Characterizing bacteria involved in fouling of spiral wound membranes

Large amounts of water are purified by high pressure driven membrane filtration processes, i.e. nanofiltration (NF) and reverse osmosis (RO). In Chapter 1, we introduce the topics discussed in this Thesis. A common but major challenge in the operation of high pressure membrane filtration systems is the formation of deposits on the membrane surface. This is defined as fouling when it leads to a certain membrane performance decrease. The deposition of abiotic particles on the membrane can often be adequately prevented via pre-treatment of the feed water steps, or reduced by cleaning the membrane. Microbial presence on the membrane surface cannot be prevented, and the consequential growth and formation of recalcitrant biofilms on the membrane surface is difficult to control via membrane cleaning.

Much scientific effort has been invested to understand which bacteria are involved in membrane fouling. This has shown that bacteria belonging to the Sphingomonadaceae family are detected frequently within membrane fouling layers. How bacteria function and interact on the membrane surface, and respond to cleaning agents is not well studied. This information might prove valuable to design alternative cleaning strategies. In this Thesis, we have physiologically characterized members of the Sphingomonadaceae family that were isolated from fouled membranes, and investigated the ability to remove membrane fouling layers using an alternative cleaning strategy.

In Chapter 2, we provide on overview of the studies that have investigated the microorganisms involved in fouling of high-pressure membrane systems. Most of these studies focus on the bacterial community composition. The diversity of eukaryotic and archaeal species within membrane fouling layers is rarely studied. Comparison between the microbial diversity of the feed water and pre-treatment steps in relation to that of the membrane surface reveals large differences within the bacterial and archaeal diversity, but not within the fungal. Membrane filtration thus provides a selective environment for bacteria and archaea, but not for fungi. The close resemblance between the fungal diversity before and after membrane cleaning confirms the tolerance of particular fungal species to bactericidal agents (e.g. chlorine). Studies that have investigated the role fungi in membrane biofouling collectively show that their role has been underestimated.

In Chapter 3, we investigate the physiological traits of 21 Sphingomonadaceae (Sph) strains that were isolated from membranes used in practise and in fouling simulation experiments. As a reference, we used 21 Sphingomonadaceae type strains, to mark the close physiological between the 21 Sph membrane isolates. The growth requirements (temperature, pH, salt concentration) of the 21 Sph membrane isolates are irreconcilable with members of the Sphingomonadaceae family that were not isolated from fouled membranes. Bacterial behaviour on the membrane surface will thus be incorrectly predicted when assuming that membrane isolates behave similar to bacteria that are phylogenetically closely related.

In Chapter 4, we show that simulation of membrane fouling can be studied under axenic yet representative and conditions. Physiological differences between Sph membrane isolates substantially affects membrane performance. Examining the activity of three bacterial strains under static and representative membrane fouling conditions, reveals that surface adhesion poorly predicts the ability of a bacterium to cause membrane fouling. Due to the stochastic nature of biofilm formation, considerable differences in membrane performance emerge, and membrane biofouling experiments should be repeated to recognize experimental outliers. In this manner, true effective biofilm control strategies can be identified.

In Chapter 5, we investigate nutrient limitation as sustainable manner to control unwanted biofilm growth on synthetic membrane filters. The final stage of biofilm formation is the active and regulated release of cells to the planktonic lifestyle. Pure culture studies have identified the factors (e.g. nutrient limitation) that stimulate biofilm dispersion of individual bacterial species, but the response of multispecies biofilms to these factors is not well studied. We show that nutrient limitation leads to multispecies biofilm dispersal from membranes. However, of flushing the (membrane) surface leads to biomass release is due to a biological response, as well as due to the mechanical force of the water flow. The cycle of biofilm growth, nutrient limitation and flushing the membrane was repeated three times to identify that the bacterial community composition stabilized after the second cycle. Because this coincided with a substantial decrease in biomass release during the third cycle, we conclude that repetitive cleaning leads to a stable, resilient bacterial community with increased mechanical tolerance.

In Chapter 6, we discuss our main findings in respect to existing literature. Standardized methods, with experimental conditions representative for membrane used in practise, are required to increase scientific relevancy and improve interlaboratory comparison. Often, literature reports on novel fouling control strategies that under laboratory conditions are successful, but which often are not implemented in practise. A standardized membrane fouling method could remove large evidence gaps, and as such be a bridge between successful laboratory control strategies and their implementation in practise. Hitherto, fouling control strategies are designed to be effective in their own regard. We propose that control strategies can be combined to increase the overall effectivity. The combination of nutrient limitation and a modified feed-spacer provides a blend that might successfully control membrane biofouling.