Open Collections

UBC Undergraduate Research

Towards a testing protocol to assess CO2 sequestration potential in mine waste Carroll, Kate Jessica Apr 30, 2015

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

Item Metadata

Download

Media
52966-Carroll_Kate_EOSC_449_2015.pdf [ 1.76MB ]
Metadata
JSON: 52966-1.0053619.json
JSON-LD: 52966-1.0053619-ld.json
RDF/XML (Pretty): 52966-1.0053619-rdf.xml
RDF/JSON: 52966-1.0053619-rdf.json
Turtle: 52966-1.0053619-turtle.txt
N-Triples: 52966-1.0053619-rdf-ntriples.txt
Original Record: 52966-1.0053619-source.json
Full Text
52966-1.0053619-fulltext.txt
Citation
52966-1.0053619.ris

Full Text

   TOWARDS A TESTING PROTOCOL TO ASSESS CO2 SEQUESTRATION  POTENTIAL IN MINE WASTE   by  KATE JESSICA CARROLL      A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF  THE REQUIREMENTS FOR THE DEGREE OF   BACHELOR OF SCIENCE (HONOURS)  in   THE FACULTY OF SCIENCE  (Geological Sciences)           This thesis conforms to the required standard  ……………………………………… Dr. Greg Dipple   THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)  APRIL 2015    © Kate Jessica Carroll, 2015    ii  ABSTRACT Carbon sequestration via fixation of CO2 into carbonate minerals, a process known as carbon mineralization, has the potential to offset greenhouse gas emissions from mining operations through the precipitation of stable carbonates from dissolved ultramafic mine tailings. Under certain chemical conditions, dissolution rates of Mg-bearing tailings minerals control the overall rate of carbon mineralization. A testing protocol is proposed to assess the CO2 sequestration potential of mine tailings by investigating the capability to extract individual rate parameters of brucite and serpentine from dissolution of a brucite-serpentine mixture. Brucite and serpentine solids were well-characterized prior to experimentation and were selected to broadly mimic mineralogy and surface area of ultramafic tailings at the Mount Keith Nickel Mine in Western Australia. Two brucite, two serpentine, and one mixed brucite-serpentine dissolution experiment was carried out in a continuously-stirred flow-through reactor at circumneutral pH (~8) or acidic pH (~1.2) and with a residence time of 10 or 100 min. Mg concentration decayed from early, elevated transients reflecting both brucite and serpentine dissolution; short-lived plateaus in pH reflected only brucite behaviour; and long-term Si concentration reflected saturation of an amorphous silica phase. Mineral surface area adjustments in PHREEQC geochemical models were used to fit short-lived plateaus in pH and aqueous Mg concentration reflecting brief steady-state, and the data were used to constrain dissolution rates of brucite and serpentine, respectively. A long-lived steady-state regime could not be inferred for brucite or serpentine dissolution, yet rate parameters extracted from the mixed experiment agreed well with those of pure minerals and with Mount Keith tailings surface area data. A dichotomy exists between optimization of experimental design for rate extraction and simulation of real tailings chemistry for extrapolation to natural scenarios. Both options merit investigation, but the relative implications for assessment of tailings carbonation potential requires further research.    iii  TABLE OF CONTENTS ABSTRACT ........................................................................................................................ ii TABLE OF CONTENTS ..................................................................................................... iii LIST OF FIGURES.............................................................................................................v LIST OF TABLES ............................................................................................................ vi LIST OF SUPPLEMENTARY MATERIALS ................................................................ vii ACKNOWLEDGEMENTS ............................................................................................ viii 1. INTRODUCTION ......................................................................................................1 1.1. CO2 Sequestration via Carbon Mineralization ........................................................1 1.2. Carbon Mineralization in Ultramafic Mine Tailings ...............................................1 1.3. Mineral Dissolution Rate Controls on Carbon Mineralization ................................2 1.4. Thesis Objectives ..................................................................................................2 2. METHODS .................................................................................................................3 2.1. Sample collection and preparation .........................................................................3 2.2. Analytical Methods ...............................................................................................4 2.2.1. X-ray fluorescence spectroscopy................................................................4 2.2.2. X-ray powder diffraction ...........................................................................4 2.2.3. Surface area analysis and particle size distribution .....................................5 2.2.4. Scanning electron microscopy ...................................................................5 2.2.5. Analysis of water samples .........................................................................5 2.3. Experimental Methods ...........................................................................................5 2.4. Geochemical Modelling .........................................................................................7 3. RESULTS AND DISCUSSION .................................................................................9 3.1. Characterization of Solids ......................................................................................9 3.1.1. Chemical composition and mineralogy ......................................................9 3.1.2. BET surface area and particle size ........................................................... 10 3.1.3. Textural evolution of serpentine .............................................................. 11  iv  3.2. Preliminary Modelling Results ............................................................................ 12 3.3. Experimental Results and Discussion................................................................... 13 3.3.1. Brucite single-mineral experiments .......................................................... 15 3.3.2. Serpentine single-mineral experiments..................................................... 19 3.3.3. Serpentine and brucite multi-mineral experiment ..................................... 22 4. IMPLICATIONS...................................................................................................... 25 4.1. Extraction of Rate Parameters from Mineral Mixtures ......................................... 25 4.1.1. Rate extraction from a brucite-serpentine mixture .................................... 25 4.1.2. Extrapolation of single-mineral rate parameters ....................................... 26 4.1.3. Geochemical model predictions ............................................................... 27 4.1.4. Rate extraction using trace elements ........................................................ 28 4.2. Developing a Protocol to Assess Mine Tailings Dissolution................................. 28 4.2.1. Advantages of a flow-through time-resolved analysis technique .............. 28 4.2.2. Improvements and limitations to experimental design and geochemical modelling .......................................................................................................... 29 4.2.2.1. Optimization for rate extraction ................................................. 29 4.2.2.2. Simulation of real tailings ......................................................... 30 REFERENCES .................................................................................................................. 32 APPENDIX ........................................................................................................................ 36     v  LIST OF FIGURES Figure 1. Schematic of experimental apparatus. ....................................................................6 Figure 2. Particle size distributions of washed and unwashed solid samples. ....................... 10 Figure 3. A photograph and representative scanning electron micrographs of initial and reacted serpentine. ............................................................................................... 12 Figure 4. pH dependence of brucite and serpentine dissolution rate laws.. ........................... 14 Figure 5. Experimental and model data for Brucite-DI. ....................................................... 15 Figure 6.  Contour plot of Brucite-DI data weighted sums-of-squares analysis. ................... 17 Figure 7. Brucite-HCl [Mg] and pH data plotted over the range of brucite surface areas predicted by PHREEQC models. ......................................................................... 17 Figure 8. Experimental and model data for Brucite-HCl. .................................................... 18 Figure 9. Experimental and model data for Serp1. .............................................................. 19 Figure 10. Experimental and model data for Serp2. ............................................................. 19 Figure 11. Evolution of aqueous Mg/Si stoichiometry over time for all three experiments containing serpentine.. ......................................................................................... 22 Figure 12. Experimental and model data for Mixed............................................................. 23 Figure 13. Mixed [Mg] and pH data plotted over the range of brucite surface areas predicted by PHREEQC models.. ....................................................................................... 23 Figure A1. X-ray diffraction pattern and Rietveld refinement of initial brucite sample. ....... 40 Figure A2. X-ray diffraction pattern and Rietveld refinement of initial serpentine sample... 41 Figure A3. X-ray diffraction pattern for Serp2 reacted solids.. ............................................ 42 Figure A4. X-ray diffraction pattern for Mixed reacted solids. ............................................ 42     vi  LIST OF TABLES Table 1. Summary of experimental conditions. .....................................................................7 Table 2. Major oxide composition of solid samples. ..............................................................9 Table 3. Summary of preliminary PHREEQC modelling data. ............................................ 13 Table 4. Summary of PHREEQC model-fit data. ................................................................ 14 Table A1. Mineralogical data for initial solid samples. ........................................................ 36 Table A2. Summary of all PHREEQC modelling data for weighted sums of squares analysis of Brucite-DI experiment. ................................................................................. 37      vii  LIST OF SUPPLEMENTARY MATERIALS  Digital Media 1. Contains representative input and output files from PHREEQC models presented in this thesis. Files are in PDF format and are named according to experiment. ‘Preliminary’ folder contains preliminary investigative models. ‘Experiments’ folder contains experimental data fits. ‘Surface Area’ folder contains models used to investigate the relationship between surface area and both pH and Mg concentration. Surface areas are noted in the file names. The CD-ROM is attached to the inside back cover of this thesis.   viii  ACKNOWLEDGEMENTS Numerous thanks are due to those whose help and support have made this thesis possible. Firstly, I must thank my supervisor, Greg Dipple, for giving me a project and the support I needed to complete it. Your insight and feedback has helped me understand the big picture; the science behind my results. Ian Power and Anna Harrison have contributed immensely to the completion of this thesis by sharing their time, knowledge, and kindness. Ian: thank you for helping with ICP standards and SEM, and for your noted willingness to answer my questions, provide feedback on my work, teach me lab techniques, and talk about science in general. Anna: I must thank you for your help with modelling and dilutions, and for the inspiring quality of your own thesis. I cannot thank you both enough for your invaluable support over the past eight months.  A sincere thank you to Bart De Baere (and the Belgomatic) for facilitating my experiments and being a mine of knowledge on the practicalities of mineral dissolution experiments. Thanks are due to Maureen Soon, who ran some of my samples on the ICP-OES, and to Mariko Ikehata for doing ICP-MS on my samples at such short notice. Training and assistance with the BET instrument was generously afforded by Kristal Li. Appreciation for organization of the BSc. Honours Thesis course is justly due to Mary Lou Bevier. I am thankful to the SEG Canada Foundation for awarding me an undergraduate thesis scholarship and to the Department of Earth, Ocean and Atmospheric Sciences for providing some funding for my project.  I would like to thank my parents, John and Anthea, for always asking me about my thesis even though they don’t understand what I’m talking about; your hollow enthusiasm for minerals is heartwarming. And to my long-suffering roommate, Marie-Claire: thank you for getting me interested in research in the first place and for understanding and echoing my inclination to talk about lab things for hours on end. Finally, cheers and good luck to the other BSc. Honours students and to my friends in geology. Hard work is worthwhile if you have people to share and celebrate it with.    1. INTRODUCTION 1.1. CO2 Sequestration via Carbon Mineralization Motivation to mitigate anthropogenic CO2 emissions has emerged due to concerns that rising concentrations of greenhouse gases (GHGs) in Earth’s atmosphere are driving global climate change (IPCC, 2007). Long-term CO2 storage via carbon sequestration may partially offset emissions until alternative clean energy sources can replace current fossil fuels (Hoffert et al., 2002; Lackner, 2003). One such strategy - carbon mineralization - involves fixation of CO2 into carbonate minerals through the dissolution of alkaline earth silicate and hydroxide minerals in the presence of CO2 and subsequent precipitation of carbonates. By mimicking natural weathering processes (Kump et al., 2000; Wilson et al., 2006, 2009), active carbon mineralization capitalizes on the long-term stability of alkaline earth carbonate products (Lackner et al., 1995). Natural carbonation processes are sluggish, thus previous studies have investigated methods to accelerate carbonation rate, including the use of elevated CO2 partial pressure (pCO2), biological catalysts, or high pressures and temperatures (e.g., Krevor and Lackner, 2009, 2011; Power et al, 2011, 2013; Harrison et al., 2013).   1.2. Carbon Mineralization in Ultramafic Mine Tailings Ultramafic mine tailings are an advantageous feedstock for carbon sequestration because they contain abundant Mg-bearing minerals and have high reactive surface areas due to milling during ore processing (Wilson et al., 2009; Bobicki et al., 2012 and references therein). Exploitation of industrial waste materials for carbon mineralization capitalizes on coupling excess material generation with the need for long-term, environmentally benign storage. Wilson et al. (2014) estimate that carbonation of 10% of tailings at Mount Keith Nickel Mine (MKM) in Western Australia has the capacity to fully offset CO2 emissions from its mining operations. Mineralogy of MKM tailings includes major serpentine group minerals, comprising primarily antigorite [(Mg,Fe)3Si2O5(OH)4] and lizardite [Mg3Si2O5(OH)4], and minor brucite [Mg(OH)2] (Wilson et al., 2014). Serpentine and brucite  2  are common components of ultramafic mine tailings and offer significant sequestration potential due to their abundance and reactivity (Harrison et al., 2013), respectively.   1.3. Mineral Dissolution Rate Controls on Carbon Mineralization Mineral dissolution rates significantly influence carbonation progression, thus knowledge of mineral dissolution rates in a multi-mineral tailings sample enlightens its overall potential for carbon sequestration. Complexities of modelling tailings dissolution, including the inability to measure individual mineral surface areas within a mixture, are currently prohibitive. An empirical test for assessment of tailings dissolution is proposed here. Somewhat analogous to humidity cell testing (Sapsford et al, 2009) for assessing acid rock drainage at mine sites,  a flow-through time-resolved analysis (FT-TRA) technique (De Baere et al., 2013) for tailings dissolution could be used for assessment of mine tailings’ carbonation potential. Extensive research has been conducted into dissolution rate law extraction using flow-through reactors (e.g., Samson et al., 2000; Thom et al., 2013; Daval et al., 2013; Hariharan et al., 2014), primarily focusing on dissolution of well-characterized pure mineral samples at a range of pH.   1.4. Thesis Objectives The objective of this thesis is to design a protocol to assess the CO2 sequestration potential of mine tailings by investigating the capability to extract individual mineral specific rate parameters from dissolution of mineral mixtures. Specifically, extraction of rate parameters from dissolution of a brucite-serpentine mixture is attempted based on dissolution of well-characterized individual mineral samples. Flow-through dissolution experiments were conducted in conjunction with PHREEQC (Parkhurst and Appelo, 2013) geochemical modelling such that experimental data were used to constrain the model and enhance its predictive capabilities.  This is meant to be a preliminary study upon which a testing protocol for industry may be further developed. Ultimately, the optimization of experimental design, analytical techniques, and geochemical modelling may facilitate a test to quantify dissolution rates for complex mineral mixtures in natural tailings samples. Implementation of such a test would allow mine engineers to make scientifically informed decisions about the efficacy of  3  integrating accelerated carbon mineralization techniques at their mine in order to reduce or offset CO2 emissions. 2. METHODS 2.1. Sample collection and preparation Samples of serpentine, brucite and MKM tailings were acquired and processed prior to characterization. The serpentine was obtained from a serpentinite float sample collected at the active Swift Creek landslide in northwest Washington, U.S.A. in August 2014. The rock sample was crushed with a hammer, pulverized in a ring mill for 2 min and then sieved to ≤106 μm in diameter. The finest particles were removed by agitation in an ethanol suspension and subsequent decanting of the cloudy supernatant; this process was repeated 4-5 times on several ~2 g aliquots. A final rinse was done with deionized water and the sample was dried at room temperature.  Brucite obtained from Minerals Unlimited by Power et al. (2013) was pulverized with a ring mill for 1 minute. The finest particles were removed using the same washing method as the serpentine. Tailings sample 06MK30-7 collected by Wilson et al. (2014) by backhoe at MKM in 2006 was used here for analysis. The tailings were washed in deionized water to remove halite that had precipitated from hypersaline ([Na+] = 0.78 mol/L; Wilson et al., 2014) pore fluids at MKM. Aliquots of tailings were suspended in deionized water in 50 mL tubes and centrifuged for 1 minute. The electrical conductivity of the supernatant combined from all tubes was measured. This process was repeated until consecutive conductivity measurements were within 100 μS of one another, and less than 250 μS. Brucite, serpentine and MKM tailings samples were characterized for their chemical compositions, mineralogy, reactive surface areas and particle size distributions prior to experimentation. Although the MKM tailings were not used in the experiments, their characteristics served as a target for the pure mineral samples to simulate tailings.  Final solids collected after the serpentine and mixed experiments were analyzed by scanning electron microscopy (SEM) and qualitative X-ray diffraction (XRD). Additionally,  4  water samples collected periodically for the duration of the experiments were analyzed with the objective of tracking mineral dissolution through time.  2.2. Analytical Methods 2.2.1. X-ray fluorescence spectroscopy Major oxide geochemistry of the brucite, serpentine and MKM tailings samples was determined by X-ray fluorescence (XRF) spectroscopy at ALS Global Laboratories, North Vancouver, British Columbia. The samples were dissolved by lithium borate fusion in preparation for analysis.  2.2.2. X-ray powder diffraction  The solid phases present in each of the initial samples were identified and quantified using the Rietveld method for X-ray diffraction (XRD) (Rietveld, 1969; Bish and Howard, 1988). An aliquot of each sample with a 10 wt.% annealed fluorite (CaF2) spike was ground under anhydrous ethanol for 7 minutes in a McCrone® micronizing mill with agate grinding elements in order to reduce particle size and ensure homogenization. Samples were dried overnight at room temperature then lightly disaggregated using a corundum mortar and pestle. Back-loading cavity powder mounts were prepared against rough sandpaper to minimize preferred crystallographic orientation. Measurements were collected using a Bruker D8 Focus Bragg-Brentano diffractometer with LynxEye detector. A long, fine focus cobalt X-ray tube was operated with CoKα radiation at 35 kV and 40 mA. Data were collected with a 0.03°2θ step size and a counting time of 7 s/step over a range of 3°-80°2θ. Search/match qualitative phase identification was completed with DIFFRACplus Eva 14 software (Bruker AXS, 2008) using the International Centre for Diffraction Data PDF-4+ 2010 database. Topas Version 3 software (Bruker AXS, 2004) was used for quantitative Rietveld refinement of identified phases. Mineral content of solid material remnants from dissolution experiments were identified using qualitative XRD. An aliquot of each sample was finely ground under anhydrous ethanol in a corundum mortar and pestle then smear-mounted onto a round glass cover slide and allowed to dry at room temperature. Qualitative XRD data was collected  5  under the same operating conditions used for Rietveld XRD, except that a rotation of 60 rpm was applied. Rietveld refinement was not performed.  2.2.3. Surface area analysis and particle size distribution Surface areas of the samples were determined by multipoint BET with N2 adsorption using a Quantochrome Autosorb-1 surface area analyzer. A Malvern Mastersizer 2000 Laser Diffraction Particle Size Analyzer was used to determine particle size distributions for each of the samples and their unwashed precursors (for comparison).   2.2.4. Scanning electron microscopy Scanning electron microscopy (SEM) of initial serpentine and final reaction solid samples was performed at the Centre for High-Throughput Phenogenomics at The University of British Columbia.  Samples were coated with 7 nm of iridium using a Leica EM MED020 coating system and imaged using a FEI Helios NanoLab 650 operating at 1.0 kV voltage.  2.2.5. Analysis of water samples Water samples were collected regularly during the experiments and analyzed for Mg and Si concentrations ([Mg] and [Si], respectively). All samples, standards and blanks were acidified to a 2% ultrapure nitric acid (HNO3) matrix prior to analysis. The circumneutral-pH brucite experiment (Brucite-DI) samples were analyzed by inductively coupled plasma optical emission spectrometry (ICP-OES) using a Varian 725-ES Optical Emission Spectrometer. Samples from all other experiments were analyzed by inductively coupled plasma mass spectrometry (ICP-MS) using a Perkin Elmer NexION® 300D Mass Spectrometer. Standards were prepared in the range of 1 to 100 ppm for ICP-OES and 10 to 1,000 ppb for ICP-MS using a 1,000 ppm multi-element standard.  2.3. Experimental Methods Five flow-through mineral dissolution experiments were carried out at room temperature using a cylindrical 50 mL Teflon® chamber containing no head space and 500 mg of powdered sample that was continuously mixed using a magnetic stir bar. Eluent solutions of either deionized water (pH ≈ 5) or 0.1 N HCl (pH ≈ 1), both with a 0.001 M  6  NaCl background electrolyte, were prepared in 2 L plastic bottles open to the atmosphere. Fluid was pumped from the eluent bottle at a constant flow rate of 5 mL min-1 or 0.5 mL min-1 through plastic tubing entering into the bottom of the chamber using a DIONEX ICS-5000 DP gradient pump. The reacted fluid then exited the chamber through the top lid into the effluent line. An in-line pH sensor took approximately one measurement every 5 s and an ensuing 0.2 μm syringe filter (preceded by a 0.45 μm filter in Serp2 and Mixed experiments) removed particulate matter. Effluent samples of 8 to 13 mL volumes were collected by a Foxy®R2 fraction collector in plastic tubes at intervals of 10 to 32 minutes depending on flow rate. Tubes were capped within 4 hours, on average, to minimize evaporation. Experiments were conducted for a duration of 4, 20 or 40 hours (Fig. 1). Parameters for each experiment are shown in Table 1.       Figure 1. Schematic of experimental apparatus.    7  Table 1. Summary of experimental conditions. Experiment name Mineral reactant(s) Duration (h) Flow rate (mL min-1) Residence time (min) Sample mass (mg) Eluent pH Brucite-DI brucite   4 5.0   10 500.30 4.9 Brucite-HCl brucite 20 0.5 100 500.76 1.3 Serp1 lizardite 20 0.5 100 500.49 1.3 Serp2 lizardite 40 0.5 100 500.35 1.2          Mixed brucite 40 0.5 100   50.11 1.1 lizardite 450.43   The experiments were subject to numerous problems affecting the data, including filter clogging, pH data loss, and inhomogeneous mixing. Fines from the unwashed serpentine used in Serp1 clogged the eluent filter, dramatically reducing flow rate. A second filter also clogged within the first 2 h, so the filter was removed entirely, allowing fine particles to remain in the fluid and continuing dissolving in sample tubes until manual filtering up to 24 h later. These incidents had an unknown effect on the data, so a second serpentine experiment (Serp2) was conducted. During Serp2, pH was monitored but the data could not be recovered after the experiment. Inhomogeneous mixing within the reaction chamber is interpreted to be the source of non-systematic scattering of [Mg] and [Si] data points, particularly that observed in the first 20 h of Serp2. Though the reaction chamber is opaque, occasional settling of larger particles and inconsistent stirring are consistent with mixing tests carried out without a lid and with a clear view into the reactor.  2.4. Geochemical Modelling Geochemical modelling of kinetic mineral dissolution was conducted using PHREEQC V.3 (Parkhurst and Appelo, 2013) with the LLNL database. Models simulated a 1 L reaction chamber into which discrete volumes of solution reacted at 25°C with the chosen solid mineral phase(s) for one residence time each, never mixing with one another. Each simulation lasted 24 residence times, so duration depended on residence time. Solutions were  8  either pure water or dilute HCl, each on a background 0.001 M NaCl electrolyte, and were equilibrated with atmospheric CO2. Models were constructed using the following pH-dependent kinetic dissolution rate laws for brucite (Pokrovsky and Schott, 2004) and lizardite (Daval et al., 2013):  0.2 20[ ] (1 )brc brcr SA k H   0.530[ ] (1 )liz lizr SA k H     where SA is the total reactive surface area (m2), k0 is the rate constant (mol m-2 s-1), and Ω is the saturation ratio. The rate constants derived from Pokrovsky and Schott (2004) and Daval et al. (2013) are, respectively:  6.380 10brck   mol m-2 s-1 9.630 10lizk   mol m-2 s-1  Total reactive surface area was based on a reactor volume of 1 L, with the same mineral mass to fluid volume ratio as used in the experiments. PHREEQC modelling was employed toward two distinct purposes: 1) preliminary modelling prior to the experiments to scope out experimental parameter space and determine suitable experimental design (referred to as investigative experiments) and 2) post-experiment model fitting to experimental data (referred to as retroactive models). PHREEQC simulations were used to design experiments where samples contained Mg and Si concentrations in the range of 1 to 100 ppm. The pCO2 and/or total surface area were used to fit experimental pH or [Mg] and [Si] data. Representative input and output files for PHREEQC models are given in Digital Media 1.  9  3. RESULTS AND DISCUSSION 3.1. Characterization of Solids 3.1.1. Chemical composition and mineralogy Results of bulk chemical analysis and mineralogy of solid brucite, serpentine and MKM tailings are presented in Table 2 and A1, respectively. ‘Single-mineral’ samples are not pure: the brucite sample contains minor dolomite [CaMgCO3] and calcite [CaCO3]; the lizardite sample contains minor magnetite [Fe3O4]. Evidence for these phases is present in XRD patterns (Figs. A1, A2) and Ca and Fe oxide data (Table 2). A portion of amorphous solids in the serpentine sample likely represent additional, poorly crystalline serpentine (Wilson et al., 2006). MKM tailings contain 81.6 wt.% serpentine (comprising antigorite and lizardite) and 3.4 wt.% brucite with minor or trace carbonate, oxide, halide and hydrotalcite-group minerals (Wilson et al., 2014).     Table 2. Major oxide composition of solid samples.     Brucitea Serpentine MKM tailings Al2O3        0.14                   0.36                   0.19            BaO      <0.01              0.01                   0.01            CaO        2.50                   0.05                   0.19            Cr2O3      <0.01              0.28                   0.21            Fe2O3        0.40                   7.72                   6.64            K2O      <0.01            <0.01            <0.01       MgO      61.55                 38.80                 41.00            MnO        0.02                   0.09                   0.08            Na2O      <0.01              0.15                   0.16            P2O5        0.01                 <0.01              0.01            SO3      <0.01              0.18                   0.65            SiO2        1.74                 39.40                 34.20            SrO      <0.01            <0.01            <0.01       TiO2        0.01                 <0.01              0.01            Total      98.83               100.25               100.60            LOI 1000 b      32.47                 12.89                 16.84            a       Data from Power et al. (2013).   b       Indicates loss on ignition at 1000°C for one hour.  10  Qualitative XRD of final reaction solids indicates the complete loss of brucite in the Mixed experiment and does not show evidence for precipitation of new crystalline phases in any experiment. Solids were completely exhausted in the Brucite-HCl experiment, so there were no final solids to analyze. XRD patterns for final reaction solids are given in the appendix (Figs. A3 and A4).   3.1.2. BET surface area and particle size BET reactive surface area measurements for washed brucite, serpentine and MKM tailings were 6.6 m2 g-1, 13.5 m2 g-1, 9.8 m2 g-1, respectively. Washing the serpentine sample reduced the BET measurement from 18.5 m2 g-1 to 13.5 m2 g-1, expressing the loss of very fine particles with high surface area. The effect of washing on particle size distribution is captured in Figure 2. Sieving and washing of serpentine, in particular, restricted the range of particle sizes around the median size of 56 μm. Washing of MKM tailings was meant to removed halite, not fines, so the particle size distribution was relatively unaffected by washing and had a median value of 10 μm for washed tailings. Slightly to clearly bimodal Figure 2. Particle size distributions of washed and unwashed solid samples. MKM tailings were washed to remove halite, not fines.   11  distributions (unwashed and washed samples, respectively) for brucite are interpreted to be the result of aggregation of smaller particles. Dissolution is likely to disrupt these aggregates, but it remains possible that aggregation eliminates some reactive surface area. Disregarding aggregates, the median particle size of washed brucite is 5 μm.  3.1.3. Textural evolution of serpentine  Complete consumption of brucite in most experiments precluded the comparison of brucite surface textures before and after dissolution. Representative scanning electron micrographs of serpentine initial and final reaction solids are presented in Figure 3 alongside a photograph of the hand specimen prior to milling (Fig. 3A). The massive texture observed in hand specimen is not correlative to the far more fibrous microscopic texture. Implications of a fibrous micron-scale serpentine texture are twofold: 1) measured BET reactive surface area may be an underestimate and 2) a geometric model is not sufficient for modelling evolution of surface area during reaction progression. Fibrous aggregates may have internal surface area that, due to the size of the gas adsorbate, is excluded from BET measurement but is available for reaction in solution. True reactive surface area of a fibrous mineral cannot be represented by a simple geometric model, which represents particles as a uniform population of perfect spheres (Mayer et al., 2002). This has implications for the development of a dissolution-dependent surface area update in geochemical models to account for evolution of surface area during reaction progression. A difference in texture is observed between micrographs of initial serpentine (Fig. 3B) and final solids (Figs. 3C and 3D), though there is no conclusive evidence supporting or rejecting the precipitation of an amorphous phase. Reacted solids appear more ragged and fragmental than initial solids. Fragmentation of fibres is interpreted as surface area evolution associated with the accessing of surface area internal to fibrous bundles and not accounted for by the BET measurement on precursor materials. The result is an increased reactive surface area that must be accounted for in geochemical modelling.   12   Figure 3. A photograph and representative scanning electron micrographs of serpentine showing macroscopic and microscopic texutres. A) Photograph of serpentinite hand specimen before milling. Rulings represent 1 cm. B) Initial washed serpentine solids. C) Final reaction solids of Serp2. D) Final reaction solids of Mixed.  3.2. Preliminary Modelling Results Preliminary PHREEQC (Parkhurst and Appelo, 2013) modelling was performed using the range of possible experimental and analytical capabilities to plan experiments. Results of simulations upon which experiments were designed are given in Table 3. Flow rates of between 1 μL min -1 and 10 mL min-1 limited the acceptable residence times to between ~35 days and 5 minutes, respectively. A practical upper limit of 10 g powdered sample per litre was used to limit settling of particles. Addition of HCl to the eluent controlled pH. Experiments were designed to have [Mg] and [Si] in the calibration range of the analytical instrumentation. All input surface areas were based on BET specific surface measurements except for Brucite-HCl. Preliminary models for brucite-HCl and Mixed experiments were based on an adjusted value for brucite surface area of 0.17 m2 g-1 based on  13  experimental modelling of the Brucite-DI experiment. All experimental simulations reached steady-state; pH, [Mg], and [Si] reported in Table 3 are steady-state values.  3.3. Experimental Results and Discussion Progress of the experiments was tracked by measurements of pH and Mg, and Si concentration for experiments containing serpentine. A summary of all model-fit data are given in Table 4. Recognition of transient versus steady-state dissolution in the data was based on time-dependence of concentration, Mg/Si molar ratio, and the adjustments to surface area required for model fitting. Temporally evolving concentrations, high Mg/Si ratios and large adjustments to modelled surface areas generally indicated non-steady-state dissolution. Dissolution rate laws for brucite (Pokrovsky and Schott, 2004) and lizardite (Daval et al., 2013) (used in PHREEQC modelling), and chrysotile (Thom et al., 2013) (another serpentine-group mineral) are plotted over the stated pH range for which they apply (Fig. 4). Note that all low-pH experiments in this thesis were conducted outside of the relevant pH range. Relevance of the literature rate laws to these experiments is informed by model adjustments to fit data.  Table 3. Summary of preliminary PHREEQC modelling data. Experiment name Mineral reactant(s) Specific surface area (m2 g-1) Residence time (min) pHSS a [Mg]SS (ppm) [Si]SS (ppm) Brucite-DI b, c brucite   6.6   10 10.20 2.1  Brucite-HCl brucite     0.17 100   1.12 62.46  Serp1 lizardite 18.5 100   1.09   5.01 3.88 Serp2 lizardite 13.5 100   1.09   3.65 2.81          Mixed d brucite     0.17 100   1.09   9.36 2.53 lizardite  13.5 a       SS indicates steady-state values. b       Eluent was pure water; all other simulations had a 0.1 N HCl eluent solution. c       Simulation duration was 4 h; all other simulations lasted 40 h. d       Mass ratio of brucite to serpentine was 1:9.    14   Table 4. Summary of PHREEQC model-fit data. Experiment name Mineral reactant(s) Residence time (min) Surface area (m2 g-1) pHSS a [Mg]SS (ppm) [Si]SS (ppm) Brucite-DI brucite   10       0.17 9.78          0.38  Brucite-HCl brucite 100     2.8 1.71 920  Serp1 lizardite 100 130.0 1.10     35.0 4.9 Serp2 lizardite 100 111.0 1.10     29.9 4.9          Mixed brucite 100     6.6 1.10     29.1 4.9 lizardite 120.0 a         SS indicates steady-state values.    Figure 4. pH-dependence of  dissolution rate laws. Experimentally-determined rate laws for brucite, chrysotile and lizardite were taken from Pokrovsky and Schott (2004), Thom et al. (2013), and Daval et al. (2013), respectively. Ranges of experimental pH are shaded.  15  3.3.1. Brucite single-mineral experiments Brucite dissolution experiments illustrate characteristic features of mineral dissolution processes in continuous flow reactions.  Initial transient peaks in pH and [Mg], which are conventionally interpreted to reflect the dissolution of mineral fines and the adjustment of the mineral surface reaction sites (e.g., Thom et al., 2013), are evident in the Brucite-DI experiment (Fig. 5). Following this transient, pH and [Mg] decay to relatively constant values that can be taken to represent a steady-state brucite dissolution regime from which bulk mineral dissolution rates can be inferred. Brucite dissolution at circumneutral pH reached steady-state after approximately 2 h and significantly buffered pH, stabilizing at 8.6 after an initial pH of 4.9. In the Brucite-HCl experiment, low pH and a tenfold increase in residence time caused rapid dissolution and exhaustion of brucite after approximately 13 h (Fig. 6). The rapid consumption of brucite leads to long term declines in both pH and [Mg] attributable to declining bulk surface area during extensive brucite dissolution; a steady-state regime is difficult to infer.  In Brucite-DI, initial pH prior to the introduction of solid brucite to the chamber (i.e., pH of the eluent) was 4.9, whereas models predicted a pH of 5.5 for water in equilibrium with laboratory air (600 ppm CO2). Addition of brucite buffered the pH up to 9.3 in the first five minutes and then slowly declined to a steady-state pH of approximately 8.6 after 2 h. Although time-resolution for tracking transient [Mg] early in the experiment is fairly poor, the data clearly show an early peak of approximately 1.25 ppm Mg in the first five minutes (Fig. 5). A steady-state Mg concentration of approximately 0.43 ppm was reached after 2.5 h.  PHREEQC reactive transport modelling was used to fit the steady-state pH of 8.6 and the steady-state Mg concentration of 0.43 ppm by adjusting brucite surface area and CO2 partial pressure (pCO2), which served as a proxy for pH in these less acidic conditions. Simultaneous fitting of both steady-state pH and initial (i.e., eluent) pH was not possible. A weighted sums of squares analysis was applied to the model data in order to determine the surface area and pCO2 combinations that generated both [Mg] and pH within tolerance (Fig.6). Measurement error on pH was estimated to be ± 0.20 based on discrepancies between two different pH sensors measuring the eluent solution. Error on [Mg] based on standard deviation of hidden ICP-OES standards was estimated to be ± 0.22 ppm. Weighting 16    Figure 5. Experimental and PHREEQC model-fit [Mg] and pH data from Brucite-DI experiment. Eluent was deionized water and residence time was 10 min. Model surface area and pCO2 was adjusted to fit [Mg] and pH plateaus within measurement error.  on pH was 462 times higher than on [Mg] due to the difference in absolute sums of squares, reflecting the superior abundance and time-resolution of pH measurements. All retroactive modelling results for this sums of squares analysis are given in Table A2. An acceptable range of pCO2 for lab air in equilibrium with the eluent is 500 to 700 ppm (Yanes and Yapp, 2010; You et al., 2012). Within this range, [Mg] and pH were matched within tolerance by reducing the specific surface area of brucite to 1/40 of the BET measurement, a finding consistent with modelling of brucite carbonation reactions performed by Harrison et al. (2015). BET surface area measurements often overestimate the reactive surface area by including internal and unreactive surfaces (Brantley and Mellot, 2000; Gautier et al., 2001), justifying the reduction of surface area in the model. The final accepted model-fit parameters for Brucite-DI were SAbrc = 0.17 m2 g-1 and pCO2 = 631 ppm (Fig. 6).  17   Figure 6.  Contour plot of weighted sums-of-squares analysis of Brucite-DI data using pCO2 and brucite surface area to fit plateaus in experimental pH and [Mg]. SSq indicates sums-of-squares values.  Figure 7. Brucite-HCl [Mg] and pH data plotted over the range of brucite surface areas predicted by PHREEQC models.  18   The Brucite-HCl experiment (Fig. 8) exhibited a large [Mg] peak at between 800 and 1000 ppm and a slight pH peak at 1.68 within 1 h, followed by decline of [Mg] after approximately 13 h, indicating exhaustion of brucite. Brucite dissolution did not reach steady-state, so PHREEQC modelling was performed to fit transient [Mg] and pH. Eluent containing 0.1 N HCl strongly constrained reaction pH, and therefore model fitting was accomplished by adjusting brucite surface area only. Reduction of specific surface area to 2.8 m2 g-1, ~ 58% of BET measurement, resulted in a 920 ppm transient [Mg] and a pH of 1.71, both within the estimated measurement error (Table 4, Fig. 7).    Figure 8. Experimental and PHREEQC model-fit [Mg] and pH data from Brucite-HCl experiment. Eluent was 0.1N HCl and residence time was 100 min. Model surface area was adjusted to fit the short-lived pH plateau and peak in [Mg].    19  3.3.2. Serpentine single-mineral experiments Serpentine dissolution experiments presented a number of challenges. Two experiments (Serp1 and Serp2) were conducted to overcome these difficulties. Experimental and PHREEQC modelling results of each are presented in Figures 9 and 10. Note that pH data is not available for Serp2, but that serpentine dissolution results in a relatively small fluctuation in pH during Serp1. Serp1 was conducted using unwashed solids with a slightly higher BET surface area and lasted only 20 h, half as long as Serp2. Several anomalies in the data must be accounted for prior to interpretation. A spike in both [Mg] and [Si] in the Serp1 prior to 2 h coincided with clogging of the effluent filter by fine particles, causing a severe decrease in flow rate (i.e., increase in fluid residence time). The decreased flow rate was not reflected in pH data. The filter was removed and particles were able to exit the chamber for the remainder of the experiment, decreasing total serpentine surface area within the reactor  Figure 9. Experimental and model-fit [Mg], [Si], and pH data for Serp1. Eluent was 0.1 N HCl and residence time was 100 min. Adjustment of serpentine surface area was used to fit relatively constant late [Mg]. [Si] is consistent with precipitation of an amorphous silica phase akin to chalcedony. Figure 10. Experimental and model-fit [Mg], [Si], and pH data for Serp2. Eluent was 0.1 N HCl and residence time was 100 min. Adjustment of serpentine surface area was used to fit a short-lived [Mg] plateau. [Si] is consistent with precipitation of an amorphous silica phase akin to chalcedony. Note that experimental pH was not recovered for this experiment.  20   and permitting fine particles to continue dissolving in the sample tubes until manual filtering up to 24 h later. The extent to which these incidents affected the data is not known. In the first 20 hours of Serp2, [Mg] and [Si] show obvious trends with scattered anomalously low data points. These data are thought to be the result of temporarily poor mixing within in the chamber that effectively reduced the total surface area for reaction, thus decreasing the rate of dissolution. Data from neither serpentine experiment conclusively indicate occurrence of steady-state dissolution.  [Mg] in Serp1 is significantly higher than in Serp2 due to the use of unwashed serpentine with more fine particles and thus more surface area. In both cases, [Si] stabilized between 3 and 5 ppm within 10 h while [Mg] declined asymptotically, but did not demonstrably stabilize within the experimental time frame. The initial transient in [Mg] lasts for 10’s of hours and evolves as a series of steps that could represent a series of steady-state Figure 10. Experimental and model-fit [Mg], [Si], and pH data for Serp2. Eluent was 0.1 N HCl and residence time was 100 min. Adjustment of serpentine surface area was used to fit relatively constant late [Mg]. [Si] is consistent with precipitation of an amorphous silica phase akin to chalcedony. Not that experimental pH data was not recovered.  21  regimes, changes in reaction processes due to incomplete mixing within the reactor, or other unidentified processes. Data collected from 20 to 40 h in Serp2 show a further decrease in [Mg] to approximately 30 ppm, but one sample taken at 48 h is further depleted in aqueous Mg, casting doubt on the attainment of steady-state anytime within the 40-hour experiment. Chrysotile dissolution experiments conducted at low pH by Thom et al. (2013) reached steady-state within ten residence times. In comparison, Serp2 lasted 24 residence times without reaching steady-state. This suggests that while steady-state is achievable, experimental design must be optimized in order to achieve it. Residence times used by Thom et al. (2013) were between 16 and 24 hours, more than 10 times longer than the 100 min used in Serp1 and Serp2. Modeled serpentine dissolution establishes steady-state conditions which can be used to infer reaction rates during short-lived plateaus in [Mg] during dissolution experiments. Total serpentine surface area was adjusted in order to match the [Mg] plateau at 35 ppm in Serp1 and 30 ppm in Serp2. This fit required specific surface areas of 130 and 111 m2 g-1, respectively; a seven- to eightfold increase of the BET measurements (Table 4).  Species concentrations from both experiments revealed strongly non-stoichiometric dissolution. Experimental Mg/Si molar ratios between 7 and 27 contrast sharply with the theoretical Mg/Si molar ratio of 1.5 for stoichiometric lizardite dissolution (Fig. 11). Though non-stoichiometric dissolution does not preclude steady-state dissolution, especially at low pH (Thom et al., 2013), the combination of variable [Mg] and high Mg/Si ratios justify rejection of a steady-state interpretation for the serpentine experiments.  Allowing precipitation of chalcedony [SiO2] in the models dropped the aqueous concentration of Si from over 20 ppm to 5 ppm, within 2 ppm of the experimental observations. This suggests that precipitation of amorphous silica may, in part, account for the high Mg/Si molar ratio observed in both serpentine experiments by consuming aqueous Si.  22   Figure 11. Evolution of aqueous Mg/Si stoichiometry over time for all three experiments containing serpentine. The ideal mineral stoichiometric ratio of 1.5 is plotted for comparison.  3.3.3. Serpentine and brucite multi-mineral experiment Initial solids of the Mixed experiment contained a 9:1 mass ratio of washed serpentine to brucite. Experimental and modelling data from the Mixed experiment are given in Figure 12. Based on the Brucite-HCl experiment, it was anticipated that brucite would be exhausted in the Mixed experiment. An early peak in [Mg] and short-lived (~1 h) plateau in pH were concurrent in the data. A decline in pH was observed between 1 h and 10 h, followed by fluctuation around a constant value of approximately 1.1. Between 10 h and 40 h, [Mg] declined steadily from 50 ppm to 30 ppm.  A fit of the short-lived [Mg] plateau reflecting dissolution and rapid consumption of brucite and the subsequent plateau reflecting ‘steady-state’ serpentine dissolution was accomplished using PHREEQC. Total surface area of brucite and serpentine were adjusted to fit these Mg concentrations. Modelling parameters and results are listed in Table 4. BET   23   Figure 13. Experimental and model-fit [Mg], [Si], and pH data from the Mixed experiment. Eluent was 0.1 N HCl and residence time was 100 min. The serpentine to brucite mass ratio was 9:1. Adjustment of serpentine surface area was used to fit relatively constant late [Mg] and brucite surface area was used to fit the [Mg] peak. [Si] is consistent with precipitation of an amorphous silica phase akin to chalcedony. Figure 12. Mixed [Mg] and pH data plotted over the range of brucite surface areas predicted by PHREEQC models.  24  surface area was used for brucite and nearly nine times BET surface area was used for serpentine. Modelled steady-state pH was 1.10, within measurement uncertainty of the ‘steady-state’ experimental pH. An early peak in model pH coincided with early dissolution and consumption of brucite. As with the ‘pure’ serpentine experiments, precipitation of chalcedony in the model decreased [Si], improving the fit between model and data. Given the ambiguity of [Mg] data, distinguishing brucite dissolution from serpentine dissolution in a mixed system may require careful consideration of pH, [Si], and/or trace element data. The early, short-lived plateau of 1.22 in the pH data and subsequent steady decline to around 1.13 is consistent with early dissolution and exhaustion of brucite (Fig. 13). After 10 h, the pH data show no obvious trends, indicating that brucite dissolution does control pH whereas serpentine dissolution does not, a finding consistent with results of the ‘pure’ brucite and serpentine experiments. Dissolution of fines during clogging in Serp1 (~2 h) (Fig. 9) lead to nearly doubling [Mg] while pH remained unaffected, supporting the usage of pH data to infer brucite behaviour. Fluctuations in pH around a value consistent with serpentine dissolution in the Mixed experiment is attributable to inhomogeneous mixing within the reaction chamber. Use of [Si] to track serpentine dissolution in a brucite-serpentine mixture eliminates the issue of confounding data from multiple sources, but as with [Mg], apparently non-stoichiometric dissolution complicates the relationship between elemental flux and serpentine molar flux. Even late in the experiment, after exhaustion of brucite, the Mg/Si molar ratio remained strongly non-stoichiometric (Fig.11). Near the termination of the experiment, the Mg/Si ratio fluctuated between approximately 15 and 20, over ten times greater than the mineral stoichiometric ratio. This must be taken into account if [Si] is to be used as a proxy for serpentine dissolution. Thom et al. (2013) achieved stoichiometric dissolution when performing experiments at a pH of approximately 4, so decreasing the acidity of experiments may be advantageous. Precipitation of an amorphous silica phase akin to chalcedony may be controlling [Si], decoupling aqueous Mg and Si concentrations. Experimental pH and flow rate must be selected such that chalcedony is well below saturation to eliminate precipitation effects.  25  4. IMPLICATIONS 4.1. Extraction of Rate Parameters from Mineral Mixtures Empirical extraction of dissolution rate parameters requires attainment of far-from-equilibrium steady-state dissolution. Rate laws from multi-mineral mixtures comprising phases of various solubilities can be attained if all minerals reach steady-state in the given experimental conditions and if dissolution of individual minerals can be distinguished in the data. Substantial loss of reactive surface area as a result of mineral consumption during reaction progression hinders attainment of steady-state. Cation concentrations are not always a useful measure of dissolution rate due to transient conditions and confounding of concentrations from mixtures of minerals containing the same metal cations. Despite difficulties arising from experimental design, rate parameters derived from single-mineral systems show a robust correlation with the mixed system. Furthermore, predictive modelling of MKM tailings dissolution using model-fit parameters from the Mixed experiment yields results broadly consistent with MKM surface area data.  4.1.1. Rate extraction from a brucite-serpentine mixture The brucite-serpentine mixture studied in this thesis has important implications for tailings dissolution. Based ostensibly on the composition of MKM tailings (Wilson et al., 2014), serpentine was selected due to its natural abundance in ultramafic tailings and brucite, though commonly a minor tailings component, was chosen due to its high reactivity and potential for carbon sequestration (Harrison et al., 2013). As such, dissolution of these minerals may strongly control the carbonation potential of more complex mixtures in ultramafic tailings.  In a brucite-serpentine system, meaningful rate parameters for brucite dissolution can be extracted from short-lived plateaus in pH. Dissolution of brucite decreases acidity of the reaction fluid while serpentine does not have a substantial effect on pH in a 0.1 N HCl fluid and with relatively short residence times (~100 min). Even dissolution of high-surface area serpentine fines in unwashed powders has a negligible effect on pH (Fig. 9). Steady-state brucite dissolution is characterized by a plateau in pH preceded by an initial transient spike. To ensure that the plateau is not due to exhaustion of brucite, the plateau pH must be more  26  alkaline than the eluent pH and detailed mineralogy of the final solids must confirm the presence of brucite. If the rate of brucite dissolution can be modelled to fit pH, PHREEQC modelling can then be used to infer the contribution of serpentine dissolution to total [Mg]. Both brucite and serpentine affect [Mg], which decays from elevated transient concentrations over time. Modelling of [Mg] was complicated by lack of steady-state behaviour in the experiment, but long-term [Mg] values generally reflect serpentine dissolution and can be used for extraction of rate parameters fairly consistent with serpentine-only experiments. Caution must be employed if using [Si] as a proxy for serpentine dissolution. Non-stoichiometric dissolution of serpentine during transient regimes (Thom et al., 2013) means the surface flux of Si is not stoichiometrically related to the serpentine molar flux. If steady-state coincides with saturation of an amorphous silica phase (e.g., chalcedony), [Si] will represent the balance between dissolution of serpentine and precipitation of silica. To avoid this problem, chalcedony must be below saturation during steady-state at the given experimental conditions. Chalcedony’s saturation index can be calculated using PHREEQC modelling. High [Mg] and low [Si] observed in the experimental data are consistent with pore fluid chemistry at MKM (Stolberg, 2005) and Clinton Creek asbestos mine, Yukon (Power et al., 2010), suggesting that [Si] in pore fluids of ultramafic tailings is controlled by solubility of an amorphous silica phase. Deliberate experimental precipitation of silica may be advantageous if simulation of true tailings chemistry is a greater priority than rate extraction.  4.1.2. Extrapolation of single-mineral rate parameters Geochemical modelling of mineral dissolution was achieved by adjusting reactive surface area in order to fit experimental dissolution rates. Comparison of modelled surface areas from single-mineral experiments with those from the Mixed experiment is therefore a justified method of assessing the validity of using rate parameters derived from individual brucite and serpentine experiments to infer behaviour of their counterparts in the mixture. Modelled surface areas from Brucite-HCl, Serp1 and Serp2 and their respective Mixed model values are broadly consistent with one another and are within acceptable variation of reactive surface area (Table 4).  27  Brucite surface area was increased from 2.8 m2 g-1 in Brucite-HCl to 6.6 m2 g-1 in the Mixed experiment. A factor of 2.5 is trivial considering experimentally determined brucite dissolution rates can vary by up to two orders of magnitude (Pokrovsky and Schott, 2004) and adjustments to brucite BET surface area by a factor of 20 are common for model-fitting (Harrison et al., 2015). Serpentine surface area in the Mixed model lies between that of Serp1 and Serp2 (Table 4), suggesting that the rate of serpentine dissolution extracted from a brucite-serpentine mixture is indistinguishable from serpentine-only dissolution. Although a steady-state regime was elusive in all serpentine experiments, a variation in dissolution rate of approximately 30% could account for a difference of 10 ppm in [Mg], encompassing all short-lived plateaus in [Mg] data. Mineral-specific rates for both brucite and serpentine correlate robustly with mixture rate parameters despite numerous experimental difficulties (e.g., filter clogging, inhomogeneous mixing and substantial consumption of solids precluding steady-state). Thus it appears possible to infer from experiments on mineral mixtures individual mineral reaction parameters that are generally consistent with parameters extracted from experiments on pure minerals.  4.1.3. Geochemical model predictions Experimental data were used to adjust PHREEQC model parameters to improve predictive capabilities. In order to assess progress to this end, data from MKM are compared to retrieved reaction parameters from the Mixed experiment. Serpentine and brucite are present in a ratio of 9 to 1, respectively, in the Mixed solids and 24 to 1 in MKM tailings. The MKM bulk BET measurement of 9.8 m2 g-1 includes surface area contributions from all present minerals and is close in value to the brucite and serpentine BET measurements of 6.6 and 13.5 m2 g-1, respectively. The difference in BET surface area between serpentine and MKM tailings is over an order of magnitude smaller than the adjustment to BET surface area required for modelling of serpentine dissolution in the Mixed system. Similarity of mineral abundances and surface area measurements between MKM tailings and the brucite-serpentine mixture suggest that parameters retrieved from the Mixed model can be used to make robust predictions about dissolution behaviour of MKM tailings.   28  4.1.4. Rate extraction using trace elements As most mineral components of ultramafic tailings contain many of the same major metals (e.g., Mg and Fe), it may be useful to investigate trace element fluxes as a proxy for bulk mineral dissolution. Many minerals will incorporate trace amounts of transition metals into their structure as substitutions within their crystal lattice, such as trace nickel in olivine [(Mg,Fe)2SiO4] (Simkin and Smith, 1970). It follows that measurement of trace element concentrations from flow through analysis can be used to track dissolution of their host phases. Necessary considerations drawn from present findings are twofold. Firstly, non-stoichiometric dissolution, even at steady-state, may disrupt the correlation between trace element flux and molar flux of the mineral. Secondly, bulk chemistry alone cannot verify the presence of certain trace elements as substitutions within the crystal lattice; they may comprise a discreet minor phase. For example, bulk chemical analysis of the serpentine in this study indicated the presence of trace chrome, but it is unclear whether the chrome is contained within lizardite or minor magnetite (i.e., chrome magnetite) (Burns and Burns, 1975). This ambiguity can be eliminated by the use of electron microprobe analysis to unquestionably determine the chemical composition of each discrete phase in the solid sample to be dissolved.  4.2. Developing a Protocol to Assess Mine Tailings Dissolution 4.2.1. Advantages of a flow-through time-resolved analysis technique The flow-through time-resolved analysis (FT-TRA) method (De Baere et al., 2013) employed in the present study shows promise for future use in a testing protocol for assessment of mine tailings dissolution. Its advantages span experimental design, analytical capabilities and modelling simplicity.  Automation of pumping and sample collection programs allows for extensive temporal data coverage and the possibility of varying conditions within a single experiment. Automated ‘offline’ sampling using a fraction collector can eliminate time delays between consecutive samples and does not require constant experiment supervision. In the ‘online’ format, fluid can be pumped directly from the reaction chamber to a quadrupole ICP-MS for ‘real-time’ measurements of cation concentrations in the 1-1000 ppb range. Superior time-resolution capability is advantageous in optimizing the fit between model and data. ‘Online’  29  data collection is more time-efficient than ‘offline’ collection because experiment and data collection are concurrent, but the cost of ‘online’ analysis is significantly higher. The eluent solution can be interchanged or mixed during an experiment by switching or simultaneously pumping from multiple eluent bottles, allowing for variation of pH with time. The capability for ‘pH-jump’ experiments is valuable for not only extracting pH-dependant dissolution rate laws (Thom et al., 2013) but also for studying reaction mechanisms and kinetic pathways (Samson and Eggleston, 1998; Samson et al., 2000; Pokrovsky and Schott, 2004). Flow rate, and thus residence time, can be adjusted to a certain degree during an experiment as well.  4.2.2. Improvements and limitations to experimental design and geochemical modelling A dichotomy exists between optimization of experimental design for rate extraction and simulation of real tailings chemistry for extrapolation to natural scenarios. These two end-member lines of investigation have distinct motivations, but which is most useful for assessment of tailings carbonation potential is unknown. Each has merits towards a certain goal; the former for laboratory determination of dissolution rates and the latter for empirical investigation of complex tailings chemistry.  4.2.2.1. Optimization for rate extraction Designing experiments optimized for rate extraction involves shifting parameters beyond their natural ranges in order to achieve ideal steady-state dissolution. Removal of fines from the solid sample benefits experimental design in three ways: preventing clogging of the effluent filter, minimizing length of transient dissolution, and preventing substantial consumption of solid due to rapid dissolution of fines. The highly contrasting dissolution rates of brucite and serpentine make it difficult to establish conditions under which they may simultaneously achieve steady-state. Successful rate law extraction may require a high brucite to serpentine ratio, a lengthy residence time at circumneutral pH, or sequential leaching of brucite followed by steady-state dissolution of serpentine using different eluent solutions. Accelerated dissolution due to acidic conditions permits less sensitive analytical techniques and shorter experimental duration.   30  Experimental (Table 4) and field data (Stolberg, 2005; Power et al., 2010) have shown that [Si] is controlled by solubility of an amorphous silica phase akin to chalcedony. Therefore, utility of [Si] for rate extraction depends on maintenance of silica concentration below saturation during steady-state. Optimization of pH and residence time is necessary to prevent silica saturation during brucite-serpentine dissolution. Rate extraction is relatively simple under these ideal conditions, but justification of rate law extrapolation to natural tailings is difficult.  4.2.2.2. Simulation of real tailings Results of experiments designed to closely simulate real tailings chemistry can be justifiably extrapolated to natural scenarios. However, complexity of tailings samples significantly encumbers rate extraction from their dissolution. Presence of fines may cause prolonged transient dissolution, hampering of steady-state due to significant loss of surface area during dissolution, and clogging of the effluent filter. Like MKM tailings (Fig. 2), mine tailings generally contain a high abundance of fine material. Early dissolution of these fine particles non-geometrically reduces total surface area in the reactor. A dissolution-dependant surface area-reduction algorithm must be developed for the PHREEQC reactive transport model to account for this. Effluent filter clogging can be avoided by joining progressively finer filters to catch particles, but loss of surface area remains an issue with this technique. If elemental chemistry and mineralogy of a fine versus coarse fraction of a sample are sufficiently similar, washing to remove fines will overcome the problem of surface area loss to the filter. The decrease in specific surface area due to washing must be accounted for in the model. Additionally, BET surface area measurements of tailings samples reflect bulk surface area, so individual mineral contributions to surface area are unknown. Regulation of [Si] by saturation of an amorphous silica phase may be typical of pore fluid chemistry in ultramafic tailings piles (Wilson et al., 2014). Therefore, deliberate precipitation which convolutes rate extraction may be advantageous for extrapolation of experimental results. Pore fluids at MKM have circumneutral pH (i.e., ~ 5-8) (Wilson et al., 2014); this pH range is more typical of tailings fluid chemistry than acidic conditions used in the present study. However, sluggish dissolution at circumneutral pH would decrease  31  concentrations of dissolved species, potentially causing analytical difficulties, and extend the necessary experimental duration.  Findings of the present study demonstrate that relatively strong correlations can be drawn between single-mineral dissolution rates and mineral mixtures in spite of significantly non-ideal experimental design (e.g., filter clogging, inhomogeneous mixing, silica saturation, and significant consumption of solids). Further research can be accomplished by running an MKM tailings flow-through dissolution experiment and applying findings from this study to discern dissolution of brucite and serpentine using plateaus in pH and [Mg] data, respectively.    32  REFERENCES Bish, D.L., and Howard, S.A., 1988. Quantitative phase analysis using the Rietveld method. Journal of Applied Crystallography, v. 21, p. 86-91. Bobicki, E.R., Liu, Q., Xu, Z., and Zeng, H., 2012. Carbon capture and storage using alkaline industrial wastes. Progress in Energy and Combustion Science, v. 38, p. 302–320 (doi:10.1016/j.pecs.2011.11.002). Brantley, S.L., and Mellot, N.P., 2000. Surface area and porosity of primary silicate minerals. American Mineralogist, v. 85, p. 1767–1783. Bruker AXS, 2008. DIFFRACplus EVA 14 Release 2008. Bruker AXS, Germany. Bruker AXS, 2004. Topas V.3.0: General Profile and Structural Analysis Software for Powder Diffraction Data. Bruker AXS, Germany. Burns, V.M., and Burns, R.G., 1975. Mineralogy of chromium. Geochimica et Cosmochimica Acta, v. 39, p. 903-910. Daval, D., Hellmann, R., Martinez, I., Gangloff, S., and Guyot, F., 2013. Lizardite serpentine dissolution kinetics as a function of pH and temperature, including effects of elevated pCO2. Chemical Geology, v. 351, p. 245–256 (doi:10.1016/j.chemgeo.2013.05.020). De Baere, B., Mayer, K.U., and Francois, R., 2013. Measuring mineral dissolution kinetics using an automated flow-through module. Mineralogical Magazine, v. 77, p. 957 (doi:10.1180/minmag.2013.077.5.4). Gautier, J.M., Oelkers, E.H., and Schott, J., 2001. Are quartz dissolution rates proportional to B.E.T. surface areas? Geochimica et Cosmochimica Acta, v. 65, p. 1059–1070 (doi:10.1016/S0016-7037(00)00570-6). Hariharan, S., Werner, M., Hänchen, M., Zingaretti, D., Baciocchi, R., and Massotti, M., 2014. Dissolution kinetics of thermally activated serpentine for mineralization at flue gas conditions. Energ Procedia, v. 63, p. 5887-5891 (doi:10.1016/j.egypro.2014.11.622). Harrison, A.L., Dipple, G.M., Power, I.M., and Mayer, K.U., 2015. Influence of surface passivation and water content on mineral carbonation rates: Implications for CO2 sequestration in mine tailings. Geochimica et Cosmochimica Acta, v. 148, p. 477–495 (doi:10.1016/j.gca.2014.10.020).  33  Harrison, A.L., Power, I.M., and Dipple, G.M., 2013. Accelerated carbonation of brucite in mine tailings for carbon sequestration. Environmental Science and Technology, v. 47, p. 126–134 (doi:10.1021/es3012854). Hoffert, M.I., Caldeira, K., Benford, G., Criswell, D.R., Green, C., Herzog, H., Jain, A.K., Kheshgi, H.S., Lackner, K.S., Lewis, J.S., Lightfoot, H.D., Manheimer, W., Mankins, J.C., Mauel, M.E., Perkins, L.J., Schlesinger, M.E., Volk, T., and Wigley, T.M.L., 2002. Advanced technology paths to global climate stability: energy for a greenhouse planet. Science, v. 298, p. 981–987 (doi:10.1126/science.1072357). IPCC, 2013. Climate change 2013: The physical science basis. Working group 1 contribution to the fifth assessment report of the Intergovernmental Panel on Climate Change. eds. Stocker, T., Qin, D., Plattner, G.-K., TIgnor, M., Allen, S., Boschung, A., Xia, Y., Bex, V., and Midgley, P., Cambridge University Press, Cambridge, UK and New York. Krevor, S.C., and Lackner, K.S., 2009. Enhancing process kinetics for mineral carbon sequestration. Energy Procedia, v. 1, p. 4867–4871 (doi:10.1016/j.egypro.2009.02.315). Krevor, S.C.M., and Lackner, K.S., 2011. Enhancing serpentine dissolution kinetics for mineral carbon dioxide sequestration. International Journal of Greenhouse Gas Control, v. 5, p. 1073–1080 (doi:10.1016/j.ijggc.2011.01.006). Kump, L.R., Brantley, S.L., and Arthur, M.A., 2000. Chemical weathering, atmospheric CO2, and climate. Annual Review of Earth and Planetary Sciences, v. 28, p. 611–667. Lackner, K.S., 2003. A Guide to CO2 Sequestration. Science, v. 300, p. 1677–1678 (doi:10.1126/science.1079033). Lackner, K.S., Wendt, C.H., Butt, D.P., Joyce, E.L., and Sharp, D.H., 1995. Carbon dioxide disposal in carbonate minerals. Energy, v. 20, p.1153–1170. Mayer, K.U., 2002. Multicomponent reactive transport modeling in variably saturated porous media using a generalized formulation for kinetically controlled reactions. Water Resources Research, v. 38, p. 1174 (doi:10.1029/2001WR000862). Parkhurst, D.L., and Appelo, C.A.J., 2013. Description of Input and Examples for PHREEQC Version 3 — A Computer Program for Speciation, Batch-Reaction, One- 34  Dimensional Transport, and Inverse Geochemical Calculations. In U.S. Geological Survey Techniques and Methods, book 6, chap. A43. Denver, Colorado. p. 497. Pokrovsky, O.S., and Schott, J., 2004. Experimental study of brucite dissolution and precipitation in aqueous solutions: Surface speciation and chemical affinity control. Geochimica et Cosmochimica Acta, v. 68, p. 31–45 (doi:10.1016/S0016-7037(03)00238-2). Power, I.M., Dipple, G.M., and Southam, G., 2010. Bioleaching of ultramafic tailings by Acidithiobacillus spp. for CO2 sequestration. Environmental Science and Technology, v. 44, p. 456–462 (doi:10.1021/es900986n). Power, I.M., Harrison, A.L., Dipple, G.M., and Southam, G., 2013. Carbon sequestration via carbonic anhydrase facilitated magnesium carbonate precipitation. International Journal of Greenhouse Gas Control, v. 16, p. 145–155 (doi:10.1016/j.ijggc.2013.03.011). Power, I.M., Wilson, S.A., Small, D.P., Dipple, G.M., Wan, W., and Southam, G., 2011. Microbially mediated mineral carbonation: Roles of phototrophy and heterotrophy. Environmental Science and Technology, v. 45, p. 9061–9068. (doi:10.1021/es201648g). Rietveld, H.M., 1969. A profile refinement method for nuclear and magnetic structures. Journal of Applied Crystallography, v. 2, p. 65–71 (doi:10.1107/S0021889869006558). Samson, S.D., Stillings, L.L., and Eggleston, C.M., 2000. The depletion and regeneration of dissolution-active sites at the mineral-water interface: I. Fe, Al, and In sesquoxides. Geochimica et Cosmochimica Acta, v. 64, p. 3471-3484. Samson, S.D., and Eggleston, C.M., 1998. Active sites and the non-steady-state dissolution of hematite. Environmental Science and Technology, v. 32, p. 2871–2875 (doi:10.1021/es9803097). Sapsford, D.J., Bowell, R.J., Dey, M., and Williams, K.P., 2009. Humidity cell tests for the prediction of acid rock drainage. Minerals Engineering, v. 22, p. 25-36 (doi:10.1016/j.mineng.2008.03.008). Simkin, T., and Smith, J.V., 1970. Minor-element distribution in olivine. The Journal of Geology, v. 78, p. 304-325.  35  Stolberg, D.J., 2005. Rehabilitation studies on tailings storage facilities in an arid hypersaline region. Ph. D. thesis. The University of Queensland, Brisbane, Australia. Thom, J.G.M., Dipple, G.M., Power, I.M., and Harrison, A.L., 2013. Chrysotile dissolution rates: Implications for carbon sequestration. Applied Geochemistry, v. 35, p. 244–254 (doi:10.1016/j.apgeochem.2013.04.016). Wilson, S.A., Harrison, A.L., Dipple, G.M., Power, I.M., Barker, S.L.L., Mayer, U.K., Fallon, S.J., Raudsepp, M., and Southam, G., 2014. Offsetting of CO2 emissions by air capture in mine tailings at the Mount Keith Nickel Mine, Western Australia: Rates, controls and prospects for carbon neutral mining. Internation Journal of Greenhouse Gas Control, v. 25, p. 121–140 (doi:10.1016/j.ijggc.2014.04.002). Wilson, S.A., Raudsepp, M., and Dipple, G.M., 2006. Verifying and quantifying carbon fixation in minerals from serpentine-rich mine tailings using the Rietveld method with X-ray powder diffraction data. American Mineralogist, v. 91, p. 1331–1341 (doi:10.2138/am.2006.2058). Wilson, S.A., Dipple, G.M., Power, I.M., Thom, J.M., Anderson, R.G., Raudsepp, M., Gabites, J.E., and Southam, G., 2009. Carbon dioxide fixation within mine wastes of ultramafic-hosted ore deposits: Examples from the Clinton Creek and Cassiar Chrysotile deposits, Canada. Economic Geology, v. 104, p. 95–112 (doi:10.2113/gsecongeo.104.1.95). Yanes, Y., and Yapp, C.J., 2010. Indoor and outdoor urban atmospheric CO2: Stable carbon isotope constraints on mixing and mass balance. Applied Geochemistry, v. 25, p. 1339–1349 (doi:10.1016/j.apgeochem.2010.06.004). You, Y., Niu, C., Zhou, J., Liu, Yating., Bai, Z., Zhang, J., He, F., and Zhang, N., 2012. Measurement of air exchange rates in different indoor environments using continuous CO2 sensors. Journal of Environmental Sciences, v. 24, p. 657-664 (doi:10.1016/S1001-0742(11)60812-7).    36  APPENDIX   A. MKM tailings a Sample 06MK30-7 serpentine 81.6 brucite   3.4 hydrotalcite-group   4.7 magnetite and chromite    3.8 halite    2.4 magnesite    2.9 dolomite    1.0 calcite    0.3 Total   100 a    Modified from Wilson et al. (2014).  B. Brucite sample brucite                  94.4                  dolomite                    5.1                  calcite                    0.5                  Total                100.0   C. Serpentine sample lizardite                  94.8                  magnetite                    5.2                  Total                 100.0  Table A1. Mineralogical data in wt.% obtained by XRPD with Reitveld refinement for: A) MKM tailings, B) brucite, and C) serpentine samples, showing minor impurities in mineral samples. Values are renormalized without amorphous and spike (fluorite) phases. Halite was removed from MKM tailings by washing. XRD patterns with Rietveld refinement are given in Figures A1 and A2.    37   Table A2. Summary of all PHREEQC modelling data for weighted sums of squares analysis of Brucite-DI experiment.  MODEL INPUT MODEL OUTPUT  SAbrc (m2 g-1) a pCO2 (ppm) b [Mg]SS (ppm) pHSS Weighted SSq c 0.08 316 0.19 8.40 1.21E-03 0.10 316 0.22 8.68 6.01E-04 0.13 316 0.25 8.84 1.16E-03 0.08 355 0.20 8.30 1.83E-03 0.10 355 0.23 8.63 4.73E-04 0.13 355 0.26 8.80 8.77E-04 0.08 398 0.21 8.14 3.41E-03 0.10 398 0.24 8.57 4.21E-04 0.13 398 0.27 8.76 6.37E-04 0.15 398 0.30 8.89 1.34E-03 0.18 398 0.32 8.98 2.08E-03 0.20 398 0.35 9.06 2.94E-03 0.23 398 0.37 9.12 3.69E-03 0.25 398 0.39 9.17 4.40E-03 0.08 447 0.22 7.88 7.50E-03 0.10 447 0.26 8.48 5.42E-04 0.13 447 0.29 8.71 4.03E-04 0.15 447 0.31 8.85 1.00E-03 0.18 447 0.34 8.95 1.75E-03 0.20 447 0.36 9.04 2.67E-03 0.23 447 0.38 9.10 3.40E-03 0.25 447 0.41 9.16 4.24E-03 0.10 501 0.27 8.35 1.13E-03 0.13 501 0.30 8.65 2.27E-04 0.15 501 0.33 8.81 7.17E-04 0.18 501 0.35 8.92 1.45E-03 0.20 501 0.38 9.01 2.30E-03 0.23 501 0.40 9.08 3.12E-03 0.25 501 0.43 9.14 3.93E-03 0.10 562 0.29 8.16 2.85E-03 0.13 562 0.32 8.56 1.67E-04 0.15 562 0.35 8.75 3.83E-04 0.18 562 0.37 8.88 1.10E-03 0.20 562 0.40 8.98 1.96E-03  38  0.23 562 0.42 9.05 2.73E-03 0.25 562 0.44 9.11 3.51E-03 0.10 631 0.30 7.85 7.78E-03 0.13 631 0.34 8.44 4.45E-04 0.15 631 0.37 8.68 1.33E-04 0.17 631 0.38 8.78 4.62E-04 0.18 631 0.39 8.83 7.29E-04 0.20 631 0.42 8.94 1.56E-03 0.23 631 0.44 9.02 2.38E-03 0.25 631 0.47 9.08 3.12E-03 0.10 708 0.32 7.47 1.74E-02 0.13 708 0.36 8.24 1.81E-03 0.15 708 0.39 8.59 2.13E-05 0.18 708 0.42 8.77 3.92E-04 0.20 708 0.44 8.89 1.14E-03 0.23 708 0.47 8.98 1.96E-03 0.25 708 0.49 9.05 2.78E-03 0.10 794 0.33 7.20 2.66E-02 0.13 794 0.38 7.92 6.27E-03 0.15 794 0.41 8.45 3.08E-04 0.18 794 0.44 8.68 8.74E-05 0.20 794 0.47 8.83 7.31E-04 0.23 794 0.49 8.93 1.52E-03 0.25 794 0.52 9.01 2.36E-03 0.10 891 0.34 7.00 3.46E-02 0.13 891 0.40 7.51 1.60E-02 0.15 891 0.44 8.21 2.05E-03 0.18 891 0.47 8.56 3.92E-05 0.20 891 0.50 8.75 3.54E-04 0.23 891 0.52 8.87 1.08E-03 0.25 891 0.55 8.96 1.91E-03 0.10 1000 0.35 6.84 4.19E-02 0.13 1000 0.41 7.22 2.57E-02 0.15 1000 0.46 7.82 8.22E-03 0.18 1000 0.50 8.38 7.03E-04 0.20 1000 0.53 8.64 1.32E-04 0.23 1000 0.55 8.79 6.66E-04 0.25 1000 0.58 8.90 1.48E-03 0.10 1122 0.36 6.72 4.77E-02 0.13 1122 0.43 7.01 3.41E-02 0.15 1122 0.48 7.42 1.88E-02  39  0.18 1122 0.53 8.04 4.34E-03 0.20 1122 0.56 8.47 4.22E-04 0.23 1122 0.59 8.68 3.77E-04 0.25 1122 0.61 8.82 1.05E-03 0.10 1259 0.37 6.61 5.35E-02 0.13 1259 0.44 6.85 4.13E-02 0.15 1259 0.50 7.15 2.84E-02 0.18 1259 0.55 7.58 1.42E-02 0.20 1259 0.60 8.17 2.81E-03 0.23 1259 0.63 8.52 5.38E-04 0.25 1259 0.66 8.71 7.59E-04 0.10 1413 0.38 6.52 5.84E-02 0.13 1413 0.45 6.73 4.72E-02 0.15 1413 0.52 6.96 3.64E-02 0.18 1413 0.58 7.25 2.48E-02 0.20 1413 0.63 7.69 1.16E-02 0.23 1413 0.67 8.23 2.50E-03 0.25 1413 0.70 8.54 8.82E-04 a     Values given are specific surface area. Model inputs were total surface area, the product of specific surface area and total grams of brucite. b     The PHREEQC input for CO2 concentration was saturation index (SICO2).  c     Bolded values are within tolerance.    262 10 COSIpCO  40        Figure A1. X-ray diffraction pattern and Rietveld refinement of initial brucite sample. The blue line represents observed data and is overlain by the calculated data in red. The grey line shows the residual pattern and vertical lines below show Bragg diffractions for each phase.     41           Figure A2. X-ray diffraction pattern and Rietveld refinement of initial serpentine sample. The blue line represents observed data and is overlain by the calculated data in red. The grey line shows the residual pattern and vertical lines below show Bragg diffractions for each phase. The fluorite was added to allow calculation of amorphous content in the sample.      42   Figure A3. X-ray diffraction pattern for Serp2 reacted solids. The black line is the observed data and vertical lines represent Bragg diffractions. Red and blue lines correspond to lizardite and magnetite, respectively.    Figure A4. X-ray diffraction pattern for Mixed reacted solids. The black line is the observed data and vertical lines represent Bragg diffractions. Red and blue lines correspond to lizardite and magnetite, respectively. Brucite was not detected. 

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.52966.1-0053619/manifest

Comment

Related Items