Ozone-Surface Interactions: Investigations of Mechanisms, Kinetics, Mass Transport, and Implications for Indoor Air Quality

In this dissertation, I present the results of laboratory investigations and mathematical modeling efforts designed to better understand the interactions of ozone with surfaces. In the laboratory, I exposed carpet and duct materials to ozone and measured ozone uptake kinetics and the ozone induced emissions of volatile organic compounds. To understand the results of my experiments, I developed mathematical models to describe dynamic indoor aldehyde concentrations, mass transport of reactive species to smooth surfaces, the equivalent reaction probability of whole carpet due to the surface reactivity of fibers and carpet backing, and ozone aging of surfaces.

Carpets, separated carpet fibers, and separated carpet backing all tended to release aldehydes when exposed to ozone. Secondary emissions were mostly n-nonanal and several other smaller aldehydes. The pattern of emissions suggested that vegetable oils may be precursors for these oxidized emissions. I discuss several possible precursors and experiments in which linseed and tung oils were tested for their secondary emission potential. Dynamic emission rates of 2-nonenal from a residential carpet may indicate that intermediate species in the oxidation of conjugated olefins can significantly delay aldehyde emissions and act as reservoir for these compounds. The ozone induced emission rate of 2-nonenal, a very odorous compound, can result in odorous indoor concentrations for several years.

Surface ozone reactivity, is a key parameter in determining the flux of ozone to a surface, is parameterized by the reaction probability, which is simply the probability that an ozone molecule will be irreversibly consumed when it strikes a surface. In laboratory studies of two residential and two commercial carpets, I determined the ozone reaction probability for carpet fibers, carpet backing and the equivalent reaction probability for whole carpet. Typically reaction probability values for these materials were 10^-7,10^-5, and 10^-5 respectively.

To understand how internal surface area influences the equivalent reaction probability of whole carpet, I developed a model of ozone diffusion into and reaction with internal carpet components. This was then be used to predict "apparent" reaction probabilities for carpet. I combine this with a modified model of turbulent mass transfer developed by Liu, et al. to predict deposition rates and indoor ozone concentrations. The model predicts that carpet should have an equivalent reaction probability of about 10^-5, matching laboratory measurements of the reaction probability.

For both carpet and duct materials, surfaces become progressively quenched ("aging"), losing the ability to reactor otherwise take up ozone. I evaluated the functional form of aging and find that the reaction probability follows a power function with respect to the cumulative uptake of ozone. To understand ozone aging of surfaces, I developed several mathematical descriptions of aging based on two different mechanisms. The observed functional form of aging is mimicked by a model which describes ozone diffusion with internal reaction in a solid. I show that the fleecy nature of carpet materials in combination with the model of ozone diffusion below a fiber surface and internal reaction may explain the functional form and the magnitude of power function parameters observed due to ozone interactions with carpet.

The ozone induced aldehyde emissions, measured from duct materials, were combined with an indoor air quality model to show that concentrations of aldehydes indoors may approach odorous levels. I show that ducts are unlikely to be a significant sink for ozone due to the low reaction probability in combination with the short residence time of air in ducts.