Operation of membrane bioreactor with powdered activated carbon addition

The effect of powdered activated carbon (PAC) addition to the activated sludge (AS) in a membrane bioreactor (MBR) has been investigated. The long term nature of the tests allowed the PAC to gradually incorporate into the biofloc forming biologically activated carbon (BAC). One series of tests involved 4 bench scale (2 L) MBRs operated at sludge retention times (SRTs) of 30 days with PAC inventories of 0, 1, 3 and 5 g/L and steady state biomass concentrations of 12.0 +/- 1.0 g/L. The characteristics of the mixed liquors (MLSS) from the 4 reactors were compared. Short term filtration tests, including measurement of specific cake resistance (SCR), flux decline profile, and irreversible fouling resistance in an unstirred cell and "sustainable flux (by monitoring transmembrane pressure (TMP) rise) in a crossflow cell all showed better filtration performance for the MLSS with BAC compared with the AS alone. In terms of SCR and flux decline profile the 1 g/L PAC addition performed best, but in terms of minimizing irreversible membrane fouling and maximizing "sustainable flux the 5 g/L PAC was best. All 4 systems showed lower total organic carbon (TOC) in the permeate compared to the bioreactors, but the lowest permeate TOC (and the best removal) was for the highest PAC loading. The benefit of PAC addition was confirmed in a second series of tests with two 20 L MBRs with submerged hollow fibers, one operated without PAC, the MBR(AS), and the other with 5 g/L PAC, the MBR(BAC). For an SRT of 30 days (which involved 3.3% sludge wastage per day and 3.3% new PAC addition per day) and a fixed flux of 21 L/m(2) hr the MBR(AS) showed a TMP rise of about 2.4 kPa/day whereas the MBR(BAC) showed a rise of only 0.8 kPa/day. However when the MBRs were operated without wastage the performance of the MBR(BAC) was worse than the MBR(AS). Thus the improved performance of the MBR(BAC) requires regular replenishment of aged BAC with fresh PAC.