Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Sidestream control of dissolved nutrients in anaerobically digested sludge centrate using anammox and… Galvagno, Giampiero 2016

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

Item Metadata


24-ubc_2017_february_galvagno_giampiero.pdf [ 5.66MB ]
JSON: 24-1.0339890.json
JSON-LD: 24-1.0339890-ld.json
RDF/XML (Pretty): 24-1.0339890-rdf.xml
RDF/JSON: 24-1.0339890-rdf.json
Turtle: 24-1.0339890-turtle.txt
N-Triples: 24-1.0339890-rdf-ntriples.txt
Original Record: 24-1.0339890-source.json
Full Text

Full Text

SIDESTREAM CONTROL OF DISSOLVED NUTRIENTSIN ANAEROBICALLY DIGESTED SLUDGE CENTRATEUSING ANAMMOX AND CHEMICAL PRECIPITATIONbyGiampiero GalvagnoM.A.Sc., University of British Columbia, 1998A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE COLLEGE OF GRADUATE STUDIES(Civil Engineering)THE UNIVERSITY OF BRITISH COLUMBIA(Okanagan)November, 2016© Giampiero Galvagno, 2016iiThe undersigned certify that they have read, and recommend to the College of Graduate Studiesfor acceptance, a thesis entitled:SIDESTREAM CONTROL OF DISSOLVED NUTRIENTS IN ANAEROBICALLYDIGESTED SLUDGE CENTRATE USING ANAMMOX AND CHEMICALPRECIPITATIONsubmitted by  Giampiero Galvagno  in partial fulfilment of the requirements of the degree ofDoctor of Philosophy .Dr. Cigdem Eskicioglu, School of EngineeringSupervisor, ProfessorDr. Bahman Naser, School of EngineeringSupervisory Committee Member, ProfessorDr. Rehan Sadiq, School of EngineeringSupervisory Committee Member, ProfessorDr. Jonathan Holzman, School of EngineeringUniversity Examiner, ProfessorDr. Wayne Parker, University of WaterlooExternal Examiner, ProfessorNovember 18, 2016iiiAbstractThe objective of this research was to assess specific side-stream processes for biodegradationand precipitation of dissolved nutrients in dewatering centrate and support the seamlessintegration of an anaerobic digester (AD) into a biological nutrient removal wastewater treatmentplant (WWTP). Anaerobic digestion processes release reactive and non-reactive dissolvednutrients which are returned to the mainstream treatment process in the dewatering centrate.Conventional side-stream treatment processes are typically designed for removal of reactivenutrients (ie, nitrate/nitrite, ammonia and orthophosphate). However, many WWTPs with lowtotal nitrogen and total phosphorus criteria will also be impacted by the non-reactive, difficult-to-degrade nutrient forms such as polyphosphates (poly-P), dissolved organic phosphorus (DOP)and dissolved organic nitrogen (DON).In this study, characterization was made of a conventional suspended growth deammonificationtreatment (Anammox) process for transforming poly-P, DOP and DON in two types ofdewatering centrate. The first centrate feed studied was from the full-scale Annacis IslandWWTP (AIWWTP), Canada. The second centrate feed was from a lab-scale AD fed wastesludge from the existing City of Kelowna Wastewater Treatment Facility (KWTF), Canada. TheAnammox process showed similar treatment characteristics for both the KWTF and AIWWTPcentrates with excellent DON removal and poor non-reactive dissolved phosphorus (NRDP)removal. A statistical comparison of the DOP and poly-P through the Anammox processsuggests that DOP has a higher biodegradation potential. Utilization of a post-Anammox, polyaluminum chloride chemical dosing, optimized based on zeta potential, was able to achieve theobjective of precipitating residual DON and NRDP and producing an effluent with lowerdissolved nutrients than the pre-digestion KWTF dewatering centrate scenario.ivAdditional testing was conducted on final effluent to characterize the dissolved phosphorus anddetermine an optimal coagulant dose. The testing suggests that the dissolved phosphorus ineffluent could be associated with colloidal dissolved organic matter (DOM). Furthermore,coagulant batch dosing experiments using poly aluminium chloride (PACL) and polyepichlorohydrin amine (polyepiamine) provides strong support for the use of zeta potentialmeasurements as a way to optimize coagulant dose.vPrefaceThe research program, including design and operation of the reactors and data analyses, wasexecuted by myself with assistance from Dr. Cigdem Eskicioglu. From time-to-time,characterization testing was conducted by collaborators, as described below.A version of Chapter 3 was presented at WEFTEC 2014 and the paper was published as part ofthe proceedings (Galvagno, G., Eskicioglu, C., Abbott, T., Cella, M., Gosselin, M., 2014. Thefate of recalcitrant nutrients through an anaerobic digester and anammox sidestream treatmentprocess. Proceedings of Water Environment Federation Technology and Exhibition Conference2014. Alexandria, Va: Water Environment Federation). The paper writing and all data analyseswere performed by myself. Dr. Cigdem Eskicioglu provided feedback and review of themanuscript. The majority of the lab data on which the paper is based was acquired by the firstauthor. Tim Abbott assisted in characterizing waste sludge (COD, VS/TS and biogas) from theanaerobic reactor. Monica Cella prepared samples and coordinated acquisition of scanningelectron micrograph images. Mike Gosselin, the managing supervisor of the KelownaWastewater Treatment Facility provided access to their in-house lab data and coordinatedadditional sampling to complete the effluent and centrate characterization.   A version of Chapter 4 was published by Water Research (Galvagno, G., Eskicioglu, C.,Abel-Denee, M., 2016. Biodegradation and chemical precipitation of dissolved nutrients inanaerobically digested sludge dewatering centrate. Water Research, 96, pp.84-93). Theexperimental design, reactor operation and all analyses were performed by myself. Marco Abel-Denee assisted in lab testing.A version of Chapter 5 was accepted as a paper for WEFTEC 2016 (Galvagno, G., Stevens, G.,Abel-Denee, M., Eskicioglu, C., Abbott, T.L., 2016. Effluent particulate phosphorusvifractionation and coagulation by optimized chemical dosing. Proceedings of Water EnvironmentFederation Technology and Exhibition Conference 2016. Alexandria, Va: Water EnvironmentFederation). The paper was written by myself, including all the analyses, and was presented atthe conference on September 28, 2016. The experimental setup was designed by myself withinput from Gerry Stevens of AECOM. Gerry provided the initial motivation to characterizecolloidal phosphorus and supported the research as a sponsor for the ENGAGE grant. Dr.Cigdem Eskicioglu provided important feedback and review through development of themanuscript. Marco Abel-Denee assisted in the coagulation dosing, phosphorus testing andfiltration steps.viiTable of contentsAbstract ................................................................................................................................... iiiPreface ....................................................................................................................................... vTable of contents .....................................................................................................................viiList of tables ............................................................................................................................xiiList of figures ......................................................................................................................... xiiiList of abbreviations .............................................................................................................. xviAcknowledgements ...............................................................................................................xviiiDedication ............................................................................................................................... xixChapter 1: Introduction ............................................................................................................ 11.1 Background and context ..............................................................................................11.2 Research goals & experimental hypotheses ..................................................................61.3 Research scope and methodology ................................................................................71.4 Novelty of research ......................................................................................................91.5 Thesis organization ......................................................................................................9Chapter 2: Literature review .................................................................................................. 102.1 Sidestream characteristics .......................................................................................... 112.2 Importance of sidestream management ...................................................................... 112.3 Potential impacts on final effluent .............................................................................. 132.4 Moving towards “Plant of the Future” ........................................................................ 162.5 Sidestream treatment processes .................................................................................. 172.5.1 Physical and chemical treatment ........................................................................ 18viii2.5.1.1 Ammonia stripping ..................................................................................... Acid stripping ........................................................................................ Steam and hot air stripping ..................................................................... Vacuum flash distillation ....................................................................... Summary of ammonia stripping options ................................................. Oxidation of ammonia ................................................................................ Break-point chlorination of ammonia ..................................................... Catalytic oxidation of ammonia.............................................................. Electrochemical oxidation of ammonia .................................................. Chemical precipitation................................................................................ Struvite precipitation .................................................................................. Membrane filtration .................................................................................... Forward osmosis .................................................................................... Liqui-Cel® membrane contactor ............................................................ Ion exchange and adsorption ...................................................................... 312.5.2 Biological treatment .......................................................................................... 332.5.2.1 Cannibal® sidestream process .................................................................... 332.5.2.2 Nitrification/denitrification ......................................................................... 352.5.2.3 Nitritation ................................................................................................... 372.5.2.4 Partial nitritation and anaerobic ammonium oxidation (Anammox) ............ 382.5.2.5 Bioaugmentation ........................................................................................ 472.5.2.6 Algae production ........................................................................................ 492.5.3 Summary of the literature review ....................................................................... 50ixChapter 3: Start-up, acclimation and characterization of lab-scale anaerobic digester andAnammox bioreactor performance ........................................................................................ 523.1 Context ...................................................................................................................... 533.2 Methodology ............................................................................................................. 553.2.1 Anaerobic digester ............................................................................................. 553.2.2 Sidestream nitritation-Anammox bioreactor ....................................................... 563.2.3 Sample preparation and testing .......................................................................... 643.3 Results ....................................................................................................................... 673.3.1 AD and Anammox reactor operation.................................................................. 673.3.2 Effluent and centrate characterization results ..................................................... 733.3.3 KWTF nutrient mass flow measurements .......................................................... 753.4 Discussion ................................................................................................................. 783.5 Summary of the bioreactor start-up phase .................................................................. 81Chapter 4: Biodegradation and chemical precipitation of dissolved nutrients inanaerobically digested BNR sludge ........................................................................................ 824.1 Context ...................................................................................................................... 834.2 Material and methods................................................................................................. 854.3 Anaerobic digester set-up/operation ........................................................................... 854.3.1 Sidestream partial nitritation-deammonification bioreactor set-up/operation ...... 854.3.2 Sample preparation and analysis ........................................................................ 874.3.2.1 Nitrogen & phosphorus testing ................................................................... 874.3.3 Dose optimization for nutrient precipitation ....................................................... 924.4 Results ....................................................................................................................... 92x4.4.1 Existing KWTF centrate characterization results ................................................ 924.4.2 Anammox performance results .......................................................................... 964.4.3 Centrate biodegradation through the Anammox process ..................................... 994.4.3.1 AIWWTP centrate feed .............................................................................. 994.4.3.2 Lab-scale KWTF AD centrate .................................................................. 1014.4.4 Summary of centrate biodegradation study ...................................................... 1034.5 Dose optimization for nutrient precipitation ............................................................. 1044.6 Summary of the centrate treatment characterization ................................................. 108Chapter 5: Final effluent particulate phosphorus fractionation and coagulation byoptimized chemical Dosing ................................................................................................... 1095.1 Research objectives & approach .............................................................................. 1115.2 Methodology ........................................................................................................... 1125.2.1 Phosphorus testing ........................................................................................... 1125.2.2 Particle size and zeta potential characterization ................................................ 1135.2.3 Coagulant Dosing ............................................................................................ 1135.2.4 Conventional filtration apparatus ..................................................................... 1145.2.5 Ultra-filtration (UF) fractionation .................................................................... 1155.3 Results & discussion ................................................................................................ 1155.4 Summary of the effluent particulate characterization ................................................ 123Chapter 6: Conclusions ......................................................................................................... 1246.1 Summary and conclusions ....................................................................................... 1246.2 Limitations .............................................................................................................. 126xiBibliography .......................................................................................................................... 128Appendices ............................................................................................................................ 150Appendix A – sample calibration curves ............................................................................... 150Appendix B – data & statistical analyses for autoanalyzer and IC measurements .................. 155Appendix C – data & statistical analyses for dissolved phosphorus testing ........................... 157xiiList of tablesTable 2.1 – Typical sub-processes and associated sidestreams ................................................ 11Table 2.2 – Characteristics of high-strength wastewater liquid sidestreams ............................. 12Table 2.3 – Typical acids used for ammonia stripping and associated reaction products.......... 20Table 2.4 – Operating parameters and observed inhibition thresholds for Anammox processes .......................................................................................................................... 41Table 2.5– Performance of nitritation-anammox systems treating AD digester sidestream ...... 46Table 3.1– Annacis Island WWTP centrate feed characteristics .............................................. 64Table 3.2 – Mesophilic anaerobic digester average feed and effluent characteristics ............... 70Table 3.3 – Comparison of influent/effluent characteristics for the bioreactors ....................... 73Table 4.1 – Full-scale AIWWTP and lab-scale Kelowna AD centrate feed characteristics ...... 87xiiiList of figuresFigure 1.1 – Generalized wastewater treatment schematic ........................................................1Figure 1.2 – Stages of anaerobic digestion ................................................................................3Figure 1.3 – BNR process schematic ........................................................................................5Figure 2.1 – Dissolved nutrient speciation .............................................................................. 15Figure 2.2 – Sidestream treatment options .............................................................................. 18Figure 2.3 – Fraction of unionized ammonia in solution at various pH and temperatures ........ 19Figure 2.4 – FO/RO process schematic ................................................................................... 29Figure 2.5 – Membrane contactor process schematic for ammonia removal ............................ 30Figure 2.6 – Cannibal sidestream treatment process schematic ............................................... 34Figure 2.7 – Biological transformation of nitrogen ................................................................. 36Figure 2.8 – Minimum residence time for ammonium and nitrite oxidizers ............................. 37Figure 2.9 – Schematic of Anammox attached growth structure .............................................. 44Figure 2.10 – Schematic of a bioaugmentation sidestream process.......................................... 48Figure 3.1 – Process schematic for the two-stage nitritation-Anammox reactor ....................... 57Figure 3.2 – Photo of the two-stage nitritation-Anammox setup at start-up ............................. 57Figure 3.3 – Process flow schematic for the lab-scale Anammox reactor ................................ 59Figure 3.4 – Photo of the single stage nitritation-Anammox reactor used in the research ......... 59Figure 3.5 – Anammox Sequence Batch Reactor Operation .................................................... 60Figure 3.6 – Photo of Anammox reactor sludge in suspension prior to hydrocycloning ........... 62Figure 3.7 – Particle size distribution of waste sludge after a cyclone and 180 micron sieve ... 62Figure 3.8 – Scanning electron micrograph of Anammox granules from the UBC reactor ....... 63Figure 3.9 – Operating data for the anaerobic digester ............................................................ 68xivFigure 3.10 - SEM images of digester struvite precipitate ....................................................... 70Figure 3.11 – Ammonia removal efficiency and effluent characteristics for the Anammoxbioreactor .......................................................................................................... 72Figure 3.12 – DON and NRDP flux through the Anammox SBR ............................................ 74Figure 3.13 – DON and NRDP flux through the mesophilic AD ............................................. 74Figure 3.14 – Non Reactive Dissolved Phosphorus Mass Flow Diagram for KWTF ............... 76Figure 3.15 – Dissolved Organic Nitrogen Mass Flow Diagram for KWTF ............................ 77Figure 4.1 - Kelowna WWTP final effluent nutrient speciation ............................................... 83Figure 4.2 – Comparison of IC and autoanalyzer for measurement of TN ............................... 88Figure 4.3 – Comparison of IC and probe for measurement of nitrate ..................................... 89Figure 4.4 – Comparison of nitrate measurements made using a probe and IC ........................ 90Figure 4.5 – Centrate feed and Anammox effluent dTN and dTP ............................................ 91Figure 4.6 – Nutrient fractionation for the existing KWTF dewatering centrate ...................... 94Figure 4.7 – Existing KWTF dewatering centrate NRDP and relationship to dTP ................... 95Figure 4.8 – Operating characteristics and results for the Anammox process .......................... 98Figure 4.9 – AIWWTP centrate and Anammox effluent dissolved nutrient characterization.. 100Figure 4.10 – KWTF AD centrate and Anammox dissolved nutrient characterization ........... 102Figure 4.11 - Modified KWTF schematic incorporating anaerobic digestion......................... 105Figure 4.12 – Impact of PACL coagulant dosing on the precipitation of dissolved nutrients . 106Figure 5.1 – Example of zeta potential response curve using a 2:1 ratio of PACL topolyepiamine ................................................................................................... 114Figure 5.2 – Average zeta potential and particle size of Westside Regional WWTP finaleffluent samples ............................................................................................... 116xvFigure 5.3 – Particle size distribution of Westside Regional WWTP final effluent samples ... 117Figure 5.4 – Fractionation of composite final effluent sample from Westside Regional WWTP ........................................................................................................................ 118Figure 5.5 – P fractionation of composite final effluent from Westside Regional WWTP usingultrafiltration ................................................................................................... 120Figure 5.6 – Concentrations of coagulant blends that result in neutral zeta potential andassociated costs ............................................................................................... 121Figure 5.7 – P fractionation of coagulated effluent using a blend of 2 parts PACL to 1 partpolyepiamine ................................................................................................... 122xviList of abbreviationsAD Anaerobic DigestionAnammox Anaerobic Ammonia OxidizingAIWWTP Annacis Island Wastewater Treatment PlantBNR Biological Nutrient RemovalBOD Biological Oxygen DemandCH4 MethaneCO2 Carbon dioxideCOD Chemical Oxidation DemandDAF Dissolved Air FlotationDEMON® DEaMmONificationDON Dissolved Organic NitrogendTN Dissolved Total NitrogendTP Dissolved Total PhosphorusDOP Dissolved Organic PhosphorusFPS Fermented Primary SludgeGC Gas ChromatographGHG Greenhouse GasHRT Hydraulic Retention TimeIC Ion ChromatographKWTF Kelowna Wastewater Treatment FacilityML Mixed LiquorxviiN NitrogenNRDP Non-Reactive Dissolved PhosphorusP PhosphorusPACL Poly Aluminum ChloridePAO Phosphorus Accumulating OrganismsPHA Polyhydroxyalkanoatepoly-P Polyphosphatepolyepiamine poly epichlorohydrin aminePS Primary SludgeSBR Sequence Batch ReactorSRT Sludge Retention TimeTKN Total Kjeldahl NitrogenTP Total PhosphorusTS Total SolidsTWAS Thickened Waste Activated SludgeUF Ultra FiltrationUV Ultra VioletVFA Volatile Fatty AcidVS Volatile SolidWAS Waste Activated SludgeWWTP Wastewater Treatment PlantxviiiAcknowledgementsI would like to express my gratitude to my supervisor, Dr. Cigdem Eskicioglu, for hermentorship, feedback and support over the years this research was undertaken. Her patience andencouragement were important for moving past some difficult technical challenges. I would alsolike to thank my committee members, Dr. Sadiq Rehan, Dr. Naser Bahman and Dr. HewageKasun for their feedback; and Dr. Wayne Parker for taking the time to be an examiner for mythesis defence and his thorough review. I am also very grateful for the assistance provided byoperations and management staff of the Kelowna WWTP, Annacis Island WWTP, WestsideRegional WWTP, Okanagan Falls WWTP and Penticton WWTP.Financial support for this work was provided by the Natural Science and Engineering Council(NSERC) Strategic Project Grant #396519-10, NSERC Engage Grant (F15-00588) and PacificCentury Graduate Scholarship.xixDedicationTo my wife, children, family and friends who gave me the courage to try and the inspiration topersevere.1Chapter 1: IntroductionThis chapter provides a general description of the wastewater treatment context that serves asthe motivation for the research.1.1 Background and contextAll WWTPs produce organics-rich sludge as a by-product. In general, the treatment processresults in two waste sludge streams (Figure 1.1). Primary sludge (PS) consists of the settleablesolids removed from the primary clarifier. Waste activated sludge (WAS) is the sludge removedfrom the secondary clarifier and is the by-product of biological treatment. The WAS consistspredominately of the micro-organisms (biomass) which are used in the biological treatmentprocess.Figure 1.1 – Generalized wastewater treatment schematicAs part of the wastewater treatment process, the sludge produced must be stabilized. Solidsstabilization involves converting sludge to biosolids which are low in odour and pathogens. ADis a common sludge management technique which produces relatively stable biosolids. The2anaerobic degradation of a complex wastewater or sewage sludge is a multi-step processcomprising four major reactions as illustrated in Figure 1.2. Complex organic polymers in thewaste are first hydrolyzed by extracellular enzymes of facultative or obligate anaerobic bacteria.The hydrolysis step provides monomeric/oligomeric constituents small enough to allow transportacross the cell membrane and is the limiting step in the AD process. These simple solublecompounds are then fermented, or anaerobically oxidized, to short chain fatty acid intermediates,alcohols, carbon dioxide, hydrogen and ammonia (acidogenesis). The short chain fatty acids(other than acetate) are converted to acetate, hydrogen and carbon dioxide (acetogenesis).Finally, methanogenesis occurs from carbon dioxide reduction by hydrogen, and form acetateresulting in the methane and carbon dioxide mixture that constitutes biogas. The biogas consistsof 60-70% methane, with the remaining 30-40% consisting of carbon, nitrogen, water vapour andreduced sulphur compounds. In wastewater context, WAS and PS serve as the substrate for AD.The WAS and PS are held in a sealed, oxygen free tank for 10 to 30 days, depending on thetemperature, to achieve 40 to 60% volatile solids destruction (Metcalf & Eddy, 2014). During theprocess biogas is continually removed from the reactor headspace.3Figure 1.2 – Stages of anaerobic digestionIn the United States, it is estimated that approximately 22% of the 16,000 WWTPs utilizeanaerobic digesters for stabilizing wastewater residuals. Furthermore, of treatment plants withanaerobic digesters, only 70 (2%) utilize the methane for generating electricity or heat (Lewis etal., 2008). While it is not clear how many facilities utilize the methane for other beneficialpurposes, the statistics suggest a low utilization of the methane generating potential in the UnitedStates and Canada.The low utilization of AD for biosolids stabilization reflects the historically poor economicreturn of constructing and operating an anaerobic digester. Conventionally, the economics of adigester have been evaluated based on comparison of the capital costs against the energy savings.However, rising energy costs, introduction of environmental commodities and markets changethe evaluation and outcome (Scanlan et al., 2010).As the economics and environmental benefits of implementing an anaerobic digester improvein the coming years, technical innovation must also advance to overcome operational challenges,HydrolysisAcidogenesis(Fermentation)AcetogenesisMethanogenesis1234Organic MatterSoluble Organics(sugars, amino acids, fatty acids)VolatileFatty AcidsCH4 + CO2H2 + CO2Acetic Acid2134 44improve on the efficiencies and fill the knowledge gaps. From an Okanagan perspective,additional work is required to reliably incorporate AD into the biological nutrient removal(BNR) processes which are common in the Okanagan Valley to prevent eutrophication in theOkanagan Lake.In the Okanagan Valley, dewatered sludge from the BNR treatment plant is taken off-site andcomposted (Figure 1.3). This approach to biosolids management prevents re-release ofphosphorus into the treatment process. Phosphorus is biologically removed in the BNR reactor ina two-step process (WERF, 2010). First, under anaerobic conditions, phosphorus accumulatingorganisms (PAOs) release their stored phosphate. Volatile fatty acids (VFAs) are simultaneouslyabsorbed and stored as a food source in the form of polyhydroxyalkanoate (PHA). In the secondstep, the PAOs are transferred to aerobic conditions where they utilize their stores of PHA forcell growth and re-absorb the soluble phosphorus previously released. If the conditions arecarefully controlled and sufficient VFAs are available in the anaerobic zone, the PAOs willabsorb phosphorus in excess of the amount originally released.5Figure 1.3 – BNR process schematicIf stored in the bacterial cell, net removal of phosphorus can be accomplished through normalwasting of biomass. However, if exposed to un-aerated conditions as part of the wasting processthe PAOs will re-release the stored phosphorus as with the first step. Consequently, any attemptto anaerobically digest the WAS triggers release of more than 90% of the total phosphorus asphosphate and solubilizes 45% of the total total kjeldahl nitrogen (TKN) (Mavinic et al., 1995).If allowed to return to the mainstream process, the nutrient-rich digester supernatant wouldeventually overwhelm the WWTP and result in elevated effluent nutrient levels.Use of composting as a sludge stabilization process eliminates the need to return phosphorus tothe main treatment process. The dewatered or thickened WAS is mixed with wood chips andaerated to produce a stabilized biosolids. There are currently six BNR facilities operating in theOkanagan Valley. The stabilized compost material from these facilities is marketed as a soilamendment.6Recycling biosolids as a soil amendment has significant benefits within a BNR context(Pitman, 1999). However, there are disadvantages to the process including relatively high energyexpenditure for operation of air blowers in the compost facility and transportation and thepotential for foul air emissions.An alternative option to composting for stabilizing biosolids is AD. In their comparison ofanaerobic and aerobic treatment technology, Cakir and Stenstrom (2005) show that from agreenhouse gas perspective, anaerobic treatment is more favourable at higher chemical oxygendemand (COD) concentrations because of the ability to recover methane.Methane gas could be utilized in a beneficial way (i.e., electricity generation) and digestioncould reduce the mass of biosolids by approximately 50%. The smaller volume would reduce thecost and energy associated with transportation and composting while retaining the nutrient valueof the final composted product.The benefits of implementing an anaerobic digester versus maintaining the status quo(composting biosolids) are attractive. However, there are some key challenges that would need tobe overcome in a BNR process context that will determine the feasibility of such a system. Inparticular, the sidestream treatment processes treating digester supernatant need to be able todegrade or remove the recalcitrant species that the mainstream process is not able to efficientlyremove.1.2 Research goals & experimental hypothesesThe main goal of this research was to fill the knowledge gaps associated with dissolvednutrients released through AD of BNR sludge and provide a management approach that wouldeliminate the risk of the sidestream nutrient loading on the mainstream WWTP. The research isbased on the hypothesis that both the reactive and recalcitrant nutrients can be removed through7existing biological sidestream processes or chemical precipitation. However, published researchand design manuals do not cover this issue and only a focused lab-scale bioreactor study willprovide the required basis for design. Specifically, objectives for this current research are to:1. Provide baseline characterization of the City of Kelowna’s Wastewater TreatmentFacility (KWTF) to understand the existing mass loading of nutrients within themainstream and sidestream flow streams;2. Characterize the performance of a mesophilic digester fed waste sludge from the KWTFwith emphasis on measuring dissolved nutrient concentrations in the waste sludgesupernatant;3. Assess the efficacy of a conventional deammonification sidestream centrate treatmentsystem for bio-transforming or removing dissolved, recalcitrant nutrients; and4. Characterize final effluent particulate and dissolved phosphorus and determine theefficiency of an aluminum-based coagulant for precipitating dissolved nutrients.1.3 Research scope and methodologyTo achieve the objectives stated in Section 1.2, a research program was proposed utilizing alab-scale anaerobic digester fed waste sludge from the KWTF and biological deammonificationprocess. The lab-scale KWTF digester would represent the digested sludge characteristics andassociated centrate sidestream. Centrate treatment to remove dissolved phosphorus and nitrogenspecies could then be assessed. The literature review identified the advantage of an ANaerobicAMMonia OXidation (Anammox) process over others for removal of ammonia. The centrate forthe lab-scale AD could be fed to a sidestream Anammox process to allow assessment ofdissolved recalcitrant nutrient removal. Anammox has the potential to biodegrade recalcitrantnutrients but this has never been measured. Subsequent dissolved phosphorus removal by8precipitation was assessed using aluminum salts. Ideally, phosphate removal by struviteprecipitation would be included as part of the process train but was impractical. Based on thegeneral description of the scope provided above, the following methodology was adopted:1. Collect samples of KWTF mainstream and sidestream flows to assess the mass flow of thevarious nutrient species. The data serves as a baseline to determine the level of treatmentrequired for the AD centrate. The testing also allows assessment of where the majority ofrecalcitrant dissolved nutrient originates or is removed.2. Establish a mesophilic AD using waste mixed sludge from KWTF as a substrate and assessthe dissolved nutrient flux through the digestion process.3. Start-up an Anammox process using centrate from Annacis Island WWTP (AIWWTP). TheAIWWTP centrate was used for the start-up to avoid limitations on centrate supply from thelab-scale AD since it was available in unlimited quantities.4. Once the Anammox reactor stabilizes, collect dissolved nutrient data for the AIWWTPcentrate feed and Anammox effluent to assess removal efficiency.5. After collecting dissolved nutrient data for AIWWTP, slowly change the substrate feed tocentrate generated from the lab-scale digester fed mixed sludge form KWTF. Once theAnammox process stabilized, collect dissolved nutrient data for the lab-scale KWTF ADcentrate and Anammox effluent.6. Assess the efficacy of using aluminum coagulants and flocculants for precipitatingdissolved nutrients in Anammox effluent and final effluent. Provide a detailedcharacterization of dissolved phosphorus in final effluent to determine size range andoptimized coagulant dosing for enhanced phosphorus removal.91.4 Novelty of researchThe current body of published research has not clearly documented nutrient flux through ADsystems fed WWTP sludge or sidestream treatment processes treating centrate. There isconsiderable available published research that is focused on removal of dissolved recalcitrantnutrients in the final effluent. However, the approach of optimizing sidestream treatmentprocesses to achieve the same result has not been documented. More specifically, a large amountof published research effort has been devoted to characterizing the performance of variousAnammox sidestream processes for removing ammonia but there are no available studiesshowing the propensity of this process for degrading dissolved recalcitrant nutrients.1.5 Thesis organizationChapter 2 is a detailed literature review which provides context to the issue of sidestreamtreatment and includes a description of the various treatment options for high-strength nitrogenand phosphorus sidestreams. The literature review lays the ground-work for selecting aconventional biological deammonification process for assessing recalcitrant nutrient degradation.In Chapter 3 and Chapter 4, characterization of the feed and waste sludges for the anaerobicdigester and deammonification bioreactors is made. In Chapter 5, particle characterization ofNRDP is presented along with the efficacy of coagulants for removing these species.Conclusions and recommendations for additional research are provided in Chapter 6.10Chapter 2: Literature reviewThe purpose of this chapter is to provide a detailed review of the various conventional andemerging sidestream treatment options for municipal WWTPs. Optimization of wastewatertreatment facilities from energy consumption and emissions stand-point necessitatesconsideration of the impact of the various internal sidestreams. Sidestreams from anaerobicsludge  digesters,  in  particular,  have  the  potential  to  be  a  significant  ammonia  and  phosphorusload to the mainstream treatment process. However, the literature suggests that managingsidestreams through their treatment in the mainstream process is not the most energy efficientapproach, nor does it allow for practical recovery of nutrients. Furthermore, as effluent criteriabecome more stringent in some jurisdictions and sludge hydrolysis pre-treatment for digestersmore common, an understanding of the fate of recalcitrant carbonaceous and nutrient species inthe digester supernatant becomes more important. A variety of sidestream treatment processesdescribed in the literature is reviewed in the context of these considerations.As a by-product of biological wastewater treatment, various minor and major sidestreams areproduced and must be managed through recycling or separate treatment. Sidestreams associatedwith the solids stabilization processes have the highest potential impact on the mainstreamprocess. Solids stabilization can release significant amounts of nutrients and organics to theliquid stream which must be further treated. As the trend towards greater energy efficiency andhigher effluent standards increases, more careful attention of the sidestream characteristics andmethod of treatment have become necessary. In order to gain a better perspective, a detailedsummary from the literature of the various sidestream treatment management options ispresented.112.1 Sidestream characteristicsAs part of a biological wastewater treatment process, several liquid and solids wastesidestreams are created that are not part of the main liquid or solids treatment trains. Table 2.1provides a list of conventional sub-processes and associated liquid and sludge sidestreams in atypical WWTP. The sidestreams occur as part of the requirements of the various treatment stagesand are typically recycled back or removed from the main treatment process.Table 2.1 – Typical sub-processes and associated sidestreamsTreatment StageSidestreams(Off-SiteRemoval)Potential Recycle SidestreamScreening Screenings Screenings dewatering underflowGrit removal Grit Grit wash waterPrimary clarification Scum (FOGs)a Scum dewatering underflowBioreactor (activated sludge)Bioreactor (membrane) Membrane backwashEffluent filter Filter backwashSludge thickener Thickener underflowSludge dewatering Dewatering centrateSludge composting LeachateOdour control Biofilter leachate or scrubberunderflowNotes:  a. FOG: Fat, oil and grease2.2 Importance of sidestream managementTreatment stages within the WWTP can produce high-strength waste streams which ifunmanaged can degrade the final effluent quality or increase energy demand. In particular,sludge dewatering centrate associated with AD processes are known to contain relatively highconcentrations of COD, ammonium and phosphate, as well as other nitrogen and phosphorusspecies (Table 2.2).12Table 2.2 – Characteristics of high-strength wastewater liquid sidestreamsLiquid SidestreamFlow(%ofInfluent)pHAlkalinity(mg/L)Temperature(°C)COD(mg/L)TSS(mg/L)TotalNitrogen(mg/L)Ammonia(mgN/L)TotalPhosphorus(mgP/L)Orthophosphate(mgP/L)ReferenceDAF underflow(WAS thickening)4.4 - - - - 6.6 - 2.0 - 0.35 Kelowna(2009/10)fDewatering centrate:BNR without sludgestabilizationb0.48 6.3 - - 2,150 674 4.9 27 208 181 Kelowna(2009/10)fEBPR with alumtrimming & mesophilicAD- - - - - - - 1,200 - 316-632eBauer 2010EBPR with mesophilicAD- 7.3-7.5- - - - - 756 - 207 Fujimoto et al.1991Trickling filter withmesophilic AD- 8.1 2,087 - - 250 - 989 - 88e Mavinic et al.2007Trickling filter withthermophilic AD- 8.0 3,295 - - 500 - 1,050 - 140e Mavinic et al.2007Trickling filter withATAD & alum dosing0.6 6.0 - - - 500 - 700 - 40 Salmon ArmWWTPgMesophilic AD &thermal hydrolysis- 7.8 9,500 -11,500- - <300 - 2,500-3,000e- - Figdore et al.2010aActivated sludge withmesophilic ADd- 7.9 - - - - - 707e - 157e Uysal et al.2010Activated sludge withmesophilic AD- 7.7 2,090 31.8 430 580 1,360 590 - - Carrio et al.2003Biofilter leachate - 7.2 –7.8- - - - 176 -245154 -210- - Shanchayan etal. 2004Notes: a. Published data provided are design valuesb. Sludge taken off-site and composted with no return streamc. “-“ indicates data not reportedd. Characterization based on the clear liquid component of lab-centrifuged sludge samplese. Calculated from Authors’ published dataf. Data provided by City of Kelowna, Canadag. Data provided by City of Salmon Arm, Canadah. DAF = dissolved air flotation; EBPR = enhanced biological phosphorus removal; BNR = biological nutrientremoval; AD = anaerobic digester; ATAD = autothermal thermophilic aerobic digester; TSS = total suspendedsolidsA review based on data from full-scale facilities and modelling studies (Jones and Takács,2004) show that substantial ammonia and phosphate is released in anaerobic digesters through13hydrolysis of WAS. Janus and van der Roest (1997) suggests that the nitrogen load from thesludge digestion sidestream contains up to 25% of the total nitrogen load while contributing only2% of the total flow to a treatment plant. This additional stream increases loading to themainstream process, resulting in larger bioreactors, increased energy expenditure and apotentially decreased effluent quality.2.3 Potential impacts on final effluentRecent research has highlighted the need to control recalcitrant nitrogen and phosphate speciesin order to optimize nutrient removal processes and reduce impacts on receiving waters. Thetrend of a decreasing effluent nitrogen and phosphorus objective for wastewater effluentdischarged to surface water bodies has resulted in a focus on understanding the formation andremoval of dissolved organic nitrogen (DON) (Pagilla et al., 2008) and non-reactive dissolvedphosphorus (NRDP) (Tooker et al., 2010). DON and NRDP make up a significant proportion ofthe effluent nitrogen and phosphorus in treated wastewater effluent and can be difficult todegrade within a mainstream wastewater treatment process (Neethling and Stensel, 2013). Thechallenge of managing effluent DON and NRDP is complicated by the secondary objective ofbeneficially reusing wastewater biosolids through AD. Stabilization of wastewater biosolidsthrough the use of AD can derive significant benefits to the overall WWTP operation in terms ofcost and carbon expenditures (Cakir and Stenstrom, 2005; Berg et al., 2013). AD of wastewatersludge represents an opportunity to generate biogenic methane which could be used to offsetdemand for conventional petroleum-based energy, thereby reducing carbon emissions (Verstraeteet al., 2005). In addition, AD is able to reduce the solids content of the sludge, reducingtransportation and disposal requirements of the dewatered biosolids. However, AD releases avariety of dissolved nutrients including DON and NRDP (Wilson et al., 2011; Wild et al., 1997).14If returned to the mainstream treatment process, dewatering centrate associated with AD couldnegatively impact the final effluent quality.DON consists of the nitrogen fraction which can be difficult to degrade. DON originates indomestic wastewater influent in a variety of forms including urea, amino acids, proteins,aliphatic N compounds and synthetic compounds, such as EDTA, N-containing pesticides andpharmaceuticals; however, DON can also be released into the treatment process through cellmetabolism processes that excrete biomolecules, cell decay and cell lysis (WERF, 2009). Non-reactive dissolved phosphorus (NRDP) is also difficult to remove with conventional treatment,requiring advanced treatment processes with multiple stages including filtration, coagulation andadsorption (Neethling et al., 2007). As with DON, NRDP can increase through biologicalproduction in the mainstream process (Neethling and Stensel, 2013).Published data from various full-scale WWTPs indicate a wide range of measured finaleffluent DON and NRDP concentrations. A survey of 197 nitrogen removal WWTPs in Virginaand Maryland reported an average final effluent DON of 0.93 mg N/L with a range of 0 to 2.5mg N/L (WERF, 2009). For WWTPs with biological phosphorus removal systems, the range offinal effluent NRDP measured in a separate study was 0.015 to 0.050 mg P/L (Neethling andStensel, 2013). While most of the DON and NRDP enters the mainstream bioreactor as part ofthe raw wastewater, internal recycle streams, like dewatering centrate and sludge thickenerunderflow, can also represent significant contributions. Published data of DON and NRDPconcentrations in dewatering centrate is not available. However, given the high potential forsoluble microbial products produced through cell metabolism and decay of the anaerobic bacteriaand WAS (Barker and Stuckey, 1999) it is expected that centrate will contain relatively highDON and NRDP concentrations.15DON is measured as the difference between total dissolved nitrogen and the sum of ammonia,nitrate and nitrite (Figure 2.1). Similarly, NRDP is the difference between total dissolvedphosphorus and dissolved reactive phosphorus. The orthophosphate concentration (molybdatereactive phosphate) is generally accepted as a measure of the reactive phosphorus (APHA, 4500-P A).Figure 2.1 – Dissolved nutrient speciationDON and NRDP can occur as part of the raw wastewater and resist degradation or increasethrough the biological treatment process, thereby contributing to the final effluent nutrient load(Arnaldos and Pagilla, 2010). Once in the receiving water environment DON can becomebioavailable, contributing to algae growth and negatively impacting the receiving environment.Using a 14-day bioassay, Urgun-Demirtas et al. (2008) were able to show that up to 61% of theeffluent DON was available for algae growth. The NRDP in WWTP effluent is bioavailable(Ekholm and Krogerus, 2003). Through algae bioavailability studies Liu et al. (2011)demonstrated that 67-90% of the dissolved organic phosphorus (DOP) is bioavailable. This studyalso showed that DOP is the most recalcitrant form of phosphorus remaining in highly treatedeffluents and contributes 22% to 89% of the bioavailable soluble effluent phosphorus.16New treatment technologies incorporated into plants can also inadvertently change theformation of DON and NRDP species. For example, Dwyer et al. (2008) determined that thecoloured compounds formed during high temperature sludge hydrolysis was found to bemelanoidins, a DON compound. The melanoidins were implicated in reduced ultraviolettransmissivity (UVT) and increased DON in the final effluent. Consequently, in this case study,the untreated digester return stream caused a failure of the UV effluent disinfection system.Follow-up research suggested that alum dosing could reduce melanoidin type humic substancesin the effluent (Dwyer et al., 2009). Arnaldos and Pagilla (2010) were also able to achieve 69%DON removal using alum at a molar ratio of 1.5. While chemical dosing could be used to controlthe increased DON production, additional research is needed to better understand and control theformation of these compounds.2.4 Moving towards “Plant of the Future”In a Water Environment Research Foundation (WERF) publication, the concept of the “Plantof the Future” is introduced (WERF, 2010). The “Plant of the Future” is a technology conceptbrainstormed by participants of a workshop attended by a range of experienced researchers,regulators, consultants and equipment manufacturers. The “Plant of the Future” is one thatprovides for energy self-sufficiency with greatest possible recovery of nutrients while providingeffluent of sufficient quality to allow for reuse. These new and innovative approaches serve asvision for sustainable wastewater treatment and identify research needs. Ultimately, the principleobjectives are to minimize the carbon footprint and greenhouse gas (GHG) emissions ofWWTPs.In order to achieve sustainability objectives for wastewater treatment, AD will need to featureprominently in the future (WERF, 2010; Verstraete et al., 2005). However, the potential17advantages of AD can be outweighed by impacts of the sidestream processes. For example,implementation of the ANaerobic AMMonium Oxidation (Anammox) process to removenitrogen from the digester return stream was a key feature in changing the energy balance at theStrass WWTP (Austria) from a net energy consumer to net energy producer (Wett et al., 2007a).Furthermore, selection of a particular sidestream treatment process needs to be consideredholistically to overcome the potential dichotomy of principles. Depending on the situation,minimizing energy consumption of the treatment plant may result in the most sustainableapproach. In other situations, resource extraction of nutrients or reduction in GHG emissionsmay need to serve as the priority objective despite the potential increase in energy consumption.Therefore, a clear understanding of the various potential sidestream processes and theirimplication on the energy balance, effluent quality, emissions and resource recovery arerequired.2.5 Sidestream treatment processesIn this section, potential sidestream treatment options documented in the literature arepresented and reviewed (Figure 2.2). Particular attention is paid to the implications for resourcerecovery and energy demand. Many of the sidestream processes documented address only oneparticular constituent and may necessitate more than one process to address the full spectrum ofnitrogen, phosphorus and organic loading. Consequently, this review can serve as a starting pointfor any specific assessment.18Figure 2.2 – Sidestream treatment options(adapted from Constantine and Johnson, 2006)2.5.1 Physical and chemical treatmentMany of the sidestream processes that have been researched or implemented rely on physicalor chemical removal of nutrients. This section provides an overview of conventional andpotential physical and chemical sidestream treatment processes. Ammonia strippingAmmonia gas reacts with water to form ammonium hydroxide ions and exist in equilibriumbased on the pH and temperature of the solution. Based on this chemistry, one approach toammonia removal is to shift the equilibrium toward the gaseous phase and then air stripammonia from the solution. At high pH and temperature, the equilibrium approaches 100%ammonia in the solution (Figure 2.3). While the stripped ammonia gas can be exhausted to theatmosphere, ammonia gas contributes to smog formation and can impact local ecosystems(Krupa, 2003). Ammonia reacts with strong acidic species in the atmosphere such as nitric and19sulphuric acids, by-products of vehicle and industrial combustion processes, to form ammoniumsalts. The ammonium salts become fine particulate matter or aerosols. In addition to the healthimplications associated with breathing the particulate matter, the ammonium salts are known tobe deposited locally and can contribute to eutrophication of water bodies and negatively impactterrestrial ecosystems (Krupa, 2003). The Intergovernmental Panel on Climate Change (IPCC)reports that ammonia can also lead to the formation of N2O from atmospheric chemicalreactions. However, N2O formation from NH3 is not fully understood and there is no methodavailable for estimating conversion in the atmosphere (Intergovernmental Panel on ClimateChange, 2006).Figure 2.3 – Fraction of unionized ammonia in solution at various pH and temperatures(adapted from Emerson et al., 1975)Various approaches can be used to strip ammonia gas from a sidestream. However, given thenegative impacts associated with uncontrolled ammonia gas discharges, the focus of this sectionwill be on processes which have potential for ammonia recovery; these include acid stripping,steam stripping and vacuum flash distillation processes. Acid strippingAir stripping in combination with absorption, can be used to remove and recover ammoniafrom dewatering centrate. First, the centrate pH is increased to convert the ammonium toammonia. Through agitation and turbulence the ammonia gas dissolved in the centrate istransferred to an air stream. In a second vessel, the ammonia gas is absorbed from the air into astrong acid solution to generate an ammonium salt. Selection of the acid depends on economicsand intended use of the ammonium salt solution. Typical acids and their associated ammoniumsalt are listed in Table 2.3. Removal of scale inside the tank can add to maintenance of such asystem. Also, low temperature conditions can affect performance by reducing the reaction rate.Table 2.3 – Typical acids used for ammonia stripping and associated reaction productsAcid used forammoniastrippingAssociated ammonium saltproduced by reaction withacidPotential reuse ofammonium salt as abulk solution or driedpowderSulphuric acid Ammonium sulphate(NH4)2SO4Fertilizer,flame retardantPhosphoric acid Mono-ammonium phosphateNH₄H₂PO₄FertilizerAmmonium phosphate(NH4)2SO4Fertilizer,flame retardantHydrochloric acid Ammonium chloride(NH4Cl)FertilizerAcetic acid Ammonium acetate(NH4C2H3O2)De-icerNitric acid Ammonium nitrate(NH4NO3)FertilizerLei et al. (2010) used acetic acid to remove ammonia at a pilot scale and at 25ºC achievedalmost complete removal (99.95%) of ammonia from the wastewater based on an inletconcentration of 21,006 mg/L ammonia nitrogen. However, it is unclear from the data presentedwhat proportion of the ammonia reacted with the acetic acid and was recovered as ammonium21acetate. The use of phosphoric acid to recover ammonia from a pilot-scale air stripping columnwas assessed by Minocha and Rao (1988). In this study, two reactors in series were operated tomaximize ammonium recovery and minimize ammonia gas emissions.As part of a sidestream treatment process selection exercise, District of Columbia Water andSewer Authority (USA) conducted a market assessment for production and sale of ammoniumsulphate generated from its Blue Plains Advanced WWTP (Eschborn et al., 2010; Figdore et al.,2010). Based on these assessments, air stripping combined with acid absorption and limeaddition for the production of ammonium sulphate, the most economical process of thetechnologies evaluated. However, product marketing and supply management represented riskswhich eliminated ammonia stripping as the preferred option. Consequently, biological treatmentremains the preferred primary method for sidestream treatment since the nitrogen is destroyedwithin the plant (Figdore et al., 2010).The VEAS WWTP in Oslo has been operating a full-scale acid stripping sub-process for morethan 10 years to remove ammonia from its anaerobic digester sidestream (Sagberg et al., 2006).The process produces ammonia nitrate as a by-product which is reused as a fertilizer. Tominimize operational costs, supply of the acid is tendered and the low bidder is able to receiveany volume of the ammonium nitrate solution at a reduced cost. The removal efficiency of theVEAS WWTP ammonia stripping process is limited on a practical basis to approximately 88%.Operating the system beyond a removal efficiency of 88% results in a rapid increase in energyconsumption and diminishes return. Steam and hot air strippingAmmonia can be stripped from centrate through contact with a high temperature carrier liquidor air stream. Typically, a packed tower is used to provide sufficient contact time. As centrate22flows down the packing, the counter-current steam or hot air contacts and vaporizes theammonia. The vapourized ammonia is transferred to a condensation column by the carrier liquidor air stream to form a concentrated ammonia solution. Alternatively, an acid reaction columncould be incorporated into the process to generate a stable ammonium salt.New York City has evaluated steam stripping of ammonia from dewatering centrate by meansof bench and pilot scale facilities and achieved 80 – 85% removal without pH control (Carrio etal., 2003). However, to minimize clogging of the packing, centrate screening and solids removalis required. A similar evaluation of hot air stripping achieved 90% removal but required additionof alkalinity (ibid.). Vacuum flash distillationVacuum flash distillation can be used as part of an ammonia stripping process to increase theefficiency of ammonia gas liberation from solution (Orentlicher et al., 2009). Flash distillation isapplied to the centrate after the pH has been elevated. Dissolved ammonia gas is removed byincreasing the temperature and lowering the pressure of the centrate stream. The process istemperature dependent. However, to achieve greater than 80% removal within a 10 minute batchprocess, the temperature of the centrate must be maintained at 43ºC (Orentlicher et al., 2009). Summary of ammonia stripping optionsAmmonia stripping has the potential to provide a robust and potentially energy efficientprocess for centrate treatment. A key challenge that must be overcome for such systems is howto manage the recovered ammonia. Reaction with acid to form ammonium salts has greatpotential for recovering a fertilizer product provided that the WWTP owner has the resources tomarket and distribute the product. Under specific circumstances, use of stripped ammonia as acompound for reducing emissions from fuelled electrical generating stations or incinerators may23be possible. Under controlled conditions, ammonia can be reacted with SO2, NOx and CO2 in theflue gases (Resnik et al., 2004). The by-product of this scrubbing is ammonium sulphate,ammonium nitrate, and ammonium bicarbonate, all of which could be reused as a fertilizer. If theammonia stripping process associated with a WWTP were in close proximity to a fossil fuelledgenerating station, the surplus aqueous ammonia could be used to scrub the flue gases.Alternatively, the ammonia could be used for scrubbing exhaust gases from a sludge incinerationor biogas co-generation facility. Additional research is required to determine whether usingammonia for exhaust gas scrubbing is a practical solution. Oxidation of ammoniaThis section provides a brief overview of options for chemically oxidizing ammonia. Break-point chlorination of ammoniaThe addition of chlorine to effluent either as gaseous chlorine or liquid sodium hypochlorite isconventionally used as a disinfectant but if sufficiently dosed will result in ammonia oxidationand removal. Chlorine reacts with ammonia in water to form various chloramines. However,more readily oxidizable compounds such as Fe2+, Mn2+, H2S and organics will react first with thechlorine. If the dose is further increased, chlorine will react with ammonia to form chloramineswhich contribute to an increase in the combined chlorine residuals. As the chloramines aredegraded, the total chlorine residual approaches a minimum which is associated with a maximumammonia removal. Following this minima, the chlorine residual begins to increase again ifchlorine dosing increases. The local minimum that indicates completion of ammonia oxidation isknown as the break-point chlorination demand. Chlorine dosing must be sufficient to achieve thecharacteristic break-point to provide for ammonia removal (Tchobanoglous et al., 2003).24The required chlorine dosing to achieve ammonia removal is dependent on pH and the amountof readily oxidizable compounds. In reasonably clean water with low suspended solids and a pHin the range of 6 - 8, the minimum chlorine dosing on a molar ratio was measured to beapproximately 8:1 (Pressley et al., 1972; Brooks, 1999). High alkalinity can increase the chlorinedemand to as high as 14.8:1 (Tchobanoglous et al., 2003). The presence of high suspended solidswould further increase chlorine demand and increase chlorine by-products.There do not appear to be any full-scale use of chlorine for ammonia removal in digestersidestreams. However, given that dewatering centrate typically has a moderate level ofsuspended solids and alkalinity, the chlorine dose would be at the high end of the publishedranges. At a dosing rate of 15:1, the amount of required chlorine would be prohibitivelyexpensive to provide removal of the typical range of anaerobic digester centrate ammonia (700 –4,000 mg/L NH4). Catalytic oxidation of ammoniaAt a temperature of 800ºC and in the presence of a platinum catalyst, ammonia gas can beoxidized to form nitric oxide (NO) which is further oxidized to nitrogen dioxide (NO2). Thereaction steps are:4NH3 + 5O2 → 4NO + 6H2O (Equation 2.1)2NO + O2 → 2NO2 (Equation 2.2)A catalytic oxidation system could be used to reduce ammonia gas removed from centratestreams. Under this kind of process, the centrate pH would need to be increased to shift the ionequilibrium toward the gaseous phase. Once liberated from the wastewater stream, the ammoniagas can be oxidized catalytically.25Nitrogen dioxide (NO2) is a photochemical oxidant in the atmosphere and in the presence ofhydrocarbons UV light is the primary cause of urban smog (World Health Organization, 2005).Consequently, uncontrolled release of nitrogen dioxide into the atmosphere as part of a centratetreatment system could contribute to existing environmental and health impacts in an urban area.To mitigate these impacts, catalytic oxidation of ammonia may require a post-treatment processto mitigate the local impacts of nitrogen dioxide. Electrochemical oxidation of ammoniaElectrochemical oxidation of ammonia could be accomplished by passing a current through ahigh-strength centrate stream. Lab-scale experiments by Lei and Maekawa (2007) resulted innearly 100% removal of ammonia after 5 hours of reaction time. The electrolysis processgenerates hypochlorous acid in the presence of sodium chloride. Hypochlorous acid reducesammonia to nitrogen gas (N2). However, in order to sustain the reaction, the wastewater streammust be dosed with sodium chloride and the electrode must be able to adsorb ammonia (Kim etal., 2006).Currently, there do not appear to be any full-scale applications of electrochemical oxidation forremoval of ammonia in centrate. The potentially high operating energy cost and issues aroundsodium chloride carry-over into the mainstream process would need to be addressed priorimplementing such a system. Chemical precipitationChemical precipitation of phosphorus using a cationic coagulant (iron, aluminum or lime) is aconventional approach in wastewater treatment systems is well documented in the literature.With these systems precipitated phosphorus is typically recovered in the dewatered sludge ordewatering centrate. Depending on the form of phosphorus (ie, dissolved, colloidal or suspended26particulate), coagulation may occur through adsorption destabilization, bridging mechanism orprecipitating enmeshment (Bratby, 2006). Metcalf & Eddy ( 2014) notes that historicallydissolved phosphorus coagulation using aluminum and iron salts has been described principallyin terms of reactions forming ferric or aluminum phosphate. However, the literature shows thatin wastewater systems, the adsorption of reactive and non-reactive phosphorus on to Al(OH)3whether in solution or as part of an alum sludge may be more significant (Galarneau and Gehr,1996). In wastewater, the reaction of aluminum and orthophosphate formed aluminumhydroxyphosphate which could enhance further phosphorus removal or contribute to organicsremoval through the formation of soluble and colloidal hydroxyl-aluminum-tannate complexes(Omoike et al., 1999). The finding that surface charge on alum sludge can be used forphosphorus precipitation has led to the notion of re-using water treatment residuals forphosphorus removal in wastewater (Babatunde & Zhao, 2010). Recycled alum sludge introducedinto a full-scale wastewater facility was shown to efficiently adsorb soluble phosphorus andcontinued to be effective even with increasing alum sludge age with the treatment process(Maher et al., 2015). Struvite precipitationStruvite (MgNH4PO46H2O) is a white, crystalline substance composed of magnesium,ammonium and phosphate. Struvite can precipitate under conditions where the pH is above 7.5and magnesium (Mg2+), ammonium (NH4+) and phosphate (PO43-) occur in equal parts (de-Bashan and Bashan, 2004). Historically, struvite formation in WWTPs has been a nuisance andassociated with increased maintenance. However, recent research has led to techniques forharvesting struvite from digester supernatant sidestreams, thereby serving as a way to removeboth phosphate and ammonia. Digester supernatant tends to be high in ammonia and moderately27high in phosphate but low in magnesium. Consequently, adding magnesium to the supernatantstream in a purposely designed reactor can sustain struvite precipitation and achieve 90%phosphate removal (Ueno and Fuji, 2001). However, due to the stoichiometry and molar mass ofphosphorus and nitrogen ammonia will always be less. For every gram of phosphorus removedas struvite, the nitrogen removal will be 0.45 g N. In anaerobic digestion scenarios, struviterecovery provides only nominal ammonia control. For example, in a mesophilic AD scenario thatproduces centrate with ammonia of 1,050 mg N/L and phosphate of 140 mg P/L (Mavinic et al.,2007), 90% removal of phosphorus as struvite translates to only 5.4% nitrogen removal. Wherelow nutrient concentrations are desired, struvite recovery would normally need additionalammonia treatment.A secondary benefit of struvite recovery from centrate is that the recovered struvite can be usedas a fertilizer, thereby displacing conventionally produced phosphate fertilizer. Barnard (2007)has highlighted the fact that phosphate is a limiting nutrient in agricultural production withworld-wide supplies estimated to be as low as 50 years. The phosphate conversion efficiencyincreases with pH and optimal struvite occurs at pH 9.5 (Hanhoun et al., 2009). However, controlof the removal efficiency can also be achieved through control of the supersaturation ratio(Mavinic et al., 2007). In this experiment, operating the reactor at pH 8 but with a highmagnesium dosage achieved over 90% phosphate removal efficiency.Air stripping of CO2, thereby increasing pH has also been studied for process optimization.Stumpf et al. (2009) used a bench-scale, airlift reactor to achieve 90% removal of the availablephosphate through struvite precipitation. The availability of a seeding material and adequatemixing energy are also important factors for controlling and optimizing struvite precipitation(Wang et al., 2006).28A commercialization of the struvite recovery process has led to a full-scale implementation atthe Durham WWTP (North Carolina, USA) (Bauer, 2010). Bauer describes a successfulimplementation of the Ostara® process (originally developed by the University of BritishColumbia, Canada) which achieves 89% phosphorus removal and 20% ammonia removal in thedewatered digested sludge centrate stream. Struvite recovery allows the Durham plant to meet itslow effluent phosphate limit without the use of chemicals. Membrane filtrationAdvances in membrane filtration technology have included processes for centrate treatment.Two processes which have been proposed for centrate treatment include forward osmosis andLiqui-Cel®, a proprietary membrane process. Forward osmosisForward osmosis (FO) is a technique for using osmotic pressure to concentrate a feed solution.In reverse osmosis (RO), hydraulic pressure is used to overcome the osmotic pressure of anaqueous feed solution and force the movement of water molecules across a membrane to producepurified water. In contrast to this, FO relies on the natural osmotic pressure between the aqueousfeed solution and an engineered draw solution (DS) (Cath et al., 2006). The osmotic pressuregradient between the feed solution and DS forces water molecules across a membrane, therebyconcentrating the feed solution.Holloway et al. (2007) applied the principle of FO to centrate treatment to produce aconcentrated centrate solution. Because FO is an osmotically driven process that requires verylow hydraulic pressure, the effects of fouling are reduced and membrane support and compactionare of less concern. To sustain the osmotic pressure gradient, Holloway et al. introduced a ROmembrane (Figure 2.4). In this system, reject water from the RO system is used to maintain a29constant concentration in the DS. Based on Holloway et al.’s research, an optimal water recoveryrate of 70% was established. At this rate, power consumption was approximately 4 kWh/m3.Figure 2.4 – FO/RO process schematic(adapted from Holloway et al., 2007)There are no reported full-scale facilities utilizing forward osmosis as a sidestream treatmentmethod. The ability to concentrate the centrate stream would potentially allow for a costeffective sidestream treatment if the concentrated centrate could be used as a liquid fertilizerwithout additional treatment. While this is technically possible, there would continue to bepotential challenges. Suspended solids and other organic constituents would be concentratedtogether with the nutrients. The odour potential, stability and utility of the concentrated solutionis required to assess whether additional pre or post treatment is required. Liqui-Cel® membrane contactorLiqui-Cel® is a proprietary process which utilizes a membrane contactor to selectively removeammonia from a wastewater stream (Liqui-Cel, 2009). In this process a hydrophobic membraneseparates the centrate solution at high pH (>11.3) and a sulphuric acid (H2SO4) solution (Figure2.5). At a pH of 11.3 or higher, the ammonia will be present in the centrate solution as freeammonia gas. The affinity with sulphuric acid causes ammonia to migrate across the porous30membrane and react to form ammonium sulphate ((NH4)2SO4). The membrane pore size andhydrophobic properties are designed to minimize the migration of water molecules or otherconstituents.Figure 2.5 – Membrane contactor process schematic for ammonia removal(adapted from Liqui-Cel, 2009)As suggested by Figdore et al. (2010), operational costs for such a treatment process will belargely dictated by chemical supply costs due to its heavy reliance on caustic for pH control andsulphuric acid to achieve the ammonia removal. The potential for membrane fouling and theneed to pre-filter the centrate have also been identified as potential limitations of this technology(ibid.). However, biofouling on the feed side of the membrane should not be problematic giventhe feed wastewater is at an elevated pH (> pH 11.3). The ability for the process to be containedwithin a compact, closed vessel and the potential to produce a high quality liquid fertilizer serveas advantages. These characteristics would allow for control of noxious air emissions andpotentially provide a revenue source from fertilizer sales.312.5.1.6 Ion exchange and adsorptionSidestream processes that utilize ion exchange or adsorption as a principle removal mechanismhave been investigated or implemented as a sidestream process. Ion exchange processes useresins which have the ability to exchange an ion adsorbed on the resin surface with a specificcation or anion in the centrate. The ion targeted by the process has a stronger affinity for thecharged location occupied by the ion and displaces it. In order to be effective, the exchanged ionneeds to have little or no impact to the main treatment process or effluent. Adsorption is similarto ion exchange except that no ion is exchanged as part of the process.The idea of using the ammonium adsorbing properties of vermiculite to create a fertilizerappears to have been first investigated by Allison et al. (1953). Vermiculite is a clay mineral witha high cation exchange capacity for ammonium. Naturally occurring magnesium on thevermiculite surface can be replaced by other cationic elements. The stability of the adsorbedcation is determined by the expansion of the lattice and its hydration properties. Ammonium isrelatively stable but can be replaced by sodium cations (ibid.). Akerback et al. (2009)investigated the use of spent vermiculite ion exchange media as a fertilizer. Sand-size particles ofvermiculite were used to adsorb ammonium and then used as a growing medium for pineseedlings and trees. The average maximum nitrogen content of the vermiculite was 1.44%,suggesting a removal rate of 14.4 NH3-N/kg. The growth characteristics of the ammonium-vermiculite amended seedlings were better than the untreated trees and those receiving nutrientirrigation.The use of vermiculite as a single use ion exchange media and fertilizer has potential as a low-cost method for control of ammonia for small scale facilities. However, on a large scale the massof required vermiculite makes this approach impractical. Assuming a sidestream flow of 0.5% of32total plant inflow and ammonia nitrogen concentration of 3,000 mg/L, a conventional anaerobicdigester for a moderately sized 20,000 m3/d treatment plant could produce 300 kg NH3-N/d for atreatment plant. At this rate and assuming an adsorption capacity of 1.44% ammonia andremoval rate of 90%, 19,000 kg/d of vermiculite would be required to treat the centrate. Theeffort and cost to transport, stockpile and reuse this mass of vermiculite would be onerous unlessa source of vermiculite and demand for the spent material were in close proximity to thetreatment facility. Carr et al. (2009) present results from a lab-scale experiment using ochre toadsorb phosphate from wastewater and then as a phosphorus fertilizer substitute. Depending onthe source of ochre, adsorption capacities up to 30.5 mg P/g ochre can be achieved. Based on thismaximum adsorption rate, 32.8 kg of ochre would be required to remove one kilogram ofphosphorus. As with vermiculite for removing ammonia, the economics of a full-scaleapplication of this technology would depend largely on the proximity of the mined ochre.A more conventional approach is to use a regenerative ion exchange media to remove ammoniaor phosphorus ions from the waste stream. Design considerations for ion exchange systemsinclude competition with non-target ions and management of target ion after regeneration.Clinoptilolite, a naturally occurring zeolite, was proposed as a sorption medium by Zorpas et al.(2010) to remove ammonium from digester sidestreams. However, clinoptilolite has a strongeraffinity for potassium than ammonium, thereby potentially reducing its efficiency as a centratetreatment method (Guo et al., 2008). Using centrate from a belt press as a waste stream,Thornton et al. (2007) was able to achieve a total loading capacity of 47 to 51 g NH4+-N /kg forMesoLite, a manufactured clay-aluminum ion exchange media. The ion exchange media wasregenerated using 5% NaOH solution. At this elevated pH (> 12), the ammonia-ammoniumequilibrium would shift to the gas phase. Consequently, regeneration of the ion exchange media33would result in release of ammonia gas. To mitigate the potential health and environmentalimpacts discussed previously, this approach to ammonium removal should include a secondaryprocess to capture the ammonia gas.Various approaches have been studied for regeneration of the ion exchange media. Nitric acidand sulphuric acid were assessed to allow nitrogen recovery as ammonium nitrate andammonium sulphate (Pawlowski, 1979). Biological regeneration of the media has also beenconsidered (Lahav and Green, 1998). Sengupta and Pandit (2010) showed the feasibility of usinga polymeric anion exchanger impregnated with iron oxide nanoparticles to selectively removephosphate and allow for phosphorus recovery. In these tests, regeneration with sodium chlorideand sodium hydroxide consistently recovered more than 95.0% of sorbed phosphate within tenbed volumes. Phosphorus can be recovered from the spent regenerant through addition of achemical to precipitate the phosphate.2.5.2 Biological treatmentBiological processes have been applied to sidestream treatment. In this section, conventionaland new systems are summarized. Cannibal® sidestream processThe Cannibal® process is a sidestream process used to reduce the solids production of aconventional activated sludge treatment plant. By reducing the sludge volume, the process ispotentially able to minimize the sidestream loading from the solids digestion process. In theCannibal® process, approximately 10% of the return activated sludge stream from the secondaryclarifier is routed through an anaerobic bioreactor (Figure 2.6) with an optimum hydraulicretention time of 7 days (Easwaran et al., 2009).34Figure 2.6 – Cannibal sidestream treatment process schematic(adapted from Novak et al., 2007)At a lab scale, the Cannibal® system has resulted in a 60 - 63% reduction in sludge productioncompared to a control treatment system with aerobic digestion (Novak et al., 2007; Goel andNoguera, 2006a). The lower sludge production results in lower power consumption for operatingthe aerobic digester; reduced sludge disposal costs; and reduced polymer consumption (Siemens,2011). The literature does not specifically assess whether there is an increased energyrequirement associated with the additional readily degradable organic load. However, Novak etal. (2011) notes an increased air utilization rate in the aeration phase of the mainstreambioreactor.The reduction in solids is strongly associated with release, under anaerobic conditions, of ironand protein bound up in the extracellular polymeric network of the floc (Park et al., 2006). Underthese conditions, the released organic matter, principally proteins are rapidly degraded whenreturned to the activated sludge tank. Novak et al. (2011) also showed that feed patterns can havea significant impact on the iron release: under a fast feed (high substrate pressure) scenario a35higher iron and protein release was detected in the Cannibal® reactor, resulting in a higher solidsreduction than the slow feed scenario.The fermentation that is associated with a sidestream anaerobic process has shown to havebenefit to biological phosphorus removal processes by increasing the volatile fatty acid contentof the return stream (Goel and Noguera, 2006a; Goel and Noguera, 2006b). The approach offermenting a slip-stream of return activated sludge to enhance biological phosphorus removalwas explored by Barnard et al. (2010a).An approach similar to the Cannibal® process for solids reduction which involved recirculatingpart of the sludge stream from a thermophilic anaerobic digester at the Rockaway WWTP (NewYork, US) was reported by the Torpey et al. (1984). In this study, 67% of the sludge on a volatilesolids basis from the mesophilic-thermophilic digester was returned to the mainstream activatedsludge treatment process and resulted in a 55% reduction in the overall sludge production fromthe plant. During the five month testing period the overall biogas production increased slightly.Following this pilot study, similar sludge reduction benefits were measured at other treatmentplants in the New York area (Carrio et al., 1985). However, the same approach of recyclinganaerobically digested sludge was tested at the Hanover WWTP (Pennsylvania, US) and did notresult in any overall solids reduction (Prakasam et al., 1990). Due to the reduced operating costs,the Cannibal® may be an attractive sidestream process for treatment plants that employ aerobicdigesters for sludge stabilization. However, for treatment plants that rely on AD of sludge, thenet benefit of recycling anaerobically digested sludge is not assured. Nitrification/denitrificationIn AD, many of the by-products are nitrogen-free (CH4, CO2, H2, H2S) with most of the boundnitrogen released as ammonia (Wett et al., 2010). Conventional biological treatment of ammonia36nitrogen involves nitrification of ammonia to nitrate (ܱܰଷି) based on Equation 2.3 and Equation2.4:2 ∙ ܰܪସା + 3 ∙ ܱଶ → 2 ∙ ܱܰଶି + 4 ∙ ܪା + 2 ∙ ܪଶܱ (Equation 2.3)2 ∙ ܱܰଶି + ܱଶ → 2 ∙ ܱܰଷି (Equation 2.4)Following nitrification, nitrate (ܱܰଷି) is reduced to nitrogen gas under anoxic conditions withcarbon serving as the electron donor (Equation 2.5):C10H19O3N + 10∙NOଷି® 5×N2+ 10×CO2 + 2×H2O +NH3 +10×OHି (Equation 2.5)The process converts the ammonia nitrogen to nitrogen gas (Figure 2.7) but requires relativelyhigh energy expenditures. The observed oxygen requirement for removing biochemical oxygendemand (BOD) varies between 0.90 – 1.3 kg O2/kg BOD; this compares to 4.57 kg O2/kg N foroxidizing ammonia to nitrate (Metcalf and Eddy, 2014).Figure 2.7 – Biological transformation of nitrogen(adapted from Stinson, 2001)372.5.2.3 NitritationNitritation is the first step in the nitrification process in which ammonia is oxidized to nitrite(Equation 2.3). The SHARON (Single reactor High activity Ammonia Removal over Nitrate) is acontinuous flow, nitritation process originally developed to treat high ammonia in dewateringcentrate streams. The process operates at a low sludge retention time (SRT) and relatively hightemperature of 35ºC (Figure 2.8). The elevated temperature conditions results in high specificgrowth rates which obviates the need for sludge retention; in addition, ammonia oxidizingbacteria (AOB) out-compete the nitrite oxidizing bacteria (Hellinga et al., 1998). Operating apartial nitrification process to treat high ammonia waste streams reduces aeration requirementsand bioreactor sizing over a complete nitrification process.Figure 2.8 – Minimum residence time for ammonium and nitrite oxidizers(adapted from Hellinga et al., 1998)Since ammonium (NH4+) rather than ammonia (NH3) serves as the substrate, pH becomes animportant operational parameter (Hellinga et al., 1999). An increased pH will result in a lowerammonium concentration and elevated effluent ammonia concentration (van Dongen et al.,382001). Furthermore, nitrous acid (HNO2) is inhibitory to the AOB and used as a substrate fornitrite oxidizing bacteria (NOB). To minimize the inhibition effects of NHO2 and ensure theAOB’s continue to out-compete the NOB’s, the nitritation reactor needs to be operated at anoptimum pH between 6.5 and 8 (van Hulle et al., 2007). To provide for pH control, adenitrification stage can be added to replace the lost alkalinity. However, this necessitatesaddition of methanol as a carbon source. Van Hulle et al. (2007) determined that provision of adenitrification stage with addition of methanol to control pH was more economical than chemicalpH control (i.e., NaOH). Recently, Mayer et al. (2009) was able to satisfy the denitrificationcarbon requirements using a slip-stream of the anaerobic digester feed sludge, thereby furtherimproving the economics.Garrido et al. (1997) investigated the stability of a fixed film reactor for nitritation as part of anitrification/denitrification process. At a loading rate of 5 kg NH3-N/m3, the lab-scale biofilmairlift suspension (BAS) reactor was able to achieve a stable conversion of ammonia to equalparts nitrite and nitrate.The nitritation processes took on new significance with the discovery of the anaerobicammonium oxidation (Anammox) process. This feature is discussed below in greater detail as itis also the sidestream process used in this thesis. Partial nitritation and anaerobic ammonium oxidation (Anammox)In 1995, Mulder et al. published a paper that described a new bacterium which can oxidizeammonia under anaerobic (anoxic) conditions. The results were significant because anaerobicoxidation of ammonium was not thought to be feasible (Kuenen, 2008) or at best a theoreticalpossibility as was suggested by Broda in 1977 (as cited by Kuenen, 2008). The Anammoxbacteria were eventually classified as belonging to the group Planctomycetes and contain a39membrane-bound organelle in which ammonium and nitrite are converted to nitrogen gas usinghydrazine as an intermediate (Kartal et al., 2010). The newly discovered bacteria were assignedto the sub-group Candidatus of the Planctomycetes phylum within which five genera havecurrently been identified: Anammoxoglobus, Brocadia, Jettenia, Kuenenia, and Scalindua(Kuenen, 2008). From a sidestream treatment perspective, processes which utilize the Anammoxbacteria are showing great potential for removing ammonia-N with low consumption of energyand no supplemental carbon addition.After Mulder’s original paper, additional research confirmed that the anoxic oxidation ofammonium is biologically mediated (van de Graaf et al., 1995) and that nitrite was the preferredelectron acceptor (van de Graaf et al., 1996). The original research had suggested nitrate wasbeing used as an electron acceptor. The role of nitrate in the ammonia oxidation process is notclear. Research has indicated that under certain conditions Anammox bacteria can also convertnitrate to nitrogen gas while oxidizing an organic energy source (Kartal et al., 2007).In the late-1990’s, the SHARON process was coupled to an Anammox process to achievedenitrification without addition of a supplemental COD source or methanol (Jetten et al., 1997;1998). The SHARON reactor was used to partially nitritate an anaerobic digester supernatantwaste stream. Effluent from the SHARON reactor was then used as an influent to the Anammoxreactor following the partial nitritation (Equation 2.6) and anaerobic ammonia oxidation(Equation 2.7) (van Dongen et al., 2001).NHସା + HCOଷି + 0.75×Oଶ → 0.5×NHସା + 0.5×NOଶି + COଶ + 1.5×HଶO (Equation 2.6)0.5×NHସା + 0.5×NOଶି → 0.5×Nଶ + HଶO       (Anammox) (Equation 2.7)Equation 2.8 below describes the Anammox step and accounts for the nitrogen mass balanceand cell growth (Strous et al., 1998).40NHସା + 1.32×NOଶି + 0.066×HCOଷି + 0.13×ܪା→ 1.02×Nଶ + 0.26×NOଷି+0.066×CHଶO଴.ହN଴.ଵହ + 2.03×HଶO (Equation 2.8)Based on the above reaction rates and piloting of the SHARON-Anammox configuration, thecost of removing nitrogen from waste streams containing high ammonium and low organicmatter, like digester supernatant, could be reduced by up to 90% (Jetten et al., 2001).Factors reported in the literature which have been shown to impact the Anammox processinclude temperature, dissolved air, pH, nitrite, nitrate, ammonia, phosphate and methanol. Table2.4 summarizes the range of operation and inhibition values reported in the literature. Theavailable data indicates that dissolved air and nitrite can have a strong inhibitory impact on theAnammox process. Nitrite toxicity is a rate limiting criterion for many Anammox reactorconfigurations and results in an over-designed bioreactor (van der Star et al., 2007).Consequently, a better understanding of the factors that result in nitrite, an electron acceptor foroxidation of ammonium, being toxic to the Anammox bacteria could lead to more efficientbioreactor designs. Phosphate and nitrate can also be inhibitory at higher concentrations.However, the large range of inhibition values reported for nitrite and phosphate should motivatefurther investigation. Process configurations, testing protocol or the dominant planctomycetegenera could all be factors in the variation of nitrite and phosphate inhibition values indicated inTable 2.4.41Table 2.4 – Operating parameters and observed inhibition thresholds for AnammoxprocessesParameter Criteria Observed Impacts TestConfigurationReferenceTemperature, 20 – 43 -Stable SBR Strous et al. (1999)ºC 37 -Optimum at pH 8 Fixed volume assays  Egli et al. (2001)25 -Stable MBBR Szatkowska et al. (2007)Dissolved Air,mg/L>0.2 -Increasing inhibition UASBb Jung et al. (2007)1.2c -Irreversible inhibition after 64h Fixed volume assays Egli et al. (2001)pH 6.7 – 8.3 -Stable SBR Strous et al. (1999)6.5 -61% better N-removalefficiency than pH between 7.5and 8.1MBBR Jaroszynski et al. (2011)8.0 -Optimum at 37ºC Fixed volume assays Egli et al. (2001)8.5 -Elevated nitrite concentration SBR Lopez et al. (2008)TSS, mg/L na - - -Ammonium-N,mg N/L980 -Stable SBR Strous et al. (1999)770 -50% inhibition Fixed volume assays Dapeena-Mora et al. (2007)90 -No effect SBR Bettazzi et al. (2010)Ammonia-N,mg N/L13 - 90 -Inhibition observed Batch test Waki et al. (2007)>2 -Increasing inhibition UASBb Jung et al. (2007)Nitrite-N, 350 -50% inhibition Fixed volume assays Dapeena-Mora et al. (2007)mg N/L 274 -Threshold inhibition Gel entrapped media Kimura et al. (2010)185 -Complete, reversible inhibition Fixed volume assays  Egli et al. (2001)102 -Decreased activity SBR Lopez et al. (2008)100 -Complete inhibition SBR Strous et al. (1999)70 -Decreased activity UASBb Jung et al. (2007)60 -Strong inhibition SBR Fux et al. (2002)60 -Decreased activity (spiked) SBR Bettazzi et al. (2010)30 -Decreased activity (prolonged) SBR Bettazzi et al. (2010)Nitrate-N, 980 -Stable SBR Strous et al. (1999)mg N/L 630 -50% inhibition Fixed volume assays Dapena-Mora et al. (2007)57 -No effect SBR Bettazzi et al. (2010)Phosphate, 1,990 -50% inhibition Fixed volume assays Dapeena-Mora et al. (2007)mg/L 1,900a -No inhibition Fixed volume assays Egli et al. (2001)155 -Complete inhibition Fluidized bed van de Graaf et al. (1996)95 -No inhibition Fluidized bed van de Graaf et al. (1996)HydrogenSulphide, mg/Lna -Increased activity UASBb Jung et al. (2007)Acetate, mg/L 118a -16% increase in nitrite-reducingactivityBatch test withpurified cultureGűven et al. (2005)Propionate, mg/L 73a -24% increase in nitrite-reducingactivity increaseBatch test withpurified cultureGűven et al. (2005)Methanol, mg/L 106a -51% inhibition Batch test with gelentrapped mediaIsaka et al. (2008)16a -Complete loss of activity Batch test withpurified cultureGűven et al. (2005)Notes: a. Calculated from authors’ published datab. Upflow Anaerobic Sludge Blanket (UASB); Sequence Batch Reactor (SBR); MBBR: Moving Bed BiofilmReactorc. Calculated based on 18% oxygen saturation value provided by authorBased on the research, phosphate inhibition is an important factor in the design of a full-scaleAnammox process treating centrate associated with digested BNR sludge. Bauer (2010) notes42that peak phosphate in the centrate from the Durham Advanced WWTP digester which is fedsludge from an enhanced biological phosphorus removal process can range between 900 and1,800 mg/L PO4 (300 – 600 mg/L PO4-P). The peak centrate phosphate concentrations measuredat the Durham facility are within the range of inhibitory values documented in Table 2.4.Ammonium has some inhibitory impact but this may be influenced by shifts in the ammonia-ammonium equilibrium caused by changes in pH. Jung et al. (2007) were able to show a clearcorrelation between free ammonia and Anammox inhibition and suggested a maximum freeammonia concentration less than 2 mg/L. At high ammonium-nitrogen loading and temperatureconditions associated with digester supernatant treatment using Anammox, pH control becomesan important operating variable. Under these conditions, small pH increases can result in arelatively large increase in free ammonia potentially leading to ammonia toxicity. Furtherresearch is required to support this argument.Many studies use an operating temperature above 30ºC. However, Anammox bioreactors canbe operated at lower temperatures but achieve less than optimal nitrogen removal rates.Szatkowska et al. (2007) successfully operated a pilot Anammox reactor at 25ºC which was feddigester supernatant. At this temperature the need for external heating was eliminated therebyoff-setting the reduced biological process efficiency.In the presence of propionate, Anammox activity increased, resulting in increased CO2production and higher level of nitrite consumption (Gűven et al., 2005). The results suggest thatthe propionate was oxidized to CO2 using nitrite or nitrate as an electron acceptor. Other researchshows that hydrogen sulphide can also enhance Anammox activity by reducing the oxygenreduction potential (Jung et al., 2007).43Single tank nitritation-Anammox processes have been developed as an alternative to the tworeactor processes where nitritation is controlled in a separate aeration tank. A completelyautotrophic nitrogen removal over nitrite (CANON) process was first described by Third et al.(2001) and utilized a SBR to develop a mixed culture of nitrosomonas-like aerobic andplanctomycete-like anaerobic ammonium oxidizing bacteria. The CANON concept of using asingle stage reactor has continued to expand to include attached growth fixed film reactors,moving bed biofilm reactor (MBBR), membrane and SBR systems.DEMON (DeamMONification) is an alternative mixed culture SBR process proposed by Wettet al. (2007) which utilizes Anammox bacteria and incorporates two significant advancements.First, the process control was simplified, using pH as a primary control variable. Duringnitritation, the mixed liquor pH is allowed to decrease to a lower set-point, after which theaeration is turned off creating the anoxic condition for deammoniafication which in-turnincreases the pH. Similarly, when the pH reaches an upper set-point, the aeration turns on. Bycycling the aeration on and off in this way, the nitritation and deammoniafication stages can bebalanced. The second advancement was incorporation of a mechanism for retaining Anammoxculture. Anammox bacteria grow slowly, having a doubling time of 10 – 12 days at 35°C (Kartalet al., 2010) which necessitates a long sludge age. However, in order to ensure nitritation isselected over nitrification, a sludge age of approximately 1.5 d is required for the aerobic culture(nitrosomonas population). To recover Anammox bacteria in the waste activated sludge whilestill achieving nitritation, a technique was developed for using a cyclone-like device as a selector(Wett et al., 2010). The cyclone recovers the granular sludge that is associated with theAnammox bacteria, thereby allowing for a mechanism to increase Anammox biomassconcentration.44Anammox attached growth processes have shown good potential because of their ability toretain biomass. Several biofilm reactor configurations have been developed, including granularsludge, fluidized bed biofilm reactors, airlift reactors, and rotating biological contactors. Oncethe bacteria are established, Anammox systems are robust (van der Star et al., 2007).Characteristics of an Anammox biofilm were reported by Siegrist et al. (1998) in a rotatingbiological contactor and later confirmed by Egli et al. (2001). Furukawa et al. (2001) reported theresults of research on a biofilm process which utilized Anammox bacteria. Further work byFurukawa et al. (2006) led to a patent for Single-stage Nitrogen removal using Anammox andPartial nitritation (SNAP). Veolia (2011) has incorporated the Anammox culture into a singlestage moving bed biofilm reactor (MBBR), known as ANITATM MOX process. The Anammoxbacteria form in an interior layer of the biofilm, below the aerobic layer where any dissolvedoxygen is consumed and nitrite is produced through oxidation of ammonium (Furukawa et al.,2001). A schematic of this biofilm concept is provided as Figure 2.9.Figure 2.9 – Schematic of Anammox attached growth structure(adapted from Veolia Water, 2016)The ability of the Anammox culture to form compact, fast settling granules has been studiedextensively because of the potential to achieve high biomass concentrations. Microbes grow ingranules in response to high velocities that would otherwise wash out the lighter floc and in45effect results in biofilm-type growth (Beun et al., 2000). Imajo et al. (2004) first demonstratedthe feasibility of using granular Anammox sludge in an upflow reactor configuration at both thelab and pilot scale. Fernandez et al. (2008) later investigated factors associated with theformation of the granules. Influent with high salt content which promote precipitate formation orzeolite particles which serve as support for biofilm growth resulted in increased biomassretention compared to the control (ibid.). Active Anammox granules form gas tunnels as theygrow and can result in trapped pockets of gas, leading to sludge floatation (Chen et al., 2010). Toavoid the potential reactor instabilities associated with floating sludge, Chen (2010) showed thatit was possible to mechanically break-up the granules thereby re-establishing a well-settlinggranular biomass.Membrane bioreactors have also been configured to utilize an Anammox process. Themembrane is well-suited to biomass retention and results in the fastest start-up of any reactorsystem at the lab scale (Wang et al., 2009) with low membrane fouling if a granular biomass isachieved (Trigo et al., 2006). However, the high retention efficiency of membrane systems canresult in accumulation of non-volatile, inert material which can inhibit Anammox growth (Trigoet al., 2006).Table 2.5 provides a summary of operating and nitrogen removal data from select pilot-scaleand full-scale facilities. Data on large-scale facilities is limited. In some cases the mean reportednitrogen removal efficiencies in Table 2.5 likely under-represent the actual removal efficiencies.For example, the Rotterdam (NL) facility shows exceptional ammonia oxidation efficiency in theAnammox stage (van der Star et al., 2007) but the nitritation stage is likely over-sized since itwas originally designed to nitritate full ammonium load (Mulder et al., 2001), a feature notrequired in an Anammox system. Consequently, the calculated efficiency on a total tank volume46basis of the Rotterdam sidestream treatment system is likely under-estimated. As the collectiveoperational experience increases, it is expected that a more accurate representation of thetreatment efficiency for different configurations will become available.Table 2.5– Performance of nitritation-anammox systems treating AD digester sidestreamFacility/LocationReactorTypeaNitritationStageReactorVolume(m3)AnammoxStageReactorVolumee(m3)Temp.(ºC)MeanNitrogenRemovalCapacityd(kgN/m3×d)ReferenceRotterdam(NL)GSR 1,800 70 30-40 0.37 van der Star et al., 2007;Mulder et al., 2001Hattingen(DE) MBBR 104 158 20-300.26 –0.32b Rosenwinkel et al., 2005Himmerfjarden(SE) MBBR - 2.1f 25 0.36 Szatkowska et al., 2007Strass (AT) SBR - 500 28 0.61c Wett et al., 2007bZurich (CH) SBR 2.0 1.6 30 0.30b Fux et al., 2002Notes: a. Granular Sludge Reactor (GSR); Moving Bed Biofilm Reactor (MBBR); Sequence BatchReactor (SBR)b. Calculated from authors’ published datac. NH4-N removal reportedd. Calculation based on sum of nitritation and Anammox reactor volumese. Total volume for single stage reactorsf. Does not include volume of final settling tankGiraldo (2009) measured bursts of nitrous oxide (N2O) generation which was associated withaccumulation of nitrite in the culture. The burst of N2O is proportional to the concentration ofnitrite and is also repetitive. Kampschreur et al. (2009) reported a N2O emission rate equal to1.2% of the nitrogen load to a full-scale nitritation-Anammox sequence batch reactor. However,Sliekers et al. (2002) reported negligible N2O emissions (0.05%) from a lab-scale Anammoxreactor. Based on the literature, the potential exists for nitrous oxide emissions from thenitritation stage associated with Anammox processes. The research suggests that higher aerationrates increases N2O emissions (Kampschreur et al., 2009). As a result, minimizing aeration couldminimize N2O emissions. In the case of the Zurich WWTP, Joss et al. (2009) show that even47assuming a higher N2O emission rate, the total CO2 equivalent is more favourable for thenitritation-Anammox process over conventional nitrification/denitrification. In large part, this isa consequence of the CO2 associated with the external carbon source requirement forconventional nitrification/denitrification.While the factors which contribute to N2O production in an Anammox process are not fullyunderstood, the propensity for higher N2O emissions could make selection of Anammoxunfavourable on a carbon footprint basis under specific circumstances. In particular, if a local,waste by-product were available as a carbon source, the nitritation-Anammox process couldresult in a higher CO2 equivalent value than conventional nitrification/denitrification. BioaugmentationSidestream treatment of high ammonia centrate has been used to augment the mainstreamtreatment process. In these nitrification/denitrification sidestream treatment systems, excessnitrification sludge is returned to the mainstream process to further increase the overall nitrogenremoval capacity of the plant (Figure 2.10).48Figure 2.10 – Schematic of a bioaugmentation sidestream process(adapted from Kos, 1998)Centrate treatment combined with bioaugmentation of the mainstream process is particularlyadvantageous for wastewater treatment facilities located in colder climates since nitrification ishighly temperature dependent. The dewatering return stream is at an elevated temperature andallows for a compact nitrification process. Kos (1998) presented the theoretical basis for abioaugmentation process which suggested a 60% reduction in the required SRT in themainstream process over a conventional facility. Consequently, a sidestream nitrification processwith bioaugmentation could reduce the bioreactor volume of the mainstream treatment plant.Salem et al. (2002, 2003) undertook model studies to assess the benefits of a bioaugmentationbatch enhanced (BABE) sidestream process for upgrade of the Walcheren WWTP and HoutrustWWTP in Netherlands. In the BABE process, part of the return activated sludge (RAS) streamalong with the dewatering return stream is routed through a sidestream reactor. The results fromthese studies indicate that the BABE process could reduce the mainstream bioreactor volume49requirements by lowering the effective SRT and allow for nitrification at lower wintertemperatures. A modelling study reported by Sova et al. (2004) showed similar benefits.Baker et al. (2009) utilized a SBR to treat 1,600 kg/d of ammonia nitrogen from dewateringcentrate associated with mesophilic digesters at the North End Water Pollution Control Centre inWinnipeg (Canada). The system was designed to allow the mainstream high purity oxygenactivated sludge treatment process to operate at a low SRT time of two days while still achievingnitrogen removal. While the full-scale sidestream SBR operated as designed, the nitrogenremoval efficiency of the mainstream treatment process did not change, indicating thatbioaugmentation was not taking place. In a separate study, a detailed microbiological assessmentof a similar bioaugmentation failure attributed the loss to a shift in the predator-prey dynamics(Bouchez et al., 2000). Bouchez speculates that adding a massive population to the mainstreamprocess in order to induce a biological activity results in a disturbance of the ecosystemequilibrium. Consequently, a successful bioaugmentation process needs to provide for protectionof the bacteria from grazing by protozoa. The BABE process addresses this by using a smallRAS fraction from the mainstream treatment process and short SRT. This maximizes thecultivation of organisms and a floc structure best suited to the mainstream process (Salem et al.,2003). Algae productionCentrate has been shown to be a good substrate for algae production and is an approach thathas attracted research interest in recent years (Arita et al., 2015). Algae treatment systems havethe potential to be low cost, simple and robust. Furthermore, algae grown on centrate could beused as a feed-stock for biofuel production (Kong et al., 2010) or co-digested with municipalsludge to produce methane gas as a fuel (McGinn et al., 2011; Cecchi et al., 1996).50The green algae Chlorella sp. has been shown to be well suited to being grown in dewateringcentrate from a municipal WWTP (Min et al., 2011; Li et al., 2011). The work suggests that thealgae systems treating centrate can remove 89.1% of the total nitrogen and 80.9% of the totalphosphorus (Li et al., 2011). The feasibility of recovering algae and the impact of anyunrecoverable algae in the effluent on the mainstream process are key considerations foroperating such systems (Barnard et al., 2010b). In addition to harvesting and dewateringchallenges, other bottlenecks in algae production include changing microbial diversity andabundance with seasonal variations in light, resulting in performance instabilities as measured bynutrient removal (Zhang et al., 2015) and energy output of AD fed excess algae. Also, the moreproductive algae grown in wastewater have strong cell walls to protect themselves frompredators and require pretreatment maximize methane yield of anaerobic digestion (Ras et al.,2011).2.5.3 Summary of the literature reviewAs a commodity, ammonia is necessary for production of numerous industrial and agriculturalproducts. Most industrial ammonia is produced through conversion of a hydrocarbon source,usually natural gas but also coal and heavy oil, and requires significant energy input. Anydeammoniafication process ultimately sustains the demand for ammonia from conventionalsources and the associated CO2 emissions. The potential for N2O emissions fromdeammoniafication processes (i.e., either through atmospheric or biologically mediatedconversion of ammonia) further shifts the balance in favour of ammonia recovery processes forhigh strength sidestream treatment. As our understanding of advanced biological treatmentprocesses and their impact on the environment improves, the best application of a specificprocess should become clearer. The Anammox process, for example, is showing great potential51as an energy efficient ammonia removal system, therefore has been implemented as the processfor the next chapters of this thesis. However, a better understanding of the factors which effectN2O emission rates will undoubtedly result in a more optimized process. The impact of thephysical and biological sidestream processes on recalcitrant nutrient species (i.e., DON andNRDP) is an area that requires further research. In the context of treatment plants with lownutrient limits, sidestream treatment processes which are able to reduce the recalcitrant nutrientload to the mainstream process will be more desirable. The criteria used for selection of aparticular sidestream treatment will be unique to each treatment plant.52Chapter 3: Start-up, acclimation and characterization of lab-scale anaerobicdigester and Anammox bioreactor performanceThe objective of this chapter is to provide detailed information on start-up, acclimation anddissolved nutrient characterization data from a lab-scale partial nitrification, deammonification(Anammox) bioreactor and anaerobic digester (AD). In addition, results are presented of asampling program aimed at mapping out the mass flow of dissolved nutrients within thesidestreams of the full-scale City of Kelowna Wastewater Treatment Facility (KWTF). In effect,this chapter provides the context for a more detailed assessment of the effectiveness ofAnammox for treating dissolved nutrients described in Chapter 4.The lab-scale AD was fed waste sludge from the existing KWTF. In parallel to the ADpiloting, the Anammox sidestream treatment reactor was operated to assess its efficacy forremoving various nutrient species from dewatering AD centrate, including dissolved organicnitrogen (DON) and non-reactive dissolved phosphorus (NRDP). For comparison, dissolvednutrient characterization was made of sidestream and mainstream flows from the KWTF. TheKWTF data are intended to serve as a baseline for the research and to understand the mass fluxof DON and NRDP through a full-scale process. Reactor studies were conducted at UBC’sBioreactor Technology Laboratory at Okanagan Campus from 2012 to 2016. The results showthat the Anammox process is able to degrade a large fraction of the DON (90%) and to a lesserextent the NRDP (12%). The research suggests that additional removal of sidestream dissolvedphosphorus (orthophosphate and NRDP) would be required to allow an AD to be incorporatedinto the KWTF without compromising effluent quality.533.1 ContextUp until the 1980’s, increasing nutrient loading to Okanagan Lake (Canada) from the dischargeof secondary treated wastewater was resulting in algae blooms and fish kills. Construction ofBNR tertiary wastewater treatment plants (WWTPs) in the Okanagan starting in the 1980s wasinstrumental in reversing the water quality deterioration of the Okanagan Lake system (Oldhamand Rabinowitz, 2002). However, the current trend in the Okanagan is towards more energyefficient sub-processes while maintaining or improving the quality of the effluent discharged toOkanagan Lake. Anaerobic digestion of wastewater sludge represents an opportunity to generatebiogenic methane which could be used to offset demand for conventional petroleum-basedenergy, thereby reducing carbon emissions (Verstraete et al., 2005). In addition, AD is able toreduce the solids content of the sludge, reducing transportation and disposal requirements of thedewatered biosolids. Anaerobic digestion is not used as a sludge stabilization process in BNRWWTPs located in the Okanagan because of the potential to re-release nutrients taken up by theprocess. Instead dewatered sludge is hauled offsite and composted.The literature review (Chapter 2) highlighted the need to control dissolved recalcitrant nitrogenand phosphate species in order to optimize nutrient removal processes with high effluent qualitycriteria. In particular, dissolved organic nitrogen (DON) and non-reactive dissolved phosphorus(NRDP) appear to be the more recalcitrant nutrient forms and have the most risk to the finaleffluent quality. In addition to reactive forms of nitrogen and phosphorus, AD is expected toproduce a relatively large fraction of DON and NRDP. However, the literature does not providea characterization of the DON and NRDP in AD centrate nor the degree of biodegradability insubsequent sidestream treatment processes.54The primary objective of the research described in this chapter was to measure the capacity of amesophilic AD and Anammox sidestream reactor for producing or removing DON and NRDP.This first part of the research fills a gap in the literature by providing the characterization dataneeded to assess the impact of an AD on a BNR WWTP. The AD is expected to release a largeamount of nutrients into the supernatant, including DON and NRDP. The Anammox process hasshown to be an effective and energy efficient process for removing ammonia from ADdewatering centrate (Wett et al., 2007b). However, the literature does not demonstrate ifAnammox has the capability to also remove DON and NRDP species. If this potential exists thenan Anammox sidestream treatment reactor could support the introduction of an AD in a BNRfacility by simultaneously removing ammonia and recalcitrant dissolved nutrients.To achieve the primary research objective, characterization was made of the supernatant from alab scale AD fed mixed sludge from the Kelowna Wastewater Treatment Facility (KWTF). TheKWTF was selected to serve as the test case; but the proposed approach could be applicable toother similar wastewater and sludge management scenarios. KWTF is one of 6 tertiary WWTPswhich operate in the Okanagan Lake basin (BC, Canada), a phosphorus-sensitive watershed. Tomitigate impacts to the Lake, the KWTF utilizes a BNR (modified Bardenpho) process. TheKWTF has effluent criteria of total nitrogen (TN) < 6.0 mg N/L and total phosphorous (TP) <0.25 mg /L on an annual average basis. Other WWTPs discharging effluent to the lake systemhave a similar standard.The KWTF mixed sludge feed consisted of fermented primary sludge (FPS) and thickenedwaste activated sludge (TWAS). The AD supernatant characterization attempts to measure therelative concentrations of the various forms of nitrogen and phosphorus species in the AD sludgefeed and effluent sludge. A lab scale Anammox reactor was also operated to assess its ability to55remove dissolved nutrient species from dewatering centrate taken from a full-scale thermophilicAD. Comparison of the DON and NRDP characterization results between the lab-scale reactorsand KWTF centrate was used to determine on a preliminary basis the feasibility of incorporatingan AD at KWTF and assessing the need for additional treatment or research.A second research objective was to understand the mass flux of DON and NRDP in the fullscale KWTF and provide a baseline context to the research. To this end, testing of mainstreamand internal sidestream flows was conducted for the KWTF. The data were intended to establisha maximum target concentration for DON and NRDP in the return stream and to understandwhere the recalcitrant nutrients originate.3.2 MethodologyIn this section, a description of the research setup and testing protocols is provided. In the firsttwo sub-sections, setup and operation of the AD and Anammox reactors are outlined. Followingthis, testing instruments, protocols and sample preparation techniques are documented.3.2.1 Anaerobic digesterThe 5.0 L anaerobic digester was started up on January 16, 2012. The digester was operated asa mesophilic process (37oC) and fed mixed TWAS and FPS from KWTF at a ratio of 66:34percent by volume until it was shut-down on June 27, 2015. The solids content of the mixedsludge feed was typically 4.0 to 5.0% by weight. The digester operated at a sludge retention time(SRT) of 20 days. The organic loading rate averaged 9.19 g VS/d (1.84 g VS/L/d).A New Brunswick BioFlo® 115 glass fermenter/mixer and control centre was used for thedigester reactor setup. The digester produced on average 250 mL/d of effluent sludge. Thecontrol centre allowed for a semi-continuous automated feeding on an hourly basis. Peristaltic56pumps allowed for sludge wasting followed by a feed interval. The waste and feed volumes werechecked daily and pump rates adjusted accordingly.Inoculum used for the mesophilic digester was taken from the City of Penticton (Canada)WWTP. The City of Penticton digesters were operated as mesophilic digesters with an SRT of14 – 20 days and temperature of 36 - 37°C. The digesters were fed FPS. Total and volatile solidsof the mesophilic inoculum was 2.35% and 1.64% at the time of AD start-up.3.2.2 Sidestream nitritation-Anammox bioreactorThe nitritation-Anammox bioreactor was originally designed to be a two-stage reactor with thenitritation and Anammox processes in separate tanks (Figure 3.1, Figure 3.2). The nitritationstage was designed with a settler integrated into the outer annular space. The Anammox tank wasan upflow reactor designed to retain the heavier granules in the lower section of the tank.Centrate was pumped into the nitritation tank at regular intervals with the effluent flowing to theAnammox reactor by gravity. A recycle pump transferred Anammox effluent back to thenitritation tank. By controlling the recycle flow rate, the hydraulic retention time of the centratein the nitritation reactor could be varied depending on the nitrite concentration. A second internalrecycle pump was included to control the upflow velocity, thereby allowing for growth of largergranular sludge. Inline heaters were used to maintain a reactor temperature at 35˚C. In theory,separating the nitritation and Anammox stages would allow for more efficient process control.57Figure 3.1 – Process schematic for the two-stage nitritation-Anammox reactorFigure 3.2 – Photo of the two-stage nitritation-Anammox setup at start-up58The two stage design was started mid-February, 2011 and by July, 2011, the system wasabandoned for a more conventional single stage design. Poor sludge retention in the nitritationreactor resulted in intermittent effectives of the two-stage design. The nitritation stage wasseeded with activated sludge from the KWTF and initially good aerobic ammonia oxidationcapacity was achieved. However, over a week or two the aerobic ammonia oxidizer sludge waswashed out and collected in the Anammox reactor resulting in deteriorating ammonia removalperformance.The single stage nitritation-Anammox bioreactor that ultimately proved successful wasconfigured as a 5.5 L sequence batch reactor (SBR), operated at a temperature of 34oC. A NewBrunswick BioFlo® 115 glass fermenter/mixer and controller were used for this purpose and theprocess was started-up in October, 2012. The bioreactor controller and associated computersoftware allows for automated operation of a centrate feed pump, waste pump, alkalinity feedpump and air valve. Operation of the nitritation-Anammox SBR is based on the DEMONprocess, as detailed by Wett et al. (2007), also in the literature review Section A flowschematic of the process and photo of the setup used this study is provided in Figure 3.3 andFigure 3.4.59Figure 3.3 – Process flow schematic for the lab-scale Anammox reactorFigure 3.4 – Photo of the single stage nitritation-Anammox reactor used in the researchThe Anammox SBR operated on 8 hour cycles with 7 hours of biological reaction and one hourfor sludge wasting and effluent decant (Figure 3.5). The 7-hour react phase was further dividedinto 35, 12 minute increments during which an equal volume of centrate was fed to the SBR. The60centrate would cause the pH to increase approximately 0.01 to 0.02 units. For 4 minutes thereactor would slowly mix, allowing for an anoxic phase during which the Anammox bacteriaoxidized ammonia using the available nitrite.Figure 3.5 – Anammox Sequence Batch Reactor OperationFollowing the anoxic phase, the air valve would open to provide an aerobic phase for the next 8minutes. The air flow would be controlled by pH and dissolved oxygen (DO). A DOconcentration of 0.20 to 0.25 mg/L would be maintained for the duration of the 8 minute aerobicphase. The low DO combined with the relatively high operating temperature (34°C) allows forgrowth of aerobic ammonia oxidizing bacteria while suppressing growth of nitrite oxidizingbacteria. As a result, if the process is working properly some of the ammonia will have beenconverted to nitrite during the aerobic cycle and be available for the Anammox bacteria at thestart of the anoxic period. Introduction of the centrate at the start of the anoxic period may also61enhance the Anammox process by either providing a source of VFA or reducing the oxygenreduction potential, as suggested by the literature (Gűven et al., 2005, Jung et al., 2007).Nitrate was monitored on a daily basis with ammonia and nitrite assessed 2-3 times per week.Monitoring of the MLSS and SRT were not necessary for process control and were measuredintermittently. Sludge wasting was based on measurements of nitrite and nitrate. In general, anincrease in nitrate usually signaled the need for additional wasting through the hydrocyclone andscreen.The average hydraulic retention time (HRT) and SRT of the Anammox reactor was 2.6 daysand 57 days, respectively, at an average mixed liquor suspended solids concentration of 13,700mg/L. HRT was measured as the reactor working volume divided by the average volume ofcentrate feed. SRT was measured as the mass MLSS at the start of the react phase (ie, afterwasting is complete) divided by the total wasted mass TSS.The SBR sludge wasting technique utilized in this research was modified compared to thepublished DEMON operation in order to resolve issues associated with scale. The DEMONprocess relies on a hydrocyclone for retention of granular sludge. However, the relatively smallwaste flow rates of the UBC lab-scale pilot cannot generate sufficient g-forces in thehydrocyclone to provide for adequate retention of the granular sludge. For this lab-scale reactor a180 micron screen was successfully used to further process the sludge and maximize retention ofthe granules. This approach of using screens for physically isolating granular sludge wasreported by de Clippeleir et al. (2013). A Mastersizer 3000 with a (Malvern Instruments) sizerange of 0.01 – 3500 micron was used to monitor the size distribution of the mixed liquor in theAnammox bioreactor and the waste sludge. A photo of the reactor sludge is provided in Figure623.6 and shows the granules in suspension. The relative influence of the hydrocylone and 180micron sieve on the retention efficiency of granular sludge is depicted in Figure 3.7.Figure 3.6 – Photo of Anammox reactor sludge in suspension prior to hydrocycloningFigure 3.7 – Particle size distribution of waste sludge after a cyclone and 180 micron sieve01020304050607080901000 100 200 300 400 500 600 700 800 900 1000 1100 1200PortionofParticlesSmaller,%Particle Size, micronsWaste Sludge After 180μ ScreenWaste Sludge After CycloneSBR Mixed Liquor63A scanning electron micrograph of a typical sludge granule isolated using the 180 micronscreen is provided as Figure 3.8. The micrograph shows an asymmetrical structure that appearsto be an agglomeration of smaller granules. However, no attempt was made in this study toidentify the factors affecting granule morphology.Figure 3.8 – Scanning electron micrograph of Anammox granules from the UBC reactorAs part of the process control, an additional sludge return step was incorporated under specificconditions to control effluent nitrite concentrations. Nitrite levels in the DEMON process abovea mean level of 1.7 mg/L have been shown to decrease Anammox activity (Wett et al., 2007b).In this research, it was observed that addition of settled, stored overflow sludge from the 180micron screen helped to reduce nitrite levels and stabilize the process. Consequently, whennitrite increased above 1.7 mg/L in the SBR, a small amount of reserve sludge was returned tothe bioreactor. The flow schematic shown in Figure 3.3 includes this additional, intermittent step.The Anammox reactor was started in November, 2012 using centrate from the Annacis IslandWWTP (AIWWTP), Metro Vancouver, Canada, as a substrate feed. Given the limited volume of64centrate output from the lab-scale AD, centrate from AIWWTP was used as a feedstock tostabilize the Anammox process. Characteristics of the Annacis Island WWTP centrate feed areprovided in Table 3.1. The Annacis Island WWTP uses a trickling filter for mainstreamsecondary treatment and a thermophilic AD process for sludge stabilization. A target centrateloading of 2,750 mg N/d (0.5 g N/L bioreactor volume) to the Anammox reactor was reached onJune 28, 2013. Seed Anammox sludge was acquired from World Water Works from theirDEMON pilot plant at Pierce County, Washington. The Anammox inoculant was supplied as adewatered cake at 16% total solids concentration.Table 3.1– Annacis Island WWTP centrate feed characteristicsParameterAverage ofMeasuredValues (StandardDeviation,Number ofSamples)pH 8.1 (0.21,11)Alkalinity, mg/L as CaCO3 5,160 (670, 11)Total Suspended Solids (TSS), mg/L 340 (51.8, 4)Ortho Phosphate, mg P/L 181 (7.1, 10)Ammonia, mg N/L 1,192 (64.0, 27)Biological Oxygen Demand, mg/L 368 (80.5, 16)Volatile Fatty Acids, mg/L 32.4 (20.9, 4)3.2.3 Sample preparation and testingDuring the data collection phase, nutrient testing for the AD and Anammox reactor (feed andeffluent) was typically undertaken one or two times per week. For consistency, a set of feed andwaste (effluent) sample was taken on the same day. Other parameters (i.e., alkalinity, TS/VS,TSS, COD, BOD, VFAs) were sampled monthly to monitor reactor performance. Formeasurement of the mainstream and sidestream dissolved nutrients at the full scale KWTF,sampling was coordinated with the City’s sampling program between March 11, 2014 and July10, 2014. Composite influent and final effluent samples collected by KWTF staff were used for65this research. In addition, KWTF’s dissolved phosphorus test results on the composite finaleffluent samples were utilized in the mass flow calculations. KWTF uses the same analyticalmethods as the UBC lab. The remaining sidestream grab samples were collected on the same daythat the composite samples were taken and processed at the UBC lab.For sample characterization, Standard Methods procedures 2540 B and 2540 E (APHA, 2005)were used for TS and VS determination. TSS was measured according to Standard Methodsprocedure 2540D. The closed reflux colorimetric chemical oxygen demand (COD)measurements were performed based on Standard Methods procedure 5250D (APHA, 2005)with a Thermo Genesys 10S spectrophotometer and 600 nm wavelength absorbance. Alkalinitywas measured according to Standard Methods 2320B (APHA, 2005). Biochemical OxygenDemand (BOD) was measured using an OxiTop® system, based on pressure measurements in aclosed system.Total volatile fatty acids, calculated as the sum of acetic, propionic and butyric acids, weremeasured by injecting supernatants (filtered through membrane discs with 0.2 µm pore size) intothe Agilent 7890A Gas Chromatograph (GC) with a capillary column (Agilent 19091F-112, HP-FFAP polyethylene glycol TPA column length x ID: 25 m × 320 μm) and a flame ionizationdetector (oven, inlet and outlet temperatures: 200, 220 and 300oC, respectively, carrier gas flowrate: 25 mL helium/min) equipped with an autosampler (Agilent 7693A). The method developedby Ackman (1972) used iso-butyric acid as an internal standard.A standard capacity (1.5 - 800 L/d) Wet Tip gas meter was used for measurement of biogasflow from the AD. Biogas volumes were then converted to standard temperature and pressure(STP = 0°C, 1 atm). Biogas composition in the AD headspace was determined in terms of CO2,66CH4,  N2 and  O2 percent using an Agilent 7820A GC with a packed column (Agilent G3591-8003/80002) and a thermal conductivity detector using helium as the carrier gas.Samples analyzed for total dissolved nitrogen and total dissolved phosphorus were first filteredusing a 0.45 micron syringe tip filter (Millex nylon, 13 mm or 25 mm) and then diluted to matchthe digestion test range (0.0 to 1.0 mg P/L and 0.0 to 2.0 mg N/L). For liquid streams with highTS, like digested sludge, the sample was first centrifuged at 5,000 ×g for 10 min to extract asupernatant.Digestion for total nitrogen was conducted using a persulfate method (APHA, 4500-P J). Whilethis method is intended for simultaneous total nitrogen and total phosphorus analysis, it provedinsufficient for digestion of dewatering centrate and AD effluent samples for total phosphorus.Consequently, an alternate digestion procedure for total phosphorus was adopted. The digestedtotal nitrogen sample was analyzed for nitrite/nitrate using an Astoria Method A2 autoanalyzerfitted with Multi-Test Cartridge A000-AMT and a cadmium reactor. Ammonia and nitrite/nitratein the undigested sample were measured using the same autoanalyzer.Digestion for total phosphorus analysis was based on the Persulfate Digestion Method (APHA,4500-P B.5). Samples and persulfate/acid solution were dispensed into glass test tube, cappedand autoclaved at 120°C for 2 hours. For complex analytes like dewatering centrate and ADeffluent, the 2 hour digestion time was necessary to achieve consistent results. The digestedsolution was cooled and analyzed immediately for orthophosphate using the Ascorbic AcidMethod (APHA, 4500-P E) and a Thermo Genesys 10S spectrophotometer with 1.0 cm lightpath. The original undigested sample was analyzed for orthophosphate in the same way.Acid hydrolyzable phosphorus (polyphosphates) was measured by digesting samples in a mildacid solution followed by measurement of phosphate (APHA, 4500-P B.2). The polyphosphate67(poly-P) concentration was calculated as the difference between the phosphate concentration ofthe acid hydrolyzed sample and original undigested sample. Organic phosphorus was measuredas the difference between the measured TN and poly-P concentration of the sample.Characterization of bioreactor influents and effluents were tested for nitrite, nitrite/nitrate,ammonia and orthophosphate using the Astoria Method A2 autoanalyzer. In addition, dailychecks of the Anammox SBR effluent nitrate were made using an Accumet nitrate probe(APHA, 4500-NO3¯ - D). Representative calibration curves for the nitrogen and phosphorustesting are provided in Appendix A.3.3 ResultsPresentation of the results is divided into three sub-sections. First, general operational data onthe lab-scale AD and Anammox reactors are presented. Following this, results of the lab-scalebioreactor (Anammox and AD) effluent and full-scale KWTF centrate dissolved nutrientcharacterization are provided. The final sub-section presents results of the NRDP and DON massflow measurements through the full-scale KWTF.3.3.1 AD and Anammox reactor operationThe lab-scale anaerobic digester was operated continuously for more than two years (Figure3.9). The digester reached stable operation in March, 2012 and showed good biogas productionuntil April, 2013 when the daily production crashed. The large variability in biogas productionthrough year 2013 is attributed to the changes in the sludge feed characteristics and a leak in thebioreactor headspace. During this time, the KWTF was commissioning a new upgrade to itstreatment plant. Start-up challenges with the new fermenter and operational changes resulted influctuations in the solids content and VFA concentration of the mixed feed sludge through muchof 2013. Also, rubber gaskets on the reactor assembly were replaced in September, 2013 to68address a suspected gas leak. Since the latter part of year 2013, biogas generation values showedimprovement. While the biogas volume generation measurements fluctuated considerably, theCOD and TS/VS removal rates remained relatively stable. Biogas composition through theperiod of testing, as measured in the AD headspace, averaged 67.1% methane (n = 28, standarddeviation = 1.02%). This indicates that organic removal data are more reliable than the biogasvolume measurements, a feature which is common for AD operation.Figure 3.9 – Operating data for the anaerobic digester69Feed and effluent sludge characteristics through the period between April 1, 2012 andApril 1, 2014 are provided in Figure 3.10 - SEM images of digester struvite precipitateTable 3.2. The digester showed a 54.0% reduction in COD, 52.9% reduction in VS and 44.2%in TS. In addition, on average, ortho phosphate decreased by 131 mg P/L or 22.2%. It ispresumed that the phosphate precipitated as struvite (MgNH4PO4·6H2O). A common feature ofthe digester operation was the observation of small crystals in the waste sludge. Scanningelectron microscope images of the crystals are provided as Figure 3.10. X-ray diffractionconfirms the presence of a high proportion of oxygen, phosphorus magnesium and nitrogen, allelements associated with struvite. In addition, the x-ray diffraction also detected sodium, ironand calcium as part of the crystal structure suggesting a variety of chemical precipitates.70Figure 3.10 - SEM images of digester struvite precipitateTable 3.2 – Mesophilic anaerobic digester average feed and effluent characteristicsParameterFeed Sludge Effluent SludgeAverageValue(StandardDeviation,Number ofData Points)AverageValueStandardDeviation,Number ofData Points)pH 5.47 (0.47, 52) 8.06 (0.20, 35)Alkalinity, mg/L as CaCO3 896 (305, 42) 5,240 (1,090, 35)Total Solids (TS), % by wt. 4.41 (0.55, 57) 2.46 (0.26, 56)Volatile Solids (VS), % by wt. 3.67 (0.47, 57) 1.73 (0.13, 56)Chemical Oxygen Demand(COD), mg/L57,000 (8,880, 61) 26,200 (3,300, 62)Ortho Phosphate, mg P/L 591 (130, 28) 460 (105,31)Ammonia, mg N/L 203 (57, 30) 1,210 (215, 32)Volatile Fatty Acids, mg/L 1,590 (804, 31) 29.9 (20.9, 21)Figure 3.11 provides summary plots of the Anammox operation through to the end of thecurrent testing. A series of failures necessitated turning off the bioreactor for most of the month71of August, 2013. Through the months of September and October, 2013, the effluent nitrate andammonia concentrations remained high. However, by January 1, 2014 the effluent ammonia andnitrate concentrations had stabilized allowing dissolved nutrient testing to commence. During theperiod between January 1 and April 1, 2014 when the first set of effluent characterization testswere conducted, the ammonia removal rate remained between 83% and 92%. Ammonia removalwas calculated based on a mass balance of ammonia-nitrogen loading compared to the mass ofeffluent nitrogen. Since the main objective of the research was to measure the change indissolved nutrients through the Anammox process, total nitrogen removal was not monitored.The Annacis Island centrate feed to the Anammox reactor was measured to be 1,566 mg N/L (n= 9, σ  = 74.0) compared to a dissolved total nitrogen of 1,320 mg/L and ammonia of 1,130 mgN/L.72Figure 3.11 – Ammonia removal efficiency and effluent characteristics for the Anammox bioreactor733.3.2 Effluent and centrate characterization resultsTable 3.3 provides a summary of the dissolved nutrient testing that was conducted on theanaerobic digester waste sludge, Anammox feed/effluent, and KWTF dewatering centrate. Thetesting  was  undertaken  to  compare  the  relative  concentrations  of  DON  and  NRDP  and  assesswhether introduction of an anaerobic digester or Anammox sidestream reactor would have anappreciable impact on the current treatment plant operation.A summary of the DON and NRDP measurements through the Anammox and AD is alsoprovided schematically as Figure 3.12 and Figure 3.13. The results are discussed in Section 3.4.Table 3.3 – Comparison of influent/effluent characteristics for the bioreactorsLiquid StreamDissolved Nitrogen Species Dissolved PhosphorusSpeciesDissolvedTotalN(mgN/L)Ammonia(mgN/L)Nitrite/Nitrate(mgN/L)DissolvedOrganicN(mgN/L)TotalDissolvedP(mgP/L)Ortho-Phosphate(mgP/L)NonReactiveDissolvedPhosphorus(mgP/L)Anaerobic DigesterEffluent Sludge1,400[8,106]1,190[8, 59]0.12[8, 0.06]210[8, 96]557[10, 169]525[10, 157]34.2[10, 23]Centrate Feed(Annacis Is. WWTP)1,320[12, 50]1,130[12, 73]0.13[12, 0.1]195[12, 75]191[10, 9.3]181[10, 7.1]10.0[10, 4.4]Anammox Effluent 170[19, 30]115[19, 21]36.0[19, 15]19.2[19, 7.9]183[17, 8.3]174[17, 7.5]8.79[17, 4.5]KWTF DewateringCentrate68.0[8, 18]55.8[8, 20]0.47[8, 0.26]12.0[8, 6.0]331[13, 115]317[13, 109]14.5[13, 10]Note: Values in brackets denote sample size and standard deviation, [#, stdev]74Figure 3.12 – DON and NRDP flux through the Anammox SBRFigure 3.13 – DON and NRDP flux through the mesophilic AD753.3.3 KWTF nutrient mass flow measurementsAs a way to understand the origin of the DON and NRDP species within the KWTF, testingwas conducted on the internal recycle and influent streams of the full scale WWTP to measurethe mass flow of these species (Figure 3.14 and Figure 3.15). To calculate the DON and NRDPloadings for the KWTF, testing was conducted on composite samples of influent wastewater,primary effluent and final effluent. In addition, grab samples of dewatering centrate, TWAS, FPSand WAS thickener underflow (DAF underflow) were also tested. Flow data for each of thestreams tested was provided by KWTF.76Figure 3.14 – Non Reactive Dissolved Phosphorus Mass Flow Diagram for KWTF77Figure 3.15 – Dissolved Organic Nitrogen Mass Flow Diagram for KWTF783.4 DiscussionThe resulting mass flow diagrams indicate that the majority of the DON and NRDP in theKWTF primary effluent originate in the raw wastewater. However, approximately 23% of theprimary effluent NRDP mass flow is contributed by the combined internal recycle stream (i.e.,dewatering centrate and DAF underflow). A smaller proportion of the DON in the primaryeffluent (approximately 5%) originates from the internal recycle stream.The UBC test results for total dissolved phosphorus on the KTWF effluent was inconsistentwith the total phosphorus values being measured by the City and therefore were neglected. Onthe same composite effluent sample, the UBC testing consistently measured higher totaldissolved phosphorus than the total phosphorus measured by the City of Kelowna’s commerciallaboratory (i.e., 0.61 mg P/L versus 0.17 mg P/L, average values). Based on these results, it wasconcluded that the filtration process probably contributed a positive background interference ofpolyphosphate and the results were neglected. The KWTF effluent was not filtered using asyringe filter. Since the effluent required no dilution for testing, the relatively large requiredsample volume meant that a conventional vacuum filtration apparatus was more practical than asyringe filter. Additional final effluent samples were collected and analyzed for dissolved totalphosphorus and orthophosphate by the KWTF’s lab which uses an autoanalyzer to measureorthophosphate and a commercial lab for TP. The calculated effluent NRDP values in Figure3.14 are based on these results.The nutrient characterization confirmed that introduction of the untreated AD waste sidestreamwould impose a substantial additional nutrient load on the KWTF. The pilot testing alsoindicated that elevated DON and NRDP concentrations occur in the UBC mesophilic AD sludgeeffluent. Both DON and NRDP in the digester effluent were significantly higher than the KWTF79centrate. The AD supernatant DON was 210 mg N/L compared to the KWTF centrate of 12.0 mgN/L, 17.5 fold increase. Similarly, NRDP in the AD supernatant was 34.2 mg P/L compared to14.5 mg P/L for the KWTF centrate, a 2.36 fold increase. Ammonia and orthophosphateconcentration also increased significantly through the digestion process. However, a variety oftreatment processes for removing centrate ammonia and phosphate are documented in theliterature. It is assumed that any increase to the effluent nitrogen or phosphorus concentrationdue to this loading increases the risk of exceedances.The influent/effluent characterization indicates that the Anammox process is able to degrade alarge fraction of the DON and to a lesser extent the NRDP. On average, the Anammox bioreactorwas able to reduce DON by 90% (i.e., from 195 mg N/L to 19.2 mg N/L). The NRDP wasreduced on average by 12% (i.e., from 10.0 mg P/L to 8.79 mg P/L). The reduction in both DONand NRDP are statistically significant, based on a paired t-test at a 95% confidence interval (α =0.05). A more detailed statistical assessment is provided in Chapter 4 for a similar set of testresults. The work described in Chapter 4 focused on measuring the effluent characteristics fromthe Anammox reactor fed centrate derived from the pilot AD fed KWTF sludge centrate. Thisresearch is more relevant to Okanagan Valley requirements.Another research consideration is to try to understand the transformation of the removedNRDP and DON compounds through the Anammox process, which is also studied in Chapter 4.In particular, while significant DON removal was measured through the Anammox reactor, itmay be shifted to the particulate form (i.e., greater than the 0.45 micron size cut-off) ormetabolized and transformed to nitrogen gas. Similarly, the proportion of dissolved organicphosphorus and inorganic condensed phosphorus that is measured as NRDP may change throughthe AD or Anammox reactor.80Based on results of the lab-scale reactors, the Anammox process shows great promise for DONremoval from the return stream of an anaerobic digester. However, in order to mitigate theimpacts of the return stream on a BNR process, additional treatment may be required to removeNRDP compounds. Various combinations of coagulation, filtration and absorption have beenshown to successfully remove dissolved phosphorus in wastewater effluent (Neethling et al.,2007; Arnaldos and Pagilla, 2010; Gu et al., 2011). Further research is required to assess thesuitability of this particular treatment approach for a sidestream flow that is composed ofdewatering centrate and WAS thickener underflow. Furthermore, since the return sidestreamNRDP at KWTF is greater than the mass contained in the final effluent, removal of most of theNRDP in the sidestream may also reduce the final effluent NRDP concentration. Whileadditional research is required to test this hypothesis (refer to Chapter 4), if it is feasible thereduction in final effluent NRDP could be achieved while treating only a fraction of themainstream flow.Although this research is focused on the removal of DON and NRDP, the characterization dataindicates the presence of high orthophosphate concentration in the AD effluent sludge. Thiswould translate into a high orthophosphate concentration in the dewatered sludge centrate of afull-scale AD. The orthophosphate concentration in the supernatant of the pilot AD is 525 mgP/L, which is significantly higher than in the Annacis Island WWTP centrate or KWTFdewatering centrate. Since the Anammox reactor had only a very small reduction in theorthophosphate concentration, an alternative removal process is required. The most attractivesolution to address the orthophosphate issue is to incorporate a struvite recovery stage upstreamor downstream of the Anammox reactor. In one specific lab-scale configuration where thestruvite reactor was located downstream of the Anammox reactor, an orthophosphate removal81rate of 86% was achieved (Hassan et al., 2013). Struvite is an equimolar precipitate ofammonium, phosphate and magnesium (ibid.). Consequently, a struvite recovery process wouldhave the added benefit of reducing the ammonia concentration in the centrate or Anammoxeffluent, depending on the process configuration.3.5 Summary of the bioreactor start-up phaseResults from a lab-scale anaerobic digester fed mixed waste sludge from a BNR treatmentplant confirms that the sidestream contains high levels of dissolved nutrients, including NRDPand DON. If untreated the sidestream return flow would likely have a negative impact on theeffluent quality of the wastewater treatment plant.A nitritation-Anammox sidestream process is able to remove a large proportion of theinorganic nitrogen (ammonia and nitrate/nitrite), as well as most of the DON in the AnnacisWWTP centrate. Inorganic nitrogen removal through the period of testing averaged 87% whileDON removal averaged 90%. However, the Anammox process was only able to reduce NRDP inthe centrate by 12% and did not appreciably reduce the centrate orthophosphate concentration.The research suggests that the Anammox sidestream reactor holds promise as a way of reducingDON. Additional removal of sidestream NRDP and orthophosphate would be required to allowan anaerobic digester to be incorporated into the KWTF without compromising effluent quality.Further testing, including measurements of the inorganic condensed phosphorus and particulatespecies, is required to understand the transformation pathways of DON and NRDP through thesidestream and anaerobic digestion processes. Additional research could also be conducted toassess the suitability of a coagulation/flocculation/absorption and struvite removal process forremoving dissolved phosphorus from the Anammox effluent. The following Chapter 4 wasintended to answer questions raised as a result of lab testing in Chapter 3.82Chapter 4: Biodegradation and chemical precipitation of dissolved nutrientsin anaerobically digested BNR sludgeThe objective of this chapter was to further explore processes for biodegradation andprecipitation of dissolved nutrients in dewatering centrate. In this study, characterization wasmade of a conventional suspended growth deammonification treatment process for transformingdissolved polyphosphate (poly-P), dissolved organic phosphorus (DOP) and dissolved organicnitrogen (DON) in two types of dewatering centrate. The experiments were designed  todetermine if the initial results generated with Annacis Island WWTP centrate documented inChapter 3 could be repeated with centrate derived from a nutrient removal treatment plant typicalof the Okanagan Valley. The characterization of dissolved phosphorus species through theAnammox reactor was also aimed at determining if the poor NRDP degradation was linked to achange in the relative proportion of DOP and poly-P. Furthermore, poly aluminum chloride(PACL) dosing was assessed to determine if coagulant dosing stage can achieve the objective ofprecipitating any residual DON and NRDP and producing an effluent that has lower dissolvednutrients than the pre-digestion, dewatering centrate at the Kelowna Wastewater TreatmentFacility (KWTF).834.1  ContextThe KWTF final effluent contains measured DON that represents 29% of the total nitrogen(TN) on average; similarly, the NRDP represents 52% of the total phosphorus (TP) on average(Figure  4.1).  Based  on  these  results,  a  significant  proportion  of  the  effluent  nitrogen  andphosphorus consists of DON and NRDP.Figure 4.1 - Kelowna WWTP final effluent nutrient speciationAs effluent nutrient standards become more stringent in the Okanagan Lake watershed,understanding the formation, fate and treatment characteristics of NRDP and DON will becomemore important. Furthermore, as the KWTF service population grows, more energy efficient andcompact sludge stabilization techniques are becoming desirable. However, the uncertainty of thefate of DON and NRDP through AD processes and the potential associated impacts on theKTWF effluent quality have served as an obstacle to adopting these technologies.84Suspended growth partial nitritation coupled with an anaerobic ammonia oxidizing(Anammox) step has shown to be an effective and energy efficient processes for removingammonia from AD dewatering centrate (Wett et al., 2007a). Previous research described inChapter 3 demonstrated the capacity of a similarly configured deammonification process forremoving a large proportion of DON but to a lesser extent NRDP from dewatering centrate.Therefore, the primary objective of this research was to assess the transformation ofpolyphosphates (poly-P) and DOP through the deammonification process. This aspect of theresearch was intended to understand the poor NDRP biodegradation capacity of the process. Thecurrent investigation focusses on determining whether the relative fraction of poly-P and DOPchanges through the deammonification process and if there is a gain in either constituent. Thebiodegradability of DON and NRDP centrate generated from a bench-scale AD fed BNR sludgewas also characterized.Finally, the success of implementing an AD within a BNR environment will ultimately bemeasured  by  the  quality  of  the  sidestream  treatment  effluent.  Under  ideal  conditions,  anAnammox process coupled with a struvite recovery process should remove more than 90% of thedissolved nitrogen and reactive phosphorus contained in the centrate (Hassan et al., 2013). Underupset conditions or where the NRDP needs further control, a final treatment barrier would bebeneficial. Therefore, as a secondary objective of this research, the impact of dosing theAnammox effluent with poly aluminum chloride (PACL) was assessed. PACL was selected forthe following reasons. PACL is supplied as a pre-polymerised liquid with the active aluminumspecies hydrolyzed with a base which reduces alkalinity demand (Jiang and Graham, 1998). Inaddition, previous research demonstrated that PACL is superior for TP removal compared toaluminum sulphate (Hatton and Simpson, 1985).854.2 Material and methods4.3 Anaerobic digester set-up/operationThe 5.0 L (effective volume) anaerobic digester (New Brunswick BioFlo® 115 glassfermenter) was started up on January 16, 2012 and had been operating continuously throughoutthe current testing. The digester was operated as a mesophilic process (37oC)  at  a  sludgeretention time (SRT) of 20 days and fed mixed sludge from KWTF. The mixed sludge feedconsisted of thickened waste activated sludge (TWAS) and fermented primary sludge (FPS) at apercentage ratio of 66:34 by volume, at a total solids (TS) content between 4.0% and 5.0% byweight. Organic loading rate of the digester ranged between 1.73 and 1.93 grams volatile solids(VS)/L of digester/d. Mesophilic sludge inoculum was taken from the City of Penticton WWTPwhich operated a digester fed FPS as substrate. More detailed set-up/operating data for the AD isdocumented in Chapter 3.Waste sludge from the AD was manually dewatered using BASF Zetag 7587 as a flocculant ata dosage of 50 mL, 0.5% solution per litre of sludge. Dewatering was undertaken using aThermo Sorvall XT centrifuge operated at 3,900 rpm, 3,300 ×g for 20 minutes in 750 mLcontainers. Prior to dewatering, the waste sludge was stored in a 20 litre plastic container at 4°C.When the container was full, the sludge was batch dewatered to provide approximately 20 litresof centrate feed for the Anammox reactor which was stored in 4 L containers. This approach ofbatch processing sludge was intended to minimize the variability in centrate feed characteristics.4.3.1 Sidestream partial nitritation-deammonification bioreactor set-up/operationThe partial nitritation-Anammox bioreactor was configured as a 5.5 L (effective volume)sequence batch reactor (SBR), operated at a temperature of 34oC. The New Brunswick BioFlo®115 glass fermenter, controller and associated computer software allowed for automated86operation of a centrate feed pump, waste pump, alkalinity feed pump and air valve. TheAnammox reactor configuration used for the experiment in this Chapter was identical to previouswork. A more detailed description of the Anammox reactor is provided in Chapter 3.At the start of operation, the Anammox reactor was operated using centrate from the AnnacisIsland WWTP (AIWWTP), owned by Metro Vancouver (Canada), as a substrate feed. TheAIWWTP uses a trickling filter for mainstream secondary treatment and a thermophilic ADprocess for sludge stabilization. Once sufficient characterization data was obtained, the substratefeed was switched from AIWWTP centrate to lab-scale KWTF AD centrate. During thetransition period, a mixture of AIWWTP and lab-scale KWTF AD centrate was used as feed.The proportion of AIWWTP centrate was progressively decreased over several days to minimizeany potential inhibitory effects. In particular, high phosphate concentrations have shown toinhibit the anaerobic ammonia oxidizing bacteria (van de Graaf et al., 1996).Characteristics of the full-scale AIWWTP and lab-scale KWTF AD centrate feed are providedin Table 4.1. For this study, a target centrate loading of 2,750 mg N/d (0.5 g N/L bioreactorvolume) to the Anammox reactor was used.87Table 4.1 – Full-scale AIWWTP and lab-scale Kelowna AD centrate feed characteristicsParameterFull-scaleAIWWTPCentrateLab-scaleKelowna ADCentrateMean(StandardDeviation,Number ofSamples)Mean(StandardDeviation,Number ofSamples)pH 8.21 (0.10, 40) 7.84 (0.20, 22)Alkalinity, mg/L as CaCO3 5,047 (147, 11) 3,780 (300, 18)Total Suspended Solids, mg/L 140 (51.8, 11) 175 (39.4, 21)Ortho-Phosphate, mg P/L 157 (9.6, 16) 459 (53.0, 21)Ammonia, mg N/L 1,184 (74.2, 14) 1,045 (147, 16)Biological Oxygen Demand, mg/L 388 (74.4, 12) 315 (127, 12)Volatile Fatty Acids, mg/L 32.4 (21, 4) 32.9 (8.1, 4)4.3.2 Sample preparation and analysisTesting protocols in this experiment were carried out as described in Chapter 3 with thefollowing modification and observations. Nitrogen & phosphorus testingInitially, an Astoria Method A2 autoanalyzer fitted with Multi-Test Cartridge A000-AMT anda cadmium reactor was used to measure nitrate and TN. The persulfate method used for TNmeasurement (APHA, 4500-P-J) converts nitrogenous compounds to nitrate. TN is measured asthe total nitrate concentration of the digested sample. Consequently, accurate TN testingnecessitates a reliable method for nitrate. The Astoria autoanalyzer uses the cadmium reactor toreduce nitrate to nitrite. Following reduction of nitrate, nitrite is measured colorimetrically. Inthis way, TN testing required careful management of the cadmium reactor which over time lostefficiency due to oxidation of the metal surface. However, to simplify the testing and avoiderrors associated with drifting of the cadmium reactor efficiency, use of an ion chromatograph(IC) was investigated. The digested TN sample was analyzed for nitrate using a Dionex IC,88Model ICS-2100. By this approach, nitrate could be measured directly and prone to less error. Acomparison of TN results obtained using the autoanalyzer and IC showed excellent correlationfor a total number (n) of 8 and 15 samples for nitrate and TN, respectively with a correlationcoefficient (R2) of 0.9978 (Figure 4.2).Figure 4.2 – Comparison of IC and autoanalyzer for measurement of TNThe use of an Accumet nitrate probe was also tested as a method for measuring nitrate insamples or digested TN samples. Measurements using the nitrate probe were found to not be areliable measure of the nitrate concentration when compared with the autoanalyzer or IC. Ineffect, the test results were inconsistent. Figure 4.3 provides a plot of nitrate measurements usingthe probe and IC. In general, the probe measured nitrate consistently higher than the IC with therelative error increasing with decreasing concentration.y = 0.9935xR² = 0.9978020040060080010001200140016000 200 400 600 800 1000 1200 1400 1600IonChromatographMeasurement(mgN/L)Autoanalyzer Measurement (mg N/L)Nitrate (n=8)Total Nitrogen (n=15)Linear Correlation89Figure 4.3 – Comparison of IC and probe for measurement of nitrateWhile results from the probe are a poor measure of the absolute nitrate concentration, itnevertheless provided good value for operation of the Anammox reactor. The probemeasurements reliably captured relative changes in the nitrate concentration in the day-to-dayoperation of the reactor (Figure 4.4). The time required to test Anammox effluent nitrate usingthe probe was a fraction of the time required to measure a sample using the IC. Effluent nitrateconcentration is an important process variable, providing early indication of unusual loadingconditions and signaling the need to adjust aeration or sludge wasting. Consequently, the probewas used for convenience to monitor relative changes in nitrate concentration from day-to-dayand to allow timely process changes. Duplicate samples were stored and measurements of nitratemade using the IC were used as a check and to calculate nitrogen balance. In addition, ammoniawas measured using APHA 4500-NH3 F and nitrite was measured using APHA-ܰOଶି B.y = 0.0049x2 + 0.052x + 9.0R² = 0.980204060801001201401601802000 20 40 60 80 100 120 140 160 180 200ICMeasurement(mgN/L)NO3 Probe Measurement (mg N/L)90Figure 4.4 – Comparison of nitrate measurements made using a probe and ICFurthermore, to assess the accuracy of the TN and TP digestion and analysis procedures used inthis research, 6 centrate feed and Anammox effluent samples were split and tested by both theUBC lab and a commercial lab. The samples were collected in two clusters within 5 days of eachother in late February, 2015 and March, 2015. The commercial lab measurement was based onthe summation of total Kjeldahl nitrogen (TKN) (APHA 4500-NORG D) and nitrite/nitrate; TPwas measured using persulphate method (APHA 4500-P). A comparison of the measureddissolved TN (dTN) and dissolved TP (dTP) values for 6 centrate feed and 6 Anammox effluentsamples is provided in Figure 4.5. The centrate feed for both clusters was part of the same batchof dewatered sludge and the nutrient content would be theoretically similar. The Anammoxeffluent characteristics would also have been similar for each cluster of samples but the effluentnitrogen had decreased between February and March.020406080100120140160180200NitrateConcentration(mgN/L) Ion ChromatographNitrate Probe91Figure 4.5 – Centrate feed and Anammox effluent dTN and dTPBased on a paired t-test comparison of the dTN measurements, the UBC lab results were notstatistically different to the commercial lab result at a confidence limit of 95% (α = 0.05). Rawdata and summary statistics are provided in Appendix B. The dTP measurements between theUBC lab and commercial lab were also not statistically different at a lower confidence limit of90% (α = 0.10). However, dTP results from the commercial lab exhibit more variability than theUBC lab results. Also, the commercial lab data underestimates the dTP compared to the UBClab. Given that NRDP is calculated based on the difference of dTP and orthophosphate, thehigher variability and tendency for lower values in the commercial lab measurements risksunder-estimating NRDP.11501200125013001350140014501500155016001650UBC Lab Commercial LabTotalNitrogen(mgN/L)a.) Centrate Feed, dTNt Stat = -0.51, T = 2.57, α = 0.05, n = 6130135140145150155160165170175180185UBC Lab Commercial LabTotalPhosphorus(mgP/L)c.) Centrate Feed, dTPt Stat = 2.28, T = 2.57, α = 0.05, n = 6130135140145150155160165170175180185UBC Lab Commercial LabTotalPhosphorus(mgP/L)d.) Anammox Effluent, dTPt Stat = 1.64, T = 2.57, α = 0.05, n = 6190200210220230240250260270280290UBC Lab Commercial LabTotalNitrogen(mgN/L)b.) Anammox Effluent, dTNt Stat = -0.38, T = 2.57, α = 0.05, n = 6Measuredda ta rangeMea n &standarddeviation bar924.3.3 Dose optimization for nutrient precipitationFor assessing the effectiveness of PACL on precipitating dissolved nutrients, 100 mL batchvolumes of Anammox effluent were used. A pre-mixed PACL product, Isopac (KlearwaterEquipment and Technologies Corp.), was used for the testing. Isopac is a non-sulphated PACLsolution, 18% Al2O3 by weight and 45% basicity. The PACL solution was dosed in the effluentand flash mixed in a glass beaker using a magnetic stir rod at 200 rpm for 30 seconds. Followingflash mixing, the stir speed was reduced to 25 rpm and the sample flocculated for 25 minutes. Asample was drawn off after settling for 5 minutes for dissolved nutrient testing and measurementof zeta potential. Zeta potential is a measure of the electrostatic charge repulsion and attractionbetween particles and can be used to monitor coagulant dosing efficiency (Nobbmann et al.,2010). The coagulant dose was incrementally increased and new batch volumes started until aneutral zeta potential was reached. A Malvern Zetasizer (nano ZS) was utilized for measurementof zeta potential. All zeta potential measurements were made at a temperature of 25°C.4.4 Results4.4.1 Existing KWTF centrate characterization resultsThe potential for increased effluent nutrient concentration associated with introducing an ADinto a tertiary WWTP can be minimized by designing a sidestream treatment process to achieve apre-digestion dewatering centrate quality. The KWTF currently dewaters its primary and wasteactivated sludge and transports the biosolids off-site for composting. Introducing an AD wouldimpact the WWTP process through changes in the dewatering centrate quality. The existing(baseline) dewatering centrate nutrient load was characterized to assess NRDP and DONconcentrations. This exercise was necessary to allow comparison of the Anammox effluentNRDP and DON concentrations with the baseline characterization. The comparison could be93used to assess whether centrate deammonification would be sufficient to reduce nutrientconcentrations to baseline levels if KWTF were to implement AD for biosolids stabilization andenergy recovery. A graphical representation of the baseline KWTF centrate nutrient fractionationdata (no sludge AD) is provided in Figure 4.6. Standard deviation and number of data collectedare represented by σ and n, respectively. Samples were collected from July to September, 2015.To further characterize NRDP, the samples were analyzed for poly-P and DOP. The dataindicates that poly-P and DOP make up approximately an equal proportion of the NRDP in theexisting KWTF centrate.The KWTF centrate also contains a relatively high fraction of ammonia (81%) and phosphate(99%). The ammonia is associated with the fermented primary sludge (FPS). As part of the BNRprocess, primary sludge is fermented to generate VFAs which are used to enhance phosphateremoval in the mainstream process. The fermentation process also releases some ammonia as aby-product. The residual VFAs in the FPS also stimulate release of phosphate in the TWASwhen the two sludges are mixed just prior to centrifugation.94Figure 4.6 – Nutrient fractionation for the existing KWTF dewatering centrateBased on the 6 samples of KWTF dewatering centrate that were fractionated to measuredissolved phosphorus species, the NRDP on average was 2.88 mg P/L (σ = 1.8 mg P/L). Thismeasured NRDP is less than the 14.5 mg P/L measured over the February-July, 2014 timeperiod. The decrease in NRDP may be explained by process improvements that wereimplemented at KWTF over this time. In particular, the mixing system in a sludge vault used toDissolved Organic PPolyphosphateOrtho-PDissolved Organic NNitrate/NitriteAmmonian = 6250255260265270275Phosphorus SpeciesPhosphorusConcentration(mgP/L)n= 1030. SpeciesNitrogenConcentration(mgN/L)1.47 mg/L(σ = 1.2)1.41 mg/L(σ = 1.0)267 mg/L(σ = 24)12.2 mg/L(σ = 5.5)52.7 mg/L(σ = 18)0.420 mg/L(σ = 0.29)95store thickened waste activated sludge (TWAS) when the centrifuge dewatering is off-line (i.e.,evenings and weekends) was upgraded to eliminate short-circuiting and suspected fermentingTWAS conditions. After this change, the dTP values decreased from an average of 331 mg P/L(σ = 115 mg P/L, n =13) in year 2014 to 267 mg P/L (σ = 23.8 mg P/L, n = 6) in year 2015. Thissuggests that dissolved phosphorus release associated with fermenting TWAS also contributes tohigher NRDP. The plot of existing KWTF centrate NRDP and dTP values for year 2014 and2015 (Figure 4.7) provides support to this idea. In the plot, measured NRDP concentration showsa positive, non-linear relationship with dTP. The data suggest that the proportion of NRDPincreases as centrate dTP increases above 200 mg P/L.Figure 4.7 – Existing KWTF dewatering centrate NRDP and relationship to dTPBased on information provided by the KWTF staff, the plant was challenged to meet the finaleffluent criteria during periods when the centrate dissolved phosphorus exceeded 300 mg P/L.During these times, a backup aluminum sulphate (alum) dosing system was utilized to ensure thefinal effluent TP objective was met. The use of alum is reserved as a method of last resort due toy = 8E-07x3 - 0.0004x2 + 0.0702xR² = 0.74270510152025303540450 50 100 150 200 250 300 350 400 450 500 550 600NRDP(mgP/L)Dissolved Total Phosphorus (mg P/L)96the high cost. Once the conditions associated with fermenting TWAS were resolved the KWTFwas able to resort to it normal operation without the use of alum trimming to achieve it effluentpermit.4.4.2 Anammox performance resultsExcept for the occasional equipment and power failure event, the lab-scale Anammox reactoroperated continuously from November, 2012 to June, 2015, a period of 32 months. The lastfailure occurred on August 29, 2014 when pH control was lost resulting in excess aeration.Following this upset, the process was slowly re-established using dewatering centrate fromAIWWTP, reaching the target loading of 2,750 mg NH3-N/d on November 24 (Figure 4.8). FromNovember 24, 2014 to April 2, 2015 AIWWTP centrate feed and Anammox effluent sampleswere collected and analyzed for dissolved nutrients. Since the principal objective of the researchwas to understand the flux of dissolved nutrients through the reactor, the testing did not includetesting to assess Anammox TN removalOnce sufficient data had been collected, the feed substrate was transitioned on April 29, 2015to lab-scale KWTF AD centrate.  Through this period of testing,  the average ammonia removalrate  was  79.4% (σ =  2.1%)  for  the  AIWWTP centrate  and  85.1% (σ = 2.8%) for the lab-scaleKWTF AD centrate. The mixed liquor nitrite concentration averaged 0.38 mg N/L (σ = 0.17 mgN/L) and ammonia averaged 112 mg N/L (σ = 21 mg N/L).During the transition phase, the proportion of lab-scale KWTF AD centrate was increased overa period of 28 days to minimize the potential inhibition associated with the increasedorthophosphate loading. In previous research, van de Graaf et al. (1996) observed inhibition byphosphate on anaerobic ammonia oxidizing bacteria at concentrations as low as 155 mg P/L (i.e.,KH2PO4 at 5 mM concentration). Since the KWTF utilizes a BNR process, lab-scale KWTF AD97centrate contains significantly higher concentrations of dissolved phosphorus compared to theAIWWTP centrate. Through the transition period, the orthophosphate loading increased from anaverage of 157 mg P/L (σ = 9.6 mg P/L) in the AIWWTP centrate to 459 mg P/L (σ = 53 mgP/L) in the lab-scale KWTF centrate. The Anammox process did experience a temporarydecrease in removal efficiency near the end of the transition period. To compensate for the loss,the loading was decreased to around 2,000 mg NH3-N/d until the process stabilized. Theammonia removal rate associated with the KWTF AD centrate was slightly lower at 79.4% (σ =2.1%) compared to the AIWWTP centrate and 85.1% (σ = 2.8%). However, the Anammoxreactor operated for 5 months on the AIWWTP centrate feed compared to 2.5 months for theKWTF AD centrate feed. It’s possible that in time the difference in removal rates would haveclosed. The instability in the Anammox process that was observed as the proportion of KWTFAD centrate was increase may suggest some short-term inhibitory event did occur but to whichthe process was ultimately able to adapt.During the second testing phase, lab-scale KWTF AD centrate feed and effluent samples werecollected from the Anammox reactor and analyzed for dissolved nutrients.98Figure 4.8 – Operating characteristics and results for the Anammox process60%65%70%75%80%85%90%95%100%AmmoniaRemovalEfficiency(%) Ammonia Nitrogen Removal020406080100120140160180200220240260280EffluentNitrogenConcentration(mgN/L)Effluent AmmoniaEffluent Nitrate1,0001,2501,5001,7502,0002,2502,5002,7503,0003,250AmmoniaLoading(mgN/d)Ammonia Loading0. Effluent NitriteKelowna Lab ADCentrate FeedTransitionAnnacis Island WWTPCentrate FeedAcclimatizationPeriod994.4.3 Centrate biodegradation through the Anammox processFractionation of dissolved nutrients in the Anammox centrate feed and effluent wascharacterized to assess biodegradation capacity of dissolved nutrient species. The underlyingassumption of this work is that in a full-scale BNR facility, bio-conversion of the dissolvednutrients into a particulate form will increase the likelihood that it can be removed by theprocess. The KWTF effluent characterization documented in Section 4.1 indicates that thetreatment process is less efficient at removing dissolved nutrients compared to particulatenutrients greater than 0.45 μm in size. The testing necessary to assess TN removal efficiency andparticle size characterization of the influent and effluent was not included in the current research.However, future research designed to characterize the change in colloidal and particulate nutrientsize distribution through the process could help further optimize removal.For this part of the research, AIWWTP centrate and the lab-scale KWTF AD centrate wereused as feed substrate for the Anammox process. Previous work had shown that the Anammoxprocess was able to degrade AIWWTP centrate DON but had a limited reducing effect on theNRDP (Chapter 3). The current testing was undertaken to determine whether this result could bereplicated using the lab-scale KWTF AD centrate feed and to assess whether partitioning ofpoly-P and DOP changes through the Anammox process. AIWWTP centrate feedResults of the dissolved phosphorus and nitrogen fractionation for AIWWTP centrate feed andassociated Anammox effluent are summarized in Figure 4.9. The results show a relatively highDON removal rate of 94% by the Anammox process which is consistent with previous workdescribed in Chapter 3. Also, Wadhawan et al. (2015) measured good DON removal (70%)through a nitrification process in secondary effluent and a lesser rate through an anammox100culture under anoxic conditions (27%). The higher removal rate measured in this study may be aconsequence of the Anammox process which cycles between aerobic and anoxic processes,thereby maximizing DON degradation through both nitrification and anammox exposure.Figure 4.9 – AIWWTP centrate and Anammox effluent dissolved nutrient characterizationOn average, the AIWWTP centrate DON was decreased from 204 mg N/L to 11.3 mg N/L.Overall, the dTN was reduced through the Anammox process by 79.9% (i.e., from 1,389 mg N/Lto 279 mg N/L).Dissolved Organic P Polyphosphate Ortho-PDissolved Organic N Nitrate/Nitrite AmmoniaPhosphorusConcentration(mgP/L)n = 10n = 10140145150155160165170AIWWTP Centrate Feed Anammox EffluentPhosphorusConcentration(mgP/L) a.) Dissolved phosphorus3.90 mg/L(σ = 3.2)4.28 mg/L(σ = 2.3)158 mg/L(σ = 10)2.38 mg/L(σ = 3.1)3.93 mg/L(σ = 2.0)153 mg/L(σ = 8.7)n= 14n = 1402004006008001000120014001600AIWWTP Centrate Feed Anammox EffluentNitrogenConcentration(mgN/L)b.) Dissolved nitrogen204 mg/L(σ = 61)0.558 mg/L(σ = 0.10)1184 mg/L(σ = 74)11.3 mg/L(σ = 10)116 mg/L(σ = 15)152 mg/L(σ = 39)101A small reduction in dissolved phosphorus species was measured across the Anammoxprocess. The average removal of NRDP was 22.9% (i.e., from 8.18 to 6.31 mg P/L). Based on apaired t-test, the NRDP removal results meet the standard for rejecting the null hypothesis thatthe means are statistically equal at a confidence limit of 95% (i.e., t-statistic of 2.82 is greaterthan the T-score of 2.26). In effect, the measured NRDP removal is statistically significant. Thepoly-P concentration in the centrate feed was not statistically different from Anammox effluentpoly-P at a confidence level of 95% (i.e., t-statistic, 0.74 is less than the T-score of 2.26). At the95% confidence limit, the change in DOP is less certain. The paired t-statistic of 2.17 for DOP inthe AIWWTP centrate feed and Anammox effluent is lower than the T-score of 2.26 whichindicates the reduction is not statistically significant. However, there is an apparent contradictionin that NRDP reduction is statistically significant but by the same standard reduction in bothpoly-P and DOP which make up NRDP is not. Consequently, the paired t-test is not definitivebut does suggest that additional testing is needed to confirm the reduction potential of DOPthrough the Anammox process. Raw data and summary statistics are provided in Appendix C. Lab-scale KWTF AD centrateIn the second trial, the degradation of lab-scale KWTF AD centrate through the Anammoxprocess shows similar characteristics as with the AIWWTP centrate (Figure 4.10). Reduction inDON concentration was 74%, from 85.7 mg N/L in the centrate to 22.4 mg N/L in the Anammoxeffluent. The precise fate of DON through the Anammox reactor is unclear. The colloidalfraction of centrate DON was not measured as part this research, nor was the TN flux through thereactor assessed. As a consequence, it’s not possible to assess whether most of the reduction inDON through the Anammox reactor was associated with colloidal or non-colloidal DON; or todetermine how much of the DON was truly removed and oxidized to ammonia, nitrate/nitrite or102nitrogen gas or converted to a suspended solids form (i.e., particle size greater than 0.45 μm).Additional research could focus on testing colloidal as well as dissolved and particulate nitrogenin order to better assess the impacts. However, in a full-scale application the large reduction inKWTF AD centrate DON, as well as ammonia, through the Anammox reactor does mitigate thepotential risk of dissolved nitrogen on the mainstream process and final effluent.Figure 4.10 – KWTF AD centrate and Anammox dissolved nutrient characterizationDissolved Organic P Polyphosphate Ortho-PDissolved Organic N Nitrate/Nitrite Ammonian = 13n = 13460465470475480485490KWTF AD Centrate Feed Anammox EffluentPhosphorusConcentration(mgP/L)a.) Dissolved phosphorusn = 16n = 16020040060080010001200KWTF AD Centrate Feed Anammox EffluentNitrogenConcentration(mgN/L)b.) Dissolved nitrogen85.7 mg/L(σ = 38)0.096 mg/L(σ = 0.22)1045 mg/L(σ = 147)22.36mg/L(σ = 9.1)117 mg/L(σ = 25)38.9 mg/L(σ = 17)3.94 mg/L(σ = 3.3)4.73 mg/L(σ = 3.9)477 mg/L(σ = 14)2.02 mg/L(σ = 2.5)5.01 mg/L(σ = 3.0)465 mg/L(σ = 15)103DON in the lab-scale AD had been measured at higher levels, prior to dewatering using poly-acrylamide (PAM) polymer, similar to that used at the AIWWTP. Of the 8 AD sludge samplestested without the use of PAM dewatering polymer, DON averaged 210 mg N/L (σ = 96 mgN/L). It’s not certain what caused the apparent decrease in DON in the digested sludgesupernatant. The influence of dewatering polymer dosing on AD DON is a possible factor thatcould be tested in future experiments.The Anammox process fed lab-scale KWTF AD centrate feed, showed a small reduction ineffluent NRDP from 8.68 to 7.03 mg P/L. Based on the paired t-test at a confidence limit of 95%,the  change  in  NRDP  was  statistically  insignificant  (i.e.,  t-statistic  of  1.70  versus  a  T-score  of2.18). Using the paired t-test to compare the results, the poly-P concentration in the lab-scaleKWTF  AD  centrate  feed  is  not  different  from  the  Anammox  effluent  at  a  confidence  level  of95% (i.e., t-statistic, -0.23 is greater than the T-score of -2.18). In this case, the change in DOP ismore certain compared to the AIWWTP centrate. The paired t-statistic of 3.21 for DOP in thelab-scale KWTF AD centrate feed and Anammox effluent is significantly higher than the T-scoreof 2.18. Consequently, the measured DOP reduction of 48.7% is statistically significant and maysuggest that the Anammox process is able to degrade more DOP than poly-P. Raw data andsummary statistics are provided in Appendix C.4.4.4 Summary of centrate biodegradation studyThe bioreactor studies using both AIWWTP centrate and lab-scale KWTF AD centrate showsimilar results. The Anammox process is able to degrade most of the centrate DON. Thepropensity for DON degradability is assumed to be due more to the aerobic ammonia oxidizingbacteria than the anaerobic variety as suggested by research conducted on secondary treatmentprocesses (Wadhawan et al., 2015). Combining both aerobic and anaerobic ammonia oxidizers in104a SBR process such as in the current study may help to maximize DON degradation. It’s possiblethat the ammonia oxidizers have the enzymatic processes to efficiently degrade the organicnitrogen. However, heterotrophs may also be contributing to the higher aerobic degradation ofDON.The testing also shows that DOP is more readily degraded through the Anammox reactor thanpoly-P. This feature may have application for NRDP removal if DOP degradation can beoptimized by process changes like increased HRT.The nitrogen and phosphorus removal rates were similar for both centrate feeds. This indicatesthe high phosphate concentration in the KTWF AD centrate did not significantly impact theAnammox process performance.4.5 Dose optimization for nutrient precipitationIn addition to characterizing the biodegradability of dissolved nutrient species through theAnammox process, the current research was intended to assess requirements for additionalsidestream treatment. When fed lab-scale KWTF AD centrate, the Anammox effluent containsdissolved reactive and non-reactive phosphorus levels that are significantly higher than theexisting KWTF centrate. In particular, the Anammox effluent NRDP was 7.03 mg P/L comparedto 2.88 mg P/L in the existing KWTF undigested sludge centrate. Consequently, if the results ofthis lab-scale trial are representative of a full-scale scenario, introduction of AD into the KWTFwill need to include additional nutrient removal. In addition to removing NRDP, the centrate hasan elevated concentration of phosphate and ammonia. The sidestream treatment paradigm thathas guided this research includes struvite recovery and chemical dosing following the Anammoxstage (Figure 4.11).105Figure 4.11 - Modified KWTF schematic incorporating anaerobic digestionand side-stream treatment for centrate dissolved nutrient removalStruvite recovery is a conventional approach for phosphate removal and provides anopportunity to recover phosphorus and ammonia in conjunction with biological centratetreatment (Hassan et al., 2013). The Anammox and struvite processes together can treat forcentrate ammonia, phosphate and as this research shows a significant fraction of the DON. WhileFigure 4.11 shows the Anammox as the first stage as the work by Hassan et al. (2013) suggestedwas better suited for operation of the struvite process, an alternative staged treatment processcould prove more optimal.Dosing the effluent with an aluminum cation like PACL following the Anammox and struviterecovery could serve as a physico-chemical process to polish the treated centrate, minimizingpotential impacts of any residual NRDP on the mainstream process and final effluent quality.The effectiveness of coagulant dosing with PACL for precipitating dissolved nutrients in theAnammox effluent was assessed in a series of batch tests. The tests were conducted in theabsence of struvite precipitation and therefore illustrate a worse-case scenario for PACL dosing.106Results of the PACL dosing experiment using effluent from the Anammox bioreactor fed lab-scale KWTF AD centrate are provided as Figure 4.12.Figure 4.12 – Impact of PACL coagulant dosing on the precipitation of dissolved nutrientsThe testing indicates that a PACL dose of 400 mg Al/L effluent would reduce the NRDPconcentration in the Anammox effluent to approximately 1.0 mg P/L, well below the average2.88 mg P/L currently in the KWTF dewatering centrate. Furthermore, the PACL dosing dataalso shows a propensity for significant reductions in DON. The 400 mg Al/L effluent PACL dosealso reduced the DON concentration in the Anammox effluent to approximately 11.0 mg N/Lwhich is lower than the existing KWTF dewatering centrate. In effect, the combination ofAnammox sidestream treatment and PACL dosing together is sufficient to reduce dissolvedmv/mg PACL/L Slope = -0.029R² = 0.831R² = 0.938-24.0-22.0-20.0-18.0-16.0-14.0-12.0-10.0-8.0-6.0-4.0- 50 100 150 200 250 300 350 400 450 500Zetapotential(mV)NRDP(mgP/L)18% PACL Solution Dose (mg Al/L effluent)Non-Reactive Dissolved PZeta PotentialR² = 0.9130. Organic NitrogenR² = 0.982050100150200250300350400450500Phosphate(mgP/L) Dissolved Organic Nitrogen107phosphorus and nitrogen in dewatering centrate associated with anaerobically digested BNRsludge. If orthophosphate removal by means of struvite recovery were introduced to thistreatment train, the PACL dosing requirements would be significantly reduced.The observation of DON and NRDP precipitation through chemical dosing is supported byprevious research by Dwyer et al. (2009) and Arnaldos and Pagilla (2010). In these studies,hydrated aluminum sulphate was used to precipitate DON and NRDP species from wastewatereffluent at much lower initial concentrations. The current research concludes that the approach ofaluminum dosing the dewatering centrate at higher initial concentrations is an equally effectivecontrol for DON and NRDP. Although the dosing concentrations for centrate are significantlyhigher, there are efficiencies gained from having to dose a much smaller centrate flow comparedto the mainstream final effluent flowrate. For KWTF, dewatering centrate represents on average0.51% of the total mainstream effluent volume. Consequently, under a sludge digestion scenario,the 400 mg Al/L treated centrate dose represents an equivalent dose rate of 2.0 mg Al/L based onthe final effluent volume (i.e., 0.51% * 400 mg Al/L Anammox effluent).After dosing the treated centrate with PACL to remove DON and NRDP in a full-scale WWTP,it may be possible to avoid additional sidestream or mainstream filtration to intercept thecoagulated nutrients. Instead, the precipitated DON and NRDP could be returned to the primaryclarifier and removed through settlement. In this way, the nitrogen and phosphorus precipitateswould ultimately be returned to the digester through the normal wasting process. This approachincreases the potential for re-release of nutrients in the AD process. Removal of dissolvedreactive phosphorus through struvite recovery would reduce the mass of precipitated nutrientsreturned to the digester and associated risk of nutrient re-solubilization. There are limitedpublished studies which document the hydrolysis potential of aluminum precipitated nutrients in108AD processes. Work by Grigoropoulos et al. (1971) suggests that precipitated phosphorus isstable under mesophilic conditions but additional research is needed to assess the fate ofcoagulated DON and NRDP through the AD process.4.6 Summary of the centrate treatment characterizationBased on a detailed study of centrate treatment through the Anammox, the following summaryconclusions were drawn.· Selection of methods for dissolved nutrient testing was important for achieving consistentresults, particularly for TP and nitrate.· The elevated phosphate concentration in the lab-scale KWTF AD centrate did notappreciably affect the stability of the Anammox process, as measured by ammonia removal.· The Anammox process showed similar treatment characteristics for both the KWTF andAIWWTP centrates with excellent DON removal but poor NRDP removal.· A statistical comparison of the DOP and poly-P through the Anammox process suggests thatDOP has a higher biodegradation potential. Future research focused on understanding thevariables associated with degradation of DOP could lead to better NRDP removal throughthe Anammox process.· Both DON and NRDP may be a heterogeneous mixture of dissolved and colloidalcompounds and the form may influence biodegradability. Future research should focus oncharacterizing the relative proportion to TN and TP of dissolved, colloidal and suspendedsolids DON and NRDP.· Utilization of a post-Anammox PACL chemical dosing stage can achieve the objective ofprecipitating any residual DON and NRDP to produce an effluent that has lower dissolvednutrients than the pre-digestion KWTF dewatering centrate.109Chapter 5: Final effluent particulate phosphorus fractionation andcoagulation by optimized chemical DosingEutrophication of waterways has prompted many agencies in Canada and the United States(US) to mandate phosphorus removal processes for all the waterways identified as sensitive tophosphorus (P) enrichment. Over the past 30 years these mandates have resulted in manywastewater treatment plant (WWTP) process configurations to convert to chemical, biologicaland/or physical barriers to enhance P removal. Implementation and optimization of biologicalprocess configurations has lowered total P discharges to less than 0.5 mg/L.Biological nutrient removal (BNR) process configurations when combined with effluentfiltration, as depicted in Figure 1.3, have resulted in improved suspended solids removal andreduced total P (TP) to less than 0.25 mg/L. There is a propensity in enhanced nutrient removaltreatment plants for a large fraction of the total P in the final treated effluent to be represented byfine particulate and soluble non-reactive phosphorus (Neethling and Stensel, 2013).Consequently, increasing the level of P removal will need to consider ways to remove theseforms of effluent phosphorus.BNR treatment plants typically include a backup coagulant dosing system to chemicallyprecipitate ortho-phosphate in the event of a process upset or effluent polishing. For thisapplication, aluminum sulphate is usually utilized as the coagulant of choice due to its lower costcompared to aluminum chloride and polyhydroxy aluminum chloride (PACL). However, PACLhas other characteristics which may make it a more suitable coagulant for dosing at a wastewatertreatment plant. PACL is supplied as a pre-polymerised liquid with the active aluminum specieshydrolyzed with a base. This feature reduces alkalinity demand compared to aluminum sulphate110or aluminum chloride. The nature of the polymeric species formed depends on a variety offactors but include those with high cationic charge such as ܣ݈ଶ(ܱܪ)ଶସା, ܣ݈ଷ(ܱܪ)ସହା andܣ݈ଵଷ ସܱ(ܱܪ)ଶସ଻ା (Jiang and Graham, 1998). In combined phosphorus and nitrogen treatmentprocesses, alkalinity consumption and reduced pH due to aluminum sulphate dosing coulddestabilize nitrification (Banu et al., 2009).A further consideration for use of PACL over aluminum sulphate for removal of particulatephosphorus is provided by research conducted by Hatton and Simpson (1985). Their jar testexperimentation work comparing phosphorus removal of various aluminum coagulants showedthat PACL was superior in removal of TP and equally effective in precipitating orthophosphate.If sufficient charge is available on the PACL floc to adsorb dissolved colloidal P, as suggestedby their research, then it may be possible for the PACL floc to become attached to the biologicalfloc and settle-out in the secondary clarifiers.Within a BNR process environment, the use of PACL may be a good candidate forprecipitating the soluble, non-reactive phosphorus that currently is an obstacle for achievinglower effluent criteria. If this is the case then a further improvement in the coagulation processcould be the use of a coagulant aid to reduce the PACL dose. This approach is used in theOkanagan Valley at various water treatment plants. The surface raw water of the valley ischaracterized by a high level of colour. The common treatment method includes using PACL inconjunction with an aqueous solution of poly epichlorohydrin amine (polyepiamine) as acoagulant aid. The polyepiamine has low molecular weight and high cationic charge densitywhich is effective at neutralizing the surface charge of colloids. The proposed research wasundertaken in part to determine whether use of a polyepiamine could improve the efficiency ofTP precipitation.111Measurements of zeta potential could also be used to monitor and optimize the coagulant dose.Zeta potential is a measure of the charge potential at the plane of shear between a particle andsurrounding liquid and depending on characteristics of the suspension the magnitude of thecharge can be used as a measure of stability (Bratby, 2006). The cationic coagulants proposed inthis research destabilize the phosphorus particle charges, thereby allowing them to agglomerateand settle. By monitoring zeta potential, it may be possible to control coagulant dosing bytargeting a specific zeta potential value associated with the desired effluent total P objective. Asimilar approach has been successfully used to minimize coagulant usage in water treatmentplants (Nobbmann et al., 2010). In a wastewater treatment context, optimizing the coagulant dosecould also minimize any potential negative effects of elevated concentrations of coagulant onnutrient removal. For example, other research has highlighted a potential for reduceddenitrification capacity by PACL at 19.5 mg Al2OH3/L (Guo et al., 2015).5.1 Research objectives & approachThe research detailed in this paper focuses on measuring the efficacy of PACL (Isopac, 18%Al2O3, 45% basicity) in combination with polyepiamine (Magnafloc LT 7990, 50% solidsmedium molecular weight polyepiamine product) for coagulating particulate and dissolvedphosphorus. The main objective will be to establish an optimum coagulant dose for removingsoluble, non-reactive phosphorus in final effluent from the Westside Regional WWTP. Asecondary objective is to characterize the colloidal material that makes up dissolved, non-reactive phosphorus (NRDP). Westside Regional WWTP effluent rather than KWTF effluentwas used in this part of the experimentation work to satisfy the research partner’s (AECOM)interest in adding to their operational understanding of the Westside Regional WWTP. However,112the Westside Regional WWTP and KWTF are based on a similar design and the results would beapplicable to both facilities.5.2 MethodologyThis section provides a description of the proposed testing methods and apparatus. In general,the testing was intended to generate data of sufficient quality to characterize effluent particulatesize distribution, both dissolved (< 0.45 µm) and suspended (> 0.45 µm). As part of this work,the effluent was fractionated to determine the distribution of total P by particulate size. Theintention was to develop a baseline picture of the particulate matter in final effluent fromWestside Regional WWTP and to determine if  there is  a propensity for TP to concentrate in aspecific size range. Conventional filtration and ultrafiltration techniques at the lab scale wereused to fractionate effluent samples for testing.In addition, various blends of PACL and polyepiamine were tested to determine theirefficiency for agglomerating colloidal particles in the dissolved range (< 0.45 µm) and shiftingthe size distribution toward the suspended solids size range.5.2.1 Phosphorus testingDigestion for total phosphorus analysis was based on the Persulfate Digestion Method (APHA,4500-P B.5). Samples and persulfate/acid solution were dispensed into glass test tube, cappedand autoclaved at 120°C for 2 hours. The digested solution was cooled and analyzedimmediately for orthophosphate using the Ascorbic Acid Method (APHA, 4500-P E) and aThermo Genesys 10S spectrophotometer with 1.0 cm light path. The raw undigested sample wasanalyzed for orthophosphate in the same way.Except where noted, 24 hour composite samples were used for testing. Composite sampleswere collected in the morning from the Westside Regional WWTP and tested the same day.1135.2.2 Particle size and zeta potential characterizationA Malvern Zetasizer, (nano ZS) was utilized for measurement of zeta potential and particlesize. The Zetasizer is able to characterize the size distribution for a sample between 0.01 μm and10 μm. All Zetasizer measurements were made at a temperature of 25°C.5.2.3 Coagulant DosingCoagulant dosing was conducted in two ways. For assessing the effectiveness of various PACLand polyepiamine blends, 100 mL batch volumes of final effluent were used. The coagulantblend was dosed and flash mixed in a glass beaker using a magnetic stir rod at 200 rpm for 30seconds. Following flash mixing, the stir speed was reduced to 25 rpm and the sampleflocculated for 25 minutes. After settling for 5 minutes, a 5 mL sample of supernatant was drawnoff using a syringe and used to measure of zeta potential. The settled, coagulated particles wouldnot contribute to the zeta potential measurement. The coagulant dose was incrementallyincreased and new batch volumes started until a neutral zeta potential was reached. An exampleof the zeta potential response curve for a 2:1 coagulant blend of PACL to polyepiamine isprovided as Figure 5.1.114Figure 5.1 – Example of zeta potential response curve using a 2:1 ratio of PACL topolyepiamineFor fractionation of effluent, dosing was conducted using a jar testing apparatus (Phipps &Bird) and 1 litre glass beakers. The same time interval for flash mixing and flocculation wereused with the jar testing apparatus as with the batch experiments.5.2.4 Conventional filtration apparatusA conventional vacuum filtration system was used to fractionate suspended particulate usingthree filter cut-off sizes, 0.10, 0.22 and 0.45 μm (Millipore, Durapore). The pore sizes for theconventional filtration were selected based on the manufacturers’ availability of membranes withhydrophilic and low particle binding properties. To minimize error associated with hold-up ofparticles on the membranes, filtration was conducted in parallel (Barker et al., 1999). For eachcut-off, a starting volume of 100 mL was used to obtain sufficient filtered sample volume forphosphorus testing. The phosphorus concentration for each filter cut-off was obtained throughsubtraction of concentrations between cut-offs (for example, 0.45 μm  >  TP  >  0.22 μm wascalculated as the difference between the 0.45 μm filtrate TP and 0.22 μm filtrate TP).1155.2.5 Ultra-filtration (UF) fractionationEffluent fractionation was conducted using UF membranes housed in a 400 mL Amicon model8400 stirred cell [Amicon Corp., MA, USA] using flat (7.6 cm diameter), high recovery, loworganic adsorption hydrophilic membranes [Millipore, MA, USA]. Membranes with molecularweight cut-off of 10,000 Da (YM10), 100,000 Da (YM100) and 300,000 Da (PES300) were setup in parallel to minimize hold-up of particles and cross-contamination of samples. Cut-offconcentrations were calculated using the subtraction method described in Section 5.2.4. UFmembranes were first rinsed with distilled water in a beaker for at least 1 hour with the waterchanged three times. Rinsed membranes were placed into the Amicon model 8400 cell followedby a 200 mL sample volume and the head space pressurized with nitrogen gas. Nitrogen gaspressures were maintained at a maximum 10 psi for PES300 and YM100 membranes and 50 psifor YM10 membranes as recommended by the manufacturer.5.3 Results & discussionResults of the effluent particulate characterization and coagulant dosing experiments arepresented in this section. Baseline zeta potential and particle measurements were made on 9composite effluent samples (Figure 5.2). The average zeta potential measurement was -10.9 mV(σ = 2.4 mV) and average particle size was 195 nm (σ = 11.3 nm). Both average zeta potentialand average particle size values are provided as output by the Zetasizer instrument. Thevariability was calculated from the results obtained from the 9 samples. The size distributionplots suggest a trimodal distribution with peaks occurring between 30 to 100 nm, 200 to 800 nmand 4,000 to 6,000 nm (Figure 5.3). Based on a cumulative average of these data, 75% of theparticulate fraction occurs in the dissolved range (i.e., less than 0.45 μm).11624March/1501April/1507April/1514April/1528April/1505May/1512May/1526May/1523June/1516017018019020021022024March/1501April/15 07April/15 14April/15 28April/15 05May/15 12May/15 26May/15 23June/15AverageParticleSizeDiameter(microns)Average Measured Particle Size-16.0-14.0-12.0-10.0-8.0-6.0-4.0-2.00.0 24March/1501April/1507April/1514April/1528April/1505May/1512May/1526May/1523June/15MeanZetaPotential(mV)Average Measured Zeta PotentialFigure 5.2 – Average zeta potential and particle size of Westside Regional WWTP finaleffluent samples1170%20%40%60%80%100% 100 1,000 10,000CumulativeVolume(%)Volume(%)Particle Size (nanometre)Measured DistributionAverageAverage CumulativeFigure 5.3 – Particle size distribution of Westside Regional WWTP final effluent samples118Fractionation  of  final  effluent  samples  was  conducted  to  determine  the  propensity  for  P  tooccur around the various peaks observed in Figure 5.3. Filter cut-offs of 0.10 μm, 0.22 μm and0.45 μm were used to assess phosphorus distribution (Figure 5.4). Of the average measured TP,43% consists of particles having a diameter greater than 0.45 μm. Based on the Zetasizermeasurements (Figure 5.3), on average only 25% of the total particulate volume is greater than0.45 μm. This suggests that in the fractionated samples, a large part of the TP is concentrated inthe suspended fraction (> 0.45 μm diameter).Figure 5.4 – Fractionation of composite final effluent sample from Westside RegionalWWTPAn unexpected outcome of the filter fractionation was the observation that a relatively smallamount of measured TP occurs in the size range between 0.45 μm and 0.10 μm. The Zetasizermeasurements (Figure 5.3) suggested a relatively large proportion of the colloidal particle sizerange volume occurred in this range (ie, approximately 30%). However, according to the testing,119most of the dissolved P is associated with colloidal paricle sizes smaller than 0.10 μm. To furthercharacterize dissolved P, final effluent was fractionated using ultrafiltration membranes andtested for TP. Molecular weight cut-offs of 300 kDa, 100 kDa, 10 kDa and 1 kDa were used forthis testing. The results indicate a relatively high proportion (22%) of the TP measured in thecomposite effluent sample occurred within the 1 kDa to 10 kDa molecular weight range (Figure5.5). Due to the complexity of the testing, the results are based on one effluent sample conductedin triplicated. Additional testing was not conducted to determine additional chemical propertiesof the dissolved material. However, the research indicates that dissolved organic matter (DOM)in wastewater effluents includes a variety of recalcitrant organic compounds which can vary withthe raw water characteristics, land-use composition within the sewerage area and treatmentprocess (Michael-Kordatou et al., 2015). In other research characterizing wastewater effluentDOM, 64% of dissolved organic matter occurs at a MW less than 1 kDA with 10.9% occurringin the 1 kDa to 10 kDa range (Kim et al, 2009).  Humic-like colloidal material was also shown tobe a dominant in three wastewater effluents and occurred at a MW size range of 1.6 to 2.6 kDa(Worms et al., 2010). Other research characterizing aquatic organic matter suggests that themolecular weight range between 1 kDa and 10 kDa is associated with fulvic and humic acids(Thurman et al., 1982).The ability of DOM and humic substances to complex a variety of dissolved compounds,including metals (Brown et al., 2000), micropollutants (Zheng et al., 2008) and phosphorus(Lévesque, 1969) has also been described. Consequently, in addition to removal of phosphorus,coagulation of humic substances using PACL and polyepiamine could result in removal of manyother potentially harmful compounds.120Figure 5.5 – P fractionation of composite final effluent from Westside Regional WWTPusing ultrafiltrationVarious mixtures of PACL and polyepiamine were added to 100 mL batch volumes of theWestside Regional WWTP final effluent to test their performance. After dosing a specificconcentration of coagulant blend and allowing for flocculation and settlement time, the zetapotential was measured on the supernatant. The dose was increased as required and the processrepeated until a neutral zeta potential was reached. Based on this testing, the concentration ofPACL and polyepiamine required to achieve neutral zeta potential could be recorded for avariety of mixtures. A plot of this data is provided as Figure 5.6. The dosing performance plot onits own provides no clear optimum dose preference. However, as a secondary performancecriteria cost was also assessed. Based on chemical supply cost information provided by thesupplier (Klearwater Equipment and Technologies Corp.), a 2 to 1, PACL to polyepiamine ratio121by active chemical mass was calculated to be the least cost scenario to achieve a neutral zetapotential (Figure 5.6)Figure 5.6 – Concentrations of coagulant blends that result in neutral zeta potential andassociated costsUsing the results of the batch dosing tests, a blend of 2 parts PACL to 1 part polyepiamine wastested to assess its ability to coagulant dissolved and suspended particulate P. A jar testingapparatus was used to dose 1 litre beakers of Westside Regional WWTP final effluent. Theflocculated effluent was fractionated using conventional filtration equipment using filter cut-offsof 5 μm, 0.45 μm and 0.10 μm. The filtrate was tested for P and zeta potential and from thesevalues the fraction of P within each size range was calculated. The fractionated P results areprovided as Figure 5.7. The results show the proportion of TP associated with ortho-P andcolloidal particles less than 0.10 μm decrease with the first coagulant dose. This indicates thatthe dissolved P which may be associated with colloidal material (ie, DOM) could be removedprovided is can be filtered or settled.122Figure 5.7 – P fractionation of coagulated effluent using a blend of 2 parts PACL to 1 part polyepiamine123The results also indicate that after Dose 3, the effect on total P precipitation diminishes eventhough the zeta potential is well below a neutral charge. Therefore, optimization of the coagulantdosing for the purposes of total P precipitation could target a specific zeta potential. In the caseof Westside Regional WWTP, the target zeta potential would be approximately -8.0 mV.Additional testing would be required to optimize the coagulant dose to target this zeta potentialsince the current testing was optimized for zeta potential neutrality.5.4 Summary of the effluent particulate characterizationThe research supports the idea that a significant proportion of the final effluent non-reactivedissolved phosphorus consists of colloidal particles, less than 0.10 μm in size. Ultrafiltrationcharacterization of the final effluent further suggests that a relatively large fraction of dissolvedTP may be bound up with DOM. However, this is speculative and additional TP fractionationcombined with DOM characterization is needed to test this theory.Batch dose testing of PACL and polyepiamine, two common coagulants used for potable watertreatment, were effective at precipitating the dissolved phosphorus. The testing showed thatcoagulation of dissolved phosphorus was correlated to zeta potential. These results indicate thatcoagulant dosing for final effluent phosphorus trimming at BNR facilities could be optimizedthrough zeta potential measurements.Given the effectiveness of PACL and polyepiamine for coagulating TP and the potential use ofzeta potential for optimizing dose, it may be possible to use this combination at BNR facilitiesfor dissolved phosphorus control. Further research at the full-scale would help to clarify theimpacts of dosing a blend of PACL and polyepiamine on the BNR process. Specifically, it wouldbe useful to know how the coagulated PACL/polyepiamine floc interacts with the biological flocand how sludge wasting would need to change in order to optimize treatment.124Chapter 6: Conclusions6.1 Summary and conclusionsThe research described in this thesis fills a gap in the literature by providing a detailedcharacterization of DON and NRDP production from a mesophilic anaerobic digestion process.Furthermore, the current research provides a measure of an Anammox process for thebiodegradation potential of DON and NRDP species. Based on the experimental work describedin the foregoing chapters, the following conclusions can be drawn:1. NRDP and DON are significant sources of nitrogen and phosphorus in the existing KWTFfinal effluent and additional input through production associated with an anaerobic digesterwill risk effluent quality. In particular, the KWTF is more sensitive to additional NRDPloading to the lower effluent objective for total phosphorus. In addition, a higher proportionof the existing total effluent was measured as NRDP (ie, 52%), thereby providing less of abuffer capacity.2. The Anammox process showed similar treatment characteristics for both the KWTF andAIWWTP centrates. The side-stream treatment process showed excellent DON removaland poor NRDP removal of the dewatering centrate that would represented by a full-scaleanaerobic sludge digestion scenario. The Anammox process biodegraded 74% of theKWTF AD centrate DON and 94% of the AIWWTP centrate DON. However, the NRDPremoval rate was 19% for the KWTF AD centrate and 23% for the AIWWTP centrate.These results suggest that the Anammox would work well as a sidestream treatment processto control all nitrogen concentrations in the return stream but an additional treatmentstrategy is required for the NRDP.1253. The elevated orthophosphate concentration in the lab-scale KWTF AD centrate did notappreciably affect the stability of the Anammox process, as measured by the ammoniaremoval efficiency.4. A statistical comparison of the DOP and poly-P through the Anammox process suggeststhat DOP has a higher biodegradation potential. Future research focused on understandingthe variables associated with degradation of DOP could lead to better NRDP removalthrough the Anammox process.5. Utilization of a post-Anammox PACL chemical dosing stage can achieve the objective ofprecipitating any residual DON and NRDP and producing an effluent that has lowerdissolved nutrients than the pre-digestion KWTF dewatering centrate scenario.6. Based on a detailed characterization of final effluent particulate phosphorus, most of theNRDP appear to occur in a size range that is associated with DOM and humic substances.This result further lends itself to the idea that the same conventional approach for colourremoval in water treatment systems could work well for controlling NRDP. Coagulantdosing of the final effluent with polyepiamine and PACL, combined with phosphorustesting shows a strong propensity for shifting the colloidal phosphorus associated with theNRDP to the non-dissolved size range (i.e., > 0.45 µm). Furthermore, the testing suggestszeta potential could be used as a monitor to optimize the coagulant dosing.Based on the results of the experimental work, it is possible to integrate an anaerobic digesterinto a BNR wastewater treatment plant provided sufficient care is taken to controllingrecalcitrant dissolved nutrients from the supernatant back to the mainstream process. Use of aconventional Anammox sidestream treatment process coupled with PACL dosing provides therequired level of nutrient reduction, as compared to the existing centrate.1266.2 LimitationsAdditional work is suggested by the research work but was not pursued due to limitations in timeand resources.· The current research did not incorporate a method for phosphate removal from the ADcentrate but it is assumed that struvite recovery will provide the required level of removal.While PACL can effectively remove all dissolved phosphorus, additional research is requiredto confirm the optimal staging of the Anammox and struvite recovery processes for thedigested BNR sludge centrate.· Due to lab space limitations the experimental setup used in this research utilized an AD andAnammox process that were mismatched in terms of capacities. The Anammox centrate feedrate was higher than the centrate production capacity of the AD reactor. This feature requiredthat AD sludge be stored in order to generate a sufficient volume of centrate for continuousAnammox operation. Future research that provided for a larger AD reactor would produce amore representative centrate and remove the variable associated with the lag-time.· The experimental results indicate that DOP does degrade through the Anammox process andmay be influenced by hydraulic retention time. Future work could test this theory bysystematically decreasing the feed rate in conjunction with DOP and poly-P measurements.Under this scenario, there likely will be a practical limit where the ammonia removalcapacity of the Anammox process would be negatively affected but the testing may hint at anoptimal operating strategy.· Due to funding constraints, the research did not incorporate characterization of the Anammoxbacterial activity in the bioreactor using molecular techniques. Availability of moleculartesting for specific genetic biomarkers would have provided a means to measure the presence127of Anammox genera and allowed for a more complete understanding of the factors forachieving optimal nitrogen removal rates.128BibliographyAckman, R. G, 1972. Porous Polymer Bead Packing and Formic Acid Vapor in the GLC ofVolatile Free Fatty Acids, Journal of Chromatographic Science, 10, pp. 560-565.Akerback, N., Engblom, S., Sahlén, K., 2009. Ammonium absorption in reject water usingvermiculite. Proceedings of International Conference on Nutrient Recovery fromWastewater Streams, Vancouver (pp. 711-719). London, UK: IWA Publishing.Allison, F.E., Kefauver, M., Roller, E.M., 1953. Ammonium Fixation in Soils. Soil ScienceSociety Proceedings, NY, pp. 107-110.APHA, AWWA and WEF. Standard Methods for the Examination of Water & Wastewater. 21stEdition. Washington: American Public Health Association, 2005.Arita, C.E.Q., Peebles, C., Bradley, T.H., 2015. Scalability of combining microalgae-basedbiofuels with wastewater facilities: A review. Algal Research, 9, pp.160-169.Arnaldos, M., Pagilla, K., 2010. Effluent dissolved organic nitrogen and dissolved phosphorusremoval by enhanced coagulation and microfiltration. Water research, 44(18), pp.5306-5315.Babatunde, A.O., Zhao, Y.Q., 2010. Equilibrium and kinetic analysis of phosphorus adsorptionfrom aqueous solution using waste alum sludge. Journal of Hazardous Materials, 184,pp.746-752.Baker S., Lee, Y., Taniguchi, D., Szoke, N., Zaleski, A., Oleszkiewicz, J., 2009. Design andoperation of a full-scale sidestream bioaugmentation plant. Proceedings of WaterEnvironment Federation Technology and Exhibition Conference 2009, Orlando (pp. 139-148). Alexandria: Water Environment Federation.129Banu, J.R, Do, K., Kaliappan, S, Yeom, I., 2009. Effect of alum on nitrification duringsimultaneous phosphorus removal in anoxic/oxic reactor. Biotechnology and BioprocessEngineering, 14, pp. 543-548.Barker, D.J., Stuckey, D.C. (1999). A review of soluble microbial products (SMP) in wastewatertreatment systems. Water Research, 33(14), pp.3063-3082.Barnard, J.L., 2007. Elimination of eutrophication through resource recovery – National WaterResearch Institute Clarke Award lecture. Retrieved from Last accessedApril 25, 2016.Barnard, J., Houweling, D., Analla, H., Steichen, M., 2010a. Fermentation of mixed liquor forphosphorus removal. Proceedings of Water Environment Federation Technology andExhibition Conference 2010, New Orleans (pp. 59-71). Alexandria: Water EnvironmentFederation.Barnard, J.L., Holakoo, L., Meiring, P.G.J., 2010b. Recovery of algae from pond treatmentsystems. Proceedings of Water Environment Federation Technology and ExhibitionConference 2010, New Orleans (pp. 1856-1873). Alexandria: Water EnvironmentFederation.Bauer, R., 2010. Operating North America’s First Full Scale Nutrient Recovery Facility.Proceedings of Water Environment Federation Technology and Exhibition Conference2010, New Orleans (pp. 59-71). Alexandria: Water Environment Federation.Berg, M., Cole, B., Wesson, K., Ray, J., Currie, T., Covington, J., Buff, B., Whitlock, D., 2013.Making the case for sustainable biosolids management solutions: A roadmap for fourteenfacilities. Proceedings of Water Environment Federation Technical Exhibition andConference 2013 (n.p.). Alexandria, Va: Water Environment Federation.130Bettazzi, E., Caffaz, S., Vannini, C., Lubello, C., 2010. Nitrite inhibition and intermediateseffects on Anammox bacteria: A batch-scale experimental study. Process Biochemistry,45(4), pp. 573-580.Beun, J.J., van Loosdrecht, M.C.M., Heijnen, J.J., 2000. Aerobic granulation. Water Science andTechnology, 41(4-5) pp. 41-48.Bouchez, T., Patureau, D., Dabert, P., Juretschko, S., Doré, J., Delgenès, P., Moletta, R., 2000.Ecological study of a bioaugmentation failure. Environmental Microbiology, 2(2), pp.179-190.Bratby, J. Coagulation and flocculation in water and wastewater treatment. London, UK: IWAPublishing, 2006.Brooks, M., 1999. Breakpoint chlorination as an alternate means of ammonia-nitrogen removalat a water reclamation plant (M.A.Sc. Thesis). North Virginia Center, Falls Church, Va.Brown, P.A., Gill, S.A., Allen, S.J., 2000. Metal removal from wastewater using peat. WaterResearch, 34(16), pp. 3907-3916.Cakir, F.Y., Stenstrom, M.K., 2005. Greenhouse gas production: A comparison between aerobicand anaerobic wastewater treatment technology. Water Research, 39(17), pp.4197-4203.Carr, S.T.D., Dobbie, K.E., Heal, K.V., Smith, K.A., 2009. The use of phosphorus-saturatedochre as a fertilizer. Proceedings of International Conference on Nutrient Recovery fromWastewater Streams, Vancouver (pp. 623-633). London, UK: IWA Publishing.Carrio, L.A., Lopez, A.R., Krasnoff, P.J., Donnellon, J.J., 1985. Sludge reduction by in-plantprocess modification: New York City’s experiences. Journal of Water Pollution ControlFederation, 57(2), pp. 116-121.131Carrio, L., Sexton, L., Lopez, A., Gopalakrishnam, K., Sapienza, V., 2003. Ammonia-nitrogenremoval from centrate 10 years of testing and operating experience in NYC. Proceedings ofWater Environment Federation Technology and Exhibition Conference 2003, (np).Alexandria, Va: Water Environment Federation.Cath, T.Y., Childress, A.E., Elimelech, M., 2006. Forward osmosis: principles, applications, andrecent development. Journal of Membrane Science, 281, pp. 70–87.Cecchi, F., Pavan, P., Mata-Alvarez, J., 1996. Anaerobic co-digestion of sewage sludgeapplication to the macroalgae from the Venice lagoon. Resources, Conservation andRecycling, Vol. 17, pp. 57-66.Chen, J., Ji, Q., Zheng, P., Chen, T., Wanga, C., Mahmood, Q., 2010. Floatation and control ofgranular sludge in a high-rate Anammox reactor. Water Research, 44, pp. 3321–3328.Constantine, T., Johnson, B., 2006. Innovative Approaches for Treatment of Return Liquorsfrom Anaerobic Digestion. Proceedings of Residuals and Biosolids Management Conference2006 (pp. 608-617). Alexandria, Va: Water Environment Federation.Dapena-Mora, A., Fernández, I., Campos, J.L., Mosquera-Corral, A., Méndez, R., Jetten,M.S.M., 2007. Evaluation of activity and inhibition effects on Anammox process by batchtests based on nitrogen gas production. Enzyme and Microbial Technology, 40,, L.E., Bashan, Y., 2004. Recent advances in removing phosphorus from wastewaterand its future use as fertilizer (1997–2003). Water Research, 38(18) pp. 4222– Clippeleir, H., Jimenez, R., Giraldo, E., Wett, B., Dockett, N., Riffat, R., Al-Omari, A.,Murthy, S., 2013. Screens as a method for selective anammox retention in single stagedeammonification processes. Proceedings of Water Environment Federation Nutrient132Removal and Recovery Conference 2013, Vancouver, B.C. (pp. na). Alexandria: WaterEnvironment Federation.Dwyer, J., Starrenburg, D., Tait, S., Barr, K., Batstone, D.J., Lanta, P., 2008. Decreasingactivated sludge thermal hydrolysis temperature reduces product colour, without decreasingdegradability. Water Research, 42(18), pp. 4699 –4709.Dwyer, J., Griffiths, P., Lant, P., 2009. Simultaneous colour and DON removal from sewagetreatment plant effluent: Alum coagulation of melanoidin. Water Research, 43(2), pp. 553–561.Easwaran S.P., Novak, J.T., Doyle, M.L., 2009. Mechanism and optimization of the Cannibalprocess. Proceedings of Water Environment Federation Technology and ExhibitionConference, Orlando (pp. 7164-7180). Alexandria, Va: Water Environment Federation.Ekholm, P., Krogerus, K., 2003. Determining agal-available phosphorus of differing origin:routine phosphorus analyses versus algal assays. Hydrobiologia, 492, pp. 29-42.Egli, K., Fanger, U., Alvarez, P.J.J., Siegrist, H., van der Meer, J.R., Zehnder, A.J.B., 2001.Enrichment and characterization of an Anammox bacterium from a rotating biologicalcontactor treating ammonium rich leachate. Archives of Microbiology, 175, pp. 198-207.Emerson, K., Lund, R.E., Thurston, R.V. and Russo, R.C., 1975. Aqueous ammonia equilibriumcalculations: effect of pH and temperature. Journal of the Fisheries Research Board ofCanada, Vol. 32, pp. 2379-2383.Eschborn, R., Tepe-Sencayir, N., Bodniewicz, B., Brandt, R.C., Johnston, T., Derminassian, R.,Kharkar, S., Murthy, S. Bailey, W., 2010. Sidestream ammonia recovery, ammonium sulfatefertilizer production and market assessment of the US Mid-Atlantic Region. Proceedings of133Water Environment Federation Technology and Exhibition Conference 2010, New Orleans(pp. 5958-5962). Alexandria, Va: Water Environment Federation.Fernandez, I., Vázquez-Padín, J.R., Mosquera-Corral, A., Campos, J.L., Méndez, R., 2008.Biofilm and granular systems to improve Anammox biomass retention. BiochemicalEngineering Journal, 42, pp. 308-313.Figdore, B., Bowden, G., Bodniewicz, B. Bailey, W., Derminassian, R., Kharkar, S., Murthy, S.,2010. Impact of thermal hydrolysis solids pretreatment on sidestream treatment processselection at the DC Water Blue Plains AWTP. Proceedings of Water EnvironmentFederation Technology and Exhibition Conference 2010, New Orleans (pp. 5927-5949).Alexandria, Va: Water Environment Federation.Fujimoto, N., Mizuochi, T., Togami, Y., 1991. Phosphorus fixation in the sludge treatmentsystem of a biological phosphorus removal process. Water Science and Technology, 23,pp.635-640.Furukawa, K., Rouse, J.D., Imajo, U., Sugino, H., Fujii, T., 2001. Establishment of an anaerobicammonium oxidizing culture in continuous flow treatment with non-woven biomass carrier.Proceedings of Water Environment Federation Technology and Exhibition Conference 2001(n.p.). Alexandria, Va.: Water Environment Federation.Furukawa, K., Lieu, P.K., Tokitoh, H., Fujii, T., 2006. Development of single-stage nitrogenremoval using Anammox and partial nitritation (SNAP) and its treatment performances.Water Science and Technology, 53, pp. 83-90.Fux, C., Boehler, M., Huber, P., Brunner, I., Siegrist, H., 2002. Biological treatment ofammonium-rich wastewater by partial nitritation and subsequent anaerobic ammoniumoxidation (Anammox) in a pilot plant. Journal of Biotechnology, 99, pp. 295-306.134Galaraneau, E., Gehr, R., 1996. Phosphorus removal from wastewaters: experimental andtheoretical support for alternative mechanisms. Water Research, 31(2), pp.328-338.Garrido, J.M., van Benthum, W.A.J., van Loosdrecht, M.C.M., Heijnen, J.J., 1997. Influence ofdissolved oxygen concentration on nitrite accumulation in a biofilm airlift suspensionreactor. Biotechnology and Bioengineering, 53, pp. 168–178.Giraldo, E., 2009. Nitrous oxide emissions from wastewater treatment plants. A balancing act.Proceedings of Water Environment Federation Nutrient Removal Conference (pp. 357-376).Alexandria, Va: Water Environment Federation.Goel, R.K., Noguera, D.R., 2006a. Evaluation of sludge yield and phosphorus removal in aCannibal solids reduction process. Journal of Environmental Engineering, 132, pp. 1331-1337.Goel, R.K., Noguera, D.R., 2006b. Enhanced biological phosphorus removal in a Cannibal typeactivated sludge process. Proceedings of Water Environment Federation Technology andExhibition Conference 2006 (pp. 5933-5940). Alexandria, Va: Water EnvironmentFederation.Grigoropoulos, S. G., Vedder, R. C., Max, D. W., 1971. Use of alternative aluminium basedchemicals in coagulation with particular reference to phosphorus removal. Water PollutionControl Federation, Vol. 43, No. 12, pp. 2366-2382.Gu, A. Z., Liu, L., Neethling, J. B.,  Stensel H. D., Murthy, S., 2011. Treatability and fate ofvarious phosphorus fractions in different wastewater treatment processes. Water Science andTechnology, 63(4), pp. 804-810.135Guo, S., Qu, F., Ding, A., Bai, L. and Li, G. (2015). Effects of poly aluminum chloride dosingpositions on the performance of a pilot scale anoxic/oxic-membrane bioreactor (A/O-MBR).Water Science and Technology, 72(5), pp.689-695.Guo, X., Zeng, L., Li, X., Park, H., 2008. Ammonium and potassium removal for anaerobicallydigested wastewater using natural clinoptilolite followed by membrane pretreatment.Journal of Hazardous Materials, 151, pp. 125–133Gűven, D., Dapena, A., Kartal, B., Schmid, M.C., Maas, B., van de Pas-Schoonen, Sozeen, S.,Mendez, R., Op den Camp, H.J.M., Jetten, M.S.M., Schmidt, I., 2005. Propionate oxidationby and methanol inhibition of anaerobic ammonium oxidizing bacteria. Applied andEnvironmental Microbiology, February, pp. 1066-1071.Hanhoun, M., Azzaro-Pantel, C., Biscans, B., Frèche, M., Montastruc, L., Pibouleau, L. andDomenech, S., 2009. A Thermochemical Approach for Struvite Precipitation Modeling fromWastewater. Proceedings of International Conference on Nutrient Recovery fromWastewater Streams, Vancouver (pp. 121-130). London, UK: IWA Publishing.Hassan, P., Mavinic, D.S., Rezania, B., Kelly, H., Lo, K.V., 2013. ANAMMOX combined withstruvite crystallization: a sustainable solution for nitrogen and phosphorus removal andrecovery from side stream processes. Proceedings of Water Environment FederationNutrient Removal and Recovery Conference 2013, Vancouver, B.C. (pp. na). Alexandria:Water Environment Federation.Hatton, W. & Simpson, A.M. (1985) Use of alternative aluminium based chemicals incoagulation with particular reference to phosphorus removal, Environmental TechnologyLetters, 6:1-11, pp.225-230.136Hellinga, C., Schellen, A.A.J.C., Mulder, J.W., van Loosdrecht, M.C.M., Heijnen, J.J., 1998. TheSharon process: An innovative method for nitrogen removal from ammonium-rich wastewater. Water Science and Technology, 37(9), pp 135–142.Hellinga, C., van Loosdrecht, M.C.M., Heijnen, J.J., 1999. Model based design of a novelprocess for nitrogen removal from concentrated flows. Mathematical and ComputerModelling of Dynamical Systems, 5(4), pp. 351-371.Holloway, R.W., Childress, A.E., Dennett, K.E., and Cath, T.Y., 2007. Forward osmosis forconcentration of anaerobic digester centrate. Water Research, 41, pp. 4005-4014.Imajo, U., Tokutomi, T., and Furukawa, K., 2004. Granulation of Anammox organisms inupflow reactors. Water Science and Technology, 49, pp. 155-163.Intergovernmental Panel on Climate Change, 2006. 2006 IPCC Guidelines for NationalGreenhouse Gas Inventories, Prepared by the National Greenhouse Gas InventoriesProgramme, Eggleston H.S., Buendia L., Miwa K., Ngara T. and Tanabe K. (eds).Published: IGES, Japan.Isaka, K., Suwa, Y., Kimura, Y., Yamagishi, T., Sumino, T., Tsuneda, S., 2008. Anaerobicammonium oxidation (Anammox) irreversibly inhibited by methanol. Applied Microbiologyand Biotechnology, 81, pp. 379–385.Janus, H.M., van der Roest, H.F., 1997. Don't reject the idea of treating reject water. WaterScience and Technology, 35, pp. 27-34.Jaroszynski, L.W., Cicek, N., Sparling, R., Oleszkiewicz, J.A., 2011. Importance of operating pHin maintaining Anammox activity in moving bed biofilm reactors. Bioresource Technology,102, pp. 7051–7056.137Jetten, M.S.M., Horn, S.J., van Loosdrecht, M.C.M., 1997. Towards a more sustainablemunicipal wastewater treatment system. Water Science and Technology, 31(9), pp. 171-180.Jetten, M.S.M., Strous, M., van de Pas-Schoonen, K.T., Schalk, J., van Dongen, U.G.J.M., vande Graaf, A.A., Logemann, S., Muyzer, G., van Loosdrecht, M.C.M., Kuenen, J.G., 1998.The anaerobic oxidation of ammonium. FEMS Microbiology Reviews, 22, pp. 421-437.Jetten, M.S.M., Wagner, M., Fuerst, J., van Loosdrecht, M., Kuenen, G., Strous, M., 2001.Microbiology and application of the anaerobic ammonium oxidation process. EnvironmentalBiotechnology, 12, pp. 283–288.Jiang, J., Graham, N. J. D., 1998. Pre-polymerised inorganic coagulants and phosphorus removalby coagulation - A review. Water SA, Vol. 24 No. 3, pp. 237-244.Jones, R., Takács, I., 2004. Modeling the impact of anaerobic digestion on the overallperformance of biological nutrient removal wastewater treatment plants. Proceedings ofWater Environment Federation Technology and Exhibition Conference 2004, (n.p.).Alexandria, Va: Water Environment Federation.Joss, J., Salzgeber, D., Eugster, J., König, R., Rottermann, K., Burger, S., Fabijan, P.,Neleumann, S., Mohn, J., Siegrist, H., 2009. Full-scale nitrogen removal from digesterliquid with partial nitritation and anammox in one SBR. Environmental Science andTechnology, 43, pp. 5301–5306.Jung, J.Y., Kang, S.H., Chung, Y.C., Ahn, D.H., 2007. Factors affecting the activity of anammoxbacteria during start up in the continuous culture reactor. Water Science and Technology,55(1), pp. 459-468.Kampschreur, M.J.,  Poldermans, R., Kleerebezem, R., van der Star, W.R.L., Haarhuis, R.,Abma, W.R., Jetten, M.S.M., van Loosdrecht, M.C.M., 2009. Emissions of nitrous oxide138and nitric oxide from a full-scale single-stage nitritation-Anammox reactor. Water Scienceand Technology, 60(12), pp. 3211-3217.Kartal, B., Marcel, M., Kuypers, M., Lavik, G., Schalk, J., Op den Camp, H.J.M., Jetten,M.S.M., Strous, M., 2007. Anammox bacteria disguised as denitrifiers, nitrate reduction todinitrogen gas via nitrite and ammonium. Environmental Microbiology, 9(3), pp. 635–642.Kartal, B., Kuenen, J.G., & van Loosdrecht, M.C.M., 2010. Sewage treatment with Anammox.Science, Vol. 328, pp. 702-703.Kim, K., Kim, Y., Kim, I., Park, G., Lee, E., 2006. Electrochemical conversion characteristics ofammonia to nitrogen. Water Research, 40(7), pp. 1431-1441.Kim, H. and Dempsey, B.A., 2012. Comparison of two fractionation strategies forcharacterization of wastewater effluent organic matter and diagnosis of membrane fouling.Water Research, 46, pp. 3714-3722.Kimura, Y., Isaka, K., Kazama, F., Sumino, T., 2010. Effects of nitrite inhibition on anaerobicammonium oxidation. Applied Microbiology and Biotechnology, 86, pp.359-365.Kong, Q., Li, L., Martinez, B., Chen, P., Ruan, R., 2010. Culture of microalgae chlamydomonasreinhardtii in wastewater for biomass feedstock production. Applied Biochemistry andBiotechnology, 160, pp. 9-18.Kos, P., 1998. Short SRT (solids retention time) nitrification process/flowsheet. Water Scienceand Technology, 38(1), pp 23-29.Krupa, S.V., 2003. Effects of atmospheric ammonia on terrestrial vegetation: A review.Environmental Pollution, 124, pp. 179–221Kuenen, J.G., 2008. Anammox bacteria, from discovery to application. Nature, 6, pp. 320-326.139Lahav, O., Green, M., 1998. Ammonium removal using ion exchange and biologicalregeneration. Water Research, 32(7), pp. 2019-2028.Lei, X., Maekawa, T., 2007. Electrochemical treatment of anaerobic digestion effluent using aTiPt–IrO2 electrode. Bioresource Technology, 98, pp. 3521–3525.Lei, C., Ma, J., Li, D., Zhang, W., Guang, X., Ma, J., 2010. Composite denitrification reagent forhigh concentration ammonia removal by air stripping. Environmental Science andTechnology, 55, pp. 2657-2661.Lévesque, M. (1969). Characterization of model and soil organic matter metal-phosphatecomplexes. Canadian Journal of Soil Science, 49, pp 365-373.Lewis, F.M, Haug, R.T., Poosti, A, Redd, K.R., 2008. A powerful byproduct. WaterEnvironment and Technology. 20(1), p. 34.Li, Y., Chen, Y., Chen, P., Min, M., Zhou, W., Martinez, B., Zhu, J., Ruan, R., 2011.Characterization of a microalga chlorella sp. well adapted to highly concentrated municipalwastewater for nutrient removal and biodiesel production. Bioresource Technology, 102, pp.5138–5144.Liu, L., Nejadghafar, E., Smith, D.S., Bracken, M., Neethling, J.B., Stensel, H.D., 2011.Occurence, implication and bioavailability of dissolved organic phosphorus (DOP) inadvanced wastewater effluents. Proceedings of Water Environment Federation TechnicalExhibition and Conference 2011 (n.p.). Alexandria, Va: Water Environment Federation.Lopez, H., Puig, S, Ganigue, R., Ruscalleda, M., Balaguer, M.D., Colprim, J., 2008. Start-up andenrichment of a granular anammox SBR to treat high nitrogen load wastewaters. Journal ofChemical Technology and Biotechnology, 83, pp. 233-241.140Liqui-Cel, 2009. Technical Brief, Using ‘TransMembraneChemiSorption’ (TMCS) for AmmoniaRemoval. Retrieved from Last accessed May 12, 2016.Maher, C., Neethling, J.B., Murthy, S., Pagilla, K., 2015. Kinetics and capacities of phosphorussorption to tertiary stage wastewater alum solids, and process implications for achievinglow-level phosphorus effluents. Water Research, 85, pp.226-234.Mavinic, D.S., Knezevic, Z., Anderson, B.C. and Oleszkiewicz, J.A. (1995). Fate of nutrientsduring anaerobic co-digestion of sludges from a biological phosphorus removal process.Environmental Technology, 16(12), pp. 1165-1173.Mavinic, D.S., Koch, F.A., Hall, E.R., Abraham, K., Niedbala, D. (1998). Anaerobic Co-Digestion of Combined Sludges from a BNR Wastewater Treatment Plant. EnvironmentalTechnology, Vol. 19(1), pp. 35-44.Mavinic, D.S., Koch, F.A., Huang, H., Lo, K.V., 2007. Phosphorus reduction in an anaerobicdigester through struvite harvesting. Journal of Environmental Engineering and Science,34(6), pp. 561-571.Mayer, M., Smeets, W., Braun, R., Fuchs, W., 2009. Enhanced ammonium removal from liquidanaerobic digestion residuals in advanced sequence batch reactor system. Water Science andTechnology, 60(7), pp. 1649-1660.McGinn, P.J., Dickinson, K.E., Bhatti, S., Frigon, J., Guiot, S.R., O’Leary, S.J.B., 2011.Integration of microalgae cultivation with industrial waste remediation for biofuel andbioenergy production opportunities and limitations. Photosynthesis Research, 109, pp. 231-247.Metcalf & Eddy. Wastewater Engineering – Treatment and Resource Recovery (Fifth Edition).New York: McGraw Hill Education, 2014.141Michael-Kordatou, I., Michael, C., Duan, X., He, X., Dionysiou, D.D., Mills, M.A., Fatta-Kassinos, D., 2015. Dissolved effluent organic matter: characteristics and potentialimplications in wastewater treatment and reuse applications. Water Research, 77, pp. 213-248.Min, M., Wang, L., Li, Y., Mohr, M.J., Hu, B., Zhou, W., Chen, P., Ruan, R., 2011. Cultivatingchlorella sp. in a pilot-scale photobioreactor using centrate wastewater for microalgaebiomass production and wastewater nutrient removal. Applied Biochemistry andBiotechnology, 165, pp. 123-137.Minocha, V.K., Rao, A.V.S.P., 1988. Ammonia removal and recovery from urea fertilizer plantwaste. Environmental Technology Letters, 9, pp. 655-664Mulder, A., van de Graaf, A.A., Robertson, L.A., Kuenen, J.G., 1995. Anaerobic ammoniumoxidation discovered in a denitrifying fluidized bed reactor. FEMS Microbiology Ecology,17, pp. 177-184.Mulder, J.W., van Loosdrecht, M.C.M., Hellinga, C., van Kempen, R., 2001. Full-scaleapplication of the SHARON process for treatment of rejection water of digested sludgedewatering. Water Science and Technology, 43(1), pp. 127-134.Neethling, J.B., Benisch, M., Clark, D., Fisher, D., Gu, A.Z., 2007. Phosphorus speciationprovides direction to produce 0.01 μgL. Proceedings of Water Environment FederationTechnology and Exhibition Conference 2007 (pp. 1607-1624). Alexandria, Va: WaterEnvironment Federation.Neethling, J.B., Stensel, H.D., 2013. Nutrient species implications for technology performanceand water quality impacts. Proceedings of Water Environment Federation Nutrient Removal142and Recovery Conference 2013, Vancouver, B.C. (pp. na). Alexandria: Water EnvironmentFederation.Nobbmann, U., Morfesis, A., Billica, A. and Gertig, K. 2010. The role of zeta potential in theoptimizatin of water treatment. Proceedings of Nanotech Conference and Exposition,Anaheim, CA (pp. 605-607). Cambridge, MA: TechConnect World.Novak, J.T., Chon, D.H., Curtis, B., Doyle, M., 2007. Biological solids reduction using theCannibal process. Water Environment Research, 79(12), pp. 2380-2386.Novak, J.T., Khanthongthip, P., Doyle, M., 2011. Impact of substrate feed patterns on solidsreduction by the Cannibal™ process. Proceedings of Water Environment FederationResiduals and Biosolids Conference 2011, (pp. 660-673). Alexandria, Va: WaterEnvironment Federation.Oldham, W.K., Rabinowitz, B. (2002). Development of biological nutrient removal technologyin western Canada. Canadian Journal of Civil Engineering, 28(S1), pp.92-101.Omoike, A.I., Vanloon, G.W., 1999. Removal of phosphorus and organic matter removal byalum during wastewater treatment, Water Research, 33(17), pp.3617-3627.Orentlicher, M., Brown, S., Hall, S., 2009. Centrate ammonia reduction with ARP®: pilot datawith synthetic and actual wastewaters. Proceedings of Water Environment FederationNutrient Removal Conference 2009 (pp. 546-555). Alexandria, Va: Water EnvironmentFederation.Pagilla, K.R., Czerwionka, K., Urgun-Demirtas, M. and Makinia, J., 2008. Nitrogen Speciationin Wastewater Treatment Plant Influents and Effluents - The US and Polish case Studies.Water Science and Technology, 57(10), pp. 1511–1517.143Park, C., Abu-Orf, M.M., Novak, J.T., 2006. The digestibility of waste activated sludges. WaterEnvironment Research, 78, pp. 59-68.Pawlowski, L., 1979. Pilot plant study on ammonia and water recycling in nitrogen industry.Resource Recovery and Conservation, Vol. 4, pp. 43-57.Pitman, A.R., 1999. Management of Biological Nutrient Removal Plant Sludges – Change theParadigms. Water Research, Vol. 33, No. 5, pp. 1141-1146.Prakasam, T.B.S., Soszynski, S., Zenz, D.R., Lue-Hing, C., Blyth, L., Sernel, G., 1990. Effectsof recycling thermophilic sludge on the activated sludge process. United StatesEnvironmental Protection Agency, September, 1990.Pressley, T.A., Bishop, D.F., Roan, S., 1972. Ammonia-nitrogen removal by breakpointchlorination. Environmental Science and Technology, 6(7), pp. 622-628.Ras,  M.,  Lardon,  L.,  Bruno,  S.,  Bernet,  N.,  Steyer,  J.,  2011.  Experimental  study  on  a  coupledprocess of production and anaerobic digestion of chlorella vulgaris. BioresourceTechnology, 102(1), pp.200–2006.Resnik, K.P., Yeh, J.T., Pennline, H., 2004. Aqua ammonia process for simultaneous removal ofCO2, SO2 and NOx. International Journal of Environmental Technology and Management,4(1), pp. 89–104.Rosenwinkel, K., Cornelius, A., 2005. Deammonification in the moving bed process fortreatment of wastewater with high ammonia content. Chemical Engineering Technology, 28,pp. 49-52.Sagberg P., Ryrfors, P., Berg, K.G., 2006. 10 Years of operation of an integrated nutrientremoval treatment plant, part 2: sludge and side stream treatment. Retrieved from, May 15, 2010.144Salem, S., Berends, D., Heijnen, J.J., van Loosdrecht, M.C.M., 2002. Model-based evaluation ofa new upgrading concept for N-removal, Water Science and Technology, 45(6), p. 169–176.Salem, S., Berends, D.H.J.G., Heijnen, J.J., van Loosdrecht, M.C.M., 2003. Bio-Augmentationby nitrification with return sludge. Water Research, 37(8), pp. 1794-1804Scanlan, P., Hong, M., Olson, S., 2010. Energy Recovery Options: What is Their True Value?,Proceeding of Water Environment Federation Residuals and Biosolids Conference 2010 (pp.1-12). Alexandria, Va.: Water Environment Federation.Shanchayan, B., Parker, W., Pride, C., 2004. Pilot and full scale studies of the treatment ofATAD offgases with biofilters. Proceedings of the Water Environment Federation TechnicalExhibition and Conference 2004 (n.p.). Alexandria, Va: Water Environment Federation.Sengupta, S., Pandit, A., 2010. Selective removal of phosphorus from wastewater combined withits recovery as a solid-phase fertilizer. Proceedings of Water Environment FederationTechnology and Exhibition Conference 2010 (pp. 7060-7079). Alexandria, Va: WaterEnvironment Federation.Siegrist, H., Reithaar, S., Koch, G., Lais, P., 1998. Nitrogen loss is a nitrifying rotating contactortreating ammonmium-rich wastewater without organic carbon. Water Science andTechnology, 38, pp. 241-248.Siemens, 2011. Cannibal solids reduction process. Retrieved at, May24, 2016.Sliekers, A.O., Derwort, N., Gomez, J.L.C., Strous, M., Kuenen, J.G., Jetten, M.S.M., 2002.Completely autotrophic nitrogen removal over nitrite in one single reactor. Water Research,36(10), pp. 2475–2482.145Sova, R., Neethling, J.B., Kinnear, D., Bakke, B., Brandt, G., Wilson, R., Crisler, S., 2004.Prenitrification and seeding for enhanced nitrification. Proceedings of the WaterEnvironment Federation Technical Exhibition Conference 2004 (n.p.). Alexandria, Va:Water Environment Federation.Stinson, B., 2001. Overview of biological nitrogen and phosphorus removal. Presentation madeto New Jersey Water Environment Association, May 3, 2001.Strous, M., Heijnen, J.J., Kuenen, J.G., Jetten, M.S.M., 1998. The sequencing batch reactor as apowerful tool for the study of slowly growing Anammox. Applied Microbiology andBiotechnology, 50, pp. 589-596.Strous, M., Kuenen, J.G., Jetten, M.S.M., 1999. Key physiology of anaerobic ammoniumoxidation. Applied and Environmental Microbiology, 65, pp. 3248-3250.Stumpf, D., Heinzmann, B., Schwarz, R., Gnirss, R., Kraume, M., 2009. Induced struviteprecipitation in an airlift reactor for phosphorus recovery. Proceedings from IWAInternational Conference on Nutrient Recovery from Wastewater Streams, Vancouver (pp.179-192).Szatkowska, B., Cema, G., Plaza, E., Trela, J., Hultman, B., 2007. A one-stage system withpartial nitritation and anammox processes in the moving-bed biofilm reactor, Water Scienceand Technology, 55, pp. 19-26.Tchobanoglous, George. & Burton, Franklin L. & Metcalf & Eddy, 2003. Wastewaterengineering: treatment, disposal, and reuse, 4th edition. New York: McGraw-HillScience/Engineering.Third, K.A., Sliekers, A.O., Kuenen, J.G., Jetten, M.S.M., 2001. The CANON system(completely autotrophic nitrogen-removal over nitrite) under ammonium limitation:146interaction and competition between three groups of bacteria. Systematic and AppliedMicrobiology, 24, pp. 588-596.Thornton, A., Pearce, P., Parsons, S.A., 2007. Ammonium removal from digested sludge liquorsusing ion exchange. Water Research, 41(2), pp. 433-439.Thurman, E.M., Wershaw, R.L., Malcolm, R.L., Pinchney, D.J., 1982. Molecular size of aquaticsubstances. Organic Geochemistry, 4, pp. 27-35.Tooker N., Guswa, S., Horton, J., Hastings, M., Devalk, C., 2010. Pilot testing for low-levelphosphorus removal: confronting non-reactive phosphorus challenges. Proceedings of WaterEnvironment Federation Technology and Exhibition Conference 2010, New Orleans (pp.959-975). Alexandria, Va: Water Environment Federation.Torpey, W.N., Andrews, J., Basilico, J.V., 1984. Effects of multiple digestion on sludge. JournalWater Pollution Control Federation, 56, pp. 62-68.Trigo, C., Campos, J.L., Garrido, J.M., Méndez, R., 2006. Startup of the Anammox process in amembrane bioreactor. Journal of Biotechnology, 126, pp. 475-487.Ueno, Y. Fuji, M., 2001. Three years experience of operating and selling recovered struvite fromfull-scale plant. Environmental Technology, 22, pp. 1373-1381.Urgun-Demirtas, M., Pagilla, K.R., Kunetz, T.E., Sobanski, J.P., Law, K.P., 2008.Bioavailability of dissolved organic nitrogen in treated effluents. Water EnvironmentResearch, 80(5), pp. 397-406.Uysal, A., Yilmazel, Y.D., Demirer, G.N., 2010. The determination of fertilizer quality of theformed struvite from effluent of a sewage sludge anaerobic digester. Journal of HazardousMaterials, 181, pp. 248-254.147van de Graff, A.A., Mulder, A., de Bruijn, P., Jetten, M.S.M., Robertson, L.A., Kuenen, J.G.,1995. Anaerobic oxidation of ammonium is a biologically mediated process. Applied andEnvironmental Microbiology, April, pp. 1246–1251van de Graaf, A.A., de Bruijn, P., Robertson, L.A., Jetten, M.S.M., Kuenen, J.G., 1996.Autotrophic growth of anaerobic ammonium oxidizing micro-organisms in a fluidized bedreactor. Microbiology, 142, pp. 2187-2196.van der Star, W.R.L., Abma, W.R. Blommers, D., Mulder, J.,Tokutomi, T., Strous, M.,Picioreanu, C., van Loosdrecht, M.C.M., 2007. Startup of reactors for anoxic ammoniumoxidation: experiences from the first full-scale anammox reactor in Rotterdam. WaterResearch, 41(18), pp. 4149-4163.van Dongen, U., Jetten, M.S.M., van Loosdrecht, M.C.M. (2001). The SHARON®-Anammox®process for treatment of ammonium rich wastewater. Water Science and Technology, 44(1),pp 153-160.van Hulle, S.W.H., Volcke, E.I.P., Teruel, J.L., Donckels, B., van Loosdrecht, M.C.M.,Vanrolleghem, P.A., 2007. Influence of temperature and pH on the kinetics of the Sharonnitritation process. Journal of Chemical Technology and Biotechnology, 82, pp. 471-480.Veolia Water, 2011. ANITA™ Mox - A BioFarm solution for fast start-up of deammonifyingMBBRs. Accessed,Abstract_ANITA_Mox_BioFarm-1.pdf, May, 2016.Verstraete, W., Morgan-Sagastume, F., Aiyuk, S., Waweru, M., Rabey, K., Lissens, G., 2005.Anaerobic digestion as a core technology in sustainable management of organic matter.Water Science and Technology, 52, pp. 59-66.148Wadhawan, T., McEvoy, J., Pruess, B., Al-Omari, A., Murthy, S. and Khan, E., 2015.Biodegradation of DON under different biological wastewater treatment process conditions.Proceedings of Water Environment Federation Technology and Exhibition Conference 2015,Chicago, (pp. 6047-6060). Alexandria: Water Environment Federation.Waki, M., Tokutomi, T., Yokoyama, H., Tanaka, Y., 2007. Nitrogen removal from animal wastetreatment water by Anammox enrichment. Bioresource Technology, 98, pp. 2775-2780.Wang, J., Burken, J.G., Zhang, X., 2006. Effect of seeding material and mixing strength onstruvite precipitation. Water Environment Research, 78(2), pp. 125-132.Wang, T., Zhang, H., Yang, F., Liu, S., Fu, Z., Chen, H., 2009. Start-up of the Anammox processfrom the conventional activated sludge in a membrane bioreactor. Bioresource Technology,100, pp. 2501-2506.Water Environment Research Foundation (WERF), 2009. Dissolved Organic NitrogenCharacterization and Bioavailability in Wastewater Effluents. Alexandria, Va.: IWAPublishing.Water Environment Research Foundation (WERF), 2010. Technology Roadmap for SustainableWastewater Treatment Plants in a Carbon Constrained World. Alexandria, Va.: IWAPublishing.Wett, B., Buchauer, K., Fimml, C., 2007a. Energy Self-Sufficiency as a Feasible Concept forWastewater Treatment Systems. Proceedings of IWA Leading Edge Technology Conference,2007 (pp. 21-24). Singapore: Asian Water.Wett, B., Murthy, S., Takács, I., Hell, M., Bowden, G., Deur, A., O’Shaughnessy, M.O., 2007b.Development and Implementation of a Robust Deammonification Process. Water Scienceand Technology, 56(7), pp 81–88.149Wett, B., Nyhuis, G., Takács, I., Murthy, S., 2010. Development of Enhanced DeammonificationSelector. Proceedings of Water Environment Federation Technology and ExhibitionConference 2010 (pp. 5917-5926). Alexandria, Va: Water Environment Federation.Wild, D., Kisliakova, A., Siegrist, H., 1997. Prediction of recycle phosphorus loads fromanaerobic digestion. Water Research, 31(9), pp.2300-2308.Wilson, C.A., Tanneru, C.T., Banjade, S., Murthy, S.N. and Novak, J.T., 2011. Anaerobicdigestion of raw and thermally hydrolyzed wastewater solids under various operationalconditions. Water Environment Research, 83(9), pp.815-825.World Health Organization, 2005 Air quality guidelines for particulate matter, ozone, nitrogendioxide and sulfur dioxide. Geneva: WHO Press.Worms, I.A.M., Szigeti, Z.A., Dubascoux, S., Lespes, G., Traber, J., Sigg, L., Slaveykova, V.I.,2010. Colloidal organic matter from wastewater treatment plant effluents: characterizationand role in metal distribution. Water Research, 44, pp. 340-350.Zhang, S., Kim, T., Han, T.H., Hwang, S., 2015. Influence of light conditions of a mixture of redand blue light sources on nitrogen and phosphorus removal in advanced wastewatertreatment using scenedesmus dimorphus. Biotechology and Bioprocess Engineering, 20,pp.760-765.Zheng, Z. He, P.J., Zhang, H., Shao, L.M., 2008. Role of dissolved humic substances surrogateson phthalate esters migration. Water Science & Technology, 57(4), pp. 607-612.Zorpas, A.A., Inglezakis, V.J., Stylianou, M., Iren, V., 2010. Sustainable treatment method of ahigh concentrated NH3 wastewater by using natural zeolite in closed-loop fixed bed systems.Open Environmental Sciences, 4, pp. 1-7.150AppendicesAppendix A – sample calibration curves151152153154155Appendix B – data & statistical analyses for autoanalyzer and IC measurementsAverage of Duplicates(mg N/L)Average of Duplicates(mg N/L)Nov. 25 Centrate Feed 1347 1358Dec. 4 Centrate Feed 1450 1365Dec. 8 Centrate Feed 1387 1413Dec. 11 Centrate Feed 1391 1381Dec. 11 Centrate Feed 1363 1383Dec. 16 Centrate Feed 1365 1354Jan. 9 Centrate Feed 1345 1344Jan. 9 Centrate Feed 1449 1445Nov. 25 Anammox Effluent 287 284Dec. 4 Anammox Effluent 299 275Dec. 8 Anammox Effluent 308 280Dec. 11 Anammox Effluent 286 284Dec. 1 Anammox Effluent 289 287Dec. 16 Anammox Effluent 279 277Jan. 9 Anammox Effluent 300 277Total Nitrogen MeasurementsAutoanalyzerMeasurementIon ChromatographMeasurementDate Sample IDComparison of Measured SBR Effluent Total Nitrogent-Test: Paired Two Sample for MeansAutoanalzyerMeasurementsIonChromatographMean 292.3 280.4Variance 99.19 20.49Observations 7 7Pearson Correlation -0.299Hypothesized Mean Difference 0df 6t Stat 2.587P(T<=t) one-tail 0.021t Critical one-tail 1.943P(T<=t) two-tail 0.041t Critical two-tail 2.447156Comparison of Measured Centrate Total Nitrogent-Test: Paired Two Sample for MeansAutoanalzyerMeasurementsIonChromatographMean 1387 1380Variance 1757 1145Observations 8 8Pearson Correlation 0.604Hypothesized Mean Difference 0df 7t Stat 0.542P(T<=t) one-tail 0.302t Critical one-tail 1.895P(T<=t) two-tail 0.605t Critical two-tail 2.365157Appendix C – data & statistical analyses for dissolved phosphorus testingTotalDissolvedPhosphorusAcidDigestedPhosphorusortho-P Poly-P DOP nrDPTotalDissolvedPhosphorusAcidDigestedPhosphorusortho-P Poly-P DOP nrDPmg P/L mg P/L mg P/L mg P/L mg P/L mg P/L mg P/L mg P/L mg P/L mg P/L158 156 154 1.91 1.64 3.5 150 149 148 1.46 0.17 1.63148 146 143 3.31 2.03 5.3 147 145 143 1.68 1.80 3.48151 143 140 3.00 8.35 11.3 143 142 137 4.74 0.95 5.69160 158 154 4.17 2.18 6.3 159 158 154 4.40 0.80 5.20161 161 156 4.92 0.08 5.0 152 151 148 3.14 0.42 3.55175 165 165 0.07 10.27 10.3 164 153 151 1.29 10.95 12.24178 174 168 5.58 4.03 9.6 170 168 163 4.75 2.66 7.40173 171 166 4.95 2.45 7.4 167 165 160 5.27 1.39 6.66173 170 162 7.95 3.34 11.3 168 166 161 4.96 1.59 6.55181 176 170 6.90 4.67 11.6 172 169 161 7.61 3.03 10.64485 484 479 4.92 1.33 6.3 445 445 442 3.26 0.30 3.56481 477 474 3.02 3.93 7.0 458 456 452 4.13 2.59 6.72486 479 477 2.12 6.82 8.9 467 466 463 3.49 1.14 4.63486 485 485 0.38 0.66 1.0 475 475 473 2.38 0.22 2.60482 478 475 3.47 4.15 7.6 473 473 466 7.38 0.17 7.54510 508 506 1.75 2.30 4.1 496 494 490 4.39 2.11 6.49510 510 505 4.37 0.58 5.0 497 497 494 2.44 0.13 2.57478 468 463 4.22 10.83 15.0 487 479 473 6.40 7.57 13.97488 479 474 5.15 8.66 13.8 474 467 462 4.83 6.74 11.58478 476 474 1.84 2.20 4.0 465 462 460 2.21 2.81 5.03474 474 459 15.11 0.48 15.6 469 468 465 2.76 1.36 4.12481 475 470 4.88 5.89 10.8 469 468 458 10.36 0.45 10.81477 474 464 10.33 3.44 13.8 466 465 454 11.13 0.66 11.79Centrate Feed Anammox EffluentAIWWTPCENTRATELAB-SCALEKWTFADCENTRATEComparison of AIWWTP Centrate Feed & Anammox Poly-Pt-Test: Paired Two Sample for MeansCentrate Feed Anammox EffluentMean 4.28 3.93Variance 5.42 4.08Observations 10 10Pearson Correlation 0.778Hypothesized Mean Difference 0df 9t Stat 0.742P(T<=t) one-tail 0.238t Critical one-tail 1.833P(T<=t) two-tail 0.477t Critical two-tail 2.262158Comparison of AIWWTP Centrate Feed & Anammox DOPt-Test: DOPCentrate Feed Anammox EffluentMean 3.90 2.38Variance 9.96 9.90Observations 10 10Pearson Correlation 0.749Hypothesized Mean Difference 0df 9t Stat 2.17P(T<=t) one-tail 0.03t Critical one-tail 1.83P(T<=t) two-tail 0.06t Critical two-tail 2.26Comparison of AIWWTP Centrate Feed & Anammox NRDPt-Test: Paired Two Sample for Means, NRDPCentrate Feed Anammox EffluentMean 8.18 6.30Variance 9.06 10.49Observations 10 10Pearson Correlation 0.776Hypothesized Mean Difference 0df 9t Stat 2.819P(T<=t) one-tail 0.010t Critical one-tail 1.833P(T<=t) two-tail 0.020t Critical two-tail 2.262159Comparison of Lab-scale KWTF AD Centrate Feed & Anammox Poly-Pt-Test: Paired Two Sample for MeansCentrate Feed Anammox EffluentMean 4.73 5.01Variance 15.60 8.88Observations 13 13Pearson Correlation 0.250Hypothesized Mean Difference 0df 12t Stat -0.230P(T<=t) one-tail 0.411t Critical one-tail 1.782P(T<=t) two-tail 0.822Comparison of Lab-scale KWTF AD Feed & Anammox DOPt-Test: Paired Two Sample for MeansCentrate Feed Anammox EffluentMean 3.94 2.02Variance 10.72 6.07Observations 13 13Pearson Correlation 0.751Hypothesized Mean Difference 0df 12t Stat 3.208P(T<=t) one-tail 0.004t Critical one-tail 1.782P(T<=t) two-tail 0.008160Comparison of Lab-scale KWTF AD Centrate Feed & Anammox NRDPt-Test: Paired Two Sample for MeansCentrate Feed Anammox EffluentMean 8.68 7.03Variance 22.50 14.69Observations 13 13Pearson Correlation 0.686Hypothesized Mean Difference 0df 12t Stat 1.70P(T<=t) one-tail 0.06t Critical one-tail 1.78P(T<=t) two-tail 0.12


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items