Aerosol Science and Technology, 45:244–261, 2011
Copyright © American Association for Aerosol Research
ISSN: 0278-6826 print / 1521-7388 online
DOI: 10.1080/02786826.2010.532178
Secondary Organic Material Produced by the Dark
Ozonolysis of ␣-Pinene Minimally Affects the
Deliquescence and Efflorescence of Ammonium Sulfate
M. L. Smith,1 M. Kuwata,1 and S. T. Martin1,2
1
2
School of Engineering and Applied Sciences, Harvard University, Cambridge, Massachusetts, USA
Department of Earth and Planetary Sciences, Harvard University, Cambridge, Massachusetts, USA
Downloaded by [Harvard College] at 13:11 07 September 2012
1.
The hygroscopic phase transitions and growth factors of mixed
particles having as components ammonium sulfate and secondary
organic material (SOM) were measured. The SOM was generated by the dark ozonolysis of α-pinene, and organic particle mass
concentrations of 1.63 and 12.2 µg m−3 were studied. The hygroscopic properties were investigated using a 1×3 tandem differential mobility analyzer (1×3-TDMA). The 1×3-TDMA takes
advantage of the hysteresis between solid-to-aqueous and aqueousto-solid phase transitions to determine the efflorescence and deliquescence relative humidities (ERH and DRH, respectively) of
materials. Overall, the influence of the SOM produced by the dark
ozonolysis of α-pinene on the ERH and DRH of ammonium sulfate
was small, shifting for example the DRH from 80% for pure ammonium sulfate to 77% for organic volume fractions of 0.96. The
ERH likewise shifted by only a small amount across this composition range, specifically from 31 to 29%. The SOM produced at the
lower organic particle mass concentrations shifted ERH and DRH
even less, indicating an influence of SOM chemical composition
on phase transitions. The hygroscopic growth factors of the mixed
particles were adequately modeled across the range of studied RH
(50 to 83%) using volume-averaged growth factors of the pure materials. The results for ERH, DRH, and the growth factors were all
consistent with a model of phase separation between the inorganic
and organic phases in individual particles, at least for the studied
RH values (<83%) and for SOM prepared by α-pinene ozonolysis.
[Supplementary materials are available for this article. Go to the
publisher’s online edition of Aerosol Science and Technology to
view the free supplementary files.]
Received 2 August 2010; accepted 23 September 2010.
This material is based upon work supported by the National Science Foundation under Grant No. 0925467. Any opinions, findings, and
conclusions or recommendations expressed in this material are those
of the authors and do not necessarily reflect the views of the National
Science Foundation. MLS is the recipient of a Graduate Research Environmental Fellowship from the Global Change Education Program
of the Department of Energy.
Address correspondence to S. T. Martin, School of Engineering
and Applied Sciences & Department of Earth and Planetary Sciences, Harvard University, 29 Oxford Street, Cambridge, MA 02138,
USA. E-mail: scot martin@harvard.edu; httn://www.seas.harvard.edu/
environmental-chemistry
244
INTRODUCTION
Aerosol particles significantly affect the balance of global
radiation and the pathways of atmospheric chemistry by scattering and absorbing incoming solar radiation, serving as cloud
condensation nuclei, and participating in heterogeneous chemical reactions (Seinfeld and Pandis 2006; Solomon et al. 2007).
The effects depend on particle phase, composition, and size,
all of which are influenced by particle deliquescence, efflorescence, and hygroscopic growth (Hanel 1976). Deliquescence is
the phase transition that occurs when a solid spontaneously takes
up water at a specific relative humidity (RH) to form a saturated
solution with water (Martin 2000). The deliquescence relative
humidity (DRH) is the RH at which the Gibb’s free energy of
an aqueous particle equals that of a solid particle surrounded by
water vapor. Efflorescence, in contrast to deliquescence, is governed by kinetics. The efflorescence relative humidity (ERH) is
the RH at which a solid crystallizes from solution, accompanied
by evaporation of water from the droplet. Because a nucleation
barrier must be overcome for crystallization to occur, the solute
does not crystallize until a critical supersaturation is reached.
The consequence is that the ERH is typically significantly below the DRH, resulting in a hysteresis effect for particle phase.
This hysteresis has been predicted to alter global annual sulfate radiative forcing by up to 16% (Martin et al. 2004; Wang
et al. 2008a,b). The RH-dependent hygroscopic growth factor
g(RH), defined as the ratio of particle diameter at an elevated
RH to that at a low reference RH, quantifies diameter growth
resulting from water uptake.
The hygroscopic phase transitions and growth factors of
many inorganic particles have been studied extensively over
a wide range of composition and diameter (Tang and Munkelwitz 1984; Cziczo et al. 1997; Ha and Chan 1999; Onasch et al.
1999; Martin 2000). The measured DRH values and growth
factors of these compounds generally agree well with thermodynamic predictions, including shifts for nano-sized materials
(Hameri et al. 2000, 2001; Biskos et al. 2006a,b). The dependence of ERH on inorganic composition has been more difficult
to codify because of the stochastic nature of nucleation; however, there are measurements and useful parameterizations in
Downloaded by [Harvard College] at 13:11 07 September 2012
PHASE TRANSITIONS OF SOM-AS PARTICLES
the literature for some compositions (Martin 2000; Martin et al.
2003).
In addition to inorganic components, atmospheric particles
also consist of organic components. With respect to fine-mode
particle mass concentration, the organic components can be
dominant at many times and locations (Zhang et al. 2007; Hallquist et al. 2009). Several groups have studied the DRH and
ERH properties of particles generated by mixing pure organic
compounds with inorganic salts in solution with water (Cruz
and Pandis 2000; Peng et al. 2001; Brooks et al. 2002; Parsons et al. 2004; Marcolli and Krieger 2006). For the most part,
the organic compounds chosen for study have been carboxylic
acids because they have been observed in real atmospheric
organic material and therefore may represent constituents of
secondary organic material (SOM). The studies have shown
that the effects of specific organic molecules on the hygroscopic phase transitions of the inorganic component depend
on the interactions between both components. For example, in
regard to ERH Choi and Chan (2002) showed that succinic
acid (HOOCCH2 CH2 COOH), which is relatively insoluble in
water, increased the ERH of sodium chloride (NaCl) and ammonium sulfate (AS) by 8 and 12% RH, respectively, through a
mechanism of heterogeneous nucleation of NaCl and AS crystals. In comparison, the same authors showed that glutaric acid
(HO2 C(CH2 )3 CO2 H), which is highly soluble in water, also increased the ERH of NaCl by 14% and increased the ERH of AS
by 22% RH but must have done so by a homogeneous mechanism that influenced germ formation. As another example, Marcolli and Krieger (2006) investigated particles composed of five
carboxylic acids (including malic, maleic, malonic, glutaric, and
methysuccinic acids) and ammonium sulfate and found that the
ERH of micrometer-sized particles decreased with increasing
organic volume fraction, inferring that AS served as heterogeneous nuclei for carboxylic acid crystallization. In a third example Parsons et al. (2006), using malonic acid (CH2 (COOH)2 )
as the organic component, found that the ERH decreased with
increasing organic volume fraction until becoming completely
suppressed above an organic volume fraction of approximately
0.5.
The DRH values of mixed inorganic–organic particles have
also been studied, and some generalities with regard to the observations are possible (Choi and Chan 2002; Brooks et al.
2003; Parsons et al. 2004). Water-soluble organic compounds
decrease the DRH compared to pure inorganic particles, and significant water uptake caused by partial dissolution occurs prior
to final deliquescence. In comparison, water-insoluble organic
molecules affect the DRH of the inorganic component only in a
minor way, if at all. As an example, Choi and Chan (2002) reported that particles of AS mixed with the insoluble compound
succinic acid deliquesced at the DRH of AS and exhibited no
measurable hygroscopic growth for RH < DRH. Parsons et al.
(2004) extended these results by showing that the DRH was
statistically identical to that of pure AS for small amounts of organic material in mixed systems of malonic acid, levoglucosan,
245
or glycerol. The DRH, however, decreased by up to 10% RH as
the organic mole ratio increased from 0.35 to 0.6. These studies
are consistent with the thermodynamic prediction that particles
of mixed composition deliquesce at an RH less than the DRH
of the pure components (Wexler and Seinfeld 1991).
What emerges from the literature of both ERH and DRH is
that the organic volume fraction is a key variable for predicting
the influence of specific organic compounds on the hygroscopic
phase transitions of inorganic materials. The influence, however, of secondary organic material does not necessarily follow
as a direct extension of the results using pure compounds. SOM
forms by the multi-generation OH, O3 , and NO3 oxidation of
volatile organic compound (VOC) precursors (Hallquist et al.
2009), and the resulting SOM in the particle phase consists of
the low-volatility fraction of the products (Donahue et al. 2006).
Tens to hundreds of different organic molecules occur in significant quantities (Hamilton et al. 2004), and these mixtures
can behave much differently than isolated pure-compound constituents (Marcolli et al. 2004). The prevailing hypothesis in
regard to ERH discussed in Martin et al. (2004), Parsons et al.
(2004), and Marcolli et al. (2004) has been that the multiple compounds in SOM combine in a way that should reduce the ERH of
mixed SOM-inorganic particles and eventually eliminate crystallization entirely. In regard to DRH, the prevailing hypothesis
has been that SOM is water soluble, as inferred from cloudcondensation-nuclei activity of the material (King et al. 2007,
2009, 2010) as well as its hygroscopic growth (Virkkula et al.
1999; Cocker et al. 2001; Saathoff et al. 2003; Baltensperger
et al. 2005; Varutbangkul et al. 2006; Meyer et al. 2009), and
should therefore considerably influence the deliquescence properties of AS.
To date, there have been few studies on the phase transitions
of mixed SOM-inorganic particles, and the aforementioned hypotheses could be tested only relatively weakly by the available
data. A summary of the studies is provided in Table 1. Kleindienst et al. (1999) found that the DRH and ERH of AS did
not shift for mixed particles having AS and SOM components
for organic volume fractions of 0.05 to 0.16. The SOM was
produced by oxidation of toluene, p-xylene, and 1,3,5-TMB.
Saathoff et al. (2003) placed limits of between 70 and 80% RH
on the deliquescence of particles composed of AS and SOM
products of α-pinene ozonolysis for an organic volume fraction
of 0.18. Takahama et al. (2007) determined that the ERH values
did not differ from the reference value for pure AS, at least within
the measurement uncertainty, for AS particles coated with the
ozonolysis products of α-pinene for organic volume fractions
of 0.54 to 0.72 or with those of limonene ozonolysis for organic
volume fractions of 0.59 to 0.94. For α-pinene photooxidation,
Meyer et al. (2009) inferred that the DRH was below 75% for
mixed particles of AS and SOM for organic volume fractions
greater than 0.2. A summary statement, as apparent from an examination of Table 1, is that these previous studies did not carry
out systematic evaluations of the influence of the SOM organic
volume fraction on the ERH and DRH of AS. First-principle
246
M. L. SMITH ET AL.
TABLE 1
Summary of the literature concerning the deliquescence and efflorescence phase transitions of internally mixed particles of
ammonium sulfate and secondary organic material
VOC
precursor
Oxidation
pathway
Organic
fraction
Photo-oxidation
0.16
0.13
0.05
0.18
0.54–0.72
0.59–0.94
0.0–0.2
0.2–0.9
Toluene
p-xylene
1,3,5-trimethylbenzene
α-pinene
α-pinene limonene
Dark ozonolysis
Dark ozonolysis
α-pinene
Photo-oxidation
DRH (%)
ERH (%)
80
80
80
Between 70 and 80∗
NA
30
30
30
NA
28–34 ± 2.5
28–34 ± 2.5
NA
Between 75 and 85∗
<75∗
Reference
Kleindienst et al. (1999)
Saathoff et al. (2003)
Takahama et al. (2007)
Meyer et al. (2009)
Downloaded by [Harvard College] at 13:11 07 September 2012
Reference values of DRH and ERH of ammonium sulfate are 79.5% and 30–35%, respectively (Martin 2000). ∗ Experiments in these studies
were conducted using protocols that did not allow more precise determination of the RH of the phase transition.
considerations, however, as well as results reported in the literature using pure organic compounds, support the importance of
organic volume fraction in quantifying the influence of SOM on
the phase transitions of inorganic materials.
Herein, results are presented of DRH, ERH, and g(RH) for
particles consisting of AS and SOM components over the full
range of organic volume fractions. The SOM is produced from
the dark ozonolysis of α-pinene. Because the chemical composition of SOM components in particles changes with organic
particle mass concentration (Shilling et al. 2009; Chhabra et al.
2010), results are presented for studies at two mass concentrations. The hygroscopic phase transitions and growth factors are
investigated using a 1×3 tandem differential mobility analyzer
(1×3-TDMA) (Martin et al. 2008; Rosenoern et al. 2009). The
1×3-TDMA takes advantage of the hysteresis between solidto-aqueous and aqueous-to-solid phase transitions to determine
ERH and DRH. The results show that both of the aforementioned hypotheses regarding the effects of SOM on the ERH
and DRH of ammonium sulfate are not accurate (at least for the
studied SOM material) in that DRH and ERH do both decrease
but do so only slightly (e.g., by 2–3% RH) across a range of
0.0–0.99 in organic volume fraction. Explanations and implications for the chemical nature of SOM and its interactions with
inorganic materials are discussed.
2.
EXPERIMENTAL
2.1. Aerosol Generation and Characterization
The Harvard Environmental Chamber (HEC), operated in
a continuous-flow mode as described in detail in King et al.
(2009), was used to coat ammonium sulfate particles with secondary organic material. The HEC consists of a 4.7 m3 Teflon
bag attached within a temperature-controlled housing maintained at 25◦ C. The total flows (22 Lpm) in and out of the
bag were balanced, implying a reactor residence time of 3.6 h.
The relative humidity and the ozone concentration in the bag
were 40% and 50 ppb, respectively, and were kept at steady state
by feedback control.
Aqueous polydisperse ammonium sulfate particles were generated outside the chamber by nebulizing a 0.05 wt% solution
(TSI 3076), and the resulting particles were dried to below 10%
RH using a silica-packed diffusion drier. A nano-differential
mobility analyzer (nDMA; TSI, 3085), operated with an aerosol
flow rate of 2 Lpm and a sheath flow rate of 10 Lpm, was employed to select a quasi-monodisperse size fraction, which was
then continuously injected into the bag. A solution (1:600 v/v)
of (+)-α-pinene (Aldrich, 98% purity) in 2-butanol (SigmaAldrich, ≥99.5% purity) was introduced into a gently warmed
glass bulb outside the chamber, and the vapor was then continuously swept into the bag.
The α-pinene reacted with ozone inside the bag, forming
secondary organic material. Two experimental conditions were
used to achieve a wide range of organic volume fractions: (1)
76.5 nm seed particles with 2 ppb injected α-pinene (i.e., in
absence of reaction) and (2) 49.5 nm seed particles with 20 ppb
injected α-pinene. The data analysis included treatment of the
multiply charged seed particles, such as +2 particles having a
physical diameter of 113 nm that passed through the nDMA
alongside +1 particles having a physical diameter of 76.5 nm.
Flow was continuously withdrawn from the HEC for analysis. The withdrawn particles were on the lower branch of
the water uptake hysteresis curve. The particles in the outflow were characterized by a scanning mobility particle sizer
(SMPS, TSI 3936; Wang and Flagan 1990) to obtain the numberdiameter distribution and by a high-resolution time-of-flight
Aerodyne Aerosol Mass Spectrometer (HR-ToF-AMS; DeCarlo
et al. 2006) to assay organic particle mass concentration and
O:C and H:C elemental ratios (Shilling et al. 2009). A redundant estimate of the mass concentration was obtained from the
AMS-determined SOM density (1.62 and 1.4 g cm–3 for 2 and
PHASE TRANSITIONS OF SOM-AS PARTICLES
Downloaded by [Harvard College] at 13:11 07 September 2012
20 ppb injected α-pinene, respectively) and the SMPSdetermined SOM volume concentration (Shilling et al. 2008).
The SOM density was confirmed by measurement with a DMAAPM (Kanomax 3600 Aerosol Particle Mass Analyzer (APM);
Ehara et al. 1996; Kuwata and Kondo 2009; Malloy et al. 2009).
Experiments were conducted at organic particle mass concentrations of 1.63 (1.41) and 12.2 (14.6) µg m−3 , with the SMPS- and
AMS-determined values shown outside and inside the parentheses, respectively.
2.2. 1×3-TDMA
One part of the flow withdrawn from the HEC was continuously sampled by the 1×3-TDMA (2.0 Lpm). The 1×3-TDMA,
described in detail by Rosenoern et al. (2009) and depicted in
Figure 1, takes advantage of the hysteresis between DRH and
ERH to measure irreversible water uptake. In brief, particles
sampled into the 1×3-TDMA were first pre-conditioned to a
reference RH0 (Perma Pure Nafion tube, MD 110). A quasimonodisperse size fraction was subsequently selected by the
first DMA (DMAmono ; TSI, 3081; 9.6:1 sheath:sample flow ratio; 25◦ C) (Figure 1). Following DMAmono , the flow was split
and directed into reference and test arms. Each arm consisted
of two Nafion tubes, a differential mobility analyzer, and a condensation particle counter (CPC; TSI, 3010) in series.
2.2.1. Reference Arm
In the reference arm, both Nafion tubes were maintained at
RH0 . In this case, the particles were not exposed to any RH
perturbations, and the particle diameter remained unchanged.
With DMAmono and DMA0 (9.38:1 sheath:sample flow ratio;
25◦ C) set to the same mobility diameter, the particles passed
through DMA0 to be counted downstream by CPC0 .
2.2.2. Deliquescence Test
In the test arm of a deliquescence-mode experiment, the particles, which exited the HEC on the lower branch of the water
uptake hysteresis curve, were exposed in the first Nafion tube to
elevated RH (denoted RH+δ ; Figure 1a) for approximately 1 s
and in the second Nafion tube were returned to RH0 . The particles were classified by DMA+δ , with DMAmono and DMA+δ
set to the same mobility diameter. In the case of a positive test
for deliquescence, the particle diameter increased because of
the hysteresis effect induced by this RH sequence: the larger
diameter particles did not pass DMA+δ and therefore were not
counted by CPC+δ . By contrast, in the absence of deliquescence
(i.e., a negative test), particle diameter was unchanged, and the
particles passed through DMA+δ to be counted by CPC+δ .
For the DRH measurements, RH0 and RH+δ were scanned
from 65 to 80% with a constant offset +δ of 5%. In one scan,
the sampled mobility diameter was held constant. Across a set
of experiments, a combination of seed particle diameters in the
HEC and particle diameters sampled by the 1×3-TDMA was
used to systematically vary the organic volume fraction from 0
to 0.96 (Table 2).
247
2.2.3. Efflorescence Test
For the efflorescence-mode experiments, prior to sampling
by the 1×3-TDMA, the dry particles exiting the HEC were deliquesced by exposure to 84% RH, as set by a saturated solution
of potassium chloride. The particles, now on the upper branch of
the water uptake hysteresis curve, were then equilibrated to RH0
as they entered the 1×3-TDMA. In an efflorescence test, the particles were exposed in the first Nafion tube of 1×3-TDMA to
a lowered RH (denoted RH−δ ; Figure 1b) for approximately 1
s and in the second Nafion tube were returned to RH0 . This
RH sequence decreased the particle diameter irreversibly if efflorescence occurred. The particles were classified by diameter
using DMA−δ , and particles passing DMA−δ were counted by
CPC−δ .
For the performed measurements, RH−δ was scanned from
40% to 20% while RH0 was held constant at 50%, corresponding to –δ of 10% to 30% during the scan. This approach was
a modification to the experimental protocol of Rosenoern et al.
(2009), who varied RH0 as in the DRH experiments and conducted simultaneous ERH and DRH scans. This approach was
possible for Rosenoern et al. (2009) because the composition of
the studied particles did not vary with diameter. In our study, the
revised approach of fixed RH0 in ERH scans was utilized so that
the organic volume fraction of particles selected by DMAmono
remained the same throughout one scan. As a result of this
requirement for different RH0 profiles, the DRH and ERH measurements in this study were carried out separately. As for the
DRH measurements, the organic volume fraction for the ERH
measurements was varied using a combination of seed particle
diameters and sampled particle diameters (Table 2).
2.2.4. Transmission Ratio
The transmission ratio of the 1×3-TDMA is defined
as CPC+δ :CPC0 for a deliquescence-mode experiment and
CPC−δ :CPC0 for an efflorescence-mode experiment. Particles
entering DMA0 exited in the flow that continued to CPC0 and
were counted by that instrument. Likewise, in the absence of a
phase transition in the efflorescence or deliquescence test arms,
particles entering DMA−δ and DMA+δ exited in the flows that
continued to CPC−δ and CPC+δ . In these cases, the transmission ratios CPC+δ :CPC0 or CPC−δ :CPC0 were ideally unity.
Because particles that deliquesced (effloresced) had a changed
mobility diameter, these particles wholly to partially drifted
laterally from the flow that exited DMA+δ (DMA−δ ) and that
subsequently entered CPC+δ (CPC−δ ). In the case of complete
drift, the counts recorded by CPC+δ (CPC−δ ) correspondingly
dropped to zero. The transmission ratio accompanying a phase
transition therefore ideally also dropped to zero.
In practice, ideal behavior was not always observed, both for
practical as well as scientific reasons. For instance, as a practical reason, the arms were at times slightly unbalanced. In this
case, the transmission ratio for changing RH was constant in
the absence of phase transitions but different from unity (e.g.,
between 0.85 and 1.12). A scaling factor was applied in the
M. L. SMITH ET AL.
Downloaded by [Harvard College] at 13:11 07 September 2012
248
FIG. 1. Schematic representation of the 1×3 Tandem Differential Mobility Analyzer (1×3-TDMA). Panels A and B depict deliquescence and efflorescence tests,
respectively. The apparatus is represented by boxes and arrows. Alongside, the colored circles of different sizes represent the physical processes of deliquescence,
efflorescence, and hygroscopic growth of aerosol particles as they move through the apparatus. Items labeled in the figure include the seed-particle generation
system (SGS), the Harvard Environmental Chamber (HEC), Nafion RH0 , RH+δ , and RH−δ conditioners, differential mobility analyzers DMAmono , DMA0 ,
DMA+δ , DMA−δ , and condensation particle counters CPC0 , CPC+δ , and CPC−δ . In each panel, particles pass the reference arm and are detected by CPC0 . For
the test arms, two possible pathways are considered, namely one for which a phase transition occurs and another for which it does not. For the former case, the
change in diameter shifts the particles out of the mobility filter of the downstream DMA, and these particles are thus not passed for counting to the CPC further
downstream. In the figure, circles represent aerosol particles having as components solid ammonium sulfate (AS), secondary organic material (SOM), water (H2 O),
and mixed AS-H2 O. The circles represent different particle types in a generic sense. They should not be interpreted as representative of particle morphologies,
which can be much more complex (Figure 10).
249
PHASE TRANSITIONS OF SOM-AS PARTICLES
TABLE 2
List of experiments and investigated conditions. The secondary organic material was generated in the Harvard Environmental
Chamber at 25◦ C, 40% RH, 50 ppbv ozone, and 3.7 h residence time
Downloaded by [Harvard College] at 13:11 07 September 2012
Experiment
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Initial αpinene (ppbv)
SOM mass
concentration (µg m−3 )
Seed
+1
dm,+1
(nm)
Seed
+2
dm,+1
(nm)
Selected
dm,+1 (nm)
ε̂+1
ε̂+2
Experiment
type
NA
2
20
2
2
20
20
20
NA
20
2
20
2
20
20
20
20
NA
1.63
12.2
1.63
1.63
12.2
12.2
12.2
NA
12.2
1.63
12.2
1.63
12.2
12.2
12.2
12.2
NA
76.5
49.5
76.5
76.5
49.5
49.5
49.5
NA
49.5
76.5
49.5
76.5
49.5
49.5
49.5
49.5
NA
113
72
113
113
72
72
72
NA
72
113
72
113
72
72
72
72
100
85
60
100
120
78
110
150
100
62∗
103∗
78∗
135∗
105∗
198∗
78
110
0.00
0.20
0.42
0.51
0.72
0.74
0.90
0.96
0.00
0.18
0.26
0.67
0.76
0.88
0.99
0.74
0.90
NA
NA
NA
NA
0.15
0.22
0.73
0.89
NA
NA
NA
0.22
0.23
0.66
0.82
0.22
0.73
D
D
D
D
D
D
D
D
E
E
E
E
E
E
E
R
R
+1
+2
Legend: seed dm,+1
, mobility diameter of sulfate seed particles having +1 charge; seed dm,+1
, mobility diameter of seed particles having +2
charge; selected dm,+1 , mobility diameter classified by the 1×3-TDMA; ε̂+1 , mode value of the organic volume fraction of particles having
+1
+2
sulfate core (cf. Figure A1); ε̂+2 , mode value of organic volume fraction of particles having a dm,+1
core; D, deliquescence test; E,
a dm,+1
efflorescence test; R: deliquescence test of recrystallized particles. An asterisk (*) indicates that the classified particle was on the upper branch
of the water uptake hysteresis curve, and in these experiments the sampling RH was 50% (§2.2.3).
data analysis based on the constant transmission ratio observed
across a broad RH range. Another consideration is that uniform
RH within an instrument, both because of mixing of flows and
more importantly because of minor temperature gradients (i.e.,
1 K corresponds to 6% RH), was not routinely obtained in a
laboratory setting (Weingartner et al. 2002). As a result, the
transmission ratio dropped from unity to zero across a range
of approximately 1–2% RH for the deliquescence of pure inorganic compounds (Rosenoern et al. 2009), even though the
underlying phase transition was expected at a single RH. For
these reasons, herein transmission curves (i.e., transmission ratios for increasing RH) that were as sharp as 2% are taken to
have a single DRH (ERH) value at the relative humidity for
which 50% of the particle population had undergone a phase
transition (i.e., corresponding to a scaled transmission ratio of
0.5). Transmission curves broader than 2% RH are interpreted
in a scientific context, meaning that reasons are sought that a
particle population had a broad phase-transition response, such
as heterogeneity of particles within the population (§3.3).
In addition to practical reasons, scientific reasons can also
cause a deviation from ideal 1×3-TDMA response. For example, in the case of a sufficiently small change in diameter with
a phase change, overlap occurred between the mobility filter of
DMA+δ (DMA−δ ) and the mobility distribution of the phase-
changed particles. In this case, although a phase transition occurred, the particles only partially drifted out of the flow that
exits DMA+δ (DMA−δ ) and that subsequently entered CPC+δ
(CPC−δ ). The transmission ratio therefore did not drop fully to
zero. Examples of this effect are shown in the data sets presented
in this study (§3.3), especially for particles having high organic
volume fractions. Rosenoern et al. (2009) presented equations
that define the minimum growth factor needed for complete
filtering by DMA+δ (DMA−δ ).
2.3. Alternate Modes of 1×3-TDMA Operation
Compared to the original description of the 1×3-TDMA
in Rosenoern et al. (2009), in this study two alternate modes
were developed for data collection by the instrument, including (1) measurement of the transmission ratio for increasing particle diameter and (2) measurement of the hygroscopic
growth factor associated with a phase transition. Measurement
of the transmission ratio for increasing diameter can provide
information about whether a particle population has diameterdependent phase transitions. In this mode, RH0 , RH+δ , and
RH−δ were held constant while the mobility diameter selected
by DMAmono , DMA0 , DMA+δ , and DMA−δ was swept in unison. In the present study, we report data in this mode only for the
Downloaded by [Harvard College] at 13:11 07 September 2012
250
M. L. SMITH ET AL.
deliquescence test. RH0 and RH+δ were set to 60% and 90%,
respectively. In our study, the sweep in particle diameter corresponded to a systematic increase of organic volume fraction.
The control software of the 1×3-TDMA was extended in this
study so that the number-diameter distributions prior to and after
a phase transition were recorded. From these data, the hygroscopic growth factor gwas determined (Swietlicki et al. 2008). In
this mode, rather than being set to pass the same mobility diameter as DMAmono , the voltages of DMA0 and DMA+δ (DMA−δ )
were scanned at a rate of 6 V s−1 and the number concentrations
were recorded by CPC0 and CPC+δ (CPC−δ ). Number-diameter
distributions were constructed by inverting the raw data using
the transfer functions of DMA0 , DMA+δ , or DMA−δ (Wang and
Flagan 1990). The number-diameter distributions corresponded
to particles having undergone RH sequences of RH0 → RH0 →
RH0 for the reference arm and RH0 → RH+δ (RH−δ ) → RH0
for the test arm. By design, the value of RH+δ (RH−δ ) was
sufficient to cause deliquescence (efflorescence).
In the case of deliquescence, the number-diameter distribution recorded for the reference arm corresponded to undeliquesced particles at RH0 and that recorded for the test arm corresponded to deliquesced particles at RH0 . Growth factors (>1)
at RH0 were determined from the increase in the mode diameter
between the two distributions (Biskos et al. 2006b). For the experiments of this study, RH0 and RH+δ values of 60% and 90%,
respectively, were used to determine g(60%). A growth factor at
83% RH was also determined by using the upper-branch distribution at 83% RH and the lower-branch distribution at 60% RH.
In the case of efflorescence, shrinkage factors (<1, i.e., 1/g) at
RH0 were determined from the decrease in the mode diameter
between the two distributions. In the present study, settings of
RH0 and RH−δ held constant at 50% and 20%, respectively,
were used to determine g(50%).
3.
RESULTS AND DISCUSSION
3.1. Phase-Transition Hysteresis in Size Distributions
The diameter change resulting from the phase-transition hysteresis of mixed SOM-AS particles is illustrated for a deliquescence test in Figure 2a and an efflorescence test in Figure 2b. The
number-diameter distributions were obtained by running the instrument in the voltage-scanning mode described in §2.3. As a
reference point in each panel, the transfer function of the DMA
downstream of DMAmono (i.e., DMA0 , DMA+δ , or DMA−δ )
for the normal fixed-voltage mode of 1×3-TDMA operation is
superimposed as a dotted line over the measured distribution.
For the deliquescence test, the normalized number-diameter
distribution at 60% RH is shown in the left panel of Figure 2a for
particles traveling through the reference arm for an RH history
of 60% → 60% → 60%. For this RH sequence, the ammonium sulfate component of the particles remains solid. The size
distribution overlaps with the transfer function of DMA0 , and
the particles therefore pass through DMA0 and are counted by
CPC0 . By comparison, the distribution is shown in the right
FIG. 2. Number-diameter distributions illustrating the hysteresis effect of
phase transitions. (a) Deliquescence test applied to 120-nm diameter (76.5nm seed diameter) solid sulfate-organic particles that are on the lower branch
of the water uptake hysteresis curve. (b) Efflorescence test applied to 78-nm
diameter (49.5-nm seed diameter) sulfate-organic particles that are on the upper
branch of the water uptake hysteresis curve. Number-diameter distributions in
this figure (solid lines) were obtained by running the instrument in the voltagescanning mode (cf. §2.3). Dotted lines represent the DMA transfer functions in
each arm when operated in normal fixed-voltage mode.
panel for an RH history in the test arm of 60% → 90% → 60%.
For this RH sequence, the sulfate component deliquesces at 90%
RH and remains on the upper branch of the hysteresis curve at
the end of sequence at 60% RH. The size distribution is therefore shifted to larger diameters, as shown in the right panel. The
resulting size distribution overlaps to a limited extent with the
transfer function of DMA+δ . Based on this overlap, the transmission ratio CPC+δ :CPC0 is 0.10. For this example, representing
particles of a high organic volume fraction, the shift of the size
distribution after the phase transition is insufficient to move the
distribution fully out of the classification window of DMA+δ .
The right panel of Figure 2a shows that the numberdistribution of the particles on the upper branch of the hysteresis
curve has two modes. For the shown experiment, DMAmono selected 120-nm particles exiting the HEC in experiments that
employed 76.5-nm (+1 charge) and 113-nm (+2 charge) seed
Downloaded by [Harvard College] at 13:11 07 September 2012
PHASE TRANSITIONS OF SOM-AS PARTICLES
particles. The 120-nm particles grown on larger seed particles
have a smaller organic volume fraction and thus a larger hygroscopic growth factor, leading to the second mode of the
distribution.
For the efflorescence test, the number-diameter distribution
at 50% RH is shown in the left panel of Figure 2b for particles traveling through the reference arm for an RH history of
50% → 50% → 50% and therefore remaining on the upper
branch of the hysteresis curve. The distribution shown in the
right panel represents particles passing through the test arm for
an RH history of 50% → 20% → 50%. For this RH sequence,
the sulfate component of the particles effloresces and becomes
solid. Because the particles thus move to the lower branch of
the hysteresis curve, the size distribution shifts to smaller diameters following efflorescence, as shown in the right panel. The
overlap between the transfer function of DMA−δ and the size
distribution of effloresced particles implies a transmission ratio
CPC−δ :CPC0 of 0.23.
3.2. Control Experiments using Ammonium Sulfate
The phase transitions of particles of pure ammonium sulfate
were investigated with the 1×3-TDMA. Times series of RH
profiles and CPC counts for efflorescence and deliquescence
tests are shown in Figure 3. Raw data of this type are discussed
in more detail in Rosenoern et al. (2009). The corresponding
transmission ratios plotted against relative humidity are shown
in Figure 4. The particles deliquesce at 80.6 ± 1% and effloresce at 31 ± 1%. These values agree with measurements in
the literature for aerosol particles as well as with the thermodynamic DRH of 79.5% RH for bulk materials at 298 K (Biskos
et al. 2006c; Rosenoern et al. 2009).The hygroscopic growth factors g(RH) (data not shown) measured from 35 < RH < 80%
also agree with model predictions. For example, the measured
g(60%) is 1.26 ± 0.03 compared to a model prediction of 1.27.
The hygroscopic model is described in Biskos et al. (2006c).
3.3. Deliquescence and Efflorescence of Mixed Particles
The 1×3-TDMA transmission ratios obtained for the application of the deliquescence test to mixed particles of ammonium
sulfate and secondary organic material are shown in Figure 5a.
The transmission ratios with increasing relative humidity are
plotted for eight experiments. Table 2 lists the diameters sampled by the 1×3-TDMA and the corresponding mode value ε̂+1
of the organic volume fraction of experiments 1 to 8, which are
listed in order of a monotonic increase from 0.0 to 0.96 in ε̂+1 .
Figure 5a shows that the DRH shifts from 80.6% to 77.3% RH
across this range of ε̂+1 . The ε̂+1 values are based on the probability density function porg,ss (ε) of organic volume fraction ε
at the diameter sampled by the 1×3-TDMA. The calculation of
porg,ss (ε) and ε̂+1 is described in the Appendix.
The dependence of DRH on organic volume fraction is revealed in Figure 5b, which shows as closed symbols the DRH
(ε̂+1 ) values deduced from Figure 5a. Data are further clas-
251
sified for SOM at organic particle mass concentrations of 1.63
(squares) and at 12.2 µg m−3 (triangles). The data show two features: DRH decreases both for higher organic volume fraction
and for higher particle organic mass concentration. By contrast,
the DRH values have no independent correlation with sampled
particle size (analysis not shown), suggesting the absence of
size-dependent DRH values, at least for particles in the studied
size range of 60 to 150 nm.
The larger decrease of DRH for SOM generated at higher organic particle mass concentrations implies a shift in SOM chemical composition with concentration. In agreement, Shilling et al.
(2009) reported that as organic mass concentration decreases the
organic material becomes more oxygenated. In the data of the
present study (not shown), the AMS-derived O:C ratio increased
from 0.39 at 12.2 µg m−3 to 0.44 at 1.63 µg m−3 . These values
are consistent with those reported previously in Shilling et al.
(2009).
For some experimental conditions, a significant second mode
ε̂+2 resulting from doubly charged seed particles is also sampled
by the 1×3-TDMA (cf. appendix). These values are also listed
in Table 2 and appear as shoulders in Figure 5a. Experiment
7 therein provides a prominent example. Values of DRH(ε̂+2 )
are plotted as open symbols in Figure 5b. The figure shows that
the values of DRH(ε̂+1 ) and DRH(ε̂+2 ) are consistent with each
another.
For a closer investigation of the dependence of DRH on
organic volume fraction, we employed an alternate mode of
1×3-TDMA measurement in which the transmission ratio is
monitored while the sampled particle diameter is increased in
steps (cf. §2.3). The RH profiles in the instrument are held
constant during this sweep. The values of ε̂+1 are determined
for each sampled diameter (cf. Appendix), and a plot of the
measured transmission ratio for increasing ε̂+1 is constructed
(Figure 6). These measurements test experimentally if deliquescence might be fully inhibited above a threshold organic volume
fraction, meaning that the transmission ratio would increase to
1. The data of Figure 6 show that this inhibition does not occur.
Ambiguity in this regard at the highest organic volume fractions
in the plot of Figure 6 is removed by a close examination of
the data for ε̂+1 of 0.96 of experiment 8 displayed in Figure 5a,
which clearly reveals deliquescence.
One feature of Figure 6 is that for ε̂+1 > 0.65 the transmission ratio steadily increases with greater organic volume
fractions. In these cases, the hygroscopic growth of the particle
population is not sufficient to shift the number-diameter distribution fully out of the transfer function of DMA+δ (cf. §3.1),
as is explained by the decreasing diameter difference between
the upper and lower branches of the hysteresis curve for increasing organic volume fraction. The effect is confirmed by
modeling the response expected for the transmission ratio of
the 1×3-TDMA (solid line, Figure 6). The model includes the
transmission function of the instrument (Rosenoern et al. 2009)
and the expected hygroscopic growth of mixed SOM-sulfate
particles (see Equation (1) below). The shaded region around
M. L. SMITH ET AL.
Downloaded by [Harvard College] at 13:11 07 September 2012
252
FIG. 3. Example of raw data recorded by the 1×3-TDMA for particles of ammonium sulfate. Columns a and b correspond to efflorescence and deliquescence
tests, respectively. Rows 1 and 2 show time series of relative humidity and CPC counts, respectively, in both the reference and test arms. The dotted lines are
drawn at the locations in the data sets at which phase transitions occur. For the shown measurements, the mobility diameter (+1 charge) of the selected particles is
100 nm.
the solid line in Figure 6 represents the model response to ±1 nm
uncertainty in the classification by the DMAs. The conclusions
from the good agreement between the model and the observations are that deliquescence occurs across the entire range of
organic volume fractions, that the increase in transmission ratio
for ε̂+1 > 0.65 is explained by decreased hygroscopic growth
of the mixed particle, that full inhibition of deliquescence is not
observed even to the highest organic volume fractions, and that
the DRH values recorded are consistent with a thermodynamic
effect rather than a kinetic effect of SOM coatings.
As an elaboration of the final point, the hypothesis of a kinetic effect, which is ruled out by the data of Figure 6, is that
deliquescence would have occurred at lower RH values if the
exposure time had been longer, with the associated idea that the
limitation in observing deliquescence was the rate of H2 O diffusion through an organic coating. This hypothesis is ruled out
Downloaded by [Harvard College] at 13:11 07 September 2012
PHASE TRANSITIONS OF SOM-AS PARTICLES
253
FIG. 4. The transmission ratios CPC−δ :CPC0 (left line; efflorescence test) and
CPC+δ :CPC0 (right line; deliquescence test) for increasing relative humidity.
The abscissa is RH−δ for the efflorescence test and RH+δ for the deliquescence
test, corresponding to the data displayed in row 1 of Figure 3. The ordinate of
Figure 4 is calculated from the ratio of the data displayed in row 2 of Figure 3.
The transmission ratio drops to zero at the ERH of ammonium sulfate in the
efflorescence test and at the DRH in the deliquescence test.
by the good agreement between the hygroscopic growth model
(i.e., solid line of Figure 6) and the collected data: the functional
dependence of a kinetic effect on water uptake with increasing
organic volume fraction would be expected to differ from that
of the hygroscopic growth model. More specifically, a kinetic
effect, if present, would have increased the transmission ratio
at a lower organic volume fraction than the upper-limit value
imposed by the hygroscopic growth model and observed in the
data. A related complementary observation is that a kinetic effect of a non-water-permeable organic coating that would limit
deliquescence (even though the data of Figure 6 argue against
this effect) would likewise inhibit efflorescence. The data of Figure 7, however, show facile efflorescence, even to high organic
volume fractions.
In regard to efflorescence, Figure 7a shows the transmission
ratios recorded for seven experimental conditions that represent
ε̂ in monotonic order from 0.0 to 0.99 (Table 2). Across this
range, ERH decreases slightly from 30.8% to 29.0% RH. Figure 7a shows that the transmission ratio does not fall to zero for
high values of ε̂. The explanation is that the growth factors at
50% RH are not sufficient to move the transmitted particles fully
out of the transfer function of DMA−δ (cf. §3.1), as explained
above for Figure 6 for high values of ε̂ in the deliquescence
test. In experiment 15 of Figure 7a, for example, the particles
selected for analysis are composed almost entirely of SOM (i.e.,
ε̂+1 = 0.99). According to the model, approximately 5% of the
effloresced particles shift out of the DMA−δ transfer function.
The resulting predicted transmission ratio of 0.95 is in agreement within uncertainty with the observed transmission ratio of
0.97.
FIG. 5. Observations for deliquescence tests. (A) Transmission ratios measured for increasing RH+δ and constant +δ of 5%. The transmission ratios are
scaled to 1, with scaling factors ranging from 0.85 to 1.12 (not shown). Eight
experimental conditions of selected mobility diameter dm,+1 and corresponding
mode organic volume fraction ε̂+1 are investigated. Table 2 lists the exact values for each experiment. In some experiments, a significant fraction of particles
selected at dm,+1 is derived from larger sulfate seed particles (i.e., derived from
doubly charged particles in the seed distribution that became coated with organic material in the chamber). In these cases, the transmission ratio decreases
stepwise (e.g., experiments 5 and 7). For these cases, the calculated ε̂+2 are also
listed in Table 2. (B) Deliquescence relative humidity for increasing organic
volume fraction. Results are shown for pure ammonium sulfate (diamond) as
well as for mixed particles prepared at organic particle mass concentrations
of 1.63 (squares) and 12.2 µg m–3 (triangles). Mode organic volume fractions
derived from singly charged seed particles (ε̂+1 ; closed symbols) as well as doubly charged seed particles (ε̂+2 ; open symbols) are shown. The box represents
the quartiles of the probability density function porg,ss (ε) of organic volume
fraction, with the count median ε̄¯ marked by the vertical bar.
The summary results of ERH(ε̂+1 ) are presented in Figure 7b.
These results show that, as for DRH(ε̂) in Figure 5b, ERH(ε̂) of
Figure 7b depends on both organic particle mass concentration
and organic volume fraction, with lower ERH at higher organic
volume fraction and higher organic particle mass concentration.
In the case of efflorescence, the absence of shoulders in the
data of Figure 7a, in contrast to their presence in the data for
deliquescence of Figure 5a, precludes the possibility to plot
Downloaded by [Harvard College] at 13:11 07 September 2012
254
M. L. SMITH ET AL.
FIG. 6. Plot of transmission ratio with increasing organic volume fraction.
In this experiment, RH0 and RH+δ are held constant at 60 and 90%, respectively, as dm,+1 is incrementally increased. The value on the abscissa represents
ε̂+1 . Results are shown for organic particle mass concentrations of 1.63 (open
squares) and 12.2 µg m–3 (closed squares). The model (black line) predicts
the transmission ratio assuming that all particles fully deliquesce and that their
growth factors follow Equation (1). The shaded region around the line represents
the range of modeled transmission ratios, taking into account an uncertainty of
±1 nm in the DMA particle sizing.
both ERH(ε̂+1 ) and ERH(ε̂+2 ) within Figure 7b. For experiment 13, the second mode ε̂+2 resulting from doubly charged
seed particles is slightly larger than ε̂+1 , and this point is correspondingly plotted as ERH(ε̂+2 ) (open symbol). Given the
greater concentration of singly charged seed particles for all
other efflorescence experiments, the remaining points in Figure
7b are plotted as ERH(ε̂+1 ), with the implied assumption that
differences between ERH(ε̂+1 ) and ERH(ε̂+2 ) are sufficiently
small that they are not resolved by the 1×3-TDMA. This conclusion of an embedded but unresolved effect of ERH(ε̂+1 ) and
ERH(ε̂+2 ) is supported by the broad response of the transmission curve across 3.7% RH that is apparent in Figure 7a for
experiment 13. This experiment also has the widest distribution
of ε values among experiments 9 to 15 (Figure 7b, box sizes).
For comparison, the breadth of the response of the transmission
curves varies from 2.1 and 2.8% RH for the other efflorescence
tests, which correspondingly have narrower distributions of
ε values
3.4. Recrystallization
Morphology can influence phase transitions, as shown, for
example, in the case of a less hygroscopic salt surrounding a
more hygroscopic one (Mifflin et al. 2009). Cycling relative
humidity from low to medium back to low values can lead
to changes in the shapes of hygroscopic particles as they adjust from high-energy morphologies associated with their initial
synthesis and locked into place at low RH toward low-energy
euhedral shapes that can be reached when adsorbed water layers increase ionic mobility (Mikhailov et al. 2004; Biskos et al.
FIG. 7. Observations for efflorescence tests. (a) Transmission ratios measured for decreasing RH−δ . Shown data correspond to a –δ of 10 to 30%. The
transmission ratios are scaled to 1, with scaling factors ranging from 0.88 to
1.05 (not shown). The selected mobility diameter dm,+1 and the corresponding
mode organic volume fraction ε̂+1 for experiments 9 to 15 are listed in Table
2. (b) Efflorescence relative humidity for increasing organic volume fraction.
Results are shown for pure ammonium sulfate (diamonds) as well as for mixed
particles prepared at organic particle mass concentrations of 1.63 (squares) and
12.2 µg m–3 (triangles). Mode organic volume fractions derived from either
singly charged seed particles (ε̂+1 ; closed symbols) or doubly charged seed
particles (ε̂+2 ; open symbols) are shown. The box represents the quartiles of
porg,ss (ε), with the count median ε̄¯ marked by the vertical bar. Stars represent
data of Takahama et al. (2007).
2006a; Rosenoern et al. 2008). Cycling relative humidity can
also alter the morphology and deliquescence behavior of some
but not all organic-coated inorganic particles (Chan et al. 2006;
Chan and Chan 2007; Ciobanu et al. 2009). In some cases, the
different effects can depend on the ability of water to permeate
an outer coating (e.g., Mifflin et al. 2009).
Although no imaging was conducted of the particles exiting
the HEC, a particle morphology is expected of a solid sulfate
core surrounded by an organic coating because the mixed particles were prepared by the condensation of SOM molecules onto
sulfate seed particles. Other morphologies are certainly possible under some conditions, such as an incomplete coating of
SOM across the surface of the ammonium sulfate that might
Downloaded by [Harvard College] at 13:11 07 September 2012
PHASE TRANSITIONS OF SOM-AS PARTICLES
be represented by a partially engulfing morphology (Kwamena
et al. 2010). Chan et al. (2006) observed that glutaric acid coats
supermicron AS particles in a morphology of towers of glutaric
acid serving as a shell on a sulfate core. Nevertheless, the density measurements carried out by us with the APM showing a
diameter-independent SOM density, as well as the agreement
between the AMS- and SMPS-derived mass concentrations, are
all consistent with a spherical particle. Furthermore, in our experiments the AMS-measured sulfate increased substantially as
soon as SOM condensation began, indicative of a liquid-phase
coating of SOM that eliminates particle bounce in the AMS
(Bahreini et al. 2005; Virtanen et al. 2010). These results are
all consistent with a particle morphology of a solid sulfate core
surrounded by an organic coating for the particles exiting the
HEC.
The influence of possible differences in particle morphology,
such as a morphology resulting from condensational growth
in the HEC compared to a perhaps different morphology that
might result after interaction with water and efflorescence during atmospheric aging, was investigated by recrystallizing the
mixed particles (i.e., deliquescing and re-drying them) prior to
sampling into the 1×3-TDMA. Results of these experiments
are shown in Figure 8, in which the RH-dependent transmission ratios are plotted both for particles sampled directly from
the chamber and for recrystallized particles. The experiments
investigated two organic volume fractions, specifically experiments 6, 7, 16, and 17 of Table 2. The results plotted in
Figure 8 show that the RH-dependent transmission ratio for
directly sampled compared to recrystallized particles are in
agreement with each other within the RH uncertainty (±1%)
of the 1×3-TDMA. The conclusion is either that the particle morphologies are similar between the two sets of experiments or that any differences in morphology do not influence
deliquescence.
3.5. Growth Factor Measurements
Growth factors g were obtained from data such as those
shown in Figure 2 as the ratio of the mode diameter in the
right panel of Figure 2 to that in the left panel. The values
of g(50%), g(60%), and g(83%) are plotted in Figure 9 for
increasing organic volume fraction. The data show that, within
experimental uncertainty, the growth factors do not depend on
the organic particle mass concentration in the chamber (i.e.,
squares and triangles), in contrast to the observations for DRH
(Figure 5) and ERH (Figure 7). The implication is that DRH and
ERH are more sensitive to changes in the chemical composition
of the SOM than is g(RH).
The hygroscopic growth factor of mixed particles can be
modeled based on volume averaging of the growth factors of
the end members. The equation is as follows:
1
g(RH )predicted = (1 − ε) g(RH )3AS + ε g(RH)3SOM 3
[1]
255
FIG. 8. Demonstration that recrystallization does not influence deliquescence. The transmission ratio of recrystallized particles, corresponding to experiments 16 and 17 of Table 2 for ε̂+1 of 0.74 and 0.90, agree within an
experimental uncertainty of ±1% RH with the transmission ratio of particles
extracted directly from the chamber (i.e., experiments 6 and 7).
This equation applies equally to two physical cases: (1) the
mixed particle is composed of separate and non-interacting
phases (i.e., phase-separated particle) or (2) the mixed particle is composed of a single phase in which any solute-solute
interactions in the mixed solution are either negligible with regard to water uptake or alternatively cancel out with one another
(Stokes and Robinson 1966).
In use of Equation (1), we take growth factors g(RH )AS of
pure AS from Biskos et al. (2006c). Growth factors g(RH)SOM of
FIG. 9. Dependence of growth factor on organic volume fraction. Results are
shown for 50%, 60%, and 83% RH. The value on the abscissa represents ε̂+1 .
Squares and triangles represent organic particle mass concentrations of 1.63 and
12.2 µg m–3 , respectively. Circles represent pure ammonium sulfate. Vertical
bars show the uncertainty of the measurement of g(RH). For comparison to the
data points, model results using Equation (1) are shown as shaded regions, with
the uncertainty based on upper and lower limits of the hygroscopic response of
pure SOM reported in the literature. The main text provides a further description.
Downloaded by [Harvard College] at 13:11 07 September 2012
256
M. L. SMITH ET AL.
SOM produced by α-pinene ozonolysis products are taken both
from Saathoff et al. (2003) and in Varutbangkul et al. (2006) for
comparison. Saathoff et al. measured the hygroscopic response
of SOM prepared in the absence of seed particles. Varutbangkul
et al. measured the response of seeded but nearly pure SOM
(ε = 0.996) and corrected for the small contribution by ammonium sulfate to determine g(RH)SOM (i.e., an inversion analysis
based on Equation (1)). The shaded regions in Figure 9 represent the range of values modeled using the two different formulations of g(RH)SOM , with the upper side corresponding to that
of Saathoff et al., and the lower to that of Varutbangkul et al.
Examination of Figure 9 shows that the predicted and observed growth factors are in satisfactory agreement. The conclusion is that there is only limited interaction between the SOM
and AS materials and one of the two physical cases mentioned in
regard to Equation (1) is indicated. That said, there must be, however, at least some non-negligible interactions given the existent
small shifts in DRH and ERH. Another implication of agreement between the observed growth factors and g(RH )predicted
is that the particles on both sides of the water uptake hysteresis curve are spherical (i.e., Equation (1) implicitly assumes
spherical particles) (Biskos et al. 2006b).
A cautionary note in the use of g(RH)SOM of Saathoff et al.
and Varutbangkul et al. for our application is the difference in
organic particle mass concentrations, specifically 65 and 199 µg
m−3 in those respective studies compared to 1.63 and 12.2 µg
m−3 in the present study. Across this range of concentrations,
there are significant differences in composition, with elemental
O:C shifting from 0.45 at 1 µg m−3 to 0.33 at 40 µg m−3
(Shilling et al. 2009). Although we did not observe differences
in hygroscopic behavior between 1.63 and 12.2 µg m−3 , the
shift in composition accompanying the shift to the higher mass
concentrations of Saathoff et al. and Varutbangkul et al. may also
affect g(RH)SOM to some extent. This aspect is a recommended
focus of future studies.
4.
CONCLUSIONS
The experimental findings are summarized as follows: (1)
the sulfate component of mixed SOM-AS particles had DRH
and ERH values within 4% RH of the reference values of pure
ammonium sulfate up to the highest organic volume fractions
investigated, specifically up to 0.96 for DRH and 0.99 for ERH;
(2) the DRH and ERH values of the sulfate component decreased approximately linearly for increasing organic volume
fraction; (3) the DRH and ERH values were consistently lower
for particles having an SOM component that was generated at
higher organic particle mass concentration (i.e., 1.63 compared
to 12.2 µg m−3 ); (4) the recrystallization of the mixed particles
did not affect the DRH behavior of the sulfate component; and
(5) the hygroscopic growth factor of the mixed particles was
described within measurement uncertainty by a non-interacting
volume average of the growth factors of pure AS and pure SOM.
These findings taken together indicate that the overall chem-
ical interactions between SOM generated by α-pinene dark
ozonolysis and ammonium sulfate, both in the presence and
absence of water, are weak over the range of relative humidities studied (20–85%). Within the mixed particles, in regard
to deliquescence, efflorescence, and hygroscopic growth factor,
the ammonium sulfate behaves largely as it does in its pure
state. The SOM acts as a separate material with very limited
synergistic behavior with AS, namely with respect to the small
decreases in the DRH and the ERH. These results are consistent
with a model of phase separation between the organic and inorganic materials within a single particle, both in the presence
and absence of water (Chou et al. 1998; Marcolli and Krieger
2006; Ciobanu et al. 2009). With respect to the arguments in
favor of a biphasic particle, the weak effect of SOM on the
ERH of water-containing particles is especially noteworthy. For
a counter proposal of a uniphasic particle, the high concentration
of SOM molecules would strongly interfere with nucleation of
AS crystals from the ionic medium (Marcolli and Krieger 2006;
Parsons et al. 2004, 2006), thereby eliminating efflorescence, in
contradistinction to the observations.
On the lower side of the hysteresis loop, the biphasic particle
consists of a first phase that is SOM material and a second that is
solid AS. The biphasic particle on the upper side of the hysteresis
loop is more complicated. The first phase is largely composed
of ammonium sulfate and water, and the second phase consists
mostly of SOM and smaller amounts of water. According to
the thermodynamic principles underlying liquid-liquid phase
diagrams, these two phases are not entirely pure with respect to
one another. Small amounts of SOM should be present in the
water-sulfate phase to explain the observed decreases in DRH
and ERH.
Proposed phase-separated chemical morphologies for both
crystalline and deliquesced AS are illustrated in Figure 10.
While Figure 10 shows a specific example of a largely inorganic salt phase engulfed by a largely SOM phase, any phaseseparated spherical morphology is consistent with our results.
Additional possible particle configurations are shown in Figure S1. The actual morphology can be expected to vary with
a dependence on the organic volume fraction, the particle size,
and the relative surface tensions of the two phases against air as
well against each other (Kwamena et al. 2010).
Some possibilities that might complicate the interpretation
of the results, such as a diameter-dependent composition of the
SOM or diameter-dependent nano-size effects on phase transitions, can be ruled out. The variability of parameters in Table 2
is sufficient such that linear trends would not be observed in Figures 5b and 7b if diameter were more important than organic volume fraction in explaining the DRH and ERH results. Furthermore, in regard to ruling out a diameter-dependent composition,
Shilling et al. (2009) reported that the composition of SOM
prepared in the HEC by a similar protocol of α-pinene dark
ozonolysis was independent of diameter, as expected for particle growth by uniform condensation. The AMS data recorded
in the present study are consistent with that earlier result (data
Downloaded by [Harvard College] at 13:11 07 September 2012
PHASE TRANSITIONS OF SOM-AS PARTICLES
257
FIG. 10. Proposed phase-separated morphologies for particles having SOM, AS, and H2 O as components and an organic volume fraction of 0.5. (Left panel)
Illustrations of phase-separated compared to solid- or liquid-solution morphologies. The illustrations depict particles on the lower branch of hysteresis curve as
they exit the HEC (i.e., prior to a deliquescence test using the 1×3-TDMA). (Right panel) Illustrations of phase-separated compared to solid- or liquid-solution
morphologies for particles on the upper branch of hysteresis curve at 35 and 80% RH. The data sets of DRH, ERH, and growth factors rule out a solution-phase
chemical morphology, as emphasized by the black crosses over these illustrations in both the left and right panels.
not shown). In regard to nano-size effects, according to this
hypothesis, because lower organic volume fractions in the conducted experiments correspond to smaller particle diameters and
because nano-sized particles can have increased DRH and ERH
values (Biskos et al. 2006a), the result showing that DRH and
ERH are greater for lower organic volume fractions might arise
from a nano-size effect. The smallest diameter of the present
study, however, was 60 nm, which can be compared to a size
of 6 nm at which the DRH and ERH of ammonium sulfate are
equal to those of their bulk counterparts (Biskos et al. 2006c).
Even for NaCl nanoparticles, which have a much stronger nanosize effect, the DRH and ERH values agree with bulk values
for particles of 40 nm and larger (Biskos et al. 2006a). Thus,
the study-minimum diameter of 60 nm is much larger than the
known threshold for nanosize effects.
The result that SOM does not significantly shift the DRH
and ERH of ammonium sulfate even up to organic volume fractions of greater than 0.95, at least for SOM prepared by the
dark ozonolysis of α-pinene, causes a new picture to emerge
of the effects of organic material on the phase transitions of
atmospheric particles. As mentioned in the introduction, the
presence of tens to hundreds of different organic molecules in
SOM has previously led to a prevailing hypothesis that the mul-
tiple compounds in SOM should combine to reduce the ERH of
mixed SOM-inorganic particles and eventually eliminate crystallization entirely. The present study’s results across a wide
range of organic volume fractions indicate that this hypothesis
is incorrect, at least for the SOM generated from ozonolysis of
α-pinene. Furthermore, the prevailing hypothesis in regard to
DRH, which is based on a water-soluble classification that is inferred both from the cloud-condensation-nuclei (CCN) activity
and the hygroscopic growth of SOM, has been that there should
be a considerable influence on the deliquescence properties of
inorganic salts. The results of the present study again reinforce
that this hypothesis on the effect on DRH is incorrect for the
studied SOM.
Our work on CCN activity concluded that SOM generated
under similar experimental conditions was fully water soluble,
with the implication of a single aqueous phase of AS and SOM
(King et al. 2007, 2009). Our results reported herein, however,
indicate that SOM and AS remain largely phase separated in
the presence of condensed-phase H2 O. These apparently contradictory observations are reconciled once the differences in
water activity between the studies are taken into account. At
the point of CCN activation, SOM, AS, and H2 O are present
as a molecular uniphase solution at water activities near 1 (i.e.,
Downloaded by [Harvard College] at 13:11 07 September 2012
258
M. L. SMITH ET AL.
the water activities typically required for CCN activation). At
lower water activities (e.g., <0.85 in the present study), there
is insufficient water for complete shells of solvation to form,
and as a result inorganic and organic phases (both containing
some water) are thermodynamically predicted to form below a
separation relative humidity (SRH) (Pankow 2003; Zuend et al.
2010).
This conceptual difference of SOM solubility with a shift in
water activity is chemically reasonable. Oxalic acid, for example, has a saturation water activity of 0.93 (Brooks et al. 2002),
implying that in the CCN regime this material is fully water soluble and miscible with an inorganic aqueous phase. By contrast,
in a hygroscopic growth regime below 0.93 water activity, this
material in pure form is fully water insoluble (unless metastable)
and, in mixed solution with ammonium sulfate, phase separates
between the initial and final DRH values. It is fully phase separated at the eutonic RH. The conclusion is that understanding
the behavior of SOM, AS, and H2 O in concentrated solutions
(i.e., 20 to 85% RH of the present study) has limited application
to the regime of CCN activation.
The present study’s findings of limited chemical interactions
between SOM generated by α-pinene dark ozonolysis and ammonium sulfate in the presence of water for 20% < RH <
85% could simplify the treatment of the phase changes of atmospheric mixed particles within global models of climate and
atmospheric chemistry. The O:C ratios of 0.39 and 0.44 measured in this study for SOM are in the middle of the range of
0.20 to 0.89 reported for ambient particles (Ng et al. 2010). The
deviations in DRH and ERH with increasing organic volume
fraction are less than 4% with respect to the reference values of
AS, implying that the phase transitions of internally mixed particles of SOM and AS can be modeled to first order as occurring
at the DRH and ERH of pure AS. Our results taken together with
existing literature, however, show that not all SOM exhibits the
same behavior. Further work is needed to quantify the effects
of secondary organic material as a function of composition by
varying the VOC precursor, the dominant oxidation pathway,
and the organic particle mass concentration.
REFERENCES
Bahreini, R., Keywood, M. D., Ng, N. L., Varutbangkul, V., Gao, S., Flagan,
R. C., Seinfeld, J. H., Worsnop, D. R., and Jimenez, J. L. (2005). Measurements of Secondary Organic Aerosol from Oxidation of Cycloalkenes,
Terpenes, and m-Xylene Using an Aerodyne Aerosol Mass Spectrometer.
Environ. Sci. Technol. 39:5674–5688.
Baltensperger, U., Kalberer, M., Dommen, J., Paulsen, D., Alfarra, M. R., Coe,
H., Fisseha, R., Gascho, A., Gysel, M., Nyeki, S., Sax, M., Steinbacher,
M., Prevot, A. S. H., Sjogren, S., Weingartner, E., and Zenobi, R. (2005).
Secondary Organic Aerosols from Anthropogenic and Biogenic Precursors.
Faraday Discuss. 130:265–278.
Biskos, G., Malinowski, A., Russell, L. M., Buseck, P. R., and Martin, S. T.
(2006a). Nanosize Effect on the Deliquescence and the Efflorescence of
Sodium Chloride Particles. Aerosol Sci. Technol. 40:97–106.
Biskos, G., Russell, L. M., Buseck, P. R., and Martin, S. T. (2006b). Nanosize
Effect on the Hygroscopic Growth Factor of Aerosol Particles. Geophys. Res.
Lett. 33:L07801.
Biskos, G., Paulsen, D., Russell, L. M., Buseck, P. R., and Martin, S. T. (2006c).
Prompt Deliquescence and Efflorescence of Aerosol Nanoparticles. Atmos.
Chem. Phys. 6:4633–4642.
Brooks, S. D., Wise, M. E., Cushing, M., and Tolbert, M. A. (2002). Deliquescence Bhavior of Organic/Ammonium Sulfate Aerosol. Geophys. Res. Lett.
29:1917.
Brooks, S. D., Garland, R. M., Wise, M. E., Prenni, A. J., Cushing, M., Hewitt,
E., and Tolbert, M. A. (2003). Phase Changes in Internally Mixed Maleic
Acid/Ammonium Sulfate Aerosols. J. Geophys. Res. 108:4487.
Chan, M. N., Lee, A. K. Y., and Chan, C. K. (2006). Responses of Ammonium
Sulfate Particles Coated with Glutaric Acid to Cyclic Changes in Relative Humidity: Hygroscopicity and Raman Characterization. Environ. Sci. Technol.
40:6983–6989.
Chan, M. N., and Chan, C. K. (2007). Mass Transfer Effects on the Hygroscopic
Growth of Ammonium Sulfate Particles with a Water-Insoluble Coating.
Atmos. Environ. 41:4423–4433.
Chhabra, P. S., Flagan, R. C., and Seinfeld, J. H. (2010). Elemental Analysis
of Chamber Organic Aerosol Using an Aerodyne High-Resolution Aerosol
Mass Spectrometer. Atmos. Chem. Phys. 10:4111–4131.
Choi, M. Y., and Chan, C. K. (2002). The Effects of Organic Species on the Hygroscopic Behaviors of Inorganic Aerosols. Environ. Sci. Technol. 36:2422–
2428.
Chou, T. J., Tanioka, A., and Tseng, H. C. (1998). Salting Effect on the LiquidLiquid Equilibria for the Partially Miscible Systems of n-Propanol-Water and
i-Propanol-Water. Ind. Eng. Chem. Res. 37:2039–2044.
Ciobanu, V. G., Marcolli, C., Krieger, U. K., Weers, U., and Peter, T. (2009).
Liquid-Liquid Phase Separation in Mixed Organic/Inorganic Aerosol Particles. J. Phys. Chem. A. 113:10966–10978.
Cocker, D. R., Clegg, S. L., Flagan, R. C., and Seinfeld, J. H. (2001).
The Effect of Water on Gas-Particle Partitioning of Secondary Organic
Aerosol. Part 1: Apha-Pinene/Ozone System. Atmos. Environ. 35:6049–
6072.
Cruz, C. N., and Pandis, S. N. (2000). Deliquescence and Hygroscopic Growth
of Mixed Inorganic-Organic Atmospheric Aerosol. Environ. Sci. Technol.
34:4313–4319.
Cziczo, D. J., Nowak, J. B., Hu, J. H., and Abbatt, J. P. D. (1997). Infrared
Spectroscopy of Model Tropospheric Aerosols as a Function of Relative
Humidity: Observation of Deliquescence and Crystallization. J. Geophys.
Res. 102:18843–18850.
DeCarlo, P. F., Kimmel, J. R., Trimborn, A., Northway, M. J., Jayne, J. T.,
Aiken, A. C., Gonin, M., Fuhrer, K., Horvath, T., Docherty, K. S., Worsnop,
D. R., and Jimenez, J. L. (2006). Field-Deployable, High-Resolution, Timeof-Flight Aerosol Mass Spectrometer. Anal. Chem. 78:8281–8289.
Donahue, N. M., Robinson, A. L., Stanier, C. O., and Pandis, S. N. (2006). Coupled Partitioning, Dilution, and Chemical Aging of Semivolatile Organics.
Environ. Sci. Technol. 40:2635–2643.
Ehara, K., Hagwood, C., and Coakley, K. J. (1996). Novel Method to Classify
Aerosol Particles According to Their Mass-to-Charge Ratio: Aerosol Particle
Mass Analyser. J. Aerosol Sci. 27:217–234.
Ha, Z. Y., and Chan, C. K. (1999). The Water Activities of MgCl2 ,
Mg(NO3 )2 , MgSO4, and Their Mixtures. Aerosol Sci. Technol. 31:154–
169.
Hallquist, M., Wenger, J. C., Baltensperger, U., Rudich, Y., Simpson, D., Claeys,
M., Dommen, J., Donahue, N. M., George, C., Goldstein, A. H., Hamilton,
J. F., Herrmann, H., Hoffmann, T., Iinuma, Y., Jang, M., Jenkin, M. E.,
Jimenez, J. L., Kiendler-Scharr, A., Maenhaut, W., McFiggans, G., Mentel,
T. F., Monod, A., Prevot, A. S. H., Seinfeld, J. H., Surratt, J. D., Szmigielski,
R., and Wildt, J. (2009). The Formation, Properties and Impact of Secondary
Organic Aerosol: Current and Emerging Issues. Atmos. Chem. Phys. 9:5155–
5236.
Hameri, K., Vakeva, M., Hansson, H. C., and Laaksonen, A. (2000). Hygroscopic Growth of Ultrafine Ammonium Sulphate Aerosol Measured Using an Ultrafine Tandem Differential Mobility Analyzer. J. Geophys. Res.
105:22231–22242.
Downloaded by [Harvard College] at 13:11 07 September 2012
PHASE TRANSITIONS OF SOM-AS PARTICLES
Hameri, K., Laaksonen, A., Vakeva, M., and Suni, T. (2001). Hygroscopic
Growth of Ultrafine Sodium Chloride Particles. J. Geophys. Res. 106:20749–
20757.
Hamilton, J. F., Webb, P. J., Lewis, A. C., Hopkins, J. R., Smith, S., and Davy,
P. (2004). Partially Oxidised Organic Components in Urban Aerosol Using
GCXGC-TOF/MS. Atmos. Chem. Phys. 4:1279–1290.
Hanel, G. (1976). The Properties of Atmospheric Aerosol Particles as Functions
of the Relative Humidity at Thermodynamic Equilibrium with the Surrounding Moist Air. Adv. Geophys. 19:73–188.
King, S. M., Rosenoern, T., Shilling, J. E., Chen, Q., and Martin, S. T. (2007).
Cloud Condensation Nucleus Activity of Secondary Organic Aerosol Particles Mixed with Sulfate. Geophys. Res. Lett. 34:L24806.
King, S. M., Rosenoern, T., Shilling, J. E., Chen, Q., and Martin, S. T. (2009).
Increased Cloud Activation Potential of Secondary Organic Aerosol for Atmospheric MassLoadings. Atmos. Chem. Phys. 9:2959–2971.
King, S. M., Rosenoern, T., Shilling, J. E., Chen, Q., Wang, Z., Biskos, G.,
McKinney, K. A., Poschl, U., and Martin, S. T. (2010). Cloud Droplet Activation of Mixed Organic-Sulfate Particles Produced by the Photooxidation
of Isoprene. Atmos. Chem. Phys. 10:3953–3964.
Kleindienst, T. E., Smith, D. F., Li, W., Edney, E. O., Driscoll, D. J., Speer, R. E.,
and Weathers, W. S. (1999). Secondary Organic Aerosol Formation From the
Oxidation of Aromatic Hydrocarbons in the Presence of Dry Submicron
Ammonium Sulfate Aerosol. Atmos. Environ. 33:3669–3681.
Kuwata, M., and Kondo, Y. (2009). Measurements of Particle Masses of Inorganic Salt Particles for Calibration of Cloud Condensation Nuclei Counters.
Atmos. Chem. Phys. 9:5921–5932.
Kwamena, N. -O. A, Buajarern, J., and Reid, J. P. (2010). Equilibrium Morphology of Mixed Organic/Inorganic/Aqueous Aerosol Droplets: Investigating the
Effect of Relative Humidity and Surfactants. J. Phys. Chem. A. 114:5787–
5795.
Malloy, Q. G. J., Nakao, S., Qi, L., Austin, R., Stothers, C., Hagino, H., and
Cocker, D. R. (2009). Real-Time Aerosol Density Determination Utilizing a
Modified Scanning Mobility Particle Sizer Aerosol Particle Mass Analyzer
System. Aerosol Sci. Technol. 43:673–678.
Marcolli, C., Luo, B. P., and Peter, T. (2004). Mixing of the Organic Aerosol
Fractions: Liquids as the Thermodynamically Stable Phases. J. Phys. Chem.
A. 108:2216–2224.
Marcolli, C., and Krieger, U. K. (2006). Phase Changes During Hygroscopic Cycles of Mixed Organic/Inorganic Model Systems of Tropospheric Aerosols.
J. Phys. Chem. A. 110:1881–1893.
Martin, S. T. (2000). Phase Transitions of Aqueous Atmospheric Particles.
Chem. Rev. 100:3403–3453.
Martin, S. T., Schlenker, J. C., Malinowski, A., Hung, H. M., and Rudich, Y.
(2003). Crystallization of Atmospheric Sulfate-Nitrate-Ammonium Particles.
Geophys. Res. Lett. 30:2102.
Martin, S. T., Hung, H. M., Park, R. J., Jacob, D. J., Spurr, R. J. D., Chance,
K. V., and Chin, M. (2004). Effects of the Physical State of Tropospheric
Ammonium-Sulfate-Nitrate Particles on Global Aerosol Direct Radiative
Forcing. Atmos. Chem. Phys. 4:183–214.
Martin, S. T., Rosenoern, T., Chen, Q., and Collins, D. R. (2008). Phase Changes
of Ambient Particles in the Southern Great Plains of Oklahoma. Geophys.
Res. Lett. 35: L22801.
Meyer, N. K., Duplissy, J., Gysel, M., Metzger, A., Dommen, J., Weingartner,
E., Alfarra, M. R., Prevot, A. S. H., Fletcher, C., Good, N., McFiggans, G.,
Jonsson, A. M., Hallquist, M., Baltensperger, U., and Ristovski, Z. D. (2009).
Analysis of the Hygroscopic and Volatile Properties of Ammonium Sulphate
Seeded and Unseeded SOA Particles. Atmos. Chem. Phys. 9:721–732.
Mifflin, A. L., Smith, M. L., and Martin, S. T. (2009). Morphology Hypothesized to Influence Aerosol Particle Deliquescence. Phys. Chem. Chem. Phys.
11:10095–10107.
Mikhailov, E., Vlasenko, S., Niessner, R., and Poschl, U. (2004). Interaction
of Aerosol Particles Composed of Protein and Salts with Water Vapor: Hygroscopic Growth and Microstructural Rearrangement. Atmos. Chem. Phys.
4:323–350.
259
Ng, N. L., Canagaratna, M. R., Zhang, Q., Jimenez, J. L., Tian, J., Ulbrich,
I. M., Kroll, J. H., Docherty, K. S., Chhabra, P. S., Bahreini, R., Murphy,
S. M., Seinfeld, J. H., Hildebrandt, L., Donahue, N. M., DeCarlo, P. F., Lanz,
V. A., Prevot, A. S. H., Dinar, E., Rudich, Y., and Worsnop, D. R. (2010).
Organic Aerosol Observed in Northern Hemispheric Datasets from Aerosol
Mass Spectrometry. Atmos. Chem. Phys. 10:4625–4641.
Onasch, T. B., Siefert, R. L., Brooks, S. D., Prenni, A. J., Murray, B., Wilson,
M. A., and Tolbert, M. A. (1999). Infrared Spectroscopic Study of the Deliquescence and Efflorescence of Ammonium Sulfate Aerosol as a Function of
Temperature. J. Geophys. Res. 104:21317–21326.
Pankow, J. F. (2003). Gas/Particle Partioning of Neutral and Ionizing Compounds to Single and Multi-Phase Aerosol Particles. Part 1: Unified Modeling
Framework. Atmos. Environ. 37:3323–3333.
Parsons, M. T., Knopf, D. A., and Bertram, A. K. (2004). Deliquescence and
Crystallization of Ammonium Sulfate Particles Internally Mixed with WaterSoluble Organic Compounds. J. Phys. Chem. A. 108:11600–11608.
Parsons, M. T., Riffell, J. L., and Bertram, A. K. (2006). Crystallization of
Aqueous Inorganic-Malonic Acid Particles: Nucleation Rates, Dependence
on Size, and Dependence on the Ammonium-to-Sulfate. J. Phys. Chem. A.
110:8108–8115.
Peng, C., Chan, M. N., and Chan, C. K. (2001). The Hygroscopic Properties of
Dicarboxylic and Multifunctional Acids: Measurements and UNIFAC Predictions. Environ. Sci. Technol. 35:4495–4501.
Rosenoern, T., Schlenker, J. C., and Martin, S. T. (2008). Hygroscopic Growth of
Multicomponent Aerosol Particles Influenced by Several Cycles of Relative
Humidity. J. Phys. Chem. A. 112:2378–2385.
Rosenoern, T., Paulsen, D., and Martin, S. T. (2009). The 1-by-3 Tandem Differential Mobility Analyzer for Measurement of the Irreversibility of the
Hygroscopic Growth Factor. Aerosol Sci. Technol. 43:641–652.
Saathoff, H., Naumann, K. H., Schnaiter, M., Schock, W., Mohler, O., Schurath,
U., Weingartner, E., Gysel, M., and Baltensperger, U. (2003). Coating of Soot
and (NH4 )2 SO4 Particles by Ozonolysis Products of Alpha-Pinene. J. Aerosol
Sci. 34:1297–1321.
Seinfeld, J. H., Kleindienst, T. E., Edney, E. O., and Cohen, J. B. (2003).
Aerosol Growth in a Steady-State, Continuous Flow Chamber: Application
to Studies of Secondary Aerosol Formation. Aerosol Sci. Technol. 37:728–
734.
Seinfeld, J. H., and Pandis, S. N. (2006). Atmospheric Chemistry and Physics:
From Air Pollution to Climate Change. John Wiley & Sons, New York.
Shilling, J. E., Chen, Q., King, S. M., Rosenoern, T., Kroll, J. H., Worsnop,
D. R., McKinney, K. A., and Martin, S. T. (2008). Particle Mass Yield in
Secondary Organic Aerosol Formed by the Dark Ozonolysis of Alpha-Pinene.
Atmos. Chem. Phys. 8:2073–2088.
Shilling, J. E., Chen, Q., King, S. M., Rosenoern, T., Kroll, J. H., Worsnop,
D. R., DeCarlo, P. F., Aiken, A. C., Sueper, D., Jimenez, J. L., and Martin,
S. T. (2009). Loading-Dependent Eemental Composition of Alpha-Pinene
SOA Particles. Atmos. Chem. Phys. 9:771–782.
Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Avery, K. B., Tignor,
M., and Miller, H. L., eds. (2007). IPCC Climate Change 2007: The Physical
Science Basis. Contribution of Working Group I to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge.
Stokes, R. H., and Robinson, R. A. (1966). Interactions in Aqueous Nonelectrolyte Solutions. Part 1: Solute-Solvent Equilibria. J. Phys. Chem. Pre-1997.
70:2126–2131.
Swietlicki, E., Hansson, H. C., Hämeri, K., Svenningsson, B., Massling, A.,
McFiggans, G., McMurry, P. H., Petäjä, T., Tunved, P., Gysel, M., Topping, D.,
Weingartner, E., Baltensperger, U., Rissler, J., Wiedensohler, A., and Kulmala,
M. (2008). Hygroscopic Properties of Submicrometer Atmospheric Aerosol
Particles Measured with H-TDMA Instruments in Various Environments: A
Review. Tellus: Series B. 60:432–469.
Takahama, S., Pathak, R. K., and Pandis, S. N. (2007). Efflorescence Transitions
of Ammonium Sulfate Particles Coated with Secondary Organic Aerosol,
Environ. Sci. Technol. 41:2289–2295.
Downloaded by [Harvard College] at 13:11 07 September 2012
260
M. L. SMITH ET AL.
Tang, I. N., and Munkelwitz, H. R. (1984). An Investigation of Solute Nucleation
in Levitated Solution Droplets. J. Colloid and Interface Sci. 98:430–438.
Varutbangkul, V., Brechtel, F. J., Bahreini, R., Ng, N. L., Keywood, M. D.,
Kroll, J. H., Flagan, R. C., Seinfeld, J. H., Lee, A., and Goldstein, A. H.
(2006). Hygroscopicity of Secondary Organic Aerosols Formed by Oxidation
of Cycloalkenes, Monoterpenes, Sesquiterpenes, and Related Compounds.
Atmos. Chem. Phys. 6:2367–2388.
Virkkula, A., Van Dingenen, R., Raes, F., and Hjorth, J. (1999). Hygroscopic
Properties of Aerosol Formed by Oxidation of Limonene, Alpha-Pinene, and
Beta-Pinene. J. Geophys. Res. 104:3569–3579.
Virtanen, A., Joutsensaari, J., Koop, T., Kannosto, J.,Yli-Pirilä, P., Leskinen,
J., Mäkelä, J. M., Holopainen, J. K., Pöschl, U., Kulmala, M., Worsnop,
D. R., and Laaksonen, A. (2010). An Amporphous Solid State of Biogenic
Secondary Organic Aerosol Particles. Nature, 467:824–827.
Wang, J., Hoffmann, A. A., Park, R. J., Jacob, D. J., and Martin, S. T. (2008a).
Global Distribution of Solid and Aqueous Sulfate Aerosols: Effect of the
Hysteresis of Particle Phase Transitions. J. Geophys. Res. 113:D11206.
Wang, J., Jacob, D. J., and Martin, S. T. (2008b). Sensitivity of Sulfate Direct
Climate Forcing to the Hysteresis of Particle Phase Transitions. J. Geophys.
Res. 113:D11207.
Wang, S. C., and Flagan, R. C. (1990). Scanning Electrical Mobility Spectrometer. Aerosol Sci. Technol. 13:230–240.
Weingartner, E., Gysel, M., and Baltensperger, U. (2002). Hygroscopicity of
Aerosol Particles at Low Temperatures. Part 1: New Low-Temperature HTDMA Instrument: Setup and First Applications. Environ. Sci. Technol.
36:55–62.
Wexler, A. S., and Seinfeld, J. H. (1991). 2nd-Generation Inorganic Aerosol
Model. Atmos. Environ. 25:2731–2748.
Zhang, Q., Jimenez, J. L., Canagaratna, M. R., Allan, J. D., Coe, H., Ulbrich,
I., Alfarra, M. R., Takami, A., Middlebrook, A. M., Sun, Y. L., Dzepina,
K., Dunlea, E., Docherty, K., DeCarlo, P. F., Salcedo, D., Onasch, T., Jayne,
J. T., Miyoshi, T., Shimono, A., Hatakeyama, S., Takegawa, N., Kondo,
Y., Schneider, J., Drewnick, F., Borrmann, S., Weimer, S., Demerjian, K.,
Williams, P., Bower, K., Bahreini, R., Cottrell, L., Griffin, R. J., Rautiainen, J.,
Sun, J. Y., Zhang, Y. M., and Worsnop, D. R. (2007). Ubiquity and Dominance
of Oxygenated Species in Organic Aerosols in Anthropogenically-Influenced
Northern Hemisphere Midlatitudes. Geophys. Res. Lett. 34:L13801.
Zuend, A., Marcolli, C., Peter, T., and Seinfeld, J. H. (2010). Computation
of Liquid-Liquid Equilibria and Phase Stabilities: Implications for RHDependent Gas/Particle Partitioning of Organic-Inorganic Aerosols. Atmos.
Chem. Phys. 10:7795–7820.
FIG. A1. Measured (long-dashed line) and modeled (short-dashed line)
number-diameter distributions of particles exiting the chamber. The model is
described in the Appendix. The solid line shows the number-diameter distributions of seed particles recorded prior to formation of secondary organic material.
Panel (a) shows a distribution measured for low organic particle mass concentrations (1.63 µg m–3 ) and large seed particles (76.5 nm), which can be compared
to the distribution displayed in panel (b) for high organic particle mass concentrations (12.2 µg m–3 ) and small seed particles (49.5 nm). The two sets of
concentrations imply low and high organic volume mass fractions, respectively,
for particles sampled at 90 nm (vertical line). The calculated probability density functions porg,ss (ε) of organic volume fraction for the 90-nm particles are
shown in the insets. Mode values ε̂+1 and ε̂+2 are labeled. The inset illustrates
the basis for the box quartiles appearing within Figures 5b and 7b.
APPENDIX: SECTIONAL MODEL FOR p org,ss (␧; dme)
A probability density function nseed (dme ) describes the
number-diameter distribution of the mass-equivalent diameter dme of the seed particles that are initially injected into the
chamber. For nonporous particles, mass-equivalent and volumeequivalent diameters are identical. As the sulfate seed particles are introduced, α-pinene oxidation products condense onto
the particles. This condensation leads to diameter growth. The
amount of organic material condensed on a particle, as expressed
by a theoretical mass-equivalent diameter dme,org , depends on a
particle’s residence time because growth continues until a particle exits the chamber. The chamber is operated in a continuousflow mode, meaning that seed particles are continuously injected while grown particles are continuously withdrawn. A
probability density function presidence (τ ) then exists of a particle’s residence time τ in the chamber. The distributions nseed
and presidence imply that particles of common dme can be physically quite different: some can have smaller dme,seed but larger
dme,org than others of larger dme,seed but smaller dme,org (i.e.,
3
3
+ dme,org
]1/3 ). There is therefore a probability
dme = [dme,seed
density function porg,ss (ε;dme ) of organic volume fraction ε for
particles of identical dme exiting the chamber at steady state,
where ε = 1 – (dme,seed /dme )3 .
A numerical sectional model is used to estimate the probability density function porg,ss (ε;dme ) of organic volume fraction. The number-diameter distribution nseed (dme ) is represented
by i sections. The distribution nseed (dm ) is measured before
ozonolysis is initiated (Figure A1, solid lines), and our analysis assumes nseed (dme ) = nseed (dm ). Mobility diameter dm
and mass-equivalent diameter dme are equal for spherical nonporous particles. In the model, each section ni,0 (dme ) is described by a normal distribution defined by a mean diameter µi
and a standard deviation σ i . The amplitude of each section is
weighted by the measured value of nseed (µi ) such that the sum
of the sections matches the original distribution, specifically
PHASE TRANSITIONS OF SOM-AS PARTICLES
|1 − ni,0 (dme )/nseed (dme )| < 0.02 for all dme . This sectional
approach accounts for multiply charged particles within the seed
distribution.
For each section, the size distribution ni,ss (dme ) exiting the
chamber is given by the analytical solution to the steady-state
distribution of spherical particles undergoing condensational
growth in a continuous-flow chamber, as described by Seinfeld
et al. (2003):
ni,ss (dme ) = (1−ατ )θ4 e−θ1 θ2θ3 ni,0 (dme ) + θ5 e−θ6 (dme ) θ7 (dme )1−θ3
dme
×
e−θ6 (s) θ7 (s)ni,0 (s)ds
[A1]
Downloaded by [Harvard College] at 13:11 07 September 2012
0
The overall number-diameter distribution nss (dme ) exiting the
chamber is then given by ni,ss (dme ). In Equation (A1), α is the
particle wall loss coefficient and the θ i are algebraic relationships among α, τ , β, λ, and dme , as defined in Seinfeld et al. The
terms α and τ are obtained by independent measurement in the
laboratory and are therefore known quantities for use in Equation (A1). The terms β and λ represent the diameter-dependent
particle growth rate I (dme ) arising from the condensation of
secondary organic material and can be expressed in the following form: I (dme ) = β/(dme + λ). These terms are obtained by
optimizing the modeled to the measured nss at each organic
particle mass concentration (cf. short-dashed and long-dashed,
Figure A1).
The probability density function porg,ss (ε) of organic volume
fraction for fixed dme directly follows as the derivative of the
cumulative distribution function Porg,ss (ε), which is calculated
as follows:
ni,ss δ(εi < ε)
Porg,ss (ε) = i [A2]
i ni,ss
where εi = 1 – (µi /dme )3 for a monodisperse approximation. The term δ(εi < ε) is a step function that is defined
as zero when the stated condition is false and as unity when
it is true. The count median ε̄¯ , count mode ε̂, and arithmetic
mean ε̄ are defined by ε:(Porg,ss = 0.5), ε:max(porg,ss ), and
1
0 porg,ss (ε)ε dε, respectively. In cases of significant concentra-
261
tions of multiply charged seed particles, there are two modes, ε̂+1
and ε̂+2 .
Examples of porg,ss (ε;dme ) for low and high organic particle mass concentrations (1.63 and 12.2 µg m–3 , respectively)
are shown in the insets of panels A and B of Figure A1 for
DMAmono set to pass 90-nm particles. The top main panel of
Figure A1 (solid line) shows that doubly charged seed particles
of 113 nm accompany the 76.5-nm singly charged seed particles. In panel B, 72-nm doubly charged particles accompany the
49.5-nm singly charged particles. The inset of panel A shows
that porg,ss (ε) has a single mode ε̂+1 . The explanation is that the
doubly charged seed particles are larger than 90 nm. In comparison, the inset of panel B shows that porg,ss (ε) has two modes
ε̂+1 and ε̂+2 , as explained by 72-nm seed particles that serve as
cores to 90-nm mixed particles.
Although two particle types are sufficient to describe the
1×3-TDMA data shown in Figures 5a and 7a, additional particle types exist in the population. For example, DMAmono set
to pass 90-nm particles also selects doubly charged particles of
diameter 184 nm. These particles, like their 90-nm counterparts,
can also contain small and large seed particles, corresponding
to the +1 and +2 seed particles selected initially by the nDMA.
In principle, the list of particle types increases further if +3 particles are taken into account, though the statistical distribution
of number concentrations drops off precipitously. In the case
of DMA-CCN activation curves reported by King et al. (2009),
four particle types of highest statistical occurrence (i.e., +1/+1,
+2/+1, +1/+2, and +2/+2, where x/y represents the charge
x on the seed particle entering the HEC and the charge y on the
particle classified from the flow exiting the HEC) were required
to explain the observations. For the data of Figure 5a, however,
only two particle types (i.e., +1/+1 and +2/+1) are needed
to explain the observed transmission curves. The reduction in
the number of needed particle types can be understood in that
deliquescence occurs over a small RH range that approaches the
RH stability of the 1×3-TDMA and the resolution of multiple
particle types is therefore obscured. Because the influence of
these particles is not observed in the data of the 1×3-TDMA,
they are also not included in the modeling of organic volume
fraction.