The CF3C(O)O2 Radical. Its UV Spectrum, Self-Raction Kinetics, and Reaction with NO

Article abstract of DOI:10.1021/jp9532935

Flash photolysis combined with time-resolved UV spectroscopy and transient infrared absorption is used to investigate the reactions of CF3C(O)O2 with itself, CF3O2, and nitric oxide.The UV spectrum of CF3C(O)O2 exhibits two bands, the stronger short wavelength component of which has a maximum cross section of 7.1E-18 cm² at 207 nm.These bands are used to monitor the disappearance of CF3C(O)O2 and the secondary formation of CF3O2, yielding a self-reaction rate constant of (3.7 +5/-2)E-12 e^{(270 +/- 200)/T} cm³ s^-1.The cross reaction between CF3C(O)O2 and CF3O2 is found to be slow, having a rate constant of 3 s^-1.Transient IR monitoring of the loss of NO and concomitant formation of NO2 leads to a rate constant of (4.0 +2.2/-1.4)E-12 e^{(563 +/- 115)/T} cm³ s^-1 for the reaction between CF3C(O)O2 and NO.This result implies that CF3C(O)O2 radicals formed as intermediates in the atmospheric degradation of hydrofluorocarbons (HFCs) are rapidly converted into CF3O2 radicals, which are in turn converted into carbonyl fluoride and FNO.

Full text of DOI:10.1021/jp9532935

4514
J. Phys. Chem. 1996, 100, 4514-4520
The CF3C(O)O2 Radical. Its UV Spectrum, Self-Reaction Kinetics, and
Reaction with NO
M. Matti Maricq* and Joseph J. Szente
Research Laboratory, Ford Motor Company, P.O. Box 2053, Drop 3083, Dearborn, Michigan 48121
Gregory A. Khitrov† and Joseph S. Francisco
Department of Earth and Atmospheric Sciences and Department of Chemistry, Purdue UniVersity,
West Lafayette, Indiana 47907-1397
ReceiVed: NoVember 7, 1995; In Final Form: December 12, 1995X
Flash photolysis combined with time-resolved UV spectroscopy and transient infrared absorption is used to
investigate the reactions of CF3C(O)O2 with itself, CF3O2, and nitric oxide. The UV spectrum of CF3C(O)O2
exhibits two bands, the stronger short wavelength component of which has a maximum cross section of 7.1
× 10-18 cm2 at 207 nm. These bands are used to monitor the disappearance of CF3C(O)O2 and the secondary
+5
formation of CF3O2, yielding a self-reaction rate constant of (3.7 ) × 10-12 e(270(200)/T cm3 s-1. The cross
-2
reaction between CF3C(O)O2 and CF3O2 is found to be slow, having a rate constant of e2 × 10-12 cm3 s-1.
Transient IR monitoring of the loss of NO and concomitant formation of NO2 leads to a rate constant of
+2.2
(4.0 ) × 10-12 e(563(115)/T cm3 s-1 for the reaction between CF3C(O)O2 and NO. This result implies that
-1.4
CF3C(O)O2 radicals formed as intermediates in the atmospheric degradation of hydrofluorocarbons (HFCs)
are rapidly converted into CF3O2 radicals, which are in turn converted into carbonyl fluoride and FNO.
I. Introduction
(∼300 mJ in a 1-cm2 cross section beam) of either F2 or Cl2 in
a slowly flowing gas mixture containing trifluoroacetaldehyde,
oxygen, nitrogen, and, when required, nitric oxide. F2 is
preferred over Cl2 as a radical source, because fluorine atoms
react more rapidly with CF3CHO than do chlorine atoms;
however, F2 cannot be used in conjunction with NO because of
the spontaneous reaction
Amongst the atmospheric degradation products of the partially
halogenated hydrocarbons that are currently being used as
alternatives to the chlorofluorocarbons (CFCs) are carbonyl
compounds of the form CF3C(O)X, where X ) F, Cl, H. The
fully halogenated species are relatively unreactive with other
gas-phase atmospheric species and are primarily removed by
photolysis1 or by incorporation into cloud droplets.2 Trifluo-
roacetaldehyde, however, is susceptible to attack by hydroxyl
radicals to form CF3CO and water.3 Although CF3CO radicals
collisionally dissociate4 into CF3 + CO, the rate of bond
cleavage under atmospheric conditions is considerably smaller
than the rate of adding oxygen;5 thus, their principal fate is
conversion to trifluoroacetylperoxy radicals.
F2 + NO f F + FNO
Following photolysis of the halogen, the peroxy radicals are
formed via
X + CF3CHO f CF3CO + XH
(2)
(3)
Possible fates of CF3C(O)O2 induce reactions with NO, NO2,
and HO2. The lower bound of k1 > 1 × 10-11 cm3 s-1 reported
by Wallington et al.5 for the reaction
CF3CO + O2 + M f CF3C(O)O2 + M
where X ) F, Cl. Sufficient oxygen (∼15 Torr) is added to
the gas mixture to ensure that reaction 3 overwhelms CF3CO
dissociation and that it is essentially instantaneous on the >100-
µs time scale of the subsequent peroxy radical reactions. With
fluorine initiation (k2(F) ) 2.3 × 10-11 cm3 s-1 at 295 K),5
enough trifluoroacetaldehyde can be added to ensure that
reaction 2 is also rapid on this time scale. This is more difficult
to achieve for the slower chlorine initiation (k2(Cl) ) 2.4 ×
10-12 cm3 s-1 at 295 K);6 even with ∼15 Torr of CF3CHO, a
slight delay in CF3C(O)O2 formation is noted. The delay is
taken into account by including reaction 2 in the model used to
fit the data.
CF3C(O)O2 + NO f CF3C(O)O + NO2
(1)
suggests that this represents a significant removal process for
the peroxy radical. The present study derives temperature-
dependent rate constants for reaction 1 by monitoring in real
time both the disappearance of NO and the appearance of NO2.
The paper also presents results for the rate constant of the
CF3C(O)O2 self-reaction, which competes under our experi-
mental conditions with reaction 1. It is concluded that trifluo-
roacetaldehyde, a degradation product of some hydrofluorocar-
bon compounds, will not adversely affect the ozone layer.
Both time-resolved UV spectroscopy7,8 and transient IR
spectroscopy9 are used to interrogate in real time the reaction
progress. More complete descriptions of the two experimental
setups are available in refs 7 and 8 and in ref 9, respectively.
The loss of CF3C(O)O2 radicals via
II. Experimental Section
The formation of CF3C(O)O2 radicals, in concentrations of
(3-10) × 1014 cm-3, is initiated by the 351-nm photolysis
* Author for correspondence.
CF3C(O)O2 + CF3C(O)O2 f 2CF3C(O)O + O2 (4)
† Present address: Department of Chemistry, Eastern Michigan Univer-
sity.
X Abstract published in AdVance ACS Abstracts, February 15, 1996.
and the secondary generation of CF3O2 radicals and their
0022-3654/96/20100-4514$12.00/0 © 1996 American Chemical Society

The CF3C(O)O2 Radical Reaction with NO
subsequent loss via
J. Phys. Chem., Vol. 100, No. 11, 1996 4515
CF3C(O)O f CF3 + CO2
(5)
(6)
CF3 + O2 + M f CF3O2 + M
CF3C(O)O2 + CF3O2 f CF3C(O)O + CF3O + O2 (7)
CF3O2 + CF3O2 f 2CF3O + O2 (8)
are followed by observing the changing contribution of these
peroxy radicals to the UV spectrum of the reaction mixture.
Concentration vs time profiles of the individual peroxy species
are determined by fitting the spectra at various times following
the photolysis pulse to the expression
Figure 1. UV absorption spectrum of CF3C(O)O2. The spectrum (solid
line) consists of two bands: a strong feature with an absorption
maximum of 7.1 × 10-18 cm2 at 210 nm and a weaker band with an
onset of about 300 nm and a maximum that is obscured by the strong
band. The dashed line represents a fit of the spectrum to two Gaussian-
shaped absorption bands (see text).
Abs(λ,t) ) σCF C(O)O (λ)l[CF3C(O)O2]t +
3
2
σ
CF O (λ)l[CF3O2]t + c(t)Abs(λ,∞) (9)
3 2
where σn(λ) represents the optical cross section of species n
and l is the path length. The last term in eq 9 accounts for the
increasing contributions of stable products to the absorbance
as they accumulate. c(t), which is proportional to the concentra-
tion of products, is varied along with [CF3C(O)O2]t and [CF3O2]t
in order to achieve the best fit to the absorbance. Its affect on
the peroxy radical concentrations is included as part of the 2σ
fitting error.
The reaction between CF3C(O)O2 radicals and NO is studied
by measuring NO loss and NO2 formation using transient IR
absorption. This procedure was chosen in lieu of using UV
spectroscopy to monitor CF3C(O)O2, because it is more specific
to reaction 1. In contrast, peroxy radical loss occurs both via
reaction 1 and by the competing peroxy radical self-reaction.
NO is monitored via lines in the P branch of the V ) 0 f 1
transition in the vicinity of 1850 cm-1. Absorption features in
the R branch of the ν3 asymmetric stretch near 1630 cm-1 are
used to probe NO2. Radiation at the required IR wavelengths
is provided by a lead-salt diode laser, which is frequency locked
to the desired absorption line and directed through the reaction
vessel counter to the photolysis pulse. The IR light passes
through a monochromator for mode selection and to filter stray
light and is then detected by a HgCdTe detector with a 0.2-µs
response time.
of peroxy radicals. This was ensured by examining the
absorbance of the reaction mixture 20 µs after F2 photolysis as
a function of CF3CHO and O2 concentrations. The absorbance
increased with the addition of each reactant, reaching a plateau
for > 1 Torr of CF3CHO and > 10 Torr of O2, at which
concentrations the peroxy radical formation has an expected
half-life of ≈1 µs. Loss of CF3C(O)O2 by self-reaction occurs
with a rate constant of 7.9 × 10-12 cm3 s-1 at 295 K (see section
IIIB), implying that after 20 µs approximately 9% of the radicals
are converted to CF3O2, given a typical initial concentration of
3 × 1014 cm-3 CF3C(O)O2. Using the known UV spectrum12
of CF3O2, the measured absorbance at 20 µs is corrected to zero
time in order to extract the CF3C(O)O2 spectrum.
The formation of FO2 by the addition of O2 to fluorine atoms
(k ) 4.4 × 10-33 cm6 s-1 13
)
poses a possible source of
interference in the determination of the CF3C(O)O2 UV
spectrum, since it absorbs strongly, about 3 times that of a
typical peroxy radical, over the 190-225-nm region. This
interference is minimized by the use of moderate total pressures
(≈170 Torr) and sufficient quantities of CF3CHO (>2 Torr),
under which conditions [FO2]/[CF3C(O)O2] < 0.005.
Absolute absorption cross sections are assigned by comparing
the CF3C(O)O2 absorbance, corrected to zero time, to the
absorbance of ethylperoxy radicals8 recorded under identical
conditions, except for the substitution of ≈2 Torr of ethane for
≈2 Torr of CF3CHO. This radical is ideal for intensity
calibration because it is rapidly formed, but its removal via self-
reaction is very slow.14 Furthermore, three recent diode array
spectra of C2H5O2 yield cross sections in very good agreement
with each other;8,15,16 thus, its contribution to the error in the
intensity of the CF3C(O)O2 band is estimated at <5%. An
additional 5% error is allowed for the small correction of the
trifluoroacetylperoxy absorbance to zero time.
The UV spectrum of CF3C(O)O2 is illustrated in Figure 1
(solid line), and absorption cross sections over the 190-300-
nm range are collected in Table 1. The spectrum appears to
consist of two unresolved bands, analogous to what is ob-
served17,18 for CH3C(O)O2 (for which the bands are resolved).
It has been noted that the UV spectra of peroxy radicals such
as CH3O2 and C2H5O2 are nearly Gaussian in shape,19 with
CF3CHO was prepared by the method of Berney.10 It was
purified by several freeze-pump cycles, checked by comparing
its IR spectrum against those reported by Berney10 and by Dodd
et al.11 and stored at -196 °C. During the experiments, it was
kept at 0 °C to minimize its polymerization. Nitrogen (99.999%)
and oxygen (99.8%) were obtained from Michigan Airgas. Nitric
oxide (99%) was obtained from Matheson as a 10.1% mixture
in N2. Nitrogen, oxygen, and NO gas flows were controlled
by Tylan flow controllers, whereas the flows of CF3CHO, F2,
and Cl2 were set by needle valves. Flow rates were determined
by measuring the rate of pressure increase into a constant
volume. The total pressure was monitored at the entrance and
exit of the reaction cell. The cell temperature was maintained
by a Neslab ULT-80dd temperature regulator, with the gas
mixture precooled/preheated prior to entering the cell.
III. Results
A. UV Spectrum. The formation of CF3C(O)O2 radicals
proceeds in two steps: attack of CF3CHO by F atoms followed
by O2 addition to the resultant trifluoroacetyl radical.
σ(λ) ) σm exp(-R[ln(λm/λ)]2)
(10)
A
quantitative measurement of UV absorption cross sections
requires that these formation steps be fast compared to the loss
Here, λm marks the position of the band maximum, σm represents
the UV cross section at this position, and R provides a measure

4516 J. Phys. Chem., Vol. 100, No. 11, 1996
Maricq et al.
TABLE 1: CF3C(O)O2 UV Absorption Cross Section
cross section,
10-20 cm2
cross section
wavelength
wavelength
10-20 cm2
300
295
290
284
279
274
269
264
259
254
248
4
23
23
39
58
66
79
105
127
149
169
243
238
233
228
223
218
212
207
202
197
192
195
224
249
315
415
564
693
700
592
459
330
of the width of the band. Assuming the CF3C(O)O2 spectrum
to consist of two such bands leads to a very good fit of the
spectrum, as indicated by the dashed line in Figure 1. The
parameters σ1m ) 3.2 × 10-18 cm2, λ1m ) 207.4 nm, R1 ) 20.1
and σ2m ) 4.0 × 10-18 cm2, λ2m ) 209.2 nm, R2 ) 289 afford
the advantage of providing a compact representation of the UV
cross sections of CF3C(O)O2. The short-wavelength band
occurs at nearly the same location as for CH3C(O)O2 (206 nm);
however, the long-wavelength band, which occurs at 250 nm
for acetylperoxy radical, shifts to 209 nm for CF3C(O)O2 and
broadens considerably. This can be compared to aliphatic per-
oxy radicals, the UV spectra of which typically undergo a shift
to the blue upon fluorine or chlorine substitution.20 Also shown
in Figure 1 are the measurements of Wallington et al.5 (filled
circles). These are in good agreement with the present results,
except for the points at 230 and 240 nm. It is possible that
these small discrepancies occur due to the corrections for FO2
formation and for CF3CO dissociation that must be introduced
by Wallington et al., because of the large radical concentrations
and the 750 Torr of total pressure used in their study.
Figure 2. Time-resolved UV spectra following the photolysis of F2 in
the presence of CF3CHO and O2. The more rapid decay in the
absorbance of the long-wavelength band near 240 nm as compared to
the feature at 210 nm indicates the formation of secondary UV
absorbing species, in this case CF3O2.
B. Self-Reaction Kinetics. The loss of CF3C(O)O2 radicals
is followed by time-resolved UV spectroscopy. As shown in
Figure 2, the absorbance of the reaction mixture decays as a
function of time; however, the decay appears faster over the
230-260-nm range than for 200-230 nm. This behavior is
explained by the secondary formation of CF3O2 radicals via
reactions 4-6. The latter radical exhibits a single Gaussian-
shaped absorption feature centered at 210 nm12 but lacks the
long-wavelength tail characteristic of CF3C(O)O2. Thus, sec-
ondary formation of CF3O2 masks the decay of the (trifluoro-
acetyl)peroxy radical in the vicinity of 210 nm, but not at
240 nm.
Deconvolution of the time-resolved spectra according to eq
9 yields CF3C(O)O2 and CF3O2 concentration vs time profiles,
such as those illustrated in Figure 3. As expected from the
reaction model proposed in Table 2, the loss of trifluoroacetyl-
peroxy radicals is accompanied by the formation of CF3O2,
which itself subsequently decays. This behavior is satisfactorily
described by reactions 4, 7, 8, and
Figure 3. Time dependence of the CF3C(O)O2 and CF3O2 concentra-
tions as deconvolved from the time-resolved UV spectra. Error bars
represent 2σ of the fitting error. Solid lines illustrate best fits of the
reaction model to the data. Both sets of concentrations are fit
simultaneously.
TABLE 2: CF3C(O)O2 Self-Reaction Mechanism
reactiona
rate constant
4.
CF3C(O)O2 + CF3C(O)O2 f
2CF3C(O)O + O2
k ) 3.7 × 10-12e270/T
cm3 s-1 b
5.
6.
CF3C(O)O f CF3 + CO2
CF3 + O2 + M f
k > 5 × 104 s-1 b
k ) 4.0 × 10-12(T/300)-1
cm3 s-1
CF3O2 + M
7.
8.
CF3C(O)O2 + CF3O2 f
CF3C(O)O + CF3O + O2
CF3O2 + CF3O2 f
2CF3O + O2
k ) (0-2) × 10-12
cm3 s-1 b
k ) 1.8 × 10-12
cm3 s-1 20
11.
CF3O2 + CF3O f
CF3OOOCF3
k ) 1.4 × 10-11
cm3 s-1 20
CF3O + CF3CHO f
k < 2 × 10-14
cm3 s-1 b
CF3CO + CF3OH
CF3O + CF3O f products
k ) 1.4 × 10-11
cm3 s-1 21
CF3O2 + CF3O f CF3OOOCF3
(11)
a Reaction numbers correspond to those used in the text. Rate
constants are taken from ref 13 unless otherwise specified. b Measured
in the present study.
with the latter two reactions accounting for the CF3O2 decay.
Reaction 6, the addition of O2 to CF3, is sufficiently fast at the
concentrations of O2 employed in the experiments to be
essentially instantaneous. Although only a lower limit5 has been
reported for the rate of CF3C(O)O decomposition, a rough
estimate suggests that it, too, is sufficiently rapid to appear
instantaneous on our experimental time scale. Ab initio
computations by Francisco22 reveal the dissociation to proceed
with an activation energy of 4.9 kcal/mol, which, when
combined with an A factor of 2 × 1013 s-1, leads to an estimated
high-pressure limiting rate of 5 × 109 s-1 at 295 K. The present
data are consistent with this estimate; a lower limit of 5 × 104
s-1 can be established for the dissociation rate at 150 Torr of
N2 from the lack of an observable delay in the CF3O2 rise
relative to the CF3C(O)O2 decay.
Temporal profiles of the peroxy radicals, such as those
depicted in Figure 3, were compared to predictions from the

The CF3C(O)O2 Radical Reaction with NO
J. Phys. Chem., Vol. 100, No. 11, 1996 4517
TABLE 3: Self-Reaction Rate Constants
conditions
temp, CF3CHO, F2,
O2, Ptot,
[F]0,
results
K
Torr
Torr Torr Torr 1014 cm-3 k4 10-11 cm3 s-1
210
213
233
253
273
295
3.2
3.7
3.6
4.3
4.4
3.3
3.3
3.2 11
3.5
3.9
3.9
3.9
8.3 151
148
9.0 159
9.2 170
9.4 172
9.2 175
3.8
3.6
3.3
7.8
7.7
3.6
1.2 ( 0.5
1.4 ( 0.4
1.1 ( 0.3
1.2 ( 0.3
1.1 ( 0.3
0.8 ( 0.2
model of Table 2, treating the rate constants for CF3C(O)O2
self-reaction and its cross reaction with CF3O2 as adjustable
parameters. In all cases, the best fit to the data is obtained with
a rate constant for the cross reaction of k7 ) 0. After accounting
for possible systematic errors, an upper limit of k7 e 2 × 10-12
cm3 s-1 is reported. Rate constants for the self-reaction, k4,
are tabulated in Table 3 along with the experimental conditions.
The overall 2σ error bars range from 20% to 40%. The major
fraction (15-30%) arises from noise in the data (fitting error),
which originates from separating the overlapping contributions
of CF3C(O)O2 and CF3O2 to the overall absorption of UV light
by the reaction mixture. Smaller contributions to the error bars
of about 10% each arise from uncertainties in the CF3C(O)O2
and CF3O2 UV cross sections and from the initial radical con-
centrations. As described in section IIIA, the latter is ascertained
by calibration against the production of ethylperoxy radicals.
The temperature dependence of the CF3C(O)O2 self-reaction
Figure 4. Variation of CF3C(O)O2 self-reaction rate constant with
temperature. Error bars and dotted lines represent deviations of 2σ.
rate constant is illustrated in Figure 4. It has at best a slight
+5
negative temperature dependence given by k4 ) (3.7 ) ×
-2
10-12 e(270(200)/T cm3 s-1
. Both the magnitudes of the rate
constants and their variation with temperature are typical of
peroxy radical self-reactions.14 For comparison, the room-
temperature rate constant is about 50% smaller than that of the
analogous CH3C(O)O2 self-reaction.17,18
C. Reaction of CF3C(O)O2 with NO. The reaction between
CF3C(O)O2 and NO was followed by IR probes of NO loss
and NO2 formation. Figure 5 illustrates the concentration vs
time dependence of these species at 324 K and for two initial
NO concentrations. The NO and NO2 traces were obtained in
“back-to-back” experiments run under identical conditions; only
the frequency of the diode laser was adjusted. As expected,
the rate of NO2 formation matches the rate of NO loss, and the
rates are faster at the higher initial NO concentration. Of note
are the relative values of the initial radical concentration, the
amount of NO lost, and the NO2 formed. At the higher [NO]0
level in Figure 5, [NO2]∞ > [Cl]0 and ∆[NO]∞ = -2[Cl]0;
clearly additional chemistry besides reaction 1 plays a role in
determining the NOx levels.
Figure 5. Time dependence of NO loss and NO2 formation for the
reaction of CF3C(O)O2 with NO. Both sets of data are fit simulta-
neously, treating k1 and k5 as adjustable parameters.
self-reaction and cross reactions, (b) loss of peroxy radicals by
reaction with chlorine atoms, (c) reactions of the peroxy radicals
CF3C(O)O2 and CF3O2 with the NO2 formed by reactions 1
and 12, and (d) a competition between the dissociation of
CF3C(O)O and its reaction with NO.
The first two possibilities involve reactions that compete with
the reaction between CF3C(O)O2 and NO. The second of these,
Nitric oxide is involved in two secondary reactions. The
CF3C(O)O product of the primary CF3C(O)O2 + NO reaction
rapidly dissociates into CF3 and CO2. The former fragment adds
oxygen to generate a secondary peroxy radical which itself
converts NO to NO2:
Cl + RO2 f RO + ClO
(14)
is effective when peroxy radical formation is limited by Cl attack
of the precursor, in this case CF3CHO, as opposed to the O2
addition step, reaction 3. When this occurs, peroxy radicals
and chlorine atoms are present simultaneously and are removed
via reaction 14, which is typically very fast.28 Both a and b
can explain a decrease in ∆[NO]∞ and [NO2]∞ relative to [Cl]0;
however, neither can effectively rationalize a higher than
expected ratio of ∆[NO]∞ to [NO2]∞. Furthermore, since these
reactions compete with reaction 1, they should become less
effective at influencing the NO and NO2 levels as [NO]0 is
raised; this is not observed.
CF3O2 + NO f CF3O + NO2
(12)
Furthermore, the trifluoromethoxy radical that is formed sub-
sequently reacts with nitric oxide
CF3O + NO f CF2O + FNO
(13)
to generate carbonyl fluoride and FNO. Thus, one expects from
reactions 1, 12, and 13 to observe a [Cl]0:∆[NO]∞:[NO2]∞ ratio
of 1:-3:2. Instead, the observed ratio is closer to 1:-2.3:1.3.
Possible reasons are (a) removal of peroxy radicals via their
Possibilities c and d can explain a smaller than expected NO2
yield as compared to NO loss. According to c, NO2 already

4518 J. Phys. Chem., Vol. 100, No. 11, 1996
Maricq et al.
TABLE 4: CF3C(O)O2 + NO Reaction Mechanism
reactiona
rate constant
Initiation
2.
3.
Cl + CF3CHO f CF3CO + HCl
k ) 6.1 × 10-11e-955/T cm3 s-1 23
k ) 7.3 × 10-13 cm3 s-1 5
CF3CO + O2 + M f CF3C(O)O2 + M
Self-Reactions
CF3C(O)O2 + CF3C(O)O2 f 2CF3C(O)O + O2
4.
5.
6.
7.
8.
k ) 3.7 × 10-12e270/T cm3 s-1 b
k > 5 × 104 s-1 b
CF3C(O)O f CF3 + CO2
CF3 + O2 + M f CF3O2 + M
k ) 4.0 × 10-12(T/300)-1 cm3 s-1
k ) (0-2) × 10-12 cm3 s-1 b
k ) 1.8 × 10-12 cm3 s-1 20
k ) 1.4 × 10-11 cm3 s-1 20
k < 2 × 10-14 cm3 s-1 b
CF3C(O)O2 + CF3O2 f CF3C(O)O + CF3O + O2
CF3O2 + CF3O2 f 2CF3O + O2
CF3O2 + CF3O f CF3OOOCF3
CF3O + CF3CHO f CF3CO + CF3OH
CF3O + CF3O f products
11.
k ) 1.4 × 10-11 cm3 s-1 21
NOx Reactions
CF3C(O)O2 + NO f CF3C(O)O + NO2
CF3O2 + NO f CF3O + NO2
1.
12.
15.
k ) 4.0 × 10-12e(563)/T cm3 s-1 b
k ) 1.5 × 10-11(T/298)-1.2 cm3 s-1 20
k ) 3 × 10-11 cm3 s-1 24
CF3C(O)O + NO f products
CF3C(O)O2 + NO2 f CF3C(O)O2NO2
CF3O2 + NO2 f CF3O2NO2
k ) 6.6 × 10-12(T/298)-1 cm3 s-1 25
k ) 8.9 × 10-12(T/298)-0.72 cm3 s-1 20
k ) 4.2 × 10-11e120T cm3 s-1 26,27
13.
CF3O + NO f CF2O + FNO
RO2 + Cl
Cl + CF3C(O)O2 f CF3C(O)O + ClO
Cl + CF3O2 f CF3O + ClO
14.
14.
k ) 1 × 10-10 cm3 s-1 (see text)
k ) 1 × 10-10 cm3 s-1 (see text)
a Reaction numbers correspond to those used in the text. Rate constants are taken from ref 13 unless otherwise specified. b Measured in the
present study.
TABLE 5: CF3C(O)O2 + NO Reaction Constants
conditions
Ptot Torr
results
k1, 10-11 cm3 s-1
temp, K
CF3CHO, Torr
Cl2, Torr
O2, Torr
[Cl]0 1014 cm-3
NO, 1015 cm-3
k5, 105 s-1
220
220
254
254
295
295
324
324
20
20
17
17
16
16
12
12
0.29
0.29
0.28
0.28
0.28
0.28
0.34
0.34
12
12
12
12
12
12
13
13
104
104
100
100
102
102
101
101
8.4
8.4
6.8
6.8
5.7
5.7
7.5
7.5
6.7
12.4
5.7
11.0
5.2
9.7
4.8
5.6 ( 1.0
5.3 ( 1.3
3.4 ( 0.6
3.3 ( 0.8
2.7 ( 0.5
2.8 ( 0.5
2.6 ( 0.5
2.1 ( 0.4
1.3
0.9
1.9
0.9
2.3
2.5
2.2
1.9
9.3
formed is removed by reaction with peroxy radicals, thereby
lowering its yield relative to ∆[NO]∞. By explanation d, the
secondary peroxy radical CF3O2 is not formed. Instead, CF3C-
(O)O is removed by reaction with NO, whereby the ratio of
∆[NO]∞ to [NO2]∞ is -2:1 rather than -3:2. Both mechanisms
lead to good fits of the ∆[NO]t and [NO2]t data, yielding
essentially the same values of k1. However, the rate constant
required for the reaction between CF3C(O)O2 and NO2 in order
to obtain a good fit to the data using explanation c varies from
5 to 30 times the accepted value.
5 shows, the values of k5 derived in this manner are consistent
with the lower limit of 5 × 104 s-1 found above from the
CF3C(O)O2 self-reaction experiments and the lower limit of 6
× 104 s-1 reported by Wallington et al.5
Contributions to the error in k1 arise from signal noise, the
measurement of [Cl]0, uncertainties in σir(NO) and σir(NO2),
and uncertainties in the reaction model, primarily in k2, k4, and
k14. Noise in the absorption vs time traces introduces ap-
proximately a 4-20% error into k1. A 5% uncertainty in [Cl]0,
arising from the ethylperoxy UV cross section, contributes a
9% error to the rate constant. Uncertainty in the IR cross section
of NO adds another 10% to the error in k1, whereas the
contribution from σir(NO2) is about 3% (it has a larger effect
on k5). A delay in CF3C(O)O2 formation owing to the slow
attack of CF3CHO by Cl was found to lead to underprediction
of k1. For example, reduction of CF3CHO from 16 to 4.3 Torr
leads to an apparent decrease of k1 from 2.8 × 10-11 to 1.5 ×
10-11 cm3 s-1 at 295 K. At the higher CF3CHO concentration,
a 50% uncertainty in k2 leads to an error of <10% in k1. Owing
to the relatively large values of [NO]0 and the high CF3CHO
concentrations, uncertainties of 30% in the CF3C(O)O2 self-
reaction and 100% in the Cl + RO2 rate constants contribute
only about 3% each to the overall error. Combining statistically
the various contributions yields overall 2σ error bars in the range
of (20%.
We find that the observed NOx changes are best explained
by d. Thus, we postulate that the reaction
CF3C(O)O + NO f products
(15)
competes with CF3C(O)O dissociation and fit simultaneously
the NO loss and the NO2 formation data to the model in Table
4, treating k1 and k5 as fitting parameters. [Cl]0 is measured
separately by substituting ethane for CF3CHO, as explained in
section IIIA. The results are collected in Table 5, and
representative fits are illustrated by the solid lines in Figure 5.
The slight delay observed in NO2 formation may be due to a
fraction being formed in excited vibrational states, which rapidly
relax via collisions to the ground state. However, the ability
to fit successfully both the NO and NO2 traces with the same
values of k1 and k5 adds confidence to the reaction model
adopted in Table 4. Because it is based on an assigned rate
constant for reaction 15,24 k5 must be considered as a relative
as opposed to an absolute value. However, as reference to Table
The room-temperature rate constant of k1 ) 2.8 × 10-11 cm3
s-1 derived in the present study for the reaction between
CF3C(O)O2 and NO is considerably higher than the lower limit

The CF3C(O)O2 Radical Reaction with NO
J. Phys. Chem., Vol. 100, No. 11, 1996 4519
atm of SF6.25 Taking this as an upper limit to the rate constant
under atmospheric conditions and assuming [NO2] ) 4 × 108
cm-3 provides a lifetime of >6 min for CF3C(O)O2 with respect
to reaction with NO2; assuming the rate constant to remain
constant with altitude, the lifetime at 20 km is about 2 min.
The rate constant for the reaction between HO2 and
CF3C(O)O2 is unknown; as an estimate, we assume at 295 K
the value of 1.4 × 10-11 cm3 s-1 from comparison to
CH3C(O)O2. Using an HO2 concentration29 of 2 × 108 cm-3
leads to a lifetime of about 6 min. If the temperature
dependence of CF3C(O)O2 + HO2 mimics that of CH3C(O)O2,
then at 20 km, where the HO2 concentration has fallen to 5 ×
106 cm-3, the lifetime of CF3C(O)O2 with respect to HO2
increases to 60 min. The only CF3C(O)O2-peroxy radical
reaction that has been measured is the self-reaction. If the rate
constant of ∼1.5 × 10-11 cm3 s-1 is taken as typical and if the
atmospheric RO2 concentration is assumed to be comparable
to the HO2 concentration, then the lifetime for removal by RO2
reactions will also be roughly 6 min.
Comparing the above atmospheric lifetimes indicates that,
in the lower troposphere CF3C(O)O2 radicals will mainly be
removed by reaction with NO, but the reactions with NO2, HO2,
and RO2 will contribute significantly. Based on the presently
available data, approximately 55% of the removal will be by
NO, 15% by NO2, 15% by HO2, and 15% by RO2. In contrast,
the contributions from HO2 and RO2 will decrease significantly
at 20 km, where 85% of the removal is expected to by via NO
reaction and 15% by NO2.
The CF3C(O)O radical that is produced in reaction 1 rapidly
dissociates. Even at 233 K, the self-reaction data and the NO
reaction experiments are consistent with a dissociation rate of
>5 × 104 s-1. This can be compared with a possible reaction
between CF3C(O)O and NO, for which, even assuming a rate
constant of 5 × 10-11 cm3 s-1, the atmospheric rate would be
of the order 10-2 s-1. Thus, in the atmosphere, NO effectively
converts CF3C(O)O2 radicals into CF3O2 radicals and CO2. The
chemistry of CF3O2, and the CF3O radical that is subsequently
formed, has itself been the subject of intense scrutiny over the
past few years due to suggestions that these radicals can
participate in ozone depletion cycles.20 However, subsequent
work has shown that CF3O2 radicals are principally converted,
via CF3O, to CF2O and FNO;26,27 thus, the major fate of
CF3C(O)O2 is conversion into CO2, CF2O, and FNO.
Figure 6. Temperature dependence of the CF3C(O)O2 + NO reaction
rate constant. Error bars and dotted lines represent deviations of 2σ.
of k1 > 9.9 × 10-12 cm3 s-1 reported from the pulsed radiolysis
work of Wallington et al.5 As illustrated by Figure 6, the
reaction exhibits a negative temperature dependence with k1 )
+22
(4.0 ) × 10-12 e(563(115)/T cm3 s-1. This dependence sug-
-1.4
gests that the reaction proceeds via complex formation followed
by rapid rearrangement and dissociation. While an increase in
rate constant with declining temperature is generally observed
for peroxy radical-NO reactions, the typical dependence, with
EA ∼ 250 K, is more modest than found here.14
IV. Atmospheric Implications
The atmospheric oxidation of HFCs and HCFCs initiated by
the reaction of these compounds with OH radicals generates a
variety of halogen-substituted aldehydes and acid aldehydes,
amongst them CF3CHO. This molecule is itself subject to OH
attack, yielding a series of CF3COx radicals. The purpose of
this, and a previous paper,4 has been to investigate the kinetics
of reactions relevant to the atmospheric fate of these radicals.
The CF3CO radical formed by the reaction of OH with CF3-
CHO collisionally dissociates with an activation energy of
approximately 12 kcal/mol. Thus, the dissociation rate constant
falls from a value of about 105 s-1 at the earth’s surface (T )
295 K, P ) 760 Torr) to roughly 300 s-1 at an altitude of 20
km (T ) 215 K, P ) 45 Torr). The dissociation process
competes with oxygen addition to CF3CO to produce the
corresponding peroxy radical. While the latter rate constant is
unknown, oxygen addition to alkyl radicals is typically of the
order 4 × 10-12 cm3 s-1. This reaction, too, is temperature
and pressure dependent, but both dependencies are expected to
be modest compared to those for the dissociation process.
Furthermore, the expected negative temperature dependence
should cancel the effect of decreasing pressure, rendering the
rate constant relatively independent of altitude. A comparison
of the dissociation rate to the O2 addition rate reveals that
∼99.5% of the CF3CO radicals are converted to peroxy radicals
at 0 km and >99.9% are converted at 20 km.
V. Conclusion
As typical of peroxy radicals, CF3C(O)O2 exhibits a strong
UV absorption in the range 190-300 nm. The spectrum
consists of two Gaussian-shaped bands, such as is observed for
CH3C(O)O2; however, in this case, the band centers are almost
coincident at 207 and 209 nm, but with the latter band much
broader than the first. The long-wavelength tail of the
CF3C(O)O2 spectrum allows it to be distinguished from CF3O2
produced by self-reaction of the former radicals. Time-resolved
UV spectroscopy provides a means of monitoring the progress
of the self-reaction and of the cross reaction between
CF3C(O)O2 and CF3O2 radicals. The self-reaction exhibits a
slight negative temperature dependence typical of peroxy
radicals. In contrast, the cross reaction is significantly slower
than expected.
As observed by both the loss of NO and the production of
NO2, CF3C(O)O2 reacts rapidly with NO, with rate constants
ranging from 2.4 × 10-11 cm3 s-1 at 324 K to 5.5 × 10-11 cm3
s-1 at 220 K. Consideration of the atmospheric concentrations
of the principal species involved in peroxy radical removal
reactions, NO, NO2, HO2, and RO2, along with the available
Peroxy radicals are removed from the atmosphere by reactions
with NO, NO2, HO2, and other peroxy radicals. Assuming a
tropospheric NO concentration29 of 4 × 108 cm-3 leads to an
atmospheric lifetime of ∼1.5 min for CF3C(O)O2 with respect
to removal by NO; the lifetime falls to about 0.3 min at 20 km,
with [NO] = 109 cm-3 and k1 ) 5 × 10-11 cm3 s-1. For
comparison, the reported rate constant for the reaction of
CF3C(O)O2 with NO2 is 6.6 × 10-12 cm3 s-1 at 295 K and 1

4520 J. Phys. Chem., Vol. 100, No. 11, 1996
Maricq et al.
(15) Bauer, D.; Crowley, J.; Moortgat, G. K. J. Photochem. Photobiol.
A: Chem. 1992, 65, 329.
rate constants (or estimates), suggests that NO plays the major
role in the fate of CF3C(O)O2 in the lower troposphere and
dominates its removal in the stratosphere.
(16) Fenter, F. F.; Catoire, V.; Lesclaux, R.; Lightfoot, P. D. J. Phys.
Chem. 1993, 97, 3530.
(17) Moortgat, G.; Veyret, B.; Lesclaux, R. J. Phys. Chem. 1989, 93,
2362.
References and Notes
(18) Maricq, M. M.; Szente, J. J. J. Phys. Chem. 1996, 100, 4507.
(19) Joens, J. A. J. Phys. Chem. 1994, 98, 1394.
(1) Maricq, M. M.; Szente, J. J. J. Chem. Phys. 1995, 99, 4554.
(2) George, Ch.; Saison, J. Y.; Ponche, J. L.; Mirabel, Ph. J. Phys.
Chem. 1994, 98, 10857.
(3) Dobe, S.; Kachatryan, L. A.; Berces, T. Ber. Bungenges. Phys.
Chem. 1989, 93, 847. Scollard, D. J.; Treacy, J. J.; Sidebottom, H. W.;
Balestra-Garcia, C.; Laverdet, G.; Lebras, G.; MacLeod, H.; Teton, S. J.
Phys. Chem. 1993, 97, 4683.
(4) Maricq, M. M.; Szente, J. J.; Khitrov, G. A.; Dibble, T. S.;
Francisco, J. S. J. Phys. Chem. 1995, 99, 11875.
(5) Wallington, T. J.; Hurley, M. D.; Nielsen, O. J.; Sehested, J. J.
Phys. Chem. 1994, 98, 5686.
(6) Wallington, T. J.; Hurley, M. D. Int. J. Chem. Kinet. 1993, 25,
819.
(7) Maricq, M. M.; Szente, J. J. J. Phys. Chem. 1992, 96, 10862.
(8) Maricq, M. M.; Wallington, T. J. J. Phys. Chem. 1992, 96, 986.
(9) Maricq, M. M.; Szente, J. J.; Kaiser, E. W. J. Phys. Chem. 1993,
97, 7970.
(10) Berney, C. V. Spectrochim. Acta, Part A, 1969, 25, 793.
(11) Dodd, R. E.; Roberts, H. L.; Woodward, C. A. J. Chem. Soc. 1957,
2783.
(12) Maricq, M. M.; Szente, J. J. J. Phys. Chem. 1992, 96, 4925.
(13) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampton, R. F.;
Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina,
M. J. Chemical Kinetics and Photochemical Data for Use in Stratospheric
Modeling. JPL Publication 92-20, 1992.
(20) Francisco, J. S.; Maricq, M. M. AdV. Photochem. 1995, 20, 79.
(21) Taken as “typical” alkoxy self-reaction rate constant.
(22) Francisco, J. S. Chem. Phys. Lett. 1992, 191, 7.
(23) Based on the ab initio activation energy of 1.9 kcal/mol from
Dibble, T. S.; Francisco, J. S. Chem. Phys. Lett. 1993, 215, 409, and on
the room-temperature rate constant reported by Wallington and
Hurley.6
(24) Taken by comparison with the rate constant for FC(O)O + NO:
Maricq, M. M.; Szente, J. J.; Dibble, T. S.; Francisco, J. S. J. Phys. Chem.
1994, 98, 12294.
(25) Wallington, T. J.; Sehested, J.; Nielsen, O. J. Chem. Phys. Lett.
1994, 226, 563.
(26) Turnipseed, A. A.; Barone, S. B.; Ravishankara, A. R. J. Phys.
Chem. 1994, 98, 4594.
(27) Dibble, T. S.; Maricq, M. M.; Szente, J. J.; Francisco, J. S. J. Phys.
Chem. 1995, 99, 17394.
(28) Maricq, M. M.; Szente, J. J.; Kaiser, E. W.; Shi, J. J. Phys. Chem.
1994, 98, 2083.
(29) Cantrell, C. A.; Shetter, R. E.; Calvert, J. G.; Parrish, D. D.;
Fehsenfeld, F. C.; Goldan, P. D.; Kuster, W.; Williams, E. J.; Westberg,
H. H.; Allwine, G.; Martin, R. J. Geophys. Res. 1993, 98, 18355.
(14) 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.
JP9532935