Temperature dependence of the product branching ratio of the CN + O2 reaction

Article abstract of DOI:10.1021/jp990373q

The reaction of CN radicals with O₂ was studied using infrared diode laser absorption spectroscopy. CO and CO₂ products were detected directly, while the yield of NCO products was inferred by measuring the N₂O yield upon addition of excess NO. Experiments and kinetic modeling calculations were performed to examine the extent of secondary chemistry in this system. The following branching ratios of the CN + O₂ reaction at 296 K were determined: ?(N+CO₂) = 0.02 ± 0.01 and ?(CO+NO) = 0.22 ± 0.02. The branching ratio into the CO + NO channel has a strong negative temperature dependence over the range 239-643 K.

Full text of DOI:10.1021/jp990373q

J. Phys. Chem. A 1999, 103, 3721-3725
3721
Temperature Dependence of the Product Branching Ratio of the CN + O2 Reaction
Kwang Taeg Rim and John F. Hershberger*
Department of Chemistry, North Dakota State UniVersity, Fargo, North Dakota 58105
ReceiVed: February 1, 1999; In Final Form: March 16, 1999
The reaction of CN radicals with O2 was studied using infrared diode laser absorption spectroscopy. CO and
CO2 products were detected directly, while the yield of NCO products was inferred by measuring the N2O
yield upon addition of excess NO. Experiments and kinetic modeling calculations were performed to examine
the extent of secondary chemistry in this system. The following branching ratios of the CN + O2 reaction at
296 K were determined: φ(N+CO2) ) 0.02 ( 0.01 and φ(CO+NO) ) 0.22 ( 0.02. The branching ratio into
the CO + NO channel has a strong negative temperature dependence over the range 239-643 K.
Introduction
channel 1b contributes ∼6% of the total reaction at 298 K.3
More recently, in our laboratory, we used infrared diode laser
The chemical kinetics of the CN radical are of substantial
importance because of the role this species plays in combustion
chemistry. For example, CN radicals are intermediates in the
oxidation of HCN by OH radicals, which is an important part
of NOx formation mechanisms in both fuel-rich hydrocarbon
flames (the prompt-NO mechanism) and flames containing fuel
nitrogen.1 The CN + O2 reaction has been widely studied
previously by both experimental2-17 and computational18-20
methods. Part of the interest is due to the fact that this is one of
the simplest radical-radical reactions known, proceeding over
an attractive potential energy surface without a barrier in the
entrance channel.17,21-22 The total rate constant has been
measured by numerous groups,5-16 which together span the
temperature range 13-3800 K, wider than any other reaction.
Recent reviews of this kinetic data may be found in refs 17 and
23. The Baulch review recommends k ) (1.20 × 10-11) exp-
(210/T) cm3 molecule-1 s-1 over the temperature range 290-
4500 K.23 Agreement between different groups is good although
not perfect.
absorption to estimate φ1b ) 0.23 ( 0.10 at 298 K, with a very
significant temperature dependence.24 Mohammad et al. used
infrared emission spectroscopy and reported φ1b ) 0.29 (
0.02,25 in reasonable agreement with our results. The most
obvious route to channel 1b involves a four-centered transition
state. MP2 calculations of Mohammad et al. indicated that such
a transition state is energetically unfavorable, and they proposed
a dynamic mechanism for CO + NO formation in which NCO
+ O are initially formed with a high degree of NCO rotational
excitation followed by N-atom abstraction as the complex
dissociates.25 No report of any contribution by channel 1c has
previously appeared.
Our earlier work24 on this reaction suffered from large
uncertainties. That study was primarily concerned with a
measurement of the branching ratio of NCO + NO:
NCO + NO f N2O + CO
f N2 + CO2
(2a)
(2b)
In contrast to the wealth of total rate constant data, much
less information is available regarding product branching ratios
of this reaction. This is an important question, leading to
significant uncertainties in modeling. For example, the temper-
ature dependence of the branching ratio is a crucial parameter
in recent attempts to evaluate empirical rate laws of the low-
temperature kinetics of this reaction.22 Three exothermic product
channels are possible:
where NCO was formed in reaction 1a by photolyzing a CN-
radical precursor in the presence of O2. The existence of channel
1b was inferred by the observation of a modest excess of CO
yields over N2O yields in this system. Determination of φ1b in
this way required subtracting two comparable product yields,
leading to large error bars. The observation of a large temper-
ature dependence in φ1b was therefore tentative. In addition, all
previous experimental reports of φ1b, both in our laboratory and
in others, have suffered from the possibility that some of the
observed CO may have been formed by secondary chemistry,
of which the most serious is the NCO + O reaction:
CN + O2 f NCO + O
f CO + NO
∆H ) -56 kJ/mol
∆H ) -456 kJ/mol
(1a)
(1b)
(1c)
NCO + O f CO + NO
f N + CO2
(3a)
(3b)
f N + CO2
∆H ) -354 kJ/mol
where the thermochemical data from ref 17 was used. Channel
1a has generally been considered the most important product
channel even though it is least exothermic because only an O-O
bond breaking of an initially formed NCOO collision complex
is required.
Several reports of the existence of channel 1b as a minor
pathway have appeared. In an early report, Schmatjko and
Wolfrum used CO laser absorption spectroscopy to estimate that
The experiments reported here are similar but not identical
to our previous approach. First, we obtain φ1b without the
subtraction of two comparable numbers inherent in our previous
determination and therefore obtain substantially improved
precision. Second, we investigate in greater detail the possible
role of secondary chemistry and, in particular, demonstrate that
our results are not affected by reaction 3. These improvements
10.1021/jp990373q CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/15/1999

3722 J. Phys. Chem. A, Vol. 103, No. 19, 1999
Rim and Hershberger
permit a more precise determination of the temperature depen-
dence of the branching ratio.
Experimental Section
The experimental procedure is similar to that described in
previous publications.24,26 A schematic of the experimental
apparatus is shown in Figure 1. Photolysis light of 248 or 193
nm was provided by an excimer laser (Lambda Physik, Compex
200). Several lead salt diode lasers (Laser Photonics) operating
in the 80-110 K temperature range were used to provide tunable
infrared probe laser light. The IR beam was collimated by a
lens and combined with the UV light by means of a dichroic
mirror, and both beams were copropogated through an absorp-
tion cell. After the UV light was removed by a second dichroic
mirror, the infrared beam was then passed into a 1/4 m
monochromator and focused onto a 1 mm InSb detector
(Cincinnati Electronics, ∼1 µs response time). Transient infrared
absorption signals were recorded on a LeCroy 9310A digital
oscilloscope and transferred to a computer for analysis.
A 1.46 m static absorption cell was used for experiments at
room temperature and above. Resistive heating was used to
achieve elevated temperatures. For measurements at T < 296
K, a 1.21 m jacketed cell was used. Cooling was achieved by
pumping chilled solvent through the outer jacket. For CO2
product measurements, the infrared laser beam path was purged
with N2 to remove atmospheric CO2.
Figure 1. Schematic of the experimental apparatus.
Two different CN precursors were used in these experiments.
For room temperature and elevated temperatures, ICN dissocia-
tion at 248 nm was used. ICN did not have sufficient vapor
pressure for reliable measurements at temperatures below
ambient, however. C2N2 photolysis at 193 nm was therefore
used for experiments at these temperatures. Typical photolysis
laser pulse energies were ∼5-15 mJ/pulse, producing CN
Figure 2. Transient infrared absorption signals for N2O, CO, and CO2
product molecules. Each transient was obtained from a single photolysis
laser shot. Reaction conditions: PICN ) 0.05 Torr (CO and N2O signals
only), PICN ) 0.08 Torr (CO2 signal only), PO ) 2.0 Torr, PSF6 ) 1.5
number densities of ∼1013 cm-3
.
Torr (N2O and CO2 signals only), PCF ) 1.52 Torr (CO signal only),
SF6 and CF4 (Matheson) were purified by repeated freeze-
pump-thaw cycles at 77 K. NO (Matheson) were purified by
repeated freeze-pump-thaw cycles at 163 K to remove NO2
and N2O. ICN (Aldrich) was purified by vacuum sublimation
to remove dissolved air. C2N2 was synthesized by the reaction
of copper sulfate with aqueous sodium cyanide27 and purified
by freeze-pump-thaw cycles at 77 K. Traces of CO2 were
removed from O2 (Matheson, Research Grade) and SF6 by the
use of an ascarite trap.
4
PNO ) 0.2 Torr (N2O signal only).
Torr. To prevent signal degredation due to buildup of product
molecules, the reaction cell was evacuated and refilled after only
a few (two to five) excimer laser shots.
Results
Time-resolved transient absorption signals of product mol-
ecules at 296 K are shown in Figure 2. The lower and middle
traces show the CO and CO2 transient signals produced by the
photolysis of ICN in the presence of O2 and buffer gas. The
upper trace shows an N2O transient signal obtained when NO
reagent is included in the reaction mixture. Given the fast rate
constant of the title reaction, one expects the reaction to occur
on a time scale of a few microseconds under the experimental
condition of 2.0 Torr of O2. The slower rise time observed in
the transient absorption signals is due to the fact that products
are produced in excited vibrational and rotational states and that
vibrational relaxation to the probed ground state is significantly
slower than the reaction rate. Previous experiments in our
laboratory29 as well as measurements of vibrational relaxation
rates30-33 have demonstrated that SF6 is an effective buffer gas
for the relaxation of CO2 and N2O vibrational excitation. For
CO, CF4 is a more efficient relaxer of vibrational excitation
and was therefore used as the buffer gas whenever CO was
probed. Under the experimental conditions used (1-2 Torr
buffer gas), relaxation of the nascent vibrational distribution to
a Boltzmann distribution occurs on a time scale of ∼20-40 µs
for CO2 and N2O, and ∼200 µs for CO. In all cases this
CO, N2O, and CO2 product molecules were probed by
observing the following absorption lines:
CO(V)0,J)9) f CO(V)1,J)8)
P(9) line at 2107.423 cm-1
N2O((0000),J)13) f N2O((0001),J)12)
P(13) line at 2212.326 cm-1
CO2((0000),J)14) f CO2((0001),J)13)
P(14) line at 2337.658 cm-1
In some of the experiments at elevated temperatures, higher
rovibrational transition lines P(10) at 2103.270 cm-1 and P(24)
at 2201.749 cm-1 were probed for CO and N2O product
molecules, respectively. The HITRAN molecular database was
used to locate and identify the spectral lines of product
molecules.28
Typical experimental conditions were PICN or PC N ) 0.05
2
2
Torr, PO ) 2.0 Torr, PSF or PCF ) 1.5 Torr, and PNO ) 0-0.4
2
6
4

CN + O2
J. Phys. Chem. A, Vol. 103, No. 19, 1999 3723
TABLE 1: Product Branching Ratios of the CN + O2
Reaction at 296 K
product channel
branching ratio
NCO + O
CO + NO
CO2 + N
φ1a ) 0.76 ( 0.02
φ1b ) 0.22 ( 0.02
φ1c ) 0.02 ( 0.01
described above. Since φ2a is accurately known, the NCO
concentration is then readily calculated as [NCO] ) [N2O]/φ2a.
The branching ratio φ1b is then [CO]/([CO] + [CO2] + [NCO]),
where CO and CO2 yields are measured without the addition
of NO, as shown in Figure 3. Similarly, φ1c ) [CO2]/([CO] +
[CO2] + [NCO]) and φ1a ) 1.0 - φ1b - φ1c. Table 1 shows
the resulting branching ratios obtained at 296 K in this way
upon averaging of four separate experimental runs. The
uncertainties quoted represent two standard deviations. Note that
both methods produce the same result for φ1b, indicating that
our estimate of [CN]0 was reliable.
Figure 3. Product yields of CO, CO2, and N2O. N2O yields are shown
as a function of NO pressure. ICN, SF6, CF4, and O2 pressures are
identical to those in Figure 2.
By use of the second method, the branching ratio between
channels 1a and 1b was investigated as a function of temper-
ature. Channel 1c, which was only a very minor contribution
at 296 K, was ignored during these variable temperature
measurements. Figure 4 shows the resulting branching ratio into
channel 1b vs T over the range 239-643 K.
B. Secondary Chemistry. The branching ratio measurements
described above depend on the assumption that most or all of
the observed CO is formed by reaction 1b and that contributions
to the CO yield by secondary reactions are small. The most
serious potential secondary reaction is that between the products
of channel 1a:
relaxation was fast compared to the ∼1 ms time scale for
diffusion of product molecules out of the probed beam volume,
which appears in Figure 2 as a slow decay.
A. Branching Ratio Determination. Absorption signals were
converted to number densities using tabulated line strengths28
and equations described previously.24 Calibration experiments
in our laboratory have demonstrated that these line strengths
are accurate to better than 10% (and probably better than 5%)
for the molecules probed in these experiments. For CO and N2O,
peak amplitudes were used to calculate number densities. The
CO2 transient signals displayed both a fast and a slow rise
component, with rise times of ∼30 and ∼300 µs, respectively.
For reasons described below in section B, only the fast rise
components were used to calculate branching ratios. Figure 3
shows the resulting product yields. As shown, the N2O yield in
the absence of added NO is essentially zero, as expected. When
NO is included in the reaction mixture, N2O formation is
NCO + O f CO + NO
f N + CO2
(3a)
(3b)
In addition, reaction 2 may occur to some extent even in the
absence of added NO reagent, since NO is formed in channels
1b and 3a. To examine these possibilities, we performed
experiments in which an additional reagent was included to
remove any NCO formed in channel 1a, effectively quenching
reactions 2 and 3. Among possible NCO quenching reagents,
saturated hydrocarbons are not ideal because NCO + alkane
reactions are generally rather slow except at elevated temper-
atures.36,37 As a result, we chose silane (SiH4) as a quencher.
SiH4 reacts quickly with both NCO and CN without forming
CO:
expected via reaction 2a, which has a branching ratio of φ2a
)
0.44 ( 0.07 over the temperature range 296-623 K.24 In the
limit of high NO pressure, every NCO radical formed in channel
1a reacts with NO. As shown in Figure 3, this occurs at ∼0.1
Torr, beyond which the N2O yield is essentially constant. Note
that an excess of 2.0 Torr of O2 was used and that CN reacts
only very slowly with NO at low total pressures, with a rate
constant of 4.3 × 10-13 cm3 molecule-1 s-1 at 296 K and 5
Torr.34 Since k1 is ∼50 times greater than this and [O2] > [NO],
essentially all of the photolytically created CN radicals are
removed by reaction 1.
CN + SiH4 f HCN + SiH3
(4)
(5)
The branching ratio can be obtained from our data in two
different ways. The first method is the less accurate approach
and was only applied in T ) 296 K experiments. In this
approach, we calculate the initial radical yield [CN]0 from the
measured photolysis laser pulse energy and the 248 nm
NCO + SiH4 f HNCO + SiH3
Measurements in our laboratory indicate that k4 ) 2.2 × 10-10
cm3 molecule-1 s-1 and k5 ) 7.0 × 10-12 cm3 molecule-1 s-1
at 296 K.38 Reaction 5 effectively competes with reaction 3 when
even a small amount of SiH4 is present. For example, if [SiH4]
) 0.1 Torr ≈ 3 × 1015 molecules cm-3, most NCO radicals
formed will react via reaction 5 rather than reaction 3, since
[O] ≈ 1013 molecules cm-3 under typical experimental condi-
tions. As a result, any CO formed via reactions 2 or 3 is expected
to be almost completely suppressed by the addition of small
amounts of SiH4. CO formed by the title reaction 1b will also
be suppressed by the competition for CN radicals (reactions 1b
vs reaction 4); however, this requires higher SiH4 concentrations,
absorption coefficient of ICN (R ) 0.009 cm-1 Torr-1).24
A
dissociation quantum yield of unity is assumed. If secondary
chemistry is insignificant, then φ1b is simply [CO]/[CN]0. Using
this method, we obtain φ1b ) 0.22 ( 0.02 at 296 K, where the
uncertainty represents two standard deviations.
The second approach, which we believe is potentially more
accurate, is to measure the product yields of all active channels.
Unfortunately, quantification of channel 1a by direct NCO
detection is difficult, even though infrared transition wavelengths
near 1900 cm-1 are known,35 because these transitions are quite
weak and precise absorption coefficients are unknown. The
addition of excess NO to the reaction mixture, however,
efficiently converts any NCO into N2O via reaction 2, as
since [O2] ≈ 1016 molecules cm-3
.
The squares of Figure 5 show the experimentally measured
CO yield as a function of added silane reagent. As shown, the

3724 J. Phys. Chem. A, Vol. 103, No. 19, 1999
Rim and Hershberger
Figure 4. Temperature dependence of the branching ratio into channel
1b, CO + NO.
Figure 6. Experimental and modeled dependence of CO2 yield on
added SiH4 reagent. See text for details. PICN ) 0.08 Torr. PO ) 2.0
2
Torr. PSF ) 1.5 Torr.
6
model to be suppressed by 0.1-0.2 Torr SiH4, in sharp contrast
to the experimental data. As shown, a satisfactory agreement
between the model and the experimental data is obtained only
on the assumption that essentially all of the observed CO did
indeed originate from reaction 1b.
Similar modeling calculations were performed for the CO2
products. A slight complication is that the CO2 transient signals
displayed both a fast and a slow rise, as shown in Figure 2.
The fast rise time scale of ∼30 µs is comparable to that apparent
in previous measurements of CO2 reaction products in our
laboratory24,29 and is attributed to rapid vibrational relaxation
of nascent CO2 produced by channel 1c to the probed ground
vibrational state. The slow rise occurs on a longer time scale
than can be attributed to vibrational relaxation, and this rise
time was unaffected by the SF6 buffer gas pressure. We attribute
this slow component of the signal to formation of CO2 by
secondary chemistry such as reaction 2b and possibly reaction
3b. The silane quenching experiments shown in Figure 6 support
this hypothesis. The fast rise component of the CO2 signal (open
squares) decays only gradually with increasing [SiH4], as
predicted by the modeling calculations (filled circles) for CO2
originating from channel 1c. The slow rise components, however
(open triangles), are greatly affected by the addition of small
quantities of SiH4, in fair but not perfect agreement with
modeling predictions for CO2 originating from channel 2b (filled
triangles). Part of the disagreement between the experiment and
the model for the slow rise component may be due to the
difficulty in distinguishing the slow and fast components on
small, noisy signals. On the basis of the these observations, we
used only the fast rise components of the CO2 signals in
determining the branching ratio φ1c listed in Table 1. We note
that these CO2 signals are significantly affected by secondary
chemistry only because φ1c is so small.
Figure 5. Experimental and modeled dependence of CO yield on added
SiH4 reagent. Experimental results are represented by open squares.
Model 1 (triangles) assumes that all CO originated from secondary
chemistry. Model 2 (circles) assumes that all CO originated from
channel 1b. See text for details. PICN ) 0.05 Torr. PO ) 2.0 Torr. PCF
2
4
) 1.0 Torr.
TABLE 2: Reactions Used in Kinetic Quenching Modeling
Simulations of CO and CO2 Yields
k(298K)
reaction
cm3 molecule-1 s-1
ref
CN + O2 f products
2.3 × 10-11
7.0 × 10-11
1.50 × 10-11
1.90 × 10-11
2.2 × 10-10
7.0 × 10-12
23a
43
24
24
38
38
NCO + O f CO + NO
NCO + NO f N2O + CO
NCO + NO f CO2 + N2
CN + SiH4 f HCN + SiH3
NCO + SiH4 f HCN + SiH3
a This work used for the branching ratio
product yield declines to nearly zero only as the SiH4 pressure
exceeds 1.0 Torr. Comparison of these data to the argument
above suggests that little of the observed CO could have been
formed from any secondary reaction involving NCO because
such CO should have been suppressed by a silane pressure as
low as 0.1 Torr.
To quantitatively test these arguments, kinetic modeling
simulations were performed on this system. Table 2 shows the
reactions included in the kinetic model. The ACUCHEM kinetic
modeling program39 was used to predict the CO yield under
two contrasting assumptions: that all CO was formed by the
title reaction 1b and that all CO was formed by secondary
reactions 2a and 3a. These modeling predictions are shown in
Figure 5 as circles and triangles, respectively. As expected, any
CO produced by the secondary reactions is predicted by the
Discussion
Our value of φ1b ) 0.22 ( 0.02 at 296 K is in excellent
agreement with our previous measurement,24 which was ob-
tained in a somewhat less direct manner by measuring the excess
of CO over N2O yields in the NCO + NO reaction, where NCO
was formed by the CN + O2 reaction. Our result is in fair
although not perfect agreement with the value of φ1b ) 0.29 (
0.02 obtained by Mohammed et al. using infrared emission.25
In general, we believe that absorption spectroscopy has fewer
uncertainties in the determination of absolute concentrations than
emission spectroscopy. For example, our experiment is not

CN + O2
J. Phys. Chem. A, Vol. 103, No. 19, 1999 3725
(2) Phillips, L. F.; Smith, I. W. M.; Tuckett, R. P.; Whitham, C. J.
Chem. Phys. Lett. 1991, 183, 254.
(3) Schmatjko, K. J.; Wolfrum, J. Ber. Bunsen-Ges. Phys. Chem. 1978,
82, 419.
(4) Wright, S. A.; Dagdigian, P. J. J. Chem. Phys. 1995, 100, 6479.
(5) Reisler, H.; Mangir, M.; Wittig, C. Chem. Phys. 1980, 47, 49.
(6) Jensen, R. C.; Walton, D. B.; Coombe, R. D. Ber. Bunsen-Ges.
Phys. Chem. 1990, 169, 441.
(7) Anastasi, C.; Hancock, D. U. J. Chem. Soc., Faraday Trans. 2 1988,
84, 9.
sensitive to the nascent vibrational distribution of product
molecules because our use of polyatomic buffer gases relaxes
the detected products to an ambient Boltzmann distribution. In
any case, both techniques are in qualitative agreement that
channel 1b represents a significant but not dominant pathway
of the CN + O2 reaction at room temperature.
The temperature dependence of φ1b is an important result of
this experiment. Comparison with our earlier work24 shows that
the decrease in φ1b as T increases is slightly less pronounced
(8) Lichtin, D. A.; Lin, M. C. Chem. Phys. 1985, 96, 473.
(9) Louge, M. Y.; Hanson, R. K. Int. J. Chem. Kinet. 1984, 16, 231.
than previously thought (Figure 4 of this work shows φ1b
)
0.07 at T ) 640 K, while Figure 7 of ref 24 shows φ1b ≈ 0.02
at a similar temperature). Our uncertainties of (0.02, however,
are much improved compared to ref 24, which reported (0.10
error bars. Furthermore, our newer data show that at tempera-
tures below 296 K, φ1b increases quite dramatically. Although
extremely low temperatures are not attainable in our experi-
ments, it appears quite likely that channel 1b is in fact the
dominant channel as T f 0. In any case, it is remarkable that
the branching ratio changes so much over a modest temperature
range. This is in contrast to many other radical-radical reactions
in combustion chemistry, such as NCO + NO24,40 and NH2 +
NO,41,42 in which significant changes in branching ratio are
observable only when comparing experiments performed over
a much wider temperature range.
(10) You, Y. Y.; Wang, N. S. J. Chin. Chem. Soc. (Taipei) 1993, 40,
337.
(11) Durant, J. L., Jr.; Tully, F. P. Chem. Phys. Lett. 1989, 154, 568.
(12) Balla, R. J.; Castleton, K. H. J. Phys. Chem. 1991, 95, 2344.
(13) Sims, I. R.; Smith, I. W. M. Chem. Phys. Lett. 1988, 151, 481.
(14) Sims, I. R.; Queffelec, J.-L.; Defrance, A.; Rebrion-Rowe, C.;
Travers, D.; Bocherel, P.; Rowe, B. R.; Smith, I. W. M. J. Chem. Phys.
1994, 100, 4229.
(15) Atakan, B.; Jacobs, A.; Wahl, M.; Weller, R.; Wolfrum, J. Chem.
Phys. Lett. 1989, 154, 449.
(16) Burmeister, M.; Gulati, S. K.; Natarajan, K.; Thielen, K.; Mozzhukin,
E.; Roth, P. Symp. (Int.) Combust., (Proc.) 1989, 22, 1083.
(17) Smith, I. W. M. In Chemical Dynamics and Kinetics of Small
Radicals; Liu, K. Wagner, A., Eds.; World Scientific: Singapore, 1995;
Part I, p 214.
(18) Vallance, C.; Maclagan, R. G. A. R.; Phillips, L. F. Chem. Phys.
Lett. 1996, 250, 59.
In principle, ab initio and RRKM calculations should be able
to predict the temperature dependence of the branching ratio.
Unfortunately, all the high-level calculations to date have
essentially ignored channel 1b and instead concentrated on either
the entrance channel to the NCOO collision complex or the
transition state leading to NCO + O products.18-20 In addition,
Phillips has suggested that this system may be too small for a
statistical treatment of the NCOO complex decay kinetics.20
Prediction of our results thus represents a substantial challenge
for ab initio calculations. Our observation of a dramatic
temperature dependence in φ1b suggests that the transition states
leading to product formation have energies fairly close to that
of the initial CN + O2 reactants. Thus, it is somewhat surprising
that a recent G2 calculation indicates that the transition state to
NCO + O lies 71.5 kJ/mol below the reactants.18 It is clear
that more computational work on this system, especially in
regard to transition states leading to CO + NO, are needed.
(19) Klippenstein, S. J.; Kim, T.-W. J. Chem. Phys. 1993, 99, 5790.
(20) Phillips, L. F. J. Phys. Chem. A 1998, 102, 31.
(21) Smith, I. W. M. Int. J. Mass Spectrom. Ion Processes 1995, 149,
231.
(22) Hessler, J. P. J. Chem. Phys., submitted.
(23) Baulch, D. L.; Cobos, C. J.; Cos, R. A.; Frank, P.; Hayman, G.;
Just, Th.; Kerr, J. A.; Murrells, T.; Pilling, M. J.; Troe, J.; Walker, R. W.;
Warnatz, J. J. Phys. Chem. Ref. Data 1994, 23 (Suppl. 1), 847.
(24) Cooper, W. F.; Park, J.; Hershberger, J. F. J. Phys. Chem. 1993,
97, 3283.
(25) Mohammad, F.; Morris, V. R.; Fink, W. H.; Jackson, W. M. J.
Phys. Chem. 1993, 97, 11590.
(26) Rim, K. T.; Hershberger, J. F. J. Phys. Chem. A 1998, 102, 5898.
(27) Janz, G. J. Inorg. Synth. 1957, 5, 43.
(28) Rothman, L. S.; et al. J. Quant. Spectrosc. Radiat. Transfer 1992,
48, 469.
(29) Cooper, W. F.; Hershberger, J. F. J. Phys. Chem. 1992, 96, 771.
(30) Fakhr, A.; Bates, R. D., Jr. Chem. Phys. Lett. 1980, 71, 381.
(31) Stephenson, J. C.; Moore, C. B. J. Chem. Phys. 1970, 52, 2333.
(32) Richman, D. C.; Millikan, R. C. J. Chem. Phys. 1975, 63, 2242.
(33) Green, W. H.; Hancock, J. K. J. Chem. Phys. 1973, 59, 4326.
Conclusion
(34) Sims, I. R.; Smith, I. W. M. J. Chem. Soc., Faraday Trans. 1993,
89, 1.
(35) Brueggemann, R.; Petri, M.; Fischer, H.; Mauer, D.; Reinert, D.;
Urban, W. Appl. Phys. B 1989, 48, 105.
(36) Park, J.; Hershberger, J. F. Chem. Phys. Lett. 1994, 218, 537.
Infrared absorption spectroscopy was used to determine the
product branching ratios of the CN + O2 reaction. CO2 + N is
a very minor channel, as expected. CO + NO is a significant
channel with a surprisingly large temperature dependence,
becoming more important at lower temperatures.
(37) Schuck, A.; Volpp, H.-R.; Wolfrum, J. Combust. Flame 1994, 99,
491.
(38) Edwards, M. A. Ph.D. Thesis, North Dakota State University, Fargo,
ND, 1999.
(39) Braun, W.; Herron, J. T.; Kahaner, D. Int. J. Chem. Kinet. 1988,
20, 51.
(40) Lin, M. C.; He, Y.; Melius, C. F. J. Phys. Chem. 1993, 97, 9124.
(41) Glarborg, P.; Kristensen, P. G.; Dam-Johansen, K.; Miller, J. A. J.
Phys. Chem. A 1996, 101, 3741.
Acknowledgment. We thank Jan Hessler of Argonne
National Laboratory for providing a preprint prior to publication.
This work was supported by the Division of Chemical Sciences,
Office of Basic Energy Sciences of the Department of Energy,
Grant DE-FG03-96ER14645.
(42) Park, J.; Lin, M. C. J. Phys. Chem. 1996, 100, 3317.
References and Notes
(43) Baulch, D. L.; Cobos, C. J.; Cox, R. A.; Esser, C.; Frank, P.; Just,
Th.; Kerr, J. A.; Pilling, M. J.; Troe, J.; Walker, R. W.; Warnatz, J. J.
Phys. Chem. Ref. Data 1992, 21, 411.
(1) Miller, J. A.; Bowman, C. T. Prog. Energy Combust. Sci. 1989,
15, 287.