Thermal decomposition of NCN3 as a high-temperature NCN radical source: Singlet-triplet relaxation and absorption cross section of NCN( 3Σ)

Article abstract of DOI:10.1021/jp1043046

The potential of the thermal decomposition of cyanogen azide (NCN ₃) as a high-temperature cyanonitrene (NCN) source has been investigated in shock tube experiments. Electronic ground-state NCN( ³Σ) radicals have been detected by narrow-bandwidth laser absorption at overlapping transitions belonging to the Q₁ branch of the vibronic ³Σ⁺-³Π subband of the vibrationally hot A³Π_u(010)-X ³Σ_g^-(010) system at ν = 30383.11 cm^-1 (329.1302 nm). High-temperature absorption cross sections σ have been directly measured at total pressures of 0.2-2.5 bar, log[σ/(cm² mol^-1)] = 8.9-8.3 × 10^-4 × T/K (±25%, 750 < T < 2250 K). At these high temperatures, NCN(³Σ) formation is limited by a slow electronic relaxation of the initially formed excited NCN(¹Δ) radical rather than thermal decomposition of NCN₃. Measured temperature-dependent collision-induced intersystem crossing (CIISC) rate constants are best represented by k_CIISC/(cm³ mol^-1 s ^-1) = (1.3 ± 0.5) × 10¹¹ exp[-(21 ± 4) kJ/mol/RT] (740 < T < 1260 K). Nevertheless, stable NCN concentration plateaus have been observed, showing that NCN₃ is an ideal precursor for NCN kinetic experiments behind shock waves.

Full text of DOI:10.1021/jp1043046

J. Phys. Chem. A 2010, 114, 12963–12971
12963
Thermal Decomposition of NCN3 as a High-Temperature NCN Radical Source:
Singlet-Triplet Relaxation and Absorption Cross Section of NCN(3Σ)†
J. Dammeier and G. Friedrichs*
Institut fu¨r Physikalische Chemie, Olshausenstrasse 40, Christian-Albrechts-UniVersita¨t zu Kiel,
D-24098 Kiel, Germany
ReceiVed: May 11, 2010; ReVised Manuscript ReceiVed: June 25, 2010
The potential of the thermal decomposition of cyanogen azide (NCN3) as a high-temperature cyanonitrene
(NCN) source has been investigated in shock tube experiments. Electronic ground-state NCN(3Σ) radicals
have been detected by narrow-bandwidth laser absorption at overlapping transitions belonging to the Q1 branch of
3
+
3
3
3
-
-1
˜
˜
the vibronic Σ - Π subband of the vibrationally hot A Πu(010)-X Σg (010) system at ν˜ ) 30383.11 cm
(329.1302 nm). High-temperature absorption cross sections σ have been directly measured at total pressures of
0.2-2.5 bar, log[σ/(cm2 mol-1)] ) 8.9-8.3 × 10-4 × T/K ((25%, 750 < T < 2250 K). At these high temperatures,
NCN(3Σ) formation is limited by a slow electronic relaxation of the initially formed excited NCN(1∆) radical
rather than thermal decomposition of NCN3. Measured temperature-dependent collision-induced intersystem
crossing (CIISC) rate constants are best represented by kCIISC/(cm3 mol-1 s-1) ) (1.3 ( 0.5) × 1011 exp[-(21
( 4) kJ/mol/RT] (740 < T < 1260 K). Nevertheless, stable NCN concentration plateaus have been observed,
showing that NCN3 is an ideal precursor for NCN kinetic experiments behind shock waves.
I. Introduction
have resolved this issue by showing that another spin-allowed,
endothermic (∆rH° ) 82 kJ/mol) reaction pathway exists. On the
0K
The nitrogen oxides NO and NO2 (NOx) are main pollutants
of combustion processes and play an important role in tropo-
spheric ozone formation as well as stratospheric ozone depletion
by triggering two different chain reaction cycles.1 Additionally,
nitrous oxide (N2O), a secondary product of NOx reactions, is
of environmental concern due to its high global warming
potential.2 From several distinct NO formation pathways,
especially under rich combustion conditions, prompt-NO is
generated by reactions of small hydrocarbon radicals with
nitrogen stemming from the combustion air. Until recently, it
was well accepted that the most important initiation step of
prompt-NO formation, according to the Fenimore mechanism,3
is the spin-forbidden reaction
basis of quantum chemical exploration of the potential energy
surface followed by RRKM and master equation modeling, they
concluded that the NCN-radical-forming channel
CH(2Π) + N2(1Σ+) f H(2S) + NCN(3Σ-g )
(R1b)
dominates the overall reaction. Very recently, another high-
temperature reaction channel passing through an alternative
intermediate structure (HNNC instead of HCNN) has been
discovered.5,11 This pathway eventually yields NCN(3Σ) radicals
and H(2S) atoms as well.
The theoretical predictions of Moskaleva et al. have stimu-
lated much experimental effort in order to clarify the role of
NCN radicals in NO formation processes. First experimental
evidence for the existence of NCN in flames was provided by
LIF detection of NCN reported by Smith12 and, more recently,
by Sutton et al.13 and Klein-Douwel et al.14 In all three studies,
NCN concentrations were found to be strongly correlated with
the corresponding CH signal. However, it was not until the
carefully conducted shock tube experiments of Vasudevan et
al.15 that NCN was shown to be the main product (>70%) of
the reaction CH + N2 at temperatures of 2228 < T < 2905 K.
The branching fraction of reaction R1b was deduced from
detailed modeling of CH concentration profiles measured during
the pyrolysis of C2 H6/Ar mixtures in the presence of N2 behind
shock waves. Vasudevan et al. also detected NCN profiles by
laser absorption, which were used to infer NCN absorption cross
sections at ν˜ ) 30383 cm-1 as well as to estimate the high-
temperature rate constant of the reaction H + NCN. So far, to
the best of our knowledge, the only other experimental study
on NCN high-temperature kinetics (i.e., for temperatures T >
1000 K) was reported by Busch and Olzmann,16,17 who generated
NCN radicals behind shock waves using the fast thermal
CH(2Π) + N2(1Σ+) f N(4S) + HCN(1Σ+)
(R1a)
In subsequent reaction steps, N atoms are then transformed into
NO.4 Assuming that the initially formed energetically excited
addition complex undergoes fast intersystem crossing from the
doublet to quartet potential energy surface, this only slightly
endothermic (∆rH° ) 9 kJ/mol)5 reaction seemed to be a
0K
reasonable candidate for N2 activation. Moreover, measured
activation energies of Ea ≈ 75 kJ/mol were found to be consistent
with a reaction barrier stemming from an elimination step of the
postulated quartet intermediate.6,7 However, detailed calculations
of the intersystem crossing probability led to the conclusion that
the overall reaction rate should be much smaller than that measured
experimentally.8 In the year 2000, Moskaleva et al.9,10 seemed to
† Originally submitted for the “30th Free Radical Symposium” special
section, published in the April 15, 2010, issue of J. Phys. Chem. A (Vol.
114, No. 14).
* To whom correspondence should be addressed. E-mail: friedrichs@
phc.uni-kiel.de.
10.1021/jp1043046  2010 American Chemical Society
Published on Web 07/15/2010

12964 J. Phys. Chem. A, Vol. 114, No. 50, 2010
Dammeier and Friedrichs
decomposition of NCN3 as a NCN radical source. At temper-
atures of 1800 < T < 2900 K, rate constants of the thermal
unimolecular decomposition of NCN were obtained from time-
resolved concentration profiles of the reaction products C and
N. NCN radicals were not detected.
nitrogen mixtures as the driver gas using 30 or 80 µm thick
aluminum diaphragms. Experimental temperatures and pressures
were calculated from pre-shock conditions and the shock wave
velocity, which was measured by a fast count unit wired to four
piezoelectric sensors (PCB Piezotronics M 113A21). The
registered damping of the shock wave on the order of 1% per
meter as well as real gas effects was taken into account when
calculating the shock wave parameters based on a one-
dimensional, frozen chemistry computer code.
Other recent work was concerned with the experimental
investigation of NCN radical reactions at low temperatures.16-22
For example, Baren and Hershberger18 measured the rate
constant of the reaction NCN + NO in the temperature range
of 298 < T < 573 K. NCN radicals were generated by 193 nm
UV flash photolysis of diazomethane in the presence of
cyanogen (CH2N2/C2N2) and were detected by LIF at a
wavenumber of ν˜ ) 30395 cm-1 (λ ≈ 329 nm). Other authors
employed 193 nm photolysis of cyanogen azide (NCN3) as an
efficient NCN source,19,20,22 in which the initially formed elec-
tronically excited NCN(1∆) was subsequently deactivated by
collisions to its NCN(3Σ) ground state.23 Theoretical work on
bimolecular NCN reactions has been mainly performed in the M. C.
Lin group. Mostly scaled to fit low-temperature experimental rate
constant data, high-temperature extrapolations were reported for
the reactions with NO and NS,19,24 NO2,20 O2,25 OH,26 and O.27
Finally, first mechanistic studies on NOx formation by implement-
ing NCN chemistry were conducted by El Bakali et al.,28 Sutton
and Fleming,29 Gersen et al.,30 and Konnov.31
Although these studies serve as a good starting point for future
prompt NO modeling, NCN kinetics model refinements and
theoretical predictions should be based on or validated against
direct high-temperature measurements of NCN rate constants
as well. A prerequisite of such direct measurements is the
availability of a clean NCN source. In this work, the thermal
decomposition of cyanogen azide, NCN3, has been studied with
regard to its potential to serve as such a precursor in shock tube
experiments. According to Bock and Dammel32 and Benard et
al.,33 the initiating pyrolysis step of NCN3 can be assumed to
yield NCN quantitatively
Room-temperature spectra were recorded using a 45 cm long
slow flow photolysis cell equipped with quartz windows. The
193 nm UV photolysis of cyanogen azide in Ar mixtures was
performed collinearly with the detection laser beam by com-
bining and dividing the photolysis and detection laser by an
optical setup with two dichroic mirrors. Although continuously
flushed with pure nitrogen, deposition of polymeric material
took place at the quartz windows, and therefore, regular cleaning
of the cell windows was necessary.
NCN3 Synthesis and Gas Mixtures. Cyanogen azide is an
extremely explosive substance; for that reason, no attempt was
made to condense or to purify the gas after the synthesis. Instead,
care was taken to directly synthesize NCN3 as pure as possible
using a variant of a synthesis described by Milligan et al.35
Basically, solid sodium azide (NaN3) and gaseous cyanogen
bromide reacted to form gaseous cyanogen azide and solid sodium
bromide (NaBr). Cyanogen bromide was purified by resublimation
and further dried by passing it through a molecular sieve (3 Å).
Sodium azide was pestled and then degassed in vacuo overnight
at pressures down to 1 × 10-4 mbar. Small amounts (∼20 mbar)
of gaseous cyanogen bromide were allowed to expand into an
evacuated flask, which held a huge excess of sodium azide. After
8 h of reaction time and repeated careful shaking of the reaction
flask, the gaseous product was let into another pre-evacuated storage
flask. Immediate analysis of the gaseous reaction product by means
of FTIR spectroscopy showed that no water (indicator for a gas
leak) and only negligible amounts of carbon dioxide (∼0.01%)
were present. Levels of remaining cyanogen bromide were always
<3%, but in some syntheses, an overall purity of >99% NCN3 was
achieved. Subsequently, cyanogen azide was diluted with argon,
and the 0.1-0.5% NCN3/Ar mixtures, which were stored in a 10
L glass flask, were used within 3 days after preparation. Since
noticeable decomposition of typically 10% per day was found to
take place, NCN3 concentrations were checked by FT-IR on a daily
basis. For shock tube experiments, the supply mixtures were further
diluted with argon using calibrated mass flow controllers (Aera
FC-7700C). Although no noticeable dependence of the absorption
signal on the sample preparation procedure was found, the shock
tube was flushed with the particular mixture at a pressure of ∼50
mbar to minimize any wall adsorption effects.
NCN3 + M f NCN + N2 + M
(R2)
We detected NCN, which can be assumed to be initially formed
in its first excited electronic singlet state (1∆) due to spin
conservation, in its electronic 3Σ ground state by means of time-
resolved laser absorption spectroscopy at wavenumbers around
ν˜ ) 30395 cm-1. NCN(3Σ) formation rate constants have been
extracted from concentration-time profiles measured during
pyrolysis experiments of NCN3 behind incident and reflected
shock waves. Moreover, absorption cross sections of NCN(3Σ)
have been directly determined from the observed absorption
plateaus at long reaction times. The unimolecular decomposition
of NCN3 was analyzed based on quantum chemical calculations
and statistical rate calculations combined with master equation
modeling. A comparison of predicted and experimental tem-
perature dependences revealed that a rather slow collision-
induced intersystem crossing pathway, NCN(1∆) + M f
NCN(3Σ) + M, plays a decisive role.
Gases and chemicals used were argon (99.999%), hydrogen
and nitrogen (99.99%, as the driver gas), sodium azide (Merck,
99%), and cyanogen bromide (Acros, 97%).
Laser Absorption Setup. NCN was detected in its triplet
ground state at ν˜ ≈ 30395 cm-1 using difference amplification
laser absorption spectroscopy. In order to clarify the role of
singlet NCN relaxation, some measurements were also per-
formed at ν˜ ) 30045.7 cm-1. Tunable narrow-bandwidth (<10
MHz) UV laser radiation was provided by internal frequency
doubling of a frequency-stabilized continuous-wave ring dye
laser (Coherent 899) operated with DCM-Special dye. The dye
laser was pumped by 8 W of a frequency-doubled Nd:YVO4
solid-state laser (Coherent Verdi V10) at a wavelength of 532
nm, resulting in 1.5 mW of UV output. The wavelength of the
laser fundamental was measured by means of an interferometric
II. Experimental Section
Shock Tube and Slow Flow Photolyis Reactor. High-
temperature experiments were carried out in an electropolished
stainless steel shock tube of 8 m overall length.34 The 4.4 m
long test section with an inner diameter of 81 mm could be
evacuated by a combination of oil-free turbomolecular drag and
diaphragm pumps down to pressures of p ≈ 10-7 mbar. The
shock tube was operated by hydrogen, nitrogen, or hydrogen/

Thermal Decomposition of NCN3 as a NCN Radical Source
J. Phys. Chem. A, Vol. 114, No. 50, 2010 12965
type wavemeter referenced to a HeNe laser beam (MetroLux
WL200). The readout of the wavemeter was corrected for a
small systematic wavelength offset of 0.02 cm-1 at ν˜ ) 15129
cm-1, which was determined by measuring the known wave-
lengths of several absorption lines of NH2, resulting in an overall
accuracy of the wavelength determination on the order of 5 ×
10-7 corresponding to ∆ν˜ ≈ (0.015 cm-1 at ν˜ ) 30395 cm-1
.
The UV laser beam was split into the detection and reference
beam by a 50:50 beam splitter plate, where the intensity of the
reference beam could be fine-tuned using a variable neutral
density filter. The detection beam was passed through the
photolysis cell and the shock tube, respectively. The beams were
coupled into two separate optical fibers (Thorlabs BF H22-550),
which were connected to a balanced photodetector and amplifier
(Thorlabs PDB150A-EC). For shock tube measurements, the
resulting difference signal was low-pass-filtered (1.4 MHz),
further amplified (Ortec Fast Preamp 9305, 18 dB), and finally
stored by an analog input board (Measurement Computing PCI-
DAS4020/12, 12 bit, 20 MHz). For the measurement of room-
temperature absorption spectra, the wavelength of the detection
laser was slowly scanned (20 GHz mode-hop free scan range),
and the measured absorption signals were processed in a boxcar
amplifier (Stanford Research Systems, SR250, 235, 280). The
differences of the boxcar integrals measured after and before the
UV photolysis laser pulse were taken as a measure of absorption
and were stored in a digital oscilloscope (LeCroy WaveSurfer 454,
500 MHz, 8 bit), which was synchronized with the scanning laser.
Numerical Methods. Spectral simulations of the NCN
absorption spectra were performed with the PGOPHER program,
which is a general-purpose program for simulating and fitting
rotational, vibrational, and electronic spectra developed by the
Bristol laser group.36 The complex electronic spectrum of NCN
was calculated based on accurate spectroscopic constants taken
from Beaton et al.,37,38 which have been recently recast and
refitted in the work of Lamoureux et al.39 Using the same
spectroscopic parameters as have been used and explained in
the work of Lamoureux et al., it was possible to predict
temperature-dependent spectral contours in order to select a
suitable detection wavelength for high-temperature measure-
ments of NCN concentration-time profiles.
Thermal rate constants of the unimolecular decomposition
of NCN3 were modeled within the framework of statistical rate
theories. Standard RRKM calculations followed by solution of
the energy-grained master equation were performed using the
NIST ChemRate program package that enables one to calculate
time-dependent and steady-state rate coefficients for thermally
or chemically activated unimolecular reactions.40 Input param-
eters for the RRKM model have been obtained from ab initio
quantum chemical calculations on the G3 (energies) and MP2/
6-311+G(d,p) (rotational constants, vibrational frequencies)
levels of theory41 using the program suite Gaussian 03.42 The
harmonic vibrational wavenumbers were scaled by a factor of
0.9496 according to Scott and Radom.43 The transition state was
located using the transient-guided quasi-Newton method (QST3)
and was verified by vibration frequency calculations and intrinsic
reaction coordinate analysis.
Figure 1. Normalized high-resolution spectrum of NCN(3Σ) around
30383 cm-1. Upper graph: Experimental room-temperature spectrum
at p ) 25 mbar in argon. Lower graph: Simulated room (T ) 298 K,
p ) 25 mbar) and high-temperature (T ) 1600 K, p ) 1000 mbar)
spectra. Spectral assignments were taken from Beaton et al.37
Renner-Teller components in the excited state, have to be taken
into account as well. So far, high-temperature detection of NCN
in flames12,13,39 as well as the shock tube work of Vasudevan et
al.15 has been mostly performed at wavenumbers ν˜ around 30385
cm-1. Therefore, as a starting point, we recorded a high-
resolution room-temperature spectrum of NCN at 30381 < ν˜ <
30386 cm-1 following the 193 nm photolysis of NCN3. The
spectrum is shown in the upper graph of Figure 1 and has been
pieced together from six successive 20 GHz scans. The
individual spectral intensities have been rescaled in order to
account for different NCN concentrations associated with
variable photolysis intensities caused by the slowly forming
deposits at the cell windows. Due to the unknown NCN
photolysis quantum yield as well as the variable photolysis
intensities in the reactor, it was not feasible to measure absolute
room-temperature absorption cross sections of NCN. Also note
that the initially formed NCN is NCN(1∆), which is subsequently
deactivated to the triplet ground state. From NCN signal rise
times as well as NCN(1∆) decays measured at 30045.72 cm-1
(332.8261 nm) (Q branch band head of the 1Πu(000)-1∆g(000)
transition),44 we inferred a rather slow electronic relaxation time
constant of krelax(Ar, 298 K) ≈ 8 × 108 cm3 mol-1 s-1 corre-
sponding to τrelax ≈ 300 µs at p ) 100 mbar.
At room temperature, the triplet spectrum is dominated by
the intense but unresolved 3Π1 subband origin of the
3
3
Q
Q
Q
˜
˜
A Πu(000)-X Σg(000) ( P21, Q22, and R23 branches) at 30385
cm-1. Other, more isolated lines can be assigned to branches
3
3
III. Results and Discussion
originating from the Π0 and Π2 subband, where the assign-
ments indicated in Figure 1 were adopted from the work of
Beaton et al.37 Moreover, Beaton and Brown attributed the barely
resolved spectral feature at 30383.1 cm-1 to the Q1 branch of
the vibronic 3Σ+-3Π subband of the vibrationally hot
A. Room-Temperature Photolysis. The NCN radical is
linear in its ground and first excited triplet electronic states.
The electronic spectrum is complicated due to the partial overlap
3
3
˜
˜
of 27 branches originating from the A Πu(000)-X Σg(000)
transition. Especially at high temperatures, spectral contributions
from the (010)-(010) hot band, which is split into three
3
3 -
˜
˜
A Πu(010)-X Σg (010) system of NCN. The lower graph in
Figure 1 illustrates the corresponding PGOPHER simulation of

12966 J. Phys. Chem. A, Vol. 114, No. 50, 2010
Dammeier and Friedrichs
Figure 2. Typical NCN absorption profile measured behind an
incident and reflected shock wave. Experimental conditions: 11 ppm
NCN3; ν˜laser ) 30383.11 cm-1
.
Figure 3. Part of the high-temperature NCN absorption spectrum,
3
3
˜
˜
which is dominated by the Q1 band head of the A Πu(010)-X Σu
(010) transition at ν˜ ) 30383.11 cm-1. Symbols refer to measured
absorbances normalized with respect to the band maximum. The solid
and dashed curves denote original and slightly shifted PGOPHER
simulations, respectively.
the mainly Doppler-broadened room-temperature spectrum. A
high-temperature spectrum calculated for T ) 1600 K at p )
1.0 bar is shown as well. Allowance was made for a small
pressure-broadening effect, which was estimated to be identical
to that previously measured for the three-atom radical NH2 (2.27
× (T/295 K)-0.78 GHz/bar).45 Note that all spectra have been
normalized with respect to the highest-occurring spectral
intensity. Except for minor spectral shifts, the agreement
between the experimental and simulated room-temperature
spectrum is very good, and thus, it can be concluded that the
high-temperature extrapolation is reliable as well. As can be
seen from a comparison of the two simulated spectra in the
lower graph of Figure 1, at high temperatures, the Q1 band head
indicated by a strong Schlieren signal, a fast rise of the
absorption signal was observed that can be attributed to NCN
radical formation. During the 160 µs reaction time behind the
incident wave, a plateau absorption value was reached, indicat-
ing almost complete decomposition of NCN3. An accurate
plateau value was extracted from the absorbance-time profiles
by fitting a first-order production kinetics according to
x ) x0(1 - exp[-k(t - τOffset)])
(1)
at around 30383.1 cm-1 provides much stronger absorption than
3
the congested Π1 subband transitions at around 30385 cm-1
,
which are strongest at room temperature. This finding is in
agreement with the LIF excitation flame spectra reported by
Lamoureux et al.,39 which was measured at a lower spectral
resolution of 0.15 cm-1. Also, Vasudevan et al.15 observed an
absorption maximum at 30383.06 cm-1 in their shock tube work
(T ≈ 2640 K) but mistakenly assigned the observed absorption
where x denotes NCN absorbance, x0 denotes the plateau
absorbance, and k is the first-order rate constant. At lower
experimental temperatures, short induction times were visible
in the NCN signals (due to NCN(1∆) f NCN(3∆) relaxation;
see below). These induction times were excluded from the fit
by offsetting time zero using a suitable τOffset in eq 1. Following
the arrival of the reflected shock wave indicated by the second
Schlieren signal, an almost constant NCN absorbance was ob-
served, showing (i) that at this high temperature, the NCN3
decomposition is complete within the time resolution of the
experiment and (ii) that secondary chemistry of the NCN radical
is almost negligible on the time scale of the experiment. The slow
decay of NCN is most probably due to the onset of the unimolecular
decomposition of NCN at these high temperatures. Despite the
higher total density behind the reflected wave, the observed NCN
absorbance is significantly lower, clearly indicating a strong
negative temperature dependence of the NCN absorption cross
section.
The NCN absorption spectrum has been mapped out via
repeated absorbance measurements by setting the laser to
different detection wavelengths. The resulting NCN spectrum
is plotted in Figure 3. Measured absorbances have been
normalized with respect to initial NCN3 concentrations and are
corrected to average temperatures of T ) 850 and 1600 K by
adopting the temperature dependence of the absorption cross
section outlined below. The solid and dashed curves correspond
to PGOPHER simulations, in which the wavenumber scale of
the latter has been shifted by 0.04 cm-1. Apart from this obvious
wavenumber offset, there is very good agreement between the
3
peak to the Πu(000)-3Σg(000) band head.
It turns out from the PGOPHER simulation shown in Figure
1 that the overall spectral shape of the vibrationally hot Q1 band
head remains narrow even at high temperatures. This causes
problems because quantitative narrow-bandwidth laser absorp-
tion detection of NCN in shock tube experiments relies on a
correct, fixed detection wavelength. The laser has to be set to
this specific wavelength before the experiment; therefore the
exact band head position must be known. Unfortunately, as it
becomes clear from a comparison of the calculated with the
experimental room-temperature spectrum, the heavily congested
spectral region at wavenumbers around 30383 cm-1 is not
quantitatively reproduced by the simulation. The remaining
discrepancies are small but may indicate a significant offset of
the underlying spectral position of the simulated Q1 band.
B. High-Temperature Absorption Cross Section. In order
to measure the exact peak position of the Q1 band head, pyrolysis
experiments of NCN3 have been performed at variable detection
laser frequencies with the temperature set to T ) 850 ( 50 K
behind the incident and T ) 1600 ( 100 K behind the reflected
shock waves. Figure 2 illustrates a typical absorption-time
profile measured for a mixture containing 11 ppm of NCN3 in
Ar. After the arrival of the incident shock wave, which is

Thermal Decomposition of NCN3 as a NCN Radical Source
J. Phys. Chem. A, Vol. 114, No. 50, 2010 12967
trapolated values but are consistent within error limits. As a
comparison of their flame data with the data of Vasudevan et
al., Sutton et al. provided an absorption cross section value of
σ ) 7.6 × 106 cm2/mol at T ) 2490 K (solid diamond), which
is based on LIF measurements of CH, OH, and NCN profiles
in low-pressure methane flames.13 Given the reported uncertainty
estimate of a factor of 3, the nearly perfect agreement with our
measurement is fortuitous. Similarly, Lamoureux et al. presented
a quantitative determination of NCN radical concentration-time
profiles in a low-pressure methane flame, where LIF intensities
have been calibrated using cavity ringdown spectroscopy.39
Laborious spectral simulations, identical to the PGOPHER
simulations performed in this work, were used to calculate
absolute absorption cross section values starting from the
transition dipole moments of the (000) and (010) bands derived
from fluorescence lifetime measurements by Smith et al.46 Less
temperature-dependent absorption cross sections have been
reported for temperatures between 1500 and 1900 K (line
marked with triangles), yielding a factor of 2 higher value at
T ) 1700 K. Altogether, keeping in mind the well-known
difficulties in determining accurate high-temperature absorption
cross sections of transient species, the overall good agreement
of all above-mentioned indirect studies with our first direct
measurement is quite remarkable.
C. Singlet-Triplet Relaxation and Unimolecular Decom-
position of NCN3. Concentration-time profiles of NCN(3Σ)
formation following shock heating of mixtures of 5-20 ppm
of NCN3 in Ar have been recorded at temperatures between
546 and 1257 K at two distinct density ranges of 3.0-4.0 ×
10-6 (F ≈ 3.5 × 10-6 mol/cm3) and 6.5-7.5 × 10-6 mol/cm3
(F ≈ 7.0 × 10-6 mol/cm3). Validation experiments were
performed beforehand in order to exclude interfering absorption
or secondary chemistry of CN radicals that may be formed from
BrCN impurities. Room-temperature photolysis experiments of
pure BrCN did not show visible absorption of CN radicals at
the used detection wavelength nor did shock tube experiments
performed with reaction mixtures containing up to 50% BrCN
yield noticeably different absorption-time traces.
As illustrated in Figure 2, first-order NCN formation rate
constants k were determined by fitting eq 1 to the time profiles.
Short induction times, which have been observed at low
temperatures (e.g., 30 µs at T ≈ 740 K), were excluded from
the fit by slightly offsetting time zero, τOffset. As a potential chain-
accelerated reaction mechanism can be excluded due to the very
low reactant concentrations, these induction times have to be
attributed to relaxation processes. Possible candidates are the
delayed heating of NCN3 behind the shock wave and, probably
more important, the relaxation of the initially formed NCN(1∆)
radicals into the 3Π (010) vibronic state of NCN(3Σ) via
intersystem crossing and vibrational cooling. A kinetic simula-
tion assuming the simple consecutive reaction scheme NCN3
(+M) f NCN(1∆) + N2 (+M) f NCN(3Σ) + N2 (+M)
smoothly reproduced the signals. For example, at T ≈ 820 K,
a ratio of kfast/kslow ≈ 5 was obtained for the two rate constants
describing the two-step process. At not too high temperatures, good
agreement was found between the rate constant of the rate-limiting
step described by kslow and the corresponding rate constant k
obtained by fitting eq 1. However, a systematic evaluation of the
observed induction times turned out to be imprecise because
the observed delays were short and the rate constants of the two
consecutive steps were found to be strongly correlated. More-
over, as a consequence of the simple consecutive mechanism, the
extracted rate constants were interchangeable in the simulations,
and thus, it was not possible to assign their values unambiguously
Figure 4. High-temperature NCN absorption cross section of the Q1
band head at ν˜ ) 30383.11 ( 0.01 cm-1 (λ ) 329.1302 ( 0.0001 nm)
in comparison with literature data.
simulated and experimental band shape. A very slight shift of
the band maximum to lower wavenumbers with increasing
temperature (∼0.01 cm-1/1000 K) is also consistent with our
measurements. Note that the necessary wavenumber offset of
0.04 cm-1 exceeds the (0.015 cm-1 precision of our wavemeter.
Furthermore, the spectral positions measured for the isolated
room-temperature lines shown in Figure 1, which are (0.01
cm-1 within the corresponding line positions reported by Beaton
et al.37 (estimated accuracy (0.006 cm-1), verify the accuracy
of the wavemeter readout. We thus conclude that our measure-
ment of the Q1 band position are accurate and recommend a
high-temperature detection wavelength of ν˜max ) 30383.11 (
0.01 cm-1 (329.1302 ( 0.0001 nm) for future shock tube work.
Assuming quantitative formation of NCN according to
reaction R2 and relying on a known initial concentration of
NCN3 inferred from the FT-IR measurements, the observed
plateau absorptions behind incident and reflected shock waves
could be directly converted to high-temperature NCN absorption
cross sections. Measured temperature-dependent values, which
were obtained from thermal decomposition experiments of
NCN3 between 766 and 2241 K, are shown as open symbols in
Figure 4 and are listed in Table 1. All measurements have been
performed at ν˜ ) 30383.11 cm-1. The strongly temperature-
dependent absorption cross section can be approximated by the
expression (solid line in Figure 4)
σ
log
) 8.9 - 8.4 × 10-4 × T/K
((25%)
2
(
)
cm /mol
where the error represents the combined uncertainties related
to statistical scatter and initial NCN3 concentrations. Comparing
the data points corresponding to measurements behind the
incident wave (0.2 < p < 0.8 bar, open squares) with the data
points taken from measurements behind the reflected wave (0.8
< p < 2.5 bar, open circles), no significant pressure dependence
is noticeable within the scatter of the data.
Our absorption cross section data are compared with literature
values in Figure 4. The line marked with solid circles represents
the data obtained from shock tube measurements reported by
Vasudevan et al.15 They performed indirect measurements of
the absorption cross section at temperatures between 2378 and
2492 K based on quantitative detection of CH profiles and on
the determination of the branching fraction of reaction R1. The
reported values are approximately 60% higher than our ex-

12968 J. Phys. Chem. A, Vol. 114, No. 50, 2010
Dammeier and Friedrichs
TABLE 1: Selected Experimental Dataa
T/K
p/mbar
106 × F/(mol/cm3)
10-4 × k/s-1
x(NCN3)/ppm
10-8 × σ/(cm2/mol)
Incident Shock Wave: F ≈ 3.5 × 10-6 mol/cm3
588
593
623
743
759
766
772
788
815
823
859
864
877
916
920
978
1127
1240
134
136
148
196
202
200
208
212
224
225
243
245
250
290
261
291
351
395
2.75
2.76
2.86
3.17
3.20
3.14
3.24
3.23
3.31
3.30
3.40
3.42
3.43
3.81
3.41
3.58
3.75
3.84
0.07b
0.07b
0.21b
1.82
1.88
-
1.96
1.80
2.41
2.35
2.22
2.14
2.25
4.11
3.34
3.64
6.10
8.51
∼10
∼10
∼10
∼10
∼5
c
c
c
c
c
1.60
1.95
1.71
c
2.41
1.71
c
1.25
c
c
7.14
6.39
6.39
∼5
14.9
6.39
∼10
6.39
∼15
∼30
6.39
1.19
0.955
0.649
7.14
6.39
Incident Shock Wave: F ≈ 7.0 × 10-6 mol/cm3
546
583
672
701
742
764
788
804
816
865
963
993
1073
1153
1155
1257
236
264
335
358
391
408
426
438
451
487
570
590
659
724
726
678
5.19
5.45
5.99
6.14
6.34
6.42
6.51
6.54
6.66
6.77
7.12
7.15
7.39
7.55
7.56
6.49
0.01b
0.14b
1.43b
1.93b
2.99
2.77
3.03
4.46
4.37
3.28
6.80
7.94
5.46
∼10
c
c
c
c
c
c
2.08
c
c
1.94
c
1.55
c
c
c
c
∼10
∼10
∼10
∼10
∼20
5.46
∼10
∼20
5.46
∼10
11.4
∼10
∼15
∼5
11.1
9.98
15.0
∼15
Reflected Shock Wave
1395
1431
1432
1514
1600
1617
1643
1921
2241
773
794
1599
866
953
1588
993
6.66
6.67
13.4
6.88
7.17
14.3
7.27
15.5
8.30
6.39
0.615
0.407
0.606
0.601
0.392
0.462
0.260
0.269
0.115
6.39
5.46
14.86
6.39
5.46
6.39
11.4
7.14
2478
1547
a First-order NCN(3Σ) formation rate constants k and high-temperature NCN absorption cross sections σ at 30383.11 cm-1 b Excluded from
.
the Arrhenius fit of k; see text. c The initial mole fraction was not promptly checked by FT-IR, and hence no absorption cross section is
specified.
to the NCN3 unimolecular decomposition or the relaxation step.
Obtained temperature-dependent first-order rate constants are listed
in Table 1 and are summarized in Figure 5 as filled (F ≈ 3.5 ×
10-6 mol/cm3) and open (F ≈ 7.0 × 10-6 mol/cm3) symbols. Data
points within 740 < T < 1260 K can be approximated by the
following Arrhenius expressions (solid lines)
is consistent with both a unimolecular decomposition close
to its low-pressure regime and a collision-induced relaxation
process. Note that the measured rate constants nearly collapse
at temperatures T < 700 K, causing a distinct kink in the
Arrhenius plot. Obviously, at a temperature of 700 K, the
NCN formation rate changes from being singlet-triplet-
relaxation-rate-limited at high temperatures to being rate-
limited by the thermal decomposition of NCN3 at lower
temperatures. Hence, in order to further support this conclu-
sion, the unimolecular decomposition rate constant of NCN3
in Ar was estimated based on statistical rate theory. On the
basis of quantum chemical input structures and energies,
specific rate constants were calculated using standard RRKM
procedures followed by modeling of the energy-transfer step
by solving the energy-grained master equation. Input param-
eters are summarized in Table 2. A reasonable value for the
k(F ≈ 3.5 × 10-6 mol/cm3)/s-1
(7.8 ( 2) × 105 exp[-(24 ( 2) kJ/mol/RT]
)
k(F ≈ 7.0 × 10-6 mol/cm3)/s-1
)
(1.3 ( 0.3) × 106 exp[-(24 ( 2) kJ/mol/RT]
A clear density dependence of k has been observed, k(F ≈
7.0 × 10-6 mol/cm3)/k(F ≈ 3.5 × 10-6 mol/cm3) ≈ 1.6, which

Thermal Decomposition of NCN3 as a NCN Radical Source
J. Phys. Chem. A, Vol. 114, No. 50, 2010 12969
crossing (CIISC) step and can be subsumed in a bimolecular
rate constant expression according to
kCIISC/(cm3 mol-1 s-1) )
(1.3 ( 0.5) × 1011 exp[-(21 ( 4) kJ/mol/RT]
Of course, the data points close to the intercept point at T ≈
700 K are influenced by both thermal decomposition and
relaxation. Therefore, true CIISC rate constants at these
temperatures can be expected to be slightly higher than those
inferred from our approximate analysis, thus somewhat
reducing the overall temperature dependence of the relaxation
process. More accurate measurements of the CIISC deactiva-
tion process by monitoring NCN(1∆) concentration-time
profiles are underway and will be reported elsewhere.
Depending on the inert collision partner, CIISC rate constants
for small molecules can vary over several orders of magnitude.
For example, room-temperature rate constants for the process
CH2(1A1) + M f CH2(3B1) + M have been reported to be fast
for both Ar (k ) 3.6 × 1012 cm3 mol-1 s-1) and Xe (9.6 × 1012
cm3 mol-1 s-1),49 whereas significantly different rates have been
found for the deactivation of singlet NH, NH(1∆) + M f
NH(1Σ-) + M. Here, the CIISC is slow for Ar (<6 × 108 cm3
mol-1 s-1) but very efficient in the case of Xe (7.2 × 1012 cm3
mol-1 s-1).50,51 Quantitative prediction of CIISC probabilities
remains challenging, and even general trends are difficult to
work out. Pertaining to CH2 deactivation, which has been quite
extensively studied due to the importance of singlet CH2 chemistry
in high-temperature combustion, CIISC rates have been found to
show weak temperature dependences according to kCIISC ∝ Tn with
0.5 < n < 1.0.52 Hence, with k ) 1.0 × 1010 cm3 mol-1 s-1 at
T ) 1000 K, the NCN relaxation rates found in this work are
at the lower bound of the expected range but exhibit an
unusually strong temperature dependence. Additional measure-
ments using different collision bath gases, which would have
been beyond the scope of this work, are helpful to further
elucidate the nature of the relaxation process.
Figure 5. Experimental rate constant data measured at two different
total density ranges (symbols), fitted Arrhenius expressions (solid lines),
and RRKM/ME prediction of NCN3 unimolecular decomposition.
TABLE 2: Input Parameters for the RRKM/ME Model of
NCN3 Unimolecular Decomposition
NCN3
NCN‡3
°
∆fHT)0
511 kJ/mol
155
394
620 kJ/mol
71
107
vibrations, ν˜/cm-1
431
235
433
471
633
531
876
1256
2079
1056
1973
2032
2193
rotational constants/cm-1
0.0956
0.1030
1.3339
9138 cm-1
0.8
0.0798
0.0870
0.9557
critical energy, E0
energy transfer efficiency, ꢀE
energy transfer, 〈∆E〉all
-160 cm-1
444 kJ/mol
565 kJ/mol
∆fHT)0(NCN(3Σ))
°
∆fHT)0(NCN(1∆))
°
Finally, the role of a possible but spin-forbidden NCN3
decomposition path directly yielding NCN(3Σ) should briefly
be discussed. Benard et al.33 performed quantum chemical
calculations of the ground-state singlet and excited triplet
potential energy surfaces of NCN3 at the B3LYP/6-311+G(2df)
level of theory. They reported a singlet dissociation barrier for
the reaction NCN3(1A′) f NCN(1∆) + N2(1Σ) of E0 ) 114 kJ/
mol, which is in good agreement with our G3 result of E0 )
109 kJ/mol. The triplet manifold, which is repulsive and yields
NCN(3Σ) + N2(1Σ), was found to cross the singlet manifold
near the maximum of the dissociation barrier. In order to
compete with the spin-allowed NCN(1∆) path, the crossing point
would have to be located much below the singlet dissociation
barrier because the singlet-triplet crossing of NCN3 prior to
dissociation can be assumed to be slow. Note that also the
observed induction times at low temperatures are less compatible
with a markedly direct formation of NCN(3Σ). Therefore, we
conclude that nonadiabatic branching to ground-state NCN(3Σ)
does not play a significant role.
average energy transferred per collision, 〈∆E〉all ) -160 cm-1
at E0 ) 109 kJ/mol, was based on the simplified ergodic
collision theory on intermolecular energy transfer (ECT).47
The energy-transfer efficiency parameter ꢀE for argon was
set to a reasonable value of 0.8,48 thus assuming that the
energy is almost completely freely distributed among all
degrees of freedom of both collision partners. Modeling
results are shown in Figure 5 as dashed curves.
The calculated activation energy of Ea ) 93 kJ/mol, which
is significantly lower than E0, and the pronounced density
dependence (factor of 1.8 for twice the density) indicate a
unimolecular decomposition close to the low-pressure limit.
Good agreement is found between the calculated unimolecular
decomposition rate constants and the few experimental data
points at T < 700 K. Of course, the absolute values of the
predicted rate constants are critically dependent on the exact
values of the estimated collision efficiency ꢀE and the G3
energy barrier height. However, as the low-temperature
experimental data are also less accurate due to the observed
slow NCN formation rates, no attempt was made to further
improve the agreement between calculation and experiment.
Clearly, at T < 700 K, the rate-limiting step is the decom-
position of NCN3. In this spirit, the experimental data at T >
700 K reflect the rate of the collision-induced intersystem
IV. Conclusion
On the basis of quantitative formation of NCN during the
thermal decomposition of NCN3, we were able to directly
measure high-temperature absorption cross sections of ground-
state NCN(3Σ). At high temperatures, the (010) vibrational level
is substantially populated and provides stronger absorption than

12970 J. Phys. Chem. A, Vol. 114, No. 50, 2010
Dammeier and Friedrichs
the corresponding (000)-(000) transitions. For quantitative
absorption detection of NCN behind shock waves and in flames,
we recommend the use of the spectral feature centered around
30383.11 cm-1 (329.1302 nm) corresponding to the Q1 branch
of the vibronic 3Σ+-3Π subband of the vibrationally hot
energy surfaces. These can result in both completely different
reaction products as a result of separated singlet and triplet
reactivities (e.g., for the reaction NCN(1∆,3Σ) + O(3P)) as well
as variable dynamics for a reaction on a shared potential energy
surface (e.g., in case of NCN(1∆,3Σ) + NO2(2A1)).20,27 Shock
tube measurements of the temperature dependences of both
singlet and triplet bimolecular reactions are currently underway
and will contribute to an improved understanding of the role of
NCN chemistry for NOx formation in combustion processes.
Preliminary data on the formation and decay of NCN(1∆) during
the thermal decomposition of NCN3 are consistent with the
conclusion drawn in this paper, namely, that CIISC is the rate-
determining step for NCN(3Σ) formation at high temperatures.
3
3 -
˜
˜
A Πu(010)-X Σg (010) system. A strongly temperature-depend-
ent narrow-bandwidth absorption cross section
σ
log
) 8.9 - 8.3 × 10-4 × T/K
((25%)
2
(
)
cm /mol
is recommended for temperatures within 750 < T < 2250 K.
The absorption band shape is dominated by Doppler broadening
of the underlying transitions, and no significant pressure
broadening effect has been detected at pressures between 0.2
and 2.5 bar.
Acknowledgment. This work was supported by the Deutsche
Forschungsgemeinschaft. We would like to thank Nancy
Fassheber and Benjamin Oden for their active help with careful
wavemeter calibration, high-resolution NCN absorption spectra
acquisition, and preliminary measurement of room-temperature
NCN relaxation rates. Many thanks to Xavier Mercier and Colin
Western for providing their PGOPHER input file for NCN
spectral simulations. We further acknowledge Friedrich Temps
for helpful discussions and advice.
The stable NCN plateau signals observed behind the reflected
shock waves clearly reveal the fact that the thermal decomposi-
tion of NCN3 serves as an almost ideal high-temperature source
of NCN(3Σ). As illustrated for an experiment with an initial
NCN3 mole fraction of 11 ppm in Figure 2, on the one hand,
excellent signal-to-noise ratios are attainable. On the other hand,
the resulting NCN concentrations are still low enough to make
possible the investigation of bimolecular NCN reactions by
means of simple pseudo-first-order kinetics using an excess of
the respective reactant. Care should be taken, however, to ensure
complete singlet-triplet relaxation of the initially formed
NCN(1∆) radicals, which, at temperatures between 740 and 1260
K, is best represented by the expression
References and Notes
(1) Warnatz, J.; Maas, U.; Dibble, R. W. Combustion, 3rd ed.; Springer:
Berlin/Heidelberg, Germany, 2001.
(2) Forster, P.; Ramaswamy, V.; Artaxo, P.; Berntsen, T.; Betts, R.;
Fahey, D.; Haywood, J.; Lean, J.; Lowe, D.; Myhre, G.; Nganga, J.; Prinn,
R.; Raga, G.; Schulz, M.; Dorland, R. V. Climate Change 2007: The
Physical Science Basis. Contribution of Working Group I to the Fourth
Assessment Report of the IntergoVernmental Panel on Climate Change;
Solomon, S., Qin, D., Manning, M., Chen, Z., Averyt, M. M. K., Tignor,
M., Miller, H., Eds.; Cambridge University Press: Cambridge, U.K., 2007.
(3) Fenimore, C. P. Proc. Combust. Inst. 1971, 13, 373.
(4) Miller, J. A.; Bowman, C. T. Prog. Energy Combust. Sci. 1989,
15, 287.
kCIISC/(cm3 mol-1 s-1) )
(1.3 ( 0.5) × 1011 exp[-(21 ( 4) kJ/mol/RT]
(5) Harding, L. B.; Klippenstein, S. J.; Miller, J. A. J. Phys. Chem. A
2008, 112, 522.
At p ) 1 bar, relaxation times of τCIISC < 10 µs are obtained for
temperatures T > 875 K, which should be sufficiently short for
most kinetic experiments. A useful upper temperature limit is
set by the thermal stability of NCN. From an experiment
performed at T ) 2303 K and p ) 3.2 bar, which already
showed a fast decay of NCN behind the reflected shock wave,
we deduced a unimolecular decomposition rate constant for
(6) Lindackers, D.; Burmeister, M.; Roth, P. Proc. Combust. Inst. 1990,
23, 251.
(7) Dean, A. J.; Hanson, R. K.; Bowman, C. T. Proc. Combust. Inst.
1990, 23, 259.
(8) Cui, Q.; Morokuma, K.; Bowman, J. M.; Klippenstein, S. J.
J. Chem. Phys. 1999, 110, 9469.
(9) Moskaleva, L. V.; Lin, M. C. Proc. Combust. Inst. 2000, 28, 2393.
(10) Moskaleva, L. V.; Xia, W. S.; Lin, M. C. Chem. Phys. Lett. 2000,
331, 269.
(11) Berman, M. R.; Tsuchiya, T.; Gregusova, A.; Perara, A.; Bartlett,
R. J. J. Phys. Chem. A 2007, 111, 6894.
(12) Smith, G. P. Chem. Phys. Lett. 2003, 367, 541.
(13) Sutton, J. A.; Williams, B. A.; Fleming, J. W. Combust. Flame
2008, 153, 465.
(14) Klein-Douwel, R. J. H.; Dam, N. J.; ter Meulen, J. J. Opt. Lett.
2008, 33, 2620.
(15) Vasudevan, V.; Hanson, R. K.; Bowman, C. T.; Golden, D. M.;
Davidson, D. F. J. Phys. Chem. A 2007, 111, 11818.
(16) Busch, A.; Olzmann, M. Shock-Tube Study of the Thermal
Decomposition of NCN, paper p810138; Proceedings of the European
Combustion Meeting, Vienna, Austria, 2009.
(17) Busch, A.; Gonza´lez-Garc´ıa, N.; Olzmann, M. Thermal Decomposi-
tion of NCN: Shock-Tube Study and Master-Equation Modeling, poster
abstract; 30th International Symposium on Free Radicals, Savonlinna,
Finland, 2009.
(18) Baren, R. E.; Hershberger, J. F. J. Phys. Chem. A 2002, 106, 11093.
(19) Huang, C.-L.; Tseng, S. Y.; Wang, T. Y.; Wang, N. S.; Xu, Z. F.;
Lin, M. C. J. Chem. Phys. 2005, 122, 184321.
(20) Yang, T.-J.; Wang, N. S.; Lee, L. C.; Xu, Z. F.; Lin, M. C. J.
Phys. Chem. A 2008, 112, 10185.
(21) Tzeng, S.-Y.; Chen, P.-H.; Wang, N. S.; Lee, L. C.; Xu, Z. F.;
Lin, M. C. J. Phys. Chem. A 2009, 113, 6314.
(22) Welz, O.; Olzmann, M. An experimental and theoretical study on
the kinetics of the NCN + NO reaction, Montreal, poster W5P089; 32nd
International Symposium on Combustion, Montreal, Canada, 2008.
(23) Kroto, H. W.; Morgan, T. F.; Sheena, H. H. Trans. Faraday Soc.
1970, 66, 2237.
NCN + M f C + N2 + M of k ) 2.0 × 109 cm3 mol-1 s-1
.
This value is in good agreement with the recent experimental
determination of Busch and Olzmann, who performed atomic
resonance absorption spectroscopy (ARAS) of C and N atoms
following the thermal decomposition of NCN3 at temperatures
above 1800 K.16 In agreement with the theoretical prediction
of Moskaleva and Lin,53 C atoms were observed as the main
cleavage product. Relying on the reported rate expression of
Busch and Olzmann, a feasible upper temperature limit of T ≈
2350 K (corresponding to a 100 µs NCN lifetime at p ) 1 bar)
can be specified for studying bimolecular reactions behind shock
waves with NCN3 as the NCN precursor.
Though interfering with kinetic NCN(3Σ) measurements, a
benefit of the rather low CIISC rates and with it the long
lifetimes of NCN(1∆) behind shock waves is that it opens the
door for studying singlet NCN chemistry. This is in in contrast
with singlet CH2 radical chemistry, for example, which is
difficult to assess in shock tube experiments due to the very
fast singlet-triplet deactivation and hence thermal equilibration
with the triplet ground sate.54,55 Comparative studies of singlet/
triplet reactivity of NCN hold the potential to map out different
reaction pathways and pecularities of the underlying potential

Thermal Decomposition of NCN3 as a NCN Radical Source
J. Phys. Chem. A, Vol. 114, No. 50, 2010 12971
(24) Chen, H.-T.; Ho, J.-J. J. Phys. Chem. A 2005, 109, 2564.
(25) Zhu, R. S.; Lin, M. C. Int. J. Chem. Kinet. 2005, 37, 593.
(26) Zhu, R. S.; Nguyen, M. T.; Lin, M. C. J. Phys. Chem. A 2009,
113, 298.
(27) Zhu, R. S.; Lin, M. C. J. Phys. Chem. A 2007, 111, 6766.
(28) El Bakali, A.; Pillier, L.; Desgroux, P.; Lefort, B.; Gasnot, L.;
Pauwels, J. F.; da Costa, I. Fuel 2006, 85, 896.
(29) Sutton, J. A.; Fleming, J. W. Combust. Flame 2008, 154, 630.
(30) Gersen, S.; Mokhov, A. V.; Levinsky, H. B. Combust. Flame 2008,
155, 267.
(31) Konnov, A. A. Combust. Flame 2009, 156, 2093.
(32) Bock, H.; Dammel, R. Angew. Chem., Int. Ed. Engl. 1987, 26, 504.
(33) Benard, D. J.; Linnen, C.; Harker, A.; Michels, J. J.; Addison, J. B.;
Ondercin, R. J. Phys. Chem. B 1998, 102, 6010.
(34) Colberg, M.; Friedrichs, G. J. Phys. Chem. A 2006, 110, 160.
(35) Milligan, D. E.; Jacox, M. E.; Bass, A. M. J. Chem. Phys. 1965,
43, 3149.
(36) Western, C. M. PGOPHER, A Program for Simulating Rotational
Structure, Version 5.2.343; University of Bristol, U.K., June 14, 2007;
available at http://pgopher.chm.bris.ac.uk.
(37) Beaton, S. A.; Ito, Y.; Brown, J. M. J. Mol. Spectrosc. 1996, 178,
99.
(38) Beaton, S. A.; Brown, J. M. J. Mol. Spectrosc. 1997, 183, 347.
(39) Lamoureux, N.; Mercier, X.; Western, C.; Pauwels, J. F.; Desgroux,
P. Proc. Combust. Inst. 2009, 32, 937.
Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,
X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.;
Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;
Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.;
Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich,
S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.;
Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson,
B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03,
revision D.01; Gaussian, Inc.: Wallingford, CT, 2004.
(43) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502.
(44) Kroto, H. W. Can. J. Phys. 1967, 45, 1439.
(45) Friedrichs, G.; Colberg, M.; Fikri, M.; Huang, Z.; Neumann, J.;
Temps, F. J. Phys. Chem. A 2005, 109, 4785.
(46) Smith, G. P.; Coppeland, R. A.; Crosley, D. R. J. Chem. Phys.
1989, 91, 1987.
(47) Nilsson, D.; Nordholm, S. J. Chem. Phys. 2002, 116, 7041.
(48) Nilsson, D.; Nordholm, S. J. Phys. Chem. A 2006, 110, 3289.
(49) Ashfold, M. N. R.; Fullstone, M. A.; Hancock, G.; Ketley, G. W.
Chem. Phys. 1981, 55, 245.
(50) Rohrer, F.; Stuhl, F. Chem. Phys. Lett. 1984, 111, 234.
(51) Hack, W.; Wilms, A. J. Phys. Chem. 1989, 93, 3540.
(52) Bley, U.; Koch, M.; Temps, F.; Wagner, H. Gg. Ber. Bunsen-Ges.
Phys. Chem. 1989, 93, 833.
(40) Mokrushin, V.; Bedanov, V.; Tsang, W.; Zachariah, M.; Knyazev,
V. ChemRate, Version 1.5.2; National Institute of Standards and Technology:
Gaithersburg, MD, 1996-2006.
(41) Curtiss, L. A.; Raghavachari, K.; Redfern, P. C.; Rassolov, V.;
Pople, J. A. J. Chem. Phys. 1998, 109, 7764.
(42) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.;
Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
(53) Moskaleva, L. V.; Lin, M. C. J. Phys. Chem. A 2001, 105, 4156.
(54) Friedrichs, G.; Wagner, H. Gg. Z. Phys. Chem. 2000, 214, 1723.
(55) Friedrichs, G.; Wagner, H. Gg. Z. Phys. Chem. 2001, 215, 1601.
JP1043046