Azide Pyrolyses
FULL PAPER
while a quartet centred at d=4.27 ppm is assigned to the methylene
group. In the 13C NMR spectrum, run in CDCl3 solution, three peaks
were observed: a peak at d=13.9 ppm with respect to TMS due to the
methyl carbon atom, a peak at d=64.6 ppm assigned to the methylene
carbon, with the third at d=157.3 ppm assigned to the carbonyl carbon
atom. The IR spectrum of the pure compound recorded between KBr
For studies on this compound, the sample was contained in two small
glass vials placed in a region immediately above the pyrolysis zone on a
small pad of glass wool, where the temperature is high enough to give a
suitable vapour pressure. The inlet system itself consisted of two coaxial
quartz tubes, and pyrolysis took place in the final few centimetres of the
inner tube of this inlet system, above the photoionisation region in the
spectrometer. With this arrangement, run times of about 30 min could be
achieved, and it was possible to reach a temperature of 5508C without
any major loss of resolution. The distance between the end of the pyroly-
sis region and the photoelectron beam was between 1 and 2 cm, which at
the typical pressure of about 10ꢀ4 mbar normally achieved in the experi-
ments, corresponds to a flight time between the centre of the pyrolysis
region and the photoelectron beam of about 5–10 ms.
plates showed peaks at 2986 cmꢀ1, assigned to C H stretching absorp-
ꢀ
tions, and strong bands at 1724 cmꢀ1 (C=O stretching), 1220 cmꢀ1, and
1021 cmꢀ1. A doublet at 2179/2132 cmꢀ1 was assigned to the N3 group.
2-azido-N,N-dimethylacetamide was similarly prepared from the chloro
derivative. In
a typical preparation, 2-chloro-N,N-dimethylacetamide
(FW 121.57) was added slowly to a solution of sodium azide (FW 65.01;
3 equiv) in water, and the mixture stirred for 2 h in an oil bath at 608C.
After cooling, the product was extracted with ethyl acetate, and the or-
ganic phase dried over anhydrous sodium sulfate. The solvent was then
removed using a rotary evaporator, and the desired product purified by
distillation at reduced pressure.
Calibration of the vertical ionisation energies (VIEs) of the parent azide
photoelectron bands was achieved by introducing methyl iodide and
argon into the ionisation chamber together with the parent azide. The re-
lease of N2 and other clearly identifiable pyrolysis products also served
to provide internal spectral calibrants.
At room temperature, 2-azido-N,N-dimethylacetamide (N3CH2CONMe2)
is a viscous liquid with a low vapor pressure. It was characterized in the
vapour phase by UV-photoelectron spectroscopy and electron impact
mass spectrometry, and in the liquid phase by 1H and 13C NMR and by
IR spectroscopy. The nitrogen matrix IR spectrum was also recorded.
The absence of detectable impurities arising from the starting materials
used in the preparation was confirmed by running PE spectra of the vola-
tile reagents used, and, in order to assist assignment of the bands appear-
ing on pyrolysis, PE spectra were acquired for 2-oxazolidone and
HCONMe2.
The 70-eV electron impact mass spectrum displayed a parent peak at
128 amu (3.4%) and prominent peaks at 28, (N2+, CH2N+, 100%); 72,
(CONMe2+, 100%), 43, (HNCO, MeN2, MeCH2N, 22.2%), (42, 18.9%),
and 99, (NCH2CONMeCH2+, 28.9%). Signals were also present at 15
(Me+) and 85 (CHCONMe2+) amu. As found previously for azidoaceta-
mide ,[11] the 20 eV electron impact mass spectrum showed enhanced in-
Molecular orbital calculations
Molecular orbital calculations were carried out with the Gaussian98 pro-
gram on methyl azidoformate, ethylazidoformate, 2-azido-N,N-dimethyl-
acetamide and dimethyliminoacetamide, Me2NCOCH=NH, at the MP2/
6–31G** level to establish equilibrium geometries, and subsequently to
calculate vertical ionisation energies (VIEs) and infrared frequencies and
intensities. For the VIEs, Koopmans’ theorem was applied to the SCF or-
bital energies obtained at the MP2/6–31G** geometry and the values ob-
tained were scaled[13, 14] by a factor of 0.92. Harmonic vibrational frequen-
cies were calculated at the MP2/6–31G** level by second-derivative cal-
culations. These frequencies are expected to be higher than the experi-
mental values not only because no anharmonicity correction was intro-
duced but also because in the method used only partial allowance was
made for electron correlation.
+
tensities for most of the above ions with respect to the N2 signal. The
1H NMR spectrum, recorded in CDCl3 solution showed a partially re-
solved doublet at d=2.69 ppm (relative to TMS) which was assigned to
two chemically inequivalent methyl groups; and a singlet at d=3.68 ppm
due to the methylene protons. The intensity ratio between the two
groups was 3:1, consistent with the relative intensity expected for (CH3)2
and CH2 protons. In the 13C-{H} NMR spectrum, (also in CDCl3) so-
lution, a doublet peak at d=35.6 and 36.3 ppm (relative to TMS) is as-
signed to the methyl groups; a peak at d=50.4 ppm to the methylene
carbon, and a feature at d=167.3 ppm to the carbonyl carbon.
The IR spectrum of the liquid recorded between KBr plates showed a
relatively broad band with peaks at 2934 cmꢀ1 and 2925 cmꢀ1, (C H
ꢀ
stretch), 2107 cmꢀ1 (N3 stretch), 1660 cmꢀ1 (C=O stretch), 1403, 1284 and
Results and Discussion
1147 cmꢀ1
.
Calculated equilibrium geometries: assignment of PE spectra
Matrix isolation IR studies
The methodology of our matrix isolation infrared experiments followed a
very similar pattern to that described in previous studies on organic
azides.[1–3,11] The inlet and pyrolysis parts of the apparatus were identical,
as were the low temperature Displex and IR spectrometers. Spectra of
the parent azides and of their decomposition products were obtained in
nitrogen matrices, and supporting N2 matrix experiments were also car-
ried out on 2-oxazolidone, MeNCO, Me2NH and HCONMe2 to augment
our N2 matrix infrared data bank of known molecular vibration frequen-
cies. Matrix ratios were estimated to be well in excess of 1000:1. Deposi-
tion times were typically 30 mins at a particular superheater temperature,
and any changes occurring during this period were monitored by spectral
subtraction. Spectral subtraction was also employed to remove the three
(weak) IR bands of matrix-isolated H2O, which were routinely observed
in all experiments, and which are believed to arise from (variable) traces
of water adsorbed on the inner walls of the sample inlet system.
Methyl azidoformate: For methyl azidoformate, four mini-
mum-energy conformers were located for the closed-shell
singlet state, depending on the relative positions of the car-
bonyl, methyl and azide groups. The three most stable struc-
tures lie within 4.3 kcalmolꢀ1, whilst the fourth lies
12.7 kcalmolꢀ1 higher in energy than the most stable con-
former. Because of the small differences in energy between
the three lowest conformers, it is likely that all three struc-
tures contribute to the experimental PE spectrum. The ex-
perimental photoelectron spectrum was assigned by apply-
ing Koopmans’ theorem to the computed orbital energies
for the lowest energy conformer, although the valence VIEs
computed for the three lowest energy structures were found
to be very similar.
Photoelectron spectroscopy
The PE spectrometer used a HeI photon source, and its mode of opera-
tion has been discussed elsewhere.[12] However, although the azidofor-
mates studied here have sufficient vapour pressure to be introduced di-
rectly into the ionisation chamber via a needle valve, 2-azido-N,N-di-
methylacetamide is a liquid which does not have a sufficiently high
vapour pressure at room temperature to allow PE spectra with sufficient
signal-to-noise to be obtained in this way.
Ethyl azidoformate: For the ethyl azidoformate, seven mini-
mum-energy conformers were located for the closed-shell
singlet state, depending on the different relative positions of
the carbonyl, methyl, methylene and azide groups. The
Chem. Eur. J. 2005, 11, 1665 – 1676
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1667
Dyke, John M.
