Temperature dependence of the gas-phase reactions F + CHFO, CFO + F, and CFO + CFO

Article abstract of DOI:10.1002/(SICI)1097-4601(1998)30:5<329::AID-KIN2>3.0.CO;2-U

The rate coefficients for the reactions CHFO + F, CFO + F and the self-reaction of CFO were determined over the temperature range of 222-298 K A computer controlled discharge-flow system with mass spectrometric detection was used. The results are expressed in the Arrhenius form (with energies in J) CHFO + F → CFO + HF: k₁,(T) = (9.7 ± 0.7) · 10^-12 exp[-(5940 ± 150)/RT] cm³ molecule cm^-1 s^-1 CFO + F + M → CF₂O + M: k₂(T) = (260 ± 1.17)·10^-10 exp[-(10110 ± 1250)/RT cm³ molecule^-1 s^-1 CFO + CFO → CF₂O + CO: k₃(T) = (3.77 ± 2.7)·10^-10 exp[-(8350 ± 28001/RT] cm³ molecule^-1 s^-1

Full text of DOI:10.1002/(SICI)1097-4601(1998)30:5<329::AID-KIN2>3.0.CO;2-U

Temperature Dependence
of the Gas-Phase Reactions
F ϩ CHFO, CFO ϩ F, and
CFO ϩ CFO
PETER BEHR, CORNELIA KAUPERT, EDUARD SHAFRANOVSKI*, HORST HEYDTMANN
Institut f u¨ r Physikalische und Theoretische Chemie, J. W. Goethe University, Marie Curie Str. 11, 60439 Frankfurt,
Germany
Received 21 February 1997; accepted 22 July 1997
ABSTRACT: The rate coefficients for the reactions CHFO ϩ F, CFO ϩ F and the self-reaction
of CFO were determined over the temperature range of 222–298 K. A computer controlled
discharge-flow system with mass spectrometric detection was used. The results are expressed
in the Arrhenius form (with energies in J):
CHFO ϩ F !: CFO ϩ HF:
Ϫ12
3
Ϫ1
Ϫ1
k1(T) ϭ (9.7 Ϯ 0.7)·10 exp[Ϫ(5940 Ϯ 150)/RT] cm molecule
s
CFO ϩ F ϩ M !: CF O ϩ M:
2
Ϫ10
3
Ϫ1
Ϫ1
k2(T) ϭ (2.60 Ϯ 1.17)·10 exp[Ϫ(10110 Ϯ 1250)/RT cm molecule
s
CFO ϩ CFO !: CF O ϩ CO:
2
Ϫ10
3
Ϫ1
Ϫ1
k3(T) ϭ (3.77 Ϯ 2.7)·10 exp[Ϫ(8350 Ϯ 2800)/RT] cm molecule
s
᭧
1998 John Wiley & Sons, Inc. Int J Chem Kinet 30: 329–333, 1998
INTRODUCTION
F ϩ CHFO !: HF ϩ CFO
(1)
The fluoroformyl radical, CFO, is a species of interest
for atmospheric processes [1]. For example it is
The rate of this reaction and of the consecutive pro-
cesses
formed by photolysis of CF O and CFClO which are
2
intermediates in the degradation of fluorinated hydro-
carbons [2,3]. In two preceding publications [4,5] we
have shown that a convenient source for CFO is the
reaction of fluorine atoms with formyl fluoride:
F ϩ CFO ϩ M !: CF O ϩ M
(2)
(3)
2
CFO ϩ CFO !: CF O ϩ CO
2
have been studied earlier at room temperature in a flow
system. We note that k was measured at 298 K by our
group [5] in excellent agreement with the value re-
ported by Wallington et al. [6] who used an entirely
different experimental method. In the present publi-
*
Guest from the Institute of Chemical Physics, Russian Acad-
3
emy of Sciences, Moscow
Correspondence to: H. Heydtmann
Contract grant Sponsor: Deutsche Forschungsgemeinschaft
1998 John Wiley & Sons, Inc.
᭧
CCC 0538-8066/98/050329-05

3
30
BEHR ET AL.
cation we have investigated the temperature depen-
dence of these three processes between 222 K and
[7]. The temperature was checked by a thermometer
inside the cooling liquid of the cryostat and by a ther-
mocouple on the wall of the flow tube. A calibration
was performed to correlate cryostat temperatures with
thermocouple readings. During experiments the ther-
mocouple was removed. The estimated temperature
accuracy is Ϯ0.5 K.
2
98 K. This can be considered as a step towards using
reaction (1) in laboratory experiments to simulate re-
actions of stratospheric relevance at stratospheric tem-
peratures.
The flow tube has three side-arms and is equipped
with a movable injector. For the kinetic studies of the
F ϩ CHFO reaction system CHFO was introduced
through the moveable injector. F atoms were generated
in a quartz side-arm of the flow tube by a microwave
EXPERIMENTAL
A low-pressure discharge-flow reactor which is cou-
pled to a quadrupole mass spectrometer (Vacuum
Generator SXP 400) was used in the experiments. The
MS has a range up to 400 amu and it was used for
continuous sampling and detection of gases in the re-
actor. A differential pumping technique was applied
to reduce the pressure from typically 2.8 mbar in the
flow system to less than 10 mbar in the mass spec-
trometer chamber [7]. The discharge-flow apparatus
employing modulated molecular-beam sampling has
been described earlier in detail [5]. The typical linear-
discharge using a mixture of F (5%) in Helium (purity
2
99.9998%). The interior of the side arm was lined with
alumina to minimize the attack of F-atoms on quartz
and to suppress the formation of SiF . Wall recombi-
4
nation of F-atoms was found to be less than 5% with
a tube coated by Teflon. At least 95% dissociation of
F2 could be achieved judging from the m/z 38 peak
Ϫ6
(F2). The absolute F-atom concentration was deter-
mined by titration using the well studied fast reactions
F ϩ H S and F ϩ HS [5,8].
2
Ϫ1
flow velocity was 1600 cm s . For high-detection
sensitivity the single-ion counting technique was ap-
plied using a channeltron as ion detector. The chan-
neltron pulses were fed to an amplifier-discriminator
and an IBM-Computer with a sixteen-channel counter.
This allowed up to 16 masses to be selected for si-
multaneous analysis. Stable species could be detected
down to densities 10 cm at S/N ϭ 2 for counting
periods of 512 s. All gas flows were measured by
mass-flow controllers (Tylan). During experiments
pressure and flow parameters were controlled and re-
corded by the computer. To study the reactions at dif-
ferent temperatures a cryostat was connected to the
cooling tube surrounding the flow reactor. The cooling
tube is pictured in Figure 1 of our earlier publication
To synthesize CHFO we used the preferred [4] and
convenient fluorination of formic acid with cyanuric
acid [9]. CHFO was diluted with He in the range of
2% to 5%. The dilution was controlled by measure-
ment of the partial pressures of the components with
Baratron pressure gauges. Mass spectra were obtained
using the ionization energy of 70 eV. Product and ed-
uct peaks were obtained at m/z 19, 20, 47, 48, and 66.
These peaks were assigned to the species F, HF, CFO,
9
Ϫ3
CHFO, and CF O, respectively, and used to follow
2
their consumption and formation rates. CF O obtained
2
from P.C.R (purity of 98%) was used to calibrate the
signal at m/z at 66.
RESULTS
1. CHFO ؉ F : HF ؉ CFO
The rate constant of this reaction was obtained as de-
scribed previously for the room temperature measure-
ments. Fluorine atoms were added in approximately
tenfold excess using the side inlet. Following the de-
cay of the formyl fluoride concentration the LARKIN
evaluation program was used to fit the rate constant to
the experimental data. Table I gives reagent initial
concentrations, reactor temperatures and the resulting
k1 values. Figure 1 shows the Arrhenius plot for the
twelve experiments. This number was sufficient since
there is no need to vary [F ]/[CHFO ] as for the de-
0
0
Figure 1 Reaction CHFO ϩ F: Plot of ln k as a function
termination of k and k . The following rate equation
2 3
1
of inverse temperature.
was obtained:

GAS-PHASE REACTIONS
331
Table I Experimental Conditions and the Results for
the CHFO ϩ F Reaction
Table II Initial Conditions to Study the Temperature
Dependence of COF ϩ F : Product and COF ϩ
COF : Products. The Rate Constants Obtained with
these Experiments are Given in Table III
1
3
14
[
CHFO] /10
[cm ]
[F] /10
k1(T)
0
0
Ϫ3
Ϫ
3
3
Ϫ1
Ϫ1
T (K)
[cm ]
[cm molecule s ]
1
4
14
[
CHFO] /10
[cm ]
[F] /10
0
98.2a
Ϫ13a
Ϫ13
Ϫ13
Ϫ13
Ϫ13
Ϫ13
Ϫ13
Ϫ13
Ϫ13
Ϫ13
Ϫ13
Ϫ13
Ϫ13
0
2
2
2
2
2
2
2
2
2
2
2
2
2
–
–
(8.8 Ϯ 0.16)*10
(7.44 Ϯ 1.11)*10
(6.28 Ϯ 0.94)*10
(6.27 Ϯ 0.94)*10
(5.46 Ϯ 0.82)*10
(5.71 Ϯ 0.86)*10
(5.16 Ϯ 0.77)*10
(5.22 Ϯ 0.78)*10
(4.78 Ϯ 0.72)*10
(4.61 Ϯ 0.69)*10
(4.27 Ϯ 0.64)*10
(4.17 Ϯ 0.63)*10
(3.9 Ϯ 0.59)*10
Ϫ3
Ϫ3
Exp. No. T [K]
[cm ]
[F] /[CHFO]0
0
75.6
61.4
61.4
52.4
52.4
43.8
43.8
35.3
35.3
27.4
27.4
23.4
3.05
4.41
3.22
4.48
3.33
4.67
3.45
4.87
3.57
5.1
4.11
3.82
4.33
3.96
4.48
4.1
4.64
4.25
4.81
4.39
4.98
4.47
1
2
3
4
5
6
7
8
9
10
11
2
3
4
5
275.6
1.385
1.04
1.928
3.087
4.617
1.496
2.996
4.512
2.15
1.39
2.97
6.68
0.73
1.93
4.35
0.94
1.88
4.24
11.29
0.72
1.93
4.35
0.92
1.88
4.24
11.29
0.57
1.52
3.43
0.72
1.48
3.34
8.90
0.43
1.14
2.57
0.39
0.27
0.18
0.691
2.059
1.549
1.037
2.299
1.728
1.158
0.583
2.185
1.648
1.105
2.461
1.851
1.241
0.624
3.00
2.246
1.497
3.241
2.439
1.743
0.899
4.305
3.367
2.286
1.507
2.42
257.0
3.253
4.906
6.582
1.584
3.186
4.808
2.259
3.485
5.256
7.044
1.713
3.419
5.129
2.343
3.618
5.814
7.997
1.844
3.845
5.874
0.586
0.659
0.756
3.7
5.22
1
1
1
1
a
The value at room temperature is taken from ref. [4].
239.2
Ϫ12
k1(T) ϭ (9.7 Ϯ 0.7)·10 exp[Ϫ(5940
16
17
3
Ϫ1
Ϫ1
Ϯ 150)/RT] cm molecule
s
1
1
2
2
2
2
2
2
2
2
8
9
0
1
2
3
4
5
6
7
The activation energy is given in Joules and the errors
in the equation are purely statistical. For each temper-
ature we estimate the possible systematic errors less
than 15%.
223.4
222.1
2. Secondary Reactions with CFO
With the knowledge of the temperature dependence of
the CFO formation reaction (1) we are able to deter-
mine the Arrhenius parameters of the secondary re-
actions (2 and 3).
28
29
30
284.5
257.0
223.4
4.166
The possible reaction
F ϩ CFO !: CF O ϩ F
(4)
2
2
is not important under our experimental conditions be-
cause of the low F concentration [5].
2
Table III Rate Constants k and k as a Function of
the Temperature
2
3
It is impossible to determine the two rate constants
of the reactions (2) and (3) simultaneously using the
CF O product profile measured with a fixed [F] /
k2(T)
k3(T)
2
0
3
Ϫ1
Ϫ1
3
Ϫ1
Ϫ1
T [K] [cm molecule s ] T [K] [cm molecule s ]
[
CHFO] ratio.
0
As a result we obtain an infinite set of two depend-
98.2a (5.00 Ϯ 1.4)*10Ϫ
12a
12
12
12
12
12
12
12
12
a
Ϫ
Ϫ
Ϫ
Ϫ
11a
2
2
298.15 (1.70 Ϯ 0.35)*10
ant rate constants. Therefore, we had to study the
CF O production at different [F] /[CHFO] ratios be-
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
12
75.6 (2.50 Ϯ 0.7)*10
284.5 (8.11 Ϯ 1.22)*10
256.95 (7.55 Ϯ 1.13)*10
223.35 (4.41 Ϯ 0.66)*10
12
2
0
0
275.6 (3.30 Ϯ 0.9)*10
257.0 (2.00 Ϯ 0.5)*10
257.0 (2.80 Ϯ 0.8)*10
39.2 (1.30 Ϯ 0.4)*10
39.2 (1.80 Ϯ 0.5)*10
23.4 (1.10 Ϯ 0.3)*10
22.1 (1.20 Ϯ 0.3)*10
12
cause reaction (2) is preferred at high [F] and the re-
action (3) is favored at low F concentrations. To obtain
the temperature dependence of these two reactions we
changed the [F] /[CHFO] value at each temperature.
2
2
2
2
0
0
The initial concentrations of F and CHFO and their
ratios are reported in Table II. The concentration ratio
a
for a set of experiments, the corresponding CF O prod-
The values at room temperature are from ref. [5].
2

3
32
BEHR ET AL.
stantially higher than the value obtained in our work.
The more recent result, however, by Meagher et al.
Ϫ12
3
Ϫ1
Ϫ1
[
11] who report 0.83·10 cm molecule
s agrees
with our work. Principal drawbacks of the chemilu-
minescence method are: (1) Emission of the
2.3–3.0 ␮m region is observed corresponding to a
multitude of vibrational-rotational states of HF and (2)
the emission curve is fitted to a double exponential
expression corresponding to formation and deactiva-
tion of vibrationally excited HF. This is an oversim-
plification since deactivation is different for each col-
lision partner and also for each vibrational state; the
higher vibrational states (v ϭ 2, 3) will preferably re-
lax with ⌬v ϭ Ϫ1 to form the lower vibrational states
and these still will give emission [12].
Figure 2 Reaction CFO ϩ F: Plot of ln k as a function of
inverse temperature.
2
uct concentration, and time led to a three dimensional
Because of this cascading effect the evaluation is
not straightforward and in our opinion is the main rea-
son for the discrepancies concerning the rate constants.
No obvious source of error comes to our mind, how-
ever, when we consider the relative rate method of
Meagher et al. [11]. This supports the higher quality
of the results of Meagher et al. and also of our tem-
perature dependent results. Although Francisco and
plot from which both rate constants k and k can be
2
3
obtained. The procedure of fitting the experimental
data with the new 3D-software is described in detail
in our previous article [5]. At first the evaluation of
the experiments 28–29 (Table II) using preliminary
values of k yield the values of k , which are summa-
2
3
rized in Table III. These k values together with
3
k3, 298 from [5] resulted in the Arrhenius expression
Zhao report too high values of k at all temperatures
1
Ϫ
10
their Arrhenius activation energy is roughly in agree-
ment with our result.
Reaction (2) is expected to be a third-order reaction
at low pressures. Cobos et al. [13] have measured the
pressure dependence between 20 and 700 torr at
k3(T) ϭ (3.77 Ϯ 2.7)·10 exp[Ϫ(8350
3
Ϫ
1
Ϫ1
Ϯ 2800)/RT] cm molecule
s
The final values of k (T) were then obtained from ex-
2
periments 1–27 using the 3D-software as described
before and were listed in Table II. The logarithms of
these values were plotted as a function of the inverse
temperature as shown in Figure 2. The Arrhenius
equation for the recombination of F with CFO is rep-
resented by
2
96 K with CF as a bath gas. At 20 torr they find a
4
Ϫ11
cm3
second-order rate constant of 21·10
molecule s . This agrees reasonably with our value
at ambient temperature, the higher value is due to the
Ϫ1
Ϫ1
higher pressure and the more efficient collider CF . In
4
our experiments [5] the pressure dependence remained
concealed due to the small pressure variation possible.
For the closely related recombination reaction in
which a CF bond is formed
Ϫ10
k2(T) ϭ (2.60 Ϯ 1.17)·10 exp[Ϫ(10110
3
Ϫ1
Ϫ1
Ϯ 1250)/RT cm molecule s .
F ϩ CF ϩ M !: CF ϩ M
(5)
3
4
DISCUSSION
data are available in temperature dependence from a
shock-tube study between 1700 and 3000 K [14].
From the analysis of the rather complex reaction sit-
uation
The temperature dependence of reaction (1) was in-
vestigated before by Francisco and Zhao [10]. They
used a pulsed photolysis of SF to form F atoms and
6
followed the decay of IR-chemiluminescence emitted
by the excited HF molecules. The Arrhenius equation
extracted by these authors was
Ϫ16
Ϫ4.64
k5 ϭ 2.72·10
T
exp
Ϫ1434/T] cm molecules
6
Ϫ
2
Ϫ1
[
s
Ϫ
11
k(T) ϭ 4.4 Ϯ 2.6·10 expϪ[7530
was derived. We note that the temperature dependence
is represented by two terms which appear to compen-
sate largely around 300 K. Therefore, a further dis-
cussion is not useful. The activation energy as deter-
mined by us for process (2) appears to be rather high,
3
Ϫ1
Ϫ1
Ϯ 1670]/T cm molecule
s
which yields
a
room temperature value of
.3·10 cm molecule s . This latter value is sub-
Ϫ12
3
Ϫ
1
Ϫ1
2

GAS-PHASE REACTIONS
333
however, for an atom-radical recombination process.
This may be due to a contribution of a second micro-
scopic channel or to some temperature dependent het-
erogeneous recombination.
Reaction (3) was studied at several temperatures
before by following the CFO concentration in a flash-
photolysis experiment [15]. Maricq et al. [15] claimed
this process to be nearly temperature independent:
BIBLIOGRAPHY
1. Z. Li, T. S. Dibble, and J. S. Francisco, in Progress and
Problems in Atmospheric Chemistry, J. R. Barker, Ed.,
World Scientific, Singapore, 1995.
2
3
4
. A. N o¨ lle, H. Heydtmann, R. Meller, W. Schneider, and
G. Moortgat, Geophys. Res. Lett., 19, 281 (1992).
. M. Hermann, A. N o¨ lle, and H. Heydtmann, Chem. Phys.
Lett., 226, 559 (1994).
. P. Behr, K. Goldbach, and H. Heydtmann, Int. J. Chem.
Kinet., 25, 957 (1993).
Ϫ11
k3 ϭ (3.7 Ϯ 1.2)·10 exp[Ϫ(160
cm3
Ϫ1
Ϫ1
5. P. Behr and H. Heydtmann, Ber. Bunsenges. Phys.
Chem., 100, 553 (1996).
Ϯ 80)/T]
molecule
s
s
6
7
8
9
. T. J. Wallington, T. Ellermann, O. J. Nielsen and J.
Sehested, J. Phys. Chem., 98, 2346, (1994).
. U. W o¨ rsd o¨ rfer and H. Heydtmann, Ber. Bunsenges.
Phys. Chem., 93, 1132 (1989).
. G. Sch o¨ nle, M. M. Rahmann, and R. N. Schindler, Ber.
Bunsenges. Phys. Chem., 91, 66 (1987).
Ϫ
11
3
Ϫ
1
Ϫ1
and k3, 293 ϭ 1.9·10 cm molecule s .
This Arrhenius equation disagrees with our results:
The rate constant at room temperature, however, is in
satisfactory agreement. As noted by Maricq et al. a
temperature dependence is visible for T Ն 273 K (fig.
of [15]). We wish to point out that for these mea-
surements at higher temperatures an Arrhenius equa-
. G. A. Olah, M. Nojima, and I. Kerekes, Synthesis, 487
4
(1973).
1
0. J. S. Francisco and Y. Zhao, J. Chem. Phys., 93, 276
Ϫ1
tion can be calculated yielding E ϭ 6100 J mol
A
(1990).
which is only a little lower than our value.
11. R. J. Meagher, M. E. McIntosh, M. D. Hurley, and T.
The lack of temperature dependence in the experi-
ments of Maricq et al. at T Ͻ 273 K was not explained
by these authors. The reason for the discrepancy in
this regime is not clear.
J. Wallington, Int. J. Chem. Kinet. 29, 619 (1997).
1
1
1
2. E. Arunan, D. Raybone, and D. W. Setser, J. Chem.
Phys., 97, 6348 (1992) and references cited therein.
3. C. J. Cobos, A. E. Croce, and E. Castellano, Chem.
Phys. Lett., 239, 320 (1995).
4. A. P. Modica and S. J. Sillers, J. Chem. Phys., 48, 3283
(1968).
The continuous financial support by the Deutsche For-
schungsgemeinschaft is gratefully acknowledged.
15. M. M. Maricq, J. J. Szente, G. A. Khitrov, and J. S.
Francisco, Chem. Phys. Lett., 199, 71 (1992).