(430a) Effect of Bacteria and Virus Shape on Rejection by Microfiltration Membranes: Comparison of Experiment with Hindered Transport Theory

Microfiltration (MF) and ultrafiltration (UF) membranes are increasingly employed in the food processing, biotechnology, and pharmaceutical industries, for water and wastewater treatment, etc. Preventing microbial contamination of the filtered water (physical disinfection and sterile filtration) is integral to these applications. To date, several studies have empirically demonstrated the incomplete rejection of bacteria and viruses by microfilters. However, quantitative analyses of colloid passage across MF membranes have predominantly focused on spherical colloids, even though rod-shaped bacteria and non-spherical viruses are frequently encountered in membrane feed waters. Detailed information on removal of non-spherical colloids by membranes is not generally available. The objective of this research was to experimentally measure rejection of various organism with different aspect ratio (=length/diameter) from microfiltration membranes and to compare experimental observations to theoretical predictions. The rejection of two rod-shaped bacteria, Brevundimonas (formerly Pseudomonas ) diminuta and Serratia marcescens, two bacterial viruses (PRD1 and T4), and spherical silica colloids from clean track-etched membranes was measured. The silica particles ranged in size from ~ 0.2 micron to 0.5 micron with a very narrow size distribution. The PRD1 bacteriophage was nearly spherical in shape, with diameter of 0.08 micron. The use of different growth periods provided us with two different B. diminuta samples, one with aspect ratio ~ 4; the other considerably longer with aspect ratio ~ 9. The S. marcescens bacteria were comparable in size to the B. diminuta , with an aspect ratio ~ 3.5. The T4 bacteriophage has six contractile tail fibers, a head and a tail. The aspect ratio of the head is ~ 2; considering the ratio of the head plus tail length to the head diameter gives an aspect ratio of ~ 2.7. Millipore polycarbonate membranes with pore sizes ranging from 0.1 micron to 5 micron were used for all rejection measurements. Pore sizes were determined using porometry as well as scanning electron microscope imaging. A stirred cell operating at high Peclet number conditions was used for rejection experiments, with samples collected before the membrane was fouled. Results obtained with spherical silica colloidal particles as well as with the PRD1 bacteriophage were in good agreement with hindered transport theory developed for rigid spherical particles1, confirming the validity of our pore size determinations and our experimental methods. Experimental results obtained with the non-spherical particles were quantitatively compared to predictions from hindered transport theory developed for rod shaped particles.2 This theory considers only configurational effects on particle transport, assuming that all possible particle orientations in the pore are equally probable. This theory predicts that the rejection coefficient for a rod shaped particle is greater than that for a spherical particle when particle size is described using a volume equivalent pore diameter, with rejection coefficient increasing as particle aspect ratio increases. In contrast to theoretical predictions, experimental results for the microorganisms with different sizes and aspect ratios were in general agreement with the prediction for a spherical particle when the volume equivalent sphere diameter is used to characterize particle size. When rod diameter is used to characterize particle size, rejection was found to be greater than the prediction for a sphere with the same diameter. Our results indicate that rod shaped particles are transported through the pore with a trajectory that is biased towards an ?end on' orientation, indicating that all particle orientations are not equally sampled. The unexpected conclusions are that the important particle size parameter appears to be the volume equivalent sphere diameter and that results show similar rejection behavior for particles with different aspect ratios. The ?take home' message from this work is that one can safely design a filtration system for microbial decontamination by using the rod diameter as the characteristic particle size parameter when selecting appropriate membrane pore size.

1 Bungay, P.M. and Brenner, H., ?The Motion of a Closely Fitting Sphere in a Fluid-Filled Tube,? Int. J. Multiph. Flow , 1, 25 (1973). 2 Anderson, J.L., ?Configurational Effect on the Reflection Coefficient for Rigid Solutes in Capillary Pores,? J. Theor. Biol. , 90, 405 (1981).