A study of the IR and UV-Vis absorption cross-sections, photolysis and OH-initiated oxidation of CF3CHO and CF3CH2CHO

Article abstract of DOI:10.1039/B315941H

Infrared and ultraviolet-visible absorption cross-sections, effective quantum yields of photolysis and OH reaction rate coefficients for CF ₃CHO and CF₃CH₂CHO are reported. Relative rate measurements at 298(2) K and 1013(10) hPa, give k(OH + CF₃CHO)/k(OH + CH₃CH₃) = 2.00(13), k(OH + CF₃CH ₂CHO)/ k(OH + CH₃CH₂OH) = 1.21(5) and k(OH + CF₃CH₂CHO)/k(OH + HC(O)OC₂H₅) = 3.51(9) (2σ). The effective quantum yield of photolysis was measured under pseudo-natural conditions in the European simulation chamber, Valencia, Spain (EUPHORE). Over the wavelength range 290-400 nm, the effective quantum yields of photolysis for CF₃CHO and CF₃CH₂CHO are less than 2 × 10^-2 and 4 × 10^-2, respectively. The tropospheric lifetimes are estimated to be: τ_OH(CF₃CHO) ～ 26 days; τ _photol(CF₃CHO) > 27 days; τ_OH(CF ₃CH₂CHO) ～ 4 days; τ_photol(CF ₃CH₂CHO) > 15 days.

Full text of DOI:10.1039/B315941H

View Article Online / Journal Homepage / Table of Contents for this issue
R E S E A R C H P A P E R
A study of the IR and UV-Vis absorption cross-sections, photolysis
and OH-initiated oxidation of CF3CHO and CF3CH2CHOy
a
Stig R. Sellevag, Tanya Kelly, Howard Sidebottom and Claus J. Nielsen*
b
b
a
˚
a
Department of Chemistry, University of Oslo, P.O. Box 1033, Blindern, N-0315 Oslo, Norway.
E-mail: claus.nielsen@kjemi.uio.no; Fax: þ47 22 85 54 41; Tel: þ47 22 85 56 80
Department of Chemistry, University College Dublin, Belfield Dublin 4,
Ireland
b
Received 9th December 2003, Accepted 27th January 2004
F|rst published as an AdvanceArticle on the web 23rd February 2004
Infrared and ultraviolet-visible absorption cross-sections, effective quantum yields of photolysis and
OH reaction rate coefficients for CF3CHO and CF3CH2CHO are reported. Relative rate measurements at
298(2) K and 1013(10) hPa, give k(OH þ CF3CHO)/k(OH þ CH3CH3) ¼ 2.00(13), k(OH þ CF3CH2CHO)/
k(OH þ CH3CH2OH) ¼ 1.21(5) and k(OH þ CF3CH2CHO)/k(OH þ HC(O)OC2H5) ¼ 3.51(9) (2s). The
effective quantum yield of photolysis was measured under pseudo-natural conditions in the European
simulation chamber, Valencia, Spain (EUPHORE). Over the wavelength range 290–400 nm, the effective
quantum yields of photolysis for CF3CHO and CF3CH2CHO are less than 2 ꢀ 10ꢁ2 and 4 ꢀ 10ꢁ2, respectively.
The tropospheric lifetimes are estimated to be: tOH(CF3CHO) ꢂ 26 days; tphotol(CF3CHO) > 27 days;
tOH(CF3CH2CHO) ꢂ 4 days; tphotol(CF3CH2CHO) > 15 days.
From pulsed laser photolysis-resonance fluorescence measure-
ments, Laverdet et al.9 reported an Arrhenius expression equal
to 3.5(1.0) ꢀ 10ꢁ12exp[ꢁ488(57) K/T] cm3 moleculeꢁ1 sꢁ1 over
the temperature range 233–313 K giving a rate constant of
6.8 ꢀ 10ꢁ13 cm3 moleculeꢁ1 sꢁ1 at 298 K. As can be seen, there
is a relatively large scatter among the results of these studies.
Theoretical studies on the OH reaction with CF3CHO have
been carried out by Francisco,10 Rayez et al.11,12 and Chandra
et al.13 Francisco and Williams14 have measured the infrared
absorption intensities of CF3CHO in the wavenumber region
2890–801 cmꢁ1. Ultraviolet-visible absorption cross-sections
of CF3CHO have been determined by Francisco and
Williams,14 Libuda15 and Meller et al.16
The photolysis of CF3CHO was first reported by Dodd and
Smith.17 At 313 nm and room temperature, they determined a
quantum yield equal to 0.021 for the primary process yielding
CHF3 and CO, and a quantum yield of 0.12 for the process
yielding CF3 and HCO in the pressure range 40–53 hPa. Pearce
and Whytock18 investigated the importance of the molecular
channel at the same wavelength, and found a quantum yield
within experimental error of zero. Richer et al.19 studied the
1. Introduction
Partially fluorinated alcohols have been suggested as new
replacement compounds for hydrochlorofluorocarbons
(HCFCs) and hydrofluorocarbons (HFCs). A number of inves-
tigations on the kinetics and mechanism for the atmospheric
degradation of CF3CH2OH have been reported.1–5 Trifluoro-
acetaldehyde (CF3CHO) has been shown to be the primary
degradation product in the OH-initiated oxidation of
CF3CH2OH.4,5 CF3CHO is also a key intermediate in the
atmospheric removal of HFC-143a (CF3CH3).6 Similarly,
CF3CH2CHO has been reported as the primary oxidation
product of CF3CH2CH2OH.4 In order to obtain a more
complete picture of the environmental burden of fluorinated
industrial compounds, it is necessary to have information
about the atmospheric lifetimes and global warming potentials
of the degradation products from the oxidation of the
alcohols. This information is deficient for CF3CHO and for
CF3CH2CHO no previous studies on OH kinetics or absorp-
tion cross-sections in the infrared (IR) and ultraviolet-visible
(UV-Vis) regions have been published.
A few experimental and theoretical studies of the reaction
´ ´
between OH and CF3CHO are available in the literature. Dobe
photolysis of CF3CHO in air at lmax ¼ 253.7 nm and lmax
¼
366 nm. The product yields at 253.7 nm of CHF3 (14%),
CF2O (80%), CO (65%) and CO2 (45%) suggest the mole-
cular channel is significant at this wavelength. However,
CF3CHO was reported to undergo only a small degree of
dissociation at lmax ¼ 366 nm after more than six hours
photolysis.
et al.7 measured a reaction rate coefficient of 6.7(4.0) ꢀ 10ꢁ13
cm3 moleculeꢁ1 sꢁ1 at 299 K by using the discharge flow-
´ ´
resonance fluorescence technique. Dobe et al. reported an
activation energy of 5.9 kJ molꢁ1 for this reaction. However,
this was based on an estimated pre-exponential factor of
7 ꢀ 10ꢁ12 cm3 moleculeꢁ1
s
ꢁ1. Scollard et al.8 reported two
values for the OH reaction rate coefficient of CF3CHO:
In order to provide better estimates of the atmospheric life-
times of CF3CHO and CF3CH2CHO, we have measured their
OH reaction rate coefficients using the relative rate method.
We have also determined their effective quantum yield of
photolysis under pseudo-natural conditions at the European
simulation chamber, Valencia, Spain (EUPHORE). Finally,
measurements of infrared and ultraviolet-visible absorption
cross-sections are presented. Based on these data, an assess-
ment of the environmental impact including the radiative
forcing and global warming potentials of CF3CHO and
CF3CH2CHO is made.
6.5(5) ꢀ 10ꢁ13 and 5.5(1.2) ꢀ 10ꢁ13 cm3 moleculeꢁ1
s
ꢁ1, the
former determined from laser photolysis-resonance fluore-
scence experiments and the latter measured relative to acetone.
y Electronic supplementary information (ESI) available: absorption
cross-sections of CF3CHO (Table S1), CF3CH2CHO (Table S2), and
FACSIMILE kinetic model of the CF3CHO/CH3CH3/O3/H2
reaction system (Table S3). See http://www.rsc.org/suppdata/cp/
b3/b315941h/
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 2 4 3 – 1 2 5 2
1243

View Article Online
the OH reaction with CF3CH2CHO, hydroxyl radicals were
2. Experimental
generated by photolysis of 2-propylnitrite-1,1,1,3,3,3-d6 (10–
15 ppm) employing two Philips TLD 18W/08 fluorescence
lamps (lmax ꢂ 375 nm) mounted in a quartz tube inside the
reaction chamber. The lamps were turned off during recording
of the spectra. Photolysis was carried out in intervals of 1–5
min. The mechanism for OH production from photolysis of
2-propylnitrite-1,1,1,3,3,3-d6 is as follows:
2.1. Measurements of IR and UV-Vis absorption
cross-sections
The absorption cross-section of a compound J at a specific
~
wavenumber n is according to Beer–Lambert’s law given by
s(n~) ¼ Ae(n~)/nJl, where Ae(n~) ¼ ꢁ ln t(n~) is the naperian
absorbance, t is the transmittance, nJ is the number density
of J and l is the path length where the absorption takes place.
The integrated absorption intensity, Sint , is given by:
Z
CD3CHðONOÞCD3 þ hn ! CD3CHðOÞCD3 þ NO ð1Þ
CD3CHðOÞCD3 þ O2 ! CD3CðOÞCD3 þ HO2
HO2 þ NO ! OH þ NO2
ð2Þ
ð3Þ
Sint
¼
sðn~Þdn~
ðIÞ
band
Absolute integrated absorption intensities of CF3CHO and
CF3CH2CHO were measured at 298(2) K in the region
4000–400 cmꢁ1. Three independent experiments were per-
formed. Fourier-transform infrared (FTIR) spectra of the pure
vapours were recorded using a Bruker IFS 113v spectrometer
employing a nominal resolution of 1.0 cmꢁ1 and Blackman-
Harris 3-Term apodization of the interferograms. A Ge/KBr
beamsplitter was used to cover the spectral region. To ensure
optical linearity, a deuterated triglycine sulfate (DTGS) detec-
tor was used. Eight single channel spectra each recorded with
32 scans were averaged to yield one background or sample
spectrum. A gas cell of 2.34(2) cm length equipped with KBr
windows was employed. The partial pressures of the gases were
in the range between 1 and 10 hPa, and were measured using
an absolute pressure transducer (MKS Baratron Type 122A)
with a stated accuracy of ꢃ0.15%.
2-Propylnitrite-1,1,1,3,3,3-d6 could not be used as a pre-
cursor for OH radicals in the reaction between OH and
CF3CHO because this reaction was too slow. Instead, OH
radicals were generated by photolysis of O3 in the presence
of H2 (reactions (4) and (5)). Ozone was produced by discharge
of oxygen, where approximately two percent of the oxygen gas
was converted to ozone. Typical mixing ratios of ozone and
hydrogen were 3 ꢀ 102 ppm and 5 ꢀ 103 ppm, respectively.
1
O3 þ hnðlmax ꢂ 310 nmÞ ! Oð DÞ þ O2
ð4Þ
ð5Þ
1
Oð DÞ þ H2 ! OH þ H
Photolysis of ozone was carried out in intervals of 1–2 min
using two Philips TL 20W/12 fluorescence lamps (lmax ꢂ 310
nm). Both CF3CHO and CF3CH2CHO were stable in the dark
in the reaction chamber. Some surface adsorption of CF3CHO
(ca. 3%) was observed for about 15 min after admission to the
cell. The aldehydes did not photolyse to any observable degree
during the time scale of the kinetic experiments.
Absorption cross-sections in the UV-Vis region were
measured at 298(2) K using a Agilent 8453E photodiode array
spectrophotometer having a spectral resolution of 2 nm. The
spectra were recorded in the wavelength range from 190 nm
to 1100 nm at sampling intervals of 1 nm. The integration time
was set to 0.5 s. The pressures of the pure vapours were in the
range 3 to 95 hPa, and were measured using a MKS Baratron
Type 122A pressure transducer. A gas cell of 8.0(1) cm length
with quartz windows was used.
The ‘‘Dublin experiments’’ were performed in a FEP Teflon
reaction chamber with a volume of approximately 50 L. All
experiments were performed in purified air at atmospheric
pressure (973–1013 hPa) and at 298(2) K. Electric fans posi-
tioned below the chamber ensured that a uniform temperature
was maintained during irradiation of the reaction mixtures.
Measured amounts of substrate and reference compounds
were flushed from calibrated bulbs into the partly inflated reac-
tion chamber by a stream of zero-grade air. All pressure read-
ings were made using MKS Baratron pressure transducers
(Model 122 A). When all the reactants had been added to
the chamber it was subsequently filled to maximum capacity
at ca. 1 atm pressure and kept in the dark for 1 h to allow
complete mixing of the reactants. A homogeneous reaction
mixture was confirmed by constant, reproducible gas chroma-
tographic analysis. Hydroxyl radicals were generated by
photolysis of ozone in the presence of water vapour using four
germicidal lamps (Philips TUV, 15 W):
2.2. Relative rate measurements
In the relative rate method (RR), the reaction rate coefficient
for the compound of interest is measured relative to a reference
compound with a known rate coefficient. If the reactants react
solely with the same radical and the reactants are not regener-
ated in the system, the relative rate coefficient, krel , is given
according to the following expression:20
ꢀ
ꢁ
ꢀ
ꢁ
½Aꢄ
½Rꢄ
kA
kR
½Aꢄ0t
½Rꢄ0t
ln
¼ krel ln
; krel
¼
;
ðIIÞ
where A is the compound of interest and R is the reference
compound. [A]0 , [R]0 , [A]t and [R]t are the concentrations of
A and R at the start and at the time t, respectively, and kA
and kR are the rate coefficients.
1
O3 þ hnðlmax ꢂ 254 nmÞ ! Oð DÞ þ O2
ð6Þ
ð7Þ
1
Oð DÞ þ H2O ! OH þ OH
The ‘‘Oslo experiments’’ were carried out in purified air at
298(2) K and 1013(10) hPa in a 250 L reaction chamber of elec-
tro-polished stainless steel with online FTIR detection. The
reaction chamber was equipped with White multi-reflection
optics and had an optical path length of 120 m. The spectro-
meter was a Bruker IFS 88 instrument and a mercury–cad-
mium–telluride (MCT) detector was used. In all experiments,
spectra were recorded in the wavenumber range 4500–400
cmꢁ1. Each spectrum was recorded by adding 100 scans and
employing a resolution of 0.5 cmꢁ1 and Happ–Genzel apodi-
zation. Typically, it took ca. two minutes to record one
spectrum.
Ozone was produced by passing zero-grade air through an
ozone generator (Monitor Labs) directly into the reaction
chamber, at a flow rate of 1 L minꢁ1 for 10 min. Triply distilled
water was injected directly into the reaction chamber. Mixtures
of substrate, reference and O3/H2O were photolysed until
about 50% depletion of the substrate or reference compound
had occurred. Typical initial concentrations employed were
[substrate]0 ¼ [reference]0 ¼ 15–80 ppm, [O3] ꢂ 50 ppm and
[H2O] ꢂ 2 ꢀ 103 ppm. Quantitative analyses were carried out
using gas chromatography (Shimadzu 8A, incorporating a
flame ionisation detector). A Valco gas-sampling valve was
used to remove samples of the reaction mixtures from the
reaction chamber for GC analysis. Photolysis of CF3CH2CHO
The initial mixing ratios of the reactants were 2–4 ppm.
Hydroxyl radicals were produced in two different ways. In
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
1244
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 2 4 3 – 1 2 5 2

View Article Online
The integration was carried out over the wavelength range
290–400 nm. The actinic flux was calculated from the photo-
lysis rate coefficient of NO2 , J(NO2), and the photolytic
production rate coefficient of O(1D), J(O(1D)). The specifi-
cations of the J(NO2) and J(O(1D)) radiometers are given by
Becker.21
was negligible over the time scale of the kinetic experiments
and the aldehyde was stable in the chamber in the dark.
2.3. Experiments performed at EUPHORE
A detailed description of the EUPHORE facility and the exist-
ing analytical instruments is given by Becker.21 Here, a brief
description of the installation is given based on that report.
The present photolysis experiments were carried out in a hemi-
spherical outdoor simulation chamber of volume about 200 m3
made of FEP foil with a thickness of 0.127 mm. The FEP foil
has a transmission of more than 80% of the solar radiation in
the wavelength range between 280 and 640 nm. The chamber
was equipped with a Nicolet Magna 550 FTIR spectrometer
coupled with a White multi-reflection mirror system for in situ
analysis. The optical path length was 553.5 m. FTIR spectra
were recorded every ten minutes by adding 900 interferograms
CF3 radicals are formed in the degradation of CF3CHO17,18
and the subsequent reaction with O2 will lead to the generation
of CF3O radicals. It is possible that CF3O could react rapidly
with CF3CHO and hence, ca. 125 ppm of NO was added to
the reaction chamber in order to provide a sink for CF3O
radicals:23
CF3O þ NO ! CF2O þ FNO
ð14Þ
2.4. Chemicals
with a resolution of 1.0 cmꢁ1
.
CF3CHO was synthesised by adding 1-ethoxy-2,2,2-trifluoro-
ethanol (Aldrich, 90%) to concentrated sulfuric acid (95%).
After ca. one hour mixing, CF3CHO was distilled off under
vacuum and trapped in a container at liquid nitrogen tempera-
ture. The purity was estimated to be better than 98%.
CF3CH2CHO (Fluorochem, > 97%), CH3CH2OH (Aldrich,
96%) and HC(O)OC2H5 (Aldrich, > 97%) were used without
further purification. Purified air containing 80% N2 and 20%
O2 (CO þ NOx < 100 ppb and CnHm < 1 ppm), oxygen gas
(99.95%), hydrogen gas (99%) and ethane (99.0%), used in
the relative rate experiments performed in Oslo, were delivered
from AGA. In the ‘‘Dublin experiments’’, synthetic air was
zero-grade from Air Products. The preparation of 2-propyl-
nitrite-1,1,1,3,3,3-d6 from 2-propanol-1,1,1,3,3,3-d6 followed
the procedure reported for n-butyl nitrite.24 2-Propanol-
1,1,1,3,3,3-d6 was prepared from acetone-d6 (Cambridge
Isotope Laboratories, Inc., 99.9%) by reduction with NaBH4
in a basic water solution. All organic compounds except
ethane, were purified/degassed by three freeze–pump–thaw
cycles.
The photolysis of CF3CHO and CF3CH2CHO were studied
in purified air (see Becker21 for a description of the drying and
purification system). The mixing ratios of CF3CHO and
CF3CH2CHO in the chamber were ca. 1 ppm. Di-n-butyl ether
(DNBE), ꢂ0.2 ppm, was added to the reaction chamber as a
tracer to monitor the OH radical activity, k(OH þ
DNBE) ¼ 2.89 ꢀ 10ꢁ11 cm3 moleculeꢁ1 sꢁ1 22
.
The aldehydes
may be removed from the chamber by photolysis (8), reaction
with OH radicals (9) and leakage (10). The loss of DNBE is
solely due to its reaction with OH radicals (11) and to leakage
(12). Approximately 20 ppb of SF6 was added to the reaction
chamber to determine the leak rate coefficient (reaction (13)),
kleak
:
lnf½SF6ꢄ0=½SF6ꢄtg ¼ kleak
t
ðIIIÞ
where [SF6]0 and [SF6]t are the initial SF6 concentration and
that after a time t, respectively.
Aldehyde þ hn ! Products
Aldehyde þ OH ! Products
Aldehyde ! Loss by leakage
Di-n-butyl ether þ OH ! Products
Di-n-butyl ether ! Loss by leakage
SF6 ! Loss by leakage
Jobs
kald
ð8Þ
ð9Þ
ð10Þ
3. Results
kleak
3.1. IR and UV-Vis absorption cross-sections
kDNBE ð11Þ
kleak
In the determination of IR absorption cross-sections, single
channel spectra of the empty cell were recorded before and
after each sample spectrum. An average of the two transmit-
tance spectra was used in the succeeding analysis. The inte-
grations over the absorption bands were carried out using a
method that defines the baseline from an average of two points
on one side of the band, and the average of two points on the
other side of the band.
ð12Þ
ð13Þ
kleak
Thus, it can be shown that the observed photolysis rate
coefficient, Jobs , of the aldehydes can be obtained from the
expression:
ꢀ
ꢁ
ꢀ
ꢁ
The integrated absorption intensities of the absorption
bands, or regions of overlapping bands, were determined by
plotting the integrated absorbance against the product of the
number density and the path length. None of the regression
lines had a y-intercept significantly different from zero. We
therefore used a least-squares method that forced the re-
gression line to go through zero in order to determine the
absorption intensities. We have only quantified uncertainties
in pressure measurements, path length and temperature as
systematic errors. These are 0.15%, 0.90% and 0.67%, respec-
tively, where the uncertainty in path length includes both
geometrical and optical errors.
½Aldꢄ
kald
½DNBEꢄ
½DNBEꢄ0t
½Aldꢄ0t
kDNBE
ln
ꢁ
ln
ꢄ
ꢂ
ꢃ
ꢅ
kDNBE ꢁ kald
¼
kleak
þ Jobs
t
ðIVÞ
kDNBE
From the observed photolysis rate, the effective quantum yield
for the photolysis of the aldehyde under study can be calcu-
lated according to the following expression:
Feff ¼ Jobs=Jmax
where the maximum photolysis rate coefficient, Jmax , is
ðVÞ
given by:
The absorption cross-sections (base e) of CF3CHO and
CF3CH2CHO in the 4000–400 cmꢁ1 region are shown in Fig.
1a and b, respectively, and the integrated absorption intensities
are given in Tables 1 and 2, respectively. As can be seen from
Table 1, the estimated uncertainty in the total absorption
intensity of CF3CHO is less than two percent, and includes
error from the least-squares fit and the above-mentioned
Z
Jmax
¼
sðlÞfðlÞFðlÞdl
ðVIÞ
Here s(l) is the absorption cross-section (base e) of the
aldehyde in units of cm2 moleculeꢁ1, f(l) is the quantum yield
(f(l) ¼ 1) and F(l) is the solar actinic flux (photons cmꢁ2 sꢁ1).
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 2 4 3 – 1 2 5 2
1245

View Article Online
Absolute integrated absorption intensities, Sint of
Table
2
,
CF3CH2CHO in the mid-infrared region. See text for a discussion
on the uncertainties
Wavenumber range/cmꢁ1
Sint/10ꢁ17 cm moleculeꢁ1
585–505
705–595
905–800
0.094 ꢃ 0.004
0.220 ꢃ 0.030
0.010 ꢃ 0.008
11.7 ꢃ 0.6
1475–915
1850–1675
2950–2650
1.77 ꢃ 0.06
0.74 ꢃ 0.11
26% lower. The UV spectrum of CF3CH2CHO (Fig. 2b) is
not shifted compared to CH3CH2CHO,25 but again the
cross-sections are approximately 26% lower.
3.2. OH reaction rate coefficients
Losses of reactant and reference compounds were monitored
by FTIR spectroscopy in the ‘‘Oslo experiments’’. Spectral
subtraction was used to determine the concentrations of the
substrate and reference compounds at different time intervals
during the relative rate experiments. The relative rate coeffi-
cients were determined according to eqn. (II) by a weighted
least-squares method that includes uncertainties in the concen-
trations of both reactants obtained by the spectral subtraction
procedure.26 Each relative rate coefficient was determined from
three independent measurements. The reported uncertainties in
this work represent 2s from the statistical analyses and do not
include any systematic errors or uncertainties in the reference
Fig. 1 Infrared absorption cross-sections (base e) of pure vapour at
298(2) K of (a) CF3CHO and (b) CF3CH2CHO.
systematic errors. For CF3CH2CHO the uncertainty is
somewhat larger, most of which we cannot account for.
UV-Vis absorption cross-sections of CF3CHO and
CF3CH2CHO at 298(2) K were determined from three
independent measurements. Their UV-Vis spectra are shown
in Fig. 2a and b, respectively. The absorption bands of both
aldehydes corresponds to the weak p* n transition of the
carbonyl group. At the wavelength of maximum absorption,
the uncertainty in the absorption cross-section is 0.5% for
CF3CHO and 2.8% for CF3CH2CHO (2s; error from the
least-squares fit only). When systematic errors in pressure mea-
surements, temperature, optical path length and instrumental
drift also are taken into account, it is estimated that the
absolute error limits of the integrated cross-sections are in
the order of ꢃ5%. The absorption cross-sections of CF3CHO
and CF3CH2CHO are given as Electronic Supplementary
Information (ESI)y in Tables S1 and S2, respectively.
As can be seen from Fig. 2a, the UV spectrum of CF3CHO
is red-shifted by approximately 10 nm as compared to
CH3CHO.25 Further, the cross-sections are on average ca.
Table 1 Absolute integrated absorption intensities, Sint , of CF3-
CHO in the mid-infrared region. See text for a discussion on the
uncertainties
Wavenumber range/cmꢁ1
Sint/10ꢁ17 cm moleculeꢁ1
600–465
750–670
990–775
0.259 ꢃ 0.006
0.508 ꢃ 0.009
0.617 ꢃ 0.009
10.42 ꢃ 0.12
1.085 ꢃ 0.014
0.614 ꢃ 0.016
Fig. 2 (a) UV-Vis absorption cross-sections (base e) of pure vapour
1450–1050
1950–1600
3000–2610
of CF3CHO at 298(2) K: (——) this work; (---) Meller et al.;16 (ꢅ ꢅ ꢅ)
14
Libuda;15 (– – ) Francisco and Williams. (b) UV-Vis absorption
.
cross-sections of CF3CH2CHO (pure vapour) at 298(2) K.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
1246
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 2 4 3 – 1 2 5 2

View Article Online
rate coefficients. Based on the residuals in the spectral sub-
traction, the uncertainty in the relative concentrations of the
reactants is estimated to be 1%.
The rate coefficient for the reaction OH þ CF3CHO was
measured using CH3CH3 as reference compound. The
wavenumber region 2950–2700 cmꢁ1, i.e., part of the C–H
stretching region, was analysed in order to determine the rela-
tive concentrations of both CF3CHO and CH3CH3 . FTIR
reference spectra of CF3CHO, CH3CH3 , O3 , HCHO,
HCOOH (very small amounts) and a sloping baseline were
included in the spectral subtractions. An example of the
residuals after the spectral subtractions is given in Fig. 3.
Fig. 4 shows the decay of CF3CHO versus CH3CH3 in the
presence of hydroxyl radicals, plotted according to eqn. (II).
From these data a relative rate coefficient of 2.00(13) was
extracted. The latest JPL data evaluation27 has recommended
a rate coefficient of 2.4 ꢀ 10ꢁ13 cm3 moleculeꢁ1 sꢁ1 for the
reaction between OH and CH3CH3 at 298 K. On an absolute
scale, the derived OH reaction rate coefficient of CF3CHO is
Fig. 4 Decay of CF3CHO versus CH3CH3 in the presence of OH
radicals at 298(2) K as measured from three independent experiments.
The uncertainty in each data point is based on an estimated uncer-
tainty of 0.01 in the relative concentrations. The uncertainty of the lin-
ear regression coefficients y ¼ ꢁ0.005(6) þ 1.208(25) represents 1s.
therefore 4.80(31) ꢀ 10ꢁ13 cm3 moleculeꢁ1 sꢁ1
.
The OH-initiated oxidation of CF3CHO results in genera-
tion of CF3O radicals. It is known that CF3O reacts with
ethane with a rate coefficient of 1.23 ꢀ 10ꢁ12 cm3 moleculeꢁ1
s .
ꢁ1 27 The CF3O radicals may therefore react significantly with
this also holds for the CF3O radicals, we have estimated
that k(CF3O þ H2) ¼ 2 ꢀ 10ꢁ14 cm3 moleculeꢁ1
both CF3CHO and the reference compound. We have assumed
the large concentrations of H2 in the system, which was used as
the OH source by reaction with O(1D), would act as a scaven-
ger for CF3O radicals. To support this, we modelled the
CF3CHO/CH3CH3/O3/H2 reaction system using FACSI-
MILE.28 Eighty-three reactions were included in the model
(see Table S3y for details). Rate coefficient data were taken
from the IUPAC,23 NIST29 and JPL27 databases. NOx chem-
istry have been included in the model together with reactions
describing the formation of CH3C(O)O2NO2 (peroxy acetyl
nitrite, PAN) and CF3C(O)O2NO2 . However, since the
concentration of NOx in the reaction chamber is very low,
the model showed that these reactions do not make any signifi-
cant contribution in describing the ongoing chemistry in the
chamber.
s
ꢁ1. Similar
arguments for the reactivity of CF3O towards CF3CHO
versus that towards CH3CH3 justifies the rate coefficient
s
ꢁ1. Based
k(CF3O þ CF3CHO) ¼ 3 ꢀ 10ꢁ12 cm3 moleculeꢁ1
on these estimates and the relative concentrations of H2 ,
C2H6 and CF3CHO employed in the experiments, the model
predicted that the reaction of CF3O with H2 is 10–30 times
more frequent than its reactions with CF3CHO and CH3CH3 .
Variations in the concentrations of H2 had no effect on the
value of the ratio k(OH þ CF3CHO)/k(OH þ CH3CH3) lend-
ing support to the assumption that hydrogen acted as a
scavenger for CF3O radicals.
The OH reaction rate coefficient of CF3CH2CHO was
measured relative to CH3CH2OH. Relative concentrations of
both reactants were determined by analysing parts of the
C–H stretching regions of CF3CH2CHO and CH3CH2OH
(3100–2600 cmꢁ1). FTIR reference spectra of CF3CH2CHO,
CH3CH2OH, CD3CH(ONO)CD3 , CF3CHO, CH3CHO,
HCHO, NO2 and a sloping baseline were included in the
spectral subtractions. In addition, we also included two weak
spectra of yet unidentified compounds in the spectral subtrac-
tions. The first spectrum had a band at 2880 cmꢁ1 that came
up together with the injection of CD3CH(ONO)CD3 into the
reaction chamber, suggesting that the synthesised CD3CH-
(ONO)CD3 contained an impurity. The second spectrum had
a band at 3031 cmꢁ1 and originated from a product as it grew
in during the reaction. The two spectra were obtained in a
separate experiment, that is, not in one of the experiments that
Some of the key reactions used in the model are given in
Table 3. As can be seen, the rate coefficients for the reactions
CF3O þ H2 and CF3O þ CF3CHO have been estimated. The
estimate of k(CF3O þ H2) is based on the rate coefficient for
the reaction CF3O þ CH4 . The reactivity of OH radicals
towards H2 is similar to that towards CH4 .23 Assuming that
Table 3 Some key reactions used in the FACSIMILE kinetic model
of the CF3CHO/CH3CH3/O3/H2 reaction system. The complete set
of reactions are given in Table S3y
k298 K/10ꢁ13 cm3
Reactions
moleculeꢁ1 sꢁ1
OH þ H2 ! H2O þ H
0.067a
2.40b
4.80c
0.2d
OH þ CH3CH3 ! H2O þ CH2CH3
OH þ CF3CHO ! H2O þ CF3CO
CF3O þ H2 ! CF3OH þ H
Fig. 3 FTIR spectra of the reaction mixture CF3CHO/CH3CH3/
H2/O3 : (A) before reaction with OH; (B) after reaction with OH;
(C) residual after spectral subtraction analysis of spectrum A, see text
for a list of reference spectra included in the subtraction; (D) residual
after spectral subtraction analysis of spectrum B; (E) reference
spectrum of CF3CHO; (F) reference spectrum of CH3CH3 . The
spectra C–F are shifted for clarity.
CF3O þ CH3CH3 ! CF3OH þ CH2CH3
CF3O þ CF3CHO ! CF3OH þ CF3CO
12.3a
30.0d
Atkinson et al.23 Sander et al.27 This work. Estimated value
(see text for details).
a
b
c
d
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 2 4 3 – 1 2 5 2
1247

View Article Online
were used to determine the relative OH reaction rate coefficient
of CF3CH2CHO. Fig. 5 shows an example of the course of
reaction from one of the relative rate experiments together
with the residuals after the spectral subtractions.
A plot of the logarithm of the relative concentrations of
CF3CH2CHO versus those of CH3CH2OH is shown in
Fig. 6a. From the slope of the plot, a relative reaction rate
coefficient of 1.21(5) was determined for the reaction between
CF3CH2CHO and OH radicals. The latest JPL data evalua-
tion27 gives the OH reaction rate coefficient for CH3CH2OH
as 3.2 ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1 and hence the derived
absolute OH reaction rate coefficient for CF3CH2CHO is
3.87(16) ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1
.
In the ‘‘Dublin experiments’’, the OH reaction rate coeffi-
cient of CF3CH2CHO was measured relative to that for
reaction with HC(O)OC2H5 . The experimental values of
[A]0/[A]t and [R]0/[R]t determined by GC analysis have esti-
mated errors of 2%. The concentration–time data were plotted
according to eqn. (II) and gave a linear relationship with
near-zero intercept, shown in Fig. 6b. From the slope of the
plot, it was found that k(OH þ CF3CH2CHO)/k(OH þ
HC(O)OC2H5) ¼ 3.51(9). The OH reaction rate coefficient of
HC(O)OC2H5 has been measured by Wallington et al.30
(1.02(14) ꢀ 10ꢁ12 cm3 moleculeꢁ1
s
s
ꢁ1) and Le Calve et al.31
ꢁ1). Using an average of
(8.52(75) ꢀ 10ꢁ13 cm3 moleculeꢁ1
these two values, 0.94 ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1, an absolute
rate coefficient of 3.30(8) ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1 is cal-
culated for CF3CH2CHO. The errors quoted in the rate
coefficients reported in this work are twice the standard devia-
tion arising from the least-squares fit of the relative rate data
and do not include an estimate of the error in the reference
rate coefficient.
3.3. Effective quantum yield of photolysis
Photolysis experiments with CF3CHO and CF3CH2CHO have
been carried out at the EUPHORE simulation chamber in
Valencia, Spain (longitude ¼ ꢁ0.5ꢆ, latitude ¼ 39.5ꢆ) during
the month of June. The solar actinic fluxes during the two
experiments are shown in Fig. 7a and b in terms of the photo-
lysis rate coefficient of NO2 , J(NO2), and the photolytic
production rate coefficient of O(1D), J(O(1D)). The pressure
and the temperature inside the chamber were not constant
during the experiments. The observed losses of SF6 , DNBE,
CF3CHO and CF3CH2CHO were therefore corrected
Fig. 6 (a) Decay of CF3CH2CHO versus CH3CH2OH in the presence
of OH radicals at 298(2) K as measured from three independent experi-
ments. The uncertainty in each data point is based on an estimated
uncertainty of 0.01 in the relative concentrations, while the uncertainty
of the linear regression coefficients y ¼ ꢁ0.0071(23) þ 1.208(25) repre-
sents 1s. Two data points were excluded in the analysis, see text. (b)
Decay of CF3CH2CHO versus HC(O)OC2H5 in the presence of OH
radicals at 298(2) K: krel ¼ 3.51(9) 2s .
according to the changes in pressure and temperature using
the ideal gas law.
From a least-squares analysis of the disappearance of SF6
(Fig. 8a), the leak rate coefficient during the CF3CHO photo-
lysis experiment was found to be 7.51(32) ꢀ 10ꢁ6 sꢁ1 (2s). At
the same time the total removal rate coefficient of DNBE
was determined to be 8.38(30) ꢀ 10ꢁ5 sꢁ1. In Fig. 8b, the decay
of CF3CHO is given as a plot of ln{[CF3CHO]0/[CF3CHO]t}
versus the photolysis time. The total removal rate coefficient
of CF3CHO was found to be 7.74(54) ꢀ 10ꢁ6 sꢁ1 which is
within experimental error of the leak rate. A least-squares ana-
lysis of the disappearance of CF3CHO according to eqn. (IV),
i.e. correcting for loss of CF3CHO caused by leakage and reac-
tion with OH radicals, give an upper limit for Jobs of 8.5 ꢀ 10ꢁ7
s
ꢁ1. The time averaged maximum photolysis rate coefficient,
Jmax (see eqn. (VI)), during the experiment was calculated to
be 5.5 ꢀ 10ꢁ5
s
ꢁ1. We therefore suggest that Feff < 2 ꢀ 10ꢁ2
,
which provides an upper limit for the effective quantum yield
for the photolysis of CF3CHO.
Fig. 5 FTIR spectra of the reaction mixture CF3CH2CHO/
CH3CH2OH/CD3CH(ONO)CD3 : (A) before reaction with OH; (B)
after reaction with OH; (C) residual after spectral subtraction analysis
of spectrum A, see text for a list of reference spectra included in the
subtraction; (D) residual after spectral subtraction analysis of
spectrum B; (E) reference spectrum of CH3CH2OH; (F) reference
spectrum of CF3CH2CHO. The residual spectra are magnified 10
times. The spectra C–F are shifted for clarity.
During the CF3CH2CHO experiment, the leak rate was
found to be 5.18(62) ꢀ 10ꢁ6 sꢁ1 as measured from the dis-
appearance of SF6 (Fig. 9a). The observed total removal rates
of DNBE and CF3CH2CHO were 2.72(30) ꢀ 10ꢁ5 sꢁ1 and
7.52(29) ꢀ 10ꢁ6
s
ꢁ1, respectively (Fig. 9b). After correcting
for loss of CF3CH2CHO caused by leakage and reaction with
s
ꢁ1. During
OH radicals, the upper limit for Jobs is 1.5 ꢀ 10ꢁ6
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
1248
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 2 4 3 – 1 2 5 2

View Article Online
Fig. 8 (a) Leakage as measured from the disappearance of SF6
during the photolysis experiment with CF3CHO in the EUPHORE
Fig. 7 Solar actinic fluxes during the photolysis experiments of
CF3CHO (June 21, 2002) and CF3CH2CHO (June 26, 2002) at
EUPHORE in terms of (a) the photolysis rate coefficient of NO2 ,
J(NO2), and (b) the photolytic production rate coefficient of O(1D),
J(O(1D)). The steep increases in the actinic fluxes are due to opening
of the reaction chamber.
simulation chamber. Leakage rate coefficient: kleak ¼ 7.51(16) ꢀ 10ꢁ6
s
ꢁ1. (b) Decay of CF3CHO during exposure to pseudo-natural sun-
light. Total removal rate coefficient: ktotal ¼ 7.74(27) ꢀ 10ꢁ6
s
ꢁ1. The
quoted uncertainty in the rate coefficients is 1s from the statistical
analyses. The data points have been corrected for changes in tempera-
ture and pressure during the experiment.
the experiment, Jmax was calculated to be 3.4 ꢀ 10ꢁ5 sꢁ1
.
This suggests that
CF3CH2CHO.
F
eff < 4 ꢀ 10ꢁ2 for the photolysis of
CF3CHO and CF3CH2CHO are not affected by any large
systematic errors.
As can be seen from Fig. 2a, the UV-Vis absorption cross-
sections of CF3CHO measured in this work are within the
experimental uncertainties reported by Meller et al.16 and
Francisco and Williams.14 The absorption cross-sections mea-
sured by Libuda15 seem to be too low. The UV-Vis spectra of
CF3CHO and CF3CH2CHO (Fig. 2a and b) show a weak
absorption below 220 nm. This is clearly different from
CH3CHO and CH3CH2CHO.25 The corresponding acids have
bands in this region.16 However, we have not seen any infrared
absorption bands that could be attributed to either CF3COOH
or CF3CH2COOH. Unfortunately, Meller et al.,16 Libuda15
and Francisco and Williams14 did not report absorption
cross-sections in this wavelength region, but in Fig. 2 of the
paper by Francisco and Williams,14 it can be seen that the
absorbance for CF3CHO is non-zero below 220 nm.
4. Discussion
4.1. IR and UV-Vis absorption cross-sections
As aforementioned, the only previous quantitative measure-
ment of the infrared absorption intensities of CF3CHO was
the study by Francisco and Williams.14 Over the wavenumber
range 2890–801 cmꢁ1, Francisco and Williams reported that
the integrated absorption intensity was 2.7(3) ꢀ 10ꢁ17 cm
moleculeꢁ1 for pure vapour and 10.7(6) ꢀ 10ꢁ17 cm moleculeꢁ1
with argon added. This is significantly different from the mea-
surement reported in this work. Over the wavenumber range
3000–775 cmꢁ1, we measured an integrated absorption inten-
sity of 12.74(12) ꢀ 10ꢁ17 cm moleculeꢁ1 for pure vapour. We
offer no explanation to this discrepancy. We can only state that
we checked for impurities and that we have tested our experi-
mental setup against HCFC-22. The absorption intensities of
HCFC-22 have been critically evaluated by Ballard et al.32
and are therefore well known. Our measurements of HCFC-
22 were within 5% of the absorption intensities reported by
Ballard et al. We therefore believe that our measurements of
4.2. OH-initiated oxidation
We have measured a rate coefficient equal to 4.80(30) ꢀ 10ꢁ13
cm3 moleculeꢁ1 sꢁ1 for the reaction between OH and
CF3CHO, relative to CH3CH3 . This value is somewhat lower
7
than the rate coefficients reported by Dobe et al., Scollard
´ ´
et al.8 and Laverdet et al.9 The OH reaction rate coefficient
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 2 4 3 – 1 2 5 2
1249

View Article Online
Two different experimental setups have been used to
measure the relative OH reaction rate coefficient of CF3-
CH2CHO. A rate coefficient of 3.87(16) ꢀ 10ꢁ12 cm3 mole-
culeꢁ1 sꢁ1 was measured using CH3CH2OH as the reference
compound (Oslo). As can be seen in Fig. 6a, there is a slight
curvature on the logarithmic plot at the highest conversion
of reactants. Because of this, two of the data points were
excluded in the least-squares analysis. This curvature could
be due to secondary reactions involving the CF3O radical since
no scavenger was used. Even though CF3O will be produced
only at a late stage in the reaction event,4 its reactivity towards
CF3CH2CHO and CH3CH2OH may be larger than its reacti-
vity towards the other compounds in the chamber. We do
not believe however, that our results are seriously affected by
this. When HC(O)OC2H5 was used as reference compound
(Dublin), a rate coefficient of 3.30(8) ꢀ 10ꢁ12 cm3 moleculeꢁ1
sꢁ1 was found. As can be seen, the difference between these
two rate coefficients lie outside the quoted combined experi-
mental errors. However, the quoted error limits reflect preci-
sion only and do not include errors in the reference rate
constants, which probably add about a further 10% to the
quoted errors. We prefer to quote a value of k(OH þ CF3-
CH2CHO) ¼ 3.6(3) ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1 for the rate
coefficient.
The major reaction pathway for reaction of OH radicals
with aliphatic aldehydes is hydrogen atom abstraction from
the aldehydic group.34 The rate coefficient for reaction of
OH with CF3CHO is around a factor of 30 lower than for
reaction with acetaldehyde.30 This result can be rationalized
in terms of changes in the overall enthalpy of reaction and/
or destabilizing polar effects in the transition states for the
reactions. The reported experimental values for the aldehydic
C–H bond dissociation energies in CH3CHO and CF3CHO
are 355 and 381 kJ molꢁ1, respectively,35 and thus reaction
with OH radicals are both strongly exothermic, D(H–OH) ¼
4.91 kJ molꢁ1 35 Rayez et al.12 have estimated that because
.
of the difference in enthalpy changes for the two reactions,
the rate coefficient for the reaction of OH radicals with
CF3CHO will be lower than the value for the corresponding
reaction with CH3CHO by a factor of less than 2.5 at 298 K.
They attributed the significant reduction in the reactivity of
CF3CHO compared to CH3CHO to destabilization of the
transition state for reaction of OH with CF3CHO by the
electron withdrawing inductive effect of the–CF3 group.
The rate coefficient for the reaction of OH with
CF3CH2CHO is about five times smaller than for reaction with
CH3CH2CHO.34 Since the aldehydic C–H bond strengths in
these two molecules would be expected to be quite similar,35
the reduction in reactivity of the b-substituted fluorinated alde-
hyde must reflect the long range destabilizing inductive effect
of the –CF3 group in the transition state. Kwok and Atkin-
son36 have previously found that rate coefficients estimated
from simple structure–activity relationships (SAR), which con-
sider only next-neighbour atomic groups, result in higher
values than experimentally observed when applied to reactions
of OH with fluorinated compounds. They pointed out that the
long range deactivating effect of fluorinated groups must be
incorporated into the relationships to obtain reasonable
agreement between calculated and experimental rate data.
Fig. 9 (a) Leakage as measured from the disappearance of SF6 dur-
ing the photolysis experiment with CF3CH2CHO in the EUPHORE
simulation chamber. Leakage rate coefficient: kleak ¼ 5.18(31) ꢀ 10ꢁ6
s
ꢁ1. (b) Decay of CF3CH2CHO during exposure of pseudo-natural
sunlight. Total removal rate coefficient: ktotal ¼ 7.52(15) ꢀ 10ꢁ6 sꢁ1
.
The quoted uncertainty in the rate coefficients is 1s from the statistical
analyses. The data points have been corrected for changes in tempera-
ture and pressure during the experiment.
for CH3CH3 is one of the best-known rate coefficients
available in the literature, so it is unlikely that the discrepancy
is due to the choice of CH3CH3 as reference compound.
Scollard et al.8 reported that k(OH þ CF3CHO)/k(OH þ
CH3C(O)CH3) ¼ 2.43(53) and k(Cl þ CF3CHO)/k(Cl þ CH3-
C(O)CH3) ¼ 1.14(4). In the latest JPL data evaluation,27 the
recommended OH reaction rate coefficient for CH3C(O)CH3
is: k(T ) ¼ [1.33 ꢀ 10ꢁ13 þ 3.82 ꢀ 10ꢁ11exp(ꢁ2000 K/T )] cm3
moleculeꢁ1 ꢁ1. Using this expression, we have recalculated
s
the rate coefficient from the relative rate measurements of
Scollard et al.8 to be k(OH þ CF3CHO) ¼ 4.4(9) ꢀ 10ꢁ13 cm3
moleculeꢁ1 sꢁ1 at 298 K. The work by Scollard et al. is now in
quite excellent agreement with the present work. It appears that
the three absolute rate coefficients7–9 determined for the reaction
of OH with CF3CHO are all slightly higher than the relative rate
measurements. This may be due to small amounts of reactive
impurities in the CF3CHO used in the three studies.
It is of interest to compare the OH reaction rate coefficient
for reaction with CF3CHO with the OH rate coefficient for
CF3CF2CHO recently determined by Sulbaek Andersen et al.33
Using an indirect relative rate technique, Sulbaek Andersen
et al.33 measured a rate coefficient of 5.26(80) ꢀ 10ꢁ13 cm3
moleculeꢁ1 sꢁ1 and as may be expected CF3CHO and
CF3CF2CHO have similar reactivity towards the OH radical.
4.3. Photolysis
Under pseudo-natural conditions in the EUPHORE simu-
lation chamber, it has been found that the effective quantum
yield of photolysis for CF3CHO and CF3CH2CHO are
<2 ꢀ 10ꢁ2 and <4 ꢀ 10ꢁ2, respectively. For CF3CHO, this is
in contrast to the quantum yields reported by Dodd and
Smith,17 who observed a quantum yield of 0.12 for the radical
channel yielding CF3 and HCO following photolysis at 313 nm
and 40–53 hPa at room temperature. A possible explanation
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
1250
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 2 4 3 – 1 2 5 2

View Article Online
Table 4 Estimated global warming potentials, GWP(t), of CF3CHO
and CF3CH2CHO for a 20 year time horizon, relative to CFC-11.
The instantaneous cloudy-sky radiative forcings (IF) for a 1 ppbv
increase in atmospheric concentrations was calculated according to
the procedure given by Pinnock et al.47 The data on CFC-11 were
taken from the paper by Pinnock et al. See text for details on the
calculation of the atmospheric lifetimes, t, of CF3CHO and
CF3CH2CHO
for this discrepancy is that the intensity of the sunlight is quite
low at wavelengths below 320 nm. Further, the photodisso-
ciation quantum yield has been shown to be strongly pressure
dependent, decreasing with increasing pressure.18
The effective quantum yields are considerably lower than
those found for CH3CHO and CH3CH2CHO. In the EU
project ‘‘RADICAL’’,37 it was found that the effective quan-
tum yields of CH3CHO and CH3CH2CHO were 0.06 ꢃ 0.1
and 0.25 ꢃ 0.04, respectively, determined during weather
conditions similar to those during this work. Yadav and
Goddard38 and Francisco39 have studied the dissociation
reactions of CH3CHO and CF3CHO, respectively, by ab initio
calculations of quite similar levels. The results indicate that the
barriers towards dissociation do not change much upon fluori-
nation. However, both calculations were carried out using
rather small basis sets without including diffuse functions
which are important when describing loosely bound
electrons.40 Further, the geometries were only optimised at
the Hartree–Fock level of theory.
Wavenumber
region/cmꢁ1
IF/
W mꢁ2
t/
year
Compound
GWP(20)
CF3CHO
CF3CH2CHO
CFC-11
1450–465
1475–505
0.129
0.132
0.26
0.36
0.008
50.0
0.0015
0.0003
1.0000
warming potentials, GWP(t), for CF3CHO and CF3CH2CHO,
relative to CFC-11, can then be calculated from the following
expression:48
An interesting question is what effect fluorine substitution in
the CH3 groups of CH3CHO and CH3CH2CHO has on the
relative importance of internal conversion (S1 ! S0) or inter-
system crossing (S1 ! T1), compared to the efficiency of relaxa-
tion of the S1 state by fluorescence. In laser-induced
fluorescence experiments with CF3CHO, fluorescence has been
observed up to 37 000 cmꢁ1 (270 nm).41 In contrast, the fluor-
escence excitation spectrum of CH3CHO has very low intensity
above 31 700 cmꢁ1 (315 nm).42 According to Robb and
co-workers,43,44 emission from excited states implies that there
is no accessible surface crossing promoting fast radiation-less
decay: the existence or lack of a crossing depends on the elec-
tronic structure of the two states and is not a simple function
of the energy gap. However, the very low effective quantum
yield of CF3CH2CHO compared to CH3CH2CHO is surpri-
sing. We are therefore currently investigating these issues
further.
IFald
tald
Mald
GWPðtÞ ¼
IFCFCꢁ11
t
CFCꢁ11 MCFCꢁ11
ꢂ
ꢃ
ð7Þ
1 ꢁ expðꢁt=tald
Þ
ꢀ
;
1 ꢁ expðꢁt=tCFCꢁ11
Þ
where M is the molecular mass and t is the time horizon over
which the instantaneous forcing is integrated. Instantaneous
forcings and global warming potentials for a 20-year time
horizon for CF3CHO and CF3CH2CHO are collected in
Table 4. The data on CFC-11 were taken from the work of
Pinnock et al.47 Although the instantaneous forcings of
CF3CHO and CF3CH2CHO are relatively large compared to
that of CFC-11, their global warming potentials are negligible
due to the short lifetimes.
Care must be exercised when applying these results. In order
to provide realistic predictions, advanced three-dimensional
chemical tracer modelling and radiative forcing calculations
are needed. However, we justify the calculations because they
provide a reasonable estimate of the atmospheric lifetimes
and global warming potentials of CF3CHO and CF3CH2CHO.
This information is important when assessing the total envir-
onmental burden of possible HCFC/HFC replacement
compounds.
4.4. Atmospheric lifetimes and global warming potentials
The atmospheric lifetimes, t, of CF3CHO and CF3CH2CHO
due to removal by reaction with OH radicals and photolysis
may be estimated from the data obtained in this study.ꢁT1h4e5
ꢁ1
atmospheric lifetime is given by tꢁ1 ¼ tOH þ tphotol
.
Using the determined OH rate coefficients, a global averaged
concentration of OH radicals equal to 9.4 ꢀ 105 radicals
cmꢁ3 (ref. 46) and the photolysis rates measured in this work,
the following lifetimes of CF3CHO and CF3CH2CHO in the
gas-phase are found: tOH(CF3CHO) ꢂ 26 days; tphotol(CF3-
CHO) > 27 days; tOH(CF3CH2CHO) ꢂ 4 days; tphotol(CF3-
CH2CHO) > 15 days (the photolytic lifetimes are lower limits
and the true photolytic lifetimes are likely to be longer). The
results suggest that reaction with OH or photolysis may both
be important sinks for CF3CHO.
Fluorinated aldehydes are fairly soluble in water and
undergo hydrolysis and oxidation forming the corresponding
carboxylic acids. Since the lifetimes of CF3CHO with respect
to loss by reaction with OH or photolysis are relatively long,
uptake in rain water or cloud droplets may be an important
sink for CF3CHO. The tropospheric lifetime of soluble species
is of the order of 20 days and hence CF3CHO could provide a
source of CF3COOH in the atmosphere.
The major fate of CF3CH2CHO in the troposphere would
appear to be reaction with OH although photolysis may also
be of some importance. The available data suggest that the
primary product of both these sinks is CF3CHO4 and hence
the atmospheric degradation of CF3CH2CHO may provide
an additional source of CF3COOH in the environment.
Pinnock et al.47 have provided a simple method for esti-
mating the instantaneous cloudy-sky radiative forcing (IF)
directly from a molecule’s absorption cross-sections. Global
Acknowledgements
This work is part of the project ‘‘Impact of Fluorinated
Alcohols and Ethers on the Environment’’, and has received
support from the Commission of the European Communities
under the Energy, Environment and Sustainable Development
Programme through contract EVK2-CT-1999-00009. We
thank Klaus Wirtz and his team for all their help during the
stay at EUPHORE. SRS acknowledges F. Temps for helpful
discussions concerning the photolysis of CF3CHO.
References
1
2
T. J. Wallington, P. Dagaut and M. J. Kurylo, J. Phys. Chem.,
1988, 92, 5024–5028.
K. Tokuhashi, H. Nagai, A. Takahashi, M. Kaise, S. Kondo,
A. Sekiya, M. Takahashi, Y. Gotoh and A. Suga, J. Phys. Chem.
A, 1999, 103, 2664–2672.
3
4
V. C. Papadimitriou, A. V. Prosmitis, Y. G. Lazarou and P.
Papagiannakopoulos, J. Phys. Chem. A, 2003, 107, 3733–3740.
I. Magneron, V. Bossontrot, A. Mellouki, G. LeBras, T. Kelly,
J. Treacy and H. Sidebottom, Kinetics and mechanisms for the oxi-
dation of fluoroalcohols under atmospheric conditions, manuscript
in preparation.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 2 4 3 – 1 2 5 2
1251

View Article Online
Evaluation Number 14, National Aeronautics and Space Admini-
stration, Jet Propulsion Laboratory, California Institute of
Technology, Pasadena, California, 2003.
˚
5
6
S. R. Sellevag, C. J. Nielsen, O. A. Søvde, G. Myhre, J. K. Sundet,
F. Stordal and I. S. A. Isaksen, Atmos. Environ., 2004, submitted.
T. J. Wallington, W. F. Schneider, O. J. Nielsen, J. Sehested,
D. R. Worsnop, W. J. De Bruyn and J. A. Shorter, in Halon
Replacements: Technology and Science, ed. A. W. Miziolek and
W. Tsang, American Chemical Society, Washington DC, 1995,
pp. 16–30.
28 FACSIMILE Version 3.0.30, AEA Technology plc, 1998.
29 NIST Chemical Kinetics Database, NIST Standard Reference
Database 17-2Q98, National Institute of Standards and Technol-
ogy, Gaithersburg, MD 20 899 USA.
´
S. Dobe, L. A. Khachatryan and T. Berces, Ber. Bunsen-Ges.
Phys. Chem., 1989, 93, 847–852.
D. J. Scollard, J. J. Treacy, H. W. Sidebottom, C. Balestra-
Garcia, G. Laverdet, G. LeBras, H. MacLeod and S. Teton,
J. Phys. Chem., 1993, 97, 4683–4688.
G. Laverdet, G. LeBras, H. MacLeod, G. Poulet, S. Teton, D. J.
Scollard, J. J. Treacy and H. W. Sidebottom, Proc. SPIE-Int. Soc.
Opt. Eng., 1993, 1715, 100–112.
´
7
8
30 T. J. Wallington, P. Dagaut, R. Liu and M. J. Kurylo, Int. J.
Chem. Kinet., 1988, 20, 177–186.
31 S. Le Calve, G. Le Bras and A. Mellouki, J. Phys. Chem. A, 1997,
101, 5489–5493.
32 J. Ballard, R. J. Knight, D. A. Newnham, J. Vander Auwera,
M. Herman, G. Di Lonardo, G. Masciarelli, F. M. Nicolaisen,
J. A. Beukes, L. K. Christensen, R. McPheat, G. Duxbury,
R. Freckleton and K. P. Shine, J. Quant. Spectrosc. Radiat.
Transfer, 2000, 66, 109–128.
33 M. P. Sulbaek Andersen, M. D. Hurley, T. J. Wallington, J. C.
Ball, J. W. Martin, D. A. Ellis, S. A. Mabury and O. J. Nielsen,
Chem. Phys. Lett., 2003, 379, 28–36.
9
10 J. S. Francisco, J. Chem. Soc., Faraday Trans., 1992, 88,
1943–1947.
11 M. T. Rayez, J. C. Rayez, T. Berces and G. Lendvay, J. Phys.
Chem., 1993, 97, 5570–5576.
12 M. T. Rayez, D. J. Scollard, H. W. Sidebottom, C. Balestra-
Garcia, S. Teton and G. Le Bras, Chem. Phys. Lett., 1994, 223,
452–458.
34 R. Atkinson, J. Phys. Chem. Ref. Data, Monogr., 1994, 2, 1–216.
35 D. F. McMillen and D. M. Golden, Annu. Rev. Phys. Chem.,
1982, 33, 493–532.
13 A. K. Chandra, T. Uchimaru and M. Sugie, Phys. Chem. Chem.
Phys., 2001, 3, 3961–3966.
36 E. S. C. Kwok and R. Atkinson, Atmos. Environ., 1995, 29,
1685–1695.
14 J. S. Francisco and I. H. Williams, Mol. Phys., 1992, 76,
1433–1441.
15 H.-G. Libuda, PhD Thesis, Bergischen Universita¨t-Gesamtho-
chschule, Wuppertal, 1992.
37 G. K. Moortgat, Evaluation of Radical Sources in Atmospheric
Chemistry through Chamber and Laboratory Studies:
RADICAL, Final report on EU project ENV4-CT97-0419, Max-
Planck-Institut fur Chemie, Atmospheric Chemistry Division,
¨
Mainz, Germany, 2000.
16 R. Meller, D. Boglu and G. K. Moortgat, STEP-HALOCSIDE/
AFEAS Workshop, Dublin, 23–25 March, 1993, pp. 130–138.
17 R. E. Dodd and J. W. Smith, J. Chem. Soc., 1957, 1465–1473.
18 C. Pearce and D. A. Whytock, J. Chem. Soc., Chem. Commun.,
1971, 1464–1466.
19 H. R. Richer, J. R. Sodeau and I. Barnes, STEP-HALOCSIDE/
AFEAS Workshop, Dublin, 23–25 March, 1993, pp. 182–188.
20 T. J. Wallington and O. J. Nielsen, in General Aspects of the
Chemistry of Radicals, ed. Z. B. Alfassi, John Wiley & Sons,
Chichester, 1999, pp. 19–50.
21 K. H. Becker, The European Photoreactor EUPHORE. Design
and Technical Development of the European Photoreactor and
First Experimental Results, Final Report of the EC-Project
Contract EV5V-CT92-0059, Wuppertal, Germany, 1996.
22 F. Kramp and S. E. Paulson, J. Phys. Chem. A, 1998, 102,
2685–2690.
23 R. Atkinson, D. L. Baulch, R. A. Cox, R. F. Hampson, Jr.,
J. A. Kerr, M. J. Rossi and J. Troe, J. Phys. Chem. Ref. Data,
1997, 26, 521–1013.
38 J. S. Yadav and J. D. Goddard, J. Chem. Phys., 1986, 84,
2682–2690.
39 J. S. Francisco, Chem. Phys., 1992, 163, 27–36.
40 T. Helgaker, P. Jørgensen and J. Olsen, Molecular Electronic-
Structure Theory, John Wiley & Sons Ltd, Chichester, 2000.
41 F. Temps, private communication.
42 G.-H. Leu, C.-L. Huang, S.-H. Lee, Y.-C. Lee and I. C. Chen,
J. Chem. Phys., 1998, 109, 9340–9350.
43 F. Bernardi, M. Olivucci and M. A. Robb, Chem. Soc. Rev., 1996,
25, 321–328.
44 M. J. Bearpark and M. A. Robb, J. Phys. Chem. A, 2000, 104,
1075–1077.
45 R. P. Wayne, Chemistry of Atmospheres, Oxford University Press,
Oxford, UK, 3rd edn., 2000.
46 R. G. Prinn, J. Huang, R. F. Weiss, D. M. Cunnold, P. J. Fraser,
P. G. Simmonds, A. McCulloch, C. Harth, P. Salameh, S.
O’Doherty, R. H. J. Wang, L. Porter and B. R. Miller, Science,
2001, 292, 1882–1887.
24 W. A. Noyes, Org. Synth., 1943, Collect. Vol. 2, 108–109.
25 R. D. Martinez, A. A. Buitrago, N. W. Howell, C. H. Hearn and
J. A. Joens, Atmos. Environ., Part A, 1992, 26, 785–792.
26 D. York, Can. J. Phys., 1966, 44, 1079–1086.
27 S. P. Sander, R. R. Friedl, D. M. Golden, M. J. Kurylo, R. E.
Huie, V. L. Orkin, G. K. Moortgat, A. R. Ravishankara, C. E.
Kolb, M. J. Molina and B. J. Finlayson-Pitts, Chemical Kinetics
and Photochemical Data for Use in Atmospheric Studies.
47 S. Pinnock, M. D. Hurley, K. P. Shine, T. J. Wallington and T. J.
Smyth, J. Geophys. Res. [Atmos.], 1995, 100, 23 227–23 238.
48 Climate Change 1994: Radiative Forcing of Climate Change and
An Evaluation of the IPCC IS92 Emission Scenarios,
ed. J. T. Houghton, L. G. Meira Filho, J. Bruce, H. Lee, B. A.
Callander, E. Haites, N. Harris and K. Maskell, Intergovernmental
Panel on Climate Change, Cambridge University Press, Cambridge,
1995.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
1252
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 2 4 3 – 1 2 5 2