Atmospheric aqueous phase radical chemistry of the isoprene oxidation products methacrolein, methyl vinyl ketone, methacrylic acid and acrylic acid-kinetics and product studies

Article abstract of DOI:10.1039/c3cp54859g

Kinetic and mechanistic studies were conducted on the isoprene oxidation products methacrolein, methyl vinyl ketone, methacrylic and acrylic acid reacting with hydroxyl and nitrate radicals and sulfate radical anions in aqueous solution by use of the laser flash photolysis technique and a reversed-rate method for kinetics. High-performance liquid chromatography/mass spectrometry was applied for product analysis. The kinetic investigations show the highest reactivity of the hydroxyl radical followed by sulfate and nitrate radicals. For methacrolein and methyl vinyl ketone the following rate constants have been determined at 298 K: k_{(OH+methacrolein)} = (9.4 ± 0.7) × 10⁹ M^-1 s^-1, k _{(OH+methylvinylketone)} = (7.3 ± 0.5) × 10⁹ M^-1 s^-1, k(NO₃+methacrolein) = (4.0 ± 1.0) × 10⁷ M^-1 s^-1, k(NO ₃+methylvinylketone)= (9.7 ± 3.4) × 10⁶ M^-1 s^-1, k _{(SO4-+methacrolein)} = (9.9 ± 4.9) × 10⁷ M^-1 s^-1 and k _{(SO4-+methylvinylketone)} = (1.0 ± 0.2) × 10⁸ M^-1 s^-1. Temperature and pH dependencies of the reactions of OH, NO₃ and SO₄ ^- with methacrolein, methyl vinyl ketone, methacrylic and acrylic acid as well as Arrhenius parameters have been obtained and discussed. Product studies were performed on the OH radical induced oxidation of methacrolein and methyl vinyl ketone. In the reaction of methacrolein + OH methylglyoxal and hydroxyacetone were identified as first oxidation products with yields of 0.099 and 0.162. Methylglyoxal was primarily produced in the oxidation of methyl vinyl ketone with a yield of 0.052. For both precursor compounds the formation of glycolaldehyde was observed for the first time with yields of 0.051 and 0.111 in the oxidation of methacrolein and methyl vinyl ketone, respectively. Furthermore, highly functionalised C₄ compounds were determined from the oxidation of both precursor compounds, but for the first time for methyl vinyl ketone. Reaction schemes were developed based on known peroxyl radical reaction mechanisms. The aqueous phase conversion of the first generation isoprene oxidation products can potentially contribute to tropospheric aqueous phase budgets of important carbonyl and dicarbonyl components which are expected to be conducive to the formation of aqSOA. This journal is

Full text of DOI:10.1039/c3cp54859g

PCCP
View Article Online
View Journal | View Issue
PAPER
Atmospheric aqueous phase radical chemistry of
the isoprene oxidation products methacrolein,
methyl vinyl ketone, methacrylic acid and acrylic
acid – kinetics and product studies†
Cite this: Phys.Chem.Chem.Phys.,
2014, 16, 6257
Luisa Schone,a Janine Schindelka,a Edyta Szeremeta,b Thomas Schaefer,a
¨
Dirk Hoffmann,a Krzysztof J. Rudzinski,b Rafal Szmigielskib and
Hartmut Herrmann*a
Kinetic and mechanistic studies were conducted on the isoprene oxidation products methacrolein, methyl
vinyl ketone, methacrylic and acrylic acid reacting with hydroxyl and nitrate radicals and sulfate radical
anions in aqueous solution by use of the laser flash photolysis technique and a reversed-rate method for
kinetics. High-performance liquid chromatography/mass spectrometry was applied for product analysis.
The kinetic investigations show the highest reactivity of the hydroxyl radical followed by sulfate and nitrate
radicals. For methacrolein and methyl vinyl ketone the following rate constants have been determined at
298 K: k(OH+methacrolein) = (9.4 ꢀ 0.7) ꢁ 109 Mꢂ1
s
ꢂ1, k(OH+methylvinyl ketone) = (7.3 ꢀ 0.5) ꢁ 109 Mꢂ1 sꢂ1
,
k(NO +methacrolein) = (4.0 ꢀ 1.0) ꢁ 107 Mꢂ1
s
ꢂ1, k(NO +methylvinylketone) = (9.7 ꢀ 3.4) ꢁ 106 Mꢂ1 sꢂ1
,
3
3
k(SO
ꢂ
= (9.9 ꢀ 4.9) ꢁ 107 Mꢂ1 sꢂ1 and k(SO
ꢂ
4 +methylvinyl ketone)
= (1.0 ꢀ 0.2) ꢁ 108 Mꢂ1 sꢂ1
.
4 +methacrolein)
ꢂ
Temperature and pH dependencies of the reactions of OH, NO3 and SO4 with methacrolein, methyl vinyl
ketone, methacrylic and acrylic acid as well as Arrhenius parameters have been obtained and discussed.
Product studies were performed on the OH radical induced oxidation of methacrolein and methyl vinyl
ketone. In the reaction of methacrolein + OH methylglyoxal and hydroxyacetone were identified as first
oxidation products with yields of 0.099 and 0.162. Methylglyoxal was primarily produced in the oxidation
of methyl vinyl ketone with a yield of 0.052. For both precursor compounds the formation of
glycolaldehyde was observed for the first time with yields of 0.051 and 0.111 in the oxidation of
methacrolein and methyl vinyl ketone, respectively. Furthermore, highly functionalised C4 compounds
were determined from the oxidation of both precursor compounds, but for the first time for methyl vinyl
ketone. Reaction schemes were developed based on known peroxyl radical reaction mechanisms. The
aqueous phase conversion of the first generation isoprene oxidation products can potentially contribute
to tropospheric aqueous phase budgets of important carbonyl and dicarbonyl components which are
expected to be conducive to the formation of aqSOA.
Received 18th November 2013,
Accepted 4th February 2014
DOI: 10.1039/c3cp54859g
www.rsc.org/pccp
formation of organic particulate matter.9–19 While the reactivity
and degradation mechanisms of isoprene in the gas phase as
1. Introduction
Isoprene represents one of the largest single sources of organic well as its contribution to organic particulate mass have been
carbon in the atmosphere.1,2 The chemistry of biogenically studied for decades in many experiments (ref. 9, 18, 20–22 and
emitted compounds affects numerous atmospheric processes references therein), only very few studies exist dealing with the
and parameters such as regional concentrations of ozone,3 further fate of the main direct isoprene degradation products
the overall oxidation capacity of the atmosphere4–8 and the methacrolein and methyl vinyl ketone.22–29 In turn, the aqueous
chemistry of short-chain degradation products of later genera-
tions, such as glyoxal, methylglyoxal and hydroxyacetone, has
a Leibniz-Institute for Tropospheric Research, Permoserstraße 15, 04318 Leipzig,
been the main subject of several laboratory and modelling
studies.30–38 These studies have shown that aqueous processes
Germany. E-mail: herrmann@tropos.de
b Institute of Physical Chemistry of the Polish Academy of Sciences,
can contribute to the formation of secondary organic aerosol mass
Laboratory of Environmental Chemistry, 01-224 Warsaw, Poland
which might be addressed as ‘‘aqSOA’’. Degradation processes of
methacrolein and methyl vinyl ketone in the aqueous phase,
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c3cp54859g
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 6257--6272 | 6257

View Article Online
Paper
PCCP
ꢂ
however, have largely been neglected because of the small Henry OH and SO4 (Coherent, 0222-583-00) or 635 nm for NO3
coefficients of both compounds (Hmethacrolein = 6.5 M atmꢂ1 and (He–Ne, Spindler & Hoyer, 35-2) is lead several times through
Hmethyl vinyl ketone = 41 M atmꢂ1),39 and hence generally only the measurement cell by a White-mirror configuration resulting
their gas-phase chemistry was considered relevant. However, in high sensitivity with optical path lengths of 84 cm for OH
a recent study by van Pinxteren (2005) showed that some and SO4ꢂ and 18 cm for NO3. The intensity of the analysis laser
carbonyl compounds can be found in much higher concentra- beam behind the measurement cell is detected by a photo-
tions in the aqueous phase than expected only from their Henry diode. The signal is passed on to an oscilloscope (Gould
coefficients.40 Recently, a study by Liu et al. (2012) found a Instrument System, Classic Delta)ꢂand a computer for further
change in the reversible to irreversible uptake of methacrolein analysis and – in the case of SO4 and NO3 measurements –
with increasing sulfuric acid concentration in aqueous solution can be converted to an absorbance–time-profile to determine
due to acid-catalysed oxidation with added hydrogen peroxide.41 the pseudo-first order rate constant (Fig. A0 in the ESI†).
Hence, as the phase transfer of methyl vinyl ketone and The slope of the regression line of a plot of pseudo-first
methacrolein might be governed by reactive uptake, their order rate constants as a function of reactant concentration
degradation and transformation processes in the tropospheric gives the second-order rate constant of the reaction (Fig. A8
aqueous phase might be of greater importance which has been in ESI†).
shown lately by the studies of Liu et al. (2009, 2012) and Zhang
As a precursor for SO4ꢂ, a 5 ꢁ 10ꢂ4 M solution of potassium
et al. (2010).41–43 In order to implement the above-mentioned peroxodisulfate was used. To form nitrate radicals in the
aqueous phase chemistry in atmospheric models and to inves- aqueous solutions potassium peroxodisulfate and sodium nitrate
tigate the atmospheric relevance of these multiphase reactions, were added in concentrations of [K2S2O8] = 0.03 M and [NaNO3] ꢂ=
2ꢂ
systematic laboratory measurements under defined conditions 0.1 M. Photolysis of S2O8 and subsequent reaction with NO3
are necessary as a first step.
then forms the nitrate radicals. As OH radical absorption bands
Within the present study, both the kinetics and the products in aqueous solution overlap with organic substances, the method
of free radical oxidation reactions of OH, NO3 and SO4ꢂ with the of competition kinetics with thiocyanate as a reference system
isoprene degradation products methacrolein (C4H6O), methyl was used.44 For OH radical reactions, 1 ꢁ 10ꢂ4 M of hydrogen
vinyl ketone (C4H6O), methacrylic acid (C4H6O2) and acrylic acid peroxide as well as [KSCN] = 1.59 ꢁ 10ꢂ5 M were added to the
(C3H4O2) were studied in aqueous solution. Based on the solution. The thiocyanate system is based on the following
obtained laboratory data, a degradation scheme for tropospheric reactions (R1)–(R4):
aqueous phase chemistry is proposed.
OH + SCNꢂ " SCNOHꢂ
SCNOHꢂ " SCN + OHꢂ
(R1)
(R2)
(R3)
(R4)
2. Experimental
2.1. Kinetic studies
SCN + SCNꢂ " (SCN)2
ꢂ
Kinetics were studied using the laser flash photolysis setup
similar to the one of Gligorovski et al. (2009) shown in Fig. 1.46
The centre of the setup is a flow cell containing a solution with
both the precursors for the radical and the organic reactant
to be investigated. Photolysis of the precursor by excimer laser
(Lambda LPX 100) pulses at 248 nm (OH, SO4ꢂ) or 351 nm
(NO3) forms the radicals that react with the organic compound.
For detection a continuous-wave laser operating at 407 nm for
OH + reactant - products
The formation of the dithiocyanate radical anion (SCN)2ꢂ is
recorded at 407 nm because of its strong absorbance in the blue
region of the spectrum. Chin and Wine (1992) reported the
temperature-dependent rate constant of reaction (R1) that
was used as the reference value.44 Herrmann et al. (2010)
recommended a quantum yield for the H2O2 photolysis of
FOH = 1.0 ꢀ 0.2.47 The initial concentration of OH radicals
formed by photolysis depends on the wavelength and laser
energy as well as the extinction properties of H2O2. Such levels
of OH will be safely converted to (SCN)2ꢂ. Resulting from
ꢂ
reactions (R1) to (R3), the (SCN)2 concentration is equal to
the initial OH radical concentration [OH]0 when a complete
conversion is assumed without any other reaction partners.
In the presence of a reactant, the yield of (SCN)2ꢂ is reduced by
the fraction of OH that is consumed in reaction (R4). More
details of the competition kinetics method have been described
elsewhere.45–48 As the reactants used absorb light at the photo-
lysis wavelength, a correction of the inner filter effect was done
according to Schaefer et al. (2012).35 This correction considers
the absorption of photolysis energy by the reactant leading to
less available energy for the formation of radicals. The rate
constant decreases due to this correction.
Fig. 1 Schematic working plan of a laser flash photolysis setup for kinetic
investigations of fast radical reactions. Red continuous-wave laser at 635 nm
for the nitrate radical NO3 and blue continuous-wave laser at 40ꢂ7 nm for
reactions with hydroxyl radicals OH and sulfate radical anions SO4
.
6258 | Phys. Chem. Chem. Phys., 2014, 16, 6257--6272
This journal is © the Owner Societies 2014

View Article Online
PCCP
The organic compounds investigated were methacrolein,
Paper
Table 1 Experimental conditions of the product studies
methyl vinyl ketone, methacrylic and acrylic acid, the latter
two both in their protonated and deprotonated forms. The
pH was adjusted using NaOH, a phosphate buffer or HClO4.
Temperature dependent studies were conducted in the range of
278 r T r 328 K.
Methyl vinyl
ketone + OH
Experiment
Methacrolein + OH
[Reactant]0 in mM
[H2O2]0 in mM
0.1
9.0
107
248
0.05
5.0
63
248
1–5, 10, 20, 50
17
0.7
Laser fluence in mJ cmꢂ2
llaser in nm
The unusual influence of temperature on reaction kinetics
Number of laser pulses
e248nm in Mꢂ1 cmꢂ1
Absorption of light by
reactant in %
1–5, 10, 15, 20, 30, 50
126
5.2
ꢂ
of carbonyl compounds with SO4 radical anions observed in
laser flash photolysis experiments motivated cross-checking
with another experimental method. A reversed-rates method
was used, which derives a relative rate constant for a reaction of
sulfate radical anions with a compound inhibiting the radical-
chain autoxidation of inorganic sulfite S(IV) catalysed by Fe(III)
ions.49,50 The principle of the method is that two inhibitors are
examined in separate runs in a well-stirred and isothermal
reactor operated batch-wise under equivalent conditions. For
each inhibitor, several runs are carried out which differ only in
the initial concentration of the inhibitor. The conditions have to
be suitably set so that the reactions develop pseudo-stationary
states and the inversed pseudo-stationary rates observed are
linear against the initial concentrations of the inhibitors used.
Quasi-stationary analysis of the chain-radical mechanism of the
inhibited autoxidation showed that in such cases the ratio of
slopes of the linear relations for two inhibitors is equal to the
ratio of the rate constants for reactions of sulfate radical anions
with these inhibitors (the mechanism and assumptions of
the analysis are listed in ESI†). Thus, one can determine an
unknown rate constant against a certified value of a known
constant for a reference inhibitor:
[OH]est. per pulse in mM
0.088
0.034
2.2. Oxidation experiments and analytical procedures
Product studies on the OH radical initiated degradation of
methacrolein and methyl vinyl ketone in the aqueous phase
were initialised by means of laser-induced photolysis of hydro-
gen peroxide (lphotolysis = 248 nm). A quartz cuvette with a
volume of 3 ml and an irradiated area of 3 cm2 served as the
reaction vessel where the reactant solution was irradiated with
a varying number of laser pulses. The experimental conditions
of the oxidation experiments are listed in Table 1. For the
analysis of carbonyl compounds formed, a sample aliquot of
2 ml was derivatised with 2,4-dinitrophenylhydrazine. The
reaction solution was stored in the dark for 2 h, cleaned via
solid phase extraction and analysed using a high-performance
liquid chromatography/mass spectrometry system (Ultimate 3000,
Dionex). The separation of the analytes was performed using
a reversed phase C18-column (Waters SunFire 2.1 ꢁ 100 mm,
3.5 mm) and a binary eluent (H2O:CH3CN) with the addition of
0.2% acetic acid each. The electrospray ionisation quadrupole
mass spectrometer (Thermofisher) was used in the negative
mode. The reaction conditions were chosen carefully so that
predominately hydrogen peroxide was photolysed by the UV
light and the reactants were degraded by OH radicals. Control
experiments were performed without addition of hydrogen
peroxide to exclude major contributions of the direct photolysis
of the reactants. Measurements showed a loss of methacrolein
of about 0.6% per laser pulse and methyl vinyl ketone of about
0.4% per laser pulse, respectively.
slope
k ¼ kref
(1)
sloperef
Uncertainties of the rate constants determined are estimated
using the exact differential method, with standard errors of the
slopes, s, and the otherwise known uncertainty of the reference
rate constant, Dkref
:
ꢀ
ꢁ
s
sslope
sloperef
sloperef sloperef
slope
Dk ¼ kref
þ
þ
Dkref
(2)
2
sloperef
More details of the method were given elsewhere.51 In the past,
the method was successfully applied to several terpenes, alcohols,
carboxylic acids, chlorophenols and nitrophenols.49,51,52
2.3. Reagents
In the present work, the reactions were followed in a 250 ml The following chemicals were used without further purifica-
¨
glass reactor with PTFE cover, magnetically stirred, jacket- tion: NaNO3 (Riedel-de Haen, 99.5%), K2S2O8 (Fluka, Z99.0%),
thermostated and equipped with a pH and a pO2 electrode. The H2O2 (Fluka, Z30%), KSCN (Fluka, Z99.0%), methacrolein
reactor contained no gas phase, while the reacting solutions were (Aldrich, 95%), methyl vinyl ketone (Aldrich, 99%), methacrylic
initially saturated with oxygen from air. The acidity of solutions acid (Aldrich, 99%), acrylic acid (Aldrich, 99%), HClO4 (J. T.
was adjusted to pH 3 with dilute H2SO4. The rate of the Na2SO3 Baker Analyzed, 70–72%), Na2HPO4 (Fluka, Z98%), NaH2PO4
¨
autoxidation was followed by measuring the concentration of (Riedel-de Haen, 99%), NaOH (Fluka, 1 M), methylglyoxal
oxygen in reacting solutions. Ethanol was chosen as a reference (Sigma, 40% in water), hydroxyacetone (Fluka, o90%), glycol-
inhibitor, with necessary rate constants taken from Clifton and aldehyde dimer (Aldrich), pyruvic acid (Aldrich, 98%), Fe(ClO4)3
Huie (1989).53 Initial concentrations of reactants were: 2 mM nonahydrate (Fluka, 497.5% analysed), Na2S2O5 (Merck, EMSUREs
s
¨
Na2SO3 (added as Na2S2O5); 0.25 mM O2; 0.01 mM Fe(III) (added ACS, Reag. Ph Eur. 498%), H2SO4 (Riedel-de Haen, Fixanal
as Fe(ClO4)3); (0.05–10.20) mM EtOH; and (0.0095–5.37) mM 0.005 M). All solutions were freshly prepared using MilliQ water
carbonyl compounds. Methacrolein and methyl vinyl ketone were (18 MO cmꢂ1 from Millipore). For derivatisation of the carbonyl
studied successfully, while the acids had no inhibitive effect.
compounds, 2,4-dinitrophenylhydrazine was used, which was
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 6257--6272 | 6259

View Article Online
Paper
PCCP
purchased from Fluka (Z99% assay, delivered in 50% water). constants for the reaction with OH radicals of 3.5 ꢁ 109 Mꢂ1 sꢂ1
It was recrystallised in acetonitrile (HPLC-grade) two times and and 4.2 ꢁ 109 Mꢂ1 sꢂ1 are similar to those of the present work.55
dried in a desiccator for a minimum of one day.
However, as the structures differ from each other, the rate
constants are also expected not to be the same. Acrolein and
methacrolein differ by a methyl group that shows electron
pushing properties resulting in a higher electron density at the
CQC double bond which attracts the oxygen of the OH radical
leading to a faster addition of OH to the double bond compared
3. Results and discussion
3.1. Kinetics at 298 K
Table 2 summarises the kinetic results obtained for the reactions to acrolein. It has to be noted that these two measurements were
of methacrolein, methyl vinyl ketone, methacrylic and acrylic acid conducted at 293 K (compared to 298 K in this work), so the
(both at pH 1 and pH 8) with the hydroxyl radical OH, the nitrate values must be somewhat smaller.
ꢂ
radical NO3 and the sulfate radical anion SO4 in aqueous
Maruthamuthu (1980) obtained rate constants of
solution at 298 K together with available literature data. The 1.57 ꢁ 1010 Mꢂ1 sꢂ1 and 5.74 ꢁ 109 Mꢂ1 sꢂ1 for the reaction
uncertainties stated are statistical errors with the Student-t factor of methacrylate and acrylate with OH radicals at pH 7–7.2,
for the 95% confidence level derived from linear regression respectively.56 The slightly smaller value for acrylate compared
analysis of k1st values over reactant concentrations.
to this work may result from the somewhat smaller pH, but it lies
The kinetic investigation of the reaction of methacrolein + OH in the range of error of our measurements. The rate constant for
has already been the subject of several studies. In this work the methacrylate + OH measured by Maruthamuthu (1980) is slightly
rate constant was determined to be (9.4 ꢀ 0.7) ꢁ 109 Mꢂ1 sꢂ1 higher than in the present work, even exceeding the diffusion
which is about 30% lower than in Gligorovski et al. (2009, limit (cf. discussion part 3.3.1 and Table 5).56 Maruthamuthu
k = (14 ꢀ 3) ꢁ 109 Mꢂ1
s
ꢂ1).46 As methacrolein absorbs (1980) also determined the kinetic behaviour of methacrylate
at the photolysis wavelength with a molar absorptivity of and acrylate exposed to SO4ꢂ being a factor of two faster than in
126 ꢀ 13 Mꢂ1 cmꢂ1, a correction due to the inner filter effect this work for methacrylate and showing a similar value for
of the reactant is essential as has recently been demonstrated acrylate (cf. Table 2).56 For methacrylic acid, a more than four
by our laboratory.35 Thus, this correction was applied to both times larger rate constant is found. The behaviour that the
the measurements of Gligorovski et al. (2009) and of the present protonated form of an acid reacts faster compared to the anion
work, decreasing the rate constant slightly in both cases.46 form is very unusual and was not confirmed during this work’s
However, this correction alone cannot explain the whole dis- measurements nor in any other study (to the authors’ knowledge).
crepancy between both measurements but only a difference in Missing error ranges for any of the given kinetic data, a too brief
k of 1 ꢁ 109 Mꢂ1
s
ꢂ1. The values of Buxton et al. (2000) and presentation of the reaction conditions and no detailed discus-
Liu et al. (2009) of k = (8.0 ꢀ 0.7) ꢁ 109 Mꢂ1 sꢂ1 at 293 K and sion part also hamper a firm comparison with Maruthamuthu’s
k = (5.8 ꢀ 0.9) ꢁ 109 Mꢂ1 sꢂ1 at 279 K are smaller than in this early results.
work which is not surprising because of the lower temperature
Buxton (2000) used the laser flash photolysis method to
the cited authors chose for their experiments.42,54 Both results obtain a rate constant for the reaction of sulfate radicals with
match quite well with the Arrhenius plot of this work (cf. Table A1 methacrolein, which is higher by an order of magnitude than the
and Fig. A1 in the ESI†). A rough comparison can be drawn to the rate constants consistently determined in the present study using
similarly structured unsaturated aldehydes crotonaldehyde and two distinct methods – laser flash photolysis and reversed-rates.54
acrolein measured by Lilie and Henglein (1970) at 293 K which is The difference was in the experimental conditions: Buxton (2000)
the most suitable comparison available.55 The corresponding rate used a two times higher concentrations of the K2S2O8 precursor
Table 2 Measured second order rate constants at 298 K for the reactions with hydroxyl, nitrate and sulfate radicals. Bold: laser flash photolysis
experiments of this work, for those labelled as * reversed-rates method was used; Gligorovski et al. (2009);46 Buxton et al. (2000), 293 K;54 Liu et al.
(2009), 279 K;42 Maruthamuthu (1980)56
k2nd/L molꢂ1 sꢂ1
ꢂ
Compound
OH
NO3
SO4
Methacrolein
(9.4 ꢀ 0.7) ꢁ 109
(14 ꢀ 3) ꢁ 109 46
(8.0 ꢀ 0.7) ꢁ 109 54
(5.8 ꢀ 0.9) ꢁ 109 42
(7.3 ꢀ 0.5) ꢁ 109
(4.0 ꢀ 1.0) ꢁ 107
(9.9 ꢀ 4.9) ꢁ 107
(5.5 ꢀ 1.4) ꢁ 107
*
(1.18 ꢀ 0.1) ꢁ 109 54
(1.0 ꢀ 0.2) ꢁ 108
Methyl vinyl ketone
Methacrylic acid pH 1
Methacrylate pH 8
(9.7 ꢀ 3.4) ꢁ 106
(9.2 ꢀ 1.6) ꢁ 107
(1.7 ꢀ 1.2) ꢁ 108
(8.7 ꢀ 2.2) ꢁ 107
(2.5 ꢀ 1.2) ꢁ 108
*
(1.1 ꢀ 0.1) ꢁ 1010
1.06 ꢁ 109 (pH 2.9)56
(3.5 ꢀ 1.1) ꢁ 108
7 ꢁ 108 (pH 6.9)56
(9.5 ꢀ 0.8) ꢁ 107
(9.9 ꢀ 2.0) ꢁ 107
1.06 ꢁ 108 (pH 6.5)56
(1.1 ꢀ 0.1) ꢁ 1010
1.57 ꢁ 1010 (pH 7–7.2)56
(5.1 ꢀ 0.8) ꢁ 109
Acrylic acid pH 1
Acrylate pH 8
(6.9 ꢀ 1.0) ꢁ 106
(5.9 ꢀ 0.9) ꢁ 109
(4.4 ꢀ 0.6) ꢁ 107
5.74 ꢁ 109 (pH 7–7.2)56
6260 | Phys. Chem. Chem. Phys., 2014, 16, 6257--6272
This journal is © the Owner Societies 2014

View Article Online
PCCP
Paper
of sulfate radicals and saturated the solutions with Ar to remove
oxygen.54 The lack of oxygen simplified the chemistry involved,
but could have provided an additional sink of sulfate radicals
through their possible reaction with organic radicals primarily
formed from methacrolein. Furthermore, to determine the rate
constant, the regression line in the plot of k1st vs. [reactant] was
forced through the origin of ordinates. During the present
works’ measurements, [reactant] = 0 M was measured by
default for all radicals indicating y-intercepts being different
from zero (cf. Fig. A8 in the ESI†) due to the reaction of the
radical with the precursor and itself – which Buxton et al. (2000)
judges as negligible54 – as well as with water.57 For a-terpineol
(as methacrolein is not shown in the paper of Buxton et al.,
2000) the obtained rate constants differ by 5% from the laser
flash photolysis data and by about 8% from PR data, once
calculated with and without considering forcing the regression
through the origin.54 Altogether, this variation in the analysis
also does not explain the discrepancy of an entire order of
magnitude in the rate constants of both studies. A comparison
of the available data for unsaturated compounds indicates that
the value of Buxton et al. (2000) is rather too large.54
The reaction rate constants in Table 2 indicate that the
reactions between isoprene oxidation products and sulfate
radicals are faster than those with NO3, on average by a factor
of about 5.5. Apparently, the reactivity of OH radicals is shown
to be significantly higher compared to the other two radicals
examined with k(OH) E 350ꢃk(NO3) and k(OH) E 60ꢃk(SO4ꢂ),
respectively. Based on the results, the following trend in reac-
tivity can be given: k(OH) c k(SO4ꢂ) 4 k(NO3). Furthermore,
the dissociated forms of the acids react faster than the undis-
sociated forms of all three radicals. Other studies confirm this
trend regardless of whether H abstraction, addition or electron
transfer reactions of OH, NO3 or SO4ꢂ with saturated, unsaturated
or aromatic compounds were considered.45,58–62 The only expla-
nation of this behaviour is slightly differing electron densities
occurring at the CQC double bond due to electron-withdrawing
properties of the deprotonated group. This leads to slightly
differing rate constants between protonated and deprotonated
acid forms as observed in the present study.
Fig. 2 Top: Arrhenius plots for acrylic acid (m) and methacrylic acid (K)
exposed to NO3. Bottom: Arrhenius plots for methacrylate (m) and acrylate
(K and J) exposed to SO4ꢂ. White dots (J) refer to values of acrylate that
were not used for averaging. The dotted lines (---) refer to the average
value of the measured temperature range.
mean rate constants of k0 = (9.8 ꢀ 1.2) ꢁ 107 Mꢂ1 sꢂ1 and
k0 = (3.3 ꢀ 0.5) ꢁ 108 Mꢂ1 sꢂ1 in the temperature range
3.2. Temperature-dependent kinetic studies
A detailed description of the temperature dependency of rate 278 r T r 318 K and 278 r T r 298 K are proposed,
constants according to the Arrhenius formula as well as the respectively.
calculation of activation parameters can be found in Laidler
(1987).63 A short version of the necessary equations is available vinyl ketone, acrylic acid and acrylate are summarised in
in the ESI.† Table 3. For the reaction of methacrolein and methyl vinyl
Arrhenius parameters obtained for methacrolein, methyl
3.2.1 Nitrate radical NO3. Only in some cases a clear ketone with NO3 the parameters show only minor differences
temperature dependence could be measured for the nitrate indicating a similar behaviour with changing ambient tempera-
radical reactions. Depending on the organic substance investi- ture. Except for the Gibbs’ free enthalpy of activation DG‡, the
gated, reaction rate constants with NO3 indicated a temperature calculated Arrhenius parameters are much larger for acrylic
dependence being distinct for acrylic acid and minor for acrylate, acid than for acrylate. Especially the slightly positive value for
methyl vinyl ketone and methacrolein or non-existent for both the entropy of activation DS‡ (usually {0 J molꢂ1
K
ꢂ1) in the
forms of methacrylic acid. Fig. 2 shows the Arrhenius plots for reaction of acrylic acid with NO3 is conspicuous.
both a clear (acrylic acid) and no (methacrylic acid) temperature
3.2.2 Sulfate radical anion SO4ꢂ. Generally, the measured rate
dependence as examples. The other Arrhenius plots are constants for sulfate radical reactions are not strongly dependent
found in the ESI.† For methacrylic acid and methacrylate + NO3, on temperature. Only for methacrylic acid, methacrylate and
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 6257--6272 | 6261

View Article Online
Paper
PCCP
Table 3 Arrhenius parameters obtained during the investigation of methacrolein, methyl vinyl ketone, methacrylic acid, methacrylate, acrylic acid and
acrylate exposed to NO3, SO4 and OH. A = collision parameter, EA = energy of activation, DG‡ = Gibbs’ free energy of activation, DH‡ = enthalpy of
ꢂ
activation, DS‡ = entropy of activation
Compound
Radical
NO3
A/L molꢂ1 sꢂ1
EA/kJ molꢂ1
DG‡/kJ molꢂ1
DH‡/kJ molꢂ1
DS‡/J molꢂ1 Kꢂ1
Methacrolein
(5.8 ꢀ 0.5) ꢁ 108
—
7 ꢀ 4
—
30 ꢀ 20
—
4.2 ꢀ 2.5
—
ꢂ(85 ꢀ 7)
—
ꢂ
SO4
OHa
OHb
(5.6 ꢀ 0.6) ꢁ 1011
10 ꢀ 7
16 ꢀ 12
8 ꢀ 5
ꢂ(28 ꢀ 3)
(4.9 ꢀ 0.2) ꢁ 1013
20 ꢀ 3
15 ꢀ 3
18 ꢀ 3
9 ꢀ 1
Methyl vinyl ketone
Methacrylic acid
Methacrylate
Acrylic acid
NO3
SO4
OH
(6.2 ꢀ 1.1) ꢁ 108
—
10 ꢀ 8
—
12 ꢀ 3
33 ꢀ 33
—
17 ꢀ 4
8 ꢀ 6
—
9 ꢀ 2
ꢂ(85 ꢀ 14)
—
ꢂ(24 ꢀ 1)
ꢂ
(9.0 ꢀ 0.4) ꢁ 1011
NO3
—
—
—
—
—
ꢂ
SO4
(1.2 ꢀ 0.4) ꢁ 1010
11 ꢀ 19
26 ꢀ 56
8 ꢀ 15
9 ꢀ 3
ꢂ(60 ꢀ 20)
OH
(1.0 ꢀ 0.1) ꢁ 1012
11 ꢀ 3
16 ꢀ 5
ꢂ(23 ꢀ 1)
NO3
—
—
—
—
—
SO4
(3.5 ꢀ 1.1) ꢁ 109
6 ꢀ 17
25 ꢀ 72
10 ꢀ 10
ꢂ(70 ꢀ 21)
ꢂ
OH
(8.0 ꢀ 0.5) ꢁ 1012
16 ꢀ 6
16 ꢀ 5
14 ꢀ 4
ꢂ(6.2 ꢀ 0.4)
NO3
(2.2 ꢀ 0.4) ꢁ 1013
(2.0 ꢀ 0.2) ꢁ 108
(9.4 ꢀ 0.1) ꢁ 1010
37 ꢀ 12
2 ꢀ 4
34 ꢀ 17
28 ꢀ 61
18 ꢀ 2
34 ꢀ 12
ꢂ(1 ꢀ 1)
5 ꢀ 1
2.0 ꢀ 0.3
ꢂ(94 ꢀ 8)
-(43 ꢀ 1)
ꢂ
SO4
OH
7 ꢀ 1
Acrylate
NO3
SO4
OH
(2.2 ꢀ 0.2) ꢁ 109
(1.8 ꢀ 0.1) ꢁ 1010
10 ꢀ 5
30 ꢀ 19
8 ꢀ 4
ꢂ(74 ꢀ 8)
ꢂ
—
—
—
—
—
3 ꢀ 1
17 ꢀ 10
0.3 ꢀ 0.2
ꢂ(55 ꢀ 1)
Methacrolein + OH: This work. Calculated from the k(T) values given in Gligorovski et al. (2009).46
a
b
acrylic acid, Arrhenius parameters could be obtained (cf. Table 3). than its protonated form. This behaviour was observed during all
Especially for EA, DG‡ and DH‡ errors are quite large due to the measurements at room temperature (cf. Table 2) and is in line
uncertainties in the rate constants together with the small with numerous other kinetic studies (e.g. ref. 41 and references
temperature dependencies observed. For methacrolein, methyl therein). Regarding the temperature dependencies of the reac-
vinyl ketone and acrylate no change in rate constants with tions of OH radicals with acrylic acid and acrylate, respectively,
increasing temperature was observed. Some measurements at (Fig. 3, top), rate constants are found to be more susceptible
higher temperature had to be discarded (cf. bottom of Fig. 2) towards changes in temperature if the undissociated form is
because of the apparently too small values measured. It was considered. On the other hand, for the reactions of methacrylic
assumed whether this might be due to the thermal decomposition acid and methacrylate with OH radicals (Fig. 3, bottom),
of the radical precursor and/or the reaction with the reactant since no differences in rate constants or temperature behaviour are
K2S2O8 itself is a strong oxidising agent. However, model calcula- observed.
tions were not able to support these suppositions since such
reactions are much too slow.
Gligorovski et al. (2009) presented rate constants dependent
on temperature for the reaction of methacrolein with OH.46
A comparison of the activation parameters reveals equally high As no activation parameters are given, they were calculated in
negative changes of entropy and similar values for DG‡ and DH‡ this contribution from the given k(T) values. Table 3 shows the
for all reactions. EA and A decrease in the order methacrylic acid 4 activation parameters (a) determined in the present work
methacrylate 4 acrylic acid, with one order of magnitude change and (b) calculated from the k(T) of Gligorovski et al. (2009).46
in the A-factor in each step. For the reaction of methacrolein, Whereas the changes in Gibbs free energy are similarly large,
methyl vinyl ketone, acrylic acid and acrylate with SO4ꢂ, mean the enthalpy of activation as well as the activation energy and
rate constants of kmethacrolein = (8.7 ꢀ 4.1) ꢁ 107 Mꢂ1 sꢂ0 1
,
entropy are much larger than reported in Gligorovski et al.
(2009).46 In the preexponential factors, a difference of two
orders of magnitude can be found which explains the difference
0
= (8.8 ꢀ 1.9) ꢁ 107 Mꢂ1
s
ꢂ1, kacrylic
=
acid
0
kmethyl
vinyl ketone
7
(9.2 ꢀ 0.6) ꢁ 107 Mꢂ1 sꢂ1 and kacrylate = (9.3 ꢀ 0.4) ꢁ 10 Mꢂ1 sꢂ1
,
0
respectively, over the whole temperature range (acrylate: in activation entropy. Comparing the activation parameters for
278 r T r 308 K) can be given since no or only small the reaction of methacrolein + OH to the other reactions that
T-dependencies of the rate constant were observed.
were investigated here, the values of this work match quite well.
3.2.3 Hydroxyl radical OH. In the present study, all reac- As can be seen from Table 3, all reactions studied in the present
tions of the unsaturated compounds with hydroxyl radicals work show a negative entropy change, the slightly positive value
were studied dependent on temperature. Fig. 3 (top) shows the of Gligorovski et al. (2009) is rather unusual and hence it is
Arrhenius plots for both forms of acrylic acid. As observed before, concluded that the newly determined preexponential factor is
acrylate reacts slightly faster in the considered temperature range more realistic than the older value.46
6262 | Phys. Chem. Chem. Phys., 2014, 16, 6257--6272
This journal is © the Owner Societies 2014

View Article Online
PCCP
Paper
according to the formula of Tyn and Calus using the critical
volume Vc at the boiling point (eqn (5)).68,69 To obtain the molar
volume at measurement temperature, the approach of the ideal gas
(Vm/T = constant) was utilised. The compounds’ radii are calculated
using the procedure of Kojima and Bard (1975) (eqn (6)).70
1.048
Vm = 0.285 ꢁ Vc
(5)
(6)
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
3 ꢁ Vm
3
r ¼
4 ꢁ p ꢁ NA
Table 4 lists the molar volumes Vm, radii r and calculated
diffusion coefficients D for all reactants and radicals. The molar
volumes and radii decrease whereas the diffusion coefficients,
which are a measure of the mobility, increase. Small molecules are
more mobile than larger ones. The radicals’ radii were adopted
from Buxton et al. (1988) for OH and from Nightingale et al. (1959)
for SO4 and NO3, respectively.58,71
ꢂ
Table 5 summarises the calculated rate constants for fully
diffusion controlled reactions kdiff for each case, together with
the measured rate constants (298 K) as well as the contribution
of kdiff to the measured overall rate constant. For kdiff, an overall
Table 4 Molar volumes at 298 K, radii and diffusion coefficients for the
reactants and the oxidants
Compound/radical
Vm/cm3 molꢂ1
r/Å
D/m2 sꢂ1
Methacrolein
Methyl vinyl ketone
Methacrylic acid
Acrylic acid
83.9
76.9
67.7
55.3
46.4
61.5
26.9
3.22
3.12
2.99
2.80
2.64
2.90
2.20
1.02 ꢁ 10ꢂ9
1.05 ꢁ 10ꢂ9
1.00 ꢁ 10ꢂ9
1.16 ꢁ 10ꢂ9
1.54 ꢁ 10ꢂ9
1.30 ꢁ 10ꢂ9
2.20 ꢁ 10ꢂ9
NO3
ꢂ
SO4
OH
Fig. 3 Top: Arrhenius plots for acrylic acid (m) and acrylate (K) exposed
to OH. Bottom: Arrhenius plots for methacrylic acid (m) and methacrylate
(K) exposed to OH.
Table 5 Rate constants for a completely diffusion controlled reaction
diff, measured rates at room temperature k298K and the proportion of kdiff
k
to k298K. For those labelled as * the reversed-rates method was used,
otherwise laser flash photolysis
3.3. Diffusion limit
A reaction is controlled by diffusion if each collision leads to a
reaction and hence the diffusion of the reaction partners
towards each other constitutes the rate-determining step. The
rate constant of a diffusion controlled reaction can be calculated
using the equation of Smoluchowski:64
kdiff
/
% of kdiff
to k298K
Radical Reactant
NO3 Methacrolein
Mꢂ1 sꢂ1
k298K/Mꢂ1 sꢂ1
1.04 ꢁ 1010 (4.0 ꢀ 1.0) ꢁ 107 0.35
Methyl vinyl ketone 1.04 ꢁ 1010 (9.7 ꢀ 3.4) ꢁ 106 0.08
Methacrylic acid
Methacrylate
Acrylic acid
Acrylate
9.93 ꢁ 109 (9.2 ꢀ 1.6) ꢁ 107 0.84
9.93 ꢁ 109 (1.7 ꢀ 1.2) ꢁ 108 1.55
1.03 ꢁ 1010 (6.9 ꢀ 1.0) ꢁ 106 0.06
1.03 ꢁ 1010 (4.4 ꢀ 0.6) ꢁ 107 0.39
kD = 4 ꢁ 103 ꢁ p ꢁ NA ꢁ (Dradical + Dreactant) ꢁ (rradical + rreactant
)
(3)
ꢂ
SO4
Methacrolein
Methacrolein*
1.09 ꢁ 1010 (9.9 ꢀ 4.9) ꢁ 107 0.91
with NA being Avogadros number, D the diffusion coefficient
and r the radii of the radical and the reactant, respectively.
Diffusion coefficients D are determined using the Stokes–Einstein
relationship that has been modified by Wilke and Chang.65,66
1.09 ꢁ 1010 (5.5 ꢀ 1.4) ꢁ 107 0.51
Methyl vinyl ketone 1.08 ꢁ 1010 (1.0 ꢀ 0.2) ꢁ 108 0.92
Methyl vinyl ketone* 1.08 ꢁ 1010 (8.7 ꢀ 2.2) ꢁ 107 0.80
Methacrylic acid
Methacrylate
Acrylic acid
Acrylate
1.04 ꢁ 1010 (2.5 ꢀ 1.2) ꢁ 108 2.41
1.04 ꢁ 1010 (3.5 ꢀ 1.1) ꢁ 108 3.37
1.08 ꢁ 1010 (9.5 ꢀ 0.8) ꢁ 107 0.88
1.08 ꢁ 1010 (9.9 ꢀ 2.0) ꢁ 107 0.92
ðXMÞ0:5T
D ¼ 7:4 ꢁ 10ꢂ8
(4)
Vm0:6Z
OH
Methacrolein
1.32 ꢁ 1010 (9.4 ꢀ 0.7) ꢁ 109 76.98
in cmꢂ2 sꢂ1 with X being the association parameter referring to
the solvent (X = 2.26 for water),67 M the molar mass of the
solvent, T the absolute temperature, Vm the molar volume and
Z the dynamic viscosity of water. The molar volume is determined
Methyl vinyl ketone 1.31 ꢁ 1010 (7.3 ꢀ 0.5) ꢁ 109 56.53
Methacrylic acid
Methacrylate
Acrylic acid
Acrylate
1.26 ꢁ 1010 (1.1 ꢀ 0.1) ꢁ 1010 92.94
1.26 ꢁ 1010 (1.1 ꢀ 0.1) ꢁ 1010 95.41
1.27 ꢁ 1010 (5.1 ꢀ 0.8) ꢁ 109 40.08
1.27 ꢁ 1010 (5.9 ꢀ 0.9) ꢁ 109 46.36
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 6257--6272 | 6263

View Article Online
Paper
PCCP
error of 4.6% can be given, calculated by the exact differential
According to the calculations by the method of Benson
method (cf. ESI†). For nitrate and sulfate reactions the measured (1976),73 the BDE of about 410 kJ molꢂ1 for methacrolein and
rate constants range below the diffusion limit by up to two and a methacrylic acid indicates that hydrogen abstraction is rather
half orders of magnitude. Hence, these reactions are clearly unlikely to occur during the reaction with the radicals investi-
chemically controlled. Hydroxyl radical reactions show values gated (o3% contribution). Looking at methyl vinyl ketone,
of kdiff that are similar to the measured rate constants indicating the BDE calculated after Benson (1976)73 indicates a hydrogen
ꢂ
partly (40–96%) diffusion controlled reactions.
abstraction reaction for NO3 and SO4 (100%). The BDE
3.3.1 Bond dissociation energy. The radical reactions con- adopted from Ochando-Pardo et al. (2005)74 indicates a possible
sidered here can occur via different mechanisms: addition to a contribution of 50 and 14% for an H atom abstraction to the
double bond, H atom abstraction or electron transfer. If a overall rate constant of methacrolein and methyl vinyl ketone for
particular mechanism is suspected, reactivity correlations can nitrate radical reactions, respectively. For SO4ꢂ, Ochando-Pardo’s
give a hint of the type. A possible tool to estimate the extent to BDE indicates that the H abstraction is the most probable
which an H atom abstraction could dominate the reaction is to reaction mechanism for methacrolein, but only a 27% contri-
calculate a theoretical rate constant kH adopting Evans–Polanyi bution of H abstraction is calculated for methyl vinyl ketone.
correlations from the literature.48 Evans and Polanyi found a Due to the reversed data for both approaches no serious
`
correlation between the activation energy and the bond disso- conclusions can be drawn. Furthermore, Noziere et al. (2010)
ciation energy (BDE) of an exothermic H atom abstraction observed the formation of organosulfates during the reaction of
ꢂ
reaction.72 As there can be several H atoms to be abstracted methyl vinyl ketone and methacrolein with SO4 and suggest
in a molecule, the measured rate constant needs to be divided by the addition of the sulfate radical anion to the unsaturated
the number nH of equivalent and most weakly bound H atoms double bond of the molecules.75 This logical mechanism was
present in the molecule to obtain kH.
confirmed by a recent study of Schindelka et al. (2013).76
The small contributions (max. 25%) of H abstraction to the
overall rate constant during the reaction of unsaturated com-
pounds with OH radicals is consistent with the generally
accepted view that OH radicals rather add to the double bond
which is mentioned by Buxton et al. (2000) who assume only
small amounts of H abstraction.54
3.3.2 Energy of the highest occupied molecular orbital.
Due to the large number of substances existing in atmospheric
liquid water the development of a predictive mechanism for the
reactivity of certain classes of compounds can be helpful. One
possibility is the determination of log(k298K) as a function of
EHOMO in eV, with EHOMO being the energy of the highest
occupied molecular orbital.77 Here, the molecular geometry of
a molecule is optimised using different initial geometries and
seeking a global energy minimum. This is done by means of
quantum mechanical calculations on a semi-empirical level,
where PM3 parameterisations of the atomic wave functions are
used.78 Table 7 lists the energy of the highest occupied mole-
cular orbital of the investigated compounds calculated using
k298K
kH
¼
(7)
nH
The plot log(kH) vs. BDE considers the number of equivalent
weakly bound H atoms to be abstracted. The regression of the
data points allows suppositions of the reaction type (whether an H
atom is possibly abstracted during the reaction or not), distinguish-
ing between the investigated radicals.48 For the investigated sub-
stances, several H atoms can possibly be abstracted. As there are no
values available for the resulting increments to calculate BDEs after
Benson (1976),73 only H atoms of the methyl group have been taken
into account knowing that these are not the most easily
abstractable H atoms in the molecule (cf. Section 3.5.1). Table 6
summarises the BDEs calculated according to the incremental
method of Benson (1976)73 and adopted from Ochando-Pardo
et al. (2005)74 as well as the calculated rate constant kcalc = kH ꢁ nH
of a pure H abstraction reaction and the amount of kcalc of the
measured rate constant k298K. Usually, small BDEs indicate larger
rate constants for H abstraction reactions.
Table 6 Bond dissociation energy (BDE) calculated according to the incremental method of Benson (1976)73 and adopted from Ochando-Pardo et al.
(2005)74 and the possible amount of H abstraction reaction with the total measured rate constant. MACR = methacrolein, MVK = methyl vinyl ketone,
MAA = methacrylic acid
BDE/kJ molꢂ1
Benson
k298K/Mꢂ1 sꢂ1
kcalc/Mꢂ1 sꢂ1
k298K/kcalc/%
Ochando
This work
Benson
Ochando
Benson
Ochando
NO3
MACR
MVK
MAA
410.7
383.4
410.0
380.2
393.6
—
(4.0 ꢀ 1.0) ꢁ 107
(9.7 ꢀ 3.4) ꢁ 106
(9.2 ꢀ 1.6) ꢁ 107
4.4 ꢁ 104
1.1 ꢁ 107
5.1 ꢁ 104
2.0 ꢁ 107
1.4 ꢁ 106
—
0.1
107.9
0.1
50.2
14.2
—
ꢂ
SO4
OH
MACR
MVK
MAA
410.7
383.4
410.0
380.2
393.6
—
(9.9 ꢀ 4.9) ꢁ 107
(1.0 ꢀ 0.2) ꢁ 108
(2.5 ꢀ 1.2) ꢁ 108
(9.4 ꢀ 0.7) ꢁ 109
(7.3 ꢀ 0.5) ꢁ 109
(1.1 ꢀ 0.1) ꢁ 1010
2.9 ꢁ 106
1.1 ꢁ 108
3.2 ꢁ 106
1.4 ꢁ 108
1.7 ꢁ 109
1.5 ꢁ 108
1.6 ꢁ 108
2.8 ꢁ 107
—
3.0
105.3
1.3
163.1
27.8
—
MACR
MVK
MAA
410.7
383.4
410.0
380.2
393.6
—
2.4 ꢁ 109
6.8 ꢁ 108
—
1.5
23.8
1.4
25.0
9.4
—
6264 | Phys. Chem. Chem. Phys., 2014, 16, 6257--6272
This journal is © the Owner Societies 2014

View Article Online
PCCP
Paper
reactions in aqueous solution. The phenols studied by Schaefer82
in the liquid phase exhibit a similar behaviour as observed for
the unsaturated compounds in this work meaning a rather small
dependence of the rate constant on the energy of the highest
occupied molecular orbital. The data from Kind83 however,
show a significantly steeper regression line meaning that small
variations in EHOMO lead to large differences in the rate con-
stants in the gas phase.
Table 7 Energy of the highest occupied molecular orbital EHOMO of the
investigated compounds calculated using the program HyperChem
Release 5.7, Hypercube Inc
Compound
EHOMO/eV
Methacrolein
ꢂ10.48
ꢂ10.72
ꢂ10.56
ꢂ11.15
Methyl vinyl ketone
Methacrylic acid
Acrylic acid
Schaefer (2007) indicated that phenols either react by
H abstraction or by electron transfer with nitrate radicals.82
From Fig. 4, the values of EHOMO obtained in this work are
rather large (in the negative) and therefore imply that electron
transfer is presumably even less probable than for phenols.
Relating log(k298K) to EHOMO was initially developed for the
gas phase and was applied here to check for another possibility
of making better predictions of rate constants. To further
develop this approach, a larger number of measurements are
necessary to improve the regression and minimise errors.
Apparently, the distinction between the different structures of
the compounds is also important.
the program HyperChem Release 5.7, Hypercube Inc. There is
no literature data available to validate the obtained values for the
investigated compounds. Nevertheless, EHOMO of the phenols
simulated with the above mentioned method compare very well
with data reported by Hayashi et al. (1999, R = 0.9990, n = 6).79
Based on this, we are regarding the EHOMO for the unsaturated
compounds investigated in the present studies valid as well.
According to Koopmans’ theorem the energy of the highest
occupied molecular orbital is equal to the negative ionisation
energy:80
EHOMO = ꢂEi
(8)
3.4. Oxidation products of the reactions of methacrolein or
methyl vinyl ketone + OH
If log(k298K) is plotted vs. EHOMO in eV, the following correla-
tions can be found for each radical:
The reaction of methacrolein and methyl vinyl ketone with
OH radicals results, besides carboxylic acids and higher mole-
cular weight compounds, in the formation of smaller C2 or C3
carbonyl compounds namely methylglyoxal, hydroxyacetone,
glycolaldehyde and pyruvic acid as well as polyfunctional C4
compounds.
3.4.1 Methacrolein. The time profiles in Fig. 5 show that
methacrolein was degraded fast within the first five laser pulses.
As products were present in higher amounts, the methacrolein
degradation was slowed down because of competing reactions of
OH radicals and oxidation products.
Hydroxyacetone and methylglyoxal are the first major oxida-
tion products from methacrolein. Both had their concentration
maximum at 10 laser pulses whereas hydroxyacetone had a
broader one over the course of 5 laser pulses. Furthermore,
hydroxyacetone was degraded faster than methylglyoxal. Another
first generation oxidation product was glycolaldehyde which
NO3: log(k298K/Mꢂ1 sꢂ1) = (1.4 ꢀ 3.5) ꢁ EHOMO/eV + (22.4 ꢀ 37.6)
R = 0.77
(9)
SO4: log(k298K/Mꢂ1 sꢂ1) = (0.3 ꢀ 2.0) ꢁ EHOMO/eV + (10.8 ꢀ 21.4)
R = 0.36
(10)
Due to the limited number of measurement points (only four)
and the high uncertainty, a statement whether electron transfer
occurs is not possible. The considerations were not made for the
hydroxyl radicals since OH radicals abstract an H atom or add to
the molecule but will not react via electron transfer.81
A comparison with existing data in the literature for reac-
tions with the nitrate radical can be seen in Fig. 4. While open
symbols denote gas-phase reactions, solid symbols represent
Fig. 4 Plots of log (k2nd) vs. EHOMO in eV for nitrate radical reactions with
different classes of substances.82,83
Fig. 5 Concentration profiles of methacrolein and the oxidation products
depending on the number of laser pulses.
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 6257--6272 | 6265

View Article Online
Paper
PCCP
reached the concentration maximum after 3 laser pulses. The
formation of pyruvic acid started with the degradation of
methylglyoxal after 10 laser pulses and therefore it is considered
a secondary oxidation product.
Polyfunctional C4 compounds which have been determined
before by Liu et al. (2009)42 were not quantified but detected
in the present study. The single derivatised 2,3-dihydroxy-
methylpropanal was detected depending on the number of
laser pulses (cf. Fig. 7). Similar to the profile of glycolaldehyde
it showed a concentration maximum after 4 laser pulses. Liu
et al. also measured 2-hydroxy-2-methyl-malonaldehyde which
was not detected here.42 It probably formed the double deriva-
tised 2,4-dinitrophenylhydrazine product. These osazones are
known to be insoluble in common solvents.84–86 As the forma-
tion of these two species should not be the main reaction
pathway it is likely that they were generated in lower amounts
than hydroxyacetone and methylglyoxal.
3.4.2 Methyl vinyl ketone. For the first five laser pulses
methyl vinyl ketone showed a fast decay similar to the one of
methacrolein. However, with ongoing formation of reaction
products, methyl vinyl ketone decay did not cease like the
methacrolein concentration resulting in a complete depletion
of methyl vinyl ketone after 50 laser pulses. The main first
generation oxidation products were glycolaldehyde and methyl-
glyoxal as can be seen in Fig. 6. These substances were
produced in lower amounts than the oxidation products of
Fig. 7 Concentration profiles of 2,3-dihydroxy-methylpropanal (DHMP),
oxidation product of methacrolein and 1-hydroxy-dimethylglyoxal (HDMG)
or 2-hydroxy-3-oxobutanal (HOB) and 3,4-dihydroxybutan-2-one (DHB),
oxidation products of methyl vinyl ketone, depending on the number of
laser pulses.
vinyl ketone and was found to show similar intensities like the
2,3-dihydroxy-methylpropanal.
3.5. Reaction mechanisms of the reactions of methacrolein or
methyl vinyl ketone + OH
methacrolein. Glycolaldehyde reached its concentration maxi- An explanation for the formation of these oxidation products
mum after 3 laser pulses, whereas methylglyoxal was formed utilises known aqueous phase peroxyl radical reaction chemi-
until 10 laser pulses. A secondary oxidation product was pyruvic stry. In general, OH radicals react with certain reactants under
acid which was detected after 2 laser pulses and showed a H abstraction or addition to a double bond. Both cases lead
strong increase up to 50 laser pulses. From the oxidation of to the formation of carbon-centred radicals which can react
methyl vinyl ketone polyfunctional C4 compounds were also further with oxygen to form peroxyl radicals. In the aqueous
formed (Fig. 7). Compared to the oxidation of methacrolein, phase the latter can recombine forming a tetroxide as a short-
single derivatised dicarbonyl compounds were detected. Two lived intermediate which then decomposes. For the structures
peaks were found for the latter ones, where two different of a tetroxide six- or five-membered ring structures have been
constitutions (1-hydroxy-dimethylglyoxal and 2-hydroxy-3- suggested but are still not understood.87,88 The decomposition
oxobutanal) with two stereoisomers each are possible. Both of the tetroxides can proceed via three different pathways.
signals could not be attributed to the different species due to Either they can decompose to form a carbonyl compound and
missing standard compounds. As for methacrolein, the diol an alcohol under elimination of oxygen, as in the gas phase or,
compound 3,4-dihydroxybutan-2-one was detected for methyl different from gas phase chemistry, they form two carbonyl
compounds and hydrogen peroxide. In the third pathway two
alkoxy radicals and oxygen just as in the radical-retaining
pathway in gas phase chemistry are produced. Depending on
their structure alkoxy radicals are quite unstable and decom-
pose to form smaller compounds and other radicals by C–C
bond cleavage or isomerise and react further with oxygen to
form more oxidized compounds.89 The formation of alkoxy
radicals is the only possible degradation reaction for a tetroxide
formed by two tertiary peroxyl radicals.90
3.5.1 Methacrolein
Initial oxidation steps. Methacrolein is degraded by the
addition of the OH radical to the double bond or hydrogen
abstraction of the aldehydic hydrogen atom. From gas phase
oxidation experiments it is known that 50% of the oxidation
proceeds via H abstraction of the aldehydic hydrogen atom.22,28
Fig. 6 Concentration profiles of methyl vinyl ketone and the oxidation
products depending on the number of laser pulses.
However, in the aqueous phase this reaction seems to be of
6266 | Phys. Chem. Chem. Phys., 2014, 16, 6257--6272
This journal is © the Owner Societies 2014

View Article Online
PCCP
Paper
minor relevance as Liu et al. (2009)42 stated that it contributes in the formation of three different tetroxides which then can
only 4.8% to the OH radical initiated methacrolein degradation. decompose to form either functionalised C4 or smaller C3
Comparing the two possible positions for the addition, the compounds as first generation products.
terminal carbon should be the favoured one because of the less
sterical hindrance and the more stable tertiary carbon-centred
Identified and quantified oxidation product: hydroxyacetone.
radical that is formed. In line with this, in the gas phase oxidation As described in the study of Liu et al. (2009), the primary formation
the addition takes place mainly at the terminal carbon (85%).22,28 of functionalised C3 compounds like hydroxyacetone and methyl-
For the aqueous phase oxidation of methacrolein the contribution glyoxal can only be explained by the decomposition of alkoxy
of the inner and terminal addition was not determined so far, as radicals involving a C–C bond cleavage.42 The only reaction to form
several oxidation products are formed on both reaction pathways. hydroxyacetone as a primary oxidation product is the decomposi-
From the addition of an OH radical to the inner or terminal tion of a tertiary alkoxy radical (Scheme 1, reaction pathway D).
carbon atom hydroxy-alkyl radicals are produced which react
further with oxygen to form a tertiary or primary peroxyl radical.
Identified and quantified oxidation product: methylglyoxal.
Scheme 1 shows the two different peroxyl radicals that can Methylglyoxal can be produced by the decomposition of a primary
be formed. The combination of two peroxyl radicals can result or tertiary alkoxy radical (Scheme 1, reaction pathways A and D),
Scheme 1 Mechanism of methacrolein degradation (observed products in blue).
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 6257--6272 | 6267

View Article Online
Paper
PCCP
whereas the formation of a primary alkoxy radical is of minor
importance.42 A secondary source of methylglyoxal is the reaction
3.5.2 Methyl vinyl ketone
Initial oxidation steps. As with methacrolein, OH radicals will
of hydroxyacetone with OH radicals. Due to an H atom abstrac- preferentially add to one carbon of the double bond. Although
tion and subsequent reaction with oxygen an a-hydroxy-peroxyl the terminal carbon should be again the favoured one due to
radical is produced which can stabilise under elimination of an the formation of the more stable secondary carbon-centred
HO2 radical to form methylglyoxal (R5) and (R6).
radical, gas phase oxidation experiments have shown that the
addition of OH radicals to the inner carbon atom can account
for 28% of the methyl vinyl ketone oxidation (Scheme 2).29 As
for methacrolein, no branching ratio was determined so far for
aqueous phase oxidation as some oxidation products are
formed on both reaction pathways.
CH3C(O)CH2OH + OH + O2 - CH3C(O)CH(OH)OO + H2O
(R5)
CH3C(O)CH(OH)OO - CH3C(O)CHO + HO2 (R6)
Again, three different tetroxides can be formed by the recombi-
nation of peroxyl radicals.
Identified and quantified oxidation product: pyruvic acid.
Methylglyoxal is the precursor compound for the production of
pyruvic acid.34,35,91 As it is mainly present in its hydrated form,92
another hydrogen atom is abstracted from the dihydroxy-
substituted carbon. After reaction with oxygen the a-dihydroxy-
peroxyl radical can decompose under HO2-elimination to form
pyruvic acid (R7)–(R9).
Identified and quantified oxidation product: glycolaldehyde.
The main oxidation products methylglyoxal and glycolaldehyde
can be explained by the decomposition of alkoxy radicals as the
primary formation of these C3 and C2 compounds require a
reaction involving a C–C bond cleavage. The formation of the
glycolaldehyde is shown in Scheme 2 on reaction pathway D,
where a secondary alkoxy radical decomposes to form glycol-
aldehyde by the loss of an acyl radical.
CH3C(O)CHO + H2O " CH3C(O)CH(OH)2
(R7)
CH3C(O)CH(OH)2 + OH + O2 - CH3C(O)C(OH)2OO + H2O
(R8)
Identified and quantified oxidation product: methylglyoxal. As
with glycolaldehyde, only one reaction mechanism explains the
formation of methylglyoxal. The C–C bond cleavage of the secondary
peroxyl radical can occur also for the other adjacent C–C bonds to
form methylglyoxal (Scheme 2, reaction pathway D).
CH3C(O)C(OH)2OO - CH3C(O)COOH + HO2 (R9)
Identified oxidation product: 2,3-dihydroxy-methylpropanal.
The formation of 2,3-dihydroxy-methylpropanal can only be
explained with the decomposition of a tetroxide formed by two
primary peroxyl radicals. Following a decomposition with loss of
oxygen, compounds with an alcohol and carbonyl function,
in this case the polyfunctional C4 compounds 2,3-dihydroxy-
methylpropanal and 2-hydroxy-2-methyl-malonaldehyde, are formed
(Scheme 1, reaction pathway B). Another route for the 2-hydroxy-
2-methyl-malonaldehyde formation is the decomposition of the
tetroxide under loss of hydrogen peroxide whereby two 2-hydroxy-
2-methyl-malonaldehyde molecules can be formed (Scheme 1,
reaction pathway C). The reaction of a primary alkoxy radical
(Scheme 1, reaction pathway A) is possibly of minor importance
and leads to the formation of a carbon centred hydroxy-alkyl
radical via intramolecular H transfer. The latter can further react
with oxygen to form an a-hydroxy-peroxyl radical. As described
above these radicals can stabilise to form a carbonyl function,
in this case the 2-hydroxy-2-methyl-malonaldehyde.
Identified and quantified oxidation product: pyruvic acid. Again,
methylglyoxal can be further degraded to pyruvic acid by attack of
OH radicals, a primary peroxyl radical formation and HO2 elimina-
tion, as described above (Scheme 2, reaction pathway D).
Identified oxidation products: 1-hydroxy-dimethylglyoxal, 2-hydroxy-
3-oxobutanal and 3,4-dihydroxybutan-2-one. Different to meth-
acrolein all tetroxides can decompose to form polyfunctional C4
compounds, namely 1-hydroxy-dimethylglyoxal, 2-hydroxy-3-
oxobutanal and 3,4-dihydroxybutan-2-one. Scheme 2 shows the
simplified reaction pathways A and C. In principle a tetroxide
can decompose to form 2-hydroxy-3-oxobutanal or 1-hydroxy-
dimethylglyoxal together with 3,4-dihydroxybutan-2-one under
loss of oxygen (Scheme 2, reaction pathway A). The other
possible decomposition is the loss of hydrogen peroxide and
formation of either two molecules of 2-hydroxy-3-oxobutanal or
1-hydroxy-dimethylglyoxal. Possible reactions, but of minor
importance, are the isomerization of an alkoxy radical with
subsequent reaction with oxygen and HO2 elimination to form
either 2-hydroxy-3-oxobutanal or 1-hydroxy-dimethylglyoxal on
reaction pathways B and D in Scheme 2.
Identified and quantified oxidation product: glycolaldehyde.
The observation of glycolaldehyde as a primary oxidation
product of methacrolein was not expected. The formation of
this C2 compound requires the cleavage of two C–C-bonds. Liu
et al. (2009) proposed a simultaneous cleavage of two C–C
bonds for the formation of acetic acid, but the mechanism
for the formation of glycolaldehyde is not yet understood
(Scheme 1, reaction pathway E).42 The direct photolysis of
3.6. Product yields of the reactions of methacrolein or methyl
vinyl ketone + OH
methacrolein as a source of glycolaldehyde can be excluded The ratio of the formation of primary oxidation products depen-
because of the low absorption coefficient at l = 248 nm of dent on the degradation of the reactants – the molar yield – was
e = 126 Mꢂ1 cmꢂ1 of the reactant (Table 1) which cannot explain calculated. As can be seen in Table 8 the yields confirm the range
this high yield of glycolaldehyde.
of results found by Liu et al. and Zhang et al.,42,43 respectively.
6268 | Phys. Chem. Chem. Phys., 2014, 16, 6257--6272
This journal is © the Owner Societies 2014

View Article Online
PCCP
Paper
Scheme 2 Mechanism of methyl vinyl ketone degradation (observed products in blue).
Table 8 Yields of reaction products
Overall, only 31 and 15% of the initial concentrations of
methacrolein and methyl vinyl ketone were found in the
products. The polyfunctionalised C4 compounds may add
significantly to the organic mass but quantification was not
possible due to a lack of standards. Molar yields obtained with
standards of hydroxyacetone for 2,3-dihydroxy-methylpropanal
and 3,4-dihydroxybutan-2-one as well as with standards of
methylglyoxal for 1-hydroxy-dimethylglyoxal and 2-hydroxy-
3-oxobutanal were 20–25% and 25–47%. As these yields are
significantly higher than obtained by Liu et al. (2009) further
Methyl vinyl
ketone + OH
Methacrolein + OH
This work Ref.
This work
Ref.
Methylglyoxal
0.099 ꢀ 0.020 0.060–0.091 42 0.052 ꢀ 0.010 0.045 43
0.075 43
Hydroxyacetone 0.162 ꢀ 0.008 0.098–0.150 42
—
—
Glycolaldehyde 0.051 ꢀ 0.004 —
0.111 ꢀ 0.004 —
However, the present study indicates that the methylglyoxal studies are necessary to clarify their contribution to the mass
yields are somewhat higher than published before for both the balance and degradation by OH radicals.42
methacrolein and methyl vinyl ketone oxidation. It is known
Finally, branching ratios can be suggested for most of the
from gas phase chemistry of methyl vinyl ketone29 that glycol- reaction pathways in Schemes 1 and 2 with the calculated and
aldehyde is formed and it was measured here as the main estimated yields. For the OH radical initiated degradation of
primary oxidation product with a yield of about 10%. The methacrolein (Scheme 1) the ratios for the pathways A, B, D and E
production of 5% glycolaldehyde during the oxidation of are 0.05, 0.2–0.25, 0.21 and 0.05, assuming equimolar formation
methacrolein was unexpected and cannot be explained so far of methylglyoxal on both suggested pathways. The branching
by peroxyl radical chemistry or photolysis.
ratios for the reaction pathways A, C and D of the degradation of
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 6257--6272 | 6269

View Article Online
Paper
PCCP
methyl vinyl ketone with OH radicals are 0.25, o0.2 and 0.16 supported by the ACCENT network. Further support of this
with all branching ratios referring to the degradation of the work by the PEGASOS project (FP7-ENV-2010-265148) is also
mother compounds methacrolein and methyl vinyl ketone.
gratefully acknowledged. Special acknowledgements go to
Bastian Stieger, Tina Hertel, Markus Hartmann, and Christian
Weller for their help with the laboratory work, their encourage-
ment, expert advice and proof-reading. Rafal Szmigielski and
Krzysztof J. Rudzinski gratefully acknowledge the financial
support by the European Commission through the Marie Curie
European Reintegration Grant (ISOMASSKIN project), and by
the Polish Ministry of Science and Higher Education through
the science funds 2011–2012 granted for the realisation of the
international co-financed project.
4. Conclusion
The results of the present study unsurprisingly exhibit the highest
ꢂ
reactivity of the OH radical, followed by SO4 and NO3 radicals
independent of the reactant. Thus, OH radical reactions will
constitute the main degradation pathway for methacrolein, methyl
vinyl ketone, methacrylic and acrylic acid in the tropospheric
aqueous phase. Some investigated reactions with OH radicals are
at least partly diffusion controlled, while others are fully diffusion
limited (methacrylic acid and methacrylate). Reactions with the
nitrate radical and the sulfate radical anion are not diffusion
controlled at all. It should be noted that to fully describe aqueous
phase degradation processes of organics non-radical reactions
should also be considered.93
Another issue occurs comparing the measured rate constants
of this work with existing gas phase rate constants. The reactions
of OH radicals with either methacrolein or methyl vinyl ketone
are roughly as fast as their gas phase analogues. Regarding the
nitrate radical reactions, both species react faster in the aqueous
phase than in the gas phase. No comparison can be given for the
acids because of missing kinetic gas phase data.
The findings of product studies suggest the addition of OH
radicals to unsaturated molecules in the presence of O2, forming a
peroxyl radical. Studies on the formation of carbonyl compounds
from the OH radical induced oxidation of methacrolein and
methyl vinyl ketone support the formation yields of methylglyoxal
and hydroxyacetone reported in former studies. Furthermore, the
production of glycolaldehyde was observed for the oxidation of
both precursor compounds whereas the yield was highest for
the oxidation of methyl vinyl ketone. Highly functionalised C4
compounds reported so far only for methacrolein were also
found to form during the oxidation of methyl vinyl ketone. Due
to missing standards a quantification of these compounds will
be the objective of the forthcoming work.
References
1 A. Guenther, T. Karl, P. Harley, C. Wiedinmyer, P. I. Palmer
and C. Geron, Atmos. Chem. Phys., 2006, 6, 3181–3210.
2 T. D. Sharkey, A. E. Wiberley and A. R. Donohue, Ann. Bot.,
2008, 101, 5–18.
3 W. L. Chameides, F. Fehsenfeld, M. O. Rodgers, C. Cardelino,
J. Martinez, D. Parrish, W. Lonneman, D. R. Lawson, R. A.
Rasmussen, P. Zimmerman, J. Greenberg, P. Middleton and
T. Wang, J. Geophys. Res., [Atmos.], 1992, 97, 6037–6055.
4 F. Paulot, J. D. Crounse, H. G. Kjaergaard, A. Kurten, J. M.
St Clair, J. H. Seinfeld and P. O. Wennberg, Science, 2009,
325, 730–733.
5 J. Peeters and J. F. Muller, Phys. Chem. Chem. Phys., 2010, 12,
14227–14235.
6 J. Peeters, T. L. Nguyen and L. Vereecken, Phys. Chem. Chem.
Phys., 2009, 11, 5935–5939.
7 X. R. Ren, J. R. Olson, J. H. Crawford, W. H. Brune, J. Q. Mao,
R. B. Long, Z. Chen, G. Chen, M. A. Avery, G. W. Sachse,
J. D. Barrick, G. S. Diskin, L. G. Huey, A. Fried, R. C. Cohen,
B. Heikes, P. O. Wennberg, H. B. Singh, D. R. Blake and
R. E. Shetter, J. Geophys. Res., [Atmos.], 2008, 113, D05310.
8 D. Taraborrelli, M. G. Lawrence, J. N. Crowley, T. J. Dillon,
S. Gromov, C. B. M. Gross, L. Vereecken and J. Lelieveld,
Nat. Geosci., 2012, 5, 190–193.
The kinetic measurements undertaken provide new data to
improve existing prediction tools for rate constants in the atmo-
spheric aqueous phase. Furthermore, the product studies evidence
the formation of various atmospherically relevant carbonyl com-
pounds like methylglyoxal and pyruvic acid as well as transient
radicals that take part in chemical transformation processes in
aqueous particles and rain droplets. All findings from this work
are aimed to be implemented in the chemical aqueous phase
radical mechanism CAPRAM3.0 to gain a better understanding of
the impact of the introduced reactions on the budget of the
compounds of interest within the atmosphere.
¨
9 O. Boge, Y. Miao, A. Plewka and H. Herrmann, Atmos.
Environ., 2006, 40, 2501–2509.
10 M. Claeys, B. Graham, G. Vas, W. Wang, R. Vermeylen,
V. Pashynska, J. Cafmeyer, P. Guyon, M. O. Andreae,
P. Artaxo and W. Maenhaut, Science, 2004, 303, 1173–1176.
11 M. Claeys, W. Wang, A. C. Ion, I. Kourtchev, A. Gelencser
and W. Maenhaut, Atmos. Environ., 2004, 38, 4093–4098.
12 M. Hallquist, J. C. Wenger, U. Baltensperger, Y. Rudich,
D. Simpson, M. Claeys, J. Dommen, N. M. Donahue,
C. George, A. H. Goldstein, J. F. Hamilton, H. Herrmann,
T. Hoffmann, Y. Iinuma, M. Jang, M. E. Jenkin, J. L.
Jimenez, A. Kiendler-Scharr, W. Maenhaut, G. McFiggans,
T. F. Mentel, A. Monod, A. S. H. Prevot, J. H. Seinfeld,
J. D. Surratt, R. Szmigielski and J. Wildt, Atmos. Chem. Phys.,
2009, 9, 5155–5236.
Acknowledgements
Part of the present study was performed with the project MISOX
funded by the German Research Foundation (DFG) under HE 13 M. Kalberer, D. Paulsen, M. Sax, M. Steinbacher,
3086/13-1. Some of the studies at TROPOS were also financially
J. Dommen, A. S. H. Prevot, R. Fisseha, E. Weingartner,
6270 | Phys. Chem. Chem. Phys., 2014, 16, 6257--6272
This journal is © the Owner Societies 2014

View Article Online
PCCP
Paper
V. Frankevich, R. Zenobi and U. Baltensperger, Science, 36 K. E. Altieri, S. P. Seitzinger, A. G. Carlton, B. J. Turpin, G. C.
2004, 303, 1659–1662. Klein and A. G. Marshall, Atmos. Environ., 2008, 42, 1476–1490.
14 T. E. Kleindienst, E. O. Edney, M. Lewandowski, 37 A. G. Carlton, B. J. Turpin, H. J. Lim, K. E. Altieri and
J. H. Offenberg and M. Jaoui, Environ. Sci. Technol., 2006,
40, 3807–3812.
S. Seitzinger, Geophys. Res. Lett., 2006, 33, L06822.
38 A. G. Carlton, C. Wiedinmyer and J. H. Kroll, Atmos. Chem.
Phys., 2009, 9, 4987–5005.
39 L. T. Iraci, B. M. Baker, G. S. Tyndall and J. J. Orlando,
J. Atmos. Chem., 1999, 33, 321–330.
40 D. van Pinxteren, A. Plewka, D. Hoffmann, K. Mu¨ller,
H. Kramberger, B. Svrcina, K. Bachmann, W. Jaeschke,
S. Mertes, J. L. Collett and H. Herrmann, Atmos. Environ.,
2005, 39, 4305–4320.
15 T. E. Kleindienst, M. Lewandowski, J. H. Offenberg, M. Jaoui
and E. O. Edney, Geophys. Res. Lett., 2007, 34, L01805.
16 J. H. Kroll, N. L. Ng, S. M. Murphy, R. C. Flagan and
J. H. Seinfeld, Geophys. Res. Lett., 2005, 32, L18808.
17 J. H. Kroll, N. L. Ng, S. M. Murphy, R. C. Flagan and
J. H. Seinfeld, Environ. Sci. Technol., 2006, 40, 1869–1877.
18 J. D. Surratt, M. Lewandowski, J. H. Offenberg, M. Jaoui,
T. E. Kleindienst, E. O. Edney and J. H. Seinfeld, Environ. Sci. 41 Z. Liu, L. Y. Wu, T. H. Wang, M. F. Ge and W. G. Wang,
Technol., 2007, 41, 5363–5369. J. Phys. Chem. A, 2012, 116, 437–442.
19 J. D. Surratt, S. M. Murphy, J. H. Kroll, N. L. Ng, 42 Y. Liu, I. El Haddad, M. Scarfogliero, L. Nieto-Gligorovski,
L. Hildebrandt, A. Sorooshian, R. Szmigielski, R. Vermeylen,
W. Maenhaut, M. Claeys, R. C. Flagan and J. H. Seinfeld,
J. Phys. Chem. A, 2006, 110, 9665–9690.
B. Temime-Roussel, E. Quivet, N. Marchand, B. Picquet-
Varrault and A. Monod, Atmos. Chem. Phys., 2009, 9,
5093–5105.
20 J. H. Kroll and J. H. Seinfeld, Atmos. Environ., 2008, 42, 43 X. Zhang, Z. M. Chen and Y. Zhao, Atmos. Chem. Phys., 2010,
3593–3624. 10, 9551–9561.
21 S. N. Pandis, S. E. Paulson, J. H. Seinfeld and R. C. Flagan, 44 M. Chin and P. H. Wine, J. Photochem. Photobiol., A, 1992,
Atmos. Environ., Part A, 1991, 25, 997–1008. 69, 17–25.
22 E. C. Tuazon and R. Atkinson, Int. J. Chem. Kinet., 1990, 22, 45 P. G. de Semainville, D. Hoffmann, C. George and H. Herrmann,
591–602. Phys. Chem. Chem. Phys., 2007, 9, 1247–1260.
23 A. W. H. Chan, M. N. Chan, J. D. Surratt, P. S. Chhabra, 46 S. Gligorovski, D. Rousse, C. H. George and H. Herrmann,
C. L. Loza, J. D. Crounse, L. D. Yee, R. C. Flagan, P. O. Int. J. Chem. Kinet., 2009, 41, 309–326.
Wennberg and J. H. Seinfeld, Atmos. Chem. Phys., 2010, 10, 47 H. Herrmann, D. Hoffmann, T. Schaefer, P. Brauer and
7169–7188. A. Tilgner, ChemPhysChem, 2010, 11, 3796–3822.
24 B. Chuong and P. Stevens, Int. J. Chem. Kinet., 2004, 36, 48 D. Hoffmann, B. Weigert, P. Barzaghi and H. Herrmann,
12–25. Phys. Chem. Chem. Phys., 2009, 11, 9351–9363.
25 B. Chuong and P. S. Stevens, J. Phys. Chem. A, 2003, 107, 49 J. Ziajka and W. Pasiuk-Bronikowska, Atmos. Environ., 2003,
2185–2191. 37, 3913–3922.
26 J. D. Crounse, H. C. Knap, K. B. Ornso, S. Jorgensen, 50 J. Ziajka and W. Pasiuk-Bronikowska, Atmos. Environ., 2005,
´
F. Paulot, H. G. Kjaergaard and P. O. Wennberg, J. Phys.
Chem. A, 2012, 116, 5756–5762.
39, 1431–1438.
51 J. Ziajka and K. J. Rudzinski, Environ. Chem., 2007, 4, 355–363.
´
ˇ
27 N. L. Ng, J. H. Kroll, M. D. Keywood, R. Bahreini, 52 I. Grgic, B. Podkrajsek, P. Barzaghi and H. Herrmann,
V. Varutbangkul, R. C. Flagan, J. H. Seinfeld, A. Lee and Atmos. Environ., 2007, 41, 9187–9194.
A. H. Goldstein, Environ. Sci. Technol., 2006, 40, 2283–2297. 53 C. L. Clifton and R. E. Huie, Int. J. Chem. Kinet., 1989, 21,
28 J. J. Orlando, G. S. Tyndall and S. E. Paulson, Geophys. Res.
Lett., 1999, 26, 2191–2194.
29 E. C. Tuazon and R. Atkinson, Int. J. Chem. Kinet., 1989, 21,
1141–1152.
677–687.
54 G. V. Buxton, G. A. Salmon and J. E. Williams, J. Atmos.
Chem., 2000, 36, 111–134.
55 J. Lilie and A. Henglein, Ber. Bunsen-Ges. Phys. Chem., 1970,
74, 388–393.
30 D. O. De Haan, A. L. Corrigan, M. A. Tolbert, J. L. Jimenez,
S. E. Wood and J. J. Turley, Environ. Sci. Technol., 2009, 43, 56 P. Maruthamuthu, Makromol. Chem., Rapid Commun., 1980,
8184–8190. 1, 23–25.
31 B. Ervens, A. G. Carlton, B. J. Turpin, K. E. Altieri, 57 H. Herrmann, A. Reese and R. Zellner, J. Mol. Struct., 1995,
S. M. Kreidenweis and G. Feingold, Geophys. Res. Lett.,
2008, 35, L02816.
32 H. J. Lim, A. G. Carlton and B. J. Turpin, Environ. Sci.
Technol., 2005, 39, 4441–4446.
33 Y. B. Lim, Y. Tan, M. J. Perri, S. P. Seitzinger and
B. J. Turpin, Atmos. Chem. Phys., 2010, 10, 10521–10539.
34 Y. B. Lim, Y. Tan and B. J. Turpin, Atmos. Chem. Phys.
Discuss., 2013, 13, 4687–4725.
35 T. Schaefer, J. Schindelka, D. Hoffmann and H. Herrmann,
J. Phys. Chem. A, 2012, 116, 6317–6326.
348, 183–186.
58 G. V. Buxton, C. L. Greenstock, W. P. Helman and A. B. Ross,
J. Phys. Chem. Ref. Data, 1988, 17, 513–886.
59 G. V. Buxton, J. Wang and G. A. Salmon, Phys. Chem. Chem.
Phys., 2001, 3, 2618–2621.
´
60 P. G. de Semainville, B. D’Anna and C. George, Z. Phys.
Chem., 2010, 224, 1247.
61 B. Ervens, S. Gligorovski and H. Herrmann, Phys. Chem.
Chem. Phys., 2003, 5, 1811–1824.
62 H. Herrmann, Chem. Rev., 2003, 103, 4691–4716.
This journal is ©the Owner Societies 2014
Phys. Chem. Chem. Phys., 2014, 16, 6257--6272 | 6271

View Article Online
Paper
PCCP
63 K. J. Laidler, Chemical Kinetics, Pearson Education, New York, 79 M. Hayashi, Y. Nakamura, K. Higashi, H. Kato, F. Kishida
3rd edn, 1987.
and H. Kaneko, Toxicol. in Vitro, 1999, 13, 915–922.
80 P. W. Atkins, Physikalische Chemie, VCH Verlagsgesellschaft
mbH, Weinheim, 1990.
81 C. von Sonntag, Water Sci. Technol., 2008, 58, 1015–1021.
82 T. Schaefer, Diploma thesis, Leibniz Institute for Tropo-
spheric Research, Leipzig, 2007.
64 M. Smoluchowski, Z. Phys. Chem., 1917, 62, 129–168.
65 P. E. Liley, R. C. Reid and E. Buck, Chemical engineers’
handbook, McGraw-Hill, New York, 4th edn, 1963.
66 C. R. Wilke and P. Chang, AIChE J., 1955, 1, 264–270.
67 W. Hayduk and H. Laudie, AIChE J., 1974, 20, 611–615.
¨
68 K. G. Joback and R. C. Reid, Chem. Eng. Commun., 1987, 57, 83 I. Kind, T. Berndt and O. Boge, Chem. Phys. Lett., 1998, 288,
233–243. 111–118.
69 B. E. Poling, J. M. Prausnitz and J. O’ Connell, The Properties of 84 L. G. Egyu¨d, J. A. McLaughlin and A. Szent-Gyorgyi,
Gases and Liquids, McGraw-Hill Professional, 6th edn, 2000. Proc. Natl. Acad. Sci. U. S. A., 1967, 57, 1422–1425.
70 H. Kojima and A. J. Bard, J. Am. Chem. Soc., 1975, 97, 85 G. Fodor, R. Mujumdar and A. Szentgyorgyi, Proc. Natl. Acad.
¨
6317–6324.
Sci. U. S. A., 1978, 75, 4317–4319.
71 E. R. Nightingale, J. Phys. Chem., 1959, 63, 1381–1387.
86 H. Reich and L. Hefle, J. Org. Chem., 1956, 21, 708.
72 M. G. Evans and M. Polanyi, Trans. Faraday Soc., 1938, 34, 87 J. E. Bennett and R. Summers, Can. J. Chem., 1974, 52,
11–24. 1377–1379.
73 S. W. Benson, Thermochemical kinetics, Wiley, New York, 1976. 88 G. A. Russell, J. Am. Chem. Soc., 1957, 79, 3871–3877.
74 M. Ochando-Pardo, I. Nebot-Gil, A. Gonzalez-Lafont and 89 P. Ulanski, E. Bothe, K. Hildenbrand, J. M. Rosiak and
J. M. Lluch, Chem. Phys. Lett., 2005, 409, 255–259.
C. von Sonntag, J. Chem. Soc., Perkin Trans. 2, 1996, 23–28.
75 B. Noziere, S. Ekstrom, T. Alsberg and S. Holmstrom, 90 C. von Sonntag and H. P. Schuchmann, Angew. Chem., Int.
Geophys. Res. Lett., 2010, 37, L05806. Ed. Engl., 1991, 30, 1229–1253.
76 J. Schindelka, Y. Iinuma, D. Hoffmann and H. Herrmann, 91 M. I. Stefan and J. R. Bolton, Environ. Sci. Technol., 1999, 33,
Faraday Discuss., 2013, 165, 237–259. 870–873.
77 A. Sabljic and H. Gusten, Atmos. Environ., Part A, 1990, 24, 92 D. J. Creighton, M. Migliorini, T. Pourmotabbed and
73–78. M. K. Guha, Biochemistry, 1988, 27, 7376–7384.
78 G. Marston, P. S. Monks, C. E. Canosa-Mas and R. P. Wayne, 93 A. Tilgner and H. Herrmann, Atmos. Environ., 2010, 44,
J. Chem. Soc., Faraday Trans., 1993, 89, 3899–3905. 5415–5422.
`
¨
¨
6272 | Phys. Chem. Chem. Phys., 2014, 16, 6257--6272
This journal is © the Owner Societies 2014