Formation of secondary aerosols from the ozonolysis of styrene: Effect of SO2 and H2O

Accepted Manuscript Formation of secondary aerosols from the ozonolysis of styrene: Effect of SO2 and H2O Yolanda Diaz-de-Mera, Alfonso Aranda, Ernesto Martínez, Ana Angustias Rodríguez, Diana Rodríguez, Ana Rodriguez PII:

S1352-2310(17)30671-4

DOI:

10.1016/j.atmosenv.2017.10.011

Reference:

AEA 15605

To appear in:

Atmospheric Environment

Received Date: 6 July 2017 Revised Date:

3 October 2017

Accepted Date: 6 October 2017

Please cite this article as: Diaz-de-Mera, Y., Aranda, A., Martínez, E., Rodríguez, A.A., Rodríguez, D., Rodriguez, A., Formation of secondary aerosols from the ozonolysis of styrene: Effect of SO2 and H2O, Atmospheric Environment (2017), doi: 10.1016/j.atmosenv.2017.10.011. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

RI PT

ACCEPTED MANUSCRIPT

SO2

AC C

EP

TE D

M AN U

SC

O3

H2O

SOA

ACCEPTED MANUSCRIPT 1

Formation of secondary aerosols from the ozonolysis of styrene:

2

effect of SO2 and H2O

3

Diaz-de-Mera, Yolanda1; Aranda, Alfonso1; Martínez, Ernesto2; Rodríguez, Ana Angustias2;

4

Rodríguez, Diana3; Rodriguez, Ana3.

5

1

6

Camilo José Cela s/n, 13071. Ciudad real. Spain.

7

2

8

Atmosférica, Camino Moledores, s/n, 13071 – Ciudad Real. Spain.

9

3

RI PT

Universidad de Castilla-La Mancha. Facultad de Ciencias y Tecnologias Quimicas. Avenida

Universidad de Castilla-La Mancha. Instituto de Investigación en Combustión y Contaminación

SC

Universidad de Castilla-La Mancha. Facultad de Ciencias Ambientales y Bioquímica. Avenida

Carlos III s/n, 45071 Toledo, Spain.

11

Correspondence to: Alfonso Aranda ([email protected])

M AN U

10

12

Abstract. In this work we report the study of the ozonolysis of styrene and the reaction

14

conditions leading to the formation of secondary aerosols. The reactions have been carried out

15

in a Teflon chamber filled with synthetic air mixtures at atmospheric pressure and room

16

temperature. We have found that the ozonolysis of styrene in the presence of low

17

concentrations of SO2 readily produces new particles under concentrations of reactants lower

18

than those required in experiments in the absence of SO2. Thus, nucleation events occur at

19

concentrations around (5.6±1.7)x108molecule cm-3 (errors are 2σ±20%) and SO2 is consumed

20

during the experiments. The reaction of the Criegee intermediates with SO2 to produce SO3

21

and then H2SO4 may explain (together with OH reactions’ contribution) the high capacity of

22

styrene to produce particulate matter in polluted atmospheres.

23

The formation of secondary aerosols in the smog chamber is inhibited under high H2O

24

concentrations. So, the potential formation of secondary aerosols under atmospheric

25

conditions depends on the concentration of SO2 and relative humidity, with a water to SO2 rate

26

constants ratio kH2O/kSO2= (2.8±0.7)×10-5 (errors are 2σ±20%).

AC C

EP

TE D

13

27 28

1. Introduction.

29

Styrene is one of the most important monomers worldwide. The double bond in the vinyl

30

group enables polymerisation and so the production of widespread used materials such as

1

ACCEPTED MANUSCRIPT polystyrene, styrene–butadiene rubber, styrene–butadiene latex, etc. Despite the high

32

reactivity of styrene in the atmosphere, it can be considered as an ubiquitous species due to

33

the wide range of different anthropogenic sources. Thus, it has been measured in remote

34

areas (Kos et al., 2014 ), in indoor air samples (Sarigiannis et al., 2011; Xu et al., 2016), or in

35

ambient air (Kabir and Kim, 2010; Li et al., 2014).

36

Styrene is naturally produced as a degradation product in cinnamic acid containing plants

37

(Duke, 1985) and as a by-product of fungi and microbe metabolism (Shimada et al., 1992).

38

Nevertheless, the presence of styrene in the atmosphere is mainly due to fugitive emissions

39

from petrochemical facilities and industrial processes involving styrene or its polymers

40

(Knighton et al., 2012; Zhang et al., 2017). Other sources include the release from styrene-

41

based materials, incineration of styrene polymers, or cigarette smoke for indoor air. The

42

combustion of traditional petroleum-based fuels or the alternative newly developed biofuels

43

are also responsible of important exhaust emissions of styrene mainly in populated areas

44

(Lopes et al., 2014; Li et al., 2015; Hu et al., 2017).

45

Styrene is a hazardous air pollutant and so its emissions are regulated by the environmental

46

protection agencies (EPA, 2017). The exposure to inhalation or intake of styrene through food

47

or drinks may lead to severe health effects (ATSDR, 2017).

48

The main gas phase reactions of styrene in the atmosphere have been studied previously,

49

being the reported room-temperature kinetic rate constants with OH, NO3 and ozone (in cm3

50

molecule-1 s-1): 5.8x10-11, 1.51x10-13 (Atkinson and Aschmann, 1988) and 1.5x10-17 (Le Person et

51

al., 2008), respectively. The results obtained by Le Person et al. (Le Person et al., 2008) show

52

that reactions with OH, NO3 and O3 are all important atmospheric loss processes for styrene in

53

the gas phase.

54

On the other hand, photolysis of styrene constitutes a slower sink during daylight hours.

55

Recently, exposure to particulate matter is being associated to the decrease of life expectancy

56

and numerous diseases, especially cardiopulmonary illness (Bates et al., 2015). The size of

57

particles is a key factor for health effects since the potential ability to access the lungs

58

becomes higher for the smaller ultrafine particles (Nel, 2005). This is a problem in urban air

59

where particles with size diameter below 100 nm may account for 80–90 % of the total particle

60

number concentration (Mejía et al., 2008). Apart from direct primary sources like vehicle

61

exhausts, one of the main sources of ultrafine particles is the formation of secondary organic

62

aerosols (SOA) from the reactions of the emissions of gas-phase pollutants (Cheung et al.,

63

2013). In this sense, it has been found that aromatic compounds readily yield SOA under urban

AC C

EP

TE D

M AN U

SC

RI PT

31

2

ACCEPTED MANUSCRIPT air conditions (Ng et al., 2007). The contribution of aromatic compounds to the total SOA

65

formation may account from 78.5 to 91.9 %, well above those of alkanes and alkenes (Sun et

66

al., 2016). Thus, given their high SOA formation potential, new approaches to predict the

67

amount of aerosol formed from the oxidation of aromatic hydrocarbons are being developed

68

(Li et al., 2016). In the case of styrene, it was identified as the second most efficient species

69

forming SOA (15%), only after toluene (16%) (Sun et al., 2016).

70

For styrene, a theoretical model shows that its reaction with OH could contribute to the

71

formation of SOA in polluted atmospheres (Wang et al., 2015). Furthermore, a previous

72

experimental work has also reported the formation of SOA from the reaction of styrene with

73

ozone and the effect of the presence of ammonia and water (Na et al., 2006).

74

Being styrene an alkene, its reaction with ozone proceeds through the formation of a Criegee

75

intermediate (CI). Very recently, it has been found that stabilised CI (sCI) can undergo reactions

76

with SO2 several orders of magnitude faster than assumed so far (Welz et al., 2012) producing

77

SO3, which contributes efficiently to the formation of ground level sulfuric acid (Mauldin et al.,

78

2012; Newland et al., 2015). Likewise, mixing SO2 with gasoline vehicle exhausts has been

79

shown to increase the formation of secondary aerosols (Liu et al. 2016). Thus, the reaction

80

with SO2 could be the key to explain the high SOA formation potential of styrene under

81

polluted conditions. On the other hand, the competition of SO2 and water vapour in the

82

removal of sCI in the atmosphere depends on the CI structure (Berndt et al., 2014a; Díaz-de-

83

Mera et al., 2017) and may play an important effect on the net SOA formation of this aromatic

84

compound.

85

In this work we study the ozonolysis reaction of styrene following the conditions that lead to

86

the formation and growth of new particles. The effect of water vapour and SO2 concentrations

87

during the process are also studied and discussed.

SC

M AN U

TE D

EP

AC C

88

RI PT

64

89

2. Experimental.

90

The experimental system consisted on a collapsible 200L Teflon reactor coupled to different

91

detectors to analyse both the gas phase and particulate matter. The reactants’ samples,

92

styrene, cyclohexane and SO2, were prepared in a volume-calibrated glass bulb with 100 and

93

1000Torr full scale capacitance pressure gauges, MKS 626AX, and then flushed to the reactor

94

using synthetic air. Concentrations for each species were obtained from their partial pressures

95

in the glass bulb and the final volume of the Teflon reactor.

3

ACCEPTED MANUSCRIPT 96

The ozone was generated in-situ by passing pure oxygen through a high voltage electric

97

discharge, BMT Messtechnik 802N, and then fed to the sampling bulb and to an absorption 10

98

cm long quartz cell to measure the concentration using a UV-vis Hamamatsu spectrometer,

99

C10082CAH at 255nm. Once the concentration in the cell and bulb were known, it was diluted with synthetic air to get the required concentration in the reactor.

101

Cyclohexane in excess was used as OH radical scavenger in concentration ratios so that >97%

102

of OH formed in the reactions was removed (Ma et al., 2009).

103

Water vapour in the reactor was introduced by passing the air flows carrying the reactants

104

through a glass bubbler containing the calculated amount of water required for any given

105

relative humidity (RH). The sampled water was then evaporated and flushed inside the reactor

106

through the flow of synthetic air. For the series of experiments carried out under “dry

107

conditions” the synthetic air was used directly from the cylinder, with a water mixing ratio

108

below 2ppm, which is the concentration stated by the supplier.

109

Typically, the reactants were sequentially introduced as follows: first, styrene, second

110

cyclohexane (OH scavenger), then SO2 and finally ozone. During the injection of ozone, the

111

reactor was shaken to improve mixing and then the valves to the sampling detectors were

112

open. Liquid water (when necessary) was introduced in the bubbler at the beginning, being

113

flushed by the air carrying the rest of gaseous reactants.

114

The concentration of particles was monitored by an scanning mobility particle sizer (SMPS), TSI

115

SMPS 3938 with a CPC particle counting 3775 and a nano-differential mobility analyser which

116

improves size resolution over the particle size range 4.0-150nm. The SMPS was operated with

117

3.0 L min-1 sheath flow and 0.3 L min-1 sampling flow rate. Size distributions were recorded at

118

1 minute intervals from the injection of ozone and the total aerosol mass concentration was

119

derived from the measured particle size distribution assuming unit density and spherical

120

particles. Prior to each experiment the Teflon bag was repetitively washed with clean synthetic

121

air until the level of particles was below 1 cm-3.

122

SO2 concentrations were measured by a fluorescence SO2 analyser, Teledyne Instruments 101-

123

E. Previous runs with only SO2/air samples in the reactor showed that SO2 wall losses were

124

negligible. Additionally, no interference in the SO2 fluorescence signal from the rest of co-

125

reactants was observed in the range of concentrations used in this study.

126

Ozone concentrations were measured with an ozone analyser, Environment O342M. Several

127

tests showed that O3 losses in the reactor were negligible within the time scale of the

128

experiments. Styrene also shows UV absorption at 255nm. Thus, several samples of styrene in

AC C

EP

TE D

M AN U

SC

RI PT

100

4

ACCEPTED MANUSCRIPT air were introduced in the reactor to calibrate the O3 analyzer for the styrene measurements.

130

After a few minutes an important decrease in the styrene signal was observed (attributed to

131

polymerisation within the UV detection cell) what prevents from monitoring the styrene

132

temporal profiles in the reactor continuously. Nevertheless, when sampling of styrene was

133

carried out intermittently for short time periods (and letting the UV absorption cell to

134

completely evacuate the previous sample) the signal always recovered its initial intensity.

135

Under dark conditions with different styrene concentrations in the reactor, no measurable

136

consumption of styrene was observed from the periodic samples measured in the UV

137

spectrometer for 80 minutes runs.

138

Reagents.

139

All the substances used were of the highest commercially available purity. Liquid reagents

140

were purified by successive trap to trap distillation. Styrene >99%, Sigma-Aldrich; SO2 99,9%,

141

Fluka; cyclohexane 99.5%, Sigma-Aldrich; synthetic air 3X, Praxair; O2 4X, Praxair.

142

M AN U

SC

RI PT

129

3. Results and discussion.

144

3.1. Nucleation conditions.

145

For asymmetric alkenes like styrene, the ozonolylis leads to the formation of two different

146

Criegee intermediates and the corresponding aldehydes: C6H5C·HOO· and HCHO or ·CH2OO·

147

and C6H5CHO. For the title reaction, previous results show that the two pathways (leading to

148

formaldehyde or benzaldehyde) are quantitative and with yields around 40 % (Tuazon et al.,

149

1993). The subsequent reactions of the primary products, especially the Criegee intermediates

150

(Na et al., 2006, Bernd et al., 2014a), may lead to the formation of secondary aerosols.

151

In order to approach real ambient air conditions, a series of experiments were carried out with

152

ozone and styrene using concentrations lower than those of the study by Na et al., 2006. Thus

153

for example, for initial styrene and ozone concentrations in the range of 10 ppbs, no particles

154

were observed in the reactor after 60 minutes runs. The concentrations were then

155

progressively increased and, for example, for experiments with concentrations of 100 ppb of

156

styrene and 200 ppb of ozone, particles generated after 15 minutes. Since the gas-phase

157

reaction of styrene with ozone has been reported previously with k295K=(1.5±0.3)x10-17cm3

158

molecule-1 s-1 (Le Person et al., 2008), this kinetic rate constant can been used to simulate the

159

ozone and styrene concentrations profiles. Thus, from the previous experiments, for 15

160

minutes reaction time the results show that the concentration of reacted styrene required for

AC C

EP

TE D

143

5

ACCEPTED MANUSCRIPT the formation of new particles is very high, (1.6±0.5)x1011 molecule cm-3 (errors are 2σ±20%)

162

and so the reaction of styrene with ozone is unlikely, by itself, to drive nucleation events under

163

usual atmospheric conditions.

164

On the other hand, when carrying the experiments in the presence of SO2, particles were

165

generated at lower concentrations of reactants. Figure 1 shows the typical experimental

166

profiles of the particle number concentration (PNbC), particle mass concentration (PMC) and

167

the simulated profiles for styrene concentration and total reacted styrene. As we can see, after

168

the initial time required to complete nucleation, the PMC increases linearly. Likewise, since the

169

ozone and styrene reaction is slow, the reactant concentrations remain essentially constant

170

and the reacted styrene increases linearly, like PMC. This behaviour suggests that the initial

171

reaction is the rate limiting step in the formation of condensed matter under these

172

experimental conditions.

AC C

173

EP

TE D

M AN U

SC

RI PT

161

174

Figure 1. Experimental temporal profiles for particles (PNbC and PMC) and simulated profiles

175

of styrene and reacted styrene. Initial concentrations: styrene, 10ppb; O3, 20ppb and SO2,

176

10ppb. RH=0%. Styrene and reacted styrene concentrations are shown multiplied by the

177

required factors to display them in the same figure.

178 179

A series of experiments was conducted to measure the reacted concentration of styrene

180

required for the generation of particles in the presence of SO2. Thus, Figure 2 shows the

181

concentrations of reacted styrene calculated at the nucleation events (assumed as the first

182

scan with PNbC above 100 cm-3) for experiments with SO2 initial concentrations in the range 56

ACCEPTED MANUSCRIPT 100 ppbs and fixed initial concentrations of styrene (10 ppb) and ozone (20 ppb). The burst of

184

particles takes place at approximately the same reaction time and concentration of reacted

185

styrene. So, increasing SO2 concentration doesn’t result into further acceleration of the

186

nucleation event, which is found to occur once the reacted styrene amounts (5.6±1.7)x108

187

molecule cm-3 (errors are 2σ±20%). This value is three orders of magnitude lower than in the

188

absence of sulphur dioxide, showing that the SO2 reactivity must play a key role in the

189

nucleation steps. As shown above, lowering the concentrations of SO2 rises the nucleation

190

threshold (as inferred from the 5 ppb SO2 experiment) up to 1.6x1011 molecule cm-3 for SO2

191

zero concentration.

RI PT

183

SC

192 2.0E+09

M AN U

1.6E+09 1.4E+09 1.2E+09 1.0E+09 8.0E+08 6.0E+08 4.0E+08 2.0E+08 0.0E+00 0

TE D

styrene reacted concentration for nucleation (molecule/cm3)

1.8E+09

20

40

60

80

100

193

EP

[SO2] (ppb)

Figure 2. Concentration of reacted styrene at the instant of particles’ appearance for

195

experiments with different initial SO2 concentrations. For this series of experiments,

196

[styrene]=10ppb and [O3]=20ppb. The displayed errors are σ.

197

AC C

194

198

3.2. Effect of styrene and ozone initial concentrations.

199

A series of experiments was carried out to assess the effect of styrene concentration on the

200

formation of secondary aerosols. Figure 3 shows the results for PNbC, PMC and the mean

201

particle size diameter for different [styrene]. The particles grew larger for higher

202

concentrations of styrene with average diameters up to 45 nm and the number concentration

203

of new particles rose up to 4x105 cm-3 for experiments carried out under 10 ppb and 20 ppb

204

initial concentrations of SO2 and ozone respectively. 7

ACCEPTED MANUSCRIPT 205

To assess the conversion of gas-phase reactants into aerosols we can use the fractional aerosol

206

yield (Y) which may be defined (Odum et al., 1996) as the fraction of the reagent (styrene for

207

this work), that is converted to aerosol: =

∆

Where PMC is the aerosol mass concentration (µg m-3) produced for a given amount of reacted

209

styrene, ∆

RI PT

208

(in µg m-3).

210

4.0E+05

45 40

M AN U

3.0E+05 2.5E+05 2.0E+05 1.5E+05 1.0E+05 5.0E+04 0.0E+00

0.1

EP

0

TE D

PNbC (#/cm3)

3.5E+05

50

0.2

0.3 0.4 [styrene] (ppm)

35 30

Diameter (nm)

4.5E+05

211

Diameter (nm)

25 20 15 10 5

PMC (µg/m3)

PMC (microg/m3)

SC

PNbC (#/cm3)

0 0.5

0.6

Figure 3. Particle mass concentration and mean particle size diameter versus styrene

213

concentration at reaction time=40 min for 10 ppb and 20 ppb initial concentrations of SO2 and

214

ozone respectively. The maximum particle number concentration (PNbC) measured for each

215

experiment are also plotted.

216

AC C

212

217

According to Odum’s model, the growth of the aerosol’s mass may be described as a gas-

218

particle partitioning absorption mechanism where organic products may partition a portion of

219

their mass at concentrations below their saturation concentration.

220

Figure 4 displays the particle mass concentration growth curves for several experiments with

221

different initial styrene concentrations. After the first few minutes, once nucleation has 8

ACCEPTED MANUSCRIPT 222

occurred, the system exhibits nearly constant yields for the rest of the experiment.

223

Furthermore, the yields from the different experiments are very similar.

224

The linearity of the fractional yield suggests that the aerosol mass comes quantitatively from a

225

single product with very low volatility. This behaviour has been found also for the formation of

226

SOA from other aromatic compounds such as benzene, m-xylene or toluene (Ng et al., 2007).

RI PT

227 12 0.1ppm

10

0.2ppm

0.4ppm 0.5ppm

M AN U

PMC (µg m-3)

SC

0.3ppm

8

6

4

0 0

5

TE D

2

10

15

20

∆C8H8 (µg

228

25

30

35

40

m-3)

Figure 4. Particle mass concentration growth curves versus the reacted styrene for different

230

initial styrene concentrations (see the legend). Fixed initial concentrations of sulphur dioxide

231

and ozone were used: [SO2]=10 ppb and [O3]=20 ppb.

AC C

232

EP

229

233

Figure S1 shows the yields obtained from the slopes of the linear fittings (straight part after the

234

first few minutes) of the plot PMC versus ∆

235

concentration affording an average fractional aerosol yield of 0.39±0.10 (errors are 2σ±20%).

236

The dependence of the aerosol fractional yield on the initial concentration of styrene is weak

237

as it is found in the plot. Nevertheless, for low styrene concentrations the yield values are

238

slightly higher.

(Figure 4) versus the styrene initial

239

The effect of increasing ozone concentrations has been also studied in this work. A series of

240

experiments fixing the initial concentrations of styrene and SO2 was carried out with different 9

ACCEPTED MANUSCRIPT 241

initial concentrations of ozone in the range 10 ppb to 0.6 ppm. The increase of ozone led to

242

the increase of PNbC, PMC and average particle size diameters, with a behaviour similar to the

243

series changing styrene concentration described previously. The results are found in the

244

supplementary material, Figures S2-S4.

245 3.3. Effect of SO2 initial concentration.

247

As shown previously, the presence of SO2 enhances the production of secondary aerosols from

248

the ozonolysis reaction of styrene in air. To find the effect of SO2 concentration in the

249

mechanism, different experiments were performed fixing the initial concentrations of styrene

250

and ozone while the initial concentration of SO2 was changed within the range 5 to 100 ppb.

251

Even though the reacted styrene profile must be the same for all these experiments (since the

252

same initial concentrations of O3 and styrene are used), a significant increase in PNbC and PMC

253

was observed for increasing SO2 concentrations (mainly in the range 0-25 ppb), Figure 5. For

254

[SO2] above approximately 25 ppb, further increases of SO2 didn´t result in significant changes

255

in the aerosol production. The calculated fractional aerosol yield also increased up with [SO2]

256

to values around 2, as it is shown in Figure S5. From the definition of the fractional yield,

257

values above unity are consistent since styrene is not the only species involved in the

258

formation of secondary aerosol and so the mass of the particulate matter can be larger than

259

the mass of the reacted styrene.

SC

M AN U

TE D

EP AC C

260

RI PT

246

10

ACCEPTED MANUSCRIPT PNbC (#/cm3)

PMC (microg/m3)

Diameter (nm)

35 30

1.0E+05

25

8.0E+04

RI PT 20

6.0E+04

15

4.0E+04

10

2.0E+04 0.0E+00 0.02

0.04

0.06 [SO2] (ppm)

0.08

0.1

0 0.12

M AN U

0

SC

5

10 x PMC (µg/m3)

PNbC (#/cm3)

1.2E+05

Diameter (nm)

40 1.4E+05

261

Figure 5: Effect of SO2 concentration on the production of secondary aerosols. The maximum

263

particle number concentration (PNbC) measured for each experiment are shown in the main

264

axis. PMC and mean particle size diameter versus SO2 concentration are shown at a fixed

265

reaction time, 50 min (secondary axis), for 10 ppb and 20 ppb initial concentrations of styrene

266

and ozone respectively.

267

TE D

262

SO2 doesn’t react with the initial reactants, cyclohexane, styrene or ozone within the time

269

scale of the experiments nor with the aldehydes HCHO, C6H5CHO coming from the ozonolysis

270

reaction. On the other hand, it is known that Criegee intermediates do undergo fast reactions

271

with SO2 (Taatjes et al., 2013). Nevertheless, since the styrene-ozone reaction is slow, under

272

the experimental conditions, Figure 5, the production rate of the Criegee intermediates is very

273

low and a slight consumption of SO2 was expected. Consistently, the experimental SO2

274

temporal profiles during most of the experiments were nearly constant. Thus, in some

275

experiments, to magnify the SO2 consumption and make it measurable, it was necessary to

276

increase the styrene and ozone concentrations (up to 100-200 ppb) while keeping [SO2] in

277

lower levels (around 20 ppb). For this series of experiments, the production of sCI was

278

relatively high and the SO2 concentration was observed to decrease with time (experimental

279

SO2 decays, Figure 6), which confirms that SO2 reacts with the products from the ozonolysis.

AC C

EP

268

11

ACCEPTED MANUSCRIPT Previous studies have shown that the first step of the reaction of sCI (R1R2COO) with SO2 is the

281

formation of a cyclic secondary ozonide which can undergo isomerisation (only possible if R2 is

282

an H-atom), leading to an organic acid (R1COOH) and SO2, or decomposition to a carbonyl

283

compound (R1R2CO) and SO3 (Jiang et al.,2010; Kurten et al., 2011). The organic acids will

284

generally be less volatile than their alkene precursors and SO3 readily produces H2SO4 in the

285

presence of water vapour. Such products are expected to be involved in the formation and

286

growth of new particles.

287

In the isomerisation channel, SO2 would behave as a catalyst in the production of acids (Díaz

288

de Mera et al., 2017). On the other hand, in this work we have found, Figure 6, that there is a

289

net consumption of SO2 and that the amount of reacted SO2 increases with the reaction time

290

coordinate. This fact suggests that for the case of styrene ozonolysis the reaction of the sCI

291

with SO2 proceeds at least in part through the oxidation channel leading to SO3.

292

Some of the previous experiments were carried out under the presence of water under

293

different RH. For example, as shown in Figure 6, when water vapour is present (RH=5 % in the

294

plotted experiment) the decrease in SO2 was significantly lower than under drier conditions.

295

These results suggest that H2O and SO2 constitute competitive sinks for the stabilised Criegee

296

intermediates.

297

For these experiments we assume a simple mechanism as follows:

298

Styrene + O3

299

sCI + SO2

intermediates (SO3)

300

sCI+ H2O

P3 (3)

301

Where sCI includes both ·CH2OO· and C6H5C·HOO· (Tuazon et al., 1993), the latter being

302

possible as conformer syn or anti. (Vereecken and Francisco, 2012).

303

For the case of ·CH2OO·, the reaction with SO2 has been studied previously and the available

304

experimental data show that it leads to the formation of SO3 with a unity yield (Berndt et al.,

305

2014a; Berndt et al., 2014b). In this study, for the experiments carried out in the drier

306

conditions (RH close to zero), the water concentration ratio is in the order of 2ppm (5x1013

307

molecule cm-3). The H2O-SO3 gas-phase reaction is very fast, k= 3.90x10-41exp(6830.6/T)[H2O]2

308

(Jayne et al., 1997) and the gas-phase water concentration is high enough to react with SO3,

309

leading to H2SO4 (Díaz-de-Mera et al., 2017), which is the species assumed to generate new

310

particles in the system.

TE D

M AN U

SC

RI PT

280

HCHO + C6H5CHO + CI + sCI

(1)

AC C

EP

secondary aerosols (2)

12

ACCEPTED MANUSCRIPT For the other sCI formed from the styrene ozonolysys, C6H5C·HOO·, the oxidation channel

312

leading to SO3, and then H2SO4, is also possible and can contribute to the nucleation step.

313

Recent works have shown that the oxidation channel also generates H2SO4 quantitatively in

314

the ozonolysis of other alkenes such as trans-2-butene and 2,3-dimethyl-2-butene when it

315

happens in the presence of SO2 (Bernd et al., 2014a). Furthermore, the crossed reaction

316

between the biradical C6H5C·HOO· and formaldehyde has been suggested to produce 3,5-

317

diphenyl-1,2,4-trioxolane (Na et al., 2006) and the secondary ozonide may also undergo partial

318

conversion into a hydroxyl-substituted ester (Tuazon et al., 1993). These low volatile products

319

may contribute to the growing of particles.

320

In this work we have studied the direct yield for the reaction between sCI and SO2. As shown

321

previously, the correlation between the measured MPC and the reacted styrene points out the

322

ozonolysis as the rate limiting step for SOA production. Thus, the sCI/SO2 reaction is expected

323

to be much faster than the styrene/ozone reaction and, so, the amount of consumed SO2

324

should correlate with the total reacted styrene. In this sense, Figure 6 shows the results for

325

several experiments with different initial concentrations of styrene and ozone under dry

326

conditions. The simulated profiles for total reacted styrene are nearly straight lines since its

327

precursors, styrene and ozone remain essentially constant during the experiment. The total

328

reacted SO2, measured experimentally, also shows linear temporal profiles except for the

329

experiments with the higher concentrations of styrene and ozone. From the ratio of the slopes

330

of the linear fits (reacted SO2 and reacted styrene) we derived an average yield for reaction (2)

331

with respect to the initial reaction (1), Table 1, η= 0.51±0.13 (errors are 2σ±20%).

332

For the experiments with higher concentrations, for example 200 ppb of both styrene and

333

ozone, the SO2 concentration dropped from 20 to 10 ppb after 25 minutes lowering the rate of

334

the sCI/SO2 reaction and driving to the curvature of the temporal profile of the reacted SO2.

335

For such experiments the yield for reaction (2) was obtained from the initial data (20 min in

336

the shown experiment) where the SO2 data can still be fitted to a straight line.

337

The measurement of the yield for reaction 2 could not be carried out within a wide range of

338

[SO2] initial concentrations since the ozonolysis reaction is slow and increasing SO2 leads to

339

very small changes in the SO2 signal.

340

The results show that sCI react quantitatively through reaction (2) in the presence of SO2 with

341

a constant yield regardless the initial concentrations of styrene and ozone. However, as sCI can

342

react through different competitive sinks, it is expected that the yield of reaction (2) depends

343

on the concentration of SO2, water and other possibly involved species.

AC C

EP

TE D

M AN U

SC

RI PT

311

13

ACCEPTED MANUSCRIPT 344 345

Table 1. Yield of reaction (2) (consumption of SO2) after the ozonolysis of styrene (reaction 1).

346

The yields are obtained from the ratio of the reacted SO2, and the reacted styrene. Errors are

347

2σ. [O3] /ppb

[SO2] /ppb

RH

η

200

100

20

0%

0.47± 0.05

100

200

20

0%

0.55±0.06

100

100

20

0%

0.54±0.05

200

200

20

0%

0.48±0.06

SC

RI PT

[Styrene]/ppb

4.0E+11 3.5E+11

15

TE D

3.0E+11

20

2.5E+11 2.0E+11 1.5E+11 1.0E+11

10

5

EP

Concentration (molecule/cm3)

4.5E+11

5.0E+10

AC C

0.0E+00

0

5

10

0 15

20

25

30

35

40

45

Time (min)

349

Figure 6. SO2 experimental decays for 100 ppb of styrene and ozone (initial concentrations)

350

under dry conditions and with RH=5 %. Simulated profiles for total reacted styrene under

351

different initial concentrations of styrene and ozone (see the legend) and the corresponding

352

experimental profiles for total reacted SO2. The initial SO2 concentration was 20 ppb for all the

353

experiments displayed.

354 355

3.4. Effect of water vapour. 14

50

[SO2] (ppb)

5.0E+11

348

Reacted SO2, [Sty]=[O3]=200ppb reacted SO2, [Sty]=200 ppb; [O3]=100ppb Reacted SO2, [Sty]=[O3]=100ppb SO2 decay. RH=0%

M AN U

Reacted styrene, [Sty]=[O3]=200ppb Reacted styrene, [Sty]=200 ppb; [O3]=100ppb Reacted styrene, [Sty]=[O3]=100ppb SO2 decay. RH=5%

ACCEPTED MANUSCRIPT A series of experiments with concentrations or reactants around 10 ppb was carried out to

357

study the effect of RH on the production of SOA. For these experiments the PMC values were

358

very small, always below 0.5 µg m-3 and the concentrations had to be increased to get more

359

accurate mass data. Figures 7 and S6-S8 show the results for PNC, PMC and average particles’

360

size diameters for experiments with 20 ppb of styrene and SO2, 40 ppb of ozone and RH in the

361

range 0-35 %.

362

As it is found in Figure 7, the addition of water, induces a significant drop (mainly in the RH

363

range below 10 %) on the PNbC and inhibits their growing leading to lower production of SOA.

364

These results are consistent with the suggested mechanism where SO2 and H2O compete to

365

scavenge the Criegee intermediates (reactions (2) and (3)). The reaction with water vapour is

366

expected to proceed through the addition of the water oxygen atom to the carbon atom in the

367

carbonyl oxide and the simultaneous transfer of one hydrogen atom from the water molecule

368

to the terminal O atom in the CI (Anglada et al., 2011), leading to α-hydroxi-hydroperoxides

369

that can subsequently decompose or react with other atmospheric species.

M AN U

SC

RI PT

356

PMC

2.0

PMC (µg/m3)

12

TE D

2.5

16

DeltaPMC/PMC

1.5

EP

1.0

8

AC C

0.5

4

0.0

0

370

5

∆PMC/PMC

3.0

0

10

15

20

25

30

35

40

% RH

371

Figure 7. Water effect on the particle mass concentration. For this series of experiments, the

372

concentration ratios were 20 ppb, 20 ppb and 40 ppb for styrene, SO2 and ozone, respectively.

373

The plotted PMC data plotted for each experiment are those obtained at 40’ reaction time.

374

15

ACCEPTED MANUSCRIPT The rates of the sCI reactions towards SO2 and H2O depend on their concentrations and their

376

kinetic rate constants, kSO2 and kH2O. Since both [H2O] and [SO2] are known (and essentially

377

constant during the experiments), experimental data of the products of these reaction can

378

provide the ratio of theses competitive reactions (kH2O/kSO2). Thus, considering the dry

379

experiments (0% RH) as reference, we can assume that for the experiments carried out in the

380

presence of water the decrease in mass concentration, ∆PMC = PMCRH=0 – PMC, is due

381

exclusively to the reaction of the sCI with water molecules. Thus, the ratio (kH2O

382

[H2O])/(kSO2[SO2]) can be obtained as the ratio ∆PMC /PMC

383

assuming that the reaction of sCI with SO2 is the sole source of SOA. Then, the linear fit of the

384

plot (∆PMC /PMC) versus [H2O], Figure 7, provides the (kH2O/(kSO2[SO2])) ratio from the slope,

385

and then the ratio between the rate constants,

386

concentration is known, 20 ppb=4.9x1011 molecule cm-3, errors are 2σ±20%).

387

This result is similar to other reported values for different CI (Berndt et al., 2014a, Díaz-de-

388

Mera et al., 2017) and may be used to assess the effects of SO2 and water vapour in the

389

formation of SOA under different atmospheric conditions.

390

4. Conclusions.

391

Previous studies have shown the outstanding potential of SOA formation due to the emissions

392

of aromatic compounds to the atmosphere. Among them, styrene is one of the aromatics with

393

the highest contribution to SOA. Although the formation of new particles from the ozonolysis

394

of styrene has been reported from laboratory experiments, this reaction can’t justify such high

395

capacity to form new particles since it would require high concentrations of reactants

396

compared to those of real polluted atmospheres. Generally, atmospheric models based on

397

laboratory chamber studies predict far less SOA than is observed in field studies (Kroll and

398

Seinfeld, 2008). This fact suggest that other yet unknown pathways may lead to the formation

399

of new particles more effectively. In this sense, SO2 may play a key role since, as found in this

400

work, the ozonolysis of styrene in the presence of low concentrations of SO2 readily produces

401

new particles. The nucleation events occur at concentration around (5.6±1.7)x108 molecule cm-

402

3

403

The particle mass concentration temporal profiles correlate with the reacted styrene showing

404

that the ozonolysis of styrene is the rate limiting step in the formation of new particles. The

405

fractional yields for different initial concentrations of reactants have been obtained based on

406

the amount of reacted styrene. They are independent on the concentrations of styrene or

RI PT

375

(Díaz-de-Mera et al., 2017),

(the SO2

AC C

EP

TE D

M AN U

SC

kH2O/kSO2= (2.8±0.7)×10-5

, that is, 3 orders of magnitude below that of the reaction in the absence of SO2.

16

ACCEPTED MANUSCRIPT ozone but increase from nearly zero in the absence of SO2 to approximately 2 when SO2 is in

408

excess.

409

The experimental temporal decays of SO2 show that it reacts with the products of the

410

ozonolysis of styrene. From the simulated profiles of reacted styrene, under low water

411

concentration conditions (below 2 ppm) the yield for the consumption of SO2 was η=

412

0.51±0.13. Thus, the formation of SO3 from the reaction of styrene with the Criegee

413

intermediates is expected to be responsible of the nucleation events through the formation of

414

H2SO4. On the other hand, the presence of water in the reaction media competes with SO2 and

415

inhibits the formation of SOA. So the potential formation of SOA under atmospheric conditions

416

depends on the concentration of SO2 and relative humidity.

417

As shown in a previous work (le Person et al. 2006) OH-based and ozone-based lifetimes for

418

styrene are 2.4h and 25h, respectively and so both reactions contribute as efficient gas-phase

419

sinks. Although the initial OH reaction is faster than ozonolysis, it is not clear whether the OH

420

reaction products contribute efficiently to the formation of aerosols, as it has been found for

421

ozonolysis in the presence of SO2. In this sense, the study of particles generation from OH

422

reactions is also required to fully understand the contribution of styrene to the formation of

423

secondary organic aerosols in urban areas.

424

Acknowledgments.

425

This work was supported by the Spanish Ministerio de Economía y Competitividad (project

426

CGL2014-57087-R, FEDER co-funding) and by the University of Castilla La Mancha (projects

427

GI20152950 and GI20163433).

EP

TE D

M AN U

SC

RI PT

407

AC C

428 429

References.

430

Anglada, J.M., González, J., Torrent-Sucarrat, M, 2011. Effects of the substituents on the

431

reactivity of carbonyl oxides. A theoretical study on the reaction of substituted carbonyl oxides

432

with water. Physical Chemistry Chemical Physics, 13, 13034–13045.

433

Atkinson, R. and Aschmann, S.M., 1988. Kinetics of the reactions of acenaphthene and

434

acenaphthylene and structurally-related aromatic compounds with OH and NO3 at 296±2K.

435

International Journal of Chemical Kinetics, 20, 513-539.

436

ATSDR, 2017. https://www.atsdr.cdc.gov/toxprofiles/tp53-c3.pdf accesed June 2017.

17

ACCEPTED MANUSCRIPT Bates, J. T., Weber, R. J., Abrams, J., Verma, V., Fang, T., Klein, M., Strickland, M. J., Sarnat, S.

438

E., Chang, H. H., Mulholland, J. A., Tolbert, P. E., Russell, A. G., 2015. Reactive oxygen species

439

generation linked to sources of atmospheric particulate matter and cardiorespiratory effects.

440

Environmental Science & Technology, 49(22), 13605-13612. DOI: 10.1021/acs.est.5b02967.

441

Berndt, T., Jokinen, T., Sipilä, M., Mauldin, R. L., Herrmann, H., Stratmann, F., Junninen, H.,

442

Kulmala, M., 2014a. H2SO4 formation from the gas-phase reaction of stabilized Criegee

443

Intermediates with SO2: Influence of water vapour content and temperature. Atmospheric

444

Environment, 89, 603-612. DOI: org/10.1016/j.atmosenv.2014.02.062.

445

Berndt, T., Voigtländer, J., Stratmann, F., Junninen, H., Mauldin, R. L., Sipilä, M., Kulmala, M.,

446

Herrmann, H., 2014b. Competing atmospheric reactions of CH2OO with SO2 and water vapour.

447

Physycal Chemistry Chemical Physics, 16, 19130-19136. DOI: 10.1039/C4CP02345E.

448

Cheung H.C., Chou C.C.K., Huang, W.R., Tsai, C.Y., 2013. Characterization of ultrafine particle

449

number concentration and new particle formation in an urban environment of Taipei, Taiwan.

450

Atmospheric Chemistry and Physics, 13(17), 8935–8946. DOI: 10.5194/acp-13-8935-2013.

451

Diaz-de-Mera, Y., Aranda, A., Bracco, L., Rodriguez, D., Rodriguez, A., 2017. Formation of

452

secondary organic aerosols from the ozonolysis of dihydrofurans. Atmospheric Chemistry and

453

Physics, 17(3), 2347-2357. DOI: org/10.5194/acp-17-2347-2017.

454

Duke, J.A., 1985, CRC handbook of medical herbs. Boca Raton, FL, CRC Press, p. 323.

455

EPA

456

https://www.epa.gov/haps/initial-list-hazardous-air-pollutants-modifications

457

2017.

458

Hu, N., Tan, J., Wang, X., Zhang, X., Yu, P., 2017. Volatile organic compound emissions from an

459

engine fueled with an ethanol-biodiesel-diesel blend. Journal of the Energy Institute, 90, 101-

460

109. DOI: org/10.1016/j.joei.2015.10.003.

461

Jayne, J. T., Pöschl, U., Chen, Y., Dai, D., Molina, L.T., Worsnop, D.R., Kolb, C.E., Molina, M. J.,

462

1997. Pressure and temperature dependence of the gas-phase reaction of SO3 with H2O and

463

the heterogeneous reaction of SO3 with H2O/H2SO4 surfaces. Journal of Physical Chemistry A,

464

101, 10000-10011. DOI: 10.1021/jp972549z.

465

Jiang, L., Xu, Y., Ding, A., 2010. Reaction of stabilized Criegee intermediates from ozonolysis of

466

limonene with sulfur dioxide: ab initio and DFT study. J. Phys. Chem. A, 114, 12452-12461.

467

DOI: 10.1021/jp107783z.

Initial

list

of

hazardous

air

pollutants

with

modifications. accesed

June

AC C

EP

2017,

TE D

M AN U

SC

RI PT

437

18

ACCEPTED MANUSCRIPT Kabir, E. and Kim, K. H., 2010. An on-line analysis of 7 odorous volatile organic compounds in

469

the ambient air surrounding a large industrial complex. Atmospheric Environment, 44(29),

470

3492-3502. DOI: org/10.1016/j.atmosenv.2010.06.021.

471

Knighton, W. B., Herndon, S.C., Wood, E.C., Fortner, E.C.,Onasch, T. B., Wormhoudt, J., Kolb, C.

472

E., Lee, B.H., Zavala, M., Molina, L., Jones, M., 2012. Detecting fugitive emissions of 1,3-

473

butadiene and styrene from a petrochemical facility: An application of a mobile laboratory and

474

a modified proton transfer reaction mass spectrometer. Industrial & Engineering Chemistry

475

Research, 51 (39), 12706–12711. DOI: 10.1021/ie202794j.

476

Kos, G., Adechina, N., Kanthasamy, V., Ariya, P. A.,2014. Volatile organic compounds in Arctic

477

snow: Concentrations and implications for atmospheric processes.

478

Science: Processes & Impacts 16 (11), 2592-2603. DOI: 10.1039/C4EM00410H.

479

Kroll, J. H., Seinfeld, J. H., 2008. Chemistry of secondary organic aerosol: Formation and

480

evolution of low-volatility organics in the atmosphere. Atmospheric Environment, 42(16),

481

3593-3624.

482

Kurtén, T., Lane, J. R., Jørgensen, S., Kjaergaard, H. G.,2011. A Computational Study of the

483

Oxidation of SO2 to SO3 by Gas-Phase Organic Oxidants. J. Phys. Chem. A, 115 (31), 8669–8681.

484

DOI: 10.1021/jp203907d.

485

Le Person, A., Eyglunent, G., Daele, V., Mellouki, A., Mu, Y. 2008. The near UV absorption

486

cross-sections and the rate coefficients for the ozonolysis of a series of styrene-like

487

compounds. Journal of Photochemistry and Photobiology A: Chemistry, 195(1), 54-63. DOI:

488

org/10.1016/j.jphotochem.2007.09.006.

489

Li, L., Li, H., Zhang, X., Wang, L., Xu, L., Wang, X., Yu, Y., Zhang, Y., Cao, G. 2014. Pollution

490

characteristics and health risk assessment of benzene homologues in ambient air in the

491

northeastern urban area of Beijing, China. Journal of Environmental Sciences, 26(1), 214-223.

492

https://doi.org/10.1016/S1001-0742(13)60400-3.

493

Li, L., Ge, Y., Wang, M., Peng, Z., Song, Y., Zhang, L., Yuan, W., 2015. Exhaust and evaporative

494

emissions from motorcycles fueled with ethanol gasoline blends. Science of the Total

495

Environment, 502, 627-631. DOI: org/10.1016/j.scitotenv.2014.09.068.

496

Li, L., Tang, P., Nakao, S., Kacarab, M., Cocker III, D. R., 2016. Novel approach for evaluating

497

secondary organic aerosol from aromatic hydrocarbons: Unified method for predicting aerosol

498

composition and formation. Environmental Science & Technology, 50(12), 6249-6256. DOI:

499

10.1021/acs.est.5b05778.

RI PT

468

AC C

EP

TE D

M AN U

SC

RSC, Environmental

19

ACCEPTED MANUSCRIPT Liu, T., Wang, X., Hu, Q., Deng, W., Zhang, Y., Ding, X., Fu, X., Bernard, F., Zhang, Z., Lü, S., He,

501

Q., Bi, X., Chen, J., Sun, Y., Yu, J., Peng, P., Sheng, G., and Fu, J. 2016. Formation of secondary

502

aerosols from gasoline vehicle exhaust when mixing with SO2, Atmos. Chem. Phys., 16, 675-

503

689, doi:10.5194/acp-16-675-2016.

504

Lopes, M., Serrano, L., Ribeiro, I., Cascão, P., Pires, N., Rafael, S., Tarelho,L., Monteiro, A.,

505

Nunes, T., Evtyugina, M., Nielsen, O.J., Gameiro da Silva, M., Miranda, A.I., Borrego, C., 2014.

506

Emissions characterization from EURO 5 diesel/biodiesel passenger car operating under the

507

new

508

org/10.1016/j.atmosenv.2013.11.071.

509

Ma, Y., Porter, R.A., Chappell, D., Andrew, A.T., Russell, T., Marston, G., 2009. Mechanisms for

510

the formation of organic acids in the gas-phase ozonolysis of 3-carene. Physical Chemistry

511

Chemical Physics, 11, 4184–4197.

512

Mauldin, R. L., Berndt, T., Sipilä, M., Paasonen, P., Petäjä, T., Kim, S., Kurtén, T., Stratmann, F.,

513

Kerminen, V. M., Kulmala, M., 2012. A new atmospherically relevant oxidant of sulphur

514

dioxide. Nature, 488, 193–196. DOI: 10.1038/nature11278.

515

Mejía, J. F., Morawska, L., and Mengersen, K., 2008. Spatial variation in particle number size

516

distributions in a large metropolitan area, Atmos. Chem. Phys., 8, 1127-1138. Doi:10.5194/acp-

517

8-1127-2008.

518

Na, K., Song, C., Cocker, D. R., 2006. Formation of secondary organic aerosol from the reaction

519

of styrene with ozone in the presence and absence of ammonia and water. Atmospheric

520

Environment, 40(10), 1889-1900. DOI: org/10.1016/j.atmosenv.2005.10.063.

521

Nel, A. 2005. Air pollution-related illness: effects of particles. Science, 308(5723), 804-806. DOI:

522

10.1126/science.1108752.

523

Newland, M. J., Rickard, A. R., Alam, M. S., Vereecken, L. Muñoz, A., Ródenas, M., Bloss, W.J.,

524

2015. Kinetics of stabilised Criegee intermediates derived from alkene ozonolysis: reactions

525

with SO2, H2O and decomposition under boundary layer conditions. Physical Chemistry

526

Chemical Physics, 17, 4076-4088. DOI: 10.1039/C4CP04186K.

527

Ng, N. L., Kroll, J. H.; Chan, A. W. H., Chhabra, P. S., Flagan, R. C., Seinfeld, J. H., 2007.

528

Secondary organic aerosol formation from mxylene, toluene, and benzene. Atmospheric

529

Chemistry and Physics. 2007, 7 (14), 3909− 3922. DOI: 10.5194/acp-7-3909-2007.

530

Odum, J.R., Hoffman, T., Bowman, F., Collins, D., Flagan, R.C., Sienfeld, J.H., 1996. Gas/particle

531

partitioning and secondary aerosol yields. Enviromental Science & Technology, 30, 2580-2585.

cycle.

Atmospheric

Environment,

84,

339-348.

DOI:

SC

driving

AC C

EP

TE D

M AN U

European

RI PT

500

20

ACCEPTED MANUSCRIPT 532

Sarigiannis, D.A., Karakitsios, S.P., Gotti, A., Liakos, I.L., Katsoyiannis, A., 2011. Exposure to

533

major volatile organic compounds and carbonyls in European indoor environments and

534

associated

535

org/10.1016/j.envint.2011.01.005.

536

Shimada, K., Kimura, E., Yasui, Y., Tanaka, H., Matsushita, S., Hagihara, H., Nagakura, M,

537

Kawahisa, M., 1992. Styrene formation by the decomposition by Pichia carsonii of trans-

538

cinnamic acid added to a ground fish product. Applied Environmental Microbiology, 58, 1577–

539

1582.

540

Sun, J., Wu, F., Hu, B., Tang, G., Zhang, J., Wang, Y., 2016. VOC characteristics emissions and

541

contributions to SOA formation during hazy episodes. Atmospheric Environment, 141, 560-

542

570. DOI: org/10.1016/j.atmosenv.2016.06.060.

543

Taatjes C.A., Welz, O., Eskola, A.J., Savee, J.D., Scheer A.M., Shallcross, D.E., Rotavera, B., Lee,

544

E.P., Dyke, J.M., Mok, D.K., Osborn, D.L., Percival, C.J., 2013. Direct measurements of

545

conformer-dependent reactivity of the Criegee intermediate CH3CHOO. Science.

546

340(6129), 177-180. DOI: 10.1126/science.1234689.

547

Tuazon, E. C., Arey, J., Atkinson, R., Aschmann, S. M., 1993. Gas-phase reactions of 2-

548

vinylpyridine and styrene with hydroxyl and NO3 radicals and ozone. Environmental Science &

549

Technology, 27(9), 1832-1841. DOI: 10.1021/es00046a011.

550

Vereecken, L. and Francisco, J. S., 2012. Theoretical studies of atmospheric reaction

551

mechanisms in the troposphere. Chemical Society Reviews, 41(19), 6259-6293. DOI:

552

10.1039/C2CS35070J.

553

Wang, H., Ji, Y., Gao, Y., Li, G., An, T. 2015. Theoretical model on the formation possibility of

554

secondary organic aerosol from OH initialed oxidation reaction of styrene in the presence of

555

O2/NO. Atmospheric Environment, 101, 1-9. DOI: org/10.1016/j.atmosenv.2014.10.042.

556

Welz, O., Savee, J. D., Osborn, D. L., Vasu, S. S., Percival, C. J, Shallcross, D., Taatjes, C., 2012.

557

Reaction of CH2I with O2 forms Criegee Intermediate: Direct Measurements of CH2OO Kinetics.

558

Science, 335, 204-207. DOI: 10.1126/science.1213229.

559

Xu, J., Szyszkowicz, M., Jovic, B., Cakmak, S., Austin, C. C., Zhu, J., 2016. Estimation of indoor

560

and outdoor ratios of selected volatile organic compounds in Canada. Atmospheric

561

Environment, 141, 523-531. DOI: org/10.1016/j.atmosenv.2016.07.031.

562

Zhang, Z., Wang, H., Chen, D., Li, Q., Thai, P., Gong, D., Li, Y., Zhang, C., Gu, Y., Zhou, L.,

563

Morawska, L., Wang, B., 2017. Emission characteristics of volatile organic compounds and their

risk.

Environment

International,

37,

743–765.

DOI:

12,

AC C

EP

TE D

M AN U

SC

RI PT

health

21

ACCEPTED MANUSCRIPT 564

secondary organic aerosol formation potentials from a petroleum refinery in Pearl River Delta,

565

China.

566

10.1016/j.scitotenv.2017.01.179.

of

the

Total

Environment,

584-585,

1162-1174.

DOI:

AC C

EP

TE D

M AN U

SC

RI PT

Science

22

ACCEPTED MANUSCRIPT 1

Formation of secondary aerosols from the ozonolysis of styrene:

2

effect of SO2 and H2O.

3

Diaz-de-Mera, Yolanda1; Aranda, Alfonso1; Martínez, Ernesto2; Rodríguez, Ana Angustias2;

4

Rodríguez, Diana3; Rodriguez, Ana3.

5

1

6

Camilo José Cela s/n, 13071. Ciudad real. Spain.

7

2

8

Atmosférica, Camino Moledores, s/n, 13071 – Ciudad Real. Spain.

9

3

RI PT

Universidad de Castilla-La Mancha. Facultad de Ciencias y Tecnologias Quimicas. Avenida

Universidad de Castilla-La Mancha. Instituto de Investigación en Combustión y Contaminación

SC

Universidad de Castilla-La Mancha. Facultad de Ciencias Ambientales y Bioquímica. Avenida

Carlos III s/n, 45071 Toledo, Spain.

11

Correspondence to: Alfonso Aranda ([email protected])

M AN U

10

12

Abstract. In this work we report the study of the ozonolysis of styrene and the reaction

14

conditions leading to the formation of secondary aerosols. The reactions have been carried out

15

in a Teflon chamber filled with synthetic air mixtures at atmospheric pressure and room

16

temperature. We have found that the ozonolysis of styrene in the presence of low

17

concentrations of SO2 readily produces new particles under concentrations of reactants lower

18

than those required in experiments in the absence of SO2. Thus, nucleation events occur at

19

concentrations around (5.6±1.7)x108molecule cm-3 (errors are 2σ±20%) and SO2 is consumed

20

during the experiments. The reaction of the Criegee intermediates with SO2 to produce SO3 and

21

then H2SO4 may explain (together with OH reactions’ contribution) the high capacity of styrene

22

to produce particulate matter in polluted atmospheres.

23

The formation of secondary aerosols in the smog chamber is inhibited under high H2O

24

concentrations. So, the potential formation of secondary aerosols under atmospheric conditions

25

depends on the concentration of SO2 and relative humidity, with a water to SO2 rate constants

26

ratio kH2O/kSO2= (2.8±0.7)×10-5 (errors are 2σ±20%).

AC C

EP

TE D

13

27 28

1. Introduction.

29

Styrene is one of the most important monomers worldwide. The double bond in the vinyl group

30

enables polymerisation and so the production of widespread used materials such as polystyrene, 1

ACCEPTED MANUSCRIPT styrene–butadiene rubber, styrene–butadiene latex, etc. Despite the high reactivity of styrene

32

in the atmosphere, it can be considered as an ubiquitous species due to the wide range of

33

different anthropogenic sources. Thus, it has been measured in remote areas (Kos et al., 2014 ),

34

in indoor air samples (Sarigiannis et al., 2011; Xu et al., 2016), or in ambient air (Kabir and Kim,

35

2010; Li et al., 2014).

36

Styrene is naturally produced as a degradation product in cinnamic acid containing plants (Duke,

37

1985) and as a by-product of fungi and microbe metabolism (Shimada et al., 1992). Nevertheless,

38

the presence of styrene in the atmosphere is mainly due to fugitive emissions from

39

petrochemical facilities and industrial processes involving styrene or its polymers (Knighton et

40

al., 2012; Zhang et al., 2017). Other sources include the release from styrene-based materials,

41

incineration of styrene polymers, or cigarette smoke for indoor air. The combustion of

42

traditional petroleum-based fuels or the alternative newly developed biofuels are also

43

responsible of important exhaust emissions of styrene mainly in populated areas (Lopes et al.,

44

2014; Li et al., 2015; Hu et al., 2017).

45

Styrene is a hazardous air pollutant and so its emissions are regulated by the environmental

46

protection agencies (EPA, 2017). The exposure to inhalation or intake of styrene through food

47

or drinks may lead to severe health effects (ATSDR, 2017).

48

The main gas phase reactions of styrene in the atmosphere have been studied previously, being

49

the reported room-temperature kinetic rate constants with OH, NO3 and ozone (in cm3 molecule-

50

1 -1

51

respectively. The results obtained by Le Person et al. (Le Person et al., 2008) show that reactions

52

with OH, NO3 and O3 are all important atmospheric loss processes for styrene in the gas phase.

53

On the other hand, photolysis of styrene constitutes a slower sink during daylight hours.

54

Recently, exposure to particulate matter is being associated to the decrease of life expectancy

55

and numerous diseases, especially cardiopulmonary illness (Bates et al., 2015). The size of

56

particles is a key factor for health effects since the potential ability to access the lungs becomes

57

higher for the smaller ultrafine particles (Nel, 2005). This is a problem in urban air where

58

particles with size diameter below 100 nm may account for 80–90 % of the total particle number

59

concentration (Mejía et al., 2008). Apart from direct primary sources like vehicle exhausts, one

60

of the main sources of ultrafine particles is the formation of secondary organic aerosols (SOA)

61

from the reactions of the emissions of gas-phase pollutants (Cheung et al., 2013). In this sense,

62

it has been found that aromatic compounds readily yield SOA under urban air conditions (Ng et

63

al., 2007). The contribution of aromatic compounds to the total SOA formation may account

TE D

M AN U

SC

RI PT

31

AC C

EP

s ): 5.8x10-11, 1.51x10-13 (Atkinson and Aschmann, 1988) and 1.5x10-17 (Le Person et al., 2008),

2

ACCEPTED MANUSCRIPT from 78.5 to 91.9 %, well above those of alkanes and alkenes (Sun et al., 2016). Thus, given their

65

high SOA formation potential, new approaches to predict the amount of aerosol formed from

66

the oxidation of aromatic hydrocarbons are being developed (Li et al., 2016). In the case of

67

styrene, it was identified as the second most efficient species forming SOA (15%), only after

68

toluene (16%) (Sun et al., 2016).

69

For styrene, a theoretical model shows that its reaction with OH could contribute to the

70

formation of SOA in polluted atmospheres (Wang et al., 2015). Furthermore, a previous

71

experimental work has also reported the formation of SOA from the reaction of styrene with

72

ozone and the effect of the presence of ammonia and water (Na et al., 2006).

73

Being styrene an alkene, its reaction with ozone proceeds through the formation of a Criegee

74

intermediate (CI). Very recently, it has been found that stabilised CI (sCI) can undergo reactions

75

with SO2 several orders of magnitude faster than assumed so far (Welz et al., 2012) producing

76

SO3, which contributes efficiently to the formation of ground level sulfuric acid (Mauldin et al.,

77

2012; Newland et al., 2015). Likewise, mixing SO2 with gasoline vehicle exhausts has been shown

78

to increase the formation of secondary aerosols (Liu et al. 2016). Thus, the reaction with SO2

79

could be the key to explain the high SOA formation potential of styrene under polluted

80

conditions. On the other hand, the competition of SO2 and water vapour in the removal of sCI in

81

the atmosphere depends on the CI structure (Berndt et al., 2014a; Díaz-de-Mera et al., 2017)

82

and may play an important effect on the net SOA formation of this aromatic compound.

83

In this work we study the ozonolysis reaction of styrene following the conditions that lead to the

84

formation and growth of new particles. The effect of water vapour and SO2 concentrations

85

during the process are also studied and discussed.

86

EP

TE D

M AN U

SC

RI PT

64

2. Experimental.

88

The experimental system consisted on a collapsible 200L Teflon reactor coupled to different

89

detectors to analyse both the gas phase and particulate matter. The reactants’ samples, styrene,

90

cyclohexane and SO2, were prepared in a volume-calibrated glass bulb with 100 and 1000Torr

91

full scale capacitance pressure gauges, MKS 626AX, and then flushed to the reactor using

92

synthetic air. Concentrations for each species were obtained from their partial pressures in the

93

glass bulb and the final volume of the Teflon reactor.

94

The ozone was generated in-situ by passing pure oxygen through a high voltage electric

95

discharge, BMT Messtechnik 802N, and then fed to the sampling bulb and to an absorption 10

96

cm long quartz cell to measure the concentration using a UV-vis Hamamatsu spectrometer,

AC C

87

3

ACCEPTED MANUSCRIPT 97

C10082CAH at 255nm. Once the concentration in the cell and bulb were known, it was diluted

98

with synthetic air to get the required concentration in the reactor.

99

Cyclohexane in excess was used as OH radical scavenger in concentration ratios so that >97% of OH formed in the reactions was removed (Ma et al., 2009).

101

Water vapour in the reactor was introduced by passing the air flows carrying the reactants

102

through a glass bubbler containing the calculated amount of water required for any given

103

relative humidity (RH). The sampled water was then evaporated and flushed inside the reactor

104

through the flow of synthetic air. For the series of experiments carried out under “dry

105

conditions” the synthetic air was used directly from the cylinder, with a water mixing ratio below

106

2ppm, which is the concentration stated by the supplier.

107

Typically, the reactants were sequentially introduced as follows: first, styrene, second

108

cyclohexane (OH scavenger), then SO2 and finally ozone. During the injection of ozone, the

109

reactor was shaken to improve mixing and then the valves to the sampling detectors were open.

110

Liquid water (when necessary) was introduced in the bubbler at the beginning, being flushed by

111

the air carrying the rest of gaseous reactants.

112

The concentration of particles was monitored by an scanning mobility particle sizer (SMPS), TSI

113

SMPS 3938 with a CPC particle counting 3775 and a nano-differential mobility analyser which

114

improves size resolution over the particle size range 4.0-150nm. The SMPS was operated with

115

3.0 L min-1 sheath flow and 0.3 L min-1 sampling flow rate. Size distributions were recorded at 1

116

minute intervals from the injection of ozone and the total aerosol mass concentration was

117

derived from the measured particle size distribution assuming unit density and spherical

118

particles. Prior to each experiment the Teflon bag was repetitively washed with clean synthetic

119

air until the level of particles was below 1 cm-3.

120

SO2 concentrations were measured by a fluorescence SO2 analyser, Teledyne Instruments 101-

121

E. Previous runs with only SO2/air samples in the reactor showed that SO2 wall losses were

122

negligible. Additionally, no interference in the SO2 fluorescence signal from the rest of co-

123

reactants was observed in the range of concentrations used in this study.

124

Ozone concentrations were measured with an ozone analyser, Environment O342M. Several

125

tests showed that O3 losses in the reactor were negligible within the time scale of the

126

experiments. Styrene also shows UV absorption at 255nm. Thus, several samples of styrene in

127

air were introduced in the reactor to calibrate the O3 analyzer for the styrene measurements.

128

After a few minutes an important decrease in the styrene signal was observed (attributed to

129

polymerisation within the UV detection cell) what prevents from monitoring the styrene

AC C

EP

TE D

M AN U

SC

RI PT

100

4

ACCEPTED MANUSCRIPT temporal profiles in the reactor continuously. Nevertheless, when sampling of styrene was

131

carried out intermittently for short time periods (and letting the UV absorption cell to

132

completely evacuate the previous sample) the signal always recovered its initial intensity. Under

133

dark conditions with different styrene concentrations in the reactor, no measurable

134

consumption of styrene was observed from the periodic samples measured in the UV

135

spectrometer for 80 minutes runs.

136

Reagents.

137

All the substances used were of the highest commercially available purity. Liquid reagents were

138

purified by successive trap to trap distillation. Styrene >99%, Sigma-Aldrich; SO2 99,9%, Fluka;

139

cyclohexane 99.5%, Sigma-Aldrich; synthetic air 3X, Praxair; O2 4X, Praxair.

SC

RI PT

130

140 3. Results and discussion.

142

3.1. Nucleation conditions.

143

For asymmetric alkenes like styrene, the ozonolylis leads to the formation of two different

144

Criegee intermediates and the corresponding aldehydes: C6H5C·HOO· and HCHO or ·CH2OO· and

145

C6H5CHO. For the title reaction, previous results show that the two pathways (leading to

146

formaldehyde or benzaldehyde) are quantitative and with yields around 40 % (Tuazon et al.,

147

1993). The subsequent reactions of the primary products, especially the Criegee intermediates

148

(Na et al., 2006, Bernd et al., 2014a), may lead to the formation of secondary aerosols.

149

In order to approach real ambient air conditions, a series of experiments were carried out with

150

ozone and styrene using concentrations lower than those of the study by Na et al., 2006. Thus

151

for example, for initial styrene and ozone concentrations in the range of 10 ppbs, no particles

152

were observed in the reactor after 60 minutes runs. The concentrations were then progressively

153

increased and, for example, for experiments with concentrations of 100 ppb of styrene and 200

154

ppb of ozone, particles generated after 15 minutes. Since the gas-phase reaction of styrene with

155

ozone has been reported previously with k295K=(1.5±0.3)x10-17cm3 molecule-1 s-1 (Le Person et al.,

156

2008), this kinetic rate constant can been used to simulate the ozone and styrene concentrations

157

profiles. Thus, from the previous experiments, for 15 minutes reaction time the results show

158

that the concentration of reacted styrene required for the formation of new particles is very

159

high, (1.6±0.5)x1011 molecule cm-3 (errors are 2σ±20%) and so the reaction of styrene with

160

ozone is unlikely, by itself, to drive nucleation events under usual atmospheric conditions.

AC C

EP

TE D

M AN U

141

5

ACCEPTED MANUSCRIPT On the other hand, when carrying the experiments in the presence of SO2, particles were

162

generated at lower concentrations of reactants. Figure 1 shows the typical experimental profiles

163

of the particle number concentration (PNbC), particle mass concentration (PMC) and the

164

simulated profiles for styrene concentration and total reacted styrene. As we can see, after the

165

initial time required to complete nucleation, the PMC increases linearly. Likewise, since the

166

ozone and styrene reaction is slow, the reactant concentrations remain essentially constant and

167

the reacted styrene increases linearly, like PMC. This behaviour suggests that the initial reaction

168

is the rate limiting step in the formation of condensed matter under these experimental

169

conditions.

170

TE D

M AN U

SC

RI PT

161

Figure 1. Experimental temporal profiles for particles (PNbC and PMC) and simulated profiles of

172

styrene and reacted styrene. Initial concentrations: styrene, 10ppb; O3, 20ppb and SO2, 10ppb.

173

RH=0%. Styrene and reacted styrene concentrations are shown multiplied by the required

174

factors to display them in the same figure.

AC C

175

EP

171

176

A series of experiments was conducted to measure the reacted concentration of styrene

177

required for the generation of particles in the presence of SO2. Thus, Figure 2 shows the

178

concentrations of reacted styrene calculated at the nucleation events (assumed as the first scan

179

with PNbC above 100 cm-3) for experiments with SO2 initial concentrations in the range 5-100

180

ppbs and fixed initial concentrations of styrene (10 ppb) and ozone (20 ppb). The burst of

181

particles takes place at approximately the same reaction time and concentration of reacted

182

styrene. So, increasing SO2 concentration doesn’t result into further acceleration of the 6

ACCEPTED MANUSCRIPT nucleation event, which is found to occur once the reacted styrene amounts (5.6±1.7)x108

184

molecule cm-3 (errors are 2σ±20%). This value is three orders of magnitude lower than in the

185

absence of sulphur dioxide, showing that the SO2 reactivity must play a key role in the nucleation

186

steps. As shown above, lowering the concentrations of SO2 rises the nucleation threshold (as

187

inferred from the 5 ppb SO2 experiment) up to 1.6x1011 molecule cm-3 for SO2 zero

188

concentration.

RI PT

183

189 2,0E+09

SC

1,6E+09 1,4E+09 1,2E+09 1,0E+09 8,0E+08 6,0E+08 4,0E+08 2,0E+08 0,0E+00

0

20

M AN U

styrene reacted concentration for nucleation (molecule/cm3)

1,8E+09

40

60

80

100

190

TE D

[SO2] (ppb)

Figure 2. Concentration of reacted styrene at the instant of particles’ appearance for

192

experiments with different initial SO2 concentrations. For this series of experiments,

193

[styrene]=10ppb and [O3]=20ppb. The displayed errors are σ.

194

EP

191

3.2. Effect of styrene and ozone initial concentrations.

196

A series of experiments was carried out to assess the effect of styrene concentration on the

197

formation of secondary aerosols. Figure 3 shows the results for PNbC, PMC and the mean

198

particle size diameter for different [styrene]. The particles grew larger for higher concentrations

199

of styrene with average diameters up to 45 nm and the number concentration of new particles

200

rose up to 4x105 cm-3 for experiments carried out under 10 ppb and 20 ppb initial concentrations

201

of SO2 and ozone respectively.

202

To assess the conversion of gas-phase reactants into aerosols we can use the fractional aerosol

203

yield (Y) which may be defined (Odum et al., 1996) as the fraction of the reagent (styrene for

204

this work), that is converted to aerosol:

AC C

195

7

ACCEPTED MANUSCRIPT 205

𝑌𝑌 =

𝑃𝑃𝑃𝑃𝑃𝑃 ∆𝐶𝐶8 𝐻𝐻8

206

Where PMC is the aerosol mass concentration (µg m-3) produced for a given amount of reacted

207

styrene, ∆𝐶𝐶8 𝐻𝐻8 (in µg m-3). PNbC (#/cm3)

PMC (microg/m3)

Diameter (nm)

RI PT 50 45

4,0E+05

40

2,5E+05 2,0E+05 1,5E+05 1,0E+05 5,0E+04

209

0

0,1

0,2

0,3 0,4 [styrene] (ppm)

30 25 20 15 10 5

0,5

0,6

0

TE D

0,0E+00

35

SC

3,0E+05

M AN U

PNbC (#/cm3)

3,5E+05

Diameter (nm)

4,5E+05

PMC (µg/m3)

208

Figure 3. Particle mass concentration and mean particle size diameter versus styrene

211

concentration at reaction time=40 min for 10 ppb and 20 ppb initial concentrations of SO2 and

212

ozone respectively. The maximum particle number concentration (PNbC) measured for each

213

experiment are also plotted.

AC C

214

EP

210

215

According to Odum’s model, the growth of the aerosol’s mass may be described as a gas-particle

216

partitioning absorption mechanism where organic products may partition a portion of their

217

mass at concentrations below their saturation concentration.

218

Figure 4 displays the particle mass concentration growth curves for several experiments with

219

different initial styrene concentrations. After the first few minutes, once nucleation has

220

occurred, the system exhibits nearly constant yields for the rest of the experiment. Furthermore,

221

the yields from the different experiments are very similar.

8

ACCEPTED MANUSCRIPT 222

The linearity of the fractional yield suggests that the aerosol mass comes quantitatively from a

223

single product with very low volatility. This behaviour has been found also for the formation of

224

SOA from other aromatic compounds such as benzene, m-xylene or toluene (Ng et al., 2007).

225 12

RI PT

0.1ppm

10

0.2ppm 0.3ppm 0.4ppm

SC

0.5ppm

6 4 2 0

0

5

10

M AN U

PMC (µg m-3)

8

15

20

25

30

35

40

∆C8H8 (µg m-3)

TE D

226

Figure 4. Particle mass concentration growth curves versus the reacted styrene for different

228

initial styrene concentrations (see the legend). Fixed initial concentrations of sulphur dioxide

229

and ozone were used: [SO2]=10 ppb and [O3]=20 ppb.

EP

227

230

Figure S1 shows the yields obtained from the slopes of the linear fittings (straight part after the

232 233

first few minutes) of the plot PMC versus ∆𝐶𝐶8 𝐻𝐻8 (Figure 4) versus the styrene initial

234

The dependence of the aerosol fractional yield on the initial concentration of styrene is weak as

235

it is found in the plot. Nevertheless, for low styrene concentrations the yield values are slightly

236

higher.

AC C

231

concentration affording an average fractional aerosol yield of 0.39±0.10 (errors are 2σ±20%).

237

The effect of increasing ozone concentrations has been also studied in this work. A series of

238

experiments fixing the initial concentrations of styrene and SO2 was carried out with different

239

initial concentrations of ozone in the range 10 ppb to 0.6 ppm. The increase of ozone led to the

240

increase of PNbC, PMC and average particle size diameters, with a behaviour similar to the series 9

ACCEPTED MANUSCRIPT 241

changing styrene concentration described previously. The results are found in the

242

supplementary material, Figures S2-S4.

243 3.3. Effect of SO2 initial concentration.

245

As shown previously, the presence of SO2 enhances the production of secondary aerosols from

246

the ozonolysis reaction of styrene in air. To find the effect of SO2 concentration in the

247

mechanism, different experiments were performed fixing the initial concentrations of styrene

248

and ozone while the initial concentration of SO2 was changed within the range 5 to 100 ppb.

249

Even though the reacted styrene profile must be the same for all these experiments (since the

250

same initial concentrations of O3 and styrene are used), a significant increase in PNbC and PMC

251

was observed for increasing SO2 concentrations (mainly in the range 0-25 ppb), Figure 5. For

252

[SO2] above approximately 25 ppb, further increases of SO2 didn´t result in significant changes in

253

the aerosol production. The calculated fractional aerosol yield also increased up with [SO2] to

254

values around 2, as it is shown in Figure S5. From the definition of the fractional yield, values

255

above unity are consistent since styrene is not the only species involved in the formation of

256

secondary aerosol and so the mass of the particulate matter can be larger than the mass of the

257

reacted styrene.

M AN U

SC

RI PT

244

1,0E+05

40 35 30 25

AC C

PNbC (#/cm3)

1,2E+05

Diameter (nm)

EP

1,4E+05

PMC (microg/m3)

8,0E+04

20

6,0E+04

15

4,0E+04

10

2,0E+04

5

0,0E+00

0

0

0,02

0,04

0,06 [SO2] (ppm)

259

10

0,08

0,1

0,12

10 x PMC (µg/m3)

PNbC (#/cm3)

Diameter (nm)

TE D

258

ACCEPTED MANUSCRIPT 260

Figure 5: Effect of SO2 concentration on the production of secondary aerosols. The maximum

261

particle number concentration (PNbC) measured for each experiment are shown in the main

262

axis. PMC and mean particle size diameter versus SO2 concentration are shown at a fixed

263

reaction time, 50 min (secondary axis), for 10 ppb and 20 ppb initial concentrations of styrene

264

and ozone respectively.

RI PT

265 SO2 doesn’t react with the initial reactants, cyclohexane, styrene or ozone within the time scale

267

of the experiments nor with the aldehydes HCHO, C6H5CHO coming from the ozonolysis reaction.

268

On the other hand, it is known that Criegee intermediates do undergo fast reactions with SO2

269

(Taatjes et al., 2013). Nevertheless, since the styrene-ozone reaction is slow, under the

270

experimental conditions, Figure 5, the production rate of the Criegee intermediates is very low

271

and a slight consumption of SO2 was expected. Consistently, the experimental SO2 temporal

272

profiles during most of the experiments were nearly constant. Thus, in some experiments, to

273

magnify the SO2 consumption and make it measurable, it was necessary to increase the styrene

274

and ozone concentrations (up to 100-200 ppb) while keeping [SO2] in lower levels (around 20

275

ppb). For this series of experiments, the production of sCI was relatively high and the SO2

276

concentration was observed to decrease with time (experimental SO2 decays, Figure 6), which

277

confirms that SO2 reacts with the products from the ozonolysis.

278

Previous studies have shown that the first step of the reaction of sCI (R1R2COO) with SO2 is the

279

formation of a cyclic secondary ozonide which can undergo isomerisation (only possible if R2 is

280

an H-atom), leading to an organic acid (R1COOH) and SO2, or decomposition to a carbonyl

281

compound (R1R2CO) and SO3 (Jiang et al.,2010; Kurten et al., 2011). The organic acids will

282

generally be less volatile than their alkene precursors and SO3 readily produces H2SO4 in the

283

presence of water vapour. Such products are expected to be involved in the formation and

284

growth of new particles.

285

In the isomerisation channel, SO2 would behave as a catalyst in the production of acids (Díaz de

286

Mera et al., 2017). On the other hand, in this work we have found, Figure 6, that there is a net

287

consumption of SO2 and that the amount of reacted SO2 increases with the reaction time

288

coordinate. This fact suggests that for the case of styrene ozonolysis the reaction of the sCI with

289

SO2 proceeds at least in part through the oxidation channel leading to SO3.

290

Some of the previous experiments were carried out under the presence of water under different

291

RH. For example, as shown in Figure 6, when water vapour is present (RH=5 % in the plotted

292

experiment) the decrease in SO2 was significantly lower than under drier conditions. These

AC C

EP

TE D

M AN U

SC

266

11

ACCEPTED MANUSCRIPT 293

results suggest that H2O and SO2 constitute competitive sinks for the stabilised Criegee

294

intermediates.

295

For these experiments we assume a simple mechanism as follows:

296

Styrene + O3  HCHO + C6H5CHO + CI + sCI

297

sCI + SO2  intermediates (SO3)  secondary aerosols (2)

298

sCI+ H2O  P3 (3)

299

Where sCI includes both ·CH2OO· and C6H5C·HOO· (Tuazon et al., 1993), the latter being possible

300

as conformer syn or anti. (Vereecken and Francisco, 2012).

301

For the case of ·CH2OO·, the reaction with SO2 has been studied previously and the available

302

experimental data show that it leads to the formation of SO3 with a unity yield (Berndt et al.,

303

2014a; Berndt et al., 2014b). In this study, for the experiments carried out in the drier conditions

304

(RH close to zero), the water concentration ratio is in the order of 2ppm (5x1013 molecule cm-3).

305

The H2O-SO3 gas-phase reaction is very fast, k= 3.90x10-41exp(6830.6/T)[H2O]2 (Jayne et al.,

306

1997) and the gas-phase water concentration is high enough to react with SO3, leading to H2SO4

307

(Díaz-de-Mera et al., 2017), which is the species assumed to generate new particles in the

308

system.

309

For the other sCI formed from the styrene ozonolysys, C6H5C·HOO·, the oxidation channel

310

leading to SO3, and then H2SO4, is also possible and can contribute to the nucleation step. Recent

311

works have shown that the oxidation channel also generates H2SO4 quantitatively in the

312

ozonolysis of other alkenes such as trans-2-butene and 2,3-dimethyl-2-butene when it happens

313

in the presence of SO2 (Bernd et al., 2014a). Furthermore, the crossed reaction between the

314

biradical C6H5C·HOO· and formaldehyde has been suggested to produce 3,5-diphenyl-1,2,4-

315

trioxolane (Na et al., 2006) and the secondary ozonide may also undergo partial conversion into

316

a hydroxyl-substituted ester (Tuazon et al., 1993). These low volatile products may contribute

317

to the growing of particles.

318

In this work we have studied the direct yield for the reaction between sCI and SO2. As shown

319

previously, the correlation between the measured MPC and the reacted styrene points out the

320

ozonolysis as the rate limiting step for SOA production. Thus, the sCI/SO2 reaction is expected to

321

be much faster than the styrene/ozone reaction and, so, the amount of consumed SO2 should

322

correlate with the total reacted styrene. In this sense, Figure 6 shows the results for several

323

experiments with different initial concentrations of styrene and ozone under dry conditions. The

324

simulated profiles for total reacted styrene are nearly straight lines since its precursors, styrene

AC C

EP

TE D

M AN U

SC

RI PT

(1)

12

ACCEPTED MANUSCRIPT and ozone remain essentially constant during the experiment. The total reacted SO2, measured

326

experimentally, also shows linear temporal profiles except for the experiments with the higher

327

concentrations of styrene and ozone. From the ratio of the slopes of the linear fits (reacted SO2

328

and reacted styrene) we derived an average yield for reaction (2) with respect to the initial

329

reaction (1), Table 1, η= 0.51±0.13 (errors are 2σ±20%).

330

For the experiments with higher concentrations, for example 200 ppb of both styrene and

331

ozone, the SO2 concentration dropped from 20 to 10 ppb after 25 minutes lowering the rate of

332

the sCI/SO2 reaction and driving to the curvature of the temporal profile of the reacted SO2. For

333

such experiments the yield for reaction (2) was obtained from the initial data (20 min in the

334

shown experiment) where the SO2 data can still be fitted to a straight line.

335

The measurement of the yield for reaction 2 could not be carried out within a wide range of

336

[SO2] initial concentrations since the ozonolysis reaction is slow and increasing SO2 leads to very

337

small changes in the SO2 signal.

338

The results show that sCI react quantitatively through reaction (2) in the presence of SO2 with a

339

constant yield regardless the initial concentrations of styrene and ozone. However, as sCI can

340

react through different competitive sinks, it is expected that the yield of reaction (2) depends

341

on the concentration of SO2, water and other possibly involved species.

M AN U

TE D

342

SC

RI PT

325

Table 1. Yield of reaction (2) (consumption of SO2) after the ozonolysis of styrene (reaction 1).

344

The yields are obtained from the ratio of the reacted SO2, and the reacted styrene. Errors are

345

2σ. [Styrene]/ppb

EP

343

[O3] /ppb

[SO2] /ppb

RH

η

100

20

0%

0.47± 0.05

100

200

20

0%

0.55±0.06

100

100

20

0%

0.54±0.05

200

200

20

0%

0.48±0.06

AC C

200

346

13

ACCEPTED MANUSCRIPT Reacted styrene, [Sty]=[O3]=200ppb Reacted styrene, [Sty]=200 ppb; [O3]=100ppb Reacted styrene, [Sty]=[O3]=100ppb SO2 decay. RH=5%

5,0E+11

Reacted SO2, [Sty]=[O3]=200ppb reacted SO2, [Sty]=200 ppb; [O3]=100ppb Reacted SO2, [Sty]=[O3]=100ppb SO2 decay. RH=0% 20

4,5E+11

Concentration (molecule/cm3)

4,0E+11

2,5E+11 2,0E+11

1,0E+11

0

5

10

M AN U

5,0E+10 15

20

25

30

35

5

40

45

Time (min)

347

Figure 6. SO2 experimental decays for 100 ppb of styrene and ozone (initial concentrations)

349

under dry conditions and with RH=5 %. Simulated profiles for total reacted styrene under

350

different initial concentrations of styrene and ozone (see the legend) and the corresponding

351

experimental profiles for total reacted SO2. The initial SO2 concentration was 20 ppb for all the

352

experiments displayed.

3.4. Effect of water vapour.

EP

354

TE D

348

353

10

SC

1,5E+11

[SO2] (ppb)

3,0E+11

0,0E+00

15

RI PT

3,5E+11

A series of experiments with concentrations or reactants around 10 ppb was carried out to study

356

the effect of RH on the production of SOA. For these experiments the PMC values were very

357

small, always below 0.5 µg m-3 and the concentrations had to be increased to get more accurate

358

mass data. Figures 7 and S6-S8 show the results for PNC, PMC and average particles’ size

359

diameters for experiments with 20 ppb of styrene and SO2, 40 ppb of ozone and RH in the range

360

0-35 %.

361

As it is found in Figure 7, the addition of water, induces a significant drop (mainly in the RH range

362

below 10 %) on the PNbC and inhibits their growing leading to lower production of SOA. These

363

results are consistent with the suggested mechanism where SO2 and H2O compete to scavenge

364

the Criegee intermediates (reactions (2) and (3)). The reaction with water vapour is expected to

365

proceed through the addition of the water oxygen atom to the carbon atom in the carbonyl

AC C

355

14

50

0

ACCEPTED MANUSCRIPT 366

oxide and the simultaneous transfer of one hydrogen atom from the water molecule to the

367

terminal O atom in the CI (Anglada et al., 2011), leading to α-hydroxi-hydroperoxides that can

368

subsequently decompose or react with other atmospheric species. 3,0

PMC

16

DeltaPMC/PMC

RI PT

2,5

1,5

8

SC

PMC (µg/m3)

2,0

0,5 0,0

0

5

10

M AN U

1,0

15

20

25

30

∆PMC/PMC

12

4

0

35

40

% RH

369

Figure 7. Water effect on the particle mass concentration. For this series of experiments, the

371

concentration ratios were 20 ppb, 20 ppb and 40 ppb for styrene, SO2 and ozone, respectively.

372

The plotted PMC data plotted for each experiment are those obtained at 40’ reaction time.

TE D

370

EP

373

The rates of the sCI reactions towards SO2 and H2O depend on their concentrations and their

375

kinetic rate constants, kSO2 and kH2O. Since both [H2O] and [SO2] are known (and essentially

376

constant during the experiments), experimental data of the products of these reaction can

377

provide the ratio of theses competitive reactions (kH2O/kSO2). Thus, considering the dry

378

experiments (0% RH) as reference, we can assume that for the experiments carried out in the

379

presence of water the decrease in mass concentration, ∆PMC = PMCRH=0 – PMC, is due exclusively

380

to the reaction of the sCI with water molecules. Thus, the ratio (kH2O [H2O])/(kSO2[SO2]) can be

381

obtained as the ratio ∆PMC /PMC (Díaz-de-Mera et al., 2017), assuming that the reaction of sCI

382

with SO2 is the sole source of SOA. Then, the linear fit of the plot (∆PMC /PMC) versus [H2O],

383

Figure 7, provides the (kH2O/(kSO2[SO2])) ratio from the slope, and then the ratio between the rate

384

constants, kH2O/kSO2= (2.8±0.7)×10-5 (the SO2 concentration is known, 20 ppb=4.9x1011 molecule

385

cm-3, errors are 2σ±20%).

AC C

374

15

ACCEPTED MANUSCRIPT This result is similar to other reported values for different CI (Berndt et al., 2014a, Díaz-de-Mera

387

et al., 2017) and may be used to assess the effects of SO2 and water vapour in the formation of

388

SOA under different atmospheric conditions.

389

4. Conclusions.

390

Previous studies have shown the outstanding potential of SOA formation due to the emissions

391

of aromatic compounds to the atmosphere. Among them, styrene is one of the aromatics with

392

the highest contribution to SOA. Although the formation of new particles from the ozonolysis

393

of styrene has been reported from laboratory experiments, this reaction can’t justify such high

394

capacity to form new particles since it would require high concentrations of reactants compared

395

to those of real polluted atmospheres. Generally, atmospheric models based on laboratory

396

chamber studies predict far less SOA than is observed in field studies (Kroll and Seinfeld, 2008).

397

This fact suggest that other yet unknown pathways may lead to the formation of new particles

398

more effectively. In this sense, SO2 may play a key role since, as found in this work, the ozonolysis

399

of styrene in the presence of low concentrations of SO2 readily produces new particles. The

400

nucleation events occur at concentration around (5.6±1.7)x108 molecule cm-3, that is, 3 orders

401

of magnitude below that of the reaction in the absence of SO2.

402

The particle mass concentration temporal profiles correlate with the reacted styrene showing

403

that the ozonolysis of styrene is the rate limiting step in the formation of new particles. The

404

fractional yields for different initial concentrations of reactants have been obtained based on

405

the amount of reacted styrene. They are independent on the concentrations of styrene or ozone

406

but increase from nearly zero in the absence of SO2 to approximately 2 when SO2 is in excess.

407

The experimental temporal decays of SO2 show that it reacts with the products of the ozonolysis

408

of styrene. From the simulated profiles of reacted styrene, under low water concentration

409

conditions (below 2 ppm) the yield for the consumption of SO2 was η= 0.51±0.13. Thus, the

410

formation of SO3 from the reaction of styrene with the Criegee intermediates is expected to be

411

responsible of the nucleation events through the formation of H2SO4. On the other hand, the

412

presence of water in the reaction media competes with SO2 and inhibits the formation of SOA.

413

So the potential formation of SOA under atmospheric conditions depends on the concentration

414

of SO2 and relative humidity.

415

As shown in a previous work (le Person et al. 2006) OH-based and ozone-based lifetimes for

416

styrene are 2.4h and 25h, respectively and so both reactions contribute as efficient gas-phase

417

sinks. Although the initial OH reaction is faster than ozonolysis, it is not clear whether the OH

418

reaction products contribute efficiently to the formation of aerosols, as it has been found for

AC C

EP

TE D

M AN U

SC

RI PT

386

16

ACCEPTED MANUSCRIPT ozonolysis in the presence of SO2. In this sense, the study of particles generation from OH

420

reactions is also required to fully understand the contribution of styrene to the formation of

421

secondary organic aerosols in urban areas.

422

Acknowledgments.

423

This work was supported by the Spanish Ministerio de Economía y Competitividad (project

424

CGL2014-57087-R, FEDER co-funding) and by the University of Castilla La Mancha (projects

425

GI20152950 and GI20163433).

RI PT

419

426 References.

428

Anglada, J.M., González, J., Torrent-Sucarrat, M, 2011. Effects of the substituents on the

429

reactivity of carbonyl oxides. A theoretical study on the reaction of substituted carbonyl oxides

430

with water. Physical Chemistry Chemical Physics, 13, 13034–13045.

431

Atkinson, R. and Aschmann, S.M., 1988. Kinetics of the reactions of acenaphthene and

432

acenaphthylene and structurally-related aromatic compounds with OH and NO3 at 296±2K.

433

International Journal of Chemical Kinetics, 20, 513-539.

434

ATSDR, 2017. https://www.atsdr.cdc.gov/toxprofiles/tp53-c3.pdf accesed June 2017.

435

Bates, J. T., Weber, R. J., Abrams, J., Verma, V., Fang, T., Klein, M., Strickland, M. J., Sarnat, S. E.,

436

Chang, H. H., Mulholland, J. A., Tolbert, P. E., Russell, A. G., 2015. Reactive oxygen species

437

generation linked to sources of atmospheric particulate matter and cardiorespiratory effects.

438

Environmental Science & Technology, 49(22), 13605-13612. DOI: 10.1021/acs.est.5b02967.

439

Berndt, T., Jokinen, T., Sipilä, M., Mauldin, R. L., Herrmann, H., Stratmann, F., Junninen, H.,

440

Kulmala, M., 2014a. H2SO4 formation from the gas-phase reaction of stabilized Criegee

441

Intermediates with SO2: Influence of water vapour content and temperature. Atmospheric

442

Environment, 89, 603-612. DOI: org/10.1016/j.atmosenv.2014.02.062.

443

Berndt, T., Voigtländer, J., Stratmann, F., Junninen, H., Mauldin, R. L., Sipilä, M., Kulmala, M.,

444

Herrmann, H., 2014b. Competing atmospheric reactions of CH2OO with SO2 and water vapour.

445

Physycal Chemistry Chemical Physics, 16, 19130-19136. DOI: 10.1039/C4CP02345E.

446

Cheung H.C., Chou C.C.K., Huang, W.R., Tsai, C.Y., 2013. Characterization of ultrafine particle

447

number concentration and new particle formation in an urban environment of Taipei, Taiwan.

448

Atmospheric Chemistry and Physics, 13(17), 8935–8946. DOI: 10.5194/acp-13-8935-2013.

AC C

EP

TE D

M AN U

SC

427

17

ACCEPTED MANUSCRIPT 449

Diaz-de-Mera, Y., Aranda, A., Bracco, L., Rodriguez, D., Rodriguez, A., 2017. Formation of

450

secondary organic aerosols from the ozonolysis of dihydrofurans. Atmospheric Chemistry and

451

Physics, 17(3), 2347-2357. DOI: org/10.5194/acp-17-2347-2017.

452

Duke, J.A., 1985, CRC handbook of medical herbs. Boca Raton, FL, CRC Press, p. 323.

453

EPA

454

https://www.epa.gov/haps/initial-list-hazardous-air-pollutants-modifications

455

2017.

456

Hu, N., Tan, J., Wang, X., Zhang, X., Yu, P., 2017. Volatile organic compound emissions from an

457

engine fueled with an ethanol-biodiesel-diesel blend. Journal of the Energy Institute, 90, 101-

458

109. DOI: org/10.1016/j.joei.2015.10.003.

459

Jayne, J. T., Pöschl, U., Chen, Y., Dai, D., Molina, L.T., Worsnop, D.R., Kolb, C.E., Molina, M. J.,

460

1997. Pressure and temperature dependence of the gas-phase reaction of SO3 with H2O and the

461

heterogeneous reaction of SO3 with H2O/H2SO4 surfaces. Journal of Physical Chemistry A, 101,

462

10000-10011. DOI: 10.1021/jp972549z.

463

Jiang, L., Xu, Y., Ding, A., 2010. Reaction of stabilized Criegee intermediates from ozonolysis of

464

limonene with sulfur dioxide: ab initio and DFT study. J. Phys. Chem. A, 114, 12452-12461. DOI:

465

10.1021/jp107783z.

466

Kabir, E. and Kim, K. H., 2010. An on-line analysis of 7 odorous volatile organic compounds in the

467

ambient air surrounding a large industrial complex. Atmospheric Environment, 44(29), 3492-

468

3502. DOI: org/10.1016/j.atmosenv.2010.06.021.

469

Knighton, W. B., Herndon, S.C., Wood, E.C., Fortner, E.C.,Onasch, T. B., Wormhoudt, J., Kolb, C.

470

E., Lee, B.H., Zavala, M., Molina, L., Jones, M., 2012. Detecting fugitive emissions of 1,3-

471

butadiene and styrene from a petrochemical facility: An application of a mobile laboratory and

472

a modified proton transfer reaction mass spectrometer. Industrial & Engineering Chemistry

473

Research, 51 (39), 12706–12711. DOI: 10.1021/ie202794j.

474

Kos, G., Adechina, N., Kanthasamy, V., Ariya, P. A.,2014. Volatile organic compounds in Arctic

475

snow: Concentrations and implications for atmospheric processes. RSC, Environmental Science:

476

Processes & Impacts 16 (11), 2592-2603. DOI: 10.1039/C4EM00410H.

477

Kroll, J. H., Seinfeld, J. H., 2008. Chemistry of secondary organic aerosol: Formation and

478

evolution of low-volatility organics in the atmosphere. Atmospheric Environment, 42(16), 3593-

479

3624.

hazardous

air

pollutants

with

modifications. accesed

June

RI PT

of

SC

list

M AN U

Initial

AC C

EP

TE D

2017,

18

ACCEPTED MANUSCRIPT 480

Kurtén, T., Lane, J. R., Jørgensen, S., Kjaergaard, H. G.,2011. A Computational Study of the

481

Oxidation of SO2 to SO3 by Gas-Phase Organic Oxidants. J. Phys. Chem. A, 115 (31), 8669–8681.

482

DOI: 10.1021/jp203907d.

483

Le Person, A., Eyglunent, G., Daele, V., Mellouki, A., Mu, Y. 2008. The near UV absorption cross-

484

sections and the rate coefficients for the ozonolysis of a series of styrene-like compounds.

485

Journal

486

org/10.1016/j.jphotochem.2007.09.006.

487

Li, L., Li, H., Zhang, X., Wang, L., Xu, L., Wang, X., ... & Cao, G. 2014. Pollution characteristics and

488

health risk assessment of benzene homologues in ambient air in the northeastern urban area of

489

Beijing,

490

https://doi.org/10.1016/S1001-0742(13)60400-3.

491

Li, L., Ge, Y., Wang, M., Peng, Z., Song, Y., Zhang, L., Yuan, W., 2015. Exhaust and evaporative

492

emissions from motorcycles fueled with ethanol gasoline blends. Science of the Total

493

Environment, 502, 627-631. DOI: org/10.1016/j.scitotenv.2014.09.068.

494

Li, L., Tang, P., Nakao, S., Kacarab, M., Cocker III, D. R., 2016. Novel approach for evaluating

495

secondary organic aerosol from aromatic hydrocarbons: Unified method for predicting aerosol

496

composition and formation. Environmental Science & Technology, 50(12), 6249-6256. DOI:

497

10.1021/acs.est.5b05778.

498

Liu, T., Wang, X., Hu, Q., Deng, W., Zhang, Y., Ding, X., Fu, X., Bernard, F., Zhang, Z., Lü, S., He, Q.,

499

Bi, X., Chen, J., Sun, Y., Yu, J., Peng, P., Sheng, G., and Fu, J. 2016. Formation of secondary aerosols

500

from gasoline vehicle exhaust when mixing with SO2, Atmos. Chem. Phys., 16, 675-689,

501

doi:10.5194/acp-16-675-2016.

502

Lopes, M., Serrano, L., Ribeiro, I., Cascão, P., Pires, N., Rafael, S., Tarelho,L., Monteiro, A., Nunes,

503

T., Evtyugina, M., Nielsen, O.J., Gameiro da Silva, M., Miranda, A.I., Borrego, C., 2014. Emissions

504

characterization from EURO 5 diesel/biodiesel passenger car operating under the new European

505

driving

506

org/10.1016/j.atmosenv.2013.11.071.

507

Ma, Y., Porter, R.A., Chappell, D., Andrew, A.T., Russell, T., Marston, G., 2009. Mechanisms for

508

the formation of organic acids in the gas-phase ozonolysis of 3-carene. Physical Chemistry

509

Chemical Physics, 11, 4184–4197.

Environmental

RI PT

of

Sciences,

26(1),

214-223.

SC

Journal

DOI:

AC C

EP

TE D

China.

and Photobiology A: Chemistry, 195(1), 54-63.

M AN U

of Photochemistry

cycle.

Atmospheric

Environment,

19

84,

339-348.

DOI:

ACCEPTED MANUSCRIPT Mauldin, R. L., Berndt, T., Sipilä, M., Paasonen, P., Petäjä, T., Kim, S., Kurtén, T., Stratmann, F.,

511

Kerminen, V. M., Kulmala, M., 2012. A new atmospherically relevant oxidant of sulphur dioxide.

512

Nature, 488, 193–196. DOI: 10.1038/nature11278.

513

Mejía, J. F., Morawska, L., and Mengersen, K., 2008. Spatial variation in particle number size

514

distributions in a large metropolitan area, Atmos. Chem. Phys., 8, 1127-1138. Doi:10.5194/acp-

515

8-1127-2008.

516

Na, K., Song, C., Cocker, D. R., 2006. Formation of secondary organic aerosol from the reaction

517

of styrene with ozone in the presence and absence of ammonia and water. Atmospheric

518

Environment, 40(10), 1889-1900. DOI: org/10.1016/j.atmosenv.2005.10.063.

519

Nel, A. 2005. Air pollution-related illness: effects of particles. Science, 308(5723), 804-806. DOI:

520

10.1126/science.1108752.

521

Newland, M. J., Rickard, A. R., Alam, M. S., Vereecken, L. Muñoz, A., Ródenas, M., Bloss, W.J.,

522

2015. Kinetics of stabilised Criegee intermediates derived from alkene ozonolysis: reactions with

523

SO2, H2O and decomposition under boundary layer conditions. Physical Chemistry Chemical

524

Physics, 17, 4076-4088. DOI: 10.1039/C4CP04186K.

525

Ng, N. L., Kroll, J. H.; Chan, A. W. H., Chhabra, P. S., Flagan, R. C., Seinfeld, J. H., 2007. Secondary

526

organic aerosol formation from mxylene, toluene, and benzene. Atmospheric Chemistry and

527

Physics. 2007, 7 (14), 3909− 3922. DOI: 10.5194/acp-7-3909-2007.

528

Odum, J.R., Hoffman, T., Bowman, F., Collins, D., Flagan, R.C., Sienfeld, J.H., 1996. Gas/particle

529

partitioning and secondary aerosol yields. Enviromental Science & Technology, 30, 2580-2585.

530

Sarigiannis, D.A., Karakitsios, S.P., Gotti, A., Liakos, I.L., Katsoyiannis, A., 2011. Exposure to major

531

volatile organic compounds and carbonyls in European indoor environments and associated

532

health risk. Environment International, 37, 743–765. DOI: org/10.1016/j.envint.2011.01.005.

533

Shimada, K., Kimura, E., Yasui, Y., Tanaka, H., Matsushita, S., Hagihara, H., Nagakura, M,

534

Kawahisa, M., 1992. Styrene formation by the decomposition by Pichia carsonii of trans-

535

cinnamic acid added to a ground fish product. Applied Environmental Microbiology, 58, 1577–

536

1582.

537

Sun, J., Wu, F., Hu, B., Tang, G., Zhang, J., Wang, Y., 2016. VOC characteristics emissions and

538

contributions to SOA formation during hazy episodes. Atmospheric Environment, 141, 560-570.

539

DOI: org/10.1016/j.atmosenv.2016.06.060.

540

Taatjes C.A., Welz, O., Eskola, A.J., Savee, J.D., Scheer A.M., Shallcross, D.E., Rotavera, B., Lee,

541

E.P., Dyke, J.M., Mok, D.K., Osborn, D.L., Percival, C.J., 2013. Direct measurements of conformer-

AC C

EP

TE D

M AN U

SC

RI PT

510

20

ACCEPTED MANUSCRIPT dependent reactivity of the Criegee intermediate CH3CHOO. Science. 12, 340(6129), 177-180.

543

DOI: 10.1126/science.1234689.

544

Tuazon, E. C., Arey, J., Atkinson, R., Aschmann, S. M., 1993. Gas-phase reactions of 2-

545

vinylpyridine and styrene with hydroxyl and NO3 radicals and ozone. Environmental Science &

546

Technology, 27(9), 1832-1841. DOI: 10.1021/es00046a011.

547

Vereecken, L. and Francisco, J. S., 2012. Theoretical studies of atmospheric reaction mechanisms

548

in the troposphere. Chemical Society Reviews, 41(19), 6259-6293. DOI: 10.1039/C2CS35070J.

549

Wang, H., Ji, Y., Gao, Y., Li, G., An, T. 2015. Theoretical model on the formation possibility of

550

secondary organic aerosol from OH initialed oxidation reaction of styrene in the presence of

551

O2/NO. Atmospheric Environment, 101, 1-9. DOI: org/10.1016/j.atmosenv.2014.10.042.

552

Welz, O., Savee, J. D., Osborn, D. L., Vasu, S. S., Percival, C. J, Shallcross, D., Taatjes, C., 2012.

553

Reaction of CH2I with O2 forms Criegee Intermediate: Direct Measurements of CH2OO Kinetics.

554

Science, 335, 204-207. DOI: 10.1126/science.1213229.

555

Xu, J., Szyszkowicz, M., Jovic, B., Cakmak, S., Austin, C. C., Zhu, J., 2016. Estimation of indoor and

556

outdoor ratios of selected volatile organic compounds in Canada. Atmospheric Environment,

557

141, 523-531. DOI: org/10.1016/j.atmosenv.2016.07.031.

558

Zhang, Z., Wang, H., Chen, D., Li, Q., Thai, P., Gong, D., Li, Y., Zhang, C., Gu, Y., Zhou, L., Morawska,

559

L., Wang, B., 2017. Emission characteristics of volatile organic compounds and their secondary

560

organic aerosol formation potentials from a petroleum refinery in Pearl River Delta, China.

561

Science of the Total Environment, 584-585, 1162-1174. DOI: 10.1016/j.scitotenv.2017.01.179.

AC C

EP

TE D

M AN U

SC

RI PT

542

21