Aerosol hygroscopicity in the marine atmosphere : a closure study using high-resolution , size-resolved AMS and multiple-RH DASH-SP data

S. P. Hersey, A. Sorooshian, S. M. Murphy, R. C. Flagan, and J. H. Seinfeld Departments of Chemical Engineering and Environmental Science and Engineering, Caltech, 1200 E. California Blvd, Pasadena, CA, 91125, USA Received: 14 July 2008 – Accepted: 6 August 2008 – Published: 4 September 2008 Correspondence to: J. H. Seinfeld (seinfeld@caltech.edu) Published by Copernicus Publications on behalf of the European Geosciences Union.


Introduction
Atmospheric aerosols change size with fluctuations in relative humidity, with a magnitude dictated by chemical composition.Because this hygroscopic response determines particle size, it influences direct climate forcing attributed to aerosols.Further, subsaturated hygroscopic growth factor (D p,wet /D p,dry ) is strongly correlated with CCN activity (Prenni et al., 2001).Given the importance of aerosol water uptake on both the direct and indirect light scattering properties of aerosols, incomplete understanding of aerosol hygroscopicity has been identified as a major limitation in estimations of climate forcing (IPCC, 2007).
Closure studies, which attempt to reconcile simultaneously measured hygroscopic and chemical data, link laboratory studies of hygroscopicity, theoretical models for water uptake, and field measurements of aerosol-water interactions.The standard method for predicting hygroscopic growth from composition data is based on volume-Introduction

Conclusions References
Tables Figures

Back Close
Full weighted water uptake by the individual chemical constituents.While it is usually possible to predict water uptake for the inorganic fraction of atmospheric aerosols, the wealth of organic species in the atmosphere, combined with limited understanding of organic aerosol hygroscopicity, has led investigators to assign the water uptake of organics as the particle growth not explained by inorganic constituents (e.g., Malm et al., 2005).
An important approach in hygroscopicity closure is to represent organics as a combination of water-soluble and insoluble fractions, corresponding to oxygenated (OOA) and hydrocarbon-like (HOA) carbon, respectively.One notable study in an urban atmosphere (Gysel et al., 2007) used high-resolution chemical data from the aerosol mass spectrometer (AMS, Jayne et al., 2000) to separate the organic component into OOA and HOA by using m/z 44 and 57 signals, respectively, with a deconvolution method developed by Zhang et al. (2005).In the study, division of organics between hygroscopically active and inactive fractions provided good agreement between predicted and measured hygroscopicity values.This study represents an improvement in the hygroscopic treatment of organics over prior work, but applies only to urban aerosol and is limited by both low time resolution and single-RH conditions inherent in HTDMA systems.
The current study presents data obtained during seven flights in the marine atmosphere off the coast of Central California during the second Marine Stratus/Stratocumulus Experiment (MASE-II).The dataset is the first to combine hightime-resolution, size-resolved AMS chemistry with high-time-resolution, size-resolved hygroscopic data at multiple RH values from the Differential Aerosol Sizing and Hygroscopicity Spectrometer Probe (DASH-SP).Hygroscopic studies have previously been carried out in the marine atmosphere (see Table 1), but none with chemical and hygroscopic data as highly size-, time-, and RH-resolved as those presented here.et al., 2007), and MASE II was undertaken in 2007 in the same region.Both experiments were carried out in the month of July, when marine stratocumulus are prevalent over the region, and utilized the CIRPAS Twin Otter aircraft.In each campaign, comprehensive airborne measurements were made of aerosol and cloud properties in areas both perturbed and unperturbed by local anthropogenic emissions.Tables 2  and 3 list the flights carried out during MASE II and the instrument payload onboard the Twin Otter, respectively.The present study addresses measurements of the hygroscopic properties of marine aerosols during MASE II.Other flights probed emissions from a large bovine source and a large container ship, and these data are presented elsewhere (Sorooshian et al., in preparation; Murphy et al., in preparation).

Aerosol composition measurements
Non-refractory aerosol chemical species were characterized by the Aerodyne compact Time of Flight Aerosol Mass Spectrometer (cToF-AMS, Drewnick et al., 2004a,b).In the AMS, particles with vacuum aerodynamic diameters (D p,va ) 50 nm≤D p,va ≤800 nm are focused by an aerodynamic lens, pass through a 3.5% chopper, and are vaporized at 500 • C. The chopper has three modes of operation, which detect background mass spectra, ensemble average mass spectra over all particle sizes, or size-resolved mass spectra.After vaporization, all molecules are ionized via electron impact, and undergo

Hygroscopicity measurements
Hygroscopicity measurements were carried out with the Differential Aerosol Sizing and Hygroscopicity Spectrometer Probe (DASH-SP, Brechtel Mfg; Sorooshian et al., 2008).Ambient particles pass through a nafion dryer before receiving a uniform charge distribution in a 210 Po neutralizer.A cylindrical, single-classification differential mobility analyzer (DMA) then size selects particles into narrow ranges of mobility-equivalent diameters (D p,em ) between 0.1 and 1.0 µm.The resulting monodisperse aerosol is split into five separate flows.One channel provides a redundant measurement of total particle concentration at the DMA-selected size with a water condensation particle counter (TSI Model 3831).The remaining four channels consist of parallel nafion humidification chambers (Perma Pure, LLP, Model MD-070-24FS-4), followed by cor-Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc

Printer-friendly Version
Interactive Discussion respondingly humidified custom optical particle counters (OPCs).In the OPC sample volume, particles pass through a focused laser beam (λ=532 nm, World Star Technologies, Model TECGL-30) and scatter light in proportion to size (D p ) and refractive index (RI).Forward-scattered light is collected and focused on a photomultiplier tube, and the resulting electrical pulse is recorded by a high-speed data acquisition computer.
An iterative data processing algorithm, based on laboratory calibrations with salts of known refractive indices, is used to determine the best fit on a solution surface relating electrical pulse height, size, and refractive index.The hygroscopic growth factor (GF =D p,wet /D p,dry ) is corrected for the RI change caused by particulate liquid water at elevated RH.In the current study, hygroscopicity was measured at dry sizes corresponding to D p,em of 150, 175, and 200 nm.Multiple RH sensors in the nafion tubes and OPCs controlled RHs to dry (<8%), 74%, 85%, and 92%, with RH uncertainty of 1.5%.Low particle loadings inherent in the marine atmosphere required increased online collection times at each DMA size step, but usually ≤ one minute was sufficient to overcome counting statistic limitations.Overall uncertainty in GF calculations is 4.5%.
Assuming particles to be uniform, non-light-absorbing spheres allows the assumption that the intensity of scattered light is a function of only RI and D p .This assumption also allows calculation of dry, 'effective' RI from the known DMA-selected D p and measured scattered light intensity.

Hygroscopic closure
A volume-weighted mixing rule was used to perform a hygroscopic closure using AMS and DASH-SP data, under the assumption of independent and additive water uptake by individual constituents in each particle: (1) Introduction

Conclusions References
Tables Figures

Back Close
Full where GF mixed is the overall particle GF , a w is the water activity, GF i is the hygroscopic growth factor for pure species i , and i is the volume fraction of species i .At equilibrium, a w =RH (Seinfeld and Pandis, 2006).Values for i were calculated for the following species, using AMS masses of ammonium (NH 4 ), and total organic: ammonium nitrate (NH 4 NO 3 ), ammonium sulfate ((NH 4 ) 2 SO 4 ), ammonium bisulfate (NH 4 HSO 4 ), sulfuric acid (H 2 SO 4 ), and organic.Partitioning between sulfate species was determined on the basis of the ammonium to sulfate molar ratio.(NH 4 ) 2 SO 4 is assumed to have a GF of unity at 74% since particles are exposed to RH well below its effloresence point before subsequent rehumidification.The organic fraction was assumed to be hydrophilic WSOC, based on evidence of a high degree of organic oxygenation from AMS mass spectra (see Sect. 3.3).Values of GF i for the organic fraction were calculated as those necessary to minimize the root mean square error in comparing predictions with measured hygroscopicity.
Table 4 presents measured GF s at each RH and each D p,dry .Typically, multiple measurements were made on each leg, at each size, for each flight.The error reported is ± one standard deviation in these multiple measurements.When only one measurement was made at a given size in a given leg, it is reported without error.Ship plumes were encountered on flights 10 and 16, as evidenced by brief, significantly elevated particle number concentration.Analysis of GF measurements in these presumptive plumes is not presented in the present work.nm D p,dry particles averaged 1.61±0.14, as compared with 1.91±0.07for all other flights.These low-GF flights corresponded to significantly elevated total organic, as measured by the AMS.Mass concentration averaged 1.97±1.71µg/m 3 organic (as opposed to 0.58±0.63µg/m 3 for all other flights), corresponding to volume fraction organic (V F organi c ) of 0.46±0.22(as opposed to 0.24±0.18for all other flights).
Back-trajectory analysis suggests that the MASE-II flights can be categorized by airmass origin as either clean/marine (flights 7, 10, 11, 13, 14) or polluted/continental (flights 12, 16). Figure 1 shows 92% GF measurements for 200 nm D p,dry particles, with corresponding 48-hr HYSPLIT (available at http://www.arl.noaa.gov/ready/hysplit4.html)back-trajectories identifying airmass origin.Note that the low GF s and high V F organic measured on flights 12 and 16 correspond with airmass origins over the continental United States, while higher GF s and lower V F organic measured on other flights correspond to airmass origins over the clean marine environment.It is interesting to note that trajectories at sea level have marine origins for all flights, including low-GF flights 12 and 16, suggesting that aloft airmass origin is a more significant factor in determining aerosol characteristics.Further, airmasses corresponding to clean, high-GF flights often pass either directly over or very near the San Francisco Bay Area, but these trajectories do not appear to significantly affect aerosol GF or V F organic .It is also noteworthy that GF values at low RH were not significantly suppressed, with values of 1.31±0.06at 74% (compared with 1.31±0.07for all other flights).GF values at intermediate RH were moderately suppressed in the continental airmass, measuring 1.53±0.10 at 85% (compared with 1.58±0.08for all other flights).In other words, the effect of the high-V F organic , polluted/continental airmass is to significantly suppress GF at high RH, while having no measurable effect on aerosol water uptake at low RH and a moderate impact at intermediate RH.

Hygroscopicity trends
No size-dependent hygroscopicity was observed over the range of measured D p,dry , despite slightly elevated V F organic at smaller sizes.This suggests that minor size-Introduction

Conclusions References
Tables Figures

Back Close
Full dependent changes in composition had too small an impact on particle GF to be detected by the DASH-SP.
Figure 2 shows below-and corresponding above-cloud 92% GF values for all flights.There exists a ubiquitous trend of higher below-cloud aerosol GF values at high RH when compared with top of cloud hygroscopicity values (1.88±0.14below cloud, versus 1.78±0.18above cloud).Marker sizes, proportional to V F organic suggest a trend of higher organic loading above cloud in several flights (0.27±0.15 below cloud, versus 0.31±0.21above cloud).While variability is rather large in these measurements, the trend of elevated V F organic above-cloud does correspond to suppressed GF at high RH.There is a trend of more moderate GF suppression at intermediate RH in the higher V F organic above-cloud layer (1.53±0.06above cloud, as compared with 1.58±0.07below cloud).Unlike the continental-influenced flights, there is evidence of more significant suppression of low-RH GF s in above-cloud legs, with 74% GF values of 1.31±0.05above cloud, compared with 1.36±0.05below.It appears, then, that the elevated organic loadings typical of above-cloud legs are correlated with GF suppression at high RH, and more moderate GF suppression at lower RHs.

Hygroscopic closure
A hygroscopic closure was performed, using volume-weighted hygroscopic contributions from each chemical constituent identified by the AMS.O:C ratios for the flights presented were 0.92±0.33,with very similar ratios for below, above, and free troposphere legs on all flights.This result indicates a consistently uniform, highly oxidized organic component (Aiken et al., 2007), with little'fresh,' or HOA-type organic carbon.With this evidence, the organic fraction was treated as a bulk, water-soluble constituent (Kondo et al., 2007), as opposed to being speciated into a soluble and insoluble OOA and HOA according to Zhang et al. (2005).As described in Sect.2.4, GF values were calculated for the organic component of the aerosol by minimizing root mean square error when comparing measured GF s with volume-weighted closure predictions.Mission-averaged organic GF s were determined to be 1.20, 1.43, and 1.46 at 16798 Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version Interactive Discussion 74, 85, and 92% RH, respectively.These values agree well with those of hydrophilic multifunctional and dicarboxylic acids studied by Peng et al. (2001).Closure results are presented in Fig. 3. Markers in Fig. 3 are color-coded according to relative humidity, and marker size is proportional to V F organic .The volume-weighted hygroscopic closure utilizing size-resolved AMS chemistry achieves good agreement with the 675 DASH-SP GF measurements, with an R 2 of 0.79.Agreement is better at lower values of RH, owning to smaller GF magnitudes and less overall GF variability.Aside from the obvious RH dependence of GF values, the clearest trend in Fig. 3 is that of larger markers (higher V F organic ) at low GF transitioning to smaller markers (lower V F organic ) at high values of GF for the same RH.The clarity and regularity of this trend reveal the importance of the organic fraction in determining GF values, and suggest that, of all the chemical species considered in the closure, the organic fraction is a particularly crucial parameter in predicting GF .

Simplified parameterization
To further investigate the relative importance of each parameter in quantifying aerosol water uptake, and to detemine the simplest statistical model still capable of accurately predicting GF , a backward stepwise linear regression was performed.The process, which eliminates predictors one-by-one to generate increasingly simplified linear representations of data, started with over 60 predictors, ranging from PILS and AMS chemical parameters to atmospheric data.The result is a two-parameter model that predicts GF as a function of RH and V F organic : addition to organic fraction with its associated GF ).So even with information on the specific nature of organics in the atmosphere, it appears as though accurate predictions of particle water uptake in the marine atmosphere can be made simply on the basis of relative humidity and the relative abundance of organics in the aerosol.While this model accurately predicts GF over the range of RH and V F organic measured during MASE-II, it should not be used when RH is outside the range 74%≤RH≤92%, or where V F organic is less than 0.1.

Discussion
Comparing mixed organic-inorganic particles with those comprised entirely of inorganic salts, there is a strong RH-dependence in the effect of organics on hygroscopicity (Peng et al., 2001).During the MASE-II field campaign, GF values at 74% RH averaged ∼1.3.Organic GF s were inferred to be 1.20 at 74% RH, suggesting that they contributed significantly to overall aerosol water uptake at low RH.GF values at 85% RH averaged ∼1.6 during the campaign, and so the organic GF of 1.43 calculated for 85% RH suggests that organics played a less significant role, but still influenced water uptake at intermediate RH.An organic GF value of 1.46 was calculated for 92% RH, while 92% GF measurements averaged ∼1.8, suggesting that organics contributed little to overall GF at high RH.Inorganic salts exhibit deliquescent behavior as RH is increased.Many organics do not deliquesce or crystallize when RH is increased or decreased, respectively, but instead retain water at RH values well below the RH of deliquescence (RHD) of the inorganic salts with which they often co-exist in ambient particles.As a result, at RH values below the salt RHD, the presence of organics enhances water uptake.Thus, the effect of organics is to contribute significantly to overall water uptake at low values of RH in mixed organic-inorganic particles.At high RH, on the other hand, organics tend to take up significantly less water than the inorganic constituents with which they coexist in ambient particles.Therefore, at RH values above the inorganic RHD, organics Introduction

Conclusions References
Tables Figures

Back Close
Full appear to suppress water uptake relative to that which a pure inorganic particle would exhibit.
Figure 5 shows the behavior of a pure ammonium sulfate (AS) particle, pure organic acid (OA) particle, and a mixed organic acid and ammonium sulfate (OA/AS) particle over the range 50%≤RH≤94%.Note that the mixed OA/AS particle shows smooth growth with RH, as opposed to the deliquescent behavior exhibited by the pure AS particle.The tendency of organics to retain water at low RH causes water uptake behavior for the OA/AS particle to follow that of the descending (i.e.efflorescence) branch of the pure AS growth curve.Since pure OA takes up less water at high RH than does pure AS, the growth curve for the OA/AS particle is suppressed, compared with that of pure AS.The overall result, as predicted by thermodynamic theory, is that the presence of OA leads to enhanced water uptake at low RH and suppressed GF at high RH.Virkkula et al. (1999) concluded that the most important factor contributing to GF suppression at high RH was the volume fraction of organic present in an aerosol.Others have suggested that the exact chemical identity of the organic constituents is not especially important, and that for an organic component classified as either oxidized or hydrocarbon-like, its relative abundance determines its effect on GF values (McFiggans et al., 2005;Moore and Raymond, 2008).Results presented here from a stepwise linear regression on GF data from the marine atmosphere suggest that the single most important factor in predicting GF (aside from RH) is, indeed, V F organic .
In most instruments that measure aerosol hygroscopicity, residence times for humidification are on the order of seconds, much longer than the equilibration time for most inorganics with water vapor.Sjogren et al. (2007) noted, however, that particles with high volume fraction organic material may require as long as 40 s to achieve equilibrium with water vapor.If such long times are necessary to achieve equilibrium, hygroscopic measurement methods suitable for the field will tend to overpredict GF at low RH (water vapor does not evaporate completely from the particle during the drying process), while underpredicting GF at high RH (insufficient humidification time is Introduction

Conclusions References
Tables Figures

Back Close
Full provided for the organic fraction to achieve thermodynamic equilibrium with water vapor).In electrodynamic balance (EDB, (Cohen et al., 1987a,b,c)) studies, suspended particles are subjected to extended exposure to water vapor (minutes to hours), establishing equilibrium.Some organics exhibit extremely high deliquescence relative humidities (DRH) (e.g., oxalic acid), while others exhibit gradual hygroscopic growth at low RH and substantial growth at high RH (e.g., malonic acid) (Peng et al., 2001).It is possible, given the wide range of organic species in the atmosphere and correspondingly wide range of hygroscopic properties associated with those species, and the relatively short humidification times in the DASH-SP and other similar instruments, that the effects attributed to organics may reflect some kinetic instrumental limitations.

Conclusions
We report a hygroscopic closure study for the marine aerosol, using high-resolution AMS chemical data coupled with highly time-resolved, multiple-RH hygroscopicity measurements from the DASH-SP.
Airmasses originating from continental locations showed elevated organic loading, and corresponded to significant GF suppression at high RH.More moderate GF suppression was measured at intermediate RH and no impact was observed at low RH.A comparison of above-cloud with below-cloud aerosol indicated that a slightly organicenriched layer above cloud corresponded with suppressed GF s at high RH.A volumeweighted hygroscopic closure provided excellent agreement with measured GF s, and mission-averaged organic GF s at 74, 85, and 92% were calculated to be 1.20, 1.43, and 1.46, respectively.These GF values are relatively high when compared with many previous estimations of organic GF , but agree well with values reported for dicarboxylic and multifunctional acids.These high organic GF s are indicative of the highly oxidized state of the aged organic fraction.A simplified parameterization for predicting GF was developed using a stepwise linear regression method.This parameterization is a function of only RH and V F organic , and explains only 2% less variability than does  Gysel, M., Alfarra, M. R., Dommen, J., Metzger, A., Prevot, A. S. H., Weingartner, E., Laaksonen, A., Raatikainen, T., Good, N., Turner, S. F., McFiggans, G.,     Comparison of growth curves for pure ammonium sulfate, pure hydrophilic organic, and mixed organic/inorganic particles.Ammonium sulfate curves calculated from AIM (Clegg and Seinfeld, 2006).

3. 1
Airmass originRelative to the other flights, RF 12 and 16 exhibited significantly suppressed water uptake at high RH for all dry sizes.During these two flights, 92% GF values for 200

Figure 4
Figure 4 demonstrates the accuracy with which this model predicts DASH-SP GF values over measured the range of RH and V F organic .It is noteworthy that the R 2 for this model is 0.77, indicating that the simple, two-parameter model explains only 2% less variability than the full volume-weighted chemical closure, which contains significantly more information (i.e.multiple inorganic chemical species and their individual GF s, in 16799
The data presented here were obtained during a series of seven cloud probing flights carried out as part of the second Marine Stratus/Stratocumulus Experiment (MASE-II) field campaign during July 2007.The MASE II experiment was the second of two airborne field campaigns directed toward measurement of aerosol-cloud relationships in marine stratocumulus in the eastern Pacific Ocean.The Marine Stratus/Stratocumulus Experiment (MASE) was carried out in 2005 off the coast of Monterey, California (Lu weighted chemical closure.The GF predictions and parameterizations presented here accurately predict GF s from high-resolution chemical inputs, and may be broadly applicable to the marine environment, where some oxidized organics are present in an otherwise clean, atmosphere.The richness of this hygroscopic/chemical data set underlies the significance of coupling the DASH-SP instrument with the AMS.The improved time resolution available with the DASH-SP eliminates much of the time-averaging of AMS chemical data that is otherwise necessary with less highly time-resolved HTDMA hygroscopicity measurements.The importance of simultaneous GF measurements at multiple RH values is demonstrated by a simplified parameterization for predicting GF as a function of RH and V F organic ; a result potentially important for efficiently representing aerosol-water interactions in global models.Introduction

Table 4 .
GF results for below, above, free troposphere (FT), and ship plume measurements.