Product branching ratio of the HCCO + NO reaction

Article abstract of DOI:10.1021/jp9922209

Excimer laser photolysis of ketene precursor molocules followed by IR absorption spectroscopy were performed to study the reaction of ketenyl radical (HCCO) radicals with NO at room temperature. CO and CO₂ were produced. Considering possible secondary chemistry, CO + (HCNO) was the main product channel, with a branching ratio of 0.88 ± 0.04:1 at 296 K. CO₂ + (HCN) was a minor channel, with a branching ratio of 0.12 ±0.04:1. The relative quantum yield for HCCO production in the 193 nm photolysis of CH₂CO was estimated to be 0.17 ± 0.02.

Full text of DOI:10.1021/jp9922209

J. Phys. Chem. A 2000, 104, 293-296
293
Product Branching Ratio of the HCCO + NO Reaction
Kwang Taeg Rim† and John F. Hershberger*
Department of Chemistry, North Dakota State UniVersity, Fargo, North Dakota 58105
ReceiVed: June 29, 1999; In Final Form: October 29, 1999
The reaction of HCCO radicals with NO was studied at room temperature by excimer laser photolysis of
ketene precursor molecules followed by infrared absorption spectroscopic detection of CO and CO2 product
molecules. After quantification of product yields and consideration of secondary chemistry, we obtain the
following product branching ratios (1σ error bars) at 296 K: 0.12 ( 0.04 for CO2 + (HCN) and 0.88 ( 0.04
for CO + (HCNO). In addition, we estimate a relative quantum yield for HCCO production in the 193 nm
photolysis of CH2CO to be 0.17 ( 0.02.
Introduction
Measurement of the product branching ratio of the title
reaction represents a challenge because no ideal photolytic
source of HCCO is known. Typically, previous experiments
have used the reaction of oxygen atoms with acetylene to form
HCCO:
The spectroscopy1-4 and kinetics4-8 of the ketenyl radical
(HCCO) is of recent interest, partly because of the role this
species plays in combustion chemistry. HCCO is formed in
flames primarily by the oxidation of acetylene.9,10 It has been
observed in laboratory studies by infrared absorption4 and laser-
induced fluorescence spectroscopy.1,2 Several kinetic studies
involving HCCO have appeared recently, with total rate constant
measurements reported for reactions with NO,4-7 NO2,5 O2,5
O,8 and C2H2.5 The reaction with NO is of particular interest
because of the role it plays in NO-reburning mechanisms11-15
for the reduction of NOx emissions from fossil-fuel combustion
processes. Several product channels are possible:
O + C2H2 f HCCO + H
f CH2 + CO
(2a)
(2b)
This approach has two problems: reaction 2 is quite slow at
moderate temperatures,8,9 and the presence of channel 2b
complicates quantification of product yields. Unfried et al.,
however, reported that direct photolysis of ketene at 193 nm
produces ketenyl radicals as well as the well-known methylene
+ CO channel:4
HCCO + NO f CO2 + (HCN)
∆H° ) -527.2 kJ/mol
(1a)
f CO + (HCNO)
∆H° ) -200.8 kJ/mol
CH2CO + hν (193 nm) f CH2 + CO
f HCCO + H
(3a)
(3b)
(1b)
f NCO + HCO ∆H° ) -71.1 kJ/mol
(1c)
We have chosen this approach in the experiments reported here.
Reaction 3b represents a faster source of HCCO than (2a),
although we must still contend with the production of CH2 +
CO in channel 3a.
The thermochemistry shown is for HCN and HCNO, but several
other isomers of these species represent possible, albeit unlikely,
product channels as well. Several reports of the total rate
constant of this reaction have appeared. Unfried et al. used
infrared absorption near 2023 cm-1 to detect HCCO and
reported k1 ) (3.9 ( 0.5) × 10-11 cm3 molecule-1 s-1 at 298
K.4 Temps et al. used far-infrared laser magnetic resonance to
detect HCCO and obtained (2.2 ( 0.6) × 10-11 cm3 molecule-1
s-1 at 298 K.5 Boullart et al. used discharge flow-mass
spectrometry to obtain k1 ) (1.0 ( 0.3) × 10-10 exp [-350 (
150)/T] cm3 molecule-1 s-1 over the temperature range 290-
670 K and also reported the following product branching ratios
at 700 K: φ1a ) 0.23 ( 0.09, φ1b ) 0.77 ( 0.09.6 These data
are contradicted by recent calculations and flow reactor studies.
Miller et al. used statistical theories and a QCISD potential
surface of Nguyen et al.7 to predict φ1a ) 0.81 at 300 K, with
a moderate temperature dependence, decreasing to about 0.32
at 2000 K.12 Kinetic modeling of flow reactor experiments at
1100-1400 K were best fit by a branching ratio of φ1a ) 0.65.13
Experimental Section
The experimental procedure is similar to that described in
previous publications.16,17 A schematic of the experimental
apparatus is shown in Figure 1. Photolysis light of 193 nm was
provided by an excimer laser (Lambda Physik, Compex 200).
Several lead salt diode lasers (Laser Photonics, Analytics
Division) 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 copropagated
through a 1.46 m absorption cell. After the UV light was
removed by a second dichroic mirror, the infrared beam was
1
then passed into a /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.
† Current address: Department of Chemistry, Columbia University, 3000
Broadway MS 3109, New York, NY 10027.
10.1021/jp9922209 CCC: $19.00 © 2000 American Chemical Society
Published on Web 12/07/1999

294 J. Phys. Chem. A, Vol. 104, No. 2, 2000
Rim and Hershberger
Figure 1. Schematic of the experimental apparatus.
Figure 2. Transient infrared absorption signals for CO and CO2 product
molecules. CO transients were obtained from a single photolysis laser
shot. CO2 transients were averages of two photolysis laser shots.
SF6, Xe, and CF4 (Matheson) were purified by repeated
freeze-pump-thaw cycles at 77 K. A trace of CO2 was
removed from SF6 by the use of an ascarite trap. NO (Matheson)
and 15N18O (Isotec) were purified by repeated freeze-pump-
thaw cycles at 163 K to remove NO2 and N2O. Ketene was
synthesized by the pyrolysis of acetic anhydride at 700 °C and
purified by freeze-pump-thaw cycles at 77 K. The purity of
the ketene samples was checked by FTIR. A substantial CO2
impurity in the ketene sample was observed, and attempts to
remove this using ascarite were unsuccessful, apparently due
to the reactivity of ketene. The use of labeled 15N18O reagents
to produce 16O12C18O products in the title reaction minimized
(but not completely eliminated) interference from this static CO2
impurity. Note that only the oxygen labeling is important in
the experiments reported here. Furthermore, isotopically labeled
NO was not used for the CO detection experiments, because
channel 1b would probably produce HC15N18O + CO, and not
C18O.
Reaction conditions: PCH CO ) 0.1 Torr, PNO ) 0.2 Torr, PXe ) 1.0
2
Torr, PSF ) 2.0 Torr (CO2 signals only), PCF4 ) 3.0 Torr (CO signals
6
only).
impurity in the ketene samples produced some interference even
when probing 16O12C18O due to the 18O natural abundance. This
factor as well as the small CO2 product yields are responsible
for the poor S/N of the CO2 signals. Typical changes in
transmittance (∆I/I0) values are 10-15% for the CO transient
signals but only 1-2% for 16O12C18O signals.
Previous experiments in our laboratory16,19 as well as
vibrational energy transfer measurements20-23 have demon-
strated that SF6 is an efficient buffer gas for the relaxation of
vibrationally excited CO2 or N2O to a Boltzmann distribution
but that CF4 is a more efficient relaxer of vibrationally excited
CO. The relatively slow rise times observed in the transient
signals represent a convolution of the fast reaction rate k1 with
slower vibrational relaxation rates. In any case, the vibrational
populations are relaxed to a Boltzmann distribution within 100-
200 µs. The slow ∼1 ms decay is attributed to diffusion of
product molecules out of the probed region of the reaction cell.
Attempts to detect N2O product molecules in this reaction
system were unsuccessful. If channel 1c is active, one expects
N2O formation in high yield via the secondary reaction:
CO, N2O, and isotopically labeled 16O12C18O product mol-
ecules were probed by observing the following absorption lines:
CO (V ) 1 r V ) 0)
N2O (0001) r (0000)
P(8), 2111.543 cm-1
P(13), 2212.326 cm-1
16O12C18O (0001) r (0000)
R(13), 2341.812 cm-1
NCO + NO f N2O + CO
f CO2 + N2
(4a)
(4b)
The HITRAN molecular database was used to locate and
identify the spectral lines of product molecules.18 The spectral
lines used are near the peak of the rotational Boltzmann
distribution, minimizing sensitivity to small heating effects. For
CO2 product molecule measurements, the infrared laser beam
path was purged with N2 to remove atmospheric CO2.
Previous experiments have shown that φ4a ) 0.44 at 298 K.16
Our failure to detect N2O indicates that channel 1c does not
contribute significantly to the title reaction. This is a more
sensitive test for channel 1c than direct NCO detection by
infrared spectroscopy24 because N2O has much greater infrared
line strengths.
Absorption signals were converted to number densities using
tabulated line strengths14 and equations described previously.16
The line strength of R(13) of 16O12C18O was obtained by
dividing the tabulated line strength, Svj ) 6.632 × 10-21 (which
assumes a natural abundance sample), by the isotopic abundance
of 16O12C18O, 3.947 × 10-3. The product yields of CO and
16O12C18O are shown in Figure 3 as a function of reagent NO
pressure. As shown, significant amounts of CO are formed in
the direct photolysis of ketene at 193 nm via channel 3a. Upon
addition of NO to the reaction mixture, however, a significant
and reproducible increase in the CO yield is observed due to
reaction 1b. In addition, a small but significant yield of
16O12C18O was observed when 15N18O was included in the
Typical experimental conditions were PCH CO ) 0.1 Torr, PSF
2
6
) 2.0 Torr, PCF ) 3.0 Torr, and PNO ) 0-0.4 Torr. Typically,
4
several (two to five) laser shots were averaged when probing
the weaker CO2 signals, but no signal averaging was used for
the stronger CO signals.
Results
Time-resolved transient absorption signals of product mol-
ecules at 296 K are shown in Figure 2. The CO transient signals
were produced by the photolysis of CH2CO in the presence of
NO reagents and CF4 buffer gas. CO signals were also observed
in the absence of NO reagent and are attributed to direct ketene
photolysis, channel 3a.
The 16O12C18O transient signals were obtained by photolyzing
CH2CO in the presence of labeled 15N18O and SF6 buffer gas
(no signals were obtained in the absence of 15N18O). The CO2

Product Branching Ratio of the HCCO + NO Reaction
J. Phys. Chem. A, Vol. 104, No. 2, 2000 295
1
1
TABLE 3: Collisional Quenching of CH2 and NH
reaction
rate const, cm3 molecule-1 s-1
ref
34
32, 34
34
35
35
1CH2 + He
1CH2 + Ar
1CH2 + N2
1NH + N2
1NH + Xe
2.5 × 10-12
(4.8-5.2) × 10-12
7.0 × 10-12
7.9 × 10-14
1.1 × 10-11
of ketene and that the more well-known CH2 + CO channel
dominates the photolysis. This fact is partly responsible for the
large background [CO]0 in this study.
Discussion
The above analysis is dependent on the assumption that only
reactions 1 and 3 are important sources of CO or CO2 when
CH2CO/NO/buffer gas mixtures are photolyzed. Several second-
ary reactions must be considered. Some of these are quite slow:
Figure 3. Product yields of CO and CO2 product molecules are shown
as a function of NO pressure. Experimental conditions are identical to
those of Figure 2. Filled triangles: CO yield. Filled circles: CO2 yields.
Reaction conditions: PCH CO ) 0.1 Torr, PNO ) 0-0.4 Torr, PXe
)
1.0 Torr, PSF ) 2.0 Torr 2(CO2 only), PCF ) 3.0 Torr (CO only).
H + CH2CO f products
(6)
(7)
6
4
TABLE 1: Product Branching Ratio of the HCCO + NO
3CH2 + CH2CO f products
Reaction at 296 Ka
where k6 ) 6.2 × 10-14 cm3 molecule-1 s-1 and k7 < 3.5 ×
10-13 cm3 molecule-1 s-1 at 298 K.25,26 Although CO is a likely
product of both of these reactions, under our experimental
conditions of 0.1 Torr of CH2CO, these reactions occur on time
scales of >1 ms, which is much greater than the observed
transient signal rise times of ∼100 µs and, in fact, is comparable
to or slower than the time scale for diffusion of products out of
the probed volume. Thus, these reactions have little effect on
our results.
product channel branching ratio without Xe branching ratio with Xe
CO + (HCNO)
CO2 + (HCN)
0.84 ( 0.07
0.16 ( 0.07
0.88 ( 0.04
0.12 ( 0.04
a Error bars represent one standard deviation.
TABLE 2: Relative Branching Ratio for the Photolysis of
CH2CO at 193 nma
product channel branching ratio without Xe branching ratio with Xe
CH2 + CO
HCCO + H
0.84 ( 0.02
0.16 ( 0.02
0.83 ( 0.02
0.17 ( 0.02
A faster secondary reaction is
3CH2 + NO f HCNO + H
(8)
a Error bars represent one standard deviation.
where k8 ) 2.9 × 10-11 cm3 molecule-1 s-1 27
.
Although other
reaction mixture. The product yields are essentially unchanged
over the range 0.2-0.4 Torr of NO, indicating that under these
conditions virtually all of the HCCO radicals formed in (1a)
are removed by reaction with NO.
product channels than shown are possible in reaction 8, HCNO
formation has been previously identified as the dominant
product.28,29 This reaction will therefore not significantly affect
our measured CO or CO2 yields.
If only reactions 1 and 3 are considered, the branching ratio
of reaction 1 may be calculated as follows. We define [CO]0 as
the CO yield observed in the absence of NO, i.e., that produced
by (3a). Let [CO]tot be the total CO yield observed in the
presence of NO. The difference [CO]dif ) [CO]tot - [CO]0 is
the CO yield from channel 1b of the title reaction. If we assume
that only channels 1a and 1b are active, the product branching
ratio φ1b is then [CO]dif/([CO]dif + [CO2]), and φ1a ) 1 - φ1b.
Table 1 shows the results of this analysis, averaged over nine
different measurements.
An additional piece of information that can be obtained from
our data is the relative photodissociation quantum yields of
channels 3a and 3b in the 193 nm ketene photolysis. Let x )
[CO]tot/[CO]0 - 1. Since [CO]tot/[CO]0 ) ([CO]0 + [HCCO]0φ1b)/
[CO]0, we have
A further issue is the possibility of excited singlet state
methylene production in the photolysis of ketene. 1CH2 produc-
tion in high yield is well-known in the near-UV photolysis of
ketene near 300 nm,30 but to our knowledge the relative amount
of 1CH2 vs 3CH2 formation at 193 nm has not been reported. If
1CH2 is a major component of 3a, then several fast reactions
are possible:
1CH2 + CH2CO f products
1CH2 + NO f products
(9)
(10)
where k9 ) (2.0-2.7) × 10-10 cm3 molecule-1 s-1 and k10
)
1.6 × 10-10 cm3 molecule-1 s-1 at 298 K.31,32 Either of these
reactions could form CO. To test for these possibilities, we
performed some experiments with an excess of xenon included
in the reaction mixture in order to electronically relax any 1CH2
formed:
x ) [HCCO]0φ1b/[CO]0
(5)
The relative quantum yields for CO and HCCO formation in
the photolysis of ketene are φ3a ) [CO]0/([CO]0 + [HCCO]0)
and φ3b ) [HCCO]0/([CO]0 + [HCCO]0), respectively. By
solving eq 5 for [CO]0 and substituting, we obtain φ3a ) φ1b/(x
+ φ1b) and φ3b ) 1 - φ3a. Note that these are only relative
quantum yields, as we do not know the fluorescence quantum
yield of ketene at 193 nm. The result of this analysis is shown
in Table 2. It is apparent that ketenyl radical formation is a
relatively minor but not insignificant channel in the photolysis
1CH2 + Xe f 3CH2 + Xe
(11)
Unfortunately, no rate constants for reaction 11 have been
reported. For comparison, Table 3 shows literature values of
1
kinetic data for removal of CH2 as well as the isoelectronic
species NH(1∆) with several collision partners.31-35 As shown,
Xe is a faster relaxer of NH(1∆) than N2 by more than 2 orders

296 J. Phys. Chem. A, Vol. 104, No. 2, 2000
Rim and Hershberger
of magnitude. This is presumably due to the well-known heavy
atom effect in promoting spin-forbidden processes. Similar
effects are expected to be present in singlet methylene relaxation
processes as well. Although explicit experimental data would
clearly be desirable, it is likely than reaction 11 occurs with a
large rate constant, probably at least an order of magnitude
greater than the 1CH2 + Ar value of 5.2 × 10-12 cm3 molecule-1
s-1. Addition of Xe to our reaction mixtures is therefore
supported by the Division of Chemical Sciences, Office of Basic
Energy Sciences of the Department of Energy, Grant DE-FG03-
96ER14645.
References and Notes
(1) Brock, L. R.; Mischler, B.; Rohlfing, E. A.; Bise, R. T.; Neumark,
D. M. J. Chem. Phys. 1997, 107, 665.
(2) Brock, L. R.; Mischler, B.; Rohlfing, E. A. J. Chem. Phys. 1999,
110, 6773.
(3) Osborn, D. L.; Mordaunt, D. H.; Choi, H.; Bise, R. T.; Neumark,
D. M.; Rohlfing, C. M. J. Chem. Phys. 1996, 105, 10087.
(4) Unfried, K. G.; Glass, G. P.; Curl, R. F. Chem. Phys. Lett. 1991,
177, 33.
(5) Temps, F.; Wagner, H. Gg.; Wolf, M. Z. Phys. Chem. 1992, 176, 27.
(6) Boullart, W.; Ngugen, M. T.; Peeters, J. J. Phys. Chem. 1994, 98,
8036.
1
expected to be effective in relaxing CH2.
Tables 1 and 2 show data obtained both with and without
1.0 Torr of xenon in the reaction mixture. As shown, the
measured branching ratios of the title reaction were only very
slightly affected by the addition of Xe. Several interpretations
of this result are possible. One is that the 1CH2 yield in the 193
nm photolysis of ketene is small. It is more likely, however,
(7) Nguyen, M. T.; Boullart, W.; Peeters, J. J. Phys. Chem. 1994, 98,
8030.
(8) Peeters, J.; Boullart, W.; Devriendt, K. J. Phys. Chem. 1995, 99,
3583.
1
that CH2 is in fact produced in our system but is efficiently
relaxed even in the absence of Xe, possibly by nonreactive
channels of (9) or (10) or by collisions with CF4 or SF6 buffer
gas. In any case, our data show that secondary chemistry of
1CH2 only has a minor effect on our results. We believe that
the data obtained with Xe included are the more reliable.
Comparing our results with the previously reported values,
we are in reasonable agreement with the study of Boullart et
al., which reported φ1a ) 0.23 ( 0.09 and φ1b ) 0.77 ( 0.09
at 700 K.6 Our results are in marked disagreement, however,
with more recent calculations12 and modeling studies,13,15 which
have suggested that φ1a ) 0.81 at 300 K and decreases
moderately with temperature. The potential energy surface of
this reaction7 involves initial formation of an HC(NO)CO
adduct. Formation of (1b) proceeds through a C-C bond fission,
while a four-centered cyclic intermediate is involved in the route
to channel (1a). These pathways are comparable to similar
pathways in the potential surface of the isoelectronic NCO +
NO reaction, which produces N2O + CO and N2 + CO2 in
comparable yield.16 Our results and those of ref 6 suggest,
however, that the HCCO + NO potential surface is not known
with sufficient accuracy to quantitatively predict the branching
ratio.
(9) Vinckier, C.; Schaekers, M.; Peeters, J. J. Phys. Chem. 1985, 89, 508.
(10) Schmoltner, P. M.; Chu, P. M.; Lee, Y. T. J. Chem. Phys. 1989,
91, 5365.
(11) Chen, S. L.; McCarthy, J. M.; Clark, W. D.; Heap, M. P.; Seeker,
W. R.; Pershing, D. W. Symp. (Int.) Combust Proc. 1986, 21, 1159.
(12) Miller, J. A.; Durant, J. L.; Glarborg, P. Symp. (Int.) Combust. Proc.
1998, 27, 235.
(13) Glarborg, P.; Alzueta, M. U.; Dam-Johansen, K.; Miller, J. A.
Combust. Flame 1998, 115, 1.
(14) Miller, J. A.; Bowman, C. T. Prog. Energy Combust. Sci. 1989,
15, 287.
(15) Prada, L.; Miller, J. A. Combust. Sci. Technol. 1998, 132, 225.
(16) Cooper, W. F.; Park, J.; Hershberger, J. F. J. Phys. Chem. 1993,
97, 3283.
(17) Rim, K. T.; Hershberger, J. F. J. Phys. Chem. A 1998, 102, 5898.
(18) Rothman, L. S.; et al. J. Quant. Spectrosc. Radiat. Transfer 1992,
48, 469.
(19) Cooper, W. F.; Hershberger, J. F. J. Phys. Chem. 1992, 96, 771.
(20) Fakhr, A.; Bates, R. D., Jr. Chem. Phys. Lett. 1980, 71, 381.
(21) Stephenson, J. C.; Moore, C. B. J. Chem. Phys. 1970, 52, 2333.
(22) Richman, D. C.; Millikan, R. C. J. Chem. Phys. 1975, 63, 2242.
(23) Green, W. H.; Hancock, J. K. J. Chem. Phys. 1973, 59, 4326.
(24) Brueggemann, R.; Petri, M.; Fischer, H.; Mauer, D.; Reinert, D.;
Urban, W. Appl. Phys. B 1989, 48, 105.
(25) Michael, J. V.; Nava, D. F.; Payne, W. A.; Stief, L. J. J. Phys.
Chem 1979, 70, 5222.
(26) Darwin, D. C.; Moore, C. B. J. Phys. Chem. 1995, 99, 13467.
(27) Darwin, D. C.; Moore, C. B. J. Phys. Chem. 1989, 93, 1074.
(28) Bauerle, S.; Klatt, M.; Wagner, H. Gg. Ber Bunsen-Ges. Phys.
Chem. 1995, 99, 97.
Conclusion
(29) Grussdorf, J.; Temps, F.; Wagner, H. Gg. Ber. Bunsen-Ges. Phys.
Chem. 1997, 101, 134.
(30) Okabe, H. Photochemistry of Small Molecules; J. Wiley: New York,
1978; p 310.
(31) Langford, A. O.; Petek, H.; Moore, C. B. J. Chem. Phys. 1983, 78,
6650.
(32) Hancock, G.; Heal, M. R. J. Phys. Chem. 1992, 96, 10316.
(33) Hayes, F.; Lawrance, W. D.; Staker, W. S.; King, K. D. J. Phys.
Chem. 1996, 100, 11314.
Photolysis (193 nm) of ketene was used to investigate the
reaction of HCCO radicals with NO. By detection of CO and
CO2 products and consideration of possible secondary chemistry,
we find that CO + (HCNO) is the major product channel, with
a branching ratio of 0.88 ( 0.04, and that CO2 + (HCN) is a
minor channel, with a branching ratio of 0.12 ( 0.04.
(34) Wagener, R. Z. Naturforsch. 1990, 45a, 649.
(35) Nelson, H. H.; McDonald, J. R.; Alexander, M. H. J. Phys. Chem.
1990, 94, 3291.
Acknowledgment. We thank Michael Edwards for as-
sistance during the early stages of this project. This work was