Evolution of fouling deposition and removal on hollow fibre membrane during filtration with periodical backwash Yun Ye, Vicki Chen ⁎, Pierre Le-Clech UNESCO Centre for Membrane Science and Technology, University of New South Wales, 2052 Sydney, NSW, Australia abstract article info Article history: Received 13 December 2010 Received in revised form 29 March 2011 Accepted 31 March 2011 Available online 27 April 2011 Keywords: Direct observation Hollow fibre membrane Backwash Air scouring Fouling deposition and removal during filtration with short periodical backwash was investigated by direct observation coupled with hydraulic resistance measurement. Using a model mixture of bentonite and alginate, it was found that short periodical backwash only (i.e. without air scouring) expanded the foulant cake, which is subsequently recompressed back to membrane surface when the filtration resumes. With air- scouring aided backwash, the fouling material is convected away from the effective filtration zone, thus limiting re-deposition. However, due to the potential fractionation of foulant species at high air scouring rate, the deposition of more highly resistant components was observed over multiple filtration and backwash cycles. The results indicated that the changes in cake height, composition and structure were affected not only by the initial composition of the feed, but also by the hydrodynamics of the backwash and air scouring present. © 2011 Elsevier B.V. All rights reserved. 1. Introduction Hollow fibre membrane systems are widely used in water and wastewater treatment. However, this process is still limited by membrane fouling, generally due to the cake formation or pore blocking by particulate and macromolecular substances in the feed water. Membrane fouling is commonly monitored by simple flux- transmembrane pressure (TMP) measurements, and the fouling resistance is calculated by using Darcy's law. Fouling mechanisms such as cake filtration, pore blocking are inferred by applying different filtration laws, but these laws are usually based on broad assumptions regarding the uniformity of the particle/pore size and cake struc- tures [1]. In reality, the heterogeneity of the foulant layers has been shown in a number of previous studies [2,3]. Application of high and/or turbulent shear via crossflow, gas sparging, and backwash (reversal of permeate flow through the pores) are common physical approaches to remove particulate and macromolecular foulants. However, there is little detailed composition and structure information about foulant nature and cake characteristics and how they evolve during cyclical filtration and physical cleaning. These aspects are crucial to provide a better understanding of the membrane fouling and mitigate the effects of fouling to a sustainable level. In the past decades, a number of non-invasive techniques have been developed and applied to in-situ real-time monitoring of membrane processes. These techniques include: direct observation through the membrane (DOTM) [4–6], direct visual observation (DVO) [7,8], nuclear magnetic resonance (NMR) imaging [9], ultrasonic time domain reflectometry (UTDR) [10] and direct observation using laser sheet [3,11]. Most of the methods have focused on flat sheet membranes, and only a limited number of studies have been carried out with hollow fibre membranes [3,11–13]. Airey et al. [9] applied the NMR imaging technique to observe the cake formation and dissipation of particle concentration polarization layers on inside/out tubular membranes during the filtration of colloidal silica suspensions at constant pressure. Their study provided the first direct experimental evidence for a flowing concentration polarization particle layer, obtained by means of a non-invasive technique. Recently, Marselina et al. [13] adopted the DOTM method, originally developed by Fane et al. [4], to visualize cake formation on the external surface of hollow fibre membranes. Via direct observation (DO), the in-situ, real time visualization and quantification of the cake height on the membrane surface was measured, reducing some assumptions required to estimate the specific cake resistance. The velocity profiles of particles near the membrane surface during filtration could also be estimated during the formation of stagnant and fluidised fouling layers [13]. Mendret et al. [3] also investigated the time variations of the thickness and porosity of deposits on hollow fibres using an optical method with laser sheet. For constant pressure operation, two different stages were observed during the deposit thickness growth: first, a thin and dense layer developed, and then a thicker and more permeable layer was generally formed. Their study demonstrated the link between deposit structure and process performance [3]. In addition to estimating cake height, the DO technique also provides useful information on the membrane cleaning and foulant removal. Mores and Davis [8] observed the fouling and cleaning of Desalination 283 (2011) 198–205 ⁎ Corresponding author. Tel.: + 61 2 9385 4813; fax: + 61 2 9385 5966. E-mail address: v.chen@unsw.edu.au (V. Chen). 0011-9164/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.desal.2011.03.087 Contents lists available at ScienceDirect Desalination journal homepage: www.elsevier.com/locate/desal