2A'-2A' detection of vibrationally excited HCO produced by the O(3P) + C2H4 reaction

Article abstract of DOI:10.1063/1.1774984

The distribution of rotational and vibrational energy in HCO produced by the O(³P) + C2H4 reaction has been measured using laser-induced fluorescence detection via the <*>²A'-<*>²A' transition. Over a detection wavelength range of 248-290 nm, our experiments have shown that HCO is formed in both the ground state and in at least six vibrationally excited states with up to two quanta of energy in the C-O stretch and the bending mode. Dispersed fluorescence experiments were conducted to positively assign all of the HCO vibrational bands. The experiments confirmed that many bands, including the <*>(000)-<*>(000) band, are affected by overlap with other HCO bands. Spectral modeling was used to separate the contributions of overlapping HCO <*>-<*> bands and to determine a nascent HCO rotational temperature of ca. 600 K, corresponding to ca. 6 percent of the total energy from the O(³P) + C2H4 reaction. HCO vibrational distributions were determined for two different average collision energies and were fit with vibrational temperatures of 1850+/-80 K and 200+/-100 K, corresponding to ca. 15 percent of the total energy. The observed Boltzmann distribution of vibrational energy in HCO indicates that HCO and CH3 are formed by the dissociation of an energized intermediate complex.

Full text of DOI:10.1063/1.1774984

B̃
2 A ′ -X̃ 2 A ′ detection of vibrationally excited HCO produced by the O ( 3 P)+ C 2 H 4
reaction
Jennifer L. Gardner and Steven M. Miller
Citation: The Journal of Chemical Physics 121, 5920 (2004); doi: 10.1063/1.1774984
View online: http://dx.doi.org/10.1063/1.1774984
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/121/12?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Experimental investigation of the Jahn-Teller effect in the ground and excited electronic states of the tropyl
radical. Part II. Vibrational analysis of the A
̃ E 3 ″ 2 - X ̃ E 2 ″ 2 electronic transition
J. Chem. Phys. 128, 084311 (2008); 10.1063/1.2829471
Fully state-resolved photodissociation of formaldehyde, H 2 CO → H + HCO : K conservation and a rigorous test
of statistical theories
J. Chem. Phys. 122, 194312 (2005); 10.1063/1.1902863
A stimulated emission pumping study of the first excited singlet state of germylidene ( H 2 C = Ge )
J. Chem. Phys. 121, 257 (2004); 10.1063/1.1758699
Discovery of the optically forbidden S 1 –S 0 transition of silylidene ( H 2 C = Si )
J. Chem. Phys. 118, 1642 (2003); 10.1063/1.1531618
Fluorescence excitation spectroscopy of the Ar–HCO (X̃ 2 A′,B̃ 2 A′) van der Waals complex
J. Chem. Phys. 107, 680 (1997); 10.1063/1.474469
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: 193.0.65.67
On: Wed, 10 Dec 2014 14:34:42

JOURNAL OF CHEMICAL PHYSICS
VOLUME 121, NUMBER 12
22 SEPTEMBER 2004
2
2
˜
˜
B A -X A detection of vibrationally excited HCO produced
Ј
Ј
3
by the O„ P…¿C2H4 reaction
Jennifer L. Gardner
Stewart Radiance Laboratory, Bedford, Massachusetts 01730
Steven M. Miller
Air Force Research Laboratory, Space Vehicles Directorate, Hanscom Air Force Base,
Massachusetts 01731-3010
͑Received 6 October 2003; accepted 1 June 2004͒
The distribution of rotational and vibrational energy in HCO produced by the O(3P)ϩC2H4
2
2
˜
˜
reaction has been measured using laser-induced fluorescence detection via the B A -X A
Ј
Ј
transition. Over a detection wavelength range of 248–290 nm, our experiments have shown that
HCO is formed in both the ground state and in at least six vibrationally excited states with up to two
quanta of energy in the C-O stretch and the bending mode. Dispersed fluorescence experiments were
conducted to positively assign all of the HCO vibrational bands. The experiments confirmed that
˜
˜
many bands, including the B(000)-X(000) band, are affected by overlap with other HCO bands.
Spectral modeling was used to separate the contributions of overlapping HCO B˜-˜X bands and to
determine a nascent HCO rotational temperature of ϳ600 K, corresponding to ϳ6% of the total
energy from the O(3P)ϩC2H4 reaction. HCO vibrational distributions were determined for two
different average collision energies and were fit with vibrational temperatures of 1850Ϯ80 K and
2000Ϯ100 K, corresponding to ϳ15% of the total energy. The observed Boltzmann distribution of
vibrational energy in HCO indicates that HCO and CH3 are formed by the dissociation of an
energized intermediate complex. © 2004 American Institute of Physics.
͓DOI: 10.1063/1.1774984͔
I. INTRODUCTION
good agreement and Arrhenius plots of the overall rate tend
to display curvature at high temperatures.12,13 Some uncer-
tainty remains in the product distributions of reaction ͑1͒
with regard to the pressure dependence of the branching
ratios.3,10,14
The O(3P)ϩC2H4 reaction is an important intermediate
step in the combustion of many different hydrocarbon fuels.1
Experimental studies have determined that channels ͑1a͒ and
͑1b͒ are the dominant reactive pathways:2–5
HCO, produced by channel ͑1a͒, is an important inter-
mediate in the combustion of hydrocarbon fuels. HCO also
plays a role in the atmospheric oxidation of hydrocarbons
and is a product of the photodissociation of aldehydes such
as formaldehyde. Vaidya15 first identified the presence of
HCO in an ethylene flame based on the observation of a
series of emission bands from 250–410 nm. The first thor-
ough rotational and vibrational analysis of the HCO emis-
sions was performed by Dixon16 using an acetylene flame.
The spectroscopy of HCO has received attention due to fun-
damental interest as well as the need to develop sensitive
detection techniques for in situ measurements of HCO in
flames. HCO is an asymmetric top molecule and a near pro-
late rotor. HCO has three infrared-active modes (␯1 ,␯2 ,␯3):
␯1ϭC-H stretch, ␯2ϭC-O stretch, ␯3ϭbending mode. The
⌬Hؠ298 K kJ/mol͒
͑
3
O P͒ϩC H →HCOϩCH
Ϫ114
͑1a͒
͑1b͒
͑
2
4
3
→HϩCH2CHO Ϫ71.
Figure 1 shows a potential energy diagram for the O(3P)
ϩC2H4 reaction based on the calculations of Fueno et al.6
and Wortmann-Saleh et al.7 Cvetanovic8,9 first proposed a
mechanism for reaction ͑1͒ in which the electrophilic oxygen
atom attacks the carbon-carbon double bond and adds to one
of the olefinic carbons in C2H4 to create a triplet diradical
CH2CH2O. The triplet diradical can dissociate to form a hy-
drogen atom and a vinoxy radical ͑1b͒. An intersystem cross-
ing of the triplet diradical intermediate to the singlet state
can also occur. A 1,2-hydrogen migration in the singlet
diradical will produce an energized acetaldehyde complex
that can dissociate to form HCO and CH3 ͑1a͒.4,10 Experi-
mental studies of the branching ratios of the O(3P)ϩC2H4
reaction have determined that channels ͑1a͒ and ͑1b͒ are both
significant reaction pathways of about equal importance.2,3,11
Figure 1 also shows the energies of some proposed minor
products from the O(3P)ϩC2H4 reaction. Measurements of
the overall rate of the O(3P)ϩC2H4 reaction are in fairly
2
˜
first excited electronic state of HCO is the A A state which
Љ
has a linear equilibrium geometry and is not well suited for
laser-induced fluorescence ͑LIF͒ detection due to rapid
predissociation.17,18 In 1990, Sappey and Crosley19 demon-
2
2
˜
˜
strated for the first time that the HCO B A -X A transition
Ј
Ј
2
˜
could produce strong LIF signals. The B A state has three
Ј
vibrational modes (␯1 ,␯2 ,␯3): ␯1ϭC-H stretch, ␯ and ␯
2
3
have mixed bending mode and C-O stretch character, with
0021-9606/2004/121(12)/5920/9/$22.00
5920
© 2004 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: 193.0.65.67
On: Wed, 10 Dec 2014 14:34:42

J. Chem. Phys., Vol. 121, No. 12, 22 September 2004
Vibrationally excited HCO
5921
FIG. 1. O(3P)ϩC2H4 potential energy diagram. Solid lines represent the
triplet surface. Dashed lines represent the singlet surface.
FIG. 2. Schematic diagram of experimental apparatus.
bands has not been properly addressed, potentially leading to
errors in the analyses.
the bending mode more dominant in ␯ and the C-O stretch
2
In order to accurately determine the distribution of en-
ergy in the HCO product of the O(3P)ϩC2H4 reaction, it
was necessary to develop experimental and spectral analysis
techniques for the interpretation of complex, overlapped
HCO LIF vibrational spectra. Dispersed fluorescence ͑DF͒
experiments have been conducted to identify overlapping
more dominant in ␯3 .17 Numerous studies have investigated
the photodissociation dynamics of molecules such as
acetaldehyde,20–23 propionaldehyde,24 formaldehyde,25 and
propenal26 using B-X LIF to detect HCO. Most of these
˜ ˜
experiments have focused on the detection of HCO in the
˜
vibrational ground state X(0,0,0). However, two studies by
bands and to positively assign all of the vibrational bands.
Gejo et al.21 and Kao et al.26 have observed vibrationally
Many new B˜-˜X bands have been observed and our experi-
˜
˜
excited HCO in the X(0,0,1) and X(0,1,0) states from the
photolysis of acetaldehyde and propenal.
ments show that HCO is formed by the O(3P)ϩC2H4 reac-
tion in at least six vibrationally excited states with up to two
quanta of energy in the C-O stretch and the bending mode.
For spectral analysis of the HCO data we have utilized two
This paper reports the first measurements of the distribu-
tion of rotational and vibrational energy in HCO produced by
the O(3P)ϩC2H4 reaction. The reaction is highly exother-
mic and many different HCO vibrational states are poten-
tially accessible due to the large amount of available energy.
33
computer models, ASYTOP and IAR95.34 These models have
been used to help separate the contributions of different over-
lapping HCO bands and to investigate the distribution of
rotational and vibrational energy in HCO formed by the
O(3P)ϩC2H4 reaction.
˜ ˜
B-X LIF detection was used to monitor the production of
2
2
˜
˜
HCO. The spectroscopy of HCO via the B A -X A transi-
Ј
tion is complicated because the B˜-˜X bands exhibit hybrid a,
Ј
b-type character with contributions from both ⌬Kaϭ0 and
⌬KaϭϮ1.27 The intensity of the branches in the LIF spectra
are also affected by axis switching,28,29 Coriolis coupling,29
and predissociation.30,31 In order to properly model HCO
II. EXPERIMENTAL SECTION
A schematic diagram of the experimental apparatus is
presented in Fig. 2. The reaction cell had an internal diameter
of 4.5 cm. A shelf was machined into the top of the cell body
to allow a 2.54 cm diameter, f/1 lens to be placed 2.5 cm
above the laser beams in order to improve the LIF collection
efficiency of the photomultiplier tube ͑PMT͒. The reaction
cell was pumped by a Heraeus-Engelhard model DK 180
roughing pump, with the line constricted to reduce the pump-
ing speed. The pressure in the cell was monitored using a 10
Torr full scale MKS Baratron capacitance manometer and
was typically maintained between 0.5–4 Torr. Photolysis of
nitrogen dioxide was used to produce a source of fast O(3P)
atoms for reaction with ethylene. Under typical conditions, a
2 STP cm3 minϪ1 flow of NO2 was mixed with a 100 STP
cm3 minϪ1 flow of C2H4 , and the flows were maintained
using calibrated MKS flow controllers. The concentration of
C2H4 was kept in large excess over NO2 .
˜ ˜
B-X LIF spectra, all of these effects must be taken into ac-
˜
˜
count. The HCO B(000)-X(000) transition has been widely
studied and used to investigate the photodissociation dynam-
ics of aldehydes. In contrast, there have only been a few
measurements of vibrationally excited HCO using B˜-X˜ LIF,
and the spectroscopy of vibrationally excited HCO has re-
ceived less attention.
˜
A variety of B(␯1 ,␯2 ,␯3) upper states are accessible via
LIF.19,32 However, only relative Franck–Condon factors have
been measured for the B˜-˜X transitions, and the experiments
˜
have only been conducted for a few of the B(␯1 ,␯2 ,␯3)
upper states.32 Over the detection wavelength range used in
our experiments, more than 50 different HCO
˜
˜
B(␯1 ,␯2 ,␯3)-X(␯1 ,␯2 ,␯3) bands are potentially accessible.
Many of the HCO vibrational bands are overlapped and in
numerous cases the relative strengths of the bands are not
known, making the analysis and even the assignment of the
bands difficult. In previous papers on the production of vi-
brationally excited HCO from the photolysis of
aldehydes,21,26 the issue of overlapping HCO vibrational
The laser pulses used to photodissociate NO2 were gen-
erated by a Spectra-Physics PDL-2 dye laser side pumped by
a Spectra-Physics DCR-2A Nd:YAG laser operating at 10
Hz. The dye laser output was frequency doubled using an
Inrad Autotracker II system with a compensated KDP crystal
in combination with a four-prism separator. Using LDS 698
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: 193.0.65.67
On: Wed, 10 Dec 2014 14:34:42

5922
J. Chem. Phys., Vol. 121, No. 12, 22 September 2004
J. L. Gardner and S. M. Miller
and LDS 765 laser dyes, ultraviolet light with a wavelength
ranging from 348 to 395 nm was produced. At these wave-
lengths, NO2 dissociates to produce:
3
NO ϩh␯→O P͒ϩNO X 2⌸ , j͒.
͑2͒
͑
͑
v
2
The NO produced by this reaction can be formed in vibra-
tionally excited states depending on the photolysis wave-
length. For example, the threshold for producing NO
(X 2⌸) ϭ1 is 370.3 nm.35 The kinetic energy of the O(3P)
v
3
˜ ˜
FIG. 3. B-X LIF excitation scan of HCO produced by the O( P)ϩC2H4
reaction. The HCO vibrational bands are labeled based on the dominant
character of the band.
atoms produced by reaction ͑2͒ can be varied by changing
the NO2 photolysis wavelength, and the distribution of en-
ergy depends on the distribution of vibrational and rotational
energy in NO.36 In our experiments, the photodissociation
laser energy was typically around 1 mJ/pulse, and the laser
beam was focused into the center of the cell using a 30 cm
focal length lens.
III. RESULTS AND DISCUSSION
A. LIF excitation spectra
Numerous B˜-˜X LIF excitation scans of HCO formed by
The O(3P) atoms were allowed to react with an excess
of C2H4 , and then the production of HCO from reaction ͑1a͒
was monitored using LIF detection. HCO molecules were
the O(3P)ϩC2H4 reaction were taken over a range of NO2
photodissociation wavelengths. These experiments are the
first measurements of vibrationally excited HCO from the
O(3P)ϩC2H4 reaction. A sample excitation scan is pre-
sented in Fig. 3, which demonstrates the range of HCO
2
˜
excited to the B A electronic state by laser pulses generated
Ј
by a second Nd:YAG pumped dye laser. Using Coumarin
500 and Coumarin 540A laser dyes, ultraviolet light with a
wavelength ranging from 248–290 nm was produced. The
HCO detection wavelength was determined by sending a
small portion of the laser beam into a calibrated Burleigh
wavemeter model WA-5500. The detection laser energy typi-
cally ranged from 50–200 ␮J/pulse, and the laser beam was
focused into the center of the reaction cell using a 50 cm
focal length lens. The delay time between the photodissocia-
tion laser pulse and the detection laser pulse was controlled
using a Stanford Research Systems model DG535 delay gen-
erator.
˜
˜
B(␯1 ,␯2 ,␯3)-X(␯1 ,␯2 ,␯3) bands accessible using Cou-
marin 540A laser dye in the detection laser. The bands are
labeled based on the dominant character of the band. In this
figure, the data has not been corrected for variations in the
LIF detection laser energy. Under typical experimental con-
ditions, the HCO LIF signal was partially saturated. Power
dependence studies were conducted in order to develop cor-
rection factors to account for the partial saturation in the data
analysis. The correction factors typically ranged from 1.3 to
1.8. Over a detection wavelength range of 248–290 nm, our
experiments have shown that HCO is formed in both the
ground state and in at least six vibrationally excited states
with up to two quanta of energy in the C-O stretch and the
bending mode. The C-H stretch of HCO is not accessible by
2
2
˜
˜
Fluorescence emission from the HCO B A →X A
Ј
Ј
transition was detected using a Hamamatsu model R105 pho-
2
2
˜
˜
tomultiplier. In initial experiments, the HCO B A -X A
Ј
Ј
LIF detection method was tested and optimized using the
photolysis of acetaldehyde between 280–315 nm to produce
2
2
˜
˜
LIF via the B A -X A transition due to low Franck–
Ј
Ј
Condon factors.32
X(␯1 ,␯2 ,␯3) states and vibrational energy levels
A
summary of observed HCO
2
˜
HCO. The HCO B A state emission is spread out from
Ј
32,37
˜
is pre-
250–450 nm, with maximum intensity near 320 nm. For the
excitation scans, Andover standard bandpass filters 307FS10-
25, 350FS10-25, or 400FS40-25, with center wavelengths of
307, 350, or 400 nm, respectively, were used to eliminate
laser scatter depending on the wavelength of the photolysis
laser. In the dispersed fluorescence experiments, an Acton
Research SpectraPro-150 monochromator with an 1800
g/mm grating was inserted above the reaction cell window.
˜
sented in Fig. 4. The HCO X(012) state is the highest energy
vibrational level observed in our study. Presumably, the HCO
2
2
˜
˜
An f/4 lens was used to focus the HCO B A -X A fluores-
Ј
Ј
cence onto the entrance slit, and the Hamamatsu R105 PMT
was placed at the exit slit. The monochromator entrance and
exit slit widths were set at 0.5 mm corresponding to a spec-
tral resolution of 1 nm. The HCO LIF signal was sent
through a fast Stanford Research Systems model SR240 cur-
rent amplifier to a LeCroy model 9310AM oscilloscope. The
wavelength scanning and the collection of data were con-
trolled by a PC using LabVIEW data acquisition software.
˜
FIG. 4. HCO X(␯1 ,␯2 ,␯3) vibrational energy levels. Boxes indicate vibra-
tional states observed in this study.
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: 193.0.65.67
On: Wed, 10 Dec 2014 14:34:42

J. Chem. Phys., Vol. 121, No. 12, 22 September 2004
Vibrationally excited HCO
TABLE I. HCO B A -X A band origins.
5923
2
2
˜
˜
3
Ј
Ј
˜
˜
X(003) and X(004) states are also formed by the O( P)
ϩC2H4 reaction, but were not detected in our experiments
due to low Franck–Condon factors.32
Band
origin
No.
B A
X A
͑cmϪ1
͒
2
2
˜
˜
Ј
Ј
As demonstrated in Fig. 3,
a variety of HCO
˜
B(␯1 ,␯2 ,␯3) upper states are available via LIF. The lifetimes
1
2a
͑100͒
͑010͒
͑011͒
͑001͒
͑002͒
͑100͒
͑011͒
͑100͒
͑010͒
͑011͒
͑002͒
͑000͒
͑001͒
͑002͒
͑100͒
͑010͒
͑011͒
͑100͒
͑011͒
͑010͒
͑001͒
͑002͒
͑002͒
͑001͒
͑000͒
͑100͒
͑011͒
͑100͒
͑010͒
͑011͒
͑002͒
͑010͒
͑000͒
͑001͒
͑002͒
͑001͒
͑000͒
͑100͒
͑010͒
͑011͒
͑010͒
͑001͒
͑002͒
͑001͒
͑100͒
͑000͒
͑011͒
͑000͒
͑010͒
͑002͒
͑000͒
͑001͒
͑000͒
͑010͒
͑001͒
͑000͒
͑001͒
͑000͒
͑001͒
͑010͒
͑010͒
͑002͒
͑001͒
͑002͒
͑010͒
͑000͒
͑001͒
͑002͒
͑011͒
͑010͒
͑011͒
͑003͒
͑003͒
͑002͒
͑010͒
͑011͒
͑003͒
͑002͒
͑001͒
͑020͒
͑020͒
͑012͒
͑011͒
͑012͒
͑020͒
͑003͒
͑010͒
͑011͒
͑012͒
͑003͒
͑002͒
͑021͒
͑020͒
͑021͒
͑012͒
͑020͒
͑021͒
͑012͒
͑030͒
͑011͒
͑030͒
͑003͒
͑021͒
͑030͒
͑020͒
͑021͒
͑012͒
͑030͒
40 211
40 075
40 044
39 760
39 728
39 423
39 257
39 150
38 994
38 983
38 940
38 695
38 679
38 667
38 350
38 207
38 183
38 121
37 954
37 933
37 891
37 867
37 638
37 618
37 614
37 583
37 416
37 295
37 133
37 128
37 100
36 904
36 826
36 818
36 812
36 589
36 553
36 508
36 366
36 342
36 078
36 051
36 025
35 763
35 762
35 753
35 595
35 524
35 292
35 279
34 986
34 976
34 698
34 545
˜
of the B states decrease rapidly with increasing vibrational
energy due to predissociation.31,38 As a result, LIF signals
3
4a
˜
˜
˜
from states such as B(002), B(011), and B(100) tend to be
5
˜
˜
weaker than from states such as B(000), B(001), and
B(010). The LIF signal strength also depends on the initial
6a
7a
˜
8
HCO population, the Franck–Condon factor, the detection
laser energy, and the wavelength range of the emission band-
pass filter. Over the detection wavelength range used in our
9a
10
11a
12a
13a
14
15
16a
17
18
19
20
21a
22
23
24a
25a
26
27
28
29a
30
31
32
33a
34a
35
36
37a
38
39a
40
41
42a
43
44
45
46a
47
48
49
50
51a
52
53a
54
experiments,
˜
more
than
50
different
HCO
˜
B(␯1 ,␯2 ,␯3)-X(␯1 ,␯2 ,␯3) bands are potentially accessible,
as shown in Table I. The band origins listed in Table I have
been calculated from vibrational term energies reported by
Tobiason et al.32 and Cool and Song.39 There are many cases
where the band origins of two different HCO bands are sepa-
rated by less than 20 cmϪ1. As a result, it is difficult to
positively identify overlapped bands in the LIF excitation
scans. In order to determine the distribution of energy in
HCO from the O(3P)ϩC2H4 reaction, it is essential to de-
termine which bands are overlapped and to separate the con-
tributions from each band.
˜
˜
The HCO B(000)-X(000) band is an example of a band
that may be contaminated due to overlap with the
˜
˜
˜
˜
B(001)-X(001) band. The B(000)-X(000) band origin is at
38 695 cmϪ1, while the B(001)-X(001) band origin is only
˜
˜
16 cmϪ1 away at 38 679 cmϪ1. There are no measurements
˜
˜
of even the relative band strength of the B(001)-X(001)
band in the literature. Previous papers on the production of
vibrationally excited HCO from the photolysis of aldehydes
have not addressed this potential band overlap problem.21,26
In order to determine if the B(001)-X(001) band makes a
significant contribution to the LIF signal in this region, ex-
periments on the production of HCO from the photodissocia-
tion of acetaldehyde were conducted for comparison. Photo-
dissociation of CH3CHO between 308.7 and 320.5 nm
produces HCO in only the ground vibrational state.20,21 LIF
excitation scans of the HCO B(000)-X(000) band were
taken following the photolysis of CH3CHO at 315 nm. A
comparison of the HCO data from the CH3CHO experiments
with the O(3P)ϩC2H4 experiments shows the onset of sig-
nificant spectral differences in the O(3P)ϩC2H4 data start-
˜
˜
˜
˜
ing at 38 674 cmϪ1, near the B(001)-X(001) band origin.
These experiments demonstrate that the B(001)-X(001)
band is an important source of LIF signal in this region,
presumably due to favorable Franck–Condon factors for
˜
˜
˜
˜
a
2
2
˜
˜
HCO B A -X A bands observed in this study.
Ј
Ј
˜
transitions to the B(001) state. Furthermore, these experi-
ments indicate that band overlap is most likely a significant
problem throughout the HCO detection region.
2
˜
tify any overlapping bands. The HCO B A state emission is
Ј
spread out over a wide spectral region from 250–450 nm,
with maximum intensity near 320 nm. Each HCO
B(␯1 ,␯2 ,␯3) upper state has a unique emission spectrum. By
scanning the monochromator and examining the wavelength
dependence of the emission, it is possible to positively de-
B. Dispersed fluorescence spectra
˜
Dispersed fluorescence experiments using a monochro-
mator were conducted to positively assign all of the HCO
3
˜ ˜
B-X vibrational bands in the O( P)ϩC2H4 data and to iden-
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: 193.0.65.67
On: Wed, 10 Dec 2014 14:34:42

5924
J. Chem. Phys., Vol. 121, No. 12, 22 September 2004
J. L. Gardner and S. M. Miller
FIG. 5. Dispersed fluorescence spectra of HCO pro-
duced by the O(3P)ϩC2H4 reaction: ͑a͒ Pure B(000)
˜
dispersed fluorescence spectrum taken at an excitation
wave number of 38 693 cmϪ1. ͑b͒ Pure B(001) dis-
˜
persed fluorescence spectrum taken at an excitation
wave number of 37 889 cmϪ1. ͑c͒ Dispersed fluores-
cence spectrum taken at an excitation wave number of
38 674 cmϪ1. The spectrum is a mixture of the B(000)
˜
˜
and B(001) spectra due to the overlap of the
˜
˜
˜
˜
B(000)-X(000) and B(001)-X(001) bands. ͑d͒ Fit to
˜
spectrum ͑c͒ made up of 57% B(000) and
˜
43% B(001).
termine which upper state produces the fluorescence. Sample
dispersed fluorescence spectra are shown in Fig. 5. These
spectra were collected in order to confirm the analysis in the
fluorescence experiments and the results of Tobiason et al.32
demonstrate that all of the HCO B(␯1 ,␯2 ,␯3) upper states
˜
emit over a wide spectral region. The problem of overlapping
bands could potentially be solved by using a very narrow
bandpass filter. However, this approach was not taken be-
cause it would greatly reduce the LIF signal and potentially
bias the measurement of the HCO rotational energy distribu-
tion by creating preferential detection of some rotational
states.
˜
˜
preceding section that the B(001)-X(001) band significantly
˜
˜
contributes to the LIF signal in the B(000)-X(000) band
region.
As shown in Fig. 5, the fluorescence emission patterns
˜
˜
for the HCO B(000) and B(001) states are very different.
˜
Figure 5͑a͒ is a pure B(000) dispersed fluorescence spectrum
taken from the clean region of the B(000)-X(000) band. Our
HCO B(000) dispersed fluorescence data is in very good
˜
˜
˜
C. Spectral modeling
agreement with the results of a previous study by Tobiason
Spectral modeling has been used to analyze the over-
et al.32 Figure 5͑b͒ is a pure B(001) spectrum taken from the
˜
˜ ˜
scans taken during the O(3P)ϩC2H4 experiments. A com-
lapped HCO B-X vibrational bands in the LIF excitation
˜
˜
clean region of the B(001)-X(010) band. For the spectrum
in Fig. 5͑c͒ the detection laser was set to an excitation wave
number of 38 674 cmϪ1, close to the band origin of the
˜ ˜
puter program that can simultaneously fit two HCO B-X
bands was developed to model the overlapped bands. In the
preceding section, dispersed fluorescence experiments were
used to positively identify underlying bands in overlapped
regions. However, the technique can only probe the compo-
sition of an overlapped band one excitation wavelength at a
time. In order to determine the distribution of energy in
HCO, it is essential to have a quantitative measurement of
how much each band contributes to the observed LIF signal
in an overlapped region. For spectral analysis of the HCO
data we have utilized two computer models, ASYTOP devel-
oped by Sears33 and IAR95 developed by Plusquellic,34 in
combination with our own two-band least-squares fitting rou-
tine. Molecular constants for HCO used in the computer
models were taken from the literature.40–46 Molecular con-
˜
˜
B(001)-X(001) band. The dispersed fluorescence spectrum
˜
˜
in Fig. 5͑c͒ is clearly a mixture of the B(000) and B(001)
spectra in Figs. 5͑a͒ and 5͑b͒. Figure 5͑c͒ can be easily fit by
a linear combination of the two spectra made up of 57%
˜
˜
B(000) and 43% B(001) as shown in Fig. 5͑d͒. The dis-
persed fluorescence experiments have confirmed that the
˜
˜
B(001)-X(001) band strongly contributes to the observed
LIF excitation signal and must be accounted for when ana-
˜
˜
lyzing B(000)-X(000) spectra.
The issue of overlapping HCO bands is a concern
throughout the entire detection wavelength range for the LIF
excitation experiments. Fluorescence emission scans were
˜
˜
˜
recorded for the HCO B(010), B(100), and B(002) upper
˜
states following excitation in clean regions of the
stants are not available for all of the X(␯1 ,␯2 ,␯3) states
˜
˜
˜
˜
˜
˜
˜
˜
˜
B(010)-X(010), B(100)-X(010), and B(002)-X(010)
observed in this study. In the case of the X(011) and X(002)
states, molecular constants were estimated based on trends in
known HCO constants32,40 and then adjusted slightly to im-
prove fits to the HCO data.
bands, respectively. These scans, in addition to the B(000)
and B(001) spectra shown in Figs. 5͑a͒ and 5͑b͒, were used
˜
as a reference data set for comparison with spectra from
overlapped band regions. Dispersed fluorescence has been
used to positively identify all of the HCO
The ASYTOP model was designed to calculate transition
frequencies and intensities for asymmetric top free radical
molecules using an effective Hamiltonian operator that in-
cludes many effects including vibrational, rotational, cen-
trifugal distortion, and spin-rotation interactions. The IAR95
program has similar features, but also incorporates the effect
of axis switching in the calculations. Excitation of HCO to
˜
˜
B(␯1 ,␯2 ,␯3)-X(␯1 ,␯2 ,␯3) LIF excitation bands observed in
the O(3P)ϩC2H4 experiments. A total of 21 bands have
been observed, and these bands are identified in Table I.
Overall, the dispersed fluorescence experiments show that
˜
˜
˜
˜
˜
˜
the B(000)-X(000), B(000)-X(001), and B(000)-X(010)
2
˜
bands all have significant band overlap issues. Our dispersed
the B A state alters the geometry of HCO resulting in a
Ј
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: 193.0.65.67
On: Wed, 10 Dec 2014 14:34:42

J. Chem. Phys., Vol. 121, No. 12, 22 September 2004
Vibrationally excited HCO
5925
˜ ˜
FIG. 6. Comparison of HCO B-X LIF excitation data
with the results of a two-band model: ͑a͒ This data set
was obtained with the NO2 photodissociation ␭ϭ348
nm. ͑b͒ Fit to spectrum ͑a͒ using a two-band model. ͑c͒
˜
˜
Contribution to the model from the B(000)-X(000)
band ͑53%͒. ͑d͒ Contribution to the model from the
˜
˜
B(001)-X(001) band ͑47%͒.
rotation of the principal inertial axes. Lee and Chen29 have
determined the axis-switching angle to be 2.6°. The axis
switching has an effect on the distribution of intensity in the
branches of the LIF spectra. In order to properly model HCO
those of the IAR95 model. Overall, the fits using the IAR95
model are similar to or slightly better than the fits using the
ASYTOP model. Both models give similar results when calcu-
lating the percent contributions to the overlapped spectra
˜ ˜
˜
˜
˜
˜
B-X LIF spectra, some additional effects must be taken into
from the B(000)-X(000) band and the B(001)-X(001)
band. Thus, the details of the spectral modeling with regard
to axis switching will not have a significant effect on the
determination of HCO vibrational populations. It is impor-
tant to consider that nascent HCO produced by the O(3P)
ϩC2H4 reaction is not necessarily formed with a Boltzmann
2
2
˜
˜
account. For example, the B A -X A bands exhibit hybrid
Ј
Ј
a, b-type character with contributions from both ⌬Kaϭ0 and
⌬KaϭϮ1 because the electric dipole transition moment lies
in the molecular plane between the a- and b-inertial axes.
Lee and Chen29 have determined that the angle between the
electric dipole transition moment and the a axis is 39° for the
˜
˜
rotational energy distribution. In our HCO B(000)-X(000)
˜
˜
B(000)-X(000) band. Intensities in the HCO B˜-X˜ spectra
data, most of the rotational lines are blended due to insuffi-
2
˜
˜
˜
are also influenced by Coriolis coupling between the B A
and A A states. A correction for Coriolis coupling has been
applied in our model.29 In addition, the B A state is predis-
Ј
cient resolution and due to the overlap of the B(001)-X(001)
band. Therefore, it is not possible to determine the distribu-
tion of rotational energy based on the intensity of individual
rotational transitions. However, using a Boltzmann distribu-
tion for the HCO rotational energy does give a reasonable fit
to the data and also provides an estimate of the amount of
energy in rotation. The best fits to the data give nascent ro-
tational temperatures of ϳ600 K, corresponding to ϳ6% of
the total energy from the O(3P)ϩC2H4 reaction. As a con-
sistency check, we have also modeled HCO data taken at a
laser delay of several microseconds, and both models fit the
data with temperatures close to 300 K. This result is in good
agreement with the fact that at long delay times HCO in our
experiments is fully rotationally relaxed due to collisions
with room temperature C2H4 .
2
˜
Љ
2
˜
2
Ј
˜
sociative, and the lifetimes of the B A states display a
Ј
strong dependence on KЈ and a weaker dependence on N .
Ј
a
Measurements of the lifetimes of the B A states29,31 have
2
˜
Ј
been used to correct for the effect of predissociation in our
model.
A comparison of HCO B˜-˜X LIF data taken during the
O(3P)ϩC2H4 experiments with results from the IAR95
model is presented in Fig. 6. Figure 6͑a͒ is the experimental
˜
˜
data taken in the B(000)-X(000) band region. Figure 6͑b͒ is
a two-band fit to the data using the IAR95 model with contri-
˜
˜
˜
˜
butions from both the B(000)-X(000) and B(001)-X(001)
bands. In Figs. 6͑c͒ and 6͑d͒, the contributions to the model
The spectral modeling has confirmed the results of the
dispersed fluorescence experiments by verifying that the
˜
˜
˜
˜
from the B(000)-X(000) and B(001)-X(001) bands are dis-
played separately. The modeling shows that it is necessary to
˜
˜
B(001)-X(001) band makes a significant contribution to the
˜
˜
include a significant contribution from the B(001)-X(001)
band in order to fit the data. In fact, Fig. 6͑b͒ is made up of
˜
˜
observed HCO LIF signal in the B(000)-X(000) band re-
gion. In previous papers on the production of vibrationally
excited HCO from the photolysis of aldehydes, the issue of
overlapping HCO bands has not been given sufficient atten-
˜
˜
˜
˜
53% B(000)-X(000) and 47% B(001)-X(001). The impor-
˜
˜
tance of the B(001)-X(001) band is particularly evident in
the good fit of the large peak near 38 674 cmϪ1. In addition,
the angle between the electric dipole transition moment and
the a axis is not known for the B(001)-X(001) band. As a
result, this angle was allowed to float in our model, and the
best fits were produced by using an angle close to 45°.
The HCO spectra have been fit assuming a Boltzmann
rotational energy population distribution. The ASYTOP model,
which does not include the effect of axis switching, tends to
give rotational temperatures that are 20–50 K lower than
tion. For example, Gejo et al.21
used B˜-˜X LIF to measure the
distribution of vibrational energy in HCO from the photoly-
sis of acetaldehyde as a function of the photolysis wave-
length from 280–310 nm. They determined relative popula-
˜
˜
˜
˜
˜
tions of HCO in the X(000), X(001), and X(010)
vibrational states by measuring the integrated intensity of the
˜
˜
LIF excitation spectra for the B(000)-X(000),
˜
˜
˜
˜
B(000)-X(001), and B(000)-X(010) bands. Although they
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: 193.0.65.67
On: Wed, 10 Dec 2014 14:34:42

5926
J. Chem. Phys., Vol. 121, No. 12, 22 September 2004
J. L. Gardner and S. M. Miller
TABLE II. HCO vibrational populations.
Relative population
did not specify which bandpass filters were used in their
experiments, it is likely that their measurements were af-
fected by the overlap of the B(001)-X(001) band with the
Relative population
NO2 photolysis at
348 nm
˜
˜
Vibrational state
HCO
NO2 photolysis at
395 nm
˜
˜
B(000)-X(000) band, possibly leading to errors in their cal-
culations. In another study, Kao et al.26 used B-X LIF to
˜ ˜
1.00Ϯ0.14
0.39Ϯ0.08
0.23Ϯ0.04
0.22Ϯ0.06
0.09Ϯ0.02
0.05Ϯ0.01
0.05Ϯ0.01
1.00Ϯ0.15
0.58Ϯ0.11
0.31Ϯ0.05
0.29Ϯ0.09
0.12Ϯ0.03
0.08Ϯ0.02
0.06Ϯ0.02
X(000)
X(001)
X(010)
X(002)
X(011)
X(020)
X(012)
˜
determine the relative populations of HCO in the X(000),
˜
˜
X(001), and X(010) vibrational states following the pho-
tolysis of propenal at 193 nm. They monitored only the
˜
˜
˜
˜
˜
˜
B(000)-X(000), B(000)-X(001), and B(000)-X(010)
bands of HCO. However, at 193 nm there is more than 200
kJ/mol of available energy, and HCO is most likely formed
in additional vibrational states with higher energy. As a re-
sult, there are many possible overlapping bands that could
have affected the results of the experiments.
Ave collision E
Vibrational temperature
Total vibrational E
3 kJ/mol
1850Ϯ80 K
17.0 kJ/mol
12 kJ/mol
2000Ϯ100 K
18.5 kJ/mol
The Boltzmann distribution of vibrational energy in
HCO indicates that the product is formed from the dissocia-
tion of a complex, rather than a direct reaction. This result is
consistent with the mechanism outlined in Fig. 1 in which
O(3P)ϩC2H4 react to form a diradical that transforms into
an energized acetaldehyde intermediate via intersystem
crossing and a 1,2-hydrogen migration. HCO and CH3 are
then produced by the dissociation of the energized CH3CHO
complex. The experimental results for the HCO vibrational
energy distribution are in reasonably good agreement with
the results of a statistical model. The model assumes a prior
distribution in which all quantum states of the products are
equally probable.47 The model is based on the principle of
maximizing entropy, and therefore does not include any dy-
namical information. Using the rigid-rotor-harmonic-
oscillator ͑RRHO͒ approximation for a polyatomic, the ex-
pected prior distribution for HCO vibrational energy is given
by48
D. Vibrational energy distribution
The nascent distribution of vibrational energy in HCO
produced by the O(3P)ϩC2H4 reaction has been measured.
˜
In order to simplify the analysis, only HCO B(000)-
˜
X(␯1 ,␯2 ,␯3) bands were used for determining the energy
distribution. Since all of the fluorescence originated from the
same upper state, no corrections were necessary for varia-
tions in the emission spectrum or the predissociation of the B
state. In overlapped band regions, spectral modeling was
used to separate the contributions of different bands. Correc-
tions were made for Franck–Condon factors,32 variations in
laser power, partial saturation of the LIF signal, and day-to-
day variability.
˜
Data sets have been analyzed for two different NO2 pho-
todissociation wavelengths. In our experiments, O(3P) at-
oms are produced by NO2 photolysis, and the kinetic energy
of the O(3P) atoms can be varied by changing the photolysis
wavelength. The amount of energy available for O transla-
tion depends on the distribution of vibrational and rotational
energy in the NO fragment. For example at 355 nm, state-
resolved photofragment imaging has shown that the O(3P)
atoms have a bimodal velocity distribution with peak speeds
of 0.9 and 1.4 km sϪ1 corresponding to NO (X 2⌸) produced
in ϭ1 and ϭ0.36 The first NO2 photolysis wavelength in
P0 ͒ϰ EϪE ͒␥.
͑v
͑3͒
͑
v
E is the total available energy and ␥ϭsϩ(␣/2)Ϫ1, where s
is the number of vibrational degrees of freedom of the CH3
coproduct, and ␣ is the number of translational-rotational
degrees of freedom of both products in the center of mass
v
v
our experiments was 395 nm, just below the NO2 dissocia-
tion threshold, and the second was 348 nm, just above the
threshold for formation of NO ( ϭ2). The difference in
v
average O(3P) atom kinetic energy is estimated to be ϳ15
kJ/mol, resulting in an average difference in O-C2H4 colli-
sion energy of ϳ9 kJ/mol. This amount of additional energy
is small compared to the 114 kJ/mol of energy released due
to the exothermicity of the O(3P)ϩC2H4 reaction. Nonethe-
less, the additional collision energy does seem to have a
small effect on the observed HCO vibrational populations as
shown in Table II. A Boltzmann plot of the relative HCO
vibrational populations for the 395 nm data is presented in
Fig. 7. Both data sets are well fit by a Boltzmann distribution
with a vibrational temperature of 1850Ϯ80 K for the 395 nm
data and 2000Ϯ100 K for the 348 nm data. The additional
collision energy available in the 348 nm experiments is re-
flected in a slightly hotter HCO vibrational energy distribu-
tion.
FIG. 7. Boltzmann plot of the HCO nascent vibrational distribution pro-
duced by the O(3P)ϩC2H4 reaction. The HCO data was fit with a vibra-
tional temperature of 1850Ϯ80 K. This data set was obtained with the NO2
photodissociation ␭ϭ395 nm.
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: 193.0.65.67
On: Wed, 10 Dec 2014 14:34:42

J. Chem. Phys., Vol. 121, No. 12, 22 September 2004
Vibrationally excited HCO
5927
TABLE III. Average partitioning of energy in the HCOϩCH3 products of the O(3P)ϩC2H4 reaction
Experimental energy Experimental energy
HCOϩCH3
statistical model
͑%͒
CH3CHO photolysis
impulsive model
͑%͒
partitioning
partitioning
͑kJ/mol͒
͑%͒
HCO rotational E
HCO vibrational E
CH3 RotϩVib E
7.5
18.5
56
6%
15%
44%
35%
10%
20%
50%
20%
19%
11%
10%
60%
Total translational E
44
than the highest average collision energy in this study. In
Table III, the amount of internal energy in the CH3 product
has been calculated by subtracting the HCO internal energy
and the estimated translational energy of both products from
the total available energy.
system. In our case, sϭ3NϪ6ϭ6 for CH3 and ␣ϭ9 for
HCOϩCH3 , so ␥ϭ9.5. In the limit EvӶE, the RRHO ap-
proximation is more accurate, and P0( ) can be approxi-
v
mated by
P0 ͒ϰexp ϪE /kT ͒,
͑v
͑4͒
͑
v
v
A very simple statistical calculation, assuming equipar-
tition of energy in the O(3P)ϩC2H4 reaction, is also pre-
sented in Table III. In this simple statistical model it is as-
sumed that (3/2)kT goes into the rotation of each product,
(3/2)kT goes into the translation of each product and (3N
Ϫ6)kT goes into the vibration of each product, for a total
energy of 15 kT. Although this model is extremely simplistic,
it matches up reasonably well with the experimental results.
For comparison, a direct counting of internal states of HCO
and CH3 was conducted using the DenSUM program from
the MultiWell software developed by Barker.49 In this statis-
tical calculation, all internal states of the products are as-
sumed to be equally probable. The direct counting method
partitions a little more of the internal energy into CH3 com-
pared to the simple statistical model presented in Table III.
For further comparison, the result of an impulsive model
calculation for CH3CHO photodissociation by Terentis
et al.22 is also presented in Table III. The impulsive model
assumes that CH3CHO breaks apart into HCOϩCH3 due to
vibrational motion, without general equilibration of internal
energy. The model predicts energy partitioning in the prod-
ucts based on simple equations describing forces and torques
between the separating fragments, including conservation of
energy, linear and angular momentum. The impulsive model
is in good agreement with the experimental results of Ter-
entis et al. for the photolysis of acetaldehyde, but the calcu-
lation does not agree well with our experimental results for
the O(3P)ϩC2H4 reaction.
with kTvϭE/␥. Thus, the statistical model predicts a
Boltzmann-like vibrational energy distribution. For the HCO
data collected using a NO2 photolysis wavelength of 348 nm,
the statistical model predicts an HCO vibrational temperature
of 1600 K, in reasonable agreement with the experimentally
measured temperature of 2000 K. The agreement supports
the theory that HCO is formed by the dissociation of an
intermediate CH3CHO complex that lives long enough for
the energy to be redistributed statistically.
The overall partitioning of energy in the HCOϩCH3
channel of the O(3P)ϩC2H4 reaction can be analyzed by
comparing our results on the distribution of internal energy
in HCO with previous results on the translational energy dis-
tribution. Table III summarizes the average amount of energy
partitioned into HCO rotation and vibration as determined by
our experimental measurements. The results listed in Table
III are for experiments conducted using a NO2 photolysis
wavelength of 348 nm to produce the O(3P) atoms. The
partitioning of energy in HCO for the 395 nm data is very
similar to the results listed in Table III. Thus, the small in-
crease in the average available energy at 348 nm, due to the
increased average O(3P) atom kinetic energy, does not have
a significant impact on the overall partitioning of energy. The
results are consistent with a mechanism in which energy gets
redistributed statistically in the HCOϩCH3 products. The
HCO rotational energy in Table III was calculated from the
formula ERotϭ(3/2)kTR , where TR is 600 K as determined
by spectral modeling of the HCO data. The vibrational en-
ergy was calculated by summing the product of the normal-
ized HCO population for each level by the vibrational energy
of each level, EVibϭ͚vNvEv . The HCO vibrational data is
very well fit by a Boltzmann distribution, as shown in Fig. 7.
Therefore, the populations of HCO vibrational states that
IV. CONCLUSIONS
Vibrationally excited HCO from the O(3P)ϩC2H4 reac-
tion has been detected using B˜-˜X LIF. HCO is produced by
the reaction in a number of vibrationally excited states, re-
sulting in many overlapping B˜-˜X bands. Dispersed fluores-
˜
were not observed in this study, such as X(100), were esti-
cence experiments were conducted to positively assign all of
mated from the Boltzmann fit to the data. In Table III, the
amount of energy partitioned into translation of the HCO
ϩCH3 products comes from the crossed molecular beam
measurements of Schmoltner et al.4 They found the average
translational energy of the HCOϩCH3 products to be 50
kJ/mol, corresponding to 35% of the total available energy. It
should be noted that the average collision energy in the mo-
lecular beam experiments was about a factor of 2 greater
3
˜ ˜
the HCO B-X vibrational bands in the O( P)ϩC2H4 data
and to identify any overlapping bands. The experiments con-
˜
˜
firmed that the B(001)-X(001) band strongly contributes to
˜
˜
the observed LIF excitation signal in the B(000)-X(000)
band region. The issue of overlapping bands needs to be
addressed in any future studies involving the detection of
vibrationally excited HCO using B˜-˜X LIF. Spectral modeling
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: 193.0.65.67
On: Wed, 10 Dec 2014 14:34:42

5928
J. Chem. Phys., Vol. 121, No. 12, 22 September 2004
J. L. Gardner and S. M. Miller
was used to separate the contributions of overlapping HCO
B-X bands and to investigate the distribution of rotational
19 A. D. Sappey and D. R. Crosley, J. Chem. Phys. 93, 7601 ͑1990͒.
20 T. Kono, M. Takayanagi, and I. Hanazaki, J. Phys. Chem. 97, 12793
͑1993͒.
˜ ˜
and vibrational energy in HCO formed by the O(3P)
ϩC2H4 reaction. The nascent HCO data were fit with a ro-
tational temperature of ϳ600 K, corresponding to ϳ6% of
the total available energy, and a vibrational temperature of
ϳ2000 K, corresponding to ϳ15% of the total available en-
ergy. The results indicate that HCO and CH3 are formed by
the dissociation of an energized CH3CHO intermediate com-
plex, and the reaction energy is distributed fairly statistically
in the products.
21 T. Gejo, M. Takayanagi, T. Kono, and I. Hanazaki, Chem. Phys. Lett. 218,
343 ͑1994͒.
22 A. C. Terentis, M. Stone, and S. H. Kable, J. Phys. Chem. 98, 10802
͑1994͒.
23 S.-H. Lee and I.-C. Chen, J. Chem. Phys. 105, 4597 ͑1996͒.
24 A. C. Terentis, P. T. Knepp, and S. H. Kable, J. Phys. Chem. 99, 12704
͑1995͒.
25 A. C. Terentis, S. E. Waugh, G. F. Metha, and S. H. Kable, J. Chem. Phys.
108, 3187 ͑1998͒.
26 Y.-T. Kao, W.-C. Chen, C.-H. Yu, and I.-C. Chen, J. Chem. Phys. 114,
8964 ͑2001͒.
27 G. W. Adamson, X. Zhao, and R. W. Field, J. Mol. Spectrosc. 160, 11
͑1993͒.
ACKNOWLEDGMENT
28 J. T. Hougen and J. K. G. Watson, Can. J. Phys. 43, 298 ͑1965͒.
29 S.-H. Lee and I.-C. Chen, J. Chem. Phys. 105, 2583 ͑1996͒.
30 M. R. Manaa, J. Chem. Phys. 112, 8789 ͑2000͒.
31 J. D. Tobiason and E. A. Rohlfing, Chem. Phys. Lett. 252, 333 ͑1996͒.
32 J. D. Tobiason, J. R. Dunlop, and E. A. Rohlfing, J. Chem. Phys. 103,
1448 ͑1995͒.
This work was supported by the Air Force Office of
Scientific Research under task 2303ES/92VS04COR.
1 W. C. Gardiner, Gas-Phase Combustion Chemistry ͑Springer, New York,
2000͒.
33 T. J. Sears, Comput. Phys. Commun. 34, 123 ͑1984͒.
34 D. F. Plusquellic, User Guide to the Jb95 Spectral Fitting Program ͑Na-
tional Institute of Standards and Technology, Gaithersburg, 2002͒, http://
physics.nist.gov/jb95
2 Y. Endo, S. Tsuchiya, C. Yamada, and E. Hirota, J. Chem. Phys. 85, 4446
͑1986͒.
3 U. Bley, P. Dransfeld, B. Himme, M. Koch, F. Temps, and H. Gg. Wagner,
Twenty-Second Symposium (International) on Combustion ͑Combustion
Institute, Pittsburgh, 1988͒, p. 997.
35 D. C. Robie, M. Hunter, J. L. Bates, and H. Reisler, Chem. Phys. Lett.
193, 413 ͑1992͒.
4 A. M. Schmoltner, P. M. Chu, R. J. Brudzynski, and Y. T. Lee, J. Chem.
Phys. 91, 6926 ͑1989͒.
36 V. P. Hradil, T. Suzuki, S. A. Hewitt, P. L. Houston, and B. J. Whitaker, J.
Chem. Phys. 99, 4455 ͑1993͒.
5 P. Casavecchia, G. Capozza, E. Segoloni, and G. G. Volpi ͑unpublished͒.
6 T. Fueno, Y. Takahara, and K. Yamaguchi, Chem. Phys. Lett. 167, 291
͑1990͒.
37 H.-M. Keller, H. Floethmann, A. J. Dobbyn, R. Schinke, H.-J. Werner, C.
Bauer, and P. Rosmus, J. Chem. Phys. 105, 4983 ͑1996͒.
38 Y. J. Shiu and I.-C. Chen, Chem. Phys. Lett. 222, 245 ͑1994͒.
39 T. A. Cool and X.-M. Song, J. Chem. Phys. 96, 8675 ͑1992͒.
40 G. W. Adamson, Ph.D. thesis ͑MIT, Cambridge, 1994͒.
41 S.-H. Lee, I.-C. Chen, G. W. Adamson, and R. W. Field, J. Mol. Spectrosc.
182, 385 ͑1997͒.
7 D. Wortmann-Saleh, B. Engels, and S. D. Peyerimhoff, J. Phys. Chem. 98,
9541 ͑1994͒.
8 R. J. Cvetanovic, J. Chem. Phys. 23, 1375 ͑1955͒.
9 R. J. Cvetanovic, J. Phys. Chem. 74, 2730 ͑1970͒.
10 V. D. Knyazev, V. S. Arutyunov, and V. I. Vedeneev, Int. J. Chem. Kinet.
24, 545 ͑1992͒.
42 S.-H. Lee and I.-C. Chen, J. Chem. Phys. 103, 104 ͑1995͒.
43 Y. J. Shiu and I.-C. Chen, J. Mol. Spectrosc. 165, 457 ͑1994͒.
44 J. M. Brown, H. E. Radford, and T. J. Sears, J. Mol. Spectrosc. 148, 20
͑1991͒.
11 M. L. Morton, D. E. Szpunar, and L. J. Butler, J. Chem. Phys. 115, 204
͑2001͒.
12 K. Mahmud, P. Marshall, and A. Fontijn, J. Phys. Chem. 91, 1568 ͑1987͒.
13 D. L. Baulch, C. J. Cobos, R. A. Cox et al., J. Phys. Chem. Ref. Data 23,
847 ͑1994͒.
45 A. R. W. McKellar, J. B. Burkholder, J. J. Orlando, and C. J. Howard, J.
Mol. Spectrosc. 130, 445 ͑1988͒.
46 A. R. W. McKellar, Faraday Discuss. Chem. Soc. 71, 63 ͑1981͒.
47 A. Ben-Shaul, R. D. Levine, and R. B. Bernstein, J. Chem. Phys. 57, 5427
͑1972͒.
14 U. C. Sridharan and F. Kaufman, Chem. Phys. Lett. 102, 45 ͑1983͒.
15 W. M. Vaidya, Proc. R. Soc. London 147A, 513 ͑1934͒.
16 R. N. Dixon, Trans. Faraday Soc. 65, 3141 ͑1969͒.
48 A. Ben-Shaul, Y. Haas, K. L. Kompa, and R. D. Levine, Lasers and
Chemical Change ͑Springer, New York, 1981͒.
49 J. R. Barker, MultiWell-1.3.2 software, 2003, http://aoss.engin.umich.edu/
multiwell
17
´
L. Serrano-Andres, N. Forsberg, and P. Malmqvist, J. Chem. Phys. 108,
7202 ͑1998͒.
18 B. M. Stone, M. Noble, and E. K. C. Lee, Chem. Phys. Lett. 118, 83
͑1985͒.
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: 193.0.65.67
On: Wed, 10 Dec 2014 14:34:42