Secondary Organic Aerosol (SOA) from Nitrate Radical Oxidation of

Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Article

Secondary Organic Aerosol (SOA) from Nitrate Radical Oxidation of Monoterpenes: Effects of Temperature, Dilution, and Humidity on Aerosol Formation, Mixing, and Evaporation Christopher M. Boyd, Theodora Nah, Lu Xu, Thomas Berkemeier, and Nga Lee Ng Environ. Sci. Technol., Just Accepted Manuscript • Publication Date (Web): 19 Jun 2017 Downloaded from http://pubs.acs.org on June 20, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Environmental Science & Technology

Secondary Organic Aerosol (SOA) from Nitrate Radical Oxidation of Monoterpenes: Effects of Temperature, Dilution, and Humidity on Aerosol Formation, Mixing, and Evaporation Christopher M. Boyd1; Theodora Nah1; Lu Xu1; Thomas Berkemeier1; Nga Lee Ng1,2* 1

School of Chemical and Biomolecular Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA 2

School of Earth and Atmospheric Sciences, Georgia Institute of Technology, Atlanta, GA 30332, USA

*

Correspondence to: Nga Lee Ng ([email protected])

Keywords: Aerosol Yield, Biogenic Secondary Organic Aerosol, Nitrate Radical, Volatility

1 ACS Paragon Plus Environment


Page 2 of 40

19

Abstract

20

Nitrate radical (NO3) oxidation of biogenic volatile organic compounds (BVOC) is important

21

for nighttime secondary organic aerosol (SOA) formation. SOA produced at night may evaporate

22

the following morning due to increasing temperatures or dilution of semivolatile compounds. We

23

isothermally dilute the oxidation products from the limonene+NO3 reaction at 25 °C and observe

24

negligible evaporation of organic aerosol via dilution. The SOA yields from limonene+NO3 are

25

approximately constant (~174%) at 25 °C and range from 81-148% at 40 °C. Based on the

26

difference in yields between the two temperatures, we calculated an effective enthalpy of

27

vaporization of 117-237 kJ mol-1. The aerosol yields at 40 °C can be as much as 50% lower

28

compared to 25 °C. However, when aerosol formed at 25 °C is heated to 40 °C, only about 20%

29

of the aerosol evaporates, which could indicate a resistance to aerosol evaporation. To better

30

understand this, we probe the possibility that SOA from limonene+NO3 and β-pinene+NO3

31

reactions is highly viscous. We demonstrate that particle morphology and evaporation is

32

dependent on whether SOA from limonene is formed before or during the formation of SOA

33

from β-pinene. This difference in particle morphology is present even at high relative humidity

34

(~70%).

35


Page 3 of 40


36

1) Introduction

37

The nitrate radical (NO3) oxidation of biogenic volatile organic compounds (BVOC) is an

38

important contributor to the secondary organic aerosol (SOA) burden due to its high reactivity1

39

and large SOA mass yields.2-13 The reaction of BVOC with NO3 radical, which is formed by the

40

reaction of NO2 with O3, provides a direct link between observations that organic aerosol is well

41

correlated with anthropogenic pollutants14 but contain carbon that is mostly biogenic in origin.15,

42

16

Modeling studies estimate that 5-21% of global SOA is produced by NO3 radical chemistry.17,

43

18

Monoterpene+NO3 chemistry is especially important at night because the NO3 radical is the

44

dominant nocturnal oxidant and monoterpenes are typically emitted during the day and at

45

night.19, 20 Limonene and β-pinene are two of the most important monoterpenes due to their high

46

global abundance18 and aerosol mass yields. 2-4, 7, 8, 11, 12 In the southeastern United States, recent

47

studies show that NO3 radical oxidation of monoterpenes can produce a substantial fraction of

48

organic aerosol.12, 21-24 Based on coordinated laboratory and field studies, Xu et al.21 shows that

49

this reaction contributes 50% of total nighttime OA production in the southeastern US, a large

50

fraction of which is from β-pinene+NO3 chemistry.

51 52

Aerosol lifetimes in the atmosphere typically range between 1 to 2 weeks, spanning multiple

53

day/night cycles. Thus, it is important to study the changes in the physicochemical properties of

54

SOA formed by BVOC+NO3 reactions during the night-to-day transition and atmospheric

55

transport where environmental parameters (e.g. T, RH, dilution, etc) can change. Since

56

BVOC+NO3 reactions produce large amounts of SOA and organic nitrates, these changes could

57

have impacts on the NOx cycle and total aerosol loading. Nah et al.25 shows that SOA and

58

organic nitrates formed from the nighttime monoterpene+NO3 reaction can serve as a NOx sink

59

or reservoir upon photochemical oxidation, depending on the precursor hydrocarbons. During the 3 ACS Paragon Plus Environment


Page 4 of 40

60

day, in addition to photochemical oxidation, surface temperature and boundary layer height

61

increase, which lead to dilution of gas- and particle-phase organics. Aerosol evaporation is

62

expected to occur as a result of dilution26 and increasing temperatures, due to the semi-volatile

63

nature of some components.27 Understanding evaporation of nighttime SOA upon daybreak is

64

important to accurately predict the dynamics of aerosol loading and properties over its lifetime.

65 66

Aerosol evaporation can be predicted by either the Odum two-product model28 or the volatility

67

basis set.

68

temperature.31-35 The temperature-dependent partitioning coefficient (K) at one temperature (T1)

69

can be approximated using the partitioning coefficient at a reference temperature (T2) and the

70

Clausius-Clapeyron equation27 if the effective enthalpy of vaporization (∆Hv) for the SOA is

71

known:

29, 30

Previous studies have shown that bulk aerosol volatility is highly dependent on

=

∆ 1 1 exp −

(1)

72

Although previous studies have shown that the effective enthalpy of vaporization increases with

73

decreasing saturation concentration (C*),32, 36-38 a single value is often assumed for a single SOA

74

system27,

75

Furthermore, many chamber and thermodenuder studies31, 34 that evaluate the effective enthalpy

76

of vaporization typically use mass loadings that are higher than what is considered to be

77

atmospherically relevant. As a result, these studies may underestimate the enthalpies of

78

vaporization and the dependence

79

underpredicted when these values are used in atmospheric models. To accurately predict aerosol

80

loadings, both the thermodynamics and kinetics of aerosol evaporation must be understood.

81

Typically, aerosol is assumed to be liquid-like, with no kinetic resistance and evaporation

31, 33

or used to estimate aerosol loadings in global models.

39 and references therein

of ambient aerosol mass on temperature may be


Page 5 of 40


82

controlled solely by thermodynamics (i.e., governed by the Clausius-Clapeyron equation).

83

However, a number of recent studies showed that SOA formed in some reactions exists as a

84

viscous semi-solid,40-43 which may significantly slow the evaporation of organic aerosol.

85 86

In this study, we investigate the effect of temperature on SOA formed from NO3 radical

87

oxidation of limonene and evaluate the effective enthalpy of vaporization for SOA formed in the

88

limonene+NO3 system. Furthermore, we study the effect of isothermal dilution and increasing

89

temperatures on the evaporation of organic aerosol. The evaporation of aerosol mixtures

90

produced from limonene+NO3 and β-pinene+NO3 is also investigated as a function of relative

91

humidity and aerosol morphology. This study provides fundamental data for understanding the

92

changes in physical properties for organic aerosol formed by BVOC+NO3 reactions and

93

highlights the importance of exploring the effects of temperature under a wide range of

94

conditions and aerosol systems.

95

2) Experimental

96

All experiments are performed as batch-reactions in the Georgia Tech Environmental Chamber

97

Facility, which consists of two 12 m3 Teflon chambers.12 Experimental conditions are shown in

98

Table 1. Reactions are conducted at either 25 oC or at 40 oC under dry conditions (RH < 3%).

99

The limonene+NO3 reaction is also performed under humid conditions (RH ~ 50%)

100

(Experiments 13 and 14) to examine the effects of humidity on aerosol yield. Prior to each

101

experiment, the chambers are flushed with zero air for at least 24 hours.

102 103

All experiments are performed by first injecting formalin solution (Sigma-Aldrich, 37% HCHO)

104

into a glass bulb, where it evaporates as clean air passes over the solution and introduces 5 ACS Paragon Plus Environment


Page 6 of 40

105

formaldehyde into the chamber. The concentration of formaldehyde ranges from 1.2 ppm to 13.0

106

ppm (with targeted formaldehyde:VOC molar ratio of 500:1). For seeded experiments,

107

ammonium sulfate seed aerosol is injected into the chamber via atomization of an 8 mM

108

(NH4)2SO4 seed solution. The initial seed number and mass concentrations are approximately

109

20000 cm-3 and 21 µg m-3, respectively. After the concentration of seed aerosol has stabilized,

110

either limonene (Sigma-Aldrich, 97%) or β-pinene (Sigma-Aldrich, > 99%) is injected by

111

passing air over a known volume of liquid inside a glass bulb. NO2 (Matheson Tri Gas, 500 ppm)

112

and O3 (generated by passing purified air through a UV photochemical cell) are then introduced

113

into a flow tube (1.3 L min-1 flow rate, 71 s residence time) to produce NO3 radicals and N2O5,12

114

which are injected into the chamber and typically marks the beginning of the reaction. The

115

[N2O5]:[Limonene] ratio for a typical limonene+NO3 experiment is usually about 8:1 but can be

116

as high as 18:1.

117 118

Experiments 4, 5, 9, and 10 are performed with a slightly different experimental protocol. To

119

simulate the ambient atmosphere, NO2 and O3 are injected separately into the chamber after

120

aerosol seed injection. NO2 and O3 react and continuously generate NO3 radicals, which reacts

121

with NO2 and HCHO to form N2O5 and HO2 respectively, for approximately 30 minutes prior to

122

injection of limonene. The [NO2]:[O3] ratio at the time of limonene injection is approximately

123

1:1

124 125

The addition of formaldehyde leads to formation of HO2, which opens up the RO2+HO2 reaction

126

pathway for organic peroxy radicals, which competes with the RO2+NO3, RO2+NO and

127

RO2+RO2 reaction pathways.12, 44 To evaluate the RO2 fate, a kinetic model based on the Master


Page 7 of 40


128

Chemical Mechanism (MCM) v.3.3.145, 46 is set up as described in the SI. The model results

129

show that the extent of the RO2+HO2 reaction may depend on the manner in which the

130

experiments are conducted. In typical experiments where the NO2 and O3 are pre-mixed in a flow

131

tube prior to injection into the chamber, RO2+HO2 reactions account for 43-80% of the RO2 fate,

132

RO2+RO2 reactions account for 2-38% of the fate, and the RO2+NO channel is negligible with

133

the remainder proceeding via RO2+NO3 reactions. In typical experiments where NO2 and O3 are

134

mixed within the chamber, the RO2+HO2 channel is more prominent as it accounts for 72% of

135

the oxidation. In these simulations, RO2+RO2 accounts for 20%., and only minor contributions

136

from the RO2+NO3 and RO2+NO channels can be observed.

137 138

A typical time profile of aerosol volume concentration is shown in Figure S3. About 3-5 hours

139

after peak SOA growth, the SOA in the chamber is either isothermally diluted or heated. In

140

isothermal dilution experiments, dry zero air is introduced into the chamber at a rate of 40 L min-

141

1

142

heating experiments, the chamber temperature is increased from 25 to 40 °C with a typical

143

temperature profile shown in Figure S4.

. This dilutes the particle- and gas-phase products by approximately a factor of two. In aerosol

144 145

To examine the extent and effect of mixing on evaporation between aerosol from two different

146

BVOC+NO3 reactions, we perform two types of experiments to mix SOA from the

147

limonene+NO3 reaction (hereafter referred to as ‘limonene SOA’) and SOA from the β-

148

pinene+NO3 reaction (hereafter referred to as ‘β-pinene SOA’). In the first type of experiment

149

(Experiments 15, 16, and 17), subsequently referred to as ‘Limonene Core’ experiments,

150

limonene is first introduced into the chamber and oxidized by NO3 radicals. Approximately 3



Page 8 of 40

151

hours after peak aerosol growth, β-pinene is then introduced into the chamber containing

152

limonene SOA and is followed by a second injection of N2O5. Aerosol is either heated

153

immediately after (Experiment 15) or an hour after (Experiments 16 and 17) β-pinene SOA

154

growth. In the second set of experiments (Experiment 18 and Experiment 19), subsequently

155

referred as the ‘Mixed’ experiments, limonene and β-pinene are oxidized simultaneously in order

156

to achieve as homogenous an aerosol mixture as possible. Aerosol is heated 3 hours after peak

157

SOA growth.

158 159

O3 and NOx concentrations are measured using an O3 analyzer (Teledyne T400) and a

160

chemiluminescence NOx monitor (Teledyne 200EU), respectively. Hydrocarbon concentrations

161

are measured using Gas Chromatography-Flame Ionization Detector (GC-FID) (Agilent 7890A).

162

Aerosol number and volume distributions are measured with a Scanning Mobility Particle Sizer

163

(SMPS, TSI) consisting of a differential mobility analyzer (DMA) (TSI 3040) and Condensation

164

Particle Counter (CPC) (TSI 3775). Bulk aerosol composition is measured using an Aerodyne

165

High Resolution Time-of-Flight Aerosol Mass Spectrometer (HR-ToF-AMS).47 The HR-ToF-

166

AMS NO family ions (mainly NO+ and NO2+) are included in the calculation of the N:C ratio but

167

are excluded in the calculation of the O:C ratio.

168

3) Results and Discussion

169

3.1) Aerosol yields and composition for limonene+NO3

170

Figures 1a and 1b show the typical high resolution aerosol mass spectra for the limonene+NO3

171

and β-pinene+NO3 systems, respectively. The mass spectrum for the β-pinene+NO3 system

172

(Figure 1b) is similar to that reported in Boyd et al.12 The average NO+/ NO2+ ratios for limonene

173

SOA and β-pinene SOA are both 6.3. This ratio is on the lower end of reported NO+/ NO2+ ratios 8 ACS Paragon Plus Environment

Page 9 of 40


174

in Boyd et al12 and is much higher than typical values for inorganic nitrates.22, 48, 49 Therefore, the

175

nitrate species in the aerosol phase are likely organic nitrates. The average molar N:C ratio of

176

limonene SOA in all seeded experiments is ~0.104. This is equivalent to the particle-phase

177

oxidation products containing 1.04 nitrate groups per organic molecule, if we assume each

178

organic nitrate compound has ten carbon atoms.12For the limonene+NO3 reaction, there can be

179

more than one nitrate group per molecule because limonene has two double bonds that can react

180

with NO3 radicals via radical addition. The N:C ratio of SOA formed from the β-pinene+NO3

181

reaction is about 0.076, in good agreement with the N:C ratio of 0.074 reported in Boyd et al.12

182

Using the approach described in Boyd et al.12, the rate of organic nitrate hydrolysis can be

183

inferred from the change in the AMS NO3:Org ratio (ratio of the nitrate mass to the organic

184

mass, Figure S5). Similar to the β-pinene+NO3 system, particulate organic nitrates produced by

185

the limonene+NO3 system appear to undergo slow hydrolysis.12 This is evident by a small

186

change in the nitrate mass with respect to the organic mass. Therefore, the products of the

187

limonene+NO3 reaction are likely primary/secondary organic nitrates (SI). 50, 51

188 189

The aerosol mass yields (defined as the mass of aerosol produced divided by the mass of

190

hydrocarbon reacted)28 at 25 and 40 °C under dry conditions for the limonene+NO3 system are

191

shown in Figure 2 (calculation shown in the SI). The aerosol mass yield at 25 °C for the

192

limonene+NO3 reaction is approximately constant as a function of mass loading, at 174%. This is

193

higher than aerosol mass yields measured in previously studied BVOC+NO3 systems, including

194

previous studies of the limonene+NO3 reaction. 2-5, 7, 8, 11, 12 A constant SOA mass yield suggests

195

that the volatility of the aerosol-forming products of the limonene+NO3 reaction are low enough

196

to be completely in the particle phase under the conditions studied. Aerosol mass yields for the



Page 10 of 40

197

limonene+NO3 reaction under humid conditions (RH ~ 50%) are shown in Figure S6. In contrast

198

to the β-pinene+NO3 reaction,12 aerosol mass yields are enhanced at higher humidity for the

199

limonene+NO3 reaction. This suggests that water may have different effects on different

200

BVOC+NO3 reactions and warrants further study. Aerosol mass yields for experiments without

201

seed are also shown in Figure S6 for both dry and humid conditions. The aerosol mass yields are

202

comparable to seeded experiments. Since seed aerosol does not enhance the aerosol mass yield

203

and limonene oxidation by NO3 is rapid, it is likely that the SOA formation from limonene+NO3

204

reaction is not significantly affected by vapor phase wall loss.

205

enhance yields because the products of the limonene+NO3 reaction are of sufficiently low

206

volatility that condensation of products occurs immediately upon formation. This is evidenced by

207

the nucleation of aerosol even in seeded experiments.

12, 52

Seed aerosol likely does not

208 209

Aerosol mass yields at 25 °C for previous limonene+NO3 studies3, 4, 8, 11 range from 20-40 %

210

(also shown in Figure 2). Spittler et al.4 show that the aerosol mass yield for the limonene+NO3

211

can range from 21-40%, depending on the use of either inorganic ammonium sulfate seed or

212

organic seed produced by ozonolysis of limonene. The differences between the previously

213

reported yields and this study are likely due to the difference in reaction conditions between this

214

study and previous work. For example, the limonene in previous work may have unreacted

215

double bonds3,

216

experiments are designed to oxidize both of limonene’s double bonds. Based on a kinetic box

217

model (SI), the majority (68-96%) of these oxidation products have both double bonds oxidized

218

(Figure S7), with greater than 87% of those double bonds oxidized by NO3 radicals. Although

219

SOA yields reported in previous studies may be relevant for BVOC-rich environments where

4, 11

or have one or both double bonds oxidized by ozone.8 In this work,


Page 11 of 40


220

only one of limonene’s two double bonds reacts,53 these yields may underpredict SOA formed

221

from the limonene+NO3 reaction in regions where NOx and NO3 radical concentrations are high

222

enough to oxidize both of limonene’s double bonds.

223 224

Unlike the SOA yields at 25 °C, SOA yields at 40 °C are not constant. The aerosol mass yields at

225

40 °C can be fitted using a variant of the Odum two-product model.28 Fitting the traditional two-

226

product model requires a high equilibrium partitioning coefficient for one of the two products

227

and suggests the presence of non-volatile products. By assuming one of the volatility bins (with

228

coefficient α1) is non-volatile, the data at 40 °C are fitted using Eq. 2: = +

1 +

(2)

229

Fitting Eq. (2) to the experimental data results in the following yield parameters: [α1, α2, K2]=

230

[0.673, 1.11, 0.0139]. The coefficient for α1 is larger than the second term on the right side of Eq.

231

(2) within the two-product model for mass loadings below 110 µg m-3, indicating that the

232

majority of SOA at 40 °C is non-volatile under the conditions studied. The abundance of low-

233

volatility products has important implications for the evaporation of SOA and will be discussed

234

further in the next section.

235 236

3.2) Isothermal Dilution

237

Figures 3a and 3b show the SMPS Volume:AMS SO4 ratio and the AMS Orgtotal:SO4 ratio (the

238

total organics measured by the AMS divided by the total sulfate measured by the AMS) for all

239

aerosol dilution experiments, respectively. Time zero is defined as the time when the diluting air

240

is introduced into the chamber. We use these two ratios as bounding parameters for aerosol

241

evaporation. The SMPS volume has contribution from inorganic seeds and thus the SMPS 11 ACS Paragon Plus Environment


Page 12 of 40

242

Volume:AMS SO4 ratio is less sensitive to evaporation as would be if only organics were

243

considered. However, the AMS Orgtotal:SO4 ratio can be affected by diffusion-limited growth,54

244

where nucleated organic-only particles are smaller than sulfate-containing particles and are lost

245

to the walls faster. The AMS Orgtotal:SO4 therefore could decrease even if there is no evaporation

246

present. Therefore, both of these are likely bounding parameters for the organic evaporation.

247

Normalizing the SMPS volume and AMS Orgtotal by AMS SO4 accounts for the decrease in

248

aerosol concentration that is caused by an increase of chamber volume associated with

249

isothermal dilution. While the decrease in the SMPS volume and AMS Orgtotal is due to both

250

dilution and evaporation of organic aerosol, AMS SO4 is expected to decrease solely due to

251

dilution because it is non-volatile below 40 °C.55 Therefore, any organic evaporation induced by

252

isothermal dilution will result in a decrease in the SMPS Volume:AMS SO4 and the AMS

253

Orgtotal:SO4.

254 255

The SMPS Volume:AMS SO4 ratio is approximately constant during isothermal dilution,

256

suggesting that the aerosol does not evaporate appreciably, as a lower bound.

257

Orgtotal:SO4 ratio does decrease slightly during the early stages of dilution, likely due to faster

258

loss of nucleated pure organic particles to chamber walls. Nevertheless, both the SMPS

259

Volume:AMS SO4 ratio and the AMS Orgtotal:SO4 ratio show an insignificant decrease over the

260

time of dilution, which suggests that a negligible portion of the limonene SOA is evaporating.

261

These results are somewhat unsurprising since the majority of the limonene SOA is composed of

262

low-volatility products and the mass loading of aerosol is not expected to affect the gas-particle

263

partitioning substantially. These observations suggest that the evaporation of limonene SOA

264

from dilution is thermodynamically unfavorable under the conditions studied here.

The AMS


Page 13 of 40


265

3.3) Aerosol Heating and Effective Enthalpy of Vaporization

266

Figure 2 shows that aerosol mass yields for the limonene+NO3 system depend strongly on

267

temperature, by as much as a factor of two. As the difference in aerosol yield (between the two

268

temperatures) is not constant across the entire range of organic mass loadings, a single enthalpy

269

of vaporization could not be applied to one yield curve to obtain the other yield curve.

270 271

The effective enthalpy of vaporization (∆Hv) can be used to obtain a yield curve, Y, at one

272

temperature (40 °C) from the yields at another temperature (25 °C):

273 Y (25 °C) = Y (∆Hv, 40 °C)

(3)

274

From these yield curves, we propose a method for determining the effective enthalpy of

275

vaporization as a function of mass loading. Based on Figure 2, Y (25 °C) is a constant while Y

276

(40 °C) can be fitted to Eq. (2). Substitution of the Clausius-Clapeyron27 equation for K2 in Eq.

277

(2) yields:

278 = + 279 280

298 ∆ 1 1 exp − 313 298 313

298 ∆ 1 1 1 + exp − 313 298 313

(4)

Solving Eq. (4) for ∆Hv gives the following dependence of ∆Hv on Mo:

281



∆ = +

− !" $ 298 + − ∗ 313 1 1 1 − 298 313

Page 14 of 40

(5)

282

The results of this analysis are shown in Figure 4. The effective enthalpy of vaporization ranges

283

between 117-237 kJ mol-1 and decreases with increasing mass loading. The effective enthalpy of

284

vaporization likely decreases with mass loading due a larger fraction of products with low C*

285

(which correlates with high effective enthalpy of vaporization for a wide arrange of organic

286

compounds)37 at low mass loadings. These results highlight the importance of determining the

287

effective enthalpy of vaporization. Currently, most atmospheric models use enthalpies that range

288

from 18-156 kJ mol-1 with over 80% of them using a value of less 60 kJ mol-1 39 and references therein.

289

This is lower than the enthalpies of vaporization calculated for SOA produced by the

290

limonene+NO3 reaction in this study (117-237 kJ mol-1). This highlights the importance to

291

examine other aerosol systems to ensure they too are not similarly underpredicted, as this may

292

have a significant effect on the aerosol mass calculated in atmospheric models.

293 294

When heated from 25 °C to 40 °C, the SOA formed from the limonene+NO3 evaporates less than

295

expected based on the difference in yield at 25 °C and 40 °C (Figure S8, TOC). To quantify this

296

difference, we define the ‘Heating Ratio’ and ‘Formation Ratio’. The ‘Heating Ratio’ is

297

determined to be the mass of OA remaining after heating the aerosol formed at 25 °C to 40 °C

298

divided by the mass of OA formed at 25 °C. This is calculated by dividing the AMS Orgtotal:SO4

299

(15-minute averaged) after the chamber has achieved a temperature of 38 °C by the AMS

300

Orgtotal:SO4 prior to heating. The ‘Formation Ratio’ is determined to be the mass of aerosol

301

formed at 40 °C divided by the mass of aerosol formed at 25 °C. At low mass loadings, there is a 14 ACS Paragon Plus Environment

Page 15 of 40


302

large difference between the ‘Heating Ratio’ and ‘Formation Ratio’, which suggests that there is

303

a resistance to aerosol evaporation. These ratios are calculated for experiments over a range of

304

mass loadings and we find that the ‘Heating Ratio’ for limonene SOA is fairly independent of

305

mass loading (Figure S9). There are several possibilities for this observation: 1) the products

306

produced at 25 °C and 40 °C have large differences in chemical composition, 2) oligomerization

307

following condensation may produce low volatility particle-phase compounds,56-58, or 3) the

308

aerosol has a kinetic limitation and does not evaporate on the time scales of hours, which is

309

longer than the predicted time scales of minutes in other SOA systems57,

310

components.59,

311

limonene+NO3 reaction formed at 25 °C and 40 °C have substantial chemical differences (SI).

312

Both oligomerization and the presence of a kinetic limitation are possible explanations for the

313

observed data. Oligomerization would be possible if a subset of the reaction products only

314

condense at high mass loadings or at temperatures much below 40 °C (e.g., 25 °C). Once formed

315

in the gas phase and subsequently partitioned to the particle phase, these compounds could

316

oligomerize to form products that do not evaporate after the temperature is increased to 40 °C.

317

Although oligomerization of products is possible, a kinetic limitation is also a feasible

318

explanation for the results if the limonene SOA is semi-solid, with the outer layers of aerosol

319

hindering the evaporation of inner layers. In Section 3.4, experiments using SOA produced by

320

two reaction systems (limonene and β-pinene) further support the hypothesis that the outer layers

321

of an SOA particle may hinder the evaporation of the inner layers.

60

59

or for pure

It is unlikely that the total (gas- and particle-phase) products of the

322



Page 16 of 40

323

3.4) Limitations to Aerosol Evaporation

324

It is challenging to quantify the degree of homogeneity within an aerosol particle formed from a

325

single VOC oxidation system (i.e., either pure β-pinene SOA or pure limonene SOA). It has been

326

demonstrated in previous studies that evaporation of organic aerosol may be hindered if it is

327

coated with organic aerosol from a different precursor.61 In this study, experiments to mix SOA

328

are performed to examine the interactions between SOA formed from the limonene+NO3 and β-

329

pinene+NO3 reactions. We heat up the mixed SOA produced from two different methods, which

330

are described in Section 2 as the “Limonene Core” and “Mixed” experiments.

331 332

We use the differences in the mass spectra of limonene SOA and β-pinene SOA (Figure 1) to

333

quantify the extent of evaporation for each SOA type upon heating.

334

characterized by an abundance of the C2H3O+ ion (m/z 43), whereas the β-pinene SOA spectrum

335

has notably high C5H7+ (m/z 67) and high C7H7+ (m/z 91). The mass spectrum of total OA, [MS],

336

is assumed to be a linear combination of limonene SOA, [Lim], and β-pinene SOA, [Bpin]

337

spectra, as represented by Eq. (6).

%&' = ( ∙ %*+,' + - ∙ %./+!'

Limonene SOA is

(6)

338

The coefficients of this linear combination, a and b, (summed to 1) represent the mass fractions

339

of limonene SOA and β-pinene SOA in the total aerosol, respectively. From these mass fractions

340

and the AMS Orgtotal:SO4 ratio, the Orglim:SO4 (the organic aerosol formed from limonene

341

divided by the sulfate measured by AMS) and Orgβpin:SO4 (the organic aerosol formed form β-

342

pinene divided by the sulfate measured by AMS) ratios over the course of the experiments can be

343

determined. By normalizing to SO4, we can account for changes in AMS collection efficiency

344

and particle wall loss. Since heating is slow right after increasing the temperature set point


Page 17 of 40


345

(Figure S4), the evaporation of aerosol is not considered until the chamber temperature has

346

reached 26 °C to avoid noise in measurements. The Orglim:SO4 and Orgβpin:SO4 ratios are

347

normalized by the average Orglim:SO4 and Orgβpin:SO4 ratios during the first five minutes to

348

facilitate comparison between experiments. The change in the Orglim:SO4 and Orgβpin:SO4 ratios

349

can therefore be used as a proxy for SOA evaporation for each precursor. The results of this

350

analysis are shown in Figure 5. The fractions of limonene SOA (5a) and β-pinene SOA (5b) that

351

remain after increasing the chamber temperature are dependent on how the initial SOA is formed

352

(i.e., “Limonene Core” vs. “Mixed” experiments).

353 354

There is less evaporation of limonene SOA and more evaporation of β-pinene SOA in the

355

‘Limonene Core’ experiments than in the ‘Mixed’ experiments as shown in Figure 5. If

356

evaporation were controlled solely by thermodynamics, the ‘Limonene Core’ and ‘Mixed’

357

experiments should have no difference in evaporation for the two types of aerosol. Possible

358

reasons for our observation include: 1) in the ‘Limonene Core’ experiment, the products from the

359

β-pinene+NO3 reaction nucleate into new particles that are smaller than the existing particles of

360

limonene SOA and are thus lost to the chamber wall more rapidly, and 2) in the “Limonene

361

Core” experiment, the β-pinene SOA forms a shell over the existing limonene SOA particles and

362

hinders the evaporation of the limonene SOA core. Possibility 1) can be eliminated because the

363

particle concentration does not increase appreciably ( 100 is negligible.

Figure 1: Typical high-resolution mass spectra of SOA formed by the reactions of

818 819

Figure 2: Aerosol mass yield as a function of organic loading for the limonene+NO3 reaction at

820

25 °C and 40 °C. The density of aerosol used to calculate the aerosol mass yield is determined in

821

the limonene+NO3 experiments without inorganic seed aerosol.81 The aerosol mass yield at 25

822

°C is approximately constant while the aerosol mass yield at 40 °C is fitted using Eq. 4, which is

823

modified from the two-product model proposed by Odum et al.28 The aerosol mass yields

824

obtained in this study are compared to those by Hallquist et al.3 and Fry et al.7, 8 The x-axis error

825

bars (which are smaller than the size of the data points) are calculated using one standard

826

deviation of volume measured by SMPS at peak aerosol growth. The y-axis error bars are

827

calculated with an 8% uncertainty in hydrocarbon injection and one standard deviation of the

828

aerosol volume measured by SMPS at peak aerosol growth.

829 830

Figure 3: (a) Aerosol volume or (b) AMS Orgtotal normalized by AMS SO4 for all experiments

831

that undergo isothermal dilution. Normalizing the data by AMS SO4 accounts for any decrease in

832

aerosol mass that may be caused by particle wall loss in addition to dilution. Under all

833

conditions, aerosol evaporation caused by isothermal dilution is negligible.

834


Page 33 of 40


835

Figure 4: The effective enthalpy of vaporization for the limonene+NO3 reaction as a function of

836

mass loading. At low mass loadings, the least volatile reaction products dominate the particle

837

phase. These products are more likely to have a higher enthalpy of vaporization than the high

838

volatility products. As the mass loading of aerosol increases, volatile products with high enthalpy

839

of vaporization will contribute more to the aerosol phase and lower the overall effective enthalpy

840

of vaporization.

841 842

Figure 5: The fraction of aerosol remaining for a) limonene SOA and b) β-pinene SOA. Data

843

taken at low humidity (RH < 2%) are represented with closed circles while data taken at high

844

humidity (RH = 70 %) are represented with open circles. Only the time after the chamber has

845

reached a temperature of 26 °C is considered for this analysis. The limonene SOA evaporates

846

less in the ‘Limonene Core’ experiments than in the ‘Mixed’ experiments. The β-pinene SOA

847

evaporates more in the ‘Limonene Core’ experiments than in the ‘Mixed’ experiments.

848 849 850



851 852

Page 34 of 40

Figure 1:

20

30

40

m/z 50 60

70

80

90

100

0.24 Limonene+NO3 0.20 Cx CxHy CxHyO CxHyOz CxHyN CxHyON CxHyOzN NOz

0.16 0.12

Fraction of Total Signal

0.08 0.04 0.00 0.16 0.14

β-pinene+NO3

0.12 0.10 0.08 0.06 0.04 0.02 0.00 20

853

30

40

50

60 m/z

70

80

90

100


Page 35 of 40

854 855


Figure 2:

200 180 Aerosol Mass Yield (%)

160 140 120 100 80

Limonene+NO3 Yield Curve (T = 25 °C) (This Study) Limonene+NO3 Yield Curve (T = 40 °C) (This Study) T = 25 °C (This Study) T = 40 °C (This Study)

60 40

T = 25 °C (Hallquist et al., 1999) T = 25 °C (Fry et al., 2009) T = 25 °C (Fry et al., 2014)

20 0 0

40

80

120

160

200

3

856

Organic Mass Loading (µg/m )



Figure 3:

20

a)

16 12.0 ppb Limonene; 25 °C

3

3

3

SMPS Volume / AMS SO4 (µm / cm / µg / m )

857

Page 36 of 40



4

4.9 ppb Limonene; 40 °C

0 0 14

100 200 300 400 500 Time Since Dilution Began (minutes)

600

b)

AMS Orgtotal:SO4

12 10 12.0 ppb Limonene; 25 °C


6 4


2


0 0 858 859

100 200 300 400 500 Time Since Dilution Began (minutes)

600


Page 37 of 40

860 861


Figure 4: 260 240

Enthalpy of Vaporization (kJ/mol)

220

Enthalpy of Vaporization Limonene+NO3 ∆Mo (40 °C)

200 180 160 140

High Volatility Products Low ∆H, High C*

120 100 80 60

Low Volatility Products High ∆H, Low C*

40 20 0 0

20

40

60

80

100

120

140

160

180

200

3

862 863 864

Organic Mass Loading (µg/m )



Figure 5:

Limonene SOA Fraction Remaining

865 866

Page 38 of 40

a)

1.00 0.95 0.90 0.85 0.80 0.75 0.70

"Limonene Core" Immediate Heating "Limonene Core" Delayed Heating "Mixed"

0.65 0.60 0.55 0.50

β-pinene SOA Fraction Remaining

0

40

80 120 160 Time Since 26 °C (min)

1.00

200

b)

0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0

867 868

40

80 120 160 Time Since 26 °C (min)

200


Page 39 of 40

869 870 871


Table 1: Experimental conditions for all experiments

RH

Seed

VOC

1 2 3 4b 5b 6 7 8 9b 10b 11 12 13 14

Reaction Temperature (°C) 24.3 24.4 24.4 25.0 24.4 24.2 38.5 39.4 39.5 38.7 38.7 24.3 24.2 24.9

< 3% < 2%

Secondary Organic Aerosol (SOA) from Nitrate Radical Oxidation of

Recommend Documents