The performance of different two-stage systems was compared for the treatme
nt of synthetic wastewater. The first stage was a completely mixed reactor
without sludge retention for the stimulation of dispersed bacterial growth.
The second stage was an activated sludge system in which growth of protozo
a and metazoa was stimulated. Solid-liquid separation was achieved either b
y sedimentation (conventional activated sludge (CAS) system) or submerged m
embrane filtration (membrane-assisted bioreactor (MBR) system). Some 80% of
the chemical oxygen demand (COD) was removed in the first stage at volumet
ric loading rates up to 1.6 g COD l(-1) d(-1). In case the subsequent stage
was conventional activated sludge, a further COD decrease was noticed. In
case of a subsequent MBR system, however, the dissolved COD concentration t
ended to increase due to the retention of organics by the membrane. In the
first stage of both systems, most of the nitrogen and phosphate was used fo
r biomass incorporation. Increased mineralization by protozoa and metazoa i
n the second stage resulted in a partial release of nitrogen and phosphate
to the effluent. The MBR system yielded a 20-30% lower sludge production th
an the CAS system under similar conditions of solids retention time and org
anic loading rate. This was attributed to the higher amount of predators in
the second stage of the MBR configuration. However, the increased grazing
of predators on nitrifying bacteria can result in overgrazing of the latter
population. Overall, the two-phase system based on conventional activated
sludge had as major point of weakness the wash-out of suspended solids, whi
le the one based on the MBR was hampered by too intensive grazing on the ni
trifiers, increased N and P concentrations and wash-out of soluble but humi
fied COD. (C) 1999 Elsevier Science Ltd. All rights reserved.