In today’s progressively urbanised society, there is an increasing need for cost-effective, environmentally sound technologies for the removal of nutrients (carbon, phosphorous, nitrogen) from polluted water. Nitrogen removal from wastewater is the focus of this thesis. Conventional nitrogen removal requires the two processes of aerobic nitrification followed by anoxic denitrification, which is driven by remaining reducing power. While most wastewaters contain a significant fraction of reducing power in the form of organic substrate, it is difficult to preserve the reducing power required for denitrification, due to the necessary preceding aerobic oxidation step. Consequently, one of the major limitations to complete N-removal in traditional wastewater treatment systems is the shortage of organic carbon substrate for the reduction of oxidised nitrogen (NO2-, NO3-), produced from nitrification. This thesis followed two main research themes that aimed to address the problem of organic carbon limitation in nitrogen removal from wastewater, by management of the oxygen supply. The first theme was the study of N-removal by simultaneous nitrification and denitrification (SND) in the novel reactor type, the sequencing batch reactor (SBR). It was aimed to increase understanding of PHB metabolism and the limiting factors of SND and then to develop a suitable on-line control strategy to manage the oxygen supply and optimise nitrogen removal by SND. The second main research theme was the application of the CANON(Completely Autotrophic Nitrogen-removal Over Nitrite) process for nitrogen removal from wastewater; a novel process that requires minimal oxygen supply and has the potential to completely circumvent the requirement for organic substrate in nitrogen removal because it is catalysed by autotrophic microorganisms – Anammox (anaerobic ammonium oxidisers) and aerobic nitrifiers. For study of the SND process, a completely automated 2 L sequencing batch reactor was developed with on-line monitoring of the dissolved oxygen concentration, pH and oxidation-reduction (ORP) potential. The SBR was operated continuously for up to 2 years and, due to its separation of different phases by time, enabled the study and optimisation of different microbial activities, including acetate uptake and conversion to PHB (feast phase), PHB hydrolysis and consumption (famine phase), nitrification and denitrification (and SND). All experimental work was performed using a mixed culture Project summary and acetate as the organic substrate. Acetate consumption and PHB production was studied under different oxygen supply rates to establish conditions that allow maximum conversion of acetate to PHB during the feast phase. Lower DO supply rates (kLa 6 – 16 h-1) resulted in preservation of a higher proportion of acetate as PHB than at higher DO supply rates (kLa 30 and 51 h-1). Up to 77 % of the reducing equivalents available from acetate were converted to PHB under O2-limitation, as opposed to only 54 % under O2-excess conditions, where a higher fraction of acetate was used for biomass growth. A metabolic model based on biochemical stoichiometry was developed that could reproduce the trends of the effect of oxygen on PHB production. Experimental findings and simulated results highlighted the importance of oxygen control during the feast phase of an SBR in preserving reducing power as PHB. To develop an oxygen management strategy for the aerobic famine phase,the effect of the dissolved oxygen (DO) concentration on SND, using PHB as the electron donor, was investigated. There was a clear compromise between the rate and the percentage of SND achieved at different DO concentrations. A DO setpoint of 1 mg L-1 was optimal for both the percentage of SND (61 %) and rate of SND (4.4 mmol N. Cmol X-1. h-1). Electron flux analysis showed that most SND activity occurred during the first hour of the aerobic famine period, when the oxygen uptake rate (due to NH4 + and PHB oxidation) was highest. Aerated denitrification ceased as soon as NH4 + was depleted. The presence of NH4 + provided an oxygen “shield”, preventing excessive penetration of oxygen into the flocs and creating larger anoxic zones for SND. PHB degradation was first order with respect to the biomass PHB concentration (dfPHB/dt = 0.19 . fPHB). The slow nature of PHB degradation made it a suitable substrate for SND, as it was degraded at a similar rate to ammonium oxidation. While DO control during the aerobic famine phase could increase nitrogen removal via SND, total N-removal in the SBR was still limited by the availability of reducing power(PHB) in the anoxic phase. The length of the aerobic phase needed to be minimised to prevent over-oxidation of PHB after NH4 + depletion. The specific oxygen uptake rate (SOUR) was found to be an effective on-line parameter that could reproducibly detect the end-point of nitrification. A simple method was developed for continuous, on-line measurement of the SOUR, which was used for automated adjustment of the aerobic phase length. Minimisation of the aerobic phase length by feedback control of the Project summary SOUR improved nitrogen removal from 69 % (without phase length control) to 86 %, during one cycle. The SOUR-aeration control technique could successfully adapt the aerobic phase length to varying wastewater types and strengths and to varying aeration conditions. The medium- and long-term effects of oxygen management on nitrogen removal was investigated by operating the SBR continuously for up to one month using DO control throughout all stages of the SBR, i.e. oxygen-limitation during the feast phase, a DO setpoint of 1 mg L-1 during the famine phase and SOUR controlled aerobic phase length. Complete oxygen management resulted in minimisation of the amount of PHB that was oxidised aerobically in each SBR cycle and caused an accumulation of cellular PHB over time. The increased availability of PHB during aeration resulted in a higher SOUR and increased N-removal by SND from 34 to 54 %. After one month of continuous SBR operation, the settling efficiency of the biomass improved from 110 mL . g-1X to less than 70 mL . g-1X and almost complete N-removal (9 %) was achieved via SND during aeration, however at a reduced rate (1.5 mmol Cmol X-1 h-1). Therefore, long-term oxygen management resulted in biomass with improved settling characteristics and a higher capacity for SND. Results of the first main research theme highlighted the importance of aeration control throughout all stages of the SBR for maximum N-removal via SND. The CANON process was investigated as an alternative to the use of conventional activated sludge for treatment of wastewaters limited by organic carbon substrate. The initial study of the CANON process was performed at the Kluyver Laboratory in Delft, the Netherlands, using an already established Anammox enrichment culture. The effect of extended periods of NH4 +-limitation on the CANON microbial populations was studied, to examine their ability to recover from major disturbances in feed composition. The CANON process was stable for long periods of time until the N-loading rate reached below 0.1 kg N m 3 day-1, when a third population of bacteria developed in the system (aerobic nitrite oxidisers), resulting in a decrease in N-removal from 92 % to 57 %. Nitrite oxidisers developed due to increased levels of oxygen and nitrite. This highlighted the requirement for oxygen control during the CANON process to prevent increased DO levels and growth of undesired microbes. To initiate the CANON process from a local source, Anammox was enriched from local activated sludge (Perth, Western Australia). FISH analysis (fluorescence in situ hybridisation) of the enriched Anammox strain showed that it belonged to the Order Planctomycetales, Project summary the same as all other identified Anammox strains, but represented a new species of Anammox. The enrichment culture was not inhibited by repeated exposure to oxygen, allowing initiation of an intermittently-aerated CANON process to achieve sustained, completely autotrophic ammonium removal (0.08 kg N m-3 day-1) for an extended period of time, without any addition of organic carbon substrate. Dissolved oxygen control played a critical role in achieving alternating aerobic and anaerobic ammonium oxidation. The main conclusion drawn from the study is the important role of oxygen management in achieving improved nitrogen removal. A careful oxygen management strategy can minimise wastage of reducing power to improve PHB-driven SND by activated sludge and can prevent major disturbances to the population balance in the CANON system. Oxygen management has the potential to reduce aeration costs while significantly improving nitrogen removal from wastewaters limited by organic carbon.

A pilot scale nitrogen (N) removal system was constructed and operated for approximately 365 days and was designed to remove inorganic total ammonia nitrogen (TAN) from solids-separated dairy manure. An anaerobic fermenter, upstream of the N removal reactor, produced volatile fatty acids (VFAs), to be used as an electron donor to fuel denitrification, and TAN at a COD:N ratio of 18:1. However, sufficient amounts of non-VFA COD was produced by the fermenter to fuel the denitrification reaction at an average NO3- removal rate of 5.3 ± 2 mg/L NO₃--N. Total ammonia N was removed from the fermenter effluent in an N removal reactor where a series of aerobic and anoxic zones facilitated aerobic TAN oxidation and anoxic NO₃- and NO₂- reduction. The minimum dissolved oxygen (DO) concentration allowing for complete TAN removal was found to be 0.8 mg/L. However, TAN removal rates were less than predicted using default nitrifying kinetic parameters in BioWin®, a biological modeling simulator, which indicated the presence of a nitrification inhibitor in fermented dairy manure. Furthermore, an N balance during the aerobic zone indicated that simultaneous nitrification-denitrification (SND) was occurring in the aerobic zone of the N removal reactor and was most apparent at DO concentrations below 1.3 mg/L. A series of nitrite generation rate (NGR) experiments confirmed the presence of an inhibitor in fermented dairy manure. A model sensitivity analysis determined that the most sensitive ammonia oxidizing bacteria (AOB) kinetic parameters were the maximum specific growth rate, , and the substrate half saturation coefficient, . Nitrifying inhibition terms of competitive, non-competitive, mixed competitive, and un-competitive were applied to the growth rate equation in BioWin® but an accurate representation of the observed TAN removal rates in the pilot scale system could not be found by adjusting the kinetic parameters alone. Reducing the default BioWin® hydrolysis rate by approximately 50% produced a more accurate calibration for all inhibition terms tested indicating that the hydrolyization of organic N in dairy manure is less than typical municipal waste water. / Master of Science