Medium-Throughput Screening Method for Assessing Pervaporation Performance of Small Flat-Sheet Membrane Samples

Like most pervaporation research groups, we have wrestled with the question of how to screen membranes. The nature of the pervaporation bench-scale test system, normally with liquid nitrogen trapping of the vapor permeate, a sufficiently high feed liquid flow rate to avoid concentration polarization, and a start-up period to reach steady state, makes it difficult to test a large number of membranes in a short period of time. Evaluation of a single pervaporation membrane usually takes 1 day, more if duplicate experiments are run. Thus, screening multiple membrane materials is quite time consuming. Some research groups increase their testing throughput by daisy chaining multiple membrane cells together with a common feed stream, but individual permeate traps. Multi-sample test cells have been used for nanofiltration and gas separation testing.

High-throughput screening (HTS) methods have been developed in a variety of fields to speed identification and optimization of important process/material variables. HTS generally involves high-tech equipment for the preparation and evaluation of a large number of small samples. Relatively few HTS studies have been published in the membrane separation field. The difference between the normal means of preparing membrane materials and the means of testing those materials makes a fully HTS approach difficult, although not impossible, for membranes. Pervaporation presents challenges to HTS because of the phase change which occurs in the process and the potential for liquid and/or vapor boundary layer formation.

The objective of this project was to assess the potential of a low-tech, vial-based method to provide screening-level pervaporation performance information with a sample throughput rate significantly higher than that offered by traditional pervaporation tests. The pieces of equipment used for this method were: 

1. Standard 2 mL gas chromatography (GC) liquid sample vials with Aluminum crimp caps and Viton septa
2. Biological incubator shaker modified with a tray to hold the sample vials horizontally and to accept a dry-air purge. Provides sample agitation, temperature control, and dry/purged permeate environment.
3. Arch hole punches to create uniform holes in the vial septa and to create membrane disks to fit the vial caps
4. Instrument air system to provide dry air (dew point of -40 deg C)
5. Karl Fischer coulometric titrator for accurate and precise water determination.
6. Analytical balance with 0.0001 g precision

The method is based on the accurate and precise measurement of the mass loss from a vial and the concentration change in the liquid in the vial over a period of time. In theory, the rate of loss of each species from the vial is dependent on the permeance of the membrane and the driving force for transport. In this way, multiple vials can be evaluated at the same time. Vials can be prepared with different initial concentrations and different membrane materials. The temperature of the incubator shaker can be changed between sample sets to evaluate the effect of temperature, within the limits of the incubator. 

The method was evaluated for the separation of ethanol-water mixtures by pervaporation at 50 deg C at three water concentrations: 2, 5, and 20 wt%. The tests were performed on three water-selective membranes with significantly different performance characteristics. The membranes were unsupported films. Five replicates were prepared for each membrane and each initial water concentration to assess sample-to-sample variability. Water permeability was determined for all samples with good precision. The precision of ethanol permeability values was dependent on the duration of testing and water permselectivity. Leak-tight crimp sealing of the vials was critical for good results, particularly for membranes with a low water permeance or high water-ethanol permselectivity. Modifications to the method are being considered to improve permeance precision, particularly for the low permeance species, ethanol in our case.

 The method can be used for any liquid mixture in which concentrations in the liquid can be determined with a very high accuracy and precision.

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