- 8:30 AM

Selective Recovery of Ethanol from Water by Pervaporation: Zeolite-Silicone Rubber Mixed Matrix Membranes

Vasudevan V. Namboodiri, Travis C. Bowen, and Leland M. Vane. U.S. EPA, 26 W. Martin Luther King Drive, M/S443, Cincinnati, OH 45268

The production of fuel-grade ethanol from renewable resources, such as biomass, is gaining attention due to the phase out of methyl t-butyl ether (MTBE) as a fuel oxygenate, national security issues related to non-domestic sources of fuels, and the effect of fossil fuel combustion on earth's climate. Most biomass-to-ethanol conversion processes involve the fermentative production of ethanol from sugars released from the biomass. Depending on the biomass source and hydrolysis procedures, the concentration of ethanol in the fermentation broth can range from 1 to 15 wt%. Thus, in order to produce fuel-grade ethanol, the water content must be reduced to less than 1.3 wt% water. Distillation is the traditional technology for performing the bulk separation of ethanol from these dilute biomass fermentation broths with molecular sieve adsorption used to meet the 1.3 wt% water specification for the product. Pervaporation with water-selective membranes competes with molecular sieves to meet fuel water limits. In addition, pervaporation with ethanol-selective membranes, the subject of this paper, is an alternative to distillation and may have energy and capital cost advantages, especially for smaller-scale systems [1].

Benchmark hydrophobic membranes for ethanol removal from water are constructed with silicone rubber (polydimethylsiloxane - PDMS) as the selective layer. Silicone rubber has a higher flux and ethanol-water selectivity than most other hydrophobic ethanol-permeable polymers. Unfortunately, this separation performance is slightly less than that offered by a single vapor-liquid equilibrium stage. Thus, in order to improve energy efficiency, higher selectivity membranes are required. Pure zeolite membranes (specifically silicalite-1) have the highest ethanol-water separation factor reported – over 10 times that of silicone rubber, but inorganic membranes are more expensive and difficult to produce than polymer membranes. Mixed matrix membranes, in which zeolite particles are dispersed in a polymer matrix, deliver separation performances intermediate to that of the two materials but with a cost and ease of fabrication comparable to polymer membranes. If mixed matrix membranes with adequate efficiency could be produced, then energy efficiency needs and cost demands can both be met.

The ability of a pervaporation membrane to separate species 1 from species 2 is generally reported as the separation factor, A12, where: A12 = (Y1/Y2)/(X1/X2) and Xi and Yi are the feed and permeate compositions of component i, respectively. The reported ethanol-water separation factor for “pure” silicone rubber membranes ranges from 4.4 to 14.4 with an average of about 8 [1]. Based on a review of the pervaporation literature, inorganic membranes with a ZSM-5 zeolite selective layer in which the Si:Al ratio is high, particularly silicalite-1, are capable of delivering separation factors up to 125 [1]. Mixed matrix membranes consisting of high silica ZSM-5 zeolite particles dispersed in silicone rubber have been shown to deliver ethanol-water separation factors as high as 59 [1]. However, most other reported separation factors for such systems were between 10 and 15 – marginally better than those of silicone rubber alone.

The objective of this study was to review preparation methods reported in the literature for zeolite-silicone rubber mixed matrix membranes and to identify material characteristics and procedural changes which might result in improved ethanol-water separation performance. Membranes prepared following these procedures were tested with a benchmark aqueous solution containing 5 wt% ethanol. Membrane variables studied included: molecular weight of the vinyl-PDMS chain, the chain length of the crosslinking agent, the density of reactive hydride groups on the crosslinking agent chain, and the ratio of hydride to vinyl groups in the system. Additionally, the source and properties of the zeolite particles, zeolite loading, and zeolite pretreatment procedures were investigated. The benchmark particle loading was 50 wt% high-silica ZSM-5 zeolite particles. At this loading, ethanol-water separation factors ranging from 24 to 36 were observed. Increasing particle loading to 60 wt% pushed the separation factor over 40. The ramifications of these separation factors on the energy required for several ethanol-water separation scenarios will be discussed and compared to that of distillation.

[1] L.M. Vane, A review of pervaporation for product recovery from biomass fermentation processes, J. Chem. Technol. Biotechnol., 80 (2005) 603-629.

*This is an abstract of a proposed presentation and does not necessarily reflect U.S. EPA policy.