A laser photolysis/time-resolved Fourier transform infrared emission study of OH(X 2Π, v) produced in the reaction of alkyl radicals with O(3P)

Article abstract of DOI:10.1063/1.475575

The emission spectra of vibrationally excited hydroxyl radical products formed in the reactions of alkyl radicals with O(³P) atoms are detected using a laser photolysis/time-resolved Fourier transform infrared spectroscopy technique. For the reaction between oxygen atoms and ethyl, the radicals are produced simultaneously by the 193 nm photolysis of the precursors SO₂ and diethyl ketone, respectively. The observed initial OH(v) product vibrational state distribution for the C₂H₅+O(³P) reaction is 0.18±0.03, 0.23±0.04, 0.29±0.05, 0.23±0.07, and 0.07±0.04 for v = 1 to 5, respectively. The population inversion is best explained by a direct abstraction mechanism for this radical-radical reaction. Vibrationally excited hydroxyl radicals are also observed in the O+ethyl, O+n-propyl, and O+i-propyl reactions when using alkyl iodides as precursors of the alkyl radicals, although quantitative detail is not obtained due to competing reaction processes.

Full text of DOI:10.1063/1.475575

A laser photolysis/time-resolved Fourier transform infrared emission study of OH (X 2
Π,v) produced in the reaction of alkyl radicals with O ( 3 P)
Jörg Lindner, Richard A. Loomis, Jody J. Klaassen, and Stephen R. Leone
Citation: The Journal of Chemical Physics 108, 1944 (1998); doi: 10.1063/1.475575
View online: http://dx.doi.org/10.1063/1.475575
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/108/5?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Reaction dynamics of O(1D) + HCOOD/DCOOH investigated with time-resolved Fourier-transform infrared
emission spectroscopy
J. Chem. Phys. 141, 154313 (2014); 10.1063/1.4897418
Photolysis of CH3CHO at 248 nm: Evidence of triple fragmentation from primary quantum yield of CH3 and HCO
radicals and H atoms
J. Chem. Phys. 140, 214308 (2014); 10.1063/1.4878668
Infrared absorption of gaseous benzoylperoxy radical C6H5C(O)OO recorded with a step-scan Fourier-transform
spectrometer
J. Chem. Phys. 135, 224302 (2011); 10.1063/1.3664304
Two-dimensional cross-spectral correlation analysis and its application to time-resolved Fourier transform
emission spectra of transient radicals
J. Chem. Phys. 123, 184104 (2005); 10.1063/1.2074147
The gaseous reaction of vinyl radical with oxygen
J. Chem. Phys. 115, 1742 (2001); 10.1063/1.1382814
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.192.114.19 On: Wed, 17 Dec 2014 21:15:53

A laser photolysis/time-resolved Fourier transform infrared emission study
2
3
of OH„X ⌸,v… produced in the reaction of alkyl radicals with O„ P…
Jorg Lindner, Richard A. Loomis,b) Jody J. Klaassen,c) and Stephen R. Leoned)
¨
JILA, University of Colorado and National Institute of Standards and Technology and University of Colorado,
Department of Chemistry and Biochemistry, University of Colorado, Boulder,
Colorado 80309-0440
a)
͑Received 5 August 1997; accepted 24 October 1997͒
The emission spectra of vibrationally excited hydroxyl radical products formed in the reactions of
alkyl radicals with O(3P) atoms are detected using a laser photolysis/time-resolved Fourier
transform infrared spectroscopy technique. For the reaction between oxygen atoms and ethyl, the
radicals are produced simultaneously by the 193 nm photolysis of the precursors SO2 and diethyl
ketone, respectively. The observed initial OH( ) product vibrational state distribution for the
v
C2H5ϩO(3P) reaction is 0.18Ϯ0.03, 0.23Ϯ0.04, 0.29Ϯ0.05, 0.23Ϯ0.07, and 0.07Ϯ0.04 for
v
ϭ1 to 5, respectively. The population inversion is best explained by a direct abstraction mechanism
for this radical–radical reaction. Vibrationally excited hydroxyl radicals are also observed in the
Oϩethyl, Oϩn-propyl, and Oϩi-propyl reactions when using alkyl iodides as precursors of the
alkyl radicals, although quantitative detail is not obtained due to competing reaction processes.
© 1998 American Institute of Physics. ͓S0021-9606͑98͒01605-5͔
the BAC-MP4 heat of formation calculations of Melius.9 The
INTRODUCTION
total rate constant is 2.2Ϯ0.4ϫ10Ϫ10 cm3 moleculeϪ1 sϪ1
and the branching ratios for the observed product channels
were determined to be 0.23Ϯ0.07, 0.40Ϯ0.04, and 0.32
Ϯ0.07, for the OH channel ͓including ͑1a͒ and ͑1d͔͒, pro-
cesses ͑1b͒ and ͑1c͒, respectively.7 The observation of OH
and the branching ratios of the products both add to the
earlier investigation of Hoyermann and Sievert.4 There is
significantly less information available for the reactions of
larger alkyl radicals with oxygen atoms. A total rate constant
of 1.6ϫ10Ϫ10 cm3 moleculeϪ1 sϪ1 was determined for both
the reactions of n- and i-propyl radicals with O(3P).10 Some
of the product channels for these reactions were identified
The reactions of alkyl radicals with oxygen atoms occur
in the combustion of hydrocarbon species1,2 and during chain
reactions.3 Despite their ubiquitous nature, the dynamics of
oxygenϩalkyl reactions have not been studied extensively.
Kinetic information such as rate constants and vibrational
distributions is frequently not available for various product
branches. An early investigation of alkylϩO reactions was
performed by Hoyermann and Sievert;4 the alkyl radical re-
actants studied included CH3, C2H5, n-C3H7, i-C3H7, and
t-C4H9. These authors determined several product channels
and branching ratios, but OH was not observed. For the
CH3ϩO reaction, Slagle et al.5 observed H2CO while an ad-
and branching ratios were also reported.4
ditional channel for this reaction that produces CO( ) was
2
v
In the present work, vibrationally excited OH(X ⌸, )
v
established by Fourier transform infrared emission studies of
products are directly detected by infrared emission following
the reactions of ethyl, n-propyl, and i-propyl radicals with
O(3P) atoms. The reactions are initiated by laser photolysis
of precursors for oxygen atoms and alkyl radicals, and time-
resolved ͑TR͒ Fourier transform infrared ͑FTIR͒ emission
spectroscopy is used to detect the vibrationally excited hy-
droxyl radical products.11–14 As in many previous
studies,5–7,15–17 the O(3P) atoms are generated by 193 nm
photodissociation of SO2.18–22 The alkyl radicals, R•, are
formed by simultaneous photodissociation of diethyl ketone,
ethyl iodide, n-propyl iodide, or i-propyl iodide.
Seakins and Leone.6
Slagle et al.7 also studied the ethyl radical plus oxygen
atom reaction,
3
C H ϩO P͒→OHϩC H ϭCH ⌬HϭϪ264.5 kJ/mol,
͑
2
5
2
4
2
͑1a͒
→CH3CHOϩH ⌬HϭϪ304.4 kJ/mol, ͑1b͒
→CH2OϩCH3 ⌬HϭϪ318.6 kJ/mol, ͑1c͒
→OHϩCH3CH ⌬Hϭϩ150.7 kJ/mol. ͑1d͒
In experiments using alkyl iodides, RI, as precursors for
the alkyl radicals, the radical–radical reaction to form OH is
clearly observed when using ethyl iodide, n-propyl iodide,
and i-propyl iodide. However, the competing reactions
The values for the heats of formation listed above are mostly
taken from the literature,8 while that of CH3CH is taken from
a͒
¨
¨
Present address: Institut fur Quantenoptik, Universitat Hannover, Welfen-
garten 1, D-30167 Hannover, Germany.
3
b͒
RIϩO P͒→HOI ͒ϩR•,
͑2͒
͑
͑v
National Research Council Research Associate, National Institute of Stan-
dards and Technology.
c͒
Present address: SVT Associates, 7620 Executive Drive, Eden Prairie, MN
55344.
and possibly
d͒
Staff member, Quantum Physics Division, National Institute of Standards
and Technology.
HIϩO͑3P͒→OH ͒ϩI ,
͑v
͑3͒
*
1944
J. Chem. Phys. 108 (5), 1 February 1998
0021-9606/98/108(5)/1944/9/$15.00
© 1998 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.192.114.19 On: Wed, 17 Dec 2014 21:15:53

Lindner et al.: Reaction of alkyl radicals with O(3P)
1945
complicate the observed kinetics and, therefore, prohibit a
careful study of the mechanism for OH formation of the
radical–radical reaction using alkyl iodides as precursors.
Reaction ͑2͒ is most likely the largest competing reaction
the reactions of ethyl, n-propyl, and i-propyl radicals with
oxygen atoms preclude a quantitative assessment of the na-
scent distributions. The magnitude of the OH signal as well
as the relaxed vibrational distributions from the n- and
i-propyl radical reactions are similar to those obtained when
ethyl is produced from ethyl iodide, which in turn can be
compared to diethyl ketone. The results suggest that direct
abstraction may also be the mechanism for these longer
chain alkyl radicals.
since the total rate of reaction is comparable to that of reac-
7,23
tion ͑1͒
and the HOI( ) vibrational bands, which have
v
been characterized previously,24,25 are significantly more in-
tense than the OH( ) emission from the R•ϩO radical–
v
radical reaction. The HOI( ) emission ͑⌬ ϭϪ1 of the OH
v
v
stretch, spanning from 3200 to 3800 cmϪ1͒ overlaps a sig-
nificant portion of the diatomic OH emission spectra.25 Re-
action ͑3͒ most likely plays only a minor role, since HI is
only a 5% product in the 193 nm photolysis of alkyl
iodides26 and the rate constant of the reaction is only 1.6
ϫ10Ϫ12 cmϪ3 moleculeϪ1 s1.10
EXPERIMENT
Vibrationally and rotationally resolved emission spectra
of the products of alkyl radical reactions with oxygen atoms
are recorded using the technique of TR-FTIR spectros-
copy.11–14 The experimental apparatus consists of a pulsed
excimer laser, a flow cell photolysis chamber with fluores-
cence collection optics, and a commercial 0.02 cmϪ1 resolu-
tion, continuous scan FTIR spectrometer with dynamic beam
alignment. Oxygen atoms and alkyl radicals are produced
simultaneously by photodissociation of SO2 and either di-
ethyl ketone or an alkyl iodide with the 193 nm output of a
pulsed ArF excimer laser. The emission from the reaction
mixture is modulated by the interferometer mirror scan of the
FTIR spectrometer. The infrared ͑IR͒ fluorescence is re-
corded at multiple time delays and corresponding time-
dependent emission spectra are obtained using commercial
fast Fourier transform software. Since the methodology of
the data acquisition has been considerably improved re-
cently, a detailed description is given here.
By utilizing a different alkyl radical source, such as di-
ethyl ketone,27 reaction ͑1͒ can be monitored without com-
plications from ͑2͒, ͑3͒, or other parallel processes that
would obscure the OH-forming mechanism for ͑1͒. Diethyl
ketone has been used as a precursor in the C2H5ϩCl(2P)
studies of Seakins et al.28 and in the C2H5ϩO(3P) kinetics
experiments of Slagle et al.7 Although, the photolysis of di-
ethyl ketone produces excited CO, which emits between
27
1950 and 2250 cmϪ1
,
the transition energies of these rovi-
brational lines are precisely known and lie at lower wave
numbers than emission from vibrationally excited OH. An
additional broad and structureless emission feature is ob-
served, spanning from ϳ2700 to ϳ3100 cmϪ1, which is at-
tributed to the C–H stretching vibrations of the ethyl radical
photoproduct.27 As discussed in more detail below, the vi-
brationally excited ethyl most likely does not contribute sig-
Data acquisition
nificantly to the OH( ) population distribution and the ethyl
v
emission only appears as an offset in the baseline for the
narrow OH transition lines.
The infrared light emitted from vibrationally excited
species within the photolysis chamber is focused by a tele-
scope onto both the entrance of the FTIR and a normaliza-
tion detector. Details of the normalization aspect of the ex-
periment can be found in the work of Lindner et al.30
Emission modulated by the interferometer is detected by a 77
K InSb detector and is amplified with commercial current ͑2
␮A/V͒ and voltage (ϫ20) amplifiers resulting in an instru-
ment response time of ϳ3 ␮s; in the following, this modu-
lated light intensity is referred to as the FT signal. A portion
(ϳ20%) of the infrared light is blocked by the optics that
couple the reference HeNe laser into the interferometer and
cannot be detected. This portion of the emission is split off
by a mirror located between the telescope and the entrance of
the interferometer and is focused onto another 77 K InSb
detector to measure the total fluorescence in the chamber;
this is referred to as the FT normalization signal.30
Following amplification, the FT signal is directed to one
or more gated integrators and to an input channel of a com-
mercial multichannel analog to digital ͑A/D͒ board ͑16 bit,
Ͻ7 ␮s sampling time͒ located within a data acquisition com-
puter. The outputs of the gated integrators are connected to
consecutive input channels of the A/D card. The amplified
FT normalization signal is input to a gated integrator, the dc
output of which is recorded by another channel of the A/D
card. The width of the normalization gate is set to 15 ␮s to
For the reaction of OϩC2H5, using diethyl ketone as the
ethyl precursor, OH( ) with up to five quanta of vibrational
v
excitation is observed with a partially inverted vibrational
state distribution. Non-Boltzmann rotational excitation up to
Jϭ19.5 is also observed. Although both direct abstraction
and addition/elimination mechanisms are possible, the vibra-
tional state population inversion is more typical of a direct
abstraction. A comparison of the OH( ) vibrational state dis-
v
tribution for OϩC2H5 with that observed for many direct
abstraction reactions, e.g., HCl( ) formed in reactions of
v
translationally hot H atoms with Cl2O,29 show qualitative
similarities. In contrast to the OH( ) vibrational state distri-
v
bution for OϩC2H5 is the distribution of HCl( ) from
v
Cl(2P)ϩC2H5,28 which decreases monotonically with in-
creasing and is attributed to an addition/elimination reac-
v
tion. In the previous work of Slagle et al.,7 the 0.23 yield for
the OHϩC2H4 channel could not be theoretically justified on
the basis of an addition/elimination mechanisms. Our find-
ings from the product vibrational distribution are in accord
with this conclusion and indicate that the majority of hy-
droxyl radical products are formed via a direct abstraction
mechanism.
Using alkyl iodide precursors, rotational and vibrational
relaxation of the OH products and obscuring emissions over
J. Chem. Phys., Vol. 108, No. 5, 1 February 1998
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.192.114.19 On: Wed, 17 Dec 2014 21:15:53

Lindner et al.: Reaction of alkyl radicals with O(3P)
1946
reduce the influence of high frequency noise and thereby
increase the S/N ratio of the FT normalization channel. All
time-dependent FT signals are then normalized with respect
to the FT normalization signal and stored in the memory of
the PC. The interferograms are Fourier transformed by a
commercial software package using two times zero filling.
The spectra are divided by the instrument response function
of the FTIR, interference filters, and detector so that accurate
OH( ) rotational line intensities are obtained. The instru-
v
ment response function, including filter dependencies, is de-
rived from a separately recorded spectrum of a black body
radiator.
FIG. 1. Flow cell arrangement and Welsh cell optics for time-resolved FTIR
emission experiments. The reagents are flowed coaxially into the upper half
of the reaction cell, mixed with baffles and then photolyzed between the
halves of the cell using the 193 nm output of an ArF excimer laser. A Welsh
cell is used to collect the total emission within the reaction region and to
image it into the FTIR spectrometer.
Because the excimer laser is limited to a 150 Hz repeti-
tion rate, even at the slowest possible mirror speed of the
FTIR, the interferograms have to be taken in an interleaved
mode, i.e., several scans of the moving mirror in the inter-
ferometer are required to complete one set of time-resolved
interferograms. The details about the recording of the
interleaved interferograms have been described pre-
viously.11,12,28,31 Synchronization of the free running interfer-
ometer with the laser and the data acquisition is achieved by
a home built interface board inserted into the interferometer
and a commercial timer/counter board inside the data acqui-
sition computer. By electronically monitoring the voltage
signals corresponding to several parameters, the zero path
difference of the FTIR spectrometer internal white light
source, the reference laser HeNe fringes, the translational
scan speed of the moving arm of the spectrometer, as well as
the direction of the arm motion, the interface board generates
two TTL signals: one is high at times when the FTIR is
ready to record data and the other consists of pulses occur-
ring at times correlating to the zero crossings of the sinu-
soidal interference pattern of the internal HeNe reference la-
ser during the time the gate is high. These two signals are
applied via simple TTL logic to the A/D card and a counter/
timer board. The counter/timer board produces pulses which
are used to trigger the laser and the data acquisition after a
specified number of HeNe fringes at the beginning of each
mirror scan and then repetitively after another specified num-
ber of HeNe fringes determined by the repetition rate of the
laser and the scan speed of the spectrometer. The counter/
timer board also produces the trigger pulses that start the
A/D conversion of the FT signals that are applied to the
channels of the A/D card. An original data acquisition pro-
gram controls the timing of the experiment, the A/D conver-
sion with pulse-to-pulse normalization, and the coadding and
storing of the time-resolved data; it also permits the acquisi-
tion of continuous FTIR spectral data.
tions of the above modes allows the setup of complex timing
schemes.
In the experiments using diethyl ketone as the ethyl pre-
cursor, the time-resolved interferograms were recorded by
using six different boxcar integrators. The integration gates
had widths of 2 ␮s with the first one delayed 5 ␮s after the
laser photolysis pulse and the remaining five successively
delayed by an additional 6 ␮s. The gate widths are kept
small compared to the time delay between two HeNe fringes,
32 ␮s, to avoid averaging the interference pattern due to the
continuously moving mirror, while the delay between suc-
cessive gates is kept larger than the time response of the
detectors, ϳ3 ␮s. In the qualitative studies using the alkyl
iodides as precursors for the different alkyl radicals, the
modulated signals at the HeNe zero crossings ͑at 32, 64, 96,
128, and 160 ␮s͒ and those from two boxcars ͑at 7 and 19 ␮s
time delays͒ were recorded. A signal recorded prior to the
emission of the laser was also monitored in all of the experi-
ments and was used to obtain a background reference spec-
trum.
3
Alkyl؉O„ P… reaction flow system
The R•ϩO(3P) atom reactions are carried out in a pho-
tolysis flow cell24 which is housed within a larger stainless
steel vacuum chamber containing the fluorescence collection
optics that couple the reaction product emission into the
FTIR spectrometer as shown in Fig. 1. The flow cell is a
stainless steel cylinder 63.5 mm in diameter with a total
length of 110 mm; the cylinder is divided into two 50 mm
halves, separated by a 10 mm gap. Reactant gases flow into
the top of the photolysis flow cell through two concentric
stainless steel tubes. The tubes are of identical length and
end ϳ15 mm before the gap separating the two halves of the
cell. To insure efficient mixing of the reactants, two plates
are inserted into the upper half of the cell, which serve as
mixing baffles ͑see Fig. 1͒. The uppermost plate is located
7.5 mm from the ends of the reactant tubes and has two 4
mm wide arcs ͑160°͒ at a radial distance of 21 mm from the
center of the plate. The lower plate is positioned at the bot-
In contrast to the previous data acquisition scheme,11,12
the new setup allows several modes for recording time-
resolved interferograms: by directly recording the FT signal
voltage at the times of the HeNe fringe zero crossings ͑as
described in Refs. 12, 28, 31͒, or at equally spaced time
delays Ͼ7 ␮s, or using multiple boxcar integrators, with
gate widths ranging from 2 ns to 15 ␮s and independently
variable time delays, the system can integrate emission volt-
ages over different time windows. Furthermore, combina-
J. Chem. Phys., Vol. 108, No. 5, 1 February 1998
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.192.114.19 On: Wed, 17 Dec 2014 21:15:53

Lindner et al.: Reaction of alkyl radicals with O(3P)
1947
tom of the upper half of the reaction cell and has a 10 mm
diameter hole in the center of the plate. The flow of the
reactant gases is therefore very turbulent, with a pathway that
proceeds from the reactant tubes outwards to the arc and then
through the center again. Both the chamber and the flow cell
are evacuated by a 300 l /s roots blower pump, backed by an
80 l /s mechanical pump, through a 25.4 mm diameter stain-
less steel tube exiting the bottom of the lower half of the
flow cell. The outside of the small flow cell is pressurized
with Ar. In past experiments performed in this laboratory,
the argon was used to purge both the mirrors and windows,32
however, here this additional background gas serves to con-
fine the products and the reactants within the flow cell and
observation region; this confinement gives an enhancement
of alkyl radical–O collisions and a higher OH( ) product
v
fluorescence intensity, but compromises nascent rotational
state information and perhaps vibrational state distributions,
which will be considered below.
The gap in the flow cell allows transmission of the pho-
tolysis laser and detection of the IR fluorescence. The pho-
tolysis beam, that of a high repetition rate ArF excimer laser
͑150 mJ/pulse maximum energy @ 150 Hz with a pulse du-
ration of ϳ20 ns͒, is collimated to an area of ϳ5 mmϫ5 mm
and is passed twice through the photolysis chamber. The
energy density within the flow cell is measured to be
ϳ140 mJ/cm2 for these experiments. As indicated in Fig. 1,
the infrared light emitted from vibrationally excited species
within the reaction vessel is collected via a Welsh cell33
composed of silver coated mirrors (Rϭ98%). The IR fluo-
rescence is then focused by a telescope ͑f1ϭϩ100 mm, f2
ϭϩ150 mm, CaF2 lenses͒ onto the entrance of the FTIR.
The spectrometer is operated at a resolution of 0.2 cmϪ1 and
the mirror speed is selected to be 0.5 cm/s with sampling of
the modulated fluorescence twice per zero fringe of the
HeNe laser wavelength. At this resolution and speed, each
interferogram consists of ϳ120 000 points with increments
correlating to mirror positions spaced by 158 nm.
FIG. 2. The time evolution of the total infrared fluorescence from the laser
initiated reaction of a diethyl ketone/SO2 mixture. A band pass filter is used
to limit the emission to that of excited vibrational levels of the hydroxyl
radical. The positions and widths of the boxcar integration gates used to
record the six different time-resolved OH spectra are superimposed.
diethyl ketone and 40 Pa ͑300 mTorr͒ for SO2. The overall
pressure in the vacuum chamber due to the addition of Ar
was 17 Pa ͑130 mTorr͒.
As expected, CO and C2H5 emission from the photodis-
sociation of diethyl ketone27 as mentioned above, was ob-
served in the low wave number region of the spectra. The
strong CO emission was discriminated with respect to the
OH emission through the use of a bandpass filter which has
ϳ70% transmission from 2800 to 4200 cmϪ1. The observed
time-resolved transient, as shown in Fig. 2, has 1/e rise and
decay times of ϳ10 and ϳ90 ␮s. The time evolution of the
product spectra was obtained with six boxcar gated integra-
tors, as mentioned above. The signals of the boxcar gates at
various time delays relative to the emission intensity profile
indicate that a full distribution of times about the most in-
tense region of the emission is being sampled. It was not
possible to set a gate at shorter delays since the emission
signal becomes too weak to obtain spectra that can be quan-
titatively analyzed, while spectra recorded at longer times do
not offer additional information about the OH(v) nascent
distributions.
In the C2H5ϩO experiments, optimal conditions for
probing the reaction mechanism were determined by varying
the concentrations of the precursors and the pressure of the
argon background gas while monitoring both the temporal
profile of the total emission and the OH( ) product spectra.
v
In general, the OH product signals increased with increasing
concentrations of the precursors within the reaction vessel.
However, clear indications of rotational and vibrational de-
activation of the OH( ) products, induced by collisions with
v
the remaining precursor molecules, buffer gas, photofrag-
ments, and/or other reaction products, was observed when
the densities became too high, especially that of the diethyl
ketone; the 1/e rise and fall times of the emission profile
decreased with increasing densities and an obvious time de-
pendence in the relative intensities of the rovibrational lines
was observed in the spectra at these higher pressures. The
maximum inlet-side pressures, for which the collisional de-
activation seemed minimal, were 112 Pa ͑0.85 Torr͒ for
͑C2H5͒2CO and 263 Pa ͑2.0 Torr͒ for SO2. The partial pres-
sures within the reaction region of the flow cell were esti-
mated with a thermocouple gauge to be 12 Pa ͑90 mTorr͒ for
The time-resolved emission spectra of the vibrationally
excited products formed in the reactions of the alkyl radicals,
C2H5, n-C3H7 and i-C3H7, with O(3P) were recorded by the
simultaneous photolysis of SO2 and one of the alkyl iodides,
C2H5I, n-C3H7I, and i-C3H7I. Because the total emission
signal is comprised of at least OH(v) and HOI(v), as dis-
cussed above, optimizing the partial pressures and concen-
trations of the reactants was very difficult. By repeatedly
performing low resolution scans, ϳ1 cmϪ1, the maximum
OH(v) line intensities were obtained with pressures within
the reaction vessel of approximately, 13 Pa ͑100 mTorr͒ for
SO2, 6.4 Pa ͑50 mTorr͒ for the alkyl iodide and 133 Pa ͑1
J. Chem. Phys., Vol. 108, No. 5, 1 February 1998
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.192.114.19 On: Wed, 17 Dec 2014 21:15:53

Lindner et al.: Reaction of alkyl radicals with O(3P)
1948
Torr͒ for the Ar buffer gas. The very high buffer gas pressure
was critical for observing the OH spectra with sufficient S/N
for resolving the narrow rotational lines in the presence of
the HOI, ⌬ ϭϪ1, emission bands. For all of the reactions
v
performed at these pressures, the total time-resolved transient
has 1/e rise and decay times on the order of ϳ10 and
ϳ120 ␮s, respectively, although these values are expected
to be controlled by the OϩRI→HOI( )ϩR• reaction pro-
v
cess and the HOI( ) emission. Both the rise and decay times
v
were very dependent on the pressures of the gases, especially
that of the buffer gas and alkyl iodides; with increasing pres-
sures of either of these two gases, both the rises and decays
become shorter, indicating different reaction and relaxation
dynamics. The integrated intensity of the HOI( ) product
v
bands was considerably larger than that of the OH( ) lines,
v
hence the reason that changes in the temporal fluorescence
profile are most likely indicative of changes in the
RIϩO(3P) chemistry and the HOI deactivation.
RESULTS AND ANALYSIS
Time-resolved spectra of the products formed in the re-
action of the C2H5 radicals with O(3P) atoms were recorded
by simultaneously photolyzing ͑C2H5͒2CO and SO2. These
experiments provide direct information about the reaction
mechanism for OH production. Regions of typical time-
dependent spectra recorded at 11, 23, and 35 ␮s after the
laser pulse are shown in Fig. 3 from bottom to top, respec-
tively. A total of 50 coadditions, mostly in blocks of four
interferograms were recorded for these spectra. The narrow
emission lines are assigned to rovibrational transitions within
FIG. 3. The time-resolved FTIR emission spectra shown here are obtained
11, 23, and 35 ␮s after the laser photolysis of a diethyl ketone/SO2 mixture,
bottom to top panels, respectively. The OH emission is attributed to the
products of the ethyl radical ϩO(3P) reaction. The tic marks above the t
ϭ35 ␮s spectrum indicate the energies of the P rotational lines used in the
population analysis; the long and short tics indicate transition lines within
the 2⌸3/2 and 2⌸1/2 spin–orbit states, respectively. The rotational quanta, J,
2
and vibrational quanta, , of the excited states levels are indicated above the
v
the X ⌸ state of the OH radical and can only be observed in
tics. The total OH emission increases with time, and only a slight shift of
relative intensity of the lines towards the higher wave number region of the
spectrum is recognizable and is mostly attributed to rotational relaxation of
the OH(v) products. The broad background is attributed to C–H stretches of
the ethyl radical.
the presence of both precursors, diethyl ketone and SO2. The
time-dependence of the integrated intensity of the OH lines
is nearly identical to the intensity of the total emission tem-
poral signal, indicating that nearly all of the observed emis-
sion is from OH and that OH vibrational relaxation is mini-
mal. A slight change in the relative intensities of the OH
rotational lines within a given vibrational band is observed
with increasing time, which indicates that some rotational
relaxation is occurring in these experiments.
The OH rotational lines are identified from the line list
of Maillard et al.34 or determined from the ground state en-
ergies given by Coxon35 and are labeled above the 35 ␮s
spectrum in Fig. 3. The ⌳-doubling can only be partly re-
solved at the 0.2 cmϪ1 resolution of the spectra. The rovibra-
where A( ,J,F ) denotes the Einstein A coefficients and i
v
i
ϭ1 or 2 indicates the corresponding spin–orbit state of OH.
The Einstein A coefficients for OH are taken from Nelson
et al.36,37 Those values not given in either of these references
are extrapolated from the available Einstein coefficients.
Only P lines are used for the analysis of the population since
the Einstein coefficients of the R lines are, in general, sig-
nificantly smaller than for the P lines,36,37 resulting in poor
S/N ratio of these intensities. The error in population,
tional line intensities, I( ,J,F ), for the two transitions of
v
i
⌬n( ,J,F ), is calculated according to
v
the multiplet components, F1 and F2 for 2⌸3/2 and 2⌸1/2 , of
i
the OH ground state with total angular momentum and vi-
⌬n ,J,F ͒ϭI
͑v
/A ,J,F ͒,
͑v
͑5͒
i
noise
i
brational quantum numbers, J and , respectively, are taken
v
where Inoise denotes the noise at the baseline of the spectrum.
Population in OH (X ⌸) rotational levels with 1 to 5
quanta of vibrational excitation are observed. Some popula-
from the peak intensities of each line with the average of the
local baseline subtracted. A comparison of the peak intensi-
ties of the individual OH rotational lines with the integrated
areas of each of the lines revealed the same relative values,
within the error limits.
2
tion may also exist in the ground ϭ0 state of OH, but in
v
these experiments only emission from vibrationally excited
species is detected. The rotational product state distributions
within each vibrational level are broad non-Boltzmann dis-
tributions with populations in J ranging from as low as 1.5,
The rovibrational distributions, n( ,J,F ), are deter-
v
i
mined by
n
,J,F ͒ϭI ,J,F ͒/A ,J,F ͒,
͑4͒
in the ϭ3 state, up to as high as 19.5, in the ϭ1 state. For
v v
͑v
͑v ͑v
i
i
i
J. Chem. Phys., Vol. 108, No. 5, 1 February 1998
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.192.114.19 On: Wed, 17 Dec 2014 21:15:53

Lindner et al.: Reaction of alkyl radicals with O(3P)
1949
population in time observed, while ϭ2 shows a small in-
v
crease and ϭ4 has a small decrease in population with
v
increasing time ͑both of which are still less than the popula-
tion error bars͒. However, the time-dependent distributions
for a given vibrational state agree almost always within the
error bars and no obvious vibrational relaxation can be ob-
served in the populations. The vibrational OH( ) popula-
v
tions at the earliest time delay are 0.18Ϯ0.03, 0.23Ϯ0.04,
0.29Ϯ0.05, 0.23Ϯ0.07, and 0.07Ϯ0.04 for ϭ1 to 5, re-
v
spectively, and are taken to be the nascent values.
As mentioned above, the broad feature spanning ϳ2700
to ϳ3100 cmϪ1 is attributed to C–H stretching modes of
ethyl and is observed only when diethyl ketone is present in
the flow cell. The vibrational state distribution of the C2H5
products following photolysis of ͑C2H5͒2CO has not been
measured experimentally, however, Hall et al.27 have esti-
mated that the average C–H vibrational energy is
ϳ9.6 kJ/mol (ϳ805 cmϪ1). The shape of the broad feature
in our experiments changes significantly within 30 ␮s, in
accord with the observations reported in Ref. 27; the chang-
ing structure corresponds to vibrational relaxation and there-
fore to lowering of the vibrational temperature of the ethyl
radical. The time dependence of the OH vibrational state
populations do not exhibit significant time dependence out to
FIG. 4. The time dependence of the relative populations, n( ), found in the
v
vibrational levels of the OH product following the laser initiated reaction of
ethyl radicals with O(3P) atoms. At each time, the total population is nor-
malized to 1. Although, the ϭ2 and ϭ4 states increase and decrease in
v
v
time, respectively, no clear vibrational relaxation is observed outside the
limits of the error bars. The tϭ5 ␮s populations are taken to be the nascent
values.
at least 35 ␮s. Although contributions to the OH( ) popula-
v
tions from reactions involving vibrationally excited ethyl
radicals cannot be ruled out, the lack of a temporal correla-
the Jϭ19.5, ϭ1 state, a significant amount of energy,
tion between the OH( ) distribution and the temperature of
the vibrationally excited ethyl radicals suggests the contribu-
tions are only minor.
v
v
Ͼ78 kJ/mol (Ͼ6500 cmϪ1), is found in rotation, while for
2
the observed Jϭ9.5, ϭ5, ⌸ state, 79% of the available
v
1/2
energy, 266 kJ/mol (ϳ22 200 cmϪ1), is found in internal
Experiments using n- and i-propyl iodides as precursors
for the two-carbon length propyl radicals have also been per-
formed to gain some information on the reaction mechanism
for larger alkyl radicals with O(3P) atoms. A typical emis-
sion spectrum obtained from the n-C3H7ϩO reaction re-
corded with a gated integrator, 3 ␮s width and delayed by 19
␮s after the laser pulse, is shown in Fig. 5. The sharp emis-
sion lines between 3100 and 3600 cmϪ1 are assigned to rovi-
energy of the OH product. Although, rotational levels with
six quanta of vibrational excitation are energetically acces-
sible, no emission from these states was observed above the
background black body radiation.
The error bars for each rotational level population range
from 10% up to 50% with a general trend of increasing error
with decreasing transition wave number. This tendency is
attributable to the presence of black body light which con-
tributes to the emission signal at lower frequencies. Even
though the black body is simultaneously recorded with the
spectra and is later subtracted from the time-resolved product
spectra, the resultant noise, because of subtraction of a larger
background in this spectral region, becomes worse. No clear
rotational state distribution nor time-dependent trends in the
rotational distributions within each vibrational state can be
characterized, due in part to the error associated with each
rotational level population. However, since rotational states
are observed up to Jϭ19.5, it is likely that either direct re-
action dynamics or vibration-to-rotation relaxation processes
play a role in the rotational distributions.38,39
2
brational transitions within the X ⌸ state of OH. The spec-
trum also shows three broad features centered about 3625,
3455, and 3290 cmϪ1. These features are the excited
,
v1
O–H, stretches of the HOI molecule.25 Vibrationally excited
HOI is a product of the reaction of O atoms with alkyl io-
dides when the alkyl chain is two carbons or longer.24 Very
similar spectra are obtained when studying the two reactions
using the other alkyl iodide precursors as well.
Clearly, there are at least two reactions occurring within
the reaction vessel: reaction between alkyl iodides and oxy-
gen atoms producing HOI( )24,25 and reactions between
v
alkyl radicals and oxygen atoms yielding OH( ). As men-
v
tioned above, in order to observe the OH lines with sufficient
intensity for analysis, the partial pressures within the flow
cell had to be changed significantly compared to the experi-
ments using diethyl ketone as a precursor for ethyl. The in-
creased densities of the reactants, due mostly to the confine-
ment from the Ar buffer gas, significantly increases the
degree of rotational and vibrational relaxation of the product
species, both HOI and OH. Furthermore, although the con-
The final OH( ) vibrational product state distributions
v
are obtained by summing over all the observed rotational
populations of both spin–orbit states, F1 and F2 , within a
given vibrational quantum, . These distributions are shown
v
in Fig. 4 for the six recorded time delays. At all times, the
ϭ3 state is the most populated while the ϭ5 state is the
v
v
least. For ϭ1, 3, and 5, there is a negligible variation of the
v
J. Chem. Phys., Vol. 108, No. 5, 1 February 1998
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.192.114.19 On: Wed, 17 Dec 2014 21:15:53

Lindner et al.: Reaction of alkyl radicals with O(3P)
1950
obtain insight into the reaction mechanism involving the
three-carbon length alkyl radicals.
DISCUSSION
In the following, we first discuss the vibrational distri-
butions of the reaction of ethyl radicals with O(3P) deter-
mined in this work in the context of the kinetics measure-
ments by Slagle et al.7 The OH( ) distributions determined
v
here are compared with the HCl( ) results determined in
v
experiments probing the mechanism of the HϩCl2O
reaction29 and in similar investigations of Seakins et al.28 on
the reaction of ethyl radicals with Cl(2P) atoms. The goal is
to consider the role of the two possible reaction mechanisms:
direct abstraction and addition/elimination.
In the work of Slagle et al.,7 the rate constant for the
reaction of C2H5ϩO(3P) was determined to be 2.2Ϯ0.4
ϫ10Ϫ10 cm3 moleculeϪ1 sϪ1, independent of temperature
and density. The chemical branching of the reaction was also
characterized both experimentally and theoretically. The
FIG. 5. Emission spectrum observed from the reaction of n-C3H7 and
n-C3H7I with O(3P) 19 ␮s after laser photolysis. The three, partially re-
solved broad features are known to be excitation of the 1 , O–H, stretch of
v
HOI. The sharp lines between 3100 and 3600 cmϪ1 are assignable rovibra-
channel for OH( )ϩC H was measured to have a yield of
v
tional transitions of OH within the X 2⌸ electronic state.
2
4
0.23Ϯ0.07. They also calculated the potential energy barri-
ers for several reaction paths beginning with the formation of
the ethoxyl radical, C2H4OH, which is the addition process.
The direct H-atom metathesis to produce OH and C2H4 was
not considered. The calculated branching fraction for the OH
channel, 0.02, was much lower than observed. Reasonable
changes in the relative barrier heights along the reaction co-
ordinate could not account for the discrepancy between
theory and experiment. This disparity was interpreted as the
formation of OHϩC2H4 via a direct H atom abstraction re-
action.
centration of oxygen atoms is greater than the alkyl radical
concentration, the alkyl iodide concentration is significantly
higher than the oxygen concentration; thus the kinetics of the
OϩRI reaction dominate over those of the OϩR reaction
being studied here.
For each of the product emission spectra recorded using
the different alkyl iodide precursors, only rotational lines
originating from the OH ϭ1–3 levels are observed. In
v
addition, the OH ϭ1 populations are the highest, followed
The OH( ) vibrational state distribution observed here
v
v
by ϭ2 and then the barely populated ϭ3. No apparent
shows a clear population inversion with the ϭ3 state being
v
v
v
time-dependence is observed in the OH rovibrational line
intensities and corresponding vibrational populations out to
at least 39 ␮s. Since the results with ethyl iodide are very
different from the results with diethyl ketone, we believe the
distributions in all cases for the alkyl iodide precursors are
accompanied by severe relaxation. In previous studies prob-
ing HOI formation,24 the partial pressures within the reaction
region were ϳ20 Pa ͑150 mTorr͒ for SO2 and ϳ9 Pa ͑67.5
mTorr͒ for the alkyl iodides, but no Ar buffer gas was used
for confinement. A clear time-dependence was observed for
the most populated, as shown in Fig. 6. If there is a small
amount of vibrational relaxation, the populations may be
slightly more inverted than shown, and ϭ6 may have some
v
small population as well. As pointed out by Holmes and
Setser,40 Agrawalla and Setser,41 and Sloan,42 an inverted
vibrational distribution is typical of most direct abstraction
reactions. Therefore, reaction ͑1a͒ is likely to proceed
through a direct abstraction, in accord with the analysis of
Slagle et al.7
A surprisal analysis of the OH vibrational energy dis-
posal was performed assuming the vibrational degrees of
freedom of the polyatomic fragment could be neglected. This
treatment, labeled model II by Bogan and Setser,43 usually
agrees with a complete treatment of all degrees of freedom
the HOI( ) vibrational state distribution which was attrib-
v
uted to vibrational relaxation of the reaction products. No
OH rotational lines were observed under the reactant pres-
sure conditions for the HOI studies. In the OH studies re-
ported here, vibrational relaxation is not observed for OH( )
within 3%. The surprisal plot, when including only the
v
v
and it is presumed that the products have been relaxed to a
thermal distribution within the time of the first measure-
ments, ϳ7 ␮s; this is the shortest time for which a signal-
to-noise level suitable for analysis could be achieved. Due to
ϭ1 to 4 populations, is well represented by a straight line,
given by I(f )ϵϪln P( )/P0( ) ϭ␭ ϩ␭ f as described in
͓
v
v ͔
v
0
v v
detail in Ref. 44. In this analysis, fv is the fraction of the
total energy in the OH vibration, P( ) is the observed popu-
v
the enhanced relaxation of the OH( ) product, nascent dis-
lation distribution and P0( ) is the prior or statistical distri-
v
v
tributions cannot be obtained for the n- and i-propyl plus
bution which is obtained by using the equations in Ref. 43
oxygen reactions; only qualitative deductions can be made
for model II. The value of ␭ determined in the fit is Ϫ5.7
v
through comparisons of the OH( ) product spectra observed
Ϯ0.2; the negative sign indicates that the vibrational distri-
v
using n- and i-propyl iodides with the ethyl iodide results to
bution is inverted with the maximum population observed at
J. Chem. Phys., Vol. 108, No. 5, 1 February 1998
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.192.114.19 On: Wed, 17 Dec 2014 21:15:53

Lindner et al.: Reaction of alkyl radicals with O(3P)
1951
produces an inverted HCl( ) distribution that peaks at
v
v
ϭ3.29 The HCl( ) product distribution observed in a similar
v
investigation of ethyl radicals reacting with chlorine atoms,
C H ϩCl→HCl ͒ϩC H ,
͑v
͑7͒
2
5
2
4
is also shown in Fig. 6. The HCl( ) distribution from reac-
v
tion ͑7͒ shows a monotonic decrease in population with in-
creasing vibrational quantum number, which is more indica-
tive of
a long-lived intermediate state followed by
dissociation. The OH( ) distribution reported here is quali-
v
tatively similar to the HCl( ) distribution for the direct ab-
v
straction reaction ͑6͒ while it is significantly different from
the distribution for the addition/elimination reaction ͑7͒.
Prediction of the OϩC2H5 reaction mechanism based on
the energetics of the potential energy surfaces of similar re-
action systems is difficult. The reactions of saturated hydro-
carbons with atomic oxygen have various activation energies
that are in general higher than those involving hydrogen
atom abstraction by chlorine atoms.7,41 Less is known about
the possible reaction pathways of O and Cl with radicals.
The direct abstraction mechanism for the OH-producing re-
action, OϩC2H5, and the addition-elimination pathway for
the ClϩC2H5 reaction cannot readily be confirmed. Although
it is difficult to perform theoretical calculations that accu-
rately interrogate the differences in small activation energies,
as expected for these reactions, such calculations could prove
invaluable to the interpretation of the experimental observa-
tions.
FIG. 6. Comparison of the OH( ) vibrational distribution from
v
C H ϩO(3P) with the HCl( ) vibrational distributions from the Cl OϩH
v
2
5
2
and C2H5ϩCl(2P) reactions ͑values taken from Refs. 29 and 28, respec-
tively͒. The OH( ) distribution from C H ϩO and the HCl( ) distribution
v
v
2
5
from the Cl2OϩH reaction are inverted, suggesting direct abstraction reac-
tion mechanisms. The HCl( ) distribution from the C H ϩCl reaction is
v
2
5
indicative of an addition/elimination mechanism with monotonically de-
creasing population with
.
v
ϭ3. When including the ϭ5 population, the value of
v
v
P(5) is too low and does not fit the linearity of the plot.
In the experiments using the alkyl iodides as alkyl radi-
cal precursors, the data suggest that the most important
Note, however, that the error in the population of the ϭ5
v
state is significant (0.07Ϯ0.04) and does span the linear fit
to the surprisal plot.
source of OH( ) is from the reactions of ethyl, n-propyl and
v
i-propyl radicals with O(3P) atoms. However, the OH distri-
butions formed in the C2H5ϩO reactions are observed to be
significantly different when using ethyl iodide as the precur-
sor compared to diethyl ketone because of vibrational relax-
ation. The emission intensities of the OH rotational lines and
The y intercept of the surprisal analysis, ␭0ϭ1.46, can
be used to estimate the ϭ0 population and the average
v
fraction of the total available energy that is released as vi-
brational energy, f (OH) . The renormalized OH( ) distri-
v
͗
͘
v
bution, calculated using the ϭ0 population obtained from
v
OH( ) distributions observed in the reactions of the n- and
v
␭0 , is 0.10Ϯ0.04, 0.16Ϯ0.03, 0.21Ϯ0.04, 0.26Ϯ0.04, 0.21
i-propyl radicals with oxygen are nearly the same as those
observed from the ethyl radical reaction using ethyl iodide.
This may suggest that the reaction mechanism for the reac-
tion of the three-carbon length chain alkyl radicals, n- and
i-propyl, with oxygen atoms is similar to that for the two-
carbon long ethyl reaction.
Ϯ0.06, 0.06Ϯ0.04, for ϭ0 to 5, respectively. Reactions
v
proceeding via unimolecular elimination mechanisms typi-
cally yield HX( ) distributions, where XϭF, Cl, Br, O, that
v
monotonically decline with increasing fv and values of
f (HX) that lie within the 0.15 to 0.25 range.29,45 The
͗
͘
v
HX( ) distributions resulting from direct abstraction over
v
repulsive potentials are sharply inverted and are rather nar-
row with f (HX) Ϸ0.4.45 Using the vibrational state distri-
͗
͘
CONCLUSIONS
v
bution including the ϭ0 state population determined in the
v
In this work, the reactions of alkyl radicals with atomic
oxygen in the P state have been investigated using an im-
surprisal analysis, the O(3P)ϩC2H5 reaction is found to have
3
f (OH) ϭ0.38Ϯ0.03. The high fraction of available en-
͗
͘
v
proved experimental setup for TR-FTIR spectroscopy. Vi-
ergy that is transferred into the OH product strongly suggests
that the dominant mechanism in the reaction of oxygen at-
oms with ethyl radicals is direct abstraction.
2
brationally excited hydroxyl radicals within the ground X ⌸
electronic state are observed as a product of this reaction
when using ethyl as well as n-propyl, and i-propyl radicals
as reagents. For the C2H5ϩO reaction, populations in OH
Comparison of the OH( ) distribution with those of
v
HCl( ) formed in both a direct abstraction reaction and an
v
vibrational levels from ϭ1 to 5 are determined with broad
v
addition/elimination reaction supports this thesis. As shown
in Fig. 6, the reaction of Cl2O molecules with translationally
hot H atoms,
rotational distributions, from Jϭ1.5 up to as high as 19.5 in
ϭ1, observed in each vibrational state. An inverted OH( )
v
v
vibrational distribution has been observed with the ϭ3
v
Cl OϩH→HCl ͒ϩClO,
͑v
2
͑6͒
state being the most populated. Together with the results
J. Chem. Phys., Vol. 108, No. 5, 1 February 1998
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.192.114.19 On: Wed, 17 Dec 2014 21:15:53

Lindner et al.: Reaction of alkyl radicals with O(3P)
1952
from the kinetic studies of Slagle et al.,7 it is likely that the
14 J. J. Sloan, in Advances in Spectroscopy ͑Wiley, New York, 1989͒, Vol.
18.
reaction mechanism for C H ϩO→OH( )ϩC H proceeds
v
2
5
2
4
15 I. R. Slagle, I. J. Kalinovski, D. Gutman, and L. B. Harding ͑private
communication͒.
through a direct abstraction mechanism. The similar HCl( )
v
16 G. Black, R. L. Sharpless, and T. G. Slanger, Chem. Phys. Lett. 90, 55
͑1982͒.
vibrational product state distribution observed in the direct
abstraction reaction, Cl OϩH→HCl( )ϩCl, and the very
v
2
17 C. J. Cobos, H. Hippler, and J. Troe, J. Phys. Chem. 89, 1778 ͑1985͒.
18 A. Freedman, S.-C. Yang, and R. Bersohn, J. Chem. Phys. 70, 5313
͑1979͒.
different HCl( ) vibrational product state distribution ob-
v
served in the addition/elimination reaction, C2H5ϩCl
→HCl( )ϩC H , is further support for this interpretation.28
19 M. Kawasaki, K. Kasatani, and H. Sato, Chem. Phys. 73, 377 ͑1982͒.
20 M. Kawasaki and H. Sato, Chem. Phys. Lett. 139, 585 ͑1987͒.
21 P. Felder, C. S. Effenhauser, B. M. Haas, and J. R. Huber, Chem. Phys.
Lett. 148, 417 ͑1988͒.
v
2
4
Similar experiments using alkyl iodides as precursors for
n-propyl and i-propyl radicals verify that OH is produced
when oxygen atoms react with longer alkyl chains as well.
However, complications due to reactions of the precursors
prohibit a quantitative analysis similar to that of the diethyl
ketone experiments.
22 Y.-L. Huang and R. J. Gordon, J. Chem. Phys. 93, 868 ͑1990͒.
23 M. K. Gilles, A. A. Turnipseed, R. K. Talukdar, Y. Rudich, P. W. Villalta,
L. G. Huey, J. B. Burkholder, and A. R. Ravishankara, J. Phys. Chem.
100, 14005 ͑1996͒.
24 R. A. Loomis, J. J. Klaassen, J. Lindner, P. G. Christopher, and S. R.
Leone, J. Chem. Phys. 106, 3934 ͑1997͒.
25 J. J. Klaassen, J. Lindner, and S. R. Leone, J. Chem. Phys. 104, 7403
͑1996͒.
ACKNOWLEDGMENTS
The authors would like to thank the Department of En-
ergy for support of this research and the National Science
Foundation for additional equipment. R.A.L. and J.J.K.
gratefully acknowledge the support of the National Research
Council for Postdoctoral Fellowships with the National Insti-
tute of Standards and Technology. The authors would also
like to thank J.B. Jeffries and his colleagues for useful dis-
cussions concerning OH vibrational relaxation.
26 J. G. Calvert and J. N. Pitts, Photochemistry ͑Wiley, New York, 1966͒.
27 G. E. Hall, H. W. Metzler, J. T. Muckerman, J. M. Preses, and R. E.
Weston, Jr., J. Chem. Phys. 102, 6660 ͑1995͒.
28 P. W. Seakins, E. L. Woodbridge, and S. R. Leone, J. Phys. Chem. 97,
5633 ͑1993͒.
29 S. J. Wategaonkar and D. W. Setser, J. Chem. Phys. 90, 251 ͑1989͒.
30 J. Lindner, J. K. Lundberg, R. M. Williams, and S. R. Leone, Rev. Sci.
Instrum. 66, 2812 ͑1995͒.
31 E. L. Woodbridge, M. N. R. Ashfold, and S. R. Leone, J. Chem. Phys. 94,
4195 ͑1991͒.
32 J. Lindner, J. K. Lundberg, C. M. Lovejoy, and S. R. Leone, J. Chem.
Phys. 106, 2265 ͑1997͒.
1 J. A. Miller, R. J. Kee, and C. K. Westbrook, Annu. Rev. Phys. Chem. 41,
345 ͑1990͒.
33 L. Welsh, C. Cumming, and E. J. Stansbury, J. Opt. Soc. Am. 41, 712
͑1951͒.
2 J. Warnatz, in Combustion Chemistry, edited by J. C. Gardiner ͑Springer,
New York, 1993͒.
34 J. P. Maillard, J. Chauville, and A. W. Mantz, J. Mol. Spectrosc. 63, 120
͑1976͒.
3 Comprehensive Chemical Kinetics, edited by C. H. Bamford and C. F.
Tipper ͑Elsevier, Amsterdam, 1970͒, Vol. 2.
35 J. A. Coxon, Can. J. Phys. 58, 933 ͑1980͒.
4 K. Hoyermann and R. Sievert, in Symposium (International) on Combus-
tion ͑The Combustion Institute, Pittsburgh, 1979͒ Vol. 17, p. 517.
5 I. R. Slagle, D. Sarzynski, and D. Gutman, J. Phys. Chem. 91, 4375
͑1987͒.
36 D. D. Nelson, Jr., A. Schiffman, and D. J. Nesbitt, J. Chem. Phys. 90,
5455 ͑1989͒.
37 D. D. Nelson, Jr., A. Schiffman, D. J. Nesbitt, J. J. Orlando, and J. B.
Burkholder, J. Chem. Phys. 93, 7003 ͑1990͒.
6 P. W. Seakins and S. R. Leone, J. Phys. Chem. 96, 4478 ͑1992͒.
7 I. R. Slagle, D. Sarzynski, D. Gutmann, J. A. Miller, and C. F. Melius, J.
Chem. Soc. Faraday Trans. 84, 491 ͑1988͒.
38 H. K. Haugen, W. H. Pence, and S. R. Leone, J. Chem. Phys. 80, 1839
͑1984͒.
39 C. A. Taatjes and S. R. Leone, J. Chem. Phys. 89, 302 ͑1988͒.
40 B. E. Holmes and D. W. Setser, in Physical Chemistry of Fast Reactions,
edited by I. W. M. Smith ͑Plenum, New York, 1980͒, Vol. 2, p. 83.
41 B. Agrawalla and D. W. Setser, in Gas-Phase Chemiluminescence and
Chemi-ionization, edited by A. Fontijn ͑North Holland, Amsterdam,
1985͒, p. 157.
8 L. V. Gurvich, I. V. Veyts, and C. B. Alcock, Thermodynamic Properties
of Individual Substances, 4th ed. ͑Hemisphere, New York, 1989͒.
9 C. F. Melius ͑private communication͒.
10 W. G. Mallard, F. Westley, J. T. Herron, and R. F. Hampson, NIST
Chemical Kinetics Database: Version 6.0 ͑NIST, Gaithersburg, MD,
1994͒.
42 J. J. Sloan, J. Phys. Chem. 92, 18 ͑1988͒.
11 S. R. Leone, Acc. Chem. Res. 22, 139 ͑1989͒.
43 D. J. Bogan and D. W. Setser, J. Chem. Phys. 64, 586 ͑1976͒.
44 R. D. Levine and R. B. Bernstein, Molecular Reaction Dynamics and
Chemical Reactivity ͑Oxford University Press, New York, 1987͒.
45 E. Arunan, R. Rengaragan, and D. W. Setser, Can. J. Chem. 72, 568
͑1994͒.
12 S. A. Rogers and S. R. Leone, Appl. Spectrosc. 47, 1430 ͑1993͒.
13 G. Hancock and D. E. Heard, in Advances in Photochemistry, edited by D.
H. Volman, G. S. Hammond, and D. C. Neckers ͑Wiley, New York, 1993͒
Vol. 18, p. 1.
J. Chem. Phys., Vol. 108, No. 5, 1 February 1998
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.192.114.19 On: Wed, 17 Dec 2014 21:15:53