(333b) Modeling and Simulation of Dynamic Bio-Ethanol Adsorption

The recovery of ethanol from fermentation broth is a major economic limiting step which prevents the wide-spread use of bio-ethanol as a gasoline additive or fuel alternative. While distillation is effective in separating the ethanol from the broth to about 85%, it is not energy efficient.

An auxiliary adsorption system for the enhancement of bio-ethanol production is believed to have economic potential. The proposed external adsorption system would treat the bioreactor effluent following the fermentation. The goal of the system is to selectively adsorb ethanol from the solution as it flows through the column. The process is made cyclical by adding parallel columns and alternating between adsorption and desorption cycles. If the adsorbent is highly selective in its preference for ethanol, the desorbed solution could be highly concentrated and thus the downstream energy requirement could be reduced.

To implement the proposed system, it is important to first understand the ethanol mass transfer which takes place between the liquid and solid phases in the column during the adsorption cycle. In this study, dynamic column experiments were performed on binary ethanol-water solutions with various activated carbon adsorbents in order to monitor the mass transfer. The effect of the particle size on the adsorption mass transfer is considered.

A mass transfer model, solved by finite difference method, is proposed to simulate the adsorption of ethanol from a binary aqueous solution. The model incorporates three rate controlling steps: external film mass transfer, intra-particle diffusion and adsorption site kinetics. The progressive development of this three-parameter model is discussed. These simulation results are compared to the dynamic experiments carried out to improve the models used for simulation of the adsorption column..

Results indicate that particle size plays a major role in the mass transfer of ethanol from the liquid to the solid phase. For large particles mass transfer becomes diffusion controlled requiring the two- or three-parameter model to adequately represent the experimental data, whereas for smaller particles, the results of the dynamic column experiments are modelled with good agreement with a single parameter.