The initial phase (solid or aqueous droplet) of aerosol particles prior to activation is among the critical factors in determining their cloud condensation nuclei (CCN) activity. Single-particle levitation in an electrodynamic balance (EDB) was used to me

Environ. Sci. Technol. 2008, 42, 3602–3608 Measurements of the Hygroscopic and Deliquescence Properties of Organic Compounds of Different Solubilities in Water and Their Relationship with Cloud Condensation Nuclei Activities MAN NIN CHAN,† SONIA M. KREIDENWEIS,‡ AND C H A K K . C H A N * ,† Department of Chemical Engineering, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, and Department of Atmospheric Science, Colorado State University, Ft. Collins, Colorado 80523-1371 Received September 17, 2007. Revised manuscript received February 28, 2008. Accepted February 29, 2008. The initial phase (solid or aqueous droplet) of aerosol particles prior to activation is among the critical factors in determining their cloud condensation nuclei (CCN) activity. Single-particle levitationinanelectrodynamicbalance(EDB)wasusedtomeasure the phase transitions and hygroscopic properties of aerosol particles of 11 organic compounds with different solubilities (10-1 to 102 g solute/100 g water). We use these data and other literature data to relate the CCN activity and hygroscopicity of organic compounds with different solubilities. The EDB data show that glyoxylic acid, 4-methylphthalic acid, monosaccharides (fructose and mannose), and disaccharides (maltose and lactose) did not crystallize and existed as metastable droplets at low relative humidity (RH). Hygroscopic data from this work and in the literature support earlier studies showing that the CCN activities of compounds with solubilities down to the order of 10-1 g solute/100 g water can be predicted by standard Köhler theory with the assumption of complete dissolution of the solute at activation. We also demonstrate the use of evaporation data (or efflorescence data), which provides information on the water contents of metastable solutions below the compound deliquescence RH that can be extrapolated to higher dilutions, to predict the CCN activity of organic particles, particularly for sparingly soluble organic compounds that do not deliquesce at RH achievable in the EDB and in the hygroscopic tandem differential mobility analyzer. Introduction Organic compounds contribute a substantial amount to the aerosol mass and play an important role in radiative forcing of atmospheric aerosols (1). Knowledge of the hygroscopicity of organic particles is essential in gaining a better understanding of the hygroscopicity and cloud condensation nuclei (CCN) activities of atmospheric particles. Recently, experi* Corresponding author phone: (852) 2358-7124; fax: (852) 23580054; e-mail: keckchan@ust.hk. † Hong Kong University of Science and Technology. ‡ Colorado State University. 3602 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 10, 2008 mental and theoretical studies have examined the relationship between the hygroscopicity and CCN activity of laboratory-generated pure organic particles. Organic compounds in atmospheric particles have a wide range of solubilities in water (2). The activation of highly soluble organic compounds (e.g., malonic acid and glutaric acid (solubility > 100 g solute/ 100 g water)) can be predicted by the standard Köhler theory assuming complete dissolution of the solute at activation (3, 4). There are still uncertainties in predicting the CCN activity of sparingly soluble organic compounds that have solubilities on the order of 10-1 to 100 g solute/100 g water. The activation of some sparingly soluble organic compounds (e.g., aspartic acid, glutamic acid, homophthalic acid, and phthalic acid) agrees well with the predictions of standard Köhler theory, similar to more-soluble organic compounds (5, 6). On the other hand, the observed activation behavior of some sparingly soluble organic compounds (e.g., adipic acid and suberic acid) is not consistent with the predictions of standard Köhler theory. In addition to factors such as surface tension, wettability (5), aerosol generation method (7), aerosol morphology (8), curvature enhanced solubility (9), and impurities (10), the initial phase (solid or aqueous droplet) of single-component organic particles prior to activation has been taken into account in explaining the difference between modeled and experimental CCN data (6, 8, 10). Hori et al. (8) and Bilde and Svenningsson (10) have shown that metastable, highly concentrated aqueous–organic solution droplets had a much lower critical supersaturation than did dry solid particles of the same organic compounds. Hartz et al. (6) have observed that the existence of metastable solution droplets prior to activation can explain why some sparingly soluble organic compounds were highly CCN active, and why their critical supersaturations can be predicted by standard Köhler theory. Laboratory hygroscopicity studies have shown that when the organic particles are initially formed as solution droplets, some sparingly soluble organic compounds can sustain a high level of supersaturation before crystallization occurs (11, 12). Furthermore, crystallization was not observed in some organic compounds with solubilities as low as less than 100 g solute/100 g water (6). Hence, the occurrence of crystallization of droplets containing organic compounds cannot be directly inferred from their solubility. Phase transitions and hygroscopic growth measurements of organic particles with different solubilities are needed for interpreting and predicting the hygroscopicity and CCN activity of organic particles. In this study, we measure the hygroscopicity of 11 pure organic particles with different solubilities (10-1 to 102 g solute/100 g water) using single-particle levitation in an electrodynamic balance (EDB). These compounds were selected for their atmospheric relevance and available measured CCN activities. We report the hygroscopicity measurements and discuss how the phase transitions and hygroscopic growth of pure organic particles can be used to understand the organic particles’ CCN behaviors at supersaturation and attempt to explain some of the differences between the modeled and experimental CCN data reported in the literature. We also compare the experimental data with predictions of the Universal Functional Group Activity Coefficients (UNIFAC) model, which has been used for estimating the water activity of different organic droplets (13, 14). 10.1021/es7023252 CCC: $40.75  2008 American Chemical Society Published on Web 04/16/2008 Experimental Section An EDB was used to levitate an aerosol particle of roughly 20–40 µm in diameter by electric fields. Hygroscopic measurements were performed by equilibrating the levitated particle of interest at different relative humidities for in situ relative mass determination. A detailed description is given in the Supporting Information (S2). For saccharides and the compounds that crystallize, we assume that an anhydrous crystal is formed, and the data are presented in the form of the mass fraction of solute (mfs) as a function of relative humidity (RH) (15, 16). The uncertainties in this study were (0.01 and (0.03 in the mfs for the solution droplets and the solid particles, respectively. For the compounds that did not crystallize, the mass ratio, m/mo in response to the RH change is reported, where m is the aerosol mass at a given RH and mo is the aerosol mass at the reference RH (∼6%RH). The uncertainties in this study were (0.01 in m/mo for the solution droplets and (0.03 for the solid particles. Results and Discussions Deliquescent Organic Compounds. C6-C9 Dicarboxylic Acids. Dicarboxylic acids have been identified as a major group of water-soluble secondary organic compounds in atmospheric particles. Here, we report on hygroscopic measurements of adipic acid (C6), pimelic acid (C7), suberic acid (C8), and azelaic acid (C9) particles. The discussion on UNIFAC predictions is given in the Supporting Information (S3). First, the EDB was equilibrated at 85%RH before introducing the particles. For solution droplets of adipic acid, suberic acid, and azelaic acid, they crystallized immediately when introduced into the EDB and formed solid particles. The levitated particles exhibited irregular light scattering, which did not resemble anything from the Mie scattering pattern of droplets. Furthermore, the balancing levitation voltage of the solid particles did not change with RH. Hence, these droplets have crystallization RH (CRH) larger than 85%RH. Once the solution droplets of adipic acid and azelaic acid crystallized, the solid particles did not deliquesce at RH < 90% (Figure 1a). Solid suberic acid particles were small because it has a low solubility (0.242 g/100 g water). The solubility lowers the initial size of the particles that can be trapped and levitated in the balance after equilibration even at a high RH (e.g., 85%RH). The size of the particles is so small that the particles cannot be levitated stationary in the balance. Our adipic acid measurements were consistent with those of Prenni et al. (17) and Hameri et al. (18), who reported ¨ that there was no detectable increase in the size of solid adipic acid particles at RH < 93% using a hygroscopic tandem differential mobility analyzer (HTDMA). Corresponding to their low solubility, the water activities, aw,sat of saturated adipic acid, suberic acid, and azelaic acid solutions were determined to be 0.998, 0.999, and 0.999, respectively (Table 1).Solutiondropletsofpimelicacidcrystallizedat51.5-53.0%RH, and the resulting solid pimelic acid particles did not deliquesce at RH < 90%. The absence of deliquescence of solid pimelic acid particles below 90%RH was consistent with the high deliquescence RH (DRH) value inferred from the measured aw,sat of the saturated pimelic acid solution (99.5%RH) (Table 1). Table 1 summarizes the hygroscopicity of the C2-C9 diacids determined in this work and in prior studies (11, 19-21). For the deliquescent compounds which crystallized in the EDB experiments but did not deliquesce at RH < 90%, such that the DRH cannot be observed directly, the measured aw in bulk saturated solution experiments (see column in Table 1) was used for determining the DRH. There is no corresponding DRH or aw,sat for the nondeliquescent compounds which do not crystallize in EDB experiments. FIGURE 1. Deliquescent compounds: (a) the hygroscopicity of adipic acid, pimelic acid, and azelaic acid particles; (b) the hygroscopicity of cis-pinonic acid particles. cis-Pinonic Acid. cis-Pinonic acid is a major product of the ozone oxidation of R-Pinene and is a major biogenic secondary compound. Figure 1b shows that the pinonic acid solution droplets crystallized to form solid particles at 33.9-37.1%RH. The solid pinonic acid particles did not deliquesce at RH < 90%. Cruz and Pandis (20) did not observe particle growth at RH less than 95% using HTDMA. The measured aw of the saturated pinonic acid solution in this study was 0.999 (Table 1), which is in good agreement with the aw of the pinonic acid solution with 0.5 wt% which was equal to 0.9993 (22). According to the solubility determined by Hartz et al. (6), the saturation concentration of pinonic acid solution is about 0.64–0.71 wt%. Nondeliquescent Organic Compounds. Monosaccharides (Fructose, Mannose) and Disaccharides (Maltose, Lactose). Monosaccharides (e.g., glucose, fructose, and mannose) and disaccharides (e.g., maltose, lactose, and sucrose) have been detected in atmospheric particles originating from biomass burning (23, 24). We measured the hygroscopicity of two monosaccharides (fructose and mannose) (Figure 2a) and two disaccharides (maltose and lactose) (Figure 2b). The bulk solution data for fructose, maltose, and lactose were obtained from other studies (25–27). As shown in Figure 2a,b, a smooth curve is observed for these four saccharides, suggesting that the aerosol particles absorbed and desorbed water reversibly and existed as liquid at RH as low as ∼6%. These reversible water sorption and desorption characteristics without hysteresis were observed in other saccharides, such as glucose (15) and sucrose (16). Data on glucose particles from Peng et al. (15) were included for comparison (Figure 2a). There is no significant difference in the mfs-aw curves of the different monosaccharides (Figure 2a). The disaccharides also exhibit very similar hygroscopicity VOL. 42, NO. 10, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3603 3604 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 10, 2008 99 99 >98.0 98 98 oxalic acid (C2) malonic acid (C3) succinic acid (C4) glutaric acid (C5) adipic acid (C6) pimelic acid (C7) suberic acid (C8) azelaic acid (C9) cis-pinonic acid 99 >99.0 99 >99.5 >99 >99.5 purity (%) 90.04 104.06 118.09 132.12 146.16 160.17 174.20 188.22 184.23 molecular wt (g/mol) 180.2 180.2 180.2 342.3 342.3 342.3 166.10 178.11 178.11 74.04 DRH (%RH) crystal density (g/cm3) 1.566n 1.600n 1.539n 1.538n 1.543n 1.589n 1.593l 1.41l 1.41l 1.342 90.9n 407.4n 250n 9.3n 20n 200n 0.41l 0.46l 0.4l - N.A.o N.A. N.A. N.A. N.A. N.A.p N.A.l N.A.l N.A. N.A. CRH (%RH) N.A.o N.A. N.A. N.A. N.A. N.A.p N.A.l N.A.l N.A. N.A. DRH (%RH) 51.8–56.7f, 85 51.5–53 >85 >85 33.9–37.1 Nondeliquescent Compounds >90f, 96.8–98g not obs.f, 69–91g >90f 83–85f,k not obs.m not obs.m not obs.m not obs.m not obs.m CRH (%RH) solubility (g/100 g water) 1.90f 1.63f 1.552f 1.429f 1.362l 1.321l 1.272l 1.251l 0.781k, 1.169l 12f 161f 8.8f 116f 2.49l 6.73l 0.242l 0.5l 0.64–0.71l molecular wt (g/mol) crystal density (g/cm3) solubility (g/100 g water) Deliquescent Compounds 0.059 ( 0.006 0.081 ( 0.008 0.182 ( 0.017 0.306 ( 0.029 0.185 ( 0.017 0.180 ( 0.017 0.183 ( 0.017 0.085 ( 0.008 0.082 ( 0.008 Kb (EDB) 97.3f, 97.1i, 97.8j 65.2f, 71.9i, 72.4j 98.8f, 97.6i, 99.1j 88.5f, 88.9i, 88.2j 99.8, 99.9j 99.5 99.9 99.9 99.9 aw,sat (%RH) 0.059 0.081 Kc (TDMA) 0.17 0.17 0.17 0.093 0.093 0.093 0.051 0.094 0.147 Kd (CCN) 0.224 ( 0.017 0.092 ( 0.009 0.587 ( 0.061 0.303 ( 0.029 0.276 ( 0.026 0.176 ( 0.016 Kb (EDB) 0.018 0.106 0.227 0.231 0.195 0.096 Kd (CCN) 0.168 0.171 0.170 0.091 Ke (bulk soln)