Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A conceptual model for apparent free ammonia inhibition in wastewater systems Simm, Robert 2004

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2005-99550x.pdf [ 16.83MB ]
Metadata
JSON: 831-1.0063410.json
JSON-LD: 831-1.0063410-ld.json
RDF/XML (Pretty): 831-1.0063410-rdf.xml
RDF/JSON: 831-1.0063410-rdf.json
Turtle: 831-1.0063410-turtle.txt
N-Triples: 831-1.0063410-rdf-ntriples.txt
Original Record: 831-1.0063410-source.json
Full Text
831-1.0063410-fulltext.txt
Citation
831-1.0063410.ris

Full Text

A C O N C E P T U A L M O D E L F O R A P P A R E N T F R E E A M M O N I A I N H I B I T I O N I N W A S T E W A T E R S Y S T E M S by R O B E R T A N T H O N Y S I M M B . A . S c , The University o f British Columbia, 1985 M . A . S c . , The University of Brit ish Columbia, 1988 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F P H I L O S O P H Y in T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Department of C i v i l Engineering) We accept this thesis as conforming to t h e rennirp^ s t a n d a r d T H E U N I V E R S I T Y OF B R I T I S H C O L U M B I A October 2004 © Robert Anthony Simm, 2004 A B S T R A C T The original hypothesis for this research program was that free hydroxylamine, and not free ammonia, is the true cause of nitrite accumulation in wastewater treatment systems. This hypothesis was developed based .upon the observation that nitrite accumulation appears to be a function of both free ammonia and dissolved oxygen concentration, as well as the work of Yang and Alleman (1992) who demonstrated hydroxylamine production in batch experiments. The starting point of the research program was the development of a suitable hydroxylamine analysis method for wastewater applications. The method developed as part of this thesis combines simplicity, good yields, and a relatively low detection limit. Measurable concentrations of hydroxylamine were not detected in completely stirred tank reactor (CSTR) or sequencing batch reactor (SBR) systems treating synthetic waste and exhibiting nitrite accumulation. In addition, even though hydroxylamine production coincided with nitrite accumulation in batch experiments, several lines of evidence suggested it was unlikely that hydroxylamine was the primary cause of observed nitrite accumulation. Batch tests conducted using mixed liquor samples collected from reactors operated for this research program indicated free ammonia was probably not inhibitory to nitrite-oxidizers at commonly reported concentrations. Therefore, free ammonia inhibition trials were undertaken using pure cultures o f Nitrospira moscoviensis, since Nitrospira spp. were the predominant nitrite-oxidizers in the systems studied for this research program. The results of these experiments support the conclusion that free ammonia is probably not inhibitory to nitrite-oxidizing organisms in wastewater systems. i i Nitrous oxide emissions coincided with nitrite accumulation in every case studied. Several experiments suggested autotrophic denitrification of nitrite by ammonia-oxidizing organisms as the most likely source. F I S H analyses conducted for this research program showed that ammonia and nitrite-oxidizing organisms grew in colonies in close proximity to each other. It is believed that the increase in oxygen utilization and denitrification of nitrite to nitrous oxide by ammonia-oxidizers following a free ammonia perturbation, which results in reduced substrate availability for nitrite-oxidizers, is the true cause o f apparent free ammonia inhibition. A conceptual model explaining apparent free ammonia inhibition of nitrite-oxidizing organisms has been developed and is presented. 111 T A B L E O F C O N T E N T S Abstract i i Table of Contents iv List o f Tables v i List of Figures v i i List of Symbols, Nomenclature and Abbreviations x Preface x i i i Acknowledgements xv i i 1.0 Thesis Background and Objectives 1 1.1 Introduction , 1 1 1.2 Literature Review 4 1.2.1 Introduction 4 1.2.2 Biochemistry of nitrifying organisms. 4 1.2.3 Documented causes of nitrite accumulation in wastewater systems. 17 1.2.3.1 Dissolved oxygen as a factor in nitrite accumulation. 18 1.2.3.2 Temperature as a factor in nitrite accumulation 21 1.2.3.3 The role of p H in nitrite accumulation. 24 1.2.3.4 Free ammonia and nitrous acid. 26 1.2.3.5 Free hydroxylamine. 29 1.2.3.6 Anaerobiosis. 32 1.2.3.7 Solids retention time. 34 1.2.3.8 Acute process loading. 36 1.2.4 The acclimation of nitrite oxidizers to free ammonia. 37 1.3 Objectives and hypotheses to be tested. 41 1.4 Thesis Outline. 43 Bibliography 47 2.0 Hydroxylamine analysis of wastewater samples v ia gas chromatography. 55 3.0 Pre l iminary evaluation of the use of fatty acid ratios for t racking the potential for nitrite accumulation in nitrifying reactors wi th low carbon to nitrogen ratio. 63 4.0 Ni t r i f ier population dynamics following a dissolved oxygen perturbation. 93 5.0 A targeted study on possible free ammonia inhibi t ion of Nitrospira. I l l 6.0 Ni t r i f ier population dynamics in a bench scale conventional activated sludge reactor following an induced free ammonia perturbation. 148 7.0 Mechanisms responsible for apparent free ammonia inhibi t ion in a sequencing batch reactor. 185 8.0 The role of hydroxylamine as a potential inhibi tor of nitrite oxidizers iv in wastewater treatment systems. 231 9.0 Conclusions and Recommendations for Future Research. 287 9.1 Discussion relating manuscripts to each other. 287 9.2 Development of a unified ammonia oxidation model to explain apparent free ammonia inhibition. 292 9.3 Conclusions. 300 9.4 Recommendations and Comments on future research requirements. 303 Bibliography 307 v LIST OF TABLES Table 3.1 - Reactor Influent/Effluent conditions 86 Table 4.1 - Synthetic feed composition for bench scale continuous stirred tank reactor 108 Table 5 . 1 - Feed components for continuous stirred tank reactor system(CSTR) 139 Table 5.2 - Summary of treatments applied to pure cultures of Nitrospira moscoviensis in batch experiments. 140 Table 5.3 - Data summary for pure culture experiment. •• . * 141 Table 6.1 - Influent and effluent characteristics prior to the start of the free ammonia stress. 175 Table 6.2 - Cumulative nitrate in reactor after start of the free ammonia stress 176 Table 6.3 - Estimated ammonia oxidation, nitrite production, and nitrite oxidation rates with time over the first 71 hours after the start of the free ammonia stress 177 Table 7.1 - Summary of tracking study results 224 Table 8.1 - Synthetic feed composition for bench scale continuous stirred tank reactor 275 Table 8.2 - Synthetic feed composition for bench scale sequencing batch reactor (SBR) . 276 Table 8.3 - Summary of experimental series for In situ and prewashed biomass tests with schedule and objectives for supplementary experiments. 277 Table 8.4 - Experimental series for shaker table experiments. 278 Table 8.5 - Experimental series for chemodentrification trials. 279 Table 8.6 - Summary of results of chemodenitrification and autotrophic denitrification trials with hydroxylamine 280 Table 8.7 - Summary of results of S B R perturbation trials relative to base case. 281 LIST OF F IGURES Figure 2.1 - Typical chromatogram for hydroxylamine recovered from a reactor sample. 60 Figure 3.1 - Reactor setup schematic. 81 Figure 3.2 - Average percentage of major fatty acid components during preliminary assessment period. 82 Figure 3.3 - Ratio of 16:1 (9)/16:1(11) versus nitrite-N as percent N O x for R3 (February 1, 2002 - March 3, 2002) 83 Figure 3.4 - Ratio of 16:1(9)/16:1(11) versus headspace nitrous oxide concentration for R3 (February 1, 2002 - March 3, 2002) 84 Figure 3.5 - Ratio of 16:1(9)/16:1(11) versus nitrite-N as a percentage of N O x for reactors R I through R6. 85 Figure 4.1 - Time series plots of results for dissolved oxygen perturbation. 104 Figure 4.2 - F ISH Images for reactor before and after perturbation. 105 Figure 4.3 - F I S H Images of test reactor following dissolved oxygen perturbation. 106 Figure 4.4 - F I S H Images for a C S T R operating with identical feed, high D.O. , and 10-day SRT for several months. 107 Figure 5 . 1 - Batch nitrification with and without free ammonia inhibition. 132 Figure 5.2 - Summary of September 6, 2002 batch test. 133 Figure 5.3 - Summary of September 11, 2002 batch test. 134 Figure 5.4 - Summary of January 23, 2003 washed biomass test. 135 Figure 5.5 - Ammonia, nitrite, and nitrate for January 23, 2003 washed biomass test. 136 Figure 5.6 - Summary of February 26, 2003 washed biomass test. 137 Figure 5.7 - Time series of batch nitrate-nitrogen concentration for pure culture trial. 138 Figure 6.1 - Time series plots of reactor ammonia, nitrate, and nitrite concentrations following a free ammonia stress. 169 Figure 6.2 - Time series plots of nitrous oxide emission rate (mg/day) and reactor nitrate nitrogen after the start of the free ammonia stress. 170 Figure 6.3 - System S R T versus time after the start of the free ammonia stress. 171 Figure 6.4 - Ratio of cis(9) hexedecenoic acid/cw(l 1) hexedecenoic acid versus time after the start of the free ammonia stress. 172 Figure 6.5 - R N A concentration in the reactor after the start of the free ammonia stress. 173 Figure 6.6 - F I S H images with anticipated substrate gradients. 174 Figure 7 . 1 - Peak anoxic phase free ammonia concentration. 212 Figure 7.2 - Time series plots of effluent nitrite, nitrate, percentage nitrite and ammonia nitrogen. 213 Figure 7.3 - Time series plots of reactor and effluent mixed liquor suspended solids concentration and system SRT. 214 Figure 7.4 - Time series plot of total ammonia removal over an entire cycle. 215 Figure 7.5 - System performance parameters for March 30, 2003 tracking study. 216 Figure 7.6 - System performance parameters for March 31, 2003 tracking study. 217 Figure 7.7 - System performance parameters for Apr i l 2, 2003 tracking study. 218 Figure 7.8 - System performance parameters for Apr i l 12, 2003 tracking study. 219 Figure 7.9 - System performance parameters for A pr i l 22, 2003 tracking study (set 1). 220 Figure 7.10 - System performance parameters for Apr i l 22, 2003 tracking study (set 2). 221 Figure 7 . 1 1 - Summary of R N A data. 222 Figure 7 . 1 2 - F ISH Images. 223 Figure 8.1 - Nitrosomonas electron transport model (Schmidt, 2003). 268 Figure 8.2 - November 22, 2002 batch test. 269 V l l l Figure 8.3 - Summary of experimental results from shaker table table experiment (trial 3). 270 Figure 8.4 - Summary of experimental results from nitrite and hydroxylamine step study. 271 Figure 8.5 - Time series plot o f nitrous oxide for prewashed biomass test with N 0 2 / N H 3 spikes. 272 Figure 8.6 - Summary of tracking study results for S B R base case prior to perturbation. 273 Figure 8.7 - Summary of tracking study results for influent ammonia increase and anoxic/aerobic phase p H control for S B R . 274 Figure 9.1 - Conceptual model explaining apparent free ammonia inhibition. 306 ix LIST OF SYMBOLS, NOMENCLATURE AND ABBREVIATIONS A D P - Adenosine Diphosphate A T P - Adenosine Triphosphate A N A M M O X - Anaerobic Ammonia Oxidation A O B - Ammonia Oxidizing Organisms B C A - Bicinchoninic acid B O D - Biochemical Oxygen Demand B S A - Bovine Serum Albumein C O D - Chemical Oxygen Demand C S T R - Continuous Stirred Tank Reactor D G G E - Denaturing Gradient Gel Electrophoresis D N A - Deoxyribonucleic A c i d D O - Dissolved Oxygen D S M Z - Deutsche Sammlung von Mikroorganismen und Zellkulturen Gmb H F A M E - fatty acid methyl esters FID - flame ionization detector F I S H - Fluorescent In-Situ Hybridization G C - gas chromatogram G C / M S - gas chromatography/mass spectrometry HNO2 - Nitrous acid H R T - Hydraulic retention time Ks - half-saturation coefficient M C R T - Mean cell residence time N A D - Nicotinamide adenine dinucleotide N A D P - Nicotinamide adenine dinucleotide phosphate NH2OH - Hydroxylamine N O - Nitr ic Oxide N 0 2 " - Nitrite NO3" 2 -Nit ra te N O x - oxidized nitrogen (sum of nitrate + nitrite) N2O - Nitrous Oxide N O B - Nitrite oxidizing bacteria NH3 - Free ammonia N H / - Ammonium Ion O R P - Oxidation-reduction potential ppb - parts per bi l l ion ppbv - parts per bi l l ion by volume R F L P - restriction fragment length polymorphism R N A - Ribonucleic acid r R N A - Ribosomal Ribonucleic acid S B R - Sequencing Batch Reactor S H A R O N - Single Reactor High Activity over Nitrite (name of a proprietary process) S N D - Simultaneous Nitrification-Denitrification SRT - Solids retention time T C A - Tricarboxylic A c i d cycle T K N - t o t a l Kjeldahl nitrogen , < . U B A F - Upflow biological aerated filter U.S . E P A - United States Environmental Protection Agency V A S - Volatile attached solids Y - bacterial yield p - maximum specific growth rate u M - micromolar concentration u L - microliters P R E F A C E This thesis has been prepared in manuscript-based format. A manuscript-based thesis, as described by the Faculty o f Graduate Studies at the University o f British Columbia, is a collection of published, in-press, accepted, submitted or draft manuscripts. The body of this thesis has been separated into nine main chapters. Chapter 1 is an introductory chapter covering the main objectives of the thesis and providing background to the problem. The results of the research program are presented in Chapters 2 through 8. Chapter 9 summarizes the thesis conclusions in the form o f a conceptual model, outlines the thesis contribution, and lists recommendations for further research. Section 1.3 - Literature Review has been submitted for publication as "a state-of-the art review/critque of nitrification". Portions of chapter 3 have been published previously in a journal paper, whereas variations of chapters 2 and 4 and 5 through 8 have recently been submitted for publication. While the general theme and content are the same, the material presented in chapters 2 through 9 has been modified based upon additional analysis and editing completed while the individual papers were under review. The following is a summary of the submitted and published material that pertain to this thesis: 1. Simm, R . A . , Ramey, W . D . , and Mavinic, D.S. 2004. Preliminary evaluation of the use of fatty acid ratios for tracking nitrite accumulation in nitrifying reactors with low carbon to nitrogen ratio. Journal of Environmental Engineering and Science, N R C Research Press. (Volume 3, pages 31-40). x i i i 2. Simm, R . A . , Ramey, W . D . , and Mavinic, D.S. 2004. Nitrifier population dynamics in a bench scale activated sludge reactor following an induced perturbation, (accepted for publication with minor revisions in Journal of Environmental Engineering and Science June 2004). 3. Simm, R . A . , Ramey, W.D. , and Mavinic, D.S. ' 2004. Mechanisms responsible for apparent free ammonia inhibition in a Sequencing Batch Reactor, (submitted to the A S C E Journal of Environmental Engineering July 2004). 4. Simm, R . A . , Ramey, W . D . , and Mavinic, D.S. 2004. A targeted study on possible free ammonia inhibition of Nitrospira (submitted to the Journal of Environmental Engineering and Science for publication M a y 2004). 5. Shiskowski, D . M . , Simm, R . A . , Mavinic, D.S. 2003. A n experimental procedure for identifying the biological source of nitrous oxide in anoxic/aerobic biological wastewater treatment systems, (accepted for publication in the Journal of Environmental Engineering and Science, M a y 2004). 6. Simm, R . A . , Ramey, W . D . , and Mavinic, D.S. 2004. The role of hydroxylamine as an inhibitor of nitrite oxidizers in wastewater treatment systems (submitted for publication in Journal of Environmental Engineering and Science, M a y 2004). 7. Simm, R . A . , Ramey, W . D . , and Mavinic, D.S. 2004. Nitrifier population dynamics following a dissolved oxygen perturbation, (submitted to W E F Research Journal for publication July 2004). xiv 8. Simm, R . A . , Ramey, W . D . , and Mavinic, D.S. 2004. State of the art review of the biochemistry of nitrifiers and the potential impact upon nitrogen removal practice in wastewater systems, (submitted to W E F Research Journal for publication August 2004). 9. Simm, R . A . , Ramey, W . D . , and Mavinic , D.S. 2004. Changes in nitrifier population and reactor gas dynamics as a result of changes in SRT. (submitted to W E F Research Journal for publication August 2004). 10. Simm, R . A . , Parkinson, P., Ramey, W . D . , and Mavinic , D.S. 2004. Hydroxylamine analysis of wastewater samples via gas chromatography, (submitted to Environmental Technology for publication July 2004). 11. Simm, R . A . , Ramey, W . D . , and Mavinic, D.S. 2004. A unified ammonia oxidation model to explain commonly observed nitrification phenomenon, (submitted to Water Research for publication July 2004). xv The contents of the submitted and published papers (and of this thesis) represent a collection of my own work. Elements of the papers listed as items 1-4, 6-8, 10, and 11 have been included directly in this thesis. Credit to co-authors acknowledges the advice and review of my supervisors, and any additional insight provided has been referenced appropriately within this thesis. Signature: Signature. Dr. D.S. Mavinic Dr. W . D . Ramey (Senior Co-Authors of published material) Signature: Signature. D . M . Shiskowski, P.Eng. P. Parkinson xv i A C K N O W L E D G M E N T S The Natural Sciences and Engineering Research Council ( N S E R C ) , University of British Columbia, and Stantec Consulting Limited are gratefully acknowledged for providing the funding for this study. M s . Susan Harper and Ms . Paula Parkinson of the Environmental Engineering Laboratory at the University of British Columbia provided essential logistical support for which I am grateful. I would also like to thank Dean Shiskowski, Janice Y i m , and Anindita Tjahjadi for their assistance in collection of laboratory data; my committee members B i l l Oldham, Eric Ha l l , and Victor Lo for their review and comments on the thesis; and also, Fred Koch and Venkat Mahendraker for their support and suggestions. Special thanks are extended to my supervisors Don Mavinic and B i l l Ramey without whom this project would not have been possible. They have both provided much appreciated support and thoughtful advice. I would finally like to thank my wife Janis and children Scott and Rachel who provided endless support and encouragement over the past five-years. xv i i 1.0 T H E S I S B A C K G R O U N D A N D O B J E C T I V E S 1.1 Introduction Nitrogen in its various forms can deplete dissolved oxygen levels in receiving waters, exhibit toxicity towards aquatic life, present a public health hazard, stimulate aquatic growth and affect the suitability of wastewater for reuse. High.nitrogen concentration wastewater is found in the following fields: alcohol production, pectin industry, potato processing industry, slaughterhouses, starch processing industry, metallurgy, landfill leachates, and the petrochemical industry. The current trend in wastewater engineering is to remove nitrogen biologically using single sludge activated sludge systems (U.S. E P A - Nitrogen Control Manual, 1993). The conventional biological process involves the sequential oxidation of ammonia to nitrite and then nitrate, a process referred to as nitrification, followed by reduction of oxidized nitrogen species to nitrogen gas. This latter process is referred to as denitrification. The process inputs include oxygen for the nitrification process and in many cases an external carbon source for denitrification. There is a great deal of interest in optimizing the nitrification /denitrification process by stopping the oxidation of ammonia at nitrite and reducing nitrite alone to nitrogen gas. Numerous authors (Voest et al. 1975; Turk and Mavinic. 1986, 1987, 1989a, 1989b; Balmelle 1992; Chen et al. 1991; Fdz- Polanco et al. 1994, 1996; Garrido et al. 1997a, 1997b; Hyungseok Yoo et al. 1999) have reported the potential benefits of such a process as a 25% reduction in aeration requirements, a 40% reduction in external carbon addition for denitrification, and reduction in anoxic zone volume. A s indicated by Focht and Chang (1975), the process represents an 1 improvement over the general nitrification-denitrification procedure in terms of simplicity and time. The process of nitrification/denitrification via nitrite is referred to here as the nitrate shunt. The nitrate shunt has not yet been applied at full scale and the majority of the work conducted to date, on the process, has been done with laboratory scale reactors treating synthetic wastewater. The primary reason for this limitation appears to be that the mechanisms responsible for sustaining nitrite accumulation- a precondition for nitrification/denitrification via nitrite- are not fully understood. A s indicated by Schmidt et al. (2003) when discussing partial nitrification to a stable nitrite end product: "it is unclear why nitrite oxidizers are inhibited; inhibition of nitrite oxidizers by ammonia and lower affinity for oxygen and/or nitrite have been suggested as possible explanations, but we still lack mechanistic evidence". Studies that elucidate these mechanisms could potentially help in determining the reason(s) for apparent nitrite oxidizer acclimation, a commonly reported cause of process failure (Turk and Mavinic. 1986, 1987, 1989a, 1989b and Fdz- Polanco et al. 1996) and ultimately widespread application of this process technology at full scale. The need for this research was identified over thirty years ago (Focht and Chang 1975). Once the mechanism(s) responsible for nitrite accumulation are fully understood it may be possible to develop appropriate kinetic models that can be used to design a full-scale system. A s indicated by Grady et al. (1999), no kinetic relationships are available to depict all of the inhibitory effects o f free ammonia, believed to be the primary cause of nitrite accumulation (Anthonisen et al. 1976). They conclude that the simple Monod equation is not adequate to depict the kinetics of nitrification when the concentration of ammonia exceeds the level that is normally found in domestic wastewater and alternative expressions should be sought. None of the current kinetic models for nitrification considers the possibilities of hydroxylamine 2 accumulation, autotrophic biological denitrification, and chemodenitrification that are potentially important at high free ammonia concentrations. This incomplete modeling may explain why the desired kinetic relationship that Grady et al. (1999) allude to, has yet to be developed. Aside from assisting in the design of a system operating via the nitrate shunt, the determination of the mechanism(s) responsible for nitrite accumulation would also help in addressing problems associated with nitrite build-up, when such a build-up is unwanted; providing additional economic and environmental benefits. The potential problems associated with unwanted nitrite build-up include increased nitrogenous oxygen demand on the receiving water body, increased chlorine demand during effluent disinfection, and the possibility o f nitrosamine formation. High concentrations of nitrite are believed to be toxic to the aquatic biota (Kelso et al. 1999), and can result when there are large inputs of agriculturally derived nitrogen substrates. The potential benefits associated with the elucidation of the mechanism(s) responsible for nitrite accumulation were the impetus for this study. 3 1.2 Literature Review 1.2.1 Introduction The literature review is separated into three subsections. A state-of-the art review of the biochemistry of nitrifying organisms is presented in Section 1.2.2. This review is followed by a summary of literature on the reported causes of nitrite accumulation (Section 1.2.3). Finally, literature evidence on the reported acclimation response of nitrite-oxidizers to high concentrations of free ammonia is presented and discussed/critiqued in Section 1.2.4. 1.2.2 Biochemistry of Nitrifying Organisms A s indicated by Mobarry et al. (1996), improved process control in nitrifying systems wi l l almost certainly depend upon a better understanding of the microbiology of nitrifying organisms. There have been significant advances in our understanding of the bioenergetics and biochemistry of ammonia and nitrite oxidizers over the last three decades, many of which are just now changing the way in which environmental engineers and soil scientists view both groups of organisms. In the traditional environmental engineering literature, ammonia and nitrite oxidizing organisms are viewed as strict chemolithoautotrophs using ammonia or nitrite for energy and carbon dioxide as sole carbon source. However, several novel principles have been recognized in the microbial conversion of nitrogen compounds (Helmer et al. 1999). These include heterotrophic nitrification, autotrophic denitrification by both ammonia and nitrite-oxidizing organisms, and anaerobic ammonium oxidation with ammonia as electron donor and nitrite as electron acceptor ( A N A M M O X ) . We now know that many nitrifiers can use organic substrates, albeit to a limited extent, in addition to ammonia and nitrite. For example, Krummel and Harms (1982) 4 demonstrated that Nitrosospira grown mixotrophically on formate and acetate had an enhanced growth rate that was approximately 130% of the autotrophic growth rate. Frijlink et al. (1992) reported that endogenous substrates are oxidized by Nitrosomonas europaea between p H 5 to 8 resulting in the generation of a considerable proton motive force. The fact that the metabolic diversity of nitrifying organisms is greater than envisaged by the traditional view has led Prosser (1989) to conclude the conventional view is simplistic at best. Although many traditional engineering texts refer to bacteria of the genera Nitrosomonas and Nitrobacter when referring to ammonia and nitirite-oxidizers respectively, recent research (Wagner et al. 1999; Schramm et al. 1999; and Okabe et al. 1999) indicates that Nitrospira-Ytke bacteria are the main nitrite oxidizing organisms in wastewater systems and not Nitrobacter. In addition, some investigators (Hiorns et al. 1995 and Schramm et al. 1999) report Nitrosospira, and not Nitrosomonas, may be the primary ammonia oxidizers in some wastewater systems. The widespread use of molecular methods in bacterial classification has also changed our understanding of the relationship between these and other organism groups. Although classified together as nitrifiers by most environmental engineers, we now know that the lithotrophic ammonia and nitrite oxidizing organisms are phylogenetically unrelated (Bock et al. 1991). The lithotrophic ammonia oxidizers can be placed into five genera based on cell shape. These genera include Nitrosomonas, Nitrosococcus, Nitrosospira, Nitrosolobus, and Nitrosovibrio. Phylogenetically, ammonia oxidizers are divided into at least two different groups (P and y subdivisions of the Proteobacteria) that are not closely related. The organisms in the P-subdivision of the Proteobacteria include Nitrosomonas europaea, Nitrococcus mobilis, Nitrosopira briensis, Nitrosolobus multiformis and Nitrosovibrio tenuis. Nitrosococcus oceanus is classified in the y subdivision. O f these, the most studied strain is Nitrosomonas europaea. 5 Prosser (1989) argued there was a significant need for more detailed study of different strains and genera of ammonia oxidizers given the significant qualitative and quantitative physiological differences between them. The fact that the electron transport systems for known ammonia oxidizers appear to be identical, as determined by difference spectroscopy (Giannakis et al., 1985; Koops et al., 1991) suggests that our knowledge of the oxidation biochemistry o f Nitrosomonas europaea can generally be applied to most ammonia oxidizers. The majority of the research quoted here was conducted with pure cultures of Nitrosomonas europaea. The specific species used is noted in those cases where Nitrosomonas europaea was not the studied species. The half saturation coefficient for ammonia oxidation decreases with increasing p H , suggesting that NH3 (free ammonia) is the real substrate for ammonia-oxidizing organisms (Suzuki et al. 1974). This possibility is consistent with the finding that the ammonia-oxidizing system of Nitrosomonas is located in the membrane and membranes are highly permeable to uncharged NH3. If ammonia is transported by passive diffusion, the rate o f uptake w i l l increase at high p H values due to an increased concentration gradient across the cell membrane. This implies that ever higher p H values should be advantageous to ammonia-oxidizers; however, Frijlink et al. (1992), found that actively oxidizing cells of Nitrosomonas europaea do not maintain a constant internal p H when the external pH is varied from 5 to 8. This suggests that above the optimum p H value for growth, the advantages of increased availability of N H 3 must be counterbalanced by the need to maintain an internal p H value below that of the external medium. Hydroxylamine oxidation was found to be only moderately pH-sensitive. The overall reaction of ammonia oxidation to nitrite is reported as a two-stage process and the generally accepted reactions are summarized as follows (Hooper, 1989; White, 1995): 6 1) 2 H + + N H 3 + 2e + 0 2 -> N H 2 O H + H 2 0 2 ) N H 2 O H + H 2 0 - > H N 0 2 +4H + + 4e" 3) 2 H + + 0.5 0 2 + 2 e " - » H 2 0 In the first reaction, catalyzed by the enzyme ammonia monooxygenase, ammonia is oxidized to hydroxylamine with one oxygen atom incorporated into hydroxylamine and the other incorporated into water. The cell invests reducing power into the conversion of ammonia to hydroxylamine and consequently this reaction does not result in the net generation of metabolic energy (Frijlink et al., 1992). Schmidt et al. (2001a, 2001b) recently reported that dinitrogen tetroxide, and not oxygen, may be the electron acceptor for Nitrosomonas eutropha. The second reaction, the oxidation of hydroxylamine to nitrous acid, is catalyzed by the enzyme hydroxylamine oxidoreductase that is believed to be located in the periplasmic space. This second reaction is believed to be the energy producing part o f the ammonia oxidation process. The bulk of this energy is used to generate reducing power, which is needed for carbon dioxide fixation. On thermodynamic grounds, 80% of the energy generated by autotrophs is used to fix carbon. Although ammonia-oxidizing organisms can oxidize hydroxylamine, they do not appear to be able to grow solely on this substrate (Frijlink et al.1992, de Bruijn et al. 1995). Frijlink et al. (1992) suggested this was likely the result of this substrate's toxic nature. Ammonia and hydroxylamine oxidation are in a steady state in exponentially growing cells, suggesting the ammonia monooxygenase and hydroxylamine oxidoreductase reactions are coupled. It is hypothesized that this steady state is controlled by the concentration of 7 hydroxylamine since increasing concentrations of hydroxylamine eventually inhibit the ammonia monooxygenase enzyme (Hymann and Woods 1984). Most ammonia-oxidizers have more than one copy of the genes for ammonia monooxygenase and hydroxylamine oxidoreductase respectively (Bock and Wagner, 2003). Hooper et al. (1990) reported the genome of Nitrosomonas contained three copies of the gene for the hydroxylamine oxidoreductase with no plasmids detected. McTavish et al. (1993) found two copies of the gene coding for the presumed active site polypeptide of ammonia monooxygenase in Nitrosomonas europaea. Hooper et al. (1990) speculated that, in all likelihood, there are three different forms of the hydroxylamine oxidoreductase enzyme, each having a different biochemical role depending upon growth conditions. They argue that it is highly unlikely that cells have three identical copies of the gene to facilitate rapid transcription, given that growth on ammonia and CO2 is slow enough that having only one copy of the gene should not provide a rate-limiting barrier to cell growth. Hommes et al. (1998) have suggested the presence of multiple gene copies might allow more-rapid generation of m R N A when the in situ ammonia concentration increases substantially. Bergman et al. (1994) suggest the multiple gene copies allow the organism to modify their gene product ratio to suit specific environmental conditions. Hooper (1989) proposed that ammonia oxidation takes place in the cytoplasm and, assuming cytoplasmic oxidation, hydroxylamine diffuses across the cell membrane to the periplasm where it is oxidized by hydroxylamine oxidoreductase. The electrons then travel to a periplasmic cytochrome c. Here the electron pathway branches. Two electrons travel to ubiquinone and then the ammonia monooxygenase enzyme. The other two electrons travel to oxygen via cytochrome aa3, that is believed to act as a proton pump translocating protons into the periplasm. The enzyme A T P synthase is then able to phosphorylate A D P under the impetus of the higher proton 8 concentration in the periplasm. Schmidt (2003) has recently hypothesized that there is an additional branch point following the periplasmic cytochrome c and that electrons may travel to nitrite reductase, nitric oxide reductase, and possibly nitrous oxide reductase in addition to oxygen. The fact that ammonia-oxidizing organisms can denitrify has been known for some time. Hooper (1968) isolated the enzyme nitrite reductase from cells of Nitrosomonas, demonstrating gas production with hydroxylamine as electron donor for dissimilative nitrite reduction. Mi l le r and Wood (1983) purified a soluble copper protein from Nitrosomonas with cytochrome oxidase and nitrite reductase activity. This protein was found to be elevated in cells grown at a low dissolved oxygen concentration and has a high affinity for O2. Hooper and Terry (1979) proposed that, in nitrifiers, production of N O occurs primarily by the aerobic oxidation of hydroxylamine. They also noted that both N O and N 2 0 were emitted as a result of nitrifier denitrification; however, with hydroxylamine as the electron donor, N2O predominated. Ritchie and Nicholas (1972) report that hydroxylamine was oxidized by cell free extracts to yield nitrite and N 2 0 aerobically, but to yield N 2 0 and N O anaerobically. Cel l extracts also aerobically and anaerobically reduced nitrite to N 2 0 and N O with hydroxylamine as an electron donor. Washed cells of Nitrosomonas also produced N2O by the reduction of nitrite under anaerobic conditions. Ritchie and Nicholas (1972) suggested that the results with washed cells provided evidence for the existence of a nitrite reductase in this organism; this was active either under anaerobic conditions or aerobically, with concomitant oxidation of ammonia to nitrite. This suggests that nitrite may serve as an alternative electron acceptor to oxygen. 9 Anderson and Levine (1986) reported an inverse relationship between the dissolved oxygen concentration and the production of N 2 0 when working with pure cultures of Nitrosomonas europaea. They did not observe the same relationship with N O . They determined N 2 0 emissions depend upon competition between nitrite and oxygen for electrons removed from ammonium during nitrification; the more oxygen available the less N 2 0 released. Under anaerobic conditions, they observed emissions of N O and N 2 0 was directly proportional to the nitrite concentration in the growth medium. Anderson and Levine (1986) reported that, under most conditions, the addition of mercuric chloride to Nitrosomonas europaea cultures caused a rapid decline in the emissions of both N O and N 2 0 , suggesting the gases were of biological origin. However, when high concentrations of nitrite had accumulated or were added to the medium, chemodenitrification was responsible for the production of much of the N O and some of the N 2 0 gas. Stuven et al. (1992) studied pure cultures of Nitrosomonas and Nitrosovibrio and mixed cultures of Nitrosomonas and Nitrobacter to determine the effect of organic matter on the production of nitrogen oxides by ammonia oxidizing organisms. These experiments were performed with and without organic compounds to elucidate the pathway of N O and N 2 0 formation. These investigators showed that organic substances, such as pyruvate or formate, are suitable electron donors for N O and N 2 0 production. In the absence of oxygen only slow pyruvate consumption with concomitant nitrite reduction was measurable. Stuven et al. (1992) showed the production of N O and N 2 0 is the result of two reactions, one enzymatic (the production of hydroxylamine) and one chemical reaction (chemodenitrification of nitrite to N O and N 2 0 ) . In addition, N 2 0 is chemically formed during rapid decomposition of N H 2 O H . Stuven et al. (1992) hypothesized that, whenever additional electrons are generated by the oxidation of pyruvate or formate, they 10 also feed the monooxygenase reaction leading to an imbalance between ammonia and hydroxylamine oxidation and the resulting release of hydroxylamine into the medium. Once hydroxylamine is released into the medium, it can reduce nitrite to nitric oxide and nitrous oxide via chemodenitrification. Bock et al. (1995) have reported that ammonia oxidizers have high denitrification rates when grown in the presence of organic matter. Bock and Wagner (2003) reported that Zart et al. (1996) found that ammonia oxidation rates are in fact low under these conditions. Although N 2 0 and N O production by ammonia-oxidizers can be significant (Hopper et al. 1990), there is some debate on its physiological significance. Ritchie and Nicholas (1972) have suggested this mechanism allows ammonia oxidizers to survive temporary conditions of anaerobiosis. Poth and Focht (1985), suggest that the process functions to: conserve oxygen for use by ammonia monooxygenase, reduce production of nitrite (which may accumulate to toxic levels), and decrease competition for oxygen by nitrite oxidizers, by denying them their source of substrate. Stein and Arp (1998b) studied the impact of nitrite on ammonia-oxidizers and found that Nitrosomonas europaea lost an increasing amount of ammonia oxidation activity upon exposure to increasing concentrations of nitrite. The loss of activity was specific to the ammonia monooxygenase enzyme. The known nitrite oxidizing organisms include Nitrobacter, Nitrococcus, Nitrospina, and Nitrospira. For these organisms, the oxidation of nitrite to nitrate is the energy generating process. Nitrobacter (a-Proteobacteria) can grow heterotrophically, while the remaining nitrite oxidizers, Nitrospina (h-Proteobacteria), Nitrococcus (y-Proteobacteria) and Nitrospira (Nitrospira phylum), are unable to grow heterotrophically (Ehrich et al. 1995). 11 To date, most of the studies on nitrite oxidation have been carried out with pure cultures of Nitrobacter. One of the reasons for this is the difficulty associated with isolating the members of other genera from soil or wastewater. The genera Nitrococcus, Nitrospina, and Nitrospira are generally restricted to marine environments and grow optimally at high salt concentrations. One exception is a Nitrospira strain isolated from soil. The generally accepted reactions of nitrite oxidation by Nitrobacter are as follows: 1) N 0 2 " + H 2 0 - > N 0 3 " + 2 H + +2e 2) 2H + +2e"+l /2 0 2 - > H 2 0 It has been well established that the oxygen atom in the nitrate molecule is derived from water (Kumar et al., 1983, and Hollocher, 1984). It has also been established that nitrite oxidation is coupled to A T P generation in the cytochrome c: 0 2 oxidoreductase region of the Nitrobacter electron transport chain, and that the phosphorylation process does not involve the participation of the pyridine nucleotides or flavins. The generation of reduced pyridine nucleotides, by the electron transfer from ferro-cytochrome c or from nitrite in Nitrobacter, has been shown to be energy-or A T P - dependent and involves the sequence of oxidation phosphorylation reactions. Aleem and Sewell (1981) conducted experiments on the possible coupling mechanisms of nitrite with the Nitrobacter electron transport chain, as well as characteristics of the various oxidoreductase enzyme systems in this organism. These authors concluded that there are two 12 nitrite-oxidizing systems in Nitrobacter. The first system is described by reactions (1) and (2) above. The reactions carried out by the second nitrite oxidizing system are summarized as follows: • 3 ) N 0 2 " + H 2 0 - > N 0 3 - + 2 H + +2e" 4) 2 H + +2e" + N A D ( P ) + -> N A D H + H + The second system depends on the first system for energy. Nitrobacter must carry out the reaction in the second system to generate the required reducing power to fix C 0 2 via the Calvin cycle. The reactions in this system are catalyzed by the A T P dependent N 0 2 - : N A D + oxidoreductase. The other two enzyme systems that appear to play a role in the energy metabolism of the Nitrobacter are N A D H : 0 2 oxidoreductase and N A D H : N 0 3 " oxidoreductase. Aleem and Sewell (1981) concluded the functions of the electron transport chain are regulated by the phosphorylation state of the adenine nucleotides and the ratio of the oxidized to the reduced state of the pyridine nucleotides. When the C0 2 -based biosynthetic metabolism exerts pressure for the utilization of N A D H and A T P , the ratios of A T P / A D P and N A D H / N A D are lowered; this results in the generation of A T P and subsequently N A D H by the circulation of proton current between the oxidoreductase loops of the N 0 2 " : 0 2 oxidoreductase, ATP-dependent N 0 2 * : N A D oxidoreductase, and the ATPase systems. Therefore, Nitrobacter engage their proton-translocating respiratory chain partly in the energy generating forward electron flow and partly in the energy-driven reversed electron flow. Coupling nitrite oxidation to the generation of a proton motive force is theoretically possible but the detailed mechanism has yet to be elucidated. There has been disagreement on whether nitrite 13 oxidation takes place on the cytoplasmic or periplasmic side of the membrane. Spieck et al. (1998) used electron microscopic immunocytochemistry and monoclonal antibodies to show that the nitrite-oxidizing system in Nitrobacter is located on the inner side of the cytoplasmic membrane. Nitrite (NO2") is thought to enter the Nitrobacter cell through an antiport system via a simple exchange between nitrite (NCV) and nitrate (NO3"). In contrast, Spieck et al. (1998) showed that the nitrite-oxidizing system of Nitrospira moscoviensis is located on the periplasmic side of the membrane. These authors indicated that, from an evolutionary point of view, it is not clear i f the cytoplasmic nitrite oxidoreductase of Nitrobacter or the periplasmic nitrite-oxidizing system of Nitrospira moscoviensis represents the ancient form. . There is debate on whether or not the membrane cytochrome oxidase of Nitrobacter can pump protons. Several electron transport models have been proposed for Nitrobacter and the reader is referred to Hooper (1989) for a summary of a number of the proposed models. Although no published electron transport model is currently available for Nitrospira, it is interesting to note that Hopper and Dispirito (1985) suggest that the extracytoplasmic oxidation of substrates in chemoautotrophic bacteria, as appears to be the case in Nitrospira moscoviensis, allows the generation of a proton gradient without an energy dependent permease system for nitrite Under anaerobic conditions Nitrobacter cells, isolated membranes and purified nitrite oxidoreductase reduce nitrate to nitrite with N A D H as the physiological, or reduced methyl or benzyl viologen as the artificial, electron donor (Tanaka et al. 1983; Sundermeyer-Klinger et al. 1984; Freitag et al. 1987). Similarly, Ehrich et al. (1995) demonstrated that Nitrospira moscoviensis has a hydrogenase activity and is able to reduce nitrate to nitrite. Therefore, nitrite oxidation appears to be a reversible process in both organisms, although it is certainly not as well established in Nitrospira. 14 A membrane-bound nitrite reductase has been co-purified with the nitrite oxidoreductase of Nitrobacter. Ahlers et al (1990) found that, at reduced oxygen partial pressure, the nitrite reductase transformed nitrite to N O according to one of the following equations: 5) N 0 2 " + 2 H + +e" -> N O + H 2 0 6) N 0 2 " + H 2 0 + e~ N O +'2 OH" • ... - . Bock et al. (1991) have indicated that it seems reasonable that the two reaction centers are oriented in different positions on the cytoplasmic membrane. Ahlers (1989) has indicated the nitrite reductase activity seems to be located on the periplasmic side of the cytoplasmic membrane. Freitag and Bock (1990) demonstrated Nitrobacter use nitric oxide (NO) as a substrate for N A D H generation. The N A D H generated can then be used for A T P synthesis. The N A D H formation induced by nitric oxide was an oxygen independent reaction. These authors reported that compared to nitrite, nitric oxide was the more efficient electron donor. The metabolic diversity of Nitrobacter has been demonstrated in various studies. Freitag et al. (1987) grew Nitrobacter as biofilms on gas-permeable silicone tubing. The biofilm was investigated by ultrathin sectioning. Transmission electron microscopy revealed two morphologically distinguishable cell types in the biofilm. A high amount of carboxysomes and intracytoplasmic membranes indicated lithotrophic growth of the cells next to the oxygen permeable surface of the silicone substratum (nitrite-oxidizing cells). Numerous poly-fi-15 hydrobutyric acid granules, virtually no carboxysomes and intracytoplasmic membranes were found in cells oriented to the bulk liquid and in the anaerobic medium. These cells were shown to grow by nitrate reduction. Aleem and Sewell (1981) observed an NAD- l inked P-hydroxybutyrate dehydrogenase activity in cell-free extracts of Nitrobacter. The electron transfer from the resultant N A D H to either O 2 catalyzed by the N A D H : 0 2 oxidoreductase or to nitrate, under oxygen limiting conditions, catalyzed by the N A D H i N C V oxidoreductase yields A T P required for the maintenance energy for the survival of the chemolithotroph in environments deficient in nitrite and/or oxygen. These authors concluded that it was unlikely these enzyme systems play a significant role in the physiology and energetics of actively growing Nitrobacter cells, since the reduction of an N A D molecule by nitrite would require 4-to-6 equivalents of A T P . The N A D , thus reduced, could barely cope with the demand of the C O 2 reduction cycle, and the reduced pyridine nucleotide would never be in excess for oxidation either by O 2 or by NO3" under the physiological growth conditions. Based on their work, Bock et al.(1991) concluded Nitrobacter belongs to the group of organisms favoured by alternating aerobic/anoxic conditions. Consequently, the nitrification capacity of the biocommunity developing in a system in which aerobic and anoxic conditions vary, should be higher than the capacity of a community which develops under entirely aerobic conditions, provided nitrite does not accumulate and does not inhibit further growth of Nitrobacter. 16 1.2.3 Documented causes of Nitr i te accumulation in biological wastewater systems. The successful implementation of the nitrate shunt is contingent on the development and maintenance of a nitrite build-up in the aerobic zone or phase of the nitrification/denitrification process. Both short and long term nitrite accumulation have been reported in biological wastewater treatment systems. Short-term nitrite accumulation is typically observed during plant start-up when nitrite accumulates as a result of the lag associated with the development of a sufficient nitrite-oxidizer population relative to the ammonia-oxidizer population. Long-term nitrite accumulation is believed to be associated with nitrite-oxidizer inhibition. The most frequently cited causes of nitrite accumulation in wastewater systems include: reduced temperatures, limiting oxygen or carbon dioxide concentration, elevated p H , free ammonia, elevated solids wastage, acute process loading, and cryptic nitrate reduction (Alleman 1985). The conditions most likely to lead to success in achieving the nitrate shunt were summarized by Hyungseok Yoo et al. (1999), who conclude that: one must use a simultaneous and/or alternating nitrification/denitrification process in the same reactor, maintain a low aerobic phase dissolved oxygen concentration, keep the microorganisms in direct contact with the influent wastewater under oxygen-deficient conditions (to induce contact with high concentrations of free ammonia and/or free hydroxylamine), raise the p H , add hydroxylamine to the reactor, and maintain the reactor temperature near 25° C . The reported cause(s) of nitrite accumulation are discussed in the text below. 17 1.2.3.1 Dissolved Oxygen as a Factor in Nitr i te Accumulat ion Numerous investigators (Alleman, 1985; Payne, 1973; Boon and Laudelot, 1962; Jones and Paskiris, 1982; Hanaki et al. 1990b; Balmelle, 1992; Cecen and Gonenc, 1994; Garrido et al. 1997a and 1997b; Hyungseok Yoo et al. 1999) have reported nitrite accumulation in nitrification/denitrification systems as a result o f low dissolved oxygen concentrations. Although there appears to be general consensus on the effect of a low dissolved oxygen concentration on nitrite accumulation, there are conflicting reports on the contributing factors and mechanism responsible for observed phenomenon. Alleman (1985) suggested that nitrite accumulates partly because of the higher K s (DO) of Nitrobacter, compared with Nitrosomonas, and partly because nitrate formation is the second step. Jones and Paskins (1982) suggested nitrite accumulation was the result o f a change in the mode of energy synthesis by nitrite oxidizing organisms at low dissolved oxygen concentration. Laanbroek and Gerards (1993), who conducted chemostat experiments with mixed cultures of Nitrosomonas europaea and Nitrobacter winogradskyi, used specific antibodies and microscopy to measure the number of organisms. They reported the steady state numbers of Nitrosomonas europaea were not significantly affected by growth rate or oxygen tension, whereas the numbers of Nitrobacter winogradskyi were. They reported the ratio between numbers of Nitrosomonas europaea and Nitrobacter winogradskyi increased with decreasing oxygen concentration and increasing growth rate (i.e. lower SRT). 18 Hanaki et al. (1990b) concluded that nitrite accumulation in a mixed nitrifier population was due to an increase in ammonia oxidizer yield under low dissolved oxygen concentration that compensated for a reduction in substrate utilization rate. Goreau et al. (1980) have also reported an increase in ammonia-oxidizer yield at low dissolved oxygen concentration when working with pure cultures. Hanaki et al. (1990b) reported nitrite oxidation was strongly inhibited at low dissolved oxygen concentration and that the increase in ammonia-oxidizer yield coupled with the difference in saturation constants between'ammonia and nitrite oxidizers contributed to the observed nitrite accumulation. Laanbroek and Gerards (1993) reported the growth yield of Nitrosomonas europaea was not significantly affected by either growth rate or oxygen tension. However, the growth yield of the nitrite-oxidizing organisms was highly dependent upon both factors. Laanbroek and Gerards (1993) reported that changes in these factors were accompanied by large differences in nitrite consumption rates per cell (Nitrobacter winogradskyi cells consumed only 0.13 fmol'nitrite cell" 1 h"1 at the highest growth rate in combination with the lowest oxygen concentration and 58 fmol nitrite cell" 1 h"1 at the lowest growth rate and highest oxygen concentration). Zheng et al. (1994) conducted studies on N 2 0 production in nitrification processes, using bench scale chemostats and a synthetic substrate with high ammonia concentration and no carbon. The primary objective of their work was to estimate the amount of N 2 0 production in nitrification processes at various dissolved oxygen and S R T values. Although the highest concentration of nitrous oxide ( N 2 0 - N ) was not found with the highest nitrite ( N 0 2 - N ) , a higher amount of nitrous oxide was generally found when nitrite accumulated. Nitrite accumulation took place at low dissolved oxygen concentrations. Zheng et al. (1994) indicated that the mechanism of nitrous oxide production could not be clarified. 19 Cecen and Gonenc (1994) and Cecen and Ipek (1998) reported that nitrite accumulation is not dependent upon dissolved oxygen concentration alone, but on the ratio of dissolved oxygen concentration to free ammonia concentration. While working with bench scale nitrifying activated sludge reactors, Cecen and Gonenc (1994) found that nitrite accumulates at dissolved oxygen to free ammonia concentration ratios lower than five. No nitrite occurrence was encountered when the dissolved oxygen to ammonia ratio exceeded five. Cecen and Ipek (1998) used a fed-batch reactor to assess the degree of nitritification and nitratification in a short time period without allowing adaptation of bacteria to high substrate concentrations. They reported that the nitrite nitrogen concentrations were above 10 mg N0 2 -N/1, when the dissolved oxygen to free ammonia nitrogen ratio was less than one. At very low dissolved oxygen to ammonia nitrogen ratios (0.2 mg 0 2/mg NFL.-N), N 0 2 - N concentrations rose above 30 mg/1. They did not observe any nitrite accumulation when urea was used as the nitrogen feed. Cecen (1996) concluded that the possibility of achieving low effluent ammonium concentrations and maintaining a high nitrite build-up appeared low. Hyungseok Yoo and co-workers (1999) concluded that nitrite accumulation depends upon the rate of increase or decrease in dissolved oxygen concentration in addition to the maximum concentration. They reported that the maximum dissolved oxygen concentration during the aeration period of an anoxic/aerobic process should be kept to a minimum and the rate of increase in dissolved oxygen between the anoxic and aerobic phase should be neither too slow nor too fast. The authors concluded that an aerobic phase median dissolved oxygen concentration of 1.0 -1.5 mg/1 appeared to be optimal for their studied process. 20 1.2.3.2 Temperature as a Factor in Nitrite Accumulation. There are many inconsistencies in the literature regarding the impact of temperature on nitrite accumulation in biological wastewater treatment systems. The reason for this variation appears to be related to the difficulty in separating the effect of temperature from other factors impacting nitrite accumulation (Quinlan, 1980 and Fdz-Polanco et al.1996). In general, nitrite accumulation occurs at low wastewater temperatures in the absence of either free ammonia inhibition or other substrate limitations (Quinlan, 1980; Randall and Buth, 1984b; Alleman, 1985). The reverse appears to be true under high levels of free ammonia or substrate limiting conditions. However, a number of exceptions can be found in the literature. Some of the more pertinent literature on temperature effects on nitrite accumulation under inhibiting and/or substrate limiting conditions is summarized here. Randall and Buth (1984b) studied the factors that suppress nitrification in the activated sludge process with a series of continuous-flow, laboratory-scale reactors treating a synthetic wastewater. These authors concluded that for a specific biological system, there is a critical temperature below which the rate of nitrate formation is less than the rate of nitrite formation. This results in a build-up of nitrites in the reactor until the temperature drops low enough to almost completely suppress nitrification. Boon and Laudelout (1962) demonstrated that the oxidation of nitrite by a pure culture of Nitrobacter proceeded much more slowly with increasing temperature when the oxygen concentration had fallen below the K o value for Nitrobacter. 21 Balmelle et al. (1992) conducted batch experiments under conditions that are typically considered inhibitory for nitrite oxidizing organisms (i.e. free ammonia concentration in the range of 2 - 5 mg/1). They reported that nitrite-oxidizing organisms are active over a temperature range o f 10 to 20° C . Under these conditions, nitrite build-up remained low, which Balmelle et al. suggested was due to the fact that the effect of Nitrobacter activation by temperature prevails over its inhibition by free ammonia. On the other hand, in the temperature range of 20-25 C> a slowing of the nitratating activity was observed, together with an activation of the nitritating activity resulting in nitrite buildup. Balmelle and co-workers suggested that the presence of free ammonia inhibition was the reason these results varied from those o f other investigators who have reported nitrite accumulation at low temperatures. Balmelle et al. (1992) found that the effect of free ammonia inhibition on Nitrosomonas was not observed until a temperature of 25 C. It should be noted that free ammonia concentration is a function of temperature, as well as pH. Fdz-Polanco et al. (1994) state that temperature is a key parameter in the nitrification process producing two opposite effects: bacteria activation and free ammonia inhibition. They used an up-flow biological aerated filter ( U B A F ) in order to study both effects. These investigators indicated the plug flow nature of this reactor allows one to study both effects simultaneously, since nitrogen and biomass concentration profiles are established within the reactor. The experimental results from the bottom part of the reactor (where the inhibition by free ammonia was believed to be dominant) were used to study free ammonia inhibition and results from the top of the reactor (where temperature activation was believed to dominate) were used to explain temperature activation. Fdz-Polanco et al. (1994) reported the activity of nitrite oxidizing organisms increased with temperature at the top of the reactor. They also reported the activity of 22 nitrite oxidizing organisms decreased substantially for unit loading values above 1 mg N H 3 - N / mg V A S where V A S is taken as volatile attached solids. The nitrite oxidizing activity increased exponentially with increasing temperature for all loadings below this threshold. Fdz-Polanco et al. (1996), also studied the combined effect of temperature, p H , and ammonium ion concentration on nitrite accumulation. These investigators varied each of the three primary parameters while maintaining a constant specific free ammonia concentration (free ammonia/biomass concentration ratio measured as mg NH3-N/g V A S ) . They reported an increase in the percent nitrite accumulation (measured as the percentage ratio N02 -N/NO x -N) from 52 to 95%, when the temperature increases from 10 to 25° C and ammonium ion concentration decreases from 80 to 25 mg/1; this implies a preponderance of the temperature effect over the ammonium concentration effect. They also reported an increase in nitrite accumulation from 25 to 52% when increasing the pH from 7 to 7.8, even when decreasing the temperature from 25 to 10 C; this indicates the p H effect dominates over the temperature effect. Fdz-Polanco et al. (1996) concluded that, in the absence of free ammonia inhibition, nitrite accumulation is more likely at low temperature; however, due to the lower activity of ammonia and nitrite oxidizers at low temperatures, the nitrite accumulation w i l l be smaller than that obtained at higher temperatures under the effect of free ammonia inhibition. Hellinga et al. (1998) reported on measurements of the minimum sludge retention times required for both ammonia and nitrite oxidizers conducted by Hunik (1993). They reported that at the 0 normal temperatures in wastewater treatment plants (5-20 C) , nitrite oxidizers grow faster than ammonia oxidizers, which means that ammonia is completely oxidized to nitrate. However, the reverse is true at elevated temperatures. This relationship appears to have been developed for 23 relatively high ammonia concentrations (130 mg N H 4 -N/1), making it consistent with the observations of Balmelle et al. (1992). 1.2.3.3 The role of pH in nitrite accumulation. The pH has a significant affect upon the concentrations of both free ammonia and nitrous acid. These molecules are widely reported in the literature as being inhibitory to both ammonia and nitrite oxidizing organisms. High pH environments result in higher concentrations of free ammonia. The formation of nitrous acid is favored as the pH decreases. Many authors (Sauter and Alleman, 1981, Alleman, 1980; Sutherson and Ganczarcsyk, 1986; Balmelle et al., 1992; Surmacz-Gorska et al., 1997; Hellinga et al., 1998) have reported on the effect of pH on nitrification under high ammonia concentration. Most report no significant nitrite accumulation at neutral pH, nitrite accumulation in the pH range of approximately 8 to 9 and no significant nitrite or nitrate accumulation beyond pH 10. This is widely believed to be the result of the dependence of the free ammonia concentration upon pH (free ammonia concentration increases with increasing pH). There have been few attempts to separate the effect of pH and free ammonia on nitrite accumulation. Sutherson and Ganczarczyk (1986) carried out experiments with various constant free ammonia concentrations and varying pH conditions at room temperature. To maintain a constant free ammonia concentration in each reactor, ammonium chloride solution was fed to each at a rate equal the ammonia oxidation rate. They found that in the pH range of 8 to 8.8, pH played an additive role in the inhibition of Nitrobacter with higher pH values favoring nitrite accumulation. 24 The inhibitory effect of p H was more pronounced at lower free ammonia concentrations (0.5 mg/1) relative to higher free ammonia concentrations (1.5 mg/1). Balmelle et al. (1992) showed that under the operating conditions for their batch tests 0 (temperature = 25 C, p H = 8.1, dissolved oxygen = 2.5 mg/1, and initial biomass = 90 +/- 10 mgMSS/1) the nitrite concentration was practically independent of the p H . They concluded that the inhibitory effect on nitrite oxidation was due to the initial free ammonia concentration, which was in the range of 2.5 to 25 m g - M i V N / l . These concentrations lead to 100% inhibition. However, a p H dependent variation in nitrite ion production was observed. The rate of nitrite production increased with p H to a maximum at p H 8.5 and then decreased thereafter. This optimum corresponds to the results presented by W i l d (1971) and Jones et al. (1982). Fdz-Polanco et al. (1996) reported that, in the absence of free ammonia inhibition, nitrite accumulation is more likely at low pH when the p H is controlled within a narrow interval from 6 to 7.5. They reported that, as the p H decreases, both ammonia and nitrite oxidizers activities decrease, but the quotient o f the activities decreases. This indicates that, in the absence o f free ammonia inhibition, the nitrite accumulation may happen at low p H values. Hellinga et al (1998) argue p H is the most important process parameter for the application of the S H A R O N process, since it affects the availability of the inorganic substrate for both ammonia-and nitrite-oxidizing organisms. At high pH more added ammonium is available as NH3, the true substrate for ammonia oxidizers, while less HNO2 (which they refer to as the true substrate for nitrite oxidizers) is available. A lower HNO2 concentration also limits inhibition of ammonia oxidizers. Hellinga et al. (1998) state that the margins in reactor residence time to maintain ammonium oxidizers and washout nitrite oxidizers are, therefore, much larger at high p H . 25 1.2.3.4 Free Ammonia and Nitrous Acid Various investigators have reported high concentrations of free ammonia and free nitrous acid as inhibitory to ammonia and/or nitrite oxidizing organisms. To date, there are conflicting reports on the actual concentration of either free ammonia or nitrous acid needed for inhibition of ammonia and/or nitrite oxidizing organisms. In the case of free ammonia inhibition, evidence suggests that acclimation of nitrite oxidizers to free ammonia prevents sustained nitrite build-up (Section 1.2.4). Focht and Chang (1975) suggest Meyerhof was the first to report substrate inhibition of ammonia and nitrite-oxidizing bacteria in 1916. Meyerhof showed that optimal oxygen consumption rates for nitritifying and nitratifying declined at respective ammonium and nitrite concentrations beyond 60 and 350 mg-N/liter, respectively. Anthonisen et al. (1976) conducted one of the most extensive works on nitrification inhibition. They investigated the effect of free ammonia and free nitrous acid inhibition on ammonia and nitrite oxidizing bacteria. Their experiments were conducted with a number of wastes including synthetic wastes, agricultural wastes and soil. The data reviewed encompassed a broad range o f microbial environments they believed would help to identify the common basis of nitrification inhibition in biological systems. r The work of Anthonisen and co-workers resulted in the development of an operational chart to illustrate the inhibitory concentrations of free ammonia and free nitrous acid. The inhibitory 26 range of free ammonia was reported as 10 to 150 mg/1 for Nitrosomonads and 0.1 to 1 mg/1 for Nitrobacter. The inhibitory range of free nitrous acid was reported as 0.22 and 2.8 mg/1. Anthonisen et al. (1976) stressed that because of factors such as acclimation, number of active organisms, and the effect o f temperature on reaction rates, these ranges were likely to be situation specific. Hunki et al (1993) could not find significant NH3 inhibition of chemostat cultures of Nitrobacter agilis when the activity was compared with that for a salt solution blank. Only at a p H of 6.5 could the inhibition of ammonium be clearly distinguished from the inhibition of a N a C l solution. On this basis, Hunik et al. (1993) concluded inhibition of Nitrobacter agilis by NH3 was unlikely. They did not observe severe inhibition of Nitrobacter agilis by N H V and found this bacterium is less sensitive to osmotic pressure at high salt concentrations than Nitrosomonas europaea. Chen et al. (1991) conducted laboratory studies using a nitrifying fluidized bed reactor to study the feasibility of nitrification/denitrification via nitrite. They used influent NH 4 -N/1 concentrations o f 50, 100, 200, 400 and 1000 mg/1 in their work. The authors did not report the reactor p H for each test run. Chen and co-workers reported that the percentage of N 0 2 as a proportion of the total oxidized nitrogen species increased with increasing ammonium concentration, with nitrite representing 65% and 95% of the total oxidized nitrogen species at the lowest and highest ammonium loadings, respectively. Abeling and Seyfried (1992) used an anoxic/aerobic process to study nitrification/denitrification via nitrite. The p H of the aerobic reactor was controlled via N a O H addition (pH 8.2- 8.4) and the ammonium content was measured continuously and adjusted to a predetermined value. The 27 nitrification tank was filled with plastic media with a specific area of 150 m 2 /m 3 . A secondary treatment reactor was installed for residual nitrification and nitratation. The p H in this secondary reactor was regulated between 7-7.5. These investigators stated that, in order to attain the highest nitritation rate, it was crucial to keep the free ammonia concentration high enough to inhibit Nitrobacter but low enough to prevent the inhibition of Nitrosomonas. A t a p H of 8.5 0 and temperature of 20 C, the optimal free ammonia concentration for maximum nitritation was found to be approximately 5.0 mg/1. The inhibition of nitritation was observed at approximately 7.0 mg/1. Neufeld et al (1980) confirmed the beginning of nitritation inhibition at 10 mg/1. Mauret et al (1996) concluded that high free ammonia concentrations inhibit Nitrobacter, in the range of 6.6 to 8.9 mg N H 3 - N / 1 . Fdz-Polanco et al. (1996) used bench scale Up-flow Biological Aerated Filter ( U B A F ) reactors inoculated with seed from a pilot plant treating domestic wastewater, to treat a synthetic wastewater consisting of ammonium sulphate, sodium bicarbonate, and trace nutrients. They reported different combinations of temperature, p H and ammonium concentration bring about different nitrite accumulations for the same specific free ammonia concentration. Fdz-Polanco et al. (1996) hypothesized this is due to the fact that each parameter (temperature, pH, and ammonium concentration) can exert an effect of activation-deactivation over the metabolism of ammonia and nitrite oxidizing organisms. They concluded that this fact could explain the different free ammonia inhibition thresholds reported in the literature and recommended that any inhibition situation should be defined by the values of free ammonia, ammonium and biomass concentrations, p H and temperature. 28 1.2.3.5 Free Hydroxylamine Several investigators have implicated hydroxylamine as an inhibitor of ammonia and nitrite oxidation. A s indicated previously (Section 1.2.2), free hydroxylamine is an intermediary in the ammonia oxidation process. The inhibition of Nitrobacter by hydroxylamine has been reported by Castell and Mapplebeck, (1956), Castignetti and Gunner (1980, 1982), Stuven et al. (1992), and Yang and Alleman (1992). Castignetti and Gunner (1980, 1982) studied the production of nitrite from pyruvic acid and hydroxylamine by an Alcaligenes species, with subsequent oxidation by Nitrobacter. In monoculture, Nitrobacter was very sensitive to levels of hydroxylamine, which were tolerated by Alcaligenes strains. Castignetti and Gunner (1981) studied the differential tolerance of an Alcaligenes sp., a heterotrophic nitrifier, and Nitrobacter agilis to increasing concentrations of hydroxylamine. They reported Nitrobacter agilis cultured separately in the presence of hydroxylamine, at concentrations as low as 5 mg/1 of NH2OH-N, were subsequently unable to convert nitrite to nitrate. Interestingly, a transient exposure of Nitrobacter agilis resulted in the loss of metabolic activity of the organisms as determined by the lack of nitrite oxidation by Nitrobacter agilis transferred to a fresh medium. The Nitrobacter agilis group subjected to 5 mg/1 of NH2OH-N for one day produced only 22 mg/1 of nitrate-nitrogen after 22 days when transferred to a medium having an initial nitrite concentration of 303 mg/1. The control group, which was not subjected to hydroxylamine, produced 322 mg/1 of nitrate-nitrogen over the same time period. 29 Joint culturing of Alcaligenes sp and Nitrobacter agilis in a medium containing 18 mg/1 NH2OH-N resulted in the eventual synthesis of nitrite and nitrate. Pure cultures of Alcaligenes sp. were able to tolerate hydroxylamine concentrations as high as 235 mg/1 NH2OH-N. Other heterotrophic nitrifiers have also been shown to be resistant to hydroxylamine toxicity. Stuven et al. (1992) confirmed that hydroxylamine inhibits Nitrobacter cells. Stuven et al. demonstrated hydroxylamine production by mixotrophically growing cells of Nitrosomonas and Nitrosovibrio in pure culture as well as in mixed culture with Nitrobacter. Pure cultures were grown at relatively high oxygen and ammonium concentrations (dissolved oxygen 7-8 mg/1 and ammonium concentration approximately 500 mg/1 as N) . H u (1990) investigated acute substrate-intermediate-product related inhibition of nitrification and Nitrobacter. using respirometry. Hu's work was carried out using enriched nitrifying cultures (Nitrosomonas and Nitrobacter) as well as enriched Nitrobacter cultures. He reported that both hydroxylamine and nitrite had approximately the same degree o f acute inhibition to nitrification but Nitrobacter was more susceptible to hydroxylamine than ammonia. Based upon oxygen uptake rate (OUR) , ammonia produced 57% inhibition of Nitrobacter at 91.6 mg/1 N H 3 - N when compared to a control while hydroxylamine created 82% inhibition at 90.7 mg/1 NH2OH-N. Only nitrite was found to inhibit hydroxylamine oxidation. Yang and Alleman (1992) investigated the cause of nitrite build-up in batch systems enriched with a nitrifying culture. The stock culture was supplied with a solution containing ammonium chloride (NH4CI) and sodium bicarbonate (NaHCOs), in a continuous flow system. A 500 ml sample was taken from the culturing system, washed twice to remove the residual inorganic nitrogen species, and then dosed with the required concentration of NH4CI (either 50, 200, or 500 30 mg/1) to achieve the desired initial ammonia concentration in each experiment. The samples were mixed and aerated to one of two dissolved oxygen levels (referred to here as high or low). The target p H (7.0, 7.5, 8.0, or 8.5) value was maintained during each batch experiment via a p H control system. Three different parameters (total ammonia concentration, D O level, and pH) composed of ten different conditions were used for the batch experiments. Samples were taken on an hourly basis during each batch test and analyzed for total ammonia, total hydroxylamine, nitrite, and nitrate. The end of each experiment was marked by the achievement of a constant nitrate concentration. For experiments conducted with an initial ammonia concentration of 200 mg N/1 controlled at p H of 7.5 (free ammonia level of 3 mg N/1) nitrite buildup was observed for the low D O system (DO = 0.5 mg/1) but not the high D O system (DO = 6 mg/1). A n increase in the initially applied ammonia concentration to 500 mg/1 (free ammonia concentration of 7.2 mg/L) for the high D O system did not result in a nitrite buildup; however, nitrite accumulated in the low D O system even when the initially applied ammonia concentration was reduced to 50 mg/1 (free ammonia concentration of 0.92 mg/1). Yang and Alleman (1992) reported that nitrite buildup generally occurs with a low dissolved oxygen concentration (0.5 mg/1), whereas no nitrite buildup was observed with a batch system operated at high dissolved oxygen (6 mg/1). Nitrite peaks were however found in both the high and low dissolved oxygen systems at p H values of both.8.0 and 8.5. However, no nitrite accumulation was observed in either the low or high dissolved oxygen system at p H 7.0. These investigators concluded, therefore, that dissolved oxygen concentration was not the dominant factor behind a nitrite build-up, and that the correlation with the free ammonia concentration was erratic. Based on Hooper's mechanism of nitrification, hydroxylamine was believed to accumulate in the systems at low dissolved oxygen concentration. Therefore, Yang and Alleman (1992) concluded that free hydroxylamine most likely plays a role in nitrite build-up in nitrifying systems. The 31 results of their research showed that variable levels of hydroxylamine accumulated in the batch systems under all conditions. The presence of unionized or free hydroxylamine appeared to have a consistent correlation with incomplete nitrification and the authors concluded that free hydroxylamine was likely the major cause, of nitrite build-up,in their batch nitrification systems. These authors hypothesized that molecular free hydroxylamine is able to penetrate the membrane of Nitrobacter to exert an inhibitory effect on the activity of the enzyme nitrite oxidoreductase. Hyungseok Yoo et al (1999) surmised that the process they used to remove nitrogen via nitrite -an intermittently aerated and decanted single-reactor - was successfully demonstrated due to the establishment of an oxygen deficit in both the settling and decant phases. They reported that when Nitrobacter is in direct contact with the free ammonia in the influent wastewater, there is a high probability the ammonia is converted to free hydroxylamine by Nitrosomonas. Influent wastewater was fed to the laboratory scale batch reactors from the bottom on a continuous basis to create the desired conditions. 1.2.3.6 Anaerobiosis A period of anaerobiosis, coupled with free ammonia inhibition, upstream of the aerobic zone has been reported in the literature as a requirement for nitrite accumulation, consistent with nitrification /denitrification via nitrite. Turk and Mavinic (1983) conducted studies using a laboratory scale reactor treating synthetic wastewater. One of two reactors was operated under conditions that were believed to be inhibitory to Nitrobacter namely: low dissolved oxygen (less than 1 mg/1) and an inhibitory free 32 ammonia concentration in at least two of the four reactors. The control reactor was operated in a pre-denitrification mode with an internal recycle rate of 100%. Nitrite accumulation was observed in the control but not in the experimental condition. A gradual change in operation of the experimental reactor, from fully aerobic to a pre-denitrification mode, resulted in nitrite accumulation. These authors attributed the observed nitrite accumulation in the predenitrification mode to an increase in free ammonia concentration in the first cell of the four-cell reactor system following the switch to the predentrification mode. Specifically, the nitrite accumulation was believed to be the result of the intermittent contact with the high free ammonia level in the anoxic zone. Van Benthum et al. (1998) conducted bench scale nitrogen removal studies using a nitrifying airlift suspension reactor, referred to as the B A S reactor, coupled with a denitrifying chemostat. The authors suggested this process configuration is attractive for concentrated ammonia containing wastewaters. Basalt particles served as carriers for the biofilms in the B A S reactor. The dissolved oxygen concentration in the airlift suspension reactor was maintained at 3 mg/1 and nitrogen gas was sparged through the chemostat to keep it anoxic. Initially, the process was run in a post-denitrification mode, during which nitrification and denitrification was via nitrate. A change to a pre-denitrification operating mode, resulting in recirculation of the denitrifiers between both reactors, resulted in nitrification/denitrification primarily via the nitrite route. Nitrite was observed in the B A S reactor from the beginning, while the nitrate concentration in the effluent decreased considerably. The maximum specific activity of ammonia did not decrease compared to the post denitrification period; however, the maximum specific activity o f the nitrifiers decreased considerably compared with the post denitrification period. Van Benthum and co-workers (1998) hypothesized that nitrite accumulation in the pre-denitrification mode was the result of competition between nitrite oxidizing and reducing organisms, with nitrite reducing 33 organisms having a higher reaction rate. The authors further hypothesized that nitrite oxidizers were washed out of the system at the high recirculation flow (recirculation ratio 6.5:1). The pre-denitrification experiments were run for a total o f 65 days, 16 of which were at a high ammonia load (8 kg/m 3-d~l) with the remainder at a loading referred to by the researchers as 'normal' (5 kg/m 3-d~l). These investigators pointed out that previous experiments with a B A S reactor and the same dissolved oxygen concentration of 3 mg/1 did not result in nitrite accumulations, suggesting oxygen limitation was not the cause of the observed accumulation in the predenitrification mode. 1.2.3.7 Solids Retention Time Several investigators (Beccari et al., 1979; Randall et al., 1984a; Hellinga et al., 1998) have shown that stable partial nitrification can be achieved by carefully controlling sludge age. Hellinga et al. (1998) have demonstrated that in practice this procedure can be used at o temperatures above 15-20 C. Beccari et al. (1979) determined that effluent nitrogen species could be switched between NH4, NO2, and NO3 by controlling the mean cell residence time ( M C R T ) of continuous flow reactors. They found that retention of nitrifiers in mixed liquor, depends upon their respective growth characteristics. They reported that Nitrosomonas and Nitrobacter are not retained below a mean cell residence time of two days and until about four days, only Nitrosomonas are present. Control o f solids wastage within this intermediate M C R T range could consequently result in a predominant nitrite product. 34 Hanaki et al. (1990a, 1990b), used bench scale chemostats treating synthetic wastewaters at high (6 mg/1) and low (0.5 mg/1) dissolved oxygen concentrations. Each of the systems was run at o various retention times (1.5 to 8 days) in a constant temperature room (25 C). These investigators reported the percentage of nitrite oxidation was always lower in the low dissolved oxygen (0.5 mg/1) than for the high dissolved oxygen (6 mg/1) reactors. A retention time of 3.8 days was found to be sufficient to ensure complete nitrite oxidation (i.e. no nitrite accumulation) at high dissolved oxygen (6 mg/1), while even a retention time of 6.5 days was insufficient to achieve complete nitrite oxidation at low dissolved oxygen (0.5 mg/1) concentration. The nitrite level reached a high of approximately 60 mg/1 (equivalent to 75% of the influent NH3 -N) at a retention time of 2 to 3.8 days in the low dissolved oxygen concentration system; however, low nitrite accumulation resulted at the shorter retention times due to poor nitrite production by the ammonia oxidizers. Hanaki et al. (1990a, 1990b) indicated that there is a relationship between solids retention time and nitrite accumulation and this relationship appears to depend upon other factors such as dissolved oxygen concentration. A z i m i and Horan (1991) studied the impact of reactor mixing characteristics on the rate of nitrification in the activated sludge process. These investigators operated two continuous flow activated sludge systems with mixing regimes approximating completely mixed and plug-flow. A z i m i and Horan (1991) reported that plug-flow reactors operated at solid retention times that are too low for complete nitrification, are able to partially nitrify and produce an effluent high in nitrite. They found the critrical S R T was 4 days. Zheng et al. (1994) found that, when the dissolved oxygen concentration was maintained above 5 •mg/1 and S R T was varied between 3 and 20 days, N2O production increased with decreasing 35 SRT. In addition, significant N 0 2 production occurred at an S R T of 3 days. Influent ammonia was completely oxidized to NO3 at all other SRT's . 1.2.3.8 Acute Process Loading Prakasam and Loehr (1972) both documented the effect of process loading on nitrification behavior. They reported acute process loading was likely to result in nitrite accumulation. Turk and Mavinic (1989) found that the impact of shock ammonia loadings on nitrite accumulation in a pre-denitrification system diminished with time, as the biomass appeared to acclimate to the high ammonia load. Horan and A z i m i (1992) investigated the impact of nitrogen and hydraulic loading transients on the nitrification performance of bench scale plug flow and completely mixed activated sludge reactors. Both reactors were pulsed with concentrated stock solutions of ammonium chloride and urea for 3 hours each day. The authors found that influent ammonia loads over the range of 50-200 mg-N/1 produced only a slight increase in effluent ammonia concentration for both the plug flow and completely mixed reactors. However, there was a large increase in effluent ammonia and nitrite concentrations for influent ammonia concentrations greater than 200 mg-N/1. Ghyoot et al. (1999) studied start-up conditions for a membrane bioreactor treating sludge reject water. They reported that during start-up, sludge loading rates could be rapidly increased due to the complete retention of nitrifying bacteria in the reactor. However, ammonia-oxidizing 36 bacteria are better able to cope with increasing nitrogen loadings than nitrite oxidizers, resulting in a rapid increase in nitritation capacity that is not followed by an increase in nitratation capacity. Ghyoot et al. (1999) found that, i f complete nitrification is desired, the nitrifying capacity should be carefully monitored to determine the reactor loading rate and an increase of the NH3 concentration above 0.1 mg N/1 should be avoided. 1.2.4 The acclimation of nitrite oxidizers to free ammonia. Turk and Mavinic (1987) demonstrated the feasibility of removing nitrogen from highly nitrogenous wastewater by blocking the nitrification process at the intermediary nitrite level with free ammonia then subsequently reducing the nitrite to nitrogen gas. The authors found that nitrite accumulation could be induced and sustained for a finite period, using a plug-flow, pre-denitrifying activated sludge process configuration. However, nitrite accumulation could not be sustained indefinitely. Turk and Mavinic (1987) concluded that acclimation of nitrite oxidizers to free ammonia may have played a major role in limiting the successful implementation of this process. However, the impact of the apparent acclimation effect did not manifest itself for several SRT ' s and the authors concluded that it could only be properly assessed by long-term experiments. Turk and Mavinic (1989) investigated several methods for overcoming the effects of acclimation including: reduction df the sludge age to under 4 days, extension of the contact time with high free ammonia levels to more than 50% of the total retention time, raising free ammonia concentrations to 40 mg NH3 -N/I , use of a more complex substrate, double substrate inhibition, internal recirculation, temporary reduction of free ammonia levels, and temporary stoppage of feed. O f these, the most effective was internal recirculation. These investigators reported that 37 nitrite build-up could be maintained for an extended period of time (125 days) in the presence of a nitrifier biomass acclimated to high levels of free ammonia, by intermediary denitrification. Several other researchers have suggested that Nitrobacter in mixed culture systems could acclimate to increasing concentrations of free ammonia. Wong-Chong and Loehr (1978) reported that Nitrobacter acclimated to free ammonia could tolerate concentrations as high as 40 mg/1, while unacclimated ones were inhibited at concentrations of 3.5 mg/1. Ford et al. (1980) reported total inhibition of nitrification activity at free ammonia levels of 24 mg/1, but noted that system recovery was possible, even at levels as high as 56 mg/1. Sutherson and Ganczarczyk (1986) introduced the concept of 'recovery time' into their experiments. They hypothesized Nitrobacter inhibition can be readily reversed when the organisms are returned to favorable conditions following free ammonia inhibition. These investigators theorized that by retaining the biomass in the aeration tank for a time that is less than the recovery time of the inhibited nitrite oxidizing organisms, the suppression of nitrite oxidation could be achieved. The percentage of inhibition was calculated as the ratio between the accumulated nitrite and the combined nitrate and nitrite species. The 'recovery time' was arbitrarily defined as the time taken for the oxidation of five percent of accumulated nitrite following relief o f the inhibitory condition. Higher free ammonia concentrations significantly increased the recovery time at a constant p H value and at a particular free ammonia concentration, the percent increase in recovery time, with increasing p H , was not as significant as the increase in percent inhibition under the same conditions. Sutherson and Ganczarczyk found that, by gradually increasing the free ammonia concentration in the laboratory scale reactors a free ammonia concentration of 2.5 mg/1 could be tolerated. 38 Villaverde et al. (2000) studied nitrifying biofilm acclimation to free ammonia in submerged biofilters. The primary objective of their research was to determine whether the threshold specific free ammonia concentration, defined as the applied free ammonia concentration per unit biomass weight, to which nitrite-oxidizing organisms were inhibited changes with time. The impact of biofilm startup conditions on the variation in threshold concentration was also monitored. The biomass concentration was measured as volatile attached solids and the specific activity of the nitrifying organisms, pmax/Y, was measured as substrate removal rate per biomass unit and time (mg N g V A S " 1 h"1). Two reactors were used in this study. During startup, one reactor was fed continuously with synthetic waste containing 100 mg NHV-N/1 in an attempt to set up a concentration profile along the reactor. The second reactor was fed a concentrated solution distributed homogeneously throughout the column once per day, in order to setup a homogeneous ammonia distribution along the column. The continuous reactor was operated in a once through mode during startup whereas the batch reactor was operated with liquid recycle. Both reactors were operated in continuous flow mode (operating parameters were D O = 2 mg/1, temperature = 25 C°, and p H = 7.5), without recycle, following the 20 day start up period. After two months of operation the percentage ammonium oxidation for both reactors was approximately 75%, however the concentrations of nitrite and nitrate varied. For the reactor started in batch mode the nitrite accumulation, measured as N 0 2 - N / N O x - N , was 65%, compared to a value of only 30% for the reactor started in continuous mode operation. Over a four month period, there was a progressive increase in nitrate concentration and increase in ammonia oxidation percentage for both reactors although the nitrifying population, measured as volatile attached solids, remained essentially the same. Furthermore, the threshold value for specific free ammonia inhibition over the activity of nitrite oxidizers rose to 0.5 and 0.7 mg N g V A S " 1 h"1 for the reactor started in batch and continuous flow mode respectively. This compares to an initial 39 threshold inhibition value of 0.2 mg N g V A S " 1 h"1. Villaverde et al. (2000) suggested that the observed increase in nitrate concentration and threshold free ammonia concentration was the result of acclimation of nitrite oxidizing organisms. They also observed, via activity tests, the specific ammonium oxidation rate increased during the duration of the test suggesting ammonia oxidizers also acclimated to higher free ammonia concentrations. These investigators concluded the nitrification-denitrification process via nitrite is not likely to be stable or feasible for long periods of time, due to the acclimation of nitrite oxidizers to free ammonia. 40 1.3 Objectives and Hypotheses to be Tested Long-term nitrite accumulation, a precursor to the successful application of the nitrate shunt, is believed to be associated with the inhibition of nitrite oxidation. A number of factors have been suggested to cause and/or control inhibition of nitrite oxidation and the accumulation of nitrite in a biological wastewater treatment system. The most commonly suggested cause of nitrite accumulation is considered to be a high free ammonia concentration. The evidence in the literature that casts doubt on this explanation is summarized as follows: • Yang and Alleman (1992) report nitrite accumulation in a batch reactor with free ammonia concentration of 0.92 mg/1 and 0.5 mg/1 dissolved oxygen concentration, but not in a reactor with free ammonia concentration of 3 mg/1 and dissolved oxygen concentration of 6 mg/1. The correlation of nitrite build up with free ammonia concentration appears erratic and free hydroxylamine, an intermediary in the ammonia oxidation process, is implicated as the likely inhibitor. • Hunki et al (1993) conducted chemostat experiments using pure cultures of Nitrobacter agilis. The inhibition of ammonia oxidation could only be clearly distinguished at pH 6.5 from the inhibition of a NaCl solution. On this basis, Hunik et al. concluded inhibition of Nitrobacter agilis by NH3 was unlikely. • Zheng et al (1993) report nitrite accumulation in chemostat reactors treating high ammonia concentration while studying N2O-N production in nitrifying systems. The nitrite accumulation was found to occur only under low dissolved oxygen concentrations. 41 • Cecen and Gonenc (1994) and Cecen and Ipek (1998) report that nitrite accumulation is not dependent upon free ammonia concentration alone, but on the ratio of dissolved oxygen to free ammonia concentration. It is also unlikely that a low dissolved oxygen concentration is the sole cause of nitrite accumulation, since each of the cited studies found that a low dissolved oxygen concentration alone could not explain the observed nitrite accumulation. Okada et al. (1999) reported that they have frequently observed increases in nitrite peak concentrations in surface biofilms with increases in the bulk oxygen concentration. Increases in the bulk oxygen concentration increase the ammonia oxidation rates in the surface biofilms but not the nitrite oxidation rates, because of the limitation of nitrite-oxidizing bacteria. Furthermore, active nitrite oxidation zones were found at oxygen concentrations below 50 u M in all cases. Interestingly, in the work of Okada et al. (1999), the active ammonia oxidizing zones in the biofilm were separated from the active nitrite oxidizing zones. The overall objective o f this study was to investigate the mechanisms responsible for nitrite accumulation in nitrification/denitrification systems. The initial hypothesis was that hydroxylamine, not free ammonia, was the true cause of nitrite accumulation. Based upon Yang and Allemans (1992) work using batch systems, it was believed that hydroxylamine might also be produced in continuous flow systems to cause nitrite oxidizer inhibition and ultimately nitrite accumulation. If hydroxylamine was indeed the cause of nitrite-oxidizer inhibition, then observed nitrite oxidizer acclimation could potentially be the result of a reduction in hydroxylamine concentration due to biological and/or chemical denitrification. The initial objectives of the research program were to confirm the primary cause(s) of nitrite accumulation in nitrification systems, the conditions required for hydroxylamine production, and the 42 mechanism(s) responsible for nitrite oxidizer acclimation in a system(s) operating via the nitrate shunt. The overall aim of the study was to develop a conceptual model that could potentially explain the phenomenon observed here and elsewhere. The specific hypotheses to be tested were as follows: Hypothesis No. 1 - Nitrite accumulation in continuous flow systems is due to the formation of free hydroxylamine, which can be formed by ammonia oxidizing organisms under conditions of low dissolved oxygen and high free ammonia. Nitrite accumulation is not the result of free ammonia inhibition of nitrite oxidizing organisms, as commonly reported. Hypothesis No. 2 - A high concentration of nitrite in any zone/phase not being used specifically for denitrification is detrimental to the nitrate shunt, since autotrophic denitrification or chemodenitrification can occur. Both processes would reduce the concentration of hydroxylamine responsible for selective inhibition of nitrite oxidizers. 1.4 THESIS OUTLINE As indicated in the Preface, this thesis is presented in manuscript-based format. The individual papers (manuscript-chapters) are presented in a logical manner that directs the reader towards the development of a conceptual model while addressing the objectives and testing the hypotheses presented in Section 1.3. Hydroxylamine is an unstable compound that autodecomposes to one of several decomposition products. One of the primary conclusions of Yang's PhD work (1990), was that a better 43 analytical method was needed for the hydroxylamine assay. Therefore, the starting point of this study was the development of an appropriate analytical method for hydroxylamine. The development of a suitable hydroxylamine assay took a great deal of effort. The work is presented in the manuscript attached as Chapter 2.0. To this authors' knowledge, no one has yet demonstrated hydroxylamine accumulation in a wastewater system during relatively stable operation. Yang and Alleman (1992) had demonstrated hydroxylamine accumulation in batch but not in continuous flow systems. Therefore, the next logical step in the research program was to create conditions believed to be conducive to hydroxylamine production. Several reactors were operated at various dissolved oxygen and solids retention times (SRTs) in an attempt to create the required conditions. The result of this work is summarized in Chapters 3.0 and 4.0. N o hydroxylamine was measured in these systems even though the ideal conditions of low dissolved oxygen and free ammonia stress were present in some cases. One potential explanation for the observed phenomenon was that hydroxylamine was being used as electron donor for autotrophic denitrification; however, i f this was the case, then hydroxylamine was unlikely to be the cause of the observed nitrite accumulation since the system operated for several months never appeared to acclimate. Recent research evidence suggests that Nitrospira most likely has a higher affinity for oxygen than ammonia oxidizers, so it is unlikely that low dissolved oxygen is the primary cause of the accumulation (Schramm et al. 2001). The experiments presented in Chapters 3.0 and 4.0 suggested that nitrous oxide might play a significant role in nitrite accumulation. Since hydroxylamine did not appear to be the cause of observed nitrite accumulation it was necessary to confirm whether or not free ammonia was actually inhibitory to nitrite-oxidizers. Most environmental engineering investigators have only used mixed cultures to confirm the 44 inhibitory affects o f free ammonia on nitrite-oxidizers. This approach ignores the potential competition for nitrite and oxygen between ammonia and nitrite oxidizers, when the concentration of free ammonia is high. In addition, Nitrospira is now considered the prevalent nitrite-oxidizer in wastewater systems. Most of the references to free ammonia toxicity refer to Nitrobacter rather than Nitrospira. Therefore, a combination of pure and mixed culture experiments was used to address the issue of free ammonia inhibition. The approach used and results obtained are presented in the manuscript attached as Chapter 5.0. The pure culture experiments conducted as part of the experiments described in Chapter 5.0, do not support free ammonia inhibition of Nitrospira. N o measurable quantities of hydroxylamine were found in the initial experiments (presented in Chapters 3.0 and 4.0). Therefore, it was decided to carry out a number of system perturbations that could result in nitrite accumulation and measure the response of the system and nitrifier population. Close monitoring of the acclimation response could potentially assist in determining the actual mechanism(s) responsible for nitrite accumulation. These experiments included a sustained free ammonia perturbation using a C S T R and anoxic zone p H control with an S B R system by using a similar operating strategy to that described by Turk and Mavinic (1986, 1987, 1989a, 1989b). These experiments did not produce any measurable concentrations of hydroxylamine. However, the possibility that nitrous oxide and specifically substrate competition may play a role in nitrite accumulation became apparent/ The results from the anoxic zone p H control indicated the control strategy itself likely biased the initial system response. The methodology used and results obtained from these experiments are presented in Chapters 6.0 and 7.0. 45 Yang and Alleman (1992) were able to produce measurable concentrations of hydroxylamine in batch systems. Therefore, several batch experiments were conducted to ascertain the mechanism responsible in order to assess the potential role of hydroxylamine, in observed nitrification phenomenon. Several experiments were also conducted to confirm hydroxylamine toxicity. The methodology used and results obtained from these experiments are presented in Chapter 8.0. Based upon the observations made throughout the studies a conceptual model was developed. This model is presented in Chapter 9.0 and is referred to as the "Unified Ammonia Oxidation Model". The implications of the model are discussed and future research requirements are presented. 46 Bibliography Abeling, U., and Seyfried, C F . 1992. Anaerobic-aerobic treatment of high-strength ammonium wastewater - nitrogen removal via nitrite., Wat. Sci. Technol., 26(5-6):1007-10T5. Ahlers, B., Konig, W., and Bock, E. 1990. Nitrite reductase activity in Nitrobacter vulgaris. FEMS Microbiol. Lett. 67(1-2): 121-126. Alleman, J.E. 1985. Elevated nitrite occurrence in biological wastewater treatment systems. Wat. Sci. Technol. 17 (2/3): 409-419. Anderson, I.C., and Levine, J.S. 1986. Relative rates of nitric oxide and nitrous oxide production by nitrifiers, denitrifiers and nitrate respierers. Appl. Environ. Biotechnol. 51(5): 938-945. Anthonisen, A.C., Loehr, R.C., Prakasam, T.B.S., and Srinath, E.G. 1976. Inhibition of nitrification by ammonia and nitrous acid. J. Water Pollut. Control Fed. 48: 835-851. Azimi, A. A., and Horan, N.J. 1991. The influence of reactor mixing characteristics on the rate of nitrification in activated sludge process. Water. Res. 25(4), 419-423. Balmelle, B., Nguyen, K.M. , Capdeville, B., Cornier, J.C., and Deguin, A. 1992. Study of factors controlling nitrite build-up in biological processes for water nitrification. Wat. Sci. Technol. 26(5-6), 1017-1025. Beccari, M. , Marani, D., and Ramodori, R. (1979), A critical analysis of nitrification alternatives. Water Res. 13: 185-192. Bergmann, D.J., Arciero, D., and Hopper, A.B. 1994. Organization of the HAO gene cluster of Nitrosomonas,europaea: Genes for two tetra heme cytochromes. J. Bacteriol. 176: 3148-3153. Bock, E., Koops, H.P., Harms, H., and Ahlers, A. 1991. The biochemistry of nitrifying organisms, pp. 171-199. J.M. Shively and L.L. Barton (Editors) In: Variations in autotrophic life., Academic Press, Ltd., London, England. .. := ». Bock E., Stueven R., and Zart D. 1995. Nitrogen loss caused by denitrifying Nitrosomonas cells using ammonium or hydrogen as electron donors and nitrite as electron acceptor. Arch. Microbiol. 163: 16-20. Bock, E., and Wagner, M . 2003. Oxidation of inorganic nitrogen compounds as an energy source, In: The Prokaryotes an evolving electronic resource for the microbial community (Release 3.14). Springer-Verlarg, Heidelberg, Germany. Boon, B., and Laudelout, H. 1962. Kinetics of nitrite oxidation by Nitrobacter winogradskyi. Biol. Biochem. 10: 377-381. Castell, C.H., and Mapplebeck, E.G. 1956. A note on the production of nitrite from hydroxylamine by heterotrophic bacteria. J. Fish Research Bd. Canada. 13: 201-206. 47 Castignetti, D., and Gunner, H.B. 1980. Sequential nitrification by : an Alcaligenes sp. and Nitrobacter agilis. Can. J. Microbiol. 26: 1114-1119. ' ' Castignetti, D., and Gunner, H.B. 1982. Differential tolerance of hydroxylamine by an Alcaligenes sp., a heterotrophic nitrifier, and by Nitrobacter agilis, Can'. J.'Microbiol. 28: 148-150. Cecen, F. 1996. Investigation of partial and full nitrification characteristics of fertilizer wastewaters in a submerged biofilm reactor. Wat. Sci. Technol. 34(11): 77- 85. Cecen, F., and Gonenc, I.E. 1994. Nitrogen removal characteristics of nitrification and denitrification filters. Wat. Sci. Technol. 29(10-11): 409-416. Cecen, F., and Ipek, S. 1998. Determination of the inhibition of ammonia-N and urea-N oxidations by the fed-batch reactor (FBR) technique. Wat. Sci. Technol. 38(1): 141-148. Chen, S.K., Juaw, C.K., and Cheng, S.S. 1991. Nitrification and denitrification of high strength ammonium and nitrite wastewater with biofilm reactors. Wat. Sci. Technol. 23(7-9): 1417-1425. de Bruijn P., van de Graaf, A.A. , Jetten, M.S.M., Robertson, L.A. , and Kuenen, J.G. 1995. Growth of Nitrosomonas europaea on hydroxylamine. FEMS Microbiol. Lett. 125:179-184. Ehrich, S., Behrens, D., Lebedeva, E., Ludwig W., and Bock E. 1995, A new obligately chemolithotrophic, nitrite-oxidizing bacterium, Nitrospira moscoviensis sp. nov. and its phlogenetic relationship. Arch. Microbiol. 164:16-23. Fdz-Polanco, F., Garcia P.A., and Villaverde, S. 1994. Temperature effect over nitrifying bacteria activity in biofilters: activation and free ammonia inhibition. Wat. Sci. Technol. 30(11):121-130. Fdz-Polanco, F., Villaverde, S., and Garcia P.A. 1996. Nitrite accumulation in submerged biofilters - combined effects. Wat. Sci. Technol. 34(3-4):371-378. Focht, D.D., and Chang, A.C. 1975. Nitrification and denitrification process related to wastewater treatment. Adv. Appl. Microbiol. 19:153-186. Ford, D.L., Churchwell, R.L., and Kachtick, J.W. 1980. Comprehensive analysis of nitrification of chemical processing wastewaters. J. Water Pollut. Control Fed. 51(11):2726-2746. Freitag, A., Rudert, M . , and Bock, E. 1987. Growth of Nitrobacter by dissimilatoric nitrate reduction. FEMS Microbiol. Lett. 48( 1 -2): 105-109. Freitag, A., and Bock, E. 1990. Energy conservation in Nitrobacter. FEMS Microbiol. Lett. 66:157-162. Frijlink, M.J., Abee, T., Lanbroek H.J., de Boer, W., and Konings, W.N. 1992. The bioenergetics of ammonia and hydroxylamine oxidation in Nitrosomonas europaea at acid and alkaline pH. Arch. Microbiol.l57:194-199. 48 Garrido, J . M . , Campos, J .L. , Mendez, R., and Lema, J . M . 1997a,.Nitrous oxide production by nitrifying biofilms in a biofilm airlift suspension reactor. Wat. Sci . Technol. 36(1):157-163. Garrido, J . M . van Benthum, W . A . J . , van Loosdrecht, M . C . M , and Heijnen, J.J. 1997b. Influence of dissolved oxygen concentration on nitrite accumulation in a biofilm airlift suspension reactor. Biotechnol. Bioeng. 53:168-178. Ghyoot, W. , Vandaele, S., and Verstraete, W. 1999. Nitrogen removal from sludge reject water with a membrane-assisted bioreactor. Water Res. 33(1): 23-32. Giannakis, C , Mil le r , D.J . , and Nicholas, D.J .D. 1985. Comparitive studies on redox proteins from ammonia-oxidizing bacteria. F E M S Microbiol . Lett. 30: 81-85. Goreau, T.J. , Kaplan, W . A . , Wofsy, S . C , McElroy , M . B . , Valois F .W. , and Watson, S.W. 1980. Production of N O x and N2O by nitrifying bacteria at reduced concentrations of oxygen. App l . Environ. Microbiol . 40(3): 526-532. Grady, C .P .L . , Daigger, G.T. , and Henry, C . L . 1999. Biological Wastewater Treatment (Second Edition). Marcel Dekker Inc. Hanaki, K . , Chalermraj, W. , and Shinichiro, O. 1990a. Effects of heterotrophs on nitrification in a suspended-growth reactor. Water Res. 24(3): 289-296. Hanaki, K . , Wantawin, C , and Ohgaki, S. 1990b. Nitrification at low levels of dissolved oxygen with and without organic loading in a suspended-growth reactor. Water Res. 24(3): 297-302. Hellinga, C , Schellen, A . A . J . C , Mulder, J.W., van Loosdrecht, M . C . M . , and Heijnen, J.J. 1998. The S H A R O N process: an innovative method for nitrogen removal from ammonium-rich waste water. Wat. Sci. Technol. 37(9): 135-142. Helmer, C , Kunst, S., Juretschko, S., Schmid, M . C , Schleifer, K . H . , and Wagner, M . 1999. Nitrogen loss in a nitrifying biofilm system. Wat. Sci. Technol. 39(7): 13-21. Hiorns, W . D . , Hastings, R . C , Head, I . M . , McCarthy, A . J . , Saunders, J.R., Pickup, R .W. , and Hal l , G . H . 1995. Amplification of 16S ribosomal R N A genes of autotrophic ammonia-oxidizing bacteria demonstrates the ubiquity of Nitrospirds in the environment. Microbiol . 141: 2793-2800. Hollocher, T . C 1984. Source of the oxygen atoms of nitrate in the oxidation of nitrite by Nitrobacter agilis and evidence against a P - O - N anhydride mechanism in oxidative-phosphorylation. Arch . Biochem. Biophys. 233(2): 721-727. Hommes, N . G . , Sayavedra-Soto, L . A . , and Arp, D.J . 1998. Mutagenesis and expression of amo, which codes for ammonia monooxygenase in Nitrosomonas europaea. J. Bacteriol. 180:3353-2259. Hooper A . B . , 1968, A nitrite reducing enzyme for Nitrosomonas europaea. Preliminary characterization with hydroxylamine as electron donor. Biochim. Biophys. Acta, 162, 49-65. 49 Hooper, A .B . , 1989. Biochemistry of the nitrifying lithoautotrophic bacteria. Autotrophic Bacteria, pp. 239-265. H.G. Schlegel and B. Bowien (Editors). Science Tech. Publishers, Madison, Wisconsin. Hooper, A .B . , Arciero, D .M. , DiSpirito, A .A . , Fuchs, J., Johnson, M . , LaQuier, F., Mundform, G., and McTavish, H . 1990. Production of nitrite and N 2 0 by ammonia-oxidizing nitrifiers. In: Nitrogen fixation: achievements and objectives. Gresshoff, Roth, Stacey & Newton (Editors). Chapman and Hall, New York. Hooper, A .B . , and Dispirito, A . 1985. In bacteria which grow on simple reductants, generation of a proton gradient involves extracytoplasmic oxidation of substrate. Microbiol. Rev. 49(2): HO-IST. Hooper, A .B . , and Terry, K.R. 1979. Hydroxylamine oxidoreductase of Nitrosomonas production of nitric oxide from hydroxylamine. Biochim. Biophys. Acta. 571(1): 12-20. Horan, N.J., and Azimi, A . A . 1992. The effects of transient nitrogen loadings on nitrifying activated sludge in completely mixed and plug-flow reactors. Water Res. 2 6 : 279-284. Hu, S.S. 1990. Acute substrate-intermediate-product related inhibition of nitrifiers. M.S. Thesis. School of Civil Engineering, Purdue University, West Lafayette, Indiana. Hunik, J.H., Meijer, H.J.G., and Tramper, J. 1993, Kinetics of Nitrobacter agilis at extreme substrate, product and salt concentration. Appl. Microbiol. Biotechnol. 40: 442-448. Hyman, M.R., and Wood, P .M. 1984. pp. 49-52. In: Microbial Growth on C\ Compounds. (R.L. Crawford, and R.S. Hanson, (Editors). Proc. 4 t h Int. Symp. American Society of Microbiology, Washington, D.C. Hyungseok Yoo, Kyu-Hong Ahn, Hyung-Jib Lee, Kwang-Hwan Lee, Youn-Jung Kwak and Kyung-Guen Song. 1999. Nitrogen removal from synthetic wastewater by simultaneous nitrification and denitrification (SND) via nitrite in an intermittently-aerated reactor. Water Res. 33(1): 145-154. Jones, G.J., and Paskins, A.R. 1982. Influence of high partial pressure of carbon dioxide and/or oxygen on nitrification. J. Chem. Tech. Biotechnol. 3 2 : 213-223. Kavlentis, E. 1988. Spectrophotometric determination of hydroxylamine using the DPGH-Iron (II)-NH 2OH ternary complex. Microchem. J. 3 7 : 22-24. Kelso, B.H.L., Smith, R.V., and Laughin, R.J. 1999. Effects of carbon substrates on nitrite accumulation in freshwater sediments. Appl. Environ. Microbiol. 65(1): 61-65. Koops, H.P., and Moller, U.C. 1991. The lithotrophic ammonia oxidizing bacteria, pp. 2625-2637. In: The Prokaryotes, 2 n d edn. A . Balows, H.G. Truper, M . Dworkin, W. Harder and K . H . Schleifer (Editors)., New York, Fischer-Verlag. Krummel, A . and Harms, H. 1982. Effect of organic matter on growth and cell yield of ammonia-oxidizing bacteria. Arch. Microbiol. 1 3 3 : 50-54. 50 Kumar, S., Nicholas, D.J .D. , Will iams, E . H . 1983. Definitive 1 5 N evidence that water serves as a source of O during nitrite oxidation by Nitrobacter agilis. F E B S . Lett. 152(1): 71-74. Laanbroek, H.J . , and Gerards, S. 1993. Competition for limiting amounts of oxygen between Nitrosomonas europaea and Nitrobacter winogradskyi grown in mixed continuous cultures. Arch. Microbiol . 159: 453-459. Mauret, M . , Paul, E . , Puech-Costes, E . , Maurette, M . T . and Baptiste, P. 1996. Application of experimental research methodology to the study of nitrification in mixed culture., Wat. Sci . Technol. 34(1-2): 245-252. McTavish, H . , Fuchs, J .A. , and Hooper, A . B . 1993. Sequence of the gene coding for ammonia monooxygenase in Nitrosomonas europaea. J. Bacteriol. 175(8): 2436-2444. Mil ler , D.J . , and Wood, P . M . 1983. Co-binding c-type cytochromes and a high potential cytochrome-c in Nitrosomonas europaea. Biochem. J. 211(2): 503-506. Mobarry, B . K . , Wagner, M . , Urbain, V . , Rittmann, B . E . , and Stahl, D . A . 1996. Phylogenetic probes for analyzing abundance and spatial organization of nitrifying bacteria. App l . Environ. Microbiol . 62(6): 2156-2162. Neufeld, R .D . , H i l l , A . J . , and Adekoya, D.O. 1980. Phenol and free ammonia inhibition to Nitrosomonas activity. Water. Res. 14:1695-1703. Okabe, S., Satoh, H . , and Watanabe, Y . 1999. In situ analysis of nitrifying biofilms as determined by in situ hybridization and the use of microelectrodes. App l . Environ. Microbiol . 65(7): 3182-3191. Payne, W.J . 1973. Reduction of nitrogenous oxides by microorganisms. Bacteriol. Rev. 37: 409-452. Poth, M . , and Focht, D . D . 1985. 1 5 N kinetic analysis o f N2O production by Nitrosomonas europaea: ah examination of nitrifier denitrification. App l . Environ. Microbiol . 49: 1134-2241. Prakasam, T.B.S. , and Loehr, R . C . 1972. Microbial nitrification and denitrification in concentrated wastes. Water Res. 6: 859-869. Prosser, J. I. 1989. Autotrophic nitrification in bacteria. Adv . Microbiol . Physiol. 30: 125-181. Quinlan, A . 1980. The thermal sensitivity of nitrification as a function of the concentration of nitrogen substrate. Water Res. 14: 1501-1507. Randall, C .W. , and Buth, D . 1984a. Nitrite build up in activated sludge resulting from combined temperature and toxicity effects. J. Water Pollut. Control Fed. 56(9): 1045- 1049. Randal, C .W. , and Buth, D . 1984b. Nitrite build-up in activated sludge resulting from temperature effects. J. Water Pollut. Control Fed. 56(9): 1039-1044. 51 Ritchie, G . A . F . , and Nicholas, D.J .D. 1972. Identification of the sources of nitrous oxide produced by oxidative and reductive processes in Nitrosomonas europaea. Biochem. J. 126: 1181-1191. Sauter, L . J . , and Alleman, J. E . 1980. A streamlined approach to biological nitrogen removal., Proc. A m . Soc. C iv . Engineers Environ. Eng. Div . pp. 296-306. Schmidt, I., and Bock,. E . 1997. Anaerobic ammonia oxidation with nitrogen dioxide by Nitrosomonas eutropha. Arch . Microbiol . 167(2-3): 106-111. Schmidt, I., Zart, D . , and Bock, E . 2001a. Gaseous NO2 as a regulator o f ammonia oxidation for Nitrosomonas eutropha. Antonie Van Leeuwenhoek. 79(3-4): 311-318. Schmidt, I., Bock, E . , and Jetten, M . S . M . 2001b. Ammonia oxidation by Nitrosomonas eutropha with NO2 as oxidant is not inhibited by acetylene. Microbiology. 147: 2247-2253. Schmidt, I., 2003. In: The Prokaryotes an evolving electronic resource for the microbial community (Release 3.14). Springer-Verlarg, Heidelberg, Germany. Schmidt, I., Sliekers, O., Schmid, M . , Bock, E . , Fuerst, J., Kuenen, J .G. , Jetten, M . S . M . , and Strous, M . 2003. N e w concepts of microbial treatment processes for the nitrogen removal in wastewater. F E M S Microbiol . Rev. 27(4):481-492. Schramm A . , de Beer D . , van den Heuvel, J . C , Ottengraf, S., and Amann, R. 1999. Microscale distribution of populations and activities of Nitrospira and Nitrospira spp. along a macroscale gradient in a nitrifying bioreactor: quantification by in situ hybridization and the use of microsensors. App l . Environ. Microbiol . 65(8): 3690-3696. Schramm, A . , de Beer, D . , Gieseke, A . , and Amann, R. 2001. Microenvironments and distribution of nitrifying bacteria in a membrane-bound biofilm. Environ. Microbiol . 2(6): 680-686. Spieck, E . , Ehrich, S., Aamand, J., and Bock, E . 1998. Isolation and immunocytochemical location of the nitrite-oxidizing system in Nitrospira moscoviensis. Arch . Microbiol . 169: 225-230. Stein, L . Y . , and Arp , D.J . 1998. Loss of ammonia monooxygenase activity in Nitrosomonas europaea upon exposure to nitrite. App l . Environ. Microbiol . 64(10): 4098-4102. Stuven, R., Vollmer, M . , and Bock, E . 1992. The impact of organic matter on nitric oxide formation by Nitrosomonas europaea. Arch. Microbiol . 158: 439-443. Sundermeyer-Klinger, H . , Meyer, W. , Warninghoff, B . , and Bock, E . 1984. Membrane-bound nitrite oxidoreductase of Nitrobacter-evidence for a nitrate reductase system. Arch. Microbiol . 140(2-3): 153-158. Surmacz-Gorska, J. , Cichon, A . , and Miksch, K . 1997. Nitrogen removal from wastewater with high ammonia nitrogen concentration via shorter nitrification and denitrification. Wat. Sci . Technol. 36(10): 73-78. 52 Suzuki, I., Dular, U., and Kwok, S.C. 1974. Ammonia or ammonium ion as substrate for oxidation by Nitrosomonas europaea cells and extracts. J. Bacteriol. 120(1): 556-558.. Sutherson, S., and Ganczarczyk, J.J. 1986. Inhibition of nitrite oxidation during nitrification some observations. Water Poll. Res. J. Canada. 21(2). Tanaka, Y., Fukumori, Y., Yamanaka, T. 1983. Purification of cytochrome-AlCl from Nitrobacter agilis and characterization of nitrite oxidation system of the bacterium. Arch. Microbiol. 135(4): 265-271. Turk, O. 1986. Ph.D. thesis. The feasibility of a shortened pathway for nitrogen removal from highly nitrogenous wastes. University of British Columbia., Vancouver. Turk, O., and Mavinic, D.S. 1986. Preliminary assessment of a shortcut in nitrogen removal from wastewater, Can. J. Civil Eng. 13: 600-605. Turk, O., and Mavinic, D.S. 1987. Selective inhibition: a novel concept for removing nitrogen from highly nitrogenous wastes. Env. Technol. Lett. 8: 419. Turk, O. and Mavinic, D.S. 1989a. Stability of nitrite build-up in an activated sludge system. J. Water Pollut. Control Fed. 61(8): 1440-1448. Turk, O., and Mavinic, D.S. 1989b. Influence of process changes on maintaining nitrite build-up in an activated sludge system acclimated to free ammonia. Water Res. 23(11): 1383-1388. U.S. Environmental Protection Agency, (1993), Nitrogen Control Manual. van Benthum, W.A.J., Derissen, B.P., van Loosdrecht, M.C.M., and Heijnen, J.J. 1998. Nitrogen removal using nitrifying biofilm growth and denitrifying suspended growth in a biofilm airlift suspension reactor coupled with a chemostat. Water Res. 32(7): 2009-2018. Villaverde, S., Fdz-Polanco, F., and Garcia, P.A. 2000. Nitrifying biofilm acclimation to free ammonia in submerged biofilters start-up influence. Water Res. 34(2): 602-610. Voets, J.P., Vanstaen, H.., and Verstraete, W. 1975. Removal of nitrogen from highly nitrogenous wastewaters. J. Water Pollut. Control Fed. 47: 394-398. Wagner, M. , Rath, G., Koops, H.P., Flood, J., and Amann, R. 1996. In situ analysis of nitrifying bacteria in sewage treatment plants. Wat. Sci. Technol. 34(1-2): 237-244. Wild, H.E. 1971. Factors affecting nitrification kinetics. J. Water Pollut. Control Fed. 43:(9): 1845-1971. White, D. 1995. The physiology and biochemistry of prokaryotes, Oxford University Press, New York, Oxford. Wong-Chong, G.M., and Loehr, R.C. 1978. Kinetics of microbial nitrification: nitrite-nitrogen oxidation. Water Res. 12: 605-609. 53 Yang, L . , and Alleman, J .E. 1992. Investigation of batchwise nitrite build-up by an enriched nitrification culture. Wat. Sci . Technol. 26 (5-6): 997-1005. Yang, L . 1990. Investigation of nitrite build-up within an enriched nitrification process. Ph.D. Dissertation, School of C i v i l Engineering, Purdue University, West Lafayette, Indiana. Zheng, H . , Hanaki, K . , and Matsuo, T. 1994. Production of nitrous oxide gas during nitrification of wastewater. Wat. Sci . Technol. 30(6): 133-141. 54 2.0 Hydroxylamine analysis of wastewater samples via gas chromatography1 Introduction Several investigators (Stuven et al. 1992; Yang et Alleman 1992; and Hyungseok Yoo et al. 1999) have suggested that hydroxylamine - an intermediate in the ammonia oxidation process -may be the true cause of nitrite oxidizer inhibition and therefore of nitrite accumulation in wastewater systems. The work of Yang and Alleman (1992) is of particular interest since these investigators reported that nitrite accumulation in batch nitrification systems was coincidental with the accumulation of free hydroxylamine. One of the primary recommendations of Yangs' PhD thesis, on which the paper of Yang and Alleman (1992) is based, is that a better method for hydroxylamine analysis was required. However, Yang does not provide any indication of his reasons for this conclusion. Yang and Alleman (1992) used the spectrophotometric method described by Kavlentis (1988) in their research work. The method of Kavlentis (1988) is based upon the reaction between Di-(2-pyridyl)ketone guanylhydrazone (DPGH) and iron (II) in ammonia medium to form a yellow complex. This complex is decolorized by the addition of sodium hydroxide but in the presence of hydroxylamine a stable blue-colored complex is formed. It is the colored reaction that is employed for the spectophotometric determination of hydroxylamine. According to Kavlentis (1988) hydroxylamine can be determined in the range 5-150 ppm using this method. It was decided to pursue an alternative method for this research, partly due to the recommendation presented by Yang (1990) but primarily due to the fact that the inhibitory concentration of 1 A version of this chapter has been submitted for publication. Simm, R.A., Parkinson, P., Mavinic, D.S., Ramey, W.D. Hydroxylamine analysis of wastewater samples via gas chromatography. Environ. Technol., July, 2004. 55 hydroxylamine to nitrite-oxidizers was unknown. There was some concern that the method of Kavlentis (1980) might not provide a suitable method detection limit i f a hydroxylamine concentration significantly below 5 ppm was found to be biologically significant. For this reason, we decided to investigate alternative methods for the hydroxylamine assay. These included the rc-hexyl chlorformate-mediated dervatization developed by Angelino et al., (1998), the spectrophotometric method of Frear and Burrell (1955), and the acetone derivatization procedure described by Darke (1980). The rc-hexyl chloroformate method described by Angelino et al., (1998) was tried first since it appeared to be most sensitive (reported method detection limit of 10 ppb in water samples). In this procedure the sample is mixed with «-hexyl chloroformate under sonication and subsequently a saturated pyridine solution of dicyclohexylcarbodimide is added. The resulting solution is extracted with «-hexane, allowed to separate, and 1 p L of the hexane is immediately injected into a G C / M S . We were unable to recover hydroxylamine spikes in distilled water or filtered reactor supernatant using this method and this method was abandoned. The spectrophotometric method of Frear and Burrell (1955) is very sensitive. These authors used the method to measure hydroxylamine in millimolar quantities (reported concentration range of between 0 and 0.33 mg NH2OH/L) in plant tissue extract. This method is based upon the reaction of hydroxylamine with an excess of 8-quinolinol in the presence of ethyl alcohol and sodium carbonate to form 5,8-quinolinequinone-5-(8-hydroxy-5-quinolylimide) or Indooxine. The reaction is carried out for 1 minute in a boiling water bath after which time the reaction vessel is allowed to cool for fifteen minutes for color development. The color complex is stable for a short period of time (Frear and Burrell (1955) suggest 30 minutes following the 15 minute 56 cooling period). We found the method to be sensitive for hydroxylamine fortified distilled water, however, background absorbance was a significant problem in reactor supernatant. In addition, the analytical procedure was considered logistically problematic in a batch test situation wherein the investigator might want to take hydroxylamine samples every 10 to 15 minutes in addition to samples for several other analytes. The most promising analytical method was the acetone derivatization procedure developed by Darke (1980) for colonic fluid. Darke (1980) was able to recover hydroxylamine added to colonic distillate in the range of 0.1 mg/L to 1 g/L. We have adapted this method to measure hydroxylamine in bench scale reactors treating synthetic wastewater. The details of the reactors and the synthetic feed are summarized in Simm et al., (2004). The developed method is presented in the text below. Experimental Materials Hydroxylamine hydrochloride, acetone, and acetone oxime were obtained from Fischer, Sigma, and Kodak respectively. Apparatus A n HP 5890 Series II gas chromatograph fitted with a flame ionization detector (FID) was used for this analysis. A 900-pL-injection port liner with single taper was used. Specific attention 57 was paid to column insertion depth into the injection port. A n insertion depth of 2 mm was found to be optimum in order to avoid split peaks. Although the insertion depth reported here was optimum for our analysis, column insertion depth would likely have to be optimized for a specific G C and sample matrix. A splitless injection was used for this analysis. A D B - W a x column was used (15 meters long, 0.53 mm I.D., with a fi lm thickness of 0.5 um) for this analysis. The injection port temperature was 120 °C. The carrier gas was helium and the carrier gas flow was 20 mL/min. The initial purge came on at 0.25 minutes. The oven temperature program included an initial temperature of 45 °C for 0.5 minutes, a temperature ramp to 90°C at 5 °C/min, followed by a temperature ramp to 200 °C at 40 °C/min. Procedure M i x e d liquor samples were collected from operating bench scale reactors using a gas tight syringe. Samples collected from reactors having a low mixed liquor suspended solids concentration (<700 mg/L) were filtered (approx. 3 mL) through a 0.45 um filter directly into a 10 m L glass vial containing 60 uL of acetone (derivatizing agent approx. 2% of total). In those cases where direct filtration was difficult or impossible, the sample was discharged directly into a 10 m L glass vial , centrifuged for. 30 seconds, the supernatant was poured into a syringe capped at the end, and then approximately 3 m L was filtered through a 0.45 um filter into a clean glass vial containing 60 uL of acetone. - The glass vial containing acetone and filtered sample was capped using a teflon screw cap and the contents were shaken vigorously for 15 to 20 seconds. The derivatized sample was allowed to sit for 5 minutes and then 1-2 u L was transferred to a G C vial for analysis. Sample injections were limited to 1 uL, since injections of more than 2 uL 58 resulted in problems associated with sample 'flash back". The recovery values were 100% in distilled water and between 70 and 100% in reactor supernatant. The method detection limit is 10 ppb in distilled water and between 10 and 50 ppb in wastewater samples collected from bench scale reactors treating synthetic wastewaters. A typical chromatogram, from the analysis of a reactor supernatant sample, is presented as Fig...2.1. Discussion A s indicated by Darke (1980), primary and secondary amines, as well as nitrite, nitrate, and ammonia do not interfere with the detection of acetone oxime. We were able to confirm that neither of these compounds in fact interfered with the assay. In addition, several p H trials confirmed the derivatization reaction is not sensitive to p H in the range of p H used in the experiments conducted as part of this research. The acetone oxime produced in the derivatization reaction was relatively stable and there was no significant difference in chromatographic response over an eight hour period making this test well suited for batch test or reactor tracking studies where samples may be collected for several parameters every ten or fifteen minutes. The test's simplicity, good yields, and relatively low detection limit, made it ideally suited for measuring hydroxylamine in synthetic wastewater matrices. This assay was used to measure biologically produced hydroxylamine at concentrations as high as 5 ppm (Simm et al., 2004). 59 Figure 2.1 - Typical chromatogram for hydroxylamine recovered from a reactor sample. 60 Bibliography Angelino, S., Maurino, V., Minero, C , Pelizzetti, E., and Vincenti, M . 1998. Improved procedure for «-hexyl chloroformate-mediated derivatization of highly hydrophilic substances directly in water: hydroxylamine compounds. J. Chromat. A. 7 9 3 : 307-316. Darke, D.J. 1980. Method for the measurement of hydroxylamine in colonic fluid using derivatisation and gas chromatography. J. Chromat. 181: 449-452. Frear, D.S., and Burrell, R.C. 1955. Spectrophotometric method for determining hydroxylamine reductase activity in higher plants. Biochemical J. 27(10): 1664-1665. Hyungseok Yoo, Kyu-Hong Aim, Hyung-Jib Lee, Kwang-Hwan Lee, Youn-Jung Kwak, and Kyung-Guen Song. 1999. Nitrogen removal from synthetic wastewater by simultaneous nitrification and denitrification (SND) via nitrite in an intermittently aerated reactor. Water Res. 33(1): 145-154. Kavlentis, E. 1988. Spectrophotometric determination of hydroxylamine using the DPGH-Iron(II)-NH2OH ternary complex. Microchem. J. 3 7 : 22-24. Simm, R.A., Mavinic, D.S., and Ramey, W.D. 2004. The role of hydroxylamine as a potential inhibitor of nitrite oxidizers in wastewater treatment systems, (submitted for publication to Journal of Environmental Engineering and Science, February 2004). 61 Stuven, R., Vollmer, M . , and Bock, E . 1992. The impact of organic matter on nitric oxide formation by Nitrosomonas europaea. Arch. Microbiol . 1 5 8 : 439-443. Yang, L . 1990. Investigation of nitrite build-up within an enriched nitrification process. PhD Dissertation, School of C i v i l Engineering, Purdue University, West Lafayette, Indiana. Yang, L . , and Alleman, J .E. 1992. Investigation of batchwise nitrite build-up by an enriched nitrification culture. Wat. Sci . Technol. 26(5-6): 997-1005. 62 3.0 Preliminary Evaluation of the Use of Fatty Acid Ratios for Tracking the Potential for Nitrite Accumulation in Nitrifying Reactors with Low Carbon to Nitrogen Ratio2 Introduction There has' been a considerable amount of interest in the past ten to fifteen years in the design of nitrification/denitrification systems employing nitrite as the primary intermediate - thereby eliminating the formation of nitrate - in both the nitrification and denitrification steps. Numerous authors (Voets et al. 1975; Turk and Mavinic 1986, 1987, 1989a, 1989b; Balmelle 1992; Chen et al. 1991; Fdz-Polanco et al. 1996; Garrido et al. 1997; Hyungseok Yoo et al. 1999) have reported the capital and operational benefits of the process, referred to here as the nitrate shunt; to include a 25% reduction in aeration requirements, a 40% reduction in external carbon addition for denitrification, a potential reduction in anoxic zone volume and a significant reduction in sludge production. To date, free ammonia is the consensus cause of nitrite oxidizer inhibition and therefore considered the key to process operation. However, Cecen and Ipek (1998) have called into question this conclusion, suggesting the ratio of dissolved oxygen to free ammonia is of primary importance when attempting to induce nitrite accumulation. The work of others (Stuven et al. 1992; Yang et Alleman 1992; and Hyungseok Yoo et al. 1999) has suggested that hydroxylamine - an intermediate in the ammonia oxidation process - may be the true cause of nitrite oxidizer inhibition and therefore of nitrite accumulation. 2 A version of this chapter has been accepted for publication. Simm, R.A., Ramey, W.D., and Mavinic, D.S. 2004. Preliminary evaluation of the use of fatty acid ratios for tracking the potential for nitrite accumulation in nitrifying reactors with low carbon to nitrogen ratio, published, J. Environ. Eng. Sci. 3: 31-40. 63 Although there is some disagreement on the cause of nitrite accumulation resulting from nitrite oxidizer inhibition, the general consensus on the long term viability of the process is aptly summarized by Van Loosdrecht and Jetten (1998) who conclude: "nitrogen removal via the oxidation of ammonia to nitrite that is subsequently denitrified does not appear stable for long term operation." The primary limitation of the process appears to be the apparent acclimation o f nitrite oxidizing organisms to increasing concentrations of free ammonia. This has been reported by numerous researchers, including Turk and Mavinic (1986, 1987, 1989a, 1989b) and Villaverde et al. (2000). The acclimation response is typically the result of proliferation of small populations, change in the occurrence of toxins, predation by protozoa, appearance of new genotypes, diauxie, and enzyme induction (Alexander 1999). Any useful method to characterize the population should assist in distinguishing between these effects. The problems associated with using standard culturing methods for studies of this nature have been widely reported (Bergey's Manual o f Systematic Bacteriology 2001). Nitrifying bacteria have proven particularly difficult to study by cultivation techniques such as most probable number and selective plating, because of their long generation times and poor counting efficiencies for these types of bacteria. Over the past ten years, molecular methods have gained wider acceptance in environmental engineering research, providing valuable insights into the makeup of the complex mixed microbial communities responsible for wastewater treatment. These methods include slot and dot blot hybridization, fluorescent in situ hybridization (FISH), denaturing gradient gel electrophoresis ( D G G E ) , and restriction fragment length polymorphism (RFLP) to name but a few. The primary barriers to using these methods for engineering research include their cost, 64 complexity and the time requirements relative to conventional analyses'. In fact, many of these methods are so time consuming and tedious that they are simply impractical in most laboratories. A n alternative cheaper method that has been used in the Environmental Engineering Laboratory at the University of British Columbia (Werker and Hal l 2002) for the analysis of mixed microbial populations is based upon the analysis of fatty acid methyl esters ( F A M E ) . Fatty acids are the lipids of widest distribution within the bacterial cell. They are mainly found with carbon chain lengths between C H and C20, though both longer and shorter chains are not unusual. Fatty acids may be fully saturated and may be linear, branched, or contain alicyclic rings. The analysis of fatty acids involves saponification, acidification, extraction and methylation. Fatty acid methyl esters are identified using gas chromatograpy (GC) or gas chromatography/mass spectrometry ( G C / M S ) , which are commonly available in most environmental engineering laboratories. The objective of this study was to determine the potential utility of F A M E analysis for tracking conditions/circumstances that lead to a change in the nitrite oxidizer population. The overall objective of the research program was to determine whether hydroxylamine (produced under combined free ammonia and low dissolved oxygen stress) plays a role in nitrite accumulation. If the results the F A M E analyses prove promising, this technique could/can be used, in conjunction with standard liquid and gas analyses, to answer a number o f fundamental questions with respect to the acclimation response such as: do we have a gradual slow increase in the nitrite oxidizer population size or do certain chemical/reactor conditions actually precede the acclimation/inhibition response? Monitoring changes in the microbial population in conjunction with hydroxylamine analyses would assist in the determination of the role of hydroxylamine, i f any, in nitrite accumulation and population changes that precede acclimation. 65 Background The process of nitrification is carried out by two phylogenetically unrelated groups of autotrophic bacteria, the ammonia oxidizers and the nitrite oxidizers. Characterized ammonia-oxidizing bacteria are restricted to the y and subdivisions of the Proteobacteria (purple bacteria). Members of the genera Nitrosospira, Nitrosovibrio, Nitrosolobus, Nitrosomonas, and Nitrosococcus mobilis are within the P subclass of the Proteobacteria (Juretschko et al. 1998). Nitrosococcus oceanus and N. halophilus belong to the y subdivision (Juretschko et al. 1998). Nitrite-oxidizing genera occur in the alpha, gamma, and delta subclasses of the Proteobacteria (Teske et al. 1994) as well as the Nitrospira-phylum (Enrich et al. 1995; Spieck and Bock 2001). Although Nitrobacter has historically been considered to be the dominant nitrite-oxidizing organism in wastewater treatment systems, recent research using modern molecular techniques (Schramm et al. 1998; Juretschko et al. 1998; Daims et al. 2000) indicates that Nitrospira is the predominant nitrite-oxidizer. Although there has been a great deal of work conducted On the development of molecular probes for ammonia and nitrite-oxidizing organisms (Mobarry et al. 1996; Hovanec and DeLong 1996; and Hovanec et al. 1998) there has been very little work done on collecting chemotaxonomic data with the exception of the excellent work on the fatty acid composition of nitrite-oxidizers done by Lipski et al. (2001). The only previous major work has been the work conducted by Blumere ta l . (1969). 66 Blumer et al. (1969) analyzed the fatty acid composition of nineteen marine and terrestrial nitrifying bacteria. The terrestrial or fresh-water strains included Nitrosolobus multiformis, Nitrosomonas europaea, Nitrosomonas sp., and two unidentified strains from a Chicago sewer. Two strains of Nitrobacter agilis were studied as part o f this research. Blumer et al., reported that the lipid composition of nitrifying bacteria, especially o f the,ammonia oxidizers, is less complex than that found in most other bacterial groups. In fact, cis-9 hexadecenoic acid accounted for 60 to more than 80% of the total acids, whereas the sum of hexadecanoic acid (16:0) and the cis-9 hexadecenoic acid (16:1) made up 96 to 100% o f the acids in the lipids of the cell. The fatty acid composition of Nitrobacter was found to cover a much wider range of fatty acids, from C u to C 19; however the fatty acid composition was dominated by two to four fatty acids that account for more than 80% of the total acid content. The most abundant fatty acids for the Nitrobacter strains studied were the 18:1 (approximately 50 to 60% o f the total) and 19:1 fatty acids (approximately 30%). The most prevalent fatty acid for a marine strain o f Nitrospina was found to be a 16:1 acid (approximately 40 to 50% of total). Lipski et al. (2001) analyzed the fatty acid profiles of representative strains for all described species of the nitrite-oxidizing genera Nitrobacter, Nitrococcus, Nitrospina, and Nitrospira. The analyses were done on autotrophic and mixotrophically grown cells harvested under both exponential and stationary growth phases. These authors found the genera studied can be differentiated by genus specific fatty acid profiles. Additionally, they found only minor changes in fatty acid profiles related to different growth conditions (lithoautotrophic, mixotrophic, or chemoorganotrophic). The genus Nitrobacter was characterized by up to 92% vaccenic acid (18:1 cis 11) and the absence of hydroxy fatty acids. The representative strains of the two Nitrospira species studied were characterized by the presence of cis-1 and cis-11 hexadecenoic acid (16:1 cis 7 and 16:1 cis 11), representing between 5-34% and 15-37% of the total 67 respectively. The strain Nitrospira moscoviensis was characterized by a large proportion of a 16-carbon fatty acid identified as 11-methyl branched palmitate, which dominated the profiles of autotrophically grown cells with portions higher than 30%. Mixotrbphic growth resulted in a decrease in this fatty acid, whereas the amount of 16:0 increased. Enrichment cultures of Nitrospira moscoviensis, grown at 42 C° and 47 C°, contained high concentrations of the 11-methyl branched palmitate but no measurable amount of cis-\ 1 hexadecenoic acid. Lipski et al., indicate that this fatty acid has not yet been reported in any other organism and therefore can be regarded as a specific l ip id marker for this taxon. The use o f F A M E profiles for characterizing changes in the nitrifier population of mixed bacterial communities has not received serious attention in the past, due to a number of perceived difficulties with this technique. A s indicated above, nitrifiers generally possess relatively common bacterial fatty acids, with the exception of the 11-methyl branched palimate for Nitrospira and 19:1 acid for Nitrobacter. This limitation of available lipid markers is particularly significant given that these bacteria typically represent a relatively small proportion of the total in most mixed populations. In addition, the amount and type of fatty acids in nitrifiers, like that in all bacteria, varies according to how the microorganism is grown (temperature, nature o f carbon and nitrogen nutrients, oxygen levels, presence and absence of trace metals, etc.), or at what stage during growth the organism is taken for analysis. A t first glance, these facts would tend to suggest the technique might be of questionable utility when studying nitrifiers in mixed microbial communities. However, the limited research conducted to date suggests nitrifiers may dominate the fatty acid composition of a mixed microbial community treating high strength ammonia waste so that their contribution would be less affected by changes in environmental conditions than systems with 68 many other organisms in the community. The proportion of nitrifiers would be relatively high as a result of a low C O D / T K N ratio and their individual contribution to the specific fatty acid totals would likely be greater. The fatty acid composition of most nitrifiers appears to be dominated by one or two fatty acid species, and lipids make up a much larger percentage of the total bacterial weight (typically 18-20%) in nitrifiers relative to other bacterial species (e.g. 2% for E . coli); this is due to their extensive intercytoplasmic membranes (Gunstone et al. 1986). In addition, the preliminary information presented by both Blumer et al. (1969) and Lipsk i et al. (2001) suggests that relative to other organisms, variation in growth conditions may have a lesser impact upon the fatty acid composition of nitrifiers. Methodology Reactor System Six, 10.8-liter bench scale reactors were operated at various dissolved oxygen levels, ambient temperatures (21-23 C°), and different solids retention times in order to produce varying degrees of partial nitrification. Each of the six reactors was designed with a gas tight headspace, to allow collection of headspace gas samples. Reactor headspace was vented through a gas collection bulb and beaker filled with water to allow development of a constant headspace pressure in each reactor. A l l reactors were run as continuous stirred tank reactors (CSTRs) , with secondary clarification having a two-hour settling time. Each reactor was mixed with a 100-rpm motor. The influent feed to each reactor was 24 liters per day, giving a reactor hydraulic retention time (HRT) of approximately 11 hours. The dissolved oxygen concentration in each reactor was controlled manually at one of three levels: < 0.5 mg/L, 0.8 to 1.2 mg/L, and 2.0-3.0 mg/L. Once the rotameter was set and the system was operating within the predetermined dissolved oxygen 69 range, there was little need to adjust the aeration setting. The dissolved oxygen in each reactor was measured with a dissolved oxygen probe and logged with a DataTaker™ data logger. The pH in each reactor was measured once daily, using a hand held pH probe and meter. The measured pH in the six reactors was 7.4-7.5, 7.4-7.7, 7.3-7.4, 7.4-7.5, 7.5-7.6, and 7.7-7.8 for reactors R l , R2, R3, R4, R5, and R6 respectively. Reactors R3 through R6 were operated at a reduced oxygen concentration in order to induce partial nitrification of influent ammonia creating a constant free ammonia stress, which in combination with low dissolved oxygen concentration would hopefully lead to hydroxylamine accumulation. The Garrett configuration in which biomass is wasted directly from the reactor (Grady et al. 1999) was used for solids control and solid wasting was carried out daily. The reactor setup is illustrated in Fig. 3.1. The original mixed culture seed for all six reactors was taken from a laboratory scale reactor operating via the University of Cape Town (UCT) process to biologically remove phosphorus. Synthetic Feed Each of the six reactors was fed a synthetic feed that simulated a relatively weak landfill leachate with low carbon to nitrogen ratio. The feed consisted of 250 mg/L sodium acetate (NaCHaCOOH), 556 mg/L ammonium chloride (NH4CI) , 1,936 mg/L sodium bicarbonate (NaHCOs), 50 mg/L yeast extract, 56 mg/L potassium phosphate ( K 2 H P O 4 X 37.5 mg/L magnesium chloride (MgCl2 -6H20), 24 mg/L calcium chloride (CaCl2-2H20), 1.9 mg/L ferric chloride (FeCl3-6H20), 0.09 mg/L manganese sulphate ( M n S 0 4 H 2 0 ) , 0.008 mg/L sodium molybdate (Na 2Mo0 4-2H 20), 0.375 mg/L zinc sulphate (Zn S0 4 -7H 2 0), and 0.001 mg/L cobalt chloride (CoCl2-2H20). 70 Analytical Methods Fatty Acid Methyl Esters (FAME) One to five mL of mixed liquor suspended solids was collected from each reactor and placed in a 15 mL glass tube, with a Teflon screw cap. The collected mixed liquor was then centrifuged at 3600 rpm (1,100 g) for ten minutes and the supernatant was decanted. The sludge pellet was stored at - 20 °C until analysis. Upon analysis, the sample was thawed and distilled water was added to make up the volume to the original. A 1-mL aliquot of 1 N NaOH in 50% HPLC grade methanol, spiked with O-methylpodocarpic acid (O-MPCA), was added to each sample. The O-MPCA served as an internal standard. The sample tubes were then incubated at 90 °C for thirty minutes in order to saponify the cellular fatty acids. The sample was then acidified with 1 mL of 1 N H2SO4, in order to reduce the aqueous solubility of the fatty acids prior to extraction. Following acidification, the sample was extracted twice with 1-mL of methyl-tert-butyl-ether (MTBE). The sample was centrifuged following each extraction and solvent containing fatty acids was transferred directly to a 2 mL GC vial and dried under vacuum. After the second drying step, a 100 pl aliquot of MTBE spiked with heneicosanoic acid methyl ester (HAC-ME) and tricosanoic acid (TCA) was added to each vial (HCA-ME was used as the internal standard and TCA was used as the control for methylation). The sample was then methylated by dispensing and vortexing 400 pl of chilled MTBE and HPLC grade methanol (80:20) containing excess dissolved diazomethane. The extracted microbial fatty acids were identified and measured by GC/FID (HP 5890 Series II with a DB5 column of 30 meters length and having a 0.32 mm ID and 0.25 pm film thickness) with 1 pl splitless injection. The carrier gas for the analysis was hydrogen. The inlet and detector temperatures were 250 °C and 300 °C, 71 respectively. The temperature program included an initial four minute hold at 130 °C followed by a ramp of 4°C/min. to 250 °C and finally a post-run temperature of290 °C for five minutes. N 2 0 The headspace of each of the six reactors was fully enclosed thus allowing for the collection of reactor off gases. Off gas samples were collected using a gas tight syringe and analyzed for nitrous oxide by gas chromatography, using a Hewlett-Packard 5880A equipped with an electron capture detector. In the GC analysis, nitrogen was used as the carrier gas (at 20 mL/min) with a column packing material of Haycep C. The injector, oven, and detector temperatures were 80 °C, 80 °C, and 250 °C, respectively. Chemical Analyses The analysis of ammonia, total Kjeldahl nitrogen, nitrate, nitrite, total organic carbon, and total dissolved solids was carried out as described in the 19th Edition of Standard Methods for the Examination of Water and Wastewaters (Eaton et al. 1995). Hydroxylamine was measured according to the method of Simm et al. (2004a). Molecular Methods RNA for slot blotting was extracted from mixed liquor samples immediately following collection. RNA extractions were conducted using Trizol™ reagent and bead beating, according to the manufacturers protocols, using a Bio-Spec mini-bead beater. This procedure is based upon the RNA extraction method developed by Chomczynski and Sacchi (1987). Oligonucleotide 72 probes were synthesized at the University of British Columbia Nucleic A c i d and Protein Synthesizing (NAPS) unit and unincorporated nucleotides were removed from the probes using an ammonia-butanol purification step. The lyophilized probes were solubilized in RNase-free water prior to dilution for use in probe labeling. A l l oligonucleotide probes were labeled to a specific activity of 10 8-to-10 9 C P M / u g with 3 2 P by using a 5' end labeling kit as supplied by Amersham Biosciences ( R P N 1509). Slot blot hybridizations were conducted in general accordance with the method outlined by Stahl et al. (1988), Raskin et al. (1994) and Mobarry et al. (1996). The molecular probes used for this preliminary assessment included EUB-338 (5'-GCT GCC TCC CGT AGG AGT-3) targeting all eubacteria, Nso 190 (5'-CGA TCC CCT GCT TTT TCT CC-3') targeting all characterized ammonia-oxidizers in the P subdivision of the Proteobacteria (purple bacteria), Nb 1000 (5'-TGC GAC CGG TCA TGG-3) targeting Nitrobacter, and Ntspa-685 (5'-CAC CGG GAA TTC CGC GCT CCT C-3') targeting Nitrospira moscoviensis and Nitrospira marina. The EUB-338 , N b 1000, and Nso 190 probes have been previously described by Mobarry et al. (1996). The Ntspa-685 probe for Nitrospira has been previously described by Hovanec et al. (1998). Fluorescent In Situ Hybridization (FISH) was conducted using the Nit r i -VIT™ kit supplied by Vermicon Inc., according to the manufacturers instructions. Results This part o f the research program was conducted between February 1, 2002 and March 3, 2002. The reactor conditions and influent/effluent results for each of the six reactors are summarized in Table 3.1. Reactors R l and R2 were operated at a relatively high dissolved oxygen concentration (2.0-3.0 mg/L) to ensure complete nitrification; reactors R3 through R6 were 73 operated at lower dissolved oxygen concentration in an attempt to induce varying degrees of partial nitrification. Although reactors R3 and R4 were both operated at a dissolved oxygen concentration of 0.8-1.2 mg/L, they were operated at different dissolved oxygen concentrations prior to the start of this assessment. Reactor R4 was operated at a dissolved oxygen concentration of 1.5-1.7 mg/L for one month, prior to the start of the run; reactor R3 was operated at a dissolved oxygen concentration of <0.5 mg/L for one month, prior to the start of the run. No measurable concentrations of hydroxylamine were recovered from any of the reactors even though four of the six reactors were operated with a low dissolved oxygen concentration and a sustained free ammonia stress (R3- NH 3 -N of 0.02 to 0.1 mg-N/L, R4 - NH 3 -N of 0.08 to 0.35, R5 - NH 3 -N of 0.21 to 0.71, R6 - NH 3 -N of 1.94 to 2.5). Analysis of the fatty acid composition of the six reactors indicated that the fatty acid profile for mixed liquor samples taken from each reactor was dominated by four fatty acids, accounting for 50-80% of the fatty acid composition of the mixed liquor samples. One fatty acid with a chain length of 16 carbons was found in significant quantities in reactors with high effluent nitrate concentrations. The concentration of this particular fatty acid was relatively low or absent from reactors where nitrite, but no nitrate, was present. This major fatty acid was identified as cis-l 1 hexadecenoic acid by its mass spectra and equivalent chain length calculation. Lyophilized samples were sent to the University of Guelph Food Sciences Department (Guelph, Ontario) for confirmatory analysis, using the MIDI™ system. The four remaining fatty acids that dominated the fatty acid profile included cis-9 hexadecenoic acid, hexadecanoic acid, and vaccenic acid. The dominant fatty acid in all cases was cis-9 hexadecenoic acid. The relative proportion of each of the major fatty acids is summarized in Fig. 3.2. The 95% confidence limit (9 samples) for each fatty acid percentage is indicated on the figure. There is some variation in the percentage that each fatty acid makes up relative to the total but it is quite clear that the pattern in 74 the high dissolved oxygen reactors ( R l & R2) is distinctly different from the medium (R3 & R4) to low dissolved (R5 & R6) oxygen reactors. The difference between medium (R3 & R4) dissolved oxygen reactors and the low dissolved oxygen reactors is in the proportion of cis-11 hexadecanoic acid. The primary difference with respect to the fatty acid profile of the six reactors appeared to be the relative proportions of cis(9) and cis (77)-hexadecenoic acids. Given that cis (9)- and cis (11) are reported to be major fatty acids associated with a number of ammonia-oxidizers in the P subgroup of the Proteobacteria and Nitrospira respectively, and that Nitrospira are believed to be the primary nitrite-oxidizers in wastewater systems, a correlation between the ratio of cis (9) hexadecenoic acid to cis (11) hexadecenoic acid and nitrite as a percentage o f the total N O x was investigated. With the exception of R3 , which exhibited an appreciable change in both nitrite and nitrate concentrations during the preliminary assessment period, plots of the cis (9) hexadecenoic acid/c/5 (11) hexadecenoic acid versus nitrite, as a percentage of total N O x , clustered around a localized area. The applicable plots for reactor R3 are presented in Fig. 3.3 and 3.4. These plots indicated that the ratio of these two fatty acids might be useful for tracking the acclimation of nitrite-oxidizers to changing reactor conditions. Interestingly, there was a strong correlation between this ratio and the concentration o f nitrous oxide in the reactor headspace. A t this point in time it is difficult to determine whether the measured nitrous oxide was produced by autotrophic or partial heterotrophic denitrification. However, the ability o f ammonia oxidizing organisms to produce nitrous oxide under low dissolved oxygen conditions has been well established (Anderson and Levine 1986). With the results of the correlations for reactor R3, and given the fact that each of the six reactors were receiving identical feed stock, it was decided to plot the fatty acid ratio for all the reactors 75 versus the percentage of effluent nitrite ((NO2-N/NO2-N+NO3-N)*100%)) data for each on one graph. The resulting plot is presented as Fig.. 3.5. A semi-logarithmic curve fit was applied to the data presented in Fig. 3.5. The Pearson's correlation coefficient was computed to determine the degree of correlation between these two variables. The correlation coefficient was 0.974 and the ninety-five percent confidence limit is (0.954, 0.985). Figure 3.5 suggests that the ratio of 16:1 (9)/16:1(11) must be less than two, if one wants to ensure complete nitrification and greater than 30 if the goal is nitrite accumulation. The plot also confirms the potential merit of using fatty acid profiles to track nitrifier populations in low C:N wastes. Discussion The results suggest that fatty acid profiling may serve as a useful tool for tracking nitrifier populations in high strength ammonia wastes with low carbon to nitrogen ratio. It would be useful as a process development tool for tracking nitrite-oxidizer acclimation for processes operating via the nitrate shunt. In addition, it may be possible to use fatty acid analysis as a tool for assessing the stability of nitrification processes where nitrification to a nitrate end point is desired. Based upon this data, it appears that as the ratio of 16:1(9) to 16:1(11) increases a particular system becomes more prone to nitrite accumulation. This tendency is predicated upon Nitrospira-type organisms being the dominant nitrite-oxidizer. The fatty acid results for reactor R3 indicate a gradual decrease in the fatty acid ratio with time that correlated with the reduction in headspace nitrous oxide and effluent nitrite concentrations (Figures 3.3 and 3.4). This reduction in fatty acid ratio was primarily the result of a slow, but 76 steady increase, in the relative proportion of cis-11 hexadecenoic acid in the overall total, suggesting that this technique may. be useful for tracking the proliferation of small Nitrospira populations. Surprisingly the 11-methyl branched palmitate, which is suggested as a possible taxon specific marker for Nitrospira by Lipski et al. (2001), was not found in measurable quantities in the fatty acid samples taken for this assessment. However, Lipski et al. (2001) indicate in their work that mixotrophic growth resulted in a decrease in the amount of this l ipid marker. The fact that this study was conducted at a temperature o f 21-23 °C, as opposed to the culture temperature o f 37 °C used by Lipski and his colleagues may also have been a factor, since lipid concentration can be temperature dependant (Gunstone et al., 1986). N o measurable quantities of hydroxylamine were recovered, even in reactors operated under conditions considered favorable for its production, suggesting hydroxylamine decomposed prior to measurement, or was not the cause of observed nitrite accumulation. Those reactors exhibiting nitrite accumulation were operated under both free ammonia and low dissolved oxygen stresses indicating that either one or both of these factors may be responsible for observed phenomenon. This preliminary trial indicated that nitrous oxide emissions were generally coincidental with nitrite accumulation. The fact that headspace nitrous oxide concentration declined with increasing cis-11 hexadecenoic acid (Fig. 3.4) suggests nitrous oxide production might some how be related to Nitrospira inhibition, given the fact the cis-11 hexadecenoic acid was believed to be a direct marker of Nitrospira organisms. The two reported nitrite-oxidizing groups generally associated with wastewater treatment systems are Nitrobacter and Nitrospira, although recent research suggests Nitrospira is the primary nitrite oxidizer in the majority of cases (Schramm et al. 1998; Juretschko et al. 1998; Daims et al. 2000). The preliminary results using the fatty acid ratio and limited probing with 77 the Nitrospira specific Ntspa-685 probe and Nitrobacter specific N b 1000 probe, indicate that Nitrospira was, in fact, the dominant nitrite-oxidizer in the systems studied for this assessment. However, in order for fatty acid analysis to be useful for its intended purpose in this research program, it must allow discrimination between a Nitrospira and Nitrobacter dominated nitrite-oxidizing population. Although the results o f this assessment do not provide evidence that this sort of resolution is possible, a later trial with reactor R5 suggests the fatty acid technique might in fact provide the required information (Chapter 6 and Simm et al. 2004b). Reactor R5 was subjected to a free ammonia transient under relatively high dissolved oxygen (>3mg/L) conditions. The fatty acid ratio increase, during a period of free ammonia stress, was consistent with R N A analyses using the Nso 190 and Ntspa 454 probes (the Ntspa 454 probe is specific for Nitrospira-Wks nitrite-oxidizers). The ratio stabilized once complete ammonia oxidation had been established and did not change appreciably during a time period when percent nitrite oxidation was increasing. Probing with the Nb 1000 probe indicated the Nitrobacter population was increasing relative to the Nitrospira population during this time suggesting Nitrobacter was primarily responsible for additional nitrite oxidation and supporting the specificity of the fatty acid ratio as an indirect measure of the ammonia oxidizer population relative to the Nitrospira population. The above result suggests that one possible cause o f the reported acclimation response in processes operating via the nitrate shunt may, in fact, be a switch in the dominant nitrite-oxidizer population from Nitrospira to Nitrobacter. This switch is supported by the facts that Nitrospira appears to be the dominant nitrite-oxidizer in most municipal wastewater systems and is less tolerant to high concentrations of nitrite than Nitrobacter (Spieck, E . , and Bock, E . 2001). Given that Nitrobacter is a relatively common soil nitrifier and that soil organisms are thought to be the dominant source of nitrifiers in municipal wastewater treatment systems, it seems likely that 78 some Nitrobacter would be present in the mixed liquor from municipal wastewater plants, even though it is not dominant. Therefore, any process operated via the nitrate shunt, whereby significant nitrite accumulation took place, would likely create an environment more favorable for Nitrobacter relative to Nitrospira. The very long time periods required for the appearance of significant effluent nitrate may wel l be the result of the very small number o f Nitrobacter species present in the original seed. The important point in the context of this paper is that these data support the utility of the fatty acid technique for tracking this type of population shift. Assuming Nitrospira was the dominant nitrite-oxidizer in most sewage treatment systems it is likely that previous researchers, who have taken seed from a municipal wastewater treatment system almost without exception, may have initially created process conditions that inhibited Nitrospira, only to eventually create conditions that favored Nitrobacter. This response would manifest itself as a very slow and gradual increase in the contribution of nitrate to the overall oxidized nitrogen species. This change would take a very long time to produce a measurable result. This is consistent with the observations of both Turk and Mavinic (1986) and Velliverde et al. (2000) in their studies of the nitrate shunt. Although fatty acid profiling does not directly measure enzyme induction or diauxie, these acclimation responses would likely manifest themselves as a relatively stable fatty acid profile under conditions of changing process performance. This is not unlike the way in which molecular methods would be used in this same situation. It is not being suggested here that the fatty acid analysis would serve as a replacement for more sophisticated molecular techniques. On the contrary, its proposed use for this research program is as a low cost routine monitoring tool that can be used to prescreen samples to be subjected to molecular analysis. In this way, the analytical burden can be substantially reduced. 79 Conclusions The following conclusions have been drawn based upon this research: • In systems with low carbon to nitrogen ratios and relatively high ammonia concentrations the nitrification activity showed a strong correlation to the fatty acid profile. This correlation can be used to assist in tracking the population acclimation response and in evaluating nitrification stability. • This preliminary evaluation was conducted on a waste having a low carbon to nitrogen ratio. The conclusions regarding the absolute value of the nitrifier fatty acid ratio would not apply to other wastes and the relationships presented here would be waste specific. The important point is that the technique appears useful for tracking population changes and would therefore be useful as a routine research tool. • Fatty acid analyses could serve as an inexpensive screening tool in support of molecular analysis in process development studies. In this case, one would use fatty acid analysis on a routine basis to select previously collected molecular samples for analysis, thus optimizing the analytical program. • N o measurable quantities of hydroxylamine were recovered from reactors under conditions considered ideal for its production (i.e. free ammonia and low dissolved oxygen stress) even though nitrite accumulation was observed, suggesting hydroxylamine may not be the cause of the observed phenomenon. 80 Stirrer Motor Reactor offgas Gas Tight Headspace Reactor outlet pas collection M b Clarifier Scrapper Motor Effluent Secondary Clarifier Feed Bucket Figure 3.1 - Reactor setup schematic 81 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0 • 16:1(9) •16:1(11) D16:0 D18:1 o E-o a o a -R1 R2 R3 R4 R5 R6 Reactor Fig. 3.2 - Average percentage of major fatty acid components during preliminary assessment period (9 samples) to 100 40 30 0 5 10 15 20 25 30 35 40 45 50 Fatty Acid Ratio 16:1(9)/16:1(11) Fig. 3.3 - Ratio of 16:1(9)/16:1(11) versus nitrite-N as percent NOx for R3 (February 1, 2002-March 3, 2002) 0.00 , 0 5 10 15 20 25 30 35 40 45 Fatty Acid Ratio Fig. 3.4 - Ratio of 16:1(9)/16:1(11) versus headspace nitrous oxide concentration for R3 (February 1,2002 - March 3, 2002) oo 4^ oo Table 3.1 - Reactor influent/effluent conditions Dissolved Oxygen Reactor Solids Total Reactor SRT Concentration Suspended T O C Ammonia-N T K N - N Nitrite-N Nitrate-N (days) (mg/L) (mg/L) Sample Statistic Solids (mg/L) (mg/L) (mg/L) (mg/L) (mg/L) (mg/L) RI 10 2-3 1220-1688 Influent Range N A 77.4-105.6 147.7-155.9 166.3-188.7 Note (1) Note(l) Avg. = 1429 Effluent Average N A • 98.1 152 176.4 Note(l) Note(l) Range 2-32 2.8-7.7 0 0-2.2 0-1.13 121.2-156.5 R2 Average 16 4.7 0 0.93 0.27 137.4 5 2-3 425-795 Influent Range N A 76.7-139.4 138.2-155 171-182.3 Note(l) Note(l) Avg. = 626 Effluent Average N A 105.6 147.9 175.5 Note(l) Note(l) Range 1-57 3.3-18.8 0-43.6 0-49.8 0-21.3 76.9-150.6 R3 Average 19 9 8.2 15.9 8.6 118.3 5 0.8-1.2 554-840 Influent Range N A 77.4-105.6 147.7-155.9 166.3-188.7 Note (1) Note(l) Avg. = 646 Effluent Average N A 98.1 152 176.4 Note(l) Note(l) Range 14-64 4-18.2 1.8-8.1 3-89.2 69.8-109.5 7-51.3 R4 Average 45 8.7 5 16 89.8 30.7 5 0.8-1.2 513-895 Influent Range N A 76.7-139.4 138.2-155 171-182.3 Note(l) Note(l) Avg. = 724 Effluent Average N A 105.6 147.9 175.5 Note(l) Note(l) Range 5-93 3.1-17.8 6.5-23 9.8-27.4 1-41.3 62.8-126.5 R5 10 Average 44 7.4 16.5 18 27 85.9 <0.5 680-1020 Influent Range N A 77.4-105.6 147.7-155.9 166.3-188.7 Note (1) Note(l) Avg. = 860 /Effluent Average N A 98.1 152 176.4 Note(l) Note(l) Range 9-45 2.3-18 13.4-37.2 12.7-43.8 61.5-114.6 3.9-12.8 R6 Average 26 8 30.3 32 78.1 8.5 5 <0.5 242-479 Influent Range N A 76.7-139.4 138.2-155 171-182.3 Note(l) Note(l) Avg. = 347 Effluent Average N A 105.6 147.9 175.5 Note(l) Note(l) Range 8-47 2.4-27.1 80.8-103.3 91.8-123.6 25.2-35.5 0-4.4 Average 21 10.7 88.7 107.4 31.2 2 NA - Not applicable. < for all samples taken. oo 0\ Bibliography Alexander, M . 1999. Biodegradation and bioremediation (Second Edition). Academic Press, San Diego, California. Anderson, I.C., and Levine, J.S. 1986. Relative rates of nitric oxide and nitrous oxide production by nitrifiers, denitrifiers and nitrate respierers. Appl. Environ. Biotechnol., 51(5): 938-945. Balmelle, B., Nguyen, K.M. , Capdeville, B., Cornier, J.C, and Deguin, A. 1992. Study of factors controlling nitrite build-up in biological processes for water nitrification. Wat. Sci. Technol. 26(5-6): 1017- 1025. Blumer, M. , Chase, T., and Watson, S.W. 1969. Fatty acids in the lipids of marine and terrestrial nitrifying bacteria. J. Bacteriol. August 1969: 366-370. Boone, D.R., and Castenholz, R.W. (Editors), Bergy's manual for systematic bacteriology, 2 n d edition. Springer, New York. Cecen, F., and Ipek, S. 1998. Determination of the inhibition of ammonia-N and urea-N oxidations by the fed-batch reactor (FBR) technique. Water Sci. Technol. 38(1): 141-148. Chen, S.K., Juaw, C.K., Cheng, S.S. 1991. Nitrification and denitrification of high strength ammonium and nitrite wastewater with biofilm reactors, Water Sci. Technol. 23(7-9): 1417-1425. 87 Chomczynski, P., arid. Sacchi, N . 1987. Single-step method for. R N A isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal . Biochem. 162: 156-159. Daims, H . , Neilsen, P .H . , Juretschko, S., and Wagner, M . 2000. Novel Nitrospira-like bacteria as dominant nitrite-oxidizers in biofilms from wastewater treatment plants: diversity and in situ physiology. Water Sci . Technol. 41: 85-90. Eaton, A . D . , Clesceri, L .S . , Greenberg, A . E . (Editors) 1995. Standard methods for the examination of water and wastewater. 19 t h ed. American Public Health Association, Inc., New York. Enrich, S., Behrens, D . , Lebedeva, E . , Ludwig, W. , and Bock, E . 1995. A new obligately chemolithoautotrophic, nitrite-oxidizing bacterium, Nitrospira moscoviensis sp., nov. and its phylogenetic relationship. Arch . Microbiol . 164: 16-23. Fdz-Polanco, F. , Villaverde, S., and Garcia, P . A . 1996. Nitrite accumulation in submerged biofilters - combined effects. Water Sci. Technol. 34(3-4): 371-378. Garrido, J . M . , van Bethum, W . A . J . , van Loosdrecht, M . C . M . , and Heijnen, J.J. 1997. Influence of dissolved oxygen concentration on nitrite accumulation in a biofilm airlift suspension reactor. Biotechnol. Bioeng. 53: 168-178. Grady, C .P .L . , Daigger, G.T. , L i m , H . C . 1999. Biological wastewater treatment (Second Edition). Marcel Dekker, Inc., N e w York, N Y . 88 Gunstone, F.D., Harwood, J.L., and Padley, F.B. (Editors). 1986. The lipid handbook. Chapman and Hall, London and New York. Hovanec, T.A. and DeLong, E.F. 1996. Comparative analysis of nitrifying bacteria associated with freshwater and marine aquaria. Appl. Environ. Microbiol. 62: 2888-2896. Hovanec, T.A., Taylor, L.T., Blakis, A., and DeLong, E.F. 1998. Nitrospira-like bacteria associated with nitrite oxidation in freshwater aquaria. Appl. Environ. Microbiol. 64: 258-264. Hyungseok Yoo, Kyu-Hong Ahn, Hyung-Jib Lee, Kwang-Hwan Lee, Youn-Jung Kwak, and Kyung-Guen Song. 1999. Nitrogen removal from synthetic wastewater by simultaneous nitrification and denitrification (SND) via nitrite in an intermittently aerated reactor. Water Res. 33(1): 145-154. Juretschko, S., Timmermann, G., Schmid, M. , Schleifer, K.H., Pommerening-Roser, A., Koops, H.P., and Wagner, M . 1998. Combined molecular and conventional analyses of nitrifying bacterium in activated sludge: Nitrosococcus mobilis and Nitrospira-like bacteria as dominant populations. Appl. Environ. Microbiol. 64: 3042-3051. Lipski, A., Spieck, E., Makolla, A., and Altendorf, K. 2001. Fatty acid profiles of nitrite-oxidizing bacteria reflect their phylogenetic heterogeneity. Syst. Appl. Microbiol. 24: 377-384. Mobarry, B.K., Wagner, M. , Urbain, V., Rittmann, B.E., and Stahl, D.A. 1996. Phylogenetic probes for analyzing abundance and spatial organization of nitrifying bacteria. Appl. Environ. Microbiol. 62:2156-2162. 89 Raskin, L., Stromley, J.M., Rittmann, B. E., and Stahl, D. 1994. Group-specific 16S rRNA hybridization probes to describe natural communities of methanogens. Appl. Environ. Microbiol. 60: 1232-1240. Schramm, A., De Beer, D., Wagner, M., Amman, R. 1998. Identification and activities in situ of Nitrospira and Nitrospira spp. as dominant populations in a nitrifying fluidized be reactor. Appl. Environ. Microbiol. 64: 3480-3485. Simm, R.A., Parkinson, P., Mavinic, D.S., and Ramey, W.D. 2004a. Hydroxylamine analysis of wastewater samples via gas chromatography, (submitted for publication to J. Environ. Eng. Sci. February 2004). Simm, R.A., Ramey, W.D., and Mavinic, D.S. 2004b. Nitrifier population dynamics in a bench scale activated sludge reactor following an induced perturbation, (accepted for publication with minor revisions in Journal of Environmental Engineering and Science June 2004). Spieck, E., and Bock, E. 2001. Genus Nitrospira. In Bergy's Manual of Systematic Bacteriology. 2 n d ed., Vol. 1, Edited by D.R. Boone, R.W., Gastenholz, and G.M., Garrity. New York, Springer-Verlag. pp. 451-453 Stahl, D.A., Flesher, B., Mansfield, H.R., and Montgomery, L. 1988. Use of phylogenetically based hybridization probes for the studies of ruminal microbial ecology. Appl. Environ. Microbiol. 54: 1079-1084. 90 Stuven, R., Vollmer, M . , and Bock, E . 1992. The impact of organic matter on nitric oxide formation by Nitrosomonas europaea. Arch . Microbiol . 1 5 8 : 439-443. Teske, A . , A i m , E . , Regan, J . M . , Toze, S., Rittmanri, B : E . , Stahl, D . A . 1994. Evolutionary relationships among ammonia- and nitrite-oxidizing bacteria. J. Bacteriol. 1 7 6 : 6623-6630. Turk, O., and Mavinic , D.S . 1986. Preliminary assessment of a shortcut in nitrogen removal from wastewater. Can. J. C i v i l Eng. 1 3 : 600-605. Turk, O., and Mavinic , D.S. 1987. Selective inhibition: a novel concept for removing nitrogen from highly nitrogenous wastes. Env. Technol. Lett., 8 : 419. Turk, O., and Mavinic , D.S. 1989a. Stability of nitrite build-up in an activated sludge system. J. Water Pollut. Control Fed. 61(8): 1440-1448. Turk, O., and Mavinic , D.S . 1989b. Maintaining nitrite buildup in a system acclimated to free ammonia. Water Res. 23(11): 1383-1388. Van Loosdrecht, M . C . M . , and Jetten, M . S . M . 1998. Microbial conversions in nitrogen removal. Water Sci. Technol. 38 (1) : 1-7. Villaverde, S., Fdz-Polanco, F. , and Garcia, P .A . 2000. Nitrifying biofilm acclimation to free ammonia in submerged biofilters start-up influence. Water Res. 34 (2) : 602-610. 91 Voets, J.P., Vanstaen/H. and Verstraete, W. 1975. Removal of nitrogen from highly nitrogenous wastewaters. J. Water Pollut. Control Fed. 47: 394-398. Werker, A . , and Hal l , E . 2002. Development and application of a fatty acid based microbial community structure similarity index. Envirometrics, 13(4): 347-363. Yang, L . , and Alleman, J.E. 1992. Investigation of batch wise nitrite build-up by an enriched nitrification culture. Water Sci . Technol. 26(5-6): 997-1005. 92 4.0 Nitrifier Population Dynamics Following A Dissolved Oxygen Perturbation Introduction The work of Anthonisen et al. (1976) is considered by many to be the definitive work that established the inhibitory nature of free ammonia toward nitrite oxidizers. Several investigators have called into question the true cause of nitrite oxidizer inhibition and, as indicated by Schmidt et al. (2003), when discussing partial nitrification to a stable nitrite end product: "it is unclear why nitrite oxidizers are inhibited; inhibition of nitrite oxidizers by ammonia and lower affinity for oxygen and/or nitrite have been suggested as possible explanations, but we still lack mechanistic evidence". Cecen and Ipek (1998) have suggested the dissolved oxygen to free ammonia ratio and not the free ammonia concentration itself, is of primary importance when attempting to induce nitrite accumulation. The work of others (Stuven et al. 1992; Yang et Alleman 1992; and Hyungseok Yoo et al. 1999) has suggested that hydroxylamine - an intermediate in the ammonia oxidation process - may be the true cause of nitrite oxidizer inhibition and therefore of nitrite accumulation. The work of Yang and Alleman (1992) is of particular interest since these investigators reported that nitrite accumulation in batch nitrification systems was coincidental with the accumulation of free hydroxylamine. They surmised that free hydroxylamine, and not free ammonia, was the likely cause of observed nitrite accumulation. The work of Yang and Alleman (1992) therefore provides some theoretical basis for the conclusions of Cecen and Ipek 3 A version of this chapter has been submitted for publication. Simm, R.A., Mavinic, D.S., and Ramey, W.D. 2004. WEF Research Journal, July 2004. 93 (1998) and Hyungseok Yoo et al. (1999), who both reported that free ammonia and oxygen together appeared to play a key role in nitrite accumulation. The work reported here was the direct result of our inability to recover measurable concentrations of hydroxylamine under conditions believed to encourage hydroxylamine accumulation, namely free ammonia stress and low dissolved oxygen concentration (Simm et al., 2004a). The reactor used in this portion of the research accomplished partial nitrification and provided a stable nitrite end product for several months. The purpose of this portion of the research program was to measure the change in reactor and population dynamics, following an increase in dissolved oxygen concentration, and confirm whether the originally observed nitrite accumulation was caused by free ammonia stress or low dissolved oxygen. Methodology Reactor Systems This study was carried out using a bench-scale, continuous stirred, tank reactor (CSTR). The reactor was designed with a gas tight headspace, to allow collection of headspace gas samples. The reactor was operated at a 10-day solids retention time (SRT) and low dissolved oxygen concentration (<0.5 mg/L) to induce partial nitrification, the build up of a free ammonia stress, and therefore hydroxylamine accumulation. The details of the test reactor were presented in Chapter 3 and a previous publication (Simm et al., 2004a). 94 Synthetic Feed The C S T R was fed a synthetic feed that simulated a relatively weak landfill leachate with low carbon to nitrogen ratio, while providing a nitrifier-enriched culture. The feed composition is presented in Table 4.1. Chemical Analyses The analysis of ammonia, total Kjeldahl nitrogen ( T K N ) , nitrate, nitrite, total organic carbon, and total dissolved solids was carried out as described in the 19 t h Edition of Standard Methods for the Examination of Water and Wastewaters (Eaton et al. 1995). The free ammonia and nitrous acid concentrations were estimated using the relationships presented by Anthonisen et al. (1976). Hydroxylamine was measured using the G C method described in Chapter 2 (Simm et al. 2004b). Fatty Acid Methyl Esters (FAME)/RNA Slot Blotting Mixed liquor suspended solids microbial communities were characterized using fatty acid methyl esters ( F A M E ) and R N A slot blotting. The methodology employed is described in Chapter 3 (Simm et al. 2004a). Fluorescent In Situ Hybridization (FISH) A limited number of samples were analyzed via fluorescent in situ hybridization. Fluorescent In Situ Hybridization (FISH) was conducted using the Nitr i -VIT™ kit supplied by Vermicon Inc., according to the manufacturers instructions. 95 N 2 0 Reactor off gas samples were collected using a gas tight syringe and analyzed for nitrous oxide by gas chromatography, using a Hewlett-Packard 5880A equipped with an electron capture detector. In the GC analysis, nitrogen was used as the carrier gas (at 20 mL/min) with a column packing material of Haycep C. The injector, oven, and detector temperatures were 80 °C, 80 °C, and 250 °C, respectively. Nitric Oxide (NO) Off gas samples were collected for nitric oxide analysis using a gas tight syringe. The collected samples were immediately injected into a Sievers Model 280i Nitric Oxide Analyzer (NOA™). This instrument uses a high-sensitivity detector for measuring nitric oxide based on a gas-phase chemiluminescent reaction between nitric oxide and ozone. A thermoelectrically cooled, red-sensitive photomultiplier tube detects the emissions from the activated nitrogen dioxide produced. The detection limit of the NOA for measurement of gas-phase NO is approximately 0.5 ppbv. Results This reactor was operated between January 2001 and January 2002, at both a low dissolved oxygen concentration and a 10-day SRT. The overall performance data for the time period February 1 to March 3, 2002 were presented in a previous publication (referred to as Reactor R5 96 in Chapter 3 and Simm et al., 2004a). This reactor was operated at low dissolved oxygen in order to retard nitrification to the point where a free ammonia stress would be maintained and hydroxylamine could potentially accumulate. N o measurable concentrations of hydroxylamine were recovered from this reactor. B y May 2002, the aeration rate for this reactor had been reduced from 1.2 L /min to 0.38 L/min in an attempt to reduce ammonia oxidation, increasing residual free ammonia, and therefore induce hydroxylamine accumulation. The resulting reactor p H and dissolved oxygen concentrations were 7.7-7.9 and 0.2-0.3 mg/L, respectively. Ammonia oxidation declined to between 15 and 20%. The average reactor free ammonia (NH3) concentration ranged between 2.4 to 4.5 mg-N/L. The reactor was sampled once per week to ensure a free ammonia stress was present and that nitrite was the primary detectable oxidized nitrogen product. Between November 5, 2002 and November 28, 2002 an intensive sampling was done to establish baseline conditions prior to increasing the dissolved oxygen concentration. N o measurable amounts of hydroxylamine were recovered during this time. The headspace nitrous oxide concentration during this time period ranged between 700 and 1000 ppm. On November 28, 2002 the aeration rate was increased from 0.38 to 1.2 L/min. This was implemented to allow nitrification to recover and track the events leading up to the complete recovery of nitrite oxidation. The results of this aeration rate perturbation are summarized as Fig. 4.1 (November 28, 2002 is referred to as day 0 in this figure). The residual ammonia concentration in the reactor had dropped by over 60% within 6-hours of the aeration rate increase (reactor ammonia concentration dropped from 117.5 to 43.6 mg-N/L) and complete ammonia oxidation was essentially achieved within twenty-four hours (>95% ammonia removal). The 97 decrease in residual ammonia concentration was accompanied by a decline in headspace nitrous oxide concentration (Fig. 4 . IB) that continued until nitrite disappeared from the reactor on day 12. The initiation of nitrite oxidation (as measured by the presence of nitrate) did not appear to commence until day 7. This is 4 days after the reactor free ammonia concentration had declined to less than 0.1 mg-N/L (Fig. 4.1 A ) . Both effluent nitrite (Fig. 4.1 A ) and headspace nitrous oxide concentration (Fig. 4. IB) appeared to level off between days 3 and 7. The effluent nitrate concentration increased from 11.4 to 156.7 mg-N/L between days 7 and 12. This was coincidental with a decline in effluent nitrite from 147.6 to 4.2 mg-N/L (Fig. 4.1A) and headspace nitrous oxide from 107 to 20 ppm (Fig. 4 .IB). During the perturbation, emissions of ammonia, nitrite, nitrate, and nitrous oxide accounted for between 80 and 100% of the influent nitrogen loading to the reactor. The incorporation of nitrogen into biomass accounted for between 2 and 4% of the total. The nitric oxide emissions accounted for less than 1% of the measured total. The percentage of each nitrogen species (100% taken as the sum of ammonia, nitrite, nitrate, nitrous oxide, and synthesis) is summarized as Fig. 4.1C. These data suggest that the nitrate mass production rate does not start increasing until after it exceeds the nitrous oxide mass emission rate. This is supported by a scatter plot of nitrate versus nitrous oxide presented as Fig. 4.1D. The R N A slot blotting indicated no significant change in the concentration of P Proteobacterial ammonia-oxidizers, as measured by the Nso 190 R N A probe, before and after the aeration perturbation (Fig. 4.IE). The concentration of Nitrospira-\ik& nitrite-oxidizer R N A , as measured 98 by the Ntspa 454 R N A probe, increased by over four-fold by day 7 of the perturbation (Fig. 4.IF). The R N A slot blotting results indicate Nitrospira-like organisms survived in this system prior to the perturbation, even in the presence of long-term (>30 SRTs) free ammonia and dissolved oxygen stresses. N o Nitrobacter-like nitrite-oxidizing organisms were detected in this reactor, based upon probing results using the N b 1000 probe. The presence of nitrite-oxidizing organisms is clearly shown in the F I S H image dated September 19, 2002 presented in Fig. 4.2a (ammonia oxidizer colonies are presented in red and nitrite oxidizer colonies are presented in green). Unfortunately, this particular sample exhibited a significant autofluorescence in the green; however, it clearly shows nitrite-oxidizer colonies in the middle of the floe. The F I S H analyses conducted on December 13, 2002 (two samples Fig . 4.2b and Fig. 4.3a) and January 10, 2003 (Figure 4.3b) nitrite-oxidizer colonies are not significantly larger, but appear more numerous, than indicated in Fig. 4.2a. Previous research (Chapter 3 and Simm et al., 2004a) had indicated that fatty acid analyses, and particularly the ratio between 16:1(9) and 16:1(11) might be a good indicator of nitrite accumulation. Specifically, this ratio was found to increase with increasing percentage o f effluent nitrite. There was no change in the ratio of these two fatty acids during the first 13 days of the perturbation. The ratio was consistent with expectation, however, more than thirty days following the perturbation. Discussion The reactor used for this study sustained a stable nitrite end-point for several months, even though ammonia-oxidation itself was incomplete. The mixed liquor culture in this reactor was 99 subjected to both free ammonia and low dissolved oxygen stresses. This combination was believed to be required to produce measurable quantities of hydroxylamine, which Yang and Alleman (1992) have reported as inhibitory to nitrite-oxidizers. N o measurable quantities of hydroxylamine (detection limit <0.1 mg/L) were recovered from this reactor, suggesting that it is unlikely that hydroxylamine was responsible for the observed nitrite accumulation. A high headspace nitrous oxide concentration was coincidental with nitrite accumulation, and nitrous oxide is a known hydroxylamine decomposition product; however, chemodenitrification trials conducted as part of this research (Chapter 8 and Simm et al., 2004c) suggested that it is unlikely that hydroxylamine decomposition was the source of measured nitrous oxide. A small population of Nitrospira-\\ke nitrite-oxidizers was maintained in this reactor, even in the presence of a severe dissolved oxygen limitation and a free ammonia concentration (widely believed to be inhibitory). The fact that nitrite-oxidizers were able to survive for a long period, under such conditions, does not support free ammonia inhibition as the cause of the observed phenomenon. Pure cultures of Nitrospira moscoviensis were subjected to a free ammonia stress as part of this research, tb test the inhibitory properties of free ammonia to these organisms (Chapter 5 and Simm et al., 2004d). The results of the pure culture study do not support free ammonia inhibition. In addition, one would have expected a visible recovery in nitrite oxidation prior to day 7 if, in fact, free ammonia is inhibitory. If dissolved oxygen limitation was truly the cause of the observed phenomenon, one would have expected nitrite oxidation (as measured by a nitrate end point) to recover far before day 7. In addition, the work o f Schramm et al. (2000), suggests that the oxygen half saturation coefficient for Nitrospira may, in fact, be significantly lower than for Nitrobacter. If this is the case and the 100 half saturation coefficients for Nitrosomonas and Nitrobacter are similar, one must ask why is it then that Nitrosomonas is able to out-compete Nitrospira in this particular experiment? This reactor was used for a follow up experiment with a mixed liquor culture grown at a higher carbon to nitrogen ratio, but lower influent ammonia concentration (Chapter 6 and Simm et al. 2004e). However, in this case a high dissolved oxygen concentration (> 3 mg/L) and a sustained free ammonia perturbation was applied. The nitrite-oxidizer population in this experiment was compromised as a result o f the perturbation, even though free ammonia does not appear to be inhibitory and the bulk dissolved oxygen concentration was not limiting. A batch test conducted with another C S T R reactor supported the possibility that nitrite-oxidizers have a higher affinity for oxygen than ammonia oxidizers (Chapter 8 and Simm et al., 2004c). The dissolved oxygen perturbation indicated that nitrous oxide emissions were coincidental with nitrite accumulation. The potential sources for nitrous oxide emissions include chemodenitrification of hydroxylamine, heterotrophic denitrification, and autotrophic denitrification. Chemodenitrification of hydroxylamine was previously discounted as the primary source. Although heterotrophic denitrification cannot be completely discounted, the coincidental drop in reactor ammonia and headspace nitrous oxide concentration supports autotrophic denitrification as the most likely source. This is also supported by the R N A data, which indicates no significant difference in ammonia-oxidizer R N A (as measured by the Nso 190 probe) following the dissolved oxygen perturbation. The nitrogen balance (Fig. 4.1C) and nitrous oxide/nitrate scatter plot (Fig. 4 .ID) suggest the initiation of nitrite oxidation was coincidental with a threshold reduction in nitrous oxide emissions. One must ask, however, why wasn't nitrite oxidation initiated sooner than day 7, 101 given the fact that both dissolved oxygen and nitrite substrates were plentiful? The answer can be partially obtained from the F I S H images presented in Figures 4.2 and 4.3. The F I S H image for September 19, 2002 shows a very limited number of nitrite-oxidizer colonies in close proximity to several ammonia oxidizer colonies. These nitrite-oxidizer colonies would have to compete for oxygen and nitrite (assuming ammonia-oxidizers denitrify nitrite to nitrous oxide). In contrast, the F I S H images for samples taken on December 13, 2002 show a large number of nitrite-oxidizer colonies spread throughout the microscope field. Many of these nitrite-oxidizer colonies would be in a more opportune location to take advantage of increased oxygen and nitrite concentrations. Therefore, the delay in observed production of nitrate could simply represent the time required for nitrite-oxidizer colonies to establish themselves in these more favorable microenvironments. The fatty acid results were particularly surprising. What was particularly surprising is that the fatty acid ratio (cis-9 hexadecenoic acid/cis-11 hexadecenoic acid) did not change, even though nitrite oxidation was complete. What is interesting is that the lower ratio anticipated was finally observed almost one month after the dissolved oxygen perturbation. This suggests that the observed difference in the ratio is dependant upon oxygen concentration, that a different nitrite oxidizer population had established itself, or that the nitrite-oxidizer population had not yet stabilized, following the perturbation. F I S H images for a C S T R receiving identical feed to the reactor studied here are presented in Fig. 4.4. This reactor had been achieving complete nitrification at the same S R T for several months. The F I S H images in Fig . 4.4 indicate the nitrite-oxidizer colonies for this reactor are significantly larger than for the test reactor discussed in this manuscript; this supports the possibility that the nitrite-oxidizer population in the perturbed system had not yet stabilized to its new conditions, by December 13, 2003. 102 Conclusions .. The following conclusions have been drawn based upon this research: • Nitrite oxidizer populations were present in a reactor for a sustained period of time (>30 SRTs) even under a combination of free ammonia and low dissolved oxygen stress. • R N A slot blotting suggests that the ammonia oxidizer population did not change significantly between the low and high dissolved oxygen perturbation. This suggests that there were some organisms that were present but inactive and/or that there was a shift in ammonia oxidizer metabolism, from using nitrous acid as a terminal electron acceptor, to using dissolved oxygen. • Hydroxylamine was not detected in the bulk liquid, suggesting that it was being used as an electron donor for denitrification and that the cell enzyme levels had adapted as required. • The results support the possibility of substrate (nitrite and oxygen) competition between ammonia and nitrite oxidizers. 103 A. Ammonia-N, Nitrite-N and Nitrate-N versus Time Elapsed Time (days) C. Fraction of Primary Nitrogen Species following Perturbation 3 5 : 9 Elapsed Time (days) • Ammonia • N i t n l a JNi l ra ta a S v n m s a s « Nitrous oxidei Nso 190 RNA •a te Headspace Nitrous oxide (ppm) Elapsed Time (days) Scatter Plot of Nitrate vs Nitrous oxide 10 •a Jsoo i, )ooo § . 2500 iooo = 500 200 400 300 100 Nitrous oxide mass amission (mg-N/day) Ntspa 454 3NA Oata Fig. 4.1 - Time series plots of results for dissolved oxygen perturbat ion 104 R5VITSept 19-02 Red = ammonia oxidizer Green - nitnte oxidizer 5 0 p i x e l a = 1 5 . a microns Fig. 4.2a R5VIT Dec 13-02 Red = ammonia oxidizer G r een = nilnle oxidizer 5Q p i x e l 3 - 7 . 5 x i c r o o s Fis. 4.2b Figure 4.2. FISH images for test reactor before and after perturbation. Fig. 4.3a R5 Dec 13-02 Red - ammonia oxidizer Green - nitrite oxidizer Figure 4.3. FISH images of test reactor after dissolved oxygen perturbation. 106 RI S e p t 2 5 - 0 2 Red-ammonia o x i d i z e r s green=nitrite o x i d i z e r s 50 pijtola=9.4 microns r , . , , ri<z. 4.4 Figure 4.4. FISH images for a CSTR operating with identical feed, high DO, and 10-day SRT for several months. Table 4.1 - Synthetic feed composition for bench scale continuous stirred tank reactor (CSTR). Feed Component Concentration (mg/L) Sodium acetate ( N a C H 3 C O O H ) 250 mg/L Ammonium chloride (NH4CI) 556 mg/L Sodium bicarbonate (NaHC0 3 ) 1,936 mg/L Yeast extract 50 mg/L Potassium phosphate (K2HPO4) 56 mg/L Magnesium chloride (MgCl2-6H20) 37.5 mg/L Calcium chloride (CaCl 2 -2H 2 0) 24 mg/L Ferric chloride (FeCl 3 -6H 2 0) 1.9 mg/L Manganese sulphate ( M n S O y F t O ) 0.09 mg/L Sodium molybdate ( N a 2 M o 0 4 - 2 H 2 0 ) 0.008 mg/L Zinc sulphate (Zn S0 4 -7H 2 0) 0.375 mg/L Cobalt chloride (CoCl2-2H20) 0.001 mg/L 108 Bibliography Anthonisen, A.C., Loehr, R.C., Prakasam T.B.S., and Srinath, E.G. 1976. Inhibition of nitrification by ammonia and nitrous acid. J. Water Pollut. Control Fed. 48: 835-851. Cecen, F., and Ipek, S. 1998. Determination of the inhibition of ammonia-N and urea-N oxidations by the fed-batch reactor (FBR) technique. Water Sci. Technol. 38(1): 141-148. Eaton, A.D., Clesceri, L.S., Greenberg, A.E. (Editors) 1995. Standard methods for the examination of water and wastewater. 19th ed. American Public Health Association, Inc., New York. Hyungseok Yoo, Kyu-Hong Ahn, Hyung-Jib Lee, Kwang-Hwan Lee, Youn-Jung Kwak, and Kyung-Guen Song. 1999. Nitrogen removal from synthetic wastewater by simultaneous nitrification and denitrification (SND) via nitrite in an intermittently aerated reactor. Water Res. 33(1): 145-154. Schramm, A., De Beer, D., Gieseke, A., Amman, R. 2000. Microenvironments and distribution of nitrifying bacteria in a membrane-bound biofilm. Environ. Microbiol. 2(6): 680-686. Simm, R.A., Ramey, W.D., and Mavinic, D.S. 2004a. Preliminary evaluation of the use of fatty acid ratios for tracking the potential for nitrite accumulation in nitrifying reactors with low carbon to nitrogen ratio. J. Environ. Eng. Sci. 3: 31-40. 109 Simm, R . A . , Parkinson, P., Mavinic , D.S., Ramey, W . D . 2003b. Hydroxylamine analysis of wastewater samples via gas chromatography (submitted to Environmental Technology for publication July 2004). > Simm, R . A . , Mavinic , D.S. , and Ramey, W . D . 2004c. The role of hydroxylamine as an inhibitor of nitrite oxidizers in wastewater systems (submitted for publication in Journal of Environmental Engineering and Science, M a y 2004). Simm, R . A . , Ramey, W . D . , Mavinic , D.S. 2004d. A targeted study on possible free ammonia inhibition of Nitrospira (submitted to J. Environ. Eng. Sci . for publication May 2004). Simm, R . A . , Ramey, W . D . , and Mavinic , D.S. 2003e. Nitrifier population dynamics in a bench scale activated sludge reactor following an induced perturbation (accepted for publication with minor revisions in Journal of Environmental Engineering and Science June 2004). Stuven, R., Vollmer, M . , and Bock, E . 1992. The impact of organic matter on nitric oxide formation by Nitrosomonas europaea. Arch. Microbiol . 158: 439-443. Yang, L . , and Alleman, J .E. 1992. Investigation of batchwise nitrite build-up by an enriched nitrification culture. Water Sci . Technol. 26(5-6): 997-1005. 110 5.0 A Targeted Study on Possible Free Ammonia Inhibition oi Nitrospira4 Introduction and Background The work of Anthonisen et al. (1976) is considered by many to be the definitive work that established the inhibitory nature of free ammonia toward nitrite oxidizers. Anthonisen et al. (1976) hypothesized that, depending upon the operating conditions and initial loading rates, varying nitrite concentrations will persist without subsequent oxidation to nitrate, especially in solutions or wastewater having high organic nitrogen or ammonia-nitrogen concentrations. These authors postulated that inhibition was specifically related to free ammonia and free nitrous acid and that certain concentrations of free ammonia would inhibit nitrite but not ammonia oxidizers. In their original paper, they provide schematic representations of batch nitrification with and without inhibition, that are presented here as Fig. 5.1. Anthonisen et al. (1976) used the associated concepts to develop an operational chart indicating the various concentration ranges where free ammonia and nitrous acid would likely be inhibitory to ammonia and/or nitrite oxidizers. These authors reported free ammonia was likely to be inhibitory towards nitrite oxidizers in the range 0.1 to 1 mg-NJiyL (0.08 to 0.8 mg NH3-N/L) and to ammonia oxidizers in the range 10 to 150 mg-NH3/L (8.2 to 123.5 mg NH3-N/L). There has been a great deal of interest in free ammonia inhibition of nitrite-oxidizers since Anthonisen et al. (1976) published their seminal paper. Much of this interest has been driven by a desire to design nitrification/denitrification systems that selectively inhibit nitrite oxidation -thereby eliminating the formation of nitrate. Numerous authors (Voets et al. 1975; Turk and 4 A version of this chapter has been submitted for publication May 2004. Simm, R.A., Mavinic, D.S., Ramey, W.D. A targeted study on possible free ammonia inhibition of Nitrospira. J. Env. Eng. Sci. I l l Mavinic 1986, 1987, 1989a, 1989b; Balmelle 1992; Chen et al. 1991; Fdz-Polanco et al. 1996; Garrido et al. 1997; Hyungseok Yoo et al. 1999) have reported the capital and operational benefits of the process, referred to here as the nitrate shunt, to include: a 25% reduction in aeration requirements, a 40% reduction in external carbon addition for denitrification in low C : N wastes, a potential reduction in anoxic zone volume and a significant reduction in sludge production. Although free ammonia is still the consensus cause of nitrite oxidizer inhibition and therefore considered the key to process operation, there are reported discrepancies in reported inhibitory ranges, as well as the actual cause of observed nitrite oxidizer inhibition. For instance, Turk and Mavinic (1986, 1987, 1989a, 1989b) reported no significant nitrite-oxidizer inhibition for free ammonia concentrations below 10 mg N H 3 - N / L , whereas Mauret et al. (1996) concluded that free ammonia inhibits Nitrobacter in the range of 6.6 to 8.9 mg N H 3 - N / L . Several investigators have called into question the true cause of nitrite oxidizer inhibition. For example, Cecen and Ipek (1998) have suggested that the dissolved oxygen to free ammonia ratio, and not the free ammonia concentration itself, is of primary importance when attempting to induce nitrite accumulation. The work of others (Stuven et al. 1992; Yang et Alleman 1992; and Hyungseok Yoo et al. 1999) has suggested that hydroxylamine - an intermediate in the ammonia oxidation process - may be the true cause of nitrite oxidizer inhibition and therefore of nitrite accumulation. In addition, it is now believed that Nitrospira, and not Nitrobacter, is typically the dominant nitrite oxidizer in most environmental matrices including wastewater systems (Schramm et al. 112 1998; Juretschko et al. 1998; Daims et al. 2000; Bartosch et al. 2002). This is in spite of the fact that described species of Nitrospira grow significantly slower in pure culture than Nitrobacter (Juretschko et al. 1998). Interestingly, Gieseke et al. (2003) state that it is not known whether bacteria of the genus Nitrospira are inhibited by free ammonia. The work reported here was the direct result of the authors' inability to reproduce nitrite inhibition effects at concentrations reported as inhibitory in the engineering literature. The primary objective of this part of the research program was to elucidate whether or not free ammonia is truly inhibitory to nitrite oxidizers. A s indicated previously, a number of researchers have called into question the inhibitory nature of free ammonia. If free ammonia inhibition can be eliminated as the primary cause of nitrite accumulation, one can then focus on coincidental observations that may help explain the true cause of the observed phenomenon. This work was carried out with a combination of mixed and pure cultures. Methodology The initial work was done using biomass from a bench-scale, continuous stirred tank reactor (CSTR), being fed a synthetic wastewater mixture: This was followed up with inhibition studies carried out using static cultures of Nitrospira moscoviensis. Reactor System A 10.8-liter bench scale reactor was operated at a dissolved oxygen level of greater than 2 mg/L, ambient temperatures (21-23 °C), and a solids retention time of 5 days. The reactor was operated 113 in this manner in order to achieve complete nitrification. The reactor was designed with a gas tight headspace, to allow collection of headspace gas samples. Reactor headspace was vented through a gas collection bulb and beaker filled with water to allow development of a constant headspace pressure. The reactor was run as a C S T R , with secondary clarification having a two-hour settling time. The reactor was mixed with a 100-rpm motor. The influent feed to the reactor was 24 liters per day, giving a reactor hydraulic retention time (HRT) of approximately 11 hours. The C S T R was used for in situ batch tests; each test was carried out by stopping the influent feed and adding a combination of ammonium chloride and sodium bicarbonate to the reactor, in order to achieve the predetermined free ammonia concentration. The dissolved oxygen concentration was controlled manually at a preset level for the duration of the test. The complete reactor setup is illustrated elsewhere (Chapter 3 and Simm et al. 2004b). Synthetic Feed The bench scale reactor was fed a synthetic feed that simulated a relatively weak landfill leachate (part of ongoing studies at the University of British Columbia) with a low carbon to nitrogen ratio and provided a nitrifier-enriched culture. The feed components and their respective concentrations are presented in Table 5.1. Prewashed Biomass Tests A limited number of washed biomass tests were carried out in order to eliminate the potential confounding effect of heterotrophic denitrification upon the interpretation of results. In addition, there was some concern that higher free ammonia concentrations might compromise the mixed microbial population in the test reactors, resulting in long lead times between experiments while waiting for system recovery. 114 When carrying out the prewashed biomass tests, the mixed liquor was collected, centrifuged, and washed with distilled water. The prewashed biomass tests were carried out using a 1.5-liter batch test unit that allowed for gas collection. Preliminary tests, described below, were carried out using a 3-liter Erlenmyer flask. The washed biomass tests were conducted with synthetic feed devoid of carbon to which a predetermined amount of ammonia was added. The reactor was stirred with a magnetic stirrer and pH was controlled manually, using a dilute sodium hydroxide solution. Samples were collected using a syringe. Preliminary Batch Test A preliminary batch test was carried out with biomass from a reactor with a ten-day SRT and dissolved oxygen concentration of 2 to 3 mg/L. The synthetic feed recipe used was the one described earlier. Biomass characterization was not conducted on this biomass using molecular methods; however, the fatty acid profile for this mixed liquor was similar (>90% similarity) to that from the test unit described below. Given that this reactor received the same synthetic feed and the biomass had a similar fatty acid profile, Nitrospira-\\kz organisms were likely the dominant nitrite oxidizers in this reactor. The preliminary batch test was carried out using a 3-liter Erlenmeyer flask and magnetic stirrer arrangement. The test apparatus was previously described by Comeau (1989). The preliminary batch test was operated in an anaerobic/aerobic sequence. Anaerobic conditions were established via nitrogen sparging. The biomass was cycled through an anaerobic cycle to get a preliminary assessment of the role that anaerobic free ammonia stress might play on nitrite 115 accumulation. Turk and Mavinic (1986, 1987, 1989a, 1989b) have suggested that periodic exposure to free ammonia stress, under anaerobic conditions, might play a role in nitrite accumulation. This test sequence provided the added benefit o f providing an opportunity to determine whether or not hydroxylamine would be produced at the anaerobic/aerobic interface. Yang and Alleman (1992) have suggested hydroxylamine is the true inhibitor of nitrite oxidizers. The biomass used in this test was collected from the seed reactor, centrifuged, and washed twice with distilled water. The washed biomass was added to 2 liters of synthetic feed devoid o f carbon and ammonia. Subsequently, an ammonium bicarbonate spike was added to the mixed liquor, in order to achieve a target free ammonia concentration of between 10 and 20 mg-N/L. Chemical Analyses The analysis of ammonia, total Kjeldahl nitrogen (TKN) , nitrate, nitrite, total organic carbon, and th total dissolved solids was carried out as described in the 19 Edition of Standard Methods for the Examination of Water and Wastewaters (Eaton et al. 1995). The free ammonia and nitrous acid concentrations were estimated, using the relationships presented by Anthonisen et al. (1976). Hydroxylamine was measured using the G C method described in Chapter 2 (Simm et al. 2004a). Fatty Acid Methyl Esters (FAME) Mixed liquor samples were collected and subjected to fatty acid methyl ester analysis ( F A M E ) . The analysis procedure was described previously (Chapter 3 and Simm et al. 2004b). 116 N 2 0 The headspace of the reactor and test unit was fully enclosed thus allowing for the collection o f reactor off-gases. Off-gas samples were collected using a gas tight syringe and analyzed for nitrous oxide by gas chromatography, using a Hewlett-Packard 5880A equipped with an electron capture detector. In the G C analysis, nitrogen was used as the carrier gas (at 20 mL/min) with a column packing material of Haycep C. The injector, oven, and detector temperatures were 80 °C, 80 °C, and 250 °C, respectively. RNA Slot Blotting R N A samples were collected, centrifuged, and stored at -80 °C prior to extraction. R N A extractions were conducted using Trizol™ reagent and bead beating, according to the manufacturers protocols, using a Bio-Spec mini-bead beater. This procedure is based upon the R N A extraction method developed by Chomczynski and Sacchi (1987). Oligonucleotide probes were synthesized at the University of British Columbia Nucleic A c i d and Protein Synthesizing (NAPS) unit and unincorporated nucleotides were removed from the probes using an ammonia-butanol purification step. The lyophilized probes were solubilized in RNase-free water prior to dilution for use in probe labeling. A l l oligonucleotide probes were labeled to a specific activity of 10 8-to-10 9 C P M / p g with 3 2 P by using a 5' end labeling kit as supplied by Amersham Biosciences ( R P N 1509). Slot blot hybridizations were conducted in general accordance with the method outlined by Stahl et al. (1988), Raskin et al. (1994) and Mobarry et al. (1996). The molecular probes used for this study included Nso 190 (5'-CGA TCC CCT GCT TTT TCT CC-117 3') targeting all characterized ammonia-oxidizers in the P subdivision of the Proteobacteria (purple bacteria), Nb ,1000 (5'-TGC GAC CGG TCA TGG-3'). targeting Nitrobacter, and Ntspa-454 (5'- TCC ATC TTC CCT CCC GAA AA -3') targeting Nitrospira moscoviensis and related Nitrospira-like organisms. Mobarry et al. (1996) previously, described the Nb 1000 and Nso 190 probes. The Ntspa-454 probe for Nitrospira has been previously described by Hovanec et al. (1998). Pure Cul ture Strains Pure cultures of Nitrospira moscoviensis (DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen Gmb H) strain DSM 10035) were kindly provided by Dr. Eva Spieck of the University of Hamburg. Pure Cul ture Feed The pure culture cells were grown in DSMZ mineral medium 756d. The mineral medium has the following composition: 899 mL distilled water, 0.5 g NaN0 2; 1 mL trace element solution; and 100 mL of stock solution. The medium pH was adjusted to 8.6 with NaOH. The media was autoclaved following pH adjustment and allowed to stand for two to three days to allow the pH to adjust itself to pH 7.4-7.6. The trace element solution had the following composition: 33.8 mg MnS0 4«H 20; 49.4 H 3 B0 3 ; 43.1 mg ZnS0 4»7H 20; 37.1 mg (NH 4 ) 6 Mo 7 0 2 4 ; 97.3 mg FeS04*7H20; 25 mg CuS0 4»5H 20; and 1000 mL distilled water. The stock solution consisted of 0.07 g CaC0 3; 5 g NaCl; 0.5 g MgS0 4»7H 20, 1.5 g KH 2 P0 4 ; and 1000 ml distilled water. 118 Cultures grown under free ammonia stress were amended with the appropriate volume of a filter-sterilized solution of ammonium chloride. The salt controls were prepared by adding the appropriate weight of sodium chloride to the growth media, prior to autoclaving. The heterotrophic Nitrospira media included the same ingredients as D S M Z 756, as well as 1.5 g yeast extract, 1.5 g peptone, and 0.55 g sodium pyruvate. These are the carbon sources used for the heterotrophic Nitrobacter media ( D S M Z Medium 756). Pure Culture Experiments The Nitrospira moscoviensis strain was cultured in eleven separate flasks under one of five different culture conditions. Four of the eleven were grown under free ammonia stress. The objective of this experiment was to determine i f there was any significant difference in nitrite oxidation rate between Nitrospira cultures grown in the presence and absence of free ammonia stress. Salt controls were used to eliminate possible salt effects associated with ammonium chloride addition. The treatments applied to each batch culture are summarized in Table 5.2. The eleven cultures were grown in 2-liter Erlenmyer flasks. Each flask contained 1000 m L of sterile media and, with the exception of flask 11 (40 m L inoculum), was inoculated with 100 m L stationary phase Nitrospira moscoviensis culture from the same inoculum source. A l l o f the cultures were grown as static cultures at 37 °C. The flasks were wrapped in aluminum foil to eliminate possible inhibition from U V light. 119 Protein Assay Protein analyses were carried out using the bicinchoninic acid ( B C A ) assay described by Smith et al. (1985). This sensitive method was selected due to the low protein concentrations of the pure cultures. Ten-milliliter triplicate samples were aseptically taken from each flask and filtered using a 0.2 micron sterile filter. The filter concentrated cells were suspended using a sterile rubber policeman and sterile distilled deionized water then transferred to a 5 m L reaction vessel. The filter concentrated cells were diluted to 1 m L volume with distilled deionized water and then vortexed. One milliliter of the B C A reagents were added to the cell concentrate, the mixture was vortexed, and then incubated at 60 °C for 30 minutes. Standards were made up using crystallized Bovine Serum Albumein (BSA) . Several attempts were made to use bacterial laden filters directly in the reaction/incubation; however, material in the filters themselves provided high background interference. Results Preliminary Batch Test The initial ammonia concentration in the preliminary batch test was approximately 300 mg-N/L. The test unit p H increased during the test, likely as the result o f carbon dioxide stripping. N o appreciable nitrite accumulation was observed during the test, and apparent nitrite-oxidizer inhibition was not observed, even though the measured free ammonia concentration was between 10 and 18 mg-N/L. It should be noted that no measurable quantities (detection limit <0.1 mg/L) of hydroxylamine could be recovered during the test. The possible cause(s) of the observed 120 result could have included the fact that free ammonia is not inhibitory, as widely believed, or the nitrite-oxidizer population was large enough to overcome the inhibition. CSTR Characterization Several tests were carried out using biomass from a seed reactor being operated in the laboratory and used in the tests described below. The reactor, referred to here as R2, was operated at a five day S R T with a dissolved oxygen concentration of between 2.0 and 3.0 mg/L. The biomass from this reactor was used for several exploratory tests between September, 2002 and Apr i l , 2003. The performance and population dynamics of this reactor were characterized, to allow performance comparisons with other systems. The influent ammonia concentration for R2 was 151 mg/L +/- 20 mg/L (two standard deviations). The reactor provided essentially complete ammonia oxidation to nitrate, with nitrate making up 99 % of the measured oxidized nitrogen species, on average. The average mixed liquor suspended solids concentration for this reactor was 925 mg/L +/- 250 mg/L (two standard deviations). The nitrite oxidizer population in the reactor was characterized via R N A slot blotting and F A M E . R N A slot blotting, using the N b 1000 and Ntspa 454 probes, indicated that Nitrospira-\\ke organisms were the dominant nitrite oxidizers in the test reactor. In fact, on average, these organisms accounted for approximately 11.3 % +/- 8% (two standard deviations) of the total R N A in this community. This estimate was also supported by the F A M E analysis based upon the large proportion of cis(\\) hexadecenoic acid, relative to cis(9) hexadecenoic acid. 121 In situ Batch Tests September 6, 2002 batch test (high dissolved oxygen) The first in situ batch test was carried out on September 6, 2002. An ammonia spike of approximately 40 mg-N/L was added to the reactor after terminating the influent feed. The reactor pH was adjusted to an initial value of approximately 8.4, in order to obtain the desired free ammonia stress and the pH was allowed to drop following the spike. The aeration rate was adjusted until a constant dissolved oxygen concentration of approximately 3 mg/L (aeration rate of 1.6 liters/min.) was obtained; the test was terminated when ammonia all but disappeared and the dissolved oxygen concentration began to climb. The results from this test are summarized as Fig. 5.2. An inhibitory, free-ammonia stress was maintained in the reactor for approximately the first eighty minutes of the test (NH3-N 0.1 to 3.5 mg/L). The combination of nitrate, nitrite, and emitted nitrous oxide accounted for approximately 82% of the oxidized ammonia. The measured ammonia, nitrite, and nitrate accumulated during the test are summarized as Fig. 5.2H. The nitrate accumulated during the test is defined as the measured nitrate concentration minus the nitrate concentration at the start of the test. The delay in nitrite oxidation and characteristic accumulation in nitrite that, according to Anthonisen et al. (1976), is characteristic of free ammonia inhibition, was not observed for this test although nitrite accounted for approximately thirty-six percent of the oxidized nitrogen species at its peak concentration. The nitrite concentration continued to build as the test progressed and peaked at the point where free 122 arnmonia disappeared. The headspace concentration of emitted nitrous oxide increased as the test progressed. No measurable concentrations of hydroxylamine were obtained during this test. September 11,2002 batch test (low dissolved oxygen) A second in situ batch test was carried out on September 11, 2002. The test conditions were the same as those for the September 6, 2002 test; however, the dissolved oxygen concentration was controlled at less than 0.5 mg/L. This was undertaken to ascertain whether a lower dissolved oxygen concentration could induce nitrite accumulation, as indicated by Cecen and Ipek (1998), and whether or not a measurable hydroxylamine concentration could be produced. The results for this test are summarized as Fig. 5.3. The ammonia oxidation rate was severely impaired by the low dissolved oxygen concentration. The measured ammonia oxidation rate for the low dissolved oxygen test was 2.6 mg/L-hr, versus 16 mg/L-hr for the high dissolved oxygen test. This test, unlike other tests conducted in the laboratory, produced a measurable concentration of hydroxylamine, the peak concentration of which coincided with a drop in the reactor ORP. The difficulty with this test is that there was a significant variation in the measured nitrate-nitrogen concentration during the test duration. Given the low ammonia oxidation rate, and the fact that the measured nitrite and nitrous oxide concentrations account for approximately 90% of the oxidized ammonia, it is unlikely that any nitrite oxidation actually took place. The fact that ammonia oxidation was limited, and that heterotrophic denitrification may have taken place, makes it difficult to rule out oxygen limitation as the cause of observed nitrite accumulation. 123 Washed Biomass Tests January 23, 2003 batch test (low dissolved oxygen) A low dissolved oxygen batch test was carried out using washed biomass from the test reactor (R2). The test unit p H was controlled between 8.3 and 8.5 for the first 280 minutes, using a syringe and dilute sodium hydroxide solution. The estimated free ammonia concentration was between 6 and 8.5 mg/L during this time period. The dissolved oxygen concentration was maintained below 0.5 mg/L for the duration of the test. The test results are summarized in Fig . 5.4. The ammonia, nitrite, and nitrate results are summarized in Fig. 5.5. This figure indicates that nitrite oxidation was not inhibited, even at free ammonia concentrations far in excess of the inhibitory range reported by Anthonisen et al. (1976). It should be noted, however, that the nitrite concentration began to decline once p H control was abandoned. There was a low concentration of hydroxylamine measured during this test. The hydroxylamine concentration increased with increasing nitrite and nitrous oxide concentrations, declining when p H control was abandoned. The observed concentration of hydroxylamine did not appear to negatively affect nitrite oxidation. Although the correlation between free ammonia and hydroxylamine was poor (R =0.48), there appeared to be an inverse relationship between the two parameters. The removal of two points that appeared to be outliers improved the correlation coefficient significantly (R 2=0.74). The estimated ammonia oxidation rate during the test was 4.3 mg/L-hr. 124 February 26, 2003 washed biomass test (high dissolved oxygen). A second washed biomass test was carried out using mixed liquor from the test reactor. The test conditions were essentially the same; with the exception that the higher dissolved oxygen concentration of 3 mg/L and higher free ammonia concentrations were used. The estimated free ammonia concentration during the test period was between 3.8 and 14:8 mg/L. The test results are summarized as Fig . 5.6. The ammonia, nitrate, and nitrite results are summarized in Fig . 5.6H. Again, free ammonia did not appear to be inhibitory to nitrite oxidizers at the free ammonia concentrations used. It should be noted that the nitrite concentration peaked at approximately the same time as the headspace nitrous oxide concentration. N o hydroxylamine was measured in the reactor fluid during the test. The estimated ammonia oxidation rate was 19.5 mg/L-hr. Pure Culture Test The results obtained from the preceding tests, and others, suggested that free ammonia might not be inhibitory as suggested in the literature; these observed phenomena might be due to dissolved oxygen limitation, the difference in kinetic rates between ammonia and nitrite oxidizers, a combination of the first two factors, or some yet undetermined cause. Hunik et al. (1993) conducted chemostat studies using pure cultures of Nitrobacter and concluded that free ammonia was not inhibitory to these nitrite oxidizers. The work conducted for this research program, and others, has indicated Nitrospira-Wke, organisms are the predominant nitrite oxidizers in wastewater systems. To the knowledge of these authors, no one had yet studied Nitrospira 125 inhibition in pure cultures. The overwhelming majority of literature dealing with free ammonia inhibition of nitrite oxidizers comes from the environmental. engineering field, where mixed cultures and municipal sewage mixed liquor seed are used almost exclusively. One of the difficulties associated with this approach is that it does not eliminate the possibilities that observed nitrite accumulation is simply the result of partial heterotrophic denitrification, localized substrate competition (oxygen, nitrite, or carbon dioxide) between ammonia and nitrite oxidizers, or production of inhibitory compounds (e.g. hydroxylamine) by ammonia oxidizers or other organisms. For these reasons, it was decided to confirm the mixed culture results with pure cultures of the nitrite oxidizer Nitrospira moscoviensis. A s indicated in the methodology section, ten 2-liter Erlenmyer flasks were inoculated with an equal volume of stationary phase Nitrospira moscoviensis culture (100 mLs + 1 liter). Two, 2-liter flasks were inoculated with a smaller inoculum (40 m L + 1 liter). The experimental results are summarized in Table 5.3 (the results for the culture grown on the heterotrophic media were not included in Table 5.3, since the media used appeared to interfere with the protein assay) and Fig. 5.7. The measured nitrate-nitrogen concentration with time for each experimental treatment is also presented in Fig . 5.7. The results suggest that there was no significant difference between treatments, with the exception of the cultures that received the mixotrophic Nitrobacter media and smaller inoculum. The results with the mixotrophic Nitrobacter media suggest organic matter is inhibitory to Nitrospira moscoviensis. This is consistent with the findings of Ehrich et al. (1995) who reported that concentrations as low as 0.75 grams organic matter/L or higher, inhibited growth o f Nitrospira moscoviensis cells. 126 These results corroborate the observations from the mixed culture reactor experiments, suggesting that free ammonia is not inhibitory to Nitrospira in a concentration range commonly reported as being inhibitory. Although there was a significant amount of variation in the protein measurements, the results indicate that there was no significant difference in protein concentrations between the free ammonia treatments and their respective salt controls. The protein analysis for Salt Control 2-2 indicated the initial and final concentrations were the same. The authors believe this to be a sampling anomaly and cannot offer any explanation for the discrepancy. Discussion The results presented here do not support free ammonia inhibition of nitrite oxidizers as reported in the literature. A s indicated above, others have also suggested that free ammonia inhibition may not be the primary cause of nitrite accumulation. The lack of apparent free ammonia inhibition may be the result of population acclimation, either due to the large number of nitrite-oxidizers, strain variability, or the presence of a common mutation that confers resistance. One of the primary strengths of the work of Anthonisen et al. (1976) is that the data encompass a broad range of microbial environments, including several agricultural wastewaters and soil. In virtually every study on free ammonia inhibition reported in the literature, nitrite-oxidizers appear to acclimate to ever-higher concentrations of free ammonia. If mutation was the cause of the observed acclimation response, one would expect, based upon literature evidence, that such a mutation would be a common one. The fact that Anthonisen et al. (1976) were able to 127 demonstrate what appeared to be free ammonia inhibition for a broad range of environments does not support the existence of such a common mutation nor strain variability. The results of the pure culture study reported on here rule out the possibility that the number of organisms is at issue. The initial Nitrospira moscoviensis concentration of approximately 2-pg protein/mL (average of ten flasks) is equivalent to an R N A concentration of 750 ng/mL, assuming protein and R N A make up 55 and 20.5 percent of the cell dry weight, respectively. These percentages are based upon the commonly reported chemical composition of prokaryotic cells (Madigan et al., 1997). This concentration of Nitrospira R N A is of the same order of magnitude as the measured concentration for a C S T R reactor used in this research, whereby apparent free ammonia inhibition was observed (data presented in Chapter 6). The reactor in question was operated with an influent ammonia concentration of approximately 40 mg-N/L, prior to a step increase in ammonia to 150 mg-N/L (maximum free ammonia concentration of approximately 9 mg-N/L) , but with no change in influent carbon concentration (the T O C : T K N ratio declined from approximately 3:1 prior to the step to 1:1 following the step). This step resulted in an immediate drop in nitrate-nitrogen concentration in the reactor, as well as a drop in Ntspa 454 R N A concentration that was consistent with the wasting rate. The drop in nitrate-nitrogen concentration was coincidental with an increase in headspace nitrous oxide concentration and a rapid increase in ammonia oxidation rate (Simm et al. 2004c). The majority of the decline in nitrite oxidizer activity, relative to the original activity prior to the perturbation, could be accounted for as emitted nitrous oxide. The evidence from that particular study suggested competition for nitrite (specifically nitrous acid) and potentially oxygen, between ammonia oxidizers and Nitrospira-Mke, organisms, was the 128 primary cause of reduced nitrite oxidizer activity. The reduction in nitrite oxidizer activity occurred in spite of a substantial increase in reactor nitrite concentration, suggesting that free ammonia is unlikely to be a competitive inhibitor. Nitrite oxidation recovered once free ammonia levels in this reactor had essentially dropped to zero and this was coincidental with a significant decline in headspace nitrous oxide concentrations. One of the primary differences between the step study discussed above, and the pure culture study presented in this chapter, is the presence of ammonia oxidizers that provide competition for nitrous acid, oxygen, and potentially carbon dioxide. The primary differences between the aforementioned step study and test reactor used for the batch tests was the initial number of nitrite oxidizers and the carbon to nitrogen ratio of the initial feed. Much of the work conducted by Anthonisen et al. (1976) was carried out using agricultural wastes (e.g. poultry manure, mink manure, and cattle manure) that one would expect to have higher influent ammonia and carbon concentrations than the synthetic waste used for this research. One would also expect ammonia and nitrite oxidizer colonies, located within mixed liquor floe particles taken from such systems, to be located deeper inside the floe, due to competition for oxygen and the potential for carbon inhibition. A combination of a sudden increase in ammonia oxidizer substrate, diffusional resistance, stoichiometric reduction in nitrous acid concentration at a higher p H , and ammonia oxidizer denitrification of nitrous acid to nitrous oxide, could potentially explain the delay in the appearance in nitrate and apparent nitrite oxidizer inhibition reported by these investigators. The same spatial relationship between heterotrophs and nitrifiers would also be expected in typical municipal mixed liquor that, coupled with relatively low numbers, would support apparent free ammonia inhibition in batch systems. 129 The importance of the initial seed can be demonstrated by one of the experiments conducted for the present research program, whereby an S B R was used to mimic the conditions set up by Turk and Mavinic (1986, 1987, 1989a, 1989b) to induce nitrite accumulation (Chapter 7 and Simm et al., 2004d). These investigators were able to induce aerobic nitrite accumulation via periodic exposure of a mixed microbial population to inhibitory levels of free ammonia in the anoxic zone of a predenitrification, plug-flow reactor. The ammonia level in the anoxic zone was controlled via p H adjustment. We were able to duplicate these results using a similar strategy. R N A slot blotting suggested that the initial perturbation, which impacted the nitrite oxidizers most negatively, and not free ammonia, was the most likely cause of observed nitrite accumulation discussed in Chapter 7 and Simm et al (2004d). Nitrous oxide emissions were coincidental with the accumulation, suggesting that nitrite reduction by ammonia oxidizers likely contributed to the time required for the nitrite oxidizer acclimation to manifest itself. Therefore, in essence, the control strategy biased the initial seed population. The mixed liquor acclimated to the high free ammonia levels (average 10 mg N H 3 - N / L ) within one SRT. A review of Turk's PhD work (1986) shows that, on several occasions when the reactor p H was increased to boost free ammonia levels, ammonia oxidation itself was typically lost for a couple of days. This was typically followed by the recovery in ammonia oxidation and nitrite accumulation. Villeverde et al. (2000) also previously demonstrated the importance of culture history, upon observed nitrite accumulation. 130 Conclusions The following conclusions have been drawn based upon this research. • Nitrospira-like organisms were the dominant nitrite-oxidizer population in a bench scale C S T R , operated at low carbon to nitrogen ratios. • In-situ and washed biomass tests conducted with the mixed microbial population from a bench scale reactor did not show classical free ammonia inhibition of nitrite-oxidizers, at free ammonia concentrations as high as 14.8 mg N H 3 - N / L . • The combination of low dissolved oxygen and high free ammonia resulted in the recovery of a measurable concentration of hydroxylamine in batch tests. This was coincidental with a significant reduction in ammonia oxidation rate. • Free ammonia concentrations as high as 10 mg N H 3 - N / L were not toxic or inhibitory to pure cultures of Nitrospira moscoviensis grown in batch cultures. 131 Nitrogen transformations during non inhibited nitrification. mg-N/L N 0 3 N O , Timf Nitrogen transformations during inhibited nitrification. mg-N/L Time Figure 5.1. Batch Nitrification with and without free ammonia inhibition. A. Dissolved Oxygen J 7 .00 £ 6.00 s 5 .00 I? 4 .00 D 3.00 2 2 .00 > o 1.00 »i <n £ 0.00 5 0 100 Elapsed Time (minutes) [ c T NH 4 -N 45 -i 40 i 35 % S 30 V. y = -0.2662X + 32.749 25 * RJ = 0.9335 z 20 1 .2 15 "5 10 1 5 E o .5 20 4 0 60 80 (t)0 120 140 1 0 Elapsed Time (minutes) 40 60 80 100 120 Elapsed Time (minutes) Headspace Nitrous oxide 50 100 Elapsed Time (minutes) PH 40 60 80 100 120 140 Elapsed Time (minutes) 40 60 80 100 120 Elapsed Time (minutes) NO,-N 40 60 80 100 120 140 Elapsed Time (minutes) Major Nitrogen Species versus Time 20 40 60 80 100 120 Elapsed Time (minutes) - A m m o n i a - N —01—Nitrite-N •Accumulated Nitrate-N Figure 5.2 - Summary of September 6, 2002 batch test 133 Dissolved Oxygen o [c~| 45 40 •a 35 B, 30 z 25 CQ 20 'S o 15 S 10 B 5 < 200 300 4(H) Elapsed Time (minutes) NH 4 -N y = •0.0434.x + 34.099 HH) 200 300 Elapsed Time (minutes) N0 3 -N UK) 200 300 400 Elapsed Time (minutes) ID ORP 140 12(1 100 \ a a O 80 60 41) 20 ( 100 200 300 400 500 Elapsed Time (minutes) Headspace Nitrous Oxide 100 200 300 400 Elapsed Time (minutes) 3.5 3.0 £, 211 Z 1.5 3? 1.0 z X o £ • z pH 100 200 300 400 E l a p s e d T i m e (minutes) NHi-N KH) 2(H) 3(H) 4(H) 5(H) Elapsed Time (minutes) N O r N 7 1> 5 E 4 z 1 1 9 2 z 0 100 200 300 400 5(H) Elapsed Time (minutes) Hydroxylamine Elapsed Time (minutes) Liquid Nitrous Oxide „ 4 O 3 s ^ •} < / — d Nitroi (mg/L / d Nitroi (mg/L / d Nitroi (mg/L / Sr 0 5 / J 0 KH) 2(H) 3(H) 4(H) 5(H) Elapsed Time (minutes) Figure S.3 - Summary of September 11, 2002 Batch Test Figure 5.4 - Summary of January 23. 2003 washed biomass test 135 c. G . N O , - N 50 100 150 200 250 Elapsed Time (minutes) 300 NO3-N 50 100 150 200 250 Elapsed Time (minutes) 300 %Nitrite vs Time 45 40 35 — 30 25 ~~ 20 15 *• 0 50 100 150 200 250 Elapsed Time (minutes) 300 Nitrous Acid 0.000160 0.000140 0.000120 ^ 0.000100 50 100 150 200 250 Elapsed Time (minutes) N H . - N 50 100 150 200 250 Elapsed Time (minutes) N H 3 - N 50 100 150 200 250 Elapsed Time (minutes) F. Headspace Nitrous Oxide Elapsed Time (minutes) Nitrogen Species versus time 50 100 150 200 250 Elapsed Time (minutes) - A m m o n i a - N -N i t r i t e -N • -Ni t ra te-N Figure 5.6 - February 26, 2003 washed biomass test 137 140.0 M a y 7 -03 M a y 9 - 0 3 M a y 1 2 - 0 3 M a y 1 5 - 0 3 • C o n t r o l 1 • C o n t r o l 2 • F A - 1 m g / L ( 1 ) B F A = 1 m g / L (2) • Sa l t C o n t r o l 1-1 B Sa l t C o n t r o l 1 2 • F A - 1 0 m g / L (1) El F A = 1 0 mg/L (2) • S a i l C o n t r o l 2 1 • Sa l t C o n t r o l 2-2 • I l e l e i t r oph i c C o n t r o l F igure 5.7 - T ime ser ies of batch ni t rate-ni t rogen concent ra t ion for pure cu l ture trial OO Table 5.1. Feed components for continuous stirred tank reactor (CSTR) system. Feed Component Concentration Sodium acetate ( N a C H 3 C O O H ) 250 mg/L Ammonium chloride (NH4CI) 556 mg/L Sodium bicarbonate ( N a H C 0 3 ) 1,936 mg/L Yeast extract 50 mg/L Potassium phosphate ( K 2 H P 0 4 X 56 mg/L Magnesium chloride ( M g C l 2 - 6 H 2 0 ) 37.5 mg/L Calcium chloride ( C a C l 2 - 2 H 2 0 ) 24 mg/L Ferric chloride (FeCl 3 -6H 2 0) 1.9 mg/L Manganese sulphate ( M n S 0 4 H 2 0 ) 0.09 mg/L Sodium molybdate ( N a 2 M o 0 4 - 2 H 2 0 ) 0.008 mg/L Zinc sulphate (Zn S 0 4 - 7 H 2 0 ) 0.375 mg/L Cobalt chloride ( C o C l 2 - 2 H 2 0 ) 0.001 mg/L 139 Table 5.2. Summary of treatments applied to pure cultures of Nitrospira moscoviensis in batch experiments. Flask Number Treatment Designation Notes 1 Control-1 Standard Nitrospira mineral media only. 2 Control-2 Standard Nitrospira mineral media only. 3 Free Ammonia lmg/L (1) Standard Nitrospira media amended with sterile ammonium chloride to provide a free ammonia concentration of 1 mg/L. 4 Free Ammonia 1 mg/L (2) Standard Nitrospira media amended with sterile ammonium chloride to provide a free ammonia concentration of 1 mg/L. 5 Salt Control 1-1 Standard Nitrospira media amended with sterile NaCl having equal molarity to sterile ammonium chloride added to flasks 3 & 4. 6 Salt Control 1-2 Standard Nitrospira media amended with sterile NaCl having equal molarity to sterile ammonium chloride added to flasks 3 & 4. 7 Free Ammonia 10 mg/L (1) Standard Nitrospira media amended with sterile ammonium chloride to provide a free ammonia concentration of 10 mg/L. 8 Free Ammonia 10 mg/L (2) Standard Nitrospira media amended with sterile ammonium chloride to provide a free ammonia concentration of 10 mg/L. 9 Salt Control 2-1 Standard Nitrospira media amended with sterile NaCl having equal molarity to sterile ammonium chloride added to flasks 7 & 8. 10 Salt Control 2-2 Standard Nitrospira media amended with sterile NaCl having equal molarity to sterile ammonium chloride added to flasks 7 & 8. 11 Heterotrophic Media Standard Nitrospira media amended with carbon ingredients from mixotrophic Nitrobacter media. 140 Table 5.3. Data summary for pure culture experiment. Flask N 0 3 (mg-N/L) Est. Protein (|xg/mL) Day 1 Day 3 Day 9 D a y l Day 9 Control-1 16.4 49.2 109.3 0.6 2.6 Control-2 15.5 45.9 113.9 1.6 4.0 Free Ammonia lmg/L(l) 16.3 48.7 111 1.9 2.8 Free Ammonia 1 mg/L (2) 16.5 45.5 116.2 1.9 4.5 Salt Control 1-1 16.3 51.3 113.5 1.9 3.0 Salt Control 1-2 16.2 46.7 111.4 2.1 3.5 Free Ammonia 10 mg/L(l) 17.3 49 112 2.6 3.5 Free Ammonia 10 mg/L (2) • .16.3 . 48 112.8 2.0 3.8 Salt Control 2-1 16.8 54.1 110.7 2.3 3.3 Salt Control 2-2 16.7 49.9 110.7 3.6 3.6 141 Bibliography Anthonisen, A.C., Loehr, R.C., Prakasam T.B.S., and Srinath, E.G. 1976. Inhibition of nitrification by ammonia and nitrous acid. J. Water Pollut. Control Fed. 48: 835-851. Balmelle, B., Nguyen, K.M. , Capdeville, B., Cornier, J.C, and Deguin, A. 1992. Study of factors controlling nitrite build-up in biological processes for water nitrification. Wat. Sci. Technol. 26(5-6): 1017- 1025. Bartosch, S., Hartwig, C , Spieck, E., and Bock, E. 2002. Immunological Detection of Nitrospira-like bacteria in various soils. Microb. Ecol. 43: 26-33. Cecen, F., and Ipek, S. 1998. Determination of the inhibition of ammonia-N and urea-N oxidations by the fed-batch reactor (FBR) technique. Wat. Sci. Technol. 38(1): 141-148. Chen, S.K., Juaw, C.K., Cheng, S.S. 1991. Nitrification and denitrification of high strength ammonium and nitrite wastewater with biofilm reactors, Wat. Sci. Technol. 23(7-9): 1417-1425. Chomczynski, P., and Sacchi, N . 1987. Single-step method for RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162: 156-159. Comeau, Y. 1989. The role of carbon storage in biological phosphate removal from wastewater, PhD thesis, University of British Columbia, Vancouver. 142 Daims, H., Neilsen, P.H., Juretschko, S., and Wagner, M . 2000. Novel Nitrospira-like bacteria as dominant nitrite-oxidizers in biofilms from wastewater treatment plants: diversity and in situ physiology. Wat. Sci. Technol. 41: 85-90. Eaton, A.D., Clesceri, L.S., Greenberg, A.E. (Editors) 1995. Standard methods for the examination of water and wastewater (19th Edition)., American Public Health Association, Inc., New York. Ehrich, S., Behrens, D., Lebedeva, E., Ludwig, W., and Bock, E. 1995. A new obligately chemolithoautotrophic, nitrite-oxidizing bacterium, Nitrospira moscoviensis sp., nov. and its phylogenetic relationship. Arch. Microbiol. 164: 16-23. Fdz-Polanco, F., Villaverde, S., and Garcia, P.A. 1996. Nitrite accumulation in submerged biofilters - combined effects. Wat. Sci. Technol. 34(3-4): 371-378. Garrido, J.M., van Bethum, W.A.J., van Loosdrecht, M.C.M., and Heijnen, J.J. 1997. Influence of dissolved oxygen concentration on nitrite accumulation in a biofilm airlift suspension reactor. Biotechnol. Bioeng. 53: 168-178. Gieseke, A., Bjerrum, L., Wagner, M. , and Amann R. 2003. Structure and activity of multiple nitrifying bacterial populations co-existing in a biofilm. Environ. Microbiol. 5(5): 355-369. Hovanec, T.A., Taylor, L.T., Blakis, A., and DeLong, E.F. 1998. Nitrospira-\\ke bacteria associated with nitrite oxidation in freshwater aquaria. Appl. Environ. Microbiol. 64: 258-264. 143 Hunik, J.H., Meijer, H.J.G., and Tramper, J. 1993. Kinetics of Nitrobacter agilis at extreme substrate, product and salt concentrations. Appl. Microbiol. Biotechnol. 40:442-448. Hyungseok Yoo, Kyu-Hong Ahn, Hyung-Jib Lee, Kwang-Hwan Lee, Youn-Jung Kwak, and Kyung-Guen Song. 1999. Nitrogen removal from synthetic wastewater by simultaneous nitrification and denitrification (SND) via nitrite in an intermittently aerated reactor. Water Res. 33(1): 145-154. Juretschko, S., Timmermann, G., Schmid, M. , Schleifer, K.H., Pommerening-Roser, A., Koops, H.P., and Wagner, M . 1998. Combined molecular and conventional analyses of nitrifying bacterium in activated sludge: Nitrosococcus mobilis and Nitrospira-like bacteria as dominant populations. Appl. Environ. Microbiol. 64: 3042-3051. Madigan, M.T., Martinko, J.M., and Parker, J. 1997. Brock Biology of Microorganisms (8th Edition). Prentice Hall, Upper Saddle River, New Jersey. Mauret, M. , Paul, E., Puech-Costes, E., Maurette, M.T., and Baptiste, P. 1996. Application of experimental research methodology to the study of nitrification in mixed culture. Wat. Sci. Technol. 34(1-2): 245-252. Mobarry, B.K., Wagner, M. , Urbain, V., Rittmann, B.E., and Stahl, D.A. 1996. Phylogenetic probes for analyzing abundance and spatial organization of nitrifying bacteria. Appl. Environ. Microbiol. 62:2156-2162. 144 Raskin, L., Stromley, J.M.; Rittmahn, B. E., and Stahl, D. 1994: Group-specific 16S rRNA hybridization probes to describe natural communities of methanogens. Appl. Environ. Microbiol. 6 0 : 1232-1240. Schramm, A., De Beer, D., Wagner, M. , Amman, R. 1998. Identification and activities in situ of Nitrospira and Nitrospira spp. as dominant populations in a nitrifying fluidized be reactor. Appl. Environ. Microbiol. 6 4 : 3480-3485. Simm, R.A., Mavinic, D.S., Ramey, W.D., Parkinson, P. 2004a. Hydroxylamine analysis of wastewater samples via gas chromatography, (submitted to Environmental Technology for publication July 2004). Simm, R.A., Ramey, W.D., and Mavinic, D.S. 2004b. Preliminary evaluation of the use of fatty acid ratios for tracking nitrite accumulation in nitrifying reactors with low carbon to nitrogen ratio. J. Environ. Eng. and Sci. 3: 31-40. Simm, R.A., Ramey, W.D., and Mavinic, D.S. 2004c. Nitrifier population dynamics in a bench scale activated sludge reactor following an induced perturbation (accepted for publication with minor revisions in Journal of Environmental Engineering and Science June 2004). Simm, R.A., Ramey, W.D., and Mavinic, D.S. 2004d. Mechanisms responsible for apparent free ammonia inhibition in a sequencing batch reactor (submitted to the ASCE Journal of Environmental Engineering July 2004). 145 Smith, P .K . , Krohn, R.I. , Hermanson, G.T., Mal l ia , A . K . , Gartner, F . H . , Provenzano, M . D . , Fujimoto, E . K . , Goeke, N . M . , Olson, B.J . , and Klenk, D . C . 1985. Measurement of protein using bicinchoninic acid. Analytical Biochemistry, 1 5 0 : 76-85. Stahl, D . A . , Flesher, B . , Mansfield, H.R. , and Montgomery, L . 1988. Use of phylogenetically based hybridization probes for the studies of ruminal microbial ecology. App l . Environ. Microbiol . 5 4 : 1079-1084. Stuven, R., Vollmer, M . , and Bock, E . 1992. The impact of organic matter on nitric oxide formation by Nitrosomonas europaea. Arch. Microbiol . 1 5 8 : 439-443. Turk, O. 1986. The feasibility of a shortened pathway for nitrogen removal from highly nitrogenous wastes, PhD Thesis, University of British Columbia, Vancouver. Turk, O., and Mavinic , D.S . 1986. Preliminary assessment of a shortcut in nitrogen removal from wastewater. Can. J. C i v i l Eng. 1 3 : 600-605. Turk, O., and Mavinic , D.S. 1987. Selective inhibition: a novel concept for removing nitrogen from highly nitrogenous wastes. Env. Technol. Lett., 8 : 419. Turk, O., and Mavinic , D.S . 1989a. Stability of nitrite build-up in an activated sludge system. J. Water Pollut. Control Fed., 61(8): 1440-1448. .146 Turk, O., and Mavinic , D.S. 1989b. Maintaining nitrite build-up in a system acclimated to free ammonia. Water Res. 23(11): 1383-1388. Villaverde, S., Fdz-Polanco, F., and Garcia, P .A . 2000. Nitrifying biofilm acclimation to free ammonia in submerged biofilters start-up influence. Water Res. 34(2): 602-610. Voets, J.P., Vanstaen, H . and Verstraete, W. 1975. Removal of nitrogen from highly nitrogenous wastewaters. J. Water Pollut. Control Fed. 47: 394-398. Yang, L . , and Alleman, J .E. 1992. Investigation of batchwise nitrite build-up by an enriched nitrification culture. Wat. Sci . Technol. 26(5-6): 997-1005. 147 6.0 Ni tr i f ier Populat ion Dynamics in a Bench Scale Convent ional Act ivated Sludge Reactor Fol lowing an Induced Free Ammon ia Perturbat ion 5 Introduction There has been a considerable amount of interest in the past ten to fifteen years in the design of nitrification/denitrification systems that selectively inhibit nitrite oxidation - thereby eliminating the formation of nitrate. Numerous authors (Voets et al. 1975; Turk and Mavinic 1986, 1987, 1989a, 1989b; Balmelle et al. 1992; Chen et al. 1991; Fdz-Polanco et al. 1996; Garrido et al. 1997; Hyungseok Yoo et al. 1999) have reported the capital and operational benefits of the process, referred to here as the nitrate shunt. These benefits include a 25% reduction in aeration requirements, a 40% reduction in external carbon addition for denitrification, a potential reduction in anoxic zone volume and a significant reduction in sludge production. To date, free ammonia is the consensus cause of nitrite oxidizer inhibition and therefore a high free ammonia concentration is considered the key to process operation. The work of Anthonisen et al. (1976), who studied the effect of free ammonia on environmental samples from activated sludge and soil systems, is considered by many to be the definitive work that established the inhibitory nature of free ammonia toward nitrite oxidizers. However, some investigators have called into question the true cause of nitrite oxidizer inhibition. For example, Cecen and Ipek (1998) suggested that the dissolved oxygen to free ammonia ratio and not the free ammonia concentration itself is important, when attempting to induce nitrite accumulation. The work of 5 A version of this chapter has been accepted for publication with minor revisions J. Environ. Eng. Sci. June 2004. Simm, R.A., Ramey, W.D., Mavinic, D.S. Nitrifier population dynamics in a bench scale conventional activated sludge reactor following an induced perturbation. 148 others (Stuven et al. 1992; Yang et Alleman 1992; and Hyungseok Yoo et al. 1999) has suggested that hydroxylamine - an intermediate in the ammonia oxidation process - may be the true cause of nitrite oxidizer inhibition and therefore of nitrite accumulation. Free ammonia inhibition is one of the most misunderstood topics in environmental engineering and science. Gieseke et al. (2003) studied the structure and activity of multiple nitrifying bacterial populations co-existing in a biofilm. These authors take the work of Anthonisen et al. (1976) literally, reporting free ammonia as inhibitory to Nitrobacter and use the published inhibition limits as proof of possible inhibition of these organisms. However, Anthonisen et al. (1976) used mixed liquor from wastewater plants and soil cultures in their research. Based upon the conventional knowledge at the time, they assumed Nitrobacter was the dominant nitrite oxidizer but did not confirm this assumption by any measurement technique. It is now believed that Nitrospira, and not Nitrobacter, is typically the dominant nitrite oxidizer in both mixed liquor and soil cultures (Schramm et al. 1998; Juretschko et al. 1998; Daims et al. 2000; Bartosch et al. 2002). This dominance occurs even though the studied species of Nitrospira grow significantly slower in pure culture than Nitrobacter (Juretschko et al. 1998). There is also considerable discrepancy in the published range of free ammonia concentrations that are believed to be inhibitory to nitrite oxidizers. For instance, Anthonisen et al. (1976) reported free ammonia inhibition of nitrite-oxidizing organisms at free ammonia concentrations of 0.1 to 1 mg/L, Turk and Mavinic (1986) reported no significant nitrite-oxidizer inhibition for free ammonia concentrations below 10 mg N H 3 - N / L , whereas Mauret et al. (1996) concluded that free ammonia inhibits Nitrobacter, in the range of 6.6 to 8.9 mg N H 3 - N / L . On the other hand, Gieseke et al. (2003) state that it is not known whether bacteria of the genus Nitrospira are 149 inhibited by free ammonia. Recent work conducted at the University of British Columbia, using pure cultures of Nitrospira moscoviensis indicated that free ammonia was not inhibitory to these organisms and casts doubt on whether this compound is inhibitory to Nitrospira-like organisms in general (Chapter 5 and Simm et al. 2004a). If Nitrospira is the predominant nitrite oxidizer in environmental matrices and free ammonia is not inhibitory to these organisms, one must ask what is the true cause of nitrite accumulation? Although there is some disagreement on the cause of nitrite accumulation resulting from nitrite oxidizer inhibition, the general consensus on the long term viability of the process is aptly summarized by Van Loosdrecht and Jetten (1998) who conclude: "nitrogen removal via the oxidation of ammonia to nitrite that is subsequently denitrified does not appear stable for long term operation." The primary limitation of the process appears to be the apparent acclimation o f nitrite oxidizing organisms to increasing concentrations of free ammonia. This acclimation has been reported by numerous researchers, including Turk and Mavinic (1986, 1987, 1989a, 1989b) and Villaverde et al. (2000). The acclimation response o f microbial populations is typically attributed to the proliferation of small populations, presence/absence of toxins, predation by protozoa, appearance of new genotypes, diauxie, and enzyme induction (Alexander 1999). The purpose of this study was to quantify the changes in the nitrite oxidizer population during a period of free ammonia stress, due to an increase in influent ammonia loading. These populations were studied using a combination of slot blot hybridization and fatty acid analyses (Simm et al, 2004b). . . . ' 150 Methodology Reactor System A 10.8-liter, bench-scale reactor, operating at a solids retention time (SRT) of five days (120 hours) was used in this study. The reactor was designed with a gas tight headspace, to allow collection of headspace gas samples. Reactor headspace was vented through a gas collection bulb and beaker filled with water to maintain a constant headspace pressure. The reactor was run as a continuous stirred tank reactor (CSTR), with secondary clarification having a two-hour settling time. The reactor was mixed with a 100-rpm motor. The influent feed to the reactor was 24 liters per day with a recycle of 100%, giving a nominal reactor hydraulic retention time (HRT) of approximately 11 hours. The dissolved oxygen concentration in the reactor was controlled manually between 4 and 5 mg/L with a rotameter. Once the rotameter was set, the aeration setting was only changed when the measured dissolved oxygen was outside the desired range. The dissolved oxygen in the reactor was measured with a dissolved oxygen probe. Oxygen values were stored and logged every thirty seconds using a computer and data logging system. The reactor p H was also logged and controlled between p H 7.9 and 8.4, by adjusting influent alkalinity through a combination of sodium bicarbonate (0.6 M ) and sodium hydroxide (2 M ) addition. The Garrett configuration, in which biomass is wasted directly from the reactor (Grady et al. 1999), was used for solids control; solids wasting was carried out daily. The reactor setup has been presented in previous publication (Chapter 3 and Simm et al. 2004b). The original seed for the reactor was taken from an S B R system operating at the Kent wastewater 151 treatment plant in Agassiz, British Columbia, Canada. The typical operating regime of the Kent facility is described by Louzeiro et al. (2002). Synthetic Feed The initial feed components consisted of 254 mg/L sodium acetate (NaCHaCOOH), 161.5 mg/L ammonium chloride (NH4CI), 1.7 g/L sodium bicarbonate (NaHCOa), 48.75 mg/L yeast extract, 13.65 mg/L potassium phosphate (K2HPO4X 62.1 mg/L magnesium sulphate (MgSC^), 62.1 mg/L calcium chloride (CaCl2'2H20), and 4 ml of a trace metals mix per twenty liters of feed. The trace metals mix was identical to that used by Turk and Mavinic (1986, 1987, 1989a, 1989b) for their research program. The trace metals mix included: 4.9 grams FeCb, 1 gram M n S O 4 H 2 0 , 0.8 grams Z n C l 2 , 0.5 grams C u C l 2 - 2 H 2 0 , 0.73 grams C o C l 2 H 2 0 , 0.7 grams Na2MoCy2H20, 0.3 grams Na2B40r IOH2O, and 44.3 grams of sodium citrate in 250 ml of water. The concentration of ammonium chloride was increased to achieve the desired transient free ammonia concentration in the reactor at the start of the acclimation period. Analy t ica l Methods Fatty A c i d M e t h y l Esters ( F A M E ) Mixed liquor samples were collected and subjected to fatty acid methyl ester analysis ( F A M E ) . The analysis procedure was described previously (Chapter 3 and Simm et al. 2004b). 152 Nitrous Oxide (N 20) The reactor headspace was fully enclosed, to allow for the collection of reactor off gases. Of f gas samples were collected using a gas tight syringe and analyzed for nitrous oxide by gas chromatography, using a Hewlett-Packard 5880A equipped with an electron capture detector. In the G C analysis, nitrogen was used as the carrier gas (at 20 mL/min) with a column packing material of Haycep C. The injector, oven, and detector temperatures were 150 °C, 100 °C, and 250 °C, respectively. Nitric Oxide (NO) Off gas samples were collected for nitric oxide analysis using a gas tight syringe. The collected samples were immediately injected into a Sievers Model 280i Nitr ic Oxide Analyzer (NOA™). This instrument uses a high-sensitivity detector for measuring nitric oxide based on a gas-phase chemiluminescent reaction between nitric oxide and ozone. The emissions from the activated nitrogen dioxide produced are detected by a thermoelectrically cooled, red-sensitive photomultiplier tube. The detection limit of the N O A for measurement of gas-phase N O is approximately 0.5 ppbv. Chemical Analyses The analysis of ammonia, total Kjeldahl nitrogen ( T K N ) , nitrate, nitrite, total organic carbon, and total dissolved solids was carried out as described in the 19 t h Edition of Standard Methods for the 153 Examination of Water and Wastewaters (Eaton et al. 1995). Hydroxylamine was measured by gas chromatography according to the method described in Chapter 2 (Simm et al. 2004c). R N A Slot Blot t ing Mixed liquor samples were collected and subjected to R N A slot blotting analysis. The analysis procedure was previously described in Chapter 3 (Simm et al. 2004b). The molecular probes used for this assessment included Nso 190 (5'-CGA TCC CCT GCT TTT TCT CC-3') targeting all characterized ammonia-oxidizers in the P subdivision of the Proteobacteria (purple bacteria), Nb 1000 (5 '-TGC GAC CGG TCA TGG-3') targeting Nitrobacter, and Ntspa-454 (5'- TCC ATC TTC CCT CCC GAA AA -3') targeting Nitrospira moscoviensis and related Nitrospira-like organisms. The N b 1000, and Nso 190 probes are described by Mobarry et al. (1996). The Ntspa-454 probe for Nitrospira is described by Hovanec et al. (1998). Results The reactor system was seeded in early A p r i l , 2003 and operated with an influent total ammoniacal nitrogen ( N H 3 - N + N H / - N abbreviated as T A N ) concentration of approximately 40 mg/L as nitrogen. The system stabilized by early M a y and the step increase in the influent ammonia was initiated on M a y 7, 2003. The average influent and effluent concentrations prior to the step are summarized in Table 6.1. The average concentrations o f nitrous and nitric oxide in the reactor headspace were <1 ppm and <0.5 ppm respectively, prior to the step increase. 154 To simulate an instantaneous increase in ammonia loading the reactor was spiked to 110 mg-N/L when the influent ammonia concentration was increased from approximately 40 to 150 mg-N/L. The ammonia spike resulted in an initial reactor free ammonia concentration of approximately 9 mg-N/L. The combination o f the spike and increase in the influent concentration resulted in a sustained period of free ammonia stress, with concentrations well above the inhibitory range reported for nitrite oxidizers of 0.1 to 1 mg/L (Anthonisen et al. 1976). Manual p H control was maintained during the first 24 hours of the test, in order to maintain the desired p H range; the influent alkalinity was adjusted on the second day, following the spike, to account for the increase in ammonia oxidation. Chemical Data and Gas Emissions Ini t ial Per iod Fo l lowing the Perturbation (Hour 1 - 72) Time series plots o f the reactor concentrations o f ammonia; nitrate and nitrite for the entire study duration are presented in Fig . 6.1. Time series plots of both nitrous oxide emission rate (mg-N/day) and reactor nitrate concentration are presented in Fig. 6.2. The measured concentrations of headspace nitric oxide were relatively low (< 5 ppm) throughout the duration of the study. A measurable free ammonia concentration was maintained in the reactor over the first seventy-two hours of the study. The increase in reactor free ammonia coincided with an immediate increase in nitrous oxide emissions and a gradual decline in the concentration of nitrate in the reactor. The dissolved oxygen concentration in the reactor declined steadily during the first 48 hours following the perturbation (dissolved oxygen concentration of 3.8 mg/L by hour 48), at 155 which time the aeration rate was increased from 1.48 to 1.98 L/min. The latter aeration rate was maintained for the remainder of the study. The headspace nitrous oxide concentration increased from a concentration of less than 1 ppm to a concentration of at least 70 ppm in the first twenty-four hours following the spike. Thirty hours after the perturbation the reactor nitrate concentration had essentially leveled off between 8 and 12 mg-N/L and remained relatively stable until the free ammonia concentration dropped to zero (by hour 72). B y hour thirty the concentration of nitrous oxide in the headspace had started to decline and it leveled off after hour 72. The cumulative nitrate flux from the reactor during the first 72 hours o f the study is presented in Table 6.2. This value is compared to the value predicted by the original nitrite oxidation activity, assuming no change in nitrite oxidation rate (i.e. the reactor nitrate concentration of 35.9 mg-N/L prior to the perturbation remained unchanged). What is interesting about the data presented in Table 6.2 is that they indicate that essentially all (>95%) of the measured decrease in nitrate, relative to the initial concentration, over the first 7.7 hours can be accounted for by the estimated cumulative nitrous oxide emissions. This balance suggests that nitrate was being extensively denitrified to nitrous oxide by heterotrophs or that nitrite was denitrified to nitrous oxide either autotrophically or heterotrophically, before the nitrite oxidizers had an opportunity to convert it to nitrate. Even though nitrous oxide can also be generated by chemodenitrification via the autodecomposition of hydroxylamine to nitrous oxide, this source was probably negligible in this study because hydroxylamine samples taken during the first twenty-four hours, following the perturbation did not contain measurable quantities of hydroxylamine. 156 B y hour 23.8, the actual cumulative sum of nitrate and emitted nitrous oxide only accounted for approximately 60% of the projected values. However, for logistical reasons (i.e. overnight run) no nitrous oxide samples were taken between hour 7.7 and 23.8. If the headspace nitrous oxide concentration had continued to increase during that time period the estimated cumulative nitrous oxide emitted would "under estimate" the actual value. If one assumes that the projected cumulative nitrate value for hour 23.8 equals the cumulative of the actual nitrate plus nitrous oxide, and then applies the actual measured increases in nitrous oxide emissions from that point forward the sum (referred to in Table 6.2 as Actual + adjusted N2O-N (without considering wasting)) accounts for over 90% of the projected cumulative nitrate between hour 23.8 and 29.8. The adjusted cumulative total would account for 75% and approximately 60% of the projected total by hours 48 and 71.3 respectively. However, a portion of the original nitrite oxidizer activity was lost v ia wasting and therefore the column labeled actual + adjusted N2O-N (with wasting) in Table 6.2 should be used to balance the nitrogen in the system. When these adjusted values are used the agreement between the projected (with wasting) and actual + adjusted N2O (with wasting) is greater than 90% by hour 48 and greater than 80% by hour 72. The potential significance of Table 6.2 is discussed in more detail in the Discussion section of this manuscript. The ammonia oxidation rate increased approximately twenty percent, five hours after the perturbation (see Table 6.3). The ammonia oxidation rate had increased by 75%, relative to the original, after the first twenty-four hours and reached the maximum rate by 71 hours following the original perturbation. The increase in the ammonia oxidation rate was estimated via a mass balance on a time step basis. It was assumed that most of the ammonia loss was due to ammonia oxidation. The nitrite and nitrate production rates were estimated in a similar fashion. The nitrite production rate was estimated as the sum o f the cumulative nitrite and nitrate since all the 157 nitrate arises from nitrite. The nitrite production rate increased marginally and stayed relatively stable immediately following the step (first 7.7 hours), whereas the nitrate production rate steadily declined. B y hour 23.8, the nitrate production rate stabilized and remained relatively unchanged until hour 71. The nitrite production rate increased proportional to the ammonia oxidation rate. Post Free Ammonia Stress Period (Hour 72 to Hour 900) B y hour 72, the free ammonia concentration in the reactor was essentially zero. B y this point in time, all o f the continuously supplied ammonia was being completely oxidized to nitrite and the mass emission rate of nitrous oxide was approximately one third of the maximum measured rate (refer to Fig . 6.1 and 6.2). The nitrate concentration in the reactor increased steadily between hour 72 and 300 signifying a recovery of the nitrite oxidation activity. The reactor nitrate concentration became erratic between hour 300 and 500 when there was an inexplicable increase in effluent solids concentration that resulted in a sharp drop in system solids retention time (SRT dropped to approximately 3 days between these time periods). The system appeared to recover by hour 500 and the result was a gradual increase in the reactor nitrate concentration until complete nitrification was established by hour 900. The increase in the nitrate concentration between hour 72 and 300 was mirrored by a steady decline in the nitrite concentration. This rate of decline slowed between hour 300 and 500, coincidental with the drop in system SRT; however, the decline in nitrite concentration continued from hour 560 onwards, until hour 900, when the reactor nitrite concentration was essentially zero. The mass emission rate of nitrous oxide appeared to decline between hour 72 and 200, was 158 relatively stable between hour 226 and 410, and then declined steadily to zero from hour 410 onwards due to an inexplicable increase in effluent suspended solids. Hour 410 also coincided with the starting point for the recovery in system SRT. The variation in system SRT with time is presented in Fig. 6.3. Fatty Acid and RNA Slot Blotting Analyses F A M E Analyses Results Membrane fatty acid and R N A slot blotting analyses were used to characterize the nitrifying population during the duration of the study. Previous work (Chapter 3 and Simm et al. 2004b) indicated that measuring membrane fatty acids for the ratio of cis(9) hexedecenoic acid to cis(ll) hexedecenoic acid was a useful tool for tracking changes in nitrifier populations in mixed cultures. That study indicated that the ratio of cis(9) hexedecenoic acid to cis(ll) hexedecenoic acid increased as the proportion of nitrite increased in the reactor effluent. Fatty acid analysis of pure culture samples indicated cis(ll) hexedecenoic acid and cis(9) hexedecenoic acid accounted for a significant proportion of the fatty acid total in Nitrospira moscoviensis and Nitrosomononas europaea, resepectively. This finding was consistent with the work of others (Blumer et al. 1969; Lipski et al. 2001). The ratio of cis(9) hexedecenoic acid to cis(ll) hexedecenoic acid in this study is presented in Fig. 6.4. The ratio of cis(9) hexedecenoic acid to cis(ll) hexedecenoic acid increased significantly up until hour 192, suggesting that the ammonia oxidizer population was increasing relative to the Nitrospira population during this time period. Hour 192 was roughly the time where the reactor 159 nitrate concentration had returned to its original value prior to the perturbation. The ratio dropped between hour 192 and 288, suggesting that the Nitrospira population was increasing relative to the ammonia oxidizer population. Between hour 288 and 600, the ratio was relatively stable and then it dropped between hour 600 and 900. Surprisingly, the ratio did not return to the original value prior to the perturbation. However, the fatty acid ratio does appear to track the nitrous oxide profile fairly closely. RNA Slot Blotting Results The total RNA concentration (ug/mL) and concentrations of Nso 190 (pg/mL), Ntspa 454 (pg/mL), and Nb 1000 (ug/mL) RNA, are presented in Fig. 6.5A through 6.5D, respectively. The total RNA concentration (Fig. 6.5A) dropped slightly until approximately hour 300 and then appeared to increase between hour 300 and 500, then dropping again after hour 500. The increase between hours 300 and 500 is consistent with an expected increase in biological growth with the drop in the system SRT. The concentration of Nso 190 RNA (probe Nso 190 targets all characterized ammonia-oxidizers in the P subdivision of the Proteobacteria) increased approximately 3 fold by hour 72; this is consistent with the fact that the additional ammonia added at the start of the perturbation was completely oxidized by this time. Interestingly, there was no appreciable difference between the 24-hour RNA concentration for ammonia-oxidizers and the pre-perturbation values. In general, the concentration of RNA for ammonia-oxidizers continued to increase until hour 360, after which time the concentration was approximately constant at 1.5 u.g/mL and represented approximately 10% of the total system RNA. The relatively high values between hour 300 and 160 500 are consistent with the expected increase in nitrification activity to maintain functionality, when system S R T drops. The relatively low concentration of R N A for ammonia-oxidizers at hour 192 is believed to be a sampling anomaly. The concentration of Ntspa 454 R N A (probe Ntspa 454 targets Nitrospira moscoviensis and related Nitrospira-like organisms) declined approximately 35% during the first twenty-four hours following the ammonia step and declined an additional 28%) (total approximately 46%) by seventy-two hours following the step (Fig. 6.5C). The observed decrease matches the decline predicted i f Nitrospira-like organisms stop growing and are removed from the system via wasting. Morgenroth et al. (2000) probed mixed liquor samples taken from S B R systems, having varying idle periods o f 6-to-20 days without feeding; and reported the signal intensities of the ammonia-oxidizers and nitrite-oxidizers. Nitrospira obtained by in situ hybridization were not significantly reduced during the idle periods. The work of Morgenroth et al. suggests that the R N A content of the nitrite oxidizers could only be reduced via growth inhibition or wasting from the system during starvation. The R N A concentration for the Nitrospira-like organisms started to increase again starting at hour 72, once the free ammonia was all but gone; however, the increase stopped between hour 288 and 504, when the reactor S R T dropped below 5 days. There was a slight increase in the R N A concentration for Nitrospira-like organisms after hour 504; however, the concentration did not exceed the pre- perturbation values. The R N A concentration for Nitrospira-like organisms accounted for 8 to 10 % of the total R N A at hour 816. The concentration o f NblOOO R N A (probe N b 1000 targets Nitrobacter) stayed relatively constant at approximately 0.2 pg/mL (approximately 2% of the total) until hour 216, at which time the concentration began to increase. The increase in R N A concentration for Nitrobacter 161 was coincidental with the drop in system SRT (hour 300 to 500) and the leveling off of the R N A concentration for Nitrospira-Wke organisms. The R N A concentration for Nitrobacter accounted for approximately 6% of the total R N A at hour 900. Discussion The population of Nitrospira-like nitrite-oxidizing organisms was adversely affected following the ammonia perturbation. The bulk of the impact, based upon observed nitrate levels in the reactor, had manifested itself within the first 24 hours. Interestingly, the reactor nitrate concentration was relatively stable between hour 24 and 71, even in the presence of free ammonia concentrations that are widely believed to be inhibitory to nitrite oxidizers in general. The fact that the NblOOO R N A concentration stayed relatively constant during the free ammonia period suggests Nitrobacter-like organisms were not as affected as the Nitrospira-Wke. organisms by the perturbation and that this population rather than the Nitrospira-Yike organisms was responsible for the limited nitrite oxidation (approximately 30% of the original activity prior to the perturbation) that took place between hour 24 and 72. The decline in Ntspa 454 R N A concentration over the first 72 hours is consistent with the wasting rate as i f there was a stoppage or complete impairment in growth of Nitrospira due to inhibition or starvation. A s indicated by Neidhardt et al. (1990), a shift-down (reduction in growth) brings on an immediate decrease in the rate of R N A synthesis. A decline in nutrient availability for Nitrospira-\\kt organisms could have resulted in a decline in observed R N A synthesis. This would result in a decrease in Nitrospira R N A i f the Nitrospira grew sufficiently slowly that it was wasted from the system faster than it grew. 162 Despite the presence of Nitrospira and Nitrobacter (Fig. 6.5C and 6.5D) the nitrate concentration dropped during the free ammonia spike. On the surface it appears that free ammonia was inhibitory to Nitrospira-like organisms and caused the decline in reactor nitrate concentration. However, free ammonia inhibition of Nitrospira appears unlikely based upon the results of a pure culture inhibition study recently completed in the U B C laboratory (Chapter 5). During that study when pure cultures of Nitrospira moscoviensis were subjected to free ammonia concentrations of 1 and 10 mg-N/L there was no observed difference in nitrite oxidation rate between these organisms and a control group that was not subjected to a free ammonia stress. These results are not unlike those of Hunik et al. (1993), who conducted chemostat experiments using pure cultures of Nitrobacter agilis. These investigators reported that, only at p H 6.5, could the inhibition of ammonia be clearly distinguished from the inhibition of a N a C l solution and concluded that inhibition of Nitrobacter agilis by N H 3 was unlikely. These observations suggest that the normal flow of ammonia-nitrogen to nitrate was diverted following the spike. The increase in nitrous oxide emissions coincides with the decline in nitrite oxidation and might be the true cause of apparent free ammonia inhibition. The present study shows that an increase in nitrous oxide emissions is coincidental with nitrite accumulation, in virtually every studied case. A s indicated in Table 6.2, the majority of the decline in nitrate formation (>95%), relative to the original, over the first 7.7 hours, can be accounted for by the increase in nitrous oxide emission. This implies that nitrite, that would have been oxidized to nitrate prior to the spike, was reduced to nitrous oxide instead. The potential sources of the nitrous oxide include heterotrophic denitrification of nitrate to nitrite and then nitrous oxide, enhanced heterotrophic denitrification o f nitrite to nitrous oxide, autotrophic denitrification o f nitrite, and chemodenitrification of hydroxylamine. 163 The R N A data for Nitrospira does not support heterotrophic denitrification of nitrate to nitrite and nitrous oxide. If the oxidation of nitrite to nitrate continued and the nitrate decreased due to heterotrophic activity, then the population of Nitrospira should have increased; however, Nitrospira R N A declined in a manner that suggests this population declined due to the fact it was not growing and was wasted from the system. The heterotrophic denitrification of nitrite to nitrous oxide by other organisms cannot be entirely eliminated as a possibility, although it is difficult to imagine establishing an active denitrifying community in a reactor with an average dissolved oxygen concentration of over 5 mg/L, prior to the ammonia spike. Nitrous oxide is one of the end products of the autodecomposition of hydroxylamine. Although samples were collected and analyzed for hydroxylamine over the first twenty-four hours, no hydroxylamine was measured. In addition, the production of hydroxylamine by ammonia oxidizers appears to be associated with a limitation in potential electron acceptor concentration (both dissolved oxygen and nitrite) and that limitation did not appear to be the case here. The ability of ammonia oxidizers to denitrify nitrite to nitrous oxide has been well established (Anderson and Levine, 1986). Hooper (1968) isolated the enzyme nitrite reductase from cells of Nitrosomonas, demonstrating gas production with hydroxylamine as an electron donor for nitrite reduction. Poth and Focht (1985) have hypothesized the nitrite reductase system in Nitrosomonas functions to: conserve oxygen for use by ammonia monooxygenase, reduce production o f nitrite (which may accumulate to toxic levels), and decrease competition for oxygen by nitrite oxidizers, by denying them their source of substrate. The natural habitat o f 164 many nitrifiers is the oxic/anoxic interface where oxygen is often limiting, which makes the ability to denitrify nitrite to nitrous oxide advantageous. Therefore, the competition for nitrite between ammonia oxidizers and the Nitrospira-like nitrite oxidizers appears to be a potential cause of the observed decline in nitrite oxidation. Interestingly, Gieseke et al. (2003) concluded that 'nitrifier' denitrification by Nitrosomonas europaea/eutropha might play a role in the nitrifying function of the biofilm of highly ammonium-loaded systems. A localized decrease in oxygen availability following the ammonia spike at time zero could also have potentially contributed to the drop in nitrate concentration. A s indicated previously, the initial increase in ammonia oxidation rate, following the perturbation, was almost 20%. The oxygen uptake rate associated with ammonia oxidation would likely result in localized lowering of the oxygen gradient inside the floe particles. Ammonia oxidizers use approximately-three times as much oxygen-as nitrite oxidizers on a stoichiometric basis, which implies that a 20% increase in localized oxygen consumption would leave 60% less oxygen for the equivalent nitrite oxidation, on a nitrogen basis. Work with membrane bound biofilms (Schramm et al., 2000), demonstrated spatial separation between Nitrobacter and Nitrospira populations in response to both oxygen and nitrite gradients. The oxic part o f the biofi lm, which was subjected to high ammonium and nitrite concentrations, was dominated by members o f the genus Nitrobacter while Nitrospira species were virtually absent in this part o f the film. Conversely, Nitrospira were most abundant at the oxic-anoxic interface where oxygen and nitrite concentrations were significantly lower. A similar spatial separation inside a floe particle would increase the vulnerability o f Nitrospira-like organisms to a decline in bulk liquid dissolved oxygen 165 concentrations relative to Nitrobacter-like organisms as a result o f a drop in the bulk liquid dissolved oxygen concentration. Nitrous oxide generation by Nitrosomonas would also be more likely to take place in these localized low dissolved oxygen zones, with a resulting reduction in nitrite and oxygen supply for Nitrospira-like organisms. The combination of a localized reduction in the availability of nitrite, oxygen, and possibly carbon dioxide, may have caused the majority of the measured impact on Nitrospira-like relative to Nitrobacter-like nitrite oxidizers. A s noted by Wood (1986), the oxidation-reduction potential of ammonia and nitrite are similar (+ 400 mV) , yet approximately 2.8 times more CO2 is fixed per ammonia oxidized to nitrite in Nitrosomonas, than per nitrite oxidized to nitrate by Nitrobacter (Baas Becking and Parks, 1927). Although there are no similar published figures for Nitrospira, relative to Nitrosomonas, it is likely that the relative difference in carbon dioxide fixed is similar. The competition for substrate between ammonia and nitrite oxidizers is, therefore, considered as a likely cause of the observed impact upon Nitrospira-like organisms. The substrate competition would be exacerbated by the structure of ammonia and nitrite oxidizer colonies. A F I S H image taken from another reactor studied as part o f this research program is presented as Fig . 6.6 to illustrate the impact of the free ammonia perturbation on the nitrifier population (the ammonia oxidizer colony is labeled with the red fluorochrome whereas the nitrite-oxidizer colonies are labeled with green). A s indicated in Fig . 6.6, an increase in free ammonia, results in a decline in dissolved oxygen in the center o f the ammonia oxidizer colony. The ammonia oxidizers in the center of the colony then use nitrite as a terminal electron acceptor producing nitrous oxide, in turn making less nitrite locally available for nitrite oxidizers. The combination of localized reduction in nitrite and oxygen concentrations, coupled with the slower 166 growth rate of nitrite-oxidizers relative to ammonia-oxidizers, put the nitrite-oxidizers at a competitive disadvantage. Nitrospira-like organisms are more likely to be located at the fringe of oxygen availability and grow more slowly than Nitrobacter and are therefore most affected. A s indicated in a review by Philips et al. (2002) numerous researchers have begun reporting the importance o f the juxtaposition of the ammonia and nitrite oxidizer colonies to nitrite accumulation. In the initial stages of the ammonia perturbation nitrite appeared to be diverted to nitrous oxide as opposed to nitrate potentially explaining the drop in Nitrospira-like R N A concentrations (prior to the perturbation these organisms are more likely to be located in the lowest dissolved oxygen regions of the floe). The increase in bulk liquid nitrite was likely due in part to the lower yield of nitrite oxidizers relative to ammonia oxidizers and an insufficient nitrite oxidizer population in,areas of the floe where both dissolved oxygen and nitrite concentrations were suitable. The fact that most of the impact on nitrite oxidizers, based upon bulk nitrate levels in the reactor, occurred over the first 24 hours is consistent with this hypothesis. Conclusions • R N A slot blotting results indicate that Nitrospira R N A declined whereas Nitrobacter R N A remained unchanged suggesting Nitrospira were more adversely impacted than Nitrobacter by a period of free ammonia stress following an increase in influent ammonia loading. • A drop in system S R T corresponded to a relative increase in Nitrobacter population, as measured by R N A slot blotting, consistent with the higher in situ growth rate of Nitrobacter relative to Nitrospira. 167 The initial decline in nitrite oxidizer activity was accompanied by an increase in reactor nitrous oxide emissions. The competition between ammonia oxidizers and Nitrospira-like organisms for nitrite and potentially oxygen is believed to be the primary cause of reduced nitrite oxidizer activity. The consistent correlation between the results of the fatty acid analyses and the results from the RNA slot blotting reinforces the potential utility of this technique for tracking nitrifiers in mixed populations. 168 160 140 120 SRT drops below 5 davs s s DD 80 O i-1 60 . a o • • • 4 • • • — SRT recovers • Nitrite (mg-N/L) •Nitrate (mg-N/L) A Ammonia (mg-N/L) 100 200 300 400 500 600 700 Elapsed Time (hours) • 800 4-* 900 1000 Figure 6.1 - Time series plot of reactor ammonia, nitrate, and nitrite concentrations following a free ammonia stress ON 200 100 200 300 700 400 500 600 Elapsed Time (hours) Figure 6.2 - T i m e series plots of nitrous oxide emission rate (mg/day) and reactor nitrate-nitrogen after the start of the free ammonia stress o 6.0 Elapsed Time (hours) F igure 6.3 - System S R T versus time after the start of the free ammonia stress 30 1 Note: Darker bars correspond to time period when SRT dropped below 5 days. nu 24 120 192 216 288 312 336 600 624 648 672 696 720 744 792 Elapsed Time (hours) Figure 6.4 - Ratio of cis (9) hexadecanoiec acid/cis (11) hexadecenoiec acid versus time after the start of the free ammonia stress -0 to 3 1.0 E O) 3, 0.8 < O. ° 6 04 02 0 0 Total RNA Elapsed Time (hours) Note; Darker bars correspond to time period when SRT dropped below S days. Ntspa 454 RNA (Nitrospira RNA) CN CO CN «=. CN CO 3 8 3 5° A co I— co Elapsed Time (hours) 2.5 O) 2.0 < Z 1.5 D. _ 1.0 - i £ < z o.e or Nso 190 (Beta-Proteobacteria Ammonia-oxidizers) RNA Elapsed Time (hours) Nb 1000 RNA {Nitrobacter RNA) Mil CM Ol «T 55 co oo Elapsed Time (hours) Figure 6.5 - RNA concentration in the reactor after the start of the free ammonia stress (a) Before free amrnonia perturbation -i » Figure 6.6 - FISH Images with predicted substrate gradient overlay. Table 6.1. Influent and Effluent Characteristics prior to the start of the free ammonia stress. Sample Nitrogen (mg/L as N ) N H 3 + N I T 4 N 0 2 " N 0 3 " Influent 35 0 0 Effluent 0.2 1.2 36.6 175 Table 6.2. Cumulative Nitrate in Reactor after the start of the free ammonia stress. Elapsed Time (hours) Cumulative Nitrate (mg as N) Actual Nitrate production Actual Nitrate + N 2 0 Projected based upon original rate of nitrate production Actual Nitrate + adjusted N 2 0 (without wasting) Actual Nitrate + adjusted N 2 0 (with wasting) Original activity lost by wasting only. 4.2 15.3 18.4 17.6 17.6 17.6 17.6 5.0 40.1 46.0 45.6 45.6 45.6 45.6 5.7 62.5 69.7 71.4 69.7 69.7 71.4 6.3 79.7 87.5 91.2 87.5 87.5 91.2 6.9 98.3 108.4 113.1 108.4 108.4 113.1 7.3 110.0 120.9 127.1 120.9 120.9 127.1 7.7 122.2 134.9 142.5 134.9 134.9 142.5 23.8 350.8 445.5 720.5 720.5 604.9 604.9 25.0 367.9 476.7 763.6 751.7 636.1 639.4 26.6 389.3 505.7 820.0 780.7 665.1 684.5 27.6 403.6 528.9 858.0 803.9 688.3 714.9 29.8 431.5 575.6 935.2 850.6 735.0 776.7 48.2 641.2 915.5 1595.4 1190.5 1074.9 1172.8 51.5 679.8 954.2 1714.9 1229.2 1113.6 1244.5 52.6 689.7 964.1 1753.4 1239.1 1123.5 1267.6 59.4 760.2 1091.2 1996.8 1366.2 1250.6 1413.6 71.3 865.6 1236.5 2426.8 1511.5 1395.9 1671.6 176 Table 6.3. Estimated Ammonia Oxidation, Nitrite Production, and Nitrate Oxidation Rates with Time over the first 71 hours after the start of the free ammonia stress. Elapsed Time (hours) Rate (mg-N/L-hr) Ammonia oxidation rate Nitrate Production rate Nitrite Production rate 4.2 40.5 35.6 38.0 5.0 49.4 31.8 39.1 5.7 51.6 31.1 40.0 6.3 48.5 31.3 41.5 6.9 48.6 30.4 41.8 7.3 48.6 30.1 42.2 7.7 51.6 28.2 40.4 23.8 69.9 14.2 53.2 25.0 70.2 14.3 56.5 26.6 70.5 13.6 57.9 27.6 71.5 13.5 59.1 29.8 78.6 13.0 61.9 48.2 110.2 11.4 98.9 51.5 113.0 11.6 103.6 52.6 115.3 9.2 104.7 59.4 127.5 10.4 118.2 71.3 147.5 8.8 141.0 177 Bibliography Alexander, M . 1999. Biodegradation and bioremediation (Second Edition). Academic Press, San Diego, California. Anderson, I.C., and Levine, J.S. 1986. Relative rates of nitric oxide and nitrous oxide production by nitrifiers, denitrifiers and nitrate respierers. App l . Environ. Biotechnol. 51(5): 938-945. Anthonisen, A . C . , Loehr, R . C . , Prakasam T.B.S. , and Srinath, E . G . 1976. Inhibition of nitrification by ammonia and nitrous acid. J. Water Pollut. Cont. Fed. 48: 835-851. Balmelle, B . , Nguyen, K . M . , Capdeville, B . , Cornier, J.C., and Deguin, A . 1992. Study o f factors controlling nitrite build-up in biological processes for water nitrification. Water Sci. Technol. 26(5-6): 1017- 1025. Bartosch, S., Hartwig, C , Spieck, E . , and Bock, E . 2002. Immunological Detection o f Nitrospira-like bacteria in various soils. Microbial Ecology, 43: 26-33. Bass Becking, L . G . M . and Parks, G.S. 1927. Energy relations in the metabolism of autotrophic bacteria. Physiol. Rev. 7: 85-106. Blumer, M . , Chase, T., and Watson, S.W. 1969. Fatty acids in the lipids of marine and terrestrial nitrifying bacteria. J. Bacteriol. August 1969: 366-370. 178 Cecen, F., and Ipek, S. 1998. Determination of the inhibition of ammonia-N and urea-N oxidations by the fed-batch reactor (FBR) technique. Water Sci . Technol. 38(1): 141-148. Chen, S.K., Juaw, C . K . , Cheng, S.S. 1991. Nitrification and denitrification of high strength ammonium and nitrite wastewater with biofilm reactors, Water Sci . Technol. 23(7-9): 1417-1425. Daims, H . , Neilsen, P . H . , Juretschko, S., and Wagner, M . 2000. Novel Nitrospira-like bacteria as dominant nitrite-oxidizers in biofilms from wastewater treatment plants: diversity and in situ physiology. Water Sci . Technol. 41: 85-90. Eaton, A . D . , Clesceri, L .S . , Greenberg, A . E . (Editors) 1995. Standard methods for the examination of water and wastewater. 19 ed., American Public Health Association, Inc., N e w York. Fdz-Polanco, F. , Villaverde, S., and Garcia, P . A . 1996. Nitrite accumulation in submerged biofilters - combined effects. Water Sci. Technol. 34(3-4): 371-378. Garrido, J . M . , van Bethum, W . A . J . , van Loosdrecht, M . C . M . , and Heijnen, J.J. 1997. Influence of dissolved oxygen concentration on nitrite accumulation in a biofilm airlift suspension reactor. Biotechnol. Bioeng. 53: 168-178. Gieseke, A . , Bjerrum, L . , Wagner, M . , and Amann R. 2003. Structure and activity of multiple nitrifying bacterial populations co-existing in a biofilm. Environ. Microbiol . 5(5): 355-369. 179 Grady, C .P .L . , Daigger, G.T. , L i m , H . C . 1999. Biological wastewater treatment. 2 n d ed. Marcel Dekker, Inc., N e w York, N Y . Hooper A . B . , 1968, A nitrite reducing enzyme for Nitrosomonas europaea. Preliminary characterization with hydroxylamine as electron donor, Biochim. Biophys. Acta. 162: 49-65. Hovanec, T . A . and DeLong, E .F . 1996. Comparative analysis of nitrifying bacteria associated with freshwater and marine aquaria. App l . Environ. Microbiol . 62: 2888-2896. Hovanec, T . A . , Taylor, L .T . , Blakis, A . , and DeLong, E .F . 1998. Nitrospira-like bacteria associated with nitrite oxidation in freshwater aquaria. App l . Environ. Microbiol . 64: 258-264. Hunik J .H. , Meijer H.J .G. , and Tramper J. 1993. Kinetics of Nitrobacter agilis at extreme substrate, product and salt concentration, App l . Microbiol . Biotechnol. 40: 442-448. Hyungseok Y o o , Kyu-Hong A h n , Hyung-Jib Lee, Kwang-Hwan Lee, Youn-Jung Kwak, and Kyung-Guen Song. 1999. Nitrogen removal from synthetic wastewater by simultaneous nitrification and denitrification (SND) via nitrite in an intermittently aerated reactor. Water Res. 33(1): 145-154. Juretschko, S., Timmermann, G . , Schmid, M . , Schleifer, K . H . , Pommerening-Roser, A . , Koops, H.P. , and Wagner, M . 1998. Combined molecular and conventional analyses of nitrifying 180 bacterium in activated sludge: Nitrosococcus mobilis and Nitrospira-like bacteria as dominant populations. App l . Environ. Microbiol . 64: 3042-3051. Lipski , A . , Spieck, E . , Makol la , A . , and Altendorf, K . 2001. Fatty acid profiles of nitrite-oxidizing bacteria reflect their phylogenetic heterogeneity. Syst. A p p l . Microbiol . 24: 377-384. Louzeiro, N . R . , Mavinic , D.S. , Oldham, W . K . , Meisen, A . , and Gardner, L S . 2002. Methanol-induced biological nutrient removal kinetics in a full-scale sequencing batch reactor. Water Res. 36(11): 2721-2731. Mauret, M . , Paul, E . , Puech-Costes, E . , Maurette, M . T . , and Baptiste, P. 1996. Application of experimental research methodology to the study of nitrification in mixed culture. Water Sci . Technol. 34(1-2): 245-252. Mobarry, B . K . , Wagner, M . , Urbain, V . , Rittmann, B . E . , and Stahl, D . A . 1996. Phylogenetic probes for analyzing abundance and spatial organization of nitrifying bacteria. App l . Environ. Microbiol . 62:2156-2162. Morgenroth, E . , Obermayer, A . , Arnold, E . , Bruhl, A . , Wagner, M . , and Wilderer, P .A . 2000. Effect of long-term idle periods on the performance of sequencing batch reactors. Water Sci . Technol. 41(1): 105-113. Neidhardt, F .C . , Ingraham, J .L. , and Schaechter, M . 1990. Physiology of the bacterial cell - a molecular approach. Sinauer Associates Inc., Sunderland, Massachusetts, U . S . A . 181 Philips, S., Laanbroek, H.J . , and Verstraete, W . 2002. Origin, causes and effects of increased nitrite concentrations in aquatic environments. Re/Views in Environmental Science & Bio/Technology. 1 : 115-141. Poth, M . and Focht, D . D . 1985. I 5 N kinetic analysis of N 2 0 production by Nitrosomonas europaea: an examination of nitrifier denitrification. App l . Environ. Microbiol . 49:1134-1141. Schramm, A . , De Beer, D . , Wagner, M . , Amman, R. 1998. Identification and activities in situ of Nitrospira and Nitrospira spp. as dominant populations in a nitrifying fluidized be reactor. App l . Environ. Microbiol . 64: 3480-3485. Simm, R . A . , Mavinic , D.S. , and Ramey, W . D . 2004a. A targeted study on possible free ammonia inhibition of Nitrospira. (submitted for publication to J. Environ. Eng. Sci . M a y 2004). Simm, R . A . , Ramey, W . D . , and Mavinic , D.S. 2004b. Preliminary evaluation of the use of fatty acid ratios for tracking the potential for nitrite accumulation in nitrifying reactors with low carbon to nitrogen ratio. J. Environ. Eng. Sci. 3: 31-40. Simm, R . A . , Parkinson, P., Mavin ic , -D.S . , Ramey, W . D . 2004c. Hydroxylamine analysis of wastewater samples v ia gas chromatography, (submitted to Environmental Technology for publication July 2004). Stuven, R., Vollmer, M . , and Bock, E . 1992. The impact o f organic matter on nitric oxide formation by Nitrosomonas europaea. Arch. Microbiol . 158: 439-443. 182 Turk, O., and Mavinic , D.S . 1986. Preliminary assessment of a shortcut in nitrogen removal from wastewater. Can. J. C i v i l Eng. 13: 600-605. Turk, O., and Mavinic , D.S. 1987. Selective inhibition: a novel concept for removing nitrogen from highly nitrogenous wastes. Env. Technol. Lett., 8: 419. Turk, O., and Mavinic , D.S . 1989a. Stability o f nitrite build-up in an activated sludge system. J. Water Pollut. Control Fed. 61(8): 1440-1448. Turk, O., and Mavinic , D.S . 19896. Maintaining nitrite buildup in a system acclimated to free ammonia. Water Res. 23(11): 1383-1388. Van Loosdrecht, M . C . M . , and Jetten, M . S . M . 1998. Microbial conversions in nitrogen removal. Water Sci.Technol. 38(1): 1-7. Villaverde, S., Fdz-Polanco, F., and Garcia, P .A . 2000. Nitrifying biofilm acclimation to free ammonia in submerged biofilters start-up influence. Water Res. 34(2): 602-610. Voets, J.P., Vanstaen, H . and Verstraete, W. 1975. Removal of nitrogen from highly nitrogenous wastewaters. J. Water Pollut. Control Fed. 47: 394-398. 183 Wood, P . M . 1986. Nitrification as a bacterial energy source, pp. 39-62 in Prosser, J.I. (Editor), Nitrification, Special Publications of the Society for General Microbiology, Volume 20, I R L Press, Oxford, Washington, D . C . Yang, L . , and Alleman, J.E. 1992. Investigation of batchwise nitrite build-up by an enriched nitrification culture. Water Sci. Technol. 26(5-6): 997-1005. 184 7.0 Mechanisms Responsible for Apparent Free Ammonia Inhibition in a Sequencing Batch Reactor6 Introduction In the past ten to fifteen years there has been a considerable amount of interest in the design of nitrification/denitrification systems that selectively inhibit nitrite oxidation - thereby eliminating the formation of nitrate. Numerous authors (Voets et al. 1975; Turk and Mavinic 1986, 1987, 1989a, 1989b; Balmelle 1992; Chen et al. 1991; Fdz-Polanco et al. 1996; Garrido et al. 1997; Hyungseok Yoo et al. 1999) have reported the capital and operational benefits of this nitrate shunt; to include a 25% reduction in aeration requirements, a 40% reduction in external carbon addition for denitrification, a potential reduction in anoxic zone volume and a significant reduction in sludge production. Free ammonia, currently believed to be the cause of nitrite oxidizer inhibition, and level of ammonium are thought to be the key to process operation. The work o f Anthonisen et al. (1976), who studied the effect of free ammonia on environmental samples from activated sludge and soil systems, is considered by many to be the definitive work that established the inhibitory nature of free ammonia toward nitrite oxidizers. However, some investigators have called into question the true cause of nitrite oxidizer inhibition. For example, Cecen and Ipek (1998) have suggested that the ratio of dissolved oxygen to free ammonia rather than the actual concentration of free ammonia is of primary importance when attempting to induce nitrite accumulation. The work of 6 A version of this chapter has been submitted for publication July 2004. Simm, R.A., Mavinic, D.S., and Ramey, W.D. Mechanisms responsible for apparent free ammonia inhibition in a sequencing batch reactor. A S C E Env. Eng. J. 185 others (Stuven et al. 1992; Yang and Alleman 1992; and Hyungseok Yoo et al. 1999) has suggested that hydroxylamine - an intermediate in the ammonia oxidation process - may be the true cause of nitrite oxidizer inhibition that causes nitrite accumulation. Anthonisen et al. (1976) reported that free ammonia was inhibitory to Nitrobacter and many researchers use the published inhibition limits as proof of possible inhibition. However, based upon the conventional knowledge at the time, Anthonisen et al. assumed that Nitrobacter was the dominant nitrite oxidizer but did not confirm this assumption. This is important because Anthonisen et al. used mixed liquor from wastewater plants and soil cultures in their research. It is now believed that Nitrospira, and not Nitrobacter, is typically the dominant nitrite oxidizer in both matrices (Schramm et al. 1998; Juretschko et al. 1998; Daims et al. 2000; Bartosch et al. 2002). This is in spite of the fact that described species of Nitrospira grow significantly slower in pure culture than Nitrobacter (Juretschko et al. 1998). Interestingly, Gieseke et al. (2003) state that it is not known whether bacteria of the genus Nitrospira are inhibited by free ammonia. Recent work conducted at the University of British Columbia (U.B.C. ) using pure cultures of Nitrospira moscoviensis, indicates that free ammonia does not appear to be inhibitory to these organisms and casts doubt on whether this compound is inhibitory to Nitrospira-like organisms in general (Simm et al. 2004a and Chapter 5.0). If Nitrospira is the predominant nitrite oxidizer in environmental matrices and free ammonia is not inhibitory to these organisms, one must ask what is the true cause o f nitrite accumulation? Although there is some disagreement on the cause of nitrite accumulation resulting from nitrite oxidizer inhibition, the general consensus on the long term viability of the process is aptly summarized by Van Loosdrecht and Jetten (1998) who conclude: "nitrogen removal via the 186 oxidation of ammonia to nitrite that is subsequently denitrified does not appear stable for long term operation." The primary limitation of the process appears to be the apparent acclimation of nitrite oxidizing organisms to the conditions causing the nitrate shunt. This acclimation has been reported by numerous researchers, including Turk and Mavinic (1986, 1987, 1989a, 1989b) and Villaverde et al. (2000). However, acclimation responses of microbial populations result from proliferation of small populations, presence/absence of toxins, predation by protozoa, appearance of new genotypes, diauxie, and enzyme induction (Alexander 1999). The primary objective of this study was to track the specific events leading to the acclimation of a system operating via the nitrate shunt. Aerobic phase nitrite accumulation was induced by periodically subjecting the nitrite oxidizer population in a sequencing batch reactor to an expected inhibitory concentration of free ammonia under anoxic conditions as previously described by Turk and Mavinic (1986, 1987, 1989a, 1989b). Previous work associated with this research project had shown that increased nitrous oxide emissions coincide with nitrite accumulation. Therefore, both gas and population dynamics were studied as part of this research. M i x e d microbial populations were studied using a combination of slot blot hybridization and fatty acid analyses. The potential utility of the fatty acid technique was established in a previous study (Chapter 3 and Simm et al. 2004b). 187 Methodology Reactor System Using a plug flow nitrification/denitrification system, Turk and Mavinic (1986, 1987, 1989a, 1989b) showed that nitrite accumulation could be maintained by periodically subjecting a mixed population to free ammonia stress (5 to 10 mg N H 3 - N / L ) under anoxic conditions. A sequencing batch reactor (SBR) system can be operated to mimic the conditions established by Turk and Mavinic, accomplishing in time what a plug flow system accomplishes in space. The S B R system provides the added advantage of allowing an investigator to conduct in situ batch tests, during individual cycles. This testing is particularly useful for studying free ammonia inhibition since the change in apparent free ammonia inhibition limits can be tracked with time. The reactor system used for this study was a 10.6-liter (operating volume) bench scale S B R . The bench scale S B R was operated with a total cycle time of four hours. The cycle included a 3-minute fill, one hour anoxic, two hour aerobic, thirty minutes settling, fifteen minute decant, and fifteen minute idle. The reactor was mixed with a 100-rpm motor. The influent feed to the reactor was 19.8 liters per day or 3.3 liters per cycle. Reactor control was accomplished using a four channel (one channel each for influent pump, stirrer motor, air solenoid valve, and decant solenoid valve) programmable Chrontrol electronic timer. S R T control was achieved by wasting once daily from the reactor, ten minutes prior to the end of the aerobic phase. The target S R T for this study was 10 days. 188 The reactor was designed with a gas-tight headspace, to allow collection of headspace gas samples. Reactor headspace was vented through a gas collection bulb and beaker filled with water in order to maintain a constant headspace pressure. The original seed for the reactor was taken from the pilot University of Cape Town (UCT) process operating at the University of British Columbia pilot plant facility in Vancouver, British Columbia, Canada. Seed was collected on March 5, 2003 and was stored at 4 °C for two days, prior to use. The anoxic period p H (and therefore free ammonia concentration) was controlled by adjusting the influent feed p H via a combination of sodium hydroxide and sodium bicarbonate addition. This is the same p H control scheme used by Turk and Mavinic (1986, 1987, 1989a, 1989b) to control anoxic zone p H and cause nitrite oxidizer inhibition in the aerobic phase. Decisions on required influent adjustments were made by considering anoxic free ammonia and aerobic dissolved oxygen concentrations. The target anoxic zone free ammonia concentration was 8-to-10 mg N H 3 - N / L . The aeration rate was initially kept constant and the aerobic phase dissolved oxygen level was used as a gauge of nitrifier health. A n y significant increase in aerobic phase dissolved oxygen following a change in influent conditions was used as a cue that the process had likely been impaired and changes were required. In most cases, the approach was to reduce influent ammonia and/or p H for one or two cycles or to remove ammonia from the feed completely following a significant increase in aerobic dissolved oxygen level. This 189 adjustment was deemed to be necessary since unoxidized ammonia simply builds up in the system. Synthetic Feed The bench scale S B R was fed a synthetic feed that simulated a relatively weak sanitary landfill leachate, since this wastewater was considered a likely candidate for application of the nitrate shunt. The carbon required for denitrification was added directly to the influent bucket since influent feeding was done under anoxic conditions. The initial feed components consisted of 520 mg/L sodium acetate (NaCH 3COOH), 340 mg/L ammonium chloride (NH 4 C1), 500 mg/L sodium bicarbonate (NaHCOs), 100 mg/L yeast extract, 28 mg/L potassium phosphate (K2HPO4X 62.1 mg/L magnesium sulphate (MgSOx), 62.1 mg/L calcium chloride (CaCi2-2H.20), and 4 milliliters of a trace metals mix per 20 liters of feed. The trace metals mix included the following ingredients in 250 milliliters of water: 4.9 grams F e C U , 1 gram MnSO -4H 2 0, 0.8 grams Z n C l 2 , 0.5 grams C u C l 2 - 2 H 2 0 , 0.73 grams C o C l 2 H 2 0 , 0.7 grams N a 2 M o 0 4 - 2 H 2 0 , 0.3 grams Na2B4Oyl0H 2 O, and 44.3 grams of sodium citrate. The concentrations of sodium acetate and ammonium chloride were eventually increased in order to achieve the desired anoxic zone free ammonia concentration, as well as to provide the required carbon for denitrification. The final feed components consisted of 650 mg/L sodium acetate (NaCHsCOOH) , 612 mg/L ammonium chloride (NH 4 C1), 1,750 mg/L sodium bicarbonate (NaHCOs), 125 mg/L yeast extract, and 35 mg/L potassium phosphate (K2HPO4). A l l other feed components remained the same as in the initial mixture. Sodium hydroxide was added to the feed bucket to adjust influent p H and, ultimately, the anoxic zone p H . 190 Analy t ica l Methods Fatty A c i d M e t h y l Esters ( F A M E ) Mixed liquor samples were collected and subjected to fatty acid methyl ester analysis ( F A M E ) . The analysis procedure was described previously (Chapter 3 and Simm et al. 2004b). Nitrous Oxide ( N 2 0 ) The reactor headspace was fully enclosed, thus allowing for the collection of reactor off gases. Off gas samples were collected using a gas tight syringe and analyzed for nitrous oxide by gas chromatography, using a Hewlett-Packard 5880A equipped with an electron capture detector. In the G C analysis, nitrogen was used as the carrier gas (at 20 mL/min) with a column packing material of Haycep C. The injector, oven, and detector temperatures were 150 °C, 100 °C, and 250 °C, respectively. Ni t r i c Oxide (NO) Off gas samples were collected for nitric oxide analysis using a gas tight syringe. The collected samples were immediately injected into a Sievers Model 280i Nitr ic Oxide Analyzer (NOA™). This instrument uses a high-sensitivity detector for measuring nitric oxide based on a gas-phase chemiluminescent reaction between nitric oxide and ozone. A photomultiplier tube is used to detect the photon emissions from the activated nitrogen dioxide produced. The detection limit of the N O A for measurement o f gas-phase N O is approximately 0.5 ppbv. 191 Chemical Analyses The analysis of ammonia, total Kjeldahl nitrogen ( T K N ) , nitrate, nitrite, total organic carbon, and total dissolved solids was carried out as described in the 19 t h Edition of Standard Methods for the Examination of Water and Wastewaters (Eaton et al. 1995). Automated methods were used for routine analyses of all nitrogen species. However, a limited number of effluent and anoxic phase samples were taken periodically and analyzed for ammonia via the manual Phenate method. The manual method was used for process control purposes only. Hydroxylamine was measured using the G C method described previously (Chapter 2 and Simm et al. 2004c). R N A Slot Blot t ing Mixed liquor samples were collected and subjected to R N A slot blotting analyses. The analysis procedure was previously described in Simm et al. (2004b). The molecular probes used for this assessment included Nso 190 (5*-CGA TCC CCT GCT TTT TCT CC-3') targeting all characterized ammonia-oxidizers in the (3 subdivision of the Proteobacteria (purple bacteria), N b 1000 (5'-TGC GAC CGG TCA TGG-3') targeting Nitrobacter, Ntspa-454 (5'- TCC ATC TTC CCT CCC GAA AA -3') targeting Nitrospira moscoviensis and related Nitrospira-like organisms, and Ntspa-685 (5'-CAC CGG GAA TTC CGC GCT CCT C-3') targeting Nitrospira moscoviensis and Nitrospira marina. Mobarry et al. (1996) previously described the N b 1000 and Nso 190 probes. The Ntspa-454 and Ntspa-685 probes for Nitrospira have been previously described by Hovanec etal . (1998). 192 Fluorescent In situ Hybridization (FISH) A limited number of samples were analyzed via fluorescent in situ hybridization. Fluorescent in situ hybridization (FISH) was conducted using the Nit r i -VIT™ kit supplied by Vermicon Inc., according to the manufacturers instructions. Results Reactor Performance and Nitrite Accumulation The bench scale reactor was seeded on March 8, 2003 (day 0). The initial seed p H was 5.5. Initially, the reactor was fed with the conventional synthetic feed with an ammonia concentration of 90 mg-N/L and a p H of 7.6. The ammonia concentration and p H of the influent were adjusted in order to control anoxic-zone, free ammonia. The calculated peak anoxic phase free ammonia concentration (NH3-N) during the course of this study is presented in Figure 7.1. On March 11, 2003 the feed p H was increased to 8.6 via the addition of sodium hydroxide to the feed bucket. On March 11, 2003 (day 3) the anoxic zone total ammoniacal nitrogen (NH3-N+NH4+-N) concentration was 32 mg/L (pH = 8.4 and temperature of 24 C°). The calculated anoxic free ammonia (NH3-N) concentration was 3.7 mg/L. The effluent ammonia concentration was essentially zero by March 12, 2002 (day 4) and the influent alkalinity was increased in preparation for an increase in the influent ammonia concentration. 193 The influent ammonia-nitrogen concentration and p H were increased to 159 mg/L and 8.9, respectively, prior to the 1 A . M . cycle (1 AM-to-5 A M cycle) on March 14, 2003 (day 6). In addition, the aeration rate was increased (1.25 liters/min. to 1.38 liters/min.) in anticipation of the increased ammonia oxidation. The anoxic p H and temperature at the start of this cycle were 8.7 and 24 °C respectively, resulting in a free ammonia concentration of 10.4 mg/L. During the 9 PM-to-1 A M cycle, the aeration rate was increased to 1.83 liters/min. and was not changed for the duration of the study. The effluent ammonia-nitrogen concentration on May 14, 2003 (day 6) was 54 mg/L. The measured anoxic ammonia-nitrogen concentration on M a y 15, 2003 (day 7) was 101 mg/L. The anoxic temperature and p H were 24 °C and 8.5 respectively, resulting in an anoxic free ammonia concentration of 14.5 mg/L. The ammonia was eliminated from the feed by the 9 PM-to-1 A M cycle on March 15, 2003 (day 7), due to an increase in aerobic phase dissolved oxygen concentration (0.5 mg/L to 3.8 mg/L). On March 16, 2003 (day 8), the measured ammonia concentration in the anoxic zone was 81.7 mg/L (9 AM-to-1 P M cycle). The measured p H and temperature were 8.4 and 24 °C respectively. The calculated free ammonia concentration in the anoxic zone was 11.7 mg/L. B y the 1 PM-to-5 P M cycle, the measured anoxic zone ammonia concentration was 63.3 mg/L, with an anoxic p H of 8 and temperature of 24 °C. Ammonia was added back into the feed for the 5 PM-to-9 P M cycle and the influent p H was increased from 7.6 to 8.55, for the 9 PM-to-1 A M cycle. 194 The feed composition was not changed again until the 5 PM-to-9 P M cycle on March 19, 2003 (day 11). A t this point, the influent ammonia concentration was increased to 125 mg/L and the influent acetate concentration was increased from 520 mg/L to 650 mg/L as sodium acetate ( N a C H 3 C O O H ) . The feed was adjusted again on March 24, 2003 (day 16), when the influent p H was increased to 8.7 for the 5 PM-to-9 P M cycle and then the feed p H and ammonia concentrations were increased to 8.9 and 150 mg/L, respectively, for the 9 PM-to-1 A M cycles. On March 25, 2003 (day 17), there was a steady increase in the aerobic dissolved oxygen concentration from 1.8 to 3.8 mg/L. Therefore, the influent ammonia and p H concentrations were adjusted to 125 mg/L and 8.31, respectively, to allow the system to recover. This change was initiated prior to the 9 A M cycle, therefore the system was fed the higher ammonia feed for two cycles. This timing becomes especially important when interpreting the R N A data. The anoxic free ammonia concentration was approximately 14 mg-N/L during these two cycles. The final adjustment in the feed was carried out on March 28, 2003 (day 20), when the influent ammonia concentration and p H were adjusted to 150 mg/L and 8.8, respectively. The anoxic free ammonia concentration was 11.8 mg-N/L. Complete nitrification was essentially established by March 31, 2003 (day 23), with nitrite as the primary oxidized species (more than 80%). There was a rapid recovery in nitrite oxidation starting on A pr i l 8, 2003 (day 31), with full recovery by A p r i l 12, 2003 (day 35). The process was, therefore, operating via the shunt for approximately 10 days. Time series plots of effluent nitrite, nitrate, percentage nitrite ((NO2"-N/(N02"-N+N03"-N)) X 100%) and ammonia-nitrogen are presented as Figure 7.2. The largest • 195 relative drop in effluent nitrate concentration was on days 8 and 17, following an increase in anoxic phase free ammonia concentration (Figure 7.1). There was a progressive increase in effluent suspended solids and a decline in reactor mixed liquor suspended solids, that appeared to start on March 26, 2003 (day 18). The effluent suspended solids concentration peaked by Apr i l 3, 2003 (day 26). B y A p r i l 2, 2003 (day 25)the reactor S R T had dropped below 10 days and did not recover until A p r i l 7, 2003 (day 30), at which time the reactor mixed liquor suspended solids concentration had stabilized at approximately 2000 mg/L. Time series plots of the reactor and effluent mixed liquor suspended solids concentrations and system SRT are presented in Figure 7.3. The total ammonia oxidation was not adversely affected by the drop in SRT; however, nitrite was the primary oxidized nitrogen species at the end of the aerobic phase (>80%), signifying nitrite oxidation had been impacted. A time series plot of the total oxidized ammonia-nitrogen over the entire cycle (anoxic ammonia concentration minus effluent ammonia concentration) is presented as Figure 7.4. S B R T rack ing Studies One of the primary advantages of using the sequencing batch reactor as a research tool is that it offers the flexibility of doing online batch tests, tracking the dynamics of the biochemical reactions in situ. Whereas one must do batch tests with grab samples from a C S T R , process kinetics can be measured with an S B R in situ. In addition, the point in time where "apparent free ammonia inhibition" has been lifted can be determined. Several tracking studies were carried out during the shunt period and following the collapse of the shunt, in order to elucidate the 196 differences in process response, over and above nitrite accumulation, between the shunt and non-shunt periods. M a r c h 30, 2003 (day 22) T rack ing Study The first tracking study was carried out on March 30, 2003 (day 22). Complete ammonia oxidation had not yet been established at this time. The measured system performance parameters for this tracking study are presented as Figure 7.5. The hydroxylamine analysis results are not presented in the figure since no measurable quantities of hydroxylamine were found in either of the anoxic or aerobic phases for this, or any other, tracking study. The maximum free ammonia concentration during the cycle was 10.7 mg/L and a measurable free ammonia concentration was carried to the end of the cycle. The free ammonia concentration throughout the cycle exceeded the inhibitory concentration for nitrite oxidizers reported by Anthonisen et al. (0.1 to 1 mg/L). The ammonia oxidation and nitrite production rates during the aerobic phase were 21.9 mg/L-hr and 19.2 mg/L-hr, respectively. Ammonia plots as a straight line in Figure 7.5, indicating ammonia oxidation kinetics are zero order. Although there was no measurable dissolved oxygen in the reactor during the first twelve minutes of the aerobic phase, the ammonia oxidation rate appeared unaffected. The dissolved oxygen concentration in the reactor increased steadily to approximately 3 mg/L (average 1.3 mg/L), but the characteristic dissolved oxygen breakthrough that normally signifies ammonia disappearance was not observed. 197 Nitrate accounted for twelve percent of the oxidized nitrogen in the effluent; however, there was no measurable nitrate in the reactor until eighteen minutes into the aerobic phase. The appearance of nitrate-nitrogen appeared to coincide with an observed measurable oxygen concentration and an estimated free ammonia concentration of 5.8 mg-N/L . Nitrous oxide was emitted from the reactor for essentially the entire aerobic portion of the cycle. The emitted nitrous oxide accounted for approximately 4 mg or one percent of the oxidized ammonia. The total emitted nitric oxide was less than 0.3 mg as nitrogen and was not considered significant. March 31, 2003 (day 23) Tracking Study A second tracking study was carried out on March 31, 2003 (day 23). B y this time, essentially complete ammonia oxidation had been achieved. The measured system performance parameters for the tracking study are presented as Figures 7.6. The maximum free ammonia concentration in the anoxic phase was 7 mg-N/L. A free ammonia residual was carried to the end of the aerobic phase during this cycle and the measured residual was within or exceeded the range reported as being inhibitory by Anthonisen et al. (1976). Nitrate accounted for sixteen percent of the oxidized nitrogen in the effluent; however, there was no measurable nitrate in the reactor until fifteen minutes into the aerobic phase. The appearance of nitrate-nitrogen coincided with an estimated free ammonia concentration of 5.2 mg-N/L. 198 Nitrous oxide was again emitted from the reactor for essentially the entire aerobic portion of the cycle. The emitted nitrous oxide accounted for approximately 8 mg or two percent of the oxidized ammonia. The total emitted nitric oxide was less than 0.15 mg as nitrogen and was not considered significant. April 2, 2003 (day 25) Tracking Study Yang and Alleman (1992) have reported hydroxylamine as an inhibitor of nitrite oxidation. Previous experiments in our laboratory had indicated that high p H and low dissolved oxygen concentration could result in hydroxylamine production and that nitrous oxide is a product of hydroxylamine decomposition. Therefore, the aerobic phase aeration rate was reduced on A p r i l 2, 2003 (day 25) in an attempt to induce hydroxylamine production. It was hoped that a lower aeration rate would minimize any autodecomposition that might be taking place helping us to ascertain whether or not hydroxylamine was the source of measured nitrous oxide during the March 30 (day 22) and March 31, 2003 (day 23) tracking studies. N o measurable quantities of hydroxylamine were observed. The lower aeration rate resulted in a significant delay in ammonia oxidation (ammonia oxidation started thirty-eight minutes into the cycle as opposed to eighteen and sixteen minutes into the cycle for the March 30 (day 22) and March 31 (day 23) studies, respectively) as well as a reduction in nitrous oxide production. The measured system performance parameters for the tracking study are presented as Figure 7.7. 199 April 12,2003 (day 35) Tracking Study Nitrite oxidation had essentially recovered by this time, despite a significantly lower aerobic phase dissolved oxygen concentration of 0.1 mg/L. N o attempt was made to increase the dissolved oxygen concentration to match the shunt period, since nitrification was accomplished within the allotted aerobic period. The measured system performance parameters are presented in Figures 7.8. The nitrate-nitrogen concentration was 3 mg/L seventeen minutes into the aerobic phase. The fact that the nitrate production rate was approximately 10.5 mg/L-hr by this time suggests that nitrite oxidation essentially started at the outset of the aerobic phase. The free ammonia concentration at the start of the aerobic phase was approximately 9.3 mg-N/L . Nitrite made up 50% o f oxidized nitrogen species until minute 99, when the concentration of nitrate-nitrogen exceeded the nitrite-nitrogen concentration. This point appears to approximately coincide with the peak headspace nitrous oxide concentration and the change in slope in the free ammonia concentration curve. The nitrate production rate for this point onward was 30.4 mg/L-hr. April 22,2003 (day 44) Tracking Study Two, back-to-back tracking studies were carried out on A p r i l 22, 2003 (day 44). The primary purpose of these studies was to elucidate the source of emitted nitrous oxide. It was hoped that confirming the source of emitted nitrous oxide might shed light on the cause(s) of observed 200 nitrite accumulation, since the headspace nitrous oxide concentration continued to build as long as nitrite formed a large proportion of the oxidized nitrogen species. The tracking studies were carried out with and without ammonia addition. The initial tracking study was carried out in an identical manner to the previous studies; however, ammonia was eliminated from the feed for the second study. The p H and dissolved oxygen regimes of the first tracking study were matched in the second study. The p H control was accomplished via the addition of a dilute hydrochloric acid solution. Several on and off line titrations were carried out to establish the required normality and feed rate for the p H control chemical. It was critical that the dissolved oxygen concentration in the second tracking study match that of the first and that the gassing rates be identical, to allow direct comparison between the two studies. The oxygen demand dropped substantially when ammonia was eliminated from the feed. A number of on and off line oxygen uptake rate tests, with and without ammonia and nitrite, were carried out to confirm the aeration requirements without ammonia. The process airflow was supplemented with nitrogen in order to match the gassing rate from the first cycle. A nitrite spike was added to the reactor during the aerobic phase, approximately fifteen minutes into the cycle and an acetate spike was added approximately ten minutes after the nitrite spike. It was hypothesized that i f heterotrophic denitrification were the source of emitted nitrous oxide the nitrite would be denitrified to nitrous oxide. The acetate was added as a carbon source in the event there was insufficient carbon to drive the denitrification reaction. The logistical details associated with the second tracking study are presented elsewhere (Shiskowski et al. 2004). The measured system performance parameters for the two tracking studies are presented in Figures 7.9 and 7.10. Tracking study one was carried out between minute 0 and 168 and 201 tracking study two encompasses minute 263 to 449. The observations made for the first cycle tracked on A p r i l 22, 2003 (day 44) are similar to those made for the A p r i l 12, 2003 (day 35) cycle. The total nitrous oxide emitted from the reactor as nitrogen over the entire time was 16.8 mg, or 4.4 percent o f the total ammonia oxidized during the aerobic phase. The headspace nitrous oxide concentration peaked at minute 108, which coincides with the change in slope of the free ammonia curve. Nitrate became the dominant oxidized nitrogen species, approximately 13 minutes into the aerobic phase. A l l of the nitrite added to the aerobic phase during the second tracking study was recovered as nitrate indicating the nitrite oxidizers had the capacity to oxidize nitrite under the tested conditions. There was virtually no nitrous oxide detected in the reactor headspace at this point in time, thus eliminating heterotrophic denitrification as the source nitrous oxide emission source. The aeration rate was returned to the original setting at minute 360 and ammonia was added to the reactor. This addition resulted in an immediate rise in headspace nitrous oxide concentration, confirming that ammonia was required for nitrous oxide generation and suggesting that autotrophic denitrification was the most likely source. Summary of T r a c k i n g Study Results The tracking study results are summarized in Table 7.1. The primary findings of the tracking studies are summarized as follows: • Nitrite oxidation was not completely eliminated even during the shunt period. In fact, nitrite oxidation commenced at a free ammonia concentration of approximately 5 mg-202 N/L. This concentration is far in excess of the reported inhibitory range reported by Anthonisen et al. (1976). Nitrite oxidation commenced at even higher free ammonia concentrations (approximately 9 mg/L) during the non-shunt period. The ammonia oxidation rate was approximately 20% higher during the non-shunt period relative to the shunt period. The nitrite oxidation rate was, however, approximately five times higher during the non-shunt period relative to the shunt period. No measurable concentrations of hydroxylamine were observed; even when the aerobic phase dissolved oxygen concentration was dropped to less than 0.1 mg/L (April 2, 2003 batch test), thus suggesting that hydroxylamine was not likely involved in the observed nitrite accumulation. Although emitted nitrous oxide accounted for a small proportion of the oxidized ammonia (<5%), the headspace nitrous oxide concentration increased as long as nitrite was the dominant oxidized nitrogen species. In all cases, the peak nitrous oxide concentration coincided with the change in slope in the free ammonia curve. In all cases, there was a linear relationship between the concentrations of nitrous acid and nitrate. The April 22, 2003 (day 44) tracking study suggests that ammonia was required in order to produce nitrous oxide. The ammonia either indirectly provided substrate to ammonia 203 oxidizers that denitrified nitrite to nitrous oxide or resulted in a localized decrease in oxygen concentration, thereby creating the required conditions for heterotrophic or autotrophic denitrification. Either set of conditions would give ammonia oxidizers a relative advantage over nitrite oxidizers. The fact that exogenously added nitrite was oxidized to nitrate at an equivalent dissolved oxygen concentration, but in the absence of ammonia, suggests that autotrophic denitrification is the most likely source o f measured nitrous oxide. • The average aerobic phase dissolved oxygen concentration dropped as nitrite oxidation recovered. R N A Slot Blot t ing Analysis The R N A slot blotting results are summarized in Figure 7.11. M i x e d liquor samples were probed with the Nso 190, Ntspa 454, Ntspa 685, and N b 1000 probes described in the methods section. The probing data indicated that the measured R N A concentration dropped more than four fold for all measured species between March 24, 2003 (day 16) and March 26, 2003 (day 18). A drop of this magnitude, over such a short period, suggests cell death and lysis or deflocculation o f mixed liquor solids. This author has found that pure cultures of Nitrosomonas, Nitrobacter, and Nitrospira are extremely difficult to pellet via centrifugation. This difficulty is likely the result of their very small size and possibly entrapped gases. If these organisms were to be present in the reactor in the planktonic state, following mixed liquor deflocculation, they would probably not settle out and in all likelihood be washed out o f the reactor with the effluent. Deflocculation would affect nitrification most, on a functional basis, due to the longer doubling 204 times for nitrifiers relative to heterotrophs. Cel l lysis and/or deflocculation are consistent with the observed increase in effluent suspended solids starting on day 18. The concentration of all measured R N A species remained essentially unchanged from day 18 to day 35. The data for Nso 190 probe suggests that the new operating conditions were less favorable for ammonia oxidizers or that an entirely new species of ammonia oxidizer had now become dominant. It should be noted that although nitrite oxidation recovered between day 30 and 34 there was no significant increase in the nitrite oxidizer population as measured by R N A concentration binding to the Ntspa 685, Ntspa 454, and N b 1000 R N A probes. R N A slot blotting of mixed liquor suspended solids samples collected between Day 35 and 48 indicate that the concentration of Ntspa 685 increased during this time period. These data suggest that a nitrite oxidizer, other than one of the organisms targeted by Ntspa 454, Ntspa 685, or N b 1000, may have been dominant prior to and during the time the shunt recovered. It is also possible that the number of nitrite oxidizers required to achieve complete nitrite oxidation was lower than the measurement threshold of the probing assay. FISH Analyses A limited number of samples were analyzed via fluorescent in situ hybridization (FISH). F I S H samples were collected on March 9 (day 1) and Apr i l 8, 2003 (day 31). The F I S H images are presented as Figure 7.12. These images indicate the presence o f ammonia oxidizer colonies (red) and nitrite oxidizer colonies (green) growing in relatively close proximity. 205 Discussion The control of the anoxic free ammonia concentration by adjusting the influent p H resulted in a brief (10 day) period of nitrite accumulation followed by eventual process acclimation and complete nitrite oxidation. Nitrite oxidation, albeit limited, was initiated at a free ammonia concentration far in excess of the inhibitory range reported by Anthonisen et al. (1976). B y the time the shunt had collapsed, nitrite oxidation was initiated at free ammonia concentrations exceeding 9 mg-N/L. The R N A and suspended solids data suggest that the nitrifier population was severely compromised when the feed p H adjustment was made on day 16, prior to the shunt period. The drop in R N A binding Nso 190, Ntspa 454, Ntspa 685 and N b 1000 probes after day 16, coincided with the increase in effluent suspended solids. A drop in R N A concentration of this magnitude suggests the death and lysis of cells. The drop in total R N A suggests the entire population was compromised; however, the Ntspa 454, Ntspa 685 and N b 1000 R N A concentrations were most affected. The establishment of a new suspended solids equilibrium in the reactor supports the possibility that the change in p H may have changed overall process yield, due to a reduction in the ability to produce a proton motive force and/or a nutrient limitation brought on by the high p H . A review of Turk's PhD work (1986) shows that, on several occasions when the reactor p H was increased to boost free ammonia levels, ammonia oxidation was typically lost for a couple of days. This was then followed by a recovery in ammonia oxidation and nitrite accumulation. The R N A data presented here suggest that the effect observed by Turk and Mavinic likely resulted 206 from the initial perturbation that would have affected the nitrite oxidizer population most and not from free ammonia inhibition, as reported. If the above were, indeed, the case, the length of the apparent shunt period would, in effect, be governed by the number of viable nitrite oxidizer cells relative to ammonia oxidizer cells and competitive conditions following the perturbation. The nature of the process recovery in this study suggests that the recovery was the result of the proliferation of a small population of organisms (i.e. the population reached a certain threshold above which the populations impact upon process performance could actually be measured). This is partially supported by the way in which nitrite oxidation recovered over a four-day period, with the successive doubling of effluent nitrate over consecutive days. Interestingly, complete nitrite oxidation was achieved, even though the concentrations of Ntspa 685, Ntspa 454 and N b 1000 R N A had not increased significantly; this suggests a nitrifier, not detected by either probe, became dominant or that the R N A measurement threshold was not high enough. The probing results for days 38 and 43 with Ntspa 685 support the latter conjecture. Competitive conditions within the reactor w i l l affect the recovery of nitrite oxidation. The tracking study results, as well as results from other research within the U B C laboratory, suggest autotrophic denitrification may play a role in extending the length of time required for nitrite oxidizer recovery. The ability of ammonia oxidizers to denitrify nitrite to nitrous oxide has been well established (Anderson and Levine, 1986). Hopper (1968) isolated the enzyme nitrite reductase from cells o f Nitrosomonas, demonstrating gas production with hydroxylamine as the electron donor for nitrite reduction. Poth and Focht (1985) have hypothesized that the nitrite reductase system in Nitrosomonas functions to: conserve oxygen for use by ammonia 207 monooxygenase, reduce production of nitrite (which may accumulate to toxic levels), and decrease competition for oxygen by nitrite oxidizers, by denying them their source of substrate. The natural habitat of many nitrifiers is the oxic/anoxic interface, where oxygen is often limiting; this makes the ability to denitrify nitrite to nitrous oxide advantageous. Previous works conducted as part of this research program support the possibility of competition for nitrite between ammonia and nitrite oxidizers under high free ammonia concentrations (Chapter 6 and Simm et al. 2004d). Although nitrous oxide emissions are not believed to be the cause of the decline in the nitrite oxidizer population observed here, it is believed that ammonia oxidizer denitrification was the source of the nitrous oxide and therefore increased the time required for nitrite oxidizer acclimation. It should be noted that allowing ammonia to temporarily build up in the system, following a significant perturbation, would have likely resulted in a larger disparity in ammonia and nitrite oxidizer populations; this could have magnified the impact of autotrophic denitrification and extended the length of acclimation even further. This would certainly be expected i f ammonia oxidation were lost, as reported by Turk (1986), following a process perturbation. F I S H analyses o f mixed liquor samples show that ammonia and nitrite oxidizer colonies grow in close association. Under the right conditions, ammonia oxidizers w i l l denitrify nitrite to nitrous oxide, with hydroxylamine as electron donor (Anderson and Levine 1986). The oxidation of ammonia to nitrite is carried out by two enzymes, namely ammonia monooxygenase and hydroxylamine oxidoreductase (Bock and Wagner 2003). The half saturation coefficient for ammonia oxidation decreases with increasing p H , suggesting that free ammonia is the actual substrate for ammonia oxidizers (Suzuki et al. 1974). The oxidation of free ammonia to hydroxylamine takes place in the periplasmic space (Bock and Wagner 2003), raising the possibility that the hydroxylamine concentration could 208 build within the organism under oxygen limitation. The availability o f hydroxylamine and nitrite would allow ammonia oxidizers to denitrify nitrite to nitrous oxide, in essence limiting localized substrate availability for nitrite oxidizers. The fact that measured headspace nitrous oxide concentrations peaked at a point that appeared to coincide with the change in slope of the free ammonia curve supports this hypothesis. This result also suggests that higher free ammonia concentrations would favor ammonia oxidizer denitrification since hydroxylamine was not detected during any o f the tracking studies suggesting the amount of terminal electron acceptor (whether oxygen or nitrite) was sufficient for the ammonia oxidation process. Finally, the effect of carbon loading on nitrification recovery must also be considered. Turk and Mavinic (1986, 1987, 1989a, 1989b) used a high strength synthetic feed for their work, with a complex carbon source. The C O D of their waste exceeded that required for denitrification, suggesting there was likely carbon carry over into the first aerobic cell of their plug flow reactors. Pure culture works completed as part of the present research program have confirmed that complex carbon compounds are inhibitory to Nitrospira moscoviensis. Hanaki et al. (1990) confirmed the carbon form (whether simple or complex) greatly impacts nitrite oxidizers. A n y reduction in the number of heterotrophs, as observed in this study, would increase the possibility of an increased localized carbon concentration in and around Nitrospira cultures that survived the original perturbation. The combination of carbon concentration and ammonia oxidizer denitrification would extend the length of time required for nitrite oxidizer recovery. 209 Conclusions The following conclusions have been drawn based upon this research. • The control of anoxic zone p H resulted in a short period of nitrite accumulation in a bench scale S B R . R N A slot blotting results suggest that the initial perturbation, prior to the nitrite accumulation, had a negative impact upon the entire microbial population in the reactor and that the impact was greatest upon the nitrite oxidizers. This pattern implies that the initial perturbation rather than a subsequent free ammonia inhibition is the true cause of nitrite accumulation. • The results of several tracking studies showed that nitrite oxidation was initiated at free ammonia concentrations of over 5 mg-N/L even though these concentrations are far in excess of the inhibitory range reported by Anthonisen et al. (1976). • The observed acclimation of the nitrite oxidizer population was the result of the proliferation of small populations, following the original perturbation. • There were no measurable concentrations of hydroxylamine in the reactor during the anoxic or aerobic cycles, suggesting that free hydroxylamine does not play a direct role in the observed phenomenon. 210 Back-to-back tracking studies, with and without ammonia, suggest autotrophic denitrification was the source of observed nitrous oxide emissions. The reduction of nitrite by ammonia oxidizers, although not believed to be the cause of nitrite accumulation in this particular case, likely contributed to the time required for the nitrite oxidizer acclimation to manifest itself since nitrite reduced to nitrous oxide is not available to nitrite-oxidizers. 211 Elapsed Time (days) Figure 7.1 - Peak anoxic phase free ammonia concentration K i Anoxic pH and effluent oxidized nitrogen species 8.8 8.6 8.4 X a. o 8.2 X o c < 8.0 7.8 7.6 7.4 10 15 20 25 30 Elapsed Time (days) 35 • pH • Nitrite-N -*-Nitrate-N • • • • • p • • a I • I • • • • / . • 7 \ / \ f * 1 • / • • • • U-50 45 40 35 30 25 20 15 10 5 0 CO C "o 0> Q. V) g? f l £ ; O) z E. •o Q> N 'x 40 B. SBR 3 - Effluent Ammonia vs Time (March 9 to April 13, 2003) -4*4-4. 10 15 20 Elapsed Time (days) 30 35 Figure 7.2 - Time series plots of effluent nitrite, nitrate, percentage nitrite and ammonia nitrogen SBR 3 - Reactor Solids with Time (March 9 - April 13,2003) Ol 2000 E, in f> 1500 A A A — • • < • • • • • — • — s — • • • • • s> ^ ^ Shunt Period Non Shunt Period 15 20 25 Elapsed Time (days) 35 40 Effluent Suspended Solids • • • • • Shunt Period Non Shunt • Period • > $— -5 • • • 15 20 25 Elapsed Time (days) 35 40 i E 3 System SRT a. in — • • • — * — • • — • — * • • • • • • • • • • 4 • • • • • ^ e r ^ Shunt Period Non Shunt Period 15 20 25 Elapsed Time (days) 35 40 Figure 7.3 - Time series plots of reactor and effluent mixed liquor suspended solids and system SRT. 214 * ^ <b NN <b N<3 < V N°> rV> rf> rf? $ c f > Elapsed Time (days) Figure 7.4 - Time series plot of total ammonia removal over an entire cycle. D.O. Trace Fill « amwic 45 40 35 2 30 » 25 « 2 0 B> 15 E 10 5 0 -5 Elapsed Time (minutes) Total Ammonia 70 .—. Rn • • v ? 50 O) £ 40 IS c 30 O £ 20 E < 10 n 0 50 100 150 200 Elapsed Time (minutes) Oxidized Nitrogen Elapsed Time (minutes) \-*-NitriU-Nltn>gim -O-Nltnte-N | 12 N/L] 10 D l E 8 « E 6 1 I 4 2 2 Li. 0 45 a •o 40 o 35 (A 3 30 itro ? 25 z a. « 3 20 15 a T3 10 n X 5 0 g pH Data 8.5 I 8 a. 7 5 7 50 100 150 200 Elapsed Time (minutes) Free Ammonia 50 100 150 Elapsed Time (minutes) Headspace Nitrous Oxide 50 100 150 Elapsed Time (minutes) Figure 7.5 - System performance parameters for March 30,2003 tracking study Dissolved Oxygen - « 7 -J |> 6 T 5 « 0 3 1 2 o a 1 F i l l & Anoxic Aerobic Phase Q ^ \ Tl 1 ^ " ^ , Elapsed Time (minutes) Total Ammonia 50 100 150 Elapsed Time (minutes) Oxidized Nitrogen Elapsed Time (minutes) | - 0 - Nitrite-Nitrogen -m- Nitrate-Nitrogen \ 70 © 2 60 O in 50 3 £ •=• 40 Z Q. » S 30 u ra O. 20 •D S 10 X 0 Free Ammonia 50 100 150 Elapsed Time (minutes) Headspace Nitrous Oxide 50 100 150 Elapsed Time (minutes) Figure 7.6 - System performance parameters for March 31, 2003 tracking study 217 Total Ammonia 50 100 150 Elapsed Time (minutes) Oxidized Nitrogen Elapsed Time (minutes) \—*-NrtrttoJiltrogon —g-NrtrateJtHrogen | i S 30 Free Ammonia 50 100 150 Elapsed Time (minutes) Headspace Nitrous Oxide 50 100 150 Elapsed Time (minutes) Figure 7.7 - System performance parameters for April 2,2003 tracking study 218 Dissolved Oxygen Elapsed Time (minutes) Total Ammonia 50 45 - r 40 - J Z 35 \ | 30 «T 25 O 20 I 15 E < 10 5 0 0 50 100 150 200 Elapsed Time (minutes) Oxidized Nitrogen Elapsed Time (minutes) -Nitrite-Nitrogen • -Nitrate-Nitrogen pH 90 140 Elapsed Time (minutes) Free Ammonia 50 100 150 Elapsed Time (minutes) Headspace Nitrous Oxide 50 100 150 Elapsed Time (minutes) Figure 7.8 - System performance parameters for Apri l 12, 2003 tracking study 219 OO (SBR3 Tracking Study April 22-03) Elapsed Time (minutes) Total Ammonia (SBR3 Tracking Study April 22-03) 100 200 300 400 Elapsed Time (minutes) 500 - Regular Cycle --»- Cycle with Delayed Ammonia Addition | Oxidized Nitrogen (SBR3 Tracking Study April 22-03) 000 Elapsed Time (minutes) -Nitrite Regular Cycle -Nitrate Regular Cycle - Nitrite Delayed Ammonia Addition - Nitrate Delayed Ammonia Addition pH (SBR3 Tracking Study April 22-03) 50.00 100.00 150.00 200.00 250.00 300.00 350.00 400.00 450.00 Elapsed Time (minutes) Free Ammonia (SBR3 Tracking Study April 22-03) 100 200 300 400 Elapsed Time (minutes) 500 -Regular Cycle - a-Cycle with Delayed Ammonia Addition | Nitrous Acid (SBR3 Tracking Study April 22-03) 0.000160 0.000140 0.000120 0.000100 0.000080 0.000060 0.000040 0.000020 0.000000 -0.000020 4- ^ 5 -we «ee see -tee-Elapsed Time (minutes) seo -Regular Cycle -m-Cycle with Delayed Ammonia Addition I Figure 7.9 - System performance parameters for April 22, 2003 tracking study (set 1) 220 Oxidized Nitrogen (SBR3 Tracking Study April 22-03) 5 S 4 O) 8~r 3 •4-1 —I z co 2 a ~ 1 1 o--12 W6- -949" 2 9 9 — Elapsed Time (minutes) -999-• Nitrite Delayed Ammonia Addition •Nitrate Delayed Ammonia Addition -9S0 Headspace Nitrous Oxide (SBR3 Tracking Study April 22-03) 0) T3 250 o 200 in 3 itro ? 150 z a a> Q a 100 n Q. 50 (0 •o (0 Q> X 0 50 100 150 Elapsed Time (minutes) 200 •Regular Cycle • Delayed Ammonia Addition Figure 7.10 - System performance parameters for April 22, 2003 tracking study (set 2) to to Total RNA ijig/mg solids) n j) > 31 Elapsed Time (days) 120 100 Tatai 3NA Lig/mi_, >? <> C i , s Vs 1? V '? »> V> S5 iiaguHd "Ima aavai B. SNA Conianc iiig/mg ioildSi 3> 3 0> v N * %N v> v V ^ # Elapsad Time (days) 2 -Vf o 30 *N,4 % Nso 190 and Ntsoa 454 SNA ^ . N ^ Elapsed Time idays) 1? * 3> 4 5 NSO ! 90 SNA .jiq/rtiL N^> ^ 0 £ vN fa 1? V # V> -5° -3 Nb '000 ANA [ig/mu II I J 9 -\ N \ Vs V? i> y 1? b s elapsed Tims ;days) Ntspa 4Sa RNA iigymu. )i3 ill 1 i i <? N? S £ 1? # 1> 1? * $ Elapsed Time (days) Ntspa *3S RNA Lug/mL, 1 ' 5 | 1.0 1.5 1.0 £lapsed "me daysi Figure 7. U - Summary of R.N A Data 222 A. SBRJ-March * Z1)»J (ordinal set.il) Kid • Vmini»ni;i Oxiiluern-(rrt'i'ii -NiiriU* 0 \ i i | j / i »r i . 1 1 Fig. 7.1.2 50 fwwli "'.ii nm-> B. SBRJ- Ipril S. 2003 Rt-d * V m i n i m i . i 0\i tli/L'r> GlVCM - Nilrili* (Kit i/t*rs I i 5*) 11 i s. * -1 - - ".(i nut i i ii' F g . 7.12 Figure 7.12. FISH Images. 223 Table 7.1 - Summary of Tracking Study Results. Tracking Ammonia NH 3 -N Avg. Nitrite Approximate Approximate Study oxidation Concentration aerobic oxidation [N02-N] start [HNO2-N] start rate start of end phase rate nitrite nitrite (mg/L-hr) N02-N aerobic D.O.(mg/L) (mg/L-hr) oxidation oxidation oxidation phase (mg/L) (Hg/L) March 30, 21.9 5.8 0.5 1.3 4 10.4 0.154 2003 March 31, 22.6 5.2 0.11 1.3 5.5 9.1 0.116 2003 April 2, 2003 22.6 5.45 0.5 0 4.5 12.2 0.182 April 12, 26 approx. 9 0 0.1 23.9 <3 <0.029 2003 April 22, 27 approx. 0 <0.1 22 <4.5 <0.011 2003 9.7 224 Bibliography Alexander, M . 1999. Biodegradation and bioremediation. 2 n ed. Academic Press, San Diego, California. Anderson, I.C., and Levine, J.S. 1986. Relative rates of nitric oxide and nitrous oxide production by nitrifiers, denitrifiers and nitrate respierers. App l . Environ. Biotechnol., 51(5): 938-945. Anthonisen, A . C . , Loehr, R . C . , Prakasam T.B.S. , and Srinath, E . G . 1976. Inhibition of nitrification by ammonia and nitrous acid. J. Water Pollut. Cont. Fed. 48: 835-851. Balmelle, B . , Nguyen, K . M . , Capdeville, B . , Cornier, J . C , and Deguin, A . 1992. Study of factors controlling nitrite build-up in biological processes for water nitrification. Water Sci. Technol. 26(5-6): 1017- 1025. Bartosch, S., Hartwig, C , Spieck, E . , and Bock, E . 2002. Immunological Detection of Nitrospira-like bacteria in various soils. Microbial Ecology, 43: 26-33. Bock, E . , and Wagner, M . 2003. Oxidation of inorganic nitrogen compounds as an energy source, In: The Prokaryotes an evolving electronic resource for the microbial community (Release 3.14). Springer-Verlag, Heidelberg. Cecen, F., and Ipek, S. 1998. Determination of the inhibition of ammonia-N and urea-N oxidations by the fed-batch reactor (FBR) technique. Water Sci . Technol. 38(1): 141-148. 225 Chen, S.K., Juaw, C . K . , Cheng, S.S. 1991. Nitrification and denitrification of high strength ammonium and nitrite wastewater with biofilm reactors, Water Sci . Technol. .23(7-9): 1417-1425. Daims, H . , Neilsen, P . H . , Juretschko, S., and Wagner, M . 2000. Novel Nitrospira-like bacteria as dominant nitrite-oxidizers in biofilms from wastewater treatment plants: diversity and in situ physiology. Water Sci . Technol. 41: 85-90. Eaton, A . D . , Clesceri, L .S . , Greenberg, A . E . (Editors) 1995. Standard methods for the examination of water and wastewater (19 Edition)., American Public Health Association, Inc., New York. Fdz-Polanco, F. , Villaverde, S., and Garcia, P .A . 1996. Nitrite accumulation in submerged biofilters - combined effects. Water Sci . Technol. 34(3-4): 371-378. Garrido, J . M . , van Bethum, W . A . J . , van Loosdrecht, M . C . M . , and Heijnen, J.J. 1997. Influence of dissolved oxygen concentration on nitrite accumulation in a biofilm airlift suspension reactor. Biotechnol. Bioeng. 53: 168-178. Gieseke, A . , Bjerrum, L . , Wagner, M . , and Amann R. 2003. Structure and activity of multiple nitrifying bacterial populations co-existing in a biofilm. Environ. Microbiol . 5(5): 355-369. 226 Hooper A . B . , 1968, A nitrite reducing enzyme for Nitrosomonas europaea. Preliminary characterization with hydroxylamine as electron donor, Biochim. Biophys. Acta. 162: 49-65. Hanaki, K . , Wantawin, C , and Ohgaki, S. 1990. Nitrification at low levels of dissolved oxygen with and without organic loading in a suspended-growth reactor. Water Res. 24(3): 297-302. Hovanec, T . A . and DeLong, E .F . 1996. Comparative analysis of nitrifying bacteria associated with freshwater and marine aquaria. App l . Environ. Microbiol . 62: 2888-2896. Hovanec, T . A . , Taylor, L .T . , Blakis, A . , and DeLong, E .F . 1998. Nitrospira-like bacteria associated with nitrite oxidation in freshwater aquaria. App l . Environ. Microbiol . 64: 258-264. Hyungseok Yoo , Kyu-Hong Ahn , Hyung-Jib Lee, Kwang-Hwan Lee, Youn-Jung Kwak, and Kyung-Guen Song. 1999. Nitrogen removal from synthetic wastewater by simultaneous nitrification and denitrification (SND) via nitrite in an intermittently aerated reactor. Water Res. 33(1): 145-154. Juretschko, S., Timmermann, G . , Schmid, M . , Schleifer, K . H . , Pommerening-Roser, A . , Koops, H.P., and Wagner, M . 1998. Combined molecular and conventional analyses of nitrifying bacterium in activated sludge: Nitrosococcus mobilis and Nitrospira-like bacteria as dominant populations. App l . Environ. Microbiol . 64: 3042-3051. 227 Mobarry, B . K . , Wagner, M . , Urbain, V . , Rittmann, B . E . , and Stahl, D . A . 1996. Phylogenetic probes for analyzing abundance and spatial organization of nitrifying bacteria. App l . Environ. Microbiol . 62: 2156-2162. Poth, M . and Focht, D . D . 1985. 1 5 N kinetic analysis of N2O production by Nitrosomonas europaea: an examination of nitrifier denitrification. App l . Environ. Microbiol . 49:1134-1141. Schramm, A . , De Beer, D . , Wagner, M . , Arnman, R. 1998. Identification and activities in situ of Nitrospira and Nitrospira spp. as dominant populations in a nitrifying fluidized be reactor. App l . Environ. Microbiol . 64: 3480-3485. Shiskowski, D . M . , Simm, R . A . , Mavinic, D.S. 2004. A n experimental procedure for identifying the biological source of nitrous oxide in anoxic/aerobic biological wastewater treatment systems, (accepted for publication with revisions, J. Env. Eng. Sci. January 2004). Simm, R . A . , Ramey, W . D . , and Mavinic , D.S . 2004a. A targeted study on possible free ammonia inhibition of Nitrospira. (submitted to J. Env. Eng. Sci. M a y 2004). Simm, R . A . , Ramey, W . D . , and Mavinic, D.S. 2004b. Preliminary evaluation of the use of fatty acid ratios for tracking nitrite accumulation in nitrifying reactors with low carbon to nitrogen ratio. J. Env. Eng. Sci . 3 : 31-40. 228 Simm, R . A . , Parkinson, P., Mavinic , D.S. , Ramey, W . D . 2004c. Hydroxylamine analysis of wastewater samples via gas chromatography, (submitted to Environmental Technology for publication July 2004). Simm, R . A . , Ramey, W. , and Mavinic , D.S. 2004d. Nitrifier population dynamics in a bench scale activated sludge reactor following an induced perturbation, (accepted for publication with minor revisions in Journal of Environmental Engineering and Science June 2004). Stuven, R., Vollmer, M . , and Bock, E . 1992. The impact of organic matter on nitric oxide formation by Nitrosomonas europaea. Arch. Microbiology, 158: 439-443. Suzuki, I., Dular, U . , and Kwok , S.C. 1974. Ammonia or Ammonium Ion as Substrate for Oxidation by Nitrosomonas europaea cells and extracts. J. Bacteriol. 120(1): 556-558. Turk, O. 1986. The feasibility of a shortened pathway for nitrogen removal from highly nitrogenous wastes, PhD Thesis, University of British Columbia, Vancouver. Turk, O., and Mavinic , D.S. 1986. Preliminary assessment of a shortcut in nitrogen removal from wastewater. Can. J. C i v i l Eng. 13: 600-605. Turk, O., and Mavinic , D.S . 1987. Selective inhibition: a novel concept for removing nitrogen from highly nitrogenous wastes. Env. Technol. Lett. 8: 419. 229 Turk, O., and Mavinic , D.S. 1989. Stability o f nitrite build-up in an activated sludge system. J. Water Pollut. Control Fed. 61(8): 1440-1448. Turk, O., and Mavinic , D.S. 19896. Maintaining nitrite buildup in a system acclimated to free ammonia. Water Res. 23(11): 1383-1388. Van Loosdrecht, M . C . M . , and Jetten, M . S . M . 1998. Microbial conversions in nitrogen removal. Water Sci . Technol. 38(1): 1-7. Villaverde, S., Fdz-Polanco, F., and Garcia, P . A . 2000. Nitrifying biofilm acclimation to free ammonia in submerged biofilters start-up influence. Water Res. 34(2): 602-610. Voets, J.P., Vanstaen, H . and Verstraete, W. 1975. Removal of nitrogen from highly nitrogenous wastewaters. J. Water Pollut. Control Fed. 47: 394-398. Yang, L . , and Alleman, J .E. 1992. Investigation of batchwise nitrite build-up by an enriched nitrification culture. Water Sci . Technol. 26(5-6): 997-1005. 230 8.0 The Role of Hydroxylamine as a Potential Inhibi tor of Ni t r i te Oxidizers in Wastewater Treatment Systems . Introduction and Background The work of Anthonisen et al. (1976) is considered by many to be the definitive work that established the inhibitory nature of free ammonia toward nitrite oxidizers. Several investigators have called into question the true cause of nitrite oxidizer inhibition. A s indicated by Schmidt et al. (2003): "it is unclear why nitrite oxidizers are inhibited; inhibition of nitrite oxidizers by ammonia and lower affinity for oxygen and/or nitrite have been suggested as possible explanations, but we still lack mechanistic evidence". Cecen and Ipek (1998) have suggested the dissolved oxygen to free ammonia ratio, and not the free ammonia concentration, itself is of primary importance when attempting to induce nitrite accumulation. The work o f others (Stuven et al., 1992; Yang et Alleman, 1992; and Hyungseok Yoo et al., 1999) has suggested that hydroxylamine - an intermediate in the ammonia oxidation process - may be the true cause o f nitrite oxidizer inhibition, and therefore, of nitrite accumulation. Yang and Alleman (1992) used an enriched nitrifying culture to investigate the cause of nitrite build-up in batch, by controlling dissolved oxygen, ammonia, and p H levels. They concluded that neither dissolved oxygen nor free ammonia were the primary cause of observed nitrite accumulation. They found that nitrite build-up increased with increasing p H , and thus increased in direct relation to the amounts of free hydroxylamine. They surmised that free hydroxylamine, and not free ammonia, was the likely cause of observed nitrite accumulation. This work therefore provides some theoretical 7 A version of this chapter has been submitted for publication May 2004. Simm, R.A., Mavinic, D.S., and Ramey, W.D. The Role of Hydroxylamine as a Potential Inhibitor of Nitrite Oxidizers in Wastewater Treatment Systems. J. Env. Eng. Sci. 231 basis for the conclusions of Cecen and Ipek (1998) and Hyungseok Y o o et al. (1999), who both reported that free ammonia and oxygen together appeared to play a key role in nitrite accumulation. The generally accepted reactions responsible for ammonia oxidation to nitrite are summarized as follows (Hooper, 1989; White, 1995): 2 H + + N H 3 + 2e" + 0 2 = N H 2 O H + H 2 0 N H 2 O H + H 2 0 = H N 0 2 +4H + + 4e" 2 H + + 0.5 Q 2 + 2e' = H 2 Q Total: 2 N H 3 + 3 0 2 = 2 H N 0 2 + 2 H 2 0 The cell invests reducing power into the conversion of ammonia to hydroxylamine and the reaction has a Gibbs free energy close to zero; consequently, this reaction does not result in the generation of metabolic energy (Frijlink et al., 1992). The oxidation of hydroxylamine is catalyzed by the enzyme hydroxylamine oxidoreductase, that is believed to be located in the periplasmic space. Thermodynamically, the oxidation of hydroxylamine to nitrous acid is the energy-producing part of the ammonia oxidation process. Yang (1990) hypothesized that hydroxylamine ( N H 2 O H ) might accumulate in a system with limiting amounts of oxygen; the process of hydroxylamine accumulation could be explained by the following equations: 4 H + + 2 N H 3 + 4e' + 2 0 2 = 2 N H 2 O H + 2 H 2 0 232 N F b O H + H9O = HNO9 +4H + + 4e Total: 2 N H 3 + 202 = N H 2 0 H + H N 0 2 + H 2 0 Bock and Wagner (2003) suggest there is uncertainty in where the two electrons from hydroxylamine oxidation (assuming two are returned to ammonia monooxygenase) would go, under oxygen limitation, since three other systems (nitrite reductase, cytochrome oxidase, and N A D H production) are fed with electrons from the oxidation of hydroxylamine to nitrite. Schmidt (2003) recently proposed a model of electron flow in ammonia-oxidizing bacteria, based upon literature evidence. This model is reproduced here as Fig . 8.1. Since both ammonia and hydroxylamine seem to be balanced at steady state, cytochrome C 5 5 4 is thought to be the first electron transfer branch point (Bock and Wagner, 2003). Two of the four electrons released from the hydroxylamine oxidation must pass to the monooxygenase reaction, the latter two flow to the second branch point, cytochrome C 5 5 2 , and then to one of the terminal oxidases cytochrome aa 3 or nitrite reductase. A s indicated by Wood (1986), two electrons are assumed to enter a reverse electron transfer pathway for N A D H production once per 5.7 cycles. The important point here is that the electrons can flow to different paths, with nitrite serving as an alternate electron acceptor under oxygen limitation. Anderson and Levine (1986) reported an inverse relationship between the dissolved oxygen concentration and the production of N 2 0 , when working with pure cultures of Nitrosomonas europaea. They did not observe the same relationship with N O . The more oxygen available, the less N 2 0 released. Under anaerobic conditions, they observed that the emissions of N O and N 2 0 were directly proportional to the nitrite concentration in the growth medium. They determined 233 N 2 0 emissions depend upon competition between nitrite and oxygen for electrons removed from ammonium during nitrification, consistent with Schmidt's (2003) model. Poth and Focht (1985) suggested several functional roles for ammonia-oxidizer denitrification, including the possibility it results in a decrease in competition for oxygen by nitrite oxidizers, by denying them their source of substrate. In the present research program, it was originally hypothesized that hydroxylamine, and not free ammonia, is the true inhibitor of nitrite oxidizers. Free ammonia was previously discounted as the true source o f inhibition; the supporting evidence was presented in Chapter 5 (Simm et al., 2004a). However, continuous flow studies indicated that hydroxylamine was unlikely to be the cause of sustained nitrite accumulation (Simm et al., 2004b, 2004c, 2004d). The primary objective of this work was to investigate the mechanism(s) responsible for hydroxylamine production in perturbed systems and, therefore, its possible role in nitrification phenomenon. Methodology This study was carried out with a number of batch and semi-batch tests, using mixed liquor from a continuous stirred tank reactor (CSTR) and sequencing batch reactor (SBR) systems. This work was augmented with shaker table and chemical denitrification trials. Reactor Systems The reactor system used was previously described in detail by Simm et al. (2003b, 2003c, and 2003d). The design allowed both C S T R and S B R operation and included a gas-tight headspace, 234 allowing collection o f headspace gas samples. Reactor headspace was vented through a gas collection bulb and beaker-filled with water, to allow development of a constant headspace pressure. Two bench scale sequencing batch reactor systems (SBR3 and SBR4) were used for the S B R perturbation studies. These systems were operated with an anoxic f i l l and twenty minutes anoxic and one-hundred and sixty minutes aerobic reaction time. The system solids retention time (SRT) was 10 days and the average aerobic phase dissolved oxygen concentration was maintained between 0.3 and 0.5 mg/L, in an attempt to induce hydroxylamine and/or nitrite accumulation. Synthetic Feed The C S T R was fed a synthetic feed with a low carbon to nitrogen ratio. The feed composition described in Table 8.1 was selected to provide a nitrifier-enriched culture. The S B R feed was a synthetic feed, that simulated a medium strength wastewater with moderate ammonia concentration. The S B R feed components are presented in Table 8.2. In situ Perturbation Studies Several in situ perturbation experiments were conducted with both the C S T R and S B R systems. The initial C S T R perturbations were primarily undertaken in an effort to induce hydroxylamine production. These experiments were similar to the batch experiments conducted by Yang and Alleman (1992). Whereas Yang and Alleman conducted a series of batch tests, in an effort to develop a correlation between free hydroxylamine and nitrite accumulation, a series of supplementary tests, with waste biomass outside the reactor systems as well as in situ was 235 conducted for this research, in order to elucidate the mechanism(s) responsible for the observed phenomenon. This approach was based upon the fundamental tenant that-correlation does not necessarily establish cause and effect. The in situ experiments, and the associated supplementary experiments, are summarized in Table 8.3. The objectives and basic approach of each set of supplementary experiments are discussed in the text below. Details o f the tests, as well as basic findings, are presented together in the Results section, for the sake o f continuity and ease o f reading. Prewashed Biomass Tests Prewashed biomass was used for exploratory work and initial trials. Waste mixed liquor was collected, centrifuged, and washed with distilled water. The prewashed biomass was then placed in a 1.5-liter batch test unit or bench scale reactor, identical to the C S T R and S B R . A l l reactors were equipped for gas collection. The washed biomass tests were conducted with synthetic feed devoid of carbon, to which predetermined amounts of ammonia, nitrite and/or hydroxylamine were added. The 1.5-liter, batch test unit was stirred with a magnetic stirrer and p H was controlled manually, using a dilute sodium hydroxide solution. Samples were collected using a gas tight syringe. The rationale and objectives for each test were previously presented in Table 8.3. Shaker Table Experiments -Several shaker table experiments were conducted using waste mixed liquor from the bench scale C S T R . These experiments were conducted to get an initial indication of hydroxylamine toxicity, 236 without compromising the integrity of the C S T R culture that was used as seed for several experiments. M i x e d liquors were settled, centrifuged, washed, and resuspended in distilled water amended with trace metals, bicarbonate alkalinity, phosphate buffer (as required), and substrate (either nitrite and/or hydroxylamine). Sixty milliliters of washed biomass was then placed in one of several 125 m L Erlenmeyer flasks. The flasks were placed in an incubator equipped with a reciprocating shaker table and set at 25 °C and 150 rpm. The objective(s) and details of each experimental series are presented in Table 8.4. Chemodenitrification Trials Several chemodenitrification trials were carried out in a 10.8-liter reactor, identical to the C S T R and S B R reactors. These tests were carried out with feed and with treated effluent, with and without prewashed biomass. The primary objectives of these tests were to determine the rate of hydroxylamine decomposition in the various matrices, the potential decomposition products, and the potential contribution that the chemodentrification of hydroxylamine could make to nitrous oxide emissions. Nitrous oxide emissions were found to be coincidental with nitrite accumulation in the experiments discussed here, as well as in previous work (Simm et al. 2004b, 2004c, and 2004d). The rate of hydroxylamine decomposition, and its potential contribution to the measured nitrous oxide concentration, would assist in determining whether or not hydroxylamine could decay to non-detectable concentrations during sample collection. Two tests were also carried out with and without boiled biomass, in order to determine the mechanisms responsible for biological decomposition of hydroxylamine under anoxic conditions. The objectives and details of each experimental test series are presented in Table 8.5. 237 Reactor Step Studies Three reactor step studies were carried out with the C S T R . The first two studies were conducted at p H 7.5 and 8.5, with various nitrite addition rates. The reactor was taken off line and fed a concentrated nitrite solution at 0.25, 2.5, and 5 mg-N/min (each nitrite addition rate was sustained for 1 to 2 hours). The first two-step studies were conducted to confirm the impact of p H on the nitrite oxidation rate, as well as to quantify the size of the nitrite oxidizer enzyme pool. A third step study was conducted with nitrite and hydroxylamine additions. The reactor was fed with the nitrite solution of 2.5 mg-N/min for the entire test duration; after two hours, a hydroxylamine solution was added. The hydroxylamine addition rate (1 mg-N/min) was selected based upon the results o f the chemodenitrification trials, to ensure a steady increase in reactor hydroxylamine concentration. Hydroxylamine, nitrite, and nitrate samples were collected every ten to fifteen minutes. Nitrous oxide samples were collected continuously. The hydroxylamine concentration, at which the nitrite oxidation rate changed, was deemed to be inhibitory. Chemical Analyses The analysis of ammonia, total Kjeldahl nitrogen ( T K N ) , nitrate, nitrite, total organic carbon, and total dissolved solids was carried out as described in the 19 t h Edition of Standard Methods for the Examination of Water and Wastewaters (Eaton et al. 1995). The free ammonia and nitrous acid concentrations were estimated using the relationships presented by Anthonisen et al. (1976). Hydroxylamine was measured using the G C method described in Chapter 2 (Simm et al. 2004e). The free hydroxylamine concentration was estimated using the relationship presented in Yang and Alleman (1992). 238 Fatty Acid Methyl Esters (FAME) and RNA Slot Blotting Microbial communities, in mixed liquor suspended solids, were characterized using fatty acid methyl esters ( F A M E ) and R N A slot blotting. The methodology employed is described in Chapter 3 (Simm et al. 2004b). N 2 0 The headspace of the reactors and test unit was fully enclosed, thus allowing for the collection o f reactor off-gases. Samples were collected using a gas-tight syringe and analyzed for nitrous oxide by gas chromatography, using a Hewlett-Packard 5880A equipped with an electron capture detector. In the G C analysis, nitrogen was used as the carrier gas (at 20 mL/min), with a column packing material of Haycep C . The injector, oven, and detector temperatures were 80 °C, 80 °C, and 250 °C, respectively. The headspace nitrous oxide concentration was also monitored continuously during the chemodentrification trials using a Bacharach™ (Model 3030) nitrous oxide analyzer. Nitric Oxide (NO) Off-gas samples were collected for nitric oxide analysis, using a gas tight syringe. The collected samples were immediately injected into a Sievers Model 280i Nitr ic Oxide Analyzer (NOA™). This instrument uses a high-sensitivity detector for measuring nitric oxide, based on a gas-phase chemiluminescent reaction between nitric oxide and ozone. A photomultiplier tube is used to 239 detect the photon emissions from the activated nitrogen dioxide produced. The detection limit of the N O A , for measurement of gas-phase N O , is approximately 0.5 ppbv. Results Several in situ batch tests were conducted with the C S T R (referred to as R2) and the SBRs (referred to as SBR3 and SBR4) . Studies using C S T R Biomass Overall, performance information for Reactor R2 is presented in detail in Chapter 5 (Simm et al., 2004a). R N A slot blotting data from Chapter 5 indicated that Nitrospira-like organisms were the dominant nitrite oxidizers in the test reactor. On average, these organisms accounted for approximately 11.3 % +/- 8% (two standard deviations) o f the total R N A in this community. This fact was also supported by the F A M E analysis, based upon the large proportion of cis(l 1) hexadecenoic acid relative to cis(9) hexadecenoic acid. The potential utility o f using fatty acid analyses to study C S T R communities was discussed in Chapter 3 (Simm et al., 2004b). In situ Batch Tests wi th a C S T R Two in situ batch tests were conducted at high and low dissolved oxygen concentrations, using reactor R2. The first in situ batch test was carried out on September 6, 2002. A n ammonia spike, of approximately 40 mg-N/L, was added to the reactor after terminating the influent feed. The reactor p H was adjusted to an initial value of approximately 8.4, in order to obtain the desired free ammonia stress, and the p H was allowed to drop following the spike. The aeration 240 rate was adjusted until a constant dissolved oxygen concentration of approximately 3 mg/L (aeration rate o f 1.6 liters/min.) was obtained; the test was terminated when the exhaustion of ammonia substrate caused the reactor dissolved oxygen concentration to increase rapidly. A n inhibitory, free-ammonia stress was maintained in the reactor for approximately the first eighty minutes of the test (NH3 -N 0.1 to 3.5 mg-N/L) . The combination of nitrate, nitrite, and emitted nitrous oxide accounted for approximately 82% of the oxidized ammonia. The delay in nitrite oxidation and characteristic accumulation of nitrite that was reported to be characteristic of free ammonia inhibition (Anthonisen et al., 1976) was not observed for this test, although nitrite did account for approximately thirty-six percent of the oxidized nitrogen species at its peak concentration. The nitrite concentration continued to build as the test progressed and peaked at the time where free ammonia disappeared. The headspace concentration of emitted nitrous oxide increased as the test progressed. N o measurable concentrations of hydroxylamine were obtained during this test. A second in situ batch test was carried out on September 11, 2002. The test conditions were identical to the September 6, 2002 test, with the exception that the dissolved oxygen concentration was controlled at less than 0.5 mg/L. This test was done to determine whether a lower dissolved oxygen concentration could induce nitrite accumulation, as reported by Cecen and Ipek (1998), and whether a measurable hydroxylamine concentration would be produced. The ammonia oxidation rate was severely impaired at the lower dissolved oxygen concentration. The measured ammonia oxidation rate for the low dissolved oxygen test was only 2.6 mg/L-hr, versus 16 mg/L-hr for the high dissolved oxygen test. This test, unlike other tests conducted, produced a measurable concentration of hydroxylamine (55 ppb as NH2OH - N ) . The analysis of 241 these tests was complicated by a significant variation in the measured nitrate-nitrogen concentration during the test duration. Given the low ammonia oxidation rate, and the fact that the measured nitrite and nitrous oxide concentrations accounted for approximately 90% of the oxidized ammonia, makes it unlikely that significant nitrite oxidation actually took place. The fact that ammonia oxidation was limited, and that heterotrophic denitrification may have taken place, makes it difficult to rule out oxygen limitation as the cause. In addition, it was not clear i f hydroxylamine disappeared as a result of a drop in free ammonia or an increase in nitrous acid concentration. The low dissolved oxygen test was repeated on November 22, 2002, using a higher free ammonia concentration and p H control. The test results are summarized in Fig. 8.2. A n ammonia spike of 75 mg/L was added to the reactor, to give a free ammonia concentration of approximately 8.5 mg-N/L . This concentration is far in excess of the inhibitory concentration reported by Anthonisen et al. (1976). The reactor p H was controlled between 8.2 and 8.4 for the first 285 minutes of the test (Fig. 8.2B). During this time, there was a progressive increase in the hydroxylamine concentration (Fig. 8.2H). This increase was accompanied by an increase in both headspace nitrous oxide (Fig. 8.2G) and nitric oxide (Fig. 8.21) concentrations. The nitrous acid (Fig. 8.2J) concentration remained relatively low (< 0.3 ug/L) during this time period. Nitrite accounted for only 60% of the reduction in the initial ammonia-nitrogen concentration during the first 285 minutes of the test. Hydroxylamine and nitrous and nitric oxide emissions accounted for the remainder. A linear regression (data not presented) analysis of free ammonia versus hydroxylamine concentration for this time period indicated a strong correlation between the two (R = 0.83). N o such correlation was observed after p H control was abandoned. This was not unexpected, since free ammonia is the substrate for the ammonia monooxygenase enzyme that 242 produces hydroxylamine. Regression analyses of headspace nitrous oxide concentration versus both hydroxylamine and nitrite-nitrogen indicated r-squared values of 0.97 and 0.96, respectively. The measured ammonia oxidation rate during the first 285 minutes of the test was higher, at 6.2 mg/L-hr. The reactor nitrate concentration dropped by approximately 5% during the first 285 minutes o f the test, from a starting concentration of about 164 mg/L to 155 mg/L (Fig. 8.2E). The bulk of the decline (approximately 5 mg/L) can be accounted for by dilution, from the addition of dilute sodium hydroxide solution used for p H adjustment; this suggests there was little or no nitrite oxidation during this time period and the nitrite-oxidizer population was being inhibited. The reactor nitrate-nitrogen did not begin to increase again until after minute 410, a full two hours after p H control was abandoned (even though the dissolved oxygen concentration had started to increase (Fig. 8.2A)). The rate of nitrite-nitrogen accumulation also started declining at this point (Fig. 8.2F). The increase in nitrate concentration was coincidental with the disappearance of hydroxylamine and drop in both the free ammonia (Fig. 8.2D) and headspace nitrous oxide (Fig. 8.2G) concentrations. Both the nitrous (Fig. 8.2G) and nitric oxide (Fig. 8.21) emissions began to drop once p H control was abandoned. The reactor dissolved oxygen concentration (Fig. 8.2A) started to increase at approximately minute 100, with a further increase in this rate following the abandonment of p H control (Fig. 8.2A). N o attempt was made to control the dissolved oxygen level following the abandonment of p H adjustment. This would suggest a persistent change in the microbial population, a switch in terminal electron acceptor, and/or a reduction in the concentration of oxygen demanding material. The fact the hydroxylamine concentration began declining suggests 243 that the production rate had declined or that it had stopped being produced all together. The ammonia-nitrogen concentration dropped 12.4 mg/L between minute 285 and minute 497, during which time the hydroxylamine concentration dropped to zero. The drop in ammonia concentration was matched by a corresponding 12.3 mg/L increase in the nitrate plus nitrite-nitrogen concentration, suggesting that the production of excess hydroxylamine had essentially stopped. A n additional 88.5 mg of nitrous oxide as nitrogen was emitted from the reactor during this time period (between minute 285 and minute 497) and measured hydroxylamine accounted for 21.5 mg, or approximately twenty-five percent of this total. The remainder was likely sparged liquid nitrous oxide. The re-establishment of nitrite oxidation after minute 410 was coincidental with the drop in p H and associated increase in nitrous acid concentration, the drop in both free ammonia and hydroxylamine concentrations, as well as a subsequent drop in both headspace nitrous and nitric oxide concentrations. Previous experiments (Chapter 5 and Simm et al., 2004a) had indicated free ammonia is unlikely to be inhibitory to Nitrospira-like organisms at the concentrations used here. In addition, nitrite step studies, in which nitrite solution was fed into this reactor in situ at various rates under two different p H regimes (pH 7.5 and 8.5), indicated the p H used in the November 22, 2002 experiment was unlikely to be the primary cause of impaired nitrite oxidation. These data suggest the drop in hydroxylamine, nitric oxide, and/or nitrous oxide concentration may well be associated with the re-establishment of nitrite oxidation. N o measurable concentrations of hydroxylamine were recovered in continuous flow systems that exhibited nitrite accumulation (Chapter 3 (Simm et al., 2004b), Chapter 6 (Simm et al. 2004c), and Chapter 7 (Simm et al. 2004d), suggesting hydroxylamine was unlikely the cause of the 244 observed phenomena in these experiments. However, nitrous oxide emissions were coincidental with nitrite accumulation. Nitrous oxide is one of the known autodecomposition products of hydroxylamine. A series of supplementary experiments were carried out to confirm the source of hydroxylamine as being ammonia-oxidizing organisms, the potential contribution "of hydroxylamine autodecomposition and/or its reaction with nitrous acid towards nitrous oxide generation, the rate of hydroxylamine decomposition in various matrices, as well as the toxicity of hydroxylamine towards nitrite-oxidizing organisms. Confirmation that ammonia-oxidizer organisms were the likely source of measured hydroxylamine would provide insights into the physiological state of this population, and therefore of reactor conditions at the colony level. If the source of nitrous oxide was indeed chemodenitrification of hydroxylamine, and measured hydroxylamine levels were proven to be inhibitory, then the presence of nitrous oxide could indicate hydroxylamine is present locally inside a floe particle, even though it might not be measured in the bulk liquid. Nitrous and nitric oxide are generated by both heterotrophic and autotrophic denitrification. Partial heterotrophic denitrification of nitrate to nitrite would result in nitrite accumulation. Autotrophic denitrification of nitrous acid to nitric oxide and/or nitrous oxide by ammonia-oxidizers would result in a localized reduction in substrate concentration for nitrite-oxidizers and potentially nitrite accumulation. Therefore, i f the source of measured nitrous oxide was, in fact, biological, it was important to confirm whether it was the result of heterotrophic and/or autotrophic denitrification. 245 Confirmation That Ammonia Oxidation was the Source of Measured Hydroxylamine The tests conducted on September 11 and November 22, 2002 produced measurable quantities of hydroxylamine. The possibility exists that ammonia oxidation was not the source of hydroxylamine production and that cell rupture or possibly hydroxylamine contamination of the sodium hydroxide used for p H control might be the true source. Therefore, several supplementary experiments were conducted to confirm the previous results. Dilute sodium hydroxide from the same batch used in the November 22, 2002 experiment was added to treated effluent from reactor R2, in an equal proportion to the amount added to the reactor during the test. In addition, mixed liquor samples taken from reactor R2, SBR4 , and another C S T R being operated in the laboratory were subjected to sonication, bead beating, and rapid p H increase, to induce individual cell rupture. The supernatant from all tests was filtered through a 0.45 pm filter, derivatized and analyzed for hydroxylamine. These tests indicated that neither cell rupture nor hydroxylamine contamination of the N a O H used for p H control were the source of hydroxylamine in these tests. Two, small-scale, batch tests were then conducted, using waste mixed liquor from reactor R2. These tests were conducted in a 1.5-liter test unit. A n ammonium bicarbonate spike was added at the start of each test and samples were collected for ammonia, nitrite, nitrate, and hydroxylamine. Headspace gas samples were collected from the test unit and analyzed for nitrous oxide. The first test was used as confirmation o f hydroxylamine production. A predetermined concencentration of allylthiourea ( A T U ) , a known inhibitor of the ammonia monooxygenase enzyme, was added to the second test to confirm the impact upon 246 hydroxylamine production. The concentration of hydroxylamine dropped significantly following A T U addition (data not presented), confirming ammonia oxidation as the likely source of hydroxylamine production. Chemodenitrification and Autotrophic Denitrification Trials with Hydroxylamine The results of chemodenitrification and autotrophic denitrification trials with hydroxylamine are summarized in Table 8.6. Chemodenitrification trials 8 and 9 (Table 8.6) were conducted with live biomass samples taken from the C S T R . The biomass was spiked with hydroxylamine (as hydroxylamine hydrochloride) and hydroxylamine decay rate and nitrous oxide generation were measured. Tests 8 and 9 were conducted under high and low dissolved oxygen conditions, respectively. The initial spike concentration was approximately twice the maximum measured concentration in the November 22, 2002 batch test. The maximum nitrous oxide mass emission rate from tests 8 and 9 were approximately 0.046 and 0.0046 mg-N/min, respectively. The maximum mass emission rate for the November 22, 2002 test was 0.46 mg-N/min, at approximately one half the maximum hydroxylamine concentration. These data suggest it is unlikely that hydroxylamine decomposition (whether biological or chemical) was the primary source o f measured nitrous oxide. The soil science literature suggests that hydroxylamine w i l l react with nitrous acid (HNO2) to form nitrous oxide (Arnold, 1954). This reaction could result in a local reduction in nitrite oxidizer substrate. Both hydroxylamine and nitrite spikes were added in chemodenitrification 247 trials 1, 2, and 7. Tests 1 and 2 were conducted in the presence of live and boiled biomass, respectively. Test 7 was conducted in reactor feed. A l l tests were carried out at high p H (8.3-8.6) with nitrogen sparging. The reactor was sparged with nitrogen to eliminate oxygen from the reactor and allow only the reaction of hydroxylamine and nitrite to be studied. There was no significant change in nitrite concentration during any of these three tests, suggesting a reaction between hydroxylamine and nitrite was unlikely to be the source o f measured nitrous oxide during the November 22, 2002 test. According to Chalk and Smith (1983), the reaction between nitrite and hydroxylamine is more likely to occur under acidic or neutral conditions. This is consistent with the findings presented here. The measured hydroxylamine decomposition rate in the November 22, 2002 batch test, following the abandonment o f p H adjustment, was 1.12 mg-N/L-hr. This rate is of the same order of magnitude as the average decomposition rates for tests 8 (1.6 mg-N/L-hr) and 9 (1.1 mg-N/L-hr), thus supporting the conclusion that significant hydroxylamine production was unlikely, following the abandonment o f p H control on November 22, 2002. The measured hydroxylamine decomposition rates reported in Table 8.6 range from 0.1 to 2.6 mg-N/L-hr. In general, reported rates are higher in biomass and increase with increasing dissolved oxygen concentration. The results of test 10 (neutral pH) suggest the rate is significantly lower at neutral p H values. The results presented suggest that the reactive nature of hydroxylamine is not a significant barrier to measurement in the experimental matrices studied here and, taken together with the mass emission data for nitrous oxide, these data suggest localized areas of high hydroxylamine concentration are unlikely to be the source o f nitrite oxidizer inhibition in previous experiments. 248 The results of the chemodenitrification tests suggest biological denitrification (whether heterotrophic or autotrophic) as the most likely source of measured nitrous oxide in the November 22, 2002 batch test. Previous tests conducted in the 1.5-liter test unit with waste biomass under nitrogen sparging, with and without mercuric chloride addition, indicated no apparent reaction between hydroxylamine and nitrite. These tests also suggested the chemical decomposition of hydroxylamine as the source of measured nitrous oxide, since there was no significant difference in nitrous oxide generation or hydroxylamine decay rate between tests with and without mercuric chloride addition. There was concern that the mercury may have killed the bacteria but not altered enzyme activity. For this reason, it was decided to repeat these experiments using heat-killed biomass (trials 1 and 2 in Table 8.6). Added nitrite was not consumed in trial 1 or trial 2. Both trials resulted in hydroxylamine consumption; however, the consumption rate was approximately three times higher in the live biomass test (trial 1) relative to the killed biomass test (trial 2). A l l of the hydroxylamine consumed in trial 1 could be accounted for as emitted nitrous oxide. In trial 2, emitted nitrous oxide accounted for less than 30% of hydroxylamine consumption. These data suggest that hydroxylamine was either biologically denitrified directly to nitrous oxide or that the physical act o f boiling modified the biomass surface sufficiently to change both the rate and stoichiometry of hydroxylamine decomposition. Shaker Table Experiments and Substrate Step Studies Hydroxylamine toxicity was studied via a combination of shaker table experiments and substrate step studies. The shaker table test series was previously presented in Table 8.4. Test series 2 indicated no significant difference in nitrite oxidation rate between a control and experimental 249 flask having an initial hydroxylamine concentration of 1.95 mg-N/L (99% as free hydroxylamine). There was a 50% reduction in nitrite oxidation between the control and the flask, with an initial hydroxylamine concentration of 18.6 mg-N/L (99% as free hydroxylamine). This test was repeated with a wider range of initial hydroxylamine concentrations as part of test series 3 (Table 8.4). The test was conducted at p H 8; therefore, virtually all of the hydroxylamine is in the free form. The results of test series 3 are summarized as Fig. 8.3. The results of test series 3 (Fig. 8.3) indicate no significant difference in nitrite oxidation rate, relative to the control, up to a hydroxylamine concentration of 4.2 mg-N/L . The maximum measured concentration during the November 22, 2002 batch test was 2.3 mg-N/L (at p H 8.2-8.4 essentially all o f the measured hydroxylamine would be present as free hydroxylamine). The shaker table experiments suggested the measured hydroxylamine concentration during the November 22, 2002 test was unlikely the cause o f observed nitrite accumulation. However, the initial hydroxylamine concentrations would not have been sustained, since hydroxylamine decomposition would have taken place during the test duration. The results from the shaker table and chemodenitrification experiments were used to design a supplementary experiment, to characterize the significance of the measured hydroxylamine concentration on November 22, 2002. The reactor step studies indicated no significant difference in nitrite oxidation rate between p H values 7.5 and 8.5, suggesting that p H was not a contributing factor in measured nitrite accumulation during the November 22, 2002 test. The maximum measured rate for both tests was approximately 3 mg-N/min. This rate is considered a direct measurement of the size of the nitrite oxidizer enzyme pool. 250 A third step study was carried out with both nitrite and hydroxylamine additions. The results of step study 3 are presented in Fig . 8.4. These data indicate that the nitrite oxidation rate began to change at a hydroxylamine concentration of 2.5 mg-N/L (Fig. 8.4B). The hydroxylamine feed was stopped soon after this point; however, the nitrite oxidation rate had been reduced by 30% from this point onward, indicating the nitrite-oxidizer population had been irreparably damaged. The reactor was placed back on line following the test; however, both ammonia and nitrite oxidation were severely impaired for two days (more than 50% decline in ammonia oxidation rate 24 hours later). This test confirmed that the level of hydroxylamine measured on November 22, 2002 was likely toxic to both the ammonia and nitrite oxidizer populations. However, the measured concentration did not appear to be high enough to result in a full cessation of nitrite-oxidize activity as previously observed. Therefore, it is unlikely that hydroxylamine alone could fully account for the observed results on November 22, 2002. Interestingly, the difference between the added and measured hydroxylamine concentrations (i.e. unaccounted hydroxylamine) was accounted for as nitrite. This suggests the majority of the residual hydroxylamine was oxidized to nitrite, most of which was not oxidized by the nitrite-oxidizer population (Fig. 8.4D). Confirmation of Autotrophic Denitrification as a Source of Nitrous Oxide and Nitr i te/Oxygen Dynamics in a M i x e d M i c r o b i a l Populat ion. The 10.8-liter reactor was used for a large-scale, washed biomass test. This test was conducted to investigate the primary source of measured nitrous oxide (since it was shown that chemodenitrification of hydroxylamine was unlikely to be the primary source), to ascertain the role of nitrite in hydroxylamine formation, and to study the competition for oxygen between 251 ammonia and nitrite oxidizers. Normally, free ammonia inhibition studies are carried out with a free ammonia spike, with nitrite becoming available to nitrite oxidizers as quickly as ammonia oxidizers can make it available. Waste mixed liquor was collected from reactor R2 over a one-week period. The waste mixed liquor was settled, supernatant decanted, washed twice with distilled water and then refrigerated until use. The mixed liquor was then placed in an empty reactor, identical to R2, and filled with feed devoid of carbon and ammonia. The reactor contents were then aerated for twelve hours, to remove any stored carbon from the biomass. The oxygen was then sparged from the reactor using nitrogen and the following treatments were applied in succession: sodium nitrite was added to give a nitrite-nitrogen concentration of approximately 30 mg-N/L; an ammonium-nitrogen spike was then added; aeration was commenced and dissolved oxygen was controlled between 0.3 and 0.6 mg/L (previously shown to result in nitrous oxide emissions); the p H was increased from approximately 7.5 to 8.3; and, a second ammonia spike was added to give a final free ammonia concentration of approximately 10 mg-N/L. The reactor was sampled for headspace nitrous oxide and nitric oxide, hydroxylamine, nitrite, nitrate, and ammonia. A low dissolved oxygen concentration was selected in an attempt to induce hydroxylamine accumulation and study competition for oxygen between ammonia and nitrite oxidizers. The reactor was sparged with nitrogen for approximately forty minutes, following the addition of the nitrite spike. The concentration of nitrous oxide was less than 2 ppm and remained essentially constant during this time period (Fig. 8.5). After forty minutes, an ammonia spike of 50 mg-N/L (approximate free ammonia concentration of 1 mg-N/L) was added to the reactor, under nitrogen sparging. This resulted in a slow and steady increase in head-space nitrous oxide (Fig. 8.5). This further supported the possibility that autotrophic 252 denitrification was the source of measured nitrous oxide in previous tests. The increase in nitrous oxide concentration was accompanied by a measurable decline in nitrite-nitrogen concentration. The nitrogen sparging was stopped and the reactor contents were aerated approximately one hour after the initial ammonia addition. The measured nitrate-nitrogen accounted for essentially all o f the oxidized nitrite and ammonia. This time period was accompanied by a leveling off in the headspace nitrous oxide concentration (Fig. 8.5). The p H increase from 7.5 to 8.2 (approximately 2 hours after the start of aeration) increased the reactor free ammonia concentration by approximately four-fold,, from approximately 0.7 mg-N/L to 3 mg-N/L. There was a dramatic increase in headspace nitrous oxide concentration (Fig. 8.5) following the p H adjustment; this was accompanied by a decline in the measured rate of nitrite oxidation (rate decline from 9.8 mg/L-hr to 6 mg/L-hr). A second ammonia spike was added at minute 580 approximately one hour after the p H adjustment. The second spike raised the free ammonia from 3 to approximately 10 mg-N/L; this resulted in a further increase in nitrous oxide generation rate (Fig. 8.5) and a decline in nitrite oxidation rate (decline from 6 mg/L-hr to 4.2 mg/L-hr). There was only a trace amount of hydroxylamine measured in bulk liquid samples following the second ammonia spike (< 0.01 mg/L). The measured ammonia oxidation rate was 1.2 mg/L-hr following aeration. This occurred at a free ammonia concentration of approximately 1 mg-N/L. This test suggests that the nitrite oxidizer population had a higher affinity for oxygen than ammonia oxidizers, since the nitrite oxidation rate was almost ten times higher than the ammonia oxidation rate. The ammonia oxidation rate had essentially doubled, once the free ammonia concentration had increased to 10 mg-N/L. This increase was accompanied by a significant decline (essentially 60%) in nitrite-253 oxidation rate and a sharp increase in the nitrous oxide generation rate. This test suggested that nitrous oxide emissions increase with increasing free ammonia concentration and the autotrophic denitrification was likely a primary source of measured nitrous oxide in the November 22, 2002 test. The rate of nitrite oxidation appeared to decrease with increasing nitrous oxide concentration, suggesting that nitrite substrate was being diverted from nitrite-oxidizing to ammonia-oxidizing organisms. SBR Perturbation Studies The experimental results indicated that nitrous oxide generation was coincidental with nitrite accumulation. Supplementary testing suggested that it was unlikely that hydroxylamine was the primary cause of nitrite accumulation or observed nitrous oxide emissions. In addition, the prewashed biomass test with NO2 and NH3 spikes and low dissolved oxygen trials with a C S T R (Simm et al., 2004b, 2004c), suggested hydroxylamine formation is unlikely, in the presence of a high nitrous acid concentration. The experimental evidence suggested that the presence of hydroxylamine in batch systems exhibiting nitrite accumulation, as reported by Yang and Alleman (1992), might in fact be coincidental, as opposed to the primary cause of the observed nitrite accumulation. A series of S B R perturbations were conducted as part of an investigation to elicit the mechanism(s) responsible for hydroxylamine production in mixed microbial cultures. A n S B R system was considered an ideal candidate for such experiments, based upon the general observation that systems operating under a substrate gradient are more likely to sustain nitrite accumulation. The perturbations included lowering the aeration rate during the aerobic phase, 254 doubling the ammonia load during the anoxic phase, adjusting the anoxic p H upwards to increase the anoxic free ammonia concentration, and controlling the aerobic p H in an effort to limit nitrite oxidation by limiting the nitrous oxide concentration. The results of the S B R perturbation studies are summarized in Table 8.7 and discussed in the text below. Anoxic/aerobic tracking studies were conducted once the two SBRs had appeared to reach steady state. The purpose of this first set of tracking studies was to characterize the time course of nitrification/denitrification, when complete nitrification (i.e. oxidation to an nitrate end point) occurred. This first set of tracking studies was carried out on August 27, September 4, and September 17, 2002. The results for the September 17, 2002 tracking study, summarized in Table 8.7 (Trial 1) and presented in Fig . 8.6, are representative of pre-perturbation conditions. The concentrations of aerobic phase nitrite and headspace nitrous oxide were less than 1 mg/L and 14 ppm, respectively, on September 17, 2002. The first perturbation (Trial 2, Table 8.7) involved a reduction in aerobic phase aeration rate, to encourage hydroxylamine and/or nitrite accumulation. In this perturbation study, the airflow rate for the aerobic phase was reduced from 957 mL/min to 537 mL/min . The change caused dissolved oxygen concentration to drop by approximately one half compared with previous measurements. The ammonia oxidation rate decreased to 7 mg/L-hr (base case ammonia oxidation rate was 11 mg/L-hr) and ammonia oxidation was only seventy-five percent complete at the end of the aerobic phase. The nitrous oxide concentration had not yet peaked by the end o f the aerobic phase. In addition, there was no measurable concentration of hydroxylamine or any significant nitrite accumulation relative to the base case; this is reflected in the fact that the nitrite oxidation rate essentially equaled the ammonia oxidation rate. 255 A n ammonia spike o f approximately 29 mg NH4-N/L was added to S B R 3 at the start of the anoxic phase (Trial 3, Table 8.7), essentially doubling the steady state ammonia load. The purpose of this study was to determine whether or not hydroxylamine and/or nitrite accumulation could be encouraged by an increase in the influent ammonia. The characteristic dissolved oxygen breakthrough during the aerobic phase (indicating the disappearance of ammonia) was not observed. Although the additional ammonia remained essentially unoxidized at the end of the aerobic phase, the rates of ammonia and nitrite oxidation were essentially unchanged, relative to the base case. The chromatograms for the hydroxylamine analysis indicated a trace of hydroxylamine was present, approximately 20 minutes into the aerobic period; interestingly, nitrous oxide also appeared in the reactor headspace earlier than in previous tracking studies reaching a maximum concentration of 84 ppm (compared to only 14 ppm on September 17, 2002). The characteristic decline in the headspace nitrous oxide concentration was not observed for this run (typically observed just prior to the dissolved oxygen breakthrough and coincidental with the disappearance o f ammonia). There was no significant difference in nitrite accumulation in this run, relative to the base case. The fourth perturbation involved an increase in anoxic p H in S B R 4, from approximately 8.3 to 8.8, using sodium hydroxide, and the p H was allowed to fall as nitrification took hold (Trial 4, Table 8.7). The intent here was to raise the anoxic free ammonia concentration to determine whether or not an increase would impair downstream nitrite oxidation. This was the method commonly used by Turk and Mavinic (1986, 1987, 1989a, 1989b) to»induce nitrite accumulation in their plug flow reactor system. The maximum anoxic free ammonia concentration during this run was 5.5 mg-N/L. The perturbation did not result in any significant aerobic accumulation of 256 nitrite. It did, however, appear to reduce the ammonia oxidation rate at the start of the aerobic phase, when a measurable concentration of hydroxylamine was recovered (Trial 4, Table 8.7). The interesting point about this test is that hydroxylamine was measured at the start of the aerobic phase, following the anoxic p H adjustment. The overall rates of ammonia and nitrite oxidation remained unchanged relative to the base case. A second p H adjustment perturbation was conducted for trial 5 (Table 8.7). However, p H control was applied throughout the anoxic cycle and throughout the first 110 minutes of the aerobic cycle. The p H was manually controlled between 8.5 and 8.8, using a dilute sodium hydroxide solution. The intent was to confirm that hydroxylamine production could be sustained with aerobic p H control. A measurable quantity of hydroxylamine was observed during the time of p H control. Although the maximum concentration of hydroxylamine was lower than that observed for trial 4 (likely due to the lower free ammonia concentration of 4.4 mg-N/L) , the length of time over which hydroxylamine was measured was longer, coinciding with the longer period of p H adjustment. Again, the presence of hydroxylamine was coincidental with a reduction in the ammonia oxidation rate. Trials 4 and 5 produced measurable quantities of hydroxylamine, although neither test resulted in significant nitrite accumulation (as a percentage of total N O x ) or nitrous oxide emissions. The concentration of nitrous oxide in the headspace was low relative to the measurements taken during the initial tracking studies. The most important result is that hydroxylamine production was associated with reduced ammonia oxidation and nitrous oxide production rates. 257 A combination of ammonia spike and anoxic/aerobic p H control was applied for trial 6 (Table 8.7). The results of trial 6 are presented as Fig . 8.7. The ammonia spike doubled the steady-state ammonia loading in the anoxic phase, resulting in a maximum anoxic free ammonia concentration of approximately 7.5 mg-N/L (Fig. 8.7C). This perturbation resulted in an even higher measured concentration of hydroxylamine (Fig. 8.71), relative to trials 4 and 5 (approximately 300 ppb (127 ppb as N ) compared to 80 to 100 ppb), as well as a much higher proportion of nitrite (Fig. 8.7D) at the end of the aerobic period (30% for trial 6 versus 0% for trial 4 and approximately 17% for trial 5). What was surprising is that there was a significant increase in the headspace nitrous oxide concentration (Fig. 8.7G), once p H control was abandoned and the p H started to drop. Approximately 12 mg o f nitrous oxide (as nitrogen) was emitted in the reactor headspace between the time when p H control was abandoned and the end of the cycle. This is equivalent to a reactor hydroxylamine concentration of approximately 1 mg-N / L or seven times the maximum measured concentration. The actual reduction in measured hydroxylamine concentration during this time period was only 54 ppb as N . This result suggests, that either hydroxylamine continued to be produced and subsequently autodecomposed to nitrous oxide at the lower pH', or that the rapid "increase in nitrous acid concentration, coupled with available ammonia, allowed ammonia oxidizers to denitrify nitrous acid to nitrous acid. Autotrophic denitrification appears to be the most probable explanation, based upon the hydroxylamine data presented in Table 8.6. The S B R perturbation studies demonstrate several important points, one of them being that the nitrite oxidation rate was virtually equal to the ammonia oxidation rate, even at free ammonia concentrations in excess of 7.5 mg-N/L. This suggests that nitrite oxidizers essentially oxidized nitrite as quickly as the ammonia oxidizer produced it. This differs from the results of the March 258 29, 2003 batch test, whereby the nitrite oxidation rate declined relative to the ammonia oxidation rate, at higher free ammonia concentrations. This decline was coincidental with an increase in nitrous oxide and virtually no recovered hydroxylamine, suggesting that ammonia oxidizers were denitrifying nitrite to nitrous oxide. The results from trial 3 (doubling of the ammonia loading) suggest that the ammonia oxidation rate is dependant upon the size of the given enzyme pool and available concentration o f terminal electron acceptors. This is significantly different from what one would expect with heterotrophs, since an increased loading would have likely resulted in a change from first order to zero order kinetics. The combination of increased free ammonia and p H adjustment resulted in the recovery of a measurable concentration of hydroxylamine. This was coincidental with reduced ammonia oxidation and nitrous oxide production rates. The reduction in nitrous oxide emissions for trials 4 through 6 suggest that hydroxylamine was recovered, as opposed to autodecomposing, or the nitrous acid concentration was insufficient for autotrophic denitrification. The result from trial 6 supports autotrophic denitrification as the most likely source of nitrous oxide. In addition, one would not expect hydroxylamine to be expelled from the cell, unless the nitrous acid concentration was insufficient, based upon the electron transport model proposed by Schmidt (2003). Discussion Although hydroxylamine was measured in systems with low dissolved oxygen and high p H following a free ammonia perturbation, no appreciable concentrations of hydroxylamine was measured in perturbed systems with high dissolved oxygen concentration or in unperturbed continuous and semi-continuous flow systems exhibiting significant nitrite accumulation (Simm et a l , 2004b, 2004c, 2004d). The results of the N 0 2 / N H 3 step study (Fig. 8.4) suggest that 259 hydroxylamine is toxic to both ammonia and nitrite-oxidizers, but do not explain the total cessation of nitrite oxidation observed during the November 22, 2002 batch test. The results of the chemodenitrification trials (Table 8.6) suggest the reactive nature of hydroxylamine is not a significant barrier to measurement in the matrices studied here and, taken together with the mass emission data for nitrous oxide, suggest localized areas of high hydroxylamine concentration are unlikely to be the source of nitrite oxidizer inhibition. In addition, the results of the NO2/NH3 spike study support the assertion that hydroxylamine is unlikely to accumulate under high nitrous acid conditions. N o significant concentration of hydroxylamine was recovered under low dissolved oxygen and high p H conditions in the presence of nitrite in this test; this suggests that hydroxylamine may have been used in the denitrification of nitrous acid to nitrous oxide, instead of being expelled from the cell. The results of the S B R perturbation studies (Table 8.7) provide insight into the mechanisms responsible for hydroxylamine production and therefore, the physiological state of the ammonia-oxidizer population. Hydroxylamine was recovered under high p H (low nitrous acid), high free ammonia, and low dissolved oxygen conditions. Literature evidence suggests ammonia-oxidizing organisms can use both oxygen and nitrite as terminal electron acceptors (Anderson and Levine, 1986). Three other systems are also fed with electrons from the oxidation of hydroxylamine to nitrite, under oxygen limitation (Bock and Wagner, 2003). The fact that hydroxylamine was expelled from the cell (November 22, 2002 batch test and S B R trials 4,5,and 6) suggests there is a deficiency in both oxygen and nitrous acid for all or part of the ammonia oxidizer population. In theory, this set of circumstances could be initiated or. exacerbated by a high free ammonia concentration, since free ammonia is the true substrate of the ammonia monooxygenase enzyme and therefore ammonia-oxidizing organisms (Suzuki et al., 1974). 260 Therefore, an increasing free ammonia concentration in an environment deficient in available electron acceptor (whether it be oxygen and/or nitrite) could result in hydroxylamine being expelled from the cell (assuming an insufficient enzyme pool). The absence of hydroxylamine in stable nitrification systems is likely due to one or a combination of physiological adaptations. Ammonia-oxidizing organisms are known to have more than one copy of the genes for hydroxylamine oxidoreductase and ammonia monooxygenase (Hooper et al., 1990, McTavish et al., 1993). It has been speculated that the presence of multiple genes might allow more rapid regeneration of the respective messenger R N A during ammonia flushes (Hommes et al, 1998) or may be responsible for maintaining a certain ratio of gene products (Bergmann et al, 1994). Bock et al. (1995) reported that Nitrosomonas europaea and Nitrosomonas eutropha grown under oxygen limitation were physiologically different from those grown with sufficient aeration. In contrast to the densely stacked intracytoplasmic membranes of well-aerated cells, the distance between the membrane layers of cells grown under oxygen limitation had increased, with cells grown under oxygen limitation containing layers of high and low electron density between the intercytoplasmic membranes. Bock et al. (1995) proposed the areas of high electron density contain high amounts of unidentified enzyme. The data presented here raise two important questions. Firstly, what is the primary cause(s) o f nitrite accumulation in nitrification systems, i f hydroxylamine and free ammonia are not responsible for the observed phenomenon? Several mixed and pure culture experiments indicated free ammonia itself is not inhibitory to Nitrospira (Chapter 5 and Simm et al., 2004a). Finally, 261 what potential role(s) could hydroxylamine play where long-term nitrite accumulation is observed? Nitrous oxide emissions were coincidental with nitrite accumulation in every case studied as part of this research program (Simm et al., 2004b, 2004c, 2004d). The results of the prewashed biomass test with NO2/NH3 spikes implicate autotrophic denitrification as the primary source of measured nitrous oxide. The fact that hydroxylamine was measured during the November 22, 2002 batch test suggest both nitrite and oxygen were limiting in parts of the mixed liquor culture. The significant nitrous oxide emissions during the test suggest a portion of the ammonia-oxidizer population was likely denitrifying nitrous acid to nitrous oxide. The electron transport model proposed by Schmidt (2003) (presented in Fig . 8.1) suggests that, under oxygen limitation, ammonia could be metabolized to hydroxylamine, nitrous acid, and ultimately nitrous oxide within the ammonia-oxidizer cell, therefore denying adjacent nitrite-oxidizer colonies of their substrate. This possibility is supported by the results of chemodenitrification trial 1 (Table 8.6), whereby live biomass from R2 appeared to convert hydroxylamine to nitrous oxide, with no change in exogenousjy added nitrite. It is the ability of ammonia-oxidizers to metabolize ammonia directly to nitrous oxide, which is believed to be the most likely cause of observed nitrite accumulation in nitrification systems. The first step of the ammonia oxidation process (oxidation of free ammonia to hydroxylamine catalyzed by the ammonia monooxygenase enzyme) requires oxygen. The rate of ammonia oxidation, and therefore oxygen utilization, w i l l increase with increasing free ammonia concentration, since free ammonia is the true substrate for ammonia-oxidizers (Suzuki et al., 1974). The second step in the ammonia oxidation process (the oxidation of hydroxylamine to nitrous acid) does not require molecular oxygen. The fact that the localized oxygen concentration is lowered by the increase in free ammonia oxidation rate results 262 in all or part of the ammonia-oxidizer cells metabolizing nitrous acid to nitrous oxide. The localized drop in both oxygen and nitrite concentration result in a significant decrease in nitrite-oxidizer metabolism and, therefore, nitrite accumulation. The model proposed here is referred to as the "substrate competition model" for apparent free ammonia inhibition of nitrite oxidizers. The model provides a plausible explanation for the data presented in other studies conducted as part of this research program (Simm et al., 2004b, 2004c, 2004d). When the data of other researchers is viewed under the context of the "substrate competition model", and hydroxylamine is viewed as an indicator of terminal electron acceptor deficiency (and not an inhibitor), a very different picture emerges to that commonly reported in the literature. For instance, Yang and Alleman (1992) reported nitrite peaks at p H 7.5 and low dissolved oxygen (0.5 mg/L), at p H values of 8.0 and 8.5 at both high (6 mg/L) and low (0.5 mg/L) dissolved oxygen, but no nitrite accumulation at either high (6 mg/L) or low (0.5 mg/L) dissolved oxygen at p H 7.0. They reported measuring hydroxylamine in each test exhibiting significant nitrite accumulation, suggesting that some or all o f the ammonia-oxidizer population was subjected to an electron acceptor limitation. A t p H 7.5, the concentration of free ammonia is three times higher than at p H 7.0, whereas the concentration of nitrous acid is essentially three times lower, relative to the concentration at p H 7.0. The fact that animonia oxidizers are subjected to an electron acceptor limitation makes it a certainty that nitrous acid is being denitrified to nitrous oxide, via an internal process that limits the localized substrate concentration for nitrite oxidizers. A t p H values of 8.0 and 8.5, the free ammonia concentration is 9 and 25 times higher, respectively, than at p H 7.0, whereas the nitrous acid concentration is 10 and 32 times lower, respectively. Therefore, one would expect 263 nitrite accumulation to take place at higher dissolved oxygen concentrations, since the nitrous acid concentration is already limited on stoichiometric grounds. A competition model also explains the observations of Cecen and Ipek (1998) who suggested the dissolved oxygen to free ammonia ratio was most important for nitrite accumulation. A s the dissolved oxygen to free ammonia concentration ratio declines for a given p H value, the possibility of autotrophic denitrification, and therefore, nitrite accumulation increases. Substrate competition also explains the data of Hung Soek Yoo et al. (1999), who have suggested that nitrite accumulation in an S B R was the result of periodic exposure to free ammonia and possibly hydroxylamine. Hung Soek Yoo et al. (1999) did not assay hydroxylamine and the experimental results collected for this research do not support the assertion that a high enough hydroxylamine concentration could be produced and maintained to sustain nitrite accumulation and still maintain complete ammonia oxidation. However, these investigators reported simultaneous nitrification denitrification (SND) taking place in the aeration phase of their S B R system and reported nitrite accumulation to be dependant upon both the dissolved oxygen concentration and aeration rate. This correlation is certainly consistent with the " substrate competition model", since both dissolved oxygen concentration and aeration rate would impact the availability of oxygen and nitrous acid, rate of ammonia oxidation, and amount and timing of autotrophic nitrous oxide generation. The experiments conducted as part of this research program (Chapter 3 (Simm et al. 2004b), Chapter 5 (Simm et al. 2004c), and Chapter 7 (Simm et al. 2004d)) suggest autotrophic nitrous oxide generation as the most likely source of the S N D reported by Hung Soek Yoo et al. (1999). 264 This "substrate competition model" also supports the observation that plug flow systems, such as the one used by Turk and Mavinic (1986, 1987, 1989a, 1989b), are more likely to exhibit sustained nitrite accumulation, since the dissolved oxygen and nitrous acid concentrations w i l l be relatively low, and free ammonia relatively high, at the exit of the anoxic zone. The effect would be magnified at lower anoxic zone O R P values, since more oxygen is used in the first aerobic zone to meet the demand of highly reduced substances. These investigators reported that nitrite accumulation is dependant upon aerobic retention time and the number of aerobic stages, which is consistent with the competition model. The data collected for this research program suggests that accumulation of hydroxylamine to inhibitory levels is unlikely to play a significant role in nitrification processes under stable operating conditions (Simm et al., 2004b, 2004c, 2004d). However, the S B R perturbation studies indicate hydroxylamine can be expelled from ammonia-oxidizer cells when the concentration of terminal electron acceptors is limited, relative to the enzyme pool. This would certainly be possible under transient loading conditions that cause both high oxygen demand and free ammonia concentration. The expulsion of hydroxylamine from the cell would starve some ammonia-oxidizer cells and k i l l both nitrite and ammonia-oxidizer cells, i f the concentration were to build up to sufficient levels. The end-result would be a severely compromised nitrifier population with nitrite-oxidizers likely being most affected. The result would be a transient accumulation of nitrite, the length and severity of which would depend upon the relative difference in number o f ammonia and nitrite-oxidizer cells following the original upset. The response of a given nitrification system to such a transient accumulation would be dependant upon the size of the nitrifier enzyme pool. This explains why acclimation to high ammonia loads must be carried out in a step-wise fashion. A n enzyme or population assay of the starting seed 265 should allow one to make an educated assessment of the allowable step size, for a given waste and seed. The "substrate competition model" has significant implications, not only for apparent free ammonia inhibition, but for the modeling of nitrification processes, in general. A s indicated in previous manuscripts (Simm et al., 2004a, 2004d), ammonia and nitrite oxidizer colonies grow in close proximity to each other. The localized availability o f substrate for nitrite oxidizing organisms w i l l , therefore, depend upon the way in which ammonia-oxidizers respond to the relative amounts of both oxygen and nitrite as determined by aeration rate, mixing conditions, and p H . This, in turn, impacts measured kinetic parameters for both organism groups when studied in mixed cultures. Conclusions The following conclusions have been drawn based upon this research. • The combination of low dissolved oxygen and high p H , resulting in low nitrous acid concentration, appear to be required tojnduce hydroxylamine accumulation under batch and semi-batch conditions. • Autotrophic denitrification was the most likely source of nitrous oxide in both C S T R and S B R systems. 266 It is unlikely that hydroxylamine is the primary source of nitrite-oxidizer inhibition in biological nitrification systems. Substrate competition between ammonia and nitrite oxidizers is the most likely cause of apparent free ammonia inhibition of nitrite oxidizing organisms. The "substrate competition model" is consistent with research findings of other researchers and has potentially profound implications upon the way in which we view nitrification and ultimately model these processes. 267 Periplasmic Space Cytoplasmic membrane Cytoplasm Figure 8.1- Nitrosomonas electron transport chain (Schmidt, 2003) 268 |A. | Dissolved Oxygen 200 400 600 800 Elapsed Time (minutes) • _ 80 NH4-N 200 400 600 800 Elapsed Time (minutes) NO,-N 200 400 600 Elapsed Time (minutes) Headspace Nitrous Oxide 200 400 600 Elapsed Time (minutes) Nitric Oxide 200 400 600 800 Elapsed Time (minutes) ° - 7.8 7.6 E pH 200 400 600 800 Elapsed Time (minutes) NH3-N 200 400 600 800 Elapsed Time (minutes) N02-N 200 400 600 800 1000 Elapsed Time (minutes) Hydroxylamine Elapsed Time (minutes) Nitrous Acid 200 400 600 800 Elapsed Time (minutes) Figure 8.2 - November 22,2002 Batch Test 269 60 50 0 I G ) £ g 40 3 O <D 0) •2 30 o 20 8 10 0 0.0 2.1 4.2 8.5 17.0 42.4 Hydroxylamine concentration (mg-N/L) 42.4 Figure 8.3- Summary of experimental results from shaker table experiment (trial 3) to J 7 * 6 Nitrite-nitrogen |NOz addition (2.5 mg-N/mln) | ; ^ r \ iNHjOH addition (1 mg-Wmin) I \ • c / -j \ \ \ \ A j • 215 265 315 Elapsed Time (minutes) Hydroxylamine 260 330 Elapsed Time (minutes) • •»- Projected—a— Actual Headspace Nitrous Oxide 30 25 I 20 a. * w Q. • 15 Oxid itrous z o -5 50 100 150 200 250 300 350 400 4 0 Elapsed Time (minutes) G . 6.5 Dissolved Oxygen Trace | 6 c 5.5 «• O) 1 5 1 « I 1 A if 0 50 100 150 200 250 300 350 400 450 500 Elapsed Time Nitrate and Hydroxylamine Nitrogen 165 215 265 315 Elapsed Time (minutes) -Proj, Nitrate -- Actual Nitrate - Hydroxylamine (mg-N/L) j Nitrite and Unaccounted for Hydroxylamine Elapsed Time (minutes) -Nitrite-N -Unaccounted for Hydroxylamine —+- Nitrite minus baseline | Cummulative Nitrous Oxide Elapsed Time (minutes) Ammonia (mg-N/L) Elapsed Time (minutes) Figure 8.4 - Summary of Experimental Results from nitrite and hydroxylamine step study 271 1000 i 100 200 300 400 500 600 700 Elapsed Time (minutes) Figure 8.5 - Time series plot of nitrous oxide for prewashed biomass test with N0 2/NH 3 spikes to to Figure 8.6 - Summary of tracking study results for SBR base case prior to perturbation 273 G. | Headspace Nitrous Oxide 500 <J> 450 p 400 £ p" 350 Z O. 300 » S 250 « « 200 •Sg 100 S 50 I o -50 • / / J 7 1 * * • * • t *-• ^ * * * * * Elapsed Time (minutes) • Q. 300 Q. —* 250 01 C 200 i 1 5 0 >* 100 Hydroxylamine Elapsed Time (minutes) Nitrous Acid Concentration versus Time 0.0004 0 00035 "O o •— 0.0003 51 « 0 0 0 0 2 5 „ h <a 0.0002 • Sd 0.00015 J C p 0.0001 z o S 0.00005 f j j j I / t i T- * -0.00005 • I • Elapsed Time (minutes) pH Elapsed Time (minutes) Elapsed Time (minutes) • N02 14 (mg/L) / (mg/L) / litrate-N / litrate-N J litrate-N z -50 100 150 200 250 Elapsed Time (minutes) Head space NO Elapsed Time (minutes) Elapsed Time (minutes) SBR4 (NaOH addition - October 18,2002) O E 40 5 "~- 30 s*> <£> # £ *? \<> # <fc N<V ^ ^ ^ Elapsed Time (minutes) Figure 8.7 - Summary of tracking study results for influent ammonia increase and anoxic/aerobic phase pH control for SBR 274 Table 8.1- Synthetic feed composition for bench scale completely st irred tank reactor ( C S T R ) . Feed Component Concentrat ion (mg/L) Sodium acetate ( N a C H 3 C O O H ) 250 mg/L Ammonium chloride (NH4CI) 556 mg/L Sodium bicarbonate (NaHCOa) 1,936 mg/L Yeast extract . 50 mg/L Potassium phosphate (K2HPO4) 56 mg/L Magnesium chloride (MgCl2-6H20) 37.5 mg/L Calcium chloride (CaCl 2 -2H 2 0 ) 24 mg/L Ferric chloride (FeCl 3 -6H 2 0) 1.9 mg/L Manganese sulphate ( M n S C v F ^ O ) 0.09 mg/L Sodium molybdate ( N a 2 M o 0 4 - 2 H 2 0 ) 0.008 mg/L Zinc sulphate (Zn S 0 4 - 7 H 2 0 ) 0.3.75 mg/L Cobalt chloride ( C o C l 2 - 2 H 2 0 ) 0.001 mg/L 275 Table 8.2- Synthetic feed composition for bench scale sequencing batch reactor (SBR) . Feed Component Concentrat ion (mg/L) Sodium acetate ( N a C H 3 C O O H ) 520 mg/L Ammonium chloride (NH4CI) 340 mg/L Sodium bicarbonate ( N a H C 0 3 ) 500 mg/L Yeast extract 100 mg/L Potassium phosphate ( K 2 H P 0 4 ) 28 mg/L Magnesium sulphate (MgSC^) 62.1 mg/L Calcium chloride ( C a C l 2 - 2 H 2 0 ) 62.1 mg/L Ferric chloride (FeCl 3 -6H 2 0) 1.9 mg/L Manganese sulphate ( M n S C y F b O ) 0.09 mg/L Sodium molybdate ( N a 2 M o C y 2 H 2 0 ) 0.008 mg/L Zinc sulphate (Zn S 0 4 7 H 2 0 ) 0.375 mg/L Cobalt chloride ( C o C l 2 - 2 H 2 0 ) 0.001 mg/L 276 Table 8.3 - Summary of Experimental Series for In-situ and Prewashed Biomass Tests with schedule and objectives for supplementary experiments. Reactor In Situ Test(s) Objective(s) Supplementary Test(s) Supplementary Test Objective(s) CSTR NH 3 spike at low D.O. To induce NH 2OH production. NH 3 spike at low D.O. with pH control. To induce NH 2OH production. Cell breakage experiment Analysis of pH control chemical Inhibition trials with ATU. Confirm source of measured NH 2OH. Prewashed biomass tests with N0 2 /NH 3 spikes. Chemodenitrification trials Confirm the primary source(s) of nitrous oxide and the potential role of N 0 2 in NH 2OH production. Shaker table experiments and reactor step studies Confirm role of pH and measured [NH2OH] on observed N0 2 accumulation. SBR Reduced aerobic phase DO To elucidate the mechanism(s) responsible for NH 2 OH production. Increase influent NH 3 . Anoxic pH control. Anoxic/aerobic pH control. NH 3 perturbation + phase pH control. 277 Table 8.4 - Experimental Series for Shaker Table Experiments. Test Series Objectives Description Initial [N02-N] (mg/L) Initial [NH2OH-N] (mg/L) 1 Determine appropriate incubation time for hydroxylamine trials. Individual flasks incubated for 0.5, 1, 2, 3, and 5 hours. Experiment included a media blank having no biomass and a replicate for the 3-hour incubation. The media pH before and after incubation was approx. 8.0. 100 0 2 Obtain a preliminary indication of the toxicity of hydroxylamine to nitrite oxidizers. Flasks were incubated for 3 hours (based upon results of test series 1) with various hydroxylamine concentrations. 100 0,1,1.95, and 18.6. 3 Determine the toxic level of hydroxylamine to nitrite oxidizers. Flasks were incubated for 3 hours (based upon test series 1) with various hydroxylamine concentrations. The media pH was approximately 8. 100 0,2.1,4.2,8.5 ,17, 42.4. 4 Determine the affect of pH on observed phenomenon in test series 2 and 3 and November 22, 2002 batch test. Flasks were incubated at 1,2,3, and 5 hours at pH 7 and 7.8. 100 0 278 Table 8.5 - Experimental Series for Chemodenitrification Trials. Trial Description Objective(s) 1 Live biomass with N H 2 O H and N 0 2 spike under N 2 sparging. Confirm biological consumption of N H 2 O H and stoichiometry of autotrophic denitrification. 2 Same as test series 1 with boiled biomass. 3 N H 2 O H spike in CSTR effluent with high D.O. (approx. 8 mg/L) Confirm the affect of dissolved oxygen concentration on N H 2 O H decomposition rate. 4 Same as 3 with median D.O. (2 mg/L). 5 Same as 3 with low D.O. (0.3-0.5 mg/L). 6 N H 2 O H spike in feed with high D.O. (approx. 8mg/L) Confirm impact of feed components on N H 2 O H decomposition. 7 N H 2 O H and N 0 2 spikes in feed with nitrogen sparging. Confirm chemodenitrification reaction between N H 2 O H and H N 0 2 . 8 Same as test 3 with biomass Confirm affect of biomass on N H 2 O H decomposition rate. 9 Same as test 4 with biomass 10 Same as 3 but at neutral pH. Confirm the impact of pH on N H 2 O H decomposition rate. 279 Table 8.6 - Summary of Results of Chemodenitr i f icat ion and Autotrophic Denitr i f ication Tr ia ls with Hydroxylamine. Trial Description D.O. (mg/L) A [NH 2OH-N] (mg/L-hr) N 2 0 (avg.) (max.) (max. ppm) % A [NH 2 OH-N] 1 Live biomass with N H 2 O H and N 0 2 spike under N 2 sparging. 0 0.6 0.6 200 >100 2 Same as test series 1 with boiled biomass. 0 0.17 0.17 20 28.6 3 N H 2 O H spike in CSTR effluent with high D.O. 8 1.1 2.4 94 34.5 4 Same as 3 with median D.O. 2 0.13 0.13 5 Same as 3 with low D.O. 0.3-0.5 0.33 0.33 20 22.8 6 N H 2 O H spike in feed with high D.O. 8 0.98 0.98 11 13 . 7 N H 2 O H and N 0 2 spikes in feed with nitrogen sparging. 0 approx. 0 approx. 0 7 <0.5 8 Same as test 3 with biomass 8 1.6 2.6 51 14.3 9 Same as test 5 with biomass 0.3-0.7 1.1 2 41 7.6 10 Same as 3 but at neutral pH. 8 0.4 0.4 26 31.5 280 Table 8.7 - Summary of results of S B R perturbation trials relative to base case. Tria 1 Description [NH 3] max. N H 4 oxid. rate (mg/L-hr) N 0 2 oxid. rate (mg/L-hr) [NH 2OH]ma X. Heaspace N20 trend [ppm] max 1 Base case (no perturbation) 2.2 11 10.3 0 14 2 Reduce aerobic phase D.O. 2.0 1 I 0 t 25 3 Double influent ammonia load 2.7 Same as base case Same as base case trace t 84 4 Anoxic phase pH adj. 5.5 Approx. same Approx. same 110 t 30 5 Anoxic & partial aerobic phase pH control 4.4 1 80 0 6 Increase influent ammonia load & anoxic and aerobic pH control. 7.5 1 1 300 450 (N 2 0 did not appear in significant quantities until pH control abandoned). 281 Bibl iography Anderson, L C , and Levine, J.S. 1986. Relative rates of nitric oxide and nitrous oxide production by nitrifiers, denitrifiers and nitrate respierers. Appl. Environ. Biotechnol., 51(5): 938-945. Anthonisen, A . C , Loehr, R .C, Prakasam T.B.S., and Srinath, E.G. 1976. Inhibition of nitrification by ammonia and nitrous acid. J. Wat. Pollut. Control Fed. 4 8 : 835-851. Arnold, P.W. 1954. Losses of nitrous oxide from soils. J. Soil Sci. 5: 116-126. Bergmann, D.J., Arciero, D., and Hooper, A.B., 1994, Organization of the HAO gene cluster of Nitrosomonas europaea: Genes for two tetraheme cytochromes. J. Bacteriol. 176: 3148-3153. Bock, E., Schmidt, I., Stuven, R., and Zart, D. 1995. Nitrogen loss caused by denitrifying Nitrosomonas cells using ammonium or hydrogen as electron donors and nitrite as electron acceptor, Arch. Microbiol. 163(1): 16-20. Bock, E., and Wagner, M . 2003. Oxidation of inorganic nitrogen compounds as an energy source, In: The Prokaryotes an evolving electronic resource for the microbial community (Release 3.14). Springer-Verlag Heidelberg. Cecen, F., and Ipek, S. 1998. Determination of the inhibition of ammonia-N and urea-N oxidations by the fed-batch reactor (FBR) technique. Wat. Sci. Technol. 38(1): 141-148. 282 Chalk, P . M . , and Smith, C .J . 1983. Chemodenitrification. In: Developments in plant and soil sciences, Volume 9, Gaseous loss o f nitrogen from plant-soil systems. J.R. Freney and J.R. Simpson (eds). Martinus Nijhoff / Dr. W . Junk Publishers (Kluwer Academic Publishers Group), The Hague, pp. 65-89. Eaton, A . D . , Clesceri, L .S . , Greenberg, A . E . (eds) 1995. Standard methods for the examination of water and wastewater (19 t h Edition)., American Public Health Association, Inc., New York. Frijlink, M . J . , Abee, T., Lanbroek, H.J . , de Boer, W. , and Konings, W . N . 1992. The bioenergetics of ammonia and hydroxylamine oxidation in Nitrosomonas europaea at acid and alkaline p H . Arch . Microbio l . 157: 194-199. Hooper, A . B . 1989. Biochemistry of the nitrifying lithoautotrophic bacteria., Autotrophic Bacteria, Ed . by H . G . Schlegel and B . Bowien, Science Tech. Publishers, Madison, Wisconsin, pp. 239-265. Hooper, A . B . , Arciero, D . M . , DiSpirito, A . A . , Fuchs, J., Johnson, M . , LaQuier, F., Mundform, G. , and McTavish , H . 1990. Production of nitrite and N2O by ammonia-oxidizing nitrifiers. In: Nitrogen fixation: achievements and objectives. Gresshoff, Roth, Stacey & Newton (eds). Chapman and Hal l , N e w York. Hommes, N . G . , Sayavedra-Soto, L . A . , and Arp , D.J . 1998. Mutagenesis and expression of amo, which codes for ammonia monooxygenase in Nitrosomonas europaea. J. Bacteriol. 180: 3353-3359. 283 Hyungseok Yoo, Kyu-Hong Ahn, Hyung-Jib Lee, Kwang-Hwan Lee, Youn-Jung Kwak, and Kyung-Guen Song. 1999. Nitrogen removal from synthetic wastewater by simultaneous nitrification and denitrification (SND) via nitrite in an intermittently aerated reactor. Water Res. 33(1): 145-154. McTavish, H., Fuchs, J.A., and Hooper, A.B. 1993. Sequence of the gene coding for ammonia monooxygenase in Nitrosomonas europaea. J. Bacteriol. 175(8): 2436-2444. Poth, M. , and Focht, D.D. 1985. 1 5 N kinetic analysis of N2O production by Nitrosomonas europaea: an examination of nitrifier denitrification. Appl. Environ. Microbiol.-49: 1134-1141. Schmit, I. 2003. In: The Prokaryotes an evolving electronic resource for the microbial community (Release 3.14). Springer-Verlag Heidelberg. Schmidt, I., Sliekers, O., Schmid, M. , Bock, E., Fuerst, J., Kuenen, J.G., Jetten, M.S.M., and Strous, M. 2003. New concepts of microbial treatment processes for the nitrogen removal in wastewater. FEMS Microbiol. Rev. 27(4): 481-492. Simm, R.A., Ramey, W.D., Mavinic, D.S. 2004a. A targeted study on possible free ammonia inhibition of Nitrospira. (submitted to J. Env. Eng. Sci. May 2004). 284 Simm, R . A . , Ramey, W . D . , and Mavinic , D.S. 2004b. Preliminary evaluation of the use of fatty acid ratios for tracking the potential for nitrite accumulation in nitrifying reactors with low carbon to nitrogen ratio. J. Environ. Eng. Sci . 3: 31-40. Simm, R . A . , Ramey, W . D . , and Mavinic , D.S. 2004c. Nitrifier population dynamics in a bench scale activated sludge reactor following an induced perturbation (accepted for publication with minor revisions in Journal of Environmental Engineering and Science June 2004). Simm, R . A . , Ramey, W . D . , and Mavinic , D.S. 2004d. Mechanisms responsible for apparent free ammonia inhibition in a sequencing batch reactor, (submitted to the A S C E Journal of Environmental Engineering July 2004). Simm, R . A . , Parkinson, P., Mavinic , D.S. , Ramey, W . D . 2004e. Hydrooxylamine analysis of wastewater samples v ia gas chromatography, (submitted to Environmental Technology for publication July 2004). Stuven, R., Vollmer, M . , and Bock, E . 1992. The impact of organic matter on nitric oxide formation by Nitrosomonas europaea. Arch. Microbiol . 158: 439-443. Suzuki, I., Dular, U . , and Kwok , S.C. 1974. Ammonia or ammonium ion as substrate for oxidation by Nitrosomonas europaea cells and extract. J. Bacteriol. 120(1): 556-558. Turk, O., and Mavinic , D.S. 1986. Preliminary assessment of a shortcut in nitrogen removal from wastewater. Can. J. C i v i l Eng. 13: 600-605. 285 Turk, O., and Mavinic, D.S. 1987. Selective inhibition: a novel concept for removing nitrogen from highly nitrogenous wastes. Env. Technol. Lett., 8: 419. Turk, O., and Mavinic, D.S. 1989a. Stability of nitrite build-up in an activated sludge system. J. Water Pollut. Control Fed. 61(8): 1440-1448. Turk, O., and Mavinic, D.S. 1989b. Influence of process changes on maintaining nitrite build-up in an activated sludge system acclimated to free ammonia. Water Res. 23(11): 1383-1388. White, D. 1995. The physiology and biochemistry of prokaryotes, Oxford University Press, New York, Oxford. Yang, L., and Alleman, J.E. 1992. Investigation of batchwise nitrite build-up by an enriched nitrification culture. Water Science and Technology, 26(5-6): 997-1005. 286 9.0 Conclusions and Recommendations for Future Research 9.1 Discussion Relating Manuscripts To Each Other. The original hypothesis for this research was that free hydroxylamine, and not free ammonia, is the true cause of nitrite accumulation. This hypothesis was developed based upon the observation that nitrite accumulation is a function of free ammonia and dissolved oxygen concentration (Cecen and Ipek, 1998, Hyugseok Yoo et al., 1999) and the work of Yang and Alleman (1992), who reported that the combination of high free ammonia and low dissolved oxygen results in hydroxylamine accumulation in batch systems. Hydroxylamine is a reactive compound that, once formed, autodecomposes to one of several decomposition products. One of the primary recommendations from Yang's (1990) PhD work is that an improved method for hydroxylamine analysis was required. Therefore, the starting point of this research program was the development of an appropriate analytical method for hydroxylamine analysis. The development work for the required analytical method is the subject of the manuscript presented in Chapter 2. To this authors' knowledge, no one has yet demonstrated hydroxylamine accumulation in a wastewater system exhibiting relatively stable operation. Yang and Alleman (1992) had demonstrated hydroxylamine accumulation in batch systems subjected to extreme ammonia and dissolved oxygen stresses, conditions which would have likely overwhelmed the available enzyme pool. Therefore, the next logical step in the research program was to create conditions believed to be conducive to hydroxylamine production. The coincidental observation of nitrite and hydroxylamine accumulation would be taken as first direct evidence that hydroxylamine 287 might well be the true inhibitor of nitrite oxidation. Several reactors were operated at various dissolved oxygen levels and solids retention times (SRTs). The reactors operated under low dissolved oxygen conditions resulted in partial nitrification and the establishment of a stable free ammonia stress. The results of this work are summarized in the manuscript provided as Chapter 3. N o hydroxylamine was recovered from any of the systems studied. N o hydroxylamine was recovered from the test reactors discussed in Chapter 3, either because it was not produced, or possibly because it had decomposed prior to measurement. Nitrous oxide emissions were coincidental with nitrite accumulation and nitrous oxide is a known hydroxylamine decomposition product. Increasing the dissolved oxygen concentration in one the reactors exhibiting partial nitrification and nitrite accumulation and tracking the reactor and population dynamics associated with its recovery, would provide insights into the cause of the original nitrite inhibition over and above oxygen limitation. The results of such an experiment are summarized in the manuscript provided as Chapter 4. N o hydroxylamine was measured before or during the dissolved oxygen perturbation. These test results indicated a delay in appreciable nitrite oxidation (as measured by a nitrate end-product) until after the dissolved oxygen concentration had increased and free ammonia had disappeared. This suggested that low dissolved oxygen and/or free ammonia inhibition may have played a primary role in the establishment o f a stable nitrite end product. Interestingly, the establishment of significant nitrite oxidation was coincidental with a decline in nitrous oxide emissions. Several batch tests conducted using mixed liquor samples collected from reactors operated for this research program indicated free ammonia might not be inhibitory to nitrite-oxidizers at commonly reported inhibitory concentrations (Chapter 5). The majority of research conducted on free ammonia inhibition of nitrite-oxidizing organisms has been with mixed cultures from full 288 scale and laboratory reactors. In most cases, an ammonia spike is added to the culture under controlled dissolved oxygen and p H conditions. This approach does not account for potential interactions between ammonia and nitrite-oxidizer colonies or the impact of colony structure and spatial orientation upon the observed phenomena. This is particularly relevant when one considers the fact that ammonia-oxidizing organisms can use nitrite as a terminal electron acceptor to produce nitrous oxide, which coincidentally had been recovered from reactors exhibiting nitrite accumulation. Hunik et al., (1993) had previously demonstrated that free ammonia was unlikely to be inhibitory to pure cultures of Nitrobacter agilis. Recent research has indicated the primary nitrite-oxidizers in wastewater systems are Nitrospira and not Nitrobacter. Therefore, it was decided to undertake an inhibition trial using pure cultures of Nitrospira moscoviensis in order to confirm free ammonia toxicity toward these organisms in the absence of ammonia and nitrite-oxidizer interactions. The results of this study, reported in Chapter 5, support the conclusion that free ammonia is unlikely to be inhibitory to nitrite-oxidizing organisms in wastewater systems. The results of the dissolved oxygen perturbation, presented in Chapter 4, coupled with the results presented in Chapter 5, suggest the limiting bulk liquid dissolved oxygen concentration may have been the primary cause of observed nitrite accumulation. A n ammonia step study was conducted at high dissolved oxygen concentration, to elucidate the mechanism(s) responsible for nitrite accumulation, i f any, in the presence of a free ammonia concentration considered as inhibitory, based upon literature evidence. The results of this study are presented in Chapter 6. The instantaneous, and sustained, increase in ammonia loading resulted in a gradual decrease in nitrite oxidizer activity that was coincidental with an increase in nitrous oxide emissions from the reactor, even at bulk liquid dissolved oxygen concentrations well above 3 mg/L. Microbial community analysis, using slot blotting techniques, indicated that Nitrospira, and not 289 Nitrobacter, populations were most affected during the perturbation. This study provided evidence that localized competition for nitrite and possibly oxygen at the bacterial colony level may be the true cause of nitrite accumulation. Once again, no measurable concentrations of hydroxylamine were recovered from the test reactor during this particular study. Turk and Mavinic (1986, 1987, 1989a, 1989b) completed one o f the most quoted pieces o f research in the area of free ammonia inhibition. These authors controlled p H during the anoxic period, and therefore, free ammonia concentration, in a plug flow nitrification-denitrification system to induce nitrite-oxidizer inhibition and, therefore, aerobic phase nitrite accumulation. This particular strategy produced high concentrations of aerobic phase nitrite for several SRTs, but always resulted in nitrite-oxidizer acclimation. Experiments conducted as part of the current research (Chapter 8), indicated a combination of free ammonia, p H , and dissolved oxygen manipulation could result in hydroxylamine production. Therefore, it was decided to emulate the work of Turk and Mavinic (1986, 1987, 1989a, 1989b) using a sequencing batch reactor (SBR) system, to study the initial conditions leading to nitrite-oxidizer inhibition and subsequent nitrite-oxidizer acclimation. It was believed that the study of the acclimation response would assist in elucidating the true cause o f the nitrite oxidizer inhibition observed by Turk and Mavinic (1986, 1987, 1989a, 1989b). The results of this particular study are presented in Chapter 7. The experiments described in Chapter 7 indicated the initial perturbation, which impacted nitrite oxidizers most severely, and not free ammonia, was the dominant cause of observed nitrite accumulation. Nitrous oxide emissions were once again coincidental with the accumulation of nitrite. A spike experiment, described by Shiskdwski et al., (2004) indicated ammonia was required for nitrous oxide generation, suggesting that autotrophic denitrification was likely the primary source of measured nitrous oxide. The consumption of nitrite by ammonia oxidizers likely prolonged the time required for the nitrite oxidizer population to acclimate to the change 290 and re-establish nitrate accumulation in the aerobic phase. N o measurable concentrations of hydroxylamine were recovered from the S B R during the period of nitrite accumulation, confirming that hydroxylamine is unlikely to be the cause o f sustained nitrite accumulation. To this point, several lines of evidence suggested free ammonia was unlikely to be the primary cause of nitrite oxidizer inhibition as reported in the literature (Chapter 5). However, continuous flow studies (Chapters 3, 4, 6, and 7) indicated that hydroxylamine was unlikely to be the cause of sustained nitrite accumulation, although nitrous oxide production was coincidental in every case of nitrite accumulation. Since nitrous oxide is a known autodecomposition product of hydroxylamine, it was necessary to eliminate hydroxylamine decomposition as the primary source of nitrous oxide and the possibility that hydroxylamine had decomposed to non-detectable concentrations prior to measurement. In addition, by studying the mechanisms responsible for hydroxylamine production in nitrifying systems it might be possible to develop an alternate explanation for the inhibition of nitrification observed of Yang and Alleman (1992). This explanation might well have more universal application. The results of experiments conducted to determine the mechanisms associated with hydroxylamine production are presented in Chapter 8. Consistent with our current understanding of the biochemistry of ammonia-oxidizing organisms, a combination of low dissolved oxygen and low nitrous acid concentrations was required to produce measurable concentrations of hydroxylamine. This work indicated that decomposition of hydroxylamine is unlikely to be the primary cause of observed nitrite accumulation. In fact, the presence of hydroxylamine was 'deemed to be an indicator of conditions at the bacterial colony level, as opposed to the primary cause of nitrite accumulation. Several lines of evidence presented in Chapters 3 through 8, suggest that substrate competition between ammonia and nitrite oxidizing organisms is the most likely cause of observed nitrite 291 accumulation. These data, and other observations made during the course of this research program, have led to the development of a conceptual model that this author believes helps to explain not only free ammonia inhibition, but also the variation in inhibitory limits presented in the literature. A summary o f this model, and its implications, is presented in the text below. This model is considered as the reference point for future research on the topic of nitrite accumulation in nitrification systems. 9.2 Development of a Unified Ammonia Oxidation Model to Explain Apparent Free Ammonia Inhibition. The experiments conducted as part of this research suggest that free ammonia is not inhibitory to nitrite-oxidizing organisms at concentrations widely reported as inhibitory in the engineering literature (Chapter 5). Nitrous oxide emissions, which are believed to be primarily autotrophic in origin, were coincidental with nitrite accumulation. There are five fundamental tenets of the biochemistry of nitrifiers that receive little or no attention in the engineering literature. These include: (1) the complexity of both biological and chemical nitrogen transformations as they relate to nitrification, (2) substrate availability with changing p H , (3) the potential for substrate competition (specifically nitrite and oxygen) between ammonia and nitrite oxidizers, (4) the impact of the enzyme pool size on observed phenomena, and (5) the impact that ammonia and nitrite-oxidizer colony structure and spatial orientation has on nitrification performance. The half saturation coefficient for ammonia oxidation decreases with increasing p H , therefore free ammonia is the real substrate of ammonia-oxidizing organisms (Suzuki et al., 1974). This implies that as p H increases the rate of free ammonia oxidation to hydroxylamine increases, assuming no other substrate limitations. The free ammonia perturbation, discussed in Chapter 6, 292 resulted in an immediate increase in ammonia oxidation rate of over twenty percent. Frijlink et al., (1992) found that actively growing cells of Nitrosomonas europaea do not maintain a constant internal p H when the external p H is varied from 5 to 8; this suggests that above the optimum p H value for growth, the advantages of increased availability of free ammonia must be counterbalanced by the need to maintain an internal p H value below that of the external medium. Any increase in substrate utilization rate, as a result of free ammonia availability, should result in an increase in the oxygen utilization rate. The form of nitrite that serves as the true substrate for nitrite-oxidizers is not as well established. Hellinga et al., (1999) have indicated that nitrous acid is the true substrate for nitrite oxidizers and argue that high p H results in a decrease in nitrite oxidizer substrate and, therefore, nitrite oxidation rate, thus making the S H A R O N (Single Reactor High Activi ty Over Nitrite) process feasible. Recent research on the biochemistry of nitrite oxidizers supports this claim; however, more work needs to be done. Spieck et al., (1998) reported the nitrite oxidoreductase in Nitrospira moscoviensis was located on the periplasmic side of the cell membrane. Hopper and Dispirito (1985) suggest that the extracytoplasmic oxidation of substrates, specifically nitrous acid, in nitrite-oxidizers would allow the generation of a proton gradient, without an energy dependent permease system for nitrite. Although the location of the nitrite oxidoreductase enzyme supports the possibility that nitrous acid is the true substrate, it is unlikely that p H would have a significant impact upon nitrite oxidation rate as a result of reduced nitrous acid availability. The nitrite step studies, discussed in Chapter 8, in which nitrite was fed into a C S T R at p H 7.5 and 8.5, resulted in no appreciable change in nitrate production rate. This suggests the half saturation coefficient for nitrous acid is so low that it is unlikely to make a significant difference in nitrite oxidation rate. The fact that nitrite rarely accumulates in nitrification systems suggests nitrite oxidizers have a much higher affinity for nitrous acid than 293 do ammonia oxidizers; therefore, a high p H would have a greater impact upon the ability of ammonia-oxidizers to use this substrate. This is supported by the experimental results presented in Chapter 8, whereby sustained p H control was required to induce hydroxylamine accumulation in an S B R system. The reader is reminded that hydroxylamine oxidation to nitrous acid is believed to take place in the periplasmic space and therefore, the nitrous acid concentration wi l l be affected by external p H . A n increase in p H w i l l therefore reduce the nitrous acid concentration in the periplasmic space of ammonia-oxidizers, resulting in a significant reduction in nitrous oxide generation. This was, in fact, observed for the S B R system perturbed in the experiments discussed in Chapter 8. The fact that hydroxylamine production was coincidental with nitrous oxide generation in the C S T R experiments (November 22, 2002 batch test discussed in Chapter 8) suggests conditions were not uniform for all members of the ammonia-oxidizer population. This was most likely due to both mass transfer limitations at the colony level, as well as spatial variation in substrate availability. Both of these factors are discussed in further detail in the text below. Nitrous oxide emissions were coincidental with nitrite accumulation in every case studied as part of this research program. The experiments discussed in Chapters 6, 7, and 8 suggest autotrophic denitrification as the most likely source o f nitrous oxide generation. The fact that ammonia-oxidizing organisms can denitrify nitrite to nitrous oxide has been known for some time (Hopper, 1968). Ammonia-oxidizers can use both nitrite and oxygen as terminal electron acceptors. A s indicated by Anderson and Levine (1986), nitrous oxide generation by pure cultures of Nitrosomonas europaea increased with decreasing oxygen concentration. The fact that nitrous oxide emissions were coincidental with nitrite accumulation, in all cases studied for this research, suggests that alb or part of the ammonia-oxidizer population was subjected to an oxygen limitation. The electron transport model for ammonia-oxidizers proposed by Schmidt 294 (2003) implies nitrous acid, produced by the oxidation of hydroxylamine, can be directly reduced to nitrous oxide, without leaving the ammonia-oxidizer cell. The fact that hydroxylamine was reduced to nitrous oxide, while the concentration of exogeneously added nitrite remained the same in the live biomass test conducted under nitrogen sparging (Chapter 8) supports this portion of Schmidts' model. Therefore, nitrogen reduced to nitrous oxide by ammonia-oxidizers is simply unavailable to nitrite-oxidizing organisms. The fact that free ammonia is the true substrate of ammonia-oxidizing organisms and these organisms can use nitrite as substrate for nitrous oxide generation has profound implications on the way in which free ammonia inhibition tests are interpreted. A transient increase in free ammonia, as would be the case in a typical batch test, would result in an immediate increase in oxygen utilization rate of ammonia-oxidizers and increased potential for localized oxygen limitations, depending upon mixing conditions and aeration rate. The fact that most ammonia-oxidizers possess multiple copies of the ammonia monooxygenase and hydroxylamine oxidoreductase enzymes would likely magnify the response of the ammonia-oxidizer population. If there were a sufficient decline in local dissolved oxygen concentration, some members of the ammonia-oxidizer population would denitrify nitrite to nitrous oxide. This could potentially result in a localized limitation in the concentration of the two complementary nitrite-oxidizer substrates, nitrite and oxygen. A s indicated by Grady et al., (1999), relatively little is known about how microorganisms respond to simultaneous limitation of two or more complementary nutrients. However, Bae and Rittmann (1996) have shown that the interactive model is more appropriate when the two limiting constituents are the electron donor and the electron acceptor, as would be the case here. A n interactive model is based upon the assumption that two complementary nutrients can both influence the specific growth rate at the same time (for example, i f two complementary nutrients are present at concentrations equal to their half 295 saturation coefficients, then each together w i l l reduce the specific growth rate to half the maximum, resulting in an overall reduction to one-fourth the maximum). If an interactive type model were to apply to nitrite-oxidizers, and Nitrospira specifically, the combined effect of even a small decline in both localized oxygen and nitrite concentrations could have a significant impact upon nitrite oxidation kinetics. The fact that free ammonia was not inhibitory to Nitrospira moscoviensis, in the absence of ammonia oxidizing organisms, (as demonstrated in Chapter 5) supports the contention that it is the complex interaction between the two groups of organisms that is most important. It is unlikely that the biochemistry alone is responsible for observed phenomena, since nitrite occasionally accumulates under high dissolved oxygen conditions. The reader is referred to the experiments presented in Chapter 6. The F ISH images presented throughout this thesis show ammonia and nitrite-oxidizing organisms growing in colonies in relatively close proximity. The series of events discussed above is made possible, and in fact is likely exacerbated, by this tendency. The F I S H image presented in Chapter 6, as were all F I S H images collected for this research, was taken using a scanning confocal microscope that allows one to get a three dimensional view of the ammonia-oxidizer colonies. The data collected indicated that ammonia-oxidizer colonies constituted a solid mass of cells. Therefore, it is not difficult to imagine that ammonia-oxidizer cells in the middle of the colony could receive less oxygen, relative to those on the outer perimeter. A transient increase in free ammonia concentration, as would be the case during a batch test, would result in an increase in ammonia oxidation rate and, therefore, oxygen utilization. Three zones would likely be established inside the ammonia oxidizer colony (refer to Fig. 9.1) with the size o f each dependant upon local environmental conditions. In the first, or outer zone of the colony, ammonia would be oxidized to nitrite. The increase in ammonia oxidation rate would result in a localized reduction in dissolved oxygen concentration for nitrite 296 oxidizers in close proximity to the ammonia oxidizer colony. In the second zone of the ammonia-oxidizer colony, located deeper in the colony mass, oxygen would be limited and free ammonia and hydroxylamine from the most inner zone would be converted to a combination of nitrite and nitrous oxide. There would be a trace amount (just enough to allow a minimum of free ammonia oxidation) of oxygen in the most inner zone and free ammonia would be converted to hydroxylamine here. The net effect would be a reduction in both nitrite and oxygen concentrations for adjacent nitrite oxidizer colonies and a significant localized reduction in nitrite oxidation rate. In extreme cases (refer to Chapter 6), the reduction in localized nitrite concentration would be high enough to completely deny the nitrite-oxidizing population of their substrate. Although the ammonia oxidation rate of some cells w i l l increase as a result of a free ammonia transient, the enzyme pool has a finite size that increases at a relatively slow rate due to the low yield of nitrifying organisms. The nitrite that escapes the immediate vicinity of the nitrite oxidizer colonies, indicated in Figure 9.1, w i l l be available to other nitrite oxidizers located in a more favorable location with respect to oxygen availability, assuming such colonies exist. The experiment with transient change in dissolved oxygen (Chapter 4) is a case in point. Nitrite oxidation didn't recover until several days after both dissolved oxygen and free ammonia stresses were removed. F I S H images taken before and after this perturbation suggest a difference in the number of nitrite-oxidizer colonies, but not in their size; this suggests the delayed recovery was a function of the time required for new nitrite-oxidizer colonies to be established in these more favorable microenvironments. The acclimation of nitrite-oxidizers to free ammonia in a continuous flow system is, therefore, likely to be a function of the time required for the ammonia-oxidizer enzyme pool to adjust to the new conditions, as well as the 297 time required for nitrite-oxidizer colonies to become established in favorable microenvironments. The model presented here to explain apparent free ammonia inhibition also explains a number of anomalies in current nitrification literature. For instance, Anthonisen et al., (1976) used mixed liquor samples from wastewater systems treating agricultural wastes. These wastes are known to have extremely high nitrogen loading and, i f the treatment system is providing complete nitrification, large nitrifier populations. If free ammonia were inhibitory to nitrite-oxidizers, why is it that Anthonisen et al., (1976) report nitrite-oxidizer inhibition at free ammonia concentrations more than five times lower than those reported by Turk and Mavinic (1986, 1987, 1989a, 1989b) who seeded their reactor with mixed liquor from a pilot plant treating a relatively dilute municipal waste, having a much lower influent ammonia concentration? The answer to this question can be deduced when one looks at the F I S H images for the S B R reactor presented in Chapter 7. The increased ammonia loading eventually resulted in larger ammonia oxidizer colonies. Oxygen starvation o f portions o f an ammonia-oxidizer colony is more likely to take place in a larger colony, with a large ammonia-oxidizer population, than in a smaller colony. This, in turn sets up a localized reduction in both dissolved oxygen and nitrite concentration for adjacent nitrite-oxidizer colonies. Culture history ( C : N ratio, bulk liquid dissolved oxygen concentration, mixing conditions) w i l l determine the size and relative location of both ammonia and nitrite-oxidizer colonies and, therefore, the apparent susceptibility to a free ammonia stress. If free ammonia and dissolved oxygen concentrations coupled with colony structure govern observed nitrite accumulation phenomenon what, i f any, role does hydroxylamine play? In retrospect, the fact-that hydroxylamine was not measured in, continuous flow studies, except under the most extreme of conditions, should not have come as a surprise. The ammonia-298 oxidizing cell derives essentially no energy from the oxidation of free ammonia to hydroxylamine. In addition, hydroxylamine is a toxic compound that, as indicated by the results o f the nitrite/hydroxylamine step study (Chapter 8), kil ls ammonia-oxidizer organisms as well as nitrite-oxidizers. The production of a toxic compound that starves the producing cell, while poisoning its' neighbors within the colony, does not appear to be, a wise long-term strategy. The results of the S B R perturbation trials presented in Chapter 8 illustrate that hydroxylamine is produced only under extreme conditions. How much is produced, and whether it is produced at all , w i l l be dependant upon the size and severity of the free ammonia and dissolved oxygen stresses, relative to the size of the ammonia-oxidizer enzyme pool. Under the most extreme conditions, hydroxylamine could be produced, resulting in the starvation of ammonia-oxidizers and poisoning of both nitrifier groups. This would bias the starting seed of any long-term experiment. Such an affect was observed in the S B R trials presented in Chapter 7; however, it is unclear whether hydroxylamine was the cause. It is hypothesized here, however, that hydroxylamine production is the most likely source for free ammonia inhibition of ammonia-oxidizer organisms reported by Anthonisen et al., (1976). Most importantly, however, the presence of hydroxylamine in the bulk liquid can be viewed as an indication that part or all o f the ammonia-oxidizer population has insufficient electron acceptor, relative to available free ammonia substrate. A s indicated in Chapter 8, the hydroxylamine measured by Yang and Alleman (1992) was most likely an indicator of conditions at the colony level, as opposed to the primary cause of observed nitrite accumulation. The conceptual model presented here has several important implications, the most important o f which concerns the way in which nitrification processes are modeled in wastewater systems. It is this authors' contention that fundamental research on the various facets of the model presented 299 should lead to mechanistically-based, simulation models, with greater utility than the empirical models currently available. 9.3 Conclusions The following conclusions have been drawn based upon this research program: • GC analysis of filtered acetone derivatized samples is a suitable method for hydroxylamine analyses of wastewater samples. The method is easy to use, requires little sample preparation, provides sensitive detection, and provides a relatively stable end product for analysis. The method is particularly suited to batch test experiments whereby the researcher wishes to analyze a large number of samples over a short time period. • Preliminary evaluation of fatty acid analysis indicates this method holds promise for routine analysis of nitrifier populations in wastes having a low carbon to nitrogen ratio. • Nitrous oxide emissions were coincidental with nitrite accumulation in every case studied as part of this research. Therefore, any application of the nitrate shunt should consider the negative environmental impacts of nitrous oxide generation. • Autotrophic denitrification was the primary source of measured nitrous oxide in both completely stirred tank reactor (CSTR) and sequencing batch reactor (SBR) systems. 300 N o measurable concentrations of hydroxylamine were found in stable nitrifying and nitrification/denitrification systems exhibiting nitrite accumulation, suggesting hydroxylamine is unlikely to be the source of observed nitrite accumulation. High free ammonia and low electron donor (both nitrous acid and oxygen) concentrations, relative to the ammonia-oxidizer enzyme pool appear to be required to induce hydroxylamine accumulation. Although hydroxylamine production is possible under transient conditions, it is unlikely to occur in functionally stable systems. The results of both pure and mixed culture studies conducted as part of this research suggest free ammonia does not appear to be inhibitory to nitrite-oxidizing organisms, at the concentrations commonly reported in the engineering literature. Nitrospira-like organisms, and not Nitrobacter, were the dominant nitrite-oxidizer in C S T R and S B R reactors treating synthetic wastewater. However, data collected as part of this research suggests that Nitrobacter may become dominant under high dissolved oxygen and short S R T conditions. Chemodenitrification of hydroxylamine is unlikely to be a significant source of nitrous oxide generation in wastewater systems. Several lines of evidence presented as part of this research program suggest substrate (both nitrite and oxygen) competition between ammonia- and nitrite-oxidizing organisms is l ikely the primary cause of observed nitrite accumulation in nitrification systems. The 301 unified ammonia oxidation model presented in the previous section explains many observed nitrification phenomena, including free ammonia inhibition. • The control of anoxic zone pH and, therefore, free ammonia concentration in a predentrifying nitrification/denitrification system (as practiced by Turk and Mavinic, 1986, 1987, 1989a, 1989b) is unlikely to provide long-term nitrogen removal with nitrite at the primary oxidized nitrogen species. Data collected as part of this research program have demonstrated the initial application of this strategy, which impacts nitrite-oxidizers most severely, and not free ammonia, is the dominant cause of nitrite accumulation. • Long-term nitrite accumulation is achievable in a nitrifying system; however, it would likely be in conjunction with only partial nitrification of ammonia and significant nitrous oxide emissions. The effluent from such a process would then be discharged to an anoxic reactor whereby ANNAMOX would be performed. This particular operating mode has been referred to as the CANON process. The benefits of such a process would have to be weighed against the potentially negative environmental impacts of the emission of nitrous oxide, a known green house gas. The primary contributions of this research program to engineering practice and research include: the development of an analytical method for hydroxylamine analysis in wastewater matrices, a preliminary indication on the feasibility of fatty acid analysis for analyzing mixed microbial populations in nitrifying systems operated under low carbon to nitrogen ratio, confirmation that free ammonia is not inhibitory to nitrite organisms as commonly reported, and the development of a conceptual model that explains apparent free ammonia inhibition. It is strongly hoped, that the results of this research program will change the engineering community's perspective on free 302 ammonia inhibition and nitrification processes, in general, and help refocus the approach taken by fliture researchers in this field. 9.4 Recommendations and Comments on Future Research Requirements This research program, like many before it, raises a number of unanswered questions that cannot be addressed during the tenure of a single PhD student. The recommendations are summarized as action items for future researchers and include: • The limited number of FISH images analyzed for this research program suggest that ammonia- and nitrite-oxidizer colony size and location within a floe particle will likely play a significant role in localized mass transfer, substrate (oxygen and nitrite) availability, and therefore the measured kinetic response to nitrogen transients. The role of reactor ammonia loading on nitrifier colony size and relative location should be studied using FISH and digital imaging technology. The affect that average colony size and relative location has on nitrous oxide generation and nitrifier kinetics should be investigated. This research could eventually lead to the development of a new generation of dynamic simulation models that are based upon actual physical phenomena, as opposed to an empirical best fit. • The results of this research suggest that conditions at the nitrite-oxidizer colony level dictate observed phenomena in nitrifying systems. There is a need for further studies, using a combination of microelectrodes and molecular methods, to further elucidate the important interactions between ammonia- and nitrite-oxidizers. 303 There is a wide spread need for research on the impact of multiple substrate limitations on bacteria in general and nitrifiers, specifically. A targeted study using a high nitrite feed could be used to measure the kinetic response of a nitrite-oxidizer enriched population to various nitrite and oxygen concentrations in order to confirm the nitrite-oxidizer response to dual substrate limitation. The results from this study would be of immediate use, when interpreting the results of the microelectrode studies discussed above. A l l o f R N A measured via slot blotting and/or F I S H is not necessarily from active organisms. There is a significant need for research programs that result in the quantification of active and non-active nitrifier R N A . Current literature evidence suggests that much of measured R N A ; particularly Nitrospira R N A , recovered from wastewater systems may not be participating in observed phenomena. The data presented as part of this research support the feasibility of partial nitrification, (which is the first part of the C A N O N process) as a viable option for nitrification/denitrification via nitrite. However, it appears that nitrous oxide generation w i l l be a byproduct of such a process. Research should be conducted to quantify the mass emissions of nitrous oxide from such processes and methods of reducing these emissions should be studied. The data collected as part of this research program suggest that hydroxylamine may play a significant role during transient nitrogen loadings (particularly when such transients result in a combination of low dissolved oxygen, high free ammonia, and low nitrous acid concentration). Research should be conducted with various ammonia transients, in 304 order to quantify the potential for short-term hydroxylamine production in wastewater systems and the potential impact of these on nitrifying systems. The microbiology community has made tremendous contributions to our understanding of the biochemistry of Nitrosomonads and Nitrobacter. We now know that Nitrospira are the dominant nitrite oxidizers in most wastewater systems; however, we still know very little about the biochemistry of these organisms. There is an immediate need for further research on the fundamental biochemistry of Nitrospira-like nitrite-oxidizers. Further research is required on the species diversity of Nitrospira-like nitrite-oxidizing organisms within wastewater systems. The culmination of the recommended research projects outlined above should lead to a more realistic nitrification model, most likely a flux model for wastewater systems, based upon organism interaction, as opposed to Monod kinetics. Such a model would be of significant utility but is unlikely to be developed until some or all of the other research projects, discussed above, have been completed. 305 Figure 9.1 - Conceptual Model Explaining Apparent Free Ammonia Inhibition 306 Bibl iography Anderson, L C , and Levine, J.S. 1986. Relative rates of nitric oxide and nitrous oxide production by nitrifiers, denitrifiers and nitrate respierers. App l . Environ. Biotechnol. 51(5): 938-945. Anthonisen, A . C , Loehr, R . C , Prakasam, T.B.S. , and Srinath, E . G . 1976. Inhibition of nitrification by ammonia and nitrous acid. J. Water Pollut. Control Fed. 48: 835-851. Bae, W . arid Rittmann, B . E . 1996. A structured model for dual-limitation kinetics. Biotechnol. Bioeng. 49: 683-689. -Cecen, F. , and Ipek, S. 1998. Determination of inhibition of ammonia-N and urea-N oxidations by the fed-batch reactor (FBR) technique. Wat. Sci . Technol. 38(1): 141-148. Frijlink, M . J . , Abee, T., Lanbroek, H.J. , de Boer, W. , and Konings, W . N . 1992. The bioenergetics of ammonia and hydroxylamine oxidation in Nitrosomonas europaea at acid and alkaline p H . Arch . Microbio l . 157: 194-199. Grady, C .P .L . , Daigger, G.T. , and Henry, C L . 1999. Biological wastewater treatment (2 n d Edition). Marcel Dekker Inc., N e w York. Hellinga, C , Schellen, A . A . J . C , Mulder, J.W., van Loosdrecht, M . C . M . , and Heijnen, J.J. 1998. The S H A R O N process: an innovative method for nitrogen removal from ammonium-rich waste water. Wat. Sci . Technol. 37(9): 135-142. 307 Hopper, A.B. 1968. A nitrite reducing enzyme for Nitrosomonas europaea. Preliminary characterization with hydroxylamine as electron donor. Biochim. Biophys. Acta. 162: 49-65. Hopper, A.B., and Dispirito, A. 1985. In bacteria that grow on simple reductants, generation of a proton gradient involves extracytoplasmic oxidation of substrate. Microbiol. Rev. 49(2): 140-157. Hunik, J.H., Meijer, H.J.G., and Tramper, J. 1993. Kinetics of Nitrobacter agilis at extreme substrate, product and salt concentrations. Appl. Microbiol. Biotechnol. 40: 442-448. Hyungseok Yoo, Kyu-Hong Ahn, Hyung-Jib Lee, Kwang-Hwan Lee, Youn-Jung Kwak, and Kyung-Guen Song. 1999. Nitrogen removal from synthetic wastewater by simultaneous nitrification and denitrification (SND) via nitrite in an intermittently-aerated reactor. Water Res. 33(1): 145-154. Schmidt, I. 2003. Proposed electron transport model for Nitrosomonas. In: The prokaryotes an evolving electronic resource for the microbial community (Release 3.14). Springer-Verlag, Heidelberg. Shiskowski, D.M. , Simm, R.A., and Mavinic, D.S. 2004. An experimental procedure for identifying the aerobic-phase biological source of nitrous oxide in anoxic/aerobic wastewater treatment systems (In press, J. Env. Eng. Sci. January 2004). 308 Spieck, E . , Enrich, S., Aamand, J., and Bock, E . 1998. Isolation and immunocytochemical location of the nitrite-oxidizing system in Nitrospira moscoviensis. Arch . Microbiol . 169: 225-230. Stuven, R., Vollmer, M . , and Bock, E . 1992. The impact of organic matter on nitric oxide formation by Nitrosomonas europaea. Arch. Microbiol . 158: 439-443. Suzuki, I., Dular, U . , and Kwok, S.C. 1974. Ammonia or ammonium ion as substrate for oxidation by Nitrosomonas europaea cells and extracts. J. Bacteriol. 120(1): 556-558. Turk, O., and Mavinic , D.S . 1986. Preliminary assessment o f a shortcut in nitrogen removal from wastewater, Can. J. C i v i l Eng. 13: 600-605. Turk, O., and Mavinic , D.S. 1987. Selective inhibition: a novel concept for removing nitrogen from highly nitrogenous wastes. Env. Technol. Lett. 8: 419. Turk, O., and Mavinic , D.S. 1989a: Stability of nitrite build-up in an activated sludge system. J. Water Pollut. Control Fed. 61(8): 1440-1448. Turk, O., and Mavinic , D.S. 1989b. Influence of process changes on maintaining nitrite build-up in an activated sludge system acclimated to free ammonia. Water Res. 23(11): 1383-1388. Yang, L . 1990. Investigation o f nitrite build-up within an enriched nitrification process. PhD Dissertation, School of C i v i l Engineering, Purdue University, West Lafayette, Indiana. 309 Yang, L. , and Alleman, J.E. 1992. Investigation of batchwise nitrite build-up by an enriched ntirification culture. Wat. Sci. Technol. 26(5-6): 997-1005. 310 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0063410/manifest

Comment

Related Items