Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Evaluating the desalination performance and efficiency of capacitive deionization with activated carbon… Chung, Ting-Chih 2018

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

Item Metadata

Download

Media
24-ubc_2018_may_chung_ting-chih.pdf [ 4.13MB ]
Metadata
JSON: 24-1.0365275.json
JSON-LD: 24-1.0365275-ld.json
RDF/XML (Pretty): 24-1.0365275-rdf.xml
RDF/JSON: 24-1.0365275-rdf.json
Turtle: 24-1.0365275-turtle.txt
N-Triples: 24-1.0365275-rdf-ntriples.txt
Original Record: 24-1.0365275-source.json
Full Text
24-1.0365275-fulltext.txt
Citation
24-1.0365275.ris

Full Text

  EVALUATING THE DESALINATION PERFORMANCE AND EFFICIENCY OF CAPACITIVE DEIONIZATION WITH ACTIVATED CARBON ELECTRODES  by TING-CHIH CHUNG  B.Sc., University of Toronto, 2015   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF APPLIED SCIENCE in  THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Chemical and Biological Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   April 2018   © Ting-Chih Chung, 2018   The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, a thesis/dissertation entitled:  EVALUATING THE DESALINATION PERFORMANCE AND EFFICIENCY OF CAPACITIVE DEIONIZATION WITH ACTIVATED CARBON ELECTRODES  submitted by Ting-Chih Chung  in partial fulfillment of the requirements for the degree of Master of Applied Science in Chemical and Biological Engineering   Examining Committee: Madjid Mohseni, Chemical and Biological Engineering Co-supervisor David Wilkinson, Chemical and Biological Engineering Co-supervisor  Anthony Lau, Chemical and Biological Engineering Supervisory Committee Member  Additional Examiner   Additional Supervisory Committee Members: Jongho Lee, Civil Engineering Supervisory Committee Member  Supervisory Committee Member     iii  Abstract   Capacitive deionization (CDI) is an incipient desalination technology based on the principle of electrical double layer capacitors. When a constant voltage is applied to high surface area and electrically conductive electrodes, electrodes become oppositely charged and ions are adsorbed onto the electrode surfaces under the presence of the electric field, thereby producing a purified stream of water. When the electrodes are saturated with ions, the applied voltage is removed or the polarity is reversed to desorb the ions and generate a stream of waste concentrate.   For brackish water with intermediate salinities, CDI technology has advantages over conventional desalination technologies because of operation at ambient temperatures and pressures, high water recovery, and no chemical usage. However, there are issues with translating CDI technology from laboratory to practice because of the lack of experience with its operation and uncertainties about its robustness and durability. To address these challenges, this thesis investigated a laboratory-scale CDI cell with the aim to holistically evaluate its desalination performance and efficiency using desalination metrics, kinetic models, and circuit models. The activated carbon electrodes used in the CDI cell were purchased commercially as well as fabricated in-house, and were analyzed with cyclic voltammetry, scanning electron microscopy and energy-dispersive X-ray spectroscopy. Operating parameters including applied voltage, NaCl concentration, and flow rate were varied to study their effects. Lastly, the effect of including ion-exchange membranes was examined and preliminary tests were performed to explore the long-term desalination performance and efficiency of CDI technology.  Commercial electrodes were found to be superior to the in-house fabricated electrodes. For operating parameters, higher applied voltages were found to increase the salt adsorption and capacitance but decrease the energy efficiency. Increasing NaCl concentration also increased salt adsorption but did not affect capacitance or energy efficiency. No trends were observed for flow rate and kinetic parameters. Ion-exchange membranes boosted the electrode performance considerably, with salt adsorption improving by 1.70 – 1.94 times and energy efficiency by 1.11 – 1.35 times. Long-term tests showed that electrode performance degraded steadily and reached half its original performance at 40 cycles but could be regenerated with NaOH washing. iv  Lay Summary   In light of the issues of water scarcity around the world, there is a need for alternative supplies of drinking water to be found. Desalination, the process in which salt water is converted to drinking water, is an attractive option for the supply of drinking water because of the vast quantity of salt water resources on Earth. Capacitive deionization (CDI) is an emerging technology for desalination that uses electricity to induce a charge onto high surface area, electrically conductive electrodes, thereby causing salt ions in water to adsorb onto the electrode surfaces. As a result, purified drinking water can be produced. This research investigates the performance and efficiency of capacitive deionization technology with cost-effective activated carbon electrodes for desalination of low concentration, synthetic salt waters. The results have implications for assessing and evaluating the feasibility of capacitive deionization when compared to already existing desalination technologies.              v  Preface   The author, Ting-Chih (Mike) Chung, under the supervision of Dr. Madjid Mohseni and Dr. David Wilkinson from the CHBE Department at UBC Vancouver, was responsible for the research project including the identification of research objectives, literature review, experiment design, data collection and analysis, and presentation of findings in this thesis.  Adrian Serrano helped obtain the data for the CV experiments described and reported in Chapters 3 and 4.   The work in this thesis has been presented and published in the proceedings of the following conferences:  • Mike Chung, David Wilkinson, Madjid Mohseni.  “Evaluating the Desalination Performance of Capacitive Deionization with Activated Carbon Electrodes.” CSChE 2017 (Oral Presentation), 2017 October 22-25. • Mike Chung, David Wilkinson, Madjid Mohseni. “Capacitive Deionization with Activated Carbon Electrodes for Desalination of Brackish Water.” RES’EAU AGM 2017 (Poster Presentation), 2017 May 26-27.   • Mike Chung, Adrian Serrano, Macarena Cataldo, David Wilkinson, Madjid Mohseni. “Capacitive Deionization with Activated Carbon Electrodes for Desalination of Brackish Water.” NDWC 2017 (Poster Presentation), 2016 October 14-16. • Mike Chung, David Wilkinson, Madjid Mohseni. “Capacitive Deionization (CDI) – A Promising Technology for Desalination of Brackish Water”. RES’EAU AGM 2016 (Poster Presentation), 2016 April 20 – May 1. • Mike Chung, David Wilkinson, Madjid Mohseni. “Capacitive Deionization (CDI) for Desalination of Brackish Water”. IC-IMPACTS 2016 AGM & Research Conference (Poster Presentation), 2016 March 10-11.  vi  Parts of Chapters 1 and 2 were adapted from sections of a manuscript of a review paper currently being prepared for submission. Only the author’s contribution to the manuscript was used. The details of the review paper are as follows: • “Challenges and opportunities in the development of capacitive deionization (CDI) for brackish water treatment: From lab to practice.” Dehkhoda, A., Chung, M., Serrano, A., Cataldo, M., Ellis, N., Wilkinson, D., Mohseni, M.                vii  Table of Contents  Abstract ......................................................................................................................................... iii Lay Summary ............................................................................................................................... iv Preface ............................................................................................................................................ v Table of Contents ........................................................................................................................ vii List of Tables ................................................................................................................................ xi List of Figures .............................................................................................................................. xii List of Abbreviations .................................................................................................................. xv List of Symbols ......................................................................................................................... xviii Acknowledgements .................................................................................................................... xxi Dedication .................................................................................................................................. xxii Chapter 1: Introduction ............................................................................................................... 1 1.1. Drinking Water Treatment ................................................................................................... 1 1.2. Current State of Desalination ............................................................................................... 2 1.3. Capacitive Deionization as an Alternative Desalination Technology .................................. 5 1.4. Research Objectives ........................................................................................................... 10 Chapter 2: Literature Review .................................................................................................... 11 2.1. Defining Desalination Performance and Efficiency........................................................... 11 2.1.1. Salt Removal Efficiency and Water Recovery ............................................................ 11 2.1.2. Specific Capacitance.................................................................................................... 12 2.1.3. Salt Adsorption Capacity ............................................................................................. 14 2.1.4. Charge Efficiency ........................................................................................................ 15 2.1.5. Specific Energy Consumption ..................................................................................... 15 2.1.6. Model Parameters ........................................................................................................ 16 2.2. Electrode Materials and Processing ................................................................................... 16 2.2.1. Ideal Electrode Properties for Electrosorption ............................................................ 16 2.2.2. Activated Carbon Powder Electrodes .......................................................................... 18 2.2.2.1. Activated Carbon Powder .................................................................................... 18 2.2.2.2. Polymeric Binders, Conductive Additives, and Current Collectors .................... 19 2.2.2.3. Fabrication Procedures......................................................................................... 21 viii  2.2.3. Other Materials and Nanotechnology .......................................................................... 22 2.3. Cell Architecture ................................................................................................................ 26 2.3.1. Flow-by and Flow-through Geometries ...................................................................... 26 2.3.2. Membrane Capacitive Deionization ............................................................................ 27 2.3.3. Flow-Electrode Capacitive Deionization and Inverted Capacitive Deionization ........ 28 2.4. System Operation ............................................................................................................... 29 2.4.1. Process Flows .............................................................................................................. 30 2.4.2. Applied Voltage ........................................................................................................... 32 2.4.3. Flow Rate ..................................................................................................................... 32 2.4.4. Salt Concentration ....................................................................................................... 33 2.4.5. Environmental Factors ................................................................................................. 34 2.4.6. Modes of Operation ..................................................................................................... 35 2.5. Electrosorption Modelling.................................................................................................. 37 2.5.1. Adsorption Isotherms .................................................................................................. 37 2.5.2. Lagergren Adsorption Kinetics ................................................................................... 38 2.5.3. Electrical Double Layer and Ion Transport ................................................................. 40 2.6. Degradation and Long-Term Stability ............................................................................... 42 2.6.1. Fouling and Scaling ..................................................................................................... 42 2.6.2. Parasitic Electrochemical Reactions ............................................................................ 43 Chapter 3: Methodology............................................................................................................. 47 3.1. Activated Carbon Electrodes .............................................................................................. 47 3.1.1. Materials ...................................................................................................................... 47 3.1.2. Fabrication procedures ................................................................................................ 49 3.1.3. Characterization ........................................................................................................... 51 3.2. Capacitive Deionization Experimental Setup .................................................................... 52 3.2.1. Capacitive Deionization Cell ....................................................................................... 52 3.2.2. Process Flow Setup ...................................................................................................... 52 3.2.3. Electrical Circuit and Measurement Setup .................................................................. 53 3.3. Desalination Experiments .................................................................................................. 55 3.3.1. Comparison of Activated Carbon Electrodes .............................................................. 55 3.3.2. Varying of Operating Parameters ................................................................................ 55 3.3.3. Long-Term and Regeneration Tests ............................................................................ 56 ix  3.4. Desalination Metrics and Modelling .................................................................................. 57 3.4.1. Adsorption Isotherms .................................................................................................. 58 3.4.2. Lagergren Adsorption Kinetics ................................................................................... 59 3.4.3. Equivalent Electrical Circuit ....................................................................................... 59 Chapter 4: Results and Discussion ............................................................................................ 62 4.1. Electrode Fabrication ......................................................................................................... 62 4.1.1. Structural Integrity and Mass Loading ........................................................................ 62 4.1.2. Scanning Electron Microscopy and Energy Dispersive X-Ray Spectroscopy ............ 71 4.1.3. Cyclic Voltammetry .................................................................................................... 78 4.2. Desalination Experiments .................................................................................................. 79 4.2.1. Processing of Raw Data ............................................................................................... 79 4.2.2. Charge Efficiency and Specific Energy Consumption vs. Cycle Time ....................... 81 4.2.3. Adsorption Kinetics and Electrical Circuit Modelling ................................................ 83 4.2.4. Commercial vs. Fabricated Electrodes ........................................................................ 85 4.3. Effect of Operating Parameters .......................................................................................... 87 4.3.1. Effect of Applied Voltage and Ion-Exchange Membranes ......................................... 87 4.3.2. Effect of NaCl Concentration ...................................................................................... 92 4.3.3. Effect of Flow Rate ...................................................................................................... 95 4.3.4. Combined Effects of Applied Voltage and NaCl Concentration ................................ 97 4.4. Degradation and Regeneration ........................................................................................... 99 4.4.1. Long-Term Experiments.............................................................................................. 99 Chapter 5: Conclusion and Recommendations ...................................................................... 105 5.1. Summary of Work ............................................................................................................ 105 5.2. Future Development and Recommendations ................................................................... 106 Bibliography .............................................................................................................................. 110 Appendices ................................................................................................................................. 124 Appendix A. Graphite Foil Specifications .............................................................................. 124 A.1. Pureechem Electrode Graphite Foil Specifications ..................................................... 124 A.2. Mineral Seal Corporation Flexible Graphite 2010A Specifications ............................ 124 Appendix B. Electrical Circuit for CDI Setup ........................................................................ 125 B.1. Electrical Circuit for Zero Voltage Discharge ............................................................. 125 B.2. Electrical Circuit for Reverse Polarity Discharge ........................................................ 125 x  Appendix C. Statistical Formulas............................................................................................ 126 C.1. Correlation Coefficient (r2) .......................................................................................... 126 C.2. Root-Mean-Square Error (RMSE) ............................................................................... 126 Appendix D. Raw Data Processing and Modelling Calculations............................................ 127 D.1. pH Contribution to Conductivity Calculation of NaCl Concentration ........................ 127 D.2. NaCl Concentration to Conductivity Calibration Curve ............................................. 128 D.3. Salt Adsorption Capacity, Charge Efficiency, and Specific Energy Consumption Calculation ........................................................................................................................... 128 D.4. Adsorption Isotherm Linear Plots and Calculation of Parameters .............................. 130 Appendix E. SEM Images and EDS Spectra of Regenerated Electrodes ............................... 131 E.1. Deionized Water Washed Electrode ............................................................................ 131 E.2. Citric Acid Washed Electrode ...................................................................................... 131 Appendix F. Electrode Potential vs. Hg/HgO Reference Electrode ....................................... 132            xi  List of Tables Table 1.1. Desalination separation processes and TDS range for cost-effective application, adapted from [9]. ............................................................................................................................. 3 Table 1.2. Energy requirements and costs of conventional desalination technologies, adapted from [9]. .......................................................................................................................................... 4 Table 1.3. List of active vendors for CDI technology. ................................................................... 9 Table 2.1. Advantages and disadvantages of different binders for CDI electrodes. .................... 20 Table 2.2. Summary of operating parameters and desalination performance of materials for CDI electrodes. Slashes indicate composite materials. Compiled from sources: [15], [18], [74]. ....... 25 Table 3.1. Specifications of ACPs from Pureechem, Calgon Carbon and Yihuan Carbon. ......... 49 Table 3.2. Key specifications and parameters for in-house fabricated ACP electrodes. .............. 55 Table 4.1. Desalination metrics of the ACP electrodes at baseline conditions. ........................... 86 Table 4.2. Lagergren adsorption kinetics and equivalent electrical circuit model parameters for CDI (top) and MCDI (bottom) systems with varying applied voltage. ........................................ 90 Table 4.3. Lagergren adsorption kinetics and equivalent electrical circuit model parameters for different NaCl concentrations. ...................................................................................................... 94 Table 4.4. Lagergren adsorption kinetics and equivalent electrical circuit model parameters for different flow rates. ....................................................................................................................... 96 Table 4.5. Salt adsorption capacity (top), charge efficiency (middle), and specific energy consumption (bottom) results for 3 x 3 matrix of experiments with varying applied voltage and NaCl concentration at a flow rate of 20 mL/min. ......................................................................... 97         xii  List of Figures Figure 1.1. Process diagram of typical drinking water treatment system. ..................................... 2 Figure 1.2. Schematic representation of the basic principle of CDI separation............................. 5 Figure 1.3. EDL model showing surface and solution interfacial interactions. ............................. 6 Figure 2.1. Flow-by (a) and flow-through (b) geometries for CDI systems. .............................. 27 Figure 2.2. Prevention of co-ion expulsion with an anion-exchange membrane (AEM) and cation-exchange membrane (CEM) in MCDI............................................................................... 28 Figure 2.3. Process flow diagrams of batch (a), recirculating batch (b) and single pass (c) processes for CDI experimental setups; sources: [133]–[135]. .................................................... 31 Figure 2.4. EDL structure according to the GCS model (a) and the mD model (b); source: [15]........................................................................................................................................................ 41 Figure 3.1. 3D drawing of Pureechem ACP electrode with dimensions. .................................... 48 Figure 3.2. In-house ACP electrode fabrication procedure and completed 8 x 8 cm2 electrode (a) and 15 x 10 cm2 electrode (b). ...................................................................................................... 51 Figure 3.3. Schematic of stack in CDI cell (left) and picture of CDI cell (right). ....................... 52 Figure 3.4. Process diagram of CDI desalination experimental setup showing process flow, electric circuit, and measurement instrumentation. ...................................................................... 54 Figure 3.5. Picture of CDI desalination experimental setup. ....................................................... 54 Figure 3.6. Electrical circuit diagrams of the RC series circuit (a) and Randles circuit (b). ....... 60 Figure 4.1. YEC-8A/PVDF electrodes with ACP-to-PVDF ratios of 85:15 (left), 90:10 (middle), and 95:5 (right). Electrodes were fabricated using the EC method with a wet mass loading of 31.3 mg/cm2 and a solvent-to-solids ratio of 5:1. ......................................................................... 62 Figure 4.2. YEC-8A/PVDF electrodes with solvent-to-solids ratios of 4:1 (left) and 5:1 (right). Electrodes were fabricated with the EC method with an ACP-to-PVDF ratio of 85:15 and a wet mass loading of 31.3 mg/cm2. ....................................................................................................... 64 Figure 4.3. YEC-8A/PVDF electrodes fabricated with the DBC method using a solvent-to-solids mass ratio of 3:1 (left) and 4:1 (right). The wet thickness was set to 200 µm and the ACP-to-PVDF ratio was 85:15. .................................................................................................................. 65 Figure 4.4. WPC/PVDF electrodes fabricated with the EC method using ACP-to-PVDF ratios of 85:15 (left), 90:10 (middle), and 95:5 (right). The solvent-to-solids ratio was 1.5:1 and the wet mass loading was 31.25 mg/cm2. .................................................................................................. 66 Figure 4.5. Relationship between dry mass loading and wet mass loading for WPC/PVDF and YEC-8A/PVDF electrodes fabricated with the EC method. WPC/PVDF electrodes were fabricated with a solvent-to-solids ratio of 1.5:1 and YEC-8A/PVDF electrodes with 5:1. ........ 67 xiii  Figure 4.6. Relationship between dry mass loading and wet thickness for YEC-8A/PVDF electrodes fabricated with the DBC method. Conditions were kept constant with an ACP-to-PVDF ratio of 85:15 and a solvent-to-solids ratio of 5:1. ............................................................. 68 Figure 4.7. YEC-8A/PVDF electrodes fabricated using the EC method with dry mass loadings of 39.1 mg/cm2 (left) and 46.9 mg/cm2. ACP-to-PVDF ratio was 85:15 and solvent-to-solids ratio was 5:1. ................................................................................................................................. 69 Figure 4.8. WPC/PVDF electrodes fabricated using the EC method with dry mass loadings of 14 mg/cm2 (left), 28 mg/cm2 (middle) and 40 mg/cm2. ACP-to-PVDF ratio was 85:15 and solvent-to-solids ratio was 1.5:1. ............................................................................................................... 69 Figure 4.9. YEC-8A/PVDF electrode on carbon paper with a dry mass loading of 2.30 mg/cm2, ACP-to-PVDF ratio of 85:15 and solvent-to-solids ratio of 3:1. Graphite foil was taped onto the 10 x 10 cm2 electrode to prevent leaking when placed in the CDI unit........................................ 70 Figure 4.10. SEM images of graphite foil (a), Pureechem electrode (b), WPC/PVDF electrode (c) and YEC-8A/PVDF electrode (d) with ACP-to-PVDF ratios of 85:15. ................................. 72 Figure 4.11. EDS spectra and SEM images of YEC-8A/PVDF electrodes with ACP-to-PVDF ratios of 85:15 (top), 90:10 (middle), and 95:5 (bottom). Solvent-to-solids ratio was 5:1 and dry mass loading was 5 mg/cm2. ......................................................................................................... 73 Figure 4.12. EDS spectra of WPC/PVDF electrode (top) and Pureechem electrode (bottom). The WPC/PVDF electrode had an ACP-to-PVDF ratio of 85:15, solvent-to-solids ratio of 1.5:1 and dry mass loading of 13.75 mg/cm2. .............................................................................................. 75 Figure 4.13. SEM image of carbon paper (a) and YEC-8A/PVDF electrode on carbon paper (b) with ACP-to-PVDF ratio of 85:15, solvent-to-solid ratio of 3:1, and dry mass loading of 2.30 mg/cm2. Below is the EDS spectrum of the YEC-8A/PVDF electrode on carbon paper (c). ...... 76 Figure 4.14. Cross-sectional SEM images of the YEC-8A/PVDF electrode on graphite foil (a), YEC-8A/PVDF electrode on carbon paper (b), and Pureechem electrode (c). ............................ 77 Figure 4.15. Cyclic voltammograms of the ACP electrodes at a scan rate of 5 mV/s and an electrolyte concentration of 1000 ppm NaCl. ............................................................................... 79 Figure 4.16. Example of raw data generated from desalination experiments. Data shown is from desalination experiment with Pureechem ACP electrodes without membrane at baseline conditions of 1.2 V, 1000 ppm NaCl, 20 mL/min, and 60 min cycle time. Conductivity and pH vs. time (top) and current vs. time (bottom) data were logged in 10 s intervals........................... 80 Figure 4.17. Charge efficiency (top) and specific energy consumption (bottom) vs. cycle time of desalination experiments with Pureechem electrodes at baseline condition. ............................... 82 Figure 4.18. Theoretical results from the Lagergren adsorption kinetics model and experimental results of the desalination experiments with Pureechem electrodes at baseline conditions. ........ 84 Figure 4.19. Theoretical results from the equivalent electrical circuit models and experimental results of the desalination experiments with Pureechem electrodes at baseline conditions. ........ 85 xiv  Figure 4.20. Column graphs of salt adsorption capacity (a), charge efficiency (b), and specific energy consumption (c) for CDI and MCDI systems with varying applied voltages. Experiments were performed with 1000 ppm NaCl concentration and 20 mL/min flow rate. Error bars are shown for n = 3 trials. ................................................................................................................... 88 Figure 4.21. Column graphs of salt adsorption capacity (a), charge efficiency (b), and specific energy consumption (c) at varying NaCl concentrations. Experiments were performed with 1.2 V applied voltage and 20 mL/min flow rate. Error bars are shown for n = 3 trials. ......................... 92 Figure 4.22. Graph of adsorption isotherms and experimental data (left) and table of model parameters (right). ......................................................................................................................... 93 Figure 4.23. Column graphs of salt adsorption capacity (a), charge efficiency (b), and specific energy consumption (c) at varying flow rates. Experiments were performed with 1.2 V applied voltage and 1000 ppm NaCl concentration. Error bars are shown for n = 3 trials. ...................... 95 Figure 4.24. Salt adsorption capacity as a function of cycle number using Pureechem electrodes at baseline conditions. Regenerated electrode salt adsorption capacities are also displayed. ...... 99 Figure 4.25. Snapshot of long-term desalination experiment showing pH fluctuations and tendency toward acidity. ............................................................................................................. 101 Figure 4.26. Electrode potential of the anode and cathode in the CDI cell over time at baseline conditions. ................................................................................................................................... 102 Figure 4.27. SEM image and EDS spectra of 0.01 M NaOH washed Pureechem electrode after degradation. ................................................................................................................................. 103           xv  List of Abbreviations Abbreviation Definition 2D 3D AC ACC ACF ACP AEM AGM BET CA Ca2+ CDI CEM CHBE Cl2 Cl- CMC CNT CO2 CV CVD DBC DC DMA EC ED EDL EDLC Two-Dimensional Three-Dimensional Alternating Current Activated Carbon Cloth Activated Carbon Fiber Activated Carbon Powder Anion-Exchange Membrane Annual General Meeting Brunauer-Emmett-Teller  Carbon Aerogel calcium (ion) Capacitive Deionization Cation-Exchange Membrane Chemical and Biological Engineering chlorine (gas) chloride (ion) carboxymethyl cellulose Carbon Nanotube carbon dioxide (gas) Cyclic Voltammetry Chemical Vapor Deposition Doctor Blade Casting Direct Current N,N-dimethylacetamide Evaporative Casting Electrodialysis Electrical Double Layer Electrical Double Layer Capacitor xvi  Abbreviation Definition EDS EIS EOAS FCDI Fe3+ FTIR GCS GR H2 H+ Hg HgO HgSO4 HNO3 H2O H2O2 i-CDI IC-IMPACTS  IEM IUPAC KOH MC MCDI MED MnO2 MSF N2 Na+ NaCl Energy-Dispersive X-Ray Spectroscopy Electrochemical Impedance Spectroscopy Earth, Ocean and Atmospheric Sciences Flow-Electrode Capacitive Deionization ferric or iron (III) (ion) Fourier-Transform Infrared Gouy-Chapman-Stern Graphene hydrogen (gas) hydrogen (ion) mercury mercury (II) oxide mercury (II) sulfate nitric acid water hydrogen peroxide Inverted Capacitive Deionization India-Canada Centre for Innovative Multidisciplinary Partnerships to Accelerate Community Transformation and Sustainability Ion-Exchange Membrane International Union of Pure and Applied Chemistry potassium hydroxide Mesoporous Carbon Membrane Capacitive Deionization Multi-Effect Distillation manganese (IV) oxide Multi-Stage Flash Distillation nitrogen (gas) sodium (ion) sodium chloride xvii  Abbreviation Definition Ni NDWC O2 OH- -OH PF PTFE PU PVA PVDF RO SEM SHE SO42- SRP TDS TOC Ti TiO2 UBC USA USD ZnO nickel National Drinking Water Conference oxygen (gas) hydroxide (ion) hydroxyl  phenolic resin polytetrafluoroethylene polyurethane polyvinyl alcohol polyvinylidene fluoride Reverse Osmosis Scanning Electron Microscope Standard Hydrogen Electrode sulfate (ion) Standard Reduction Potential Total Dissolved Solids Total Organic Carbon titanium titania or titanium (IV) oxide University of British Columbia United States of America United State Dollar zinc (II) oxide       xviii  List of Symbols Symbol Definition Unit % $ │ │ ∫ ∑ √ ∆ Λ µ (prefix) ω ata ºC C c Da d E e F F g h I i K k (prefix) kWh κ percent dollar absolute value integral summation square root change charge efficiency micro angular frequency atmospheres absolute degree Celsius concentration specific capacitance Dalton differential change potential Euler’s number (= 2.718) Farad Faraday constant (= 96,485) gram hour current observations degree Kelvin kilo kilowatt hour conductivity - - - - - - - % - s-1 ata ºC mg L-1 F g-1 - - V - F C mol-1 g h A - K - kWh S xix  Symbol Definition Unit L log ln λ M (prefix) M m m (prefix) m min n (prefix) n Pa ppm Q q R RMSD S SAC SRE s σ t v V V W WR wt. % Litre logarithm  natural logarithm equivalent conductivity mega molar mass meter milli mass minute nano number of observations Pascal parts per million charge salt adsorption per electrode mass resistance Root-Mean-Square Deviation Siemen salt adsorption capacity salt removal efficiency scan rate capacitance time volumetric flow rate Volt volume  Watt Water Recovery weight percentage L - - S cm2 mol-1 - g mol-1 g m - min - - Pa ppm C mg g-1 Ω - S mg g-1  % V s-1 F s, min L min-1 V L W % % xx  Symbol Definition Unit y ŷ Z z observed value predicted value impedance charge number - - Ω -                    xxi  Acknowledgements   I would like to give my genuine thanks to my supervisors, Dr. Madjid Mohseni and Dr. David Wilkinson, for their guidance. They were the root of many of the ideas presented in this thesis and always challenged me to look at the inner workings behind my experiments and observations. They also trusted me and gave me free rein in many aspects of my research project, and I attribute a large part of my growth in the past years as a researcher to them.  Thank you to the Dr. Mohseni’s and Dr. Wilkinson’s team for their knowledge transfer, technical assistance and collaborative environment. Adrian Serrano and Dr. Macarena Cataldo obtained the adsorption isotherm and CV results, and contributed and discussed ideas for the research project along with Dr. Amir Dehkhoda and Dr. Arman Bonakdarpour. Also, thank you to Lan Kato and Elisabetta Pani from the EOAS department for their assistance with the SEM and EDS instruments.  The CHBE Department staff has also been helpful in many tasks throughout my time spent at UBC Vancouver. These people include Richard Ryoo, Doug Yuen, Richard Zhang, Helsa Leong, Keyvan Maleki, Heidi Backous, Amber Lee, Lori Tanaka, Gina Abernethy and Marlene Chow.  The research project received funding from IC-IMPACTS, as well as indirectly from RES’EAU. These two organization also organized events from which I was able to attend to build my network and gain experience. Dr. Sathish Kumar from Eureka Forbes in India was a collaborator for this research project as part of the IC-IMPACTS program, and provided much useful advice during our periods of contact. The Faculty of Applied Science’s financial assistance through the Faculty of Applied Science Graduate Award is also much appreciated.  Finally, I sincerely thank my grandparents for teaching me the virtues of family and hard work, and so my thesis is dedicated to them. I am also lucky to have been born to gracious parents who have unconditionally supported my living during the past years. My older brother has been an example to me of being persistent with your goals, and my younger brother always pushes me to be better. I also thank the Lord Jesus Christ for supplying me with His life and placing me in the church life. To the members of the church, your portions each week supplied and encouraged me, and I am grateful for your constant cherishing and nourishing. xxii  Dedication     To my grandparents1  Chapter 1: Introduction 1.1. Drinking Water Treatment  Drinking water of sufficient quantity and quality is necessary for life. According to the United Nations Sustainable Development Goals, currently more than 40% of the world faces water scarcity and around 1.8 billion of the population drinks from a contaminated water source [1]. On a local level, even developed countries can face challenges with providing safe, accessible drinking water for communities. In the province of British Columbia, Canada, there were 528 boil water advisories in effect in 2011 [2]. In particular, small communities in Canada face numerous challenges with their drinking water because of remoteness, which leads to inadequate monitoring and treatment, inexperienced operators, poor infrastructure, and lack of funding [3].    Treatment of water is needed to provide safe and accessible drinking water for the human population. According to the World Health Organization drinking water quality standards, drinking water must meet the criteria of microbial, radiological, inorganic and organic chemical aspects, as well as be acceptable to consumers in taste, odor, and appearance [4]. Natural source waters can contain many species which are harmful to human health, from disease-causing microorganisms to toxic metals and chemicals. Furthermore, with climate change and heightened industrial activity in the past few decades, water quality standards are constantly being updated to account for emerging contaminants such as algal bloom toxins or pharmaceutical pollutants [5]. To separate these harmful elements and compounds, drinking water treatment typically involves a series of stages depending on the source water quality and the size scale of the water treatment system. A general scheme of a drinking water treatment system includes screening, coagulation and flocculation, sedimentation or dissolved air flotation, filtration, and disinfection [6].  2   Figure 1.1. Process diagram of typical drinking water treatment system.  Extra stages may be required if the source water quality warrants additional treatment, or stages may be omitted if the source water does not contain certain contaminants. Additional treatment processes include pH adjustment, softening and ion removal, membrane processes such as ultrafiltration and nanofiltration, and adsorption processes such as activated carbon among many others [7].  1.2. Current State of Desalination  Water treatment technology must evolve through innovation to meet the goal of providing safe and accessible drinking water for everyone. Desalination has emerged as an attractive solution because over 97% of the total volume of water resources on the Earth is salt water, mostly as sea water in the oceans but also in saline lakes, estuaries, mangroves, and marshes [8]. Salt water can be categorized based on the concentration of total dissolved solids (TDS). Freshwater is classified as having TDS concentrations of 500 ppm and below, brackish water has TDS concentrations ranging between 500 – 15,000 ppm, and sea water has TDS concentrations of over 15,000 ppm 3  [9]. For drinking water quality, the Canadian Drinking Water Guidelines sets as an aesthetic objective less than 500 ppm TDS, although even lower TDS levels may be more palatable to consumers [4], [10]. Different separation processes are cost-effective for treating the various salinities of salt water resources in the world. Separation Process Cost-Effective Range of TDS Concentrations [ppm] Distillation 20,000 – 100,000 Reverse Osmosis  50 – 46,000  Electrodialysis 200 – 3000  Ion-Exchange 1 – 800  Table 1.1. Desalination separation processes and TDS range for cost-effective application, adapted from [9].  Conventional desalination technologies are energy-intensive because they depend on high temperatures or pressures for the removal of ions from water. Traditional distillation-based processes such as Multi-Stage Flash Distillation (MSF), Multi-Effect Distillation (MED) and Vapor Compression (VC) can require up to 15.0 kWh/m3 of total energy if waste heat is not utilized, corresponding to a cost of 4 $USD/m3. Reverse Osmosis (RO), the predominant membrane-based process used for desalination, also suffers from high total energy requirements of up to 4.0 kWh/m3, or 3 $USD/m3, because of high pressures needed to drive water through the semi-permeable membrane [9]. The energy and cost of desalination also varies considerably because it is heavily dependent on the size of the plant. Large plants that can benefit from economies of scale and spreading overhead costs can have water production costs as cheap as 0.50 $USD/m3 for desalination of sea water, whereas smaller plants generally cost more than 1.30 $USD/m3 [11]. 4   MED MSF VC RO (Brackish Water) RO  (Sea Water) Steam Pressure [ata] 0.2 – 0.4  2.5 – 3.5 n/a n/a n/a Electrical Energy Equivalent [kWh/m3] 4.5 – 6.0 9.5 – 11.0 n/a n/a n/a Electricity Consumption [kWh/m3] 1.2 – 1.8 3.2 – 4.0 8.0 – 12.0 0.3 – 2.8  2.5 – 4.0 Total Energy Use [kWh/m3] 5.7 – 7.8 12.7 – 15.0 8.0 – 12.0 0.3 – 2.8 2.5 – 4.0 Water Production Costs [$USD/m3] 0.7 – 3.5 0.9 – 4.0 1.0 – 3.5  0.2 – 1.8 0.5 – 3.0 Table 1.2. Energy requirements and costs of conventional desalination technologies, adapted from [9].  For the operation and maintenance of conventional desalination technologies, chemicals are also required for cleaning and conditioning of the source water, and are therefore another cost that should be accounted for. For RO, operation and maintenance can be especially expensive because the membranes tend to foul easily and thus warrant frequent chemical cleaning and replacements that can be costly because of material and manufacturing costs [11].  Despite these weaknesses, desalination is dominated by RO with a global capacity share of 64%, followed by MSF with 23%, MED with 8%, and electrochemical and other technologies with a 5% share [12]. According to the International Desalination Association, as of 2015, there were 18,426 desalination plants worldwide treating a capacity of 86.8 million m3/day of water for more than 300 million people in 150 countries [13].  5  1.3. Capacitive Deionization as an Alternative Desalination Technology  Capacitive Deionization (CDI) is an incipient technology which has shown promise to be cost-effective for the desalination of brackish water with TDS concentrations between 200 – 5,000 ppm [14]. In CDI, a direct current (DC) voltage is applied to a cell consisting of a pair of electrodes, like a capacitor (hence the name). Saline water is passed through the cell via an insulating spacer to prevent short circuiting and ions are adsorbed onto the charged electrode surfaces; this process is coined as electrosorption to describe the applied voltage as the driving force for adsorption. Once the electrode surfaces are saturated with ions, the applied voltage can be removed or reversed in polarity to desorb the ions. The adsorption (charge) and desorption (discharge) steps can then be alternated to generate purified water and concentrated brine, respectively. Multiple CDI cells can be placed on top of each other in a stack to increase the capacity [15].   Figure 1.2. Schematic representation of the basic principle of CDI separation.  6  The higher the applied voltage, the higher the ion electrosorption and consequential removal. However, low applied voltages below 1.23 V are usually used to prevent electrochemical reactions such as the electrolysis of water which decreases the charge efficiency (the total charge of the adsorbed salt divided by the total charge supplied to the CDI system) of the desalination process [16]. The ion electrosorption phenomenon can be described by the formation of electrical double layers (EDLs) at the interface of an electrolyte solution and a charged surface. In the simplest EDL model, the Helmholtz model, counter-ions (ions with an opposite charge to that of the surface) are attracted to, and adsorb onto the surface to compensate for the charge [15].  Figure 1.3. EDL model showing surface and solution interfacial interactions.  More complex EDL models such as the Gouy-Chapman-Stern model and the modified Donnan model have also been extensively researched. These EDL models are typically combined with models describing ion transport through pores to predict the desalination performance of CDI systems [17]. For optimal desalination performance in CDI, it is essential that electrodes have large 7  surface areas, high electrical conductivity, excellent wettability, high  mechanically stability, chemical and biological inertness, and low cost [18]. Carbon materials meet many of these criteria, and thus studies have been performed for activated carbons, activated carbon cloths, activated carbon fibers, carbon aerogels, mesoporous carbons, carbon nanofibers, carbon nanotubes, graphene and etc. [19].  Compared to RO, the present dominating desalination technology, CDI does not use high pressures, instead it uses low applied voltages to separate ions from water. Moreover, it is possible to recover energy from CDI systems during the discharge step since the CDI cell essentially behaves as a capacitor. For salinities below 2000 ppm TDS, it was found that less than 0.5 kWh/m3 was needed to produce 1 ppm TDS of purified water. Feasibility studies have shown that CDI can be competitive with RO at concentrations below 5000 ppm TDS at energy efficiencies of 60-70%, and can even be competitive for sea water desalination at energy efficiencies above 85% [11]. Therefore, CDI can potentially consume less energy for desalination compared to RO systems, leading to cost savings [20]. One of the major reasons CDI has shown potential over RO is that higher water recoveries can be attained. RO systems generally have water recoveries of less than 50%, whereas CDI systems can have water recoveries upwards of 90%. As a result, less concentrate is generated as waste and it can be more easily disposed of through brine disposal methods such as evaporation ponds [14].  The other main competing technology for CDI desalination of brackish water is Electrodialysis (ED), which is cost-effective in a similar TDS concentration range as CDI. ED is an electrochemical technology for desalination that also applies a voltage across electrodes in a cell; however, ions migrate toward the opposite charged electrode through cation and anion exchange membranes into concentrate streams to separate them from the purified water stream, 8  instead of being directly electrosorbed onto the electrodes [21]. There are many limitations which ED faces as a desalination technology which may not apply to CDI. Firstly, ED is unsuitable for low salinities below 400 ppm TDS due to low electrical conductivity, unless large electrode and membrane areas are used. Also, low molecular weight charged species are preferentially separated, leaving a larger fraction of the high molecular weight charged species in the purified water. Ion exchange membranes can also foul and pretreatment of the influent water is required to remove species which precipitate onto or form coatings on the membranes [11]. Lastly, CDI has slightly better water recoveries, although both can reach over 90% [22].  Despite considerable research in CDI, there has been difficulty transforming research findings into real-world applications. In recent years, there have been some developments and applications of CDI in the industry as a water solution. In the 2000s, several companies commercializing CDI emerged but some have abandoned the market or went bankrupt because of lack of market identification and technical challenges [14]. Nonetheless, there are companies which remain or have appeared in the current decade that supply CDI technology, most still in the research or start-up phase. A few of these companies directly sell water purification products with CDI technology to consumers. As of 2017, Pureechem supplies the Ecomite series, a line of three CDI products ranging from a single cell with a capacity of 0.03 L/min and 1000 ppm TDS to a system with multiple stacks capable of handling 2 L/min and 3000 ppm TDS [23]. Aqua EWP also supplies products with CDI technology for household use as well as small systems with a maximum treatable level of 3000 ppm TDS [24]. Other companies commission the design of their CDI technology. Atlantis Technologies provides a system called Radial Deionization that can process high salinity influents by allowing water to flow across up to 10 m of continuous electrode material [25]. Enpar Technologies calls its CDI technology Electro-Static Deionization and has 9  stated that they are commencing the design and procurement phase of a large desalination plant for Chemsbro in Saudi Arabia [26]. They have also had experience with ammonia, nitrate, hardness and TDS removal from industrially contaminated ground waters, surface waters and mining waste waters from countries in the Middle East and the Canadian provinces of Ontario and Quebec [27]. Lastly, Voltea’s CapDI technology can be commissioned for industrial, commercial and residential applications, and can deliver up to 537 m3/day of purified water [28]. Name Location Website References Enpar Technologies Guelph, Ontario Canada http://www.enpar-tech.com/  [29] Aqua EWP San Antonio, Texas U.S.A. http://aquaewp.com/ [30]–[32] Atlantis Technologies Dana Point, California U.S.A. http://www.atlantis-water.com/ - Pureleau North Saanich, British Columbia Canada http://www.pureleau.ca/ - Pureechem Cheongju, Chungcheong South Korea http://www.pureechem.com/en/ [33] Siontech Yuseong, Daejon South Korea http://www.siontech.co.kr/eng/ - Voltea Sassenheim Netherlands Farmers Branch, Texas U.S.A. http://voltea.com/en/ - Idropan Milan Italy http://www.idropan.com/en/ - Table 1.3. List of active vendors for CDI technology.   CDI pilot plants have been tested in the field in previous studies, although there is a strong demand for more pilot and field studies to evaluate their feasibility. AlMarzooqi et al. claimed in 2014 that there have only been five demonstrations of CDI plants with the largest capacity being 3,785 m3/day compared to that of a typical RO plant with capacity 100,000 m3/day, and highlighted 10  the need for more studies outside the laboratory [12]. For the successful transition of CDI technology from laboratory to the field, the robustness to source water of various qualities, and the stability of its performance over time must be determined and improved upon. Also, a stronger understanding of how to design and operate CDI systems to optimize performance and efficiency holistically must be realized and communicated for CDI to bridge the gap from research to practice. 1.4. Research Objectives  To contribute to the advancement of CDI as a practical and viable desalination technology, this research project focused on studying the design and operation of a laboratory scale CDI cell with activated carbon electrodes using synthetic NaCl solution. The results would serve as a preliminary step and baseline for further research and development of CDI technology for point-of-use water systems or small water systems. Although there have been studies on the effect of operating parameters [16], [30], [34], electrode properties, and fabrication procedures [35], [36], the results are sometimes contrary and cannot be compared with one another because of differences in electrode materials, CDI experimental setups, and methodologies. Moreover, there are only a few studies on the inefficiencies of the CDI process and how they relate to desalination performance degradation [31], [37]. Therefore, with the aim to provide a holistic approach to investigating CDI systems, the following research objectives were formulated: 1. Characterize the properties and evaluate the desalination performance and efficiency of activated carbon electrodes in a laboratory scale CDI cell using synthetic NaCl solution. 2. Assess the effect of operating parameters on CDI desalination performance and efficiency using a range of metrics and models. 3. Analyze for inefficiencies and degradation of performance in CDI systems by operating continuously for multiple cycles. 11  Chapter 2: Literature Review 2.1. Defining Desalination Performance and Efficiency  To evaluate the feasibility of CDI technology, reliable metrics or indicators must be developed for desalination performance and efficiency. This section describes and comments on the various metric definitions of desalination performance and efficiency. 2.1.1. Salt Removal Efficiency and Water Recovery Traditionally, salt removal efficiency (SRE) has been utilized to assess desalination technologies, which is defined as: 𝑆𝑅𝐸 [%] =  𝐶𝑖− 𝐶𝑓𝐶𝑖 𝑥 100%       (1) where Ci [mg/l] is the initial or influent concentration and Cf [mg/L] is the final or effluent concentration. Another metric that is common to all desalination technologies is water recovery (WR), which is a measure of the volume of purified water as a percentage of volume of feed water inputted into the system. Mathematically, this is described as:  𝑊𝑅 [%] =  𝑉𝑓𝑉𝑖 𝑥 100%        (2) where Vi [L] is the volume of feed water and Vf [L] is the volume of purified water. In desalination, concentrate is generated as waste, and it is beneficial to increase the WR of the system to improve the water production efficiency, particularly in water scarce areas. Indeed, although membrane-based technologies can have high SRE, their WR is typically in the range of 40 – 55%, which means that approximately for each cup of water produced, another cup is wasted. Therefore, there has been a demand for emerging desalination technologies with high WR such as CDI with reports of 12  over 90% recovery [38]. SRE and WR are valuable in comparing CDI technology to other desalination technologies because they are simple and use the same variables for their calculation. However, the size and scale of the desalination systems must be similar for proper comparison using SRE and WR, since larger and more sophisticated systems have higher SRE and WR [39].  2.1.2. Specific Capacitance  Because of the limited insight of SRE and WR into the desalination performance of CDI systems, further metrics have been developed and implemented. Since a major component of CDI systems is the electrode, many of these metrics pertain to the electrode performance. CDI shares many commonalities with supercapacitors, and thus performance metrics from the supercapacitor field have been utilized for CDI electrodes. A universal metric for supercapacitors is specific capacitance (c), which is an indicator of the charge storage capacity normalized by the mass of the electrode. This can be measured with electrochemical techniques such as galvanostatic charge-discharge cycles, cyclic voltammetry (CV), and electrochemical impedance spectroscopy (EIS). Galvanostatic charge-discharge cycles involve charging a two-electrode setup (working electrode and counter electrode) to a desired voltage, and then discharging at a constant current. Using the obtained data, c can be calculated from equation (3), or equation (4) if the change in potential vs. the change in time is treated linearly.  𝑐 [𝐹/𝑔] =  𝐼𝑚 𝑑𝐸𝑑𝑡          (3)  𝑐 [𝐹/𝑔] =  𝐼 ∆𝑡𝑚 ∆𝐸=  𝑄𝑚 ∆𝐸        (4) Here, I [A] is the discharge current dE/dt [V/s] is the slope of the discharge curve, Q [C] is the overall charge released during discharge, ∆E [V] is the decrease in potential during discharge, ∆t 13  [s] is the time of the discharge period and m [g] is the mass of a single electrode [40]. In CV, the current is recorded while linearly sweeping the cell voltage over a range for multiple cycles in a three-electrode setup (working electrode, counter electrode, and reference electrode). Equation (5) can then be used to find c, assuming the working electrode is an ideal capacitor.  𝑐 [𝐹/𝑔] =  𝐼𝑖𝑛𝑠𝑡𝑚 𝑑𝐸𝑑𝑡=  𝐼𝑖𝑛𝑠𝑡𝑚 𝑠        (5) Here, Iinst [A] is the instantaneous current and s [V/s] is the scan rate. Since Iinst requires ideal capacitor behavior which is rarely seen in real world systems, the average current is generally used instead in the following equation: 𝑐 [𝐹/𝑔] =  1𝑚 𝑠 ∆𝐸  ∫ 𝐼 𝑑𝐸        (6) where I [A] is the current as a function of potential and ∆E [V] is the change in potential of the half cycle [40], [41]. The integral is either taken over the half cycle, or over the full cycle and multiplied by ½ if the graph is asymmetrical over the potential axis. Lastly, EIS can also measure specific capacitance and moreover, indicate whether the electrodes behave like an ideal capacitor. In EIS, a small perturbation is caused by imposing an alternating current (AC) signal, and the voltage and current response of the system is measured. From the obtained data, the impedance, which is a complex representation of the ability of a circuit to resist the flow of current from an applied voltage, can be found. Further modelling with equivalent circuits can then yield information on the capacitance, as well as inductance and resistance, of the system. Nyquist plots are usually used to graphically present the real and imaginary components of the impedance, with the real component corresponding to circuit elements like resistors and the imaginary component 14  corresponding to circuit elements like capacitors and inductors. By making certain assumptions in EIS, the calculation to find c can be simplified to equation (7):  𝑐 [𝐹/𝑔] =  1𝑚 𝜔 |𝑍′′|        (7) where Z’’ [Ω] is the imaginary part of the impedance and ω [s-1] is the angular frequency [40], [42]–[44]. The three electrochemical techniques can produce different results, and thus their respective assumptions must be taken into careful consideration. 2.1.3. Salt Adsorption Capacity  Salt adsorption capacity (SAC) is a widespread metric specifically for CDI and has become nearly universal for studies of CDI electrodes. In sum, SAC describes the mass of salt adsorbed normalized by the mass of electrode. For CDI experiments, the amount of salt adsorbed can be calculated from the decrease in salt concentration and the volume:  𝑆𝐴𝐶 [𝑚𝑔/𝑔] =  𝑣 ∫(𝐶𝑖 − 𝐶𝑓) 𝑑𝑡𝑚 =  𝑉(𝐶𝑖 − 𝐶𝑓) 𝑚      (8) where v is the volumetric flow rate [L/min], V [L] is the volume of salt solution, Ci [mg/L] is the initial salt concentration and Cf [mg/L] is the final salt concentration. The integral form of the equation must be used if the volume of salt solution in the CDI system is not constant with time, such as in single pass experimental setups. The valid measurement of SAC requires that the charge step is long enough for the final salt concentration to reach equilibrium. Moreover, to standardize the SAC metric, it is proposed that synthetic NaCl solution is used for the CDI experiment, and the quantity used for electrode mass is the combined mass of both electrodes including the binder and additives, but without the spacer [39]. If other salt solutions are tested, it is recommended to express SAC in units of mmol/g to account for the differences in molar mass [45]. SAC is an 15  indicator of the propensity of the material for electrosorption because it is irrespective of the mass and is thus convenient for comparing the desalination performance of different electrode materials. It is also effective for evaluating the cost-effectiveness of materials, particularly because materials are often priced based on their mass. 2.1.4. Charge Efficiency  To shed light on the energy efficiency, charge efficiency (Λ) has been created as a performance metric unique to CDI systems. Λ is mathematically defined as:  𝛬 [%] =  𝑧 𝐹 𝑣 ∫(𝐶𝑖 − 𝐶𝑓) 𝑑𝑡𝑀 ∫ 𝐼 𝑑𝑡 𝑥 100% =  𝑧 𝐹 𝑉 (𝐶𝑖 − 𝐶𝑓) 𝑀 ∫ 𝐼 𝑑𝑡 𝑥 100% (9) where z is the charge number of the ions in the salt, M [g/mol] is the molar mass of the salt, F is the Faraday constant and equal to 96,485 C mol-1 and I [A] is the current supplied to the CDI cell [16]. In an ideal CDI process, for each unit of charge supplied to the CDI cell, one unit of equivalent charge of salt is adsorbed, leading to a Λ of 100%. In practice, however, inefficiencies such as co-ion desorption, resistive heating, and parasitic electrochemical reactions can occur, resulting in Λ of less than 100%. Energy consumption is strongly affected by Λ, and CDI systems are therefore operated at low voltages, despite increased SAC with increased voltages, to maximize Λ [46].  2.1.5. Specific Energy Consumption To compare the energy requirements of CDI with pre-existing desalination technologies, the specific energy consumption (η) is used to describe the amount of energy needed to remove an amount of salt from the saline water. For CDI systems, this is mathematically expressed as 16   𝜂 [𝑘𝑊ℎ/𝑔] =  𝐸𝑐𝑒𝑙𝑙  ∫ 𝐼𝑑𝑡𝑣 ∫(𝐶𝑖 − 𝐶𝑓) 𝑑𝑡=  𝐸𝑐𝑒𝑙𝑙  ∫ 𝐼𝑑𝑡𝑉 (𝐶𝑖 − 𝐶𝑓)    (10) where Ecell [V] is the voltage applied to the CDI cell and the other variables are as defined previously. η can also be described on the basis of purified water volume produced by using purified water volume instead of amount of salt in the denominator of equation (10) [47].  2.1.6. Model Parameters  Desalination performance and efficiency metrics can also be derived from model parameters that provide insight into the principles and mechanisms of the CDI system. Modelling of electrosorption for CDI systems is discussed in detail in Section 2.5.  2.2. Electrode Materials and Processing  Electrode materials for CDI technology have been researched extensively in the past decades, and there have even been multiple review papers dedicated solely to evaluating the properties and characteristics of novel electrode materials [18], [19], [48].  2.2.1. Ideal Electrode Properties for Electrosorption Carbon materials have displayed excellent applicability as CDI electrodes because of large surface areas due to their porosity, high electrical conductivities, and potentially low material and processing costs [15]. A larger surface area means that there are potentially more adsorption sites for ions to adsorb and EDLs to form. The most widely used metric for surface area of porous materials is the Brunauer-Emmett-Teller (BET) theory. In the BET theory, it is assumed that the adsorbate (the adsorbed species) can either be adsorbed onto the adsorbent (the species the adsorbate adsorbs onto) or onto other adsorbates. Using noble gas adsorption isotherms and 17  making further assumptions about adsorbate interactions, the BET surface area can be calculated in units of m2/g [49]. Surface area cannot fully describe the electrosorption efficiency because pores may be inaccessible to ions or unsuitable for EDL formation. Indeed, it has been found that BET surface area does not always have a direct relationship with salt adsorption capacity [17]. Carbon materials can have a variety of pore sizes and structures which can facilitate ion transport and formation of EDLs. According to the International Union of Pure and Applied Chemistry (IUPAC), pore sizes can be roughly divided into three categories. By definition, macropores are pores with diameters greater than 50 nm, mesopores are between 2 to 50 nm, and micropores are less than 2 nm [50]. Pore size and structure are crucial to electrosorption because they affect ion transport and the formation of EDLs. On one hand, smaller pores produce a greater surface area because of the surface area to volume ratio. On the other hand, if the pore sizes begin to approach the size of the ion diameters in solution, there is the possibility for EDL overlap and thus reduction in electrosorption efficiency [51]. In a study based on activated carbons with various BET surface areas and pore characteristics, it was found that mesopores had the greatest contribution to electrosorption efficiency because of the accessibility of ions to the surface adsorption sites [52]. On the contrary, there have also been theoretical and experimental studies that show the correlation between micropores and electrosorption is greater than that of mesopores [17]. It is likely, therefore, that pore structure, such as whether they are open or closed or connected, is also an important factor and a combination of mesopores and micropores is the optimal structure. Pore structure, however, is rarely studied because it requires 3D nano-scale imaging, which is currently an unestablished technique.  18  2.2.2. Activated Carbon Powder Electrodes 2.2.2.1. Activated Carbon Powder Activated carbon powder (ACP) is arguably the most cost-efficient material for electrodes because it is readily available and inexpensive. With BET surface areas of around 1000 – 2000 m2/g, ACP is commonly mixed with a polymeric binder in a solvent to form a slurry and placed onto a current collector to form CDI electrodes [53]. ACP is produced by subjecting carbon resources such as coal, peal, lignite, wood, sawdust, bamboo, and coconut shells to either physical activation at high temperatures of above 500 ˚C with steam or other gases, or chemical activation with reagents such as acids, bases, or salts [54]. Due to the multiple possible raw materials and activation methods, ACPs can vary in quality, and therefore suitable ACPs should be chosen based on their surface area, pore size and structure, and other relevant properties. ACP electrodes can be treated to increase their electrosorption efficiency by modifying their properties. By physical activation under flowing CO2, the surface area of the ACP could be increased at the expense of increased hydrophobicity, translating to an improved specific capacitance and salt adsorption capacity of the electrode [55]. Also, physical activation with H2 at 400 ˚C increased the surface area and specific capacitance of ACP electrodes, whereas an atmosphere of N2 decreased the surface area and specific capacitance [56]. Another study found that treating ACP with HNO3 at 90 ˚C resulted in more oxygen containing functional groups on the electrode surface, leading to increased salt removal efficiency and electrosorption kinetics because of more adsorption sites and improved hydrophilicity [57]. Chemical activation of ACP with KOH at temperatures over 100 ˚C increased the number of –OH groups and therefore the wettability of the electrodes by decreasing the contact angle between water droplets and the electrode surface. Furthermore, KOH activation could affect the pore size and structure by tuning the mesopore and micropore ratios of the ACP. 19  Overall, this amounts to increases of salt removal efficiency by 13 – 20% and salt adsorption capacity of up to 9.72 mg/g [58], [59].  2.2.2.2. Polymeric Binders, Conductive Additives, and Current Collectors The polymeric binder serves to give structural integrity to the ACP and adhere it onto the current collector. Common binders are polytetrafluoroethylene (PTFE) or polyvinylidene fluoride (PVDF). Since large amounts of binder can decrease the surface area by blocking access to the pores, the minimum amount that provides enough mechanical strength to the electrode is generally used. This is around 5 wt. % (dry weight, not including the solvent) for PTFE and 10 wt. % for PVDF [36], [60]. PTFE and PVDF have the drawback of being hydrophobic and can therefore reduce wettability, a desired property because water needs to have close contact with the surfaces. Also, other weaknesses are that organic solvents must be used to dissolve the binder and their rigidness can generate cracks on the electrodes [61]. New binders have been studied because of the limitations of PTFE and PVDF. Polyvinyl alcohol (PVA) is hydrophilic because of the presence of multiple –OH groups and it has been found that water droplets are completely immersed in ACP electrodes made with PVA, as opposed to PTFE and PVDF bound electrodes in which a contact angle is formed between the water droplet and the surface [62]. However, hydrophilic binders tend to swell and can dissolve in aqueous solutions. Other binders that have been tested in CDI systems include carboxymethyl cellulose (CMC), phenolic resin (PF), and polyurethane (PU) [61]–[63].     20  Binder Advantages Disadvantages References PTFE Thermal stability, chemical resistance, uniform dispersion in water as a solvent and on the electrode Blocks access to pores, defects can occur on electrode [62] PVDF High mechanical strength, thermal stability, chemical resistance Some blocking of pores, increases electrical resistance, cracks can form on electrode, requires organic solvent to dissolve [62] PVA Excellent wettability, soluble in water as a solvent Weak thermal stability and chemical resistance, swelling of electrode [62] CMC Mechanical strength, good wettability, soluble in water as a solvent Requires additive and reaction for crosslinking, some degradation of electrode [64] PF High mechanical strength, good wettability Requires high temperature and pressure to make electrode, blocks access to pores, increases electrical resistance [62] PU High adhesion strength, flexible electrode structure, low electrical resistance, stable desalination performance Some blocking of pores, poor wettability, complex synthesis [61] Table 2.1. Advantages and disadvantages of different binders for CDI electrodes.  Additionally, conductive additives like carbon black and graphite powder may also be added to improve the electrical conductivity of the AC electrode. Empirical studies have found that 12 wt. % of carbon black is the optimal amount for highest specific capacitance [60]. Other studies have found that there was no statistically significant difference for the salt adsorption capacity of ACP electrodes prepared with varying amounts of carbon black [16]. It is likely that the type of carbon black used determines the effect on the desalination performance of the CDI system. This is supported by a study displaying that mesoporous conductive carbon black, which contain a higher fraction of mesopores, increased the specific capacitance of the electrode by over three times [65]. Polyaniline, a highly researched conducting polymer, has also been tested as a conductive additive 21  for CDI electrodes, and it was found that the conductivity (the inverse of the resistivity) was increased by more than four times [66]. The current collector is an electrically conductive material selected such that it is electrochemically stable in the working applied voltage range of the CDI system. Graphite foil, sheet, or paper is known to be resistant to corrosion and is thus a normally used current collector for ACP electrodes [67], [68]. Other current collectors include stainless steel plate or meshes, carbon fiber paper, Ti foils, and Ni foam or meshes [35], [61], [69]–[72]. 2.2.2.3. Fabrication Procedures   After making the slurry of ACP, binder, and/or conductive additive in a suitable solvent, the electrode is fabricated by casting or coating onto the current collector. Common coating methods include roll coating and spray coating. In roll coating, the slurry is added to the current collector and passed between two rollers to produce compressed electrodes with a uniform thickness. In spray coating, the slurry is filled into a nozzle and sprayed onto the current collector at a specified pressure. Another simple method, called evaporative casting, is to place the current collector on a heated glass plate, drop a certain amount of the slurry onto the current collector, and allow the solvent to evaporate. The casting or coating method can affect the characteristics and properties of the CDI electrode. According to Lu et al. [35], electrodes fabricated by roll coating and spray coating had denser, more compact structures which made the materials fall off more easily. However, they found that the evaporative casting method resulted in a loose structure and many cracks on the electrode surface which reduced surface stresses, improved durability, and increased electrode surface area and wettability. A similar method to evaporative casting is doctor blade or film applicator casting, which uses a tool to cast the slurry at a specified wet thickness, usually around 200 – 300 µm [72]. After casting or coating, the electrode is dried at a temperature near the boiling point of the solvent, sometimes under vacuum, until all the solvent is evaporated 22  [45]. For the thickness of the electrode, it has been shown that increasing thickness slows down the rate at which maximum desalination is reached, i.e. when the electrodes are saturated [17]. Furthermore, it has been found that the desalination performance can be improved significantly by varying the electrode thickness asymmetrically, such as by doubling or tripling the anode [73]. There have also been some contrary results which show that salt removal kinetics were not affected by changes in electrode thickness, and the amount of salt removed does not increase linearly [16]. All in all, optimal electrode thickness requires balancing of salt adsorption capacity, electrosorption kinetics, and material costs. 2.2.3. Other Materials and Nanotechnology  An assortment of other materials has also been applied as CDI electrodes. To name a few, activated carbon cloth (ACC), activated carbon paper, carbon felt, carbon veil, carbon foam and carbon aerogel (CA) are all commercially available and have been evaluated as CDI electrodes. Nonetheless, ACP electrodes have exhibited superior properties for salt adsorption capacity, specific capacitance, conductivity, and wettability [16], [60]. Activated carbon fiber (ACF), which is known for its high mechanical strength and thermal stability, is another prospective CDI electrode material which can be prepared via electrospinning of carbon precursors. If the fibers are made such that their size is at the nano-scale, they are commonly called carbon nanofibers (CNFs) [19]. Both ACFs and CNFs have been successfully tested for CDI electrodes, although their applications are limited by production costs [48]. Since desalination performance has been known to be affected by the pore size distribution of the electrosorption material, mesoporous carbon (MC) has been synthesized and prepared as CDI electrodes. MCs are produced by either the hard-template method, where the carbon precursors are attached to templates with 3D pores structures and the template is removed after carbonization, or the soft-template method, where organic 23  surfactants or block copolymer templates are directly assembled with the carbon precursors. In both methods, the synthesis and preparation can be costly, time-consuming and thus impractical for CDI systems [74]. With the advent of nanotechnology, CDI researchers have also studied incorporating nanomaterials into CDI electrodes either as is or as a composite for ACP. Carbon nanotubes (CNTs) are typically produced through chemical vapor deposition (CVD) and are composed of mostly mesopores with large surface area and excellent electrical conductivity. However, they are hydrophobic and tend to agglomerate, which can block access to surface adsorption sites [75]. Graphene (GR) is another nanomaterial, which can also be produced through CVD, that theoretically has large surface areas of 2630 m2/g and electrical conductivities of 7200 S/m [76]. The challenge with nanotechnology is that the material and production costs are too great to justify their use in CDI electrodes, unless they are composited in small amounts with ACP. Moreover, it is difficult to produce nanomaterials with reliable quality in terms of being uniform and defect-free [77]. Nevertheless, the maturation of novel materials and nanotechnology may bring about exciting opportunities for CDI systems in the future. An overview of the various electrode materials and their performances are presented in Table 2.2. A major part of the development of electrodes for CDI is based on and reflects research in the field of electrical double layer capacitors (EDLCs), or supercapacitors, because of the many similarities of the technologies. Electrodes with high energy storage capacity generally also have high desalination performances. However, CDI systems desalinate water as their primary goal as opposed to storing energy in EDLCs, and therefore also require excellent wettability and ion transport [39]. Furthermore, since there are water flows, systems must be designed to optimize solution transport to the electrodes in the CDI cell.  24  Material Salt Concentration (mg/L) Applied Voltage (V) SAC (mg/g) Reference ACP                 ACP / MnO2 ACP / TiO2  ACP / GR 59 200 200 292 25 292 292 292 1170 25 50 100 200 500 1000 1500 2000 25 100 584 500 1.2 1.5 1.5 1.2 1.4 1.2 1.4 1.2 1.4 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 2.6 3.7 5.3 (membrane) 10.5 6.1 6.9 8.4 10.9 13.0 0.25 0.27 6.10 8.00 9.72 10.80 11.00 11.76 0.99 8.05 17 2.94 [36] [78]  [79] [80]   [81]  [82] [83] [84]      [85] [86] [87] [88] ACC   ACC / TiO2  ACC / ZnO 550 5500 5.85 5.844 5844 100 1.1 1.1 1.0 1.0 1.0 1.2 10.0 7.7 1.75 4.68 4.3 8.5 [89]  [90]  [91] [92] ACF   ACF / CNF CNF   CNF / CNT 60 500 1000 25 90 95 45 10 50 100 1.2 1.2 1.2 1.2 1.6 1.6 1.2 1.2 1.2 1.2 1.7 2.57 3.71 17.19 9.13 4.6 1.91 3.32 1.61 3.87 [93]   [94] [95] [96] [97] [98] [99] 25  Material Salt Concentration (mg/L) Applied Voltage (V) SAC (mg/g) Reference CA 50 500 2922 58.5 140 50 50.8 50.8 2000 1.2 1.2 1.5 1.2 1.2 1.2 1.5 1.7 1.3 1.4 2.9 9.6 1.34 4.51 3.33 2.81 3.76 7.0 [100]  [101] [102]  [103]  [104] [105] MC  MC / CNT  MC / GR 25 50 40 40 45 34 1.2 0.8 1.2 1.2 2.0 1.6 0.68 0.93 0.63 0.69 0.73 2.3 [82] [83] [106]  [107] [108] CNT       CNT / GR 3000 23 60 500 1000 1500 2000 27 50 770 1.2 1.2 1.2 2.0 1.2 1.2 1.2 2.0 1.6 2.0 1.7 1.3 0.7 2.57 3.71 4.76 5.24 1.41 0.88 26.42 [109] [76] [93]  [110]   [111] [112] [113] GR       GR / TiO2 25 25 250 65 30 53 25 500 6000 2.0 2.0 2.0 2.0 1.6 1.6 1.2 1.2 1.2 1.8 1.36 8.6 3.23 2.26 2.9 6.18 15.1 24.2 [114] [115] [116] [117] [118] [119] [120] [121] Table 2.2. Summary of operating parameters and desalination performance of materials for CDI electrodes. Slashes indicate composite materials. Compiled from sources: [15], [18], [74]. 26  2.3. Cell Architecture 2.3.1. Flow-by and Flow-through Geometries  The most prevalent cell architectures for CDI technology are planar electrodes with flow-by and flow-through geometries. In both these geometries, electrodes are placed parallel with a gap in between, usually formed by an insulating spacer to prevent short circuiting. The gap is around 100 – 300 µm if a spacer is used or greater than 1 mm if left as an open channel. In the flow-by geometry, water flows into the planar gap which acts as a channel between the electrodes. On the other hand, in the flow-through geometry, water flows straight through the planar gap and the electrodes, which must be permeable to water [15]. There are advantages and disadvantages for both geometries. Flow-through geometries have faster rates of desalination because the water comes into intimate contact with the electrodes and allows the spacer thickness to be minimized, leading to increased c and reduction of cell volume [101]. However, there may be the problem of resistance to the solution flow which may lead to lower flow rates and hence smaller capacities for water treatment [122]. Additionally, flow-by geometries have been found to be more stable over time than flow-through geometries, likely due to the lower diffusion rate of dissolved O2 into the electrodes, resulting in less oxidation of the carbon functional groups [37].    27   Figure 2.1. Flow-by (a) and flow-through (b) geometries for CDI systems.  2.3.2. Membrane Capacitive Deionization  Ideally, the CDI electrosorption process involves only counter-ions adsorbing onto the electrode. In reality, other phenomenon can also occur when a voltage is applied. Co-ions, ions with the same charge as that of the electrode surface, can also be expelled to balance out the charge from the applied voltage. Consequently, the charge efficiency of the electrosorption process is compromised because each unit of charge transported is not directly converted to one unit of ions adsorbed onto the electrode [123]. To mitigate co-ion expulsion and reduced charge efficiencies, it was found that an ion-exchange layer could be incorporated on the electrode; this modified process is called Membrane Capacitive Deionization (MCDI). The ion-exchange layer serves to selectively allow only counter-ion transport while blocking co-ion transport [124]. Most commonly, the ion-exchange layer is applied as an ion-exchange membrane (IEM) which is stacked between the electrodes and the spacer [78]. However, another method is to coat the electrode with a polymer containing ion-exchange functional groups to act as the ion-exchange layer, resulting in a lower contact resistance between the ion-exchange layer and the electrode 28  [125]. Ion-exchange layers are generally used only for flow-by CDI systems because water flows through one electrode at a time in flow-through systems. Experimental studies have shown that compared to CDI cells, MCDI cells can greatly enhance the desalination performance by increasing salt adsorption capacity by 27 – 56% and charge efficiency by 69 – 95%, depending on the operating conditions of the system [68], [78].  Figure 2.2. Prevention of co-ion expulsion with an anion-exchange membrane (AEM) and cation-exchange membrane (CEM) in MCDI.  2.3.3. Flow-Electrode Capacitive Deionization and Inverted Capacitive Deionization Innovative cell architectures have also been proposed and studied. One interesting study displayed that CDI desalination can be achieved with rigid carbon wire electrodes [58]. The wire electrodes can be movable to allow for separation of purified water and concentrated water in a CDI cell without the need for charge and discharge cycles [126]. Another creative cell architecture that has received considerable attention is flow-electrode capacitive deionization (FCDI). Instead of using rigid electrodes, a suspension of carbon materials in an electrolyte solution is used as the 29  electrodes. The electrodes can be flowed through the CDI cell, constantly replacing the saturated carbon materials for continuous electrosorption and removal of ions, and therefore no discharge step is needed [127]. Results have shown that the flow electrode CDI process is able to effectively desalinate high salinity feeds, which may bring in applications of CDI technology for sea water desalination [128]. However, FCDI systems can be significantly more complex to design and operate, and is currently in its infancy as a technology since there have been no pilot or field studies and commercialization attempts to the best of the author’s knowledge. Another inventive CDI cell architecture is inverted capacitive deionization (i-CDI). In i-CDI, the electrodes are chemically modified by introducing negative functional groups to the anode through acid treatment or sulfonation, and positive functional groups to the cathode through amination [129]. Resultantly, the anode has a net negative surface charge while the cathode has a net positive surface charge, causing EDLs to spontaneously form without an applied voltage. This results in an inverted operation where ions desorb during the charge step and adsorb during the discharge step. The advantage of i-CDI is that it counteracts performance degradation due to anode oxidation, and thus has better long-term stability. Nevertheless, more research work is necessary to increase the working range and assess the desalination performance of i-CDI systems [130]. 2.4. System Operation  Conventional CDI systems have a charge step, where a voltage is applied and ions are adsorbed in the cell, and a discharge step, where the voltage is reversed in polarity or short circuited to desorb the ions from the saturated electrodes. This occurs while salt water is passed through the CDI cell. The operation of conventional CDI systems is affected by many parameters such as applied voltage, salt concentrations, flow rate, temperature, and etc. These operating parameters which affect the CDI system change depending on the water treatment application and can 30  sometimes be varied to achieve the desired desalination performance or optimize water production costs. Additionally, there are novel operation modes, such as applying a constant current instead of a constant voltage, which have emerged to make CDI a diverse desalination technology. 2.4.1. Process Flows Experimental setups for CDI can be broadly classified into three types of process flows: batch, recirculating batch, and single pass. The simplest setup is the batch process well-known to chemical engineers. In the CDI batch process, a single vessel is used to house the entire system, including the CDI cell, mixer, measurement probes, and saline water. The vessel is then capped to prevent influences from the surroundings on the system. In a typical CDI batch process experiment, the electrodes are attached to a mechanically rigid current collector or support and suspended on the cap of the vessel. The cap also has slots for electrical connections to the current collector, as well as ports for the measurement probes [71]. The limitation of the batch process is that there is a volume of water outside the CDI cell (i.e. not in the channel between the electrodes) where the measurement probes are located. Since electrosorption only occurs in the channel between the electrodes, there is the possibility of poor ion and solution transport in the system. Furthermore, the gap between the electrodes is generally larger which leads to smaller capacitances because of their inverse relationship. Lastly, the batch process is unrepresentative of the real world because water treatment systems usually involve water flows. Next, the recirculation batch process is probably the most frequently studied and reported CDI experimental setup. A fixed volume of saline water is pumped from a recycling tank or reservoir into the CDI cell and back. The measurement probes are placed in the recycling tank or inline measurement probes are used [131], [132]. Finally, in the single pass process, saline water is pumped from the inlet tank into the CDI cell and then into the outlet tank. Measurements are taken from the water before and after passing 31  through the CDI cell. Most CDI systems in the real world are designed and operated as single pass systems because of the need for continuous water treatment. However, single pass processes can be impractical for research and laboratory studies because larger quantities of water is required, particularly for long term experiments and when sources of saline water are scarce [31].   Figure 2.3. Process flow diagrams of batch (a), recirculating batch (b) and single pass (c) processes for CDI experimental setups; sources: [133]–[135].  32  2.4.2. Applied Voltage  As predicted by EDL theory, increasing the voltage applied to the CDI cell increases the salt adsorption capacity because it increases the charge at the electrode-solution interface. Salt adsorption capacity does not increase linearly with increased applied voltage but rather increases exponentially from 0.6 V to 1.2 V. However, above 1.2 V the salt adsorption capacity does not increase as significantly, likely due to parasitic electrochemical reactions [133]. This is supported by the water electrolysis standard reduction potential (SRP) at 1.23 V and evidence showing that charge efficiency decreases and specific energy consumption increases with increasing applied voltage [44]. To optimize the various desalination metrics such as salt adsorption capacity, charge efficiency, specific energy consumption, and electrosorption rate, different applied voltages have been reported. Zornitta et al. (2016) claimed based on their results that 1.2 V, just slightly below the SRP for water electrolysis, is the optimal point when accounting for all the aforementioned metrics [16]. On the other hand, Zhao et al. (2014) conducted a study using response surface methodology that found that 1.57 V was best for maximum salt adsorption capacity [44]. Still more studies reported that 1.8 V yielded the fastest electrosorption rate [34] and 2.0 V produced the lowest specific energy consumption [136]. All in all, an applied voltage of 1.0 – 2.0 V is appropriate for most CDI systems. 2.4.3. Flow Rate  In single pass experimental setups, flow rate has a critical role because it determines the time period which the salt can interact with the electrodes. At slower flow rates, the effluent concentration is more purified because the saline water has a longer time period of contact with the electrodes, thus resulting in higher salt removal efficiencies. However, this must be balanced as flow rate also has a positive relationship with salt adsorption capacity and charge efficiency, 33  and an inverse relationship with energy consumption [30], [34], [46], [47]. For a single pass CDI system with 200 sheets of activated carbon electrodes and dimensions of 158 x 174 x 0.3 mm3 for each electrode, a flow rate of 7 L/min was recommended for optimal salt removal efficiency and energy consumption [32]. However, this optimal value of 7 L/min is inapplicable to other CDI systems because of differences in size, scale, and design. Resultantly, it has been proposed that retention time, which is the time period in which the saline water is in contact with the electrode, be used in place of flow rate as an operating parameter. Generally, if the electrodes are not yet saturated, it is beneficial for salt removal efficiency to increase the retention time either by reducing flow rate or adding more CDI cells [133]. In batch recirculation experimental setups, flow rate has a less consequential effect because the same volume of water is in contact with the electrodes. Nevertheless, there is still some effect because of mixing in the CDI cell and the recycling tank. Salt adsorption capacity in batch recirculation systems increases with flow rate up to a certain point and then begins to decrease, suggesting an optimal value [44], [137]. 2.4.4. Salt Concentration  CDI technology is competitive at brackish water salinities are therefore usually tested below 3,000 ppm TDS with synthetic NaCl solution to simplify the water matrix. As the influent salt concentration increases, salt adsorption capacity also increases as predicted by adsorption isotherm models. For a detailed discussion of adsorption isotherm models and their application to CDI systems, refer to Section 2.6.1. The charge efficiency tends to decrease as NaCl concentration increases, suggesting that low salinity water is preferable for CDI technology [47]. However, the effect of NaCl concentration on charge efficiency is slight and studies have even found it to be insignificant [46]. From EDLC theory, it is expected that higher salt concentrations increase the amount of ions between the electrodes and thus the capacitance. Higher influent NaCl 34  concentration also has a positive effect on certain desalination metrics such as energy consumption and electrosorption kinetics, likely because of the greater conductivity of the water [32], [138]. The ion species also matters because source water matrices generally have salts other than NaCl. Determining the selectivity of ion electrosorption based on theory is quite complicated and has often yielded contradicting results. A simplified explanation is ions with higher charge numbers have stronger electrostatic interactions and therefore are more selective for electrosorption. For ions with the same charge number, the size of the ion as defined by the hydrated radius determines selectivity, with smaller ions being preferential to electrosorption. The electrode’s pore size and structure is also considered to have a significant effect on the electrosorption selectivity of ions [30], [139]. When ions have different molar concentrations, the ion with the highest concentration is usually most competitive for electrosorption [45]. Another study argued that hydration ratio, which is the ratio of hydrated radius to ion radius, is a stronger indicator for electrosorption selectivity than hydrated radius alone [140]. Finally, kinetics and thermodynamics both play roles in ion electrosorption, and competition and substitution of adsorbed ions can occur over time [141].  2.4.5. Environmental Factors Environmental factors can also influence the desalination performance and efficiency of CDI systems. In practice, saline water temperatures for desalination systems hover around room temperature (25 ºC or 298 K) but can reach cold extremes near the freezing point of water (0 ºC or 273 K) and hot extremes of over 35 ºC [142]. Temperature can affect CDI systems in numerous ways. As temperature increases from 15 – 27 ºC, the maximum salt adsorption capacity, which is an adsorption isotherm model parameter, decreases [143]. However, increased temperature resulted in faster kinetics of electrosorption [144]. Salt removal efficiencies followed a decreasing 35  trend as temperature increased from 20 – 50 ºC [30]. The pH of saline water has been found to have no significant effect on the salt adsorption capacity or charge efficiency [145]. However, dissolved gases can have considerable influence on CDI systems, especially the long-term stability. An in-depth discussion of the effect of dissolved gases on long-term stability can be found in Section 2.7.2. As a result, the saline water is sometimes purged with an inert gas such as N2 prior to conducting desalination experiments with CDI systems [136]. However, purging saline water with inert gas can be expensive and complicated because of the transport and use of compressed gases. To the best of the author’s knowledge, the practicality of inert gas purging CDI systems remains to be seen and may be an important study for the future. 2.4.6. Modes of Operation There have been innovative operation modes which have been researched to improve CDI desalination performance and efficiency. Using a constant current, as opposed to a constant voltage, to operate the CDI system has been investigated in previous studies. The advantage of constant current operation is that it produces a steady effluent salt concentration that does not vary with time for single pass systems, and by tuning the current supplied to the CDI system, the effluent salt concentration can be adjusted and controlled [123], [146]. Moreover, the charge efficiency and energy consumption was found to be lower for constant current operation when compared with constant voltage operation [67]. Hybrid constant current and constant voltage operation has also been developed and studied for CDI technology. Constant voltage operation results in an abrupt decrease in the effluent salt concentration at the beginning of the charge step but starts increasing once the lowest point is reached. In contrast, constant current operation reaches the lowest point later in time but can maintain the effluent salt concentration at the lowest point. Therefore, it was shown that a CDI system could be successfully operated with constant voltage at the beginning of 36  the charge step and constant current near the end of the charge step [147]. Another study showed that energy efficiency can be optimized by operating at constant current during the charge step and including a buffer step at constant voltage before the discharge step [69]. Operation modes which optimize the cycle (i.e. charge and discharge step) times have also been researched. Demirer et al. (2013) used steady cycle tests to find characteristic times of minimum effluent salt concentration, maximum specific energy consumption, and maximum electrosorption rate. With these characteristic times, the CDI system was operated with transient cycles to optimize its efficiency [138]. This practice may be limited in real world application because it results in long cycle times. For the discharge step in the CDI process, the CDI cell can either be short circuited or reversed in polarity to desorb the ions. Applying a voltage with reverse polarity can only be performed with MCDI because it is possible for the ions to desorb from the electrode and adsorb onto the opposite electrode, which now has an opposite charge to the ions, without the IEM acting as a barrier. The benefit of reversing the polarity of the applied voltage is that there is more rapid desorption kinetics, charge efficiency is enhanced, and energy consumption is overall reduced [148]. On the other hand, short circuiting allows for the possibility of recovering the energy stored in the CDI cell during the charge step. Energy recovery can be accomplished directly through a buck-boost convertor to control the energy transfer from the CDI cell to the supercapacitor [149]. Another method is via an external load which stores the discharge current as energy. It has been demonstrated that up to 83% of the energy can be recovered during the discharge step for MCDI systems when using constant current operation [150]. For single pass systems, the discharge step can be performed with no flow of saline water to reduce the volume of concentrate produced, although desorption kinetics may be negatively affected [151]. Lastly, a new operation mode has been investigated where the discharge step is performed with brine, which has higher conductivity. 37  After the charge step, the CDI cell is drained and brine is fed into the CDI cell under a negative applied current and significant improvements in energy efficiency were discovered [152]. 2.5. Electrosorption Modelling Several research studies have been dedicated to modelling the electrosorption process, with different levels of mathematical complexity, to provide tools for the design and operation of CDI systems [15]. Basic, simplistic models include adsorption isotherms and Lagergren adsorption kinetics which are used to model the adsorption phenomenon. More complex models focus on EDL coupled with ion transport theory through porous materials. 2.5.1. Adsorption Isotherms Adsorption isotherm models thermodynamically describe the retention and mobility of the adsorbate from gaseous or aqueous media onto the adsorbent at constant conditions. Generally, they are used for equilibrium calculations of the adsorbent and thus can reveal information about the capacity of absorbent systems [153]. For CDI systems, the Langmuir model and Freundlich model for adsorption isotherms have commonly been used to describe the salt adsorption capacity as a function of the salt concentration [15]. The Langmuir isotherm model is an empirical model of adsorption that assumes monolayer adsorption onto a homogeneous surface with no interactions between the adsorbates in the media or on the adsorbent. On the other hand, the Freundlich isotherm model assumes adsorbate multilayer adsorption over the heterogeneous adsorbent surface. Both these isotherm models require the estimation of two parameters using the experimental data [153]. The Langmuir isotherm model and Freundlich isotherm model as applied to CDI systems are shown below in equations (11) and (12) respectively as: 38  𝑞𝑒 =𝑞𝑚𝐾𝐿𝐶𝑒1 + 𝐾𝐿𝐶𝑒 (11) 𝑞𝑒 = 𝐾𝐹𝐶𝑒1𝑛 (12) where qe [mg/g] is the salt adsorption per electrode mass at equilibrium, Ce [mg/L] is the salt concentration at equilibrium, qm [mg/g] is the maximum salt adsorption per electrode mass corresponding to monolayer coverage, KL is the Langmuir constant related to the heat of adsorption, KF is the Freundlich constant related to the adsorption capacity of the adsorbent, and 1/n is the tendency of the adsorbate to be adsorbed [47]. In general, the Langmuir isotherm model has been shown to have better fit for NaCl solutions compared to the Freundlich isotherm model, which is indicative of monolayer adsorption onto a homogeneous surface [32], [34], [47], [52], [84], [133], [144], [154]. However, studies have also shown that the Freundlich isotherm model is a better fit for certain ions [155]. Adsorption isotherm models illustrate the relationship between desalination performance, as described by the salt adsorption per electrode mass at equilibrium (i.e. salt adsorption capacity) and the salt concentration. One of its uses is that the selectivity of different ions in solution can be compared using adsorption isotherms [45]. This is particularly useful in the water treatment field because the source water quality is often variable and subject to complex water matrices and uncontrollable environmental factors.  2.5.2. Lagergren Adsorption Kinetics The kinetics of adsorption is considered to be governed by two phenomena: 1) the diffusion of the adsorbate in the media to the absorbent and 2) the surface interactions between the adsorbate and the adsorbent. The Lagergren adsorption kinetic model is a lumped description of adsorption 39  systems close to equilibrium that uses experimental data of the adsorbate concentration in the media and the absorbent capacity [156]. Lagergren adsorption kinetics is largely empirical and includes a pseudo-first order equation and pseudo-second order equation, with each equation having better applicability for different adsorption systems. The pseudo-first order equation works best when adsorbate interactions are negligible, such as for protein on silica systems, while the pseudo-second order equation is a better fit for adsorption systems with adsorbate interactions, such as for heavy metals or dyes on various materials [157]. For CDI systems, the differential equations and their corresponding solutions are shown in equations (13) and (14) for the Lagergren pseudo-first order equation and (15) and (16) for the Lagergren pseudo-second order equation.  𝑑𝑞𝑑𝑡= 𝑘1(𝑞𝑒 − 𝑞) (13) 𝑙𝑜𝑔(𝑞𝑒 − 𝑞) = 𝑙𝑜𝑔 𝑞𝑒  −  𝑘1𝑡2.303 (14) 𝑑𝑞𝑑𝑡=  𝑘2(𝑞𝑒 − 𝑞)2 (15) 𝑡𝑞 =  1𝑘2𝑞𝑒2 +  𝑡𝑞𝑒 (16) Here, t [min] is time, q [mg/g] is salt adsorption per electrode mass at time t, qe [mg/g] is salt adsorption at equilibrium, k1 [min-1] is the pseudo-first order rate constant, and k2 [g mg-1 min-1] is the pseudo-second order rate constant [144]. The experimental data used for the estimation of parameters should be recorded until the electrode reaches equilibrium (i.e. electrode saturation) 40  for the most accurate results. Some studies have reported that the pseudo first-order equation provided a closer fit than the pseudo second-order equation [52], [84], [158], while others have found the opposite to be true [47], [144]. The major reason for using the Lagergren adsorption kinetic model is that the kinetics of the electrosorption process can be quantified with the rate constant as a parameter. This allows for the prediction of salt removal as a function of time and the comparison of electrosorption kinetics as a desalination performance metric across different operating conditions, electrodes, and CDI systems.  The major drawback of adsorption isotherms and Lagergren adsorption kinetics is that they are an incomplete description of the electrosorption process and do not include crucial operating parameters and environmental factors as variables or parameters. Therefore, the model is only applicable at specific conditions (i.e. applied voltage, flow rate, salt concentration, temperature, and etc.) that the CDI system has been tested at. Adsorption isotherms and Lagergren adsorption kinetics are inherently derived from results from the phenomenon of physical adsorption, and therefore do not include the impact of the applied electric potential and the resulting ion EDL formation. The typical CDI electrode is also porous and therefore ion transport theory through  porous materials should also be applied when considering electrosorption kinetics [15]. Lastly, adsorption isotherms and Lagergren adsorption kinetics only account for a few aspects of the desalination performance and efficiency of the CDI system, and therefore a more holistic model needs to be established. Because of these limitations, models based on EDL coupled with ion transport theory have been developed and represent the next advances to CDI modelling. 2.5.3. Electrical Double Layer and Ion Transport  The Helmholtz model is the most elementary electrosorption model based on EDL theory, and it assumes that on the solution side of the interface, the ions form an EDL on a fixed plane to 41  counter the charge on the electrode side. The Gouy-Chapman-Stern (GCS) model is a step further in complexity and divides the EDL into the inner Stern layer and the outer diffuse layer. The Stern layer is a thin and compact layer composed of ions adsorbed onto the electrode, and the diffuse layer is the layer between the Stern layer and the bulk solution with decaying ion concentration as it gets closer to the bulk solution [39]. The issue with the Helmholtz model and GCS model is that they fail to account for EDL overlapping when the electrode pore dimensions are similar in magnitude to the Debye length, which is an indicator of the thickness of the EDL. To compensate, the modified Donnan (mD) model has been developed with the assumption of strongly overlapping EDLs in the micropores so that electric potentials are constant [15]. Further adjustments are made by classifying the pores as macropores, where ions are transported, and mesopores and micropores, where EDLs form [80]. EDL models are coupled with ion transport models through porous materials to arrive at a more complete mathematical description of CDI systems [39].   Figure 2.4. EDL structure according to the GCS model (a) and the mD model (b); source: [15].  The mD model has been implemented in many studies and has yielded profitable results that have been verified by experimental data. Desalination performance of CDI systems can be 42  predicted with metrics like salt adsorption capacity, charge transfer, and kinetics by using a 2D mD model [17]. Additionally, the mD model has been used by discretizing the flow channel between the electrodes into ideally-stirred volume cells to determine the lowest effluent concentration [159], and optimal operating parameters for average salt adsorption rate and water recovery [160]. Modelling results have also shown that charge efficiency and energy consumption can be predicted under different charging and discharging conditions [123], [148]. The effect of electrode modifications has also been investigated with electrosorption modelling, including the effect of thickness [73], [161] and the enhancement from incorporating chemical surface charges [162]. Membranes can be accounted for by including the interparticle porosity, transport resistance and stagnant diffuse layer [79], [124]. Finally, ion selectivity and pH changes have also been studied with electrosorption models [141], [163]. Electrosorption models based on EDL coupled with ion transport theory requires the non-trivial solving of a series of differential equations and algebraic equations, which is generally completed through numerical analysis. Research to validate current models and develop new models for electrosorption is currently a trending topic. 2.6. Degradation and Long-Term Stability 2.6.1. Fouling and Scaling  Fouling is defined as the settling of solids on a surface to create a layer of electrical, thermal, or physical resistance. In contrast, scaling is defined as a physico-chemical change that causes crystallization, precipitation, or solidification of components from the solution onto the surface [12]. Organics has been identified as one of the main sources of fouling and degradation for CDI systems. Zhang, Mossad & Zou (2013) compared two brackish groundwaters, one with hardness and one with total organic carbon (TOC). Hardness was found to have a limited effect on the long-term performance and efficiency of the CDI system, while a TOC level of just 2 mg/L caused rapid 43  degradation, with the salt removal efficiency decreasing from 86% to 55% after 15 days of operation [164]. Mossad & Zou (2013) further discovered that synthetic solutions of humic acid with a TOC level of 3.1 mg/L drastically reduced the salt removal efficiency and flow rate. They also found that Fe3+ ions contributed significantly to scaling, and resulted in falling salt removal efficiencies and flow rates [31]. Wang et al. (2015) studied the effect on CDI systems of a bio-treated wastewater effluent with at TOC of 7 mg/L. They observed that the CDI system adsorbed less and less salt per cycle, and electrical resistance and salt removal efficiency rapidly decreased. Additionally, by using a 3D fluorescence technique, they deduced that protein-like and humic acid-like substances were the major reason for fouling [34]. Lado et al. (2015) noticed that ions can either be removed by electrosorption or physical adsorption. From their study, they found that SO42- interacted mostly through physical adsorption and thus was the main cause of poor regenerability, whereas Na+, Ca2+, and Cl- interacted mostly through electrosorption [135]. To recover the desalination performance and efficiency of CDI systems, it has been shown that alkaline cleaning could remove organic fouling while acid cleaning could reverse inorganic scaling [31]. It should be noted that biofouling, the accumulation of micro-organisms, plants and algae on surfaces, has been rarely investigated for CDI systems [12]. There is a general lack of research on fouling and scaling of CDI systems, and the implementation of CDI technology will likely be obstructed unless further research is performed to answer questions of practicality. 2.6.2. Parasitic Electrochemical Reactions  In addition to fouling and scaling from the source water matrix, the long-term stability of CDI systems can be degraded even in synthetic solutions by electrochemical, or Faradaic, reactions. Haro et al. (2011) observed that ACC and ACP electrodes underwent decreases in BET surface area and regeneration efficiency over 30 cycles with 1500 ppm TDS solution and an applied 44  voltage of 1.2 V [165]. Furthermore, Cohen et al. (2013) discovered that the degradation of long-term stability for CDI systems was because of oxidation of the anode, or the positive electrode. This anode oxidation resulted in the system becoming asymmetric and an inversion effect where desorption of co-ions occurred during charging and re-adsorption occurred during discharging [166]. In another study, Cohen et al. (2015) further expanded on the results by showing evidence and suggesting that faster diffusion rates of dissolved O2 and a local basic environment at the cathode, or the negative electrode, compromised the long-term stability [37]. By conducting CDI experiments with 500 ppm NaCl at 1.2 V for 50 continuous cycles, Duan et al. (2015) found similar results where performance degradation of ordered MC electrodes was because of increases in oxygen functional groups and reductions of BET surface area [167].  The possible oxidation half reactions that can occur at the anode is shown below. All SRPs are given versus the standard hydrogen electrode (SHE).  2H2O →  O2 +  4H+ +  4e−     E0 = 1.229 V   (17)  H2O2  →  O2 + 2H+ + 2e−    E0 = 0.69 V   (18)  C + 2H2O →  CO2 +  4H+ +  4e−    E0 = 0.7 – 0.9 V  (19)  2Cl−  →  Cl2 +  2e−     E0 = 1.36 V   (20) Reactions (18) and (19) are the most probable electrochemical reactions to occur because their SRP is below 1.2 V, the typical applied voltage a CDI system is operated at. This is supported by experimental evidence from He et al. (2016) that showed that oxidation of Cl-  was negligible, whereas H2O2 was generated and pH variations occurred [168]. Anode oxidation of the carbon electrode has also been widely known to occur, leading to asymmetry and destabilization of the performance of the CDI system [169].  45  For the cathode, multiple reduction half reactions may occur based on the species present in the CDI system.  O2 +  2H2O +  4e−  → 4OH−   E0 = 0.401 V   (21)  O2 + 4H+  +  4e−  → 2H2O    E0 = 1.229 V   (22)  O2 + 2H+ +  2e−  →  H2O2    E0 = 0.69 V   (23)  H2O2 + 2H+ +  2e−  → 2H2O    E0 = 1.78 V   (24)  2H+ + e−  →  H2     E0 = 0 V   (25) H2O2 can be produced via reaction (23), where it can then undergo reduction through reaction (18), further oxidation through reaction (24), or disproportionation to H2O and O2. Tang et al. (2017) found experimentally that during charging, H2O2 increased rapidly and then decreased in CDI cells, whereas no H2O2 was generated in MCDI cells [170]. Aside from reactions (21) to (23), dissolved O2 can also undergo a peroxide pathway in both acidic and alkaline solutions where a cascade of electrochemical and disproportionation reactions occur [171]. Since these electrochemical reactions contain H+ and OH- ions, variations in pH are caused during the CDI process. Variations in pH are also caused by the different mobility of H+ and OH- ions, and chemical surface charge groups in the micropores of the electrodes, although electrochemical reactions arguably play the biggest role [163]. The pH variations can be localized near the electrode, which may be problematic since the cell potential is affected by the pH according to the Nernst equation [172]. The electrochemical reactions can be somewhat prevented by either purging the saline water with an inert gas such as N2 or applying a smaller voltage [173].  46  SRP only describes the thermodynamics of the electrochemical reactions at standard state. According to the Nernst equation, localized conditions such as the ion concentrations, pH, and temperature determine the actual electrode potential and thus also govern whether the electrochemical reaction has the potential to occur. Kinetically, an overpotential is also typically required for an electrochemical reaction and is usually experimentally determined [174]. As can also be seen, predicting electrochemical reactions is a difficult process because there are many half reactions which may happen, even for ideal, synthetic solutions. In real world CDI systems, there is also a source water matrix which further complicates the analysis.             47  Chapter 3: Methodology 3.1. Activated Carbon Electrodes 3.1.1. Materials  Activated carbon powder (ACP) electrodes were obtained from Pureechem (Cheongju, South Korea). The ACP used for the electrodes had an approximate BET surface area of 550 m2/g. Aside from the ACP, the electrodes consisted of sodium carboxymethyl cellulose (CMC) as the polymeric binder and graphite foil as the support or current collector. The specifications of the graphite foil are described in Appendix A.1. The coating layer (ACP and CMC) of the electrodes had dimensions of 10.0 x 10.0 cm2 with a 1 cm diameter circular hole cut in the center for water to flow out. The average thickness and average density was 414 µm and 0.73 g/cm3 respectively. The whole electrode had dimensions of 15.0 x 10.0 cm2 with a 5.0 x 10.0 cm2 section of uncoated graphite foil to serve as a pathway for the electrical connection. The uncoated section of graphite foil was covered with plastic tape to prevent tearing and had three circular holes of 1 cm diameter for the screws. A diagram of the Pureechem ACP electrodes with its dimensions can be found in Figure 3.1. According to the vendor, the simplified fabrication procedure for the electrodes included mixing, coating, drying, pressing, and cutting performed sequentially. No other details were provided by the supplier. 48   Figure 3.1. 3D drawing of Pureechem ACP electrode with dimensions.   ACP electrodes were also fabricated in-house for comparison with the Pureechem ACP electrodes. An activated carbon called WPC was obtained from Calgon Carbon (Moon Township, PA, U.S.A.) and another called YEC-8A was obtained from Yihuan Carbon (Fuzhou, China). Both ACPs were derived from coconut shells and their specifications are displayed in Table 3.1. The binder used for in-house fabrication of ACP electrodes was polyvinylidene fluoride (PVDF) in powder form with an average molecular mass of 534,000 Da. Other binders such as PTFE and CMC were not used because of issues of structural integrity, namely problems with inadequate adhesion to the support, resulting in material losses when rubbed with a gloved finger or when placed in flowing water. To dissolve the PVDF, anhydrous N,N-dimethylacetamide (DMA) was used as the solvent. Both PVDF and DMA were purchased from Sigma-Aldrich (Oakville, ON, Canada). Graphite foil with a thickness of 250 µm was used as the current collector or support and was supplied as Flexible Graphite 2010A by Mineral Seal Corporation (Tuscon, AZ, U.S.A.). More detailed specifications of the graphite foil can be found in Appendix A.2. Additionally, Toray 49  carbon paper with 190 µm thickness was also tested as the support for the ACP electrode and was supplied by Fuel Cell Store (College Station, TX, U.S.A.).  WPC YEC-8A Pureechem ACP Electrode BET Surface area [m2/g] ~ 850 ≥ 2000 ~ 550 Specific pore volume [mL/g] - ≥ 1.2 - Average particle size [µm] - ≤ 10 ≤ 10 Moisture content [wt. %] ≤ 8 ≤ 5 - Ash content [wt. %] ≤ 18 ≤ 0.5 - Iron concentration [ppm] - ≤ 50 - pH - 6 - Bulk density [g/mL] - 0.38 – 0.40 ~ 0.73 Specific capacitance [F/g] - 260 (aqueous electrolyte) - Table 3.1. Specifications of ACPs from Pureechem, Calgon Carbon and Yihuan Carbon.  For the synthetic salt solutions used in the desalination experiments, analytical-grade NaCl crystals was purchased from Sigma-Aldrich and deionized water was obtained from a Purelab Option-Q water purification system by Elga LabWater capable of delivering water with an inorganic purity of < 18.2 MΩ cm. Solutions of 500 ppm, 1000 ppm, 2000 ppm, and 3000 ppm NaCl were prepared by measuring NaCl crystals on an analytical balance and diluting in a volumetric flask. 3.1.2. Fabrication procedures  Slurries with ACP-to-PVDF solids ratios of 85:15, 90:10, and 95:5 were prepared in glass vials. First, binder solutions were prepared by dissolving PVDF in DMA. Then, ACP was added 50  to the binder solution slowly and the mixture was stirred for 4 h at 900 rpm to form the slurry. To determine the amount of solvent required to form a slurry with a suitable viscosity for casting, solvent-to-solids mass ratios of 1.5:1 to 5:1 were tested. The slurry was cast onto a 15.0 x 10.0 cm2 sheet of graphite foil with a 5.0 x 10.0 cm2 section covered with plastic tape to form a coating layer with an area of 10.0 x 10.0 cm2. The coating layer thickness was controlled through either of two methods: by applying a predetermined mass of slurry as the coating, or by using a manual doctor blade to make a coating with a uniform wet thickness. In the first method, which is referred to as evaporative casting (EC), the slurry was spread as evenly as possible onto the coating layer area with a spatula using wet mass loadings of 23.4 – 46.8 mg/cm2. In the latter method, which is referred to as doctor blade casting (DBC), a 100 mm doctor blade with adjustable thickness from MTI Corporation was used to apply the slurry to the coating layer area at uniform wet thicknesses of 50 – 700 µm. After casting, the electrode was left standing at ambient conditions for 1 h and then placed in an oven held at 80 ºC to dry for 18 h. After drying, the electrode was placed in ambient conditions for 1 h to cool to room temperature and 1 cm holes were cut in the electrodes according to the same design as the Pureechem ACP electrodes. For electrodes with carbon paper, fully-coated 10.0 x 10.0 cm2 electrodes were prepared in a similar manner to the electrodes with graphite foil. However, the carbon paper electrodes were taped to 5.0 x 10.0 cm2 sheets of graphite foil to prevent water leaking in the CDI cell since carbon paper is permeable to water. Fully-coated 8 x 8 cm2 electrodes were also prepared and cut into small, approximately 5 x 5 mm2 pieces for other measurements and characterization. The mass of the graphite foil or carbon paper support and fabricated electrode was measured to determine the dry mass loading. 51                                                                                                             Figure 3.2. In-house ACP electrode fabrication procedure and completed 8 x 8 cm2 electrode (a) and 15 x 10 cm2 electrode (b).  3.1.3. Characterization  Scanning electron microscopy (SEM) and energy-dispersive x-ray spectroscopy (EDS) were used to analyze the surface characteristics and morphologies, and elemental compositions of the pristine ACP electrodes. The instrument used for the analysis was a Philips XL30 SEM with a Bruker Quantax 200 EDS. Cyclic voltammetry (CV) with a three-electrode setup was performed on the ACP electrodes to investigate their voltage, current, and capacitance relationships. Two 2.5 x 6.0 cm2 pieces of the electrodes were pressed onto Ni mesh and taped with Kapton on the back. Then, the electrodes were placed in a glass sample holder with a separation distance of 15 mm and filled to a height of 5.0 cm (2.5 x 5.0 cm2 wet area) with 1000 ppm NaCl. A Hg/Hg2SO4 reference electrode was placed in between the two electrodes. A Bio-Logic VMP3 potentiostat was connected via electrical wires to the electrodes, and the scan rate was set to 5 mV/s with a voltage range of -0.050 V to 0.650 V. Data was recorded and processed with EC-Lab software. Mixing Casting Drying and Cutting 52  3.2. Capacitive Deionization Experimental Setup  3.2.1. Capacitive Deionization Cell A bench-scale CDI cell was purchased from Pureechem (Cheongju, South Korea). The CDI cell consisted of a pair of activated carbon electrodes as described in Section 3.1.1., a cation-exchange membrane and anion-exchange membrane, and a nylon mesh as the insulating spacer. These components were stacked and screwed tightly between a pair of acrylic plates with clear rubber gaskets for lining and sealing. The CDI cell was designed with a flow-by geometry where the water enters from two corners of the electrode and leaves through a hole in the center.   Figure 3.3. Schematic of stack in CDI cell (left) and picture of CDI cell (right).  3.2.2. Process Flow Setup  A recirculating batch process flow was used to conduct desalination experiments. The inlets and outlet of the CDI cell were connected to the reservoir via Tygon plastic tubing with 3/16 inch outer diameter and 1/16 inch inner diameter. The total length of the tubing in the setup was 122 cm. A Masterflex peristaltic pump connected to a Cole-Parmer solid-state speed controller 53  was used to recirculate a constant volume of 250 mL of synthetic NaCl solution from the reservoir through the CDI cell and back into the reservoir. The NaCl solution to be treated was recirculated throughout the CDI system for 30 min prior to each desalination experiment so start-up effects could be neglected. The total volume displaced by the tubing and the CDI cell was approximately 21 mL. 3.2.3. Electrical Circuit and Measurement Setup  A BK Precision (Yorba Linda, CA, U.S.A.) 1688B DC power supply provided the applied voltage to the CDI cell. The power supply was connected to a ProXR USB relay from National Control Devices (Osceola, MO, U.S.A.) to switch between the charge and discharge stages. The relay was connected to a computer and controlled with Relay Pros Quick Timer software. For a detailed circuit describing how the relay works, see Appendix B. An Amprobe (Everett, WA, U.S.A.) 38XR-A digital multimeter was connected in series with the power supply, relay, and CDI cell to measure current continuously over time. To measure the conductivity and pH of the synthetic NaCl solution in the reservoir continuously over time, a Hanna Instruments (Laval, QC, Canada) HI5522 meter with a conductivity probe, pH probe, and temperature probe for automatic temperature correction of conductivity and pH was used. The electrode potential of the electrodes in the CDI cell was also measured for some of the desalination experiments using a Hg/HgO reference electrode placed in the reservoir and connected to the digital multimeter. The conductivity and pH data was logged over time with HI92000 software from Hanna Instruments, and the current and voltage data was logged with 38SW-A software from Amprobe. All data was logged with 10 s time intervals in between measurements. A process diagram and picture of the setup can be found in Figure 3.4. and 3.5., respectively.  54   Figure 3.4. Process diagram of CDI desalination experimental setup showing process flow, electric circuit, and measurement instrumentation.   Figure 3.5. Picture of CDI desalination experimental setup.  55  3.3. Desalination Experiments 3.3.1. Comparison of Activated Carbon Electrodes In-house fabricated and commercially obtained ACP electrodes were placed in the CDI cell without ion-exchange membranes and desalination experiments were performed with operating parameters of 1.2 V applied voltage, 1000 ppm NaCl concentration, and 20 mL/min flow rate as the baseline condition. Cycle times were 60 min for the charge/adsorption stage and 60 min for the discharge/desorption stage, and at least 3 complete cycles were performed for each experiment. In total, the desalination performance of four ACP electrodes was tested based on structural integrity observations and the SEM and EDS analysis: commercially obtained Pureechem electrodes, WPC/PVDF on graphite foil, YEC-8A/PVDF on graphite foil, and YEC-8A/PVDF on carbon paper. Details of the tested in-house fabricated ACP electrodes can be found in Table 3.2. Electrode Fabrication Method Dry Mass Loading (mg/cm2) ACP-to-PVDF solids ratio Solvent-to-solids mass ratio WPC/PVDF on graphite foil EC 13.75 85:15 1.5:1 YEC-8A/PVDF on graphite foil DBC 1.94 85:15 3:1 YEC-8A/PVDF on carbon paper DBC 2.30 85:15 3:1 Table 3.2. Key specifications and parameters for in-house fabricated ACP electrodes.   3.3.2. Varying of Operating Parameters  Using the baseline conditions of 1.2 V, 1000 ppm NaCl, and 20 mL/min, one of the operating parameters of applied voltage, NaCl concentration, and flow rate were sequentially varied. The applied voltage was varied to 0.8 V, 1.0 V, 1.2 V, and 1.5 V by adjusting the setting 56  on the power supply. Solutions of 500 ppm, 1000 ppm, 2000 ppm, and 3000 pmm NaCl were tested in desalination experiments with the CDI experimental setup separately to determine the effect of NaCl concentration on desalination performance and efficiency. Finally, the flow rate was set to either 20 mL/min, 40 mL/min, or 60 mL/min with the solid-state speed controller. Moreover, the flow rate was measured to confirm the controller setting using a stopwatch and graduated cylinder. All experiments for determining the effect of operating parameters were conducted using Pureechem ACP electrodes without ion-exchange membranes with cycle times of 60 min. Also, a 3 x 3 matrix of experiments with applied voltages of 0.8 V, 1.0 V, and 1.2 V, and NaCl concentrations of 500 ppm, 1000 ppm, and 2000 ppm were performed to further investigate the combined effects of these operating parameters. The range of operating parameters chosen was based on the size and scale of the CDI cell and experimental setup, and the application that it was to be used for (i.e. brackish water desalination at low applied voltages).  Experiments were also performed with the ion-exchange membranes on the Pureechem ACP electrodes in the CDI cell. For these experiments, the NaCl concentration and flow rate was kept constant at 1000 ppm and 20 mL/min respectively, while the applied voltage was varied to 0.8 V, 1.0 V, 1.2 V, or 1.5 V. 3.3.3. Long-Term and Regeneration Tests Long-term desalination experiments were performed for the CDI cell with Pureechem ACP electrodes without ion-exchange membranes at the baseline condition of 1.2 V applied voltage, 1000 ppm NaCl concentration, and 20 mL/min flow rate. The experiments were run for a total of 70 cycles of 60 min charge and 60 min discharge. SEM and EDS analyses were performed for the pristine electrode before the desalination experiments and for the degraded electrodes after. 57  Degraded electrodes were then soaked in either 0.01 M citric acid or 0.01 M NaOH solution for 18 h. After soaking, the electrodes were further washed with deionized water until the pH of the wash water was between 6 and 8, followed by drying in an oven held at 80 ºC for 4 h. The chemically regenerated electrode was then re-tested in the CDI system under the same baseline conditions mentioned previously. Also, a blank electrode was prepared by soaking a degraded electrode in deionized water for 18 h. SEM and EDS measurements were performed on the blank electrode and the regenerated electrodes. Additionally, two experiments were conducted with the Hg/HgO reference electrode connected to the digital multimeter and either the anode or cathode of the CDI cell under the same baseline conditions. This was to determine the standard electrode potentials and explore the possibility of electrochemical reactions. 3.4. Desalination Metrics and Modelling  To provide a holistic evaluation of the desalination performance and efficiency of the CDI system, salt adsorption capacity (SAC), charge efficiency (Λ), and specific energy consumption (η) was utilized and are defined as follows: 𝑆𝐴𝐶 [𝑚𝑔/𝑔] =  𝑉(𝐶𝑖 − 𝐶𝑓) 𝑚        (8)  𝛬 [%] =  𝑧 𝐹 𝑉 (𝐶𝑖 − 𝐶𝑓) 𝑀 ∫ 𝐼 𝑑𝑡 𝑥 100%       (9) 𝜂 [𝑘𝑊ℎ/𝑔] =  𝐸𝑐𝑒𝑙𝑙  ∫ 𝐼𝑑𝑡𝑉 (𝐶𝑖 − 𝐶𝑓)        (10) 58  Here, V [L] is the volume of the water, Ci and Cf [mg/L] are the initial and final TDS concentrations of the water respectively, m [g] is the mass of the electrode, z is the charge number of the ions in the salt, M is the molar mass of the salt [g/mol], F [96,485 C/mol] is the Faraday constant, Ecell [V] is the voltage applied, and I [A] is the current supplied to the CDI cell. For more details regarding these desalination metrics, please refer to Section 2.1.  Modelling can offer a deeper understanding of the phenomenon in the CDI process through quantifiable parameters, and therefore should also be considered in the evaluation of desalination performance and efficiency. In this research project, adsorption isotherms, Lagergren adsorption kinetics, and equivalent electrical circuit models and their parameters are considered. 3.4.1. Adsorption Isotherms  Langmuir and Freundlich adsorption isotherms were applied to explore the effect of NaCl concentration on SAC. The equations describing the Langmuir adsorption isotherm and its linearized form are shown below. 𝑞𝑒 =𝑞𝑚𝐾𝐿𝐶𝑒1+𝐾𝐿𝐶𝑒          (11) 1𝑞𝑒=1𝑞𝑚𝐾𝐿𝐶𝑒+1𝑞𝑚         (26) Similarly, the equations describing the Freundlich adsorption isotherm and its linearized form is as follows:   59  𝑞𝑒 = 𝐾𝐹𝐶𝑒1𝑛          (12) 𝑙𝑛(𝑞𝑒) = 𝑙𝑛(𝐾𝐹) +1𝑛𝑙𝑛(𝐶𝑒)       (27) Adsorption isotherm model parameters qm, KL, n, and KF were calculated from the slopes and y-intercepts of the linearized equations. The correlation coefficient (r2) value was utilized to assess the goodness of fit of the model (see Appendix C.1). Further details of adsorption isotherms can be found in Section 2.6.1. 3.4.2. Lagergren Adsorption Kinetics  Lagergren pseudo-first order and pseudo-second order adsorption kinetics were applied to investigate the electrosorption kinetics. The two equations describing Lagergren adsorption kinetics are displayed below. 𝑙𝑜𝑔(𝑞𝑒 − 𝑞) = 𝑙𝑜𝑔 𝑞𝑒  −  𝑘1𝑡2.303       (14) 𝑡𝑞 =  1𝑘2𝑞𝑒2  +  𝑡𝑞𝑒         (16) The parameters qe, k1, and k2 were estimated using the Goal Seek function in Microsoft Excel. The goodness of fit for the model was assessed with the root-mean-square error (RMSE), which is defined in Appendix C.2. 3.4.3. Equivalent Electrical Circuit  Equivalent electrical circuit models provide helpful information by simplifying the system into circuit elements with quantifiable parameter values. For the CDI cell, the current vs. time data 60  during the charge stage was modelled with two equivalent electrical circuits: the modified RC series circuit with constant leakage current and the Randles circuit.          Figure 3.6. Electrical circuit diagrams of the RC series circuit (a) and Randles circuit (b).  The modified RC series circuit has the power supply (Es) connected to a resistor (Rs) and capacitor (σ) in series, while the Randles circuit is identical except for an additional resistor (Rp) in parallel with the capacitor. Since a leakage current is common to electrical systems, the RC series circuit model was modified to include a constant leakage current term (Ileak). For the modified RC series circuit with constant leakage current, the mathematical equation describing the equivalent circuit’s current has been analytically derived as  𝐼(𝑡) =  𝐸𝑠𝑅𝑠 𝑒−𝑡 𝜎𝑅𝑠⁄ + 𝐼𝑙𝑒𝑎𝑘        (28) where I is the current, t is time, and the circuit elements are as defined above [175]. Likewise, the mathematical equation for the current of the Randles circuit has been solved to be  𝐼(𝑡) =  𝐸𝑠𝑅𝑠 + 𝑅𝑝+ 𝐸𝑠𝑅𝑝𝑅𝑠(𝑅𝑠 + 𝑅𝑝) 𝑒((−𝑡(𝑅𝑠 + 𝑅𝑝)) (𝜎𝑅𝑠𝑅𝑝))⁄     (29) (a) (b) Es Rs σ σ  Es Rs Rp Ileak 61  where the variables and parameters are as defined previously [176]. The equivalent circuit model parameters Rs and Rp were calculated using the following boundary conditions.  𝑡 → 0:   𝐼(0) =  𝐸𝑠𝑅𝑠         (30)  𝑡 → ∞:   𝐼(∞) =  𝐸𝑠𝑅𝑠 + 𝑅𝑝        (31) Goal Seek from Microsoft Excel was then used to find σ. To assess goodness of fit, RMSE was used (see Appendix C.2.).           62  Chapter 4: Results and Discussion 4.1. Electrode Fabrication 4.1.1. Structural Integrity and Mass Loading  Activated carbon powder (ACP) electrodes are required to have structural integrity for them to be used in CDI cells. More specifically, electrodes must be able to withstand the shear stress of flowing salt water as well as be indissolvable in the salt water [62]. The effect of different types of ACPs and fabrication methods on electrode structural integrity was observed. Also, the structural integrity of ACP electrodes with varying ACP-to-PVDF solids ratio, solvent-to-solids mass ratio, and slurry mass loading was investigated.   First, it was found that a higher PVDF binder content generally resulted in stronger structural integrities. Pictures of ACP electrodes with ACP-to-PVDF solids ratios of 85:15, 90:10, and 95:5 are shown in Figure 4.1.  Figure 4.1. YEC-8A/PVDF electrodes with ACP-to-PVDF ratios of 85:15 (left), 90:10 (middle), and 95:5 (right). Electrodes were fabricated using the EC method with a wet mass loading of 31.3 mg/cm2 and a solvent-to-solids ratio of 5:1.  63  As seen in Figure 4.1., electrodes with an ACP-to-PVDF ratio of 90:10 exhibited minor cracking while the 95:5 ratio electrode showed large cracks. Additionally, slight material loss when rubbing with a gloved finger was observed for the 90:10 ratio electrode, while considerable material loss was observed for the 95:5 ratio electrode. On the other hand, the electrode with an 85:15 ratio had a relatively smooth, uniform surface with no material loss when rubbed with a gloved finger. Since the role of PVDF was to provide rigidity to the electrode, it was expected that increased PVDF content also improved the structural integrity of the electrode [61]. On the other hand, too high a PVDF content could decrease the accessible surface area, double layer capacitance, and electrosorption efficiency and therefore it must also be balanced. Most studies have found that an ACP-to-PVDF ratio of around 90:10 produces the optimal desalination performance without sacrificing structural integrity [36].  The solvent-to-solids mass ratio mainly affected the viscosity of the slurry, which could also influence the structural integrity of the electrode for both the evaporative casting (EC) method and the doctor blade casting (DBC) method. Pictures of ACP electrodes fabricated using the EC method with two solvent-to-solids mass ratios are shown in Figure 4.2. 64   Figure 4.2. YEC-8A/PVDF electrodes with solvent-to-solids ratios of 4:1 (left) and 5:1 (right). Electrodes were fabricated with the EC method with an ACP-to-PVDF ratio of 85:15 and a wet mass loading of 31.3 mg/cm2.  In the EC method of electrode fabrication, a viscous slurry was detrimental to the uniformity of the electrode surface because there was no tool to ensure a uniform wet thickness. The fabricated ACP electrode with a solvent-to-solids ratio of 5:1 had a well-formed surface with no cracks or bald spots, whereas the 4:1 ratio electrode had a non-uniform surface with cracks and bald spots.  65   Figure 4.3. YEC-8A/PVDF electrodes fabricated with the DBC method using a solvent-to-solids mass ratio of 3:1 (left) and 4:1 (right). The wet thickness was set to 200 µm and the ACP-to-PVDF ratio was 85:15.  On the other hand, in the DBC method, the doctor blade served to provide a uniform wet thickness and therefore more viscous slurries could be used without sacrificing surface uniformity. Both 3:1 and 4:1 solvent-to-solids ratio slurries could be utilized to fabricate ACP electrodes with uniform surfaces using the DBC method as evident in Figure 4.3. Finally, another factor which was observed to influence the surface uniformity through the slurry viscosity was the characteristics of the ACP. When slurry with WPC was made, it was noticed that the slurry was substantially less viscous compared to YEC-8A when the solvent-to-solids ratio was identical. With WPC as the ACP, uniform electrodes could be fabricated using the EC method at a solvent-to-solids ratio of 1.5:1 compared to 5:1 for YEC-8A.  66   Figure 4.4. WPC/PVDF electrodes fabricated with the EC method using ACP-to-PVDF ratios of 85:15 (left), 90:10 (middle), and 95:5 (right). The solvent-to-solids ratio was 1.5:1 and the wet mass loading was 31.25 mg/cm2.  The reason for the lower viscosity could be explained by the particle sizes and bulk density of the ACP. The YEC-8A was a fine powder with reported particle sizes of 10 µm. On the other hand, WPC was a coarse powder and thus had a considerably higher bulk density than YEC-8A. As a result, for the same mass of ACP, there was a larger volume of YEC-8A and therefore a more viscous slurry. Moreover, since WPC was a coarser powder, the WPC/PVDF electrodes held together easier. Even at a low ACP-to-PVDF ratio of 95:5, barely any cracks were observed and no materials loss occurred when rubbing with a gloved finger as seen in Figure 4.4. Lastly, the slurry mass loading and its effect on structural integrity was explored for both the EC and DBC methods. For the EC method, the slurry mass loading was controlled using the wet mass loading, which was defined as the mass of wet slurry casted per area of the support. On the other hand, the slurry mass loading was controlled in the DBC method by changing the height of the opening to adjust the wet thickness. The two methods could be compared using the dry mass loading, which was calculated by subtracting the mass of the support from the mass of the fabricated electrode and dividing by the area of the coating layer. The correlation between wet 67  mass loading in the EC method and dry mass loading can be visualized graphically as in Figure 4.5.  Figure 4.5. Relationship between dry mass loading and wet mass loading for WPC/PVDF and YEC-8A/PVDF electrodes fabricated with the EC method. WPC/PVDF electrodes were fabricated with a solvent-to-solids ratio of 1.5:1 and YEC-8A/PVDF electrodes with 5:1.   For the fabricated electrodes, results suggested that dry mass loading was positively and linearly correlated with wet mass loading. However, the strength of the correlation differed depending on the slurry characteristics. For the YEC-8A/PVDF slurry, there was a high solvent-to-solids ratio and therefore a weaker correlation because the slurry consisted mostly of solvent which evaporated. There was no observed difference for the relationship between dry mass loading and wet mass loading at various ACP-to-PVDF ratios for both the WPC/PVDF and YEC-8A/PVDF electrodes. Similarly, the relationship between dry mass loading and wet thickness in the DBC method can be illustrated as in Figure 4.6. 051015202530354045500 20 40 60 80 100Dry Mass Loading (mg/cm2)Wet Mass Loading (mg/cm2)WPC:PVDF 85:15WPC:PVDF 90:10WPC:PVDF 95:5YEC-8A:PVDF 85:15YEC-8A:PVDF 90:10YEC-8A:PVDF 95:568   Figure 4.6. Relationship between dry mass loading and wet thickness for YEC-8A/PVDF electrodes fabricated with the DBC method. Conditions were kept constant with an ACP-to-PVDF ratio of 85:15 and a solvent-to-solids ratio of 5:1.   As expected, dry mass loading was positively correlated and had a mostly linear relationship with wet thickness. From the results, it was found that a wet thickness of 700 µm in the DBC method is comparable in dry mass loading to a wet mass loading of 31.25 mg/cm2 in the EC method.  As dry mass loading increased, the structural integrity worsened. YEC-8A/PVDF electrodes fabricated with the DBC method using wet thicknesses of 50 µm to 700 µm (dry mass loadings below around 6.5 mg/cm2) all had intact and robust surfaces, although the ones with low wet thicknesses were slightly inhomogeneous because of the lack of material to coat the entire area of the support. On the other hand, as seen in Figure 4.7., YEC-8A electrodes with dry mass loadings above around 6.5 mg/cm2 had uneven and cracked surfaces. 012345670 100 200 300 400 500 600 700 800Dry Mass Loading (mg/cm2)Wet Thickness (µm)69   Figure 4.7. YEC-8A/PVDF electrodes fabricated using the EC method with dry mass loadings of 39.1 mg/cm2 (left) and 46.9 mg/cm2. ACP-to-PVDF ratio was 85:15 and solvent-to-solids ratio was 5:1.  Contrarily, WPC/PVDF electrodes were able to hold higher dry mass loadings without compromising surface structure. This was possibly due to the coarser nature of the WPC particles making it easier for PVDF to hold together the coating layer as well as adhere it to the support. However, slight cracking could be observed at exceptionally high dry mass loadings at and above 28 mg/cm2 as evident in Figure 4.8.  Figure 4.8. WPC/PVDF electrodes fabricated using the EC method with dry mass loadings of 14 mg/cm2 (left), 28 mg/cm2 (middle) and 40 mg/cm2. ACP-to-PVDF ratio was 85:15 and solvent-to-solids ratio was 1.5:1.  70   Lastly, the YEC-8A/PVDF electrodes fabricated on carbon paper support had noticeably better structural integrities than those fabricated on graphite foil supports. An explanation is that since carbon paper was permeable to liquid, the slurry was in much more intimate contact with the support, resulting in stronger adhesion. This was supported by dry mass loading results which showed higher material retention in carbon paper electrodes (2.30 mg/cm2) compared to graphite foil electrodes (1.94 mg/cm2) at the same wet thickness of 200 µm. A picture of a YEC-8A/PVDF electrode on carbon paper is shown in Figure 4.9.  Figure 4.9. YEC-8A/PVDF electrode on carbon paper with a dry mass loading of 2.30 mg/cm2, ACP-to-PVDF ratio of 85:15 and solvent-to-solids ratio of 3:1. Graphite foil was taped onto the 10 x 10 cm2 electrode to prevent leaking when placed in the CDI unit.   The results of structural integrity and mass loading were significant because they provide experimental evidence and limitations. Studies often expressed electrode mass loading in terms of electrode thickness and density. A comprehensive study of electrode thickness found that although thicker electrodes had higher salt removals, thinner electrodes had higher salt adsorption capacities which suggest they were more efficient in their use of the electrode mass [16]. Therefore, the thickness of the electrode must be optimized to balance salt removal with salt adsorption capacity. 71  Furthermore, the electrode cannot be increased in thickness without limit lest there be a breakdown in the surface structure. The results presented in this section suggest that higher amounts of PVDF and solvent, and lower mass loadings provided electrodes with improved structural integrity. However, the results can be dependant on materials and fabrication procedures and therefore cannot be generalized to all cases of ACP electrodes. 4.1.2. Scanning Electron Microscopy and Energy Dispersive X-Ray Spectroscopy  Scanning Electron Microscopy (SEM) images provided a qualitative view of the surface morphology of the electrodes. From the SEM images in Figure 4.10., the Pureechem electrodes had the finest ACP particles when compared to the fabricated WPC/PVDF and YEC-8A/PVDF electrodes. The YEC-8A ACP particles had comparable sizes below 10 µm as described by the specifications, although still noticeably larger than the ACP particles from the Pureechem electrode. The WPC ACP particles, on the other hand, were coarser and contained impurities as evident by the brighter specks on the surface. This agrees with the WPC specifications which claimed a maximum ash content of 18 wt. % compared to YEC-8A at 0.5 wt. %. Also, the WPC/PVDF and YEC-8A/PVDF electrodes differed from the Pureechem electrode in that there were specks and strands of PVDF binder throughout the electrode surface holding the ACP particles together. The Pureechem electrode, on the other hand, used CMC as the binder which functioned to thicken the ACP electrode [177]. Lastly, the SEM images show that the Pureechem electrode had a more flat and compact structure, likely because of the pressing step in the electrode fabrication procedure.   72    Figure 4.10. SEM images of graphite foil (a), Pureechem electrode (b), WPC/PVDF electrode (c) and YEC-8A/PVDF electrode (d) with ACP-to-PVDF ratios of 85:15.   The SEM and EDS analyses also confirmed higher PVDF contents for electrodes with lower ACP-to-PVDF ratios. Figure 4.11. displays the EDS spectra and SEM images of the fabricated YEC-8A/PVDF electrodes with varying ACP-to-PVDF ratios. The electrode with the lowest ratio at 85:15 also had the largest fluorine peak in the EDS spectra, and more specks and strands of PVDF were observed in the SEM image. In contrast, the 95:5 electrode had a small fluorine peak in the EDS spectra, and the SEM image displayed mostly ACP particles with little amounts of PVDF specks and strands. (a) (b) (c) (d) 73     Figure 4.11. EDS spectra and SEM images of YEC-8A/PVDF electrodes with ACP-to-PVDF ratios of 85:15 (top), 90:10 (middle), and 95:5 (bottom). Solvent-to-solids ratio was 5:1 and dry mass loading was 5 mg/cm2.  74    Figure 4.11. also reveals that YEC-8A/PVDF electrodes had carbon and oxygen elements as expected. Moreover, the absence of other peaks showed that the amount of impurities from other elements were untraceable. In comparison, the WPC/PVDF electrode had many impurities with considerably large peaks for silicon and potassium, and small peaks for sodium, magnesium, aluminum and sulfur as seen in Figure 4.12. The WPC ACP was of lower quality than the YEC-8A ACP with a higher ash content of 18 wt. % compared to 0.5 wt. %. The higher ash content, which could be constituted of many different elements, was probably the reason for the increased amount of impurities observed in the EDS spectra. The EDS spectra of the Pureechem electrode did not have a fluorine peak because PVDF was not used as the binder. However, small chlorine, silicon, and sulfur peaks could be observed, which suggested some impurities were also present.     75    Figure 4.12. EDS spectra of WPC/PVDF electrode (top) and Pureechem electrode (bottom). The WPC/PVDF electrode had an ACP-to-PVDF ratio of 85:15, solvent-to-solids ratio of 1.5:1 and dry mass loading of 13.75 mg/cm2.  76   Finally, SEM and EDS analysis were also performed for the YEC-8A/PVDF electrode on carbon paper as shown in Figure 4.13. As expected, the surface morphology and elemental composition of the electrode on carbon paper was similar to the same electrode on graphite foil since there was no change in the ACP and binder materials or ratios used.   Figure 4.13. SEM image of carbon paper (a) and YEC-8A/PVDF electrode on carbon paper (b) with ACP-to-PVDF ratio of 85:15, solvent-to-solid ratio of 3:1, and dry mass loading of 2.30 mg/cm2. Below is the EDS spectrum of the YEC-8A/PVDF electrode on carbon paper (c).  (a) (b) (c) 77  The carbon paper consisted of fibers randomly interwoven between each other with some empty spaces in between as displayed in Figure 4.13a. In comparison, the graphite foil in Figure 4.10a. was mostly smooth and compact with some flakes. Cross-sectional SEM images of the Pureechem electrode, the YEC-8A electrode on graphite foil, and the YEC-8A electrode on carbon paper are shown in Figure 4.14.  Figure 4.14. Cross-sectional SEM images of the YEC-8A/PVDF electrode on graphite foil (a), YEC-8A/PVDF electrode on carbon paper (b), and Pureechem electrode (c).    (a) (b) (c) 78  The cross-sectional SEM images showed that the Pureechem electrode had the largest electrode thickness, followed by the YEC-8A/PVDF electrode on carbon paper and the YEC-8A/PVDF electrode on graphite foil. It can be estimated from the scaling of the SEM images that the electrode thicknesses were 250 µm, 120 µm, and 100 µm for the Pureechem, YEC-8A/PVDF on graphite foil, and YEC-8A/PVDF on carbon paper electrodes, respectively. Qualitatively, the YEC-8A/PVDF electrode on graphite foil also had a looser and more uneven structure. The Pureechem electrode, in contrast, had a compact and uniform structure. Lastly, it can be observed that graphite foil consisted of many flakes or layers stacked on each other, while carbon paper consisted of fibers of random length and orientation tangled together. 4.1.3. Cyclic Voltammetry  CV tests were performed to confirm and compare the capacitive behavior of the commercially obtained Pureechem electrode and the in-house fabricated YEC-8A/PVDF electrode on graphite foil. Figure 4.15. shows the cyclic voltammograms with the electrode potential converted from vs. Hg/Hg2SO4 to vs. SHE (Hg/Hg2SO4 vs. SHE = 0.68 V [174]).    79   Figure 4.15. Cyclic voltammograms of the ACP electrodes at a scan rate of 5 mV/s and an electrolyte concentration of 1000 ppm NaCl.  Both the Pureechem electrode and the YEC-8A/PVDF electrode on graphite foil displayed capacitive behavior with an absence of peaks, denoting that no electrochemical reactions occurred. Additionally, the Pureechem electrode reached higher current densities at the same electrode potential compared to the YEC-8A/PVDF electrode on graphite foil, suggesting that the Pureechem electrode had a lower resistance.  4.2. Desalination Experiments 4.2.1. Processing of Raw Data  The CDI desalination experiments yielded data for conductivity, pH, and current vs. time. Also, the temperature vs. time data was logged and used to automatically correct for the effects of temperature via linear compensation. Furthermore, pH corrections were made to the measured conductivities (see Appendix D.1.) and a calibration curve for conductivity and NaCl concentration was made (see Appendix D.2.).  -30-20-1001020300 0.2 0.4 0.6 0.8 1 1.2 1.4Current Density (A/cm2)E vs. SHE (V)PureechemYEC-8A/PVDF 85:1580    Figure 4.16. Example of raw data generated from desalination experiments. Data shown is from desalination experiment with Pureechem ACP electrodes without membrane at baseline conditions of 1.2 V, 1000 ppm NaCl, 20 mL/min, and 60 min cycle time. Conductivity and pH vs. time (top) and current vs. time (bottom) data were logged in 10 s intervals.  012345671.91.9522.052.12.150 5000 10000 15000 20000 25000pHConductivity (mS/cm)Time (s)ConductivitypH-0.15-0.1-0.0500.050.10.150 5000 10000 15000 20000 25000Current (A)Time (s)81   From the conductivity and pH vs. time and current vs. time data, salt adsorption capacity (SAC) was calculated from the maximum and minimum NaCl concentrations of each cycle. Furthermore, the charge efficiency (Λ) and specific energy consumption (η) was calculated by integrating the current for each cycle. In-depth descriptions of how these calculations were performed can be found in Appendix D.3.  4.2.2. Charge Efficiency and Specific Energy Consumption vs. Cycle Time  Desalination experiments were conducted at long cycle times of 60 min to ensure that the electrodes were at their full capacity and equilibrium was reached. However, the electrodes were saturated, i.e. the conductivity reached its minimum, in considerably less time than 60 min. Once the electrodes were saturated, current still moved through the system because of leakage currents, i.e. electrochemical reactions, resistive losses, or other pathways. Thus, when the charge efficiency and specific energy consumption was calculated at the end of the cycle time, worse values were obtained. To illustrate this, a graph of the charge efficiency and specific energy consumption plotted against cycle time is displayed in Figure 4.17.    82    Figure 4.17. Charge efficiency (top) and specific energy consumption (bottom) vs. cycle time of desalination experiments with Pureechem electrodes at baseline condition.   During the charge stage, charge efficiency started out near the ideal maximum value of 100%, then fell to a minimum pf 25%, rose back up to 50%, and then began steadily declining. This showed that there were start-up effects as charge built up on the electrode and ions began to migrate to them. The maximum point at approximately 20 min (1200 s) corresponded to the minimum conductivity, or electrode saturation. After electrode saturation, the charge efficiency 00.20.40.60.810 1000 2000 3000 4000 5000 6000 7000 8000Charge EfficiencyTime (s)00.00050.0010.00150.0020.00250.0030.00350 1000 2000 3000 4000Specific Energy Consumption (kWh/g)Time (s)Charge Discharge 83  steadily declined as ions could not be adsorbed onto the electrode anymore, but current was still supplied to the system. In the discharge stage, ions took a while to begin desorbing from the electrode as the system was short-circuited. After all the ions were desorbed, the charge efficiency similarly reached a maximum and steadily declined because of leakage currents. The specific energy consumption graph showed that a minimum is reached at approximately 15 min (900 s) and then slowly increased because energy is wasted on the leakage currents.  Overall, the charge efficiency and specific energy consumption results suggested that energy efficiency can be optimized by running the CDI system between 15 – 20 min. Moreover, it was more practical to compare charge efficiency and specific energy consumption values at their optimal values instead of at the end of the cycle time. Therefore, charge efficiency and specific energy consumption values were compared at 20 min instead of 60 min for the desalination experiments.  4.2.3. Adsorption Kinetics and Electrical Circuit Modelling  Model parameters were estimated with the Goal Seek function in Microsoft Excel, which used an iterative linear search method to converge upon a solution [178]. For the Lagergren pseudo-first order equation, the equilibrium salt adsorption (qe) and pseudo-first order rate constant (k1) were varied to minimize the RMSE. Similarly, qe and the pseudo-second order rate constant (k2) were varied to minimize the RMSE for the Lagergren pseudo-second order equation. Figure 4.18. shows a graph of the theoretical results calculated from the Lagergren adsorption kinetics model beside the experimental results.   84   Figure 4.18. Theoretical results from the Lagergren adsorption kinetics model and experimental results of the desalination experiments with Pureechem electrodes at baseline conditions.  With the Pureechem electrodes at the baseline experiment conditions, the pseudo-first order equation RMSE was 0.064, whereas the pseudo-second order RMSE was 0.165. However, as seen from the graph, both equations described the experimental data excellently. This suggested that the electrosorption of ions had negligible interactions between adsorbates [157].  The parameters for the modified RC series circuit with constant leakage current model was found by first using the current value at the end of the cycle as the leakage current term (Ileak). Then, the resistance (Rs) and capacitance (σ) were varied to minimize the RMSE. For the Randles circuit model, the series resistance (Rs) and parallel resistance (Rp) was found from the boundary conditions at the beginning and end of the cycle. Then, the capacitance (σ) was found by varying the parameter to minimize the RMSE. The fit of the two equivalent electrical circuit models with the experimental data is shown in Figure 4.19. 00.511.522.533.50 500 1000 1500 2000 2500 3000 3500 4000Salt Adsorption (mg/g)Time (t)ExperimentPseudo-1st-orderPseudo-2nd-order85   Figure 4.19. Theoretical results from the equivalent electrical circuit models and experimental results of the desalination experiments with Pureechem electrodes at baseline conditions.  Both models agreed reasonably with the experimental data. However, there were some instances where the models were off with their predictions. For the modified RC series with constant leakage current, the current was underpredicted at the beginning of the cycle but then closely followed the experimental data afterwards. On the other hand, the Randles circuit began by overpredicting the current but then underpredicted the current after 10 min (600 s). Quantitatively, the modified RC series model had a RMSE of 0.00256 and the Randles model had a RMSE of 0.00353. Therefore, this suggested the leakage current is better described by a constant current term than a resistor in parallel. 4.2.4. Commercial vs. Fabricated Electrodes The desalination performance and efficiency of the commercially obtained Pureechem electrode was tested and compared with three in-house fabricated electrodes in Table 4.1. 00.020.040.060.080.10.120.140 500 1000 1500 2000 2500 3000 3500 4000Current (A)Time (s)ExperimentRC SeriesRandles86   Pureechem YEC-8A/PVDF on carbon paper YEC-8A/PVDF on graphite foil WPC/PVDF on graphite foil SAC [mg/g] 2.43 ± 0.09 0.91 ± 0.08 0.24 ± 0.02 0.16 ± 0.01 Λ (20 min) 65 ± 2 % 25 ± 1 % 48 ± 1 % 5 ± 1 % η (20 min) [kWh/kg] 1.07 ± 0.06 3.23 ± 0.08 5.56 ± 0.09 34.3 ± 0.4 Table 4.1. Desalination metrics of the ACP electrodes at baseline conditions.  For salt adsorption capacity, charge efficiency, and specific energy consumption, the Pureechem electrode had exceptionally better values. The best in-house fabricated electrode was the YEC-8A/PVDF electrode on carbon paper, but it still performed more than two times worse than the Pureechem electrode for all desalination metrics measured. Since the Pureechem electrode was from a commercial vendor, the materials, such as type of activated carbon and polymeric binder, and fabrication processes were already optimized to yield maximum desalination performance and efficiency. The in-house fabricated electrodes, on the other hand, performed poorly for the desalination metrics measured, and further experimentation and optimization of the electrode materials and fabrication procedures is necessary to make electrodes that are on par with commercially available ones.  The salt adsorption capacity of the Pureechem electrode was 2.43 mg/g, a value that was consistent with previous results with activated carbon powder electrodes [36], although other results have shown higher values [84]. The specific energy consumption of the Pureechem electrode was 1.07 kWh/kg, which could be converted to a value of 0.86 kWh/m3 for treating 1000 ppm NaCl to a target concentration of 200 ppm NaCl. However, this disregarded the energy consumption of other equipment such as the pump. In comparison, RO for brackish water in real-world systems can have a total energy use between 0.3 – 2.8 kWh/m3 as seen from Table 1.2. [9]. Therefore, CDI has the potential to be competitive with RO for brackish water desalination, 87  especially at lower TDS concentrations. Further optimization of CDI technology may even lead to reduced total energy use and application to higher TDS concentrations. Since the Pureechem electrode had vastly superior desalination performance and efficiency, further desalination experiments were performed with the Pureechem electrodes to yield useful research results for practical applications. 4.3. Effect of Operating Parameters  To improve the understanding of operation of CDI systems, operating parameters were investigated which include applied voltage, ion-exchange membranes, NaCl concentration, and flow rate. Additionally, a 3 x 3 matrix of desalination experiments was performed for applied voltage and NaCl concentration to explore combined effects. 4.3.1. Effect of Applied Voltage and Ion-Exchange Membranes  The graphs of the desalination metrics at incremental applied voltages, and with and without ion-exchange membranes is shown in Figure 4.20. 88   Figure 4.20. Column graphs of salt adsorption capacity (a), charge efficiency (b), and specific energy consumption (c) for CDI and MCDI systems with varying applied voltages. Experiments were performed with 1000 ppm NaCl concentration and 20 mL/min flow rate. Error bars are shown for n = 3 trials.  As applied voltage increased, the salt adsorption capacity also increased significantly because a thicker electrical double layer was formed, and thus more ions were adsorbed on the electrode. However, the charge efficiency generally decreased, and the specific energy consumption mostly increased with increased voltages, although for some of the data points that are closer together the change was insignificant. This suggested that unwanted electrochemical reactions and resistive 01234560.8 1 1.2 1.5SAC [mg/g]Applied Voltage (V)CDIMCDI(a)50607080901000.8 1 1.2 1.5Λ (20 min) [(%]Applied Voltage (V)(b)00.20.40.60.811.21.40.8 1 1.2 1.5η (20 min) [kWh/kg]Applied Voltage (V)(c)89  losses were more abundant at higher voltages, and not only for applied voltages above 1.23 V, the SRP of water. Overall, the results obtained confirmed the findings from past studies for salt adsorption capacity [133]. On the other hand, for charge efficiency, past studies showed differing results. Mossad & Zou (2013) found that charge efficiency increased when applied voltage increased for their study, which contradicts the obtained results. However, their findings concluded that only a small change was observed from 65.5 % to 70.0 % when applied voltage was varied from 0.8 V to 1.6 V, and some of the changes were likely insignificant. Interestingly, they also found that specific energy consumption increased with higher voltages which agreed with the obtained results [47]. Huyskens et al. (2013), on the other hand, found that there was a reduction in charge efficiency with increasing voltage, which agreed with the obtained results [46].   MCDI, which includes ion-exchange membranes on the electrodes, was found to be superior to CDI for desalination performance and efficiency. Salt adsorption capacity was improved by 1.70 – 1.94 times and charge efficiency by 1.11 – 1.24 times for MCDI when compared with CDI. Moreover, specific energy consumption was higher in CDI systems by 1.21 – 1.35 times. By including ion-exchange membranes, co-ion expulsion was prevented so more ions could be retained in the electrodes and the electrosorption process became more energy efficient [124]. In comparison, Kim & Choi (2010) also observed improvements in salt adsorption capacity at 1.3 – 1.6 times and current efficiency at 2.1 – 2.4 times with ion-exchange membranes[179]. Therefore, ion-exchange membranes were extremely beneficial, and arguably indispensable, for boosting the performance and efficiency of CDI systems.  Table 4.2. shows the adsorption kinetics and electrical circuit model parameters at varying applied voltages for CDI and MCDI systems.  90  CDI  0.8 V 1.0 V 1.2 V 1.5 V Pseudo-first order kinetics qe = 1.81 mg/g k1 = 0.00102 RMSE = 0.075 qe = 2.14 mg/g k1 = 0.00096 RMSE = 0.144 qe = 2.39 mg/g k1 = 0.00140 RMSE = 0.064 qe = 3.28 mg/g k1 = 0.00142 RMSE = 0.138 Pseudo-second order kinetics qe = 1.99 mg/g k2 = 0.000866 RMSE = 0.117 qe = 2.13 mg/g k2 = 0.000968 RMSE = 0.189 qe = 2.55 mg/g k2 = 0.000905 RMSE = 0.165 qe = 3.51 mg/g k2 = 0.000811 RMSE = 0.296 Modified RC series circuit σ = 32.9 F Rs = 16.2 Ω Ileak = 6.05 mA RMSE = 0.0007 σ =28.3 F Rs = 19.6 Ω Ileak = 9.55 mA RMSE = 0.0005 σ = 32.0 F Rs = 17.8 Ω Ileak = 13.2 mA RMSE = 0.0015 σ = 30.6 F Rs = 17.5 Ω Ileak = 21.5 mA RMSE = 0.0058 Randles circuit σ = 42.6 F Rs = 13.3 Ω Rp = 119.9 Ω RMSE = 0.0009 σ = 41.6 Ω Rs = 16.6 Ω Rp = 88.1 Ω RMSE = 0.0006 σ = 44.6 F Rs = 12.4 Ω Rp = 78.6 Ω RMSE = 0.0027 σ = 45.4 F Rs = 9.2 Ω Rp = 60.7 Ω RMSE = 0.0079  MCDI  0.8 V 1.0 V 1.2 V 1.5 V Pseudo-first order kinetics qe = 2.78 mg/g k1 =0.00103 RMSE = 0.122 qe = 3.23 mg/g k1 = 0.00115 RMSE = 0.063 qe = 4.46 mg/g k1 = 0.00112 RMSE = 0.054 qe = 5.16 mg/g k1 = 0.00106 RMSE = 0.098 Pseudo-second order kinetics qe = 3.56 mg/g k2 = 0.000274 RMSE = 0.054 qe = 4.34 mg/g k2 = 0.000212 RMSE = 0.113 qe = 5.56 mg/g k2 = 0.000212 RMSE = 0.094 qe = 6.61 mg/g k2 = 0.000160 RMSE = 0.130 Modified RC series circuit σ = 50.2 F Rs = 17.6 Ω Ileak = 2.51 mA RMSE = 0.0025 σ = 51.6 F Rs = 13.2 Ω Ileak = 4.78 mA RMSE = 0.0036 σ = 44.9 F Rs = 18.2 Ω Ileak = 10.7 mA RMSE = 0.0052 σ = 45.1 F Rs = 15.9 Ω Ileak = 14.0 mA RMSE = 0.0065 Randles circuit σ = 54.71 F Rs = 12.7 Ω Rp = 305.6 Ω RMSE = 0.0027 σ = 58.1 F Rs = 7.6 Ω Rp = 201.6 Ω RMSE = 0.0088 σ = 50.9 F Rs = 8.5 Ω Rp = 145.4 Ω RMSE = 0.010 σ = 55.3 F Rs = 7.7 Ω Rp = 99.3 Ω RMSE = 0.015 Table 4.2. Lagergren adsorption kinetics and equivalent electrical circuit model parameters for CDI (top) and MCDI (bottom) systems with varying applied voltage.  As expected, the equilibrium salt adsorption (qe) predicted by both adsorption kinetics models increased when applied voltage was increased and when ion-exchange membranes were included, 91  which confirms the salt adsorption capacity results. The salt adsorption rate, as quantified by the rate constants k1 and k2, was found to be independent of the applied voltage. A previous study by Mossad & Zou (2013) using Lagergren adsorption kinetics models found that the rate constants increased when applied voltage was increased between 0.8 V and 1.6 V [47]. However, most of the other studies which used the same equations found no correlation between applied voltage and the rate constants, which suggested that different CDI systems may behave differently [34], [84], [180]. Although k1 values were comparable for MCDI and CDI, k2 values were lower for the MCDI experiments. This suggested that the salt adsorption rate was slower for MCDI systems, which is plausible because ions also must take time to diffuse across the membrane. The pseudo-first order equation was overall a better fit for the salt adsorption data with lower RMSE values in most cases. In the research literature, the verdict on which equation is a better fit is split with some claiming pseudo-first order [84], [143], [158] while others claim pseudo-second order [47], [144].  For the equivalent electrical circuit models, the modified RC series circuit with constant leakage current model was a better fit than the Randles circuit model from the RMSE values. This suggests that a constant leakage current term was a better descriptor of the circuit process than the parallel resistance term. Furthermore, some trends were deduced from the results. As voltage increased, the leakage current (Ileak) increased and parallel resistance (Rp) decreased, thus showing that more energy was wasted at higher voltages. No trend was observed for capacitance (σ) and series resistance (Rs) with higher voltages; however, the MCDI experiments had noticeably higher capacitances as well as lower leakage currents and higher parallel resistance compared to the CDI experiments. It should be noted that the electrical circuit models were imperfect since the model parameters did not remain constant with changing applied voltages even though voltage was included as a variable. Further work can be done to develop more accurate electrical circuit models. 92  4.3.2. Effect of NaCl Concentration  The desalination metrics results for NaCl concentrations ranging from 500 – 3000 ppm is shown in Figure 4.21.    Figure 4.21. Column graphs of salt adsorption capacity (a), charge efficiency (b), and specific energy consumption (c) at varying NaCl concentrations. Experiments were performed with 1.2 V applied voltage and 20 mL/min flow rate. Error bars are shown for n = 3 trials.   As NaCl concentration increased, salt adsorption capacity also increased. There was no obvious correlation for charge efficiency. In the literature, the effect of NaCl concentration on 00.511.522.53500 1000 2000 3000SAC [mg/g]NaCl Concentration [ppm](a)020406080500 1000 2000 3000Λ(20 min) [%]NaCl Concentration [ppm](b)00.20.40.60.811.21.4500 1000 2000 3000η(20 min) [kWh/kg]NaCl Concentration [ppm](c)93  charge efficiency was also not concrete, with some results that claimed lower NaCl concentrations have higher charge efficiencies [47] and others that found no effect [46]. Although errors were high, the graphs show that specific energy consumption decreased when going from 500 ppm to 1000 ppm but increased for 3000 ppm. For CDI, it has been established that the technology is more suitable for brackish water concentrations, which is lower than sea water concentrations. However, the results suggested that, at brackish water concentrations between 500 – 3000 ppm, there was a point around 2000 ppm where a minimum was reached for specific energy consumption. This was in line with experimental results from other researchers and electrical double layer model predictions [47].  The salt adsorption capacity at different NaCl concentrations can be expressed with adsorption isotherm models as shown in Figure 4.22.   Figure 4.22. Graph of adsorption isotherms and experimental data (left) and table of model parameters (right).   00.511.522.530 1000 2000 3000 4000Equilibrium Salt Adsorption [mg/g]NaCl Concentration [ppm]ExperimentLangmuirFreundlich94  The linearized plot of the equations and the calculation of model parameters is described in Appendix D.4. Both the Langmuir and Freundlich model predicted the experimental data quite accurately, with the Freundlich model having a better fit with a r2 value closer to unity. This was indicative of multilayer adsorption and adsorbate interactions, which seems plausible given that ions had electrostatic interactions in electrical double layers. Although most studies have found the opposite where the Langmuir model was a better fit, there have been a couple of results that have shown the Freundlich model was more appropriate [155].  The adsorption kinetics and electrical circuit model parameters while varying NaCl concentration is displayed in Table 4.3.  500 ppm 1000 ppm 2000 ppm 3000 ppm Pseudo-first order kinetics qe = 2.14 mg/g k1 = 0.00108 RMSE = 0.083 qe = 2.39 mg/g k1 = 0.00140 RMSE = 0.064 qe = 2.78 mg/g k1 = 0.00071 RMSE = 0.080 qe = 2.89 mg/g k1 = 0.00125 RMSE = 0.118 Pseudo-second order kinetics qe = 2.38 mg/g k2 = 0.000859 RMSE = 0.162 qe = 2.55 mg/g k2 = 0.000905 RMSE = 0.165 qe = 3.14 mg/g k2 = 0.000898 RMSE = 0.184 qe = 3.41 mg/g k2 = 0.000886 RMSE = 0.193 Modified RC series circuit σ = 29.1 F Rs = 29.45 Ω Ileak = 12.1 mA RMSE = 0.0014 σ = 32.0 F Rs = 17.8 Ω Ileak = 13.2 mA RMSE = 0.0015 σ = 32.0 F Rs = 26.4 Ω Ileak = 11.9 mA RMSE = 0.0003 σ = 31.1 F Rs = 25.0 Ω Ileak = 11.8 mA RMSE = 0.0004 Randles circuit σ = 38.4 F Rs = 19.5 Ω Rp = 79.5 Ω RMSE = 0.0008 σ = 44.6 F Rs = 12.4 Ω Rp = 78.6 Ω RMSE = 0.0027 σ = 43.4 F Rs = 21.0 Ω Rp = 80.1 Ω RMSE = 0.0002 σ = 43.4 F Rs = 21.8 Ω Rp = 79.6 Ω RMSE = 0.0002 Table 4.3. Lagergren adsorption kinetics and equivalent electrical circuit model parameters for different NaCl concentrations.  As NaCl concentration increased, the equilibrium salt adsorption (qe) also increased as expected from the salt adsorption capacity results. No other trends in the model parameters were observed. In the context of the research literature, both rate constants have been found to increase with higher 95  NaCl concentrations [59], [180]. On the flip side, there have also been cases where no trends were observed for both the rate constants with varying NaCl concentrations [143], [158]. 4.3.3. Effect of Flow Rate  The desalination metrics for the set of experiments investigating the effect of flow rate is displayed in Figure 4.23.    Figure 4.23. Column graphs of salt adsorption capacity (a), charge efficiency (b), and specific energy consumption (c) at varying flow rates. Experiments were performed with 1.2 V applied voltage and 1000 ppm NaCl concentration. Error bars are shown for n = 3 trials.  00.511.522.533.520 40 60Salt Adsorption Capacity (mg/g)Flow Rate (mL/min)(a)505560657020 40 60Λ(20 min) [%]Flow Rate [mL/min](b)00.20.40.60.811.21.420 40 60η (20 min) [kWh/kg]Flow Rate [mL/min](c)96  Flow rate in batch recirculation systems have a lesser effect than in single-pass systems because the same volume of water is treated. There was no significant effect of flow rate found for the desalination metrics. Theoretical models have found that salt removal was larger at higher flow rates for batch recirculation systems because of improved mixing [137], but this was not observed in the obtained results. A possible reason is that the model results were not applicable at the range of flow rates tested, which were quite high given the size of the CDI cell. The adsorption kinetics and electrical circuit model parameters are displayed in Table 4.4. Again, no significant correlation was found for the effect of flow rate on the model parameters, but there was noticeable variation between the results.  20 mL/min 40 mL/min 60 mL/min Pseudo-first order kinetics qe = 2.39 mg/g k1 = 0.00140 RMSE = 0.064 qe = 2.50 mg/g k1 = 0.00104 RMSE = 0.107 qe = 2.33 mg/g k1 = 0.00172 RMSE = 0.099 Pseudo-second order kinetics qe = 2.55 mg/g k2 = 0.000905 RMSE = 0.165 qe = 2.85 mg/g k2 = 0.000778 RMSE = 0.189 qe = 2.84 mg/g k2 = 0.000811 RMSE = 0.215 Modified RC series circuit σ = 32.0 F Rs = 17.8 Ω Ileak = 13.2 mA RMSE = 0.0015 σ = 32.6 F Rs = 16.7 Ω Ileak = 14.2 mA RMSE = 0.0022 σ = 27.9 F Rs = 21.6 Ω Ileak = 16.2 mA RMSE = 0.0005 Randles circuit σ = 44.6 F Rs = 12.4 Ω Rp = 78.6 Ω RMSE = 0.0027 σ = 45.6 F Rs = 11.1 Ω Rp = 73.3 Ω RMSE = 0.0033 σ = 42.4 F Rs = 16.8 Ω Rp = 57.3 Ω RMSE = 0.0005 Table 4.4. Lagergren adsorption kinetics and equivalent electrical circuit model parameters for different flow rates.  For further desalination experiments, a flow rate of 20 mL/min was chosen because lower flow rates were less variable and reduced the risk of clogging.  97  4.3.4. Combined Effects of Applied Voltage and NaCl Concentration  A 3 x 3 matrix of experiments was performed for applied voltages of 0.8 V, 1.0 V, 1.2 V and NaCl concentrations of 500 ppm, 1000 ppm, 2000 ppm. The desalination metrics results are tabulated in Table 4.5. Salt Adsorption Capacity [mg/g]  Applied Voltage (V) NaCl Concentration [ppm] 0.8 1.0 1.2 500 1.21±0.15 1.37±0.1 2.23±0.02 1000 1.68±0.01 1.85±0.07 2.42±0.06 2000 1.70±0.10 1.82±0.18 2.64±0.17  Charge Efficiency [%]  Applied Voltage (V) NaCl Concentration [ppm] 0.8 1.0 1.2 500 54.0±1.0 52.2±1.5 46.3±1.6 1000 69.1±1.7 68.3±1.9 64.8±1.4 2000 75.8±1.0 67.9±1.1 62.1±1.3  Specific Energy Consumption [kWh/kg]  Applied Voltage (V) NaCl Concentration [ppm] 0.8 1.0 1.2 500 0.68±0.06 0.88±0.06 1.22±0.09 1000 0.90±0.04 0.95±0.04 1.07±0.06 2000 0.91±0.03 0.92±0.04 0.94±0.08 Table 4.5. Salt adsorption capacity (top), charge efficiency (middle), and specific energy consumption (bottom) results for 3 x 3 matrix of experiments with varying applied voltage and NaCl concentration at a flow rate of 20 mL/min.  Since errors were high, some of the values overlapped and thus were not significantly different. Nonetheless, some information could still be extrapolated from the data. Salt adsorption capacity generally increased as applied voltage and NaCl concentration increased. The only 98  exception is at 1.0 V going from 1000 ppm to 2000 ppm NaCl, which was likely affected by high errors. Charge efficiency decreased as applied voltage increased, which was unsurprising because electrochemical reactions and resistive losses were more prominent at higher voltages. However, again, there was no trend between charge efficiency and NaCl concentration. Finally, specific energy consumption was like charge efficiency in that an increasing trend was observed for higher voltages, but no trend was observed for NaCl concentrations. At 0.8 V, it was noted that specific energy consumption decreased at lower NaCl concentrations, which agrees with trends found in previous studies [47]. It is possible that trends in desalination metrics are delicately dependant on the other operating parameters, which explains why contradicting results were often found between studies. The results shed some light on the combined effects of operating parameters. Generally, combined effects are difficult to investigate experimentally because of the large number of experiments required to deduce effects. Liu et al. (2015) conducted 189 CDI desalination experiments and used a multiplicative power-law relationship to model electrosorption as a function of applied voltage, spacer thickness, and retention time. Their findings revealed that applied voltage was the most important factor for electrosorption, and they could reproduce data with their model at r2 values above 0.85 [133]. Another approach is to use theoretical models to derive the combined effect of operating parameters, as was done by Zhao et al. (2013) in their study to optimize the salt adsorption rate [160]. The results obtained in this section add on to the body of research by providing experimental evidence that some individual trends stayed constant while others changed when operating parameters were varied in a matrix format. 99  4.4. Degradation and Regeneration  4.4.1. Long-Term Experiments  Long-term experiments were performed with pristine Pureechem electrodes at the baseline condition for 70 cycles. After 70 cycles, the Pureechem electrode was washed with either deionized water, 0.01 M citric acid, or 0.01 M NaOH to see if regeneration of the electrode’s desalination performance was possible. Figure 4.24. shows the salt adsorption capacities plotted against cycle number and the regenerated electrode results.  Figure 4.24. Salt adsorption capacity as a function of cycle number using Pureechem electrodes at baseline conditions. Regenerated electrode salt adsorption capacities are also displayed.   Salt adsorption capacity started out at 3.01 mg/g but fell rapidly to around 2.00 mg/g in approximately 10 cycles. Then, it started decreasing steadily and quite linearly up to the experimental limit of 70 cycles. Electrodes regenerated with deionized water and citric acid did 00.511.522.533.50 20 40 60 80 100Salt Adsorption Capacity (mg/g)Cycle NumberDI WaterCitric AcidNaOH100  not recover their desalination performance and maintained the downward trend of salt adsorption capacity. On the other hand, NaOH regenerated the electrode salt adsorption capacity to 2.35 mg/g where it remained relatively stable for at least 5 cycles.  The degradation of electrode performance has been attributed to oxidation of the anode (positive electrode) because of electrochemical reactions. Anode oxidation can lead to pore damage and loss of symmetry of the CDI cell [166]. One of the proposed electrochemical reactions for anode oxidation is as follows [168]: C + 2H2O →  CO2 +  4H+ +  4e−    E0 = 0.7 – 0.9 V  (19) The exact SRP of the reaction depends on the activated carbon molecular structures, which can vary widely. Furthermore, oxidation of organic carbon molecules can also produce functional groups such as alcohols, aldehydes, ketones, and carboxylic acids. However, organic electrochemistry is complex and not well-studied, so no detailed reactions and SRPs could be found. As seen from reaction (19), H+ ions are produced, which makes the solution more acidic. This was also observed experimentally, as the pH fluctuated in phase with the charge and discharge steps, and tended to become acidic over time. Cohen et al. (2013) also observed the same pH drift and fluctuation in accordance with the cycling [166]. 101   Figure 4.25. Snapshot of long-term desalination experiment showing pH fluctuations and tendency toward acidity.  The mechanism by which NaOH reverses the anode oxidation was hypothesized to be neutralization of the surface carboxylic acid functional groups and recovery of symmetry. Carboxylic acids and NaOH react to form water and a Na-carboxylic acid salt. The Na-carboxylic acid salt is soluble and therefore dissolves in the wash solution, reversing the pore damage and recovering the symmetry. For electrode regeneration, it has also been established that acid cleaning was effective in reversing inorganic scaling, while alkaline cleaning could remove organic fouling [31]. It is possible that the mechanism for removing organic fouling is similar to the mechanism for regenerating degraded electrodes. Humic acid is one of the main components of organic matter and has been investigated as a foulant for CDI electrodes. Wang et al. (2015) conducted electrode regeneration tests and discovered that humic-like substances were mostly found in alkaline wash waters [34]. This lended credibility to the hypothesis of oxidation as the electrode degradation and regeneration mechanism. However, there has also been contradictory evidence that showed that oxidation of electrodes before desalination increased the performance and efficiency [57]. 01234560 1000 2000 3000 4000 5000 6000 7000pHTime (s)102  Therefore, loss of symmetry probably also played a large role and the mechanisms may be more complex than what was hypothesized.  4.4.2. Electrode Potential vs. Reference Electrode  Electrode potential against a Hg/HgO reference electrode was measured at baseline conditions to determine whether electrochemical reactions were thermodynamically possible. Figure 4.26. displays the plot of electrode potential over the course of a desalination experiment.  Figure 4.26. Electrode potential of the anode and cathode in the CDI cell over time at baseline conditions.  Electrode potential was converted to vs. SHE for easier analysis of the results. The original plot vs. Hg/HgO can be found in Appendix F. The electrode potential for the anode ranged from around 0.2 – 0.8 V, while the cathode ranged from -0.4 – 0.2 V. Additionally, the electrode potential was below the SRP of water electrolysis at 1.23 V and did not drift for the three cycles. The electrode -0.6-0.4-0.200.20.40.60.811.20 5000 10000 15000 20000 25000Electrode Potential  vs. SHE (V)Time (s)vs. cathodevs. anode103  potential of the anode could reach up to 0.8 V, and therefore it was likely that carbon oxidation as described in equation (19) could thermodynamically occur.  4.4.2. SEM Images and EDS Spectra  To further explore the electrode degradation mechanism, SEM and EDS analysis was performed on the regenerated electrodes. The deionized water and citric acid washed electrodes had indistinguishable results with the pristine electrodes, and are shown in Appendix E. The SEM and EDS results for the NaOH washed electrode is shown in Figure 4.27.  Figure 4.27. SEM image and EDS spectra of 0.01 M NaOH washed Pureechem electrode after degradation.  No observable differences were seen in the SEM images, meaning that surface morphology changes did not occur at the micron scale. For the EDS spectra, on the other hand, a large Na peak was seen which was not present for the pristine electrode. The presence of Na could be because 104  they were retained on the carboxylic acid groups on the activated carbon particles. It is possible that only certain Na-carboxylic acids could dissolve in water, such as those with lower molecular weight and were thus more polar. Overall, the SEM and EDS results supported the possible hypothesis for oxidation as the electrode degradation and regeneration mechanism; nevertheless, more concrete evidence is required to further cement the hypothesis.              105  Chapter 5: Conclusion and Recommendations  5.1. Summary of Work  The results of this research project contribute to the body of knowledge in the literature for conventional CDI desalination using activated carbon electrodes. Also, the findings represent the preliminary step toward more advanced research such as using electrodes with novel materials, innovative cell designs, and new operation modes.   The conclusions that were drawn can be summarized in the following points. • Structural integrity of activated carbon powder electrodes could be improved by including more polymeric binder, using coarser activated carbon powders, and lowering the mass loading. However, coarse particles also had lower surface areas and low mass loadings which limited the amount of adsorbent material. Therefore, a balance must be struck. • Commercially obtained Pureechem electrodes had superior desalination performance to even the best in-house fabricated YEC-8A/PVDF electrodes on carbon paper. According to the SEM and CV results, the Pureechem electrode had smaller particle sizes, more compact structure, and lower electrical resistance. • Including ion-exchange membranes greatly improved the desalination performance and efficiency of the CDI process. Increasing applied voltage increased the salt adsorption capacity but decreased the charge efficiency and increased the specific energy consumption. Salt adsorption capacity also increased when NaCl concentration increased, but there was no effect on charge efficiency and specific energy consumption. No effect on the desalination metrics was observed for flow rate. 106  • Lagergren adsorption kinetics and equivalent electrical circuit models theoretically and mathematically described the electrosorption process. In general, model parameters varied greatly, and no conclusions could be reliably drawn for trends. • An experimental matrix with applied voltage and NaCl concentration revealed that individual trends could be observed such as increased salt adsorption capacity at higher applied voltages and NaCl concentrations, and increased charge efficiency and decreased specific energy consumption at lower applied voltages. • It was discovered through long-term experiments that electrode degradation happened mostly at a steady rate, reaching half of its original performance at approximately 50 cycles. The degradation mechanism was hypothesized to be anode oxidation based on pH fluctuations and reference electrode measurements.  • Degraded electrodes could be regenerated to an acceptable desalination performance by washing with NaOH solution, as opposed to citric acid solution or deionized water. It was further hypothesized based on pH data, SEM and EDS results that regeneration occurred because NaOH neutralized the carboxylic acid functional groups on the electrode, which then dissolved into solution, reversed pore damage and recovered symmetry in the electrodes.  5.2. Future Development and Recommendations  Further research can build upon the preliminary findings in this project. A list of the areas of the research project that could be considered incomplete is shown and described below. • The in-house fabricated electrodes performed poorly compared to the commercially purchased ones. It is prudent to make more careful selection of the activated carbon powder, 107  polymeric binder, and current collector to fabricate electrodes with excellent desalination performance and efficiencies. Activated carbon with large mesopore volumes, which are ideal for electrical double layer formation, can be used. Also, hydrophilicity, which is a desired quality for CDI electrodes, can be improved by using a water-stable hydrophilic binder. Analytical techniques such as SEM, EDS, and CV can yield valuable insight on the selection of these materials. Furthermore, fabrication processes can also be optimized by exploring the use of sophisticated equipment such as automatic film applicators. • More work can be done by testing more operating parameter values in a larger range with further trials to clarify the effects and extend the applicability. In the research literature, there are often contradictory results where some studies find significant effects while others find no effects. In addition, the matrix of experiments with different operating parameters can be better investigated with improved experimental design and statistical methods. • The degradation results are preliminary and more evidence to support the claims about the electrode degradation and regeneration mechanism is needed. Investigating the mechanism should then give insight on how to delay or prevent degradation. Also, the electrode regeneration should be further explored with the goal to establish a practical regeneration procedure, so maintenance plans can be developed once the CDI system is in application. Real-world CDI systems must be able to operate for as long as possible, and therefore there is much needed work to improve the long-term desalination performance efficiency for practical applications. Additionally, potential directions and recommendations for the research and development of CDI as a desalination technology is listed below. 108  • There have been a wide variety of novel materials which have been explored for CDI desalination. As nanomaterials and nanotechnology becomes more practical and common in the real-world, it is likely they will start finding applications in electrodes for CDI since they have excellent surface area, pore size distribution, and conductivity properties. Other materials that are promising is biochar, which is renewable and low in cost. Biochar materials have exhibited exceptional electrical double layer properties in previous studies and thus are a strong candidate for application as CDI electrodes [181]. • The Lagergren adsorption kinetics and equivalent electrical circuit models are relatively crude descriptions of the electrosorption process. Both have assumptions that do not necessary apply to electrosorption, and therefore its accuracy suffers accordingly. Even though applied voltage was included as a variable in the equivalent electrical circuit models, the model parameters were still changed when applied voltage was varied. This suggests that the electrosorption process was more complex, and more rigorous models based on electrical double layer and porous media transport theory could potentially improve the prediction of results. • Practical CDI systems will operate in salt water with complex matrices of inorganic ions and organic matter. Thus, an issue that arises is fouling and scaling of the electrode in the water matrix. Future studies can focus on the predominant species which are present in brackish or sea water, and the mechanisms by which electrode fouling and scaling occur. Further studies can then investigate prevention of degradation or regeneration of electrodes, or pretreatment steps to eliminate species that are particularly damaging.  • Lastly, pilot and field demonstrations of CDI technology is scarce, and therefore CDI as a desalination technology is extremely inexperienced even though there is a wealth of 109  laboratory studies in the research literature. Through pilot and field studies, essential information can be revealed about the scalability, costs, and limitations of the technology in the real-world [12].                 110  Bibliography  [1] United Nations, “Clean Water and Sanitation: Why It Matters,” p. 2, 2016. [2] B. Ronayne, C. Chiu, C. Hyman, R. Alberto, D. Gagnon, A. Horvath, J. Lawn, and K. Sawatzky, Fit to Drink: Challenges in Providing Safe Drinking Water in British Columbia, no. 32. 2008. [3] British Columbia Ministry of Health (BCMOH), “Small Water System Guidebook,” no. December, 2013. [4] H. G. Gorchev and G. Ozolins, WHO guidelines for drinking-water quality., 4th ed. Geneva, Switzerland: World Health Organization, 2011. [5] Health Canada, “Guidelines for Canadian Drinking Water Quality Summary Table Prepared by the Federal-Provincial-Territorial Committee on Drinking Water of the Federal-Provincial-Territorial Committee on Health and the Environment March 2006,” Environments, no. October 2014. pp. 1–16, 2012. [6] J. C. Crittenden, R. R. Trussell, D. W. Hand, K. j. Howe, and G. Tchobanoglous, “Introduction,” in MWH’s Water Treatment: Principles and Design, 3rd ed., Hoboken, NJ: John Wiley & Sons, Inc., 2012, pp. 1–16. [7] Alaska Department of Environmental Conservation, “Introduction to Water Treatment,” in Introduction to Small Water Systems, 2015, pp. 95–145. [8] I. Shiklomanov, “World fresh water resources,” in Water in Crisis: A Guide to the World’s Fresh Water Resources, P. H. Gleick, Ed. New York: Oxford University Press, 1993, pp. 13–24. [9] N. Voutchkov, “Desalination Engineering: An Overview,” in Desalination Engineering: Planning and Design, New York City, NY: The McGraw-Hills Companies, Inc., 2012, pp. 1–12. [10] Health Canada, “Total Dissolved Solids (TDS),” Guidelines for Canadian Drinking Water Quality - Technical Documents. pp. 22–24, 1991. [11] M. A. Anderson, A. L. Cudero, and J. Palma, “Capacitive deionization as an electrochemical means of saving energy and delivering clean water. Comparison to present desalination practices: Will it compete?,” Electrochim. Acta, vol. 55, no. 12, pp. 3845–3856, 2010. [12] F. a. AlMarzooqi, A. a. Al Ghaferi, I. Saadat, and N. Hilal, “Application of Capacitive Deionisation in water desalination: A review,” Desalination, vol. 342, pp. 3–15, 2014. [13] “Desalination by the Numbers,” International Desalination Association, 2017. [Online]. Available: http://idadesal.org/desalination-101/desalination-by-the-numbers/. [Accessed: 22-Aug-2017]. [14] L. Weinstein and R. Dash, “Capacitive Deionization: Challenges and Opportunities,” Desalin. Water Reuse, no. December, pp. 34–37, 2013. 111  [15] S. Porada, R. Zhao,  a. Van Der Wal, V. Presser, and P. M. Biesheuvel, “Review on the science and technology of water desalination by capacitive deionization,” Prog. Mater. Sci., vol. 58, no. 8, pp. 1388–1442, 2013. [16] R. L. Zornitta, J. J. Lado, M. A. Anderson, and L. A. M. Ruotolo, “Effect of electrode properties and operational parameters on capacitive deionization using low-cost commercial carbons,” Sep. Purif. Technol., vol. 158, pp. 39–52, 2016. [17] S. Porada, L. Borchardt, M. Oschatz, M. Bryjak, J. S. Atchison, K. J. Keesman, S. Kaskel, P. M. Biesheuvel, and V. Presser, “Direct prediction of the desalination performance of porous carbon electrodes for capacitive deionization,” Energy Environ. Sci., vol. 6, no. 12, p. 3700, 2013. [18] Y. Liu, C. Nie, X. Liu, X. Xu, and L. Pan, “Review on carbon-based composite materials for capacitive deionization,” RSC Adv., vol. 5, pp. 15205–15225, 2015. [19] Z.-H. Huang, Z. Yang, F. Kang, and M. Inagaki, “Carbon electrodes for capacitive deionization,” J. Mater. Chem. A, vol. 5, no. 2, pp. 470–496, 2017. [20] R. Zhao, S. Porada, P. M. Biesheuvel, and A. Van der Wal, “Energy consumption in membrane capacitive deionization for different water recoveries and flow rates, and comparison with reverse osmosis,” Desalination, vol. 330, pp. 35–41, 2013. [21] H. Strathmann, “Electrodialysis, a mature technology with a multitude of new applications,” Desalination, vol. 264, no. 3, pp. 268–288, 2010. [22] E. García-Quismondo, C. Santos, J. Palma, and M. a. Anderson, “On the challenge of developing wastewater treatment processes: capacitive deionization,” Desalin. Water Treat., no. August 2015, pp. 1–10, 2014. [23] “Product - Ecomite,” Pureechem, 2014. [Online]. Available: http://www.pureechem.com/en/product/sub02.html. [Accessed: 23-Aug-2017]. [24] “Commercial Systems - Products - Aqua EWP,” Aqua EWP, 2016. [Online]. Available: http://aquaewp.com/product-category/commercial-systems/. [Accessed: 23-Aug-2017]. [25] “Radial Deionization - Atlantis Technologies,” Atlantis Technologies, 2017. [Online]. Available: http://www.atlantis-water.com/technoloogy/. [Accessed: 23-Aug-2017]. [26] “ENPAR COMMENCES CONSTRUCTION OF WORLD CLASS ESD WATER TREATMENT PLANT - RECEIVES $800,000 INITIAL PAYMENT FROM CHEMSBRO,” Enpar Technologies, 2017. [Online]. Available: http://www.enpar-tech.com/news/~70-ENPAR-COMMENCES-CONSTRUCTION-OF-WORLD-CLASS-ESD-WATER-TREATMENT-PLANT---RECEIVES-800-000-INITIAL-PAYMENT-FROM-CHEMSBRO. [Accessed: 23-Aug-2017]. [27] Y. Oren, “Capacitive deionization (CDI) for desalination and water treatment - past, present and future (a review),” Desalination, vol. 228, no. 1–3, pp. 10–29, 2008. [28] “INDUSTRIAL - Voltea,” Voltea, 2017. [Online]. Available: http://voltea.com/industrial-english/. [Accessed: 23-Aug-2017]. 112  [29] J. B. Lee, K. K. Park, S. W. Yoon, P. Y. Park, K. Il Park, and C. W. Lee, “Desalination performance of a carbon-based composite electrode,” Desalination, vol. 237, no. 1–3, pp. 155–161, 2009. [30] M. Mossad and L. Zou, “A study of the capacitive deionisation performance under various operational conditions,” J. Hazard. Mater., vol. 213–214, pp. 491–497, 2012. [31] M. Mossad and L. Zou, “Study of fouling and scaling in capacitive deionisation by using dissolved organic and inorganic salts,” J. Hazard. Mater., vol. 244–245, pp. 387–393, 2013. [32] M. Mossad, W. Zhang, and L. Zou, “Using capacitive deionisation for inland brackish groundwater desalination in a remote location,” Desalination, vol. 308, no. October 2011, pp. 154–160, 2013. [33] J. S. Kim, C. S. Kim, H. S. Shin, and J. W. Rhim, “Application of synthesized anion and cation exchange polymers to membrane capacitive deionization (MCDI),” Macromol. Res., vol. 23, no. 4, pp. 360–366, 2015. [34] C. Wang, H. Song, Q. Zhang, B. Wang, and A. Li, “Parameter optimization based on capacitive deionization for highly efficient desalination of domestic wastewater biotreated effluent and the fouled electrode regeneration,” Desalination, vol. 365, pp. 407–415, 2015. [35] G. Lu, G. Wang, P.-H. Wang, Z. Yang, H. Yan, W. Ni, L. Zhang, and Y.-M. Yan, “Enhanced capacitive deionization performance with carbon electrodes prepared with a modified evaporation casting method,” Desalination, vol. 386, pp. 32–38, 2016. [36] C. H. Hou, J. F. Huang, H. R. Lin, and B. Y. Wang, “Preparation of activated carbon sheet electrode assisted electrosorption process,” J. Taiwan Inst. Chem. Eng., vol. 43, no. 3, pp. 473–479, 2012. [37] I. Cohen, E. Avraham, Y. Bouhadana, A. Soffer, and D. Aurbach, “The effect of the flow-regime, reversal of polarization, and oxygen on the long term stability in capacitive de-ionization processes,” Electrochim. Acta, vol. 153, pp. 106–114, 2015. [38] A. Subramani and J. G. Jacangelo, “Emerging desalination technologies for water treatment: A critical review,” Water Res., vol. 75, pp. 164–187, 2015. [39] M. Suss, S. Porada, X. Sun, M. Biesheuvel, J. Yoon, and V. Presser, “Water desalination via capacitive deionization: what is it and what can we expect from it?,” Energy Environ. Sci., vol. 8, no. Cdi, pp. 2296–2319, 2015. [40] H. Li, J. Wang, Q. Chu, Z. Wang, F. Zhang, and S. Wang, “Theoretical and experimental specific capacitance of polyaniline in sulfuric acid,” J. Power Sources, vol. 190, no. 2, pp. 578–586, 2009. [41] A. Allagui, T. J. Freeborn, A. S. Elwakil, and B. J. Maundy, “Reevaluation of Performance of Electric Double-layer Capacitors from Constant-current Charge/Discharge and Cyclic Voltammetry,” Sci. Rep., vol. 6, no. 1, p. 38568, 2016. [42] “Testing Electrochemical Capacitors Part 1: CV, EIS, and Leakage Current,” Gamry 113  Instruments, 2017. [Online]. Available: https://www.gamry.com/application-notes/battery-research/testing-electrochemical-capacitors-cyclic-voltammetry-leakage-current/. [Accessed: 20-Sep-2017]. [43] “Basics of Electrochemical Impedance Spectroscopy,” Gamry Instruments, 2017. [Online]. Available: https://www.gamry.com/application-notes/EIS/basics-of-electrochemical-impedance-spectroscopy/. [Accessed: 20-Sep-2017]. [44] Y. Zhao, X. min Hu, B. hui Jiang, and L. Li, “Optimization of the operational parameters for desalination with response surface methodology during a capacitive deionization process,” Desalination, vol. 336, no. 1, pp. 64–71, 2014. [45] C. H. Hou and C. Y. Huang, “A comparative study of electrosorption selectivity of ions by activated carbon electrodes in capacitive deionization,” Desalination, vol. 314, pp. 124–129, 2013. [46] C. Huyskens, J. Helsen, and A. B. de Haan, “Capacitive deionization for water treatment: Screening of key performance parameters and comparison of performance for different ions,” Desalination, vol. 328, pp. 8–16, 2013. [47] M. Mossad and L. Zou, “Evaluation of the salt removal efficiency of capacitive deionisation: Kinetics, isotherms and thermodynamics,” Chem. Eng. J., vol. 223, pp. 704–713, 2013. [48] S. Song, Y. Zhang, S. Bao, and W. E. I. Huang, “Desalination By Capacitive Deionization With Carbon-Based Materials As Electrode: a Review,” Surf. Rev. Lett., vol. 20, no. 6, p. 1330003, 2013. [49] J. B. Cordon, “An Overview of Physisorption,” in Surface Area and Porosity Determinations by Physisorption: Measurements and Theory, Amsterdam: Elsevier, 2006, pp. 1–28. [50] B. Zdravkov, J. Čermák, M. Šefara, and J. Janků, “Pore classification in the characterization of porous materials: A perspective,” Open Chem., vol. 5, no. 2, pp. 385–395, 2007. [51] C. Tsouris, R. Mayes, J. Kiggans, K. Sharma, S. Yiacoumi, D. Depaoli, and S. Dai, “Mesoporous carbon for capacitive deionization of saline water,” Environ. Sci. Technol., vol. 45, no. 23, pp. 10243–10249, 2011. [52] G. Wang, B. Qian, Q. Dong, J. Yang, Z. Zhao, and J. Qiu, “Highly mesoporous activated carbon electrode for capacitive deionization,” Sep. Purif. Technol., vol. 103, pp. 216–221, 2013. [53] J. Choi, “Fabrication of a carbon electrode using activated carbon powder and application to the capacitive deionization process,” vol. 70, pp. 362–366, 2010. [54] M. K. B. Gratuito, T. Panyathanmaporn, R. A. Chumnanklang, N. Sirinuntawittaya, and A. Dutta, “Production of activated carbon from coconut shell: Optimization using response surface methodology,” Bioresour. Technol., vol. 99, no. 11, pp. 4887–4895, 2008. 114  [55] M. Aslan, M. Zeiger, N. Jäckel, I. Grobelsek, D. Weingarth, and V. Presser, “Improved capacitive deionization performance of mixed hydrophobic/hydrophilic activated carbon electrodes,” J. Phys. Condens. Matter, vol. 28, no. 11, p. 114003, 2016. [56] I. Villar, S. Roldan, V. Ruiz, M. Granda, C. Blanco, R. Menéndez, and R. Santamaría, “Capacitive deionization of NaCl solutions with modified activated carbon electrodes,” Energy and Fuels, vol. 24, no. 6, pp. 3329–3333, 2010. [57] W. Huang, Y. Zhang, S. Bao, R. Cruz, and S. Song, “Desalination by capacitive deionization process using nitric acid-modified activated carbon as the electrodes,” Desalination, vol. 340, no. 1, pp. 67–72, 2014. [58] D. Wang, C. Li, J. Guo, and T. Li, “Carbon electrode modified by KOH solution to improve performance of capacitive desalination,” Desalin. Water Treat., vol. 3994, no. October, pp. 1–7, 2015. [59] C.-L. Yeh, H.-C. Hsi, K.-C. Li, and C.-H. Hou, “Improved performance in capacitive deionization of activated carbon electrodes with a tunable mesopore and micropore ratio,” Desalination, vol. 367, pp. 60–68, 2015. [60] K. K. Park, J. B. Lee, P. Y. Park, S. W. Yoon, J. S. Moon, H. M. Eum, and C. W. Lee, “Development of a carbon sheet electrode for electrosorption desalination,” Desalination, vol. 206, no. 1–3, pp. 86–91, 2007. [61] C. H. Fang, P. I. Liu, L. C. Chung, H. Shao, C. H. Ho, R. S. Chen, H. T. Fan, T. M. Liang, M. C. Chang, and R. Y. Horng, “A flexible and hydrophobic polyurethane elastomer used as binder for the activated carbon electrode in capacitive deionization,” Desalination, vol. 399, pp. 34–39, 2016. [62] Y. Zhang, J. Guo, and T. Li, “Research Progress on Binder of Activated Carbon Electrode,” Adv. Mater. Res., vol. 549, pp. 780–784, 2012. [63] B. M. Asquith, J. Meier-Haack, and B. P. Ladewig, “Poly(arylene ether sulfone) copolymers as binders for capacitive deionization activated carbon electrodes,” Chem. Eng. Res. Des., vol. 104, pp. 81–91, 2015. [64] P. D. A. McCrae, T. Zhang, and D. R. B. Walker, “Method of Making Shaped Activated Carbon,” US 6,573,212 B2, 2003. [65] S. Nadakatti, M. Tendulkar, and M. Kadam, “Use of mesoporous conductive carbon black to enhance performance of activated carbon electrodes in capacitive deionization technology,” vol. 268, pp. 182–188, 2011. [66] C. Yan, L. Zou, and R. Short, “Polyaniline-modified activated carbon electrodes for capacitive deionisation,” Desalination, vol. 333, no. 1, pp. 101–106, 2014. [67] J.-H. Choi, “Comparison of constant voltage (CV) and constant current (CC) operation in the membrane capacitive deionisation process,” Desalin. Water Treat., vol. 3994, no. November, p. doi:10.1080/19443994.2014.942379, 2014. [68] Y. J. Kim and J. H. Choi, “Improvement of desalination efficiency in capacitive deionization using a carbon electrode coated with an ion-exchange polymer,” Water Res., 115  vol. 44, no. 3, pp. 990–996, 2010. [69] M. A. Anderson, “Optimizing the Energy Efficiency of Capacitive Deionization Reactors Working under Real-World Conditions,” 2013. [70] Z. Huang, L. Lu, Z. Cai, and Z. J. Ren, “Individual and competitive removal of heavy metals using capacitive deionization,” J. Hazard. Mater., vol. 302, pp. 323–331, 2016. [71] A. M. Dehkhoda, “Development and Characterization of Activated Biochar as Electrode Material for Capacitive Deionization,” The Universeity of British Columbia, 2016. [72] P.-I. Liu, L.-C. Chung, C.-H. Ho, H. Shao, T.-M. Liang, M.-C. Chang, C.-C. M. Ma, and R.-Y. Horng, “Comparative insight into the capacitive deionization behavior of the activated carbon electrodes by two electrochemical techniques,” Desalination, vol. 379, pp. 34–41, 2016. [73] S. Porada, M. Bryjak, A. Van Der Wal, P. M. Biesheuvel, A. Van Der Wal, P. M. Biesheuvel, A. Van Der Wal, and P. M. Biesheuvel, “Effect of electrode thickness variation on operation of capacitive deionization,” Electrochim. Acta, vol. 75, pp. 148–156, 2012. [74] B. Jia and W. Zhang, “Preparation and Application of Electrodes in Capacitive Deionization (CDI): a State-of-Art Review,” Nanoscale Res. Lett., vol. 11, no. 1, p. 64, 2016. [75] S. Chung, H. Kang, J. D. Ocon, J. K. Lee, and J. Lee, “Enhanced electrical and mass transfer characteristics of acid-treated carbon nanotubes for capacitive deionization,” Curr. Appl. Phys., vol. 15, no. 11, pp. 1539–1544, 2015. [76] H. Li, L. Pan, T. Lu, Y. Zhan, C. Nie, and Z. Sun, “A comparative study on electrosorptive behavior of carbon nanotubes and graphene for capacitive deionization,” J. Electroanal. Chem., vol. 653, no. 1–2, pp. 40–44, 2011. [77] A. Zurutuza and C. Marinelli, “Challenges and opportunities in graphene commercialization,” Nat. Nanotechnol., vol. 9, no. 10, pp. 730–734, 2014. [78] Y. J. Kim and J. H. Choi, “Enhanced desalination efficiency in capacitive deionization with an ion-selective membrane,” Sep. Purif. Technol., vol. 71, no. 1, pp. 70–75, 2010. [79] P. M. Biesheuvel, R. Zhao, S. Porada, and A. van der Wal, “Theory of membrane capacitive deionization including the effect of the electrode pore space,” J. Colloid Interface Sci., vol. 360, no. 1, pp. 239–248, 2011. [80] S. Porada, L. Weinstein, R. Dash,  der W. van  A., M. Bryjak, Y. Gogotsi, and P. M. Biesheuvel, “Water Desalination Using Capacitive Deionization with Microporous Carbon Electrodes.,” ACS Appl. Mater. Interfaces, vol. 4, no. 3, pp. 1194–1199, 2012. [81] R. Zhao, P. M. Biesheuvel, H. Miedema, H. Bruning, and A. van der Wal, “Charge efficiency: A functional tool to probe the double-layer structure inside of porous electrodes and application in the modeling of capacitive deionization,” J. Phys. Chem. Lett., vol. 1, no. 1, pp. 205–210, 2010. 116  [82] L. Zou, L. Li, H. Song, and G. Morris, “Using mesoporous carbon electrodes for brackish water desalination,” Water Res., vol. 42, no. 8–9, pp. 2340–2348, 2008. [83] L. Li, L. Zou, H. Song, and G. Morris, “Ordered mesoporous carbons synthesized by a modified sol-gel process for electrosorptive removal of sodium chloride,” Carbon N. Y., vol. 47, no. 3, pp. 775–781, 2009. [84] Z. Chen, C. Song, X. Sun, H. Guo, and G. Zhu, “Kinetic and isotherm studies on the electrosorption of NaCl from aqueous solutions by activated carbon electrodes,” DES, vol. 267, no. 2–3, pp. 239–243, 2011. [85] J. Yang, L. Zou, H. Song, and Z. Hao, “Development of novel MnO2/nanoporous carbon composite electrodes in capacitive deionization technology,” Desalination, vol. 276, no. 1–3, pp. 199–206, 2011. [86] P. I. Liu, L. C. Chung, H. Shao, T. M. Liang, R. Y. Horng, C. C. M. Ma, and M. C. Chang, “Microwave-Assisted ionothermal synthesis of nanostructured anatase titanium dioxide/activated carbon composite as electrode material for capacitive deionization,” Electrochim. Acta, vol. 96, pp. 173–179, 2013. [87] C. Kim, J. Lee, S. Kim, and J. Yoon, “TiO2 sol-gel spray method for carbon electrode fabrication to enhance desalination efficiency of capacitive deionization,” Desalination, vol. 342, pp. 70–74, 2014. [88] H. Li, L. Pan, C. Nie, Y. Liu, and Z. Sun, “Reduced graphene oxide and activated carbon composites for capacitive deionization,” J. Mater. Chem., vol. 22, no. 31, p. 15556, 2012. [89] Y. Bouhadana, E. Avraham, A. Soffer, and D. Aurbach, “Several Basic and Practical Aspects Related to Electrochemical Deionization of Water,” AlChE J., vol. 56, no. 3, pp. 779–789, 2010. [90] M. W. Ryoo, J. H. Kim, and G. Seo, “Role of titania incorporated on activated carbon cloth for capacitive deionization of NaCl solution,” J. Colloid Interface Sci., vol. 264, no. 2, pp. 414–419, 2003. [91] M. W. Ryoo and G. Seo, “Improvement in capacitive deionization function of activated carbon cloth by titania modification,” Water Res., vol. 37, no. 7, pp. 1527–1534, 2003. [92] M. T. Z. Myint, S. H. Al-Harthi, and J. Dutta, “Brackish water desalination by capacitive deionization using zinc oxide micro/nanostructures grafted on activated carbon cloth electrodes,” Desalination, vol. 344, pp. 236–242, 2014. [93] M.-X. Wang, Z.-H. Huang, L. Wang, M.-X. Wang, F. Kang, and H. Hou, “Electrospun ultrafine carbon fiber webs for electrochemical capacitive desalination,” New J. Chem., vol. 34, no. 9, pp. 1843–1845, 2010. [94] Y. Zhan, C. Nie, H. Li, L. Pan, and Z. Sun, “Enhancement of electrosorption capacity of activated carbon fibers by grafting with carbon nanofibers,” Electrochim. Acta, vol. 56, no. 9, pp. 3164–3169, 2011. [95] G. Wang, Q. Dong, Z. Ling, C. Pan, C. Yu, and J. Qiu, “Hierarchical activated carbon nanofiber webs with tuned structure fabricated by electrospinning for capacitive 117  deionization,” J. Mater. Chem., vol. 22, no. 41, p. 21819, 2012. [96] G. Wang, C. Pan, L. Wang, Q. Dong, C. Yu, Z. Zhao, and J. Qiu, “Activated carbon nanofiber webs made by electrospinning for capacitive deionization,” Electrochim. Acta, vol. 69, pp. 65–70, 2012. [97] A. G. El-Deen, N. A. M. Barakat, K. A. Khalil, M. Motlak, and H. Yong Kim, “Hollow carbon nanofibers as an effective electrode for brackish water desalination using the capacitive deionization process Ahmed,” New J. Chem., vol. 40, no. 38, pp. 198–205, 2014. [98] X. Z. Wang, M. G. Li, Y. W. Chen, R. M. Cheng, S. M. Huang, L. K. Pan, and Z. Sun, “Electrosorption of NaCl Solutions with Carbon Nanotubes and Nanofibers Composite Film Electrodes,” Electrochem. Solid-State Lett., vol. 9, no. 9, p. E23, 2006. [99] X. Z. Wang, M. G. Li, Y. W. Chen, R. M. Cheng, S. M. Huang, L. K. Pan, and Z. Sun, “Electrosorption of ions from aqueous solutions with carbon nanotubes and nanofibers composite film electrodes,” Appl. Phys. Lett., vol. 89, no. 5, 2006. [100] J. C. Farmer, D. V Fix, G. V Mack, R. W. Pekala, and J. F. Poco, “Capacitive deionization of NaCl and NaNO3 solutions with carbon aerogel electrodes,” J. Appl. Electrochem., vol. 143, no. 1, p. 159-, 1996. [101] M. E. Suss, T. F. Baumann, W. L. Bourcier, C. M. Spadaccini, K. A. Rose, J. G. Santiago, and M. Stadermann, “Capacitive desalination with flow-through electrodes,” Energy Environ. Sci., vol. 5, no. 11, p. 9511, 2012. [102] T.-Y. Ying, K.-L. Yang, S. Yiacoumi, and C. Tsouris, “Electrosorption of ions from aqueous solutions by nanostructured carbon aerogel.,” J. Colloid Interface Sci., vol. 250, no. 1, pp. 18–27, 2002. [103] J. Farmer, “Method and Apparatus for Capacitive Deionization, Electrochemical Purification, and Regeneration of Electrodes,” 5,425,858, 1995. [104] H. H. Jung, S. W. Hwang, S. H. Hyun, K. H. Lee, and G. T. Kim, “Capacitive deionization characteristics of nanostructured carbon aerogel electrodes synthesized via ambient drying,” Desalination, vol. 216, no. 1–3, pp. 377–385, 2007. [105] P. Xu, J. E. Drewes, D. Heil, and G. Wang, “Treatment of brackish produced water using carbon aerogel-based capacitive deionization technology,” Water Res., vol. 42, no. 10–11, pp. 2605–2617, 2008. [106] Z. Peng, D. Zhang, L. Shi, and T. Yan, “High performance ordered mesoporous carbon/carbon nanotube composite electrodes for capacitive deionization,” J. Mater. Chem., vol. 22, no. 14, p. 6603, 2012. [107] D. Zhang, X. Wen, L. Shi, T. Yan, and J. Zhang, “Enhanced capacitive deionization of graphene/mesoporous carbon composites,” Nanoscale, vol. 4, no. 17, p. 5440, 2012. [108] H. Wang, L. Shi, T. Yan, J. Zhang, Q. Zhong, and D. Zhang, “Design of graphene-coated hollow mesoporous carbon spheres as high performance electrodes for capacitive deionization,” J. Mater. Chem. A, vol. 2, no. 13, pp. 4739–4750, 2014. 118  [109] K. Dai, L. Shi, J. Fang, D. Zhang, and B. Yu, “NaCl adsorption in multi-walled carbon nanotubes,” Mater. Lett., vol. 59, no. 16, pp. 1989–1992, 2005. [110] S. Wang, D. Wang, L. Ji, Q. Gong, Y. Zhu, and J. Liang, “Equilibrium and kinetic studies on the removal of NaCl from aqueous solutions by electrosorption on carbon nanotube electrodes,” Sep. Purif. Technol., vol. 58, no. 1, pp. 12–16, 2007. [111] D. Zhang, T. Yan, L. Shi, Z. Peng, X. Wen, and J. Zhang, “Enhanced capacitive deionization performance of graphene/carbon nanotube composites,” J. Mater. Chem., vol. 22, no. 29, p. 14696, 2012. [112] H. Li, S. Liang, J. Li, and L. He, “The capacitive deionization behaviour of a carbon nanotube and reduced graphene oxide composite,” J. Mater. Chem. A, vol. 1, no. 21, p. 6335, 2013. [113] Y. Wimalasiri and L. Zou, “Carbon nanotube/graphene composite for enhanced capacitive deionization performance,” Carbon N. Y., vol. 59, pp. 464–471, 2013. [114] H. Li, T. Lu, L. Pan, Y. Zhang, and Z. Sun, “Electrosorption behavior of graphene in NaCl solutions,” J. Mater. Chem., vol. 19, no. 37, p. 6773, 2009. [115] H. Li, L. Zou, L. Pan, and Z. Sun, “Novel graphene-like electrodes for capacitive deionization,” Environ. Sci. Technol., vol. 44, no. 22, pp. 8692–8697, 2010. [116] B. Jia and L. Zou, “Graphene nanosheets reduced by a multi-step process as high-performance electrode material for capacitive deionisation,” Carbon N. Y., vol. 50, no. 6, pp. 2315–2321, 2012. [117] Z. Wang, B. Dou, L. Zheng, G. Zhang, Z. Liu, and Z. Hao, “Effective desalination by capacitive deionization with functional graphene nanocomposite as novel electrode material,” Desalination, vol. 299, pp. 96–102, 2012. [118] H. Wang, D. Zhang, T. Yan, X. Wen, L. Shi, and J. Zhang, “Graphene prepared via a novel pyridine–thermal strategy for capacitive deionization,” J. Mater. Chem., vol. 22, no. 45, p. 23745, 2012. [119] H. Wang, D. Zhang, T. Yan, X. Wen, J. Zhang, L. Shi, and Q. Zhong, “Three-dimensional macroporous graphene architectures as high performance electrodes for capacitive deionization,” J. Mater. Chem. A, vol. 1, no. 38, p. 11778, 2013. [120] X. Wen, D. Zhang, T. Yan, J. Zhang, and L. Shi, “Three-dimensional graphene-based hierarchically porous carbon composites prepared by a dual-template strategy for capacitive deionization,” J. Mater. Chem. A, vol. 1, no. 39, p. 12334, 2013. [121] H. Yin, S. Zhao, J. Wan, H. Tang, L. Chang, L. He, H. Zhao, Y. Gao, and Z. Tang, “Three-dimensional graphene/metal oxide nanoparticle hybrids for high-performance capacitive deionization of saline water,” Adv. Mater., vol. 25, no. 43, pp. 6270–6276, 2013. [122] I. Cohen, E. Avraham, A. Soffer, and D. Aurbach, “Water Desalination by Capacitive Deionization - Advantages Limitations and Modification,” ECS Trans., vol. 1, no. 17, pp. 43–59, 2013. 119  [123] R. Zhao, P. M. Biesheuvel, and  a. van der Wal, “Energy consumption and constant current operation in membrane capacitive deionization,” Energy Environ. Sci., vol. 5, no. 11, p. 9520, 2012. [124] P. M. Biesheuvel and A. van der Wal, “Membrane capacitive deionization,” J. Memb. Sci., vol. 346, no. 2, pp. 256–262, 2010. [125] J. Y. Lee, S. J. Seo, S. H. Yun, and S. H. Moon, “Preparation of ion exchanger layered electrodes for advanced membrane capacitive deionization (MCDI),” Water Res., vol. 45, no. 17, pp. 5375–5380, 2011. [126] S. Porada, B. B. Sales, H. V. M. Hamelers, and P. M. Biesheuvel, “Water desalination with wires,” J. Phys. Chem. Lett., vol. 3, no. 12, pp. 1613–1618, 2012. [127] A. Rommerskirchen, Y. Gendel, and M. Wessling, “Single module flow-electrode capacitive deionization for continuous water desalination,” Electrochem. commun., vol. 60, pp. 34–37, 2015. [128] S. I. Jeon, H. R. Park, J. G. Yeo, S. Yang, C. H. Cho, M. H. Han, and D. K. Kim, “Desalination via a new membrane capacitive deionization process utilizing flow-electrodes,” Energy Environ. Sci., vol. 6, no. 5, pp. 1471–1475, 2013. [129] T. Wu, G. Wang, F. Zhan, Q. Dong, Q. Ren, J. Wang, and J. Qiu, “Surface-treated carbon electrodes with modified potential of zero charge for capacitive deionization,” Water Res., vol. 93, pp. 30–37, 2016. [130] X. Gao, A. Omosebi, J. Landon, and K. Liu, “Surface charge enhanced carbon electrodes for stable and efficient capacitive deionization using inverted adsorption–desorption behavior,” Energy Environ. Sci., vol. 8, p. 897, 2015. [131] Y. Bian, X. Yang, P. Liang, Y. Jiang, C. Zhang, and X. Huang, “Enhanced desalination performance of membrane capacitive deionization cells by packing the flow chamber with granular activated carbon,” Water Res., vol. 85, pp. 371–376, 2015. [132] A. Omosebi, X. Gao, J. Rentschler, J. Landon, and K. Liu, “Continuous operation of membrane capacitive deionization cells assembled with dissimilar potential of zero charge electrode pairs,” J. Colloid Interface Sci., vol. 446, pp. 345–351, 2015. [133] D. Liu, K. Huang, L. Xie, and H. Tang, “Relation between operating parameters and desalination performance of capacitive deionization with activated carbon electrodes,” Environ. Sci. Water Res. Technol., vol. 1, pp. 516–522, 2015. [134] C. Yan, Y. W. Kanaththage, R. Short, C. T. Gibson, and L. Zou, “Graphene/Polyaniline nanocomposite as electrode material for membrane capacitive deionization,” Desalination, vol. 344, pp. 274–279, 2014. [135] J. J. Lado, R. E. Pérez-Roa, J. J. Wouters, M. I. Tejedor-Tejedor, C. Federspill, and M. A. Anderson, “Continuous cycling of an asymmetric capacitive deionization system: An evaluation of the electrode performance and stability,” J. Environ. Chem. Eng., vol. 3, no. 4, pp. 2358–2367, 2015. [136] F. Duan, Y. Li, H. Cao, Y. Wang, J. C. Crittenden, and Y. Zhang, “Activated carbon 120  electrodes: Electrochemical oxidation coupled with desalination for wastewater treatment,” Chemosphere, vol. 125, pp. 205–211, 2015. [137] Y. A. C. Jande and W. S. Kim, “Modeling the capacitive deionization batch mode operation for desalination,” J. Ind. Eng. Chem., vol. 20, no. 5, pp. 3356–3360, 2014. [138] O. N. Demirer, R. M. Naylor, C. A. Rios Perez, E. Wilkes, and C. Hidrovo, “Energetic performance optimization of a capacitive deionization system operating with transient cycles and brackish water,” Desalination, vol. 314, pp. 130–138, 2013. [139] Z. Chen, H. Zhang, C. Wu, Y. Wang, and W. Li, “A study of electrosorption selectivity of anions by activated carbon electrodes in capacitive deionization,” Desalination, vol. 369, pp. 46–50, 2015. [140] Y. Li, C. Zhang, Y. Jiang, T. J. Wang, and H. Wang, “Effects of the hydration ratio on the electrosorption selectivity of ions during capacitive deionization,” Desalination, vol. 399, pp. 171–177, 2016. [141] R. Zhao, M. van Soestbergen, H. H. M. Rijnaarts, A. van der Wal, M. Z. Bazant, and P. M. Biesheuvel, “Time-dependent ion selectivity in capacitive charging of porous electrodes,” J. Colloid Interface Sci., vol. 384, no. 1, pp. 38–44, 2012. [142] “Sea Temperature Water Temperatures,” World Sea Temperature, 2017. [Online]. Available: https://www.seatemperature.org/. [Accessed: 25-Sep-2017]. [143] G. Wang, B. Qian, Q. Dong, J. Yang, Z. Zhao, and J. Qiu, “Highly mesoporous activated carbon electrode for capacitive deionization,” Sep. Purif. Technol., vol. 103, pp. 216–221, 2013. [144] Y. Wimalasiri, M. Mossad, and L. Zou, “Thermodynamics and kinetics of adsorption of ammonium ions by graphene laminate electrodes in capacitive deionization,” Des, vol. 357, pp. 178–188, 2015. [145] S. Porada, M. Bryjak, A. van der Wal, and P. M. Biesheuvel, “Effect of electrode thickness variation on operation of capacitive deionization,” Electrochim. Acta, vol. 75, pp. 148–156, 2012. [146] Y. a C. Jande and W. S. Kim, “Desalination using capacitive deionization at constant current,” Desalination, vol. 329, pp. 29–34, 2013. [147] M. W. Saleem, Y. A. C. Jande, M. Asif, and W.-S. Kim, “Hybrid CV-CC operation of capacitive deionization in comparison with constant current and constant voltage,” Sep. Sci. Technol., vol. 6395, no. March, pp. 1–7, 2016. [148] T. Kim, J. E. Dykstra, S. Porada, A. van der Wal, J. Yoon, and P. M. Biesheuvel, “Enhanced charge efficiency and reduced energy use in capacitive deionization by increasing the discharge voltage,” J. Colloid Interface Sci., vol. 446, pp. 317–326, 2015. [149] J. Kang, T. Kim, H. Shin, J. Lee, J. I. Ha, and J. Yoon, “Direct energy recovery system for membrane capacitive deionization,” Desalination, vol. 398, pp. 144–150, 2016. [150] P. Długołecki and A. Van Der Wal, “Energy recovery in membrane capacitive 121  deionization,” Environ. Sci. Technol., vol. 47, no. 9, pp. 4904–4910, 2013. [151] Q. Yao and H. L. Tang, “Effect of Desorption Methods on Electrode Regeneration Performance of Capacitive Deionization,” vol. 143, no. 9, pp. 1–6, 2017. [152] E. García-Quismondo, C. Santos, J. Soria, J. Palma, and M. A. Anderson, “New Operational Modes to Increase Energy Efficiency in Capacitive Deionization Systems,” Environ. Sci. Technol., vol. 50, no. 11, pp. 6053–6060, 2016. [153] K. Y. Foo and B. H. Hameed, “Insight into the modeling of adsorption isotherm systems,” Chem. Eng. J., vol. 156, pp. 2–10, 2010. [154] H. Li, Y. Gao, L. Pan, Y. Zhang, Y. Chen, and Z. Sun, “Electrosorptive desalination by carbon nanotubes and nanofibres electrodes and ion-exchange membranes,” Water Res., vol. 42, no. 20, pp. 4923–4928, 2008. [155] L. Chang, Y. Zhou, and X. Duan, “Kinetics and equilibrium studies on the electrosorption of anions with activated carbon electrodes,” Desalin. Water Treat., vol. 52, no. 34–36, pp. 6549–6555, 2014. [156] W. Rudzinski and W. Plazinski, “Theoretical description of the kinetics of solute adsorption at heterogeneous solid / solution interfaces On the possibility of distinguishing between the diffusional and the surface reaction kinetics models,” J. Phys. Chem. C, vol. 111, pp. 15100–15110, 2007. [157] K. V. Kumar, “Linear and non-linear regression analysis for the sorption kinetics of methylene blue onto activated carbon,” J. Hazard. Mater., vol. 137, no. 3, pp. 1538–1544, 2006. [158] R. Niu, H. Li, Y. Ma, L. He, and J. Li, “An insight into the improved capacitive deionization performance of activated carbon treated by sulfuric acid,” Electrochim. Acta, vol. 176, pp. 755–762, 2015. [159] Y. A. C. Jande and W. S. Kim, “Predicting the lowest effluent concentration in capacitive deionization,” Sep. Purif. Technol., vol. 115, pp. 224–230, 2013. [160] R. Zhao, O. Satpradit, H. H. M. Rijnaarts, P. M. Biesheuvel, and A. van der Wal, “Optimization of salt adsorption rate in membrane capacitive deionization,” Water Res., vol. 47, no. 5, pp. 1941–1952, 2013. [161] J. E. Dykstra, R. Zhao, P. M. Biesheuvel, and A. Van der Wal, “Resistance identification and rational process design in Capacitive Deionization,” Water Res., vol. 88, pp. 358–370, 2016. [162] X. Gao, S. Porada, A. Omosebi, K.-L. Liu, P. M. Biesheuvel, and J. Landon, “Complementary surface charge for enhanced capacitive deionization,” Water Res., vol. 92, pp. 275–282, 2016. [163] J. E. Dykstra, K. J. Keesman, P. M. Biesheuvel, and A. van der Wal, “Theory of pH changes in water desalination by capacitive deionization,” Water Res., vol. 119, pp. 178–186, 2017. 122  [164] W. Zhang, M. Mossad, and L. Zou, “A study of the long-term operation of capacitive deionisation in inland brackish water desalination,” Desalination, vol. 320, pp. 80–85, 2013. [165] M. Haro, G. Rasines, C. MacIas, and C. O. Ania, “Stability of a carbon gel electrode when used for the electro-assisted removal of ions from brackish water,” Carbon N. Y., vol. 49, no. 12, pp. 3723–3730, 2011. [166] I. Cohen, E. Avraham, Y. Bouhadana, A. Soffer, and D. Aurbach, “Long term stability of capacitive de-ionization processes for water desalination: The challenge of positive electrodes corrosion,” Electrochim. Acta, vol. 106, pp. 91–100, 2013. [167] F. Duan, X. Du, Y. Li, H. Cao, and Y. Zhang, “Desalination stability of capacitive deionization using ordered mesoporous carbon: Effect of oxygen-containing surface groups and pore properties,” Desalination, vol. 376, no. 1, pp. 17–24, 2015. [168] D. He, C. E. Wong, W. Tang, P. Kovalsky, and T. D. Waite, “Faradaic Reactions in Water Desalination by Batch-Mode Capacitive Deionization,” Environ. Sci. Technol. Lett., p. acs.estlett.6b00124, 2016. [169] Y. Bouhadana, E. Avraham, M. Noked, M. Ben-Tzion, A. Soffer, and D. Aurbach, “Capacitive deionization of NaCl solutions at non-steady-state conditions: Inversion functionality of the carbon electrodes,” J. Phys. Chem. C, vol. 115, no. 33, pp. 16567–16573, 2011. [170] W. Tang, D. He, C. Zhang, P. Kovalsky, and T. D. Waite, “Comparison of Faradaic reactions in capacitive deionization (CDI) and membrane capacitive deionization (MCDI) water treatment processes,” Water Res., vol. 120, 2017. [171] B. Shapira, E. Avraham, and D. Aurbach, “Side Reactions in Capacitive Deionization (CDI) Processes: The Role of Oxygen Reduction,” Electrochim. Acta, vol. 220, pp. 285–295, 2016. [172] N. Holubowitch, A. Omosebi, X. Gao, J. Landon, and K. Liu, “Quasi-Steady-State Polarization Reveals the Interplay of Capacitive and Faradaic Processes in Capacitive Deionization,” ChemElectroChem, vol. 40506, pp. 1–11, 2017. [173] Y. Bouhadana, M. Ben-Tzion, A. Soffer, and D. Aurbach, “A control system for operating and investigating reactors: The demonstration of parasitic reactions in the water desalination by capacitive de-ionization,” Desalination, vol. 268, no. 1–3, pp. 253–261, 2011. [174] A. Bard and L. Faulkner, Electrochemical Methods: Fundamentals and Applications, 2nd ed. Wiley, 2001. [175] C. K. Alexander and M. N. O. Sadiku, “First-Order Circuits,” in Fundamentals of Electric Circuits, 5th ed., New York City, NY: McGraw-Hill, 2013, pp. 253–312. [176] A. Omosebi, X. Gao, and K. Liu, “Membrane Assisted Capacitive Deionization with Binder-Free Carbon Xerogel Electrodes,” ECS Trans., vol. 61, no. 21, pp. 9–17, 2014. [177] D. M. Drobny, S. A. Tychyna, Y. A. Maletin, N. G. Stryzhakova, and S. A. Zelinsky, 123  “Methods for Manufacturing Carbon Electrodes for Supercapacitors: Pros and Cons,” Proc. Int. Conf. Nanomater. Appl. Prop., vol. 22, no. 42, pp. 4–17, 2013. [178] “XL: Method Used by Goal Seek to Find a Solution,” Microsoft Support, 2017. [Online]. Available: https://support.microsoft.com/en-us/help/100782/xl-method-used-by-goal-seek-to-find-a-solution. [Accessed: 22-Dec-2017]. [179] Y.-J. Kim and J.-H. Choi, “Enhanced desalination efficiency in capacitive deionization with an ion-selective membrane,” Sep. Purif. Technol., vol. 71, no. 1, pp. 70–75, 2010. [180] K.-H. Park and D.-H. Kwak, “Electrosorption and electrochemical properties of activated-carbon sheet electrode for capacitive deionization,” J. Electroanal. Chem., vol. 732, pp. 66–73, 2014. [181] A. M. Dehkhoda, N. Ellis, and E. Gyenge, “Electrosorption on activated biochar: Effect of thermo-chemical activation treatment on the electric double layer capacitance,” J. Appl. Electrochem., vol. 44, pp. 141–157, 2014. [182] T. Chai and R. R. Draxler, “Root mean square error (RMSE) or mean absolute error (MAE)? -Arguments against avoiding RMSE in the literature,” Geosci. Model Dev., vol. 7, no. 3, pp. 1247–1250, 2014. [183] D. C. Stone and J. Ellis, “The Correlation Coefficient,” Statistics in Analytical Chemistry, 2011. [Online]. Available: http://www.chem.utoronto.ca/coursenotes/analsci/stats/CorrCoeff.html#section3-2-1. [Accessed: 02-Nov-2017]. [184] H. Wagner, “Influence of Temperature on Electrical Conductivity of Diluted Aqueous Solutions,” 2012.            124  Appendices  Appendix A. Graphite Foil Specifications  A.1. Pureechem Electrode Graphite Foil Specifications  Specification Value Thickness 200 µm Carbon content 99% Compressive strength 160 MPa Specific resistivity 4.5 µΩ m Bulk density 1.0 ± 0.05 g/cm3  A.2. Mineral Seal Corporation Flexible Graphite 2010A Specifications  Specification Value Thickness 250 µm Carbon content 99.5% Sulfur content < 300 ppm Chloride content (leachable) < 20 ppm Compressive strength 240 MPa Tensile strength 4.9 MPa Specific resistivity 0.9 µΩ m Bulk density 1.0 g/cm3 Working temperature range -240 to 510 ºC     125  Appendix B. Electrical Circuit for CDI Setup  B.1. Electrical Circuit for Zero Voltage Discharge   When all Relays are off, CDI system is shut off. When all Relays are on, CDI system is in charge stage. When Relays 3 and 4 are on, and Relays 1 and 2 are off, CDI system is in discharge stage.  B.2. Electrical Circuit for Reverse Polarity Discharge   When all Relays are off, CDI system is shut off. When all Relays are on, CDI system is in charge stage. When Relays 3 and 4 are on, and Relays 1 and 2 are off, CDI system is in discharge stage. 126  Appendix C. Statistical Formulas.  C.1. Correlation Coefficient (r2)  𝑟 =  𝑛 ∑ 𝑥𝑖𝑦𝑖𝑛𝑖=1 − ∑ 𝑥𝑖𝑛𝑖=1 ∑ 𝑦𝑖𝑛𝑖=1√(𝑛 ∑ 𝑥𝑖2𝑛𝑖=1 − (∑ 𝑥𝑖𝑛𝑖=1 )2)(𝑛 ∑ 𝑦𝑖2𝑛𝑖=1 − (∑ 𝑦𝑖𝑛𝑖=1 )2)  C.2. Root-Mean-Square Error (RMSE)  𝑅𝑀𝑆𝐸 =  √∑ (ŷ𝑖 − 𝑦𝑖)2𝑛𝑖=1𝑛 Here, i is the observations, n is the number of predictions, ŷi is the model predicted value of the dependant variable, yi is the observed value of the dependant variable, and xi is the value of the independent variable [182], [183].            127  Appendix D. Raw Data Processing and Modelling Calculations  D.1. pH Contribution to Conductivity Calculation of NaCl Concentration  Theoretically, the conductivity (κ) of a solution can be calculated by summing up the contribution from each ion. 𝜅 =  ∑ 𝐶𝑖𝑧𝑖𝜆𝑖 Here, Ci [mol cm-3] is the concentration of ion i, zi is the absolute value of the charge number of ion i, and λi [S cm2 mol-1] is the molar conductivity of ion i. The molar conductivity of relevant ions has been experimentally determined for dilute solutions at temperatures of 25 ⁰C and is shown in the table below [184]. Ion Na+ Cl- H+ OH- Molar conductivity [S cm2 mol-1] 50.9 75.5 350 198  The pH correction was performed by calculating the conductivity contribution from the H+ and OH- ions and subtracting the contribution from the measured conductivity. For example, at a pH of 5.0 and a measured conductivity (κ) of 2.2000 mS/cm, the calculation of the pH-corrected conductivity (κ*) is as follows.  [H+] = 10 ^ -pH = 10-5 mol/dm3   [OH-] = (10-14) / (10-5) = 10-9 mol/dm3  𝜅∗ =  𝜅 − (𝜆𝐻+𝐶𝐻+ + 𝜆𝑂𝐻−𝐶𝑂𝐻−) 𝜅∗ = 2.2000𝑚𝑆𝑐𝑚– ((350𝑆 𝑐𝑚2𝑚𝑜𝑙) (10−8𝑚𝑜𝑙𝑐𝑚3) + (198 𝑆 𝑐𝑚2𝑚𝑜𝑙) (10−12  𝑚𝑜𝑙𝑐𝑚3)) (1000𝑚𝑆𝑆)  𝜅∗ = 2.2000𝑚𝑆𝑐𝑚− 0.0035𝑚𝑆𝑐𝑚= 2.1965𝑚𝑆𝑐𝑚  128  D.2. NaCl Concentration to Conductivity Calibration Curve    D.3. Salt Adsorption Capacity, Charge Efficiency, and Specific Energy Consumption Calculation  Salt adsorption capacity is calculated from the following equation. 𝑆𝐴𝐶 =  𝑉(𝐶𝑖  −  𝐶𝑓) 𝑚 Here, V [L] is the volume, Ci [mg/L] is the initial concentration, Cf [mg/L] is the final concentration, and m is the electrode mass. For the desalination experiment with Pureechem electrodes at the base line condition of 1.2 V applied voltage, 1000 ppm NaCl concentration, 20 mL/min flow rate, and 60 min cycle time:   y = 2.010020xR² = 0.999792010002000300040005000600070000 500 1000 1500 2000 2500 3000 3500Conductivity (µS/cm)NaCl concentration (ppm)129   V = 0.250 L  Ci = 1045 mg/L Cf = 959 mg/L  m = 8.4 g  𝑆𝐴𝐶 =  (0.25 𝐿)(1045−959 𝑚𝑔/𝐿)8.4 𝑔 = 2.53 mg/g Charge efficiency is calculated for the charge cycle as follows:  𝛬 =  𝑧 𝐹 𝑉 (𝐶𝑖 − 𝐶𝑓) 𝑀 ∫ 𝐼 𝑑𝑡 𝑥 100% Here, z is the charge number, M [g/mol] is the molar mass, F [96,485 C mol-1] is the Faraday constant, I [A] is the current, and t [s] is time. The integral ∫ 𝐼 𝑑𝑡 was calculated from the data using the mid-point rule.  z = 1  M = 58.44 g/mol  V = 0.250 L  Ci = 1.045 g/L   Cf  = 0.959 g/L ∫ 𝐼 𝑑𝑡 = 88.2 C  𝛬 =  (96,485𝐶𝑚𝑜𝑙 )(0.25 𝐿)(1.045−0.959𝑔𝐿)(58.44𝑔𝑚𝑜𝑙)(88.2 𝐶) 𝑥 100% =  40% Specific energy consumption is calculated from the equation below. 𝜂 =  𝐸𝑐𝑒𝑙𝑙  ∫ 𝐼𝑑𝑡𝑉 (𝐶𝑖  −  𝐶𝑓) Here, Ecell [V] is the cell potential and the other variables are defined previously.  Ecell = 1.2 V   ∫ 𝐼 𝑑𝑡 = 88.2 C  V = 0.250 L  Ci = 1.045 g/L  Cf  = 0.959 g/L 𝜂 =  (1.2 𝑉)(88.2 𝐶)(0.25 𝐿)(1.045 − 0.959𝑔𝐿)= 4,922.79𝐽𝑔= 0.00137 𝑘𝑊ℎ/𝑔 130  D.4. Adsorption Isotherm Linear Plots and Calculation of Parameters  Langmuir       y-int = 0.3598 = 1𝑞𝑚   ->   qm = 1 / 0.3598 = 2.7790 slope = 39.179 = 1𝑞𝑚𝐾𝐿  ->   KL = 1 / (2.7790 * 39.179) = 0.0091846 Freundlich   slope = 0.1052 = 1/n y-int = 0.1615 = ln KF  ->   KF = e ^ 0.1615 = 1.1752 y = 39.179x + 0.3598R² = 0.959800.10.20.30.40.50 0.0005 0.001 0.0015 0.002 0.00251/q_e1/Cy = 0.1052x + 0.1615R² = 0.995400.10.20.30.40.50 0.5 1 1.5 2 2.5 3 3.5 41/q_eln C131  Appendix E. SEM Images and EDS Spectra of Regenerated Electrodes  E.1. Deionized Water Washed Electrode  E.2. Citric Acid Washed Electrode    132  Appendix F. Electrode Potential vs. Hg/HgO Reference Electrode   -1.2-1-0.8-0.6-0.4-0.200.20.40.60 5000 10000 15000 20000 25000Electrode Potential vs. Hg/HgO (V)Time (s)vs. cathodevs. anode

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.24.1-0365275/manifest

Comment

Related Items