UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development of control strategies for the operation of a struvite crystallization process Fattah, Kazi Parvez 2010

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

Item Metadata

Download

Media
24-ubc_2010_fall_fattah_kazi.pdf [ 4.83MB ]
Metadata
JSON: 24-1.0062917.json
JSON-LD: 24-1.0062917-ld.json
RDF/XML (Pretty): 24-1.0062917-rdf.xml
RDF/JSON: 24-1.0062917-rdf.json
Turtle: 24-1.0062917-turtle.txt
N-Triples: 24-1.0062917-rdf-ntriples.txt
Original Record: 24-1.0062917-source.json
Full Text
24-1.0062917-fulltext.txt
Citation
24-1.0062917.ris

Full Text

1  DEVELOPMENT OF CONTROL STRATEGIES FOR THE OPERATION OF A STRUVITE CRYSTALLIZATION PROCESS by Kazi Parvez Fattah  M.A.Sc. (Civil Engineering), The University of British Columbia, Vancouver, Canada, 2004 B.Sc. (Civil Engineering), Bangladesh University of Engineering and Technology, Dhaka Bangladesh, 2001   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (CIVIL ENGINEERING)    THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  July 2010    © Kazi Parvez Fattah, 2010 Abstract ii  ABSTRACT In this study, a dynamic control model for struvite crystallization process was developed that incorporated both chemistry and control software, which could be used to increase the efficiency and ease process operation. This process model was the basis of an automatic controller that had the capability to manipulate flows and chemical additions, and thereby control the system at a desired set point. The control model was then used as a prediction tool to determine conditions that influence the supersaturation ratio of the process.  A pilot  scale  crystallizer  was  operated  at  a  local  treatment  plant  to  test  the  model.  The struvite produced from the operation of the process was then examined to determine the influence of various operating parameters on its quality. Supersaturation ratio (SSR) and upflow velocity in the crystallizer were found to influence the size and shape of the pellets developed. Mid-sized pellets (2.0-2.5 mm), had the highest crushing strengths; SSR did not appear to influence the crushing strength of pellets formed. High concentration of magnesium in the crystallizer was related to the formation of pellets having greater crushing strengths.  To determine a single solubility constant for struvite, a study was conducted under varying experimental conditions. Results showed that, for a particular temperature, and in the working pH range of 7.0-9.0, the solubility constant was independent of the pH and water matrix. These experimental values, along with values found in literature, were used to derive a universal constant and a linear equation relating solubility product (pKsp) with temperature.  In this study, the effectiveness of two carbon dioxide strippers, in reducing caustic usage, were evaluated. Results showed that carbon dioxide stripping was efficient in reducing caustic costs, by as much as 46%. The potential saving in caustic cost due to CO2 stripping was calculated to be as high as 38 cents per thousand liters treated.  Abstract iii  The determination of the concentration of Mg in a struvite crystallization process is important because of its influence on SSR, the associated operational cost and its struvite forming potential when unused chemical is passed back to the treatment plant. In this study, methods were tested, with acceptable degree of accuracy, which can provide information, on- site, on the concentration and rate of application of the element.      Table of Contents iv  TABLE OF CONTENTS ABSTRACT……………………………………………………………………………...ii TABLE OF CONTENTS .............................................................................................iv LIST OF TABLES.........................................................................................................x LIST OF FIGURES......................................................................................................xi LIST OF ABBREVIATIONS......................................................................................xv ACKNOWLEDGEMENTS.......................................................................................xvii DEDICATION. .........................................................................................................xviii CHAPTER ONE: INTRODUCTION...........................................................................1 1.1 Background........................................................................................................1 1.2 Organization of the Dissertation .........................................................................4 CHAPTER TWO: BACKGROUND AND LITERATURE REVIEW........................6 2.1 Introduction .......................................................................................................6 2.2 Motivation for Phosphorus Removal ..................................................................6 2.3 Methods of Phosphorus Recovery ....................................................................10 2.4 Advantages of Phosphorus Recovery as Struvite ..............................................11 2.5 Chemistry of Struvite .......................................................................................13 2.6 Solubility of Struvite ........................................................................................14 2.7 Factors Affecting Struvite Solubility ................................................................16 2.7.1 pH ………………………………………………………………………….16 2.7.2 Temperature .............................................................................................16 Table of Contents v  2.8 Chemical Requirements for Struvite Crystallization Process.............................17 2.8.1 Caustic .....................................................................................................17 2.8.2 Magnesium addition .................................................................................18 2.9 Reduction of Gases by Stripping ......................................................................19 2.10 Determination of Magnesium Usage for Struvite Precipitation .........................21 2.11 Morphology of Struvite Pellets.........................................................................24 2.12 Quality of Struvite Pellets and its Applications.................................................25 2.13 Operation of Crystallizer for Struvite Formation ..............................................27 2.14 Modeling and Control of Struvite Precipitation ................................................28 CHAPTER THREE: OBJECTIVES OF THE STUDY.............................................31 CHAPTER FOUR: METHODS AND METHODOLOGY .......................................34 4.1 Struvite Recovery Process................................................................................34 4.2 Chemicals, Storage Tanks and Pumps ..............................................................36 4.2.1 Centrate....................................................................................................36 4.2.2 Magnesium feed .......................................................................................36 4.2.3 pH control ................................................................................................36 4.3 Sample Collection, Storage and Analysis .........................................................37 4.4 Lulu Island Wastewater Treatment Plant ..........................................................39 4.5 Instrumentation and Process Monitoring ..........................................................42 4.6 Struvite Control Process...................................................................................43 4.7 Coding Controller Program in Matlab ..............................................................46 4.7.1 Coding to relate hydrogen ion concentration to SSR (in solvepH.m).........46 4.8 Characteristics of Process Feed during Pilot Scale Studies with Instrumentation ...............................................................................................49 4.9 Experimental Setup ..........................................................................................52 Table of Contents vi  4.9.1 Purpose of the study – solubility tests .......................................................53 4.9.2 Experimental setup and methodology for solubility tests...........................53 4.9.3 Purpose of the study - carbon dioxide stripping ........................................55 4.9.4 Experimental setup for carbon dioxide stripping .......................................55 4.9.5 Methodology to determine efficiency of the carbon dioxide strippers .......59 4.9.6 Purpose of the study - Determination of magnesium usage for struvite precipitation ........................................................................................61 4.9.7 Experimental setup and methodology for the determination of magnesium usage for struvite precipitation.............................................................62 4.10 Product Quality Determination.........................................................................63 4.10.1 Development of strength tester .................................................................64 4.11 Terminology ....................................................................................................66 4.11.1 Recycle ratio ............................................................................................66 4.11.2 Upflow velocity........................................................................................67 4.11.3 Removal efficiency...................................................................................67 4.11.4 Confidence limit/Error..............................................................................67 4.11.5 Root mean square …………………...................................................…….68 CHAPTER FIVE: RESULTS AND DISCUSSIONS .................................................69 5.1 Testing of Control Program..............................................................................69 5.1.1 Laboratory simulations .............................................................................69 5.1.2 Experimental runs ....................................................................................71 5.2 Application of Model to Predict Different Scenarios for Control of Struvite Formation Potential.........................................................................................73 5.3 Application of Model for Decision Making ......................................................76 5.4 Carbon Dioxide Stripping Model......................................................................80 5.5 Struvite Crystallization Operation Control Windows........................................80 5.6 Graphical User Interfaces (GUI) used for Operation of Crystallization Process............................................................................................................81 Table of Contents vii  5.7 Performance Comparison of the Two Systems .................................................87 5.8 Solubility Tests ................................................................................................91 5.8.1 pKsp prediction using speciation model....................................................92 5.8.2 Summary from solubility tests ..................................................................96 5.9 Carbon Dioxide Stripping.................................................................................96 5.9.1 Run No 1..................................................................................................96 5.9.2 Run No. 2.................................................................................................98 5.9.3 Run No. 3.................................................................................................99 5.9.4 Run No. 4...............................................................................................101 5.10 Potential for Stripper Fouling .........................................................................103 5.10.1 Clogging of the compact media stripper..................................................103 5.10.2 Cost analysis ..........................................................................................105 5.10.3 Conclusions............................................................................................106 5.11 Prediction of Magnesium Requirements from Conductivity-pH Measurements ...............................................................................................107 5.11.1 Theoretical versus practical change in conductivity ................................107 5.11.2 Experimental runs to determine bending point for magnesium addition ..110 5.12 Use of Chemicals to Reduce Phosphate Interference ......................................117 5.12.1 Use of polyaluminum chloride (PAC).....................................................117 5.12.2 Use of alum ............................................................................................119 5.13 Hardness Test Results ....................................................................................122 5.14 Composition of Struvite Formed During Pilot Scale Operation.......................124 5.14.1 Magnesium to phosphate (Mg:P) molar ratio ..........................................124 5.14.2 Ammonium-nitrogen (N:P)  to phosphate molar ratio .............................125 5.14.3 Purity .....................................................................................................126 5.14.4 Shell formation – composition, SEM pictures.........................................128 5.14.5 Compactness of struvite pellets...............................................................130 5.14.6 Influence of upflow velocity on compactness .........................................131 5.14.7 Struvite composition for presence of heavy metals .................................131 Table of Contents viii  5.15 Influence of Parameters on Physical Structure of Struvite Pellets ...................132 5.15.1 Possible influence of crystallizer supersaturation ratio ............................133 5.15.2 Upflow velocity......................................................................................136 5.15.3 Flow patterns in the crystallizer ..............................................................137 5.16 Influence of Parameters on Crushing Strength of Struvite Pellets ...................140 5.16.1 Influence of size on strength...................................................................140 5.16.2 Influence of struvite composition on strength .........................................141 5.16.3 Influence of SSR on strength ..................................................................143 5.16.4 Influence of reactor magnesium concentration on strength......................144 5.16.5 Shell formation and two peak strength....................................................145 5.16.6 Conclusions on struvite quality...............................................................145 CHAPTER SIX:  CONCLUSIONS ..........................................................................148 6.1 Development of Control Programs .................................................................148 6.2 Solubility Tests ..............................................................................................149 6.3 Carbon Dioxide Stripping Tests .....................................................................149 6.4 Magnesium Prediction Techniques.................................................................150 6.4.1 Conductivity-pH measurements..............................................................150 6.4.2 Determination of magnesium concentration by hardness test method......151 6.5 Influence of Variables on Structure of Struvite Pellets....................................151 6.6 Determination of Crushing Strength...............................................................151 CHAPTER SEVEN: RECOMMENDATIONS FOR FURTHER STUDY .............153 REFERENCES …………………………………………………………………….…  155 APPENDICES…………………………………………………………………………167 APPENDIX A: SUMMARY OF PROGRAM CODE TO DETERMINE REQUIRED pH OF A STRUVITE CRYSTALLIZER SYSTEM (solvepH.m) ......................................................................................168 Table of Contents ix  APPENDIX B: VALUES OF PARAMETERS TO DETERMINE THEIR INFLUENCES ON THE SUPERSATURATION RATIO.............172 APPENDIX C: EQUATIONS USED IN DERIVING STRIPPER MODEL ..........173 APPENDIX D: PERFORMANCE OF CONTROLLER – EFFLUENT DATA.....174 APPENDIX E: SOLUBILITY TESTS .....................................................................176 APPENDIX F: CARBON DIOXIDE TESTS ...........................................................184 APPENDIX G: INDIRECT METHOD (USING pH AND CONDUCTIVITY) TO DETERMINE EXTERNAL Mg ADDITION REQUIREMENTS FOR STRUVITE CRYSTALLIZATION ......188 APPENDIX H: APPLICATION OF CHEMICALS FOR THE REDUCTION OF PHOSPHATE CONCENTRATION IN WATER MATRIX...193 APPENDIX I: HARDNESS TEST RESULTS .........................................................195 APPENDIX J: OPERATIONAL DATA – REACTOR 1 ........................................197 APPENDIX K: COMPOSITION OF STRUVITE PELLETS – ICP-MS TEST RESULTS ........................................................................................209 APPENDIX L: CRUSHING STRENGTH DATA ...................................................211  List of Tables x                                      LIST OF TABLES Table 2.1. Competing reactions in struvite formation in a distilled water system..................13 Table 2.2. Sources of magnesium studied with their corresponding pH values .....................19 Table 4.1. Dimensions of the reactor ...................................................................................36 Table 4.2. Sets of equations used to account for the impact of temperature on the equilibrium constants. The equations are based on Equation 2.6 and values in Table 2.1............................................................................................................45 Table 4.3. List of files written and developed in Matlab with their uses ...............................48 Table 4.4. Characteristics of the process feed used during pilot-scale operation (2009)........49 Table 4.5. Operating conditions in each struvite crystallizer ................................................52 Table 4.6. Test conditions in the crystallizers and strippers for the comparison of strippers .............................................................................................................61 Table 4.7. Characteristics of centrate used during carbon dioxide stripping tests..................61 Table 4.8. Summary of operating conditions in the crystallizers ..........................................61 Table 5.1. Summary of results from the operation of the two crystallizers ...........................89 Table 5.2. pKsp values for each water matrix as calculated using the Speciation Model.......93 Table 5.3. pKsp values used to derive Equation 5.2 .............................................................95 Table 5.4. Summary of findings from four stripping tests ..................................................103 Table 5.5. Cost analysis for caustic usage in a pilot scale struvite crystallizer fitted with a cascade stripper .............................................................................................106 Table 5.6. Composition of struvite formed at LIWWTP.....................................................127 Table 5.7. Composition of shell-structured struvite pellet ..................................................128 Table 5.8. Concentration of heavy metals found in struvite pellets formed at LIWWTP and in nature ....................................................................................................132 Table 5.9. Influence of SSR on crushing strength ..............................................................144 Table 5.10. Influence of reactor magnesium concentration on crushing strength of struvite.............................................................................................................145 1 List of Figures xi  LIST OF FIGURES Figure 2.1. (a) Struvite in pipe leading to centrate pump at Lulu Island WWTP, (b) Struvite in adjacent piping to the digester pump at Annacis Island WWTP (courtesy of Brian Hystad, Metro Vancouver, 2007)...........................................8 Figure 4.1. Pilot-scale struvite crystallizer reactor process design. .......................................35 Figure  4.2.  Setup  of  crystallizers,  clarifiers  and  stripping  columns  at  the  LIWWTP  (a) crystallizer (R#1) with compact media stripper attached and without instrumentation and control – the manually controlled system (b) crystallizer (R#2) with cascade stripper attached and with instrumentation and control – the automated system (c) transmitters, pH meter and pH pump for R # 2 system (d) location of conductivity and pH probe in harvest zone of R#2. ........38 Figure 4.3. Trends in important parameters in centrate at LIWWTP over the last few years. (a) pH, (b) temperature and (c) conductivity. ..........................................40 Figure 4.4. Schematic of phosphorus recovery process with instrumentation locations at the Lulu Island Wastewater Treatment Plant. ...................................................42 Figure 4.5. Schematic of the control strategy for struvite crystallization process. .................44 Figure 4.6. Centrate characteristics at the Lulu Island Wastewater Treatment Plant during pilot-scale validation of process control in 2009 (a) phosphate, (b) ammonium. ......................................................................................................50 Figure 4.7. Setup of solubility determination. ......................................................................54 Figure 4.8. (a) Compact media stripper schematic, (b) stripper at setup at LIWWTP, (c) hollow plastic balls used and (d) final arrangement of balls in the stripper........57 Figure 4.9. (a) Dimensions of the cascade stripper, (b) stripper at setup at LIWWTP and (c) dimensions of individual baffles..................................................................58 Figure 4.10. Centrate characteristics during experiments testing the efficiency of the carbon dioxide stripper (a) phosphate, (b) ammonium and (c) magnesium........60 Figure 4.11. Device used to determine crushing strength of struvite (a) sketch (not to scale) and (b) actual device. .............................................................................65 List of Figures xii  Figure 4.12. Program window used to graphically illustrate force as a function of time and to determine peak load...............................................................................66 Figure 5.1. Effectiveness of the SSR controller in calculating and maintaining required pH due to (a) changing phosphate concentration, (b) changing temperature......70 Figure 5.2. (a) Variation of conductivity and temperature in the crystallizer, (b) variation between set point pH (for particular SSR) and measured pH in a pilot-scale struvite crystallizer controlled by the SSR controller. .......................................72 Figure 5.3. Influence of temperature and Mg:P molar ratios on calculated system SSR........74 Figure 5.4. Influence of temperature on supersaturation ratio at different phosphate concentrations. .................................................................................................75 Figure 5.5. Influence of temperature on supersaturation ratio at different magnesium concentrations. .................................................................................................75 Figure 5.6. (a) Influence of magnesium concentration and temperature on SSR and (b) influence of Mg:P molar ratio and temperature; all other parameters constant. .77 Figure 5.7. Graphical representation showing the level of influence each parameter has (while  keeping  others  constant)  on  supersaturation  ratio  (shown  in  the  last graph). .............................................................................................................79 Figure 5.8. First level of control window for struvite crystallization process. .......................83 Figure 5.9. Second level control block containing the SSR and pH control codes. ...............84 Figure 5.10. Graphical user interface for reactor operation...................................................85 Figure 5.11. Graphical user interface for calculating the supersaturation ratio......................86 Figure 5.12. Graphical user interface for determining the efficiency of a cascade carbon dioxide stripper. ...............................................................................................87 Figure 5.13. Effluent characteristics from the two systems (a) phosphate, (b) ammonium and (c) magnesium concentrations. Sample numbers represents a day’s sample..............................................................................................................90 Figure 5.14.  pPS eq curve of a batch experiment trial of supernatant at different temperatures (Adapted from Forrest, 2004). .....................................................92 Figure 5.15. Typical pKsp predictions using Speciation Model. ...........................................93 Figure 5.16. Variation of pKsp values with temperature. Error bars at 95% confidence interval. ............................................................................................................94 List of Figures xiii  Figure 5.17. Molar phosphorus removal and caustic use in R#1 and R#2 – Run No. 1. ........98 Figure 5.18. Molar phosphorus removal and caustic use in R#1 and R#2 – Run No. 2. ........99 Figure 5.19. Molar phosphorus removal and caustic use in R#1 and R#2 – Run No. 3. ......101 Figure 5.20. Molar phosphorus removal and caustic use in R#1 and R#2 – Run No. 4. ......102 Figure 5.21. Struvite accumulation in stripper during operation. ........................................105 Figure 5.22. Influence of magnesium chloride addition on the conductivity – (a) predicted vs. measured in distilled water (with no P and N), and (b) predicted vs. measured in centrate sample...........................................................................109 Figure 5.23. (a) Influence of magnesium chloride addition on pH and conductivity of a synthetic wastewater and (b) change in pH and conductivity as a function of magnesium chloride addition -  Run 1. ...........................................................111 Figure 5.24. Influence of magnesium chloride addition on pH and conductivity in centrate sample for three different runs........................................................................114 Figure 5.25. Influence of magnesium chloride addition on conductivity and pH of centrate – determination of transition point.....................................................115 Figure 5.26. Change of pH and conductivity with magnesium chloride addition in centrate matrix. The primary data used is the same as for Figure 5.25.............116 Figure 5.27. Relationship between moles of phosphate removed and Mg:P molar ratio at the transition point. ........................................................................................117 Figure 5.28. Removal of phosphate from centrate with addition of PAC. ...........................119 Figure 5.29. Influence of alum on phosphate concentration. ..............................................120 Figure 5.30. Influence of alum addition on magnesium concentration................................121 Figure 5.31. Comparison of calcium concentrations via hardness method and AA. The line depicts the equivalence line. ....................................................................123 Figure 5.32. Comparison of magnesium concentrations via hardness method and AA. The line depicts the equivalence line. ....................................................................123 Figure 5.33. Measured magnesium to phosphate molar ratio of struvite samples................125 Figure 5.34. Measured ammonium-nitrogen to phosphate molar ratio of struvite samples. .126 Figure 5.35. Shell formation of pellets (a) actual struvite and (b) SEM picture taken at x 50 magnification. ...........................................................................................129 List of Figures xiv  Figure 5.36. SEM pictures of struvite pellets: (a) illustrating the fusing of struvite crystals (500 magnification) and (b) compactness of the crystals (at 250 magnification). ...............................................................................................130 Figure 5.37. Typical SSR profiling in UBC’s crystallizer with crystallizer height (adapted from Forrest, 2004). .......................................................................................134 Figure 5.38. Formation of struvite pellets: individual crystals at (a) 2000 X magnification and (b) 500 X magnification, (c) agglomeration of crystals at 350 X magnification, (d) fused and smooth outer surface of pellet at 100 X magnification. ................................................................................................135 Figure 5.39. Struvite pellets harvested at LIWWTP in crystallizer with instrumentation. ...136 Figure 5.40. Breakage of struvite pellets possibly due to high upflow velocity...................137 Figure 5.41. (a) Elongated pellets possibly formed due to flow restrictions in the crystallizer and (b) pellets from Ostara’s reactor.............................................139 Figure 5.42. Influence of pellet size on crushing strength. Error bars: 95% confidence interval. ..........................................................................................................141 Figure 5.43. Relationship between Mg:P molar ratio and the crushing strength of struvite pellet. Error bars: 95% confidence interval. ....................................................142 Figure 5.44. Relationship between N:P molar ratio and the crushing strength of struvite pellet. Error bars: 95% confidence interval. ....................................................143 Figure 5.45. Crushing strength graphs showing formation of single peaks {(a) and (b)} for normal pellets and dual peaks for pellets with shell-structure {(c) and (d)}. ...............................................................................................................147  List of Abbreviations xv  LIST OF ABBREVIATIONS AA Atomic Absorption AIWWTP Annacis Island Wastewater Treatment Plant ANN Artificial Neural Network ASR Air Supply Rate ATM Atmosphere, Unit of pressure AWWTP Advanced Wastewater Treatment Plant BC British Columbia BN Baffle Number BOD Biochemical Oxygen Demand CAD Canadian Dollar Cond. Conductivity (E)BNR (Enhanced) Biological Nutrient Removal EC Electric Conductivity (mS/cm) EDTA Ethylene diamine tetraacetic acid ERR Effluent Recycle Ratio GUI Graphical User Interface HP Horse power HRT Hydraulic Retention Time (min) IAP Ionic Activity Product IBC Influent Buffering Capacity ICP Inductively Coupled Plasma IFR Influent Flow Rate ISE Ion Selective Electrode IT Influent Temperature K Kelvin, Unit of Temperature Ksp Solubility Product Kspeq Equilibrium Ksp LIWWTP Lulu Island Wastewater Treatment Plant List of Abbreviations xvi  MAP Magnesium Ammonium Phosphate Hexahydrate; Struvite MGD Millions Gallons per Day MLD Million Liters per Day Mol Moles MS Mass Spectrometer mS/cm Milli Siemens per centimeter P Phosphorus PAC Polyaluminum chloride PM Peroxide Microwave Ps Conditional Solubility Product Pseq Equilibrium Ps PVC Polyvinyl chloride Qf Feed flow Qr Recycle flow Qt Total flow R Gas constant R #1 Reactor (system) 1 R #2 Reactor (system) 2 RCF Relative Centrifugal Force RMS Root Mean Square RR Recycle Ratio SD Standard Deviation SEM Scanning Electron Microscope SSE Sum of Squared Error SSR Supersaturation Ratio SSRc Conditional Supersaturation Ratio T Temperature in degrees Kelvin TSS Total Suspended Solids UBC University of British Columbia WWTP Wastewater Treatment Plant XRD X-ray diffraction  Acknowledgements xvii  ACKNOWLEDGEMENTS  I would like to acknowledge the following people for the assistance, support and encouragement that I have received throughout my research. Without their cooperation, this study would not be possible. ƒ Firstly, I would like to convey my gratitude to my supervisor, Prof. D. S. Mavinic, for his encouragement, patience, understanding and firm support throughout the course of the research. ƒ I appreciate the critical suggestions and guidance provided to me by my committee members,  Dr  Eric  Hall,  Dr.  Victor  Lo,  Dr.  John  Grace  and  Dr.  Bhushan  Gopaluni., throughout the study period, and for their review and comments on the thesis. ƒ Appreciation also goes to our research associate, Frederic Koch, and the other graduate students in this research group, Iqbal Bhuiyan, Saifur Rahaman and Nandini Sabrina,  for  sharing  their  ideas  and  knowledge  with  me.  I  am  also  grateful  to  my undergraduate co-op students, Jeffrey Ho and Larissa Lam, who helped me with some of the laboratory work. Special thanks go to Doug Hudnuik, John Wang, Paula Parkinson and Susan Harper of the department. ƒ The people  at  Lulu  Island  Wastewater  Treatment  Plant,  for  allowing  me to  conduct my research at the plant and for encouraging me all along, and Cristina Jacobs, Brian Hystad and Vince Chu of Metro Vancouver. ƒ Metro Vancouver, BC Hydro and NSERC, for their generous funding of this research project. ƒ Alexander Forrest and Nandini Sabrina, with whom I have co-authored journal articles and conference proceedings. ƒ My parents, for their encouragement and support throughout my degree. And my lovely beautiful daughter of 19 months, who sacrificed her playtime with daddy. ƒ My wife, Farah Laj Chowdhury, for her patience, encouragement, time and support, and who sacrificed the most in seeing me through graduate school. ƒ Above all, this research was possible only by the grace of GOD Dedication xviii  2 DEDICATION  To my wife  Farah Laj Chowdhury For all her love, support and patience Chapter One: Introduction 1  1 CHAPTER ONE: INTRODUCTION 1.1 Background The recovery of phosphorus, as struvite, from wastewater provides an environmentally sound and renewable nutrient source to the agricultural, landscaping and recreational industry, as well as solving wastewater treatment plant problems. Although there are different phosphorus recovery processes in use, struvite is usually grown in fluidized bed reactors by a crystallization process.  The pilot-scale crystallizer used at the University of British Columbia (UBC) is different from others reported in the literature in that the reactor has different sections of varying diameters, as opposed to the more common conical shape. Previous studies at the UBC have successfully demonstrated the applicability of struvite precipitation as a means of reducing, and recovering, phosphates from wastewater. Although different control parameters can be used for struvite (magnesium ammonium phosphate, MgNH4PO4.6H2O) precipitation, the primary control parameter used in most of the studies was the supersaturation ratio (SSR), which is the degree of struvite saturation in the crystallizer. This value is dependent on the struvite solubility product, or Ksp,  more  commonly  referred  to  as  the  pKsp (-log Ksp) value. Numerous solubility constants are found in the literature with respect to struvite formation. Due  to  these  wide  range  of  reported  values,  experiments  are  needed  to  determine  if  there exists a correlation between values obtained with different water matrices (distilled water, tap water, digester supernatant and centrate), at different temperatures and pH. By using several combinations of experimental conditions involving the aforementioned variables, the struvite solubility product was determined in the present study.  The acceptance of struvite precipitation as a means of wastewater treatment and a commercial  source  of  phosphorus  will,  to  a  large  extent,  depend  on  the  economics  of  the process. One of the primary expenses arises from the need to increase the pH of the system by the addition of a caustic substance, usually sodium hydroxide. Carbon dioxide stripping Chapter One: Introduction 2  can help increase the pH of the system, thereby providing a viable and cheaper option than the traditional method of only chemical addition.  Although carbon dioxide stripping is not a new  phenomenon,  its  use,  efficiency  and  cost  benefits,  when  applied  to  a  struvite crystallization process, have not been studied in great detail. The present study examined the efficiency of two types of carbon dioxide strippers in reducing chemical cost, as well as, determined operating conditions and issues arising from their use.  Struvite precipitation in a fluidized bed crystallizer is a continuous process. Therefore, variations in the constituent ions (magnesium, ammonium and phosphate) can result in the formation, or dissociation, of the compound. Consequently, it is important to monitor the concentration of these ions. Since most wastewater used for phosphorus recovery as struvite lacks magnesium (Mg), an external source is usually applied to increase the concentration in the system. The determination of the concentration of Mg during this addition is important, because not only does it determine the value of the supersaturation ratio, unused Mg in the effluent can increase struvite formation potential elsewhere in the treatment plant. Although ammonium and phosphate ions can be measured in real time using online analyzers, there are no suitable analyzers for measuring real-time magnesium concentrations in wastewater and in the external source. The usual method of measuring magnesium [by atomic absorption (AA) spectroscopy] is a time consuming and expensive process. Therefore, a suitable on-site method that is quick and cheap would go a long way in optimizing the magnesium addition. In the present study, two methods to determine magnesium addition requirements on-site were proposed and tested for their efficiency.  Despite numerous studies being conducted worldwide on phosphorus precipitation as struvite,  most  of  the  research  is  based  on  the  chemistry  of  the  process,  and  few have  dealt with actual operation and optimization of the process at the pilot- or full-scale installations. Studies carried out have demonstrated the need for tighter control of process variables, such as temperature, conductivity, pH and constituent ions, for improved product quality and ease of  process  control  of  a  struvite  crystallization  process.  In  order  to  maintain  this  “narrow working window” in process conditions, it is important that the effects of the variations in these process variables be as limited as possible. The variations in variables are natural Chapter One: Introduction 3  during wastewater treatment processes, and as struvite crystallization uses the effluent from these processes, it is natural to expect these variations during struvite processes as well. The changes in the variables are responsible for changing the primary control parameter – the supersaturation ratio in a struvite crystallizer. The problem with the present method of manual control of a struvite crystallization process is that any sudden and unexpected change in any of the variables during its operation (when no operator is present on-site) is not noticed until the next day. Consequently, struvite growth and quality can be compromised.  As mentioned above, the production of struvite, from wastewater, in a crystallizer depends on a number of variables, which bring about complex and non-linear changes in the process chemistry. This complex situation can be handled diligently by gaining an insight into the process using models. Models can help in determining the effects of individual, or the combined effects, of process variables. Models also allow operators to determine the best possible method of operating a phosphorus recovery process. In order to keep a process in a desired state, it is necessary to counteract influences that disturb the process. Although various chemical-based models are present that predict formation of struvite, there are no known automatic controller programs that are capable of maintaining a struvite crystallization system at the desired set point, such as a particular supersaturation ratio. It is widely believed that the efficiency and economy of process operations can be increased by providing continuous, suitable, and stable conditions by automatic control. During the last decade, interest in the recovery of phosphate from wastewater has grown tremendously, and methods are being developed in laboratory, pilot- and full-scale operations at treatment plants are beginning to show up worldwide. This rapid growth shows the significance of the process in recovering the phosphorus from wastewater. Consequently, development of on-line instrumentation, modeling of the process, and the integration of the two, promises to enhance the efficiency of a struvite crystallization system and improve product quality. The present study attempted to develop a control system that can be used to efficiently operate a phosphorus recovery/struvite crystallizer at a pilot-scale installation. Through computer modeling of the chemical process, the effects of different variables were also determined. Chapter One: Introduction 4  1.2 Organization of the Dissertation This dissertation has been written in the classical style, and contains all pertinent experimental methodology, data and results. The following provides an outline of the organization of the various chapters in this dissertation.  Chapter One provides an introduction to the topic of phosphorus recovery from wastewater as struvite, and the reasons for pursuing the different topics dealt with in this research. A summary of the different chapters presented in this dissertation is also provided.  Chapter Two provides a background and literature review of phosphorus removal- recovery and struvite chemistry, as well as, presenting an overview of the various factors affecting struvite formation. This chapter provides information on struvite modeling, struvite solubility product, carbon dioxide stripping and magnesium determination. It points out some of the areas where knowledge gaps exist in struvite crystallization process operation. It provides the basis and motivation for carrying out the present study.  Based on information gathered, and with the intention of addressing the knowledge gap identified in Chapter Two, Chapter Three lists the hypothesis and objectives of the present study, and the rationale for the various supplementary studies carried out.  Chapter  Four  provides  a  description  of  the  different  experiments  conducted,  the  water matrix and instruments used in each experiment, and the samples collected and analyzed. Because of the variation in the nature of the experiments and studies carried out, each set of experiments have been separated under a new heading for the sake of clarity. The purpose of each experiment has also been provided in brief.  The first part of the chapter introduces the struvite process system as used in the present study, followed by the development of the methodology and formulation of the codes and programs used to develop a process controller for struvite precipitation. Methods and materials used for the supplementary experiments, namely, struvite solubility product, carbon Chapter One: Introduction 5  dioxide stripping, indirect measurements of concentration and application rates of external magnesium during struvite crystallization process are then introduced and elaborated.  Chapter Five provides a summary of the experimental results that support the hypothesis of the research and the need and effectiveness of the supplementary studies. The chapter also discusses the impact(s) of the research and results on struvite crystallization process.  Chapter Six summarizes the main findings from the present study. Similar to the organization in Chapter Four, conclusions for each set of experiments have been placed separately. This chapter provides an identification of the significance and contribution of the thesis research to the field of phosphorus removal-recovery as struvite.  Chapter Seven provides a list of recommendations and directions for future study on phosphorus recovery from wastewater as struvite.  Chapter Two: Background and Literature Review 6  2 CHAPTER TWO: BACKGROUND AND LITERATURE REVIEW 2.1 Introduction Although phosphorus (P) is one of the key elements that sustain all life forms, it is a non- renewable resource and has the least reserves/resources globally. Phosphorus is primarily used by the agricultural fertilizer industry, followed by household products. Phosphorus originating from detergents, metabolic processes, diffuse runoff from agricultural land and inputs  from  the  air  find  their  way  into  the  domestic  wastewater  system.  Even  at  low concentrations of approximately 10 ȝg/L of phosphorus, secondary reactions, also known as eutrophication, may occur in the receiving water. Discharge of partially treated wastewater into water bodies has been linked to increased numbers of algal blooms around the world (Barnard, 2009). Algal blooms reduce the amenity value and compromise ecological health of water bodies such as lakes, slow moving rivers and drinking water reservoirs. 2.2 Motivation for Phosphorus Removal The current source for commercial phosphorus is still primarily phosphate rock. However, given that around 161 million tonnes of phosphate (expressed as P2O5) were extracted in 2008 (Gilbert, 2009), and the highest-grade (with respect to percentage of phosphorus) deposits are rapidly being depleted, it is expected that economically viable mining could last or be depleted within the next few decades. With global oil reserves dwindling and increased cost for petroleum products, new technologies involving the use of corn as a biofuel have brought about an increased demand for fertilizers. A recent study suggested that by 2033, the demand for phosphate-based fertilizers would exceed supply, forcing an increase in the price of fertilizers, which in turn would increase the cost of food to the rapidly growing population (Cordell et al. 2009). Another factor that influences the price of fertilizers is the transportation cost, and that is related to the increasing global price of oil. Although phosphorus in the ground is limited, other potential sources, such as human sewage and animal manure and slurries, could be used in the future to supplement the demand for the Chapter Two: Background and Literature Review 7  element. For phosphorus to be useful, it must be in a form which is both technically and economically recoverable. Although current recovery methods are not always economical, the processes offer a sustainable source of phosphorus. Numerous studies have already been completed and more are being conducted every day, to identify new sources of phosphorus.  In most cases, phosphorus is one of the limiting nutrients that may cause eutrophication in freshwaters. The growth of algae leads to a decrease in the aesthetic values and usefulness of the water bodies. Phosphorus concentrations as low as 0.01 mg/L have been known to initiate eutrophication (Lee, 1970); on the other hand, 4-15 mg/L of phosphorus in untreated domestic wastewater is common (Tchobanoglous et al. 2003). It was estimated that as a result of human activity, more than 12,000 tonnes of P entered fresh, ground and coastal waters of Canada in 1996. The largest point source was municipal sewage, which added an estimated 5600 tonnes of P (Chambers et al. 2001). In general, the current discharge limit on total phosphorus in North America ranges from 2 to 0.1 mg/L (Tchobanoglous et al. 2003).  To combat the increase in eutrophication occurrences and other related effects from discharge of municipal sewage into water bodies, governments all over the world are placing stringent regulations on nutrient discharge criteria; this has brought about the growth of several biological nutrient removal (BNR) plants. Despite being removed from wastewater and into sludge during treatment, phosphorus is re-released into the liquid phase during anaerobic digestion of the sludge. Various studies show that 26% to 90% of the phosphorus at the head of the treatment plant is due to phosphorus feedback, that is, phosphorus in the return liquors (Jaffer et  al. 2002; Mavinic et  al. 1998). Some plants have even reported recirculated phosphorus loads of up to 100% (Pitman et al. 1991). This liquid phase, called supernatant or centrate, depending on dewatering process at the plant, is usually re-routed back to the front of the treatment process. The net result is that phosphorus is never fully removed by the treatment process, and it accumulates in the treatment plant. This leads to an increase in the phosphorus inventory in the plant, and is one of the primary causes of scaling or encrustation of piping and equipment. Although the excess phosphorus can be chemically precipitated, it produces extra sludge volumes that are costly to dispose of. In addition, Chapter Two: Background and Literature Review 8  increased phosphate loads can harm the optimal operation of a BNR process, which depends heavily on the BOD:P ratio of the wastewater (Mavinic et al. 1998).   A major problem in many wastewater treatment plants is the accumulation of different forms of phosphate precipitates, such as struvite, vivianite, hydroxyapatite to name a few, in various sections of the treatment system. Among the precipitates, struvite tends to be most prevalent. Struvite can form in locations having high turbulence, such as pump impellers and pipe bends and in piping and other equipment (e.g. digestion tanks pumps, valves, etc). Struvite precipitation occurs when the combined concentrations of Mg2+, NH4+ and PO43- exceeds the struvite solubility product. The removal of struvite from these locations is difficult, expensive and time consuming, and often requires the parts to be taken out of service, repaired or replaced. Cleaning of parts often requires the use of corrosive and concentrated acids and the use of a hammer and chisel. It was reported that the annual costs for a mid-size treatment plant (3785 cubic meter or 25 MGD) related to struvite accumulation exceeded 100,000 US dollars (Doyle et al. 2000). Error! Reference source not found. illustrates the presence of struvite in clogging post-digestion piping treatment plants.   Figure 2.1. (a) Struvite in pipe leading to centrate pump at Lulu Island WWTP, (b) Struvite in adjacent piping to the digester pump at Annacis Island WWTP (courtesy of Brian Hystad, Metro Vancouver, 2007). Chapter Two: Background and Literature Review 9   Although the most common practice for removing phosphorus in wastewater treatment plants is the use of chemicals to precipitate out the element, the process increases sludge production, and consequently increases sludge disposal costs.  Another method is the use of enhanced BNR (EBNR) process. In recent times, focus has shifted from removal to recovery of phosphorus through different technologies (CEEP, 2001; Jeanmaire, 2001; Seckler et al. 1996).  In a feasibility study, Woods et al. (1999) estimated that, when phosphorus recovery is carried out, sludge volumes could be reduced by up to 30% (compared to EBNR) and up to 49% (compared to chemical phosphorus precipitation). Another study conducted by Jeanmaire and Evans (2001) concluded that a decrease in sludge mass of 2-8% could be expected, if phosphorous recovery was undertaken at an operating BNR facility, with anaerobic sludge digestion.  The above mentioned problems, related to the increase of phosphorus inventory in wastewater treatment plants, have made recovery of the element an alternative option to the traditional chemical phosphorus precipitation. Although the value in recovering phosphorus from wastewater is straight forward, two basic considerations have to be considered before the method(s) is implemented. The first is the potential for cost savings - in terms of reduced chemical additions in treatment plants (where used), reduced struvite occurrences leading to downtime, and clean up and sludge handling costs – versus the cost of production that includes capital and operating costs. The second is the demand for this phosphorus (from wastewater), which in turn will depend on the availability, demand and cost of mining phosphorus from ores, and the potential revenue from the sale of phosphorus as struvite. There is a general consensus among wastewater experts that the cost of sludge disposal will increase significantly in the coming years, because of limitations to agricultural spreading of sludge, costly land filling and the increase in transportation costs. The recovered product may be used either as a source of phosphorus (instead of phosphate rock) or as a fertilizer (in case of struvite precipitation). Most of the recovered product is of higher purity than the phosphate rock and contains lower quantities of heavy metals (Fattah et al. 2008b), a factor that may increase the value of the product. Struvite has also been found to be a good fertilizer because it can release nutrients slowly. Extensive studies focusing on comprehensive evaluation of the economics related to phosphorus removal and recovery have shown that the Chapter Two: Background and Literature Review 10  value of phosphorus-based fertilizers continues to increase, and recovery of the element is crucial for a sustainable future (Dockhorn, 2009; von Horn and Sartorius, 2009). 2.3 Methods of Phosphorus Recovery  Within the last decade, the interest in phosphorus recovery has grown tremendously, with  research  topics  ranging  from  use  of  chemicals  in  wastewater,  to  recovery  studies involving both domestic and animal wastewater. Despite the widespread interest, there are only a handful of full scale plants in the world recovering phosphorus. Potential technologies for phosphorus recovery from wastewater include: calcium phosphate precipitation, struvite precipitation, aluminum and iron precipitation and membrane or ion exchange technologies, followed by precipitation.  Among the methods available for phosphorus recovery, precipitating phosphorus as calcium phosphate may be the most promising (Driver et al. 1999), because the forms produced are close to the forms found in mined phosphate, and can thus be recycled by the existing thermal or wet route processes of the P-industry. Numerous studies involving the use of calcium as a medium to remove and recover phosphate have been conducted using both domestic and animal wastewater. Rendl (2007) provides a summary of the various studies regarding the use of calcium to precipitate phosphate, as hydroxyapatite. A full-scale fluidized bed reactor, the DHV Crystalactor TM, was in operation in the Netherlands, where the solid was precipitated on a seeding grain, usually sand, with up to 11% P content (Seckler et al. 1996). The Kurita fixed bed crystallization column process uses phosphate rock as seed material and is based on the same chemistry as that of the CrystalactorTM process. The difference between the two processes is that the Kurita process uses magnesium hydroxide as the source of magnesium, whereas the DHV process uses magnesium chloride (Greaves et al. 1999). Vanotti and Szogi (2009) applied lime to recover phosphorus from livestock wastewater both at pilot-scale and full-scale, in North Carolina, USA.  Although struvite cannot replace mined phosphorus used by the P-industry, it can reduce the demand for the element, as it can be used directly as a fertilizer. In recent times, several Chapter Two: Background and Literature Review 11  studies have been conducted involving the removal-recovery of phosphate as struvite. A number of full-scale plants, utilizing fluidized bed reactors, are already in operation to recover phosphorus as struvite. The available processes include the Unitika Phosnix Process (Stratful et al. 1999)  and  the  Ostara  Nutrient  Recovery  Process.  The  latter  process  was developed  at  The  University  of  British  Columbia  (UBC),  Canada.  Details  of  recovering phosphate in the form of struvite are described in subsequent sections.   Ferric chloride is more widely used in Europe for P-removal, but the process is not sustainable because it cannot be used in existing P-industry processes and probably has low or zero fertilizer value (Jeanmaire, 2001). The AlPO4 form  is  more  promising  as  it  can  be recycled in the P-industry, using the thermal route.  Technologies involving the use of membrane or ion exchange followed by precipitation have also been tested successfully and commercialized. An example of this technology is the REM-NUT process (Liberti et al. 2001) that uses ion exchange, followed by struvite precipitation. This process removes phosphate and ammonium ions from tertiary wastewater. 2.4 Advantages of Phosphorus Recovery as Struvite The application of phosphorus recovery from municipal wastewater, through struvite (magnesium ammonium phosphate, MAP) precipitation, provides a viable and sustainable alternative to mined rocks as a source of phosphorus (Jaffer et al. 2002; Berg, 1982), and offers both environmental and economical benefits. Controlled and intentional struvite precipitation in wastewater treatment plants, especially those employing biological nutrient removal (BNR) technologies, also provides necessary phosphorus removal from the system. The reduction of phosphorus inventories can also reduce the probability of unintentional struvite formation, a costly nuisance that is common in wastewater treatment plants. Studies have shown that more than 90% of dissolved phosphorus can be removed from anaerobic digester supernatant and centrate in the form of struvite precipitation (Fattah et  al. 2008b; Britton et al. 2005). Although effective, this method is costly, due to the need for large amounts of caustic to keep the pH of the system steady. The pH is especially important in Chapter Two: Background and Literature Review 12  preserving the reactor supersaturation ratio (SSR) (Fattah et al., 2008b). However, struvite crystallization at wastewater treatment plants employing biological phosphorus processes contains high amounts of phosphate in the digester effluent, and therefore, may require little or no caustic addition. Thus, the success of introducing struvite crystallization processes in wastewater treatment plants will mostly depend on its economical sustainability. The operational costs of struvite crystallization mainly depend on two factors - costs of chemicals and  energy  requirements  for  pumping.  In  their  study,  (Jaffer et  al. 2002) showed that, compared to the cost of chemicals, the energy costs are quite insignificant, and 97% of the total chemical cost was due to the addition of caustic, which was used to achieve the desired operating pH. However, when determining the economics of struvite recovery, attention must be given to the potential savings, such as lower operational downtime, lower sludge production, and improved process operation, to name a few.  Struvite is an effective slow release fertilizer that can be used for agriculture (Owen et al. 2009; Ponce and De Sa, 2007; Gaterell et al. 2000). The beneficial effect of the slow release is that it allows the possibility of lower rates of application. With lower phosphate reserves and increasing fertilizer costs, the application of crystallization to recover phosphate as struvite from wastewater provides a long term solution and a sustainable approach. A major advantage of struvite recovered from wastewater for use as fertilizer is the presence of low heavy metal concentrations, such as for cadmium or uranium (Fattah et al. 2008b). With the quality degradation of phosphate rock, the heavy metal content is likely to increase; current fertilizer production from phosphate rocks does not eliminate heavy metals, and upgrading the processing plants is not very economical (von Horn and Sartorius, 2009).  Human activities during the past century, particularly intensive forest practices, fishing, urbanization, industrialization and impoundment construction, have had a negative impact on the health of British Columbia’s numerous wild salmonid stocks (Slaney et al. 1996). Overharvesting and alteration of salmonid habitat reduces the return of spawning adults which naturally fertilize streams for their progeny (Larking and Slaney, 1996). Consequently, there is a lack of marine-derived nutrients to freshwater habitats, resulting in nutrient deficient streams. Oligotrophic stream conditions, both human-induced and naturally- Chapter Two: Background and Literature Review 13  occurring in granitic coastal systems, can be made adequately fertile with low level addition of nutrients. Struvite has been found to be an effective, slow release fertilizer, to increase the nutrient levels in streams, and more tests are underway to fertilize inland water bodies in British Columbia (Sterling and Ashley, 2003). 2.5 Chemistry of Struvite Struvite, magnesium ammonium phosphate (MgNH4PO4•6H2O) is a sparingly soluble crystal that contains equimolar amounts of magnesium, ammonium and phosphate, bound together by six waters of hydration, and forms according to Equation 2.1. During the formation of struvite in wastewater systems, many side reactions also occur concurrent to struvite formation. These include the interactions of each of the various species present in the wastewater; some common ones are summarized in Table 2.1. In wastewater systems, many other different species are present which may indirectly influence struvite equilibrium. Other researchers have suggested interaction of various other ionic species, such as H2CO3, CH3COO-, CH3COOH, Mg3(PO4)2.8H2O (Loewenthal et  al. 1994), K+, Cl-, Ca2+, Na+,  K+ (Gadekar et al. 2009).  Mg+2 + NH4+ + PO4-3 + 6 H2O ļ MgNH4PO4•6H2O                                                 (2.1) Table 2.1. Competing reactions in struvite formation in a distilled water system Equilibrium pKsp (25 oC) ǻHo (kJ/mole) Reference MgOH+ ļ Mg+2 + OH- 2.56 -10.89 Morel et al. 1983 NH4+ļ H+ + NH3 9.3 52.21 Snoeyink et al. 1980 H3PO4ļ H2PO4- + H+ 2.15 -10.14 Martell et al. 1989 H2PO4-ļ HPO4-2  + H+ 7.2 4.90 Martell et al. 1989 HPO4-2ļ PO4-3  + H+ 12.35 13.87 Martell et al. 1989 MgH2PO4+ļ H2PO4- + Mg+2 0.45 -14.25 Morel et al. 1983 MgHPO4ļ HPO4-2 + Mg+2 2.91 -13.83 Taylor et al. 1963 MgPO4-ļ PO4-3 + Mg+2 4.8 -12.99 Childs, 1970 H2O ļ H+ + OH- 14 55.94 Martell et al. 1989  Chapter Two: Background and Literature Review 14  2.6 Solubility of Struvite Struvite production for recovering phosphorus from domestic wastewater has gained substantial interest and progress in recent times. However, discrepancies continue to exist between reported values of some of the most important operating parameters for struvite crystallization. One such parameter is the solubility product or Ksp, more commonly referred to  as  the  pKsp (-log Ksp) value. This parameter is important in the determination of supersaturation ratio (SSR), the parameter that determines if struvite formation is possible or not. SSR is defined as the ratio of the ion activity product (IAP) to the equilibrium solubility product of struvite (Kspeq) (Equation 2.2a). Sometimes, for simplicity, a supersaturation ratio, called conditional SSR, SSRc, (Equation 2.2b) is also calculated as the ratio of conditional solubility (PS) to the equilibrium conditional solubility product (PSeq). In the conditional product, the ionic activity of the species is not considered.  SSR = IAP/Kspeq      (2.2a) SSRc = PS/ PSeq          (2.2b) where, IAP = {Mg+2}{NH4+}{PO4-3}     (2.3a)                (2.3b)  Į is the ionization fraction and Ȗ is the activity coefficient.  The [ ] brackets indicate concentration, while the {} brackets indicate ionic concentration in moles per liter, corrected for activity.  The determination of ion activity product involves the speciation of analytically determined concentrations using published acid and base dissociation constants, as well as an adjustment for activity. The activity is a function of the concentration of the ion and its activity coefficient, Ȗ. The activity is given by the Güntelberg approximation of the Debye-Hückel equation shown in Equation 2.4 (Sawyer et al. 1994). The ionic strength of the solution can be determined based on electrical conductivity (EC) 3- 4 3- 4 + 4 + 4 22 4 POPONHNH 3- 4 +2+ Ksp = ][PO ][NH ][Mg = Ps JDJDJD  MgMg Chapter Two: Background and Literature Review 15  measurements. Additional coefficients have been suggested relating ionic strength and EC. By processing data from various studies carried out at the University of British Columbia using centrate, supernatant and synthetic waters, Rahaman et al. (2006) suggested a value of 5×10-6 for the activity coefficient. Considering the temperature dependence of EC, an activity coefficient of 7.22×10-6 was suggested by Bhuiyan et al. (2009).                                       (2.4) where, Ȗ = the activity coefficient for the species of interest z = the ionic charge of the species of interest ȝ = ionic strength  = 1.6×10-6EC  Although Ksp is theoretically constant, Kspeq is highly correlated with pH, due to the changing component concentrations, each time a new equilibrium is reached. If SSR is to be used as a control variable for struvite recovery, it is essential to know the true value of Ksp for the  pH range  that  the  systems are  expected  to  operate  in.  The  SSR of  the  bulk  fluid  is  the primary control variable used by the P-recovery team at the University of British Columbia (Fattah et al. 2008b; Forrest et al. 2008a; Forrest et al. 2007). Although extensive studies on the  value  of  Ksp of struvite have been conducted, there still exists significant variation between reported values: 5.50 x 10-14 - 3.89 x 10-10 corresponding to pKsp values of 9.41 to 13.27 (Rahaman et al. 2006). This variation may be related to the large range of experimental methodologies. The standard method for the experimental determination of a Ksp value of a particular reaction involves either the formation of precipitate or the dissolution of a previously formed salt in distilled water. In either approach, experiments are conducted under carefully controlled conditions in which constant mixing energy, constant pH, constant temperature or a set conductivity is maintained. Widely varying procedures in experimental methodologies account for much of the discrepancy that exists between the reported values for the struvite solubility product. In addition, some of the studies neglected the influence of ionic strength in the determination of the solubility product. Some additional factors that may also influence the value of Ksp are discussed by Rahaman et al. (2006). P PJ   1 5.0 log 2z Chapter Two: Background and Literature Review 16  2.7 Factors Affecting Struvite Solubility 2.7.1 pH Struvite solubility depends on the concentrations of the constituent ions. Formation or dissolution of struvite in a water matrix depends on its solubility, at a particular pH, to the equilibrium solubility product. The availability of the species in IAP depends on the pH of the  system;  at  high  pH,  the  concentration  of  ammonium  ion  (NH4+) decreases rapidly, converting to ammonia (NH3). The proportion of orthophosphate (PO4-3) ions also varies with  the  pH,  and  at  a  pH  value  of  9.5  and  above,  magnesium  ion  converts  to  MgOH+. In addition, depending on the pH, and in the presence of phosphate, magnesium forms a number of complexes such as MgH2PO4+, MgPO4- and MgHPO4.   In general, the solubility of struvite decreases with increasing pH of the solution. However, above pH 9.0, an increase is solubility was suggested because of the decrease in ammonium concentration (Booker et al. 1999; Buchanan et al. 1994; Snoeyink and Jenkins, 1980). Other studies found a minimum solubility at a pH of 10.3 (Ohlinger et al. 1998). One of the main reasons for this discrepancy in solubility is the selection of different Ksp values. 2.7.2 Temperature Similar to the pH effect, there are different opinions on the effect of temperature on solubility of struvite. Maximum solubilities occurring at 20°C (Durrant et  al. 1999), 30°C (Bhuiyan et al. 2007) and 50°C (Aage et al. 1997) have been suggested. Doyle and Parsons (2002) found that at high temperature, the structure of struvite pellets changed, which affected its solubility. Burns and Finlayson (1982) determined the Ksp at different temperatures and showed that Ksp increased with increasing temperature in the range from 25°C to 45°C.  The effect of temperature on the solubility of a substance can be related to the enthalpy of reaction. The enthalpy change of a chemical reaction (ǻH°) is the amount of heat that is released or taken up during the course of the reaction (Snoeyink and Jenkins, 1980). Burns Chapter Two: Background and Literature Review 17  and Finlayson (1982) reported an enthalpy value of 24.23 kJ.mol-1 for struvite formation, showing that the formation of struvite is endothermic. Standard enthalpy change values, ǻH°, for reactions are most commonly used in water chemistry to determine the effect of temperature on the position of equilibrium. A useful expression relating ǻH° to Ksp is given by the van’t Hoff equation (Equation 2.5 to 2.7) (Snoeyink et al. 1980).  2 °ln RT H dT Kspd '    (2.5) or, 21 21 1 2 ° )( )(ln TT TT R H TKsp TKsp u u'       (2.6) or,  ¸ ¸ ¸ ¸ ¹ · ¨ ¨ ¨ ¨ © §   '  12 11 10ln° 12 TT RH pkpk    (2.7) where,  T = temperature (in K) pk1 and pk2 = solubility constants at temperature T1 and T2, respectively ǻH° = enthalpy of reaction (J/mol)  R = ideal gas constant (8.314 J/mol.K). 2.8 Chemical Requirements for Struvite Crystallization Process 2.8.1 Caustic The formation of struvite is accompanied by a lowering of pH in the system, and for efficient  operation  of  a  struvite  crystallization  process,  control  of  the  pH is  paramount.  pH plays an important role in maintaining the supersaturation ratio, one of the key operating parameters, and controlling it by manipulating the pH has been preferable preferred approach Unintentional struvite growth in wastewater treatment plants is sometimes removed by changing  the  pH  of  the  location  -  providing  acidic  conditions,  as  struvite  is  soluble  under acidic conditions and highly insoluble under alkaline solutions; minimum solubility of Chapter Two: Background and Literature Review 18  struvite occurs at pH 10.3 (Ohlinger et al. 1998). Several methods have been used to increase the  pH of  a  struvite  crystallization  system –  from using  caustic  (NaOH) to  carbon dioxide stripping (Fattah et al. 2008a; Jaffer et al. 2002; Battistoni et al. 2001; Munch and Barr 2001; Ohlinger et al. 1999), although the former is more prevalent in use. For optimized struvite crystallization from wastewater, a pH in the range of 7.5-9.0 is usually suggested. Some recommended pH values found in the literature are given in Table 2.2. 2.8.2 Magnesium addition For phosphorus precipitation as struvite, equimolar concentrations of phosphorus and magnesium are necessary. In most wastewaters, struvite precipitation is limited due to insufficient magnesium, and, as a result, an external source of magnesium is added during intentional  struvite  precipitation.  Primarily,  there  are  two  sources  commonly  used  to supplement Mg requirements – magnesium hydroxide (Mg(OH)2) and magnesium chloride (MgCl2). The chloride form is sometimes preferable because it is easier to transport as it can be obtained as pellets, and it dissociates faster than the hydroxide, therefore requiring a shorter reaction time (Jaffer et al. 2002).  Mg(OH)2 has the advantage of being cheaper and can also help raise the pH, thereby requiring less caustic. However, using magnesium hydroxide to serve both the functions of magnesium dose and pH increase means that one cannot be optimized independent of each other (Jaffer et al. 2002; Munch and Barr, 2001). Sea water has also been used successfully as a magnesium source, without affecting the overall performance of the process (Kumashiro et al. 2001). Some sources of magnesium with their corresponding pH values are given in Table 2.2. For optimal P-removal, there should be an excess of soluble magnesium, that is, greater than stoichiometric requirements. Although, theoretically, a Mg:P molar ratio of 1:1 is required for struvite precipitation, in most cases for efficient growth of struvite, the required ratio is in the range of 1.3-2.0 (Jaffer et al. 2002; Munch and Barr, 2001). The influence of magnesium on struvite formation is discussed in detail in Chapter Five.     Chapter Two: Background and Literature Review 19  Table 2.2. Sources of magnesium studied with their corresponding pH values Base added    Magnesium source Suggested pH value Reference NaOH MgCl2 7.7-8.2 Fattah et al. 2008b; Britton et al. 2005; Forrest et al. 2008a NaOH and CO2 stripping MgCl2 7.8-8.3 Fattah et al. 2008a NaOH, Mg(OH)2 MgCl2, Mg(OH)2 pH value • 8.5 Jaffer et al. 2002 NaOH MgCl2, MgO 8.5< pH value < 9 Celen and Turker, 2001 NaOH Seawater pH value ~ 7.7 Kumashiro et al., 2001 Only CO2 stripping (if alkalinity is low) Not required 8.2< pH value <8.8 Battistoni et al. 2001 2.9 Reduction of Gases by Stripping In wastewater treatment plants, struvite precipitation usually occurs at locations where carbon dioxide is stripped from the solution, with a corresponding increase in pH. This reaction (Equation 2.8) also takes place naturally in aquatic environments through the uptake of CO2 by algae. Detailed chemistry, relating the increase of pH through carbon dioxide stripping, can be found in Cohen and Kirchmann (2004). Areas of high turbulence, such as pipe elbows, mixer blades, valves, and pumps are main locations of struvite deposits (Neethling and Benisch, 2004). In these locations, a reduction of partial pressure of CO2 takes place. The relationship between partial pressure of a gas and its concentration in liquid is given by Equation 2.9. Hence, Loewenthal et al. (1994) concluded that the partial pressure of CO2 is one of the driving forces for struvite precipitation. In their study, Pitman et al. (1991) demonstrated the possibility of increasing pH with aeration. This phenomenon was attributed to CO2 stripping. In a different study, Loewenthal et al. (1994) showed that the partial  pressure  of  CO2 controls struvite precipitation inside an anaerobic digester. Using anaerobic supernatant as feed, Battistoni et al. (1997) investigated the application of air Chapter Two: Background and Literature Review 20  stripping to raise the pH in a struvite crystallizer using two modes of stripping - external gradual aeration (EGA) and external continuous aeration (ECA). The study found that the pH of  the  system could  be  increased  from 7.9  to  8.6  with  increase  in  upflow rate,  (which  was increased from 1.8 L/min to 5 L/min), while keeping the airflow rate constant at 15 L/min. They were able to obtain up to 80% phosphorus removal by the ECA method only. The EGA method was not as efficient and was unable to remove phosphorus rapidly.  H+ + HCO3-ĺ H2CO3ĺ CO2Ĺ + H2O    (2.8) Pi = Xi . P   (2.9) where, Pi is the partial pressure of the individual gas in the gas mixture Xi is the mole fraction of the individual gas component in the gas mixture P is the total pressure of the gas mixture  Carbon dioxide stripping to increase pH in a pilot scale struvite crystallization system using centrate as process feed has been shown to be an efficient method in reducing overall process cost, by as much as 46-65% (Fattah et al. 2008a). Hiroyuki and Toru (2003) used 1 L bench scale reactors to demonstrate the affect of aeration on phosphorus precipitation. Along with aeration, NaOH solution was also fed continuously to increase the pH. They found that, by increasing the aeration intensity from 2.l to 10.5 mg/L, the rate of phosphorus removal also increased. The authors concluded that aeration influenced the quantities of CO2 in solution, which helped to raise pH, and resulted in an increased rate in phosphorus removal.  Ammonium ions in wastewater exist in equilibrium with gaseous ammonia. As the pH of the wastewater is increased above 7, ammonium ion is converted to ammonia, which may then be removed by air/gas stripping. However, the Henry’s law constant of ammonia is only 0.75  atm  (mol  H2O/mol air), which makes it hard to strip this gas (Tchobanoglous, 2003; Musvoto et al. 2000). During stripping of gases from wastewater, ammonia and CO2, the stripping rate for CO2 is higher by two orders of magnitude than for ammonia. This happens since the Henry’s law constant for ammonia is much lower than for CO2 (the dimensionless Henry’s law constants for ammonia and CO2 are 0.011 and 0.95, respectively). Chapter Two: Background and Literature Review 21  Various  factors  determine  the  rate  of  ammonia  stripping  from  wastewater,  such  as  pH, temperature, relative ammonia concentrations, and agitation of the air-water interface. Theoretically, the rate of stripping is proportional to the increase in these factors. At room temperature, a high efficiency ammonia removal usually takes place at pH up to 10.5-11.0 (Liao et al. 1995). With a decrease in temperature, the amount of air required increases significantly for the same degree of removal (Tchobanoglous, 2003). However, it is important to limit the amount of air flow for stripping because it can lead to a cooling effect, which in turn reduces stripping (Liao et al. 1995). One of the benefits of removing ammonia by stripping is that it can be recovered by absorption in sulfuric acid, and then used by the fertilizer industry. Therefore, control of the pH is important when choosing air stripping of gas for the purpose of increasing the pH, as a high operating pH tends to reduce higher amounts of ammonia, which in turn suppresses the tendency for pH to increase, due to carbon dioxide stripping. 2.10 Determination of Magnesium Usage for Struvite Precipitation As mentioned earlier, in wastewater treatment systems, magnesium is normally the limiting ion for struvite precipitation, and it is commonly supplied by adding an external magnesium source. In the present study, magnesium deficiency was overcome by adding magnesium chloride. However, it is important to determine the magnesium requirements as insufficient addition limits phosphate removal, while excess addition results in higher concentration in the effluent; this, in turn, increases struvite formation potential downstream of the system and reduce struvite purity (Demeestere et al. 2001).  The common method for the determination of Mg involves costly atomic absorption (AA) or inductively coupled plasma (ICP) tests; the downside of these methods in a continuous struvite crystallization process is the lag time between sampling and analysis. For efficient process control and optimal product quality, it is imperative that the Mg concentration in the reactor and process feed be known in real time, or at least be measured by  a  process  that  requires  little  time  to  provide  results.  Although  the  determination  of magnesium is important because of its widespread use, little work has been reported on the Chapter Two: Background and Literature Review 22  development of ion-selective electrodes (ISEs) for magnesium (Gupta et  al. 2002). Some applications of the sensors are in molten metal processing and for biological liquids, such as blood. Of the few electrodes that have been reported, the reliability in measuring in-line magnesium is reduced due to interference from other metals, especially calcium and potassium (Gunzel and Schlue, 2002). An added disadvantage of the use of magnesium ISE in the wastewater industry is the fact that these few probes are easily fouled.  Magnesium can also be determined indirectly by calculating the water hardness, assuming that the hardness is a function of only calcium and magnesium ions. By determining the total hardness and calcium hardness, magnesium hardness can be estimated, which can then be used to estimate the magnesium concentration. However, this technique has not been used in the operation of a struvite crystallization process till now.  In the absence of a suitable and reliable online analyzer for the determination of magnesium, the quantity of magnesium required for suitable struvite growth can be determined indirectly. This can be accomplished by introducing the phenomenon of pH and electric conductivity changes in accordance with the struvite reaction. Conductivity is a direct measurement of the ions in solution, and is proportional to the type and number of ions (Shepherd et al. 2009). Since formation of struvite reduces the number of ions, this in turn is expected to reduce the conductivity of the solution. The relationship between ionic concentrations and specific conductivity, ț , can be calculated from Equation 2.10 (Shepherd et al. 2009).  ț = Ȟ * N/1000   (2.10) where, N = normality of the solution (eq/L) ț = specific conductivity (S/cm) Ȟ =  equivalent conductance of the solution (S m2/ mol)  For simplicity, the theoretical conductivity of the present system was assumed to derive from only magnesium, chloride, ammonium and phosphate ions (Equation 2.11). The initial Chapter Two: Background and Literature Review 23  theoretical conductivity was corrected to correspond to the actual conductivity of centrate, so that both theoretical and practical conductivity values had the same starting point.  ț  = [Mg+2] Ȟ0 Mg+2 + [Cl-] Ȟ0 Cl- + [PO4-3] Ȟ0 PO4-3 + [NH4+] Ȟ0 NH4+    (2.11) where, ț = specific conductivity (S/cm) Ȟ0 =  Molar (equivalent) conductance of the solution at infinite dilution (S m2/mol). Values are taken from Lide (1991).  The MAP reaction with magnesium chloride (MgCl2) is obtained from Equation 2.12. The theory behind this indirect determination is as follows. When MgCl2 is added to centrate (or crystallizer constituents), electric conductivity increases as the chloride ion concentration increases. However, it has been shown (Sasai et al. 1995) that the increase in conductivity up to the completion of the MAP reaction, when theoretically all phosphate has been removed, is slower than after the equivalence point – the point where theoretically all the phosphate has been precipitated. As a result, there exists a bending point (details in Section 5.11.2) at the equivalence point. On the other hand, since hydroxide ion is consumed during the reaction, the pH decreases as MAP reaction proceeds. That is, there exists another bending point for pH. Thus, by determining the bending points for completion of the MAP reaction, the amount of magnesium needed can be determined. In order to apply this approach, it will be necessary to calculate the location of the bending point, and then calculate the magnesium concentration required. One advantage of this method is that there is no need to measure the individual magnesium concentration in the wastewater (digester supernatant or centrate) and the matrix present in the reactor.  Mg+2 + 2 Cl- + NH4+ + PO4-3 + 6 H2O  ļ MgNH4PO4 · 6H2O + 2 Cl-    (2.12) Despite the need for a quick and on-site determination of magnesium addition, there are no known methods available to accomplish this requirement. Therefore, the approaches mentioned above could possibly be used in determining the concentration and amount of external magnesium addition to a struvite crystallization process. Chapter Two: Background and Literature Review 24  2.11 Morphology of Struvite Pellets Among the methods used to detect the presence of struvite in precipitates from phosphorus recovery processes, characterization by X-ray diffraction (XRD) and by scanning electron microscopy (SEM) is common. Struvite is a white crystalline substance that has a distinctive orthorhombic structure (Doyle and Parsons, 2002). The structure can be identified via X-ray diffraction (XRD) by matching the intensity and positions of the peaks produced to a database for the crystal structure. Depending on the conditions during growth, struvite pellets can be variable from equant, wedge shaped, short prismatic, to thick tubular (Durrant et al. 1999). One of the drawbacks of this determination is that it does not provide information  on  the  composition  of  the  struvite  particles.  To  determine  actual  ionic concentrations, the common practice of dissolving the struvite, and then measuring the concentrations, still has to be carried out. Reports on the spontaneous formation of struvite in supersaturated solutions suggest that the pellet structure of struvite depends on the solution pH, the solution supersaturation, the Mg:P molar ratio, impurities in solution and the kinetic factors (Bouropoulos and Koutsoukos, 2000; Wierzbicki et al. 1997; Abbona and Boistelle 1979).   Struvite solubility is dependent on the solution pH and temperature, and these factors highly influence struvite formation and morphology (Al-Jibbouri et al. 2002; Booker et al. 1999). Abbona and Boistelle (1979) found that, at very high levels of supersaturation, bi- dimensional and tri-dimensional twinned pellets were formed, and on changing the supersaturation from high to low, the pellet structure changed from tubular to increasing elongation. However, studies made by Bouropoulos and Koutsoukos, (2000) found that the degree of supersaturation did not have any impact on the morphology of the pellets.   Studies on struvite pellets have found that the sizes of the pellets were influenced by the Mg:P molar ratio and the magnesium concentration in the effluent (Fattah et al. 2008b; Bouropoulos and Koutsoukos 2000). A study by Hirasawa et al. (1997) found that at Mg:P molar  ratio  of  2,  the  crystals  agglomerated,  resulting  in  large  crystals.  On  the  other  hand, when the molar ratio was increased to 4, together with fine crystals, needle-like crystals were Chapter Two: Background and Literature Review 25  formed. The presence of calcium in the wastewater was found to influence the growth of struvite pellets by acting as a binding agent (Le Corre et al. 2005). Another important parameter that influences crystal growth is the mixing energy in the crystallizer, which, depending on the value, can enhance or reduce pellet size. Adequate mixing energy is required to enhance mass transfer of the constituent ions to form struvite. On the other hand, a high mixing energy, brought about by higher upflow velocities in the crystallizer, can increase pellet-pellet collisions, which in turn can lead to attrition of the larger pellets. 2.12 Quality of Struvite Pellets and its Applications Although the recovery of the phosphate is relatively simple, conditions required for the production of “good quality” struvite have not been studied in great deal. The quality of struvite has to be of certain grade to be used as a fertilizer or for nutrient enrichment of water bodies. Quality can refer to the composition, size, shape and crushing strength of the pellets formed. It is important to determine and understand how and why crystals bind to form pellets. Pellet formation in animal feed preparation has been documented (Thomas and vander Poel, 1996) where a three-phase condition was suggested. It was concluded that the binding strength of the pellets was higher with decreasing radius of the particles, and that the finer the grind, the better the production of pellets. In struvite formation in a crystallizer, a liquid-solid two-phase condition is normally assumed, but the assumption of better pelletization with smaller crystals probably holds true. In addition, solid-solid interactions between particles, through collisions, also factor into pelletization when the distance between the two is small, such as when the struvite crystallizer’s pellet inventory is high.  The dissolution of struvite, a sparingly soluble material, is generally controlled by surface reaction and therefore, the size (0.5-2.0 mm) of the struvite pellets influences the dissolution kinetics (Bhuiyan et al. 2008). Smaller pellets, having a greater surface area to volume ratio, are  more  easily  dissolved  than  larger  pellets,  under  the  same  conditions.  These  dissolution rates are important for struvite applications. The use of struvite as an agricultural fertilizer may demand lower dissolution rates for slow release, where a single high-dose may be applied without burning of the crop. On the other hand, small pellets are desirable for stream Chapter Two: Background and Literature Review 26  nutrient enhancement that will provide “relatively faster”, slow-release of nutrients. In addition to the size requirements, the formation of pellets with stronger crushing strengths is desirable for transport and application of the struvite in the field. The pellets also have to be durable, since fine or crushed struvite may lead to loss of material during manufacture, harvesting and application.  The quality of struvite can be described in terms of both composition/purity and crushing strength, and the requirements are different for varying applications. Struvite needs to have a considerably greater crushing strength when applied with a fertilizer spreader, than when used for stream nutrient enhancement. Thomas and Poel (1996) described different devices used to determine hardness of pellets for animal feed manufacturing.  There are also instruments in the pharmaceutical industry that measure crushing strength of pharmaceutical pills, but this equipment cannot be applied for struvite due to the struvite size range and the relatively lower crushing strengths of struvite pellets (Key International Inc.). In order to quantify crushing strength of struvite, it is necessary to develop a device that can be used on- site as a quick means of checking the strength for product quality assurance.  Various studies have developed struvite pellets from wastewater with sizes varying from less than 0.5 mm to as big as 6.5 mm (Le Corre et al., 2007; Fattah, 2004; Ueno and Fujii, 2001). However, few studies have been able to handle the production of fines and growing large pellets. There are different techniques to control and handle of the fines. A two-stage tank, where fines are grown to 300 µm in a sub-reaction tank before being fed to the main reaction tank, has been proposed by Shimamura et al. (2003). Others have used these fines as a seeding material in the crystallization process (Ueno and Fujii, 2001). A coagulant was also used by Le Corre et al. (2007) to increase the sizes of the fines so that they could be used in fluidized beds. Studies at UBC have successfully and consistently produced pellets as large as 4-5 mm (Fattah et al. 2008b; Forrest et al. 2008a). In most applications, larger pellets are desirable because they are easier to handle, transport and apply, whereas managing fine precipitates may result in loss of the material during the recovery to application processes. However,  there  are  still  needs  for  smaller  and  softer  struvite  pellets  as  well,  such  as  when applied to increase the nitrogen and phosphorus content of a nutrient deficient water body. Chapter Two: Background and Literature Review 27  Therefore, there is a need to determine conditions that can produce, consistently and without interruption, struvite of a particular quality and size. In addition to the determination of conditions,  another  important  factor  is  the  operation  of  the  struvite  crystallization  process. The importance of this topic is discussed in the next section. 2.13 Operation of Crystallizer for Struvite Formation Various studies involved the use of struvite precipitation as a means of removing and recovering phosphorus from both domestic and dairy wastewater. Struvite formation can be easily accomplished once the supersaturation (SSR) of the system, with respect to struvite, is greater than unity. For efficient precipitation and pelletization, the SSR value is normally kept between 3-5 (Fattah et  al. 2008b; Forrest et al. 2008a).  In  pilot-scale  or  full-scale application at wastewater treatment plants, variation in process feed characteristics is normal. Therefore, the development of a control system is important and necessary to: (i) counteract the variations in the process water, especially concentrations of those species that influence struvite precipitation and SSR directly, such as Mg+2, NH4+ and PO4-3, the pH, temperature and  conductivity  of  the  water  matrix;  and  to  (ii)  optimize  the  precipitation  process. Optimization is to provide conditions for smooth operation of the process, and to develop products that are of good quality.  As mentioned earlier, although a number process variables determine the supersaturation ratio of the system, SSR in a struvite crystallization process is usually controlled by either, or a combination of, three methods – controlling the pH, adjusting the external magnesium loading and controlling the recycle flow (the effluent flow from the crystallizer that is put back into the system, as illustrated in Figure 4.1). Controlling the pH seems to be preferable for running the process for phosphorus recovery as struvite, as the other methods for control are more complex and difficult. Varying the magnesium loading influences the SSR in two ways – by changing magnesium concentration in the system, and by changing the pH. Therefore,  it  is  difficult  to  optimize  one  without  influencing  the  other.  Control  via  recycle flow is also complex because it changes the mass loading of all the component ions, as well as influences the conductivity and temperature in the crystallizer. The optimum operational Chapter Two: Background and Literature Review 28  pH for different wastewaters varies greatly, depending on the particular waste stream. Although some of the literature cite pH between 8.2 and 9.0, to ensure higher (above 80%) phosphorus removals (Munch and Barr, 2001; Stratful et al. 2001), others were able to achieve over 90% phosphorus removal using different feeds, such as synthetic water, digester supernatant and centrate, at pH of 7.3 to 7.8 (Fattah et al. 2008b; Forrest et al. 2008a). Since the use of caustic for pH adjustment is expensive, a large fraction of the struvite crystallization operational costs can be reduced by controlling and efficiently using the chemical, as well as using other methods to increase the pH. One such method is air stripping of CO2 to aid in raising the pH. Results from a study by Fattah et  al. (2008a) found that, depending on the operating conditions, the stripper was able to reduce 46% to 65% of caustic chemical addition. In a study at the Treviso wastewater treatment plant, Italy, Cecchi et al. (2003) used air stripping to raise the pH to 8.5. The results demonstrated that, depending on the type of air stripping process applied, costs associated with caustic addition could be reduced significantly. 2.14 Modeling and Control of Struvite Precipitation Struvite crystallization is a complex process that involves a number of process variables, as well as complex process chemistry. In most cases, control of all the factors involved with the process is not possible. However, there exist some control variables, such as pH and flows,  which can effectively keep the process running smoothly and efficiently. As mentioned earlier, control of the process is usually accpmplished by keeping a constant SSR in the system. Predicting formation potentials and controlling struvite precipitation in crystallizers are important for both designers and operators at a treatment plant. Although various physico-chemical, equilibrium-based mathematical models related to struvite formation and its kinetics have been developed previously (Gadekar et al, 2009; Ali and Schneider 2008; Harada et al. 2006; Ohlinger et al. 1998; Loewenthal et al. 1994), none of the models actually involves process optimization or control.  Loewenthal et al. (1994) tested their struvite model on a synthetically-prepared solution that was designed to mimic anaerobic digester effluent. The model developed by Ohlinger et Chapter Two: Background and Literature Review 29  al. (1998) included more chemical reactions than the model of Loewenthal et al. (1994), and also included the influence of ionic strengths on species concentrations. Harada et al. (2006) used his model to predict formation of struvite in urine, and included more solid formations. Ali and Schneider (2008) developed a model to predict struvite formation and incorporated growth kinetics. They tested their model in batch experiments using synthetic solutions in a pilot scale reactor. At UBC, different chemical-based models have been developed that have been used successfully in process control, both in bench-scale and pilot-scale studies (Fattah et al. 2008b; Forrest et  al. 2008a; Adnan et al. 2003). However, since the models were not used in developing any control system, the studies lacked real-time control and the possibility of automation.  Many different software packages, such as, Matlab (Hanhoun et al. 2009, PHREEQC (Bhuiyan et al. 2008), MINTEQA2 (Ali, 2005), MINELQL+ (Ohlinger et al. 1999), ChemEQL (Bouropoulos et al. 2000) have been used to solve the equilibrium problem.  Each model performs an iterative analysis, using an internal thermodynamic database and user- defined input concentrations, to calculate the equilibria of all considered complexes. Kinetic- based modeling approaches have also been used to develop a kinetic model of struvite formation. A three-phase (aqueous/solid/gas) model was proposed by Musvoto et al. (2000) for application in anaerobic digester liquors. Black box models using artificial neural networks (ANN) have also been suggested for applications when it is difficult to determine, with accuracy, the constituent ionic characteristics (Forrest et  al. 2007). However, the downside of this method is that operators are unaware of the relationship(s) between various variables and the process condition, and therefore it is difficult to control the process. Despite their ability to predict struvite formation, none of the programs mentioned above are able to handle changes in process variables on a continuous basis. One group of researchers attempted to provide a solution to the change in variable problem mentioned by using Matlab’s fsolve function to solve 16 chemical reactions, but they did not have much success (Hanhoun et al. 2009). Later, although they reduced the problem size by substituting some variables, they found solutions ranging in an unfeasible physical domain. They finally found a solution, based on a rearranged set of equations so that a linear relationship was developed, Chapter Two: Background and Literature Review 30  and using multiobjective genetic algorithm to determine a good initial solution point, but concluded that the method was very inefficient. Chapter Three: Objectives of the Study 31  3 CHAPTER THREE: OBJECTIVES OF THE STUDY Based on literature review on phosphorus recovery from wastewater through struvite precipitation, it was found that major knowledge gaps still exist in terms of actual crystallizer operation. Although major components in the chemistry related to recovery of phosphate as struvite are well documented, few studies have actually been carried out with respect to the control  of  the  crystallization  process.  The  present  study  gave  emphasis  on  the  efficient operation of a struvite crystallization system as a whole.  This dissertation is based on the hypothesis that “development of a process controller is essential for efficient operation of a phosphorus removal-recovery process by struvite crystallization”.  The reason for building such a controller was to provide a means of automatic control of the crystallization process. This would be accomplished by controlling the supersaturation ratio of the system through manipulating the required pH setpoint.  The efficiency of the struvite crystallization process, as defined in the present study, was based on the following.  x The ability to detect and take action against changes in process variables, such as temperature, pH, conductivity and ionic concentrations; x Possible increase in struvite quality through greater process stability brought about by the use of a process controller; x Increase in ease of operation, due to lower operator time required to maintain the struvite formation process and the crystallizer system; and x Use of the models developed as a means of predicting the influence of various variables and operating conditions that favor struvite formation.   Chapter Three: Objectives of the Study 32  In order to develop a universal supervisory control system and efficient operation of a struvite crystallization process, some related studies had to be carried out. These, supplementary studies, and their objectives, are the following.  1. Solubility Test Objective:  To determine struvite solubility constant using four different water matrices, under varying pH and temperature, and to determine the influence of temperature on the value. Reasoning: Depending  on  the  nature  of  the  experiment  carried  out,  there  still  exist different values in the literature for struvite solubility product. However, the need for a single and universally applicable/acceptable constant is felt in the struvite recovery community.  2. Carbon Dioxide Stripping Tests Objective:  (i) To evaluate the effectiveness of two carbon dioxide strippers in reducing caustic usage, and consequently reducing chemical cost, during the operation of a struvite crystallization process. (ii) To develop a chemistry-based model for the operation of one of the strippers. Reasoning: (i) Chemical cost related to pH increase (by adding caustic) in the operation of  a  struvite  crystallization  process  is  high,  and  methods  to  reduce  it  are needed. (ii) The development of the stripping model is needed to incorporate carbon dioxide stripping in the main supervisory control system developed in the present study, and to determine the extent of carbon dioxide stripping possible and its related pH increase.  3. Magnesium Prediction Techniques Objective:  To develop methods that can provide information quickly, and on-site, on the concentration and rate of application of external magnesium addition in a struvite crystallization process. Reasoning: One of the primary operational costs of a struvite crystallization process arises from the need to add external magnesium to the system. In addition to Chapter Three: Objectives of the Study 33  chemical cost, maximum utilization of the external magnesium added is necessary so that the process effluent does not contain high amounts of the chemical.  4.  Influence of Variables on Structure of Struvite Pellets Objective:  To identify engineering factors that affect struvite pellet growth and its characteristics. Reasoning: The quality of struvite formed in a crystallizer depends on engineering factors, such as upflow velocity, in-reactor magnesium concentration and supersaturation ratio.  5. Determination of Crushing Strength Objective:  To develop a device that can be used as a quick method to determine the crushing strength of struvite pellets.     Reasoning:  Due to lack of a suitable instrument, a device is needed for quick, on-site measurement of pellet strength. The device is needed to determine the relationship(s) between process variables and crushing strength of struvite. Chapter Four: Methods and Methodology 34  4 CHAPTER FOUR: METHODS AND METHODOLOGY 4.1 Struvite Recovery Process The process used in this study for the removal and recovery of phosphorus (as struvite) was developed in the Civil Engineering Department of the University of British Columbia, Vancouver, BC, Canada. Pilot scale testing of the crystallization process was conducted at the Lulu Island Wastewater Treatment Plant (LIWWTP) in Richmond, BC. The setup contains two parallel processes, each of which contains a crystallizer, an external clarifier, tanks for magnesium feed and caustic, pumps for the centrate, recycle flow, magnesium flow and  a  pH controller.  Centrate  from the  plant  was  used  as  a  process  feed  that  was  stored  in tanks to reduce suspended solids contents. The basic flow diagram of the individual process is shown in Figure 4.1.  Centrate is fed at the bottom of the reactor along with the recycle stream. Magnesium chloride and sodium hydroxide is added to the reactor through the injection ports, just above the feed and recycle flows. The pH of the system is monitored by a pH meter at the top of the first section and controlled using a pH controller.  The  reactor  has  four  distinct  zones  of  different  pipe  diameters.  The  bottom  part  of  the fluidized reactor is called the ‘harvest zone’, because this is the section from which struvite is harvested. The second, ‘active zone’, is the location where flows are turbulent enough to allow for agglomeration and formation of larger pellets. The last fluidized section, called the ‘fines zone’, is the location where individual crystals and small pellets reside until they grow large enough to fall to the lower sections. To avoid the loss of fines from the crystallizer, a settling zone, called the ‘seed hopper’, is located at the top. The dimensions of the various sections are given in Table 4.1. The crystallizer, when first developed, was unique in the sense that the diameter of each section was varied, increasing with increasing height. This setup was considered to develop varying degrees of turbulence in each section, and to help in classifying the fluidized particles according to size. As the pellets increase in size (and mass), Chapter Four: Methods and Methodology 35  they are able to overcome higher upflow velocities and, thereby, move down to the lower sections of the reactor. As mentioned, harvesting of struvite pellets is accomplished by emptying the harvest zone (the lowest section), ensuring that only the largest pellets are harvested.  The crystallizer was constructed of clear PVC piping connected with standard Schedule 40 or Schedule 80 PVC fittings.  In the construction of the reactor, the inside joints between piping  and  fittings  were  kept  as  smooth  as  possible,  to  minimize  dead  zones  where  the fluidized particles could settle and struvite encrustation problems could occur.  Clear piping was used, in order to monitor, visually, the behavior of the struvite crystals in the fluidized bed and to watch for signs of plugging or encrustation.                   Figure 4.1. Pilot-scale struvite crystallizer reactor process design. Recycle          MgCl2 NaOH Clarifier Effluent Sludge Stripper pH probe Harvest zone Seed hopper Reaction zone Active zone     Downpipe  Centrate holding tank Chapter Four: Methods and Methodology 36  Table 4.1. Dimensions of the reactor Section Length (mm) Diameter (mm) Volume (dm3) Harvest zone 749 76 3.42 Active zone 1549 102 12.6 Fines zone 1270 152 23.2 Seed hopper 457 381 52.1 Below harvest zone 521 -- --  4.2 Chemicals, Storage Tanks and Pumps 4.2.1 Centrate Centrate used for the study was stored in three 5600 liter capacity holding tanks, which were filled every day from LIWWTP centrate holding tank. These holding tanks served two purposes – (i) help in reducing the total suspended solids by settling them and (ii) control feed composition variations. The centrate was pumped to the reactor through 1.9 cm (¾ inch) tubing, using a ½ HP motor Monyo (Model 500 332) progressive cavity pump that was connected to a digital speed controller. 4.2.2 Magnesium feed The magnesium feed for the study was made from commercial grade magnesium chloride (MgCl2. 6H2O) pellets. The pellets were dissolved in water and stored in a 1600 L tank and pumped into the reactor using a MasterFlex L/S variable speed peristaltic pump with a standard pump head. 4.2.3 pH control Caustic solutions made from sodium hydroxide pellets was stored in two 120 L tanks and used to control the pH of the system. A carbon dioxide trap (concentrated caustic in a bottle) Chapter Four: Methods and Methodology 37  was used to absorb carbon dioxide from the air, before the air entered the caustic tank. This setup was made to ensure minimal loss of caustic strength in the holding tank and which, in turn, could reduce the amount of caustic needed. The pH in the crystallizers and external clarifiers were monitored continuously with Oakton pH monitors, each equipped with an Oakton gel filled, epoxy body pH probe. The pH meters were regularly calibrated by the two point method, using standard buffer solutions of pH 7 and pH 10. Figure 4.2 shows the crystallizer setup of the study area at the Lulu Island Wastewater Treatment Plant. 4.3 Sample Collection, Storage and Analysis Samples were collected daily from each of the clarifiers, one of the centrate holding tanks (because centrate was pumped into the crystallizer from one tank only) and from the magnesium holding tank. Samples were first centrifuged and filtered onsite with 0.45 micron membrane filter papers, before being preserved according to Standard Methods (APHA et al., 1995). For NH4+ and PO4-3,  5 mL samples were taken in small  test  tubes to which was added one drop of 5% v/v sulfuric acid, to lower the pH to below 2. Two drops of concentrated nitric acid were added to each 10 mL of sample to preserve the metal samples. Analytical measurements of phosphorus and ammonium were carried out in the laboratory using colorimetric flow injection analysis (model LaChat QuikChem® 8000). Magnesium and sodium ions were analyzed using flame atomic absorption spectroscopy (AA) (model Varian Inc. SpectrAA220®). Caustic samples were collected for analysis from the individual (one each for Reactors 1 and 2) tanks, and the caustic usage (in terms of volume) by the crystallization processes was recorded daily. By measuring the sodium concentration in the caustic tanks, the amount of caustic (NaOH) used was calculated. Dissolved CO2 samples were collected from the seed hopper and clarifier effluent. The concentration of dissolved gaseous carbon dioxide, in liquid solution, was measured by an Accumet Gas-Sensing Combination ISE. The hardness test methods were followed according to Standard Methods. The pH, temperature and conductivity of the samples were recorded, using a Horiba D54 portable meter onsite.   Chapter Four: Methods and Methodology 38   (a) (b)     (c) (d) Figure 4.2. Setup of crystallizers, clarifiers and stripping columns at the LIWWTP (a) crystallizer (R#1) with compact media stripper attached and without instrumentation and control – the manually controlled system (b) crystallizer (R#2) with cascade stripper attached and with instrumentation and control – the automated system (c) transmitters, pH meter and pH pump for R # 2 system (d) location of conductivity and pH probe in harvest zone of R#2. Chapter Four: Methods and Methodology 39  4.4 Lulu Island Wastewater Treatment Plant The  Lulu  Island  Wastewater  Treatment  Plant  (LIWWTP)  is  a  secondary  wastewater treatment plant that has an average flow of 75 million liters per day. It is operated by Metro Vancouver  of  the  Province  of  British  Columbia,  Canada,  and  primarily  serves  residents  of Richmond, B.C. Biosolids at the treatment plant are digested using mesophilic anaerobic digestion, at 35°C, and thickened with two centrifuges. The centrate is then recycled back to the  headworks  of  the  treatment  plant.  The  biosolids  dewatering  processes  result  in  soluble phosphate being released from the sludge, which then accumulates in the centrate. Due to the buildup of phosphorus, and the presence of ammonia and some magnesium in the liquid centrate, the plant has, in recent times, encountered rapid growth of struvite in the piping system. The current method used to reduce the struvite formation potential at the plant is to dilute the centrate with treated water. However, this only delays the inevitable formation of struvite, as the phosphate and magnesium in the centrate are never fully removed from the system. Figure 4.3 illustrates the trends in some of the important parameters during studies carried out at LIWWTP by the research group. Each point represents a day’s data. The gaps in the graphs represent periods when no study was conducted, and hence no samples were taken. The characteristics of the trends are important to note, because they help determine the struvite  formation  potential  at  any  time  and  the  change  over  the  years.  It  also  provides background information for future studies. As illustrated in Figure 4.3, depending on the time of the year, the temperature of the centrate can vary from 15°C to 35°C. Similarly, high variability in the concentrations of magnesium (zero to 22 mg/L), ammonium (400 mg/L to above 900 mg/L) and phosphate (30 mg/L to 100 mg/L) can be found. Since struvite crystallization efficiency depends on having a stable process in the crystallizer, and these parameters control the process, the variations show the need for developing a control system which will address these variations. Chapter Four: Methods and Methodology 40  7 7.2 7.4 7.6 7.8 8 8.2 8.4 8.6 1-Jan-04 7-Sep-04 15-May-05 20-Jan-06 27-Sep-06 4-Jun-07 9-Feb-08 16-Oct-08 23-Jun-09 pH  (a) 15 20 25 30 35 1-Jan-04 7-Sep-04 15-May- 05 20-Jan-06 27-Sep-06 4-Jun-07 9-Feb-08 16-Oct-08 23-Jun-09 Te m pe ra tu re  (° C )  (b) 5 5.5 6 6.5 7 7.5 8 8.5 1-Jan-04 7-Sep-04 15-May-05 20-Jan-06 27-Sep-06 4-Jun-07 9-Feb-08 16-Oct-08 23-Jun-09 Co nd uc tiv ity  (m S) Date  (c) Figure 4.3. Trends in important parameters in centrate at LIWWTP over the last few years. (a) pH, (b) temperature and (c) conductivity. Chapter Four: Methods and Methodology 41  0 5 10 15 20 25 1-Jan-04 7-Sep-04 15-May- 05 20-Jan-06 27-Sep-06 4-Jun-07 9-Feb-08 16-Oct-08 23-Jun-09 M ag ne si um  (m g/ L)  (d) 30 40 50 60 70 80 90 100 1-Jan-04 7-Sep-04 15-May- 05 20-Jan-06 27-Sep- 06 4-Jun-07 9-Feb-08 16-Oct-08 23-Jun-09 O rth o- Ph os ph at e (m g P/ L)  (e) 400 500 600 700 800 900 1000 1-Jan-04 7-Sep-04 15-May-05 20-Jan-06 27-Sep-06 4-Jun-07 9-Feb-08 16-Oct-08 23-Jun-09 Am m on iu m- N (m g/ L) Date  (f) Figure 4.3 cont’d. Trends in important parameters in centrate at LIWWTP over the last few years. (d) magnesium concentration, (e) ortho-phosphate concentration and (f) ammonium concentration. Chapter Four: Methods and Methodology  42  4.5 Instrumentation and Process Monitoring In order to efficiently apply a control model to a process, a fully functional and complete process  monitoring  system  should  be  in  place.  In  this  study,  the  important  variables  that determined  the  condition  of  the  process  were  monitored  in  the  pilot-scale  setup.  These included the pH, conductivity, temperature, flows, and the ionic concentrations of magnesium, ammonium and phosphate. Figure 4.4 illustrates the basic process components and the locations of sensors and other control variables for monitoring the process. Conductivity, pH and temperature were monitored in-situ continuously, while grab samples were tested for phosphate, magnesium and ammonium concentrations. Conductivity, pH and temperature were measured using sensors that were connected to individual analog transmitters, respectively. Magnesium, ammonium and phosphate concentrations were measured in the laboratory.  Figure 4.4. Schematic of phosphorus recovery process with instrumentation locations at the Lulu Island Wastewater Treatment Plant. Chapter Four: Methods and Methodology  43  4.6 Struvite Control Process The struvite process control model was designed to control the reactor supersaturation ratio by manipulating the pH in the crystallizer. This was achieved by calculating the effects of different chemical reactions that can take place in the reactor. These well-established chemical reactions, along with their equilibrium constants, are given in Table 2.1. The effect of temperature on the equilibrium constants is given in Table 4.2. The equations in Table 4.2 have been derived from Equation 2.6 and the enthalpy values have been taken from Parkhurst and Appelo (1998). Although numerous other chemical reactions have been suggested, the present study included the ones that were most likely to affect the process at the given pH of operation and to simplify the code. The efficiency of the process using the reactions specified has been proven in earlier studies (Fattah et  al. 2008b).  The effect  of temperature on these reactions was also considered in deriving the equilibrium constants. The influences of these reactions, as well as pH, conductivity and temperature, on the system, were taken into account. The model was based on two programs (mentioned in Table 4.3) to control the crystallization process at the desired, optimum condition. The control model was developed using Matlab and Simulink software from MathWorks™. The software was chosen because of  its  wide  use  in  development  of  industrial  process  designs  and  the  ability  to  control complicated chemical reactions, as well as to provide tools for the process. The program also had the ability to smoothly integrate with current industrial software and hardware. A schematic of the control methodology is illustrated in Figure 4.5.  The total concentrations of the reacting species were calculated according to Equations 4.1-4.3. The concentrations were based on a mass balance that included both influent and the recycled effluent, that is, species present in the crystallizer.  T-PO4=[H3PO4]+[H2PO4-]+[HPO4-2]+[PO4-3]+[MgH2PO4+]+[MgPO4-]+[MgHPO4]     (4.1) T-NH4-N = [NH3] + [NH4+]                                                              (4.2) T-Mg = [Mg+2] + [MgOH+] + [MgH2PO4+] + [MgPO4-] + [MgHPO4]                             (4.3) Chapter Four: Methods and Methodology  44   Get data from sensors and analyzers (P,N,Mg, crystallizer pH, temp,cond, pH set point,flows, desired SSR, current SSR ) Calculate SSR (using solveSSR.m) Is current SSR equal to desired SSR Leave pH at current set point Calculate required pH set point (using solvepH.m) Send new pH setpoint to pH meter (using topHcontroller.m) No Yes Save current values (using signalLogger.m) Save current values (using signalLogger.m)  Figure 4.5. Schematic of the control strategy for struvite crystallization process. Chapter Four: Methods and Methodology  45  Table 4.2. Sets of equations used to account for the impact of temperature on the equilibrium constants. The equations are based on Equation 2.6 and values in Table 2.1                     Chapter Four: Methods and Methodology  46  4.7 Coding Controller Program in Matlab In the present study, Matlab was used as the program to write the codes for the SSR controller  (which  determined  the  pH  required  for  set  SSR)  and  the  SSR  calculator  (which calculates the SSR based on the reactor pH). As mentioned in Section 2.14, although different programs have been developed and tried, none have had much success in using their models to operate a struvite crystallizer. The beauty of the codes written in the present study is the simplification of the nonlinear feature of the chemical reactions. In order to reduce the number of variables (from the different nonlinear equations), the codes written for the present study related all phosphate and magnesium-phosphate complexed species to phosphate; this new phosphate equation, together with the nitrogen and magnesium species, was then related to pH only (Section 4.7.1). By setting an initial H+ value,  pH  was  calculated  by  using  the built-in function fzero in Matlab. The activities of the reacting species were used in the determination of SSR, with temperature corrections applied in calculating the equilibrium constants. By substituting and rearranging the various equations, an equation was then developed where SSR was a function of hydrogen ion only. Data collection and processing was performed in Simulink, which used codes written in Matlab to calculate the SSR and required pH. Three graphical user interfaces (GUIs) were also developed for ease of operation and control of the process and to calculate some process parameters. The various codes written are listed in Table 4.3, along with their application(s). 4.7.1 Coding to relate hydrogen ion concentration to SSR (in solvepH.m) The simplified coding used to determine the hydrogen ion activity for a struvite crystallization process is given below. The different symbols and abbreviations used are defined in Appendix A.       Chapter Four: Methods and Methodology  47  Hfunction =  @(H)... (function of H)               ...% C1:               a/(b + c/H + (d + e*H + f*H^2) * ...                 ( -( (n + o*H + p*H^2)*a + (b + c/H)*(j + k*H^3 + l*H^2 + m*H) - (d + e*H + f*H^2)*i ) + ...                     sqrt( ( (n + o*H + p*H^2)*a + (b + c/H)*(j + k*H^3 + l*H^2 + m*H) - (d + e*H + f*H^2)*i )^2 - ...                         4*( (j + k*H^3 + l*H^2 + m*H) * (d + e*H + f*H^2) )*( -i * (b + c/H) ) ) ) ...                 / (2* ( (j + k*H^3 + l*H^2 + m*H) * (d + e*H + f*H^2) ) ) )...               ...% C2:               * g / (1 + h/H ) ...               ...% C3:               * ( -( (n + o*H + p*H^2)*a + (b + c/H)*(j + k*H^3 + l*H^2 + m*H) - (d + e*H + f*H^2)*i ) + ...                     sqrt( ( (n + o*H + p*H^2)*a + (b + c/H)*(j + k*H^3 + l*H^2 + m*H) - (d + e*H + f*H^2)*i )^2 - ...                         4*( (j + k*H^3 + l*H^2 + m*H) * (d + e*H + f*H^2) )*( -i * (b + c/H) ) ) ) ...                 / (2* ( (j + k*H^3 + l*H^2 + m*H) * (d + e*H + f*H^2) ) ) ...               ...% SSR * KMAP:               - SSR * KMAP;           Chapter Four: Methods and Methodology  48  Table 4.3. List of files written and developed in Matlab with their uses Code Number Program code name and extension Application Link to other codes 1 fromHEX.m Converts HEX code (from pH meter) to regular decimal number Used by code 3 2 toHEX.m Converts regular decimal number to HEX code (for pH meter) Used by code 4 3 frompHcontroller.m Used to acquire data from the pH meter 4 topHcontroller.m Used to set required pH in pH meter 5 solvepH_block.m Set block in Simulink environment for using Matlab code Used by code 6 6 solvepH.m Used  to  calculate  the  desired  pH  set  point for particular SSR  7 solveSSR_block.m Set block in Simulink environment for using Matlab code Used by code 8 8 solveSSR.m Calculates the SSR for reactor conditions Used by code 10,11 13 9 strippermodel.m Calculates efficiency of carbon dioxide stripping and potential pH increase 12 10 SSRcalculator.m Calculates the SSR on a system 11 SSRcalculator.fig Code for developing SSR calculator GUI 12 strippermodelgui.fig Code for Carbon dioxide stripping GUI. 13 reactoroperation.fig Code for developing reactor operation GUI 14 signalLogger_block.m For logging the signals of the process 15 15 signalLogger.m Logs process data from listed signals into Excel file  16 System.mdl Simulink window for operation of a struvite crystallization process  Chapter Four: Methods and Methodology  49  4.8 Characteristics of Process Feed during Pilot Scale Studies with Instrumentation During the course of the study, some major changes occurred in the characteristics of the centrate that was used as process feed. It should be noted that the process feeds for the two crystallizers were obtained from the same centrate tank. Some of the characteristics of the centrate are shown in Table 4.4 and illustrated in Figure 4.6.  In general, the high variability in phosphate and ammonium concentrations illustrated in Figure 4.6 is not common. At the initial part of the study, from the start to the end of July, 2009, it was found that centrate from the plant was being diluted by processed water to reduce the potential for struvite formation; this in turn, reduced the concentrations of the ions.  In  general,  phosphate  and  ammonium  concentrations  at  the  plant  varied  from  60-120 mg/L and 700-100 mg/L, respectively (Fattah et al. 2008b). Table 4.5 provides the values of the operating parameters used for the study. Since the objective of this study was to develop and validate a control system, operating conditions were based on a previous study undertaken at the same location (Fattah et al. 2008b).  Table 4.4. Characteristics of the process feed used during pilot-scale operation (2009)  pH Conductivity mS/cm Temperature C PO4-P mgP/L NH4-N mg/L Mg mg/L N:P Minimum 7.1 2.2 23.1 9.2 63.5 2.3 14.7 Maximum 8.4 7.3 38.9 101.0 920.0 22.0 25.2 Average 7.4 4.7 28.5 57.7 507.5 9.5 19.8   Chapter Four: Methods and Methodology  50  0 20 40 60 80 100 120 26-Jun 6-Jul 16-Jul 26-Jul 5-Aug 15-Aug 25-Aug 4-Sep Ph op sh at e (m g P/ L) Date (a)  0 100 200 300 400 500 600 700 800 900 1000 26-Jun 6-Jul 16-Jul 26-Jul 5-Aug 15-Aug 25-Aug 4-Sep A m m on iu m  (m g N /L ) Date (b) Figure 4.6. Centrate characteristics at the Lulu Island Wastewater Treatment Plant during pilot-scale validation of process control in 2009 (a) phosphate, (b) ammonium.  Chapter Four: Methods and Methodology  51  0 5 10 15 20 25 26-Jun 6-Jul 16-Jul 26-Jul 5-Aug 15-Aug 25-Aug 4-Sep M ag ne si um  (m g /L ) Date (c)  12 14 16 18 20 22 24 26 26-Jun 6-Jul 16-Jul 26-Jul 5-Aug 15-Aug 25-Aug 4-Sep N :P  m ol ar  ra tio Date  (d) Figure 4.6 cont’d. Centrate characteristics at the Lulu Island Wastewater Treatment Plant during pilot-scale validation of process control in 2009 (c) magnesium and (d) N:P molar ratio.      Chapter Four: Methods and Methodology  52   Table 4.5. Operating conditions in each struvite crystallizer Parameter Unit Val1ue Centrate flow L/min 2.51 Recycle ratio (RR) - 6 Recycle flow L/min 15.63 Mg:P ratio - 1.3 Total flow in crystallizer L/min 18.24 Magnesium feed concentration mg/L 2000 pH - Depends on feed concentration (7.8-8.2) Caustic feed concentration N 0.3 Upflow velocity (in harvest zone) mm/sec 67 Number of baffles in stripper - 10 SSR in crystallizer - 3  4.9 Experimental Setup The present study encompassed experiments carried out in both laboratory- and pilot- scale setups over several months. It was found that centrate characteristics were different during different experimental runs, and therefore, the characteristics have been placed in different sections. In addition, several different water matrices were used during the course of the study in the laboratory. The following section has been sub-divided according to the type of experiment performed. Chapter Four: Methods and Methodology  53  4.9.1 Purpose of the study – solubility tests The water matrix for typical struvite recovery installations range from advanced wastewater treatment plant (AWWTP) supernatant to centrate, making it necessary to eliminate as much of the chemical variability from the operating control loop as possible. Experiments conducted in the present study aimed at determining a working value for Ksp, which reflects the operating conditions (in terms of pH and temperature) of the various field installations. These experiments were designed to establish equilibrium between struvite crystals in the solid phase and their dissolved component ions in an aqueous system. Ideally, this would be done under a closed system to prevent the absorption and escape of gases, such as carbon dioxide and ammonia; however, due to experimental constraints, tests were conducted under open conditions, as the top of the beakers were not covered. Digester supernatant and centrate from local treatment plants, and artificial solutions formed from distilled and tap water, were used as water matrix. Multiple matrices were used in order to quantify appreciable differences, if any, in Ksp (due to inherent chemical composition). 4.9.2 Experimental setup and methodology for solubility tests For all the solubility tests, a six-station paddle stirrer (Phipps and Bird) was used with square jars in a temperature-controlled room (Figure 4.7). About 10 g of struvite (previously grown  in  the  water  matrix  used)  was  added  to  1.5  L  of  the  water  matrix  (either  distilled water, tap water, centrate or digester supernatant) in each jar. The idea was to insert enough struvite in the jars to attain saturation. The paddle stirrers were set to operate at 70 ± 2 RPM. Samples  of  20  L  of  each  water  matrix  were  brought  to  the  testing  temperature  by  storing them in the controlled temperature room before the experiment. Since high levels of ammonia and phosphate were already present in the wastewater, only magnesium chloride was added, in order to supersaturate the solutions with respect to each of the component ions. In the distilled and tap water systems, additional diammonium hydrogen phosphate {(NH4)2HPO4} and magnesium chloride (MgCl2)  were  added  to  achieve  the  same concentration levels as in the centrate and digester supernatant water matrices. The rationale for using struvite previously grown in a given matrix was to maintain the chemical uniqueness of a water matrix, as much as possible. The final concentration ranges that were Chapter Four: Methods and Methodology  54  being targeted for Mg+2, NH4-N and PO4-P were 180-200 mg/L, 150-180 mg/L and >300mg/L respectively. The prepared solutions were kept at a pH of 4.0, to discourage struvite formation.  Since Ksp is highly temperature dependant, trials were conducted at 10, 15 and 20ºC (chosen to reflect seasonal operational temperatures in Vancouver, BC). As Ksp is also pH dependant, tests for each of the trial temperatures were conducted for multiple pH values, starting from 6.0 to 9.0 in 0.2 - 0.3 pH increments. In order to keep the pH at the desired value, dilute hydrochloric acid or sodium hydroxide solutions were added to each of the jars. Previous studies (Fattah, 2004) showed that equilibrium could be achieved after 24 hours; however, in order to provide a further degree of safety, sampling was undertaken when the pH had remained constant for 3 hours after assumption of equilibrium (after 24 hours). Once equilibrium was considered to be established, the pH and conductivity in each jar were measured. Samples were taken from each jar to determine magnesium, ammonium and ortho- phosphate concentrations.    Figure 4.7. Setup of solubility determination. Chapter Four: Methods and Methodology  55  4.9.2.1 Speciation Model The model used to calculate the supersaturation ratio was an in-house program coded in Microsoft Excel, entitled SimpleMAP v1.0, to perform the chemical speciation of the tested solutions. The chemical equilibrium reactions (Table 2.1) and values (corrected for temperature) were used in the program. The ionic activity coefficient was determined using the empirical formula presented in Chapter Two, which relates the coefficient with conductivity measurements.  The pKsp was determined using the speciation model, by entering the measured values for each of the three species of interest (ortho phosphate, magnesium and ammonium), as well as the measured pH, conductivity and temperature. An iterative analysis was then conducted in order to calculate the value of pKsp that would set the SSR value to 1.0. The assumption made here is that each batch test reached equilibrium. Details on the model can be found in Fattah (2004). 4.9.3 Purpose of the study - carbon dioxide stripping As mentioned earlier, carbon dioxide stripping may reduce the amount of caustic needed to increase the pH in a struvite crystallization process. Experiments were carried out at Metro Vancouver’s Lulu Island Wastewater Treatment Plant in Richmond, B.C. Canada, using centrate generated at the plant. In order to determine the effects of stripper configurations, two types of strippers were used in the present study – a compact media stripper and a cascade stripper.  The strippers were located between the overflow pipe of the crystallizer and the clarifier. Some of the parameters used in this section were sampled, measured and evaluated by Sabrina (2007). 4.9.4 Experimental setup for carbon dioxide stripping 4.9.4.1 Compact media stripper One of the two strippers used in this study was a compacted media type developed by Ostara Nutrient Recovery Technologies Inc., Canada, and connected to R#1 (Figure 4.2a). It consisted of a circular stripping tower, a supporting plate for the packing material near the bottom, a liquid distributor system located above the packing material, and a fan at the bottom of the stripping tower (Figure 4.8a and b). The total height of the stripping tower was Chapter Four: Methods and Methodology  56  1.8 m, 1.2 m of which were packed with 25.4 mm diameter hollow plastic balls, as packing material. The balls were placed to provide a larger specific surface area and allow the flowing liquid stream adequate water/air contact time to increase the rate of carbon dioxide stripping. Initially the balls (Figure 4.8c) were suspended between plates, but due to clogging of the surface (detailed in Chapter Five), the setup was changed, whereby the balls were attached to strings (Figure 4.8d). The stripper was placed directly on top of the external clarifier. 4.9.4.2 Cascade Stripper The second type of stripper used in this project was a cascade stripper (Figure 4.9) that was designed at UBC. Details of the design can be found in Zhang (2006). Essentially, this was a rectangular vessel with baffles placed at an inclination of 10°. This stripper was connected to R#2 (Figure 4.2b). The cascade stripper was designed for a maximum hydraulic loading  rate  of  20  L/min,  and  was  incorporated  into  the  crystallizer  system  just  before  the external clarifier, substituting almost 1/3 of the reactor downpipe. The top of the stripper was sealed by placing a plate that had a connection for external air pump tubing. The bottom of the stripper was placed 30 cm above the external clarifier.  Chapter Four: Methods and Methodology  57    (a) (b)   (c) (d) Figure 4.8. (a) Compact media stripper schematic, (b) stripper at setup at LIWWTP, (c) hollow plastic balls used and (d) final arrangement of balls in the stripper. Fan mount Clarifier Feed into stripper Chapter Four: Methods and Methodology  58     (a) (b)  (c) Figure 4.9. (a) Dimensions of the cascade stripper, (b) stripper at setup at LIWWTP and (c) dimensions of individual baffles. Clarifier Air Chapter Four: Methods and Methodology  59  4.9.5 Methodology to determine efficiency of the carbon dioxide strippers During the course of the present study, four experiments (Table 4.6) were conducted – one without strippers, and the rest with strippers under different operating conditions. The first experiment was carried out to determine if the two reactors used in this study were similar, regarding overall performance, so that a direct comparison could be made between the two. The results showed that they were very similar and hence, all performances of the two strippers were compared directly. In the first run (Run1) the struvite crystallizers were operated without strippers. In subsequent runs, the strippers operated with air (Run 2) and without air (Run 3), to determine the influence of air on CO2 stripping. In the last test run (Run 4), the recycle ratio (defined in Section 4.11.1) and upflow velocity (defined in Section 4.11.2) were set to 9 and 75 mm/sec, respectively and the external air supply was resumed.  Air was pumped into the cascade stripper through an airflow meter from the top of the stripper; the maximum airflow rate was 107.5 L/min. The air into the compact media stripper could not be controlled, since it was provided by a fan having fixed speed. While running without external air  supply,  the top surface of the cascade stripper was kept open. With the introduction of the airflow, the top surface of the stripper was covered with a plexiglass lid. The total flow of water matrix, which is the recycle flow from the crystallizer, through the strippers was 18.24 L/min. The characteristics of the centrate that was used for the process is given in Table 4.7 and the variation illustrated in Figure 4.10. Each point in the figure represents concentration based on daily grab samples. Table 4.8 summarizes the operating conditions for the crystallizers; these values were based on previous studies carried out at the site. As expected, the magnesium concentration in the centrate was the limiting factor for struvite  crystallization.  The  molar  ratio  of  Mg:P  was  always  below  1,  the  minimum  ratio required for struvite precipitation. Hence, magnesium feed (in the form of MgCl2) was injected into the reactors, to raise the Mg:P ratio. The average supersaturation ratio (SSR) of centrate during the study period was 0.96.    Chapter Four: Methods and Methodology  60  0 20 40 60 80 100 120 15-Aug 4-Sep 24-Sep 14-Oct 3-Nov 23-Nov 13-Dec Ph os ph at e (m g/ L) Date  (a) 0 200 400 600 800 1000 15-Aug 4-Sep 24-Sep 14-Oct 3-Nov 23-Nov 13-Dec A m m on iu m   ( m g/ L) Date  (b) 0 5 10 15 20 15-Aug 4-Sep 24-Sep 14-Oct 3-Nov 23-Nov 13-Dec M ag ne siu m  (m g/ L) Date  (c) Figure 4.10. Centrate characteristics during experiments testing the efficiency of the carbon dioxide stripper (a) phosphate, (b) ammonium and (c) magnesium.  Chapter Four: Methods and Methodology  61  Table 4.6. Test conditions in the crystallizers and strippers for the comparison of strippers  Stripper Air Recycle Ratioa Upflow velocityb (mm/sec) Run #1 × × 6 67 Run #2 ¥ ¥ 6 67 Run #3 ¥ × 6 67 Run #4 ¥ ¥ 9 75 a Recycle Ratio = recycle flow/feed flow b Upflow velocity = flow velocity in the harvest zone of the crystallizer Table 4.7. Characteristics of centrate used during carbon dioxide stripping tests  pH Temp  Cond  Mg  PO4-P NH3-N Molar Ratio    (oC) (mS/cm) (mg/L) (mg/L) (mg/L) Mg:P N:P Minimum 7.2 15.3 4.11 4.1 42.6 500 0.01 18 Maximum 8.1 34 12.48 17.4 100.0 916 0.39 36 Average 7.6 25 7.01 9.8 77.7 782 0.17 23  Table 4.8. Summary of operating conditions in the crystallizers  Total feed Centrate flow Mg feed flow Recycle ratio Recycle flow Total flow pH Upflow velocity Unit (L/min) (L/min) (mL/min)  (L/min) (L/min)  (mm/sec) Value 2.61  2.51  100  6 15.63  18.24  8.1    67 For cascade stripper the air flow rate was 107.5 L/min; the air in the compact media stripper could not be controlled 4.9.6 Purpose of the study - Determination of magnesium usage for struvite precipitation Common methods of determining the concentration of magnesium in a water matrix involves the use of expensive and time consuming methods, such as use of atomic absorption or inductively coupled plasma. In a continuous process, such as in a struvite crystallizer, Chapter Four: Methods and Methodology  62  methods that provide results quickly are needed. In the present study, two methods that could potentially be used to provide information on magnesium concentration and addition requirements were tested. The first method involved the use of pH and conductivity, while the second one involved the use of hardness test methods. 4.9.7 Experimental setup and methodology for the determination of magnesium usage for struvite precipitation 4.9.7.1 Use of pH and conductivity Several experiments, using synthetic water and centrate from Lulu Island wastewater treatment  plant  (LIWWTP),  were  carried  out  to  determine  the  applicability  of  using conductivity and pH for magnesium requirements during struvite precipitation. The synthetic water was prepared by adding ammonium dihydrogen phosphate (NH4H2PO4), ammonium chloride (NH4Cl) and magnesium chloride (MgCl2.6H2O), for a final solution having phosphate, ammonium and magnesium concentrations of 300 ppm, 180 ppm and 180 ppm, respectively. The phosphate concentration was deliberately set high so that it would not be a limiting ion during the precipitation process. A magnesium standard of 3000 ppm was prepared in water using MgCl2.  All experiments were carried out in well-mixed 1L Nalgene beakers using 500 mL of water sample. The prepared magnesium chloride standard was added to the solution in 0.5 mL increments every minute; conductivity and pH were logged one minute after chloride addition. Conductivity was measured using a Hanna Instruments HI9033 multi range conductivity meter, while pH was measured with an Oakton® pH meter.  Chemical coagulants, such as ferric chloride, ferric sulfate and alum, and polymers (such as poly aluminum chloride), are commonly used to reduce phosphate concentrations in wastewater. The same principle was used for the present study to reduce phosphate concentrations, so that its presence did not interfere with hardness tests. Phosphate concentrations of around 20 mg/L and higher can interfere with hardness tests (APHA et al., 1995). Therefore, exploratory tests were carried out with two chemicals – polyaluminum Chapter Four: Methods and Methodology  63  chloride (PAC) and alum – to determine the dosages required to reduce phosphate concentrations to values that would not interfere with the hardness tests.  For the second set of experiments, phosphate interference was removed by the following method: a 200 mL centrate sample was taken in a beaker and mixed with PAC or alum in a jar tester for 1.5 hours. After settling the mixture for an hour, the sample was centrifuged at 8000 relative centrifugal force (RCF) for 15 minutes. The supernatant was then used for the hardness tests. The standard alum was prepared by dissolving 5 g of aluminum sulphate (Al2(SO4)3.  16  H2O)  in  500  mL  of  distilled  water,  producing  a  concentration  of approximately 855 mg/L of Al. The standard PAC concentration was 2 × 10-5 % (v/v) of 30% PAC as Al2O3. 4.9.7.2 Use of hardness tests Hardness tests were carried out according to Standard Methods (APHA et al., 1995) for both calcium (Method 3500-Ca D) and total hardness (Method 2340 C). The magnesium concentration was calculated based on the difference between total and calcium hardness, assuming that the hardness was derived from only Ca+2 and  Mg+2 ions. Due to low concentrations usually found in centrate (Fattah et al., 2008b), the presence of interfering cations, such as cadmium, zinc, iron, nickel and cobalt, were ignored. The interference due to aluminum was reduced by adding a magnesium salt of 1, 2-cyclohexanediaminetetraacetic acid (MgCDTA). 4.10 Product Quality Determination Struvite produced from the pilot-scale operation was tested for different characteristics. With  the  intention  of  harvesting  only  relatively  large  size  (3-5  mm  and  above)  pellets, harvesting was carried out when there was sufficient quantity in the reactor, or when necessary (e.g. to clean the injection port). The harvested crystals were collected and dried in air on trays.   Chapter Four: Methods and Methodology  64   The quality of the harvested pellets was determined by analyzing their composition, morphology and crushing strength. In order to determine the composition and purity of the pellets, approximately 0.2 g of struvite pellets was dissolved in 50 mL of distilled water and 10 drops of concentrated nitric acid. In order to accelerate dissolution, the samples were mixed in a mechanical shaker for 24 hours, after which samples were analyzed for magnesium, ammonium, ortho-phosphate, calcium, aluminum and iron. Inductively coupled plasma (ICP)/mass spectrophotometer (MS) analysis of harvested struvite was also carried out  on  some  samples.  The  morphology  of  the  harvested  crystals  was  examined,  using  a scanning electron microscope (SEM) (model Hitachi S-3000N). As described in the next section, the crushing strength was determined using a newly developed device. 4.10.1 Development of strength tester The absence of a strength tester device in the laboratory made it necessary to construct a new “gadget” that could be used to quantify crushing strength of struvite, which, until now, was described only qualitatively – ability/unability to crush a pellet between fingers (Fattah et al. 2008b). The device made for this struvite study consisted of a load cell at the tip of a handle (Figure 4.11) that was connected to a computer via a data acquisition box (PMD-1208 LS, Measurement ComputingTM). A program was developed in DASYLab®9.0 (Measurement ComputingTM)  to  collect  continuous  force  data  during  each  particle  test (Figure 4.12). The peak load required to crush a particle was recorded first in the program and then saved as an ASC file which was accessible from Microsoft Excel for data processing and developing graphics. This device allowed for quick and easy measurement of crushing strength of pellets. On average, over 100 pellets were randomly picked and crushed for each data set/struvite harvest.  Chapter Four: Methods and Methodology  65  Sp rin g  (a) (b)  Figure 4.11. Device used to determine crushing strength of struvite (a) sketch (not to scale) and (b) actual device.  Chapter Four: Methods and Methodology  66   Figure 4.12. Program window used to graphically illustrate force as a function of time and to determine peak load. 4.11 Terminology Some uncommon terms used in this study are described in this section. 4.11.1 Recycle ratio Effluent from the crystallizer was collected in the clarifier and then recycled back into the process. The recycle ratio represents the ratio of the flow from the clarifier (by using the recycle pump) to the combined flow from the centrate and chemical pumps.  This recycling had two purposes - to dilute the centrate so that the operating pH was around 7.5-8.0, and to provide adequate total water matrix flow into the crystallizer. The recycle ratio was calculated according to Equation 4.4.  Recycle Ratio = Qr/Qf = (Qt-Qf)/Qf    (4.4) where, Qr is the recycle flow Qt is the total combined flow into the crystallizer (feed + recycle), (mm3/sec or L/min). Chapter Four: Methods and Methodology  67  Qf is the feed flow (magnesium and centrate flow) (mm3/sec or L/min) 4.11.2 Upflow velocity The upflow velocity (Equation 4.5), as used in the present study, is the velocity of the liquid in the harvest (lowest section of the crystallizer) zone. This parameter was used to calculate the feed and recycle flows into the system.  Upflow velocity (mm/sec) = Qt/AH                                     (4.5)  where, AH is the area of the harvest zone (mm2) 4.11.3 Removal efficiency Two of the primary parameters that defined the efficiency of the system were the removal efficiencies of phosphate and ammonium from the system. They are given by Equations 4.6 and 4.7. %P removal = {[Pinfluent] - [Peffluent]}/ [Pinfluent] × 100  (4.6) %N removal = {[Ninfluent]- [Neffluent]}/ [Ninfluent] × 100 (4.7) [Pinfluent], [Ninfluent] = concentrations of centrate ortho-phosphate and ammonium, respectively,  at  the  inlet  (mg/L)  (i.e.  multiplying  the  respective  concentrations  with  the centrate flow rate and dividing by the total feed flow) [Peffluent], [Neffluent] = concentration of ortho-phosphate and ammonium, respectively, in the reactor effluent collected from the external clarifier (mg/L) 4.11.4 Confidence limit/Error The 95% confidence limit has been calculated for some data. The relationship between standard deviation and the error, or confidence limit, is given by Equation 4.8.  95 % Confidence Limit = 1.96 * SD/¥n                                                                    (4.8)  where, Chapter Four: Methods and Methodology  68  SD = standard deviation n = number of samples 4.11.5 Root mean square The root mean square (RMS) value was used to determine the efficiency of the automated controller developed in the present study in keeping the actual process pH similar to the desired process pH. For a set of n values { }, the RMS value is given by Equation 4.9.                                                                      (4.9)  Chapter Five: Results and Discussions  69  5 CHAPTER FIVE: RESULTS AND DISCUSSIONS 5.1 Testing of Control Program Once the control program was developed, it was tested in the laboratory before being used at the pilot-scale process at LIWWTP. 5.1.1 Laboratory simulations The SSR controller program was initially evaluated in the laboratory by simulating changes in the variables, or process parameters, and determining their influences on the controlled variable SSR. The SSR was calculated based on actual measurements of pH, conductivity  and  temperature.  The  program  was  then  tested  at  pilot-scale,  at  the  treatment plant. The laboratory simulation was carried out by changing one variable at a time and then allowing the control program to adjust the required pH for a set SSR. The pH of tap water in a 500 ml beaker was reduced by adding dilute hydrochloric acid to mimic struvite formation, and then adjusted by adding caustic from a pump, simulating the crystallization process. Although all factors involved with the struvite crystallization can bring about a change in the SSR (and consequently the pH required to keep the SSR steady) only two variables were tested, since it can be concluded that the program would work the same with other variable changes. Figure 5.1 shows the effects of phosphate and temperature changes on the required pH, and the pH of the system as controlled by the pump. Phosphate was chosen because it is the parameter that needs to be optimized with respect to removal-recovery, and temperature was found to have the greatest influence on SSR. As illustrated in the figure, the controller was able to detect the change in the variable (phosphate or temperature) and calculate the required pH. Once the pH was calculated,  the appropriate signal was sent to the pH pump, which added caustic to the water matrix to reach the desired pH.    Chapter Five: Results and Discussions  70  8.05 8.1 8.15 8.2 8.25 8.3 8.35 8.4 8.45 8.5 0 10 20 30 40 50 60 70 80 90 0 200 400 600 800 pH Ph os ph at e ( m g P/ L) Sample time (s) Phosphate Concentration Actual pH Required pH  (a) 0 5 10 15 20 25 30 35 40 500 700 900 1100 1300 1500 Sample time (s) Te m pe ra tu re  (° C ) 8.15 8.2 8.25 8.3 8.35 8.4 8.45 8.5 pH Temperature Actual pH Required pH  (b) Figure 5.1. Effectiveness of the SSR controller in calculating and maintaining required pH due to (a) changing phosphate concentration, (b) changing temperature. Chapter Five: Results and Discussions  71  5.1.2 Experimental runs Once the SSR controller was successfully tested at the laboratory-scale through simulations-batch tests, the system was applied to the operation of a crystallizer (R#2) at LIWWTP – the automated crystallizer system. This system (R#2) differed from the R#1 system,  which  was  manually  controlled.   Due  to  economic  constraints,  not  all  the  required real-time data could be obtained, and hence, the full applicability of the model could not be realized. It is expected that in the future, when this model is applied in a full-scale installation, with adequate instrumentation, the process can be operated at a higher efficiency. Despite the need for more real-time data, the model was still able to operate the crystallizer with a high degree of efficiency, by keeping the SSR, and the pH, at the desired levels. Figure 5.2a illustrates the variation in the measured conductivity and temperature in the crystallizer, while Figure 5.2b illustrates the set point and the actual measured pH in the crystallizer. Other than the conductivity, temperature and pH of the crystallizer system, all process variables were kept constant during this period. Figure 5.2b also shows the absolute error in pH. The sudden jump in the set point pH was attributed to the sudden increase in conductivity, as illustrated in Figure 5.2a. The regular variations in the desired set point pH could be linked to the combined influences of temperature and conductivity on the crystallizer  SSR.  It  must  be  remembered  that  the  pH  values  measured  in  the  process  were sometimes limited by the accuracy of the pH meter and its transmitter, so an absolute error less than 0.03 (RMS value of 0.098) seemed satisfactory. The crystallizer was operated continuously for few months using the control program, with great success. Struvite harvested from the system was of better quality (as mentioned in later sections) compared with that formed in previous studies, at the same location. Chapter Five: Results and Discussions  72   22.5 23.0 23.5 24.0 24.5 25.0 25.5 26.0 26.5 27.0 27.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 0 5 10 15 20 25 30 35 40 Te m pe ra tu re  ( C ) C on du ct iv ity  (m S/ cm ) Time (hour) Conductivity Temperature 0.00 0.03 0.05 0.08 0.10 8.0 8.1 8.2 8.3 0 5 10 15 20 25 30 35 40 A bs ol ut e er ro r i n pH pH Time (hour) pH error Set point pH Measured pH (b) Figure 5.2. (a) Variation of conductivity and temperature in the crystallizer, (b) variation between set point pH (for particular SSR) and measured pH in a pilot-scale struvite crystallizer controlled by the SSR controller. Chapter Five: Results and Discussions  73  5.2 Application of Model to Predict Different Scenarios for Control of Struvite Formation Potential Although the primary use of the developed model was in operating a struvite crystallization process, it could also be used as a predictive tool. Many studies (Fattah et al. 2008b; Forrest et al. 2008a; Bhuiyan et  al. 2007; Ohlinger et al. 1999) have reported the influence of chemical species’ concentration on the solubility of struvite, system supersaturation and struvite crystallization feasibility. However, few have reported the effect of temperature and conductivity on supersaturation ratio. Among the controlling conditions for struvite product quality, the influence of the Mg:P molar ratio has been suggested (Hanhoun et al., 2009; Wang et al. 2006).  Therefore, by using the SSRcalculator.m model (Table 4.3), the sensitivity of SSR to temperature was tested with varying Mg:P molar ratios. The results are illustrated in Figure 5.3. What the figure illustrates is the strong inter-dependence of the variables, and that changing one parameter alone (while others are low) does not necessarily increase the SSR, or the struvite formation potential. This figure can be used by practitioners to determine their system’s struvite formation potential and to assess what particular change(s) could lead to lowering the SSR value. Chapter Five: Results and Discussions  74  0 1 2 3 4 5 6 7 8 10 15 20 25 30 35 Temperature (°C) SS R Mg:P = 0.064; Mg 5 mg/L, PO4 = 100 mg P/L Mg:P = 0.08; Mg 5 mg/L, PO4 = 80 mg P/L Mg:P = 0.106; Mg 5 mg/L, PO4 = 60 mg P/L Mg:P = 0.127; Mg 10 mg/L, PO4 = 100 mg P/L Mg:P = 0.159; Mg 10 mg/L, PO4 = 80 mg P/L Mg:P = 0.212; Mg 10 mg/L, PO4 = 60 mg P/L Mg:P = 0.191; Mg 15 mg/L, PO4 = 100 mg P/L Mg:P = 0..239; Mg 15 mg/L, PO4 = 80 mg P/L Mg:P = 0.319; Mg 15 mg/L, PO4 = 60 mg P/L  Figure 5.3. Influence of temperature and Mg:P molar ratios on calculated system SSR.  Both Figure 5.4 and Figure 5.5 illustrate the strong influence of temperature on SSR. The SSR values shown in Figure 5.4 and Figure 5.5 were calculated based on the variables values shown in the figures. It is seen that there was more than a 90% increase in SSR due to a 5°C drop in temperature. The variation was more pronounced at lower temperatures, where the SSR nearly doubled (between 15 and 20°C).  The variable values used for this simulation are representative of those found at the LIWWTP. Since phosphate was not the limiting ionic species for struvite growth at the treatment plant, its influence (Figure 5.4) on SSR was less than that of the more limiting Mg ion (Figure 5.5). As illustrated in Figure 5.5, irrespective of the temperature, at the lower concentrations (between 5-10 mg/L Mg), the SSR nearly doubled (a 200% increase), whereas the increase in SSR at higher concentrations was only 100%. Thus, treatment plants that have low magnesium concentrations may not have Chapter Five: Results and Discussions  75  encountered struvite in there system as yet, but if other conditions are present, a small increase in magnesium concentration will increase struvite formation many folds. 0.67 1.67 0.85 0.44 0.23 2.15 1.09 0.56 0.29 2.60 1.31 0.35 0.0 0.5 1.0 1.5 2.0 2.5 3.0 10 15 20 25 30 35 Temperature (°C) SS R PO4 = 60 mg P/L PO4 = 80 mg P/L PO4 = 100 mg P/L NH4-N = 800 mg/L; Mg = 5 mg/L  Figure 5.4. Influence of temperature on supersaturation ratio at different phosphate concentrations. 1.67 0.85 0.44 0.23 3.31 1.67 0.86 0.46 4.91 2.48 1.28 0.67 0 1 2 3 4 5 6 10 15 20 25 30 35 Temperature (°C) SS R Mg =5 mg/L Mg = 10 mg/L Mg =15 mg/L NH4-N = 800 mg/L; PO4-P = 60 mg/L  Figure 5.5. Influence of temperature on supersaturation ratio at different magnesium concentrations. Chapter Five: Results and Discussions  76  5.3 Application of Model for Decision Making The decrease in SSR, due to an increase in temperature, was shown above (Figure 5.4 and Figure 5.5). This is probably why treatment plants having high temperature anaerobic digestion are less likely to see struvite forming in the digester, during early years of operation. However, as mentioned earlier, in later years, the concentrations of phosphate and magnesium can increase due to recirculation within the treatment plant, and these may increase struvite formation potential. Since the temperature is not the only driving force, substantial increases in the other controlling parameters can easily drive the SSR above unity, and provide conditions for rapid struvite formation, even at high temperatures. This fact is illustrated in Figure 5.6 where SSR was above 1 at 30°C. This 3-D graphical representation provides a more visual summary of the relationship between the different parameters, and can help practitioners make decisions regarding their operating schemes. Chapter Five: Results and Discussions  77    (a)  (b) Figure 5.6. (a) Influence of magnesium concentration and temperature on SSR and (b) influence of Mg:P molar ratio and temperature; all other parameters constant. SS R  SS R  Chapter Five: Results and Discussions  78  Another application of the model could be to perform various simulations, as well as to see the combined effects of all the parameters. The simulated parameter values are provided in Appendix B and the effects of the parameter change on SSR is illustrated in Figure 5.7. The  parameter  values  used  for  the  simulation  are  based  on  ranges  found  at  LIWWTP.  As struvite formation is heavily dependent on concentrations of the constituent ions, temperature and conductivity, it is reasonable to use these parameters simultaneously, under different scenarios, to determine their combined effects on the supersaturation ratio in the system (shown by the last graph in Figure 5.7). As previously shown with the LIWWTP simulation, the system temperature and limiting magnesium ion concentration had the most influence on the SSR. The pH of the simulation in Figure 5.7 was kept constant at 7.7 throughout the simulation. The sample number represents the number of data collected by the data acquisition toolbox in Simulink software; 180 data points were collected each second. Chapter Five: Results and Discussions  79   0 5 10 15 20 25 30 35 0 10000 20000 30000 40000 50000 Te m pe ra tu re  (C ) Sample number  0 5 10 15 20 25 30 35 0 10000 20000 30000 40000 50000 M ag ne siu m   ( m g/ L) Sample number  0 50 100 150 200 0 10000 20000 30000 40000 50000 Ph os ph at e  (m g/ L) Sample number  0 200 400 600 800 1000 1200 1400 0 10000 20000 30000 40000 50000A m m on iu m -N   ( m g/ L) Sample number  0 1000 2000 3000 4000 5000 0 20000 40000 60000 Co nd uc tiv ity  (m S/ cm ) Sample number  0 2 4 6 8 10 0 20000 40000 60000 SS R Sample number    Figure 5.7. Graphical representation showing the level of influence each parameter has (while keeping others constant) on supersaturation ratio (shown in the last graph).  Chapter Five: Results and Discussions  80  5.4 Carbon Dioxide Stripping Model Previous studies (Fattah et al. 2008a) at this same treatment plant found that the cascade stripper was very effective in saving caustic usage, ranging from 35% to 86%, depending on the operating conditions. A carbon dioxide stripper model was proposed. In the present study, the proposed model was coded in Matlab, so that it could be easily integrated into the process control model. This model takes into account various factors that influence the stripping of carbon dioxide, namely, the baffle number, the effluent recycle ratio, the influent flow rate, the air supply rate, the influent temperature, the influent dissolved CO2 concentration and the influent buffering capacity. This model facilitates prediction of the change in pH within the stripper. The basic equation of the stripping model is given by Equation 5.1. Detailed expressions for A~G, based on a statistical evaluation of the results by Zhang (2006), are provided in Appendix C.                                   SE = SEԧ*A*B*C*D*E*F*G                                                 (5.1) where,       SE = predicted CO2 stripping efficiency, %  SEԧ = CO2 stripping efficiency under reference conditions, 74 %  A = coefficient for baffle number (BN)  B = coefficient for effluent recycle ratio (ERR)  C = coefficient for influent flow rate (IFR)  D = coefficient for air supply rate (ASR)  E = coefficient for influent temperature (IT)  F = coefficient for influent’s dissolved CO2 concentration ([CO2]inf)  G = coefficient for influent’s buffering capacity (IBC) 5.5 Struvite Crystallization Operation Control Windows As mentioned in Chapter Four, various program codes were written in Matlab to be used in the Simulink environment, to control and monitor the pilot scale struvite crystallization process at the LIWWTP. This section provides the various graphical user interface tools, or Chapter Five: Results and Discussions  81  screens, developed for the purpose. The first level control window (Figure 5.8) is used to calculate the in-reactor species concentrations (in terms of mass) and also to calculate the external magnesium flow rate required. The in-reactor concentrations are determined by calculating the centrate and recycle ionic concentrations and using the flow rates. Since the system did not have any flow meters or online analyzers, these values were manually inserted, based on determinations at the pilot plant and in the laboratory. This window provides the necessary information required for the second level window (Figure 5.9), which is the actual heart of the control system.  The second level control block, named reactor pH regulator in the first level block, contains  the  SSR  and  pH  calculator  codes  written  in  Matlab.  In  essence,  this  window calculates the present SSR value in the reactor and compares it to the desired SSR. A difference of SSR= 0.1 between the two SSR values is used as the criterion to trigger a change in the pH set point; this is then passed on to the pH meter attached to the crystallizer. 5.6 Graphical User Interfaces (GUI) used for Operation of Crystallization Process The use of a graphical user interfaces provides a simple and efficient method of operating any system, and therefore, three GUIs were coded in this study – the reactor operation model, the SSR calculator model and the stripper efficiency model. These three GUIs are illustrated in Figure 5.10 to Figure 5.12. The purposes of the reactor operation GUIs were to provide information regarding the conditions required for operating the crystallizer, as well as to provide some useful information, such as struvite production and expected effluent concentrations. As inputs, it requires the concentrations of the various ionic concentrations, the pH, temperature and conductivity, as well as the desired upflow velocity and recycle ratio. The GUI uses the solvepH.m code to determine required pH and calculates the required feed rates, based on flow characteristics. The SSR calculator GUI is a more simplistic window that calculates SSR value, given the required information. Recalling that struvite growth is SSR dependent, this GUI can be used to determine the struvite formation potential at any treatment plant. The carbon dioxide stripper GUI is based on the carbon dioxide stripping model presented earlier, and can be used to predict potential stripping efficiency Chapter Five: Results and Discussions  82  and the expected pH increase. This model can be used to predict if the installation of a stripper will actually be economical, before installing one and spending time, money and resources in carrying out studies.                          Chapter Five: Results and Discussions  83   Figure 5.8. First level of control window for struvite crystallization process. Chapter Five: Results and Discussions  84   Figure 5.9. Second level control block containing the SSR and pH control codes.  Chapter Five: Results and Discussions  85   Figure 5.10. Graphical user interface for reactor operation.  Chapter Five: Results and Discussions  86   Figure 5.11. Graphical user interface for calculating the supersaturation ratio. Chapter Five: Results and Discussions  87   Figure 5.12. Graphical user interface for determining the efficiency of a cascade carbon dioxide stripper. 5.7 Performance Comparison of the Two Systems The two crystallizer systems, identified as Manual Control (System 1) for Reactor 1 and Automated Control for Reactor 2 (R#2), were operated in parallel, with the intention of determining if continuous control of SSR was beneficial for phosphate removal-recovery and struvite growth. Conductivity, pH and temperature were continuously monitored in Reactor 2, and the required pH, for a particular SSR, was controlled in real-time by using the SSR controller program previously described (Chapter Four). No operating parameters were measured or logged in real-time in Reactor 1. Based on the temperature and conductivity at the pilot-scale setting, the required pH in Reactor 1 was calculated manually each day using the SSR controller program. Due to lack of automation, and in the absence of an operator, variations during the day in the required pH were ignored in Reactor 1. Detailed comparison data are provided in Appendix D. Chapter Five: Results and Discussions  88  Figure 5.13 illustrates the difference between the effluents from the two crystallizer systems. Each sample number in the figure represents days on which a sample was taken and tested; due to operational problems, samples could not be taken every day the crystallizers were operated. It should be recalled that the operating centrate was the same for both systems. As shown in Table 5.1, System 2 exhibited better performance in most instances in reducing the phosphate concentration. The difference was significant. Since the concentration of ammonium in the centrate was high and the removal efficiency low, there was little difference  in  ammonium  concentrations  between  the  effluents  from  the  two  systems.  As mentioned above, the centrate phosphate and ammonium concentrations were much lower than expected at the beginning of the study; consequently, a higher-than-required amount of magnesium was added to each of the crystallizers, resulting in high magnesium concentrations in the effluent. This showed that continuous and real-time data collection arevital for process operation efficiency. The latter part (after sample 12) of the magnesium graph is more representative of the effluent magnesium concentrations that would be expected from operating the process.  In terms of purity (% struvite in pellets), System 2 pellets were slightly higher in struvite content.. This was expected, since the two reactors were operated at very similar operating conditions, and the pH of the manually controlled crystallizer was physically changed each day to produce good quality struvite. Although the struvite quality was not very different, operation of Reactor 2 was easier than Reactor 1, since the pH did not need be monitored and controlled. In addition, the effects of temperature and conductivity could not be accounted for in Reactor 1, and therefore, data from Reactor 2 were used. As shown previously, both conductivity and temperature have large impacts on the supersaturation ratio, and therefore, it is important to monitor these parameters. Table 5.1 provides a summary of different parameters observed the two systems.      Chapter Five: Results and Discussions  89   Table 5.1. Summary of results from the operation of the two crystallizers Parameter System 1 – Reactor 1 (without instrumentation and control) System 2 – Reactor 2 (with instrumentation and real-time control) Average (and maximum)a removal efficiency  Phosphate (%) 52 ± 11 (80) 73 ± 8 (88) Ammonium-N (%) 13 ±  4 (30) 15 ± 4 (34) Magnesium (%) 54 ±  10 (100) 63 ± 8 (100) Amount of struvite in pellet % (based on magnesium concentration) 88 ± 2 90 ± 2 Average crushing strength of struvite pellet harvested (g) 2180 ± 670 2460 ± 240 aValues in brackets are the maximum obtained Error at 95% confidence interval Chapter Five: Results and Discussions  90  0 10 20 30 40 50 60 70 0 5 10 15 20 25 30 35 40 Ph os ph at e ( m g/ L) Sample Number Without control With control 0 200 400 600 800 1000 0 5 10 15 20 25 30 35 40 A m m on iu m -N   ( m g/ L) Sample Number Without control With control 0 20 40 60 80 100 0 5 10 15 20 25 30 35 40 M ag ne siu m   ( m g/ L) Sample Number Without control With control  Figure 5.13. Effluent characteristics from the two systems (a) phosphate, (b) ammonium and (c) magnesium concentrations. Sample numbers represents a day’s sample. (a) (b) (c) Chapter Five: Results and Discussions  91  5.8 Solubility Tests Although extensive studies on the value of solubility constants for struvite have been conducted, there still exist significant variations between reported values. The objective of this part of the study was to determine the struvite solubility constant in four water matrices, and to examine the influence of both temperature and water matrix on the results. It was expected that a single value of solubility constant could be derived that could be used in the control model. Solubility experiments were carried out with struvite pellets in a six-station paddle stirrer with square jars in a temperature-controlled room. Based on the results, the Speciation Model was then used to examine the influence of both temperature and water matrix on the results. Experimental data are provided in Appendix E.  Test data showed very similar trends for each of the study temperatures (Figure 5.14). As demonstrated, there are two distinct regions in the data, with the change in slope at a pH of approximately 7.0. As the typical pH operating range of the crystallizer is 7.0 and 8.5 (Fattah et al., 2008b; Forrest et al., 2008a), pKsp determination was restricted to this region.     Chapter Five: Results and Discussions  92  5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 pH pP s 10°C 15°C 20°C Figure 5.14.  pPS eq curve of a batch experiment trial of supernatant at different temperatures (Adapted from Forrest, 2004). 5.8.1 pKsp prediction using speciation model At  each  of  the  testing  temperatures  and  water  matrix,  the  dependence  of  pKsp on pH produced very similar results. Typical results are shown in Figure 5.15. Table 5.2 provides a breakdown of each of the pKsp values determined using the Speciation Model.  The predicted values compared well with values in the literature, and demonstrated that the value was independent (within the standard deviation) of the water matrix used. The results for the four water matrices tested in the present study were 13.58 r 0.11, 13.50 r 0.11 and 13.35 r 0.09 for 10, 15 and 20ºC, respectively. This temperature dependence can then be used to calculate the enthalpy change (ǻH°) for struvite, which is defined as the amount of heat absorbed or released during the course of the reaction; reported values for struvite ǻH° display as much discrepancy as do reported pKsp values.  Chapter Five: Results and Discussions  93  13.0 13.2 13.4 13.6 13.8 14.0 14.2 14.4 14.6 14.8 5.5 6 6.5 7 7.5 8 8.5 9 pH pK sp Supernatant @ 10°C Tap water @ 15°C Centrate @ 25°C Distilled water at 10°C Centrate @ 15°C  Figure 5.15. Typical pKsp predictions using Speciation Model.  Table 5.2. pKsp values for each water matrix as calculated using the Speciation Model 10ºC (n = 6) 15ºC (n = 6) 20ºC (n = 6) Model Average S.D. Average S.D. Average S.D. Distilled Water 13.63 0.19 n/a n/a 13.39 0.20 Tap Water 13.65 0.20 13.44 0.16 13.26 0.06 Supernatant 13.42 0.14 13.63 0.24 13.30 0.03 Centrate 13.63 0.19 13.44 0.16 13.45 0.17 Average of the matrices 13.58 0.11 13.50 0.11 13.35 0.09 Prediction using Equation 5.2 13.60  13.49  13.38  The experimentally determined pKsp values for struvite were combined with values from the literature (Table 5.3) (Bhuiyan et al. 2007; Babiü-IvanþLü et al. 2002; Ohlinger et al. 1999; Burns et al. 1982) in Figure 5.16 and a linear regression of the data was carried out to Chapter Five: Results and Discussions  94  derive an equation (Equation 5.2) relating the influence of temperature on solubility constant. A strong correlation (R2 = 0.88) was found between the results from the different tests. This relationship probably provides the most updated value amongst published data, since it was derived using six different studies with eight different water matrices. From the above conclusions, it can be hypothesized that the pKsp value is independent of the water matrix. This conclusion is important, because it allows a single parameter to be used in developing a control model for the struvite crystallization processes. The average enthalpy value (ǻHo) from all the tests was calculated to be 31.62 kJ/mol, which is similar to values published in the literature (23.62 kJ/mol by Bhuiyan et al. 2007; 34.48 kJ.mol by Aage et al. 1997; 24.23 kJ/mol by Burns et al. 1982).  13.82 + C)( Temp  0.022-   qu sppK    (5.2) y = -0.022x + 13.82 R² = 0.88 12.6 12.8 13.0 13.2 13.4 13.6 13.8 14.0 0 10 20 30 40 50 pK sp Temperature (°C) Centrate Distilled water Supernatant Tap water Ohlinger Babic Burns and Finlayson Bhuiyan  Figure 5.16. Variation of pKsp values with temperature. Error bars at 95% confidence interval.  Chapter Five: Results and Discussions  95  Table 5.3. pKsp values used to derive Equation 5.2 Reference* Water matrix    Temperature          (°C) pKsp Standard deviation Error  Centrate 10 13.63 0.19 0.15  Centrate 15 13.44 0.16 0.13 Fattah (2004) Centrate 20 13.45 0.17 0.14   Distilled water 10 13.63 0.19 0.15 Fattah (2004) Distilled water 15 13.65  Distilled water 20 13.39 0.20 0.16  Forrest (2004) Supernatant 10 13.42 0.14 0.11 Forrest (2004) Supernatant 15 13.63 0.24 0.19 Forrest (2004) Supernatant 20 13.3 0.03 0.02  Forrest (2004) Tap water 10 13.65 0.20 0.16  Tap water 15 13.44 0.16 0.13  Tap water 20 13.26 0.06 0.05  Ohlinger et al. 1998 Synthetic Water 25 13.25  Babiü-IvanþLü et al. 2002 Synthetic Water 25 13.26  Burns and Finlayson, 1982  25 13.12   35 12.97   40 12.94   45 12.84  Bhuiyan et al. 2007 Distilled water 30 13.17 0.05 * Unless mentioned, experiments were carried out in the present study. Error at 95 % confidence limit Chapter Five: Results and Discussions  96  5.8.2 Summary from solubility tests This investigation used an in-house Speciation Model, named SimpleMAP v1.0, to predict values of pKsp. The predicted values compared well with values taken from the literature and demonstrated that the value was independent of the working fluid. This work also revealed that the pKsp is independent of pH in the typical working range of the crystallization process. The data from this study were combined with those found in literature, and a linear relationship between pKsp and temperature was determined. The enthalpy value determined was also similar to those from other studies. The relationship between pKsp and temperature has a significant impact on potential control systems that can be developed for the struvite crystallization process. 5.9 Carbon Dioxide Stripping The  purpose  of  these  experiments  was  to  determine  the  applicability  of  strippers  to reduce caustic usage in a struvite crystallization process, by stripping of carbon dioxide gas from the system. Two types of strippers were operated in parallel to compare the efficiency in their application.  The efficiency of the strippers was determined by comparing the amount and percentage of caustic saved. At the start of the study, the reactors were operated without the strippers, to confirm that they performed identically, so that a direct comparison could be made between the performances of the strippers. The study was operated under four different conditions (Table 4.6), the conditions being the same for both reactors that were operated in parallel in this study. Table 4.7 lists the characteristics of the centrate used. Detailed operational data are provided in Appendix F. Some of the data are part of a study carried out by Sabrina (2007). 5.9.1 Run No 1. In this run, the crystallizers were operated without any stripper. The recycle ratio was set to 6 and the upflow velocity in the crystallizer was 400 cm/min. Chapter Five: Results and Discussions  97   In this run, the removal rates achieved for magnesium, phosphorus and ammonium were 69%, 88% and 7%, respectively in R#1, and 66%, 90% and 10%, respectively, in R#2. On an average, more caustic was used by R#1 than R#2. The average difference of caustic used by the two crystallizers was 91%. Although both systems saw an average pH increase of 0.65, R#1 needed an average 2.35 kg/d of caustic, compared to an average 1.23 kg/d of caustic used by R#2.  Figure 5.17 illustrates the relationship between molar P-removal and the molar caustic use. Although both reactors were identical and were expected to operate similarly, it was found that there were differences with respect to caustic usage. One possible explanation for this is that both strippers were already connected to the respective reactors (compact media stripper with R#1 and cascade stripper with R#2) during the test period. Although the process flow was going directly from the seed hopper to the clarifier, by bypassing the strippers, the compact media stripper was installed directly over clarifier #1, thereby totally blocking the top of the clarifier. On the other hand, the top of clarifier #2 was open, as the cascade stripper was installed about 30 cm above the top surface. The consequence of this setup is that stripped CO2 (while falling along the downpipe) could escape through this opening. As the compact media stripper sealed R#1 system, stripped CO2 eventually dissolved back into the liquid stream. As a result of this, recycle flow from the clarifier #1 carried a solution higher in CO2 to R#1, which reduced the pH of the return stream, and therefore, R#1 required more caustic. Since there was a background difference in the quantity of caustic used by the two systems, the efficiency of the crystallizers was based on the individual systems’ Run #1 data.  Chapter Five: Results and Discussions  98  0 20 40 60 80 100 0.00E+00 5.00E-04 1.00E-03 1.50E-03 2.00E-03 2.50E-03 3.00E-03 1 2 3 4 5 M ol ar  c au st ic  u se M ol ar  P  re m ov al Day R#1 R#2 NaOH#1 NaOH#2  Figure 5.17. Molar phosphorus removal and caustic use in R#1 and R#2 – Run No. 1. 5.9.2 Run No. 2 In this Run, the strippers were connected to the crystallizers and operated with air. The recycle ratio and upflow velocity were 6 and 400 cm/min, respectively.  During this test period, 90% phosphorus removal was achieved by both systems. The average removal efficiency of ammonium was 6% and 5% in R#1 and R#2, respectively. Average magnesium removal efficiency was 74% in R#1 and 75% in R#2. Throughout the test period, the amount of caustic used by R#2 was consistently lower than by R#1. On average, R#2 used 0.84 kg/d of caustic, whereas in case of R#1, this amount was 1.41 kg/d. It should be noted that almost the same amount of phosphorus was removed by both the crystallization processes during this period. Figure 5.18 illustrates the relationship between molar caustic consumption and molar P-removal during the test run.  Chapter Five: Results and Discussions  99  0 10 20 30 40 50 1.98E-03 2.08E-03 2.18E-03 2.28E-03 2.38E-03 2.48E-03 2.58E-03 1 2 3 4 M ol ar  c au st ic  u se M ol ar  P  re m ov al Day R#1 R#2 NaOH#1 NaOH#2  Figure 5.18. Molar phosphorus removal and caustic use in R#1 and R#2 – Run No. 2.  Comparing these results with the results obtained by running the crystallizers without strippers, it can be seen that, by introducing the cascade stripper into the system, an average of 32% savings in caustic use was achieved, while the compact media stripper saved 40%. However, the absolute amount of caustic used by R#1 was higher than for R#2. In addition to the problem mentioned earlier regarding placement of the stripper, it was found that the compact media stripper was also prone to clogging, which may explain the reason for limited carbon dioxide stripping. These two factors may explain why consistently higher caustic was used in R#1. Thus, the cascade stripper proved to be more effective in stripping CO2 than the compact media stripper, and lowered the daily requirement of caustic. The overall CO2 removal rate for R#1 was 11%, and 20% for R#2. 5.9.3 Run No. 3 In this experiment, the strippers were operated without an external air supply. The recycle ratio and upflow velocity was kept unchanged from the previous two runs.  Chapter Five: Results and Discussions  100  The  other  operating  parameters  remained  the  same  as  in  the  second  run.  Both  systems were able to remove around 90% of phosphorus during this period, as well. The ammonium removal efficiency increased in both reactors, with a value of approximately 18% in R#1 and 15% in R#2, compared to 6% and 5%, respectively, during the second run. As expected, without an external air supply, the caustic use rate increased in both systems. Again, the cascade media stripper used smaller amounts of caustic than the compact media stripper. The molar caustic use per mole of phosphorus removed showed the same trend as in previous runs, as illustrated in Figure 5.19. During this period, the cascade media stripper saved an average 26% of caustic chemical, compared to 32% in the previous run. On a daily basis, the compact media stripper used an extra 4% of caustic, compared to the previous run.  Without an external air supply, the CO2 removal efficiency was expected to decrease. Surprisingly, a slightly higher amount of CO2 removal was achieved by the compact media stripper. One explanation for this could be that the stripper media was cleaned before the start of the run, resulting in higher CO2 stripping efficiency. The average CO2 removal efficiency for R#1 was around 14% during this test, whereas this amount was only about 11% in the previous  run,  with  the  air  supply.  On  the  other  hand,  the  CO2 removal efficiency for R#2 decreased, as expected, from 20% to 17%.      Chapter Five: Results and Discussions  101  0 10 20 30 40 50 0.00E+00 5.00E-04 1.00E-03 1.50E-03 2.00E-03 2.50E-03 1 2 3 4 M ol ar  c au st ic  u se M ol ar  P  re m ov al Day R#1 R#2 NaOH#1 NaOH#2  Figure 5.19. Molar phosphorus removal and caustic use in R#1 and R#2 – Run No. 3. 5.9.4 Run No. 4 In this test run, the recycle ratio and the upflow velocity were 9 and 75 mm/sec, respectively, and the external air supply resumed.  In this run, R#1 was able to remove an average 89% of phosphorus. The R#2 system was slightly better in removing phosphorus, achieving an average 92% during this test period. Both systems showed improvement in magnesium removal, compared to the previous runs. The magnesium removal efficiency for R#1 and R#2 averaged 55% and 42%, respectively. In this run, the ammonium removal efficiency averaged 9% in R#1 and 8% in R#2. Regarding daily caustic use, higher recycle ratio and upflow velocity proved to have positive impacts on the performance of the strippers. R#1 used an average 1.35 kg/d of caustic during this time. The improvement was more pronounced in case of R#2, where the average caustic used during this run was only 0.66 kg/d. Both systems were removing almost same amount of molar phosphorus from the system (Figure 5.20).  Comparing results from this run with the results from Run No. 1 showed that the amount of caustic saved by the cascade stripper was 46%, and by the compact media stripper 42%. In Chapter Five: Results and Discussions  102  Run 4, the cascade stripper showed an improvement under higher recycle ratio and upflow velocity. Under a higher recycle ratio and upflow velocity, both strippers performed better in stripping CO2 than in the previous runs, and, as a result, less caustic was used by the systems.  0 10 20 30 40 50 60 70 0.00E+00 5.00E-04 1.00E-03 1.50E-03 2.00E-03 2.50E-03 1 2 3 4 5 M ol ar  c au st ic  re m ov al M ol ar  P  re m ov al Day R#1 R#2 NaOH#1 NaOH#2  Figure 5.20. Molar phosphorus removal and caustic use in R#1 and R#2 – Run No. 4.  Both strippers proved effective in reducing caustic usage. The amount of caustic saved by the cascade stripper ranged from 26% (without an external air supply) to 46% (with air, higher  recycle  ratio  and  upflow  velocity).  In  the  previous  study  at  the  LIWWTP,  the reduction in caustic addition ranged from 46% to 65% (Fattah et al. 2009). Even though the previous study did not use an external air supply, the operators were able to obtain a higher amount of caustic savings than in the present study, where only 26% caustic was saved when the stripper was run without an external air supply (the amount was 32% with an external air supply at the recycle ratio of 6.0). However, it should be noted that the operating pH during the previous study was lower (7.9 and 8.0) than in the current study, where the operative pH was maintained at 8.1; hence there was a higher caustic saving. Therefore, it can be concluded that, with a lower operating pH (if conditions satisfy all criteria of struvite formation and recovery), the cascade stripper will be more effective in saving caustic; thus, a Chapter Five: Results and Discussions  103  lower cost of struvite production can be expected. A summary of all the test results is provided in Table 5.4.  Table 5.4. Summary of findings from four stripping tests  Without Stripper With Stripper Run No 2a Run No 3b Run No 4c R #1 R #2 Compact media stripper Cascade stripper Compact media stripper Cascade stripper Compact media stripper Cascade stripper P removal (%) 88 90 90 90 90 90 89 92 Caustic use (kg/d) 2.35 1.23 1.41 0.84 1.51 0.91 1.35 0.66 Caustic savingsd (%)  -  40 32 36 26 42 46 CO2 stripping (%)  - 11 20 14 17 14 21 NH3 stripping (%)  - 2 1 9 7 6 5 a b c d  : With air, Recycle ratio (RR) = 6, Upflow velocity = 400 cm/min : Without air, RR = 6, Upflow velocity = 400 cm/min : With air, RR = 9, Upflow velocity = 450 cm/min : Caustic savings were calculated comparing the results of the Runs (2nd, 3rd and 4th) to Run 1 5.10 Potential for Stripper Fouling 5.10.1 Clogging of the compact media stripper Comparing the two strippers, operational problems in the compact media stripper were more frequent. From the beginning of the study, the major problem that was faced in operating the compact media stripper was its susceptibility to clogging. When the experiments began, the plastic balls were suspended between the liquid distributor and the Chapter Five: Results and Discussions  104  supporting plate. However, it was soon noticed that flow restrictions occurred and, upon investigation, it was found that the balls were totally coated with fine struvite particles and suspended solids. Little success was achieved in removing the solids, despite applying a hot water jet to the balls inside the stripping column. In order to reduce the coating effects, the balls were then rearranged by attaching them to vertical strings, as illustrated in Figure 4.8d. Although this setup permitted easier cleanup, the stripper required constant monitoring and had to be cleaned more regularly (on alternate weeks) than the cascade stripper.  The cascade stripper, which was made of plexiglass, was subject to lower solid buildup due to its relatively smooth surface along which the water flowed. As expected, with the stripping of carbon dioxide, some struvite and suspended solids build-up was visible on the sides, after prolonged operation. This can be explained in terms of the increase in pH of the water matrix. As the pH increased, the SSR of the water matrix increased, which provided suitable conditions for struvite growth.  However, since the water matrix passing through each baffle is low in magnesium and phosphate – most of it had been removed in the crystallizer – there was a lower potential for struvite growth. The amount of struvite present did  not  hamper  the  operation  or  clog  the  stripper.  The  path  of  the  water  flow  was  always clean, as struvite did not accumulate in place. Moreover, cleanup of the stripper was straightforward as the accumulation was easily removed by hosing it with hot water. Figure 5.21 shows a cross-section of the stripper, showing the struvite accumulation.  Chapter Five: Results and Discussions  105   Figure 5.21. Struvite accumulation in stripper during operation. 5.10.2 Cost analysis Due to the higher efficiency and ease of operation of the cascade stripper, it was decided that future studies would be made with this stripper. Hence, the following sections contain details on this stripper only.  The construction cost of the stripper was reasonable (about $ 2000), relative to the cost of the pilot-scale struvite reactor, which was around $ 20,000 (including materials and labor). Once installed, there was little or no cost associated with its maintenance. As mentioned previously, the incorporation of the stripper resulted in lower caustic use to keep the crystallizer at the desired pH level. Lower caustic use can be directly correlated to operational costs. Lulu Island WWTP is a relatively small secondary treatment plant treating, on average, seventy five million liters per day (75 MLD). The crystallizer in the present study operated with a centrate flow of only 2.61 L/min, which is equivalent to approximately Struvite and suspended solid accumulation along water matrix Chapter Five: Results and Discussions  106  3800 L/day. This is only 3% of the total centrate production at the treatment plant. Based on the current caustic price of $ 2.57/kg, the potential yearly plant savings in caustic is substantial. Based on this calculation, the potential caustic saving at the larger (average 469 MLD) treatment plant in Metro Vancouver’s Annacis Island could be as high as $ 79,200. A summary of the cost study is given in Table 5.5.  Table 5.5. Cost analysis for caustic usage in a pilot scale struvite crystallizer fitted with a cascade stripper  Without Strippers With Strippers    Run 1 Run 2a Run 3b Run 4c  Caustic use (kg/d) 1.23 0.84 0.91 0.66 Caustic savings (kg/d) - 0.39 0.32 0.57 Caustic savings (%)  32 26 46 Current caustic savingsd $/day)  1.00 0.82 1.46 Potential annual savings at Lulu Island WWTPe ($) 12,200 10,000 17,800 Potential annual savings at Annacis Island WWTPe ($) 54,500 44,500 79,200 a Stripper run with external air b Stripper run without external air c Stripper run with higher upflow velocity and RR d Caustic cost $ 2.57/kg e For full scale installations 5.10.3 Conclusions Based on the results from the pilot-scale study of phosphorus recovery with carbon dioxide stripping, the following conclusions can be drawn.  Chapter Five: Results and Discussions  107  Through the process, over 90% of phosphorus was easily removed from the centrate at a pH of 8.1, with or without air stripping. With and without air stripping, an average of 32% and 26% caustic savings was achieved using the cascade stripper, compared to the no- stripper condition, respectively. For the same conditions, the compact media stripper saved, on average, 40% and 36%. It is expected that, with a lower operating pH (if conditions satisfy all criteria for struvite formation), the cost of the crystallizer could be recovered within 3-4 years; consequently, the cost of producing struvite pellets would be reduced. Although struvite did accumulate on the sides and baffles of the stripper, there were no plugging problems related to the presence of struvite. The accumulation was easily removed by washing with hot water. On the other hand, although the compact media stripper was more efficient in reducing caustic usage, its operation and maintenance was more difficult and time consuming. The study confirms that stripping of carbon dioxide, to raise pH in the production of struvite, is a viable means of reducing caustic usage, and thereby reducing operating and production costs. 5.11 Prediction of Magnesium Requirements from Conductivity-pH Measurements Since the quantification of magnesium in the process feed and crystallizer is time consuming, a new detection method was needed, that would be able to provide information regarding magnesium concentrations and application rates in real-time, or at least require little time to provide results. One of the two methods examined in the present study is the use of conductivity and pH to determine magnesium application rates. 5.11.1 Theoretical versus practical change in conductivity In the present study, experiments were carried out in the laboratory to predict (using Equation 2.10) the conductivity increase with the addition of external magnesium chloride, and then to compare this value to the actual change in conductivity. As illustrated in Figure 5.22a, the actual conductivity in distilled water could be predicted according to this equation, with a high degree of accuracy. However, when the same principle was used to predict the conductivity increase in real centrate solution, the values did not increase as predicted (Figure 5.22b). This was assumed to be due to the formation of struvite, and removal of other Chapter Five: Results and Discussions  108  ions through subsidiary reactions. The initial (Measured 1 in Figure 5.22b) rate of change in conductivity was lower compared to the latter part, indicating that the rate of increase in ionic concentrations was lower. After 30 mL (Mesured 2 in Figure 5.22b) of magnesium chloride addition, the rate of conductivity increase changed and became higher than that at the beginning. The location of the increased rate of change was then assumed as the bending/transition point where all phosphate is expected to have precipitated out. The process of locating the bending point is further explained in the following section.  Chapter Five: Results and Discussions  109  0 200 400 600 800 1000 0 2 4 6 8 10 12 C on du ct iv ity  (m ic ro S/ cm ) MgCl2 added (ml) Predicted Measured Conductivity  (a) y = 0.0514x + 8.0196 y = 0.0245x + 7.978 R2 = 0.986 y = 0.0714x + 6.5886 R2 = 0.982 7 7.5 8 8.5 9 9.5 10 10.5 11 0 10 20 30 40 50 Volume of MgCl2 added (mL) C on du ct iv ity  (m S/ cm ) Predicted Measured 1 Measured 2  (b) Figure 5.22. Influence of magnesium chloride addition on the conductivity – (a) predicted vs. measured in distilled water (with no P and N), and (b) predicted vs. measured in centrate sample. Chapter Five: Results and Discussions  110  5.11.2 Experimental runs to determine bending point for magnesium addition The objective of this section was to determine the presence-absence of a bending or transition point in the conductivity and pH trends, when magnesium chloride is added to the water matrix.  Figure 5.23 and Figure 5.24 illustrate the influence of magnesium chloride additions on the conductivity and pH in experimental runs with synthetic and LIWWTP centrate. Detailed analytical data are provided in Appendix G. By using the first derivatives of pH and conductivity, a more graphic and ‘easy-to-comprehend’ location of the bending point can be determined. The bending point is the location where the dpH/dMg ratio is highest, while the dcon/dMg ratio is lowest. This corresponds to 1.5 mL of MgCl2 in Figure 5.23. Figure 5.25 is expanded to provide an explanation of the phenomena occurring, and the method to detect the bending point (which is enhanced in Figure 5.26). In this experiment, three conductivity zones were distinguishable – the initial rapid increase in conductivity, the middle low conductivity rise and a final increased rate of conductivity change. By theory, the slope of the conductivity curve in the final section should be higher than that at the beginning, but as was found from the test, this was not the case. Three zones in the pH-magnesium chloride addition was also noticeable – the initial slow pH change, the rapid pH decrease and the final slow pH decrease. The addition of magnesium chloride to water would normally increase the pH of the system into which it is added. However, since struvite formation is accompanied by lowering of the pH, the end of the middle section was hypothesized as the end of struvite formation.  Since  the  final  pH  section  did  not  show  an  increase  in  pH,  as  expected,  due  to continued addition of magnesium chloride, it could be hypothesized that certain reactions, which reduces the pH of the system, were still occurring. The equations shown in the figure are based on linear trend lines of the three distinguishable zones. Chapter Five: Results and Discussions  111   1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 6.5 7 7.5 8 8.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 C on du ct iv ity  (m S/ cm ) pH Volume of MgCl2 added, mL pH conductivity  (a) -0.06 -0.01 0.04 0.09 0.14 0.19 0.24 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 dc on /d M g dp H /d M g Volume of MgCl2 added, mL dpH/dMg dcon/dMg (b) Figure 5.23. (a) Influence of magnesium chloride addition on pH and conductivity of a synthetic wastewater and (b) change in pH and conductivity as a function of magnesium chloride addition -  Run 1. Chapter Five: Results and Discussions  112  1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 6.5 7 7.5 8 8.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 C on du ct iv ity  (m S/ cm ) pH Volume of MgCl2 added, mL pH conductivity  (a) -0.06 -0.01 0.04 0.09 0.14 0.19 0.24 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 dc on /d M g dp H /d M g Volume of MgCl2 added, mL dpH/dMg dcon/dMg   (b) Figure 5.23. cont’d. (a) Influence of magnesium chloride addition on pH and conductivity of a synthetic wastewater and (b) change in pH and conductivity as a function of magnesium chloride addition - Run 2. Chapter Five: Results and Discussions  113  1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 6.5 7 7.5 8 8.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 C on du ct iv ity  (m S/ cm ) pH Volume of MgCl2 added, mL pH conductivity  (a) -0.08 -0.03 0.02 0.07 0.12 0.17 0.22 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 dc on /d M g dp H /d M g Volume of MgCl2 added, mL dpH/dMg dcon/dMg (b) Figure 5.23.cont’d. (a) Influence of magnesium chloride addition on pH and conductivity of a synthetic wastewater and (b) change in pH and conductivity as a function of magnesium chloride addition - Run 3. Chapter Five: Results and Discussions 114   7 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.60 7.70 7.80 7.90 8.00 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10 C on du ct iv ity  (m S/ cm ) pH Volume of MgCl2 added, mL pH conductivity  6 6.1 6.2 6.3 6.4 7.60 7.65 7.70 7.75 7.80 7.85 7.90 0 1 2 3 4 5 6 7 8 9 10 C on du ct iv ity  (m S/ cm ) pH Volume of MgCl2 added, mL pH  6.8 6.9 7 7.1 7.2 8.50 8.55 8.60 8.65 8.70 8.75 8.80 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 C on du ct iv ity  (m S/ cm ) pH Volume of MgCl2 added, mL pH conductivity  Figure 5.24. Influence of magnesium chloride addition on pH and conductivity in centrate sample for three different runs. Chapter Five: Results and Discussions 115  y = -0.0057x + 7.9171 R² = 0.8 y = -0.1x + 8.2067 R² = 0.9494 y = -0.0129x + 7.8462 R² = 0.9512 y = 0.0779x + 7.1775 R² = 0.9921 y = 0.02x + 7.3433 R² = 0.75 y = 0.0424x + 7.2446 R² = 0.9963 7 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.60 7.70 7.80 7.90 8.00 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10 C on du ct iv ity  (m S/ cm ) pH Volume of MgCl2 added, mL pH pH_2 pH3 conductivity conductivity_2 conductivity3 Figure 5.25. Influence of magnesium chloride addition on conductivity and pH of centrate – determination of transition point. Chapter Five: Results and Discussions 116  0 0.01 0.02 0.03 0.04 0.05 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.510 -d co n/ dM g dp H /d M g Volume of MgCl2 added, mL dpH/dMg dcon/dMg  Figure 5.26. Change of pH and conductivity with magnesium chloride addition in centrate matrix. The primary data used is the same as for Figure 5.25.  Several tests were carried out to determine if the process of finding the transition point could be replicated, and each time the same transition location was found. Upon determining the molar concentrations at the bending points, it was found that, for each mole of phosphate removed, the Mg:P molar ratio was between 1.3 and 2.0 (Figure 5.27). This ratio is important, because, although it does not signal the end of the reaction, it determines the quantity of magnesium chloride required to remove one mole of phosphate. By knowing the phosphate concentration through use of online analyzers, the amount of Mg required to completely precipitate phosphate can then be calculated more easily; that is, the phosphate molar concentration would be multiplied by a factor of 1.3-2.0 to get the required magnesium dosages.  Although pure struvite crystallization occurs at a Mg:P molar ratio of 1:1, the Mg:P molar ratio determined here can be explained in terms of the activity and concentration of MgCl2 in the standard. Normal detection of Mg, by AA, provides values for the total magnesium concentration. However, the portion of the measured Mg that actually takes part in the reaction is represented by the activity of soluble magnesium ion, which, with most ions Chapter Five: Results and Discussions 117  concerned with struvite precipitation, is lower than the actual concentration. Therefore, this technique provides a better estimate of the dosages required. Conductivity changes, due to magnesium chloride addition on swine manure slurries by Shepherd et al. (2009) found that an inflection point occurred around 50% of the stoichiometric magnesium demand. However, they  were  unable  to  use  their  data  to  provide  any  ratio  or  estimate  the  rate  of  magnesium addition that would provide for optimized application.  0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.5 1.0 1.5 2.0 2.5 M ol es  o f P  re m ov ed Mg:P molar ratio at transition point  Figure 5.27. Relationship between moles of phosphate removed and Mg:P molar ratio at the transition point. 5.12 Use of Chemicals to Reduce Phosphate Interference As previously mentioned, for hardness tests, it is important that phosphate concentrations are low enough to reduce interference during the test. Two chemicals – polyaluminum chloride (PAC) and alum - were tested to determine their efficiency in reducing the phosphate concentration to levels that would be acceptable for hardness testing. 5.12.1 Use of polyaluminum chloride (PAC)  Results from multiple tests (data presented in Appendix H) using centrate as working fluid were used to derive an equation that could be used to predict the quantity of PAC Chapter Five: Results and Discussions 118  required, given the initial phosphate concentration. Data from these tests were then compared to tests by Forrest (2004) to determine if the method was feasible.  During the present study, the concentration of PAC was lower than that used by Forrest (2004), so that more data points could be obtained within the range below the point of phosphate re-release.  Figure 5.28 illustrates the results of multiple tests, and it was found that the initial removal of phosphate was high and reached an asymptotic curve with 1 ml of PAC addition. In all experiments, PAC additions of more than 1 ml increased the phosphate concentration. However, since phosphate concentrations below 20 mg/L are not known to cause major interference during hardness tests, this method of phosphate reduction can be efficiently conducted. The first two runs produced a linear relationship (Figure 5.28) in the range tested; however, the third run reduced phosphate concentrations more rapidly than the first two. Therefore, the relationship between PAC addition and phosphate removal should be determined for each the water matrix used for struvite crystallization. This linear relationship, once developed, could be applied in subsequent experiments, to determine the amount of PAC required. However, it must be pointed out that the amount of PAC required may be dependent on the total solids present in the water matrix, as some amount will be consumed in reducing the solids.  Chapter Five: Results and Discussions 119  Run 1, y = -102.4x + 97.77 R² = 0.949 Run 2, y = -126.2x + 94.729 R² = 0.8855 Run 3, y = -178.73x + 95.98 R² = 0.9131 0 20 40 60 80 100 120 0 0.2 0.4 0.6 0.8 1 Volume of PAC added (ml) Ph os ph at e co nc en tra tio n (m g/ L) Run 1 Run 2 Run 3  Figure 5.28. Removal of phosphate from centrate with addition of PAC. 5.12.2 Use of alum Multiple tests were carried out to determine the quantity of standard alum (855 mg/L as Al) to be added to the LIWWTP centrate and to reduce phosphate concentration. The data are presented in Appendix H. The influence of alum addition on phosphate concentration is illustrated in Figure 5.29. The trendline depicts data from several tests carried out in the laboratory. As illustrated, a high linear correlation (R2 = 0.9) was found between phosphate concentration and the volume of standard alum (0.0159 M) added. Given the initial phosphate concentration, this equation (Figure 5.29 for centrate) can now be used in subsequent tests to decide on the amount of alum to be added. In order to predict the applicability of the method with different water matrices, centrate from the Annacis Island wastewater treatment plant (AIWWTP) was also used. Although the two water matrices did not produce the same slope, the rate of phosphate removal was linear in both instances. Therefore, this method can be used for initial testing of new wastewater samples to determine the equation, which can then be used to determine the amount of alum needed to reduce phosphate concentrations.  Chapter Five: Results and Discussions 120  y = -15.139x + 187.25 R² = 0.9797 y = -19.564x + 100.31 R² = 0.8973 0 50 100 150 200 250 0 2 4 6 8 10 12P ho sp ha te  c on ce nt ra tio n (m g P/ L) Volume of standard alum added, mL AIWWTP centrate LIWWTP centrate  Figure 5.29. Influence of alum on phosphate concentration.  The influence of alum addition on magnesium concentration in the sample was also studied (Figure 5.30) by measuring the concentrations over the range of alum addition. Results showed that there was little influence of adding alum, and so using the chemical as a phosphate precipitator did not lower the magnesium concentration. Although the results indicate that both PAC and alum can efficiently remove phosphate from the water matrix, the final choice of chemical will depend on the availability and choice of the experimenter. In the present study, the use of alum was deemed to be more suitable.  Chapter Five: Results and Discussions 121  0 2 4 6 8 10 12 14 16 18 0 1 2 3 4 5M ag ne si um  c on ce nt ra tio n (m g/ L) Volume of standard alum added, mL Test 1 Test 2 Test 3 Test 4  Figure 5.30. Influence of alum addition on magnesium concentration. Chapter Five: Results and Discussions 122  5.13 Hardness Test Results The hardness method for determining the magnesium concentration required the determination of both total hardness and Ca-hardness. The result for calcium detection has been included in this section to illustrate the use and efficiency of the method in determining the concentration of the metal.  From Ca-hardness test, the calcium concentration was calculated and compared with values determined by the AA. Figure 5.31 illustrates that, in most instances, the hardness method gave higher than actual values; the average absolute error in the values was only 3.8 mg/L (15%). Given that hardness test results, by themselves, have limitations (such as in detecting the end point), the predictability of this method was deemed acceptable.  Similar to the calcium values, it was found that, on most occasions, magnesium concentrations determined by the hardness test method were higher than those measured by the AA (Figure 5.32); the absolute average error between the two tests was only 2.0 mg/L (20%). The relatively higher experimental values can probably be attributed to, in addition to titration limitations/errors, the addition of MgCDTA inhibitor during the hardness test. Tests carried out to determine magnesium concentration in a parallel study found an absolute error of 3.9 mg/L (Dirk, B. pers. comm., 2009). Therefore, although not 100% accurate, this method provides a good estimate of the Mg+2 concentration in the sample; also the method can be performed on site, and is much quicker and cheaper than AA or ICP. Detailed data for the hardness tests are provided in Appendix I.  Chapter Five: Results and Discussions 123  0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70C al ci um  c on ce nt ra tio n fr om  A A  (m g/ L) Calcium concentration from hardness method (mg/L)  Figure 5.31. Comparison of calcium concentrations via hardness method and AA. The line depicts the equivalence line.  0 5 10 15 20 25 30 0 5 10 15 20 25 30 M ag ne si um  c on ce nt ra tio n by  A A  (m g/ L) Magnesium concentration by hardness method (mg/L)  Figure 5.32. Comparison of magnesium concentrations via hardness method and AA. The line depicts the equivalence line. Chapter Five: Results and Discussions 124  5.14 Composition of Struvite Formed During Pilot Scale Operation Struvite formed during the study was tested for various metals, ammonium-nitrogen and phosphate  to  determine  the  composition  and  purity  of  the  product.  ICP  tests  were  also performed on certain samples to determine the concentration of heavy metals. It is important to note the heavy metal concentrations, because they may determine whether the struvite product can be used as a fertilizer, either on land or in a nutrient-deficient water body. Previous analysis of struvite from the same location found that the concentrations of heavy metals in the struvite grown were much lower than those present in commercial fertilizers and phosphate ore (Fattah et al. 2008b). Struvite harvested from the crystallizer with automated process control was consistently of larger size, and with higher breaking strength than that harvested at the same location in a previous (Fattah et al. 2008b) study. Operational data collected during the pilot-scale study at LIWWTP are detailed in Appendix J. 5.14.1 Magnesium to phosphate (Mg:P) molar ratio Various studies (Fattah et  al. 2008b; Stratful et  al. 2001) have determined that the quantity of magnesium in a pellet influences the morphology and purity of the struvite. For pure struvite pellets, the Mg:P molar ratio should be unity; however, as illustrated in Figure 5.33, for all the samples tested, this ratio was lower than unity, inferring that the pellets/solids tested were not 100% struvite; this indicated that other phosphate compounds may be present. It was not possible to test what other compounds may have been present, but studies carried out previously (Bhuiyan et al., 2008) have suggested the presence of calcium phosphate, hydoxyapatite (HAP) and dolomite.  Although the potential for formation of these non-struvite compounds is low at the operating pH of 8.1, the mode of operating the crystallizer, with respect to pH stabilization, may help to explain their presence. At the injection zone, where concentrated caustic is added to the system, the localized pH may be much greater than the set point. This high pH can lead to the utilization of calcium in forming a phosphate compound. By simulating potential for solid formation at different conditions, Bhuiyan et al. (2007) found that, at very low Ca:P Chapter Five: Results and Discussions 125  ratios, the solution could be supersaturated with dolomite and HAP. It has been suggested that excess magnesium in the reactor can lead to greater phosphate percentage removal (Stratful et al. 2001), but via formation of struvite of lower purity (Demeestere et al. 2001). Although purity of recovered struvite is important, there are other side benefits (as explained below) of having excess magnesium in the effluent.  0.70 0.75 0.80 0.85 0.90 0.95 1.00 1.00 1.98 2.36 3.33 4.70 M g: P m ol ar  ra tio Pellet size (mm) Sample 1 Sample 2 Sample 3 Sample 4 Sample 5  Figure 5.33. Measured magnesium to phosphate molar ratio of struvite samples. 5.14.2 Ammonium-nitrogen (N:P)  to phosphate molar ratio Figure 5.34 illustrates the measured molar ratios of ammonium-N to phosphate in harvested struvite. As was the case with Mg:P molar ratios, the N:P molar ratios are below unity, suggesting again the existence of a phosphate compound that does not contain nitrogen. Possible compounds include, but are not limited to, anapaite {Ca2Fe(PO4)2.4H2O} and vivianite {Fe3(PO4)2.8H2O}. The presence of vivianite crystals was confirmed, along with  struvite,  in  the  anaerobic  digester  at  the  Annacis  Island  WWTP  (Brian  Hystad,  pers. comm., 2009).  Chapter Five: Results and Discussions 126  0.80 0.85 0.90 0.95 1.00 1.00 1.98 2.36 3.33 4.70 N :P  m ol ar  ra tio Pellet size (mm) Sample 1 Sample 2 Sample 3 Sample 4 Sample 5  Figure 5.34. Measured ammonium-nitrogen to phosphate molar ratio of struvite samples. 5.14.3 Purity Although the loss of ammonium from solution could be used as an indication of the amount of struvite precipitated, the purity (% struvite) in the present study was calculated based on the amount of magnesium present in the pellet. The reasoning for this is that there is less magnesium present in the pellet than required to fulfill the ammonium loss conversion to struvite (Mg:N molar ratio is not unity). However, it was assumed that no other magnesium- based phosphates were formed. Overall, based on magnesium concentration alone, the pellets were on average 90 ± 2% struvite. Table 5.6 provides details of the struvite analyses.          Chapter Five: Results and Discussions 127  Table 5.6. Composition of struvite formed at LIWWTP Molar ratios Sample number Mass of sample (g) P (mg/L) N (mg/L) Mg (mg/L)     Fe (mg/L)  (mg/L)    Ca (mg/L) Mg:P N:P Ca:Mg % Purity (based on Mg) 1 0.209 574 236 366 0.6 0.5 0.813 0.908 0.001 89 2 0.203 536 228 363 0.6 7.1 0.863 0.941 0.012 91 3 0.212 553 238 386 0.6 3.7 0.890 0.951 0.006 93 4 0.206 541 229 411 0.5 4.0 0.967 0.934 0.006 102* 5 0.238 583 242 394 0.5 5.6 0.861 0.916 0.009 85 6 0.209 516 222 344 0.7 9.2 0.851 0.953 0.016 84 7 0.224 580 255 391 1.4 7.5 0.859 0.973 0.012 89 8 0.218 552 245 373 1.5 8.3 0.860 0.982 0.013 87 9 0.213 525 216 357 0.8 9.0 0.866 0.909 0.015 86 10 0.218 560 237 370 0.6 7.7 0.840 0.935 0.013 87 11 0.212 539 238 359 1.6 7.6 0.850 0.977 0.013 87 12 0.209 530 237 373 1.5 5.8 0.897 0.989 0.009 91 13 0.207 550 237 360 1.4 15.5 0.835 0.953 0.026 89 14 0.203 529 229 358 1.2 6.9 0.861 0.958 0.012 90 15 0.214 562 244 369 1.6 5.3 0.837 0.960 0.009 88 16 0.205 550 239 368 1.7 5.2 0.853 0.961 0.008 91 17 0.135 364 156 257 1.1 15.2 0.899 0.948 0.036 97 18 0.200 516 223 361 1.4 8.3 0.892 0.956 0.014 92 19 0.222 563 245 385 1.5 7.7 0.871 0.963 0.012 88 Average       0.867  0.951 0.008 90 95 % confidence ± 0.02 ± 0.01 ± 0.00 ± 2  *Possible analytical error Chapter Five: Results and Discussions 128  5.14.4 Shell formation – composition, SEM pictures During one harvest of struvite, pellets with a distinct outer shell were formed (Figure 5.35). Visual inspection of these pellets showed that they had one dense inner part similar to other  harvests,  while  a  second  outer  shell  was  coated  on  top  of  the  inner  crystal.  SEM pictures (Figure 5.35b) confirmed the two-layer structure, with a definite boundary between the two layers. However, the inner core was not as fused as the rest of the pellet. Although a detailed explanation could not be provided, it was thought that this abnormality may be due to changes in the operating conditions. During the few days preceding the harvest, the reactor SSR fluctuated considerably due to pH pump-related problems, resulting in an unstable reactor pH. In addition, the suspended solids content in the centrate was also very high during this period; normal total suspended solids (TSS) content in the centrate ranged from 600-800 mg/L, but during this particular event, the TSS increased to as high as 2500 mg/L. These fluctuations may have caused the inner pellet to act as a seed for the outer shell, hence showing the importance of operational stability, with respect to the growth of good quality pellets. In terms of composition, the two layers did not show significant variation (Table 5.7). The practical consequence of this shell-structured struvite pellets is explained in Section 5.16.5.  Table 5.7. Composition of shell-structured struvite pellet  Mg:P N:P Inner part 0.861 0.916 Exterior shell 0.857 0.912   Chapter Five: Results and Discussions 129   (a)                                                     (b) Figure 5.35. Shell formation of pellets (a) actual struvite and (b) SEM picture taken at x 50 magnification. Boundary 1 Boundary 2 Chapter Five: Results and Discussions 130  5.14.5 Compactness of struvite pellets The compactness of struvite pellets was determined through SEM imagery (Figure 5.36). Figure 5.36a also shows the fusing of crystals in the pellet. It was observed that the compactness of the pellets was “different”, and depended on factors such as size and upflow velocity. Struvite crystals, which have a negative zeta potential in the pH ranges suitable for phosphorus recovery, tend to repel each other (Le Corre et al. 2007). It was found that metal salts, such as alum and ferric chloride, and cationic polymers, were efficient in growing struvite flocs. Although the probability of excess magnesium chloride in the reactor acting as a binding agent could not be determined conclusively, it is postulated that it may act as a destabilization agent in reducing the zeta potential of struvite crystals. This reduction in zeta potential would allow individual struvite crystals to come into contact with each other and allow increased compactness, and therefore density. The practical importance of this variety of  struvite  pellet  is  that  they  will  contain  greater  nutrient  per  unit  volume,  and  therefore reduce the space required for transportation and storage.  (a) (b)  Figure 5.36. SEM pictures of struvite pellets: (a) illustrating the fusing of struvite crystals (500 magnification) and (b) compactness of the crystals (at 250 magnification). Chapter Five: Results and Discussions 131  5.14.6 Influence of upflow velocity on compactness Higher flow, in a fully loaded crystallizer, allows for greater impacts and collisions between pellets, which, in turn, can be responsible for the compactness of the solids. A fully loaded crystallizer, that is, a crystallizer in which each section is filled with fluidized struvite pellets, can also allow individual crystals to come in close contact with each other, leading to higher collisions and agglomeration, or growth, of the particles. However, based on practical experience, there seems to be an upper limit to upflow velocity at which a crystallizer should be operated at, in order to produce good quality pellets. The present study operated the crystallizers at mainly two upflow velocities – 400 cm/min and 500 cm/min. An upper limit of 400 cm/min upflow velocity was found to be “efficient” for the crystallizer used. Operating the process at 500 cm/min caused too many strong inter-particle collisions that resulted in the formation of broken pellets. These broken pellets destroy the shape and reduce the crushing strength of the pellets, hence reducing the marketability of this product. Upflow velocity is discussed further in Section 5.15.2. 5.14.7 Struvite composition for presence of heavy metals Struvite pellets were also analyzed, using inductively coupled plasma – mass spectrometry (ICP-MS), to determine the presence/absence of metals.  The quantification of the metals is very important because it will ultimately determine if the product can be used as a fertilizer. The concentrations of some of the major metals of concern are given in Table 5.8. Detailed analytical composition is provided in Appendix K. Comparing the metal contents in the struvite formed in the present study to those found in literature, it can be seen that the those harvested at the LIWWTP is much richer (having higher percentages of constituent elements) in terms of quality. Therefore, from a heavy metal content perspective, the use of struvite formed at this treatment plant may not pose any health concerns, or at least have lower concerns than fertilizer formed from phosphate ore. However, it must be mentioned that the quality of struvite in Table 5.8 may not be representative of struvite grown at other treatment plants; similar analyses need to be conducted to determine suitability of the product for use as a local fertilizer. As mentioned above, calcium phosphate compounds are most likely  to  be  present,  in  addition  to  struvite,  in  a  pellet.  The  data  show  that,  although  little Chapter Five: Results and Discussions 132  (0.2%) calcium is present in the struvite pellet formed, part or all of calcium may be present as phosphate compounds. Table 5.8. Concentration of heavy metals found in struvite pellets formed at LIWWTP and in nature Element Present studya (ppm) Moroccob P-rocks (ppm) Geestmerambachtb Ca3(PO4)2 (ppm) Aluminum (Al) < 0.01 (as %) 200 950 Arsenic (As) 0.3 5 2 Cadmium (Cd) < 0.01 40 <6 Calcium (Ca) 0.2 (as %) Copper (Cu) 5.3 23 17 Chromium (Cr) 4.2 357 8 Iron (Fe) < 0.04 (as %) 1600 1260 Lead (Pb) 0.6 Lithium (Li) 0.4 Manganese (Mn) 236 10 560 Mercury  < 0.01 Nickel (Ni) 0.32 67 8 Sodium (Na) 0.01 (as %) 1700 360 Tin (Ti) 9.1 108 8 Uranium  < 0.05 Zinc (Zn) 4 880 310 a Value indicates the average of six samples b values given in Jeanmaire (2001). 5.15 Influence of Parameters on Physical Structure of Struvite Pellets In crystallization kinetics, two phases are generally considered – nucleation and growth. The increase in pellet size is a result of growth occurring due to the assimilation of ions in the lattice structure, established by the crystal embryo foundation (Ohlinger et al. 1999). Mass Chapter Five: Results and Discussions 133  transport  from  solution  to  the  crystal  surface  and  incorporation  of  material  into  a  crystal lattice are two processes that dictate the growth of crystals. Among the various parameters that influence the particle size in a fluidized bed reactor, SSR, magnesium concentration in the reactor and fluid dynamics are among the most important (Fattah et al. 2008b) and will be discussed further. 5.15.1 Possible influence of crystallizer supersaturation ratio In a parallel study, Bhuiyan et al. (2007) suggested that, at high supersaturation ratios, the growth of smaller crystals was faster than that of larger ones. In addition to SSR, the growth of larger pellets is also strongly influenced by the upflow (fluidization) velocity, as discussed below. During the present study, it was determined visually that low reactor SSR (around 2.0) did not significantly reduce the size (around 3-5 mm) of pellets formed, as long as they were fluidized and the upflow velocity was in the range that provided collision-adherence of the particles, but not breakage.  Figure 5.37 illustrates a typical SSR profiling of UBC’s second generation crystallizer. The difference between this crystallizer and the system used in the present study is that the former has six sections as opposed to five sections. Since the headspace in the pilot-scale setup in the present study was very low, it was not possible to sample inside the crystallizer and  develop  a  profile  of  the  SSR.  However,  the  sections  of  the  two  reactors  were approximately the same, and hence the data were used in this section to provide an explanation of the SSR impact on size and shape of struvite pellets formed. The objective of this section was to show, qualitatively, how the sectioning of the crystallizer used in the present study could be related to the growth of struvite pellets.   In the present study, the initial or starting SSR was higher in the injection zone than in the seed hopper (measured from effluent characteristics). Therefore, it can be assumed that the SSR decrease, as the water flows from the bottom to the top of the crystallizer. Therefore, it is reasonable to assume that, in the bottom section of the crystallizer, which has higher SSR and upflow velocity, tiny crystals are formed that are whisked away almost immediately to the top of the crystallizer; here the SSR and upflow velocity are lower and allow Chapter Five: Results and Discussions 134  agglomeration and growth of the pellets. Figure 5.38 illustrates the possible mechanism of pellet growth – from individual crystals that become fused, to being agglomerated into larger pellets with a smooth surface. Struvite pellets grown in the crystallizer with instrumentation and control (R #2) at the pilot-scale were measured for size with a scale and digital caliper. As illustrated in Figure 5.39, pellets as large as 6mm and above were consistently found.  0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0 1 2 3 4 5 SS R Crystallizer height (m)  Dimensions at the top indicate the crystallizer section diameter Figure 5.37. Typical SSR profiling in UBC’s crystallizer with crystallizer height (adapted from Forrest, 2004).        51 mm          76 mm         102 mm            152mm             380 mm Chapter Five: Results and Discussions 135   (a) (b)  (c) (d)  Figure 5.38. Formation of struvite pellets: individual crystals at (a) 2000 X magnification and (b) 500 X magnification, (c) agglomeration of crystals at 350 X magnification, (d) fused and smooth outer surface of pellet at 100 X magnification.  Chapter Five: Results and Discussions 136     Figure 5.39. Struvite pellets harvested at LIWWTP in crystallizer with instrumentation. 5.15.2 Upflow velocity Omar and Ulrich (2003) suggested that fluid dynamics has a great effect on the growth of particles in a fluidized bed reactor, with higher velocities resulting in higher mass transfer and consequently, bigger pellet size. A previous study (Fattah et al. 2008b) determined that an upflow velocity of about 400 cm/min was high enough to remove and recover phosphate, with a high degree of efficiency. During the course of the present study, the reactors were operated at a steady 400 cm/min upflow velocity. This value was increased to 500 cm/min to determine if upflow velocity had any effect on the quality of struvite pellets formed and to increase the treatment capacity of the reactors. Samples collected during the increased flow conditions were found to have broken structures – parts were missing from the parent pellet Chapter Five: Results and Discussions 137  (Figure 5.40). This could imply that agglomeration of crystals, resulting from collisions between particles, was overcome by destructive, colliding forces due to the higher upflow velocity.  Therefore,  there  is  a  maximum  upflow  velocity  that  a  crystallizer  should  be operated, in order to have good quality harvest. Upflow velocity also plays an important role in the operation of the process as it determines the mass loading of the reactor. As such, this operating limit should be confirmed at each full-scale installation.    Figure 5.40. Breakage of struvite pellets possibly due to high upflow velocity. 5.15.3 Flow patterns in the crystallizer The development of struvite pellets can be described as a two-phase phenomenon – the formation  of  pure  struvite  crystals  (nucleation)  and  the  agglomeration,  or  growth,  of  these crystals through physical collision and particle bridging (Ohlinger et al. 1999). Previous pellets grown with the same crystallizer were mostly spherical in shape. However, during one episode of the operation, when part of the reactor was clogged due to struvite accumulation Chapter Five: Results and Discussions 138  (both  by  pellets  and  formation  of  a  layer  of  struvite  on  the  sides  of  the  reactor),  elongated pellets were formed (Figure 5.41a). This was attributed to the variation of hydrodynamics within the reactor and the existence of dead zones. The fluid hydrodynamics in such a system is important because it determines the local solids and ionic concentrations, mixing energies and the rate and intensity of particle collisions. Struvite pellets grown in Ostara’s (Ostara Nutrient, Ltd.) reactors tend to be less spherical in nature (Figure 5.41b). One of the reasons for this may be the higher diameter to height ratio, producing a flow pattern that does not allow sufficient collisions and grinding/polishing of pellets. However, this hypothesis could not be verified in the present study. Chapter Five: Results and Discussions 139     (a)  (b) Figure 5.41. (a) Elongated pellets possibly formed due to flow restrictions in the crystallizer and (b) pellets from Ostara’s reactor. Chapter Five: Results and Discussions 140  5.16 Influence of Parameters on Crushing Strength of Struvite Pellets Variations in the crushing strength of struvite pellets (Table 5.19 and 5.10) grown in the crystallizer show that they are strongly influenced by process parameters. In the present study, although detailed experiments were not possible to determine the exact impact of these parameters on the crushing strength, based on the observations and analytical data obtained during the pilot-scale operation of the two crystallizers, some probable relationships have been suggested. These relationships provide the background for further studies. Variations of size, shape and orientation of flaws, such as pores and defects, may result in differences in the crushing strengths of harvested pellets. The formations of fine crystals that fuse together, to form a tight internal structure, were probably what provided the high strength of struvite pellets formed in the present study. Although mechanical compaction can increase crushing strength,  it  can  result  in  the  formation  of  smaller  pellet  sizes.  The  following  discussion  is based on struvite formed in both crystallizers. 5.16.1 Influence of size on strength Figure 5.42 illustrates the relationship between size of pellets and crushing strength. The crushing strength data are based on strengths obtained from crushing 100 or more pellets of each category. The strength initially increased with increase in the size of the pellets (till 2.5 mm); pellets in the 2.0-2.5 mm range had the greatest strengths. In a compression study using nitrogen-phosphorus-potassium (NPK) fertilizers, Walker et al., (1997) found that, in the size range (2 mm to 4 mm) tested, the mechanical crush strength of granular fertilizer was linearly related to the increase in size of fertilizer granules. However, in the present study, beyond the 2.5  mm  pellet  size,  the  strength  showed  a  decreasing  pattern.  This  could  be  related  to  the lower denser structure found for larger pellets than for smaller ones. The consequence of this information is that, when larger and stronger pellets are required, it may be necessary to harden the outside of the pellet with other coating material, or soak the pellets in a liquid that will fill the inner pores of the pellets and provide higher crushing strength. The numerical crushing strength value also provides valuable information that can be used when the struvite is to be machine spread as fertilizer.  Chapter Five: Results and Discussions 141   0 500 1000 1500 2000 2500 3000 3500 4000 1.00 1.98 2.36 3.33 4.70 C ru sh in g st re ng th  (g ) Pellet size (mm) Sample 1 Sample 2 Sample 3 Sample 4  Figure 5.42. Influence of pellet size on crushing strength. Error bars: 95% confidence interval. 5.16.2 Influence of struvite composition on strength Several samples of struvite were tested for their crushing strength and analyzed for composition. As illustrated in Figure 5.43, the Mg:P molar ratio was not 1:1 at the highest strength, indicating that the strongest struvite pellet may contain material that is not struvite in nature, but one that has a higher crushing strength. However, it was not possible to isolate the compound(s) that may lead  to  the  stronger  pellets.  Similar  to  the  Mg:P  molar  ratio,  the  strongest  struvite  pellets were not 100% pure, with respect to the N:P molar ratio (Figure 5.44). However, the ratio is higher than the Mg:P molar ratio. Detailed analytical data are provided in Appendix L.  Chapter Five: Results and Discussions 142  0 500 1000 1500 2000 2500 3000 3500 4000 0.800 0.850 0.900 0.950 1.000 C ru sh in g st re ng th  (g ) Mg:P molar ratio  Figure 5.43. Relationship between Mg:P molar ratio and the crushing strength of struvite pellet. Error bars: 95% confidence interval.  Chapter Five: Results and Discussions 143  0 500 1000 1500 2000 2500 3000 3500 4000 0.8 0.85 0.9 0.95 1 1.05 C ru sh in g st re ng th  (g ) N:P molar ratio  Figure 5.44. Relationship between N:P molar ratio and the crushing strength of struvite pellet. Error bars: 95% confidence interval. 5.16.3 Influence of SSR on strength Various studies (Fattah et al. 2008b; Forrest et al. 2008a; Britton et al. 2005; Adnan et al. 2003; Ohlinger et  al. 1998) have indicated the use of SSR as a control parameter in the removal- recovery of phosphorus from wastewater. However, there is little information on how this parameter affects the composition and quality of the product. In the range (1.5 – 7.1) covered in the present study, SSR did not appear to affect the crushing strength of struvite pellets. It is probable that the struvite pellets grow according to the SSR (as mentioned previously) and once formed, they maintain their integrity, as long as the SSR is above unity. Table 5.9 provides some evidence to support this hypothesis.  The SSR value given in Table 5.9 was the actual value in the harvest section of the crystallizer (R#2) when the pellets, whose crushing strengths are given, were harvested. Although SSR does not have a huge impact on strength, it is critical to control the parameter in the crystallizer during the initial agglomeration process. Also, it is essential to maintain a profile similar to Figure 5.37 Chapter Five: Results and Discussions 144  to produce a good quality product – with a high SSR in the bottom of the reactor, which then decreases in the top sections. Table 5.9. Influence of SSR on crushing strength SSR Crushing Strength (g) 1.46  2460  ± 120 1.62  2640  ± 90 3.36  2660 ± 160 7.11  2340  ± 80  5.16.4 Influence of reactor magnesium concentration on strength Although previous studies (Fattah et  al. 2008b; Mavinic et al. 2007) hypothesized that the  magnesium  content  in  the  crystallizer  influences  the  quality  of  struvite  pellets,  no  data were provided to support the theory. In the present study, magnesium concentrations in the crystallizer were correlated to the breaking strength of struvite (Table 5.10). An increase of 15% – 27% in crushing strengths were determined in the present study due to an increase of 7% (22.5 to 24.1 mg/L) and 132% (19 – 44 mg/L) of excess magnesium concentration in the crystallizer, respectively Based on the data, it is possible that higher Mg concentrations may have been responsible for higher strengths. The hypothesis that magnesium concentration can influence crushing strength is backed by the fertilizer industry, which uses, among other things, magnesium sulphate to granulize mineral fertilizer pellets in order to develop larger and  stronger  material.  It  is  important  to  note  that  the  upflow  velocities,  when  the  samples were taken, were similar, to negate its influence. The average magnesium concentration in Table 5.10 represents the expected Mg concentration in the harvest zone of the crystallizer, based on the average of two days’ concentration before the pellets were harvested. This value was calculated to allow for the fact that the pellets may not have grown in a day, and may have been exposed to this concentration before being harvested.    Chapter Five: Results and Discussions 145  Table 5.10. Influence of reactor magnesium concentration on crushing strength of struvite  Average* Mg concentration (mg/L) Mg concentration on day of harvest (mg/L) Strength (g) 22.5 11.7 2145 ± 55 System 1 (R #1; without instrumentation and control) 24.1 18.1 2460 ± 120 19.0 16.7 2637 ± 90 System 2 (R # 2; with instrumentation and control) 44.0 43.7 3345 ± 124 * Preceding two days before harvest  Error at 95% confidence interval. 5.16.5 Shell formation and two peak strength As mentioned previously, struvite with an outer shell also formed in this study. These pellets were tested for crushing strength and, unlike single pellets (Figure 5.45 a and b), these pellets did not have a distinct strength peak (Figure 5.45 c and d). The first peak corresponds to the breakage of the outer shell, while the second peak corresponds to the inner pellet. Given the higher crushing strength and compactness of the outer shell, the major implication of this pellet strength is that, if used with a mechanical spreader, there may be less loss of the material. Concurrently, when placed in a water matrix, the inner part, which had a lower dense structure and had lower crushing strength, may dissolve faster. 5.16.6 Conclusions on struvite quality The composition of pellets is important and needs to be verified before being applied as a fertilizer. The product formed at LIWWTP had good struvite purity (in terms of magnesium content) and contained much lower heavy metals concentration than those found in ores used as a source of phosphate. The physical quality of the pellets formed is important for a number of reasons, such as, ease of transportability and applicability, lower loss of material and release rates. Pellets grown in the present study were as large as 6 mm; this size is among the largest formed from domestic wastewater.  Supersaturation ratio and upflow velocity in the Chapter Five: Results and Discussions 146  crystallizer were found to influence the size and shape of the pellets developed. Crushing strengths of struvite pellets were correlated to the size and composition of struvite, the reactor magnesium concentration and supersaturation ratio.  Chapter Five: Results and Discussions 147  0 500 1000 1500 2000 2500 3000 3500 4000 0 500 1000 1500 2000 C ru sh in g st re ng th  (g ) Sample Number  0 200 400 600 800 1000 1200 1400 0 500 1000 1500 2000 Cr us hi ng  st re ng th  (g ) Sample Number  (a) (b) 0 200 400 600 800 1000 1200 1400 0 500 1000 1500 2000 Cr us hin g s tre ng th (g ) Sample Number  0 200 400 600 800 1000 1200 0 500 1000 1500 2000 Cr us hi ng  st re ng th  (g ) Sample Number  (c) (d) Figure 5.45. Crushing strength graphs showing formation of single peaks {(a) and (b)} for normal pellets and dual peaks for pellets with shell-structure {(c) and (d)}. Chapter Six: Conclusions 148  6 CHAPTER SIX:  CONCLUSIONS Based on experience in operating two struvite crystallization processes at the pilot-scale setup at Lulu Island Wastewater Treatment Plant, one with manual control and another with automated control (using the controlled program developed in the present study), and process data results, it was determined that the hypothesis on which the present study was based on was valid. As detailed in Chapter Five, and summarized below, changes in the process variables were detected, and actions were easily and timely taken in the controlled struvite crystallizer, which lead to the production of struvite of higher quality and decreased operation  time.  Supplementary  studies  carried  out  also  proved  to  expand the  knowledge  of struvite crystallization process.  Based on the results from experiments carried out at both laboratory- and pilot-scale, and development of a control technology, the following conclusions can be drawn. 6.1 Development of Control Programs x Laboratory and pilot-scale operation showed that the control program, developed to keep the SSR of a struvite crystallization system at the desired value was efficient, and could take into account various changes in process variables, such as ortho- phosphate concentration, temperature and conductivity. x The automated crystallizer system (R#2) that had instrumentation and was controlled by the control programs developed in the present study exhibited better performance than the manual (uncontrolled) crystallizer system (R#1) in reducing the phosphate concentration. R#2 had an increase, with respect to R#1, of 40% and 15% in removing ortho-P and ammonium, respectively. x In terms of purity (% struvite in pellets), compared to R#1, R#2 pellets were slightly higher (by 2%) in struvite content, and exhibited 13% higher average crushing strength. Chapter Six: Conclusions 149  x Analyses of the composition of struvite pellets formed at the Lulu Island Wastewater Treatment Plant showed that the pellets were never 100% pure with respect to struvite, as they contained different metal salts. However, the heavy metals content in the pellets were found to be much lower than reported in literature for phosphate ores. x Struvite pellets harvested were, on average, 90% pure (based on magnesium content in the pellets). x The models programmed in the present study can be used as a predictive tool for treatment plant operators to determine struvite formation potential due to changes in the process variables. x A 5°C change in temperature can bring about an increase of more than 90% in the SSR value. The variation is enhanced at lower temperatures, where the SSR nearly doubles (between 15-20°C). x The variation of SSR is more pronounced with limiting parameters than those present in excess. For example, at the lower concentration region (between 5-10 mg/L Mg), the SSR nearly doubles, whereas the rate of change is 50% at higher concentrations. x Various graphical-user-interfaces developed provided visual and easier control of the struvite crystallization process. 6.2 Solubility Tests x Within the working pH range (7.0-8.5) of a struvite crystallization process, the struvite solubility product, pKsp,  was determined to be independent of the working fluid matrix and pH. x The struvite solubility product at 25°C was determined to be 13.26. x An equation relating pKsp to temperature was developed. 6.3 Carbon Dioxide Stripping Tests x Over 90% of phosphorus was easily removed from the centrate at a pH of 8.1, with or without air stripping. Chapter Six: Conclusions 150  x With air and without air stripping, an average of 32% and 26% caustic savings was achieved using the cascade stripper, compared to the no-stripper condition, respectively. x Both the compact media and cascade strippers were equally efficient (average 90% phosphate removal) in their application. However, due to ease of operation (lower plugging and easier cleaning), the cascade stripper proved to be more suitable. x The cost of operating a struvite crystallization process could be significantly reduced by increasing the pH of the system through carbon dioxide stripping. The potential saving in caustic cost, due to CO2 stripping, was calculated to be as high as 38 cents per thousand liters treated. 6.4 Magnesium Prediction Techniques 6.4.1 Conductivity-pH measurements x For a struvite crystallization system, the combined use of pH and conductivity can be an easy and quick method to determine the amount of external magnesium that need to  be  added  to  a  water  sample.  The  amount  required  was  determined  by  locating  a bending point in the pH-conductivity-external magnesium added graph. x At the bending point, for each mole of phosphate removed, the Mg:P molar ratio was 1.3-2.0.  This  ratio  is  important,  because,  although  it  does  not  signal  the  end  of  the reaction, it determines the quantity of magnesium chloride required to remove one mole of phosphate. x Since the activity of an ion is responsible for actual reaction, and Mg determination by AA provides total concentration, the pH-conductivity method of determining magnesium addition provides a better estimate of the dosages required.  Chapter Six: Conclusions 151  6.4.2 Determination of magnesium concentration by hardness test method x Since phosphate interferes with hardness tests, removal of the ion by both PAC and alum was suitable in reducing phosphate concentrations (to lower than 20 mg/L) from the water matrix, prior to the hardness tests. x Tests showed that there was a linear relationship between the amount of alum added and phosphate reduction. Although the relationship (rate of change) was different as the water matrices varied, by developing an equation for each matrix, the amount of alum required can be predicted. x Determination of magnesium concentration by the hardness test method was found to be suitable for quick, on-site testing. The absolute average error between results from the hardness test and AA was only 2.0 mg/L. 6.5 Influence of Variables on Structure of Struvite Pellets x Pellets grown in this study were as large as 6 mm; this size is among the largest formed from domestic wastewater.  Upflow velocity in the crystallizer was found to influence the size and shape of the pellets. x The compactness of struvite pellets is a function of size of the pellets and the upflow velocity. At an upflow velocity of above 400 cm/min, the surface smoothness and sphericity of the struvite pellets formed were lowered. Higher upflow velocity can possibly lead to the formation of denser, but smaller and less spherical pellets. x Flow patterns in the crystallizer possibly influences the shape of the struvite pellets formed. 6.6 Determination of Crushing Strength x Mid-sized pellets, in the 2.0-2.5 mm range, exhibited the highest crushing strengths. This could be related to the denser structural composition found for smaller pellets than larger ones. x Within the range (SSR of 1.5 – 7.1) tested,  supersaturation ratio did not appear to influence the crushing strength of pellets formed. Chapter Six: Conclusions 152  x A high concentration (above 20 mg/L) of unused magnesium in the crystallizer can lead to the formation of pellets having greater crushing strengths. An increase of 15% – 27% in crushing strengths were determined in the present study due to an increase of 7% (22.5 to 24.1 mg/L) and 132% (19 – 44 mg/L) of excess magnesium concentration in the crystallizer, respectively. However, it should be remembered that there is a limit to how much excess magnesium one should have in the crystallizer. High concentration of unused magnesium can lead to higher concentrations in the system’s effluent, which can then increase struvite formation potential downstream of the treatment. In addition, inefficient use of magnesium leads to an increase in operating costs.  To  summarize,  the  overall  major  contributions  of  the  present  study  to  the  field  of phosphorus removal-recovery as struvite were:  y The  development  of  a  process  controller  and  graphical  tools  for  use  in  operating  a struvite crystallization process, and predicting struvite formation potential in wastewater treatment plants. y The determination of a universal solubility product for struvite and the influence of temperature on the value. y The applicability of a carbon dioxide stripper in reducing chemical costs, and development of program to determine operating stripping efficiency in increasing the pH of the system. y The development of a protocol for quick and on-site measurement of external magnesium dosing requirements in struvite crystallization process. y The development of a device that can be used as a quick means of measuring struvite pellet crushing strength.  Chapter Seven: Recommendations for Further Study 153  CHAPTER SEVEN: RECOMMENDATIONS FOR FURTHER STUDY x The prediction and equation developed in this study for struvite solubility product should be tested further with water matrices from more treatment plants, and should cover higher temperature ranges. x The applicability of using gas stripping as a means of increasing the pH of the water matrix in a struvite crystallization process has been confirmed. The carbon dioxide stripping model developed should be tested at other pilot-scale setups, to increase its prediction power. x The impact of operating pH on the degree and efficiency of carbon dioxide stripping should be tested. x The presence of ammonium/ammonia in wastewater treatment plant effluent can increase the toxic effect of the gas in the water body into which it is discharged. Although struvite precipitation alone can reduce ammonium concentrations in the centrate/supernatant, the percentage removal is small (10-20%). Stripping of ammonia gas is another viable option to reduce the concentration. Therefore, the applicability of ammonia stripping, using the carbon dioxide cascade stripper, should be investigated. x The application of indirect magnesium addition requirement (by using pH and conductivity) has been proposed and found to be useful in the laboratory in this study. The efficiency, and deficiencies, of this method should be tested at a larger pilot scale facility. x Since  the  lack  of  instrumentation  in  this  study  could  not  utilize  the  full  potential  of the control program developed, it should be tested in a fully instrumented, full-scale struvite crystallization process. x Tests should be carried out to determine the types of compounds, other than struvite, that are present in a struvite pellet, and conditions that increase their formation potential. Chapter Seven: Recommendations for Further Study 154  x Agglomeration studies, including formulation for struvite growth, should be investigated and incorporated into the model developed in this study. x The application of the struvite crystallization process mentioned in this study should be investigated using dairy wastewater. At UBC, the application of peroxide and microwave (PM) to enhance solubility of phosphates and nitrogen from dairy wastewater has been studied. A treatment scheme that integrates PM technology with the crystallization process should be studied at a pilot scale. References 155  REFERENCES Aage, H.K., Anderson, B.L., Blom, A. and Jensen, I. (1997). The solubility of struvite. J Radioanal Nucl. Chem. 223, 213–215.  Adnan, A., Mavinic, D.S. and Koch, F.A. (2003). Pilot-scale study of phosphorus recovery through struvite crystallization – examining the process feasibility. J. Environ. Eng. Sci. 2, 315-324.  Ali,  M.I.  (2005).  Struvite  crystallization  from  a  nutrient  rich  wastewater.  Ph.D.  Thesis, Department of Chemical Engineering. Townsville, James Cook University, Australia.  APHA, AWWA, WPCF (1995). Standard Methods for the Examination of Water and Wastewater, 19th Edition. American Public Health Association, Washington, D.C.  Babiü-IvanþLü,  V.,  Kontrec,  J.Kralj,  D.  and  Breþeivü,  L.  (2002).  Precipitation  diagrams  of struvite and dissolution kinetics of different struvite morphologies. Croatica Chemica Acta, 75 (1), 89-106.  Barnard, J.L. (2009) Elimination of eutrophication through resource recovery. International conference on nutrient recovery from wastewater streams. Vancouver, BC, Canada. IWA publishing, UK.  Battistoni, P., Fava, G., Pavan, P., Musacco, A. and  Cecchi, F. (1997). Phosphate removal in anaerobic liquors by struvite crystallization without addition of chemicals: preliminary results. Water Research. 31 (11), 2925-2929.  Berg, E. (1982). Handbook for Sampling and Sample Preservation for Water and Wastewater. United States Environmental Monitoring and Support Laboratory, USEPA, Cincinnati, Ohio, U.S.A. References 156  Bhuiyan, M.I.H. and Mavinic, D.S. (2009). Temperature dependence on electrical conductivity and its relationship with ionic strength for struvite precipitation system. International conference on nutrient recovery from wastewater streams. Vancouver, BC, Canada. IWA publishing, UK.  Bhuiyan, M.I.H., Mavinic, D.S. and Beckie, R.D. (2008). Nucleation and growth kinetics of struvite in a fluidized bed reactor. J. Crystal Growth. 210, 1187-1194.  Bhuiyan, M. I. H., Mavinic, D. S. and Beckie, R. D. (2007). A solubility and thermodynamic study of struvite. Environ Technol. 28(9), 1015-1026.  Bhuiyan, M. I. H., Mavinic, D. S. and Beckie, R. D. (2009). Determination of Temperature Dependence of Electrical Conductivity and Its Relationship with Ionic Strength of Anaerobic Digester Supernatant, for Struvite Formation. Journal of Environmental Engineering-ASCE. 135(11), 1221-1226.  Booker, N. A., Priestley, A. J. and Fraser, I. H. (1999). Struvite formation in wastewater treatment plants: Opportunities for nutrient recovery. Environ Technol. 20(7), 777-782.  Bouropoulos, N. C. and Koutsoukos, P. G. (2000). Spontaneous precipitation of struvite from aqueous solutions. J Cryst Growth. 213(3-4), 381-388.  Britton, A., Koch, F. A., Mavinic, D. S., Adnan, A., Oldham, W. K. and Udala, B. (2005). Pilot-scale struvite recovery from anaerobic digester supernatant at an enhanced biological phosphorus removal wastewater treatment plant. Journal of Environmental Engineering and Science. 4(4), 265-277.  Buchanan, J.R., Mote, C.R. and Robinson, R.B. (1994). Thermodynamics of struvite formations. Transactions of the ASEA. 37 (2), 617-621.  References 157  Burns, J.R. and Finlayson, B. (1982). Solubility products of ammonium magnesium phosphate hexahydrate at various temperatures. The Journal of Urology. 128, 426-428.  CEEP (2001). Phosphate Recovery: Where do we stand today? Special issue of the Scope Newsletter, published in preparation to the 2nd International Conference on P-recovery from human and animal wastes. 12-14 march, 2001. Noordwijkkerhout, The Netherlands.  Celen, I. and Turker, M. (2001). Recovery of ammonia as struvite from anaerobic digester effluents. Environmental Technology. 22, 1263-1272.  Chambers, P.A., Guy, M., Roberts, E.S., Charlton, M.N., Kent, R., Gagnon, C., Grove, G. and Foster, N. (2001). Nutrients and their impact on the Canadian Environment. Agriculture and Agri-Food Canada, Environment Canada, Fisheries and Oceans Canada, Health Canada and Natural Resources Canada. pp 241.  Childs, C. W. (1970). A potentiometric study of equilibria in aqueous divalent metal orthophosphate solutions. Inorg. Chem. 9(11), 2465–2469.  Cohen, Y. and Kirchmann, H. (2004). Increasing the pH of wastewater to high levels with different gases-Co2 stripping. Water Air and Soil Pollution. 159(1-4), 265-275.  Cordell, D., Schmid-Neset, T., White, S. and Drangert, J.O. (2009). Preferred future phosphorus scenarios: a framework for meeting long-term phosphorus needs for global food demand. International conference on nutrient recovery from wastewater streams. Vancouver, BC, Canada. IWA publishing, UK.  Demeestere,  K.,  Smet,  E.,  Van  Langenhove,  H.  and  Galbacs,  Z.  (2001).  Optimalisation  of magnesium  ammonium  phosphate  precipitation  and  its  applicability  to  the  removal  of ammonium. Environ Technol. 22(12), 1419-1428.  References 158  Dockhorn, T. (2009). About the economy of phosphorus recovery. International conference on nutrient recovery from wastewater streams. Vancouver, BC, Canada. IWA publishing, UK.  doro, R., Bolzonella, D. and Pezzoli, S. (2001). Optimization of chemical and physical pretreatments in a platform for the treatment of liquid industrial wastes. Ind Eng Chem Res. 40(21), 4506-4512.  Doyle, J. D. and Parsons, S. A. (2002). Struvite formation, control and recovery. Water Res. 36(16), 3925-3940.  Doyle, J. D., Philp, R., Churchley, J. and Parsons, S. A. (2000). Analysis of struvite precipitation in real and synthetic liquors. Process Saf Environ Prot. 78(B6), 480-488.  Driver, J., Lijmbach, D. and Steen, I. (1999). Why recover phosphorus for recycling, and how? Environ Technol. 20(7), 651-662.  Durrant,  A.  E.,  Scrimshaw,  M.  D.,  Stratful,  I.  and  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(7), 749-758.  Fattah,  K.P.,  Mavinic,  D.S.,  Rahaman,  M.S.  and  Koch,  F.A.  (2009).  Development  of  a process control system for online monitoring and control of a struvite crystallization process. Proceeding of International Conference on Nutrient Recovery from Wastewater Streams, Vancouver, B.C., May 10-13, 2009.  Fattah, K. P., Zhang, Y., Mavinic, D. S. and Koch, F. A. (2008a). Application of carbon dioxide stripping for struvite crystallization - I: Development of a carbon dioxide stripper model to predict CO2 removal and pH changes. Journal of Environmental Engineering and Science. 7(4), 345-356.  References 159   Fattah, K. P., Mavinic ,D. S., Koch, F. A. and Jacob, C. (2008b). Determining the feasibility of phosphorus recovery as struvite from filter press centrate in a secondary wastewater treatment plant. Journal of Environmental Science and Health Part A-Toxic/hazardous Substances and Environmental Engineering. 43(7), 756-764.  Fattah,  K.P.  (2004).  Pilot  scale  struvite  recovery  potential  from  centrate  at  Lulu  Island Wastewater Treatment Plant. M.A.Sc. Thesis, Department of Civil Engineering, University of British Columbia, Vancouver, BC, Canada.  Forrest,  A.  L.,  Fattah,  K.  P.,  Mavinic,  D.  S.  and  Koch,  F.  A.  (2008a).  Optimizing  struvite production for phosphate recovery in WWTP. Journal of Environmental Engineering-Asce. 134(5), 395-402.  Forrest, A. L., Mavinic, D. S. and Koch, F. A. (2008b). The measurement of magnesium: A possible key to struvite production and process control. Environ Technol. 29(6), 603-612.  Forrest, A. L., Fattah, K. P., Mavinic, D. S. and Koch, F. A. (2007). Application of artificial neural networks to effluent phosphate prediction in struvite recovery. Journal of Environmental Engineering and Science. 6(6), 713-725.  Forrest, A. (2004). Process Optimization of a Technical Scale Phosphorus Recovery System Through Struvite Crystallization at the City of Penticton Advanced Wastewater Treatment Plant. Vancouver, BC: University of British Columbia, M.A.Sc. Thesis.  Gadekar, S., Pullammanappallil, P. and Varshovi, A. (2009). Validation of a comprehensive chemical equilibrium model for predicting struvite precipitation. International conference on nutrient recovery from wastewater streams. Vancouver, BC, Canada. IWA publishing, UK.    References 160  Gaterell, M. R., Gay, R., Wilson, R., Gochin, R. J. and Lester, J. N. (2000). An economic and environmental evaluation of the opportunities for substituting phosphorus recovered from wastewater treatment works in existing UK fertiliser markets. Environ Technol. 21(9), 1067- 1084.  Gilbert, N. (2009). The disappearing nutrient. Nature. 461, 8 October 2009.  Gunzel, D. and Schlue, W. R. (2002). Determination of [Mg-2+](i) - an update on the use of Mg2+-selective electrodes. Biometals. 15(3), 237-249.  Gupta, V. K., Chandra, S. and Mangla, R. (2002). Magnesium-selective electrodes. Sensors and Actuators B-Chemical. 86(2-3), 235-241.  Hanhoun, M., Azzaro-PBiscans, B., Freche, M., Montastruc, L., Pibouleau, L. and Domenech, S. (2009). A thermochemical approach for struvite precipitation modeling from wastewater. International conference on nutrient recovery from wastewater streams. Vancouver, BC, Canada. IWA publishing, UK.  Harada, H., Shimizu, Y., Miyagoshi, Y., Matsui, S., Matsuda, T. and Nagasaka, T. (2006). Predicting struvite formation for phosphorus recovery from human urine using an equilibrium model. Water Science and Technology. 54(8), 247-255.  Hirasawa,  I.,  Nakagawa,  H.,  Yosikawa,  O.  and  Itoh,  M.  (1997).  Phosphate  recovery  by reactive crystallization of magnesium ammonium phosphate: Application to wastewater. Separation and Purification by Crystallization. 667, 267-276.  Hiroyuki, H. and Toru, S. (2003). Nitrogen and Phosphorus recovery from a piggery waste water treatment facility. Proc. 2003 Third International Symposium on Environmentally Conscious Design and Inverse Manufacturing, Tokeyo, Japan, 422-426.  References 161  Jaffer, Y., Clark, T. A., Pearce, P. and Parsons, S. A. (2002). Potential phosphorus recovery by struvite formation. Water Res. 36(7), 1834-1842.  Jeanmaire, N. (2001). Recycling of removed phosphorus: Analysis of the potential interest in wastewater treatment plants. Centre European d’Etudes des Polyphosphates (CEEP). World Wide Web.     http://www.nhm.ac.uk/mineralogy/phos/Jeanmaire_report.doc  Jeanmaire, N. and T. Evans (2001). Technico-economic feasibility of P-Recovery from municipal wastewaters.  Environmental Technology. 22, 1355-1361.  Kumashiro, K., Ishiwatari, H. and Nawamura, Y. (2001). A pilot-scale study on using seawater as a magnesium source for struvite precipitation. In proceedings from Second International Conference on the recovery of phosphorus from sewage and animal wastes, Noodwijikerhout, Holland, March 2001.  Larkin, G.L. and Slaney, P.A. (1996). Trends in marine derived nutrient sources to south coastal British Columbia streams: impending implications to salmonid production. Watershed Restoration Management Report No. 3, B.C. Ministry of the Environment, Lands and Parks.  Le Corre, K. S., Valsami-Jones, E., Hobbs, P. and Parsons, S. A. (2007). Kinetics of struvite precipitation: Effect of the magnesium dose on induction times and precipitation rates. Environ Technol. 28(12), 1317-1324.  Le Corre, K. S., Valsami-Jones, E., Hobbs, P. and Parsons, S. A. (2005). Impact of calcium on struvite crystal size, shape and purity. J Cryst Growth. 283(3-4), 514-522.  Le Corre, K. S., Valsami-Jones, E., Hobbs P., Jefferson, B. and Parsons, S. A. (2007). Agglomeration of struvite crystals. Water Res. 41(2), 419-425.  References 162  Lee, G.F. (1970). Eutrophication. Prepared for the supplementation to the encyclopedia of chemical technology. John Wiley and Sons Inc. N.Y.  Liao, P.H., Chen, A. and Lo, K.V. (1995). Removal of nitrogen from swine manure wastewaters by ammonia stripping. Bioresource Technology. 54 (1), 17-20.  Liberti, L., Petruzzelli, D. and De Florio, L. (2001). REM NUT ion exchange plus struvite precipitation process. Environ Technol. 22(11), 1313-1324.  Lide, D. R. (1991). Handbook of chemistry and physics. 71st Edition. CRC Press.  Loewenthal, R. E., Kornmuller, U. R. C. and Vanheerden, E. P. (1994). Modeling struvite precipitation in anaerobic treatment systems. Water Science and Technology. 30(12), 107- 116.  Martell, A. E. and Smith, R. M. (1989) Critical Stability Constants. Plenum Press, New York  MathWorks website. The MathWorks, Inc. (1994-2006)       http://www.mathworks.com  Mavinic, D. S., Koch, F. A., Hall, E. R., Abraham, K. and Niedbala, D. (1998). Anaerobic co-digestion of combined sludges from a BNR wastewater treatment plant. Environ Technol. 19(1), 35-44.  Mavinic, D. S., Koch, F. A., Huang, H. and Lo, K. V. (2007). Phosphorus recovery from anaerobic digester supernatants using a pilot-scale struvite crystallization process. Journal of Environmental Engineering and Science. 6(5), 561-571.  Morel, F.M.M. (1983) Principles of Aquatic Chemistry. New York, NY: John Wiley and Sons.  References 163  Munch, E. V. and Barr, K. (2001). Controlled struvite crystallisation for removing phosphorus from anaerobic digester sidestreams. Water Res. 35(1), 151-159.  Musvoto, E. V., Ekama, G. A., Wentzel, M. C. and Loewenthal, R. (2000). Extension and application of the three-phase weak acid/base kinetic model to the aeration treatment of anaerobic digester liquors. Water Sa. 26(4), 417-438.  Neethling, J. B., Benisch, M. (2004). Struvite control through process and facility design as well as operation strategy. Water Science and Technology. 49(2), 191-199.  Ohlinger, K. N., Young, T. M. and Schroeder, E. D. (1999). Kinetics effects on preferential struvite accumulation in wastewater. Journal of Environmental Engineering-Asce. 125(8), 730-737.  Ohlinger, K. N., Young, T. M. and Schroeder, E. D. (1998). Predicting struvite formation in digestion. Water Res. 32(12), 3607-3614.  Omar, W. and Ulrich, J. (2003). Influence of crystallization conditions on the mechanism and rate of crystal growth of potassium sulphate. Crystal Research and Technology. 38(1), 34-41.  Owen, J., Stoven, H., Kowalski, J. and Phillips, K. (2009). Assessment of struvite containing controlled release fertilizer as a source of phosphorus for containerized ornamental crops. Hortiscience. 44 (4), 1131-1132.  Parkhurst, D.L. and Appelo, C.A.J. (1998). User's guide to PHREEQC (Version 2)--A computer program for speciation, batch-reaction, one-dimensional transport, and inverse geochemical calculations.…………………………………………………………………….. http://wwwbrr.cr.usgs.gov/projects/gwc_coupled/phreeqc/html/final.html  Pitman A. R., Deacon, S. L. and Alexander, W. V. (1991). The Thickening and Treatment of Sewage Sludges to Minimize Phosphorus Release. Water Res. 25(10), 1285-1294. References 164   Ponce,  R.  G.  and  De  Sa,  M.  E.  G.  L.  (2007).  Evaluation  of  struvite  as  a  fertilizer:  a comparison with traditional P sources. Agrochimica. 51(6), 301-308.  Rahaman, M. S., Mavinic, D. S., Bhuiyan, M. I. H. and Koch, F. A. (2006). Exploring the determination of struvite solubility product from analytical results. Environ Technol. 27(9), 951-961.  Rahman,  M.  J.,  Bozadjiev,  P.  and  Pelovski,  Y.  (1994).  Studies  on  the  Effects  of  some Additives on the Physicomechanical Properties of Urea Ammonium-Sulfate (Uas) Pellets. Fertilizer Research. 38(2), 89-93.  Sabrina, N. (2007) Comparing the effectiveness of CO2 strippers to reduce operating costs for struvite formation. M.A.Sc. Thesis, Department of Civil Engineering, The University of British Columbia Vancouver, B.C., Canada.  Sasai,  H.,  Kameyama,  K.  and  Sugimori,  N.  (1995).  An  innovative  option  for  nitrogen  and phosphorus recovery in sludge treatment process. In Proceedings of Water Environment Federation. WEFTEC’ 95.  Sawyer,  C.,  McCarty,  P  and  Parkin,  G.  (1994).  Chemistry for Environmental Engineering. McGraw-Hill Series in water Resources and Environmental Engineering, New York, U.S.A.  Seckler, M.M., Bruinsma, O.S.L. and Van Rosmalen, G.M. (1996). Calcium phosphate precipitation in a fluidized bed in relation to process conditions: a black box approach. Water Res. 30 (7), 1677-1685.  Shepherd, T. A., Burns, R. T., Moody, L. B., Raman, D. R. and Stalder, K. J. (2009). Investigating Conductivity to Predict Magnesium Addition Requirements for Struvite Precipitation in Swine Manure Slurries. Appl Eng Agric. 25(1), 103-108.  References 165  Shimamura, K., Tanaka, T., Miura, Y. and Ishikawa, H. (2003). Development of a high- efficiency phosphorus recovery method using a fluidized-bed crystallized phosphorus removal system. Water Science and Technology. 48(1), 163-170.  Sterling, M.S. and Ashley, K. I. (2003). Evaluations of a slow-release fertilizer for rehabilitating oligotrophic streams. American Fisheries Society Symposium. 34, 237-243.  Stratful, I., Brett S., Scrimshaw, M. D. and Lester, J. N. (1999). Biological phosphorus removal, its role in phosphorus recycling. Environ Technol. 20(7), 681-695.  Stratful I., Scrimshaw M. D. and Lester J. N. (2001). Conditions influencing the precipitation of magnesium ammonium phosphate. Water Res. 35(17), 4191-4199.  Slaney, T.L., Hyatt, K.D., Northcote, T.G. and Fielden, R.J. (1996). Status of anadromous salmon and trout in British Columbia and Yukon. Fisheries. 21, 20-35.  Snoeyink, V. and Jenkins, D. (1980). Water Chemistry. John Wiley and Sons, NY.  Taylor, A.W., Frazier, A.W. and Gurney, E.L. (1963). Solubility products of magnesium ammonium and magnesium potassium phosphates. Transactions of the Faraday Society. 59, 1580-1584.  Tchobanoglous, G., Burton, F.L. and Stensel, H.D. (2003). Wastewater engineering: treatment and reuse. McGraw-Hill, New York, U.S.A.  Thomas, M., vanderPoel, A. F. B. (1996). Physical quality of pelleted animal feed .1. Criteria for pellet quality. Anim Feed Sci Technol. 61(1-4), 89-112.  Ueno, Y and Fujii, M. (2001). Three years experience of operating and selling recovered struvite from full-scale plant. Environmental Technology.  22 (11), 1373-1381.  References 166  Vanotti, M. and Szogi, A. (2009). Technology for  recovery of phosphorus from animal wastewater through calcium phosphate precipitation. International conference on nutrient recovery from wastewater streams. Vancouver, BC, Canada. IWA publishing, UK.  von Horn, J. and Sartorius, C. (2009). Impact of supply and demand on the price development of phosphate (fertilizer). International conference on nutrient recovery from wastewater streams. Vancouver, BC, Canada. IWA publishing, UK.  Wang,  J.,  Burken,  J.  G.  and  Zhang,  X.  Q.  (2006).  Effect  of  seeding  materials  and  mixing strength on struvite precipitation. Water Environ Res. 78(2), 125-132.  Wierzbicki A., Sallis J. D., Stevens E. D., Smith M. and Sikes C. S. (1997). Crystal growth and molecular modeling studies of inhibition of struvite by phosphocitrate. Calcif Tissue Int. 61(3), 216-222.  Woods, N.C., Sock, S.M. and Daigger, G.T. (1999). Phosphorus recovery technology modeling and feasibility evaluation for municipal wastewater treatment plants. Environmental Technology. 20, 663-679.  Yoshino, M., Yao, M., Tsuno, H. and Somiya, I. (2003). Removal and recovery of phosphate and ammonium as struvite from supernatant in anaerobic digestion. Water Science and Technology. 48(1), 171-178.  Zhang, Y. (2006). Struvite crystallization from digester supernatant – reducing caustic chemical addition by CO2 stripping. M.A.Sc. Thesis, Department of Civil Engineering, University of British Columbia, Vancouver, BC, Canada. Appendices 167  APPENDICES Appendix A 168  APPENDIX A: SUMMARY OF PROGRAM CODE TO DETERMINE REQUIRED pH OF A STRUVITE CRYSTALLIZER SYSTEM (solvepH.m) Ionic species Considered. 1. Mg2+ 2. NH4+ 3. PO43- 4. MgOH+ 5. NH3 6. H3PO4 7. H2PO4- 8. HPO42- 9. MgPO4- 10. MgHPO4 11. MgH2PO4+  Equilibrium Constant H3PO4ĺ H+ + H2PO4 - K1 H2PO4 -ĺ H+ + HPO42 - K2 HPO42 -ĺ H+ + PO43 - K3 MgH2PO4+ĺ Mg2+ + H2PO4 - KMg MgHPO4ĺ Mg2+ + HPO42 - KMg1 MgPO4 -ĺ Mg2+ + PO43 - KMg2 NH4+ĺ H+ + NH3 KN H2O ĺ H+ + OH- Kw Appendix A 169   Z (n) = ionic charge of nth species A (n)  = activity coefficient of nth species AW (n) = atomic weight of the nth species  The following formulae for the coefficients have been derived based on thermodynamic equilibrium relationships between species. a = totalMg/AWMg/1000 * A(1)*A(4)*A(9)*A(11); b = A(4)*A(9)*A(11); c = Kw/KMgOH * A(1)*A(9)*A(11); d = 1/KMg3 * A(1)*A(4)*A(11); e = 1/K3/KMg2 * A(1)*A(4)*A(9)*A(11); f = 1/K2/K3/KMg1 * A(1)*A(4)*A(9); g = totalN/AWN/1000 * A(2); h = Kn * A(2); i = totalP/AWP/1000 * A(3)*A(7)*A(8)*A(9)*A(11); j = A(7)*A(8)*A(9)*A(11); k = 1/K1/K2/K3 * A(3)*A(7)*A(8)*A(9)*A(11); l = 1/K2/K3 * A(3)*A(8)*A(9)*A(11); m = 1/K3 * A(3)*A(7)*A(9)*A(11); n = 1/KMg3 * A(3)*A(7)*A(8)*A(11); o = 1/K3/KMg2 * A(3)*A(7)*A(8)*A(9)*A(11); p = 1/K2/K3/KMg1 * A(3)*A(7)*A(8)*A(9);   Quadratic Formula coefficients a_quad = ( (j + k*H^3 + l*H^2 + m*H) * (d + e*H + f*H^2) ); b_quad = ( (n + o*H + p*H^2)*a + (b + c/H)*(j + k*H^3 + l*H^2 + m*H) - (d + e*H + f*H^2)*i ); c_quad = ( -i * (b + c/H) );  Appendix A 170  The following equation solves the quadratic equation for activity of phosphorus based on equation derived above. Solves quadratic for activity of P P1 = (-b_quad + sqrt(b_quad^2 - 4*a_quad*c_quad)) / (2*a_quad); P1 = @(H) ( -( (n + o*H + p*H^2)*a + (b + c/H)*(j + k*H^3 + l*H^2 + m*H) - … (d + e*H + f*H^2)*i ) + sqrt( ( (n + o*H + p*H^2)*a + (b + c/H)*(j + k*H^3 + l*H^2 + m*H) - … (d + e*H + f*H^2)*i )^2 - 4*( (j + k*H^3 + l*H^2 + m*H) * (d + e*H + f*H^2) )*( -i * (b + c/H) ) ) ) .../ (2* ( (j + k*H^3 + l*H^2 + m*H) * (d + e*H + f*H^2) ) ); P2 = (-b_quad - sqrt(b_quad^2 - 4*a_quad*c_quad)) / (2*a_quad);  pH calculation; Takes the SSR as a function of H and solves for H  C1  = a/(b + c/H + (d + e*H + f*H^2)* C3); C2  = g / (1 + h/H ); C3  = P1;  Equation for struvite formation - MgNH4PO. 6H20 <> Mg2+ + NH4+ + PO43- + 6H2O SSR = {Mg2+}*{NH4+}*{PO43-} / KMAP  SSR * KMAP = MAPproduct ; MAPproduct = C1 * C2 * C3; Hfunction =  @(H)... (function of H)               ...% C1:               a/(b + c/H + (d + e*H + f*H^2) * ...                 ( -( (n + o*H + p*H^2)*a + (b + c/H)*(j + k*H^3 + l*H^2 + m*H) - (d + e*H + f*H^2)*i ) + ...                     sqrt( ( (n + o*H + p*H^2)*a + (b + c/H)*(j + k*H^3 + l*H^2 + m*H) - (d + e*H + f*H^2)*i )^2 - ...                         4*( (j + k*H^3 + l*H^2 + m*H) * (d + e*H + f*H^2) )*( -i * (b + c/H) ) ) ) ...                 / (2* ( (j + k*H^3 + l*H^2 + m*H) * (d + e*H + f*H^2) ) ) )...               ...% C2: Appendix A 171                * g / (1 + h/H ) ...               ...% C3:               * ( -( (n + o*H + p*H^2)*a + (b + c/H)*(j + k*H^3 + l*H^2 + m*H) - (d + e*H + f*H^2)*i ) + ...                     sqrt( ( (n + o*H + p*H^2)*a + (b + c/H)*(j + k*H^3 + l*H^2 + m*H) - (d + e*H + f*H^2)*i )^2 - ...                         4*( (j + k*H^3 + l*H^2 + m*H) * (d + e*H + f*H^2) )*( -i * (b + c/H) ) ) ) ...                 / (2* ( (j + k*H^3 + l*H^2 + m*H) * (d + e*H + f*H^2) ) ) ...               ...% SSR * KMAP:               - SSR * KMAP; Appendix B 172  APPENDIX B: VALUES OF PARAMETERS TO DETERMINE THEIR INFLUENCES ON THE SUPERSATURATION RATIO  Sample Number Influence of Cond. SSR Temp. PO4 NH4 Mg   mS/cm  °C mg/L mg/L mg/L 0 Temp. 3500 8.65 15 70 800 10 1800 Temp. 3500 8.65 15 70 800 10 3600 Temp. 3500 4.33 20 70 800 10 5400 Temp. 3500 4.33 20 70 800 10 7200 Temp. 3500 1.15 30 70 800 10 9000 Mg 3500 1.15 30 70 800 10 10800 Mg 3500 1.70 30 70 800 15 12600 Mg 3500 3.24 30 70 800 30 14400 PO4 3500 1.15 30 70 800 10 16200 PO4 3500 1.27 30 80 800 10 18000 PO4 3500 1.39 30 90 800 10 19800 PO4 3500 1.56 30 105 800 10 21600 PO4 3500 1.71 30 120 800 10 23400 PO4 3500 1.75 30 125 800 10 25200 PO4 3500 1.85 30 135 800 10 27000 PO4 3500 1.89 30 140 800 10 28800 PO4 3500 1.97 30 150 800 10 30600 PO4 3500 2.02 30 155 800 10 32400 PO4 3500 2.05 30 160 800 10 34200 NH4 3500 1.60 30 100 850 10 36000 NH4 3500 1.88 30 100 1000 10 37800 NH4 3500 2.16 30 100 1150 10 39600 NH4 3500 2.44 30 100 1300 10 41400 Cond. 3500 1.88 30 100 1000 10 43200 Cond. 4500 1.61 30 100 1000 10 45000 Cond. 3000 1.04 30 100 1000 10   Appendix C 173  APPENDIX C: EQUATIONS USED IN DERIVING STRIPPER MODEL Coefficient Expression A (6.72 Ln(BN) + 69.36)/76.74Note* B (-0.57(ERR)2 + 5.76(ERR) + 74.98)/74.98 C (-6.27(IFR) + 92.96)/78.53 D (-8×10-6 (ASR)3 + 0.003(ASR)2 – 0.168(ASR) + 64.64)/78.19 E (0.67 × (IT -18) + 59.22)/ 59.22 F (20.21 Ln([CO2]inf) – 67.14)/55.85 G (-8×10-10 (IBC)3 + 9×10-6(IBC)2 – 0.029(IBC) + 96.67)/73.07 Note*: the expression for coefficient A is obtained from (6.72 Ln(BN) + 69.36)/ (6.72 Ln(BNș) + 69.36), where BNԧ = 3. Equations for B~ G are obtained in the similar way.  Appendix D 174  APPENDIX D: PERFORMANCE OF CONTROLLER – EFFLUENT DATA Sample Number Reactor 1(without instrumentation and control)  Reactor 2 (with instrumentation and control)  PO4-P (mg/L) NH4-N (mg/L) Mg (mg/L)  PO4-P (mg/L) NH4-N (mg/L) Mg (mg/L) 1 26.9 387.5 85.7  47.2 432.5 10.1 2 30.2 195.0 11.7  24.0 184.5 7.2 3 35.1 381.0 47.9  28.3 281.0 39.3 4 34.4 321.0 57.0  39.1 344.0 12.5 5 13.4 54.6 72.3  9.7 49.3 54.0 6 18.0 54.1 14.7  15.0 50.1 58.0 7  291.0 57.6   296.0 43.5 8  281.5 64.5   244.5 41.7 9 33.8 331.5 53.2    39.1 11 27.5 466.0 49.0    32.8 12 15.6 416.5 17.4    18.4 14 8.2 110.5 31.9    28.6 15 14.2 292.0 21.8 16 12.8 117.5 33.7 18 13.6 133.0 32.2 19   33.7 20 15.0 218.5 26.8  14.0 199.0 21.6 21 10.5 333.0 34.5  12.8 310.0 25.5 22 9.1 274.5 41.6  5.2 273.0 27.3 23 12.2 504.0 40.9  25.3 440.0 42.8 24 34.6 395.0 44.3  6.3 262.5 30.3 25 13.3 400.5 27.8  8.0 396.5 15.2 26 61.6 552.5 6.1  50.4 529.5 0.9 Appendix D 175  Sample Number Reactor 1(without instrumentation and control)  Reactor 2 (with instrumentation and control)  PO4-P (mg/L) NH4-N (mg/L) Mg (mg/L)  PO4-P (mg/L) NH4-N (mg/L) Mg (mg/L) 27 53.2 656.5 3.8  9.8 546.5 6.5 28 27.3 730.5 8.1  9.2 693.5 6.7 29        13.0 727.0 23.6 30   720.5 2.2  18.3 795.0 6.1 31 37.7 779.0 18.1  22.2 746.5 10.4 32 42.1 796.5 10.4  26.6 787.5 4.3 33 34 6.8 705.0 14.1  8.3 713.5 10.3 35 34.5 705.50 10.9  18.2 655.5 21.5 36 9.4 687.0 5.9  10.5 708.0       9.1    Appendix E 176  APPENDIX E: SOLUBILITY TESTS Table E.1. Centrate at 15°C  Sample pH  NH4-N (mg/L) Average N mg/L PO4-P mg/L Average P mg/L Mg mg/L Average Mg mg/L Cond. ms/cm 1.00 1120.0 468.0 295.8 2.00 1100.0 465.0 295.9 S0 initial day 1 3.00 4.80 1090.0 1103.3 461.0 464.7 288.9 293.6 10.25 4.00 963.0 312.0 252.0 5.00 977.0 310.0 254.6 S1 6.00 5.95 1000.0 980.0 347.0 323.0 234.3 247.0 10.15 7.00 925.0 193.0 175.7 8.00 903.0 177.0 174.1 S2 9.00 6.26 938.0 922.0 166.0 178.7 173.4 174.4 9.85 10.00 875.0 88.1 113.0 11.00 885.0 83.5 113.2 S3 12.00 6.52 891.0 883.7 83.9 85.2 111.0 112.4 9.75 13.00 922.0 55.6 80.8 14.00 939.0 50.4 80.4 S4 15.00 6.73 877.0 912.7 47.9 51.3 79.2 80.2 9.65 16.00 876.0 33.6 67.6 17.00 899.0 33.9 67.6 S5 18.00 6.88 881.0 885.3 32.1 33.2 67.3 67.5 9.65 19.00 908.0 28.0 62.2 20.00 909.0 27.4 61.9 S6 21.00 7.02 903.0 906.7 26.8 27.4 60.7 61.7 9.60 22.00 543.0 13.6 84.5 23.00 717.0 9.4 84.8 S7 24.00 7.15 793.0 684.3 9.0 10.7 86.0 85.1 8.75 25.00 857.0 7.6 83.6  26.00 876.0 7.0 88.1   S8 27.00   7.28 883.0   872.0 6.8   7.1 82.6         84.8   8.83 Appendix E 177   Sample pH  NH4-N (mg/L) Average N mg/L PO4-P mg/L Average P mg/L Mg mg/L Average Mg mg/L Cond. ms/cm 28.00 833.0 5.5 81.4 29.00 831.0 5.3 81.2 S9 30.00 7.46 801.0 821.7 5.1 5.3 81.141 81.3 8.80 31.00 744.0 5.1 81.842 32.00 752.0 4.9 81.6035 S10 33.00 7.55 748.0 748.0 4.8 5.0 82.1335 81.9 8.65 34.00 650.0 4.2 80.439 35.00 647.0 4.0 79.287 S11 36.00 7.72 648.0 648.3 4.2 4.2 78.812 79.5 8.55 37.00 537.0 3.0 76.219 38.00 537.0 3.4 76.024 S12 39.00 8.21 546.0 540.0 3.1 3.2 88.417 80.2 8.40 40.00 1060.0 373.0 351.4 41.00 1030.0 362.0 340.2 S0-2 initial for second set 42.00 5.93 1040.0 1043.3 398.0 377.7 338.7 343.4 9.40 43.00 895.0 17.6 89.5032 44.00 866.0 17.8 89.3148 S13 45.00 7.32 855.0 872.0 14.0 16.5 88.6218 89.1 10.40 46.00 875.0 12.3 87.8658 47.00 854.0 10.6 89.895 S14 48.00 7.42 882.0 870.3 10.4 11.1 87.8892 88.6 10.20 49.00 829.0 8.4 85.8546 50.00 857.0 7.8 85.7814 S15 51.00 7.59 862.0 849.3 7.5 7.9 86.514 86.1 10.50 52.00 777.0 6.2 83.7954 53.00 774.0 6.0 85.2294 S16 54.00 7.71 768.0 773.0 6.0 6.1 84.9228 84.6 10.40 55.00 700.0 5.3 83.6244 56.00 698.0 4.7 83.0964 S17 57.00 8.00 682.0 693.3 4.1 4.7 82.5468 83.1 9.80 58.00 625.0 5.1 80.7192 59.00 873.0 3.8 79.7058 S18 60.00 8.34 575.0 691.0 3.6 4.2 80.6688 80.4 9.90 Appendix E 178  Table E.2. Distilled water at 10°C.  Sample pH  NH4-N (mg/L) Average N mg/L PO4-P mg/L Average P mg/L Mg mg/L Average Mg mg/L Cond ms/cm  1 295.0 397.0 210.9  2 297.0 404.0 213.3 Initial for set 1 S0  3 2.64 293.0 295.0 400.0 400.3 218.1 214.1 3.35 4  312.0 437.0 252.0  5 314.0 433.0 258.9 S1  6 6.14 311.0 312.3 438.0 436.0 257.4 256.1 3.20  7 289.0 388.0 214.4  8 286.0 379.0 214.4 S2  9 6.25 284.0 286.3 376.0 381.0 218.3 215.7 3.00  10 241.0 275.0 122.9  11 241.0 277.0 123.3 S3  12 6.50 241.0 241.0 276.0 276.0 124.9 123.7 2.80  13 216.0 204.0 46.9  14 213.0 199.0 50.1 S4  15 6.90 216.0 215.0 201.0 201.3 50.0 49.0 2.90  16 194.0 176.0 26.0  17 192.0 174.0 26.2 S5  18 7.20 191.0 192.3 173.0 174.3 26.9 26.4 2.90  19 182.0 166.0 9.4  20 182.0 168.0 10.1 S6  21 7.32 182.0 182.0 168.0 167.3 10.2 9.9 2.8/  22 189.0 161.0 10.2  23 192.0 159.0 10.3 S7  24 7.30 182.0 350.0 157.0 159.0 10.4 10.3 2.95  25 185.0 155.0 6.7  26 191.0 154.0 6.8 S8  27 7.50 190.0 187.7 154.0 154.3 7.0 6.8 3.03  28 180.0 151.0 4.3  29 181.0 153.0 4.4 S9  30 7.70 179.0 188.7 152.0 152.0 4.6 4.4 2.90  31 160.0 154.0 4.6  32 161.0 155.0 4.7 S10  33 7.74 157.0 180.0 155.0 154.7 4.8 4.7 2.88  34 147.0 152.0 1.6  35 142.0 151.0 1.6 S11  36 8.30 143.0 159.3 151.0 151.3 1.6 1.6 2.91  37 124.0 150.0 1.1  38 123.0 146.0 1.2 S12  39 8.60 127.0 144.0 144.0 146.7 1.2 1.2 2.85  40 346.0 534.0 261.2  41 353.0 544.0 262.9 initial for set 2 S13  42 5.90 351.0 350.0 540.0 539.3 264.8 263.0 3.30 Appendix E 179   Table E.3. Distilled water at 20°C Sample pH  NH4-N (mg/L) Average N (mg/L) PO4-P (mg/L Average P (mg/L) Mg (mg/L) Cond (mS/cm) 1.00 536.0 468.0 2.00 544.0 488.0 3.00 initial - 5.92 538.0 539.3 483.0 479.7 245.0 7.19  4.00 424.0 130.0 5.00 420.0 122.0 6.00 7.46 416.0 420.0 119.0 123.7 12.5 7.14  7.00 441.0 222.0 8.00 421.0 185.0 9.00 7.53 407.0 423.0 158.0 188.3 11.9 7.10  10.00 380.0 125.0 11.00 376.0 120.0 12.00 7.67 378.0 378.0 121.0 122.0 7.5 7.14  13.00 353.0 114.0 14.00 345.0 111.0 15.00 7.88 351.0 349.7 114.0 113.0 5.2 6.98  16.00 296.0 108.0 17.00 289.0 106.0 18.00 8.13 287.0 290.7 106.0 106.7 3.3 7.00  19.00 233.0 108.0 20.00 233.0 110.0 21.00 8.33 236.0 234.0 110.0 109.3 2.4 7.02   Appendix E 180  Table E.4.  Tap water at 15°C     Sample pH  NH4-N (mg/L) Average N (mg/L) PO4-P (mg/L) Average P (mg/L) Mg (mg/L) Average Mg (mg/L) Cond (ms/cm) 1.00 294.0 492.0 302.8 2.00 288.0 490.0 305.5 S0 15.00  4.44 287.0 289.7 477.0 486.3 309.2 305.8 4.24 3.00 355.0 533.0 325.4 4.00 352.0 524.0 321.2 S1 16.00 6.00 356.0 354.3 522.0 526.3 319.6 322.1 5.37 5.00 307.0 394.0 210.5 6.00 297.0 397.0 209.5 S2 17.00 6.30 297.0 300.3 401.0 397.3 213.1 211.0 5.01 22.00 175.0 223.0 76.8 23.00 175.0 225.0 77.4 S7 36.00 6.75  175.0  224.0 77.3 77.2 4.20 7.00 203.0 251.0 69.3 8.00 195.0 237.0 70.9 S3 18.00 6.80 198.0 198.7 230.0 239.3 69.9 70.0 4.55 24.00 138.0 175.0 37.5 25.00 139.0 175.0 37.6 S8 37.00 7.10  138.5  175.0 37.2 37.5 4.00 9.00 171.0 183.0 21.8 10.00 173.0 176.0 21.9 S4 19.00 7.30 169.0 171.0 174.0 177.7 22.1 21.9 4.51 11.00 150.0 154.0 17.7 12.00 151.0 157.0 17.5 S5 20.00 7.50 155.0 152.0 156.0 155.7 17.3 17.5 4.25 Appendix E 181   Sample pH  NH4-N (mg/L) Average N (mg/L) PO4-P (mg/L) Average P (mg/L) Mg (mg/L) Average Mg (mg/L) Cond (ms/cm) 26.00 118.0 150.0 17.1 27.00 118.0 147.0 17.1 S9 38.00 7.51  118.0  148.5 16.8 17.0 4.00 28.00 111.0 141.0 12.6 29.00 111.0 140.0 12.5 S10 39.00 7.66  111.0  140.5 12.5 12.5 4.10 13.00 134.0 145.0 10.8 14.00 128.0 139.0 10.8 S6 21.00 7.70 132.0 131.3 142.0 142.0 10.7 10.8 4.18 30.00 102.0 148.0 10.5 31.00 99.9 146.0 10.7 S11 40.00 7.80  101.0  147.0 10.7 10.6 4.20 32.00 91.9 137.0 6.1 33.00 91.1 138.0 6.2 S12 41.00 8.10  91.5  137.5 6.2 6.1 4.10 34.00 308.0 509.0 302.8 35.00 320.0 504.0 305.5 S0-2 initial for second set 42.00 3.81  314.0  506.5 309.2 305.8 4.82  Appendix E 182  Table E.5. Tap water at 20°C    Sample pH  NH4-N (mg/L) Average N (mg/L) PO4-P (mg/L) Average P (mg/L) Mg (mg/L) Average Mg (mg/L) Cond (ms/cm) 1 260.0 491.0 399.2 2 269.0 534.0 381.2 Initial for set 1 S0 3 5.50 261.0 263.3 520.0 515.0 377.8 386.1 4.50 4 266.0 522.0 381.2 5 273.0 514.0 380.9 S1 6 6.06 271.0 270.0 514.0 516.7 377.8 380.0 4.75 7 213.0 380.0 279.7 8 214.0 381.0 271.3 S2 9 6.26 214.0 213.7 381.0 380.7 274.9 275.3 4.40 10 152.0 228.0 167.2 11 154.0 225.0 165.5 S3 12 6.59 154.0 153.3 217.0 223.3 163.2 165.3 4.05 13 107.0 127.0 94.3 14 109.0 127.0 93.7 S4 15 6.93 110.0 108.7 128.0 127.3 92.6 93.5 3.00 16 88.0 97.0 72.0 17 87.2 99.0 71.8 S5 18 7.20 87.7 87.6 98.5 98.2 68.5 70.8 3.90 19 69.6 80.8 72.9 20 72.7 80.6 60.5 S6 21 7.21 73.8 72.0 81.5 81.0 61.1 64.8 3.95 25 74.9 79.4 58.8 26 76.6 79.1 61.7 S7 27 7.36 77.4 76.3 79.3 79.3 61.5 60.7 4.70 Appendix E 183   Sample pH  NH4-N (mg/L) Average N (mg/L) PO4-P (mg/L) Average P (mg/L) Mg (mg/L) Average Mg (mg/L) Cond (ms/cm) 28 68.1 63.8 52.9 29 68.2 64.5 51.7 S8 30 7.55 67.4 76.3 64.5 64.3 52.1 52.2 4.75 31 55.8 64.1 50.6 32 57.2 63.9 51.7 S9 33 7.69 56.9 67.9 63.4 63.8 51.9 51.4 5.50 34 47.1 49.5 42.0 35 47.9 50.0 40.1 S10 36 7.89 48.1 56.6 46.1 48.5 39.7 40.6 4.65 37 31.6 43.4 32.1 38 31.5 12.3 33.2 S11 39 8.18 31.2 47.7 42.6 32.8 32.8 32.7 4.60 40 27.8 34.0 25.4 41 29.2 34.9 25.6 S12  42 8.45 29.5 31.4 34.0 34.3 25.4 25.5 4.85 22 335.0 690.0 508.6 23 333.0 688.0 494.1  initial for set 2 S13 24 5.24 337.0 335.0 686.0 688.0 489.3 497.4 5.45 Appendix F 184  APPENDIX F: CARBON DIOXIDE TESTS Table F.1. Run #1. Total feed: 2.5 L/min, Recycle Ratio R = 6, Operating pH  = 8.1  R #1 Inflow to reactor Effluent Day pH Cond Mg PO4-P NH3-N pH Cond Mg PO4-P NH3-N Molar P removal Caustic use  Molar     (mS/cm) (mg/L) (mg/L) (mg/L)   (mS/cm) (mg/L) (mg/L) (mg/L)   (kg/day) NaOH usage 1 7.53 6.9 99.4 90.4 878 8.15 6.9 5.5 21 819 2.24E-03 2.91 72.8 2 7.61 7.2 84.5 85.8 838 8.22 7.53 33.2 5.3 810 2.60E-03 1.66 41.5 3 7.68 7.5 83.7 81.1 797 8.40 7.27 13.3 10.5 741 2.28E-03 2.30 57.5 4 7.63 7.4 81.0 81.6 800 8.20 7.76 74.9 2.7 736 2.55E-03 1.93 48.3 5 7.58 7.36 60.4 82.0 802 8.29 7.13 2.4 12.3 723 2.25E-03 2.93 73.3  R #2 Influent to reactor Effluent Day pH Cond Mg PO4-P NH3-N pH Cond Mg PO4-P NH3-N Molar P removal Caustic use  Molar     (mS/cm) (mg/L) (mg/L) (mg/L)   (mS/cm) (mg/L) (mg/L) (mg/L)   (kg/day) NaOH usage 1 7.53 6.9 82.1 90.4 878 8.16 7.11 10 16.2 841 2.40E-03 0.43 10.8 2 7.61 7.2 86.1 85.8 838 8.3 7.81 57.6 3.3 787 2.66E-03 0.81 20.3 3 7.68 7.5 82.3 81.1 797 8.1 7.39 24.2 16.1 734 2.10E-03 1.65 41.3 4 7.63 7.4 81.0 81.6 800 8.3 8.48 31.2 2.0 649 2.57E-03 2.14 53.5 5 7.58 7.36 76.0 82.0 802 8.4 7.0 18.3 5.9 703 2.46E-03 1.12 28.0     Appendix F 185  Table F.2. Run # 2.  Total feed: 2.61 L/min,  Recycle ratio: 6, Operating pH: 8.1 Harvest zone up flow velocity: 400 cm/min,  With air  R #1 Influent to reactor Effluent Day pH Cond Mg PO4-P NH3-N pH Cond Mg PO4-P NH3-N Molar P removal Caustic use  Molar     (mS/cm) (mg/L) (mg/L) (mg/L)   (mS/cm) (mg/L) (mg/L) (mg/L)   (kg/day) NaOH usage 1 7.8 7.62 81.8 88.3 826 8.29 7.13 15.2 9.56 740 2.54E-03 0.92 23.0 2 7.75   87.9 78.0 814 8.3   21.8 7.5 780 2.28E-03 1.55 38.8 3 7.78 7.22 90.0 84.1 845 8.41 6.73 20.5 6.5 791 2.51E-03 1.65 41.3 4 8.4 7.4 78.8 82.2 775 8.42 7.04 30.2 8.5 761 2.38E-03 1.53 38.3   R # 2 Inflow to reactor Effluent Day pH Cond Mg PO4-P NH3-N pH Cond Mg PO4-P NH3-N Molar P removal Caustic use  Molar     (mS/cm) (mg/L) (mg/L) (mg/L)   (mS/cm) (mg/L) (mg/L) (mg/L)   (kg/day) NaOH usage 1 7.8 7.62 91.3 88.3 826 8.2 7 21.0 9.1 772 2.56E-03 0.61 15.3 2 7.75   87.9 78.0 814 8.1   24.3 8.6 790 2.24E-03 0.34 8.5 3 7.78 7.22 90.0 84.1 845 8.3 7.11 22.1 7.8 779 2.46E-03 0.93 23.3 4 8.4 7.4 92.0 82.2 775 8.5 7.13 72.0 7.1 752 2.42E-03 1.46 36.5     Appendix F 186   Table F.3. Run # 3 Total feed: 2.61 L/min, Recycle ratio: 6, Operating pH: 8.1 Harvest zone upflow velocity: 400 cm/min, No air   R # 1 Inflow to reactor Effluent Date pH Cond Mg PO4-P NH3-N pH Cond Mg PO4-P NH3-N Molar P removal Caustic use     (mS/cm) (mg/L) (mg/L) (mg/L)   (mS/cm) (mg/L) (mg/L) (mg/L)   (kg/day) 1 8.3   78.3 69.3 634 8.51 7.21 55.9 7.1 370 2.01E-03 1.69 2 7.78 7.16 77.1 50.6 827 8.44 7.26 26.2 6.9 761 1.41 E-03 1.64 3 8 6.83 75.0 29.7 698 8.49 5.18 54.4 3.1 573 0.86 E-03 1.02 4 7.82 7.15 72.4 63.1 760 8.45 7.09 27.0 5.0 718 1.88 E-03 1.7   R # 2 Inflow to reactor Effluent Day pH Cond Mg PO4-P NH3-N pH Cond Mg PO4-P NH3-N Molar P removal Caustic use     (mS/cm) (mg/L) (mg/L) (mg/L)   (mS/cm) (mg/L) (mg/L) (mg/L)   (kg/day) 1 8.3   78.3 69.3 634 8.46   33.7 6.9 600 2.01E-03 0.57 2 7.78 7.16 82.7 50.6 827 8.48 7.5 76.0 7.0 672 1.41E-03 0.94 3 8 6.83 69.7 29.7 698 8.36 8.09 46.8 2.4 505 8.81E-04 0.48 4 7.82 7.15 72.4 63.1 760 8.3 7.14 31.7 4.5 713 1.89E-03 1.17      Appendix F 187   Table F.4. Run # 4. Total feed: 2.05 L/min, Recycle ratio:9, Operating pH: 8.1 Harvest zone up flow velocity: 450 cm/min, With air   R #1 Inflow to reactor Effluent Day pH Cond Mg PO4-P NH3-N pH Cond Mg PO4-P NH3-N Molar P removal Caustic use     (mS/cm) (mg/L) (mg/L) (mg/L)   (mS/cm) (mg/L) (mg/L) (mg/L)   (kg/day) 1 8 6.99 100.2 67.3 799 8.42 7.04 44.4 4.6 696 2.03E-03 0.87 2 8.1 12.48 80.7 38.2 553 8.45 5.25 16.9 4.5 483 1.09 E-03 1.05 3 7.86 5.09 81.7 42.6 575 8.48 5.33 44.4 4.9 525 1.22 E-03 1.15 4 7.93 5.99 59.8 49.0 670 8.46 5.94 32.5 6.3 652 1.38 E-03 1.40 5 7.92 5.68 75.2 52.6 696 8.25 5.8 38.5 9.0 637 1.41 E-03 2.29   R # 2 Inflow to reactor Effluent Day pH Cond Mg PO4-P NH3-N pH Cond Mg PO4-P NH3-N Molar P removal Caustic use     (mS/cm) (mg/L) (mg/L) (mg/L)   (mS/cm) (mg/L) (mg/L) (mg/L)   (kg/day) 1 8 6.99 75.58 67.3 799 8.42 6.98 34.29 4.42 732 2.03E-03 0.38 2 8.1 12.48 80.71 38.2 553 8.54 8.61 77.7 2.47 545 1.15E-03 0.53 3 7.86 5.09 83.55 42.6 575 8.23 8.47 76 2.96 221 1.28E-03 0.48 4 7.93 5.99 78.54 49 670 8.22 6.14 52.6041 5.94 653 1.39E-03 0.08 5 7.92 5.68 99.49 52.6 696 8.31 5.94 49.1331 5.13 624 1.53E-03 0.55 Appendix G 188  APPENDIX G: INDIRECT METHOD (USING pH AND CONDUCTIVITY) TO DETERMINE EXTERNAL Mg ADDITION REQUIREMENTS FOR STRUVITE CRYSTALLIZATION Table G.1. Synthetic water                                                                       Run 1 ml of MgCl2 added pH Conductivity (mS/cm) dpH/dml dcon/dml 0 8.21 1.886 0.5 8.2 1.916 0.02 0.06 1 8.2 1.944 0 0.056 1.5 7.76 1.919 0.88 -0.05 2 7.82 1.933 0 0.028 2.5 7.72 1.949 0.2 0.032 3 7.61 1.952 0.22 0.006 3.5 7.55 1.967 0.12 0.03 4 7.49 1.981 0.12 0.028 4.5 7.43 2.02 0.12 0.078 5 7.38 2.03 0.1 0.02 5.5 7.34 2.05 0.08 0.04 6 7.29 2.07 0.1 0.04 6.5 7.26 2.1 0.06 0.06 Run 2 ml of MgCl2 added pH  Conductivity (mS/cm) dpH/dml dcon/dml 0 8.25 2.01 0.5 8.23 2.05 0.04 0.08 1 8.23 2.08 0 0.06 1.5 7.86 2.05 0.74 -0.06 2 7.78 2.06 0.16 0.02 2.5 7.71 2.08 0.14 0.04 3 7.56 2.08 0.3 0 3.5 7.49 2.1 0.14 0.04 4 7.44 2.13 0.1 0.06 4.5 7.38 2.15 0.12 0.04 5 7.33 2.18 0.1 0.06  Appendix G 189  Run 3 ml of MgCl2 added pH Conductivity (mS/cm) dpH/dml dcon/dml 0 8.23 2.03 0.5 8.23 2.07 0 0.08 1 8.22 2.1 0.02 0.06 1.5 7.88 2.06 0.68 -0.08 2 7.79 2.08 0.18 0.04 2.5 7.69 2.09 0.2 0.02 3 7.58 2.1 0.22 0.02 3.5 7.52 2.12 0.12 0.04 4 7.44 2.13 0.16 0.02 4.5 7.38 2.16 0.12 0.06 5 7.32 2.18 0.12 0.04 Appendix G 190  Table G.2. Centrate matrix Run 1 ml of MgCl2 added pH Conductivity (mS/cm) dpH/dml dcon/dml 0 7.92 7.18 0.5 7.91 7.21 0.01 0.03 1 7.91 7.25 0 0.04 1.5 7.91 7.3 0 0.05 2 7.91 7.34 0 0.04 2.5 7.9 7.38 0.01 0.04 3 7.9 7.4 0 0.02 3.5 7.87 7.42 0.03 0.02 4 7.8 7.42 0.07 0 4.5 7.8 7.43 0 0.01 5 7.78 7.46 0.02 0.03 5.5 7.77 7.48 0.01 0.02 6 7.76 7.5 0.01 0.02 6.5 7.76 7.52 0 0.02 7 7.75 7.54 0.01 0.02 7.5 7.75 7.56 0 0.02 8 7.74 7.58 0.01 0.02 8.5 7.74 7.6 0 0.02 9 7.73 7.62 0.01 0.02 9.5 7.73 7.65 0 0.03 10 7.72 7.68 0.01 0.03             Appendix G 191  Run 2  ml of MgCl2 added pH Conductivity (mS/cm) dpH/dml dcon/dml 0 7.78 6.09 0.5 7.78 6.12 0 0.06 1 7.78 6.14 0 0.04 1.5 7.75 6.14 0.06 0 2 7.7 6.12 0.1 -0.04 2.5 7.69 6.13 0.02 0.02 3 7.66 6.15 0.06 0.04 3.5 7.66 6.16 0 0.02 4 7.65 6.17 0.02 0.02 4.5 7.63 6.19 0.04 0.04 5 7.62 6.2 0.02 0.02 5.5 7.61 6.22 0.02 0.04 6 7.6 6.24 0.02 0.04 6.5 7.59 6.25 0.02 0.02 7 7.58 6.27 0.02 0.04 7.5 7.57 6.29 0.02 0.04 8 7.57 6.32 0 0.06 8.5 7.56 6.33 0.02 0.02 9 7.55 6.36 0.02 0.06 9.5 7.56 6.38 -0.02 0.04              Appendix G 192  Run 3 ml of MgCl2 added pH Conductivity (mS/cm) dpH/dml dcon/dml 0 8.65 6.92 0.5 8.66 6.94 -0.02 0.04 1 8.66 6.95 0 0.02 1.5 8.65 6.95 0.02 0 2 8.64 6.96 0.02 0.02 2.5 8.63 6.96 0.02 0 3 8.62 6.95 0.02 -0.02 3.5 8.62 6.96 0 0.02 4 8.61 6.97 0.02 0.02 4.5 8.61 6.98 0 0.02 5 8.6 6.99 0.02 0.02 5.5 8.59 7 0.02 0.02 6 8.59 7.02 0 0.04 6.5 8.58 7.04 0.02 0.04 7 8.58 7.05 0 0.02 7.5 8.57 7.11 0.02 0.12 8 8.57 7.13 0 0.04 8.5 8.56 7.18 0.02 0.1 Appendix H 193  APPENDIX H: APPLICATION OF CHEMICALS FOR THE REDUCTION OF PHOSPHATE CONCENTRATION IN WATER MATRIX Table H.1. Effects on phosphate concentrations due to PAC addition to centrate Test Number PAC (mL added) Phosphate (mg/L) 1 0.1 91.7  0.2 74.6  0.3 71.2  0.4 61.7  0.5 32.5  0.6 38.3  0.8 9.93  1.0 2.88 2 0.0 106  0.1 90.9  0.2 63.7  0.3 38  0.4 9.71  0.5 56.5*  0.6 7.69  0.8 9.68 3 0.0 108  0.2 48.8  0.4 11.3  0.6 1.35  0.8 1.03  1.0 1.04  * Possible outlier   Appendix H 194   Table H.2. Effects on pH, phosphate and aluminum concentrations due to alum addition to centrate  Alum added  PO4-P pH Al  (ml) (mg/L)   (mg/L) Test 1 0 196 8.26 0.66  2 157 8.28 1.94  4 119 8.22 1.91  6 94.9 8.13 14.76  8 67.1 8.13 12.95  10 45.5 8.06 13.79  Test  2  0 199 8.3 1.07  2 151 8.26 0.67  4 116 8.19 5.34  6 87.4 8.13 15.57  8 62.8 8.07 16.03  10 43 7.99 15.97   Appendix I 195  APPENDIX I: HARDNESS TEST RESULTS Table I.1. Comparison of calcium concentrations in LIWWTP centrate between samples tested by AA and by hardness test method Ca by hardness test (mg/L) Ca by AA Ca from AA Absolute error (mg/L) (mg/L) 44.0 35.9 8.1 36.0 33.3 2.7 28.0 33.0 4.9 28.0 31.4 3.4 44.0 30.5 13.5 40.0 52.8 12.8 40.0 37.2 2.9 36.0 32.3 3.7 32.0 34.8 2.8 32.0 32.8 0.7 32.0 33.6 1.6 31.3 33.9 2.6 14.4 14.8 0.3 11.2 10.1 1.1 10.4 8.7 1.7 10.4 7.6 2.8 8.4 8.3 0.1 Appendix I 196  Table I.2. Comparison of magnesium concentrations in LIWWTP centrate between samples tested by AA and by hardness test method  Mg by hardness test (mg/L) Mg by AA Mg from AA Absolute error (mg/L) (mg/L) 18.8 16.3 2.6 13.7 12.7 1.0 13.7 12.3 1.4 15.2 12.5 2.7 15.2 12.1 3.1 9.7 13.9 4.2 14.6 13.7 0.9 14.6 13.0 1.6 4.9 7.8 2.9 7.3 5.2 2.1 9.7 6.2 3.5 6.3 5.5 0.8 5.3 5.7 0.3 4.6 5.8 1.2 4.6 6.2 1.6 Appendix J 197  APPENDIX J: OPERATIONAL DATA – REACTOR 1   Flows Date   Mg inflow Centrate  Total influent Recycle  Total Upflow RR     mL/min L/min L/min L/min L/min cm/min 1-Jul Wednesday 0.0 2.70 2.70 18.30 21.00 461 6.8 2-Jul Thursday 0.0 2.40 2.40 -2.40 3-Jul Friday 0.0 2.80 2.80 18.80 21.60 474 6.7 7-Jul Tuesday 84.0 2.52 2.60 15.40 18.00 395 5.9 8-Jul Wednesday 90.0 2.41 2.50 15.50 18.00 395 6.2 9-Jul Thursday 84.0 2.52 2.60 15.40 18.00 395 5.9 10-Jul Friday 90.0 2.41 2.50 15.50 18.00 395 6.2 11-Jul Saturday 90.0 2.51 2.60 14.80 17.40 382 5.7 12-Jul Sunday 90.0 2.54 2.63 14.47 17.10 375 5.5 13-Jul Monday 90.0 2.41 2.50 15.80 18.30 401 6.3 14-Jul Tuesday 54.0 2.65 2.70 15.60 18.30 401 5.8 15-Jul Wednesday 54.0 2.87 2.92 15.80 18.72 411 5.4 16-Jul Thursday 54.0 2.45 2.50 14.50 17.00 373 5.8 17-Jul Friday 54.0 2.85 2.90 14.10 17.00 373 4.9 19-Jul Sunday 54.0 2.75 2.80 14.90 17.70 388 5.3 20-Jul Monday 54.0 2.45 2.50 15.50 18.00 395 6.2 21-Jul Tuesday 54.0 2.60 2.65 16.25 18.90 414 6.1 22-Jul Wednesday 54.0 2.55 2.60 15.40 18.00 395 5.9 23-Jul Thursday 54.0 2.55 2.60 16.00 18.60 408 6.2 24-Jul Friday 54.0 2.61 2.66 15.64 18.30 401 5.9  Appendix J 198     Flows Date   Mg inflow Centrate Total influent Recycle  Total Upflow RR    mL/min L/min L/min L/min L/min cm/min 25-Jul Saturday 68.0 2.43 2.50 16.10 18.60 408 6.4 26-Jul Sunday       16.10 16.10 353 recycle mode 27-Jul Monday 62.0 2.41 2.47 14.93 17.40 382 6.0 28-Jul Tuesday   3.30 3.30 15.40 18.70 410 4.7 29-Jul Wednesday   2.50 2.50 15.80 18.30 401 6.3 30-Jul Thursday 0.0 2.50 2.50 15.20 17.70 388 6.1 31-Jul Friday 0.0 2.60 2.60 -2.60 1-Aug Saturday 60.0 2.59 2.65 15.05 17.70 388 5.7 11-Aug Tuesday 60.0 -0.06   0.00 12-Aug Wednesday 60.0 2.60 2.66 26.74 29.40 645 10.1 13-Aug Thursday 60.0 2.54 2.60 15.40 18.00 395 5.9 17-Aug Monday 66.0 1.08 1.15 14.15 15.30 336 12.3 18-Aug Tuesday 0.0 1.70 1.70 14.80 16.50 362 8.7 19-Aug Wednesday       16.00 16.00 351 recycle mode 28-Aug Friday 78.0 2.47 2.55 15.75 18.30 401 6.2 29-Aug Saturday 60.0 2.34 2.40 16.20 18.60 408 6.8 30-Aug Sunday 31-Aug Monday 80.0 1.47 1.55 14.95 16.50 362 9.6 1-Sep Tuesday 102.0 3.15 3.25 15.95 19.20 421 4.9 2-Sep Wednesday 102.0 3.45 3.55 16.05 19.60 430 4.5     Appendix J 199   Centrate Influent pH Conductivity Temp PO4-P NH4-N Mg Mg:P N:P PO4-P NH4-N Mg PO4-P NH4-N Mg   mS/cm C mg/L mg/L mg/L     mg/L mg/L mg/L moles/L moles/L moles/L 7.35 1.26 23.1 13.6 105.0 5.4 0.51 17.08 13.6 105.0 5.4 4.39E-04 7.50E-03 2.22E-04 7.50 1.37 24.7 16.2 120.5 4.4 0.35 16.45 16.2 120.5 4.4 5.23E-04 8.61E-03 1.81E-04 7.44 3.38 27.8               0.0 7.30 0.95 24.7 9.3 63.5 4.8 0.66 15.10 9.0 61.4 67.6 2.91E-04 4.39E-03 2.78E-03 7.30 0.94 25.3   69.1 5.8     0.0 66.6 75.8 0.00E+00 4.76E-03 3.12E-03 7.44 3.99 28.0   402.0 12.6     0.0 389.0 75.2 0.00E+00 2.78E-02 3.09E-03 7.34 2.84 31.0   233.0 8.8         78.7 7.47 4.36 30.0 53.6 357.0 11.4 0.27 14.73 51.7 344.6 78.5 1.67E-03 2.46E-02 3.23E-03 7.45 3.84 29.0 48.7 396.0 9.8 0.26 17.99 47.0 382.4 76.1 1.52E-03 2.73E-02 3.13E-03 7.52 4.84 26.5 59.1 513.0 12.4 0.27 19.20 57.0 494.5 82.1 1.84E-03 3.53E-02 3.38E-03 7.49 4.85 30.0 67.0 481.5 12.9 0.24 15.90 65.7 471.9 51.6 2.12E-03 3.37E-02 2.12E-03 7.50 3.07 26.5 36.1 317.0 8.4 0.30 19.43 35.4 311.1 44.3 1.14E-03 2.22E-02 1.82E-03 7.30 1.12 38.9 9.2 84.3 2.4 0.33 20.27 9.0 82.5 44.4 2.91E-04 5.89E-03 1.83E-03 7.45 2.89 34.5 15.6 109.3 6.4 0.52 15.50 15.3 107.3 42.5 4.94E-04 7.66E-03 1.75E-03 7.20 1.03 36.3     5.4     0.0 0.0 42.9 0.00E+00 0.00E+00 1.77E-03 7.5 2.17 34.0 21.8 173.5 4.2         46.2     1.90E-03 7.4 2.47 33.7 24.7 198.0 6.1 0.31 17.73 24.2 194.0 45.7 7.81E-04 1.39E-02 1.88E-03                 0.0 0.0 40.5 0.00E+00 0.00E+00 1.67E-03 7.5 2.29 29.6 32.7 246.0 5.5 0.21 16.64 32.0 240.9 45.9 1.03E-03 1.72E-02 1.89E-03 7.5 4.22 28.2 52.7 442.5 9.6 0.23 18.57 51.6 433.5 49.0 1.67E-03 3.10E-02 2.02E-03 7.54 3.00 26.8 44.4 361.0 9.1 0.26 17.99 52.4 690.2 61.9 1.69E-03 4.93E-02 2.55E-03                       0.00E+00 0.00E+00 0.00E+00 7.63 4.90 27.7 60.4 565.5 11.7 0.25 20.71 58.9 551.3 60.3 1.90E-03 3.94E-02 2.48E-03 7.48 4.87 28.4 48.6 408.5 11.5 0.30 18.59 48.6 408.5 11.5 1.57E-03 2.92E-02 4.73E-04 Appendix J 200    Centrate Influent pH Conductivity Temp PO4-P NH4-N Mg Mg:P N:P PO4-P NH4-N Mg PO4-P NH4-N Mg   mS/cm C mg/L mg/L mg/L     mg/L mg/L mg/L moles/L moles/L moles/L 7.59 4.61 30.3 53.5 514.0 11.1 0.26 21.25 53.5 514.0 11.1 1.73E-03 3.67E-02 4.57E-04 7.5 4.85 29.7 71.3 604.0 8.5 0.15 18.74 71.3 604.0 8.5 2.30E-03 4.31E-02 3.50E-04 7.65 5.91 29.8 70.5 559.5 4.4 0.08 17.56 70.5 559.5 4.4 2.28E-03 4.00E-02 1.81E-04 7.6 5.80 30.0 58.0 682.0 3.6 0.08 26.01 56.7 666.6 47.7 1.83E-03 4.76E-02 1.96E-03 7.84 6.95 32.3 83.5 800.5 19.1 7.1   28.5 83.5 900.5 16.2 0.25 23.86 81.6 880.2 59.8 2.64E-03 6.29E-02 2.46E-03 8.08 6.90 27.4 101.0 920.0 17.4 0.22 20.15 98.7 898.8 62.0 3.19E-03 6.42E-02 2.55E-03 8.2 7.28 29.6 91.5 904.0 22.0 0.31 21.86 86.2 852.1 135.5 2.78E-03 6.09E-02 5.58E-03 8.26 6.69 28.7 75.0 796.5 8.3 0.14 23.49 75.0 796.5 8.3 2.42E-03 5.69E-02 3.42E-04 8.2 6.60 28.0 95.7 880.0 12.3 0.16 20.34       0.00E+00 0.00E+00 0.00E+00 8.36 7.05 30.1 92.1 867.0 8.1 0.11 20.82 89.3 840.5 69.0 2.88E-03 6.00E-02 2.84E-03     26.0 88.0 840.0 2.3 0.03 21.12 85.8 819.0 52.2 2.77E-03 5.85E-02 2.15E-03       . 8.35 6.39 28.7 67.0 762.5 14.9 0.28 25.18 63.5 723.1 117.4 2.05E-03 5.17E-02 4.83E-03     29.5 81.3 745.0 9.6 0.15 20.27 78.7 721.6 72.1 2.54E-03 5.15E-02 2.97E-03 8.3 6.96   66.4 690.0 8.4 0.16 22.99 64.5 670.2 65.6 2.08E-03 4.79E-02 2.70E-03       Appendix J 201   Effluent Expected In reactor Conditions PO4-P NH4-N Mg PO4-P NH4-N Mg Temp Conductivity PO4-P NH4-N Mg pH mg/L mg/L mg/L moles/L moles/L moles/L pH °C mS/cm mg/L mg/L mg/L 8.1             8.1 25.0 3.10 1.7 13.5 0.7 8.2             8.2 25.0 1.23 8.1             8.1 26.0 3.39     0.0 8.2 13.4 54.6 72.3 4.33E-04 3.90E-03 2.98E-03 8.2 25.0 1.50 12.8 55.6 71.6 8.4 18.0 54.1 14.7 5.81E-04 3.86E-03   8.4 25.0 0.90 15.5 55.8 23.2 8.1   291.0 57.6 0.00E+00 2.08E-02   8.1 25.0 3.92   305.2 60.1 7.9   281.5 64.5   2.01E-02   7.9 30.2 3.67     66.5 8.0 33.8 331.5 53.2 1.09E-03 2.37E-02 2.19E-03 8.0 32.0 4.46 36.5 333.5 56.9 8.0 33.5 278.5 63.3 1.08E-03 1.99E-02 2.60E-03 8.0 25.7 3.39 35.6 294.5 65.3 8.1 27.5 466.0 49.0 8.88E-04 3.33E-02 2.02E-03 8.1 24.2 4.90 31.5 469.9 53.5 8.0 15.6 416.5 17.4 5.04E-04 2.98E-02 7.14E-04 8.0 18.5 4.62 23.0 424.7 22.4 8.0 10.4 217.0 19.8 3.36E-04 1.55E-02 8.15E-04 8.0 26.0 2.86 14.3 231.7 23.6 8.0 8.2 110.5 31.9 2.65E-04 7.89E-03 1.31E-03 8.0 37.8 1.74 8.3 106.4 33.7 8.0 14.2 292.0 21.8 4.59E-04 2.09E-02 8.95E-04 8.0 27.8 3.25 14.4 260.5 25.3 7.3             7.3 27.7 1.94     6.8   13.6 133.0 32.2       8.0 27.7 1.78     34.1       33.7     1.39E-03 7.9 27.5 2.15 3.4 27.2 35.4               7.9 27.0 1.56     5.9   15.0 218.5 26.8 4.84E-04 1.56E-02 1.10E-03 7.9 28.7 2.65 17.4 221.6 29.5   10.5 333.0 34.5 3.39E-04 2.38E-02 1.42E-03 7.9 26.0 4.37 16.5 347.6 36.6   9.1 274.5 41.6 2.94E-04 1.96E-02 1.71E-03 8.0 26.1 3.21 13.7 284.8 44.3               8.0 26.1 3.21   12.2 504.0 40.9 3.94E-04 3.60E-02 1.68E-03 8.0 28.8 5.22 18.8 510.7 43.7   34.6 395.0 44.3 1.12E-03 2.82E-02 1.82E-03 7.4 27.0 3.85 37.1 397.4 38.5 Appendix J 202    Effluent Expected In reactor Conditions pH PO4-P NH4-N Mg PO4-P NH4-N Mg pH Temp Conductivity PO4-P NH4-N Mg  mg/L mg/L mg/L moles/L moles/L moles/L  °C mS/cm mg/L mg/L mg/L   13.3 400.5 27.8 4.29E-04 2.86E-02 1.14E-03 8.0 32.6 4.75 18.8 416.0 25.5   61.6 552.5 6.1 1.99E-03 3.95E-02 2.51E-04 8.1 29.8 4.88 63.0 559.8 6.4   53.2 656.5 3.8 1.72E-03 4.69E-02 1.56E-04 8.1 28.1 5.96   27.3 730.5 8.1 8.81E-04 5.22E-02 3.33E-04 8.1 27.7   31.7 720.9 14.0  8.4 22.5 795.0 11.4 7.27E-04 5.68E-02 4.69E-04 8.0   7.58 27.8 802.7 15.8 8.3 23.8 800.5 25.7 7.68E-04 5.72E-02 1.06E-03 7.2 26.6 6.45 34.6 814.7 30.9 9.1 21.9 818.0 14.1 7.07E-04 5.84E-02 5.80E-04 7.9 28.8 7.12 26.7 820.6 23.2   79.7 916.5 4.6 2.57E-03 6.55E-02 1.89E-04 7.9     79.2 904.1 5.0 9.8       0.00E+00 0.00E+00 0.00E+00 8.0 33.7 5.36 8.5 37.7 779.0 18.1 1.22E-03 5.56E-02 7.45E-04 7.3 29.9 7.09 44.9 787.6 25.2   42.1 796.5 10.4 1.36E-03 5.69E-02 4.28E-04 7.2   6.70 47.7 799.4 15.8               8.0   6.8 705.0 14.1 2.57E-03 6.55E-02 2.98E-03 7.9 28.9   12.1 706.7 23.8   34.5 705.5 10.9 2.57E-03 6.55E-02 2.98E-03   28.3   42.0 708.2 21.3   9.4 687.0 5.9 6.28E-04 2.35E-02 8.67E-04   24.0 3.34 19.4 684.0 16.7 Appendix J 203  APPENDIX J: OPERATIONAL DATA – REACTOR 2   Flows Date   Mg inflow Centrate Total influent Recycle  Total Upflow RR     mL/min L/min L/min L/min L/min cm/min 1-Jul Wednesday  3.0 3.0 13.2 16.2 355 4.4 2-Jul Thursday   2.5 2.5 15.5 18.0 395 6.2 3-Jul Friday   2.7 2.7 18.3 21.0 461 6.8 7-Jul Tuesday 102.0 2.5 2.6 15.4 18.0 395 5.9 8-Jul Wednesday   2.5 2.5 14.9 17.4 382 6.0 9-Jul Thursday 84.0 2.5 2.6 15.4 18.0 395 5.9 10-Jul Friday 84.0 2.0 2.1 15.3 17.4 382 7.3 11-Jul Saturday 84.0 2.0 2.1 15.9 18.0 395 7.6 12-Jul Sunday 84.0 2.4 2.5 14.7 17.1 375 6.0 13-Jul Monday 84.0 2.3 2.4 15.6 18.0 395 6.5 14-Jul Tuesday 60.0 2.1 2.2 -2.2   0 -1.0 15-Jul Wednesday 60.0 2.8 2.9 16.6 19.5 428 5.7 16-Jul Thursday 54.0 2.4 2.5 16.2 18.6 408 6.6 17-Jul Friday       16.0 16.0 351 18-Jul Saturday       16.0 16.0 351 recycle mode 19-Jul Sunday       16.0 16.0 351 recycle mode 20-Jul Monday       16.0 16.0 351 recycle mode 21-Jul Tuesday       13.0 13.0 285 recycle mode 22-Jul Wednesday 42.0     13.0 13.0 285 recycle mode 23-Jul Thursday 54.0 2.7 2.8 15.3 18.0 395 5.5  Appendix J 204    Flows  Mg inflow Centrate Total influent Recycle  Total Upflow RR Date  mL/min L/min L/min L/min L/min cm/min 24-Jul Friday 54.0 2.8 2.8 15.2 18.0 395 5.4 25-Jul Saturday 56.0 2.6 2.7 15.4 18.0 395 5.8 26-Jul Sunday 0.0 0.0 0.0 15.4 15.4 338 recycle mode 27-Jul Monday 62.0 2.0 2.1 15.1 17.2 378 7.2 28-Jul Tuesday 48.0 2.2 2.3 15.4 17.7 387 6.8 29-Jul Wednesday 48.0 2.3 2.3 15.7 18.0 395 6.8 30-Jul Thursday 0.0 2.5 2.5 15.5 18.0 395 6.2 31-Jul Friday 48.0 2.6 2.6 15.4 18.0 395 5.9 11-Aug Tuesday             recycle mode 12-Aug Wednesday 48.0 2.7 2.7 16.8 19.5 428 6.2 13-Aug Thursday 48.0 1.9 2.0 14.0 15.9 349 7.2 16-Aug Sunday 0.0 0.0 17-Aug Monday 48.0 0.0 0.0 0.0 18-Aug Tuesday 19-Aug Wednesday 102.0 2.2 2.4 15.4 17.7 388 6.5 20-Aug Thursday 48.0 2.6 2.6 15.4 18.0 395 5.9 28-Aug Friday 48.0 2.0 2.0 15.7 17.7 388 7.9 29-Aug Saturday 48.0 2.4 2.4 15.2 17.6 386 6.3 30-Aug Sunday             recycle mode 31-Aug Monday 50.0 2.7 2.7 14.9 17.6 386 5.5 1-Sep Tuesday 96.0 2.5 2.6 15.7 18.3 401 6.0 2-Sep Wednesday 96.0 2.5 2.6 17.8 20.4 447 6.9   Appendix J 205    Centrate Influent pH Cond Temp PO4-P NH4-N Mg Mg:P N:P PO4-P NH4-N Mg PO4-P NH4-N Mg   mS/cm C mg/L mg/L mg/L     mg/L mg/L mg/L moles/L moles/L moles/L 7.35   23.1 13.6 105.0 5.4 0.51 17.08 13.6 105.0 5.4 4.39E-04 7.50E-03 2.22E-04 7.50   24.7 16.2 120.5 4.4 0.35 16.45 16.2 120.5 4.4 5.23E-04 8.61E-03 1.81E-04 7.44 3.38 27.8           0.0 0.0 0.0 7.30   24.7 9.3 63.5 4.8 0.66 15.10 8.9 61.0 83.1 2.89E-04 4.36E-03 3.42E-03 7.30   25.3   69.1 5.8       69.1 5.8   4.94E-03 2.39E-04 7.44 3.99 28.0   402.0 12.6       389.0 76.8   2.78E-02 3.16E-03 7.34 2.84 31.0   233.0 8.8       223.7 88.4 7.47 4.36 30.0 53.6 357.0 11.4 0.27 14.73 51.5 342.7 90.9 1.66E-03 2.45E-02 3.74E-03 7.45 3.84 29.0 48.7 396.0 9.8 0.26 17.99 47.0 382.4 78.0 1.52E-03 2.73E-02 3.21E-03 7.52 4.84 26.5 59.1 513.0 12.4 0.27 19.20 57.0 495.0 81.9 1.84E-03 3.54E-02 3.37E-03 7.49 4.85 30.0 67.0 481.5 12.9 0.24 15.90 65.2 468.4 67.0 2.10E-03 3.35E-02 2.76E-03 7.50 3.07 26.5 36.1 317.0 8.4 0.30 19.43 35.4 310.4 49.6 1.14E-03 2.22E-02 2.04E-03 7.30   38.9 9.2 84.3 2.4 0.33 20.27 9.0 82.4 46.4 2.91E-04 5.89E-03 1.91E-03 7.45 2.89 34.5 15.6 109.3 6.4 0.52 15.50  7.20   36.3     5.4 7.45 2.17 34.0 21.8 173.5 4.2 7.41 2.47 33.7 24.7 198.0 6.1 0.31 17.73  7.51 2.29 29.6 32.7 246.0 5.5 0.21 16.64 32.1 241.2 43.7 1.04E-03 1.72E-02 1.80E-03 7.54 4.22 28.2 52.7 442.5 9.6 0.23 18.57 51.7 434.1 46.6 1.67E-03 3.10E-02 1.92E-03 7.54 3.00 26.8 44.4 361.0 9.1 0.26 17.99 43.5 353.4 50.1 1.40E-03 2.52E-02 2.06E-03 Appendix J 206   Centrate Influent pH Cond Temp PO4-P NH4-N Mg Mg:P N:P PO4-P NH4-N Mg PO4-P NH4-N Mg   mS/cm C mg/L mg/L mg/L     mg/L mg/L mg/L moles/L moles/L moles/L 7.63 4.90 27.7 60.4 565.5 11.7 0.25 20.71 58.6 548.8 68.9 1.89E-03 3.92E-02 2.83E-03 7.48 4.87 28.4 48.6 408.5 11.5 0.30 18.59 47.6 399.8 52.8 1.54E-03 2.86E-02 2.17E-03 7.59 4.61 30.3 53.5 514.0 11.1 0.26 21.25 52.4 503.3 51.5 1.69E-03 3.59E-02 2.12E-03 7.50 4.85 29.7 71.3 604.0 8.5 0.15 18.74 71.3 604.0 8.5 2.30E-03 4.31E-02 3.50E-04 7.65 5.91 29.8 70.5 559.5 4.4 0.08   69.2 549.2 40.3 7.84 6.95 32.3 83.5 800.5 19.1 7.10   28.5 83.5 900.5 16.2 0.25 23.86 82.0 884.5 50.6 2.65E-03 6.32E-02 2.08E-03 8.08 6.90 27.4 101.0 920.0 17.4 0.22 20.15 98.5 897.4 64.9 3.18E-03 6.41E-02 2.67E-03  8.20 7.28 29.6 91.5 904.0 22.0 0.31 21.86 0.0 0.0 1949.0 0.00E+00 0.00E+00 8.02E-02 8.26 6.69 28.7 75.0 796.5 8.3 0.14 23.49 8.20 6.60 28.0 95.7 880.0 12.3 0.16 20.34 91.5 841.8 96.4 2.96E-03 6.01E-02 3.97E-03 8.21 6.90 28.5 88.0 866.5 8.8 0.13 21.78 86.4 850.5 44.6 2.79E-03 6.08E-02 1.84E-03 8.36 7.05 30.1 92.1 867.0 8.1 0.11 20.82 89.9 846.2 54.7 2.90E-03 6.04E-02 2.25E-03     26.0 88.0 840.0 2.3 0.03 21.12 86.2 823.2 41.2 2.78E-03 5.88E-02 1.70E-03  8.35 6.39 28.7 67.0 762.5 14.9 0.28 25.18 65.8 748.4 50.7 2.12E-03 5.35E-02 2.09E-03     29.5 81.3 745.0 9.6 0.15 20.27 78.3 717.5 81.2 2.53E-03 5.12E-02 3.34E-03 8.34 6.96   66.4 690.0 8.4 0.16 22.99 63.9 664.2 80.9 2.06E-03 4.74E-02 3.33E-03   Appendix J 207  Effluent Expected In reactor Conditions PO4-P NH4-N Mg PO4-P NH4-N Mg Conductivity Temp PO4-P NH4-N Mg mg/L mg/L mg/L moles/L moles/L moles/L pH mS C mg/L mg/L mg/L           8.15 2.76 24.7 13.6 105.0 1.0          8.04 1.30 24.4 16.2 120.5 0.6          7.81 1.78 26.4     0.0 9.7 49.3 54.0 0.00031 0.00352 0.00222 8.15 2.28 24.1 66.4 353.0 58.2 15.0 50.1 58.0 0.00048 0.00358 0.00239 7.86 2.02 24.0 89.4 367.7 50.5   296.0 43.5 0.00000 0.02114 0.00179 7.66 2.12 25.0  2142.2 48.3   244.5 41.7   0.01746 0.00172 7.98 2.29 25.0   2005.0 47.3     39.1   0.00161 8.06 2.53 25.4 51.5 342.7 45.1     54.0   0.00222 8.1 2.71. 26.2 47.0 382.4 57.4     32.8   0.00135 8.13 3.49 25.7 57.0 495.0 39.3     18.4   0.00076 8.09 4.60 24.2 65.2 468.4     30.7   0.00126 7.6 2.62 24.5 35.4 310.4 33.5     28.6   0.00117 8.1 1.67 26.3 9.0 82.4 30.9          8.07 2.96 27.8           8.13   27.7      194.8   0.00802 8.14 2.79 27.5   194.8          8.14       6.3 14.0 199.0 21.6 0.00045 0.01421 0.00089 8.12 2.14 28.2 16.8 205.4 25.0 12.8 310.0 25.5 0.00041 0.02214 0.00105 8.15 2.32 26.0 18.9 329.5 28.8 5.2 273.0 27.3 0.00017 0.01950 0.00112 8.18 2.7 26.1 10.8 284.8 30.7              8.21 2.47 26.2 Appendix J 208   Effluent  Expected In reactor conditions PO4-P NH4-N Mg PO4-P NH3-N Mg pH Conductivity Temp PO4-P NH4-N Mg mg/L mg/L mg/L moles/L moles/L moles/L  mS/cm C mg/L mg/L mg/L 25.3 440.0 42.8 0.00082 0.03143 0.00176 8.3 3.08 27.4 29.4 453.3 46.0 6.3 262.5 30.3 0.00020 0.01875 0.00125 8.17 3.5 27.2 11.6 280.0 33.2 8.0 396.5 15.2 0.00026 0.02832 0.00063 7.57 4 29.7 13.7 410.1 19.8 50.4 529.5 0.9 0.00163 0.03782 0.00004 8.15 4.85 26.5 53.3 539.8 2.0 9.8 546.5 6.5 0.00032 0.03904 0.00027 8.13   28.0 18.4 546.9 11.4  29.1 825.5 14.9 0.00094 0.05896 0.00061 7.89 2.15 26.0 36.4 833.7 19.8 44.7 862.5 15.1 0.00144 0.06161 0.00062 7.66 3.39 25.6 51.3 866.8 21.2   13.0 727.0 23.6 0.00042 0.05193 0.00097 8.27 6.05 26.5 23.4 742.2 33.3 18.3 795.0 6.1 0.00059 0.05679 0.00025 7.6 5.46 26.4 28.1 803.0 11.7 22.2 746.5 10.4 0.00072 0.05332 0.00043 8.14 4.1 27.2 29.8 757.8 15.4 26.6 787.5 4.3 0.00086 0.05625 0.00018 8.15 4.16 25.0 34.7 792.4 9.3             8.5 3.2 26.0 8.3 713.5 10.3 0.00163 0.06161 0.00802 8.13 3.72 25.6 17.1 718.9 16.5 18.2 655.5 21.5 0.00000 0.00000 0.00000 7 6.28 28.3 26.7 664.3 30.0 10.5 708.0 9.1 0.00035 0.01748 0.00082 8     17.2 702.5 18.1 Appendix K 209  APPENDIX K: COMPOSITION OF STRUVITE PELLETS – ICP-MS TEST RESULTS Sampling date   2-Sep 2-Sep 2-Sep 21-Aug 21-Aug 20-Aug  Unit MDL 4.7mm 3.33mm 1.98mm 3.33mm 1.98mm 1.98mm Aluminum (Al)  % 0.01  <0.01  <0.01  <0.01  <0.01  <0.01  <0.01 Arsenic (As)  ppm 0.1 0.3 0.3 0.6 0.3 0.3 0.4 Barium (Ba)  ppm 10  <10  <10  <10  <10  <10  <10 Berylium (Be)  ppm 0.05  <0.05  <0.05  <0.05  <0.05  <0.05  <0.05 Bismuth (Bi)  ppm 0.01 0.01  <0.01  <0.01  <0.01  <0.01  <0.01 Cadmium (Cd)  ppm 0.01  <0.01  <0.01  <0.01  <0.01  <0.01  <0.01 Calcium (Ca)  % 0.01 0.19 0.19 0.2 0.2 0.19 0.2 Cerium (Ce)  ppm 0.02 0.17  <0.02 0.03  <0.02  <0.02  <0.02 Cesium (Cs)  ppm 0.05  <0.05  <0.05  <0.05  <0.05  <0.05  <0.05 Cobalt (Co)  ppm 0.1  <0.1 0.4  <0.1  <0.1  <0.1  <0.1 Copper (Cu)  ppm 0.2 16.4 3.2 4.3 2.5 2.9 2.3 Cromium (Cr)  ppm 1 4 4 5 4 4 4 Gallium (Ga)  ppm 0.05 0.8 0.69 0.73 0.65 0.71 0.7 Germanium (Ge)  ppm 0.05  <0.05  <0.05  <0.05  <0.05  <0.05  <0.05 Indium (In)  ppm 0.005  <0.005  <0.005  <0.005  <0.005  <0.005  <0.005 Iron (Fe)  % 0.01 0.04 0.04 0.04 0.04 0.04 0.04 Lanthanum (La)  ppm 0.2  <0.2  <0.2  <0.2  <0.2  <0.2  <0.2 Lead (Pb)  Ppm 0.2 0.7 0.5 1.5 0.2 0.3 0.2 Lithium (Li)  ppm 0.1 0.6 0.3 0.6 0.3 0.2 0.2 Manganese (Mn) ppm         5 218 216 229 240 248 266 Appendix K 210  Sampling date   2-Sep 2-Sep 2-Sep 21-Aug 21-Aug 20-Aug  Unit MDL 4.7mm 3.33mm 1.98mm 3.33mm 1.98mm 1.98mm Mercury (Hg)  ppm 0.01  <0.01  <0.01  <0.01  <0.01 0.12  <0.01 Molybdenum (Mo)  ppm 0.05  <0.05  <0.05  <0.05  <0.05  <0.05  <0.05 Nickle (Ni)  ppm 0.2 0.6  <0.2 0.4 0.2 0.2 0.2 Potassium (K)  % 0.01 0.03 0.03 0.03 0.03 0.03 0.03 Rhenium (Re)  ppm 0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001 Rubidium (Rb)  ppm 0.1 0.9 0.9 1 1 1 1 Scandium (Sc)  ppm 0.1  <0.1  <0.1  <0.1  <0.1  <0.1  <0.1 Selenium (Se)  ppm 0.2  <0.2  <0.2  <0.2  <0.2  <0.2  <0.2 Silver (Ag)  ppm 0.01  <0.01  <0.01  <0.01  <0.01  <0.01  <0.01 Sodium (Na)  % 0.01 0.01 0.01 0.01  <0.01  <0.01 0.01 Strontium (Sr)  ppm 0.2 1.1 1.4 1.4 1.3 1.6 1.7 Sulphur (S)  % 0.01  <0.01  <0.01  <0.01  <0.01  <0.01  <0.01 Thalium (Tl)  ppm 0.02  <0.02  <0.02  <0.02  <0.02  <0.02  <0.02 Tin (Sn)  ppm 0.2 28 5.1 9.3 4.4 5.4 2.5 Titanium (Ti)  % 0.005 0.025 0.025 0.023 0.025 0.022 0.023 Tungsten (W)  ppm 0.05 0.19 2.66 0.16 0.12 0.08 0.06 Uranium (U)  ppm 0.05  <0.05  <0.05  <0.05  <0.05  <0.05  <0.05 Vanadium (V)  ppm 1  <1           <1           <1           <1           <1           <1 Zinc (Zn)  ppm 2 4 4 5 3 4 4 Zirconium (Zr)  ppm 0.5  <0.5  <0.5  <0.5  <0.5  <0.5  <0.5 Appendix L 211  APPENDIX L: CRUSHING STRENGTH DATA Average Strength Std. Dev Sample Number P (mg/L) N (mg/L) Mg (mg/L) Fe (mg/L) Ca (mg/L) g N g 1 573.5 235.5 365.8 0.6 0.5 1503 14.7 305.5 2 536 228 362.8 0.6 7.1 3346 32.8 636.2 3 552.5 237.5 385.8 0.6 3.7 1736 17.0 325.4 4 541 228.5 410.6 0.5 4.0 2714 26.6 745 5  583 241.5 393.7 0.5 5.6 1868 18.3 464 6 515.5 222 344.0 0.7 9.2 2804 27.5 433 7 580 255 390.77 1.4 7.5 2187 21.5 370 8 552 245 372.48 1.5 8.3 2110 20.7 329 9 524.5 215.5 356.5 0.8 9.0 2544 25.0 557 10  560.5 237 369.5 0.6 7.7 2283 22.4 379 11 364 156 256.65 1.1 15.2 2828 27.7 620.7 12 516 223 361.22 1.4 8.3 2950 28.9 493.7 13 563 245 384.73 1.5 7.7 2420 23.7 289 

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:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0062917/manifest

Comment

Related Items