Atmos. Chem. Phys., 9, 2129–2140, 2009www.atmos-chem-phys.net/9/2129/2009/© Author(s) 2009. This work is distributed underthe Creative Commons Attribution 3.0 License.

AtmosphericChemistry

and Physics

Reactions of isoprene and sulphoxy radical-anions – a possiblesource of atmospheric organosulphites and organosulphates

K. J. Rudzinski1, L. Gmachowski2,1, and I. Kuznietsova1

1Department of Catalysis on Metals, Institute of Physical Chemistry of the Polish Academy of Sciences,01-224 Warsaw, Poland2Institute of Chemistry, Warsaw University of Technology, 09-400 Płock, Poland

Received: 24 September 2008 – Published in Atmos. Chem. Phys. Discuss.: 12 December 2008Revised: 2 March 2009 – Accepted: 14 March 2009 – Published: 24 March 2009

Abstract. Transformation of isoprene coupled with auto-oxidation of SIV in aqueous solutions was studied exper-imentally and by chemical-kinetic modelling over a broadrange of solution acidities (pH=3–9) to complement the re-search on aqueous-phase and heterogeneous transformationof isoprene reported recently by many laboratories. Iso-prene significantly slowed down the auto-oxidation in acidicand basic solutions, and accelerated it slightly in neutral so-lutions. Simultaneously, production of sulphate ions andformation of solution acidity were significantly reduced.Formation of sulphite and sulphate derivatives of isoprene- sulphurous acid mono-(2-methyl-4-oxo-but-2-enyl) ester(m/z=163), sulphurous acid mono-(4-hydroxy-2-methyl-but-2-enyl) ester (m/z=165), sulphuric acid mono-(2-methyl-4-oxo-but-2-enyl) ester (m/z=179), sulphuric acid mono-(4-hydroxy-2-methyl-but-2-enyl) ester (m/z=181), and possi-ble structural isomers of these species – was indicated byelectrospray ionisation mass spectrometric analysis of post-reaction mixtures. The experimental results were explainedby changes in a subtle quantitative balance of three super-imposed processes whose rates depended in different man-ner on the acidity of reacting solutions – the scavengingof sulphoxy radical-anions by isoprene, the formation ofsulphoxy radical-anions during further reactions of isopreneradicals, and the auto-oxidation of SIV itself. A chemi-cal mechanism based on this idea was explored numericallyto show good agreement with experimental data. In ba-sic and neutral solutions, the model overestimated the con-sumption of isoprene, probably because reactions of primarysulphite and sulphate derivatives of isoprene with sulphoxyradical-anions were not included. Interaction of isoprene

Correspondence to:K. J. Rudzinski(kjrudz@ichf.edu.pl)

with sulphur(IV) species and oxygen can possibly result information of new organosulphate and organosulphite com-ponents of atmospheric aerosols and waters, and influencethe distribution of reactive sulphur and oxygen species inisoprene-emitting organisms exposed to SIV pollutants.

1 Introduction

Isoprene(C5H8) and sulphur dioxide(SO2) are importanttrace components of the atmosphere that are emitted frombiogenic and anthropogenic sources (Georgii and Warneck,1999; Sharkey et al., 2007). Sulphur dioxide is also pro-duced directly in the atmosphere, for instance by oxidation ofdimethylsulphide (Barnes et al., 2006). The recognised fateof SO2 is conversion to sulphuric acid and sulphates, in thegas-phase, heterogeneous or multiphase processes (Seinfeldand Pandis, 2006). On the other hand, isoprene is quickly ox-idised in the gas phase by radicals (OH, NO3) and by ozone,to several primary and many secondary products, such asmethacrolein and methylvinyl ketone or methylglyoxal andorganic nitrates, respecively (LeBras and the LACTOZ Steer-ing Group, 1997; Fan and Zhang, 2004; Paulot et al., 2009;Taraborrelli et al., 2008).

The discovery of tetrols in ambient aerosol samples col-lected in Amazon rainforest and boreal K-Puszta forest(Claeys et al., 2004a,b) spawned intensive research on for-mation of secondary organic aerosol (SOA) in heteroge-neous and multiphase reactions of isoprene. For review ofthis research see (Rudzinski, 2006, 2008). In summary, nu-merous experiments in simulation chambers, unseeded orseeded with ammonium sulphate particles, showed isopreneinduced formation of aerosol that contained such compoundsas methyltetrols, C5 alkene triols, 2-methylglyceric acid,

Published by Copernicus Publications on behalf of the European Geosciences Union.

2130 K. J. Rudzinski et al.: Reactions of isoprene and sulphoxy radical-anions

glyoxal, methylglyoxal, acetals and hemiacetals, organosul-phates and organonitrates, as well as oligomers of some ofthese (Surratt et al., 2007a,b, 2008). Yield and compositionof aerosol depended on the presence of NOx or SO2 in thegas phase, or H2SO4 added to acidify the seed particles.

In particular, sulphuric acid and sulphur dioxide increasedthe yields of aerosol (Jang et al., 2002; Limbeck et al., 2003;Kleindienst et al., 2006; Surratt et al., 2007b). Sulphuricacid alone promoted the heterogeneous and aqueous forma-tion of isoprene derivatives containing sulphate functionalgroups. Formation of these organosulphates was originallyexplained by direct reaction of the acid with products of iso-prene oxidation such as aldehydes and polyols (Liggio et al.,2005; Surratt et al., 2007a). However, a recent experimentalstudy showed direct esterificaton of alcohols was not feasi-ble kinetically (Minerath et al., 2008). Reactions of glyoxalon ammonium sulphate aerosol produced glyoxal sulphateonly upon irradiation (Galloway et al., 2008), which likelycould initiate the formation of sulphate radical-anions. Onthe other hand, isoprene was shown to inhibit the aqueous-phase auto-oxidation of dissolved SO2, or rather sulphiteions (Rudzinski, 2004). The suggested mechanism of in-hibition included direct reaction of isoprene with sulphateradical-anions, which led to formation of sulphate esters ofisoprene derivatives.

The aim of this work was to show that aqueous reactions ofsulphoxy radical-anions with isoprene can produce sulphur-containing derivatives of isoprene, and to show how thesereactions depend on and influence the acidity of aqueous so-lutions during oxidation of dissolved SIV species (sulphurdioxide, sulphite ions) to SVI species (sulphuric acid, sul-phate ions).

2 Experimental

2.1 Method

The experimental method we used to study the transforma-tion of isoprene coupled with auto-oxidation of SIV catalysedby manganous sulphate was described in detail byRudzinski(2004). We used a well stirred glass reactor of 0.785 dm3

volume, closed with a Teflon cover and thermostatted withina water jacket. The reactor was operated homogenously (nogas phase was present), and in a batch manner. Each experi-mental run was prepared by filling the reactor with aqueoussolution of manganese catalyst and oxygen, and adding analiquot of aqueous solution of isoprene. Then, the reactorwas sealed and the run was started by injecting an aliquotof aqueous solution of sodium sulphite. This solution wasprepared from Na2SO3 and Na2S2O5 used in different pro-portions in order to obtain the desired initial acidities of thereacting mixtures (Reactions R1–R3). Only for the lowest

initial acidity (pH=2.9), we additionally acidified the mixturewith a small amount of dilute sulphuric acid.

(Na2SO3)s + H2O → 2Na++ SO2−

3 + H2O (R1)

(Na2S2O5)s + H2O → 2Na++ 2HSO−

3 (R2)

HSO−

3 H++ SO2−

3 (R3)

The variables recorded in each experimental run includedpH, concentration of dissolved oxygen, temperature and highresolution UV spectra of reacting solutions. The pH andconcentration of oxygen were measured continuously, usinga Ross Ultra glass combination pH electrode and a 525A+pH-meter from Thermo Electron, and a 9708-99 oxygenprobe (Clark type) with 920 pH-meter from Orion (now alsoThermo Electron). The oxygen electrode was calibrated be-fore each experiment against the electronic zero and the localatmospheric pressure, using a built-in circuitry. The concen-tration of dissolved oxygen was read directly from the pH-meter, in ppmO2. Voltage signals from both pH-meters wererecorded every 0.1 s, and converted to pH and oxygen con-centration in mol dm−3, using a computer system equippedwith a M6281 data acquisition card and a LabView applica-tion, both from National Instruments. The UV spectra wererecorded periodically with a Jasco V570 spectrophotometer(0.2 nm bandwidth, 0.5 nm data pitch, 1 cm light path), us-ing a closed sampling loop constructed from a T valve, a sy-ringe and a Hellma Suprasil cell. Temperature in the reactorwas measured with a mercury thermometer. Post-reactionsolutions were analysed off-line, and in negative ion modeusing an API 365 triple quadrupole mass spectrometer withelectrospray ionisation from Applied Biosystems. The sam-ples were taken directly after each experiment, using a gas-tight syringe to prevent the access of air, and diluted withmethanol before injection to the MS apparatus.

Concentrations of isoprene, sulphite ions and bisulphiteions in the reacting solutions were obtained from therecorded UV spectra by subtraction of reference spectra ofthese species. The accuracy of determination of subtractioncoefficients was better than 12%. The reference spectra andthe analysis of subtraction errors were given in (Rudzinski,2004).

2.2 Chemicals

Most of chemicals were purchased from Merck and usedwithout further purification: C5H8 for synthesis grade, sta-bilised with 100 ppm of 4-tert-butylpyrocatechol, Na2SO3,Na2S2O5 and MnSO4×H2O ACS pro analysi grade. Sul-phuric acid, pro analysis grade, was obtained from CHE-MAN. Buffer standards used for daily calibration of pH elec-trodes were from Thermo Electron.

All solutions were prepared using the Milli-Qplus Milli-pore water. Oxygen was dissolved in water by equilibrationwith atmospheric air. Liquid isoprene was dissolved in waterwith the aid of an ultrasonic bath.

Atmos. Chem. Phys., 9, 2129–2140, 2009 www.atmos-chem-phys.net/9/2129/2009/

K. J. Rudzinski et al.: Reactions of isoprene and sulphoxy radical-anions 2131

Table 1. Experimental conversion ratios in transformation of iso-prene coupled with auto-oxidation of SIV .

Initial pH 1[SIV

]/1

[O2

]1

[C5H8

]/1

[SIV

]8.3 1.53 0.217.7 1.28 0.257.0 1.52 0.255.8 1.63 0.404.0 1.61a 1.21a

a the values have inreased uncertainity, because conversion of eachreactant was lower than 1%.

3 Results and discussion

All experiments were carried out at 25.0±0.2◦C. Initial con-centrations of reactants were set to(0.950±0.020) × 10−3

mol dm−3 (SIV ), (0.250±0.030) ×10−3 mol dm−3 (O2),(0.117±0.050) ×10−3 mol dm−3 (C5H8), 1.0×10−5

mol dm−3 (MnSO4). The initial acidity of reacting solutionsvaried from pH=2.9 to 8.5. Each experiment with isoprenewas paired with another experiment carried out at practicallythe same conditions but without isoprene. Figure 1 showssample time traces of reactant concentrations and pH,obtained from a single experimental run in neutral solution,with isoprene present.

3.1 Influence of isoprene on auto-oxidation of SIV

Auto-oxidation of SIV catalysed by transition metals andundisturbed by addition of foreign substances is charac-terised by a fixed ratio of SIV conversion to oxygen con-version, which reflects the 2:1 stoichiometry of the overallreaction (Reaction R4). In the presence of isoprene, the con-version ratio dropped to lower values, due to formation ofoxygenated derivatives of isoprene (Table 1, see also mecha-nism in Sect. 4, Figs. 7–9).

2Na2SO3 + O2MnSO4→ 2Na2SO4 (R4)

Our previous work showed that isoprene slowed downthe auto-oxidation of SIV in solutions of pH≥8 (Rudzinski,2004). Here, we confirmed this observation for higher con-centrations of isoprene, and extended the experiment to neu-tral and acidic solutions (pH=3÷8).

We found that isoprene slowed down the auto-oxidationof SIV in acidic and basic solutions (pHo< 7 and pHo>8),but accelerated it in neutral solutions (pHo=7÷8). Figures 2and 3 compare time traces of the dimensionless concentrationof oxygen (xO2=[O2]/[O2]o) obtained from experiments inacidic and basic solutions, with and without isoprene added(red and cyan lines, respectively). In all these cases, isopreneslowed down the auto-oxidation. In solutions of pH<5, theuninhibited auto-oxidation was so slow itself, that isoprene

Fig. 1. Transformation of isoprene coupled with auto-oxidation ofSIV – time traces of reactant concentrations and pH obtained froman experiment in neutral solution (SIV is sum of HSO−3 and SO2−

3ions).

Fig. 2. Inhibition of SIV auto-oxidation by isoprene in acidic solu-tions – time traces of dimensionless concentration of oxygen.

could stop it almost completely. On the contrary, Fig. 4shows that in all experiments in neutral solutions, isopreneaccelerated the auto-oxidation. The diverse influence of iso-prene could result from subtle changes in a quantitative bal-ance of three superimposed processes whose rates depended

www.atmos-chem-phys.net/9/2129/2009/ Atmos. Chem. Phys., 9, 2129–2140, 2009

2132 K. J. Rudzinski et al.: Reactions of isoprene and sulphoxy radical-anions

Fig. 3. Inhibition of SIV auto-oxidation by isoprene in basic solu-tions – time traces of dimensionless concentration of oxygen.

in a different manner on the acidity of reacting solutions –the scavenging of sulphoxy radicals by isoprene, the forma-tion of sulphoxy radicals during further transformation ofisoprene, and the auto-oxidation of SIV itself. The chemicalmechanism utilising this concept was discussed in Sect. 4.

Figures 2–4 show that in all experiments, the additionof isoprene reduced the acidification of reacting solutions.During uninhibited auto-oxidation of SIV , the solutions wereacidified due to changes in the acid-base dissociation balancethat took place when sulphite and bisulphite ions were con-verted into sulphate ions. Isoprene always reduced this acid-ification markedly, or even reversed it in experiments start-ing at pH>8.2. This influence was a natural consequence ofslower auto-oxidation of SIV , but also reflected the forma-tion of isoprene derivatives capable of associating with freeprotons in solutions.

3.2 Transformation of isoprene

The analysis of UV spectra of reacting solutions showed thatin all experiments isoprene decayed proportionally to the de-cay of sulphite. The ratio of these conversions ranged from0.21 to 0.40, and to 1.21, the latter value being rather uncer-tain as calculated from conversions lower than 1% (Table 1).In neutral and slightly acidic solutions, we saw weak lightabsorption by non-sulphate reaction products, with a broadpeak at 240 nm (Fig. 5). The peak was positioned at longerwavelengths than the 225 nm peak of isoprene (Rudzinski,

Fig. 4. Acceleration of SIV auto-oxidation by isoprene in neutralsolutions – time traces of dimensionless concentration of oxygen.

Fig. 5. UV absorption spectra of products of isoprene transforma-tion coupled with SIV auto-oxidation in an experiment starting atpH=6.6. Inset shows the time trace of absorbance at 240 nm.

2004), and close to the 239 nm peak of the unsaturated alde-hyde 3-methylbut-2-enal,(CH3)2C=CH−CHO, calculatedusing the increment method (Hesse et al., 1997): 207 nmfor the base structure -C=C-CHO in methanol+24 nm for twoCH3 groups+8 nm for water replacing methanol as a solvent.

Atmos. Chem. Phys., 9, 2129–2140, 2009 www.atmos-chem-phys.net/9/2129/2009/

K. J. Rudzinski et al.: Reactions of isoprene and sulphoxy radical-anions 2133

Two structural isomers of this aldehyde had calculated peaksat positions a little more distant: CH3CH=C(CH3)-CHO at237 nm, and CH2=C(CH3)-C(CH3)O at 233 nm. All threecarbonyls are similar in structure to possible products of iso-prene transformation in our experiments, which additionallycontain sulphate or sulphite substituents. We cautiously sug-gest that spectra in Fig. 5 indicate these products, because theinfluence of sulphate or sulphite groups on the UV absorp-tion above 200 nm is likely negligible. Unfortunately, otherpossible products of isoprene transformation, which containhydroxyl groups instead of carbonyl oxygens, would not ab-sorb the UV light between 200 and 300 nm, and, if present,were invisible to our absorbance measurements.

Electrospray ionisation mass spectra of post-reaction solu-tions contained peaks of deprotonated molecules that couldbe attributed to products of isoprene transformation initi-ated by sulphite and sulphate radical-anions (Fig. 6), suchas sulphurous acid mono-(-2-methyl-4-oxo-but-2-enyl) ester(m/z=163), sulphurous acid mono-(4-hydroxy-2-methyl-but-2-enyl) ester (m/z=165), sulphuric acid mono-(2-methyl-4-oxo-but-2-enyl) ester (m/z=179), and sulphuric acid mono-(4-hydroxy-2-methyl-but-2-enyl) ester (m/z=181). Test anal-yses excluded the possibility of analytical artefacts, showingthat the peaks did not appear in spectra of freshly preparedsolutions containing isoprene and sodium sulphate or iso-prene and sodium sulphite. The compounds shown in Fig. 6are examples of several structural isomers, which could beproduced in our experiments and attributed to the observedMS peaks. A sulphoxy radical can add to double bonds inisoprene at four different positions. Consequently, furtherreactions can produce five -oxo- isomersHOxSO-CH2-C(CH3)=CH-CHO,CH2=CH-(CH3)C(OSOxH)-CHO,CH2=C(CH3)-CH(OSOxH)-CHO,HOxSO-CH2-CH=C(CH3)-CHO,CH2=C(CH3)-C(O)-CH2-OSOxH,and six -hydroxy- isomersHOxSO-CH2-(CH3)C(OH)-CH=CH2,HOxSO-CH2-C(CH3)=CH-CH2OH,CH2=CH-(CH3)C(OSOxH)-CH2OH,CH2=C(CH3)-CH(OSOxH)-CH2OH,HOxSO-CH2-CH=C(CH3)-CH2OH,CH2=C(CH3)-CH(OH)-CH2-OSOxHfor each type of sulphoxy radical involved (x=2 for sulphiteradical, and x=3 for sulphate radical). Identification of in-dividual isomers was not possible in this work, and wouldrequire higher yields of the products, and commercial or syn-thesised standards that were not available. However, three ofthe -oxo- compounds were also indicated by the UV spectradiscussed in the previous subsection.

Whatever their exact structure is, the tentatively identi-fied compounds have double carbon-carbon bonds that canreact further, for instance with radical species. Saturatedorganosulphates produced in such reactions would be similarto sulphate esters of methylglyceric acid and 2-methyltetrols

5 0 1 0 0 1 5 0 2 0 00

1

2

3

4

5

6

1 6 3

O S O 3

O

-

O S O 3

O H

-O S O 2

O H

-

O S O 2

O

-

1 3 7 1 4 3 1 8 7

1 2 21 3 0

1 1 91 1 7

1 1 39 9

9 5

6 56 17 5

4 5

9 1

8 1

9 7

1 6 51 7 9

����

����

��� ����

m / z , a m u

1 8 1

Fig. 6. Peaks corresponding to products of isoprene transformationcoupled with auto-oxidation of SIV appeared in negative ion elec-trospray ionisation mass spectra of post-reaction solutions.

that had been detected in field samples of atmosphericaerosols (Gomez Gonzalez et al., 2007; Surratt et al., 2007a,2008).

4 Chemical mechanism and simulation of experiments

Auto-oxidation of SIV catalysed by manganese was stud-ied extensively in the context of both atmospheric chem-istry and technology of cleaning the industrial waste gasesor waste-waters (Pasiuk-Bronikowska et al., 1992; Berglundand Elding, 1995; Brandt and van Eldik, 1995; Fronaeuset al., 1998; Grgic and Bercic, 2001; Ermakov and Purmal,2002; Kuo et al., 2006). It is a chain reaction initiated byMnIII and propagated by sulphoxy radical-anions which alsoregenerate MnIII . In the absence of inhibitors, the chain ter-minates via several radical-radical reactions.

The chemical mechanism constructed to explain our ex-periments with transformation of isoprene coupled with auto-oxidation of SIV consisted of three groups of reactions –auto-oxidation of MnII to MnIII in presence of sulphite orbisulphite ions (Fig. 7), auto-oxidation of sulphite and bisul-phite ions

(SIV

)catalysed by MnSO4 (Fig. 8), and trans-

formation of isoprene initiated by sulphate radical-anions(Fig. 9) or by sulphite radical-anions (not shown, since analo-gous to the scheme for sulphate radical-anions). The individ-ual reactions in the schemes were identified by specific sym-bols of the rate constants rather than by cosecutive numbers,to simplify the presentation and provide better mnemonic as-sociation. Values of the rate constants used in simulationswere collected in Table 2, along with available references.

www.atmos-chem-phys.net/9/2129/2009/ Atmos. Chem. Phys., 9, 2129–2140, 2009

2134 K. J. Rudzinski et al.: Reactions of isoprene and sulphoxy radical-anions

Table 2. Rate constants used for simulation of chemical mechanism shown in Figs. 7–9.

Constant Value Constant Value Constant Valuein Fig. 7 M−1 s−1 in Fig. 8 M−1 s−1 in Fig. 9 M−1 s−1

kass 1.0×1011 kass 1.0×1011 ks1k 2.19×109

kdysa 6.75×103 kdys

a 6.75×103 ks2 1.0×109

kkf b 6×106 kib 1.0×103 ks3b 1.0×109

kkbb 1 kia 3.0×101 ks3a 1kkf a 6×106 kp1

d 1.5×109 ks4b 1.0×109

kkba 1 kp21be 3.25×106 ks4a 1.0×109

kaf b 6 kp22be 9.75×106 ks5 (1 ÷ 5)×104

kabb 1 kp21ae 2.5×104 ks6

l 3.5×109

kaf a 8 ÷ 10 kp22ae 7.5×104

kaba 1 kp3bf 1.4×107 Analogous constants for

kdf b 6 kp3af 1.4×107 reactions of SO−3 radicals

kdbb 1 kox11 3.2×106 (not shown in figures)kdf a 8 ÷ 10 kox12 1.2×102

kdba 1 kox2 1.0×106 k2s1 1.0×109

ko2b 6 kt22g 1.0×108 k2s2 1.0×109

ko2a 8 ÷ 10 kcas 1.0×1010 k2s3b 5.0×104

khpbb 0.19 kcdy

a 3.98 k2s3a 1.0×106

khpac 3.4×107 kf b 7.5×103 k2s4b (1.2 ÷ 2.5)×104

kf a 7.5×103 k2s4a 5.0×104

kt1h 4.0×108 k2s5 1

kt21i 1.0×107

kt3 1.0×106

kt4j 5.0×108

a s−1. b Drexler et al.(1992). c M−2 s−1, rate=khpa[H+][H2O2][HSO−

3 ],Drexler et al.(1991). d Huie and Neta(1984). e Huie and Neta

(1987). f Averaged inCAPRAM. g 5.5×106 Herrmann et al.(1995), ∼ 2.2×108 Buxton et al.(1996). h Waygood and McElroy(1992).i 1.3×108 Herrmann et al.(1995), 4.8×107 Buxton et al.(1996). j McElroy and Waygood(1990). k Rudzinski (2004). l 1.7×109 Buxtonet al.(1996).

4.1 Group 1 – auto-oxidation of MnII in presence of sul-phite or bisulphite ions (Fig. 7)

Chemical-kinetic modeling of our experiments requiredsome mechanism for the initial oxidation of MnII to MnIII

in the presence of sulphite or bisulphite ions. Common re-action pathways suggested in literature included oxidationof MnII by such impurities as other transition metal ions(Grgic and Bercic, 2001), and several oxidation schemesbased on formation of manganese-sulphite complexes andoxygen adducts (Bassett and Parker, 1951; Berglund andElding, 1995). Here, we used a mechanism first intro-duced byRudzinski and Pasiuk-Bronikowska(2000), whichutilised the idea of formation ofµ-peroxo-dimer of metalcomplexes with organic ligands (Wilkins, 1971; Mimoun,1980), followed by intramolecular one-electron oxidation ofMnII and decomposition to MnIII , H2O2 and hydroxyl ions(Gubelmann and Williams, 1983). The scheme comprisedreversible formation of MnII -sulphite complexes (kkf a/kkba

and kkf b/kkbb), reversible addition of oxygen molecule tothis complex followed by intramolecular one-electron oxi-

dation of MnII (kaf a/kaba and kaf b/kabb), then formationof dimer from the adduct and the complex (kdf a/kdba andkdf b/kdbb) and intramolecular oxidation of second MnIII ,and finally decomposition of the dimer to H2O2 and Mn3+

(ko2a/ko2b). Unoxidised sulphite ions were released fromthe dimer, while oxygen molecule was used to build H2O2,which could oxidise sulphite ions directly to sulphate ions(khpa andkhpb). Two hydroxyl ions were also formed, whichresulted in some alkalisation of the reaction mixture. Theoxidation of MnII could take two parallel paths, one start-ing with SO2−

3 ions and the other – with HSO−3 ions, thetwo bound by a reversible association-dissociation reaction(kass/kdys).

The extent of MnIII auto-oxidation was very minor, and itspurpose was to start the auto-oxidation of SIV by supplyingthe very initial amount of Mn3+ in the absence of any otherinitiator. When the auto-oxidation chain grew longer, MnII

was oxidised back to MnIII in several reactions with sulphoxyradicals and ions. This is exactly why manganese is alsocalled a catalyst of auto-oxidation.

Atmos. Chem. Phys., 9, 2129–2140, 2009 www.atmos-chem-phys.net/9/2129/2009/

K. J. Rudzinski et al.: Reactions of isoprene and sulphoxy radical-anions 2135

kasskdys

kkbb kkbakkfb kkfa

kabb kabakafb kafa

kdbbkdfb kdbakdfa

ko2b ko2a

−+− + 323 HSOHSO

3IISOMn +HSOMn 3

II

2O

3III

2 SO)Mn(O +HSO)Mn(O 3III

2

3IIIIII

3 SOOOMnSMnO ++ HSOOOMnSMnHO 3IIIIII

3

++++ 3-2322

- Mn22SOOH2OH

+2Mn+ +

+ +

+ +

−+ 3HSO+−23SO

+− ++ HOHSO 224

khpb khpa

Fig. 7. Mechanism of auto-oxidation of MnII in the presence ofsulphite and bisulphite ions.

4.2 Group 2 – auto-oxidation of SIV catalysed by man-ganese (Fig. 8)

Chain auto-oxidation of SIV was initiated by formation ofsulphite radical-anions in reaction of Mn3+ with sulphite orbisulphite ions (kib andkia), and proceeded via several prop-agation steps – formation of peroxymonosulphate radical-anions from sulphite radical-anions and oxygen (kp1), splitreactions of peroxymonosulphate radical-anions with sul-phite or bisulphite ions that led to formation of peroxymono-sulphate ions and regeneration of sulphite radical-anions(kp21a andkp21b), or to formation of sulphate radical-anionsand sulphate ions (kp22a andkp22b), and reaction of sulphateradical-anions with sulphite or bisulphite ions, which regen-erates sulphite radical-anions (kp3a andkp3b). Peroxymono-sulphate ions SO2−

5 were reversibely protonated to HSO−

5ions (kcas /kcdy). Sulphate ions were also produced in re-action of peroxymonosulphate ions with sulphite ions (kf a

andkf b), which could include formation of an intermediatecomplex that had not been included in the present mecha-nism. Chain termination occurred via radical-radical reac-tions (kt1, kt21, kt3, kt4). Reaction of two peroxymonosul-phate radical-anions could also regenerate sulphate radical-anions (kt22). The chain initiator, Mn3+, was regeneratedin reactions of Mn2+ cations with sulphite radical-anions(kox11), sulphate radical-anions (kox2) and peroxymonosul-phate ions (kox12).

4.3 Group 3 – transformation of isoprene initiated bysulphoxy radical-anions (Fig. 9)

In this mechanism, isoprene could react with sulphite andsulphate radical-anions. Each radical could add to one oftwo double bonds of isoprene at any of four locations (1, 2, 3or 4). We assumed there were no kinetic differences betweenthese additions, so they could be represented by a single reac-tion for each type of sulphoxy radicals. The final products ofisoprene transformation included in the mechanism were thecompounds indicated by mass spectra of post-reaction solu-tions (Fig. 6).

Figure 9 shows the mechanism of isoprene transformationinitiated by addition of sulphate radicals (ks1). Alkyl radicalswere formed that accepted oxygen molecules to form perox-yalkyl radicals (ks2). The peroxyalkyl radicals could oxidisesulphite (ks3b) or bisulphite ions (ks3a), while turning intoalkoxy radicals. The alkoxy radicals reacted with sulphite(ks4b) or bisulphite ions (ks4a) to form a hydroxy derivativeof isoprene and to regenerate sulphite radical-anions. In aparallel reaction with oxygen, the alkoxy radicals turn into anoxy derivative of isoprene and produce hydroperoxy radicals(ks5). Hydroperoxy radicals reacted with sulphate radical-anions, producing sulphate ions (ks6).

An analogous scheme (not shown here) was constructedfor transformation of isoprene starting with the addition of asulphite radical-anion to a double bond. The rate constantsfor the respective reactions were distinguished by adding “2”in the subscript (k2s1, k2s2, k2s3a, k2s3b, k2s4a, k2s4b, k2s5).The isoprene derivatives were similar to those obtained inthe transformation initiated by sulphate radical-anions, butcontained -OSO2 substituent groups instead of -OSO3.

4.4 Simulation of experiments

Basing on the chemical mechanism described above, a cor-responding set of ordinary differential equations was con-structed and solved using a Mathematica 2.3 package. Re-versible reactions were treated kinetically, each one as a pairof independent reactions. Rate constants used for simulationswere collected in Table 2. Constants, for which no exper-imental estimates were available, were adjusted using trialand error guesses.

www.atmos-chem-phys.net/9/2129/2009/ Atmos. Chem. Phys., 9, 2129–2140, 2009

2136 K. J. Rudzinski et al.: Reactions of isoprene and sulphoxy radical-anions

kib kia

kp1

kp21akp22b kp22a

kp3b kp3a

kt22

+2Mn

kass −+− + 323 HSOHSO

+3Mn+ +

+•+ ++ HSOMn 32 -

2O +

−•5SO

−•+ 5SO

kp21b

+−−• ++ HSOSO 243

+−23SO −+ 3HSO

+−−• ++ HSOSO 253

+−−• ++ HSOSO 244

+−23SO −+ 3HSO

244 OSOSO ++ −•−•

kdys

kcas −+− + 525 HSOHSO

kcdy+−2

3SO −+ 3HSO

++− ++ HH2SO24

kfb kfa

228255 OOSSOSO ++ −−•−•

−−•−• + 26233 OSSOSO

−−•−• + 28244 OSSOSO

226253 OOSSOSO ++ −−•−•

kt21

kt1kt4

kt3

+−•5SO

+−•4SO

+−5HSO

−+ 25SO

−+ 24SO

−−• ++ OHSO 4

+3Mn

kox11

kox2

kox12H+

Fig. 8. Mechanism of SIV auto-oxidation catalysed by manganese.

The results of simulation were compared to matching ex-periments in figures arranged according to the initial valuesof pH of reacting mixtures. Time traces of reactant concen-trations were shown in Figs. 10 and 11, while changes of pHof reacting solutions were shown in Fig. 12. Generally, timetraces of SIV and oxygen concentrations were reproducedwith good accuracy at all acidities. Consumption of isoprenewas simulated quite accurately in acidic solutions, but wasoverestimated in basic and neutral solutions. The simulationwas still accurate qualitatively, reproducing the right shapesof isoprene time traces from all experiments. The quanti-tative disagreement was explained by the deficiency of themechanism used, which treated the primary nonradical prod-ucts of isoprene degradation as nonreactive. Each of theseproducts contained a double C-C bond, which could reactwith radicals in solutions to influence the overall inhibiting oraccelerating action. In the present mechanism, “all the work”had to be done by isoprene, which consequently appeared to

decay faster than in the experiments. The effect was morepronounced in basic and neutral solutions, because conver-sion of isoprene was higher than in acidic solutions. In ad-dition, the mechanism would probably gain still more quan-titaive flexibility, if included all primary nonradical productsof isoprene transformation (Sect. 3.2), not the four necessaryrepresentatives (Sect. 4.3).

The ratio of simulated conversions of sulphite and oxygenwas 2.0 in the absence of isoprene and 1.64–1.66 in the pres-ence of isoprene, while the ratio of simulated conversionsof isoprene and sulphite ranged from 0.28 to 0.31 (Table 3).The calculated ratios agreed well with those determined ex-perimentally (Table 1). In the absence of isoprene all sulphiteand bisulphite ions were converted into sulphate ions. Simu-lation showed that isoprene reduced this conversion by about30%, as part of SIV species became the substituent groups inproduced esters.

Atmos. Chem. Phys., 9, 2129–2140, 2009 www.atmos-chem-phys.net/9/2129/2009/

K. J. Rudzinski et al.: Reactions of isoprene and sulphoxy radical-anions 2137

854 HCSO +−•

2O+

+−23SO −+ 3HSO

−+ ++ 242 SOOH

2O+

+−•4SO

ks1

ks2

ks3aks3b

ks4b ks4a ks5

ks6

2323 HCHC)C(CHSOCHO •− =

•− = OOCHCH)C(CHSOCHO 2323

+•− ++= HSOOCHCH)C(CHSOCHO -242323

+−•−− ++= HSOOCHCH)C(CHSOCHO 32323

+−23SO −+ 3HSO

•− += 2323 HOCHCHO)C(CHSOCHO

Fig. 9. Mechanism of transformation of isoprene initiated by sul-phate radical-anions. Similar scheme (not shown) was used fortransformation initiated by sulphite radicals, with reaction costantsnamedk2s1, k2s2, k2s3a, k2s3b, k2s4a, k2s4b, k2s5, respectively.

The acidity of reacting mixtures was reproduced quitewell, with the exception of experiments that started atpHo=8.3 and included isoprene, in which the experimen-tal acidity decreased, while the simulated acidity slightlyincreased. This disagreement could be explained by theability of functional groups in isoprene derivatives to bindfree protons from solutions, which had not been includedin the present mechanism. In acidic and basic solutions,the effect was masked by high yields of free protons re-leased upon oxidation of HSO−3 ions to SO2−

4 . It becamevisible in basic solutions, just because the initial reservoirof dissociable protons was smaller by orders of magnitude,while the extremely low initial concentration of free pro-tons (∼5×10−9 M) was very sensitive to even small shifts inthe protonation-deprotonation balance of species whose con-centration was higher by three or four orders of magnitude.Thus, the prediction of hydrogen ion concentration in basicsolutions should improve, if the protonation and deprotona-tion reactions of isoprene esters are added to the mechanism.

Table 3. Model conversion ratios in transformation of isoprene cou-pled with auto-oxidation of SIV .

Initial pH 1[SIV

]/1

[O2

]1

[C5H8

]/1

[SIV

]8.3 1.65 0.297.7 1.65 0.287.0 1.65 0.305.8 1.65 0.314.0 1.64 0.31

Fig. 10. Chemical-kinetic simulation of temporal concentrationprofiles for experiments starting in basic and neutral solutions(pHo≥7).

In summary, the results of simulation showed that the pro-posed mechanism of isoprene transformation during autoxi-dation of SIV was a reasonable approximation of the real-lifechemistry in our experiments. The experimental and simu-lated data agreed very well qualitatively, and well quantita-tively. The shapes of the concentration time traces, and thestoichiometry of transformation were reproduced accurately.The qualitative performance of the mechanism reflected itsadequate structure. The quantitative performance can stillimprove within this structure, for instance if a better set ofreaction rate constants is used. The working set of rate con-stants used for the presented simulation contained many con-stants, whose values were guessed by fitting, and should bereplaced by values determined in future experiments. Fur-ther improvement is expected, if the mechanism is extended

www.atmos-chem-phys.net/9/2129/2009/ Atmos. Chem. Phys., 9, 2129–2140, 2009

2138 K. J. Rudzinski et al.: Reactions of isoprene and sulphoxy radical-anions

Fig. 11. Chemical-kinetic simulation of temporal concentrationprofiles for experiments in acidic solutions (pHo<7).

by adding the reactions indicated in this section. Despite thepresent limitations, the mechanism is ready for application infuture studies.

5 Conclusions

We showed, by kinetic experiments, product analysis andchemical-kinetic modelling, that isoprene was capable ofreacting with radicals in aqueous solutions, and turningitself into reactive radicals. Subtly regulating the bal-ance of sulphoxy radical-anions, isoprene influenced theauto-oxidation of SIV catalysed by manganese. The auto-oxidation was significantly slowed down in basic and acidicsolutions, and slightly accelerated in neutral solutions. How-ever, production of sulphate ions and build-up of solutionacidity were reduced in all cases. Products of isoprene trans-formation included unsaturated sulphite and sulphate esters– sulphurous acid mono-(2-methyl-4-oxo-but-2-enyl) ester,sulphurous acid mono-(4-hydroxy-2-methyl-but-2-enyl) es-ter, sulphuric acid mono-(2-methyl-4-oxo-but-2-enyl) ester,sulphuric acid mono-(4-hydroxy-2-methyl-but-2-enyl) ester,and their possible structural isomers. These esters are likelyprecursors of saturated organosulphur compounds similar toorganosulphates detected in aerosol samples from field cam-paigns and from simulation chamber experiments.

Reactions of isoprene with sulphoxy radical-anions havea few potential implications for atmospheric chemistryand atmosphere-biosphere interactions. They are another

Fig. 12. Chemical simulation of temporal pH profiles for experi-ments at all aciditiies.

possible source of new organosulphur components of atmo-spheric aerosols and waters, which add to heterogeneoussources postulated recently. They can also slow down theformation of sulphuric acid and sulphate ions in atmosphericwaters, and consequently reduce the formation of acidity andinorganic sulphate aerosol. Probably, reactions of isoprenewith sulphoxy radical-anions can also influence the distribu-tion of reactive sulphur and oxygen species inside isoprene-emitting organisms (plants, animals and humans) and ontheir surfaces.

Realisation of reactions of isoprene with sulphoxy radical-anions in real atmospheric systems has to be evaluated quan-titatively by further experimental and modelling study. Thepreferred systems are characterised by high availability ofisoprene and SIV species (sulphur dioxide), either emittedlocally or transported from remote areas. Good examplesare rainforests, urban areas with high emissions of vehicleexhausts, sea-coastal regions, sea-atmosphere interface andsurfaces of isoprene-emitting plants. The laboratory workshould focus on the uptake of isoprene by aqueous solu-tions, which is limited by low solubility of isoprene in wa-ter on one hand, but enhanced by aqueous reactions of iso-prene and microphysical redistribution of reactants on theother hand. Products of reactions of isoprene with sulphoxyradical-anions should be searched for in field samples of at-mospheric aerosols and waters.

Further research is carried out to better identify productsof isoprene reactions with sulphoxy radical-anions, and toinvestigate further reactions of these products with sulphoxyradical-anions and other radicals.

Atmos. Chem. Phys., 9, 2129–2140, 2009 www.atmos-chem-phys.net/9/2129/2009/

K. J. Rudzinski et al.: Reactions of isoprene and sulphoxy radical-anions 2139

Acknowledgements.The authors acknowledge the opportunity ofdoing this work within the framework of European Network ofExcellence ACCENT, and thank Polish Ministry of Science andHigher Education for the financial support of this work in relationto ESF Program INTROP. Ms Edyta Szeremeta is acknowledged forhelp with carrying out part of the experimental work.

References

Barnes, I., Hjorth, J., and Mihalopoulos, N.: Dimethyl sulfide anddimethyl sulfoxide and their oxidation in the atmosphere, Chem.Rev., 106, 940–975, 2006.

Bassett, H. and Parker, W. G.: The oxidation of sulphurous acid,J. Chem. Soc., 1951, 1540–1560, doi:10.1039/JR9510001540,1951.

Berglund, J. and Elding, L. I.: Manganese-catalysed autoxidationof dissolved sulfur dioxide in the atmospheric aqueous phase,Atmos. Environ., 29, 1379–1391, 1995.

Brandt, C. and van Eldik, R.: Transition metal-catalyzed oxidationof sulfur(IV) oxides. atmospheric-relevant processes and mech-anisms, Chem. Rev., 95, 119–190, online available at:http://pubs.acs.org/doi/pdf/10.1021/cr00033a006, 1995.

Buxton, G. V., McGowan, S., Salmon, G. A., Williams, J. E., andWood, N. D.: A study of the spectra and reactivity of oxysulphur-radical anions involved in the chain oxidation of S(IV): A pulseand gamma – radiolysis study, Atmos. Environ., 30, 2483–2493,1996.

CAPRAM: Chemical Aqueous Phase RAdical Mechanism, onlineavailable at: http://projects.tropos.de/capram/, last access: 16July 2008.

Claeys, M., Graham, B., Vas, G., Wang, W., Vermeylen, R., Pashyn-ska, V., Cafmeyer, J., Guyon, P., Andreae, M. O., Artaxo, P.,and Maenhaut, W.: Formation of secondary organic aerosolsthrough photooxidation of isoprene, Science, 303, 1173–1176,doi:10.1126/science.1092805, online available at:http://www.sciencemag.org/cgi/content/abstract/303/5661/1173, 2004a.

Claeys, M., Wang, W., Ion, A. C., Kourtchev, I., Gelencser, A.,and Maenhaut, W.: Formation of secondary organic aerosolsfrom isoprene and its gas-phase oxidation products through reac-tion with hydrogen peroxide, Atmos. Environ., 38, 4093–4098,2004b.

Drexler, C., Elias, H., Fecher, B., and Wannowius, K. J.: Kineticinvestigation of sulfur(IV) oxidation by peroxo compounds R-OOH in aqueous solution, Fresen, J. Anal. Chem., 340, 605–616,1991.

Drexler, C., Elias, H., Fecher, B., and Wannowius, K. J.: Kineticsand mechanism of sulfur(IV) oxidation by hydrogen peroxide inaqueous phase: the non-linear parts of the pH-profile, Ber. Bun-sen Gesell., 96, 481–485, 1992.

Ermakov, A. N. and Purmal, A. P.: Catalysis of HSO−

3 /SO2−

3 oxi-dation by manganese ions, Kinet. Catal., 43, 249–260, 2002.

Fan, J. and Zhang, R.: Atmospheric oxidation mechanism of iso-prene, Environ. Chem., 1, 140–149, 2004.

Fronaeus, S., Berglund, J., and Elding, L.: Iron-manganese re-dox processes and synergism in the mechanism for manganese-catalyzed autoxidation of hydrogen sulfite, Inorg. Chem., 37,4939–4944, online availabel at:http://pubs.acs.org/doi/pdf/10.1021/ic980225z, 1998.

Galloway, M. M., Chhabra, P. S., Chan, A. W. H., Surratt, J. D.,Flagan, R. C., Seinfeld, J. H., and Keutsch, F. N.: Glyoxal up-take on ammonium sulphate seed aerosol: reaction products andreversibility of uptake under dark and irradiated conditions, At-mos. Chem. Phys. Discuss., 8, 20799–20838, 2008,http://www.atmos-chem-phys-discuss.net/8/20799/2008/.

Georgii, H. W. and Warneck, P.: Global aspects of atmosphericchemistry, in: Chemistry of the tropospheric aerosol and ofclouds, edited by: Zellner, R., Steinkopf, Darmstadt, Germany,111–179, 1999.

Gomez Gonzalez, Y., Vermeylen, R., Szmigielski, R., Surratt, J. D.,Kleindienst, T. E., Jaoui, M., Lewandowski, M., Offenberg, J. H.,Edney, E. O., Maenhaut, W., and Claeys, M.: Characterisation oforganosulphates from the photo-oxidation of isoprene in ambi-ent PM2.5 aerosol by LC/(-)ESI-linear ion trap mass spectrom-etry, in: European Aerosol Conference, Salzburg, Austria 9-14September 2007, T01A018, 2007.

Grgic, I. and Bercic, G.: A simple kinetic model for autoxidation ofS(IV) oxides catalyzed by iron and/or manganese ions, J. Atmos.Chem., 39, 155–170, 2001.

Gubelmann, M. H. and Williams, A. F.: The structure and reactivityof dioxygen complexes of the transition metals, in: TransitionMetal Complexes – Structures and Spectra, no. 55 in Structureand Bonding, Springer-Verlag, Berlin, Germany, 1983.

Herrmann, H., Reese, A., and Zellner, R.: Time-resolved UV/VISdiode array absorption spectroscopy of SO−

x (x=3,4,5) radicalanions in aqueous solution, J. Mol. Struct., 348, 183–186, 1995.

Hesse, M., Meier, H., and Zech, B.: Spectroscopic Methods inOrganic Chemistry, Georg Thieme Verlag, Stuttgart, New York,1997.

Huie, R. E. and Neta, P.: Chemical behavior of SO−

3 and SO−5 radi-cals in aqueous solutions, J. Phys. Chem., 88, 5665–5669, 1984.

Huie, R. E. and Neta, P.: Rate constants for some oxidations ofS(IV) by radicals in aqueous solutions, Atmos. Environ., 21,1743–1747, 1987.

Jang, M., Czoschke, N. M., Lee, S., and Kamens, R. M.: Aerosolproduction by acid-catalyzed particle-phase reactions, Science,298, 815–817, 2002.

Kleindienst, T. E., Edney, E. O., Lewandowski, M., Offenberg,J. H., and Jaoui, M.: Secondary organic carbon and aerosol yieldsfrom the irradiations of isoprene andα-pinene in the presence ofNOx and SO2, Environ. Sci. Technol., 40, 3807–3812, 2006.

Kuo, D. T. F., Kirk, D. W., and Jia, C. Q.: The chemistry of aqueousS(IV)-Fe-O2 system: state of the art, J. Sulfur Chem., 27, 461–530, online available at:http://www.informaworld.com/10.1080/17415990600945153, 2006.

LeBras, G. and the LACTOZ Steering Group: Oxidation mecha-nism of isoprene, in: Chemical Processes in Atmospheric Oxida-tion: Laboratory Studies of Chemistry Related to TroposphericOzone. Vol. 3: Transport and Chemical Transformation of Pollu-tants in the Troposphere, Springer-Verlag, Berlin, Germany, 68–72, 1997.

Liggio, J., Li, S.-M., and McLaren, R.: Heterogeneous reactions ofglyoxal on particulate matter: Identification of acetals and sulfateesters, Environ. Sci. Technol., 39, 1532–1541, 2005.

Limbeck, A., Kulmala, M., and Puxbaum, H.: Secondary organicaerosol formation in the atmosphere via heterogeneous reactionof gaseous isoprene on acidic particles, Geophys. Res. Lett., 30,1996, doi:10.1029/2003GL017738, online available at:http://dx.

www.atmos-chem-phys.net/9/2129/2009/ Atmos. Chem. Phys., 9, 2129–2140, 2009

2140 K. J. Rudzinski et al.: Reactions of isoprene and sulphoxy radical-anions

doi.org/10.1029/2003GL017738, 2003.McElroy, W. J. and Waygood, S.: Kinetics of the reactions of the

SO−

4 radical with SO−4 , S2O2−

8 , H2O and Fe2+, J. Chem. Soc.Faraday T., 86, 2557–2564, 1990.

Mimoun, H.: The role of peroxymetallation in selective oxidationprocesses, J. Mol. Catal., 7, 1–29, 1980.

Minerath, E. C., Casale, M. T., and Elrod, M. J.: Kinetics feasibilitystudy of alcohol sulfate esterification reactions in troposphericaerosols, Environ. Sci. Technol., 42, 4410–4415, 2008.

Pasiuk-Bronikowska, W., Ziajka, J., and Bronikowski, T.: Autoxi-dation of Sulphur Compounds, Ellis Horwood, New York, USA,1992.

Paulot, F., Crounse, J. D., Kjaergaard, H. G., Kroll, J. H., Seinfeld,J. H., and Wennberg, P. O.: Isoprene photooxidation: new in-sights into the production of acids and organic nitrates, Atmos.Chem. Phys., 9, 1479–1501, 2009,http://www.atmos-chem-phys.net/9/1479/2009/.

Rudzinski, K. J.: Degradation of isoprene in the presence ofsulphoxy radical anions, J. Atmos. Chem., 48, 191–216, 2004.

Rudzinski, K. J.: Heterogeneous and aqueous-phase transformationof isoprene, in: Environmental Simulation Chambers: Appli-cation to Atmospheric Chemical Processes, edited by: Barnes,I. and Rudzinski, K. J., Springer, Dordrecht, The Netherlands,261–277, 2006.

Rudzinski, K. J.: Undiscovered chemistry – is it important formechanisms and models?, in: Simulation and Assessment ofChemical Processes in a Multiphase Environment, edited by:Barnes, I. and Kharytonov, M. M., Springer, Dordrecht, TheNetherlands, 231–253, 2008.

Rudzinski, K. J. and Pasiuk-Bronikowska, W.: Inhibition of SO2oxidation in aqueous phase:, Work. Stud. I. Environ. Eng. Pol.Acad. Sci., 54, 175–191, 2000.

Seinfeld, J. H. and Pandis, S. N.: Atmospheric Chemistry andPhysics, John Wiley & Sons, Inc., Hoboken, New Jersey, sec-ond edn., USA, 2006.

Sharkey, T. D., Wiberley, A. E., and Donohue, A. R.: Isopreneemission from plants: why and how, Ann. Bot.-London, 1–14,doi:10.1093/aob/mcm240, 2007.

Surratt, J. D., Kroll, J. H., Kleindienst, T. E., Edney, E. O., Claeys,M., Sorooshian, A., Ng, N. L., Offenberg, J. H., Lewandowski,M., Jaoui, M., Flagan, R. C., and Seinfeld, J. H.: Evidence fororganosulfates in secondary organic aerosol,, Environ. Sci. Tech-nol., 41, 517–527, 2007a.

Surratt, J. D., Lewandowski, M., Offenberg, J. H., Jaoui, M., Klein-dienst, T. E., and Edney, E. O.: Effect of acidity on secondaryorganic aerosol formation from isoprene, Environ. Sci. Technol.,41, 5363–5369, 2007b.

Surratt, J. D., Gomez-Gonzalez, Y., Chan, A. W. H., Vermeylen, R.,Shahgholi, M., Kleindienst, T. E., Edney, E. O., Offenberg, J. H.,Lewandowski, M., Jaoui, M., Maenhaut, W., Claeys, M., Flagan,R. C., and Seinfeld, J. H.: Organosulfate Formation in BiogenicSecondary Organic Aerosol, J. Phys. Chem. A, 112, 8345–8378,doi:10.1021/jp802310p, online available at:http://pubs.acs.org/doi/abs/10.1021/jp802310p, 2008.

Taraborrelli, D., Lawrence, M. G., Butler, T. M., Sander, R., andLelieveld, J.: Mainz Isoprene Mechanism 2 (MIM2): an isopreneoxidation mechanism for regional and global atmospheric mod-elling, Atmos. Chem. Phys. Discuss., 8, 14033–14085, 2008,http://www.atmos-chem-phys-discuss.net/8/14033/2008/.

Waygood, S. and McElroy, W. J.: Spectroscopy and decay kineticsof the sulfite radical anion in aqueous solution, J. Chem. Soc.Faraday T., 88, 1525–1530, 1992.

Wilkins, R. C.: Uptake of oxygen by cobalt(II) complexes in solu-tion, in: Bioinorganic Chemistry, no. 100 in Advances in Chem-istry Series, ACS, Washington DC, USA, 1971.

Atmos. Chem. Phys., 9, 2129–2140, 2009 www.atmos-chem-phys.net/9/2129/2009/