Introduction
The pH is a fundamental aerosol property that affects aerosol formation and
composition through reactions that involve the hydronium ion (e.g., Jang et
al., 2002; Eddingsaas et al., 2010; Surratt et al., 2010) and gas–particle
partitioning of semivolatile acids and bases (e.g., Fridlind and Jacobson,
2000; Young et al., 2013; Guo et al., 2016, 2017). Acidity also modulates
aerosol toxicity and atmospheric nutrient supply to the oceans by augmenting
the solubility of transition metals and other nutrient species
(Meskhidze et al., 2003; Nenes et al., 2011; Longo et al., 2016; Stockdale
et al., 2016; Fang et al., 2017). Despite its importance, challenges in
measuring fine-mode-particle pH have led to the adoption of measurable
aerosol properties as acidity proxies, such as aerosol ammonium-sulfate ratio
or ion balances with a priori assumed dissociation states (e.g., Paulot and
Jacob, 2014; Wang et al., 2016; Silvern et al., 2017). Recent work has shown
that such proxies are not uniquely related to pH because they do not capture
variability in particle water content, ion activity coefficients, or
dissociation state of polyprotic acids and bases (Guo et al., 2015, 2016; Hennigan
et al., 2015; Song et al., 2018). An alternative approach
that better constrains aerosol pH is a thermodynamic analysis of semivolatile
acid (or base) measurements, whose partitioning is observably sensitive to
shifts in aerosol acidity (pH is optimally constrained when gas–particle
concentration ratios approach 1:1), and with the aerosol water content or
phase state constrained as well (Guo et al., 2015; Hennigan et al., 2015).
NH3–NH4+, HNO3–NO3-, and
HCl–Cl- pairs often meet this condition for a wide range of
atmospherically relevant pH. The method has been utilized for a range of
meteorological conditions (RH, T) and gas–aerosol concentrations
demonstrating that model predictions are often in agreement with observations
(Bougiatioti et al., 2016; Guo et al., 2016, 2017; Liu et al., 2017; Murphy
et al., 2017; Song et al., 2018).
Despite their skill and widespread use in regional and global models, aerosol
thermodynamic models can predict ammonium-sulfate molar ratios (Kim et al.,
2015; Weber et al., 2016; Silvern et al., 2017) that departs from
observations in seemingly counterintuitive ways. In the southeastern US,
where total ammonium (NHx = NH3 +
NH4+) is in large excess of particle sulfate, observed
NH4+/SO42- molar ratios are in the range of 1–2 (Hidy et
al., 2014; Guo et al., 2015; Kim et al., 2015). Thermodynamic models predict
very low pH (0.5 to 2) (Guo et al., 2015) and molar ratios always close to 2
(Kim et al., 2015; Weber et al., 2016; Silvern et al., 2017). This
predicted–observed molar ratio discrepancy has led to the hypothesis that
thermodynamic predictions are incorrect because they do not consider
interactions with organic species, either in the form of films that inhibit
gas-to-particle mass transfer of NH3 or other mechanisms that are
not accounted for (Silvern et al., 2017). Such limitations, if prevalent, are
suggested to oppose the validity of aerosol thermodynamic equilibrium with
significant impacts on aerosol chemistry and acidity-mediated processes
worldwide (Silvern et al., 2017), especially given the expected increasing
organic mass fractions in the future due to reduced anthropogenic emissions,
as seen with SO2 emission reductions in the eastern US (Hand et
al., 2012; Attwood et al., 2014; Hidy et al., 2014).
The effect of organic species on gas–particle equilibrium of inorganic
species has been the subject of many past studies. Organic films are
often hypothesized to act as barriers for gas–particle mass transfer, which
given their ubiquity means they require special attention in studies. For
example, Anttila et al. (2007) reports the formation of ∼10 nm thick
organic films in regions with monoterpene emissions, which are the largest
source of summertime organic aerosol (OA) in the southeastern US (Zhang et al.,
2018). Lab studies have shown that organic films may significantly slow down
mass transfer of NH3 from gas to particle at low relative humidity
(less effect at higher RH, such as the southeastern US) (Daumer et al., 1992;
Liggio et al., 2011) but have little effect on water vapor uptake for a large
RH range (Garland et al., 2005). Such films, as noted by Silvern et
al. (2017), would have important implications for partitioning of
NH3 and other larger semivolatile molecules, such as H2O,
HNO3, and organic acids. However, in contrast, numerous studies show
that NH3, water vapor, and HNO3 equilibrate with
organic-rich atmospheric aerosols (Ansari and Pandis, 2000; Moya et al.,
2001; Morino et al., 2006; Fountoukis et al., 2009; Guo et al., 2015, 2016,
2017; Liu et al., 2017; Murphy et al., 2017; Paulot et al., 2017), which
suggest organic films, if present, do not impose considerable delays in mass
transfer and gas–particle equilibration.
At low temperature and low relative humidity, particles may be in a
semi-liquid or glassy state characterized by a very low molecular diffusivity
throughout its volume (e.g., Zobrist et al., 2008; Bertram et al., 2011; Tong
et al., 2011; Zobrist et al., 2011; Bones et al., 2012; Reid et al., 2018).
When in this state, gas–particle mass transfer of all semivolatile components
may be severely limited and require much longer timescales to equilibrate
than the ∼20 min typically thought to apply for PM1
(Dassios and Pandis, 1999; Cruz et al., 2000; Fountoukis et al., 2009).
However, such an effect has not been observed for the conditions in the
eastern US, as there is good agreement between observed and predicted
particle water, and partitioning of NH3–NH4+ and
HNO3–NO3-, especially in cases where RH is
sufficiently high (greater than 40 %) to maintain the aerosol in a
deliquesced (completely liquid) state (Guo et al., 2015, 2016).
Other reasons, unrelated to the presence of organic aerosol, may drive the
observed molar ratio discrepancy. Analyses of aerosol acidity, molar ratios,
and partitioning of semivolatile species often neglect the variations of
composition with size, especially in the PM1 to PM2.5 range
(Keene et al., 1998; Fridlind and Jacobson, 2000; Nenes et al., 2011;
Young et al., 2013; Bougiatioti et al., 2016; Fang et al., 2017). If acidity
across size changes sufficiently, average equilibrium composition (including
molar ratios) may deviate considerably against observations owing to the
nonlinear dependence of partitioning with acidity (e.g., Guo et al., 2016).
Soluble nonvolatile cations (NVCs, such as Na+, K+,
Ca2+, Mg2+), potentially present in large quantities in
PM2.5 and to a lesser extent in PM1, can strongly modulate acidity
and molar ratios. NVCs are often omitted from thermodynamic calculations
because of their relatively minor contribution to aerosol mass and ion charge
balance; for similar reasons, NVCs are not routinely included in aerosol
composition measurements; when they are, proximity to level of detection
(LOD) often increases their concentration uncertainty. Here we show, based on
analysis of observational aerosol and gas datasets, that excluding even
small amounts of NVC in thermodynamic analyses results in predicted
NH4+/SO42- molar ratios close to 2, whereas including them
brings model-predicted molar ratios into agreement with observed levels. We
also assess the implications of using specific datasets on molar ratios and
the impact of adopting a size-averaged (bulk) thermodynamic analysis
against one that considers an incomplete mixing (size-dependent composition)
of ambient aerosols.
Methods
Molar ratios definition
Two ammonium-sulfate aerosol molar
ratios (mol mol-1) are used in the following analysis:
R=NH4+SO42-,RSO4=NH4+-NO3-SO42-.
Both are based on mole concentrations in units of mol m-3.
RSO4 is a more narrowly defined molar ratio that excludes
NH4+ associated with NO3-, because some fractions
of ammonium sulfate and ammonium nitrate can be associated with different
sized particles (Zhuang et al., 1999) and molar ratios are calculated based
on bulk composition data (PM2.5 or PM1). This issue is discussed in
more detail below. The upper limit for R and RSO4 is 2 for a
particle composition of pure (NH4)2SO4 and a lower limit of 0
for R when SO42- is associated with other cations instead of
NH4+ (e.g., Na2SO4) or if there is free
H2SO4 in the aerosol. A negative RSO4 can occur for
conditions of high NO3- and low NH4+,
SO42- concentrations (e.g., NaNO3) but is rare for
ambient fine particles (at least not seen in the three datasets studied in
this paper). R or RSO4 is typically observed in the range of 1
and 2 in the southeastern US (Hidy et al., 2014; Guo et al., 2015; Weber et
al., 2016). In cases where NO3- levels are low relative to
SO42-, the two ratios, RSO4 and R, are equivalent,
as is observed in the summertime southeastern US, where NO3- is
typically ∼0.2 µg m-3, NH4+
∼1 µg m-3, and SO42-
∼3 µg m-3 (Blanchard et al., 2013).
Data
Two datasets are mainly used for analysis: the Southern
Oxidant and Aerosol Study (SOAS) and the Wintertime Investigation of
Transport, Emissions, and Reactivity (WINTER). The SOAS study was conducted
from 1 June to 15 July in the summer of 2013 at a rural ground site in
Centreville (CTR), AL, representative of the southeastern US background
atmosphere in summer. PM2.5 ions were determined with a
particle-into-liquid sampler coupled with an ion chromatograph (PILS-IC). The
PILS-IC detects aerosol water-soluble anions and cations collected and
diluted by deionized water to the extent of complete deprotonation of
H2SO4 in the aqueous sample (Orsini et al., 2003). NH3
was obtained from chemical ionization mass spectrometer measurements (You et
al., 2014). In the following, we only use PM2.5 ion data from a 12-day
period (11–23 June) of the SOAS campaign. (PILS PM1 data were collected
in the second half of the study and are not used here). Periods of rainfall
are not included in the analysis, as equilibrium does not apply. The same
dataset was used to study pH sensitivity to sulfate and ammonia (Weber et
al., 2016). PM2.5 anion and cation data along with NH3 and
HNO3 were also collected with a Monitor for AeRosols and Gases in
ambient Air (MARGA) during SOAS (Allen et al., 2015). The WINTER data were collected
during 13 research aircraft flights from 1 February to 15 March 2015 mainly
sampling over the northeastern US. We use PM1 aerosol data collected
with a high-resolution time-of-flight aerosol mass spectrometer (hereafter
referred to as AMS) (Schroder et al., 2018), which have been extensively compared to the PILS anion
measurements also made in that study (Guo et al., 2016). Details of the these
two campaigns and instruments, and calculations and verification of pH based
on the observation datasets, have been described in Guo et al. (2015, 2016),
respectively.
In the following analysis, we focus on R for summertime datasets since
NO3- was generally low, and RSO4 for wintertime datasets where higher NO3- concentrations were observed.
Thermodynamic analyses of both datasets indicate highly acidic aerosols with
an average pH ∼1 (Guo et al., 2015, 2016). At these pH levels,
aerosol sulfate can be in the partially deprotonated form of
HSO4- instead of SO42-. For example, 10 % of the
total sulfate is predicted to be HSO4- for the SOAS condition
(see Fig. S1 in the Supplement). Free-form H2SO4, which requires
even lower pH, is rare. To avoid any confusion, we note that in this study
SO42- refers to the sum of total aqueous aerosol sulfate
(SO42-, HSO4-, and H2SO4), i.e.,
S(VI). Similarly, NH4+ refers to the sum of total aqueous
ammonium (NH4+, NH3) and NO3- refers to
the sum of total nitrate (NO3-, HNO3) in aqueous
aerosols. SO42-, NH4+, and NO3- are
reported by PILS-IC. However, PILS-IC cannot distinguish the in situ aerosol
ion forms for collecting aerosols in diluted deionized water (i.e., the ionic
strength is altered) (Orsini et al., 2003). The AMS vaporizes aerosols and
ionizes non-refractory species with a 70 eV electron impact ionization and
also cannot distinguish the dissociation states of inorganic ions (DeCarlo et
al., 2006).
In addition to the SOAS and WINTER datasets, the Southeastern Aerosol
Research and Characterization (SEARCH) CTR sampling site (the same as SOAS)
historical data from year 1998 to 2013 are reanalyzed to show that the
thermodynamic model can reproduce the observed decreasing trend of
RSO4 when NVCs are considered. Molar ratios determined from the
Chemical Speciation Network (CSN), which were utilized and discussed by
Silvern et al. (2017) and Pye et al. (2018), are not used in this work
because of a significant low bias when compared to the SEARCH and SOAS data
(see Table S1 and S2 in the Supplement). The discrepancy is likely due to the
loss of semivolatile NH4+ collected on the CSN nylon filters (Yu
et al., 2006; Silvern et al., 2017) and can result in an underestimation or
under-measurement in R, compared to online measurements, by as much as 1
unit (Table S1). Other than the tentative explanation that sulfate aerosols
are coated by organic material, Silvern et al. (2017) found that NVCs would
modify R on average by 0.11 for the ensemble of CSN sites and NVC
concentrations were too low to significantly affect the charge balance, as
previously shown by Kim et al. (2015). Therefore, NVCs could not explain the
overprediction of R by the thermodynamic model due to the large low bias of
CSN R shown above. In contrast, this study investigates the effects of NVCs
by three datasets, including SOAS, WINTER, and SEARCH, and concludes the
importance of NVCs in accurately predicting ammonium-sulfate molar ratios.
Thermodynamic analysis of observations
We have used the thermodynamic model ISORROPIA II (Fountoukis and Nenes, 2007) to determine
the liquid water content and composition (including H+) of an
NH4+–SO42-–NO3-–Cl-–Na+–Ca2+–K+–Mg2+–water
inorganic aerosol (or a subset therein) and its partitioning with
corresponding gases. A molality-based definition of pH is used:
pH=-log10γH+Haq+=-log101000γH+Hair+Wi+Wo≅-log101000γH+Hair+Wi,
where γH+ is the hydronium ion activity coefficient
(assumed = 1; note that the binary activity coefficients of ionic pairs,
including H+, are calculated in the model), and Haq+
(mol kg-1) and Hair+ (µg m-3) are the
hydronium ion concentration in particle liquid water and volume of air,
respectively. Wi and Wo (µg m-3) are particle water
concentrations associated with inorganic and organic species, respectively.
The pH predicted solely with Wi is systematically lower by 0.15–0.23 units
but highly correlated (r2=0.97) to pH predicted with measured total
particle water (Wi+Wo) for the southeast US (which includes the
SOAS study), where Wo accounted for 35 % of total particle water (Guo
et al., 2015). For simplicity, we therefore use only Wi for the
following pH calculations. ISORROPIA II was run in forward mode to
calculate gas–particle equilibrium concentrations based on the input of total
concentration of various inorganic species (e.g., NH3 +
NH4+). In all cases we also chose a metastable (not
stable) solution, which assumes inorganic ions are associated with the
aerosol components that are completely aqueous and contain no solid
precipitate forms, other than CaSO4 (Haq+ is
meaningless in a completely effloresced aerosol). Given this phase state
requirement, we restrict the analysis to conditions where RH >40 %.
Mixing state
Because the aerosol composition data are bulk PM1 or PM2.5, and
used as input to ISORROPIA II, the thermodynamic analysis implicitly assumes
that all particle species were internally mixed, so that one value of pH
represents the aerosols and governs the gas–particle partitioning. The
existence of externally mixed particles may quantitatively and qualitatively
affect the bulk thermodynamic analysis. To address this, we begin from the
bulk analysis, then repeat the same calculation, augmenting each time the
degree of external mixture of NVCs and sulfate. Direct measurements of
aerosol mixing state during SOAS suggest that ambient particles indeed
exhibit a range of mixing states (Bondy et al., 2018). In the external mixing
analysis, the bulk aerosol is split into two subgroups: (1) species largely
found in PM1 (e.g., NH4+ and SO42-) and
(2) species found in PM1-2.5, which contains mostly the NVCs,
NO3-, and some SO42- and NH4+. These
two external mixtures are in equilibrium with gaseous NH3 and
HNO3 and so interact through these species (i.e.,
NH4+ and NO3- can move between the two).
Nonvolatile species, such as SO42- and NVCs (Na+),
remain in the original size class assumed at the start of the analysis. To
determine the composition of the two subgroups, we iteratively solve for the
equilibrium conditions, by sequentially calling ISORROPIA for each subgroup.
The solution is found when the composition of each group no longer changes
with iteration and both are in equilibrium with the gas-phase species (in
this case, NH3, HNO3, and H2O (water vapor)).
The mass of each species (gas plus particle) is conserved at all times and
constrained by the observations. Given that pH is size dependent and
generally higher at larger sizes (Fridlind and Jacobson, 2000; Young et al.,
2013; Bougiatioti et al., 2016; Fang et al., 2017), bulk pH is compared
against an aerosol liquid water-weighted pH:
pH=-log101000Hair,subgroup1++Hair,subgroup2+Wi,subgroup
1+Wi,subgroup 2.
Discussion
Internal vs. external mixtures
Our thermodynamic analysis up to
this point has been based on the assumption that all ions were internally
mixed (e.g., bulk PM2.5 or PM1). Although over time gas–particle
and particle–particle interactions will lead to complete internally mixed
systems (Seinfeld and Pandis, 2016), aerosol near their source regions tend
to be externally mixed. Typical ambient conditions can be expected to exist
somewhere between these two extreme cases
(Bondy et al., 2018) owing to chemistry, coagulation, cloud processing,
dilution, and gas-to-particle mass transfer (Zaveri et al., 2010). We address
this here by studying how the conclusions described above are affected by the
degree of mixing of NVCs with sulfate – as the other species,
being semivolatile, quickly equilibrate.
PM2.5 Na+, K+, Ca2+, and Mg2+
from sea salt (or dust) are often not well mixed with ammonium and sulfate
because of their different sources and sizes. NVC from sea salt and dust are
largely produced by mechanical means and so are mainly in the coarse mode,
with a tail extending into the fine mode (Whitby, 1978). Biomass burning and
biogenic K+ is emitted into the fine mode (Bougiatioti et al., 2018); however,
ammonium and sulfate are formed through gas-phase processes
and mostly reside in the accumulation mode (e.g., Whitby, 1978; Seinfeld and
Pandis, 2016). For the SOAS PILS-IC dataset, NH4+ and
SO42- were highly correlated (r2=0.88), but
NH4+ and Na+ (r2=0.07) or SO42-
and Na+ (r2=0.17) were not. In contrast, PM2.5
Na+ and NO3- (r2=0.82) or Na+ and
Cl- (r2=0.64) were highly correlated, which is consistent with
internal mixing of most Na+, NO3-, and Cl-
ions, leading to the depletion of some Cl- through evaporation of HCl
(e.g., Katoshevski et al., 1997; Seinfeld and Pandis, 2016). Rapid scavenging
of HNO3 by sea-salt aerosols is well established (Hanisch and
Crowley, 2001; Meskhidze et al., 2005), with equilibrating timescales of 3–10 h
for HNO3 uptake by 1–3 µm sea spray aerosols
(Meng and Seinfeld, 1996; Fridlind and Jacobson, 2000) and subsequent
evaporation of HCl.
NVCs can also be associated with small amounts of sulfate. For example,
sea-salt aerosols are largely composed of NaCl but also include sulfate,
approximately 8 % (g g-1) of all ions (∼25 %
SO42-/Na+ mass ratio) (DOE, 1994). In addition, sulfur
enrichment and chloride depletion in aged sea-salt aerosols are possible by
uptake of H2SO4 or oxidation of dissolved SO2 by
O3 (McInnes et al., 1994; O'Dowd et al., 1997). These secondary
sulfates are normally referred as non-sea-salt sulfates, to be distinguished
from sea-salt sulfate that is naturally in sea waters (Tang et al., 1997).
Many studies have reported sulfate-containing sea-salt aerosols with some
degree of internal mixing (Andreae et al., 1986; McInnes et al., 1994;
Murphy et al., 1998; Laskin et al., 2002; Bondy et al., 2018). In summary, a
realistic external mixing state of the SOAS fine particles is that most of
NH4+ and SO42- are in PM1, whereas
Na+ with associated anions (NO3- and Cl-)
and at least small amounts of NH4+ and SO42- are
associated in PM1-2.5 (particles with sizes 1–2.5 µm). This
is consistent with the single particle mixing state observations by Bondy et
al. (2018) from the SOAS study. The interactions between aerosols with gases
are illustrated in Fig. 7a. Particle size distributions measured in the
southeast US also support these types of particle mixing state (Fang et al.,
2017).
(a) Schematic of assumed internally and externally mixed
aerosols. NVCs (here represented by Na+) are all assumed in
PM1-2.5 for the external mixing case. The two externally mixed aerosol
groups (PM1 and PM1-2.5) are in equilibrium with the same gases.
The internal mixed case has bulk PM2.5 compositions (PM1+
PM1-2.5) together with gases as model input. The predicted molar ratio
(R), pH, and liquid water (Wi) of the internally and
externally mixed aerosols are summarized in (b), (c), and
(d), respectively. The x axis is the sulfate (mass) fraction
assumed in PM1-2.5, with the remaining sulfate in PM1. For the
analysis shown here only data for which measured Na+ was above the
LOD are utilized. Lower Na+ concentrations require smaller
fractions of SO42- in the PM1-2.5 range for agreement with
the bulk analysis (e.g., 5 % for the PILS-IC Na+ LOD of
0.07 µg m-3). Standard deviations of the data are shown as
error bars or shaded zones.
Explanation for role of NVCs in R based on bulk (internal
mixture) analysis
Recapping the bulk analysis above where ions are all
assumed to be internally mixed, we have shown that the observations relating
R to NVCs, and deviations in R between models and observations, can be
readily explained. First, when NVCs such as Na+ are present in the
ambient aerosol and not included in the thermodynamic model, but some
fraction of the associated anion pair is, the thermodynamic model will
predict higher NH4+ than observed because the model will
partition greater levels of available semivolatile cations (i.e.,
NH3) to the particle phase (NH4+) to conserve
NHx and make up for the missing NVCs. This leads to a
predicted R near 2. The trends in measured R with measured Na+
are also expected. As noted before, measured R becomes increasingly less
than 2 as measured Na+ increases because at higher Na+
bulk aerosol pH increases (Fig. 3c), resulting in lower ε(NH4+) (see NH4+ S curve in Fig. S9 in the Supplement), shifting NH4+ to gas-phase NH3. Other NVCs have
similar effects as Na+, as long as soluble forms of the salts are
observed (e.g., NaNO3, Na2SO4, KNO3,
K2SO4, Ca(NO3)2, Mg(NO3)2). We have
shown with this bulk analysis that accurately including NVCs in the
thermodynamic analysis appears to largely resolve the disparity in predicted
and measured R for the datasets we analyzed. But the bulk analysis is only
an approximation of the actual aerosol mixing state. We next test if assuming
an internal mixture will roughly represent the behavior of externally mixed
aerosols in terms of the effect of NVCs on R, pH, and partitioning of
semivolatile species. To assess this, we consider the behavior of external
mixing cases.
Explanation for the role of NVCs in R based on external mixture
analysis
An extreme (and unrealistic at the timescale of aerosol lifetime
Zaveri et al., 2010) external mixture is where PM1 is composed of all
the measured NH4+, SO42- and PM1-2.5 is
composed of all the measured Na+ (all NVCs), NO3-, and
Cl-.
NH3, HNO3, HCl, and H2O (water
vapor) can still equilibrate between these externally mixed particle types
(see Fig. 7a), given the relatively short equilibrating timescales for these
sizes of particles (Dassios and Pandis, 1999; Cruz et al., 2000; Fountoukis
et al., 2009). As Fig. 7b shows, for the extreme external mixing case (i.e.,
0 % sulfate in PM1-2.5), predicted R, combined from PM1 and
PM1-2.5, is close to 2, deviating from the lower predicted R of 1.66
±0.13 from the internal mixture. This is due to the vastly different pH
of PM1 (0.6) and PM1-2.5 (4.1) (Fig. 7c), where all
NH4+ is predicted to be in PM1 and all NO3-
is predicted to be in PM1-2.5.
For more realistic mixing cases, where some fraction of the sulfate is mixed
with NVCs (Bondy et al., 2018), the combined R of the external mixture
decreases rapidly as more SO42- is mixed with Na+ in
PM1-2.5. At ∼20 % SO42- fraction in PM1-2.5,
the average levels of predicted R start to converge between external and
internal mixtures (Fig. 7b). The difference in pH between PM1 and
PM1-2.5 is also reduced to within one pH unit (Fig. 7c). With these
small differences in pH, NH4+ can condense on both
externally mixed aerosol groups. For example, PM1 and PM1-2.5
NH4+ are predicted to be 0.67 and 0.04 µg m-3,
respectively (equal to the sum of the measured PM2.5 NH4+
of 0.71 µg m-3). From this analysis, based only on data when
Na+ was above the LOD, predicted R for the bulk and external
mixture are the same when on average 18±7 % (by mass) of the
PM2.5 SO42- is in the PM1-2.5 size range (i.e., mixed
with Na+). This is comparable to inferences of mixing based on
size-resolved aerosol measurements in the southeast (e.g., Fang et al.,
2017 shows ∼30 % PM2.5 SO42- mass in
PM1-2.5). Less internal mixing of SO42- with Na+
is needed when Na+ concentrations are lower. For the SOAS 12-day
Na+ average level of 0.07 µg m-3, only 5 % of the
SO42- (by mass) when mixed with Na+ produces the same
results as the bulk totally internal mixture case (see
Fig. S11 in the Supplement). Note that higher Na+ concentrations generally require
more SO42- to obtain agreement in R between external and
internal mixtures (scatter plots are shown in Fig. S12 in the Supplement).
The difference between the internally and externally mixed system is not as
great as may be expected, especially for particle pH and liquid water
(Wi) (Fig. 7c and d). Since liquid water levels are determined as the
sum of the water associated with the various salts, the bulk liquid water
generally equals the sum of the two externally mixed liquid water
concentrations, based on the Zdanovskii–Stokes–Robinson (ZSR) relationship
(Zdanovskii, 1936; Stokes and Robinson, 1966). Because the most hygroscopic
salts (i.e., NH4+ and SO42-; NO3-
concentrations are low) are in PM1, PM1 liquid water dominates over
PM1-2.5, making the combined pH of the external mixture nearly identical
to the PM1 pH (see Eq. 4 for combined pH calculation). The combined pH of
the external mixture is also similar to that of internal mixture, regardless
of the SO42- fractions (see Fig. 7c).
Summary
We have shown that including NVCs in the thermodynamic model largely resolves
the ammonium-sulfate molar ratio (R= NH4+/SO42-)
discrepancy, based on our datasets. (We have not utilized the CSN dataset
as other researchers have due to a large low bias in R.) Since only small
amounts of NVC can significantly affect R, measurement limitations, such as
high NVC LODs or NVCs not measured at all (e.g., AMS measurements), can lead
to substantial differences in observed and thermodynamic-model-predicted R.
We show that this bias in R (ISORROPIA-predicted R with Na+
minus ISORROPIA-predicted R without Na+) is correlated with and
highly sensitive to measured Na+, but not correlated with organic
aerosol mass or mass fraction. Similarly, the difference in measured R from
a ratio of 2 (2 minus observed R) is correlated to measured Na+
(NVCs) and not correlated with organic aerosol mass or mass fraction. If
organic films were limiting mass transfer, the discrepancy in R should
worsen as the films become thicker. We find the opposite. Furthermore,
ISORROPIA-predicted NH3–NH4+ partitioning (with
measured Na+ as input) agrees well with the observation, showing an
equilibrium state of the partitioning and no significant NH3 mass
transfer limit caused by organic films. These results provide evidence for the
role of NVCs, but not bulk organic aerosol species or organic films in the
molar ratio discrepancy observed in the southeastern US.
Excluding minor amounts of fine-mode NVC in thermodynamic calculations
results in a predicted R near 2, which is generally higher than observed
values. This results from the model criteria for aerosol electrical
neutrality and because semivolatile NH4+ has to be increased to
compensate for the missing NVCs. Less absolute discrepancy is associated with
predicted particle pH with or without NVC because pH is on a logarithmic
scale of Haq+ and the range of pH is larger than that of R (or
RSO4) in the eastern US. However, neglecting NVC can induce pH
biases that could result in significant partitioning errors for semivolatile
species like ammonium, nitrate, chloride, and even organic acids, under
certain conditions. Because NVCs are often minor constituents of fine
particles, especially for submicron particles, implying low ambient
concentrations and high measurement uncertainties, assessing thermodynamic
model predictions through molar ratios is problematic. If NVCs were not
measured or significantly below the measurement LOD, an ion charge balance
could be used to infer an upper limit on NVC concentrations, but addition of
measurement uncertainties can lead to uncertain results. Note that the ion
charge balance on its own generally cannot be used to infer H+
since the H+ concentrations are generally very low, even at the low
pH of the southeastern US aerosols, and the dissociation states of acids must
be known (e.g., proportions of HSO4- and SO42-),
which requires a full thermodynamic analysis.
A motivation for the organic effects on ammonia partitioning (Silvern et al.,
2017) was the observed RSO4 decreasing trend over the past
15 years in the southeastern US. Fully considering NVCs does not change the
finding of nearly constant fine-particle pH in the southeast (summertime)
despite the large sulfate reductions in the past 15 years, but it does now
lead to agreement with the observed RSO4 decreasing trend.
Although the analysis was performed assuming internal mixtures of aerosol
constituents, since only bulk PM2.5 composition data were available, we
show that external mixtures of NVCs and sulfate produce similar molar ratios,
with the requirement that only small amounts of sulfate are needed to be
mixed with the NVC-rich particle, which is qualitatively consistent with the particle
mixing state measured for the SOAS study reported by Bondy et al. (2018). In
contrast to molar ratio, the average pH for externally mixed aerosol is not
sensitive to the mixing fraction of SO42- and Na+.
Further assessments on possible effects of organic species on semivolatile
partitioning of inorganic species should be carried out, especially for
regions that are chemically different from the eastern US conditions
evaluated in this study.