As water temperature decreases, the membrane performance measured by permeability naturally decreases because water viscosity increases (see the equation here). The cyclical behavior of permeability well coincides with water temperature as shown in Fig. 1, where long-term behavior of permeability is plotted against water temperature.
One thing remarkable is that the amplitude of permeability variation is much larger than the viscosity variation caused by the temperature swing. For instance, viscosity increases from 0.89 cP to 1.31 cP when temperature drops from 25oC to 10oC, which is a 47% increase of viscosity that can reduce permeability by 32% according to the resistance in series model here. Meanwhile, in the actual field data (Fig. 1), permeability decreases 40-60% from 250 LMH/bar to 100-150 LMH/bar depending on where we look at.
The more severe membrane performance drop than predicted based on water viscosity is well presented in Fig. 3, where a pilot MBR with 85 L reactor volume treating municipal wastewater using flat sheet immersed membranes (Kubota, Japan) were used to compare membrane fouling propensities at different water temperatures (van den Brink, 2011). It is very apparent that the membrane fouling represented by the Rm+Rf increasing rate (see resistance in series model here) is disproportionally high at low temperature relative to water viscosity increase. It is also noticeable that the temperature effect on membrane fouling is more apparent at high flux conditions.
The discrepancy might be able to be explained by the partial defloccuation of microbial floc at low temperature. As shown in Fig. 3, supernatant total organic carbon (TOC) contents, which represent free fine particles that are not flocculated, increases during the winter to spring time (Wu, 2011). In other study shown in Fig. 4 (van den Brink, 2011), it was found that the TOC increase in supernatant was mainly contributed by polysaccharides rather than proteins. Since such fine particles / macromolecules have higher tendencies to deposit on membrane surface than larger particles, membrane fouling can be more significant in cold water than in warmer water.
Other plausible cause of permeability is the lowered shear rate on membrane surface due to the slower upflow velocity in the viscous medium at low temperatures, but experimental evidences are yet to be published. Low particle diffusivity at low temperature also can reduce particle back-diffusion from the membrane surface, which increases the thickness of concentration polarization layer and in turn the filtration resistance.
Fig. 1. Weekly average permeability and temperature in the MBR plant of Nordkanal, Germany, in 2004-2009. Permeabilities are not normalized against temperature (Brepols, 2011).
Fig. 3. Average fouling rates for different flux steps at three different temperatures (van den Brink, 2011).
Fig. 3. Seasonal variation of supernatant organic concentration in the MBR and CAS systems (WU, 2011)
Fig. 4. Ratio between organic carbon concentrations in the soluble fraction of the experimental system and MBR pilot (van den Brink, 2011).
© Seong Hoon Yoon