Wednesday, May 7, 2008 - 9:00 AM
9-03

Energy efficient recovery and dehydration of ethanol from fermentation broths by Membrane Assisted Vapor Stripping technology

Leland M. Vane1, Franklin R. Alvarez1, Ivy (Yu) Huang2, and Richard W. Baker2. (1) National Risk Management Research Laboratory, U.S. Environmental Protection Agency, 26 W. Martin Luther King Dr., M/S443, Cincinnati, OH 45268, (2) Membrane Technology & Research, Inc., 1360 Willow Rd., Suite 103, Menlo Park, CA 94025-1516

Distillation combined with molecular sieve dehydration is the current state of the art for fuel grade ethanol production from fermentation broths.  To improve the sustainability of bioethanol production, energy efficient separation alternatives are needed, particularly for lower fermented ethanol concentrations and smaller scale systems.  A process, termed Membrane Assisted Vapor Stripping (MAVS), integrating steam stripping with a vapor compression step and a vapor permeation membrane separation step is proposed.  The stripping column provides high ethanol recoveries and low effluent concentrations; the vapor compression-vapor permeation membrane component enables the efficient recovery of latent and sensible heat from both the ethanol-enriched retentate and water-enriched permeate streams from the membrane system.  Based on computer simulations, recovery of 1 kg of fuel grade ethanol (0.5 wt% water) from a 1 wt% ethanol feed stream can be achieved using 8.9 MJ of fuel-equivalent energy.  For a 5 wt% ethanol feed stream, only 2.5 MJ are needed.  This represents an energy savings of at least 43% relative to standard distillation producing azeotropic ethanol (7 wt% water).  To produce a 99.5 wt% ethanol product from a broth containing 11.5 wt% ethanol, a conventional distillation/mol sieve system is expected to use 5.5 MJ-fuel equiv/kg-ethanol; the MAVS process would require as little as 2.2 MJ-fuel equiv./kg-ethanol.  Overall costs are also estimated to be lower for a MAVS system.  Process simulation results and the results from ongoing pilot-scale validation tests will be reported.