Anionic, Cationic, and Nonionic Surfactants in Atmospheric Aerosols

Feb 19, 2016 - analysis was applied to PM2.5 aerosols collected at the Baltic station of. Askö ..... Baltic sea were available in the SHARK database ...
0 downloads 0 Views 940KB Size
Subscriber access provided by RUTGERS UNIVERSITY

Article

Anionic, Cationic, and Non-ionic Surfactants in Atmospheric Aerosols from the Baltic Coast at Askö, Sweden: Implications for Cloud Droplet Activation Violaine Myriam Francine Gérard, Barbara Noziere, Christine Baduel, Ludovic Fine, Amanda Ann Frossard, and Ronald Carl Cohen Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.5b05809 • Publication Date (Web): 19 Feb 2016 Downloaded from http://pubs.acs.org on February 22, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 34

Environmental Science & Technology

1

Anionic, Cationic, and Non-ionic Surfactants in

2

Atmospheric Aerosols from the Baltic Coast at

3

Askö, Sweden: Implications for Cloud Droplet

4

Activation

5

Violaine Gérard,a Barbara Nozière,a * Christine Baduel,b† Ludovic Fine,a Amanda A. Frossard,c

6

and Ronald C. Cohenc

7

a

8

Université Lyon 1, 69626 Villeurbanne, France.

9

b

Institut de Recherches sur la Catalyse et l’Environnement de Lyon (IRCELYON), CNRS and

Department of Applied Environmental Science (ITM), Stockholm University, Stockholm,

10

Sweden

11

c

12

California, Berkeley, USA.

13

ABSTRACT. Recent analyses of atmospheric aerosols from different regions have demonstrated

14

the ubiquitous presence of strong surfactants and evidenced surface tension values, σ, below 40

15

mN m-1, suspected to enhance the cloud-forming potential of these aerosols. In this work, this

16

approach was further improved and combined with absolute concentration measurements of

17

aerosol surfactants by colorimetric titration. This analysis was applied to PM2.5 aerosols

18

collected at the Baltic station of Askö, Sweden, from July to October 2010. Strong surfactants

Departments of Chemistry & Department of Earth and Planetary Science, University of

ACS Paragon Plus Environment

1

Environmental Science & Technology

Page 2 of 34

19

were found in all the sampled aerosols, with σ = (32 – 40) ± 1 mNm-1 and concentrations of at

20

least 27 ± 6 mM or 100 ± 20 pmol m-3. The absolute surface tension curves and Critical Micelle

21

Concentrations (CMC) determined for these aerosol surfactants show that 1) surfactants are

22

concentrated enough in atmospheric particles to strongly depress the surface tension until

23

activation, and 2) the surface tension does not follow the Szyszkowski equation during

24

activation, but is nearly constant and minimal, which provide new insights on cloud droplet

25

activation. In addition, both the CMCs determined and the correlation (R2 ~ 0.7) between aerosol

26

surfactant concentrations and chlorophyll-a seawater concentrations suggest a marine and

27

biological origin for these compounds.

28

INTRODUCTION

29

Clouds are essential components of the Earth’s hydrological and climate systems but

30

predicting their formation is still beyond the capacity of current models.1 These limits are due in

31

part to computational challenges but also to an incomplete understanding of some fundamental

32

physical-chemical processes.1 The formation and growth of cloud droplets from the condensation

33

of water onto aerosol particles is described by Köhler theory, in which surface tension is one of

34

the two fundamental parameters.2 But, beside a few early studies,3-7 the role of surfactants on

35

cloud formation has been largely considered negligible. The surface tension of growing droplets

36

is thus assumed equal to that of pure water in all the investigations of cloud droplet formation.8-11

37

Recent laboratory and field studies are starting to bring evidence of the contrary. The importance

38

of surface properties or surfactant fractions for the growth of water droplets has thus been

39

demonstrated in laboratory, both with model surfactants and with authentic aerosol samples.12-15

40

The observation of Cloud Condensation Nuclei (CCN) numbers that were 30-50 % larger than

41

expected in different regions of the atmosphere with a modified on-line instrument was

ACS Paragon Plus Environment

2

Page 3 of 34

Environmental Science & Technology

42

accounted for by the presence of strong surfactants in the particles, with σ ~ 50 mN m-1.16-17 The

43

presence of surfactants in atmospheric aerosols was also confirmed by the direct analysis of

44

aerosol samples. Their concentrations have been determined with colorimetric techniques18-21

45

and relative electrochemical techniques.22-23 More recently, a targeted extraction method allowed

46

the surfactant fraction of atmospheric aerosols to be isolated, opening the possibility for more

47

direct analyses of these compounds.24-26 Surface tension measurements of such extracts revealed

48

the presence of much stronger surfactants than previously expected (σ = 30 – 40 mN m-1).24-25

49

They also showed their slow equilibration kinetics in atmospheric particles, which is suspected to

50

account for their lack of detection by most on-line instruments.26

51

In this work, this extraction method was further improved and combined with concentration

52

measurements to fully characterize the physical-chemical properties of aerosol surfactants and

53

provide unique new information for their role in cloud formation, which can not be obtained by

54

on-line techniques. This analysis was applied to PM2.5 aerosol fractions collected at the marine

55

research station of Askö, Sweden, a site mostly influenced by marine and biogenic emissions.

56

EXPERIMENTAL

57

Sampling

58

The aerosols analyzed in this work were collected at the marine research station of Askö,

59

Sweden (58° 49.5’ N, 17° 39’ E), on the Baltic coast, 80 km south of Stockholm from July to

60

October 2010 (see sample list on Table 1). The samples were collected at ground level on 47

61

mm-Quartz fiber filters (SKC) with a Leckel SEQ 47/50 Sequential Filter Sampler equipped with

62

PM2.5 inlets. Each sample was collected over 72 h at 2.3 m3 h-1, corresponding to about 165 m3

63

of air. Prior to sampling, the Quartz filters were heated at 873 K for 12 h to remove

64

contaminants. For quality analysis, blank samples (about 1 for every 3 filter samples) were also

ACS Paragon Plus Environment

3

Environmental Science & Technology

Page 4 of 34

65

taken as clean filters placed in the sampler for 72 h with the sampling flow off. After sampling,

66

the filters were packed in plastic Petri boxes and stored in a freezer (255 K) until analysis. The

67

total aerosol volume sampled on each filter was determined by weighting the filters before and

68

after sampling under controlled temperature (T = 293 K) and humidity conditions (RH = 50 %),

69

with an accuracy of 10-3 mg and assuming a density of 1 g cm-3. For the extractions, the filter

70

samples were grouped together to obtain a sampled mass of about 2 mg, to exceed the detection

71

limit for surface tension measurements,25 which resulted in a total of 11 samples.

72

Improvement of the extraction method

73

The extraction method developed recently24-25 involved the following steps:

74

- 1) extraction of the filters in MilliQ® water followed by filtration,

75

- 2) microextraction of the water extracts onto silicon tubes,

76

- 3) recovery of the compounds of interest by drying the silicon tubes, eluting with methanol,

77 78

evaporating, and re-dissolving in MilliQ water. In the present work, the silicon tubes were replaced by solid phase extraction (SPE) cartridges

79

using a derivative of silicon, silice C18,25,

80

extraction procedure thus involved the following steps:

81 82 83 84

27-28

providing better reproducibility. The modified

- 1) extraction of the-Quartz filters in 7 mL of MilliQ water for 2 h at 279 ± 1 K, followed by filtration, - 2) a solid phase extraction (SPE) on C18-E cartridges (500 mg / 3 mL, Phenomenex) and elution with 4 mL of acetonitrile,

85

- 3) evaporation of the eluted solution with N2 and redissolution in 60 µL of MilliQ water.

86

The same procedure was applied to the blank filters. The extraction efficiency for this

87

extraction procedure, and for each class of surfactant (anionic, cationic and non-ionic), was

ACS Paragon Plus Environment

4

Page 5 of 34

Environmental Science & Technology

88

determined as the ratio of the concentrations of reference compounds in known solutions before

89

and after extraction, the latter being quantified by the colorimetric techniques described below.

90

For this, 10-9 to 10-4 moles of reference surfactants were spiked onto clean Quartz filters, which

91

were dried for 24 h, then extracted following the steps described above. The efficiency was

92

determined for nine reference surfactants, representing the three classes of surfactants: Sodium

93

dodecyl sulfate (SDS), and Dioctyl sulfosuccinate sodium (AOT) representing anionic

94

surfactants, benzyltetradecyl dimethylammonium (zephiramine), cetyltrimethyl ammonium

95

chloride (CTAC) representing cationic surfactants, and (1,1,3,3-Tetramethylbutyl) phenyl-

96

polyethylene glycol (Triton X114), Polyethylene glycol dodecyl ether (Brij®35), Surfactin,

97

Rhamnolipid, and L-α-Phosphatidylcholine, representing non-ionic surfactants. For anionic

98

surfactants, this extraction efficiency was 65 ± 10 %, for cationic surfactants 20 ± 5 %, and for

99

non-ionic surfactants, 90 ± 10 %. These efficiencies were taken into account in the detection

100

limits and concentration values obtained with this method.

101

For the aerosol samples, the extraction technique was shown to remove the entire surfactant

102

fraction of these samples, i.e. all the aerosol components contributing to the surface tension,25 by

103

measuring the surface tension of the residual solutions obtained after the first and second

104

extraction steps. After the first step (water extraction) the surface tension of the residual

105

solutions was around 50 mN m-1, which was consistent with the surface tension of aerosol

106

samples resulting from a simple water extraction.5 After the second extraction step, the surface

107

tension of the residual aqueous solutions had increased back to 72.8 ± 1 mN m-1, i.e. the value of

108

pure water, thus demonstrating that all the surface-active compounds had been transferred to the

109

extracts.

ACS Paragon Plus Environment

5

Environmental Science & Technology

Page 6 of 34

110

The 60 µL-extract obtained from each aerosol sample were used as parent solution for both

111

surface tension and concentration measurements. The surface tension of the extracts were

112

measured first, then the corresponding surface tension curves were determined by successively

113

diluting the extracts in MilliQ water. The diluted solutions (10 mL) obtained at the end of this

114

procedure were then separated in 3 aliquots and used for the measurement of anionic, cationic,

115

and non-ionic surfactant concentrations.

116

Surface tension and surface tension curves measurements

117

The surface tension of reference solutions and sample extracts was measured with the hanging

118

droplet method using a Dataphysics OCA 15EC tensiometer and Dataphysics SCA software for

119

OCA version 4-4.1. In this method, the surface tension of a solution is obtained by fitting the

120

Young-Laplace equation to the shape of a drop of solution hanging from a syringe. Syringes with

121

a tip of 0.30 mm of diameter were used for the solutions with low surface tension (< 50 mN m-1)

122

and 0.51 mm of diameter for solutions with larger surface tension (> 50 mN m-1). This produced

123

droplets with diameters between 1.4 and 2.4 mm. The tensiometer was calibrated with MilliQ

124

water and the measurements were carried out at 297 ± 2 K. Before each measurement the droplet

125

was left to equilibrate for 30 to 60 s.26 Each measurement was repeated 5 times and the

126

reproducibility between the results was ± (1 – 3) %. The instrument also allowed the volume of

127

the droplet to be monitored in real time, and ensured that the latter did not significantly evaporate

128

during the measurements. The overall uncertainties on each surface tension measurement were ±

129

(0.3 – 1.0) mN m-1.

130

These surface tension measurements were used not only to determine the surface tension of the

131

sampled aerosols, but also to determine the surface tension curves (curves σ as function of

132

surfactant concentration) and Critical Micelle Concentrations (CMC, concentration at which the

ACS Paragon Plus Environment

6

Page 7 of 34

Environmental Science & Technology

133

particle surface is saturated and excess surfactant molecules start to arrange themselves in

134

micelles in the bulk) for the first time for atmospheric surfactants. The CMC was determined

135

graphically on the surface tension curves, as the intersection between the sharp slope and the

136

minimum surface tension level (Figure 1). The surface tension curves, including the minimum

137

surface tensions and CMCs, are characteristic of specific surfactant molecules. Before

138

determining such curves for aerosol surfactants, they were measured for the 9 reference

139

surfactants listed above: SDS, AOT, Zephiramine, CTAC, Triton X114, Brij35, Surfactin,

140

Rhamnolipid, and L-α-Phosphatidylcholine. The surface tension curves, minimum surface

141

tensions, and CMCs obtained were in excellent agreement with the literature.29-30

142

Surface tension curves were then determined for the aerosol samples. First, the surface tension

143

of the 60-µL extracts obtained from each aerosol sample was measured. The extracts were then

144

successively diluted with MilliQ water, and the surface tension was measured at each dilution

145

step, until the surface tension value for pure water was reached (Figure 1). For each sample, the

146

absolute position of the curve on the X-axis was given by the surfactant concentration obtained

147

by the colorimetric techniques described below, and the CMC was determined graphically, by

148

the same method than described above for the reference surfactants.

149

In spite of the extraction, in a few atmospheric samples the surfactant concentration was low,

150

i.e. close to or slightly larger than the CMC, and their surface tension was slightly larger than

151

their minimum surface tension. Overestimating the minimum surface tension in surface tension

152

curves, in turn, slightly underestimated the CMC value. The uncertainties on these CMCs were

153

thus determined as the combination (square root of the sum of the squares) of the uncertainties

154

on the concentrations (20 %, see below) and those on the minimum surface tension values. The

155

latter were taken as the relative value of the first derivative of the surface tension curves (=

ACS Paragon Plus Environment

7

Environmental Science & Technology

Page 8 of 34

156

δσ/δC as function of C) at the largest concentration measured, which should be zero if the actual

157

minimum surface tension is reached. In nearly all the atmospheric samples studied in this work,

158

these uncertainties were less than 3.5 %, corresponding to about 1 mN m-1, and indicating that

159

the surface tension measured for the sample was equal to the minimum surface tension for the

160

surfactant, within uncertainties. Only for two samples, b and g (see list in Table 1), these

161

uncertainties were larger, 19 and 6 %, respectively. But in all the samples, the overall

162

uncertainties on the CMC values were mostly due to those on the concentrations.

163

Concentration measurements

164

Colorimetric techniques were chosen in this work to determine surfactant concentrations in

165

aerosols, because they provide absolute concentrations and have been shown to be sensitive

166

enough for aerosol surfactants.18-21 However, because there is no dye reacting with all types of

167

surfactants, it was necessary to measure anionic, cationic and non-ionic surfactants separately to

168

obtain the total surfactant concentration. To our knowledge, this is the first time that

169

concentrations for non-ionic surfactants in atmospheric aerosols are reported, and the results of

170

this work show that they represent a major fraction of aerosol surfactants. The principle of these

171

colorimetric methods is to titrate the surfactants with an ionic dye specific to the surfactant class

172

(anionic, cationic or non-ionic). The resulting surfactant-dye complex is then extracted in an

173

organic phase and its concentration determined by UV-Vis absorption spectroscopy, using

174

calibration curves established with reference compounds. The advantage of this technique is to

175

provide a unique calibration curve for each class of surfactants, thus making possible to

176

determine the concentration of unknown surfactants belonging to each class.

177

Anionic surfactant concentration. The dye used to quantify anionic surfactants was ethyl violet

178

(C31H42N3).18, 31-33 The reactions were carried out by adding to 3 mL-aqueous samples 200 µL of

ACS Paragon Plus Environment

8

Page 9 of 34

Environmental Science & Technology

179

acetate buffer (pH = 5), 100 µL of EDTA solution 0.1 M, 500 µL of 1 M sodium sulfate solution,

180

and 200 µL of ethyl violet solution (0.49 g L-1).31-33 2.5 mL of toluene were then added and the

181

solutions were stirred for one hour. Once the aqueous and organic phases were separated, the

182

toluene phase was removed and analyzed by UV-vis spectroscopy (see below).

183

Cationic surfactant concentration. The dye chosen to quantify cationic surfactants was

184

disulfine blue (C27H32N2O6S2).18,

185

aqueous samples 1 mL of acetate buffer pH = 5 and 500 µL of disulfine blue solution (2.58 g L-1

186

in a mixture of 90:10 water ethanol solution), and 2.5 mL of chloroform. The mixture was stirred

187

for one hour, after which the chloroform phase was removed and analyzed.

188

Non-ionic surfactant concentrations. The quantification of non-ionic surfactants was more

189

challenging than for anionic and cationic ones because there is no known dye able to react with

190

all types of non-ionic surfactants. In this work, we chose cobalt thiocyanate (Co(NCS)2) as

191

reagent because it reacts with compounds containing a wide range of organic groups, in

192

particular ethoxylated-polyoxyethylene groups (or “EO-PO”), -(CH2)n-O)-, that are common in

193

surfactants.35-36 The reactions were performed by adding to the 3 mL-aqueous samples 1 mL of

194

cobalt thiocyanate solution (6.2 g of ammonium thiocyanate, 2.8 g cobalt nitrate hexahydrate in

195

10 mL water) and 2 mL of chloroform. After stirring for one hour the chloroform phase was

196

removed and analyzed.

197

Quantification, uncertainties, and detection limits. The concentrations of the surfactant-dye

198

complexes were quantified by UV-Vis absorption spectroscopy, by placing small amounts of

199

solutions in a 1-cm quartz cell and measuring the absorption over 190 – 1100 nm with an Agilent

200

8453 UV-Vis Spectrophotometer. Calibration curves for each surfactant class were established

201

by measuring the maximum absorbance of known solutions of reference surfactants. For anionic

34

The reactions were performed by adding to the 3 mL-

ACS Paragon Plus Environment

9

Environmental Science & Technology

Page 10 of 34

202

surfactants, the maximum absorbance was at 612 nm and the reference compounds were SDS

203

and AOT. They resulted in a single calibration curve with a slope of 0.37 ± 0.02 µM-1 (Figure 2),

204

the uncertainties being mostly due to the linear regression, and in a detection limit (intercept on

205

the Y-axis + uncertainties) of 0.054 µM, or 0.016 mg L-1 for SDS. For cationic surfactants, the

206

maximum absorbance was at 628 nm and the reference compounds were Zephiramine and

207

CTAC, which resulted in a single calibration curve with slope of 0.35 ± 0.05 µM-1, and in a

208

detection limit of 0.059 µM, thus 0.011 mg L-1 for Zephiramine. Thus, for anionic and cationic

209

surfactants, these colorimetric methods are more sensitive than relative techniques for surfactant

210

concentration measurement.22-23 For non-ionic surfactants, two peaks of maximum absorbance

211

could be used, at 317 and 621 nm, and five reference compounds were used, Triton X114,

212

Brij35, Surfactin, Rhamnolipid, and L-α-Phosphatidylcholine. They resulted in calibration curves

213

containing each ± 10 % of uncertainties, but displaying different slopes, depending on the

214

structure of the surfactant molecule (number of EO-PO units) and spanning over nearly a factor

215

of 10 between Triton X114 (7-8 EO-PO units) and Brij35 (23 EO-PO units). Previous works

216

reported a similar range of variability between seven industrial surfactants measured with cobalt

217

thiocyanate.35 Because the objective of this work was to determine the importance of surfactants

218

for cloud formation, we chose to determine lower limits for their concentrations by using the

219

calibration curve with the largest slope (i.e. giving the smallest concentrations), 0.013 ± 0.001

220

µM-1, which was the one for Brij35. With this curve, the detection limit for non-ionic surfactants

221

was estimated to 0.3 µM. As different non-ionic surfactants with different structures are likely to

222

be present in aerosols, using the calibration curve for Brij35 potentially underestimated the

223

overall non-ionic concentrations in aerosols. The extent of this potential underestimation

224

(systematic errors on the measurements) was determined from the calibrations slopes of non-

ACS Paragon Plus Environment

10

Page 11 of 34

Environmental Science & Technology

225

ionic reference compounds with surface tension curves similar to those found for aerosol

226

surfactants in this work (CMC ~ 10-4 M, σmin ≤ 40 mNm-1, see Figure 5). Those were Surfactin,

227

Rhamnolipid, and Triton X114. Their calibration slopes spanned over a factor 3.5 in total, thus

228

implying an average underestimation of the concentrations of a mixture of such compounds by a

229

factor 1.75. But because the dye could also miss some non-ionic surfactants entirely (those not

230

containing EO-PO units, for instance), the potential underestimation on the non-ionic

231

concentrations in aerosols was estimated to be a factor 2 on average. The resulting systematic

232

errors in the total surfactant concentrations were, however, lower (see below).

233

Determination of total surfactant concentrations. Before determining the total surfactant

234

concentration in aerosols, it was confirmed, using reference compounds, that the cationic method

235

did not detect any anionic surfactants and vice-versa, and that the non-ionic method detected

236

neither anionic nor cationic surfactants. The cationic method was also confirmed not to detect

237

any non-ionic surfactants. Only the anionic method was found to weakly detect some biological

238

non-ionic surfactants (10 - 30 % of the calibration slope). But as these compounds were also

239

weakly detected by the non-ionic method (10 - 50 % of the calibration slope for Brij35),

240

summing up their concentrations obtained with the anionic and the non-ionic methods still

241

accounted for less than 100 % of their concentration. Therefore, the total surfactant concentration

242

in aerosols was determined as the sum of the concentrations of anionic, cationic and non-ionic

243

surfactants obtained with the different dyes. The uncertainties on these total concentrations were

244

estimated to be ± 20 % as the square roots of the sums of those on anionic (5 %), cationic (15 %)

245

and non-ionic surfactant concentrations (10 %) and on the extraction efficiencies. In addition to

246

these random errors, the potential underestimation of the non-ionic surfactant concentrations by a

247

factor 2 was estimated to result in potential underestimation of 33 % on the total surfactant

ACS Paragon Plus Environment

11

Environmental Science & Technology

Page 12 of 34

248

concentrations, as non-ionic surfactants contributed to about 1/3 of the total surfactant

249

concentration in the samples (see the Results and Discussion section).

250

Interferences from other ionic species. The possibility of interferences on the measured

251

concentrations due to the reaction of the ionic dyes with other ionic species than surfactants

252

present in the atmospheric samples was also studied. For this, sodium chloride (NaCl),

253

ammonium sulfate ((NH4)2SO4), and oxalic acid, representing some the most abundant ionic

254

species in atmospheric aerosols, were added in concentrations 1 mM to 1 M to known solutions

255

of reference surfactants and the surfactant concentrations were measured using the colorimetric

256

methods.

257

All these compounds were found to interfere, positively or negatively (i.e. leading to over- or

258

underestimations) with all classes of surfactants (Figure 3). In the presence of interferents, the

259

differences between the measured concentrations and the calibration curves became larger than

260

the uncertainties for interferent concentrations of the order of 0.01 - 0.05 M. This showed that

261

using such colorimetric methods directly on water extracts of atmospheric samples might lead to

262

erroneous surfactant concentrations. In our work, however, the surfactant concentrations

263

obtained after the double extraction were unaffected by the presence of these interferents, even in

264

concentrations as high as 1 M. This was because they were not retained (thus eliminated) by the

265

second (SPE) extraction step. This showed that this second step is essential for the accurate

266

measurement of surfactant concentrations.

267

Once the surfactant concentrations (for each class of surfactant and in total) were determined for

268

each sample volume (60 µL), they were determined for the corresponding aerosol volume by

269

dividing the results by the aerosol total aerosol volume determined by weighting the filters (see

270

the “sampling” section).

ACS Paragon Plus Environment

12

Page 13 of 34

271

Environmental Science & Technology

Seawater chlorophyll-a concentrations

272

Seawater concentrations of chlorophyll-a and other biological seawater markers near Askö

273

station and at other locations in the Baltic sea were available in the SHARK database (Svenskt

274

HavsARKiv, Swedish Meteorological and Hydrological Institute, SMHI). But because these data

275

were missing for too many days over the sampling period, concentrations of Chlorophyll-a, from

276

the aqua MODIS satellite instrument (Level L3, 1 Day Composite), provided by the NOAA

277

(National Oceanic and Atmospheric Administration) ERDDAP data server, were preferred.

278

Whenever they were available, they agreed within 10 % with those given by the SHARK

279

database. The daily chlorophyll-a MODIS concentrations were averaged over the time periods

280

corresponding to each aerosol sample, and over areas between 50 and 300 km2 around Askö. 10

281

% of uncertainties were attributed to the chlorophyll concentrations thus obtained, to account for

282

the fact that these data were available only for about 70% of the days over the sampling period.

283

Chemicals

284

All the chemicals used were purchased directly from the manufacturers and used without

285

further purification. Sodium dodecyl sulfate, ≥ 98.5 % Bioreagent, Sigma; Dioctyl sulfosuccinate

286

sodium salt, 98 %, Aldrich; Benzyltetradecyldimethylammonium, ≥ 99.0 % anhydrous, Fluka;

287

Cetyltrimethylammonium chloride solution, 25 wt. % in H2O, Aldrich; Triton X114, laboratory

288

grade, Sigma-Aldrich; Brij®35, Fluka Bio Chemika; L-α-Phosphatidylcholine from egg yolk,

289

Type XVI-E, lyophilized powder, ≥ 99 %, Sigma; Surfactin from Bacillus subtilis, ≥ 98 %,

290

Sigma; R-95Dd Rhamnolipid (95 % dirhamnolipid, 5 % monorhamnolipid) Aldrich; Ethyl

291

Violet, cationic triarylmethane dye, Sigma-Aldrich; Patent Blue VF, dye content 50 %, Sigma-

292

Aldrich; Ammonium thiocyanate, ≥ 99 % puriss. p.a., ACS reagent; Sigma-Aldrich; Cobalt(II)

293

nitrate hexahydrate, ≥ 98 % ACS reagent, Sigma-Aldrich; Acetic anhydride, ≥ 99 %

ACS Paragon Plus Environment

13

Environmental Science & Technology

Page 14 of 34

294

ReagentPlus®, Sigma-Aldrich; Sodium acetate, ≥ 99.0 % anhydrous Reagent Plus, Sigma-

295

Aldrich; Ethylenediaminetetraacetic acid, 99.4 - 100.6 % ACS reagent Powder, Sigma-Aldrich;

296

Sodium sulfate anhydrous, ≥ 99.0 % granulated puriss. p.a. ACS reagent, Fluka; Oxalic acid, ≥

297

99 %, Aldrich; Ammonium sulfate, ≥ 99.0 % BioXtra; Sigma-Aldrich; Ammonium sulfate, ≥

298

99.5 % puriss. p.a. ACS reagent, Fluka Chemika Sigma-Aldrich; methanol, ≥ 99.9 %

299

CHROMASOLV®

300

CHROMANORM® Reag. Ph. Eur. super gradient grade pour HPLC, VWR BDH Prolabo;

301

Chloroform 99 % stab. with 0.8-1 % ethanol, Alfa Aesar; toluene, > 99 %, Chimie Plus.

302

RESULTS AND DISCUSSION

303

Atmospheric surfactant concentrations

for

HPLC,

Sigma-Aldrich;

acetonitrile,



99,9

%

HiPerSolv

304

The surfactant concentrations obtained for the 11 atmospheric samples are shown in Figure 4.

305

They varied between 27 ± 6 and 143 ± 29 mM in the aerosol volume, and 104 ± 21 and 785 ±

306

157 pmol m-3 in the air, and, as discussed in the Experimental section, were potentially

307

underestimated by 33 %. The concentrations in volume of air reported in this work are consistent

308

with those reported for PM2.5 aerosols from the Middle Adriatic using a relative method (224 -

309

496 pmol m-3),23 and with the anionic and cationic surfactant concentrations measured in

310

aerosols from rural and semi-urban locations with similar colorimetric methods (1 – 1000 pmol

311

m-3).18-21 As shown in Figure 4, in the Askö aerosols the total surfactant fraction was dominated

312

by anionic and non-ionic compounds, while cationic surfactants were in very small concentration

313

and below the detection limit (0.06 µM) in a number of samples. However, as explained in the

314

Experimental section, the concentrations for non-ionic surfactants could be underestimated by up

315

to a factor 2, and these compounds could thus have been even more abundant in the samples than

316

suggested by Figure 4.

ACS Paragon Plus Environment

14

Page 15 of 34

Environmental Science & Technology

317

Figure 4 also shows that the relative abundance of anionic, cationic and non-ionic surfactants

318

in the aerosols remained relatively constant from July to October, 60, 8, and 32 %, on average,

319

respectively, with only 10-15 % of variability. This suggested that the sources for these

320

surfactants remained the same throughout the sampling period. However, no significant

321

correlations were found between the different classes of surfactants, suggesting that anionic,

322

cationic and non-ionic surfactants had distinct sources. The main variability observed between

323

the samples from July to October was in the total surfactant concentrations, displaying 60 to 70

324

% of standard deviation. This suggested that the surfactant sources varied mostly in intensity,

325

rather than in composition, over the sampling period.

326

Absolute surface tension curves and CMC values

327

The surface tension measurements performed in this work showed the presence of strong

328

surfactants in all the aerosols sampled, with surface tension values of σmin = 32 - 40 mN m-1

329

(Figure 5). These values are consistent with the surface tension reported for aerosols from other

330

regions, using a similar method.24-25 The surface tension values for aerosol and fog water

331

samples reported previously by other groups were, however, significantly larger (usually ≥ 50

332

mN m-1)3-7 and did not display any lower plateau, indicating that the minimum surface tension

333

was not reached. This indicated that the surfactants were not concentrated enough in the samples

334

and underlines the importance of using a targeted extraction for this type of investigation.

335

For each sample, absolute surface tension curves (Figure 5) and CMC values (Table 1) were

336

also determined, which, to our knowledge, is the first time for aerosol surfactants. In previous

337

works, surface tension was only related to the total organic fraction in aerosols or fog water as

338

the compounds responsible for the surface tension effect were not isolated.3-5 As discussed in the

339

Experimental section, the surface tension measured for the samples were equal to the minimum

ACS Paragon Plus Environment

15

Environmental Science & Technology

Page 16 of 34

340

surface tension of the corresponding surfactant, within uncertainties. These uncertainties were

341

carried onto the uncertainties on the CMC values and are reported in Table 1.

342

The minimum surface tensions reported for the sampled aerosols in this work are comparable

343

to those of strong organic surfactants, such as the reference compounds used in the calibrations

344

(SDS, AOT, Zephiramine, CTAC, Triton X114, Brij35, Surfactin, Rhamnolipid, and L-α-

345

Phosphatidylcholine). The CMCs, which are much more characteristic of specific surfactants,

346

were compared with those of known artificial and biological surfactants (Table 1). This

347

comparison shows that aerosol surfactants are generally in the range for biological surfactants,

348

suggesting their biological origin.

349

Implications for particle activation

350

The quantitative results obtained for aerosol surfactants in this work bring some unique

351

information for the prediction of cloud droplet activation, which can not be obtained by other

352

techniques. During activation an aerosol particle typically undergoes a radius increase between 3

353

and 10 (ratios of critical radius over dry radius), which is true both for inorganic salt particles37

354

and for mixed organic/inorganic particles such as Secondary Organic Aerosols.37 This

355

corresponds to a volume increase by a factor 27 to 1000. The surface tension curves determined

356

in this work (Figure 5) show that lowering the surfactant concentrations by these factors still

357

leads to low surface tension values, typically below 50 mN/m. These results thus show that

358

surfactant concentrations in atmospheric aerosols are large enough to maintain the surface

359

tension of growing droplets very low (≤ 50 mN m-1) until activation. This conclusion is

360

reinforced by the facts that the concentrations determined in this work are likely to be

361

underestimated, and that the surfactants might not have been present in all the aerosol particles

362

collected, but only in some of them, where they would have been at even larger concentrations.

ACS Paragon Plus Environment

16

Page 17 of 34

Environmental Science & Technology

363

Lowering the surface tension of growing droplets to 50 mNm-1 or less until activation is, in turn,

364

expected to have substantial effects on the activation efficiency of aerosol particles. For instance,

365

using surface tension values of about 50 mN m-1 was shown previously to predict 30 - 50 %

366

larger CCN numbers in various regions of the atmosphere.17

367

Another implication of these results is that the surface tension of forming droplets would vary

368

little or not at all with surfactant concentration during activation. This is because, as shown by

369

the ratios Caerosol/CMC in Table 1, a particle volume increase by a factor 30 to 1000 at activation,

370

corresponds to a surfactant concentration of the order of, or slightly larger than the CMC. In

371

nearly all the samples studied, the particles can undergo a volume increase by at least a factor

372

200 until activation, i.e. a radius increase by at least a factor 6, without any significant change on

373

their surface tension. Such a nearly-constant and minimal surface tension in growing droplets is

374

in contradiction with the Szyszkowski equation,8,

375

cloud droplet activation and assumes a decrease of σ with the log of surfactant concentration.

376

Using this equation is thus likely to significantly underestimate the effects of aerosol surfactants

377

on cloud droplet formation.

378

Correlations with seawater chlorophyll concentrations

10-11

which is used in nearly all models for

379

To further examine the origin of the aerosol surfactants studied in this work, their

380

concentrations were compared with those of a tracer for biological activity in seawater: seawater

381

chlorophyll-a concentrations. These concentrations were obtained from the MODIS instrument,

382

as explained in the Experimental section. When averaged over an area of 50 km × 50 km around

383

the Askö station, these concentrations displayed some correlations with anionic (r2 = 0.65),

384

cationic (r2 = 0.75), and total (r2 = 0.67) surfactant concentrations (Figure 6), the correlation with

385

total surfactants not resulting directly from those with anionic and cationic surfactants. These

ACS Paragon Plus Environment

17

Environmental Science & Technology

Page 18 of 34

386

correlations suggested a biological and marine origin for these compounds. However,

387

chlorophyll-a concentrations averaged over larger areas (100 km x 100 km) did not correlate

388

with the surfactant concentrations (r2 < 0.35), indicating that, if the surfactants were indeed

389

produced by biological sources in seawater, these sources were local.

390

A marine and biological origin for the surfactants might seem contradictory with the lack of

391

correlation between non-ionic surfactants and chlorophyll concentrations, as most biological

392

surfactants are non-ionic. This lack of correlation could be attributed either to large uncertainties

393

(mostly underestimations) in the non-ionic concentrations or to the choice of the wrong marker

394

for their sources. Chlorophyll-a was chosen in this work mostly because the available data had a

395

frequency similar to those of our samples, but has been reported not to be the best surrogate for

396

the organic matter transferred from the sea surface to atmospheric aerosols,38 or for the

397

biological processes controlling the sea-surface organic composition.39-40 In future studies,

398

correlations between aerosol surfactants and other seawater markers will thus be sought.

399

In conclusion, directly extracting and analyzing surfactants from atmospheric aerosols in this

400

work has brought unique new information for the understanding of their role in cloud formation,

401

which can not be obtained by classical on-line techniques. The results show that surfactants are

402

concentrated enough in atmospheric aerosols to keep the surface tension of growing droplets

403

very low until activation, which should enhance the cloud-forming efficiency. They also show

404

that the surface tension of growing droplets remains nearly constant and close to its minimum

405

during activation, and thus does not follow the Szyszkowski equation. Models using this

406

equation are thus expected to significantly underestimate the role of surfactants on cloud droplet

407

formation. All these conclusions are reinforced by the facts that the concentrations reported in

408

this work might be underestimated by about 30 %, and that the surfactants might have been

ACS Paragon Plus Environment

18

Page 19 of 34

Environmental Science & Technology

409

present in only a fraction of the collected particles, in which their concentration would have been

410

much larger than reported here. In particular, while the agreements between observed and

411

predicted CCN numbers in the atmosphere (or “closure”) are currently achieved by assuming the

412

surface tension of pure water for all aerosol particles (or CN),15,11, the present results suggest that

413

such closure could also be achieved with large concentrations of surfactants in only a fraction of

414

the CN, corresponding to the observed CCN/CN ratios. To determine if this is the case would

415

now require to investigate the particle size range in which surfactants are present. More

416

generally, further work is needed to investigate the presence of surfactants in smaller aerosol size

417

fractions (PM1), which are more critical for CCN numbers, and in many other regions of the

418

atmosphere. Further work would also be needed to identify the sources and chemical structures

419

of these compounds, in order to propose some proxies for atmospheric models.

420

AUTHOR INFORMATION

421

Corresponding Author

422

* CNRS and Université Lyon 1, Institut de Recherche sur la Catalyse et l’Environnement de

423

Lyon, Villeurbanne, France. Phone: +33 4 27 46 57 32. E-mail: [email protected]

424

lyon1.fr.

425

Present Address

426

† National Research Center for Environmental Toxicology, The University of Queensland,

427

Brisbane, Australia.

428

Notes

429

The authors declare no competing financial interest.

ACS Paragon Plus Environment

19

Environmental Science & Technology

430

Page 20 of 34

ACKNOWLEDGMENT

431

The authors warmly thank Susan Eriksson and the staff of the Askö Laboratory, Stockholm

432

University, Sweden, for their assistance in this project, and Prof. Corinne Ferronato for the use of

433

the SPE extraction set-up in Lyon. Drs. Barbara Deutsch, Stockholm University, and Örjan Bäck

434

and Torny Axell, Swedish Meteorological and Hydrological Institute (SMHI), are also greatly

435

thanked for the information on the seawater measurements and the Shark database (Svensk

436

Havsarkiv). This project was funded by the Swedish Research Council FORMAS (project 2009-

437

224) for the aerosol sampling and the French Agence Nationale de la Recherche (ANR) and US

438

National Science Foundation (NSF) on a joint ANR-NSF project (SONATA) for the analysis of

439

the results.

440

REFERENCES

441

1.

442

Kerminen, Y. Kondo, H. Liao, U. Lohmann, P. Rasch, S.K. Satheesh, S. Sherwood, B. Stevens

443

and X.Y. Zhang, Clouds and Aerosols. In Climate Change 2013: The Physical Science Basis.

444

Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel

445

on Climate Change, Stocker, T. F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung,

446

A. Nauels, Y. Xia, V. Bex, P.M. Midgley, Ed. Cambridge University Press: Cambridge, UK,

447

2013.

448

2.

449

1936, 32 (0), 1152-1161.

450

3.

451

of fog. Environ. Sci. Technol. 1990, 24 (5), 722-727.

Boucher, O., D. Randall, P. Artaxo, C. Bretherton, G. Feingold, P. Forster, V.-M.

Kohler, H., The nucleus in and the growth of hygroscopic droplets. Trans. Faraday Soc.

Capel, P. D.; Gunde, R.; Zuercher, F.; Giger, W., Carbon speciation and surface tension

ACS Paragon Plus Environment

20

Page 21 of 34

Environmental Science & Technology

452

4.

Facchini, M. C.; Mircea, M.; Fuzzi, S.; Charlson, R. J., Cloud albedo enhancement by

453

surface-active organic solutes in growing droplets. Nature 1999, 401 (6750), 257-259.

454

5.

455

Rissler, J.; Swietlicki, E.; Frank, G.; Andreae, M. O.; Maenhaut, W.; Rudich, Y.; Artaxo, P.,

456

Importance of the organic aerosol fraction for modeling aerosol hygroscopic growth and

457

activation: a case study in the Amazon Basin. Atmos. Chem. Phys. 2005, 5 (11), 3111-3126.

458

6.

459

like substances in connection with relaxation, dilution, and solution pH. J. Geophys. Res. 2006,

460

111 (D23), D23205.

461

7.

462

atmospheric and model humic-like substances (HULIS). Geophys. Res. Lett. 2007, 34 (16),

463

L16807.

464

8.

465

The role of surfactants in Köhler theory reconsidered. Atmos. Chem. Phys. 2004, 4 (8), 2107-

466

2117.

467

9.

468

Drewnick, F.; Hings, S.; Jung, D.; Borrmann, S.; Andreae, M. O., Size Matters More Than

469

Chemistry for Cloud-Nucleating Ability of Aerosol Particles. Science 2006, 312 (5778), 1375-

470

1378.

471

10.

472

Kulmala, M.; Laaksonen, A.; Lehtinen, K. E. J.; McFiggans, G.; Kokkola, H., Surfactant effects

473

in global simulations of cloud droplet activation. Geophys. Res. Lett. 2012, 39 (5).

Mircea, M.; Facchini, M. C.; Decesari, S.; Cavalli, F.; Emblico, L.; Fuzzi, S.; Vestin, A.;

Salma, I.; Ocskay, R.; Varga, I.; Maenhaut, W., Surface tension of atmospheric humic-

Taraniuk, I.; Graber, E. R.; Kostinski, A.; Rudich, Y., Surfactant properties of

Sorjamaa, R.; Svenningsson, B.; Raatikainen, T.; Henning, S.; Bilde, M.; Laaksonen, A.,

Dusek, U.; Frank, G. P.; Hildebrandt, L.; Curtius, J.; Schneider, J.; Walter, S.; Chand, D.;

Prisle, N. L.; Asmi, A.; Topping, D.; Partanen, A. I.; Romakkaniemi, S.; Dal Maso, M.;

ACS Paragon Plus Environment

21

Environmental Science & Technology

Page 22 of 34

474

11.

Farmer, D. K.; Cappa, C. D.; Kreidenweis, S. M., Atmospheric Processes and Their

475

Controlling Influence on Cloud Condensation Nuclei Activity. Chem. Rev. 2015, 115 (10), 4199-

476

4217.

477

12.

478

of molar volume and surfactant characteristics of water-soluble organic compounds in biomass

479

burning aerosol. Atmos. Chem. Phys. 2008, 8 (4), 799-812.

480

13.

481

Strong evidence of surface tension reduction in microscopic aqueous droplets. Geophys. Res.

482

Lett. 2012, 39 (23), L23801.

483

14.

484

Growth of Submicrometer Particles at High Relative Humidity. J. Phys. Chem. A 2014, 118 (22),

485

3952-3966.

486

15.

487

Changes in Droplet Surface Tension Affect the Observed Hygroscopicity of Photochemically

488

Aged Biomass Burning Aerosol. Environ. Sci. Technol. 2013, 47 (19), 10980-10986.

489

16.

490

Coe, H.; McFiggans, G., Consistency between parameterisations of aerosol hygroscopicity and

491

CCN activity during the RHaMBLe discovery cruise. Atmos. Chem. Phys. 2010, 10 (7), 3189-

492

3203.

493

17.

494

measurements of hygroscopic growth and critical supersaturation of aerosol particles in central

495

Germany. Atmos. Chem. Phys. 2010, 10 (23), 11737-11752.

Asa-Awuku, A.; Sullivan, A. P.; Hennigan, C. J.; Weber, R. J.; Nenes, A., Investigation

Ruehl, C. R.; Chuang, P. Y.; Nenes, A.; Cappa, C. D.; Kolesar, K. R.; Goldstein, A. H.,

Ruehl, C. R.; Wilson, K. R., Surface Organic Monolayers Control the Hygroscopic

Giordano, M. R.; Short, D. Z.; Hosseini, S.; Lichtenberg, W.; Asa-Awuku, A. A.,

Good, N.; Topping, D. O.; Allan, J. D.; Flynn, M.; Fuentes, E.; Irwin, M.; Williams, P. I.;

Irwin, M.; Good, N.; Crosier, J.; Choularton, T. W.; McFiggans, G., Reconciliation of

ACS Paragon Plus Environment

22

Page 23 of 34

Environmental Science & Technology

496

18.

Latif, M. T.; Brimblecombe, P., Surfactants in Atmospheric Aerosols. Environ. Sci.

497

Technol. 2004, 38 (24), 6501-6506.

498

19.

499

Mohamed, C. A. R.; Latif, M. T., Surfactants in the sea-surface microlayer and their contribution

500

to atmospheric aerosols around coastal areas of the Malaysian peninsula. Mar. Pollut. Bull. 2010,

501

60 (9), 1584-1590.

502

20.

503

Khan, M. F.; Tahir, N. M., Surfactants in the sea-surface microlayer and atmospheric aerosol

504

around the southern region of Peninsular Malaysia. Mar. Pollut. Bull. 2014, 84 (1–2), 35-43.

505

21.

506

marine aerosols at different locations along the Malacca Straits. Environ Sci Pollut Res 2014, 21

507

(10), 6590-6602.

508

22.

509

atmospheric surfactants in the bulk precipitation by electrochemical tools. J Atmos Chem 2010,

510

66 (1-2), 11-26.

511

23.

512

substances in atmospheric aerosol: an electrochemical approach. Tellus B 2012, 64, 12.

513

24.

514

P., A possible role of ground-based microorganisms on cloud formation in the atmosphere.

515

Biogeosci. 2010, 7 (1), 387-394.

516

25.

517

aerosols from Grenoble, France. Atmos. Environ. 2012, 47 (0), 413-420.

Roslan, R. N.; Hanif, N. M.; Othman, M. R.; Azmi, W. N. F. W.; Yan, X. X.; Ali, M. M.;

Jaafar, S. A.; Latif, M. T.; Chian, C. W.; Han, W. S.; Wahid, N. B. A.; Razak, I. S.;

Mustaffa, N.; Latif, M.; Ali, M.; Khan, M., Source apportionment of surfactants in

Orlović-Leko, P.; Kozarac, Z.; Ćosović, B.; Strmečki, S.; Plavšić, M., Characterization of

Frka, S.; Dautovic, J.; Kozarac, Z.; Cosovic, B.; Hoffer, A.; Kiss, G., Surface-active

Ekström, S.; Nozière, B.; Hultberg, M.; Alsberg, T.; Magnér, J.; Nilsson, E. D.; Artaxo,

Baduel, C.; Nozière, B.; Jaffrezo, J.-L., Summer/winter variability of the surfactants in

ACS Paragon Plus Environment

23

Environmental Science & Technology

Page 24 of 34

518

26.

Nozière, B.; Baduel, C.; Jaffrezo, J.-L., The dynamic surface tension of atmospheric

519

aerosol surfactants reveals new aspects of cloud activation. Nature Comms. 2014, 5, 4335.

520

27.

521

Environment: Problems and Challenges. Chem. Rev. 2011, 111 (9), 5667-5700.

522

28.

523

of surfactants in environmental samples. Talanta 2012, 88 (0), 1-13.

524

29.

525

potential. Microbiol. Mol. Biol. R. 1997, 61 (1), 47-64.

526

30.

527

remediation. J. Appl. Microbiol. 2002, 93 (6), 915-929.

528

31.

529

spectrophotometric determination of anionic surfactants with ethyl violet. Anal. Chem. 1982, 54

530

(3), 392-397.

531

32.

532

anionic surfactants in sea water. Analyst 1987, 112 (10), 1405-1408.

533

33.

534

Decker, Inc.: 2001; Vol. 96, p 637.

535

34.

536

Physical Methods. Wiley: 2000; p 412.

537

35.

538

Thiocyanate Ions in Aqueous Solutions of Nonionogenic Surfactants. Russ. J. Coord. Chem.

539

2003, 29 (8), 554-558.

Olkowska, E.; Polkowska, Ż.; Namieśnik, J., Analytics of Surfactants in the

Olkowska, E.; Polkowska, Ż.; Namieśnik, J., Analytical procedures for the determination

Desai, J. D.; Banat, I. M., Microbial production of surfactants and their commercial

Christofi, N.; Ivshina, I. B., Microbial surfactants and their use in field studies of soil

Motomizu,

S.;

Fujiwara,

S.;

Fujiwara,

A.;

Toei,

K.,

Solvent

extraction-

Yamamoto, K.; Motomizu, S., Solvent extraction-spectrophotometric determination of

Schmitt, T. M., Analysis of Surfactants, Second Edition (Surfactant Science). Marcel

Hummel, D. O., Handbook of Surfactant Analysis: Chemical, Physico-chemical and

Amirov, R. R.; Skvortsova, E. A.; Saprykova, Z. A., Complexation of Cobalt(II) with

ACS Paragon Plus Environment

24

Page 25 of 34

Environmental Science & Technology

540

36.

Pacheco e Silva, V., Method to measure surfactant in fluid. US Patent 2013/0337568-A1

541

2013.

542

37.

543

Fuzzi, S.; Gysel, M.; Laaksonen, A.; Lohmann, U.; Mentel, T. F.; Murphy, D. M.; O'Dowd, C.

544

D.; Snider, J. R.; Weingartner, E., The effect of physical and chemical aerosol properties on

545

warm cloud droplet activation. Atmos. Chem. Phys. 2006, 6 (9), 2593-2649.

546

38.

547

Hardenberg, J.; Ceburnis, D.; Vaishya, A.; O'Dowd, C. D.; Facchini, M. C., Is chlorophyll-a the

548

best surrogate for organic matter enrichment in submicron primary marine aerosol? J. Geophys.

549

Res.-Atmos. 2013, 118 (10), 4964-4973.

550

39.

551

Moore, K. A.; Beall, C. M.; McCluskey, C. S.; Cornwell, G. C.; Zhou, Y.; Cox, J. L.;

552

Pendergraft, M. A.; Santander, M. V.; Bertram, T. H.; Cappa, C. D.; Azam, F.; DeMott, P. J.;

553

Grassian, V. H.; Prather, K. A., Microbial Control of Sea Spray Aerosol Composition: A Tale of

554

Two Blooms. ACS Central Science 2015, 1 (3), 124-131.

555

40.

556

114.

McFiggans, G.; Artaxo, P.; Baltensperger, U.; Coe, H.; Facchini, M. C.; Feingold, G.;

Rinaldi, M.; Fuzzi, S.; Decesari, S.; Marullo, S.; Santoleri, R.; Provenzale, A.; von

Wang, X.; Sultana, C. M.; Trueblood, J.; Hill, T. C. J.; Malfatti, F.; Lee, C.; Laskina, O.;

Vaida, V., Ocean Sea Spray, Clouds, and Climate. ACS Central Science 2015, 1 (3), 112-

557 558

ACS Paragon Plus Environment

25

Environmental Science & Technology

Page 26 of 34

Table of contents Graphics and Tables

559 560 561 562

Graphical abstract

563 564

Tables:

565

-

Table 1

566 567

Figures :

568

-

Figure 1

569

-

Figure 2

570

-

Figure 3

571

-

Figure 4

572

-

Figure 5

573

-

Figure 6

574 575

ACS Paragon Plus Environment

26

Page 27 of 34

576

Environmental Science & Technology

Graphical abstract

577 578

ACS Paragon Plus Environment

27

Environmental Science & Technology

Page 28 of 34

579

Table 1: List of the aerosol samples, CMC values determined in this work, and ratios to the

580

aerosol surfactant concentration (Caerosol, in µM). Comparison with CMC ranges for known

581

microbial and artificial surfactants.

Aerosol samples

Bacterial surfactants

Artificial surfactants

Sample label

Sample date

CMC (µM)

Caerosol/CMC

a

04-10/07/10

96 ± 19

417

b

10-14/07/10

245 ± 69

139

c

14-29/07/10

129 ± 26

279

d

29/07-10/08/10

49 ± 10

580

e

10-19/08/10

118 ± 24

338

f

19/08-03/09/10

139 ± 28

268

g

14-23/09/10

95 ± 20

694

h

23/09-02/10/10

210 ± 42

268

i

02-08/10/10

212 ± 42

207

j

08-11/10/10

232 ± 46

315

k

11-13/10/10 Trehalose dicorynomycolate, Surfactin, Sophorolipids, Viscosin, Rhamnolipids Triton X114, Tween 20, CTAC, zephiramine, AOT, SDS

134 ± 27

1117

3 – 20029

200 – 10000this work,30

582

ACS Paragon Plus Environment

28

Page 29 of 34

Environmental Science & Technology

583 584

Figure 1: Determination of the surface tension curves and CMC values for sample h (see sample

585

list in Table 1). The red dot is the measurement on the initial extract, the black ones at lower

586

concentrations are those obtained from successive dilutions, and the black dot at the largest

587

concentration, corresponding to the concentration in the aerosol, is obtained from the volume

588

ratio between the extract and the aerosol. The blue dashed line represents the value for pure

589

water, and red dashed lines illustrate the graphical determination of the CMC.

590

ACS Paragon Plus Environment

29

Environmental Science & Technology

Page 30 of 34

591 592

Figure 2: UV-visible spectra of obtained from the reaction of ethyl violet with various

593

concentrations of SDS. Upper right corner: calibration curve for anionic surfactants obtained

594

from the absorbance at 612 nm on such spectra, from SDS (blue points) and AOT (orange

595

points). Grey points: blanks.

596

ACS Paragon Plus Environment

30

Page 31 of 34

Environmental Science & Technology

597 598

Figure 3: Effects of ionic interferents on measured concentrations for cationic surfactants

599

(zephiramine, 2 µM) relative to the calibration curve (horizontal line): NaCl (blue squares),

600

ammonium sulfate (green circles) and oxalic acid (red triangles). The color curves are the best

601

fits through the experimental points.

602

ACS Paragon Plus Environment

31

Environmental Science & Technology

Page 32 of 34

603 604

Figure 4: Concentration of anionic (blue and vertical lines), cationic (red), and non-ionic (green

605

and diagonal lines) surfactants in the aerosols sampled at Askö, Sweden from July to October

606

2010. (A) in the aerosol volume, (B) in sampled air. Concentrations not shown (in particular for

607

cationic surfactants) are under the detection limit.

608

ACS Paragon Plus Environment

32

Page 33 of 34

Environmental Science & Technology

609 610

Figure 5: Absolute surface tension curves for the surfactant fractions of the 11 aerosol samples

611

collected in Askö, Sweden. See sample list in Table 1.

612

ACS Paragon Plus Environment

33

Environmental Science & Technology

Page 34 of 34

613 614

Figure 6: Correlations between anionic (blue circles), cationic (red triangles), non-ionic (green

615

squares) and total surfactant (black diamonds) concentrations in atmospheric aerosols samples

616

and seawater chlorophyll-a concentrations over for the same dates, provided by aqua MODIS

617

satellite.

ACS Paragon Plus Environment

34