Kinetics and Product Study of the Reaction of CH3 Radicals with O(3P) Atoms Using Time Resolved Time-of-Flight Spectrometry

Article abstract of DOI:10.1021/jp991157k

The kinetics and product distribution for the reaction of methyl radicals, CH₃, with ground-state O(³P) oxygen atoms have been investigated. This reaction was studied with a newly constructed apparatus combining a tubular flow reactor and a time-of-flight mass spectrometer (TOFMS), using a hollow-cathode lamp for photoionization. The radicals are produced by an excimer laser pulse (λ = 193 nm) in the cophotolysis of acetone, CH₃COCH₃, or bromomethane, CH₃Br, and sulfur dioxide, SO₂, creating a homogeneous distribution of radicals along the axis of the flow reactor. A small fraction of the reaction mixture is sampled through a pinhole in the wall. Subsequent ionization and repeated extraction of ionized molecules into the TOFMS at a high repetition rate (≈20 kHz) allows the simultaneous observation of rapid changes in the concentration of multiple species in the flow reactor. In addition to the dominant product, formaldehyde (CH₂O), carbon monoxide (CO) was detected as a product with a yield of 0.17 ± 0.11. Analysis of the rate of disappearance of methyl radicals and appearance of formaldehyde for different O(³P) concentrations resulted in an overall rate coefficient for this reaction k = (1.7 ± 0.3) × 10^-10 cm³ molecule^-1 s^-1 at T = (299 ± 2) K and P = l Torr (He).

Full text of DOI:10.1021/jp991157k

5722
J. Phys. Chem. A 1999, 103, 5722-5731
Kinetics and Product Study of the Reaction of CH3 Radicals with O(3P) Atoms Using Time
Resolved Time-of-Flight Spectrometry
Christopher Fockenberg,* Gregory E. Hall, Jack M. Preses, Trevor J. Sears, and
James T. Muckerman
Chemistry Department 555A, BrookhaVen National Laboratory, P.O. Box 5000, Upton, New York 11973-5000
ReceiVed: April 6, 1999; In Final Form: May 17, 1999
The kinetics and product distribution for the reaction of methyl radicals, CH3, with ground-state O(3P) oxygen
atoms have been investigated. This reaction was studied with a newly constructed apparatus combining a
tubular flow reactor and a time-of-flight mass spectrometer (TOFMS), using a hollow-cathode lamp for
photoionization. The radicals are produced by an excimer laser pulse (λ ) 193 nm) in the cophotolysis of
acetone, CH3COCH3, or bromomethane, CH3Br, and sulfur dioxide, SO2, creating a homogeneous distribution
of radicals along the axis of the flow reactor. A small fraction of the reaction mixture is sampled through a
pinhole in the wall. Subsequent ionization and repeated extraction of ionized molecules into the TOFMS at
a high repetition rate (≈20 kHz) allows the simultaneous observation of rapid changes in the concentration
of multiple species in the flow reactor. In addition to the dominant product, formaldehyde (CH2O), carbon
monoxide (CO) was detected as a product with a yield of 0.17 ( 0.11. Analysis of the rate of disappearance
of methyl radicals and appearance of formaldehyde for different O(3P) concentrations resulted in an overall
rate coefficient for this reaction k ) (1.7 ( 0.3) × 10-10 cm3 molecule-1 s-1 at T ) (299 ( 2) K and P )
1 Torr (He).
Introduction
C2H4 + O f CH3 + HCO
(2)
The combustion of methane is of considerable importance
in the generation of energy. The oxidation of methane starts
with a chain initiation process in which a hydrogen atom is
abstracted. Under lean-to-moderately rich conditions the methyl
radicals produced react predominantly with oxygen atoms:1,2
The reactants were detected by mass spectrometry. The
reported yield of formaldehyde from reaction 1 was larger than
0.85. Water as a possible byproduct was not detected, thus ruling
out channel (1d). However, reaction 2 does not seem to be a
clean source for methyl radicals.11 According to Koda et al.12
another, equally important product of this reaction is CH2CHO,
which could also lead to CH2O in the reaction with oxygen
atoms, thus altering the measured yield of formaldehyde. Slagle
et al.6 used a different approach for generating methyl radicals
and oxygen atoms. The radicals were produced by the photolysis
of acetone and SO2, respectively. A photoionization mass
spectrometer was used as an analytical tool. Although a Ne
resonance lamp served as a source of VUV radiation (hν )
16.67 and 16.85 eV) for ionization, no attempt to detect CO
(ionization energy (IE) ) 14.01 eV)13 was reported. However,
neither HCO nor CH2 radicals were found, rendering channels
1b and 1e negligible. It was concluded that 1a was the only
open channel at temperatures between 294 and 900 K.
Fairly recently, Seakins and Leone14 performed a product
study focusing on CO as a possible reaction product. They
applied essentially the same method for generation of the
reactants as did Slagle et al., but instead used a time-resolved
FTIR emission technique to detect reaction products. From these
infrared emission studies they could clearly identify vibrationally
excited CO as a product of reaction 1 with a nascent population
of up to V ) 8. They determined an overall yield for CO from
reaction 1 of 0.4 ( 0.2. This would definitely indicate that
channel 1c cannot be neglected as a direct route for producing
CO.
CH3 + O(3P) f products
(1)
The kinetics of this reaction have been studied extensively3-7
by observing either the disappearance of methyl radicals in the
presence of oxygen atoms or the formation of formaldehyde as
the major product. The overall rate coefficient for reaction 1
has been observed to be (1.4 ( 0.3) × 10-10 cm3 molecule-1
s-1, independent of pressure and temperature.8 However, only
a few investigations with regard to the product distribution were
performed. There are several exothermic channels and one
endothermic product channel for this reaction:9,10
CH3 + O f CH2O + H
f HCO + H2
∆H ) -294.1 kJ/mol
∆H ) -354.2 kJ/mol
(1a)
(1b)
(1c)
(1d)
(1e)
f CO + H + H2 ∆H ) -228.7 kJ/mol
f CH + H2O
f OH + CH2
∆H ) -43.9 kJ/mol
∆H ) +29.2 kJ/mol
Niki et al.3 performed flow tube measurements in which
methyl radicals were produced in the reaction of ethylene with
oxygen atoms:
We have chosen to investigate reaction 1 to test the
capabilities of our new apparatus and to attempt to resolve the
discrepancy concerning the branching ratios for reaction 1. We
* Author to whom correspondence should be addressed.
10.1021/jp991157k CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/24/1999

Reaction of CH3 Radicals with O(3P) Atoms
J. Phys. Chem. A, Vol. 103, No. 29, 1999 5723
Figure 1. Schematic diagram of the apparatus.
will present results for the rate coefficient that agree well with
previously measured values and show evidence for the existence
of channel 1c, supporting the findings of Seakins and Leone.
These measurements will demonstrate the suitability of this new
technique for this type of experiment.
We want to focus our research on radical-radical reactions,
which are especially challenging to investigate because potential
side and chain reactions may complicate the data analysis
greatly. Moreover, the recombination of two radicals frequently
produces highly excited adducts that open up more than one
product channel. To fully characterize a reaction, one needs to
observe more than a single species involved in the reaction.
Ideally, all participant species should be detected simultaneously,
eliminating assumptions about the concentrations of unobserved
molecules that are made in experiments with single-species
detection.
are produced in a tubular quartz reactor by photolysis of suitable
precursor molecules. Precursors, reactants, and products are
sampled continuously by allowing gas to escape from the reactor
through a pinhole in the wall, into a region where a portion of
the sampled gas is photoionized. By switching voltages on a
grid assembly at the entrance to the time-of-flight mass
spectrometer [TOFMS: R. M. Jordan Co., D850 Reflectron with
a microchannel plate detector (MCP): C-726] ions can enter
the TOF chamber during a brief time interval (≈12 µs). The
gating/extraction procedure can be carried out at a high rate
(∼20 kHz) enabling us to take “snapshots” of the composition
of the reaction mixture in intervals of 48 µs. This interval was
chosen from the possible choices afforded by the hardware so
that species up to a mass of 150 amu can be acquired without
interference from neighboring mass spectra. Rapid changes in
concentration due to reactions can therefore be observed on a
millisecond time scale. Figure 1 shows a schematic diagram of
the apparatus.
Our experimental setup resembles the one that Slagle et
al.6,15-17 designed with one important exceptionsinstead of
using a quadrupole mass spectrometer we analyze sampled gas
from the photolysis reactor by time-of-flight mass spectrometry
(TOFMS). Thus, direct information on product distributions as
well as reaction rates can be obtained in a single experiment.
Apparatus. The reaction vessel consists of a 43 cm long
quartz tube with an inner diameter of approximately 1 cm. The
sample orifice (1 mm in diameter) is located roughly in the
middle of the tube. The whole reactor is placed inside a vacuum
chamber pumped by an oil diffusion pump backed by a rotary
pump. Connections for a gas inlet and outlet were attached to
flanges on either side of the vacuum chamber also bearing quartz
windows through which radiation from an excimer laser
Experimental Section
A detailed description of the apparatus will be published
elsewhere18 and only an overview will be given here. Radicals

5724 J. Phys. Chem. A, Vol. 103, No. 29, 1999
Fockenberg et al.
(Lambda Physik, LPX 240iMC) was directed into and out of
the reactor. The rectangular cross section of the emitted laser
radiation was reduced to a square shape by a Galilean telescope
made of cylindrical quartz lenses. Typical energies were ∼100
mJ/pulse at the laser, with 30 mJ/pulse exiting the reactor
without any absorber present in the flow tube. The main loss
of beam intensity was caused by the windows and lenses. The
remaining losses could be attributed to absorption by ambient
oxygen in the pathway and the divergence of the laser beam
leaving the flow tube. However, the divergence was sufficiently
low to produce a uniform distribution of radicals along the
tubular reactor from its entrance to the sampling orifice. Intensity
measurements made at the exit window of the flow tube
automatically take atmospheric O2 absorption into account.
A gas mixture composed of precursor molecules for the
radicals as well as bath gas was passed through the tube where
a constant flow was established by a set of three mass-flow
controllers (Tylan General, FC 260). The pressure in the flow
reactor was controlled by throttling the gas flow through a gate
valve at the downstream end of the reactor. Typical flow
velocities ranged from 10 to 20 m/s.
Figure 2. Averaged TOF spectra recorded in experiment #22 (see Table
2) using Ar emission for ionization. 150000 sweeps were co-added for
a single mass spectrum. (a) The mass spectrum of the mixture before
the laser pulse; (b) the mass spectrum taken 240 µs after the photolysis
pulse; (c) the mass spectrum after all methyl radicals were consumed.
The signal below m/e ) 35 and above m/e ) 75 is multiplied by a
factor of 15 for clarity.
Gas that left the tube through the orifice first passed through
a U-shaped skimmer that had windows cut into its wall through
which the ionizing radiation emitted from a hollow-cathode lamp
(McPherson, Model 630) was directed. The lamp was operated
with either Ne (hν )16.67 and 16.85 eV) or Ar (hν ) 11.62
and 11.83 eV) in the discharge at low pressures (p e 400
mTorr). A glass capillary, which also acts as a differential
pumping barrier, is used to guide the photons from the lamp to
the ionization region. The TOF chamber is pumped by a
turbomolecular pump. Starting a few hundred microseconds
before each laser shot, a sequence of 85-170 mass spectra were
typically acquired covering 4-8 ms. After averaging typically
100 000 to 200 000 laser shots, the counts under each mass peak
were then integrated and the sum was plotted against the time
relative to the laser initiation.
For reference measurements, mixtures of CO (Matheson,
purity > 99%) and CH2O with He were prepared in 3 L glass
bulbs. Formaldehyde was made by heating para-formaldehyde
(Aldrich) to a temperature around 100 °C. CH2O was collected
in a liquid nitrogen trap, transferred to a glass bulb on warming,
and diluted with He. Typically, this was performed a sufficiently
long time before use to ensure complete mixing (at least 5 h),
but not so long as to lose formaldehyde through polymerization.
Typical concentrations in the reactor were (0.5-1.0) × 1014
molecules cm-3 (1.5-3.0 mTorr) for SO2 or CH3Br, and (0.5-
1.5) × 1013 molecules cm-3 (0.15-0.45 mTorr) for acetone.
At laser pulse energies of 15-30 mJ/cm2 (λ ) 193 nm),
measured behind the exit window of the reactor tube, 10-20%
of the SO2, approximately 1.5% of the CH3Br, and 3-6% of
the acetone were photolyzed typically. The attenuation of the
laser beam due to absorption in the reactor was small and
therefore was not suitable for a quantitative determination of
radical concentrations formed in the photolysis. Instead, the
initial concentrations of radicals were calculated according to
the relative drop in signal multiplied by the concentration of
precursor molecules (see Figure 4). Unfortunately, the small
loss of bromomethane was almost completely buried in the
scatter of the CH3Br signal so that only estimates of the methyl
radical concentration could be given when this source of methyl
To check for fluctuations in the detection efficiency of the
apparatus, calibration measurements on absolute concentrations
of the main stable species involved in this study were done on
a daily basis. Although the determined calibration coefficients
could vary by more than 20% from day to day, the cracking
pattern of any species in the photoionization process were very
reproducible.
Experiment. For the investigation of the reaction of methyl
radicals with oxygen atoms we used either acetone or bro-
momethane as precursor species for methyl radicals, whereas
oxygen atoms were generated by the photolysis of SO2:
CH3COCH3 + hν (193 nm) f 2CH3 + CO
CH3Br + hν (193 nm) f CH3 + Br
(3)
(4)
radicals was used. On the basis of the absorption coefficients
8,19,20
SO2 + hν (193 nm) f SO + O(3P)
(5)
for CH3Br and SO2
and the actual concentration of both
species in the reactor, one can also calculate the relative drop
in the CH3Br concentration assuming a quantum efficiency of
unity for dissociation for both molecules. The measured and
the calculated values agree very well. The overall pressure in
the reactor was generally 1 Torr He at a flow velocity of 10-
20 m/s. The laser was pulsed at a rate of 8-10 Hz. During an
experiment, the energy per pulse of the excimer laser, flow
controller values, temperature, pressure, and the intensity of the
ionization radiation were constantly recorded. The TOF data
were normalized for differences in excimer pulse energy and
hollow cathode lamp emission where necessary.
All gases were stored as dilute mixtures in He in 20 L glass
bulbs, except SO2, which was purchased as a 5% mixture in
He (Praxair, purity: SO2 g 99.6%, He UHP-grade). Prior to
use, acetone (Mallinckrodt, purity 99.7%) was degassed in
several freeze-pump-thaw cycles, released into the storage
vessel, and pressurized with He (Praxair, UHP grade). The
concentration ratio was determined according to the partial
pressures (MKS Baratron Model 622). Bromomethane (Mathe-
son, purity > 99%) was used without purification.

Reaction of CH3 Radicals with O(3P) Atoms
J. Phys. Chem. A, Vol. 103, No. 29, 1999 5725
Figure 3. Ion signals plotted vs time for several species acquired in
experiment #22 (see Figure 4 and Table 2). The lines in plot (b) are
fits of simple exponential decay and rise functions to the data. The
calculated first-order rate constants are k′ ) 1390 s-1 (s) and k′′ )
1590 s-1 (- - -).
Figure 4. Ion signals associated with O(3P), SO2, and acetone plotted
vs time acquired in experiment # 5 (see Table 2) using Ne emission
for ionization.
that only reaction rates and not branching ratios could be
analyzed in these experiments.
Two noble gases were used as discharge media for the hollow
cathode lamp, delivering mainly photons of either 16.67/16.85
eV (Ne) or 11.62/11.83 eV (Ar). Even though SO2 (IE ) 12.5
eV)21 should not be ionized using Ar, a signal could be clearly
detected at m/e ) 64 (see Figure 2). This might be explained
by a contamination of the primary emission with shorter
wavelength radiation, emitted perhaps from ionized Ar. This
might also cause the occurrence of a signal at m/e ) 32 after
SO2 was photolyzed producing SO (IE ) 10.29 eV)22 and O(3P)
(IE ) 13.62 eV)23 (see Figure 3). Highly excited SO+ ions might
crack into S+ ions [IE(S) ) 10.36 eV]23 and oxygen atoms,
although there should not be enough energy available for
ionization and dissociation after absorption of one photon. The
general advantage of using longer-wavelength radiation for
ionization is that cracking is reduced or eliminated. The
dissociation of ionized acetone [IE ) 9.7 eV,24 appearance
potential (AP) for CH3+ AP(CH3+) ) 15.6 eV,25 AP(CH3CO+)
Calibration of Concentration. When Ne was used as the
discharge gas, the photon energy exceeded the threshold for
ionization and dissociation of almost all parent molecules. To
determine the quantitative amount of each cracking channel,
mass spectra of each individual species were recorded in separate
experiments. These measurements were usually carried out by
acquiring mass spectra of the selected species at different
concentrations. The counts under each mass peak were summed
up and plotted against the concentration. A linear fit through
these points gives the individual absolute calibration coefficients,
CC(a,n), for a specific cracking channel “n” of molecule “a”.
However, the absolute values of the calibration coefficients
depend on a number of different parameters (temperature,
pressure, photoionization wavelength and intensity, etc.) which
could differ quite substantially over an extended period of time.
To eliminate these variations it is easier to work with ratios of
calibration coefficients, CR(a,m:b,n) ) CC(a,m)/CC(b,n), which
are only dependent on the wavelength of the ionizing radiation.
Since one of our main interests lies in determining relative
product yields, which depend only on concentration ratios, this
restriction has no influence on the analysis. Concentration ratios
) 10.5 eV24] into CH3 ions did not occur, leaving the m/e )
+
15 channels (IE ) 9.84 eV)26 essentially free from background.
Similarly, low background signals could be collected for CH2O
(m/e ) 30, IE ) 10.88 eV)27 (see Figure 2). However, CO (IE
) 14.01 eV)13 could not be detected at this photon energy, so

5726 J. Phys. Chem. A, Vol. 103, No. 29, 1999
Fockenberg et al.
TABLE 1: Calibration Ratios for CO, CH2O, and Acetone at an Ionization Photon Energy ) 16.67/16.85 EV (Ne)a
molecule channel
acetone 15
acetone 28
acetone 29
CO 28
CH2O 28
CH2O 29
CH2O 30
CH2O 30
acetone 43
1.368 ( 0.14
0.255 ( 0.01
0.172 ( 0.014
1.237 ( 0.08
1
0.09 ( 0.01
0.00958 ( 0.00134
0.017 ( 0.0015
0.206 ( 0.04
a Ratios were calculated by dividing the individual calibration constants of molecules in top row through those in the first column.
can be determined via:
and below 1 Torr.11 In addition, the CH3 concentration is so
small compared to the oxygen concentration that the rate of
recombination of methyl radicals is negligible with regard to
the rate of their reaction with O(3P). The maximum acceleration
of the measured reaction rate due to reaction 8 can be estimated
to be only 5% for [O]/[CH3] ≈ 7.
[a]
[b]
C(a,m) C(b,n)
C(a,m)
) /CR(a,m:b,n)
)
/
CC(a,m) CC(b,n) C(b,n)
where C(i,j) denotes measured counts in channel m/e ) “j” of
molecule “i”. Some of these ratios are listed in Table 1 for
different acetone, CO, and CH2O channels compared to the m/e
) 43 channel of acetone and the m/e ) 30 channel of
formaldehyde. These calibration experiments were routinely
performed with different mixtures.
Other reactions (reactions 9-12), that might interfere directly
or indirectly are too slow under the experimental conditions to
have a measurable influence on the acquired profiles.11,30,31
CH3 + SO (+ M) f products (+ M)
CH3 + SO2 (+ M) f products (+ M)
H + CH3Br f CH3 + HBr
(9)
(10)
(11)
(12)
Kinetic Analysis. The concentrations of the precursor species
(CH3COCH3/CH3Br and SO2) were chosen so that the oxygen
atom concentration was always in excess of the methyl radical
concentration resulting in [O]/[CH3] ratios ranging from 7 to
50. For the “highest” concentrations of methyl radicals ([O]/
[CH3] e 10) the data were treated according to second-order
kinetics, whereas pseudo-first-order conditions were assumed
for lower methyl concentrations. However, the difference in rate
coefficients calculated with both methods was always smaller
than 10%, which lay well inside the range of experimental
scatter. Also, heterogeneous loss of oxygen atoms, methyl
radicals, or any other species on the reactor wall could be
neglected on the time scale of the reaction (τ e 4 ms). In cases
where Ne was used in the hollow cathode lamp, the concentra-
tion of oxygen atoms could be observed directly. Practically
no loss of oxygen atoms could be detected in any experiment;
this was also assumed to be valid when Ar emission was used
for photoionization and oxygen atoms could not be ionized.
Interestingly, mass peaks at m/e ) 80 and 96 (see Figure 2)
suggest the existence of slow side reactions involving sulfur-
containing compounds. The signal at m/e ) 96 can be explained
by the slow recombination of two SO molecules (k6 ) 4.4 ×
O + CH2O f OH + HCO
SO and SO2 might react with methyl radicals, but no decrease
in either signal was observed in any experiment. In addition,
the possible recombination channels of reactions 9 and 10 should
show up at m/e ) 63 and 79, respectively, which were not
detected. These observations are supported by findings of Slagle
et al.6 who investigated the title reaction under similar conditions
and concluded that reactions 9 and 10 can be neglected.
Therefore, in the case of pseudo-first-order conditions, both
the decay of the methyl radicals and the rise of formaldehyde
as the major product of reaction 1 could be generally fitted to
simple exponential functions according to
C(CH3) ) [CH3]0 × e-k′t
C(CH3) ) [CH2O]max × (1 - e-k′′t
)
10-31 cm6 molecule-2 s-1 28
tration compared to SO:
) generating S2O2 in small concen-
Under second-order conditions the following two expressions
were used to model the CH3 and CH2O signal profiles:
SO + SO + M f S2O2 + M
(6)
e-kh′t
[CH3]0
[O]0 - [CH3]0
C(CH3) ) [CH3]0 ×
×
The channel at m/e ) 80 can have its origin either in SO3+ (IE
) 13.15 eV)21 or S2O+ (IE ) 10.6 eV)22 ions. However, the
recombination of oxygen atoms with sulfur dioxide is too slow
to produce detectable amounts of SO3 within the observation
time (k7 ) 1.0 × 10-33 cm6 molecule-2 s-1):29
[O]0
1 -
e-kh′t
[O]0
C(CH2O) ) [CH2O]max
×
e-kh′′t
[CH3]0
[O]0 - [CH3]0
1 -
×
O + SO2 + M f SO3 + M
(7)
[O]0
1 -
e-kh′′t
(
)
[O]0
Slagle et al.6 detected very small amounts of sulfur atoms
produced in the photolysis of SO2 which might undergo
reactions with SO or SO2 to form S2O. Unfortunately, the
literature on reactions involving S atoms is very sparse, and
this particular channel remains unconfirmed.
The slight decay in the concentration of methyl radicals in
experiments without SO2 can be explained easily by the radical
recombination reaction:
where kh ) k × [O]0 × (1 - [CH3]0/[O]0). [CH3]0/[O]0 ratios
were obtained from the initial radical concentrations (see Table
2). In the data analysis, [CH3]0 and k′ were floated freely,
whereas [CH2O]max was determined by averaging the last data
points in the time profile and only k′′ was fitted (see Figures
3b and 5a,b). The first-order rate constants, k′ and k′′, were then
plotted against the oxygen atom concentration and fitted to a
straight line whose slope gave the second-order rate coefficient
(see Figure 6). For the Ar-lamp experiments, signals at m/e )
CH3 + CH3 + M f C2H6 + M
(8)
+
However, the rate coefficient of this reaction has an upper
15 and 30 were identified to be CH3 (IE ) 9.84 eV,26 AP-
value of only e 5 × 10-11 cm3 molecule-1 s-1 at pressures at
(CH2+) ) 15.09 eV32) and CH2O+ (IE ) 10.88 eV,27 AP-

Reaction of CH3 Radicals with O(3P) Atoms
J. Phys. Chem. A, Vol. 103, No. 29, 1999 5727
TABLE 2: Experimental Conditions, Calculated Product Yields, and First-Order Rate Constantsa
Ptotal
Torr
,
VUV
gas
[SO2],
[precursor],
1013cm-3
[CH3],
[O(3P)],
CO yield,
%
CH2O yield,
%
k′ (CH3),
k′′ (CH2O),
#
T, K
1013cm-3
1011cm-3
1012cm-3
s-1
s-1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2.5
2.5
2.5
1
1
1
1
1
1
1
300
300
395
300
300
300
299
299
299
298
299
299
298
298
299
297
297
298
300
300
299
299
301
300
300
295
296
Ne
Ne
Ne
Ne
Ne
Ne
Ne
Ne
Ne
Ne
Ne
Ne
Ne
Ne
Ne
Ne
Ne
Ne
Ar
Ar
Ar
Ar
Ar
Ar
Ar
Ar
Ar
10.4
15.0
8.53
7.26
10.1
6.14
6.7
8.48
10.1
5.35
6.07
8.28
4.6
10.4a
9.0a
≈1.5
≈1.3
≈0.9
14.4
19.7
7.97
≈1.0
≈1.4
≈1.7
≈0.9
6.3
7.9
3.7
13
5.3
4.9
3.5
≈0.6
17.3
16.3
3.7
20.4
27.2
14.0
12.3
14.6
8.7
18.1 ( 4.0
17.9 ( 5.1
16.1 ( 7.0
14.2 ( 7.8
14.9 ( 9.6
12.5 ( 10.1
18.0 ( 5.4
20.2 ( 5.0
17.5 ( 4.5
13.6 ( 14
29.7 ( 4.8
29.2 ( 4.6
21.8 ( 6.7
15.8 ( 7.2
19.4 ( 7.9
9.1 ( 9
2909
3887
2637
2193
3220
1678
2652
4547
4426
2194
2723
2891
1785
3698
1237
1072
2500
2900
2249
2378
3455
1590
1149
2330
2957
1209
6.27a
1.13b
1.61b
0.725b
6.46a
8.07a
10.1a
5.05a
0.65b
0.91b
0.52b
1.4b
76.6 ( 23
77.0 ( 23
72.3 ( 22
1901
2521
1099
11.2
14.1
17.2
8.84
11.7
15.1
8.43
20.9
6.46
6.75
11.6
12.4
12.6
11.0
16.8
7.41
6.1
86.8 ( 26
73.3 ( 22
90.7 ( 27
86.7 ( 26
95.0 ( 29
63.3 ( 19
117.8 ( 35
2029
2661
1363
2681
1035
881
12
3.91
5.41
5.72
6.36
8.03
7.03
8.6
3.99
3.91
6.02
10.1
5.21
0
0.46b
0.5b
3.45b
3.42a
1.58b
1.5b
28.8 ( 8.9
14.9 ( 15
1675
2069
2092
2796
1390
1003
2140
2803
0.8b
0.45b
0.5b
3.1
3.8
2.5
21.4
7.3
0.66b
2.1b
11.1
17.0
6.81
0
1
1
0.5b
18.1b
25.7
130
a Precursor: CH3Br. b Acetone.
(HCO+) ) 11.97 eV, AP(CO+) ) 14.1 eV33), respectively. In
the Ne-lamp experiments where acetone was used as the
precursor, the peak at m/e ) 14 originating from dissociation
of excited CH3+ into CH2+ was taken for the fitting procedure
because the separation of the actual methyl radical signal from
the acetone cracking into the same channel was not possible.
For bromomethane (IE ) 10.53 eV,34 AP(CH3+) ) 12.8 eV,
AP(CH2+) ) 14.7 eV35) neither channel, m/e ) 14 nor 15, could
be used so that only kinetic data from the formaldehyde
formation could be obtained. The fit for formaldehyde formation
was checked against the increase in the signal at m/e ) 29,
which should originate mainly from cracking of CH2O+ into
HCO+. For this test the formaldehyde fit for m/e ) 30 was
multiplied by the respective calibration ratio (see Table 1)
resulting in the equivalent value for the m/e ) 29 channel of
the CH2O+ dissociation. Afterward, the offsets caused by
dissociation of CH3COCH3+ were added. The resulting profiles
described the actual data very well for all experiments, confirm-
ing the fit and the identity of the m/e ) 30 channel as
formaldehyde. For the fitting procedure, the data points were
shifted in time by one time step (48 µs) to account for the finite
travel time of the molecules from the orifice to the ionization
region.
As can be seen in Figure 6, the reaction rates obtained from
the rise time of formaldehyde (shown as open triangles down)
are generally larger by up to 20% compared to the rates
calculated from the decrease in the CH3 signal (solid triangles
up). The most likely source of error lies in the fitting procedure
(see above), for which the infinity value, [CH2O]max, was fixed.
Even a slow loss reaction for CH2O molecules could easily lead
to an overestimate of the observed reaction rate. As a test we
generated curves composed of two exponential functions
mimicking a fast rise and a slow decrease and fitted single-
exponential functions to these. Even small decay rates of 1-2%
compared to the rate of production resulted in apparent reaction
rates 5% to 10% higher. Unfortunately, we were not able to
observe such small decay rates (≈10 s-1) under the experimental
conditions chosen. A candidate for this loss could be a temporary
adsorption of formaldehyde on the reactor walls. However,
second-order rate coefficients calculated separately from the
methyl decay or formaldehyde increase lie within 10% of the
second-order rate coefficient calculated considering all data. In
general the statistical error for each evaluated rate can be given
by approximately 10% (2σ) for the methyl decay and 20% (2σ)
for the formaldehyde increase.
With bromomethane as precursor, the data show a large
amount of scatter, which we attribute to wall reactions caused
by surface contamination with bromine atoms or bromomethane
itself. Therefore, only data from experiments in which acetone
as precursor was used were taken for both the CH3 decay and
CH2O increase. The overall second-order rate coefficient for
the reaction CH3 + O(3P) f products at P ) 1 Torr (He) and
T ) (299 ( 2) K is given by
k1 ) (1.7 ( 0.3) × 10-10 cm3 molecule-1 s-1
in excellent agreement with the recommended value of 1.4 ×
10-10 cm3 s-1 11
The error given here is twice the standard
.
deviation.
Product Analysis. Products of reaction 1 identified in this
investigation were CH2O and CO. The high ionization energy
of CO made it necessary to use high-energy photons from a Ne
discharge for ionization. Under these conditions there were
several sources of the signal at m/e ) 28: N2+ from background
gas, CO+ from cracking of CH2O and acetone, and CO+ from
photolysis of acetone and finally from reaction 1c. Taking the
actual counts of acetone at m/e ) 43 and multiplying the number
by the calibration ratio CR(acetone,28: acetone,43) gives the
contribution to the counts in channel 28 from cracking of
acetone. The amount of CO originating from acetone photolysis
can be determined by multiplying the drop in acetone counts at
m/e ) 43 by CR(CO,28: acetone,43). After subtraction of all
contributions, the remaining counts at m/e ) 28 were associated
with CO produced in reaction 1. Figure 5 c shows the evolution

5728 J. Phys. Chem. A, Vol. 103, No. 29, 1999
Fockenberg et al.
Figure 6. Second-order plot of first-order rate constants vs the
concentration of oxygen atoms under various conditions. (a) Ar
emission: observed were masses 15 (2) and 30 (3), both with acetone
as precursor. (b) Ne emission: observed were mass 30 (b) with CH3-
Br as precursor, and masses 14 (2) and 30 (3) both with acetone as
precursor. The solid line (s) in (a) and (b) is obtained by a linear fit
to all first-order rate constants with acetone as a precursor.
laser fires and after reaction 1 is complete, itemized by
contributions of its sources. In experiments with acetone or CH3-
Br as the precursor molecule approximately 10% or 20%,
respectively, of the total counts at this mass could be attributed
to CO from reaction 1.
The quantification of the yields was carried out in two
different ways. First, the final amount of formaldehyde produced
was compared with the drop in concentration of acetone after
photolysis, assuming a methyl quantum yield of two (see
Discussion on correction). Second, the increase in the CO
concentration was compared to the increase in CH2O. The
concentration ratios were calculated from the differences in
counts of acetone (m/e ) 43), CH2O (m/e ) 30), and CO (m/e
) 28) before the excimer laser pulse and after all methyl radicals
were consumed, and corrected according to the calibration ratios
from Table 1. The results are shown in Table 2 and Figure 7.
The average CH2O yield for reaction 1 was Φ1(CH2O)) 0.84
( 0.15 whereas for the CO yield, a value of Φ1(CO) ) 0.18 (
0.08 was found for T ) (299 ( 2) K and P ) 1 Torr (He),
confirming the findings of Seakins and Leone.14 These results
Figure 5. Ion signals of reactants and products plotted vs time recorded
in experiment # 5 (see Figure 4 and Table 2). The solid lines in plots
(a) and (b) are fits to second-order decay and rise functions. The broken
lines were obtained by normalization of the data according to the
respective calibration ratios and adjusting for different offsets. Plot (c)
shows data for channel m/e ) 28 and the contributions to the signal
counts from the different sources. First-order rate constants are (a) k′
) 2521 s-1, (b) k′′ ) 3220 s-1
.
of the signal at m/e ) 28 and the average values before the

Reaction of CH3 Radicals with O(3P) Atoms
J. Phys. Chem. A, Vol. 103, No. 29, 1999 5729
nm did not detect either CH4 or CH2CO, thus confirming the
upper limit given by Lightfoot et al. However, North et al. found
that all detected H atoms could be attributed to secondary
photolysis of methyl fragments, leaving the 3% value of the
CH2COCH3 + H channel questionable.
We attempted to detect CH2 radicals (IE ) 10.35 eV38) in
separate experiments photolyzing acetone at the highest obtain-
able laser intensities. Using Ar as the discharge gas the formation
+
of CH2 from CH3 was not possible, providing a low back-
ground for the detection of CH2 radicals. At bath gas (He)
pressures of 1 Torr, electronically excited singlet CH2 should
be quickly quenched (τ ≈ 10 µs39) into the less reactive ground
triplet state. Assuming a reactive removal of singlet methylene
on every collision with acetone, less than 50% of the CH2
produced is supposed to elude detection. However, even at
methyl radical concentration of larger than 1 × 1013 molecules
cm-3 no signal at m/e ) 14 could be observed. Assuming that
the ionization cross section of CH2 is comparable to that of
CH3, we concluded that the photon density was not high enough
to produce methylene radicals in any significant amount.
As mentioned earlier, the chemistry of the system with
bromomethane as the precursor species does not appear to be
as clean as in the case with acetone. We assume that hetero-
geneous reactions on the reactor walls involving bromine species
are responsible for the observed fluctuations in the reaction rate,
so that it is unlikely that the products are either CO or CH2O.
Therefore, the measured concentration ratio of these species
should simply reflect exactly the product distribution of reaction
1. As can be seen in Figure 7, consistent yields of CO were
determined from experiments using either CH3Br or acetone as
a methyl radical precursor.
Since ethane produces a strong dissociative ionization signal
at m/e ) 28, we also considered the effect of methyl radical
recombination to ethane as a possible contribution to the total
signal at m/e ) 28. Calculations based on rate coefficients and
measured ionization efficiencies suggest the effect should be
undetectable at our noise levels. Furthermore, no correlation
exists between the [O]/[CH3] ratio and the computed CO yield,
justifying the neglect of corrections for ethane formation and
dissociative ionization.
The surface adsorption of formaldehyde mentioned above
would definitely alter the measured CH2O and CO yields
directly. However, it is very unlikely that the loss amounted to
more than a few percent, and even with an unrealistic correction
for the formaldehyde yield to 100%, the yield for CO (which
depends on the [CO] to [CH2O] ratio), would drop only to 85%
of its calculated value.
Figure 7. Plot of product yields for CH2O and CO from reaction 1.
Average values are marked by (s) and (- - -), respectively. The data
are numbered according to Table 2.
also indicate that formaldehyde and carbon monoxide are the
dominant carbon-containing products of reaction 1.
Discussion
Two precursor substances were used for these investigations,
and each had its advantages and disadvantages. The photolysis
of acetone at λ ) 193 nm is essentially a clean source for methyl
radicals as Lightfoot et al. showed.36 According to these authors,
the channel leading to 2CH3 + CO accounts for ≈95% of the
total dissociation. The apparent CO yield of reaction 1 would
have to be corrected upward by approximately 14% whereas
the CH2O yield would have to be raised by 0.95-1 if 5% of the
photolyzed acetone went into minor channels. The reported
minor products are CH2COCH3 + H (∼3%), CH4 + CH2CO
(e 2%), and CH2 from secondary photolysis of methyl radicals.
A subsequent reaction of CH2COCH3 with O atoms might lead
directly or indirectly to CO or CH2O, thus obscuring any
production of carbon monoxide or formaldehyde from reaction
1. In turn, this would lower the apparent yield for both reaction
products by a slight amount depending on the actual product
distribution of the reaction CH2COCH3 + O(3P). Studies
performed by North et al.37 on acetone photolysis at λ ) 193
Averaging all measured product yields of CO from reaction
1, Φ1(CO), including the results for the formaldehyde yield with
Φ1(CO) ) 1 - Φ1(CH2O) results in
Φ1(CO) ) 0.17 ( 0.11
for T ) (299 ( 2) K and P ) 1 Torr (He). The error given is
one standard deviation. Correcting the product yield according
to the quantum yield measured by Lightfoot et al. would lower
Φ1(CO) by less than 10%, which lies well inside the error limits.
Another source of error might come from reactions of
vibrationally excited photolysis products. A number of authors
have investigated the energy distribution in the photofragments
produced in the photolysis of acetone at λ ) 193 nm.37,40,41
The internal energy in the methyl fragments is considerable (Eint
) 98.7 kJ/mol)37 which might influence the dynamics as well
as kinetics of reactions 1. However, deactivation through
collisions with the bath gas, He,42,43 or the reactor walls is

5730 J. Phys. Chem. A, Vol. 103, No. 29, 1999
Fockenberg et al.
sufficiently fast, resulting in a lifetime of vibrationally excited
methyl radical of τ e 100 µs. The time constant for the reaction
is up to 10 times longer than that for deactivation, so that a
possible influence on the mechanism for reaction 1 should be
minimal. Moreover, a few experiments were performed at bath
gas pressures of P ) 2.5 Torr, accelerating the deactivation
without any detectable effect on kinetics or product distribution.
The literature on the photodissociation dynamics of bro-
momethane at λ ) 193 nm is unfortunately very sparse. Van
Veen et al.44 investigated the translational energy distribution
in the methyl and bromine fragments and concluded that the
internal energy in the CH3 fragments is considerably less than
in the case of acetone photodissociation (Eint ) 30.6 and 18.5
kJ/mol for the Br/Br* channels). Most of the excitation is in
the ν2 (CH2 umbrella) vibrations, which are quenched much
faster by He atoms than the ν3 (antisymmetric C-H stretch)
vibration.42,43 In terms of dynamical properties, CH3Br is
definitely favored over acetone, but the presence of bromine
compounds tends to introduce some disadvantages (see above).
The photoionization of the CO product at Ne wavelengths
also needs additional treatment. The absorption spectrum of
ground-state CO(X1Σ+) is very structured in this region,
stemming from Rydberg transitions converging to the CO+(A2Π)
state together with non-Rydberg transitions and an underlying
photoionization continuum.45,46 According to Cook et al.45 the
bands are subject to preionization with an efficiency around
20%. Keeping in mind that the CO produced in reaction 1 can
be highly vibrationally excited14 the measured CO(V) ensemble
might exhibit a different overall ionization efficiency after
absorption of a VUV photon than a thermalized ensemble.
Although Rydberg states might be accessible from vibrationally
excited CO it seems unlikely to find this situation realized. It
is also conceivable that the vibrational excitation leads to an
enhanced dissociation of CO into neutral fragments instead of
ionization resulting in an underestimation of the measured
product yield. All these effects of the vibrational population of
CO are difficult to predict exactly but were considered to be
small and generally omitted in the error analysis. This problem
will be addressed in the future by using the He line at 21.22
eV for ionization because the absorption spectrum for CO above
60 nm becomes flat and the ion yield approaches unity. Any
vibrational excitation should therefore not alter the ionization
efficiency.
observation of the kinetics and mechanism of fast radical-
radical reactions possible. The reaction of methyl radicals and
oxygen atoms was investigated. The results of Seakins and
Leone,14 showing the existence of a second reaction channel
producing carbon monoxide, was confirmed. The 17% yield
determined in this work would indicate that a nonnegligible
amount of the CO formed in important combustion systems
would come directly from reactions 1. The remaining ≈80%
would have to be produced in an oxidation chain beginning with
CH2O. The exact influence on flame properties can only been
determined in simulation calculations. Further experiments,
especially on the temperature dependence of the branching ratio,
are necessary to establish these results under more combustion-
relevant conditions and are planned for the future.
Acknowledgment. The authors thank Dr. Herbert J. Bern-
stein for fruitful discussions and assistance in designing and
assembling the data acquisition system. This work was carried
out at Brookhaven National Laboratory under Contract DE-
AC02-98CH10886 with the U.S. Department of Energy and
supported by its Division of Chemical Sciences, Office of Basic
Energy Sciences.
References and Notes
(1) Miller, J. A.; Kee, R. J.; Westbrook, C. K. Annu. ReV. Phys. Chem.
1990, 41, 345.
(2) Warnatz, J.; Mass, U.; Dibble, R. W. Combustion: Physical and
Chemical Fundamentals, Modeling and Simulation, Experiments, Pollutant
Formation; Springer: Berlin Heidelberg New York, 1996.
(3) Niki, H.; Daby, F. E.; Weinstock, B. J. Chem. Phys. 1968, 48, 5729.
(4) Washida, N. J. Chem. Phys. 1980, 73, 1665.
(5) Plumb, I. C.; Ryan, K. R. Int. J. Chem. Kinet. 1982, 14, 861.
(6) Slagle, I. R.; Sarzynski, D.; Gutman, D. J. Phys. Chem. 1987, 91,
4375.
(7) Oser, H.; Walter, D.; Stothard, N. D.; Grotheer, O.; Grotheer, H.
H. Chem. Phys. Lett. 1991, 181, 521.
(8) Atkinson, R.; Baulch, D. L.; Cox, R. A.; R. F. Hampson, J.; Kerr,
J. A.; Troe, J. J. Phys Chem. Ref. Data 1992, 21, 1125.
(9) Tsang, W. Heats of Formation of Organic Free Radicals by Kinetic
Methods in Energetics of Organic Free Radicals; Simoes, J. A. M.,
Greenberg, A., Liebman, J. F., Eds.; Blackie Academic and Professional:
London, 1996; p 22.
(10) Chase, M. W., Jr.; Davies, C. A.; Downey, J. R., Jr.; Frurip, D. J.;
McDonald, R. A.; Syverud, A. N. J. Phys. Chem. Ref. Data, Suppl. 1 1985,
14, 1.
(11) Baulch, D. L.; Cobos, C. J.; Cox, R. A.; Frank, P.; Hayman, G.;
Just, T.; Kerr, J. A.; Murrels, T.; Pilling, M. J.; Troe, J.; Walker, R. W.;
Warnatz, J. J. Phys. Chem. Ref. Data 1994, 23, 847.
(12) Koda, S.; Endo, Y.; Tsuchiya, S.; Hirota, E. J. Chem. Phys. 1991,
95, 1241.
(13) Erman, P.; Karawajczyk, A.; Rachlew-Kallne, E.; Stromholm, C.;
Larsson, J.; Persson, A.; Zerne, R. Chem. Phys. Lett. 1993, 215, 173.
(14) Seakins, P. W.; Leone, S. R. J. Phys. Chem. 1992, 96, 4478.
(15) Slagle, I. R.; Yamada, F.; Gutman, D. J. Am. Chem. Soc. 1981,
103, 149.
(16) Slagle, I. R.; Gutman, D. J. Am. Chem. Soc. 1985, 107, 5342.
(17) Slagle, I. R.; Sarzynski, D.; Gutman, D.; Miller, J. A.; Melius, C.
F. J. Chem. Soc., Faraday Trans. 2 1988, 84, 491.
(18) Fockenberg, C.; Bernstein, H. J.; Hall, G. E.; Muckerman, J. T.;
Preses, J. M.; Sears, T. J.; Weston, R. E. J. ReV. Sci. Instr. 1999, in press.
(19) Manatt, S. L.; Lane, A. L. J. Quant. Spectrosc. Radiat. Trans. 1993,
50, 267.
(20) Phillips, L. F. J. Phys. Chem. 1981, 85, 3994.
(21) Snow, K. B.; Thomas, T. F. Int. J. Mass Spectrom. Ion Processes
1990, 96, 49.
(22) Norwood, K.; Ng, C. Y. Chem. Phys. Lett. 1989, 156, 145.
(23) Kelly, R. L. J. Phys. Chem. Ref. Data 1987, 16, 337.
(24) Trott, W. M.; Blais, N. C.; Walters, E. A. J. Chem. Phys. 1978,
69, 3150.
Concerning a possible mechanism for reactions 1, we refer
to the paper of Seakins and Leone.14 The initial step is supposed
to be the association of CH3 and O(3P) to form an activated
methoxy radical, CH3O*, with internal energy of approximately
380 kJ/mol above the ground state. This is enough energy to
overcome barriers to elimination of either H atoms, leading to
CH2O, or molecular hydrogen, leaving an energized HCO*
radical behind which might subsequently eliminate another H
atom, leading to CO. Other possible pathways include isomer-
ization of CH3O* to CH2OH* followed by stepwise elimination
of H2 and H atoms. However, as Seakins and Leone pointed
out, there is either considerable disagreement about the barrier
heights of the various transition states, or information about them
is not available at all.
Summary
(25) Powis, I.; Danby, C. J. Int. J. Mass Spectrom. Ion Phys. 1979, 32,
27.
(26) Berkowitz, J.; Ellison, G. B.; Gutman, D. J. Phys. Chem. 1994,
98, 2744.
(27) Ohno, K.; Okamura, K.; Yamakado, H.; Hoshino, S.; Takami, T.;
Yamauchi, M. J. Phys. Chem. 1995, 99, 14247.
(28) Herron, J. T.; Huie, R. E. Chem. Phys. Lett. 1980, 76, 322.
The reaction of methyl radicals with oxygen atoms has been
investigated with a new apparatus. TOF mass spectrometry
combined with a special grid assembly at the entrance of the
flight tube that allows extraction of ions at a high repetition
rate and an efficient counting apparatus make the simultaneous

Reaction of CH3 Radicals with O(3P) Atoms
J. Phys. Chem. A, Vol. 103, No. 29, 1999 5731
(29) Singleton, D. L.; Cvetanovic, R. J. J. Phys. Chem. Ref. Data 1988,
17, 1377.
(30) James, F. C.; Kerr, J. A.; Simons, J. P. J. Chem. Soc., Faraday
Trans. 1 1973, 69, 2124.
(37) North, S. W.; Blank, D. A.; Gezelter, J. D.; Longfellow, C. A.;
Lee, Y. T. J. Chem. Phys. 1995, 102, 4447.
(38) Reineke, W.; Strein, K. Ber. Bunsen-Ges. Phys. Chem. 1976, 80,
343.
(31) Callear, A. B.; Cooper, I. A. J. Chem. Soc., Faraday Trans. 1990,
86, 1763.
(32) Chupka, W. A.; Lifshitz, C. J. Chem. Phys. 1968, 48, 1109.
(33) Guyon, P. M.; Chupka, W. A.; Berkowitz, J. J. Chem. Phys. 1976,
64, 1419.
(39) Ashfold, M. N. R.; Fullstone, M. A.; Hancock, G.; Ketley, G. W.
Chem. Phys. 1981, 55, 245.
(40) Donaldson, D. J.; Leone, S. R. J. Chem. Phys. 1986, 85, 817.
(41) Hall, G. E.; Vanden Bouten, D.; Sears, T. J. J. Chem. Phys. 1991,
94, 4182.
(34) Kimura, K.; Katsumata, S.; Achiba, Y.; Yamazaki, T.; Iwata, S.
Ionization energies, Ab initio assignments, and valence electronic structure
for 200 molecules. In Handbook of HeI Photoelectron Spectra of Funda-
mental Organic Compounds; Japan Scientific Soc. Press: Tokyo, 1981.
(35) Kaposi, O.; Riedel, M.; Vass-Balthazar, K.; Sanchez, G. R.; Lelik,
L. Acta Chim. Acad. Sci. Hung. 1976, 89, 221.
(42) Callear, A. B.; Van den Bergh, H. E. Chem. Phys. Lett. 1970, 5,
23.
(43) Donaldson, D. J.; Leone, S. R. J. Phys. Chem. 1987, 91, 3128.
(44) Van Veen, G. N. A.; Baller, T.; De Vries, A. E. Chem. Phys. 1985,
92, 59.
(45) Cook, G. R.; Metzger, P. H.; Ogawa, M. Can. J. Phys. 1965, 43,
1706.
(46) Ogawa, M.; Ogawa, S. J. Mol. Spectrosc. 1972, 41, 393.
(36) Lightfoot, P. D.; Kirwan, S. P.; Pilling, M. J. J. Phys. Chem. 1988,
92 (2), 4938.