UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Optimization of phosphorus recovery from anaerobic digester supernatant through a struvite crystallization… Ghosh, Shayok 2016

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

Item Metadata

Download

Media
24-ubc_2017_february_ghosh_shayok.pdf [ 3.23MB ]
Metadata
JSON: 24-1.0340473.json
JSON-LD: 24-1.0340473-ld.json
RDF/XML (Pretty): 24-1.0340473-rdf.xml
RDF/JSON: 24-1.0340473-rdf.json
Turtle: 24-1.0340473-turtle.txt
N-Triples: 24-1.0340473-rdf-ntriples.txt
Original Record: 24-1.0340473-source.json
Full Text
24-1.0340473-fulltext.txt
Citation
24-1.0340473.ris

Full Text

   OPTIMIZATION OF PHOSPHORUS RECOVERY FROM ANAEROBIC DIGESTER SUPERNATANT THROUGH A STRUVITE CRYSTALLIZATION FLUIDIZED BED REACTOR  by   Shayok Ghosh  B.Sc. Eng. (Civil), Bangladesh University of Engineering and Technology, Dhaka, Bangladesh, 2013   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Civil Engineering)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  December, 2016   © Shayok Ghosh, 2016  ii  ABSTRACT Phosphorus is an essential element for all living organisms, but its supply is limited. On the other hand, phosphorus recovery from domestic wastewater can satisfy 15-20 % of current phosphate rock demand. Moreover, struvite scaling is a concern for wastewater engineers as it clogs various equipment. Recovering phosphorus as struvite pellets from wastewater can yield sustainable solution for both problems. Although several technologies have already been available to recover phosphorus from wastewater with reasonable P- recovery efficiency, these technologies possess a number of shortcomings such as higher capital and operating cost, production of fines instead of pellets etc. This study aimed at optimization of phosphorus recovery from wastewater by developing a sustainable and efficient technology. To accomplish this purpose, a new crystallization fluidized bed reactor (FBR) was developed and impact of different physio-chemicals (supersaturation ratio) and hydrodynamic (up-flow velocity, nozzle velocity and configurations) parameters on its performance were analyzed to determine optimum operating conditions. This reactor achieved over 90% of P removal from synthetic supernatant with up to 18% of P recovery. Lower P-recovery was resulted due to lack of proper harvesting mechanism.  Results showed that P-removal efficiency was increased with increase in initial supersaturation ratio up to a value of 6.5. But increase in supersaturation ratio yielded lower P-recovery with higher fines production. A value in the range of 5.5-6.0 was suggested by this study for optimum output. Low up-flow velocity was found to be associated with higher P-removal and recovery efficiency, where high up-flow velocity was found to be associated with the production of more large sized pellets and fines. But, higher nozzle velocity was found to be responsible for accomplishing higher P-removal and recovery efficiency. Two nozzles on opposite side yielded higher P-recovery   iii  efficiency with more large sized pellets and lower fine production. Based on these results, this study concluded that 40 cm/min up-flow velocity with 18.04 cm/min nozzle velocity and two nozzles on opposite side might be optimum operating conditions. Analysis on the performance of up-scaled reactor showed that optimum conditions for pilot scale and up-scaled reactor might be different due to different hydrodynamic conditions.       iv  PREFACE This dissertation is original, unpublished, independent work by the author, Shayok Ghosh.                    v  TABLE OF CONTENTS ABSTRACT .................................................................................................................................... ii PREFACE ...................................................................................................................................... iv TABLE OF CONTENTS ................................................................................................................ v LIST OF TABLES ......................................................................................................................... xi LIST OF FIGURES ...................................................................................................................... xii LIST OF ABBREVIATIONS ..................................................................................................... xvii ACKNOWLEDGEMENTS ......................................................................................................... xix DEDICATION .............................................................................................................................. xx CHAPTER 1: INTRODUCTION ................................................................................................... 1 1.1 Background ........................................................................................................................... 1 1.2 Organization of Thesis .......................................................................................................... 3 CHAPTER 2: LITERATURE REVIEW ........................................................................................ 4 2.1 Phosphorus Facts .................................................................................................................. 4 2.1.1 Phosphorus Cycle .......................................................................................................... 4 2.1.2 Phosphorus Scarcity ....................................................................................................... 5 2.1.3 Phosphorus and Environmental Issues .......................................................................... 6 2.1.4 Issues with WWTP ........................................................................................................ 6 2.2 Potential Sources of Phosphorus Recovery .......................................................................... 8 2.3 Phosphorus Recovery Approaches ....................................................................................... 9 2.3.1 Phosphorus Recovery from Side Stream ....................................................................... 9 2.3.2 Phosphorus Recovery from Sewage Sludge ................................................................ 10 2.4 Phosphorus Recovery as Struvite ........................................................................................ 11 2.5 Crystallization Theory ........................................................................................................ 13   vi  2.5.1 Supersaturation ............................................................................................................ 13 2.5.2 Nucleation .................................................................................................................... 14 2.5.3 Crystal Growth ............................................................................................................. 15 2.5.4 Aggregation ................................................................................................................. 16 2.6 Factors Affecting Struvite Crystallization .......................................................................... 18 2.6.1 pH ................................................................................................................................. 18 2.6.2 Supersaturation Ratio ................................................................................................... 20 2.6.3 Temperature ................................................................................................................. 22 2.6.4 Hydrodynamics and Mixing ........................................................................................ 23 2.6.5 Impurity Ions ................................................................................................................ 24 2.7 Previous Research on Phosphorus Recovery as Struvite .................................................... 24 2.8 Available Technologies ...................................................................................................... 25 2.8.1 Ostara ........................................................................................................................... 26 2.8.2 Multiform Harvest Process .......................................................................................... 27 2.8.3 DHV Crystalactor ........................................................................................................ 29 2.8.4 PHOSNIX .................................................................................................................... 30 2.8.5 AirPrex ......................................................................................................................... 31 2.8.6 PHOSPAQ ................................................................................................................... 33 2.8.7 NuReSeys ..................................................................................................................... 33 2.8.8 Limitations of Technologies ........................................................................................ 34 2.9 Modelling Application ........................................................................................................ 35 CHAPTER 3: THE PROBLEM STATEMENT AND OBJECTIVE ........................................... 38 3.1 Problem Statement .............................................................................................................. 38 3.2 Research Objective ............................................................................................................. 38   vii  3.3 Research Questions ............................................................................................................. 38 3.4 Scope of the Study .............................................................................................................. 39 CHAPTER 4: MATERIALS AND METHODOLOGY .............................................................. 40 4.1 Research Phases .................................................................................................................. 40 4.2 Synthetic Supernatant ......................................................................................................... 41 4.3 Process Layout .................................................................................................................... 41 4.4 Design of the Reactor .......................................................................................................... 43 4.4.1 Injection Port ................................................................................................................ 43 4.4.2 The Reaction Zone ....................................................................................................... 47 4.4.3 Clarifier ........................................................................................................................ 47 4.5 Storage Tanks and Pumps ................................................................................................... 49 4.6 Chemicals ............................................................................................................................ 49 4.7 Process Control, Monitoring and Maintenance .................................................................. 50 4.8 Sample Collection ............................................................................................................... 52 4.9 Analytical Methods ............................................................................................................. 53 4.9.1 Ortho-Phosphate (PO4-P) and Ammonium Nitrogen (NH4-N) ................................... 53 4.9.2 Magnesium and Sodium .............................................................................................. 53 4.9.3 Caustic (NaOH) ........................................................................................................... 54 4.10 Pellets Harvesting and Fines Collection ........................................................................... 55 4.11 Pellets and Fines Quality Determination .......................................................................... 55 4.11.1 Chemical Analysis ..................................................................................................... 55 4.11.2 XRD Analysis ............................................................................................................ 55 4.11.3 Morphology ............................................................................................................... 56 4.12 Terminology ...................................................................................................................... 56   viii  4.12.1 Supersaturation Ratio ................................................................................................. 56 4.12.2 Removal Efficiency ................................................................................................... 57 4.12.3 Struvite Recovery Efficiency ..................................................................................... 57 4.12.4 Coefficient of Variation (COV) ................................................................................. 57 4.12.5 Pellets Growth Rate ................................................................................................... 57 4.12.6 Nominal Weight of Fines ........................................................................................... 57 4.13 Quality Assurance/Quality Control ................................................................................... 58 4.13.1 Quality Assurance ...................................................................................................... 58 4.13.2 Quality Control .......................................................................................................... 58 4.13.2.1 Data Quality Objective (DQO) ............................................................................ 59 CHAPTER 5: RESULTS AND DISCUSSION ............................................................................ 60 5.1 Impact of Physiochemical Variables .................................................................................. 60 5.1.1 pH ................................................................................................................................. 61 5.1.2 Supersaturation Ratio ................................................................................................... 64 5.1.2.1 P-Removal Efficiency ............................................................................................ 65 5.1.2.2 Fine Production ...................................................................................................... 66 5.1.2.3 P-Recovery Efficiency ........................................................................................... 67 5.1.2.4 Pellet Size Distribution .......................................................................................... 69 5.1.2.5 Morphology of Pellets and Fines ........................................................................... 70 5.1.2.6 Purity and Composition of Pellets and Fines ......................................................... 73 5.1.2.7 Loss of Fines Through Effluent ............................................................................. 77 5.1.2.8 Variation of Supersaturation Level Along the Length of the Reactor ................... 78 5.1.3 Optimization of Supersaturation .................................................................................. 78 5.2 Impact of Hydrodynamics Variables .................................................................................. 79   ix  5.2.1 Nozzle and Up-flow Velocities .................................................................................... 80 5.2.1.1 P-Removal Efficiency ............................................................................................ 80 5.2.1.2 P-Recovery Efficiency ........................................................................................... 82 5.2.1.3 Pellet Size Distribution .......................................................................................... 84 5.2.1.4 Fines Production .................................................................................................... 86 5.2.1.5 Morphology of Pellets and Fines ........................................................................... 86 5.2.1.6 Variation of Supersaturation Along the Length of the Reactor ............................. 90 5.2.2 Nozzle configurations .................................................................................................. 91 5.2.3 Optimization of Hydrodynamic Conditions ................................................................ 98 CHAPTER 6: SCALING UP OF THE REACTOR ..................................................................... 99 CHAPTER 7: CONCLUSIONS AND RECOMMENDATIONS .............................................. 107 7.1 Conclusions ....................................................................................................................... 107 7.1.1 Pilot Scale Reactor ..................................................................................................... 107 7.1.2 Up-Scaled Reactor ..................................................................................................... 110 7.2 Recommendations ............................................................................................................. 111 REFERENCES ........................................................................................................................... 112 APPENDIX A: VOLUME CALCULATION ............................................................................ 120 APPENDIX B: NOZZLE VELOCITY CALCULATION ......................................................... 121 APPENDIX C: UPFLOW VELOCITY AND HRT CALCULATION ...................................... 122 APPENDIX D: REYNOLDS NUMBER CALCULATION ...................................................... 124 APPENDIX E: BEAKER TEST ................................................................................................. 125 APPENDIX F: pH METERS ...................................................................................................... 126 APPENDIX G: pH CALIBRATION CURVES ......................................................................... 127 APPENDIX H: PHOSPHORUS CONCENTRATIONS IN UNFILTERED AND FILTERED EFFLUENT SAMPLES DURING FIRST PHASE ................................................................... 128   x  APPENDIX I: AGGREGATION MECHANISM ...................................................................... 129                     xi  LIST OF TABLES Table 2.1: Side reactions involved in struvite precipitation. ........................................................ 12 Table 2.2: Full Scale performance of AirPrex .............................................................................. 32 Table 4.1: Supernatant characteristics. ......................................................................................... 41  Table 4.2: Various nozzle sizes and configurations for supernantant flow. ................................. 46 Table 4.3: Caustic flows corresponding to various runs. .............................................................. 51 Table 4.4: Summary of analytical methods. ................................................................................. 53 Table 4.5: Instrumental parameters for magnesium and sodium analysis. ................................... 54 Table 4.6: Instrumental settings for Bruker D8 Advance XRD. ................................................... 56  Table 5.1: The results of chemical analysis of pellets and fines. .................................................. 73 Table 5.2: Optimum supersaturation level for P-removal and recovery. ...................................... 79 Table 5.3: Summary of parameters during phase 2. ..................................................................... 79 Table 5.4: Optimum hydrodynamic conditions for P-removal and recovery. .............................. 98 Table 6.1: Operating conditions for up-scaled reactor................................................................ 101 Table 6.2: Synthetic supernatant characteristics for up-scaled reactor. ...................................... 101 Table B.1: Nozzle Velocity for different configurations. ........................................................... 121 Table C.1: Upflow velocity and HRT for reaction zone. ............................................................ 122 Table C.2: Up-flow velocity and HRT for clarifier. ................................................................... 123 Table D.1: Reynolds number for different sections. ................................................................... 124 Table F.1: Locations and models of pH meters. ......................................................................... 126   xii  LIST OF FIGURES Figure 2.1: Phosphorus cycle .......................................................................................................... 4 Figure 2.2: Sources of phosphorus as fertilizer. ............................................................................. 6 Figure 2.3:  Recirculation of P in BNR process where in circle represents percent of P loading coming to the plant.......................................................................................................................... 8 Figure 2.4: Phosphorus recovery from side stream. ..................................................................... 10 Figure 2.5: Phosphorus recovery from sewage sludge and sewage sludge ash (wet chemistry technology) ................................................................................................................................... 11 Figure 2.6: Supersaturation as driving force of crystallization  . .................................................. 13 Figure 2.7: Different types of nucleation ...................................................................................... 14 Figure 2.8: Effect of supersaturation on nucleation rate. .............................................................. 15 Figure 2.9: Particle formation through agglomeration. ................................................................ 17 Figure 2.10: Molar removal of Mg2+, NH4+ and PO43- by precipitation of struvite at various pH........................................................................................................................................................ 19 Figure 2.11: Effect of pH on the size of the struvite crystals ....................................................... 20 Figure 2.12: Influence of supersaturation ratio on the induction time of struvite at pH 8.50, 25°C........................................................................................................................................................ 21 Figure 2.13: Influence of supersaturation ratio on the growth rate of struvite at pH 8.50, 25°C. 22 Figure 2.14: Effect of temperature on struvite solubility product. ............................................... 23 Figure 2.15: Ostara technology to recover phosphorus from dewatering liquor  ......................... 26 Figure 2.16: Schematic diagram of Ostara Pearl process   ........................................................... 27 Figure 2.17: Multiform Harvest Process   ..................................................................................... 28   xiii  Figure 2.18: Simplified scheme of DHV-Crystalactor. ................................................................ 30 Figure 2.19: PHOSNIX process. ................................................................................................... 31 Figure 2.20:AirPrex process. ........................................................................................................ 32 Figure 2.21: NuReSys process. ..................................................................................................... 34 Figure 4.1: Different phases of research. ...................................................................................... 40 Figure 4.2: Process Layout. .......................................................................................................... 42 Figure 4.3: Pilot Scale Setup. ........................................................................................................ 43 Figure 4.4:(a) Cross section of injection port, (b) Plan view of injection port. ............................ 45 Figure 4.5: Detail drawing of FBR struvite crystallizer. .............................................................. 48 Figure 5.1: Struvite crystallization rate. ........................................................................................ 61 Figure 5.2: Relationship between influent and effluent pH. ......................................................... 62 Figure 5.3: Impact of effluent pH on P-removal efficiency.......................................................... 63 Figure 5.4: Impact of initial pH on P-removal. ............................................................................. 64 Figure 5.5: Effluent pH Vs Supersaturation ratio. ........................................................................ 65 Figure 5.6: Impact of initial supersaturation ratio on P-removal efficiency. ................................ 66 Figure 5.7: Influence of supersaturation on fine production......................................................... 67 Figure 5.8 Impact of supersaturation ratio on P-recovery efficiency. .......................................... 68 Figure 5.9: Influence of supersaturation ratio on pellets growth rate. .......................................... 69 Figure 5.10: Pellet size distribution against initial supersaturation ratio- (a) Size >1mm, (b) 1mm <Size <0.5 mm and (c) 0.5 mm <Size <0.25 mm. ........................................................................ 70   xiv  Figure 5.11: Microscopic images of pellets (Size >1 mm) under different Supersaturation condition magnified at 100X - (a) Sstruvite-4.8, (b) Sstruvite-5.34, (c) Sstruvite-6.23, (d) Sstruvite-6.40. ................ 71 Figure 5.12: Microscopic images of fines under different Supersaturation condition magnified at 400X - (a) Sstruvite-5.04, (b) Sstruvite-5.49, (c) Sstruvite-7.21. ............................................................. 72 Figure 5.13: X-ray diffraction pattern of struvite pellets (Size: 1 mm to 0.5 mm) for low supersaturation level (Sstruvite- 4.64). ......................................................................................... 75 Figure 5.14: X-ray diffraction pattern of struvite pellets (Size: 1 mm to 0.5 mm) for high supersaturation level (Sstruvite- 7.10). ............................................................................................. 75 Figure 5.15: X-ray diffraction pattern of fines for low supersaturation level (Sstruvite- 5.04). ...... 76 Figure 5.16: X-ray diffraction pattern of fines for high supersaturation level (Sstruvite- 7.21). ..... 76 Figure 5.17: Differences in P-concentrations in unfiltered and filtered effluent samples under different supersaturation levels. .................................................................................................... 77 Figure 5.18: Variation of Sstruvite along the length of the reactor. ................................................. 78 Figure 5.19: Influence of up-flow and nozzle velocities on P-removal efficiency - (a)scatter diagram and (b) bar diagram. ........................................................................................................ 81 Figure 5.20: Influence of up-flow and nozzle velocities on P-recovery. ...................................... 82 Figure 5.21: Differences in P-concentrations in unfiltered and filtered effluent samples under different up-flow and nozzle velocities. ........................................................................................ 83 Figure 5.22: Pellets growth rate under different nozzle and up-flow velocities. .......................... 84 Figure 5.23: Pellet size distribution under different up-flow and nozzle velocities. .................... 85 Figure 5.24: Production of fines under various up-flow and nozzle velocities. ........................... 86 Figure 5.25: Microscopic images of pellets magnified at 400X under different nozzle velocity-(a) high nozzle velocity with size > 1mm , (b) high nozzle velocity with 0.5 mm <size< 1.0 mm , (c) low nozzle velocity with size > 1mm , (d) low nozzle velocity with 0.5 mm <size < 1mm. ....... 87   xv  Figure 5.26: Microscopic images of pellets magnified at 400X under different up-flow velocity-(a) high up-flow velocity with size > 1mm, (b) high up-flow velocity with 0.5 mm <size< 1.0 mm, (c) low up-flow velocity with size > 1mm, (d) low up-flow velocity with 0.5 mm <size < 1mm........................................................................................................................................................ 88 Figure 5.27: Microscopic images of fines magnified at 400X under different nozzle velocity. .. 89 Figure 5.28: Microscopic images of fines magnified at 400X under different up-flow velocity. 89 Figure 5.29: Variation of supersaturation level along the length of the reactor under different up-flow and nozzle velocities. ............................................................................................................ 90 Figure 5.30: Impact of nozzle configuration on P-removal efficiency. ........................................ 91 Figure 5.31: P-recovery efficiency and pellet size distribution under different nozzle configurations. .............................................................................................................................. 92 Figure 5.32: Pellets growth rate under various nozzle configurations. ......................................... 92 Figure 5.33: Influence of nozzle configurations on fines production. .......................................... 93 Figure 5.34: Differences in P-concentrations in unfiltered and filtered effluent samples under different nozzle configurations. .................................................................................................... 94 Figure 5.35: Microscopic images of pellets (size > 1mm) magnified at 100X. ........................... 95 Figure 5.36: Eddies formed under different nozzle configurations. ............................................. 96 Figure 5.37: Microscopic images of fines magnified at 400X under different nozzle configurations........................................................................................................................................................ 97 Figure 5.38: Variation of supersaturation level along the length under various nozzle configurations. .............................................................................................................................. 97  Figure 6.1: Details drawing of FBR struvite crystallizer. ........................................................... 100 Figure 6.2: P-removal efficiency of up-scaled reactor. ............................................................... 102 Figure 6.3: P-recovery efficiency for up-scaled reactor. ............................................................ 104   xvi  Figure 6.4: Microscopic images of pellets (magnified at 100X) and fines (magnified at 400X)...................................................................................................................................................... 105 Figure 6.5: Difference in P-concentration between unfiltered and filtered effluent samples from up-scaled reactor. ........................................................................................................................ 106 Figure 6.6: Initial and effluent supersaturation ratio for up-scaled reactor. ............................... 106 Figure E.1:Beaker test result from run 7. .................................................................................... 125 Figure G.1: pH calibration curves. .............................................................................................. 127 Figure H.1: P-concentration in effluent (filtered and unfiltered) samples under different supersaturation level. .................................................................................................................. 128 Figure I.1: Aggregation mechanism of two approaching crystals. ............................................. 130     xvii  LIST OF ABBREVIATIONS ATP Adenosine Triphosphate BC British Columbia BNR Biological Nutrient Removal BPR Biological Phosphorus Removal cm Centimetre EBPR Enhanced Biological Phosphorus Removal EC Electro Conductivity FBR Fluidized Bed Reactor gm Grams HRT Hydraulic Retention Time kg Kilograms L Litre m metre   xviii  mg Milligram MGD Million Gallons Per Day ml Millilitre mm Millimetre min Minute Sstruvite Supersaturation Ratio TSS Total Suspended Solids WWTP Wastewater Treatment Plant XRD X-ray Diffraction        xix  ACKNOWLEDGEMENTS First of all, I would like to thank and convey my deepest gratitude to my research supervisor Dr. Victor Lo, Department of Civil Engineering for his continuous encouragement, understanding and firm support throughout my years at UBC.  I want to thank Dr. Sergey Lobanov for his continuous support, guidance, supervision and also for teaching me with basic knowledge on P-recovery technologies. I would like to thank Paula Parkinson and Timothy Ma for helping me with analytical works, placing equipment orders and for always attending me with a smile even at their busiest time.  Last, but not the least, I want to thank all my lab mates, friends, family, specially my parents and my wife for their endless support and encouragement. Above all, this research was possible only by the grace of GOD.      xx  DEDICATION To my parents and my wife who gave me their unconditional supports and guidance.                   1  CHAPTER 1: INTRODUCTION 1.1 Background  Phosphorus is a vital element for human body and other living organisms as it exists as phosphate in ATP (adenosine triphosphate) to produce energy. That is why phosphate is a critical ingredient in growing foods. As a result, 90% of current global demand for phosphorus is due to food production  (Cordell, 2016). Phosphate rock is the major source for extraction of phosphorus in current world (Cordell et al., 2009). But because of lower quality of current phosphorus rock, seeking for a sustainable source of phosphorus has become a point of great interest in both the agricultural and industrial sectors. On the other hand, a large portion of phosphorus are being lost in wastewater (municipal and agricultural) which offers an environmental engineering challenge of removing phosphorus from wastewaters in order to prevent eutrophication resulting in scarcity of phosphorus (Ashley et al., 2011).  Technologies have been developed to remove phosphorus from wastewater since 1950s (Huang, 2003). Typical P removal technologies are based on the principal of fixation of phosphorus in activated sludge either by a biological method or a chemical precipitation by metal salts (Le Corre et al., 2009). Although these processes achieve higher P removal, but they are associated with several disadvantages including accumulation of phosphorus in sludge, larger sludge volumes and formation and precipitation of struvite (Le Corre et al., 2009). Among these, struvite scaling is considered as the most serious problem since they are hard to dislodge. As a result, struvite scaling reduces the capacity of treatment plant (Huang et al., 2006).   2  Under these circumstances, considering wastewater as a resource where P can be recovered and reused rather than disposing off is now being accepted as the best solution to above mentioned problems (Huang et al., 2006).  A sustainable approach for P recovery is to recover phosphorus from nutrient enriched stream in the form of struvite, which will also eventually provide solution to the struvite scaling problem. Several studies have been demonstrated to evaluate the potential technologies of P recovery in the form of struvite at bench and pilot scales, and only a few has been incorporated in full scale system (Huang, 2003; Le Corre et al., 2009).  Environmental Engineering research group at The University of British Columbia (UBC) has developed an up-flow fluidized bed reactor for struvite crystallization to recover phosphorus from municipal as well as agricultural streams (Lobanov et al., 2014; Mavinic et al., 2007). Although this technology has already been commercialized by a spin off company of UBC and it gained a reasonable success, this process has a number of shortcomings like longer crystallization time (CRT), high recirculation ratio, high up-flow velocity etc. These shortcomings have made this technology energy intensive along with higher operating and maintenance cost. Other technologies such as Multiform Harvest Process, PHOSNIX, AirPrex PHOSPAQ etc. are also associated with various shortcomings such as high capital and operating cost, high energy consumption etc.  This study focussed on development of more energy efficient and financially viable (lower capital, operating and maintenance cost) technology to recover P from wastewater. In order to do so, efforts were taken to build a new struvite crystallization reactor along with the determination of its operating parameters in such a way that it would yield efficient and sustainable P recovery from anaerobic digester supernatants as struvite pellets. Synthetic supernatant was used to analyze performance of the reactor under different operation condition as a pilot scale operation.    3  1.2 Organization of Thesis  This thesis consists of five chapters. Chapter 1 gives a summary on the basic problems related to P recovery from wastewater which had to be dealt with in this research. Chapter 2 provides a literature review on the phosphorus problems associated with wastewater, different technologies used to remove and recover P from wastewater and basic crystallization theory. This chapter also provides definition of various relevant terms. Chapter 3 outlines the problem statement along with objectives and research questions. Chapter 4 describes the materials and methodologies that have been used in this research. Configuration of the new FBR, characteristics of synthetic supernatants, data quality objectives, sampling methods and descriptions of analytical methods used to determine concentration of different constituents of samples are presented in this chapter. Chapter 5 outlines the results, data analysis and relevant discussion.  Chapter 6 deals with the analysis of performance of reactor due to upscaling.  Chapter 7 contains summary of the findings of the research and relevant recommendations.     4  CHAPTER 2: LITERATURE REVIEW 2.1 Phosphorus Facts Phosphorus exists as the eleventh most abundant element on Earth (Le Corre et al., 2009). Phosphorus is a vital nutrient for both animals and plants. As mentioned in Chapter 1, phosphate has a significant role to produce energy as one of components of ATP. Phosphate is also an important constituent of DNA and RNA for the purpose of protein synthesis. Phosphorus plays a vital role to build certain components of the human and animal body like the bones and teeth (Roberts, 2016).  2.1.1 Phosphorus Cycle Phosphorus in nature passes through rocks, water, soil and organisms in a cyclic order (as shown in Figure 2.1). Rain and weathering induce rock to release phosphate in soil and water. Plants uptake phosphorus from soil and also from artificial and chemical sources (fertilizer). Then this phosphorus is transformed to animals by means of consuming plants. Death or decay of the plants and animals release phosphorus back to soil (Punmia et al., 1998).   Figure 2.1: Phosphorus cycle (adapted from Punmia et al., 1998).   5  2.1.2 Phosphorus Scarcity The demand of phosphorus for crop production was satisfied from natural level of soil phosphorus and other organic matter such as manure and human excreta before the 20th century. In order to keep pace with increased food demand from 20th century, guano and later rock phosphate was applied in abundance to food crops. Figure 2.2 outlines evolution of phosphorus fertilizer from 1800 to 2000 (Cordell et al., 2009).  The demand of phosphorus is expected to increase by 50-100% by 2050 (Cordell et al., 2009). It is very difficult to predict the timeline for the depletion of world’s phosphorus reserve. Some studies show that current phosphorus reserve can be exhausted from 100 to 250 years (Johnston and Steén, 2000). Moreover, depletion of phosphorus reserve is not the only concern; deteriorated quality of phosphorus rock is also becoming a great issue. Increasing metallic contents like cadmium, uranium, nickel, chromium, copper and zinc in rocks is making the end products unacceptable for fertilizer (Driver et al., 1999).   6   Figure 2.2: Sources of phosphorus as fertilizer (adapted from Cordell et al., 2009). 2.1.3 Phosphorus and Environmental Issues Because of being a limiting nutrient in freshwater bodies, phosphorus is closely associated with eutrophication which causes algal bloom. Algal bloom is responsible for reduction of light penetration and available oxygen in the water body. That’s why stringent regulations of phosphorus discharge are applied for point sources such as municipal sewage discharge (Huang, 2003). As a result, approximately 48 percent of European population is subject to phosphorus removal requirement since 1991 (Woods et al., 1999). 2.1.4 Issues with WWTP In order to achieve low effluent P concentration, chemical and biological P removal is increasingly being implemented around the world. Both these processes are able to achieve effluent TP levels of 0.5 mg/L. Even this level can be dropped further if tertiary filtration is applied (Oleszkiewicz, 2015) . However, there are several issues related to these P removal processes.   7   Traditional phosphorus removal technologies are very energy intensive and/or they also need large amounts of chemical which may result in emission of greenhouse gas (Oleszkiewicz, 2015).  During biological P removal (BPR) or enhanced biological P removal (EBPR) phosphorus is supposed to be re-released during sludge digestion. Sludge digestion, especially anaerobic digestion, causes re-release of a large portion (about 60%) of the stored polyphosphate (Phillips et al., 2006; Pitman et al., 1991).  Battistoni et al. (1997) suggested that sludge liquors might have phosphorus level up to 100 mg/L under anaerobic condition and presence of readily biodegradable COD.  This released P is returned to plant inlet for recycling, which results in increase in phosphorus load in the treatment plant as shown in Figure 2.3. As a result, most of the phosphorus is recirculated rather than being removed. Recycled phosphorus load can be 40% to 100% of the incoming phosphorus (Jardin and Pöpel, 1994; Pitman et al., 1991). This additional phosphorus load can project a negative impact on the performance of treatment plant, as efficiency of BNR process depends on BOD5: P ratio.   Both biological and chemical P removal are associated with high sludge production. Paul et al. (2016) reported that the amount of excess sludge production for EBPR and chemical P removal processes are 3 gTS/g P and 5 to 7 gTS/g P respectively. Operating cost for EBPR process mainly comes from sludge production (Paul et al., 2016). Struvite (MgNH4PO46H2O) accumulation on treatment equipment surfaces of anaerobic digestion and post digestion is considered to be a great issue for WWTPs worldwide (Chirmuley, 1994). Struvite deposits are hard to dislodge. As a result, they cause plugging of sludge pipes, pumps and centrifuges which lead to the reduction of plant capacity (Mavinic et al., 2007). Thereby, struvite scaling is associated with high maintenance cost and sometimes enforces costly shut down. Doyle   8  and Parsons (2002) reported that for mid sized treatment plant (25 MGD) annual cost for remediation of struvite scaling might exceed US$100,000. During anaerobic digestion, ammonium, phosphate and magnesium are released from sludge because of sludge digestion. Under favorable condition, they are combined to form struvite. As a BNR process removes phosphorus, higher concentrations of phosphorus, ammonium and magnesium are released in the wasted sludge. Thus, possibility of  formation of struvite is higher for WWTPs associated with BNR (Huang, 2003).   2.2 Potential Sources of Phosphorus Recovery Human and animal wastes are considered to be potential major sources for phosphorus recovery. Humans excrete generates annually 3.3 million tons of phosphorus around the world (Cordell et Figure 2.3:  Recirculation of P in BNR process where in circle represents percent of P loading coming to the plant (based on Oleszkiewicz, 2015).   9  al., 2009). Oleszkiewicz (2015) projected that 43,000 tons of phosphorus would be discharged annually in municipal wastewater as Canada’s population was approaching 36 millions. Some study revealed that 15-20% of world phosphate rock demand can be satisfied by recycling phosphorus from municipal wastewater (Cordell et al., 2009; Mihelcic et al., 2011).  Nutrient contents of animal wastes are relatively higher when compared to human waste. About 15 million tonnes of phosphorus are produced from livestock around the world (Oleszkiewicz, 2015). In 2006, Canadian livestock’s wastes contained about 300 thousand tonnes of phosphorus, which showed a 21% increase from the 1981 values (Statistics Canada, 2015). However, lack of proper collection system has limited the phosphorus recovery from animal manure (Huang, 2003).  2.3 Phosphorus Recovery Approaches Phosphorus can be recovered from side stream (supernatant liquor of anaerobic digestion) or from sewage sludge. 40 to 45% of incoming phosphorus load with raw wastewater can be recovered from side stream. On the other hand, recovery from sewage sludge or sewage ash can be up to 90% (Cornel and Schaum, 2009). 2.3.1 Phosphorus Recovery from Side Stream   The side stream, which is enriched with phosphorus is fed into a crystallization reactor where calcium or magnesium salt and seed crystal, if needed, are added as shown in Figure 2.4. As a result phosphorus is recovered as calcium phosphate and magnesium ammonium phosphate (MAP) (Cornel and Schaum, 2009).     10   Figure 2.4: Phosphorus recovery from side stream (based on Cornel and Schaum, 2009). 2.3.2 Phosphorus Recovery from Sewage Sludge  There exist two approaches to recover phosphorus from sewage sludge – wet chemical technology and thermochemical technology. During wet chemical technology, phosphorus in sewage sludge is dissolved by the addition of acid or base and also by changing temperature if required. After removal of insoluble compounds, phosphate can be isolated from phosphorus rich water using precipitation, ion exchange, reactive liquid-liquid extraction or nanofiltration as shown in Figure 2.5. This technology is also applicable for sludge ash (Cornel and Schaum, 2009). The thermochemical approach involves exposure of sludge ashes to chlorine containing compounds (KCl or MgCl2) and thermochemical treatment. At temperature greater than 1000°C, larger portion of heavy metals is converted into heavy metal chlorides which vaporize. The heavy metals are entrapped using flue gas treatment (Cornel and Schaum, 2009; Jardin and Pöpel, 1994).   11  40 to 45% of incoming phosphorus load with raw wastewater can be recovered from side stream. On the other hand, recovery from sludge ash can yield up to 90%. But, because of using strong chemicals recovery form sludge is proven to be uneconomical (Oleszkiewicz, 2015).   Figure 2.5: Phosphorus recovery from sewage sludge and sewage sludge ash (wet chemistry technology) (based on Cornel and Schaum, 2009). 2.4 Phosphorus Recovery as Struvite  Struvite is a white crystalline substance, which is sparingly soluble in water. It consists of equal molar concentration of magnesium, ammonium and phosphate. Struvite forms following the general equation as shown in Equation (2.1) with n = 0-2 (Le Corre et al., 2009).         𝑀𝑔2+ + 𝑁𝐻4+ + 𝐻𝑛𝑃𝑂43−𝑛 + 6𝐻2𝑂 ↔  𝑀𝑔𝑁𝐻4𝑃𝑂4. 6𝐻2𝑂 + 𝑛𝐻+                               (2.1)  Moreover, some side reactions are also involved with the precipitation of struvite which are depicted in Table 2.1(Doyle and Parsons, 2002).    12  Table 2.1: Side reactions involved in struvite precipitation. 𝑁𝐻4+ ↔ 𝑁𝐻3 + 𝐻+ 𝑀𝑔𝑂𝐻+ ↔ 𝑀𝑔2+ +  𝑂𝐻−  𝐻3𝑃𝑂4 ↔ 𝐻+ + 𝐻2𝑃𝑂4− 𝐻2𝑃𝑂4− ↔ 𝐻+ + 𝐻𝑃𝑂42− 𝐻𝑃𝑂42− ↔ 𝐻+ + 𝑃𝑂43− 𝑀𝑔𝐻2𝑃𝑂4+ ↔ 𝑀𝑔2+ +  𝐻2𝑃𝑂4−  𝑀𝑔𝐻𝑃𝑂4 ↔ 𝑀𝑔2+ +  𝐻𝑃𝑂42−  𝑀𝑔𝑃𝑂4− ↔ 𝑀𝑔2+ +  𝑃𝑂43−  𝐻2𝑂 ↔ 𝑂𝐻− + 𝐻+ Because of its composition and lower solubility, it can be used as a slow release fertilizer. A phosphorus recovery technique from supernatant of anaerobic digestion in the form of struvite can yield solution to problems mentioned in section 2.1.4 as well as can serve as a potential source of phosphorus fertilizer. Shu et al. (2006) estimated that 0.63 million tons of phosphorus could be harvested as struvite annually from WWTPs around the world which would reduce phosphate rock mining by 1.6%. A pilot scale study at Goldbar WWTP, Edmonton, Alberta showed that struvite recovery would result in 20% reduction of phosphorus load on the treatment plant (Britton et al., 2007).    13  2.5 Crystallization Theory Crystallization can be defined as a solid liquid separation process where conversion of mass of a solute from liquid phase to solid crystals occurs.  Crystallization can be reassembled with precipitation process, which results in formation of crystals. This section summarizes the conception of basic crystallization theory regarding the formation of struvite crystals.  2.5.1 Supersaturation  The term “Supersaturation” is applied to a solution when concentration of a solute is greater than its equilibrium concentration. Precipitation takes place as a result of supersaturation. Supersaturation is described as the driving force of crystallization as shown in Figure 2.6.   Figure 2.6: Supersaturation as driving force of crystallization (Adapted from Sohnel and Garside, 1992). Supersaturation can be expressed in several ways. Some of these are (Mullin, 1993) - Concentration driving force, ∆c = c-c* Supersaturation Ratio, S = 𝐶𝐶∗   14  Relative supersaturation ratio, σ = ∆c𝐶∗ Where, c indicates concentration at supersaturation, and c* indicates concentration at equilibrium. 2.5.2 Nucleation Nucleation is the process of formation of new solid phase.  It takes place when molecules of solute come closer in clusters and grow by accumulation to produce significant amount of a new phase. When nucleation in the system does not get influenced by crystalline matter itself, the nucleation is described as primary nucleation. On the other hand, if the nuclei are formed under the presence same solid phase, then this behavior is termed as secondary nucleation (Mullin, 1993). Figure 2.7 depicts different form of nucleation process.  Figure 2.7: Different types of nucleation (based on Mullin, 2001). The rate of nucleation, J is a function of supersaturation ratio, temperature and interfacial tension. Mullin (2001) suggested a general expression of nucleation rate as shown in Equation (2.2).                                                       𝐽 = 𝐴 exp [−16𝜋𝛾3𝜈23𝑘3𝑇3(ln 𝑆)2]                                                            (2.2) Where, k is the Boltzmann constant, T is temperature, S is the supersaturation ratio, A is pre-exponential factor, ν is molecular volume and γ is surface tension. A plot of Equation (2.2) is NucleationPrimaryHomogoeneous(not influnced by any solid phase)Heterogenous (induced by foreign solid phase)Secondary (induced by solid phase itself)  15  shown in Figure 2.8. From this plotting, it is apparent that higher supersaturation increases nucleation rate and theoretically a little increase in S induces larger increase in J.  Figure 2.8: Effect of supersaturation on nucleation rate (based on Mullin, 2001). Induction time is another important conception which is used to determine J. Induction time is defined as the time laps between the achievement of supersaturation and appearance of a crystal (Galbraith and Schneider, 2009).                                                                         𝑡𝑖𝑛𝑑 =  1𝐽                                                                    (2.3)           2.5.3 Crystal Growth Crystal growth is the process through which crystal embryos grow in detectable size or beyond. This process is controlled by diffusion and surface integration (Jones, 2002; Le Corre et al., 2009). This process consists of two major steps (Jones, 2002)-  Transformation of solutes from the solution to the surface of crystals, by diffusion or convection mechanism, or combination of both.   16   Incorporation of material into the crystal lattice through the mechanism of surface integration.  As crystal growth rate has a complex relationship with temperature, supersaturation, size habit, system turbulence etc., there exists no simple or generally accepted expression of crystal growth rate (Mullin, 1993).  That’s why crystal growth rate is usually correlated with environmental parameters like concentration and temperature using following power law model (Jones, 2002; Le Corre et al., 2009).                                                                 𝐺 = 𝑘𝑔𝜎𝑔                                                                                    (2.4) Where, G is the growth rate, kg represents growth constant, and σ represents the relative supersaturation and g is apparent reaction order. For diffusion-controlled growth g =1, and for surface integration-controlled growth g=1–2 or >2.  Along with nucleation, crystal growth is also an important process as this process determines the final size and the shape of crystals which are considered to be vital parameters in struvite recovery process (Le Corre et al., 2009). Abbona and Boistelle (1979) reported that flat crystals were formed with high-growth kinetics, while stick-like crystals were produced with low-growth kinetics. 2.5.4 Aggregation Aggregation is the process of clustering of separate particles to form larger particles. This process is sometimes termed as also “agglomeration”, “coagulation” and “flocculation” without having generally accepted strict definition (Mullin, 1993). This paper will term this phenomenon as “aggregation” or “agglomeration”.   17  Agglomeration is the mechanism of clustering primary particles to from strong secondary particles held together mainly by crystalline bridges (Jones, 2002). Agglomeration can be of two types - primary and secondary agglomeration. Primary agglomeration takes place when a crystalline particle goes through a mal-growth, which is related to its crystallography and constitute individual crystals within parallel units, dendrite or twin. On the other hand, collision of particles may take place depending on flow and combine together to form larger particle. These particles may undergo disruption and re-dispersion to produce a secondary agglomerate (Jones, 2002). Figure 2.9 represents the overall agglomeration process.                     Figure 2.9: Particle formation through agglomeration (based on Jones, 2002). Aggregation constant which is proportional to aggregation rate can be expressed as following-  (Linnikov, 2008)                                                          𝐾𝑎 =  𝑘1𝐷3𝑒−𝐷1𝑙𝑛𝐶𝐶𝑜+𝐷2                                                         (2.5) A detail description of the terms used in this equation is depicted in Appendix  I. Nuclei bridge   18  2.6 Factors Affecting Struvite Crystallization Struvite nucleation and growth have complex co-relation with several physio-chemical parameters such as pH, supersaturation ratio, mixing energy, temperature and presence of foreign ions (Le Corre et al., 2009).        2.6.1 pH   pH is considered to be one of the most significant factors of struvite precipitation as it is closely related with solubility. One of the reasons for struvite deposition in WWTPs is due to increase in pH which is associated with CO2 stripping (𝐻𝐶𝑂3−  → 𝐶𝑂2 ↑ +𝑂𝐻−) (Neethling and Benisch, 2004).  As per Equation (2.1), the precipitation of struvite involves a release of protons in solution, which lowers the pH. The rate of change of pH indicates the formation rate of struvite and also exerts significant impact on the quality of crystalline particles (Le Corre et al., 2009).  Moreover, solubility of struvite is highly dependent on pH. Struvite solubility decreases with increase in pH within the pH range of wastewater treatment (6.0-8.0) (Ohlinger et al., 1998). Buchanan et al. (1994) identified pH of minimum solubility as 9.0 where Ohlinger et al. (1998) reported the pH of minimum solubility as 10.3. The identified pH value associated with minimum solubility ranged from 8 to 10.7. The difference in reported values are due to the selection of different Ksp values (Ohlinger et al., 1998). Booker et al.(1999) by analyzing the molar removal of NH4+ ,Mg2+ and PO43- due to struvite crystallisation from starting solutions having an equimolar quantity of both N and P, and an excess of Mg (about 7%) demonstrated that the removal of the different species was maximized within the pH range 8.8 to about 9.4 as depicted in Figure 2.10.   19   Figure 2.10: Molar removal of Mg2+, NH4+ and PO43- by precipitation of struvite at various pH (based on Booker et al., 1999). Initial pH is also important as increase in initial pH leads to increase in supersaturation which yields increased nucleation and growth rate (Le Corre, 2006). High initial pH can also induce the transformation of ammonium ions into gaseous ammonia which results in the reduction of ammonium ion concentration, thus exerting effect on the molar ratio of magnesium , nitrogen and phosphorus (Le Corre et al., 2009).  pH also has impact on the size of the struvite crystals. Matynia et al. (2006) manifested that an increase in pH of reaction environment from 8 to 11 might lead to five times reduction in mean crystal size along with increase in growth rate as represented in Figure 2.11. 01234566.5 7.5 8.5 9.5 10.5 11.5Moles RemovedpHNH4 -N Mg PO4-PNH4- PO4-P  20   Figure 2.11: Effect of pH on the size of the struvite crystals (Adapted from Matynia et al., 2006). 2.6.2 Supersaturation Ratio As mentioned earlier, supersaturation is considered to be the driving force of crystallization. Supersaturation can have influence on both nucleation and crystal growth rate.  Bouropoulos and Koutsoukos (2000) reported that at constant pH, induction time was inversely related with supersaturation of struvite (as shown in Figure 2.12), which actually indicated that higher supersaturation induced higher nucleation rate e.g. higher supersaturation shortened the time of crystal appearance. They demonstrated that an increase in Sstruvite from 1.042 to 1.50 induced 20-times reduction in induction time. Kofina and Koutsoukos (2005)  demonstrated same results by applying same methodology but a different solvent (synthetic wastewater which was made of glucose, NaHCO3, NaCl, NaNO2, and Na2SO4 rather than de-ionized water) over the range of Sstruvite from 1.27 to 1.62.     21   Figure 2.12: Influence of supersaturation ratio on the induction time of struvite at pH 8.50, 25°C. Bouropoulos and Koutsoukos (2000) also demonstrated that struvite growth rate adopted a power law Rp = kpσn , where n was the apparent order of the reaction and kp was the precipitation constant. They reported that an increment of Sstruvite from 1.042 to 1.50 caused 55-fold increase in crystal growth rate as shown in Figure 2.13. But , Kofina and Koutsoukos (2005)  demonstrated higher crystal growth rate compared to Bouropoulos and Koutsoukos (2000)  as depicted in Figure 2.13 which confirmed the effect of nature of solution in which struvite formed.                                                                                                                                                    0204060801001201401 1.1 1.2 1.3 1.4 1.5 1.6 1.7Induction time (min)SstruviteBouropoulos and Koutsoukos (2000) Kofina and Koutsoukos (2005)  22   Figure 2.13: Influence of supersaturation ratio on the growth rate of struvite at pH 8.50, 25°C. 2.6.3 Temperature Temperature has lower implication on struvite crystallization compared to pH and supersaturation (Durrant et al.,1999). However, temperature can influence solubility and crystal configuration. Aage et al. (1997) demonstrated the impact of temperature on solubility product of struvite as presented in Figure 2.14, which implied that increase in temperature induced increment in solubility of struvite from 10°C to 50°C. Aage et al. (1997) also reported alteration of structure of struvite in the range of 64°C to 67°C,  which might exert inevitable change in solubility of struvite. Temperature is also considered to have an impact on crystal growth, as it is associated with relative rates of diffusion and surface integration. High temperatures generally induce diffusion-controlled growths. On the other hand, low temperatures cause surface integration controlled growths (Jones, 2002). High temperature can also project impact on the size, shape and type of crystal. Boistelle 0501001502002503003504004505001 1.1 1.2 1.3 1.4 1.5 1.6 1.7Growth rate (X108mol/min)SstruviteBouropoulos and Koutsoukos (2000) Kofina and Koutsoukos (2005)  23  et al. (1983) demonstrated that rectangular and prismatic crystals were formed at 25°C, while square and thick crystals were formed at 37°C.  Figure 2.14: Effect of temperature on struvite solubility product (based on Aage et al., 1997). 2.6.4 Hydrodynamics and Mixing Fluid hydrodynamics plays an important role in crystallizer as it is associated with local solids and ionic concentrations, mixing energies and inter-particle collision rate (Fattah et al., 2012). Mixing energy is the primary controlling factor for crystal growth (Ohlinger et al., 1999). Energy inputs stimulate the disruption of concentration gradients in boundary layers causing increment in crystal growth rate (Wang et al., 2006). Bhuiyan (2007) showed that induction time could be lowered by increasing mixing energy when supersaturation is constant. On the other hand, up-flow velocity plays an important role to recover phosphorus in FBR (Bhuiyan et al., 2008). Up-flow velocity can be determined by dividing total up-flow by cross-sectional area of FBR (Ye and Lou, 2016). High up-flow velocity induces strong inter-particle collisions resulted in destruction of shape and reduction in crushing strength. That’s why Fattah et al. (2012) established a limit of maximum up-flow velocity to produce better quality of struvite pellets in FBR. 05101520253035400 10 20 30 40 50 60Solubility Product (Ksp X 1014)Temperature (°C)  24  Ohlinger et al. (1999) demonstrated that mixing energy had impacts on the size and shape of the crystals as in areas of semi-quiescent zones where more elongated crystals were formed in compared with high mixing zone. During this study about 21 times more growth rate was observed in high mixing environments compared to the quiescent zone. 2.6.5 Impurity Ions Foreign ions can block the sites where crystals can form which results in inhibition growing of crystalline particles in size (Jones, 2002). Le Corre et al. (2005) manifested that molar ratio Mg:Ca 1:1 or higher could cause inhibition of formation of struvite by forming amorphous calcium phosphate. Presence of foreign ions can also affect ionic strength of solutions because of reduction in effective concentrations due to electrostatic interactions of ions. 2.7 Previous Research on Phosphorus Recovery as Struvite  Extensive studies have been carried out on struvite (MAP) recovery on pilot and full-scale basis. Most of these studies yield satisfactory performance in terms of P removal efficiency. This section summarized the performances and operating conditions of some of these studies. Abe (1995) reported that full-scale struvite recovery process at Fukuoka City accomplished 80% of phosphate removal efficiency with high initial phosphate concentration of 245 mg/L with the production of pellet size greater than 2 mm.  This reactor utilized NaOH to achieve pH value of 8.2. Another full-scale MAP process at Shimane Prefecture Lakes Shinji East Clean Centre in Japan yielded 90% P removal efficiency with a capacity of 500 m3/day. This plant was operated at pH from 8.2 to 8.8 with influent ortho-P concentration of 100-110 mg/L. This study used 10 days of retention time to harvest 0.5 to 1.0 mm struvite crystals. A combination of NaOH and CO2 air stripping was applied to maintain desired pH level (Ueno and Fujii, 2001). A study  by Munch   25  and Barr (2001) demonstrated P-recovery from centrate of anaerobically digested sludge in Oxley Creek WWTP in Brisbane using an air agitated column reactor with a reaction and settling zone based on the PHOSNIX process of Unitika Ltd., Japan. This study accomplished 94% ortho-P removal with influent concentration 61 mg/L. Mg(OH)2 was applied to maintain pH 8.5 and also as a source of Mg.  In another study, low-cost Mg2+ solution separated from seawater by nanofiltration was utilised a source magnesium for MAP process. This study also accomplished up to 95% of P-removal even with high initial phosphorus concentration (about 300 mg/L). But, most of the struvite particles harvested during this study was lower than 250μm. The authors identified short duration of the experiments (18 hours to 36 hours ) as a probable reason for this small size (Lahav et al., 2013).  Some studies on the fluidized bed reactor developed by the University of British Columbia (UBC) revealed that above 90% P –removal was achieved with effluent P concentration of 5 mg/l or less (Fattah et al., 2008). Fattah et al.(2008) also demonstrated that 80% or more phosphorus removal could be accomplished at a pH of 7.5 or less. This study also manifested that increase in supersaturation ratio increased P-removal up to an optimal value of 1.41. Fattah et al.(2008) also accomplished about 86% struvite recovery efficiency. Another study on same reactor revealed that 400 cm/min up-flow velocity was reported to be more efficient to produce good quality of struvite pellets compared to 500 cm/min as the high velocity induced more collisions between particles resulted in more broken particles (Fattah et al., 2012). 2.8 Available Technologies Ostara, Multiform Harvest, PHOSNIX, Crystalactor, AirPrex and Phospaq are widely used existing struvite recovery technologies (Oleszkiewicz, 2015). This section described the highlights of these technologies.    26  2.8.1 Ostara  Ostara process is primarily applied to recover phosphorus from dewatering liquor and centrate after anaerobic digestion as shown in Figure 2.15.  PRIMARY SEDIMENTATIONInfluentSECONDARY SEDIMENTATIONTREATED EFFLUENTPRIMARY SOLIDSSECONDARY SOLIDSANAEROBIC DIGESTIONDEWATERINGBIOSOLIDSINFLUENTCRYSTAL GREEN FERTILIZER PRODUCTOSTARA PEARL PROCESSRETURN TO PLANT THROUGH CLEAN PIPES Figure 2.15: Ostara technology to recover phosphorus from dewatering liquor (adapted from Howorth and Wirtel, 2015). Ostara technology utilizes up-flow fluidized bed reactor for struvite crystallization through a combination of chemical dosing, fluidization energy, and loading rate control. Since magnesium is the limiting precipitant in digested liquor, magnesium chloride is added in this technology. pH is controlled by dose of sodium hydro-oxide (Howorth and Wirtel, 2015). A schematic diagram of Ostara reactor is presented in Figure 2.16.   27   Figure 2.16: Schematic diagram of Ostara Pearl process (adapted from Britton et al., 2009; Oleszkiewicz, 2015)  This technology is now currently in full scale operation in seven locations around the world (Oleszkiewicz, 2015). Phosphorus and ammonia recovery using this technology was reported to achieve 80-90% and 14-42% respectively (Baur et al., 2011; Britton et al., 2009; Oleszkiewicz, 2015). Benisch et al. (2009) reported that at Durham Advanced WWTP Ostara process achieved recovery of 20% of the plant initial phosphorus load and 1.5% of the influent nitrogen load along with the production of 1.1 tons per day of dried fertilizer. The target pH for this treatment plant was 7.25 with a design P-removal efficiency of 85% which would result in 40 mg P/L of effluent orthophosphate (OSTARA, 2009). The WASSTRIP process was incorporated in this treatment plant.  2.8.2 Multiform Harvest Process Multiform harvest process is primarily developed to treat agricultural waste stream. This technology aims at overall phosphorus removal with lower operating cost rather than the purity and appearance of struvite. This reactor consists of a conical shaped fluidized bed with no recycle   28  flow. Nutrient rich stream is injected at the bottom of the reactor. Struvite crystallization is accomplished by the addition of magnesium chloride and caustic solution, which stabilizes pH. Struvite pellets are collected from bottom of the reactor (Oleszkiewicz, 2015). A schematic diagram of this reactor has been presented in Figure 2.17. Some study demonstrated that this reactor would be capable of removing 80% of phosphorus and 20% of nitrogen (Kataki et al., 2015). Currently two full scale treatment plants are using this process, one in Boise, Idaho and another in City of Yakima, Washington (Kataki et al., 2015). The plant in Yakima was reported to produce 70 ton of struvite  during the first year of operation (Oleszkiewicz, 2015). Bowers and Westerman (2005) demonstrated from lab scale experiments that phosphorus removal process was optimal at pH 7.56 with 60 mg/L Mg addition for this conical shape fluidized bed reactor.  Figure 2.17: Multiform Harvest Process (adapted from Bowers and Westerman, 2005; Oleszkiewicz, 2015).    29  2.8.3 DHV Crystalactor  This crystalactor process was developed by DHV (Ingenieursbureau Dwars, Heederik en Verhey), Netherlands and marketed by Procorp Enterprises LLC in North America. The Crystalactor is a cylindrical fluidized bed reactor where phosphate containing wastewater is pumped in an upward direction. The reactor is partially filled with a suitable seed material like sand or minerals where phosphorus precipitates as calcium phosphate, magnesium ammonium phosphate or potassium magnesium phosphate. In order to precipitate the phosphate, a driving force is created by a reagent dosage and sometimes also pH-adjustment (Bergmans, 2011). The pellets grow and start to move towards the bottom of reactor. Largest fluidised pellets are discharged at regular intervals with the addition of fresh seed material (Bergmans, 2011).  Valsami-Jones (2004) reported that this reactor reduced TP concentration from 6.7 mg/L to 0.3 mg/L which yielded 101 kg P recovery per day as pellet in Geestmerambacht WWTP, Netherlands. High crystallization rate is also reported to be achieved with concentrated solutions (> 100 mg P/L) (Oleszkiewicz, 2015). Pilot scale results indicated optimal pH would be between 8.0 and 8.5 with Ca:P molar ratio 2-3:1. This reactor could produce 0.8 mm pellet size. It was demonstrated that high turbulence  was required for mixing  and high up-flow velocity (30-50 m/h) was required to keep the pellets fluidized in the reactor (Valsami-Jones, 2004).     30   Figure 2.18: Simplified scheme of DHV-Crystalactor (adapted from Oleszkiewicz, 2015; Valsami-Jones, 2004). 2.8.4 PHOSNIX The Phosnix which was developed in Japan by Unitika Ltd. The Phosnix  is capable of removing and recovering phosphorus from side stream (Ueno and Fujii, 2001). Figure 2.19 represents a schematic diagram of the process. The wastewater is introduced into the bottom of a fluidized bed reactor. The column of the reactor contains a bed of granulated struvite as seeding material for the growth of crystals. Magnesium hydroxide is added to maintain magnesium to phosphate ratio of 1:1 and the pH is stabilized to 8.2–8.8 with the addition of sodium hydroxide and by air stripping (Ueno and Fuji, 2001). Pellets grow between 0.5 and 1.0 mm in size with a crystallization retention time of 10 days, after which they are collected from the bottom of the reactor column. In order to maintain the continuity of the process granular struvite in separate liquid in returned to the column (Ueno and Fujii, 2001).    31  This process is operating in full scale at Lake Shinji and Fukuoka WWTP (Oleszkiewicz, 2015). This process was reported to recover 90% of phosphorus in the side stream as struvite which could reduce P concentration from 100-140 mg/L to 10 mg/L (Oleszkiewicz, 2015).  Figure 2.19: PHOSNIX process ( adapted from Ueno and Fujii, 2001). 2.8.5 AirPrex Airprex is technology to optimize biosolids treatment with a secondary option of phosphorus recovery. A CSTR type of reactor is utilized in this technology where struvite is precipitated by adjusting pH around 8 through air stripping of carbon di oxide and addition magnesium chloride (as shown in Figure 2.20). Sludge is lifted upward by air bubbles in the aerated zone in middle of the reactor which enables circulating movement of sludge. After reaching the surface, the sludge settles in the tranquil zone in the outer part of the reactor. Struvite is continuously removed from the bottom of tank. Sand washing equipment is utilized to ensure cleaning and purification of the recovered struvite (Desmidt et al., 2015; Oleszkiewicz, 2015; P-Rex, 2015).    32   Figure 2.20:AirPrex process (adapted from P-Rex, 2015). This technology is in full scale operation in three WWTPs in Germany and Netherlands. These plants were reported to remove 80 to 90% of phosphate from the liquid phase of the digested sludge (Desmidt et al., 2015). Table 2.2 shows the struvite production rate from three full scale operation. Table 2.2: Full Scale performance of AirPrex (adapted from Forstner, 2015). Name of WWTP Volume of sludge (m3/day) Struvite production (kg/day) Berlin Wassmannsdorf 2000 2500 Moenchengladbach-Neuwerk 1500 1500 Echten NL 400 500   33  2.8.6 PHOSPAQ Phospaq is another commercialized struvite recovery technology. This process takes place in an aerated reactor. Struvite precipitated by maintaining pH between 8.2 and 8.3 and addition of magnesium oxide. This reactor consists of separators in order to retain struvite. The struvite is collected from the bottom of the reactor by means of a hydrocyclone, followed by a screw press and transferred into a container. The harvested crystals have an average size of around 0.7 mm (Paques, 2016). This technology is in commercial use in The Netherlands at Lomm (for processing potato factory effluent), at Olburgen (for processing sewage sludge effluent following dewatering, combined with potato factory effluent) and in the UK at Severn Trent’s Stoke Bardolph wastewater treatment works (Kataki et al., 2015). It was reported that the Olburg plant could recover 82% of the initial phosphorus. On the other hand, the plant at Lomm could accomplish 75% recovery of initial phosphorus (Oleszkiewicz, 2015). 2.8.7 NuReSeys NuReSys stands for Nutrient Recycle Systems. A CSTR type of reactor is utilized in this technology where the reactor consists of a simple blade impeller (as shown in Figure 2.21). Struvite is crystallized under the pH (8-8.5) and by the addition of magnesium chloride. pH is controlled by the addition of NaOH. The struvite pellets are harvested by intermittent purging (Moerman et al., 2009). Moerman et al. (2009) demonstrated that this technology could accomplish 76% of P removal on full-scale application.       34   Figure 2.21: NuReSys process (adapted Moerman et al., 2009 and adapted from Desmidt et al., 2015). 2.8.8 Limitations of Technologies Although the technologies mentioned in section 2.8 are capable of reducing P-level in effluent to a satisfactory level, these technologies are associated with several shortcomings. This section highlights these shortcomings. Ostara Pearls is operated on low supersaturation ratio (ranging from1 to 1.6) which leads to lower nucleation and crystal growth rate according to Equation 2.3 and 2.5. Because of lower nucleation and crystal growth rate, longer crystallization retention time (8 to 12 days) is required to produce marketable size pellets. Longer crystallization time induces larger reactor volume which leads to higher capital cost. Again, this technology requires high recycle flow in order to maintain desired supersaturation ratio. In order to achieve this high recycle flow, powerful pump is required, which induces higher capital and operating cost. Moreover, this technology involves higher up-flow   35  velocity (320 cm/min) for mixing. Powerful pumps are also involved to operate this high up-flow velocity.  On the other hand, Multiform process produce fines as final product which requires further processing to make it as a marketable product. NuReSys technology is not also free from shortcomings. This technology is associated with higher energy consumption due to use of impeller and air stripping. This technology also yields higher footprint due to requirement of several reactors as they need reactors for clarification, air stripping and also require multiple reactors as CSTR (Continuous Stirred-Tank Reactor). Low value of final product is also another drawback of this technology. This technology is also associated with difficulties in controlling crystal size distribution.  Phospaq produces more fines rather than pellets. This technology requires removal of TSS. Utilization of air-stripping may render this process expensive.  Crystalactor requires constant seeding with sand, which in turns results in low P contained final products. On the other hand, final product from Air-Prex needs to be further processed to make it as a marketable product as it produces fines or larger crystals rather than pellets. Aeration is also involved in this process which has made this technology financially expensive. 2.9 Modelling Application  Optimum conditions for struvite crystallization varied from publication to publication, due to the facts that compositions of wastewater and concentrations of nutrients were different for each of these studies, which were defining factor for struvite precipitation. More importantly, some wastewater constituents can have significant impact on equilibrium and can co-precipitate with   36  struvite under specific conditions. As P-recovery in the form of struvite is related with a complex equilibrium which is dependent on entire wastewater matrix , mathematical modelling can be the best approach to analyze this kind of equilibrium (Lobanov et al., 2013). Many researchers have utilized various computational tools such as PHREEQC, MINTEQ, gPROMS, Struvite version 3.1 etc. to study relevant chemical equilibriums regarding struvite precipitation (Ali et al., 2013; Çelen et al., 2007; Gadekar and Pullammanappallil, 2010; Lobanov et al., 2013; Wang et al., 2006b). PHREEQC version 2.18 is a computer program developed by the U.S. Geological Survey for simulating chemical reactions along with transport processes in natural or polluted water. The program is not only based on equilibrium chemistry of aqueous solutions interacting with minerals, gases, solid solutions, exchangers, and sorption surfaces, but also includes the ability to model kinetic reactions with rate equations that are completely user-specified in the form of Basic statements (Parkhurst and Appelo, 1999). This study utilized the model developed by Lobanov et al. (2013) using PHREEQC version 2.18. Lobanov et al. (2013) included a substantial number of solid phases along with aqueous species, which are expected to have significant role in struvite equilibrium. Authors selected Lawrence Livermore National Library database file (llnl.dat; included in PHREEQC package) as a base for the model. They critically analyzed thermodynamic properties from the file for the main solid phases and aqueous species of interest and in some cases modified with the values critically selected from available published literature. For the solid phases with unknown thermodynamics properties, experiments were carried out to determine these properties.  Lobanov et al. (2013) upgraded the database which is comprised of more than 60 solid phases and more than 50 aqueous species related with struvite equilibrium. This database also encompassed a wide range of   37  temperature for all of the relevant compounds. Lobanov et al. (2013) demonstrated that modelling results showed a good agreement with the published data.               38  CHAPTER 3: THE PROBLEM STATEMENT AND OBJECTIVE 3.1 Problem Statement Although the technologies mentioned in section 2.8.1 to 2.8.8 are capable of reducing P-level in effluent to a satisfactory level, these technologies are associated with several shortcomings, which have been described in section 2.8.9. These shortcomings have driven to look for a more sustainable and financially feasible technology, which can remove and recover P from anaerobic digester supernatant along with the production of more marketable product. 3.2 Research Objective Overall purpose of this research was to find out a more efficient way to recover phosphorus from anaerobic digester supernatants. To accomplish this objective, a new struvite crystallization FBR was built. The specific objective of this study was optimization of operating physio-chemical and kinetic factors of this new FBR in order to ensure optimum phosphorus removal from supernatants as well as optimum phosphorus recovery as struvite pellet along with the least amount of production of fine struvite. 3.3 Research Questions In order to accomplish aforementioned objective, two sets of research question were formulated.   Research Question 1: Investigation of impact of higher supersaturation ratio under typical concentration of supernatant of anaerobic digester at comparatively constant conditions.  Research Question 2: Evaluation of impact of various hydraulics parameters under relatively constant supersaturation ratio.    39  3.4 Scope of the Study This study put effort to relate the effects of various operation conditions (both physio-chemical and hydrodynamic) on P-removal and recovery efficiency of abovementioned reactors. Efforts were also undertaken to describe probable reasons behind these relationships based on literature reviews. In depth analysis of various factors such as nucleation rate, crystal growth rate, agglomeration rate, nature of eddies formed, flow pattern, shear forces etc., which are expected to be probable elements to contribute in these relationships were out scope of this study.               40  CHAPTER 4: MATERIALS AND METHODOLOGY 4.1 Research Phases In order to the answer of two specific research questions mentioned in Chapter 3, whole research program was divided into two phases as shown in Figure 4.1. First phase was set to determine the influence of supersaturation on P-recovery process under typical concentration of anaerobic digester supernatant. This phase was continued from run 1 to run 7 including a test run where most of the run was continued for four days except run 2 and run 5. The second phase was aimed to evaluate the influence of various hydraulics parameters such as up-flow velocity, nozzle velocity, nozzle configuration etc. on P-recovery technology. This phase lasted for run 8 to run 16. Most of the runs also continued for four days for this phase.  Figure 4.1: Different phases of research.  Influence of Supersaturation RatioInfluence of upflow velocityInfluence of nozzle velocityInfluence of nozzle configurationPhase 1 (Run 0 to Run 7) Phase 2 (Run 7 to Run 16)   41  4.2 Synthetic Supernatant Synthetic supernatant was utilized during this study in order to evaluate the performance of the FBR. The detail composition and physical properties of synthetic supernatant during various phases of research was tabulated in Table 4.1. Table 4.1: Supernatant characteristics.  Mean 95% confidence Interval Minimum Maximum Phase 1     Phosphorus (mg/L) 151.65  ±2.95  120.5   165  Ammonia (mg/L) 679.07  ±20.5  477  881.5  Magnesium (mg/L) 117.58  ±2.22  94.47  129.27  Temperature (°C) 22.17  ±0.59  13.3  27.8  Conductivity (µS/cm) 10.81  ±0.24  8.41  12.36  Phase 2     Phosphorus (mg/L) 159.36  ±2.77  93.7  179  Ammonia (mg/L) 678.14  ±11.25  512.5  775  Magnesium (mg/L) 116.91  ±1.36  84.25  124.84  Temperature (°C) 23.36  ±0.43  20.2  34.2  Conductivity (µS/cm) 9.67  ±0.19  7.265  11.13   4.3 Process Layout The process layout of this pilot scale study has been depicted in Figure 4.2. There were three tanks in the layout as shown in Figure 4.2. One of the tank was for storing caustic, which was used to stabilize the pH in the system. Another tank was used to contain concentrated feed. Concentrated   42  feed was mixed with tap water in a mixing tank in order to have 20 times dilution to obtain feed with desired composition. From mixing tank, synthetic feed was inserted from bottom of the fluidized bed reactor where it was mixed with caustic. Five pH meters were set up at various locations in the layout in order to observe the change of pH during runs as indicated in Figure 4.2. There were also four sampling points with a view to investigating the change of supersaturation along the length of the reactor. A photograph of pilot scale is shown in Figure 4.3.  Figure 4.2: Process Layout.   43   Figure 4.3: Pilot Scale Setup. 4.4 Design of the Reactor The fluidized bed reactor was designed in order to fluidize particles from 145 μm to 1.5 mm. The reactor consisted of three segments – the injection port, the reaction zone and the clarifier. The detail description of these three segments are presented in following sections. 4.4.1 Injection Port The injection port is a 3” diameter cap with an inner diameter of 1.5” (as shown in Figure 4.4). Inner side of the port was grooved so that it could be joined tightly with the reactor. The injection port was equipped with five caps in order to fit various nozzles. The centre one was for fitting the nozzle, which was used for caustic flow. In other four caps, various sizes of nozzles were inserted Concentrated Feed Tank Caustic Tank Clarifier Reaction Zone Mixing Tank   44  for synthetic supernatant flow. For phase one, four nozzles with internal diameter of 0.508 mm were in operation. On the other hand, during phase two, in order to evaluate the influence of nozzle velocity and configuration on P-recovery efficiency, various sizes of nozzles with internal diameter from 0.3556 mm to 0.889 mm were used for supernatant flow with different configuration. Table 4.2 listed the diameter of various nozzles with their configuration for supernatant flow.           45    Figure 4.4:(a) Cross section of injection port, (b) Plan view of injection port.   (a) (b)   46  Table 4.2: Various nozzle sizes and configurations for supernantant flow. Run Nozzle ID, thou Nozzle ID, mm Nozzle Velocity a, m/s Number  Configuration 1-7 20 0.508 8.98 4 All around the centre nozzle (as shown in Figure 4.2) 8 14 0.3556 18.04 4 All around the centre nozzle (as shown in Figure 4.2) 9 14 0.3556 9.02 4 All around the centre nozzle (as shown in Figure 4.2) 10 28 0.7112 4.53 4 All around the centre nozzle (as shown in Figure 4.2) 11 28 0.7112 9.06 2 On the same side of the centre nozzle 12 25 0.635 8.95  4 All around the centre nozzle (as shown in Figure 4.2) 13 28 0.7112 9.06 2 On the different side of the centre nozzle 14 25 0.635 8.95 4 All around the centre nozzle (as shown in Figure 4.2) 15 17 0.4318 18.04 4 All around the centre nozzle (as shown in Figure 4.2) 16 35 0.889 4.5 4 All around the centre nozzle (as shown in Figure 4.2) a Calculations provided in Appendix B.   47  4.4.2 The Reaction Zone The reaction zone of the reactor was designed in such a way so that reaction would be mostly completed within this zone before reaching clarifier. This zone consisted of a 1.5” internal diameter clear PVC pipe with an approximate length of 47 inch. The volume of this zone was calculated to be 1.36 L (as calculated in Appendix A). At the bottom of this zone, harvesting point was inserted as indicated in Figure 4.5. A valve was inserted at this harvesting point to aid in harvesting. Clear pipe was used to observe the behaviour of struvite crystals and to monitor the occurrence of plugging. Up-flow velocities of this section varied from 19.67 to 62.21 cm/min at various flow with hydraulic retention time (HRT) from 2 to 6.33 minutes (as calculated in Appendix C). Reynolds number corresponding to these velocities varied from 140 to 443 (see Appendix D). These Reynolds number were determined to quantify the degree of turbulence. However, it should be mentioned that Reynolds number determined here are expected to be quite different for the reaction zone with pellets. 4.4.3 Clarifier Above the reaction zone, there was a clarifier which was made of food grade high density polyethylene as shown in Figure 4.5. The clarifier consisted of two sections- a conical section at bottom with a pail section at the top. The aim of providing conical section at the bottom of clarifier was to help in returning the fine particles into the reaction zone so that they can grow into larger pellets through agglomeration with other fine particles. The approximate volume of the clarifier was calculated as 31.3 L (see in Appendix A) with a total approximate length of 25”. There were two outlets at the top of the clarifier to discharge as indicated in Figure 4.5. Up-flow velocities of this section ranged from 0.29 to 0.93 cm/min during this study with hydraulic retention time 46 to   48  145.6 min (detail calculations were shown in Appendix C). Corresponding Reynolds numbers were calculated from 17 to 53 (as calculated in Appendix D).  Figure 4.5: Detail drawing of FBR struvite crystallizer.   49  4.5 Storage Tanks and Pumps Concentrated feed was stored in a 150-gallon holding tank from where concentrated feed was pumped into the mixing tank using a Masterflex Peristaltic pump with a standard pump head as illustrated in Figure 4.2.  Mixing tank was an inductor tank with a conical bottom. From mixing tank synthetic feed was pumped into the bottom of the reactor by a double head Masterflex Peristaltic pump with a Cole Parmer motor (model no # 7553-70) except for run 8 and 15 where a Seepex pump (model# 025-6LMD) with a Leeson motor (model no # C14D6F218) was used. Caustic was stored in 105-gallon capacity holding tank and pumped into injection port using Masterflex Peristaltic pump with a standard pump head.  4.6 Chemicals Concentrated feeds were prepared from struvite, sodium chloride, ammonium chloride. Struvite used in this study were collected from previous experiments on struvite crystallization process by some other researchers (S. Lobanov, Postdoctoral Research Fellow, Civil Engineering Department, University of British Columbia, per.comm.). Concentrated hydrochloric acid was utilized in order to dissolve struvite. Technical grade sodium chloride and ammonium chloride were used for the preparation of synthetic feed. In some case technical grade phosphoric acid and magnesium chloride were utilized to maintain the desire concentration of PO4—P and Mg. Caustic feed was prepared from technical grade caustic soda (micropearls tech).   50  4.7 Process Control, Monitoring and Maintenance  Struvite formation is highly dependent on pH, as struvite is less soluble in alkaline solutions. As a result, adjustment of pH is a very important step to control struvite crystallization. During the formation of struvite, pH drops due to release of protons, so it was necessary to rise pH of the system in order to maintain a specific supersaturation ratio. Fujimoto et al.(1991) stated that addition of sodium hydroxide was more effective to control pH compared to lime or magnesium hydroxide. Aeration by stripping CO2 is also capable of controlling pH but it takes longer time (Battistoni et al., 1998). In this study, preliminary bench tests (called beaker tests) were executed to determine the caustic flow rate to maintain a desired influent supersaturation ratio. Caustic flows were determined in such a way so that the pH at 0 min could be matched with the pH determined by PHREEQC model. Table 4.3 illustrated the various caustic flow determined for various runs. A typical curve and sample calculation corresponding to Beaker test has been depicted in Appendix E. A pH controller (model no# Cole Parmer pH/ORP 350) was installed at a distance of 45.5” from bottom of the reactor (except run 0 when pH controller was installed at a distance of 1’-4” from the bottom of the reactor) as shown in Figure 4.5 in order to stabilize pH in the system.          51  Table 4.3: Caustic flows corresponding to various runs. Run Concentration of caustic (M) Feed Flow (ml/min) Caustic flow (ml/min) 0 0.545 483 12 1 0.545 437 11 2 0.545 437 14 3 0.2725 437 16.2 4 0.2725 437 21 5 0.268 437 19 6 0.268 437 23.3 7 0.268 437 17 8 0.268 430 16.7 9 0.148 215 8.4 10,11,13 0.268 432 16.8 12,14 0.268 680 26.5 15 0.268 634 24.7 16 0.268 670 26.1 pH along with temperatures at various points as shown in Figure 4.2 and 4.3 were measured regularly to monitor pH and temperature changes along the length of the system. Models of various pH meters used in this study have been tabulated in Appendix F. Every pH meters were calibrated using three points calibration points by standard pH 4, pH 7 and pH 10 buffer solutions before each run. A pH controller was calibrated using two points calibration by standard pH 7 and pH 10 buffer solutions. After each run, calibration of each pH meter was checked against standard pH 7 and pH 10 buffer solutions (except the pH meter for mixing tank which was rechecked by standard pH 4 and pH 7 buffer solutions). Typical pH calibration curves produced based on the recheck of pH meters are shown in Appendix G.  In order to have better understanding and also to monitor and control the struvite formation process, several parameters such as pump pressure, water main pressure, conductivity, flow coming at mixing tank and effluent flow etc. were measured and recorded regularly along with temperature and pH. These include grab samples from influent, filtered and non-filtered samples from effluent, filtered samples from the two points which were 1’-4” and 45.5’ (pH controlling   52  point) from bottom of the reactor. Collection and filtration procedure of samples is described in section 4.8. These samples were analyzed to determine the concentrations of NH4+, PO43-, Mg2+. Analytical methods to measure the concentration of these constituents are described in section 4.9. Flow coming at mixing tank and flow from clarifier (effluent flow) were measured using a graduated cylinder and a stop watch. As conductivity is an important parameter to estimate supersaturation ratio, conductivity at mixing tanks was measured using a portable Oakton meter with built in temperature sensor. Conductivity meter was calibrated using recommended solution before each run. Water main pressure was recorded regularly along with the pump pressure, which was used to pump feed from mixing tank to the reactor.   Occasionally injection port and reaction zone were clogged by struvite accumulation. A screw driver or a thin rod was used to scrap off struvite from the injection port and the wall of the reactor. As acid washing is a very efficient way to remove struvite from the reactor (Huang, 2003), the reactor was cleaned with acid after the end of each run. Growth of some type of bacteria was observed in mesh filters. 4.8 Sample Collection A precise and accurate sampling procedure was required to ensure the collection and preservation of representative samples. Grab samples from mixing tank, clarifier (effluent) and two intermediate points in the reactor (as discussed in section 4.7) were collected in 50 ml plastic test tubes. Samples from clarifier and two intermediate points were filtered through 0.45 microns’ membrane filters following standard methods. One or two drops of hydrochloric acid were used for all samples except the samples from mixing tank in order to dissolve struvite through lowering pH of the solutions.   53  4.9 Analytical Methods A summary of analytical techniques used in this study is presented in Table 4.4. Table 4.4: Summary of analytical methods. Parameter Analytical Technique Instruments Method Version Reference pH Electrometric Method Various pH meters as shown in Appendix G 4500 H+ APHA et al. (2005) PO4-P  H2SO4 digestion, Flow Injection Analysis Lachat QuikChem 8000 4500 APHA et al. (2005) NH4-N H2SO4 digestion, Flow Injection Analysis Lachat QuikChem 8000 4500-NH3 APHA et al. (2005) Magnesium and Sodium Atomic Absorption Spectrometry Varian Spectra AA 200 FS 3500-Mg B 3500-Na B APHA et al. (2005)  A brief discussion on analytical methods is described in following sections. 4.9.1 Ortho-Phosphate (PO4-P) and Ammonium Nitrogen (NH4-N) In addition of following standard method in order to get proper range of PO4-P/NH4-N concentration from the equipment, a sample size of 5ml were used for the analysis based on previous experience of PO4-P/NH4-N analysis (T. Ma, Department Safety and Environmental Lab Specialist, Dept. of Civil Engineering, UBC, Vancouver, B.C., pers. comm.). Calibration ranges used for this study were 0-25 mg/L for PO4-P and 0-50 mg/L for NH4-N. 4.9.2 Magnesium and Sodium In addition to following standard methods, the instrumental parameters which were used according to manufacturer’s specification have been listed in Table 4.5. Lanthanum chloride, with   54  concentration of 20000 µg/ml was used to overcome interferences for determination of magnesium concentration. An eight-point calibration curve, (0,0.5,1,2.5, 5, 10,15 and 25 mg/L Mg) was generated for magnesium analysis. On the other hand, a five-point calibration curve (0, 0.1,0.5,1 and 2 mg/L Na) was generated for sodium analysis. Table 4.5: Instrumental parameters for magnesium and sodium analysis.  4.9.3 Caustic (NaOH) Caustic concentrations were determined using standard acid titration method. Samples collected from caustic tank were titrated against 1 N HCl acid and end point was detected using pH meter. 𝐶𝑁𝑎𝑂𝐻 × 𝑉𝑁𝑎𝑂𝐻 =  𝐶𝐻𝐶𝑙 × 𝑉𝐻𝐶𝑙                                                                                                              (4.1) Where, CHCl = Concentration of HCl,  VHCl = Volume of HCl used, CNaOH = Concentration of NaOH,  VNaOH = Volume of NaOH used.  Magnesium Sodium Lamp Current (mA) 4 5 Fuel Acetylene Acetylene Support Air Air Flame Stoichiometry Oxidizing Oxidizing Wavelength(nm) 285.2 330.3 Slit width (nm) 0.5 0.5   55  4.10 Pellets Harvesting and Fines Collection Struvite pellets were collected when there were sufficient amount of pellets in the reactor or sometimes when the injection port was clogged. Struvite pellets were harvested at point which was 1.625” from bottom of the reactor (as shown in Figure 4.5) by opening a valve installed at that point and using a clear PVC pipe. Harvested pellets were collected using bucket and then kept in air for drying. Based on atmospheric condition, it actually took 1-2 days for accomplishing drying. Dried struvite crystals were sieved through 1 mm, 0.5 mm and 0.25 mm sieve size in order to classify the pellets. After sieving, each fraction of crystals was weighed with a digital balance. After each run, reactor was drained in order to collect fines by filtration through a Fisherbrand filter paper (catalog no # 09-790F) using a Millipore XX5500000 vacuum pump.   4.11 Pellets and Fines Quality Determination The quality determination of pellets and fines were accomplished through chemical analysis, XRD analysis and morphology examination. Random selection of pellets were performed for analysis based on the assumption that they were representative ones.   4.11.1 Chemical Analysis  For chemical analysis, 100 mg of struvite pellets and fines were dissolved in 100 ml of water using concentrated HCl and then analyzed for Mg, NH4-N and PO4-P.  4.11.2 XRD Analysis Bruker D8 Advance X-ray diffractometer (XRD) with CuKα radiation was applied for XRD analysis in this study. The analysis was executed on powdered struvite and fines to match the intensity and positions of the peaks with the standard peak of crystalline phase provided by the   56  powder diffraction database file, PDF-2, prepared by the International Centre for Diffraction Data. Details of instrumental settings are in tabulated in Table 4.6.  Table 4.6: Instrumental settings for Bruker D8 Advance XRD. Parameter  Type of radiation CuKα Scanning angle, 2θ  5º-70º Average scanning rate 0.019º  4.11.3 Morphology Morphology of the struvite pellets and fines was studied using Motic B3 Professional Series microscope and images were captured and processed with the help of Motic Images Plus software.  4.12 Terminology Various terminologies used in this paper have been discussed in this section. 4.12.1 Supersaturation Ratio In this study, for struvite following expression was used to calculate supersaturation ratio-          𝑆𝑠𝑡𝑟𝑢𝑣𝑖𝑡𝑒 = ({𝑀𝑔2+ }{𝑁𝐻4+}{𝑃𝑂43− }𝑘𝑠𝑝(𝑠𝑡𝑟𝑢𝑣𝑖𝑡𝑒))1/3                                                                                           (4.2) {} has been used to indicate activity of ionic species and ksp(struvite) is thermodynamic solubility product of struvite. Here, ionic activity was used instead of concentration because of existence of complex ionic systems in wastewater.    57  4.12.2 Removal Efficiency X-removal efficiency = (𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 X in influent × influent flow - Concentration of X 𝑖𝑛 𝑒𝑓𝑓𝑙𝑢𝑒𝑛𝑡×𝑒𝑓𝑓𝑙𝑢𝑒𝑛𝑡 𝑓𝑙𝑜𝑤  𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 X in influent × influent flow × 100) %    (4.3) 4.12.3 Struvite Recovery Efficiency Struvite Recovery Efficiency = (𝐴𝑚𝑜𝑢𝑛𝑡 𝑜𝑓  𝑠𝑡𝑟𝑢𝑣𝑖𝑡𝑒 𝑝𝑒𝑙𝑙𝑒𝑡𝑠 ℎ𝑎𝑟𝑣𝑒𝑠𝑡𝑒𝑑𝐴𝑚𝑜𝑢𝑛𝑡 𝑜𝑓 𝑠𝑡𝑟𝑢𝑣𝑖𝑡𝑒 𝑝𝑟𝑜𝑑𝑢𝑐𝑒𝑑 × 100) %                          (4.4) 4.12.4 Coefficient of Variation (COV) The coefficient of variation (CV) is defined as the ratio of the standard deviation (σ)  to the mean (μ).  𝐶𝑂𝑉 =  (σμ × 100) %                                                                                                                             (4.5) 4.12.5 Pellets Growth Rate Pellets growth rate can be defined as amount of sizeable pellets grow per unit time. For this study, pellets growth rate has been expressed using following expression- 𝑃𝑒𝑙𝑙𝑒𝑡𝑠 𝑔𝑟𝑜𝑤𝑡ℎ 𝑟𝑎𝑡𝑒 =𝑀𝑎𝑠𝑠 𝑜𝑓 𝑝𝑒𝑙𝑙𝑒𝑡𝑠 ℎ𝑎𝑟𝑣𝑒𝑠𝑡𝑑 𝑇𝑖𝑚𝑒 𝑒𝑙𝑎𝑝𝑠𝑒𝑑 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 ℎ𝑎𝑟𝑣𝑒𝑠𝑡𝑖𝑛𝑔                                                   (4.6) 4.12.6 Nominal Weight of Fines Nominal weight of fines is the amount of fines produced per unit volume of influent, which has been calculated using following expression- 𝑁𝑜𝑚𝑖𝑛𝑎𝑙 𝑤𝑒𝑖𝑔ℎ𝑡 𝑜𝑓 𝑓𝑖𝑛𝑒𝑠=  𝐴𝑚𝑜𝑢𝑛𝑡 𝑜𝑓 𝑓𝑖𝑛𝑒𝑠 𝑝𝑟𝑜𝑑𝑢𝑐𝑒𝑑 𝑖𝑛 𝑒𝑎𝑐ℎ 𝑟𝑢𝑛𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑓𝑙𝑜𝑤 𝑐𝑜𝑚𝑖𝑛𝑔 𝑡𝑜 𝑡ℎ𝑒 𝑟𝑒𝑎𝑐𝑡𝑜𝑟 𝑑𝑢𝑟𝑖𝑛𝑔 𝑒𝑎𝑐ℎ 𝑟𝑢𝑛 × 𝑑𝑢𝑟𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑒𝑎𝑐ℎ 𝑟𝑢𝑛        (4.7)    58  4.13 Quality Assurance/Quality Control 4.13.1 Quality Assurance The Field quality assurance involves a series of steps, procedures and practices. According to  British Columbia Ministry of Environment (2003) following steps were undertaken to ensure field quality-  Sample bottles with cap were kept cleaned and could be certified as ‘contamination free’, for the analysis.   Contamination free reagents were added.   Sample bottles were capped and kept in a clean environment, away from dust, dirt, fumes and grime. Bottles were stored in clean shipping containers both before and after the collection of the sample.   Samples were stored in a cool, dark place and shipped to the laboratory just immediately after collection.   Sample collectors used recommended safety features and kept their hands clean and were ceased from eating or smoking while working with water samples  Sample equipment were cleaned after each sampling round.  4.13.2 Quality Control Quality control is a crucial part of analysis. Duplicates samples for each analytes were prepared in order to maintain the quality of data generated. Procedures and methods for quality control and quality assurance were followed according to standard methods 1020 B and 1020 C (APHA et al., 2005).   59  4.13.2.1 Data Quality Objective (DQO) A coefficient of variation (COV) of 10% was considered in this study for magnesium analysis. On the other hand, for ammonium-nitrogen and orthophosphate analysis, a difference of 10% between duplicates was taken into consideration for this study.                 60  CHAPTER 5: RESULTS AND DISCUSSION 5.1 Impact of Physiochemical Variables  Based on Equation (2.2), (2.4) and (2.5) efforts were made to develop a hypothetical qualitative diagram to show the impact of supersaturation on nucleation, crystal growth and agglomeration rate as shown in Figure 5.1. From Figure 5.1 it is apparent that nucleation rate is lower at low supersaturation level but crystal growth rate is higher. On the other hand, at higher supersaturation level (after point “c”), nucleation rate overtakes crystal growth rate and becomes more dominant process than crystal growth and agglomeration. This section has tried to find out the optimum supersaturation level, which will produce balance among nucleation, crystal growth and agglomeration rate to obtain higher P-recovery efficiency along with lower fines production.   61   Figure 5.1: Struvite crystallization rate. 5.1.1 pH pH is one of the most important variables to exert impact on struvite crystallization as the solubility of struvite is closely related to it. Various studies were performed in order to analyze the effect of pH on P- removal through struvite crystallization reactor (Adnan et al., 2003; Battistoni et al., 2001; Fattah et al., 2008; Kumashiro et al., 2001; Munch and Barr, 2001; Ohlinger, 1999). pH of effluent was measured during each sampling time as discussed in section 4.7 and 4.8 but there was no provision to measure initial pH directly. Instead of measuring pH directly, PHREEQC model was used to generate initial pH in the reactor. Figure 5.2 illustrates the relationship between initial and effluent pH which was found to be linear within the operating range. (c)   62   Figure 5.2: Relationship between initial and effluent pH. Figure 5.3 represents the impact of pH on P-removal efficiency. It shows a general trend of increase in P-removal with increase in pH. This finding coincides with the agreement of other various studies (Fattah et al., 2008).  Although increase in pH yielded increment in P-removal efficiency, but after pH 8 this increment became negligible according to Figure 5.3. Although some researchers reported high pH values (8.2 to 9.0) to have P- removal efficiency (above 80%) (Battistoni et al., 2001; Jaffer et al., 2002; Munch and Barr, 2001; Ohlinger, 1999) , this study was capable of achieving up to 90% of P-removal at pH 8 or even at slightly lower pH. Even 80% of P-removal was achieved at pH 7.5. This result is comparable to the study by Adnan (2002) and  Fattah et al. (2008) as Adnan (2002) reported 79% of P-removal at a pH 7.1 using synthetic supernatant and Fattah et al. (2008) reported 80%  P- removal with a pH level of 7.5  using centrate at Lulu Island Wastewater Treatment Plant, Richmond, BC. y = 1.9543x - 8.8478R² = 0.863277.27.47.67.888.28.48.68.10 8.20 8.30 8.40 8.50 8.60 8.70 8.80 8.90pH effluentpH initial (model generated)  63  Figure 5.3 also shows a comparison between measured P-removal efficiency with model generated values at three different supersaturation ratios (1, 1.17 and 1.36) of effluent. From this figure it is apparent that this pilot scale study yielded result which were more closely coincided with the values generated by model using effluent supersaturation ratio of 1.17 and most of the values fallen within the curves generated by using supersaturation ratio of 1 and 1.36. This outcome was expected as it took a long time to reach complete chemical equilibrium i.e. effluent supersaturation ratio of 1...………………………………………………………………………………………….                           Figure 5.3: Impact of effluent pH on P-removal efficiency.  Efforts were made to demonstrate the impact of initial pH on P-removal performance of the reactor as initial pH affects initial supersaturation which exerts impact on nucleation and growth rate (Fang et al., 2016; Le Corre, 2006) . Figure 5.4 shows that increase in initial pH yielded increase in P-removal efficiency. Lee et al. (2015) also found the same trend by carrying out investigation on 304050607080901006 6.5 7 7.5 8 8.5 9Phosphorus Removal Efficiency (%)pH (effluent)Modelled S (effluent) = 1Modelled S (effluent) = 1.17Modelled S (effluent ) = 1.36MeasuredPoly. (Measured)  64  anaerobic digester supernatant of swine manure. Again, after pH 8.6 this increase in P-removal became insignificant. Figure 5.4 also manifests that 85% of P-removal could be accomplished by an initial pH 8.5 while Lee et al. (2015) reported that 77.8% of P-removal could be achieved by an initial pH of 13.   Figure 5.4: Impact of initial pH on P-removal. 5.1.2 Supersaturation Ratio As supersaturation ratio is considered to be the driving force of struvite crystallization process, this study has tried to investigate the impact of supersaturation ratio on P-removal efficiency, P-recovery efficiency, fine production and shape and size of pellets. This study achieved phosphorus removal by controlling supersaturation ratio in the reactor and supersaturation ratio was controlled by adjusting pH. Figure 5.5 demonstrates that pH was randomly related with supersaturation ratio compared to the model generated values at three different supersaturation values of effluent which was due to the fact that supersaturation is not only dependent on pH but also closely related with other parameters such as concentrations of Mg2+, NH4+, PO43- or temperature etc. R² = 0.8558505560657075808590951008.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9Phosphorus Removal efficiency (%)pH (initial)  65   Figure 5.5: Effluent pH Vs Supersaturation ratio. Subsequent sections have discussed the findings of this study of supersaturation ratio on P-removal and recovery performance elaborately. 5.1.2.1 P-Removal Efficiency The graph of P- removal efficiency against supersaturation ratio was found to be scattered with a tentative trend of increasing P-removal with increase in supersaturation ratio up-to supersaturation ratio of 6.5 as shown in Figure 5.6. So there is no point to increase supersaturation ratio above 6.5. This graph also suggests that optimal P-removal could be achieved within a supersaturation ratio range of 5.5 to 6.5.  Similar trends were observed in some other studies (Bhuiyan et al., 2008; Fattah et al., 2008). The scatter distribution of P-removal efficiency with supersaturation ratio 6.577.588.593 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8pH (effluent)Intial SstruviteModelled S (effluent) = 1 Modelled S (effluent) = 1.17 Modelled S (effluent) = 1.36Measured Poly.(Measured)  66  compared to the model generated curves again indicated the dependency of supersaturation ratio on various parameters including temperature, pH, concentration of struvite constituents.   Figure 5.6: Impact of initial supersaturation ratio on P-removal efficiency. 5.1.2.2 Fine Production Amount of fine produced is also an important criterion in order to establish an optimum condition for phosphorus removal and recovery. Higher amount of fine production is economically a drawback for phosphorus recovery as more processing will be required to make a marketable product where pellets can be readily used as fertilizer.   Figure 5.7 illustrates that amount of fine production was increased with increase in supersaturation ratio both for actual and theoretical calculation. These findings also can be explained based on Figure 5.1. According to Figure 5.1 higher supersaturation level induces higher nucleation rate compared to crystal growth rate which results in more fines production rather than pellets production. Difference between theoretical and actual value was resulted from the loss of fines 304050607080901003 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8Phosphorus Removal Efficiency (%)Initial SstruviteModelled S (effluent) = 1Modlled S (effluent)= 1.17Modelled S (effluent) = 1.36MeasuredPoly. (Trendline)  67  during sample collection and filtration, pellets harvesting and fines collection. In order to plot this graph (Figure 5.7), average value of supersaturation ratio for each run was considered and standard deviation of supersaturation ratio for each run has been depicted in Figure 5.7 as error bars.  Figure 5.7: Influence of supersaturation on fine production. 5.1.2.3 P-Recovery Efficiency Increasing supersaturation ratios were found to have a negative impact on P-recovery efficiency of the reactor as indicated by Figure 5.8. Figure 5.8 also suggested that optimal P-recovery efficiency as struvite was achieved within the range of supersaturation ratio of 4 to 6 while efficiency tended to be very low when supersaturation ratio was more than 6. As higher supersaturation ratio produces more fines (as shown in Figure 5.7), so higher supersaturation ratio yields lower P-recovery efficiency. This trend can also be explained with the help of Figure 5.1. According to Figure 5.1, it is apparent that at higher supersaturation level nucleation rate is more 00.20.40.60.811.24 4.5 5 5.5 6 6.5 7 7.5Nominal weight of fines (gm/L)Initial SstruviteActual Theoretical Expon. (Actual) Expon. (Theoretical)  68  than crystal growth rate, which leads to lower struvite recovery efficiency as struvite pellets.  Another point is to be noted here that this reactor yielded a relatively lower struvite recovery efficiency (maximum of 14.8%) compared to some other studies conducted by Fattah et al. (2008) and Mavinic et al. (2007) where they accomplished about 86% P-recovery efficiency. This was due to the fact that this small scale reactor was not equipped with any special harvesting system. Pellets were harvested only by opening the harvest valve at the bottom of the reactor and draining the entire content of the bottom section, which actually led to loss of a lot of fines through in-reactor liquid which was wasted during harvesting. Struvite recovery efficiency is expected to be improved with higher efficient harvesting and recovery process in a larger scale installation as some preliminary experiments with an up-scaled reactor accomplished more than 80% P-recovery (S. Lobanov, Postdoctoral Research Fellow, Civil Engineering Department, University of British Columbia, per.comm.).  Figure 5.8 Impact of supersaturation ratio on P-recovery efficiency. Trend observed in Figure 5.8 was re-confirmed by plotting the pellets growth rate against supersaturation ratio as illustrated in Figure 5.9. Figure 5.9 also indicated that higher pellets growth 02468101214163.5 4 4.5 5 5.5 6 6.5 7 7.5 8Phosphorus recovered as Struvite (%)Initial Sstruvite  69  rates were associated with supersaturation of 4 to 6 same as suggested by Figure 5.8. Here is to be noted that, in order to plot both Figure 5.8 and 5.9 average of supersaturation ratios between consecutive harvesting was considered.  Figure 5.9: Influence of supersaturation ratio on pellets growth rate. 5.1.2.4 Pellet Size Distribution The pellets harvested during this study were mostly within the size range of 1 mm to 0.5 mm as indicated by Figure 5.10. From Figure 5.10 it is apparent that there was a tentative trend of production of larger pellet size (size > 1mm) with higher supersaturation ratio. On the other hand, production trend of smaller size of pellets (0.5 mm <size< 0.25 mm) was inversely related with supersaturation ratio. For pellets with size between 1 mm to 0.5 mm optimal production was accomplished within supersaturation ratio from 4.0 to 6.0 as depicted in Figure 5.10. During operating under high supersaturation ratio, a lot of crystals become available for agglomeration due to high nucleation rate which might lead to formation of larger pellets by agglomeration. This explanation also gives an idea that agglomeration might be the dominant process over crystal growth for development of the pellets under high supersaturation ratio. 01020304050607080903.5 4 4.5 5 5.5 6 6.5 7 7.5 8Pellets Growth Rate (gm/day)Initial Sstruvite  70   Figure 5.10: Pellet size distribution against initial supersaturation ratio- (a) Size >1mm, (b) 1mm <Size <0.5 mm and (c) 0.5 mm <Size <0.25 mm. 5.1.2.5 Morphology of Pellets and Fines In order to understand the impact of supersaturation on the morphology of pellets, pellets were examined under a Motic B3 Professional Series microscope. In general, harvested pellets were found to be with satisfactory hardness as they were not easy to break apart during the process of harvesting, drying, sieving, weighting and transportation.  Figure 5.11 shows the images of pellets under different supersaturation condition which were magnified at 100X. According to Figure 5.11, pellets were found to be elongated with rough surface under low supersaturation value. On the other hand, pellets were more round with smooth surface under higher supersaturation value. As more fines were produced under high 024681012143 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8P-recoverd as struvite (%)Initial Sstruvite024681012143 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8P-recovered as struvite (%)Initial Sstruvite024681012143 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8P-recovered as struvite (%)Initial Sstruvite(a) (b) (c)   71  supersaturation level due to high nucleation rate, these fines might contribute to fill up the gaps on the surface of pellets to make it smoother. Further investigations are required to point out more specific explanation to this phenomenon.    Figure 5.11: Microscopic images of pellets (Size >1 mm) under different Supersaturation condition magnified at 100X - (a) Sstruvite-4.8, (b) Sstruvite-5.34, (c) Sstruvite-6.23, (d) Sstruvite-6.40. Morphology of fines was also examined under microscopes in order to analyze the shape of the crystals as supersaturation level can have significant impact on the shape of the crystals (Lobanov, (a) (b) (c) (d)   72  2009). Shape of the fines would give an idea regarding the formation mechanism of struvite crystals. Crystals formed under higher supersaturation ratio was found to be needle shaped with sharp edges as shown in Figure 5.12 (c). Such crystals were also detected by  Lobanov (2009) at higher supersaturation level. Under relatively lower supersaturation (Sstruvite = 5~5.5) more plate shaped crystals were formed as indicated by Figure 5.12 (a) and (b). But comparing to higher supersaturation ratio, crystals formed under lower supersaturation level were detected to have more irregular shape. Crystals formed under lower supersaturation were observed to be developed along all three dimensions rather than just in one dimension as in the higher supersaturation ratio.      Figure 5.12: Microscopic images of fines under different Supersaturation condition magnified at 400X - (a) Sstruvite-5.04, (b) Sstruvite-5.49, (c) Sstruvite-7.21. (a) (b) (c)   73  5.1.2.6 Purity and Composition of Pellets and Fines In order to determine the purity and quality of fines and pellets produced under different supersaturation level, crystals and fines were analyzed chemically. Samples (fines and pellets) were selected from lowest and highest supersaturation level in order to investigate the impact of supersaturation level on the purity and composition of pellets and fines. 100 mg of each sample was dissolved in 100 ml of distilled water and then resultant solution was analyzed to determine the concentration of Mg, NH4-N and PO4-P. Table 5.1 summarizes the result of chemical analysis of pellets and fines. From the analysis, it can be concluded that both pellets and fines were high purity struvite. It is also apparent from Table 5.1 that influence of supersaturation level on the chemical composition of both pellets and fines is negligible as both high and low supersaturation ratio yielded highly pure pellets and fines. Table 5.1: The results of chemical analysis of pellets and fines. Constituents Theoretical Value (mg/L) Analytical Value (mg/L) Pellets Fines Sstruvite-7.10 Sstruvite-4.64 Sstruvite-7.21 Sstruvite-5.04 PO4-P 126.4 122.5 127.5 115.5 121 NH4-N 57.1 57.6 61.45 55.1 57.15 Mg 99.1 97.6 101.5 94.6 95.9   74  Purity of fines and pellets were re-examined by XRD analysis. X-ray diffraction pattern for pellets and fines under different supersaturation level are illustrated from Figure 5.13 to Figure 5.16. The peak and intensities of struvite pellets produced under lower supersaturation level matched with the peak and intensities of a powder diffraction file (PDF) for struvite (PDF 01-077-2303) as shown in Figure 5.13 which conforms the purity of the pellets. On the other hand, although pellets formed under higher supersaturation level produced the peaks and intensities similar to struvite (PDF 00-015-0762), but there also existed some trace amounts of cattiite (Mg2(PO4)2.22H2O) corresponding to PDF 00-055-0828 and kovdorskite (Mg2 (PO4)(OH).3H2O) corresponding to PDF 00-045-1380 as per Figure 5.14. These results can be justified on the basis of literature reviews as various studies also reported that increment in pH can result in transformation of struvite species into cattiite (Dempsey, 1998; Taylor et al., 1963). Bhuiyan  et al., (2008b) also detected the presence of cattiite under high pH level. Again, Ma et al. (2014) observed the presence of  kovdorskite under high pH.  For fines under both lower and higher supersaturation condition, XRD analysis conformed pure struvite as indicated in Figure 5.15 and 5.16.    75    Figure 5.13: X-ray diffraction pattern of struvite pellets (Size: 1 mm to 0.5 mm) for low supersaturation level (Sstruvite- 4.64).  Figure 5.14: X-ray diffraction pattern of struvite pellets (Size: 1 mm to 0.5 mm) for high supersaturation level (Sstruvite- 7.10). Run 3 0.5-1 mm harvest (low SSR)01-077-2303 (*) - Struvite, syn - MgNH4PO4(H2O)6 - Y: 100.30 % - d x by: 1. - WL: 1.5406 - Orthorhombic - a 6.95500 - b 6.14200 - c 11.21800 - alpha 90.000 - beta 90.000 - gamma 90.000 - Primitive Operations: ImportFile: Run3_low_SSR_harvest.raw - Start: 5.000 ° - End: 64.999 ° - Step: 0.019 ° - Step time: 36.4 s - Anode: Cu - WL1: 1.5406 - WL2: 1.54439 - kA2 Ratio: 0.5 - Generator kV: 40 kV - Generator mA: 40 Lin (Cps)01002003004005006007002-Theta - Scale5 10 20 30 40 50 60Test run 0.5-1 mm harvest (high SSR)00-045-1380 (I) - Kovdorskite - Mg2(PO4)(OH)·3H2O - Y: 22.59 % - d x by: 1. - WL: 1.5406 - Monoclinic - a 4.73800 - b 12.92300 - c 10.33400 - alpha 90.000 - beta 101.480 - gamma 90.000 - Primitive -00-055-0828 (I) - Cattiite - Mg3(PO4)2·22H2O - Y: 21.57 % - d x by: 1. - WL: 1.5406 - Triclinic - a 6.93200 - b 6.92500 - c 16.15400 - alpha 82.210 - beta 89.700 - gamma 119.510 - Primitive - P-1 (2) - 1 00-015-0762 (*) - Struvite, syn - NH4MgPO4·6H2O - Y: 44.84 % - d x by: 1. - WL: 1.5406 - Orthorhombic - a 6.94500 - b 11.20800 - c 6.13550 - alpha 90.000 - beta 90.000 - gamma 90.000 - Primitive - Operations: ImportFile: Test_run_high_SSR_harvest.raw - Start: 5.000 ° - End: 64.999 ° - Step: 0.019 ° - Step time: 36.4 s - Anode: Cu - WL1: 1.5406 - WL2: 1.54439 - kA2 Ratio: 0.5 - Generator kV: 40 kV - Generator mALin (Cps)01002003004005002-Theta - Scale5 10 20 30 40 50 60  76   Figure 5.15: X-ray diffraction pattern of fines for low supersaturation level (Sstruvite- 5.04).  Figure 5.16: X-ray diffraction pattern of fines for high supersaturation level (Sstruvite- 7.21). Run 3 fines (low SSR)00-015-0762 (*) - Struvite, syn - NH4MgPO4·6H2O - Y: 80.81 % - d x by: 1. - WL: 1.5406 - Orthorhombic - a 6.94500 - b 11.20800 - c 6.13550 - alpha 90.000 - beta 90.000 - gamma 90.000 - Primitive - Operations: ImportFile: Run3_low_SSR_fines.raw - Start: 5.000 ° - End: 64.999 ° - Step: 0.019 ° - Step time: 36.4 s - Anode: Cu - WL1: 1.5406 - WL2: 1.54439 - kA2 Ratio: 0.5 - Generator kV: 40 kV - Generator mA: 40 mLin (Cps)01002003004005006007002-Theta - Scale5 10 20 30 40 50 60Run 2 High SSR fines New Crystallizer01-077-2303 (*) - Struvite, syn - MgNH4PO4(H2O)6 - Y: 86.82 % - d x by: 1. - WL: 1.5406 - Orthorhombic - a 6.95500 - b 6.14200 - c 11.21800 - alpha 90.000 - beta 90.000 - gamma 90.000 - Primitive - Operations: ImportFile: Run2_high_SSR_fines.raw - Start: 5.000 ° - End: 64.999 ° - Step: 0.019 ° - Step time: 36.4 s - Anode: Cu - WL1: 1.5406 - WL2: 1.54439 - kA2 Ratio: 0.5 - Generator kV: 40 kV - Generator mA: 40 mLin (Cps)01002003004005006007008009002-Theta - Scale5 10 20 30 40 50 60  77  5.1.2.7 Loss of Fines Through Effluent Along with the measurement of phosphorus concentrations in filtered effluent samples as discussed in section 4.8, phosphorus concentrations were also measured for unfiltered samples in order to assess the loss of fines through effluent. P- concentrations in filtered and unfiltered samples have been shown in Appendix H. Differences in P-concentrations in both type of effluent samples were plotted in Figure 5.17. Differences between filtered and unfiltered samples were laid below 10 mg/L in most cases as shown in Figure 5.17 with an average value of 8.6 mg/L along with a standard deviation of 5.2 mg/L. Some extreme outliers resulted due to some operational difficulties such as clogging of the reactor or nozzles etc. These values indicate that there were some losses of fines through effluent but these losses were not substantial in most cases.  Figure 5.17: Differences in P-concentrations in unfiltered and filtered effluent samples under different supersaturation levels.    0510152025304 4.5 5 5.5 6 6.5 7 7.5 8Difference in P-concentration, mg/LSstruvite  78  5.1.2.8 Variation of Supersaturation Level Along the Length of the Reactor Efforts were taken during this study to look into the variations of supersaturation level along various points of the reactor in order to detect speed of the reaction. Figure 5.18 shows that the supersaturation ratio reached a value approximately 1 within the length of 16 inches which implies that the reaction was fast enough to reach almost equilibrium condition within a very short time. It is also implied by the Figure 5.18 that the reaction rate was not dependent on initial supersaturation condition.  Figure 5.18: Variation of Sstruvite along the length of the reactor. 5.1.3 Optimization of Supersaturation Based on the discussion form section 5.1.2, the optimum supersaturation levels for different scenarios were tabulated in Table 5.2. In the light of summarized information presented in Table 5.2 it can be predicted that an optimum influent supersaturation ratio of 5.5 to 6.0 might produce optimal P-removal and recovery along with lesser fine production. 0123456780 12 24 36 48 60 72 84SstruviteLenght from bottom of reactor (inch)(initial) (effluent)  79  Table 5.2: Optimum supersaturation level for P-removal and recovery.  Optimum Conditions P-removal efficiency Optimum Sstruvite : 5.5 to 6.5 P-recovery efficiency and pellets growth rate Higher with lower Sstruvite. Optimum Sstruvite : 4 to 6. Production of fines Higher with higher Sstruvite. Pellet Size Optimum Sstruvite : 5.5 to 6.5 Pellet shape More smooth surface with higher Sstruvite..  5.2 Impact of Hydrodynamics Variables Impacts of various hydrodynamic variables e.g. nozzle configurations and velocities, up-flow velocities (Phase-2) on struvite crystallization process have been described in subsequent sections of this chapter. A summary of the operational parameters for this phase has been tabulated in Table 5.3. During this phase, initial Sstruvite was tried to maintain within 5.5 to 6.0 as detected to be optimum in section 5.1.3, but due to operational difficulties there existed some deviations from the expected range of Sstruvite as shown in Table 5.3. Table 5.3: Summary of parameters during phase 2. Run Influent Sstruvite Up-flow Velocity (cm/min) Nozzle Velocity and Configuration Average Standard Deviation Average Standard Deviation 7 5.49 0.14 40.21 1.01 *Referred to Table 4.1 8 5.17 0.21 39.19 1.52 9 5.38 0.15 19.58 0.54 10 5.58 0.1 39.97 0.69 11 5.59 0.72 39.06 1.22 13 5.42 0.71 38.6 0.81 14 5.82 0.14 62.79 0.97 15 5.77 0.24 58.82 1.02 16 5.42 0.25 60.49 0.55   80  5.2.1 Nozzle and Up-flow Velocities Different up-flow velocities in combination with different nozzle velocities were taken into consideration during this study to establish the optimum condition for optimum P-removal and recovery.   5.2.1.1 P-Removal Efficiency Lower up-flow velocity induces higher hydraulic retention time (HRT) providing more time to complete reaction which results in higher phosphorus removal efficiency. This fact was found to be also effective for this study as moderate up-flow velocity (~ 40 cm/min) as well as lower up-flow velocity (20 cm/min) was observed to be more efficient to remove phosphorus from synthetic anaerobic digester supernatant compared to high up-slow velocity (~ 60 cm/min) as illustrated in Figure 5.19. The only exception was observed in case of high nozzle velocity where 40 cm/min up-flow velocity was observed to be less effective to remove phosphorus compared to 60 cm/min (as shown in Figure 5.19). This might happen due to the fact that moderate up-flow velocity was operated at a lower supersaturation ratio than the optimum one as illustrated in Table 5.3 Although, theoretically nozzle velocity is not associated with P-removal but this study observed moderate nozzle velocity induced higher P-removal efficiency.       81    Figure 5.19: Influence of up-flow and nozzle velocities on P-removal efficiency - (a) scatter diagram and (b) bar diagram. 606570758085909510018 23 28 33 38 43 48 53 58 63 68P-Removal Efficiency, (%)Up-flowVelocity, (cm/min)8.95-9.06 4.53 18.04Nozzle Velocity (cm/min):051015202530354045505560657075808590Moderate Up-flow VelocityHigh Up-flowVelocityLow Up-flowVelocityModerate Up-flow VelocityHigh Up-flowVelocityModerate Up-flow VelocityHigh Up-flowVelocityLow Nozzle Velocity Moderate Nozzle Velocity High Nozzle VelocityP-Removal Efficiency, (%)Up-flow velocities (cm/min):                        Low ~ 20Moderate ~ 40High ~ 60Nozzle velocities (cm/min):                        Low- 4.53Moderate-8.95-9.06High - 18.04(a) (b)   82  5.2.1.2 P-Recovery Efficiency As illustrated in Figure 5.20, lower up-flow velocity was observed to produce more pellets to accomplish higher P-recovery efficiency. Higher up-flow velocity induces more small particles to be suspended at the top of the reactor. As a result, there is a lack of crystals at bottom for agglomeration to grow as pellets which results in lower P-recovery.  Figure 5.20: Influence of up-flow and nozzle velocities on P-recovery. On the other side, higher nozzle velocities stimulated more P-recovery efficiency as suggested by Figure 5.20. Higher nozzle velocity produces more collision and more turbulence, which induces higher P-recovery as struvite pellets. But lower P-recovery efficiency was found from high nozzle velocity when the reactor was run under high up-flow velocity. This may be due to the fact that higher up-flow velocity caused more fines to wash away with effluent resulting in less crystals available to grow as pellets by agglomeration or crystal growth. This explanation can be validated 02468101214161820Moderate Up-flow VelocityHigh Up-flowVelocityLow Up-flowVelocityModerate Up-flow VelocityHigh Up-flowVelocityModerate Up-flow VelocityHigh Up-flowVelocityLow Nozzle Velocity Moderate Nozzle Velocity High Nozzle VelocityP-recovered as struvite pellets, (%)Up-flow velocities (cm/min):                        Low ~ 20 Moderate ~ 40 High ~ 60 Nozzle velocities (cm/min):                        Low- 4.53 Moderate-8.95-9.06 High - 18.04   83  by comparing P-content between unfiltered and filtered effluent samples. Figure 5.21 illustrated that difference in p-content between two types of effluent samples became more dominant when reactor was running under high up-flow velocity with high nozzle velocity which supported the explanation of washing away of fines with effluent. But this reasoning is required more validation through further in-depth research as there existed a large variation in data which might be attributed due to operational difficulties,  Figure 5.21: Differences in P-concentrations in unfiltered and filtered effluent samples under different up-flow and nozzle velocities. Pellets growth rates were also observed to be higher with lower up-flow and higher nozzle velocity (as shown in Figure 5.22).  -10-5051015202530350 10 20 30 40 50 60 70Difference in P-content , mg/LUp-flow velocity, cm/min4.538.95-9.0618.04Nozzle velocity (cm/min):  84   Figure 5.22: Pellets growth rate under different nozzle and up-flow velocities. 5.2.1.3 Pellet Size Distribution Tarragó et al. (2016) suggests that up-flow velocity has a direct connection with the size of the struvite pellets recovered. High up-flow velocities cause a large number of smaller particles to be at suspension at the top of the reactor creating lower population density of crystals at bottom, which lead to more space for crystals to grow as larger pellets. That’s why increase in up-flow velocities resulted in increase in size of pellets recovered as suggested by Figure 5.23 with an exception for high up-flow velocity under high nozzle velocity. As discussed earlier, this exception may be due to the wash way of more fines through effluent. Tarragó et al. (2016) also achieved to recover larger struvite pellets by increasing up-flow velocity. Although Fattah et al. (2012) suggested an upper limit for up-flow velocity for better quality of pellets.   0102030405060708090100110Moderate Up-flow VelocityHigh Up-flowVelocityLow Up-flowVelocityModerate Up-flow VelocityHigh Up-flowVelocityModerate Up-flow VelocityHigh Up-flowVelocityLow Nozzle Velocity Moderate Nozzle Velocity High Nozzle VelocityPellets Growth Rate (gm/day)Up-flow velocities (cm/min):                        Low ~ 20Moderate ~ 40High ~ 60Nozzle velocities (cm/min):Low- 4.53Moderate-8.95-9.06High - 18.04  85     Figure 5.23: Pellet size distribution under different up-flow and nozzle velocities. 02468101214161820ModerateUp-flowVelocityHigh Up-flowVelocityLow Up-flowVelocityModerateUp-flowVelocityHigh Up-flowVelocityModerateUp-flowVelocityHigh Up-flowVelocityLow Nozzle Velocity Moderate Nozzle Velocity High Nozzle VelocitySize > 1mm05101520ModerateUp-flowVelocityHigh Up-flowVelocityLow Up-flowVelocityModerateUp-flowVelocityHigh Up-flowVelocityModerateUp-flowVelocityHigh Up-flowVelocityLow Nozzle Velocity Moderate Nozzle Velocity High Nozzle VelocitySize: 1-0.5 mm05101520ModerateUp-flowVelocityHigh Up-flowVelocityLow Up-flowVelocityModerateUp-flowVelocityHigh Up-flowVelocityModerateUp-flowVelocityHigh Up-flowVelocityLow Nozzle Velocity Moderate Nozzle Velocity High Nozzle VelocitySize:0.5 mm to 0.25 mm  86  5.2.1.4 Fines Production Theoretical calculations yielded that increase in up-flow velocity would increase fine production as shown in Figure 5.24. The same trend was observed with the actual measurement as illustrated in Figure 5.24 with an exception for high up-flow velocity with high nozzle velocity. On the other hand, although theoretical value showed that variations in amount of fines production under different nozzle velocities were not significant where actual measurement depicted that impacts of nozzle velocities were different depending on up-flow velocities.  Increase in nozzle velocity results in increase in fine production, which was evident at up-flow velocity level around 40 cm/min but at up-flow velocity of 60 cm/min this increase in nozzle velocity resulted in decrease in fine production. Increase in up-flow velocity might induce wash away of fines through effluent as discussed earlier, which might result in reduction of fine production.  Figure 5.24: Production of fines under various up-flow and nozzle velocities. 5.2.1.5 Morphology of Pellets and Fines In order to have an in-depth understanding about the impacts of various hydrodynamic condition on morphology of pellets and fines, pellets and fines were analyzed using microscopes. According to Figure 5.25 pellets formed under high nozzle velocity (~ 18.50 cm/min) tended to have smoother 00.20.40.60.8115 35 55Nominal weight of fines (g/L)Up-flow velocity (cm/min)Actual4.53 8.95-9.06 18.04Nozzle Velocity(cm/min) :00.20.40.60.811.215 25 35 45 55 65Nominal weight of fines (g/L)Up-flow velocity (cm/min)Theoretical 4.53 8.95-9.06 18.04Nozzle Velocity(cm/min) :  87  surface compared to lower nozzle velocity. High nozzle velocity generates higher turbulence and more inter particles collisions, which induces more attritions. This phenomenon might be the reason for smoother pellets with high nozzle velocity.   Figure 5.25: Microscopic images of pellets magnified at 400X under different nozzle velocity-(a) high nozzle velocity with size > 1mm , (b) high nozzle velocity with 0.5 mm <size< 1.0 mm , (c) low nozzle velocity with size > 1mm , (d) low nozzle velocity with 0.5 mm <size < 1mm. As illustrated in Figure 5.26, influence of up-flow velocity was observed to be negligible on the shape of the pellets. High up-flow velocity might help to form pellets with a slightly smooth surface.  (a) (b) (c) (d)   88  Overall, the pellets formed under different hydrodynamic condition were found to be hard and difficult to break.  Figure 5.26: Microscopic images of pellets magnified at 400X under different up-flow velocity-(a) high up-flow velocity with size > 1mm, (b) high up-flow velocity with 0.5 mm <size< 1.0 mm, (c) low up-flow velocity with size > 1mm, (d) low up-flow velocity with 0.5 mm <size < 1mm. High nozzle velocity produced random shaped crystals as shown in Figure 5.27 but with mostly plate shaped rather than rod or needle shaped. On contrary, low nozzle velocity produced mostly rod shaped crystals. As high nozzle velocity induced more fines, so crystals might develop on all (a) (b) (c) (d)   89  sides through agglomeration of small crystals resulting in mostly plate shaped crystals. Figure 5.27 also suggested presence of a lot of small crystals under high nozzle velocity.  Figure 5.27: Microscopic images of fines magnified at 400X under different nozzle velocity. As illustrated in Figure 5.28, rod shaped with smooth edges crystals along with some traces of plate shape crystals were observed under both low and high up-flow velocity suggesting insignificant influence of up-flow velocity. Figure 5.28: Microscopic images of fines magnified at 400X under different up-flow velocity. Low Nozzle Velocity High Nozzle Velocity Low up-flow velocity High up-flow velocity   90  5.2.1.6 Variation of Supersaturation Along the Length of the Reactor Figure 5.29 depicts that the reaction approached to almost completion within very short time and the reaction rate did not depend on hydrodynamic conditions e.g. up-flow or nozzle velocity.     Figure 5.29: Variation of supersaturation level along the length of the reactor under different up-flow and nozzle velocities. 012345670 20 40 60 80SstruviteLenght from bottom of reactor (inch)Nozzle Velocity (cm/min): 8.95-9.06402063012345670 20 40 60 80SstruviteLength from bottom of the reactor (inch)Nozzle Velocity (cm/min): 18.04395901234560 20 40 60 80SstruviteLength from bottom of the reactor (inch)Nozzle Velocity (cm/min):4.534060(influent) (effluent) (influent) (effluent) Up-flow velocity (cm/min): Up-flow velocity (cm/min): Up-flow velocity (cm/min): (influent) (effluent)   91  5.2.2 Nozzle configurations Three types of nozzle configurations, as shown in Table 4.2, were taken into account in order to have in-depth insight regarding the impact of nozzle configurations on overall P-removal and recovery performance of the reactor. Figure 5.30 illustrates that P-removal efficiency did not differ significantly depending on nozzle configurations although two nozzles on opposite side was detected to accomplish slightly lower P-removal efficiency.   Figure 5.30: Impact of nozzle configuration on P-removal efficiency. On the contrary, two nozzles on both sides was observed to achieve high P-removal efficiency as depicted by Figure 5.31. Two nozzles on opposite side also produced more large-sized pellets (size > 1mm) and medium sized pellets (0.5 mm<size< 1mm) compare to other two nozzle 05101520253035404550556065707580859095100P-removal efficiency, %4-Nozzles 2-Nozzles on same side 2-Nozzles on opposite side  92  configurations. This configuration was also found to be associated with high pellets growth rate as suggested by Figure 5.32.  Figure 5.31: P-recovery efficiency and pellet size distribution under different nozzle configurations.  Figure 5.32: Pellets growth rate under various nozzle configurations. 0246810121416Size>1mm0.5mm<Size<1mm0.25mm<Size<0.5mmTotalSize>1mm0.5mm<Size<1mm0.25mm<Size<0.5mmTotalSize>1mm0.5mm<Size<1mm0.25mm<Size<0.5mmTotal4-Nozzles 2-Nozzles on same side 2-Nozzles on opposite sideP-recovery efficiency, %010203040506070809010037.5 38 38.5 39 39.5 40 40.5Pellets growth rate, gm/dayUp-flow velocity, cm/min4-Nozzles 2-Nozzles on same side 2-Nozzles on opposite side  93  Figure 5.33 illustrates that theoretically two nozzles on opposite side would produce less fines. But actual measurement suggested that four nozzles produced less fines. As discussed earlier, these discrepancies might result from losing fines during sample collection, filtration and handling. Also, according to Figure 5.34 washing away of fines through effluent was detected to be high while operating the reactor with four nozzles.   Figure 5.33: Influence of nozzle configurations on fines production. 00.10.20.30.40.50.60.70.80.914-Nozzles 2-Nozzles on same side 2-Nozzles on opposite sideNormalized weight of fines, gm/LActual Theoretical  94   Figure 5.34: Differences in P-concentrations in unfiltered and filtered effluent samples under different nozzle configurations. Shapes of fines and pellets were analyzed under microscope in order to have an understanding regarding the impact of nozzle configuration on their shapes. Two nozzles on same sides were observed to have comparatively rough surface compared to other two configurations. It is difficult to rationalize this phenomenon without in-detail research on hydrodynamic behaviour of various nozzle configurations. Nature of eddies formed by each configurations is required to be known in order to provide accurate reasoning for these outcome. 02468101214161837 38 39 40 41 42 43Difference in P-concentration, mg/LUp-flow velocities, cm/min4-Nozzles 2-Nozzles on same side 2-Nozzles on opposite side  95   Figure 5.35: Microscopic images of pellets (size > 1mm) magnified at 100X. This paper would try to provide a tentative hypothesis regarding the type of eddies formed under different nozzle configurations. Eddies formed from four nozzles might be random. On the contrary, probable eddies formed other two configurations have been depicted in Figure 5.36. Eddies formed on both sides of the reactor as the case for four nozzles and two nozzles on opposite side might contribute to smoothness of pellet surface due to attrition effect on each side of the pellets. Two nozzles on same side Two nozzles on opposite side Four nozzles    96   Figure 5.36: Eddies formed under different nozzle configurations. Crystals formed under different nozzle configurations were found to be mostly with “rod” shaped as suggested by Figure 5.37, which revealed insignificant influence of nozzle configurations on the shape of the crystals.  Like other hydrodynamic properties (up-flow and nozzle velocities), nozzle configurations also exerted no influence on supersaturation level at various points along the length of the reactor as shown in Figure 5.38 and supersaturation ratio also approached one very quickly in spite of type of nozzle configurations which also indicated that the reaction rate was independent of nozzle configurations. Two nozzles on same side Two nozzles on opposite side side   97   Figure 5.37: Microscopic images of fines magnified at 400X under different nozzle configurations.  Figure 5.38: Variation of supersaturation level along the length under various nozzle configurations. 01234560 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0SStruviteLength from the bottom of the reactor, inch4-Nozzles 2-Nozzles on same side 2-Nozzles on different side(effluent)Four nozzles Two nozzles on same side Two nozzles on opposite side (influent)   98  5.2.3 Optimization of Hydrodynamic Conditions Based on analysis presented in Section 5.2.1 and Section 5.2.2, influence of various hydrodynamic conditions has been summarized in Table 5.4. Based on this table it can be assumed that moderate up-flow velocity with nozzles on opposite side would produce optimum outcome for P- recovery and removal. Table 5.4: Optimum hydrodynamic conditions for P-removal and recovery.  Optimum Conditions Up-flow Velocity Nozzle Velocity Nozzle configuration P-removal efficiency Low up-flow velocity (20 cm/min) Not significant Not significant P-recovery efficiency and pellets growth rate Higher with lower up-flow velocity. Optimum velocity: low up-flow velocity (20 cm/min) High nozzle velocity (18.04 cm/min) Two nozzles on opposite side Production of fines Higher with higher up-flow velocity Not significant Lower with two nozzles on opposite side Pellet Size Larger pellets with higher up-flow velocity Not significant Larger pellets with two nozzles on opposite side Pellet shape Not significant Smoother surface with high nozzle velocity Smoother surface with two nozzles on opposite side    99  CHAPTER 6: SCALING UP OF THE REACTOR During this study one run was carried out with an up scaled reactor to assess the feasibility of accomplishing optimum performance by the up-scaled reactor by running the enlarged reactor under the optimum conditions set up by pilot scale reactor. This chapter gives an insight of the operating conditions and relevant outcome from the up-scaled reactor. Up-scaled reactor consists of three segment, which is similar to pilot scaled reactor- injection port, reaction zone and clarifier. The reaction zone was made of PVC pipe with a 5 inch diameter and 48.94 inch length (as shown in Figure 6.1). Total volume of this zone was found to be 15.75 L (shown in Appendix A) with a HRT of 3 minutes. At the bottom of the zone, a harvesting device was installed and recycle line was connected with this harvesting device. Recycle flow could be adjusted in order to harvest desired sized pellets. Reynolds number was calculated to be 975 for the reaction zone. Clarifier of up-scaled reactor also consisted of two sections- a cylindrical section and a conical section (as illustrated in Figure 6.1).  Total volume of this zone was about 421.4 L with HRT around 81 minutes. Reynolds numbers for cylindrical and conical section were calculated to be around 256 and 9 respectively. There were two outlets at the top of the clarifier to discharge as indicated in Figure 6.1.   100   Figure 6.1: Details drawing of FBR struvite crystallizer. Based on the results obtained from pilot scale reactor, the operating conditions were selected as illustrated in Table 6.1 for up-scaled reactor.    101  Table 6.1: Operating conditions for up-scaled reactor. Up-flow Velocity (cm/min) Nozzle Velocity (cm/min) Nozzle Configuration Recycle Flow (L/min) Average Standard Deviation Average Standard Deviation 41.1 1.02 18.04 Two on opposite sides 0.53 0.13 Table 6.2 tabulates a summary of synthetic supernatant characteristics for this run. From Table 6.2, it is apparent that phosphorus concentrations were observed to be a little bit lower compared to phosphorus concentration during phase I (151.65 mg/L) and phase II (159.36 mg/L) of this study (as shown in Table 4.1) which indicated that dilution of concentrated feed was a little bit higher than expected. Due to this over dilution, concentrations of magnesium and ammonia were also found to be a little lower compared to other phases with small scale reactor.  Table 6.2: Synthetic supernatant characteristics for up-scaled reactor.  Average 95% Confidence Interval Maximum Minimum Phosphorus (mg/L) 137.83 ± 2.58 144.5 133 Ammonia (mg/L) 614.17 ± 12.94 655 585 Magnesium (mg/L) 111.11 ± 2.09 117.38 106.33 Temperature (°C) 24.53 ± 0.28 25 23.7 Conductivity (µS/cm) 9.91 ± 0.16 10.52 9.71   102  Figure 6.2 represents the variation of P-removal efficiency along with influent supersaturation ratio and pH (effluent) for the enlarged reactor. As suggested by Figure 6.2, the up-scaled reactor was run under a lower initial Sstruvite (with average Sstruvite of 4.86) compared to the optimum value (5.5-6.0) determined in section 5.3. This was due to over dilution of concentrated feed. In spite of this lower supersaturation level, the reactor accomplished as much as about 88% of P-removal efficiency with initial supersaturation ratio of 5-5.2 where the small reactor achieved about 80% to 90% of P-removal efficiency when operated at initial Sstruvite of 5.5 to 6.5 as suggested by Figure 5.6. This reactor also confirmed the trend of increment of P-removal efficiency with increase in Sstruvite within a value of 4.5 to 5.2 of Sstruvite . Influence of effluent pH on P- removal efficiency was found to be identical with pilot scale reactor as Figure 5.3 suggested that around 90% P-removal was accomplished by pilot scale reactor with effluent pH values around 8.  Figure 6.2: P-removal efficiency of up-scaled reactor. 024681012146 23 30 47.5 53.5 71 78 101.75505560657075808590Sstruvite(initial)pH (effluent)Time, hoursP-removal efficiency, %P-removal efficiency S pH  103  Another significant finding from this run was that up-scaled reactor was capable of obtaining higher P-removal efficiency compared to small reactor when operated under high nozzle velocity, and nozzles on opposite side even with low nozzle velocity as Figure 5.19, Figure 5.30 and Table 5.3 suggest that high nozzle velocity (Sstruvite was 5.17) was capable of achieving about 72% removal while nozzles on opposite side (Sstruvite was 5.42) achieved about 74% removal. Although having a designated harvesting system (as shown in Figure 6.1), this reactor also detected to achieve lower P-recovery efficiency as shown in Figure 6.3. Figure 6.3 also suggests that most of the pellets produced were small sized (0.5 mm <size<0.25 mm) along with no production of larger sized pellets (size > 1 mm). Several factors might contribute to these outcomes. One of the reasons might be due to low initial Sstruvite , which might lead to lower crystal growth rates and hence production of smaller particles as Figure 5.10 also suggests that lower Sstruvite was responsible for production of small sized pellets. Another point is to be noted that high nozzle velocity and nozzles on opposite side were detected to be optimum condition for pilot scale reactor, but these two conditions were not stimulated together. Different hydrodynamic scenarios might develop when these two conditions simultaneously operated which might be the case for enlarged reactor. On the other hand, hydrodynamic characteristics of up-scaled reactor might also be different from the small reactor. So the optimum conditions set up by operating pilot scale reactor might not accomplish optimum outcome for enlarged reactor.   104   Figure 6.3: P-recovery efficiency for up-scaled reactor.  The run with enlarged reactor produced 0.5842 gm/L of fines where theoretical calculation yielded 0.862261 gm/L of fines. These values were identical with the values generated by lower Sstruvite with pilot scale reactor as indicated by Figure 5.10. These values are also very close to the values generated by high nozzle velocity with moderate up-flow velocity and also with nozzle on opposite side as suggested by Figure 5.24 and Figure 5.33 respectively. Figure 6.4 depicts microscopic images of pellets and fines produced from up-scale reactor. Pellets were observed to be slightly elongated in shape with smooth surface. This shape of the pellets was found to be very similar with the pellets produced from run 8 (high nozzle velocity) and run 13 (nozzles on opposite side) as suggested by Figure 5.25 and Figure 5.35 respectively. On the other hand, although lower Sstruvite led to formation of rough surfaced pellets as shown in Figure 5.11, but up-scaled reactor produced smooth pellets under lower Sstruvite. This discrepancy might result from the effect of high nozzle velocity and nozzle configuration. Crystals formed in enlarged reactor were detected to be plate shaped as shown in Figure 6.4 which also identical with lower 024681012Size > 1mm 1 mm < Size <0.5mm0.5 mm < Size<0.25 mmTotalP-recoverey effieicency as struvite, %  105  initial Sstruvite and high nozzle velocity operation with pilot scaled reactor as illustrated in Figure 5.12 and 5.27.  Figure 6.4: Microscopic images of pellets (magnified at 100X) and fines (magnified at 400X). As illustrated by Figure 6.5, differences in P-concentration between unfiltered and filtered samples were below 2 mg/L in most cases with an average value of 1.57 mg/L, where run 8 (high nozzle velocity) produced more than 5 mg/L difference (Figure 5.21), and run 13 (nozzles on different side) produced an average 3.5 mg/L difference in P-concentration (Figure 5.34). So, losses of fines through effluent were very insignificant for enlarged reactor.     106   Figure 6.5: Difference in P-concentration between unfiltered and filtered effluent samples from up-scaled reactor. In spite of initial supersaturation level, effluent supersaturation always approached to almost one as depicted by Figure 6.5.    Figure 6.6: Initial and effluent supersaturation ratio for up-scaled reactor.  0510152025303540450 20 40 60 80 100 120P-concentration, mg/LTime, hoursEffluent (Unfiltered) Effluent (Filtered) Difference01234560 20 40 60 80 100 120SStruviteTime, hoursInitial Effluent  107  CHAPTER 7: CONCLUSIONS AND RECOMMENDATIONS 7.1 Conclusions  Based on this study on P-removal and recovery efficiency from synthetic anaerobic digester supernatant using both pilot scale reactor and up-scaled reactor, a number of conclusions can be drawn. These conclusions are discussed in this section. 7.1.1 Pilot Scale Reactor Several conclusions can be derived from the analysis of performance of pilot scale reactor under various operating conditions, which have been described below.  The pilot scale reactor was effective in removing and recovering phosphorus from synthetic supernatant (with 93.7 – 179 mg/L concentration of P).  Over 90% of P was removed successfully from supernatant along with upto 18% of P was recovered as struvite pellets after harvesting, drying and screening. The concentration of phosphorus in filtered effluent samples was found to be as low as 9.8 mg/L. Lack of proper harvesting mechanism led to relatively lower P-recovery efficiency for pilot scale reactor. Full-scale installation with proper harvesting and recovery technology is expected to have better P-recovery efficiency.   The initial supersaturation ratio was considered to be effective physio-chemical parameter to control the crystallization process. Desired Sstruvite was maintained by adjusting pH through the flow of caustic. Various hydrodynamics properties such as up-flow velocity, nozzle flow and configuration were studied to determine the optimum condition.    108   Increase in pH resulted in increase in P-removal from supernatant. But after pH 8, increment in P-removal became negligible with increase in pH and 90% of phosphorus was removed operating the reactor within pH 8.  P-removal efficiency was observed to increase with increase in Sstruvite up to a Sstruvite value of 6.5. Sstruvite value of 5.5 to 6.5 was detected to be optimum for phosphorus removal. 92% P-removal efficiency was accomplished by operating the reactor below Sstruvite 6.5. On the other hand, P-recovery efficiency was found to be decreased with increase in Sstruvite which might be due to higher nucleation rate compared to crystal growth rate. An optimum value of 4 to 6 of Sstruvite was observed to be effective for phosphorus recovery. Similarly, struvite pellets growth rate was found to be increased with decrease in Sstruvite with an optimum value of 4 to 6. However, production of fines was increased by raising Sstruvite. Moreover, larger pellets with smoother surface were also produced under higher Sstruvite. Pellets produced under different supersaturation level were very hard to break and mostly pure with presence of small amount of cattite and kovdorskite under higher Sstruvite. Furthermore, needle shaped crystals were developed under higher Sstruvite while plate shaped crystals were observed with lower Sstruvite. In order to achieve a balance between higher P- removal and recovery along with lower production of fines, an optimum value of 5.5 to 6.0 of Sstruvite was suggested in this study for an efficient operation.  Higher Up-flow velocity yielded lower P-removal efficiency due to lower HRT. Not only P-removal efficiency but also P-recovery efficiency and pellets growth rate were observed to be higher with lower up-flow velocity. On the contrary, higher up-flow velocity was associated with the development of larger struvite pellets. However,   109  higher up-flow velocity also produced more fines compared to lower up-flow velocity. On the other hand, up-flow velocity exerted no impact on pellet and crystal shape. Higher nozzle velocity yielded higher P-removal efficiency along with higher P-recovery efficiency as well as higher pellets growth rate. Fines production under different nozzle velocity yielded different scenarios based on operating up-flow velocity. At 40 cm/min up-flow velocity, 18.04 cm/min nozzle velocity produced more fines where 4.53 cm/min nozzle velocity was responsible for producing more fines at higher up-flow velocity (60 cm/min). Again, higher nozzle velocity was noticed to develop smoother surface for pellets. Furthermore, plate shaped crystals were formed under high nozzle velocity where lower nozzle velocity produced rod shaped crystals. So, in order to develop an optimum and efficient P-recovery technology, a moderate up-flow velocity (40 cm/min) with higher nozzle velocity (18.04 cm/min) was proposed by this study. This study suggested that nozzle configuration did not have significant influence on P-removal efficiency. But, two nozzles on opposite side was found to be responsible for higher P-recovery efficiency along with production of larger pellets. Higher pellet growth rate was also found to be developed under this configuration.  Furthermore, less fines produced by this configuration. Again, nozzles on opposite side were also found to develop smoother pellets.  Therefore, in order to develop an optimum and efficient P-recovery technology, a moderate up-flow velocity (40 cm/min) with higher nozzle velocity (18.04 cm/min) in combination with two nozzles on different sides were proposed by this study. One point is to be noted here, this study did not carry out any run with a combination of moderate up-flow velocity and higher nozzle velocity along with two nozzles on   110  opposite side which might yield slightly different result due to different hydrodynamics. 7.1.2 Up-Scaled Reactor Although only one run was conducted by up-scaled reactor and due to some operational difficulties failure resulted in maintain the optimum conditions, a number outlines can be made based on the knowledge gained by this one run.  Efforts were made to run the up-scaled reactor under the optimum conditions set up by the pilot scale reactor such as moderate up-flow velocity (~ 40 cm/min), high nozzle velocity (~ 18.04 cm/min) and 5.5 to 6.0 Sstruvite.  Although up-scaled reactor was run at a lower struvite (Average Sstruvite = 4.86) compared to optimum range due to over dilution of concentrated feed, this reactor was able to accomplish up to 88% P-removal with an average removal efficiency of about 83%.  This reactor succeeded to recover only just over 10% of P as struvite pellets. This low recovery efficiency might be due to the fact that hydrodynamic characteristics of up-scaled reactor is differed from the pilot scale reactor and that’s why optimum hydrodynamic condition set up by pilot scale reactor might not be optimum for enlarged reactor. Again, most of the pellets recovered were small sized (0.5 mm <size < 0.25 mm) which might be due to low initial Sstruvite.   Amount of fines produced and crystal shapes developed from enlarged reactor were identical with pilot scale reactor.  Although lower initial Sstruvite led to the formation of rough surfaced pellets for pilot scale reactor, but up-scaled reactor produced smooth surfaced pellets which might be due to the action of high nozzle velocity and nozzle configuration.   111  7.2 Recommendations A list of recommendations is made based on the analysis performed in this study.  This study was carried out on synthetic wastewater in order to assess the performance of the reactor. As P-recovery performance through struvite crystallization is site specific with respect to feed characteristics such as TSS, presence of ions etc. (Adnan, 2002), it is required to analyze of the performance of this reactor when it is applied in the actual wastewater treatment plant settings.   A better understanding on hydrodynamics within the reactor is required to have better control on struvite crystallization process. Knowledge on eddies formed under different hydrodynamic conditions, flow patterns, shear forces will be very significant for an efficient operation.  Further investigations on nucleation, crystal growth and agglomeration process will provide better understanding on crystallization.  Impact of temperature and ionic strength on struvite crystallization process is recommended to study to have better knowledge.           112  REFERENCES Aage, H.K., Andersen, B.L., Blom, A., Jensen, I., 1997. The solubility of struvite. J. Radioanal. Nucl. Chem. 223, 213–215. doi:10.1007/BF02223387 Abbona, F., Boistelle, R., 1979. Growth morphology and crystal habit of struvite crystals (MgNH4PO4 ?? 6 H2O). J. Cryst. Growth 46, 339–354. doi:10.1016/0022-0248(79)90082-4 Abe, S., 1995. Phosphate removal from dewatering filtrate by MAP process at Seibu treatment plant in Fukuoka City. Sew. Work. Japan 43, 59–64. Adnan, A., 2002. Pilot-Scale Study of Phosphorus Recovery Through Struvite Crystallization. The University of British Columbia, Vancouver,BC, Canada. Adnan, A., Mavinic, D.S., Koch, F.A., 2003. Pilot-scale study of phosphorus recovery through struvite crystallization — II: Applying in-reactor supersaturation ratio as a process control parameter. J. Environ. Eng. Sci. 2, 473–483. doi:10.1139/s03-048 Ali, M.I., Schneider, P.A., Hudson, N., 2013. Thermodynamics and solution chemistry of struvite. J. Indian Inst. Sci. 85, 141. APHA, AWWA, WEF, 2005. Standard Methods for the Examination of Water and Wastewater. American Public Health Association, Washington, D.C. Ashley, K., Cordell, D., Mavinic, D., 2011. A brief history of phosphorus: from the philosopher’s stone to nutrient recovery and reuse. Chemosphere 84, 737–746. Battistoni, P., De Angelis, A., Pavan, P., Prisciandaro, M., Cecchi, F., 2001. Phosphorus removal from a real anaerobic supernatant by struvite crystallization. Water Res. 35, 2167–2178. Battistoni, P., Fava, G., Pavan, P., Musacco, A., Cecchp, F., 1997. Research Note Phosphate Removal in Anaerobic Liquors By Struvite Crystallization Without Addition of Chemicals : Preliminary Results 31, 2925–2929. Battistoni, P., Pavan, P., Cecchi, F., Mata-Alvarez, J., 1998. Phosphate removal in real anaerobic supernatants: modelling and performance of a fluidized bed reactor. Water Sci. Technol. 38, 275–283. Baur, R., Cullen, N., Laney, B., 2011. Nutrient Recovery: One Million Pounds Recovered-With Benefits. Proc. Water Environ. Fed. 2011, 223–235. Benisch, M., Baur, R., Britton, A., Neethling, J.B., Oleszkiewicz, J.A., 2009. Startup of the First Commercial Phosphorus Recycling Facility in the US at Durham AWWTP. Proc. Water Environ. Fed. 2009, 102–119.   113  Bergmans, B., 2011. Struvite recovery from digested sludge. TUDelft. Delft University of Technology, Delft, Netherlands. Bhuiyan, M.I.H., 2007. Investigation into Struvite Solubility, Growth and Dissolution Kinetics in the Context of Phosphorus Recovery from Wastewater. The University of British Columbia, Vancouver,BC, Canada. Bhuiyan, M.I.H., Mavinic, D.S., Koch, F.A., 2008. Phosphorus recovery from wastewater through struvite formation in fluidized bed reactors: A sustainable approach. Water Sci. Technol. 57, 175–181. doi:10.2166/wst.2008.002 Bhuiyan, M.I.H., Mavinic, D.S., Koch, F.A., 2008b. Thermal decomposition of struvite and its phase transition. Chemosphere 70, 1347–1356. doi:10.1016/j.chemosphere.2007.09.056 Boistelle, R., Abbona, F., Madsen, H.E.L., 1983. On the transformation of struvite into newberyite in aqueous systems. Phys. Chem. Miner. 9, 216–222. Booker, N. a., Priestley,  a. J., Fraser, I.H., 1999. Struvite Formation in Wastewater Treatment Plants: Opportunities for Nutrient Recovery. Environ. Technol. 20, 777–782. doi:10.1080/09593332008616874 Bouropoulos, N.C., Koutsoukos, P.G., 2000. Spontaneous precipitation of struvite from aqueous solutions. J. Cryst. Growth 213, 381–388. doi:10.1016/S0022-0248(00)00351-1 Bowers, K.E., Westerman, P.W., 2005. Performance of cone-shaped fluidized bed struvite crystallizers in removing phosphorus from wastewater. Trans. ASAE 48, 1227–1234. British Columbia Ministry of Environment, 2003. Ambient Freshwater and Effluent Sampling, Field Sampling Manual : Part E -Water and Wastewater Sampling. Victoria, British Columbia, Canada. Britton, A., Prasad, R., Balzer, B., Cubbage, L., 2009. Pilot testing and economic evaluation of struvite recovery from dewatering centrate at HRSD’s Nansemond WWTP, in: Proceedings of the International Conference on Nutrient Recovery from Wastewater Streams. pp. 193–202. Britton, A., Sacluti, F., Oldham, W.K., Mohammed, A., Mavinic, D.S., Koch, F.A., 2007. Value from waste: struvite recovery at the city of Edmontons Gold Bar WWTP. Proc. IWA Spec. Conf. Mov. Forw. - Wastewater biosolids Sustain. 575–581. Buchanan, J.R., Mote, C.R., Robinson, R.B., 1994. Thermodynamics of struvite formation. Am. Soc. Agric. Eng. 3, 617–621. doi:10.13031/2013.28121 Çelen, I., Buchanan, J.R., Burns, R.T., Robinson, R.B., Raman, D.R., 2007. Using a chemical equilibrium model to predict amendments required to precipitate phosphorus as struvite in liquid swine manure. Water Res. 41, 1689–1696.   114  Chirmuley, D.G., 1994. Struvite Precipitation in WWPTS: Causes and Solutions. Water-melb. THEN ARTARMON- 21, 21. Cordell, D., 2016. The Story of Phosphorus: 7 reasons why we need to rethink the management of phosphorus resources in the global food system | Phosphorus Futures . Inst. Sustain. Futur. Univ. Technol. Sydney Aust. URL http://phosphorusfutures.net/the-phosphorus-challenge/the-story-of-phosphorus-8-reasons-why-we-need-to-rethink-the-management-of-phosphorus-resources-in-the-global-food-system/ (accessed 5.6.16). Cordell, D., Drangert, J., White, S., 2009. The story of phosphorus : Global food security and food for thought 19, 292–305. doi:10.1016/j.gloenvcha.2008.10.009 Cornel, P., Schaum, C., 2009. Phosphorus recovery from wastewater: Needs, technologies and costs. Water Sci. Technol. 59, 1069–1076. doi:10.2166/wst.2009.045 Dempsey, B.A., 1998. Removal and reuse of ammonia and phosphate by precipitation of struvite, in: Proceedings of the Industrial Waste Conference-Purdue University. Mosby Year Book Europe Ltd, pp. 369–376. Desmidt, E., Ghyselbrecht, K., Zhang, Y., Pinoy, L., Van der Bruggen, B., Verstraete, W., Rabaey, K., Meesschaert, B., 2015. Global Phosphorus Scarcity and Full-Scale P-Recovery Techniques: A Review. Crit. Rev. Environ. Sci. Technol. 45, 336–384. doi:10.1080/10643389.2013.866531 Doyle, J.D., Parsons, S.A., 2002. Struvite formation, control and recovery. Water Res. 36, 3925–3940. Driver, J., Lijmbach, D., Steen, I., 1999. Why Recover Phosphorus for Recycling, and How? Environ. Technol. 20, 651–662. doi:10.1080/09593332008616861 Durrant, A.E., Scrimshaw, M.D., Stratful, I., Lester, J.N., 1999. Review of the feasibility of recovering phosphate from wastewater for use as a raw material by the phosphate industry. Environ. Technol. 20, 749–758. Fang, C., Zhang, T., Jiang, R., Ohtake, H., 2016. Phosphate enhance recovery from wastewater by mechanism analysis and optimization of struvite settleability in fluidized bed reactor. Sci. Rep. 6, 32215. Fattah, K.P., Mavinic, D.S., Koch, F. a., 2012. Influence of Process Parameters on the Characteristics of Struvite Pellets. J. Environ. Eng. 138, 466. doi:10.1061/(ASCE)EE.1943-7870.0000576 Fattah, K.P., Mavinic, D.S., Koch, F. a, Jacob, C., 2008. Determining the feasibility of phosphorus recovery as struvite from filter press centrate in a secondary wastewater treatment plant. J. Environ. Sci. Health. A. Tox. Hazard. Subst. Environ. Eng. 43, 756–764. doi:10.1080/10934520801960052   115  Forstner, G., 2015. AirPrexTM: Biosolids Treatment Optimization Process with the option of Phosphate Recovery, in: MWEA -Annual Biosolids Conference, Michigan. USA. Fujimoto, N., Mizuochi, T., Togami, Y., 1991. Phosphorus fixation in the sludge treatment system of a biological phosphorus removal process. Water Sci. Technol. 23, 635–640. Gadekar, S., Pullammanappallil, P., 2010. Validation and applications of a chemical equilibrium model for struvite precipitation. Environ. Model. Assess. 15, 201–209. Galbraith, S.C.S.C., Schneider, P.A.P.A., 2009. A review of struvite nucleation studies. Int. Conf. Nutr. Recover. from Wastewater Streams. Vancouver, Canada IWA Publ. 69–78. Howorth, C., Wirtel, S., 2015. Nutrient Recovery Delivers Economic and Environmental Sustainability , and Helps You Meet Effluent and Biosolids Nutrient Permits. Vancouver,BC, Canada. Huang, H., 2003. Pilot Scale Phosphorus Recovery from Anaerobic Digester Supernatant. The University of British Columbia, Vancouver, BC, Canada. Huang, H., Mavinic, D.S., Lo, K. V, Koch, F. a, 2006. Production and basic morphology of struvite crystals from a pilot-scale crystallization process. Environ. Technol. 27, 233–245. doi:10.1080/09593332708618637 Jaffer, Y., Clark, T.A., Pearce, P., Parsons, S.A., 2002. Potential phosphorus recovery by struvite formation. Water Res. 36, 1834–1842. Jardin, N., Pöpel, H.J., 1994. Phosphate release of sludges from enhanced biological P-removal during digestion. Water Sci. Technol. 30, 281–292. Johnston, A.E., Steén, I., 2000. Understanding Phosphorus and Its Use in Agriculture. Brussels, Belgium. Jones, A.G., 2002. Crystallization Process Systems, First. ed. Butterworth-Heinemann, Oxford, UK. doi:10.1016/B978-075065520-0/50007-9 Kataki, S., West, H., Clarke, M., Baruah, D.C., 2015. Phosphorus recovery as struvite from farm, municipal and industrial waste: Feedstock suitability, methods and pre-treatments. Waste Manag. doi:10.1016/j.wasman.2016.01.003 Kofina, A.N., Koutsoukos, P.G., 2005. Spontaneous precipitation of struvite from synthetic wastewater solutions. Cryst. Growth Des. 5, 489–496. doi:10.1021/cg049803e Kumashiro, K., Ishiwatari, H., Nawamura, Y., 2001. A pilot plant study on using seawater as a magnesium source for struvite precipitation, in: Second International Conference about P-Recovery–submitted. Noodwijikerhout, Holland.   116  Lahav, O., Telzhensky, M., Zewuhn, A., Gendel, Y., Gerth, J., Calmano, W., Birnhack, L., 2013. Struvite recovery from municipal-wastewater sludge centrifuge supernatant using seawater NF concentrate as a cheap Mg(II) source. Sep. Purif. Technol. 108, 103–110. doi:10.1016/j.seppur.2013.02.002 Le Corre, K.S., 2006. Understanding Struvite Crystallization and Recovery. Cranfield University, Cranfield, UK. Le Corre, K.S., Valsami-Jones, E., Hobbs, P., Parsons, S.A., 2009. Phosphorus Recovery from Wastewater by Struvite Crystallization: A Review, Critical Reviews in Environmental Science and Technology. doi:10.1080/10643380701640573 Le Corre, K.S., Valsami-Jones, E., Hobbs, P., Parsons, S.A., 2005. Impact of calcium on struvite crystal size, shape and purity. J. Cryst. Growth 283, 514–522. Lee, E.Y., Oh, M.H., Yang, S.H., Yoon, T.H., 2015. Struvite crystallization of anaerobic digestive fluid of swine manure containing highly concentrated nitrogen. Asian-Australasian J. Anim. Sci. 28, 1053–1060. doi:10.5713/ajas.14.0679 Linnikov, O.D., 2008. Mechanism of aggregation and intergrowth of crystals during bulk crystallization from solutions. Cryst. Res. Technol. 43, 1268–1277. doi:10.1002/crat.200800176 Linnikov, O.D., 2004. Spontaneous crystallization of potassium chloride from aqueous and aqueous-ethanol solutions Part 2: Mechanism of aggregation and coalescence of crystals. Cryst. Res. Technol. 39, 529–539. Lobanov, S., 2009. Sewage treament to remove ammonium ions by struvite precipitation, in: Ashley, K., Mavinic, D.S., Koch, F. (Eds.), International Conference on Nutrient Recovery from Wastewater Streams. IWA publishing, Vancouver, British Columbia, Canada, pp. 579–590. Lobanov, S., Koch, F., Mavinic, D., 2014. Nutrient recovery from wastewater streams by crystallisation. J. Environ. Eng. Sci. 0, 1–4. doi:10.1680/jees.13.00008 Lobanov, S., Koch, F., Mavinic, D., 2013. The implication of aqueous equilibrium modelling to evaluate the potential for nutrient recovery from wastewater streams. WEF/IWA Nutr. Remov. Recover. Trends Resour. Recover. Use. (USB proceedings). Ma, N., Rouff, A.A., Phillips, B.L., 2014. A 31P NMR and TG/DSC-FTIR investigation of the influence of initial pH on phosphorus recovery as struvite. ACS Sustain. Chem. Eng. 2, 816–822. doi:10.1021/sc4004746 Matynia, A., Koralewska, J., Wierzbowska, B., Piotrowski, K., 2006. The Influence of Process Parameters on Struvite Continuous Crystallization Kinetics. Chem. Eng. Commun. 6445, 37–41. doi:10.1080/009864490949008   117  Mavinic, D.S., Koch, F. a, Huang, H., Lo, K. V, 2007. Phosphorus recovery from anaerobic digester supernatants using a pilot-scale struvite crystallization process. J. Environ. Eng. Sci. 6, 561–571. doi:10.1139/S07-007 Metcalf and Eddy, 2003. Wastewater engineering: treatment and reuse. McGraw-Hill, Boston, USA. Mihelcic, J.R., Fry, L.M., Shaw, R., 2011. Global potential of phosphorus recovery from human urine and feces. Chemosphere 84, 832–839. doi:10.1016/j.chemosphere.2011.02.046 Moerman, W., Carballa, M., Vandekerckhove, A., Derycke, D., Verstraete, W., 2009. Phosphate removal in agro-industry: Pilot- and full-scale operational considerations of struvite crystallization. Water Res. 43, 1887–1892. doi:10.1016/j.watres.2009.02.007 Mullin, J.W., 2001. Crystallization, Fourth. ed. Butterworth-Heinemann, Oxford, UK. Mullin, J.W., 1993. Crystallization, Third. ed. Butterworth-Heineman, Oxford, UK. Munch, E. V., Barr, K., 2001. Controlled Struvite Crystallisation for Removing Phosphorus From Anaerobic Digester Sidestreams. Water Res. 35, 151–159. Neethling, J.B., Benisch, M., 2004. Struvite control through process and facility design as well as operation strategy. Water Sci. Technol. 49, 191–199. Ohlinger, B.K.N., Member, S., Young, T.M., Member, A., Schroeder, E.D., 1999. Kinetics Effects on Preferential Struvite Accumulation in Wastewater. J. Environ. Eng. 125, 730–737. Ohlinger, K., 1999. Struvite Controls in Anaerobic Digestion and Post-digestion Wastewater Treatment Processes. University of California - Davis,U.S. Ohlinger, K.N., Young, T.M., Schroeder, E.D., 1998. Predicting struvite formation in digestion. Water Res. 32, 3607–3614. doi:10.1016/S0043-1354(98)00123-7 Oleszkiewicz, J.A.N., 2015. Options For Improved Nutrient Removal And Recovery From Municipal Wastwater in the Canadian Context, Canada. OSTARA, 2009. Case study:durham Advanced WWTP clean water services. Vancouver, British Columbia, Canada. Paques, 2016. PHOSPAQ. URL http://en.paques.nl/products/other/phospaq (accessed 6.17.16). Parkhurst, D.L., Appelo, C.A.J., 1999. User’s Guide To PHREEQC (version 2) — a Computer Program for Speciation, and Inverse Geochemical Calculations. Exch. Organ. Behav. Teach. J. D, 326. doi:Rep. 99-4259   118  Paul, E., Laval, M.L., Sperandio, M., 2016. Excess Sludge Production and Costs Due to Phosphorus Removal. Environ. Technol. 3330. doi:10.1080/09593332208618195 Phillips, H.M., Kobylinski, E., Barnard, J., Eng, P., Wallis-lage, C., 2006. Nitrogen and Phosphorus-Rich Sidestreams: Managing the Nutrient Merry-Go-Round. Water Environ. Found. 5282–5304. Pitman, A.R., Deacon, S.L., Alexander, W. V., 1991. The thickening and treatment of sewage sludges to minimize phosphorus release. Water Res. 25, 1285–1294. doi:10.1016/0043-1354(91)90069-3 Polak, A.F., 1966. Hardening of Mono-Mineral Astringent Substances. Publ. Lit. Constr. Moscow. P-Rex, 2015. AirPrex® Struvite crystallization in sludge. URL http://p-rex.eu/uploads/media/PREX_Factsheet_AIRPREX.pdf (accessed 6.17.16). Punmia, B.C., Jain, A.K., Jain, A.K., 1998. Waste Water Engineering. Firewall Media. Roberts, S., 2016. What Is the Main Function of Phosphorus in the Body? | Healthy Eating | SF Gate. URL http://healthyeating.sfgate.com/main-function-phosphorus-body-5789.html (accessed 5.11.16). Shu, L., Schneider, P., Jegatheesan, V., Johnson, J., 2006. An economic evaluation of phosphorus recovery as struvite from digester supernatant. Bioresour. Technol. 97, 2211–2216. doi:10.1016/j.biortech.2005.11.005 Sohnel, O., Garside, J., 1992. Precipitation Basic principles and industrial applications, First Edit. ed. Butterworth-Heincmann Ltd, Linacre House, Jordan Hill, Oxford OX2 8DP. Statistics Canada, 2015. Production of nitrogen and phosphorus from livestock manure, 2006. URL http://www.statcan.gc.ca/pub/16-002-x/2009001/article/10821-eng.htm (accessed 5.24.16). Tarragó, E., Puig, S., Ruscalleda, M., Balaguer, M.D., Colprim, J., 2016. Controlling struvite particles’ size using the up-flow velocity. Chem. Eng. J. 302, 819–827. doi:10.1016/j.cej.2016.06.036 Taylor, A.W., Frazier, A.W., Gurney, E.L., 1963. Solubility products of magnesium ammonium and magnesium potassium phosphates. Trans. Faraday Soc. 59, 1580–1584. Ueno, Y., Fujii, M., 2001. Three years experience of operating and selling recovered struvite from full-scale plant. Environ. Technol. 22, 1373–1381. doi:10.1080/09593332208618196 Vaganov, V.P., Amelina, E.A., Yusupov, R.K., Shchukin, E.D., Rebinder, P.A., 1974. Experimental Study of Crystallization-Contact Formation in Intergrowth of Separate   119  Crystals. Colloid J. USSR 36, 400–405. Valsami-Jones, E., 2004. Phosphorus in environmental technologies: Principles and applications. IWA Publishing, London, UK. Wang, J., Burken, J.G., Zhang, X., 2006. Effect of seeding materials and mixing strength on struvite precipitation. Water Environ. Res. 78, 125–132. doi:10.2175/106143005X89580 Wang, J., Song, Y., Yuan, P., Peng, J., Fan, M., 2006b. Modeling the crystallization of magnesium ammonium phosphate for phosphorus recovery. Chemosphere 65, 1182–1187. Woods, N.C., Sock, S.M., Daigger, G.T., 1999. Phosphorus recovery technology modeling and feasibility evaluation for municipal wastewater treatment plants. Environ. Technol. 20, 663–679. Ye, Z., Lou, Y., 2016. A comprehensive understanding of saturation index and upflow velocity in a pilot-scale fluidized bed reactor for struvite .Powder Technol. doi:10.1016/j.powtec.2016.03.022              120  APPENDIX A: VOLUME CALCULATION Pilot Scale Reactor 1. Reaction Zone Volume of Reaction zone = 𝜋 × 𝑟2 × ℎ where, r = 0.75” and h = 47”                                            = 83.056 inch3                                            = 0.0164 X 83.056 = 1.36 L    2. Clarifier Volume of clarifier = volume of conical section + volume of pail section =  13× 𝜋 × (𝑟2 + 𝑟𝑟𝑖𝑛 +  𝑟𝑖𝑛2) × ℎ +13× 𝜋 × (𝑟12 + 𝑟1𝑟2 +  𝑟22) × ℎ1 = 1906.7 𝑖𝑛3                                                                                                                                           = 31.3 𝐿  where r = 5”, rin = 0.75”, h = 8”, r1= 5”,r2= 6.115”, h1= 17.0625” Up-scaled Reactor 1. Reaction Zone Volume of Reaction zone = 𝜋 × 𝑟2 × ℎ where, r = 2.5” and h = 48.9375”                                            = 960.64 inch3                                            = 0.0164 X 960.64 = 15.75 L 2. Clarifier Volume of clarifier = volume of conical section + volume of pail section = 𝜋 × 𝑟2 × ℎ +13× 𝜋 × (𝑟12 + 𝑟1𝑟2 +  𝑟22) × ℎ1 = 25694.3 𝑖𝑛3 = 421.4 𝐿        where r = 19.25” h = 13.375”, r1= 2.5”,r2= 19.25”, h1= 22.75”   121  APPENDIX B: NOZZLE VELOCITY CALCULATION Nozzle Velocity = 𝐹𝑙𝑜𝑤 𝑁𝑜𝑧𝑧𝑙𝑒 𝐶𝑟𝑜𝑠𝑠 𝑠𝑒𝑐𝑡𝑖𝑜𝑛𝑎𝑙 𝐴𝑟𝑒𝑎∗𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑁𝑜𝑧𝑧𝑙𝑒𝑠  where nozzle cross sectional area, A = 𝜋4 × (𝐷𝑖𝑎𝑚𝑒𝑡𝑒𝑟)2 Table B.1: Nozzle Velocity for different configurations. Flow, ml/min Nozzle Diameter, mm Nozzle Area, cm2 Number of nozzles Nozzle Velocity, m/s  437 0.508 0.0020027 4 8.98 430 0.3556 0.000993 4 18.04 215 0.3556 0.000993 4 9.02 432 0.7112 0.003973 4 4.53 432 0.7112 0.003973 2 9.06 680 0.635 0.003167 4 8.95 634 0.4318 0.001464 4 18.04 670 0.889 0.006207 4 4.5     122  APPENDIX C: UPFLOW VELOCITY AND HRT CALCULATION Flow velocity = 𝐹𝑙𝑜𝑤𝐶𝑟𝑜𝑠𝑠 𝑆𝑒𝑐𝑡𝑖𝑜𝑛𝑎𝑙 𝐴𝑟𝑒𝑎 Hydraulic Retention Time = 𝑉𝑜𝑙𝑢𝑚𝑒𝐹𝑙𝑜𝑤                                         Table C.1: Upflow velocity and HRT for reaction zone. Volume (L) Cross-Sectional Area (cm2) Flow (mL/min) Velocity (cm/s) HRT (min) 1.36 10.93 483 44.19 2.82 437 39.98 3.11 430 39.34 3.16 215 19.67 6.33 432 39.52 3.15 680 62.21 2 634 58.01 2.15 670 61.3 2.03    123  Table C.2: Up-flow velocity and HRT for clarifier. Volume (L) Cross-Sectional Area (cm2) Flow (mL/min) Velocity (cm/s) HRT (min) 31.3 729.66 483 0.66 64.80331 437 0.6 71.62471 430 0.59 72.7907 215 0.29 145.5814 432 0.59 72.4537 680 0.93 46.02941 634 0.87 49.36909 670 0.92 46.71642            124  APPENDIX D: REYNOLDS NUMBER CALCULATION Reynolds Number = 𝜌 × 𝑉 ×𝐷𝜇 Where, V = velocity of fluid, m/s ρ = density of fluids, kg/m3 D = diameter, m μ = viscosity of fluid, N.s/m2 The value of ρ and μ at 25°C are respectively 997 kg/m3 and 8.9 × 10-4 N.s/m2 (Metcalf and Eddy, 2003). Table D.1: Reynolds number for different sections. Section Diameter (m) Velocity (cm/min) Velocity (m/s) Reynolds Number Reaction Zone 0.0381 19.67-62.71 0.003278-0.010368 140-443 Clarifier 0.3048 0.29-0.93 4.83×10-5- 0.000155 17-53      125  APPENDIX E: BEAKER TEST  Figure E.1:Beaker test result from run 7. Caustic flow calculation from Beaker test (for run 7): Volume of caustic = 7.8 ml Volume of feed = 200 ml Feed flow = 437 ml/min Caustic flow = 𝑉𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑐𝑎𝑢𝑠𝑡𝑖𝑐 ×𝐹𝑒𝑒𝑑 𝑓𝑙𝑜𝑤𝑉𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑓𝑒𝑒𝑑= 17𝑚𝑙𝑚𝑖𝑛.      7.67.77.87.988.18.28.38.48.58.60 20 40 60 80 100 120pHTime (min)  126  APPENDIX F: pH METERS Table F.1: Locations and models of pH meters. Location Model Mixing Tank Orion pH – 420 A 1’-4” from bottom of the reactor Oakton pH 11 2’-6” from bottom of the reactor Oakton pH 11 Effluent Oakton pH 11            127  APPENDIX G: pH CALIBRATION CURVES           Figure G.1: pH calibration curves.   y = 0.9615x + 0.23080510150 5 10 15Actual pHpH shown by pH meter1'-4” from bottom of the reactory = 0.9464x + 0.31860510150 5 10 15Actual pHpH shown by pH meterEffluenty = 1.0033x - 0.224105100 5 10Actual pHpH shown by pH meterMixing Tanky = 0.9615x + 0.16350510150 5 10 15Actual pHpH shown by pH meter1.625" from the bottom of the reactory = 0.9464x + 0.43220510150 5 10 15Actual pHpH shown by pH meterpH Controller  128  APPENDIX H: PHOSPHORUS CONCENTRATIONS IN UNFILTERED AND FILTERED EFFLUENT SAMPLES DURING FIRST PHASE  Figure H.1: P-concentration in effluent (filtered and unfiltered) samples under different supersaturation level.           010203040506070804 4.5 5 5.5 6 6.5 7 7.5 8Phosphorus concentration (mg/L)Initial SstruviteEffluent (Unfiltered) Effluent (Filtered)  129  APPENDIX I: AGGREGATION MECHANISM A brief identification of various terms used in Equation 2.5 has been depicted below (Linnikov, 2008, 2004; Polak, 1966; Vaganov et al., 1974) -                                                     𝐷1 =  4ℎ𝜎2𝑁𝐴𝑀𝜌𝑇𝑅2𝑇2                                                                                   (𝐽. 1)                                                     𝐷2 =  2𝜎𝑀ℎ𝜌𝑇𝑅𝑇                                                                                        (𝐽. 2)                                                     𝐷3 =  𝐼𝑜𝜏𝑘                                                                                              (𝐽. 3)                                                     𝑘1 =  𝑘𝑜𝑃𝑉−1𝑊1                                                                                   (𝐽. 4) Where h = Height of nucleus bridge (m) as shown in Figure J.1. σ = specific surface energy (J.m-2). NA = Avogrado constant (mol-1). M = Molar weight of a salt (kg.mol-1). 𝜌𝑇 = Density of a salt (kg.m-3). T = Temperature (K). R= Universal gas constant (J.m-2). τk = Time during which crystals are in contact (s). ko = Hydrodynamic constant based on fluid dynamics conditions of a crystallization process which determines the probability of collisions between crystals in solution. P = Probability of a suitable intergrowth orientation of crystals at the moment of collision known as orientation factor. V = volume of solution (m3).   130  W1 = probability that the nucleus-bridges formed between crystals are not destroyed at subsequent collisions.    Figure I.1: Aggregation mechanism of two approaching crystals based on Polak (1966). 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0340473/manifest

Comment

Related Items