Submitted in fulfillment of the degree of Master of Applied Science: Biotechnology, Durban University of Technology, Durban, South Africa, 2017. / Domestic wastewater contains a high nutrient load, primarily in the form of Carbon (C), Nitrogen (N), and Phosphorous (P) compounds. If left untreated, these nutrients can cause eutrophication in receiving environments. Biological wastewater treatment utilizes a suspension of microorganisms that metabolize this excess nutrient load. Nitrogen removal in these systems are due to the synergistic processes of nitrification and denitrification, each of which requires its own set of operating parameters and controlling microbial groups. An alternative N-removal pathway termed the anammox process allows for total N-removal in a single step under anoxic conditions. This process, mediated by the anammox bacterial group, requires no organic carbon, produces negligible greenhouse gases and requires almost 50 % less energy than the conventional process, making it a promising new technology for efficient and cost-effective N-removal. In this study, a sequencing batch reactor (SBR) was established for the autotrophic removal of N-rich wastewater through an anammox-centric bacterial consortia. The key microbial members of this consortia were characterized and quantified over time using molecular methods and next generation sequencing to determine if the operational conditions had any effect on the seed inoculum population composition. Additionally, local South African wastewater treatment plants were screened for the presence of anammox bacteria through 16S rRNA amplification and enrichment in different reactor types. A 3 L bench scale SBR was inoculated with active biomass (~ 5 % (v/v)) sourced from a parent anammox enrichment reactor, and maintained at a temperature of 35 °C ± 1 °C. The reactor was fed with a synthetic wastewater medium containing no organic C, minimal dissolved oxygen (< 0.5 mg/L), and N in the form of ammonium and nitrite in the ratio of 1:1.3. The reactor was operated for a period of 366 days and the effluent ammonium, nitrite and nitrate were measured during this period. The hydraulic retention time was controlled at 4.55 days from Day 1 to Day 250, and thereafter shortened to 1.52 days from Day 251 to Day 360 due to an increased nitrogen removal rate (NRR). During Phase I of operation (Day 1 to Day 150), the reactor performance gradually increased up to an NRR of ~160 mg N/day. During Phase II (Day 151 to Day 250), the overall reactor performance decreased with the NRR decreasing to ~90 mg N/day, while Phase III (Day 251 to Day 366) displayed a gradual recovery of NRR back to the reactor optimum of ~160 mg N/day. The accumulation of nitrate in the effluent during the latter parts of Phase II and Phase III, coupled with oxygen ingress (~2.1 mg/L) in the same period, indicated that it was not the anammox pathway that was dominating N-removal within the reactor, but more likely the second half of the nitrification pathway mediated by the nitrite oxidizing bacteria (NOB). This was further confirmed through molecular analysis, which indicated that the bacterial population had shifted significantly over the course of reactor operation. Quantitative PCR methods displayed a decrease in all the key N-removing population groups from Day 1 to Day 140, and a marginal increase in anammox and aerobic ammonia oxidizing bacteria from Day 140 – Day 260. From Day 300 onwards, NOB had started dominating the system, simultaneously suppressing the growth of other N-removing bacterial groups. Despite this, the NRR peaked during this period, indicating an alternative mechanism for ammonia removal within the reactor system. A total population analysis using NGS was also performed, which corroborated the QPCR results and displayed a population shift away from anammox bacteria towards predominantly NOB and members of the phylum Chloroflexi. The proliferation of aerobic NOB and Chloroflexi, and the suppression of anammox bacteria, indicated that DO ingress was indeed the primary cause of the population shift within the reactor. Despite this population shift, N-removal within the reactor remained high. New pathways have recently emerged which implicate these two groups as potential N oxidizers, with specific NOB groups showing the ability for oxidation of ammonia through the comammox process, and members of the Phylum Chloroflexi being capable of nitrite reduction. This could imply that an alternate pathway was responsible for the majority of N-removal within the system, in addition to the anammox and conventional nitrification pathways. Additionally, in an attempt to detect a local anammox reservoir, eleven wastewater systems from around South Africa were screened for the presence of anammox bacteria. Through direct and nested PCR-based screening, anammox bacteria was not detectable in any of the activated sludge samples tested. Based on the operating conditions of the source wastewater systems, a subset of three sludge samples were selected for further enrichment. After 60-110 days of enrichment in multiple reactor configurations, only one reactor sample tested positive for the presence of anammox bacteria. Although this result indicates that anammox bacteria might not be ubiquitous within every biological wastewater system, it is more likely that anammox bacteria might only be present at undetectable levels, and that an extended enrichment prior to screening is necessary for a true representation of anammox bacterial prevalence in an environmental sample. / M

In order to size the anoxic reactors in nutrient (N and P) removal activated sludge plants, it is essential to know the denitrification kinetics that are operative in such systems. To date, denitrification kinetics have been accurately defined only for systems that remove N alone; little research has been performed on denitrification in N and P removal plants.

M.S.

The effects of effluent type, effluent loading rate, dosing interval, and temperature on denitrification in low pressure distribution, on-site wastewater treatment and disposal systems (OSWTDS) were evaluated in this study. The treatments were surface and subsurface soil horizons; nitrified and non-nitrified wastewaters; 0.5, 1.0, and 1.5 times the Virginia Department of Health (VDH 1989) recommended wastewater loading rate; 24 and 48 hour dosing intervals; and summer and winter temperatures. Surface and subsurface soil cores were collected from a Groseclose silt loam soil (clayey, mixed, mesic Typic Hapludult) and subjected to the various treatments. The effects of the treatments on denitrification were evaluated based on analyses of leachate from the cores, soil chemical analyses, and microcosm studies to estimate actual denitrification activity. A model was developed from the study that estimated the mean N₂O production for each combination of experimental treatments. The results of the study and the model indicate that denitrification can be enhanced in OSWTDS by the application of non-nitrified wastewater at one-half the VDH recommended loading rate, or 1.25 cm/day, for surface soil horizons (30 min inch⁻¹ percolation rate) using a 48 hour dosing interval. A field study was conducted on a Lowell silt loam soil (fine, mixed, mesic Typic Hapludalf). Denitrification was measured at this site using acetylene blocking and the results compared to those predicted by the denitrification model developed from the laboratory data. The field measurements of denitrification based on N₂O concentration in the soil atmosphere were three orders of magnitude higher than that predicted by the model. It was concluded that the laboratory techniques can be used to determine optimum method of operation for denitrification in a low pressure distribution system, but it cannot be used to determine the field design loading rates. / Ph. D.

In the recent past more and more attention has been given by sanitary engineers to the problem of nitrogen removal in sewage treatment. This attention is brought about by several problems associated with nitrogen. First, in some locations, such as our southwest United States, there is both an essentially constant supply of water and an increasing demand for water. This calls for water recycling. or reuse, as a means of meeting the demand for potable water. An accumulation of impurities, such as compounds of nitrogen, can limit the recycling of water. One such substance is nitrate nitrogen, which when in excess of ten parts per million can cause the disease methemo-globinemia in bottle-fed infants. / Master of Science

Thesis (MScEng) -- Stellenbosch University, 2014. / ENGLISH ABSTRACT: Urine contains up to 80% of nitrogen, 50 % of phosphates and 90 % of potassium of the total load in domestic wastewater but makes up less than 1% of the total volume (Larsen et al., 1996). The source separation and separate treatment of this concentrated waste stream can have various downstream advantages on wastewater infrastructure and treated effluent quality. The handling of undiluted source separated urine however poses various challenges from the origin onward. The urine has to be transported to a point of discharge and ultimately has to be treated in order to remove the high loads of organics and nutrients. Wilsenach (2006) proposed onsite treatment of source separated urine in a sequencing batch reactor before discharging it into the sewer system. This study focused on the treatment of urine in a sequencing batch reactor (SBR) primarily for removal of nitrogen through biological nitrification-denitrification. The aim of the study was to determine nitrification and denitrification kinetics of undiluted urine as well as quantification of the stoichiometric reactions. A further objective was to develop a mathematical model for nitrification and denitrification of urine using experimental data from the SBR. The SBR was operated in 24 hour cycles consisting of an anoxic denitrification phase and an aerobic nitrification phase. The sludge age and hydraulic retention time was maintained at 20 days. pH was controlled through influent urine during volume exchanges. Undiluted urine for the study was obtained from a source separation system at an office at the CSIR campus in Stellenbosch. Conditions in the reactor were monitored by online temperature, pH and ORP probes. The OUR of the system was also measured online. One of the main challenges in the biological treatment of undiluted urine was the inhibiting effect thereof on nitrification rate. The anoxic mass fraction was therefore limited to 17 % in order to allow longer aerobic phases and compensate for the slow nitrification rates. Volume exchanges were also limited to 5% of the reactor volume in order to maintain pH within optimal range. Samples from the reactor were analysed for TKN, FSA-N, nitrite-N, nitrate-N and COD. From the analytical results it was concluded that ammonia oxidising organisms and nitrite oxidising organism were inhibited as significant concentrations of ammonia-N and nitrite-N were present in the effluent. It was also concluded that nitrite oxidising organisms were more severely inhibited than ammonia oxidising organisms as nitrate-N was present in very low concentrations in the effluent and in some instances not present at all. Ultimately the experimental system was capable of converting 66% of FSA-N to nitrite- N/nitrate-N of which 44% was converted to nitrogen gas. On average 48% of COD was removed. A mathematical model was developed in spreadsheet form using a time step integration method. The model was calibrated with measured online data from the SBR and evaluated by comparing the output with analytical results. Biomass in the model was devised into three groups, namely heterotrophic organisms, autotrophic ammonia oxidisers (AAO) and autotrophic nitrite oxidisers (ANO). It was found that biomass fractionation into these three groups of 40% heterotrophs, 30% AAO and 30% ANO produced best results. The model was capable of reproducing the general trends of changes in substrate for the various organism groups as well as OUR. The accuracy of the results however varies and nearexact results were not always achievable. The model has some imperfections and limitations but provides a basis for future work.

This doctoral research provides an advanced molecular approach for the investigation of microbial structure and function in response to operational conditions of biological nitrogen removal (BNR) processes, including those leading to direct production of a major greenhouse gas, nitrous oxide (N₂O). The wastewater treatment sector is estimated to account with 3% of total anthropogenic N₂O emissions. Nevertheless, the contribution from wastewater treatment plants (WWTPs) is considered underestimated due to several limitations on the estimation methodology approach suggested by the Intergovernmental Panel on Climate Change (IPCC). Although for the past years efforts have been made to characterize the production of N₂O from these systems, there are still several limitations on fundamental knowledge and operational applications. Those include lack of information of N₂O production pathways associated with control of aeration, supplemental organic carbon sources and adaptation of the microbial community to the repeated operational conditions, among others. The components of this thesis, lab-scale investigations and full-scale monitoring of N₂O production pathways and emissions in conjunction with meta-omics approach, have a combined role in addressing such limitations. Lab-scale experiments imposing short-term anoxic-aerobic cycling on partial- and full-nitrification based processes were conducted to investigate the microbial response to N₂O production. Interestingly, it was determined that full-nitrification systems could be a higher contributor to N₂O production and emissions than partial-nitrification. While it has been reported in the literature a higher contribution from the latter when the microbial community is not subjected to oxygen cycling conditions. Following the knowledge obtained with a single anoxic-aerobic cycle imposed to nitrifying communities, long-term adaptation of the microbial community to continued anoxic-aerobic cycling and its impact on N₂O production were investigated through a meta-omics approach. Long-term studies are particularly significant regarding engineered systems, where the microorganisms are continually subjected to cycling conditions again and again. A microbial adaptation at the RNA level was identified on both autotroph and heterotroph organisms. The transcripts of the metabolic pathways related to NO and N₂O production (nir, nor) and consumption (nor, nos) were initially induced followed by a gradual decline, leading to a parallel reduction in gaseous emissions over time. Other pathways not typically interrogated in conjunction with the nitrogen metabolism, such as electron transport chain and carbon fixation were also investigated and revealed a mechanism to overcome the imbalance in electron flow and generation of proton motive force (increased transcription of terminal oxidase genes, cco and cox) to uphold carbon fixation during continued cycling. The second part of this thesis focuses on full-scale WWTPs, where it is crucial to determine specific nuances of the systems’ dynamics and of the different types of treatment that may contribute to increased production and emissions of N₂O. For that purpose, two distinct BNR systems not usually considered and studied in terms of N₂O production and emissions were chosen. First, a separate centrate treatment (SCT) process employing glycerol as the supplemental carbon source was monitored. Significantly, this system was found to have one of the highest levels of N₂O production and emission report thus far. Glycerol revealed to foster a microbial community (i.e. Burkholderiales, Rhodobacterales and Sphingomonadales) that stores internal carbon and promote partial denitrification, leading to accumulation of nitrite and N₂O [7-11]. Second, both fixed- and moving-bed biofilm BNR systems were investigated. The overall N₂O emission fractions for the Integrated Fixed-Film Activated Sludge (IFAS)(0.09 – 1.1% infl-TKN) and denitrification filters (0.11 – 1.4% infl-TN) were similar to the reported emissions from suspended growth activated sludge systems [4-6]. For the IFAS system, aqueous and gaseous N₂O profiles paralleled the diurnal variability on influent nitrogen load. The production of N₂O was significantly correlated with ammonia concentration (p<0.05, r=0.91), suggesting the production through hydroxylamine oxidation pathway. Denitrification filters displayed a very peculiar pattern on N₂O emissions associated with intermittent operational cycles (i.e. nitrogen release cycle and backwash). These intrinsic operations of the denitrification filters contributed to transient oxygen conditions and nearly the entire N₂O emissions through gaseous stripping and production by inhibition of denitrification. Similarly to suspended growth systems, process design and operations demonstrated to also play an important role in N₂O emissions from attached growth processes. Finally, aeration strategies for energy efficient conventional nitrification based on the microbial community development and its associated performance was investigated in lab-scale. It was demonstrated that using the same air supply rate, continuous and intermittent aeration resulted in completely different microbial structure. Consequently, distinct kinetics and nitrification performance were observed. The aeration rate could be minimized (resulting in reduction in energy consumption) for high ammonia removal efficiency and lower N₂O emissions, as long as the process is designed accordingly to the microbial ecology developed in such conditions. In sum, the microbial structure, function and connection of metabolic pathways of complex engineered microbial communities as applicable to BNR systems and its operations were investigated in detail. From an engineering perspective, this dissertation provides an advancement on the molecular approach to characterize structure and function of microbial responses to engineered operations beyond the business-as-usual target genes, which can eventually result in better design and control of engineered BNR processes. This study offers more than an improved scientific understanding of the complex microbial environment and direct engineering applications. It connects sanitation with water quality and the greenhouse gas effect by prioritizing concurrent enhanced biological nitrogen removal and mitigation of N₂O production and emission. Ultimately the implications of the result presented herein can provide economical, environmental, health benefits for the society.

University of Macau / Faculty of Science and Technology / Department of Civil and Environmental Engineering

Combined treatment technologies for the removal of waste carbon, nitrogen, and/or sulfur under anoxic/anaerobic conditions have recently received considerable attention. It has been reported that nitrate and/or reduced N-oxides, such as nitrite (NO2-), nitric oxide (NO), and nitrous oxide (N2O), which are products of denitrification, suppress methanogenesis. Research was conducted to investigate the effect of N-oxides and sulfide on mixed, mesophilic (35oC) methanogenic cultures, along with the effect of the type of electron donor on the kinetics and pathway of nitrate reduction. Among all N-oxides tested, NO exerted the most and nitrate exerted the least inhibitory effect on the fermentative/methanogenic consortia. Long-term exposure of a methanogenic culture to nitrate resulted in an increase of N-oxide reduction and a decrease of methane production rates. Sulfide addition to sulfide-free enriched cultures resulted in inhibition of NO2-, NO, and N2O reduction causing accumulation of these intermediates, which in turn inhibited methanogenesis and fermentation. In nitrate-amended, sulfide-acclimated cultures, nitrate reduction occurred via dissimilatory nitrate reduction to ammonia (DNRA); thus, accumulation of N-oxides was avoided and inhibition of methanogenesis was prevented. The nitrate reduction rates in cultures fed with different electron donors followed the descending order: H2/CO2 > acetate > glucose > dextrin/peptone > propionate. Denitrification was observed in the propionate-, acetate-, and H2/CO2-fed cultures regardless of the COD/N value. Both denitrification and DNRA were observed in the dextrin/peptone- and glucose-fed cultures and the predominance of either of the two pathways was a function of the COD/N value. Nitrate reduction processes were incorporated into the IWA Anaerobic Digestion Model No. 1 (ADM1) in order to account for the effect of nitrate reduction processes on fermentation and methanogenesis. The extended ADM1 described the experimental results very well. Model simulations showed that process interactions during nitrate reduction within an overall methanogenic system cannot be explained based on only stoichiometry and kinetics, especially for batch systems and/or continuous-flow systems with periodic, shock nitrate loads. The results of this research are useful in predicting the fate of carbon-, nitrogen-, and sulfur-bearing waste material, as well as in understanding microbial process interactions, in both natural and engineered anoxic/anaerobic systems.

published_or_final_version / Civil and Structural Engineering / Master / Master of Philosophy