In order to advance overall understanding of solubilization phenomena within micelles, this work examines the equilibrium partitioning behavior of hydrophobic aroma compounds using headspace solid-phase microextraction (HS-SPME) techniques. The first focus of this research was to develop a robust analytical method to characterize d-limonene’s distribution between water and air over a 15–40˚C temperature interval. Vapor-water partition coefficients (Henry’s coefficients) were evaluated for limonene in well-sealed vials representing a closed system. Through these experimental measurements, we were also able to quantify the enthalpy of volatilization of limonene. In addition, we determined the solubility concentrations of decane, limonene, and 1-octanol, which collectively spanned four orders in magnitude. Second, HS-SPME was used to quantify micellar partitioning of limonene within aqueous mixtures containing diacyl, short-chain phosphatidylcholine with acyl chains of 6–8 carbons, as a function of temperature, solute concentration, and tail length. The three phospholipids studied were 1,2 dihexanoyl-sn-glycero-3-phosphocholine (diC6PC), 1,2 diheptanoyl-sn-glycero-3-phosphocholine (diC7PC), and 1,2 dioctanoyl-sn-glycero-3-phosphocholine (diC8PC). By varying phospholipid concentration above the appropriate critical micelle concentration, at fixed dilute limonene concentration, the micelle partition coefficient K_mw=x_(i,m)/c_(i,w) was evaluated using the pseudo-phase model, with x_(i,m) the mole fraction of solute i within micelles and c_(i,w) the solute concentration within the water continuum. Additional studies were conducted at fixed phospholipid concentration but varying limonene concentrations. For all lecithins studied, our findings show enhanced micellar partitioning occurred with increases in tail length, but reduced with an increase in temperature. In diC6PC or diC7PC micellar solutions, the partition coefficient was constant at lower limonene concentrations, indicating negligible effects of non-ideal solute interactions or solute induced micellar rearrangements. However, at higher limonene concentrations, K_mw increased with increasing solute. The effect of limonene on the phase behavior of the longest studied short-chain lecithin, diC8PC, was studied and it was observed that the addition of limonene resulted in the depression of diC8PC’s upper consulate temperature. In the last focus of this dissertation, the partitioning behavior of several aroma compounds within aqueous nonionic surfactant solutions was examined. These measurements were related to a solubilization framework grounded in fundamental thermodynamic principles. It was determined that solute molecular volume alone was a strong predictor of the maximum amount of solute that could be dissolved for a given concentration of surfactant. Comparing limonene partition coefficients for various types of micelles as measured by us or reported in literature, we find that K_mw falls in the range of 1–6 M–1, regardless of the surfactant structure. Measurements of solute vapor concentration were also used to determine the chemical potential of solute i within the micelle solution. Based on the thermodynamics of mixing, the chemical potential of studied mixtures could be related to the mole ratio of solute-to-surfactant and partition coefficient K_mw. For solutes limonene and octanol, the data indicated that K_mw was nearly constant at all molar ratios, consistent with a dilute pseudo-phase model. However, at higher mole ratios in which decane was the solute, we observed micelles lowered the chemical potential of solute below predictions of dilute theory. Thus, enhanced micellar partitioning occurred as the decane concentration was varied up to its solubility limit. Finally, through these measurements, the molar Gibbs energy of dissolution for 1-octanol, decane, and limonene was determined.

In this thesis, we examine diffusion in ternary, aqueous solutions of the nonionic surfactant decaethylene glycol monododecyl ether (C12E10) and a hydrophobic solute, either decane or limonene. In solution, the surfactant molecules self-assemble to form micelles swollen by hydrophobic solutes, with essentially no free hydrophobic solute or surfactant monomer in the surrounding solvent. The diffusive behavior of this system is very interesting in that surfactant-solute interactions are strong, and result in a highly non-diagonal diffusivity matrix [?], which depends in part on how strongly micelles grow with an increasing amount of solubilizate along the diffusion pathway. This behavior is distinct from that of colloidal dispersions comprised of polydisperse rigid hard particles, which are unable to reassemble on a molecular level to lower the system free energy as they diffuse. The goal of this work is to present experimental data and develop rigorous theoretical results that capture the influence of self-assembly on the ternary diffusion coefficient matrix [?], and on the time and static correlation functions that are commonly used to analyze light scattering data in these mixtures.

In Chapter 1, ternary diffusion coefficient matrices [?] and morphological parameters, such as the micelle aggregation number, hydrodynamic radius, and hydration index, were measured using the Taylor dispersion method and static and dynamic light scattering techniques, respectively, for C12E10/decane/water solutions. The matrix [?] for this system was found to be highly non-diagonal, and concentration dependent, over a broad domain of solute to surfactant molar ratios, and micelle volume fractions up to ? ≈ 0.25. Measurements for the average micelle radius and aggregation number indicate a weak dependence on the micelle volume fraction but a strong linear increase with the solute-to-surfactant molar ratio. Furthermore, a theoretical model, based on Batchelor’s theory for gradient diffusion in dilute, polydisperse mixtures of interacting spheres is developed and effectively used to predict [?] with no adjustable parameters. In this model, a Poisson distribution of solute molecules among micelles was assumed with a one-to-one correspondence between the number of solute to surfactant molecules distinguishing each micelle species.

In Chapter 2, experimental data for the ternary diffusion coefficient matrices [?] are presented for crowded ternary mixtures of C12E10 surfactant with either decane or limonene solute. Our theoretical model for [?], which was introduced in Chapter 1, is simplified by neglecting local polydispersity. Even though the model originates from dilute theory that incorporates pairwise hydrodynamic and thermodynamic interactions, the theoretical results were in surprisingly good agreement with experimental data for concentrated mixtures, with volume fractions up to ? ≈ 0.47. This agreement suggests that the effects of many-particle hydrodynamic and thermodynamic interactions cancel, resulting in experimental and theoretical predictions that are nearly linear over the entire range of concentration. In addition, the theory predicts eigenvalues ?− and ?+ that correspond to long-time self and gradient diffusion coefficients, respectively, for monodisperse spheres, in reasonable agreement with experimental data.

The third and final chapter of this thesis involves the development of model equations for the Rayleigh ratio and the mode amplitudes of the normalized electric field autocorrelation function, which are commonly used to analyze time averaged and photon correlation spectroscopy data, respectively. These theoretical results were derived using thermodynamic fluctuation theory applied to crowded solute-containing micellar solutions and microemulsions with negligible molecular species and polydispersity. This theory invokes nonequilibrium thermodynamics and enforces local equilibrium between molecular solute, surfactant, and the various micellar species, in order to support the influence of self-assembly on the light scattering functions for the first time. We find that micelle growth effects along the diffusion path in these mixtures, which were shown to drive strong multicomponent diffusion effects, expressed via the ternary diffusivity matrix [?], do not affect the scattering functions in the limit of zero local polydispersity. Hence, theoretical predictions for the Rayleigh ratio and the field autocorrelation function for ternary mixtures of solute-containing, locally monodisperse micellar solutions are identical to those developed for binary mixtures of monodisperse, colloidal hard spheres. However, micelle growth effects are predicted to influence the thermodynamic driving forces and eigenmodes for diffusion. In support of our theoretical results, measurements for the Rayleigh ratio and the field autocorrelation function for ternary aqueous solutions of decaethylene glycol monododecyl ether (C12E10) with either decane or limonene solute were performed for several molar ratios and volume fractions up to ? ≈ 0.25, and for binary mixtures of C12E10/water up to ? ≈ 0.5. Excellent agreement between our light scattering theory and experimental data is achieved for low to moderate volume fractions (? < 0.3) and at higher concentration when our volume fraction calculations are corrected to account for micelle dehydration.

The solubilization and retention of aromas in foods can be improved using phospholipid vesicle dispersions which protect these aromas from evaporation, degradation, and chemical reactions. In this work, a quantification method was developed for the solubilization and retention of volatile aromas in phosphatidylcholine vesicles, with headspace solid phase microextraction (HS–SPME) combined with gas chromatography and mass spectrometry (GC/MS). This method allows us to selectively sample volatile compounds in the vapor phase using a sorbent fiber. HS–SPME is fast, accurate and non-invasive and allows for in situ measurements of partition coefficients for local regions within food products. With this approach, limonene partition coefficients were determined both between vapor and water and between vesicle bilayers and water. The unilamellar, nearly monodisperse vesicles were very effective at solubilizing limonene molecules, with large bilayer–water partition coefficients of ~104 M–1. These values are 3–8 times larger than those observed for short chain phospholipid micelles, which have smaller core volumes. In addition, vesicles can solubilize limonene up to very high mole fractions x_lim≤0.8, where x_lim is the moles of limonene relative to moles of limonene + phospholipid in the bilayers. This maximum solubilization is much greater than that measured for short chain phospholipid micelles (x_lim≤0.5). The bilayer–water partitioning behavior of limonene was investigated as a function of phospholipid composition and vesicle size, using lecithin enriched in phosphatidylcholine and pure dimyristoylphosphatidylcholine (DMPC). There was no significant difference in extent of solubilization whether vesicles were made using DMPC or lecithin, despite the fact that DMPC has saturated 14-carbon fatty acid tails, while lecithin is predominantly unsaturated with 18-carbon fatty acid tails. With an increase in vesicle diameter from ~100 to 200 nm, there was a marginal increase in solubilization for DMPC and a marginal decrease in solubilization for lecithin.The vapor phase concentration of limonene was used to analyze the mixing thermodynamics through calculations of chemical potentials. For a wide range of limonene–lipid mole ratios, constant partition coefficient values could quantitatively capture solubilization in the bilayer, as predicted by ideal–dilute mixing theory for the solute. At high limonene–lipid mole ratios, however, we saw evidence of non-ideal behavior with higher measured lipid–water partition coefficients. Using regular solution theory with an interaction parameter (χ_(PL,L)) of ~1 predicted this partitioning behavior quite well, suggesting that preferential solute–solute interactions may enhance solubilization at higher limonene concentrations. With saturated phospholipids, high mole fractions of limonene in the bilayer lowered the gel–to–fluid lipid phase transition temperature through entropic contributions that could be predicted quite well with freezing point depression theory. Gel, gel/fluid coexistence, and fluid regions were observed over small to large limonene–lipid mole ratios. In the gel/fluid coexistence region, we observed constant vapor phase concentrations of limonene. At low mole fractions of limonene, formation of the gel phase decreased partition coefficients at least two-fold in comparison to the fluid phase. Above the phase transition temperature, decreases in temperature also lowered the partition coefficients, but to a much lower extent. Gel phase partition coefficients were used to modify the freezing point depression model and predict the solubilization behavior of limonene at all compositions at 15 and 20ºC. At these temperatures, experimental data agreed with the theory very well for low total limonene concentrations.