The critical flux theory indicates large particles hardly deposit on membrane surface due to high back transport velocity away from the membrane surface. Although it is not clear what are the large enough particles that hardly deposit in a typical MBR with immersed membranes, it might be the particles sized around a micron or bigger. Since the typical particle size analyzers based on laser light scattering principle have a lower detection limit of around 0.1 micron, the vast majority of particles detected by the analyzer are not the likely culprit of membrane fouling. Moreover, such particle size analyzers have a rather high inaccuracy in submicron range especially when they cover a wide range of particle sizes. In addition, macromolecules such as soluble microbial products (SMP) and extra-cellular polymeric substances (EPS), which are macromolecules mainly smaller than 0.1 micron, appear the major membrane foulant in most cases as discussed here. Then, a question arises – why do we measure particle size distribution to estimate membrane fouling potential?
Although over-micron particles are not the major membrane foulant, high occurrences of near-micron particles appears indicate high occurrences of sub-micron particles. As shown in Fig. 1, average particle sizes were monitored over time while flux of tubular membrane was measured. Depending on the type of pump, clearly different mean particle sizes were observed. The system with rotary vein pump had smaller mean floc sizes and the flux decline was more severe.
In MBR, small particle sizes somehow tend to indicate unfavorable biological conditions for membrane filtration. The mean and/or average particle sizes tend to decrease when one or more of the following conditions are met.
- Dissolved oxygen (DO) level decreases (Kang, 2003)
- F/M ratio increases (Syed, 2009)
- SRT decreases (Ahmed, 2007)
In some cases, opposite SRT effect on particle size can be observed depending on experimental conditions (Cao, 2008), but the exact causes are not known. Perhaps the differences in the experimental conditions not described in literature.
a) Mean particle size b) Flux
Fig. 1. Trend of particle size changes and flux decline over time with two different circulation pumps in MBR (Kim, 2001)
a) Effect of SRT on particle size distribution b) Effect of SRT on membrane fouling
Fig. 2. Relation among SRT, particle size distribution, and membrane fouling rate. Flat sheet Kubota membrane was used in lab-scale systems fed with synthetic feed (Ahmed, 2007)
© Seong Hoon Yoon