Observation of deposition and removal behaviour of submicron bacteria on the membrane surface during crossflow microfiltration

Abstract

The deposition and removal of submicron bacteria (SW8) has been examined by the direct observation through the membrane (DOTM) technique. The original DOTM was modified to incorporate fluorescence microscopy to better visualise the submicron material. The flux at which deposition commenced, the so-called critical flux, was visually identified before the transmembrane pressure indicated cake formation. After supercritical operation for about 15 min the flux was reduced to subcritical and then zero; slow cake removal was observed as distinct cylindrical rolling floc (about 50 μm diameter) and aggregates. The extent of cake removal varied from about 90 to <5% depending on the ionic environment with low ions resulting in better removal. Cake formed over longer periods (upto 60 min) showed negligible removal. The critical fluxes of SW8 measured with DOTM increased with crossflow, but exhibited higher values than expected from their primary particle size.

Introduction

Microfiltration (MF) is a pressure driven membrane process in which suspended colloids and particles in the approximate size range 0.1–20 μm are retained by microporous membranes. Over the past decade, MF has grown to be a mature process for industrial applications including bacterial, yeast, and mammalian cell harvesting; clarification and/or sterilisation of various fluids such as wine, vinegar, and apple juice in the food industry; clarification of medical solutions; recovery and purification of biological and pharmaceutical products and treatment of water and wastewater.

In the application of MF to feeds containing bacteria, the convection due to the filtration process causes formation of a cake layer and eventually formation of a multi-layer cake, which presents a hydraulic resistance to permeate flow. The dynamic nature of the bacteria with their capacity to physiologically modify their surface in response to environmental conditions is one of the complicating features of MF of bacteria. This complication makes it difficult to simply apply the models developed from non-biological colloidal suspensions. In addition, both surface charge and hydrophobicity, commonly used to describe surface properties of colloidal suspensions, are bulk surface properties that do not account for the molecular complexity of the bacterial surface. The molecular architecture of the bacterial surface and the nature of the bacterial extracellular polymeric substances (EPS), which may be critical to the fouling process, are physiologically variable. The extracellular matrices extend beyond the outer cell wall and are comprised of complex polymers such as acidic polysaccharides and proteins [1]. Understanding the mechanisms which mediate bacterial fouling is critical for the successful application of MF in biotechnological and wastewater separations. However, there are relatively few studies of membrane fouling behaviour by bacterial material in the literature.

A study of bacterial MF by Hodgson et al. [2] suggested that the bacterial surface must be considered when describing cake resistance and solution rejection phenomena. This study demonstrated the role of the EPS matrix in filtration resistance by changes in flux affected through matrix modification with a proteolytic enzyme and a chelating agent. Based on the evidence of their research, they proposed that the polymer matrix associated with each cell was able to effectively enmesh and confer the bulk of the resistance. The nature of this enmeshment was a function of pressure, ionic strength and microbial surface treatment.

In a separate study, Leslie et al. [3] investigated the fouling of microfiltration membranes by a protocol involving bacterial cake deposition and re-dispersion under defined hydrodynamic conditions. The degree of fouling was quantified by measuring the amount of cake re-dispersed and the recovery of the membrane original water flux. Their study showed that the re-dispersive behaviour of model bacteria (pseudomonas, SW8) grown in carbon-limited conditions was influenced by changes in bacterial physiology. The re-dispersion of SW8 decreased from approximately 70% at low ionic strength to 20% at high ionic strength. They explained the results in terms of the change of the diffused electrical double layer surrounding bacteria. At low ionic strength, the electrical double layer surrounding each bacterium is extended and molecular interactions, which stabilise a cake, are prevented by electrostatic repulsion between bacteria. At high ionic strength, the electrical double layer is compressed, and the cake is stabilised by the interaction of molecules on the bacterial surface.

In recent years, our research group has developed a continuous, in situ direct observation through the membrane (DOTM) technique [4], [5] and applied it to observe the fundamental phenomena of particle deposition, adhesion and possible re-dispersion on microfiltration membranes as a function of flux, tangential shear and, particle size of super-micron particles (3–12 μm). Using the DOTM technique, we have identified the “critical flux” at which deposition commences for a given feed and hydrodynamic condition for super-micron particles and compared those critical fluxes with predictions from various models such as the inertial lift model, shear-induced diffusivity and shear-induced hydrodynamic models [5]. Our study found that a modified shear-induced diffusivity model, which uses the shear-induced diffusivity for dilute suspension, predicts critical fluxes similar to our experimental results for 6 and 12 μm latex and yeast. However the model provided a poor prediction of the critical fluxes of 3 μm latex particles and 4 μm algae.

The work reported here extends the DOTM technique to directly observe the deposition and removal behaviour during filtration of submicron bacteria at various conditions using model bacteria SW8 (mean size 0.8 μm) with known physical and physiological properties. In order to observe the sub-micron bacteria SW8, we applied fluorescent microscope components so that fluorescent particles could be observed. The characteristic cake formation and removal patterns under various operational conditions, as well as ionic conditions will be presented in this paper. We have also estimated the critical flux of the bacteria which are compared with our previously obtained critical fluxes of biomaterials.

The purpose of this study is to observe in situ the characteristics of cake formation on the membrane surface at and above the critical flux as well as to observe the removal of the cake in response to reduced flux to below the critical flux and at zero flux with recirculation. Due to the nature of the in situ observation, estimation of cake build up and the proportion of cake removal are not expected to be precise. In addition, the stained bacteria cells lose their brightness with time under strong fluorescent light which makes it more difficult for estimation of bacteria deposition on the membrane. In the case of the filtration tests in salt solution, the refractive index of the permeate at high salt concentration varies significantly from that of water, making it difficult for the microscope objective lens to focus on the membrane surface. However the observations presented in this paper do provide hitherto unrecognised insights into the nature of fouling in membrane filtration of bacterial materials. The paper also demonstrates that special consideration should be given when applying general membrane filtration models to biomaterials.