Separating azeotrope-forming solvent−water mixtures by conventional distillation poses technical, economic, and environmental challenges. Pervaporation and vapor permeation membrane technologies using water-permselective membranes provide an efficient alternative for water removal from solvents. We present here new water-selective materials, based on 1,2polybutadiene, that address two problems reported for traditional hydrophilic membrane materials under high water activities: swelling and hydrolysis. Exposure to UV radiation and/or heat converted portions of the vinyl groups in the polybutadiene to cross-links and hydrophilic functional groups, including alcohols, ketones, and carboxylic acids. In testing with a series of alcohols, such materials displayed high water permeabilities and stable performance over several months even at the extremes of 100% water, low water (2%), and an ethanol/water vapor at 115 °C and 2.5 bar.

The ability of homogeneous and mixed matrix membranes prepared using standard silicone rubber, poly(dimethylsiloxane) (PDMS), and fluorosilicone rubber, poly(trifluoropropylmethylsiloxane) (PTFPMS), to dehydrate ethanol by pervaporation was evaluated. Although PDMS is generally considered to be the benchmark hydrophobic membrane material in pervaporation, water/ethanol molar permselectivity of a pure PDMS membrane was found to be 0.89 for a feed containing 80/20 w/w ethanol/water at 50 C, indicating a slight selectivity for water. Fluorinated groups in PTFPMS improved the water-ethanol permselectivity to 1.85, but decreased the water permeability from 9.7 Â 10 À12 kmol Á m/m 2 Á s Á kPa in PDMS to 5.1 Â 10 À12 kmol Á m/m 2 Á s Á kPa (29,000 and 15,200 Barrer, respectively). These water permeabilities are attractive, particularly since the rubbery materials should not experience the steep declines in water permeability observed with most standard dehydration membranes as water concentration in the feed decreases. However, the water selectivity is lower than desired for most applications. Particles of hydrophilic zeolite 4A were loaded into both PDMS and PTFPMS matrices in an effort to boost water selectivity and further improve water permeability. Water-ethanol permselectivities as high as 11.5 and water permeabilities as high as 23.2 Â 10 À12 kmol Á m/m 2 Á s Á kPa were observed for the PTFPMS/zeolite 4A mixed matrix membranesÀ6 times higher than for the unfilled PTFPMS membrane.

BACKGROUND Many organic solvents form difficult‐to‐separate mixtures with water and have an affinity for water, making drying a potential reuse prerequisite. Pervaporation (PV) and vapor permeation (VP) membrane technologies hold promise for energy‐efficient solvent drying. Several water‐selective membrane materials are commercially available, but performance data is limited, particularly for two recently commercialized membrane materials: chabazite (CHA) and T‐type zeolites. In this work, commercial‐grade samples of CHA and T‐type membranes, along with a NaA zeolite membrane, were evaluated for the removal of water from ethanol. RESULTS The CHA sample had the highest initial PV water permeance (6820 GPU) and water permselectivity (3430) with 5 wt% water in ethanol at 50 °C. Initial NaA membrane performance was slightly lower (6060 GPU and 3260), while the T‐type membrane had the lowest initial permeance and selectivity (4260 GPU and 1090). Performance declined over time, most notably for the NaA membrane, for which water permeance fell over 50% through 39 days of testing. The T‐type membrane exhibited the steadiest PV water permeance, but the most variable ethanol permeance. CONCLUSION The PV performance of the three membranes largely overlapped the predicted range for T‐type membranes. That performance generally exceeds the anticipated ethanol drying performance of non‐zeolitic PV membranes but is less than that predicted for NaA and CHA membranes. The present CHA membrane results, along with other recent reports, refine earlier predictions of the ethanol dehydration performance of that type of zeolite. The changing performance with time should be understood to properly design a solvent dehydration system. © 2022 Society of Chemical Industry (SCI). This article has been contributed to by US Government employees and their work is in the public domain in the USA.