Measuring Pore Size in Single Hollow Fiber Membranes by Capillary Flow Porometry
By Richard Wenman and Martin A. Thomas, Quantachrome Instruments


Figure 1. Sample mounted and glued in acrylic sample holder. The arrow shows sealed fiber end.

Hollow fiber and capillary membranes have a number of attractive properties such as flexibility, high surface area per unit volume, and unique packaging opportunities, which means they can be adapted to a variety of filtration applications. However, their physical form — small diameter flexible polymeric tubes — is a challenge for determining the sizes of the largely sub-micron pores penetrating the walls. But a simple yet reliable preparation procedure combined with state-of-the-art laboratory equipment can lead to successful characterization and differentiation of single fibers.

The excellent mass-transfer properties of the hollow fiber construction (a relatively large lumen surrounded by a large area of a thin porous membrane wall) has led it to being utilized in numerous commercial applications in widely differing fields such as medical (blood fractionation), water reclamation (purification and desalination), gas separation, and techniques using pervaporation. Other promising applications of this type of membrane are in the biochemical industry (bioseparation and bioreactors). Specifically, its beneficial features (compared with more traditional filtration and separation systems) are modest energy requirements, high volume efficiency, two modes of operation ("inside-out" and "outside-in"), and low operating costs. To a certain extent these benefits are offset by more-frequent fouling and initial capital expense. The challenge faced by those needing to determine the pore size distribution through the walls is to find a technique which can functionally transfer a fluid radially, even through a single narrow fiber; making pore size measurements across a flat sheet, even a bundle of fibers, is simple by comparison. The difficulty of analyzing hollow fibers one at a time has been overcome by a special preparation technique, which involves sealing an individual fiber into a special sample holder.


Figure 2: External Sample Manifold The red arrow indicates the attachment point for the acrylic sample holder. The cylinder is a buffer volume.

EXPERIMENTAL
Preparing a Single Hollow Fiber
Acrylic tube sample holders, 8mm O.D. and 30mm long were modified by drilling the bore to 1.5mm and countersinking each end. Four samples of polymeric hollow fiber samples identified as samples 1, 2, 3, and 4, each having an outside diameter of 1 mm and a wall thickness of 100 µm were prepared as follows: a length of each fiber was cut and "opened" by inserting a TFE coated wire down the lumen (center of the tube). This wire remained in place during the following procedure to mount the hollow fiber in the holder: the sample was glued and sealed inside the acrylic using quick-drying epoxy resin. The loose end of the fiber was also sealed with the glue. After the glue had started to harden in 2 or 3 minutes, the wire was removed.

ANALYSIS
Pore size distributions were determined using a Porometer 3Gz (Quantachrome Instruments, Boynton Beach, FL USA). The completed sample holders were installed in an external sample manifold (see Figure 2) in place of the usual sample holder assembly and block. After wetting a sample with PorofilT wetting fluid (Quantachrome) up to 256 data points were measured over the selected pressure / pore size range. The 3G z was equipped with both 10 and 100 ml/min sensors, and both ranges were used.

DISCUSSION
Measured data of flow versus pressure for wet and dry runs for all four samples are shown in Figure 3. The mean flow pore sizes were calculated in the usual manner, at the pressure intersection of half the dry flow data with the wet flow curve.
The pore size distributions were calculated from the measured pressure by the Washburn equation, assuming a zero contact angle. Three of the four samples (1, 3, and 4) have pores predominantly smaller than 0.5 micron diameter as immediately evidenced by liquid expulsion pressures spanning the range from 1 to 6 bar. The pores of the fourth sample (2) empty completely below 0.7 bar. The resulting pore size distributions reveal that sample 2 has a very narrow distribution of relatively large pores - a half width of little more than 10% of modal value - and showing slight skewing towards smaller sizes on the logarithmic plot. In contrast, the distributions of pores in the other three samples are rather broad, half widths up to 50% of nominal. Sample 1 appears to be effectively symmetrical on the logarithmic plot, whereas samples 3 and 4 show significant skewing towards larger sizes than the mode.


Figure 3: Measured “wet” and “dry” flow versus pressure. The differences between all the samples are already evident.

Figure 4: Differential percent flow versus pore size. Differences between sub-micron pores are highly resolved.

Figure 5: Differential pore number percent area versus pore size. Some significant similarities between samples 3 and 4 can now be seen.


Number distributions were calculated based on the internal geometric surface of the sample fibers (from gross dimensions). Samples 3 & 4 show even greater similarity when the pore number distributions are compared (Figure 5) in spite of their very different flow characteristics and the very evident bimodal nature of sample 4. The multimodal nature of sample 1 is also pronounced in Figure 5.

CONCLUSIONS
A new sample preparation technique has been successfully employed to demonstrate the sub-micron resolution of capillary flow porometry on what are usually considered to be "difficult" samples. This ability to measure pore size distributions in the (membrane) wall of a short sample of a single hollow fiber is of significant value to manufacturers and users of these materials, as it allows them to study in detail the consistency of membranes pores discretely along the length of a fiber.
Richard Wenman is Quantachrome's porometer specialist. He worked for Coulter Corporation for many years developing their first and second generation porometers. He subsequently formed the Xonics company to develop the third generation of porometers, which he continued at WSI before joining Quantachrome in 2010.
Martin A. Thomas has more than 28 years experience in the characterization of porous materials, the last 20 of those have been with Quantachrome. He is also a technical specialist but shares much of his time in market development especially for new products and applications..

For more information contact:
Quantachrome Instruments | 1900 Corporate Drive, Boynton Beach | FL 33426 USA
Tel: 1-561-731-4999
Email: qc.support@quantachrome.com | Website: www.quantachrome.com