The focus of this thesis was on the mechanisms of release of volatile molecules from smectite clay minerals. The weight loss was monitored by thermogravimetry, differential calorimetry, and mass spectrometry under isothermal conditions and under ramp temperature conditions. Modeling of the clay-volatile release systems was performed with the help of finite element method calculations at the macroscopic level and ab initio calculations at the molecular level. A first comparison between finite element method simulations and experimental data for the evaporation of bulk volatile liquids under a convective gas flow showed the calculations to be in good agreement with experiments. Results of the simulations were used to develop a semi-analytical model explaining the dependence of the evaporation on the total pressure, the carrier gas flow, the temperature, and material constants. In particular, its temperature dependence could be approximated to good accuracy by an Arrhenius-type equation derived from the semi-analytical model. Differential calorimetry measurements of the heats of vaporization showed equilibrium conditions at the surface of the liquids to be satisfied. The same approach was extended to the release of model volatiles (water, ethanol, ethyl acetate and toluene) from smectite clays. At high coverage, the release was found to be close to that for the bulk liquids. Its decrease with time followed the behaviour observed in the respective curves of the gas/condensed phase partition coefficients, the equilibrium desorption isotherms. Equilibrium condition at the surface of the sample was evidenced by a comparison between the measured heats of vaporization and the equilibrium desorption isotherms. The differences observed in the measured equilibrium desorption isotherms of the volatile on the smectite clays could be rationalized by the use of ab initio calculations of the binding of the volatiles on the surface of a sodium smectite clay. At low coverage, the differences could be attributed to their differences in binding energies with the clay counter ions, which could be explained in terms of the chemical nature of the interacting species. At high coverage, the differences could be related to the properties of the bulk liquids, because high coverage corresponds to high activities of the volatiles on the clay samples. The differences observed in the measured rates of release of the volatiles from the smectite clays could be rationalized by the use of finite element calculations and the semi-analytical model developed for the bulk liquids, taking as input parameters the measured equilibrium desorption isotherms, that determines the volatile gas phase concentration at the surface of the sample. The slower rates of release of ethanol and ethyl acetate from the smectite clay, compared to water, could be explained from their differences in their equilibrium desorption isotherms from the clay, though diffusion effects also played a minor role. However, the slower rate of release of toluene, compared to water, could only be explained by its slow diffusion in the smectite clay. Diffusion effects were found to be enhanced with an increase in the size of the aggregate, such that, in addition to the binding strength of the counter ions, the swelling capacity of the clay, allowing the clay to clump, was also found to be an important factor for the controlled release of volatiles from the clay. Ion-exchange of the clay with lithium cations was found to be optimal in terms of ionic strength of the cation and swelling capacity, as compared to other metallic cations and small organic cations. This clay modification was tested on the release of malodorous compounds determined in cat urine by mass spectrometry and sniffing experiments. The smectite clay modified by ion-exchange with lithium cations showed improved controlled release of volatile organic compounds found in cat urine, compared to the sodium smectite, and improved clumping capacity compared to the calcium smectite. The calcium smectite clay used by Nestlé Purina could be modified by column exchange with lithium cations in high quantities. However the potential toxicity of the lithium is unknown. Additional information was obtained by ab initio calculations on the interaction of water with the smectite. The calculations of the binding of one to six water molecules on the surface of a sodium clay showed the water molecules bind to the cation and form hydrogen bonds between each other and with the clay oxygens, leading to the formation of different configurations for the same number of water molecules on the clay surface. With an increase in the number of water molecules on the surface of the clay, intermolecular interaction between the water molecules became gradually more important than sodium-water interactions. The binding energy was also found to be dependent on the location of the charge substitution in the clay (tetrahedral or octahedral) and to the formation of hydrogen bonds between the water molecule and a basal oxygen linked to a tetrahedral replacement. With an increase in the number of water molecules bound to one counter ion, the sodium counter ion was found to move away from the clay surface. The dynamic calculations of water molecules in the clay interlayer showed both the water molecules and the cations to be mobile. The average diffusion coefficient of the water molecules was reduced compared to that of bulk water, due to restrictions imposed by both the clay platelets and the counter ions. The structure of the water molecules showed to be similar to that of bulk water with a preferential orientation of their hydroxyl group almost perpendicular to the clay surface. The difference in the state of order of the water molecules around the sodium cation was interpreted as a symptom of hysteresis as observed in the measured equilibrium sorption isotherm.
EPFL2009
