Absolute and relative rate constants for the reactions CH3C(O)O2 + NO and CH3C(O)O2 + NO2 and thermal stability of CH3C(O)O2NO2

Article abstract of DOI:10.1021/jp972881a

A pulse-radiolysis system was used to measure absolute rate constants for the reactions of CH₃C(O)O₂ radicals with NO and NO₂ at 295 K and 1000 mbar total pressure of SF₆. When the rate of formation and decay of NO₂ using its absorption at 400.5 and 452 nm were monitored, the rate constants k(CH₃C(O)O₂ + NO) = (2.0 ± 0.3) × 10^-11 and k(CH₃C(O)O₂ + NO₂) = (1.0 ± 0.2) × 10^-11 cm³ molecule^-1 s^-1 were determined. Long path-length Fourier transform infrared spectrometers were used to study the rate-constant ratio k(CH₃C-(O)O₂ + NO)/k(CH₃C(O)O₂ + NO₂) in 6-700 Torr total pressure of N₂ diluent at 243-295 K. At 295 K in 700 Torr of N₂ diluent k(CH₃C(O)O₂ + NO)/k(CH₃C(O)O₂ + NO₂) = 2.07 ± 0.21. The results are discussed in the context of the atmospheric chemistry of acetylperoxy radicals.

Full text of DOI:10.1021/jp972881a

J. Phys. Chem. A 1998, 102, 1779-1789
1779
Absolute and Relative Rate Constants for the Reactions CH3C(O)O2 + NO and CH3C(O)O2
+ NO2 and Thermal Stability of CH3C(O)O2NO2
Jens Sehested,*,† Lene Krogh Christensen, Trine Møgelberg,‡ and Ole J. Nielsen*,§
Atmospheric Chemistry, Plant Biology and Biogeochemistry Department, Risø National Laboratory,
DK-4000 Roskilde, Denmark
Timothy J. Wallington*,| and Andrei Guschin
Ford Research Laboratory, SRL-3083, Ford Motor Company, P.O. Box 2053, Dearborn, Michigan 48121-2053
John J. Orlando and Geoffrey S. Tyndall∇
Atmospheric Chemistry DiVision, National Center for Atmospheric Research, P.O. Box 3000,
Boulder, Colorado 80307
ReceiVed: September 4, 1997; In Final Form: December 2, 1997
A pulse-radiolysis system was used to measure absolute rate constants for the reactions of CH3C(O)O2 radicals
with NO and NO2 at 295 K and 1000 mbar total pressure of SF6. When the rate of formation and decay of
NO2 using its absorption at 400.5 and 452 nm were monitored, the rate constants k(CH3C(O)O2 + NO) )
(2.0 ( 0.3) × 10-11 and k(CH3C(O)O2 + NO2) ) (1.0 ( 0.2) × 10-11 cm3 molecule-1 s-1 were determined.
Long path-length Fourier transform infrared spectrometers were used to study the rate-constant ratio k(CH3C-
(O)O2 + NO)/k(CH3C(O)O2 + NO2) in 6-700 Torr total pressure of N2 diluent at 243-295 K. At 295 K
in 700 Torr of N2 diluent k(CH3C(O)O2 + NO)/k(CH3C(O)O2 + NO2) ) 2.07 ( 0.21. The results are discussed
in the context of the atmospheric chemistry of acetylperoxy radicals.
1. Introduction
O2),1 1.4 × 10-11 (R′O2 ) CH3O2, 298 K),1 1.0 × 10-11 (R′O2
) CH3O2, 298 K),4 and 1.0 × 10-11 cm3 molecule-1 s-1 (R′O2
) C2H5O2, 298 K).4 k3 was studied by Bridier et al.,5 Addison
et al.,6 and Basco and Parmar.7 The rate constants reported by
Addison et al. and Basco and Parmar are a factor of 2-3 lower
than reported by Bridier et al. As discussed elsewhere, the study
of Bridier et al.5 was more comprehensive than those of Addison
et al.6 and Basco and Parmer,7 and the data by Bridier et al. are
considered more reliable. Bridier et al.5 report k3 ) 9.6 × 10-12
cm3 molecule-1 s-1 at 298 K and 760 Torr total pressure of
N2.
Recently, two absolute studies of k2 have been reported.10,11
At room temperature Villalta and Howard11 measured k2 ) (2.0
( 0.3) × 10-11 while Maricq and Szente10 report k2 ) (1.4 (
0.2) × 10-11 cm3 molecule-1 s-1. The objective of the present
study is to improve our understanding of the reactions of
CH3C(O)O2 with NO and NO2. We have measured absolute
values of k2 and k3 using pulse radiolysis coupled with time-
resolved UV absorption spectroscopy employing two different
chemical systems. The FTIR spectrometer systems at Ford
Motor Company and the National Center for Atmospheric
Research (NCAR) were used to measure the rate constant ratio
k2/k3 and the thermal stability of CH3C(O)O2NO2.
Acetyl peroxy radicals, CH3C(O)O2, are formed during the
atmospheric degradation of oxygenated organic compounds such
as acetaldehyde, acetone, and methylglyoxal and react with HO2,
NO, NO2, or other peroxy radicals (R′O2):
CH3C(O)O2 + HO2 f products
(1)
(2)
CH3C(O)O2 + NO f CH3C(O)O + NO2
CH3C(O)O2 + NO2 + M h CH3C(O)O2NO2 + M (3,-3)
CH3C(O)O2 + R′O2 f products (4)
Reaction 3 forms CH3C(O)O2NO2 (peroxyacetyl nitrate or
PAN), which has sufficient stability toward thermal decomposi-
tion in the free troposphere to be transported over long distances
from urban high-NOx areas to remote low-NOx areas. The
efficiency of such long-range transport is determined by the
rate constants of reactions 1-4, the concentrations of HO2, NO,
NO2, and other peroxy radicals (R′O2), and the thermal stability
of PAN.
Moortgat et al.,1,2 Roehl et al.,3 and Villenave et al.4
determined k1 to be 4.3 × 10-13 exp(1040/T) cm3 molecule-1
s-1,2 and k4 to be 2.8 × 10-12 exp(530/T) (R′O2 ) CH3C(O)-
2. Experimental Section
The three different experimental systems used for this work
are described in detail elsewhere12-15 and are discussed briefly
here. The uncertainties reported in this paper are two standard
deviations unless otherwise stated. Standard error propagation
methods were used to calculate combined uncertainties.
2.1. Pulse Radiolysis System. CH3C(O)O2 radicals were
generated by radiolysis of either CH3CHO/O2/CO2/NOx or CH3-
* To whom correspondence should be addressed.
† E-mail: jens.sehested@risoe.dk.
‡ Present address: Danish Institute of Fundamental Metrology, Building
307, Anker Engelunds Vej 1, DK-2800 Lyngby, Denmark.
§ E-mail: ole-john.nielsen@risoe.dk.
| E-mail: twalling@ford.com.
E-mail: orlando@acd.ucar.edu.
∇ E-mail: tyndall@acd.ucar.edu.
S1089-5639(97)02881-8 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/12/1998

1780 J. Phys. Chem. A, Vol. 102, No. 10, 1998
Sehested et al.
CHO/O2/SF6/NOx gas mixtures in a 1-L stainless steel reactor
by a 30-ns pulse of 2-MeV electrons from a Febetron 705B
field-emission accelerator. The radiolysis dose, referred to
herein as a fraction of the maximum dose achievable, was varied
by insertion of stainless steel attenuators between the accelerator
and the chemical reactor. The analyzing light was obtained from
a pulsed Xenon arc lamp and reflected in the reaction cell by
internal White-type optics. The length of the cell is 10 cm,
and the optical path length for the analysis light was 120 cm.
The analyzing light was monitored by a photomultiplier attached
to a monochromator operated at a spectral resolution of 0.8 nm.
All transients were results of single-pulse experiments.
CO2 or SF6 was used as diluent gas. Radiolysis of CO2 and
SF6 produces oxygen and fluorine atoms, respectively:
-8 CO + O
(5)
(6)
2-MeV e
CO2
-8 F + products
2-MeV e
SF6
CO2 or SF6 was always present in excess to minimize the
relative importance of direct radiolysis of other compounds in
the gas mixtures.
The O-atom yield following radiolysis of CO2 was determined
using the absorbance of ozone at 254 nm following radiolysis
of mixtures of 50 mbar of O2 and 950 mbar of CO2.
O + O2 + M f O3 + M
(7)
The filled circles in Figure 1A show the absorbance after the
formation of ozone had ceased as a function of the radiolysis
dose. As seen from this figure, the absorbance is proportional
to the radiolysis dose. The slope is 0.441 ( 0.021. Using σ254
nm(O3) ) 1.15 × 10-17 cm2 molecule-1 16 we derive an O atom
yield of (7.75 ( 0.54) × 1014 cm-3 at full dose and 1000 mbar
of CO2. The error includes statistical uncertainty in the slope
of the data in Figure 1A and a 5% systematic uncertainty
associated with the ozone absorption cross section.
The fluorine atom yield was determined by two different
methods: first, by measuring the yield of CH3O2 radicals
following radiolysis of mixtures of 10 mbar CH4, 40 mbar O2,
and 950 mbar SF6,
Figure 1. (A) Maximum transient absorptions versus radiolysis dose.
Circles are data obtained using mixtures of 50 mbar of O2, 950 mbar
of CO2, and a montoring wavelength of 254 nm. Squares are data
obtained using mixtures of 10 mbar of CH4, 40 mbar of O2, 950 mbar
of SF6, and a montoring wavelength of 260 nm. Triangles are data
obtained using mixtures of 10 mbar of CH4, 40 mbar of O2, 950 mbar
of SF6, and a montoring wavelength of 240 nm. (B) Absorbance loss
at 452 nm (squares) and 400.5 nm (circles) as a function of dose
following radiolysis of mixtures of 0.5 mbar of NO2 and 1000 mbar of
SF6. An optical path length of 120 cm was used. The straight lines are
linear regressions through the low-dose data (filled symbols). See text
for details.
F + CH4 f CH3 + HF
(8)
(9)
CH3 + O2 + M f CH3O2 + M
σ
400.5nm(NO2) ) (6.5 ( 0.65) × 10-19 and σ452 nm(NO2) ) (4.5
and second, by measuring the loss of NO2 following radiolysis
of mixtures of 0.5 mbar of NO2 and 999.5 mbar of SF6,
( 0.45) × 10-19 cm2 molecule-1 16-19
,
the F-atom yield was
determined to be (3.25 ( 0.33) × 1015, (3.14 ( 0.32) × 1015,
(3.07 ( 0.33) × 1015, and (3.33 ( 0.34) × 1015 cm-3 at full
dose and 1000 mbar SF6. Quoted uncertainties reflect both
statistical uncertainties in the slopes from Figure 1 and a 10%
uncertainty in σ(CH3O2) and σ(NO2). It is gratifying to note
the good agreement among these four different determinations.
In the following we choose to use an average of the four
determinations with a 10% uncertainty, [F]0 ) (3.20 ( 0.32)
F + NO2 + M f FNO2 + M
(10)
The maximum transient absorbance at 240 and 260 nm
ascribed to CH3O2 radicals is plotted in Figure 1A, while the
changes in absorbance at 400.5 and 452 nm due to loss of NO2
via reaction 10 are shown in Figure 1B. As seen from Figure
1, the absorbance is proportional to the radiolysis dose up to
42% of maximum dose. At higher doses the absorbance falls
below that expected from extrapolation of the low-dose data.
We ascribe this to loss of radicals via unwanted radical-radical
reactions at high doses. Based on an optical path length of 120
cm, the slopes of the straight lines through the low-dose data
in Figure 1, 0.712 ( 0.014 (CH3O2, 240 nm); 0.494 ( 0.011
(CH3O2, 260 nm); -0.1040 ( 0.0044 (NO2, 400.5 nm);
-0.0780 ( 0.0015 (NO2, 452 nm); and σ240nm(CH3O2) ) (4.42
× 1015 cm-3
.
Reagents used were the following: 10-100 mbar CH3CHO
(>99.5%); 20-120 mbar O2 (ultrahigh purity); 850-900 mbar
CO2 (>99.9%); 870-1000 mbar SF6 (>99.9%); 0.29-0.71
mbar NO (>99.8%); 0.23-0.63 mbar NO2 (>98%). All gases
were used as received.
Four sets of experiments were performed using the pulse-
radiolysis system. First, mixtures of CH3CHO, O2, NO, and
CO2 were radiolyzed and the formation of NO2 was monitored
( 0.44) × 10-18,9 σ260nm(CH3O2) ) (3.18 ( 0.32) × 10-18 9
,

CH3C(O)O2 + NO and CH3C(O)O2 + NO2
J. Phys. Chem. A, Vol. 102, No. 10, 1998 1781
at 400.5 and 452 nm to provide measurements of k2. Second,
mixtures of CH3CHO, O2, NO2, and CO2 were radiolyzed and
the decay of NO2 was monitored at 400.5 and 452 nm to provide
measurements of k3. Third, k2 was determined from the rate of
formation of NO2 using radiolysis of mixtures of CH3CHO, O2,
NO, and SF6 and detection wavelengths of 400.5 and 452 nm.
Finally, k3 was determined from the rate of decay of NO2 upon
radiolysis of CH3CHO/O2/NO2/SF6 mixtures using detection
wavelengths of 400.5 and 452 nm.
2.2. FTIR Systems at Ford and NCAR. The experimental
systems at Ford14 and NCAR15 are described elsewhere and only
discussed briefly here. In both systems chemical analysis was
performed using FTIR spectroscopy. Experiments at Ford were
conducted at 293-308 K and 5.9-700 Torr total pressure of
N2/O2 diluent. Experiments at NCAR were performed over the
temperature range 240-295 K in 30-700 Torr total pressure
of N2/O2 diluent. The system at Ford Motor Company consists
of a 2-m-long, 140-L evacuable Pyrex chamber surrounded by
22 UV fluorescent lamps interfaced to a Mattson Instruments
Inc. Sirius 100 FTIR spectrometer. The system at NCAR
consists of a 47-L stainless steel reactor fitted with a quartz
window at one end to allow photolysis using a filtered xenon
arc lamp. A Bomem DA 3.01 FTIR spectrometer was interfaced
to a Hanst-type optical arrangement mounted within the reaction
cell. The path lengths for the analyzing infrared beam were 28
m (Ford) and 33 m (NCAR), the spectral resolutions were 0.25
cm-1 (Ford) and 0.1-0.5 cm-1 (NCAR), infrared spectra were
derived from 32 (Ford) and 100-200 (NCAR) coadded inter-
ferograms.
Figure 2. Absorption transients at 400.5 nm following pulsed radiolysis
of mixtures of 0.30-0.71 mbar NO, 50 mbar O2, 75-100 mbar CH3-
CHO, and 850-875 mbar CO2 (single pulses, full dose, and optical
path length of 120 cm). For clarity, the transients are separated vertically
by 0.01 units. The smooth solid lines are simulations using k2 ) 2 ×
10-11 cm3 molecule-1 s-1. Additional simulations of the transient where
[NO]0 ) 0.43 mbar using k2 ) 1.7 × 10-11 and 2.3 × 10-11 cm3
molecule-1 s-1 are shown with smooth dashed lines.
452 nm was observed to scale in a fashion consistent with the
absorption cross sections for NO2 at these two wavelengths,
σ400.5 nm(NO2) ) 6.5 × 10-19 and σ452 nm(NO2) ) 4.5 × 10-19
cm2 molecule-1 16-19
In addition, the formation rate increased
.
Initial concentrations (and purities) of the gas mixtures used
were 16-26 mTorr of CH3CHO (>99%), 49-460 mTorr of
chlorine (>99%), 29-100 mTorr of NOx (>99%), and 5.9-
700 Torr of O2/N2 (both >99.999%) diluent. Isotopically
labeled 13CH313CHO was used as the reactant in the experiments
at Ford. Nonlabeled acetaldehyde was used at NCAR. Products
were quantified by fitting reference spectra of the pure
compounds obtained at appropriate total pressures to the
observed product spectra using integrated absorption features.
The procedure was as follows. The CH3CHO, NO, and NO2
reactants were quantified and subtracted from the product spectra
using characteristic absorption features over the wavelength
regions 800-1500 cm-1. CO2 and CH3C(O)O2NO2 were then
identified and quantified using features at 2000-2400 and 800-
1500 cm-1, respectively. With the exception of CH3C(O)O2-
NO2, all reagent and reference compounds were obtained from
commercial sources. Reference spectra of CH3C(O)O2NO2 and
13CH313C(O)O2NO2 were obtained by irradiation of acetalde-
hyde/Cl2/NO2/O2 mixtures and equating the resulting peroxy-
acetyl nitrate features to the loss of acetaldehyde (corrected for
a small, <2%, formation of CO2 caused by the unavoidable
presence of a small amount of NO in the reaction system). At
1163 cm-1 σ(CH3C(O)O2NO2) ) 1.47 × 10-18 cm2 molecule-1
while at 1133 cm-1 σ(13CH313C(O)O2NO2) ) 1.21 × 10-18 cm2
with increasing NO concentration (see Figure 2). It seems
reasonable to assume that the observed absorption is caused by
NO2 formed from the following set of reactions:
2-MeV e
CO2
-8 CO + O
(5)
(11)
(12)
(13)
(2a)
O + CH3CHO f CH3CO + OH
OH + CH3CHO f CH3CO + H2O
CH3CO + O2 + M f CH3C(O)O2 + M
CH3C(O)O2 + NO f CH3C(O)O + NO2
CH3C(O)O2 + NO + M f CH3C(O)ONO2 + M (2b)
Consistent with previous work8 and as discussed in section 3.6,
we assume in the following that reaction 2b is unimportant and
that CH3C(O)O radicals decompose immediately after their
formation.
CH3C(O)O + M f CH3 + CO2 + M
(14)
The methyl radicals formed by reaction 14 add O2 to give
CH3O2 radicals that react with NO to produce another NO2
molecule. Therefore, a first-order expression does not describe
the NO2 formation adequately, and the experimental transients
need to be fitted using a chemical mechanism describing the
chemical reactions in the system. To model the experimental
absorption transients, the chemical mechanism shown in Table
1 and the absorption cross sections of NO2 at 400.5 and 452
nm were applied. In addition, the absorption at 400.5 nm by
CH3ONO formed by the addition of CH3O radicals to NO was
included (σ400.5 nm(CH3ONO) ) 1.7 × 10-19 cm2 molecule-1 21).
The chemical mechanism in Table 1 was constructed from the
molecule-1 20
.
3. Results
3.1. Absolute Rate Constant for the Reaction of CH3C-
(O)O2 Radicals with NO Using O Atom Initiation. The
formation of NO2 was studied by monitoring the absorbance at
400.5 and 452 nm following the radiolysis (full dose) of
mixtures of 0.30-0.71 mbar of NO, 50 mbar of O2, 75-100
mbar of CH3CHO, and 850-875 mbar of CO2. Figure 2 shows
five absorption transients obtained using a detection wavelength
of 400.5 nm. The maximum transient absorption at 400.5 and

1782 J. Phys. Chem. A, Vol. 102, No. 10, 1998
Sehested et al.
TABLE 1: Reaction Mechanism Used To Fit the Experimental Data
rate constant
reaction
(cm3 molecule-1 s-1
)
ref
O + CH3CHO f CH3CO + OH
4.5 × 10-13
3.1 × 10-14
1.6 × 10-12
9.7 × 10-12
1.0 × 10-10
0.4 × 10-10
1.9 × 10-13
5.1 × 10-12
1.5 × 10-11
1.6 × 10-11
4.8 × 10-12
1.1 × 10-11
3.2 × 10-12
2.6 × 10-11
1.6 × 10-11
2.0 × 10-11
1.66 × 10-11
2.0 × 10-11
1.0 × 10-11
5.5 × 10-12
5.5 × 10-12
1.07 × 10-12
1.04 × 10-11
1.6 × 10-11
7.7 × 10-12
7.5 × 10-12
2.5 × 10-11
1.7 × 10-11
1.8 × 10-14
2.6 × 10-11
1.3 × 10-12
2.4 × 10-15
2.6 × 10-13
2.5 × 10-11
1.9 × 10-11
4.8 × 10-11
8 × 10-12
16
25
16
16
28
28
40
41, 42
43
22
16
16
26
a
O + O2 + M f O3 + M
O + NO + M f NO2 + M
O + NO2 + M f 0.86NO + 0.86O2 + 0.14NO3 + M
F + CH3CHO f CH3CO + HF
F + CH3CHO f HC(O)CH2 + HF
F + O2 + M f FO2 + M
F + NO + M f FNO + M
F + NO2 + M f FNO2 + M
CH3CHO + OH f CH3CO + H2O
NO + OH + M f HONO + M
NO2 + OH + M f HNO3 + M
CH3CO + O2 + M f CH3C(O)O2 + M
CH3CO + NO + M f CH3C(O)NO + M
CH3CO + NO2 + M f CH3C(O)NO2 + M
CH3CO + CH3CO + M f CH3C(O)C(O)CH3 + M
CH3C(O)O2 + CH3C(O)O2 f 2CH3 + 2CO2 + O2
CH3C(O)O2 + NO f CH3 + CO2 + NO2
CH3C(O)O2 + NO2 + M f CH3C(O)O2NO2 + M
CH3C(O)O2 + CH3O2 f CH3 + CO2 + CH3O + O2
CH3C(O)O2 + CH3O2 f CH3C(O)OH + HCHO + O2
CH3 + O2 + M f CH3O2 + M
a
44
22
this work
this work
22
22
28
28
a
22
22
16
16
16
16
16
16
45
46
47
a
CH3 + NO + M f CH3NO + M
CH3 + NO2 f products
CH3O2 + NO f CH3O + NO2
CH3O2 + NO2 + M f CH3O2NO2 + M
CH3O + NO + M f CH3ONO + M
CH3O + NO2 + M f CH3ONO2 + M
O3 + NO f NO2 + O2
NO + NO3 f NO2 + NO2
NO2 + NO3 + M f N2O5 + M
CH3CHO + NO3 f HNO3 + CH3CO
HC(O)CH2 + O2 + M f HC(O)CH2O2 + M
HC(O)CH2 + NO + M f HC(O)CH2NO + M
HC(O)CH2 + NO2 + M f HC(O)CH2NO2 + M
HC(O)CH2 + HC(O)CH2 + M f HC(O)CH2CH2CHO + M
HC(O)CH2O2 + NO f HC(O)CH2O + NO2
HC(O)CH2O2 + NO2 + M f HC(O)CH2O2NO2 + M
2HC(O)CH2O2 f 1.5HC(O)CH2O + 0.25CHOCHO +
0.25HC(O)CH2OH + O2
a
a
a
6.4 × 10-12
8 × 10-12
HC(O)CH2O2 + CH3C(O)O2 f HC(O)CH2O + CH3 + CO2 + O2
HC(O)CH2O2 + CH3C(O)O2 f CHOCHO + CH3C(O)OH + O2
HC(O)CH2O2 + CH3O2 f HC(O)CH2O + CH3O + O2
HC(O)CH2O2 + CH3O2 f stable products
HC(O)CH2O + NO f stable products
HC(O)CH2O + NO2 f stable products
2.5 × 10-12
2.5 × 10-12
1.1 × 10-12
2.7 × 10-12
2.5 × 10-11
1.7 × 10-11
a
a
a
a
b
b
a Assumed equal to the analogous reactions for the acetonyl radical CH3C(O)CH2.48,49 b Assumed equal to k(CH3O + NO) and k(CH3O + NO2).
literature data where available. The rate constants not available
in the literature were obtained as follows. It was assumed that
the rate constants for the reactions of CH3CO radicals with NO
and NO2, the reaction of CH3 radicals with NO2, the self-
reactions of CH3CO, HC(O)CH2, and HC(O)CH2O2 radicals and
the cross reactions of HC(O)CH2O2 radicals with CH3C(O)O2
and CH3O2 radicals are equal to the rate constants for the
analogous reactions of CH3C(O)CH2 radicals. In addition, the
distribution between molecular and radical products from these
reactions was estimated from the product distribution of the
analogous CH3C(O)CH2 reaction. Finally, it was assumed that
the rate constants for the reactions of HC(O)CH2O radicals with
NO and NO2 are equal to the reactions of CH3O radicals with
NO and NO2.
Using this mechanism, we modeled the experimental tran-
sients in Figure 2 with k2 varied to give the best fit. As seen
from Figure 2, use of k2 ) 2.0 × 10-11 reproduces the
experimental data well. To illustrate the sensitivity of the
modeled transients to k2, we present simulated transients
obtained using k2 ) 1.7 × 10-11 and 2.3 × 10-11 cm3
molecule-1 s-1 (for [NO] ) 0.43 mbar) in Figure 2. Use of k2
) 1.7 × 10-11 and 2.3 × 10-11 gives transients in which the
formation of NO2 is too slow or too fast.
In addition to reaction 2 there are several other reactions that
can potentially influence the kinetics of NO2 formation that we
need to consider. First, the rate of formation of CH3C(O)O2
radicals can obviously impact the observed rate of NO2
formation. Under our experimental conditions the pseudo-first-
order rates of formation of CH3CO radicals (from reaction 11),
CH3C(O)O2 radicals (from reaction 13), and NO2 (via reaction
2) are (0.83-1.1) × 106 s-1, 3.9 × 106 s-1, and (1.5-3.5) ×
105 s-1. Clearly, the formation of NO2 is not sensitive to the
rate constant for the addition reaction of O2 to the CH3CO
radical. However, it is not possible to avoid some dependency
of the derived rate constant k2 on the value used for the O +
CH3CHO reaction. For this reason separate experiments were
performed in which the reaction of F atoms with CH3CHO was
used as a source of CH3CO radicals. As discussed in section
3.3, the value of k2 determined in these separate experiments
was indistinguishable from that presented in the present section.

CH3C(O)O2 + NO and CH3C(O)O2 + NO2
J. Phys. Chem. A, Vol. 102, No. 10, 1998 1783
Second, the formation of NO2 is also dependent on the rate
constant for the reaction of CH3O2 with NO. Fortunately, the
rate constant for this reaction is well established.16,22-24 It
mainly influences the formation rate of the last part of the NO2
transient while k2 determines the formation rate of NO2 just
after the electron pulse. Consequently, the CH3C(O)O2 + NO
and CH3O2 + NO reactions are partially separated in time.
Third, CH3O radicals are formed in the reaction of CH3O2
with NO. Under the present experimental conditions the fate
of CH3O radicals is addition to either NO or NO2 to give CH3-
ONO or CH3ONO2. CH3ONO has a small, but significant,
absorption at the wavelengths used to monitor NO2 formation.
CH3ONO2 does not absorb in the region 400-450 nm.21 The
absorption associated with the formation of CH3ONO was
included in the model using an absorption cross section of 1.7
× 10-19 cm2 molecule-1 21 at 400.5 nm. At 452 nm the
absorption due to CH3ONO is negligible.21 Finally, the value
of the rate constant for the self-reaction of CH3C(O)O2 radicals,
1.6 × 10-11 cm3 molecule-1 s-1, also affects the measurements.
The half-life for the self-reaction between CH3C(O)O2 radicals
is t1/2 ) 1/(2k[CH3C(O)O2]) ) 52 µs (calculated using an
average concentration of CH3C(O)O2 radicals [CH3C(O)O2] ≈
0.6 × 1015 cm-3). The half-life for the reaction of CH3C(O)-
O2 radicals with NO is 2-5 µs, at least an order of magnitude
faster than the self-reaction. Therefore, the influence of the self-
reaction of CH3C(O)O2 is small. A simulation using [NO] )
0.43 mbar and the kinetic model in Table 1 shows that 7% of
the CH3C(O)O2 radicals are removed by the self-reaction.
Figure 3. Absorption transients obtained at 400.5 nm following pulsed
radiolysis of mixtures of 0.235-0.58 mbar NO2, 50 mbar O2, 100 mbar
CH3CHO, and 850 mbar CO2 (single pulses, full dose, and optical path
length of 120 cm). The smooth solid lines are simulations using k3 )
1 × 10-11 cm3 molecule-1 s-1. Additional simulations of the transient
where [NO2]0 ) 0.305 mbar using k3 ) 0.8 × 10-11 and 1.2 × 10-11
cm3 molecule-1 s-1 are shown with smooth dashed lines.
× 10-12 cm3 molecule-1 s-1 fall well outside the noise level of
the experimental absorption transients; hence, we report k )
(10 ( 2) × 10-12 cm3 molecule-1 s-1
.
Several reactions could potentially interfere with the above
determination of k3. First, the O atoms may react with species
other than CH3CHO. O atoms generated by the radiolysis pulse
can react via the following reactions:
In summary, although the mechanism used to model the
experimental data (see Table 1) is complex, reaction 2 is the
key reaction that determines the rate of NO2 formation. Other
reactions play a minor role, and the influence of these reactions
is explicitly taken into account. As seen in Figure 2, using k2
) 2.0 × 10-11 cm3 molecule-1 s-1 gives a good fit to
experimental transients obtained using a range of NO concentra-
tions. Also shown in Figure 2 are numerical simulations using
k2 ) 1.7 × 10-11 and 2.3 × 10-11 cm3 molecule-1 s-1, which
are well outside the error limits of the experiment. In light of
the number of experimental absorption transients (nine) that are
well fitted using k2 ) 2.0 × 10-11 cm3 molecule-1 s-1 and the
relatively well-established kinetic mechanism used for the
simulations, we choose to quote a final value of k2 ) (2.0 (
O + CH3CHO f CH3CO + OH
O + O2 + M f O3 + M
O + NO2 f NO + O2
(11)
(15)
(16a)
(16b)
O + NO2 + M f NO3 + M
For our experimental conditions with [O2] ) 50 mbar, [NO2]
) 0.235-0.58 mbar, and [CH3CHO] ) 100 mbar and using
0.3) × 10-11 cm3 molecule-1 s-1
.
rate constants of k11 ) 4.5 × 10-13 16
,
k15 ) 3.1 × 10-14 25
, k16a
3.2. Absolute Rate Constant for the Reaction of CH3C-
(O)O2 Radicals with NO2 Using O-Atom Initiation. The
decay of NO2 was studied by monitoring the absorbance at 400.5
and 452 nm following the radiolysis (full dose) of 0.235-0.58
mbar NO2, 50 mbar of O2, 100 mbar of CH3CHO, and 850
mbar CO2. Figure 3 shows typical absorption transients
obtained using 400.5 nm as the detection wavelength. The
decay rate increased with increasing NO2 concentration. The
absorption loss at 400.5 and 452 nm was observed to scale in
a manner consistent with the absorption cross sections of NO2
at these wavelengths. It seems reasonable to ascribe the
observed loss of absorbance at 400.5 and 452 nm to loss of
NO2.
16
) 9.7 × 10-12 16
,
and k16b ) 1.6 × 10-12 cm3 molecule-1 s-1
,
we calculate that >85% of the O atoms react to give CH3CO
and OH radicals, <3% of the O atoms are converted into O3,
<1.7% of the O atoms are converted into NO3, and <10.6% of
the O atoms are converted into NO. The small amounts of
ozone, NO, and NO3 produced by reactions 15 and 16 are not
expected to complicate the study of reaction 3.
Second, CH3CO radicals formed via reactions 11 or 12 can
react with either O2 or NO2:
CH3CO + O2 + M f CH3C(O)O2 + M
CH3CO + NO2 + M f CH3C(O)NO2 + M
(13)
(17)
The chemical mechanism in Table 1 was used to model the
experimental transients in Figure 3 with k3 varied to give the
best fit. As seen from Figure 3 the absorption transients were
well fitted using k3 ) 1 × 10-11 cm3 molecule-1 s-1. To
illustrate the sensitivity of the modeled transients to k3, we
present simulated transients obtained using k3 ) 8 × 10-12 and
12 × 10-12 cm3 molecule-1 s-1 (for [NO2] ) 0.305 mbar) in
Figure 3. The modeled transients using k3 ) 8 × 10-12 and 12
The rate constant for reaction 13 is 3.2 × 10-12 cm3 molecule-1
s-1 26
There are no available data for reaction 17 at higher
.
pressures. Assuming that the rate constant for reaction 17 is
equal to that for the reaction between CH3COCH2 radicals and
NO2 of 2.6 × 10-11 cm3 molecule-1 s-1, we calculate that
<9.1% of the CH3CO radicals react via reaction 17 and that
>90.9% of the CH3CO radicals form CH3C(O)O2. Since

1784 J. Phys. Chem. A, Vol. 102, No. 10, 1998
Sehested et al.
reaction 17 only consumes a small fraction of the CH3CO
radicals and since CH3C(O)NO2 is a closed-shell molecule, the
formation of this species is not expected to interfere with this
study.
Third, the formation rate of CH3C(O)O2 may influence the
decay rate of NO2. The pseudo-first-order rate constants for
formation of CH3CO and CH3C(O)O2 and for loss of NO2 (from
CH3C(O)O2 + NO2) are 1.1 × 106, 3.9 × 106, and (0.58-1.4)
× 105 s-1. Hence, the NO2-loss rate is mainly dependent on k3
with only minor influences from k12 and k15.
Fourth, some CH3C(O)O2 radicals will undergo self-reaction
leading to formation of CH3 radicals, CO2, and O2. The half-
life for the self-reaction between CH3C(O)O2 radicals is t1/2
)
1/(2k[CH3C(O)O2]) ) 52 µs using an average concentration of
CH3C(O)O2 radicals, [CH3C(O)O2] ≈ 0.6 × 1015 cm-3. The
half-life for the reaction of CH3C(O)O2 radicals with NO2 is
5-12 µs or 5-10 times faster than the self-reaction of CH3C-
(O)O2 radicals. The CH3 radicals formed by the self-reaction
will either react with O2 to give CH3O2 radicals (which will
react with NO2) or react directly with NO2. Both reactions lead
to loss of another NO2. The impact of the self-reaction of
CH3C(O)O2 radicals is a slower overall decay of NO2 since the
loss of NO2 is delayed. Self-reaction is the fate of up to 25%
of the CH3C(O)O2 radicals. As seen in Figure 3, the chemical
mechanism in Table 1 with k3 ) 1.0 × 10-11 provides a good
fit for transients observed in the presence of 0.235-0.58 mbar
of NO2. The importance of the self-reaction as a loss of CH3C-
(O)O2 radicals varies by a factor of 2.5 over this range of [NO2],
while the quality of the fits is unchanged, suggesting that the
mechanism used here provides an adequate account of the
CH3C(O)O2 radical self-reaction. We choose to report k3 )
(1.0 ( 0.2) × 10-11 cm3 molecule-1 s-1 as determined above.
3.3. Absolute Rate Constant for the Reaction of CH3C(O)-
O2 Radicals with NO Using F-Atom Initiation. The formation
of NO2 was studied by monitoring the absorbance at 400.5 and
452 nm following the radiolysis (53% of full dose) of gas
mixtures of 0.29-0.61 mbar of NO, 50 mbar of O2, 10 mbar
of CH3CHO, and 940 mbar of SF6. Figure 4 shows four
absorption transients obtained at 400.5 nm. The maximum
transient absorption at 400.5 and 452 nm was observed to scale
in a fashion consistent with the absorption cross sections for
NO2 at these two wavelengths, and the formation rate increased
with increasing NO concentration. We attribute the observed
increase in absorption to NO2 formed by the reaction of peroxy
radicals with NO. However, three peroxy radicals could be
formed in the system. CH3C(O)O2 radicals are formed by the
following reactions:
Figure 4. Absorption transients at 400.5 nm following pulsed radiolysis
of mixtures of 0.29-0.61 mbar NO, 50 mbar O2, 10 mbar CH3CHO,
and 940 mbar SF6 (single pulses, 53% of full dose, and optical path
length of 120 cm). For clarity, the transients are separated vertically
by 0.01 units. The smooth solid lines are simulations using k2 ) 2 ×
10-11 cm3 molecule-1 s-1. Additional simulations of the transient where
[NO]0 ) 0.50 mbar with k(HC(O)CH2O2 + NO) varied by a factor of
2 (short-dashed lines, hardly distinguishable from the solid and the long-
dashed lines) and k2 ) 1.6 × 10-11 and 2.4 × 10-11 cm3 molecule-1
s-1 are shown with smooth long-dashed lines.
In addition to these two peroxy radicals, CH3O2 radicals are
formed via the following set of reactions:
CH3C(O)O2 + NO f CH3C(O)O + NO2
CH3C(O)O + M f CH3 + CO2 + M
CH3 + O2 + M f CH3O2 + M
(2a)
(11)
(9)
From the rate constants in Table 1 and neglecting radical-
radical chemistry, we calculate a yield of CH3O2 radicals of
59-65% relative to the initial F-atom yield.
We have a mixture of three different peroxy radicals that can
react with NO to produce NO2. It is not possible to derive k2
from a fit of an analytical expression to the experimental
absorption transient; instead, it is necessary to model the
absorption transient. The rate constant for the reaction of CH3O2
with NO is well-known and can be well represented in the
model. However, the rate constant for the reaction of HC(O)-
CH2O2 radicals with NO is not known and the fate of the
subsequently formed radical, HC(O)CH2O, is unknown. In the
following we have assumed that k(HC(O)CH2O2 + NO) )
F + CH3CHO f CH3CO + HF
(18)
(13)
k(CH3C(O)CH2O2 + NO) ) 8 × 10-12 cm3 molecule-1 s-1 27
.
CH3CO + O2 + M f CH3C(O)O2 + M
It was also assumed that HC(O)CH2O radicals are the only
product of the reaction between HC(O)CH2O2 radicals and NO.
The fate of HC(O)CH2O radicals in the system has been shown
to be reaction with NO to give HC(O)CH2ONO.28 The
experimental transients shown in Figure 4 were modeled using
the mechanism given in Table 1 with k(HC(O)CH2O2 + NO)
) 8 × 10-12 cm3 molecule-1 s-1. The model reproduced the
experimental absorption transients well. As a check of the
sensitivity of the model to k(HC(O)CH2O2 + NO), we modeled
an absorption transient with [NO]0 ) 0.50 mbar, changing k(HC-
(O)CH2O2 + NO), up and down by a factor of 2. The result is
shown in Figure 4. The modeled curves are shown as smooth,
short-dashed lines that can hardly be distinguished from the solid
and long-dashed lines. We conclude that the result is not very
From the rate constants in Table 1 and correcting the yield of
CH3C(O)O2 radicals for small losses of CH3CO radicals and F
atoms due to reaction with NO, we calculate that 64-68% of
the initially formed F atoms are converted into CH3C(O)O2
radicals. HC(O)CH2O2 radicals are formed by the reactions
F + CH3CHO f HC(O)CH2 + HF
(19)
HC(O)CH2 + O2 + M f HC(O)CH2O2 + M (20)
Using the rate constants in Table 1, we calculate that 13-18%
of the F atoms are converted to HC(O)CH2O2 radicals.

CH3C(O)O2 + NO and CH3C(O)O2 + NO2
J. Phys. Chem. A, Vol. 102, No. 10, 1998 1785
Figure 5. Absorption transients obtained at 400.5 nm following pulsed
radiolysis of mixtures of 0.285-0.555 mbar NO2, 50 mbar O2, 10 mbar
CH3CHO, and 940 mbar SF6 (single pulses, 53% of full dose, and
optical path length of 120 cm). The smooth solid lines are simulations
using k3 ) 1 × 10-11 cm3 molecule-1 s-1. Additional simulations of
the transient where [NO2]0 ) 0.397 mbar using k3 ) 0.8 × 10-11 and
1.2 × 10-11 cm3 molecule-1 s-1 are shown with smooth dashed lines.
Figure 6. Arrhenius plot for k-3 at 700 (b) Torr total pressure of N2
diluent. The solid line is a linear least-squares fit. Previous data for
k-3 reported by Tuazon et al.30 (at 740 Torr), Bridier et al.5 (at 600
Torr), Roberts and Bertman31 (at 760 Torr), Roumelis and Glavas32 (at
760 Torr), and Grosjean et al.33 (at 750 Torr) are indicated by the
squares, triangles, diamonds, inverted triangles, and dotted line,
respectively. The insert shows the observed decay of CH3C(O)O2NO2
at 293.0 (triangles), 300.6 (circles), and 308.3 K (diamonds) in 700
Torr of N2 diluent.
sensitive to k(HC(O)CH2O2 + NO). We estimate the uncer-
tainty in k(CH3C(O)O2 + NO) due to uncertainty in k(HC(O)-
CH2O2 + NO) to be 0.2 × 10-11 cm3 molecule-1 s-1
.
Finally, as a check of the sensitivity of the model to k(CH3C-
(O)O2 + NO), we modeled the absorption transient in Figure 4
with [NO]0 ) 0.50 mbar using k(CH3C(O)O2 + NO) ) 1.6 ×
10-11 and 2.4 × 10-11 cm3 molecule-1 s-1
. The derived
transients are shown in Figure 4 with long-dashed lines. As
seen from Figure 4, the fits are sensitive to the value of k(CH3C-
(O)O2 + NO). Eight experiments were well described by the
model using k(CH3C(O)O2 + NO) ) 2.0 × 10-11 cm3
molecule-1 s-1. From the F-atom-initiated experiments, we
derive k(CH3C(O)O2 + NO) ) (2.0 ( 0.5) × 10-11 cm3
molecule-1 s-1; the quoted errors reflect both statistical
uncertainties and our estimate of possible systematic uncertain-
ties associated with k(HC(O)CH2O2 + NO). This result is
consistent with, although less precise than, that reported in
section 3.1.
and 15-20% of the initial F-atom yield. The remaining F atoms
are consumed by reaction with NO2. Hence, 77-82% of the
initially formed peroxy radicals are CH3C(O)O2 radicals. To
simulate the observed absorption transient, we had to assume
that k(HC(O)CH2O2 + NO2) is 6.4 × 10-12 cm3 molecule-1
s-1 by analogy to k(CH3C(O)CH2O2 + NO2).27
An additional complication is that the self-reaction of the
CH3C(O)O2 radicals is relatively fast. Therefore, a nonnegli-
gible fraction of the CH3C(O)O2 radicals undergoes self-reaction
to give CH3C(O)O radicals that decompose to CH3 radicals and
CO2. The half-life for the self-reaction between CH3C(O)O2
radicals is t1/2 ) 1/(2kCaverage) ) 52 µs using an average
3.4. Absolute Rate Constant for the Reaction of CH3C-
(O)O2 Radicals with NO2 Using F-Atom Initiation. The
decay of NO2 was studied by monitoring the absorbance at 400.5
and 452 nm following the radiolysis (53% of full dose) of
mixtures of 0.285-0.63 mbar of NO2, 50 mbar of O2, 10 mbar
of CH3CHO, and 940 mbar of SF6. Figure 5 shows typical
absorption transients obtained at 400.5 nm. The decay rate
increased with increasing NO2 concentration. The absorption
loss at 400.5 and 452 nm scaled with the absorption cross
sections of NO2 at these wavelengths. We ascribe the observed
loss of absorbance to loss of NO2 via reaction 3. The transients
in Figure 5 were modeled using the mechanism given in Table
1. The best fits were achieved using k3 ) 1.0 × 10-11 cm3
molecule-1 s-1 and are shown as the smooth curves.
Before the NO2 loss rate can be used to determine the rate
constant for reaction 3 we need to consider possible interfering
reactions. As discussed in section 3.3, three different peroxy
radicals are formed in the system. CH3C(O)O2 and HC(O)-
CH2O2 are produced following reaction of F atoms with CH3-
CHO. CH3O2 radicals are formed following the self-reaction
of CH3C(O)O2 radicals as discussed below. We estimate the
yields of CH3C(O)O2 and HC(O)CH2O2 radicals to be 66-69%
concentration of CH3C(O)O2 radicals Caverage ≈ 0.6 × 1015 cm-3
.
The half-life for the reaction of CH3C(O)O2 radicals with NO2
is 5-10 µs or 5-10 times faster than the self-reaction of the
CH3C(O)O2 radical. The CH3 radicals formed by the self-
reaction will react with NO2 either directly or indirectly (via
formation of CH3O2). Both reactions lead to loss of NO2. The
effect of the self-reaction of CH3C(O)O2 radicals is a slower
overall decay of NO2, since the loss of NO2 is delayed.
However, since the model can fit all absorption transients with
different initial NO2 concentrations and since it is possible to
determine the fate of the products of the self-reaction of CH3C-
(O)O2 radicals, this reaction is a complication that can be dealt
with.
Finally, we checked the sensitivity of the model to k3. Figure
5 shows two dashed lines that are simulations of the experiment
with [NO]0 ) 0.397 mbar using k3 ) 8 × 10-12 and 12 × 10-12
cm3 molecule-1 s-1. Clearly, these two simulations fall well
outside the noise level of the transient. We choose to report k3
) (1.0 ( 0.2) × 10-11 cm3 molecule-1 s-1. This result is
consistent with that reported in section 3.2.

1786 J. Phys. Chem. A, Vol. 102, No. 10, 1998
Sehested et al.
a
TABLE 2: Measured Values for k-3
temp
(K)
total pressure
(Torr) (nitrogen diluent)
k-3
(10-3 s-1
)
308.3
306.8
300.6
296.4
293.0
293.1
293.1
700
700
700
700
700
100
30
1.95 ( 0.14
1.38 ( 0.09
0.557 ( 0.031
0.371 ( 0.058
0.188 ( 0.010
0.162 ( 0.015
0.141 ( 0.014
a Quoted errors are two standard deviations.
3.5. Thermal Stability of CH3C(O)O2NO2. As a prelimi-
nary exercise prior to measurement of k2/k3, control experiments
were performed to measure the thermal stability of CH3C(O)O2-
NO2. To check for heterogeneous loss of CH3C(O)O2NO2 in
the chamber at Ford, a mixture of CH3C(O)O2NO2 in the
presence of a large excess of NO2 was left in the chamber in
the dark for 64 min at 308 K. There was no loss (<2%) of
CH3C(O)O2NO2, suggesting the absence of complications
caused by heterogeneous loss processes (PAN is known to be
very stable in the NCAR chamber29). Addition of NO to
reaction mixtures containing CH3C(O)O2NO2 led to a decay of
this species. In the presence of excess NO, the loss of
CH3C(O)O2NO2 was not dependent on the NO concentration.
The loss of CH3C(O)O2NO2 in the presence of NO is not caused
by a reaction of these two compounds. Instead, NO scavenges
CH3C(O)O2 radicals formed in the thermal decomposition of
CH3C(O)O2NO2, thereby limiting the re-formation of CH3C(O)O2-
NO2 via reaction 3:
Figure 7. Plot of ∆[CO2]corr/∆[PAN]corr versus [NO]av/[NO2]av at 700
Torr total pressure and 295 K. See text for details.
k-3. Reaction -3 is a unimolecular decomposition, and a
decrease in k-3 with decreasing total pressure is to be expected.
At 293 K the IUPAC data-evaluation panel recommends that
when compared with the value at 700 Torr, the rate of PAN
decomposition decreases by 12% and 35% at 100 and 30 Torr,
respectively.22 The dependence of k-3 on total pressure
observed in the present work is indistinguishable from that
recommended by the IUPAC panel. As expected, the observed
pressure dependence of k-3 is indistinguishable from that of k3
(see Figure 8).
3.6. Measurement of k2/k3 at Ford and NCAR. The rate
constant ratio k2/k3 was measured using the FTIR systems at
Ford and NCAR in experiments employing the UV irradiation
of acetaldehyde/Cl2/NO/NO2/O2/N2 mixtures.
CH3C(O)O2NO2 + M f CH3C(O)O2 + NO2 + M (-3)
CH3C(O)O2 + NO2 + M f CH3C(O)O2NO2 + M (3)
Cl2 + hν (λ > 300 nm) f 2Cl
Cl + CH3CHO f CH3C(O) + HCl
CH3C(O) + O2 + M f CH3C(O)O2 + M
(21)
(22)
(13)
CH3C(O)O2 + NO f CH3C(O)O + NO2
(2)
The decay of CH3C(O)O2NO2 followed first-order kinetics.
Representative data are shown in the insert in Figure 6. Linear
least-squares analysis of these data gives pseudo-first-order rate
constants for the thermal decomposition of CH3C(O)O2NO2.
CH3C(O)O2NO2 can be regenerated via reaction 3. The
observed rate of CH3C(O)O2NO2 decay kobs is related to the
true rate of decay k-3 by the expression
Acetylperoxy radicals are formed via the reactions given above
and then react with either NO or NO2. As discussed above,
reaction with NO leads to the formation of CO2 while reaction
with NO2 produces peroxyacetyl nitrate (PAN). The relative
importance of reactions 2 and 3 as loss mechanisms for CH3C-
(O)O2 radicals depends on the concentration ratio [NO]/[NO2]
and the rate constant ratio k2/k3. Provided that reactions 2 and
3 are the sole sources of CO2 and PAN and that there are no
losses of these species, then
k-3 ) kobs(1 + (k3[NO2]/k2[NO]))
The observed CH3C(O)O2NO2 decay rate kobs was corrected for
regeneration via reaction 3 using k3 ) 1.0 × 10-11 and k2 )
2.0 × 10-11 cm3 molecule-1 s-1 (see previous sections). The
concentrations of NO and NO2 in the chamber were monitored
using their characteristic IR absorptions. Corrections were in
the range 2-10% and have been applied to the data given in
Table 2 and the Arrhenius plot in Figure 6.
Experiments were performed at a variety of temperatures over
the range 293-308 K. Results are listed in Table 2 and plotted
in Figure 6 together with previous measurements of k-3 by
Tuazon et al.30 at 740 Torr, Bridier et al.5 at 600 Torr, Roberts
and Bertman31 at 760 Torr, Roumelis and Glavas32 at 760 Torr,
and Grosjean et al.33 at 750 Torr. As seen from Figure 6, the
results from the present work are in excellent agreement with
all previous measurements except those of Grosjean et al.33
which lie 20-30% below the rest of the data. To test for the
effect of total pressure, experiments were performed at 293 K
with 30, 100, or 700 Torr total pressure. As seen from Table
2, there was a small, but significant, effect of total pressure on
[CO2]
k2
[NO]
)
×
( )
(
)
k3
[PAN]
[NO2]
where [CO2], [PAN], [NO], and [NO2] are the concentrations
of CO2, PAN, NO, and NO2 in the chamber and k2 and k3 are
the rate constants for reactions 2 and 3.
The first set of experiments was performed using the
experimental system at Ford in 700 Torr of N2 at 296 K using
mixtures of 16-23 mTorr of 13CH313CHO, 60 mTorr of Cl2,
9-91 mTorr of NO, 22-56 mTorr of NO2, and 200 Torr of
O2. Reaction mixtures were irradiated for 2-40 s, resulting in
the loss of 2-24% of the initial 13CH313CHO and changes in
the NO and NO2 concentrations of 1-10%. In all cases the
combined molar yields of CO2 and PAN were indistinguishable
from the observed acetaldehyde loss. Figure 7 shows a plot of

CH3C(O)O2 + NO and CH3C(O)O2 + NO2
J. Phys. Chem. A, Vol. 102, No. 10, 1998 1787
TABLE 3: Measured Values for k2/k3
total pressure
temp (K)
(Torr) (nitrogen diluent)
k2/k3
295 (Ford data)
295 (Ford data)
295 (Ford data)
295 (Ford data)
295 (Ford data)
295 (Ford data)
283 (NCAR data)
283 (NCAR data)
243 (NCAR data)
243 (NCAR data)
700
300
100
50
2.07 ( 0.21a
2.38b
2.70b
3.36b
25
3.12b
5.9
700
30
700
100
5.46b
2.2 ( 0.2
2.9 ( 0.3
2.1 ( 0.2
2.4 ( 0.2
a Obtained from linear least-squares analysis of the data in Figure
7. b Single point determinations.
[CO2]/[PAN] versus [NO]av/[NO2]av for experiments conducted
at 700 Torr and 295 K, where [NO]av and [NO2]av are the
average concentrations of NO and NO2 during the experiment.
As expected, there is a linear relationship between [CO2]/[PAN]
and [NO]av/[NO2]av. Linear least-squares analysis of the data
in Figure 7 gives k2/k3 ) 2.17 ( 0.15.
Figure 8. Pressure dependence for k3 at 295 K measured in the present
work (circles) and by Bridier et al.5 (triangles).
of Bridier et al., and a nonlinear least-squares fit to the combined
data set returns the same values of the parameters in the Troe
expression: ko ) 2.7 × 10-28 and kinf ) 1.2 × 10-11 for Fc )
0.3 and Nc ) 1.41.
Derivation of the rate-constant ratio k2/k3 relies on the
assumption that CO2 and PAN are formed solely as a result of
reactions 2 and 3 and are not lost via any process. Although it
is difficult to imagine other sources of PAN in the chamber, it
is not difficult to imagine other sources of CO2. A control
experiment was performed in which a mixture of 24 mTorr of
13CH313CHO, 77 mTorr of Cl2, 35 mTorr of NO, and 84 Torr
of O2 in 700 Torr total pressure of N2 diluent was subjected to
four successive 10-s periods of UV irradiation, leading to 21%
consumption of the acetaldehyde. The molar yield of 13CO2
(relative to loss of acetaldehyde) was 95 ( 8%; PAN was just
detectable with a yield of 1-2%. The fact that the 13CO2 yield
was not significantly greater than 100% shows that there are
no confounding sources of CO2 in this chemical system.
Although there are no possible loss mechanisms for CO2 in the
reactor there are several possible loss processes for PAN that
need consideration. PAN can be lost via thermal decomposition
(see previous section), heterogeneous reactions, and reaction
with Cl atoms or OH radicals (OH radicals are formed in the
system via reaction of HO2 radicals with NO). As discussed
in the previous section, there was no evidence for heterogeneous
loss of PAN in the chamber systems. PAN reacts very slowly
with both Cl atoms and OH radicals,34 and such reactions will
not be significant in the present work. Finally, the yields of
CO2 and PAN can be corrected for the thermal decomposition
of PAN during the 1-5 min taken for UV irradiation and data
acquisition using the value of k-3 ) 3.9 × 10-4 s-1 reported in
the previous section. Corrections were in the range 1-6% and
have been applied to the data in Figure 7. Linear least-squares
analysis of the data in Figure 7 gives k2/k3 ) 2.07 ( 0.21 at
700 Torr and 295 K. Additional experiments were performed
at reduced total pressures at 295 K. The results of these single-
point determinations are given in Table 3. On the basis of the
data scatter evident in Figure 7, we estimate that the single-
point determinations have an associated uncertainty of (15%.
The increase in k2/k3 with decreasing pressure reflects a
decrease in the association rate constant k3 with decreasing
pressure. There have been three studies of the effect of pressure
on k3.5-7 The most comprehensive study is that of Bridier et
al.5 The current data have been converted to absolute values
for k3 by use of k2 ) 2.0 × 10-11 cm3 molecule-1 s-1 measured
here and by Villalta and Howard.11 The data are shown in
Figure 8 along with those of Bridier et al., and the fitted line to
their data. The current data are indistinguishable from those
Analogous experiments were performed at NCAR at 243 and
283 K for pressures between 30 and 700 Torr. For the initial
experiments a resolution of 0.5 cm-1 was used. At the lower
pressures the dependences of the absorption due to CO2 and
NO2 became very nonlinear because of underresolution and
saturation. The resolution was increased to 0.1 cm-1, and
concentrations were determined by the use of integrated line
intensities rather than peak heights. This led to an improvement
in the accuracy of measurements of these species. At the lowest
temperature and pressure, the problem of CO2 measurement was
avoided by the measurement of CH3ONO2 instead of CO2. At
the low temperatures and low O2 partial pressure, methyl nitrate
is formed following the reaction of peroxyacetyl radicals with
NO:
CH3C(O)O2 + NO f CH3 + CO2 + NO2
CH3 + O2 + M f CH3O2 + M
(2)
(9)
CH3O2 + NO f CH3O + NO2
(23)
(24)
CH3O + NO2 + M f CH3ONO2 + M
The spectral features of methyl nitrate are smooth, and its
concentration varied linearly with that of PAN over five or six
irradiations.
The results are included in Table 3. When combined with
the temperature dependence for the rate constant k2 measured
by Villalta and Howard,11 the results at lower temperatures agree
to within 5% with those predicted by the falloff parameters given
by Bridier et al.,5 further confirming the pressure dependence
reported there. Because of the strong negative temperature
dependence of k3 at low pressure, measurements at low
temperature do not extend very far into the falloff region, and
so the paramaters for k0 could not be improved.
4. Discussion and Conclusion
Recently, two groups have reported direct kinetic studies of
the reaction of CH3C(O)O2 radicals with NO.10,11 At room
temperature (295-298 K) Villalta and Howard,11 and Maricq
and Szente10 obtain values for k2 of (2.0 ( 0.3) × 10-11 and

1788 J. Phys. Chem. A, Vol. 102, No. 10, 1998
Sehested et al.
Figure 9 is drawn at k2/k3 ) 2.17 which is the average of the
data reported by Cox et al.,35 Kirchner et al.,37 Tuazon et al.,30
Seefeld et al.,39 and the present study. The two dashed lines
represent changes in k2/k3 by (15% and encompass most of
the experimental data.
In models of urban and regional air chemistry the rate-
constant ratio k2/k3 is an important parameter that determines
the formation of PAN. We recommend the use of a tempera-
ture-independent value of k2/k3 ) 2.17 ( 0.33 in such models.
Although the rate-constant k3 is pressure-dependent, for pres-
sures near ambient the effect is modest with k3 changing by
less than 5% as the pressure is reduced from 700 to 300 Torr
at 298 K. Hence, neglecting the effect of pressure on k2/k3 in
urban and regional air-quality models is reasonable.
Acknowledgment. J.S. and O.J.N. thank the Commission
of the European Communities for financial support. NCAR is
partially supported by the National Science Foundation.
References and Notes
Figure 9. Comparison of literature data for k2/k3 at, or near, ambient
pressure with results from the present work: Cox et al.35 (diamond),
Kenley and Hendry38 (dotted line), Kirchner et al.37 (triangles), Tuazon
et al.30 (squares), Seefeld et al.39 (open circles), this work (filled circles).
The solid line is drawn at k2/k3 ) 2.17. The dashed line represent (15%.
(1) Moortgat, G. K.; Veyret, B.; Lesclaux, R. J. Phys. Chem. 1989,
93, 2362.
(2) Moortgat, G. K.; Veyret, B.; Lesclaux, R. Chem. Phys. Lett. 1989,
160, 443.
(3) Roehl, C. M.; Bauer, D.; Moortgat, G. K. J. Phys. Chem. 1996,
100, 4038.
(1.4 ( 0.2) × 10-11 cm3 molecule-1 s-1, respectively. Our
value of k2 ) (2.0 ( 0.3) × 10-11 cm3 molecule-1 s-1 clearly
supports the work of Villalta and Howard. Maricq and Szente10
measured the formation of NO2 and the loss of NO using IR
and the formation and decay of CH3C(O)O2, CH3O2, and CH3-
ONO using UV absorption. The experimental data were fitted
using a complex mechanism; hence, the value of the rate-
constant is obtained less directly than that of Villalta and
Howard but in a way very similar to the work reported here.
The origin of the discrepancy between the results of Maricq
and Szente10 and that measured here and by Villalta and
Howard11 is unknown.
(4) Villenave, E.; Lesclaux, R. J. Phys. Chem. 1996, 100, 14372.
(5) Bridier, I.; Caralp, F.; Loirat, H.; Lesclaux, R.; Veyret, B.; Becker,
K. H.; Reimer, A.; Zabel, F. J. Phys. Chem. 1991, 95, 3594.
(6) Addison, M. C.; Burrows, J. P.; Cox, R. A.; Patrick, R. Chem. Phys.
Lett. 1980, 73, 283.
(7) Basco, N.; Parmar, S. Int. J. Chem. Kinet. 1987, 19, 115.
(8) Lightfoot, P. D.; Cox, R. A.; Crowley, J. N.; Destriau, M.; Hayman,
G. D.; Jenkin, M. E.; Moortgat, G. K.; Zabel, F. Atmos. EnViron. 1992, 26,
1805.
(9) Wallington, T. J.; Dagaut, P.; Kurylo, M. J. Chem. ReV. 1992, 92,
667.
(10) Maricq, M. M.; Szente, J. J. J. Phys. Chem. 1996, 100, 12380.
(11) Villalta, P. W.; Howard, C. J. J. Phys. Chem. 1996, 100, 13624.
(12) Hansen, K. B.; Wilbrandt, R.; Pagsberg, P. ReV. Sci. Instrum. 1979,
50, 1532.
(13) Sehested, J. Risø-R-804, 1994.
Bridier et al.5 determined a value for k3 of 9.6 × 10-12 cm3
molecule-1 s-1 at 1 bar total pressure and a high-pressure limit
of (1.2 ( 0.2) × 10-11 cm3 molecule-1 s-1. k2 was determined
at 1 bar total pressure of SF6 or CO2 in this work. SF6 and
CO2 are generally more efficient third bodies than air, and it is
expected from the results of Bridier et al.5 that the value of k3
determined here should lie between 0.96 and 1.2 × 10-11 cm3
molecule-1 s-1. This prediction is in excellent agreement with
k3 ) (1.0 ( 0.2) × 10-11 cm3 molecule-1 s-1 obtained here.
The rate-constant values of Addison et al.6 and Basco and
Parmar7 fall below those obtained here and the values of Bridier
et al. As discussed elsewhere,8,9 the chemical mechanism used
in the data analysis of Bridier et al.5 was more complete than
those used by Addison et al.6 and Basco and Parmer.7 Hence,
the results of Bridier et al. are to be preferred. As seen in Figure
8, the results from the present study are in excellent agreement
with those of Bridier et al.5
There have been six previous measurements of the rate-
constant ratio k2/k3 at, or near, atmospheric pressure (Cox et
al.,35 Cox and Roffey,36 Kirchner et al.,37 Tuazon et al.,30 Kenley
and Hendry,38 and Seefeld et al.39). Data from these previous
studies are compared with the results from the present work in
Figure 9. The data from Cox and Roffey36 have a high degree
of scatter with values ranging from 1.2 to 3.0 and are not
included in the plot. All studies have concluded that the rate-
constant ratio k2/k3 is independent of temperature. With the
exception of the study of Hendry and Kenley38 there is also
broad agreement on the magnitude of k2/k3. The solid line in
(14) Wallington, T. J.; Japar, S. M. J. Atmos. Chem. 1989, 9, 399.
(15) Shetter, R. E.; Davidson, J. A.; Cantrell, C. A.; Calvert, J. G. ReV.
Sci. Instrum. 1987, 58, 1427.
(16) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.;
Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina,
M. J. Jet Propulsion Laboratory Publication, 94-26; Pasadena, CA, 1994.
(17) Davidson, J. A.; Cantrell, C. A.; McDaniel, A. H.; Shetter, R. E.;
Madronich, S.; Calvert, J. G. J. Geophys. Res. 1988, 93, 7105.
(18) Schneider, W.; Moortgat, G. K.; Tyndall, G. S.; Burrows, J. P. J.
Photochem. Photobiol. 1987, 40, 195.
(19) Me´rienne, M. F.; Jenouvrier, A.; Coquart, B. J. Atmos. Chem. 1995,
20, 281.
(20) Tsalkani, N.; Toupance, G. Atmos. EnViron. 1989, 23, 1849.
(21) Finlayson-Pitts, B. J.; Pitts, J. J. N. Atmospheric Chemistry; John
Wiley & Sons: New York, 1986.
(22) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F.; Kerr, J.
A.; Troe, J. J. Phys. Chem. Ref. Data 1992, 21, 1125.
(23) Eberhard, J.; Howard, C. J. J. Phys. Chem. 1997, 101, 3360.
(24) Sehested, J.; Nielsen, O. J.; Wallington, T. J. Chem. Phys. Lett.
1993, 213, 257.
(25) Sehested, J.; Nielson, O. J.; Egsgaard, H.; Larsen, N. W.; Pedersen,
T.; Christensen, L. K.; Wiegell, M. J. Geophys. Res. 1995, 100, 20979.
(26) Tyndall, G. S.; Orlando, J. J.; Wallington, T. J.; Hurley, M. D. Int.
J. Chem. Kinet. 1997, 29, 655.
(27) Sehested, J.; Christensen, L. K.; Nielsen, O. J.; Bilde, M.;
Wallington, T. J.; Schneider, W. F.; Orlando, J. J.; Tyndall, G. S. Int. J.
Chem. Kinet., in press.
(28) Sehested, J.; Christensen, L. K.; Nielsen, O. J.; Wallington, T. J.
Int. J. Chem. Kinet., submitted.
(29) Orlando, J. J.; Tyndall, G. S.; Calvert, J. G. Atmos. EnViron. 1992,
26A, 3111.
(30) Tuazon, E. C.; Carter, W. P. L.; Atkinson, R. J. Phys. Chem. 1991,
95, 2434.
(31) Roberts, J. M.; Bertman, S. B. Int. J. Chem. Kinet. 1992, 24, 297.
(32) Roumelis, N.; Glavas, S. Monatsh. Chem. 1992, 123, 63.

CH3C(O)O2 + NO and CH3C(O)O2 + NO2
J. Phys. Chem. A, Vol. 102, No. 10, 1998 1789
(33) Grosjean, D.; Grosjean, E.; Williams, E. L., II. J. Air Waste Manage.
Assoc. 1994, 44, 391.
(42) Wallington, T. J.; Ellermann, T.; Nielsen, O. J.; Sehested, J. J. Phys.
Chem. 1994, 98, 2346.
(34) Mallard, W. G.; Westley, F.; Herron, J. T.; Hampson, R. F. NIST
Chemical Kinetics Database, Version 6.0; NIST Standard Reference Data;
NIST: Gaithersburg, MD, 1994.
(35) Cox, R. A.; Derwent, R. G.; Holt, P. M.; Kerr, J. A. Faraday Trans.
1976, 2061.
(36) Cox, R. A.; Roffey, M. J. EnViron. Sci. Technol. 1977, 11, 900.
(37) Kirchner, F.; Zabel, F.; Becker, K. H. Bunsenges. Phys. Chem. 1990,
94, 1379.
(38) Kenley, R. A.; Hendry, D. G. J. Am. Chem. Soc. 1982, 104, 220.
(39) Seefeld, S.; Kinnison, J.; Kerr, J. A. J. Phys. Chem. 1997, 101, 55.
(40) Ellermann, T.; Sehested, J.; Nielsen, O. J.; Pagsberg, P.; Wallington,
T. J. Chem. Phys. Lett. 1994, 218, 287.
(43) Pagsberg, P.; Sillesen, A.; Jodkowski, J. T.; Ratajczak, E. Chem.
Phys. Lett. 1996, 252, 165.
(44) Tsang, W.; Hampson, R. F. J. Phys. Chem. Ref. Data 1986, 15,
1087.
(45) Lorenz, K.; Rha¨sa, D.; Zellner, R.; Fritz, B. Ber. Bunsen-Ges. Phys.
Chem. 1985, 89, 341.
(46) Gutman, D.; Nelson, H. H. J. Phys. Chem. 1983, 87, 3902.
(47) Barnhard, K. J.; Santiago, A.; He, M.; Asmar, F.; Weiner, B. R.
Chem. Phys. Lett. 1991, 178, 150.
(48) Bridier, I.; Veyret, B.; Lesclaux, R.; Jenkin, M. E. J. Chem. Soc.,
Faraday Trans. 1993, 89, 2993.
(41) Sehested, J.; Ellermann, T.; Nielsen, O. J.; Wallington, T. J. Int. J.
Chem. Kinet. 1994, 26, 615.
(49) Cox, R. A.; Munk, J.; Nielsen, O. J.; Pagsberg, P.; Ratajczak, E.
Chem. Phys. Lett. 1990, 173, 206.