Comparison of dead-end and continuous filtration conditions in a denitrification membrane bioreactor

The impact of using dead-end filtration conditions to sustain permeability in an immersed denitrification membrane bioreactor (MBR) has been studied, and the performance compared with a more conventional MBR hydrodynamic regime using constant gas sparging. Shear imparted during constant gas sparging was sufficient to reduce median floc size (d50) from 182 μm, observed during dead-end filtration, to 55 μm. Although this reduction in floc size demonstrated erosion, a concomitant generation of soluble extracellular polymeric substance (EPS) was not observed. Whilst unexpected, this finding accords with several recent investigations studying the long-term effect of shear. However, a significant concentration of soluble EPS was recorded in the biofilm that developed when using constant gas sparging (44.0% cf. 18.8% for dead-end) and was explained by the preferential transport of soluble compounds to the membrane wall during constant shear conditions. For an imposed flux (J20 °C) of 22 L m−2 h−1, a fouling rate (dP/dt) of 0.39 mbar h−1 was recorded during constant gas sparging compared to 0.11 μbar h−1 during dead-end filtration. The lower fouling rate recorded during dead-end conditions was sustained using only a 30 s gas sparge at the end of each dead-end filtration cycle (ca. 10–30 min) which appeared sufficient to almost completely remove the cake layer formed. The reversibility of the cake was attributed to the simultaneous deposition of soluble, colloidal and flocculant materials which created a more heterogeneous, porous and subsequently less tenacious deposit. Cycle length, rather than gas flow rate, was identified as the critical parameter for optimum dead-end filtration; provided filtered volume was sufficiently short to limit mass deposition, low gas flow rates were sufficient to reverse deposition. This significantly impacts on energy demand; dead-end filtration provides a net energy reduction for gas sparging from 0.19 kWh m−3 for constant sparging to 0.007 kWh m−3.

Research highlights▶ Permeability was sustained using dead-end conditions which at first glance is counter-intuitive. ▶ Shear imparted during constant gas sparging was sufficient to reduce median floc size. ▶ Floc size reduction resulted at high shear, however, soluble EPS generation was not observed. ▶ An-order of magnitude decrease in specific gas demand cf. continuous gas sparging is possible.