| Summary: | Ethanol production through biomass fermentation is one of the major technologies available to produce liquid fuel from renewable energy sources. A major problem associated with the production of ethanol through fermentation remains the inhibition of the yeast Saccharomyces cerevisiae by the produced ethanol. Currently high water dilution rates are used to keep the ethanol concentrations in the fermentation broth at low concentrations, resulting in low yields and increased downstream processing to remove the excess water. Yeast strains that have a high tolerance for ethanol have been isolated but the time and cost associated with doing so poses a challenge.
The fermentation process can be combined with pervaporation, thereby continuously removing ethanol while it is being formed. In this study a mathematical model for ethanol fermentation with yeast, Saccharomyces cerevisiae, coupled with pervaporation was developed. The fermentation of glucose was optimised in the first part of the study and experimental data were obtained to find a kinetic model for fermentation. It was found that an optimum ethanol yield can be obtained with an initial glucose concentration of 15wt%, a yeast concentration of 10 g.L–1, and a pH between 3.5 and 6. The maximum ethanol yield obtained in this study was 0.441g.g–1 (86% of the theoretical maximum) using 15wt% glucose, 10g/L yeast and a pH of 3.5.
Two kinetic models for fermentation were developed based on the Monod model. The substrate–limiting model, predicted fermentation very accurately when the initial glucose concentration was below 20wt%. The second model, the substrate–inhibition model, predicted fermentation very well when high initial glucose concentrations were used but at low glucose concentrations, the substrate–limiting model was more accurate. The parameters for both models were determined by non–linear regression using the simplex optimisation method combined with the Runge–Kutta method.
The PERVAP®4060 membrane was identified as a suitable membrane in this study. The effect of the ethanol content in the feed as well as the influence of the glucose content was investigated. The total pervaporation flux varied with ethanol content of the feed and the highest total flux of 0.853 kg/m2h was obtained at a feed with 20wt% ethanol. The addition of glucose had almost no effect on the ethanol flux but it lowered the water flux, thereby increasing the enrichment factor of the membrane.
The mass transport through the PERVAP®4060 membrane was modelled using the solution–diffusion model and Greenlaw’s model for diffusion coefficients was used. The limiting diffusion coefficient (Di0) and plasticisation coefficients (Bij) were determined by using the Nelder–Mead simplex optimisation method. The theoretical values predicted with the model showed good agreement with the measured experimental values with R2 values above 0.998.
In the third part of this investigation, the kinetic model developed for fermentation was combined with the transport model developed for pervaporation. The combined kinetic model was compared to experimental data and it was found that it could accurately predict fermentation when coupled with pervaporation. This model can be used to describe and better understand the process when fermentation is coupled with pervaporation. === Thesis (M.Ing. (Chemical Engineering))--North-West University, Potchefstroom Campus, 2012.
|
|---|
