A kinetic study of the reactions of NO3 with methyl vinyl ketone, methacrolein, acrolein, methyl acrylate and methyl methacrylate

Article abstract of DOI:10.1039/a904613e

Rate coefficients for the reactions NO₃ + CH₃C(O)CHCH₂ → products(1), NO₃ + CH₂C(CH₃)CHO → products(2), NO₃ + CH₂CHCHO → products, and NO₃ + CH₂CHC(O)OCH₃ → products were obtained by relative and absolute methods. The rate coefficient for the reaction NO₃ + CH₂C(CH₃)C(O)OCH₃ → products was determined by the relative-rate method only. Relative rate measurements agreed with results of previous studies. However, in some cases, the kinetic data obtained using the absolute method were significantly higher than those from the relative technique which may be due to secondary chemistry and reactive impurities. Product studies revealed that methyl glyoxal is a product of reactions (1) and (2) along with peroxymethacryloyl nitrate for (2) in air. A diurnally varying boundary-layer model suggests that (2) is an important loss process for methacrolein, which can lead to OH generation at night. NO₃ may be an important oxidant for methacrolein in the urban boundary layer and where urban plumes impact on forested areas.

Full text of DOI:10.1039/a904613e

View Article Online / Journal Homepage / Table of Contents for this issue
A kinetic study of the reactions of NO with methyl vinyl ketone,
3
methacrolein, acrolein, methyl acrylate and methyl methacrylate
Carlos E. Canosa-Mas,a Sinead Carr,a Martin D. King,a Dudley E. Shallcross,b
Katherine C. Thompsona and Richard P. Waynea*
a Physical and T heoretical Chemistry Department, South Parks Road, Oxford, UK OX1 3QZ
b Centre for Atmospheric Science, Department of Chemistry, University of Cambridge,
L ensÐeld Road, Cambridge, UK CB2 1EW
Received 9th June 1999, Accepted 12th July 1999
Absolute and relative-rate techniques have been used to obtain rate coefficients for the reactions:
NO ] CH C(O)CHCH ] products (1), NO ] CH C(CH )CHO ] products (2), NO ] CH CHCHO ]
3
3
2
3
2
3
3
2
products (3), and NO ] CH CHC(O)OCH ] products (4). The reaction NO ] CH C(CH )C(O)OCH ]
3
2
3
3
2
3
3
products (5), has been investigated by a relative-rate method only. The rate coefficients obtained by the
relative-rate method at T \ 296 ^ 2 K and P \ 760 Torr are k \ (4.7 ^ 1.7) ] 10~16 cm3 molecule~1 s~1,
1
k \ (3.7 ^ 1.0) ] 10~15 cm3 molecule~1 s~1, k \ (1.1 ^ 0.4) ] 10~15 cm3 molecule~1 s~1, k \ (1.0 ^ 0.6)
2
3
4
] 10~16 cm3 molecule~1 s~1 and k \ (3.6 ^ 1.3) ] 10~15 cm3 molecule~1 s~1. The rate coefficients
5
determined by the discharge-Ñow technique at low pressure (P \ 1È10 Torr) and at T \ 293È303 K are
k \ (3.2 ^ 0.6) ] 10~16 cm3 molecule~1 s~1, k \ (9.6 ^ 2.0) ] 10~15 cm3 molecule~1 s~1, k \ (8.9 ^ 2.8)
1
2
3
] 10~15 cm3 molecule~1 s~1, k \ (1.9 ^ 0.4) ] 10~16 cm3 molecule~1 s~1. The discrepancy between the
4
values obtained from the relative-rate technique and the absolute technique are discussed and explained in
terms of interference in the absolute study caused by secondary chemistry and fast-reacting impurities. Product
studies reveal that methyl glyoxal is a product of reactions (1) and (2) along with peroxymethacryloyl nitrate
(MPAN) for reaction (2) in air. A diurnally varying boundary-layer model suggests that reaction (2) is an
important loss process for methacrolein and that it can lead to the generation of OH at night.
In addition, the reaction between NO and three other a,b-
Introduction
3
unsaturated-carbonyl compounds (acrolein, CH CHCHO;
2
Vegetation provides a major source of reactive carbon enter-
ing the EarthÏs atmosphere.1 Estimates for the total emissions
of biogenic species range from 430 to 830 Tg yr~1.1 Isoprene
(C H ) is a hemiterpene emitted by many plants and trees;1
methyl acrylate, CH CHC(O)OCH ; and methyl meth-
2
3
acrylate, CH C(CH )C(O)OCH ) have also been investigated:
2
3
3
NO ] CH CHCHO ] products
(3)
(4)
(5)
3
2
5
8
estimated total emissions range from 250 to 450 Tg yr~1.
Isoprene therefore constitutes about half the total biogenic
emissions, and the higher end of the estimated range of emis-
sions is comparable with the emission of methane.2 Emissions
of isoprene are known to fall to zero at night and, since iso-
prene reacts extremely quickly with the major daytime
oxidant, the OH radical, its night-time concentration is
usually small or zero.1 It is now widely recognised that the
NO ] CH CHC(O)OCH ] products
3
2
3
NO ] CH C(CH )C(O)OCH ] products
3
2
3
3
Rate coefficients have been determined using the relative-rate
technique for reactions (1)È(5), and also by the absolute
discharge-Ñow method for reactions (1)È(4). Studies of the
reaction products were carried out for processes (1) and (2).
The atmospheric implications of these reactions are discussed
in the context of a diurnally varying planetary boundary layer
model.5
nitrate radical (NO ) could be the dominant oxidant for a
3
number of species, particularly unsaturated ones.3 Although
NO reacts rapidly with isoprene, the nitrate radical is rapidly
3
photolysed during the day and can only build up to signiÐcant
Experimental
concentrations during the night, so that the two species are in
anti-phase. During the day, the OH-initiated oxidation of iso-
prene yields methyl vinyl ketone (CH C(O)CHCH ) and
methacrolein (CH C(CH )CHO) as the major primary pro-
ducts.4 Both methyl vinyl ketone and methacrolein are suffi-
ciently long lived that they will persist into the night and can
Relative-rate experiments
3
2
The relative-rate technique was used to study reactions (1)È(5).
The experiments were carried out at atmospheric pressure
using a 56 l PTFE-lined bag housed in an insulating, dark
cabinet. The decay of the substrate and a reference compound
2
3
therefore interact with NO . The purpose of the work report-
ed in this paper was to study the kinetics of the reactions of
3
was monitored in the presence of NO . The relative concen-
3
trations of the substrate and reference compounds were deter-
NO with methyl vinyl ketone and methacrolein:
3
mined by gas chromatography with Ñame-ionisation detection
(Perkin Elmer, GC 8500 series). The reaction mixture was
sampled by expanding it into an evacuated 1 cm3 stainless-
steel sample loop and injecting it onto a column packed with
NO ] CH C(O)CHCH ] products
(1)
(2)
3
3
2
NO ] CH C(CH )CHO ] products
3
2
3
Phys. Chem. Chem. Phys., 1999, 1, 4195È4202
4195

View Article Online
Table 1 Summary of rate coefficients obtained from the relative rate experimentsa
Substrate
Reference
Ratio
k /cm3 molecule~1 s~1
s
Methyl vinyl ketone
Methacrolein
Acrolein
Methyl acrylate
Methyl methacrylate
Ethene
Propene
Ethene
Ethene
Propene
2.53 ^ 0.59
0.40 ^ 0.05
6.09 ^ 1.30
0.56 ^ 0.25
0.39 ^ 0.06
(4.7 ^ 1.7) ] 10~16
(3.7 ^ 1.0) ] 10~15
(1.1 ^ 0.4) ] 10~15
(1.0 ^ 0.6) ] 10~16
(3.6 ^ 1.3) ] 10~15
a The rate coefficients used for the reactions of NO with propene and ethene were (9.3 ^ 1.9) ] 10~15 and (1.85 ^ 0.24) ] 10~15 cm3 molecule~1
3
s~1, respectively.
Chromosorb 102 (Alltech) by means of a 10-port gas-sampling
valve (Alltech). The temperature of the column was modiÐed
for each mixture to give optimum peak separation. The initial
concentrations of substrate and reference compound were
varied with respect to each other between limits of 4.0 ] 1015
and 1.0 ] 1016 molecule cm~3, to test for interference from
secondary reactions. The organic reagents were added simul-
taneously to the bag and allowed to mix for at least one hour.
The thermal decomposition of N O was used as the source of
Materials
Acrolein (90%, containing O10% H O and cyclic dimer, with
hydroquinone stabiliser), pentane (99]%), methacrolein (95%,
2
inhibited with 0.1% hydroquinone), methyl vinyl ketone (99%,
inhibited with 0.1% acetic acid and 0.05% hydroquinone),
methyl methacrylate (99%, 10È100 ppm hydroquinone mono-
ethyl ether stabiliser) and propene (99]%) were supplied by
Aldrich. Ethene (99.8%), methane (99.5%) and methyl acrylate
(99%, 0.02% 4-methoxy-phenol stabiliser) were supplied by
BDH. Volatile impurities were removed from the organic
reagents by repeated freezeÈpumpÈthaw cycles and for the
discharge-Ñow experiments, trap-to-trap distillation was also
performed Nitrogen (99%, BOC) and oxygen (99%, BOC)
were passed through a phosphorus pentoxide trap to remove
water. N O was prepared in advance and stored at liquid
2
5
NO radicals. Five or six additions of N O were made to the
3
2 5
mixture of organic reagents in the bag, a chromatogram of the
mixture being recorded after each addition. An internal stan-
dard was used to allow for dilution due to the N O addi-
2
5
tions. The internal standard used for experiments with
methacrolein was n-pentane, whereas methane was used as the
internal standard for all other substrates studied. The rate
2
5
nitrogen temperatures until used. The preparation consisted of
allowing NO to react with O . A 3% mixture of O was
constants for the reaction of NO with pentane and methane
3
are 8.7 ] 10~17 cm3 molecule~1 s~1 6 and O3.9 ] 10~19 cm3
2
3 3
made by passing
(Argentox). NO (BDH 99%) was trapped at 196 K, any vola-
O
through a conventional ozoniser
molecule~1 s~1,7 respectively, so that little or no reaction
would be expected. All relative-rate experiments were per-
formed at T \ 296 ^ 2 K using nitrogen as the bath gas.
2
2
tile impurities were pumped o†, and the NO was allowed to
2
warm up in a stream of the O mixture. After passing through
3
a 3 l mixing volume, the N O formed and any remaining
2
5
NO was trapped in a second trap at 196 K. The process was
Discharge-Ñow (absolute) experiments
2
repeated until N O (determined by FTIR spectroscopy) con-
2
5
tained only minor amounts of NO and HNO . Nitric acid
A conventional discharge-Ñow system8 was used to determine
absolute rate constants for reactions (1)È(4). The decay of
2
3
(BDH, Anala-R}) was dried before use by preparing a 1 : 2
mixture (by volume) of nitric and sulfuric acid (BDH,
Aristar}) at ca. 258 K. Helium was passed through an
“OxisorbÏ cartridge (Messer Griesheim) to remove oxygen, a
column containing phosphorus pentoxide to remove water,
and two molecular sieves (BDH 13X and 4A) held at 77 K, to
NO as a function of time was monitored using optical
3
absorption at j \ 662 nm. The absolute concentration8 of
NO was found by titrating NO with NO in the reaction
3
3
NO ] NO ] 2NO
3
2
remove hydrocarbons, water and CO respectively. Fluorine
(5% in helium, BOC) was used as supplied.
(k
\ 2.6 ] 10~11 cm3 molecule~1 s~1)9
(6)
2
298 K
An e†ective absorption cross section p
] 10~17 cm2 molecule~1 was determined. Nitrate radicals
were generated by the reaction of F atoms with anhydrous
nitric acid
\ (1.5 ^ 0.2)
662 nm
Results
Relative-rate studies
F ] HNO ] HF ] NO
If it is assumed that the loss of the substrate and reference
3
3
compound are due solely to reaction with NO , then it can be
(k
\ 2.3 ] 10~11 cm3 molecule~1 s~1)9
(7)
3
298 K
shown that:
The minimum detectable [NO ] was found to be ca. 5 ] 1011
molecule cm~3, for a signal-to-noise ratio of unity and a 10 s
3
[substrate]
t/0
k
k
A
[reference]
t/0
B
s
r
ln
[ D \
ln
[ D
(I)
t
t
integration time. Nitrogen dioxide, a possible product of reac-
tions (1)È(4), was monitored by laser-induced Ñuorescence
(LIF), the excitation being provided by an argon-ion laser
(Coherent Innova 70) operating at j \ 488 nm with a power
of ca. 0.6 W. The LIF signal was calibrated daily by admitting
[substrate]
[reference]
t
t
where k is the rate coefficient for the reaction of NO with the
s
3
substrate, k is the rate coefficient for the reaction of NO with
r
3
the reference compound and D is the dilution term, given by
t
the equation:
a known amount of NO . The signal varied linearly with
2
[NO ]. The detection limit of the system was 5 ] 109
A
[I]
t/0
B
2
D \ ln
(II)
molecule cm~3 in the absence of any quenching species other
t
[I]
t
than He, for a signal-to-noise ratio of unity and a time con-
stant of 5 s. The experiments were performed at low pressures
(1È10 Torr) and at T \ 293È303 K. Pseudo-Ðrst order condi-
tions were employed for all kinetic studies, the organic species
where I represents an internal standard, a compound which
does not react with NO on the time-scale and conditions of
3
the experiment so that it can be used to allow for the dilution
being in at least a ten-fold excess over NO . The carrier gas
used was helium.
that occurs when the NO radical precursor, N O , is added
to the reaction vessel.
3
3
2 5
4196
Phys. Chem. Chem. Phys., 1999, 1, 4195È4202

View Article Online
Fig. 1 Relative-rate plots for the reaction of the NO radical with
Fig. 3 First-order plots obtained in Ñow-tube experiments on the
3
methyl vinyl ketone, acrolein and methyl acrylate (reference com-
reactions of NO with methyl vinyl ketone and methyl acrylate. Data
3
pound: ethene). = \ Acrolein, … \ methyl vinyl ketone, K \ methyl
for methyl vinyl ketone are displaced by 0.2 units on the ordinate.
acrylate.
… \ Methyl vinyl ketone, K \ methyl acrylate.
Products observed at atmospheric pressure
The data plotted in the form of eqn. (I) are displayed in Fig.
The products of reactions (1) and (2) were investigated; the
experimental conditions were exactly the same as for the
1 and 2, and the values of k /k are summarised in Table 1. A
s
r
weighted least-squares regression, which takes into account
the errors in the measured concentrations of both the sub-
strate and the reference compounds, was used to obtain the
gradient of the best straight line.10 The errors in the measured
concentrations were estimated by repeatedly sampling the
same mixture. The errors given for the values of k Èk deter-
relative-rate experiments, except that O was used as the bath
2
gas instead of N . Two distinct gas-chromatographic product
2
peaks were observed in each case, the retention times of the
two peaks being the same when either methyl vinyl ketone or
methacrolein was the starting compound. The retention time
of one of the peaks is identical to that of methyl glyoxal on
two di†erent columns (Alltech, Chromosorb 102 and Porapak
Q). Attempts were made to identify the second product, but
all proved unsuccessful. An FTIR-spectroscopic investigation
of the products of reactions (1) and (2) revealed the presence of
CO (presumably arising from decomposition of methyl
glyoxal) along with MPAN (peroxymethacryloyl nitrate) for
reaction (2) only.13 No other products could be unam-
biguously identiÐed in this investigation.
1
5
mined by the relative-rate technique incorporate both the
95% conÐdence limits from the slopes and the quoted error in
the value of the rate coefficient for the reference compound.
The values of the rate constants used for reactions of
NO with the reference compounds propene and ethene were
3
taken as (9.3 ^ 1.9) ] 10~15 and (1.85 ] 0.24) ] 10~16 cm3
molecule~1 s~1, respectively.11,12
Fig. 2 Relative-rate plots for the reaction of the NO radical with
3
methacrolein and methyl methacrylate (reference compound:
Fig. 4 First-order plots obtained in Ñow-tube experiments on the
propene). Data for methyl methacrylate are displaced by 0.2 units on
the ordinate. > \ Methyl methacrylate, L \ methacrolein.
reactions of NO with methacrolein and acrolein. L \ Methacrolein,
3
= \ acrolein.
Phys. Chem. Chem. Phys., 1999, 1, 4195È4202
4197

View Article Online
Fig. 5 Second-order plot obtained for reaction of NO with methyl
Fig. 7 Second-order plot obtained for reaction of NO with acrolein.
3
3
vinyl ketone.
each substrate are shown in Fig. 3 and 4. Second-order rate
coefficients were obtained from the slope of a plot of k@ as a
function of concentration of organic reagent. The second-
order plots are shown in Fig. 5È8. Table 2 summarises the rate
coefficient data obtained, the errors quoted being the 95%
Discharge-Ñow (absolute) studies
Pseudo-Ðrst order rate constants, k@, for reactions (1)È(4) were
obtained from the gradients of lnM[NO ] /[NO ] N plotted
as a function of time. Typical Ðrst-order plots obtained for
3 t/0
3 t
Fig. 6 Second-order plot obtained for reaction of NO with meth-
Fig. 8 Second-order plot obtained for reaction of NO with methyl
3
3
acrolein.
acrylate.
Table 2 Summary of rate coefficients obtained from the discharge-Ñow experiments
Pressure/
Torr
Substrate
Temperature/K
k/cm3 molecule~1 s~1
Methyl vinyl ketone
Methacrolein
Acrolein
6.6 ^ 0.6
2.5 ^ 0.3
2.4 ^ 0.1
9.0 ^ 0.4
296 ^ 1
300 ^ 7
293 ^ 2
295 ^ 2
(3.2 ^ 0.6) ] 10~16
(9.6 ^ 2.0) ] 10~15
(8.9 ^ 2.8) ] 10~15
(1.9 ^ 0.4) ] 10~16
Methyl acrylate
Table 3 Values of *[NO ]/ [ *[NO ] obtained from the discharge-Ñow system (all experiments were performed at 297 ^ 1 K)
2
3
Substrate
[Substrate]/[NO ]
Pressure/Torr
*[NO ]/ [ *[NO ]
3
2
3
Methyl vinyl ketone
Methacrolein
Acrolein
10È50
100È1000
10È50
1.4
0.45 ^ 0.06
0.95 ^ 0.30
0.54 ^ 0.02
2.3 ^ 0.3
2.5
4198
Phys. Chem. Chem. Phys., 1999, 1, 4195È4202

View Article Online
conÐdence limits of the slopes. In an attempt to determine the
overall reaction stoicheiometry, we measured, in a separate
channel) and methyl glyoxal (addition channel); a value of
*[NO ]/[*[NO ] of near to unity thus seems to be some-
2
3
series of experiments, the yield of NO as well as the loss of
what high for this process. We conclude that secondary reac-
tions are taking place and that the concentrations of organic
reagent used are insufficiently in excess to suppress the reac-
2
NO (see later for explanation). Table 3 gives the values of the
3
ratio of [NO ] observed to [NO ] consumed.
2
3
tion of organic radicals with NO . For reaction (4), with
3
methyl acrylate, we suspect that NO is not observed as a
2
Discussion
direct result of the higher Ñow tube pressures employed, which
would not only cause quenching of the NO Ñuorescence, but
Products and the yield of NO
2
2
also favour the reaction of NO with NO .
2
3
Kwok et al.14 have suggested that the abstraction pathway
dominates for the reactions of NO with methacrolein and
3
acrolein, analogous to the behaviour of OH.15 Methyl
Comparison of rate coefficients measured by the two techniques
glyoxal, CH C(O)CHO, was observed as a product in our
3
Rate coefficients measured by relative-rate and Ñow-tube
methods sometimes appear to be di†erent. That turns out to
be the case in our study with some, but not all, of the com-
pounds. We shall show that we believe the relative-rate
method to give a more reliable (or, at least, more directly
applicable) rate constant. However, we present our data for
both methods, because such di†erences as we Ðnd shed light
on the behaviour of the di†erent compounds. There are
several potential reasons for errors arising in Ñow-tube experi-
ments: (i) fast reactions of the reactant present in lower con-
centration with products of the primary process; (ii) the
reaction of this reactant with impurities present in the excess
reactant; and (iii) an increase in the reactivity of the wall upon
addition of the excess reactant. We have employed a numeri-
cal model of likely secondary chemistry in an attempt to inter-
pret our observations, but wish to emphasize that our
modelling activities were not directed at obtaining a rate con-
stant for the primary reaction from the Ñow-tube data, but
rather at understanding better the chemistry that occurs in the
Ñow tube.
studies for both reactions (1) and (2). It can be shown that, in
order to form methyl glyoxal from methacrolein, addition of
NO must occur via a mechanism such as
3
CH C(CH )CHO ] NO ] O ]
2
3
3
2
CH (ONO )COO(CH )CHO (8)
2
2
3
CH (ONO )COO(CH )CHO ] NO ]
2
2
3
3
CH (ONO )CO(CH )CHO ] NO ] O (9)
2
2
3
2
2
CH (ONO )CO(CH )CHO ]
2
2
3
CH C(O)CHO ] HCHO ] NO (10)
3
2
We also observe MPAN, which must arise from
abstraction;13 hence, both pathways seem to be occuring for
methacrolein.
For reaction of NO with methyl vinyl ketone, methyl
3
glyoxal was also observed. In this reaction, we would expect
addition to dominate: the reactions of NO with ketones
which contain no carbonÈcarbon unsaturated functionality,
and therefore presumably react via an abstraction mechanism,
are known to be very slow.16 A route can be envisaged for the
formation of methyl glyoxal following the addition of NO
3
In the experiments with methyl vinyl ketone and methyl
acrylate, there is a positive intercept in the Ðrst-order plots
(Fig. 3) which can be attributed to the initial consumption of
3
NO by a rapid reaction with impurities. So long as the impu-
3
CH C(O)CHCH ] NO ] O ]
rities react rapidly, the plots will be linear, and their slopes
3
2
3
2
will correctly represent the pseudo-Ðrst order rate coefficients.
However, the second-order plots, which are reasonably good
straight lines, show positive intercepts. We assume that these
intercepts originate from an increase in the reactivity of the
wall, which is not dependent on the concentration of the
carbonyl species in the range of the experiments. For methyl
vinyl ketone, there is quite good agreement between the rate
coefficients obtained by relative-rate and Ñow-tube methods
without the need to invoke secondary reactions. Methyl
CH C(O)CH(O )CH (ONO ) (11)
3
2
2
2
CH C(O)CHOOCH (ONO ) ] NO ]
3
2
2
3
CH C(O)CHOCH (ONO ) ] NO ] O (12)
3
2
2
2
2
CH C(O)CHOCH (ONO ) ]
3
2
2
CH C(O)CHO ] HCHO ] NO (13)
3
2
The yield of NO form the reactions studied in the Ñow tube
provides some further information about the mechanism of
2
acrylate reacts very slowly with NO (this was the slowest
3
reaction of the series that we were able to study in the Ñow
reaction. For reactions (1)È(3), NO was observed as a
2
product, indicating, at least prima facie, that NO addition to
3
system). As a result, the secondary reaction of NO with the
3
NO formed
the carbonÈcarbon double bond had occurred. In each of
2
these experiments, conducted in the absence of O , the addi-
NO ] NO ] M ] N O ] M
(14)
2
3
2
2 5
tion of NO would be expected to be followed by the elimi-
3
nation of NO to leave an epoxide. If the reaction were to
is particularly evident, especially since rather high pressures
were used. Invoking this reaction alone is sufficient to reconcile
the Ñow-tube data with the rate constant from the relative-
rate study.
The rate constants determined by relative-rate experiments
for the a,b-unsaturated aldehydes (acrolein, methacrolein)
di†er more substantially from those measured using the Ñow-
tube system; the Ñow-tube experiments give a much higher
value for the rate constants (see Tables 1 and 2). According to
our analysis of the Ñow-tube data, all three causes of error, (i),
(ii) and (iii), need to be invoked to explain the Ñow-tube data
2
proceed wholly by an addition mechanism followed by imme-
diate elimination, we would expect *[NO ]/[*[NO ] to be
very near to unity. In the case of the reaction of NO with
methyl vinyl ketone [reaction (1)], we do not anticipate any
2
3
3
abstraction by NO . In reality, *[NO ]/[*[NO ] is much
3
2
3
less than unity. Possible explanations are either that a bifunc-
tional compound (NO , NO or NO , NO ) may also be
2
3
3
3
formed, resulting in the consumption of two NO molecules
3
with no NO released, or that the initial adduct does not
2
decompose completely on the timescale of the experiment.
The value of *[NO ]/[*[NO ] of roughly 0.5 for the
on the reaction of NO with acrolein and methacrolein. The
positive intercepts in Fig. 4 are clearly indicative of the pres-
ence of fast-reacting impurities, but secondary chemistry of
2
3
3
reaction of acrolein suggests that the rate constants for addi-
tion and abstraction are roughly equal. Both abstraction and
addition must also occur for reaction with methacrolein
[reaction (2)], because we observe MPAN (abstraction
acyl radicals as well as of NO is also important. A good Ðt
2
between experimental and modelled data can be obtained
Phys. Chem. Chem. Phys., 1999, 1, 4195È4202
4199

View Article Online
using the rate coefficients from the relative-rate studies, espe-
hydrogen by the NO radical is sensitive to substituents on
the carbonyl group. This behaviour is exempliÐed by the reac-
3
cially if the wall-loss rates of NO in the presence of reactant
3
are incorporated.
tions of NO with HCHO, CH CHO, C H CHO and
3
3
2 5
C H CHO, for which the rate constants are20 0.058 ] 10~14,
3
7
0.29 ] 10~14, 0.57 ] 10~14 and 1.09 ] 10~14 cm3 molecule~1
Comparison of rate coefficients with those of other studies
s~1, respectively. The rate constant k
for acrolein
for meth-
abstraction
Table 4 details the rate constants obtained in this work and
some other recently determined values of the rate constants
would therefore be expected to be less than k
abstraction
acrolein. Furthermore, k
for acrolein is also expected to
addition
where available. The value of k determined by the relative-
be less than that for methacrolein, so that the increased reacti-
vity of methacrolein compared with acrolein can be attributed
to the operation of both e†ects.
1
rate technique lies between the lower limits obtained by Kwok
et al.14 and Rudich et al.17 The value obtained for k using
2
the relative-rate technique is less than the upper limit of
Rudich et al.,17 and in excellent agreement with the results of
Atmospheric implications
Kwok et al.,14 as is our value obtained for k using the
3
The reaction of NO with isoprene has been studied by
relative-rate technique.
3
Berndt and Bo
Ž
ge.21 These workers concluded that methyl
vinyl ketone and methacrolein were major products of the
reaction, with the yield of methyl vinyl ketone being greater
than that of methacrolein. A mechanism is proposed for the
atmospheric oxidation in which 1È2 addition takes place
Reactivity and reaction mechanisms
Inspection of Table 1 shows that the reactivity for the species
studied lies in the order methacrolein ^ methyl metha-
crylate [ acrolein [ methyl vinyl ketone [ methyl acrylate.
CH C(CH )CHCH ] NO ] O ]
2
3
2
3
2
NO is an electrophilic species and will add most readily to a
3
CH (ONO )COO(CH )CHCH (15)
carbonÈcarbon double bond which is electron rich,18,19 but
2
2
3
2
may also abstract a labile hydrogen, such as an aldehydic one.
All the species studied possess an a,b-unsaturated carbonyl
group but have di†erent substituents attached to it. The
ketone (methyl vinyl ketone) and the two esters (methyl acry-
CH (ONO )COO(CH )CHCH ] NO ]
2
2
3
2
3
CH (ONO )CO(CH )CHCH ] NO ] O (16)
2
2
3
2
2
2
CH (ONO )CO(CH )CHCH ]
late and methyl methacrylate) can react with NO only via the
addition mechanism. The presence of electron-withdrawing
substituents attached to the carbonÈcarbon double bond will
lower the electron density on the double bond and thus
reduce the reactivity; the rate constant for reaction of the
2
2
3
2
3
HCHO ] NO ] CH C(O)CHCH (17)
2
3
2
2NO ] 2O ] 2NO ] 2O
2
3
3
2
Net CH C(CH )CHCH ] 2O ]
2
3
2
3
NO radical with methyl acrylate is consequently lower than
CH C(O)CHCH ] HCHO ] 2O
(18)
3
3
2
2
that for reaction with methyl vinyl ketone since the ÈOCH
3
Alternatively, 3È4 addition may occur
group of the ester has an electron-withdrawing e†ect, whilst
the ÈCH group of methyl vinyl ketone donates electron
density to the double bond. A similar e†ect is observed in
CH C(CH )CHCH ] NO ] O ]
3
2
3
2
3
2
comparing the reactivity of methyl acrylate with methyl
methacrylate; the electron-donating inductive e†ect of the
methyl group attached to the carbonÈcarbon double bond in
methyl methacrylate increases the electron density on the
bond and hence increases the reactivity relative to that of methyl
acrylate.
CH C(CH )CHOOCH (ONO ) (19)
2
3
2
2
CH C(CH )CHOOCH (ONO ) ] NO ]
2
3
2
2
3
CH C(CH )CHOCH (ONO ) ] NO ] O (20)
2
3
2
2
2
2
CH C(CH )CHOCH (ONO ) ]
2
3
2
2
Rationalizing the reactivity of the a,b-unsaturated alde-
HCHO ] NO ] CH C(CH )CHO (21)
2
2
3
hydes towards the NO radical is less straightforward since
3
2NO ] 2O ] 2NO ] 2O
Net CH C(CH )CHCH ] 2O ]
CH C(CH )CHO ] HCHO ] 2O
the radical may either abstract the aldehydic hydrogen or add
2
3
3
2
to the carbonÈcarbon double bond. The rate constant for the
2
3
2
3
addition channel (k
the case of methacrolein than acrolein as the methyl group
would have an electron-donating e†ect on the double bond.
) would be expected to be greater in
(18)
addition
2
3
2
Oxidation of isoprene, a species that may persist into the
The rate constant for abstraction (k
) of an aldehydic
night, by NO will thus yield methyl vinyl ketone and meth-
abstraction
3
Table 4 Comparison of rate constants determined in this work to other recent determinations
Substrate
k/10~15 cm3 molecule~1 s~1
Technique
Ref.
Methyl vinyl
ketone
0.32 ^ 0.06
0.47 ^ 0.17
\0.6
Flow tube
This work
This work
14
Relative rate
Relative rate
Flow tube
\0.12
17
Methacrolein
Acrolein
9.6 ^ 2.0
3.7 ^ 1.0
4.4 ^ 1.7
O8.0
Flow tube
This work
This work
14
Relative rate
Relative rate
Flow tube
17
8.9 ^ 2.8
1.1 ^ 0.4
1.11 ^ 0.17
0.19 ^ 0.04
0.10 ^ 0.06
3.6 ^ 1.3
Flow tube
This work
This work
16
This work
This work
This work
Relative rate
Relative rate
Flow tube
Relative rate
Relative rate
Methyl
acrylate
Methyl
methacrylate
4200
Phys. Chem. Chem. Phys., 1999, 1, 4195È4202

View Article Online
acrolein. Provided [O ] is relatively high, NO itself will not
be removed. The oxidation of methyl vinyl ketone and meth-
and the oxidation of methyl glyoxal follows a similar pattern
CH C(O)CHO ] NO ] O ] CH C(O)O ] CO ] HNO
3
3
acrolein by NO in the atmosphere must therefore now be
3
3
2
3
2
3
3
considered. As we have noted earlier, oxidation of methyl
(31)
vinyl ketone by NO in the presence of O produces methyl
3
2
glyoxal,
followed by reactions (26)È(29) and 2 ] (18) leads to the net
reaction
CH C(O)CHCH ] NO ] O ]
3
2
3
2
Net CH C(O)CHO ] NO ] 3O ]
3
3
3
CH C(O)CHOOCH (ONO ) (11)
3
2
2
CO ] CO ] HCHO ] OH ] HNO ] 3O
2
3
2
CH C(O)CHOOCH (ONO ) ] NO ]
3
2
2
3
Only one molecule of NO is thus required to oxidise the pro-
ducts of isoprene degradation in the atmosphere, since MPAN
3
CH C(O)CHOCH (ONO ) ] NO ] O (12)
3
2
2
2
2
formation makes only a minor contribution.13 At the same
time, one OH radical is formed in its place. The implications
CH C(O)CHOCH (ONO ) ]
3
2
2
of the process are that NO concentrations can remain high in
CH C(O)CHO ] HCHO ] NO (13)
3
3
2
the presence of fast-reacting organic species (as a result of
NO regeneration), but that one molecule of OH may be pro-
2NO ] 2O ] 2NO ] 2O
3
2
3
3
2
duced for every NO consumed. The prerequisite for this
Net CH C(O)CHCH ] 2O ]
3
3
2
3
mechanism to operate is that O levels must be quite high,
CH C(O)CHO ] HCHO ] 2O
(18)
3
typically in excess of 30 ppb. At the very surface of the Earth
3
2
at night, O is rapidly destroyed either via dry deposition or
The addition of NO to methacrolein also yields methyl
3
3
by reaction with NO emitted at night. However, away from
glyoxal
the surface, but still in the boundary layer (the so called
CH C(CH )CHO ] NO ] O ]
residual boundary layer), O will not be deposited, and any
2
3
3
2
3
NO emitted at night will have been converted to NO near
2
the surface. In this residual boundary layer, NO can build to
3
CH (ONO )COO(CH )CHO (8)
2
2
3
very high levels (100 ppt).3 Aliwell and Jones3 have recently
CH (ONO )COO(CH )CHO ] NO ]
2
2
3
3
reported NO concentrations that are on average 89 ppt in
3
summer over Cambridge UK, and that peak at 450 ppt. A
CH (ONO )CO(CH )CHO ] NO ] O (9)
2
2
3
2
2
diurnally varying boundary layer model5 consisting of two
levels has been shown to reproduce these very high levels of
CH (ONO )CO(CH )CHO ]
2
2
3
NO in the residual boundary layer. At the very surface (up to
CH C(O)CHO ] HCHO ] NO (10)
3
approximately 300 m) in the nocturnal boundary layer, model
3
2
NO levels are low, typically zero to 20 ppt, because O
declines rapidly at night and also because of titration by NO
2NO ] 2O ] 2NO ] 2O
3
3
2
2
3
3
3
2
Net CH C(CH )CHO ] 2O ]
Ñuxes. In the residual boundary layer, where NO concentra-
3
3
CH C(O)CHO ] HCHO ] 2O
(18)
tions are high, NO will dominate the night-time oxidation of
3
2
3
methacrolein, and reaction with NO can make up a signiÐ-
cant fraction of the total overall loss of methacrolein. The
Hence, addition of NO to methyl vinyl ketone and meth-
3
3
acrolein yield methyl glyoxal and once more, provided [O ] is
3
daytime ratio of methyl vinyl ketone to methacrolein concen-
trations is greater than unity, and often around 2È2.5.22,23
During the day, oxidation of isoprene by OH generates more
methyl vinyl ketone than methacrolein, and also OH reacts
more quickly with methacrolein than methyl vinyl ketone,
driving the partition of these species apart.23 At night, if we
sufficiently high, NO will not be removed. NO also
3
3
abstracts the aldehydic hydrogen from methacrolein
CH C(CH )CHO ] NO ] O ]
2
3
3
2
CH C(CH )C(O)OO ] HNO (22)
2
3
3
ignore NO , ozonolysis of isoprene generates more meth-
3
CH C(CH )C(O)OO ] NO ]
acrolein than methyl vinyl ketone, and ozone also reacts more
2
3
3
quickly with methyl vinyl ketone than methacrolein. Martin et
al.22 observed at a rural site that the ratio of methyl vinyl
ketone to methacrolein concentrations is a maximum by day
and a minimum by night, consistent with OH oxidation by
CH C(CH )C(O)O ] O ] NO (23)
2
3
2
2
CH C(CH )C(O)O ] CH C(CH ) ] CO
(24)
(25)
(26)
(27)
(28)
(29)
2
3
2
3
2
CH C(CH ) ] 2O ] HCHO ] CH C(O)O
day and O -dominated night-time oxidation. In urban
2
3
2
3
2
3
regions, the methyl vinyl ketone to methacrolein concentra-
CH C(O)O ] NO ] CH O ] NO ] CO
3
2
3
3
2
2
2
tion ratio remains above unity throughout the day.22 Bie-
senthal and Shepson23 have shown that both methyl vinyl
ketone and methacrolein sources have a strong automotive
component, which may itself be the cause of the high ratio at
CH O ] NO ] CH O ] NO ] O
3
2
3
3
2
2
CH O ] O ] HCHO ] HO
3
2
2
night; however, NO will also drive the partitioning of methyl
3
HO ] O ] OH ] 2O
vinyl ketone and methacrolein apart. If isoprene lingers into
2
3
2
the night, NO oxidation produces more methyl vinyl ketone
3O ] 3NO ] 3NO ] 3O
3
3
2
3
2
3
than methacrolein, and, as we have seen, NO reacts consider-
3
Net CH C(CH )CHO ] NO ] 4O ]
ably faster with methacrolein than with methyl vinyl ketone,
2
3
3
2CO ] 2HCHO ] HNO ] OH ] 3O
(18)
i.e. the behaviour is the exact analogue of OH chemistry
during the day. The high ratio of methyl vinyl ketone to meth-
acrolein concentrations in urban regions may be further evi-
2
3
2
Although one NO radical is consumed, one OH radical is
3
formed in its place. MPAN can also be formed in the reaction
dence that NO plays an important role in the planetary
3
boundary layer. Measurements from the Southern Oxidant
CH C(CH )C(O)OO ] NO ] M ]
2
3
2
Study (SOS)24h27 show how the ratio of methyl vinyl ketone
to methacrolein concentrations changes when a forested area
CH C(CH )C(O)ONO ] M (30)
2
3
2
Phys. Chem. Chem. Phys., 1999, 1, 4195È4202
4201

View Article Online
7
8
A. A. Boyd, C. E. Canosa-Mas, A. D. King, R. P. Wayne and
M. R. Wilson, J. Chem. Soc., Faraday T rans., 1991, 87, 2913.
C. E. Canosa-Mas, M. D. King, R. Lopez, C. J. Percival, R. P.
Wayne, D. E. Shallcross, J. A. Pyle and V. Daele, J. Chem. Soc.,
Faraday T rans., 1996, 92, 2211.
W. B. DeMore, D. M. Golden, R. F. Hampson, M. J. Kurylo,
C. J. Howard, A. R. Ravishankara, C. E. Kolb and M. J. Molina,
Chemical Kinetics and Photochemical Data for Use in Strato-
spheric Modeling, JPL Publication 97-4, NASA, 1997.
is impacted by urban plumes. One of the conclusions drawn
from the SOS study is that NO reacts with methacrolein, and
3
that substantial amounts of MPAN are formed at night as a
product of this interaction.24
9
Conclusion
Rate coefficients for Ðve structurally similar a,b-unsaturated-
carbonyl compounds have been measured by relative and
absolute methods. The relative-rate measurements are in good
agreement with the results of previous studies where they
exist. However, the kinetic data obtained using the absolute
method are, in some cases, signiÐcantly higher than those
from the relative method. Secondary chemistry, together with
small amounts of reactive impurities, may be the cause of the
discrepancy. Product studies have enabled us to elucidate a
mechanism for the oxidation of methyl vinyl ketone and
10 T. Brauers and B. J. Finlayson-Pitts, Int. J. Chem. Kinet., 1997,
29, 665.
11 C. E. Canosa-Mas, S. J. Smith, S. J. Waygood and R. P. Wayne,
J. Chem. Soc., Faraday T rans. II, 1991, 87, 3473.
12 C. E. Canosa-Mas, S. J. Smith, S. Toby and R. P. Wayne, J.
Chem. Soc., Faraday T rans. II, 1988, 84, 263.
13 C. E. Canosa-Mas, M. D. King, D. E. Shallcross and R. P.
Wayne, Phys. Chem. Chem. Phys., 1999, 1, 2411.
14 E. C. Kwok, S. M. Aschmann, J. Arey and R. Atkinson, Int. J.
Chem. Kinet., 1996, 28, 925.
15 R. Atkinson, J. Phys. Chem. Ref. Data Monograph No. 1, 1989.
16 R. Atkinson, J. Phys. Chem. Ref. Data Monograph No. 2, 1994.
17 Y. Rudich, R. K. Talukdar, R. W. Fox and A. R. Ravishankara,
J. Phys. Chem., 1996, 100, 5374.
methacrolein by NO in the atmosphere. We conclude that
3
NO may be an important oxidant for methacrolein in the
3
urban boundary layer and may also be important where
18 R. Atkinson, J. Phys. Chem. Ref. Data, 1997, 26, 215.
19 R. P. Wayne, I. Barnes, P. Biggs, J. P. Burrows, C. E. Canosa-
Mas, J. Hjorth, G. Le Bras, G. K. Moortgat, D. Perner, G.
Poulet, G. Restelli and H. Sidebottom, Atmos. Environ., 1991,
25A, 1.
urban plumes impact on forested areas.
Acknowledgements
20 B. DÏAnna and C. J. Nielsen, Int. J. Chem. Kinet., 1997, 93, 3479.
21 T. Berndt and O. Boge, Int. J. Chem. Kinet., 1997, 29, 755.
22 R. Martin, H. Westberg, E. Allwine, L. Ashman, C. Farmer and
B. Lamb, J. Atmos. Chem., 1991, 13, 1.
23 T. A. Biesenthal and P. B. Shepson, Geophys. Res. L ett., 1997, 24,
1375.
The authors thank NERC for support under grants GST/02/
1259 and GT24/95/LSAC/2. The Centre for Atmospheric
Science is a joint initiative of the Department of Chemistry
and the Department of Applied Mathematics.
24 T. K. Starn, P. B. Shepson, S. B. Bertman, J. S. White, B. G.
Splawn, D. D. Reimer, R. G. Zika and K. Olszyna, J. Geophys.
Res., 1998, 103, 22425.
25 T. K. Starn, P. B. Shepson, S. B. Bertman, D. D. Reimer, R. G.
Zika and K. Olszyna, J. Geophys. Res., 1998, 103, 22437.
26 G. Nouaime, S. B. Bertman, C. Seaver, D. Elyea, H. Huang, P. B.
Shepson, T. K. Starn, D. D. Reimer, R. G. Zika and K. Olszyna,
J. Geophys. Res., 1998, 103, 22463.
27 J. M. Roberts, J. Williams, K. Baumann, M. P. Buhr, P. D.
Goldan, J. Holloway, G. Hubler, W. C. Kuster. A. A. McKeen,
T. B. Ryerson, M. Trainer, E. J. Williams, F. C. Fehsenfeld, S. B.
Bertman, G. Nouaime, C. Seaver, G. Grodzinsky, M. Rogers and
V. L. Young, J. Geophys. Res., 1998, 103, 22473.
References
1
F. Fehsenfeld, J. Calvert, R. Fall, P. Goldan, A. B. Guenther,
C. N. Hewitt, B. Lamb, S. Liu, M. Trainer, H. Westberg and
P. Zimmerman, Global Biogeochem. Cycles, 1992, 6, 389.
ScientiÐc Assessment of Ozone Depletion, 1994 World Meteoro-
logical Organization, Global Ozone Research and Monitoring
ProjectÈReport No. 37, 1995, Geneva, Switzerland.
S. Aliwell and R. L. Jones, J. Geophys. Res., 1998, 103, 5719.
M. E. Jenkin, S. M. Saunders and M. J. Pilling, Atmos. Environ.,
1997, 31, 81.
2
3
4
5
6
D. J. Fish, D. E. Shallcross and R. L. Jones, Atmos. Environ.,
1999, 33, 687.
J. A. Bagley, C. E. Canosa-Mas, M. R. Little, A. D. Parr, S. J.
Smith, S. J. Waygood and R. P. Wayne, J. Chem. Soc., Faraday
T rans., 1990, 86, 2109.
Paper 9/04613E
4202
Phys. Chem. Chem. Phys., 1999, 1, 4195È4202