UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Innovations in natural organic matter removal using ion exchange Dezfoolian, Maryam 2019

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

Item Metadata

Download

Media
24-ubc_2019_september_dezfoolian_maryam.pdf [ 21.2MB ]
Metadata
JSON: 24-1.0379511.json
JSON-LD: 24-1.0379511-ld.json
RDF/XML (Pretty): 24-1.0379511-rdf.xml
RDF/JSON: 24-1.0379511-rdf.json
Turtle: 24-1.0379511-turtle.txt
N-Triples: 24-1.0379511-rdf-ntriples.txt
Original Record: 24-1.0379511-source.json
Full Text
24-1.0379511-fulltext.txt
Citation
24-1.0379511.ris

Full Text

   INNOVATIONS IN NATURAL ORGANIC MATTER REMOVAL USING ION EXCHANGE   by Maryam Dezfoolian MASc., The University of Tehran, 2014  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)  June 2019  © Maryam Dezfoolian, 2019    ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, a thesis/dissertation entitled:  Innovations in natural organic matter removal using ion exchange  submitted by Maryam Dezfoolian in partial fulfillment of the requirements for the degree of Master of Applied Science in Chemical and Biological Engineering  Examining Committee: Dr. Madjid Mohseni, Department of Chemical and Biological Engineering, The University of British Columbia Supervisor  Dr. Benoit Barbeau, Department of Civil, Geological& Mining Engineering, The University of Polytechnique de Montréal   Supervisory Committee Member  Dr. Pierre Bérubé, Department of Civil Engineering, The University of British Columbia Supervisory Committee Member      iii  Abstract  Ion exchange (IEX) is a viable technology for the removal of natural organic matter (NOM) from surface water. One potential drawback of the IEX process, however, is the need for frequent regeneration of the resins using a brine solution which needs to be disposed of safely. Operating ion exchange systems in biological could potentially be a promising approach. By allowing a viable microbial community to grow on the resins, NOM removal could be achieved through biodegradation in addition to ion exchange. This process, here named biological ion exchange (BIEX), decreases the need for frequent regeneration and prolongs the life of the resins.  The aim of this study was to further explore the efficacy of the BIEX process, operating in the long-term, at removing NOM and to investigate different mechanisms involved in the process. Two parallel and identical packed bed ion exchange columns were used in pilot scale, each column with a height of 1 m was halfway filled with ion exchange resin and operated with a filtration velocity of 1 m/h, without regeneration. Also, two parallel continuous stirred tank reactors, each with a volume of 400 mL, were operated in the lab scale, in biotic and abiotic conditions. The performance of the systems was assessed by monitoring the concentrations of dissolved organic carbon, different anions, and UV absorbance of water. In addition, several other parameters (i.e., THMFP, BDOC and ATP measurements, resin morphology) were monitored on a less frequent basis.  Laboratory experiments involving biotic and abiotic conditions resulted in no significant difference in terms of performance and NOM removal. This was likely due to the type of reactor (i.e., stirred tank), as well as operating conditions that did not allow for sufficient growth of   iv  biofilm. Pilot experiment, on the other hand, had noticeable biological growth and demonstrated effective removal of NOM, after approximately 9 months. Extended NOM removal along with no chloride release after 60 days, positive ATP data through the bed, and significant sulphate release over the course of experiments, indicated that in addition to ion exchange, other mechanisms would be responsible for NOM removal.      v  Lay Summary  Presence of NOM in raw drinking water supplies brings many challenges in drinking water quality as well as the water treatment systems. Therefore, a number of processes are available for the removal of NOM from drinking water supplies. Among those processes, ion exchange (IEX) is a competitive treatment process for removing NOM. Although this process is known as an efficient method to remove NOM, it comes with a challenge related to the disposal of spent brine used for resin regeneration. Biological ion exchange is an exciting alternative to IEX processes. It is introduced as a new concept to address issues common in IEX applications in small systems, i.e., brine disposal. The main objective of this work was to investigate the efficiency of a BIEX system at removing NOM under long-term operation and in the field, with surface water, used as the source of drinking water for a small community.     vi  Preface  The work presented in this thesis has been completed by the author, Maryam Dezfoolian, under the supervision of Professor Madjid Mohseni (Chemical and Biological Engineering) at the University of British Columbia. The contribution of the author to the work involved reviewing the literature, designing and carrying out the lab or pilot scale experiment, performing various analytical tests and data collection, and analyzing and presenting the data. Dr. Mohseni contributed to the development of experimental plan and research objectives and provided guidance over the course of the study.     vii  Table of Contents  Abstract .....................................................................................................................................iii Lay Summary ............................................................................................................................. v Preface ....................................................................................................................................... vi Table of Contents ...................................................................................................................... vii List of Tables ............................................................................................................................ xii List of Figures .......................................................................................................................... xiv List of Symbols and Abbreviations ........................................................................................ xviii Acknowledgements .................................................................................................................. xxi Dedication .............................................................................................................................. xxii Chapter 1: Introduction ........................................................................................................... 1 Chapter 2: Background and Literature Review ........................................................................ 3 2.1 Natural organic matter (NOM) ................................................................................. 3 2.2 Challenges associated with natural organic matter (NOM) ....................................... 4 2.3 Removal of natural organic matter (NOM) ............................................................... 5 2.4 Ion exchange (IEX).................................................................................................. 6 2.5 Removal of natural organic matter by ion exchange ................................................. 7  Effect of resin properties on NOM removal .................................................... 10  Effect of NOM characteristics and water matrix on NOM removal ................. 11 2.6 Challenges associated with ion exchange method in small systems ........................ 16 2.7 Biological ion exchange (BIEX) ............................................................................ 17   viii  2.8 Knowledge gap ...................................................................................................... 22 Chapter 3: Research Approach and Objectives ...................................................................... 23 3.1 Research approach ................................................................................................. 23 3.2 Research objectives and hypotheses ....................................................................... 23 Chapter 4: Experimental Methodology and Procedures ......................................................... 25 4.1 Reagents and chemicals ......................................................................................... 25  Anionic ion exchange resin ............................................................................ 25  Reagents ........................................................................................................ 26 4.2 Glassware .............................................................................................................. 26 4.3 Analytical techniques and procedures .................................................................... 26  Physical water characteristics (Turbidity and pH) ........................................... 26  Dissolved organic carbon (DOC) ................................................................... 27  Biodegradable dissolved organic carbon (BDOC) .......................................... 28  UV absorbance at 254 nm (UVA254) and specific UV absorbance (SUVA) .... 28  High performance size exclusion chromatography (HPSEC) .......................... 29  Disinfection by products potential tests (THM-UFC) ..................................... 29  Chlorine demand ............................................................................................ 30  Inorganic composition of water ...................................................................... 30 4.3.8.1 Anions concentration.................................................................................. 30 4.3.8.2 Alkalinity ................................................................................................... 31  Ion exchange mechanism analysis .................................................................. 31 4.3.9.1 Chloride release ......................................................................................... 31 4.3.9.2 Mass balance .............................................................................................. 31   ix   Biological activity analysis............................................................................. 32 4.3.10.1 ATP measurements ................................................................................... 32  Morphology investigation .............................................................................. 33  Statistical analysis .......................................................................................... 33 4.4 Experimental process of pilot study ....................................................................... 34  Location, historical data and source water characteristics ............................... 34  Experimental set up ........................................................................................ 38  Operating conditions ...................................................................................... 41  Backwash ....................................................................................................... 41  Resin regeneration ......................................................................................... 42 4.5 Lab scale studies .................................................................................................... 44  Source water characteristics ........................................................................... 44  Experimental set up ........................................................................................ 46  Operating conditions ...................................................................................... 48 Chapter 5: Pilot Study Results ............................................................................................... 49 5.1 Natural organic matter (NOM) removal ................................................................. 49  Dissolved organic carbon (DOC) ................................................................... 49  UV absorbance at 254 nm (UVA254) and specific UV absorbance (SUVA) .... 51  Chloride release ............................................................................................. 52  Carbon mass balance ...................................................................................... 54 5.2 Biological ion exchange (BIEX) ............................................................................ 55  Biological activity analysis............................................................................. 55 5.3 Impact of natural water characteristics ................................................................... 58   x   Inorganic composition of water ...................................................................... 58  Alkalinity ....................................................................................................... 60 5.4 Possible mechanisms happening over the whole process ........................................ 61 5.5 Reduction in the formation of disinfection byproducts (DBP) ................................ 65 5.6 Resin regeneration ................................................................................................. 66  Impact of regeneration duration and reagents ................................................. 66  Explore the regenerated resin performance under long-term operation ........... 68  Impact of long-term usage and regeneration on resin morphology .................. 69 5.7 Resin surface analysis- Field emission scanning electron microscopy .................... 70 Chapter 6: Laboratory Study Results ..................................................................................... 73 6.1 Ion exchange performance without regeneration (Biotic/Abiotic) ........................... 74  Priest Lake-CSTR-30 min RT-Up to 4,000 BV .............................................. 74  Jericho Pond-CSTR-40 min RT-Up to 20,000 BV .......................................... 76  A comparison between ion exchange performance in pilot and lab scale studies (PBR and CSTR) ........................................................................................................... 79 6.2 Estimation of ion exchange resin exhaustion (Biotic condition) ............................. 81  CSTR experiments (Priest Lake- 30 min RT- up to 4,000 BV) ....................... 81  CSTR experiments (Jericho Pond- 40 min RT- up to 20,000 BV) ................... 83  A comparison of ion exchange resin exhaustion in pilot and lab scale studies (PBR and CSTR) ........................................................................................................... 84 Chapter 7: Conclusions and Recommendations ..................................................................... 86 7.1 Recommendations for future work ......................................................................... 87 7.2 Significance and outcome of the research............................................................... 87   xi  References ................................................................................................................................ 89 Appendices ............................................................................................................................... 95 Appendix A ....................................................................................................................... 95 Appendix B ..................................................................................................................... 101     xii  List of Tables  Table 2-1. Summary of studies investigated parameters influencing NOM removal over time. .. 13 Table 2-2. Results obtained regarding NOM removal by ion exchange process in a biological mode. ........................................................................................................................................ 21 Table 4-1. Typical physical & chemical characteristics of Purolite A860................................... 25 Table 4-2. Equipment used to measure turbidity and pH. ........................................................... 27 Table 4-3. Water source quality after filtration process- Middle River - July 2017 to April 2018. ................................................................................................................................................. 38 Table 4-4. Set up characteristics for this study and the previous works. ..................................... 39 Table 4-5. Operating conditions for this study and previous studies. .......................................... 41 Table 4-6. Different regeneration scenarios. .............................................................................. 43 Table 4-7. Water source quality: Priest lake (November 2015) and Jericho Pond (March-May 2017). ....................................................................................................................................... 46 Table 4-8. Operation conditions of experiments performed in CSTR. ........................................ 48 Table 5-1. Mass balance results of carbon in BIEX columns. .................................................... 54 Table 5-2. A comparison between two pilot studies with two different water sources. ............... 55 Table 5-3. Total sulphate removed-released over the course of experiment................................ 59 Table 5-4. Equivalents of different adsorbed compounds over 277 days of operation and the resin bed. ........................................................................................................................................... 59 Table 5-5. Chlorine demand values for raw and treated water at two different BV. .................... 66 Table 5-6. Different regeneration scenarios. .............................................................................. 67   xiii  Table 5-7. Biomass concentration in different collected resins................................................... 67 Table 5-8. DOC Removals by fresh resin and resins regenerated with different regeneration approaches. ............................................................................................................................... 68 Table 5-9. Comparison between NOM removal at 1000 BV for regenerated resin and fresh resin. ................................................................................................................................................. 69 Table 5-10. FESEM images of the fresh resin, used resin for 277 days, and regenerated resin. .. 71 Table 5-11. FESEM images of the surface of the BIEX resin (used for 277) before and after regeneration with NaCl 10 Wt.% for 2 hours, at different magnifications. ................................. 72 Table 6-1. Some of the operation conditions for two sets of experiments performed in CSTR. .. 74 Table A-1. Turbidity values in inlet and outlets of 7.5 and 30 min EBCT. ................................. 98 Table B-2. Selected parameters and their levels used for designing of experiments. ................ 101 Table B-3. Design of experiments table including conditions of 15 runs. ................................. 102     xiv  List of Figures  Figure 2-1. Hierarchy of organic matter classification in raw water sources, as commonly used in drinking water treatment. ............................................................................................................ 4 Figure 2-2. Conceptual model of an anionic ion exchange resin bead. ......................................... 9 Figure 2-3. Weekly dissolved organic carbon (DOC) monitoring in the source water (SW) and GAC, BAC, BIEX and IEX effluents during a pilot study from Polythechnique Montréal (Amini et al., 2018). .............................................................................................................................. 20 Figure 4-1. Location of the Middle River village in north-central British Columbia, site of the pilot study. ................................................................................................................................ 34 Figure 4-2. Monitoring the presence of total coliform in raw, treated and tap water. .................. 36 Figure 4-3. Monitoring the presence of E coli in raw, treated and tap water. .............................. 36 Figure 4-4. Middle River water treatment plant building before installation of experimental set up (July 2017). .......................................................................................................................... 37 Figure 4-5. Schematic of the experimental set up. ..................................................................... 40 Figure 4-6. Experimental set up in Middle River water treatment plant (2017). ......................... 40 Figure 4-7. Location of priest lake on Texada Island, British Columbia, Canada. ...................... 45 Figure 4-8. Location of Jericho Pond in Vancouver, British Columbia, Canada. ........................ 45 Figure 4-9. Schematic of the benchtop continuous stirred tank reactor....................................... 47 Figure 4-10. Experimental set up (CSTRs). ............................................................................... 47 Figure 5-1. Dissolved organic carbon (DOC) concentration vs. days of operation. .................... 50 Figure 5-2. Dissolved organic carbon (DOC) removal vs. days of operation. ............................. 50   xv  Figure 5-3. UV absorbance at 254 nm vs. days of operation. ..................................................... 51 Figure 5-4. Specific UV absorbance (SUVA) reduction vs. days of operation. .......................... 52 Figure 5-5. Chloride concentration in inlet and outlet vs. days of operation. .............................. 53 Figure 5-6. Sampling point of resin collection for ATP tests...................................................... 56 Figure 5-7. Biomass (ATP) profiles through depth of the filter bed (a) After 95 days of operation (4560 BV), (b) after 276 days of operation (13276 BV)............................................................. 57 Figure 5-8. Inlet and outlet sulphate concentration over the days of operation. .......................... 58 Figure 5-9. Cumulative capacity of the resin based on the adsorbed DOC and sulphate. ............ 60 Figure 5-10. Alkalinity removal or release. ............................................................................... 60 Figure 5-11. Acidity (pH) in source water and treated water. ..................................................... 61 Figure 5-12. Accumulation of individual and total NOM and sulphate capacity over time (in calculating sulphate capacity the released sulphate values were subtracted). .............................. 62 Figure 5-13. Hypothetic NOM concentration profile over bio degradation and bioregeneration. 65 Figure 5-14. Removal of THM precursors at different EBCT after 144 and 277 days of operation. ................................................................................................................................................. 66 Figure 5-15. Resin morphology and size order distribution for fresh resin, BIEX resin used for 277 days and regenerated resin after 277 days of operation. ...................................................... 70 Figure 6-1. DOC concentration vs. bed volume in biotic and abiotic conditions. ....................... 75 Figure 6-2. Sulphate concentration vs. bed volumes in biotic and abiotic conditions. ................ 76 Figure 6-3. DOC removal vs. bed volumes for both biotic and abiotic systems. ......................... 77 Figure 6-4. Treated water sulphate concentration at two different biotic and abiotic conditions vs. bed volume. .............................................................................................................................. 78   xvi  Figure 6-5. Accumulative removed NOM (g) over the period of each experiment per 1 mL resin. ................................................................................................................................................. 80 Figure 6-6. Cumulative removed sulphate (mole) over the period of each experiment per 1 L resin. ......................................................................................................................................... 81 Figure 6-7. Accumulative equivalents of removed DOC and removed-released sulphate vs. bed volume in biotic systems (Priest lake-CSTR). ........................................................................... 83 Figure 6-8. Accumulative DOC and sulphate equivalents over the course of experiment in biotic condition (Jericho Pond-CSTR). ............................................................................................... 84 Figure 6-9. Resin capacity (%) occupied by NOM and sulphate at different experiments over time........................................................................................................................................... 85 Figure 7-1. Middle River water treatment pilot plant installed on September 2018. ................... 88 Figure A-1. BDOC concentration in water treated at different EBCTs of 0, 7.5 and 30 min after 144 and 277 days of operation. .................................................................................................. 95 Figure A-2. Size exclusion chromatography for raw and treated water at two different EBCTs of 7.5 and 30 minutes after 277 days of operation. ......................................................................... 96 Figure A-3. Changes in temperature and NOM removal over the course of experiments vs. date. ................................................................................................................................................. 97 Figure A-4. Turbidity values at different EBCTs. (values= result of samples taken over 277 days of operation). ............................................................................................................................ 98 Figure A-5.Turbidity values over the backwash procedure. ....................................................... 99 Figure A-6. HPSEC chromatograms of raw and treated water in biotic and abiotic conditions at different bed volumes. ............................................................................................................. 100 Figure B-7. Inlet flowrate impact on NOM removal efficiency in CSTR. ................................ 103   xvii  Figure B-8. Resin dose impact on NOM removal efficiency in CSTR. .................................... 103 Figure B-9. Mixer speed impact on NOM removal efficiency in CSTR. .................................. 104 Figure B-10. Flowrate and mixer speed impacts on NOM removal. ......................................... 104 Figure B-11. Resin dose and mixer speed impacts on NOM removal....................................... 105 Figure B-12. Resin dose and flowrate impacts on NOM removal. ........................................... 105     xviii  List of Symbols and Abbreviations  AC  Activated carbon  AEC  Anion exchange capacity  AMW Apparent molecular weight AOP  Advanced oxidation process  ATP  Adenosine triphosphate  BAC  Biological activated carbon  BDOC  Biodegradable dissolved organic carbon  BIEX  Biological ion exchange  BOM  Biodegradable organic matter  BV  Bed volume  CSTR Continuous stirred tank reactor DOC  Dissolved organic carbon  DWTP  Drinking water treatment plant  EBCT  Empty bed contact time  FA  Fulvic acid  FESEM Field emission scanning electron microscopy GAC  Granular activated carbon  GCDWQ Guideline for Canadian drinking water quality HA  Humic acid  HMW  High-molecular-weight    xix  HPLC  High-performance liquid chromatography  HPSEC High performance size exclusion chromatography IC  Ion chromatography  IEX Ion exchange  LMW  Low-molecular-weight  MF  Microfiltration  MIEX  Magnetic ion exchange  MP Macro porous NF  Nanofiltration  NOM  Natural organic matter  NTU  Nephelometric turbidity unit  PAC Powdered activated carbon PBR Packed bed reactor RLU Relative light unit RT Residence time SBA Strong base anion exchange resin SR Suwannee river SSF  Slow sand filtration  SUVA  Specific ultra-violet absorbance  THM  Trihalomethane  TOC  Total organic carbon  UF  Ultrafiltration    xx  UFC  Uniform formation condition  UVA  Ultra-violet absorbance  UVA254 Ultra-violet absorbance at 254 nm UVT Ultra-violet transmittance WBA Weak base anionic ion exchange resin      xxi  Acknowledgements  I would like to thank my supervisor, Dr. Madjid Mohseni for his incredible mentorship, unwavering support, and patience during the course of my graduate studies. His continuous encouragement and motivation afforded me a wonderful graduate experience that was both incredibly productive and stimulating. I would also like to thank my committee members, Dr. Benoit Barbeau and Dr. Pierre Bérubé for their time and valuable insight throughout my research. I am also grateful to my wonderful lab members- past and present- for their feedback, assistance, and moral support.  I would also like to express my deepest gratitude to my family, particularly my parents; Monir and Ali, as well as my sisters and brother; Bahareh, Shabnam, and Mehdi for the unconditional love and utmost understanding as I completed this part of my journey in life.  Finally, I would like to thank my loving husband, Mohamad, who has stood by my side throughout all the trials, tribulations, and optimizations and who cheered me on through all of my accomplishments no matter how big or small.     xxii  Dedication       This work is dedicated to kids all around the world  who have not tried the real taste of the water.     1 Chapter 1: Introduction Natural organic matter (NOM) in raw drinking water supplies inflicts many challenges to drinking water quality as well as the water treatment systems. Therefore, a number of chemical and physical processes are available for the removal of NOM from drinking water supplies. Among those processes, ion exchange (IEX) is a competitive treatment process for removing NOM. It is particularly suitable for small and remote water systems due to the ease of operation and effectiveness. One potential drawback of the IEX process, however, is the need for regeneration of the resins using a brine solution which needs to be disposed of safely. This puts serious environmental constraints on small systems that are not equipped with proper resources for handling and management of the spent brine. While ion exchange is the primary mechanism behind NOM removal, under certain operating conditions other mechanisms like biodegradation, resulting from microbial activity, could be involved and may lead to additional degradation of NOM. Therefore, operating ion exchange systems in biological mode (no or less resin regeneration) could potentially be a promising approach as it decreases the need for frequent regeneration and prolongs the life of the resin. This process is called “biological ion exchange” (BIEX) which is a potential alternative to the conventional IEX processes. Earlier lab scale research at UBC and a pilot scale research in Quebec confirmed that efficient NOM removal can be achieved in a BIEX system (Amini et al., 2018; Winter et al., 2018). Even though this process is simple in concept, there are many reactions and   2 exchanges happening in parallel. Therefore, more research is required to assess the BIEX system performance and investigate mechanisms involved in the process. A pilot study as well as the lab experiments have been performed to further explore the efficiency of BIEX process at removing NOM. Two parallel and identical packed bed ion exchange columns were used in pilot scale and two parallel continuous stirred tank reactors were operated in the lab scale. This thesis compiles the results obtained over the course of this research and discusses them in two chapters including the pilot study and laboratory study results. The information gathered is presented as follows: Chapter 1: Provides a very high level overview of the research. Chapter 2: Presents a general background on the research topic and comprehensive literature review on NOM removal by ion exchange and concerns regarding the using brine. Review recent researches on BIEX and presents knowledge gap. Chapter 3: Presents research approach and objectives of the work. Chapter 4: Provides detailed description of experimental methodology and procedure. Chapter 5: Presents the results achieved throughout the pilot study and discuss them. Chapter 6: Presents the results achieved in laboratory experiments. Discuss and compare them with PBR performance in pilot study. Chapter 7: Highlights overall conclusions and provides recommendation for the future research. Appendix A: Provides supplementary results from the pilot or laboratory experiments. Appendix B: Presents the results of a design of experiments to optimize the effective parameters in NOM removal by ion exchange operation (without regeneration) using a CSTR.     3 Chapter 2: Background and Literature Review  2.1 Natural organic matter (NOM) The term natural organic matter (NOM) is used to describe a complex mixture of organic substances originating from natural sources, i.e., plants, animals, etc. (Stevenson, 1994; Thurman, 1985). Aquatic NOM is derived from the breakdown of plants as well as the byproducts of the biological activity of bacteria, algae, and aquatic plants or animals (Sillanpää, 2015). It can also be introduced into the water body from land (Crittenden et al., 2012a). NOM describes a broad category of compounds with varying chemical properties, whose composition may change significantly from one water body to another (Crittenden et al., 2012a). While predominantly composed of hydrophobic fractions (Sillanpää, 2015; Thurman, 1985), NOM is made up of hydrophobic, hydrophilic and neutral constituents (Leenheer and Croue, 2003). Though the majority of NOM fractions (around 90%) have molecular weights between about 500 and 3000 Da, many NOM compounds are found in a wide range of molecular weights beyond this range (Crittenden et al., 2012a). Characterization of NOM is a critical step in designing its removal during the design process for many water treatment processes. Using various methods, NOM can be separated into different fractions based on properties such as molecular weight, polarity or volatility. In drinking water treatment, NOM is often quantified by measuring total organic carbon (TOC), a surrogate that gives an indication of the NOM concentration (Crittenden et al., 2012a; Thurman, 1985). TOC concentrations for ground and surface waters are often in the ranges of 0.1 to 2 and 1 to 20 mg/L, respectively (Crittenden et al., 2012a). In addition to TOC, dissolved organic carbon (DOC),   4 biodegradable dissolved organic carbon (BDOC) and assimilable organic carbon (AOC) can aid in raw water characterization of NOM (Crittenden et al., 2012a; Matilainen et al., 2011). Figure 2-1 shows the relation between these parameters. Furthermore, UV254 absorbance and specific UV absorbance (SUVA) are often used to represent the presence and nature of NOM. SUVA represents the fraction of aromatic compounds of NOM, and is obtained through normalizing UV absorbance at 254nm over the DOC in a water sample (Weishaar et al., 2003) (see Eq. (1)). SUVA = UVA&'(DOC × 100	 (1)  Figure 2-1. Hierarchy of organic matter classification in raw water sources, as commonly used in drinking water treatment.  2.2 Challenges associated with natural organic matter (NOM) The presence of NOM in raw drinking water supplies brings many challenges, including the impact of NOM on the aesthetic quality of water such as taste or colour (Thurman, 1985). NOM is also a potential food source for bacteria and can lead to bacterial regrowth and biofilm formation. This may lead to biological instability of drinking water in water distribution systems (van der Kooij, 1992). However, the most important concern regarding NOM in drinking water is the risk of Organic Matter (OM)Total Organic Carbon (TOC)Dissolved Organic Carbon (DOC)Biodegradable Dissolved Organic Carbon (BDOC)Readily assimilable organic carbon (AOC)----  5 formation of disinfection by-products (DBPs) through the reaction of NOM with chlorine during the disinfection process (Bolto et al., 2002a; Chen and Westerhoff, 2010; Crittenden et al., 2012a). Trihalomethanes (THMs) and Halo-acetic Acids (HAAs) are some of common DBP compounds which have been subjected to stringent regulations by environmental and health authorities all around the world due to presenting adverse health effects (Singer, 2006). The limitations for THMs and HAAs concentration according to the Health Canada guideline are <100 µg/L and <80 µg/L respectively. The presence of NOM also affects the efficiency of many water treatment processes. NOM reacts with and consumes chemical disinfectants such as in the aforementioned reaction with chlorine, as well as consuming coagulants such (Crittenden et al., 2012a; Edzwald, 1993). Consequently, to achieve efficient treatment the amount of these reagents is increased beyond what would be required in the absence of NOM. NOM also causes clogging and fouling of membranes during membrane filtration, which leads to a rapid decline in membrane performance and decreased lifespan (Crittenden et al., 2012a; Kennedy et al., 2008). During UV-based disinfection treatments, NOM absorbs ultraviolet (UV) radiation, reducing the radiation dose delivered for microorganisms inactivation. In activated carbon (AC) filters, it adsorbs rapidly on the media and leads to a drop in adsorption capacity (Gibert et al., 2013; Sillanpää, 2015). The issues discussed above are important reasons for removing NOM from drinking water supplies before disinfection and other treatment processes. 2.3 Removal of natural organic matter (NOM) A number of chemical and physical processes are available for the removal of NOM from drinking water supplies. These include coagulation and flocculation (Matilainen et al., 2010), membrane   6 filtration (Sillanpää, 2015), oxidation (i.e., Ozonation) (Matilainen and Sillanpää, 2010), activated carbon (Iriarte-velasco et al., 2008) and ion exchange (Levchuk et al., 2018). However, most methods are either complex or costly, which reduces their applicability for small treatment systems. For instance, coagulation/flocculation followed by filtration is a common and economically feasible process for NOM removal in urban/large water treatment systems (Matilainen and Sillanpää, 2010); however, this approach suffers from excessive need of operation, maintenance and chemical (reagent) usage. This is especially undesirable for small and remote treatment systems. An optimal water treatment for small systems in remote areas should ideally be robust and cost-effective, simple and easy to operate, have low reagent or consumable requirements, and produce minimal sludge or waste (Amini et al., 2018). Therefore, a simpler, more efficient, affordable and robust process is required to meet drinking water treatment needs of small systems. Among the mentioned options, the anionic ion exchange (IEX) process has been previously demonstrated to be efficient at NOM removal and has received considerable attention (Bolto et al., 2002b; Levchuk et al., 2018). Not only is it recommended as a pre-treatment step in water treatment processes, but also it is a promising approach for small drinking water treatment systems and remote communities due to its ease of operation and affordability (Amini et al., 2018; Levchuk et al., 2018). Therefore, this method could serve as a viable option for drinking water treatment in small and remote communities. 2.4 Ion exchange (IEX) Ion exchange involves a reversible exchange of ions between a solid and a liquid phase. The solid phase carries exchangeable ions and the liquid phase contains ions needing to be removed   7 (C.E.Harland, 1994; Wachinski, 1997). In drinking water treatment applications, ion exchange is primarily used for water softening (Crittenden et al., 2012b). IEX has also been shown effective for the removal of inorganic anions such as nitrate and sulphate from water in drinking water treatment (Boyer and Singer, 2006; Levchuk et al., 2018). The solid phase in IEX is a resin, composed of a crosslinked polymer matrix containing covalently bonded functional groups with fixed ionic charges. Resins are normally fabricated as small beads, are insoluble in the liquid phase, and will release their counter ions to adsorb equimolar amounts of ions, commonly the contaminants of interests from the liquid (Crittenden et al., 2012b). Synthetic resins are used widely for ion exchange applications due to their relatively large available exchange capacities and ease of regeneration (Crittenden et al., 2012a). IEX resins vary in functionality and capacity and are characterized by their polymeric matrix, functional groups, ionic form (counter ion), ion exchange capacity, water content and bead size.  Ion exchange resins can be divided in two main groups of cation (acidic) and anion (basic) exchangers. Each group can further be classified as weak or strong, differing by their functional groups. Resins can also vary by porosity (gel or macro porous) or polymeric structure (styrene or acrylic) (C.E.Harland, 1994; Crittenden et al., 2012b). Since the majority of NOM fractions are negatively charged in water, anion exchange resins are used for NOM removal (Crittenden et al., 2012b; Dorfner, 1972). 2.5 Removal of natural organic matter by ion exchange Figure 2-2 shows a schematic of an anionic ion exchange resin immersed in water containing negatively-charged NOM. The figure depicts the spherical resin shape containing the crosslinked polymers and positively charged functional groups. Chloride ions (Cl-) play the role of counter   8 ions and are free to move in the pores of the matrix. When the resin is immersed in water, there is a tendency for the Cl- ions to diffuse into the bulk solution; consequently because of the concentration differences between the solution and resin phases, and also to maintain electro neutrality, NOM diffuses into the positively charged resin phase and replaces Cl- ions stoichiometrically (Crittenden et al., 2012b). Even though ion exchange is simple in concept, there are many reactions and exchanges happening in parallel. Tan and Kilduff (2007) stated that NOM removal will take place under two dominant mechanisms, ion exchange and physical adsorption. Ion exchange is about electrostatic interaction between the NOM molecules and functional groups in resin phase. Physical adsorption is caused by van der Waals connections between the neutral fractions of NOM and resin polymer backbone (Tan and Kilduff, 2007). Eq. (2), represents NOM removal from water, where R- represents a negatively charged NOM molecule (Sillanpää, 2015). Resin⋯NMe(8Cl: +	R: 	⟷ 	Resin⋯NMe(8R: + 	Cl: (2) Eventually, all exchange sites become occupied (by NOM or any other contaminant), and the IEX resin is deemed ‘exhausted’. Ion exchange can be reversed through addition of a highly concentrated counter ion (Cl-) solution (i.e., brine) to the exhausted resins. Adsorbed ions (NOM) will be released and resin can function as a fresh resin. This is known as resin regeneration (see Eq. (3)). Resin⋯NMe=8R: + 	Cl: 	⟷ Resin⋯NMe=8Cl: +	R: (3)   9  Figure 2-2. Conceptual model of an anionic ion exchange resin bead.  As discussed earlier, IEX is a competitive treatment process for removing NOM, especially for cases involving high concentrations of organics (Fettig, 1999) or in small water systems (Amini et al., 2018). Over recent years, the number of studies on NOM removal by IEX has increased significantly (Amini et al., 2018; Finkbeiner et al., 2018; Sillanpää, 2015; Winter et al., 2018). These studies evaluated different aspects of the process, including the effects of resin characteristics and source water properties on organic removal performance. Some focused on the ion exchange mechanisms, as well as the challenge of resin regeneration. Unfortunately, there are only a few studies on full scale ion exchange applications and most of the studies are limited to lab or pilot scale research (Bhatnagar and Sillanpää, 2017). Table 2-1 summarizes some of the studies’ findings regarding the effect of the resin type or NOM and water characteristics on NOM removal efficiency.   10  Effect of resin properties on NOM removal Several studies have investigated different anion exchange resins with different properties for NOM removal. The properties investigated include functional groups, polymeric matrix type, ionic form (counter ion), water content or bead size. Most studies have compared various commercially available strong or weak anion exchanger resins. The results confirmed that strongly basic resins (SBA) show better performance at NOM removal (Bazri et al., 2016a; Bolto et al., 2002a; Croué et al., 1999; Sharbatmaleki, 2010). Regarding the backbone structure of the resin, acrylic anion exchange resins have been shown more efficient at NOM removal than polystyrene anion exchange resin (Boyer and Singer, 2008). Boyer and Singer observed much greater removal of Suwannee River Fulvic Acid (SRFA) using polyacrylic resins (Boyer and Singer, 2008). This was explained by the fact that polystyrene resins are more hydrophobic than polyacrylic resins. Therefore, in water, polyacrylic resins have a more open structure, which leads to higher NOM removal (Boyer and Singer, 2008; Pürschel, 2014). The pore structure has also been used to compare resins with different backbones. Bolto et al. (2002) stated that resins with a large and open macroporous structure is more efficient at NOM removal than gel-type resins (Bolto et al., 2002b). It is reported that larger fractions of NOM like biopolymers and fulvic acid could diffuse easier within a macro porous resin as they benefit from a more open structure (Bolto et al., 2002b; Boyer and Singer, 2008). Some of the studies also related this to the water content of the resin and showed that resin with higher water content often had larger pores, and could be more effective in the removal of high MW NOM (Bolto et al., 2004, 2002b; Cornelissen et al., 2008).   11 To sum up, in most of the studies, macroporous strong base anion exchange (SBA) resins showed the best performance in removing organics from water (Bazri et al., 2016a). Among all SBA resins, magnetic ion exchange resin (MIEX) is the most studied ion exchange resin (Boyer and Singer, 2006; Drikas et al., 2011). Purolite A860 is another macro porous strong base anion exchange resin with a polyacrylic backbone and exchange capacity of 0.8 Eq/L, with excellent performance for NOM removal (Bazri et al., 2016a; Pürschel, 2014; Winter et al., 2018). In a study comparing different anion exchange resins, Bazri et al (2016) concluded that Purolite A860 showed the best performance with respect to NOM removal (Bazri et al., 2016a).  Effect of NOM characteristics and water matrix on NOM removal Aside from type and characterisation of the resin, NOM characteristics and the water matrix have significant roles in the removal efficiency. The nature of organic matter such as molecular weight (MW), charge density and polarity along with the characteristics of water such as pH, ionic strength, hardness and turbidity are important influencing factors on NOM removal (Bolto et al., 2002b; Croué et al., 1999; Tan and Kilduff, 2007).  Previous studies indicated that anionic ion exchange resins show a preferential removal of smaller, polar and highly charged compounds; hence, being more hydrophobic and having a higher MW leads to lower efficiency (Bolto et al., 2002b; Croué et al., 1999). Tan and Kilduff (2007) found the greatest uptake of NOM in ion exchange mechanism belongs to the fractions with MW of 1000 Da (Tan and Kilduff, 2007). MIEX process also resulted in a relatively high efficiency into removing organic matter in the range of low to medium MW (100–10,000 Da) which was consistent with earlier findings (Humbert et al., 2007). Bazri et al. (2016a) reported preferential removal of smaller, low UV absorbing compounds during early stages of the treatment because   12 these small molecules can occupy the inner pores of the resin. The high UV adsorbing compounds (i.e., hydrophobic high MW fractions) were removed later (Bazri et al., 2016a). Furthermore, since anionic ion exchange resins attract negatively charged inorganic constituents, some anions present in raw water sources can compete with NOM and affect the removal mechanism and efficiency of IEX resins. Among these anions, sulphate is the most important parameter influencing DOC removal compared to other anions such as bicarbonate, nitrate, and bromide. It has been shown that the sulphate ion is preferentially removed over most NOM molecules due to higher IEX resin’s preference to it (Ates and Incetan, 2013; Bazri et al., 2016b; Boyer et al., 2008; Tan and Kilduff, 2007).      13 Table 2-1. Summary of studies investigated parameters influencing NOM removal over time.  Resin type/ Water source Results (Brattebø et al., 1987) Strong base anion exchange resins (SBA), macro porous, styrene (Lewatit MP 500A, Dowex MSA-1 and Dowex MSA-2) Medium base resin, macro porous, styrene (Lewatit OC 1035) Weak base resin (WBA), macro porous, styrene (Dowex MWA-I)  Lewatit MP 500 A showed the best performance either in isotherm experiments or in column test. The empty bed contact time was the most important parameter. Declared that ion exchange can be a promising method for Scandinavian soft water treatment, however, the costs of ion exchange vs traditional processes was not calculated.   (Afcharian et al., 1997) SBA, macro porous (Lewatit MP 500 and Lewatit S632SA) Non-ionic resins, macro porous (Lewatit VP OC 1062 and Lewatit VP 1064) Cation exchange resin, macro porous (Lewatit SP 112)  Source water: Seine River (Nitrified)  Two strong base resins removed 90% and 93% of the UV absorbing compounds and 59% and 73% of the DOC. Two non-ionic resins removed 23% and 28%, of the UV absorbing compounds. Cation exchange resin did not remove UV absorbing compounds.  DOC and UVA254 removal: Anionic resin> Non-ionic resin > Cationic resin  (Croué et al., 1999) SBA, gel type, styrene (Dowex 11) SBA, macro porous, styrene (Dowex MSA-1) WBA, macro porous, styrene (Imac HP 661)  Source water: NOM extracted from two surface water sources  In general (not always) strong anion exchange was more efficient than weak anion exchange.  Stronger the hydrophobic character of the NOM, the higher its MW and the lower removal of DOC with anion exchange resin. Physical adsorption can play a role in NOM removal, depending on the pH and the resin type.  (Bolto et al., 2004, 2002b) SBAs: Gel/ macro porous, styrene/ acrylic (Different Amberlite resins) Macro porous, styrene (Lewatit MP 500, Imac HP 555 and Purolite A520E) Gel, styrene (ResinTech SIR 22P) And CSIRO (MASB/ PDAA) WBAs:  Macro porous, pyridine (Reilex 4) Gel/ Macro porous, styrene/ acrylic (Amberlite resins)  And CSIRO MAWB  Coagulant (Alum)  Source water: Surface water dosed with extracted NOM   Strong base anion exchange resins offered better results in NOM removal than weak base resins. Higher water content and more open structure resulted in higher NOM removal. Resins with a macro porous structure removed more NOM than gel-type resins. Anion exchange resin showed a preferential removal of smaller, polar and highly charged compounds. IEX very efficiently removed essentially all of the charged material, while alum preferentially removed only the larger molecules.    14  Resin type/ Water source Results (Humbert et al., 2008, 2005) SBA, macro porous, acrylic (MIEXs and IRA 958) SBA, gel type, styrene (DOWEX-11) SBA, macro porous, styrene (DOWEX-MSA and IRA-938s)   ------------------------------------------------------- Powdered activated carbon (PAC)  Source water: Villejean/Rennes drinking water treatment plant (WTP). Mixture of two water sources (dam and river)  All resins found to be very effective to remove DOC. MIEX and IRA 938s show faster kinetic removal than two other resins. IRA-938s showed higher removal of larger and more aromatic structures than MIEXs due to its aromatic backbone and its large macro porous structure. ------------------------------------------------------------------- Comparing MIEXs, IRA938s and PAC. MIEXs and IRA938s showed higher and up to 75% removal of DOC after 30 min contact time. The results need to be validated on a full-scale treatment plant.   (Tan et al., 2005; Tan and Kilduff, 2007) SBA, gel (Marathon 11and Marathon A) WBA, macro porous (Dowex M-43)   Source water: Two surface water sources (microfiltered, softened, and concentrated)  - Tomhannock Reservoir (TMK): low DOC, high chloride, high sulphate concentration - Myrtle Beach (MB): high DOC, low chloride, low sulphate concentration NOM molecular weight (size exclusion) and sulphate concentration caused preferential uptake of different fractions of NOM. Sulphate is preferentially removed over NOM. NOM removal by strong base resins was higher and equal to the weak base resin.  NOM removal: Marathon A> Marathon 11³ Dowex M-43 Water content: Marathon A> Marathon 11> Dowex M-43  When the water content was high enough performance was good regardless of whether the pores were gel or macro porous or whether the polymer backbone structure was styrene or acrylic.  (Boyer et al., 2008; Boyer and Singer, 2008) MIEX  Source water: Different raw water sources with different SUVA values and anion concentrations.  ---------------------------------------------------------- 5 SBAs Polyacrylic (MIEX, Macro-T and IRA958) Polystyrene (IRA910 and A-641)  Source water: Synthetic water, suwannee river fulvic acid (SRFA) in the presence of bicarbonate and chloride.   MIEX has the greatest affinity For NOM with high charge density.  Higher sulphate concentration lower removal of DOC and UV absorbing compounds.   ------------------------------------------------------------------- All three polyacrylic resins exhibited similar removal of DOC. The polyacrylic resins exhibited much greater removal of DOC than the polystyrene resins. The polyacrylic resins had higher water content than the polystyrene resins.      15  Resin type/ Water source Results (Cornelissen et al., 2008) Nine different commercially available resins. (No name due to the confidentiality) Weak/Strong, macro porous/gel type, acrylic/styrene  Source water: surface water treatment plant at Weesperkarspel (Netherlands)  Confirmed Bolto’s findings on the importance of water content. Removal of NOM increased with an increase in water content of the resins and a decrease in average resin size. Resins with higher water content and smaller beads had longer breakthrough time.  (Graf et al., 2014) SBA, macro porous, acrylic (MIEX)  SBA, gel type, styrene (DOWEX Marathon 11and Purofine PFA444) SBA, macro porous, styrene (Dowex Tan-1 and Purolite A500P) SBA, gel type, acrylic (Purolite A850) SBA, macro porous/gel, acrylic/styrene (Tanex) GAC & PAC  Source water: St. Johns River (SJR) and a reservoir in Virginia (VA)  MIEX resin, AERs, GACs, and PAC showed 54%, 20–30%, 5–10%, and 20% DOC removal at a 30 min contact time.  DOC reduction: MIEX > Tan-1 ³ PFA444 ³ A850 ³ Marathon 11³ Tanex > A500P UVA reduction: MIEX > PFA444 ³ Tan-1 ³ Marathon 11 ³A850 ³ Tanex > A500P   (Bazri et al., 2016a; Bazri and Mohseni, 2016) SBA, gel type, acrylic (Amberlite IRA 458 and Lewatit VPOC 1071) SBA, macro porous, acrylic (Purolite A860 and Ionac Macro-T) WBA, gel type, acrylic (Purolite A847 and Lewatit VPOC 1073)   Source water: MilleÎles River (Québec) and the synthetic water (SRNOM)  ----------------------------------------------------------- SBA, macro porous, acrylic (Purolite A860)  Source water: Synthetic water Suwannee River NOM - SRNOM Suwannee River Fulvic Acid - SRFA Suwannee River Humic Acid - SRHA Pony Lake Fulvic Acid – PLFA Strong anion exchange resin A860 showed slightly better removal than weak anion exchange resin A847. Ion exchange was the dominant mechanism for NOM removal with Purolite A860 as a strong anion exchange resin. A combination of ion exchange and surface adsorption governed the removal with Purolite A847 as a weak anion exchange resin.  ------------------------------------------------------------------- Charge density and molecular weight were the influencing parameters in removal efficiency. Low charge density and high MW (size exclusion) negatively affected the NOM removal. Charge density (meq/g-C):  SRNOM (10.16) » SRFA (10.97) > SRHA (8.89) > PLFA (6.84) Molecular weight (Da):  PLFA (760) < SRNOM (1030) » SRFA (1070) < SRHA (1520) NOM removal:  SRNOM (96%) » SRFA (97%) > PLFA (93%) > SRHA (81%)    16  Resin type/ Water source Results (Rahmani and Mohseni, 2017) SBA, polyacrylic, macro porous (Ionac MacroT)  Source water: Synthetic water Including three different organic compounds: Naphthalene acetic acid-NAA Chlorobenzoic Acid-BA Gluconic Acid-GA  A proof for significance of positive effect of hydrophobicity in NOM removal. Hydrophobicity: (NAA > BA > GA) DOC removal: (NAA (74%) > BA (67.3%) > GA (34.7%)) Sulphate was less competitive in presence of more hydrophobic molecules (Dominant role of hydrophobic effect). Inorganic anions were preferentially removed by ion exchange resin. (Regardless of the type and structure of organic compound) Sulphate was more competitive than nitrate.   2.6 Challenges associated with ion exchange method in small systems Although anionic ion exchange is known as the most efficient method to remove NOM in small systems, it comes with a challenge related to the disposal of spent brine solution used for resin regeneration. Resins used in IEX process require frequent regeneration to maintain capacity to removal further contaminants. Sodium chloride solution (Brine-NaCl 10 Wt.%) is a common and efficient regenerant, and is often used in conventional IEX treatment (Amini et al., 2018; Rokicki and Boyer, 2011). For some hard-to-treat water, adding up to 2 Wt.% NaOH to the brine solution to have extra cleaning is recommended (Purolite, 2014). Bicarbonate salt (NaHCO3) is a more environmentally friendly alternative regenerant (used in MIEX resin regeneration); however, it is not as affordable for use as a regenerant brine solution (Ness and Boyer, 2017; Rokicki and Boyer, 2011). To make the required concentrated brine solution, a huge amount of salt is used which then needs to be disposed of. Due to the environmental hazard of discharged large volumes of high concentration salt solutions, disposal is a major limitation to the IEX process. This is especially an   17 issue for small communities that are not equipped with the proper resources to manage and/or dispose the spent brine. Previous studies have explored different approaches to regeneration by examining other reagents or treating the spent brine (Bazri et al., 2016a; Brattebø et al., 1987; Chandrasekara and Pashley, 2017; Grefte et al., 2013; Höll et al., 1981; Rokicki and Boyer, 2011; Schippers et al., 2004; Yang et al., 2013); however, to make this technology affordable and environmentally friendly more investigations are required.  2.7 Biological ion exchange (BIEX) Biological ion exchange (BIEX) is an exciting alternative to the conventional IEX processes. It has been introduced as a new concept to address issues common in IEX applications in small systems (Amini et al., 2018; Schulz et al., 2017; Wray et al., 2016). Explaining the concept of BIEX requires a brief background of biological filters. A biological filter (bio-filter) is a biologically active media including various substrates such as sand, soil and granular activated carbon (GAC) and was initially developed to remove particles and reduce biodegradable organic matter (BOM) (Chen et al., 2016). BOM is usually measured as biodegradable dissolved organic carbon (BDOC) is the biodegradable fraction of NOM with lower molecular weight. BDOC provide substrate for microbial growth in distribution systems and is a food for undesirable microorganisms like pathogens; therefore, removal of this fraction of NOM is required (Korotta-Gamage and Sathasivan, 2017; Thornhill and Kumar, 2018).  Several studies reported reasonable BDOC removal from water using biological treatments (Korotta-Gamage and Sathasivan, 2017) and biological filters equipped with either sand or GAC demonstrated effective BDOC removal from a water source. Comparing with the sand media, GAC can support more biomass; therefore, biological activated carbon (BAC) leads to a higher BDOC   18 removal (Thornhill and Kumar, 2018). Being affordable and cost competitive makes BAC one of the more accepted methods over traditional water treatment approaches (Chien et al., 2008; Korotta-Gamage and Sathasivan, 2017; Wang et al., 1995; Wert et al., 2008; Williams and Pirbazari, 2007). The removal of organic compounds in BAC filters takes place through both adsorption and biodegradation mechanisms (Ross et al., 2018). In other words, the presence of an active microbial community causes biodegradation of organics and less pore blocking which leads to higher NOM removal compared to the conventional GAC.  The enhanced NOM removal performance of BAC over GAC on the one hand, and the advantages of IEX over other adsorption processes, on the other hand, led to the development of biological ion exchange (BIEX). BIEX aims to avoid the waste-producing regeneration step in the IEX process, instead allowing the natural consortium of microbial community to grow/develop on the IEX resin in the filter bed. Eventually, this community can biodegrade a portion of NOM or DOC, much the same as what happens in BAC. The initial questions around the use of BIEX were:  - How does the IEX process in a biological mode compare with BAC or conventional IEX?  - How does this system perform over long-term operation? - What is the impact of water source characteristics on BIEX performance? - What are the mechanisms of NOM removal behind this process? Recently, researchers at the University of British Columbia addressed some of these questions. The lab scale set up used in those studies included packed bed columns filled with either GAC or anionic ion exchange resin in bed depth of 0.1 m. Columns operated under biotic (BAC/ BIEX) or abiotic (GAC/IEX) conditions. Sodium azide-NaN3 (0.01 Wt.%) was added to the inlet for the abiotic tests, to eliminate microbial activity in the columns. The results demonstrated significantly   19 higher NOM removal by IEX column operating in biotic condition (i.e., BIEX) compared to BAC. It was concluded that in the BIEX column, the presence of a microbial community could lead to effective operation beyond the point of conventional IEX resin exhaustion and reduce the regeneration frequency. It was also speculated that microbial activity caused greater removal of organic material in the system; in particular, the removal of humic substances, building blocks and low molecular weight acids (Schulz et al., 2017; Winter et al., 2018; Wray et al., 2016). In 2017, to better understand the performance of BIEX, researchers at Polytechnique Montréal, performed a pilot study evaluating GAC, BAC, IEX and BIEX columns operating in parallel and similar to operating conditions used at the UBC laboratories. The pilot was installed at the Pont-Viau water treatment plant in Laval, Quebec (Amini et al., 2018). The columns with 2 meters height were half-filled with media and operated at a 2 m/h flow rate. According to the reported results by Amini et al. (2018), during the first 64 days of operation BIEX showed 80% DOC removal followed by DOC breakthrough after 92 days of operation, after which DOC removal decreased gradually. The BIEX column operated without regeneration for 331 days. Over the 331 days, BIEX removed around 51% DOC, while DOC removal in BAC was around 8% (see Figure 2-3). After almost one year of operation, BIEX was successfully regenerated with brine. The researchers concluded that operation of ion exchange in biological mode is a promising option to reduce spent brine production while high DOC removal is still achievable (Amini et al., 2018).   20  Figure 2-3. Weekly dissolved organic carbon (DOC) monitoring in the source water (SW) and GAC, BAC, BIEX and IEX effluents during a pilot study from Polythechnique Montréal (Amini et al., 2018).  Table 2-2 summarizes the results achieved by studies investigated the NOM removal using ion exchange operating in biological mode. Regarding the questions mentioned above, recent studies showed that the BIEX process performed better for NOM removal compared with GAC or BAC.     21  Table 2-2. Results obtained regarding NOM removal by ion exchange process in a biological mode.  Conditions Results (Schulz et al., 2017) Purolite A860  Bench scale Source water:  Jericho Pond- JP water  “Biological operation enhanced NOM removal by approximately 50% due to an additional degradation of smaller humic substances, building blocks and low molecular weight acids.” “Promotion of biological activity significantly increased the time to breakthrough of the filters.” “Both the biotic and abiotic columns could effectively remove a large portion of the DOC present in the raw water. However, substantially greater DOC removal was achieved for the biotic columns.”  (Winter et al., 2018) Purolite A860  GAC  Bench scale Source water:  Jericho Pond (JP) Synthetic water (SR)  Significant difference between biotic and abiotic IEX columns in the removal of organics. Abiotic columns A gradual decrease in removal over time till breakthrough for both waters. (approximately 37 days-1800 BV)  Biotic columns JP water: A gradual decrease in removal for the first 1-3 weeks. However, after that the removal remained constant for 3 months followed by decreasing and observing breakthrough after 11 months (16000 BV). SR water: The same trend as JP water; however, removal remained constant for 7-8 weeks then decreased and breakthrough happened after 94 days.  Indicating that biological ion exchange can be used without the need for frequent regeneration over extended periods.  BIEX was more effective than BAC removing disinfection by product (DBP).  (Amini et al., 2018) Purolite A860 GAC  Pilot scale Source water: des Prairies River IEX had best performance at 80% and BIEX had equivalent performance for initial 50 days then got worse. BIEX removal somewhat related to water temperature. After around 11 months of operation, BIEX was successfully regenerated. DOC removal under steady-state BAC was typically in the low range of 5-20%  DOC removal in long-term performance: IEX>BIEX>GAC » BAC    22 2.8 Knowledge gap Recent studies confirmed that BIEX could be a possible replacement to IEX as it eliminates the need for frequent regeneration of resins (Amini et al., 2018; Schulz et al., 2017; Winter et al., 2018). So, it will address the main drawback of IEX treatment. However, there remain many uncertainties and unknowns related to the fundamentals of the process and the long-term viability of the technology. More experiments should be done to understand the mechanisms involved in this process or how different water characteristics affect the performance of the system. Besides, although the pilot study at Polytechnique Montréal shows an acceptable performance of the BIEX system in long-term, more field valuation is required on the performance of the system with respect to NOM removal under different conditions representing water treatment systems in remote communities. This study aimed to address some of the specific knowledge gaps as presented below: - Performance of the ion exchange process, operating without regeneration in the long term, with raw surface water serving a small community. - Mechanisms involved in NOM removal in a long-term operation. - Impact of inorganic compounds present in the source water on NOM removal by IEX process in a biological mode. - Comparison of the IEX operated without regeneration in packed bed reactor (PBR) and continuous stirred tank reactor (CSTR).   23 Chapter 3: Research Approach and Objectives 3.1 Research approach As discussed in chapter 2, biological ion exchange (BIEX) is introduced to overcome the limitations of IEX, i.e., brine disposal. Based on the promising results from the bench and pilot-scale studies on the removal of NOM by BIEX (Amini et al., 2018; Schulz et al., 2017; Winter et al., 2018; Wray et al., 2016), RES’EAU-WaterNET and its partners installed a pilot system at a First Nations community water treatment plant. The aim of this study was to further explore the efficacy of the BIEX process at removing NOM with real raw water, used as the source of drinking water for the community. This pilot set up operated between July 2017 and April 2018. Besides, lab scale experiments were carried out to investigate the impact of the microorganisms activity on NOM removal using CSTRs. The results from this study were then used to design a full scale treatment system, involving BIEX, for the removal of NOM and disinfection by-product precursors. 3.2 Research objectives and hypotheses The main objective of this work was to “investigate the efficiency of an IEX system operated in a way that allowed for the growth of microbial community (i.e., BIEX process whereby biological activity is promoted in the bed) for the removal of natural organic matter (NOM) under field conditions”. Achieving this main objective involved working on the following tasks: - Exploring different mechanisms involved in NOM removal over the period of the study. Looking through the previous results, it was noted that even after the primary ion exchange   24 stage, with resin exhaustion (i.e., no chloride release from the resin), the system removed NOM efficiently confirming that other mechanisms were involved in addition to ion exchange. - Assessing the system performance in long-term operation (more than 13,000 BV) with respect to the NOM removal efficiency and operational conditions. Earlier studies confirmed that BIEX system allows for a higher NOM removal than BAC system.  - Investigating the impact of water source characteristics (inorganic constituents) on the BIEX performance. Presence of some inorganic anions, such as sulphate, in the water sources may influence the ion exchange selectivity and removal efficiency. Consequently, it will impact IEX performance of the system over the long-term. Another objective of this work was to investigate the efficiency of the BIEX process in lab-scale CSTRs. Although in general PBRs are more efficient than CSTRs, benefits of using CSTRs, such as less resin inventory or no pressure drop encourage us to investigate their performance in the BIEX mode. Achieving this objective involved the following two specific sub-objectives; - Investigating the impact of microorganism’s activity on NOM removal by comparing two biotic and abiotic conditions in CSTRs. It was expected that biotic systems would show higher NOM removal over longer operational time in comparison with abiotic systems. - Exploring the effect of resin performance in long-term IEX operation (without regeneration) on the quality of treated water by anionic ion exchange using a CSTR. It was assumed that biological activity would improve the performance of the resin after exhaustion and would result in higher NOM removal in long-term IEX operation.   25 Chapter 4: Experimental Methodology and Procedures This chapter describes the methodology followed to fulfill the stated objectives of this research. It contains the general methodologies and analytical techniques that were performed on the samples. Experimental procedures come in two separate sections, consisting of 1) pilot and 2) laboratory scale studies. 4.1 Reagents and chemicals  Anionic ion exchange resin Given the promising results from earlier lab and pilot scale studies (Bazri et al., 2016a; Wray et al., 2016), Purolite A860, a macroporous strongly basic anionic ion exchange resin, was selected for this study. Table 4-1 presents more details about the resin used in this study. Table 4-1. Typical physical & chemical characteristics of Purolite A860. Characteristics Value* Polymer structure Macro porous polyacrylic crosslinked with divinylbenzene Appearance Spherical Beads Functional Group Quaternary Ammonium Ionic Form Cl- form Total Capacity 0.8 eq/L (Cl- form)  Particle size range 300-1200 µm *The values are provided by the manufacturer (www.purolite.com).   26  Reagents The reagents used in this study were limited to chemicals and standard solutions used for conducting different analyses. Also, sodium chloride (NaCl) and sodium hydroxide (NaOH) purchased from Thermo Fisher Scientific (Waltham, MA), were used for resin regeneration. Sodium azide (NaN3) purchased from Sigma Aldrich (St. Louis, MO) was used to suppress microorganisms activity and make abiotic condition. Deionized water (resistivity 18.2 M.Ω.cm) was used to prepare all the reagent solutions. 4.2 Glassware All glassware items used during this study were soaked in acid (HCl 0.1 Wt.%) overnight, then rinsed three times with deionized water and sealed with aluminum foil followed by baking in a Lindberg/Blue M Box furnace (Thermo Fisher Scientific, Waltham, MA) at 400 °C for 2 hours. 4.3 Analytical techniques and procedures Parameters monitored to evaluate the system performance by sampling over the course of experiments. For the pilot study, water samples were collected from the inlet and outlet streams and shipped to UBC for analyses. Samples were taken three times per week: 2 times from the lowest sampling port (EBCT=30 min) and one time from the top sampling port (EBCT=7.5 min). For the laboratory studies, samples were taken every 100 BV. Analytical techniques are described below:  Physical water characteristics (Turbidity and pH) Turbidity and pH of the raw water and treated water were measured using field equipment (see Table 4-2)   27 Table 4-2. Equipment used to measure turbidity and pH. Parameter Equipment Turbidity HACH 2100Q portable turbidity meter pH HANNA HI2002-01 edge dedicated pH/ORP Meter  Dissolved organic carbon (DOC) Concentration of dissolved organic carbon (DOC) in the samples was determined with a Shimadzu ASI-V TOC analyzer (Shimadzu Corp., Kyoto, KP), according to the high-temperature combustion method (Standard method 5310 B) (Clesceri et al., 1999). Blank and standard solutions with a known concentration were analyzed at the start and end of each series of samples to make sure the measurements were consistent. The method detection limit of the instrument was 0.21 mg/L carbon. The TOC analyzer begins by injecting a micro portion of samples (4 injection/ 1 outlier rejection) into a heated reaction chamber packed with an oxidative catalyst. The water is vaporized, and the organic carbon is oxidized to CO2 and H2O. The CO2 from oxidation of organic and inorganic carbon was transported in the carrier-gas streams and measured by an infrared analyzer. To determine the organic carbon (TOC) concentration, the inorganic carbon was removed by acidification and sparging. Furthermore, dissolved organic carbon (DOC) was measured after the samples was filtered with 0.45 µm syringe filters (17mm Target polypropylene) (Thermo Fisher Scientific, Waltham, MA) to eliminate particulate organic matter (Clesceri et al., 1999).   28  Biodegradable dissolved organic carbon (BDOC) Biodegradable dissolved organic carbon (BDOC) measurements were carried out at Polytechnique Montréal, by Dr. Barbeau’s research laboratory. Prepared samples being kept at 4oC were shipped to Montréal. In the method used (Servais et al., 1989) water samples were filtered into 125 mL glass bottles using 0.45 µm syringe filters (17mm Target polypropylene) (Thermo Fisher Scientific, Waltham, MA) while the natural microbial community was introduced to the samples and ammonium sulphate (NH4)2SO4 (2% v/v) and mono potassium phosphate KH2PO4 have been added. DOC was analyzed (DOC initial) and the remaining samples were incubated in the dark at 20oC for 30 days. After completion of incubation, the DOC of the samples was measured again (DOC final). BDOC = DOCInitial	- DOCFinal (4) Equation (4) was used to determine BDOC value as the difference in the value of DOCinitial and DOCfinal measured over the span of 30 days and was used to determine the biodegradable fraction of dissolved organic carbon.  UV absorbance at 254 nm (UVA254) and specific UV absorbance (SUVA) UV absorbance at 254 nm was measured using a Cary 100 UV-Vis Spectrophotometer (Agilent Technologies, Santa Clara, CA) according to the standard method 5910B (Clesceri et al., 1999). The method detection limit of the instrument was 0.0040 cm-1. The machine was operated in single beam mode using a 1 cm path length quartz cuvette and three times reading. SUVA was also determined using equation (5) below. This equation describes the ratio of UV absorbance to the DOC content of the water samples. Employing SUVA helps to represent the relative amount of aromatic carbon in NOM.   29 SUVA=UVA254DOC×100	 (5)  High performance size exclusion chromatography (HPSEC) To characterize different fractions and apparent molecular weight distribution of NOM in water, high performance size exclusion chromatography (HPSEC) using a Waters 2695 HPLC separation module equipped with a Waters 2487 dual λ absorbance detector set to detection at 260 nm was used. The eluent was a phosphate buffer of 0.01 M KH2PO4, 0.01 M K2HPO4, and 0.06 M NaCl (Thermo Fisher Scientific, Waltham, MA). The column flow rate was 0.7 mL/min. The standard solution was a set of poly sulfonate standards (15 kDa PSS15K, 7 kDa PSS7K, 4 kDa PSS4K, 3 kDa PSS3K) (American Polymer Standards Corporation, Mentor, OH). Equation (6) relates elution time (t in minutes) to the NOM apparent molecular weight (AMW) (Sarathy and Mohseni, 2007). Log	(MW) = 	−0.2857t + 6.9205 (6)  Disinfection by products potential tests (THM-UFC) Formation of trihalomethanes (THMs) as a group of disinfection by-products was investigated on some water samples. THM formation was measured using the uniform formation conditions (UFC) method (Summers et al., 1996). After measuring the chlorine demand at the UBC lab, samples were shipped to Polytechnique Montréal, Dr. Barbeau’s research laboratory, where the THM-UFC analysis was carried out. The measurement of THM was performed according to the USEPA method 524.2 (USEPA, 2007), using a gas chromatograph (GC) Agilent 7890B equipped with purge and trap stratum 9800 (Agilent Technologies, Santa Clara, CA).   30  Chlorine demand Chlorine demand was measured on some samples as a prerequisite to measuring THM according to the UFC method. Chlorine demand describes the amount of chlorine that reacts with NOM (and other compounds), as described by summers et al., 1996 (Summers et al., 1996). A 24-hour chlorine demand study was performed yielding a free chlorine residue of about 1.0±0.5 mg/L of Cl2. The ratio of initial chlorine dosing to the TOC concentration can vary from 1.2:1, 1.8:2 or 2.5:1. After measuring the initial dosed chlorine, samples were kept in the dark and at room temperature (22 oC) in pH 8 for the 24 hours. The chlorine concentration after 24 hours was used to determine the chlorine demand value using the following equation: Chlorine demand = Initial chlorine concentration - Chlorine concentration after 24 hrs (7)  Inorganic composition of water 4.3.8.1 Anions concentration The concentrations of anions such as chloride, nitrite, nitrate, phosphate and sulphate, present in the inlet and outlet were monitored according to the standard method 4110 B (Clesceri et al., 1999). An ion chromatograph (Dionex ICS- 1100, Waltham, MA) equipped with an electrical suppressor, an analytical column (AS22 Fast column), a guard column and a conductivity detector, was used for the analysis. The instrument detection limit was 0.2 mg/L. Prior to each analysis, samples were filtered using 0.45 µm syringe filter (17 mm Target polypropylene, Thermos Fisher Scientific, Waltham, MA). Sodium carbonate/bicarbonate concentrate (NaHCO3/Na2CO3) (Thermo Fisher Scientific, Waltham, MA) was used as eluent (Dionex AS22 eluent concentrate) to pass the sample through a series of ion exchangers. 5 mL sample was required to fill IC poly vials (Thermos Fisher   31 Scientific, Waltham, MA). Blanks and standards were analyzed at the beginning and end of each series of samples. A Dionex seven anion standard (Thermos Fisher Scientific, Waltham, MA) was diluted 10 times to reach the required range of concentration expected in the samples. Each sample was analyzed three times and an average was reported according to the standard method 4110 B (Clesceri et al., 1999). 4.3.8.2 Alkalinity Alkalinity was measured using Aquaculture photometer- HI83303 (Hanna Instruments, Woonsocket, RI) according to the colorimetric method, USEPA method 310.2 (USEPA, 1974). By adding the Alkalinity reagents (HI 775-26) (HANNA Instruments, Woonsocket, RI) to the sample, the reaction took place and the photometer measured the concentration from the color that was produced. The results were reported as concentration of CaCO3 in mg/L.  Ion exchange mechanism analysis 4.3.9.1 Chloride release  Chloride concentration in the inlet and outlet streams was measured using an ion chromatograph (Dionex ICS- 1100, Waltham. MA). Chloride released from the columns was determined using the following equation to explore ion exchange removal mechanisms.  Released	chloride = [chloride]TUVWXV − [chloride]YZWXV (8) 4.3.9.2 Mass balance  A mass balanced was performed to understand what fraction of NOM is removed by various mechanisms other than ion exchange, i.e., biodegradation. For each test, the following parameters were known and equations (9) and (10) display the relations between them.   32 1. Amount of NOM present in the inlet and outlet of the columns over the course of the experiment (indicating the total NOM removed in this study). 2. Amount of NOM (organics) extracted from the resin during the regeneration process, assuming that all adsorbed NOM is extracted from the resin during regeneration (representing the fraction of NOM removed by IEX). Total removed NOM (mg) = Total NOM Source water (mg) - Total NOM Outlet(mg) (9) NOM removed by biodegradation (mg) = Total NOM removed (mg) - Total NOM recovered from the used resin during regeneration (mg) (10)  Biological activity analysis 4.3.10.1 ATP measurements Microbial activity of the media (i.e., resin bed) was monitored by measuring adenosine triphosphate (ATP) concentrations. ATP was measured by introducing the sample to an enzyme-containing solution. Reacting ATP with the enzyme in the sample produced light and the light was detected in a luminometer as relative light units (RLU). The final RLU value of the samples was converted to the ATP using Equation (11). These measurements were performed with ATP kit (LuminUltra, Fredericton, NB). 1 gr or 1 mL of sample was added to 5mL UltraLyse (extraction tube) (LuminUltra, Fredericton, NB) and mixed well. After incubation for at least 5 minutes, 1 mL of the solution was transferred from the extraction tube to be combined with 9 mL UltraLute in a dilution tube (LuminUltra, Fredericton, NB) and mixed well. Finally, 100 µL of sample from the dilution tube was placed into the test tube and 100 µL of Luminas (enzyme) was added to it. The test tube was swirled gently five times and inserted into the luminometer (LuminUltra, Fredericton, NB). And the RLU was recorded. Equation (11) demonstrates the calculation.    33  ATP (pg ATP gr⁄ )= RLUATPRLUATP0 × 50,000 (pg ATP)mSample(gr) or Vsample(mL) (11) where RLUATP is the number luminometer gives, and RLUATPo is the value which is obtained after measuring the standard solution (Ultra Check) (LuminUltra, Fredericton, NB) with lumniometer. Either mass or volume of the sample could be used in this equation.  Morphology investigation The morphology of the fresh resin as well as the collected resin from the columns after 277 days of operation, before and after regeneration, were studied by field emission SEM (Hitachi S-4700, Japan) at UBC bioimaging facility and a compound LED microscope with USB camera (Omax 40X-2000X lab binocular, Kent, Washington, USA) at Dr. Wilkinson’s research laboratory.  Statistical analysis All error bars on graphs represent average of values from analysis of samples taken of both columns (in pilot experiment) or replicates (laboratory tests) with the range of max and min values, unless explained otherwise. 95% confidence interval was also calculated for some water quality parameters, based on the assumption of a Normal distribution. The general formula used for calculating the 95% confidence interval is as below; µ ± t0.95,n-1 × \√^  (12) where µ is the mean value; n is the number of samples, 𝛿 is the standard deviation, and the t value for 95% confidence interval is estimated based on t-distribution (two tailed).   34 4.4 Experimental process of pilot study  Location, historical data and source water characteristics The pilot study has been carried out in the community of Middle River, a small remote reserve of the Tl’azt’en Nation. The village is in north central British Columbia, over 110 km from Fort St. James. The nearest community to the village is 60 km away on a gravel road. As it is shown in Figure 4-1, the village is next to a large river and there are 12 homes with around 5-30 people year-round; the population of the village increases to over 40 in summer months.  Figure 4-1. Location of the Middle River village in north-central British Columbia, site of the pilot study.  The community had been under boil water advisory for over 14 years. Since 1998 efforts have been underway to identify a water treatment system that was appropriate for the community. At   35 the time, the village’s water treatment system consisted of a submersible pump drawing water from the river and a package plant consisting of a multimedia filter and a set of 1-micron cartridge filter followed by a sodium hypochlorite injection system to disinfect the water. Nevertheless, there were some concerns about the reliability of the system and its capacity. Also, there was a concern about high turbidity, colour and specifically organics in the raw water which caused the formation of disinfection by-products (DBPs). The emergence of membrane technology in late 1990s, led the authorities to implement a pilot project using ceramic nanofiltration technology to address the high organics and fluctuating turbidity issues of the raw water. However, the complexity of the system along with its frequent maintenance requirements resulted in the plant to become dysfunctional and the community went back to boil water advisory. Given the lack of resources and challenging/fluctuating raw water quality, the only treatment available was cartridge filtration with inconsistent/occasional application of chlorine as primary disinfection. Bacteriological tests on the source water (raw water), treated water (post filtration) and residential water (tap water) for the period April 2014 through December 2014 indicated repeated presence of total coliform bacteria (Figure 4-2). E. coli sampling in the same period also showed repeated presence of E. Coli counts in the source water, treated water and residential water (Figure 4-3).   36  Figure 4-2. Monitoring the presence of total coliform in raw, treated and tap water.   Figure 4-3. Monitoring the presence of E coli in raw, treated and tap water.    37  Figure 4-4. Middle River water treatment plant building before installation of experimental set up (July 2017).  The biological ion exchange columns used in this study were fed with the raw water, filtered with a 5-micron cartridge filter and stored in a large tank. Average values of the inlet water characteristics over the course of the study, from July 2017 to April 2018, are shown in Table 4-3. The water contained relatively modest level of DOC and low levels of minerals. The temperature of the water fed to BIEX was around 15 to 21 oC in cold and warm seasons, respectively. However, in the month of January, due to the breakdown of the heater of the building, temperature was around 7 oC.    38 Table 4-3. Water source quality after filtration process- Middle River - July 2017 to April 2018. Characteristics Value Number of samples Dissolve Organic Carbon “DOC” (mg/L) ~ 5.10 ± 0.5 97 pH ~ 7 36 Alkalinity (mg/L CaCO3) 39 ± 4 37 Turbidity (NTU) 0.8 ± 0.1 35 UVAbs254 (cm-1) 0.16 ± 0.02 100 Chloride (mg/L) 0.7 ± 0.1 82 Sulphate (mg/L) 4.1 ± 0.3 82 Nitrate (mg/L) 0.5 ± 0.1 82  Experimental set up The experimental set-up consisted of two parallel and identical packed bed ion exchange columns (see Figure 4-6). Figure 4-5 shows a schematic of the test column which was made of 2 inch diameter polyvinyl chloride (PVC) pipe, with a height of 39.37 inches (100 cm). Half of each column (50 cm) was filled with anionic ion exchange resin, Purolite A860. Each column included 3 sampling ports at depths of 50, 25 and 12.5 cm of the resin bed corresponding to empty bed contact times (EBCT) of 30, 15 and 7.5 minutes, respectively. These ports allowed to collect water samples at different EBCTs. Since the resin particles were quite small (minimum particle size of 300 µm) and could be easily lost from the column, small screens (mesh 60 with the nominal sieve opening of 0.25 mm) were placed at top and bottom of the column as well as at the sampling ports. A Master flex L/S peristaltic pump (Cole Parmer, Montréal, QC) pumped water from the water   39 source (water tank) into the columns from the top, allowing for a down-flow operation. A Master flex L/S pump (Cole Parmer, Montréal, QC) was also used to provide an upward-flow to backwash the system. A Hiblow air pump (Septic solutions, Dieterich, IL) was used to inject air, prior to the backwash with water. The tubing was replaced a few times during the system operation due to the biofilm growth. Table 4-4. Set up characteristics for this study and the previous works. Description Lab Scale (Schulz et al., 2017; Winter et al., 2018; Wray et al., 2016) Pilot Scale (Amini et al., 2018) Current Pilot Study Column Length (cm) 50 200 100 Column ID (inch) 3/8 4 2 Filter bed depth (cm) 10 100 50 Media (Resin) Purolite A860 Purolite A860 Purolite A860 Resin Volume (mL) 7.85 8100.00 950.00 Area (cm2) 0.71 81.03 19.00    40  Figure 4-5. Schematic of the experimental set up.      Figure 4-6. Experimental set up in Middle River water treatment plant (2017).   41  Operating conditions Table 4-5 provides the operating conditions of the pilot study. It also provides a comparison between this work and earlier studies. The columns operated with a total empty bed contact time (EBCT) of 30 min and filtration velocity of 1 m/h. This allowed for the treatment of water at 2 BV/h, corresponding to 48 BV per day. The number of bed volume (BV) was calculated as amount of treated water (L) to the amount of resin (L). The study started in July 2017 and ended in April 2018; the filter columns were under operation for 277 days. Table 4-5. Operating conditions for this study and previous studies. Operating Condition Bench scale (UBC) (Winter et al., 2018) Pilot scale (Poly MTRL) (Amini et al., 2018) Pilot scale (UBC) EBCT (min) 30 30 30 Filtration velocity(m/h) 0.2 2 1 BV/h 2 2 2 Backwash No Frequently (weekly) Not frequently Treated water (L/day) 0.36 388.80 46.00  Backwash Backwash is one of the important procedures related to the operation and maintenance of the packed bed columns. It helps prevent clogging or channeling in the bed and removes particulate matters and excess biomass from the columns. Given the remote location of the pilot and the goal of the study, to assess the long-term of operation of BIEX with little operator involvement, column backwash was performed on an infrequent basis, and only few times during the 8-month operation   42 of the pilot. Increased pressure drop in the bed was one indication that the system required backwash.  The backwash procedure consists of an air spurge followed by a water backwash. The procedure starts with all valves closed and the main pump was turned off while the level of water was about 20 cm above the resin bed. After opening the top of the column, air was injected gradually from the top of the resin bed to the bottom. A narrow pipe with attached air hose was passed vertically from top to bottom through the bed. This step was done carefully to avoid crushing the resins. Once the air hose had reached the lowest level of the resin bed, air injection continued for 3-5 minutes (bed expansion was around 50%-60%). After that, the air hose was carefully removed and the air spurge procedure was complete.  The water backwash started with a flowrate of 200 mL/min and lasted for 30 min. Bed expansion was close to 100%. Ideally, treated water can be used for this process, however since treated water was discharged to the sewer in this study and there was no tank to store it, the backwashing was done using partially-treated water from after cartridge filtration. Finally, the backwashing pump was turned off and all valves were closed. The media was allowed to settle naturally after which the columns started operation in the filtration mode.  Resin regeneration At the end of the field study and after 277 days of operation, resins were collected from the column and transferred to the UBC lab to perform standard regeneration. Note, there was no regeneration during the full pilot study. Different regeneration scenarios were examined to investigate their efficiency. Sodium chloride solution (NaCl 10 Wt.%) was used as the regenerant, as the conventional and established method of regeneration. Previous studies examined different   43 concentrations (higher) of brine and observed no significant difference (Schulz et al., 2017; Winter et al., 2018). Other regeneration scenarios included the addition of sodium hydroxide solution (NaOH 0.5 Wt.%) to the brine to give extra cleaning as well as incorporating sonication of the biofilm to examine the impact of biofilm on regeneration.  In all scenarios the ratio of regenerant to resin was 10 to 1 and regeneration happened in a well-mixed batch system using a mechanical stirrer. 20 mL saturated resin was regenerated in 200 mL regenerant solution. And 4 different regeneration times have been checked to see if higher regeneration time for the resin used for a long time (277 days- 13266 BV) was required. Sonication using an ultrasonic unit (Elmasonic model S 50 round, Germany) of the resin occurred while 20 mL of saturated resin was immersed in 300 mL deionized water. To make sure that biofilm was detached after sonication, the ATP content of the resin before and after sonication was measured. Table 4-6 summarizes different regeneration scenarios. Table 4-6. Different regeneration scenarios. Scenario Regenerant Regeneration time #1 NaCl 1, 2, 3, 4 hours #2 NaCl + NaOH 1, 2, 3, 4 hours #3 NaCl (Cleaning resin with sonication) 1, 2, 3, 4 hours To examine NOM removal performance of the resin after regeneration, all regenerated resins in different scenarios have been tested and their performances compared with each other. All columns were operated under the operation conditions similar to the pilot study with an EBCT of 30 minutes and filtration velocity of 2 BV/h. All columns were operated up to 4 BV and their NOM removal   44 investigated. However, to explore the regeneration efficiency, operation of resin for longer time and higher BVs after regeneration may be required. Consecutive loading cycles of IEX in a batch system were used to test the performance of the resin used for long trem after regeneration at a higher BV (1000 BV). Due to the time limitation, only one of the regenerated resins was selected to be compared with a fresh resin. The selected resin was regenerated in the same process as is most common in other studies: NaCl 10 Wt.% for 2 hours. 4.5 Lab scale studies  Source water characteristics The raw water in lab scale studies was collected from Priest lake in Texada Island, British Columbia, and Jericho Pond in Vancouver, British Columbia. Figure 4-7 and Figure 4-8 show the locations of the lake and the pond, respectively. After filtration with 5-micron Polyester Filter Bag (Cat. No 5162K117), raw water was mixed with distilled water to achieve the desired DOC concentration of around 5 mg/L. Raw water was kept at 4 oC and in dark during each experimental run. Table 4-7 presents some of the raw water characteristics.   45  Figure 4-7. Location of priest lake on Texada Island, British Columbia, Canada.  Figure 4-8. Location of Jericho Pond in Vancouver, British Columbia, Canada.    46 Table 4-7. Water source quality: Priest lake (November 2015) and Jericho Pond (March-May 2017). Characteristics Priest lake Jericho Pond Dissolve Organic Carbon “DOC” (mg/L) 5.51 ± 0.15 5.51 ± 0.36 UVAbs254 (cm-1) 0.185 ± 0.073 0.175 ± 0.038 Chloride (mg/L)  3.14 ± 0.21 10.26 ± 0.70 Sulphate (mg/L) 31.4 ± 0.21 2.33 ± 0.41  Experimental set up The experimental set-up consisted of two parallel and identical continuous stirred tank reactors operating under the same operational conditions (flowrate, residence time, resin volume) and in either a biotic or an abiotic condition (see Figure 4-10). Figure 4-9 shows a schematic of the continuous stirred tank reactor (CSTR) which was made of optically clear cast acrylic, with a volume of 400 mL. The reactor was equipped with a screen (mesh 60 with the nominal sieve opening of 0.25 mm) to prevent resin particles escaping the reactor with the outlet stream. A Master flex L/S pump (Cole Parmer, Montréal, QC) was used to pump the inlet into the reactor with the desired flowrate. The tubing was replaced a few times during the system operation due to the biofilm growth inside of them. To provide the abiotic condition, one of the raw water sources was dosed with powder sodium azide (NaN3) at a concentration of 0.2 g/L to suppress microorganisms activity.    47  Figure 4-9. Schematic of the benchtop continuous stirred tank reactor.   Figure 4-10. Experimental set up (CSTRs).   48  Operating conditions These studies happened in two different sets. The first set was tested up to 4,000 BV using Priest Lake water. The second set was operated up to 20,000 BV using Jericho Pond water and the test performed twice between March and May 2017. The experiments were performed in CSTRs with 10 mL IEX resin (Purolite A860). The reactors operated with the inlet flowrates of 10 and 13.3 mL/min corresponding to residence times of 30 or 40 minutes, respectively. Water was pumped into the 400 mL reactors using peristaltic pump. Considering constant reactor volume of 400 mL, the flowrates of 10 and 13.3 mL/min allowed for the treatment of water at around 80 and 60 BV/h, respectively. Number of bed volume (BV) was calculated as the amount of treated water (L) divided by the amount of resin (L). Samples were taken from the inlet and outlet at around 100 BV intervals and analyzed for DOC, UVA254 and anion concentration to monitor the performance. Table 4-8 summarizes the operational conditions for the lab scale studies. Table 4-8. Operation conditions of experiments performed in CSTR.  Operation condition First set (Priest Lake) Second set (Jericho Pond) Resin volume-Purolite A860 (mL) 10 10 Inlet Flowrate (mL/min) 13.3 10 Reactor Volume (mL) 400 400 Residence time-RT (min) 30 40 BV/h 79.8 60 Operation duration (BV) 4,000 20,000 Operation duration (hours) 50 333   49 Chapter 5: Pilot Study Results 5.1 Natural organic matter (NOM) removal  Dissolved organic carbon (DOC) Figure 5-1 shows the concentration of DOC in the source water (inlet) as well as in the treated waters from the two columns over 277 days of operation. Figure 5-2 is another representation of the same results, showing the removal percentage of NOM by the ion exchange columns. The DOC concentration in the source water was relatively stable with an average of 5.1 ± 0.5 mg/L throughout the study. Since both experimental columns operated under identical conditions, the results are shown as an average with the defined error bars (min/max values) for each set of data. As expected, the concentration of DOC in the treated water (outlet) was initially very low (<1 mg/L) (~85% DOC removal), and gradually increased to just under 2 mg/L at around 60 days of operation (~59% DOC removal). On the subsequent days, DOC concentration in the treated water fluctuated around 2 mg/L, then it rose to maximum 2.6 mg/L after 190 days of operation, followed by marginal decrease by the end of experiment. Different mechanisms were involved in the good performance (i.e., low outlet DOC concentration) of the BIEX system during the long-term operation, these will be discussed in later sections of this chapter.   50  Figure 5-1. Dissolved organic carbon (DOC) concentration vs. days of operation.  Figure 5-2. Dissolved organic carbon (DOC) removal vs. days of operation.    51  UV absorbance at 254 nm (UVA254) and specific UV absorbance (SUVA) The BIEX columns provided excellent performance in terms of the removal of UV absorbing compounds. As shown in Figure 5-3, UV absorbance in the source water was constantly around 0.16 cm-1, while absorbance of the treated water was around 0.01 cm-1, then went up slightly from 0.01 cm-1 to 0.03 cm-1. Between 160 and 230 days of operation, the UV absorbance fluctuated around 0.04 cm-1 after which it decreased to 0.03 cm-1 again. The significant reduction of UV absorbance is an indication that UV transmittance (UVT) increased substantially. This by itself will make water very suitable for downstream disinfection processes such as UV disinfection. Since UV disinfection is not generally approved for low UVT waters, improving UVT makes a significant impact with respect to the potential application of this technology in small systems. The results of this study were consistent with those obtained in the pilot study at Polytechnique Montréal (Amini et al., 2018), where a gradual increase in UV absorbance at 254 nm was also detected.   Figure 5-3. UV absorbance at 254 nm vs. days of operation.   52 To gain insights regarding the removal of aromatic compounds during the treatment, SUVA reduction was calculated and plotted. It can be clearly seen from Figure 5-4 that during the first 95 days of operation, SUVA reduction was approximately 80%, but it decreased to around 44% for the remainder of the pilot study. Given the positive correlation between molecular weight and SUVA, the obtained results agreed well with the literature (Tan et al., 2005), confirming that larger molecular weight organics are removed preferentially through ion exchange process.  Figure 5-4. Specific UV absorbance (SUVA) reduction vs. days of operation.   Chloride release As mentioned before, the dominant NOM removal mechanism varies through the course of the experiment. In general, chloride release along with DOC removal is an indicator of NOM removal by ion exchange: exchanging negatively-charged NOM with chloride ions (Cl-). Thus, the absence of chloride in the outlet (i.e., no chlorine release) could be associated with DOC removal by other   53 mechanisms. Figure 5-5 demonstrates chloride concentration in both inlet and outlet streams. The outlet values correspond to the average of both columns with the defined error bars (min/max). The chloride concentration in the outlet was 30 mg/L on the first day of operation and it decreased to 0.7 mg/L after 40 days. After about 60 days of operation, the concentration of chloride in the outlet was equal to that of the source water. No significant chloride release was observed after this point, even though DOC continued to be removed, confirming that primary ion exchange (i.e., exchange of NOM with chloride ions (Cl-)) was no longer the main mechanism for NOM removal.  When all sites on the anionic ion exchange resins are occupied, no further primary ion exchange can occur, and so the resin is considered ‘exhausted’. Therefore, that point (day of 60) was considered as primary ion exchange exhaustion, even though resin saturation had not happened yet.  Figure 5-5. Chloride concentration in inlet and outlet vs. days of operation.   54  Carbon mass balance A carbon mass balance was performed to better delineate the mechanism of NOM removal. This was achieved using DOC concentrations in the inlet and outlet of the columns, as well as the DOC concentration recovered from the spent brine, assuming that organics were recovered effectively from the resin during the regeneration. Such assumption was also made in earlier studies, with BIEX systems operated for ~2,600 BV (Winter et al., 2018) and ~15,840 BV (Amini et al., 2018), which reported effective/full regeneration of resins. In this study and according to the inlet and outlet DOC concentrations, 41 g carbon was removed over the entire experiment. Out of this value, based on the DOC recovered / measured in the spent brine, 28.5 g carbon was removed by ion exchange mechanism. This value corresponded to the 69.5% of the total NOM removed over the course of the experiment. The remaining part which was about 30.5%, was removed due to the other mechanisms such as biodegradation. Table 5-1 summarizes the observations mentioned above and Table 5-2 compares the results obtained in this study and the pilot study at Polytechnique Montréal (Amini et al., 2018). The fractions of NOM removed by IEX or other mechanisms in both studies are roughly the same confirming the consistency of the results. Table 5-1. Mass balance results of carbon in BIEX columns. Total organic carbon over the course of experiment (g) DOC Charge *(meq)  DOC in the inlet 61.7 g 610  DOC in the outlet 20.7 g 207  Total carbon removed 41 g (100%) 410  By ion exchange 28.5 g (69.5%) 285  By other mechanisms 12.5 g (30.5%) 125 * Charge density estimations for SRNOM is 10.157 meq/g C (Dixit et al., 2019).   55 Table 5-2. A comparison between two pilot studies with two different water sources. Condition Middle River  (UBC) Les Prairies River (Polytechnique Montréal) Resin volume 0.95 L 8.10 L Days of operation before regeneration 277 days 331 days DOC in the inlet 61.7 g 898 g DOC in the outlet 20.7 g 354 g Removed by IEX 28.5 g (69.5%) 372 g (68.5%) Other mechanisms 12.5 g (30.5%) 171 g (31.5%) DOC measured in spent brine 3 g/L 18.6 g/L  5.2 Biological ion exchange (BIEX)   Biological activity analysis ATP tests were performed to study biofilm growth on the resin surface. Three samples were taken from each column before and after the backwash on days 95 and 276. The sampling points were at different depths of the columns, 0%, 50% and 100% from the top of the bed, corresponding to different residence times of 0 min, 15 min and 30 min, respectively (see Figure 5-6).   56  Figure 5-6. Sampling point of resin collection for ATP tests.  Figure 5-7 shows the results of the ATP test, the data points are the average of the results from both columns and the error bars correspond to the min/max values. As expected, there was a profile of high concentration at the top and low concentrations at the bottom before the backwash. After the backwash, however, ATP was more evenly distributed and likely there was also some loss of ATP.  The ATP values of different sampling points at both times were roughly the same, suggesting there were minimal changes in biological activity during the steady state biological mode of operation. One exception was the sample taken from the top of the columns on day 95 where ATP concentration was significantly higher than that on day 276. This difference should be due to the fact that before 95 days of operation, no backwash had been performed on the system and hence, there was an accumulation of biofilms on the top (i.e., at the inlet). Once the backwash was performed, the resin bed was mixed thoroughly, resulting in the microbial community to be distributed more uniformly through the bed after the backwash. However, the average ATP   57 concentration obtained after 95 days of operation before backwash was much higher than that after backwash (127 ngATP/cm3 compared to 67 ngATP/cm3). Also, the values after 276 days of operation showed a small loss of ATP (89 ngATP/cm3 vs. 76 ngATP/cm3) as expected. Interestingly, Wray et al. (2016) previously reported approximately similar values after 60 days of BIEX operation. They presented a profile of higher concentration of 119 ngATP/cm3 at top of the column to a lower concentration of 10 ngATP/cm3 at bottom. However, the ATP concentration reported by Amini et al. (2018), after 133 days of operation, were around 28, 12 and 14 ngATP/cm3 for the top, middle and bottom of the resin bed, respectively. These values were lower than those observed in this study, which might be due to the more frequent (weekly) backwashing of the system which could result in the loss of biomass. Also, different techniques used might explain this difference, as Amini et al. detached the biomass from the resin by sonication (Magic-knezev and Kooij, 2004) and then then measured the ATP (Amini et al., 2018; Wray et al., 2016).    Figure 5-7. Biomass (ATP) profiles through depth of the filter bed (a) After 95 days of operation (4560 BV), (b) after 276 days of operation (13276 BV). a b   58 5.3 Impact of natural water characteristics  Inorganic composition of water Figure 5-8 shows the results obtained on sulphate removal during the study. Initially, the BIEX columns provided very effective removal of sulphate, in agreement with higher resin affinity towards sulphate. Indeed, sulphate removal stayed constant at around 93% for the first 140 days of operation. The most striking feature of this graph is the changing performance of the columns from removing sulphate to releasing sulphate at around 160 days of operation. After the initial period of excellent sulphate removal, the columns started releasing significant amounts of sulphate. Between days 194 and 277, the release of sulphate fluctuated around 53%. The reason behind this phenomenon could be related to maintaining equilibrium in the system, as any concentration change leads to impairing equilibrium and consequently causes adsorption or desorption.  Figure 5-8. Inlet and outlet sulphate concentration over the days of operation.   59 Table 5-3. Total sulphate removed-released over the course of experiment. Total sulphate removed-released Charge (meq) Days 1-160 (Removed) 25.3 g 527 Days 160-277 (Released) 7.54 g 157.2 Because of the different reactions happening in parallel in the system, to better understand the performance of the resins (and BIEX columns) at removing NOM and inorganic compounds, mass balance was performed on NOM, as well as sulphate ions for the whole 277-day operation. Other compounds such as bicarbonates due to the very low charge density were neglected in these calculations. Table 5-4 provides the equivalents of removed NOM and removed/released sulphate from the system, as well as the resin bed capacity. Based on the charges, the amount of the adsorbed NOM and sulphate was equal to the total resin capacity. Figure 5-9 shows the amount of adsorbed compounds (NOM and sulphate ions) on the resin at 10- day intervals in equivalent base. When reporting the net adsorbed sulphate values, the released sulphate was subtracted. As it is shown, the system adsorbed compounds steadily towards equilibrium and reached full saturation after 160 days of operation (note the plateauing of the graph). Table 5-4. Equivalents of different adsorbed compounds over 277 days of operation and the resin bed. Value Mass Charge density Charge Removed DOC 41.05 g 10.16 (meq/g C)* 0.41 eq Removed-Released Sulphate 17.8 g 20.83 (meq/g SO4−2)* 0.37 eq Total capacity of the fresh resin 950 mL 800 (meq/L Purolite A 860) 0.76 eq * Charge density estimations obtained from Dixit et al. (2019).   60  Figure 5-9. Cumulative capacity of the resin based on the adsorbed DOC and sulphate.  Alkalinity Source water alkalinity was monitored over the course of experiment and it was 39±4 mgCaCO3/L. Figure 5-10 shows the percentage of calcium carbonate which was removed or released during the process. Initially system removed alkalinity efficiently. Between day 1 of experiment and day 45 of the experiment, the alkalinity removal decreased from 90% to 0%. Then, slight amounts of CaCO3 seemed to have been released from the columns. A gradual release of CaCO3 from the column was detected between 54 to 105 days of operation. Then, the release percentage fluctuated around an average of 10%.  Figure 5-10. Alkalinity removal or release.   61 The pH as a correlated parameter with alkalinity was also monitored over the course of the experiment (see Figure 5-11). The pH of the raw water remained stable at 7.19 ± 0.51 throughout the study. However, the pH of the treated water was initially very low, around 5.0, and then increased gradually to a maximum of 7.7 at around day 80 of operation, followed by leveling off around 7.0. The drop in pH at the initial stage could be related to the removal of alkalinity and removing the buffering capacity of water, as it can be clearly seen that both figures followed the same trend over time.  Figure 5-11. Acidity (pH) in source water and treated water.  5.4 Possible mechanisms happening over the whole process In previous sections the removal or release of all the organics and inorganics was presented separately. However, these all occur in parallel and should be investigated together in order to find the mechanisms behind removals and releases of different solutes. Figure 5-12 shows the adsorbed NOM, sulphate, and both NOM and sulphate together in equivalent basis over the course of the   62 experiment. It also shows the resin bed capacity based on the used resin capacity provided by the manufacturer. As shown, NOM was removed steadily over the course of the experiment. Sulphate also had a consistent removal of 92% up to 160 days of operation; however, after that point the columns started releasing sulphate to reach equilibrium.  Figure 5-12. Accumulation of individual and total NOM and sulphate capacity over time (in calculating sulphate capacity the released sulphate values were subtracted).  As shown in Figure 5-12, the entire process can be separated into 3 different stages discussed as follows: First stage The first stage was up to 60 days of operation where chloride release was observed representing ion exchange mechanism. The system removed NOM and sulphate effectively over this stage. Theoretically, as there was no chloride release at the end of this period, the end point could be   63 called IEX exhaustion; however, the system still had capacity for the removal of NOM and sulphate the end point could be called primary IEX exhaustion. Second stage The second stage was between 60 and 160 days of operation. Comparing the equivalent of the total removed NOM and sulphate with the bed capacity (~350 meq vs. 760 meq), it was concluded that although no chloride release was observed at this stage, there were still resin sites available to remove NOM and sulphate; therefore, ion exchange mechanism was believed to be taking place. At the end of this stage, the equivalent of total adsorbed NOM and sulphate was equal to the total bed capacity, which one could consider this to be when resin breakthrough happened. On the other hand, ATP measurements for this period confirmed that the system was biologically active. Therefore, biodegradation of organics might have happened in parallel with the adsorption, contributing to the overall NOM removal.  Third stage The third stage was between 160 and 277 days of operation. At this stage, columns started releasing sulphate, even though NOM removal remained roughly the same. At the same time, according to the capacity of the bed and total adsorbed NOM and sulphate equivalents, it was determined that the system had reached equilibrium. Therefore, the reason behind sulphate release could be related to maintaining the equilibrium in the system. As a result of releasing sulphate ions, more sites on the resin would become available for NOM to be adsorbed. Therefore, another possible mechanism for NOM removal was introduced as “secondary ion exchange mechanism” which was about the replacement of sulphate ions with NOM molecules, thereby contributing to the removal of NOM   64 in the long-term. So, the main mechanism of NOM removal in the third stage is secondary ion exchange. At the same time, due to the presence of microorganisms in the column, biodegradation of organics continued to take place. As a result of biodegradation, bioregeneration was also a possible mechanism. The hypothesis around bioregeneration mechanism in the BIEX system is based on the theory suggested by Cecen et al. (2011) for the BAC filters (Cecen and Aktas, 2011). This hypothesis is based on the following assumptions: - System is at equilibrium. - Biofilm coverage is not uniform; therefore, resin is only partially covered by biofilm and the remaining parts are exposed to the liquid film. So, it could be regenerated from the areas that are covered by the biofilm (via biodegradation), allowing the remaining surface to have some extra sites available for NOM to adsorb on.  In a biologically active resin bed, through the degradation of NOM by biofilm, the NOM concentration at the interface of resin and biofilm is higher than the NOM concentration at interface of biofilm and liquid film. Consequently, adsorbed NOM on the surface of resin is desorbed due to the concentration gradient between the bulk solution and resin (a profile of higher NOM concentration on the surface of the resin to the lower NOM concentration in the interface of biofilm and liquid film). Therefore, more sites on the resin would be available for adsorption. Note this is a hypothesise and additional investigations will be needed to assess the role of bioregeneration in this process. Figure 5-13 is plotted to show how this bioregeneration and biodegradation may happen on the surface of the resin.   65   Figure 5-13. Hypothetic NOM concentration profile over bio degradation and bioregeneration.  5.5 Reduction in the formation of disinfection byproducts (DBP)  THM formation were measured in the inlet and treated waters at different EBCTs (7.5 and 30 min) after 144 and 277 days of operation. Figure 5-14 shows the THM-UFC values at different EBCTs. On both occasions, the THM-UFC concentrations in treated water were below the recommended Health Canada guideline (<100 µg/L). The THM-UFC values were 16 µg/L and 40 µg/L after 144 and 277 days of operation, respectively. The results showed that even the shorter EBCT of 7.5 min was sufficient for the system to reduce the formation of DBPs effectively. Comparing the removal on days 144 and 277, indicated that efficiency of the system decreased by about 4% and 10% for 7.5 min and 30 min, respectively. Chlorine demand was also performed as a required test for measuring THM precursors. Table 5-5 shows the results of raw water and treated waters at different EBCTs and at two different BVs of 6912 and 13296.    66  Figure 5-14. Removal of THM precursors at different EBCT after 144 and 277 days of operation.  Table 5-5. Chlorine demand values for raw and treated water at two different BV. Samples Chlorine demand (mg/L) Raw water (6912 BV) 3.78 Raw water (13296 BV) 4.37 Treated at 30 min EBCT (6912 BV) 0.58 Treated at 30 min EBCT (13296 BV) 1.12 Treated at 7.5 min EBCT (6912 BV) 0.77 Treated at 7.5 min EBCT (13296 BV) 1.14  5.6 Resin regeneration  Impact of regeneration duration and reagents Considering the long-term operation of the system without regeneration and biomass growth on the surface of the resins in this study, reduced regeneration efficiency was expected. Therefore,   67 different regeneration scenarios were tested to investigate their efficiency. Regeneration scenarios differing in the regenerant, regeneration duration or prepared resin were tested as mentioned in chapter 4 and summarized in Table 5-6. Table 5-6. Different regeneration scenarios. Scenario Regenerant Regeneration duration #1 NaCl 1, 2, 3, 4 hours #2 NaCl + NaOH 1, 2, 3, 4 hours #3 NaCl (Cleaning resin with sonication) 1, 2, 3, 4 hours For condition #3, where sonication was used to detach the biofilm from the resins prior to exposure to brine, ATP was measured after sonication, to ensure biofilm was completely detached. As seen in Table 5-7, sonication was able to remove 96% of the biofilm (relative to the ATP concentration after backwash), assuming that all the biomass in the resin (including all microorganisms in the pores) were measured by the ATP measurements. Table 5-7. Biomass concentration in different collected resins. Sample ATP (ng/cm3) Used resin for 13296 BV- Before backwash-performed on site 110 Used resin for 13296 BV- After backwash-performed on site 81 Used resin for 13296 BV-Washed with Sonication method-performed in Lab 3 As seen from Table 5-8, there was no significant difference among the results. As expected fresh resin showed a higher removal than the regenerated resins because these regenerated resins had   68 been used for 277 days without any regeneration process. All the regeneration scenarios were able to release roughly the same amount of organics and regenerate the resin. However, further tests are required to investigate how effective different scenarios were in terms of recovery of ion exchange capacity and make a strong conclusion. Table 5-8. DOC Removals by fresh resin and resins regenerated with different regeneration approaches. Scenario NaCl 10 Wt.% NaCl 10 Wt%+ NaOH 0.5 Wt.% NaCl 10 Wt.% + Sonication Fresh resin Duration (hr) 1hr 2hr 3hr 4hr 1hr 2hr 3hr 4hr 1hr 2hr 3hr 4hr - Removal (%) 87.7±0.5 89.1±0.4 87.5±0.5 87.8±0.5 92.3±0.3 93.0±0.3 93.2±0.3 91.7±0.3 88.1±0.5 88.2±0.5 87.7±0.5 86.6±0.5 94.1±0.2 * All regeneration scenarios happened in a well-mixed batch system using a mechanical stirrer with 20 mL saturated resin and 200 mL regenerant. Error propagations are presented in the table.  Explore the regenerated resin performance under long-term operation As mentioned in chapter 4, to explore the regenerated resin performance in long-term operation, consecutive loading cycles of IEX in a batch system were used to test resin performance at a higher BV (1000 BV). Because of the time limitation, only one of the regenerated resins was selected to be compared with fresh resin. The selected resin was regenerated in the same process as is most common in other studies: NaCl 10 Wt.% for 2 hours. Comparing DOC removals after 1000 BV, showed that regenerated resins provided 20% lower DOC removal compared to that provided by the fresh resin, showing that resin performance was not able to be recovered to that of fresh resin. This decrease may be due to the inner pores fouling in the resin matrix. These pores might be occupied with contaminants or dead microorganisms. More frequent resin regeneration might minimize this decline.   69 Table 5-9. Comparison between NOM removal at 1000 BV for regenerated resin and fresh resin. Resin Capacity Recovery NOM removal% at 1000 BV Fresh resin 73% Regenerated resin 51%  Impact of long-term usage and regeneration on resin morphology Excessive amount shrinking and swelling during loading and regeneration of the resin, may lead to physical stresses and impact the stability of the resin (Crittenden et al., 2012b). Therefore, morphology of the resin as a key factor on its performance was studied to investigate the impact of the long-term operation without regeneration or regeneration process on the resin morphology and consequently on the resin performance. In order to explore this, the resin beads (virgin resin, used resin and regenerated resin) were observed under optical microscope and size distribution analysis was performed.  The results showed that resin morphology changed noticeably through the course of long-term use (13296 BV). Figure 5-15 demonstrates different resin morphologies and their size distributions. According to the D50 (the diameter at which 50% of a sample's mass is comprised of smaller particles), fresh resin had a more open structure than used resin or regenerated resin. Even regeneration could not expand completely the compact structure of the resin used in an extreme condition (277 days of operation without regeneration). According to the size distributions, the fresh resin and used resin had approximately normal distribution; however, the regenerated resin had a right-skewed size distribution which might be related to the overall expansion after regeneration. Considering mentioned facts, it is more clear that extending regeneration intervals affects the resin morphology and further studies should be done to optimize those intervals.   70 Figure 5-15. Resin morphology and size order distribution for fresh resin, BIEX resin used for 277 days and regenerated resin after 277 days of operation. 5.7 Resin surface analysis- Field emission scanning electron microscopy Table 5-10 contains field emission scanning electron microscopy (FESEM) images of the surface of different ion exchange resins (fresh, used and regenerated resins). And Table 5-11 shows FESEM images of the surface of the resin used for 277 days in a BIEX mode before and after regeneration with NaCl 10 Wt.% for 2 hours, in a well-mixed batch system.  Fresh Resin BIEX resin used for 277 days Regenerated BIEX Resin          D10= 545 µm D50= 700 µm D90= 804 µm UC=1.33  D10= 517 µm D50= 651 µm D90= 781 µm UC=1.32  D10= 527 µm D50= 686 µm D90= 938 µm UC= 1.38  200 µm 400 µm 400 µm 400 µm 200 µm 200 µm   71 A smooth surface on the fresh resin beads was observed, while the surface of the used resin (after 277 days of service) had noticeable cracks and was covered by different kinds of microorganisms and debris. The cracks are likely the result of critical point drying, that is, if the beads were not fully dehydrated, the residual water could heat up, expand, and cause cracking. The surface of the regenerated resin was cleaner than that prior to regeneration; however, there were still traces of microorganisms and other debris on it. The remaining cells (alive or dead) observed on the surface of the regenerated resin could be another evidence for pore blockage and differences between the performance of the regenerated resin and fresh resins. As it shows in images with different magnifications in Table 5-11, the used resin prior to regeneration, seemed to show an array of microorganisms exposed by cracks. Comparing these two series of images, specially comparing the images “b” in both resins, confirmed the presence of biofilm on the surface of the used resin. Table 5-10. FESEM images of the fresh resin, used resin for 277 days, and regenerated resin. Fresh resin BIEX resin used for 277 days Regenerated BIEX resin       1.00 mm 1.00 mm 1.00 mm   72 Table 5-11. FESEM images of the surface of the BIEX resin (used for 277) before and after regeneration with NaCl 10 Wt.% for 2 hours, at different magnifications. BIEX resin used for 277 days Regenerated BIEX resin        b b a a c c   73 Chapter 6: Laboratory Study Results The results presented in this chapter are related to the two sets of experiments carried out as initial steps of this work in lab scale. The main objective of these experiments was to investigate the impact of microorganisms activity on NOM removal during drinking water treatment using an ion exchange process without regeneration in a CSTR. Therefore, the extent of biodegradation of NOM resulting from microbial activity on the surface of the IEX resins was investigated. Using a different reactor design from the one used in the pilot study (CSTR instead of PFR) as well as using different raw waters with different characteristics (i.e., different sulphate concentration) allowed for comparison of the performance of IEX operating in BIEX mode at different conditions.  Two parallel continuously stirred tank ion exchange reactors, one biotic and one abiotic (as control), were operated for an extended period to allow for the development and growth of biofilm on the surface of the resins. As explained in chapter 4, sodium azide (NaN3) with a concentration of 0.2 g/L was used in the abiotic system to inhibit microorganisms activity. The first and second sets of experiments operated for 4,000 and 20,000 bed volumes (BV), respectively, and their performances were assessed periodically by monitoring the concentrations of dissolved organic carbon (DOC) and sulphate (SO42-). Spectrophotometric measurements (e.g., UV254 and SUVA: UV254/DOC) and molecular weight distribution analysis were also carried out for the second set of experiments. Table 6-1 summarizes some of the conditions in both sets and more details are provided in chapter 4.     74 Table 6-1. Some of the operation conditions for two sets of experiments performed in CSTR.  Parameter First set  Second set  Water source Priest Lake Jericho Pond Raw water DOC concentration (mg/L) 5.51±0.15 5.51 ± 0.36 Raw water sulphate concentration (mg/L) 31.4±0.21 2.33 ± 0.41 Reactor type CSTR CSTR Residence time (RT) 30 min 40 min Operation duration (Bed Volumes) 4,000 20,000 Operation duration (hours) 50 333  6.1 Ion exchange performance without regeneration (Biotic/Abiotic)  Priest Lake-CSTR-30 min RT-Up to 4,000 BV Figure 6-1 shows the concentration of DOC in the source water (raw water) as well as the treated waters in two different biotic and abiotic conditions over the course of the experiment. No significant difference was observed between NOM removal in biotic or abiotic conditions; Hence, it was assumed that longer operation time was required for microorganisms to act and make the difference between biotic and abiotic condition. The DOC concentration in the source water was stable at 5.51±0.15 mg/L and suspended ion exchange was able to remove over 60% DOC initially. The removal gradually decreased to approximately 50% and fluctuated around that level until the end of the experiment (4,000 BV).   75  Figure 6-1. DOC concentration vs. bed volume in biotic and abiotic conditions.  Due to the resin selectivity, sulphate removal impacts the NOM removal and elevated sulphate concentration might inhibit NOM removal (Ates and Incetan, 2013). Therefore, the concentration of sulphate anions competing with NOM in the IEX process was monitored over the course of the experiment. Figure 6-2 shows sulphate concentrations in the raw water as well as the treated waters in biotic and abiotic systems. Similar to the trend observed for NOM removal, both systems (biotic and abiotic) performed roughly the same with respect to the removal and release of sulphate. Sulphate concentration in the raw water was 32.8±1.4 mg/L. Because of the high affinity of the ion exchange resin towards sulphate, it removed sulphate rapidly. The systems provided around 92% sulphate removal. However, as shown in Figure 6-2, after around 1,500 BV, sulphate concentrations in the treated water was greater than that in the inlet. The most probable explanation is the competitive effects among the competing compounds (sulphate ions and NOM) which   76 released the previously exchanged ions off the resin (Crittenden et al., 2012b). Between 1,800 BV and 2,700 BV sulphate release was about 40% and after 2,700 BV onward it fluctuated around 11% (see Figure 6-2).   Figure 6-2. Sulphate concentration vs. bed volumes in biotic and abiotic conditions.   Jericho Pond-CSTR-40 min RT-Up to 20,000 BV In the first set of experiments, no difference in NOM removal between biotic and abiotic conditions indicated that longer time of operation would be required for microorganisms to be active on the surface of resin. Therefore, the system in the second set was operated for a longer time: up to 20,000 BV. Also, residence time was increased from 30 min to 40 min to achieve a higher NOM removal in CSTRs. As mentioned in the methodology section, the raw water was collected from Jericho Pond, Vancouver. In contrast of the first set, an approximately similar range of sulphate concentration in Jericho Pond water were detected compared to the Middle River water used during   77 the pilot study discussed previously in chapter 5. This made it possible to compare the results of different studies. Figure 6-3 shows DOC removal in both systems. According to the DOC removals, both systems performed similarly at both conditions; however, the abiotic system showed less removal for the whole period of the experiment, which may be due to the sodium azide and consequently the impact of nitrate ions on NOM removal. The DOC removal in biotic system started from less than 60% and gradually decreased and reached around 23% after 20,000 BV.  Figure 6-3. DOC removal vs. bed volumes for both biotic and abiotic systems.  During ion exchange, resins remove sulphate rapidly and almost completely (Ates and Incetan, 2013). Figure 6-4 illustrates treated water sulphate concentrations in both CSTRs (biotic and abiotic). The graph followed the same trend as that observed in previous tests. The system removed sulphate until breakthrough of sulphate happened, then similar to the first set, the outlet concentrations were higher than the inlet concentration. However, due to the lower sulphate   78 concentration this breakthrough happened at much later BVs (i.e. around 8,000-10,000 vs. around 1,000 BV). Sulphate removal during the biotic condition was initially higher and more stable than that in the abiotic system. This was also observed in the previous test which could be because of the presence of sodium azide in the water in abiotic condition. Ion chromatography analysis detected nitrate ions in the water dosed with sodium azide and nitrate as another anion to be adsorbed in IEX process might have impacted the NOM and sulphate removals.   Figure 6-4. Treated water sulphate concentration at two different biotic and abiotic conditions vs. bed volume.  The results obtained from laboratory experiments  at two different biotic and abiotic conditions confirmed that microorganisms were not able to make significant differences at different biotic and abiotic conditions even at longer time of operation (20,000 BV). However, there were a few   79 parameters that may impact the results and need to be considered for the future work. Those parameters were as follow: - The use of sodium azide to create the abiotic conditions could be one of the impacting factors on the ion exchange mechanism due to the changing conditions such as adding nitrate anions to IEX process. Applying another approach to suppress microorganisms activity may address this issue.  - Furthermore, the use of a CSTR could make it difficult for microorganisms to grow on the surface of resin due to the specific flow regime and mass transfer limitations in this reactor (Winter et al., 2018). Therefore, biofilm formation in the system was not as efficient to make any difference.  A comparison between ion exchange performance in pilot and lab scale studies (PBR and CSTR) Figure 6-5 and Figure 6-6 compare the efficiency of ion exchange mechanism in three different experiments (pilot study, and two sets of laboratory studies). As the conditions (reactor type and amount of the resin) varied in these experiments, the mass values were normalized according to the volume of the resin used in each experiment. The figures show the cumulative NOM or sulphate removed per 1 L of the resin. Figure 6-5 shows that all three experiments removed NOM steadily over time, although the efficiencies were different. PBR-Middle River and CSTR-Jericho Pond could be compared with one other as raw water used in those experiments contained the low ranges of sulphate. According to the plot, the PBR was more efficient in NOM removal than the CSTR and this difference would be more significant over the long-term (higher BVs).   80  Figure 6-5. Accumulative removed NOM (g) over the period of each experiment per 1 mL resin.  Regarding sulphate removal, Figure 6-6 shows that all the experiments (pilot study and two sets of lab-scale studies) followed the same trend, that is, the sulphate loading on 1 L of resin increased to a peak before decreasing. However, these maximums happened at different BVs and in different levels. As can be seen from Figure 6-6, the first peak belongs to the Priest Lake-CSTR test with raw water sulphate concentration around 33 mg/L and the second and third peaks are related to the Middle River-PBR and Jericho Pond-CSTR with raw water sulphate concentrations of 4.1 and 2.3 mg/L, respectively. The higher the initial sulphate concentration, the sooner the peak was observed. This was expected as generally higher concentration causes exhaustion sooner. Competition between NOM and sulphate ions impacts the ion exchange performance. Sulphate concentration impacts on the preferential removal of NOM fraction (Tan and Kilduff, 2007); however, more analysis on NOM molecular weight distribution of the raw and treated water are required in order to make a more conclusive statement.   81  Figure 6-6. Cumulative removed sulphate (mole) over the period of each experiment per 1 L resin.  6.2 Estimation of ion exchange resin exhaustion (Biotic condition) The net amount of NOM and sulphate removed/released as accumulated values in equivalents was calculated for both experiments to estimate ion exchange resin exhaustion.  CSTR experiments (Priest Lake- 30 min RT- up to 4,000 BV) As shown in Figure 6-7, the net amount of the removed NOM and sulphate in equivalent reached a semi equilibrium state between 1,000 and 2,000 BV, after which it gradually decreased. During the pilot study (refer to Figure 5-9), the net amount of NOM and sulphate equivalents removed increased steadily and reached the total resin capacity around 8,000 BV. However, in this experiment the highest capacity attained was around 5.5 meq after around 1,000 BV which was   82 lower than the total resin capacity (8 meq). Further, this value decreased gradually to around 3 meq at 4,000 BV. Differences in the cumulative capacities at resin breakthrough is possible because of the different characteristics of the PBR and CSTR. In general a PBR is more efficient than CSTR (Crittenden et al., 2012c). This can be determined by comparing the removal performance of a CSTR and a PBR in the same condition. The rate of removal in CSTR governed by concentration in the reactor which is equal to the outlet concentration however the rate in a PBR is governed by a concentration representing the average of a profile from higher concentration in inlet to the lower concentration in outlet. This results in higher rate of removal in the PBR and consequently more removal and more occupied site on the resin.  Another difference was related to the point of releasing sulphate from the system. Due to the higher sulphate concentrations in this experiment compared to the pilot (31.4 mg/L vs. 2.33 mg/L) the breakthrough of sulphate happened sooner and consequently the system started releasing sulphate earlier (~1,200 BV vs. ~8,000 BV).   83  Figure 6-7. Accumulative equivalents of removed DOC and removed-released sulphate vs. bed volume in biotic systems (Priest lake-CSTR).   CSTR experiments (Jericho Pond- 40 min RT- up to 20,000 BV) The raw water used in this set of the CSTR experiments contained sulphate with a concentration in roughly the same range as the one in raw water used in pilot study. Therefore, similarities in the results, were expected as a result of the above conditions. The cumulative DOC and sulphate uptake in equivalent basis are shown in Figure 6-8. The system removed NOM steadily over the duration of the experiment. Sulphate removal also happened steadily, however, after around 10,000 BV a release of sulphate was observed. The system reached a semi equilibrium around 11,000 BV, similar to the pilot study which was at equilibrium after 8,000 BV. However, similar to that noted in the first set of experiments, the total value of the   84 removed compounds charges did not reach the total capacity of the resin (6 meq vs. 8 meq), the same as what observed in previous experiment.  Figure 6-8. Accumulative DOC and sulphate equivalents over the course of experiment in biotic condition (Jericho Pond-CSTR).  A comparison of ion exchange resin exhaustion in pilot and lab scale studies (PBR and CSTR) To explore the rate of resin exhaustion over time in lab scale studies as well as the pilot study, the percentage of resin capacity occupied by NOM and sulphate ions for each set of experiment is calculated and plotted in Figure 6-9. According to the graph, Middle River-PBR and Jericho Pond-CSTR exhibited more similar trends in terms of resin exhaustion. In these experiments, resin saturation increased gradually before reaching an equilibrium. In the pilot scale study, at equilibrium (around 8,000 BV) resin was completely saturated. NOM removal after this point, while the resin was saturated confirms the involvement of other removal mechanisms (i.e., secondary ion exchange). Although the Jericho Pond-CSTR test showed similar trends to that of   85 the pilot, around 20% of the resin capacity was available after equilibrium in this test. In the Priest Lake-CSTR, due to the high concentration of sulphate in the raw water, the resin loaded very fast, then the system started to release sulphate and more sites were available in higher bed volumes.  Figure 6-9. Resin capacity (%) occupied by NOM and sulphate at different experiments over time.   86 Chapter 7: Conclusions and Recommendations Laboratory and pilot scale studies have been performed to investigate the efficiency of an IEX system operated in long-term without regeneration at NOM removal. The main part of this study belongs to the pilot experiment performed for 277 days in Middle river, a small remote community in north central British Columbia. The key conclusions and outcomes from this pilot study as well as the complementary lab studies are as follows: - IEX columns removed NOM efficiently for 277 days of operation (more than 13,000 BV) without regeneration in a pilot scale. It was concluded that mechanisms other than primary ion exchange (i.e., exchange of NOM with chloride ions) would be involved in NOM removal over the long-term operation of the system. The mechanisms involved in this process were introduced as primary ion exchange, secondary ion exchange (replacement of NOM with sulphate ions), bioregeneration and biodegradation. - Laboratory studies showed no significant difference between biotic and abiotic conditions in NOM removal in CSTR; This was speculated to be because the selected condition might not have been desirable for microorganisms growth.  - Comparing the results obtained from pilot study and laboratory studies, roughly the same trend for resin exhaustion observed. However, in PBR the full capacity of the resin was utilized, while in CSTRs around 80% capacity of the resin was occupied.   87 7.1 Recommendations for future work As BIEX is still an incipient technology, there are many areas and outstanding questions that require further investigation and research. Some of the topics that could be covered as part of the future research are as follows: The impact of water source characteristics on NOM removal in BIEX systems - Investigate the impact of sulphate ions in the water source on the performance.  - Study the removal of different fractions of NOM over short/long-term. - Investigate the impact of type of microorganisms in the water source by identifying them or adding selective bacteria to the water source. The impact of resin type - Explore the performance of other types of resins, such as weak or strong base resins. - Monitor changes in morphology and capacity of the resin by collecting resin samples over the course of NOM removal procedure. The fundamental of IEX - Effect of EBCT on the system performance for the purpose of optimizing it in columns treatments. - Impact of temperature on NOM removal (this is an on-going at UBC) - Study regeneration efficiency to find the optimum frequency of regeneration as well as regeneration conditions. 7.2 Significance and outcome of the research This study was intended to further investigate the long-term NOM removal efficiency by ion exchange operating in BIEX mode. It provided knowledge about possible mechanisms happening   88 through the process and impacting the NOM removal. This knowledge gained through this study, would be useful in future water treatment system designs. Also, comparing different experiments at different scales and different reactors or source water characteristics made it possible to make more comprehensive conclusions on the potentials of this technology. After recognizing the potential benefits of the BIEX system such as avoiding chemical treatment and subsequent improvement in UV disinfection for small and remote communities, RES’EAU-WaterNET and its partners decided to build the first full scale BIEX system at the same community where the pilot study was conducted. Therefore, the promising results of this study were used towards the design of a full scale BIEX system, aiming to eliminate the long-term drinking water advisory in Middle River village. The system was built, installed, and commissioned in September 2018 and is currently supplying treated water to the community. As a result, the long-term boil water advisory has been lifted for this community and BIEX process is now considered as a new solution for safe drinking water treatment.  Figure 7-1. Middle River water treatment pilot plant installed on September 2018.    89 References  Afcharian, A., Levi, Y., Kiene, L., Scribe, P., 1997. Fractionation of dissolved organic matter from surface waters using macroporous resins. Water Res. 31, 2989–2996. Amini, N., Papineau, I., Storck, V., Bérubé, P.R., Mohseni, M., Barbeau, B., 2018. Long-term performance of biological ion exchange for the removal of natural organic matter and ammonia from surface waters. Water Res. 146, 1–9. Ates, N., Incetan, F.B., 2013. Competition Impact of Sulfate on NOM Removal by Anion-Exchange Resins in High-Sulfate and Low-SUVA Waters. Ind. Eng. Chem. Res. 14261–14269. Bazri, M.M., Barbeau, B., Mohseni, M., 2016a. Evaluation of Weak and Strong Basic Anion Exchange Resins for NOM Removal. J. Environ. Eng. 142, 1–8. Bazri, M.M., Mohseni, M., 2016. Impact of natural organic matter properties on the kinetics of suspended ion exchange process. Water Res. 91, 147–155. Bazri, M.M., Sarathy, S., Mohseni, M., 2016b. Enhancement of UV H2O2 efficacy using strong base anion exchange resins. J. Am. Water Works Assoc. 108, 318–326. Bhatnagar, A., Sillanpää, M., 2017. Removal of natural organic matter (NOM) and its constituents from water by adsorption – A review. Chemosphere 166, 497–510. Bolto, B., Dixon, D., Eldridge, R., 2004. Ion exchange for the removal of natural organic matter. React. Funct. Polym. 60, 171–182. Bolto, B., Dixon, D., Eldridge, R., King, S., 2002a. Removal of THM precursors by coagulation or ion exchange. Water Res. 36, 5066–5073. Bolto, B., Dixon, D., Eldridge, R., King, S., Linge, K., 2002b. Removal of natural organic matter by ion exchange. Water Res. 36, 5057–5065. Boyer, T.H., Singer, P.C., 2008. Stoichiometry of removal of natural organic matter by ion exchange. Environ. Sci. Technol. 42, 608–613. Boyer, T.H., Singer, P.C., 2006. A pilot-scale evaluation of magnetic ion exchange treatment for removal of natural organic material and inorganic anions. Water Res. 40, 2865–2876. Boyer, T.H., Singer, P.C., Aiken, G.R., 2008. Removal of dissolved organic matter by anion exchange: effect of dissolved organic matter properties. Environ. Sci. Technol. 42, 7431–  90 7437. Brattebø, H., Ødegaard, H., Halle, O., 1987. Ion exchange for the removal of humic acids in water treatment. Water Res. 21, 1045–1052. C.E.Harland, 1994. Ion Exchange : Theory and Practice. Cecen, F., Aktas, O., 2011. Activated carbon for water and waste water treatment: Integration of adsorbtion and biological treatment, First. ed. Wiley-VCH. Chandrasekara, N.P.G.N., Pashley, R.M., 2017. Regeneration of strong acid/strong base mixed-bed resins using ammonium bicarbonate (AB) for a sustainable desalination process. Desalination 409, 1–6. Chen, B., Westerhoff, P., 2010. Predicting disinfection by-product formation potential in water. Water Res. 44, 3755–3762. Chen, F., Peldszus, S., Elhadidy, A.M., Legge, R.L., Van Dyke, M.I., Huck, P.M., 2016. Kinetics of natural organic matter (NOM) removal during drinking water biofiltration using different NOM characterization approaches. Water Res. 104, 361–370. Chien, C.C., Kao, C.M., Chen, C.W., Dong, C.D., Wu, C.Y., 2008. Application of biofiltration system on AOC removal: Column and field studies. Chemosphere 71, 1786–1793. Clesceri, L.S., Greenberg, A.E., Eaton, A.D., 1999. Standard Methods for the Examination of Water and Wastewater Part 4000, inorganic nonmetallic constituents standard methods for the examination of water and wastewater. American Public Health Association, Washington DC. Cornelissen, E.R., Moreau, N., Siegers, W.G., Abrahamse, A.J., Rietveld, L.C., Grefte, A., Dignum, M., Amy, G., Wessels, L.P., 2008. Selection of anionic exchange resins for removal of natural organic matter (NOM) fractions. Water Res. 42, 413–423. Crittenden, J.C., Trussell, R.R., Hand, D.W., Howe, K.J., Tchobanoglous, G., 2012a. water treatment: Principles and Design, in: Water Treatment: Principles and Design. John Wiley & Sons, pp. 17–71. Crittenden, J.C., Trussell, R.R., Hand, D.W., Howe, K.J., Tchobanoglous, G., 2012b. water treatment: Principles and Design, in: Water Treatment: Principles and Design. pp. 1263–1334. Crittenden, J.C., Trussell, R.R., Hand, D.W., Howe, K.J., Tchobanoglous, G., 2012c. water   91 treatment: Principles and Design, in: Water Treatment: Principles and Design. pp. 287–390. Croué, J.P., Violleau, D., Bodaire, C., Legube, B., 1999. Removal of hydrophobic and hydrophilic constituents by anion exchange resin. Water Sci. Technol. 40(9) ,207-214. Dixit, F., Barbeau, B., Mohseni, M., 2019. Science of the Total Environment Removal of Microcystin-LR from spiked natural and synthetic waters by anion exchange. Sci. Total Environ. 655, 571–580. Dorfner, K., 1972. Ion Exchangers: properties and applications, Ann Arbor science publishers. Drikas, M., Dixon, M., Morran, J., 2011. Long term case study of MIEX pre-treatment in drinking water; understanding NOM removal. Water Res. 45, 1539–1548. Edzwald, J., 1993. Coagulation in drinking-water treatment - particles, organics and coagulants. Water Sci. Technol. 27, 21–35. Fettig, J., 1999. Removal of humic substances by adsorption/ion exchange. Water Sci. Technol. 40, 173–182. Finkbeiner, P., Redman, J., Patriarca, V., Moore, G., Jefferson, B., Jarvis, P., 2018. Understanding the potential for selective natural organic matter removal by ion exchange. Water Res. 146, 256–263. Gibert, O., Lefèvre, B., Fernández, M., Bernat, X., Paraira, M., Pons, M., 2013. Fractionation and removal of dissolved organic carbon in a full-scale granular activated carbon filter used for drinking water production. Water Res. 47, 2821–2829. Graf, K.C., Cornwell, D.A., Boyer, T.H., 2014. Removal of dissolved organic carbon from surface water by anion exchange and adsorption: Bench-scale testing to simulate a two-stage countercurrent process. Sep. Purif. Technol. 122, 523–532. Grefte, A., Dignum, M., Cornelissen, E.R., Rietveld, L.C., 2013. Natural organic matter removal by ion exchange at different positions in the drinking water treatment lane. Drink. Water Eng. Sci. 6, 1–10. Höll, W., Kiehling, B., Holl, W., 1981. Regeneration of anion exchange resins by calcium carbonate and carbon dioxide. Water Res. 15, 1027–1034. Humbert, H., Gallard, H., Jacquemet, V., Croué, J.P., 2007. Combination of coagulation and ion exchange for the reduction of UF fouling properties of a high DOC content surface water. Water Res. 41, 3803–3811.   92 Humbert, H., Gallard, H., Suty, H., Croue', J.P., 2005. Performance of selected anion exchange resins for the treatment of a high DOC content surface water. Water Res. 39, 1699–1708. Humbert, H., Gallard, H., Suty, H., Croué, J.P., 2008. Natural organic matter (NOM) and pesticides removal using a combination of ion exchange resin and powdered activated carbon (PAC). Water Res. 42, 1635–1643. Iriarte-velasco, U., Jon, I.A., Juan, R., 2008. Natural organic matter adsorption onto granular activated carbons : implications in the molecular weight and disinfection byproducts formation. Ind. Eng. Chem. Res. 7868–7876. Kennedy, M.D., Kamanyi, J., Heijman, B.G.J., Amy, G., 2008. Colloidal organic matter fouling of UF membranes: role of NOM composition & size. Desalination 220, 200–213. Korotta-Gamage, S.M., Sathasivan, A., 2017. Potential of a biologically activated carbon treatment to remove organic carbon from surface waters. Int. Biodeterior. Biodegrad. 124, 82–90. Leenheer, J.A., Croue’, J.-P.E., 2003. Peer Reviewed: Characterizing aquatic dissolved organic matter. Environ. Sci. Technol. 37, 18–26. Levchuk, I., Rueda Márquez, J.J., Sillanpää, M., 2018. Removal of natural organic matter (NOM) from water by ion exchange – A review. Chemosphere 192, 90–104. Magic-knezev, A., Kooij, D. Van Der, 2004. Optimisation and significance of ATP analysis for measuring active biomass in granular activated carbon filters used in water treatment 38, 3971–3979. Matilainen, A., Gjessing, E.T., Lahtinen, T., Hed, L., Bhatnagar, A., Sillanpää, M., 2011. An overview of the methods used in the characterisation of natural organic matter (NOM) in relation to drinking water treatment. Chemosphere 83, 1431–1442. Matilainen, A., Sillanpää, M., 2010. Removal of natural organic matter from drinking water by advanced oxidation processes. Chemosphere 80, 351–365. Matilainen, A., Vepsäläinen, M., Sillanpää, M., 2010. Natural organic matter removal by coagulation during drinking water treatment: A review. Adv. Colloid Interface Sci. 159, 189–197. Ness, A., Boyer, T.H., 2017. Pilot-Scale Evaluation of Bicarbonate-Form Anion Exchange for DOC Removal in Small Systems. J. Am. Water Works Assoc. 109, 513–513.   93 Purolite, 2014. Cleaning methods for fouled ion exchange resins-application guide. Pürschel, M., 2014. Uptake of natural organic matter (NOM) fractions by anion exchangers in demineralisation and drinking water plants. Rahmani, S., Mohseni, M., 2017. The role of hydrophobic properties in ion exchange removal of organic compounds from water. Can. J. Chem. Eng. 95, 1449–1455. Rokicki, C.A., Boyer, T.H., 2011. Bicarbonate-form anion exchange: Affinity, regeneration, and stoichiometry. Water Res. 45, 1329–1337. Ross, P.S., van der Aa, L.T.J., van Dijk, T., Rietveld, L.C., 2018. Effects of water quality changes on performance of biological activated carbon (BAC) filtration. Sep. Purif. Technol. 212, 676–683. Sarathy, S.R., Mohseni, M., 2007. the impact of UV:H2O2 advanced oxidation on molecular size distribution of chromophoric natural organic matter. Environ. Sci. Technol. 41, 8315–8320. Schippers, D., Kooi, M., Sjoerdsma, P., Bruijn, F. de, 2004. Colour removal by ion exchange and reuse of regenerant by means of nanofiltration. Water Sci.  Technol. Water Supply 4, 57–64. Schulz, M., Winter, J., Wray, H., Barbeau, B., Bérubé, P., 2017. Biologically active ion exchange ( BIEX ) for NOM removal and membrane fouling prevention. Water Sci. Technol. Water Supply 17, 1178–1184. Servais, P., Anzil, A., Ventresque, C., A. Anzil, Ventresque, C., Anzil, A., Ventresque, C., 1989. Simple method for determination of biodegradable dissolved organic carbon in water. Appl. Environ. Microbiol. 55, 2732–2734. Sharbatmaleki, M., 2010. Investigation of potential pathways and multi-cycle bioregeneration of ion-exchange resin laden with perchlorate. Sillanpää, M., 2015. Natural organic matter in water,characterization and treatment methods. Singer, P.C., 2006. Disinfection byproducts in drinking water : additional science and policy considerations in the pursuit of public health protection. Am. Water Work. Assoc. 98, 73–80. Stevenson, F.J., 1994. Humus chemistry genesis, composition, reactions, 2nd ed, Willey Interscience. New York. Summers, R.S., Hooper, S.M., Shukairy, H.M., Solarik, G., Owen, D., 1996. Assessing DBP   94 yield: uniform formation conditions. Am. Water Work. Assoc. 88, 80–93. Tan, Y., Kilduff, J.E., 2007. Factors affecting selectivity during dissolved organic matter removal by anion-exchange resins. Water Res. 41, 4211–4221. Tan, Y., Kilduff, J.E., Kitis, M., Karanfil, T., 2005. Dissolved organic matter removal and disinfection byproduct formation control using ion exchange. Desalination 176, 189–200. Thornhill, S.G., Kumar, M., 2018. Biological filters and their use in potable water filtration systems in spaceflight conditions. Life Sci. Sp. Res. 17, 40–43. Thurman, E.M., 1985. Organic Geochemistry of Natural Waters, Usgs. Denver. USEPA, 1974. Method 310 . 2 : Alkalinity ( Colorimetric , Automated , Methyl Orange ). van der Kooij, D., 1992. Assimilable organic carbon as an indicator of bacterial regrowth. Am. Water Work. Assoc. 84, 57–65. Wachinski, A.M., 1997. Environmental Ion Exchange. Lewis Publishers. Wang, J.Z., Summers, R.S., Miltner, R.J., Wang, J.Z., Summers, R.S., Miltner, R.J., 1995. Biofiltration performance part 1, relationship to biouss. Am. water Work. Assoc. 87, 55–63. Weishaar, J.L., Aiken, G.R., Bergamaschi, B.A., Fram, M.S., Fujii, R., Mopper, K., 2003. Evaluation of specific ultraviolet absorbance as an indicator of the chemical composition and reactivity of dissolved organic carbon. Environ. Sci. Technol. 37, 4702–4708. Wert, E.C., Neemann, J.J., Rexing, D.J., Zegers, R.E., 2008. Biofiltration for removal of BOM and residual ammonia following control of bromate formation. Water Res. 42, 372–378. Williams, M.D., Pirbazari, M., 2007. Membrane bioreactor process for removing biodegradable organic matter from water. Water Res. 41, 3880–3893. Winter, J., Wray, H.E., Schulz, M., Vortisch, R., Barbeau, B., Bérubé, P.R., 2018. The impact of loading approach and biological activity on NOM removal by ion exchange resins. Water Res. 134, 301–310. Wray, H., Schulz, M., Winter, J., Bérubé, P.R., Barbeau, B., Bayless, W., 2016. Biological ion exchange for nom removal – unforeseen synergies, in: american water works association-Water Quality Technology Conference (WQTC) 2016 Nov., 13-17. Indianapolis,IN,USA. Yang, T., Doudrick, K., Westerhoff, P., 2013. Photocatalytic reduction of nitrate using titanium dioxide for regeneration of ion exchange brine. Water Res. 47, 1299–1307.     95 Appendices Appendix A Supplementary results related to the chapters 5 and 6.  A1. Biodegradable organic carbon (BDOC) removal in pilot study  Achieved results confirm that BIEX columns removed BDOC effectively. Figure A-1 shows the effect of EBCT on BDOC removal, after 144 and 277 days of system operation. According to the results, the EBCT of 7.5 minutes would be effective at removing BDOC and further treatment duration will not impact the outcomes significantly. Further experiments required to make a strong conclusion.  Figure A-1. BDOC concentration in water treated at different EBCTs of 0, 7.5 and 30 min after 144 and 277 days of operation.  A2. High performance size exclusion chromatography (HPSEC)- pilot study   96 Samples collected from inlet and treated water at different EBCTs (7.5 and 30 min), after 277 days were analyzed for different fractions of NOM. Figure A-2 shows the NOM fractions removed after 277 days of operation. It can be seen that the filter was showing significant removal of higher molecular weight NOM. As there was no chloride release observed at this time, mechanisms other than primary ion exchange were likely occurring.   Figure A-2. Size exclusion chromatography for raw and treated water at two different EBCTs of 7.5 and 30 minutes after 277 days of operation.  A3. Temperature Water temperature inside the columns was measured a few times over the course of pilot study. Figure A-3 shows that the temperature decreased significantly from November to January due to the seasonal temperature drop and also a malfunction of the building heating system. Figure A-3 also provides the corresponding NOM removal. Except the first two points that have roughly the same temperature, in other points the temperature and removal changes followed the same trends.   97 However, other factors should be considered in these trends and it was difficult to draw a conclusion based on these few data points.  Figure A-3. Changes in temperature and NOM removal over the course of experiments vs. date.   A4. Turbidity Turbidity was monitored for the inlet and treated samples at 7.5 and 30 minutes of EBCT. Table A-1 displays different turbidity values at different depth of the column or different EBCT. According to the results, the system performed effectively in decreasing turbidity after 30 minutes EBCT; however, 7.5 minutes is not sufficient for removing NOM. Data reports at 95% confidence interval of the observed turbidities. Figure A-4 displays different turbidity values at different depth of the column or different EBCT. According to the results, the system performed effectively in decreasing turbidity after 30 minutes EBCT; however, 7.5 minutes is not sufficient for removing NOM.    98 Table A-1. Turbidity values in inlet and outlets of 7.5 and 30 min EBCT. Sample Turbidity (NTU) Number of samples Inlet (EBCT= 0 min) 1.45±0.37 38 Treated water (EBCT= 7.5 min) 1.25±0.53 29 Treated water (EBCT= 30 min) 0.43±0.07 59   Figure A-4. Turbidity values at different EBCTs. (values= result of samples taken over 277 days of operation).  A5. Backwash-CSTR-Middle River To monitor the efficiency of the backwash procedure, turbidity of the outlet over time was measured. Figure A-5 shows that backwash turbidity fluctuated significantly and after backwashing it dropped to 0.5 NTU and stayed the same level in both columns. This information would be useful for designing a full scale system. Due to the biofilm growth on the surface of resin, longer time may be required to reach the normal turbidity in the system.   99  Figure A-5.Turbidity values over the backwash procedure.  A6. Different NOM fractions removal-PBR-Jericho Pond Figure A-6 shows the HPSEC chromatogram of the samples taken from the inlet and outlet water at different BVs (400, 2000 and 5000 BV) in biotic or abiotic conditions. Sodium azide in samples taken from abiotic conditions showed a specific peak at around 16 minutes. Looking at the changes of the apparent molecular weight distribution, no difference was observed in all treated water except an area around 16 minutes elution time related to sample taken from abiotic system at 5,000 BV. The difference between treated water after 5,000 BV in both conditions confirmed that the specific fraction of NOM has not been removed completely and according to the calibration, this fraction was the lower apparent molecular weight fraction of NOM. Comparing these results with the samples taken at 5,000 BV in biotic system, it can be concluded that biodegradation was responsible for removal of this specific range of NOM molecular weight. However, HPSEC analysis is entirely dependent on the UV absorbance of the organic molecules at 260 nm, so biodegradable organic species (i.e., non-aromatic & non-conjugated double bond   100 compounds) may not have been detected precisely. Further, the changes in biodegradable organic species concentration are in very low range (i.e., ranging between 20-40 µg/L) be able to observe significant differences.  Figure A-6. HPSEC chromatograms of raw and treated water in biotic and abiotic conditions at different bed volumes.     101 Appendix B The Box Behnken design (BBD) model was used to study the main effects and the interaction effects between operational parameters in ion exchange process using CSTR. The BBD model based on RSM is efficient and requires a minimum number of experiment runs. Table B-2 shows the parameters including flowrate, resin dose and mixer speed and their selected levels. Table B-3 presents the results of the experimental matrix, which contains 15 experiments using BBD. In each test, the NOM removal after 1000 BV was chosen as the output response variable. Figures B7, B8 and B9 demonstrate the individual effect of each parameters on NOM removal.  No interaction was found between the effects of the two variables on any of the considered parameters. To determine the effect of each factor and interactions of two factors, three-dimensional response surface plots have been constructed using the Design Expert 7.0.0 software based on model function (Figures B10, B11 and B12). As a result of very narrow ranges of parameter’s value, no interaction was seen among the selected factors and a linear relation was seen among them (Eq. 13). Table B-2. Selected parameters and their levels used for designing of experiments. Parameters Level 1 Level 2 Level 3 Flowrate (ml/min) 2 6 10 Resin dose (ml) 5 12.5 20 Mixer speed (rpm) 100 200 300      102 Table B-3. Design of experiments table including conditions of 15 runs.  Run # Flow rate (ml/min) Resin dose (ml) Mixer speed (rpm) Response (Removal %) 1 10 12.5 100 47.2 2 2 20 200 57.3 3 10 12.5 300 49.3 4 6 20 100 53.4 5 6 12.5 200 54 6 10 20 200 50.5 7 6 12.5 200 54.5 8 2 12.5 100 57.2 9 6 20 300 56.8 10 10 5 200 43.4 11 2 5 200 57.5 12 6 5 300 49.5 13 2 12.5 300 57.3 14 6 12.5 200 55.2 15 6 5 100 41.3  Removal (%) = 50.65 -1.22 ×	flowrate + 0.44 ×	Resin dose + 0.17 ×	Mixer speed (13)     103 B1. Effect of inlet flowrate  Figure B-7. Inlet flowrate impact on NOM removal efficiency in CSTR.  B2. Effect of resin dose  Figure B-8. Resin dose impact on NOM removal efficiency in CSTR.      104 Effect of mixer speed  Figure B-9. Mixer speed impact on NOM removal efficiency in CSTR.  B3. Flowrate-Mixer speed interactions  Figure B-10. Flowrate and mixer speed impacts on NOM removal.    105 B4. Resin dose- Mixer speed interactions  Figure B-11. Resin dose and mixer speed impacts on NOM removal. B5. Flowrate-Resin dose interactions  Figure B-12. Resin dose and flowrate impacts on NOM removal. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0379511/manifest

Comment

Related Items