UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The impact of fabric and surface characteristics on the engineering behavior of polymer-amended mature… Boxill, Lois Esther 2016

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

Item Metadata

Download

Media
24-ubc_2016_may_boxill_lois.pdf [ 5.65MB ]
Metadata
JSON: 24-1.0228588.json
JSON-LD: 24-1.0228588-ld.json
RDF/XML (Pretty): 24-1.0228588-rdf.xml
RDF/JSON: 24-1.0228588-rdf.json
Turtle: 24-1.0228588-turtle.txt
N-Triples: 24-1.0228588-rdf-ntriples.txt
Original Record: 24-1.0228588-source.json
Full Text
24-1.0228588-fulltext.txt
Citation
24-1.0228588.ris

Full Text

THE IMPACT OF FABRIC AND SURFACE CHARACTERISTICS ON THE ENGINEERING BEHAVIOR OF POLYMER-AMENDED MATURE FINE TAILINGS    by    LOIS ESTHER BOXILL    B.Sc.CE, The Georgia Institute of Technology, 1998 M.Sc.CE, The Georgia Institute of Technology, 1999    A THESIS SUBMITTED IN PARTIAL FULFILMENT OF   THE REQUIREMENTS FOR THE DEGREE OF   DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES  (Mining Engineering)    THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)    March 2016     © Lois Esther Boxill, 2016 ii  Abstract Management and reclamation of large inventories of legacy and fresh mature fine tailings (MFT) represent a continuing and significant challenge to surface mine operators in the Alberta Oil Sands because of the complex chemical and physical behavior of these tailings.  Suncor’s tailings recovery operations (TROTM) and Shell Canada’s atmospheric fines drying (AFD) use anionic polymers to flocculate MFT to remove fine tailings solids from aqueous suspension.  Addition of anionic polyacrylamide polymer to MFT results in the creation of a complex synthetic material, polymer-amended MFT (PA-MFT). This research investigates the fundamental properties and characteristics governing PA-MFT dewatering in an effort to better understand how these factors contribute to overall material behavior including strength development and consolidation.  This work confirms that the addition of anionic polyacrylamide polymer does little to change the zeta potential of the input raw MFT as the resulting material is colloidally stable.  The work also indicates how residual bitumen and fabric act independently and in combination to reduce the permeability of PA-MFT and enable retention of water within the material’s fabric especially when it is stored in lifts that exceed the depth at which the combination of evaporative drying and underdrainage are effective.   It is concluded that PA-MFT fabric includes abundant micropores with tortuous flow paths with little connectivity between pores.  The size and configuration of the pores effectively traps water within the PA-MFT fabric.  Residual bitumen may also block pore throats or form a hydrophobic barrier that limits both the effectiveness of evaporation for material deposited below a depth of 15 cm in thick lift deposits.  Residual bitumen may also prevent diffusion of trapped water upward through deposited material and contributes to the plasticity and compressibility exhibited by PA-MFT when it is deposited in lifts thicker than 20 cm.     iii  Preface The definition, design and scope of the research program, data collection, analysis and reporting (tabulation and graphical presentation), and preparation of this thesis manuscript were completed by the author.  All material samples provided to external laboratories for testing were sampled, prepared and submitted by the author.  Research co-supervisors Dr. Dirk van Zyl and Dr. Les Lavkulich provided critical feedback to confirm that the scope and depth of the proposed research and experimental program were commensurate with PhD level research.  Technical input and feedback on experimental work were also periodically provided by the following specialists who served on the author’s PhD committee: Dr. Bern Klein (rheology and flocculation), Dr. Dharma Wijiwickreme (geotechnical engineering and soil testing methods), and Mr. Daryl Hockley (applied civil engineering). Dr. Adrian Revington provided an in-laboratory demonstration of a method that could be used to consistently flocculate mature fine tailings (MFT) and prepare samples of polymer-amended mature fine tailings (PA-MFT).  The author used this method to generate the PA-MFT tested in this research program. Mr. Bill Chen, a student in the Civil Engineering program at UBC at the time, was trained by the author in the method so that he could assist with preparation of almost 76 kg of PA-MFT used to complete experimental test work.  The author performed the specific gravity determinations of MFT and PA-MFT and MBI determinations using the Canmet Energy and OFITE methods described in this thesis.  The author trained Mr. Chen to complete these three test methods so that he could assist with validation and verification of the results obtained by the author. Dean-Stark analyses to determine percent bitumen, water and mineral content, and duplicate percent clay determination using the Canmet slurry method using methylene blue (NRC,  iv  2008), were completed by the Exova laboratory in Calgary on subsamples of the homogenized MFT provided to the lab by the author.  The Burnaby laboratory of SGS completed mineralogical characterization of PA-MFT samples provided by the author using QEMSCAN® and XRD methods.  In both cases the author was responsible for interpretation, analysis and synthesis of the raw data provided by both laboratories.  These data sets were used to establish and complete baseline characterization of raw and flocculated MFT. Details of the rheological characterization test program were developed by the author in consultation with Dr. Babak Derakhshandeh.  The author completed determination of the flow curves for MFT and PA-MFT using a Brookfield rheometer. Oscillatory rheology test work and additional peak yield and flow curve data for select samples of hydrated polymer, AITF-MFT, and PA-MFT were completed using more advanced rheometer and viscometer equipment. This testing took place at the Burnaby facilities of Coanda Research and Development Corporation (Coanda).  Dr. Derakhshandeh completed testwork on the Anton Paar viscometer while Ms. Ashley Gonzalez completed rheometer testing on duplicate samples of PA-MFT. The author reduced, completed comparative analysis and interpretation of data provided by Coanda.  The author completed FT-IR spectra and zeta potential determinations for samples of input and secondary materials described in this thesis using equipment located in the UBC Mining Engineering Department.  Raman spectra and photographs of the material surfaces being analyzed were obtained from the LabRAM HR unit owned by the UBC Materials Engineering Department.  This unit was operated by Ms. Ye Zhu who was guided by the author to the areas of the sample for which imaging and spectra were to be developed.  The author completed the analysis and interpretation of FT-IR spectra and zeta potential data.  Dr. Xia Liu of the Canadian  v  Light Source (CLS) assisted the author with interpretation of the Raman spectra after it had been base shift corrected by Ms. Ye Zhu. Field emission scanning electron microscopy (FESEM) and cryo-SEM images were obtained by Mr. Derek Horne of UBC’s BioImaging Facility for samples of wet PA-MFT that were prepared and provided by the author.  SEM imaging was also analyzed and described by the author with confirmation of various features by Mr. Horne. Samples of wet and dry MFT and PA-MFT were provided by the author for imaging at the CLS in Saskatoon.  Dr. Julie Thompson and Mr. Jeremy Olson oversaw collection of x-ray fluorescence and computer tomography data from the three beamlines used.  The author worked closely with these researchers to complete post-processing and analysis of the collected data.  The author provided interpretation of the imaging and spectral data provided by the CLS, as presented in this thesis.  Dr. Julie Thompson provided background references to aid the author’s interpretation of x-ray fluorescence and sulfur-adsorption spectra.  The author prepared MFT and PA-MFT samples for analysis in UBC’s seepage induced consolidation test (SICT) cell.  Mr. Mathiew Estepho assisted the author with set up and operation of the SICT cell and reduction of the collected data.  The author completed an independent interpretation of the collected data and comparison against data from literature.  The author designed the evaporation test cell used in the research.  The test cell was fabricated and calibrated by RST Instruments.  Operation of the test cell, and the collection, analysis and interpretation of all data collected was completed by the author. Comparison of collected test cell data against predicted surface consolidation from the CONDES0 model were also developed by the author.     vi  Table of Contents Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iii List of Tables ................................................................................................................................ xi List of Figures .............................................................................................................................. xii List of Abbreviations .................................................................................................................. xv Acknowledgements .................................................................................................................... xvi Chapter 1: Introduction ............................................................................................................... 1 1.1 Overview .......................................................................................................................... 1 1.2 Background ...................................................................................................................... 3 1.3 Importance of This Study ................................................................................................. 6 1.4 Problem Statement ........................................................................................................... 8 1.5 Research Objectives ....................................................................................................... 10 Chapter 2: Literature Review .................................................................................................... 13 2.1 Overview of Desirable Tailings Characteristics and Management Approaches ............ 13 2.2 Fluid Fines Tailings in the Alberta Oil Sands ................................................................ 14 2.2.1 Composition of Mature Fine Tailings and Identification of “Bad Actors” ............ 18 2.2.2 Fine Tailings Management Using Chemical Additives .......................................... 20 2.2.3 Use of Anionic Polymers to Aid Fine Tailings Dewatering ................................... 22 2.3 Polymer Flocculation Processes ..................................................................................... 23 2.4 Ionic Concentration of Primary extraction Water .......................................................... 27 2.5 pH ................................................................................................................................... 28 2.6 Residual Bitumen ........................................................................................................... 29  vii  2.7 Assemblage Structure and Material Fabric .................................................................... 30 2.8 Polymer-MFT Interactions ............................................................................................. 32 2.9 Effectiveness of Polymers Used for Fine Tailings Management ................................... 35 2.10 Total Cost and Technical Considerations for Management of PA-MFT ....................... 36 2.11 Rheological and Strength Characterization of Flocculated Oil Sands Tailings ............. 38 2.11.1 Rheological Characterization .................................................................................. 38 2.11.2 Assessing Impacts of Material Handling ................................................................ 39 2.11.3 Geotechnical Strength Characterization ................................................................. 39 2.11.3.1 Development of Parametric Undrained Strength Relationships ...................... 43 2.11.3.2 Large Strain Consolidation .............................................................................. 44 2.12 Summary of Key Points from Literature Review........................................................... 47 Chapter 3: Experimental Program ........................................................................................... 50 3.1 Introduction .................................................................................................................... 50 3.2 Program Components ..................................................................................................... 50 3.3 Samples and Reagents .................................................................................................... 57 3.3.1 Mature Fine Tailings ............................................................................................... 57 3.3.2 Primary Extraction Water ....................................................................................... 58 3.3.3 Polymer ................................................................................................................... 59 3.3.4 Polymer-Amended Mature Fine Tailings ............................................................... 60 3.3.5 Other Reagents ........................................................................................................ 62 3.4 Procedures, Methods and Equipment ............................................................................. 63 3.4.1 Geotechnical Moisture Content and Percent Solids Determination ....................... 63 3.4.2 Baseline Characterization of Raw AITF-MFT ....................................................... 64  viii  3.4.3 Specific Gravity Determination for PA-MFT ......................................................... 64 3.4.4 Percent Clay Determination Using Methylene Blue Index ..................................... 65 3.4.4.1 CANMET Method for Slurries ........................................................................ 65 3.4.4.2 OFITE Method for Drilling Muds ................................................................... 66 3.4.5 Mineralogical Characterization of PA-MFT ........................................................... 67 3.4.6 Zeta Potential Determination .................................................................................. 67 3.4.7 Rheological Characterization .................................................................................. 68 3.4.7.1 Flow Curve Development ................................................................................ 68 3.4.7.2 Yield Stress Determination .............................................................................. 69 3.4.7.3 Determination of Storage and Loss Moduli of PA-MFT ................................ 71 3.4.8 Characterization of PA-MFT Using an Evaporation Test Cell ............................... 71 3.4.8.1 Strength Characterization of Material in Evaporation Test Cell ..................... 74 3.4.8.2 Undrained Strength Determination Using Field Vane .................................... 75 3.4.8.3 Undrained Strength Determination Using the Swedish Fall Cone .................. 75 3.4.8.4 Investigation into the Effect of Surface Load Application .............................. 76 3.4.9 Seepage Induced Consolidation Testing ................................................................. 76 3.4.10 Infrared Spectroscopy ............................................................................................. 77 3.4.11 Raman Spectroscopy ............................................................................................... 78 3.4.12 Synchrotron Based Computer Tomography ........................................................... 79 3.4.13 X-ray Fluorescence Microscopy and X-ray Adsorption Spectroscopy .................. 80 3.4.14 Scanning Electron Microscopy ............................................................................... 82 Chapter 4: Results....................................................................................................................... 83 4.1 Baseline Characterization of PA-MFT and its Component Materials ........................... 83  ix  4.1.1 Methylene Blue Index Results ................................................................................ 84 4.1.2 Specific Gravity and Other Bitumen Effects on PA-MFT Phase Relationships .... 85 4.1.3 Mineralogy of PA-MFT .......................................................................................... 87 4.2 Surface Characterization of a PA-MFT ......................................................................... 88 4.2.1 Zeta Potential .......................................................................................................... 88 4.2.2 Investigation of Functional Groups Using Spectroscopy ....................................... 92 4.2.3 Quantification of PA-MFT Surface Area Covered by Visible Bitumen................. 99 4.2.4 X-Ray Fluorescence Analysis ............................................................................... 100 4.2.5 X-ray Sulphur Absorption Analysis...................................................................... 101 4.2.6 Effects of Surface Characteristics on Dewatering of a PA-MFT ......................... 104 4.3 Rheological Characterization of a PA-MFT and its Component Materials ................. 106 4.3.1 Flow Curve Characteristics of BASF 5250 and Pre-Sheared AITF-MFT ............ 106 4.3.2 PA-MFT Peak Yield Stress ................................................................................... 108 4.3.3 Oscillatory Rheology ............................................................................................ 114 4.3.4 Effects of Dewatering Condition on Rheological Characteristics of a PA-MFT . 116 4.4 SICT of PA-MFT and Raw AITF-MFT ....................................................................... 118 4.5 Dewatering and Surface Deformation Characteristics of a PA-MFT .......................... 123 4.5.1 Dewatering Characteristics of a PA-MFT ............................................................ 123 4.5.2 Surface Deformation Characteristics of a PA-MFT ............................................. 129 4.5.2.1 Evaporation Test Cell Sample Surface Deformation Results ........................ 129 4.5.2.2 CONDES0 Results ........................................................................................ 132 4.6 Geotechnical Strength Characterization of a PA-MFT ................................................ 136 4.6.1 Shear Strength and Undrained Strength Characteristics of a PA-MFT ................ 136  x  4.6.2 Relating Geotechnical Properties .......................................................................... 142 4.7 Fabric Characterization for a PA-MFT ........................................................................ 144 4.7.1 Results of Synchrotron Based Computer Tomography ........................................ 144 4.7.1.1 Estimating Void Space in Wet PA-MFT ....................................................... 147 4.7.2 FESEM and cryo-SEM of Raw MFT and PA-MFT Fabric .................................. 147 4.7.3 Effects of a PA-MFT Fabric on Dewaterability ................................................... 152 Chapter 5: Discussion ............................................................................................................... 154 5.1 Overview of Research Findings ................................................................................... 154 5.2 Summary of Key Research Findings ............................................................................ 159 Chapter 6: Conclusions ............................................................................................................ 162 Chapter 7: Limitations and Contributions to Knowledge .................................................... 166 7.1 Limitations ................................................................................................................... 166 7.2 Contributions to Knowledge ........................................................................................ 167 Chapter 8: Recommendations for Future Work.................................................................... 170 Bibliography .............................................................................................................................. 174 Appendices ................................................................................................................................. 190     xi  List of Tables Table 2.1: Forces that Exist in Flocculated Suspensions (after Scales, 2013) .............................. 23 Table 2.2: Ions Present in Suncor and Syncrude Tailings Ponds ................................................. 27 Table 3.1: Summary of Characteristics Investigated and Methods Used ..................................... 54 Table 3.2: Characteristics of Primary Extraction Water (PEW) Used for Polymer Hydration .... 59 Table 3.3: Comparison of CANMET Slurry and OFITE MBI Methods ...................................... 67 Table 3.4: Summary of Investigations Completed Using the Evaporation Test Cell ................... 74 Table 4.1: Summary of Raw AITF-MFT and PA-MFT Index Properties .................................... 83 Table 4.2: Relationship between Bitumen Occupied Pores and Target Percent Solids................ 87 Table 4.3: Indication of Material Colloid Stability ....................................................................... 89 Table 4.4: Summary of Zeta Potential for Tested Suspensions .................................................... 90 Table 4.5: Summary of Observed Frequencies from IR and Raman Spectra ............................... 97 Table 4.6: Summary of Likely Sulphur Species ......................................................................... 103 Table 4.7: Summary of PA-MFT Shear Stress Test Data ........................................................... 110 Table 4.8: Summary of Slump Test Results ............................................................................... 114 Table 4.9: Summary of Input Material Characteristics and SICT Load Step Information ......... 119 Table 4.10: Summary of SICT Results and Compressibility and Permeability Relationships ... 122 Table 4.11: Summary of Dewatering Results from Evaporation Test Cell Results ................... 125 Table 4.12: Summary of Test Protocol and Findings from FESEM and Cryo-SEM ................. 149 Table 6.1: Summarized Research Conclusions ........................................................................... 164    xii  List of Figures Figure 1.1: Generalized Cross Section of Typical Oil Sands Tailings Ponds ................................ 4 Figure 1.2: Projected Accumulation of MFT at Oil Sands Surface Mining Projects ..................... 5 Figure 1.3: Fundamental Material Characteristics to Be Investigated .......................................... 12 Figure 2.1: Various Arrangements of Polymer Flocculants (after Mahmoudkhani et al, 2012) .. 24 Figure 2.2: Schematic of Bridging Flocculation (from Zeta-Meter, 1993) .................................. 25 Figure 2.3: Possible Effect of Bitumen on MFT Permeability Proposed by Scott et al. (1985)... 29 Figure 2.4: Hard and Soft Interactions (after Coussot, 2005) ....................................................... 31 Figure 2.5: Repeating Functional Unit in a Polyacrylamide Molecule (after Entry et al., 2002) . 34 Figure 2.6: Relationships between Solids Content, Tailings Volume, and Clay to Water Ratio . 36 Figure 2.7: Plasticity Chart for Flocculated Oil Sands Fine Tailings (from Beier et al. 2013) .... 41 Figure 3.1: Characterization and Behaviour Evaluation Framework ........................................... 51 Figure 3.2: Test Methods Used to Characterize Various Factors ................................................. 52 Figure 3.3: Overview of Experimental Program........................................................................... 53 Figure 3.4: Visual Comparison of (A) Raw AITF-MFT and (B) PA-MFT.................................. 60 Figure 3.5: Anton-Paar TruGap Rheometer (A) and Concentric Cylinder Geometry Detail (B) . 68 Figure 3.6: Infinite Cup and Vanes (A) Used with Brookfield Rheometer (B) ............................ 70 Figure 3.7: Evaporation Test Cell Setup (A), DrainTube® Filter Fabric (B), and Mass Used (C)....................................................................................................................................................... 72 Figure 3.8: Swedish Fall Cone Used in Test Work (A), Showing Sample Being Tested (B) ...... 76 Figure 3.9: UBC SICT Setup Showing Inner and Outer Cells Filling with PEW ........................ 77 Figure 3.10: Schematic Showing Principles of Attenuated Total Reflectance Spectroscopy ...... 78 Figure 3.11: LabRam HR Raman Spectrometer (A) with Optical Lens Above Sample (B) ........ 79  xiii  Figure 3.12:  Mounted Dry Samples Used for X-Ray Adsorption Spectroscopy (XANES) ........ 81 Figure 4.1: QEMSCAN Determination of Mineral Abundance in PA-MFT ............................... 88 Figure 4.2: FT-IR Spectra for PA-MFT and its Component Materials ........................................ 95 Figure 4.3: Raman Spectra for PA-MFT and its Component Materials ....................................... 96 Figure 4.4: Surface of Air Dried PA-MFT ................................................................................... 99 Figure 4.5: XRF Scans of Raw AITF-MFT and PA-MFT ......................................................... 100 Figure 4.6: Sulphur Absorption Spectra of PA-MFT and Its Component Materials .................. 102 Figure 4.7: Flow Curve of Hydrated Anionic Polyacrylamide Polymer .................................... 107 Figure 4.8: Flow Curve of Conditioned Raw AITF-MFT .......................................................... 108 Figure 4.9: PA-MFT Shear Stress Versus Time Curves ............................................................. 110 Figure 4.10: Storage and Loss Modulus Strain Sweep Profiles.................................................. 115 Figure 4.11: Compressibility Relationships for Raw AITF-MFT and PA-MFT ........................ 122 Figure 4.12: Permeability Relationships for Raw AITF-MFT and PA-MFT ............................. 123 Figure 4.13: Comparison of Dewatering Trends for Unloaded PA-MFT (Phase I Testing) ...... 125 Figure 4.14: Evaporation Cell Mass Balance for All Phases ...................................................... 126 Figure 4.15: Change in Percent Solids with Time Under Unloaded Conditions (Phase I) ......... 127 Figure 4.16: Evaporation Cell Test 2 – Change in Geotechnical Moisture Content with Time . 128 Figure 4.17: Change in Sample Thickness for Evaporation Cell Tests (Phase I) ....................... 130 Figure 4.18: Evaporation Cell Test 2 – Change in Sample Thickness with Time ...................... 131 Figure 4.19: Comparison of Phase I Evaporation Test Results and CONDES0 Predictions ..... 134 Figure 4.20: Comparison Between CONDES0 Predictions and EC Test 2 Results ................... 134 Figure 4.21: Change in Surface Shear Strength with Time for EC Test 1 and EC Test 2 .......... 138 Figure 4.22: Change in Undrained Shear Strength at 10 cm with Time (Phase I, Both Tests) .. 139  xiv  Figure 4.23: Change in Undrained Shear Strength at 14 cm with Time (Phase I, Both Tests) .. 140 Figure 4.24: Change in Undrained Shear Strength with Time for EC Test 2 (All Phases) ........ 141 Figure 4.25: Picture Showing Deformation of Surface PA-MFT Materials During EC Test 2 . 142 Figure 4.26: Relating Undrained Shear Strength; Solids and Geotechnical Moisture Contents 143 Figure 4.27: Wide Angle View at Interior of Wet PA-MFT Sample with Voids False Subset . 145 Figure 4.28: Wet PA-MFT Showing Voids (False Colored) in Selected Region ....................... 145 Figure 4.29: Dry PA-MFT Showing Voids (False Colored) in Selected Region ....................... 146 Figure 4.30: Cryo-SEM Image of Sublimated and Gold Sputter-Coated PA-MFT ................... 148 Figure 4.31: Potential Aging Effects of PA-MFT Kept Under Undrained Conditions .............. 151 Figure 4.32: Threadlike (A) and Globular (B) Features Visible on PA-MFT Mineral Surfaces 151 Figure 4.33: Differently Aged PA-MFT Samples in HMDS Prior to FESEM Imaging ............ 152 Figure 5.1: Summary of Connection Between Multi-Scale Processes ....................................... 155 Figure 6.1: Testing to Evaluate Dewatering Characteristics of Chemically Amended MFT ..... 162    xv  List of Abbreviations AITF  Alberta Innovates Technology Futures CLS  Canadian Light Source EC   Evaporation Cell ESRF  The European Synchrotron FESEM Field Emissions Scanning Electron Microscopy FT-IR  Fourier Transform Infrared Spectroscopy HMDS  Hexamethyldisilizane  MBI  Methylene Blue Index MFT  Mature Fine Tailings PAM  Polyacrylamide polymer PA-MFT Polymer-Amended Mature Fine Tailings PEW  Primary Extraction Water SICT  Seepage Induced Consolidation Test XANES X-ray Absorption Near Edge Structure XPS  X-ray Photoelectron Spectroscopy XRD  X-ray Diffraction XRF  X-ray Fluorescence    xvi  Acknowledgements Commencement and completion of this research journey would not have been possible without the significant contributions of wide ranging support from my wife, Dr. Tristin Wayte, my “boss” Daryl Hockley, financial support from my employer, SRK Consulting, and input from the dynamic duo of my research co-supervisors, Dr. Les Lavkulich and Dr. Dirk Van Zyl.  To each of you I owe a debt of tremendous gratitude I will never grow tired repaying. I would also like to acknowledge the support and encouragement received from my SRK colleagues who took care of key clients during the final year of my research. My ability to approach this multidisciplinary study would have been impossible without technical contributions provided by the following: Dr. Dharma Wijiwickreme and Dr. Bern Klein who served on my research committee; Dr. Dobroslav Znidarcic, Dr. Marek Pawlik, Dr. Adrian Revington, Dr. Dipo Omotoso, Dr. Heather Kaminsky, Dr. Babak Derakhshandeh, Dr. Sarah Prout, Dr. Jacob Kabel, Mr. Derrick Horne, and Mr. Garnet Martens, whose passion for and expertise in key technical areas provoked, inspired, and facilitated substantial advancement of my research; Mr. Rob Taylor who took my design sketches and specs and turned them in to working test apparatus - some of which he personally delivered to my lab; Ms. Carla Wood, Mr. Richard Fostokjian, and Mr. Pascal Saunier who provided the raw ingredients for this research.  Dr. Joseph Dove, Dr. Roger Melley, Dr. Kim Kasperski and Dr. Ali Hooshiar each provided kernels of wisdom and insight from the outset of my research that helped me get off to a running start, and in Joe’s case, insight into what would be required to successfully complete this work.  Ms. Sally Finora, Mr. Aaron Hope and Ms. Rosemarie Cocuaco provided invaluable support in helping me get my lab space up, running, and secure.  Mr. Mathiew Estepho and Mr. Bill Chen were my anchors in the lab – you both helped me take ideas of what was possible in our lab to a  xvii  new level and this research is the direct beneficiary of your efforts and dedication.  Thank you.  Dr. Esau Arinaitwe and Dr. Ahmed Moustafa also served as living examples of how one can navigate a PhD while keeping family first – thank you for your examples of rooted excellence.  The editing and formatting wizardry of Ms. Gabi Green and Mr. Aaron Fultz ensured that this document was ready for prime time in both digital and printed formats. Special thanks are extended to Dr. Julie Thompson and Mr. Jeremy Olson, scientists who made the Canadian Light Source feel like a home away from home.  The Canadian Light Source is funded by the Canada Foundation for Innovation, the Natural Sciences and Engineering Research Council of Canada, the National Research Council Canada, the Canadian Institutes of Health Research, the Government of Saskatchewan, Western Economic Diversification Canada, and the University of Saskatchewan. I would also like to express my gratitude to the Rev. Thomi Glover, Mr. John Glover, and Ms. Emily Patenaude who provided space on Mayne Island for me to take the first major plunge into the literature review required to complete my PhD proposal.  I would like to acknowledge the unparalleled prayer and emotional support received from dear friends and family, especially in the final year of my research.  I would especially like to acknowledge the encouragement provided by my grandmother Marjorie, my aunt Judy, my mother and sisters, my dear friends Dimple, Sandra, Mathieu, Ken and any of the other friends I inadvertently neglected during my research preoccupation. I close by acknowledging the constant companionship of my beloved canine friend, Juice, who taught me every day of this journey that curiosity always rewards and that companionship is for the long haul.  Thank you buddy!   xviii  Dedication  To Tris, Jaden, and Nile for being, to Juice, the best four-legged companion I could hope for, and To Great Spirit, who never places limits on possibility.    1  Chapter 1: Introduction 1.1 Overview Industrial scale production of bitumen from the Alberta Oil Sands represents one of the largest and most intensive resource development projects of modern times.  With conventional petroleum sources in decline since the late 1970s (Yergin, 2008), unconventional sources like oil sands and oil shale deposits in Canada and Venezuela have been increasingly exploited in an effort to keep pace with the continuing and increasing demand for energy rich transportation fuels (Rubin, 2012). Bitumen from the Alberta oil sands is described as “heavy oil” because of its large molecular mass and higher density when compared to conventional crude oil (Strausz, 1977).  Alberta oil sands bitumen has higher viscosity and specific gravity than petroleum products with lower molecular mass and density like West Texas Intermediate, a conventional crude oil with low sulphur content.  Bitumen production from oil sands surface mines in Alberta generates large quantities of tailings (approximately 12 volumes of total tailings for each volume of bitumen extracted) that generally fall into one of three categories: coarse tailings (mostly quartz sand) from hot water extraction (HWE); fine tailings (mostly clay minerals and silt) from HWE; and fine tailings (mostly clay minerals) from treatment of bitumen froth.  The fine tailings stream from HWE is generally characterized by widely dispersed clay minerals that are very slow to consolidate resulting in a class of materials referred to in industry as mature fine tailings (MFT).  MFT exhibits thixotropic properties meaning that its viscosity is stress and time dependent. These compounding factors produce significant challenges related to the management and storage of already vast inventories of MFT that necessitate expansion of existing large tailings storage facilities or construction of new ones.  2  In an effort to sequester fines and reduce fluid tailings inventories some operators pump stored MFT to dedicated disposal areas where the MFT is mixed using in-line injection with hydrated anionic polyacrylamide (PAM) upstream of subaerial discharge spigots.  The resulting material exhibits complex dewatering behavior that includes an initial release of water and the apparent ability to retain significant water within the material fabric.  This material also exhibits limited consolidation and strength development. This research attempts to better understand and interpret the factors contributing to the complex dewatering and consolidation behavior of the resulting geo-material, polymer-amended MFT (PA-MFT). To better understand and analyze bulk PA-MFT properties like permeability, plasticity, surface deformation, and strength development, fundamental properties like material fabric, identification of surface functional groups, and zeta potential are examined to understand the physico-chemical building blocks that combine to produce observed and measured PA-MFT properties and behavior.  Rheological characteristics of PA-MFT are also investigated to better understand the effects that physical processes like flocculation, deposition and material rehandling might have on the strength and durability characteristics of this material.  A comprehensive testing framework in which methods and results from a range of disciplines can be used to enable fundamental PA-MFT properties and behaviour to be identified and cross-referenced allows preliminary but important linkages between fundamental characteristics to be confirmed and described.    3  1.2 Background The largest concentration of oil sands accessible through surface mining techniques is located in the Wabiskaw-McMurray deposit located in the Athabasca oil sands area north of Ft McMurray, Alberta.  In terms of proven crude oil reserves, Alberta ranks third in the world behind Saudi Arabia and Venezuela (Alberta Energy, 2013).  In 2011 Alberta’s oil reserves were estimated at 170.2 billion barrels, with oil sands reserves accounting for 99% of this estimate (ERCB, 2012).  In the Athabasca region, the Wabiskaw-McMurray formation is the primary host of bitumen-producing ore and is typically encountered within 70 m of the existing ground surface.  The relative shallowness of the deposit enables surface mining methods to be used to obtain the ore which consists largely of bitumen, sand sized quartz, and clay minerals (Omotoso, 2011; Masliyah et al., 2011).  A hot water extraction process (in which water is typically heated to a minimum of 50 degrees Celsius) that incorporates flotation and dispersants, like sodium hydroxide to liberate bitumen from mineral solids, is used with these ores (Masliyah et al., 2011; Arinaitwe, 2013).  The fundamental process was developed by Dr. Karl Clark of the Research Council of Alberta in the early 1920s and first evaluated in the 1940s for viability at a commercial scale at the Bitumont site, located approximately 89 km north of Ft. McMurray (Sheppard, 1989). Effective clay dispersion is required to optimize bitumen recovery (Kaminsky, 2008; Masliyah et al., 2011; Arinaitwe, 2013).  This produces a tailings waste product consisting primarily of two distinct components: coarse sands (predominantly quartz minerals) and fluid fine tailings (a mixture of clay minerals, fine grained quartz minerals, residual bitumen, and ion rich process water).  While coarse tailings may be used for embankment construction, fluid fine tailings exhibit little or no consolidation necessitating construction of large tailings facilities to impound this waste stream.  A material balance completed by Scott et al. (1985) indicates that producing 1 m3 of bitumen requires approximately 4.8 m3 of oil sand ore, and 8.6 m3 of water (85% of which  4  comes from recycled process water) and generates 4.5 m3 of sand and 7.6 m3 of thin fine tailings (formerly referred to as “thin sludge” in the industry).  After approximately two years of storage, larger particles contained in the thin fine tailings stream settle to achieve a solids content of approximately 30%  leaving the bulk of the impoundment’s volume occupied by widely dispersed fine clay minerals (FTFC, 1995; Wells, 2011).  Materials achieving solids content in this range are typically categorized as mature fine tailings (MFT).  Figure 1.1 provides a schematic of a typical oil sands tailings pond.  Figure 1.1: Generalized Cross Section of Typical Oil Sands Tailings Ponds (after MacKinnon, Dusseault et al, and Carrier in Kasperski, 1992)   Kaolinite and illite predominate the clay mineral fraction found in MFT and typically carry a net negative surface charge at the basic pH (8.2-8.5) of the hot water extraction process (Kaminsky, 2011; Omotoso, 2011).  Industry is aware that reduction in pH generally promotes agglomeration of negatively charged clay particles (FTFC, 1995; Ferrera and Pawlik, 2009; Masliyah et al, 2011).  However, reducing the pH of oil sands process water is not viewed by industry as a viable means of addressing challenges associated with dispersed clays.  Costs associated with the chemical reagents and treatment systems that would be required for pH manipulation and management are important considerations for surface mine operators facing challenges associated with management of MFT inventories.  5  Clay dispersion in the fine tailings stream results in the retention of significant volumes of water (up to 70% by volume) within tailings ponds (McEachern, 2012).  This produces a consistent increase in MFT inventories, which by 2018 are projected to reach approximately 2 billion m3 (Figure 1.2).  Figure 1.2: Projected Accumulation of MFT at Oil Sands Surface Mining Projects Already Permitted (after McEachern 2012)   Consistent accumulation of MFT volumes resulted in implementation of Directive 74 by the Alberta Energy Regulator (AER), formerly known as the Energy Resource and Conservation Board (ERCB), between February 2009 and March 2015.  While Directive 74 has been superseded by the Tailings Management Framework (AER, 2015) both regulators and operators continue to explore options and opportunities to limit further accumulation of significant volumes of MFT on the leases of surface mine operators. One method explored by industry is the use of anionic polymers to agglomerate dispersed clay minerals in MFT.  While it appears that this combination may incrementally increase volumes of process water available for reuse in bitumen extraction processes, this research will investigate the factors that impact whether the resulting tailings solids can: a) achieve more effective  6  consolidation than MFT; b) develop a fabric that supports continued long term dewatering; and c) be rendered geotechnically stable to sustain long-term reclamation and inclusion in closure landscapes that do not threaten human or animal life. 1.3 Importance of This Study MFT results when fine tailings produced during hot water extraction of bitumen achieve a solids content that typically plateaus at 30% after a minimum two-year sedimentation period.  Use of chemicals to enhance agglomeration and flocculation of clays has logical merit from at least two fronts: First, as a means of causing dispersed clays to aggregate thereby potentially increasing the amount of water able to be recycled for use in the extraction process.  Second, addition of a material targeting the so-called “bad acting” clay minerals and causing them to aggregate could in theory enable these minerals to form stronger flocs that might potentially improve overall geotechnical material strength.  However, as will be investigated in this research, complexity of both MFT and the type of polymer commonly used in industry to flocculate MFT precludes them from being regarded as discrete additive units.  On the contrary, this research will demonstrate how these components combine in interesting ways to produce an entirely new geo-material with its own complex physico-chemical properties and fabric. Research described in this thesis investigates the fundamental properties governing the physico-chemical properties of both flocculated and unflocculated MFT from a single source.  The research explores whether the use of polymers truly enhances management of fluid fine tailings by creating a material fabric that can be effectively dewatered and develop the geotechnical strength characteristics required for this material to be used as reclamation media in dry closure landscapes (CAPP, 2013; Suncor, 2013).  By using a fundamental approach, this research aids investigation into how polymers, themselves complex materials, can influence or interrupt the ability of ultra-fines present in MFT to either agglomerate or remain electrochemically unaltered after polymer  7  addition.  The shear strength and effect of normal load on PA-MFT were also investigated to identify key parameters needed to develop effective performance criteria for this material.  Investigation into the micro and macro structure of PA-MFT is also explored to provide greater insight into physical mechanisms governing initial and longer-term dewatering of PA-MFT and into mechanisms limiting the thickness at which deposited PA-MFT can be effectively dewatered. The results of this study demonstrate that effective dewatering of PA-MFT is severely constrained by the thickness of the deposited lift and identifies limits to the effectiveness of atmospheric drying and underdrainage.  Consequently, treatment processes that solely use addition of anionic PAM polymer may continue to be feasible for sites with significant real estate for thin lift drying, but even those sites will find themselves increasingly challenged by projected and desired increases in bitumen production which directly translates into a need for greater capacity to manage even larger rates of tailings generation.  Creation of PA-MFT at commercial scale will also likely translate into continued elevated tailings management costs as rehandling of material that has not been effectively dewatered incurs additional costs that can be quite substantial. Mikula (2012) suggests that key elements to better understanding tailings behavior are not widely known outside of the research community especially since a significant portion of this information is held outside of the public domain.  As such, this research contributes to the public domain a comprehensive suite of characterization information that facilitates increased understanding of the key components of geo-materials that result from proposed amendment of input MFTs. Characterization of fluid fine tailings completed by Scott et al (1985) and the Fine Tailings Fundamentals Consortium in the 1980s and 1990s (FTFC, 1995) suggest that the addition of polymers to MFT would not result in enhanced consolidation of ultra-fines.  At the present time a  8  range of polymers with characteristics similar to commercially available polymers like Magnafloc 1011 (anionic) and FLOPAM A-3338 (anionic) are used to flocculate fine tailings that are subsequently dewatered using air drying or a centrifuge (Alamgir, 2010; Wells et al., 2011).  These methods have operational limitations associated with the thickness of lifts needed to achieve effective air drying. The energy/emissions profile associated with use of centrifuges and filter presses capable of keeping pace with fine tailings production is also a consideration for operators.  While PA-MFT produces measurements of shear strength, the roles of polymer and the resulting fabric of PA-MFT in producing measured values are unclear (Masala et al., 2013).  The role played by residual bitumen is also unclear.  Consequently, understanding the various factors impacting the shear strength and rheology of PA-MFT is important if the performance of this material is to be rigorously evaluated. Understanding the near and long-term geotechnical consolidation of flocculated tailings is required to develop effective tailings management strategies and to optimize utilization of areas designated for tailings storage.  Oil sands operators need to identify a means of creating a tailings product that can be consistently produced and managed, and can keep pace with predicted increases in tailings production. 1.4 Problem Statement In 2010 tailings ponds in the Athabasca oil sands region covered approximately 176 km2 and stored approximately 830 million m3 of MFT (AB Environment, 2012).  Such large inventories of fluid fine tailings represent significant risks to oil sands operators.  These risks exist primarily in the form of social, environmental and economic liabilities associated with the management, operation and closure of these facilities.  Moreover, effective characterization and subsequent management of oil sands fluid fine tailings continues to represent one of the most significant challenges to cost effective development of the Canadian oil sands (Mikula, 2012).  Failure of  9  widely dispersed clay minerals to significantly consolidate and be cost effectively removed from the fluid column of fine tailings streams has historically resulted in the creation of large impoundments that, when viewed from the air, look more like lakes than conventional tailings impoundments in the metal mining industry.  The scale of these tailings storage facilities and the perception that they contain millions of cubic meters of toxic waste that is harmful in the near and long term to bird, aquatic and human health are examples of the intertwined social and environmental risks associated with these facilities.  These potential liabilities are further compounded by fears associated with the potential for some form of uncontrolled breach or worse, catastrophic failure resulting in an uncontrolled release of significant volumes of these impounded tailings into the major river systems in the region.  Consequently, images of large tailings impoundments containing perceived toxic residue from the bitumen extraction process have come to symbolize many of the fears and concerns held by individuals and environmental groups about continued development and expansion of the Alberta oil sands industry (Pembina, 2013).  Large MFT inventories raise questions about oil sands producers’ ability to reclaim the landscape and habitat of the Wood Buffalo region to equivalent, self-sustaining and productive land use after exploitation of oil sands deposits using surface mining methods. Methods purported by industry to show the greatest promise of reducing fine tailings inventories involve addition of thickening agents (like gypsum), or anionic PAM polymers to flocculate fines from colloidal MFT (Mikula et al., 2009; Boxill and Hooshiar, 2012).  Thickened or flocculated tailings are then subjected to centrifugation or atmospheric drying to achieve effective dewatering of the resulting material.  Examples include Suncor’s Tailings Recovery Operations (TRO) publicly announced in 2009 (Wells and Riley, 2007), Shell’s Atmospheric Fine  10  Drying (AFD) program (Matthews and Masala, 2009), and Syncrude’s use of In-line Thickened Tailings (ILTT) (Jeeravipoolvarn, 2010; Beier et al., 2013). However, PA-MFT still appears to exhibit some of the same challenges with limited or no consolidation exhibited by MFT.  Consequently a means of quantifying the geotechnical engineering performance of PA-MFT compared to MFT is required.  Additionally, further inquiry into the role played by the resulting fabric and residual bitumen is needed to better understand how these facts effect overall PA-MFT dewaterability. Current operations realities suggest that more work is needed to optimize long-term dewatering and to reduce the significant costs associated with creating and managing flocculated MFT.  Failure to produce a sufficiently dewatered product poses a significant threat to the health and safety of humans and mammals as an ineffectively dewatered material essentially functions like a “quick mud”, borrowing from the idea of a quick sand with the consistency of the thick mud.  The failure of flocculated MFT to develop the shear strength required to support reclamation activities also violates the fundamental objective of being able to effectively reclaim fine tailings disposal sites and poses significant challenges to the effective inclusion of these materials in closure landscapes without use of robust and expensive engineered cover systems. 1.5 Research Objectives The objective of the proposed research program is to provide an understanding of how fundamental properties of PA-MFT, namely surface charge and fabric, govern its micro and meso scale behavior as evidenced through the material’s rheological and geotechnical engineering characteristics, namely its consolidation and strength gain.  This overall objective will be achieved if the following supporting objectives are completed: 1. Investigate how addition of BASF 5250 (anionic polyacrylamide polymer) changes the colloidal stability of raw MFT.  11  2. Identify functional groups on PA-MFT surfaces. 3. Produce 3-D imaging of PA-MFT fabric. 4. Investigate effects of handling and storage condition on shear strength. 5. Quantify evaporative & underdrainage losses & measure surface deformation during drying of thicker lift PA-MFT. 6. Quantify shear strength development in thicker lift deposits with and without load. 7. Develop qualitative understanding of interplay between PA-MFT chemical & physical characteristics by using a consistent characterization framework. Ultimately it is hoped that findings from this work will support future comprehensive evaluation of the true and total costs associated with the use of similar polymer amendments to manage oil sands fine tailings.  It is also hoped that findings from this research will enhance the fundamental understanding about the interplay of physical and chemical characteristics of PA-MFT (Figure 1.3) and inform future investigation into the modifications that would be required to enable these materials to be consistently dewatered to the degree required for their sustainable inclusion in closure landscapes. While a wide range of material characterization has been already completed on different types of MFT, polymers, and even flocculated MFT, this research combines characterization data from the fields of surface chemistry, rheology, clay mineralogy, computer tomography, and geotechnical engineering and considers how these characteristics interact to effect dewatering behavior.  By using a more comprehensive and repeatable testing framework it is hoped that a more complete assessment of factors influencing and governing the overall behaviour of PA-MFT from known sources of MFT, polymer and process water can be quantified and its total cost considerations better understood.  12    Figure 1.3: Fundamental Material Characteristics to Be Investigated Showing Hypothetical Relationship  It is hoped that findings from this research will not only enhance characterization and understanding of PA-MFT engineering properties but that it will support a continued bridging of the critical gap that persists between the domains of chemical, mineralogical, and geotechnical characterization of PA-MFT.    13  Chapter 2: Literature Review 2.1 Overview of Desirable Tailings Characteristics and Management Approaches Tailings typically consist of gangue and residual process effluent (water and chemicals) discharged from the mill process after the bulk of the commodity of economic interest has been removed.  In hard rock mining, ore is typically subjected to a crushing and grinding circuit prior to the addition of chemicals and air during flotation.  The resulting tailings consist of crushed and pulverized material deemed to be waste (based on the mineral economics at the time of deposition) ranging in size from about 2 mm to less than 2 µm.  Fine tailings typically refer to the portion of the tailings stream that is silt sized or smaller (usually less than 60 µm). Where site topography, environmental management considerations, and project economics support the establishment of a permanent tailings storage facility (TSF), tailings are subaerially deposited from spigots as part of a total tailings stream. Under desirable deposition conditions this results in the accumulation of the bulk of finer materials away from the active crest of the tailings impoundment and closer to a supernatant pond that tends to form.  When tailings deposition and particle segregation is effective, the bulk of saturated, low strength, fine materials exist adjacent to the supernatant pond, and are separated from the tailings embankment by a significant above water beach comprised primarily of coarser materials, and a zone of transition materials that are larger than the fines deposited with the total tailings. At other sites where dispersion of fines impacts effective utilization of TSF storage capacity or limits the amount of water that can be reused in the extraction process, chemical reagents like lime, gypsum and polymers may be added to enhance fine particle agglomeration or to promote creation of larger flocs to increase the fines settling rate.  However, at some sites, especially those with limited real estate for tailings management or where water loss to tailings must be minimized, methods to consistently thicken tailings using physical and/or chemical means  14  have been investigated and implemented.  While conventional thickeners have been used to improve dewatering rates of fine tailings since the early 1970s (Robinsky, 2000), a wide range of physical/mechanical methods including cycloning, centrifugation, and filtration have also been used.  These methods are more energy intensive and cost more to operate than conventional spigotted deposition but can result in the creation of TSFs that are essentially dry and in which fine tailings can be effectively co-disposed with coarser tailings and waste rock (Veillette and Davies, 2007).  These methods may be enhanced through use of a chemical additive to the input tailings stream.  As such the engineering behavior of tailings is generally impacted by the extent to which they are thickened and how tailings are deposited (Engels, 2014).  However, a key point to note is that in metal mining, TSFs are typically designed to provide the storage capacity needed to accommodate the tailings solids that will be generated during the permitted mine life.  Barring a change in economics that results in a reclassification of tailings waste as ore, TSFs are generally understood to exist in perpetuity once they have been constructed.  Consequently, effective tailings stewardship focusses on reducing the risk of failure of these facilities including limiting the amount of water stored within the impoundment to reduce the risk associated with a range of failure modes that could result in release of tailings to the environment. 2.2 Fluid Fines Tailings in the Alberta Oil Sands The six oil sands surface mining operations active in 2014 produced a combined total of 867,000 bpd in 2011, with most mines producing upwards of 100,000 bpd (AB Energy, 2013).  The combined daily bitumen output from all surface mine operations translates into daily production of a cumulative volume of approximately 200,000 m3 of thin fine tailings (Mikula et al., 2008; ERCB, 2012).  While each oil sands surface mine producing bitumen froth strives to optimize bitumen liberation from the ores found on their leases using various additives, the hot water extraction method pioneered by Dr. Karl Clark in the 1920s persists as the primary bitumen  15  extraction process used (Masliyah et al, 2011; AB Energy, 2013).  This process, also referred to as Clark hot water extraction (CHWE), involves mixing oil sands ore that has been crushed and screened with water heated to a minimum of 50°C to which dispersive additives like sodium hydroxide are added during hydro-transport of the slurry to the primary and secondary flotation cells (conditioning).  However, it should be noted that other solvent and non-solvent extraction methods have been and are being investigated to mitigate the effects of the adsorption of bitumen to clay minerals in oil sands ores (FTFC, 1995; Hooshiar, 2011). Well before the production of bitumen from the Alberta oil sands was commercialized, Dr. Clark identified challenges associated with clay minerals related to bitumen liberation and process water quality.  In a 1948 letter advising against the subsequent mixing of extraction water with oil sands pulp, Dr. Clark wrote: “Clay, real clay, is what will remain in suspension in the plant water and tend to build up.  And it is the clay in the pulp that knocks the yield (of bitumen).  I feel that water going into the pulp should be fresh water.  There are two advantages.  The main one is that it will not introduce clay where clay does harm.  The other is that it will slow down the accumulation of clay in the plant water.” (Sheppard, 1989). Industry widely accepts that challenges associated with clay minerals in oil sands ores are not limited to bitumen liberation (Kaminsky, 2008; Wallace, 2011; Hooshiar, 2011) but also play a role in the consolidation characteristics of fine tailings (Mikula, 1993a; Chalaturnyk et al., 2002; Cabrera et al., 2009; Miller, 2010; Wells, 2011).  The fluid fines tailings stream produced during the CHWE process consists of dispersed clay minerals and residual bitumen (2-4% by weight) suspended in ion rich process water with a pH ranging between 7 and 9 (Kessick, 1979; Masliyah et al., 2011; Gamal al-din and Liu, 2012).  Kaspersky (1992) and Miller (2010) provide a comprehensive description and assessment of the chemical composition of water typically found in oil sands tailings ponds, referred to in industry as “primary extraction water” (PEW).    16  Clay suspensions, like those found in oil sands fluid fine tailings, are a type of colloidal dispersion technically classified as a sol.  A sol consists of dispersed solid particles in a continuous liquid phase.  Clay minerals, especially those of ultra-fine size (<0.2 µm), dispersed in fluid fine tailings have been identified as key component in gel formation, water holding capacity, and overall behaviour of MFT (Kotlyar et al., 1993a; Sheeran, 1993; Mikula, 2012).  Charges on the surfaces of these clay minerals are directly impacted by their ionic environment and the pH of PEW discharged to and recycled from tailings ponds.  Elevated pH combined with the presence of kosmotropic ions like Na+ and HCO3- limit the ability of clay minerals to achieve their lowest energy state which would enhance agglomeration and consolidation of clay minerals.  Scott et al. (1985) and Jeeravipoolvarn (2010) suggested that limited compressibility of MFT was attributed to very low hydraulic conductivity and the material’s high thixotropic strength. Fundamental tailings research also suggests that the elevated capacity of oil sands fine tailings to hold water is attributed to the presence of a three-dimensional gel network structure and the presence of organic compounds not directly associated with bitumen (Scott et al., 1985; Kasperski, 1992; FTFC, 1995).  The amount of time required for MFT formation and the associated gelation was thought to depend on the initial concentration and particle size of ultra-fines (Kotlyar et al., 1998) as well as the type and concentration of ions present in the process water (Miller, 2010).  Scott et al. (1985) confirmed that after achieving a solids concentration of approximately 30% after about two years, rates of consolidation of this tailings stream slowed considerably.  In the early 1990s many researchers viewed this tailings stream as “resistant to consolidation” (Kasperski, 1992).  Consequently this aged fine tailings stream, referred to in industry as sludge during fundamental oil sands tailings research completed in the 1980s and 1990s, became known as mature fine tailings (MFT), referring to the stability of their composition once formed and the  17  persistently unconsolidating fines.  Fundamental tailings research completed in the early 1990s suggested that the primary problem to be solved in the industry was to find an “economical and environmentally acceptable treatment and disposal of tailings resulting from the Clark hot water process” (Kasperski, 1992).  Work by Znidarcic et al. (2011) investigating use of seepage induced consolidation testing (SICT) on MFT, indicated it could take in excess of 100 years to initiate the process of self-weight consolidation in these materials.  This finding suggests that the challenge of managing fluid fine tailings is likely to continue for a significant period which is consistent with conclusions drawn by Neff and Hagemann (2007), Devenny (2009), WWF (2010), Mikula (2012), and Morgenstern (2012). A review of the use of thickening technologies to manage fine tailings waste suggests that a robust waste management plan includes trade-offs enabling both operational and closure objectives to be satisfied (Boxill and Hooshiar, 2012).  The continuing accumulation of vast quantities of MFT suggests that efforts to mitigate the occurrence of slowly consolidating gels in fluid fines tailings ponds has not kept pace with efforts to optimize bitumen production (WWF, 2010; McEachern, 2012).  The direct link between measures taken to optimize bitumen liberation and challenges associated with managing the fine tailings waste stream cannot be overlooked.  Each oil sands surface mining operation must therefore be viewed as a complete and integrated life cycle process with overall success dependent on the effective handoffs between the critical elements of mining, bitumen extraction, tailings management and reclamation. Prior to Directive 74, oil sands surface mine operators placed a premium on increasing rates of bitumen production and effectively minimized fine tailings management costs by storing these materials in large tailings impoundments (George, 2012).  While it is true that research to identify ways to promote effective consolidation of fine tailings has been ongoing since the 1970s,  18  the pace of development in oil sands industry has not been materially impacted by the inability to effectively mitigate the short and long-term risks associated with accumulation of large inventories of MFT (Figure 1.2). Cost effective management of existing MFT inventories and continuing accumulation of large volumes of fine tailings is essential if oil sands surface mines are to remain financially viable.  Fine tailings management methods must also require minimal additional energy input and ideally not contribute to greenhouse gas (GHG) emissions.  As such, interest in identifying potentially “game changing” tailings dewatering technologies remains high with many of the front running studies utilizing flocculation and thickening technologies (Jeeravipoolvarn, 2008; Beier, 2013). 2.2.1 Composition of Mature Fine Tailings and Identification of “Bad Actors” Actual composition of oil sands ores varies widely both on a single site and across the different leases managed by the various surface mine operators.  This variability has a notable impact on the heterogeneity exhibited by the tailings which result, even within in single tailings pond.  MFT is a thixotropic fluid that occupies the majority of the storage capacity of TSFs in the Alberta oil sands.  Clay minerals removed from oil sands ore tends to accumulate in this waste stream such that on average, 70% of the solids fraction consists of kaolinite, mica, and mixed layer clays (Omotoso, 2011).  Yong and Sethi (1978) also provided valuable insight into component identification and characterization of fine tailings obtained from Suncor tailings ponds. Fluid fine tailings accumulate at a rate of more than 20% of the volume of oil sands initially mined (Kasperski, 1992; Mikula et al., 2008a; Masliyah et al., 2011).  They primarily consist of 20-30% fines, more than 70% of which have an effective diameter smaller than 2 µm.  The rate of tailings accumulation is further compounded by the fact that as much as 70% of water contained in these tailings cannot be recycled or reused in the extraction process due to elevated clay content (McEachern, 2012).  Laboratory research by Sheeran (1993) suggested that fines, especially ultra- 19  fines particles with diameter ranging between 0.01 µm and 1 µm, were responsible for “stable, intractable fine tailings deposits.” Ultra-fines are clay minerals contained in oil sands ores consisting primarily of kaolinite and mica (Kotlyar et al, 1993b) with increasing propensity to form gels as particle size decreases (FTFC, 1995).  It is estimated that up to 3% of the solids in the total tailings stream is comprised of particles smaller than 0.2 µm (McEachern, 2012).  Ultra-fines were identified as the component of fine tailings responsible for binding water within these materials and consequently received significant attention during fundamental fine tailings research completed prior to 1993 (FTFC, 1995).  Kotlyar et al (1995b) demonstrated that ultrafines (classified by these Kotlyar’s team as particles smaller than 0.3 µm) form gels enabling MFT to develop a structure that impedes consolidation.  That research also demonstrated that in the presence of multivalent cations (like Ca2+) these solids supported formation of large, porous aggregates that indicated initially quick settling, but retained significant water within their aggregate structure and resisted consolidation. The surfaces of ultra-fines were found to be “highly charged” and consequently very active (Mikula et al, 1993a; FTFC, 1995).  The elevated pH of the bitumen extraction process also sustains dispersion of clay minerals.  Some researchers also suggest that the capacity of MFT to retain water and its limited ability to consolidate is governed by the surface properties of the minerals (BGC, 2010; Omotoso and Melanson, 2014).  As a result, the importance of understanding the role of clay minerals and their surface chemistry in affecting the observed behavior of un-amended and flocculated fine tailings streams has been stressed by Kaminsky (2008 and 2011), and Mikula (2011 and 2012).  20  2.2.2 Fine Tailings Management Using Chemical Additives The combined volume and rate of fine tailings generated at oil sands surface mines suggests the need for robust economic tailings management solutions compatible with sustained high bitumen production rates that are projected to increase under favorable economic conditions (CERI, 2014).  Industry has experimented with a range of methods to agglomerate and flocculate fine tailings solids to increase their rate of settling and consolidation.  Effective treatment solutions would have the dual benefit of increasing the volume of water available for reuse in the extraction process and improving fine solids storage utilization in TSFs.  Treatment/amendment methods investigated for use with fine Alberta oil sands tailings include physical, chemical, biological, electrical, co-disposal and subaqueous storage, or natural processes (FTFC, 1995; Fourie et al., 2007; BGC, 2010; Boxill and Hooshiar, 2012). Caughill (1992) investigated use of coagulants and flocculants to consolidate tailings and found that polymers with high molecular weight did not significantly improve the MFT consolidation rate.  This finding validated the caution raised by Scott et al (1985) that polymers did not necessarily enhance consolidation of MFT beyond the consolidation rates naturally achieved with time.  Consequently, coagulants like gypsum and lime were investigated as a means of promoting fines agglomeration (Devenny, 2010).  Coagulants aggregate particles by changing the strength of the electrolyte solution which affects the surface charge of clay minerals.  Coagulants like gypsum were used to aggregate clay minerals to produce consolidated tailings.  However, Miller (2010) found that even though calcium sulphate (a coagulant) was observed to improve the initial consolidation rate of caustic tailings, long-term settlement appeared to converge to values similar to those observed in MFT samples in which coagulant was not used.  Consequently, in the Alberta oil sands coagulants like calcium oxide, gypsum, and lime, are often used in combination with polymers to enhance particle aggregation and water liberation.  21  The consolidated/composite tailings (CT) process involves addition of gypsum to promote the agglomeration of tailings fines in MFT to the coarse tailings fraction (sands) using a sands to fines ratio (SFR) of 5.5:1 (Lawrence and Ali, 2010).  This method was first piloted at Suncor in 1994 (Mikula et al., 2008b) with Syncrude conducting a field demonstration of the process in 1995 (Matthews et al., 2002).  While this process, sometimes referred to as MFT-CT, would result in a product with approximately 17% fines (Jeeravipoolvarn, 2010) the addition of gypsum resulted in the accumulation of Ca2+ and SO42- ions in the recycled water which impeded effective bitumen extraction. Syncrude subsequently piloted a modified version of the CT process in 2004, known as in line thickened tailings (ILTT)-CT to limit ionic buildup in tailings recycle water.  ILTT-CT involves in-line thickening of cyclone overflow from processing total tailings with coagulants and flocculants.  The thickened fine tailings are then combined with coarse tailings (Matthew et al., 2002) producing a material with up to 20 times the hydraulic conductivity of MFT (Jeeravipoolvarn et al., 2008).  Creation of ILTT-CT also requires a lower SFR (4:1) than the original CT method and produces a tailings product containing 68-70% solids by weight (Matthews et al., 2002; Sobkowicz, 2010).  ILTT-CT was also designed to consolidate faster than MFT-CT (Jeeravipoolvarn et al., 2010) enabling greater use of warm recycled water which would reduce GHG emissions associated with heating water for the bitumen extraction process. Both CT methods continue to be used by oil sands operators to manage some portion of their fine tailings stream even though its use is constrained by the process’ high demand for coarse sands that are required for TSF embankment construction.  However, neither process is able to keep pace with the rate of MFT production (Jeeravipoolvarn et al., 2010).  In addition, careful engineering and operations controls are required to prevent material segregation during tailings  22  deposition (Jeeravipoolvarn et al., 2010).  For these reasons oil sands surface mine operators continue to explore other methods of sequestering and consolidating fine tailings. 2.2.3 Use of Anionic Polymers to Aid Fine Tailings Dewatering In 2009 Suncor publicly announced field scale trials of its tailings recovery operations (TRO) process (Suncor, 2013).  In this process an anionic PAM polymer is added to MFT using an in-line thickening process (Suncor, 2011; Suncor, 2013).  While polymer dosage has been optimized to promote rapid initial water release based on the clay water ratio of MFT feedstock (Demoz et al., 2010; Suncor, 2011; Powter et al., 2011), effective ultimate dewatering of this material is achieved using natural evaporation processes.  Shell Canada also piloted a similar process called atmospheric fines drying (AFD) in 2010 (Vanderklippe, 2010).  Both processes demonstrated each company’s response to Directive 74 requirements to enhance fines sequestration and to develop materials with undrained shear strength of approximately 5 kPa one year after deposition.  A generalized description of the characteristics of flocculated fine tailings used in TRO is provided by Suncor (2011) and Charlebois (2012).  It is important to note that the quantity and degree of initial dewatering is related to flocculant dosage based on clay content (Wells et al., 2011) and shear conditions during polymer-MFT mixing and during subaerial deposition (Charlebois, 2012). While the results of both TRO and AFD have demonstrated the ability to produce dewatered fines in thin lifts (<20 cm thick), the current and anticipated increased rate of tailings production that would accompany increasing bitumen production suggests that effective management of fine tailings streams requires the ability to deposit polymer amended tailings in thicker lifts (>20 cm thick).  This is especially important if the majority of fines sequestration remains concentrated during the frost and ice free period between June and September in northern  23  Alberta.  Because effective dewatering using evaporation is depth limited, a thorough understanding of the mechanisms governing long-term dewatering of these materials is required. 2.3 Polymer Flocculation Processes Mewis and Wagner (2012) indicate that attractive inter-particle forces can have a significant effect on microstructure and on various suspension properties.  Forces typically present in flocculated systems are summarized in Table 2.1. Table 2.1: Forces that Exist in Flocculated Suspensions (after Scales, 2013) Attractive Forces Repulsive Forces van der Waals Electrostatic Polymer bridging Steric Polymer depletion Hydration Electrostatic  Hydrophobicity   Polymers contain assemblages of smaller molecules (monomers) that are manipulated by chemists to produce unique physical and chemical properties.  Polymers (also known as flocculants) generally function to destabilize colloidal systems by creating surfaces on which formerly dispersed particles can become chemically bonded (Mewis and Wagner, 2012).  Consequently, a polymer’s molecular weight, charge, and charge density are typically manipulated to achieve certain characteristics when the polymer is added.  In the Alberta oil sands, it appears as though the primary short term benefit resulting from the addition of anionic PAM polymer to MFT is the near instantaneous release of up to 40% of water previously occupied by dispersed clay minerals under carefully controlled mixing conditions (Figure 3.4). The physical structure of the polymer also impacts the structure of the flocs they form (Figure 2.1).  Ferrera and Pawlik (2009) found that the degree of polymer anionicity also effected  24  flocculation effectiveness.  A polymer’s radius of gyration affects the range of its attraction to particles, and its concentration affects the strength characteristics of the flocculated suspension (Mewis and Wagner, 2012).  Tripathy and Ranjan (2006) found that charge neutralization was not required for polymers to function as flocculating agents.  This appears to be an important consideration when considering the mechanism that may be at play when anionic polymers are mixed with MFT containing a significant quantity of clay minerals that are generally thought to have a net negative surface charge.  Figure 2.1: Various Arrangements of Polymer Flocculants (after Mahmoudkhani et al, 2012) Klein (2014) proposed that when an anionic polymer with a high molecular weight and low charge density is combined with MFT, bridging flocculation (Figure 2.2) is the flocculation mechanism that dominates.  25   Figure 2.2: Schematic of Bridging Flocculation (from Zeta-Meter, 1993)  In bridging flocculation, the polymer creates a physical bridge between mineral particles.  The higher the molecular weight of the polymer the longer the molecule chains which provides increased polymer chain surfaces on which colloids may attach (Zeta-Meter, 1993).  The high concentration of molecular chains enables polymer branches to attach to two or more particles beneath the electrostatic barrier surrounding the mineral particles (Xu and Hamza, 2003).  An anionic polymer utilizing bridging flocculation is typically used after an inorganic coagulant like alum has been added, as the alum neutralizes colloidal surface charges and creates a micro floc that can attach to polymer bridges (Zeta-Meter, 1993).  However, concerns about the potential effects of inorganic coagulants on overall recycle and seepage water quality persist (Munoz et al., 2010).  Polymer overdosing can also lead to steric repulsion between polymer strands which can inhibit creation of effective polymer bridges between mineral particles (Gregory, 1988).   Physical mixing also plays an important role in ensuring adequate dispersion of the polymer into MFT.  However, as will be described during the experimental section of this thesis, over mixing/shearing can result in floc breakage.  While an optimum mixing intensity can be identified for various flocculating systems (Yeung et al., 1997; Blanco et al., 2005; Sworska et al., 2000), heterogeneity of the input MFT, distribution of particle sizes, clay mineralogy and residual  26  bitumen, all impact the ultimate effectiveness of polymer bridging that results with the applied mixing energy.  The nature of polymer conformation impacts overall utilization of the polymer to bridge particles (i.e. the number of particles that can ultimately be bridged by each polymer chain), and the degree to which the polymer chain can itself hinder overall flocculation, especially when steric repulsion occurs during polymer overdosing.  Electrostatic repulsion between the charged segments in charged polymers causes them to exhibit a more extended conformation than a polymer without a charge (Adachi, 1995).  Consequently, having an extended confirmation increases a polymer’s hydrodynamic radius which in turn has the potential to increase the polymer’s effectives as a bridge between particles.  Polymer-particle adsorption is inhibited in particle systems that have the same charge as the polymer, especially when the polymer has low molecular weight and high charge density (Vorob’ev et al., 2008; Hogg, 1999). de Krester et al. (2003), Masliyah et al. (2011), and Mewis and Wagner (2012) indicate that very low polymer concentrations can induce bridging flocculation in which the polymer adsorbs simultaneously resulting in the flocculation of fine clay particles.  These sources also indicate that addition of too much polymer results in excess amounts of free polymer in solution which can result in the phenomena of depletion attraction.  Depletion attraction produces a material that is poorly flocculated and has its behavior dominated by the ionic properties of the polymer itself.  Additionally, polymer overdosing can cause flocs to create space-filling structures that form a gel (Mewis and Wagner, 2012; Scales, 2013). The gels that result from combining anionic low charge density but high molecular weight PAM exhibit solid-like behaviors and characteristics like yield stress and viscoelasticity.  This finding has implications for the measurement of shear strength of these materials and the development of intrinsic material strength.  The capacity of these gels to adsorb water (based on  27  the hydrophilic nature of PAM) also confounds efforts to dewater suspensions and creates significant challenges to the development of materials with desirable geotechnical strength characteristics (Yao et al., 2012). 2.4 Ionic Concentration of Primary extraction Water Connate water (from pore spaces in oil sands ore), Athabasca River water (used in bitumen extraction process), and process chemicals (surfactants and polymers) represent the three major sources of dissolved inorganic ions found in tailings ponds (Kasperski, 1992).  Dissolved ions that have been found historically in tailings ponds at both Suncor and Syncrude oil sands surface mines are summarized in Table 2.2. Table 2.2: Ions Present in Suncor and Syncrude Tailings Ponds (after Burchfield and Hepler, 1979)    Sodium is the most abundant cation while bicarbonate is the most abundant anion in Syncrude pond water (MacKinnon, 1989).  Sodium and bicarbonate ions originate from soluble ions found in oil sands ores.  The addition of sodium hydroxide to enhance dispersion of clay minerals to optimize bitumen recovery also increases the concentration of monovalent sodium ions in PEW.  Addition of divalent calcium ions has been considered as a means of limiting the effect of monovalent sodium ions on the clay minerals as the negative clay mineral surfaces appear to demonstrate an affinity for calcium ions (Mikula et al., 2008b).  These factors contribute to the  28  industry wide concern about “ionic buildup” over time, the severity of which appears to vary based on pond depth (Burchfield and Hepler, 1979; Mackinnon and Retallack, 1982; Miller 2010). The kinetics of ultra-fines aggregation in MFT is highly influenced by the concentration of ions present in tailings ponds.  Aggregation kinetics is thought to control the rate of water release during the formation of MFT (FTFC, 1995).  Water chemistry is also thought to control or influence the development of gel strength (Kasperski, 1992; Miller, 2010).  Mikula et al. (2008c) also identified a reciprocal effect of clay mineralogy on water chemistry and the associated settling and rheological properties of the tailings suspension. More recent work by Miller (2010) identified sodium adsorption ratio (SAR) as a good guide for predicting the dispersive nature of fine tailings.  Predictions using this method were consistent with the degree of dispersion measured using a hydrometer.  Miller (2010) found that compressibility of caustic fine tailings exhibited an over-consolidation effect associated with the development of thixotropic strength in MFT and that changes in pore water chemistry impacted Atterberg limits results. 2.5 pH Change in pH towards the point of zero charge can result in colloidal aggregation (Mewis and Wagner, 2012).  Kaolinite aggregates strongly at a pH of 5 when faces of kaolin minerals are negatively charged and particle edges are positively charged (Pawlik, 2011).  This results in electrostatic attraction that creates a “house of cards” aggregated structure (Laxton and Berg, 2006).  Consequently, pH reduction appears to represent a simple strategy for the dewatering of oil sands fine tailings.  Ferrera and Pawlik (2009) confirmed that flocculation of oil sand tailings was more efficient at a pH of 5.5 than at a pH of 8.3 such as typically used in CHWE processes.  The aggregating effect of lower pH observed was believed to be most likely related to a reduction of the surface charges of clay minerals.  29  However, because bitumen liberation is optimized at a pH ranging between approximately 8.3 and 9 and operators seek to optimize the volume of water reused in the extraction process, pH manipulation would both significantly increase consumable costs for ore processing as well as increase the rate of ionic buildup in process water stored in tailings ponds.  The accumulation of ions in the process water would further limit the number of times process water could be recycled before adverse effects on bitumen liberation/recovery were observed. 2.6 Residual Bitumen The role of residual bitumen on the behavior of MFT and on attempts to flocculate this material is an area of increasing investigation (Masliyah et al., 2011; Klein, 2014).  The hindered consolidation model developed by Scott et al (1985) for unflocculated MFT (Figure 2.3) draws attention to the possible role of bitumen in forming a hydrophobic link between clay particles and impeding flow within and through the material.    Figure 2.3: Possible Effect of Bitumen on MFT Permeability Proposed by Scott et al. (1985)   30  Work by Long et al (2006a), using single molecule force microscopy with an atomic force microscope, found that the attachment energy of polyacrylamide and the probability of polymer attachment was lower for bitumen than for mineral surfaces.  Klein (2014) found that reduction in bitumen content between 0.45 wt% and 0.18 wt% increased the initial settling rate of flocculated MFT but that bitumen removal below the lower value actually decreased the observed PA-MFT settling rate.  This decrease in observed settling rate when bitumen content is less than 0.18 wt%  was interpreted as an indication that a certain amount of bitumen was required to provide a certain amount of total organic carbon to work with a hydrolyzed polyacrylamide polymer such as that used to flocculate the sample MFT.26 While bitumen typically constitutes between 1 wt% and 4 wt% in MFTs found in the Alberta oil sands, its ability to coat a significant portion of clay mineral surfaces exists.  Consequently, the ability for bitumen to limit the number of polymer bridging sites as well as to provide a hydrophobic encapsulation of water contained within the fabric of PA-MFT was explored as part of this research. 2.7 Assemblage Structure and Material Fabric The fabric of a material describes both the arrangement of particles, particle groupings or clustering, and the configuration of the spaces between particles (Bennett et al., 1991; Mitchell and Soga, 2005).  The material fabric of flocculated systems is therefore necessarily impacted by the polymer’s initial spatial configuration, the degree to which previously dispersed particles become bonded to the polymer’s receptor sites, and the spatial arrangement of the attached particles. Understanding a material’s physical fabric also enhances understanding of its behaviour and enables development and refinement of analytical models used to predict or indicate anticipated behavior.  Calibrated models that can predict PA-MFT consolidation are critical for assessing not only the amount of consolidation that could likely be achieved, but to evaluate  31  whether desired closure prescriptions are even feasible within desired timeframes.  Scales (2013) noted that particle orientation appears to have a significant impact on suspension behavior.  The nature of the porous network that exists in PA-MFT were investigated as well as inquiry into how related inter-particle geometry affects dewatering characteristics and shear strength.   In addition to noting that most “dewatered’ tailings only achieve the gel point, Scales (2013) suggests that the breakup of clay minerals during the extraction process increases the aspect ratio of these particles which reduces the time for the gel point to be reached.  Once the gel point is achieved, interaction between particles resists gravity and compression (de Krester et al., 2003).  This work also suggests that the packing and spacial arrangement of these particles also influences material behaviour.  Farinato and Dubin (1999) and Scales (2013) hypothesized that the fabric of flocculated fine tailings consisted of disordered aggregates that create tortuous flow paths. Flocculation of MFT exhibits not only sol-gel processes (Duguet, 2012) but the creation of initially soft jammed material (Ovarlez and Coussot, 2007).  The fabric of these soft jammed systems varies considerably from systems in which hard particle interactions occur.  As seen in Figure 2.4 polymers and colloidal particles exhibit “soft interactions” (Coussot, 2005).  Figure 2.4: Hard and Soft Interactions (after Coussot, 2005)   32  Deposition characteristics of well-flocculated MFT indicate that this material consists of macroscopic floc assemblages with diameters ranging between a few millimeters to as much as five centimeters (Charlebois, 2012).  While ideal flocculation optimizes water release and limits polymer waste, the degree of flocculation achieved is a function of surface chemistry, use of appropriate polymer dosage, and consistency of operational systems and controls used to flocculate and deposit materials in designated areas.  Variability in these factors results in the creation and deposition of under-, over- and optimally flocculated MFT within a single lift in a deposition cell within any given deposition cycle.  Consequently the fabric of flocculated MFT deposited in a cell is influenced by both macro fabric (how floc assemblages exist in space relative to other floc assemblages) and micro fabric (the nature of the fabric of individual flocs).  The increased void ratio of flocculated MFT identified by Jeeravipoolvarn (2008) and Yao et al (2012) suggests that PA-MFT macro fabric consists of large pore openings that govern the rate at which water initially released is removed from the amended material.  However, the retention of water within flocculated MFT beyond initially released water suggests that the micro fabric exhibits lower permeability and likely contains smaller pores capable of restricting flow.  The success of air drying (Romero et al., 1999), consolidation rates (Mitchell and Soga, 2005) and depth of freezing (Pusch, 1979), have all been identified as effected by the presence of a bimodal fabric in structured soils. 2.8 Polymer-MFT Interactions Flocculated and gelled suspensions are present across a wide range of industries and result in large capital expenditures estimated in the billion to trillion dollar range (Scales, 2013).  Components capable of forming this type of suspension also occur in nature (Mewis and Wagner, 2012).  Bitumen production at oil sands surface mines using hot water extraction appears to combine elements from both nature and industry to produce large scale gelation of naturally  33  occurring fine clay minerals.  The occurrence of gelation has been identified in bitumen extraction (Arinaitwe, 2013; Omotoso, 2013) and in fluid fine tailings flocculated under several tailings management regimes (Jeeravipoolvarn, 2008; Beier, 2013).  It is therefore reasonable to consider the cost and operational implications of exploiting an ore body rich in naturally occurring minerals with a propensity for creating gelled suspensions at industrial scale. Mechanisms governing the behavior of MFT involve complex and compounded interactions of water chemistry, clay mineralogy, clay mineral size and residual bitumen (Scott et al., 1985; Miller, 2010).  The addition of an adsorbing polymer to an initially complex colloidal suspension has the ability to produce a material that exhibits even more complex behavior further impacted by polymer-particle interactions driven by polymer structure and charges on surfaces of fine clay minerals in suspension (Hiemenz and Rajagopalan, 1997; Farinato and Dubin, 1999; Mewis and Wagner, 2012).  Furthermore, it has been demonstrated that sol-gel processes are possible with the combination of nanoparticles and ions in solution (Kotlyar et al., 1993) and in the presence of polymers can result in the formation of polymerized gels (Duguet, 2013). Flocs or colloidal aggregates resulting from this combination of polymer and clay minerals are usually very open and can be easily broken and reconfigured by applying shear force (Watson et al., 2011; Mewis and Wagner, 2012; de Krester et al., 2003; Scales, 2013).  This results in a complex interplay between microstructure and flow and produces strong non-Newtonian behavior (Mewis and Wagner, 2012).  While the size, shape and compactness of the flocs formed are important descriptors of material fabric, fractals are often used to describe internal aggregate structure.  Fractals are characterized by a self-similar structure which means that the object looks similar regardless of viewing scale or magnification (Mewis and Wagner, 2012).  34  PAM is a common flocculant that has long been used in the water treatment and paper making industries.  This polymer has a very high capacity to absorb water and forms a soft gel-like material when hydrated.  The functional unit of a generic form of this polymer is shown in Figure 2.5.  Figure 2.5: Repeating Functional Unit in a Polyacrylamide Molecule (after Entry et al., 2002)  In the Alberta oil sands, anionic, partially hydrolyzed versions of this polymer have been used to sequester fines (Alamgir et al., 2010; Alagha et al., 2013).  Nabzar and Pefferkorn (1985) found that when hydrolyzed PAM was combined with kaolinite, primary attachment between the clay mineral and polymer occurred on the clay edges with minimal attachment on the basal planes of the clay minerals.  However, Alagha et al (2013) found that Magnafloc (a commercially available partially hydrolyzed PAM) adsorbed strongly to the aluminum-hydroxyl basal planes of kaolinite clay minerals while exhibiting limited, reversible adsorption to the silica basal planes of kaolinite, likely as a result of hydrogen bonding.  Alagha et al. (2013) also suggested that the limited adsorption between the polymer and the silica basal planes resulted from electrostatic repulsion caused by a combination of negative charges on the surfaces of both the polymer and the clay minerals. Long et al. (2006b) found that the addition of hydrolyzed PAM at low dosages, especially in the presence of divalent ions at low concentrations resulted in adhesion interactions.  The combination of this polymer and cations was synergistic such that adhesion forces, especially  35  between fine particles, increased with increasing polymer dose and cation concentration up to an optimal polymer dose and cation concentration (Long et al., 2006b).  Above this optimal polymer dose, adhesive forces decreased and in some instances, repulsive forces were measured.  These findings support previously described conclusions by others asserting the importance of determining appropriate polymer dose. 2.9 Effectiveness of Polymers Used for Fine Tailings Management Although fundamental fine tailings research has demonstrated a limited ability of polymers to improve the consolidation characteristics of flocculated clay minerals (Scott et al., 1985; Caughill, 1992; FTFC, 1995), use of polymers continues to be investigated and tested at commercial scale in the Alberta oil sands at surface mine operations like Syncrude, Suncor, and Shell as previously indicated.  Polymers used to flocculate fine oil sands tailings are typically either long chain or branched polyacrylamide polymers designed to form fractal assemblages with fine particles.  Soane et al (2010) formulated a polymer to selectively bond with quartz and clay minerals in oil sands tailings streams based on mineral surface charges.  A list of the wide range of chemical additives being investigated to enhance MFT consolidation is provided in CTMC (2012). Polymer use in the Alberta oil sands has been found to be most effective when it is combined with atmospheric drying of flocculated tailings deposited in thin (< 0.2 m) lifts (Wells and Riley, 2011; Suncor, 2011).  However, use of these methods present the following challenges to effective management of fine tailings streams: • Need for large drying areas (not available at each lease site); • Limited to seasons when air drying is effective in northern Alberta (optimally June through early October); and  36  • Inability to keep pace with the rate of fine tailings generation.  This results in continued increase of MFT inventories in large tailings ponds. While a range of technologies purport to effectively address the accumulation of fine tailings by increasing the settling rate of fine particles through the addition of polymers, Mikula (2012) notes that none of the reported increases in solids content approach the level of dewatering (80% solids) required to achieve true transition from a liquid tailings product to a functional solid material (Figure 2.6).  Figure 2.6: Relationships between Solids Content, Tailings Volume, and Clay to Water Ratio for Typical Oil Sands Fines Tailings (after Mikula, 2012)   2.10 Total Cost and Technical Considerations for Management of PA-MFT High level analysis of Suncor’s TROTM process (Suncor, 2011; Charlebois, 2012) indicates that production of flocculated MFT for thin lift deposition necessitates the following energy and cost input: • Construction of dedicated disposal areas/containment cells – energy and possible ore sterilization (capital expense);  37  • Pumping MFT from tailings impoundments – energy  (recurring cost and GHG emissions) and infrastructure (capital expense); • MFT conditioning and dilution – process plant infrastructure (capital expense) and energy (recurring cost and GHG emissions); • Polymer, polymer storage and mixing – recurring consumable costs, process plant infrastructure (capital expense) and energy (recurring costs and GHG emissions); • Pumping and piping MFT to deposition cells – energy (recurring cost and GHG emissions) and infrastructure (capital expense); • Pumping and piping polymer to deposition cells – energy (recurring cost and GHG emissions) and infrastructure (capital expense); and Periodic emptying of dedicated disposal areas (DDAs) – energy (costs and GHG emissions) and need for secondary diversion areas to accommodate tailings deposition when DDAs are being cleared. That industrial scale tailings treatment processes like Suncor’s TROTM utilize atmospheric drying and therefore are currently only conducted during snow and ice-free periods in northern Alberta (June–mid October) means that additional storage facilities will likely be required to store MFT to support batch processing of these materials.  Ineffective dewatering also increases material handling costs especially if waste materials cannot be stored in the facility where they will ultimately be reclaimed. Results of thickened tailings trials also suggest that the use of undrained shear strength as stand-alone performance criteria for flocculated tailings, as specified in the now defunct Directive 74, may be ineffective when used as the sole means of confirming when desirable material behavior has been achieved.  Use of shear strength data to characterize material performance is  38  investigated as part of the completed study.  Research into centrifugation completed by Syncrude indicated that the type and dosage of flocculant used, mixing technique, and the type of centrifuge used directly impacted the quality of the dewatered cake. Findings by Yao et al. (2012) and Beier (2013) that flocculated MFT occupies greater volume than unflocculated MFT point to the need for better characterization of the fabric of flocculated MFT and how it enables water retention.  The presence of water within the fabric of flocculated MFT has implications for the effective drying and associated performance of thicker deposits.  These findings also raise questions about the likelihood of flocculated MFT to adsorb additional water during deposition of a subsequent layer of flocculated MFT or from rainfall events.  Questions about persistence of polymer hydrophilicity within deposited flocculated fine tailings are also raised. For polymer-flocculated MFT to be an effective and sustainable means of reducing MFT inventories, this material would need to satisfy the following operational performance criteria: • Enable enough MFT to be flocculated in a season that additional tailings pond capacity is not required to keep pace with ongoing fine tailings production; and • Create a stackable, self-contained deposit that can either be directly reclaimed or easily mined through. 2.11 Rheological and Strength Characterization of Flocculated Oil Sands Tailings 2.11.1 Rheological Characterization Rheology is the study of how materials deform and flow as a consequence of their innate structure and in response to the effects of temperature, applied shear, and the rate at which shearing force is applied.  MFT is a non-Newtonian fluid that exhibits time-dependent reversible thixotropy (thinning with applied shear due to rupturing of interparticle bonds but rethickening with time after shear stress is removed).  Thixotropy, and its rheological counterpart rheopexy, are believed to  39  result from what is happening to interparticle bonds and from changes in the material’s structure in response to the application and relaxation of shear stress with time (Mewis, 1979; Barnes, 2000). Irreversible changes in a material’s viscosity are related to either permanent breakage of structural linkages in the material or permanent strengthening of material assemblages.  Characterization of these properties provides insight into material behaviour both at the time when shear force is applied (e.g. during material pumping, mixing, at deposition, or subsequent handling).  Rheological characterization also provides insight into how shear stress and a material’s environment affect its strength condition with time.  Rheomalaxis refers to irreversible degradation of a material’s viscosity resulting from the shear rate and shear force applied (Schramm, 2005).  This concept has particular significance when considering the long-term effects of material handling, especially on materials whose strength is impacted by its stress environment.  Long-term material performance is also a consideration for materials exhibiting rheomalaxy. 2.11.2 Assessing Impacts of Material Handling Rheological properties are an important component of the overall characterization of oil sands tailings streams as they enable evaluation of the effects which material handling, deposition, and storage under saturated conditions have on shear strength and its development.  Rheological testing on flocculated and non-flocculated concentrated suspensions of ore and tailings from the Alberta oil sands was completed by Gutierrez (2013), Arinaitwe (2013), and Estepho (2014).  Klein (1992), Barnes (2000) and de Krester et al. (2003) provide extended discussion of the empirical models commonly used to describe the flow curves of concentrated suspensions. 2.11.3 Geotechnical Strength Characterization Qiu and Sego (1998) suggest that knowledge of the basic physical properties, i.e. the consolidation and desiccation behavior of fine tailings is necessary to understand tailings behavior and to optimize storage of this waste stream.  As it relates to the use of chemical additives to  40  improve the geotechnical behaviour, Scott et al. (1985) considered the search to identify additives to promote flocculation of MFT as “misguided” based on field and laboratory results that indicated that fines consolidation proceeded at a very slow rate once MFT achieved a solids content of 30% with and without chemical amendments.  This finding was more recently confirmed by Wells et al. (2011) although the ability to air dry even a portion of flocculated fine tailings is widely heralded as an improvement in oil sands fine tailings management. Standard geotechnical characterization of oil sands tailings is consistent with many of the methods applied to natural soils and includes determination of Atterberg limits, drained and undrained strength, permeability, coefficient of consolidation, and compressibility (Sobkowicz and Boswell, 2011; Znidarcic, 2012).  Jeeravipoolvarn (2010) provides a comprehensive account of the various geotechnical characterization methods that have been applied to in-line flocculated oil sands fine tailings.  More recently, Yao et al. (2014) completed research investigating shrinkage and swelling properties of MFT flocculated with a high molecular weight polymer. Solids content, grain size distribution, clay content, mineralogy, and water chemistry have been identified as impacting tailings shear strength (Sridharan et al., 2002; Mitchell and Soga, 2005; Sobkowicz and Morgenstern, 2009; Miller, 2010).  Beier et al. (2013) suggested that the combined impact of these factors, in addition to the presence of bitumen on overall material behaviour, could be evaluated using the liquidity index, IL which is derived using Atterberg limits data.  Using this approach, flocculated MFT from both Syncrude and Suncor were consistently classified as highly plastic clays (Figure 2.7).  41   Figure 2.7: Plasticity Chart for Flocculated Oil Sands Fine Tailings (from Beier et al. 2013)  Locat and Demurs (1988) developed a relationship between undrained shear strength and liquidity index for sensitive clays.  Similar data collected from field trials using flocculated tailings from both Syncrude and Suncor were compared by Beier et al. (2013) against this trend. This indicated that development of undrained shear strength in the 5 kPa to 10 kPa range could not be consistently achieved for samples using atmospheric fines drying from a 2010 trial completed by Shell.  Failure of these samples to achieve desired undrained shear strength characteristics was attributed to problems with the program’s flocculation process and deposition methods used.   However, apart from the cautions and limitations raised by Scott et al. (1985) and Mikula (2012), Yao et al. (2012) and Beier et al (2013) identify flocculated fine tailings as providing the following advantages over unflocculated MFT: • Achieve higher initial dewatering rates; • Satisfy undrained strength criteria first specified in the now defunct Directive 74; and  42  • Provide higher permeability values. These apparent benefits must be reconciled with findings also made by Yao et al. (2012) and Beier et al. (2013) that flocculated MFT has the ability to hold more water (measured in one case to be 0.63 m3/tonne of dry solids) than unflocculated MFT.  Field trials also found that flocculated tailings were often deposited at water contents exceeding their liquid limit (Masala and Matthews, 2010; Jeeravipoolvarn, 2010). This finding led Beier et al. (2013) to conclude that while fine tailings flocculation enabled desired undrained shear strength to be reached at lower solids content than unflocculated MFT, flocculated materials were only partially dewatered and required use of additional methods to achieve desired material performance.  However, findings by Yao et al. (2012) that polymer amended MFT required greater storage volume than unflocculated MFT suggests that characterization methods beyond those typically used in geotechnical engineering may need to be utilized.  Indeed, interpretation of geotechnical measurements of consolidation and shear strength for non-Newtonian materials like MFT and PA-MFT must be carefully considered (Scott et al., 1985).  Results from completed trials (Mikula, 2008; Jeeravipoolvarn, 2008; Charlebois, 2012) suggest that the ability of flocculated MFT to achieve desired material performance and to behave like a more conventional free draining soil requires effective initial and long-term drainage. In the Alberta oil sands, measurements of undrained strength are typically obtained in the field using both cone penetrometer testing and hand held field vanes (Charlebois, 2012).  While use of a cone penetrometer rig is best suited to characterize variability in the profile of deposited PA-MFT, hand-operated field vanes are easy to operate and can be used by field personal to gauge how undrained strength is developing.  Consequently this research focuses on the measurement and interpretation of undrained shear strength values obtained using this field method.    43  Field test programs completed at Shell (Masala and Matthews, 2010; Masala and Dhadli, 2012); Syncrude (Jeeravipoolvarn, 2010) and Suncor (Charlebois, 2012) and more recent laboratory work completed by Scales (2013) suggests that the field vane enables undrained shear strength of flocculated fine tailings to be estimated.  While field vane use in the Alberta oil sands has not been standardized, use of handheld vanes provides distinct advantages associated with their portability, robust calibrated spring loading mechanism, and method simplicity.  Masala and Dhadli (2012) noted that inconsistencies in measured field vane data may be attributed to failure to implement standard operating procedures appropriate for effective vane operation and anisotropy of flocculated tailings.   2.11.3.1 Development of Parametric Undrained Strength Relationships Beier et al (2013) reported Atterberg limit test results of flocculated MFT.  However, the effects of residual bitumen and flocculation on results (namely liquid limit, wL determination and measured shear strength) are unknown and have not been quantified.  The need for correction factors to be applied to field vane measurements in flocculated oil sands tailings is also unknown (Masala and Dhadli, 2012). Use of the fall cone to determine the liquid limit and undrained shear strength (su) of soft materials has been debated in geotechnical engineering community (Prakash and Sridharan, 2006; Zentar et al, 2009; Tanaka et al., 2012).  However, the fundamental point made by Haigh (2012) that the Casagrande percussion test and fall cone method measure different mechanical soil properties is often overlooked.  Haigh (2012) notes that the fall-cone test provides a direct measure of undrained strength while the Casagrande cup provides a measure of specific strength which is defined as a fixed ratio of undrained strength to material density.  Prakash (2005) also noted that the liquid limit of kaolinitic soils is primarily impacted by its particle arrangement which is governed by interparticle/frictional forces and shear resistance at the particle level.  44  Under the assumption that soils exhibit a unique strength at each of its Atterberg limits, Wroth and Wood (1978) proposed the fall cone test as a simple strength measuring device using the method first developed by the Geotechnical Commission of the Swedish State Railways (Hansbo, 1957).  While correlations of penetration depth and undrained strength were developed for the Swedish fall cone using reference to clay standards that are not contaminated with hydrocarbons (Houlsby, 1982; Sherwood, 1970), use of this device in the research described in this thesis provides a consistent means by which the effects of moisture content and fall cone penetration depth could be consistently related for tested samples of PA-MFT.  Use of this method also provided a means by which PA-MFT percent solids information could be considered in conjunction with Plastic Limit and undrained shear strength data. 2.11.3.2 Large Strain Consolidation In geotechnical engineering, consolidation refers to the process by which a material (usually understood to be a natural soil) experiences a loss of water from its voids as it compresses.  The voids in the material are assumed to exist in a network with a degree of connectivity that is quantified in terms of permeability which facilitates the flow and expulsion of water from the material’s pores.  This assumed fabric provides a basis for the assumed validity of Darcy’s law in consolidation theory, which, it should be remembered, was initially developed for homogeneous granular soils and assumes water flow in the sample is laminar, continuous and steady (Whitaker, 1986).  While the applicability of Darcy’s law to homogeneous soils containing clay minerals has been demonstrated (Freeze and Cherry, 1979), assumption of a direct relationship between void space and permeability does not account for the existence of a material fabric that is at once both highly porous but effectively rendered impermeable based on the irregularity in shape, variety in sizes, and tortuosity associated with the material’s voids.  Furthermore, the relationship between velocity and hydraulic gradient is non-linear in extremely fine-grained and colloidal materials in  45  which pore spaces can also be very small.  This pore size effect further constrains easy application of Darcy’s law to materials with these characteristics (Freeze and Cherry, 1979; Fetter, 2001).  In addition to the assumption of the validity of Darcy’s law, Terzaghi’s theory of one dimensional consolidation developed in 1923 (Terzaghi and Peck, 1967) is considered to apply if the following assumptions are also satisfied: 1. The material is completely saturated and all voids are filled with water. 2. The material does not experience lateral strain and is assumed to exist in an infinitely wide layer. 3. Compressibility is not time dependent and no secondary consolidation occurs. 4. Solid particles and water are incompressible. 5. Any strain that exists is very small or infinitesimal. 6. Self-weight effects are ignored; and 7. Permeability and compressibility are related and constant. Based on findings by Znidarcic et al. (2011) and Estepho (2014) during their respective investigations into the consolidation behaviour of MFT, the applicability of at least four of the stated conditions governing applicability of one dimensional small strain consolidation (namely assumptions 4 through 7 above) to PA-MFT is called into question. The Seepage Induced Consolidation Test (SICT) has been successfully used since the 1990s to determine the consolidation properties of soft soil slurries and spoils from dredging operations (Znidarcic et al., 2011).  This testing apparatus was developed to significantly decrease the amount of time required to determine compressibility and permeability parameters for phosphatic clays (Abu-Hejleh and Znidarcic, 1992).  This reduction in time is achieved through  46  use of a flow pump to induce seepage in the sample (Znidarcic et al., 1992).  The SICT enables determination of the following: • Void ratio at zero effective stress (e0) and at very small effective stresses. • Measurement of changes in sample height as load is applied in a controlled and measured way. • Effective stress at the bottom and top of the sample. • Calculation of permeability and void ratio at the final load step. Collected SICT data is analyzed using the Seepage Induced Consolidation Test Analysis (SICTA) program that was developed by Abu-Hejleh and Znidarcic (1992).  SICTA uses large strain consolidation theory described by Gibson et al. (1967) and a compressibility function (Equation 2.1) developed by Liu and Znidarcic (1991) which enables the void ratio to be calculated at zero effective stress and is expressed as follows: 𝒆𝒆 = 𝑨𝑨(𝝈𝝈′ + 𝒁𝒁)𝑩𝑩  Equation 2.1  where the void ratio (e) is given as a function of stress in which the parameters A and B are unitless and Z has units of stress (e.g. N/m2 or Pa based the unit of stress applied).  Somogyi (1979) demonstrated that the permeability function used in SICTA (Equation 2.2) was suitable for low density soils.  This function is expressed as: 𝒌𝒌 = 𝑪𝑪𝒆𝒆𝑫𝑫 Equation 2.2  in which permeability (k) is a function of void ratio (e) and parameters C and D are unitless.  Estepho (2014) provides a comprehensive description of the iteration scheme used to develop the various parameters (A, B, Z, C and D) required to solve the compressibility and permeability functions used in the SICTA routine.  47  While large strain consolidation theory assumes that Darcy’s law is valid, compressibility relationships for test materials can be developed using collected empirical data such that: • Effects of self-weight are accounted for (differs from assumption #6 of small strain consolidation theory mentioned earlier in this section). • No assumptions are made about how sample compressibility and permeability are related and they are not assumed to be constant (contrary to assumption #7 of small strain consolidation theory presented earlier). • No assumptions are made about how changes in effective stress impact void ratio as these parameters are independently calculated. Estepho (2014) successfully commissioned a SICT chamber at the University of British Columbia (UBC) that provided compressibility data comparable to that developed using similar equipment designed by Znidarcic et al. (1992) at the University of Colorado.  The UBC SICT apparatus was used in the research program described in this thesis to compare how the compressibility and permeability functions of PA-MFT compared to similar functions for unflocculated MFT.  The data was also compared to other published MFT consolidation data. Consolidation profiles developed using the computer program CONDES0 (Yao and Znidarcic, 1997) were also compared to actual consolidation measurements obtained from evaporation test cell experiments conducted as part of the research described in this thesis. 2.12 Summary of Key Points from Literature Review Current management of oil sands MFT makes it necessary to store large volumes of water containing dispersed kaolinite rich sediments.  MFT is a thixotropic, colloidal gel in which ultra-fine particles (<0.2 µm in apparent diameter) appear to play a significant role in creating a fabric that binds water within it (FTFC, 1995).  Operations requirements to maximize the use of recycled  48  water while maintaining the pH of process water used for bitumen extraction at a range between 8 and 9, causes the clay-sized minerals present in these tailings storage facilities to exhibit electronegative repulsion forces.  These dispersed clay minerals exist in the form of a colloidally stable gel that limits the amount of water that can be effectively extracted from within the colloidal suspension and recycled for use in extraction.  In an attempt to sequester dispersed fines, some operators have investigated the effect of adding anionic polyacrylamide polymers.  The addition of this type of polymer results in limited aggregation of the clay-sized minerals and produces a material that occupies more volume than untreated MFT (Yao et al, 2014).  Furthermore, this material behaves like a plastic clay (Beier et al, 2013).  These findings appear to support the contention by Scott et al. (1985) that additives to promote flocculation of MFT would have a limited effect, if any.  A conceptual MFT model developed by Scott et al. (1985) also suggested that residual bitumen could adversely impact permeability if it were to block pore throats. Bridging flocculation is the mechanism thought to occur when anionic polyacrylamide polymer is combined with MFT (Klein, 2014).  However, Nabzar and Pefferkorn (1985) found that positively charged clay edges were the sites of primary polymer attachment.  This results in the creation of an open material fabric. MFT flocculation with anionic polyacrylamide has been optimized for use with thin lift (< 20 cm) deposition and air drying (Wells et al., 2011).  This would necessitate the use of large drying areas if this technology were to keep pace with current and projected rates of bitumen production.  In field trials of MFT flocculation with anionic polyacrylamide polymer, the flocculated material is deposited subaerially into dedicated drying areas where it is left to air dry between deposition campaigns.  Initially, the flocculated material appears to be comprised of soft  49  flocs that vary in size and which undergo significant deformation during deposition.  Field vanes have been used to estimate the undrained shear strength of deposited material (Charlebois, 2012).  Fall cone testing has also been effectively used to directly measure the undrained strength of soft soils (Haigh, 2012).  Large strain consolidation theory has been applied by Znidarcic et al. (2011) and Estepho (2014) to investigate the consolidation behavior of MFT.  This testing was completed using a seepage induced consolidation test that has determined the consolidation properties of soft soil slurries since the 1990s (Znidarcic et al., 2011).    50   Chapter 3: Experimental Program 3.1 Introduction The experimental program for this research was designed to provide a basis for understanding the fundamental properties governing the behavior of PA-MFT using microscale and mesoscale characterization methods.  The methods selected were based on the author’s experience as a consulting geotechnical engineer, consultation with experts in the fields of rheology, geotechnical engineering, surface chemistry, spectroscopy, and imaging indicated in the Preface, and through literature review.  To benefit from insight gained from important disciplines that are not consistently cross-referenced in published oil sands tailings research (namely findings from surface chemistry, rheology, geotechnical engineering, and high resolution material imaging), a broad characterization framework (Figure 3.1) was developed to facilitate consistent collection of intrinsic and complementary characteristic data.  The test methods used to evaluate how characteristic data function independently and in conjunction to produce observed material dewatering and strength behaviour during bench and field scale testing are shown in Figure 3.2.  Unless noted otherwise, all testing was completed in the author’s laboratory at the Point Grey campus of the University of British Columbia.  The laboratory space used was temperature controlled and provided standard work bench and floor space to conduct surface chemistry experiments and bench scale geotechnical testing. 3.2 Program Components The objective of the experimental program is to quantify how PA-MFT surface phenomena and material fabric function independently and in combination to impact the dewatering characteristics and associated strength of this synthetic geomaterial.   51   Figure 3.1: Characterization and Behaviour Evaluation Framework for Chemical Amendment of MFT  Note: The symbols 1o and 2o indicate primary and secondary, respectively.   52   Figure 3.2: Test Methods Used to Characterize Various Factors Impacting PA-MFT Behaviour   Findings in these areas would address persistent gaps in industry and support continued development of cost effective and sustainable ways of managing large volumes of oil sands fine tailings during operations, at closure, and during post closure. In this research only PA-MFT flocculated by the author using a single MFT source was evaluated.  Source materials and reagents used to flocculate PA-MFT are described later in this chapter.  External laboratories and test facilities were used to complete mineralogical, imaging and x-ray adsorption characterization of MFT and PA-MFT.   Experimental work was grouped under four areas of focus: baseline, surface, fabric and material strength (Figure 3.3).  Under each focus area, collected data where analyzed and their potential impact on the dewatering characteristics of PA-MFT evaluated.  Findings from qualitative techniques are used to identify potential constituents of interest or to confirm initial thinking about the potential for possible contributions of identified material components to observed behaviour.  Characteristics investigated using the testing framework (Figure 3.1 and  53  Figure 3.2) and the various areas of investigation (Figure 3.3) are summarized in Table 3.1.  This table also indicates the test methods used, key references, and where the tests were completed.  Figure 3.3: Overview of Experimental Program   54  Table 3.1: Summary of Characteristics Investigated and Methods Used     55  Baseline characterization (Focus Area 1) included determination of particle size distribution using laser diffraction, percent solids, and percent bitumen by an external industrial laboratory (Exova, Calgary) familiar with analysis of oil sands tailings samples.  The methylene blue index (MBI) was also determined at this external laboratory using the slurry method developed by Natural Resources Canada (2008).  These results were compared to MBI results obtained by the author using the same method.  Apparent specific gravity of PA-MFT was determined by the author.  Mineralogy of PA-MFT was determined using x-ray diffraction (XRD) and quantitative evaluation of minerals by scanning electron microscopy (QEMSCAN) at the SGS analytical test facility located in Burnaby, British Columbia. Surface characterization (Focus Area 2) included determination of the zeta potential of raw MFT, PA-MFT and bitumen impacted water released after flocculation of MFT.  Infrared spectra and Raman spectra were also obtained in an attempt to qualitatively identify the functional groups (specific groups of atoms or bonds within molecules) present on mineral surfaces pre- and post-flocculation.  X-ray fluorescence (XRF) was used to identify abundance of elements with molecular weights between 39.3 amu (Argon, Ar) and 74.9 amu (Arsenic, As) while X-ray adsorption near edge spectroscopy (XANES) was used to investigate the presence and form of sulphur in raw and flocculated MFT.  Both XRF and XANES were completed at the Canadian Light Source (CLS) with XRF investigated on the IDEAS beam line and XANES completed on the SXRM beam line.  These investigations were completed to investigate the role of surface properties on dewaterability of PA-MFT. Investigation into the fabric of PA-MFT (Focus Area 3) was accomplished by completing imaging at scales ranging between the millimeter and sub-micrometer.  Synchrotron based CT was used to image the three dimensional domain of wet and dry samples of PA-MFT at voxel  56  (volumetric pixel) resolutions of 8 µm, 4 µm and 0.65 µm.  This imaging method was used to investigate the spatial arrangement of PA-MFT pore networks at the imaging resolutions indicated.  Field emission (FE) SEM and cryo-SEM were used to obtain two dimensional images of both the pore spaces and mineral surfaces present in wet and dry samples of raw and flocculated MFT.  Using the SEM methods indicated the effects of flocculation on the surface features of clay minerals were also investigated. Strength characterization (Focus Area 4) of raw and flocculated MFT included measurement of both the rheological and geotechnical properties of raw and flocculated MFT.  Rheological characterization included: development of the flow curve for raw MFT and hydrated polymer; determination of the apparent yield stress of raw and flocculated MFT; completion of a slump test on PA-MFT; and calculation of the strength (G’) and loss (G’’) moduli of PA-MFT.  Characterization of geotechnical strength included: shear strength test using vane shear testing; measurement of surface elevation changes during material drying under self-weight and loaded conditions; and completion of consolidation tests using the UBC SICT apparatus described in Chapter 2.  A field vane was used to measure changes in peak undrained strength throughout the depth profile of a sample of PA-MFT that had been placed in an evaporation test cell at lift thickness exceeding 20 cm.  It should be noted that residual strength measurements were not measured during this research as the work focused on characterizing the maximum strength of PA-MFT. After adhesion between some of the residual bitumen in PA-MFT and the metal of the Casagrande cup was observed during attempts to determine the liquid limit using that method, a Swedish Fall cone was used to determine the liquid limit and undrained strength of flocculated MFT.  This method enabled correlations between geotechnical moisture content and undrained  57  strength to be determined as well as facilitating an evaluation of the use of solids content as a means of evaluating PA-MFT behaviour.  Undrained strength measurements obtained using the field vane and fall cone methods were compared.  The shear strength of materials exposed to air drying at the surface of the evaporation test cell was measured using a handheld Torvane.  The evaporation test cell designed by the author for this research provided a means by which the effects of lift thickness, dewatering characteristics, and geotechnical strength of PA-MFT could be measured and correlated. 3.3 Samples and Reagents 3.3.1 Mature Fine Tailings 100 L of MFT was ordered from the MFT sample bank operated by Alberta Innovates Technology Futures (AITF) in December 2013.  The raw MFT sample was prepared at AITF’s laboratory facilities by combining MFT from four producers involved in the MFT sample bank into a single vessel and homogenizing the combined MFT according to protocol developed for use at the MFT bank by leading oil sands tailings researchers working in the industry (Lloyd, 2014a).  The MFT sample bank confirmed that no additional water was introduced or removed to create the homogenized MFT bulk sample (Lloyd, 2014b).  The sample was then subdivided into five 20 L plastic buckets to facilitate the sample’s ground transportation from the sample bank located in Alberta to the author’s laboratory at UBC’s West Point Grey campus in British Columbia.  When the sample was delivered to UBC in late December 2013, the contents of the five buckets were emptied into a 100 L neoprene storage container and mixed using a hand drill with a paddle mixer blade for thirty minutes to re-homogenize the bulk sample.  The fitted neoprene lid of the bulk storage container was then secured to the top of the bulk storage container using several layers of industrial-grade plastic food wrap.  This material is referred to as raw AITF-MFT in this thesis.  The bulk AITF-MFT sample was stored in its covered neoprene container away from direct  58  sunlight in a laboratory maintained at an annual average temperature of approximately 20oC.  All subsamples of AITF-MFT were only obtained after the bulk sample had been mixed for a minimum of twenty minutes using the hand operated paddle to achieve a uniform water-like consistency throughout the bulk sample.  Solids content and geotechnical moisture content were determined for every sample of AITF-MFT obtained from the bulk storage container. Beckman Coulter laser particle size analysis provided by AITF (Appendix A) indicated that the majority of the particles in the homogenized sample had an apparent diameter ranging between 2 µm and 30 µm, with a secondary concentration of particles with an apparent diameter ranging between 0.8 µm and 2 µm.  AITF also reported a solids content of approximately 22% by weight at the time of initial sample homogenization in December 2013 (Lloyd 2014b). 3.3.2 Primary Extraction Water A single source of primary extraction water (PEW) provided by an oil sands surface operation was used to hydrate the polymer used to create PA-MFT (Table 3.2).  The water was sampled by the author and submitted for testing at ALS Laboratories in Vancouver.  When settled, the PEW used in this research appeared clear and had low concentration of total suspended solids.  For the duration of testing, PEW pH ranged between 8.3 and 8.5 and had an average conductivity of 340 mS/m.  When compared to the characteristics of tailings pond water tested in the early 1990s (Table 2.2), PEW used in this research shows an almost three-fold increase in the concentration of sodium and a fifty times increase in the concentration of chloride.  Sulphate levels have also increased.  Concentrations of silicon are also lower.  PEW was stored in the author’s laboratory in sealed 5 gallon plastic containers.  59  Table 3.2: Characteristics of Primary Extraction Water (PEW) Used for Polymer Hydration   3.3.3 Polymer Magnafloc® 5250 was the anionic flocculant used in the flocculation experiments described in this thesis.  This flocculant is described by its manufacturers as a high molecular weight anionic PAM (BASF, 2015).  A high molecular weight PAM is described as having a molecular weight ranging between 106 g/mol and 107 g/mol (Zeta-Meter, 1993).  This polymer was selected based on its similarity to other polymers used to thicken fine tailings at operating surface mines in the Alberta oil sands.  Magnafloc® 5250 is also used in the sedimentation of coal tailings, kimberlite slimes, and for thickening tailings from mineral sands and base metal mines (ACG, 2006; BASF, 2015).  BASF provided samples of the polymer to the author in its dry form as an off-white granular powder.  The polymer was stored in a metal chemical storage cabinet in the author’s UBC laboratory away from sources of moisture and from direct sunlight.  Once the polymer  60  concentration required to flocculate samples of AITF-MFT was determined (described later in this Chapter), the polymer was hydrated with PEW, consistent with hydration methods used in industry at lab and field scale.  The polymer was hydrated in small batches (<3 L/batch) using a jar tester.  Hydrated polymer was used within 10 days of hydration to ensure freshness.  The procedure used to hydrate the polymer is provided in Appendix B. 3.3.4 Polymer-Amended Mature Fine Tailings Polymer-amended mature fine tailings (PA-MFT) is the name given to the thickened tailings product resulting after the controlled combination and mixing of a subsample of raw AITF-MFT and hydrated anionic polymer which produces a noticeable change in the fabric of the AITF-MFT (Figure 3.4).  The resulting PA-MFT has an initial appearance similar to a cottage cheese (Figure 3.4B) with visible voids and is accompanied by an observed release of clear water.  Figure 3.4: Visual Comparison of (A) Raw AITF-MFT and (B) PA-MFT Immediately After Flocculation Showing Initial Water Release (~30% of the Initial MFT Volume)    To develop a means of consistently producing PA-MFT for analysis in this research, the author solicited expertise from researchers with considerable experience flocculating MFT at laboratory and industrial scales.  This included an “in-lab” demonstration of the methods used to flocculate MFT and guidance on how to determine concentrations of hydrated polymer needed to flocculate MFT with characteristics similar to raw AITF-MFT.  Stock solutions of hydrated  61  polymer were prepared at concentrations of 0.2%, 0.3% and 0.4% and used to flocculate approximately equal masses of MFT that had been measured into beakers. Using the laboratory procedure described in Appendix B.2, polymer is gradually added to the beaker containing the MFT using a 60 mL syringe while a rectangular paddle blade from a Jar Tester is operated at a low to moderate speed (30-75 rpm) with occasional pulses at high speed (>200 rpm) to ensure proper distribution of the polymer throughout the MFT.  The “flocculation” point was identified as the point at which the solids achieved a cottage cheese/curdled consistency in the presence of clear water, described as the “initial water release” from the MFT (Figure 3.4B).  The amount of water released was determined by measuring the mass of clear water released within 15 minutes of flocculation and after 24 hours.  The amount of water release was calculated by dividing the measured mass of clear water decanted by the sum of the initial mass of MFT and the volume of hydrated polymer used to achieve the desired end condition.  The 0.3% polymer stock solution consistently produced the largest quantity of water released at the time of flocculation (between 30% and 40%) and amended MFT with the desired cottage cheese structure providing the greatest resistance to pressure applied by the flat face of a 10 cm long metal palette knife.  During field trials, the quantity of water release is used as a performance indicator for fine tailings flocculation with flocculation effectiveness determined by visual comparison of the resulting material to pictures of the desired cottage cheese fabric. For ease of discussion, the term “flocculation” will be used in this thesis to describe the process that results after the desired combination of MFT and the polymer has occurred.  However, as this research progressed the complexity of characterizing the physico-chemical mechanisms at play during combination of PAM and MFT became evident and the term polymer-amended MFT (PA-MFT) was deemed to be appropriate for the resulting material.  Challenges associated with  62  use of the term flocculation to describe the likely physico-chemical processes that are involved are described in Chapter 4 and Chapter 5. On average, the 75 kg of PA-MFT produced in the author’s laboratory had an average solids content of approximately 30% of which approximately 66% was comprised of clay minerals.  On a percent solids basis, the average polymer dosage required to produce the desired flocculation condition was approximately 1,070 ppm or g/tonne of dry solids while on a percent clay minerals basis, the polymer dose averaged approximately 500 ppm.  The resulting PA-MFT had an average 24-hr initial water release of approximately 30% calculated as the total mass of water released from the PA-MFT after 24 hrs divided by the sum of the mass of raw AITF-MFT and the mass of hydrated polymer added. PA-MFT was produced in the amounts required to conduct the analyses described later in this Chapter.  Except for the samples that were immediately placed in 1 L glass mason jars that were used for flow curve and apparent peak yield strength determination at an external laboratory, PA-MFT used in both the SICT and in the evaporation test cell was stored in 5 gallon buckets that had been filled to no more than half full and fitted with a lid that provided an air-tight seal.  These buckets were stored in the in the author’s laboratory at temperatures ranging between 18°C and 20oC and were not exposed to direct sunlight.  The solids content and geotechnical moisture content of PA-MFT was taken both at the time of flocculation and at the time of material testing.  The date of PA-MFT flocculation and the date testing was completed were also recorded so that potential self-weight consolidation and storage effects could be considered during data analysis. 3.3.5 Other Reagents All chemical reagents used in the procedures described in this Chapter were either directly supplied with a test kit provided by an industrial supplier for a specific procedure (in the case of the OFITE MBI test), or obtained from commercial chemical suppliers.  Reagents not provided at  63  the required molar or normal concentrations were prepared in the author’s laboratory using the indicated dilution factors and de-ionized water.  All reagents were stored in properly sealed containers and reactive agents like sulphuric acid and sodium hydroxide were stored in non-reactive sealed enclosures apart from each other to avoid chemical interaction during storage.  Specific reagents used are itemized either in the procedures described in the following sections, or as indicated in the method descriptions provided in Appendix B. 3.4 Procedures, Methods and Equipment In an attempt to characterize the properties and behaviour of raw AITF-MFT and the resulting PA-MFT, testing outlined in Figure 3.1 was completed using unmodified whole samples of the indicated materials i.e. bitumen content of tested materials was not altered to complete laboratory experiments.  The only exception was completion of Dean Stark extraction on three samples of raw AITF-MFT submitted to Exova’s laboratory to determine bitumen, water and solids content.  The following sections provide brief summaries of the methods used, indicating where variances were made by the author to facilitate test replication and assessment by other researchers.  Standard test methods developed by the American Society for Testing and Materials (ASTM) can be accessed using the organization’s website or in the edition of the standards manual noted. 3.4.1 Geotechnical Moisture Content and Percent Solids Determination To assess the impact of evaporation and the effectiveness of homogenization of the bulk sample of raw AITF-MFT, the percent solids and geotechnical moisture content were determined every time subsamples were taken from the bulk sample storage container.  Determination of both the geotechnical moisture content and percent solids required completion of the following procedure consistent with ASTM D2216-90 as follows: The geotechnical moisture content was calculated using the following formula:  64  𝒘𝒘(%) = �𝑾𝑾𝑻𝑻𝑻𝑻𝑻𝑻+𝑾𝑾𝒆𝒆𝑾𝑾 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝒆𝒆 −𝑾𝑾𝑻𝑻𝑻𝑻𝑻𝑻+𝑫𝑫𝑫𝑫𝑫𝑫 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝒆𝒆𝑾𝑾𝑻𝑻𝑻𝑻𝑻𝑻+𝑫𝑫𝑫𝑫𝑫𝑫 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝒆𝒆 −𝑾𝑾𝑻𝑻𝑻𝑻𝑻𝑻 � ∗ 𝟏𝟏𝟏𝟏𝟏𝟏 Equation 3.1 The percent solids content was calculated using the following formula: % 𝑺𝑺𝑺𝑺𝑺𝑺𝑻𝑻𝑺𝑺𝑺𝑺 =  �𝑾𝑾𝑻𝑻𝑻𝑻𝑻𝑻+𝑫𝑫𝑫𝑫𝑫𝑫 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝒆𝒆 −𝑾𝑾𝑻𝑻𝑻𝑻𝑻𝑻𝑾𝑾𝑻𝑻𝑻𝑻𝑻𝑻+𝑾𝑾𝒆𝒆𝑾𝑾 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝒆𝒆 −𝑾𝑾𝑻𝑻𝑻𝑻𝑻𝑻� ∗ 𝟏𝟏𝟏𝟏𝟏𝟏 Equation 3.2  In both equations, w represents the mass described in the adjacent subscript. 3.4.2 Baseline Characterization of Raw AITF-MFT Samples of raw AITF-MFT were sent to the Exova industrial oil sands laboratory located in Calgary, Alberta.  Samples were shipped to Exova’s laboratory in July 2014 and in September 2014 to assess if there was any variability that might exist in the baseline characteristics of the bulk sample that could potentially impact testing completed throughout the summer of 2014.  A total of three 1 L samples were shipped to Exova after the bulk AITF-MFT had been homogenized.  Using the Dean-Stark water extraction method described by Bulmer and Starr (1979), the percentage bitumen, water, and solids content were determined.  Exova was also tasked with determining the methylene blue index (MBI) using the slurry determination method developed by CanmetENERGY (NRC, 2008) provided in Appendix B.  The laboratory completed a particle size analysis of the samples of raw AITF-MFT using a Malvern Mastersizer which employs the principles of laser diffraction to determine an apparent effective diameter of particles in suspension.  Results from this characterization are described in Chapter 4 and data presented in Appendix A.2. 3.4.3 Specific Gravity Determination for PA-MFT The specific gravity (GS) of air dried and oven dried PA-MFT was determined using ASTM D854-00.  The only variation made to the standard method was that testing was completed on  65  samples that were allowed to soak for both the specified 10 minute period and for 24 hours.  The specific gravity was calculated using the following formula: 𝑮𝑮𝑺𝑺 = 𝑾𝑾𝑷𝑷𝑺𝑺 −𝑾𝑾𝑷𝑷(𝑾𝑾𝑷𝑷𝑺𝑺 −𝑾𝑾𝑷𝑷) + (𝑾𝑾𝑨𝑨 −𝑾𝑾𝑩𝑩) = 𝑫𝑫𝑫𝑫𝑫𝑫 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝒆𝒆 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑫𝑫𝑫𝑫𝑫𝑫 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝒆𝒆 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺 + 𝑺𝑺𝑺𝑺𝑾𝑾𝑺𝑺𝑫𝑫𝑺𝑺𝑾𝑾𝒆𝒆𝑺𝑺 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝒆𝒆 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺 Equation 3.3  Specific gravity data collected from five air dried and two oven dried MFT samples are analyzed and interpreted in Chapter 4. 3.4.4 Percent Clay Determination Using Methylene Blue Index The methylene blue index (MBI) of samples of raw and flocculated AITF-MFT were determined using two methods: the Methylene Blue Procedure for Sludges and Slurries developed by CanmetENERGY (NRC, 2008), hereafter described as the CANMET slurry method, and using the methylene blue test kit for evaluation of bentonite drilling muds developed by OFI Testing Equipment (hereafter described as the OFITE method).  The CANMET slurry method was used by the author and Exova to determine the MBI of samples of raw AITF-MFT.  The author also completed MBI determination for PA-MFT samples.  MBI determination using the OFITE method was investigated because of the relative simplicity of the test compared the CANMET slurry method and the use of this test kit in oil and gas and other drilling practice to confirm adequate bentonite content in slurries and drilling muds. 3.4.4.1 CANMET Method for Slurries The Methylene Blue procedure for Sludges and Slurries (NRC, 2008) to determine the MBI of raw and flocculated AITF-MFT used in this research is provided in Appendix B.  Rather than using an MFT-analog as a reference standard, the author only used a sample of standard bentonite provided by Hoskins Scientific to confirm that the MB solution possessed the normality as specified in the procedure.  Sample masses were limited to 3 g to reduce the volume of MB solution required to complete each titration.  Boiling the diluted test sample for between 20 and 25 minutes  66  was found to consistently produce the best sample dispersion.  Also, Whatman Grade 1 filter paper was used instead of Whatman 42 ashless filter paper specified in the procedure. MBI was calculated as follows: 𝑴𝑴𝑩𝑩𝑴𝑴 � 𝑺𝑺𝒆𝒆𝒎𝒎𝟏𝟏𝟏𝟏𝟏𝟏𝟏𝟏 𝑺𝑺𝑺𝑺𝑺𝑺𝑻𝑻𝑺𝑺𝑺𝑺 � = 𝑽𝑽𝑴𝑴𝑩𝑩 (𝑺𝑺𝑺𝑺)𝒙𝒙 𝑵𝑵𝑴𝑴𝑩𝑩(𝒆𝒆𝒎𝒎)𝒙𝒙 𝟏𝟏𝟏𝟏𝟏𝟏𝑴𝑴𝑺𝑺𝑺𝑺𝑺𝑺 𝑺𝑺𝒐𝒐 𝑺𝑺𝑺𝑺𝑺𝑺𝑻𝑻𝑺𝑺𝑺𝑺 𝑻𝑻𝑻𝑻 𝑺𝑺𝑺𝑺𝑺𝑺𝑫𝑫𝑫𝑫𝑫𝑫 (𝟏𝟏) Equation 3.4  where VMB is the volume of methylene blue required to complete the titration, and NMB is the normality of the methylene blue solution used in the titration.  The percent clay by weight was always calculated using the Yong and Sethi (1984) correlation presented in Equation 3.5. 𝑾𝑾𝑾𝑾 % 𝑪𝑪𝑺𝑺𝑺𝑺𝑫𝑫 = 𝑴𝑴𝑩𝑩𝑴𝑴 + 𝟏𝟏.𝟏𝟏𝟎𝟎𝟏𝟏𝟎𝟎∗ 𝟏𝟏𝟏𝟏𝟏𝟏 Equation 3.5  3.4.4.2 OFITE Method for Drilling Muds The OFITE method used to determine the MBI of raw and flocculated AITF-MFT is presented in Appendix B.  To enable measurement and regulation of the hot plate temperature the same hot plate used to complete MBI determination using the CANMET slurry method was used, instead of the simplified hot plate provided in the OFITE test kit.  In addition, the boil duration was ultimately increased from 10 minutes to 15 minutes to enhance sample dispersion.  The primary differences between the OFITE and CANMET methods are summarized in Table 3.3.  The MBI and percent clay were calculated using Equation 3.4 and Equation 3.5 respectively.   67  Table 3.3: Comparison of CANMET Slurry and OFITE MBI Methods Component CANMET Slurry Method OFITE Drilling Fluid Method Methylene Blue Solution 0.006N 0.01N Dispersing reagent(s) NaHCO3 (1M) NaOH (10% w/w) H2O2 (3%) Acidifying Agent H2SO4 (10% v/v) H2SO4 (5N) Boiling Time 20-25 mins 10-15 mins Specified pH range prior to titration 2.5-3.8 Unspecified Filter paper used to determine titration end point Whatman Grade 1 test sheets used instead of Grade 42 ashless paper specified Whatman Grade 1 test sheets  3.4.5 Mineralogical Characterization of PA-MFT A sample of PA-MFT was submitted to SGS’ mineral characterization laboratory in Burnaby, BC.  Mineralogical characterization of the sample was completed using Quantitative Evaluation of Minerals by Scanning Electron Microscopy (QEMSCAN) and x-ray diffraction (XRD).  Chemical assay and XRD results were compared to QEMSCAN data for quality assurance and quality control.  Descriptions of QEMSCAN and XRD procedures, including sample preparation methods used by SGS, are provided in Appendix A. 3.4.6 Zeta Potential Determination Zeta potential (ζ) provides a measure of the electrokinetic potential of charged particles in a dilute suspension providing insight into the overall suspension stability i.e. how likely are particles in a given suspension to remain dispersed or aggregated under the specific chemical characteristics of the primary suspension fluid at the time of the test.  The zeta potential of primary extraction water, hydrated anionic polymer and dilute suspensions (concentration ranging between 0.1% and 0.3%) of raw AITF-MFT and wet and air dried PA-MFT was determined using a Zeta-Meter and the procedures describes in Appendix B.  68  3.4.7 Rheological Characterization 3.4.7.1 Flow Curve Development The flow curves of hydrated polymer and raw AITF-MFT were developed using an Anton-Paar Physica MCR 301 TruGap rheometer and the concentric cylinder with narrow gap geometry developed for use with that rheometer.  After the rheometer was calibrated and the cylindrical bob being used in the analysis installed in the adjustable mast head of the rheometer, approximately 19 ml of the material being characterized was placed in the cylindrical sample holder.  The cylindrical sample holder was then placed into its holding compartment located in the base of the rheometer.  The rheometer and components of the concentric cylinder geometry used are pictured in Figure 3.5.  Figure 3.5: Anton-Paar TruGap Rheometer (A) and Concentric Cylinder Geometry Detail (B)    The adjustable mast (Figure 3.5A) holding the cylindrical bob (lower region of Figure 3.5B) was lowered into the test position using the rheometer’s control software.  At the test start position the cylindrical bob was fully inserted into the test material in the sample holder and the material in the sample holder displaced such that the test fluid aligned with the fill lip of the sample holder (upper region of Figure 3.5B).  The rheometer was set to maintain a sample temperature of 22oC for the duration of the test.  The rheometer control software was used to implement a ramp function that increased the shear rate from 1 to 800 sec-1, with a 15 second interval for each shear  69  rate.  The flow curve data captured by the rheometer control software was exported for additional analysis. The MFT tested was conditioned using a pre-shearing and rest period prior to initiating the ramp, as this material exhibits shear thinning behavior when shear stress is applied but increases in viscosity as its time at rest increases (aka thixotropy).  As such, a conditioning procedure was used to enable the shear stress condition at the beginning of each test of MFT to be achieved.  This enabled test results to be compared as material shear stress at the start of testing of MFT samples would be consistent.  MFT conditioning was as follows: the rheometer was used to pre-shear the sample of raw AITF-MFT at a rate of 500 sec-1 for 15 minutes followed by a 30 minute rest period prior to proceeding with flow curve characterization using the shear rate ramp function.  To limit the effects of evaporation during sample conditioning the exposed surface of the AITF-MFT at the top of the cylinder was covered with a thin film of silicon oil.  Parafilm was used to cover the annulus between the bob spindle and the outer wall of sample holder ensuring that parafilm did not touch the spindle.  It should be noted that no sample conditioning steps were used with the polymer based on well documented flow characteristics of this material provided by its manufacturer, BASF.  Testing was completed twice for each material being characterized. 3.4.7.2 Yield Stress Determination The peak yield stress of raw AITF-MFT was determined using the Anton-Paar Physica MCR 301 TruGap rheometer and the concentric cylinder narrow gap geometry sample holder described in Section 3.4.7.1.  After the sample was loaded in the rheometer and subjected to the pre-shearing and rest procedures described earlier, the rheometer control software was used to rotate the inner cylinder at a constant rotational speed of 0.1 rpm for approximately five minutes while the corresponding torque was collected by the rheometer. Collected torque data was used to calculate the shear stress for the concentric cylinder geometry provided in Equation 3.6:  70  𝝉𝝉 = 𝑻𝑻𝟐𝟐𝟐𝟐𝑫𝑫𝑺𝑺𝟐𝟐𝑯𝑯 Equation 3.6  where shear stress (τ) is a function of the torque (T), the radius (ra) which is an average between cylinder radii r1 and r2 and the height (H) of the inner cylinder used. The yield stress profiles for PA-MFT were determined using wide gap geometry and vanes with 4 blades developed for use with a Brookfield rheometer (Figure 3.6).  Figure 3.6: Infinite Cup and Vanes (A) Used with Brookfield Rheometer (B)  The two infinite cups used in the testing were designed and fabricated by Coanda and have an inner diameter of 8.8 cm with depths of 5 cm (Figure 3.6A) and 15 cm (Figure 3.6B).  The vanes used in this research had a maximum length of 2.5 cm and a diameter of 1.3 cm.  These dimensions enable the ratio of cup depth to vane height and the ratio of cup width to vane diameter to exceed three, which enables boundary effects to be minimized (Nguyen and Boger, 1985; Fisher et al., 2007).  The methods used to determine the yield stress using the vane in cup geometry are provided in Appendix B.6. A slump test was also performed to see how yield stress values obtained using this method compared to those collected using the Brookfield rheometer with wide gap geometry.  The slump test procedure is described in Appendix B.6.    71  3.4.7.3 Determination of Storage and Loss Moduli of PA-MFT Materials exhibiting intermediate behaviors characteristic of both solids and liquids are generally classified as viscoelastic.  Oscillatory rheology provides an experimental means of characterizing the rheological behavior of these materials (Coanda, 2014c).  By subjecting a sample material to strain at a constant frequency, the storage modulus (G’) and loss modulus (G’’) can be measured and their relationship to each other at different times evaluated.  Where the storage modulus (G’) exceeds the loss modulus (G’’), the material is described as exhibiting “solid like” elastic behaviour.  Where this relationship is reversed, the material is thought to exhibit a “liquid like” viscous response.   The storage modulus (G’) and loss modulus (G’’) of PA-MFT were determined using parallel plate geometry with an Anton-Paar Physica MCR 301 TruGap rheometer at Coanda’s test facilities located in Burnaby, BC.  The method used to develop the storage and loss modulus profiles is described in Appendix B. 3.4.8 Characterization of PA-MFT Using an Evaporation Test Cell The evaporation test cell (hereafter test cell) used in this research was designed by the author and fabricated to the author’s specifications by RST Instruments at their manufacturing facility in Maple Ridge, BC.  The cell (Figure 3.7) was devised to provide a straightforward means of completing the following: measuring changes in sample mass with negligible dead weight effects as water leaves the sample due to either evaporation or underdrainage; quantifying the amount of water removed from a sample by means of underdrainage; and characterizing the impacts on drainage and undrained strength resulting from adding a normal load to the top of a lift of measured thickness exceeding 20 cm.  The cell consists of a 30 cm deep sample holding compartment (20 cm wide x 40 cm long) made from stainless steel that is mounted on a load cell.   72  The entire unit was calibrated by the manufacturer and mass measurements are obtained using a continuous digital readout box that is connected to the unit (Figure 3.7A).  Figure 3.7: Evaporation Test Cell Setup (A), DrainTube® Filter Fabric (B), and Mass Used (C)  The sample compartment was mounted on a fixed metal arm inclined at 2% to enable the base of the sample box to mimic the typical configuration of the base of fine tailings designated drying areas used in industry.  A valve-operated drainage port was installed at the lower end of the sample box to enable removal of water produced through underdrainage away from the sample when the valve is in the open position.  A braided tube was attached to the drainage port downstream of the port valve to convey underdrainage into a sealed 5 gallon bucket (Figure 3.7A). To enhance drainage and to aid clean-up of the cell between tests, the walls and bottom of the sample box were lined with DrainTube® 500P FT D25 filter fabric.  This product consists of a drainage layer and a filter layer made from short polyester or polypropylene fibers that are needle punched together with corrugated polypropylene pipes inserted at desired intervals.  The filter fabric has an apparent opening size of 120 microns while the corrugated pipes encased in the filter  73  fabric have two perforations per valley at 180° spacing with a 90° rotation per valley.  To optimize underdrainage, a single piece of polypropylene pipe was encased in the filter fabric lining the bottom of the evaporation cell and routed directly into the drainage port of the test cell.  The mass of seepage collected in the 5 gallon bucket and measurements of the sample mass in the evaporation cell were recorded daily. PA-MFT was placed at depths of 22.5 cm for the first test and 24 cm for the second test with the test cell.  These depths were selected as they are marginally higher than the maximum depths (15-20 cm) for which use of this MFT amendment method was designed.  The space between the top of the sample and top of the box enabled a distinct starting reference point to be indicated along one of the vertical walls of the cell.  This initial mark was used as a reference point enabling measurement of the incremental changes in thickness as the sample dried.  The space at the top of the test cell was also utilized during the investigation into the effects of applied surface load on undrained strength characteristics in the sample.  Table 3.4 summarizes the readings that were taken over the duration of the three major investigations completed using the test cell.   74  Table 3.4: Summary of Investigations Completed Using the Evaporation Test Cell Investigation No. Test Duration (days) Characteristic Data Collected over Test Duration Test No. 1 32 Surface Deformation • Sample thickness Water Loss (seepage & evaporation) • Mass of collected underdrainage • Mass of sample in test cell  Surface Shear Strength • Shear strength (Torvane) • Moisture content Undrained Shear Strength • Field vane (multi depth) • Swedish Fall Cone (near surface samples & sample profile at end of test) • Moisture content (profile) Test No. 2 Phase I 30 Same as for Test No. 1 Same as for Test No. 1 Test No. 2 Phase II 82 Effects of normal load • Sample thickness • Mass of sample in test cell • Mass of underdrainage • Field vane (multi depth)  3.4.8.1 Strength Characterization of Material in Evaporation Test Cell The shear strength of the surface PA-MFT placed in the test cell was measured for evaporation cell (EC) Test 1 and EC Test 2 Phase I using a Torvane ® manufactured by DGSI.  Shear strength measurements were obtained by pressing the zeroed Torvane® into a selected area on the surface of the PA-MFT that had been levelled and was a minimum of 2.5 cm away from the side walls of the test cell.  The Torvane® knob was then rotated at a constant speed and with constant vertical pressure such that failure of the sample surface developed within 5-10 seconds.  The vane was then removed and the failed sample collected for moisture content determination.  During early stages of evaporative drying, the large vane with a shear strength range up to 0.2 kg/cm2 was used.  As evaporative drying progressed, the standard vane used to measure shear strength between 0 and 1 kg/cm2 was used.  75  3.4.8.2 Undrained Strength Determination Using Field Vane A profile of undrained shear strength of PA-MFT in the test cell was obtained using a calibrated Geonor H-60 Inspection Vane with its standard 20 mm by 40 mm four-bladed vane.  To obtain measurements, the vane was inserted to the desired depth below the surface of the sample (marked on the instrument’s shaft) and the graduated scale zeroed prior to rotating the instrument handle clockwise as slowly as possible with constant speed.  The maximum shear strength value observed on the dial was then recorded.  All measurements using this device were taken by the author using the same procedure to support consistency in the data collected.  The peak shear strength was measured at a minimum of two depths in the sample at each location investigated.  At each depth the vane was extracted and test material coating the vanes was removed and used to determine the moisture content once there was enough of it to be representative of the tested material. 3.4.8.3 Undrained Strength Determination Using the Swedish Fall Cone A new, factory-calibrated Swedish Fall cone device (Figure 3.8) and its associated penetration cones, manufactured by Geonor, was used to determine the liquid limit and undrained strength of samples of PA-MFT removed from the surface of PA-MFT tested in EC Test 1 and EC Test 2 Phase I. The undrained strength of samples taken throughout the depth profile of the PA-MFT evaluated during EC Test 1 was also determined when that test was dismantled.  The Fall Cone testing procedure is provided in Appendix B.7.   76   Figure 3.8: Swedish Fall Cone Used in Test Work (A), Showing Sample Being Tested (B)  3.4.8.4 Investigation into the Effect of Surface Load Application After Phase I of EC Test 2 was complete, masses averaging approximately 4 kg (Figure 3.7C) were stacked incrementally on the surface of the sample in the test cell over a period of 32 days until a total mass of 43.6 kg had been added.  The cut-outs included in each mass functioned as sampling points enabling the effects of normal load on undrained shear strength values to be determined at four locations at different time intervals after the various masses had been applied.  Undrained shear strength measurements were determined by inserting the field vane to the desired depth of measurement through the cut-outs in the masses at various time intervals.  For each increment of mass added, changes in the surface elevation of the mass were tracked in time.  The amount of underdrainage produced with the addition of each mass increment was also measured. 3.4.9 Seepage Induced Consolidation Testing Seepage induced consolidation testing (SICT) was completed for one sample of raw AITF-MFT and two samples of PA-MFT.  Once the material sample to be tested is loaded into the inner cell, the inner and outer chambers of the SICT apparatus (Figure 3.9) are filled with PEW from the reservoir seen to the right of the image.  The test was conducted to obtain the parameters described in Section 2.11.1.4.  Estepho (2014) provides detailed descriptions and schematics of  77  the UBC SICT apparatus that was used to complete testing reported in this thesis as well as an expanded description about the test procedure, data collection and analysis using SICTA.  The specific procedures used in this research are provided in Appendix B.8.  Figure 3.9: UBC SICT Setup Showing Inner and Outer Cells Filling with PEW  3.4.10 Infrared Spectroscopy Fourier transform infrared spectroscopy (FT-IR) spectra were collected for a crushed dry granule of anionic polymer, air dried samples of raw AITF-MFT and PA-MFT, bitumen impacted water released during creation of PA-MFT, and a sample of toluene extracted bitumen (TEB) from a sample of raw AITF-MFT.  All samples were ground using a mortar and pestle to enable each specimen to be in intimate contact with and to cover the surface of the diamond-coated KRS5 crystal mounted in the Perkin Elmer Spectrum 100 FT-IR Spectrometer.  After the crystal surface had been cleaned and the background spectra for air over the KRS5 crystal was collected with the pressure arm in place, a sample of each material was used to cover the surface of the crystal before the pressure arm was placed in the locked position covering the sample.  An infrared beam was directed into the crystal as shown in Figure 3.10.  78   Figure 3.10: Schematic Showing Principles of Attenuated Total Reflectance Spectroscopy (After Perkin Elmer 2005)   The refractive index of the KRS5 crystal is very high compared to the refractive index of the samples being analyzed and results in an internal reflectance that creates an evanescent wave which propagates beyond the crystal surface by between 0.5 µm and 5 µm.  As the evanescent wave makes contact with the sample on the external surface of the crystal the attenuated infrared beam is internally reflected and eventually exits the edge of the crystal within the spectrometer.  It is the exiting wave that is measured by the spectrometer’s infrared detector.  The unit then generates the IR spectrum which was initially viewed using Perkin Elmer’s Spectrum® software.  In this research the IR spectra were collected by running continuous scans for wave numbers between 275 and 4,000 cm-1 at a 1 cm-1 increment. 3.4.11 Raman Spectroscopy Initially wet samples of raw AITF-MFT, PA-MFT, bitumen impacted water from the creation of PA-MFT, and toluene extracted bitumen were placed in clean glass Petri dishes (one sample per dish) and allowed to dry in air.  The Raman spectra for each of the dried samples and a sample of dry granules of the polymer were obtained by subsequently placing each Petri dish on the imaging platform of the LabRam HR laser Raman Spectrometer (Figure 3.11).    79   Figure 3.11: LabRam HR Raman Spectrometer (A) with Optical Lens Above Sample (B)  Using the LabSpec® software which was used to operate the LabRam HR unit, the wavenumber range was set to 500 to 8,000 cm-1.  After switching the LabRam HR unit into the “Video” mode the portion of the sample for which the spectra was to be generated was selected by adjusting the height of the imaging platform to bring the image viewed in LabSpec® into focus.  The largest magnification optical lens for which the image of the sample area of interest could be brought into focus was used.  Once the sample area of interest was selected and a screen shot of the image captured, the LabRam HR unit was switched to “Raman” mode and the LabSpec® software used to capture the spectra of the sample on the imaging stage.  To assess surface variability, spectra were generated for specific sites on the sample surface.  The spectra for each sample were then corrected for baseline shift prior to completion of additional analysis. 3.4.12 Synchrotron Based Computer Tomography Samples of dry and wet PA-MFT were prepared and loaded into polyethylene sample vials (1.26 mm thick walls, average internal diameter of 1.5 cm and approximately 4 cm long).  The vials were sealed with a fitted lid and shipped via air courier to the Canadian Light Source (CLS) located in Saskatoon.  The CLS is a third generation, 2.9 GeV storage ring operating at a ring current of 250 mA.  At the CLS, each sample vial was mounted onto a 360° rotation stage of the BMIT-BM (aka 05B1-1) beamline (Wysokinski et al., 2007).  To complete the imaging the  80  beamline monochromator was optimized to 25 keV and the sample to detector distance was optimized to 70 cm. Each vial was oriented vertically so that its cap was in contact with the horizontal surface of the imaging stage.  A 1 cm section of the sample vial was then marked for reference as the portion of the sample where the light source passing through the beamline’s monochromator would be aimed.  The test chamber was sealed prior to exposing the light source.  While the light source was focused on the randomly selected zone of interest, 3,750 image projections were taken at every 0.048 degrees and captured with a Hamamatsu camera with 8.54 µm pixel size.  The projections were first normalized and then used to create refraction and apparent absorption image data sets.  Image clean up and processing was done using ImageJ software (Schneider et al, 2012).  Reconstructions and visualizations of the computer tomographic images were done separately using SkyScan NRecon and Avizo® to support further 2D and 3D quantitative analyses (Bruker, 2015).  Variations in energy adsorption visible in the collected images were used to assign false colors to distinguish sample pore space from light and denser solids.  This differentiation was used to estimate the amount of void space in the sample visible at voxel resolutions of 8 µm and 4 µm.  To investigate the nature of voids at 0.65 µm voxel resolution, a sample of wet PA-MFT was packaged by the author and sent by CLS to the European Synchrotron Radiation Facility (ESRF) located in Grenoble, France. 3.4.13 X-ray Fluorescence Microscopy and X-ray Adsorption Spectroscopy Samples of dry polymer granules, hydrated polymer, wet and dry samples of raw AITF-MFT, and wet and dry samples of PA-MFT were sent to the CLS in capped polyethylene sample vials described earlier.  A sample of toluene extracted bitumen was provided to CLS in a covered glass sample vial.  At the CLS, dried samples were mounted on carbon tape that had been affixed  81  to conductive sample plates made out of copper or aluminum as seen (Figure 3.12).  Samples were mounted for each of the various beamlines used.  Figure 3.12:  Mounted Dry Samples Used for X-Ray Adsorption Spectroscopy (XANES) at the Sulphur Edge   The VLS-PGM beamline, which has a plane grating monochromator optimized for the 5.5 eV to 250 eV range, was initially used to investigate the absorption spectroscopy of sulphur at the L- edge for samples of dry polymer, dry AITF-MFT, and dry PA-MFT.  When use of this beamline failed to indicate the sulphur spectra at the L-edge for the three dried materials, the Soft X-ray Microanalysis beamline (SXRMB) was used to obtain Sulphur adsorption spectra at the K-edge.  This beamline uses a bending magnet source and is equipped with a double crystal monochromator with two sets of crystals, Si (111) and InSb (111), and focussing optics covering an energy range of 1.8-10 keV.  Sulphur XANES measurements were carried out with the specimens mounted on a four axis manipulator in a chamber under a continuous flow of Helium gas. The wavelength-dispersive XRF spectra of the dry samples were obtained on the IDEAS beamline.  IDEAS is a basic bending magnetic beamline on port 08B2-1 at the CLS.  The simple design of this beamline has no focusing optics and directs the light source through the mounted sample at a constant height through a double crystal monochromator with an energy range of 1.8 keV to 12.5 keV using InSb (111) and Ge (111) crystal pairs.  XRF scans of the MFT samples were taken in air using a single element silicon drift detector.  82  3.4.14 Scanning Electron Microscopy Samples of PA-MFT flocculated in September 2014 and in November 2014 were imaged in late November and in December 2014 using the Hitachi S-4700 field emission scanning electron microscope (FESEM) operated by the UBC BioImaging Facility.  To investigate the location of water within the sample fabric, samples of PA-MFT were also imaged using a cryo-stage unit mounted on the Hitachi S-4700.  Detailed description of sample dehydration, mounting and imaging procedures are provided in Appendix B.11.  83  Chapter 4: Results 4.1 Baseline Characterization of PA-MFT and its Component Materials Table 4.1 summarizes the index properties of raw AITF-MFT and PA-MFT determined by the author and from testing completed by Exova.  The MFT used in this research had an average bitumen content of 2.7% by weight, contained approximately 30% solids, with water comprising the remaining 68% of its mass.  Bitumen content was determined at Exova’s Calgary laboratory using the Dean Stark extraction procedure.  While Exova’s determination of 29% solids content was based on testing four samples, the author’s estimated solids content of 31% was based on eighty samples from the bulk raw AITF-MFT tested between May 6, 2014 and January 7, 2015. Table 4.1: Summary of Raw AITF-MFT and PA-MFT Index Properties    UBC Mining Sfc. Chem Laboratory Exova Laboratory AITF Data Raw AITF-MFT % Bitumen (by weight) Not measured 2.7 % Solids (ms/mt, by weight) 31 29 % Water (by weight) Not measured 68 Geotechnical MC (mw/ms, %) 220 Not measured  Density, ρ (g/cm3)1 1.2 Not measured MBI (meq/100 g solids, CANMET) 10 8 % Clay (wt %, Sethi 1995) 75 57 CWR = (% Clay x wt% Solid)/wt% Water 0.32 0.24 D90 (µm); D50 (µm); D10 (µm)  (Laser Diffraction) Not measured Exova: 50.1; 7.8; 1.3 AITF: 41.8; 7.4; 0.9 Gs (Yao et al., 2014) Not measured 2.3 PA-MFT Initial % Solids (by weight) 33.5 Not measured Initial MC  200 Not measured Gs 2.3 Not measured 1 At indicated solids and geotechnical moisture content  84  Additionally, Exova’s particle size distribution data determined by laser diffraction were on average 23% larger than similar data provided by AITF for the raw MFT sample.  Both sets of laser diffraction data are provided in Appendix A.   4.1.1 Methylene Blue Index Results Using the Canmet slurry method (NRC, 2008) to determine the MBI, Exova estimated an MBI of 8 meq/100 g solids based on 4 tests compared to the author’s estimate of 10 meq/100 g solids based on 19 tests.  While reasons for differences between the estimated MBI values between the two laboratories are unconfirmed, differences in sonication and sample dispersion time likely contribute to the differences between the data sets.  Using the percent clay equation developed by Sethi (1995), the Exova MBI value corresponds to a clay content of approximately 57% while the UBC MBI value estimates clay content of approximately 75%.  The difference in these values causes the calculated clay-water ratio to range between 0.32 and 0.24 (Table 4.1).  Using the OFITE method, the MBI was determined to be approximately 13 meq/100 g solids which translates into an estimated average clay content of 94%.  For the OFITE method, it was also demonstrated that increasing the dispersion time from 10 minutes to 15 minutes resulted in a 2% increase in the measured MBI.  While the range of clay contents estimated by both the Canmet and OFITE methods are consistent with findings reported by Estepho (2014) for an industrial MFT sample that exhibits consolidation characteristics similar to the AITF-MFT, the differences in the dispersion chemicals and methods used by both are likely the primary contributing factors for the range in MBI values for the two methods.  In addition, use of the 0.01N methylene blue solution specified in the OFITE method compared to the 0.006N methylene blue solution used in the Canmet method may also affect the results obtained.  These factors may combine to result in the OFITE method’s apparent  85  consistent over-prediction of MBI value compared to the results obtained using the Canmet (2008) method in this research. Samples of freshly flocculated PA-MFT had MBI values that were 10% (Canmet method) and 15% (OFITE method) higher than the MBI values of unflocculated MFT.  While it might be expected that PA-MFT would likely exhibit MBI values lower than its component raw MFT (Omotoso, 2014a), the elevated values obtained in this research may provide a means of characterizing and confirming the bridging flocculation mechanisms at play when the specific MFT and polymer used in this research are combined.  The elevated MBI values supports the understanding of bridging flocculation as a mechanism in which the polymer does not completely coat or engulf clay mineral surfaces but rather attaches at discrete locations on mineral surfaces.  The result also suggests that under the specific test conditions (pH and ionic environment) the polymer has at least select mineral surfaces on which methylene blue may adsorb leaving some of the polymer branches available to react with the methylene blue.  The MB cation may also be displacing cations included in the PAM formulation or bonding with vacant negatively charged surfaces within the polyacrylamide’s molecular structure. 4.1.2 Specific Gravity and Other Bitumen Effects on PA-MFT Phase Relationships The specific gravity of PA-MFT tested was determined to be 2.3.  Results from the six tests completed on samples of air dried and oven dried PA-MFT are summarized in Table 4.1.  This value is consistent with measurements reported for raw MFT e.g. Yao et al (2014), but is lower than values of flocculated MFT reported (Gs = 2.5) or assumed (Gs = 2.65) by Soleimani et al. (2014), Znidarcic (2014).  It should be noted that geotechnical engineers typically assign a specific gravity of 2.65 to mineral solids to complete a range of soil mechanics calculations (Mitchell and Soga, 2005) even though kaolinite, for example, can actually have a specific gravity ranging between 2.16 and 2.68 (Lambe and Whitman, 1969).  However, use of this assumed value can be  86  helpful as long as the specific gravity of all material components of PA-MFT are also accounted for. To check the validity of the result obtained, a weighted specific gravity accounting for both mineral solids and bitumen content of the tested samples was determined as follows: the combined mass of mineral solids and bitumen (determined from the Dean Stark extraction) was divided by the calculated total volume occupied by the same two components (determined using a specific gravity of 2.65 for mineral solids, and a specific gravity of 1.0 for both bitumen and water) (Read and Whiteoak, 2003).  Based on the bitumen and solid content of the sample, a combined specific gravity of 2.32 was calculated which closely corresponds to the average measured value.  While the value obtained was initially believed to be largely attributed to the numerous voids that are observed to be part of the fabric of dry PA-MFT, the validation calculation demonstrates the impact of bitumen content on measured specific gravity (DOA, 2001). As the solids content of MFT increases through either direct evaporative drying or a combination of thickening using a polymer followed by air drying,  the volume of voids occupied by bitumen increases.  Table 4.2 summarizes the results from a simple analytical spreadsheet developed by Hockley (2015) evaluating the author’s MFT characterization data.  The table shows how the void space occupied by bitumen changes as the target percent solids ranges between 70% and 85% (the identified target range for MFT treatment technologies) when MFT with the characteristics similar to that used in this research (Table 4.1) are used.  Results from this simple analysis show that while increased solids content is desirable, the amount of voids occupied by bitumen is significant and has potential implications for material plasticity (discussed later in this chapter), a barrier of flow within the material fabric, and as a possible impedance to evaporative drying evident in the formation of a distinct bituminous layer  87  immediately beneath the zone where PA-MFT crust formation occurs (typically at a depth ranging between 8 cm and 10 cm below the material surface.  Permeability is also decreased by a factor of 10 when greater than 20% of material pores are filled with heavy oil (Ashrafi et al., 2012). Table 4.2: Relationship between Bitumen Occupied Pores and Target Percent Solids Target % Solids 70 75 80 85 Mineral solids (% mass) 64 68.6 73.2 77.8 Bitumen (% mass) 6 6.4 6.8 7.2 Water (% mass) 30 25 20 15 Mineral solids (% volume) 40.2 45.2 50.7 56.9 Bitumen (% volume) 9.9 11.2 12.5 14 Water (% volume) 49.9 43.6 36.7 29.1 % Voids Occupied by Bitumen 17 20 25 33  4.1.3 Mineralogy of PA-MFT Figure 4.1 provides a visual representation of the dominant minerals found in the PA-MFT produced in the author’s UBC laboratory.  Consistency between clay mineral determination using QEMSCAN and MBI determination completed by Exova is notable.  This finding suggests that the QEMSCAN may provide a means of confirming the percent clay determination using the Canmet MBI Slurry method.  QEMSCAN quantification of carbon and organic material also suggests the presence of bitumen in PA-MFT and is on the same order of magnitude of the bitumen content determined using Dean Stark extraction. Congruence exists between the QEMSCAN and chemical assay completed by SGS to confirm the elemental composition of the samples tested.  XRD analysis of PA-MFT confirmed that kaolinite constitutes the bulk (approximately 88%) of the clay minerals present, while illite represents approximately 7% of the sample’s clay minerals.  QEMSCAN and XRD data reports prepared by the SGS assay laboratory in Burnaby are provided in Appendix B.4.    88   Figure 4.1: QEMSCAN Determination of Mineral Abundance in PA-MFT  4.2 Surface Characterization of a PA-MFT 4.2.1 Zeta Potential Generalized descriptions of the material colloidal behavior provided in Table 4.3 were developed by Zeta-Meter (1975), Greenwood and Kendall (1999) and Hanaor et al. (2012).  Actual zeta potential ranges and values recorded for each suspension tested are provided in Table 4.3 and Table 4.4 respectively.    89  Table 4.3: Indication of Material Colloid Stability   90  Table 4.4: Summary of Zeta Potential for Tested Suspensions        91  Hydrated PAM, raw AITF-MFT and freshly flocculated PA-MFT have zeta potential values ranging between -23 mV and -25 mV and exhibit incipient colloidal stability.  These results suggest that the addition of PAM polymer does little to affect the intrinsic colloidal stability or electrochemical characteristics of raw MFT even though some particle agglomeration occurs.  Decreases in moisture content of raw AITF-MFT and PA-MFT also appears to make them less electronegative by between 3 mV and 6 mV.  Water initially released after MFT flocculation and bitumen skimmed from the surface of PA-MFT, have zeta potential values ranging between -15 mV and -17 mV, making them slightly less electronegative than the input raw materials and PA-MFT. The finding that PAM and raw MFT have similar electronegative characteristics has implications for the nature and degree of bonding that is possible when these materials are combined, as well as implications for the strength and nature of the resulting material fabric.  Hydrophobic-hydrophilic interactions as well as ionic repulsion between various parts of the acrylamide polymer and hydrocarbon polymer chains present in the bitumen asphaltenes also likely have an impact on the resulting material fabric. While diluting raw MFT and PA-MFT with primary extraction water impacts the zeta potential values measured in this research, the study nonetheless enables the behaviour of colloids present in PEW to be assessed.  Moreover, zeta potential results from this research coincide with values obtained by Salehi (2010) for pre and post flocculation MFT.  These results are also consistent with the generalized finding by Zeta-Meter, Inc. that the majority of organic and inorganic colloids, especially those subjected to industrial processes, are electronegative—exhibiting a zeta potential ranging between -15 mV and -30 mV—when suspended in water with an ionic concentration between 50-1,000 µS and pH between 5.7 and 9.0.  92  Kasperski (1992), Alamgir (2011) and Najafi et al. (2014) describe research investigating the combination of cationic polymer with MFT which shows that bonding mechanisms between cationic polymers and oil sands clay minerals in suspension are no simpler than when anionic polymers are used and that flocculation challenges are not easily eliminated by combining materials whose bulk charges are opposite.  These findings suggest that any chemical additives considered for inclusion in the management of oil sands fine tailings should promote greater conformity and agglomeration between clay minerals and not be significantly impeded by the presence of residual bitumen, while significantly reducing the zeta potential of colloids in suspension at a pH ranging between 8 and 9 to optimize reuse of PEW in extraction processes. 4.2.2 Investigation of Functional Groups Using Spectroscopy  The changes in the surface characteristics of BASF 5250 and raw AITF-MFT before and after their combination to form PA-MFT were qualitatively investigated using FT-IR (Figure 4.2) and Raman spectroscopy (Figure 4.3).  Functional group identification was completed using frequency correlation charts and spectra identification data found in Socrates (1980), Stuart (1997), Murugan et al. (1998), and Michaelian (2003). Table 4.5 summarizes the distinguishable frequencies identified for PA-MFT, BASF 5250 polymer, and raw AITF-MFT.  The spectra of toluene extracted bitumen (TEB) from AITF-MFT was also obtained to identify the characteristic frequencies associated with hydrocarbons present in both raw AITF-MFT and PA-MFT.  FT-IR spectra for pure kaolinite and quartz developed by Gutierrez (2013) were referenced to identify unique frequencies associated with those minerals that are present in both raw AITF-MFT and the resulting PA-MFT.   Analysis of spectra presented in Figure 4.2 suggests that residual bitumen affects the surface characteristics of both raw and flocculated MFT.  This is most notable in the presence of bitumen’s characteristic frequency between 960 cm-1 and 1,060 cm-1 (labelled B5 on Figure 4.2)  93  in the spectra of both raw MFT and in the PA-MFT sample identified as having more visible bitumen.  The presence of aliphatic C-H stretching bonds and CH2 and CH3 functional groups present in the 2,800 cm-1 to 3,000 cm-1 frequency range (labelled B1 on Figure 4.2) was also evident in the spectra for both raw MFT and PA-MFT.  The fact that frequencies in the identified B1 range are amplified in the air dried PA-MFT spectra when compared to the spectra for raw MFT, suggests that the CH and CH2 frequencies of the polymer spectra (labelled P2 on Figure 4.2) increase the total organic content contributed by these functional groups during flocculation.  This could also indicate additional liberation of bitumen which has been observed during polymer flocculation of MFT.  That these functional groups concentrate in the flocculated solids and not to the same degree in the release water (though this effect is present in the Raman spectra discussed later in this section), is also noteworthy as it suggests that these functional groups act as a coating on the surface of flocculated minerals.  It is also notable that the intensity of frequencies in this band were least pronounced for the PA-MFT sample identified as having the most visible bitumen.  This finding suggests that visual quantification of bitumen may provide an inadequate and incomplete indication of the degree to which these hydrocarbon functional groups impact the various PA-MFT dewatering phases. Bitumen-associated C-H deformation vibrations and CH2 and CH3 functional groups in the 1,400 cm-1 to 1,500 cm-1 frequency band (labelled B3 on Figure 4.2) and frequencies associated with the CH2 functional group in the 1,225 cm-1 to 1,400 cm-1 frequency band (labelled B4 on Figure 4.2) are present in the spectra of both raw and flocculated MFT, although their presence appears to be less pronounced than the effects of both the B1 and B5 functional groups.  Hydrocarbon functional groups in the B3 range appear to be at least partially hydrophilic and soluble in water as this frequency response is present in the spectra of water initially released  94  after MFT flocculation.  Hydrocarbon functional groups in the B1, B4 and B5 frequency ranges appear to be strongly hydrophobic and concentrate in the flocculated solids.  This finding of hydrophobic functional groups on the surfaces of fine tailings solids and at least one hydrophilic hydrocarbon functional group has implications for flocculation mechanisms with respect to polymer-mineral bond locations and possible polymer-bitumen interactions.  This finding also has an effect on the dewatering characteristics exhibited by PA-MFT as will be discussed in Section 4.2.6. Finally, the frequency of the bitumen-associated C=C functional group at 1,600 cm-1 (labelled B2, Figure 4.2) was most apparent in the Raman spectra of raw and flocculated MFT.  This functional group persists relatively undiminished in both raw MFT and throughout the dewatering stages of PA-MFT (Figure 4.3).  The presence of this C=C functional group is also evident in the Raman spectra of the water initially released after MFT flocculation.  The presence of hydrocarbon functional groups in the release water is also evident in the 2,800 cm-1 to 3,000 cm-1 range of the Raman spectra.  However, as was discussed earlier in this Section, the response in the release water spectra could result from organic functional groups originating from the polymer functional groups labelled P2 on Figure 4.2.  The Raman spectra of air dried PA-MFT also indicates that the C=C functional group can persist quite strongly even in the absence of bitumen visible to the unaided human eye, as the most pronounced frequency was obtained in the sample lacking obvious visible bitumen.  The persistence of this signal also suggest that the C=C functional group may also indicate the presence of poly aromatic hydrocarbons coating PA-MFT mineral surfaces.    95   Figure 4.2: FT-IR Spectra for PA-MFT and its Component Materials  96   Figure 4.3: Raman Spectra for PA-MFT and its Component Materials  97  Table 4.5: Summary of Observed Frequencies from IR and Raman Spectra for PA-MFT and Its Component Materials      98  The presence of OH and Al-OH kaolinite functional groups were most evident in the IR spectra.  OH functional groups identified by frequencies in the 3,600 cm-1 to 3,700 cm-1 range (labelled K1 on Figure 4.2) remained present on surfaces of both raw MFT and air dried PA-MFT.  This functional group was also present in the toluene extracted bitumen spectra, suggesting the presence of kaolinite clay minerals in the extracted bitumen.  It is notable that where more visible bitumen was observed the signal intensity of this functional group diminished.  This finding suggests that the OH functional group dominates surface characteristics of fine tailings solids where visible bitumen is either completely absent or is only present in limited quantities.  Conversely, the signal of the Al-OH functional group at 900 cm-1 (labelled K2 on Figure 4.2) was most pronounced in raw MFT and in the PA-MFT sample with more visible bitumen.  This finding suggests that different kaolinite functional groups govern surface interactions based on the amount of bitumen present in the sample.  Using the spectra developed by Gutierrez (2013) for a standard quartz specimen, the presence of quartz in samples of raw MFT and PA-MFT is confirmed by the distinctive frequencies at 1,100 cm-1 and between 700 cm-1 and 800 cm-1. The spectra for BASF 5250 PAM is characterized by distinctive frequencies summarized in Table 4.5, and labelled P1-P3, P6-P7 (Figure 4.2), and P1-P2, P4 - P7 (Figure 4.3).  The effect of this polymer on the surfaces of PA-MFT appears to be limited to the effect of the P2 organic functional groups previously described, and the occasional presence of the C-O functional group visible at 1,700 cm-1 on the Raman spectra (labelled P4 on Figure 4.3).  The presence of this frequency may indicate the presence of excess polymer in one of the samples tested.  However, the limited occurrence of this frequency in the PA-MFT spectra suggests that polymer functional groups do not play a significant role in the dewatering of MFT beyond the initial water released after flocculation.  99  In addition to identifying the primary functional group frequencies of the BASF 5250 PAM, the Raman spectra also indicated the first overtones of the polymer’s CH and CH2 functional groups as well as the first overtone of C-H stretching and CH2 and CH3 functional groups associated with bitumen.  The first overtone of the bitumen-related functional groups was also present in all samples of raw and flocculated MFT tested. 4.2.3 Quantification of PA-MFT Surface Area Covered by Visible Bitumen Given the apparent role of hydrocarbon functional groups in PA-MFT surface chemistry, images taken using the 50 times magnification optic lens of the Raman spectrometer were analyzed using ImageJ, a public domain Java image processing and visual analysis program developed by the National Institutes of Health (Schneider and Rasband, 2012).  Bitumen was estimated to cover between 17% and 20% of the surfaces of dried raw MFT and PA-MFT based on estimates from four images of the sample materials for which Raman spectra were also developed.  In these images visible bitumen appears in the form of black surfaces ranging from <1 µm to 0.2 mm wide (Figure 4.4).  Figure 4.4: Surface of Air Dried PA-MFT  100  This finding of wide bitumen dispersion throughout PA-MFT is important as it suggests that merely quantifying the amount of visible bitumen by weight (estimated as 2.7% for the raw AITF-MFT used in this research) is inadequate and fails to account for either surface area coverage by bitumen or the prominence of bituminous functional groups on PA-MFT surfaces.  When combined with the IR and Raman spectra for PA-MFT, including information about the functional groups present in the sample with less visible bitumen, visual analysis data supports the earlier finding that hydrocarbon functional groups associated with bitumen potentially play a significant role in impacting the dewatering characteristics of PA-MFT after the initial release of water associated with flocculation. 4.2.4 X-Ray Fluorescence Analysis Using the IDEAS beamline described in Chapter 3 the XRF scans of samples of wet and dry MFT, and a sample of dry PA-MFT (Figure 4.5) were obtained.    Figure 4.5: XRF Scans of Raw AITF-MFT and PA-MFT  101   At energies ranging between 2 keV and 10.6 keV the relative abundance of elements with atomic mass ranging between Argon (Ar) and Arsenic (As) are identified.  Results of the scan indicate that iron (Fe) is the most abundant element in this energy range compared to the other elements able to be detected in this energy range.  The scans show that the intensity of the Fe signal is greatest for dry PA-MFT followed by the Fe signals of dry MFT and wet MFT.  This finding suggests that the signal intensity of Fe increases as it exists in increasingly oxidized states.  This finding of abundant iron is consistent with the observation made by the author of the increasingly orange color of a sample of raw AITF-MFT that had been placed in a small plastic container with a plastic threaded lid leaving the container imperfectly sealed against the ingress of atmospheric oxygen.  The intensity of the brownish-orange color increased with time and coated the walls of the plastic container and had the appearance of rust.  XRF also confirmed that Titanium (Ti), Manganese (Mn), Zinc (Zn) and Gallium (Ga) were the next four abundant elements in the samples scanned.  In addition to Zn, the scans also indicated the presence of Chromium (Cr) and Copper (Cu).  Use of these findings to identify potential contaminants of concern and to complement geochemical characterization and analysis of seepage water quality from the associated waste materials, should be explored further and will be discussed in Chapter 8. 4.2.5 X-ray Sulphur Absorption Analysis Using the SXRM beamline described in Chapter 3 the sulphur absorption spectra at the K edge were obtained for a sample of dry polymer, and for wet and oven dried samples of raw AITF-MFT and PA-MFT.  Sulphur absorption at the L edge was also attempted, but the sulphur concentration in the samples was too weak to be observed under the high vacuum and ultra-dry conditions under which that analysis must be conducted.  Figure 4.6 presents the spectra obtained.  Elemental sulphur was used as a reference material to calibrate and normalize the spectra of the  102  sample materials investigated.  Using the SXRM beamline the primary/white line absorption edge/peak for elemental Sulphur occurred at approximately 2,471.5 eV. Sarret et al. (1999) indicate that at the K-edge the energy position and the amplitude of the Sulphur absorption edge increases in the presence of a state of increasing oxidation.  Using Sulphur spectra presented in both Sarret et al. (1999) and Castillo (2009), the likely species indicated in the collected spectra are summarized in Table 4.6. .    Figure 4.6: Sulphur Absorption Spectra of PA-MFT and Its Component Materials  Analysis of the Sulphur absorption spectra indicates the following: • Very distinct spectra for polymer with a primary peak at 2,481.6 eV that likely indicates presence of sulphates. • All samples of MFT and PA-MFT share a primary peak at 2,473 eV that likely indicates the presence of dibenzothiophenes (DBT).  The prominence of this peak  103  in wet samples of MFT and PA-MFT suggests that thiophene-type molecules are dominant in these samples. • Dry samples of MFT and PA-MFT also share the following: a secondary peak at 2,475 eV indicative of sulfoxides that appears more prevalent in the sample of dry MFT; a smaller sulfone peak at 2,476.8 eV; and a more significant peak at 2,481.6 eV that coincides with the location of the primary sulfate peak of the polymer’s Sulfur spectra and that is most pronounced in the sample of dry PA-MFT.  Table 4.6: Summary of Likely Sulphur Species Identified in Sulphur K-edge Absorption Spectra  Peak energy (eV) Likely species S Oxidation state 2471.8 Disulfide, elemental S 0 2473 Dibenzothiophenes 0 2475 Sulfoxide 2 2476.8 Sulfone 4 2480 Sulfonic acids 5 2481.6 Sulfates 6 The appearance of the sulfate peak in dry MFT suggests that sulfates are present in MFT and that the presence of these compounds is amplified by the addition of sulfates contributed by the hydrated PAM polymer.  If a formulation of BASF 5250 is consistent with the composition of Magnafloc 1101, then the sulphate peak that appears in the dry PA-MFT spectra may result from interaction between the sodium acrylate component of the polymer and the sulphate present in MFT, or with the sulfate present in the PEW used to hydrate the polymer.  The presence of high oxidation state sulfur species (sulfoxide, sulfone, sulfonic acid—in  the case of dry MFT—and sulfates), suggests that the dry samples of MFT and PA-MFT become oxidized during the drying process, which may well be exacerbated by the heat associated with drying the samples in an oven.   104  The effect of heat on the degree of oxidation observed is an investigation that can be conducted in the future.  However, the abundance of oxidized sulfur forms, and sulfates in particular, could have implications for long-term seepage water quality associated with these materials and how they are incorporated into the closure landscape of oil sands surface mines. Consistent with the findings of this investigation, an investigation by Sarret et al. (1999) into the forms of Sulfur found in select asphaltene deposits using K-edge XANES, also identified DBT as the most abundant form of non-oxidized Sulfur present while oxidized Sulfur forms were dominated by sulfoxides, sulfones, and sulfates.  The similarities between the spectra of dry PA-MFT and the spectra of the Bic 374 bitumen sample from the Jebel Bichri region of Syria are also noteworthy (Sarret et al, 1999).  In that sample, reduced Sulfur species accounted for 79% of the sulfur species present while sulfoxides, sulfones, and sulfates accounted for 12%, 10% and 4% respectively of the remaining Sulfur species. 4.2.6 Effects of Surface Characteristics on Dewatering of a PA-MFT The results of the investigation into the zeta potential of colloids found in both raw AITF-MFT and PA-MFT suggests that optimized agglomeration of the colloidal particles in MFT could result if they are combined with a chemical reagent or exposed to a treatment process that significantly reduces the electronegativity exhibited by the colloids in suspension.  Ideally the chemical amendment would need to be unaffected by the presence of residual bitumen that, at a minimum, coats the surfaces of some clay minerals and exhibits a hydrophobic response.  Agglomeration would also need to occur at a pH between 8 and 9 if the released water was to be reusable in the bitumen extraction process.  Research completed by Najafi et al. (2014) and Omotoso and Melanson (2014) indicate that complex surface chemistry exhibited by MFT colloids would not be effectively overcome by the simple addition of acid to decrease pH.  However, given the need to significantly alter the surface charge of MFT colloids, it may be plausible to consider  105  treatment of MFT in a process that produces agglomeration of fine particles and treatment of release water to a quality that would be either suitable for use in steam generation, or discharge to the environment. Results of IR and Raman spectroscopy indicate and confirm that surface characteristics of PA-MFT are impacted by functional groups associated with both kaolinite minerals and bituminous asphaltenes, with bitumen and kaolinite functional groups occurring at a frequency of 2 to 1 respectively.  The finding that the prominence of different kaolinite-associated functional groups is impacted by the amount of residual bitumen and the abundance of various hydrocarbon functional groups, suggests that chemical amendments would need to be tailored to target kaolinite functional groups to promote clay mineral agglomeration, while providing a dispersing or emulsifying agent to prevent hydrophobic hydrocarbon functional groups from impeding effective clay mineral agglomeration.  However, observations by Johnson (2015) into the differences exhibited by clay minerals in oil sands ores, compared to properties of reference clay minerals, combined with the complex surface chemistry exhibited by fine tailings solids, does not preclude the existence of hydro-carbon altered clay minerals that innately exhibit both electronegative and hydrophobic responses that are not easily impacted by demulsifying agents.  The IR and Raman spectra results suggest that the presence of bitumen functional groups as a coating on the surface of clay minerals may not only impede the ability of anionic polymers like the one used in this research to create particle bridges strong enough to effectively agglomerate clay minerals but worse, that the presence of hydrophobic bitumen may further impede egress of water from PA-MFT voids and pore throats.  The ability of hydrophobic functional groups to further limit the effectiveness of the PA-MFT pore network impacts the long-term dewatering characteristics of  106  these materials and, by implication, the ability to effectively place and store this material in thick lifts for operational fine tailings management or during closure and site reclamation. While preliminary in nature, findings from XRF and K-edge Sulfur absorption spectra of PA-MFT provide input for future investigations into contaminants of potential concern, namely Fe, Cr, Cu and SO42-.  An understanding of the ability of these sources to impact the seepage water quality of PA-MFT is critical to understanding the associated closure, cost and risk implications. 4.3 Rheological Characterization of a PA-MFT and its Component Materials 4.3.1 Flow Curve Characteristics of BASF 5250 and Pre-Sheared AITF-MFT The flow curves of hydrated BASF 5250 and raw AITF-MFT are provided in Figure 4.7 and Figure 4.8, respectively.  These curves represent the averages obtained from the duplicate tests completed for both materials.  Hydrated PAM exhibits pseudoplastic shear thinning behavior described by the Ostwald Power law model (Klein and Hallbom, 2014) and expressed by Equation 4.1: 𝝉𝝉 = 𝑲𝑲?̇?𝜸𝑻𝑻  Equation 4.1 in which shear stress (τ) is taken as the product of a material consistency parameter (K) and the shear rate (?̇?𝛾) raised to the dimensionless power-law index (n).  When n is less than 1, the model indicates shear thinning behavior. Conversely, when n is greater than 1, shear thickening behavior is described.  The flow curve of hydrated polymer results in its classification as a Power Law fluid. Prior to testing, a sample of raw AITF-MFT was subjected to a period of pre-shearing and rest as described in Chapter 3. Therefore, the material’s stress state prior to testing could be described and be consistent with the conditioning provided to raw AITF-MFT prior to its combination with hydrated polymer to create PA-MFT.  Based on the average results shown in  107  Figure 4.8, raw AITF-MFT used in this research behaves like a Bingham plastic and exhibits an apparent initial yield stress of 2.4 Pa.  Using a fixed rotational speed of 0.1 rpm the peak yield stress of this material was estimated to be 7.4 Pa which was reached after approximately one minute.  While the thixotropy of MFT has been investigated and described by others (Scott et al., 1985; Jeeravipoolvarn, 2010) the flow curve of AITF-MFT was developed to investigate flow behavior of MFT as it exists at the time it is combined with anionic PAM to create PA-MFT.  Figure 4.7: Flow Curve of Hydrated Anionic Polyacrylamide Polymer   108   Figure 4.8: Flow Curve of Conditioned Raw AITF-MFT Figure 4.8 shows the fit achieved by both the Bingham plastic and Herschel Bulkley flow models to the AITF-MFT flow curve data.  The flow curve also indicates that AITF-MFT exhibits a post peak shear stress that plateaus at approximately 2 Pa, after the test had been running for approximately 10 minutes. 4.3.2 PA-MFT Peak Yield Stress The shear stress versus time profiles for samples of PA-MFT flocculated by the author in September 2014 were determined using a Brookfield DV3T rheometer with an HB spring, rated to measure shear stresses between 4 Pa and 3,200 Pa when used with V73 and V75 vane spindles fabricated by Brookfield.  The indicated vane spindles were used with the universal cup geometry designed by Coanda for use with the Brookfield rheometer.  Geometry of the vane spindles and universal cup are summarized in Table 4.7.  PA-MFT flow curve testing was completed on September 8, 2014, September 23, 2014 and on October 6, 2014 to characterize both the variability in the initial peak strength and the shear stress response of PA-MFT. The testing also assessed how  109  the initial peak strength and subsequent shear stress profiles were affected by duration between time of flocculation and rheological testing.  Figure 4.9 presents the shear stress versus time profiles obtained for the nine samples tested, while Table 4.7 summarizes the test data.  The designation ‘wet’ was given to samples that continued to produce release water in the sample container after the initial water released at the time of flocculation had been decanted, prior to the sample being poured into its long-term storage container.  Samples that did not produce additional release water after the initial release water at the time of flocculation had been decanted were designated as “dry”. In all cases, flocculated materials were directly poured from the beaker in which the MFT was flocculated into the 1 L glass mason jars used to store the samples prior to rheological testing.  Each mason jar was sealed with a metal screw cap containing a plastic vacuum seal which limited evaporative losses.  The PA-MFT samples tested were used to represent the various material conditions that exist in the field where removal of initial post-flocculation release water varies based on sample deposition conditions, effectiveness of cell configuration and the variability associated with the irregularity of PA-MFT surfaces that can easily enable local ponding of release water. Data summarized in Table 4.7 suggest that an absence of effective ways of removing water released after flocculation leads to an apparent plateau of the percent solids and geotechnical moisture content of PA-MFT even 30 days after flocculation.  This observation was explored in the evaporation test cell research described later in this Chapter.  Review of the profiles presented in Figure 4.9 indicates that both peak shear stress values and shear stress versus time profiles vary widely for the tested materials.   110  Table 4.7: Summary of PA-MFT Shear Stress Test Data   Figure 4.9: PA-MFT Shear Stress Versus Time Curves  111  Percent solids typically range between 39% and 43% for the majority of samples tested (except for the R1 wet sample tested on September 8, 2014), and measured peak shear stress ranges between approximately 350 Pa and 870 Pa.  Test results also indicate that increased time between flocculation and testing does not consistently translate to an increase in the measured peak shear stress.  The data also indicates that measured increases in peak shear stress do not necessarily correspond to increases in solids content or to corresponding decreases in geotechnical moisture content, except in the case of the R3 wet sample in which the peak shear stress increased by approximately 14% over a 15 day period.  This finding, combined with the observation that different peak shear stress values can be measured for samples with the same solids content, suggests that solids content is not the sole predictor of measured shear stress. It further suggests that sample handling, particularly when loading the sample cup prior to testing, is a significant factor in the peak shear stress value measured. In addition, the wide variability in peak shear stress values obtained suggests that PA-MFT flocs may be quite variable in nature and easily susceptible to breakage based not only on their dewatering stage or the drainage condition in which the flocs are stored, but also on the condition of the flocs themselves.  Moreover, the collected data suggests that even though a nominal increase in percent solids may occur under non-ideal drainage conditions, keeping PA-MFT solids in contact with released water can result in degradation of floc strength (Table 4.7).  This was observed in the case of the R1 wet sample where a 38% decrease in peak shear strength was observed, even though the percent solids content of the sample increased by 18% and the geotechnical moisture content decreased by 24% over the values measured initially. The three highest peak shear stress values were recorded for samples that had been minimally disturbed prior to testing and during loading of the sample cup used during testing.  The  112  shear stress profiles for testing the R1 dry and R3 wet samples completed on September 8, 2014, and the R2wet, t1 sample tested on October 6, 2014 suggests that either different flocs were encountered during testing, or that the occurrence of inter-floc aggregation resulted in the appearance of secondary and tertiary local maxima after the peak shear stress values had been measured.  The findings from this investigation, especially those related to the effects which material handling have on measured shear strength, impact the consideration of possible effects which the processes of in-line MFT flocculation and subaerial deposition have on the initial and ultimate strength of PA-MFT.  Indeed, the fact that utilization of drying cell area relies upon flow of these materials suggests that the ability for flocs broken down during deposition to regain strength is a factor when considering the ultimate strength that can be plausibly achieved by these materials, especially if undrained conditions persist.  Moreover, the general shape of the shear stress versus time profiles for PA-MFT samples tested in this research suggests that the occurrence of rheomalaxis (irreversible degradation in viscosity during shearing previously described in Section 2.11) is not a trivial consideration.  This is especially true when considered along with the challenges associated with limited particle agglomeration that occurs during flocculation, as discussed previously.  In the case of PA-MFT, this suggests that in the absence of a chemical or physical mechanism to induce strength to the mineral particles, the geotechnical strength characteristics of the bulk materials could be variable and effected by the degree of floc breakage that occurs during mixing, deposition or subsequent handling of these materials.  It is valuable to consider the potentially adverse effects of impeded drainage as the bulky, globular nature of PA-MFT results in the creation of undulating surfaces that enable water to pond locally and more widely on the surfaces of PA-MFT deposited in drying cells.  As discussed earlier,  113  the possibility of reduced permeability associated with having bitumen occupy upwards of 17% of the material voids (Ashrafi et al., 2012), must also be considered as it directly impacts the rate at which water contained in the PA-MFT fabric can leave the material. The apparent plateauing or degradation of shear strength associated with water retention further underscores the need to ensure both positive cell drainage, and also the avoidance and mitigation of areas of large ponding that tend to result at the far end of deposition cells during and after active PA-MFT deposition using spigots.  The variability in shear stress profiles independent of both percent solids and geotechnical moisture content, supports the hypothesis and recommendation by Omotoso and Melanson (2014) that while polymer dosing on a percent solids basis may be convenient, dosing on a percent clay mineral basis may be more appropriate. It provides a more direct means of accounting for variable clay mineral content present in different MFT samples extracted from a single storage pond.  This apparent intrinsic heterogeneity that exists in MFT, correlates to oil sands ore heterogeneity and supports the additional recommendation by Omotoso and Melanson (2014) and others for a need in real time to quantify clay mineralogy on MFT flocculation treatment trains. While a slump test was also used to determine the yield stress of PA-MFT, the effects of handling on the results appear to be even more pronounced than the effects observed using rheometer testing.  In the slump test the act of filling the Boger ring results in breakage of the bulk PA-MFT material.  This was most notable where attempts were made to limit the presence of large artificial voids that would serve as artificial weak points affecting slump test measurements.  For the three slump tests successfully completed, yield stress measurements ranged between 124 Pa and 187 Pa.  Test details and results are summarized in Table 4.8.   114  Table 4.8: Summary of Slump Test Results  The significant disturbance experienced by the material during filling of the Boger ring is likely responsible for the under-prediction of PA-MFT yield stress obtained by this method when compared to the results obtained using wide gap geometry with the Brookfield rheometer.  Accuracy of PA-MFT yield stress measurements obtained using this method must also account for possible effects associated with the continued release of significant water contained within the samples at the time of testing.  The effects of sample disturbance were also evident in the non-uniform way in which PA-MFT was observed to slump and a tendency for sample slumping and deformation to initiate along sample discontinuities that were artifacts from loading the sample in the Boger ring.  Experience using this method suggests it would be better suited for use with materials with a more homogeneous composition and at higher solids content than the PA-MFT samples tested. 4.3.3 Oscillatory Rheology Investigation into the viscoelastic properties of PA-MFT was limited to measurement of the material’s response during a sweep of strain between 0.1% and 500% using a frequency of 1 Hz.  This testing was completed on September 23, 2014 for two subsamples of the R3 wet PA-MFT sample which had a solids content of 40% and a geotechnical moisture content of 151% at the time of the test.  This testing was completed to facilitate preliminary investigation into the effects of deformation (correlated in the field to rehandling and reclamation activity) on PA-MFT.  PA-MFT exhibits an elastic solid-like response at strains less than approximately 220% (Figure  115  4.10).  However, once this yield strain value is exceeded, the loss modulus (G”) exceeds the storage modulus (G’) and PA-MFT behaves like a viscous liquid.    Figure 4.10: Storage and Loss Modulus Strain Sweep Profiles  This finding may explain why the author, on a separate occasion at a field test site, sunk into a subaerially deposited 40 cm thick lift of PA-MFT when a twisting motion was used to insert a sample collection tube. However, sinking did not result when the length of time spent standing at any single location on the flocculated tailings surface was limited to a single passing footstep.  Results from this laboratory investigation have potential implications for the reclamation and handling considerations associated with this material.  Exceedance of yield strain values triggered by static or dynamic load associated with operation of operations or reclamation vehicles could result in vehicle entrapment or loss of life.  While the tested materials had a solids content of approximately 40%, which is not typical of surface materials experiencing effective evaporative drying, the potential for similar materials to exist within and throughout a thick lift deposit of PA- 116  MFT could have a profound effect on the ability to reclaim the overall deposit. This is especially true if reclamation work is not limited to times when the deposit is sufficiently frozen to enable operation of mobile equipment.   4.3.4 Effects of Dewatering Condition on Rheological Characteristics of a PA-MFT PA-MFT appears to be highly sensitive to the shear stresses it experiences and exhibits rheomalaxis once the material’s peak shear stress has been attained.  This finding has direct implications for the ability of this material to develop strength when the very mechanisms used to flocculate and deposit these tailings in dedicated drying areas at operating sites likely cause significant breakage of flocs.  While test work reported by Mizani and Simms (2014) suggests that aging can help with the recovery of floc structure, results for this research suggest that shear strength development, especially of materials stored in undrained conditions, may be impacted by the degree of breakage that occurs with these effects, compounded by retention of ponded water on PA-MFT surfaces.  Consequently, the use of solids content as a basis for determining polymer dosage, combined with the mechanical processes of in-line flocculation of MFT followed by subaerial spigotted deposition of PA-MFT, and reliance on the rheological properties of PA-MFT to utilize the storage capacity of drying cells, all have the potential to undermine the ultimate strength development for PA-MFT.  This is especially true when PA-MFT is deposited in lift thicknesses at which evaporative drying on its own cannot increase the solids content to the levels at which the material exhibits a solid elastic response throughout its profile (>85% solids suggested by Mikula (2012)).  Evaporation test cell results described later in this thesis indicate that for the PA-MFT tested, evaporative drying is most effective for materials located within 10 cm to 15 cm of the material surface exposed to the atmosphere. Use of a percent clay basis to determine polymer dose appears to provide a better means of accounting for variation in clay mineral content that may result in a product that exhibits more  117  consistent shear strength properties.  However, the effect of this dosing protocol requires validation with an appropriately designed study.  Moreover, the ability to effectively implement clay-content based dosing is impacted by the absence of a reliable means of determining clay content in real time on industrial scale process streams (Omotoso and Melanson, 2014).  The role of bitumen as a surface coating that likely inhibits the ability of some clay minerals to participate effectively in flocculation processes, and the ability of these materials to decrease shear resistance between mineral particles, must also not be trivialized or discounted and warrants additional investigation. Findings from this research suggest that the as-deposited initial strength of PA-MFT can vary widely based on the degree of floc breakage. It is likely impacted by the local and overall drainage conditions that exist at the location where the spigotted PA-MFT comes to rest.  In the field this translates into variability in measured in-situ shear strength and suggests that drainage and climatic conditions at the time of testing should be collected and reported with measured shear strength data.  The plateauing of measured in-situ shear strength and solids content values could highlight the roles and importance of intrinsic (e.g. fabric), extrinsic (e.g. lift thickness or local ponding associated with deposition of a subsequent lift or rainfall event) or a hybrid combination of the two types of properties. The sensitivity of wet PA-MFT to breakage directly impacts measurements of shear stress using methods that require the material to be loaded into a cup or cylinder.  Rheological measurements on flocculated MFT are further compounded by the production of release water that can take up to two weeks to be completely decanted from a freshly flocculated PA-MFT sample. The ability for PA-MFT to exhibit viscous fluid behavior once its yield strain is exceeded is a significant consideration for reclamation of this material.  In thicker lift deposits, layers or  118  zones of low strength undrained material could results in localized geotechnical instability that could affect operations or closure activities. 4.4 SICT of PA-MFT and Raw AITF-MFT SICT was completed for one sample of raw AITF-MFT and two samples of PA-MFT using the SICT apparatus at the author’s UBC laboratory.  This apparatus was used for the analyses of various MFT samples by Estepho (2014).  Water released within the first eight to nine days after flocculation of the PA-MFT samples tested was removed from the PA-MFT solids prior to placing the test material into the test chamber.  The delay between flocculation and the start of testing enabled the bulk of the water released as a result of flocculation to be removed. This is consistent with observed field conditions following subaerial deposition of PA-MFT into drying cells designed to optimize runoff of initial release water. The SICT test program developed for PA-MFT was designed to capture changes in void ratio after the initially released water has been removed and the sample starts the process of drying resulting from evaporation and underdrainage.  Consequently, the percent solids at the start of SICT of PA-MFT is on average 10% higher than the solids content measured at the time of MFT flocculation.  To preserve the chemical environment and pH of the test material, PEW described in Chapter 3 was used to fill the triaxial chamber and to replace water removed from the test chamber during the test. Testing of raw MFT took 49 days and achieved a final solids content of 70%.  Testing of PA-MFT took approximately 40 days and achieved a final solids content of approximately 71%.  It should be noted that these solids contents are still below the 85% solids indicated by Mikula (2012) as desirable and required application of in excess of 200 kPa to be achieved.  Characteristics of the materials tested, as well as information about the loading steps included in each test, are summarized in Table 4.9.  119  Table 4.9: Summary of Input Material Characteristics and SICT Load Step Information   The initial void ratio of raw MFT was back-calculated using Equation 2.1 to be approximately 5 while the initial void ratio of PA-MFT at the start of the test averaged approximately 3.  Of interest is that the void ratio of both raw AITF-MFT and the associated PA-MFT at the end of the test were very similar and ranged between 1.03 and 1.14.  These void ratios translate into raw AITF-MFT and PA-MFT having an average porosity of 52% which represents a 31% reduction in the initial porosity of raw MFT and a 32% reduction in the porosity of PA-MFT from the start of the test. To determine the void ratio of raw MFT under steady state conditions, pump velocity was incrementally increased from 1x10-4 mm/s to 3x10-4 mm/s.  While a buildup of pore pressure could not be achieved in the first sample of PA-MFT, the remainder of the testing procedure was successfully completed allowing a compressibility relationship to be developed from which an initial void ratio could be calculated.  For the second sample of PA-MFT the pump velocity was ramped up in two stages (Stage I - 1.5 x 10-4 mm/s to 6 x 10-4 mm/sec; Stage II – 1 x 10-3 mm/s to 3 x 10-3 mm/s) to develop adequate pore pressure build up in the sample prior to commencing with the load steps used to calculate the associated permeability. In addition to the increased pump velocities used in the steady state phase of the second PA-MFT test, it was generally observed that pore pressure build up during the load steps in testing of both PA-MFT samples required pump velocities that were an order of magnitude larger than those required to induce similar pore pressure build up in the loading phases of the sample of raw  120  AITF-MFT.  This elevated pump velocity suggests a difference between the fabric of PA-MFT and that of raw MFT such that the fabric of PA-MFT provides greater impendence to water flow than unflocculated raw AITF-MFT.  Profiles showing the pore pressure responses of the various samples under steady state and loading conditions are provided in Appendix C.  The void ratio and parameters used to express the equations describing the resulting compressibility and permeability relationships are summarized in Table 4.10.  Figure 4.11 shows the compressibility relationships that were developed for raw AITF-MFT and PA-MFT. While flocculation produces a visible and measureable reduction in the initial void ratio of raw AITF-MFT, the void ratio of PA-MFT under increasing stress conditions is only marginally lower than the void ratio of unflocculated MFT.  This finding is consistent with the conclusion reported by Yao et al. (2012).  The trend for unflocculated AITF-MFT (Figure 4.11) showed good agreement with and followed the general trends of the MFT samples tests at both UBC and CU Boulder as described by Estepho (2014). PA-MFT provides a nominal increase in permeability over raw MFT (Figure 4.12).  However, the need to use increased seepage pump flow rates to obtain permeability measurements in PA-MFT for each load phase raises questions as to whether similar permeability measurements would be achieved under field conditions.  The permeability curve developed by Pollock (1988) for an MFT sludge using a step loading large strain consolidometer also shows that the addition of polymer flocculants may not result in increases in permeability that provide significant benefits over the permeability that can be achieved in untreated MFT. The SICT Analysis used to produce Figure 4.11 and Figure 4.12 was completed using a specific gravity of 2.65.  As a point of interest, using a specific gravity equal to the apparent specific gravity of 2.3 described earlier in this Chapter, shifts the compressibility curve down by a  121  void ratio increment of 0.5 for stresses less than 0.1 kPa and by a void ratio increment of approximately 0.25 for stresses greater than 10 kPa.  Similarly, use of this lower specific gravity value shifts the permeability curve nominally to the right of the curve shown on Figure 4.12 that would result in the reporting of higher estimated permeability values at each void ratio.  These findings suggest the need for further investigation into the effect that SIC testing has on samples as well as possible need for additional refinement of the formulas used in SICT analysis to estimate permeability.   122  Table 4.10: Summary of SICT Results and Compressibility and Permeability Relationships   Figure 4.11: Compressibility Relationships for Raw AITF-MFT and PA-MFT  123   Figure 4.12: Permeability Relationships for Raw AITF-MFT and PA-MFT  4.5 Dewatering and Surface Deformation Characteristics of a PA-MFT 4.5.1 Dewatering Characteristics of a PA-MFT Two experiments were conducted using the evaporation test cell described in Chapter 3 to investigate the dewatering and undrained strength characteristics of PA-MFT flocculated in the author’s UBC laboratory.  Table 4.11 summarizes the material characteristics and protocols used for both tests.  Figure 4.13 provides a graphical representation of the mass balances for the two tests investigating how samples of PA-MFT placed at a thickness greater than 20 cm dewater under the effects of evaporation and underdrainage without the addition of normal load to the surface of the test material.  As seen from the profile for the unloaded phase (Phase I) from both tests, initial dewatering is governed by drainage losses.  However, drainage losses eventually plateau and additional water loss is governed by evaporation.  While the limits to evaporation are not visible in Figure 4.13, evaporation losses also eventually plateau in the presence of any condition that impedes additional evaporation losses from surface materials (Figure 4.14).  For both tests  124  drainage losses started to plateau after approximately 14 days.  In the case of EC Test 2 in which an additional 43.6 kg was added to the surface of PA-MFT using ten loading plates that were added incrementally over a 30 day period, an additional 0.5 kg of underdrainage resulted prior to reaching a second plateau (Figure 4.14).  The addition of loading plates reduced the rate of evaporation of surface materials during the 30 day loading period (Figure 4.14), and evaporation continued after losses due to drainage had ceased prior to plateauing after a period of approximately 73 days. The general trends of underdrainage and evaporation were consistent for both evaporation cell tests.  However, the incremental increase in drainage collected in the second test likely results from the increased attention paid to ensuring direct routing of the central tube in the Draintube® filter fabric used to line the bottom of the evaporation cell into the cell’s drain port.  The natural heterogeneity that exists in PA-MFT stemming from the heterogeneity associated with the input raw MFT could also be a factor in the additional mass of water collected as drainage in the second test.       125  Table 4.11: Summary of Dewatering Results from Evaporation Test Cell Results   Figure 4.13: Comparison of Dewatering Trends for Unloaded PA-MFT (Phase I Testing)  126   Figure 4.14: Evaporation Cell Mass Balance for All Phases    127  The rate of increase in percent solids of surface materials was slightly higher during the first test than during Phase I of the second test (Figure 4.15).  The best fit curves for the surface profiles for both tests were second order polynomial functions with an R2 value of 0.88 for the test 1 curve and an R2 value of 0.99 for Phase I of the second test.  Similar trends were observed for the dewatering of materials located at a depth of approximately 10 cm beneath the sample surface, although the rate of dewatering occurred at a slower rate than was recorded for surface materials.  The best fit curve for materials at this depth in the first test had an R2 value of 0.86 while the corresponding value for Phase I of the second test was 0.98.  Figure 4.15: Change in Percent Solids with Time Under Unloaded Conditions (Phase I)  During the incremental loading (Phase IIa) of the PA-MFT sample during EC Test 2, it was observed that both surface materials and those at 10 cm depth dewatered at a similar rate that was markedly slower than the dewatering rates observed during Phase I of the test.  However, it is notable that during the loading phase the solids content of materials at 10 cm depth exceeded the solids content of surface materials located closest to the added masses by approximately 2.5%.  Figure 4.16 shows the corresponding changes in geotechnical moisture content for both tests.  Using this metric it is clear that the surface materials in the first test dried at a faster rate than  128  similar materials in Phase I of the second test.  Also in both tests materials at 10 cm depth initially dewatered at rates that were only marginally less than the rates at which surface materials dried due to evaporation.  This supports the interpretation of the profiles seen in Figure 4.14 which indicate an initial significant rate of underdrainage in overall sample dewatering but its limited effect on materials at depth in the sample over time.  Figure 4.16: Evaporation Cell Test 2 – Change in Geotechnical Moisture Content with Time  The geotechnical moisture content of surface materials reduced at a rate slightly faster than deeper materials.  Consistent with the observed percent solids trends (Figure 4.15), the best fit curves for the geotechnical moisture content for materials at both the surface and at 10 cm depth (Figure 4.15) were second order polynomials.  The best fit curve for surface materials in EC Test 1 had an R2 value of 0.89 while EC Test 2 materials had an R2 value of approximately 1.  The best fit curves for materials at a depth of 10 cm in both tests had R2 values of approximately 0.98. At the start of the incremental loading phase (IIa) for EC Test 2 the geotechnical moisture content of surface materials, which by virtue of sample loading were effectively covered by the added masses, plateaued at approximately 59%.  For materials at a depth of 10 cm the initial dewatering rate of materials in EC Test 1 was marginally faster during the first seven days of the  129  test when compared to comparable materials during the same time period in EC Test 2.  This translated into lower geotechnical moisture content values for these materials after the first 14 days of the test (Figure 4.16).  An inability to obtain representative samples of materials at depth during EC Test 1 after approximately 16 days meant that the longer term drying trends for these materials could not be confirmed.  During EC Test 2 it was observed that the rate at which the geotechnical moisture content decreased slowed appreciably after approximately 21 days.  This timing corresponds to when the effectiveness of sample under drainage appears to plateau.  Consistent with the trends observed for surface material drying, the best fit curves for these deeper materials were described by second order polynomial functions.  The best fit curve for these deeper materials in EC Test 1 had an R2 value of 0.98 while in EC Test 2, a best fit curve fitted to the combined Phase I and Phase IIa profiles had an R2 value of 0.98.  It is noteworthy that the rate at which geotechnical moisture content for materials at a depth of 10 cm in EC Test 2 decreases with time appears to be unimpacted by the addition of the masses in Phase IIa of that test.  Possible reasons for this observation are discussed later in this Chapter. 4.5.2 Surface Deformation Characteristics of a PA-MFT 4.5.2.1 Evaporation Test Cell Sample Surface Deformation Results The profile of unloaded surface deformation for both tests were similar and characterized by a distinct change in settlement rates after approximately 12 days in the EC Test 1 and after 10 days during Phase I of EC Test 2 (Figure 4.17).  During this phase of testing the elevation of the sample surface decreased at a rate of approximately 0.25 cm/day in EC Test 1 and at 0.3 cm/day during EC Test 2.  After approximately 12 days the rate of surface deformation slowed to 0.05 cm/day for EC Test 1 and to a rate of 0.1 cm/day for unloaded surface materials in EC Test 2.  130   Figure 4.17: Change in Sample Thickness for Evaporation Cell Tests (Phase I)  When EC Test 2 materials were incrementally loaded during Phase IIa over a 30 day period, the surface of PA-MFT deformed at a rate of 0.07 cm/day beginning approximately one week after the first 5 kg mass was added.  Deformation of surface materials for the duration of EC Test 2 is shown in Figure 4.18.  The “loaded” rate of deformation continued approximately one week after the last mass increment was added. This delayed deformation response by the sample suggests that PA-MFT requires time to dissipate pore pressures that build up and to release water from the material’s fabric.  It should also be noted that the almost doubling of applied load during the fourth load step in EC Test 2 did not induce either an increased rate of surface deformation (Figure 4.18) or any significant increases in the amount of drainage collected (Figure 4.14).   131   Figure 4.18: Evaporation Cell Test 2 – Change in Sample Thickness with Time (All Test Phases)   132  Limited deformation in response to added load suggests that PA-MFT’s strain and dewatering response are constrained when the amount of normal load applied does not induce strains that trigger a viscous response as indicated in the storage and loss modulus profiles (Figure 4.10).  Alternatively, deformation induced in the sample after the gradual addition of the masses likely represents a combination of the material’s elastic response and the incremental dewatering that results when the load applied is heavy enough to induce some of the water contained within the material fabric to be expelled and to report as underdrainage. 4.5.2.2 CONDES0 Results Surface deformation predictions from CONDES0 were compared to the unloaded phase deformation results for both evaporation cell tests as well as against the surface deformation profile for all phases of EC Test 2.  To run CONDES0 the empirical constants A, B, and Z defined by Liu and Znidarcic (1991) to express 1-D large strain compressibility relationships, and constants C and D defined by Somogyi (1979) to express 1-D permeability-void ratio relationships, were input into the model.  The values of these constants are summarized in Table 4.10 and resulted from analysis of the SICT data for both evaporation cell tests. The evaporation rate used to run CONDES0 was 0.33 m/year.  This value was calculated using the evaporation data collected during the unloaded phase of both evaporation cell tests (Figure 4.13).  Once the time period was set to correspond with the duration of each laboratory test, a time step of 0.01 years (3.65 days) was specified to generate enough data points to support reliable comparison between actual and CONDES0-predicted data.  The initial PA-MFT sample height for each test was entered and a specific gravity of 2.65 was assumed to complete the first round of model analysis. Figure 4.19 compares CONDES0 predictions against actual data collected during Phase I for both tests, while Figure 4.20 provides a similar comparison against surface deformation data  133  collected for all phases of EC Test 2.  The CONDES0 prediction provides good agreement with the initial consolidation trends for both tests (Figure 4.19) up to approximately Day 12 for PA-MFT in EC Test 1 and up to approximately Day 22 for materials in EC Test 2.  After both time periods indicated, CONDES0 overpredicts the amount of 1-D compression in Phase I by up to 8% and 4% in the first and second evaporation cell tests respectively.  It was also observed that CONDES0 only began to overpredict consolidation compared to the EC Test 1 results at Day 12, which corresponds to the same time when the surface deformation rate of this sample started to plateau (Figure 4.17) and the dewatering rate due to underdrainage was observed to decrease dramatically (Figure 4.13). While CONDES0 slightly underpredicts 1-D large strain consolidation during Phase IIa loading and slightly overpredicts sample thickness during Phase IIb of EC Test 2, the model’s predictions and actual measured surface consolidation are similar.  This finding suggests that the empirical constants developed to express the compression and permeability relationships of materials experiencing large strain deformations do capture some of the fundamental mechanisms governing the 1-D large strain consolidation of PA-MFT.  Additional testing of other PA-MFTs would be required to determine how consistently CONDES0 predicts the behaviour of this heterogeneous material, as well as to assist in identification of any fundamental PA-MFT dewatering mechanisms not captured by the model.     134   Figure 4.19: Comparison of Phase I Evaporation Test Results and CONDES0 Predictions   Figure 4.20: Comparison Between CONDES0 Predictions and EC Test 2 Results (All Phases)   135  The fact that the CONDES0 profile showed good agreement with laboratory measurements of surface deformation for EC Test 2 after the sample was loaded with an additional 43.6 kg (a 140% increase in the sample’s total mass at the time that loading was initiated) suggests that the application of compressive force effects the PA-MFT fabric and dewatering characteristics in such a way that is consistent with the consolidation mechanisms modelled using the finite difference implicit method included in CONDES0.  This finding might enhance understanding of why CONDES0 apparently overpredicts consolidation for unloaded samples i.e. the model assumes that continued effective draining translates into further sample consolidation which in the real test is only partially achieved when an external load is applied.  Similarity between the CONDES0 prediction and the surface deformation profile for the loaded EC Test 2, may also be related to the finding during the SICT of PA-MFT that elevated seepage rates were required to induce flow through the sample at each loading stage i.e. the test itself induces flow across the sample at rates that are higher than might necessarily occur in an actual sub-aerially deposited PA-MFT. Results from the SICT and a review of CONDES0 predictions, suggests that PA-MFT fabric impacts the material’s dewatering and consolidation characteristics.  It should also be noted that use of lower values of specific gravity resulted in an even greater overprediction of 1D large strain consolidation by CONDES0. This suggests that assumptions about PA-MFT specific gravity may need to be re-visited if CONDES0 is to be used to predict PA-MFT surface compression under loaded and unloaded conditions.  The possible effect of upward migration of bitumen in deposited PA-MFT on model predictions should also be accounted for.    136  4.6 Geotechnical Strength Characterization of a PA-MFT 4.6.1 Shear Strength and Undrained Strength Characteristics of a PA-MFT As described in Chapter 3 the shear strength at the surface of PA-MFT tested in the evaporation cell was measured using a Torvane®, while the undrained in situ strength just beneath the material’s surface, and at select depths within the test cell, were measured using a field vane.  Undrained shear stress values in excess of the 5 kPa of near surface materials measured using the field vane after 14 days (Figure 4.21) confirms the effectiveness of evaporation to dewater near surface materials, ultimately resulting in formation of a crust once evaporation is not impeded.  The accelerated shear strength gain in PA-MFT tested during the first evaporation cell test, compared to results obtained during EC Test 2, may be attributed to the incrementally higher evaporative drying rates measured (Figure 4.13).  This finding is consistent with the fact that EC Test 1 was completed during summer months when average lab temperatures exceeded those experienced in the early Fall when EC Test 2 was completed. Interestingly, the initial shear strength and undrained strength measurements of approximately 3 kPa for surface and near surface materials exceed any of the peak shear stress values obtained during the rheometer testing of PA-MFT (Figure 4.9).  This difference supports the hypothesis that drying condition effects measured strength values as the samples tested in the rheometer were not exposed to the effects of evaporative drying, optimized drainage geometry, or to the filters used in the evaporation cell.  Evaporative drying effects dewatering of materials at depths of 10 cm below the surface as increases in undrained shear strength occur between approximately 9 days and 32 days for the PA-MFT evaluated during EC Test 1, and between 10 days and 25 days during EC Test 2 (Figure 4.22).  However, in both tests the rate of undrained strength gain appreciably decreases after approximately 15 days which is slightly after the time when the amount of underdrainage starts to plateau.    137  Between 9 days and 15 days after deposition of PA-MFT in the test cell, the undrained shear strength of materials at a depth of 10 cm increased at a rate of approximately 0.4 kPa/day.  This time period corresponds to 17-23 days after PA-MFT flocculation and removal of the bulk of initial release water, prior to material placement in the test cell.  After 15 days in the test cell the rate of strength gain of these materials for both tests averaged 0.2 kPa/day for the 32 day duration of EC Test 1 and up to approximately 25 days after the start of EC Test 2.  This finding suggests that underdrainage has minimal effect on drying of materials at this depth in the tested lift. Between Day 2 and Day 14 of EC Test 1, undrained shear strength of PA-MFT located at a depth of 14 cm increased at a rate of 0.3 kPa/day and then slowed to a rate of 0.12 kPa/day after 14 days.  The undrained shear strength of material at a similar depth in EC Test 2 increased at a rate of approximately 1.5 kPa/day between Day 8 and Day 14 before it slowed to a rate of less than 0.1 kPa/day (Figure 4.23). During the incremental loading phase (Phase IIa) of EC Test 2, materials at a depth of 10 cm showed the most dramatic increase in the rate of undrained strength gain.  The rate of undrained strength gain for these materials increased from approximately 0.07 kPa/day (between Day 16 and Day 30 i.e. the last 14 days of Phase I) to a rate of approximately 0.22 kPa/day for the duration of the loading period (Figure 4.24).  The addition of surface load did not produce a measureable increase in undrained strength of materials located at a depth of 14 cm.   138   Figure 4.21: Change in Surface Shear Strength with Time for EC Test 1 and EC Test 2 (Phase I)   139   Figure 4.22: Change in Undrained Shear Strength at 10 cm with Time (Phase I, Both Tests)     140   Figure 4.23: Change in Undrained Shear Strength at 14 cm with Time (Phase I, Both Tests)   141   Figure 4.24: Change in Undrained Shear Strength with Time for EC Test 2 (All Phases)    142   4.6.2 Relating Geotechnical Properties Dewatering and undrained strength data measured using a field vane were combined with undrained strength data developed using the Swedish fall cone as described in Chapter 3.  PA-MFT samples obtained during dismantlement of EC Test 1 and from the surface of materials tested during EC Test 2 were tested using the Fall Cone.  Surface samples obtained during EC Test 2 resulted from the progressive flattening of the surface of the PA-MFT so that the load plates could have intimate contact with PA-MFT surface materials.  Compression of PA-MFT was greatest in the center of the sample surface (Figure 4.25) and followed the alignment of the central drain tube located in the filter material used to line the bottom of the test cell.    Figure 4.25: Picture Showing Deformation of Surface PA-MFT Materials During EC Test 2  As a consequence of the drying condition of the surface materials, material adjacent to the side walls of the test cell was trimmed to make it level with the sample height at the center of the sample which was used as the reference point for all deformation measurements.  The plot of undrained shear strength versus solids content shown in the left half of Figure 4.26 was developed  143  to present the information typically used in industry to characterize PA-MFT and to infer material performance.  From this plot it can be seen that materials with solids content greater than 70% can produce materials with undrained shear strength exceeding 5 kPa.  As this undrained shear strength was the value specified in the now defunct Directive 74 (AER, 2009), great attention has been historically paid to methods and processes required to achieve this undrained strength value.  However, discussion about the ultimate geotechnical performance and total cost associated with producing these materials is often limited.  Figure 4.26: Relating Undrained Shear Strength; Solids and Geotechnical Moisture Contents  To provide a means of assessing the geotechnical performance that accompanies various solids contents and undrained strength values, the plot to the right of Figure 4.26 that relates undrained shear strength and geotechnical moisture content was developed.  This graph indicates the plastic limit that was determined using the standard drying of PA-MFT threads and the liquid limit of this material using the Swedish fall cone, using the methods described in Chapter 3.  By imposing the plastic and liquid limit values on the undrained strength versus geotechnical moisture  144  content graph, it can be seen that although undrained strength values in excess of 5 kPa are achieved by materials at depths of both 10 cm and 14 cm, the geotechnical moisture content has to be much closer to the plastic limit (approximately 26%) before these materials can exhibit the characteristics of well drained natural soils. This finding has significant implications for management considerations and engineering designs associated with PA-MFT, as a significant decrease in the material’s geotechnical moisture content is needed to produce materials that can safely and economically be rehandled, reclaimed in place, or used as embankment construction material.  The dewatering characteristics of the PA-MFT evaluated in this research suggests that placement of this material in lifts thicker than approximately 15 cm in an environment in which evaporation is not optimized, will not produce a desirable waste management or material handling condition, regardless of the ability of this material to easily produce undrained strength values of almost 9 kPa at depth. 4.7 Fabric Characterization for a PA-MFT 4.7.1 Results of Synchrotron Based Computer Tomography Investigation into the 3-D fabric of the PA-MFT studied in this research was completed using the BMIT-BM beamline at the CLS.  Using the procedures described in Chapter 3, samples of wet (solids content of approximately 40%) and dry PA-MFT (solids content > 80%) located in plastic test vials were installed on the imaging platform and exposed to 25 keV.  Use of such high energy enables the 3-D domain of a 1 cm tall by 1 cm diameter portion of the samples to be imaged at both 4 µm and 8 µm voxel resolutions. Figure 4.27 provides a wide angle view of the wet PA-MFT sample imaged at an 8 µm resolution.  The area of the image that is false colored yellow provides an example of the voids that were observed throughout the wet sample.  When looked at more closely (Figure 4.28) the fabric of wet PA-MFT appears to contain a significant quantity of voids that may be connected by  145  pore throats that are not visible at either the 8 µm or 4 µm scale.  To probe the existence of smaller voids, a sample of wet PA-MFT was sent to the ESRF in France and imaged at a 0.65 µm voxel resolution (Figure 4.28B).  Figure 4.27: Wide Angle View at Interior of Wet PA-MFT Sample with Voids False Subset Colored    Figure 4.28: Wet PA-MFT Showing Voids (False Colored) in Selected Region Evidence of larger voids ranging in width between 10 µm and 250 µm (Figure 4.28A) appear to be disconnected or at best connected by pore throats that may be sub-micron in size.  Imaging investigations at the submicron scale (Figure 4.28B) supports characterization of PA-MFT as a material with abundant sub-micron sized pores that are not connected in an organized network, creating significant opportunity for retention of water within the material fabric.   146  Existence of a material fabric that is highly porous and has low effective permeability is supported by the results of the PA-MFT SICT and the results of the evaporation cell testing that showed limited effects of either evaporation or underdrainage once certain thresholds had been reached, even with the addition of a modest incremental normal load.  The presence of larger voids containing water that is effectively trapped under steady state conditions, especially if either the pore throats or void walls are partially coated by bitumen, provides a basis for conceptualizing how PA-MFT might experience strain deformation.  The ability for bitumen to significantly reduce the effective permeability of materials is also a factor to be considered (Ashrafi et al., 2012).  Once strain does not result in void rupture, PA-MFT will exhibit a solid/elastic-like response.  However, if strain exceeds the yield strain, rupture of larger voids may cause local release of water producing a viscous response and local geotechnical instability in the zone experiencing elevated strain conditions (e.g. the authors sinking into PA-MFT after standing and exerting force in a single location for an extended period). Conversely, the fabric of dry PA-MFT is very dense and experiences significant volume change as water is removed from pores during evaporative drying (Figure 4.29).   Figure 4.29: Dry PA-MFT Showing Voids (False Colored) in Selected Region  147   4.7.1.1 Estimating Void Space in Wet PA-MFT Using Image J and Avizo® the composed 3-D images of wet PA-MFT were analyzed to quantify the amount of void space detectable at both 4 µm and 8 µm voxel resolutions.  The pore space at both resolutions averaged approximately 12% (based on analysis of six sub regions of the 3-D domain for each data set).  However, the calculated average geotechnical porosity associated with an average initial void ratio of 3.3 for wet PA-MFT (determined from the SICT), which is at the same percent solids as the sample used to investigate PA-MFT fabric, suggests that the vast majority of voids in wet PA-MFT are smaller than the voxel resolution achieved on the BMIT-BM beam line, and are therefore not detectable using synchrotron-based CT at 4 µm or greater.  Porosity estimates of the sample imaged at 0.65 µm voxel resolution yielded an average porosity of 50% which supports the hypothesis that the abundance of PA-MFT voids are smaller than 1 µm.  This finding is also consistent with porosimetry investigations into flocculated kaolinite completed by Diamond (1971).   To further investigate the presence of submicron sized voids and pore throats, FESEM and cryo-SEM images of PA-MFT at similar solids content to those imaged using synchrotron-based CT were obtained.  Results from these investigations are described in the following Section. 4.7.2 FESEM and cryo-SEM of Raw MFT and PA-MFT Fabric FESEM was completed using samples of oven dried MFT and PA-MFT flocculated on both August 18, 2014 and November 11, 2014.  Moist and wet samples of PA-MFT flocculated on the indicated dates were also imaged using cryo-SEM.  Table 4.12 summarizes the testing protocol for FESEM and cryo-SEM imaging completed and provides descriptions of the most visible observations made on a selection of the images that are presented on numbered Plates provided in Appendix C.  148  FESEM and cryo-SEM images with a minimum visible reference scale of 2 µm show clay mineral surfaces, especially plate like faces, and booklets of kaolinite minerals.  Voids are irregular in shape and appear to exist along face-face, edge-face, and edge-edge contacts, or where different minerals appear to overlap with widths ranging between 0.02 µm and 3 µm (Figure 4.30).  In the images of PA-MFT the random orientations of clay minerals seems to reflect the effects of mixing, shearing and apparent breakage of some minerals that likely occurred during flocculation.  Figure 4.30: Cryo-SEM Image of Sublimated and Gold Sputter-Coated PA-MFT Comparison of the B images on Plate 2A and Plate 2B (Figure 4.31) seems to indicate that the fabric of PA-MFT becomes more quiescent (less chaotic/more ordered) the longer the flocculated solids remain in undrained and even saturated conditions.  These effects are most evident in the apparent angularity of clay mineral edges and shapes visible in the image on the left of Figure 4.31 which was taken only 13 days after the PA-MFT was created.  Figure 4.32 provides an example of the threadlike (left picture) and globular features (right picture) visible in both FESEM and cryo-SEM images, which may indicate the presence of polymer on the surfaces of mineral particles.    149  Table 4.12: Summary of Test Protocol and Findings from FESEM and Cryo-SEM of AITF-MFT and PA-MFT      150  Table 4.12 (continued)  151   Figure 4.31: Potential Aging Effects of PA-MFT Kept Under Undrained Conditions – Fresh (A), Aged (B)   Figure 4.32: Threadlike (A) and Globular (B) Features Visible on PA-MFT Mineral Surfaces If the degree of electron noise visible in the images of uncoated samples was to a level that obscured the clarity and definition of visible features, samples were sputter coated to improve resolution as noted in Table 4.12.  The sample of wet PA-MFT flocculated on November 11, 2014 appeared to release some of the hydrocarbons associated with the residual bitumen that was contained in the sample during the ethanol and HDMS series used to dehydrate wet PA-MFT samples prior to FESEM.  This is an interesting finding as similar leaching of hydrocarbons was not exhibited by the sample of PA-MFT flocculated in August 2014 (Figure 4.33).  152   Figure 4.33: Differently Aged PA-MFT Samples in HMDS Prior to FESEM Imaging (fresher sample on right shows evidence of bitumen leaching)   4.7.3 Effects of a PA-MFT Fabric on Dewaterability Results of the FESEM and cryo-SEM investigation provide support for the observation that some proportion of the voids in PA-MFT with a solids content in the order of approximately 40%, range between 0.2 µm to 2 µm.  It is also likely that an even greater percentage of the voids are comprised of voids on the Angstrom scale as Diamond (1971) identified for naturally occurring kaolinite rich sediments.  Results from both the QEMSCAN and MBI determinations confirm that kaolinite minerals are the most abundant and as such may have a fabric containing pore sizes similar to those identified by Diamond (1971) for flocculated and unflocculated kaolinite. The current CT capabilities at the CLS underestimate the total porosity of wet PA-MFT as the available imaging resolution and is unable to produce imaging of what are believed to be the majority of the submicron and nano sized voids present in these materials.  However, imaging completed at ESRF supports the hypothesis that PA-MFT consists of a bi-modal pore network in which a large quantity of the voids and pore throats are less than 1 µm in effective diameter. Experienced technicians working with the FESEM and cryo-SEM equipment used in this research also confirmed that the globular features observed in the images were not artifacts of  153  dehydration, sublimation, or sputter coating processes, and plausibly indicate the presence of polymer on mineral surfaces. The irregularity and wide range of void shapes, sizes and orientations compared to the chaotic orientation of clay mineral surfaces provides evidence for the types of tortuous flow paths that could significantly limit the ease with which water contained in nano, micro and millimeter sized voids can be removed from PA-MFT fabric.  Findings from the combined investigations into the fabric of wet PA-MFT appear to support the idea that PA-MFT is highly porous and has a low effective permeability due to the presence of a significant number of micro pores that, even if they are connected, provide significant impedance to significant additional release of water trapped in these voids.  Furthermore, if functional groups associated with residual bitumen that have been found to adhere to the surfaces of some fine tailing solids behave hydrophobically, as suggested by FT-IR and Raman spectra, this could further hinder the release of water trapped in a network that is comprised of micron to sub millimeter sized pores connected by micro and nano sized pores and pore throats.     154  Chapter 5: Discussion 5.1 Overview of Research Findings The results of this research suggest that PA-MFT is a synthetic geomaterial, the bulk properties of which are primarily influenced by the following: • Surface charge interactions, especially during initial mixing of input materials; • The nature of its void space; and • Bitumen coating on mineral surfaces dispersed throughout the material.  Results indicate how properties at the micro scale impact rheological and geotechnical engineering behavior of the bulk material (Figure 5.1).  The figure indicates that residual bitumen and PA-MFT fabric likely impact development of effective reclamation prescriptions and long-term storage utilization and PA-MFT behavior if it is placed in thicker lifts meant for inclusion in a dry closure landscape.  In brief, while PA-MFT may exhibit an initial release of water after incipient flocculation, the material fabric exhibits low permeability even though it is comprised of abundant voids.  When PA-MFT is deposited in lifts exceeding 20 cm (15 cm being the maximum desirable thickness for effective evaporative drying of the deposited lift) water retention in the material’s fabric results in elevated moisture content over the long-term which adversely impacts strength development due to the very limited effects of diffusion and underdrainage associated with the tortuous flow path in the material. PA-MFT requires characterization beyond knowledge of its solids and fines content to understand the factors at play in its dewatering, consolidation, and long-term geotechnical stability especially since all of these factors have direct implications for management and risk mitigation during operations, closure and post closure of oil sands surface mine operations.  155   Figure 5.1: Summary of Connection Between Multi-Scale Processes  Use of a comprehensive testing framework facilitated collection of a wide range of data that enabled PA-MFT to be examined through the independent yet related lenses of surface chemistry, materials science, rheology and geotechnical engineering.  The tests most critical to understanding PA-MFT dewatering were identified.  Analysis of PA-MFT fabric provides insight into characteristics that could be altered or which need to be achieved by technologies being evaluated to produce a geomaterial with more desirable dewatering and consolidation characteristics – i.e. not inadvertently creating a hydrophilic material that also can physically store water within its fabric.  The ability to draw direct comparisons between properties of reference clay minerals found outside of the Alberta oil sands (e.g. Georgia kaolinite) and those found in the Alberta Oil sands is also an area of active research and challenges the ability to make simplistic assumptions about material behavior based solely on clay mineral content.    156  Zeta potential testing indicated charge similarity between the net negative surface charge of abundant kaolinite clay minerals in tested MFT (confirmed by XRD and chemical analysis) and the bulk negative surface charge of the anionic polymer.  This charge similarity results in a degree of repulsion in the system that limits formation of strong flocs which are desirable in a material that is to be easily densified as it releases water from its matrix.  Controlled mixing of raw MFT and the tested anionic polyacrylamide polymer produces fluffy flocs that create a fabric with abundant voids separated by haphazardly arranged clay mineral particles.  PA-MFT occupies a larger volume than unflocculated MFT and traps water within its fabric (Beier et al. 2013).  These fluffy flocs are also colloidally stable, exhibit low shear strength, and have limited consolidation capacity. That kaolinite’s siloxane basal plane is hydrophobic while its hydroxyl basal plane is hydrophilic contributes to the complex surface interactions that take place in both MFT and PA-MFT.  The hydrophobic siloxane surface provides a surface to which bitumen may attach while the hydrophilic surface provides an attachment surface for both water and hydrophilic anionic polyacrylamide.  The nature of kaolinite’s two basal planes may also contribute to the colloidal stability exhibited by MFT and PA-MFT and to the chaotic arrangement of clay minerals observed in completed SEM imaging.  Hydrophilic illite (constituting between 2 and 10% by weight of PA-MFT based on XRD results) provides mechanisms for possible absorption of both water and hydrophilic polyacrylamide polymer.  Mixed layer kaolinite-smectite which commonly occurs in oil sands ore also provides a mechanism for entrapment of water in the mineral interlayer. Support for the theory of void formation resulting from the surface charge and hydrophilic-hydrophobic interactions described earlier is provided by the high resolution imagery obtained from both synchrotron-based micro-tomography and SEM (field emission and cryo stages).  These  157  images indicate abundant micropores that occur along clay mineral surface interfaces that create tortuous flow paths if and where pore connectivity exists.  At a 0.65 micron scale consistent pore connectivity, which would be a desirable feature of an effectively consolidating material, was not observed.  This supports a conclusion that PA-MFT has a pore network with limited connectivity and has permeability that is equal to or lower than raw MFT. Imaging showed that the majority of the voids in PA-MFT have an apparent effective diameter that ranges between 20 nm and 0.2 µm.  These voids are randomly arranged between the platy kaolinite mineral particles that also exhibit an apparently random orientation.  These factors combine to produce a highly tortuous flow path that limits water egress.  Both water and bitumen may become entrapped in some if not many of these abundant, small voids depending on the kaolinite basal plan exposed on the clay minerals.  This further limits effectiveness of dewatering by diffusion and likely indicates that large amounts of energy would be required to remove fluid entrapped in these voids.   Differences between MFT and PA-MFT fabric are also evident in the increased pore pressures required to induce flow across PA-MFT when compared to those used in SIC testing of raw MFT.  PA-MFT SICT data must be carefully analyzed and interpreted as increased seepage force used to develop a permeability reading across the sample may not necessarily represent field conditions and could possibly create new flow paths within the material providing a higher value of permeability than would exist in the field.  Consequently, use of SICT data could lead to under prediction of time for PA-MFT to consolidate which would directly impact reclamation planning, cost and risk.  Grounds for this concern are bolstered by the finding that CONDES0, which uses SICT output, overpredicts consolidation of deposited PA-MFT after underdrainage of PA-MFT tested in the evaporation test cell plateaus.    158  In addition to the mineralogical basis for adhesion of bitumen to the hydrophobic siloxane basal plane of abundant kaolinite minerals, bitumen coating on minerals was readily observed by the unaided eye and at 50 times magnification using the optics of the Raman spectrometer.  While bitumen may typically account for 4% of MFT by mass, this hydrophobic organic material can account for upwards of 20% of the material on a volume basis.  Bitumen also appears to act as a coating on the surfaces of mineral solids and consequently contributes to the hydrophobicity and plasticity exhibited by the bulk material.  Pore throats that exist between adjacent hydrophobic siloxane planes could be blocked by bitumen adhered to these surfaces and lead to water retention within PA-MFT.  Hydrophobic forces would also have to be overcome to induce water flow through pore throats containing bitumen.   To the unaided eye bitumen appears in the form of droplets that appear to be well-dispersed throughout PA-MFT.  However, as a component of the bulk material a portion of bitumen appears to be mobile and tends to migrate upward in subaerially deposited lifts to form a distinct layer.  Once this layer forms it acts like a sealant limiting the egress of water from the bulk PA-MFT fabric as well as limiting the depth of evaporation effectiveness.  Bitumen surface coatings also result in the increased plasticity exhibited by PA-MFT.  This likely contributes to the large plasticity index (43%) of this material and the fact that PA-MFT remains within its plastic range even at solids contents upwards of 75%.  The compressibility index of PA-MFT is consistent with values obtained for high plasticity clays, even though kaolinite is the abundant mineral.  Residual bitumen made use of a Cassagrande cup to determine liquid limit impossible due to its continued adhesion to the cup surface after more than 100 blows. While atmospheric drying appears to work well for PA-MFT deposited at thicknesses less than 20 cm, operators need a mechanism to keep up with current and projected future increased  159  volumes of MFT.  However, as a result of several compounding factors, PA-MFT as it currently exists is not a suitable candidate for stacking in thick lifts to be included in closure landforms that do not include some form of geosynthetic material (i.e. geotextile or geogrid) to provide the geotechnical stability required to create a surface layer trafficable by reclamation equipment.  Apart from the 8 to 10 cm crust that forms, water remains trapped within a porous fabric that exhibits extremely low permeability resulting in low material strength.  The creation of fluffy flocs results in ineffective utilization of storage areas like pits as the material shows limited ability to effectively consolidate.  The presence of residual bitumen appears to transform kaolinite from what geotechnical engineers typically consider to be a benign clay mineral to one that exhibits significant plasticity. This adversely impacts the bulk geotechnical engineering properties of PA-MFT and severely limits the thickness of lifts that can be entirely dried through the combination of evaporation and underdrainage.  5.2 Summary of Key Research Findings While the overall engineering behavior of PA-MFT can be observed and measured at field scale, explaining the reasons for the characteristics exhibited requires an understanding of the fundamental material properties involved.  Flocculation of MFT with anionic polyacrylamide results in limited aggregation of previously dispersed clay minerals.  The selective attachment that occurs between the anionic polymer and the previously dispersed clay minerals results in the creation of a flocculated material with an open structure.  After 24 hrs, PA-MFT has an average solids content that is between 30% and 80% higher than the solids content of raw MFT.  Zeta potential measurements reveal that the addition of this polymer only slightly decreases the bulk electronegative surface charge of clay minerals.  The limited nature of the agglomeration that results from the addition of this polymer also supports the open and fluffy nature of the flocs observed.  These observations support the finding of Nabzar and Pefferkorn (1985) that suggests  160  that the open structure observed results from the selective attachment between the positive edge charges on the clay minerals to the negatively charged surfaces of the anionic polymer.  Electronegativity of the bitumen suggests that it could also compete with the polymer for limited positively charged attachment sites on clay mineral edges.  Visual observation with both the aided and unaided eye reveals that bitumen is dispersed throughout the PA-MFT and, further, that it appears to coat the clay mineral surfaces.  The finding of an apparent affinity between bitumen and clay mineral surfaces is also indicated by the spectra obtained using FT-IR spectroscopy.  It is suggested that the bitumen surface-coating of clay minerals dispersed throughout PA-MFT impacts its plasticity.  The distribution of residual bitumen throughout PA-MFT also suggests that blockage of pore throats with bitumen is plausible and could present an added hindrance to PA-MFT dewaterability.  SICT results further suggest that the permeability and compressibility profiles of PA-MFT follow trends similar to those developed for unflocculated MFT.  Use of synchrotron based computer tomography enabled 3-dimensional imaging of 1 cm3 samples of PA-MFT.  This represents a significant advancement in the non-destructive imaging of these materials and the evaluation of PA-MFT fabric.  While abundant void space, ranging between 10 and 250 µm, could be identified at 8 µm voxel resolution, a consistent, effective connection between these voids was not visible.  SEM imaging also appeared to indicate abundant micropores ranging between 20 nm and 0.2 µm.  Clay-sized minerals in these images appeared chaotically arranged suggesting that flow pathways, where they existed, would likely be highly tortuous and not continuous.  The combined findings from these two imaging methods suggest that while PA-MFT fabric may be characterized by high porosity, its effective permeability is likely very low.  161  Rheological testing of PA-MFT suggests that floc strength is a key factor in determining the peak shear strength of this material.  Furthermore, breakage of floc structure resulting from applied energy (either during flocculation or subsequent material handling) appears to be irreversible.  PA-MFT exhibits an elastic response up to an applied yield strain of 200% at which point the material exhibits a viscous fluid-like response to applied oscillatory stress. These combined findings suggest that the overall behavior of PA-MFT is governed by fundamental characteristics associated with charges and coatings that exist on clay-sized mineral surfaces. The nature of the fabric that results when raw MFT is flocculated with anionic polyacrylamide polymers is similar to those used in this research.  These findings are significant as they not only provide a basis for understanding some of the mechanisms that impede effective dewatering of PA-MFT, but they also provide insight into the factors impacting the plasticity, strength development, and consolidation of this material.  Findings from this research are of particular importance when considering some of the factors that could impact performance and storage utilization if materials with similar characteristics were placed in thick lift deposits.    162  Chapter 6: Conclusions The overarching objective of this research was to provide an understanding of how fundamental properties of PA-MFT, namely surface charge and fabric, govern its micro and meso scale behavior as evidenced through the material’s rheological and geotechnical engineering characteristics.  To this end, this research has identified baseline characteristics and essential tests from the fields of surface chemistry, rheology and geotechnical engineering (Figure 6.1) that should be completed to better understand the interrelated physical and chemical characteristics impacting the dewatering characteristics of a geo-material resulting from chemical amendment of raw MFT.  Figure 6.1: Testing to Evaluate Dewatering Characteristics of Chemically Amended MFT as Recommended   Baseline characterization enables clear quantification of key input material properties and provides valuable information that can be used to perform high level screening for likely responses  163  to chemical amendments.  Such a framework also facilitates comparative evaluation of a range of possible MFT amendments.  Mineralogical and surface characterization expand the scope of baseline characterization as they enable identification of dominant clay minerals present as well as provide indication about the colloidal stability and the surface reactivity of the amended material.  The rheological tests recommended facilitate early assessment of initial strength of the amended material as well as provide an indication as to how the material responds when it is handled or repetitive oscillatory/rotational stress is applied.  These results provide some indication as to the amended material’s response when exposed to different loading mechanisms which is important when evaluating material performance in the near and long term.  The geotechnical testing recommended enables rudimentary geotechnical engineering behaviour to be characterized and indicates how material strength changes as a function of moisture content.  Finally, use of non-destructive imaging methods allows the material’s fabric to be evaluated to see if any physical mechanisms exist that would impede or support sustained material dewatering. In addition to achieving the overarching general objective, the research also aimed to satisfy seven specific objectives first presented in Chapter 1.  These objectives and their specific conclusions are summarized in Table 6.1.  Solids content, fines content, and clay water ratio on their own do not provide a complete basis for understanding the fundamental mechanisms controlling the dewatering and strength characteristics of PA-MFT.  This research demonstrates that establishing and understanding critical linkages between the fundamental properties governing dewatering enables the observed behaviour and performance of PA-MFT to be better conceptualized and understood.      164  Table 6.1: Summarized Research Conclusions ID No. Research Objective Conclusion 1 Investigate how addition of BASF 5250 (anionic polyacrylamide polymer) changes the colloidal stability of raw MFT. Insignificant change in electronegativity of MFT and only incipient flocculation results. 2 Identify functional groups on PA-MFT surfaces. Four distinct functional group signals associated with bitumen and two signals associated with kaolinite (confirmed by XRD) were identified. 3 Produce 3-D imaging of PA-MFT fabric. High resolution 3-D imaging of PA-MFT developed using non-destructive synchrotron-based CT.  This is a significant advancement over current state of SEM imaging as utilized in Alberta Oil Sands tailings research. 4 Investigate effects of handling and storage condition on shear strength. - PA-MFT exhibits rheomalaxis (i.e. irreversible degradation in shear strength) with handling.  Impacts shear strength development. - Storage under saturated conditions limits and may degrade strength development. 5 Quantify evaporative & underdrainage losses & measure surface deformation during drying of thicker lift PA-MFT. - Achieved using evaporation test cell. - Identified limits to underdrainage and evaporative drying for thick lift deposits. 6 Quantify shear strength development in thicker lift deposits with and without load. Gradually plateaus to ~0.2 kPa/day. 7 Develop qualitative understanding of interplay between PA-MFT chemical & physical characteristics by using a consistent characterization framework. - Hydrophobic bitumen surface covering combined with abundant micropores in a poorly connected pore network limit the achievable dewaterability of PA-MFT. - Evaluation & comparison of dewatering characteristics of proposed MFT amendments is facilitated by using a consistent characteristic framework.  Using data collected from complementary fields, PA-MFT evaluated in this research may be classified as a synthetic geo-material that is colloidally stable with dewatering characteristics that are adversely impacted by the porous but impermeable fabric that results from ineffective flocculation and results in entrapment of water within the material fabric.  Residual bitumen is widely dispersed through the resulting material and adheres to enough hydrophobic mineral surfaces in a way that can further restrict water flow in the material, limit material strength development, and cause PA-MFT to behave like high plasticity clay.  This represents a fundamental shift in how geotechnical engineers working in the Alberta oil sands should view oil sands kaolinite.  The finding also supports a conclusion that geotechnical testing on Dean Stark  165  solids, from which the bitumen has been removed, will not provide information that represents the true nature of the material encountered in the field.  Furthermore, the fluffy flocs that result from the particular chemical amendment of raw MFT evaluated are not optimally packed and due to their colloidal stability do not allow agglomeration into a more consolidated structure.  This results in this material exhibiting ineffective consolidation and poor ultimate strength development when it is deposited in lifts exceeding 20 cm.  Consequently, if this material is considered for inclusion in the closure landscape, additional engineering will be required to develop dry land forms that are trafficable by construction equipment and which will be geotechnically stable, have a reduced risk profile, and can be cost effectively managed and reclaimed.      166  Chapter 7: Limitations and Contributions to Knowledge 7.1 Limitations This research focused on evaluation of the characteristics of a product resulting from the controlled mixing of a composite source of MFT obtained from the MFT bank of Alberta Innovates and a single anionic polyacrylamide similar to that used at commercial oil sands surface mine leases.  As such, while these research findings may provide a lens with which to consider other the properties of other combinations of MFT and anionic polyacrylamide polymer, the research findings are limited to the specific material tested.  The PA-MFT evaluated was also produced using a lab scale mixing method which differs from the in-line flocculation methods used in field scale operations.  As such, further work would be required to evaluate how characteristics of lab flocculated PA-MFT compare to the product that results from field scale flocculation methods.  This comment extends to all testing as it was completed at laboratory scale and comparisons to larger scale testing would need to be based on actual data that would need to be generated. Because this research focused on how selected data obtained from surface chemistry, rheology, imaging, and geotechnical engineering may be combined to develop a more comprehensive understanding of the fundamental characteristics governing PA-MFT behavior, it does not attempt to be exhaustive.  Instead, it demonstrates how characterization methods from various fields may be used as effective tools to help shed a broader light on the material properties that drive bulk material behaviour and provide insight into material characteristics that either aid or impede intended performance objectives. Finally, while synchrotron based computer tomography facilitates non-destructive three dimensional imaging of PA-MFT fabric, characterization of the micro pore space observed is limited to 0.65 microns which is still much larger than the majority of pore sizes identified using SEM.  This means that better characterization of the total (3-D) volume of voids in PA-MFT can  167  be improved as the image resolution using this method or other nano-CT methods improves.  Quantification of pore space in wet PA-MFT was directly impacted by insight from the author’s work with and observations of wet PA-MFT in the laboratory.  However, as all researchers in this area may not have the opportunity to flocculate their own PA-MFT, development of guidelines for identification and quantification of pore space that are independent of the researcher would be helpful to see how data compares among various chemical amendments of MFT.  The inability of carbon to be seen using synchrotron-based CT also inhibits use of this method to develop imaging of the blockage of PA-MFT fabric voids postulated.  Consequently investigation of other methods of non-destructively imaging residual carbon at the scale of micro fines is needed. 7.2 Contributions to Knowledge Identification of key factors impacting dewatering enabled a discrete suite of essential test methods to be identified (Figure 6.1).  This testing schema can be applied to the screening and evaluation of a wide range of MFT amendment strategies being considered well before costly advancement to pilot scale trials and provides a consistent basis for comparison of the ability of various MFT amendments to achieve desired performance criteria e.g. steady dewatering of water contained in the material fabric.  Use of a consistent test framework also could translate into potentially significant cost savings associated with the ability to tighten the evaluation feedback loop between material amendment and desired geotechnical engineering behavior.  Evaluation of the results from the recommended tests in an integrated fashion harnesses key insights from characterization tools not consistently considered collectively in oil sands industry.  In addition, the framework enables short term and long term behaviour, cost and risk associated with using a particular type of chemical amendment to be consistently evaluated. Assessment of the interplay of fundamental PA-MFT properties like electrokinetic potential, surface chemistry pre and post flocculation, relating undrained strength development of  168  this synthetic material’s drying characteristics, and fabric are primary contributions of this research.  Consequently, broad integrated knowledge of fundamental mechanisms controlling dewatering enables key performance indicators (KPIs) like maximum initial water release and target rates of long term dewatering and consolidation to be developed for various components of a treatment process.  By using a comprehensive approach to characterization, individual KPIs (e.g. net water release at time of flocculation) can be confirmed while ensuring that they do not also confound a site’s overall ability to cost effectively produce, handle, and store chemically and geotechnically stable fine tailings that do not pose a threat to the health of ecosystems and safety of human or animal life. The integrated findings from this research suggest that current PA-MFT created at operating sites significantly limits the industry’s ability to keep pace with existing and projected increases in fine tailings production without need for significant additional capital investment in equipment and infrastructure to enhance PA-MFT dewatering characteristics.  Findings from this research suggest that significant energy inputs would be required to remove additional water from current PA-MFT fabric.  While it has been well demonstrated in industry that evaporative drying of this material is effective when it is deposited in lifts under 20 cm, results from this research suggest that depositing PA-MFT type materials in the thicker lifts - required to keep pace with existing and projected future increases in fine tailings inventories - would require a change in how these materials are currently managed or contemplated for inclusion in the closure landscape.  The retention of water in the micro and nano voids of PA-MFT suggests that management practices that involve costly handling and transport of these materials from dedicated drying areas to secondary staging locations will continue unless the management practice is adapted to account for the nature of the material being produced.  Alternatively, creation of permanently designated  169  large mud stacking areas designed to permanently store the volumes of PA-MFT produced should be considered.  While such facilities could sterilize some portion of oil sands ore located beneath their footprint and would require modification of current energy development lease requirements.  The proposed facilities would be configured such that PA-MFT could be deposited in thicker lifts requiring limited amounts of overburden to construct a shell that would form the facility’s outer perimeter.  Reclamation of this facility would likely involve installation of a passive means of aiding release of water contained in the micro and nano-pores like wick drains, while reclamation of surface materials would require use of synthetic geomaterials, similar to the geogrid used in the reclamation of Suncor’s Pond 5.  Furthermore the water reclaim system for this type of facility would be designed to enable return of water for use in additional processing or routing through a treatment system that produces water of a quality suitable for use in either energy generation or discharge to the environment.  Ultimately, creation of PA-MFT shifts the challenges associated with MFT management from a liquid phase to a solid phase problem with direct implications on environmental management and cost.  Existence of thick lifts of material with high porosity and low permeability creates a hazard to the health and safety of human operators tasked with accessing the surface of these materials as well as mammalian wildlife that would likely become trapped in these materials if they ventured on to them.     170  Chapter 8: Recommendations for Future Work Continued research into the near-term and long-term performance of materials resulting from combining chemical additives with MFT is warranted.  At a minimum, findings from this research provide a benchmark by which the geotechnical performance and long-term dewatering characteristics of resulting materials may be evaluated.  By looking at the fundamental properties impacting and governing a material’s dewatering characteristics, insight can be gained about which factors and characteristics need to be modified to produce a geotechnically better material that can be cost effectively produced, dewatered, and included in the closure landscape. Future work should be generally focused in the following areas: 1) Additional characterization of the surface chemistry of PA-MFT; 2) 3-D imaging of how macro, micro, nano and angstrom sized pores of PA-MFT relate; and 3) Seeing how CONDES0 predictions compare to actual measurements taken on field scale thick lift deposits of PA-MFT.  Specifically, completion of the following research should be considered to further advance understanding of materials resulting from MFT amendment.  In all cases it is recommended that research activities account for actual oil sands minerals and conditions in which materials exist in the field. 1. Quantification of the Hydrophobic and Absorption Effects of PA-MFT Functional Groups This work would involve extensive investigation into and quantification of how residual bitumen content covers and influences the properties of PA-MFT mineral surfaces.  To quantify the impact of bitumen concentration on surface functional groups, FT-IR analysis of standard samples of known bitumen content would be performed in a KBr cell.  X-ray photoelectron spectroscopy (XPS) may be used to quantify surface area coverage by various functional groups and to evaluate results of imaging analysis reported in this thesis.  The surface area covered by the various functional groups identified would also be quantified.  The  171  effect which bitumen content has on permeability should be quantified and related to surface area coverage by hydrocarbon functional groups.  The amount of activity exhibited by each functional group should also be quantified for PA-MFT under saturated and dry conditions.  The role and activity of functional groups associated with kaolinite should also be investigated.  Bonding mechanisms between bitumen and clay mineral surfaces and bitumen and polymer should also be characterized.  2.   3-D Imaging of the PA-MFT Pore Network and Porosimetry Additional nano scale synchrotron-based computer tomography of wet PA-MFT should be completed in an effort to better characterize the sub-micron pore network of this material.  Given the range of sizes of pores visible in images from FESEM and cryo-SEM, imaging would ideally be completed at a 0.1 µm voxel resolution.  A hydrodynamic model of the resulting pore network could also be developed and analyzed to quantify the effects of bitumen content on permeability and to see how permeability predictions compare to values obtained from the SICT.  Porosimetry of PA-MFT should also be performed in a manner that accounts for how much these materials are affected by their stress environment.  This work would likely require use of techniques outside of those typically used in geotechnical engineering.  Quantification of PA-MFT void space using both CT and porosimetry could also be compared to SICT void ratio determinations. 3. Evaluation of Applicability of CONDES0 Predictions to Field Scale Thick Lift PA-MFT Deposits In this study, consolidation data collected after a thick (40 cm) and thicker (3 m) lift of PA-MFT has been deposited in a large scale drying cell, would be compared to CONDES0  172  predictions.  Based on the results achieved, the effects of self-weight consolidation on CONDES0 predictions could be investigated further. In addition to the primary areas of additional study, further research into the following areas would enhance the preliminary findings described in this research:  Enhanced Agglomeration of Clay Minerals Addition of both anionic and cationic polymers result in limited agglomeration of oil sands clay minerals.  The formulation and investigation of chemical amendments that can effectively agglomerate clay minerals whose surfaces are likely intrinsically impacted by their hydrocarbon rich environment, and whose surfaces are partially coated by hydrocarbon functional groups, should be continued.  Even an incremental improvement of clay mineral agglomeration at basic pH would enhance management of these materials and optimize storage utilization in facilities where these materials are deposited.  Attachment mechanisms between chemical amendments and bitumen affected clay mineral surfaces should also be further investigated to see if more compact attachment configurations can be achieved.  In addition, the intrinsic hydrophobicity, hydrophilicity and surface chemistry of oil sands kaolinite should be quantified. Enhanced Rheological Characterization Given the effects of sample disturbance and floc breakage on the measurement of the rheological properties of PA-MFT, a means of quantifying handling and disturbance effects on shear strength should be developed.  A mechanism should also be devised to enable PA-MFT to be immediately poured into a test container that minimizes sample disturbance and allows initial release water to flow away from the PA-MFT solids without drying out the sample.  Additional characterization of the visco-elastic properties of PA-MFT should also be completed at different  173  frequencies and durations.  The effect of solids content on visco-elastic response should also be investigated and build on the work in this thesis and on work by Mizani and Simms (2014). Geochemical Characterization of PA-MFT Seepage Water Quality Results from standard geochemical methods should be used to identify and quantify contaminants of potential or significant concern in PA-MFT seepage water quality.  These findings should be compared to the XRF and XANES data collected in this research to see how these methods may be used to evaluate near-term and post-closure seepage water quality associated with these materials.     174  Bibliography AANDC [Aboriginal Affairs and Northern Development Canada], (2013). Treaty texts – Treaty No. 8. Retrieved from http://www.aadnc-aandc.gc.ca/eng/ on September 25, 2013.  AB Energy [Alberta Energy], (2013a). Facts and statistics. Retrieved from http://www.energy.alberta.ca/OilSands/791.asp on September 23, 2013.  AB Energy, (2013b). Oil Sands Acts and Regulations. Retrieved from http://www.energy.alberta.ca/OilSands/810.asp on September 26, 2013.  AB Env [Alberta Environment], (2013). Oil Sands Information Portal. Retrieved from http://www.oilsands.alberta.ca/index.html on September 24, 2013.  Abu-Hejleh, A. & Znidarcic, D. (1992). User Manual for Computer Program SICTA. Bartaw: Florida Institute of Phosphate Research (FIPR).  ACG (Australian Centre for Geomechanics), (2006). Paste and Thickened Tailings – A Guide. R. Jewell & A. Fourie eds. Nedlands: University of Western Australia. Adachi, Y. (1995). Dynamic aspects of coagulation and flocculation. Advances in Colloid and Interface Science, 56:1-31. AER [Alberta Energy Regulator], (2009). Directive 074: Tailings Performance Criteria and Requirements for Oil Sands Mining Schemes. February 3.  Alagha, L., Wang, S., Yan, L., Xu, Z., & Masliyah, J. (2013). Probing adsorption of polyacrylamide-based polymers on anisotropic basal planes of kaolinite using quartz crystal microbalance. Langmuir, 9 (12), 3989-3998.  Alamgir, A. (2011). Al-PAM Assisted Filtration of MFT (Master’s thesis, Department of Chemical and Materials Engineering), Edmonton: University of Alberta.  Alamgir, A., Masliyah, J., & Xu, Z. (2010). Al-PAM Assisted Filtration of Mature Fine Tailings Produced from Oil Sands Development. Proceedings, Second International Oil Sands Tailings Conference, Edmonton, December 5-8, 97-105.  Arinaitwe, E. (2013). The Dispersing Action of Polysaccharides in Oil Sands Slurries (PhD dissertation, Department of Mining Engineering), Vancouver: University of British Columbia.  Ashrafi, M., Souraki, Y., & Torsaeter, O. (2012). Effect of temperature on Athabasca type heavy oil – Water Relative permeability curves in glass bead pack. Energy and Environment Research, 2(2).   175  ASTM [American Society for Testing and Materials], (2015). Annual Book of ASTM Standards. Retrieved from http://www.astm.org/BOOKSTORE/BOS/index.html on May 20, 2015. Barnes, H. (2000). A Handbook of Elementary Rheology. Aberystwyth: University of Wales Institute of Non-Newtonian Fluid Mechanics.  BASF (2015). Spec sheet for anionic poly-acrylamide 5250, Retrieved from www.basf.com/miningsolutions on May 20, 2015.  Beier, N., Wilson, W., Dunmola A., & Sego, D. (2013). Impact of flocculation-based dewatering on the shear strength of oil sands fine tailings. Canadian Geotechnical Journal 50: 1-7.  Bennett, R., Bryant, W., & Hulbert, M. (1991). Frontiers in sedimentary geology. Microstructure of fine-grained sediments from mud to shale. New York: Springer-Verlag.  BGC [BGC Engineering Inc.], (2010). Oil Sands Tailings Technology Review Report No. TR-1, Edmonton: Oil Sands Research and Information Network (OSRIN), University of Alberta, School of Energy and the Environment, 136pp. Blanco, A., Negro, C., Fuente, E., & Tijero, J. (2005). Effect of shearing forces and flocculant overdose on filler flocculation mechanisms and floc properties. Industrial Engineering and Chemical Resources, 44 (24): 9105-9112. Boxill, L. & Hooshiar, A. (2012). A review of design and operations considerations associated with thickening technologies and their application to oil sands tailings. Proceedings Third International Oil Sands Tailings Conference, Edmonton, December 2-5.  Bruker (2015). Bruker MicroCT. Retrieved from http://www.skyscan.be/company/company.htm on February 15, 2015.  Bulmer, J. & Starr, J. (1979). Syncrude Analytical Methods for Oil Sand and Bitumen Processing. Edmonton: The Alberta Oil Sands Technology and Research Authority (AOSTRA), 46-52.  Burchfield, T. & Hepler, L. (1979). Some chemical and physical properties of tailings water from oil sands extraction plants. Fuel, 58:745f.  Cabrera, S., Bryan, J., Komishke, B., & Kantzas, A. (2009). Study of the settling characteristics of tailings using nuclear magnetic resonance technique. International Journal of Mining, Reclamation and Environment, 23.1: 33-50.  CAPP [Canadian Association of Petroleum Producers], (2013). Tailings ponds. Oil Sands Today Retrieved from http://www.oilsandstoday.ca/topics/Tailings/Pages/default.aspx on September 25, 2013.   176  Casagrande, A. (1932). The structure of clay and its importance in foundation engineering. Journal Boston Society of Civil Engineers, 19: 168-209.  Castillo, K., Parsons, J., Chavez, D. & Chianelli, R. (2009). Oxidation of dibenzothiophene to dibenzothiophene-sulfone using silica gel. Journal of Catalysis, 268: 329-334.  Caughill, D. (1992). Geotechnics of Non-Segregating Oil Sand Tailings (Master’s thesis), Edmonton: University of Alberta.   CERI [Canadian Energy Research Institute], (2014). Refining Bitumen: Costs, Benefits and Analysis. Study No. 145. December, Retrieved from  Publications tab at http://www.ceri.ca/, December 14, 2014.   Chalaturnyk, R., Scott, J. and Özüm, B. (2002). Management of oil sands tailings. Petroleum Science and Technology, 20.9-10: 1025-1046.   Charlebois, L. (2012). On the Flow and Beaching Behaviour of Sub-Aerially Deposited, Polymer Flocculated Oil Sands Tailings: A Conceptual and Energy-Based Model (Master’s Thesis, Department of Mining Engineering), Vancouver: University of British Columbia.  Chu, A., Paradis, T., Wallwork, J. & Hurdal, J. (2008). The Past, Present and Future of Tailings at Syncrude: Non Segregating Tailings at the Horizon Oil Sands Project. Proceedings 1st International Oil Sands Tailings Conference, Edmonton, December 7-10.  Clark, K. (1944). Some physical properties of Alberta bituminous sands. Canadian Journal of Research, F22: 174-180.  Clark, K. & Pasternack, D. (1932). Hot water separation of bitumen from Alberta bituminous sand. Industrial and Engineering Chemistry Research, 24 (12): 1410-1416.  Coanda (Coanda Research & Development Corporation), (2014). Oscillatory Rheology. (Unpublished Progress Report), August.  Cottrell, J. (1963). Development of anhydrous process for oil sand extraction. The K.A. Clark volume: A collection of papers on the Athabasca oil sands presented to K.A. Clark on the 75th anniversary of his birthday, M. Carrigy (Ed.), Edmonton: Alberta Research Council, 196-206.  Coussot, P. (2005). Rheometry of Pastes, Suspensions and Granular Materials Applications in Industry and Environment.  Hoboken: John Wiley & Sons.  Crickmore, P., Schutte, R. & Causgrove, J. (1988). Fractal geometry and the settling of oil sands tailings sludge. Paper No. 114, 4th UNITAR International Conference on Heavy Oil and tar Sands, August 7-12.   177  CTMC [Consortium of Oil Sands Tailings Management Consultants], (2012). Oil Sands Tailings Technology Deployment Roadmap Project Report – Component 1 Results, G. McKenna and R. Dawson (Eds.), retrieved from http://www.ai.eees.ca/media/1906-component_1-report.pdf, September 15, 2015.  de Krester, R., Boger, D., and Scales, P. (2003).  Compressive rheology: An overview. Rheology Reviews, The British Society of Rheology, 125-165.  Demoz, A. Munoz, V. & Mikula, R. (2010). Optimizing MFT dewatering by controlling polymer mixing. Proceedings, 2nd International Oil Sands Tailings Conference, Edmonton. December 5-8.  Devenny, D. (2010). A Screening Study of Oil Sands Tailings Technologies and Practices. Rev. I. Prepared for Alberta Energy Research Institute AERI Contract 2008 0326. March.  Devenny, D., & Nelson, R. (2009). Economic screening of tailings options for oil sands plants. Proceedings, Tailings and Mine Waste, Banff, November 1-4.  Diamond, S. 1970. Pore size distribution in clays. Clays and Clay Minerals, vol. 18, pp. 7-23.  DOA (Department of Agriculture), (2001). Bituminous Mixtures. Field Manual No. 5, Chapter 3, 472f.  Duguet, E. (2012). Intro to hybrid organic-inorganic materials. Lecture notes, University of Bordeaux. Retrieved from http://www.icmcb-bordeaux.cnrs.fr/duguet/ on July 6, 2012.  Engels, J. (2014). What are tailings? Their nature and production. Retrieved from  http://www.tailings.info/basics/tailings.htm, January 22, 2015.  Entry, J., Sojka, R., Verwey, S., & Ross, C. (2002). Polyacrylamide removes microorganisms and nutrients from surface water [Paper and Poster]. Retrieved from http://www.ars.usda.gov/SP2UserFiles/Place/20540000/presentations/entry/asa2002/pa removesmicrobes.html on September 27, 2015.  ERCB [Energy Resources and Conservation Board], (2012). ST98-2012 Alberta’s Energy Reserves 2011 and Supply/Demand Outlook 2012-2021. Retrieved from http://www.strategywest.com/downloads/ERCB_ST98_2012.pdf , June 23, 2015. Estepho, M. (2014). Seepage Induced Consolidation Test: Characterization of Mature Fine Tailings (Master’s thesis, Mining Engineering), Vancouver: University of British Columbia, January.  Fair, A. (2008). The past, present and future of tailings at Syncrude. Proceedings, 1st International Oil Sands Tailings Conference, Edmonton. December 7-10.  Farinato, R. & Dubin, P. eds. (1999). Colloid-Polymer Interactions: From Fundamentals to Practice. New York: John Wiley & Sons, Inc.  178  Ferrera, V., & Pawlik, M. (2009). Flocculation of oil sands tailings. Proceedings, Conference of Metallurgists, Sudbury, ON, August 23-26.  Fetter, C. (2001). Applied Hydrogeology. 4th Edition. Prentice Hall, New Jersey.  Fisher, D., Clayton, S., Boger, D., & Scales, P. (2007). The bucket rheometer for shear stress shear rate measurement of industrial suspensions. Journal of Rheology, 51(5): 821-831.  Fourie, A., Johns, D., & Jones, C. (2007). Dewatering of mine tailings using electrokinetic geosynthetics. Canadian Geotechnical Journal, 44.2: 160f.  Freeze, R. & Cherry, J. (1979). Groundwater. Prentice Hall, New Jersey.   FTFC [Fine Tailings Fundamentals Consortium], (1995). Advances in Oil Sands Tailings Research, vols. I-IV, Edmonton, Alberta: Alberta Department of Energy, June.  Galosky, J. 2012. Personal communication, August.   Gamal El-Din, M. & Liu, Y. (2012). Treatment of Oil Sands Tailings Water Using Physico Chemical and Biological Processes. CONRAD Water Workshop, Edmonton, March pp. 18-19.  George, R. (2012). Sunrise: Suncor, The Oil sands and the Future of Energy. Toronto: HarperCollins.  Gibson, R., England, G., & Hussey, M. (1967). The theory of one-dimensional consolidation of saturated clays: 1. Finite non-linear consolidation of thin homogeneous layers. Geotechnique, Vol. 14, pp. 261-273.  Goldschmidt, V. (1926). Undersokelser over lersedimenter. Nordisk jordbrugsforskning, No. 47, 434-445.  Greenwood, R. & Kendall, K. (1999). Selection of suitable dispersants for aqueous suspensions of zirconia and titania powders using acoustophoresis. Journal of the European Ceramic Society, 19 (4): 479–488.  Gregory, J. (1988). Polymer adsorption and flocculation in sheared suspensions. Colloids Surfaces, 31:231-253.  Gutierrez, L. (2013). A Role of Humic Matter and Ore Oxidation in Rheology of Oil Sand Slurries and in Bitumen Extraction (PhD thesis, Mining Engineering), Vancouver: University of British Columbia.  Haigh, S. (2012). Mechanics of the Casagrande liquid limit test. Canadian Geotechnical Journal, 49: 1015-1023.   179  Hanaor, D., Michelazzi, M., Leonelli, C. & Sorrell, C. (2012). The effects of carboxylic acids on the aqueous dispersion and electrophoretic deposition of ZrO2. Journal of the European Ceramic Society 32 (1): 235–244.  Hansbo, S. (1957). A new approach to the determination of the shear strength of clay by the fall-cone test. Proceedings, Royal Swedish Geotechnical Institute, No.14.  Hiemenz, P. & Rajagopalan, R. (1997). Principles of Colloid and Surface Chemistry. 3rd Ed. New York: Marcel Dekker.  Hockley, D. (2015). Personal communication, Vancouver, May. Hogg, R. (1999). The role of polymer adsorption kinetics in flocculation. Colloids and Surfaces A: Physicochemical and Engineering Aspects, vol. 146:253-263. Hooshiar, A., Uhlik, P., Ivey, D., Liu, Q., & Etsell, T. (2012). Clay minerals in nonaqueous extraction of bitumen from Alberta oil sands: Part 2. Characterization of clay minerals. Fuel Processing Technology, 96: 183-194.  Houlsby, G. (1982). Theoretical analysis of the fall cone test. Geotechnique, 32(2): 118f. Jeeravipoolvarn, S. (2010). Geotechnical Behavior of in-Line Thickened Oil Sands Tailings (PhD dissertation, Civil Engineering), Edmonton: University of Alberta.  Jeeravipoolvarn, S., Scott, J., Donahue, R., & Ozum, B. (2008). Characterization of oil sands thickened tailings. Proceedings, International Oil Sands Tailings Conference, Edmonton, December 7-10, p.132-142.  Johnson, C. (2015). Surface properties of oil sands clay minerals (Power Point presentation). COSIA Clay Workshop, April 28, Edmonton.  Kaminsky, H. (2008). Characterization of an Athabasca Oil Sands Ore and Process Streams (PhD thesis, Department of Chemical and Materials Engineering), Edmonton: University of Alberta.  Kaminsky, H. (2011). Clay minerals and why they’re important in oil sands [Power point]. 2nd CONRAD Clay Workshop, Edmonton, March 8.  Karlsson, R. (1961). Suggested improvements in the liquid limit test, with reference to flow properties of remoulded clays. 5th International Conference on Soil Mechanics and Foundation Engineering, Vol. 1, Paris, p. 171.  Kasperski, K. (1992). A review of the properties and treatment of oil sands tailings. Alberta Oil Sands Technology and Research Authority Journal 8: 12.  180  Kasperski, K., Wan, X., Wu, J. & Dabros, T. (2012). Fate of naphtha components in oil sand tailings. Proceedings, International Oil Sands Tailings Conference, Edmonton.  Kessick, M. (1979). Structure and properties of oil sands clay tailings. Journal of Canadian Petrol Technology, 18: 49-52.  Klein, B. (1992). Rheology and stability of magnetite dense media (PhD thesis), Vancouver: University of British Columbia.  Klein, C. (2014). Effect of Residual Bitumen on Polymer-assisted Flocculation of Fluid Fine Tailings (Master's thesis, Department of Chemical Engineering), Edmonton: University of Alberta.  Konert, M., & Vandenberghe, J. (1997). Comparison of laser grain size analysis with pipette and sieve analysis: A solution for the underestimation of the clay fraction. Sedimentology, 44, 523-535  Kotlyar, L., Deslandes, Y., Sparks, B., Kodama, H. & Schutte, R. (1993a). Characterization of colloidal solids from Athabasca fine tails. Clay and Clay Minerals, 41(3), 341-345.  Kotlyar, L., Kumar, A., Ripmeester, J., Schutte, R. & Sparks, B. (1993b). Ultrafines gelation and fine tails structure formation. Proceedings, Oil Sands Our Petroleum Future Conference, Edmonton, April 4-7.  Kotlyar, L., Sparks, B., Woods, J., Capes, C. & Schutte, R. (1995a).  Biwetted ultrafine solids and structure formation in oil sands fine tailings, Fuel, vol. 74, no. 8, pp. 1146-1149.  Kotlyar, L., Sparks, B., Woods, J. & Cuddy, G. (1995b).  Oil sands facies, ultra-fines and tailings. Special Report No. ER 1332-95S, National Research Council of Canada, Institute for Environmental Research and Technology, January.  Kotlyar, L., Sparks, B., LePage, Y, & Woods, J. (1998). Effect of particle size on the flocculation behaviour of ultra-fine clays in salt solutions. Clay Minerals, 33.1:103-107.  Lambe, T. (1953). The structure of inorganic soils. Proceedings, American Society of Civil Engineers, Separate No. 315, New York, 49f.  Lambe, T. & Whitman, R. (1969). Soil Mechanics. Hoboken: Wiley.  Lawrence, L., & Ali, Z. (2010). A new approach to oil sand tailings management. Tailings and Mine Waste, London: Taylor & Francis Group, 409-415.  Laxton, P. & Berg, J. (2006). Relating clay yield stress to colloidal parameters. Journal of Colloid and Interface Science, 296: 749-755.  181  Liu, J. & Znidarcic, D. (1991). Modeling one-dimensional compression characteristics of soils. Journal of Geotechnical Engineering, 117(1), pp.162-169. Lloyd, C. (2014a). Personal communication related to delivery of the AITF-MFT sample. January 7. Lloyd, C. (2014b). Personal communication related to delivery of the AITF-MFT sample. January 8. Locat, J. & Demers, D. (1988). Viscosity, yield stress, remolded strength, and liquidity index relationships for sensitive clays. Canadian Geotechnical Journal, 25(4): 79–806.  Long, J., Li, H., Xu, Z. & Masliyah, J. (2006a). Role of colloidal interactions in oil sand tailings treatment. AIChE Journal (American Institute of Chemical Engineers), 52 (1): 371-383.  Long, J., Xu, Z., & Masliyah, J. (2006b). Adhesion of single polyelectrolyte molecules on silica, mica, and bitumen surfaces. Langmuir, 22: 1652-1659.  Mackenzie, A. (1970). The Journals and Letters of Alexander Mackenzie. W. Kaye Lamb (Ed.), Cambridge: Hakluyt Society, p. 129.  MacKinnon, M. & Retallack, J. (1982). Preliminary characterization and detoxification of  tailings pond water at the Syncrude Canada Ltd. oil sands plant. Proceedings, 4th Land and water Issues Related to Energy Development Conference, Denver 1981, Ann Arbor Science, p. 185.  MacKinnon, M. (1989). Development of the tailings pond at Syncrude’s oil sands plant: 1978 1989. AOSTRA Journal of Research, 5:109f.  Mahmoudkhani, A., Tellakula, R., Watson, P, Fenderson, T., Stewart, K., & Wu, Y. (2012). Consolidation of oil sands process tailings by enhanced coagulation – flocculation [PowerPoint presentation].  3rd International Oil Sands Tailings Conference, Edmonton, December 2-5.  Masala, S. & Matthews, J. (2010). Predicting development of undrained shear strength in soft oil sands tailings. Proceedings, Second International Oil Sands Tailings Conference, Edmonton, December 5-8.  Masala, S. & Dhadli, N. (2012). Correlations of shear strengths of soft oil sands tailings assessed by different in situ methods. Proceedings, Third International Oil Sands Tailings Conference, Edmonton, December 2-5.  Masliyah, J., Czarnecki, J., & Xu. Z. (2011). Handbook on Theory and Practice of Bitumen Recovery from Athabasca Oil Sands. Volume I: Theoretical Basis. Kingsley Publishing Services, pp. 468.   182  Matthews, J., Shaw, W., Mackinnon, M., & Cuddy, R. (2002). Development of composite tailings technology at Syncrude. International Journal of Surface Mining, Reclamation and Environment, 16(1), 24-39.  Matthews, J., & Masala, S. (2009). Tailings research at Shell's Muskeg River Mine Tailings Testing Facility. Proceedings, Tailings and Mine Waste, Banff, Edmonton: University of Alberta, Department of Civil & Environmental Engineering.  McEachern, P. (2012). Tailings and Process Water Management [PowerPoint Presentation]. CONRAD Water Workshop, Edmonton, March 18-19.  Mewis, J. (1979). Thixotropy - A general review. Journal of Non-Newtonian Fluid Mechanics, 6: 1-20.  Mewis, J. & Wagner, N. (2012). Colloidal Suspension Rheology. Cambridge: Cambridge University Press.  Michaelian, K. (2003). Photoacoustic IR Spectroscopy – Instrumentation, Applications and Data Analysis. 2nd Ed. Weinheim: Wiley-VCH.   Mikula, R. (2011). Understanding clays in oil sands processes: Operational imperative or academic curiosity? [PowerPoint presentation]. 2nd CONRAD Clay Conference, Edmonton, March 8.  Mikula, R. (2012). Evaluation of off the shelf and off the wall tailings treatment technologies: A cautionary tale. Proceedings, Third International Oil Sands Tailings Conference, December 2-5, 231f.   Mikula, R., Angle, C., Zrobok, R., Kan, J. & Xu, Y. (1993a). Factors that Determine Oil Sands Sludge Properties, CANMET Western Research Center, Division Report No. WRC-93 40, November.  Mikula, R., Munoz, V., Lam, W., Payette, C. & MacConnachie, C.  (1993b). Structure and drying behavior of oil sands fine tails: The effect of clay type and water chemistry. Proceedings, Fine Tails Symposium, Oil Sands – Our Petroleum Future, Edmonton, April 4-7.  Mikula, R., Munoz, V., & Omotoso, O. (2008a). Water use in bitumen production: tailings management in surface mined oil sands. Canadian International Petroleum Conference, Calgary, June 17-19.  Mikula, R., Munoz, V. & Omotoso, O. (2008b). Centrifuge options for production of dry stackable tailings in surface mined oil sands tailing management. Proceedings, Canadian International Petroleum Conference, Calgary, June 17-19.   183  Mikula, R., Omotoso, O. & Kasperski, K. (2008c). The chemistry of oil sands tailings: Production to treatment. Proceedings, First International Oil Sands Tailings Conference, Edmonton, Dec 7-10.  Mikula, R., Dang-Vu, T., Omotoso, O., & Lahaie, R. (2009a). Dry stackable tailings as a tailings management option: Preliminary laboratory and field experience using centrifuges. Proceedings, Canadian International Petroleum Conference and 60th Annual Technical Meeting of the Petroleum Society, 1, 607-617.  Mikula, R., Munoz, V., & Omotoso, O. (2009b). Centrifugation options for production of dry stackable tailings in surface-mined oil sands tailings management. Journal of Canadian Petroleum Technology, 48(9), 19-23.  Miller, W. (2010). Comparison of GeoEnvironmental Properties of Caustic and Noncaustic Oil Sand Fine Tailings (PhD thesis, Department of Civil and Environmental Engineering), Edmonton: University of Alberta.   Mitchell, J. (1956). The fabric of natural clays and its relation to engineering properties. Proceedings, 35th Highway Research Board, NAS-NRC 35, pp. 693-713.   Mitchell, J. and Soga, K. (2005). Fundamentals of Soil Behavior. 3rd Ed. Hoboken: John Wiley & Sons. Mizani, S. & Simms, P. (2014). Geometry of polymer-amended fine tailings deposits: Yield stress measurements and numerical modelling. Proceedings, IOSTC 2014 Conference, Lake Louise, December 7-10. Morgenstern, N. (2012). Oil sands mine closure – The end game: An update. Proceedings, Third International Oil Sands Tailings Conference, Edmonton, December 2-5.  Munoz, V., Elliot, G., Demoz, A. & Mikula, R. (2010). Flocculation behavior of mature fine tailings: Decreasing water use in oil sands extraction. Microscopy and Microanalysis, Vol. 16, Supplement S2, July, p. 1224-1225.  Murugan, R., Mohan, S., & Bigotto, A. (1998). FTIR and polarised Raman spectra of acrylamide and polyacrylamide. Journal of the Korean Physical Society, vol. 32, no. 4: pp. 505-512.  Nabzar, L. & Pefferkorn, E. (1985). An experimental study of kaolinite crystal edge polyacrylamide interactions in dilute suspensions. Journal of Colloid and Interface Science, vol. 108, Issue 1: 243-248. Najafi, A., Mian, H., Salehi, R., Lacoste-Bouchet, P., Friedrich, K., Keele, H., McWhan, C., Romero, R., Daliri, F., Jeeravipoolvarn, S., Boswell, J. & Moore, T. (2014). Proceedings, IOSTC, Lake Louise, December 7-10.  184  Neff, J., & Hagemann, R. (2007). Environmental challenges of heavy crude oils: Management of liquid waste. Proceedings, Society of Petroleum Engineers E&P Environmental and Safety Conference, Galveston, March 5-7.   Nguyen, Q. & Boger, D. (1983). Yield stress measurement for concentrated suspensions. Journal of Rheology 27: 321-349.  Nguyen, Q. & Boger, D. (1985). Direct yield stress measurement with the vane method. Journal of Rheology, 29 (3):335-347.  Omotoso, O. (2011). Clay characterization in the oil sands industry [PowerPoint presentation]. 2nd CONRAD Clay Workshop, Edmonton, March 9-10.  Omotoso, O. (2013). Role of clays in bitumen extraction [PowerPoint presentation]. 3rd CONRAD Clay Workshop, Edmonton, February 20-21. Omotoso, O. (2014). Personal communication at Rheology Short Course, Paste 2014 Conference, Vancouver. Omotoso, O., & Melanson, A. (2014). Influence of clay mineralogy on the behavior of oil sands tailings. Proceedings, Paste, Vancouver. Ovarlez G. & Coussot, P. (2007). Physical age of soft-jammed systems.  Physical Review E (76) 011406, The American Physical Society, July 19.  Pawlik, M. (2011). Mine 488 Oil Sands Processing. Lecture notes, Mining Engineering, University of British Columbia, March.  Pembina [Pembina Institute], (2013). Oil sands. Retrieved from http://www.pembina.org/oilsands/os101/tailings, September 25, 2013.  Pollock, G. (1988). Large Strain Consolidation of Oil Sand Tailings Sludge (Master’s thesis, Civil Engineering). Edmonton : University of Alberta.  Powter, C., Biggar, K., Silva, M., McKenna, G. & Scordo, E. (2011). Review of oil sands  tailings technology options. Proceedings, Tailings and Mine Waste. London: Taylor & Francis Group.  Prakash, K. (2005). Discussion of “Plastic Limit, Liquid Limit, and Undrained Shear Strength of Soil – Reappraisal” by Binu Sharma and Padma K. Bora. Journal of Geotechnical and GeoEnvironmental Engineering, 131: 402-402.  Prakash, K. & Sridharan, A. (2006). Critical appraisal of the cone penetration method of determining soil plasticity. Canadian Geotechnical Journal, 43(8): 884-888.   185  Pusch, R. (1979). Unfrozen water as a function of clay microstructure. Division of Soil Mechanics, University of Lulea, Sweden, vol. 13, Issues 1-4, April. 157-162.  Qui, Y. & Sego, D. (1998). Engineering properties of mine tailings. Proceedings, 51st Canadian Geotechnical Conference, Edmonton, October 5-7, 1:149-154.  Rao, S. (1980). Flocculation and dewatering of Alberta oil sands tailings. International Journal of Mineral Processing, 7:245-253.  Read, J. & Whiteoak, D. (2003). The Shell Bitumen Handbook. 5th ed., Shell Bitumen. London: Thomas Telford Publishing.  Robinsky, E. (2000). Thickened tailings disposal (TTD) at Kidd Creek operations, Ontario. Proceedings, Transportation and Deposition of Thickened/Paste Tailings, October 2000, Edmonton, Alberta, Canada.  Romero, E., Gens, A. & Lloret, A. (1999). Water permeability, water retention and microstructure of unsaturated compacted Boom clay. Department of Geotechnical Engineering and Geosciences, Universidad Politecnica de Catalunya, Jordi Girona 103, Mod D-2, Volume 54, Issues 1-2, September, pp. 117-127.  Rosenqvist, I. (1959). Physico-chemical properties of soils: Soil water systems. Journal of Soil Mechanics and Foundations Division, American Society of Civil Engineers 85, No. SM 2, part I, pp. 31-53.  Rosenqvist, I. (1962). The influence of physico-chemical factors upon the mechanical properties of clays. Proceedings, 9th Conference of Clays and Clay Minerals, Pergamon Press: New York, pp. 12-27.  Rubin, Jeff. (2012). The End of Growth. Toronto: Random House Canada.  Salehi, M. (2010). Characterization of Mature Fine Tailings in the Context of its Response to Chemical Treatment (Master’s thesis, Chemical Engineering). Edmonton: University of Alberta.  Sarret, G., Connan, J., Kasrai, M., Bancroft, G., Charrie-Duhaut, A., Lemoine, S., Adam, P., Albrecht, P., & Eybert-Berard, L. (1999). Chemical forms of sulfur in geological and archeological asphaltenes from Middle East, France, and Spain determined by sulfur K and L-edge x-ray absorption near-edge structure spectroscopy. Geochimica et Cosmochimica Acta, vo. 63, No. 22: pp. 3767-3779.  Scales, P. (2013). The yield and flow of clay-non clay suspension mixtures [PowerPoint presentation]. 3rd CONRAD Clay Conference, Edmonton. February 20-21.   186  Schmitz, R., Schroeder, C. & Charlier, R. 2005. Influence of microstructure on geotechnical properties of clays. Proceedings, Physics 93, Unsaturated Soils: Experimental Studies, T. Schanz (Ed.), vol., New York: Springer, 89-100.  Schneider, C., Rasband, W., Eliceiri, K. (2012). NIH image to ImageJ: 25 years of image analysis. Nature Methods, 9, 671-675. Schramm, L. (2005). Emulsions, Foams, and Suspensions: Fundamentals and Applications. August. Toronto: Wiley-VCH, pp. 463. Scott, J., Dussealt, M. & Carrier, W. (1985). Behaviour of the clay/bitumen/water sludge system from oil sands extraction plants. Applied Clay Science (1): 207-218.  Sethi, A. (1995). Methylene blue test for clay activity determination in fine tailings. MRRT Procedures. January, 23.  Sheeran, D. (1993). An improved understanding of fine tailings structure and behaviour. Proceedings, Oil Sands Our Petroleum Future Conference, Edmonton, April 4-7, 1-11.  Sheppard, M. (1989). Oil Sands Scientist: The Letters of Karl A. Clark 1920-1949. Edmonton: University of Alberta Press.  Sherwood, P. (1970). The reproducibility of the results of soil classification and compaction tests. Transport and Road Research Laboratories, Department of Transportation, UK, Report LR339. Smoluchowski, M. (1921). Electro-osmosis and current flow. Handbuch der Elecktrizitat und  es Magnetismus, L. Graetz (Ed.), Leipzig: Barth-Verlag.  Soane, D., Ware, W., Mahoney, R. & Kincaid, K. (2010). Oil sands tailings treatment via surface modification of solids with polymers. Proceedings, 2nd International Oil Sands Tailings Conference, Edmonton, December 5-8.  Sobkowicz, J. & Morgenstern, N. (2009). A geotechnical perspective on oil sands tailings. Proceedings, Tailings and Mine Waste, Banff, November 1-9.  Sobkowicz, J. (2010). History and developments in the treatment of oil sands fine tailings. Proceedings, Tailings and Mine Waste, London: Taylor & Francis Group.  Sobkowicz, J. & Boswell, J. (2011). Influence of clay mineralogy on the geotechnical stability of tailings deposits [PowerPoint presentation]. 2nd CONRAD Clay Workshop, Edmonton, March 9-10.  Socrates, G. (1980). Infrared Characteristic Group Frequencies. New York: John Wiley & Sons.   187  Soleimani, S., Simms, P., Dunmola, A., Freeman, G. & Wilson, W. (2014). Desiccation and consolidation in thin-lift deposition of polymer amended MFT. Proceedings, Paste 2014, Vancouver, 75-86.  Somogyi, F. (1979). Analysis and Prediction of Phosphatic Clay Consolidation: Implementation Package. Lakeland: Florida Phosphatic Clay Research.  Sridharan, A., El-Shafei, A., & Miura, N. (2002). Mechanisms controlling the undrained strength behavior of remolded Ariake marine clays. Marine Georesources and Geotechnology 20: 21-50.  Stuart, B. (1997). Biological Applications of Infrared Spectroscopy. David Ando (Ed.), West Sussex: John Wiley & Sons.  Suncor [Suncor Energy], (2011). Process for drying oil sand mature fine tailings – CA 2701317. Canadian Patent Application, Canadian Intellectual Property Office, March 15.  Suncor. (2013). Tailings Management. Retrieved from http://www.suncor.com/en/responsible/3229.aspx#process on September 25, 2015. Sworska, A., Laskowski, J. & Cymerman, G. (2000). Flocculation of the Syncrude fine tailings Part I. Effect of pH, polymer dosage and Mg2+ and Ca2+ cations. International Journal of Mineral Processing, 60.2: 143-52. Takamura, K. (1982). Microscopic structure of Athabasca oil sand. Canadian Journal of Chemical Engineering, 60:538-545.  Tanaka, H., Hirabayashi, H., Matsuoka, T., & Kaneko, H. (2012). Use of fall cone test as measurement of shear strength for soft clay materials. Soils and Foundations, The Japanese Geotechnical Society, 52(4): 590-599.  Terzaghi, K. (1925). Modern conceptions concerning foundation engineering. Journal of the Boston Society of Civil Engineers, 12, No. 10, pp. 1-43. Terzaghi, K. & Peck, R. (1967). Soil Mechanics in Engineering Practice. 2nd Ed., Wiley: New York. Tripathy, T. & Ranjan, De B. (2006). Flocculation: A new way to treat waste water. Journal of Physical Sciences, 10: 93-127.  Vanderklippe, N. (2010). Shell opens plant to clean up oil sands. The Globe and Mail, August 26.  Veillette, M. & Davies, M. (2007). Pogo project Alaska – Tailings and water management for a cold regions mine. Proceedings, Mining and the Environment IV Conference, Sudbury, Ontario, October 19-27.  188  Vorob’ev, N., Kurt’ko, N., Vorob’eva, and Strnadova, N. (2008). Successive adsorption of polyacrylamide compounds from electrolyte solutions on the surface of kaolinitic clay particles. Colloid Journal 20(2):148f. Wang, X. (2009). Polymer Aids for Settling and Filtration of Oil Sands Tailings (Master’s Thesis, Department of Chemical and Material Engineering), Edmonton: University of Alberta.  Wang, X., Feng, X., Xu, Z., & Masliyah, J. (2010). Polymer aids for settling and filtration of oil sands tailings. Canadian Journal of Chemical Engineering, vol. 88, Issue 3, pgs. 403- 410, June.  Watson, P., Fenderson, T., Mahmoudkhani, A., Nair, M., Patel, A., & Roberts, G. (2011). Breakage and reformation of flocs in oil sands tailings slurries. Proceedings, Tailings and Mine Waste, Vancouver, November 6-9.  Wells, P. & Riley, D. (2007). MTF drying - Case study in the use of rheological modification and dewatering of fine tailings through thin lift deposition in the oil sands of Alberta. Proceedings, Paste.  Wells, P. (2011). Long term in-situ behavior of oil sands fine tailings in Suncor’s Pond 1A. Proceedings, Tailings and Mine Waste, Vancouver, November 7-9.  Wells, P., Revington, A. & Omotoso, O. (2011). MFT drying – Technology update. Proceedings, Tailings and Mine Waste, Vancouver, November 7-9.  Whitaker, S. (1986). Flow in porous media I: A theoretical derivation of Darcy’s law. Transport in Porous Media, 1: 3-25.  Wilson, V. 2010. Personal communication. Calgary.  Wroth, C. & Wood, D. (1978). The correlation of index properties with some basic engineering properties of soils.  Canadian Geotechnical Journal, 15(2): 137-145.  WWF [World Wildlife Fund], (2010). Tailings, A Lasting Oil Sands Legacy.  Retrieved from http://www.wwf.ca, Climate & Energy Publications & Reports section, October 12, 2014.  Wysokinski T., Chapman, D., Adams, G., Renier, M., Suortti, P. & Thomlinson, W. (2007). Beamlines of the biomedical imaging and therapy facility at the Canadian Light Source Part I. Nuclear Instruments and Methods in Physics Research A, vol. 582 (1):  73-76.  Xu, Y. & Hamza, H. (2003) Thickening and disposal of oil sand tailings. Proceedings, Mining Engineering, Littleton, Issue 55, pp 33 –39.  Yao, D. & Znidarčić, D. (1997). User's manual for computer program CONDES0. Boulder: Florida Institute of Phosphate Research.  189   Yao, Y., van Tol, F. & van Paassen, L. (2012). The effect of flocculant on the geotechnical properties of mature fine tailings: An experimental study. Proceedings, Third International Oil Sands Tailings Conference, Edmonton, December 2-5.   Yao, Y., van Tol, F., van Paassen, L. & Vardon, P. (2014). Shrinkage and swelling properties of flocculated mature fine tailings. Proceedings, Fourth International Oil Sands Tailings Conference, Lake Louise, December 7-10.  Yergin, D. (2008). The Prize: The Epic Quest for Oil, Money & Power. New York: Free Press. Yeung, A., Gibbs, A., & Pelton, R. (1997). Effect of shear on the strength of polymer-induced flocs. Journal of Colloid and Interface Science, 196, 113-115. Yong, R. & Sethi, A. (1978). Mineral particle interaction control of tar sand sludge stability. Journal of Canadian Petroleum Technology, Oct-Dec, 76-81.  Zentar, R., Abriak, N. & Dubois, V. (2009). Fall cone test to characterize shear strength of organic sediments. Journal of Geotechnical and GeoEnvironmental Engineering, vol. 135, No. 1: 153-157.  Zeta-Meter [Zeta-Meter Inc.], (1993). Everything you want to know about coagulation & flocculation. 4th Ed. Staunton: Zeta-Meter, Inc.  Znidarčić, D., Abu-Hejleh, A., Fairbanks, T. & Robertson, A. (1992). Consolidation Characteristics Determination for Phosphatic Clays: Volume 1: Seepage Induced Consolidation Test Equipment Description and User’s Manual. Bartow, Florida: Florida Institute of Phosphate Research.  Znidarčić, D., van Zyl, D., Fredlund, M. & Wells, S. (2011). Consolidation testing of oil sand fine tailings. Proceedings, Tailings and Mine Waste, Vancouver, November 6-9. Znidarčić, D. (2014). Personal communication reviewing results of SICT, November 21.     190  Appendices Appendix A:  A.1: AITF Characterization of AITF-MFT   191    192         193  A.2: Exova Characterization of AITF-MFT Table A.1: Summary of Baseline Characterization Data from Exova Labs   194   195  Appendix B: Procedures and Test Methods B.1: Polymer Hydration Procedure The required volume of PEW was obtained using a clean beaker and measuring cylinder from the covered 5 gallon buckets in which it was stored.  The required volumes of PEW were then placed in clean beakers and the masses of dry polymer required to achieve the desired polymer concentration, were measured using a calibrated scale.  A 6 blade Phipps and Bird jar tester was then used to create a vortex in the PEW contained in each of the beakers (Figure B.1).  Figure B.1: Polymer Hydration using Phipps and Bird Jar Test Mixer The dry polymer granules were gradually added to each beaker while periodically stopping the mixer to remove undissolved polymer granules (aka “fish eyes”) from the mixer blades.  Mixing continued until all polymer granules had been fully hydrated and a clear fluid of uniform viscosity resulted.  This combination of polymer and PEW required an hour of mixing using a mixing speed of approximately 100 rpm.  The hydrated polymer had a pH ranging between 8.3 and 8.7, consistent with the pH of the PEW used.  Only polymer that had been hydrated for less than 10 days was used to flocculate raw AITF-MFT.  196  B.2: Creation of Polymer-Amended Mature Fine Tailings Approximately 317 g of wet MFT (with an approximate solids content of 30%) was placed in each of three glass beakers.  The quantity of hydrated polymer of desired concentration is then added to the MFT as follows: • With the jar tester turned off, hydrated polymer was added to a beaker containing the indicated mass of raw MFT in small increments (<5 mL initially and then in 1 mL increments) using a large plastic 60 mL syringe.  • After each increment of hydrated polymer was added, the jar tester was used to mix the polymer into the raw AITF-MFT using what is described as the “fast-slow” technique.  Using this method, the paddle speed was alternated between fast (> 200 rpm) and slow (<30 rpm) speeds to distribute the polymer into the MFT while avoiding overshearing of the sample. • Polymer addition combined with “fast-slow” mixing was continued until initial signs of the desired flocculation condition appeared in the form of free water, and the beaker contents started to exhibit a curdled texture.  At that time the mixing blades of the jar tester were set to rotate at speeds less than 10 rpm to avoid overshearing of the flocs.  The effect of continued mixing was closely observed to ensure continued development of the flocs. • Once the thickened MFT had achieved the desired open “cottage cheese” fabric with abundant visible voids which is accompanied by a noticeable release of clear water, the amount of this initial released water was then measured and the volume of polymer added was recorded. The polymer solution consistently producing the largest quantity of water released at the time of flocculation (typically between 30% and 40%), and cottage cheese fabric providing the  197  greatest resistance to force applied using the flat face of a metal pallet knife, is typically the one selected in industry for additional testing. B.3: CanmetENERGY Methylene Blue Procedure: Sludges and Slurries (2008) Objective and Scope The AST methylene blue procedure is intended to address dispersion of oil sands process solids prior to methylene blue adsorption as described in general by the ASTM Standard Test Method for Methylene Blue Clay (C 837-99, re-approved in 2003). A mature fine tailings (MFT) analog is used as a calibration standard, to monitor the level of dispersion provided by the dispersion equipment used for sample preparation. The MFT analog is comprised of sand, montmorillonite, de-ionized water and bitumen. Bitumen and water are removed using Dean Stark extraction, the same procedure used for extracting bitumen and water from test samples. Purpose This procedure was developed specifically for oil sands process solids but could be used for oil-free minerals as well. Apparatus • 250 ml beakers • 1l volumetric flasks • analytical balance, accurate to 0.001 g • hotplate/magnetic stirrer and stir bars • methylene blue powder • mortar & pestle • top-loading balance, accurate to 0.01 g • disposable pipettes • ultrasonic bath (Cole Parmer 1 ½ gallon with 40kHZ transducers & built in seep frequency) • room temperature water bath • hand held pH meter • burette stand • 50 ml burette with Teflon stopcock • Whatman 42 ashless filter paper  198  • watchglasses • Dean-stark extraction apparatus – for preparation of MFT-analog calibration standard.  Reagents • Analytical grade reagents shall be used in all tests. • De-ionized water shall be used unless otherwise indicated in the procedure. • Methylene blue trihydrate (M.W. 373.9) (1 ml = 0.006 meq).  Dissolve 2.2436 g of Methylene blue powder in 1 L de-ionized water (or 1.1218 g in 500 ml). Wrap volumetric flask in aluminum foil to keep the solution from degrading. A fresh batch should be used within a day. • 10% w/w NaOH (sodium hydroxide). Dissolve 10 g of NaOH pellets in 90 g of de-ionized water. • 10% v/v H2SO4 (sulfuric acid).  Add 10 ml of concentrated H2SO4 to 90 ml de-ionized water. • 0.015M NaHCO3 (sodium bicarbonate). Add 1.26 g of dry NaHCO3 to 1 L de-ionized water. • pH 4, 7 & 10 buffer solution (calibrating pH meter). • Coarse Ottawa sand, retained on 200-mesh sieve. • Na-montmorillonite, marketed as Bentonite by Fisher Scientific. • Solids-free oil sands bitumen.  Safety Precautions Wear gloves and protective eyewear when handling caustic and acidic agents. Wear gloves and protective eyewear when handling methylene blue. It is particularly painful if it enters the eyes and will stain skin a very dark blue.  If you get methylene blue on your clothes or skin apply Sunlight dish soap immediately to the area (without water). MFT-analog Standard Preparation (not used in this research) 1. Add Ottawa sand (35 g), water (55 g), bitumen (5 g) and bentonite (5 g) into a clean glass jar.   2. Mix the 100 g mixture thoroughly. 3. Run the entire mixture through standard Dean Stark extraction.  Sub-sampling without adequate homogenization may cause the bentonite to preferentially pass through the Dean-Stark thimble.   4. Once bitumen and water have been extracted, the dried solid serves as a bulk standard sample containing 12.5 wt% bentonite.  This gives a methylene blue adsorption capacity similar to MFT.   5. The bitumen free standard is referred to as an MFT-analog.  6. Gently shatter the MFT-analog to remove clumps of materials.  199  7. Store as bulk standard.  Procedure for Testing Standards 1. Prepare 500 ml of 0.006 N methylene blue solution.  Stir solution at 400 rpm for a minimum of 10 minutes to ensure that all the dye is dissolved.  Prepare this solution fresh daily. 2. The MFT-analog must have gone through the Dean Stark extraction process or equivalent bitumen removal process. 3. Bentonite is used as received. 4. Gently shatter the MFT-analog sample with a mortar and pestle to ensure that the particles will disperse during the titration. 5. Measure out in a clean beaker, approximately 1 g of MFT-analog on a top loading balance and record the weight. 6. In a second clean beaker, measure approximately 0.2 g bentonite on a top loading balance and record the weight. 7. Add 50 ml of 0.015 M NaHCO3 to each sample.  Add NaHCO3 carefully so that the solids don’t get spread all over the inside of the beaker.  Use a disposable pipette to rinse the sides of the beaker with the NaHCO3 solution. 8. Add 2 ml of 10% w/w NaOH solution to each sample with a disposable pipette.  Place a clean dry stir bar in the mixture.  Cover the beaker with a watchglass to keep the sample from evaporating. 9. Place the two beakers (containing Bentonite and MFT-analog) on a hotplate/stirrer set to 120°C (the sample should not heat above 60°C with the hotplate set at this temperature – monitor temperature as different hotplates heat at different rates).  Set the stirrer to a minimum of 250 rpm.  Adjust the mixing speed as required.    Make sure the sample is mixing completely and that all solids are in suspension. 10. The samples shall be stirred and heated for a minimum of 20 minutes.  If the solids have not completely dispersed after this time (note: small clay balls floating on solution surface) the sample can be transferred to an ultrasonic bath for 20 minutes.  Continue this cycle of heating/stirring and ultra-sonication until the sample is completely dispersed.  NOTE:  DISPERSION IS THE MOST IMPORTANT PART OF THIS METHOD.  If the sample is not dispersed completely the titration results will be inaccurate.    For MFT-analog, 20 minutes stirring followed by 20 minutes sonication is often adequate   for complete dispersion.  11. Place the beakers in a room temperature water bath and let them cool for 3-5 minutes.  The following steps are carried out for the Bentonite first, followed by the MFT-analog. 12. Place the beaker on a stirring plate located underneath the burette and start stirring the sample (no heat). 13. Place the handheld pH meter in the solution of the sample.  Add 10% v/v H2SO4 until the pH drops to 2.5-3.8.  200  14. Titrate the sample with 0.006 N methylene blue while the sample is stirring.  Add the methylene blue solution in 1 ml increments.  Switch to 0.5 ml increments when the sample is close to reaching the endpoint.   From prior experience with pure bentonite, depending on the batch, the endpoint can  range between 28 and 34 ml. 15. With a transfer pipette, remove an aliquot of the sample (after each titrant addition) and place one droplet on a piece of Whatman 42 ashless filter paper.  Continue placing drops of sample on the filter paper after each addition of titrant until the endpoint is reached.  The endpoint is observed when there is excess of methylene blue in the water phase of the sample indicating.  A blue halo will form around the sample droplet on the filter paper.    Procedure for Slurry to be Tested  1. Prepare 500 ml of 0.006 N methylene blue solutions.  Stir solution at 400 rpm for a minimum of 10 minutes to ensure that all the dye is dissolved.  Prepare this solution fresh daily. 2. Acquire % solids data (in duplicate) for the sample prior to methylene blue analysis.  Assess the sample and make sure that it has a low amount of bitumen contamination. 3. Place a clean dry beaker on the top-loading balance and tare the weight.  Before adding the sample please take the following into consideration: i. Is the sample very sandy?  If so, use 20 g of sample. ii. Does the sample contain large proportions of clay?  If so, use 10 g of sample.  4. Shake the sample extremely well before adding to the beaker.  It is very important to acquire a representative homogenous sample for this analysis. 5. If the sample has less than 5% solids it is difficult to get enough solids in suspension for an accurate titration.  Samples with low amounts of solids will have very low endpoints (typically less than 5 ml).  Modify the procedure by adding 50 ml of sample ONLY to the beaker.  Do not add any additional water.  Continue sample preparation as described below in steps 10-19. 6. Add the sample to the beaker and record the weight of the sample used. 7. Determine the amount of solids in the sample weighed out. 8. Example: Percent Solids = 35% Sample Wt. = 10.01 g Amount of solids in sample= 10.01 x 0.35 There are 3.5035 g of solids in the sample. 9. From this we can assume that the remaining weight of the sample is water: Therefore:  10.01 g – 3.5035 g = 6.5065 g.  There are 6.5065 g of water. 10. If 10 g of sample was used ADD 40 ml of de-ionized water to the sample. 11. If 20 g of sample was used ADD 80 ml of de-ionized water to the sample.  12. Add 2 ml of 10% w/w NaOH solution to the sample with a disposable pipette.  Place a clean dry stir bar in the mixture.  Cover the beaker with a watchglass (this will keep the sample from evaporating). 13. Determine the total amount of water in the sample (g water + 40 g water added = 46.5065 g).  201  14. Calculate the amount of 1.0 M NaHCO3 that needs to be added to the slurry.  We want the concentration of NaHCO3 to be 0.015 M in the solution. 15. C1V1 = C2V2 where (46.5065 g H2O)(0.015 M NaHCO3) = (1 M NaHCO3)(X) making X = the volume of 1 M NaHCO3 in ml that should be added to the sample. 16. Place the beaker (containing the sample) on a hotplate/stirrer that has the temperature set to 120°C (the sample should not heat above 60°C with the hotplate set at this temperature – monitor temperature as different hotplates heat at different rates.  Set the stirrer to a minimum of 250 rpm.  Adjust the mixing speed as required.  MAKE sure the sample is mixing completely and that all solids are in suspension. 17. The samples should stir and heat for a minimum of 20 minutes.  If the solids have not completely dispersed after this time (note: small clay balls floating on solution surface) the sample can be transferred to the sonicator for 20 minutes.  Continue this cycle of heating/stirring and sonication until the sample is completely dispersed.  NOTE:  DISPERSION IS THE MOST IMPORTANT PART OF THIS METHOD.  If the sample is not dispersed completely the titration results will be inaccurate. 18. Evaluate the sample for dispersion.  Is there hydrophobic mineral floating on the surface?  Are there clay balls present? Does the suspension settle rapidly after stirring has stopped?  Do the mineral particles start to flocculate after stirring has stopped?  If any of these qualitative questions can be answered with a ‘YES’, the sample is not fully dispersed. Repeat step 14 until dispersion is complete. 19. Place the beaker in a room temperature water bath and let it cool for 3-5 minutes. 20. Place the beaker on a stirring plate located underneath the burette and start stirring the sample (no heat). 21. Place the handheld pH meter in the solution of the sample.  Add 10% v/v H2SO4 until the pH drops to 2.5-3.8. 22. Titrate the sample with 0.006 N methylene blue while the sample is stirring.  Add the methylene blue solution in 1 ml increments.  Switch to 0.5 ml increments when the sample is close to reaching the endpoint. 23. With a transfer pipette, remove an aliquot of the sample (after each titrant addition) and place one droplet on a piece of Whatman 42 ashless filter paper.  Continue placing drops of sample on the filter paper after each addition of titrant until the endpoint is reached.   The endpoint will be evident when a light blue halo spreads out from the sample droplet on  the paper.   Record the ml’s required to reach this endpoint.    202  Calculations    Standardization  MBI MFT -Analog (± 0.25 meq/100g) =  MBIBentonite ×  Wt% Bentonite in MFT-analog100 The precision is taken from the ASTM Standard: C837-99.  Wt% Bentonite in MFT - analog =  Mass of BMass of B +  mass of S×100 Mass of B is the mass of bentonite added to MFT-analog Mass of S is the mass Ottawa sand added to the MFT-analog IF THE MFT-ANALOG MBI IS OUTSIDE THE SPECIFIED RANGE, THE DISPERSION EQUIPMENT SHOULD BE CHECKED AND TEST SAMPLES REPEATED.  References 1. ASTM Standard: C 837 – 99 (re-approved in 2003) 2. Sethi, Amar.  January 23, 1995.  Methylene Blue Test for Clay Activity Determination in Fine Tails.  MRRT Procedures.   203  B.4: Mineralogical Characterization of PA-MFT B.4.1 QEMSCAN Analysis All samples air dried PA-MFT were photo documented and cut into smaller pieces using a rock saw to prepare polished sections.  Ethylene glycol was used as the cutting fluid and an oil based suspension used for sample polishing to prevent introduction of water into the samples (aka dry preparation method).  Polished epoxy mounts of the samples (30 mm diameter) were prepared.  The epoxy mounts were placed in the sample holding chamber of the QEMSCAN (Figure B.2) and analyzed using both the back-scattered electron (BSE) signal intensity and the energy disperse x-ray signal (EDS) at each scanning point.  Figure B.2: Set up of QEMSCAN with detectors (right) and workstation (Mines, 2015) Using the Field Scan analysis mode, a chemical spectrum map of the polished epoxy mounts were obtained for the time interval used for each field of view.  The chemical spectrum maps for each field of view were processed to develop a pseudo image which was further analyzed to produce information about mineral texture, identification and modal abundance.  To account for the presence of residual bitumen, the BSE signal intensity threshold was reduced to enable visualization of carbon which is usually not bright enough to be mapped.    204  B.4.2 Chemical Assay and XRD Analysis Samples of dried PA-MFT resulting from preparing the epoxy mounts (approximately 100 g) were collected and analyzed as follows: • Approximately 20 g was pulverized and submitted for chemical assay which included determination of sulphur species using a Leco sulphur analyzer, total carbon determination, and whole rock analysis using X-ray fluorescence; • Approximately 10 g was pulverized and submitted for XRD bulk analysis; and • Approximately 60 g was pulverized to produce 80% passing 212 µm for XRD clay speciation analysis. To complete XRD bulk analysis, approximately 2 g of each sample was prepared as a pressed mount and scanned in a Bruker AXS D8 Advance Diffractometer.  To complete the clay speciation analysis the minus 2 µm fraction was removed from the indicated subsample using a combination of sonication and centrifugation.  The oriented minus 2 µm fraction was mounted on a slide and scanned in the diffractometer to produce “untreated” results.  XRD scans of each sample were taken after each round of treatment which included addition of ethylene glycol and high temperature heating.  The various clay mineral species were identified using both their untreated diffraction patterns as well as observation of characteristic changes in their diffraction pattern resulting after each treatment stage.  205     206  B.4.2 (continued)       207  B.5: Zeta Potential Backgrounder and Procedure Background Zeta potential provides a measure of the electrokinetic potential of a suspension.  It is measured in millivolts and is typically denoted by the symbol for the Greek letter zeta (ζ).  Zeta potential is not a direct measure of charge on the surface of a material but provides a measurement of the effectiveness of the surface charge on ions in solution (Hunter, 1981).  Surface charge has the greatest effect in solutions where ionic concentration is reduced.  By creating very dilute solutions, the nature of the electrokinetic potential of the particles in suspension can be measured by subjecting them to an electric field and observing whether the particles are attracted to the anode or cathode and the speed at which article migration occurs.  Electrophoresis is the term used to describe the movement of charged particles under an electric field and is the means by which electrophoretic instruments measure the mobility of solid particles (v) using the Smoluchowski (1921) equation as follows: 𝒗𝒗 = 208  suspensions dominated by clay minerals that are plate like, deformable, and exhibit heterogeneous surface charges, cannot be overlooked (Gupta and Miller, 2010).  The heterogeneous nature of charges on the deformable surface of bitumen droplets must also be considered (Chow and Takamura, 1988; Lin, 2012). Zeta potential provides an indication of colloidal stability, which simply is an indication of the propensity for the particles in suspension to either remain in suspension or to fall out of suspension.  Zeta potential is also affected by the pH of the suspension.  Consequently, manipulation of a suspension’s pH is used to determine the isoelectric point (iep) of particles (Ferrera and Pawlik, 2009).  However, because bitumen extraction processes using hot water have been optimized for process water with pH ranging between 8 and 9, the stability of MFT, polymer and PA-MFT within this pH range are the main focus of this research. Zeta Potential Test Procedure The zeta potential of primary extraction water, hydrated anionic polymer and dilute suspensions of raw AITF-MFT and PA-MFT was determined using a Zeta-Meter 3.0® (Figure B.3).  Figure B.3: Zeta-Meter 3.0 used to determine Zeta Potential of PA-MFT and Input Materials  209  Dilute suspensions (concentration ranging between 0.1% and 0.3%) of wet and air dried samples of raw AITF-MFT and PA-MFT were prepared by adding distilled water to glass test tubes containing small subsamples of each test material.  The test tubes were stoppered to limit water loss through evaporation prior to each test.  Each suspension for which the zeta-potential was analyzed was tested as follows: • The electrophoresis cell was cleaned prior to being filled with the test suspension. The cylinder-type molybdenum anode (Figure B.3) connected to the red electrical lead was installed into the left compartment of the electrophoresis cell, followed by placement of the platinum-iridium strip-type cathode into the portion of the test suspension filling the right compartment of the electrophoresis cell. • The viewing stage was then adjusted to position the central tubular chamber of the electrophoresis cell directly beneath the optical axis of the microscope.  Then, ensuring that the Zeta-Meter was in “Standby” mode, the power supply leads were then connected to the electrodes that had been mounted in the electrophoresis cell. • Using the frontal illuminator (high light) mode, the 8x objective on the stereoscopic microscope was then focused on the front wall of the electrophoresis cell, at the exact mid-depth using the OX positioning line.  After the cell had been properly aligned, the oblique illuminator (low light) mode was selected to limit the amount of light heat transmitted to the suspension in the electrophoresis cell. • The Zeta-Meter was then operated in “Specific Conductance” mode to determine the specific conductivity of the test suspension.  This information was used to  210  determine the appropriate voltage to be applied.  The appropriate voltage setting was then selected on the Zeta-Meter power source. • The Zeta-Meter unit was then operated in the “Energize Electrodes” mode.  The direction of colloid movement i.e. towards the anode (+) or cathode (-), was observed and recorded.  Then, using a stop watch, the amount of time taken by observed colloids to traverse one micrometer division (viewed through the microscope’s eyepiece) was recorded.  The size of the grid (e.g. eighth, quarter or full scale) used was based on the speed at which colloids were observed to be moving i.e. the slower moving the colloid, the smaller the grid increment used.  The travel time of at least ten colloids across the grid increments in the test suspension was measured and recorded.  The average colloid travel time was then calculated.  This value was then used to determine the zeta potential using the data interpretation curves found in Zeta-Meter (1964).  Because the curves were developed using the Helmholtz-Smoluchowski formula assuming a suspension temperature of 22.5oC, the zeta potential values were corrected using the temperature correction factors provided with the data interpretation curves. B.6: Rheological Test Methods  B.6.1  Vane in Cup Test Method The infinite cup was filled with PA-MFT using a large plastic spatula to transfer the sample from the 1 L glass mason jars used to store the samples.  Care was taken to minimize the amount of sample disturbance associated with loading the cup.  The vane spindle was then installed into the drive shaft of the rheometer.  The cup holding the sample was then installed into the holding vice attached to the rheometer.  This cup holding vice was also designed by Coanda to ensure that  211  when locked in place the top of the vane being used is covered by a thin layer (<1 mm) of the sample being tested and that the vane is positioned in the middle of the sample, which locates the vane blades at their furthest distance from the cup sidewalls (see B in Figure 3.5).  Using the rheometer control software, the rotational speed of the vane was set to be constant at 0.1 rpm.  Selection of this rotational speed was based on findings by Nguyen and Boger (1983) and confirmed by Gutierrez (2013) who indicated that use of this speed with the vane rheometer had negligible effects on measured torque.  The torque required to maintain the indicated constant rotational speed was recorded over a period of approximately 8 minutes.  Collected torque values were used to calculate associated shear stress values using Equation 3.6.  B.6.2 Slump Test Procedure Background The slump test was originally developed to provide a simplified means of determining whether a concrete mix was suitable for use (Christensen, 1991).  Pashias et al. (1996) modified the method to measure the flow behavior of concentrated slurries.  The method involves filling an empty cylinder of known height and diameter with the material to be tested and then lifting the cylinder evenly enabling the test material to slump under its own weight, ideally uniformly (Figure B.4).  Figure B.4: Schematic Showing Slump Test with Cylindrical Ring (after Klein and Hallbom 2014)  Using a cylindrical configuration, the yield stress (τy) is expressed using the following relationship:  212  𝝉𝝉𝑫𝑫 = 𝝆𝝆𝟏𝟏𝒉𝒉𝟏𝟏𝟐𝟐     Equation B.2 where ρ is the slurry density, g is acceleration due to gravity and h0 is the height of the unslumped portion of the test material (shown in Figure B.4 as the portion of the “slumped slurry column” retaining vertical walls). Test Method A metal cylinder, also referred to as a Boger ring, 7.3 cm in diameter and 7.4 cm tall was used.  The clean cylinder was placed on a flat glass surface that was also clean.  The cylinder was filled with PA-MFT using a large plastic spatula until the test material was level with the upper rim of the cylinder.  After approximately 1 minute the cylinder was lifted evenly allowing the test material to deform under its own weight.  The height of the slump was measured and the yield stress calculated using Equation B.2. B.6.3 Determination of Storage and Loss Moduli of PA-MFT To limit the effects of slippage, a layer of fine sand paper (#400 grade) was affixed to both surfaces of the plate geometry touching the test material using a dual sided adhesive strip.  A small sample of PA-MFT was placed on the lower plate and the upper plate was adjusted using the rheometer’s control software to bring the entire surface of the upper plate into intimate contact with the test material, without causing the sample to be significantly deformed beyond the limits of the upper plate.  The small portions of the sample extending beyond the limit of the upper plate were removed using a scalpel.  When the sample had been loaded in the test geometry it had a maximum height of 3 mm as reported on the rheometer’s readout interface.  The rheometer’s control software was then used to complete a strain sweep experiment at a frequency of 1 Hz.  Using this frequency, the applied strain was allowed to range between 0.1% and 200% while the corresponding changes in shear stress were calculated and reported through the rheometer’s visual  213  interface. The resulting G’ and G’’ plots were analyzed to better understand the stress responses of PA-MFT under different strain.  This test was repeated twice using two samples of PA-MFT that were flocculated in the same batch. B.7: Fall Cone Test Procedure The sample cup provided with the device was filled to its rim with the material being tested using a spatula.  The sample cup was then placed on the test platform of the fall cone unit and the cone to be used loaded into the unit’s release mechanism.  The tip of the cone was brought into light contact with the surface of the sample in the sample cup by lowering the arm holding the cone.  The cone was then released causing it to penetrate the surface of the sample beneath (image B in Figure 3.7).  The depth of penetration was recorded after 1 minute and after 5 minutes using the readout dial located on the device.  After the penetration depth was recorded the material’s moisture content was determined.  The undrained strength was determined using the appropriated clay penetration tables found in Hansbo (1957).  A minimum of three cone penetration measurements were taken for each sample.  Remolding/sample disturbance effects were also investigated by subdividing several test samples and applying different amounts of remolding energy to them during filling of the sample cup.  The undrained strength and moisture content data were used to develop a parametric relationship that could be used to evaluate the material performance of PA-MFT. Using the Swedish fall cone the liquid limit (wL) is defined as the moisture content at which a 60 g, 60o apex cone (Figure 3.7) leaves an impression of 10 mm in the test material (Karlsson, 1961; Hansbo, 1994). The liquid limit of PA-MFT was determined using this criterion by tracking how impressions related to the indicated cone changed as initially wet samples of PA-MFT dried in air.  The liquid limit determination was based on results from fifteen tests conducted using five samples of PA-MFT.  214  B.8: SICT Procedure B.8.1 SICT of raw AITF-MFT The sample of raw AITF-MFT was homogenized as described earlier in this Chapter and a subsample used for percent solids and moisture content determination.  After the pore pressure transducer and linear variable differential transformer (LVDT) of the UBC SICT cell were calibrated, the water feed lines and pressure transducer were de-aired and functionality of the LabView software interface to control equipment valve positions and pump speed confirmed.  The surfaces of the loading piston and portions of the inner triaxial cell sample chamber were lightly sanded to enable the loading piston to freely move under its own weight, as well as in response to loads applied by the load piston.  Once the inner sample chamber cylinder was installed to hold the lower porous plastic disc covered in filter paper in place, a large measuring cup was used to transfer raw AITF-MFT into the test chamber.  Filter paper and the upper porous plastic disc were gently placed on top of the sample in the sample chamber prior to adding approximately 2 L of PEW to the test chamber using a funnel to limit the impact of water on the upper surface of the sample.  The perforated loading piston was gently placed into the sample chamber and rotated into place along the sidewalls of the cell until the bottom portion of the piston made light contact with the upper surface of the upper porous plastic disc.  The outer cell of the triaxial chamber was installed and the upper portion of the triaxial cell installed and bolted in place.  The inner and outer cells of the triaxial chamber were allowed to fill with water from the PEW reservoir attached to the SICT cell.  In its simplest terms, the SICT procedure involves an initial determination of the amount of settlement that occurs in the sample under a combination of its own weight, a small buoyant weight (approximately 0.1 kPa), and a free moving LVDT rod.  Once settlement is assumed to stop a flow pump is used to induce a small flow rate to trigger the consolidation process.  Steady  215  state settlement and pore pressure measurements that result from seepage induced flow are measured.  This stage is followed by a series of loading steps in which the loading bit is used to apply the desired load to the loading piston which is in contact with the sample.  With the desired load applied, the pump is used to apply a small flow rate.  This enables permeability under increased effective stress conditions to be determined.  At the end of the test the sample is removed from the triaxial cell and its solids content and geotechnical moisture content determined using the method described in Section 3.4.1. B.8.2 SICT of PA-MFT The difference between the preceding MFT testing procedure and that used for PA-MFT is that a large scoop was used to place the sample into the test chamber.  It was also noted that the flocculated nature of PA-MFT impacted the ability to create a uniformly even surface at the top of the sample prior to installing the upper filter paper and plastic porous disc.  It should be noted that the water initially released after flocculation of the PA-MFT was removed prior to placing the sample in the test chamber. B.9: Overview of Synchrotron-Based Computer Tomography A synchrotron is a particular type of cyclic particle accelerator that generates a brilliant light that can be used to investigate the structural and chemical properties of materials at the molecular level.  Synchrotron based computer tomography (CT) is a non-destructive imaging technique that builds on the enhanced imaging that can be generated from the light source used and from significant advances in computing capacity.  As such, the method has been widely used with great effectiveness in petrochemical engineering and materials engineering to characterize the pore network of unconsolidated porous media in three dimensions (Al-Raoush and Wilson, 2005).  The non-destructive nature of the method enables imaging of actual pore space while preserving pore connectivity and spatial variability, and configuration of pores and pore throats.   216  The method has also been used to characterize the following: pore topology and connectivity (Spanse et al., 1994); porosity, specific surface area and pore size distribution (Coker and Torquato, 1996); porosity and water content (Coles et al., 1998); and to complete geometrical analysis of a soil fabric (Lindquist and Venkutarangon, 1999).  The combination of high energy light sources from synchrotrons and enhanced computer processing capabilities has resulted in the ability to develop detailed imaging of a complete three-dimensional domain of material samples (Wysokinski et al., 2007).  As it relates to flow in porous media, this method is used to evaluate the accuracy of and to refine various algorithms that are used to simulate porous networks and to estimate the permeability of those networks.  It should be noted that the majority of algorithms currently used to analyze porous networks were developed for ideal (i.e. glass bead) and natural porous media systems like sandstone in which pores and pore necks are represented using the geometry of simple shapes like spheres, cuboids, and cylinders (Rajaram et al., 1997; Fischer and Celia, 1999).  However, Luffel et al (1991) confirmed that these idealized models failed to capture the effects that compression can have on reducing the hydraulic path to particle asperities, even in systems comprised of porous media like sandstone.  In other words, imaging using synchrotron based CT has the ability to support continued refinement of existing flow models, enabling them to better represent actual pore space as compared to that assumed by idealized geometry.  Beamlines at various synchrotron facilities have been developed to provide three dimensional imaging at voxel resolutions ranging from 30 µm to the sub-micrometer scale (Al-Raoush and Wilson, 2005; CLS, 2015; ESRF, 2015).  The quality of data obtained using this method is affected by photon energy, photon flux, sample size and type, and the nature and size of the sample features (Al-Raoush and Wilson, 2005).  Image quality and the ability to distinguish  217  sample features are directly impacted by the spatial resolution and contrast resolution (Wildenscheld et al., 2002).  Said differently, if the spatial resolution of an image is larger than the distance between objects, the objects will be undistinguishable from each other in the image.  B.10: Overview of X-ray Adsorption Near Edge Spectroscopy X-ray absorption spectroscopy (XAS) is technique that has been widely used to determine the local geometric and electronic structure of matter (ESRF, 2015).  When completed at synchrotron light sources, the method is known as x-ray absorption near edge spectroscopy (XANES) or extended x-ray adsorption fine structure (EXAFS), in which the sample of interest is exposed to a photon energy source tuned to a specific energy range to probe the free orbital space of the elements being investigated.   Figure B.5 indicates the various components of the signal generated using this method.  In general terms, photoelectrons of a given charge produce a disturbance of core electrons resulting in the expulsion of a core electron to an outer orbital shell.  As part of the research described in this thesis, the responses occurring in the core 1S orbital (also known as the K-edge) for sulphur in samples of PA-MFT and its input materials were investigated.  Because this method is element specific, it enables the valance and related form of the atoms present on the sample surface to be identified by comparing the spectra received to spectra developed using reference standards.  Use of this method in this research is described in Chapter 3. B.11: FESEM and cryo-SEM Sample Preparation and Imaging Procedure Because both samples still contained moisture from water at the time of testing, subsamples of both materials to be imaged using FESEM were obtained using cylindrical thimbles with an approximate volume of 0.8 cm3 and subjected to graded ethanol and hexamethyldisilizane (HDMS) series in an attempt to dehydrate them without collapsing the voids.  Using this method,  218   Figure B.5: Schematic Showing X-ray Absorption Edge (after Wikipedia as presented on UC Davis ChemWiki, 2015)  the subsamples were placed in small beakers containing ethanol with concentration increasing from 20% for the first test, to 100% for the final three baths with both dehydrating agents.  Each bath lasted approximately 15 minutes.  A small portion of each sample was attached to an aluminum SEM mount with hot glue prior to being placed in the imaging chamber.  Images of sample surfaces at various magnifications and inclinations were obtained.  The samples were then “sputter” coated with a 4 nm film of AuPd prior to completing a second round of imaging to see if additional sample features could be observed.  The oven dried sample imaged did not undergo a dehydration series as its moisture content was low enough so as to not damage the imaging equipment.  As such, this sample was directly mounted to the aluminum SEM plate.  After images of the uncoated sample were taken, the sample was then sputter coated with a 5 nm layer of PtPd.   219  An Emitech cryostage was mounted on the FESEM unit to facilitate imaging of wet samples of MFT and PA-MFT.  This mode of imaging enabled visual investigation of the location of water in sample voids.  A fine spatula was used to fill a hollow void in a copper sample holder with PA-MFT that had been flocculated in August 2014.  The sample was stored in an air tight container since the time of flocculation and was still moist at the time of imaging in January 2015.  The flat side of a 3 mm diameter high pressure freezer (HPF) hat/disc was used to cover a portion of the sample that had been loaded into the sample holder. This enabled creation of a fresh surface when the hat was removed after the covered sample was plunge frozen in sub-cooled nitrogen in the slush chamber of an Emitech K1250x cryo-preparation unit pictured in Figure B. 6A.  Figure B. 6: Emitech Cryo-Preparation Unit (A) and Hitachi S4700 FESEM (B) The sample was transferred under vacuum into the main preparation station (also pictured in Figure B. 6A) where the HPF hat was knocked off using a fracture blade. The loaded frozen sample holder containing the sample with the freshly revealed surface was transferred under vacuum to the imaging chamber of the Emitech cryostage used with the Hitachi S4700 FESEM pictured in Figure B. 6B.  The cryostage was maintained at an average temperature of -130oC.  After images of the uncoated sample were obtained, the sample was transferred to the main prep station of the cryo-preparation unit where it was sublimed in two stages and sputter coated with Au for 1 minute using a current of 20 mA.  The sample was then replaced in the imaging chamber to enable collection of additional images.  Images obtained for  220  both the normal FESEM and cryo-SEM operating modes ranged between the 500 nm and 1 mm scales and investigated sample surfaces, edges, and pore spaces.    221  Appendix C:  C.1: Images Analyzed to Determine Surface Area Covered by Bitumen   222  C.2: Plates from FESEM and cry-SEM Imaging    223     224     225     226     227     228  C.3: Pore Pressure Responses to Load Phases of PA-MFT and raw AITF-MFT   229    230     

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.24.1-0228588/manifest

Comment

Related Items