A kinetic and mechanistic study of the gas-phase reactions of OH radicals and Cl atoms with some halogenated acetones and their atmospheric implications

Article abstract of DOI:10.1039/b304298g

Rate coefficients for the reactions of hydroxyl radicals and chlorine atoms with a series of halogenated acetones of the type CX₃COCH ₃ (X = H, Cl, F) have been determined using a photolytic relative-rate technique at T = 298 K and at 760 Torr total pressure. The reactions studied and the rate coefficients obtained are shown in the table. Reaction Reaction number Rate coefficient/cm³ molecule^-1 s^-1 OH + CH₃COCH₃ → products (1) (2.2 ± 0.5) × 10^-13 OH + CH₂ClCOCH₃ → products (2) (4.2 ± 0.8) × 10^-13 OH + CHCl ₂COCH₃ → products (3) (3.8 ± 0.8) × 10^-13 OH + CCl₃COCH₃ → products (4) (1.5 ± 0.3) × 10^-14 OH + CH₂FCOCH₃ → products (5) (2.1 ± 0.4) × 10^-13 OH + CF ₃COCH₃ → products (6) (6.9 ± 1.3) × 10^-15 Cl + CH₃COCH₃ → products (7) (2.2 ± 0.4) × 10^-12 Cl + CH₂ClCOCH₃ → products (8) (2.0 ± 0.2) × 10^-12 Cl + CHCl ₂COCH₃ → products (9) (1.7 ± 0.3) × 10^-13 Cl + CCl₃COCH₃ → products (10) ( 1. 7 ± 0.3) × 10^-14 Cl + CH₂FCOCH₃ → products (11) (8.2 ± 1.6) × 10^-13 Cl + CF ₃COCH₃ → products (12) (8.0 ± 1.6) × 10^-15 The errors quoted reflect an estimate of the absolute uncertainty in the measured rate coefficients of ±20%. For reactions (7)-(12), Fourier transform infrared spectroscopy was used to identify products. Qualitative ultra-violet absorption spectra were also recorded for most of the halogenated species investigated in this study, and have been used together with the kinetic data to derive atmospheric lifetimes for these species.

Full text of DOI:10.1039/b304298g

View Article Online / Journal Homepage / Table of Contents for this issue
A kinetic and mechanistic study of the gas-phase reactions of OH
radicals and Cl atoms with some halogenated acetones and their
atmospheric implications
a
Sinead Carr, Dudley E. Shallcross,c Carlos E. Canosa-Mas,b John C. Wenger,y
ab
´
Howard W. Sidebottom,a Jack J. Treacya and Richard P. Wayne*b
a
Department of Chemistry, University College Dublin, Belfield, Dublin, Ireland
Physical and Theoretical Chemistry Laboratory, University of Oxford, South Parks Road,
Oxford OX1 3QZ. E-mail: wayne@physchem.ox.ac.uk
Biogeochemistry Research Centre, School of Chemistry, University of Bristol, Cantock’s Close,
Bristol BS8 1TS
b
c
Received 22nd April 2003, Accepted 14th July 2003
First published as an Advance Article on the web 19th August 2003
Rate coefficients for the reactions of hydroxyl radicals and chlorine atoms with a series of halogenated
acetones of the type CX3COCH3 (X ¼ H, Cl, F) have been determined using a photolytic
relative-rate technique at T ¼ 298 K and at 760 Torr total pressure. The reactions studied and the rate
coefficients obtained are shown in the table.
Rate coefficient/cm3
Reaction
Reaction number
moleculeꢁ1 sꢁ1
OH + CH3COCH3 ! products
OH + CH2ClCOCH3 ! products
OH + CHCl2COCH3 ! products
OH + CCl3COCH3 ! products
OH + CH2FCOCH3 ! products
OH + CF3COCH3 ! products
Cl + CH3COCH3 ! products
Cl + CH2ClCOCH3 ! products
Cl + CHCl2COCH3 ! products
Cl + CCl3COCH3 ! products
Cl + CH2FCOCH3 ! products
Cl + CF3COCH3 ! products
(1)
(2)
(2.2 ꢀ 0.5) ꢂ 10ꢁ13
(4.2 ꢀ 0.8) ꢂ 10ꢁ13
(3.8 ꢀ 0.8) ꢂ 10ꢁ13
(1.5 ꢀ 0.3) ꢂ 10ꢁ14
(2.1 ꢀ 0.4) ꢂ 10ꢁ13
(6.9 ꢀ 1.3) ꢂ 10ꢁ15
(2.2 ꢀ 0.4) ꢂ 10ꢁ12
(2.0 ꢀ 0.2) ꢂ 10ꢁ12
(1.7 ꢀ 0.3) ꢂ 10ꢁ13
(1.7 ꢀ 0.3) ꢂ 10ꢁ14
(8.2 ꢀ 1.6) ꢂ 10ꢁ13
(8.0 ꢀ 1.6) ꢂ 10ꢁ15
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
The errors quoted reflect an estimate of the absolute uncertainty in the measured rate coefficients of
ꢀ20%. For reactions (7)–(12), Fourier transform infrared spectroscopy was used to identify products.
Qualitative ultra-violet absorption spectra were also recorded for most of the halogenated species
investigated in this study, and have been used together with the kinetic data to derive atmospheric
lifetimes for these species.
long lifetimes and could penetrate into the lower stratosphere
Introduction
in significant quantities, carrying halogens with them.4 In addi-
tion, it is essential that the products of the degradation of these
replacements are themselves not potential carriers of chlorine
or bromine to the stratosphere. In the degradation of non-
halogenated hydrocarbons, ketones such as acetone are
formed, and are known to be long-lived (months).5 The accu-
mulation of acetone in the upper troposphere and lower strato-
sphere has recently been highlighted.6–8 Acetone has been
shown to be a potentially important source of HOx (OH and
HO2) in these regions. Hence, there is a particular need to
investigate the chemistry of some halogenated acetones. In this
study we have used a photolytic relative rate technique9,10 at
ˆ
The role played by CFCs in stratospheric ozone depletion is
well known.1,2 To alleviate the impact, a number of replace-
ment compounds, the hydrofluorocarbons and hydrofluoro-
chlorocarbons, HFCs and HCFCs, have been developed for
short-term use. These replacement compounds are susceptible
to attack by the OH radical and can therefore be broken down
in the troposphere.3 However, some proposed replacements
(e.g. CF2ClCH3 , HCFC 142b) have been shown to have rather
y Now at the Department of Chemistry, University College Cork,
Cork, Ireland.
3874
Phys. Chem. Chem. Phys., 2003, 5, 3874–3883
This journal is # The Owner Societies 2003
DOI: 10.1039/b304298g

View Article Online
T ¼ 298 K and 760 Torr total pressure to study the kinetics of
For the study of reactions (1)–(6), OH radicals were gener-
ated by the photolysis of ozone
the reactions
1
O3 þ hnðl ¼ 254 nmÞ ! Oð DÞ þ O2
ð13Þ
OH þ CH3COCH3 ! products
OH þ CH2ClCOCH3 ! products
OH þ CHCl2COCH3 ! products
OH þ CCl3COCH3 ! products
OH þ CH2FCOCH3 ! products
OH þ CF3COCH3 ! products
ð1Þ
ð2Þ
ð3Þ
ð4Þ
ð5Þ
ð6Þ
followed by the reaction
1
Oð DÞ þ H2O ! OH þ OH
ð14Þ
Photolysis was effected using 20 germicidal lamps (Philips TU,
15 W), which produce l ¼ 254 nm radiation.
The diluent gas used in the study of reactions (1)–(6) was
always oxygen. Measured amounts of reactants, with the
exception of water, were flushed from calibrated pyrex bulbs
into the Teflon reaction vessel by a stream of medical-grade
oxygen. Water (triply distilled, 0.5 ml), was injected in liquid
form directly into the chamber. The reaction vessel was
then filled to capacity with O2 . Mixtures of O3/H2O/
CX3COCH3/reference compound were allowed to equilibrate
for ca. 1 h and were then photolysed for ca. 15 min, or until
about 50% depletion of the substrate or reference compound
had occurred. Typical initial concentrations employed
were [CX3COCH3]0 , [reference compound]0 ¼ 10–43 ppm;
[O3]0 ¼ 100–1000 ppm and [H2O]0 ¼ 2000–10 000 ppm. More
detailed information on the initial concentrations of the
substrate and reference compound is shown in Table 1.
In the investigation of reactions (7)–(12), Cl atoms were gen-
erated by the photolysis of molecular chlorine
and the kinetics and products of the reactions
Cl þ CH3COCH3 ! products
Cl þ CH2ClCOCH3 ! products
Cl þ CHCl2COCH3 ! products
Cl þ CCl3COCH3 ! products
Cl þ CH2FCOCH3 ! products
Cl þ CF3COCH3 ! products
ð7Þ
ð8Þ
ð9Þ
ð10Þ
ð11Þ
ð12Þ
A comparison of the rate coefficients is combined with the
investigations of the nature of the products to deduce the likely
mechanism of these reactions. The lifetimes of these species
and their likely impact on the atmosphere are assessed using
the Cambridge 2-Dimensional model.11
Cl2 þ hnðl > 300 nmÞ ! Cl þ Cl
ð15Þ
using 10 ‘blacklamps’ (Philips TL 20 W/08) and 10
‘sunlamps’ (Philips TL 20 W/09) which provided UV-visible
radiation in the range l ¼ 300–450 nm.
Experimental
The diluent gas used was synthetic air or zero-grade nitro-
gen. Once equilibrated, mixtures of Cl2/CX3COCH3/reference
compound/diluent gas were photolysed for up to 20–30 min-
utes or until approximately half the contents of the reaction
bag had been sampled. A list of typical initial concentrations
of the substrate and reference compound is again shown in
Table 1.
A conventional photolytic relative-rate technique was used for
all measurements of rate coefficients. The experimental appa-
ratus and procedure has been described in detail elsewhere9,10
and will only be summarised briefly here. The apparatus con-
sisted of a FEP Teflon bag (volume ca. 50 litres) surrounded by
a bank of discharge lamps. All experiments were performed at
atmospheric pressure (730–760 Torr) and at T ¼ 298 ꢀ 2 K.
An electric fan positioned below the reaction chamber ensured
that a uniform temperature was maintained during irradiation
of the mixture.
Quantitative analysis was carried out using gas chromato-
graphy (Shimadzu 14A GC instrument, with FID detector)
and FTIR spectroscopy (Mattson, Galaxy 5000). Gastight
syringes (Hamilton) or a Valco gas sampling valve were used
Table 1 Initial concentrations of substrate and reference compounds for reactions (1)–(12)
OH reactions
CX3COCH3
Concentration/ppm
Reference compound
Concentration/ppm
CH3COCH3
CH2ClCOCH3
CHCl2COCH3
CCl3COCH3
CH2FCOCH3
CF3COCH3
28
CH2ClCH2Cl
CH2ClCH2Cl
CH2ClCH2Cl
CH3CN
14–28
21–28
14–21
42–43
14–28
21–28
14–28
21–28
28
21–28
28–43
CH2ClCH2Cl
CH3CN
Cl reactions
CX3COCH3
CH3COCH3
Concentration/ppm
Reference compound
CH3Cl
Concentration/ppm
Cl2/ppm
Diluent gas
5–20
5
5–20
10
10–15
10
10
Air
N2
Air
N2
Air
N2
Air
N2
Air
N2
CH2ClCOCH3
CHCl2COCH3
CCl3COCH3
10
20
CH3(O)CH3
CH2Cl2
5–10
10
10
10
10–15
10–20
10–15
10–30
10–30
30–40
10–15
10–15
10–15
10
10–20
5–20
10–20
10–20
14–30
CHCl3
CH2FCOCH3
CF3COCH3
CH2ClCH2Cl
CHCl3
10
42
Phys. Chem. Chem. Phys., 2003, 5, 3874–3883
3875

View Article Online
to remove samples of the reaction mixtures from the Teflon
bag for GC analysis. Infrared spectra were obtained using a
2 litre evacuable Wilks Cell containing a multipass White
mirror arrangement (3–10 m pathlength) and mounted in the
cavity of the spectrometer. Ozone was prepared using a
conventional ozoniser.
Materials
Nitrogen 99.95% (Air products), synthetic air zero grade (Air
Products), oxygen, medical grade (Air Products), CH3COCH3
> 99% (Aldrich), CH2ClCOCH3 90% (Aldrich), CHCl2-
COCH3 > 98% (Aldrich), CCl3ClCOCH3 97% (Aldrich),
CH2FClCOCH3 98%, Cl2 99% (Aldrich), CF3COCH3 97%
(Aldrich), CH2ClCH2Cl 98% (BDH), CH3CN 98% (BDH),
CHCl3 99% (Merck), CH2Cl2 99% (BDH) were used as
supplied.
Results
Assuming that the losses of the substrate (S) and reference
compounds (R) are due to reaction with X, where X is OH
or Cl, alone and that dilution due to sampling is negligible,
it can be shown that
ꢀ
ꢁ
ꢂ
ꢀ
ꢁꢃ
½Sꢃt¼0
½Sꢃt
kS
½Rꢃt¼0
½Rꢃt
ln
¼
ln
ðIÞ
kR
Fig. 1 Relative-rate plots in the form of eqn. (I) for the reactions
of OH with acetone and halogenated ketones that are not photo-
lysed: (a) CH3COCH3 ; (b) CH2ClCOCH3 ; (c) CHCl2COCH3 ;
(d) CH2FCOCH3 . The reference compound in each case was
CH2ClCH2Cl.
where [S]t ¼ 0 and [S]t are the concentrations of the substrate
compound at an initial time, t ¼ 0, and a time t later, and
[R]t ¼ 0 and [R]t are the concentrations of the reference com-
pound at the same two times. kS and kR are the rate constants
for reactions of X with S and R.
Both CCl3COCH3 and CF3COCH3 are susceptible to
photolysis at 254 nm. so that a modified procedure is needed
in the study of reactions (4) and (6). The photolysis rates
(kphot) of these species were first determined in the absence
of OH and the equation
Notario et al.31 (using the flash-photolysis technique) have
obtained a rate coefficient somewhat higher than that deter-
mined here. Apart from the value of k ¼ (5.53 ꢀ 0.63) ꢂ 10ꢁ13
cm3 moleculeꢁ1 sꢁ1 for the reaction of Cl with CH2Cl-
COCH2Cl reported by Voicu et al.,36 no other kinetic data
for the reactions of Cl atoms with halogenated acetones appear
to have been published previously.
ꢀ
ꢁ
ꢂ
ꢀ
ꢁꢃ
1
½Sꢃt¼0
½Sꢃt
1 kS
½Rꢃt¼0
½Rꢃt
ln
¼ kphot
þ
ln
ðIIÞ
t
t kR
A number of investigations have been made of k1 using both
absolute and relative rate methods. The rate coefficients deter-
was employed. Fig. 1 shows plots of (ln{[S]t ¼ 0/[S]t}) as a func-
tion of (ln{[R]t ¼ 0/[R]t}) for reactions (1), (2), (3) and (5); Fig.
2 shows plots of ((1/t)ln{[S]t ¼ 0/[S]t}) as a function of ((1/
t)ln{[R]t ¼ 0/[R]t}) for reactions (4) and (6); and Fig. 3 shows
plots of (ln{[S]t ¼ 0/[S]t}) as a function of (ln{[R]t ¼ 0/[R]t})
for reactions (7)–(12). Table 2 summarises the measured ratios
kS/kR and lists the absolute rate coefficients for reaction of OH
and Cl with the reference compounds used in this study.
mined by Zetzsch,20 Kerr and Stocker,22 Wallington and
24
Kurylo23 and Le Calve et al. show good agreement with the
´
rate coefficient obtained in this study, with the recent studies
by Vasvari et al.25 and Wollenhaupt et al.,26 Gierczak et al.,27
and Talukdar et al.,28 still in broad agreement but towards
the lower end of the range. However, the rate coefficients
measured by Chiorboli et al.21 and Cox et al.19 employing a
relative-rate technique are approximately a factor of two larger
Products observed
FTIR spectroscopy was used to monitor products for reactions
(7)–(12). Table 3 lists the oxidation products observed on
photolysis of Cl2/substrate mixtures in O2 at T ¼ 298 K and
at 1 atm total pressure. The interpretation of these findings
is provided later.
Discussion
Comparison of kinetic results from different studies
A comparison of the rate coefficients determined in this and
other studies is made in Table 4. The value obtained for k7 is
in good agreement with that obtained in the work of Walling-
ton et al.,29 who also employed a relative-rate technique. More
recently, Olsson et al.30 have obtained a slightly lower rate
coefficient than the one determined in this study, but which
is consistent with ours within the combined error limits;
Fig. 2 Relative-rate plots in the form of equation (II) for the
reactions of OH with halogenated ketones that are photolytically
labile: (a) CCl3COCH3 ; (b) CF3COCH3 . In both cases, the reference
compound was CH3CN.
3876
Phys. Chem. Chem. Phys., 2003, 5, 3874–3883

View Article Online
Fig. 3 Relative-rate plots in the form of eqn. (I) for the reactions of Cl with acetone and halogenated ketones: (a) CH3COCH3 : reference
compound CH3Cl; (b) CH2ClCOCH3 : reference compound CH3COCH3 ; (c) CHCl2COCH3 : reference compound CH2Cl2 ; (d) CH2FCOCH3 :
reference compound CH2ClCH2Cl; (e) (S), CCl3COCH3 and (D), CF3COCH3 : reference compound CHCl3 in both cases.
than that obtained in this study. In both investigations, OH
radicals were generated by photolysis of HONO using radia-
tion at wavelengths longer than 330 nm. Cox et al.19 reported
the rate coefficient as an upper limit, allowing for the possibi-
lity that some photolysis of acetone may have occurred during
their experiments. The rate coefficient for the reaction of OH
with CF3COCH3 (k6) determined in this work is significantly
lower than the value obtained by Wallington et al.,35 who used
a flash-photolysis-resonance-fluorescence technique. In the
present study, OH radicals were generated by photolysis of
ozone at l ¼ 254 nm in the presence of water vapour. The
reference compounds used are not susceptible to photolysis
at these wavelengths. It is thus difficult to envisage how the
rate coefficient obtained in the present relative-rate study could
be erroneously low, as any loss of the substrate via photolysis
would be expected to yield a rate coefficient larger than the
true value. The reaction of OH with CF3COCH3 is quite slow
and it is possible that the presence of reactive impurities in the
sample used by Wallington et al.35 may have led to an overes-
timation of the rate coefficient.
Mechanistic implications of kinetic results
It is apparent from the kinetic data for OH radicals that the
reactivity of methyl groups in acetone is similar to that in
ethane.13 This behaviour may be expected because the C–H
bond dissociation energies in both compounds are approxi-
mately equal.37,38 However, the carbonyl group is strongly
electron withdrawing, so that reaction of the electrophilic
OH radical with acetone may be considerably less facile than
Table 2 Summary of the ratios kS/kR obtained and the rate coefficients obtained from them
CX3COCH3
OH kS/kR
Reference compound
1014kR/cm3 moleculeꢁ1 sꢁ1
Ref.
1014kS/cm3 moleculeꢁ1 sꢁ1
CH3COCH3
CH2ClCOCH3
CHCl2COCH3
CCl3COCH3
CH2FCOCH3
CF3COCH3
0.94 ꢀ 0.015
1.81 ꢀ 0.12
1.66 ꢀ 0.11
0.70 ꢀ 0.07
0.89 ꢀ 0.02
0.32 ꢀ 0.05
CH2ClCH2Cl
CH2ClCH2Cl
CH2ClCH2Cl
CH3CN
23.1 ꢀ 4.2
23.1 ꢀ 4.2
23.1 ꢀ 4.2
2.16 ꢀ 0.5
23.1 ꢀ 4.2
2.16 ꢀ 0.5
12
12
12
13
12
13
22 ꢀ 5
42 ꢀ 8
38 ꢀ 8
1.5 ꢀ 0.3
21 ꢀ 4
CH2ClCH2Cl
CH3CN
0.69 ꢀ 0.13
CX3COCH3
Cl kS/kR
Reference compound
1014kR/cm3 moleculeꢁ1 sꢁ1
1013kS/cm3 moleculeꢁ1 sꢁ1
CH3COCH3
CH2ClCOCH3
CHCl2COCH3
CCl3COCH3
CH2FCOCH3
CF3COCH3
4.42 ꢀ 0.11
0.91 ꢀ 0.05
0.52 ꢀ 0.04
0.18 ꢀ 0.02
0.64 ꢀ 0.01
0.084 ꢀ 0.004
CH3Cl
49.5 ꢀ 25
220 ꢀ 44
32.2 ꢀ 9
9.6 ꢀ 2.3
129 ꢀ 40
9.6 ꢀ 2.3
14
This work
22 ꢀ 4
20 ꢀ 2
CH3COCH3
CH2Cl2
CHCl3
15
16
15
16
1.7 ꢀ 0.3
0.17 ꢀ 0.03
8.2 ꢀ 1.6
CH2ClCH2Cl
CHCl3
0.080 ꢀ 0.016
Phys. Chem. Chem. Phys., 2003, 5, 3874–3883
3877

View Article Online
Table 3 Oxidation products observed upon photolysis of Cl2/sub-
strate mixtures in O2 at 1 atm total pressure and T ¼ 298 K
reaction with ethane. It is interesting to note that the rate coef-
ficient for reaction of the electrophilic Cl atom with acetone is
around four times smaller than that for reaction with ethane,14
suggesting that the reaction of OH and Cl with acetone may
have different mechanistic features.
It has generally been assumed that the reaction of OH radi-
cals with acetone leads to the formation of the CH3COCH2
radical following H-atom abstraction
Frequency of
absorption
Compound
CH3COCH3
band/cmꢁ1
Assignment
Ref.
2143
CO
CO2
17
17
2349, 668
1105, 1775
2885
HCOOH
HCl
This work
17
OH þ CH3COCH3 ! CH3COCH2 þ H2O
ð16Þ
However, Wollenhaupt and Crowley39 and Vasvari et al.25
have reported indirect evidence for production of acetic acid.
These workers suggested that addition of the OH radical to
the carbon atom of the carbonyl group followed by elimina-
tion of a methyl radical from the adduct
1177, 1800
1249
1451
CH3COCHO
CH3COOH
Unidentified
CO
This work
This work
CH2ClCOCH3
CHCl2COCH3
CCl3COCH3
CH2FCOCH3
CF3COCH3
2143
17
17
2349, 668
1782, 738
2885
CO2
HCOCl
HCl
CO
This work
17
OH þ CH3COCH3 ! CH3COðOHÞCH3 ! CH3COOH þ CH3
ð17Þ
2143
17
17
2349, 668
1826, 849
2885
CO2
COCl2
HCl
CO
This work
17
could be an important reaction channel. In contrast, several
recent studies have provided compelling evidence that the
major product is in fact CH3COCH2 , and that formation of
acetic acid is negligible below 350 K.27,28,40,41
2143
17
17
2349, 668
1826, 849
2885
CO2
COCl2
HCl
CO
This work
17
A number of theoretical studies of the reaction of OH with
acetone have also been reported.25,40,42 Three possible reaction
pathways have been considered: adduct formation by the addi-
tion of OH to the carbonyl carbon atom; direct H-atom
abstraction; and indirect H-atom abstraction via a six-
membered hydrogen-bonded OH–acetone complex. The results
show that the barrier for OH addition is sufficiently high that,
in agreement with experimental results, H-atom abstraction
will dominate at low temperatures. The calculations also indi-
cate that the most likely pathway for abstraction involves the
initial formation of a hydrogen-bonded complex involving
the oxygen atom of the carbonyl group and the hydrogen atom
of the attacking OH radical. Intramolecular transfer of a
hydrogen atom from a –CH3 group to OH results in the forma-
tion of H2O and CH3COCH2 . Investigations of the tempera-
ture dependence of the reaction of OH with acetone show
that the reaction exhibits strong non-Arrhenius behaviour,
with the overall rate coefficient increasing above room tem-
perature, but becoming essentially constant below
250 K.24,26,27 This behaviour, together with the large primary
isotope effect observed by Gierczak et al.,27 provides support
for a mechanism involving formation of an H-bonded com-
plex.28,43 On this basis, the transition state for reaction of OH
with acetone is stabilised by hydrogen bonding in a six-
membered structure so that the deactivating inductive effect
of the carbonyl group is offset to some extent. Reaction of
Cl atoms has also been shown to involve H-atom abstrac-
tion.44 However, in this case stabilisation of the transition state
is not possible and, as expected, the rate coefficient for reaction
of Cl with acetone is significantly lower than that for reaction
with ethane.13
2143
17
17
2349, 668
1835, 1065
2885
CO2
HCOF
HCl
CO2
This work
17
2349
2885
17
17
HCl
774, 1942
1200, 1809
1170, 1256, 1294
1117, 984
COF2
CF3COCHOa
CF3O3CF3
unidentified
This work
This worka
18
a
An authentic sample of this compound could not be obtained; the
presence of this compound in the product spectrum is therefore an
assumption based on a comparison with a spectrum of a sample of
CH3COCHO. ‘This work’ denotes that the spectra were verified with
an authentic sample.
Table 4 Comparison of rate coefficients at 298 K
1014kOH/cm3
moleculeꢁ1 sꢁ1
1013kCl/cm3
CX3COCH3
CH3COCH3
moleculeꢁ1 sꢁ1
Technique
Reference
22 ꢀ 5a
<53
RR
RR
This work
19
23 ꢀ 3
FR-RF
RR
RR
20
21
62 ꢀ 9
26 ꢀ 8
22
23
21.6 ꢀ 1.6
18 ꢀ 2.4
17.3 ꢀ 0.5
17.5 ꢀ 0.9
17.7 ꢀ 1.6
RR
FP-RF
DF-RF
PLP
24
25
Inspection of Table 2 shows that complete substitution of
the hydrogens of one of the methyl groups of acetone by chlor-
ine or fluorine atoms results in a decrease in the rate coefficient
for reaction with both OH and Cl. The rate coefficients for the
reactions of Cl atoms with CCl3COCH3 and CF3COCH3
(reactions (10) and (12)) are significantly lower (by factors of
65 and 138) than the rate coefficient for reaction of Cl with
acetone (taking a per methyl rate coefficient of 1.1 ꢂ 10ꢁ12
26
27
PLP
DF-CIMS
RR
RR
28
This work
22 ꢀ 4a
23.7 ꢀ 1.2
16.9 ꢀ 3.2
30.6 ꢀ 3.8
20 ꢀ 3
29
30
RR
FP-RF
RR
RR
31
32
cm3 moleculeꢁ1 ꢁ1). The corresponding reactions with OH,
s
22 ꢀ 4
33
34
29.3 ꢀ 2.0
PLP-RF
RR
reactions (4) and (6), are lower by factors of approximately 7
and 16 with respect to reaction (1). The reduction in reactivity
observed on halogenation of acetone may be due to the
changes in the overall enthalpy and/or to destabilising polar
effects in the transition state. The C–H bond strength37 for
CH3COCH3 is 411.3 ꢀ 7.5 kJ molꢁ1 and thus reactions with
both OH radicals and Cl atoms are appreciably exothermic38
CF3COCH3
0.69 ꢀ 0.13a
1.52 ꢀ 0.14
This work
35
This work
FP-RF
RR
0.080 ꢀ 0.016a
a
Errors are twice the standard deviation and represent precision only. RR,
relative rate; FP-RF, flash photolysis–resonance fluorescence; PLP, pulsed laser
photolysis; DF-CIMS, discharge flow-chemical–ionization mass spectrometry.
3878
Phys. Chem. Chem. Phys., 2003, 5, 3874–3883

View Article Online
(D{H–OH} ¼ 491 kJ molꢁ1, D{H–Cl} ¼ 425 kJ molꢁ1).
Bond dissociation energies have not been determined for the
halogenated acetones. Information on the bond dissociation
energies for halocarbons is somewhat limited,37 but it would
appear that halogenation of the b-carbon does not lead to
significant changes in the C–H bond energies. It is therefore
reasonable to assume that, for CF3COCH3 and CCl3COCH3 ,
halogenation will not sufficiently strengthen the C–H bonds to
account for the observed changes in reactivity. We suggest that
the relatively large decreases in reactivity on substitution are
mainly a consequence of polar effects in the transition states.
The transition states of highly exothermic reactions are nor-
mally assumed to be similar in structure and energy to the
reactants.45 Charge separation in the halogenated acetones
would therefore be expected to persist in the transition state
for the reactions with both OH radicals and Cl atoms. Assum-
ing that these reactions involve overall hydrogen-atom
abstraction, the transition states involve a hydrogen atom
positioned between the attacking OH radical or Cl atom and
the CX3COCH2 radical. Halogen substitution will reduce the
electron density around the abstracted H atom, and thus desta-
bilise the transition states compared with those formed with
CH3COCH3 . Furthermore, these transition states will be
increasingly destabilised as the degree of the charge separation
increases. The reduction in reactivity upon substitution by
the more electronegative fluorine atoms is greater than that
observed in Cl atom substitution, as would be expected.
In order to investigate further the electron-withdrawing
inductive effect of halogen substitution on the reactivity of
halogenated acetones, we examined the possibility of a correla-
tion between the rate coefficients and the electronegativities of
the halogenated methyl group. The simple sum of the Pauling
electronegativities of substituent atoms can be a useful mea-
sure of the withdrawal of valence electrons from a central car-
bon atom, as demonstrated by Thomas.46 In a study by Seetula
and Gutman,47 a linear relationship was found between the
reactivities towards hydrogen iodide of a series of substituted
methyl radicals and the sum of the Pauling electronegativities
of the substituents on the methyl group. Fig. 4 shows our data
displayed in a similar way: the logarithms of the room-tempera-
ture rate coefficients (per CH3 group) for the reaction of the
halogenated species towards OH and Cl are plotted against
the group electronegativities of the halogen-substituted CX3
group. Pauling electronegativities of Cl, F and H are taken
to be48 3.16, 3.98 and 2.20. Examining first the kinetics of
the reactions with OH, Fig. 4 (circles) shows that there is
a good correlation for the reactions with CH3COCH3 ,
CCl3COCH3 and CF3COCH3 . However, the rate-coefficient
data for the mono- and di-halogenated acetones do not fit
on the same line, and appear to belong to a different correla-
tion. Hydroxyl radicals could attack substituted acetones
containing a partially halogenated methyl group in two ways
OH þ CH2XCOCH3 ! CHXCOCH3 þ H2O
ð18Þ
OH þ CH2XCOCH3 ! CH2XCOCH2 þ H2O ðX ¼ F or ClÞ
ð19Þ
It is possible that different correlations for fully and partially
substituted acetones arise because of OH attack at the carbon
atom bonded to the chlorine or fluorine atoms. Substitution of
a hydrogen atom by Cl or F atoms in a methyl group reduces
the bond dissociation energies of the remaining C–H bonds.37
Hence, on the basis of the reaction enthalpies, it might be
expected that hydrogen-atom abstraction will be dominated
by removal from the halogenated methyl group rather than
the unsubstituted methyl group. The rate coefficients (k1)–
(k6) provide some support for this argument, since the rate
coefficients per methyl group for the reactions of OH with
CH2ClCOCH3 , CHCl2COCH3 and CH2FCOCH3 are larger
than that for the reaction of OH with acetone.
We consider now the reactivity of Cl atoms towards the
ketones. Fig. 4 shows that, although there is still an evident
decrease in reactivity with increased group electronegativity,
the case for a linear correlation is less clear than it is for the
reactions of OH. Nevertheless, if a trend of reactivity is drawn
as a curve through the data points for the CH3–, CCl3– and
CF3– compounds, it is clear that the mono- and di-substituted
acetones are substantially more reactive than the compounds
that do not have both hydrogen and halogen attached to the
same carbon atom. The products detected for reaction (9) pro-
vide strong evidence that hydrogen abstraction by Cl occurs
almost exclusively at the CHCl2 group for CHCl2COCH3 , as
explained subsequently in the section discussing the observed
products. Tschuikow-Roux et al.49,50 have carried out studies
of the kinetics and products of the photochlorination of
a series of halogenated ethanes. The rate-coefficient ratios for
abstraction at the halogenated methyl group relative to the
methyl group determined at 298
K
for CH2ClCH3 ,49
49
50
CHCl2CH3 and CH2FCH3 are approximately 6, 12 and
10, respectively. These results provide further support for the
assertion that abstraction is occurring at the halogenated
methyl group for reactions (8), (9) and (11).
Table 5 displays our values for the rate coefficients deter-
mined in the present study alongside those reported for the
reactions of Cl atoms and OH radicals with the corresponding
halogenated ethanes. Similar reactivity trends in the Cl atom
and OH radical data for the ketones and alkanes are expected
on the basis of the structural similarities of the a-halogenated
acetones, CX3COCH3 , with the a-halogenated ethanes,
CX3CH3 , (X ¼ H, F, Cl). Fig. 5 shows logarithmic plots of
rate coefficients for reaction with the halogenated acetone as
a function of those for reaction with the corresponding halo-
genated ethanes for reaction with both OH and Cl. In both
cases, the apparent linearity of the correlations confirms the
suspected similarity in trends for the two reaction series.
The experimental data indicate that the reactions of OH
or Cl with CH2ClCOCH3 , CHCl2COCH3 and CH2FCOCH3
involve abstraction mainly from the halogenated methyl
group. As Table 5 shows, the rate coefficients for reaction of
OH with acetone and these halogenated acetones are quite
similar to those for reaction with the corresponding haloge-
nated alkanes. Assuming that the C–H bond strengths in the
acetones and the alkanes do not differ greatly, the similarities
in the rate coefficients imply that, as for the reaction of OH
with acetone, the deactivating effect of the carbonyl group is
Fig. 4 The dependence of the logarithm of the rate coefficient on the
group electronegativity for reactions of OH and Cl with acetone and
some of its halogenated derivatives: (S), kOH for unsubstituted or tri-
substituted ketones; (X), kOH for partially substituted ketones; (D), kCl
.
Phys. Chem. Chem. Phys., 2003, 5, 3874–3883
3879

View Article Online
Table 5 Rate coefficients for reaction of OH and Cl with halogenated
acetones and ethanes. The rate coefficients for reactions with the
ketones are those determined in this work
of OH with CCl3COCH3 and CF3COCH3 , the rate coefficients
for reaction of Cl with these compounds are larger than with
CCl3CH3 and CF3CH3 , thus providing further evidence for
the role of the carbonyl group in attenuating the reactivity
of the halogenated acetones compared to the corresponding
halogenated alkanes.
1014kOH/cm3
moleculeꢁ1 sꢁ1
1014kOH/cm3
moleculeꢁ1 sꢁ1
CX3COCH3
CX3CH3
Ref.
As expected for reactions having different mechanistic fea-
tures, the slope of Fig. 5 is different for the reactions of OH
and Cl. The slope of the plot for OH is near unity, whereas
that for Cl atoms has a slope of around 0.5.
CH3COCH3
CH2ClCOCH3
CHCl2COCH3
CCl3COCH3
CH2FCOCH3
CF3COCH3
22 ꢀ 5
42 ꢀ 8
CH3CH3
25.5
42.3
25.9
1.0
13
51
51
16
13
13
CH2ClCH3
CHCl2CH3
CCl3CH3
CH2FCH3
CF3CH3
38 ꢀ 8
1.5 ꢀ 0.3
21 ꢀ 4
20
0.69 ꢀ 0.13
0.13
Products of reactions
1013kCl/cm3
1013kCl/cm3
In this section, we attempt to rationalise in terms of acceptable
chemical reactions the products observed by FTIR spectro-
scopy (see Table 3) in the Cl-initiated oxidation of the ketones.
We commence the discussion with reference to the general
cases of the mono- and di-halogenated species, and then turn
to the special cases of CH3COCH3 , CCl3COCH3 and
CF3COCH3 .
Consider first the compound CHCl2COCH3 . If abstraction
occurs at the unsubstituted methyl group, HCOCl would be
expected as a product from the following mechanism
CX3COCH3
moleculeꢁ1 sꢁ1
CX3CH3
moleculeꢁ1 sꢁ1
Ref.
CH3COCH3
CH2ClCOCH3
CHCl2COCH3
CCl3COCH3
CH2FCOCH3
CF3COCH3
22 ꢀ 4
20 ꢀ 2
CH3CH3
569
128
14
49
49
52
50
50
CH2ClCH3
CHCl2CH3
CCl3CH3
CH2FCH3
CF3CH3
1.7 ꢀ 0.3
0.17 ꢀ 0.03
8.2 ꢀ 1.6
0.080 ꢀ 0.016
12.0
0.1
65.1
0.0025
counterbalanced by stabilisation of the reaction transition
state through hydrogen bonding in an OH–halogenated-
acetone complex possessing a six-membered ring. The rate co-
efficients for reaction of OH with CCl3CH3 and CF3CH3 are
much smaller than for reaction with ethane and can be ration-
alised largely in terms of the electron-withdrawing effect of
the –CCl3 and –CF3 groups. In a similar manner, the reduction
in reactivity of CCl3COCH3 and CF3COCH3 compared to
acetone is largely a function of the inductive effect of the halo-
genated methyl group. As expected, the reduction in reactivity
is less than for the halogenated ethanes since the abstracted
H atoms are positioned b to the halogenated methyl group
as opposed to the a position in the case of the halogenated
ethanes.
Whereas the rate coefficients for reaction of OH
with CH3COCH3 , CH2ClCOCH3 , CHCl2COCH3 and CH2F-
COCH3 are similar to those for reaction with the correspond-
ing ethanes, the rate coefficients for the reaction of Cl with the
ketones and the corresponding alkanes are quite different (see
Table 5). In all cases, the rate coefficients for the reaction with
the halogenated acetones are much smaller than for reaction
with the analogous ethanes and can be explained in terms of
the inductive effect of the carbonyl group. As for the reaction
Cl þ CHCl2COCH3 ! CHCl2COCH2 þ HCl
CHCl2COCH2 þ O2 þ M ! CHCl2COCH2O2 þ M ð21Þ
2 CHCl2COCH2O2 ! 2 CHCl2COCH2O þ O2
CHCl2COCH2O ! CHCl2CO þ HCHO
CHCl2CO ! CHCl2 þ CO
ð20aÞ
ð22Þ
ð23Þ
ð24Þ
ð25Þ
ð26Þ
ð27Þ
CHCl2 þ O2 þ M ! CHCl2O2 þ M
2 CHCl2O2 ! 2 CHCl2O þ O2
CHCl2O ! HCOCl þ Cl
On the other hand, if abstraction occurs at the halogenated
methyl group, COCl2 is the expected product
Cl þ CHCl2COCH3 ! CCl2COCH3 þ HCl
CCl2COCH3 þ O2 þ M ! CCl2O2COCH3 þ M
2 CCl2O2COCH3 ! 2 CCl2OCOCH3 þ O2
CCl2OCOCH3 ! COCl2 þ CH3CO
ð20bÞ
ð28Þ
ð29Þ
ð30Þ
COCl2 and not HCOCl is observed in the oxidation of
CHCl2COCH3 , so that abstraction must, indeed, occur at
the halogenated methyl group. The nature of the products aris-
ing from the oxidation of the monohalogenated acetones gives
no information on the site of reaction since abstraction at
either the unsubstituted or substituted methyl group produces
HCOX (X ¼ F or Cl). The reaction sequence initiated in the
first case is
Cl þ CH2XCOCH3 ! CH2XCOCH2 þ HCl
CH2XCOCH3 þ O2 þ M ! CH2XCOCH2O2 þ M
2 CH2XCOCH2O2 ! 2 CH2XCOCH2O þ O2
CH2XCOCH2O ! CH2XCO þ HCHO
CH2XCO ! CH2X þ CO
ð31aÞ
ð32Þ
ð33Þ
ð34Þ
ð35Þ
ð36Þ
ð37Þ
ð38Þ
CH2X þ O2 þ M ! CH2XO2 þ M
2 CH2XO2 ! 2 CH2XO þ O2
CH2XO þ O2 ! HCOX þ X
and in the second
Fig. 5 Reactivity of OH and Cl towards halogenated acetones as a
function of their reactivity towards halogenated alkanes: (X), OH;
(D), Cl.
Cl þ CH2XCOCH3 ! CHXCOCH3 þ HCl
CHXCOCH3 þ O2 þ M ! CHXO2COCH3 þ M
ð31bÞ
ð39Þ
3880
Phys. Chem. Chem. Phys., 2003, 5, 3874–3883

View Article Online
(62)) will also produce CO
HCO þ O2 ! HO2 þ CO
2 CHXO2COCH3 ! 2 CHXOCOCH3 þ O2
CHXOCOCH3 ! HCOX þ CH3CO
ð40Þ
ð41Þ
ð63Þ
The main products identified from the oxidation of
CF3COCH3 were CF2O, CF3O3CF3 , HCl and CO2 . The yield
of CF2O was approximately 10% of the loss of CF3COCH3 . In
addition, a major degradation product showed strong absorp-
We now turn to a discussion of the products observed in the
oxidation of CH3COCH3 , CCl3COCH3 and CF3COCH3 . In
addition to CO and CO2 , methyl glyoxal (CH3COCHO),
formic acid (HCOOH) and acetic acid (CH3COOH) were
observed as products of the oxidation of acetone. Formation
of methyl glyoxal is expected in two reactions of oxy- and
peroxy-radicals that will be intermediates in the oxidation
tion bands in the 1100–1300 cmꢁ1 region and at 1813 cmꢁ1
.
This product was tentatively identified as trifluoromethyl-
glyoxal on the basis of a comparison with the spectrum of
CH3COCHO. A mechanism which is consistent with the
observed products is
Cl þ CH3COCH3 ! CH3COCH2 þ HCl
CH3COCH2 þ O2 þ M ! CH3COCH2O2 þ M
2 CH3COCH2O2 ! 2 CH3COCH2O þ O2
ð42Þ
ð43Þ
ð44Þ
Cl þ CF3COCH3 ! CF3COCH2 þ HCl
CF3COCH2 þ O2 þ M ! CF3COCH2O2 þ M
2 CF3COCH2O2 ! 2 CF3COCH2O þ O2
ð64Þ
ð65Þ
ð66Þ
2 CH3COCH2O2 ! CH3COCHO þ CH3COCH2OH þ O2
ð45Þ
ð46Þ
2 CF3COCH2O2 ! CF3COCHO þ CF3COCH2OH þ O2
CH3COCH2O þ O2 ! CH3COCHO þ HO2
ð67Þ
CF3COCH2O þ O2 ! CF3COCHO þ HO2
CF3COCH2O ! CF3CO þ HCHO
CF3CO þ O2 þ M ! CF3COO2 þ M
ð68Þ
ð69Þ
ð70Þ
Generation of formic acid requires a secondary oxidation
process involving the hydroxyacetone (CH3COCH2OH) pro-
duced in reaction (45)
Cl þ CH3COCH2OH ! CH3COCHOH þ HCl
ð47Þ
The dominant reaction for the alkoxy radical, CF3COCH2O
will probably be C–C bond cleavage to produce CF3CO and
HCHO. This process is certainly dominant in the case of
CH3COCH2O.32 Hence, it is more likely that the trifluoro-
methylglyoxal arises from reaction (67), rather than from reac-
tion (68). Zabel et al.54 have shown that CF3CO adds to O2 in
preference to decomposition (producing CO). We observe little
or no CO, in accord with this conclusion. CF3O3CF3 and
COF2 will be formed in a sequence of reactions starting from
the CF3COO2 radical
CH3COCHOH þ O2 þ M ! CH3COCHðO2ÞOH þ M ð48Þ
CH3COCHðO2ÞOH þ RO2 ! CH3COCHðOÞOH þ RO þ O2
ð49Þ
CH3COCHðOÞOH þ O2 ! CH3COO2 þ HCOOH
ð50Þ
RO2 here represents generically any of the peroxy radicals
generated in the system. A possible source of acetic acid is
the cross reaction between the two peroxy radicals53
CH3COO2 þ CH3COCH2O2 ! CH3COOH
þ CH3COCHO þ O2 ð51Þ
CF3COO2 þ RO2 ! CF3COO þ RO þ O2
CF3COO ! CF3 þ CO2
ð71Þ
ð72Þ
ð73Þ
ð74Þ
ð75Þ
ð76Þ
CF3 þ O2 þ M ! CF3O2 þ M
CF3O2 þ RO2 ! CF3O þ RO þ O2
CF3O ! COF2
In the oxidation of CCl3COCH3 , the major carbonyl species
observed was COCl2 (phosgene) with a yield close to 100%
based on the loss of CCl3COCH3 . Formation of phosgene is
consistent with a mechanism in which the radical CCl3O is
the ultimate source
CF3O2 þ CF3O þ M ! CF3O3CF3 þ M
Cl þ CCl3COCH3 ! CCl3COCH2 þ HCl
CCl3COCH2 þ O2 þ M ! CCl3COCH2O2 þ M
2 CCl3COCH2O2 ! 2 CCl3COCH2O þ O2
ð52Þ
ð53Þ
ð54Þ
Reaction (75) is written as shown specifically to indicate the
formation of COF2 , which is always found in systems contain-
ing CF3O. In the present instance, the process might involve
interaction with Cl atoms, water clusters, or a heterogeneous
reaction. Several of the reactions in the schemes that we have
proposed involve the interaction of two peroxy radicals, and
we now compare some of these processes. Lightfoot et al.53
have discussed in some detail possible mechanisms for the
self-reaction of peroxy radicals, and have shown that there
are two main channels, one forming two alkoxy radicals (e.g.
reaction (66)), and one forming a carbonyl and an alcohol
(e.g. reaction (67)), which may proceed via a complex inter-
mediate. The products that we have observed in the present
study are consistent with the hypothesis that the self reaction
for CCl3COCH2O2 proceeds mainly via the alkoxy channel fol-
lowed by decomposition via reaction (57) of the alkoxy radical
(since CCl3COCHO is not observed), whereas the self reaction
of CF3COCH2O2 proceeds at least in part via the carbonyl/
alcohol channel (since CF3COCHO is observed). We have con-
sidered the possibility that photolysis of CCl3COCHO may
prevent its detection, but, at least to a first approximation,
the photolysis rates for CCl3COCHO and CF3COCHO can
be assumed to be comparable. It is interesting that in the case
of the self-reaction of CH2ClO2 the alkoxy channel accounts
for nearly 100% of the reaction, whereas for CH2FO2 it only
accounts for roughly 70%.53 Perhaps this difference indicates
2 CCl3COCH2O2 ! CCl3COCHO þ CCl3COCH2OH þ O2
ð55Þ
CCl3COCH2O þ O2 ! CCl3COCHO þ HO2
CCl3COCH2O ! CCl3CO þ HCHO
CCl3CO ! CCl3 þ CO
ð56Þ
ð57Þ
ð58Þ
ð59Þ
ð60Þ
ð61Þ
CCl3 þ O2 þ M ! CCl3O2 þ M
CCl3O2 þ RO2 ! CCl3O þ RO þ O2
CCl3O ! COCl2 þ Cl
The absence of CCl3COCHO suggests either that reactions
(55) and (56) are unimportant or that the carbonyl is rapidly
photo-oxidised
CCl3COCHO þ hn ! CCl3CO þ HCO
ð62Þ
The presence of CO could be evidence that photo-oxidation of
CCl3COCHO is occurring via reaction (58), a process in which
decomposition of CCl3CO has been observed to dominate over
addition of O2 .54 Oxidation of HCO (derived from reaction
Phys. Chem. Chem. Phys., 2003, 5, 3874–3883
3881

View Article Online
that the RO4R intermediate discussed by Lightfoot et al.53 is
particularly stable when R contains fluorine atoms.55
Atmospheric lifetimes
The most likely fate of the halogenated acetones in the Earth’s
atmosphere is either reaction with OH or photolysis. We have
used the Cambridge 2-dimensional model11 to determine life-
times for the species studied with respect to loss via reaction
with OH. Fig. 6 shows the globally averaged loss rate constant
(1/lifetime) for these species in reaction with OH. It can be
seen readily that in the troposphere the lifetimes of these spe-
cies with respect to loss via reaction with OH are quite large
(many months). The absorption spectrum of only acetone is
known quantitatively. To provide an estimate of the photolysis
rates, ultra-violet absorption spectra for most of the haloge-
nated acetones were recorded at low resolution (ca. 1 nm)
using 10 torr of the compound in a cell of pathlength 10 cm;
Fig. 7 shows the spectra. It is possible to estimate the absolute
absorption cross-sections and hence the photolysis rates for the
halogenated acetones by simple scaling to the cross section13
for acetone over the wavelength range of interest (300–400
nm). This analysis suggests that the photolysis rates for all
the halogenated acetones are larger than that for acetone. If
it is assumed that the quantum yields for decomposition are
the same as those for acetone as published,56 then rates of
photolysis of CF3COCH3 and CCl3COCH3 are about two
and three times that for acetone, respectively. The photolysis
rates for the other halogenated acetones lie somewhere
between these extremes. The calculations show that, for acetone
itself, photolysis is at least five times more rapid than attack by
OH (at an altitude of ca. 30 km), and up to 100 times faster at
other altitudes. It follows, then, that photolysis dominates over
chemical reaction with OH as the primary sink of all the halo-
genated acetones. Indeed, if the quantum yield for dissociation
were to approach unity for the halogenated species, then the
photolysis rates become very much greater than that for acet-
one (40 and 80 times for CF3COCH3 and CCl3COCH3), and
the lifetimes against photolysis correspondingly shorter. There
is some support for a relatively high quantum yield for loss of
halogenated acetones, as described57 by Lee and Lewis: Bowles
et al.,58,59 for example, find quantum yields for CO formation
to be 0.74–0.92 for the photolysis of two fully halogenated
acetones at l ¼ 313 nm and 254 nm. Recent work by
Burkholder et al.60 has shown that the quantum yield for
Fig. 7 UV spectra of acetone and four of its halogenated derivatives.
Optical path ¼ 10 cm; pressure of compound ¼ 10 Torr:11 (a)
CF3COCH3 ; (b) CCl3COCH3 ; (c) CH3COCH3 ; (d) CHCl2COCH3 ;
(e) CH2ClCOCH3 .
photodissociation of CH2ClCOCH3 is 0.5 at both l ¼ 308
nm and at l ¼ 351 nm (while the quantum yield for
CH2BrCOCH3 loss is unity at the longer wavelength and
apparently greater than unity at the shorter one). It seems rea-
sonable to suppose, therefore, that the quantum yields for the
halogenated acetones studied in this work are also most likely
near to unity. Unpublished results from Pilling and cowor-
kers61 suggest an atmospheric rate of photolysis of acetone
itself that is roughly two times smaller than has been calculated
previously. Even so, it is clear that photolysis is by far the most
important tropospheric loss process. In summary, then,
it would appear that the halogenated acetones studied in this
work will have tropospheric lifetimes of less than a month,
and photolysis will be the dominant loss process for the
halogenated species, particularly for CF3COCH3 .
Acknowledgements
The authors acknowledge with gratitude the partial support of
this research by the European Commission in the project
‘HALOBUD’ (ENV4-CT97-0393). They also wish to thank
Forbairt and the British Council for the award of a grant
under their joint research scheme.
The Biogeochemistry Research Centre at Bristol is a joint
initiative between the School of Chemistry and the Depart-
ments of Geology and Geography.
References
1
2
M. J. Molina and F. S. Rowland, Nature, 1974, 249, 810.
J. C. Farman, B. G. Gardiner and J. D. Shanklin, Nature, 1985,
315, 207.
3
4
5
R. P. Wayne, Chemistry of Atmospheres, Oxford University Press,
Oxford, 3rd edn., 2000.
O. Wild, O. V. Rattigan, R. L. Jones, R. A. Cox and J. A. Pyle,
J. Atmos. Chem., 1996, 25, 167.
H. B. Singh, D. O’Hara, D. Herlth, W. Sachse, D. R. Blake,
J. D. Bradshaw, M. Kanakidou and P. J. Crutzen, J. Geophys.
Res., 1994, 99, 1805.
Fig. 6 Globally averaged loss rate constants and lifetimes for attack
by OH radicals on the ketones studied. These rates were calculated
using the Cambridge 2-dimensional model:11 (a) CF3COCH3 ; (b)
CCl3COCH3 ; (c) CH2FCOCH3 ; (d) CH3COCH3 ; (e) CHCl2COCH3 ;
(f) CH2ClCOCH3 .
6
H. B. Singh, M. Kanakidou, P. J. Crutzen and D. J. Jacob,
Nature, 1995, 378, 50.
F. Arnold, G. Knop and H. Zierens, Nature, 1986, 321, 505.
F. Arnold, V. Burger, B. Drost-Fanke, F. Grimm, J. Schneider,
A. Krieger and T. Stile, Geophys. Res. Lett., 1997, 24, 3017.
7
8
3882
Phys. Chem. Chem. Phys., 2003, 5, 3874–3883

View Article Online
9
S. Carr, J. J. Treacy, H. W. Sidebottom, R. K. Connell, C. E.
Canosa-Mas, R. P. Wayne and J. Franklin, Chem. Phys. Lett.,
1994, 227, 39.
32 J. J. Orlando, G. S. Tyndall, L. Vereecken and J. Peeters, J. Phys.
Chem. A, 2000, 104, 11 578.
33 L. K. Christensen, J. C. Ball and T. J. Wallington, J. Phys. Chem.
A, 2000, 104, 345.
34 J. Albaladejo, A. Notario, C. A. Cuevas, B. Ballesteros and
10 D. J. Scollard, J. J. Treacy, H. W. Sidebottom, C. Balestra-
Garcia, G. Laverdet, G. Le Bras, H. MacLeod and S. Te´ton,
J. Phys. Chem., 1993, 97, 4683.
´
E. Martınez, J. Atmos. Chem., 2003, 45, 35.
11 K.S. Law and J.A. Pyle, J. Geophys. Res., 1993, 98, 18 377.
12 S.-B. Xing, S.-H. Shi and L.-X. Qiu, Int. J. Chem. Kinet., 1992,
24, 1.
13 R. Atkinson, D. L. Baulch, R. A. Cox, R. F. Hampson, J. A.
Kerr, M. J. Rossi and J. Troe, J. Phys. Chem. Ref. Data, 1997,
26, 521.
35 T. J. Wallington, P. Dagaut and M. J. Kurylo, J. Phys. Chem.,
1988, 92, 5024.
36 I. Voicu, I. Barnes, K. H. Becker, T. J. Wallington, Y. Inoue and
M. Kawasaki, J. Phys. Chem. A, 2001, 105, 5123.
37 CRC Handbook of Chemistry and Physics, ed. D. R. Lide, CRC
Press, Boca Raton, FL, 74th edn., 1993.
14 R. Atkinson, D. L. Baulch, R. A. Cox, R. F. Hampson, J. A. Kerr
and J. Troe, J. Phys. Chem. Ref. Data, 1989, 18, 881.
15 E. Tschuikow-Roux, F. Faraji, S. Paddison, J. Niedzielski and
K. Miyokawa, J. Phys. Chem., 1988, 92, 1488.
38 D. F. McMillen and D. M. Golden, Annu. Rev. Phys. Chem.,
1982, 33, 493.
39 M. Wollenhaupt and J. N. Crowley, J. Phys. Chem. A, 2000, 104,
6429.
16 S. P. Sander, M. J. Kurylo, V. L. Orkin, D. M. Golden, R. E.
Huie, B. J. Finlayson-Pitts, C. E. Kolb, M. J. Molina, R. R.
Friedl, A. R. Ravishankara and G. K. Moortgat, Chemical
kinetics and photochemical data for use in atmospheric studies,
JPL Publication No. 02-25, JPL/NASA, Pasadena, CA, 2003.
17 G. Herzberg, Molecular Structure and Molecular Structure, II:
Infrared and Raman Spectra of Polyatomic Molecules, Van
Nostrand Reinhold, New York, 1945; K. P. Huber and
G. Herzberg, Molecular Spectra and Molecular Structure IV.
Constants of Diatomic Molecules, Van Nostrand Reinhold,
New York, 1979.
40 S. Vandenberk, L. Vereecken and J. Peeters, Phys. Chem. Chem.
Phys., 2002, 4, 461.
41 G. S. Tyndall, J. J. Orlando, T. J. Wallington, M. D. Hurley, M.
Goto and M. Kawasaki, Phys. Chem. Chem. Phys., 2000, 4, 2189.
42 S. Aloisio and J. S. Francisco, J. Phys. Chem. A, 2000, 104, 3211.
43 I. W. M. Smith and A. R. Ravishankara, J. Phys. Chem. A, 2002,
106, 4798.
44 O. J. Nielsen, M. S. Johnson, T. J. Wallington, L. K. Christensen
and J. Platz, Int. J. Chem. Kinet., 2002, 34, 283.
45 A. Pross, Adv. Phys. Org. Chem., 1977, 14, 69.
46 T. D. Thomas, J. Am. Chem. Soc., 1970, 92, 4184.
47 J. A. Seetula and D. Gutman, J. Phys. Chem., 1991, 95, 3626.
48 P. W. Atkins, Physical Chemistry, OUP, Oxford, 6th edn., 1998.
49 E. Tschuikow-Roux, J. Niedzielski and F. Faraji, Can. J. Chem.,
1985, 63, 1093.
50 E. Tschuikow-Roux, T. Yano and J. Niedzielski, J. Chem. Phys.,
1985, 82, 65.
51 N. Cohen and K. R. Westberg, J. Phys. Chem. Ref. Data, 1991,
20, 1211.
52 J. Platz, O. J. Nielsen, J. Sehested and T. J. Wallington, J. Phys.
Chem., 1995, 99, 6570.
53 P. D. Lightfoot, R. A. Cox, J. N. Crowley, M. Destriau, G. D.
Hayman, M. E. Jenkin, G. K. Moortgat and F. Zabel, Atmos.
Environ., 1992, 26A, 1805.
18 E. C. Tuazon and R. Atkinson, J. Atmos. Chem., 1993, 17, 179.
19 R. A. Cox, R. G. Derwent and M. R. Williams, Environ. Sci.
Technol., 1980, 14, 57.
20 C. Zetzsch, Rate constants for the reactions of OH with acetone and
methylethylketone in the gas phase, Abstracts, Seventh Inter-
national Symposium on Gas Kinetics, Go¨ttingen, Aug., 1982.
21 C. Chiorboli, C. A. Bignozzi, A. Maldotti, P. F. Giardini, A.
Rossi and V. Carassiti, Int. J. Chem. Kinet., 1983, 15, 579.
22 J. A. Kerr and D. W. Stocker, J. Atmos. Chem., 1986, 4, 253.
23 T. J. Wallington and M. J. Kurylo, J. Phys. Chem., 1987, 91,
5050.
´
24 S. Le Calve, D. Hitier, G. Le Bras and A. Mellouki, J. Phys.
Chem. A, 1998, 102, 4579.
25 G. Vasvari, I. Szilagyi, A. Bencsura, S. Dobe, T. Berces,
E. Henon, S. Canneaux and F. Bohr, Phys. Chem. Chem. Phys.,
2001, 3, 551.
26 M. Wollenhaupt, S. A. Carl, A. Horowitz and J. N. Crowley,
J. Phys. Chem. A, 2000, 104, 2695.
54 F. Zabel, F. Kirchner and K.-H. Becker, Int. J. Chem. Kinet.,
1994, 26, 827.
55 P. Biggs, C. E. Canosa-Mas, J.-M. Fracheboud, C. J. Percival,
R. P. Wayne and D. E. Shallcross, J. Chem. Soc., Faraday Trans.,
1997, 93, 379.
27 T. Gierczak, M. K. Gillies, S. Bauerle and A. R. Ravishankara,
J. Phys. Chem. A, 2003, 107, 5014.
56 T. Gierczak, J. B. Burkholder, S. Bauerle and A. R.
Ravishankara, Chem. Phys., 1998, 231, 229.
28 R. K. Talukdar, T. Gierczak, D. C. McCabe and A. R.
Ravishankara, J. Phys. Chem. A, 2003, 107, 5021.
29 T. J. Wallington, J. M. Andino, J. C. Ball and S. M. Japar,
J. Atmos. Chem., 1990, 10, 301.
30 B. E. R. Olsson, M. Hallquist, E. Ljungstro¨m and J. Davidsson,
Int. J. Chem. Kinet., 1997, 29, 195.
57 E. K. C. Lee and R. S. Lewis, Adv. Photochem., 1980, 12, 1.
58 R. Bowles, J. R. Majer and J. C. Robb, Trans. Faraday Soc., 1962,
58, 1541.
59 R. Bowles, J. R. Majer and J. C. Robb, Trans. Faraday Soc., 1962,
58, 2394.
60 J. B. Burkholder, M. K. Gilles, T. Gierczak and A. R.
Ravishankara, Geophys. Res. Lett., 2002, 29, 3–1.
61 D. E. Heard, personal communication, 2003.
31 A. Notario, A. Mellouki and G. Le Bras, Int. J. Chem. Kinet.,
2000, 32, 62.
Phys. Chem. Chem. Phys., 2003, 5, 3874–3883
3883