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
Environmental Science & Technology
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
2 ACS Paragon Plus Environment
Page 3 of 40
Environmental Science & Technology
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
Environmental Science & Technology
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
4 ACS Paragon Plus Environment
Page 5 of 40
Environmental Science & Technology
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
Environmental Science & Technology
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
6 ACS Paragon Plus Environment
Page 7 of 40
Environmental Science & Technology
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
7 ACS Paragon Plus Environment
Environmental Science & Technology
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
Environmental Science & Technology
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
9 ACS Paragon Plus Environment
Environmental Science & Technology
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,
10 ACS Paragon Plus Environment
Page 11 of 40
Environmental Science & Technology
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
Environmental Science & Technology
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
12 ACS Paragon Plus Environment
Page 13 of 40
Environmental Science & Technology
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
13 ACS Paragon Plus Environment
Environmental Science & Technology
∆ = +
− !" $ 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
Environmental Science & Technology
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
15 ACS Paragon Plus Environment
Environmental Science & Technology
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
16 ACS Paragon Plus Environment
Page 17 of 40
Environmental Science & Technology
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
32 ACS Paragon Plus Environment
Page 33 of 40
Environmental Science & Technology
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
33 ACS Paragon Plus Environment
Environmental Science & Technology
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
34 ACS Paragon Plus Environment
Page 35 of 40
854 855
Environmental Science & Technology
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 )
35 ACS Paragon Plus Environment
Environmental Science & Technology
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
12 12.0 ppb Limonene; 40 °C
8 4.5 ppb Limonene; 25 °C
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
8 12.0 ppb Limonene; 40 °C
6 4
4.5 ppb Limonene; 25 °C
2
4.9 ppb Limonene; 40 °C
0 0 858 859
100 200 300 400 500 Time Since Dilution Began (minutes)
600
36 ACS Paragon Plus Environment
Page 37 of 40
860 861
Environmental Science & Technology
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 )
37 ACS Paragon Plus Environment
Environmental Science & Technology
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
38 ACS Paragon Plus Environment
Page 39 of 40
869 870 871
Environmental Science & Technology
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%
