Impact of different inoculum sources on performance of MBRs for municipal wastewater treatment: Dynamic membrane versus ultrafiltration membrane

Abstract

Water scarcity is one of the major challenges of today. Treatment technologies that are used for wastewater reuse have become more attractive in 21th century while dealing with the water scarcity. Membrane bioreactors (MBRs) provide high quality effluent which can be re-used for the beneficial purposes such as irrigation. In the MBR technology, activated sludge and membrane separation are integrated. Generally, microfiltration (MF) and ultrafiltration (UF) membranes are used in conventional MBRs to remove solids from wastewater. Since solids retention time (SRT) and hydraulic retention time (HRT) are decoupled in MBR systems, solids can be retained within the bioreactor. Moreover, compared to conventional activated sludge (CAS) process, MBRs require less area, and thus have smaller environmental footprint. Although MBRs have several advantages, fouling problem and high capital cost of the membrane appear as prominent obstacles that prevent widespread applications of this technology. Dynamic membrane (DM) technology can be an alternative option to solve limitations faced in conventional MBRs. DM is a separation layer formed on a low-cost support material by the deposition of solid particles. DM can be either pre-coated or self-forming. Pre-coated DMs are formed by passing of a pre-coating reagent such as powdered activated carbon through support material. On the other hand, self-forming DM (SFDM) is generated as a biological separation layer by colloidal particles and organics which present in filtered solution such as wastewater. This biological separation layer is also called as cake layer or secondary membrane. Support material used in DM is cheaper and more durable compared to polymeric material of conventional UF membrane. Different materials such as polyester fabric, woven and non-woven fabrics, filter cloths and ceramic materials can be used as support material for DM formation. Pore size of support material can change in a wide range of 10-200 μm. One of the most significant advantages of DM technology over conventional MBR is once DM formed, support material itself is no longer important. Besides, simple physical cleaning methods are applied to recover permeability in DM technology. Biological applications of DM can be either aerobic MBR (DMBR) or anaerobic MBR (AnDMBR). DM formation process consists of four stages: substrate layer formation, separation layer formation, fouling layer formation and filtration cake layer formation. One of the major drawbacks of DM technology is transition period between the cake layer formation and removal of cake layer due to the physical cleaning applied in case of permeability reduction. During this transition period because of the large pore size of the support material, quality of permeate is decreased and biomass is lost. Time required to reach stable permeate turbidity and almost zero suspended solids concentration in the permeate is defined as DM formation time. Therefore, shorter DM formation time is generally required to prevent low quality permeate and biomass loss during the aforementioned transition period. DMBR performance is determined by several factors such as support material characteristics, sludge characteristics and operation conditions. The aim of this study was to evaluate the impact of different inoculum sources on the treatment and filtration performance of MBRs which include hollow-fiber conventional UF membrane and DM for municipal wastewater treatment. Experimental study was carried out in two stages. Excess sludge obtained from a CAS process and high rate activated sludge (HRAS) process were utilized as inoculum in Stage-1 and Stage-2, respectively. DM and UF modules were located into the same bioreactor which had a capacity of 5.2 L. Filtration and treatment performances of DM and UF for municipal wastewater treatment were compared under same operational conditions in each stage. DM and UF membranes were operated for 67 days at an operational flux of 8 LMH (L/m2.h). There was no sludge wasted apart from sampling during the operation. System was operated in cycles of 190 seconds filtration and 35 seconds backwash. Municipal wastewater taken from the outlet of an aerated grit chamber of a full-scale preliminary wastewater treatment plant (WWTP) was used as substrate. Both inoculums were characterized to observe the changes during the operational period. Moreover, membranes were characterized before the operation and ensured that similar membrane properties were provided for each stage. Polyvinylidene fluoride (PVDF) and multi-multifilament polyester filter were used for UF and DM, respectively. Different experimental analyses were conducted to observe treatment and filtration performances of UF and DM. Environmental scanning electron microscopy (ESEM) was used for the observation of cake layer structure that was accumulated on the surface of the UF and DM. Total solids (TS), volatile solids (VS), total suspended solids (TSS), volatile suspended solids (VSS), chemical oxygen demand (COD), soluble chemical oxygen demand (sCOD) ammonium‑nitrogen (NH4+-N), total nitrogen (TN), total phosphorus (TP) and turbidity were measured to observe treatment performance in each stage. Dewaterability of sludge was indicated by capillary suction time (CST). Transmembrane pressure (TMP) was recorded during the study to analyze the filtration performance. Specific resistance to filtration (SRF) of sludge samples were measured to evaluate filterability. Extracellular polymeric substances (EPS) and soluble microbial products (SMP) were measured to understand membrane fouling behavior in each stage. High COD removal efficiencies (>86%), low turbidity values in the permeate (around 1 NTU), and high TSS removal efficiencies (>99%) were obtained during the operation in each stage. On the other hand, since there was no anoxic and anaerobic zone in the bioreactor, TN and TP removal efficiencies were low in each stage. NH4+-N removal efficiency was >99% in each stage, which showed NH4+-N was converted to the nitrate (NO3−) via nitrification process. Higher NO3− concentration in the bulk sludge compared to inoculum sludge confirmed mentioned inference. In Stage-1, turbidity in the permeate of DM decreased faster compared to Stage-2. This result indicated that cake layer formed quicker in Stage-1 compared to Stage-2, which meant that DM formation time was shorter in Stage-1 compared to Stage-2. After stable condition was obtained, average TMP was 588±33 mbar in Stage-1, and 422±3 mbar in Stage-2. SRF of sludge used in Stage-1 was higher than that in Stage-2. As a consequence, DM was operated with higher TMP in Stage-1. SMP concentration and bound EPS content of sludge were higher in Stage-1 than those in Stage-2. Higher SRF, zeta potential, and sludge volume index (SVI) values for sludge obtained in Stage-1 compared to Stage-2 might be the reason of higher TMP values determined for DM in Stage-1. However, TMP of UF membrane was not affected by the inoculum type in this study. Considering ESEM images, compact and dense DM layer formed in Stage-1, while porous and fragmented cake layer was observed in Stage-2. ESEM images showed no difference between the cake layer formed on the surface of UF membrane in Stage-1 and Stage-2. Stage-2 showed more stable operation compared to Stage-1 in terms of TMP profiles of DM. Since treatment performances of each membrane were similar, it could be concluded that DMBRs can be a good alternative to conventional MBRs. More stable operation with lower TMP obtained for DMBR when HRAS inoculum used. Thus, it could be indicated that inoculating DMBRs with excess sludge from HRAS systems can be an effective solution to stabilize TMP at a reasonable level. Large scale applications should be performed in order to examine the applicability and uncertainties of DM technology on municipal wastewater treatment. Moreover, monitoring of EPS/SMP is important subject to prevent fouling and permeability loss in DMBRs.