Rate constants for H and OH attack on propanone, butanone and pentan-3-one at 753 K, and the oxidation chemistry of the radicals formed

Article abstract of DOI:10.1039/b009378p

Elementary reactions involved in the oxidation of propanone, butanone and pentan-3-one were elucidated by adding very small amounts of the compound individually to slowly reacting mixtures of H₂ + O₂ at 753 K. Kinetic studies of the relative rates of consumption of the additive and H₂ show that (21) and (22) are the key reactions consuming the ketones. Values of k₂₁ = 1.76 × 10⁹ dm³ mol^-1 s^-1, for propanone, butanone and pentan-3-one, respectively, are obtained, together with k₂₂ = 2.34 × 10⁸, 7.2 × 10⁸ and 1.20 × 10⁹ dm³ mol^-1 s^-1, respectively. OH + ketone → H₂O + ketonyl (21) H + ketone → H₂ + ketonyl (22) The values of k₂ show a 30-40% decrease by comparison with those of the analogous rate constants obtained in this laboratory for the structurally-related alkanes, whereas the values of k₂₂ show a 10-15% increase. Although independent data are extremely limited, Arrhenius parameters were obtained by combination with the results of other workers. Arguments are presented for the occurrence of a reaction not previously reported in the literature, namely the addition of H atoms to propanone, which may occur prominently at relatively high temperatures. H + CH₃COCH₃ → CH₃C(OH)CH₃ Analytical studies of the oxidation chemistry of the ketones at 753 K were carried out by adding individually 5 Torr of each to a mixture containing 70, 140 and 285 Torr of O₂, H₂ and N₂, respectively. With this approach, each oxidation occurs under effectively the same highly controllable conditions. The mechanism for the formation of products is discussed in the context of modern understanding of hydrocarbon oxidation, and the use of a self-consistent set of kinetic data gives a good prediction of the product yields in the initial stages of reaction. The low yields of methyl vinyl ketone from butanone and ethyl vinyl ketone from pentan-3-one were rationalised through considerations of the thermochemistry and kinetics involved. Values of log(A/s^-1) = 11.72 and E = 136 kJ mol^-1 are recommended for the 1,5p H atom transfer reaction in CH₃COCH₂O₂ radicals CH₃COCH₂O₂ → CH₂COCH₂O₂H.

Full text of DOI:10.1039/b009378p

View Article Online / Journal Homepage / Table of Contents for this issue
Rate constants for H and OH attack on propanone, butanone and
pentan-3-one at 753 K, and the oxidation chemistry of the radicals
formed
R. S. Tranter and R. W. Walker*
Department of Chemistry, University of Hull, Hull, UK HU6 7RX
Received 22nd November 2000, Accepted 30th January 2001
Elementary reactions involved in the oxidation of propanone, butanone and pentan-3-one were elucidated by
adding very small amounts of the compound individually to slowly reacting mixtures of H ] O at 753 K.
2
2
Kinetic studies of the relative rates of consumption of the additive and H show that (21) and (22) are the key
2
reactions consuming the ketones. Values of k \ 1.76 ] 109, 3.16 ] 109 and 3.83 ] 109 dm3 mol~1 s~1, for
propanone, butanone and pentan-3-one, respectively, are obtained, together with k \ 2.34 ] 108, 7.2 ] 108
2
1
2
2
and 1.20 ] 109 dm3 mol~1 s~1, respectively.
OH ] ketone ] H O ] ketonyl
(21)
(22)
2
H ] ketone ] H ] ketonyl
2
The values of k show a 30È40% decrease by comparison with those of the analogous rate constants obtained
21
in this laboratory for the structurally-related alkanes, whereas the values of k show a 10È15% increase.
2
2
Although independent data are extremely limited, Arrhenius parameters were obtained by combination with
the results of other workers. Arguments are presented for the occurrence of a reaction not previously reported
in the literature, namely the addition of H atoms to propanone, which may occur prominently at relatively
high temperatures.
H ] CH COCH ] CH C(OH)CH
3
3
3
3
Analytical studies of the oxidation chemistry of the ketones at 753 K were carried out by adding individually 5
Torr of each to a mixture containing 70, 140 and 285 Torr of O , H and N , respectively. With this
2
2
2
approach, each oxidation occurs under e†ectively the same highly controllable conditions. The mechanism for
the formation of products is discussed in the context of modern understanding of hydrocarbon oxidation, and
the use of a self-consistent set of kinetic data gives a good prediction of the product yields in the initial stages
of reaction. The low yields of methyl vinyl ketone from butanone and ethyl vinyl ketone from pentan-3-one
were rationalised through considerations of the thermochemistry and kinetics involved. Values of
log(A/s~1) \ 11.72 and E \ 136 kJ mol~1 are recommended for the 1,5p H atom transfer reaction in
CH COCH O radicals
3
2 2
CH COCH O ] CH COCH O H
3
2 2
2
2 2
The OH radical plays a key role in both tropospheric chem-
istry and combustion. Rate constants over extensive tem-
perature ranges have been obtained for a series of alkanes
with sufficient accuracy to provide semi-empirical relation-
ships with structure,1,2 so that rate constants may be esti-
mated with considerable conÐdence and accuracy where no
experimental data exist. However, at temperatures in excess of
ambient, there is little information on OH attack on alkane
derivatives, of which oxygenated species such as alcohols,
aldehydes, ketones, and ethers, in particular, have important
roles in combustion. Several measurements of OH
dm3 mol~1 s~1 at 1200 K, obtained by Bott and Cohen7 for
OH ] propanone by use of a shock tube.
Kinetic data for H ] ketones are available only for propa-
none. Ambridge et al.8 reported log(A/dm3 mol~1
s~1) \ 10.27 and E \ 26.6 kJ mol~1, obtained by use of a dis-
charge tube with ESR detection of H atoms. As discussed later
these parameters are almost certainly too low for the
abstraction path.
The absence of kinetic data for reactions of key radicals
with important organic species is unfortunate not only
because modelling studies of combustion are impeded, but
also because theoretical and semi-empirical attempts to calcu-
late constants cannot be validated for a wide range of com-
pounds. Thus, for example, CohenÏs seminal work on H9 and
OH10,11 on alkanes cannot be extended to related compounds
via validation.
]
propanone were carried out at room temperature in
the 1980s,3 together with Wallington and KuryloÏs study4
between 240 and 440 K. Recently, Le Calve et al.5 studied the
reaction of OH with propanone and several other ketones
between 243 and 372 K, and Wollenhaupt et al.6 carried out
an in-depth investigation of OH ] propanone between 202
and 395 K, and Ðrmly established the existence of a minimum
in the value of the rate constant at about 250 K, which they
attributed to a change of mechanism with addition being
dominant at the lower temperatures used. At combustion tem-
peratures, the sole value available for ketones is k \ 5.3 ] 109
The H ] O addition approach has proved invaluable in
2
2
determining two types of information, despite its limited tem-
perature range of 730È780 K.
(i) From a study of the relative rate of consumption of
hydrogen and the additive over a wide range of mixture com-
position, values of rate constants have been obtained for H
1262
Phys. Chem. Chem. Phys., 2001, 3, 1262È1270
This journal is ( The Owner Societies 2001
DOI: 10.1039/b009378p

View Article Online
and OH (and sometimes HO 12) attack on the additive. As
2
demonstrated in several reviews for H abstraction by H9 and
OH,10,13 this approach gives extremely accurate relative rate
data. This is particularly pertinent9 for the H atom because
very few reliable rate constants are available for any com-
pound containing CÈH bonds, frequently due to difficulties in
establishing the stoichiometry of the overall removal of H
atoms.
(
ii) From a detailed investigation of the initial product dis-
tribution over a range of mixture composition, the oxidation
chemistry of the radicals formed directly from the additive
may be elucidated. The main virtue of the approach, unparal-
leled in most other studies, is that the oxidation chemistry
takes place in an e†ectively constant radical environment
regardless of the additive, which makes elucidation of mecha-
nism and determination of rate constants much easier and
more reliable. The quality of the data is such that several
workers have used them to validate both mechanisms14 and
modelling approaches.15
This paper is primarily concerned with the kinetics of H
abstraction by H and OH radicals from propanone, butanone
and pentan-3-one, and for this purpose detailed product
analysis is not necessary. However, a number of analytical
measurements were made, and these will be discussed in the
context of mechanism and independent work.
Experimental
The apparatus and experimental procedure are extremely well
established.16 Pressure measurements were obtained by use of
a transducer, which on its maximum sensitivity range record-
ed pressure changes or absolute pressures down to 0.01 Torr
with an accuracy of 1%. Concentrations of products were
measured quantitatively by gas chromatography using a
number of packed columns to ensure good resolution of the
peaks. Products were identiÐed by retention time and
GC-MS. No analysis was made for water or peroxides, or
Fig. 1 Plots of additive concentration as a function of *P for di†er-
ent mixture compositions. Additive pressure \ 0.120 Torr, total pres-
sure with N \ 500 Torr; 753 K. (a) Pentan-3-one: ], O \ 70,
2
2
CO in the case of butanone and pentan-3-one. A cylindrical
H \ 35; L, O \ 70, H \ 140; |, O \ 70, H \ 430; …, O \ 35,
2
2
2
2
2
2
2
H \ 140; K, O \ 360, H \ 140 Torr. (b) Butanone: |, O \ 70,
Pyrex vessel, 5.2 cm in diameter and 20 cm in length, coated
2
2
2
2
H \ 35; È, O \ 70, H \ 430; ], O \ 35, H \ 140; ), O \
with boric acid, and aged by carrying out repeated reactions
between H and O , was used at 753 K.
2
2
2
2
2
2
3
60, H \ 140 Torr. (c) Propanone: L, O \ 70, H \ 35; K, O \
2
2
2
2
7
0, H \ 430; ], O \ 35, H \ 140; |, O \ 360, H \ 140 Torr
2
2
2
2
2
2
2
(
Ðlled symbols, in the presence of 0.120 Torr butanone).
Results and discussion
Kinetic results
slight uncertainties in the product yields and therefore in
In order to achieve e†ectively negligible changes in the relative
*P(KET) are unlikely to cause errors in excess of 1% in the
5
0
concentrations of H, O, OH and HO , both during reaction
values of *P
50 .
2
and when compared with those in the absence of the additive,
It is clear from Fig. 1(b) and 1(c) that the rate of decay of
propanone is considerably smaller than that of butanone. In
consequence, for propanone, *P , the pressure change for
only 0.024% of the individual ketones was used in the kinetic
experiments. The relative rates of consumption of the ketone
and H2 were measured experimentally over a wide range of
mixture composition, for the ketone by gas chromatography,
25
25% consumption of propanone, was used as the measure of
d[KET]/d[H ]. Correction of the results was then as for
2
and for H , by measurement of the pressure change *P due to
*P
50 .
2
the reaction 2H ] O \ 2H O. Fig. 1(a) shows typical plots
A computer program17 was used to calculate the values of
*P and *P (for propanone), to correct for the positive
2
2
2
of the decay of pentan-3-one with *P for several mixtures.
Similar plots for butanone are shown in Fig. 1(b), and in Fig.
5
0
25
pressure change due to the oxidation of KET, and to obtain
the minimum root mean square (rms) deviation between
observed and calculated values. Di†erential equations are
written for reactants and products, and stationary-state equa-
tions for the radicals. The di†erential equations are solved by
the standard KuttaÈRunga integration method, and a mini-
misation routine is used for the optimisation of agreement
between calculated and observed results. A comprehensive
mechanism was used for the H ] O reaction,17 and allow-
1(c) for propanone in the presence and absence of 0.120 Torr
of butanone. As indicated, the amounts of ketone added are so
small that the decays of propanone are the same in the pres-
ence and absence of butanone. With butanone and pentan-3-
one, as in previous studies, the pressure change *P , corre-
5
0
sponding to 50% consumption of ketone (KET), was used as a
measure of d[KET]/d[H ]. The values of *P are corrected
2
50
by a computer program for the formation of hydrogen peroxi-
de and for the relatively small positive pressure change,
2
2
ance was made for the slight amount of self-heating that
occurs. As usually observed, KET is removed mainly by H
*
P(KET) , caused by the oxidation of KET. The product
yields discussed later are used as the basis for correction. The
50
and OH attack, with smaller contributions from O and HO
2
values of *P(KET) determined in this way never exceed
radicals. In calculating *P or *P , these reactions enter
50
50
25
about 7% of *P and are typically less than 3%, so that
the computer program as k /k , k /k , k /k and k /k1@2.
50
21 1 22 2 23 3
24 10
1263
Phys. Chem. Chem. Phys., 2001, 3, 1262È1270

View Article Online
Table 1 Pressure changes corresponding to 50% consumption of additivea
Additive
Propanone
Butanone
Pentan-3-one
Mixture compositionb
H2
O2
*P
25(obs)
*P
25(cal)
*P
50(obs)
*P
50(cal)
*P
50(obs)
*P
50(cal)
3
7
5
0
70
70
70
70
70
70
70
35
140
210
280
360
0.80
1.36
2.38
3.91
4.69
5.02
6.17
1.82
3.25
4.05
4.23
4.55
0.79
1.34
2.48
3.55
4.48
5.27
6.06
1.77
3.30
3.82
4.19
4.41
0.74
1.06
1.71
2.49
2.63
3.33
3.52
1.21
2.30
2.66
2.79
2.98
0.76
1.09
1.66
2.25
2.76
3.17
3.56
1.32
2.26
2.65
2.94
3.09
0.38
0.80
1.18
1.62
1.91
2.36
2.49
0.74
1.64
1.91
2.33
2.66
0.44
0.73
1.20
1.67
2.04
2.33
2.55
0.77
1.58
2.01
2.33
2.59
1
2
2
3
4
1
1
1
1
1
40
10
80
50
30
40
40
40
40
40
rms deviation
4.9
5.1
6.2
a All pressures in Torr. b Additive pressure \ 0.12 Torr; total pressure \ 500 Torr with N .
2
Table 2 Sensitivity of optimum values of k /k and k
22/k2 to
2
1 1
OH ] KET \ KETONYL ] H O
(21)¤
(22)
(23)
(24)
(1)
changes in k /k and k /k1@2
2
23 24 10
3
H ] KET \ KETONYL ] H
Optimum values
2
O ] KET \ KETONYL ] OH
k
/k
(k /k1@2)a
k
/k
k
/k
22
rms deviationb
2
3
3
24 10
21
1
2
HO ] KET \ KETONYL ] H O
2
2 2
Propanone
.0c
0.5
4.0
2
0.4c
0.4
0.4
0.8
0.2
3.9
4.0
3.8
3.6
4.3
47
48
45
38
55
4.9
5.0
5.2
5.9
5.1
OH ] H \ H O ] H
2
2
H ] O \ OH ] O
(2)
2
2
.0
O ] H \ OH ] H
(3)
2.0
2
Butanone
HO ] HO \ H O ] O
(10)
4
1
5
4
.0c
1.2c
1.2
1.2
2.4
0.6
7.0
6.9
7.1
5.9
7.7
145
145
143
139
151
5.1
5.3
5.3
6.4
5.9
2
2
2 2
2
.0
.0
.0
Consistent with previous investigations, minimisation
involving four variables is not practicable and k /k and
2
3 3
k
/k1@2 are given Ðxed values for each ketone, and then the
4.0
2
4 10
values of k /k and k /k optimised. Sensitivity analysis,
Pentan-3-one
2
1 1
22 2
4
.0c
2.0c
2.0
2.0
4.0
1.0
8.5
8.7
8.4
7.2
242
239
234
227
259
6.2
6.4
6.1
7.1
6.9
shown later, emphasises that the Ðnal values are not sensitive
1
8.0
4
4
.0
to k /k (particularly) and k /k1@2. A relatively accurate
23 3
24 10
value of k /k \ 2.0 for propanone may be calculated from
.0
.0
2
3 3
literature values.18h21 Roscoe19 has determined Arrhenius
parameters for the reactions of O atoms with butanone and
pentan-3-one at temperatures below 753 K, but the data are
highly scattered and the rate constants do not di†er signiÐ-
cantly. Although a value of k /k D 3.0 is calculated, 4.0 is
10.1
a (dm3 mol~1 s~1)1@2 units. b rms deviation between calculated and
observed values of *P
(*P
for propanone). c Recommended
50
25
values.
2
3 3
assumed here for both ketones to magnify any e†ect of O
atom attack.
compares the data with those for related alkanes. The consis-
No experimental data are available for reaction (24) for any
ketone. Based on the results obtained here for H and OH
attack, relative to those for alkanes, and on values determined
tency is striking, with the values for k /k for the ketones
21 1
showing a 30È40% decrease by comparison with those for the
alkanes structurally-related in terms of CÈH bonds, whereas
previously for HO2
reaction at speciÐc CÈH bonds in
the values of k /k show a 10È15% increase.
alkanes,22 values of k /k1@2 \ 0.4, 1.2 and 2.0 (dm3 mol~1
22 2
24
10
Independent studies of OH ] ketone are mostly restricted
s~1)1@2 are taken for propanone, butanone and pentan-3-one,
respectively. By use of the computer optimisation procedure,
minimum rms deviations were obtained between the observed
to relatively low temperatures, where Wollenhaupt et al.6 have
suggested that the mechanism probably involves both
abstraction and addition. At 440 K, the highest temperature
used by Wallington and Kurylo,4 the addition process may be
considered unimportant and k (ethane)/k (propanone) \
and calculated values of *P
(and *P ) for the three
5
0
25
ketones. Table 1 shows the comparison of the observed and
calculated values for minimum rms deviation in each case.
Table 2 shows the sensitivities of the optimum values to
changes in k /k and k /k1@2. Clearly, k /k has negligible
21
21
0
]
.51 when k (propanone) is combined with k (ethane) \ 7.2
21
21
103T 2 exp([435/T ) dm3 mol~1 s~1 given by Baulch et
2
3 3
24 10
23 3
al.,23 compared with a ratio of 0.58 calculated from a shock-
tube study7 of the two reactions at 1200 K, both in good
agreement with the value of 0.65 at 753 K obtained from the
results given in Table 3.
e†ect, but the ratios do show some sensitivity to the value of
k
/k1@2, namely for OH attack on butanone and pentan-3-
one and for H attack on propanone. Table 3 summarises the
values of the ratios obtained here, indicates error limits, and
2
4 10
Data for H abstraction by other radicals from ketones are
restricted, but King et al.24 report that Br abstracts about four
times faster from ethane than from propanone between 520
and 600 K and further, combination of the data for
¤
The reaction numbers used are consistent with those in previous
publications.17
1264
Phys. Chem. Chem. Phys., 2001, 3, 1262È1270

View Article Online
Table 3 Summary and comparison of kinetic data for H and OH radicals at 753 Ka
Additive
k
/k
k
] 10~9
k
/k
22
k
] 10~8
Ref.
2
1
1
21
2
22
CH COCH
3.9 ^ 0.4
5.9 ^ 0.5
7.0 ^ 0.9
9.9 ^ 1.2
8.5 ^ 1.4
1.76
2.66
3.16
4.46
3.83
5.95
8.16
47 ^ 7
44 ^ 5
2.34
2.19
7.2
This work
17
This work
17
This work
17
17
3
3
C H
2
6
CH COC H
145 ^ 8
125 ^ 8
242 ^ 15
217 ^ 12
309 ^ 24
3
C H
3
2 5
6.2
8
C H COC H
12.0
10.8
15.4
2
5
2 5
CH (CH ) CH
2 2
CH (CH ) CH
2 3
13.2 ^ 1.3
18.1 ^ 2.3
3
3
3
3
a Absolute values have the units dm3 mol~1 s~1, and are calculated from the rate constant ratios as discussed in the text.
k(O ] propanone) given in HerronÏs20 review with
k(O ] ethane) from Baulch et al.23 gives an analogous Ðgure
of 4È6 at 700 K for O atoms. By contrast, the rate of H
abstraction by CH3 radicals from propanone is about twice
that for ethane at 650 K.23,25 Aspects of the electrophilicity of
the radical or atom, entropy of activation, and energy barrier
must be important in assessing the wide variation in the rela-
tive rate of H abstraction by the various radicals.
pentan-3-one ] OH ]
CH CH C(OH)CH CH ] C H ] C H COOH
3
2
3
2
2 5
2 5
=
O
On the assumption that addition is not important at 753 K,
and no evidence of acid formation was found in the present
study, nor in that of Akbar and BarnardÏs26 at 673 K, the
most reliable Arrhenius parameters for the abstraction process
in pentan-3-one are obtained by assuming the same activation
energy as for OH ] butanone with a suitably raised A factor,
namely, log (A /dm3 mol~1 s~1) \ 10.43 and E \ 12.2 kJ
Use of the expression,23 k \ 1.02 ] 105T 1.6 exp([1660/T )
1
dm3 mol~1 s~1 for the range 300È2500 K with the values of
k
/k , gives the values of k shown in Table 3. These and
21 1
21
independent values for OH ] KET are plotted in Arrhenius
form in Fig. 2. Although the data at high temperatures are
very restricted, there is consistency for the OH ] propanone
reaction, which as indicated shows a minimum at 250È270 K
and which Wollenhaupt et al.6 interpret mechanistically as the
occurrence of both addition (low temperature) and abstraction
2
1
21
mol~1.
Combination of the k /k values with k /k \ 4417 at
2
2 2 22e 2
7
53 K and with23 k \ 1.5 ] 106T 1.5 exp([3730/T ) dm3
2
2e
mol~1 s~1, gives the absolute rate constants for H attack on
the three ketones; they are given in Table 3.
(
high temperature). They write a double Arrhenius expression
to account for the temperature coefficient between 202 and
95 K, k \ 5.3 ] 109 exp([1320/T ) ] 1.02 ] 107 exp(423/
H ] C H ] C H ] H
2 5 2
(22e)
2
6
3
Harris and Steacie27 used a discharge method at room tem-
perature to generate H atoms and determined their concentra-
tion in the presence and absence of propanone. However, the
experimental conditions were so poorly deÐned that their
value of 104È105 dm3 mol~1 s~1 for H ] propanone cannot
even be considered an approximation. Ambridge et al.8 used a
discharge Ñow system with ESR detection of H atoms at total
pressures of 1È2 Torr between 298 and 465 K. There was no
analysis of products, and the reaction stoichiometry was based
on the relative loss of H atoms and propanone, and an
assumed mechanism. Although extrapolation of their results
to 753 K gives excellent agreement with the present value,
their activation energy of 26.6 kJ mol~1 appears extremely
low when compared in Arrhenius form and over the same
temperature range with E \ 41 kJ mol~1 for
21
T ) dm3 mol~1 s~1, which gives k \ 9.25 ] 108 at 753 K,
2
1
compared with 1.94 ] 109 dm3 mol~1 s~1 from the present
study, and 1.75 ] 109 at 1200 K compared with Bott and
CohenÏs7 value of 5.3 ] 109 dm3 mol~1 s~1, thus providing
clear evidence of further non-Arrhenius behaviour in the
abstraction path at the higher temperatures. With an allow-
ance of 5% for addition at 395 K, Arrhenius parameters of
log(A /dm3 mol~1 s~1) \ 10.41 ^ 0.35 and E \ 16.3 ^ 4
21
21
kJ mol~1 are obtained for propanone between 395 and 1200
K, but a non-Arrhenius expression k \ 2.75 ] 105T 1.5
exp([926/T ) dm3 mol~1 s~1 is preferred for use up to 2000
K. As the relative rate of abstraction from ethane and propa-
none by OH radicals is e†ectively constant between 395 and
2
1
1
200 K, the two rate constants must have almost identical
CH3
temperature coefficients.
As seen from Fig. 2, few data are available for
OH ] butanone. There is little increase in rate constant
between 300 and 440 K, and the results are not inconsistent
with a minimum at about 300 K. By inspection, and given the
enhanced reversibility of the addition process at higher tem-
peratures, it is unlikely to contribute more than 25% to the
total value of k at 440 K, and with this correction log
2
1
(A
/dm3 mol~1 s~1) \ 10.34 and E \ 12.2 kJ mol~1 for
21
21
the abstraction path in butanone between 440 and 753 K.
Interpretation of the values for OH ] pentan-3-one is even
more difficult. Only Wallington and Kurylo4 obtained rate
constants above room temperature, and they are almost con-
stant beween 300 and 440 K, and only about a factor of 2
lower than the value at 753 K. It is possible that the addition
process is still very important at 440 K which is consistent
with the expected higher degree of stability of the adduct com-
pared with that from propanone and butanone. The adduct
may dissociate back to reactants, but the formation of propi-
onic acid and ethyl radicals is thermodynamically favourable
with an overall exothermicity of 120 kJ mol~1, and very
similar to that for the abstraction process.
Fig. 2 Kinetic data for the reaction of OH radicals with propanone,
butanone and pentan-3-one. Line A, OH ] propanone: K, this work;
|
, Wallington and Kurylo;4 L, Wollenhaupt et al.;6 …, Bott and
Cohen;7 ], Le Calve et al.5 Line B, OH ] butanone; K, this work;
|, Wallington and Kurylo;4 ], Le Calve et al.5 Line C,
OH ] pentan-3-one; K, this work; |, Wallington and Kurylo.4
Phys. Chem. Chem. Phys., 2001, 3, 1262È1270
1265

View Article Online
]
propanone25 and with E \ 45 kJ mol~1 for H ] ethane.23
important inhibition process. Given the trace amounts of
Ambridge et al. in the same study report E \ 24 kJ mol~1 for
O ] propanone, e†ectively identical with their value for
propanone used here, the perturbation of [HO ] will be negli-
2
gible. It is also worth commenting on paths (b) and (c) in the
H ] propanone and the recommended value for
O
context of the present work. At 500 Torr and 753 K, the
]
C H ,20 but which is about 20 kJ mol~1 lower than for
adduct CH CH(O)CH would undergo rapid CÈC homolysis
2
6
3
3
H ] ethane.
to ethanal and CH . However, in the present studies only
3
A possible explanation that arises from consideration of the
thermochemistry is that H atoms add rapidly to propanone,
to give an excited adduct which may either revert to the reac-
tants or alternatively be stabilised through collision,
traces of ethanal were observed so that, although path (a) may
occur, the value of k
the rate of H abstraction by H atoms from propanone.
obtained here gives a true measure of
22p
Values (Table 3) of k \ 7.2 ] 108 and 1.20 ] 109 dm3
2
2
mol~1 s~1 are obtained for butanone and pentan-3-one,
respectively, at 753 K; no independent values are available.
The closeness of the values to those for propane and butane,
respectively, provides a strong argument for the use of
CohenÏs9 self-consistent set of data, determined by use of tran-
sition state theory for H ] alkane, for H abstraction from
ketones. However, there is some evidence, despite the paucity
of the information, that the closeness of the data for H
abstraction from alkanes and ketones arises from the com-
pensation of two e†ects. Two studies by Benson and co-
workers24,28 have suggested that although the CH COCH
H ] propanone
H CH C(OH)CH * ] M ] CH C(OH)CH (a)
3
3
3
3
*
H \ [116 kJ mol~1
H CH CH(O)CH * ] M ] CH CH(O)CH (b)
3
3
3
3
*
H \ [53 kJ mol~1
]
CH CHO ] CH (c)
3
3
*
H \ [23.5 kJ mol~1
3
2
Addition to the O atom [path (a)] is highly exothermic,
approaching that for the rapid H atom addition to ethene
radical has relatively little resonance energy of stabilisation,
that in the CH COCHCH radical may approximate to 10 kJ
3
3
(
[150 kJ mol~1), and a low energy barrier is likely because
mol~1. The e†ect of this would be to lower both the entropy
of activation and the energy barrier for H abstraction from
ketones. Reductions by a factor of 2 in the A factor (per CÈH
bond) and of 5 kJ mol~1 by comparison with CohenÏs rec-
ommendations for H ] alkanes are reasonable and consistent
with the data at 753 K.
the electrophilic H atom will be attracted to the electron-rich
O atom in propanone. At 1È2 Torr total pressure, most of the
excited radicals will decompose back to the reactants even at
298 K, so that although the addition is fast, the rate of forma-
tion of CH C(OH)CH is slow by comparison but very rea-
3
3
sonably faster than abstraction at 298 K. Addition to the C
atom in propanone [path (b)] is considerably less exothermic
and is likely to be much slower, although, particularly at low
pressures, the homolysis products ethanal and CH3 may be
formed in an overall process [path (c)] which is moderately
exothermic.
Finally, as with propanone, if H addition to the O atom is
important, rapid reaction of the adduct with O
2 will simply
regenerate the ketone and convert H into HO .
2
Analytical results
The main purpose of the study was to determine the kinetic
data discussed above under conditions where addition (traces)
The consequences of the introduction of the novel path (a)
merit a brief discussion. Given that k /k
2p 22e
mined accurately at 753 K and has a value of 1.07, and that
has been deter-
2
of the ketone would perturb the basic mechanism of the H
2
]
O
reaction to a negligible extent. However, a partial
*
H
present
\ *H
within experimental error, then for the
2
22e
22p
product analysis has been carried out to elucidate the main
features of the oxidation chemistry, which Ðrst permits precise
correction of the kinetic data through computer analysis, and
secondly adds considerably to the limited information avail-
able from only a small number of independent studies.
k
\ k \ 1.5 ] 106T 1.5 exp([3730/T ) dm3
22p
22e
mol~1 s~1 (ref. 23) is a reasonable assumption. With these
parameters, it may be argued that AmbridgeÏs result at 298 K
refers mainly to the addition process, with less than 5% con-
tribution from abstraction. In the absence of experimental
data, it is assumed that at 1È2 Torr (mostly argon), as used by
Ambridge, the addition process path (a) is in its third-order
region, as is certainly the case for H ] C H ] He.23 Use of
the kinetic data for this latter reaction over the range 300È800
K, with appropriate modiÐcations for path degeneracy and
the di†erent inert gas, and Ðtted to AmbridgeÏs rate constant
Propanone. Fig. 3 shows the product yields for the mixture
O \ 70, H \ 140, N \ 285, propanone \ 5 Torr. No
2
4
2 2 2
analysis was made for H O or H O. The major products
2
2
2
CH , CO, CH OH, HCHO, and ketene are all formed in
4
3
primary processes, although CO additionally and CO are
formed rapidly from intermediate products. Very small yields
2
at 298 K gives (k ) \ [Ar] ] 1.65 ] 1012exp([1700/T ) dm3
mol~1 s~1 for the addition of H atoms to propanone in path
o a
of CH CHO, and butanone are observed, apparently formed
3
(
a) in its third-order range. This expression gives (k ) \ 5.5
directly from propanone. The product distribution is not dis-
o a
]
106 dm3 mol~1 s~1 at 753 K and 1.5 Torr of argon and,
similar to that observed by Barnard and Honeyman29 who
used a mixture containing 66.7 and 33.3 Torr of propanone
using the kinetics of H ] C H ] M as a guide,23 a rate con-
2
4
stant of (3È10) ] 108 dm3 mol~1 s~1 at 500 Torr (with the
and O , respectively, at 771 K. The key features of the present
2
mix of gases used), compared with the experimental value of
product analysis are summarised below.
k
\ 2.34 ] 108 dm3 mol~1 s~1. However, at high tem-
(i) From previous studies30,31 of the formation of CH rad-
22p
3
peratures radicals such as CH C(OH)CH , other than revert-
icals in H ] O mixtures under identical conditions with
3
3
2
2
ing to the reactants, react uniquely with O to give carbonyl
those used here, it is known that 70È72% react with H to
2
compounds and HO radicals, so that in the present work
2
2
form CH , the others almost exclusively forming HCHO and
4
propanone will be regenerated.
CH OH. The present results with CH formed in a yield of
about 75% thus indicate that one CH radical is produced for
each propanone molecule consumed. Further, allowing for the
3
4
CH C(OH)CH ] O ] CH COCH ] HO
3
3
3
2
3
3
2
H ] C H ] C H ] H
(22e)
(22p)
reactivity of HCHO and the production of CO in secondary
processes, the initial yields are consistent also with the forma-
tion of about one CO molecule and one HCHO molecule
from each propanone molecule.
2
6
2 5 2
H ] CH COCH ] CH COCH ] H
3
3
3
2
2
Path (a) under the present conditions therefore provides the
Ðrst step in the efficient conversion of H atoms to HO rad-
icals, which under many combustion conditions will be an
(ii) Ketene is formed, apparently as a primary product, but
2
in yields considerably smaller than CH and CO.
4
1266
Phys. Chem. Chem. Phys., 2001, 3, 1262È1270

View Article Online
and second, there is considerable evidence to suggest that
the value of k is very low. Benson and OÏNeal32 have esti-
2
5
mated k \ 3.2 ] 1012 exp([20 000/T ) s~1 and recently
25
Chen and Bozelli33 have calculated k \ 1.51 ] 1024 T ~3.47
2
5
exp([21 720/T ) s~1 at one atmosphere pressure, which gives
k
\ 4.6 ] 101 s~1 at 753 K, a factor of 102È103 lower than
2
5
for CÈC homolysis in 1-propyl and 1-butyl radicals.
Computer analysis conÐrms that reaction (30) does not lead
to sufficient CO and CH in the early stages, and in addition
4
it is extremely likely that k is considerably lower than the
3
0
rate constant for the analogous 1,5p H atom transfer in neo-
pentylperoxy radicals.
(
CH ) CCH OO ] (CH ) C(CH )CH OOH
3
3
2
3 2
2
2
Owing to the additional strain energy in the transition state
caused by the presence of the C2O bond (an angle of almost
1
1
20¡ in the CH COCH radical), the activation energy for the
,5p H transfer will be raised by about 15 kJ mol~1 over that
3
2
for the neopentylperoxy radical. Additionally there is a path
degeneracy di†erence of a factor of 3, and further the value of
the equilibrium constant of R ] O H RO for CH COCH
will almost certainly be somewhat less than that for neopentyl,
again because of the electron-withdrawing power of the C2O
group. With the latter factor estimated as 3, the value of
2
2
3
2
Fig. 3 Variation of product yields with consumption of propanone.
O \ 70, H \ 140, N \ 285, propanone \ 5 Torr; 753 K. L, CO;
2
2
2
È, CH ; ], HCHO; ], CO ; …, CH CHO (]10); K, CH OH; |,
Kk
(\k ) \ 3.2 ] 105 dm3 mol~1 s~1 at 753 K, a factor
4
CH 2C2O.
2
2
3
3
1,5p
30
of almost exactly 100 lower than that for the neopentyl
system.34
With both reactions (25) and (30) relatively slow, the obser-
vation that butanone is formed in noticeable yields as a
primary product, and the fact that HO radicals reach a high
concentration in H ] O mixtures at 753 K, the importance
of the radicalÈradical process (31) was examined, particularly
as it leads directly to the formation of CH , CO and HCHO
as initial products. Computer analysis shows that at 5% con-
sumption of propanone, [HO ] \ 2.0 ] 10~3 Torr \
(
(
iii) CO is formed only in secondary and tertiary processes.
iv) The observation that butanone is formed in small
2
amounts implies that the CH COCH radical reaches a rela-
3
2
2
tively high concentration, given that the only plausible route
for butanone formation involves the radicalÈradical process,
CH COCH ] CH ] CH COCH CH .
2
2
3
2
3
3
2
3
4
Mechanism. Reactions (25)È(27) of the CH COCH radical
3
2
were suggested by Barnard and Honeyman29 to explain qual-
itatively their product yields.
2
4
.5 ] 10~8 mol~1 dm~3 for the mixture used. With the data
given above, and the reasonable assignment of k \ 1.0
CH COCH ] CH ] CH 2C2O
(25)
(26)
31
3
2
3
2
]
1010 dm3 mol~1 s~1, the rates of reactions (25), (30) and
CH 2C2O ] O ] 2CO ] H O
(
31) are directly comparable as k \ 4.6 ] 101, k [O ] \ 4.3
2
2
2
2
5
30
2
]
102 and k [HO ] \ 4.5 ] 102 s~1 at 753 K.
CH COCH ] O ]
3
1
2
3
2
2
CH COCH ] HO ] CH COCH OOH ]
CH COCH ] CH O ] CO ] HCHO
3
2
2
3
2
3
2
3
=
HCHO ] OH ] CH CO
OÈO
(27)
3
B
Reaction (26) must be an overall process and is most likely to
occur, inter alia, via reactions (28) and (29), both of which are
very fast with rate constants,23 k \ 1.0 ] 1010 dm3 mol~1
CH ] CO (31)
3
Homolysis of CH COCH [reaction (25)] is therefore a minor
3
2
2
8
component under the present conditions and, within the accu-
racy of the kinetic information used for the calculations, the
reactions of the radical with O and HO are of similar
importance. Computer analysis to give best Ðt to the experi-
mental data conÐrms these conclusions, and indicates that
about 80% of CH COCH radicals are removed in reaction
s~1 and k \ 1.8 ] 1010 exp([1700/T ) dm3 mol~1 s~1 over
29
the range 300È1500 K, so that ketene will be removed rela-
tively rapidly in secondary processes. Reaction (27) would
have to occur through a concerted step and, if the dominant
process, its occurrence could not explain the high yield of
2
2
CH . Current views suggest that reaction (30) is a more likely
3
2
4
(31) at the point of maximum rate when [HO ] is at its
reaction with the Ðrst step fully equilibrated, but it does not
2
highest. The optimum value of k \ 3.5 ] 105 dm3 mol~1
30
produce CH or CO as initial products.
3
s~1, so that use of K \ 5.3 ] 103 dm3 mol~1 obtained in
OH ] CH 2C2O ] CH OH ] CO
(28)
(29)
direct experiments by Pilling et al.35 for the neopentyl ] O
2
2
2
equilibrium, but reduced by a factor of 3 for the R ] O H
]
HCHO ] HCO
2
RO equilibrium in reaction (30) for the reasons discussed
2
H ] CH 2C2O ] CH ] CO
above, gives k \ 2.22 ] 102 s~1 for 1,5p H atom transfer in
2
3
CH COCH O radicals. Given the same A factor (per CÈH
CH COCH ] O H CH COCH ]
3 2 2
bond) as obtained for the 1,5p transfer in (CH ) CCH O rad-
2
3
2
2
3
=
3
3
2 2
OÈO
CH COCH ] CH 2C2O ] HCHO ] OH
icals (1.75 ] 1011 s~1),22 then E(1,5p) in reaction (30) is 135 kJ
mol~1. As outlined above the increase of 12 kJ mol~1 over
that22 (123 kJ mol~1) for the (CH )CCH O radicals may be
2
2
2
=
3
2 2
OOH
(30)
accounted for by the increased ring strain in the transition
state due to the presence of the C2O bond.
There are two prime reasons why reaction (25) is unacceptable
as the main process removing CH COCH in the present
3
2
work. First, computer analysis, using the values of k and k
Butanone and pentan-3-one. Fig. 4 shows the results of an
analysis of the products formed when 5 Torr of butanone and
pentan-3-one are added separately to the mixture containing
2
8
29
above, shows that the relative yields of CH , CO and HCHO
4
are seriously underpredicted in the early stages of reaction,
Phys. Chem. Chem. Phys., 2001, 3, 1262È1270
1267

View Article Online
tent with the observation that [CO]/[ketone consumed] D 1
in the early stages of reaction, which indicates that reaction
paths leading to fragmentation of the ketone are dominant.
(ii) Computer analysis using the kinetic data obtained
earlier gives the following % of radicals formed in the initial
stages from the two ketones by radical attack (mainly OH and
H) under the conditions used here. The values have been
rounded to the nearest 5%.
Ketone
Radical
Number %
Butanone
Butanone
Butanone
CH COCH CH
I
25
40
35
50
50
2
2
3
CH COCHCH
II
III
IV
V
3
3
CH COCH CH
3
2
2
Pentan-3-one CH CH COCHCH
3
2
3
Pentan-3-one CH CH COCH CH
3
2
2
2
The product distribution may be understood in terms of these
percentages and the thermochemistry of the possible reactions
of radicals IÈV, which is summarised in Table 4 for CÈC
homolysis and the formation of the unsaturated ketones.
Radical I cannot form an unsaturated ketone and its only
plausible homolysis path involves a high activation energy of
ca. 170 kJ mol~1, so that reactions (40) and (41) are likely to
compete favourably.
Fig. 4 Variation of product yields with consumption. O \ 70,
I ] O ] CH COCH CH ] CH COCH CH ]
2
2
2
2
3
2
2
2
H \ 140, N \ 285, Additive \ 5 Torr; 753 K. (a) Butanone; L,
=
=
2
2
CO; K, C H ; È, CH ; ], HCHO; …, CH CHO; |, CH 2C2O;
OÈO
OOH
2
4
4
3
2
]
C H ] CO ] HCHO ] OH (40)
]
, methyl vinyl ketone. (b) Pentan-3-one; L, CO; K, C H ; È, CH ;
2
4
4
2 4
]
, HCHO; …, CH CHO; ], ethyl vinyl ketone.
I ] HO ] HOOCH COCH CH ]
3
2
2
2
2
3
C H CO ] HCHO ] OH
5
B
7
0, 140 and 285 Torr of O , H and N , respectively. Again
2
2
2
CO ] C H (]O ) ] C H (41)
the results are consistent with those obtained by Akbar and
Barnard for butanone at 723 K36 and pentan-3-one at 673
K,26 and with those of Hoare and Li37 at 673 K, although the
high yields of methanol reported by Barnard and
Honeyman29 for butanone are not observed in the other two
investigations. The limited nature of the analyses in the
present work precludes a full interpretion of the formation of
the products, but a number of key mechanistic features may
be elucidated.
2 5 2 4
2
Both reactions are favourable because they lead directly to
major initial products detected.
Radical II cannot undergo a facile CÈC homolysis and the
formation of methyl vinyl ketone is signiÐcantly less favour-
able energetically (14 kJ mol~1) than that of but-1-ene from
2-butyl radicals. Based on the reasonably well established22
variation of the rate constant for the formation of conjugate
alkenes from alkyl radicals with the enthalpy of reaction,
(
i) CO, C H , CH , HCHO and CH CHO are formed as
k
D 7.0 ] 105 dm3 mol~1 s~1 at 753 K, so that with [HO ]
2
4
4
3
33
2
major initial products, together with small amounts of
CH OH and the relevant vinyl ketone. The small yields of the
vinyl ketones, although perhaps unexpected in view of the
large yields of conjugate alkene formed from structurally-
related alkyl radicals under similar conditions, is fully consis-
controlled by the H ] O mixture and the reasonable
assumption that reactions (42) and (43) have the same rate
constants as the analogous reactions of the CH COCH rad-
icals (see earlier), then about 50% of the II radicals will lead to
methyl vinyl ketone and about 50% to CH CHO, CO and
2
2
3
3
2
3
Table 4 Thermochemistrya of butanone, pentan-3-one and related systems
Reaction
number
Reactants
Products
*H/kJ mol~1
CH COC H (I)
CH 2C2O ] C H
(32)
(33)
(33a)
(34)
(34a)
(35)
(35a)
(36)
(37)
(37a)
(38)
(38a)
(39)
]99
[31
[45
]77
]94
[48
[58
]105
[32
[47
]65
]88
[48
[59
2
2
5
2
2 5
CH COCHCH (II) ] O
CH COCH2CH ] HO
3
3
2
3
2
2
CH CH CHCH ] O
CH CH CH2CH ] HO
3
2
3
2
3
2
2
2
CH COCH CH (III)
CH CO ] C H
3
2
2
3
2 4
CH CH CH CH
C H ] C H
3
2
2
2
2
5
2 4
CH COCH CH ] O
CH COCH2CH ] HO
3
2
2
2
3
2
2
CH CH CH CH ] O
CH CH CH2CH ] HO
3
2
2
2
2
3
2
2
2
C H COCHCH (IV)
C H ] CH CH2C2O
2
5
3
2
5
3
C H COCHCH ] O
C H COCH2CH ] HO
2
5
3
2
2
5
2
2
CH CH CH CHCH ] O
C H CH2CH ] HO
3
2
2
3
2
3
7
2
2
C H COCH CH (V)
C H CO ] C H
2
5
2
2
2
5
2 4
CH CH CH CH CH
CH CH CH ] C H
3
2
2
2
2
3
2
2
2 4
C H COCH CH ] O
C H COCH2CH ] HO
2
5
2
2
2
2
5
2
2
CH CH CH CH CH ] O
C H CH2CH ] HO
(39a)
3
2
2
2
2
2
3
7
2
2
a Based on the Sandia Chemkin Thermodynamic Data, R. J. Kee, F. M. Ripley and J. M. Miller, Sandia Report 87È82153, UC-4, 1987.
1268
Phys. Chem. Chem. Phys., 2001, 3, 1262È1270

View Article Online
Table 5 Observed and calculated initial yieldsa of products from butanone and pentan-3-one
Butanone
Pentan-3-one
Calculated
Product
Calculated
Observed
Observed
CO
70
60
20
20
35
35
10
110
63
17
14
28
39
D5
125
125
50
5 (max.)
7
18
È
140
107
47
6
9
17
È
C H
2
4
CH CHO
3
RCOCH2CH
HCHO
CH
CH 2C2O
2
4
2
a % mole product/mole ketone consumed.
IV ] O ] CH CH(OO)COCH CH ]
CH , with smaller amounts of HCHO, CH OH and
2
3
2
3
4
3
CH 2C2O.
CH CH(OOH)COCHCH
2
3
3
B O
II ] O ] CH CHCOCH ] CH CHCOCH ]
2
2
3
3
3
2
CH CO ] 2OH ] CO ] CH CHO (45)
IV ] HO ] CH CH(OOH)COCH CH ]
=
=
3
3
OÈO
OOH
2
3
2
3
CH CHO ] CH 2C2O ] OH (42)
C H CO ] CH CHO ] OH
3
2
2
5
3
II ] HO ] CH CH(OOH)COCH ]
B
2
3
3
CO ] C H (]O ) ] C H
(46)
CH CO ] OH ] CH CHO
2
5
2
2 4
3
3
B
Reactions (44) and (45) involve 1,6p (primary) and 1,5s
(secondary) H atom transfers, respectively. In reaction (45), a
CO ] CH ] CH , HCHO, CH OH (43)
3
4
3
Radical III has a relatively facile CÈC homolysis reaction, (34)
compared with (34a), where the di†erence in energy barriers is
probably about 50% of the di†erence in the values of *H, and
a less exothermic formation of the unsaturated species, reac-
tions (35) and (35a). Given that 99% of C H radicals react
second O addition and H atom transfer is envisaged to facili-
2
tate the breakdown of the CH CH(OOH)COCHCH radical.
3
3
On the basis of H atom transfer in alkylperoxy radicals,22
k
/k \ 3.5 at 753 K, and further k(1,6p)/k(1,5p) \ 3. As H
4
5 44
atom transfer in IV is therefore signiÐcantly faster than in II, it
2
5
with O to give C H under the present conditions22 so that
2 4
the initial value of [C H ]/[but-1-ene] \ 2k /k [O ] \
34a 35a
.5 obtained at 753 K for the mixture used here with butane
follows that reaction of IV with HO will be relatively unim-
2
2
portant, and further that the formation of the vinyl ketone will
2
4
2
3
be slower. Accepting that there is uncertainty in assessing the
e†ect of the additional strain in the ring transition states of H
atom transfer due to the presence of the C2O group, it is
simply concluded from the data that a maximum of 10% of
the IV radicals formed react through reaction (37) to give the
vinyl ketone, and about 20 and 70% through reactions (44)
and (45), respectively.
The dominant feature concerning radical V is the energeti-
cally favourable homolysis reaction (38). Repetition of the
arguments used for radical III and a recognition22 that k(1,7p)
/k(1,6p) D k(1,6s)/k(1,5s) D 3 reveals that H atom transfer is
as additive,30 and with k /k increased over k /k
by a
3
4
35
34a 35a
factor of about 30 as a result of the change in thermochemis-
try, then [C H ]/[CH COCH \ CH ] \ 53 is predicted
2
4
3
2
H atom transfer in
for radical III. Moreover,
CH CH(OO)COCH radicals and their reaction with HO
radicals will not compete e†ectively with homolysis [reaction
34)], which can be regarded e†ectively as the sole reaction of
3
3
2
(
III radicals here and the main reaction in the medium tem-
perature range unless the pressure of O exceeds 1 atm.
2
The fate of radical IV from pentan-3-one may usefully be
compared with that of radical II from butanone. Radical IV
has a very slow CÈC homolysis [reaction (36)] and the vinyl
ketone-forming reaction (37) is considerably less favourable
than the analogous reaction of 2-pentyl radicals. The alterna-
tive reactions (44)È(46) will therefore compete with reaction
not a major route for V, and given that [HO ] is determined
2
solely by the H ] O mixture, it may be concluded that
2
2
homolysis through reaction (38) is e†ectively the sole reaction
of V radicals under the present conditions.
Table 5 shows a comparison of the observed and calculated
initial yields (at ca. 10% reaction) of the major products,
based on a self-consistent set of kinetic data for radicals IÈV
and for the analogous alkyl and alkylperoxy systems. No
attempt is made to allow for consumption of the products or
for their formation in secondary processes. For butanone, the
yield of CO is underpredicted, but rapid oxidation of HCHO,
(
37).
IV ] O ] CH CH(OO)COCH CH ]
2
3
2
3
CH CH(OOH)COCH CH
3
2
2
B
CH CHO ] CO ] C H ] OH (44)
3
2 4
Table 6 a Summary of recommended kinetic data
Reaction
Type
T range/K
A
n
E
OH ] CH COCH
Abstraction
Abstraction
Abstraction
Abstraction
1,5p
300È1200
440È800
440È800
300È1200
700È800
2.75 ] 105
2.2 ] 1010
2.7 ] 1010
7.5 ] 105
5.3 ] 1011
1.5
0
0
1.5
0
7.7
3
3
OH ] CH COC H
12.2
12.2
26.0
3
2 5
OH ] C H COC H
2
5
2 5
H ] CH COCH
3
3
CH COCH O ] CH COCH O H
136
3
2
2
2
2 2
transfer
H ] CH COC H
Abstraction
Abstraction
753
753
k \ 7.2 ] 108
k \ 1.20 ] 109
3
2 5
H ] C H COC H
2
5
2 5
a Based on k \ AT ne~E@RT, dm3 mol~1 s~1 or s~1 units as appropriate.
Phys. Chem. Chem. Phys., 2001, 3, 1262È1270
1269

View Article Online
1
2
3
R. R. Baldwin, G. R. Drewry and R. W. Walker, J. Chem. Soc.,
Faraday T rans. 2, 1986, 82, 251.
D. L. Baulch, C. J. Cobos, R. A. Cox, C. Esser, P. Frank, Th. Just,
J. A. Kerr, M. J. Pilling, J. Troe, R. W. Walker and J. Warnatz, J.
Phys. Chem. Ref. Data, 1992, 21, 411.
CH CHO and CH 2C2O will provide a ready source when
included in the mechanism. The yields of the other products
3
2
1
are well predicted, although that for CH COCH2CH may be
3
2
too high by a factor of up to about 50%. With pentan-3-one,
the very much lower yields of HCHO and unsaturated ketone
are satisfactorily predicted, together with the signiÐcantly
higher yields of CO, CH CHO and C H .
14 H. J. Curran, W. J. Pitz, C. K. Westbrook, M. W. M. Hisham and
R. W. Walker, 26th Symp. (Int.) on Combustion, The Combustion
Institute, Pittsburgh, PA, 1996, p. 641.
1
1
1
1
3
2 4
5
6
7
8
E. Ranzi, T. Faravelli, P. Ga†uri and A. Sogaro, Combust. Flame,
1
995, 102, 179.
R. R. Baldwin, D. E. Hopkins and R. W. Walker, T rans. Faraday
Soc., 1970, 66, 189.
Summary and conclusions
R. R. Baldwin and R. W. Walker, J. Chem. Soc., Faraday T rans.
2
, 1979, 75, 140.
Kinetic data have been obtained for the reaction of H and
OH with propanone, butanone, and pentan-3-one at 753 K,
and where possible combined with independent results to
obtain Arrhenius parameters. A case is presented for the
occurrence of a reaction which involves the addition of H
atoms to the O atom in ketones.
A partial product analysis was carried out for the three
ketones and a set of self-consistent kinetic data used to
explain the product distributions. The mechanism of oxida-
tion is essentially that for alkanes, with the relative impor-
tance of individual reactions modiÐed by di†erences in
thermochemical characteristics. In particular, the low yields of
unsaturated ketones and the very high yields of CO in the
initial products, observed both here and by independent
workers, have been fully explained.
J. W. Sutherland, J. V. Michael, A. N. Pirraglia, F. L. Nesbitt and
R. B. Klemm, 21st Symp. (Int.) on Combustion, The Combustion
Institute, Pittsburgh, PA, 1988, p. 929.
1
9
J. M. Roscoe, Can. J. Chem., 1986, 64, 1458.
20 J. T. Herron, J. Phys. Chem. Ref. Data, 1988, 17, 967.
21 J. H. Lee and R. B. Timmons, Int. J. Chem. Kinet., 1977, 9, 133.
2
2
R. W. Walker and C. Morley, in Comprehensive Chemical
Kinetics, ed. R. G. Compton and G. Hancock, Elsevier, Amster-
dam, 1997, vol. 35, p. 1.
D. L. Baulch, C. J. Cobos, R. A. Cox, P. Frank, G. Hayman, Th.
Just, J. A. Kerr. T. Murrells, M. J. Pilling, J. Troe, R. W. Walker
and J. Warnatz, J. Phys. Chem. Ref. Data, 1994, 23, 847.
2
3
24 K. D. King, D. M. Golden and S. W. Benson, J. Am. Chem. Soc.,
970, 92, 5541.
2
1
5
Handbook of Bimolecular and T ermolecular Gas Reactions, ed.
J. A. Kerr and S. J. Moss, CRC Press, Boca Raton, FL, 1981,
vol. I, p. 186.
Table 6 gives a summary of the kinetic data determined
from this study.
2
6
M. Akbar and J. A. Barnard, T rans. Faraday Soc., 1968, 64, 3049.
27 G. M. Harris and E. W. R. Steacie, J. Chem. Phys., 1945, 13, 554.
2
2
3
8
9
0
R. K. Solly, D. M Golden and S. W. Benson, Int. J. Chem. Kinet.,
1
970, 2, 381.
J. A. Barnard and T. W. Honeyman, Proc. R. Soc. L ondon, Ser. A,
References
1964, 279, 236.
R. R. Baker, R. R. Baldwin and R. W. Walker, J. Chem. Soc.,
Faraday T rans. I, 1975, 71, 756.
1
2
3
R. Atkinson, Int. J. Chem. Kinet., 1987, 19, 799.
R. W. Walker, Int. J. Chem. Kinet., 1985, 17, 573.
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.
T. J. Wallington and M. J. Kurylo, J. Phys. Chem., 1987, 91, 5050.
S. Le Calve, D. Hitier, G. Le Bras and A. Mellouki, J. Phys.
Chem., 1998, 102, 4597.
31 R. R. Baker, R. R. Baldwin, C. J. Everett and R. W. Walker,
Combust. Flame, 1975, 25, 285.
32 S. W. Benson and H. E. OÏNeal, Kinetic Data on Gas Phase
Unimolecular Reactions, NSRDSÈNBS 21, Washington, DC,
1970.
33 C.-J. Chen and J. Bozelli, J. Phys. Chem., 1999, 103, 9731.
34 R. R. Baldwin and M. W. M. Hisham, J. Chem. Soc., Faraday
T rans. 1, 1982, 78, 1615.
35 K. J. Hughes, P. Halford-Maw, P. D. Lightfoot, T. Turanyi and
M. J. Pilling, 24th Symp. (Int.) on Combustion, The Combustion
Institute, Pittsburgh, PA, 1992, p. 645.
4
5
6
M. Wollenhaupt, S. A. Carl, A. Horowitz and J. N. Crowley, J.
Phys. Chem., 2000, 104, 2695.
7
8
J. F. Bott and N. Cohen, Int. J. Chem. Kinet., 1991, 23, 1017.
P. F. Ambridge, J. N. Bradley and D. A. Whytock, J. Chem. Soc.,
Faraday T rans. 1, 1976, 72, 1870.
9
0
1
N. Cohen, Int. J. Chem. Kinet., 1991, 23, 683.
36 M. A. Akbar and J. A. Barnard, T rans. Faraday Soc., 1968, 64,
1
1
N. Cohen, Int. J. Chem. Kinet., 1983, 15, 503.
3035.
N. Cohen, Int. J. Chem. Kinet., 1991, 23, 397.
37 D. E. Hoare and T.-M. Li, Combust Flame, 1968, 12, 145.
1270
Phys. Chem. Chem. Phys., 2001, 3, 1262È1270