To be submitted for Environmental Pollution
Origination and effects of size-resolved aerosols in the coastal area; measurement
during KORUS-AQ campaign
Keywords: MOUDI, OPS, Benzene Carboxylic Acid, WSOC, OC
______________
* Corresponding author:
Highlight
Size-resolved chemical properties in the region by different geographic origins of air masses
Comparison of different ionic balance approaches
Comparison of optical and chemical size distributions
Graphical Abstract
Abstract
Measurements for size-resolved chemical composition were made at the Anmyeon island station (ASL:
36° 32N; 126° 19E, 46 m above sea level), a clean background site on the west coast of the Republic of Korea.
This site is a Global Atmosphere Watch supersite (GAWS) managed by the World Meteorological Organization.
This study was carried out during the KORea-US Air Quality (KORUS-AQ) campaign between May 28 to June
20, 2016. The site receives both marine and continental (polluted) air masses, depending on the prevailing wind
direction. This study determined the main chemical compositions of size-resolved particulate matter (i.e.,
organic carbon (OC), water soluble organic carbon (WSOC), elemental carbon (EC), water soluble ions (WSI),
and benzene carboxylic acids (BCA)) using micro-orifice uniform deposit impactors (MOUDI). We conducted a
residence-time weighted concentration (RTWC) analysis to assess the impact of the regional transport on the
concentrations. PM2.5 mass, OC, EC, WSOC, WSI, and BCA were determined using a SUNET carbon analyser,
total organic carbon (TOC) analyser, ion chromatography, and liquid chromatography-mass spectrometry mass
spectrometry (LC-MSMS). In addition, an optical particle sizer (OPS) and Multi Angle Absorption Photometer
(MAAP) were used to obtain the sized-resolved volume concentrations and black carbon (BC) content. The sum
of the species presents a prominent accumulation mode peak at a diameter of 0.56 μm without the coarse mode
peak.
In the accumulation mode, SO42- (49.3%) was the dominant particle component in the size range of 0.1–1.8 μm,
followed by NH4+ (19.5%) and NO3- (13.6%). OC and EC accounted for 13.5% and 0.4%, respectively. 1,4BCA (18%) was also considerably high in the accumulation mode relative to the coarse mode (13.8%). 1,2,4BCA (11.4%), 1,3,5-BCA (17.7%), and 1,2,4,5-BCA (12.6%) were also considerably present in the
accumulation mode in the size-resolved PM. The differences between the size-resolved charge equivalent
concentration ratios using all measured WSIs and the ion neutralization model (INM) using expected NH 4+ are
discussed with the presence of organic acids, such as BCAs, in submicron particles that can lead to a more
neutral. Finally, the averaged size distributions of particle mass using the OPS and the MOUDI are compared.
INTRODUCTION
Particulate matter with aerodynamic diameters less than 2.5 μm (PM2.5) are found to be strongly
associated with mortality and morbidity, and are thus gaining attention around the world (Maji et al., 2017;
Sahani et al., 2014; Tang et al., 2017). Toxicological studies have suggested that ultrafine PM (i.e., less than 0.1
μm aerodynamic diameters) can penetrate deep into the alveoli of the lung, resulting in a number of adverse
health effects (Salma et al., 2015). To assess the adverse health effects of PM, the size-resolved chemical
compositions should be investigated. Size-resolved PM analyses are useful particularly to examine their
properties across a wide range of sizes as well as their associated impact on the environment and on public
health (e.g., interaction with climate variables, impact on the respiratory system upon inhalation, and long range
transport) (Bae and Park, 2013; Degrendele et al., 2016; Gong et al., 2016; Li et al., 2017; Wu et al., 2016).
Information regarding the size-resolved chemical composition of PM can provide a pivotal clue to
determine the sources of pollution and the related atmospheric processes. Micro-Orifice Uniform Deposit
Impactors (MOUDI) have been extensively used to collect PM to analyze the size-resolved chemical
compositions (Ham and Kleeman, 2011; Park et al., 2013; Plaza et al., 2011). However, the results of such
studies have occasionally reported possible occurrences of negative artifacts, such as a significant evaporation
loss of labile species at a low pressure (Zhang and McMurry, 1992), and sulfates can also interact with the
substrate (Kim et al., 2010). In general, most mass fractions of water-soluble secondary inorganic ions (nitrate,
ammonium, or sulfate) at the coast or island were found in the coarse mode (> 1 μm) under clean marine
conditions while those found in the fine mode (< 1 μm) occurred during polluted flow conditions (Cavalli et al.,
2004; Yeatman et al., 2001). In particular, high spatial variability of many elemental constituents was observed
at the coastal site with a diameter of less than 0.25 μm and at 0.25–2.5 μm, and this was associated with motor
vehicle use (Krudysz et al., 2008). The behavior of organic aerosols is affected by marine emissions of biogenic
volatile organic compounds and primary organic aerosols (Gantt et al., 2010), and their distribution in the
marine boundary layer or at the coast is strongly tied to seasonal cycles in ocean biology (O'Dowd et al., 2004).
Two-thirds of organic carbon (OC) and elemental carbon (EC) concentrations at the marine boundary layer of
the eastern Mediterranean were found in the fine mode while being water-soluble to a great extent (70%)
(Bougiatioti et al., 2013). Products of secondary organic aerosol (SOA) from aromatic VOCs have been used as
anthropogenic SOA tracers in identifying aerosol sources and source apportionment (Al-Naiema and Stone,
2017; Kleindienst et al., 2007). For instance, organic compounds besides poly aromatic hydrocarbons (PAHs)
detected in the ultrafine size fraction of rice straw and benz[de]anthracen-7-one and 4-methylphenylacetone
were found in cigarette smoke (Kleeman et al., 2008). Aromatic acids (e.g., benzoic acid (BA) and salicylic
acids) have been shown to exist in urban particulate matter (Ho et al., 2011). BA emissions are primarily from
various anthropogenic sources (i.e., automobile fumes, heavy-duty diesel trucks, and industrial boilers burning
oil as fuel) (Rogge et al., 1997). The primary photodegradation products of petroleum hydrocarbons are benzene
carboxylic acids (BCAs) with significant levels of BA that presumably form from the decomposition of alkylsubstituted benzenes. In a recent study, phthalic acid was also proposed as a tracer for the photooxidation of
primary aromatic hydrocarbons (i.e., naphthalene and methylnaphthalenes) in PM 2.5 (Kautzman et al., 2010;
Kawamura and Kaplan, 1987; Kleindienst et al., 2012).
This study aimed to assess the influence of different geographic origins of air masses based on the
size-resolved chemical compositions of aerosol particles (i.e., organic carbon (OC), water soluble organic
carbon (WSOC), elemental carbon (EC), water soluble ions (WSI), and BCA) on the west coast of Korea. In
addition, the source origin of ambient PM was determined via residence-time weighted concentrations (RTWC)
analysis. To accomplish this goal, we (1) address the size-resolved chemical properties in the region according
to the different geographic origins of the air masses, (2) discuss comparison of different ionic balance
approaches, and (3) compare optical and chemical size distributions.
EXPERIMENTAL METHODS
Sampling site and Size-resolved PM Sampling
Size-resolved PM sampling and measurements were carried out at Anmyeon Island in South Korea
from May 28 to June 20, 2016. The sampling site is a Global Atmosphere Watch Supersite (GAWS) managed by
the World Meteorological Organization in Korea. Anmyeon (36° 32 N; 126° 190, 46 m above sea level (ASL))
station has been operating as a a clean background (i.e., West coastal Korea) to monitor the PM and gaseous
components since 2003 (supplemental Fig. S1). The site receives both marine and continental air masses by
local circulation, and the regional and long-range transport processes depend on the prevailing meteorological
conditions (Jeon et al., 2015; Kim et al., 2012; Shim et al., 2013).
The filter-based 24 hour (hr) integrated sampler was operated at the GAWS, which is a relatively lowdensity, mixed-use (residential/light commercial) neighborhood that is also strongly impacted by marine PM
sources. All samples used for each of the analyses were collected downstream of the PM2.5 impactor (BMI Inc.,
USA) operating at a flow rate of 16.7 liters per minute (lpm). The integrated samplers were equipped with a
timer to allow the sampler to start and stop at pre-programmed times. The 24-hr integrated sampler was
equipped with two sample trains downstream of the impactor. One train consisted of a single 47-mm Teflon
filter (PTFE, R2PJ047, Pall Corp., USA) and the other train consisted of an organic denuder (Sunset Laboratory
Inc., USA) followed by a 47-mm quartz fiber filter (Pallflex, 2500QATUP, Pall Corp., USA) both operated at
8.35 lpm. OC, EC, WSOC, and WSI were analyzed from samples collected on a quartz filter. Air flows through
the integrated samplers were controlled using critical orifices for each train and were checked using a dry gas
calibrator (Defender 510, Mesa Labs, Inc., USA) before and after each sampling. The organic denuder was
fabricated with replaceable parallel charcoal-impregnated filter strips to collect gas-phase organic compounds
that could be absorbed to the downstream quartz fiber filter (Mader et al., 2003). PM2.5 mass concentration (i.e.,
average of three times analysis) was determined from the Teflon filter, which was weighed with a microbalance
(DM, Sartorius Corp., Germany). Prior to weighing, the filters were conditioned for 24 hr in a clean, controlled
chamber (relative humidity less than 40% and temperature about 20 °C). Size-resolved PM sampling was
conducted using a MOUDI (MSP Corp., USA) with integrated sampling between 48 and 72 hr. During the
sampling periods, a total of eight sets were collected. The samples were collected using aerodynamic cutpoint
diameters of 0.056, 0.1, 0.18, 0.32, 0.56, 1.0, 1.8, 3.2, 5.6, 10.0, and 18.0 μm on quartz filters (Pallflex,
2500QATUP, Pall Corp., USA). Thick-washers, which are commercially available (MSP Corp., USA), were
installed between the stages on the MOUDI for quartz filter sampling. Due to analytical difficulty resulting from
the small diameter (i.e., 25 mm) and amount of mass, filters with a diameter of 0.056 μm were not applied in
this study. Each filter was carefully cut in a certain amount of sampling-dotted areas using a 1.0 cm2 punch for
OC and EC analysis, so the remaining portions were extracted for WSOC, WSI, and BCA analysis. The detailed
analytical methods can be found in prior publications (Bae and Park, 2013; Park et al., 2013).
Size-resolved Chemical Analysis
Carbonaceous analyses were performed using standardized thermal-optical transmittance (TOT) OC
and EC analyzer (Sunset Laboratory Inc., USA) (Bae et al., 2004a; Bae et al., 2004b). Briefly, the filter was
heated over a series of temperature steps based on the National Institute for Occupational Safety and Health
(NIOSH) protocol. The loaded filter was heated to 840 °C in the presence of pure (oxygen-free) helium gas, and
the organic compounds and pyrolysis products were thermally released from the aerosols on the filter. The
desorbed carbon fragments were oxidized to CO2 by using MnO2 as a catalyst. Next, converted CH4 from CO2
was measured using a flame ionization detector (FID). EC was detected in the He/O2 atmosphere. An internal
standard (i.e., CH4) and external standard (i.e., sucrose) were injected for quantification. The strength of the
laser transmittance can determine the split between OC and EC for each sample analysis. The average recovery
was 1.00 ± 0.02. The instrument blank test, which was conducted by completing the analysis without collecting
aerosols, was performed for a set of sized PM samples and five samples from the 24-hr integrated sampler to
determine the noise and contamination in the instrument and filters. Duplication analysis for PM samples was
performed for every five samples.
To determine the WSOC and WSI species, the remaining portions of the samples after OC and EC
analyses were extracted. The filter extraction process involved placing the filter in previously baked Pyrex glass
vials with 20 mL of milli-Q water each, then placing the vials in an ultrasonic bath at 20 °C controlled using a
low temperature circulator (EYELA, Tokyo Rikakikai Co., Japan) for 120 minutes. All water extracts were
filtered using a 0.45 μm membrane filter (Hydrophilic PTFE 0.45 μm pore size, Advantec, Japan) and were then
analyzed for sodium (Na+), ammonium (NH4+), potassium (K+), calcium (Ca2+), magnesium (Mg2+), chloride
(Cl−), nitrate (NO3−), and sulfate (SO42−) using ion chromatography (Metrohm 883, Switzerland). Metrohm
Metrosep A Supp-5 with eluent of 3.2mM Na2CO3 & 1.0mM NaHCO3 and Metrohm Metrosep C4-250 columns
with 1.7mM HNO3 & 0.7 mM dipicolinic acid were used to analyze the anion and cation compounds,
respectively. The injection volume was of 250 μl, and the method detection limit (MDL) was determined as 3.14
(i.e., pi) times the standard deviation of the lowest detected standard concentrations. The determined MDLs for
Na+, NH4+, K+, Ca2+, Mg2+, Cl−, NO3−, and SO42− were 2, < 1.0, 3, 10, 11, 3, < 1.0, and < 1.0 ppb, respectively.
The sample precision was less than 3.0% for all ionic compounds. The remaining portion of the water extracts
were analyzed using the TOC analyzer (Sievers 900, GE, USA) for WSOC. Briefly, the TOC analyzer measures
the organic carbon content by subtracting the measured carbon content of the two channels. In the total carbon
channel, all organic carbon was transformed into carbon dioxide with ammonium persulfate and ultraviolet light.
In the inorganic carbon channel, the sample remains unaltered. The conductivity of this water through a
semipermeable membrane was measured. However, the inorganic carbon concentration causes higher analytical
uncertainties, so an inorganic carbon remover (ICR) was installed to reduce the interference of inorganic carbon
for the WSOC concentrations. The external sucrose standard was analyzed for quantification, and the average
recovery was 1.00 ± 0.03.
An optical particle sizer (OPS, Model 3330, TSI Inc., USA) was deployed with a heating inlet to
remove the moisture. The OPS measures the particle concentrations in a maximum of 16 size channels and
detects scattered laser-light at a mean scattering angle of 90° with a parabolic mirror. With a high time resolution
and easy operability, it has been extensively used to determine the size-resolved number with converted mass
concentrations (Zhao and Stephens, 2017). However, the ratio of volume to mass (i.e., density) was required to
calculate the size-resolved PM mass concentrations. In this study, we compare 24-hr integrated PM mass using
the Teflon gravity method and volume concentrations to determine the real time size-resolved PM mass
concentrations from OPS. In this study, the PM samples were collected with the 12 optical diameter in the
ranges from 0.3 through 10 μm. A Multi Angle Absorption Photometer (MAAP) (Model 5012, Thermo Fisher
Scientific Inc., USA) was operated using the same heating inlet of OPS. The samples were analyzed
downstream of the PM2.5 impactor (BMI inc., USA). The sample flows through the downtube and deposits onto
the glass fiber filter tape. A 670-nanometer visible light source was then aimed towards the deposited aerosol
and filter tape matrix. The light transmitted into the forward hemisphere and reflected into the back hemisphere
was measured using a series of photo-detectors. The reduction in the light transmission, multiple reflection
intensities, and air sample volume were continuously integrated over the sample run period to provide a real
time data output of optically equivalent black carbon (eBC) measurements (Hitzenberger et al., 2006).
Highly sensitive analysis of underivatized benzene carboxylic acids (BCAs)
An analysis of underivatized BCA is challenging due to the difficulty in their separation as well as the
detection of these small polar analytes. In general, gas chromatography mass spectrometry (GCMS) has been
widely used for BCA quantification while an analytical technique requires extraction sequences using organic
solvent and derivatization due to the analytical limitation for organic acids with a heavy molecular weight (Bae
et al., 2014; Rogge et al., 1997). In this study, the remaining extracts after analysis of the WSOC and WSI
species were combined into a set of size-resolved PM extracts to analyze the underivatized BCAs using
LC/MSMS. The composite extracts were completely dried using a freezing dryer (il-Shin Bio Base, South Korea)
after an internal standard injection of phthalic acid - D4 (quantifier ion; 169.0 m/z). Hydrophilic interaction LC
was carried out using an Eclipse XDB-C18 4.6 mm ID x 150 mm (5μm) column as stationary phase with 10
mM Ammonium Acetate and Acetonitrile in milli-Q water. It was combined with MSMS in the multiple reaction
monitoring (MRM) mode for the separation and the detection of five underivatized BCAs (i.e., phthalic acid
(12BCA, quantifier ion; 165.1 m/z), terephtalic acid (14BCA, 164.8 m/z), trimellitic acid(124BCA, 165.2 m/z),
trimesic acid(135BCA, 209.1 m/z), and pyromellitic acid(1245BCA, 165.0 m/z)). Regression coefficients to
determine the seven point calibrations varied from 0.998 to 0.999. Absolute MDL (i.e., converted atmospheric
concentration) were in the range of 4.61–1.72 pg/m3. A fast screening method of a total 8 minutes of runtime
was developed for use with samples characterized using very low analyte concentrations and complex matrices.
The best analytical conditions in LC/MSMS and the final mass fragment transitions for quantification (i.e.,
quantifier and qualifier ions) are shown in supplemental Tables S1 and S2.
Residence-Time Weighted Concentrations (RTWC)
The RTWC approach has been generally applied into the receptor site with no major local point
sources (Han et al., 2007). Briefly, for residence time Tr (i,j,k) of the grid cells (i,j) hit by a trajectory segment,
ck was the concentration measured upon the arrival of trajectory k. Cij (i.e., µg/m3 in PM2.5 mass) was the
potential source region strength, M was the total number of trajectories k at the grid cells (i,j). The unit of
RTWC was the same as ck (i. i.e., hourly PM2.5 mass). The redistribution for,
𝐶𝑖𝑗 =
∑𝑀
𝑗=1(𝑐𝑘)𝑇𝑟𝑖𝑗𝑘
∑𝑀
𝑗=1 𝑇𝑟𝑖𝑗𝑘
- Eq. (1)
Equation (1) was applied to all individual trajectories. After redistribution, a new concentration field Cij was
computed. The detailed methods were described in a previous study (Zhou et al., 2004). In order to reduce the
effect of small values of nij that result in high RTWC values with high uncertainties, an arbitrary weight
function W(nij) was multiplied into the RTWC value to better reflect the uncertainty in the values for these cells.
The RTWC values were down-weighted when the total number of end points per specific cell was less than
about three times the average value of the end points per cell.
Ion Neutralization Model (INM)
The degree of stoichiometric neutralization for the ensemble of measured particles was obtained from
the normalized ratio of the measured NH4+ concentration (in nmol m-3) to the NH4+ concentration needed for full
neutralization of the anions (Zhang et al., 2007). The ion neutralization model (INM) was frequently utilized to
determine the acidic or neutral state of PM1.0 using an aerosol mass spectrometer (AMS). The INM uses
Equation (2) to calculate the PM neutralization condition,
NH4+ measured / NH4+ neutralized = [NH4+ / 18] / [2 × SO42- / 96 + NO3- / 62 + Cl- / 35.5]
- Eq. (2)
Numerous studies to determine the value of [NH4+ measured] and [NH4+ neutralized] using AMS have
been previously published, including simple slope estimations (Han et al., 2016; Tan et al., 2016; Zhang et al.,
2007).
RESULTS AND DISCUSSION
The measurements were analyzed over different sampling times (i.e., 24 hr integrated PM2.5 mass
using Teflon filter gravity method, 15 minute PM 2.5 volume using OPS, 30 minute eBC using MAAP, and 48 72 hr integrated size-resolved PM using MOUDI). Fig. 1 presents complete time series traces for (a) the
evolution of particle size distributions and particle mass concentrations (d(µg)/dlogdp, l/m3) indicated with
selected Periods (1-4) based on MOUDI sampling periods; (b) PM2.5 massT using the Teflon filter gravity
method and PM2.5 massP using OPS; (c) EC and eBC; (d) WSOC and WIOC, (e) NO3-, SO42-, and NH4+, and (f)
underivatized 12BCA, 14BCA, 124BCA, 135BCA, and 1245BCA in PM 2.5. The four periods (1-4) were
assessed using (1) MOUDI sampling, (2) high plume concentration events, (3) and air mass transport. The
selected periods are June 1 - June 3, June 3 - June 6, June 6 - June 16, and June 16 - June 19 named for period 1,
period 2, period 3, and period 4, respectively. The start and end time was 9:00 am for all selected periods. The
amount of data available for each measurement allows for the initial validation, and data was flagged and
removed from the analysis if any of the following conditions were known to apply: when less than 80% of the
data was available for an hour; when an instrument was malfunctioned or was under maintenance. The data
completeness was over 98% measured by all instrumental measurements of OPS and MAAP. The contour plot
in Fig. 1(a) provides an overview of the evolution of the temporal mass particle size distributions by the OPS in
the size range from 0.3–10.0 μm. The contour plot shows the mass concentration of the particles normalized to
the logarithm particle diameter (logdp) in each size bin (d(µg)/dlogdp, l/m3). To present time and size-resolved
PM mass concentrations from OPS, a slope of 2.01 (r2 of 0.71) determined by the relationship between dailyaveraged PM2.5 volume (vol.) concentrations (d(nL)/dlogdp, l/m3) using OPS (summed over the size range of
0.3–2.5 μm) and 24-hr integrated PM2.5 massT were applied to size-resolved mass concentrations (d(µg)/dlogdp,
l/m3). A pairwise correlation scatterplot (slope of 1.01±0.148 (±standard error)) between daily averaged PM2.5
massP converted from PM2.5 vol. and 24 hour integrated PM2.5 massT was shown in supplemental Fig. S2(a). The
characteristic feature of the data sets does not present a daily, strong diurnal pattern, which implies that the
sampling sites were not affected by strong local sources. The higher particle concentration events can be
associated with regional and/or middle range transport, and the size range of 0.3–0.5 μm accounts for about 41.8%
of the total concentrations in the size range of 0.3–10.0 μm. Table 1 shows the statistical results for the
concentration during the sampling period. The average concentration of PM2.5 massT, PM2.5 volume, and PM2.5
massP are 24.3 ± 2.3 μg/m3 (average ± standard error), 12.3 ± 1.6 nL/m3, and 24.8 ± 3.3 μg/m3, respectively. The
overall concentrations of PM2.5 massT and PM2.5 massP are statistically indistinguishable with a 95% confidence
level. The ratio of eBC to PM 2.5 massT over the whole sampling period is of 0.045 ± 0.003. The corresponding
concentration of eBC (1.09 ± 0.12 μg/m3), which is a tracer for combustion-generated aerosols, is about 24
times lower than the PM2.5 mass concentration. A pairwise scatterplot between 24-hr averaged eBC and EC was
shown in supplemental Fig. S2(b). A fair agreement (r2 of 0.74) was presented with a slope of 7.11±0.98 (±
standard error). The large differences in the concentration between eBC and EC can be associated with the
different chemical characteristics by the sources (Bae et al., 2007). A previous study indicated that BC measured
using an aethalometer was about five times higher than thermal EC during haze episodes (Jeong et al., 2004).
Detailed investigations of the relationship between eBC and EC through chemical components from different
origins will be discussed in a further analysis.
Table 2 presents the size distribution (i.e., 10 different aerodynamic cutpoint diameters using MOUDI)
of all species for size-resolved OC, EC, WSOC, WSI concentrations. A previous size-resolved PM study (Xu et
al., 2016) showed that organic aerosol (OA) accounts for 51%, followed by NO3- (17%), SO42- (13%), NH4+
(10%), BC (6%), and Cl- (3%), which peaked at about 0.5 μm for the wintertime in a highly polluted city of
China. Xu et al. (2015) found less OA fraction in an unpolluted area, such as a remote high-altitude (4180 m)
location in the northeast of the Tibetan Plateau, China (e.g., SO42- (36%), OC (18%), NO3- (17%), NH4+ (10%),
Ca2+ (6.6%), and EC (2.6%) in PM2.5 for summertime). However, the size distributions of chemical components
were different. The size distributions of SO42- and NH4+ peaked in the accumulation mode (0.32–0.56 μm) while
those of both NO3- and Ca2+ peaked in the range of 1.8–3.2 μm (Xu et al., 2015). In this study, the sum of the
species presents a prominent accumulation mode peak at a diameter of 0.56 μm without the coarse mode peak.
In the accumulation mode, SO42- (49.3%) dominated over the composition of particles in the size range of 0.1–
1.8 μm, followed by NH4+ (19.5%) and NO3- (13.6%). The OC and EC account for 13.5% and 0.4%,
respectively. However, OC (37.5%) dominates at smaller than 0.18 μm, with SO42- contributing 29.6%. In the
coarse mode, NO3-, the main contributor, accounts for 34.0% of the particle mass in the size range of 3.2–18.0
μm, followed by SO42- (16%), OC (15%), NH4+ (5%), the rest of the species in total accounted for 64% in this
size range. For most samples analyzed via ion chromatography, the r2 between the averaged NH4+ and SO42- and
that between NH4+ and NO3- in the size range of 0.1–18 μm were 0.999 and 0.71, respectively. These results
confirm that the major compounds are (NH4)2SO4 in all size ranges. NH4+ and SO42- account for 43.9% for all
size ranges. The highest NH4+ and SO42- mass fraction presents 73.7% at diameters of 1.0 μm, followed by 58.6%
at diameters of 1.8 μm. The equivalent ratio of SO42- to NH4+ equals 0.97. This indicates an incomplete
neutralization of SO42- due to the concurrent presence of (NH4)HSO4 and other hydrogen-containing ion
compounds. The different size distributions for different compound ratios suggest that they had different sources
and/or have undergone atmospheric transformation processes. The species that are relatively abundant in the
accumulation mode aerosols are mainly secondary species, including NH4+, SO42-, and OC, while the species
that are relatively abundant in the coarse mode aerosols could mainly be primary mineral ionic species such as
Ca2+, Na+, and Cl-. NO3- is also closely associated with dust particles as a result of its formation through the
reactions of HNO3 gas with carbonate salts (e.g., calcite and dolomite) (Sullivan et al., 2009) which could form
Ca(NO3)2 and Mg(NO3)2 (Li and Shao, 2009).
To identify the potential locations of these regional sources of the PM 2.5 mass, RTWC was utilized
along the backward trajectory analyses and hourly PM2.5 massP. Fig. 2 presents the result of residence-time
weighted concentrations (RTWC) for sampling periods 1, 2, 3, and 4. The times selected for the RTWC analysis
are those with start and end times for each selected sampling period. The numbers using the back trajectories
totaled 48, 72, 480, and 72 for periods 1, 2, 3, and 4, respectively. The contour plots in Fig. 2 present the RTWC
results. A high RTWC value in a grid cell indicates a potential source area to the sampling site. The RTWC plots
for periods 1 – 4 show the different high potential PM2.5 source areas in the middle and middle-west regions of
Korea. For period 1, the RTWC shows relatively intense smaller area, which can be a local impact, while period
2 presents scatter and wider areas from the receptor site. For period 3, the RTWC is somewhat different from the
other periods, with a higher probability in the south-west regions of Korea, including the Yellow Sea. For period
4, the southern region of the Yellow Sea and eastern China can be considered as high-potential source areas to
the receptor site. The concentrations of aerosol chemical components during periods 1 and 2 are related to the
regional sources, while those during periods 3 and 4 are related to the southwesterly winds with the interaction
of secondary production and sea salts.
To investigate the size-resolved chemical characteristics for the selected periods, Fig. 3 shows the
average size distributions of the mass concentrations of WSOC, WIOC, EC, and WSIs (including and excluding
NH4+ and SO42-), and the ratios of (WIOC+EC) to (WSOC+NH4++NO3-+SO42-) and (Na++Cl-) to
(WSOC+NH4++NO3-+SO42-) in different size bins are shown in Fig. 3. The results statistical for the
concentration of the selected sampling periods 1 – 4 are presented in Table 3. The relative fractional
contributions of the individual species in different size bins are shown in Supplemental Fig. S3 as well. For
period 1, the high PM2.5 mass (36.3 ± 20.4 μg/m3) (average ± standard deviation), OC (4.03 ± 2.02 μg/m3), and
BCAs (5.12 ± 1.65 ng/m3) was analyzed compared to the other periods. The sum of the species presents a
prominent accumulation mode peak at a diameter of 0.56 μm without a clear coarse mode peak. WSOC (28%)
dominates the composition of the particles at a diameter of 0.1 μm in the accumulation mode and then gradually
decreases into the coarse mode. The average mass fraction in the accumulation mode in the size range of 0.1–1.8
μm is 17%, while that in the coarse mode of the size range of 3.2–18.0 μm is 10%. WIOC accounts for 8.4% and
8.5% in the accumulation mode and in the coarse mode, respectively, and shows no clear statistically significant
differences between the accumulation and coarse mode. EC (1.3%) dominates in the size range of 3.2–18.0 μm
in the coarse mode rather than in the accumulation mode (0.69%). SO42- (40.6%) and NH4+ (16.2%) in the
accumulation mode present the most dominant components in the PM while NO3- (38%) in the coarse mode
shows the most dominant component in the PM. This suggests different secondary inorganic aerosol formation
pathways between nitrate and sulfate, supporting the possibility of nitrate aerosol formation through a reaction
of the precursor (HNO3) with the carbonate salts in the coarse mode (Sullivan et al., 2009). SO42- (15%) and
NH4+ (4%) in the coarse mode showed a relatively smaller fraction in the PM compared to the accumulation
mode. Cl- and Na+ clearly dominate, accounting for 2.3% and 10.9% in the coarse mode of the size range from
3.2–18.0 μm, respectively, implying a marine source. In addition, K+, Ca2+, and Mg2+ dominate in the coarse
mode, accounting for 1.2%, 1.1%, and 6.9%, respectively. Unlike period 1, the sum of the species for period 2
shows a bimodal distribution at two modes: an accumulation mode peak at a diameter of 0.56 μm and the coarse
mode peak at a diameter of 3.2 μm. Higher concentrations of NO3-, WIOC, and EC lead to a clear bi-mode
distribution in the coarse mode. Both WSOC (16%) and WIOC (11%) dominate in the accumulation mode. EC
(2.7%) dominates in the coarse mode of the size range of 3.2–18.0 μm rather than in the accumulation mode
(1.1%). Stone et al. (2011) argued that carbonaceous aerosol (OC and EC) that are typically found in fine
particles can be included in coarse particles during a dust event, being internally mixed with dust (Stone et al.,
2011). Therefore, EC domination in the coarse mode rather than in the accumulation mode might be due in large
part to dust source contributions. SO42- (31%) and NH4+ (15%) in the accumulation mode present the most
dominant components. NO3- (40%) in the coarse mode shows the most dominant components in the PM. Cl-,
Na+, K+, Ca2+, and Mg2+ clearly dominate in the coarse mode of the size range of 3.2–18.0 μm accounting for
4.5%, 11.2%, 1.5%, 1.2, and 6.3%, respectively. For period 3, the relatively low PM2.5 mass (19.8 ± 6.3 μg/m3),
OC (1.1 ± 0.6 μg/m3), and BCAs (2.4 ± 0.7 ng/m3) were analyzed (Table 3). The sum of the species presents a
prominent accumulation mode peak at a diameter of 0.56 μm without the coarse mode as well. WSOC (15%)
dominates the composition of the particles both in the accumulation mode and in the coarse mode. EC (1.32%)
dominates in the coarse mode for the size range of 3.2–18.0 μm, the same as in period 1. SO42- (46%) and NH4+
(17%) are the most dominant components in the accumulation mode. NO3- (28%) was the most dominant
component in the coarse mode. For period 4, WSOC (12%) dominates over the composition of particles both on
the accumulation mode and in the coarse mode. WIOC accounts for 5.2% and 1.7% in the accumulation mode
and in the coarse mode, respectively. EC (0.73%) dominates in the coarse mode for a size range of 3.2–18.0 μm,
which had a slightly higher concentration than the concentration in the accumulation mode (0.34%). For period
4, the southern region (i.e., Mokpo and Jeju) of Korea and east of China (Yancheng and Xuzhou areas) can be
responsible for high potential areas by the high RTWC value. Excluding SO42- and NH4+, a clear bimodal
distribution was presented at the diameter of 0.56 μm and 3.2 μm for all selected periods (Fig. 3). This means
that the overall size distributions for either mono- or bi-mode can be determined by these two major ion
compounds.
The size-resolved ratios of (WIOC+EC) to (WSOC+NH4++NO3-+SO42-) showed bimodal distributions
at 0.1-0.18 μm for the accumulation mode and 3.2-5.6 μm for the coarse mode. The ratios for the accumulation
mode and coarse mode peaks can be related to primary fresh sources and dust, respectively. The ratios in periods
1 and 2 showed relatively higher values than in periods 3 and 4. Following the results for the RTWC, regional
sources can more heavily contribute to the sampling sites than in periods 3 and 4. The size-resolved ratios of
(Na++Cl-) to (WSOC+NH4++NO3-+SO42-) provide information on the marine origin. These ratios in period 3 and
4 certainly showed higher values (> 0.4) in the coarse mode than in periods 1 and 2, while relatively lower ratios
of (WIOC+EC) to (WSOC+NH4++NO3-+SO42-) were presented. This indicates that, for these periods, the sizeresolved chemical components were affected significantly by marine sources.
The results of the RTWC indicated that secondary production and sea salt in the south-west regions of
Korea, including the Yellow Sea and eastern China for periods 3 and 4 might have significantly affected the size
distribution of the aerosol components. The size-resolved chemical characteristics of the marine sources with
minimal anthropogenic source contribution were investigated. One MOUDI dataset (from June 13 to Jun 16 in
period 3) was chosen by (1) eBC concentrations of less than 1.0 µg/m3 and (2) wind direction of western marine
side (225-315o). The time series of the eBC with selected conditions and the size distributions of the mass
concentrations of OC, EC, and WSIs, including and excluding NH4+ and SO42-, are shown in supplemental Fig.
S4. The size distributions, excluding NH4+ and SO42-, showed a clear bi-modal distribution. The concentration in
the coarse mode showed a higher concentration than in the accumulation mode. Cl- and Na+ clearly dominate in
the coarse mode, which confirms the marine source. In this period, low PM2.5 mass (17.26 µg/m3) and organic
carbon (0.64 µg/m3) were measured, and the higher ratio of WSOC to OC was 0.91. The marine-originated
organic acids (e.g., hydroxy fatty acids, free amino acids, methanesulfonic acid, etc.) can result in an increase in
the ratio of WSOC to OC (Kouvarakis and Mihalopoulos, 2002; Mandalakis et al., 2011; Tyagi et al., 2015).
Four-day averaged Chlorophyll-a information using data from Geostationary Ocean Color Imager (GOCI) (i.e.,
launch on June 27, 2010 from South Korea with an hourly repeat cycle and 500 m spatial resolution) are shown
in supplemental Fig. S5 (Ruddick et al., 2014). These images indicate that the OC concentration in the
accumulation and coarse modes during the sampling periods (period 3) for the marine sector might be correlated
to a relatively high chlorophyll-a concentration in the coastal area of the Yellow sea, which could be one of the
WSOC marine aerosol sources contributing to the sampling site (Figs. S4).
Furthermore, burning large amounts of coal and biomass in East Asia (e.g., east areas of China) adds
more anthropogenic aerosols, which alter the chemical composition of aerosols in the remote Yellow Sea
atmosphere (Bae et al., 2014). Many organic compounds are photochemically oxidized, producing organic
aerosol during long-range transport. Bae et al. (2014) showed that organic molecular markers (e.g., levoglucosan,
PAHs, and BCAs) analyzed via GCMS were certainly influenced by the long-range transport of emissions from
biomass burning on the Asian continent. Water-soluble diacids with a low molecular weight have previously
been studied (Deshmukh et al., 2016), whereas the size-resolved characteristics of BCAs have not been studied
in the western North Pacific Rim. Although only one overall size distribution of BCAs was displayed in this
study due to the limited amount of sample composite extracts, as was discussed in the methods, this figure is
reported for the first time. The sum of the BCAs indicates the accumulation mode peak at the diameter of 0.56
μm without a coarse mode, as shown in Fig. 4. 12BCA accounts for 40% and 54% in the accumulation mode
and in the coarse mode, respectively. 14BCA (18%) dominated in the accumulation mode, which indicates a
slightly higher concentration than the concentration in the coarse mode (13.8%). 124BCA (11.4%), 135 BCA
(17.7%), and 1245BCA (12.6%) are presented in the accumulation mode in the PM. Based on a good agreement
(r2=0.98) with size-resolved BCAs and WSOC, the size-resolved chemical characteristics of WSOC implies a
common origin, and detailed examinations will be shown in a further study.
Fig. 5 shows pairwise scatterplots for four periods between µeq. NH4+ and µeq. SO4+ colored by logdp.
Although the correlation of determinations (r2) between μeq. NH4+ and μeq. SO42- are 0.999, 0.998, 0.999, and
0.997 for periods 1 – 4, respectively, the slopes are statistically different. Incomplete neutralization can be due to
the concurrent presence of other hydrogen-containing ion compounds. Fig. 6 shows size-resolved charge
equivalent concentration ratios between all measured cations and anions and the INM using Eq. (2) for the
selected periods (1-4) with uncertainties. The uncertainty was calculated using the error propagation method
(Bae et al., 2006). Briefly, the MDL and instrumental uncertainty (10%) were considered to obtain the final
uncertainties. On the basis of a theoretical definition, the aerosols appear to be neutralized by ionic balance for
all size ranges. However, there were ‘‘more acidic’’ conditions observed by the INM for both the accumulation
and coarse modes for all periods. The ratio calculated by the charge equivalent concentration showed around
unity, considering uncertainties over the accumulation mode and the coarse mode. This indicates that these
particles are close to neutral or toward basic. Compared to those considered as charge equivalent concentration
ratios, the ‘‘more acidic’’ particles calculated by the INM are determined by the exclusive increase in the SO42and NH4+. However, the concentrations of Na+ and other cations cannot be disregard in the accumulation mode.
In addition, the enhancement of the organic acid concentrations can also be attributed to the acid-catalyzed
secondary organic aerosol formation for the neutralization. Therefore, many previous studies based on the INM
have higher uncertainties. In such a case, more organic acids can be expected since (1) heterogeneous reactions
consume organic acids more rapidly from smaller particles due to their higher surface area-to-volume ratios, and
(2) organic acids are likely to be enriched due to their greater variety of origin compared to sea-salt particles.
The very probable presence of organic acids, such as the BCAs in the submicron particles, indicates that these
particles are expected to be more neutral. Organic acid role becomes increasingly important downwind from
polluted areas where acidic submicron particles have limited capacity to react with acidic trace gases. The high
charge equivalent concentration ratios in the accumulation mode could be due to the presence of high watersoluble organic carbon (e.g., BCAs) in our chemical analysis.
Fig. 7 compares the OPS to the MOUDI on the average size distributions of the particle mass during
the sampling periods. Our approach is based on the fact that agreement between OPS mass distributions is a
convenient measurement of the mass concentrations. The average mass distributions obtained by the OPS and
the MOUDI agree quite well for the coarse mode, but at the accumulation mode, the MOUDI shifts into larger
size ranges when compared to OPS. Such a discrepancy was observed for all periods when comparing the OPS
and MOUDI distributions because (1) the OPS only measured the scattering (e.g., EC measured by absorption);
(2) responds differently to the particle shape as the particle becomes less spherical, which can increase the
uncertainty; and (3) the vacuum cutpoint diameter measured by MOUDI vs. the optical intensity diameter by
OPS. The ratio of the organic mass (OM) to OC and unknown oxidized trace elements for MOUDI can be
additional reasons. This effect leads to differences in the accumulation mode between the OPS and MOUDI due
to irregular particles that appear larger than they really are (and thus are much more apparent than their real
number). This comparison is consistent with the differences between the concentrations of particle mass
obtained via the OPS and MOUDI. In general, these two instruments agree on the size distributions over the size
ranges.
CONCLUSIONS
This study characterized the size-resolved main chemical compositions of aerosol particles (OC,
WSOC, EC, WSI, and BCA) using MOUDI on the western coast of Korea. The potential source origins of PM
during 4 different periods were determined using residence-time weighted concentrations (RTWC) analysis. As
a result, the sum of the species presents a prominent accumulation mode peak at a diameter of 0.56 μm, without
the coarse mode peak. In the accumulation mode, SO42- (49.29%) dominates the composition of the particles in
the size range of 0.1–1.8 μm, followed by NH4+ (19.53%) and NO3- (13.64%). OC and EC account for 13.49%
and 0.39%, respectively. 1,4-BCA (18%) dominates in the accumulation mode, which is a little higher than the
concentration in the coarse mode (13.8%). The differences between the size-resolved charge equivalent
concentration ratios and the ion neutralization model (INM) can considered to be related to heterogeneous
reactions that consume organic acids and enriched organic acids from various origins. Finally, the reasons for
the differences in the average size distributions of the particle mass using the OPS and the MOUDI
measurements are that 1) the OPS only measures scattering (e.g., EC measured by absorption); 2) the two
methods respond differently to the particle shape as the particle becomes less spherical, which can increase the
uncertainty; and 3) the vacuum cutpoint diameter measured by MOUDI vs. the optical intensity diameter by
OPS. The information and knowledge collected in this study can inform the implementation of the size-resolved
aerosol problem to source identification and/or to determine health effects. Finally, the results gained form the
detailed size-resolved chemical analysis with an organic molecular marker analysis, which was used to extract
and interpret additional carbon fractions data set, improves the understanding of sources and helps determine
their origin, which is important for future studies.
ACKOWLEDGMENT
We acknowledge the support provided by the National Research Foundation of Korea (NRF) (NRF2015R1A2A1A10053971).
References
Al-Naiema, I.M., Stone, E.A., 2017. Evaluation of anthropogenic secondary organic aerosol tracers from
aromatic hydrocarbons. Atmos. Chem. Phys. 17, 2053-2065.
Bae, M.-S., Demerjian, K.L., Schwab, J.J., 2006. Seasonal estimation of organic mass to organic carbon in
PM2.5 at rural and urban locations in New York state. Atmospheric Environment 40, 7467-7479.
Bae, M.-S., Hong, C.-S., Kim, Y.J., Han, J.-S., Moon, K.-J., Kondo, Y., Komazaki, Y., Miyazaki, Y., 2007.
Intercomparison of two different thermal-optical elemental carbons and optical black carbon during ABCEAREX2005. Atmospheric Environment 41, 2791-2803.
Bae, M.-S., Park, S.-S., 2013. Thermal Distribution of Size-resolved Carbonaceous Aerosols and Water Soluble
Organic Carbon in Emissions from Biomass Burning. Asian Journal of Atmospheric Environment 7, 95-104.
Bae, M.-S., Schauer, J.J., DeMinter, J.T., Turner, J.R., 2004a. Hourly and Daily Patterns of Particle-Phase
Organic and Elemental Carbon Concentrations in the Urban Atmosphere. J Air Waste Manag Assoc 54, 823-833.
Bae, M.-S., Schauer, J.J., DeMinter, J.T., Turner, J.R., Smith, D., Cary, R.A., 2004b. Validation of a semicontinuous instrument for elemental carbon and organic carbon using a thermal-optical method. Atmospheric
Environment 38, 2885-2893.
Bae, M.S., Shin, J.S., Lee, K.Y., Lee, K.H., Kim, Y.J., 2014. Long-range transport of biomass burning emissions
based on organic molecular markers and carbonaceous thermal distribution. Sci Total Environ 466-467, 56-66.
Bougiatioti, A., Zarmpas, P., Koulouri, E., Antoniou, M., Theodosi, C., Kouvarakis, G., Saarikoski, S., Mäkelä,
T., Hillamo, R., Mihalopoulos, N., 2013. Organic, elemental and water-soluble organic carbon in size segregated
aerosols, in the marine boundary layer of the Eastern Mediterranean. Atmospheric Environment 64, 251-262.
Cavalli, F., Facchini, M.C., Decesari, S., Mircea, M., Emblico, L., Fuzzi, S., Ceburnis, D., Yoon, Y.J., O'Dowd,
C.D., Putaud, J.P., Dell'Acqua, A., 2004. Advances in characterization of size-resolved organic matter in marine
aerosol over the North Atlantic. Journal of Geophysical Research: Atmospheres 109, n/a-n/a.
Degrendele, C., Okonski, K., Melymuk, L., Landlová, L., Kukučka, P., Audy, O., Kohoutek, J., Čupr, P.,
Klánová, J., 2016. Pesticides in the atmosphere: a comparison of gas-particle partitioning and particle size
distribution of legacy and current-use pesticides. Atmos. Chem. Phys. 16, 1531-1544.
Deshmukh, D.K., Kawamura, K., Lazaar, M., Kunwar, B., Boreddy, S.K.R., 2016. Dicarboxylic acids, oxoacids,
benzoic acid, α-dicarbonyls, WSOC, OC, and ions in spring aerosols from Okinawa Island in the western North
Pacific Rim: size distributions and formation processes. Atmos. Chem. Phys. 16, 5263-5282.
Gantt, B., Meskhidze, N., Carlton, A.G., 2010. The contribution of marine organics to the air quality of the
western United States. Atmos. Chem. Phys. 10, 7415-7423.
Gong, X., Zhang, C., Chen, H., Nizkorodov, S.A., Chen, J., Yang, X., 2016. Size distribution and mixing state of
black carbon particles during a heavy air pollution episode in Shanghai. Atmos. Chem. Phys. 16, 5399-5411.
Ham, W.A., Kleeman, M.J., 2011. Size-resolved source apportionment of carbonaceous particulate matter in
urban and rural sites in central California. Atmospheric Environment 45, 3988-3995.
Han, Y.-J., Holsen, T.M., Hopke, P.K., 2007. Estimation of source locations of total gaseous mercury measured
in New York State using trajectory-based models. Atmospheric Environment 41, 6033-6047.
Han, Y., Stroud, C.A., Liggio, J., Li, S.M., 2016. The effect of particle acidity on secondary organic aerosol
formation from α-pinene photooxidation under atmospherically relevant conditions. Atmos. Chem. Phys. 16,
13929-13944.
Hitzenberger, R., Petzold, A., Bauer, H., Ctyroky, P., Pouresmaeil, P., Laskus, L., Puxbaum, H., 2006.
Intercomparison of Thermal and Optical Measurement Methods for Elemental Carbon and Black Carbon at an
Urban Location. Environmental Science & Technology 40, 6377-6383.
Ho, K.F., Ho, S.S.H., Lee, S.C., Kawamura, K., Zou, S.C., Cao, J.J., Xu, H.M., 2011. Summer and winter
variations of dicarboxylic acids, fatty acids and benzoic acid in PM<sub>2.5</sub> in Pearl Delta River Region,
China. Atmos. Chem. Phys. 11, 2197-2208.
Jeon, W., Choi, Y., Lee, H.W., Lee, S.-H., Yoo, J.-W., Park, J., Lee, H.-J., 2015. A quantitative analysis of grid
nudging effect on each process of PM2.5 production in the Korean Peninsula. Atmospheric Environment 122,
763-774.
Jeong, C.-H., Hopke, P.K., Kim, E., Lee, D.-W., 2004. The comparison between thermal-optical transmittance
elemental carbon and Aethalometer black carbon measured at multiple monitoring sites. Atmospheric
Environment 38, 5193-5204.
Kautzman, K.E., Surratt, J.D., Chan, M.N., Chan, A.W.H., Hersey, S.P., Chhabra, P.S., Dalleska, N.F., Wennberg,
P.O., Flagan, R.C., Seinfeld, J.H., 2010. Chemical Composition of Gas- and Aerosol-Phase Products from the
Photooxidation of Naphthalene. The Journal of Physical Chemistry A 114, 913-934.
Kawamura, K., Kaplan, I.R., 1987. Motor exhaust emissions as a primary source for dicarboxylic acids in Los
Angeles ambient air. Environmental Science & Technology 21, 105-110.
Kim, K.H., Sekiguchi, K., Kudo, S., Sakamoto, K., Hata, M., Furuuchi, M., Otani, Y., Tajima, N., 2010.
Performance Test of an Inertial Fibrous Filter for Ultrafine Particle Collection and the Possible Sulfate Loss
when Using an Aluminum Substrate with Ultrasonic Extraction of Ionic Compounds. Aerosol and Air Quality
Research.
Kim, W., Doh, S.-J., Yu, Y., 2012. Asian dust storm as conveyance media of anthropogenic pollutants.
Atmospheric Environment 49, 41-50.
Kleeman, M.J., Robert, M.A., Riddle, S.G., Fine, P.M., Hays, M.D., Schauer, J.J., Hannigan, M.P., 2008. Size
distribution of trace organic species emitted from biomass combustion and meat charbroiling. Atmospheric
Environment 42, 3059-3075.
Kleindienst, T.E., Jaoui, M., Lewandowski, M., Offenberg, J.H., Docherty, K.S., 2012. The formation of SOA
and chemical tracer compounds from the photooxidation of naphthalene and its methyl analogs in the presence
and absence of nitrogen oxides. Atmos. Chem. Phys. 12, 8711-8726.
Kleindienst, T.E., Jaoui, M., Lewandowski, M., Offenberg, J.H., Lewis, C.W., Bhave, P.V., Edney, E.O., 2007.
Estimates of the contributions of biogenic and anthropogenic hydrocarbons to secondary organic aerosol at a
southeastern US location. Atmospheric Environment 41, 8288-8300.
Kouvarakis, G., Mihalopoulos, N., 2002. Seasonal variation of dimethylsulfide in the gas phase and of
methanesulfonate and non-sea-salt sulfate in the aerosols phase in the Eastern Mediterranean atmosphere.
Atmospheric Environment 36, 929-938.
Krudysz, M.A., Froines, J.R., Fine, P.M., Sioutas, C., 2008. Intra-community spatial variation of sizefractionated PM mass, OC, EC, and trace elements in the Long Beach, CA area. Atmospheric Environment 42,
5374-5389.
Li, J., Wang, X., Chen, J., Zhu, C., Li, W., Li, C., Liu, L., Xu, C., Wen, L., Xue, L., Wang, W., Ding, A.,
Herrmann, H., 2017. Chemical composition and droplet size distribution of cloud at the summit of Mount Tai,
China. Atmos. Chem. Phys. Discuss. 2017, 1-21.
Li, W.J., Shao, L.Y., 2009. Observation of nitrate coatings on atmospheric mineral dust particles. Atmos. Chem.
Phys. 9, 1863-1871.
Mader, B.T., Schauer, J.J., Seinfeld, J.H., Flagan, R.C., Yu, J.Z., Yang, H., Lim, H.-J., Turpin, B.J., Deminter,
J.T., Heidemann, G., Bae, M.S., Quinn, P., Bates, T., Eatough, D.J., Huebert, B.J., Bertram, T., Howell, S., 2003.
Sampling methods used for the collection of particle-phase organic and elemental carbon during ACE-Asia.
Atmospheric Environment 37, 1435-1449.
Maji, S., Ahmed, S., Siddiqui, W.A., Ghosh, S., 2017. Short term effects of criteria air pollutants on daily
mortality in Delhi, India. Atmospheric Environment 150, 210-219.
Mandalakis, M., Apostolaki, M., Tziaras, T., Polymenakou, P., Stephanou, E.G., 2011. Free and combined amino
acids in marine background atmospheric aerosols over the Eastern Mediterranean. Atmospheric Environment 45,
1003-1009.
O'Dowd, C.D., Facchini, M.C., Cavalli, F., Ceburnis, D., Mircea, M., Decesari, S., Fuzzi, S., Yoon, Y.J., Putaud,
J.P., 2004. Biogenically driven organic contribution to marine aerosol. Nature 431, 676-680.
Park, S.-S., Sim, S.Y., Bae, M.-S., Schauer, J.J., 2013. Size distribution of water-soluble components in
particulate matter emitted from biomass burning. Atmospheric Environment 73, 62-72.
Plaza, J., Pujadas, M., Gómez-Moreno, F.J., Sánchez, M., Artíñano, B., 2011. Mass size distributions of soluble
sulfate, nitrate and ammonium in the Madrid urban aerosol. Atmospheric Environment 45, 4966-4976.
Rogge, W.F., Hildemann, L.M., Mazurek, M.A., Cass, G.R., Simoneit, B.R.T., 1997. Sources of Fine Organic
Aerosol. 8. Boilers Burning No. 2 Distillate Fuel Oil. Environmental Science & Technology 31, 2731-2737.
Ruddick, K., Neukermans, G., Vanhellemont, Q., Jolivet, D., 2014. Challenges and opportunities for
geostationary ocean colour remote sensing of regional seas: A review of recent results. Remote Sensing of
Environment 146, 63-76.
Sahani, M., Zainon, N.A., Wan Mahiyuddin, W.R., Latif, M.T., Hod, R., Khan, M.F., Tahir, N.M., Chan, C.-C.,
2014. A case-crossover analysis of forest fire haze events and mortality in Malaysia. Atmospheric Environment
96, 257-265.
Salma, I., Füri, P., Németh, Z., Balásházy, I., Hofmann, W., Farkas, Á., 2015. Lung burden and deposition
distribution of inhaled atmospheric urban ultrafine particles as the first step in their health risk assessment.
Atmospheric Environment 104, 39-49.
Shim, C., Lee, J., Wang, Y., 2013. Effect of continental sources and sinks on the seasonal and latitudinal gradient
of atmospheric carbon dioxide over East Asia. Atmospheric Environment 79, 853-860.
Stone, E.A., Yoon, S.-C., Schauer, J.J., 2011. Chemical characterization of fine and coarse particles in Gosan,
Korea during springtime dust events. Aerosol and Air Quality Resarch 11, 31-43.
Sullivan, R.C., Moore, M.J.K., Petters, M.D., Kreidenweis, S.M., Roberts, G.C., Prather, K.A., 2009. Timescale
for hygroscopic conversion of calcite mineral particles through heterogeneous reaction with nitric acid. Physical
Chemistry Chemical Physics 11, 7826-7837.
Tan, J., Duan, J., Zhen, N., He, K., Hao, J., 2016. Chemical characteristics and source of size-fractionated
atmospheric particle in haze episode in Beijing. Atmospheric Research 167, 24-33.
Tang, G., Zhao, P., Wang, Y., Gao, W., Cheng, M., Xin, J., Li, X., Wang, Y., 2017. Mortality and air pollution in
Beijing: The long-term relationship. Atmospheric Environment 150, 238-243.
Tyagi, P., Ishimura, Y., Kawamura, K., 2015. Hydroxy fatty acids in marine aerosols as microbial tracers: 4-year
study on β- and ω-hydroxy fatty acids from remote Chichijima Island in the western North Pacific. Atmospheric
Environment 115, 89-100.
Wu, Y., Wang, X., Tao, J., Huang, R., Tian, P., Cao, J., Zhang, L., Ho, K.F., Zhang, R., 2016. Size distribution
and source of black carbon aerosol in urban Beijing during winter haze episodes. Atmos. Chem. Phys. Discuss.
2016, 1-25.
Xu, J., Shi, J., Zhang, Q., Ge, X., Canonaco, F., Prévôt, A.S.H., Vonwiller, M., Szidat, S., Ge, J., Ma, J., An, Y.,
Kang, S., Qin, D., 2016. Wintertime organic and inorganic aerosols in Lanzhou, China: sources, processes, and
comparison with the results during summer. Atmos. Chem. Phys. 16, 14937-14957.
Xu, J.Z., Zhang, Q., Wang, Z.B., Yu, G.M., Ge, X.L., Qin, X., 2015. Chemical composition and size distribution
of summertime PM<sub>2.5</sub> at a high altitude remote location in the northeast of the Qinghai–Xizang
(Tibet) Plateau: insights into aerosol sources and processing in free troposphere. Atmos. Chem. Phys. 15, 50695081.
Yeatman, S.G., Spokes, L.J., Jickells, T.D., 2001. Comparisons of coarse-mode aerosol nitrate and ammonium at
two polluted coastal sites. Atmospheric Environment 35, 1321-1335.
Zhang, Q., Jimenez, J.L., Worsnop, D.R., Canagaratna, M., 2007. A Case Study of Urban Particle Acidity and Its
Influence on Secondary Organic Aerosol. Environmental Science & Technology 41, 3213-3219.
Zhang, X., McMurry, P.H., 1992. Evaporative losses of fine particulate nitrates during sampling. Atmospheric
Environment. Part A. General Topics 26, 3305-3312.
Zhao, H., Stephens, B., 2017. Using portable particle sizing instrumentation to rapidly measure the penetration
of fine and ultrafine particles in unoccupied residences. Indoor Air 27, 218-229.
Zhou, L., Hopke, P.K., Liu, W., 2004. Comparison of two trajectory based models for locating particle sources
for two rural New York sites. Atmospheric Environment 38, 1955-1963.
Table 1. Statistical results of the concentration of PM2.5 massT, PM2.5 volume, PM2.5 massP, equivalent Black
Carbon (eBC), and ratio of eBC to PM2.5 mass for the sampling period.
95%
Standard
Average
Standard
Median
Error
Minimum
Maximum
Confidenc
Deviation
e
PM2.5 massT 1)
24.31
2.26
21.29
9.58
9.80
50.71
4.76
PM2.5 vol. 2)
12.30
1.62
10.15
6.88
4.71
29.37
3.42
PM2.5 massP 3)
24.75
3.26
20.43
13.85
9.48
59.13
6.89
1.09
0.12
1.11
0.46
0.39
1.99
0.23
0.045
0.003
0.042
0.012
0.025
0.067
0.006
eBC 4)
eBC/PM2.5
massT 5)
1) 24 hour integrated PM2.5 mass concentration using Teflon filter gravity method (unit: μg/m3)
2) Daily averaged PM2.5 volume concentration using optical particle sizer (OPS, Σ dp 0.3-2.5 μm) (unit: nL/m3)
3) Daily averaged PM2.5 mass concentration using optical particle sizer (OPS, Σ dp 0.3-2.5 μm) (unit: μg/m3)
4) Daily averaged PM2.5 equivalent Black Carbon concentration using Multi Angle Absorption Photometer (MAAP)
5) Daily averaged ratio of eBC to PM2.5 massT
Table 2. Statistical summary of OC, EC, WSOC, anion, and cation concentrations size-resolved at aerodynamic cutpoint diameters (dp) (i.e., 0.1, 0.18, 0.32, 0.56, 1.0, 1.8,
3.2, 5.6, 10.0, and 18.0 μm) using MOUDI.
dp(μm)
0.1
0.18
0.32
0.56
1.0
1.8
3.2
5.6
10
18
0.285±0.077
0.515±0.245
0.896±0.410
0.497±0.179
0.302±0.095
0.356±0.147
0.255±0.120
0.095±0.029
0.040±0.018
0.005±0.004
0.006±0.002
0.013±0.005
0.021±0.010
0.015±0.005
0.020±0.011
0.040±0.015
0.039±0.035
0.009±0.006
0.003±0.003
0.181±0.062
0.239±0.121
0.356±0.148
0.769±0.390
0.514±0.220
0.240±0.189
0.188±0.122
0.174±0.150
0.083±0.203
0.064±0.182
Cl- (μg/m3)
0.014±0.009
0.011±0.004
0.013±0.003
0.028±0.023
0.030±0.021
0.033±0.027
0.096±0.106
0.136±0.164
0.082±0.097
0.066±0.078
NO3- (μg/m3)
0.079±0.017
0.085±0.023
0.171±0.093
1.185±1.029
0.853±0.506
0.419±0.286
0.658±0.387
0.476±0.258
0.259±0.088
0.238±0.071
SO42- (μg/m3)
0.130±0.042
0.305±0.079
1.154±0.309
4.430±0.834
3.064±1.826
1.006±0.537
0.350±0.105
0.187±0.072
0.134±0.037
0.114±0.025
Na+ (μg/m3)
0.053±0.023
0.043±0.017
0.060±0.027
0.057±0.015
0.080±0.048
0.120±0.036
0.205±0.099
0.211±0.116
0.132±0.067
0.081±0.049
NH4+ (μg/m3)
0.043±0.016
0.123±0.036
0.467±0.132
1.753±0.389
1.222±0.610
0.388±0.220
0.122±0.048
0.055±0.021
0.041±0.012
0.037±0.010
K+ (μg/m3)
0.004±0.003
0.005±0.003
0.014±0.004
0.045±0.021
0.035±0.015
0.021±0.012
0.018±0.007
0.017±0.010
0.007±0.003
0.010±0.004
Mg2+ (μg/m3)
0.000±0.000
0.001±0.001
0.001±0.001
0.003±0.001
0.006±0.001
0.013±0.004
0.028±0.011
0.023±0.014
0.012±0.008
0.006±0.004
Ca2+ (μg/m3)
0.003±0.003
0.007±0.006
0.005±0.002
0.009±0.003
0.013±0.005
0.027±0.008
0.070±0.032
0.061±0.028
0.043±0.018
0.032±0.022
0.264±0.054
OC (μg/m3)
1)
EC (μg/m3)
WSOC
(μg/m3)
1)
Average ± standard deviation
Table 3. Statistical concentration results for the selected sampling period 1 - 4.
Period 1 1)
Period 2
Period 3
Period 1
36.29±20.40
26.68±2.66
19.83±6.34
28.89±9.99
OC (μg/m3)
4.03±2.02
3.33±0.99
1.12±0.62
1.91±0.55
EC (μg/m3)
0.32±0.10
0.29±0.05
0.18±0.03
0.21±0.03
EC/OC
0.08±0.02
0.09±0.01
0.19±0.08
0.12±0.02
WSOC/OC
0.91±0.08
0.87±0.14
0.82±0.31
0.83±0.08
WSOC (μg/m3)
3.73±2.15
2.80±0.47
0.98±0.70
1.60±0.51
Anion sum (μg/m3) 3)
16.24±9.26
9.51±1.13
10.11±3.58
15.87±6.64
Cation sum (μg/m3) 4)
5.23±2.58
3.30±0.25
3.00±0.81
3.90±1.28
12BCA (ng/m3)
2.17±0.78
1.73±0.39
1.09±0.29
1.76±0.78
14BCA (ng/m3)
0.76±0.18
0.62±0.15
0.39±0.11
0.76±0.45
124BCA (ng/m3)
0.74±0.21
0.60±0.15
0.29±0.15
0.63±0.16
135BCA (ng/m3)
1.16±0.47
0.87±0.21
0.48±0.19
0.94±0.31
1245BCA (ng/m3)
0.30±0.01
0.25±0.03
0.17±0.07
0.29±0.05
BCA Sum (ng/m3)
5.12±1.65
4.07±0.87
2.42±0.70
4.37±1.73
PM2.5 massT (μg/m3) 2)
1) Average ± standard deviation
2) 24 hour integrated PM2.5 mass concentration using Teflon filter gravity method (unit: μg/m3)
3) Sum of Cl−, NO3−, and SO42−
4) Sum of Na+, NH4+, K+, Ca2+, and Mg2+
Figure Captions
Fig. 1. (a) Evolution of particle size distributions and particle mass concentrations [d(µg)/dlogdp] (l/m3)
indicated with the selected Periods (1-4) based on Micro-Orifice Uniform Deposit Impactors (MOUDI)
sampling. (b) Time series of PM2.5 massT using Teflon filter gravity method and PM2.5 massP using optical
particle size (OPS). (c) Time series of elemental carbon (EC) and equivalent Black Carbon (eBC). (d) Time
series of water soluble organic carbon (WSOC) and water insoluble organic carbon (WIOC). (e) Time series of
NO3-, SO42-, and NH4+. (f) Time series of underivatized phthalic acid (12-BCA), terephtalic acid (14-BCA),
trimellitic acid (124-BCA), trimesic acid (135-BCA), and pyromellitic acid (1245-BCA) in PM2.5.
Fig. 2. (a) Result of residence-time weighted concentrations (RTWC) (unit; µg/m3) for the sampling period 1(a),
2(b), 3(c), and 4(d). (note: unit range from 0 to 60 for (a), (c), (d) and to 20 for (b)).
Fig. 3. Average size distributions for the mass concentrations of WSOC, WIOC, EC, and WSIs (including and
excluding NH4+ and SO42-) for the selected period 1(a), 2(c), 3(e), and 4(g). The ratios of (WIOC+EC) to
(WSOC+NH4++NO3-+SO42-) and (Na++Cl-) to (WSOC+NH4++NO3-+SO42-) in different size bins for the selected
period 1(b), 2(d), 3(f), and 4(h).
Fig. 4. (a) Average size distributions of the mass concentrations of benzene carboxylic acid and (b) the relative
fractional contributions of individual species in different size bins for the entire sampling period.
Fig. 5. Pairwise scatterplot between µeq. NH4+ and µeq. SO4+ for periods 1-4 colored by logdp
Fig. 6. Size-resolved ionic balance in charge equivalents and expected acidity for the selected period 1(a), 2(b),
3(c), and 4(d).
Fig. 7. Comparison between optical and chemical mass size distributions using OPS and MOUDI for the
selected period 1(a), 2(b), 3(d), and 4(d).
(a)
(b)
80
3
(μg/m )
PM2.5 mass
60
40
20
0
5
4
3
2
1
0
0.5
3
EC (μg/m )
3
eBC (μg/m )
(c)
PM2.5 mass by OPS
PM2.5 mass by 24 hr Gravity method using Teflon filter
100
0.4
5/30/2016
6/3/2016
6/7/2016
6/11/2016
6/15/2016
6/19/2016
5/30/2016
6/3/2016
6/7/2016
6/11/2016
6/15/2016
6/19/2016
5/30/2016
6/3/2016
6/7/2016
6/11/2016
6/15/2016
6/19/2016
5/30/2016
6/3/2016
6/7/2016
6/11/2016
6/15/2016
6/19/2016
5/30/2016
6/3/2016
6/7/2016
6/11/2016
6/15/2016
6/19/2016
EC
eBC
0.3
0.2
0.1
(d)
μgC/m
WSOC
3
6
WIOC
4
2
0
(e)
2-
NH4
μg/m
SO4
-
3
NO3
-
50
40
30
20
10
0
(f)
8
ng/m
3
12BCA
14BCA
124BCA
135BCA
1245BCA
6
4
2
0
Fig. 1. (a) Evolution of particle size distributions and particle mass concentrations [d(µg)/dlogdp] (l/m3)
indicated with the selected Periods (1-4) based on Micro-Orifice Uniform Deposit Impactors (MOUDI)
sampling periods.. (b) Time series of PM2.5 massT using Teflon filter gravity method and PM2.5 massP using
optical particle size (OPS). (c) Time series of elemental carbon (EC) and equivalent Black Carbon (eBC). (d)
Time series of water soluble organic carbon (WSOC) and water insoluble organic carbon (WIOC). (e) Time
series of NO3-, SO42-, and NH4+. (f) Time series of underivatized phthalic acid (12-BCA), terephtalic acid (14BCA), trimellitic acid (124-BCA), trimesic acid (135-BCA), and pyromellitic acid (1245-BCA) in PM2.5.
(a)
(b)
µg/m3
µg/m3
(c)
(d)
µg/m3
µg/m3
Fig. 2. (a) Result of residence-time weighted concentrations (RTWC) (unit; µg/m3) for the sampling period 1(a),
2(b), 3(c), and 4(d). (note: unit range from 0 to 60 for (a), (c), (d) and to 20 for (b)).
2
2-
0
0
6 8
10
15
10
0
0.1
1
6 8
4
20
2
0
(g)
0
0.1
2
4
6 8
1
2
4
6 8
dp (μm)
12
8
30
6
20
4
10
2
0
0.1
2
4
6 8
1
2
dp (μm)
4
6 8
10
2
2-
0
-3
40
+
10
dM(μg)/dlogdp, m
50
Without NH4 and SO4
60
6 8
4
6 8
4
6 8
4
6 8
10
0.2
0.1
2
4
6 8
1
2
10
dp (μm)
0.3
0.2
0.1
0.0
(h)
2
10
4
0.4
2-
-3
10
+
30
dM(μg)/dlogdp, m
40
Without NH4 and SO4
6
2
0.5
8
50
1
0.3
(f)
dp (μm)
60
6 8
dp (μm)
2
10
4
0.0
2-
0
2
0.1
SO4
-3
5
10
4
0.1
0.4
and
20
+
30
NH4
40
dM(μg)/dlogdp, m
50
2
0.2
0.5
20
6 8
+
+
0.0
Without
60
4
-
0.3
(d)
dp (μm)
2
2-
(Na +Cl ) / (WSOC+SO4 +NO 3 +NH 4 )
2
Ratio
1
4
Ratio
0.1
2
-
0.1
SO4
-3
4
10
and
6
20
6 8
+
0.4
Ratio
8
+
10
NH4
All analyzed compounds
-3
dM(μg)/dlogdp, m
14
30
4
-
(WIOC+EC) / (WSOC+SO 4 +NO 3 +NH 4 )
0.5
12
2
2-
(b)
dM(μg)/dlogdp, m
All analyzed compounds
-3
dM(μg)/dlogdp, m
NH4+
40
(e)
All analyzed compounds
-3
dM(μg)/dlogdp, m
Cl-
Na+
Ca2+
w/o NH4+ & SO42-
50
(c)
All analyzed compounds
-3
dM(μg)/dlogdp, m
WIOC
SO 42Mg2+
Without
60
EC
0.1
2
4
6 8
1
2
10
dp (μm)
1.0
0.8
Ratio
WSOC
NO3K+
(a)
0.6
0.4
0.2
0.0
0.1
2
4
6 8
1
2
10
dp (μm)
Fig. 3. Average size distributions of the mass concentrations of WSOC, WIOC, EC, and WSIs (including and
excluding NH4+ and SO42-) for the selected period 1(a), 2(c), 3(e), and 4(g). The ratios of (WIOC+EC) to
(WSOC+NH4++NO3-+SO42-) and (Na++Cl-) to (WSOC+NH4++NO3-+SO42-) in different size bins for the selected
period 1(b), 2(d), 3(f), and 4(h).
(b)
-3
dM(ng)/dlogdp, m
4
12BCA
14BCA
124BCA
135BCA
1245BCA
WSOC
40
30
3
2
20
1
10
0
0.1
2
4
6 8
1
2
dp (μm)
4
6 8
10
2
-3
0
WSOC dM(μg)/dlogdp, m
50
100
Relative percent (%)
(a)
80
60
40
20
0
0.1
2
4
6 8
1
2
4
6 8
10
2
dp (μm)
Fig. 4. (a) Average size distributions of the mass concentrations of benzene carboxylic acid and (b) the relative
fractional contributions of individual species in different size bins for the entire sampling period.
Fig. 5. Pairwise scatterplot between µeq. NH4+ and µeq. SO4+ for the period 1-4 colored by logdp
(a)
(b)
50
Cation (eq.) / Anion (eq.)
+
+
NH4 (measured) / NH4 (expected)
WSOC (%)
40
1.5
0.5
0.0
0
0.1
2
3
4 5 6
(c)
2
1
dp (μm)
3
4 5 6
10
0.1
0.5
10
0.0
0
3
4 5 6
10
2
2
1
dp (μm)
3
4 5 6
10
2
50
40
30
1.0
20
0.5
WSOC (%)
20
1
dp (μm)
4 5 6
1.5
WSOC (%)
30
1.0
0.1
3
2.0
40
2
0
2
(d)
1.5
4 5 6
10
0.0
50
3
20
2
2.0
2
30
1.0
0.5
10
Ratio
Ratio
20
40
WSOC (%)
1.0
50
1.5
WSOC (%)
30
Ratio
2.0
Ratio
2.0
10
0.0
0
0.1
2
3
4 5 6
2
1
dp (μm)
3
4 5 6
10
2
Fig. 6. Size-resolved ionic balance in charge equivalents and expected acidity for the selected period 1(a), 2(b),
3(c), and 4(d).
Fig. 7. Comparison between optical and chemical mass size distributions using OPS and MOUDI for the
selected period 1(a), 2(b), 3(d), and 4(d).
Supplemental Information
Originations and effects of size-resolved aerosols in the coastal area; measurement
during KORUS-AQ campaign
Keywords: MOUDI, OPS, Benzene Carboxylic Acid, WSOC, OC
______________
* Corresponding author:
Table S1. Analytical conditions of liquid chromatography-mass spectrometry mass spectrometry (LC-MSMS)
for benzene carboxylic acids (BCA).
Column
Eclipse XDB-C18 4.6 mm ID x 150 mm (5μm)
Column Temp.
40 ℃
Gas Temp.
300 ℃
Gas Flow
8 L/min
Nebulizer
50 psi
Capillary
2000 V
Ion source
ESI 1)
Time segments Scan Type
MRM 2)
Time segments Polarity
Negative
Eluent
25 (10 mM Ammonium Acetate in DDW) : 75 (Acetonitrile)
Solvents Flow
0.7 mL/min
Injection Volume
100 µL
1) Electrospray ionization
2) Multiple reaction monitoring
Table S2. Multiple Reaction Monitoring (MRM) conditions using LC-MSMS for benzene carboxylic acids
(BCA).
FV
Compound
Abbreviation
Precursor
PI
b)
CE
PI b)
MDL
RPD
e)
f)
r2 d)
MW
a)
ion
Quantifier
c)
Qualifier
R g)
phthalic
Phth_D4
170.16
87
169.1
169
3
125.1
-
-
-
-
12 BCA
166.14
68
165.1
165.1
1
121.1
0.999
2.63
0.34
98.2
14 BCA
166.14
75
165.1
164.8
7
121.1
0.999
2.59
0.19
98.8
124 BCA
210.14
84
209.1
165.2
6
121.2
0.998
1.72
0.34
96.2
135 BCA
210.14
129
209.1
209.1
3
165.1
0.999
3.27
0.24
102.1
1245 BCA
254.15
106
253.1
165
8
121.2
0.999
4.61
1.55
104
acid - D4
phthalic
acid
terephtalic
acid
trimellitic
acid
trimesic
acid
pyromellitic
acid
a) Fragmentor Voltage (V).
b) Product Ion
c) Collision Energy (V) for Quantifier and Qualifier.
d) Correlation coefficient of determination.
e) Method of Detection Limit: Unit of pg/m3
f) Relative Percent Differences at the MDL concentrations.
g) Recovery (%) by standard spike.
Fig. S1. Sampling site; Global Atmosphere Watch supersite of World Meteorological Organization in South
Korea.
(b)
80
5
Slope=1.012 ± 0.148
Intercept=-1.579 ± 4.230
2
r =0.736
Slope=7.106 ± 0.984
Intercept=-0.364 ± 0.225
2
r =0.733
4
3
60
eBC (μg/m )
3
OPS PM2.5 mass (μg/m )
(a)
40
20
6/15/2016
6/10/2016
6/5/2016
5/31/2016
3
2
1
0
0
0
20
40
60
80
3
24 hr filter PM2.5 mass (μg/m )
0.0
0.1
0.2
0.3
0.4
0.5
3
EC (μg/m )
Fig. S2. (a) Pairwise scatterplot between PM2.5 mass using optical particle sizer (OPS) and using gravity method
from 24-hour integrated Teflon filters. (b) Pairwise scatterplot between equivalent Black Carbon (eBC) and
elemental carbon (EC).
(a)
(b)
100
Relative percent (%)
Relative percent (%)
100
80
60
40
20
0
2
4
6 8
1
2
4
6 8
10
40
20
2
0.1
(d)
dp (μm)
2
4
6 8
2
1
4
6 8
10
2
dp (μm)
100
Relative percent (%)
100
Relative percent (%)
60
0
0.1
(c)
80
80
60
40
20
0
80
60
40
20
0
0.1
2
4
6 8
1
2
4
6 8
10
2
0.1
2
4
dp (μm)
6 8
2
1
4
6 8
10
2
dp (μm)
WSOC
NO3K+
EC
SO 4
Mg2+
2-
WIOC
Cl-
Na+
NH4+
Ca2+
Fig. S3. The relative fractional contributions of individual species in different size bins for the selected period.
Period 1(a), 2(b), 3(c), and 4(d).
(a)
5
eBC
eBC with wind direction (225 - 315
o
)
and less 1.0 μg/m
3
eBC μg/m
3
4
3
2
1
0
5/30/2016
6/2/2016
6/5/2016
6/8/2016
OC
EC
Cl
NO3
25
2-
20
SO4
+
Na
+
NH4
15
+
K
2+
Mg
2+
Ca
10
5
0
6
-3
-3
30
dM(μg)/dlogdp, m
6/14/2016
6/17/2016
4
2
6/20/2016
(c)
dM(μg)/dlogdp, m
(b)
6/11/2016
OC
EC
Cl
NO3
5
+
4
Na
+
K
2+
Mg
2+
Ca
3
2
1
0
0.1
2
4
6 8
1
2
4
6 8
10
2
0.1
2
dp (μm)
6
8
1
4
6
8
10
2
dp (μm)
Fig. S4. (a) Time series of eBC (the solid circles present eBC only with wind direction of 225-315o and less
than 1.0 µg/m3). Size distributions of the mass concentrations of OC, EC, and WSIs (b) including and (c)
excluding NH4+ and SO42- for the selected MOUDI set (June 13-Jun 16).
(a)
(b)
(c)
(d)
(e)
(f)
Fig. S5. Four-day averaged chlorophyll-a concentration obtained from the first Korean Geostationary Ocean
Color Imager (GOCI, http://kosc.kiost.ac/) for (a) May 28 – 31, (b) June 1 – 4, (c) June 5 – 8, (d) June 9 – 12, (e)
June 13 – 16, and (f) June 17 – 20.