Atmospheric Aqueous Aerosol Surface Tensions: Isotherm-Based

Feature Article pubs.acs.org/JPCA

Atmospheric Aqueous Aerosol Surface Tensions: Isotherm-Based Modeling and Biphasic Microfluidic Measurements Hallie C. Boyer and Cari S. Dutcher* Department of Mechanical Engineering, University of Minnesota, Twin Cities, Minneapolis, Minnesota 55455, United States ABSTRACT: Surface properties of atmospheric aerosol particles are crucial for accurate assessments of the fates of liquid particles in the atmosphere. Surface tension directly influences predictions of particle activation to clouds, as well as indirectly acting as a proxy for chemical surface partitioning. Challenges to accounting for surface effects arise from surface tension dependence on solution concentration and the presence of complex aqueous mixtures in aerosols, including both surface-active organic solutes and inorganic electrolytes. Also, the interface itself is varied, in that it may be a liquid−vapor interface, as in the surface of an aerosol particle with ambient air, or a liquid−liquid interface between two immiscible liquids, as in the interior surfaces that exist in multiphase particles. In this Feature Article, we highlight our previous work entailing thermodynamic modeling of liquid−vapor surfaces to predict surface tension and microscopic examinations of liquid−liquid interfacial phenomena to measure interfacial tension using biphasic microscale flows. New results are presented for binary aqueous organic acids and their ternary solutions with ammonium sulfate. Ultimately, improved understanding of aerosol particle surfaces would enhance treatment of aerosol particle-to-cloud activation states and aerosol effects on climate.

■

INTRODUCTION Mass transfer and chemical reactions occur at aerosol interfaces between particles and the ambient environment, supporting growth processes and heterogeneous chemistry. These surfacebased properties and processes govern atmospheric aerosol particle size, morphology, composition, and growth. In aqueous aerosol particles, mass exchanges are regulated by ambient conditions, especially temperature and relative humidity, thereby affecting particle water content and solute concentration. Aqueous atmospheric aerosols have been established as chemically complex microenvironments,1 containing a wide representation of organics, inorganics, neutral molecules, and ions. They can also exhibit multiple separated phases, where solids partition from liquids and the aqueous phase partitions from the organic phase.2 Figure 1 is a schematic of the chemical thermodynamics of atmospheric aqueous aerosol, showing the many possible equilibria for organic and inorganic species present among solid, gas, and one or more liquid phases. Atmospheric aerosols play a significant role in our climate regulation, yet it remains a challenge to precisely characterize their direct and indirect effects. Among atmospheric constituents, the largest uncertainty in net solar radiative forcing is attributed to aerosol particles.3 Due to the heightened presence of pollutants, aerosol particles greatly impact cloud albedo,4−7 which defines their optical properties. Cloud albedo determines the proportion of sunlight that reflects back into space or gets absorbed and scattered, known as the direct effect. Atmospheric aerosols also influence cloud formation, known as the indirect effect, by acting as seed particles or cloud condensation nuclei (CCN).8−16 Through the Kelvin equation,17 surface tension is an important property in determining the activation states of particles. In aqueous aerosol particles, positive and negative © XXXX American Chemical Society

Figure 1. Chemical thermodynamics of aerosol particles showing partitioning of solids, liquids, and gases, as well as water-rich (aqueous) and water-poor (organic) liquids. The aqueous phase is partially engulfed in the organic, although complete engulfing and sideby-side separation also may occur, depending on the interfacial energies between the liquids. Not to scale.

changes in surface tension induced by solutes must be considered.18−23 Water-soluble organics typically depress surface tension,13,7 lowering the critical supersaturation point above which particles activate to cloud droplets. In order to account for solute surface Received: April 4, 2017 Revised: May 11, 2017 Published: May 12, 2017 A

DOI: 10.1021/acs.jpca.7b03189 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Feature Article

The Journal of Physical Chemistry A

Figure 2. Thermodynamic model application of lattice adsorption of gases (BET adsorption isotherm) in solution (steps 1−3) and surfaces (step 4). After step 1, blue circles represent water molecules. Black, green, and orange circles represent solutes. The layers of waters indicate solute hydration, and the disordered molecules are in the bulk of the solution. Steps 0−3 were adapted with permission from refs 25−27. Step 4 was adapted with permission from ref 24. Copyright 2011, 2012, and 2013, American Chemical Society.

Figure 3. Thermodynamic model of the adsorption isotherm at the surface for single solute solutions (step 1), partially dissociating aqueous organic acids (step 2), and mixtures of arbitrary numbers of solutes (step 3). Blue circles represent water molecules. Green and orange circles represent solutes. The waters at the surface are displaced by surface-sorbing molecules or promoted to the surface by bulk-sorbing molecules. Adapted with permission from refs 24, 29, and 30. Copyright 2013, 2016, and 2017, American Chemical Society.

Dutcher32 measured interfacial tensions of atmospherically relevant mixtures using a biphasic microfluidic platform. In the atmosphere, an important and abundant class of particulate matter known as secondary organic aerosols (SOAs) are formed through photochemical reactions of organic material in the atmosphere.33−35 Characterization of SOA formation processes and equilibrium is poorly understood and must be studied further to predict the role of atmospheric aerosols in climate and human health. 33,36 Further, it has been demonstrated that they exhibit internal phase separation,2,31,37 resulting in nonspherical equilibrium morphologies. In this Feature Article, we discuss characterization of aqueous aerosol interfaces of solutions containing SOA chemical mimics. Measurements of liquid−liquid interfacial tension and modeling of liquid−vapor surface tension shown here provide a foundation for future studies of interfaces involving real SOA samples. Our results could inform global aerosol models that predict aerosol particle activation to cloud droplets by providing surface tensions as a function of concentration and temperature and in turn provide indicators of particle morphology.

effects in large-scale aerosol models, surface tensions and surface-bulk partitioning must be considered.22 Comprehensive thermodynamic treatment of aerosols is necessary in order to consider composition-dependent properties in solution and at the interface. In previous work, Wexler and Dutcher24 developed a thermodynamic model of surface tension. The theoretical framework was developed in solution using adsorption isotherms and statistical mechanics to predict water activity and solute activity of multicomponent aqueous solutions.25−27 The surface tension model demonstrated that for solutions containing two components, a solute and a solvent, a single expression gives surface tension as a function of solute activity calculated from the multilayer adsorption isotherm. The model applicability was then extended by parametrizing data sets for single solutes and reducing the parameters through physical interpretation,28 developing a ternary model to treat dissociated organic acids,29 and deriving the model for an arbitrary number of solutes.30 Model capabilities of predicting surface tension for a liquid−vapor interface complement experimental measurements of liquid−liquid interfacial tension. Interfacial tension measurements indicate solute effects at the surface, such as solute propensity to reside at the interface and the level of adhesion between liquids. The internal microphysical structure is described by equilibrium phase partitioning and interfacial tensions. To study the interface of aqueous solutions and immiscible organic phases, Metcalf, Boyer, and

■

COMPUTATIONAL METHODS The surface tension model is based on the adsorption isotherm model developed in solution.25−27,38,39 The theoretical framework is outlined schematically in Figure 2 for the solution thermodynamic model and in Figure 3 for the interface. The B


Feature Article

The Journal of Physical Chemistry A first step in Figure 2 is monolayer lattice adsorption on planar surfaces that originated with Langmuir.40 Brunauer−Emmett− Teller (BET) theory added a second monolayer, or multilayer, and allowed further numbers of layers with the same associated energy.41 The use of adsorption isotherms in solution was initiated by Stokes and Robinson42 for concentrated electrolyte aqueous solutions by applying BET theory. In solution, solvent molecules were assigned as adsorbates and solutes were adsorbents. Ally and Braunstein43 extended the Stokes and Robinson work to accommodate ternary solutions. A modification was presented by Guggenheim, Anderson, and de Boer, known as the GAB isotherm, in which a second layer was added.44−46 In the adsorption isotherm developed by Dutcher, Ge, Wexler, and Clegg,25 a unique energy was assigned to each layer and the number of layers was determined for each individual solute. Expressions for Gibbs free energy were derived using statistical mechanics and subsequently were used to calculate water activity and solute activity. The model successfully captured activities and osmotic coefficients across the entire concentration range by setting a reference point at pure solute. The model also successfully extended to multicomponent solutions (step 2, Figure 2)27 and included long-range Coulombic interactions to reduce parameters38 and partial dissociation of organic acids39 (step 3, Figure 2). The adsorption isotherm used in solution was then introduced to the surface by Wexler and Dutcher24 to derive a model of surface tension (step 4, Figure 2). There were three key distinctions between the solution property model and the surface tension model: (1) the roles of adsorbates and adsorbents were reversed, making solvent molecules the adsorbents and solutes the adsorbates; (2) because the solutes were allowed to adsorb at the surface in the new framework, the monolayer (surface) was separate from the multilayer (bulk); and (3) only a single layer was needed at the surface. The monolayer represents either the solutes that have adsorbed at the surface, thereby removing solvents (water molecules in the case of aqueous solutions), or solutes that have remained in the bulk and in turn have promoted a number of solvents to the surface. The Gibbs free energy and entropy were found by using statistical mechanics, where a surface partition function accounts for the available configuration states of the surface and a bulk partition function accounts for solute mixing between the surface and bulk phases. An expression for single-solute or binary aqueous solutions was derived from multilayer adsorption isotherms by Wexler and Dutcher as a function of solute concentration represented by activity (as)24 σ = σw +

⎞ 1 − Kas kT ⎛ ln⎜ ⎟ rSw ⎝ 1 − Kas(1 − C) ⎠

Equation 1 was applied to surface tension data of several binary aqueous solutions for classes of water-soluble electrolytes and organics.24 For organics that significantly lower surface tension relative to pure water, K became very small (in some cases on the order of 10−10), causing C to become very large. While these parameters were valid, they presented an opportunity to simplify the equation and eliminate one parameter, which resulted in the expression σ = σw −

kT ln(1 + K ′as) rSw

(2)

where the valid range of K was increased with its new form, K′. Equation 2 reduced another parameter by using the pure solute surface tension, σs, which is known for many liquid organics. Equation 2 shows that this framework reduces to a form similar to the Szyszkowski equation,47 which uses Langmuir adsorption theory to handle binary aqueous organics. In Boyer et al.,28 the binary model (Equation 1 and 2) was applied to an array of solutes. Parameters found from the binary model treatments correlated strongly with solute properties, which were used to reduce the number parameters down to zero. In the next step, the binary model was extended to treat partially dissociated organic acids, toward a generalized model for multicomponent solutions.29 Dissociation of organic acids in water is symbolically shown as HA ⇆ H+A−, where HA is the neutral, nondissociated form of the acid and H+A− is the deprotonated, dissociated form of the acid. In the model derivation, a single model parameter, K, was decoupled into KHA ′ and KH+A− and parameters r and σs remained the same. The dissociation model is σ = σw −

C + −K + −a + − ⎞ kT ⎛ ′ aHA + H A H A H A ⎟ ln⎜1 + KHA r ̅Sw ⎝ (1 − K H+A−a H+A−) ⎠ (3)

where aHA and aH+A− are the activities of the neutral and dissociated species, which were found by Raoult’s Law and previously identified dissociation constants.48−50 To avoid additional fitting, σs and r ̅ were determined in the binary cases using eq 2 and CH+A− was recast in the limit of pure solute to become a function of σs and r ̅ and, therefore, no longer an independent parameter. Toward a multicomponent surface tension model applicable to an arbitrary number of solutes, a ternary model was derived by decoupling all model parameters.30 The values from the binary model were applied to the new model and therefore required no parametrizations for the solutions with two solutes. Letting NWS, NAS, and NBS represent the number of waters, A molecules, and B molecules on the surface, respectively, and χ = NWS − rANAS − rBNBS, the expression for ternary aqueous solution surface tension became

(1)

where σw is the surface tension of pure water (71.98 mN/m at 298 K), k is Boltzmann’s constant, T is temperature, and Sw is the projected area of a water molecule (0.1 nm2). The remaining quantities are model parameters, defined in the statistical mechanical derivation: r is a multifactorial skip from the surface partition function and represents the average number of water molecules that a solute displaces (either negatively or positively) from the surface; K is a bulk energy parameter related to molecular energies of solutes in the bulk; and C is an energy parameter that represents equilibrium partitioning of solutes between the surface and the bulk.

σ = σW + +

⎡ χ (NWS − rANAS) ⎤ kT ln⎢ ⎥ 2rASw ⎣ NWS(NWS − rBNBS) ⎦

⎡ χ (NWS − rBNBS) ⎤ kT ln⎢ ⎥ 2rBSw ⎣ NWS(NWS − rANAS) ⎦

(4)

Also, parameter r was decoupled into rA and rB to treat the adsorption behavior of the two solutes independently. The solute pairs consisted of electrolytes, organics, and both an electrolyte and an organic. It was demonstrated that the model worked well for all types of solute combinations. In one case, C


Feature Article


surrounding continuous phase. Downstream, the droplets are unconfined and carried through a transit channel with constrictions that induce an extensional flow field and cause the droplet to deform until they are confined by the channel walls. The deformation due to the extensional flow is highly dependent on interfacial forces and the fluid viscosities. If viscosities are known, this setup enables measurements of interfacial tension. Droplet deformation (D) is quantified using minor (rminor) and major radii (rmajor) at different locations down the channel (x) provided by G. I. Taylor54 and J. M. Rallison55

model predictions agreed excellently with surface tension measurements of supermicron droplets obtained from optical tweezers, which trap picoliter particles and balance them with radiative pressure and gravity. Model predictions and optical tweezers data agreed for binary and ternary solutions at dilute and supersaturated concentrations.30 In Figure 4,30 surface tension model predictions for several data sets compared to measured values are shown. There is in excellent agreement at all concentrations for a breadth of solutes and mixtures, including binary and ternary solutions.

D(x) =

rmajor(x) − rminor(x) rmajor(x) + rminor(x)

(5) 52

Using the approach of Hudson et al. and Cabral and Hudson,53 interfacial tension, γ, is a parameter in the equation ⎛ D(x) ⎞ ⎛ 5 ∂D(x) ⎞ αηc⎜ ε(̇ x) − u(x) ⎟ ⎟ = γ⎜ ∂x ⎠ ⎝ 2η ̂ + 3 ⎝ a0 ⎠

(6)

where the other quantities are all known: ηc is the continuousphase viscosity, η̂ is the ratio of the dispersed-phase viscosity, ηd, to ηc, u(x) is the velocity of the droplet in the streamwise ∂u(x) direction, ε̇(x) is the extensional flow field ε̇(x) = ∂x , a0 is the radius of the droplet prior to deformation, and α is a (2η ̂ + 3)(19η ̂ + 16) . Interfacial tension is measured coefficient, α = 40(η ̂ + 1) by taking the left-hand side of eq 6 as the y-axis and the righthand side as the slope, γ, times the x-axis, D(x) . Exemplary data

Figure 4. Surface tension model predictions for binary and ternary solutions that were not represented in the preceding figures. Data sources: methanol, ref 59; formic acid, ref 60; citric acid, ref 29; ethanol + glycerol, ref 61; NaCl + KCl, ref 62; NaCl + succinic acid, ref 63; and NH4NO3 + (NH4)2SO4, ref 30. Adapted with permission from ref 30. Copyright 2017, American Chemical Society.

a0

for interfacial tension measurements of several aqueous solutions are provided in Figure 8,32 referred to as a Taylor plot, found through eq 6.

■

RESULTS AND DISCUSSION For electrolytes, eq 1 was applied to chlorides, nitrates, and sulfates. Estimates for hypothetical values of pure solute surface tension, σs, were found by using known surface tensions of electrolyte melts at high temperatures and linearly extrapolating down to 298 K.56 Certain electrolytes do not transition to liquids at high temperatures; for each of these, the slope and intercept for extrapolation could not be determined directly from experimental data for the individual molten salts. The slope and intercepts were rather found from empirical expressions in Dutcher et al.,56 which were based on correlations among ion valence, melting point, molar volume, and cation radius. Among the electrolytes that do not transition to liquids at high temperatures were NH4Cl, which sublimates at 338 °C, and (NH4)2SO4, which thermally decomposes at 250.15 °C. Their values for σs are thus found from the generalized expressions in Dutcher et al.56 Using physically based estimates for σs allowed reasonable predictions for concentrations beyond the solubility limit, which is the supersaturated range in which deliquesced atmospheric particles often exist. As an example, the solution surface tension for aqueous ammonium nitrate at the maximum concentration, 12.00 mol/kg, is σ = 82.00 mN/m, and the proposed pure solute limit is σs = 117.6 mN/m. Parameter K was found to have a global value of 0.99 that worked well for all electrolytes. The last parameter, C, had a strong correlation with partition coefficients, KP, calculated by Pegram and Record.57 A regression equation was found, C = (2.878 ×

■

EXPERIMENTAL METHODS Biphasic microfluidics is an advantageous technique for measuring surface or interfacial tension between two immiscible liquids because surface tension dominates gravitational and inertial forces at the microscale. Using microfluidic biphasic flows, we explored liquid-liquid interfaces by measuring interfacial tension to elucidate interfacial activity of SOA particle mimics.32 Knowledge of interfacial tensions can facilitate predictions of chemical equilibrium morphology through calculations of spreading coefficients.31 Our results could therefore apply to aerosol systems where organic lenses or shells cause deviations from spherical shapes that are challenging to observe in situ. In Metcalf et al.,32 solute mixtures included organic species methylglyoxal (MG) and formaldehyde (F), plus an inorganic electrolyte, ammonium sulfate (AS). Aqueous solutions containing these water-soluble compounds have been shown to evolve chemically through selfreactions.51 In a new study presented in this Feature Article for the first time, the aqueous solutions included binary dicarboxylic acids (succinic, malonic, maleic, and malic) and their ternary solutions with AS. Techniques were adapted from Hudson et al.52 and Cabral and Hudson53 to measure interfacial tension in microfluidic devices using biphasic flows. A microgeometry was designed to allow fluid flow of two intersecting fluids, where one fluid becomes the inner dispersed phase and the other becomes the D


Feature Article


Figure 5. Surface tension as a function of solute activity for ammonium electrolytes (a) and alcohols (b). Blue lines are model predictions using a single parameter, and black lines use zero model parameters. For electrolytes shown in (a), hypothetical estimates for pure solute surface tension, σs,56 and a global value of K were used to reduce the number of parameters down to one, resulting in the blue line; a regression equation between the final parameter, C, and partition coefficients from Pegram and Record,57 KP, (C = 2.787 × 104) resulted in the black lines. For organics shown in (b), there is a single parameter needed because σs is known for alcohols and C is not necessary in eq 2; K′ is eliminated through the relationship with molar volume, v, (K′ = e0.067v +1). Reproduced with permission from ref 28. Copyright 2015, American Chemical Society.

Figure 6. Surface tension versus mole fraction for dicarboxylic acids. The acids in (a) and (b) are shown in separate plots for clarity. Black solid lines are the binary model in eq 2 for organics, using two parameters, K′ and σs. Green lines use one parameter in eq 2, following the elimination of K′ through the regression curve with molar volume, v, (K′ = 7.0821v − 278.65). Black dotted lines are eq 3 with inputs for dissociated concentrations, requiring no further data parametrization. Reproduced with permission from ref 29. Copyright 2016, American Chemical Society.

104)e−14.0KP, thus eliminating all parameters from eq 1 for the electrolytes in the study. Equation 1 was reduced to the following expression:

( + (σ − σ ) ln( ln

σ = σw

s

w

) )

2, one of them, r, was replaced with σs, leaving a single remaining parameter, K′. For dicarboxylic acids studied (e.g., succinic acid, glutaric acid), which are solid in the absence of water, σs was allowed to vary as a parmater because they do not have a surface tension at the relevant temperatures. A strong correlation between K′ and the organic molar volume was found for all organic species studied, and regression curves were used to eliminate K′ as a parameter.28,29 Three regressions were needed as functions of molar volume, demonstrating a dependence on the number of functional groups: (1) exponential for alcohols and polyols, K′ = e0.067v + 1, (2) linear for dicarboxylic acids, K′ = 7.0812v − 278.65, and (3) exponential for carboxylic acids, K′ = 0.0086e0.197v. For each

1 − Kas 1 − Kas(1 − C) 1−K 1 − K (1 − C)

(7)

For organics, eq 2 was applied to two classes of alcohols28 and organic acids.29 For the alcohols (e.g., methanol, ethanol), polyols (e.g., 1,2-ethanediol, 1,3-propanediol), and carboxylic acids (e.g., formic, propionic), the surface tension of the pure component, σs, was already known. With two parameters in eq E


Feature Article

The Journal of Physical Chemistry A binary system, model predictions using zero parameters agree well with those using one or more parameters. Equation 2 was changed to the following expression: σ = σw − (σw − σs)

ln(1 + K ′as) ln(1 + K ′)

(8)

Figure 528 shows surface tension predictions as a function of solute activity. For ammonium electrolytes (panel a), oneparameter model predictions are made with eq 1 and zeroparameter predictions are made with eq 7. For a sample of alcohols and polyols (panel b), single-parameter model predictions are made with eq 2 and parameter-free model predictions are made with eq 8. There are mostly excellent agreements between the single- and zero-parameter model treatments. Discrepancies between the fits can be attributed to how close the parameter values were to the regression curves with the partition coefficients for electrolytes and molar volume for organics. For organic acids, because KH+A− represented the bulk energy parameter for the dissociating acid species, similar to an electrolyte, it was assigned the global value of 0.99, the same value used for binary electrolytes. K′HA was the single remaining parameter; however, estimates were available using a slight modification to the molar volume relationship from the binary case (eq 19 in Boyer and Dutcher29). Figure 629 shows several dicarboxylic acids’ surface tension predictions using the twoparameter and single-parameter binary model and the zeroparameter dissociation model. The multicomponent model was successfully applied to ternary mixtures containing water and two electrolytes, two organics, or one of each. By using parameters determined in the binary cases, no additional parameters were required for the multicomponent solutions. Further, a model was derived for an arbitrary number of solutes so that the model could be extended to include any combination and number of chemical species (see the Supporting Information for Boyer et al.30). Figure 730 shows model predictions for mixtures containing NaCl and glutaric acid, as well as predictions for each of the binary solutions. Data are provided by optical tweezers experiments performed by coauthors Bzdek and Reid outlined in Boyer et al.30 Interfacial tensions of aqueous solutions were measured with respect to silicone oil, used here as a proxy for liquid−liquid phase partitioning between water-rich and water-poor phases. In Metcalf et al.,32 the aqueous phases contained mixtures of methylglyoxal (MG), formaldehyde (F), and ammonium sulfate (AS). Aqueous MG lowered interfacial tension as the concentration increased, agreeing with bulk measurements of liquid−vapor surface tension.58 Figure 932 shows the interfacial tensions of 3.1 M AS containing MG and a few points with added F. The interfacial tensions are relative to the value for 3.1 M AS. All measurements were taken 24 h after solution preparation. Interfacial tension depression was observed with increasing MG concentration. F increased the interfacial tensions, but the effect was small. In Figure 10,32 measurements of MG in 3.1 M AS at certain times show no changes in interfacial tensions (panel a), while the addition of F inhibits interfacial tension depression (panel b). With the total organic content held constant, the ratio was varied, and when the F content dominated, interfacial tension increased over time. Changes in the interface due to aging of the mixtures were reflected here, suggesting that aerosol

Figure 7. Surface tension versus molality for binary NaCl (blue), binary glutaric acid (green), and a mixture with 1:1 mass ratio ternary solutions (black). Empty symbols are bulk measurements, and filled symbols are experimental measurements taken with optical tweezers by coauthors of Boyer et al.30 Lines are model predictions; the two binary cases are eq 1, and the mixtures are eq 4. Reproduced with permission from ref 30. Copyright 2017, American Chemical Society.

Figure 8. Taylor plot showing interfacial tensions of 3.1 M AS, water, air, and 0.4 M MG. The y-axis is the left-hand side of eq 6. Interfacial tension is the slope of the y-axis and x-axis. Reproduced with permission from ref 32. Copyright 2016, American Chemical Society.

particles tend to have a partially engulfed morphology, as determined by spreading coefficients.31 Aqueous organic acid interfacial tensions were measured for both binary solutions and ternary 3.1 M AS. In Figure 11, new data are reported showing measurements for malonic acid (3C), succinic acid (4C), and glutaric acid (5C) as 1.0 M pure solutions (panel a); measurements for ternary solutions with AS are also reported for organic concentrations of 4 and 20 mM (panel b). For the ternary solutions, the 4C organic acid is maleic acid instead of succinic acid. Succinic acid was available as a stock solution at a concentration of 1.0 M and thereby diluted mixtures when added. Data are reported in both plots where the surface is liquid−air (circles) and liquid−liquid (triangles). The circles are data taken with a pendant drop F


Feature Article


liquid interfacial tension are known, information about the particle equilibrium morphology, optical properties, and heterogeneous chemistry can be inferred. Microfluidic platforms provide unique advantages for studying fluid−fluid interfaces. For example, at the microscale, interfacial forces are more significant than gravitational or inertial forces. Also, lower sample volumes are necessary in the microfluidic realm than bulk analogues. In Metcalf, Boyer, and Dutcher,32 aged organic aerosol mimics were studied. Previously unpublished data for aqueous organic acids in pure solution and salty solutions are also presented in this Feature Article. Interfacial tension measurements were taken using biphasic flows through microchannels, in which droplet deformation depended on restoring forces dictated by interfacial stability. These measurements were complemented by surface tension modeling, and an example model fit was applied to MG + AS data in Figure 9.32 An isotherm-based surface tension model, derived using statistical mechanics of adsorption at the solution interface, was developed by Wexler and Dutcher24 as a function of solute concentration for binary (single-electrolyte or organic solute) aqueous solutions. Although model parameters were derived empirically, solute size and surface propensity were implicit in the theoretical framework, facilitating physical interpretation and reduction of these parameters.28 The model was extended to partially dissociating aqueous organic acids (two solutes assumed to be equal in size and opposite in surface propensity) by Boyer and Dutcher.29 The model was then extended to multicomponent (at least two solutes) aqueous solutions by Boyer et al.30 and did not require any new model parameters. Where data were available for comparison, excellent agreement between model predictions and measurements were observed; in the absence of data, surface tension predictions were made for multicomponent aqueous solutions. Our work in this Feature Article utilizes model predictions, experimental measurements, and chemical mimics that can ultimately be applied toward improving the accuracy of predictions of CCN activity. The thermodynamic model framework could be adapted in many ways, including temperature variations or large surfactant molecules. Model development was established at 25 °C, although temperatures relevant to the atmosphere tend to be lower. Low molecular

Figure 9. Interfacial tension measurements (γ) relative to pure water (γw) with respect to MG and F concentrations. The organics were added to a 3.1 M AS solution. The solutions were allowed to sit for ∼24 h before measurements were taken. Error bars are for statistical uncertainty over hundreds of counts, as shown in the inset plot. The solid line is a form of eq 2 normalized by the measured interfacial tension of water and silicone oil with parameters (r = 5.555 and K′ = 48.50). Reproduced with permission from ref 32. Copyright 2016, American Chemical Society.

setup (Krüss Drop Shape Analysis 4), and the triangles are microfluidic measurements. Due to the lower relative values given by the liquid−liquid interfacial tensions, there could be a direct effect that the organic acids may have on the interface that is pronounced by the presence of the oil. The organic acids thus may indirectly influence the shape of the particle.

■

SUMMARY AND FUTURE DIRECTIONS This Feature Article highlights and connects recent work in thermodynamic modeling of the surface24,28−30 and biphasic microfluidic experiments,32 where each project aims to characterize the interface/surface tensions of aqueous atmospheric aerosols. If liquid−vapor surface tension and liquid−

Figure 10. Interfacial tensions of solutions of aqueous AS and (a) MG and (b) MG and F. Reproduced with permission from ref 32. Copyright 2016, American Chemical Society. G


Feature Article


Figure 11. Relative interfacial tensions of aqueous solutions consisting of (a) binary organic acids, where γwater is the interfacial tension of water with respect to silicone oil (triangles) or air (circles), and (b) ternary organic acids with 3.1 M AS, where γ3.1 M AS is the interfacial tension with respect to silicone oil (triangles) or air (circles). In (a), the organics are malonic acid (3C), succinic acid (4C), and glutaric acid (5C). In (b), the organics are malonic acid (3C), maleic acid (4C), and glutaric acid (5C).

weight surfactants, such as alcohols and polyols, were treated in our work; however, larger surfactants can potentially be treated with a competitive adsorption framework.30 Incorporating lower temperatures and surfactants would benefit model application to CCN activity. Microfluidic tensiometry could provide insight into the surface propensity of SOA samples. On the basis of prior success of modeling and measurements described in this Feature Article, there is great potential to expand the model further and to utilize the microfluidic platform to measure surface tensions of ambient aerosols.

■

are pronounced. She received her B.A. in physics from Macalester College in St. Paul, MN. She is an NSF Graduate Fellowship recipient.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: 612-624-0428. ORCID

Cari S. Dutcher: 0000-0003-4325-9197

Cari Dutcher is the Benjamin Mayhugh Assistant Professor of Mechanical Engineering at the University of Minnesota, Twin Cities, with a graduate faculty appointment in the Department of Chemical Engineering and Materials Science. Her research interests are in dynamics of complex and multiphase flows, including chemical thermodynamics of atmospheric aerosol particles. Prior to her faculty position, Dutcher was an NSF-AGS Postdoctoral Research Fellow in the Air Quality Research Center and the Department of Chemical Engineering and Materials Science at the University of California, Davis. Dutcher received her B.S. from Illinois Institute of Technology (2004) and her Ph.D. from the University of California, Berkeley (2009), both in Chemical Engineering.

Notes

The authors declare no competing financial interest. Biographies

■

ACKNOWLEDGMENTS This material is based upon work supported by the National Science Foundation under Grant No. 1554936. We gratefully acknowledge Prof. Gordon Christopher for sharing microfluidic device design files to aid in the startup of the tensiometry work and Prof. Kevin Dorfman for allowing use of his laboratory and equipment for fabrication of the PDMS microfluidic devices. We also thank and acknowledge our previous coauthors: Dr. Andrew Metcalf for microfluidic interfacial tension measure-

Hallie Boyer is a Ph.D. candidate at the University of Minnesota in Mechanical Engineering. Her doctoral research is on atmospheric aerosols and aqueous interfaces. She works on thermodynamic modeling and experiments at the microscale, where interfacial effects H


Feature Article


Organic Compounds in Nucleating Cloud Droplets. Geophys. Res. Lett. 1996, 23 (3), 277−280. (14) Sun, J.; Ariya, P. Atmospheric Organic and Bio-Aerosols as Cloud Condensation Nuclei (CCN): A Review. Atmos. Environ. 2006, 40 (5), 795−820. (15) Topping, D. O.; Mcfiggans, G. B.; Coe, H. A Curved MultiComponent Aerosol Hygroscopicity Model Framework: Part 1 − Inorganic Compounds. Atmos. Chem. Phys. 2005, 5, 1205−1222. (16) Varga, Z.; Kiss, G.; Hansson, H. C. Modelling the Cloud Condensation Nucleus Activity of Organic Acids. Atmos. Chem. Phys. Discuss. 2007, 7 (2), 5341−5364. (17) Thomson, S. W. On the Equilibrium of Vapour at a Curved Surface of Liquid. Philos. Mag. 1871, 4, 448−452. (18) Hansen, A. M. K.; Hong, J.; Raatikainen, T.; Kristensen, K.; Ylisirnio, A.; Virtanen, A.; Petaja, T.; Glasius, M.; Prisle, N. L. Hygroscopic Properties and Cloud Condensation Nuclei Activation of Limonene-Derived Organosulfates and Their Mixtures with Ammonium Sulfate. Atmos. Chem. Phys. 2015, 15 (24), 14071−14089. (19) Hritz, A. D.; Raymond, T. M.; Dutcher, D. D. A Method for the Direct Measurement of Surface Tension of Collected Atmospherically Relevant Aerosol Particles Using Atomic Force Microscopy. Atmos. Chem. Phys. 2016, 16 (15), 9761−9769. (20) Anttila, T.; Kerminen, V.-M. Influence of Organic Compounds on the Cloud Droplet Activation: A Model Investigation Considering the Volatility, Water Solubility, and Surface Activity of Organic Matter. J. Geophys. Res. 2002, 107 (D22), 1−13. (21) Ruehl, C. R.; Davies, J. F.; Wilson, K. An Interfacial Mechanism for Cloud Droplet Formation on Organic Aerosols. Science (Washington, DC, U. S.) 2016, 351 (6280), 1447−1450. (22) Prisle, N. L.; Asmi, A.; Topping, D.; Partanen, A. I.; Romakkaniemi, S.; Dal Maso, M.; Kulmala, M.; Laaksonen, A.; Lehtinen, K. E. J.; McFiggans, G.; et al. Surfactant Effects in Global Simulations of Cloud Droplet Activation. Geophys. Res. Lett. 2012, 39 (5), 1−5. (23) Prisle, N. L.; Raatikainen, T.; Laaksonen, A.; Bilde, M. Surfactants in Cloud Droplet Activation: Mixed Organic-Inorganic Particles. Atmos. Chem. Phys. 2010, 10 (12), 5663−5683. (24) Wexler, A. S.; Dutcher, C. S. Statistical Mechanics of Multilayer Sorption: Surface Tension. J. Phys. Chem. Lett. 2013, 4 (10), 1723− 1726. (25) Dutcher, C. S.; Ge, X.; Wexler, A. S.; Clegg, S. L. Statistical Mechanics of Multilayer Sorption: Extension of the Brunauer-EmmettTeller (BET) and Guggenheim-Anderson-de Boer (GAB) Adsorption Isotherms. J. Phys. Chem. C 2011, 115 (33), 16474−16487. (26) Dutcher, C. S.; Ge, X.; Wexler, A. S.; Clegg, S. L. Statistical Mechanics of Multilayer Sorption: 2. Systems Containing Multiple Solutes. J. Phys. Chem. C 2012, 116 (2), 1850−1864. (27) Dutcher, C. S.; Ge, X.; Wexler, A. S.; Clegg, S. L. An IsothermBased Thermodynamic Model of Multicomponent Aqueous Solutions, Applicable over the Entire Concentration Range. J. Phys. Chem. A 2013, 117 (15), 3198−3213. (28) Boyer, H.; Wexler, A.; Dutcher, C. S. Parameter Interpretation and Reduction for a Unified Statistical Mechanical Surface Tension Model. J. Phys. Chem. Lett. 2015, 6 (17), 3384−3389. (29) Boyer, H. C.; Dutcher, C. S. Statistical Thermodynamic Model for Surface Tension of Aqueous Organic Acids with Consideration of Partial Dissociation. J. Phys. Chem. A 2016, 120 (25), 4368−4375. (30) Boyer, H. C.; Bzdek, B. R.; Reid, J. P.; Dutcher, C. S. Statistical Thermodynamic Model for Surface Tension of Organic and Inorganic Aqueous Mixtures. J. Phys. Chem. A 2017, 121, 198−205. (31) Kwamena, N.-O. A.; Buajarern, J.; Reid, J. P. Equilibrium Morphology of Mixed organic/Inorganic/Aqueous Aerosol Droplets: Investigating the Effect of Relative Humidity and Surfactants. J. Phys. Chem. A 2010, 114 (18), 5787−5795. (32) Metcalf, A. R.; Boyer, H. C.; Dutcher, C. S. Interfacial Tensions of Aged Organic Aerosol Particle Mimics Using a Biphasic Microfluidic Platform. Environ. Sci. Technol. 2016, 50 (3), 1251−1259. (33) Zhang, Q.; Jimenez, J. L.; Canagaratna, M. R.; Allan, J. D.; Coe, H.; Ulbrich, I.; Alfarra, M. R.; Takami, A.; Middlebrook, A. M.; Sun, Y.

ments; Prof. Anthony Wexler for the genesis of the surface tension model; and Prof. Jonathan P. Reid and Dr. Bryan R. Bzdek at the University of Bristol for their optical tweezers data, whose work was supported by the Engineering and Physical Sciences Research Council 490 (EPSRC) through Grant EP/ L010569/1. Donald Hall is acknowledged for some of the optical tweezers measurements of supersaturated sodium chloride surface tension. Graphic designer Kiley Schmidt is acknowledge for the table of contents and cover art. Part of this work was carried out in the College of Science and Engineering Minnesota Nano Center, University of Minnesota, which receives partial support from NSF through the NNIN program. Part of this work was also carried out in the College of Science and Engineering Coating Process and Visualization Laboratory and the Polymer Characterization Facility, University of Minnesota, which have received capital equipment funding from the NSF through the UMN MRSEC under Award DMR1420013. The authors acknowledge funding support for H.C.B. through a National Science Foundation Graduate Research Fellowship through NSF Grant No. 00039202.

■

REFERENCES

(1) George, I. J.; Abbatt, J. P. D. Heterogeneous Oxidation of Atmospheric Aerosol Particles by Gas-Phase Radicals. Nat. Chem. 2010, 2 (9), 713−722. (2) You, Y.; Renbaum-Wolff, L.; Carreras-Sospedra, M.; Hanna, S. J.; Hiranuma, N.; Kamal, S.; Smith, M. L.; Zhang, X.; Weber, R. J.; Shilling, J. E.; et al. Images Reveal That Atmospheric Particles Can Undergo Liquid-Liquid Phase Separations. Proc. Natl. Acad. Sci. U. S. A. 2012, 109 (33), 13188−13193. (3) IPCC, 2013: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change; Stocker, T. F., Qin, D., Plattner, G. K., Tignor, M., Allen, S. K., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P. M., Eds.; Cambridge University Press: Cambridge, U.K.; New York, 2013. (4) Twomey, S. Pollution and the Planetary Albedo. Atmos. Environ. 2007, 41, 120−125. (5) O’Dowd, C. D.; Facchini, M. C.; Cavalli, F.; Ceburnis, D.; Mircea, M.; Decesari, S.; Fuzzi, S.; Yoon, Y. J.; Putaud, J.-P. Biogenically Driven Organic Contribution to Marine Aerosol. Nature 2004, 431 (7009), 676−680. (6) Meskhidze, N.; Nenes, A. Phytoplankton and Cloudiness in the Southern Ocean. Science (Washington, DC, U. S.) 2006, 314, 1419− 1423. (7) Facchini, M. C.; Mircea, M.; Fuzzi, S.; Charlson, R. J. Cloud Albedo Enhancement by Surface-Active Organic Solutes in Growing Droplets. Nature 1999, 401 (6750), 257−259. (8) Booth, A. M.; Topping, D. O.; McFiggans, G. B.; Percival, C. J. Surface Tension of Mixed Inorganic and Dicarboxylic Acid Aqueous Solutions at 298.15 K and Their Importance for Cloud Activation Predictions. Phys. Chem. Chem. Phys. 2009, 11 (36), 8021−8028. (9) Cruz, C. N.; Pandis, S. N. The Effect of Organic Coatings on the Cloud Condensation Nuclei Activation of Inorganic Atmospheric Aerosol. J. Geophys. Res. 1998, 103 (D11), 13111−13123. (10) Farmer, D. K.; Cappa, C. D.; Kreidenweis, S. M. Atmospheric Processes and Their Controlling Influence on Cloud Condensation Nuclei Activity. Chem. Rev. 2015, 115 (10), 4199−4217. (11) Frosch, M.; Prisle, N. L.; Bilde, M.; Varga, Z.; Kiss, G. Joint Effect of Organic Acids and Inorganic Salts on Cloud Droplet Activation. Atmos. Chem. Phys. 2011, 11 (8), 3895−3911. (12) Novakov, T.; Penner, J. E. Large Contribution of Organic Aerosols to Cloud-Condensation-Nuclei Concentrations. Nature 1993, 365 (6449), 823−826. (13) Shulman, M. L.; Jacobson, M. C.; Carlson, R. J.; Synovec, R. E.; Young, T. E. Dissolution Behavior and Surface Tension Effects of I


Feature Article

The Journal of Physical Chemistry A L.; et al. Ubiquity and Dominance of Oxygenated Species in Organic Aerosols in Anthropogenically-Influenced Northern Hemisphere Midlatitudes. Geophys. Res. Lett. 2007, 34 (13), 1−6. (34) Jimenez, A. J. L.; Canagaratna, M. R.; Donahue, N. M.; Prevot, A. S. H.; Zhang, Q.; Decarlo, P. F.; Allan, J. D.; Coe, H.; Ng, N. L.; Aiken, a C.; et al. Evolution of Organic Aerosols in the Atmosphere. Science (Washington, DC, U. S.) 2009, 326 (5959), 1525−1529. (35) Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics: From Air Pollution to Climate Change, 2nd ed.; John Wiley & Sons, Inc.: New York, 2006. (36) Hallquist, M.; Wenger, J. C.; Baltensperger, U.; Rudich, Y.; Simpson, D.; Claeys, M.; Dommen, J.; Donahue, N. M.; George, C.; Goldstein, a H.; et al. The Formation, Properties and Impact of Secondary Organic Aerosol: Current and Emerging Issues. Atmos. Chem. Phys. 2009, 9 (14), 5155−5236. (37) O’Brien, R. E.; Wang, B.; Kelly, S. T.; Lundt, N.; You, Y.; Bertram, A. K.; Leone, S. R.; Laskin, A.; Gilles, M. K. Liquid−Liquid Phase Separation in Aerosol Particles: Imaging at the Nanometer Scale. Environ. Sci. Technol. 2015, 49 (8), 4995−5002. (38) Ohm, P. B.; Asato, C.; Wexler, A. S.; Dutcher, C. S. IsothermBased Thermodynamic Model for Electrolyte and Nonelectrolyte Solutions Incorporating Long- and Short-Range Electrostatic Interactions. J. Phys. Chem. A 2015, 119 (13), 3244−3252. (39) Nandy, L.; Ohm, P. B.; Dutcher, C. S. Isotherm-Based Thermodynamic Models for Solute Activities of Organic Acids with Consideration of Partial Dissociation. J. Phys. Chem. A 2016, 120 (24), 4147−4154. (40) Langmuir, I. The Adsorption of Gases on Plane Surfaces of Glass, Mica and Platinum. J. Am. Chem. Soc. 1918, 40 (9), 1361−1403. (41) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60 (2), 309−319. (42) Stokes, R. H.; Robinson, R. A. Ionic Hydration and Activity in Electrolyte Solutions. J. Am. Chem. Soc. 1948, 70 (5), 1870−1878. (43) Ally, M. R.; Braunstein, J. Statistical Mechanics of Multilayer Adsorption: Electrolyte and Water Activities in Concentrated Solutions. J. Chem. Thermodyn. 1998, 30 (1), 49−58. (44) Guggenheim, E. A. Applications of Statistical Mechanics; Clarendon Press: Oxford, U.K., 1966. (45) Anderson, R. B.; Hall, W. K. Modifications of the Brunauer, Emmett and Teller Equation. J. Am. Chem. Soc. 1948, 70 (5), 1727− 1734. (46) De Boer, J. H. The Dynamical Character of Adsorption; Clarendon Press: Oxford, U.K., 1968. (47) Szyszkowski, v. B. Experimentelle Studien Uber Kapillare Eigenschaften Der Wassrigen Losungen von Fettsauren. Z. Phys. Chem. 1908, 64, 385. (48) Clegg, S. L.; Seinfeld, J. H. Thermodynamic Models of Aqueous Solutions Containing Inorganic Electrolytes and Dicarboxylic Acids at 298.15 K. 2. Systems Including Dissociation Equilibria. J. Phys. Chem. A 2006, 110 (17), 5718−5734. (49) Perrin, D. D. Dissociation Constants of Organic Bases in Aqueous Solution; Butterworths: London, 1965. (50) Lide, D. R. CRC Handbook of Chemistry and Physics, 2003−2004, 84th ed.; 2003; Vol. 53. (51) Li, Z.; Schwier, A. N.; Sareen, N.; McNeill, V. F. Reactive Processing of Formaldehyde and Acetaldehyde in Aqueous Aerosol Mimics: Surface Tension Depression and Secondary Organic Products. Atmos. Chem. Phys. 2011, 11 (22), 11617−11629. (52) Hudson, S. D.; Cabral, J. T.; Goodrum, W. J.; Beers, K. L.; Amis, E. J. Microfluidic Interfacial Tensiometry. Appl. Phys. Lett. 2005, 87 (8), 081905. (53) Cabral, J. T.; Hudson, S. D. Microfluidic Approach for Rapid Multicomponent Interfacial Tensiometry. Lab Chip 2006, 6, 427−436. (54) Taylor, G. I. The Formation of Emulsions in Definable Fields of Flow. Proc. R. Soc. London, Ser. A 1934, 146 (858), 501−523. (55) Rallison, J. M. The Deformation of Small Viscous Drops and Bubbles in Shear Flows. Annu. Rev. Fluid Mech. 1984, 16, 45−66.

(56) Dutcher, C. S.; Wexler, A. S.; Clegg, S. L. Surface Tensions of Inorganic Multicomponent Aqueous Electrolyte Solutions and Melts. J. Phys. Chem. A 2010, 114 (46), 12216−12230. (57) Pegram, L. M.; Record, M. T. Hofmeister Salt Effects on Surface Tension Arise from Partitioning of Anions and Cations between Bulk Water and the Air-Water Interface. J. Phys. Chem. B 2007, 111 (19), 5411−5417. (58) Sareen, N.; Shapiro, E. L.; Schwier, A. N.; McNeill, V. F. Secondary Organic Material Formed by Methylglyoxal in Aqueous Aerosol Mimics - Part 2: Product Identification Using Aerosol-CIMS. Atmos. Chem. Phys. Discuss. 2009, 9, 15567−15594. (59) Vazquez, G.; Alvarez, E.; Navaza, J. M. Surface Tension of Alcohol + Water from 20 to 50 °C. J. Chem. Eng. Data 1995, 40 (3), 611−614. (60) Alvarez, E.; et al. Surface Tension of Organic Acids + Water Binary Mixtures from 20 to 50 C. J. Chem. Eng. Data 1997, 42, 957− 960. (61) Ernst, R. C.; Watkins, C. H.; Ruwe, H. H. The Physical Properties of the Ternary System Ethyl Alcohol-Glycerin-Water. J. Phys. Chem. 1935, 40 (5), 627. (62) Belton, J. W. The Surface Tensions of Ternary Solutions. Part III. Trans. Faraday Soc. 1935, 31, 1648−1652. (63) Vanhanen, J.; Hyvärinen, A. P.; Anttila, T.; Raatikainen, T.; Viisanen, Y.; Lihavainen, H. Ternary Solution of Sodium Chloride, Succinic Acid and Water; Surface Tension and Its Influence on Cloud Droplet Activation. Atmos. Chem. Phys. 2008, 8 (16), 4595−4604.

J


Atmospheric Aqueous Aerosol Surface Tensions: Isotherm-Based

Recommend Documents