| Summary: | The objective of this dissertation is to develop various multiscale modeling methods for chemical reactors. This objective stems from the demand for more accurate and predictive reactor models or the need for miniaturization of materials. The wide disparity in scales encountered in practical industrial reactors requires a hierarchical approach with a different model for each scale along with methods for coupling these models. These multiscale models entail dynamic, bi-directional coupling of quantum information with molecular simulations and continuum deterministic reactor models. The feasibility of such multiscale, hybrid models is demonstrated in a model system, that of the catalytic oxidation of CO and H2 on a Pt single crystal embedded in a continuous stirred tank reactor. Emphasis is placed on surface processes, such as lateral adsorbate-adsorbate interactions, proximity effects encountered in all bimolecular events, and surface diffusion. Significant differences in model responses have been observed between multiscale models and continuum, mean field based models when the dominant surface species is immobile (e.g., oxygen). Since multiscale models for realistic systems are currently semi-quantitative, a multistep optimization methodology has been introduced and successfully applied to the catalytic oxidation of CO on Pt that enables to refine model parameters. It has been found that the lower level continuum, mean field model can be used for preliminary optimization steps, such as identification of the important kinetic parameters and generation of initial estimates. The proper orthogonal decomposition technique has been used to obtain low dimensional approximations of such multiscale models describing epitaxial growth of materials as an example. Both the dynamics of a stagnation fluid phase and the surface morphology are successfully described by reduced mathematical models. Towards a more practical application, flame propagation in natural gas microburners is explored using a detailed two-dimensional model with emphasis on interfacial gas-phase phenomena. While radial gradients and gas-surface temperature discontinuity are found to be unimportant, it is shown that the critical quenching diameter strongly depends on the initial heat loss and radical wall quenching. Furthermore, catalytic microburners appear to be a more promising choice as compared to their homogeneous counterpart.
|
|---|
