Phosphorus removal in aerobic granular sludge process: Effect of biomass management on efficiency and bacterial diversity

Aerobic Granular Sludge (AGS) represents an innovative strategy to treat wastewater. The granules are formed by microbial consortia enclosed in a self-produced exopolymeric matrix. The advantages of this process over activated sludge are excellent settling abilities and high biomass retention. The biological phosphorus removal in AGS reactors is performed by bacteria known as Polyphosphate Accumulating Organisms (PAOs). A promising alternative to enrich AGS reactors with PAOs consists into selectively purging unfit subgroups of granules. In column-shaped reactors (H/R ratio > 10), phosphorus removal can be efficiently improved by removing selectively granules from the upper part of the sludge bed. However, this strategy is not applicable in reactors with a low H/R ratio (< 2) because the granules are more evenly distributed in the sludge bed. In this research, a new methodology of PAO enrichment was explored. First, subgroups of PAO-rich granules were identified by performing isopycnic centrifugation on granular sludge (density-based separation) combined with quantitative PCR and 16S amplicon sequencing. Second, the abundance of these granules was increased through non-segregating microbial management. The bacterial dynamics were followed throughout the whole experiment by PCR-DGGE. The lab-scale reactor was operated under feast and famine regime. Phosphate, ammonium, nitrite, nitrate, chemical oxygen demand (COD) and suspended solids (total and volatile) were measured on the influent and filtered effluent. The SBR experiment was split into three periods: granulation, maturation and biomass management. The first granules were observed after 45 days and their diameter ranged from 100 to 400 µm. The formation of granules coincided with a sharp decrease of the SVI5. Phosphorus removal improved up to 95%. During the maturation period, the average diameter of the granules increased and reached 1 to 2 mm. An important variability in phosphorus removal was observed and frequent peaks with P concentration higher than 25 mg/L were recorded in the effluent. At the end of this period, a sludge sample was collected to perform a bacterial diversity analysis on granules with different densities. The high-throughput sequencing approach revealed that the 16S rRNA gene frequency of Ca. Accumulibacter, a major PAO in AGS and activated sludge processes, was dominant in the middle fraction (50-60) where it reached 45.1%. The gene frequency was lower in both the lighter and the heavier fractions. A qPCR analysis was performed on the same samples and revealed that fractions 50-60 and 70-80 contained the highest quantity of PAO-specific 16S rRNA gene copies with 6.8 log copies per ng of DNA. On the opposite, the lighter fraction contained only 4.9 log copies per ng of DNA. Two different strategies of biomass management were applied. The first strategy involved a high selective pressure (short settling time) to select fast settling granules. It resulted in a decrease of the average phosphorus removal to ca. 18%. This approach induced the development of granules with high densities, while granular seeds were washed-out. The P content in the sludge was high (102 mg P/g dry suspended solids) which suggested that the biomass was probably close to saturation in polyphosphate. The second strategy was designed to reduce the selective pressure and let the granules with a slower settling velocity reaching the sludge bed, before performing a homogeneous purge of the solids. Hence, this procedure removed not only very slow settling (wash-out with the effluent) but also fast settling granules (removed from the settling bed). The subgroup of granules with the most abundant PAO population (intermediate densities) was consequently kept in the reactor. Quickly after the start of this procedure, P removal improved and stabilized over 90% until the end of the experiment. The average removal for total soluble nitrogen was 76% and COD removal remained stable at 96%. The P content of the biomass was 84 ± 5 mg P/g dry suspended solids, which suggested a better turnover of PAO biomass. This is likely to be due to the hold of granule seeds that were not yet fully saturated in polyphosphate. Overall, this study offers a new approach of biomass management in AGS reactor with a low H/D ratio.