The organic fraction of bubble-generated, accumulation mode Sea Spray Aerosol (SSA)

Being as a relatively new approach of signalling, moving-block scheme significantly increases line capacity, especially on congested railways. This paper describes a simulation system for multi-train operation under moving-block signalling scheme. The simulator can be used to calculate minimum headways and safety characteristics under pre-set timetables or headways and different geographic and traction conditions. Advanced software techniques are adopted to support the flexibility within the simulator so that it is a general-purpose computer-aided design tool to evaluate the performance of Abstract Recent studies have detected a dominant accumulation mode (~100 nm) in the Sea Spray Aerosol (SSA) number distribution. There is evidence to suggest that particles in this mode are composed primarily of organics. To investigate this hypothesis we conducted experiments on NaCl, artificial SSA and natural SSA particles with a Volatility-Hygroscopicity-Tandem-Differential-Mobility-Analyser (VH-TDMA). NaCl particles were atomiser generated and a bubble generator was constructed to produce artificial and natural SSA particles. Natural seawater samples for use in the bubble generator were collected from biologically active, terrestrially-affected coastal water in Moreton Bay, Australia. Differences in the VH-TDMA-measured volatility curves of artificial and natural SSA particles were used to investigate and quantify the organic fraction of natural SSA particles. Hygroscopic Growth Factor ( HGF ) data, also obtained by the VH-TDMA, were used to confirm the conclusions drawn from the volatility data. Both datasets indicated that the organic fraction of our natural SSA particles evaporated in the VH-TDMA over the temperature range 170–200°C. The organic volume fraction for 71–77 nm natural SSA particles was 8±6%. Organic volume fraction did not vary significantly with varying water residence time (40 secs to 24 hrs) in the bubble generator or SSA particle diameter in the range 38–173 nm. At room temperature we measured shape- and Kelvin-corrected HGF at 90% RH of 2.46±0.02 for NaCl, 2.35±0.02 for artifical SSA and 2.26±0.02 for natural SSA particles. Overall, these results suggest that the natural accumulation mode SSA particles produced in these experiments contained only a minor organic fraction, which had little effect on hygroscopic growth. Our measurement of 8±6% is an order of magnitude below two previous measurements of the organic fraction in SSA particles of comparable sizes. We stress that our results were obtained using coastal seawater and they can’t necessarily be applied on a regional or global ocean scale. Nevertheless, considering the order of magnitude discrepancy between this and previous studies, further research with independent measurement techniques and a variety of different seawaters is required to better quantify how much organic material is present in accumulation mode SSA. isolated from other aerosol types. Both of these studies employed size-resolved impactor sampling and subsequent chemical analysis to show that SSA organic fraction increased with decreasing particle size in their experiments. Both studies measured an organic mass

produce artificial and natural SSA particles. Natural seawater samples for use in the bubble generator were collected from biologically active, terrestrially-affected coastal water in Moreton Bay, Australia. Differences in the VH-TDMA-measured volatility curves of artificial and natural SSA particles were used to investigate and quantify the organic fraction of natural SSA particles. Hygroscopic Growth Factor (HGF) data, also obtained by the VH-TDMA, were used to confirm the conclusions drawn from the volatility data. Both datasets indicated that the organic fraction of our natural SSA particles evaporated in the VH-TDMA over the temperature range 170-200°C. The organic volume fraction for 71-77 nm natural SSA particles was 8±6%. Organic volume fraction did not vary significantly with varying water residence time (40 secs to 24 hrs) in the bubble generator or SSA particle diameter in the range 38-173 nm.
At room temperature we measured shape-and Kelvin-corrected HGF at 90% RH of 2.46±0.02 for NaCl, 2.35±0.02 for artifical SSA and 2.26±0.02 for natural SSA particles. Overall, these results suggest that the natural accumulation mode SSA particles produced in these experiments contained only a minor organic fraction, which had little effect on hygroscopic growth. Our measurement of 8±6% is an order of magnitude below two previous measurements of the organic fraction in SSA particles of comparable sizes. We stress that our results were obtained using coastal seawater and they can't necessarily be applied on a regional or global ocean scale.
Nevertheless, considering the order of magnitude discrepancy between this and previous studies, further research with independent measurement techniques and a variety of different seawaters is required to better quantify how much organic material is present in accumulation mode SSA.

Introduction
Sea Spray Aerosol (SSA) is generated when air bubbles rise to the ocean surface and burst or when seawater droplets are torn from the crests of waves. These seemingly simple processes create the largest mass emission flux to the atmosphere of all aerosol types (Andreae and Rosenfeld, 2008). SSA particles vary in size over 5 decades from tens of nanometres to hundreds of micrometres. Large super-micrometre SSA particles account for the majority of sea spray mass in the atmosphere (>95%).
However it is the sub-micrometre SSA particles that are by far the most numerous. In particular recent laboratory and field measurements have consistently detected a dominant mode in the SSA number distribution centred at ~100 nm dry diameter (Clarke et al., 2006;Martensson et al., 2003;O'Dowd and Smith, 1993;Sellegri et al., 2006;Tyree et al., 2007). In this study we will refer to this dominant mode as the SSA accumulation mode even though it extends to particle sizes traditionally placed in the Aitken mode (20-100 nm). The SSA accumulation mode is climatically very important because it means SSA can potentially account for a significant proportion of cloud condensation nuclei (CCN) in the remote marine environment, particularly under high wind conditions (Clarke et al., 2006;O'Dowd et al., 1997;Pierce and Adams, 2006).
Bursting bubbles produce SSA in the form of film drops and jet drops. The SSA accumulation mode most likely originates from film drops. These are generated when fragments of bubble film (cap) are ejected into the air as a bubble bursts. Secondary droplets created when some of these fragments collide with the air-water surface may also contribute to the SSA accumulation mode (Spiel, 1998). Milliseconds after a bubble burst (Spiel, 1995) jet drops are generated from the break-up of the upward moving jet column caused by the collapse of the bubble cavity. Jet drops are in the super-micrometre size range: they are roughly one tenth the size of their parent bubbles (Blanchard, 1989).
The composition of SSA is surprisingly complex. Seawater contains a range of inorganic salts (see section 2.2) which all exist in SSA. In addition SSA contains an organic fraction which is derived from on or near the ocean surface. The existence of an organic fraction in SSA was detected many years ago (Blanchard, 1964). A number of studies found that the concentration of organic carbon (Gershey, 1983;Hoffman and Duce, 1976) and bacteria (e.g. see Blanchard, 1989 and references therein) in bulk SSA is enriched hundreds of time relative to corresponding concentrations in source water. Enrichment of organic matter in SSA occurs because it is generated from bubbles bursting in an enriched layer of chemical and biological material on seawater surfaces known as the sea-surface micro-layer (e.g. Liss and Duce, 1997). Organic material, and in particular surface-active organic material, becomes concentrated at the sea-surface micro-layer by factors of up to 10 compared to sub-surface waters (Hunter, 1997) due to processes such as diffusion, turbulent mixing and scavenging and transport by rising air bubbles.
Decades ago it was hypothesised that the organic fraction of SSA may increase with decreasing particle size (Barker and Zeitlin, 1972;Hoffman and Duce, 1974). Three recent studies have examined this hypothesis. Oppo et al. (1999) constructed a simple model that predicted the surfactant organic fraction of SSA droplets will increase hyperbolic-like with decreasing droplet size. The model rests on the assumption that SSA droplets produced from the rough sea surface contain condensed, saturated films of surfactant material of constant thickness (independent of droplet size).
Experimentally, detailed measurements of the size-resolved organic fraction of SSA produced by flowing natural seawaters through bubble generators were conducted by Facchini et al. (2008) and Keene et al. (2007). A bubble generator mimics the bubble bursting process on ocean surfaces to generate nascent SSA isolated from other aerosol types. Both of these studies employed size-resolved impactor sampling and subsequent chemical analysis to show that SSA organic fraction increased with decreasing particle size in their experiments. Both studies measured an organic mass fraction of ~80% for the lowest stages of their impactor samples corresponding to aerodynamic diameters of 130 nm (GMD, Keene et al., 2007) and 125-250 nm (Facchini et al., 2008). However Keene et al. only measured soluble organics, and   Facchini et al. measured both soluble and insoluble organics and found insoluble components dominate (~94% of organics). While both studies had different operative definitions of solubility, this consideration still implies that the results of these separate experiments do not agree as well as they first appear to. Nevertheless, based on these studies it is currently expected that the accumulation mode of SSA produced from biologically active seawater consists of particles that are predominantly organic. Bigg and Leck (2008) go even further to suggest that the particles comprising the SSA accumulation mode (< 200 nm) are actually organic fragments with no inorganic component at all.
It is important to characterise the composition of particles in the SSA accumulation mode to correctly model their climatic influence. The organic fraction of SSA particles will affect their size as a function of RH (Ming and Russell, 2001) and therefore their scattering potential (Randles et al., 2004), their ability to act as CCN (Moore et al., 2008) and also their role in atmospheric chemistry (Zhou et al., 2008).
The purpose of this study was to investigate and quantify the organic fraction of bubble-chamber-generated accumulation mode SSA using an original, independent and on-line method: the Volatility Hygroscopicity-Tandem Differential Mobility Analyser (VH-TDMA). In addition the VH-TDMA was also able to measure the hygroscopic growth factors of SSA accumulation mode particles.

Bubble generator
A bubble generator was constructed to mimic the production of SSA by bursting bubbles on seawater surfaces. The generator is depicted in Figure 1. It consisted of a 1 m long glass cylinder (id = 2.9 cm) with a fritted glass tip (SKC midget impinger; pore size 170-220 μm) inserted at the bottom. Sample water entered the bottom of the generator from a 20 L plastic drum. The 20 L drum was placed above the generator so that gravity was the driving force of water through the system. A tap was used to control the water flow rate. Water exited the generator through a 1/4 inch plastic tube fitted with a valve to prevent external air entering the generator. The vertical position of the exit tube was used to set the height of water in the generator. In these experiments the height was set for a bubble-rise distance of ~31 cm. This corresponded to a water volume of 200 mL. Particle-free air was bubbled through the fritted tip at a flow rate of 100 mL min -1 to produce a steady stream of bubbles.
Bubble size distribution was not measured in these experiments. A sample outlet at the top of the generator was used to extract SSA produced by bursting bubbles at 2 L min -1 for VH-TDMA analysis. An inlet tube led from the top of the generator to just above the air-water interface to allow time for particle-free make-up air to mix with SSA before being sampled by the VH-TDMA. All experiments were conducted at room temperature (25°C) and SSA was dried (<10% RH) before it entered the VH-TDMA. The dry size distribution (9-379 nm) of SSA produced in our generator consisted of a dominant accumulation mode centred at ~80 nm ( Figure 2), which compares well with other bubble-generated SSA size distributions (Martensson et al., 2003;Sellegri et al., 2006;Tyree et al., 2007).

Sample water
Experiments were conducted with two main types of water in the bubble generator: artificial sea salt solution (artificial SW) and natural seawater (natural SW). In addition the results were compared to VH-TDMA measurements of NaCl particles generated from an atomised solution of NaCl in ultra-pure deionised water. Artificial sea salt solution was generated by dissolving analytical grade sodium chloride (NaCl), magnesium chloride (MgCl 2 ), sodium sulphate (Na 2 SO 4 ), calcium chloride (CaCl 2 ), potassium sulphate (K 2 SO 4 ), sodium bromide (NaBr) and potassium nitrate (KNO 3 ) in ultra-pure deionised water. Two artificial salt solutions were prepared with varying ionic composition (Seinfeld and Pandis, 2006: pge. 444;Niedermeier et al., 2008).
Natural SW was collected at high tide on 28 January 2009 from Redcliffe Jetty, which extends 200 m into the north-western section of Moreton Bay on the east coast of Australia. Two minor river systems lie ~11 km to the north-west (Caboolture River) and ~8km to the south-west (Pine River) of the sampling site. As such the sampling site is subject to significant terrestrial run-off. The salinity of the collected samples was 31.8 g L -1 , measured via electrical conductivity (Eaton et al., 2005). The organic fraction or biological activity of our natural SW samples was not measured. However, monthly chlorophyll a (chl a) measurements (absorbance spectroscopy) at three sampling sites within a ~4 km radius of the sampling point were provided by the . Therefore we assume that the DOC content of our natural seawater was less than 20 mg L -1 . Natural SW samples were refrigerated in the dark and used within 2 weeks of the collection date. They were brought to room temperature and thoroughly stirred before bubbling experiments began.

Volatility Hygroscopicity-Tandem Differential Mobility Analyser (VH-TDMA)
The Volatility Hygroscopicity-Tandem Differential Mobility Analyser (VH-TDMA) has been described in detail elsewhere (Fletcher et al., 2007;Johnson et al., 2004;Modini et al., 2009) and will only be discussed very briefly here. The VH-TDMA was used to measure the average diameter of initially monodisperse SSA particles as they were heated in a thermodenuder (residence time = 0.3 secs) from ambient to 583°C in temperature increments of 15-60°C. Even after particle shrinkage occurred at higher temperatures the particles retained a monodisperse distribution. This means the average diameters before and after volatilisation could be used to calculate average volume fraction of SSA remaining (V/Vo). Volatility curves of different particle types were constructed by plotting V/Vo versus volatilisation temperature.
In addition, the VH-TDMA simultaneously measured the Hygroscopic Growth Factor at 90% RH (HGF 90% ) of the volatilised particles at each temperature. HGF 90% measurements of non-spherical particles taken with a (V)H-TDMA should be corrected for shape effects so they can be compared with independent measurements and theoretical predictions. The non-sphericity of dry NaCl particles is well described.
For the range of NaCl particle sizes investigated in this study (65-98 nm) we applied a size-dependent shape correction factor that varied from 1.213-1.199 (Biskos et al., 2006). There is evidence to suggest that natural and artificial SSA particles are also non-spherical in shape and can be described with the same size-dependent shape correction factor as NaCl (Niedermeier et al., 2008;Wise et al., 2009). Therefore we also applied the NaCl shape correction factor to natural and artificial SSA particles in this study. To remove the influence of the Kelvin Effect on the HGF 90% measurements taken at different sizes they were converted to bulk HGF 90% values (i.e. where a w =RH=0.9) using a constant single parameter representation of hygroscopic growth (Petters and Kreidenweis, 2007). The bulk HGF 90% values are reported here. All VH-TDMA data were inverted using the TDMAinv algorithm .
Difference in the VH-TDMA volatility curves of natural and artificial SSA were used to investigate, and then quantify, the organic fraction of natural accumulation mode SSA. This approach is based on the assumptions that 1) natural SSA potentially contains a seawater-derived organic fraction that is not present in artificial SSA, 2) this organic fraction is more volatile than the inorganic fraction of SSA, 3) the inorganic composition of artificial and natural SWs used in these experiments is very similar, and 4) any organic impurities present in the artificial SSA were also present in the natural SSA. The third assumption was tested by using two types of artificial sea salt solution with varying inorganic composition so we could judge whether small differences in the inorganic composition of artificial SSA translated into measurable differences in the VH-TDMA volatility curves. The fourth assumption is considered a reasonable one because the pre-cleaning process of the bubble generator was constant for all experiments and artificial SW was prepared with ultra-pure deionised water.
The advantages of using the VH-TDMA to measure the organic fraction of accumulation mode SSA are that only relatively small concentrations of particles (~100 cm -3 ) are required for the analysis, total scan time is only 1-2 hrs and the lower size limit is ~10 nm. Therefore only a small bubble generator and sample of water are required (see Table 1), which reduces the chances of organic contamination. In addition, at RHs above the deliquescence point of SSA (~75%) organic components will decrease the hygroscopic growth factor of SSA (Ming and Russell, 2001). This means the HGF 90% measurements taken by the VH-TDMA can be used to confirm the conclusions drawn from the volatility measurements. water flow rate we used was 0.3 L min -1 . This value was large enough to ensure that the water:air flow ratio was higher and water residence time lower then studies where large concentrations of organics have been detected in the aerosol phase. At the other extreme we performed one experiment with static water in the bubble generator that was left to bubble for 24 hrs before a VH-TDMA scan was conducted.

Experimental conditions
The influence of particle size on organic fraction was investigated because, as stated above, previous studies have shown that the organic fraction of SSA increases with decreasing particle size (Facchini et al., 2008;Keene et al., 2007). Most VH-TDMA scans were performed on natural SSA particles 71-77 nm in mobility diameter because this was near the centre of the accumulation mode of SSA particles produced from our bubble generator, as measured by the VH-TDMA in scanning mobility particle sizer (SMPS) mode ( Figure 2). In addition scans were also performed for particles towards the lower end (38 nm) and upper end (173 nm) of the SSA accumulation mode.

VH-TDMA volatility curves
3.1.1 The organic fraction of 71-77 nm natural SSA particles the VH-TDMA. In this notation N refers to natural SSA particles, the subscript number refers to the sample water flow rate through the bubble generator in L min -1 and the superscript number refers to the particle mobility diameter in nm (see Table   1). In addition the A 1 , A 2 , A 3 , NaCl 1 and NaCl 2 volatility curves are included for comparison (the subscript number here is simply an index). The NaCl particles (square markers) were very stable as volatilisation temperature was increased. A significant decrease in V/Vo was only observed at the highest temperature obtained in these experiments, 583°C. This sudden decrease indicated evaporation of NaCl had begun, which is consistent with the onset temperature for particle formation in evaporation/condensation NaCl aerosol generation experiments (Scheibel and Porstendo¨rfer, 1983). The artificial SSA particles (circle markers) were more volatile than the pure NaCl particles. Artificial SSA V/Vo decreased fairly steadily as temperature increased so that only 82-83% of particle volume remained at 520°C. In contrast 96% of NaCl particle volume remained at this temperature. The increased volatility of artificial SSA compared to NaCl particles could be because the evaporation or melting point of the mixture of inorganic salts was lower than the equivalent point for any pure salt in that mixture. The volatility curves of all three artificial SSA experiments agreed within the V/Vo measurement uncertainty. This indicates that small differences in the size and inorganic composition of artificial SSA does not translate into significant changes in the volatility curves, which confirms assumption number 3 of our VH-TDMA measurement approach.
The 71-77 nm natural SSA particles (diamond markers) were even more volatile than the artificial SSA particles. The natural SSA volatility curves began to diverge from the artificial SSA curves at 170°C (see inset, difference was 8%. The standard deviation of the difference was 2% and the theoretical uncertainty was 6% (twice the V/Vo measurement error). We take the larger value of ±6% as the absolute error in the measured organic volume fraction of 8%.

Dependence of organic fraction on water flow rate through the bubble generator
The volatility curves of 71-77 nm natural SSA particles generated under varying water flow rates agreed almost completely within measurement uncertainty. That is, the organic fraction of 71-77 nm natural SSA particles produced in our generator did not depend on water flow rate. If anything, the N 0 71 curve lies slightly below all others. This indicates that the rate of transfer of organic material out of our bubble generator by SSA is not sufficient to deplete the organic content of a 200 mL sample of natural SW, even after it has been bubbling at 100 mL min -1 for 24 hrs. This is consistent with calculations of the amount of organic material exported by SSA from seawater in the bubble generator. As well as the measured accumulation mode (see Fig. 2) there should have been a second super-micrometer mode in our SSA size distribution (e.g. Keene et al., 2007;Martensson et al., 2003). Therefore, we assumed a bi-modal distribution with certain properties (accumulation mode: median diameter = 0.08 μm, concentration = 10 000 cm -3 , organic mass fraction = 80%; supermicrometer mode: median diameter = 4 μm, concentration = 200 cm -3 , organic mass fraction = 10%) and an organic density of 1.1 g cm -3 (Keene et al., 2007). Although we didn't measure such large organic fractions in this study we purposefully overestimated them for this calculation. Under these assumptions and the conditions of our bubbling experiments only 2 x 10 -4 g day -1 of organic material would be exported by SSA from our sample water. If we assume the organic content of our 200 mL sample of seawater was only 1 mg L -1 , it would take 23 hours of SSA generation to deplete all the organics in the water. If we assume the seawater organic concentration was 10 mg L -1 , full depletion would take 234 hours. The fact that organics were not depleted in our generator after bubbling for 24 hrs suggests that the organic content of our sample water was greater than 1 mg L -1 , or that we have overestimated the number and organic fraction of super-micrometer SSA particles in this calculation. curves. It appears that volatility increased slightly with decreasing particle size. In the temperature range 200-500°C where it is expected that the organic fraction of the particles has evaporated the average difference (±1 standard deviation) between the 173 nm and 38 nm curves is 6±3%. The average difference between the 173 nm and 71 nm curves and the 71 nm and 38 nm curves in the same range is 2±2% and 3±2%, respectively. However, these volatility differences could be due to differences in particle size as well as composition (organic fraction). Therefore these values do not represent the difference in organic volume fraction for the different particles sizes.
Rather they overestimate these fractions by an unknown amount equal to the percentage change in V/Vo due to the change in initial particle size. Taking this into account and the fact that the theoretical uncertainty in the calculation of organic volume fraction is 6%, we conclude that the organic fraction of natural SSA particles did not vary significantly with particle mobility diameter in the range 38-173 nm.

Hygroscopic growth factor measurements
To improve the representation of the shape-and Kelvin-corrected bulk HGF 90% data all of the measurements were first categorised as NaCl, artificial SSA or natural SSA (71-77 nm) particles. Then a number of measurements were averaged at specific temperature values to obtain average HGF 90% for each particle type as a function of temperature. These averages are plotted in Figure 5. At ambient temperature NaCl HGF 90% was 2.46±0.02. Artificial SSA HGF 90% was 4.4% lower at 2.35±0.02. These values both agree well with theoretical predictions of NaCl and artificial SSA HGF 90% (e.g. Ming and Russell, 2001). Natural SSA HGF 90% was 3.9% lower than artificial SSA (8.1% lower than NaCl) at 2.26±0.02.
It is instructive to observe how HGF 90% for each particle type varied as a function of temperature. NaCl HGF 90% was fairly constant until particle evaporation began at the highest temperatures. At this point NaCl HGF 90% decreased. Artificial SSA HGF 90% continually increased with increasing volatilisation temperature. Natural SSA HGF 90% was below artificial SSA HGF 90% up to a volatilisation temperature of 170°C. At this temperature the organic component of the natural SSA particles began evaporating (see Figure 3). Coinciding with this the HGF 90% curve started approaching the artificial SSA HGF 90% curve. At temperatures above 206°C the artificial and natural HGF 90% curves agreed almost completely within experimental variation. The shapes of the two curves were even very similar.
Under the ZSR approximation (Chen et al., 1973;Stokes and Robinson, 1966) it is possible to investigate whether the difference in the natural and artificial SSA HGF 90% curves at lower temperatures is consistent with the volume fraction of organics in the natural SSA particles as calculated from the volatility data (see section 3.1.1). When making the ZSR approximation it is assumed that the individual components of an internally mixed particle do not interact with each other and therefore they uptake water independently. In practice this means that the HGF of a mixed particle can be calculated by the volume-fraction-weighted sum of the HGF's of individual components in that particle. We can use this assumption to predict HGF 90% for our natural SSA particles assuming they are a binary mixture of an organic and inorganic (sea salt) component. For input into the ZSR approximation we use our measured organic volume fraction as a function of temperature, our measured bulk HGF 90% of artificial SSA as a function of temperature and assume a bulk HGF 90% for the organic component of 1. This leads to a ZSR predicted HGF 90% curve which is plotted in Figure 5. There is generally good agreement between the ZSR predicted and measured natural SSA HGF 90% curves. At temperatures less than 200°C the ZSR predicted curve only slightly overestimates the measurements. At temperatures greater than 200°C the ZSR predicted HGF 90% curve equals the artificial HGF 90% curve because it is assumed that all organics have evaporated from the natural SSA particles and organic volume fraction is set to 0 (Fig. 3).
In summary, the HGF 90% data are consistent with the conclusions drawn from the volatility data. Namely, that our natural SSA particles had a minor organic component that evaporated over the temperature range 170-200°C. After evaporation the natural and artificial SSA particles had similar, predominantly inorganic composition.

Implications
We have measured an organic volume fraction of 8±6% for 71-77 nm natural SSA particles that were generated from samples of coastal seawater that most likely had high organic content. Note that this means aerosol-phase organics were still enriched by tens or hundreds of times relative to the sample water, depending on the exact concentration of organics in the sample water. Assuming an organic density of 1.1 g cm -3 (Keene et al., 2007) our measurement corresponds to an organic mass fraction of only 4%. We also investigated the organic fractions of 38 and 173 nm natural SSA particles and found these did not differ significantly from the organic fraction of 71-77 nm particles. In comparison, Keene et al. (2007)  Hygroscopicity measurements in the literature also provide indirect evidence that accumulation mode SSA particles often contain only a minor organic fraction. Sea salt aerosol (i.e. purely inorganic) is very hygroscopic. If a major, non-hygroscopic organic fraction is present in SSA it will significantly decrease the hygroscopicity of that aerosol. For example an SSA particle consisting of 20% sea salt (HGF 90% = 2.35) and 80% organics (HGF 90% = 1) will have HGF 90% = 1.5 according to the ZSR approximation. A few studies have reported HGF's above deliquescence RH for accumulation mode natural SSA particles that are only slightly below (< 10%) corresponding NaCl or sea salt HGF's (Niedermeier et al., 2008;Sellegri et al., 2008;Swietlicki et al., 2008). This suggests the natural SSA particles investigated in these studies did not contain large organic fractions. In a very recent study Herich et al. (2009) detected an organic component in both fresh and aged 260 nm SSA particles at a remote continental site in the arctic circle in northern Sweden. The authors found that SSA organic content did not correlate with SSA hygroscopicity. This implies that the organic component only formed a very minor fraction of total SSA mass, because changes in the amount of organics present had no effect on particle hygroscopicity.
Although these studies do not report measurements (e.g. organic content, biological activity) of the source water from which aerosols were generated, they nevertheless suggest that accumulation mode SSA frequently contains only a minor organic fraction.
Discrepancies between the different studies could be related to not only the amount of organics present in the source waters used in each experiment, but also the composition and surface-active nature of those organics. For example Facchini et al. (2008)  However, we note that no significant differences were observed between scans completed at different times during the 2 week measurement period. Therefore any artefacts due to storage are likely to be minimal. The bubble generator employed in this study was also far smaller than those used in previous studies (see water volumes in Table 1). While this reduced the risk of external organic contamination, it also meant our generator had a high surface to volume ratio. As seawater flowed bottomto-top in the generator organics potentially adsorbed to the walls thereby reducing the amount of organics eventually transferred to the aerosol. This potential loss mechanism was not quantified or estimated. We do not believe that these methodological differences can account for the order of magnitude difference between our measured accumulation mode SSA organic fraction and the fractions measured in the Keene and Facchini studies.
O'Dowd et al. (2008) have developed a combined organic-inorganic sub-micron sea spray source function for modelling purposes. One input into this source function is the organic mass fraction of sub-micron SSA as a function of chl a concentration, which was derived from ambient measurements conducted at Mace Head, Ireland.
This function saturates at 90% organic mass fraction for chl a concentrations above These considerations point to the need for further independent, size-resolved measurements of the organic fraction of SSA produced from a variety of different seawaters. Based on the conflicting studies, it seems that there may be some additional properties of seawater (e.g. organic composition, surface-active nature of organics) that control how much organic material is transported from water to the aerosol phase during the bubble bursting process. In addition, our results suggest that if these experiments are conducted with bubble generators, it may not be necessary to cycle water through the generator to maintain a fresh supply of seawater-organics.
Bubble-generated SSA did not deplete the organic content of static seawater in our bubble generator over a 24 hr period.

Conclusion
A bubble generator was constructed and used to produce SSA particles from samples of coastal seawater collected from Moreton Bay on the east coast of Australia.   natural SSA particles (diamonds) generated using different water flow rates. Legend notation is described in text and Table 1. Error bars represent ±3% measurement uncertainty in V/Vo.
Inset graph is magnified version of main graph with error bars removed Figure 4: Volatility curves of 38 nm (upside down triangles), 71 nm (diamonds) and 173 nm (squares) natural SSA particles. Legend notation is described in text and Table 1. Error bars represent ±3% measurement uncertainty in V/Vo Figure 5: Shape-and Kelvin-corrected bulk HGF 90% values for NaCl (squares), artificial SSA (circles) and natural SSA (diamonds) particles as a function of volatilisation temperature.
Each data point represents an average of a number of measurements and error bars represent ±1 standard deviation. Measurement uncertainty in HGF 90% was ±3%. Also included is the ZSR predicted HGF 90% curve (solid red line). See text for details on the calculation of this curve