A study of the thermal decomposition of 2-azidoacetamide by ultraviolet photoelectron spectroscopy and matrix-isolation infrared spectroscopy: Identification of the imine intermediate H2NCOCHNH

Article abstract of DOI:10.1021/jp031288s

The thermal decomposition of 2-azidoacetamide (N₃CH ₂CONH₂) has been studied by matrix-isolation infrared spectroscopy and real-time ultraviolet photoelectron spectroscopy. N ₂, CH₂NH, HNCO, CO, NH₃, and HCN are observed as high-temperature decomposition products, while at lower temperatures, the novel imine intermediate H₂NCOCH=NH is observed in the matrix-isolation IR experiments. The identity of this intermediate is confirmed both by ab initio molecular orbital calculations of its IR spectrum and by the temperature dependence and distribution of products in the photoelectron spectroscopy (PES) and IR studies. Mechanisms are proposed for the formation and decomposition of the intermediate consistent both with the observed results and with estimated activation energies based on pathway calculations.

Full text of DOI:10.1021/jp031288s

J. Phys. Chem. A 2004, 108, 5299-5307
5299
ARTICLES
A Study of the Thermal Decomposition of 2-Azidoacetamide by Ultraviolet Photoelectron
Spectroscopy and Matrix-Isolation Infrared Spectroscopy: Identification of the Imine
Intermediate H2NCOCHNH
J. M. Dyke,* G. Levita, A. Morris, and J. S. Ogden
Department of Chemistry, UniVersity of Southampton, Southampton SO17 1BJ, U.K.
†
†
†
†
‡
‡
A. A. Dias, M. Algarra, J. P. Santos, M. L. Costa, P. Rodrigues, and M. T. Barros
CEFITEC/CFA, Department of Physics, Faculdade de Ciencias Tecnologia, UniVersidade NoVa de Lisboa,
2
829-215 Caparica, Portugal, and CQFB, Department of Chemistry, Faculdade de Ciencias Tecnologia,
UniVersida NoVa de Lisboa, 2829-215 Caparica, Portugal
ReceiVed: December 5, 2003; In Final Form: March 30, 2004
The thermal decomposition of 2-azidoacetamide (N
spectroscopy and real-time ultraviolet photoelectron spectroscopy. N
are observed as high-temperature decomposition products, while at lower temperatures, the novel imine
intermediate H NCOCHdNH is observed in the matrix-isolation IR experiments. The identity of this
3 2 2
CH CONH ) has been studied by matrix-isolation infrared
2
, CH NH, HNCO, CO, NH , and HCN
2
3
2
intermediate is confirmed both by ab initio molecular orbital calculations of its IR spectrum and by the
temperature dependence and distribution of products in the photoelectron spectroscopy (PES) and IR studies.
Mechanisms are proposed for the formation and decomposition of the intermediate consistent both with the
observed results and with estimated activation energies based on pathway calculations.
Introduction
proceed stepwise via the formation of the nitrene, but there is
no experimental evidence for this. Although the UV photoelec-
tron spectrum of CH3N, made from decomposition of CH3N3,
A study of the thermal decomposition of organic azides is
important for a variety of reasons, not least because of their
potential roles both as explosives and as systems for high-energy
storage. As such, their intrinsic instability makes their properties
difficult to measure, and experimental work must be carried
out with care, ideally being limited to samples in small
quantities. Nevertheless, the study of the mechanisms of organic
azide thermal decompositions is of considerable interest both
from a fundamental viewpoint and because of the important
has recently been reported by Dianxun and co-workers,17 this
species was produced by the decomposition of the parent azide
on a heated molecular sieve surface in the presence of NO. It
appears that there have been no observations of nitrene
intermediates from gas-phase azide decompositions. The py-
rolysis of CH3N3 is therefore envisaged as
1
-3
roles played by such azides in heterocycle synthesis
and
biological and pharmaceutical processes.4 They are also used
-6
in the preparation of semiconductor materials7 and as high-
,8
9
,10
energy density materials used in solid propellants
and
followed by
explosives.11
Bock and Dammel1
2-16
were among the first workers to study
the thermal decompositions of alkyl azides in the gas-phase
using UV photoelectron spectroscopy (PES). The main conclu-
sion from this work was that alkyl azide thermal decomposition
follows a path in which N2 evolution is accompanied by a 1,2
shift to form an imine, which can undergo further elimination
to produce a nitrile. This process might also be envisaged to
This basic reaction sequence involving an initial 1,2 shift
satisfactorily accounts for the range of products found in azides
of type RR′R′′CN3, where R, R′, and R′′ may be H, Me, or
1
2,18
allyl,
and can be described by a “type 1” decomposition
pathway. In this process, the migrating group need not neces-
*
To whom correspondence should be addressed.
CEFITEC/CFA, Department of Physics.
CQFB, Department of Chemistry.
†
sarily be only H. Thus Me3CN3 decomposes to form Me2Cd
‡
15,19
NMe, and ultimately MeCN and (probably) C2H6.
1
0.1021/jp031288s CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/22/2004

5300 J. Phys. Chem. A, Vol. 108, No. 25, 2004
Dyke et al.
More recently, further studies on the thermal decomposition
13C NMR spectrum (CDCl3 solution) shows peaks at 52.7 ppm
(relative to TMS) assigned to the methylene carbon and at 170.2
ppm, which is assignable to the carbonyl carbon.
of organic azides have been reported using the combined
application of PES and matrix-isolation infrared spectroscopy,
together with ab initio molecular orbital methods,20-22 and these
studies have revealed a second “type 2” mode of azide
decomposition. The first system for which this was found to
The most significant features in the Nujol mull IR spectrum
-1
(KBr plates) were a broad doublet at 3373 and 3179 cm
assigned to N-H stretching modes, C-H stretching modes at
2
0
apply was the thermal decomposition of 2-azidoacetic acid,
-1
2
980-2920 cm (overlapping with Nujol bands), and promi-
which decomposed to give CO2 and CH2NH as important
products not expected by the “type 1” mechanism.
For this system, an alternative “type 2” route was proposed
as
-1
-1
nent bands at 2117 cm (N3 stretching), 1630 cm (CdO
stretching), and 1414 cm .
-1
Matrix-Isolation IR Studies. Matrix-isolation infrared stud-
ies followed a very similar pattern to those described for
20-22
previous studies on organic azides.
The inlet and pyrolysis
parts of the apparatus were identical, as were the low-
temperature Displex and IR spectrometers. However, although
spectra could generally be obtained with the parent azide
maintained at room temperature, it was often convenient to
warm the sample and inlet system to ca. 300 K to generate a
more satisfactory flux of material. Spectra of the parent azide
and its decomposition products were obtained in nitrogen
matrices, and supporting experiments were also carried out on
NH3/N2 and H2NCHO/N2 to augment our N2 matrix infrared
data bank of known molecular vibration frequencies. Deposition
times were typically 30-60 min at a particular superheater
temperature, and any changes occurring during this period were
monitored by spectral subtraction.
This second decomposition route was subsequently found to
22
be applicable for ethyl azidoacetate decomposition and also
for some of the products identified in the pyrolysis of azido-
2
1
acetone.
This paper describes related studies on the pyrolysis of
-azidoacetamide using the techniques of real-time ultraviolet
2
photoelectron spectroscopy and matrix-isolation infrared spec-
troscopy, supported by ab initio molecular orbital calculations.
The main aim is to investigate the mechanism of thermal
decomposition by obtaining spectra at different temperatures
and so perhaps detect not only the final products of pyrolysis
but also any intermediates; this may lead to a greater under-
standing of the decomposition process.
Matrix ratios were estimated to be well in excess of 1000:1.
Photoelectron Spectroscopy. The photoelectron spectra
obtained for 2-azidoacetamide were recorded using the single-
23
detector instrument described elsewhere, specifically built for
high-temperature decomposition studies. All the spectra were
recorded using He IR radiation (21.22 eV). In contrast to
previous PES studies on the organic azides 2-azidoacetone,
Experimental Section
21,22
2
-azidoethanol, and 2-azidoethyl acetate,
2-azidoacetamide
has insufficient vapor pressure at room temperature to allow
photoelectron spectra to be recorded with acceptable signal-to-
noise by simply pumping on the sample held in a flask outside
the ionization chamber of the spectrometer. As in the case of
Preparation and Characterization of N3CH2CONH2.
Samples of 2-azidoacetamide were obtained from the reaction
of chloracetamide with sodium azide under aqueous conditions.
In a typical reaction, ca. 2.0 g of chloroacetamide was added
slowly to a solution of sodium azide (3 equiv) in water. The
mixture was stirred for 24 h in an oil bath at 60 °C. After
cooling, the product was extracted with ethyl acetate, and the
organic phase was dried over anhydrous sodium sulfate and
concentrated using a rotary evaporator.
2
0
2
-azidoacetic acid, it was necessary to preheat the sample
inside the ionization chamber to achieve a sample vapor pressure
in the photoionization region, which was high enough to give
spectra with acceptable signal-to-noise ratios. This was achieved
by placing solid azidoacetamide samples in a small, noninduc-
tively wound, resistively heated furnace placed above both the
radio frequency (rf) pyrolysis tantalum furnace and the photon
beam in the ionization chamber of the spectrometer. Photoelec-
tron spectra of the parent azide were calibrated using argon and
2-Azidoacetamide (N3CH2CONH2) is a white crystalline solid
at room temperature. It was purified by recrystallization from
dichloromethane (mp 55-56 °C) and was characterized in the
vapor phase by ultraviolet photoelectron spectroscopy (PES)
2
4
1
methyl iodide added to the ionization chamber at the same
time as the azide vapor samples. The photoelectron spectrum
of the azide precursor, chloroacetamide, was also recorded and
calibrated using methyl iodide and argon. This was done to test
for the possible presence of this impurity in the azide spectra.
In practice, all parent azide spectra were free from any detectable
trace of the precursor used in the preparation. The wall
temperatures of the resistively heated and the lower rf furnace
were measured in each case by a K-type thermocouple placed
in contact with the walls. The pressure of the azide in the
and electron-impact mass spectrometry, in solution by H and
1
3
C NMR (Bruker AMX-400), and in the solid phase by IR
spectroscopy (Nujol mull-Mattson Satellite FT-IR).
The 70 eV electron-impact mass spectrum displayed a parent
+
+
peak at 100 amu (2.1%) and stronger peaks at 28 (N2 , CH2N ,
+
+
1
00%), 44 (CONH2 , 17.1%) and 72 (NCH2CONH2 , 2.6%)
amu. Signals were present also at 29 (HCO , 4.9%) and 43
HNCO , 2.6%) amu. In the 20 eV electron-impact mass
+
+
(
spectrum, the relative intensities of all the ions were enhanced
+
with respect to the N2 intensity. The relative intensity for the
-
4
+
+
ionization region was estimated as ca. 10 Torr under condi-
tions at which pyrolysis experiments were started, and the flight
time between the pyrolysis region and the photoionization region
was estimated as ca. 5 ms, and from the length of the pyrolysis
zone, the residence time in this region is estimated as 3 ms.
parent ion was 14.8%, for CONH2 100%, for NCH2CONH2
+
+
2
4.8%, for HCO 20.9%, and for HNCO 4.8%.
The H NMR spectrum (CDCl3 solution) shows a singlet at
.01 ppm relative to TMS corresponding to the methylene group
1
4
and a broad (1:1) doublet at 6.46 ppm assigned to the NH
protons. The ratio between the CH2 and NH2 signal intensities
was 1.04:1. The doublet observed for the NH2 signal is attributed
to the two -NH2 protons being chemically inequivalent. The
The thermal decomposition of 2-azidoacetamide was moni-
tored by PES in real time by slowly increasing the temperature
of the rf pyrolysis region with a fixed temperature of the

Thermal Decomposition of 2-Azidoacetamide
J. Phys. Chem. A, Vol. 108, No. 25, 2004 5301
TABLE 1: Geometrical Parameters for the Lowest Energy
a
Structure of 2-Azidoacetamide and Total Energies of the
Four Minimum-Energy Structures of Azidoacetamide at the
MP2/6-31G** Level
Geometrical Parameters for the Lowest Energy Structure
bond
length (Å)
angle
value (deg)
N11-N10
N10-N9
N9-C3
C3-H7
C3-C1
C1-O2
C1-N4
N4-H5
1.163
N11-N10-N9
N10-N9-C3
N9-C3-H7
173.2
1.2455
1.4776
1.0913
1.5243
1.2291
1.3573
1.0059
115.52
111.25
110.65
124.87
119.1
189.3
191.6
N9-C3-C1
O2-C1-N4
C1-N4-H5
N10-N9-C3-C1
N9-C3-C1-O2
Total Energies of the Four Minimum-Energy Structures
rel. energy
-
1
Figure 1. Structure of cis-cis conformer of 2-azidoacetamide.
structure
energy (au)
(kcal mol )
trans-trans
cis-cis
-371.761 257
-371.764 376
-371.764 112
-371.756 125
1.96
0.0
0.16
5.18
resistively heated furnace, the onset of pyrolysis being marked
by the appearance of characteristic N2 bands, an associated
lowering of the parent azide bands, and an increase in other
features associated with the pyrolysis products. At this stage, a
further spectral calibration was made using the vertical ionization
cis-trans
trans-cis
a
See Figure 1 for the numbering of the atoms.
2
4
energies (VIEs) of the first bands of N2, H2O, or HCN.
Molecular Orbital Calculations. To assist interpretation of
the photoelectron and infrared spectra, molecular orbital calcula-
tions were carried out on the parent azide and on the imine
H2NCOCHdNH with the aim of establishing equilibrium
geometries and then to calculate vertical ionization energies
(VIEs) and infrared frequencies and intensities. These calcula-
tions were carried out at the MP2/6-31G** level. For the VIEs,
Koopmans’ theorem was applied to the SCF orbital energies
obtained at the MP2/6-31G** geometry, and the values obtained
were scaled by a factor of 0.92.2
5,26
Also, some of the lower
VIEs were calculated with the ∆SCF method. Harmonic
vibrational frequencies were calculated at the MP2/6-31G**
level via second-derivative calculations. These frequencies are
expected to be high (by ca. 5%) because only partial allowance
27-29
has been made for electron correlation in the method used.
Also, no corrections were made for anharmonicity in comparing
the computed harmonic frequencies with the experimental
frequencies.
Figure 2. He I PE spectrum of 2-azidoacetamide at sample temperature
of 320 K.
TABLE 2: Experimental and Computed VIEs of the Lowest
Energy Structure of 2-Azidoacetamide
Results and Discussion
Calculated Equilibrium Geometry of 2-Azidoacetamide.
At the MP2/6-31G** level of calculation, four minimum energy
conformers were found for N3CH2CONH2 in its closed-shell
singlet state, the relatively small energy differences arising from
the different relative orientations of the carbonyl, methylene,
and azide groups. The most stable predicted geometry is shown
in Figure 1 and is described as the cis-cis conformer. The most
important structural parameters of this conformer and the relative
energies of the four minimum energy conformers are sum-
marized in Table 1.
KT calcd
KT calcd
∆SCF calcd exptl VIE
VIE (eV)a 0.92 × VIE (eV)
VIE (eV)
(eV)
bandb
1
0.69
9.83
10.43
10.76
11.56
13.57
14.69
15.75
16.11
16.91
17.56
10.24
10.16
10.68
A
B
11.34
1
1
1
1
1
1.70
2.57
4.75
5.97
7.11
10.40
11.94
13.37
14.44
15.35
C
D
E
F
17.52
18.38
16.93
17.99
G
H
1
9.09
Observed and Calculated VIEs of 2-Azidoacetamide.
Figure 2 shows a typical photoelectron spectrum of 2-azido-
cetamide. Eight distinct bands may be readily identified and
are labeled A-H in this figure. VIEs were determined by
averaging the VIEs measured from nine different spectra.
Table 2 reports the first nine VIEs for the most stable
conformer (structure cis-cis), calculated using Koopmans’
theorem and scaled by a factor of 0.92 and two VIEs calculated
with the ∆SCF method by optimizing the geometry of the cation.
These calculations are in satisfactory agreement with the
a
Koopmans’ theorem (KT) values (see refs 25 and 26). b See Figure
2
for the lettering of the bands.
experimental spectrum and indicate that bands B and F both
arise from overlap of two one-electron ionizations.
Of the four minimum energy structures located at the MP2/
6-31G** level, three of these structures differ by less than 2.0
kcal mol , while the trans-cis structure lies approximately 5.2
kcal mol higher than the most stable (cis-cis) structure. The
-1
-
1

5302 J. Phys. Chem. A, Vol. 108, No. 25, 2004
Dyke et al.
Figure 3. Nitrogen-matrix IR spectrum (a) of 2-azidoacetamide and calculated IR spectra of (b) cis-cis conformer, (c) cis-trans conformer, (d)
trans-cis conformer, and (e) trans-trans conformer of 2-azidoacetamide.
description used here reflects the relative positions (cis or trans)
of the methylene hydrogens relative to the oxygen atom and of
the azide chain relative to the methylene hydrogens (cis or trans).
For example, Figure 1 shows the cis-cis structure with the
and neglect of anharmonicity; this latter effect is expected to
be particularly important for N-H and C-H stretches.
Thermal Decomposition Experiments: IR Spectroscopy.
Figure 4 shows typical nitrogen-matrix infrared spectra obtained
from a sample of N3CH2CONH2 for a series of increasing
pyrolysis temperatures. At the lowest temperature (300 K), the
only absorptions detected are those of the parent (the most
prominent bands are denoted P), together with the two most
intense IR features of matrix-isolated water-denoted by an
asterisk (/). Such traces of water are typically observed from
organic azide samples, and they most probably arise either as a
result of residual water in the parent azide or as a result of
desorption from the inner glass surfaces of the heated inlet
system.
-CH2 group and the -N3 group cis to each other and the -CH2
and CdO groups cis to each other. Because of the small
differences in energy between the conformers cis-cis, trans-
trans, and cis-trans, it is possible that these three structures all
contribute to the experimental gas-phase spectrum. Valence
vertical ionization energies have been computed for these
structures using Koopmans’ theorem, and they were found to
be very similar.
Observed and Calculated IR Spectra of 2-Azidoacetamide.
Figure 3a shows a typical nitrogen-matrix infrared (IR) spectrum
obtained for 2-azidoacetamide. Prominent bands are found at
Pyrolysis experiments were carried out in the range ca. 400-
800 K, and at the relatively low temperature of 470 K, there is
evidence for the presence of a pyrolysis product characterized
-
1
3
538, 3420, 2126, 1719, and 1574 cm , bands that are at
similar, though somewhat higher, positions to those recorded
for the Nujol mull spectrum of the solid. The most significant
difference is the considerably higher frequency of the two N-H
-1
by a weak band at 869 cm . Further heating in the range 540-
590 K shows a significant reduction in the prominent azide peak
-
1
-1
stretching modes (3538 and 3420 cm in the matrix compared
at 2126 cm and the growth of at least 10 significant features.
-1
with 3373 and 3179 cm in the solid), and this almost certainly
arises from the lack of intermolecular hydrogen bonding in the
matrix-isolated species.
One of these, labeled A, is identified as HNCO by comparison
3
0
with previous studies, but the remaining new bands, labeled
I, have not previously been seen in our IR pyrolysis studies.
They are tentatively assigned to a common species on the basis
of their reproducible relative intensity ratios over a range of
experimental conditions.
Figure 3b shows the calculated IR spectrum for the lowest
energy cis-cis conformer, and apart from the anticipated errors
in absolute band position, the overall agreement with respect
to band intensities and general appearance is satisfactory and
lends credence to the use of our calculations in supporting the
identification of molecular species of this type.
Calculations were also carried out on the three other low-
energy conformers found for 2-azidoacetamide, but these did
not reproduce the experimental spectrum quite as well. The
computed spectra for these conformers are also shown in Figure
Yet further heating (650-800 K) results in a decrease of
bands I, and the continued growth of well-established features
due to HCN (bands B), CO (band C), CH2NH (bands D), and
NH3 (band E), together with a second band of HNCO (also
labeled A). Table 3 summarizes the positions and assignments
of the bands observed in these experiments.
These pyrolysis runs were very reproducible and indicate that
although the final decomposition products are well-established,
they are formed via an intermediate (bands I) that has some
structural features in common with the parent azide, notably
the ketonic CO group and the NH2 unit. This intermediate is
present over a relatively well-defined temperature range (ca.
500-700 K), and its decomposition appears to result in the
formation of all the remaining species identified.
3. Comparison with the experimental spectrum (Figure 3a) is
least satisfactory for the trans-cis structure (the highest energy
structure) and the trans-trans structure (the next highest energy
structure), notably in the N-H stretching region (3300-3900
-
1
-1
cm ) and the 2400-1200 cm region. The computed values
are all slightly higher than the experimental values because of
only partial allowance for electron correlation in the calculations

Thermal Decomposition of 2-Azidoacetamide
J. Phys. Chem. A, Vol. 108, No. 25, 2004 5303
Figure 4. Sequence of nitrogen matrix IR spectra obtained from pyrolysis of 2-azidoacetamide over the temperature range 300-800 K (band
identification given in Table 3).
Thermal Decomposition Experiments: Photoelectron Spec-
troscopy. Figure 5a,b shows typical photoelectron spectra
obtained at increasing temperatures, reflecting an increasing
degree of pyrolysis of 2-azidoacetamide from 0% at 320 K to
ca. 100% at 870 K.
to HCN is also present. At 870 K, Figure 5b, the parent azide
bands have essentially disappeared. Further heating, ultimately
to ca. 1100 K, resulted in the eventual disappearance of all the
above bands, apart from HCN and N2, and the growth of a band
due to CO.
The first evidence of pyrolysis was the appearance of the
first band of molecular nitrogen at 15.58 eV (VIE), and by
50 K, Figure 5a, the N2 bands are relatively intense, and
It is notable that HCN and CH2NH are both observed early
in the pyrolysis at approximately the same (modest) furnace
temperature, and it is evident that, unlike the azidoacetic acid
2
4
5
2
0
features arising from three additional pyrolysis products appear
simultaneously. The broad band at 10.66 eV (VIE) and a
vibrationally resolved band at 12.50 eV (VIE) are both attributed
system, HCN is not produced here from CH2NH.
No features were observed in these PES studies that could
be attributed to either an imine intermediate, H2NCOCHNH,
or a possible nitrene precursor, H2NCOCH2N (in either a singlet
or a triplet state). However, from ∆SCF calculations at the
1
6,31
to CH2NH.
The vibrationally resolved band at 11.61 eV (VIE
32
equal to AIE) is attributed to HNCO, and a weak feature due

5304 J. Phys. Chem. A, Vol. 108, No. 25, 2004
Dyke et al.
TABLE 3: Significant IR Bands (cm-1) Observed in Matrix
Isolation Studies on the Pyrolysis of 2-Azidoacetamide
24
observed at a position (10.13 eV ) quite different from the
bands of the parent azide.
N
2
matrix
Characterization of the Imine Intermediate H2NCOCHNH.
As indicated above, the IR spectral sequence shown in Figure
4 indicates the formation and decay of an intermediate, I, in
the pyrolysis of 2-azidoacetamide. Also, although this interme-
diate was not detected in the PES experiments, it is significant
that intense bands due to molecular N2 were detected in these
experiments at an early stage in the pyrolysis sequence.
The assignment of I as the imine H2NCOCHNH, formed via
a “type 1” reaction involving the elimination of N2 and a 1,2
H-atom shift, was initially based on the observed growth and
decay patterns shown in Figure 4, and this assignment was
subsequently supported by the results of the simulated IR
spectrum of this compound.
a
(
Figure 3)
P: 3538, 3420, 2126,
719, 1574
I: 3538, 3418, 1728,
previous studies
assignment
N
H
3
2 2
CH CONH
1
2
NCOCHNH
1
1
566, 1399, 1339,
178, 869
C: 2139
A: 3483, 2265
D: 3032, 2919, 1637,
450, 1352, 1127,
065
B: 3287, 747/737
E: 970
CO34
2139
3483, 2254
b
HNCO
3033, 2920, 1637,
1450, 1353, 1128,
1065
3287, 747/737
970 (most intense)
CH
2
NH35
1
1
3
6
HCN
3
7
NH
3
a
Wavenumber accuracy (1 cm-1. b Xe matrix data.30
Figure 6a shows a nitrogen-matrix IR spectrum obtained at
a pyrolysis temperature of 590 K, in which all features except
bands believed to be associated with I have been removed. This
spectrum is unlikely to be the complete spectrum of I but should
be representative of the most intense IR features.
Calculations on the various possible conformers of H2-
NCOCHdNH yielded four distinct structures, each with all real
frequencies at the MP2/6-31G** level on a closed-shell singlet
surface. The two most stable structures, cis-cis (ground-state
-1
configuration) and trans-trans, differ by less than 1.7 kcal‚mol ,
while structures cis-trans and trans-cis lie 5.3 and 7.7
-
1
kcal‚mol , respectively, above the ground state. In these
structures, the description used reflects the relative positions of
the C-H hydrogen to the oxygen atom (cis or trans) and of the
imine hydrogen to the C-H hydrogen (cis or trans).
Figure 6b shows the IR spectrum calculated for the lowest
energy cis-cis conformer of H2NCOCHNH. Although there are
discrepancies in band positions, notably in the higher frequency
region, the overall agreement between this spectrum and that
assigned to species I (Figure 6a) is relatively good, and
comparable with that found between the obserVed and calculated
spectra for the parent azide, (Figure 2). The computed spectra
for the other three conformers show poorer agreement with the
experimental spectrum, particularly in the N-H stretching
-
1
-1
region (3800-3400 cm ) and in the region 1800-1200 cm .
Some differences, however, may be noted between the
observed and cis-cis calculated spectra. In particular, the
observed spectrum (Figure 6a) contains four features in the
-
1
1
500-2000 cm region, while only three are calculated, and
-1
additional weak features in the region 500-1500 cm are also
found in the experimental spectrum. However, all the calculated
bands can be matched to features in the experimental spectrum.
We believe that these differences do not negate the identifica-
tion of I as the imine intermediate, but rather that they suggest
the possible trapping of more than one conformer in the matrix.
Indeed, if the computed spectrum for the second lowest-lying
conformer (the trans-trans structure) is taken into account, there
is ample flexibility to fit the experimental spectrum at this level
of agreement. Support for this view comes from a detailed
comparison of the observed and calculated IR spectra for the
parent azide (Figure 2), where the observed spectrum between
1300 and 1600 cm also shows more features than predicted
by calculation and where there is a similar possibility of more
than one trapped conformer.
Mechanisms of Gas-Phase Decomposition. The PES ex-
periments show that molecular N2 is evolved at an early stage
in the pyrolysis process, whereas the IR matrix-isolation
experiments suggest that the onset of pyrolysis is characterized
by the formation of the imine intermediate, H2NCOCHNH, prior
Figure 5. PE spectra obtained from pyrolysis of 2-azidoacetamide:
a) at 550 K, estimated contributions of species are (white shadow)
-azidoacetamide, (light gray shadow) CH NH, (dark gray shadow)
HNCO, and (black shadow) N and HCN; (b) at 860 K, estimated
contributions of species are (light gray shadow) CH NH, (dark gray
shadow) HNCO, (white shadow) HCN, and (black shadow) N
(
2
2
2
2
2
.
6-31G** level, the photoelectron bands for these species are
-
1
expected to be very close in energy to those of the parent azide,
and they could easily be obscured. The computed values for
the two first VIEs of these compounds are 9.94 and 10.11 eV
1
for H2NCOCHNH in its X A′ state and 9.65 and 10.16 eV for
3
H2NCOCH2N in its X A′′ state. Another potential pyrolysis
product, formamide, (see later) was also not observed in these
experiments. If formamide had been present to any significant
extent, then a vibrationally resolved band should have been

Thermal Decomposition of 2-Azidoacetamide
J. Phys. Chem. A, Vol. 108, No. 25, 2004 5305
Figure 6. Nitrogen matrix IR spectrum (a) of imine intermediate I and calculated IR spectrum for (b) cis-cis conformer, (c) cis-trans conformer,
d) trans-cis conformer, and (e) trans-trans conformer of H NCOCHNH.
(
2
Figure 7. Energy level diagram (at 298 K) showing relative enthalpies of the most stable conformers of 2-azidoacetamide, H
principal decomposition products calculated at the MP2/6-31G** level. The energy of the optimized geometry of the triplet nitrene, obtained from
the parent azide on loss of N , is also shown in this figure. Geometry optimization calculations of the singlet nitrene all converged to the singlet
imine, HNdCHCONH
2
NCOCHNH, and
2
2
.
to the formation of any other product. Also, what is clear is
However, H2NCHO was not detected in either the PES or matrix
IR studies, and we believe that this is due to the thermal
instability of formamide produced under the conditions at which
the intermediate I begins to decompose.
that, in contrast to some earlier systems,2
0-22
CH2NH cannot
be the precursor of the relatively low-temperature formation of
HCN.
The initial thermal decomposition process is therefore con-
sistent with the traditional “type 1” route:
Matrix IR experiments on the pyrolysis of H2NCHO per-
formed in this work clearly established that this compound
partially decomposes into HNCO at ca. 600 K. This observation
is consistent with thermochemical data, from which it is
possible to show that at this temperature, the Gibbs free energy
change is close to zero for the reaction
3
8
H NCOCH N f H NCOCHNH + N
2
2
2
3
2
Subsequent reaction is then determined by the decomposition
of the imine intermediate, and by analogy with the work of Bock
12-16
and Dammel,
one might anticipate elimination of HCN and
H NCHO f HNCO + H
2
2
the production of formamide, H2NCHO.
One might therefore expect to observe at least some H2NCHO
during the pyrolysis sequence.
H NCOCHNH f H NCHO + HCN (route A)
2
2

5306 J. Phys. Chem. A, Vol. 108, No. 25, 2004
Dyke et al.
The fact that formamide was not obserVed at all during azide
pyrolysis implies a very short lifetime under the conditions used.
A similar “premature” decomposition of CH2NH (into HCN)
has previously noted by Bock and Dammel15 in the pyrolysis
of CH3N3. If a similar phenomenon is occurring here, the
observable products of imine decomposition by this route are
therefore expected to be HCN and HNCO. The third product,
molecular H2, is effectively undetectable, as discussed previ-
2
1
ously.
The overall route
H NCOCHNH f HNCO + HCN + H2
2
therefore accounts for the presence of two important high-
temperature products (HNCO and HCN) but does not account
for the formation of CH2NH, CO, or NH3.
Two alternative imine decomposition routes can, however,
be proposed that might account for these products. These may
be termed route B
and route C
Finally, it is possible to postulate a fourth competing reaction
in which azide decomposition follows a “type 2” route, which
does not involve imine intermediate I:
Figure 8. Calculated geometries (at the MP2/6-31G** level) of the
transition states for the reactions (a) azide f imine, (b) imine f CH
NH + HNCO, and (c) imine f H NCHO + HNC.
2
-
2
cis-trans form of the imine, a configuration that is achieved
from the predicted cis-cis ground state by rotation. Decomposi-
tion is predicted to occur initially to give HNC (which
isomerizes to HCN) and H2NCHO (which decomposes to
HNCO + H2). The production of HNC, HNCO, and H2 is thus
very similar to the production of HNC, CO2, and H2 via a
proposed concerted mechanism from HNCHCOOH, which is
thought to be a minor channel in the decomposition of
azidoacetic acid. For route B, decomposition to HNCO + CH2-
NH is predicted to go through a nonplanar transition state, which
can be derived directly from the cis-cis form of the imine I
(Figure 8c).
This route, however, does not account for the relatively “early”
production of HCN in the pyrolysis sequence (e.g., see Figure
4
).
To investigate the various proposed mechanisms of imine
33
decomposition, ab initio pathway calculations were carried out
at the MP2/6-31G** level, and the results are summarized in
Figure 7. Here, the computed relative energies of the reagents,
products, and transition states are depicted for the production
of the imine from the initial azide and for decomposition routes
A and B. Geometry optimization of the singlet nitrene converged
to the singlet imine. Also, the enthalpy for conversion of the
azide to the triplet nitrene and N2 was calculated as +8.5 kcal
The activation barriers for both these transition states are
predicted to be approximately the same (Figure 7), suggesting
that they are both involved simultaneously in imine decomposi-
tion, a prediction that is consistent with the experimental
observations. Route C is also favorable energetically, but
attempts to locate a possible transition state were not successful.
The computed standard relative enthalpies of the products
of mechanisms A-C are displayed on the right-hand side of
Figure 7 and are in the expected order based on the available
-1
mol , higher than that for conversion of the azide to the singlet
imine and N2 (see Figure 7), probably because of stabilization
of the imine by delocalization between the CdO and CHdNH
bonds.
The transition state linking the azide to the imine is shown
in Figure 8a and indicates that the 1,2 H-atom shift occurs before
complete N2 release. For imine decomposition via route A,
which ultimately produces HCN, HNCO, and H2, the transition
state is planar (Figure 8b) and is derived most easily from the
3
8
thermochemical data. This order places H2NCHO + HCN +
N2 lower than HNCO + H2 + HCN + N2, to the extent of ca.
13 kcal/mol. However, when entropy effects at a pyrolysis
temperature of 600 K are taken into account, the Gibbs free

Thermal Decomposition of 2-Azidoacetamide
J. Phys. Chem. A, Vol. 108, No. 25, 2004 5307
energies are almost identical. At higher temperatures, the HNCO
route is lower in free energy.
(5) Schuster, G. B.; Platz, M. S. AdV. Photochem. 1992, 117, 69.
(6) Matsumara, Y.; Shiozawa, T.; Matsushita, H.; Terao, Y. Biol.
Pharm. Bull. 1995, 18, 1805.
7) Tindall, C.; Hemminger, J. Surf. Sci. 1995, 330, 67.
For the decomposition of the imine to the products, the
barriers from the imine to the transition states would be expected
to control the product distribution for several competing routes.
However, for the two transition states located, these are
computed to be very close in energy (see Figure 7). At 600 K,
one might therefore expect to see evidence for routes A, B, and
C. However as discussed previously, we believe that total
decomposition of H2NCHO may be taking place in this system
because of the temperature used and the excess energy arising
from the azide decomposition. It is, however, acknowledged
that the activation energies, computed at the MP2 level and
shown in Figure 7, are likely to be too high because the
transition states are almost certainly better represented by a
multireference method.
(
(8) Bridges, A. S.; Greef, R.; Jonathan, N. B. H.; Morris, A.; Parker,
G. Surf. ReV. Lett. 1994, 1, 573.
(9) Kubota, N. J. Propul. Power 1995, 11, 677.
(
10) Liu, Y. L.; Hsuie, G. H.; Chiu, Y. S. J. Appl. Polym. Sci. 1995, 58,
79.
(11) Badri, M.; Mooney, H. M. Geophysics 1987, 52, 772.
12) Bock, H.; Dammel, R. Angew. Chem., Int. Ed. Engl. 1987, 26, 504.
13) Dammel, R. Ph.D. Thesis, University of Frankfurt, Germany 1985.
14) Bock, H.; Dammel, R.; Aygen, S. J. J. Am. Chem. Soc. 1983, 105,
5
(
(
(
7681.
(
(
(
15) Bock, H.; Dammel, R. J. Am. Chem. Soc. 1988, 110, 5261.
16) Bock, H.; Dammel, R.; Horner, L. Chem. Ber. 1981, 114, 220.
17) (a) Jing, W.; Zheng, S.; Xinjiang, Z.; Xiaojun, Y.; Maofa, G.;
Dianxun, W.; Angew. Chem., Int. Ed. 2001, 40, 3055. (b) Dianxun, W.
Private communication.
18) See, for example: Scrivener, E. F. V., Ed. Azides and Nitrenes;
Academic Press: New York, 1984.
(
(
19) Geiseler, G.; Konig, W. Z. Phys. Chem. 1964, 227, 81.
Conclusions
(20) Dyke, J. M.; Groves, A. P.; Morris, A.; Ogden, J. S.; Dias, A. A.;
Oliveira, A. M. S.; Costa, M. L.; Barros, M. T.; Cabral, M. H.; Moutinho,
A. M. C. J. Am. Chem. Soc. 1997, 119, 6883.
(21) Dyke, J. M.; Groves, A. P.; Morris, A.; Ogden, J. S.; Catarino, M.
I.; Dias, A. A.; Oliveira, A. M. S.; Costa, M. L.; Barros, M. T.; Cabral, M.
H.; Moutinho, A. M. C. J. Phys. Chem. A 1999, 103, 8239.
2
-Azidoacetamide has been synthesized and characterized by
1
13
IR spectroscopy, H and C NMR, and mass spectrometry,
together with IR matrix-isolation and UV photoelectron spec-
troscopy, and these latter techniques were subsequently used
to monitor its thermal decomposition. The experimental results
indicate that the dominant route for decomposition of the azide
is stepwise, involving first the formation of an imine and
molecular nitrogen, followed by at least three imine decomposi-
tion pathways. Ab initio molecular orbital calculations have been
performed, which result in satisfactory agreement between
calculated and experimental IR spectra of both of the parent
azide and the novel imine intermediate, H2NCOCHdNH,
formed initially on decomposition. At higher temperatures, this
intermediate decomposes to a number of simpler molecules,
notably HCN and HNCO, and mechanisms are proposed that
account for the sequence of experimental observations.
(22) Hooper, N.; Beeching, L. J.; Dyke, J. M.; Morris, A.; Ogden, J. S.;
Dias, A. A.; Costa, M. L.; Barros, M. T.; Cabral, M. H.; Moutinho, A. M.
C. J. Phys. Chem. A 2002, 106, 9968.
(
23) Bulgin, D. K.; Dyke, J. M.; Goodfellow, F.; Jonathan, N. B. H.;
Lee, E. P. F.; Morris, A. J. Electron Spectrosc. Relat. Phenom. 1977, 12,
7.
(24) Turner, D. W.; Baker, C.; Baker, A. D.; Brundle, C. R. Molecular
Photoelectron Spectroscopy; Wiley-Interscience: New York, 1971.
25) Basch, H.; Robin, M. B.; Kuebler, N. A.; Baker, C.; Turner, D. W.
J. Chem. Phys. 1969, 51, 52.
26) Robin, M. B.; Brundle, C. R.; Kuebler, N. A.; Ellison, G. B.;
Wiberg, K. B. J. Chem. Phys. 1972, 57, 1758.
6
(
(
(
27) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502.
(28) Pople, J. A.; Scott, A. P.; Wong, M. W.; Radom, L. Isr. J. Chem.
1
993, 33, 345.
(29) Hout, R. F.; Levi, B. A.; Hehre, W. J. J. Comput. Chem. 1982, 3,
234.
(
30) Pettersson, M.; Khriatchtev, L.; Jolkkonen, S.; Rasanen, M. J. Phys.
Chem. A 1999, 103, 9154.
31) Peel, J. B.; Willett, G. D. J. Chem. Soc., Faraday Trans. II 1975,
71, 1799.
32) Eland, J. H. D. Philos. Trans. R. Soc. London, Ser. A 1970, 268,
Acknowledgment. The work reported in this paper was
carried out as part of the Reactive Intermediates RTN EC
Network. Financial support from the Leverhulme Trust is also
acknowledged. The authors are grateful to Dr. E. P. F. Lee for
valuable discussions. Also, support from the POCTI grant (Grant
(
(
7.
8
(
33) Corderio, M. N. D. S.; Dias, A. A.; Costa, M. L.; Gomes, A. N. F.
J. Phys. Chem. A 2001, 105, 3140.
34) Ogden, J. S. Unpublished observations. See also, Dubost, H.;
Abouaf-Marguin, L. Chem. Phys. Lett. 1972, 17, 269.
35) Jacox, M. E.; Milligan, D. E. J. Mol. Spectrosc. 1975, 56, 333.
1
999/FIS/35526) is gratefully acknowledged.
(
References and Notes
(
(
1) Patai, S. The Chemistry of the Azide Group; Interscience: New
(36) King, C. M.; Nixon, E. R. J. Chem. Phys. 1968, 48, 1685.
(37) Cugley, J. A.; Pullen, A. D. E. Spectrochim. Acta A 1973, 29, 1665.
(38) See, for example: NIST Chemistry WebBook; NIST Standard
Reference Database 69, March 2003 release; National Institute for Standards
and Technology: Gaithersburg, MD, 2003. Also: DeFrees, D. J.; Hehre,
W. J. J. Phys. Chem. 1978, 82, 391.
York, 1971.
(
(
(
2) Scriven, E. F. V.; Turnbull, K. Chem. ReV. 1988, 88, 298.
3) Vantinh, V.; Stadlbauer, W. J. Heterocycl. Chem. 1966, 33, 1025.
4) Maurus, R.; Bogumil, R.; Nguyen, N. T.; Mauk, A. G.; Brayer, G.
Biochem. J. 1988, 332, 67.