WATER RESEARCH, cilt.36, sa.5, ss.1155-1166, 2002 (SCI-Expanded)
Drinking water denitrification performance of a bench scale membrane bio-reactor (MBR) was investigated as function of hydraulic and biological parameters. The reactor was a stirred tank and operated both in batch and continuous mode. The mixed denitrifying culture used in the batch mode tests was derived from the mixed liquor of a wastewater treatment plant in Erzincan province in Turkey. But the culture used in the continuous mode tests was that obtained from the batch mode tests at the end of the denitrification process. The nitrate contaminated water treated was separated from the mixed liquor suspended solids (MLSS) containing active mixed denitrifying culture and other organic substances by a membrane of 0.2 mum average pore diameter. The results indicated that the use of a membrane module eliminated the need for additional post treatment processes for the removal of MLSS from the product water. Concentration of nitrite and that of MLSS in the membrane effluent was below the detectable limits. Optimum carbon to nitrogen (C/N) ratio was found to be 2.2 in batch mode tests. Depending on the process conditions, it was possible to obtain denitrification capacities based on the reactor effluent and membrane effluent up to 0.18 kg m(-3) day(-1) and 2.44 kg m(-2) day(-1) NO3--N, respectively. The variation of the removal capacity with reactor dilution rate and membrane permeate flux was the same for two different degrees of [MLSS](0)/[NO3--N](0) (mass) ratios of 25.15 and 49.33. The present M BR was able to produce a drinking water with NO3--N concentration of less than 4 ppm from a water with NO3--N contamination level of 367 ppm equivalent to a NO3--N load of 0.310 kg m(-3) day(-1). The results showed that MBR system used was able to offer NO3--N removals of up to 98.5%. It was found that the membrane limiting permeate flux increased with increasing MLSS concentration. (C) 2002 Elsevier Science Ltd. All rights reserved.