Thermal Decomposition of CF3O2NO2

Article abstract of DOI:10.1021/jp953205g

The unimolecular decomposition rate constant of CF3O2NO2 has been measured in detail as a function of temperature, pressure, and collision partner (M = N2, O2, NO).Temperatures were between 264 and 297 K, and total pressures ranged from 3 to 1013 mbar.The first-order decay of CF3O2NO2 in the presence of excess NO was followed in a temperature-controlled DURAN glass chamber by long-path IR absorption, using the absorption bands at 1768 and 1303 cm^-1.At 1013 mbar, the first-order decomposition rate constants are best represented by the Arrhenius expression k₃ = 5.7E15exp<(-97.7 +/- 1.0) kJ mol^-1/RT> s^-1 (2?).The temperature and pressure dependencies of k₃ are well reproduced by the equation log(k₃/k_3,infinite) = log<(k_3,0/k_3,infinite)/(1 + k_3,0/k_3,infinite)> + log(F_c)<1 + 3,0/k_3,infinite)/N_c>²>^-1, N_c = 0.75-1.27 log(F_c) with the parameters k_3,0/ = 2.4E-5exp(-78.4 kJ mol^-1/RT) cm³ molecule^-1 s^-1, k_3,infinite = 1.49E16exp<(-99.3 +/- 1.3) kJ mol^-1/RT> s^-1, F_c = 0.31, and k_3,0(M=O2) ca. k_3,0(M=N2).By combining the present decomposition rate constants with recombination rate constants k_-3 from Caralp et al., the following thermochemical data for the equilibrium CF3O2NO2 <-> CF3O2 + NO2 (k₃,k_-3) are derived from second- and third-law evaluations: ΔH⁰_r,298 = 102.7 +/- 2.0 kJ mol^-1, ΔS⁰_r,298 = 163 +/- 7 J mol^-1 K^-1.The temperature dependence of the equilibrium constant between 200 and 300 K is described by the expression K_c = k₃/k_-3 = 3.80E27exp<(-12140 +/- 240)K/T> molecules cm^-3.Consistency of the data on k₃ (this work) and k_-3 is shown by comparing experimental and theoretical limiting low-pressure rate constants, which lead to the reasonable value β_c = 0.17 for the collision efficiency of N2.The present data confirm that CF3O2NO2 is thermally quite stable in the upper troposphere and lower stratosphere and that its lifetime is probably limited by photolysis in these regions of the atmosphere.

Full text of DOI:10.1021/jp953205g

J. Phys. Chem. 1996, 100, 6587-6593
6587
Thermal Decomposition of CF3O2NO2
A. Mayer-Figge, F. Zabel,* and K. H. Becker
Bergische UniVersita¨t-Gesamthochschule Wuppertal, Physikalische Chemie/FB 9, 42097 Wuppertal, Germany
ReceiVed: October 30, 1995; In Final Form: January 25, 1996X
The unimolecular decomposition rate constant of CF3O2NO2 has been measured in detail as a function of
temperature, pressure, and collision partner (M ) N2, O2, NO). Temperatures were between 264 and 297 K,
and total pressures ranged from 3 to 1013 mbar. The first-order decay of CF3O2NO2 in the presence of
excess NO was followed in a temperature-controlled DURAN glass chamber by long-path IR absorption,
using the absorption bands at 1768 and 1303 cm-1. At 1013 mbar, the first-order decomposition rate constants
are best represented by the Arrhenius expression k3 ) 5.7 × 1015 exp{(-97.7 ( 1.0) kJ mol-1/RT} s-1 (2σ).
The temperature and pressure dependencies of k3 are well reproduced by the equation log(k3/k3,∞) ) log-
{(k3,0/k3,∞)/(1 + k3,0/k3,∞)} + log(Fc){1 + [log(k3,0/k3,∞)/Nc]2}-1, Nc ) 0.75-1.27 log(Fc) with the parameters
k3,0/[N2] ) 2.4 × 10-5 exp(-78.4 kJ mol-1/RT) cm3 molecule-1 s-1, k3,∞ ) 1.49 × 1016 exp{(-99.3 ( 1.3)
kJ mol-1/RT} s-1, Fc ) 0.31, and k3,0(M)O2) ≈ k3,0(M)N2). By combining the present decomposition rate
constants with recombination rate constants k-3 from Caralp et al., the following thermochemical data for the
equilibrium CF3O2NO2 S CF3O2 + NO2 (k3,k-3) are derived from second- and third-law evaluations: ∆H°
r,298
) 102.7 ( 2.0 kJ mol-1, ∆S° ) 163 ( 7 J mol-1 K-1. The temperature dependence of the equilibrium
r,298
constant between 200 and 300 K is described by the expression Kc ) k3/k-3 ) 3.80 × 1027 exp{(-12140 (
240)K/T} molecules cm-3. Consistency of the data on k3 (this work) and k-3 is shown by comparing
experimental and theoretical limiting low-pressure rate constants, which lead to the reasonable value âc )
0.17 for the collision efficiency of N2. The present data confirm that CF3O2NO2 is thermally quite stable in
the upper troposphere and lower stratosphere and that its lifetime is probably limited by photolysis in these
regions of the atmosphere.
Introduction
Experimental Section
The experiments were performed in a temperature-controlled
420 l Duran glass photoreactor (chamber I) which was described
elsewhere13 in more detail. CF3O2NO2 was prepared in situ by
photolyzing CF3I/O2/NO2/N2 mixtures. Typical initial concen-
trations were 1 × 1015 molecules cm-3 CF3I, (1-2) × 1014
molecules cm-3 NO2 which was added in several portions during
the photolysis time in order to reduce the photolysis of NO2,
1.3 × 1016 molecules cm-3 of O2, and 1000 mbar of N2. Since
the photolysis rate of CF3I was very low in chamber I (λ g
300 nm), the photolysis was performed at 253.6 nm (σ253.6(CF3I)
) 4.3 × 10-19 cm2 molecule-1 (ref 14) in a second 420 L
chamber (chamber II) equipped with three internal 40 W low-
pressure mercury lamps. CF3O2NO2 is formed by the following
mechanism:
Trifluoromethyl peroxynitrate (CF3O2NO2) is a potential
intermediate in the atmospheric degradation of fully halogenated
hydrocarbons (e.g., CF4, CF3Cl, CF3Br, CF3I) and of substitutes
replacing these compounds (CHF3, CH3CF3, CH2FCF3 (R 134a),
CHF2CF3, CHCl2CF3, CHCl2CF2CF3, etc.). Pure CF3O2NO2
has been synthesized and characterized by Hohorst and Des-
Marteau.1 It was identified by its IR spectrum as an intermediate
in the photolysis of CF3NO in the presence of O2 and NO2 by
Chen et al.2 and Wallington et al.3 Its lifetime in the atmosphere
depends mainly on the rates of thermal decomposition and
photolysis. So far, experiments have revealed only a lower limit
for the thermal decomposition rate constant of CF3O2NO2 at
room temperature and atmospheric pressure.3,4 Experiments on
the thermal stability of other chloro(fluoro)methyl peroxy-
nitrates5-10 show that the energies of the weakest bond in these
molecules are on the order of 100 kJ mol-1. From these data
it may be estimated that CF3O2NO2 is thermally fairly long-
lived at the temperatures of the upper troposphere and lower
stratosphere. In fact, in these regions of the atmosphere the
lifetime may be limited by photolysis.11,12 However, the UV
absorption coefficients of CF3O2NO2 have been published only
up to wavelengths of 280 nm,11 and the extrapolation to longer
wavelengths introduces considerable uncertainties to the pho-
tolysis lifetimes derived from these data. For this reason, kinetic
data on both thermal decomposition and photolysis are needed
to reliably estimate atmospheric lifetimes of CF3O2NO2. In the
present work, thermal decomposition rate constants of CF3O2-
NO2 have been measured in detail as a function of temperature,
pressure, and collision partner (M ) N2, O2, NO), and the data
are evaluated in terms of unimolecular reaction rate theory.
CF3I + hν w CF3 + I
(1)
(2)
CF3 + O2 (+M) w CF3O2 (+M)
CF3O2 + NO2 (+M) S CF3O2NO2 (+M)
(3,-3)
During photolysis, a small amount of CF2O was formed in side
reactions, probably initiated by the slow photolysis of NO2. CH3-
ONO2 which could complicate the kinetic analysis of reaction
3 was observed as a major product in the photolysis of CF3I/
O2/NO/NO2/N2 mixtures with initial [NO]/[NO2] ratios of ≈1
but was not detectable under our standard photolysis conditions.
Subsequent to photolysis, a valve was opened between chamber
II and the evacuated chamber I which was kept at the desired
reaction temperature. The gas mixture in chamber I was then
pressurized with N2 to the total pressure of the respective
experiment. At the selected reaction temperatures, the concen-
X Abstract published in AdVance ACS Abstracts, March 15, 1996.
0022-3654/96/20100-6587$12.00/0 © 1996 American Chemical Society

6588 J. Phys. Chem., Vol. 100, No. 16, 1996
Mayer-Figge et al.
Figure 1. IR absorption spectrum of CF3O2NO2; the spectrum is strongly perturbed by imperfectly subtracted absorption from the precursor CF3I
between 1178 and 1192 cm-1
.
trations of CF3O2 present in the equilibrium (3,-3) were so
Research grade CF3I (Aldrich, 99%) and NO2 were used as
supplied. Research-grade NO was trapped at liquid nitrogen
temperature and purified by distillation. Pressures were mea-
sured using capacitance pressure gauges. The temperature in
the reaction chamber was measured with a platinum resistance
gauge. The temperature was constant within (0.1 K during
an experiment.
small that the self-reaction of CF3O2,
CF3O2 + CF3O2 w 2CF3O + O2
(4)
(k4,295K ) (1.8 ( 0.5) × 10-12 cm3 molecule-1 s-1 (ref 15)),
was unimportant. Thermal decomposition was initiated by
addition of 5 × 1015 molecules cm-3 of NO, thereby withdraw-
ing CF3O2 radicals from the equilibrium (3,-3) via reaction 5:
Results and Data Reduction
IR Absorption Spectrum of CF3O2NO2. The IR absorption
spectrum of CF3O2NO2 is shown in Figure 1. This spectrum
was obtained from the product spectrum of the 253.6 nm
photolysis of a mixture of 0.053 mbar of CF3I, 1.3 mbar of
synthetic air, 0.008 mbar of NO2, and 1011 mbar of N2 by
spectrally subtracting absorptions of CF3I, CF2O, and NO2. The
band maxima/double maxima were at the following positions
(in cm-1; see Figure 1 for comparison): 3057, 1768/1762, 1303,
1250/1245, ≈1190, 959, 792, 710, ≈675. These band positions
are in good agreement with all the bands denoted to be “very
strong” and “strong” by Hohorst and DesMarteau1 and with the
work of Chen et al.2 The 1190 cm-1 band is strongly perturbed
due to imperfectly subtracted absorption of the precursor CF3I.
IR absorptions at 1768 and 1303 cm-1 were used to monitor
the concentration of CF3O2NO2 as a function of time.
Kinetic Results on the Thermal Decomposition of CF3-
O2NO2. The time dependence of the IR absorbances of CF3O2-
NO2 at 1768 and 1303 cm-1 was analyzed according to a first-
order rate law (Figure 2). The resulting first-order rate constants
were identical for both wavelengths within error limits and were
averaged. A total of 56 experiments were performed in N2 at
temperatures between 264 and 297 K and total pressures
between 3.5 and 1013 mbar. At low pressures, five additional
experiments were performed in O2 and two experiments with
much higher than normal mixing ratios of NO in order to
estimate the collision efficiencies of these buffer gases.
CF3O2 + NO w CF3O + NO2
(5)
The CF3O2 loss in reaction 5 is irreversible because CF3O
rapidly reacts with NO to form the stable product CF2O:16,17
CF3O + NO w CF2O + FNO
(6)
(k6,295K ) 1.7 × 10-11 cm3 molecule-1 s-1 (ref 17)). In the
course of the reaction, the [NO]/[NO2] ratio decreased by about
a factor of 2 due to the NO2 formed in reactions 3 and 5.
However, at the end of the reaction time [NO]/[NO2] was always
larger than 10. Since k5 is higher than k-3 by at least a factor
of 1.6,10 the relationship k5[NO] . k-3[NO2] was valid
throughout the reaction. The relative CF3O2NO2 concentrations
were monitored as a function of time by long-path IR absorption
(optical path length 50.4 m), using a built-in White mirror
system and an FTIR spectrometer (Nicolet Magna 550). The
effective first-order rate constants for CF3O2NO2 loss after NO
addition, keff, were corrected for wall loss. Individual wall loss
rate constants kw were determined in each experiment from the
slow first-order loss rate of CF3O2NO2 before NO addition.
Numerical simulation of the reaction mechanism using the
Larkin kinetic modeling program18 established that CF3O2 loss
by self-reaction was not responsible for the slow decrease of
CF3O2NO2 before the addition of NO. For this reason, the
difference keff - kw was identified as the gas-phase unimolecular
decomposition rate constant k3. kw amounted to (0-3) × 10-5
s-1, showing a small systematic increase with increasing
temperature and decreasing pressure.
The temperature dependence of the first-order rate constants
k3 at different total pressures is shown in Figure 3. Rate
constants obtained at total pressures between 3.6 and 3.9 mbar

Thermal Decomposition of CF3O2NO2
J. Phys. Chem., Vol. 100, No. 16, 1996 6589
Figure 2. IR absorbance of CF3O2NO2 at 1768 cm-1 as a function of
time before (9) and after (b) NO addition; T ) 276.3 K, p ) 3.9
mbar, M ) N2.
Figure 4. Falloff curves for k3 at 273, 280, and 288 K; full lines are
calculated from molecular parameters and shifted to fit to the data points
(see text); statistical errors for the individual data points are smaller
than the size of the symbols.
TABLE 1: Arrhenius Paramters of k3 at Different Total
Pressures
M
p [mbar]
A [s-1
]
Ea [kJ mol-1
]
N2
N2
N2
N2
O2
1013
104.3
10.6
3.7
5.7 × 1015
9.1 × 1014
1.05 × 1014
1.85 × 1013
3.1 × 1013
97.7 ( 1.0
94.2 ( 1.0
90.8 ( 1.1
87.8 ( 1.6
89.0 ( 3.3
3.6
first-order rate constants in the low- and high-pressure limit,
respectively, and Fc is a parameter which determines the
broadening of the falloff curve and can be calculated from
molecular parameters of the decomposing molecule (Fc ) 0.31,
see below). The resulting reduced falloff curves were then fitted
to the data points under the constraint that the limiting low-
and high-pressure rate constants k0 and k∞ obey the Arrhenius
equation. The temperature dependencies of k0 and k∞ derived
from Figure 4 are represented by the expressions k3,0/[N2] )
2.4 × 10-5 exp(-78.4 kJ mol-1/RT) cm3 molecule-1 s-1 and
k3,∞ ) 1.49 × 1016 exp{(-99.3 ( 1.3) kJ mol-1/RT} s-1. Fc
was calculated using vibrational frequencies of CF3O2NO2 from
Hohorst and DesMarteau1 (1760, 1305, 1292, 1244, 1186, 953,
880, 783, 702, 669, 600, 562, 485, 436, 371, 284, and 253 cm-1),
Nielsen et al.20 (263 cm-1 with CF3CHCl2 as the model
compound), and estimated values for the torsional vibrations
Figure 3. Arrhenius plots for k3 at different total pressures; straight
lines are least-squares fits to the data points for M ) N2.
were converted to 3.7 mbar using the empirical relationship k3
[N2]1/2 which is deduced from the slope of the falloff curves
for [M] ≈ 1 × 1017 cm3 molecule-1 (see Figure 4 for
comparison). For larger total pressures, the measured rate
constants were directly included in the Arrhenius plots. The
scatter of the data points is very low, due to the absence of
both interfering side reactions and overlapping IR absorptions
from either main or side products. The Arrhenius parameters
for different total pressures (1013 ( 1, 104 ( 2, 10.6 ( 0.3,
3.7 mbar for M ) N2 and 3.6 ( 0.2 mbar for M ) O2) are
presented in Table 1. The activation energies decrease with
decreasing total pressure, in agreement with unimolecular
reaction rate theories.
(120 cm-1 using CH3O2NO2 as a model compound; 40 and
21
35 cm-1 (refs 9 and 22)). The selected wavenumbers of the
low-frequency vibrations representing the three hindered rota-
tions are based on a combined treatment22 of experimental data
on the thermal decomposition of CCl3O2NO2, CCl2FO2NO2, and
CClF2O2NO2 and the corresponding recombination reactions
RO2 + NO2 (+M) w RO2NO2 (+M) (see below). Since it
was not clear if it is more appropriate to approximate the
hindered rotations by free rotations (Fc ) 0.36, including weak
collision corrections23) or torsional vibrations (Fc ) 0.26), the
mean value 0.31 was adopted for Fc. For the experiments
performed in 3.6 mbar of O2, there was no significant change
of either the absolute value of k3 or the activation energy as
compared to M ) N2 (see Figure 3). For the experiments at a
total pressure of ≈3.6 mbar with NO as a major collision partner
(pNO ≈ 1.5 mbar as compared to 0.2 mbar in the other experi-
ments), there was no notable change of k3. Hence
The pressure dependence of k3 at 288, 280, and 273 K is
shown in double-logarithmic plots in Figure 4. The data points
were derived from the Arrhenius parameters in Table 1. The
full lines were calculated in a reduced form by application of
eq I:19
log(k/k∞) ) log{(k0/k∞)/(1 + k0/k∞)} +
log(Fc){1 + [log(k0/k∞)/Nc]2}-1 (I)
Nc ) 0.75 - 1.27 log(Fc)
k3(MdNO) ≈ k3(MdO2) ≈ k3(MdN2)
for p, T ) constant
k0, which is proportional to the gas density [M], and k∞ are the

6590 J. Phys. Chem., Vol. 100, No. 16, 1996
Mayer-Figge et al.
30
TABLE 2: Equilibrium Constants Kc ) k3/k-3 Calculated from k3 (This Work) and k-3
[M]
k3
k-3
Kc ) k3/k-3
Kc(T)
[1017 molecules cm-3
]
T [K]
[10-3 s-1
]
[10-12 cm3 molecule-1 s-1
]
[109 molecules cm-3
]
[109 molecules cm-3
0.193
]
0.65
1.95
3.25
0.65
1.95
3.25
0.65
1.95
3.25
273
273
273
280
280
280
288
288
288
0.236
0.399
0.495
0.627
1.07
1.35
1.81
0.121
2.09
2.57
1.12
1.88
2.29
0.102
1.68
2.03
0.195
0.191
0.193
0.562
0.572
0.587
1.77
0.574
1.86
3.12
3.95
1.85
1.94
CF3O + HX w CF3OH +
According to the previous discussion, k3 can be calculated for
all atmospheric applications by using eq I and the following
parameters:
X (e.g. X ) OH, alkyl from impurities) (8)
However, there was no indication from the IR spectra that
either CF3ONO2 or CF3OH27-29 was produced under the
2
k3,0/[N2] )
reaction conditions of the thermal decomposition experiments.
In addition, there was no change of the product distribution with
total pressure.
2.4 × 10-5 exp(-78.4 kJ mol-1/RT) cm3 molecule-1 s-1
k3,∞ ) 1.49 × 1016 exp{(-99.3 ( 1.3) kJ mol-1/RT} s-1
Discussion
Kinetics and Thermochemistry. Wallington et al.3 obtained
a lower limit for k3 of 0.019 s-1 (ref 4) at 295 K and atmospheric
pressure which is consistent with the present value 0.029 s-1
derived from Table 1.
Fc ) 0.31
k3(MdO2) ≈ k3(MdN2)
(II)
The present results on k3 can be combined with recombination
rate data of Caralp et al.30 These authors measured k-3 in the
pressure range 1.3-13 mbar at three different temperatures (233,
298, and 373 K). They analyzed their data by RRKM theory
and presented the temperature and pressure dependence of k-3
in terms of eq I. In Table 2, recombination rate constants k-3
obtained by interpolation of the data of Caralp et al. (Table 1
in ref 30) are listed together with the dissociation rate constants
k3 derived from the present work (eq I and parameter set II), in
the common temperature and pressure range of both studies.
The resulting equilibrium constants Kc ) k3/k-3 are also included
in Table 3 and shown as a van’t Hoff plot in Figure 5. The
temperature dependence of Kc between 273 and 288 K is
represented by the expression Kc ) (1.40 ( 0.23) × 1027exp-
{(-11860 ( 240)K/T} molecules cm-3, where the error limits
A consistent analysis of thermal decomposition rate constants
and thermochemical properties of peroxynitrates9,22 suggests
that for most of these compounds the activation energy at 1013
mbar is only a few tenths of a kilojoule below the limiting high-
pressure value. Hence, the error limits of Ea,∞ were identified
in the present work with the error limits of the experimental
activation energy at 1013 mbar, slightly enhanced by 0.3 kJ
mol-1 allowing for the additional uncertainty introduced by the
extrapolation to the high-pressure limit. For Ea,0, error limits
are uncertain due to the large extrapolation necessary to attain
the low-pressure limit. However, in the range 1-1000 mbar,
the rate constants calculated by using eqs I and II are uncertain
by no more than 10% at room temperature and by less than
50% over the whole range of tropospheric and stratospheric
temperatures.
reflect the uncertainties of both k3 and k-3
.
The data in Table 2 lead to ∆H0r,280 ) 100.9 ( 2.0 kJ mol-1
(2σ). Using the above-mentioned vibrational frequencies for
CF3O2NO2, CF3O2, and NO2, the reaction enthalpy can be
converted from 280 to 298 K, resulting in
Reaction Products. According to reactions 3, 5, and 6, NO2,
CF2O, and FNO are expected to be the main reaction products
of the thermal decomposition of CF3O2NO2 in the presence of
NO. In fact, the concentrations of NO2 and CF2O which were
already present before the addition of NO, increased during the
thermal decomposition of CF3O2NO2. FNO was easily detected
as a product by its prominent Q branches at 1844 and 766
∆H0r,298(CF3O2NO2wCF3O2+NO2) )
100.8 ( 2.0 kJ mol-1 (“second law”)
cm-1 16,24
However, [FNO] rapidly reached a steady-state value,
.
then decreased with a time constant close to that of CF3O2-
NO2. The loss reaction of FNO is unclear; however, it is well-
known25 that FNO readily reacts with both water, leading to
HF and HONO (which, actually, were also seen in the product
spectra), and glass surfaces, leading to SiF4. SiF4, if formed,
could not be detected by its IR spectrum26 in the present
experiments due to the strong overlap by absorption from CF3I.
In one experiment, 5 × 1013 molecules cm-3 CH4 was added
to check for possible interference by F atoms which would
rapidly react with methane, forming CH3 and HF. However,
there was no evidence from the IR spectra for the presence of
any side products originating from CH3 radicals. CF3ONO2 and
CF3OH could be formed via reactions 7 and 8:
Applying a third-law analysis based on Kc(280 K) ) (5.74
( 0.92) × 108 molecule cm-3, the frequencies from above for
CF3O2NO2 and from Butler and Snelson31 for CF3O2, and the
moments of inertia of CF3O2NO2 from Hohorst and DesMar-
teau1 and of CF3O2 from Caralp et al.:30
∆H0r,298(CF3O2NO2wCF3O2+NO2) )
104.6 ( 2.4 kJ mol-1 (“third law”)
is derived where the error limits reflect the error of Kc and, in
particular, the uncertainty of the frequencies of the three
torsional vibrations. Both values of ∆H0r,298 are in agreement
within the combined error limits, demonstrating the consistency
of the available data on k3 and k-3. Since the error limits of
the ∆H0r,298 values from the second- and third-law treatments
CF3O + NO2 (+M) w CF3ONO2 (+M)
(7)

Thermal Decomposition of CF3O2NO2
J. Phys. Chem., Vol. 100, No. 16, 1996 6591
from ref 30, ∆H0r,298 ) 102.7 ( 2.0 kJ mol-1, ∆Hr0,0 ) E0 )
TABLE 3: Comparison of Selected Kinetic and
100.5 ( 2.0 kJ mol-1, ∆Sr0,298 ) 163 ( 7 J mol-1 K-1
.
Thermochemical Data from Reaction 3 with Literature Data
Caralp et al.30 Destriau, Troe34
this work
Based on these data, the Kc values in the atmospherically
important temperature range 200-300 K are approximated by
the following simple van’t Hoff-type expression:
k3 (280 K, 60
mbar N2) [s-1
k3 (298 K, 1013
mbar N2) [s-1
k3 (220 K, 50 mbar
N2 + O2) [s-1
Kc (280 K)
[molecules cm-3
3.4 × 10-3
7.1 × 10-2
3.5 × 10-8
6.8 × 108
(101.6)b
3.3 × 10-3a (2.3 ( 0.3) × 10-3
]
]
8.2 × 10-2
3.1 × 10-8
1.19 × 109
(101.6)c
4.2 × 10-2
Kc ) k3/k-3
)
2.9 × 10-8
3.80 × 1027 exp{(-12140 ( 240)K/T} molecules cm-3
]
(5.74 ( 0.92) × 108
(200-300 K)
]
E0 ) ∆H0
100.5 ( 2.0
The internal consistency of this data analysis can be tested
by comparing the experimentally derived low-pressure value
of k3 ()k3,exp(pw0) ≡ k3,0, see Figure 4) with the value
calculated using the equation32
r,0
[kJ mol-1
]
∆H0r,298 [kJ mol-1
]
(105.0)d
(105.0)c
102.7 ( 2.0
a See also ref 37. b E0(CF3O2NO2) ) E0(CF2ClO2NO2) assumed.
c Value from ref 30 adopted. d Average of calculated ∆H0r,298 values
for reactions 3 and 11-13.
k0 ) âck0(strong collisions) )
âc[M]ZLJFvib,harm(E0)Qvib-1 exp(-E0/RT)RTFEFanhFrot (III)
where ZLJ ) Lennard-Jones collision number, Fvib,harm(E0) )
vibrational density of states at E ) E0 for harmonic oscillators,
Qvib) vibrational partition function, FE ) correction factor
allowing for the energy dependence of Fvib,harm at E > E0, Fanh
) correction factor for anharmonicity of vibrations, Frot
)
correction factor for contributions of external rotations. Defin-
ing âc by k0,exp/k0(strong collisions), âc(M)N2) was found to
be in the range 0.06-0.9 for a large number of reactions at
room temperature.32,22 For the decomposition reactions of a
total of 12 peroxynitrates where comprehensive data were
available for both the dissociation and recombination reactions,
âc(M)N2) was in the range 0.10-0.85.9 Since, for theoretical
reasons, âc is expected to be very similar for different
decomposition reactions in the same buffer gas, most of the
scatter of these âc values is due to the large number of
uncertainties inherent in both the theoretical and the experi-
mental quantities entering eq III; usually, a value of âc ≈ 0.2-
0.3 is assumed for M ) N2. Accordingly, the consistency of
the present data analysis can be checked by calculating âc from
eq III, with k0 being set equal to k3,exp(pw0) and k0(strong
collisions) being calculated from molecular parameters. For this
purpose, the above vibrational frequencies of CF3O2NO2 were
used to calculate Fvib, Qvib, FE, Fanh, and Frot; ZLJ was estimated
by a method described in the literature21,32,33 which is based on
the boiling point of CF3O2NO2.1 With the present values E0 )
100.5 kJ mol-1 and k3,0(M)N2) ) 5.7 × 10-20 cm3 molecule-1
s-1 at 280 K (see eq II and Figure 4), âc(M)N2) ) 0.17 is
derived for reaction 3 which is well within the expected range
of values.
Figure 5. van’t Hoff plots of Kc ) k3/k-3; data points: mean Kc values
from Table 2; broken line: least-squares fit of the experimental data
from Table 2; full line: “best” straight line through the slightly curved
representation of the equilibrium constants calculated for the temper-
ature range 200-300 K from the recommended set of values for
∆H0r,298, ∆Sr0,298, and the vibrational frequencies; error bars: estimated
total errors of Kc, the major part of these errors is probably temperature
independent.
are comparable, the mean value
∆H0r,298(CF3O2NO2wCF3O2+NO2) )
102.7 ( 2.0 kJ mol-1 (final value)
Data on the thermal decomposition of the analogous perox-
ynitrates CCl3O2NO2, CCl2FO2NO2, and CClF2O2NO2 from
Ko¨ppenkastrop and Zabel7 and on the reverse reactions from
Caralp et al.30 have been used by Caralp et al.30 and by Destriau
and Troe34 to accomplish a self-consistent theoretical analysis
of the following group of reactions:
has been adopted for the further discussion.
To obtain a consistent set of molecular parameters and
thermochemical quantities, the three low-frequency vibrations
were adjusted so that both the experimental value Kc ) 5.74 ×
108 molecules cm-3 at 280 K and the mean value ∆H0r,298
)
102.7 kJ mol-1 obtained from second- and third-law treatments
of the experimental data are reproduced at the same time.
The experimental data on reaction 3 are thus best described
CCl3O2NO2 (+M) S CCl3O2 + NO2 (+M)
CCl2FO2NO2 (+M) S CCl2FO2 + NO2 (+M)
CClF2O2NO2 (+M) S CClF2O2 + NO2 (+M)
CF3O2NO2 (+M) S CF3O2 + NO2 (+M)
(11,-11)
(12,-12)
(13,-13)
(3,-3)
using the following set of parameters: frequencies (in cm-1
)
CF3O2NO2: 1760, 1303, 1292, 1244, 1186, 953, 877, 783, 702,
669, 600, 562, 485, 436, 371, 287, and 264 (from ref 1); 263
(from ref 20); 75, 34, 30 (from this work, best fit to the data);
CF3O2: 1303, 1260, 1172, 1092, 870, 692, 597, 580, 448, 286,
190, 120 (from ref 31); NO2: 1666, 1358, 757 (from ref 21)
moments of inertia (in amu Å2); CF3O2NO2: 129.7, 558.9, and
600.5 (from ref 1); CF3O2: 93.0, 161.5, 162.8 (from ref 30);
NO2: 2.10, 38.6, 40.8 (from ref 21), Kc (280 K) ) (5.74 (
0.92) × 108 molecules cm-3 with k3 from this work and k-3
Since there were no experimental data for k3 at that time, k3
was estimated from the results on reactions 11-13 by analogy.
Essential features of these treatments were the following: Caralp

6592 J. Phys. Chem., Vol. 100, No. 16, 1996
Mayer-Figge et al.
TABLE 4: Comparison of Selected Kinetic and
by the bond energy and to a lesser extent by the preexponential
factor in a series of analogous compounds (see, e.g., ref 22),
and that k11, k12, k13, and k3 are close to the high-pressure limit
such that there is no notable difference in falloff (see Table 4),
the absolute values of the rate constants are more appropriate
to derive a trend in bond energies than is the activation energy.
Inspection of Table 4 shows that there is a distinct drop in the
rate constants when going from CCl3O2NO2 to CFCl2O2NO2
and a smaller decline from CFCl2O2NO2 to CF2ClO2NO2,
whereas the rate constants for CF2ClO2NO2 and CF3O2NO2 are
about the same. When the difference in rate constants is totally
ascribed to different bond energies, E0 increases by 2.6, 1.3,
and 0 kJ mol-1 when Cl is successively replaced by F in CX3O2-
NO2 (X ) Cl, F). The present data thus show that it might be
inadequate to linearly extrapolate a trend of kinetic data within
a series of compounds when very accurate values are required.
Thermochemical Data for Reactions 3 and 11-13
CCl3O2NO2 CCl2FO2NO2 CClF2O2NO2 CF3O2NO2
(ref 7)
kdis(298 K, 1013 0.20
mbar N2) [s-1
Ea(≈1000 mbar 96.8 ( 1.4 100.3 ( 1.8 98.7 ( 1.8
(ref 7)
(ref 7)
(this work)
0.070
0.041
0.042
]
97.7 ( 1.0
N2) [kJ mol-1
∆H0r,0 [kJ mol-1
]
100.1 ( 3a 102.8 ( 3a
103.9 ( 3a
100.5 ( 2.0
0.72
]
k/k∞ (298 K, 1013 0.70
0.66
0.76
mbar N2)
a Based on third-law analysis of dissociation7 and recombination30
data.
et al.30 used RRKM calculations with MNDO calculated
vibrational frequencies and moments of inertia for the perox-
ynitrates and the corresponding peroxy radicals, a single rate
constant for k-12 from their own work at 273 K and 10 mbar,
a k12 value from Ko¨ppenkastrop and Zabel35 at the same
conditions, and the assumption that the reaction enthalpies of
reactions 11, 13, and 3 are equal to that of reaction 12. On the
basis of these presuppositions, Caralp et al.30 were able to
satisfactorily fit their own data on the recombination rate
constants k-11, k-12, and k-3 as a function of temperature and
pressure by RRKM calculations, with only a few parameters to
be slightly adjusted for the individual reactions. As a result of
these calculations, k3 values were also predicted which agree
with the present results within 60% over the complete range of
experimental conditions. Destriau and Troe34 used high-pressure
recombination rate constants k-3, k-11, k-12, and k-13 as
calculated from SACM theory36 and preliminary data on k11,
k12, and k13 from Ko¨ppenkastrop and Zabel7 to derive kinetic
and thermochemical data for reactions 11-13, and 3.
When the second- and third-law analysis of kinetic data of a
dissociation-recombination system leads to different values for
the enthalpy of reaction, the third-law value is usually preferred
(see, e.g., work on the reactions ClO + ClO w ClOOCl,38
CH3C(O)O2 + NO2 w CH3C(O)O2NO2,39 F + O2 w FO240).
This is certainly adequate when the vibrational frequencies of
the dissociating molecule and the products are well-known and/
or the experimental data on the temperature dependence of Kc
are not very accurate; however, the accuracy of third-law
treatments may be overestimated in other cases. The low-
frequency vibrations and, in particular, the very low torsional
oscillations are often unknown, and estimates are subject to large
uncertainties. In the present case of the unimolecular decom-
position of CF3O2NO2, evaluation of the same kinetic data with
two different sets of vibrational frequencies and moments of
inertia, i.e., (i) the analysis by Caralp et al.30 with MNDO
calculated parameters, and (ii) the analysis in the present work
with experimentally determined parameters and estimated values
based on thermochemical data of analogous reactions, leads to
The results on reaction 3 of Caralp et al.30 and Destriau and
Troe34 are compared with the experimental data of the present
work in Table 3. Table 3 shows that the most accurate kinetic
result of this work (the k3 value at 280 K, 60 mbar, i.e., in the
center of the temperature and pressure ranges covered by the
present experiments) is quite well reproduced by the previous
estimates which were deduced from kinetic results on the
decomposition of other peroxynitrates by analogy. The general
agreement between the k3 data estimated earlier30,34 and the
experimental data from this work shows that kinetic and
thermochemical data can be estimated quite reliably using
existing theories when sufficient kinetic and thermochemical
data are known for similar molecules. For k3, the slightly
different temperature dependence of Kc leads, by accident, to
even closer agreement between the earlier estimates and the
present results for the conditions close to the tropopause (220
K, 50 mbar) which are of particular interest for atmospheric
applications.
third law values of ∆Hr0,298 which are different by 3.1 kJ mol-1
.
It has been argued30 that MNDO calculations give frequencies
which are high by 10-20% for stretching vibrations but well
approximate deformation vibrations and torsions. However,
Dewar et al.41 suggested that MNDO calculations underestimate
the torsional frequencies. In their paper, a total of nine MNDO
calculated torsional frequencies were compared with experi-
mental values, and the average deviation between both sets of
data was νobs/νcalc ≈ 1.33. On the other hand, there can also
be a large scatter in experimental values. With respect to the
present case, experimental values for the wavenumbers of the
torsions of both CF3O2NO2 and CF3O2 are not known. Fol-
lowing a suggestion of Butler and Snelson,31 CF3OF can be
used as a model compound for the CF3O2 radical. The MNDO
calculated value for the torsion in CF3OF is 80 cm-1 42
By
.
multiplying by 1.33, 106 cm-1 results as a “best” MNDO value.
On the other hand, three experimental values of 56,43 120,44
and 127,45 cm-1 exist for the torsion in CF3OF. If these
wavenumbers are adopted for CF3O2, their range translates to
Kc values different by a factor of 2 which corresponds to an
uncertainty of ∆H0r,298 of 1.6 kJ mol-1. In CF3O2NO2, there
are three of these torsions, and the uncertainty of the derived
thermochemical values is correspondingly larger. Caralp et al.30
suggested that MNDO-calculated frequencies represent a con-
sistent set of values, implying that systematic errors in these
frequencies apply to both the dissociating molecule and the
products and largely cancel in the calculation of Kc. In the case
of CF3O2NO2 and CF3O2, however, two of the three torsions in
CF3O2NO2 which strongly contribute to the vibrational partition
function disappear when the decomposition products CF3O2 and
Since there is a large body of accurate data available on both
k3 and k-3, several more general conclusions shall be drawn
with respect to the evaluation of kinetic and thermochemical
data.
The present data on k3 may be compared with earlier results
from this laboratory on k11, k12, and k13.7 The former experi-
ments were carried out in the pressure range 10-800 mbar (M
) N2) and at temperatures between 260 and 293 K. Slight
extrapolation to 1013 mbar and 298 K leads to the rate constants
collected in Table 4. Since the experimental activation energies
of reactions 11-13 and 3 agree within the combined error limits
(see Table 4), it is difficult to deduce a trend in the OO-N
bond energies of CX3O2NO2 when Cl atoms in the methyl group
are successively replaced by F atoms. Bearing in mind that
the absolute values of the rate constants are determined mainly

Thermal Decomposition of CF3O2NO2
J. Phys. Chem., Vol. 100, No. 16, 1996 6593
(14) Solomon, S.; Burkholder, J. B.; Ravishankara, A. R.; Garcia, R.
R. J. Geophys. Res. 1994, 99, 20929.
NO2 are formed and directly translate into errors of the
calculated equilibrium constants. Hence we conclude that
whenever the experimental error of a reaction enthalpy derived
from a van’t Hoff plot is a matter of only a few kJ mol-1, the
“second-law” analysis could be preferable or at least equivalent
to the “third-law” analysis. In these cases, it appears to be wise
to consider the low-frequency vibrations merely as adjustable
parameters which are useful to derive a self-consistent set of
thermochemical data.
(15) Nielsen, O. J.; Ellermann, T.; Sehested, J.; Bartkiewicz, E.;
Wallington, T. J.; Hurley, M. D. Int. J. Chem. Kinet. 1992, 24, 1009.
(16) Chen, J.; Zhu, T.; Niki, H. J. Phys. Chem. 1992, 96, 6115.
(17) 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. JPL Publ. 1994, 94-26.
(18) Deuflhard, P.; Nowak, U. Ber. Bunsen-Ges. Phys. Chem. 1986, 90,
940.
(19) Troe, J. J. Phys. Chem. 1979, 83, 114; Ber. Bunsen-Ges. Phys.
Chem. 1983, 87, 161.
Atmospheric Implications. In a recent publication by Ko
et al.,12 the implications of the CF3 radical chemistry for
stratospheric ozone were thoroughly discussed. Ko et al.
introduced CF3O2NO2 as a possible temporary reservoir of
CF3Ox which would be the major CF3X compound (CF3X )
CF3, CF3O, CF3O2, CF3OH, CF3OOH, CF3ONO2, CF3O2NO2,
CF3OOCl) in the lower stratosphere if its photolysis rate constant
were e10-4 s-1. From eq I and parameter set II it is estimated
that the thermal lifetime of CF3O2NO2 is very short at room
temperature and atmospheric pressure (ca. 1 min) but reaches a
lifetime of ca. 1 year at the temperatures and pressures of the
tropopause. For this reason, the lifetime τ of CF3O2NO2 in the
tropopause and the lower stratosphere is probably limited by
photolysis (τhν ≈ several days11 or even less46).
(20) Nielsen, J. R.; Liang, C. Y.; Smith, D. C. J. Chem. Phys. 1953, 21,
1060.
(21) Patrick, R.; Golden, D. M. Int. J. Chem. Kinet. 1983, 15, 1189.
(22) Zabel, F. Thermischer Zerfall Von Peroxynitraten; habilitation thesis,
Wuppertal, 1993.
(23) Gilbert, R. G.; Luther, K.; Troe, J. Ber. Bunsen-Ges. Phys. Chem.
1983, 87, 169.
(24) Woltz, P. J. H.; Jones, E. A.; Nielsen, A. H. J. Chem. Phys. 1952,
20, 378.
(25) Ruff, O.; Menzel, W.; Neumann, W. Z. Anorg. Allg. Chem. 1932,
208, 293.
(26) Jones, E. A.; Kirby-Smith, J. S.; Woltz, P. J. H.; Nielsen, A. H. J.
Chem. Phys. 1951, 19, 242.
(27) Klo¨ter, S.; Seppelt, K. J. Am. Chem. Soc. 1979, 101, 347.
(28) Chen, J.; Zhu, T.; Niki, H.; Mains, G. J. Geophys. Res. Lett. 1992,
19, 2215.
(29) Sehested, J.; Wallington, T. EnViron. Sci. Technol. 1993, 271, 46.
(30) Caralp, F.; Lesclaux, R.; Rayez, M. T.; Rayez, J. C.; Forst, W. J.
Chem. Soc., Faraday Trans. 2 1988, 84, 569.
Acknowledgment. Financial support of this work by the EC
(European Commission) under Contract EV5V-0024 and by the
Bundesminister fu¨r Bildung, Wissenschaft, Forschung und
Technologie (BMBF) is gratefully acknowledged.
(31) Butler, R.; Snelson, A. J. Phys. Chem. 1979, 83, 3243.
(32) Troe, J. J. Chem. Phys. 1977, 66, 4758.
(33) Reid, R. C.; Sherwood, T. K. The Properties of Gases and Liquids;
McGraw-Hill: New York, 1958.
(34) Destriau, M.; Troe, J. Int. J. Chem. Kinet. 1990, 22, 915.
(35) This rate constant for k3 was presented by Reimer and Zabel at the
9th International Symposium on Gas Kinetics, Bordeaux, 1986, but is
virtually identical to that of ref 7.
(36) Quack, M.; Troe, J. Ber. Bunsen-Ges. Phys. Chem. 1974, 78, 240.
(37) It should be noted that a reevaluation of the method applied in ref
34 in the most recent IUPAC evaluation (J. Phys. Chem. Ref. Data 1992,
21, 1125) led to k3 (280 K, 60 mbar of N2) ) 1.6 × 10-3 s-1 which is a
factor of 2 lower.
(38) Nickolaisen, S. L.; Friedl, R. R.; Sander, S. P. J. Phys. Chem. 1994,
98, 155.
(39) Bridier, I.; Caralp, F.; Loirat, H.; Lesclaux, R.; Veyret, B.; Becker,
K. H.; Reimer, A.; Zabel, F. J. Phys. Chem. 1991, 95, 3594.
(40) Campuzano-Jost, P.; Croce, A. E.; Hippler, H.; Siefke, M.; Troe,
J. J. Chem. Phys. 1995, 102, 5317.
References and Notes
(1) Hohorst, F. A.; DesMarteau, D. D. Inorg. Chem. 1974, 13, 715.
(2) Chen, J.; Young, V.; Zhu, T.; Niki, H. J. Phys. Chem. 1993, 97,
11696.
(3) Wallington, T. J.; Hurley, M. D.; Schneider, W. F. Chem. Phys.
Lett. 1993, 213, 442.
(4) The value for k(CF3O2NO2 w CF3O2 + NO2) in Table 1 of ref 3
should read 0.017 s-1 (Wallington, T. J., private communication, 1993);
the value g0.019 s-1 given here corresponds to the stated upper limit of
ref 3 for the lifetime of CF3O2NO2 in the presence of NO.
(5) Simonaitis, R.; Heicklen, J. Chem. Phys. Lett. 1979, 62, 473; 1979,
68, 245.
(6) Simonaitis, R.; Glavas, S.; Heicklen, J. Geophys. Res. Lett. 1979,
6, 385.
(41) Dewar, M. J. S.; Ford, G. P.; McKee, M. L.; Rzepa, H. S.; Thiel,
W.; Yamaguchi, Y. J. Mol. Struct. 1978, 43, 135.
(42) We thank Dr. G. Hirsch (Wuppertal) for the calculation of the
vibrational frequencies of CF3O2, using a MNDO program provided by
Prof. W. Thiel, Zu¨rich.
(43) Wilt, P. M.; Jones, E. A. J. Inorg. Nucl. Chem. 1968, 30, 2933.
(44) Buckley, P.; Weber, J. P. Can. J. Chem. 1974, 52, 942.
(45) Smardzewski, R. R.; Fox, W. B. J. Fluorine Chem. 1975, 6, 417.
(46) Libuda, H. G.; Zabel, F. Data presented at the International
Conference on Ozone in the Lower Stratosphere, May 15-20, 1995,
Halkidiki, Greece.
(7) Ko¨ppenkastrop, D.; Zabel, F. Int. J. Chem. Kinet. 1991, 23, 1.
(8) Xiong, J. Q.; Carr, R. W. J. Phys. Chem. 1994, 98, 9811.
(9) Zabel, F. Z. Phys. Chem. 1995, 188, 119.
(10) 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,
26A, 1805.
(11) Morel, O.; Simonaitis, R.; Heicklen, J. Chem. Phys. Lett. 1980,
73, 38.
(12) Ko, M. K. W.; Sze, N.-D.; Rodr´ıguez, J. M.; Weisenstein, D. K.;
Heisey, C. W.; Wayne, R. P.; Biggs, P.; Canosa-Mas, C. E.; Sidebottom,
H. W.; Treacy, J. Geophys. Res. Lett. 1994, 21, 101.
(13) Barnes, I.; Becker, K. H.; Fink, E. H.; Reimer, A.; Zabel, F.; Niki,
H. Int. J. Chem. Kinet. 1983, 15, 631.
JP953205G