Mechanisms And Modeling Of Carbon Membrane Separation

Carbon membrane, as an important member of the inorganic membrane family, is prepared by pyrolysis of various carbonaceous materials. It possesses advantages of better thermal and chemical stabilities, wear resistance and well-defined stable pore structure. Therefore, carbon membrane relevant researches have gained a worldwide interest in liquid separation and gas separation. On the one hand, the carbon membrane has been experimentally demonstrated to be feasible and effective in separation of liquid mixture. On the other hand, the carbon membrane is recognized to have an attractive prospect in separation of gas mixture to displace current separation technology in this century. However, investigations on carbon membrane separation are in the lab-scale and technology push stage with less reports of industrial application. This is due to the lack of the understanding of carbon membrane separation mechanisms.In this dissertation, a systematic research on carbon membrane separation was proposed and performed including membrane fabrication, separation experiments, mechanism model development of liquid separation and molecular dynamics simulation of gas separation. Specifically,The coal-based tubular carbon membrane and the carbon composite membrane using the former as a support were prepared based on available procedures.Experiments was carried out on microfiltration of liquid mixture using TiO₂ as the representative of rigid particles. A mathematic model describing flux variations with time was developed introducing a variable cake porosity function through a mechanism analysis. Critical diameter, cake resistance, cake porosity and cake height variation with time were successfully obtained. The model predictions show excellent agreements with the experimental data, which confirms the practicability and validity of the model developed.Experiments was conducted on microfiltration of oily wastewater with oil droplet as the representative of deformable particles. Based on the modified Hermia's model, membrane fouling mechanisms were analyzed and resistance controlling factors of permeate flow were determined. Meanwhile, the implemental approach of the model developed for the rigid particle system to be used for the deformable particle system was discussed.Experiments was performed on permeation of pure CH₄ and CO₂. TheμVT ensemble nonequilibrium molecular dynamics (μVT-NEMD) simulation was then conducted using a slit pore model to investigate the effects of temperature and pressure on permeation of pure CH₄ and CO₂ so as to fix the appropriate Lennard-Jones interaction parameters according to the experimental data. Afterwards, the effects of pore size on the fluxes and wall potential energies of a binary mixture of CH₄ and CO₂ are numerically examined. The results show that the process of gas separation in carbon membrane is affected by various factors. Molecular sieving and selective adsorption are the dominating mechanisms of gas separation.A pore network model and a pore size distribution function were constructed. Using the of critical path analysis (CPA) method, a deep insight into the internal structure of pores was attained. With the binary mixture of CH₄ and CO₂ as a prototype, different pore size ranges of the species to be transported in membrane considering the effects of pore connectivity on gas separation were numerically determined based on the flux variation with pore size stimulated by the slit pore model. The membrane flux and its separation selectivity were compromised. And then, the pore size distribution curve was stimulated, which displays an excellent agreement with the experimental measurement, suggesting that the pore network model constructed can represent the separation process of CH₄ and CO₂ of the carbon membrane. At last, the pore network model constructed was applied to predict the separation process of a ternary system of CH₄/CO₂/H₂, the typical three components in coal-bed methane. Two-stage separation strategy was designed with different pore size range determined in each stage.