Understanding and tailoring aerobic granular sludge wastewater treatment systems


Prof. Christof Holliger, Dr. Graciela Gonzalez-Gil, David Weissbrodt, Samuel Lochmatter

Funding agency

Swiss National Science Foundation

Project period

July 2008 – December 2012


Dr. N. Garin, Bio-imaging and Optics Center, EPFL, Switzerland; Prof. Eberhard Morgenroth, ETHZ, Switzerland; Dr. Thomas Neu, UFZ Magdeburg, Germany; Prof. Mark van Loosdrecht, TU Delft, The Netherlands; Mr. A. Baumann, Ara Thunersee, Uetendorf, Switzerland.


Avoiding contamination of surface waters has been the driving force to implement wastewater treatment systems and the established effluent discharge standards aim to minimize impacts to receiving waters. These standards will become stricter in the near future. The current increased needs to protect water bodies together with global energy constrains exert pressure for developing more energy-efficient and reliable treatment systems. At present, wastewaters are mainly treated using activated sludge systems. Although these systems can have acceptable performance in terms of carbon, nitrogen and phosphorus removal, the energy and land space required is considerable. To alleviate this drawback, the use of aerobic granular sludge-based systems has been recently proposed as a promising innovative alternative.
For a successful operation of this promising treatment system, formation of physically and metabolically stable granular sludge is a prerequisite. A detailed understanding of the granule formation, the bacterial populations involved, the physical structure, and the importance of PAO and GAO is still missing and therefore we propose to investigate three main objectives: (i) the competition and relative importance of PAO and GAO in granular sludge structure, (ii) granule formation and stability for optimized nitrogen and phosphorus removal, and (iii) the microbial assembly and community in relation to granular structure. A combination of process engineering approaches with the molecular characterization of the microbial communities of the granules will be applied.


PAO-GAO: their competition and relative importance in granular sludge structure
The mechanism of bacterial selection was investigated in sequencing batch reactors (SBR) using mature aerobic granular sludge (AGS) to identify the operation factors governing bacterial community dynamics, competition of polyphosphate- (PAO) and glycogen-accumulating organisms (GAO), and biological nutrient removal (BNR) under both dynamic and steady-state conditions. Whereas AGS biomass growth was limited with only acetate, the presence of propionate in the influent wastewater either as single volatile fatty acid or in mixtures with acetate enabled stable AGS bed growth and selected for the PAO Accumulibacter (38%).
A multifactorial experimental design was used to screen for main effects on PAO/GAO competition within a set of six parameters, i.e. COD concentration, Ac/Pr and COD/P ratios, pH, temperature, and the redox conditions during starvation. This multifactorial investigation revealed that Accumulibacter (47%) and dephosphatation (80-100%) were favored with alkaline pH (7.5), temperature below 20°C, and Ac/Pr of 50:50%. Competibacter (35%) were preferentially selected at 28°C, pH 6.5, with acetate, and with the lower COD concentration used. Conditions selecting for GAO were also favorable for Tetrasphaera relatives (16%) belonging to Actinobacteria and possibly important PAO of full-scale BNR systems. A mathematical model targeting the PAO/GAO competition for the carbon source under anaerobic plug-flow feeding conditions across the AGS bed showed that an alkaline pH (7.5-8.0) was preferentially selecting for anaerobic acetate uptake by PAO over GAO independently from the process temperature. In conclusion, pH was identified as the most dominant factor for bacterial selection in AGS systems. Furthermore, control of duration of anaerobic and starvation phases was required for optimal performance. Dosage of a 3-carbon substrate such as propionate in the influent can also help to enable proper AGS growth. Purge of excess AGS is required for stabilizing the sludge age, the underlying bacterial community structure, and BNR.

Granule formation and stability for optimized nitrogen and phosphorus removal
An optimization of the start-up with the goal to achieve granulation as well as BNR efficiently was carried by using a multi-parameter study according a Hadamard experimental design where “n” parameters can be tested with “n+1” experiments. For each parameter two states had to be defined and the parameters were considered to be independent. The following parameters were tested: organic substrate feeding (constant load or biomass specific load), aeration (constant rate or pseudo-anoxic phases), addition of allylthiourea during the first 10 days (with or without), air flow (1.5 m s-1 or 3.0 m s-1), biomass settling mode (decrease from 30 minutes to five minutes within 10 days or 20 days with 30 minutes followed by stepwise decrease to 5 minutes within 10 days), pH (7.0-7.3 or 7.5-7-8), and temperature (15°C or 20°C). The end points that were defined to enable the comparison of the different runs were the number of days needed to achieve certain removal rate for C, N, and P removal. The analysis of the first results obtained indicated that temperature had a great influence as well as the organic substrate feeding. This was followed by the aeration and the settling mode. The best combination of these parameters achieved granulation with efficient C, N, and P removal within 3-4 weeks starting with activated sludge from a WWTP with complete BNR (Thun, Switzerland).
The reactor that produced the optimal AGS was operated further to ensure that the granules were stable concerning structure, settling behavior as well as nutrient removal performances. During this operation phase the possibility to remove nitrogen over nitrite and not nitrate was tested too. This possibility was tested by choosing operating conditions (aeration) that favored the growth of ammonium-oxidizing bacteria and could lead to the wash-out of nitrite-oxidizing bacteria (NOB). These tests were successful and it could also be shown that this phenomenon could be reversed by changing operation conditions again. Biomass samples are currently under investigation to show whether NOB decrease and increase during these different phases.