction Inc., Irvine, CA (2006) (www.wavefun.
com), running on an AMD K7 processor at 1333 MHz], the
structures and the energetics of the starting azides 1–5 and
the products of the pericyclic reactions (2a–5a) were opti-
mised (including the calculation of the LUMO–HOMO
gap). Other possible intermediate structures involved in the
TD, such as bridged bicyclic structures that would result
from potential intramolecular 1,3-cycloadditions, were not
considered here (see Scheme 3). In all cases, the energies
separating the HOMO and LUMO orbitals (DE = 1: 5.095;
2: 4.709; 3: 4.764; 4: 4.481; 5: 4.350 eV) are in good
agreement with the activation energies (Ea) presented
and those previously reported, with the slight exception of
azide 3.
Conclusions
The potentially dangerous properties and thermo-chemical
behaviours of five ortho-substituted phenyl azides (1–5)
were investigated by readily accessible DSC, TG and C80
calorimetric combined with NMR, FTIR and mass spec-
troscopic techniques. These modern techniques require
only small amounts of the compounds to be tested and
share the advantages of being highly specific, accurate,
clean and versatile. In this case, their application allowed
investigations into thermo-chemical transformations, like
the pericyclic process that generates benzoxazoles 2a–4a
and benzoxadiazole 5a in the ‘solid phase’, and into the
properties of the starting azides and resulting products. The
rate enhancements, low activation energies and negative
activation entropies determined for the TD of azides 3 and
4; which are different from those measured for the azide 1
but were comparable with those previously measured by
means of ‘controlled’ thermal reactions in solution. This
study confirms the effectiveness of ‘solid phase’ methods
for studying thermal processes in a field where the char-
acterisation of the hazardous properties of compounds must
be gained and the mechanistic aspect could be considered.
Acknowledgements The authors gratefully acknowledge the
`
‘Ministero della Universita e Ricerca’ for the majority of the financial
support for this study, and the Ateneo di Bologna from the ‘Progetto
di Finanziamento Triennale’.
A kinetic validation of the TD of azide 3 was carried out
using a DSC test performed isothermally at 95 ꢁC for
27 min, which showed an exothermic peak at the same
Tp = 130.83 ꢁC, but with a heat of decomposition (DdH)
twofold smaller than the normal scan (see Fig. 3) due to the
partial evaporation of the resulting product. This phe-
nomenon was confirmed by TG-SDTA and C80 experi-
ments that involved warming azide 3 to 140 ꢁC and
confirming the formation of 3-methylbenzo[c]isoxazole
(3a) by examining 1H- and 13C-NMR spectra of the
residue.
References
1. Bayley H, Staros JV. Photoaffinity labelling and related tech-
niques. In: Scriven EFV, editor. Azides and nitrenes, reactivity
and utility. Orlando: Academic Press; 1984. p. 433–90.
2. Breslow DS. Industrial application. In: Scriven EFV, editor.
Azides and nitrenes, reactivity and utility. Orlando: Academic
Press; 1984. p. 491–521.
3. Brase S, Gil C, Knepper K, Zimmermann V. Organic azides: an
exploding diversity of a unique class of compounds. Angew
Chem Int Ed Engl. 2005;44:5188–240.
4. Urben PG, editor. Bretherick’s handbook of reactive chemical
hazards, vols 1 and 2. 7th ed. Amsterdam: Elsevier; 2007.
5. Cartwright M, Wilkinson J. New trends in research of energetic
materials. In: Ottis J, Pachman J, editors. Proceedings of the
Seminar, vol. 1. 11th ed. Czech Republic: Pardubice; 2008.
p. 99–112.
6. Huynh M-HV, Hiskey MA, Chavez DE, Naud DL, Gilardi RD.
Synthesis, characterization, and energetic properties of diazido
heteroaromatic high-nitrogen C-N compound. J Am Chem Soc.
2005;127:12537–43.
R
R
N
R
∆
O
O
+ N2
O
N
N+ N-
N3
X
7. Duddu R, Dave PR, Damavarapu R, Surapaneni R, Gilardi R,
Parrish D. Synthesis of azido heterocycles. Synth Commun. 2008;
38:767–74.
8. Agrawal JP, Hodgson RD. Organic chemistry of explosives.
Chichester: Wiley; 2007. p. 333–9.
R
N
N
N
R:H,Me,Ph
O
9. Walton R, Lahti PM. An efficient, simple synthesis of 4-azido-
benzaldehyde. Synth Commun. 1998;288:1087–92.
Scheme 3 Possible mechanism of the TD first step
123

198
P. Cardillo et al.
10. Cardillo P, Gigante L, Lunghi A, Fraleoni-Morgera A, Zanirato P.
Hazardous N-containing system: thermochemical and computa-
tional evaluation of the intrinsic molecular reactivity of some aryl
azides and diazides. New J Chem. 2008;32:47–53.
30. Dyall LK, Kemp JE. The infrared spectra of aryl azides. Aust
J Chem. 1967;20:1395–402.
31. Von E, Doering W, De Puy CH. Diazocyclopentadiene. J Am
Chem Soc. 1953;75:5955–7.
11. Salatelli E, Zanirato P. The conversion of furan-, thiophene- and
selenophene-2-carbonyl azides into isocyanates: a DSC analysis.
ARKIVOC 2002;xi:6–16.
32. Boyer JH, Toggweiler U, Stoner GA. Spectrophotometric rela-
tionships between furoxanes and nitroso compounds. J Am Chem
Soc. 1957;79:1748–51.
12. Stadlbauer W, Hojas G. Study of the thermal behavior of azi-
dohetarenes with differential scanning calorimetry. J Biochem
Biophys Methods. 2002;53:89–99.
33. Smith PAS, Hall JH. Kinetic evidence for the formation of azene
(electron-deficient nitrogen) intermediates from aryl azides. J Am
Chem Soc. 1962;84:480–5.
13. Smith PAS. Aryl and heteroaryl azides and nitrenes. In: Scriven
EFV, editor. Azides and nitrenes, reactivity and utility. Orlando:
Academic Press; 1984. p. 95–204.
14. Schuster GB, Platz MS. Photochemistry of phenyl azide. Adv
Photochem. 1992;17:69–73.
15. Hrovat DA, Waali EE, Thatcher Borden W. Ab initio calculations
of the singlet-triplet energy difference in phenylnitrenes. J Am
Chem Soc. 1992;114:8698–9.
34. Barton DHR, Sammes PG, Weingarten GG. Photochemical
transformations. Part XXVIII. Aryl azides as potential photo-
sensitive protecting groups. J Chem Soc C. 1971;721–8.
35. Minato M, Lahti PM. Intramolecular exchange coupling of
arylnitrenes by oxygen. J Phys Org Chem. 1994;7:495–502.
36. Watson ES, O’Neill MJ, Justin J, Brenner N. A differential
scanning calorimeter for quantitative differential thermal analy-
sis. Anal Chem. 1964;36:1233–8.
16. Albini A, Bettinetti G, Minoli G. The effect of the p-nitro group
on the chemistry of phenylnitrene. A study via intramolecular
trapping. J Chem Soc Perkin. 1999; 1:2803–7.
37. Seaton WH. Group contribution method for predicting the
potential of a chemical composition to cause an explosion.
J Chem Educ. 1989;66:A137.
17. Kvaskoff D, Bednarek P, George L, Pankajakshan S, Wentrup C.
Different behavior of nitrenes and carbenes on photolysis and
thermolysis: formation of azirine, ylidic cumulene, and cyclic
ketenimine and the rearrangement of 6-phenanthridylcarbene to
9-phenanthrylnitrene. J Org Chem. 2005;70:7947–55.
18. Fraleoni-Morgera A, Zanirato P. BF3.OEt2-promoted synthesis
of acridines via N-aryl nitrenium-BF3 ions generated by disso-
ciation of 2-oxo azidoarenes in benzene. ARKIVOC. 2006;1:
111–20.
38. Frurip D, Britton L, Fenlon W, Going J, Harison BK, Niemier J,
et al. The role of ASTM E27 methods in hazard assessment: Part
I. Thermal stability, compatibility, and energy release estimation
methods. Process Saf Prog. 2004;23:266–78.
39. Cardillo P, Gigante L, Lunghi A, Di Bari C, Ludovisi G. La
termodinamica per la sicurezza chimica: criteri di previsione
`
dell’instabilita termica. Riv Combust. 2002;56:209–23.
´
40. Sabbah R, El Watik L. Etude thermodynamique de l’acridone et
de la thioxanthone. Can J Chem. 1992;70:24–8.
41. Dyall LK, Karpa GJ. Mass spectra of 3-phenyl-2,1-benzisoxaz-
oles. Org Mass Spectrom. 1989;24:70–3.
19. Dyall LK, Kemp JE. Neighbouring-group participation in pyro-
lysis of aryl azides. J Chem Soc B. 1968; 976–9.
20. Dyall LK, Holmes A-L. Pyrolysis of aryl azides. IX. Azomethines
as weak neighboring groups. Aust J Chem. 1988;41:1677–86.
21. Hall JH, Behr FE, Reed RL. Cyclization of 2-azidobenzophe-
nones to 3-phenylanthranils. Examples of an intramolecular
1,3-dipolar addition. J Am Chem Soc. 1972;94:4952–8.
22. Noelting E, Michel O. Direkte ueberfu¨hrung von aminen in
42. Doppler T, Schmid H, Hansen H-J. Zur photochemie von 2,
1-benzisoxazolen (anthranilen) und thermischen und photo-
chemischen umsetzungen von 2-azido-acylbenzolen in stark sa-
¨
urer losung. Helv Chim Acta. 1979;62:271–303.
43. Boulton AJ, Gray ACG, Katritzky AR. Heterocyclic rearrange-
ments. Part. IV, Furoxan- and furazan-benzofuroxan. J Chem
Soc. 1965;5958–64.
44. Rauhut G, Eckert F. A computational study on the mechanism
and kinetics of the pyrolysis of 2-nitrophenyl azide. J Phys Chem
A. 1999;103:9086–92.
¨
diazoimide mittels stickstoffwasserstosaure. Chem Ber. 1893;26:
86–92.
23. Dyall LK. Pyrolysis of aryl azides. VII. Interpretation of Hammett
correlations of rates of pyrolysis of substituted 2-nitroazidobenz-
enes. Aust J Chem. 1986;39:89–101.
¨
45. Leitao MLP, Pilcher G, Acree WE Jr, Zvaigzne AI, Tucker SA,
24. Ardakani MA, Smalley RK, Smith RH. 1H- and 2H-indazoles by
thermal and photolytic decomposition of o-azidobenzoic acid and
o-azidobenzaldehyde derivatives. J Chem Soc Perkin. 1983;1:
2501–6.
Ribeiro Da Silva MDMC. Enthalpies of combustion of phenazine
N-oxide, phenazine, benzofuroxan, and benzofurazan: the disso-
ciation enthalpies of the (N–O) bonds. J Chem Thermodyn.
1990;22:923–8.
25. Smolinsky G. Notes—The vapor phase pyrolysis of several
subsituted azidobenzenes. J Org Chem. 1961;26:4108–10.
26. Sakai K, Anselme J-P. Rational synthesis of 2-aminoindazole.
J Org Chem. 1972;37:2351–2.
46. Gaughran RJ, Picard JP, Kaufman JVR. Contribution to the
chemistry of benzofuroxan and benzofurazan derivatives. J Am
Chem Soc. 1954;76:2233–6.
47. Acree WE Jr, Pilcher G, Ribeiro da Silva MDMC. The dissoci-
ation enthalpies of terminal (NO) bonds in organic compounds.
J Phys Chem. 1982;34:560–2.
48. Zhang C, Shu Y, Huang Y, Zhao X, Dong H. Investigation of
correlation between impact sensitivities and nitro group charges
in nitro compounds. J Phys Chem B. 2005;109:8978–82.
¨
27. Meisenheimer J, Senn O, Zimmermann P. Uber die oxime des
o-amino-benzo- und acetophenons. Chem Ber B. 1927;60:
1736–48.
28. Smith PAS, Brown BB, Putney RK, Reinisch RF. The synthesis
of heterocyclic compounds from aryl azides. iii. some six-mem-
bered rings and some azidobiaryls. J Am Chem Soc. 1953;75:
6335–7.
29. Monge A, Palop JA, Lopez de Cerain A, Senador V, Martinez-
Crespo FJ, Sainz Y, et al. Hypoxia-selective agents derived from
quinoxaline 1,4-di-N-oxides. J Med Chem. 1995;38:1786–92.
123