J. Phys. Chem. 1994, 98, 12294-12309
12294
Atmospheric Chemical Kinetics of FC(0)O M. Matti Maricq* and Joseph J. Szente Research Laboratory, Ford Motor Company, P.O. Box 2053, Drop 3083, Dearborn, Michigan 48121
Theodore S. Dibble and Joseph S. Francisco Department of Chemistry, Wayne State University, Detroit, Michigan 48202 Received: July 12, 1994; In Final Form: September 12, 1994@
Temperature dependent rate constant measurements over the range 233-323 K are reported for the reactions of FC(0)O with NO, N02, 03, HO2, C2H5O2, F, and FC(O)O, along with rate constants for the reaction with FCO, FC(O)O,, and C2H5 at 295 K. FC(0)O is generated by the 193 nm photolysis (FC(0)0)2. The photolysis yields 67% F atoms and 33% FC(0)O radicals at 295 K. The latter species is formed with extensive intemal excitation, principally in the v3 bending mode, but is collisionally relaxed with an overall rate of (3.5 x lO-l3)PtOts-l. An FC(0)O visible absorption cross section of (9.2 f 0.8) x cm2 at 641 nm is determined by three methods: the formation of FNO from FC(O)O, the removal of ethylperoxy radicals by FC(O)O, and comparison to CH3C1 photolysis. Kinetics measurements are made by monitoring the transient change in UV/vis absorption by FC(O)O, C2H502, HOz, 0 3 , and/or FC(0)Oz. The rate constants for the removal of FC(0)O by atmospherically important radicals are k12 = 3.6 x 10-11(T/300)-0~'cm3 s-' for NO, k27 = 7.5 x 10-12(T/300)-3,1cm3 s-' for NO2, and k25 = 1.2 x 10-11(T/300)-'.2 cm3 s-l for HO2. In contrast, the rate cm3 s-'. This precludes the suggestion that constant for reaction with ozone is very slow, kz9 < 3 x FC(0)O could participate in catalytic ozone depletion cycles.
I. Introduction Can hydroflorocarbon compounds (HFCs) lead to stratospheric ozone depletion? At first glance, the answer would seem no. The standard depletion mechanism occurs via catalytic removal of ozone by chlorine atoms.' Unlike chlorofluorocarbon compounds (CFCs), the HFCs contain no chlorine atoms. They are small chain, partially fluorinated, alkanes; therefore, even if they are not removed by reaction with OH radicals in the troposphere, these molecules are incapable of canying chlorine into the stratosphere. Upon more careful consideration, however, one finds that the atmospheric degradation of HFCs2 produces some unique products with properties distinctly different from those encountered from the typical hydrocarbon atmospheric degradation pathway^.^ Two of these are CF30 and FC(0)O. Because of their unusual chemical nature, it has been proposed4s5that they could participate in ozone depletion cycles of the form
-
+ 0, RO, + 0, RO, + 0 - RO + 0,
RO
(1)
(2)
where R represents CF3 or FC(0). Other reactions are possible, besides reaction 2, that provide alternative routes to regeneration of the RO species. These cycles have the net result of converting two odd-oxygen species to 0,. In principle, then, the question of ozone depletion by HFCs remains an open one, with potentially dire consequences for the use of these compounds as CFC replacements. Alkoxy radicals are well-known in atmospheric chemistry. Their chemistry proceeds via four possible reaction pathways3 isomerization, C-C bond fission, HC1 elimination,6or reaction with 0 2 to form HO, and a ketone. AU four paths are impossible for CF3O. Recent work has shown that this alkoxy radical rather
* Author for correspondence. @Abstractpublished in Advance ACS Abstracts, November 1, 1994. 0022-3654/94/2098-12294$04.50/0
facilely abstracts a hydrogen atom from hydrocarbon^,^^^ such as methane, but reacts only very slowly with o ~ o n e . ~ -The '~ methane reaction terminates the chain represented by reactions 1 and 2; thus, CF3O is ineffective in ozone depletion. The fluoroformyloxyl radical, FC(O)O, has only recently been identified in the lab~ratory.'~The analogous compounds with a hydrogen or chlorine atom replacing the fluorine atom are unstable. FC(0)O has a lifetime, with respect to dissociation, of a few seconds at room temperat~re'~ and considerably longer at stratospheric temperatures. This radical owes its stability to the tendency of fluorine to make strong bonds with carbon atoms. Without analogous species from which the chemistry of FC(0)O could be anticipated, it is important to investigate the reactions of this radical with atmospherically significant species in order to ascertain its role in stratospheric ozone depletion. This paper reports laser photolysis-UV/vis spectroscopic investigations of the kinetics of FC(0)O reactions with the atmospherically important species NO, NO2, 0 3 , and HO, and with C2H5O2 representing a "typical" organic peroxy radical. The reactions with NO, and peroxy radicals are fast; the slowest reaction is with NO,, which has a rate constant of 7.9 x cm3 at 295 K. In contrast, ozone reacts only very slowly with FC(0)O. The upper limit of k 3 x cm3 s-l implies that chain termination reactions will considerably dominate the catalytic ozone destruction cycle represented by reactions 1 and 2. The paper is organized as follows. A brief review of the apparatus is presented in section 11. In section I11 we report 193 nm (FC(0)0)2 photodissociation yields, the collisional relaxation rate for internally excited FC(0)O radicals, and the determination of the visible cross section for this radical. The kinetics measurements of FC(0)O reactions with a variety of reactants are presented in section IV. The atmospheric implications of these kinetics measurements form the subject of section V.
0 1994 American Chemical Society
Atmospheric Chemical Kinetics of FC(0)O
J. Phys. Chem., Vol. 98, No. 47, 1994 12295
11. Experimental Section The FC(0)O kinetics experiments were carried out using a temperature variable excimer laser flash photolysis-UV/vis spectroscopy apparatus that has been previously described.15.16 FC(0)O radicals were generated by 193 nm photolysis of the FC(0)O dimer that flowed through a 3.2 cm diameter by 51 cm long fused silica cell along with a suitable mixture of precursor gases. The excimer laser pulse was collimated by cylindrical optics into a beam approximately 2 cm x 2 cm in cross sectional area and directed longitudinally through the reaction vessel. Irradiation of about 1 Torr of dimer by a -400 mJ pulse succeeded in producing typically 3 x 1014 cm-3 FC(0)O radicals along with 6 x 1014cmP3fluorine atoms (see section 111). Real time detection of FC(O)O, FNO, HO2, C2H5O*,O3, and FC(0)02 species was accomplished by UV/vis absorption spectroscopy. Light from either a Xe arc lamp or a D2 lamp was directed through the cell anti-collinear to the photolysis beam. Apertures of 0.5 cm diameter were used to ensure overlap with the latter beam and to probe the nearly homogeneous central portion of the irradiated volume. The probe light was dispersed by a monochromator and detected by either a diode array or a photomultiplier tube. Near nascent and collisionally relaxed spectra of FC(0)O were obtained using a 0.64 m monochromator, 600 groove/" grating, and gated diode array detector at a resolution of about 0.6 A. UV spectra of (FC(0)0)2 and FNO were measured using the diode array with a 0.32 m monochromator and 147 groove/" grating at about 2 nm resolution. Kinetics measurements were performed with the lower resolution spectrometer using photomultiplier detection at fixed wavelengths of 641 nm for FC(0)O (a = 9.2 x cm2, present work), 3 11 nm for FNO (a = 6.0 x cm2, ref 17 and present work), 254 nm for O3 (a = 1.15 x cm2, ref 18), 240 nm for C2H502 ( a = 4.6 x cm2, ref 16), 240 nm for FC(0)02 (a = 2.8 x cm2, ref 19), and 230 nm for H02 (a = 2.2 x cm2, ref 20). Transient absorbances were converted into concentration versus time profiles via Beer's law. The signals at wavelengths below 350 nm had to be corrected for ozone formed both inside and outside the cell via O,+
0
193nm-0+0
+ 0, + M - 0,+ M
(3)
where the latter reaction has a rate constantI8of 6 x cm6 s-l. The transient absorption signal was corrected for the ozone formed outside the cell by subtracting a background signal obtained with dimer and 0 2 omitted from the gas mixture. For 0 2 pressures of 5 12 Torr within the cell, the density of 0 atoms formed internally was cm-3 and its effect was judged negligible. For 0 2 pressures larger than about 50 Torr, the density of 0 atoms was determined from the formation of O3 in a gas mixture from which the dimer was omitted. This concentration of 0 atoms along with the appropriate oxygen atom chemistry was then added to the model used to determine the relevant rate constant (for example for the 80 Torr of 0 2 used in the FC(0)O HOz measurements, -6 x 1013cm-3 0 atoms and the reaction 0 HOZ were added to the reaction model). The detection electronics were operated at a bandwidth of 100 kHz for most experiments. The FNO measurements at 311 nm and the FC(O)O* decay experiments were conducted with 10 kHz and 1 MHz bandwidths, respectively. For some of the faster reactions a slight delay in the rise time of the absorption
+
+
signal is attributable to the finite bandwidth. This was modeled by adding to the kinetic equations the expression dddt = kdet(Voe-"''R1- s)
which describes the action of an RC low pass filter. Here, s is the photomultiplier signal, R is the species being monitored, a is its absorption cross section, 1 is the pathlength, VOis the initial detector output voltage, and kdet is the response rate of the detector (bet = 6.2 x lo5 s-l for a 100 kHz bandwidth). Most of the experiments were carried out in mixtures of 1 Torr of dimer in an N2 buffer at a total pressure of 200 Torr. Pressures of 25, 50, and 700 Torr were utilized to investigate the pressure dependence of certain reactions (e.g., with NO and N02). Small quantities, in the range of 0.1 Torr, of NO, N02, and O3were added to the gas mixture to study the reactions of these species with FC(0)O. The formation of F atoms by the photolysis of (FC(O)O)z, in addition to FC(0)O radicals, enabled the generation of a variety of reaction partners (upon the addition of a suitable precursor) via the following schemes:
+ C2H6 - C,H, + HF C2H5 + + M - C2H502+ M F
0 2
(4)
(5)
to form ethyl and ethylperoxy radicals,
+ CO + M -,FCO + M FCO + 0, + M - FC(O)O, + M F
(6)
(7)
to form FCO and FC(0)02 radicals, and
+ H, -H + HF H + 0, + M--0, +M F
(8)
(9)
to form hydroperoxy radicals. In each case, the amounts of precursor (e.g., Hz) and 0 2 (when added) were adjusted to ensure that the reaction partner (e.g., HO2) was formed on a fast time scale compared to the subsequent reaction with FC(0)O; typically the formation half-life was on the order of a few microseconds or less. The (FC(0)0)2 precursor was synthesized by Dr. Darryl DesMarteau of Clemson University and was provided to us by the Alternative Fluorocarbon Environmental Acceptability Study. Temperature control of the reaction mixture was achieved by a Neslab ULT-80dd recirculating controller using methanol and ethylene glycol as the heat exchange fluid below and above room temperature, respectively. The gas mixture was precooledpreheated prior to entering the cell. Its temperature was determined from the set point using a correction obtained from a separate measurement of the gas temperature via a probe inserted into the center of the cell. The total pressure of the reaction mixture was measured by Baratron manometers at the entrance and exit of the cell. The gas flows were controlled by Tylan flow controllers, with the individual flow rates measured from the rate of pressure increase upon their expansion into a fixed volume.
111. (FC(0)O)z Photolysis
A. Product Distribution and FC(0)O Visible Absorption Cross Section. The UV spectrum of (FC(0)0)2 exhibits a continuous absorption band (at 2 nm resolution) that has an onset at about 250 nm and increases in intensity to 190 nm and shorter wavelength. The absorption cross section at 193 nm, as measured in this study, is 2.8 x cm2. A 193 nm photon
Maricq et al.
12296 J. Phys. Chem., Vol. 98, No. 47, I994
UV c r o s s sections
(FC(0)0)2 photolysis 0 15,
I
n
N
“6, 0.04
“1
k
C 0
.4
4
0
$
0.02
111
rn
0 L3 0
__
0- 0 0
200
250
300
350
wavelength ( n m ) 200
300
400
500
700
600
BOO
0.06
wavelength ( n m )
Figure 1. UV spectra of (FC(O)O)&% and (FC(O)O)dCzHdOz/Nz gas mixtures 20 ,us after photolysis by a 193 nm laser pulse.
n
N
5 0.04
carries 148 kcaYmo1 of energy. The absorption of such a photon provides roughly twice the energy needed to break simultaneously the 0-0 bond, with an estimated dissociation energy2‘ of 27 kcaYmo1, and to overome the 21 kcaYmol activation barriers22for dissociating each of the two FC(0)O fragments into F COP(on their ground electronic state potential surface). This suggests that a substantial fraction of 193 nm photons absorbed will yield 2F 2C02 as products. The UV spectrum of a (FC(0)0)2/N2 gas mixture recorded 10 ps following photolysis is illustrated in Figure 1. It exhibits two principal features: the spectrum13of FC(0)O in the 500750 nm region and a structureless feature at 254 nm (trace marked “w/o C2H6 and 02”). The latter is attributed to ozone formed external to the reaction cell from the 193 nm photolysis of 0 2 . When ethane and 0 2 are added to the gas mixture, a strong absorption feature appears at 240 nm, whereas the absorbance of the FC(0)O remains essentially unchanged. Comparison of this band to the reference spectra presented in Figure 2B shows that the new feature belongs to the ethylperoxy radical. The formation of this species indicates that 193 nm photolysis of (FC(0)0)2 breaks both the 0-0 and F-C bonds, as expected, to produce a mixture of fluorine atoms and FC(0)O radicals, the former which subsequently react with ethane to produce ethylperoxy radicals via reactions 4 and 5. One method to determine the quantum yield for F atom production is by comparison to the C1 atom yield from CH3Cl photolysis at 193 nm under identical conditions, except for the switch in precursor. Assuming a 100% photodissociation yield from CH3C1 provides a measure of the relative fluorine atom yield from dimer photolysis as
+
+
Here, [C2H502]dimerrepresents the ethylperoxy concentration measured after photolysis of (FC(0)0)2. Both the literature18 and our own measurements indicate that a(CH3C1) = 7.2 x cm2. [ R ~ ~ ] C H represents ,C~ the total CH302 plus C2H502 concentration obtained from the photolysis of methyl chloride in a gas mixture containing ethane and oxygen. These two species are formed by the addition of 0 2 to the methyl radical and to the ethyl radical produced from the C1 C2H6 reaction. The spectrum of CH3O2 is very similar to that of C2H5O2 (shown in Figure 2B) with an optical cross section in the region of the band maximum differing by only about 5%; thus, the absorbance provides a measure of the total RO2 concentration. Table 1
+
.-0
4
0 Q)
rn 0.02 111
rn
0 L3 0
0.00
200
250
300
wavelength ( n m ) Figure 2. Reference spectra for the molecules probed in the course of this study. Arrows indicate probe wavelengths. Panel A: spectra of FC(O)O*and FNO. Panel B: spectra of HOz, C2Hs02, and 03. The FC(0)O spectrum is shown in Figures 1 and 4. collects the results of the pholysis yield measurements. At 295 K the relative yield of fluorine atoms is @(F) = 0.70 f 0.12, where the error is a 95% confidence measurement obtained as a/& from the scatter of three measurements. This error is consistent with the uncertainties in the quantities entering eq 10. If one assumes that the absorption of a 193 nm photon by (FC(0)02 must lead at least to 0-0 bond fission, then the relative yield of FC(0)O is 0.30. A second method for ascertaining the product distribution is by photolysis of the dimer in a gas mixture containing an excess of ethane and oxygen. This produces a mixture of C2H5O2 and FC(0)O radicals, which subsequently decay by their respective self-reactions and by FC(0)O
+ C2H502- products
(1 1)
Because the self-reactions are slow compared to the cross reaction, vide infra, the initial FC(0)O concentration can be deduced from the difference between [C2H502]0 and the plateau level reached after the supply of FC(0)O is exhausted (see Figure 3A). In practice, the slow second-order decay at long time of the ethylperoxy radicals by their ~elf-reaction~~ (kobs = (1.6 x 10-13)e-1m’Tcm3 s-l) is fit for the “effective” initial concentration, Le., for [C2H502]0 - [FC(O)O]o, as shown by the nearly horizontal dashed line in Figure 3A. Upon equating [C2H502]0 = [Flo, we obtain a photolysis yield (see Table 1) of @(F) = 0.67 3~ 0.05 at 295 K, where the error is due to scatter in the three measurements and to a 5% uncertainty in a(C2H502). By difference @(FC(O)O) = 0.33 f 0.05; thus, each photon absorbed by (FC(0)0)2 leads to at least one bond fission, as expected above. Independent quantification of the FC(0)O photolysis product is provided by the measurement of the FNO concentration after
Atmospheric Chemical Kinetics of FC(0)O
. I . Phys. Chem., Vol. 98, No. 47, 1994
TABLE 1: Photolysis Yield“ Relative to
12297
from FC(0)O + C Z H ~ O ~
from FNO [FC(O)Olo
u(FC(O)O)~
0.75 f 0.12 0.72 f 0.12
7.2 f 0.5 6.4 f 0.4 7.3 f 0.5
[FC(O)Olo 2.3 f 0.4‘ 2.5 f 0.4 2.8 f 0.4
2.6 f 0.4 3.6 f 0.6
295 295 295
0.81 f 0.12 0.61 f 0.10 0.69 f 0.10
7.1 f 0.5 5.3 f 0.3 7.2 f 0.5
3.1 f 0.4 2.8 f 0.4 3.6 f 0.4
2.7 f 0.4 2.7 f 0.4 3.0 f 0.4
263 233 233
0.70 f 0.10 0.71 f 0.11 0.77 f 0.12
6.9 f 0.5 6.0 f 0.4 8.2 f 0.5
3.8 f 0.4 3.4 f 0.4
3.8 f 0.5
(10.8 f 1.1) x (8.4 f 0.8) x average = (9.6 f 1.0) x (10 f 0.8) x (8.7 i 0.7) x (8.9 f 0.8) x average = (9.2 i 0.8) x (9.0 f 1.0) x 10-18
4.5 f 0.5
(9.0 f 1.0) x 10-18
CHKl WF)
Flo
353 323 323
expt T(K)
Absolute yields have units 1014 ~ m - ~Units . are cm2. Determined from initial absorbance at 641 nm assuming u(FC(0)O) = 9.2 x
A=240nm
10%(FCO,),
10 Torr 0, I 1
6 -
4 -
[F],-[FC(0)O],=(3.f3+0.Z)101‘
0
(31n-~
u 0.0
0.5
1.0
1.5
2.0
2.5
time (ms) Figure 3. Determination of FC(0)O photolysis yield at 295 K. Panel A: decay of C2H502 due to their removal by FC(0)O radicals. Panel B: conversion of FC(0)O to FNO via reaction with nitric oxide.
irradiation of a dirner/NO/Nz gas mixture, as shown in Figure 3B. The FNO is detected by its absorption at 311 nm (Figure 2A). Our measured cross section of o(FN0, 3 11 nm) = (6.0 f 0.2) x cm2 is in excellent agreement with the value reported by Burley et al.” The initial sharp rise in FNO concentration occurs from the rapid reaction FC(0)O
+ NO - FNO + CO,
(12)
whereas the subsequent slow rise is due to
F
+ NO + M -FNO + M
(13)
The “initial” FNO yield must be corrected for the rapid reaction
F
+ FC(0)O - FC(0)OF
(14)
the kinetics of which is discussed in section IVC. A fit of the data in Figure 3B to a model consisting of reactions 12-14
yields a value of [FC(O)O]o = (3.0 k 0.2) x 1014 ~ m - ~ . Combined with [F]o = (7.2 f 0.3) x 1014 ~ m - this ~ , implies Q(FC(0)O) = 0.29 k 0.03, in very good agreement with the ethylperoxy data discussed above. The average from three data sets is Q(FC(0)O) = 0.31 f 0.03. Having independently measured both the initial FC(0)O and F atom concentrations, we can combine this data with measurements of the FC(0)O absorbance to arrive at o(FC(O)O, 641 nm) = (9.2 f 0.8) x cmz (see Table 1). The above measurements of photolysis yield and FC(0)O absorption cross section were repeated over the temperature range 233-323 K. At 353 K only the photolysis yield was determined. This was based on setting [F]o equal to the initial ethylperoxy concentration formed from the photolysis of a (FC(0)0)2/C~Hd02/N2gas mixture and finding [FC(O)O]o from the initial absorbance at 641 nm, assuming the cross section remains at 9.2 x cmz. This assumption is consistent with the result that, within the 10% error of the present measurements, the FC(0)O absorption cross section is temperature independent over the 233-323 K range. In contrast, the branching ratio of photolysis products varies somewhat with temperature. The yield of fluorine atoms increases with increasing temperature from @(F) = 0.64 at 233 K to Q(F) = 0.76 at 353 K, whereas the FC(0)O yield decreases from (P(FC(0)O) = 0.36 to (P(FC(0)O) = 0.24. This temperature dependence is consistent with the view that 193 nm photolysis of (FC(0)O)z proceeds via 0-0 bond cleavage followed by dissociation of the highly excited FC(0)O fragments that are formed (see section IIIB). Increasing the temperature raises the internal energy of the dimer molecules. This thermal energy combines with the excess energy available from the absorption of a 193 nm photon to produce slightly “hotter” FC(0)O fragments at 353 K as compared to 233 K. Consequently, the probability for dissociation of the FC(0)O radical into F CO2 increases with increasing temperature, as observed. There are two previous reports of the FC(0)O visible absorption cross section at room t e m p e r a t ~ r e . ’ ~Both . ~ ~ values are significantly larger than the one reported here, 1.5 x lo-” and 1.3 x lo-’’ cm2 at 641 nm as compared to 9.2 x cm2. The discrepancy likely arises from three problems suffered by the previous measurements. They were based, indirectly, on the initial slope of the FC(0)O absorbance versus time profile, a method more prone to error than one based on the value of the absorbance itself. They assume a branching ratio of 0.8 f 0.2 for the production of FC(0)O from the self-reaction of FC(0)02 radicals. Finally, they are sensitive to secondary chemistry, such as
+
FC(0)O
+ FC(O)O,
-
products
(15)
Maricq et al.
12298 J. Phys. Chem., Vol. 98, No. 47, 1994
1.2) x 10-13]Pt,ts-l for vibrational relaxation by nitrogen at 295 K.
( F C ( 0 ) 0 ) 2 photolysis .Laser 193 nm 5 0 0 mJ 0.14
0.12
IV. FC(0)O Reaction Kinetics
-
3:
0.10
0.08
0.06
0.04
0.02
0 . 0 0 '', 12.5
'
'
'
'
'
13.0
'
'
'
'
'
13.5
'
'
'
'
.__ '
'
'
14.0
'
'
'
'
'
3
wavenumber (10 c m
'
'
,
15.0
14.5
-1
'
)
The present experiments should provide a more reliable measure of the absorption cross section since they are based on multiple direct and self-consistent determinations of the FC(0)O concentration. B. Relaxation of FC(O)O*. The formation of both FC(0)O radicals and F atoms suggests that the absorption of a 193 nm photon by (FC(0)0)2 leads to 0-0 bond cleavage and the production of two FC(0)O fragments, some of which have sufficient internal excitation to dissociate into F and COT. The extent of internal excitation in the fluoroformyloxyl fragment is better evidenced by the comparison of near nascent and relaxed spectra of the B2A1-X2B2 transition, shown in Figure 4. The lower trace, recorded 1.3 p s after photolysis and at a total pressure of 1.5 Torr, shows the spectrum after an average of 3 collisions; the upper trace, at 200 Torr and 100 ps, shows the spectrum following about 3 x lo4 collisions. The striking features from this comparison are the nearly complete absence of the origin, marked TO,and the dominance of hot bands in the near nascent spectrum. These hot bands are primarily, but not exclusively, associated with the v3 bending mode. The near nascent spectrum also exhibits a significantly enhanced continuous spectrum underlying the sharp features. This presumably occurs because of the much larger number of transitions possible from the vibrationally hot, as compared to the relaxed, FC(0)O. Because of the internal excitation, the formation of FC(O)O, as monitored by its absorption at 641 nm, appears delayed. This is apparent in some of the concentration versus time curves displayed in section IV. The delay can be described by a phenomenological relaxation process,
+ M - FC(0)O + M
FC(0)O
F
Figure 4. Comparison of the near nascent FC(0)O spectrum, from the 193 nm photodissociation of (FC(0)0)2, to the relaxed spectrum.
FC(O)O*
A. Overview. The photolysis of (FC(0)0)2 provides a direct and essentially instantaneous source of FC(0)O radicals with which to perform real time kinetics measurements. That F atoms are also produced is both a help and a hinderance. It is a hinderance because the fluorine atoms interfere with rate constant determinations for the reactions of FC(0)O with stable species such as NO, NO2, and 0 3 . Fortunately, this is a minor nuisance since the fluorine atoms can be scavenged, e.g., by 0 2 , andor the reactions between F and other species can be incorporated into the reaction model. From another perspective, the fluorine atoms are helpful as a starting reagent for producing radical reactants such as HO2, CzH502, and other species. Below, we report rate constant measurements for the reactions of FC(0)O with itself, F, C2H5, C2H5O2, FCO, FC(0)02, HO2, NO, NO?, and 03. A complete mechanism for these reactions is provided in Table 2; however, only a much simplified subset is, in general, necessary to model the reaction between FC(0)O and a particular reactant X. When X = F or X is a stable species, such as NO, NOz, or 0 3 , the subset consists of the reactions
Measurements of FC(0)O rise times in gas mixtures with 20 < P,,,< 200 Torr yield a first-order rate constant of [(3.5 f
+ FC(0)O - FC(0)OF
FC(0)O
+ FC(0)O
F
-
(FC(O)O),
+ X - products
(14)
(17)
In this case, FC(0)O is removed not only by X but also by reactions with itself and with F atoms. If X is a radical species, such as C2H5, C2H5O2, FCO, FC(0)02, or HOz, then the simplified model becomes
F
+ precursor - X
FC(0)O
+ X - products
+ FC(0)O - (FC(O)O), X + X - products
FC(0)O
(17)
In this situation, the concentration of the precursor for X is made sufficiently large that its reaction with F dominates reaction 14, whereupon the latter can be, to first order, ignored. The mechanism, thus, consists of the two self-reactions and the cross reaction between the two radical species. The results of the kinetics measurements, along with the experimental conditions, are listed in Table 3. B. Self-Reaction. Following the photolysis of (FC(0)0)2 in nitrogen, the FC(0)O absorption at 641 nm decays quickly, as seen in Figure 5 (trace marked 0 Torr 0 2 ) . When a small amount of 0 2 is added, there is a rapid initial decay followed by a much slower one. A substantial increase in the 0 2 level eliminates the rapid initial decay, leaving only a slow, secondorder, decay of FC(0)O. The rapid initial decay originates from the reaction of F atoms with FC(0)O radicals. Added 0 2 competes with FC(0)O for the F atoms via F
(16)
+ X -products
+ 0, + M -FO, + M
(18)
It reduces the removal of FC(0)O by atomic fluorine and leads to a much longer decay time. The self-reaction rate constants extracted from the FC(0)O decay observed under a variety of
Atmospheric Chemical Kinetics of FC(0)O
J. Phys. Chem., Vol. 98, No. 47, 1994 12299
TABLE 2: Reactions to Model FC(0)O Chemistry reaction
rate constant'
+ FC(0)O - (FC(0)0)2 + F FC(0)OF FC(0)O + CzH5 - products FC(0)O + CzH502 - products FC(0)O + FCO FzCO + CO2 FC(0)O + FC(O)O2 products FC(0)O + HOz - HF + COz + 0 2 FC(0)O + NO FNO + C02 FC(0)O + NO2 -.FNOz + COz FC(0)Oz + 02 FC(0)O + FC(0)O + FOz - products FC(0)O + c2H6 -products FC(O)O* + M - FC(0)O + M
17. 14.
-
FC(0)O FC(0)O
21. 11.
24. 15. 25. 12.
27. 29. 19.
0 3
22.
16. 4.
--
F+CzH6-CzHs+HF HF F H2-H F CO M FCO M F + N O +M-FNO + M F NO2 M FNOz M F 0 2 M - FO2 M F 0 3 FO 0 2
+
8.
+ +
6. 13. 28.
+
+
-
+
+ +
18.
+
30. 5. 9. 7. 3.
-
+
+
+
+
CZHS+O~+M-C~H~OZ+M H 0 2 M-HO2 M FCO 0 2 M FC(0)Oz M 0 02 M - 03 M
- ++
+ + + + + +
+
Other 23. 20a. 20b.
+
--.
FC(0)Oz FC(0)Oz 2 FC(0)O FC(0)02 + FC(0)Oz (FC(0)O)z FCO FCO F2CO CO C2H5 C2H5 products 0 HOz- OH 02 FO FO-F F 02 0 CO M COz M 0 NOz-NO 0 2 O(lD) CO COz O('D) 0 3 2 0 2
--
+ + + + + + + + + + -. + + +
26.
+ +
+ +0 2 0 2
--
Rate constants are obtained from refs 18 and 27 unless otherwise noted. Rate constants are at 295 K unless an explicit temperature dependence is indicated. Measured in the present study. 0 2 partial pressures (an example is given in Figure 6A) have an average value of (7.1 & 0.7) x cm3 s-l at 295 K, as given in Table 3, and are independent of pressure in the range 60 < P,,,< 220 Torr. However, this value must be considered an upper limit because of the possibility that the reaction
FC(0)O
+ FO,
-
t' h
m
(19)
TL295 K PLO,=22OTorr
I
E f 0
products
h = 6 4 1n m
3
2
3
v 7
contributes to the observed decay at 641 nm. An altemative method to determine the self-reaction rate constant, which avoids interference form an unwanted radicalradical reaction, is to convert the fluorine atoms produced by photodissociation of the dimer into FC(0)O radicals. This can be accomplished by adding carbon monoxide to the gas mixture, whereupon reactions 6, 7, and FC(O)O,
+ FC(0)02- 2 F C ( 0 ) 0 + 0,
(20a)
achieve the desired conversion. Sufficient CO and 0 2 are added to ensure that reactions 6 and 7 occur on a microsecond time scale. Because it is a radical-radical reaction, the time scale of the FC(0)02 self-reaction cannot be controlled in this manner; however, it is rapid enough that the FC(0)O formation is essentially complete within 100 pus after photolysis (this is evident in Figure 6A and in Figure 10B). The resultant FC-
0 h
0
v
0
1
G Y
0
0
500
1000
1500
2000
time ( p s ) Figure 5. Decay of the FC(0)O absorbance at 641 nm under various 0 2
concentrations.
(0)O decay occurs over a 20 ms time scale and yields a rate constant of k17 = (5.5 & 0.6) x cm3 s-l at 295 K for the self-reaction. This, as opposed to the determination in the 0 2 buffer gas, is the preferred value of the rate constant. Below The error in the self-reaction rate constant arises from noise in the concentration versus time profile and from the uncertainty in u(FC(0)O). The 2u error in the latter is approximately lo%,
12300 J. Phys. Chem., Vol. 98, No. 47, 1994 TABLE 3: Kinetic Resultsn T (K) Ab (FC(0)O)z
precursor
Maricq et al. OZ
Self Reaction 56 67 199 209 11 240 208 218 11 222 11 220 205 215 52 62
353 323 323 295 295 295 295 295
64 1 641 64 1 64 1 64 1 64 1 64 1 64 1
1.o 1.o 1.o 1.o 1.0 1.o 1.o 1.o
295 295
64 1 64 1
1.1 1.1
107 CO 103 CO
11 11
263 233
64 1 64 1
1.0 1.1
112 co 114CO
353 353 323 323 295 295 263 233
64 1 64 1 64 1 64 1 64 1 64 1 64 1 64 1
1.o 0.75 1.o 1.1 1.1 1.1 1.1 1.o
11 11 F
295 295 295
64 1 64 1 64 1
1.o 1.1 1.1
113 CO
ptot
FIo
[FC(O)OIo
7.2 7.3 7.3 7.2 7.2 5.5 4.8 7.1
2.3 2.5 3.0 3.3 3.3 2.7 2.3 3.1
8.0 6.7
3.8 3.2
6.2 6.8
4.0 4.5
7.2 5.6 7.3 7.3 9.2 9.4 6.2 6.8
2.3 1.9 3.1 3 .O 5 .O 4.4 4.0 4.5
(4.9 f 1.0) x (7.2 f 1.5) x (7.7 f 1.3) x (11.5 f 1.6) x (1.1 f 0.3) x (1.3 f 0.3) x (2.0 f 0.4) x (2.7 f 0.6) x
10-12
9.2 7.2 7.2
4.5 2.9 2.9
(8.5 f 0.9) x (7.8 f 1.0) x (7.8 f 1.0) x (8.3 f 0.9) x
lo-" lo-" lo-" lo-"
rate constantd
average
2.2 (22% 2.5 (22% 25 CzH6
215 218
average
+
237 240 FC(0)O 60 220 60 218 60 170 214 214
FC(0)O + CZH5 60 240 257
average 323 323 323 295 295 295 295 263 263 233 233
240 240 64 1 240 64 1 240 64 1 240 64 1 240 64 1
1.1 1.o 1.o 0.9 1.o 1.o 1.1 1.1 1.1 1.o 1.o
2.5 Cz% 2.5 C2H6 2.5 CzH6 2 CZH6 2.3 CzH6 2.4 CzH6 2.5 CzH6 2.5 CzH6 2.5 C2H6 2.4 CzH6 2.4 CzH6
295
64 1
1.1
107 CO
FC(0)O 11 12 11 10 11 11 12 11 11 11 11
+ CzHsOz
FC(0)O
207 64 207 66 63 220 240 203 203 204 204
-
lo-" lo-"
lo-" lo-"
6.6 7.5 6.6 7.1 9.0 7.2 7.2 6.8 6.8 6.8 6.8
2.4 2.7 2.4 3.1 4.0 3.6 2.9 4.1 4.1 4.6 4.6
(1.3 f 0.2)) x (1.2 f 0.2) x lo-'[ (7.0 f 1.5) x (7.0 f 2.0) x 10-l2 (5.7 f 1.5) x (10 f 3) x 10-12 (6.9 f 1.3) x (1.25 f 0.2) x lo-" (1.15 f 0.2) x lo-" (1.25 f 0.2) x lo-" (1.4 f 0.2) x lo-"
8.0
3.8
(2.4 f 0.3) x lo-"
8.0
3.8
(3.5 f 1.0) x 10-12
5.2 5.5 6.5 6.8
2.3 2.6 3.2 3.6
(1.1 f 0.2) x (1.2 f 0.2) x (1.5 f 0.3) x (1.6 f 0.3) x
10-11 10-1' lo-" lo-"
7.0 7 .O 6.4 6.4 7.2 9.0 7.2 7.2 7.2 7.2 6.9 6.9 6.8 6.8
3.0 3.0 2.9 2.9 4.0 4.5 4.0 4.0 4.0 4.0 3.9 3.9 4.5 4.5
(3.8 & 0.6) x (3.5 f 0.8) x (3.4 f 0.5) x (3.7 f 0.7) x (3.4 f 0.5) x (2.2 i 0.5) x (3.7 f 0.7) x (3.3 f 0.6) x (3.7 f 0.7) x (3.9 f 0.7) x (3.5 f 0.6) x (3.4 f 0.6) x (3.9 f 0.7) x (3.6 f 0.7) x
lo-" 10-11
6.4 6.4 6.4
2.6 2.9 2.7
5.5
2.8
(6.6 i 1.3) x (6.4 f 1.2) x (6.2 f 1.6) x lo-':! (6.4 f 1.0) x (5.6 f 0.9) x
+ FCO 215
+ FC(0)Oz 225 FC(0)O + HOz
FC(0)O 11
295
2408~641
1.o
105 CO
323 295 263 233
2308~641 2308~641 230&641 203&641
1.1 1.o 1.o 1.1
23 H:! 11 HZ 23 Hz 23 H2
323 323 323 323 295 295 295 295 295 295 263 263 233 233
64 1 64 1 64 1 64 1 64 1 64 1 64 1 64 1 641 641 64 1 64 1 64 1
1.o 1.o 1.1 1.1 1.o 1.o 1.o 1.o 1.o 1.o 1.1 1.1 1.o 1.o
0.083 NO 0.040 NO 0.095 NO 0.048 NO 0.046 NO 0.077 NO 0.053 NO 0.053 NO 0.052 NO 0.052 NO 0.094 NO 0.049 NO 0.096 NO 0.048 NO
323 323 323
64 1 64 1 64 1
1.1 1.1 1.1
0.22 NO2 0.084 NO:! 0.033 NO2
86 244 83 234 86 243 87 245 FC(0)O NO 60 60 219 219 39 60 63 121 214 536 215 215 212 212 FC(0)O NO2 42 237 42 234 42 232
295
64 1
1.1
0.12 NO2
43
64 1
-
(15 f 2) x 10-13 (6.2 f 0.6) x (5.1 f 0.5) x (5.8 f 0.6) x (8.1 f 0.8) x 10-13 (7.4 f 0.7) x 10-13 (7.4 f 0.7) x 10-13 (7.0 f 0.7) x (7.1 & 0.7) x (5.0 & 0.5) x (5.9 5 0.6) x 10-13 (5.5 f 0.6) x (4.7 f 0.5) x 10-13 (4.5 f 0.5) x 10-13
+
+
average 55
-
lo-" lo-" lo-" lo-" lo-" lo-" lo-" lo-" lo-" lo-"
Atmospheric Chemical Kinetics of FC(0)O
J. Phys. Chem., Vol. 98, No. 47, 1994 12301
TABLE 3 (Continued) T (K) 295 295
Ab
(FC(0)0)2
precursor
0 2
Ptm'
[FIo
[FC(0)010
641 64 1
1.1 1.1
0.074 NO2 0.033 NO2
43 43
56 56
5.5 5.5
2.8 2.8
295 295 295
64 1 64 1 64 1
1.1 1.1 1.1
0.18 NO2 0.12 NO2 0.032 NO2
42 42 42
224 224 225
5.5 5.5 5.5
2.8 2.8 2.8
263 263 263
64 1 64 1 64 1
1.1 1.1 1.1
0.20 NO2 0.087 NO2 0.031 NO2
43 43 43
235 232 23 1
5.4 6.1 6.1
3.0 3.4 3.4
233 233 233
64 1 64 1 64 1
1.o 1.o 1.o
0.21 NO2 0.088 NO2 0.035 NO2
44 44 44
237 235 233
3.8 4.5 5.9
2.3 3.0 3.8
323 323 295 295 263 263 233 233
1 .o 1.0 1.o 1 .o 1 .o 1.0 1.o 1.o
64 1 254 64 1 254 64 1 254 64 1 254
FC(0)O + 0
107 CO 107 CO 103 CO 103 CO 112 co 113 CO 113 CO 113 CO
11 11 11 11 11 11 11
11
224 224 218 218 237 239 239 240
rate constant! (5.7 f 0.9) x
average
-
average
-
average
-.
average
-
(5.4 f 0.9j x 10-12 (5.6 f 0.9) x (7.0 f 0.9) x (7.5 f 0.9) x (8.4 f 1.0) x (7.6 f 1.2) x (11 f 1.8) x (7.8 f 1.3) x (10 f 1.6) x (9.8 f 1.5) x (1.9 f 0.2) x lo-" (1.8 f 0.15) x lo-" (1.6 f 0.15) x lo-" (1.8 f 0.2) x lo-"
3
6.9 6.9 6.4 6.4 6.2 6.2 6.8 6.8
2.6 2.6 3.2 3.2 3.9 3.9 4.4 4.4
7 x cm3 s-l at 295 K. Part of the reason for the larger rate constants is due to the larger value, by about 60%, used for a(FC(0)O) in the previous studies (see section IIIA). Another reason is because of the added complexity incurred when forming FC(0)O as a secondary product of FC(0)02 as opposed to forming it directly via photolysis. For example, both groups report difficulties in modeling the rise of FC(0)O given the known rate of FC(0)02 self-reaction. C. Reaction with F Atoms. The rate constant for the reaction between fluorine atoms and FC(0)O radicals was determined by monitoring the decay of the latter species following photolysis of (FC(0)0)2/N2 gas mixtures. Because the radicals decay rapidly in the absence of 0 2 relative to the FC(0)O self-reaction, as illustrated by Figure 5, and since fluorine atom recombination is very slow, only the reaction
F
+ FC(0)O - FC(0)OF
(14)
is needed to model the FC(0)O disappearance, with reaction 17 providing a small correction. An example of a fit for the single parameter k14 is provided in Figure 6B for T = 295 K and Ptot= 60 Torr; the results at other temperatures and pressures are provided in Table 3. Errors in the rate constant determinations arise from three sources: signal noise, an uncertainty of 10% in a(FC(0)O) and, hence, in [FC(O)O]o, and an uncertainty of 5% in [Flo. The latter two translate to errors of f 1 5 % and f 7 . 5 % in the rate constant, respectively. The noise error varies depending on the quality of data, but contributes