Cloud condensation nuclei activity at Jeju Island, Korea in ...

Atmos. Chem. Phys., 8, 2933–2948, 2008www.atmos-chem-phys.net/8/2933/2008/© Author(s) 2008. This work is distributed underthe Creative Commons Attribution 3.0 License.

AtmosphericChemistry

and Physics

Cloud condensation nuclei activity at Jeju Island, Korea in spring2005

M. Kuwata 1, Y. Kondo1, Y. Miyazaki1,*, Y. Komazaki1, J. H. Kim 2, S. S. Yum2, H. Tanimoto3, and H. Matsueda4

1Research Center for Advanced Science and Technology, the University of Tokyo, Tokyo, Japan2Department of Atmospheric Science, Yonsei University, Seoul, Korea3Atmospheric Environmental Division, National Institute for Environmental Studies, Tsukuba, Japan4Geochemical Research Department, Meteorological Research Institute, Tsukuba, Japan* now at: Institute of Low Temperature Science, Hokkaido University, Japan

Received: 12 October 2007 – Published in Atmos. Chem. Phys. Discuss.: 13 November 2007Revised: 14 May 2008 – Accepted: 14 May 2008 – Published: 11 June 2008

Abstract. We measured the number concentrations of cloudcondensation nuclei (CCN) and the size distributions ofCCN/CN (CN: condensation nuclei) ratios at supersatura-tions (SSs) of 0.097, 0.27, 0.58, and 0.97% at Jeju Island,Korea during March-April 2005. We made simultaneousmeasurements of aerosol inorganic ions, water-soluble or-ganic carbon (WSOC), organic carbon (OC), and elementalcarbon (EC) in PM2.5. The CCN/CN ratios increased with in-creasing particle diameter, and the diameter at CCN/CN=0.5was defined asD50. D50 represents the activation dry diame-ter of atmospheric particles. The averageD50 atSS=0.097%and 0.97% was 136±17 nm and 31±3 nm, respectively. Thetemporal variation ofD50 at SS=0.097% was correlatedwith the mass fraction of water-soluble components (inor-ganic ions + WSOC), indicating that the temporal variationof CCN activity was mainly controlled by changes in thewater-soluble components fraction. The critical dry diam-eter (Dcrit), which is the threshold dry diameter for CCNactivation, was calculated from the observed aerosol chemi-cal compositions by Kohler theory for comparison withD50.The D50 at SS=0.097% was correlated (r2=0.48) with cal-culatedDcrit, althoughDcrit was larger thanD50 by 20–29%on average. The systematic difference betweenD50 andDcritcould be caused by the size dependence of the aerosol chem-ical compositions or surface tension lowering caused by themixing of water-soluble organic compounds. This differencecorresponds to a 27±14% uncertainty in the CCN numberconcentration estimated from the observed particle numbersize distribution.

Correspondence to:Mikinori Kuwata([email protected])

1 Introduction

A subset of atmospheric particles acts as cloud condensationnuclei (CCN). An increase in CCN number concentrationcauses an increase in cloud droplet concentration and a de-crease in droplet size, which in turn impacts cloud albedo andprecipitation. Consequently, CCN can significantly influenceclimate through cloud processes (Twomey, 1974; Lohmannand Feichter, 2005, and references therein).

The number concentration of CCN is determined by thenumber concentration of aerosol particles and the fraction ofCCN active particles. This fraction is often denoted as theCCN/CN ratio because condensation nuclei (CN) countershave been used for total particle number concentration mea-surements (Seinfeld and Pandis, 2006). CCN/CN stronglydepends on the size and chemical composition of aerosols.The influence of aerosol chemical composition and numbersize distribution on the bulk (size-unresolved) CCN/CN canbe separated by measuring CCN/CN for size-selected par-ticles. In particular, when aerosol particles are internallymixed, we can determine the threshold diameter from themeasurements (Dusek et al., 2006). The threshold diame-ter can also be calculated by Kohler theory. Many laboratorystudies have measured CCN/CN ratios of size-selected parti-cles for comparison with theoretical calculations (e.g. Corri-gan and Novakov, 1999; Raymond and Pandis, 2002). Theseexperiments have shown that size-resolved CCN/CN ratiosof laboratory-generated particles can be explained by Kohlertheory. However, only a few studies have employed this mea-surement method for atmospheric observations (e.g. Dusek etal., 2006), leading to an insufficient understanding of control-ling factors for the CCN activation of atmospheric particles.

Published by Copernicus Publications on behalf of the European Geosciences Union.

http://creativecommons.org/licenses/by/3.0/

2934 M. Kuwata et al.: CCN activity at Jeju Island, 2005

Diffusion Dryer 241Am DMA

CCNC

CPCAmbient air

Diffusion Dryer 241Am DMA

CCNC

CPCAmbient air

Figure 1

Fig. 1. CCN observation system used for this study. The three-wayvalve placed upstream of the CCN counter was used to select thesampling mode (see text for detailed explanation).

CCN concentrations are often influenced by anthro-pogenic particles (Twomey, 1974; Hudson and Yum, 2002).In recent years, anthropogenic emission of aerosol precur-sors (e.g. SO2 and NOx) from Asia has increased signifi-cantly (Streets et al., 2000; Akimoto et al., 2003; Richteret al., 2005). Remote sensing studies have indicated that theaerosol particles derived from anthropogenic pollutants fromAsia have changed cloud radiative properties such as opti-cal depth in this region (Chameides et al., 2002; Kawamotoet al., 2006). Thus, it is important to study CCN activityin this region for a more accurate assessment of the impactof anthropogenic aerosol on climate. However, only a fewCCN studies have been performed in this region (e.g. Mat-sumoto et al., 1997; Yum et al., 2005), and the size-resolvedCCN/CN ratios have not been measured so far.

In this study, we have measured size-resolved CCN/CNratios and CCN number concentrations at Gosan, Jeju Is-land, Korea, in March and April 2005 during the UnitedNations Environmental Programme (UNEP) AtmosphericBrown Cloud – East Asian Regional Experiment 2005(ABC-EAREX2005) campaign. In addition to the CCNmeasurements, aerosol chemical composition was simulta-neously observed. Observed threshold diameters for CCNactivation are compared with the mass fractions of aerosolchemical compositions. Threshold diameters for activation(Dcrit) are calculated and compared with the observed thresh-old diameter to investigate the controlling factors ofDcrit.

2 Theory

The equilibrium water vapor pressure (S) of an aerosol par-ticle can be calculated by Kohler theory. According to thetheory,S is described as follows (e.g. Roberts et al., 2002;Mochida et al., 2006):

ln S =A

dwet−

Bd3dry

(d3wet − d3

dry)

A =4Mwσ

RTρw

B =Mw

ρw

∑i

viφiεimi

Mi∑i

mi

ρi

(1)

whereddry is the dry diameter of a particle anddwet is the di-ameter of a droplet under equilibrium conditions. The suffixi denotes the properties of thei-th solute compound.Mw isthe molecular weight of water;Mi is the molecular weight ofsolute;ρw is the density of water;ρi is density of solute;σ

is the surface tension;R is the gas constant;T is the temper-ature;νi is the stoichiometric number of ions and molecule;φi is the osmotic coefficient;εi is the degree of dissolution;andmi is the mass mixing ratio of thei-th solute. The firstterm on the right-hand side represents an increase in the equi-librium vapor pressure of water due to surface tension (theKelvin effect). The second term on the right-hand side de-notes the decrease in the equilibrium vapor pressure of waterbecause of solute mixing (Raoult’s effect). When we con-siderS for an aerosol particle, it has a maximum value at acertaindwet (critical droplet diameter). The supersaturation(SS: SS=S−1) that corresponds to thisS is called the criti-cal supersaturation (SSc). Particles can be activated to clouddroplets if theSS of the environment is higher than theirSSc.

If we assume thatA andB are constants anddwet is muchlarger thanddry, SSc can be expressed as follows:

ln(1 + SSc)=

√√√√ 4A3

27Bd3dry

. (2)

This equation shows thatSSc decreases with particle di-ameter. Thus, at a certainSS, there exists a threshold valueof ddry above which all aerosol particles act as CCN. We de-note this diameter as the critical dry diameter (Dcrit).

3 Experiment

3.1 CCN measurement

Figure 1 shows the CCN observation system used for thisstudy. In this system, the relative humidity of the sampleflow was reduced using two diffusion dryers in series (TSIModel 3062). Silica gel was replaced periodically (about 1time in 10 days), and no systematic change in size-resolvedCCN spectra was observed before or after the replacement.Then, particles were charged with a241Am bipolar neutral-izer and introduced to a differential mobility analyzer (DMA:TSI Model 3081). The DMA classified particles by theirelectrical mobility. The voltage applied to the DMA wasscanned stepwise to change the diameter of the classified par-ticle (Table 1). The sheath to-sample flow ratio of the DMAwas set to 10:1. Classified particles were introduced to acondensation particle counter (CPC: TSI Model 3022) anda cloud condensation nuclei counter (CCNC: Droplet Mea-surement Technologies, Inc.) (Roberts and Nenes, 2005).The CPC monitored the number concentrations of conden-sation nuclei (CN), and the CCNC measured number con-centrations of CCN. The sample flow rate of the CCNCwas set to 0.045 l/min, and the sheath flow rate was set to0.455 l/min. The temperature gradient (1T ) of the thermalgradient chamber in the CCNC was changed periodically toalter theSS in the chamber. Four1T s (3.08, 4.62, 9.24,and 15.4 K) were used in this study.SSs corresponding tothese1Ts were calibrated with ammonium sulfate particles

Atmos. Chem. Phys., 8, 2933–2948, 2008 www.atmos-chem-phys.net/8/2933/2008/

M. Kuwata et al.: CCN activity at Jeju Island, 2005 2935

Table 1. Calibration results and operating conditions of the CCN measurement system shown in Fig. 1.SS was calculated at 300 K and72 mN/m by the ideal solution approximation. The values in parentheses showSS calculated using the osmotic coefficient of Clegg etal. (1996).

1T (K) Activation dry diameter of (NH4)2SO4(nm) SS (%) Particle diameters classified by the DMA(nm)

3.08 125±3 0.097(0.10) 80, 100, 110, 120, 130, 140, 160, 180, 200, 240, 2904.62 63±2 0.27(0.30) 60, 65, 70, 75, 80, 85, 90, 100, 130, 160, 200, 240, 2909.24 38±1 0.58(0.65) 35, 40, 45, 50, 60, 70, 80, 100, 130, 160, 200, 240, 29015.4 27±1 0.97(1.1) 25, 30, 35, 40, 50, 60, 80, 100, 130, 160, 200, 240, 290

as described by Kuwata et al. (2007). The calibration wasperformed at the observation site before and after the cam-paign. There are several different methods for the calculationof the water activity of ammonium sulfate particles (Krei-denweis et al., 2005, and references therein). We chose theideal solution approximation (φ=1) in calculating theDcritof atmospheric particles (Sect. 5). Therefore we used thisapproximation for the interpretation of the calibration resultsto ensure consistency, althoughSSs calculated using the os-motic coefficient of Clegg et al. (1996) are likely to be moreplausible than the ideal solution approximation, as it is basedon precise experimental data. We also showSSs calculatedusing the osmotic coefficients by Clegg et al. (1996) for ref-erence in parentheses in Table 1. The Debye – Huckel con-stant is needed to calculate the osmotic coefficient of Clegg etal. (1996) because it is based on the Pitzer model. The Debye– Huckel constant at 300 K was calculated by the equationgiven by Clegg et al. (1994).

Multiply charged particles were included in the classifiedparticles, therefore an inverse analysis was performed forCCN and CN size distribution data for multiple-charge cor-rection. The STWOM algorithm (Markowski, 1987) wasused to obtain size distributions of CN and CCN. In thiscalculation, the equilibrium charge distribution (Wieden-sohler, 1988) and the DMA transfer function derived byKnutson and Whitby (1975) were included in the kernelfunction. The raw data for CN and CCN were linearly inter-polated so that the interval of each bin was1 logdp=0.015.Size-resolved CCN/CN ratios were calculated using the dataafter the inverse analysis.

The three-way valve in Fig. 1 was switched once every30 min for the direct measurement of the CCN number con-centration in ambient air. It took 30 min for the measurementat eachSS, and 2 h were required for the entire measurementcycle.

3.2 Measurements of aerosol composition, size distribution

Inorganic components (NH+4 , Na+, K+, Ca2+, Mg2+, NO−

3 ,SO2−

4 , Cl−) were measured by a particle-into-liquid samplercombined with ion chromatography (PILS-IC) (Orsini et al.,2003; Takegawa et al., 2005). The concentration of water-

soluble organic carbon (WSOC) was measured by PILS com-bined with a total organic carbon analyzer (PILS-WSOC)(Sullivan et al., 2004; Miyazaki et al., 2006, 2007). El-emental carbon (EC) and organic carbon (OC) were mea-sured by a semi-continuous thermal-optical carbon aerosolanalyzer (Sunset Laboratory, Inc.) (Bae et al., 2004; Kondoet al., 2006; Miyazaki et al., 2006). PM2.5 cyclones wereused for these instruments. The detection limits of thePILS-IC, PILS-WSOC, OC, and EC were estimated to be0.01µg/m3, 0.1µg/m3, 1.0µg/m3, and 0.2µg/m3, respec-tively (Takegawa et al., 2005; Kondo et al., 2006; Miyazakiet al., 2006).

Aerosol number size distribution (10–300 nm) was mea-sured with a scanning mobility particle sizer (SMPS 3936,TSI). The SMPS used in this study comprised a DMA (TSIModel 3081) and a CPC (TSI Model 3010). In addition, thenumber concentration of particles larger than 10 nm (CN)was measured by another CPC (TSI Model 3010) (Yum etal., 2007). The concentration of carbon monoxide (CO) wasmeasured using a non-dispersive infrared analyzer (HoribaAPMA-360 model) (Tanimoto et al., 2007).

3.3 Measurement site

The observations were performed between 18 March and 5April 2005 at Gosan (33.2◦ N, 126.1◦ E) on Jeju Island, Ko-rea, as part of the Atmospheric Brown Cloud – East AsianRegional Experiment 2005 campaign. The location of Gosanis shown in Fig. 2. The instruments were placed in a con-tainer located about 10 m back from the edge of a cliff. Thesampling inlets were made of stainless steel tubes with an in-ner diameter of 7 mm. The top of the inlets was located about4 m a.g.l.

The meteorological parameters at the Gosan site wereobserved by the Korean Meteorological Administration(KMA). During the observation period, the dominant windswere northerlies and north-northwesterlies (more than 50%of the observation period) associated with the Siberian high-pressure system. This led to the frequent transport of anthro-pogenic pollutants from the Korean Peninsula and China toGosan. Sawa et al. (2007) attributed the high concentrationsof CO at Gosan to the emissions from these regions based

www.atmos-chem-phys.net/8/2933/2008/ Atmos. Chem. Phys., 8, 2933–2948, 2008


60

50

40

30

20

Latit

ude

(°N

)

140130120110

Longitude (°E)

Gosan

Anmyeon

Japan

Korea

Sea ofJapan

China

Russia

Yellow Sea

Figure 2

Fig. 2. Map of Gosan and surrounding region.

on tagged CO-tracer simulations. More detailed descriptionsof the meteorological conditions and the transport of anthro-pogenic pollutants can be found elsewhere (Sawa et al., 2007;Miyazaki et al., 2007). Previous studies have also shownthat air masses observed at Gosan are heavily influenced byanthropogenic emissions from East Asia (Carmichael et al.,1997; Lee et al., 2006).

4 Results

4.1 CCN/CN size distributions

Figure 3 shows the average size-resolved CCN/CN ratio atSS=0.097%, together with the data on specific days. TheCCN/CN size distribution of ammonium sulfate obtainedduring calibration at the observation site (closed circles)is also shown in this figure for comparison. In general,CCN/CN increases with increasing diameter. At 02:00–02:30 on 26 March, the CCN/CN size distribution of ambientparticles was very similar to that of ammonium sulfate. Onthe other hand, on 28 March, the CCN/CN size distributionshifted to a larger diameter, and the rate of increase was lowerthan ammonium sulfate. In addition, the spectrum is signifi-cantly different from a sigmoid function. This shift indicatesthat the bulk aerosol chemical composition on 28 March wassignificantly different from ammonium sulfate. The slowerrate of increase indicates the co-existence of different typesof aerosol particles with different activation curves. A con-volution of sigmoid functions is not necessarily a sigmoidfunction. Thus, the non-sigmoid shape of the spectrum canbe interpreted as a result of the co-existence of different types

1.0

0.5

0.0

CCN

/CN

250200150100Mobility diameter (nm)

(NH4)2SO4 Average 3/26 (2:00 - 2:30) 3/28 (22:00 - 22:30)

Figure 3 Fig. 3. Size distributions of CCN/CN ratios of ammonium sulfateand atmospheric particles atSS=0.097%. The average, standard de-viation (gray area), and two examples of CCN/CN during the obser-vation period are shown. The solid circles denote the experimentaldata for ammonium sulfate under the observation conditions and thesolid line shows the fitting results of the experimental data.

of aerosol particles. The different width of the DMA trans-fer function atD50 in diameter space (∼25%: estimated by aKnutson-type DMA transfer function) is not large enough toexplain the lower increase rate, as the widths of the activationcurves are different by about a factor of 2.

We defined the threshold diameter for CCN activation tocompare it with the mass fraction of chemical compoundsand theoretically calculatedDcrit. In this study, the thresh-old diameter was defined as the diameter corresponding toCCN/CN=0.5(D50) because it represents the bulk chemicalcomposition. If aerosol particles were not fully activated,D50 would not necessarily correspond to the threshold diam-eter. In all cases shown in Fig. 3, the CCN/CN ratios ex-ceeded 0.85 at 250 nm. In addition, no multi-step activationwas observed. These results suggest the appropriateness ofthe definition of the unique activation diameter (D50). Previ-ous studies on CCN activity of laboratory-generated particleshave shown thatD50 can be quantitatively compared withthe theoretically calculated activation diameters (e.g. Corri-gan and Novakov, 1999; Raymond and Pandis, 2002).

4.2 Temporal variation of CCN and particle number sizedistribution

4.2.1 CCN number concentration

Figure 4a shows the time series of the CCN number con-centration. In general, the variation in the CCN numberconcentration was correlated with that of CO concentration(Fig. 4d). In particular, this correlation is clearly observedat SS=0.097% (r2=0.74). Sawa et al. (2007) have reportedhigh CO concentration between 22–24 March and 30 March–2 April due to transport of CO from the Korean Peninsulaand China. During these periods, CCN concentration alsoincreased (Fig. 4a and d). CO is emitted mainly by the in-complete combustion of fossil fuels and biomass, and it is a



12x103

10

8

6

4

2

0

NC

CN

(cm

-3)

3/19 3/21 3/23 3/25 3/27 3/29 3/31 4/2 4/4

SS 0.097% SS 0.27% SS 0.58% SS 0.97%

b)

5 0x1 03

40302010

0

cm-3

3 /1 9 3 /2 1 3/23 3/25 3/27 3/2 9 3 /3 1 4 /2 4 /4

CN CCN (0.97%)

60 0

40 0

20 0CO(p

pb)

3 /1 9 3 /21 3 /23 3 /2 5 3 /27 3 /29 3 /31 4 /2 4 /4

d)

c)

a) 12x103

10

8

6

4

2

0

NC

CN

(cm

-3)

3/19 3/21 3/23 3/25 3/27 3/29 3/31 4/2 4/4

SS 0.097% SS 0.27% SS 0.58% SS 0.97%

b)

5 0x1 03

40302010

0

cm-3

3 /1 9 3 /2 1 3/23 3/25 3/27 3/2 9 3 /3 1 4 /2 4 /4

CN CCN (0.97%)

60 0

40 0

20 0CO(p

pb)

3 /1 9 3 /21 3 /23 3 /2 5 3 /27 3 /29 3 /31 4 /2 4 /4

d)

c)

a)

Figure 4

Fig. 4. Time series of(a) CCN number concentration,(b) aerosol size distribution,D50, (c) CN concentration, CCN (SS=0.97%) concen-tration, and(d) CO concentration. The dashed lines in (b) shows theDcrit of ammonium sulfate at eachSS.

good indicator of emissions from combustion sources. Pri-mary aerosols and aerosol precursors such as SO2 and NOxare co-emitted with CO. Thus, these high CCN concentra-tions were likely caused by the enhanced concentrations ofanthropogenic aerosols transported from these regions. Theaverage values and standard deviations of CCN number con-centration are shown in Table 2. The average CCN numberconcentration atSS=0.097% and 0.97% was 1200 cm−3 and4000 cm−3, respectively. These concentrations are higherthan those of other remote areas of the world such as theIsland of Tasmania in Australia and Mace Head in Ireland byabout an order of magnitude (Covert et al., 1998; Reade et al.,2006). CCN concentrations measured at Gosan during thisobservation period were extensively compared with those ob-tained in other regions by Yum et al. (2007). The CCN con-

centrations observed in this study are close to those of An-myeon (Korea: see Fig. 2) in springtime of 2004 (SS=1%)(Yum et al., 2005).

4.2.2 D50

Figure 4b shows the time series of the number size distribu-tion of aerosol particles andD50. The Dcrit of ammoniumsulfate at eachSS is also shown as dashed lines in this fig-ure. The average values ofD50 at SS=0.097% and 0.97%were 136 nm and 31 nm, respectively (Table 2). In general,theD50 is almost equal to or slightly larger (by∼25%) thanthe Dcrit of ammonium sulfate (125 nm and 27 nm, respec-tively). This means that theDcrit of ammonium sulfate is thesmallestDcrit of the atmospheric particles during the obser-vation period.



Table 2. Average values and standard deviations of the CCN number concentration,D50, andB. The surface tension of water was assumedfor the calculation ofB.

SS (%) CCN number concentration (cm−3) D50 (nm) B

0.097 1194±746 136±17 0.61±0.170.27 2543±1277 71±6 0.55±0.120.58 3496±1510 44±3 0.50±0.090.97 3996±1686 31±3 0.55±0.13

The temporal variation ofD50 at differentSSs did not al-ways correlate. As shown in Eq. (2), the threshold diameterfor CCN activation depends onA andB, which are deter-mined by the aerosol chemical composition. Thus, this dif-ference in the temporal variations ofD50 at eachSS indicatesthe difference in temporal variation of chemical compositionin different size ranges.

Using Eq. (2), we calculatedB assuming the surface ten-sion of water. This parameter gives information on the chem-ical composition (approximate number of solute ions andmolecules included in a unit volume) atD50 for eachSS.The results are summarized in Table 2. The average valuesof B did not depend onSS significantly, indicating that thechemical composition averaged over the observation periodwas rather uniform in the diameter range considered (30 to160 nm).

4.2.3 Number size distribution

New particle formation events can have a significant impacton CCN number concentration (e.g. O’Dowd et al., 2002;Laaksonen et al., 2005). In Fig. 4b, new particle forma-tion is clearly identified on 19, 25, 29, 30 and 31 March.It can be seen more clearly in Fig .4c, which shows the parti-cle number concentration measured by CPC and CCN num-ber concentration (SS=0.97%). The concentration of CCN(SS=0.97%) is a good indicator of particle number concen-trations larger than 30 nm, as the average value ofD50 atSS=0.97% is 31±3 nm (Table 2). Thus, the difference ofthe number concentrations (CN-CCN (SS=0.97%)) repre-sents the number concentration of particles between 10 and30 nm. During the periods of new particle formation eventsdescribed above, an enhancement of the number concentra-tion of small (10–30 nm) particles was observed, as indicatedby the large gap between CN and CCN concentrations. Inparticular, the events occurring on 29 and 30 March wereimportant in that newly formed particles influenced the CCNnumber concentration as a consequence of particle growthbeyondD50. Figure 5a and b shows the CCN number con-centrations, particle size distributions, andD50 during thisevent. The peak diameter of the size distribution obtained bybimodal lognormal fitting is shown as red lines in Fig. 5b.

At 14:00 local time (LT) on 29 March, small (<20-nm)particles appeared and began to grow. The peak diametergrew to 25 nm by 18:00 LT, and some particles grew largerthan theD50 at SS=0.97% (28 nm). At this time, the CCNnumber concentration (SS=0.97%) began to increase. Thepeak diameter andD50 (SS=0.97%) were equal at 21:00 LT(blue dashed vertical line in Fig. 5b). At this time, themajority of newly formed particles began to act as CCNat SS=0.97%. In the case ofSS=0.58%, the peak diam-eter equaledD50 at 2:00 LT on 30 March , and the CCNnumber concentration increased from 1700 cm−3 (01:30 LT)to 5800 cm−3 (11:30LT). At 03:00 LT, some portion of theparticles grew large enough to act as CCN atSS=0.27%.Then, CCN number concentration atSS=0.27% increasedfrom 1000 to 4700 cm−3 with the increase in the peak diam-eter. For thisSS, the peak diameter reachedD50 at 10:00 LT.At the same time, another new particle formation event oc-curred, and this event also clearly affected the CCN num-ber concentration atSS=0.97% and 0.58%. At 12:00 LT,some fraction of the particles grew larger than theD50 atSS=0.097%, and they affected the CCN number concentra-tion at thisSS. A similar phenomenon was also observed on25 March (Fig. 4a and b). These results clearly show thatthe newly formed particles significantly increased the CCNnumber concentration.

Buzorius et al. (2004) have shown that the deliques-cence relative humidity and the hygroscopic growth of newlyformed particles at Gosan during the ACE-Asia campaignwere similar to those of ammonium sulfate. This suggeststhat newly formed particles were mainly composed of in-organic compounds because no organic compounds in theatmosphere are known to be as hygroscopic as ammoniumsulfate. In addition, according to Buzorius et al. (2004), glu-taric acid is the only organic compound that is known to havea deliquescence relative humidity similar to ammonium sul-fate. However, the hygroscopic growth of glutaric acid isless than ammonium sulfate. For instance, Cruz and Pan-dis (2000) measured the hygroscopic growth factors of glu-taric acid and ammonium sulfate particles to be 1.1 and 1.5 ata relative humidity of 85%, respectively. This also supportsthe conclusion of their study that newly formed particleswere mainly composed of ammonium sulfate at Gosan. Inthe present study, the observedD50 of newly formed particles



10

2

3

456

100

2

3

Mob

ility

Dia

met

er (n

m)

0:003/29

6:00 12:00 18:00 0:003/30

6:00 12:00 18:00 0:003/31

8000

6000

4000

2000

0

dN/d

logD

p (c

m-3

)

(a)

(b)

10000

8000

6000

4000

2000

0N

CC

N (c

m-3

)00:003/29

06:00 12:00 18:00 00:003/30

06:00 12:00 18:00 00:003/31

SS = 0.097% SS = 0.27% SS = 0.58% SS = 0.97%

10

2

3

456

100

2

3

Mob

ility

Dia

met

er (n

m)

0:003/29

6:00 12:00 18:00 0:003/30

6:00 12:00 18:00 0:003/31

8000

6000

4000

2000

0

dN/d

logD

p (c

m-3

)

(a)

(b)

10000

8000

6000

4000

2000

0N

CC

N (c

m-3

)00:003/29

06:00 12:00 18:00 00:003/30

06:00 12:00 18:00 00:003/31

SS = 0.097% SS = 0.27% SS = 0.58% SS = 0.97%

Figure 5

Fig. 5. Close-up of Fig. 4 during the new particle formation events from 29–31 March. The(a) time series of CCN number concentration and(b) the number size distribution andD50 (line with filled circles) are shown. The red solid lines in (b) denote the peak diameters obtained bybimodal lognormal fitting. The vertical dashed lines show the time at which the peak diameter andD50 were equal.

was also similar to ammonium sulfate. For instance, theD50for SS=0.97% was 27 nm at 21:00 LT, 29 March, and forSS=0.27% it was 66 nm at 10:00 LT, 30 March. TheDcrit ofammonium sulfate was 27 nm and 63 nm, respectively (Ta-ble 1).

Previous studies have suggested that new particle forma-tion has an impact on the CCN number concentration fromthe measurements of number size distributions (e.g. O’Dowdet al., 2002; Laaksonen et al., 2005) and modeling (e.g.Arnold, 2006; Sotiropoulou et al., 2006). The present ob-servations clearly demonstrate that new particle formation isone of the important processes of CCN formation at Gosan.

Figure 6 shows the number size distribution of all parti-cles and CCN averaged over the whole observation period.Average CCN/CN ratios were multiplied by the number sizedistribution measured by the SMPS averaged over the ob-servation period to obtain a rough estimate of the CCN sizedistribution. The peak diameter of the average CCN sizedistribution was about 150 nm atSS=0.097% and shifted toabout 80 nm atSS=0.97%. Detailed discussion of the num-ber size distribution of particles during the observation pe-riod has been given by Yum et al. (2007).

6000

4000

2000

0

dN/d

logD

p (c

m-3

)

102 3 4 5 6 7 8 9

1002

Mobility diameter (nm)

all particles CCN (0.097%) CCN (0.27%) CCN (0.58%) CCN (0.97%)

Figure 6

Fig. 6. Number size distribution of particles and CCN. CCN sizedistributions were estimated by multiplying CCN/CN ratios by thenumber size distribution. The data were averaged over the wholeobservation period. The data range of CCN size distributions waslimited by the scanning range of the DMA.



0.5

0.0

2SO

42-+N

O3- (μ

mol

/m3 )

0.50.0NH4

+ (μmol/m3)

intercept = 0.02 slope = 0.94

r2 = 0.91

Figure 7

Fig. 7. Ion balance between ammonium, sulfate, and nitrate ions.

4.3 Ion balance of inorganic components

The average concentrations of the inorganic componentsmeasured by PILS-IC are summarized in Table 3. NH+

4 wasthe most abundant cation (56%), and Ca2+ was the second(15%). SO2−

4 was the most abundant anion (67%). Top-ping et al. (2004) measured inorganic components using aBerner impactor at Gosan during the ACE-Asia campaign.They showed that the concentration of NH+

4 was the high-est at the 0.2–0.5µm stage, and Ca2+ was the highest at the1.5–5.5µm stage. In this study, theD50 values were in thesub-micrometer range (Fig. 4). Thus, Ca2+ was ignored inthe following discussion of CCN activity. Na+, Mg2+, andCl− were also ignored for the same reason.

Figure 7 shows the ion balance of NH+

4 , NO−

3 , and SO2−

4 .The cation and anions balance very well (slope=0.94). Thisresult shows that sulfate and nitrate in PM2.5 were neutral-ized by ammonium at Gosan. The addition of K+ to the ionbalance causes the balance to deviate from the 1:1 line. Thissuggests that neither NO−3 nor SO2−

4 were the counter ions ofK+. In addition, because the molar concentration of K+ wasonly 8% of NH+

4 , we ignored K+. In the following discus-sion, we assume that the inorganic component of sub-micronparticles was composed only of ammonium sulfate and am-monium nitrate.

5 Discussion

5.1 Temporal variation of D50 and aerosol composition

Figure 8 shows the temporal variation ofD50 (SS=0.097%)and the mass fraction of the aerosol components. In this fig-ure, the aerosol components were divided into four groups:inorganic (NH+

4 + SO2−

4 + NO−

3 ), WSOC, water-insoluble

1.0

0.8

0.6

0.4

0.2

0.0

Mas

s Fr

actio

n

3/25 3/27 3/29 3/31 4/2 4/4

180

160

140

120

100

D50 (nm

)

Inorganic WSOC WIOC EC

Figure 8

Fig. 8. Temporal variation of the mass fraction of aerosol compo-nents andD50 at SS=0.097%. Here, “inorganic” denotes the sumof ammonium sulfate and ammonium nitrate.

organic carbon (WIOC = OC – WSOC), and EC. Inorganiccomponents and WSOC are soluble in water, while WIOCand EC are insoluble in water. Note that the masses ofWSOC and WIOC are the masses of carbon in organic com-pounds and do not include the masses of other elements (e.g.oxygen and hydrogen) (Kondo et al., 2007). The temporalvariation ofD50 and the water-soluble fraction (inorganic +WSOC) are correlated. With an increase in the water-solublefraction,D50 decreases. This is consistent with Kohler the-ory because the value ofB in Eq. (2) increases with an in-crease in the water-soluble fraction. In this case, the vari-ation in D50 was as small as about 40 nm. Thus,D50 iscorrelated with the water-soluble fraction almost linearly, al-though the relationship is not always linear, as anticipatedfrom Eqs. (1) and (2). This good correlation indicates thatthe variation ofD50 was not significantly influenced by thevariation in the chemical composition of the water-solublecomponent (e.g change of inorganic/WSOC ratio). Kim etal. (2006) have shown that the hygroscopicity of aerosol par-ticles at Gosan was correlated with the mass fraction of sul-fate. This indicates thatB is mainly determined by inorganiccomponents because hygroscopicity is mainly determined byB (e.g. Kreidenweis et al., 2005; Mochida et al., 2006). Thisis consistent with the present observations because the massfraction of inorganic compounds was significantly larger thanthat of WSOC during this period (Fig. 8).

TheD50 at SS=0.097% did not necessarily correlate withthose of higherSSs (Fig. 4b), as discussed in Sect. 4.2.2.Correlation ofD50 at SS=0.097% with those at higherSSsdecreased with the increase inSS (r2=0.25 and 0.00 forSS=0.27% and 0.97%, respectively). This indicates thatthe temporal variation of the mass fraction of PM2.5 wasreflected in that ofD50 at SS=0.097% (100–200 nm) butwas not reflected inD50 at higherSSs (<100 nm). This isbecause the contribution to the PM2.5 mass concentrationof particles with diameters smaller than 100 nm was muchsmaller than that of 100–200 nm.



Table 3. Average concentration of inorganic components during the observation period.

NH+

4 Na+ K+ Ca2+ Mg2+ NO−

3 SO2−

4 Cl−

1.57 (µg/m3) 0.38 0.26 0.43 0.18 1.44 3.96 0.49

5.2 Comparison ofD50 and calculatedDcrit

In this section, we compare the observedD50 and theDcritcalculated from the simultaneously measured chemical com-position. We first discuss the assumptions used to calculateA andB in Kohler theory (Sect. 2). The surface tension ofwater (72 mN/m) was assumed for the calculation of A. Thecalculation was performed atT =300 K. For the calculationof B, the ideal solution approximation (φ=1) was used. Weused the simultaneously measured aerosol chemical compo-sition (PM2.5) assuming that chemical composition was notsize-dependent. Assumptions regarding the chemical com-position and chemical properties of organic compounds (e.g.molecular weight (MW), elemental ratio, density) are alsorequired for the calculation ofB. Kawamura et al. (2003)and Mochida et al. (2003) measured dicarboxylic acids andother water-soluble organic compounds over the Sea of Japanand Yellow Sea during the ACE-Asia campaign. The con-centration of oxalic acid was higher than that of other com-pounds by an order of magnitude. Simoneit et al. (2004) alsomeasured dicarboxylic acids at Gosan during the ACE-Asiacampaign. Adipic acid was the largest-MW linear moleculeamong the dicarboxylic acids they measured. Oxalic acid hasone of the smallest molecular weights among atmosphericaerosol organic compounds. Therefore, if we assume thatall carbon atoms in WSOC originated from oxalic acid, wewill estimate a maximum Raoult’s effect. In this study, thisassumption is called as “oxalic acid assumption”. As an ex-treme case, we calculatedDcrit using this assumption. Theaverage molecular weight of WSOC should be higher thanthat of oxalic acid because dicarboxylic acids and saccha-rides with larger molecular weights were also observed atGosan (Simoneit et al., 2004). In fact, Miyazaki et al. (2007)indicated that a significant portion of WSOC measured dur-ing the ABC-EAREX2005 campaign could be attributed toorganic compounds having a MW larger than oxalic acid.To test the sensitivity ofDcrit to the assumed WSOC com-position, we also calculatedDcrit assuming that the aver-age properties (MW, elemental composition, and density) ofwater-soluble organic compounds are equal to those of adipicacid. In this study, this assumption is called the “adipic acidassumption”, although this does not mean that all WSOCoriginated from adipic acid. In addition, recent studieshave shown that significant fractions (20–60%) of WSOCare high-molecular weight compounds such as humic-likesubstances (HULIS) (Graber and Rudich, 2006, and refer-ences therein). CalculatingDcrit under the assumption that

200

150Diam

eter

(nm

)

3/25 3/27 3/29 3/31 4/2 4/4

WSOC = oxalic acid

WSOC = adipic acid

WSOC = HULIS

Dcrit (ammonium sulfate) Dcrit (calculated) D50 (observed)

Figure 9

Fig. 9. Comparison of observedD50 and calculatedDcrit atSS=0.097%. The upper and lower blue lines were calculated us-ing the “HULIS assumption” and “oxalic acid assumption”, respec-tively. See text for details of the calculations.

all WSOC originated from HULIS (HULIS assumption) cor-responds to the lowest estimate of Raoult’s effect. As inother cases, we need the chemical properties of HULIS forthe calculation, although chemical properties of HULIS varydepending on the samples used (Graber and Rudich, 2006).For this calculation, we assumed an elemental compositionof HULIS of C: H: O=1: 1.16: 0.63 (Dinar et al., 2006a),a density of 1.5 g/cm3 (Dinar et al., 2006b; Hoffer et al.,2006), and an average molecular weight of 700 Da (Dinaret al., 2006a). In these calculations, water-soluble organiccompounds were assumed to be completely dissolved in wa-ter (ε=1).

The chemical properties of WIOC are also required forthe calculation ofB. Simoneit et al. (2004) have measuredalkanes, fatty acids, and polycyclic aromatic hydrocarbons(PAHs) at Gosan. Among these compounds, the mass frac-tions of PAHs were very limited. Most of the carbon atomsof alkanes and fatty acids are in methylene groups(−CH2−).Therefore, we assumed that all carbon atoms in WIOC orig-inated from methylene groups. The density of WIOC wasassumed to be 0.8 g/cm3 because the densities of hydrocar-bons and fatty acids are typically 0.7–0.9 g/cm3 (Pang etal., 2006). The properties of EC were assumed to be equalto those of graphite. The values used for the calculation aresummarized in Table 4.

The observedD50 and calculatedDcrit at SS=0.097% arecompared in Fig. 9. This figure shows three calculated re-sults based on the oxalic acid, adipic acid, and HULIS as-sumptions. In all cases, the temporal variations of the cal-culatedDcrit correlate with the observedD50. However, thecalculatedDcrit values are systematically larger than the ob-servations by 20–29%.

The correlations ofD50 and the calculatedDcrit (adipicacid assumption) are shown in Fig. 10. At allSSs, the



Table 4. Values used for the calculation ofDcrit. υ for WIOC and EC are not shown here because they are insoluble in water.

Assumed composition Molecular weight(MW) υ ρ(g/cm3)

Ammonium sulfate (NH4)2SO4 132.14 3 1.77Ammonium nitrate NH4NO3 80.04 2 1.72WSOC (oxalic acid) (COOH)2 90.04 1 1.9WSOC (adipic acid) HOOC(CH2)4COOH 146.14 1 1.36WSOC (HULIS) CH1.16O0.63 700 1 1.5WIOC (–CH2–) 14/carbon atom N/A 0.8EC C (graphite) 12/carbon atom N/A 2.0

2

3

4

5

6

789

100

2

Dcr

it (nm

)

2 3 4 5 6 7 8 9100

2

D50 (nm)

SS = 0.097% (r2 =0.48) SS = 0.27% (r2 =0.32) SS = 0.58% (r2 =0.11) SS = 0.97% (r2 =0.00)

Figure 10

Fig. 10. Scatter plot ofD50 (observed) and calculatedDcrit (cal-culated). The adipic acid assumption was used for the calculationshown in this figure.

calculated diameters are larger than the observations. Thedifferences between the calculatedDcrit and the observedD50 (calc–obs) are summarized in Table 5. In all cases,Dcritwas overestimated by 16–29%.

The r2 values decrease with increasingSS (diminishingD50). This is likely due to the size dependence of the tem-poral variation of the aerosol chemical composition, as dis-cussed in Sect. 4.2.2.

5.3 Possible causes of the discrepancy

In this section, we discuss the possible causes of the differ-ence betweenD50 andDcrit and identify critical assumptionsfor the calculation ofDcrit.

In calculatingB, we assumed the composition and chem-ical properties of each component. The average chemi-

cal properties of water-soluble organic compounds were as-sumed to be identical to those of oxalic acid, adipic acid, orHULIS. Nevertheless, theDcrit values were larger thanD50even in the case of the oxalic acid assumption (the maximumestimation ofB). Therefore, the assumption of the chemicalcomposition of WSOC is not the main cause of this discrep-ancy.

We assumed a density of 0.8 g/cm3 for water-insolubleorganic compounds. If the actual density were larger thanthis value, this would lead to the overestimation ofDcrit be-cause of the underestimation ofB (Eq. 1). However, a sen-sitivity study of the density of WIOC (up to 1.2 g/cm3) un-der the oxalic acid assumption showed that calculatedDcritstill overestimated the observations (Dcrit–D50=20±12 nm atSS=0.097%).

The assumption of complete dissolution corresponds to theupper limit in estimatedB, and therefore the lower limit ofDcrit. For instance, Huff Hartz et al. (2006) showed that theDcrit of slightly soluble organic compounds calculated us-ing the complete dissolution assumption is smaller thanDcritcalculated assuming limited solubility. We used the ideal so-lution approximation in the calculation. If the discrepancyis due to this effect, the osmotic coefficient at the criticaldroplet diameters needs to be increased by a factor of about2, considering the magnitude of the discrepancy (Table 5)and Eq. (2). Nevertheless, such a large change in the osmoticcoefficient is unlikely, as the solution was very dilute at thecritical droplet diameters.

In this study, the aerosol chemical composition was mea-sured at PM2.5. If the chemical compositions of PM2.5 donot represent those atD50 (<200 nm), it causes an error inthe calculation ofDcrit. We discuss this point in detail inSect. 5.3.1.

The surface tension of water was assumed for the calcula-tion of A in Eq. (2). If the decrease in surface tension dueto WSOC was significant, it may affect theDcrit (Facchini etal., 1999). This point is discussed in Sect. 5.3.2.



Table 5. Differences betweenD50 (observed) and calculatedDcrit (Dcrit–D50). The values in parentheses show the magnitude of thedifference in %. “Oxalic acid”, “adipic acid”, and “HULIS” show the oxalic acid, adipic acid, and HULIS assumptions, respectively.

SS=0.097% 0.27% 0.58% 0.97%

Oxalic acid 27±18(nm) 13±10 7± 7 6±6(20±12%) (18±13%) (16±15%) (21±19%)

Adipic acid 34±19 17±11 9±8 8±6(26±13%) (24±15%) (22±17%) (27±22%)

HULIS 39±20 17±12 10±8 8±7(29±14%) (25±16%) (23±18%) (26±23%)

5.3.1 Size dependence of chemical composition

Some studies have used size-resolved aerosol chemical com-position for CCN closure studies (e.g. Cantrell et al., 2001;Roberts et al., 2002). These studies have shown the size-dependence of aerosol chemical composition. In addition,Medina et al. (2007) have shown that the use of size-resolvedchemical composition can decrease the error in the closureof the CCN number concentration. Topping et al. (2004)measured size-resolved chemical composition at Gosan dur-ing the ACE-Asia campaign using a Berner impactor. Theyreported the size distribution of mass fractions of inorganiccompounds and WSOC, and the results clearly showed thesize-dependence. However, it is difficult to use their resultsto estimate the effect of size dependent chemical composi-tion because they did not measure water-insoluble compo-nents of the impactor samples. Mochida et al. (2007) mea-sured aerosol chemical compounds sampled with a Micro-Orifice Uniform Deposit Impactor (MOUDI) over the EastChina Sea and the Sea of Japan during the ACE-Asia cam-paign. They found that the submicron mode was enrichedby non-sea salt (nss)-SO2−

4 ([OC]/[nss-SO2−

4 ]≈0.2 at 0.18–0.56µm), whereas the supermicron mode was enriched byorganic compounds ([OC]/[nss-SO2−

4 ]≈1 at 1.8–5.6µm).The median value of [OC]/[SO2−

4 ] in the present study is 1.0,suggesting that PM2.5 chemical composition was affectedby super-micron particles by comparison with the ratio ofMochida et al. (2007).

The Dcrit for eachSS was calculated using the chemicalcomposition of the sub-micron mode particle given in Table 2in Mochida et al. (2007) assuming that OC was entirely com-posed of WIOC, because the fraction of WSOC was muchsmaller than that of WIOC (Fig. 8), and WSOC concentra-tion was not reported. The major difference of this calcu-lation is the higher sulfate fraction ([OC]/[nss-SO2−

4 ]≈0.2).At SS=0.097%, the observedD50 was 136±17 nm, and thecalculatedDcrit using the data of Mochida et al. (2007) was142 nm, as summarized in Fig. 11. In the case of otherSSs,Dcrit also agrees with the average value ofD50. This result

2

3

4

5

6

789

100

2

Dcr

it (n

m)

2 3 4 5 6 7 8 9100

2

D50 (nm)

SS = 0.58%

SS = 0.097%

SS = 0.97%

SS = 0.27%

Chemical composition from Mochida et al. (2007)

Figure 11

Fig. 11. Comparison of measuredD50 andDcrit calculated usingthe chemical composition of Mochida et al. (2007). See the text fordetailed explanation.

shows possible effects of the size-dependent chemical com-position on CCN activation.

The size dependence of aerosol chemical composition isclearer for the period of new particle formation. TheD50values of newly formed particles were very close to thatof ammonium sulfate, indicating that particles were pre-dominantly composed of inorganic compounds at the time(Sect. 4.2.3.). In particular, newly formed particles grewlarger thanD50 atSS=0.097% on 31 March (Fig. 4), andD50values were nearly equal to theDcrit of ammonium sulfate onthat day (Fig. 9). However, the mass fraction of carbonaceousaerosols (WSOC+WIOC+EC) at PM2.5 was about 30% atthis time (Fig. 8), and this led to the overestimation ofDcrit



70

60

50

40

30

20

surf

ace

tens

ion

(mN

/m)

0.0001 0.001 0.01 0.1C (mol/l)

Jeju Cloud Po Valley fog

Figure 12

Fig. 12.Decrease of surface tension calculated using the parametersfor Jeju cloud water and Po Valley fog by McFiggans et al. (2006).The colored area in the figure corresponds to the values at the criti-calSSs.

(Fig. 9). The decrease ofD50 associated with new/secondaryparticle formation was also observed on 25 March. In thiscase, the mass fraction of water-insoluble compounds wasalso 30–40%, and this caused the overestimation ofDcrit.These results suggest that the PM2.5 mass concentration wasbiased by large (>200-nm) particles, at least during these pe-riods.

5.3.2 Decrease of surface tension by organic compounds

As described above, the surface tension of water was as-sumed for the calculation in this study. Nevertheless, if thedecrease in surface tension due to organic compounds wassignificant, it may affect the criticalSS of the particles (Fac-chini et al. 1999). Surface tension lowering effects have beenobserved in various regions of the world such as Po Valleyin Italy (Facchini et al. 1999), Mace Head in Ireland (Cavalliet al. 1999), and the Great Hungarian Plain (Kiss et al. 2005.In particular, Decesari et al. (2005) measured the surface ten-sion lowering effect of aerosol, cloud water, and wet deposi-tion samples at Jeju Island during the ACE-Asia campaign.They showed that the decrease of the surface tension due toaerosol was relatively small, whereas the surface tension de-crease of cloud water and wet-deposition samples was signif-icant. McFiggans et al. (2006) compared these results, andthey showed that the magnitude of the effect for cloud waterat Jeju Island was the most significant, and the effects for PoValley and Mace Head samples were not as large as that ofcloud water at Jeju (the red and blue lines in Fig. 12).

They summarized the fitted parameters of theSzyszkowski-Langmuir Eq. (3) obtained for these ob-servation results,

σ = σ0 − bT ln(1 + aC) (3)

1.5

1.0

0.5

Dcr

it/D

50

1.00.80.60.40.20.0SS (%)

72 mN/m Po Valley fog Jeju Cloud

Figure 13

Fig. 13. Comparison ofDcrit/D50 calculated using several surfacetension parameters. The calculation was performed using the adipicacid approximation.

whereσ0 is the surface tension of pure water,C is the WSOCconcentration of the solution, anda andb are the empiricalparameters obtained by fitting the observational results. Inorder to investigate the sensitivity ofDcrit to surface tension,we calculatedDcrit at eachSS using the parameters given byMcFiggans et al. (2006). In this calculation, the adipic acidassumption was used, and the values for Jeju cloud water andPo Valley fog were employed. Substitutingσ in Eq. (1) bythat expressed by Eq. (3), theDcrit values were obtained bynumerical calculation. The results are summarized in Fig. 13.In the case of Jeju cloud water, the calculated diameters forall SSs are smaller than the observedD50 by 10–47%, andthe use of the equation for Po Valley decreased the differ-ences betweenDcrit andD50 (7 to −13%). This indicatesthat the decrease of surface tension can potentially explainthe discrepancy. The surface tension at the criticalSS wasabout 46–56 mN/m (Jeju cloud water) and 51–62 mN/m (PoValley), respectively (Fig. 12).

The discussion in this section and Sect. 5.3.1 shows thatthe size- dependence of chemical composition and the de-crease in surface tension are the critical parameters in esti-mating Dcrit. For a more quantitative assessment of theseeffects, we need simultaneous measurements of these param-eters andD50.

5.4 Impact ofDcrit on CCN number concentration

The main purpose of the calculation ofDcrit is the preciseestimation of CCN number concentration from the particlesize distribution. We estimated the difference in CCN num-ber concentration due to the difference inD50 and calculatedDcrit. The following equation was used for the assessment of



the difference:

1NCCN

NCCN= −

log(Dcrit)∫fN (logDp)d logDp

log(D50)

NCCN(4)

whereNCCN is the number concentration of CCN measuredby the CCN counter,1NCCN is the difference of the CCNnumber concentration caused by the error in the estimationof Dcrit, andfN (logDp) is the number size distribution mea-sured by the SMPS. Therefore,1NCCN/NCCN is the ratioof the difference of the CCN number concentration due tothe difference inDcrit and the observed CCN number con-centration.1NCCN/NCCN depends on the number size dis-tribution (fN (logDp)) and the uncertainty of the chemicalcomposition (Dcrit–D50). For this calculation, the calculatedDcrit (adipic acid assumption, surface tension of water) wasused. In addition to the adipic acid assumption, we also cal-culated1NCCN/NCCN using the ammonium sulfate assump-tion (Dcrit=Dcrit of ammonium sulfate) because this assump-tion has frequently been used in previous CCN studies (e.g.VanReken et al., 2003), including those for Gosan (Yum etal., 2007). The results are summarized in Fig. 14. In general,the adipic acid approximation underestimates the CCN num-ber concentration due to the overestimation ofDcrit, whereasthe ammonium sulfate approximation overestimates the CCNnumber concentration because of the underestimation of theDcrit. At SS=0.097%,1NCCN/NCCN was−0.27±0.14 and0.16±0.18 for the adipic acid and ammonium sulfate approx-imations, respectively. These values give a measure of theuncertainty associated with the CCN prediction based on theresults of this study. The absolute value of1NCCN/NCCNincreased with decreasingSS. During the observation pe-riod, the average values ofB did not show a significant de-pendence on the value ofSS as discussed in Sect. 4.2.2, andthe magnitude of the error associated with the estimation ofDcritdid not depend onSS significantly. Thus, this trend wasnot mainly due to the size dependence of the chemical com-position. Dcrit decreased with increasingSS. The particlenumber concentration betweenD50 andDcrit was smaller athigherSS in comparison with the CCN number concentra-tion because at higherSS a larger fraction of CCN is in thesize range larger thanD50, as can be seen from Fig. 6. Theseresults show that the effect of chemical composition onNCCNwas more important at lowerSS, and the aerosol number sizedistribution was important at higherSS.

6 Summary

We measured the CCN number concentration (NCCN) andthe size-resolved CCN/CN ratios atSS=0.097, 0.27, 0.58,and 0.97% on Jeju Island, Korea, during 18 March–5 April2005. The average CCN number concentrations for thewhole observational period were as high as 1194±746 cm−3

1 2 3 4 5 6 7 8 9

10 11 12 13 14 15 16 17 18 19 20

-0.4

-0.2

0.0

0.2

0.4

ΔN C

CN/N

CCN

1.00.80.60.40.20.0SS (%)

adipic acid assumption ammonium sulfate assumption

Figure 14

Fig. 14.Average values and standard deviations of1NCCN/NCCN.

(SS=0.097%) and 3966±1686 cm−3 (SS=0.97%). TheCCN/CN ratios monotonically increased with increasing par-ticle diameter, and the diameter at CCN/CN=0.5 was de-fined asD50. The averageD50 values atSS=0.097% and0.97 % were 136 nm and 31 nm, respectively. These valueswere slightly larger than those of ammonium sulfate parti-cles (125 and 27 nm atSS=0.097% and 0.97%, respectively).In particular,D50 diameters of newly formed particles andthose of ammonium sulfate particles were almost identicalfor SS=0.97, 0.58 and 0.27%, indicating that these particleswere mainly composed of ammonium sulfate. The peak di-ameter of the CCN size distribution averaged for the wholeobservational period was about 150 nm atSS=0.097% andshifted to about 80 nm atSS=0.97%.

The temporal variation ofD50 at SS=0.097% was nega-tively correlated with the variation of the water-soluble frac-tion (inorganics + WSOC) of the aerosol components. Forquantitative comparison, the threshold diameters for CCNactivation (Dcrit) were calculated by Kohler theory assum-ing the surface tension of water and PM2.5 aerosol chemi-cal composition. The calculatedDcrit values were correlatedwith D50 at SS=0.097% (r2=0.48). However,Dcrit was sys-tematically larger thanD50 by about 16–29%. Sensitivitystudies have shown that this discrepancy can be explainedby possible differences in aerosol chemical composition be-tween the sub-micron and super-micron size ranges. In addi-tion, a decrease of the surface tension due to the existence ofWSOC can also significantly decreaseDcrit.

The particle number concentrations in the size range be-tweenD50 andDcrit(1NCCN) were calculated using the ob-served size distribution. The ratios of1NCCN to NCCN(1NCCN/NCCN) were estimated to be−0.27±0.14 and−0.10±0.13 atSS=0.097% and 0.97%, respectively, assum-ing that water-soluble organic compounds are represented byadipic acid. The1NCCN/NCCN ratio gives a measure of theuncertainty in estimating CCN number concentrations using



particle number size distributions and PM2.5 chemical com-positions in the East Asia region.

Acknowledgements.We acknowledge anonymous reviewers fortheir useful comments. We also thank C. S. Hong and Y. J. Kim fortheir support of the observations. This work was supported by theMinistry of Education, Culture, Sports, Science, and Technology(MEXT), the Japanese Science and Technology Agency (JST),and the global environment research fund of the Japanese Ministryof the Environment. M. Kuwata thanks the Japan Society for thePromotion of Science (JSPS) for a JSPS Research Fellowship forYoung Scientists. J. H. Kim and S. S. Yum were supported by theKorean Ministry of Environment under the “Eco-techonopia 21project”.

Edited by: C. Chan

References

Akimoto, H.: Global air quality and pollution, Science, 302, 1716–1719, 2003.

Arnold, F.: Atmospheric aerosol and cloud condensation nuclei for-mation: A possible influence of cosmic rays?, Space. Sci. Re-views, 125, 169–186, 2006.

Bae, M.-S., Schauer, J. J., DeMinter, J. T., Turner, J. R., SmithD., and Cary, R. A.: Validation of a semi-continuous instrumentfor elemental carbon and organic carbon using a thermal-opticalmethod, Atmos. Environ., 38, 2885–2893, 2004.

Buzorius, G., McNaughton, C. S., Clarke, A. D., Covert,D. S., Blomquist, B., Nielsen, K., and Brechtel, F. J.:Secondary aerosol formation in continental outflow condi-tions during ACE-Asia, J. Geophys. Res., 109, D24203,doi:10.1029/2004JD004749, 2004.

Cantrell, W., Shaw, G., Cass, G. R., Chowdhury, Z., Hughes, L.S., Prather, K. A., Guazzotti, S. A., and Coffee, K. R.: Clo-sure between aerosol particles and cloud condensation nuclei atKaashidhoo Climate Observatory, J. Geophys. Res., 106(D22),28 711–28 718, 2001.

Carmichael, G. R., Hong, M., Ueda, H., Chen, L., Murano, K.,Park, J. K., Lee, H., Kim, Y., Kang, C., and Shim, S.: Aerosolcomposition at Cheju Island, Korea, J. Geophys. Res., 102(D5),6047–6062, 1997.

Cavalli, F., Facchini, M. C., Decesari, S., et al.: Advancesin characterization of size-resolved organic matter in marineaerosol over the North Atlantic, J. Geophys. Res., 109, D24215,doi:10.1029/2004JD005137, 2004.

Chameides, W. L., Luo, C., Saylor, R., Streets, D. G., Huang,Y., Bergin, M., and Giorgi, F.: Correlation between model-calculated anthropogenic aerosols and satellite-derived cloud op-tical depths: Indication of indirect effect?, J. Geophys. Res.,107(D10), 4085, doi:10.1029/2000JD000208, 2002.

Clegg, S. L., Rard, J. A., and Pitzer, K. S.: Thermodynamic prop-erties of 0-6 mol Kg-1 aqueous sulfuric acid from 273.15K to328.15K, J. Chem. Soc. Faraday Trans., 90, 1875–1894, 1994.

Clegg, S. L., Milioto S., and Palmer, D. A.: Osmotic and activitycoefficients of aqueous (NH4)2SO4 as a function of temperature,and (NH4)2SO4 - H2SO4 mixtures at 298.15 K and 323.15 K, J.Chem. Eng. Data, 41, 455–467, 1996.

Corrigan, C. E. and Novakov, T.: Cloud condensation nucleus activ-ity of organic compounds: A laboratory study, Atmos. Environ.,33, 2661–2668, 1999.

Covert, D. S., Gras, J. L., Wiedensohler A., and Stratmann, F.:Comparison of directly measured CCN with CCN modeled fromthe number-size distribution in the marine boundary layer dur-ing ACE 1 at Cape Grim, Tasmania, J. Geophys. Res., 103(D13),16 597–16 608, 1998.

Cruz, C. N. and Pandis, S. N.: Deliquescence and hygroscopicgrowth of mixed inorganic-organic atmospheric aerosol, Envi-ron. Sci. Technol., 34, 4313–4319, 2000.

Decesari, S., Facchini, M. C., Fuzzi, S., McFiggans, G. B., Coe, H.,and Bower, K. N.: The water-soluble organic component of size-segregated aerosol, cloud water and wet depositions from JejuIsland during ACE-Asia, Atmos. Environ., 39, 211–222, 2005.

Dinar, E., Taraniuk, I., Graber, E. R., Katsman, S., Moise, T.,Anttila, T., Mentel, T. F., and Rudich, Y.: Cloud Condensa-tion Nuclei properties of model and atmospheric HULIS, Atmos.Chem. Phys., 6, 2465–2481, 2006a,http://www.atmos-chem-phys.net/6/2465/2006/.

Dinar, E., Mentel, T. F., and Rudich, Y.: The density of humic acidsand humic like substances (HULIS) from fresh and aged woodburning and pollution aerosol particles, Atmos. Chem. Phys., 6,5213–5224, 2006b,http://www.atmos-chem-phys.net/6/5213/2006/.

Dusek, U., Frank, G. P., Hildebrandt, L., Curtius, J., Schneider, J.,Walter, S., Chand, D., Drewnick, F., Hings, S., Jung, D., Bor-rmann, S., and Andreae, M. O.: Size matters more than chem-istry for cloud-nucleating ability of aerosol particles, Science,312, 1375–1378, 2006.

Facchini, M. C., Mircea, M., Fuzzi, S., and Charlson, R. J.: Cloudalbedo enhancement by surface-active organic solutes in growingdroplets, Nature, 401, 257–259, 1999.

Grabber, E. R. and Rudich Y: Atmospheric HULIS: How humic-likeare they? A comprehensive and critical review, Atmos. Chem.Phys., 6, 729–753, 2006,http://www.atmos-chem-phys.net/6/729/2006/.

Hoffer, A., Gelencser, A., Guyon, P., Kiss, G.., Schmid, O., Frank,G. P., Artaxo, P., and Andrea, M. O.: Optical properties of humic-like substances (HULIS) in biomass-burning aerosols, Atmos.Chem. Phys., 6, 3563–3570, 2006,http://www.atmos-chem-phys.net/6/3563/2006/.

Hudson, J. G. and Yum, S. S.: Cloud condensation nuclei spectraand polluted and clean clouds over the Indian Ocean, J. Geophys.Res., 107(D19), 8022, doi:10.1029/2001JD000829, 2002.

Huff Hartz, K. F., Tischuk, J. E., Chan, M. N., Chan, C. K., Don-ahue, N. M., and Pandis, S. N.: Cloud condensation nuclei acti-vation of limited solubility organic aerosol, Atmos. Environ., 40,605–617, 2006.

Kawamoto, K., Hayasaka, T., Uno, I., and Ohara, T.: Acorrelative study on the relationship between modeled an-thropogenic aerosol concentration and satellite-observed cloudproperties over east Asia, J. Geophys. Res., 111, D19201,doi:10.1029/2005JD006919, 2006.

Kawamura, K., Umemoto, N., Mochida, M., Bertram, T., Howell,S., and Huebert, B. J.: Water-soluble dicarboxylic acids in thetropospheric aerosols collected over east Asia and western NorthPacific by ACE-Asia C-130 aircraft, J. Geophys. Res., 108(D23),8639, doi:10.1029/2002JD003256, 2003.


http://www.atmos-chem-phys.net/6/2465/2006/





Kim, J., Yoon, S.-C., Jefferson, A., and Kim, S.-W.: Aerosol hy-groscopic properties during Asian dust pollution, and biomassburning episodes at Gosan, Korea in April 2001, Atmos. Envi-ron., 40, 1550–1560, 2006.

Kiss, G., Tombacz, E., and Hanson, H.-C.: Surface tension effectsof Humic-like substances in the aqueous extract of troposphericfine aerosol, J. Atmos. Chem., 50, 279–294, 2005.

Knutson, E. O. and Whitby, K. T.: Aerosol classification by electricmobility: Apparatus, theory, and applications, J. Aerosol. Sci., 6,443–451, 1975.

Kondo, Y., Komazaki, Y., Miyazaki, Y., et al.: Temporal variationsof elemental carbon in Tokyo, J. Geophys. Res., 111, D12205,doi:10.1029/2005JD006257, 2006.

Kondo, Y., Miyazaki, Y., Takegawa, N., Miyakawa, T., Weber, R. J.,Jimenez, J. L., Zhang, Q., and Worsnop, D. R.: Oxygenated andwater-soluble organic aerosols in Tokyo, J. Geophys. Res., 112,D01203, doi:10.1029/2006JD007056, 2007.

Kreidenweis, S. M., Koehler, K., DeMott, P. J., Prenni, A. J., Car-rico, C., and Ervens, B.: Water activity and activation diametersfrom hygroscopicity data – Part I: Theory and application to in-organic salts, Atmos. Chem. Phys., 5, 1357–1370, 2005

Kuwata, M., Kondo, Y., Mochida, M., Takegawa, N., and Kawa-mura, K.: Dependence of CCN activity of less volatile particleson the amount of coating observed in Tokyo, J. Geophys. Res.,112, D11207, doi:10.1029/2006JD007758, 2007.

Laaksonen, A., Hamed, A., Joutsensaari, J., Hiltunen, L., Cavalli,F., Junkermann, W., Asmi, A., Fuzzi, S., and Facchini, M. C.:Cloud condensation nucleus production from nucleation eventsat a highly polluted region, Geophys. Res. Lett., 32, L06812,doi:10.1029/2004GL022092, 2005.

Lee, J. Y., Kim, Y. P., Kang, C.-H., Ghim, Y. S., and Kaneyasu,N.: Temporal trend and long-range transport of particulatepolycyclic aromatic hydrocarbons at Gosan in northeast Asiabetween 2001 and 2004, J. Geophys. Res., 111, D11303,doi:10.1029/2005JD006537, 2006.

Lohmann, U. and Feichter, J.: Global indirect aerosol effects: Areview, Atmos. Chem. Phys., 5, 715–737, 2005,http://www.atmos-chem-phys.net/5/715/2005/.

Markowski, G. R.: Improving Twomey’s algorithm for inversion ofaerosol measurement data, Aerosol. Sci. Technol., 7, 127–141,1987.

McFiggans, G.,Artaxo, P., Baltensperger, U., et al.: The effect ofphysical and chemical aerosol properties on warm cloud dropletactivation, Atmos. Chem. Phys., 6, 2593–2649, 2006,http://www.atmos-chem-phys.net/6/2593/2006/.

Matsumoto, K., Tanaka, H., Nagao, I., and Ishizaka, Y.: Contribu-tion of particulate sulfate and organic carbon to cloud condensa-tion nuclei in the marine atmosphere, Geophys. Res. Lett., 24(6),655–658, 1997.

Medina, J., Nenes, A., Sotiropoulou, R.-E. P., Cottrell, L. D.,Ziemba, L. D., Beckman, P. J, and Griffin, R. J.: Cloud conden-sation nuclei closure during the International Consortium for At-mospheric Research on Transport and Transformation 2004 cam-paign: Effects of size-resolved composition, J. Geophys. Res.,112, D10S31, doi:10.1029/2006JD007588, 2007.

Miyazaki, Y., Kondo, Y., Takegawa, N., Komazaki, Y., Fukuda,M., Kawamura, K., Mochida, M., Okuzawa, K., and We-ber, R. J.: Time-resolved measurements of water-soluble or-ganic carbon in Tokyo, J. Geophys. Res., 111, D23206,

doi:10.1029/2006JD007125, 2006.Miyazaki, Y., Kondo, Y., Kodama, D., Han, S., Koike, M., Ko-

mazaki, Y., Tanimoto, H., and Matsueda, H.: Chemical charac-teristics of water-soluble organic carbon in the Asian outflow, J.Geophys. Res., 112, D22S30, doi:10.1029/2007JD009116,2007.

Mochida, M., Kawamura, K., Umemoto, N., Kobayashi, M.,Matsunaga, S., Lim, H., Turpin, B. J., Bates, T. S., and Si-moneit, B. R. T.: Spatial distributions of oxygenated organiccompounds (dicarboxylic acids, fatty acids, and levoglucosan)in marine aerosols over the western Pacific and off the coastof East Asia: Continental outflow of organic aerosols duringthe ACE-Asia campaign, J. Geophys. Res., 108(D23), 8638,doi:10.1029/2002JD003249, 2003.

Mochida, M., Kuwata, M., Miyakawa, T., Takegawa, N.,Kawamura, K., and Kondo, Y.: Relationship between hy-groscopicity and cloud condensation nuclei activity for ur-ban aerosols in Tokyo, J. Geophys. Res., 111, D23204,doi:10.1029/2005JD006980, 2006.

Mochida, M., Umemoto, N., Kawamura, K., Lim, H.-J., and Turpin,B. J.: Bimodal size distributions of various organic acids andfatty acids in the marine atmosphere: Influence of anthropogenicaerosols, Asian dusts, and sea spray off the coast of East Asia, J.Geophys. Res., 112, D15209, doi:10.1029/2006JD007773, 2007.

O’Dowd, C. D., Hmeri, K., Mkel, J. M., et al.: A dedicated studyof New Particle Formation and Fate in the Coastal Environ-ment (PARFORCE): Overview of objectives and achievements,J. Geophys. Res., 107(D19), 8108, doi:10.1029/2001JD000555,2002.

Orsini, D. A., Ma, Y., Sullivan, A., Sierau, B., Baumann, K., andWeber, R. J.: Refinements to the particle–into-liquid sampler(PILS) for ground and airborne measurements of water solubleaerosol composition, Atmos. Environ., 37, 1243–1259, 2003.

Pang Y., Turpin, B. J., and Gundel, L. A.: On the importanceof organic oxygen for understanding organic aerosol particles,Aerosol. Sci . Technol., 40, 128–133, 2006.

Raymond, T. M. and Pandis, S. N.: Cloud activation ofsingle-component organic aerosol particles, J. Geophys. Res.,107(D24), 4787, doi:10.1029/2002JD002159, 2002.

Reade, L., Jennings, S. G., and McSweeney, G.: Cloud condensa-tion nuclei measurements at Mace Head, Ireland, over the period1994–2002, Atmos. Res., 82, 610–621, 2006.

Richter, A., Burrows, J. P., Nub, C. Grainer, and Niemeier, U:Increase in tropospheric nitrogen dioxide over China observedfrom space, Nature, 437, 129–132, 2005.

Roberts, G. C., Artaxo, P., Zhou, J., Swietlicki, E., and Andreae, M.O.: Sensitivity of CCN spectra on chemical and physical proper-ties of aerosol: A case study from the Amazon Basin, J. Geophys.Res., 107(D20), 8070, doi:10.1029/2001JD000583, 2002.

Roberts, G. C. and Nenes, A.: A continuous-flow streamwisethermal gradient CCN chamber for atmospheric measurements,Aerosol. Sci. Technol., 39, 206–221, 2005.

Sawa, Y., Tanimoto, H., Yonemura, S., Matsueda, H., Wada,A., Taguchi, S., Hayasaka, T., Tsuruta, H., Tohjima, Y.,and Mukai, H.: Widespread pollution events of carbonmonoxide observed over the western North Pacific during theEAREX 2005 campaign, J. Geophys. Res., 112, D22S26,doi:10.1029/2006JD008055, 2007.

Seinfeld, J. H. and Pandis, S. N.: Atmospheric Chemistry andPhysics, John Wiley and Sons, Inc., New York, 2006.





Simoneit, B. R. T., Kobayashi, M., Mochida, M., Kawamura, K.,Lee, M., Lim, H.-J., Turpin, B. J., and Komazaki, Y.: Composi-tion and major sources of organic compounds of aerosol particu-late matter sampled during the ACE-Asia campaign, J. Geophys.Res., 109, D19S10, doi:10.1029/2004JD004598, 2004.

Sotiropoulou, R. E. P., Tagaris, E., Pilinis, C., Anttila, T., and Kul-mala, M.: Modeling new particle formation during air pollu-tion episodes: Impacts on aerosol and cloud condensation nuclei,Aerosol. Sci . Technol., 40, 557–572, 2006.

Streets, D. G., Tsai, N. Y., Akimoto, H., and Oka, K.: Sulfur dioxideemissions in Asia in the period 1985–1997, Atmos. Environ., 34,4413–4424, 2000.

Sullivan, A. P., Weber, R. J., Clements, A. L., Turner, J. R., Bae,M. S., and Schauer, J. J.: A method for on-line measurementof water-soluble organic carbon in ambient aerosol particles:Results from an urban site, Geophys. Res. Lett., 31, L13105,doi:10.1029/2004GL019681, 2004.

Takegawa, N., Miyazaki, Y., Kondo, Y., Komazaki, Y., T.Miyakawa, T., Jimenez, J. L., Jayne, J. T., Worsnop, D. R., Allan,J., and Weber, R. J.: Characterization of an Aerodyne AerosolMass Spectrometer (AMS): Intercomparison with other aerosolinstruments, Aerosol Sci. Technol., 39, 760–770, 2005.

Tanimoto, H., Sawa, Y., Matsueda, H., Wada, A., Yonemura, S.,Mukai, H., Wang, T., Poon, S., Wong, A., Lee, G., Jung,J. Y., Kim, K. R., Lee, M. H., Lin, N. H., Wang, J. L.,Ou-Yang, C. F, and Wu, C. F.: Evaluation of standards andmethods for continuous measurements of carbon monoxide atground-based sites in Asia, Pap. Met. Geophys., 58, 85–93,doi:10.2467/mripapers.58.85, 2007.

Topping, D., Coe, H., McFiggans, Burgess, G. R., Allan, J., Al-farra, M. R., Bower, K., Choularton, T. W., Decesari, S., andFacchini, M. C.: Aerosol chemical characteristics from samplingconducted on the Island of Jeju Korea during ACE Asia, Atmos.Environ., 38, 2111–2123, 2004.

Twomey, S.: Pollution and the planetary albedo, Atmos. Environ.,8, 1251–1256, 1974.

Van Reken, T. M., Rissman, T. A., Roberts, G. C., Varut-bangkul, V., Jonsson, H. H., Flagan, R. C, and Seinfeld, J.H.: Toward aerosol/cloud condensation nuclei (CCN) closureduring CRYSTAL-FACE, J. Geophys. Res., 108(D20), 4633,doi:10.1029/2003JD003582, 2003.

Yum, S. S., Hudson, J. G., Song, K. Y., and Choi, B.-C.: Springtime cloud condensation nuclei concentrations onthe west coast of Korea, Geophys. Res. Lett., 32, L09814,doi:10.1029/2005GL022641, 2005.

Yum, S. S., Roberts, G., Kim, J. H., Song, K., and Kim,D.: Submicron aerosol size distributions and cloud conden-sation nuclei concentrations measured at Gosan, Korea dur-ing the ABC-EAREX 2005, J. Geophys. Res., 112, D22S32,doi:10.1029/2006JD008212., 2007.

Wiedensohler, A.: Technical Note. An Approximation of the bipo-lar charge distribution for particles in the submicron size range,J. Aerosol. Sci., 19, 387–389, 1988.


Cloud condensation nuclei activity at Jeju Island, Korea in ...

Documents