Water Interfaces: A Combined

Article pubs.acs.org/jced

Green Leaf Volatiles on Atmospheric Air/Water Interfaces: A Combined Experimental and Molecular Simulation Study Thilanga P. Liyana-Arachchi,†,‡,∇ Zenghui Zhang,†,∇ Harsha Vempati,† Amie K. Hansel,† Christopher Stevens,† Andrew T. Pham,† Franz S. Ehrenhauser,§ Kalliat T. Valsaraj,† and Francisco R. Hung*,†,∥ †

Cain Department of Chemical Engineering, and ∥Center for Computation and Technology, Louisiana State University, Baton Rouge, Louisiana 70803, United States ‡ Department of Chemistry, University of Minnesota, Minneapolis, Minnesota 55455, United States § Audubon Sugar Institute, Louisiana State University Agricultural Center, St. Gabriel, Louisiana 70776, United States S Supporting Information *

ABSTRACT: Green leaf volatiles (GLVs) are a family of oxygenated hydrocarbons emitted in large quantities by plants, especially those under mechanical stress or damage. GLVs can significantly contribute to the formation of secondary organic aerosols (SOAs) by reacting with oxidants in atmospheric water droplets (fog, mist, and rain). Here we investigated the properties of four GLVs, 2-methyl-3-buten-2-ol (MBO), methyl salicylate (MeSA), cis-3-hexen-1-ol (HxO), and cis-3-hexenylacetate (HxAc) in air/water systems at 298 K using a combination of experiments and molecular simulations. Good agreement between simulation and experimental results of the 1-octanol/water partition coefficients and the free energy of hydration of the GLVs ensure the suitability of our molecular models for these systems; likewise, surface concentrations determined from experimental measurements of the surface tensions of these systems compared favorably against trends determined from molecular simulations. Our simulations indicate that all four GLVs exhibit deep free energy minima at the air/water interfaces; the stability of the GLV solutes at the air/water interface is mainly driven by energetic interactions between these solutes and water molecules. Our results suggest that the air/water interface is the most likely site for chemical reactions between GLVs and atmospheric oxidants. Such a finding has important implications in the kinetics and mechanisms of these reactions, which can be significantly different from those observed when these reactions take place in the gas phase or in the bulk of water phases.

1. INTRODUCTION Atmospheric aerosols play an important role in air quality, public health and the climate, however much remains to be understood about the processes that contribute to their formation and composition.1−4 Organic compounds are of particular interest as they account for a substantial percentage of atmospheric particulate matter (PM).4−13 A major fraction of organic aerosols are secondary (i.e., are formed by compounds originated from chemical reactions in the atmosphere), and remain poorly understood.8,9 Many of these chemical reactions between organics and atmospheric oxidants (reactive oxygen species, ROSs, e.g., ozone, ·OH, · HO2, ·H2O2, and other radicals) leading to the formation of secondary organic aerosols (SOAs), can take place in the bulk and at the surface of atmospheric water droplets (fog, mist, and rain).2 Therefore, many efforts have concentrated on characterizing the organic fraction in fog waters,2,3,7,8 but these measurements can be complicated as nearly 40 % of watersoluble organic carbon (WSOC) is typically unidentified;14,15 © 2014 American Chemical Society

factors such as source, altitude, and latitude of where the samples were taken also influence these measurements.14−19 Many organic precursors that react and form the WSOC in fog and aerosols are biogenic (i.e., plant- or bacterial/microbialderived).20−26 Among the biogenic volatile organic compounds (BVOCs) are the green leaf volatiles (GLVs),6,27−29 a family of oxygenated hydrocarbons emitted in large quantities by plants under stress conditions (e.g., grass cutting, grazing of animals, and other mechanical damages, as well as local weather changes such as freezing spells).30−34 Typical compounds in this class (Figure 1) include 2-methyl-3-buten-2-ol (MBO), methyl salicylate (MeSA), cis-3-hexen-1-ol (HxO), and cis-3-hexenylacetate (HxAc). Special Issue: Modeling and Simulation of Real Systems Received: January 31, 2014 Accepted: June 13, 2014 Published: June 24, 2014 3025

dx.doi.org/10.1021/je500114m | J. Chem. Eng. Data 2014, 59, 3025−3035

Journal of Chemical & Engineering Data

Article

water interfaces, which were then used to estimate surface concentrations of these GLV solutes; these surface concentrations were compared against the trends determined from our simulations. The validated molecular models were then used in simulations to provide molecular-level information that is difficult (or even impossible) to determine from experiments. In particular, we determined the free energy associated with moving a molecule of HxO or HxAC between the gas phase and the bulk water phase, and compared those to the results obtained for MBO and MeSA in previous studies;35,36 we also used these calculations to obtain insights of the molecular-level interactions between the GLVs and the water molecules in these systems. We also conducted classical molecular dynamics (MD) simulations to explore some of the interfacial properties of these systems. The rest of the article is arranged as follows. In section 2 we provide details of our experimental and computational methods; our main results are presented and discussed in section 3, and in section 4 we include our concluding remarks.

Figure 1. Molecular structures of (a) 2-methyl-3-buten-2-ol (MBO), (b) methyl salicylate (MeSA), (c) cis-3-hexen-1-ol (HxO), and (d) cis3-hexenylacetate (HxAc).

Molecular simulations are uniquely positioned to complement experimental efforts in this area, as it is complicated to gain a complete picture of these interfacial systems from experiments alone. The multiple questions that typically arise when interpreting experimental results could be answered by combining the experimental findings with the atomic- and molecular-level details that can be provided by computer simulations. Our previous molecular simulation studies35,36 have shown that the GLVs MBO and MeSA, as well as oxygenated radicals such as ·OH, are thermodynamically favored to remain at the air/water interface. Such a result suggests that chemical reactions among these species are more likely to occur at this interface, rather than in the bulk of atmospheric water droplets or in the gas phase. This finding has implications in the kinetics of the reactions between organics and atmospheric oxidants, which are known to vary significantly depending on the reaction environment.2,17 In the present study, we investigated the properties of two other GLVs, HxO and HxAc, near atmospheric air/water interfaces, and compared the results with those from our previous studies with MBO and MeSA.35,36 HxO and HxAc have been identified as prevalent emissions from 29 agricultural and natural plant species,37,38 are known to be highly reactive toward ·OH radicals and ozone,39,40 and have short tropospheric lifetimes (on the order of a few hours).28,31,41 The leaf alcohol HxO is an important SOA precursor due to the high reactivity of its primary oxidation product, 3-hydroxypropanal, which can hydrate and undergo further reactions with other aldehydes, resulting in SOAs with oligomers of higher molecular weight. In contrast, the ozonolysis of HxAc produces mainly smaller SOAs, which accounts for the acetate functionality that inhibits oligomer formation in the particle phase.10,28 One issue discussed in our previous studies with GLVs35,36 is the lack of experimental data for these compounds and their oxidation products. In particular, reliable experimental values of physicochemical parameters such as Henry’s law constants and 1-octanol/water partition coefficients are needed to ensure that the combination of molecular models for the different species (GLVs, water) can reproduce the physics that are important for these systems. Following our previous studies,35,36 here we determined experimental values of the 1octanol/water partition coefficients for both HxO and HxAc, which were used to validate our molecular models. We also determined experimental values of the surface tensions of the GLVs MBO, MeSA, HxO and HxAc at the atmospheric air/

2. METHODS 2.1. Experimental Determination of the 1-Octanol/ Water Partition Coefficient. The 1-octanol/water partition coefficient of HxO and HxAc were determined using the shakeflask method according to OECD guideline 107;42 such a procedure was also used in our previous studies35,36 to determine the 1-octanol/water partition coefficient of MBO and MeSA. Water (Burdick and Jackson LC−MS grade, Honeywell, Muskegan, MI, USA), and HxO (> 98 %), HxAc (natural), MeSa (≥ 99 %), MBO (≥ 98 %), and 1-octanol (analytical grade) from Sigma-Aldrich (Milwaukee, WI, USA), were used as received with no further purification. Our partition coefficients are reported in Table 1; each reported value was averaged over six independent measurements, and each standard uncertainty was determined from the standard deviation of these measurements. Table 1. Measured 1-Octanol/Water Partition Coefficient at Temperature T = 25 °C and Pressure p = 0.1 MPa GLV

log(KOW)a

HxAC HxO MeSa MBO

2.48 1.47 2.36 0.69

(0.01) (0.01) (0.01) (0.01)

a Standard uncertainty u[log(KOW)] given in parentheses; evaluated from the standard deviation of the measurements of partition coefficients (n = 6 measurements).

2.2. Experimental Determination of Surface Tension and Surface Concentration. Aqueous solutions of each GLV were prepared using water (Burdick and Jackson LC−MS grade, Honeywell, Muskegan, MI, USA) and the appropriate GLV. The bulk concentration of the GLVs was determined via high performance liquid chromatography (HPLC) analysis using an Agilent 1100 HPLC-UV/DAD. GLV bulk concentrations in the aqueous solutions ranged from 0 mmol·L−1 to 11.6 mmol·L−1 with aqueous solubility being a limiting factor. A Kruss K14 tensiometer (Kruess GmbH, Hamburg, Germany) was used for the surface tension measurements, utilizing 45 mL of each solution placed in a 7 cm Petri dish inside the tensiometer held at 25 °C. 3026



Article

OPLS-AA force field45 for all the GLVs and 1-octanol. To ensure that our models can accurately reproduce the physics relevant for these systems, we used thermodynamic integration (TI) calculations35,36,46−48 to determine the free energy of hydration (ΔGhydration) of HxO and HxAc in the water phase, as well as the free energy of solvation (ΔGsolvation) of these GLVs in a water-saturated 1-octanol phase.49,50 These two values were then used to calculate the simulated 1-octanol/water partition coefficients for HxO and HxAc:51

The measured surface tensions σ and the bulk concentrations of the GLVs were used to calculate the surface concentrations of the GLVs, based on the slope of a linear plot of σ vs ln(Ci) according to eq 1: σC2 − σC1 ΓC2 = − RT ln(C2/C1) (1) where ΓC2 is the surface concentration values of GLV at bulk concentration C2, σC2, and σC1 refer to the surface tensions at concentration of C2 and C1, respectively, R is the universal gas constant, and T is the absolute temperature in Kelvin.43 In Table 2 we present the measured bulk GLV concentrations and

log(K OW ) =

surface tensionb σ

surface concentrationc Γ·107

GLV

mmol/kg H2O

mN·m−1

mol·m−2

HxAc

0.0 0.647 (0.001) 1.500 (0.001) 1.857 (0.001) 0.0 0.98 (0.02) 3.03 (0.03) 5.04 (0.02) 5.96 (0.03) 0.0 1.09 (0.01) 2.14 (0.01) 3.14 (0.06) 5.29 (0.06) 0.0 0.98 (0.05) 2.96 (0.08) 5.21 (0.04) 8.84 (0.05)

HxO

MeSA

MBO

70.2 66.8 62.3 58.7 71.5 69.5 65.2 63.5 61.5 71.0 70.8 70.1 68.9 65.2 70.8 70.6 69.7 69.5 67.7

(0.1) (0.1) (0.1) (0.1) (0.1) (0.1) (0.2) (0.1) (0.1) (0.1) (0.1) (0.1) (0.1) (0.2) (0.1) (0.1) (0.1) (0.1) (0.1)

2.303RT

(2)

The remainder of the details of this procedure are the same as described in detail in our previous papers and described briefly here.35,36 The target GLV molecule was first modeled under vacuum conditions by switching off all the interactions with the solvent molecules (water or wet octanol). These interactions were then gradually turned on by coupling a parameter λ to the Hamiltonian H, where λ ranges from 0 (initial state, under vacuum) to 1 (final state, fully hydrated/solvated) with a step of Δλ = 0.05, which leads to 21 intermediate states with equally spaced values of λ. The simulations were conducted in the NPT ensemble (P = 1 atm and T = 298 K) for a total of at least 2.5 ns, of which at least the last 2 ns were used to calculate the averages. ΔGhydration and ΔGsolvation were determined by numerically integrating ⟨∂H/∂λ⟩ over the 21 simulations, and eq 2 was used to finally obtain the simulated log(KOW). The Parrinello−Rahman barostat was used to couple the pressure in these calculations, and the temperature was coupled using the velocity-rescaling algorithm of Bussi and Parrinello et al.52,53 According to our calculations (see later in section 3.1), using the original OPLS-AA parameters for HxO and HxAc led to values of log(KOW) that are significantly larger than those measured in our experiments. Therefore, we used the same strategy described in previous studies35,36,54−57 and increased the values of the atomic charges taken from the OPLS-AA force field by 10 % (see Supporting Information). We need to keep in mind that the nonbonded parameters (including atomic charges) from OPLS-AA were obtained from thermodynamic and structural properties of a number of pure organic liquids, and that, rigorously speaking, OPLS-AA is not a transferable force field. Furthermore, effects such as polarization might arise when the GLV molecules are hydrated (in water) or solvated (in wet 1-octanol). Increasing the atomic charges by 10 % attempts to indirectly address those issues. This new set of charges produced values of log(KOW) for HxO and HxAc that were in better agreement with experimental values. The adjusted OPLS-AA parameters for HxO and HxAc were then used to conduct potential of mean force (PMF) calculations and classical (unconstrained) molecular dynamics simulations of these GLVs at the air/water interfaces at 298 K. The simulation setup was the same used in our previous studies,35,36 where we created two air/water interfaces by placing a slab containing all the SPC/E water molecules in the center of an elongated orthorhombic simulation box with periodical boundary conditions in three directions. In our PMF calculations, we determined the free energy associated with moving one GLV molecule (HxO or HxAc) between the gas phase and the water slab. An orthorhombic simulation box of dimensions 4 nm × 4 nm × 30 nm containing 2197 SPC/E water molecules was used for these calculations. All of the simulations in our PMF calculations were run for at least 40 ns

Table 2. Measured GLV Bulk Concentration, Surface Tension of Each Solution, And Calculated Surface Concentration Values at Temperature T = 25 °C and Pressure p = 0.1 MPa bulk concentrationa C

ΔG hydration − ΔGsolvation

21.5 69.4 15.2 14.0 48.2 4.1 13.3 28.6 3.4 1.2 14.2

a

Standard uncertainty u(C) given in parentheses; calculated from linear calibration of HPLC analysis for each concentration value. b Standard uncertainty u(σ) given in parentheses; evaluated from the standard deviation of the surface tension measurements (n = 8 measurements). cCalculation of surface concentrations requires two nonzero bulk concentration values [see eq 1]. The symbol “-” is used here to indicate that the calculation is not possible for the first two bulk concentrations.

surface tensions, and the calculated surface concentration of each GLV solution. In Table 2, the standard uncertainties associated with the determination of bulk concentrations were due to HPLC analysis and usage of a calibration curve. These samples were diluted before their introduction to HPLC, and the uncertainty refers to the diluted samples. The dilution is assumed not to be a significant source of error. Regarding surface tension, each measurement was replicated a total of eight times. The average is reported in Table 2, and the standard uncertainty was evaluated from the standard deviation of these measurements. 2.3. Computational Models and Methods. Following our previous studies on MBO and MeSA on air/water interfaces,35,36 we used the SPC/E model44 for water and the 3027



Article

using the constraint force method (where the z-distance between the center of mass of the GLV solute and the center of mass of the water slab was constrained); the remainder of the details are exactly the same as described in our previous papers.35,36 Classical MD simulations were also performed using a simulation box of dimensions 8 nm × 8 nm × 30 nm, containing 17131 SPC/E water molecules and up to 2 molecules of GLVs (HxO or HxAc), in order to accurately determine density profiles, radial distribution functions, and surface tensions. The remainder of the details are exactly the same as described in our previous work and briefly described here.35,36,54,58−61 For each simulated system, the initial configurations were relaxed first using the steepest descent energy minimization method. A time step of 1 fs was used and we collected data every 10 ps. LINCS algorithm was used to constrained bond lengths,62 and the velocity-rescaling algorithm of Bussi and Parrinello et al. was used for temperature coupling.52,53 Lennard-Jones cutoff distance was set as 0.9 nm. We used the particle-mesh Ewald (PME) method63 in 3D with cubic interpolation, a cutoff of 0.9 nm, and a grid spacing of 0.12 nm to account for long-range Coulombic interactions. All our molecular simulations were performed using GROMACS,64 and VMD65 was used to generate snapshots of our simulations results.

simulation partition coefficients are summarized in Figure 2. In addition, from reported experimental values of the Henry’s law

Figure 2. Simulated and experimental values of 1-octanol/water partition coefficients log(KOW) for the GLVs HxO, HxAc, MBO, and MeSA.

constant for MBO,67 MeSA,68 and HxAc69 in water, we determined their experimental free energy of hydration to be ΔGhydration = −19.5 kJ/mol (MBO), −16.6 kJ/mol (MeSA) and −8.92 kJ/mol (HxAc). These values are reasonably close to the free energies of hydration obtained from our simulations: ΔGhydration = −22.8 kJ/mol (MBO),35 between −20.7 kJ/mol and −29.8 kJ/mol (MeSA),36 and −5.75 kJ/mol (HxAc, this work). However, we note that the reported uncertainties in the experimental values of the Henry’s law constants for MeSA68 and HxAc 69 propagate into large uncertainties in the experimental free energy of hydration (between 3.7 kJ/mol and 11.4 kJ/mol). Overall, the favorable comparison between experiment and simulation results strongly suggests that our simulation models for MBO, MeSA, HxO, and HxAc, when used in combination with the SPC/E water model, are able to reproduce relevant physical properties of these species in our interfacial systems. 3.2. Potentials of Mean Force of the GLVs MBO, MeSA, HxO and HxAc in Air/Water Systems. Our simulation models were used to determine the PMF associated with moving one molecule of MBO, MeSA, HxO, and HxAc between the gas and the bulk water phases, and through the air/water interface. These results are shown in Figure 3a; in these calculations we followed the study of Hub et al.47 and assumed the PMF of each GLV solute to be zero when they are in the bulk water phase. The PMF profiles shown in Figure 3a exhibit deep free energy minima of −13.7 kJ/mol (MBO), −12.6 kJ/mol (MeSA), −14.8 kJ/mol (HxO), and −20.5 kJ/ mol (HxAc) at the air/water interface. The uncertainty in these calculations is less than 1 kJ/mol. These results indicate that these GLVs have a thermodynamic preference to remain at the air/water interface, as compared to being in the bulk water phase or in the gas phase. These trends are consistent with the interfacial properties of other organic species such as n-alkanes and polycyclic aromatic hydrocarbons at air/water and air/ice interfaces.54,56−60,70 Furthermore, previous studies have also determined that atmospheric oxidants such as ROSs also have a thermodynamic preference to stay at the air/water or air/ice interface.55,59,61 All these results strongly suggest that the chemical reactions between GLVs and ROSs are more likely to take place at the air/water interface, which can have implications in the kinetics and mechanisms of the reactions,

3. RESULTS AND DISCUSSION 3.1. Validation of Simulation Force Field Parameters: 1-Octanol/Water Partition Coefficients from Experiments and Simulations. From our experimental measurements, the 1-octanol/water partition coefficients were found to be log(KOW) = 1.47(0.01) for HxO, and log(KOW) = 2.48(0.01) for HxAc, with the standard uncertainties given in parentheses. From our previous studies, the experimental 1-octanol/water partition coefficients for MBO and MeSA were log(KOW) = 0.69(0.01), and 2.36(0.01), respectively.35,36 All these values and their uncertainties are summarized in Table 1. To validate our choice of force fields (see section 2.3), we performed TI calculations to compute the free energies needed (eq 3) to determine the 1-octanol/water partition coefficient of HxO and HxAc. From initial simulations using the original OPLS-AA charges for these two molecules, the simulated free energies were determined to be ΔGhydration = −12.25 kJ/mol and ΔGsolvation = −30.26 kJ/mol for HxO, and ΔGhydration = −2.41 kJ/mol and ΔGsolvation = −27.12 kJ/mol for HxAc. Using these values, the calculated 1-octanol/water partition coefficients were log(KOW) = 3.2(0.6) for HxO and log(KOW) = 4.3(0.4) for HxAc, with the standard uncertainties given in parentheses. These values are well above the experimental values reported before. Therefore, and following previous work,35,36,55−57 we increased the original atomic charges of HxO and HxAc by 10 %. Using these new charges (see Supporting Information) we repeated the TI calculations, obtaining ΔGhydration = −23.03 kJ/ mol and ΔGsolvation = −31.58 kJ/mol for HxO, and ΔGhydration = −5.75 kJ/mol and ΔGsolvation = −22.56 kJ/mol for HxAc. These new values yield log(KOW) = 1.5(0.4) for HxO, and log(KOW) = 2.9(0.5) for HxAc, which are in better agreement with the experimental values reported above. The uncertainty in our simulated 1-octanol/water partition coefficients are comparable to those obtained for n-alkanes and ionic liquids using a different simulation technique.51,66 From our previous molecular simulation results,35,36 the simulated 1-octanol/water partition coefficient of MBO is log(KOW) = 0.8(0.2), and log(KOW) = 2.6(0.4) for MeSA. All these experimental and 3028



Article

concentrations are low), the concentration of the solutes at the surface of the water droplets will be proportional to the depth of the PMF minimum at the air/water interface. Therefore, the PMF results presented in Figure 3a suggest that, for a given concentration of GLV in the bulk water phase, we should observe a larger surface concentration at the air/water interface for HxAc, followed by HxO, MBO, and MeSA. We determined experimentally the surface tensions in our GLV systems, and used those to calculate the surface concentration of GLVs at the air/water interface (see section 2.2). These experimental results are summarized in Table 2, and strongly suggest that for a given concentration of GLV solute in the bulk water phase, the largest surface concentrations are observed for HxAc, followed by HxO, MeSA, and MBO. The trends for the surface concentrations as predicted from the PMF calculations are in general agreement with the trends determined from the experimental surface tensions, with the exception of MeSA and MBO. However, we note that the difference in the PMF minima observed for MeSA and MBO at the air/water interface (∼1.1 kJ/mol) is very similar to the numerical uncertainty of our calculations (∼1 kJ/mol). We also note that the results from Table 2 suggest that among the GLV solutes, MBO produced the smallest variation in the surface tension in the range of bulk concentrations studied in our experiments. Variations in the bulk concentration of HxAc, HxO, and MeSA seem to lead to larger changes in the surface tension than for MBO, and consequently larger variations in the surface concentrations of HxAc, HxO, and MeSA are observed. To further explore the interactions between the GLVs and the water molecules near the air/water interfaces, we followed previous studies35,36,47,71 and decomposed the PMF of the GLVs into their enthalpic and entropic components. The enthalpic component of the PMF (Figure 3b) was determined as the average potential energy of the system in each of the PMF simulation runs, and the entropic component (Figure 3c) was determined as the difference between the PMF and its enthalpic component for each solute at any given value of the z coordinate. The enthalpic component of the PMF of all the GLV solutes exhibit deep free minima at the interface (Figure 3b), indicating that the thermodynamic stability of these solutes at the air/water surface is mainly driven by energy interactions. Moving the GLV solutes from the gas phase into the interface lead to favorable energy interactions of the solutes with the water molecules, making the enthalpy to become more negative. These energy interactions outweigh the entropy penalty (Figure 3c) associated with moving each GLV solute from the gas phase into the air/water interface. MeSA has the smallest entropic penalty when moving from the gas phase to the interface; similarly, MeSA exhibits the smallest increase in entropy when going from the air/water interface into the bulk water phase (Figure 3c). These observations might be caused by the presence of more oxygenated groups in MeSA than in the other GLV solutes (Figure 1), and by the fact that the aromatic ring in MeSA is less hydrophobic than the carbon chains and methyl groups present in MBO, HxO, and HxAc. To further understand the energetics of the forces driving the GLV solutes into the air/water interface, we followed previous studies36,47,71 and decomposed the enthalpic component of the PMF (Figure 3b) into the water−water and water−GLV interaction energies as a function of the z coordinate of the GLV solutes (Figure 4a,b). For each solute, the curves shown in Figure 4a,b add up to the enthalpic component of the PMF shown in Figure 3b. In general these results indicate that

Figure 3. (a) Total PMF and its (b) enthalpic and (c) entropic components associated with moving a molecule of the GLVs MeSA, MBO, HxAc, and HxO between the gas phase and the bulk water phase in an air/water system at 298 K. The PMF of all GLVs in the bulk water phase was arbitrarily assumed to be zero. The dash line represents the location of the air/water interface (arbitrarily defined as the point where the density of water in the simulation box reaches 500 kg/m3).

as those can vary significantly depending on the reaction environment.2,17 The deepest minimum at the air/water interface was observed for HxAc, followed by HxO, MBO, and MeSA (Figure 3a). The values of the free energy minima at this interface depend on the molecular structure of the GLV solutes (Figure 1), and affect the partitioning of these GLVs between the bulk water phase and the air/water interface. The study of Hub et al.47 suggests that, for any fixed concentration of different solutes in bulk water (and assuming these bulk 3029



Article

moving the GLV solutes from the air/water interface into the bulk water phase produces a sharp increase in the water−water energy (Figure 4a). This observation is a consequence of the fact that a GLV solute disrupts a larger number of water−water hydrogen bonds when it is inside the bulk water phase than when it is at the air/water surface. On the other hand, all of the GLV solutes are amphiphilic (Figure 1) and can form hydrogen bonds with water molecules; this observation is reflected in the water−GLV interactions shown in Figure 4b, which become more favorable (more negative) when moving the GLV molecules from the gas phase to the interface and then into the bulk water phase. Therefore, a GLV molecule in the gas phase can be “dragged” by water molecules into the air/water interface, as the water−GLV interactions provide a thermodynamic incentive. However, if the GLV solute moves from the interface to the bulk water phase, the penalty in the water− water energy (Figure 4a) outweighs the reduction in the water−GLV energy (Figure 4b). As a result, water molecules tend to “push” the GLV solutes from the bulk water phase back into the interface. Therefore, the air/water interface provides an ideal environment where the GLV solutes can form a significant number of hydrogen bonds and have important energy interactions with the water molecules, without disrupting a large number of water−water hydrogen bonds. These trends are similar to those observed for other organic molecules near air/water interfaces.47 Among the GLV solutes, MeSA exhibits the largest reduction in magnitude of the water−GLV energy when moving from the gas phase into the interface (Figure 4b), which is consistent with MeSA having more oxygenated groups than the other GLV solutes (Figure 1), as well as by the fact that the aromatic ring in MeSA is less hydrophobic than the carbon chains and methyl groups present in MBO, HxO, and HxAc. On the other hand, MBO yields the smallest increase in

Figure 4. Decomposition of the enthalpic components of the PMF (Figure 3b) into (a) water−water and (b) water−GLV interaction energies at 298 K.

Figure 5. Density profiles of water and the GLVs (a) MBO, (b) MeSA, (c) HxO, and (d) HxAc, obtained from classical MD simulations at 298 K. The systems have a dimension of 8 nm × 8 nm × 30 nm and contain 2 GLV molecules per interface. The density profile of each species is normalized with respect to the maximum value of their local density in the simulation box. 3030



Article

Improved statistics can be obtained by considering larger systems with the same surface concentration of GLVs as in this study (i.e., larger xy areas and a larger number of GLV molecules per interface). The density profiles of the hydrophobic and hydrophilic groups of each GLV are shown in Figure 7, which provide insights of how the GLVs orient at the air/water interfaces. Not surprisingly, the hydrophilic groups of the GLV solutes tend to remain closer to the water side of the air/water interface, while the hydrophobic groups orient closer to the gas phase. As observed by Hub et al.47 for small organic molecules near the air/water interfaces, the simulation snapshots and the density profiles shown in Figures 6 and 7 suggest that the GLV solutes try to orient at the interface in ways that maximize the dispersion interactions between their hydrophobic parts and the water molecules. Simulated surface tensions (σ) for our systems were calculated using the following equation:64

water−water energy when moving from the interface into the bulk water phase, probably because its small size causes the smallest disruption in water−water hydrogen bonds as compared to the other GLV solutes. 3.3. Interfacial Properties of GLVs in Air/Water Systems. We conducted classical MD simulations having 1 or 2 molecules of GLVs per air/water interface, which corresponds to surface concentrations ranging between 1.08· 10−6 mol/m2 and 2.56·10−6 mol/m2. These simulated surface concentrations are within the range of values determined experimentally by us from surface tension measurements (between 0.406·10−6 and 4.82·10−6, Table 2). Density profiles of the GLV and water molecules as a function of the zcoordinate (perpendicular to the air/water interface) are shown in Figure 5, and representative snapshots of these MD simulations are presented in Figure 6. These systems contain

σ=

⎞ Lz ⎛ Pxx + Pyy − Pzz⎟ ⎜ 2⎝ 2 ⎠

(3)

where Lz is the box length in the z direction, and Pxx, Pyy, and Pzz indicate the diagonal components of the pressure tensor. In Figure 8 we show the interfacial tensions of our systems containing 1 or 2 GLV molecules per air/water interface. These results suggest that the surface tension decreases as the concentrations of the GLV solutes increase, which is in agreement with the trends in the experimental data presented in Table 2. Nevertheless, the simulated values reported in Figure 8 depart significantly from the measured experimental values reported in Table 2. These simulated values of surface tension were obtained using the rescaled set of atomic charges for the GLVs (see sections 2.3 and 3.1). For completeness, we ran an additional simulation for a system with one molecule of HxAc per interface using the original atomic charges from OPLS-AA; the simulated surface tension for this system (σ = 55.0 ± 0.2 mN/m) was statistically similar to the one obtained using the rescaled atomic charges (σ = 54.9 ± 0.3 mN/m). From Figure 8, we note that our simulated surface tension in the absence of GLVs was found to be 57.9 mN/m, which is comparable to values previously reported for SPC/E water,72−75 but is nevertheless still far from the experimental value of the surface tension of water (71.7 mN/m at 300 K).74 This discrepancy might be the most important cause behind the lack of quantitative agreement between simulation and experimental results for surface tensions in our systems. We note that simulation parameters such as the LJ cutoff distance and the inclusion of long-range tail corrections to the surface tension74 can significantly affect the computed surface tension from simulations. For comparison purposes, we obtained σ = 57.9 nN/m for SPC/E water (no GLVs) at T = 298 K, using a LJ cutoff distance of 0.9 nm and excluding any long-range tail corrections to σ. In turn, Vega and de Miguel74 used a LJ cutoff distance of 1.3 nm with a switching function between 1.2 nm and 1.3 nm, and at T = 300 K they obtained σ = (60.2−60.8) mN/m (no tail corrections to σ considered), and σ = (63.5− 63.7) mN/m when tail corrections are added to σ. This comparison indicates that simulation parameters such as the LJ cutoff distance and the tail corrections to the surface tension can affect significantly the computed values of σ. However, the values of σ obtained by Vega and de Miguel for SPC/E water still depart significantly from the experimental surface tension of water (71.7 mN/m at 300 K). Vega and de Miguel74 indicate

Figure 6. Side views of representative snapshots of water and the GLVs (a) MBO, (b) MeSA, (c) HxO, and (d) HxAc, from classical MD simulations at 298 K. The systems have a dimension of 8 nm × 8 nm × 30 nm and contain 2 GLV molecules per interface.

2 GLV molecules per air/water interface. The results shown in Figure 5 and 6 indicate again that the GLVs prefer to stay at the air/water interfaces, which is consistent with the PMF results shown in Figure 3a, and again suggests that chemical reactions between GLVs and atmospheric oxidants are more likely to take place at the air/water interface. The small lumps observed in the density profiles of MBO and HxO (Figure 5a,c) suggest that, although these GLVs mainly prefer to remain at the air/ water interface, they do spend an important amount of time lying deeper inside the water slab. In contrast, the density profiles of MeSA and HxAc do not exhibit any lumps inside the water slab (Figure 5b,d). These observations are consistent with the PMF results shown in Figure 3; we note that the minima for MBO and HxO (Figure 3a) are located closer to the water side of the interface, whereas the minima for MeSA and HxAC seem to be closer to the air side of the interface. The positions of these minima seem to be driven by the enthalpic interactions. In the curves for the enthalpic component of the PMF (Figure 3b), the minima for MBO and HxO are again closer to the water side of the interface, whereas for MeSA and HxAc the minima are closer to the air side. On the other hand, we note that the results shown in Figure 5 were obtained for systems having only two GLV molecules per interface. 3031



Article

Figure 7. Decomposed density profiles of hydrophilic and hydrophobic groups of the GLVs (a) MBO, (b) MeSA, (c) HxO, and (d) HxAc, obtained from classical MD simulations at 298 K. The systems have a dimension of 8 nm × 8 nm × 30 nm and contain 2 GLV molecules per interface. The density profile of each species is normalized with respect to the maximum value of their local density in the simulation box.

experiments were performed to compute the 1-octanol/water partition coefficients [log(KOW)] for HxO and HxAc. After adjusting the atomic charges, the simulated log(KOW) of those two GLVs were in good agreement with the measured experimental values; good agreement between simulation and experiments was also found for the 1-octanol/water partition coefficients of MBO and MeSA in our previous studies.35,36 Likewise, good agreement was found between the simulated free energy of hydration of MBO, MeSA, and HxAc and the values calculated from experimental values of the Henry’s law constant of these GLVs in water. PMF calculations indicate that all these GLVs have deep free energy minima at the air/water interfaces. Classical MD simulation results further confirm these trends, suggesting that chemical reactions between GLVs and ROSs are more likely to take place at the surface of water droplets in the lower atmosphere. Such a finding has important implications in the kinetics and mechanisms of these reactions, which might be very different from the kinetics and mechanisms of reactions observed in the gas phase or in the bulk of water phases. The thermodynamic stability of the GLV solutes at the air/water interface is mainly driven by energy interactions, with the air/water interface providing an ideal environment where the GLV solutes have important energy interactions with the water molecules without disrupting a large number of water−water hydrogen bonds (which would be the case if the GLV solutes would remain inside the bulk water phase). The values of the free energy minima at the air/water interface depend on the molecular structure of the GLV solutes, and determine the partitioning of these GLVs between the bulk water phase and the air/water interface. The trends in surface concentrations as predicted from the simulated PMF minima are in general agreement with the trends observed for the

Figure 8. Simulated surface tensions for air/water systems containing 1 or 2 GLVs per air/water interface. Results for a system containing no GLVs (only SPC/E water molecules) are also shown. All systems have a dimension of 8 nm × 8 nm × 30 nm.

that the TIP4P/2005 water model gives σ = (69.1−69.5) mN/ m when tail corrections are included, which are the closest to the experimental surface tension of water; the next best models for surface tension are the SPC/E and TIP4P/Ew water models. In this study we decided to use the SPC/E water model to be consistent with our previous studies of GLVs in air/water systems;35,36 additional calculations using the TIP4P/ 2005 water model can be considered for future studies.

4. CONCLUDING REMARKS Molecular simulations and experiments were conducted to investigate the properties of four GLVs, HxO, HxAc, MBO, and MeSA, near air/water interfaces at 298 K. To ensure our choice of models can reproduce relevant physical properties of these systems, thermodynamic integration (TI) calculations and 3032



Article

(4) Rudich, Y.; Donahue, N. M.; Mentel, T. F. Aging of organic aerosol: Bridging the gap between laboratory and field studies. Annu. Rev. Phys. Chem. 2007, 58, 321−352. (5) Middlebrook, A. M.; Murphy, D. M.; Thomson, D. S. Observations of organic material in individual marine particles at Cape Grim during the First Aerosol Characterization Experiment (ACE 1). J. Geophys. Res.-Atmos. 1998, 103, 16475−16483. (6) Goldstein, A. H.; Galbally, I. E. Known and unexplored organic constituents in the earth’s atmosphere. Environ. Sci. Technol. 2007, 41, 1514−1521. (7) Kanakidou, M.; Seinfeld, J. H.; Pandis, S. N.; Barnes, I.; Dentener, F. J.; Facchini, M. C.; Van Dingenen, R.; Ervens, B.; Nenes, A.; Nielsen, C. J.; Swietlicki, E.; Putaud, J. P.; Balkanski, Y.; Fuzzi, S.; Horth, J.; Moortgat, G. K.; Winterhalter, R.; Myhre, C. E. L.; Tsigaridis, K.; Vignati, E.; Stephanou, E. G.; Wilson, J. Organic aerosol and global climate modelling: A review. Atmos. Chem. Phys. 2005, 5, 1053−1123. (8) Kroll, J. H.; Seinfeld, J. H. Chemistry of secondary organic aerosol: Formation and evolution of low-volatility organics in the atmosphere. Atmos. Environ. 2008, 42, 3593−3624. (9) Turpin, B. J.; Saxena, P.; Andrews, E. Measuring and simulating particulate organics in the atmosphere: problems and prospects. Atmos. Environ. 2000, 34, 2983−3013. (10) Andrews, E.; Saxena, P.; Musarra, S.; Hildemann, L. M.; Koutrakis, P.; McMurry, P. H.; Olmez, I.; White, W. H. Concentration and composition of atmospheric aerosols from the 1995 SEAVS experiment and a review of the closure between chemical and gravimetric measurements. J. Air Waste Manage. 2000, 50, 648−664. (11) Gray, H. A.; Cass, G. R.; Huntzicker, J. J.; Heyerdahl, E. K.; Rau, J. A. Characteristics of atmospheric organic and elemental carbon particle concentrations in Los-Angeles. Environ. Sci. Technol. 1986, 20, 580−589. (12) Shah, J. J.; Johnson, R. L.; Heyerdahl, E. K.; Huntzicker, J. J. Carbonaceous aerosol at urban and rural sites in the United States. J. Air Pollut Control Assoc. 1986, 36, 254−257. (13) White, W. H.; Roberts, P. T. Nature and origins of visibilityreducing aerosols in Los-Angeles air basin. Atmos. Environ. 1977, 11, 803−812. (14) Raja, S.; Raghunathan, R.; Yu, X.-Y.; Lee, T.; Chen, J.; Kommalapati, R. R.; Murugesan, K.; Shen, X.; Qingzhong, Y.; Valsaraj, K. T.; Collett, J. L., Jr. Fog chemistry in the Texas−Louisiana Gulf Coast corridor. Atmos. Environ. 2008, 42, 2048−2061. (15) Herckes, P.; Hannigan, M. P.; Trenary, L.; Lee, T.; Collett, J. L. Organic compounds in radiation fogs in Davis (California). Atmos. Res. 2002, 64, 99−108. (16) Cappiello, A.; De Simoni, E.; Fiorucci, C.; Mangani, F.; Palma, P.; Trufelli, H.; Decesari, S.; Facchini, M. C.; Fuzzi, S. Molecular characterization of the water-soluble organic compounds in fogwater by ESIMS/MS. Environ. Sci. Technol. 2003, 37, 1229−1240. (17) Raja, S.; Ravikrishna, R.; Kommalapati, R. R.; Valsaraj, K. T. Monitoring of fogwater chemistry in the gulf coast urban industrial corridor: Baton Rouge (Louisiana). Environ. Monit. Assess. 2005, 110, 99−120. (18) Chen, J.; Ehrenhauser, F. S.; Valsaraj, K. T.; Wornat, M. J. Uptake and UV-photooxidation of gas-phase PAHs on the surface of atmospheric water films. 1. Naphthalene. J. Phys. Chem. A 2006, 110, 9161−9168. (19) Chen, J.; Valsaraj, K. T. Uptake and UV-photooxidation of gasphase polyaromatic hydrocarbons on the surface of atmospheric water films. 2. Effects of dissolved surfactants on naphthalene photooxidation. J. Phys. Chem. A 2007, 111, 4289−4296. (20) Guenther, A.; Hewitt, C. N.; Erickson, D.; Fall, R.; Geron, C.; Graedel, T.; Harley, P.; Klinger, L.; Lerdau, M.; McKay, W. A.; Pierce, T.; Scholes, B.; Steinbrecher, R.; Tallamraju, R.; Taylor, J.; Zimmerman, P. A global-model of natural volatile organic-compound emissions. J. Geophys. Res.-Atmos. 1995, 100, 8873−8892. (21) Ortega, J.; Helmig, D. Approaches for quantifying reactive and low-volatility biogenic organic compound emissions by vegetation enclosure techniquesPart A. Chemosphere 2008, 72, 343−364.

surface concentrations calculated from experimental measurements of the surface tension. The lack of experimental data for GLVs and their oxidation products is one important issue that needs to be addressed to obtain further insights on their relevance to SOAs. Experimental data can also be used to obtain reliable molecular models for these systems, which in turn are used to provide molecular-level details to complement experimental efforts in this area. We also note that in this study we did not explicitly consider chemical reactions between GLVs and atmospheric oxidants at air/water interfaces. However, many of the interfacial properties studied here are directly relevant to these reactions. Some of the current efforts by the Valsaraj group are focused on the equilibrium, kinetics and pathways of the chemical reactions between several GLVs and atmospheric oxidants in fog, and will be reported in future manuscripts. From the simulation point of view, electronic structure calculations such as DFT could in principle be used to study these chemical reactions (see, e.g., ref 76), although reaction equilibrium between these species could be explored using classical molecular simulation techniques such as reactive Monte Carlo.77,78 These calculations are beyond the scope of this manuscript and could be explored in later studies.

■

ASSOCIATED CONTENT

S Supporting Information *

Modified atomic charges used in our models of HxO and HxAc. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions ∇

T.P.L.-A. and Z.Z. contributed equally to this work.

Funding

This project was partially supported by a grant from the National Science Foundation (AGS 1106569), and by a grant from the Gulf of Mexico Research Initiative (GoMRI), as part of the Consortium for the Molecular Engineering of Dispersant Systems (C-MEDS, http://dispersant.tulane.edu/). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Mickael Vaitilingom (LSU) and Joseph DePalma (University of Delaware) for helpful discussions. High performance computational resources for this research were provided by High Performance Computing at Louisiana State University (http://www.hpc.lsu.edu), and the Louisiana Optical Network Initiative (http://www.loni.org).

■

REFERENCES

(1) Jacobson, M. C.; Hansson, H. C.; Noone, K. J.; Charlson, R. J. Organic atmospheric aerosols: Review and state of the science. Rev. Geophys. 2000, 38, 267−294. (2) Donaldson, D. J.; Valsaraj, K. T. Adsorption and reaction of trace gas-phase organic compounds on atmospheric water film surfaces: A critical review. Environ. Sci. Technol. 2010, 44, 865−873. (3) Poschl, U. Atmospheric aerosols: Composition, transformation, climate and health effects. Angew. Chem.Int. Ed. 2005, 44, 7520− 7540. 3033



Article

(40) Jain, S.; Zahardis, J.; Petrucci, G. A. Soft ionization chemical analysis of secondary organic aerosol from green leaf volatiles emitted by turf grass. Env. Sci. Technol. 2014, 48, 4835−4843. (41) Atkinson, R.; Arey, J. Gas-phase tropospheric chemistry of biogenic volatile organic compounds: a review. Atmos. Environ. 2003, 37, S197−S219. (42) Partition Coefficient (n-octanol/water): Shake Flask Method. OECD Guideline for the Testing of Chemicals No. 107OECD: Paris, France, 1995; pp 4. (43) Valsaraj, K. T.: Elements of Environmental Engineering, 3ed.; Taylor and Francis Publishers: New York, 2009, (44) Berendsen, H. J. C.; Grigera, J. R.; Straatsma, T. P. The missing term in effective pair potentials. J. Phys. Chem. 1987, 91, 6269−6271. (45) Jorgensen, W. L.; Maxwell, D. S.; TiradoRives, J. Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. J. Am. Chem. Soc. 1996, 118, 11225−11236. (46) Shirts, M. R.; Pitera, J. W.; Swope, W. C.; Pande, V. S. Extremely precise free energy calculations of amino acid side chain analogs: Comparison of common molecular mechanics force fields for proteins. J. Chem. Phys. 2003, 119, 5740−5761. (47) Hub, J. S.; Caleman, C.; van der Spoel, D. Organic molecules on the surface of water dropletsAn energetic perspective. Phys. Chem. Chem. Phys. 2012, 14, 9537−9545. (48) Bennett, C. H. Efficient estimation of free energy differences from Monte Carlo data. J. Comput. Phys. 1976, 22, 245−268. (49) Chen, B.; Siepmann, J. I. Partitioning of alkane and alcohol solutes between water and (dry or wet) 1-octanol. J. Am. Chem. Soc. 2000, 122, 6464−6467. (50) Chen, B.; Siepmann, J. I. Microscopic structure and solvation in dry and wet octanol. J. Phys. Chem. B 2006, 110, 3555−3563. (51) Kamath, G.; Bhatnagar, N.; Baker, G. A.; Baker, S. N.; Potoff, J. J. Computational prediction of ionic liquid 1-octanol/water partition coefficients. Phys. Chem. Chem. Phys. 2012, 14, 4339−4342. (52) Bussi, G.; Donadio, D.; Parrinello, M. Canonical sampling through velocity rescaling. J. Chem. Phys. 2007, 126, 014101. (53) Bussi, G.; Zykova-Timan, T.; Parrinello, M. Isothermal-isobaric molecular dynamics using stochastic velocity rescaling. J. Chem. Phys. 2009, 130, 074101. (54) Liyana-Arachchi, T. P.; Valsaraj, K. T.; Hung, F. R. Ice growth from supercooled aqueous solutions of benzene, naphthalene, and phenanthrene. J. Phys. Chem. A 2012, 116, 8539−8546. (55) Vacha, R.; Slavicek, P.; Mucha, M.; Finlayson-Pitts, B. J.; Jungwirth, P. Adsorption of atmospherically relevant gases at the air/ water interface: Free energy profiles of aqueous solvation of N2, O2, O3, OH, H2O, HO2, and H2O2. J. Phys. Chem. A 2004, 108, 11573− 11579. (56) Vacha, R.; Jungwirth, P.; Chen, J.; Valsaraj, K. T. Adsorption of polycyclic aromatic hydrocarbons at the air−water interface: Molecular dynamics simulations and experimental atmospheric observations. Phys. Chem. Chem. Phys. 2006, 8, 4461−4467. (57) Wick, C. D.; Chen, B.; Valsaraj, K. T. Computational investigation of the influence of surfactants on the air−water interfacial behavior of polycylic aromatic hydrocarbons. J. Phys. Chem. C 2010, 114, 14520−14527. (58) Chen, J.; Ehrenhauser, F.; Liyana-Arachchi, T. P.; Hung, F. R.; Wornat, M. J.; Valsaraj, K. T. Adsorption of gas-phase phenanthrene on atmospheric water and ice films. Polycycl. Aromat. Comp. 2011, 31, 201−226. (59) Liyana-Arachchi, T. P.; Valsaraj, K. T.; Hung, F. R. Molecular simulation study of the adsorption of naphthalene and ozone on atmospheric air/ice interfaces. J. Phys. Chem. A 2011, 115, 9226−9236. (60) Liyana-Arachchi, T. P.; Valsaraj, K. T.; Hung, F. R. Adsorption of naphthalene and ozone on atmospheric air/ice interfaces coated with surfactants: A molecular simulation study. J. Phys. Chem. A 2012, 116, 2519−2528. (61) Liyana-Arachchi, T. P.; Valsaraj, K. T.; Hung, F. R. Ice growth from supercooled aqueous solutions of reactive oxygen species. Theor. Chem. Acc. 2013, 132, 1−13.

(22) Claeys, M.; Graham, B.; Vas, G.; Wang, W.; Vermeylen, R.; Pashynska, V.; Cafmeyer, J.; Guyon, P.; Andreae, M. O.; Artaxo, P.; Maenhaut, W. Formation of secondary organic aerosols through photooxidation of isoprene. Science 2004, 303, 1173−1176. (23) Henze, D. K.; Seinfeld, J. H. Global secondary organic aerosol from isoprene oxidation. Geophys. Res. Lett. 2006, 33. (24) Hoffmann, T.; Warnke, J.: Organic Aerosols. In Volatile Organic Compounds in the Atmosphere; Koppmann, R., Ed.; Blackwell Publishing Ltd.: Oxford, UK, 2007; pp 342−387. (25) Larsen, B. R.; Di Bella, D.; Glasius, M.; Winterhalter, R.; Jensen, N. R.; Hjorth, J. Gas-phase OH oxidation of monoterpenes: Gaseous and particulate products. J. Atmos. Chem. 2001, 38, 231−276. (26) Lee, A.; Goldstein, A. H.; Kroll, J. H.; Ng, N. L.; Varutbangkul, V.; Flagan, R. C.; Seinfeld, J. H. Gas-phase products and secondary aerosol yields from the photooxidation of 16 different terpenes. J. Geophys. Res. 2006, 111, D17305. (27) Hallquist, M.; Wenger, J. C.; Baltensperger, U.; Rudich, Y.; Simpson, D.; Claeys, M.; Dommen, J.; Donahue, N. M.; George, C.; Goldstein, A. H.; Hamilton, J. F.; Herrmann, H.; Hoffmann, T.; Iinuma, Y.; Jang, M.; Jenkin, M. E.; Jimenez, J. L.; Kiendler-Scharr, A.; Maenhaut, W.; McFiggans, G.; Mentel, T. F.; Monod, A.; Prevot, A. S. H.; Seinfeld, J. H.; Surratt, J. D.; Szmigielski, R.; Wildt, J. The formation, properties and impact of secondary organic aerosol: Current and emerging issues. Atmos. Chem. Phys. 2009, 9, 5155−5236. (28) Hamilton, J. F.; Lewis, A. C.; Carey, T. J.; Wenger, J. C.; Garcia, E. B. I.; Munoz, A. Reactive oxidation products promote secondary organic aerosol formation from green leaf volatiles. Atmos. Chem. Phys. 2009, 9, 3815−3823. (29) Penuelas, J.; Llusia, J. Seasonal patterns of non-terpenoid C6C10 VOC emission from seven Mediterranean woody species. Chemosphere 2001, 45, 237−244. (30) Kirstine, W. V.; Galbally, I. E. A simple model for estimating emissions of volatile organic compounds from grass and cut grass in urban airsheds and its application to two Australian cities. J. Air Waste Manage. 2004, 54, 1299−1311. (31) Olofsson, M.; Ek-Olausson, B.; Ljungstrom, E.; Langer, S. Flux of organic compounds from grass measured by relaxed eddy accumulation technique. J. Environ. Monitor 2003, 5, 963−970. (32) Arey, J.; Crowley, D. E.; Crowley, M.; Resketo, M.; Lester, J. Hydrocarbon emissions from natural vegetation in California southcoast-air-basin. Atmos. Environ. 1995, 29, 2977−2988. (33) König, G.; Brunda, M.; Puxbaum, H.; Hewitt, C. N.; Duckham, S. C.; Rudolph, J. Relative contribution of oxygenated hydrocarbons to the total biogenic VOC emissions of selected Mid-European agricultural and natural plant-species. Atmos. Environ. 1995, 29, 861−874. (34) Joutsensaari, J.; Loivamaki, M.; Vuorinen, T.; Miettinen, P.; Nerg, A. M.; Holopainen, J. K.; Laaksonen, A. Nanoparticle formation by ozonolysis of inducible plant volatiles. Atmos. Chem. Phys. 2005, 5, 1489−1495. (35) Liyana-Arachchi, T. P.; Stevens, C.; Hansel, A. K.; Ehrenhauser, F. S.; Valsaraj, K. T.; Hung, F. R. Molecular simulations of green leaf volatiles and atmospheric oxidants on air/water interfaces. Phys. Chem. Chem. Phys. 2013, 15, 3583−3592. (36) Liyana-Arachchi, T. P.; Hansel, A. K.; Stevens, C.; Ehrenhauser, F. S.; Valsaraj, K. T.; Hung, F. R. Molecular modeling of the green leaf volatile methyl salicylate on atmospheric air/water interfaces. J. Phys. Chem. A 2013, 117, 4436−4443. (37) Arey, J.; Winer, A. M.; Atkinson, R.; Aschmann, S. M.; Long, W. D.; Morrison, C. L. The emission of (Z)-3-hexen-1-ol, (Z)-3hexenylacetate and other oxygenated hydrocarbons from agricultural plant species. Atmos. Environ. A-Gen. 1991, 25, 1063−1075. (38) Ohta, K. Emission of hexenol from higher-plants. Geochem. J. 1984, 18, 135−141. (39) Atkinson, R.; Arey, J.; Aschmann, S. M.; Corchnoy, S. B.; Shu, Y. H. Rate constants for the gas-phase reactions of cis-3-hexen-1-ol, cis-3hexenylacetate, trans-2-hexenal, and linalool with OH and NO3 radicals and O3 at 296 ± 2 K, and OH radical formation yields from the O3 reactions. Int. J. Chem. Kinet. 1995, 27, 941−955. 3034



Article

(62) Hess, B.; Bekker, H.; Berendsen, H. J.; Fraaije, J. G. LINCS: A linear constraint solver for molecular simulations. J. Comput. Chem. 1997, 18, 1463−1472. (63) Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald: An N⊕ log (N) method for Ewald sums in large systems. J. Chem. Phys. 1993, 98, 10089−10092. (64) Hess, B.; Kutzner, C.; van der Spoel, D.; Lindahl, E. GROMACS 4: Algorithms for highly efficient, load-balanced, and scalable molecular simulation. J. Chem. Theory Comput. 2008, 4, 435−447. (65) Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual molecular dynamics. J. Mol. Graph. Model. 1996, 14, 33−38. (66) Bhatnagar, N.; Kamath, G.; Chelst, I.; Potoff, J. J. Direct calculation of 1-octanol-water partition coefficients from adaptive biasing force molecular dynamics simulations. J. Chem. Phys. 2012, 137, 014502. (67) Sander, R.: Compilation of Henry’s law constants for inorganic and organic species of potential importance in environmental chemistry. Max-Planck Institute of Chemistry, Air Chemistry Department, 1999. (68) Karl, T.; Guenther, A.; Turnipseed, A.; Patton, E.; Jardine, K. Chemical sensing of plant stress at the ecosystem scale. Biogeosci. Discuss. 2008, 5, 2381−2399. (69) Karl, T.; Yeretzian, C.; Jordan, A.; Lindinger, W. Dynamic measurements of partition coefficients using proton-transfer-reaction mass spectrometry (PTR−MS). Int. J. Mass Spectrom. 2003, 223, 383− 395. (70) Liyana-Arachchi, T. P.; Zhang, Z. H.; Ehrenhauser, F. S.; Avij, P.; Valsaraj, K. T.; Hung, F. R. Bubble bursting as an aerosol generation mechanism during an oil spill in the deep-sea environment: Molecular dynamics simulations of oil alkanes and dispersants in atmospheric air/salt water interfaces. Environ. Sci.: Processes Impacts 2014, 16, 53−64. (71) Caleman, C.; Hub, J. S.; van Maaren, P. J.; van der Spoel, D. Atomistic simulation of ion solvation in water explains surface preference of halides. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 6838− 6842. (72) Yuet, P. K.; Blankschtein, D. Molecular dynamics simulation study of water surfaces: Comparison of flexible water models. J. Phys. Chem. B 2010, 114, 13786−13795. (73) Wemhoff, A. P.; Carey, V. P. Surface tension prediction using characteristics of the density profile through the interfacial region. Int. J. Thermophys. 2006, 27, 413−436. (74) Vega, C.; De Miguel, E. Surface tension of the most popular models of water by using the test-area simulation method. J. Chem. Phys. 2007, 126, 154707. (75) Ismail, A. E.; Grest, G. S.; Stevens, M. J. Capillary waves at the liquid−vapor interface and the surface tension of water. J. Chem. Phys. 2006, 125, 014702. (76) DePalma, J. W.; Horan, A. J.; Hall, W. A., IV Thermodynamics of oligomer formation: implications for secondary organic aerosol formation and reactivity. Phys. Chem. Chem. Phys. 2013, 15, 6935− 6944. (77) Johnson, J. K.; Panagiotopoulos, A. Z.; Gubbins, K. E. Reactive canonical Monte Carlo. Mol. Phys. 1994, 81, 717−733. (78) Turner, C. H.; Brennan, J. K.; Lísal, M.; Smith, W. R.; Johnson, J. K.; Gubbins, K. E. Simulation of chemical reaction equilibria by the reaction ensemble Monte Carlo method: A review. Mol. Simul. 2008, 34, 119−146.

3035


Water Interfaces: A Combined

Recommend Documents