Mechanism of the gas phase reaction of chlorine atoms with butanone

Article abstract of DOI:10.1016/j.cplett.2007.03.082

Smog chamber/FTIR techniques were used to study the Cl atom initiated oxidation of CH₃C(O)CH₂CH₃ in 700-760 Torr of N₂ at 296 K. The reaction of Cl atoms with CH₃C(O)CH₂CH₃ proceeds via hydrogen abstraction with 73 ± 9% of reaction occurring at the -CH₂- group. Relative rate techniques were used to measure k(Cl + CH₃C(O)CHClCH₃) = (5.62 ± 0.81) × 10^-12 cm³ molecule^-1 s^-1. It was deduced that the CH₃C(O)CHCH₃ radical reacts with Cl₂ with a rate constant of the order of 10^-14 to 10^-13 cm³ molecule^-1 s^-1.

Full text of DOI:10.1016/j.cplett.2007.03.082

Chemical Physics Letters 439 (2007) 274–279
www.elsevier.com/locate/cplett
Mechanism of the gas phase reaction of chlorine atoms with butanone
a
a
b,
*
E. Iwasaki , F. Taketani a,1, K. Takahashi a,2, Y. Matsumi , T.J. Wallington
,
b
M.D. Hurley
a
Solar-Terrestrial Environment Laboratory, Graduate School of Science, Nagoya University, 3-13 Honohara, Toyokawa 442-8507, Japan
b
Ford Motor Company, P.O. Box 2053, Dearborn, MI 48121-2053, USA
Received 2 December 2006; in final form 16 March 2007
Available online 28 March 2007
Abstract
Smog chamber/FTIR techniques were used to study the Cl atom initiated oxidation of CH3C(O)CH2CH3 in 700–760 Torr of N2 at
296 K. The reaction of Cl atoms with CH3C(O)CH2CH3 proceeds via hydrogen abstraction with 73 9% of reaction occurring at the –
CH2– group. Relative rate techniques were used to measure k(Cl + CH3C(O)CHClCH3) = (5.62 0.81) · 10ꢀ12 cm3 moleculeꢀ1 ꢀ1. It
s
was deduced that the CH3C(O)CHCH3 radical reacts with Cl2 with a rate constant of the order of 10ꢀ14 to 10ꢀ13 cm3 moleculeꢀ1 sꢀ1
.
ꢀ 2007 Elsevier B.V. All rights reserved.
1. Introduction
ceeds via both CH3CO displacement and hydrogen abstrac-
tion channels. The displacement channel gives CH3CO
radicals and CH3C(O)Cl in molar yields of 23 2%. The
hydrogen abstraction channel accounts for the balance of
the reaction and gives CH3C(O)C(O)CH2 radicals [4].
There are no available data concerning the mechanism of
the reaction of Cl atoms with butanone. To improve our
understanding of the reaction of chlorine atoms with ketones
a study of the reaction of chlorine atoms with butanone was
conducted. The reaction proceeds via three channels:
Ketones are an important class of oxygenated volatile
organic compounds (VOCs) used as solvents and formed
during the atmospheric oxidation of organic compounds
[1,2]. Accurate kinetic data for reactions of chlorine atoms
with organic compounds are needed in atmospheric chem-
istry for two reasons. First, they are inputs into global
atmospheric models to assess the loss of organics via reac-
tion with Cl atoms in the marine boundary layer. Second,
they are used to analyze data from smog chamber experi-
ments in which chlorine atoms are used to initiate the oxi-
dation of organic compounds.
Mechanistic data concerning the reactions of chlorine
atoms with ketones are sparse. The reaction of Cl atoms
with acetone has been shown to proceed predominately
(>97%) via a hydrogen abstraction mechanism to give
CH3C(O)CH2 radicals and HCl [3]. The reaction of Cl
atoms with butadione (CH3C(O)C(O)CH3, biacetyl) pro-
Cl þ CH3CðOÞCH2CH3 ! HCl þ CH2CðOÞCH2CH3 ð1aÞ
Cl þ CH3CðOÞCH2CH3 ! HCl þ CH3CðOÞCHCH3 ð1bÞ
Cl þ CH3CðOÞCH2CH3 ! HCl þ CH3CðOÞCH2CH2 ð1cÞ
By measuring the yield of CH3C(O)CHClCH3 following
the UV irradiation of butanone/Cl2 mixtures in 700–
760 Torr of N2 diluent it was determined that k1b/
(k1a + k1b + k1c) = 0.73 0.09.
2. Experimental
*
Corresponding author.
E-mail address: twalling@ford.com (T.J. Wallington).
Present address: Frontier Research Center for Global Change, Japan
Agency for Marine-Earth Science and Technology, 3173-25 Showa-machi,
Kanazawa-ku, Yokohoma, Kanagawa 236-0001, Japan.
Present address: Kyoto University Pioneering Research Unit for Next
Generation, Gokasho, Uji 611-0011, Japan.
All experiments were performed using the smog cham-
ber system at Ford Motor Company, consisting of a
140 L Pyrex reactor interfaced to a Mattson Sirus 100 spec-
trometer [5]. The reactor was surrounded by 22 fluorescent
blacklamps (GE F15T8-BL). Relative rate techniques were
1
2
0009-2614/$ - see front matter ꢀ 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2007.03.082

E. Iwasaki et al. / Chemical Physics Letters 439 (2007) 274–279
275
used to measure the reactivity of Cl atoms towards
Cl þ CH3CðOÞCHClCH3 ! products
Cl þ C2H5Cl ! products
ð3Þ
ð4Þ
ð5Þ
CH3C(O)CHClCH3. The relative rate method is a well
established technique for measuring the reactivity of Cl
atoms with organic compounds [6]. Chlorine atoms were
generated by photolysis of molecular chlorine in 700 Torr
of N2 diluent at 296 2 K.
Cl þ CH3OCHO ! products
Reaction mixtures consisted of 22.4–28.1 mTorr of
CH3C(O)CHClCH3, 5.74–15.0 mTorr of reference, and
100 mTorr of Cl2 in 700 Torr total pressure of N2 diluent.
Cl2 þ hm ! Cl þ Cl
ð2Þ
Fig.
1
shows plots of Ln([CH3C(O)CHClCH3]to/
Kinetic data were derived by monitoring the loss of
CH3C(O)CHClCH3 relative to two reference compounds
(C2H5Cl or CH3OCHO). The decays of CH3C(O)CHClCH3
and reference were then plotted using the expression:
[CH3C(O)CHClCH3]t) vs. Ln([reference]to/[reference]t) in
the presence of Cl atoms. All plots are linear with inter-
cepts which were indistinguishable from the origin, suggest-
ing the absence of complications due to secondary
chemistry. Rate constant ratios, obtained from unweighted
linear least squares analysis of the data shown in Fig. 1
were k3/k4 = 0.712 0.062 and k3/k5 = 3.94 0.51. The
quoted uncertainties include two standard deviations from
the linear regression analyses and our estimate of uncer-
tainties in the spectral analysis.
ꢀ
ꢁ
½CH3CðOÞCHClCH3ꢁto
Ln
½CH3CðOÞCHClCH3ꢁt
ꢀ
ꢁ
k
½referenceꢁto
¼
CH3CðOÞCHClCH3 Ln
kreference
½referenceꢁt
where [CH3C(O)CHClCH3]to, [CH3C(O)CHClCH3]t, [refer-
ence]to and [reference]t are the concentrations of
CH3C(O)CHClCH3 and reference at times ‘to’ and ‘t’,
kCH3C(O)CHClCH3 and kreference are the rate constants for reac-
tions of Cl atoms with CH3C(O)CHClCH3 and reference.
Plots of Ln([CH3C(O)CHClCH3]to/[CH3C(O)CHClCH3]t)
versus Ln([reference]to/[reference]t) should be linear, pass
Using k4 = 8.04 · 10ꢀ12 [7] and k5 = 1.4 · 10ꢀ12
cm3 moleculeꢀ1 sꢀ1 [8] to place our relative rate measure-
ments on an absolute basis gives k3 = (5.72 0.50) ·
10ꢀ12 and (5.52 0.71) · 10ꢀ12 cm3 moleculeꢀ1 sꢀ1. Indis-
tinguishable values of k3 were obtained using the two dif-
ferent references. We choose to cite a final value which is
the average together with error limits which encompass
the extremes of the individual determinations;
k3 = (5.62 0.81) · 10ꢀ12 cm3 moleculeꢀ1 sꢀ1. Comparing
this result to k(Cl + CH3C(O)CH2CH3) = (4.08 0.37) ·
10ꢀ11 cm3 moleculeꢀ1 sꢀ1 (average from [9–11]) we con-
clude that substitution of a chlorine atom on the third-posi-
tion in butanone reduces the reactivity of the molecule by a
factor of approximately 7 at 296 K.
through the origin and have a slope of kCH3C(O)CHClCH3
kreference
/
.
The loss of the reactant and reference compounds was
monitored by FTIR spectroscopy using an infrared optical
path length of 27 m and a resolution of 0.25 cmꢀ1. Infrared
spectra were derived from 32 co-added interferograms.
Analysis of the IR spectra was achieved through spectral
stripping, in which small fractions of the reference spec-
trum were subtracted incrementally from the sample spec-
trum. Reagents were obtained from commercial sources at
>99% purity and were subjected to repeated freeze/pump/
thaw cycling before use. Ultra-high-purity nitrogen was
used as the diluent gas.
In smog chamber experiments it is important to check
for unwanted loss of reactants and products via photolysis,
dark chemistry and heterogeneous reactions. Control
experiments were performed in which (i) mixtures of reac-
tants (except Cl2) were subjected to UV irradiation for 10–
20 min and (ii) product mixtures obtained after the UV
irradiation of reactant mixtures were allowed to stand in
the dark in the chamber for 20 min. There was no observa-
ble loss of reactants or products, suggesting that photoly-
sis, dark chemistry, and heterogeneous reactions are not
significant complications in the present work. Unless stated
otherwise, quoted uncertainties are two standard devia-
tions from least squares regressions.
1.0
CH3OCHO
C2H5Cl
0.5
0.0
3. Results
0.0
0.5
1.0
1.5
Ln ([Reference]to/[Reference]t)
3.1. Relative rate study of k(Cl + CH3C(O)CHClCH3)
Fig. 1. Loss of CH3C(O)CHClCH3 versus C2H5Cl (circles) and CH3O-
CHO (triangles) following exposure to Cl atoms in 700 Torr of N2 at
296 K.
The rate of reaction (3) was measured relative to reac-
tions (4) and (5):

276
E. Iwasaki et al. / Chemical Physics Letters 439 (2007) 274–279
3.2. Mechanism of the reaction of Cl with
CH3C(O)CH2CH3
1.0
0.8
0.6
0.4
0.2
0.0
1.0
0.8
0.6
0.4
0.2
0.0
a: Before Irradiation
b: After Irradiation
c: Product Spectrum
d: CH3C(O)Cl
To investigate the reaction of Cl with CH3C(O)CH2CH3
experiments were performed using UV irradiation of mix-
tures of 42.5–45.1 mTorr of CH3C(O)CH2CH3 and 2010–
3951 mTorr of Cl2 in 700 Torr of N2 diluent. Reaction mix-
tures were subjected to 5–8 successive irradiations each
having a duration of 5–10 s. The CH3C(O)CHCH3 radical
formed in reaction (1b) reacts with molecular chlorine to
give 3-chloro-butanone, CH3C(O)CHClCH3. The 3-
chloro-butanone yield provides a measure of the branching
ratio k1b/(k1a + k1b + k1c).
0.2
0.1
0.0
CH3CðOÞCHCH3 þ Cl2 ! CH3CðOÞCHClCH3
0.6
0.4
0.2
0.0
þ HCl
ð6Þ
To facilitate measurement of the butanone loss a small
amount (4–8 mTorr) of C2H4 was added to the reaction
mixtures. C2H4 has highly structured IR features which
can be quantified (to a precision of 1% of the original
concentration) by FTIR spectroscopy more conveniently
than those of butanone. The relative reactivity of chlorine
atoms towards C2H4 and butanone has been established re-
cently; butanone loss was calculated from the observed
C2H4 loss using k1/k7 = 0.44 [10].
0.6
0.4
0.2
0.0
e: CH3C(O)CHClCH3
800
1000
1200
1400
1600
1800
2000
-1
Wavenumber (cm )
Fig. 2. IR spectra before (a) and after (b) UV irradiation of a mixture of
45.1 mTorr CH3C(O)CH2CH3, 3951 mTorr Cl2, and 7.8 mTorr of C2H4 in
700 Torr of N2 at 296 K. Panel (c) shows the results of subtracting IR
features attributable to CH3C(O)CH2CH3 from (b). Panels (d) and (e) are
reference spectra of CH3C(O)Cl and CH3C(O)CHClCH3.
Cl þ C2H4 ! products
ð7Þ
Typical spectra obtained before (a) and after (b) a 4 sec-
ond irradiation of a mixture containing 45.1 mTorr
CH3C(O)CH2CH3, 3951 mTorr Cl2, and 7.8 mTorr of
C2H4 in 700 Torr of N2 are shown in Fig. 2. From the
observed 40% consumption of C2H4 it was calculated that
the butanone consumption was 20%. Panel c shows the
result of subtracting the IR features attributable to buta-
none from panel b. Comparison of panel c with reference
spectra of CH3C(O)Cl and CH3C(O)CHClCH3 in panels
d and e shows the formation of these two products.
CH3CHO was also observed as a product by virtue of its
characteristic IR features at 1353, 1395, 1436, and
1746 cmꢀ1. Following the subtraction of features attribut-
able to CH3C(O)Cl, CH3C(O)CHClCH3, and CH3CHO
small unidentified features were observed at 740, 1159,
Cl þ CH3CðOÞCH2CH3 ! HCl
þ CH3CðOÞCHCH3
ð1bÞ
CH3CðOÞCHCH3 þ Cl2 ! CH3CðOÞCHClCH3 þ Cl ð6Þ
CH3CðOÞCHCH3 þ O2 þ M ! CH3CðOÞCHðOOÞCH3 þ M
ð8Þ
2CH3CðOÞCHðOOÞCH3 ! 2CH3CðOÞCHðOÅÞCH3 þ O2
ð9Þ
CH3CðOÞCHðOÅÞCH3 þ M ! CH3CðOÞ þ CH3CHO ð10Þ
1366, and 1740 cmꢀ1
.
CH3CðOÞ þ Cl2 ! CH3CðOÞCl þ Cl
CH3CðOÞ þ O2 þ M ! CH3CðOÞO2 þ M
ð11Þ
ð12Þ
As discussed above, we expect the formation of 1-, 3-,
and 4-chlorobutanone. The formation of CH3C(O)Cl was
unexpected and is quite striking (see Fig. 2). There are
two possible sources of CH3C(O)Cl. First, as a primary
product in the reaction of Cl atoms with butanone via
channel (1d).
To check for CH3C(O)Cl formation via reaction (1d) an
experiment was performed in 700 Torr of air. Using k11/
k12 = 7.91 [12] it can be calculated that reaction (12) will
account for 99% of the CH3C(O) radical loss in 700 Torr
of air. If reaction (11) is a source of CH3C(O)Cl in the
nominally pure N2 experiments then there should be a large
change in the observed CH3C(O)Cl yield on switching from
N2 to air diluent. If reaction (1d) is important then the yield
of CH3C(O)Cl is not expected to be significantly different
in N2 and air diluent. There was no observable CH3C(O)Cl
formation (<1% yield) following the UV irradiation of
Cl þ CH3CðOÞCH2CH3 ! CH3CðOÞCl þ C2H5
ð1dÞ
Second, as a secondary product following reaction of
CH3C(O)CHCH3 radicals with O2 impurity in the N2 dilu-
ent in the chamber (reaction of alkoxy and peroxy radicals
with Cl2 is endothermic and not considered in the mecha-
nism below):

E. Iwasaki et al. / Chemical Physics Letters 439 (2007) 274–279
277
CH3C(O)CH2CH3/Cl2/air mixtures. We conclude that
reaction (1d) is not a significant contribution to the
CH3C(O)Cl formation observed in the N2 experiments.
Fig. 3 shows a plot of the observed formation of
CH3C(O)CHClCH3, CH3C(O)Cl, and CH3CHO versus
the fractional loss of butanone. As seen from Fig. 3, the
concentration of CH3C(O)CHClCH3 increased linearly
with the loss of butanone. This behavior suggests that
CH3C(O)CHClCH3 is formed as a primary product and
that there are no significant losses of CH3C(O)CHClCH3
in the system. The low reactivity of CH3C(O)CHClCH3
towards Cl atoms reported in Section 3.1 is consistent with
the linearity of the CH3C(O)CHClCH3 yield plot in Fig. 3.
The line through the CH3C(O)CHClCH3 data is a linear
least squares fit which gives a molar yield of 0.54 0.03.
In stark contrast to the behavior of CH3C(O)CHClCH3,
the yield plots for CH3C(O)Cl and CH3CHO in Fig. 3 dis-
play pronounced curvature. The CH3C(O)Cl yield
increases while the CH3CHO yield decreases with butanone
consumption. The simplest chemical mechanism which
explains the observed product trends consists of reactions
(1b), (6), (8)–(13).
ond, the mechanism relies on inefficient scavenging of
CH3C(O)CHCH3 radicals by reaction with Cl2 but efficient
scavenging of CH3C(O) radicals.
In attempts to eliminate the O2 impurity the chamber
was flushed with N2 repeatedly before conducting the
experiments, different cylinders of ultra high purity N2 were
employed, and an experiment was performed using argon
diluent gas. However, in all experiments significant
amounts of CH3C(O)Cl product were observed indicating
the presence of O2. We attribute the presence of O2 to a
leak in which air entered the chamber while it was filled.
In an experiment conducted using a mixture with low
[Cl2] (100 mTorr Cl2 and 44.6 mTorr butanone in 700 Torr
N2) the yield of 3-chloro-butanone fell to 9%. In the mech-
anism proposed above there is a competition between reac-
tions (6) and (8) for the available CH3C(O)CHCH3
radicals. With low [Cl2] reaction (6) becomes a less impor-
tant fate for CH3C(O)CHCH3 radicals and the yield of 3-
chloro-butanone is expected to decrease. The observed
decrease in 3-chloro-butanone is consistent with expecta-
tions based upon the reactions mechanism proposed above
and the presence of O2 impurity in the system.
To estimate the likely magnitude of the O2 impurity an
experiment was conducted in which 20 mTorr of O2 was
added to a mixture of 2004 mTorr Cl2 and 44.7 mTorr
butanone in 700 Torr N2. In this experiment the yield of
3-chloro-butanone (29%) was approximately a factor of 2
lower than that observed in comparable experiments in
the absence of added O2 (e.g., Fig. 3). We conclude that
with regard to the competition between reactions (6) and
(8), 20 mTorr of O2 is approximately as effective as
2004 mTorr of Cl2 and hence k6/k8 is approximately 0.01
and that the O2 impurity is of the order of 20 mTorr.
The rate constants for reactions of alkyl radicals with O2
are relatively insensitive to the identity of the alkyl radical
and are typically of the order of 10ꢀ12 to 10ꢀ11 cm3 mole-
culeꢀ1 sꢀ1. For example, Kaiser reported high-pressure
Cl þ CH3CHO ! HCl þ CH3CðOÞ
ð13Þ
CH3C(O)Cl is formed by reaction of CH3C(O) radicals
with molecular chlorine. The decomposition of CH3C(O)-
CH(OÅ)CH3 radicals is both a direct, reaction (10), and
indirect, reaction (10) followed by reaction (13), source of
CH3C(O) radicals. The mechanism proposed above is
somewhat unusual in two respects. First, we need to invoke
the presence of a small O2 impurity in the chamber. Sec-
0.30
0.25
0.20
0.15
0.10
0.05
0.00
limiting rate constants of k (CH3 + O2 + M) = 1.32 ·
1
10ꢀ12 [13] and k (C2H5 + O2 + M) = (9.2 0.9) ·
1
10ꢀ12 cm3 moleculeꢀ1 sꢀ1[14]. In contrast, the rate con-
stants for reactions of alkyl radicals with Cl2 are sensitive
to the identity of the alkyl radical and span the range
10ꢀ15 (e.g., CH2ClCCl2 [15]) to 10ꢀ11 cm3 moleculeꢀ1 sꢀ1
(e.g., CH3C(O) [12]). The value of k6 = 10ꢀ14 to
10ꢀ13 cm3 moleculeꢀ1 sꢀ1 implied by our estimate of k6/
k8 = 0.01 does not appear unreasonable. In contrast to
the behavior of CH3C(O)CHCH3 radicals, CH3C(O) radi-
cals react more rapidly with Cl2 than with O2 (k11/
k12 = 7.91 0.49 in 700 Torr of N2 [12]). CH3C(O) radi-
cals are approximately 800 times less sensitive to the pres-
ence of O2 impurity than CH3C(O)CHCH3 radicals. For
experiments conducted with 2010–3951 mTorr of Cl2 the
fate of CH3C(O) radicals is reaction with Cl2 to give
CH3C(O)Cl.
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Δ[CH3C(O)CH2CH3]/[CH3C(O)CH2CH3]o
Fig. 3. Formation of CH3C(O)CHClCH3 (stars), CH3C(O)Cl (circles),
and CH3CHO (triangles) versus CH3C(O)CH2CH3 loss following UV
irradiation of a mixture of 42.5 mTorr of CH3C(O)CH2CH3, 2140 mTorr
Cl2, and 4.4 mTorr C2H4 in 700 Torr of N2 at 296 K. The lines are least
squares fits to the data, see text for details.
As discussed above, we believe there are both primary
and secondary sources of CH3C(O)Cl in the system. We
attribute curvature of the CH3C(O)Cl product yield plot

278
E. Iwasaki et al. / Chemical Physics Letters 439 (2007) 274–279
Table 1
Measured molar product yields for CH3C(O)CHClCH3 and CH3C(O)Cl
Diluenta
[butanone]o
[Cl2]o
CH3C(O)CHClCH3
CH3C(O)Clc
Sumd
b
b
Experiment
#1
#2
#3
#4
#5
#6
a
760 N2
700 N2
760 N2
760 Ar
760 N2
760 N2/0.02 O2
45.1
42.5
45.0
44.1
44.6
44.7
3951
2140
2010
1006
100
0.56
0.54
0.48
0.54
0.09
0.29
0.22
0.18
0.16
0.24
0.47
0.31
0.78
0.71
0.64
0.78
0.56
0.60
2004
Units of Torr.
Units of mTorr.
b
c
Primary component (see text for details).
Sum of CH3C(O)CHClCH3 and CH3C(O)Cl.
d
in Fig. 3 to the secondary source following reaction of Cl
atoms with CH3CHO. This secondary source increases in
importance with increasing butanone consumption. Using
the analytical solution to the set of differential equations
which describe the formation of such a secondary product
[16] we can fit expression (I) to the CH3C(O)Cl data to sep-
arate the primary and secondary components
of the combined yields of CH3C(O)Cl and CH3CHO. This
alternative approach yields k1b/(k1a + k1b + k1c) =
0.71 0.04. The results from the six sets of experiments
conducted in the present work are summarized in Table
1. As discussed above, experiments #5 and 6 were con-
ducted to test the experimental mechanism. These experi-
ments were not designed to give quantitative information
concerning k1b/(k1a + k1b + k1c) and are shown in Table 1
for completeness. The results from experiments #1–4 pro-
vide a consistent picture of the importance of reaction
channel k1b. Taking an average of these determinations
with an uncertainty which encompasses the extremes of
YðCH3CðOÞClÞ ¼ ðaꢂxÞ
n
o
(
"
#!)
k
k
13
ꢀ1
a
1
þ
ðaꢂxÞꢀ
ð1ꢀxÞ ð1ꢀxÞ
ꢀ1
k13
1ꢀ k
1
ðIÞ
the
individual
determinations
gives
k1b/
where Y(CH3C(O)Cl) is the ratio of CH3C(O)Cl concentra-
tion to the initial concentration of butanone, a is the yield
of CH3C(O)CH(OÅ)CH3 radicals in the system, x is the
fractional conversion of butanone, and k13/k1 is the ratio
of rate constants for reactions (13) and (1). Literature data
for k13 = 7.9 · 10ꢀ11 [17] and k1 = 4.08 · 10ꢀ11 cm3 mole-
culeꢀ1 sꢀ1 [9–11] were used to fix k13/k1 = 1.94. The solid
curve line in Fig. 3 shows a fit of expression (I) to the
CH3C(O)Cl data with a varied to provide a best fit value
of 0.18 0.01. As seen from Fig. 3, expression (I) provides
a good description of the trend of the CH3C(O)Cl data.
(k1a + k1b + k1c) = 0.73 0.09.
4. Discussion
We present a body of experimental data showing that
the reaction of Cl atoms with butanone proceeds predom-
inately via reaction at the –CH2– site. Combining k1b/
(k1a + k1b + k1c) = 0.73 0.09 with k1 = (4.08 0.37) ·
10ꢀ11 gives k1b = (2.98 0.46) · 10ꢀ11 cm3 moleculeꢀ1 sꢀ1
.
Based upon recent work in our laboratories [10,11] it seems
reasonable to assign k1a = 0.5 · k(Cl+acetone) = (1.05
0.10) · 10ꢀ12 [18], hence we conclude that k1c = (9.95
The dotted line in Fig. 3 is a x and is the primary compo-
*
nent of the CH3C(O)Cl yield. The dashed line shows the
predicted behavior of CH3CHO in the system from expres-
sion (I) and the chemical mechanism described above
which provides a reasonable description of the observed
CH3CHO concentrations in the system. The weak IR fea-
tures, low yield, and reactive nature of CH3CHO explain
the relatively large uncertainties on the CH3CHO data
points in Fig. 3. For reasons which are unclear, the major-
ity of the CH3CHO data points in Fig. 3 lie below the val-
ues expected (dashed line) from the fit to the CH3C(O)Cl
(solid curved line).
4.70) · 10ꢀ12 cm3 moleculeꢀ1 ꢀ1. These values give k1a/
s
(k1a + k1b + k1c) = 0.026 0.003, k1b/(k1a + k1b + k1c) =
0.73 0.09, and k1c/(k1a + k1b + k1c) = 0.24 0.09. The
reactivity of Cl atoms towards the CH3– and –CH2– groups
in n-butane is 3.0 · 10ꢀ11 and 7.3 · 10ꢀ11 cm3 mole-
culeꢀ1 sꢀ1, respectively [18]. Compared to n-butane, the
CH3– groups a and b to the ˜C@O group in butanone
are deactivated towards attack by Cl atoms by factors of
approximately 30 and 3, respectively. The –CH2– group
in butanone is approximately a factor of 2.5 less reactive
than those in n-butane.
The combined yield of 3-chloro-butanone and the pri-
mary component of the CH3C(O)Cl yield provide a mea-
surement of the fraction of Cl atoms which react with
butanone at the –CH2– group. From the experimental data
It is of interest to compare the behavior of Cl atoms with
OH radicals. The reaction of OH radicals with butanone
proceeds with a rate constant of 1.2 · 10ꢀ12 cm3 mole-
culeꢀ1 sꢀ1 at 298 K [18] with 62 2% of reaction at the
–CH2– site [19]. Hence, reaction at the –CH2– site has a
rate constant of 7.44 · 10ꢀ13 cm3 moleculeꢀ1 sꢀ1. The reac-
tion of OH with n-butane has a rate constant of 2.3 ·
shown in Fig.
3
we derive k1b/(k1a + k1b + k1c) =
(0.54 0.03) + (0.18 0.01) = 0.72 0.04. An alternative
approach is to sum the 3-chloro-butanone yield and 50%

E. Iwasaki et al. / Chemical Physics Letters 439 (2007) 274–279
279
10ꢀ12 cm3 moleculeꢀ1 sꢀ1 [18] with 87% of reaction occur-
ring at the –CH2– group [20]. Hence, reaction at each
–CH2– site has a rate constant of 1.00 · 10ꢀ12 cm3 mole-
culeꢀ1 sꢀ1. The reactivity of OH radicals towards the
–CH2– group in butanone is approximately 25% less than
that in n-butane. The reactivity of OH radicals towards
each of the CH3– groups in acetone (9.0 · 10ꢀ14 [18]) is
approximately 40% less than in propane (1.54 · 10ꢀ13 cm3
moleculeꢀ1 sꢀ1 [18,21]).
negligible for c C–H bonds. Further experiments are
required to confirm or refute this conclusion.
Acknowledgements
The Nagoya group thanks Grant-in-Aid from the Min-
istry of Education, Culture, Sports, Science, and Technol-
ogy, Japan. This work was also supported by the Global
Environment Research Fund (A-1).
As with the Cl atom reactions, the presence of the
˜C@O group deactivates adjacent CH3– and –CH2–
groups towards attack by OH radicals. However, the mag-
nitude of the deactivation is much smaller for the OH rad-
ical reactions.
References
[1] G.P. Brasseur, J.J. Orlando, G.S. Tyndall, Atmospheric Chemistry
and Global Change, Oxford University Press, Oxford, UK, 1999.
[2] B.J. Finlayson-Pitts, J.N. Pitts Jr., Chemistry of the Upper and Lower
Atmosphere: Theory, Experiments, and Applications, Academic
Press, New York, 2000.
Thermochemistry offers the simplest explanation of the
different magnitudes of the deactivating effect of the
˜C@O group on the Cl atom and OH radical reactions.
The strength of the H–OH bond (492 kJ molꢀ1 [22]) is
substantially greater than those of C–H bonds in alkanes
(e.g., primary C–H in propane = 423 kJ molꢀ1, secondary
C–H in propane = 416 kJ molꢀ1 [23]). In contrast, the
strength of the H–Cl bond (432 kJ molꢀ1) is close to those
of the C–H bonds in alkanes (e.g., primary C–H in
[3] O.J. Nielsen, M.S. Johnson, T.J. Wallington, L.K. Christensen, J.
Platz, Int. J. Chem. Kinet. 34 (2002) 283.
[4] L.K. Christensen, J. Sehested, O.J. Nielsen, T.J. Wallington, A.
Guschin, M.D. Hurley, J. Phys. Chem. A 102 (1998) 8913.
[5] T.J. Wallington, S.M. Japar, J. Atmos. Chem. 9 (1989) 399.
[6] R. Atkinson, J. Phys. Chem. Ref. Data, Monograph 1, 1989.
[7] P.H. Wine, D.H. Semmes, J. Phys. Chem. 87 (1983) 3572.
[8] T.J. Wallington, M.D. Hurley, J.C. Ball, M.E. Jenkin, Chem. Phys.
Lett. 211 (1993) 41.
[9] T.J. Wallington, J.M. Andino, J.C. Ball, S.M. Japar, J. Atmos. Chem.
10 (1990) 301.
[10] F. Taketani, Y. Matsumi, T.J. Wallington, M.D. Hurley, Chem.
Phys. Lett. 431 (2006) 257.
[11] K. Takahashi, E. Iwasaki, Y. Matsumi, T.J. Wallington, J. Phys.
Chem. A 111 (2007) 1271.
[12] E.W. Kaiser, T.J. Wallington, J. Phys. Chem. 99 (1995) 8669.
[13] E.W. Kaiser, J. Phys. Chem. 97 (1993) 11681.
[14] E.W. Kaiser, J. Phys. Chem. 99 (1995) 707.
[15] P.B. Ayscough, F.S. Dainton, B.E. Fleischfresser, Trans. Faraday
Soc. 62 (1966) 1838.
[16] R.J. Meagher, M.E. McIntosh, M.D. Hurley, T.J. Wallington, Int. J.
Chem. Kinet. 29 (1997) 619.
[17] R. Atkinson, D.L. Baulch, R.A. Cox, J.N. Crowley, R.F. Hampson,
R.G. Hynes, M.E. Jenkin, M.J. Rossi, J. Troe, Atmos. Chem. Phys.
Discuss. 5 (2005) 6295.
[18] R. Atkinson, D.L. Baulch, R.A. Cox, J.N. Crowley, R.F. Hampson
Jr., M.E. Jenkin, J.A. Kerr, M.J. Rossi, J. Troe, IUPAC Subcom-
mittee on Gas Kinetic Data Evaluation, <http://www.iupac-kine-
tic.ch.cam.ac.uk/index.html>, (accessed August 2006).
[19] R.A. Cox, K.F. Patrick, S.A. Chant, Environ. Sci. Technol. 15 (1981)
587.
[20] A.T. Droege, F.P. Tully, J. Phys. Chem. 90 (1986) 5937.
[21] A.T. Droege, F.P. Tully, J. Phys. Chem. 90 (1986) 1949.
[22] B. Ruscic, D. Feller, D.A. Dixon, K.A. Peterson, L.B. Harding, R.L.
Asher, A.F. Wagner, J. Phys. Chem. A 105 (2001) 1.
[23] NIST Standard Reference Database 25, Structures and Properties,
(1994) Version 2.02, National Institute for Standards and Technol-
ogy, Gaithersburg, Maryland.
butane = 420 kJ molꢀ1
, secondary C–H in butane =
415 kJ molꢀ1 [23]). Relatively small changes in the C–H
bond strengths resulting from the introduction of the
˜C@O group in the molecule may have significant impacts
on the activation energy barriers and hence kinetics of the
Cl atom reactions, particularly for attack on the CH3–
group. However, it is worth noting that data from the
IUPAC compilation [17] indicate that C–H bonds in
–CH3 groups a to the carbonyl group in acetaldehyde
and acetone are actually weaker than C–H bonds in
ethane and propane. Thermochemical arguments are not
able to explain the differences in reactivities of –CH3
groups in acetone, acetaldehyde, ethane, and propane.
Theoretical studies of the thermochemistry and dynamics
of the reactions of Cl atoms with ketones are needed to
provide further insight into the factors affecting the reac-
tions of Cl atoms with ketones.
Finally, we believe that the present work provides the
first compelling evidence for a deactivating effect of the
˜C@O group on C–H bonds beyond the a position. Previ-
ous reports of a large and long range deactivating effect [24]
were not confirmed in subsequent work [25,26]. We report
here that the CH3– group b to the ˜C@O group in buta-
none is approximately 3 times less reactive towards Cl
atoms than CH3– groups in alkanes. However, the magni-
tude of the deactivation observed at the b position is
approximately 10 times less than observed at the a position.
We conclude that for the reactions of Cl atoms with
ketones, the deactivating effect of the ˜C@O group is large
for a C–H bonds, modest for b C–H bonds, and probably
[24] B. Olsson, M. Hallquist, E. Ljungstro¨m, J. Davidsson, Int. J. Chem.
Kinet. 29 (1997) 195.
[25] T.J. Wallington, A. Guschin, M.D. Hurley, Int. J. Chem. Kinet. 30
(1998) 309.
[26] E. Martinez, A. Aranda, Y. Diaz-De-Mera, A. Rodriguez, D.
Rodriguez, A. Notario, J. Atmos. Chem. 48 (2004) 283.