Dokument

Summary:
This thesis presents an investigation of porous media by means of simulation techniques and morphological analysis. As a basis for the investigation throughout this work, we use three- dimensional (3D) images of porous structures obtained by imaging techniques, in particular, fo- cused ion beam scanning electron microscopy (FIB-SEM) and confocal laser scanning microscopy (CLSM) for macroporous space, and scanning transmission electron microscopy (STEM) to re- solve mesopores. A set of different morphological methods (chord length distribution (CLD), medial axis analysis (MAA), estimations of geometric, branch and diffusive tortuosities) are applied to capture averaged descriptors of the reconstructed porous samples. Because fluid dy- namics is inherent in many applications of porous media, several techniques are deployed to simulate the fluid dynamics in the reconstructions of porous media. This work includes four chapters that cover three different topics associated with the investigation of fluid dynamics in porous media. Each chapter also represents a separate journal publication. In the first chapter, we perform hydrodynamic dispersion simulations to study the morphology- flow relationship in physical reconstructions of particulate beds as well as in computer-generated packings of monosized spheres. A combination of lattice-Boltzmann and random-walk parti- cle tracking (RWPT) methods were utilized to simulate the flow and mass transport, respec- tively. Based on mean chord length μ and standard deviation σ estimated for CLD, we present a morphological descriptor, σ/μ, that can predict the longitudinal dispersion coefficient for any morphological configuration of packed beds. In the second chapter, we introduce the overall hindrance factor expression, H(λ), that de- scribes transport limitations in mesoporous space of random silica monoliths in dependence of λ, the ratio of solute size to mean pore size. The presented H(λ) is obtained through diffusion simulations of finite-size tracers applying the RWPT technique in three reconstructions of mor- phologically similar porous silica. The expression can also be utilized to assess the hindered diffusion coefficient for any material with similar morphology. In the third chapter, we adopt the lattice-gas mean field density functional theory (MFDFT) to virtually reproduce adsorption/desorption processes in a mesopore network of one of the monoliths from the second chapter. We demonstrate a good qualitative agreement of simulated boundary curves with experimental isotherms with an H2 hysteresis loop obtained for nitrogen at 77 K. We also use 3D images of the phase distribution that can be provided by MFDFT for any relative pressure value in the range 0 < p/p0 ≤ 1 to reveal the relation between hysteresis and phase distribution. In the fourth chapter, we continue using the concept of exploration of phase distribution and perform MFDFT modeling in physically reconstructed geometrical models of two ordered (SBA-15, KIT-6) and one random mesoporous silicas. We conduct a short parametric study of the MFDFT model to find optimal agreement with experimental isotherms in the hysteresis region. We also present simulated boundary curves for both ordered structures with a clear H1 hysteresis loop and for the disordered material with type H2(a) hysteresis. The phase distribution analysis as well as the shape of scanning curves reveal a highly heterogeneous morphology of the random silica. Hence, pore blocking and cavitation phenomena are identified and analyzed.

Das Dokument ist im Internet frei zugänglich - Hinweise zu den Nutzungsrechten