The reaction of photochemically generated CF3O radicals with CO

Article abstract of DOI:10.1002/(SICI)1097-4601(1997)29:8<579::AID-KIN3>3.0.CO;2-R

CF₃O₂CF₃ was photolyzed at 254 nm in the presence of CO in 760 torr N₂ or air at 296 K in a static reactor. In N₂, the products CF₃OC(O)C(O)OCF₃ and CF₃OC(O)O₂C(O)OCF₃ were detected by FTIR spectroscopy. In air, the only observed products were CF₂O and CO₂ and a chain process, initiated by CF₃O, was invoked for the conversion of CO to CO₂. From both product studies, a mechanism for the CF₃O initiated oxidation of CO was derived, involving the addition reaction CF₃O₂ + CO → CF₃OC(O). The rate constant for the reaction CF₃O + CO at 296 K at a total pressure of 760 torr air was determined to be k(CF₃O + CO) = (5.0 ± 0.9) ± 10^-14 cm³ molecule^-1 s^-1.

Full text of DOI:10.1002/(SICI)1097-4601(1997)29:8<579::AID-KIN3>3.0.CO;2-R

The Reaction of
Photochemically Generated
CF3O Radicals with CO
RICHARD MELLER, GEERT K. MOORTGAT
Max-Planck-Institut für Chemie, Atmospheric Chemistry Department, Postfach 3060, 55020 Mainz, Germany
Received 10 July 1996; accepted 5 December 1996
ABSTRACT: CF3O2CF3 was photolyzed at 254 nm in the presence of CO in 760 torr N2 or air at
296 K in a static reactor. In N2 , the products CF3OC(O)C(O)OCF3 and CF3OC(O)O2C(O)OCF3
were detected by FTIR spectroscopy. In air, the only observed products were CF2O and
CO2 and a chain process, initiated by CF3O, was invoked for the conversion of CO to CO2 .
From both product studies, a mechanism for the CF3O initiated oxidation of CO was derived,
involving the addition reaction CF3O2 ϩ CO
CF3OC(O). The rate constant for the reaction
CF3O ϩ CO at 296 K at a total pressure of 760 torr air was determined to be k(CF3O ϩ CO) ϭ
(5.0 Ϯ 0.9) ϫ 10Ϫ14 cm3 moleculeϪ1
s
Ϫ1. © 1997 John Wiley & Sons, Inc. Int J Chem Kinet 29:
579–587, 1997.
with CO under atmospheric conditions received very
little attention. To our knowledge only three kinetic
INTRODUCTION
studies were recently undertaken [23–25]. Saathoff
and Zellner [24] studied reaction (3) at 298 K in a
LP/LIF system.
The interest in the chemistry of CF3O radicals with
different atmospheric trace gases increased strongly
since different hydrochlorofluorocarbons (HCFCs)
and hydrofluorocarbons (HFCs) containing a CF3
group were proposed to replace the fully halogenated
chlorofluorocarbons (CFCs). In the atmosphere CF3
radicals formed in the OH-initiated degradation of
HCFCs or HFCs will react with oxygen and conse-
quently with NO to form CF3O radicals [1–3].
CF3O ϩ CO
products
(3)
They reported a value for k3 ϭ 4.4 ϫ 10Ϫ14 cm3 mol-
eculeϪ1 sϪ1 at 298 K in 50 torr He. With the similar
technique Turnipseed et al. [23] measured the temper-
ature dependence of k3 over the range 333–240 K.
These authors found k3 to be pressure dependent and
determined a high-pressure limiting rate coefficient
in SF6 to be k3,ϱ(298 K) ϭ (6.8 Ϯ 1.2) ϫ 10Ϫ14 cm3
moleculeϪ1 sϪ1. Using IR absorption and a relative
rate technique, Wallington and Ball [25] reported
values of k3(700 torr) ϭ (7.2 Ϯ 0.7) ϫ 10Ϫ14 cm3
moleculeϪ1 sϪ1 and k3(100 torr) ϭ (4.6 Ϯ 0.5) ϫ
10Ϫ14 cm3 moleculeϪ1 sϪ1 for the reaction of CF3O
with 13CO in air. In the latter two investigations, at-
tempts were made to distinguish between the three
possible exothermic reaction channels:
CF3 ϩ O2 ϩ M
CF3O2 ϩ NO
CF3O2 ϩ M
(1)
(2)
CF3O ϩ NO2
Even though very intensive studies were undertaken
on the reaction of CF3O with NO [4–8], NO2 [9], O3
[7, 10–15], saturated and unsaturated hydrocarbons
[5,8,16–21], and H2O [22,23], the reaction of CF3O
Correspondence to: G. Moortgat
© 1997 John Wiley & Sons, Inc.
CCC 0538-8066/97/080579-09

580
MELLER AND MOORTGAT
ϩM
RESULTS AND DISCUSSION
CF3O ϩ CO
CF3OC(O)
⌬H° Ϸ Ϫ35 kcal molϪ1
(3a)
Product Studies of the Photolysis of
CF3O2CF3 in the Presence of CO
CF3 ϩ CO2
⌬H° Ϸ Ϫ22.6 kcal molϪ1 (3b)
In this study it is anticipated that photolysis of
CF3O2CF3 at 254 nm results only in the formation of
two CF3O radicals according to reaction (4a) [14,27].
Turnipseed et al. [23] proposed a contribution of
0.08 Ϯ 0.06 for channel (4b) to the total loss of
CF3O2CF3 at 248 nm.
CF2O ϩ FCO
⌬H° Ϸ Ϫ10.6 kcal molϪ1 (3c)
In a static photolysis reactor photolysis experiments
of CF3O2CF3 at ␭ ϭ 254 nm in the presence of CO
in N2 /O2 mixtures were carried out at a total pres-
sure of 760 torr and at 296 K. From product studies
using FTIR absorption spectroscopy, the mechanism
for the CF3O induced oxidation of CO was de-
rived. From the observed loss rate of CO and the
rate of the recombination of CF3O radicals in the
photolysis of CF3O2CF3 in the presence of CO in 760
torr air, the rate constant for reaction (3) was deter-
mined.
CF3O2CF3 ϩ hv
2 CF3O
(4a)
CF3O ϩ CF2O ϩ F (4b)
In order to establish the photolysis rate of CF3O2CF3
under the experimental conditions, mixtures of
CF3O2CF3/C2H6 /air were irradiated. Similar experi-
ments are described elsewhere [14]. A photolysis rate
constant of k4a ϭ (1.20 Ϯ 0.02) ϫ 10Ϫ5 sϪ1 was ob-
tained in that way.
Mixtures of (0.3–5.0) ϫ 1015 molecule cmϪ3
CF3O2CF3 and (0.3–3.6) ϫ 1015 molecule cmϪ3 CO
were irradiated in 760 torr N2 . Traces of O2 were
present in small amounts in the photolysis cell due to
impurities in the N2 used and/or through very small
leakages. Since these traces of O2 could not be quan-
tified only mechanistic information could be evalu-
ated from these experiments.
After the irradiation of CF3O2CF3 in the presence
of CO in N2 , the typical absorption bands of CF2O
(772, 920–990, and 1880–1980 cmϪ1) and CO2 (668
and 2280–2390 cmϪ1) were observed. In addition,
bands of another product were observed at 826, 1081,
1117, 1184, 1258, 1286, 1823, and 1845 cmϪ1. In
Figure 1 the IR-spectra of this product is presented in
EXPERIMENTAL
The experimental set-up used to carry out these mea-
surements is described elsewhere [26] and is only
briefly presented here. Photolysis experiments were
performed at ambient temperatures (296 K) in a long-
path multiple reflection cell (43.2 m optical path) sur-
rounded with 7 UV-lamps (Philips). The reactor is
made of quartz glass and is coupled to a Bomem DA8
series FTIR interferometer. Infrared emission from a
globar source traversed the KBr beamsplitter of the
spectrometer, and the interferenced signal was moni-
tored with a Cu–Ge detector, cooled to liquid He
temperature. Spectra recorded with a spectral resolu-
tion of 1 cmϪ1 between 450 and 4000 cmϪ1 were
taken before the photolysis and after each irradiation
interval. The total photolysis time varied between
1800 and 3600 s, during which most of the initial CO
was consumed.
Reactant and product concentrations were cali-
brated against reference spectra from pure samples of
CO (Linde, purity ϭ 99.999%), CF3O2CF3 (Fluo-
rochem, purity Ͼ98%), and CF2O (PCR, purity
Ͼ97%) recorded under the same conditions in the
same experimental set-up. Both fluorinated samples
were purified by bulb to bulb distillation before use.
No contaminations could be identified in the IR-ab-
sorption spectra. CO was used without purification.
Used bath gases were N2 (Linde, purity 99.9996%)
and synthetic air (hydrocarbon free).
Figure 1 IR spectrum of CF3OC(O)C(O)OCF3 in the
range 800–2000 cmϪ1
.

PHOTOCHEMICALLY GENERATED CF3O RADICALS
581
the range 800–2000 cmϪ1. This substance was iso-
lated by first freezing the reaction mixture at 77 K,
and then warming it up to 233 K to boil off
CF3O2CF3 and the products CO2 and CF2O. At room
temperature an additional product evaporated. Also
the absorbances at 1034 cmϪ1 due to SiF4 and the IR
features between 1300 and 2000 cmϪ1 due to H2O
can be seen in this figure.
This product was identified to be bis-perflu-
oromethyl-oxalate (CF3OC(O)C(O)OCF3), which has
been previously observed by Varetti and Aymonino
[28]. They reported IR-bands at 604 (CF3-Def.), 824,
861, 885, (CO-Stretch) 1077, 1110, 1179, 1253, 1283
(CF3-Stretch), 1819, and 1842 cmϪ1 (CO-Stretch).
All frequencies given by Varetti and Aymonino [28]
are shifted by about 2–7 cmϪ1 to shorter wavenum-
bers compared to those found here.
Figure 2 Product spectrum after the irradiation of
CF3O2CF3 and CO in 760 torr N2 containing traces of O2 in
the range 1780–1920 cmϪ1
.
In some experiments, additional absorption fea-
tures other than those of CF3OC(O)C(O)OCF3 could
be detected. In Figure 2, the product spectrum of one
of these experiments is presented. The absorptions
due to CF2O were stripped. Beside the bands of
CF3OC(O)C(O)OCF3 at 1823 and 1845 cmϪ1, ab-
sorption features at 1867 and 1889 cmϪ1 are ob-
served. Aymonino [27] reported absorptions at 1867
and 1888 cmϪ1 which appeared in the photolysis of
CF3O2CF3 and the following-up reactions with CO
and O2 . He assigned these absorptions to the car-
bonyl groups of CF3OC(O)O2C(O)OCF3 .
In all experiments in N2 , also CO2 and CF2O were
observed products. It is believed that the formation of
CO2 proceeds according to the same mechanism
which occurs in air, and that CF2O is formed by het-
erogeneous reactions of CF3O radicals on the reactor
walls [29,30].
additional experiments. In the last three experiments,
the amount of CO2 formed corresponded exactly to
the amount of CO consumed.
From the results of the product studies in N2 and
in air, a mechanism for the CF3O initiated oxidation
of CO can be postulated. This mechanism is pre-
sented in Table I. The photolysis of CF3O2CF3 gener-
ates two CF3O radicals, which react with CO. From
the three possible reaction channels (3a–3c), the ad-
dition reaction (3a) is most likely. The most plausible
path for the formation of CF3OC(O)C(O)OCF3 , the
product found in N2 , is the combination of two
CF3OC(O) radicals, which are formed in the addition
of CO to a CF3O radical.
CF3O ϩ CO
CF3OC(O)
(3a)
(6)
2 CF3OC(O)
CF3OC(O)C(O)OCF3
CF3O ϩ wall
CF2O
(5)
The other reactions (3b) and (3c) are unlikely, since
no products were observed which are expected to be
produced from radicals generated in these reactions.
In reaction (3b) CF3 radicals are formed, which com-
bine in N2 to C2F6 (reaction (7)), or in air add O2 to
form CF3O2 , which in turn reacts with CF3O to form
CF3O3CF3 (reactions (1) and (8)) [14,29]. Both com-
pounds C2F6 and CF3O3CF3 could not be detected in
these experiments.
If mixtures of CF3O2CF3/CO/air were irradiated at a
total pressure of 760 torr, CF2O and CO2 were the
only products observed. A great emphasis was put on
the detection of the carbonyl compounds which were
observed in N2 , but none of these compounds were
detected. In the experiments in air, a very fast conver-
sion of CO to CO2 was monitored. Indeed, as it will
be shown later in Table III, 3–9 molecules CO are
oxidized per CF3O radical formed, indicating that a
chain mechanism reforming CF3O radicals is occur-
ring. The yield of CO2 , relative to the amount of CO
consumed, was slightly over 1.0. Hydrocarbons stick-
ing to the walls from previous experiments might be
responsible for the additional CO2 formation. The ad-
ditional CO2 formation decreased with the number of
CF3 ϩ CF3
C2F6
(7)
(8)
CF3O2 ϩ CF3O
CF3O3CF3
FCO radicals produced in reaction (3c) are expected
to result in the formation of FC(O)O2C(O)F or

582
MELLER AND MOORTGAT
Table I Mechanism of the CF3O Initiated CO Oxidation in the Presence of O2
Reaction
Rate Const.
Reference
CF3O2CF3 ϩ hv
CF3O ϩ CO
CF3OC(O) ϩ O2
2 CF3OC(O)O2
CF3OC(O)O
2 CF3O
2 CF3O
CF3OC(O)
CF3OC(O)O2
2 CF3OC(O)O ϩ O2
CF3O ϩ CO2
CF3O2CF3
1.2 ϫ 10Ϫ5
5.0 ϫ 10Ϫ14
5.0 ϫ 10Ϫ12
1.7 ϫ 10Ϫ11
Ͼ1.0 ϫ 103
2.1 ϫ 10Ϫ11
0.45
this work
this work
[31]a
(4a)
(3a)
(13)
(14)
(15)
(17)
(5)
[31]b
assumed
[32]
this work
CF3O ϩ wall
CF2O
a Assumed to be equal k(CH3C(O) ϩ O2
b Assumed to be equal k(2 CH3C(O)O2
CH3C(O)O2).
2 CH3C(O)O).
CF3O2C(O)F in air [33,34]. Indeed, it is known that
FCO adds O2 to form FC(O)O2 . These radicals un-
dergo a self-reaction predominantly via reaction (10)
[35]. The FC(O)O radical in turn reacts either with
another FC(O)O or with a CF3O radical to form
FC(O)O2C(O)F or CF3O2C(O)F, respectively. Again,
none of these products could be detected, indicating
that reaction (3c) does not take place.
since the reaction of CF3OC(O) radicals with O2 is
favoured over the reaction with CF3OC(O)O2 . These
peroxy radicals are expected to undergo a self-reac-
tion to form an alkoxy radical CF3OC(O)O, which
decomposes immediately into CF3O and CO2 .
Turnipseed et al. [23] used PLP/PLIF and
DF/CIMS to study the reaction of CF3O with CO.
They did not observe any products formed in
reactions (3a)–(3c), but from the observed pres-
sure dependence, they concluded that reaction (3a)
is the most likely channel. Additionally, no evi-
dence was found for reactions (3b) and (3c), but
both reactions were not ruled out. Wallington and
Ball [25] employed a relative rate method using
FTIR spectroscopy to study the reaction of CF3O
with CO in air. From their endproduct study, using
13CO and using results of Czarnowski and Schu-
macher [36], they concluded that reaction (3) pro-
ceeds predominantly via reaction (3a). These results
of both studies are in agreement with those found
here.
FCO ϩ O2 ϩ M
2FC(O)O2
FC(O)O2 ϩ M
(9)
2 FC(O)O ϩ O2 (10)
2FC(O)O
FC(O)O2C(O)F (11)
FC(O)O ϩ CF3O
CF3O2C(O)F
(12)
The chain reaction for the CF3O initiated CO oxida-
tion in air can easily be explained by the sequence of
the following reactions:
CF3OC(O) ϩ O2
2 CF3OC(O)O2
CF3OC(O)O
CF3OC(O)O2
2 CF3OC(O)O ϩ O2 (14)
CF3O ϩ CO2 (15)
(13)
Determination of the Rate Constant for the
Reaction CF3O ؉ CO
The reaction of CF3OC(O) with O2 leads to
CF3OC(O)O2 . The existence of the intermediate
CF3OC(O)O2 was established through the formation
of CF3OC(O)O2C(O)OCF3 in the experiments carried
out in N2 containing traces of O2 . Its formation pro-
ceeds via the combination of the CF3OC(O)O2 radical
with another CF3OC(O) radical.
The rate constant for the reaction (3) k3a was de-
termined from 6 photolysis experiments of (0.32–
2.0) ϫ 1015 molecule cmϪ3 CF3O2CF3 in the presence
of 3.2 ϫ 1014 molecule cmϪ3 CO in 760 torr syn-
thetic air. The concentration of the reactants and the
products are listed in Table II. The first column gives
the time of irradiation. The initial concentration of
CF3O2CF3 and the concentrations of CO, CO2 and
CF2O were determined by their IR absorptions. The
concentrations of CF3O2CF3 after each interval of ir-
radiation were calculated from the difference of the
initial CF3O2CF3 concentration and the observed
CF2O concentration.
CF3OC(O) ϩ CF3OC(O)O2
CF3OC(O)O2C(O)OCF3
(16)
In the experiments performed in air, however, the for-
mation of CF3OC(O)O2C(O)OCF3 was not observed,

PHOTOCHEMICALLY GENERATED CF3O RADICALS
583
Table II Reactant and Product Concentrations in the Photolysis of CF3O2CF3 in the Presences of CO in 760 torr Air
Time
[s]
[CO]
[CF3O2CF3]
[CO2]
[CF2O]
[CF3O2CF3]rec
[molecule cmϪ3
]
0
900
1800
2700
0
600
1200
1800
2400
3000
3600
0
600
1200
1800
0
600
1200
1800
0
600
1200
1800
2400
0
600
1200
1800
2400
3.15 ϫ 1014
9.16 ϫ 1013
1.09 ϫ 1013
2.23 ϫ 1012
3.19 ϫ 1014
2.57 ϫ 1014
2.18 ϫ 1014
1.69 ϫ 1014
1.37 ϫ 1014
1.07 ϫ 1014
8.13 ϫ 1013
3.20 ϫ 1014
2.36 ϫ 1014
1.55 ϫ 1014
1.12 ϫ 1014
3.19 ϫ 1014
1.34 ϫ 1014
5.01 ϫ 1013
1.23 ϫ 1013
3.13 ϫ 1014
1.86 ϫ 1014
1.03 ϫ 1014
5.06 ϫ 1013
2.07 ϫ 1013
3.12 ϫ 1014
2.05 ϫ 1014
1.21 ϫ 1014
6.67 ϫ 1013
3.50 ϫ 1013
1.96 ϫ 1015
1.95 ϫ 1015
1.95 ϫ 1015
1.94 ϫ 1015
3.20 ϫ 1014
3.19 ϫ 1014
3.18 ϫ 1014
3.16 ϫ 1014
3.15 ϫ 1014
3.14 ϫ 1014
3.13 ϫ 1014
7.20 ϫ 1014
7.18 ϫ 1014
7.16 ϫ 1014
7.14 ϫ 1014
1.73 ϫ 1015
1.73 ϫ 1015
1.72 ϫ 1015
1.72 ϫ 1015
9.40 ϫ 1014
9.37 ϫ 1014
9.35 ϫ 1014
9.33 ϫ 1014
9.31 ϫ 1014
6.50 ϫ 1014
6.48 ϫ 1014
6.46 ϫ 1014
6.44 ϫ 1014
6.41 ϫ 1014
–.––
–.––
–.––
2.17 ϫ 1014
3.15 ϫ 1014
3.51 ϫ 1014
–.––
1.17 ϫ 1013
2.51 ϫ 1013
3.70 ϫ 1013
–.––
1.51 ϫ 1013
2.92 ϫ 1013
4.37 ϫ 1013
–.––
4.56 ϫ 1013
9.20 ϫ 1013
1.43 ϫ 1014
1.80 ϫ 1014
2.14 ϫ 1014
2.57 ϫ 1014
–.––
8.71 ϫ 1013
1.58 ϫ 1014
2.19 ϫ 1014
–.––
1.72 ϫ 1014
2.68 ϫ 1014
3.09 ϫ 1014
–.––
1.18 ϫ 1014
2.14 ϫ 1014
2.69 ϫ 1014
2.97 ϫ 1014
–.––
2.14 ϫ 1012
4.46 ϫ 1012
6.98 ϫ 1012
9.27 ϫ 1012
1.20 ϫ 1013
1.40 ϫ 1013
–.––
3.82 ϫ 1012
7.38 ϫ 1012
1.11 ϫ 1013
–.––
6.93 ϫ 1012
1.33 ϫ 1013
2.11 ϫ 1013
–.––
5.21 ϫ 1012
9.99 ϫ 1012
1.41 ϫ 1013
1.87 ϫ 1013
–.––
1.23 ϫ 1012
2.34 ϫ 1012
3.35 ϫ 1012
4.45 ϫ 1012
5.30 ϫ 1012
6.50 ϫ 1012
–.––
3.26 ϫ 1012
6.61 ϫ 1012
9.85 ϫ 1012
–.––
8.93 ϫ 1012
1.81 ϫ 1013
2.64 ϫ 1013
–.––
4.14 ϫ 1012
8.41 ϫ 1012
1.31 ϫ 1013
1.73 ϫ 1013
–.––
9.61 ϫ 1013
1.79 ϫ 1014
2.36 ϫ 1014
2.74 ϫ 1014
4.57 ϫ 1012
8.37 ϫ 1012
1.26 ϫ 1013
1.74 ϫ 1013
2.38 ϫ 1012
5.10 ϫ 1012
7.60 ϫ 1012
9.80 ϫ 1012
[CF3O2CF3]t ϭ [CF3O2CF3]0 Ϫ 1 [CF2O]t (Eq. 1)
In this figure it can be seen that the concentration
of CO decreases exponentially with photolysis
time. The symbols represent the measured concen-
trations and the solid lines corresponds to the best
least-square fits for the observed pseudo-first-
order rate constant for the removal of CO (kobs) using
Eq. 3.
2
Since there were less CF2O molecules observed than
CF3O radicals formed (calculated from the photolysis
rate of CF3O2CF3, determined in a separate experi-
ment) a fraction of the CF3O radicals must have re-
combined to CF3O2CF3.
2 CF3O
CF3O2CF3
[CO]t ϭ [CO]0 exp(Ϫkobst)
(Eq. 3)
(17)
The last column in Table II gives the concentrations
of CF3O2CF3 which are formed by recombination.
These values were calculated using the following
equation:
The concentrations for CF2O and CF3O2CF3 (recom-
bined) increase linearly with time. In Table III the ini-
tial concentrations of CO and CF3O2CF3, the ob-
served pseudo-first-order rate constant for the
removal of CO (kobs), as well as the formation rate for
CF2O (v(CF2O) ϭ d[CF2O]/dt) and the recombined
CF3O2CF3 (v(CF3O2CF3)rec ϭ d[CF3O2CF3]rec /dt) are
presented.
[CF3O2CF3]rec ϭ [CF3O2CF3]0 exp(Ϫk4at)
1
2
Ϫ
[CF2O] (Eq. 2)
In Figure 3, a plot of the CO concentrations vs. reac-
tion time is presented for various initial mixtures.
The last column in this table gives the chain
length, which corresponds to the number of CO mol-

584
MELLER AND MOORTGAT
to measure the CF3O radical concentration directly in
this apparatus, a reference reaction with a known rate
constant must be used. The chosen reference reaction
was the recombination of two CF3O radicals. The ad-
vantage of using this reference reaction is that the re-
combination reaction takes place anyway and there-
fore no other substances need to be added.
The concentration of the CF3O radicals was calcu-
lated using the following equation and the value for
k17 ϭ 2.1 ϫ 10Ϫ11 cm3 moleculeϪ1 sϪ1 given by Batt
and Walsh [32] and confirmed by Turnipseed et al.
(k17 ϭ (2 Ϫ 3) ϫ 10Ϫ11 cm3 moleculeϪ1 sϪ1) [23].
v(CF3O2CF3)rec ϭ d[CF3O2CF3]rec/dt ϭ k17 [CF3O]2
Figure 3 Concentration-time-profiles for CO at different
initial concentrations of CF3O2CF3 . Symbols represent the
measured concentrations and the lines the best least-square
fits (eq. (3)).
(Eq. 6)
In Figure 4 the calculated rate constant kobs is plotted
vs. the calculated CF3O concentration. From the
ecules oxidized per CF3O radical formed. This num-
ber was calculated using Eq. 4.
slope of this plot
a
value for k3a ϭ 5.2Ϯ
0.6) ϫ 10Ϫ14 cm3 moleculeϪ1 sϪ1 could be derived,
whereby the quoted error reflects two standard devia-
tions. In the same way, the rate constant for the for-
mation of CF2O (k5) could be calculated. A plot of the
rate of formation of CF2O (v(CF2O) ϭ d[CF2O]/dt)
vs. the calculated CF3O concentration is shown in
Figure 5 and a value for k5 ϭ 0.45 Ϯ 0.06 sϪ1 was
deduced. This rate constant has no influence on k3a .
Alternatively, using the mechanism listed in Table
I and the simulation program FACSIMILE, the rate
constants k3a and k5 were calculated by fitting simul-
taneously the observed concentrations of CO,
CF3O2CF3 , and CF2O in all experiments. From these
calculations, values of k3a ϭ (4.7 Ϯ 0.6) ϫ 10Ϫ14 cm3
moleculeϪ1 sϪ1 and k5 ϭ 0.45 Ϯ 0.04 sϪ1 were ob-
tained. It could be also shown in these simulations
that the concentration of CF3O remains constant dur-
ing the duration of each experiment. Within the first
few seconds of the irradiation an equilibrium of
CF3OC(O), CF3OC(O)O2 , and CF3OC(O)O radicals
is established and the amount of CF3O radicals lost
by reaction (3a) is equal to the amount of CF3O re-
formed in reaction (15). For these calculations it was
[CO]0 Ϫ [CO]end
chain length ϭ
(Eq. 4)
2k4a[CF3O2CF3]0
As it can be seen from this column, a mean value of
3–9 CO molecules (averaged over the whole duration
of the experiment) are oxidized by a single CF3O rad-
ical. The chain length is dependent on the CO con-
centration actually present in the system, and there-
fore on the initial concentration of CF3O2CF3 .
The removal of CO can be described by the fol-
lowing expression:
Ϫd[CO]/dt ϭ kobs[CO] ϭ k3a[CF3O][CO]
(Eq. 5)
The CF3O concentration is in a steady state during
each run, since the consumption of CF3O2CF3 under
the experimental conditions is very small (less than
3%) and the CF3O radicals reacting with CO are re-
formed in the oxidation mechanism. Therefore k3a can
be derived as the slope from a plot of kobs vs. the
mean CF3O concentration. Since it was not possible
Table III Pseudo-First-Order Rate Constants (kobs) for the Removal of CO, Rates of Formation of CF2O and CF3O2CF3
and Mean Chain Length in the Photolysis of CF3O2CF3 in the Presences of CO in 760 Torr Air
[CO]0
[CF3O2CF3]o
kobs
v(CF2O)
v(CF3O2CF3)rec
Chain Length
molecule cmϪ3
sϪ1
molecule cmϪ3 sϪ1
3.15 ϫ 1014
3.19 ϫ 1014
3.20 ϫ 1014
3.19 ϫ 1014
3.13 ϫ 1014
3.12 ϫ 1014
1.96 ϫ 1015
3.20 ϫ 1014
7.20 ϫ 1014
1.73 ϫ 1015
9.40 ϫ 1014
6.50 ϫ 1014
1.46 ϫ 10Ϫ3
3.56 ϫ 10Ϫ4
5.76 ϫ 10Ϫ4
1.52 ϫ 10Ϫ3
9.54 ϫ 10Ϫ4
8.08 ϫ 10Ϫ4
1.38 ϫ 1010
3.96 ϫ 109
6.14 ϫ 109
1.16 ϫ 1010
7.71 ϫ 109
7.14 ϫ 109
1.61 ϫ 1010
1.77 ϫ 109
5.48 ϫ 109
1.47 ϫ 1010
7.26 ϫ 109
4.14 ϫ 109
3
9
7
4
6
8

PHOTOCHEMICALLY GENERATED CF3O RADICALS
585
low CF3O2CF3 concentration with longer photolysis
times (more data points) contribute stronger. The op-
timum value between both measurements is
k3a ϭ (5.0 Ϯ 0.9) ϫ 10Ϫ14 cm3 moleculeϪ1 sϪ1.
For the determination of k3a three parameters are
needed, which all possess an uncertainty or experi-
mental error. These parameters represent the observed
CO removal rate constant (kobs), the amount of
CF3O2CF3 recombined and the rate constant k17 . The
experimental error in the determination of kobs is
Ϯ5%. Several factors contribute to the error in the de-
termination of the total amount of recombined
CF3O2CF3 , which are the error in the determination
of the photolysis rate constant k4a , and the error asso-
ciated in the determination of the initial CF3O2CF3
and the CF2O concentration. The error within k4a and
the error associated with the determination of the
concentrations are Ϯ5% and Ϯ10%, respectively.
The accuracy of the rate constant k17 [35] must
also be considered. In this study, k17 was accepted as
cited in the literature [35]. However, a change of a
factor X in k17 results in a change of a factor X1/2
in k3a . The influence of the error associated within
kobs and k17 , as well as the error in the determina-
tion of the CF2O concentration on k3a , are approxi-
mately within the statistical error of k3a . The overall
experimental error in k3a amounts to about 20%. If
there are sideproducts formed which could not
be detected (e.g., CF3OH from the reaction of
CF3O with hydrocarbons sticking to the reactor
walls from previous experiments), this would reduce
the amount of CF3O2CF3 (recombined) and therefore
also results in a lower rate constant k3a according to
Eq. 6.
Figure 4 kobs (eq. (3)) vs. calculated CF3O concentration
(eq. (6)).
not necessary to know the exact rate constants for re-
actions (13)–(15); it was satisfying to know the order
of magnitude for these rate constants.
With both methods (experimental and simulation),
rate constants of k3a(exp) ϭ (5.2 Ϯ 0.6) ϫ 10Ϫ14 cm3
moleculeϪ1 sϪ1
and
k3a(sim) ϭ (4.7 Ϯ 0.6) ϫ
10Ϫ14 cm3 moleculeϪ1 sϪ1 were determined. Both val-
ues are in agreement within their error limits. The
small deviation is caused by different weights of the
single measurements. In method 1, the measurements
at large initial CF3O2CF3 concentration and short
photolysis times are weighted stronger than in the
FACSIMILE simulations, where the experiments at
Another possible error needs to be considered.
Turnipseed et al. [23] reported a small contribution of
F atoms produced in the photolysis of CF3O2CF3 (see
reaction (4b)). F atoms also react with CO to form
FCO.
F ϩ CO ϩ M
FCO ϩ M
(18)
The subsequent reactions of FCO were presented be-
fore in reactions (9)–(12). As mentioned before, none
of the expected end products FC(O)O2C(O)F or
FC(O)O2CF3 were found in this study. The IR spectra
of both species are known and show specific features
at 1902 and 1919 cmϪ1 [37,34]. Those absorptions
bands were not observed in any product spectrum. In
addition Turnipseed et al. [23] give a contribution of
0.08 Ϯ 0.06 for channel (4b), which corresponds to a
ratio in the formation of F atoms to CF3O radicals of
0.04 Ϯ 0.03. Taking also into account that CF3O re-
Figure 5 v(CF2O) vs. calculated CF3O concentration
(eq. (6)).

586
MELLER AND MOORTGAT
3. P. D. Lightfoot, R. A. Cox, J. N. Crowley, M. Destriau,
G. D. Hayman, M. E. Jenkin, G. K. Moortgat, and F.
Zabel, Atmos. Environ., A 26, 1805 (1992).
4. J. Chen, T. Zhu, and H. Niki, J. Phys. Chem., 96, 6115
(1992).
5. T. J. Bevilacqua, D. R. Hanson, and C. J. Howard, J.
Phys. Chem., 97, 3750 (1993).
6. J. Sehested and O. J. Nielsen, Chem. Phys. Lett., 206,
369 (1993).
7. A. A. Turnipseed, S. B. Barone, and A. R. Ravi-
shankara, J. Phys. Chem., 98, 4594 (1994).
8. N. R. Jensen, D. R. Hanson, and C. J. Howard, J. Phys.
Chem., 98, 8574 (1994).
9. J. Chen, V. Young, T. Zhu, and H. Niki, J. Phys. Chem.,
97, 11696 (1993).
10. O. J. Nielsen and J. Sehested, Chem. Phys. Lett., 213,
433 (1993).
11. T. J. Wallington, M. D. Hurley, and W. F. Schneider,
Chem. Phys. Lett., 213, 442 (1993).
acts with CO with a chain length of 3–9 under the
experimental conditions, the expected contribution of
reaction (18) to the total loss of CO will be less than
1%. Even if reaction (4b) would take place, it would
have only a very little effect on the rate constant k3a .
In three earlier studies rate constants for the reac-
tion CF3O ϩ CO have been determined. Saathoff and
Zellner [24] reported a rate constant of k3 ϭ 4.4 ϫ
10Ϫ14 cm3 moleculeϪ1 sϪ1 at 298 K in 50 torr He. In
the more recent studies by Turnipseed et al. [23] and
by Wallington and Ball [25] a pressure dependence of
reaction (3) was observed. High pressure rate con-
stants have been reported for k3,ϱ ϭ (6.8 Ϯ 1.2) ϫ
10Ϫ14 cm3 moleculeϪ1 sϪ1 and k3(760 torr) ϭ
(7.2 Ϯ 0.7) ϫ 10Ϫ14 cm3 moleculeϪ1 sϪ1, respec-
tively. The value obtained in this work k3a ϭ
(5.0 Ϯ 0.9) ϫ 10Ϫ14 cm3 moleculeϪ1 sϪ1 is somehow
lower than the reported high pressure rate constants.
This lower rate constant might be caused by an over-
estimation of the CF3O concentration due to its indi-
rect determination in this study as stated earlier in
this article.
12. M. M. Maricq and J. J. Szente, Chem. Phys. Lett., 213,
449 (1993).
13. C. Fockenberg, H. Saathoff, and R. Zellner, Chem.
Phys. Lett., 218, 21 (1994).
14. R. Meller and G. K. Moortgat, J. Photochem. Photo-
biol. A: Chem., 86, 15 (1995).
No pressure dependent studies were performed in
this study, but Turnipseed et al. [23] found that the
high pressure limit for this reaction is reached at
about 300 torr. Therefore, the rate constant obtained
here at 760 torr can be used for calculations in the
troposphere. It has been state elsewhere [10] that re-
action with CH4 represents the dominant sink for
CF3O radicals in the troposphere, leading to an aver-
age lifetime for CF3O of 1–2 s. In addition, reaction
of CF3O with H2O might need to be considered, but
only upper limits for the reaction of CF3O with H2O
are reported [22,23]. Choosing an average value for
the CO concentration of about 100 ppb in the tropos-
phere and using the k3a determined in this work, it can
be calculated that the lifetime of CF3O towards the
reaction with CO is about 8 s, and that therefore only
a minor fraction of the CF3O radicals will react with
CO. It must be also considered that CF3O radicals are
reformed in the follow-up reactions (no net loss), so
that it can be concluded that CO is not an important
sink for CF3O.
15. T. J. Wallington and J. C. Ball, Chem. Phys. Lett., 234,
187 (1995).
16. J. Chen, T. Zhu, H. Niki, and G. J. Mains, Geophys.
Res. Lett., 19, 2215 (1992).
17. S. B. Barone, A. A. Turnipseed, and A. R. Ravi-
shankara, J. Phys. Chem., 98, 4602 (1994).
18. J. Chen, T. Zhu, V. Young, and H. Niki, J. Phys. Chem.,
97, 7174 (1993).
19. H. Saathoff and R. Zellner, Chem. Phys. Lett., 206, 349
(1993).
20. C. Kelly, J. Treacy, and H. W. Sidebottom, Chem. Phys.
Lett., 207, 498 (1993).
21. C. Kelly, H. W. Sidebottom, J. Treacy, and O. J.
Nielsen, Chem. Phys. Lett., 218, 29 (1994).
22. T. J. Wallington, M. D. Hurley, W. F. Schneider, J. Se-
hested, and O. J. Nielsen, J. Phys. Chem., 97, 7606
(1993).
23. A. A. Turnipseed, S. B. Barone, N. R. Jensen, D. R.
Hanson, C. J. Howard, and A. R. Ravishankara, J.
Phys. Chem., 99, 6000 (1995).
24. H. Saathoff and R. Zellner, presented at the 12th Inter-
national Symposium on Gas Kinetics, July 1992, Uni-
versity of Reading.
25. T. J. Wallington and J. C. Ball, J. Phys. Chem., 99,
3201 (1995).
BIBLIOGRAPHY
26. W. H. Raber and G. K. Moortgat, in Progress and
Problems in Atmospheric Chemistry, World Scientific
Publ. Co, Singapore, J. Barker, Ed., 1996, pp. 318–373.
27. P. J. Aymonino, J. Photochem. Photobiol. 7, 761
(1968).
28. E. L. Varetti and P. J. Aymonino, An. As. Quim. Arg.,
55, 153 (1967).
29. O. J. Nielsen, T. Ellermann, J. Sehested, E. Bart-
1. World Meteorological Organization, Global Ozone Re-
search and Monitoring Project-Report No. 20 Scientific
Assessment of Stratospheric Ozone, 1989, Vol. 2, Ap-
pendix: AFEAS report, 1989.
2. F. Caralp, R. Lesclaux, and A. M. Dognon, Chem.
Phys. Lett., 129, 433 (1986).

PHOTOCHEMICALLY GENERATED CF3O RADICALS
587
kiewicz, T. J. Wallington, and M. D. Hurley, Int. J.
Chem. Kinet., 24, 1009 (1992).
33. T. J. Wallington, M. D. Hurley, and M. M. Maricq,
Chem. Phys. Lett., 205, 62 (1993).
30. J. Czarnowski and H. J. Schumacher, Z. physik. Chem.
Neue Folge, 92, 329 (1974).
34. E. C. Tuazon and R. Atkinson, J. Atm. Chem., 16, 301
(1993).
31. R. Atkinson, D. L. Baulch, R. A. Cox, R. F. Hampson,
J. A. Kerr, and J. Troe, Evaluated Kinetics and Photo-
chemical Data for Atmospheric Chemistry, J. Phys.
Chem. Ref. Data 21, Supplement IV (1992).
32. L. Batt and R. Walsh, Int. J. Chem. Kinet., 14, 933
(1982).
35. T. J. Wallington, T. Ellermann, O. J. Nielsen, and J. Se-
hested, J. Phys. Chem., 98, 2346 (1994).
36. J. Czarnowski and H. J. Schumacher, Int. J. Chem.
Kinet., 13, 639 (1981).
37. S. I. Ahonkhai, A. Nölle, P. Behr, and H. Heydtmann,
J. Photochem. Photobiol. A: Chem., 80, 389 (1994).