1 THE REGENERATION AND FOULING OF BICARBONATE FORM ANION EXCHANGE RESINS By CHRISTOPHER ADAM ROKICKI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FO R THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013
2 2013 Christopher Adam Rokicki
3 To my mom, without her this would not have been possible
4 ACKNOWLEDGMENTS I must thank the National Sci ence Foundation ( Grant No.1209962 ) the Australian Academy of Science and the Florida Department of Environmental Protection for their generous funding and support. I need to thank Mary Drikas, Jim Morran and everyone at the Australian Water Quality Centr e that helped me while I was abroad. Everyone in my research team has been exceptionally support ive and help ful, without them my time at UF would have been much more difficult I need to also thank g uiding me to answering my questions led me here S pecial thanks go to Dr. Paul Chadik, Dr. Eric McLamore, and Dr. David Mazyck for their encouragement and help their expertise has helped developed my skills and greatly benefitted the work presented here In particular I must thank Dr. Treavor Boyer for all he did to help me complete this dissertation I learned so much from Treavor over the years and I am sure I will continue to learn more from him Last but not least I need to thank my family and friend s, in particular my mother Ewa for believing in me and making sacrifices for me my entire life; and Tori, without her constant support and encouragement finishing this dissertation would have been near impossible
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF ABBREVIATIONS ................................ ................................ ........................... 11 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 14 1.1 Natural Organic Matter ................................ ................................ ...................... 14 1.2 Anion Exchange and Regeneration ................................ ................................ .. 15 1.3 Fouling Mechanisms ................................ ................................ ......................... 16 1.3.1 Inorganic Fouling with Bicarbonate ................................ .......................... 16 1.3.2 Biological Fouling ................................ ................................ .................... 17 1.4 Research Objectives ................................ ................................ ......................... 17 1.5 Organization of Dis sertation ................................ ................................ .............. 18 2 BICABONATE FORM ANION EXCHANGE: AFFINITY, REGENERATION, AND STOICHIOMETRY ................................ ................................ ................................ .. 19 2.1 Background ................................ ................................ ................................ ....... 19 2.2 Experimental Section ................................ ................................ ........................ 22 2.2.1 Materials ................................ ................................ ................................ .. 22 2.2.2 Jar Tests and Regeneration ................................ ................................ .... 23 2.2.3 CO 2 Regeneration ................................ ................................ ................... 25 2.2.4 Titration of NOM ................................ ................................ ...................... 26 2.2.5 Analytical Me thods ................................ ................................ .................. 27 2.3 Results and Discussion ................................ ................................ ..................... 28 2.3.1 Affinity ................................ ................................ ................................ ...... 28 2.3.2 Rege neration ................................ ................................ ........................... 30 126.96.36.199 Salt regeneration ................................ ................................ ............ 30 188.8.131.52 CO2 regeneration ................................ ................................ ........... 32 2.3.3 Stoichiometry ................................ ................................ ........................... 33 2.4 Chapter Summary ................................ ................................ ............................. 37 3 IMPACT OF REGENERATION ON FRACTIONATION OF NOM IN MIEX TREATED SOUTH AUSTRALIAN WATERS ................................ .......................... 43 3.1 Background ................................ ................................ ................................ ....... 43
6 3.2 Materials and Methods ................................ ................................ ...................... 47 3.2.1 Trea tment Plants and Raw Waters ................................ .......................... 47 3.2.2 Anion Exchange Resins ................................ ................................ .......... 48 3.2.3 Multiple Loading Ion Exchange Jar Tests ................................ ................ 49 3.2.4 Resin Regeneration ................................ ................................ ................. 49 3.2.5 Analytical Methods ................................ ................................ .................. 51 3.3 Results and Discussion ................................ ................................ ..................... 51 3.3.1 Impact of regeneration agent on DOC removal ................................ ....... 51 3.3.2 Impact of Regeneration Agent on NOM Fractionation ............................. 55 3.3.3 Comparison of Bench scale to Full scale Systems ................................ 63 3.4 Chapter Summary ................................ ................................ ............................. 64 3.5 A cknowledgments ................................ ................................ ............................. 65 4 THE ROLE OF DIVALENT CATIONS IN THE FOULING OF ANION EXCHANGE RESINS ................................ ................................ ............................. 74 4.1 Background ................................ ................................ ................................ ....... 74 4.2 Materials and Methods ................................ ................................ ...................... 77 4.2.1 Materials ................................ ................................ ................................ .. 77 4.2.2 Experimental methods ................................ ................................ ............. 78 4.2.3 Ion Exchange Regeneration ................................ ................................ .... 78 4.2.4 Ion Exchange Treatment ................................ ................................ ......... 78 4.2.5 Ana lytical Methods ................................ ................................ .................. 79 4.2.6 Visual MINTEQ ................................ ................................ ........................ 80 4.3 Results and Discussion ................................ ................................ ..................... 81 4.3.1 Effect of Divalent Metal Cations on Target Anion Removal ..................... 81 184.108.40.206 DOC removal ................................ ................................ ................. 81 220.127.116.11 Divalent metal cation remova l ................................ ........................ 83 4.3.2 Effect of Divalent Metal Cations on Target Anion Removal in the Absence of NOM ................................ ................................ ........................... 85 18.104.22.168 Nitrate removal ................................ ................................ ............... 85 22.214.171.124 Divalent metal cation removal ................................ ........................ 88 126.96.36.199 Chemical speciation modeling ................................ ....................... 90 4.4 Chapter Summary ................................ ................................ ............................. 91 5 DETERMINATION OF BIOLOGICAL FOULING ON ANION EXCHANGE USED IN WASTEWATER TREATMENT ................................ ................................ ......... 103 5.1 Background ................................ ................................ ................................ ..... 103 5.2 Materials and Methods ................................ ................................ .................... 106 5.2.1 Resin ................................ ................................ ................................ ..... 106 5.2.2 Water Sour ce ................................ ................................ ........................ 107 5.2.3 Multiple Loading Resin Exhaustion ................................ ........................ 107 5.2.4 Biofilm Analysis ................................ ................................ ..................... 108 5.3 Results and Discussion ................................ ................................ ................... 108 5.3.1 Evaluation of biofilm growth ................................ ................................ ... 108 5.3.2 Follow up Work ................................ ................................ ..................... 110
7 5. 4 Chapter Summary ................................ ................................ ........................... 112 6 CONCLUSIONS ................................ ................................ ................................ ... 115 6.1 Conclusions of D issertation ................................ ................................ ............ 115 6.2 Implications of Findings ................................ ................................ .................. 115 APPENDIX A S UPPORTING INFORMATION FOR CHAPTER 3 ................................ .............. 117 B SUPPORTING INFORMATION FOR CHAPTER 4 ................................ .............. 118 LIST OF REFERENCES ................................ ................................ ............................. 120 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 129
8 L IST OF TABLES Table page 2 1 Average Measured Composition of Synthetic Model Waters used in Ion Exchange Experiments ................................ ................................ ....................... 38 3 1 Characteristics of raw water sources ................................ ................................ .. 66 3 2 ................................ ............................. 67 3 3 Change in DOC for full scale and lab oratory scale treatments ........................... 68 4 1 Solubility products of likely precipitates. ................................ ........................... 100 4 2 Theoretical model water compositions ................................ ............................. 101 4 3 Synthetic model water compositions for waters using nitrate in lieu of NOM, with varying divalent cations. Used for Visual MINTEQ input. .......................... 101 4 4 Results of Visual MINTEQ Simulations ................................ ............................ 102 A 1 ................................ ........................... 117
9 LIST OF FIGURES Figure page 2 1 Effect of mobile counter ion on affinity of MIEX resin using synthetic inorganic model water (C/C 0 = 4.6 for Cl ). ................................ ................................ ........ 38 2 2 Comparison of bicarbo nate form and chloride form MIEX resin in the absence and presence of SFR NOM. ................................ ................................ 39 2 3 Effect of multiple regeneration cycles on the removal efficiency of bicarbonate form MIEX resin using synth etic inorganic model water ................. 39 2 4 Effect of multiple regeneration cycles on the removal efficiencies of MIEX resin using synthetic NOM containing model water. ................................ ........... 40 2 5 Generation of bicarbonate form MIEX resin by CO 2 gas sparging. .................... 40 2 6 Charge density of SFR NOM derived from titration. ................................ ........... 41 2 7 Stoichiometry of various forms of MIEX resin using synthetic inorganic model water. All results are experimental data except calc. R2 CO3, which was calculated using R HCO3 data. ................................ ................................ .......... 41 2 8 Effect of regeneration on the stoichiometry of bicarbonate form and chloride form MIEX resin using synthetic NOM containing model water .......................... 42 3 1 Effect of increasing bed volume loading on DOC removal in Mount Pleasant water by resin for the first and second regeneration s cycles .............................. 69 3 2 Effect of increasing bed volume loading on DOC removal in Middle River water by resin for the first and second regeneration cycles. ............................... 70 3 3 Molecular weight distribution by HPSEC for resin and waters from Mount Pleasant WTP. ................................ ................................ ................................ .... 71 3 4 Molecular weight distribution by HPSEC for resin and waters from Middle River WTP. ................................ ................................ ................................ ......... 72 3 5 Molecular weight distribution by HPSEC of regenerant brine ............................. 73 4 1 Kinetic plots showing percent DOC removal over time for each resin in either the chloride or bicarbonate form in the presence of either: Mg, Ca, or Cd in the NOM containing water. ................................ ................................ ................. 92 4 2 Absolute magnesium concentration for each resin in the synthetic water containing NOM and magnesium ................................ ................................ ....... 93
10 4 3 Absolute calcium concentration for each resin in the synthetic water containing NOM and calcium ................................ ................................ .............. 94 4 4 Absolute cadmium concentration for each resin in the synt hetic water containing NOM and c admium ................................ ................................ ........... 95 4 5 Kinetic plots showing percent nitrate removal over time for each resin in either the chloride or bicarbonate form in the presence of either: Mg, Ca, or Cd in the nitrate containing water. ................................ ................................ ...... 96 4 6 Absolute magnesium concentration for each resin in the synthetic water containing nitrate and magnesium ................................ ................................ ...... 97 4 7 Absolute calcium concentration for each res in in the synthetic water containing nitrate and calcium ................................ ................................ ............ 98 4 8 Absolute cadmium concentration for each resin in the synthetic water containing nitrate and cadmium ................................ ................................ .......... 99 5 1 Total particle counts for resins with allowed biofilm growth for 5 days. ............. 113 5 2 Total particle counts for resins with allowed biofilm growth for 7 days ............. 113 5 3 Total particle counts for resins with allowed biofilm growth for 5 days in disinfected water, or filtered water. ................................ ................................ ... 114 B 1 Kinetic isotherm showing pH change for each resin in either the chloride or bicarbonate form in the presence of either: Mg, Ca, or Cd for waters using NOM as the targeted anion for removal ................................ ............................ 118 B 2 Kinetic isotherm showing pH change for each resin in either the chloride or bicarbonate form in the presence of either: Mg, Ca, or Cd for waters using nitrate as the targeted anion for removal ................................ .......................... 119
11 LIST OF ABBREVIATI ONS ACS American Chemical Society AMW Apparent molecular weight BV Bed Volumes Da Daltons DBP Disinfection by products DI Deionized water DOC Dissolved organic carbon HPSEC High performance size exclusion chromatography. IC Ion chromatography IEX Ion exch ange MIEX Magnetic ion exchange resin Mn Number average of molecular weight fractions Mw Weight average of molecular weight fractions NOM Natural organic matter NTU Nephelometric turbidity units SFR NOM Santa F e R iver NOM isolate S R NOM Suwannee R iver NOM Isolate SUVA Specific Ultraviolet Absorbance TOC Total organic carbon UV 254 Ultra violet light absorbance at 254nm WTP Water treatment plant polydispersity
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE REGENERATION AND FOULING OF BICARBONATE FORM ANI ON EXCHANGE RESINS By Christopher Adam Rokicki August 2013 Chair: Treavor H. Boyer Major: Environmental Engineering Sciences Anion exchange is a n increasingly utilized method for water treatment Growing water demands and decreasing water quality requir e th e increased removal of natural organic matter (NOM) from water NOM is a n important constituent in the disinfection stage of potable water treatment due to formation of disinfection byproducts (DBPs), with DBPs being of human health concern A nion exch ange is appealing because it is possible to regenerate and reuse the resin s The drawback of anion exchange is that it generates high salinity waste with high concentrations of organic matter, which is difficult to treat or dispose of. Utilizing resin in t he bicarbonate form would bypass the production of a high NaCl concentration waste, which would be easier to dispose of. To fully understand the potential of anion exchange process for NOM removal bicarbonate regeneration and how it impacts resin fouling was investigated. In this work m agnetic i on e xchange (MIEX) resin was used for anion exchange experiments Series of jar test experiments, run in single loading or multiple loading configurations were systematically utilized to investigate the impact that resin mobile ion had on resin affinity, organic matter removal, and fouling. It was confirmed that b icarbonate had a similar affinity for the resin as chloride Multiple regeneration cycles
13 of the resin found that the strength of the regeneration solution governed the ion exchange process, rather than the identity of the mobile ion and that certain organic fractions particularly within the 1000 1500 Da AMW fraction were remaining on the resin causing the resin to lose ion exchange capacity D ivalent cation s were added to model waters to determine how solubility of carbonate minerals impact ed fouling and it was found that while magnesium minerals are most soluble they tended to foul resin the mos t due to direct interference with the resin rather than precip itation on resin W astewater treatment applications were investigated to investigate biofouling potential and a method to quantify biofouling by particle counting was investigated and preliminary results showed mobile ions had minimal impact on biofouling of resins This dissertation provides an understanding on how the regeneration process can impact the performance and fouling of resin when used for NOM removal.
14 CHAPTER 1 I NTRODUCTION 1. 1 N atural O rganic M atter Removal of n atural organic matter (NOM) is an important aspect of water treatment, as high quality ground water supplies are becoming scarcer and more surface waters are being used for potable water supplies. It has been shown that interaction of chlorine in the disinfection stage of water tre atment with NOM lead s to the formation of DBPs, many of which are of human health concern NOM is a complex mixture o f organic compounds including, humic and fulvic acids that contain a variety of functional groups including carboxylic acid functional gro ups which can cause NOM to behave as an anion in solution (Collins et al. 1986). The functional groups and structure associated with NOM can vary depending on the source water, and as such can impact how the NOM reacts with divalent cations or ion exchange resins. Organic matter has been shown in previous studies to react with metal cations in aqueous solutions (Dudal and Gerard, 2004; Sharma et al., 2010) NOM is also known for its ability to interact and complex with trace elements, such as As, Pb or U th at may impact the toxicity of the organic matter (Iskrenova Tchoukova et al., 2010; Sharma et al., 2010) The formation of metal ligands and complexes can impact the NOM, having an effect on the mobility and aggregation ability of the NOM in solution. The presence of carboxylic acid groups within the structural composition of NOM is one of the most important interaction sites for metal and NOM interactions (Rey Castro et al., 2009) The nature of the specific relationships between different metal cations a re different depending on the metal cation and the specific nature of the NOM. Research has indicated that some metal ions, such as Ca 2+ bonds directly with NOM,
15 while other ions such as Mg 2+ impact the solubility of NOM by entering into hydration spheres of water (Ahn et al., 2008) It is known that divalent cations, such as Mg 2+ and Ca 2+ lead to increased membrane fouling both by precipitation (Yoon et al., 1998) but also by hydrolyzing with organic matter (Ahn et al., 2008) but there is limited work el ucidating the impact that divalent cations have on ion exchange resins themselves. The removal of NOM during water treatment has been shown to have an impact on many of the treatment techniques commonly used in water treatment (Krasner et al. 2006). One su ch technology that has been shown to be effective for the removal of NOM is the use of anion exchange (Boyer and Singer, 2006, 2008; Drikas et al., 2009; Hsu and Singer, 2010; Kitis et al., 2007; Mergen et al., 2008) 1.2 Anion Exchange and Regeneration I on exchange has been used as a means to remove dissolved organic carbon (DOC) which makes up a large proportion of NOM (Thurman and Malcolm, 1985) in water treatment systems. A wide variety of resin characteristics and properties allow for the deliberate targeting of water contaminants, with DOC being a particular contaminant of interest within this work due to the concerns related to NOM removal from potable water. As it is used now, anion exchange for the removal of NOM has been predominantly used in the chloride form (Boyer et al., 2008) The use of anion exchange resins for the removal of NOM from drinking waters has been thoroughly investigated (Boyer and Singer, 2006, 2008; Drikas et al., 2009; Hsu and Singer, 2010; Kitis et al., 2007; Mergen et al ., 2008) and systems can be tailored with different resins to specifically target NOM of different characteristics. Previous work has shown that
16 certain anion exchange resins have a greater affinity for NOM with a high charge density and high aromaticity ( Boyer et al., 2008) In addition to extensive academic investigations into anion exchange for NOM removal, there have been many pilot and full scale plants in operation (Boyer and Singer, 2006; Warton et al., 2007) The use of anion exchange is advantageou s when attempting to decrease DOC in natural waters (Hammann et al., 2004; Singer et al., 2009; Warton et al., 2007) It is common for most anion exchange systems treating n atural waters to utilize a resin saturated with chloride as the mobile counter ion. This is likely due to the low cost of sodium chloride and the relative inertness of chloride in water chemistry reactions. The drawback to the use of chloride is the use of concentrated salt brine to regenerate resin leads to a waste product high in salinity and organic matter, thus generating another waste stream to process. 1. 3 Fouling Mechanisms 1. 3 .1 Inorganic F ouling with B icarbonate The carbonate species are known t o react with many divalent metal cations. Among those are Ca 2+ and Mg 2+ which will be the primary focus of this work. In traditional water treatment systems, it is well known that hardness, in the form of calcium and magnesium, has been known to react wit h bicarbonate in waters and leads to clogging of filters (Letterman, 1999) and fouling of membranes (Yoon et al., 1998) It is due to the nature of the precipitation potential of divalent cations that it is necessary to determine how the interactions with bicarbonate may adversely affect anion exchange.
17 1. 3 .2 Biological F ouling The issue of biofouling on ion exchange resins used in water treatment has been acknowledged for a long time (Flemming, 1987) The problem associated with biofouling of ion exchange resin is the fact that a biofilm may limit the ion exchange capacity of the resins and more importantly the microbes present may pose health risks if waters are not properly disinfected after ion exchange. Bacteria tend to accumulate on anion exchange res ins because bacteria tend to be slightly negatively charged (Flemming, 1987) The combination of a growing medium and a carbon source (NOM or bicarbonate) make it so that ion exchange resins are ideal growing medi a for bacteria. It was found that the bacte rial growth associated with ion exchange was directly correlated with the frequency of use (Eisman et al., 1949) .The resin beds that used in the Eisman study were susceptible to bacterial growth, which supports the argument that ion exchange resins may ser ve as a growth medium. 1. 4 Research Objectives The goal of this research was to improve the understanding of the regeneration of an ion exchange and to gain a better understanding of how th e fouling of the bicarbonate form of the resin is impact ed by div alent cations and biological growth This work was completed through the use of bench scale experiments using synthetic waters, raw drinking waters, and waters taken from wastewater treatment systems. Synthetic waters were created using NOM isolates and sa lts, or metal oxides The objectives of this dissertation were: (1 ) d etermine whether bicarbonate form of the resin was a viable alternative to chloride form resins in regards to the stoichiometry and affinity of the ion exchange reaction; (2) evaluate the regeneration and how the
18 regeneration technique impacts the targeted organic matter fraction the resin removes; (3) investigate the mechanisms associated with inorganic fouling of bicarbonate form resins; and (4) develop a means to quantify the extent of biofouling on the resin and determine if the resin mobile ion impacts the exten t of biofouling. 1. 5 O rganization of Dissertation The dissertation is organized into 6 different chapters in various stages of publication and preparation. Chapters 1 and 6 are the Introduction and Conclusion which unite and tie together the remaining chapters. Chapter 2 : Bicarbonate form anion exchange: Affinity, Regeneration, and Stoichiometry is based on work that was published in Water Research Chapter 3 : Impact of regenera tion on fractionation of NOM in MIEX treated South Australian waters is work that has been completed and submitted to Water Research, but has not yet been accepted for publication. Chapter 4 : The role of divalent cations in the fouling of anion exchange re sins is in the final stages of completion and is intended to be submitted to Reactive & Functional Polymers. Chapter 5 : Determination of biological fouling on anion exchange used in wastewater treatment is the foundation of work that will be continued by o ther students and submitted sometime shortly thereafter to Water Science and Technology
19 CHAPTER 2 B ICABONATE FORM ANION EXCHANGE : AFFINITY, REGENERATION AND STOICHIOMETRY 2.1 Background Magnetic ion exchange (MIEX) is an effective process for trea ting water that contains dissolved organic carbon (DOC) and other pollutants (Boyer and Singer 2005, Boyer and Singer 2006, Kitis et al. 2007, Mergen et al. 2008, Drikas et al. 2009 Hsu and Singer 2010 ). MIEX resin and most conventional anion exchange res ins use chloride as the mobile counter ion, and are referred to as chloride form resin. Chloride is used as the mobile counter ion because it is relatively inert with respect to water chemistry reactions and the low cost of sodium chloride salt for regener ation. Chloride, however, is not ideal in every situation and can have several adverse outcomes, such as enhanced corrosion of plumbing, detrimental effects on biological wastewater treatment processes, and increased salinity loading to surface waters. Re search has shown that an increase in the chloride to sulfate mass ratio (CSMR) can increase lead corrosion in water distribution systems ( Edwards and Triantafyllidou 2007 ). Chloride form anion exchange resin is expected to increase the CSMR due to release of chloride and uptake of sulfate, and thus has the potential to increase the corrosivity of treated water towards lead. Chloride form resin also presents potential problems related to the regeneration process, whereby a high ionic strength sodium chloride solution (i.e., brine) is used to regenerate the resin. Waste brine is often Reproduced with permission from Rokicki, C.A., Boyer, T.H. 2011 Bicarboante form anion exchange: Affinitey, regeneration, and stoichiometry. Water Research 45(3), 1329 1337, DOI: 10.1016/j.watres.2010.10.018. Copyright 2010 Elsevier.
20 difficult to dispose of due to the detrimental effects of high salinity on biological wastewater treatment processes ( Panswad and Anan 1999 ). Furthermore, wastewater effluent tha t receives waste brine is a threat to receiving waters because the increased loading of salt can degrade water quality over time. Altering the mobile counter ion on anion exchange resin to a non chloride ion has the potential to negate the problems associ ated with chloride form anion exchange discussed above. Investigations into the performance of non chloride anion exchange are limited, and focus mostly on hydroxide form resin and, to a lesser extent, bicarbonate form resin (Hll and Kiehling 1981). Bicar bonate is an attractive alternative to chloride because it is harmless to humans and ecosystems, even when present at elevated levels ( Jelinek et al. 2004 ). Bicarbonate form anion exchange resin is expected to have similar treatment efficiency as chloride form resin based on the equal preference of chloride form resin for bicarbonate and chloride (Boyer and Singer 2008). This is supported by the work of Hll and Kiehling ( 1981), Matosic et al. (2000), and Jelinek et al. (2004), which shows bicarbonate form anion exchange can effectively treat water for the removal of nitrate and sulfate. In addition, bicarbonate form anion exchange has been shown to decrease the corrosivity of treated water ( Takasaki and Yamada 2007 ). This is also seen in the Langelier and L arson Skold indices, which consider bicarbonate to inhibit corrosion in iron containing water distribution systems (Larson 1975). The improved treatment outcomes of bicarbonate form anion exchange relative to chloride form anion exchange are treated water with lower corrosivity and waste brine that does not add chloride to engineered or natural systems. Finally, as demonstrated by Hll and Kiehling (1981), bicarbonate form resin has the potential to
21 be regenerated by sparging water with carbon dioxide (CO 2 ) gas, which would generate aqueous bicarbonate and would reduce the need for salts used during regeneration. Greenleaf and SenGupta (2009) demonstrated a similar idea in which CO 2 sparged water generated hydrogen ions that were used to regenerate calcium l oaded cation exchange resin. The potential of CO 2 sparging for regeneration of anion exchange resin can ultimately lead to a water treatment process that uses less chemicals and sequesters waste CO 2 gas for a beneficial purpose ( Greenleaf and SenGupta 2009 ). Regeneration of anion exchange resin with CO 2 gas also points to the synergies of co locating power generation and water treatment, both of which are major users of water and energy ( Carrillo and Frei 2009 Blackhurst et al. 2010). Despite the appeal of bicarbonate form anion exchange, the depth of knowledge about non chloride anion exchange is rather limited. There has been a growing interest in non chloride anion exchange which has caused an increase in the number of pilot studies and investigations in to the performance of bicarbonate form MIEX treatment (Dahlke et al. 2007). Selectivity of resins in the chloride form for various inorganic and organic anions is well established ( Li and Sengupta 1998 Tripp and Clifford 2006 Tan and Kilduff 2007 ). Howev er, it is unknown if using non chloride mobile counter ions will alter the preference of a resin for other target anions in the presence of DOC. For example, there is a limited amount of research testing bicarbonate form resin for nitrate removal ( Jelinek et al. 2004 Matosic et al. 2000, H ll and Kiehling 1981 ); there is limited published work testing bicarbonate form resin for DOC removal (Dahlke et al. 2007). This is important because anion exchange, in particular MIEX, is increasingly being used for DOC removal ( Warton et al. 2007 Neale and Schafer 2009, Singer et al.
22 2009). Other major unknowns related to non chloride anion exchange include treatment efficiency over multiple regeneration cycles; whether it is possible to convert an anion exchange resin to the bicarbonate form by sparging CO 2 gas in water, without the addition of chemicals; and the precise mechanism of bicarbonate form anion exchange. The overall goal of this work was to increase the depth of knowledge pertaining to bicarbonate from ani on exchange. First, the affinity of non chloride form anion exchange resin was evaluated by using resins loaded with various mobile counter ions. The affinity was tested using two synthetic model waters. Both model waters had the same composition of chlori de, bicarbonate, nitrate, and sulfate. One of the model waters also contained a natural organic matter (NOM) isolate. Second, the continued reuse of bicarbonate form resin was investigated to determine the effectiveness of the regeneration process. Experim ents were conducted using both the synthetic inorganic model water and the synthetic NOM containing model water. Third, the use of CO 2 gas alone as a means of producing bicarbonate form resin was examined. Finally, the stoichiometry of bicarbonate form ani on exchange process was quantified to ensure that the process occurring was in fact ion exchange and not adsorption or complex formation. 2.2 Experimental Section 2.2.1 Materials MIEX resin from Orica Watercare was used in this work and is referred to as MIEX Cl because chloride is the mobile counter ion. All data for MIEX Cl resin is virgin resin as received from the manufacturer, which is the generally accepted practice with MIEX Cl resin. MIEX Cl resin was regenerated, as described in Section 2.2, to o btain
23 resins with other mobile counter ions. The nomenclature used in this paper is MIEX Z or R Z, where R is the MIEX resin phase and Z is the mobile counter ion placed on the resin during regeneration. For example, MIEX HCO 3 refers to resin with bicarbon ate as the mobile counter ion. MIEX resin was also saturated with a pre determined mobile counter ion, to simulate exhaustion, and then regenerated with a different mobile counter ion. The nomenclature used for this resin is MIEX (Y) Z, where Y is the coun ter ion used to saturate the resin and Z is the counter ion used to regenerate the resin. For example, MIEX (SO 4 ) HCO 3 refers to resin that was saturated with sulfate and then regenerated with bicarbonate. A subscript 2 after MIEX is used to show that diva lent anions occupy two sites on the resin (e.g., MIEX 2 SO 4 ). Two synthetic model waters were used in this work: (i) inorganic ions only and (ii) inorganic ions with NOM. The composition of the synthetic model waters is listed in Table 1. The model waters w ere prepared in deionized (DI) water. Chloride, bicarbonate, nitrate, and sulfate were added as sodium salts and were ACS reagent grade purity. The NOM was a freeze dried sample that was isolated from the Santa Fe River (SFR), FL, USA by Davis and co worke rs ( Davis et al. 1999 ). 2.2.2 Jar T ests and R egeneration All experiments were conducted at ambient laboratory temperature and open to the atmosphere, unless noted otherwise. Jar tests were performed using a Phipps & Bird PB 700 jar tester with 2 L square jars. Two liters of synthetic model water was used in all jar tests. MIEX resin was measured as the volume of wet settled resin in a graduated cylinder. MIEX resin doses of 1, 2, and 4 mL /L were tested. The jar tests were conducted at 100 rpm for 30 min an d the resin was allowed to settle for 10 min
24 before samples were taken from a spigot in the jar. Each MIEX resin (e.g., MIEX Cl, MIEX HCO 3 MIEX NO 3 etc.) was tested in duplicate or triplicate doses using the synthetic inorganic model water. The complete jar tests for MIEX HCO 3 and MIEX (SO 4 ) HCO 3 resins, using the synthetic inorganic model water, were replicated. No replicates were tested for the MIEX resins using the synthetic NOM containing model water because of a limited supply of NOM isolate. MIEX r esin was regenerated to alter the mobile counter ion available for exchange on the resin. Sodium salts of chloride, bicarbonate, nitrate, and sulfate were added to DI water to prepare the regeneration solution. The equivalent capacity of MIEX Cl resin, whi ch was previously determined to be 0.52 milliequivalents (meq) per mL resin ( Boyer and Singer 2008 ), was used to calculate the concentration of mobile counter ion in the regeneration solution. Regeneration experiments were conducted using regeneration solu tions that contained 10 and 100 times the equivalent capacity of the resin (i.e., 10 and 100). Apell and Boyer (2010) used a similar approach to investigate regeneration of magnetic cation exchange resin. As an example of a 10 regeneration solution, 4 m L of MIEX Cl resin was regenerated in 1 L of salt solution with an equivalent strength of 20.8 meq/L, which required 20.8 mmol/L of bicarbonate to produce MIEX HCO 3 resin or 10.4 mmol/L of sulfate to produce MIEX 2 SO 4 resin. The resin was mixed in the rege neration solution using a magnetic stir plate and stir bar for 30 min, and settled for 10 min before the supernatant liquid was decanted. To ensure that there was no excess salt remaining in solution, the resin was washed with 1 L of DI water, by mixing 10 min, settling 10 min, and then decanting the supernatant liquid. The
25 washing procedure was repeated a second time before the resin was stored in a closed container in DI water. 2.2.3 CO 2 Regeneration CO 2 gas (Instrument Grade, Airgas) was used as an alte rnative method of producing MIEX HCO 3 resin, given the fact that CO 2 gas would generate carbonate species when bubbled in DI water ( Mills and Urey 1940 Benjamin 2002). 500 mL of DI water was added to a 1000 mL beaker and placed on a magnetic stir plate wi th stir bar. A pH probe, tube from the compressed gas cylinder, and 20 mL of MIEX Cl resin (i.e., 40 mL /L) were added to the DI water. The beaker was covered with parafilm to minimize atmospheric interference in the headspace of the beaker. Prior to mixing or sparging, a sample was taken from the beaker and served as the initial measurement. The mixing was started and CO 2 gas was sparged so that there was vigorous bubbling within the vessel with a CO 2 partial pressure equal to atmospheric pressure. A consta nt flow rate of gas was maintained using a single stage regulator (Fisherbrand) set at a 140 kPa. At 5, 15, 30, and 60 min, the stirring and sparging were temporarily stopped, MIEX resin was allowed to settle for 30 sec, and then a sample was taken before resuming mixing and sparging. The process was repeated with nitrogen (N 2 ) gas (Industrial Grade, Airgas), as a control in place of CO 2 gas, under identical experimental conditions. All experiments were conducted in duplicate. All samples were filtered and analyzed for chloride.
26 2.2.4 Titration of NOM The SFR NOM isolate was titrated to provide detailed data about the charge density as a function of pH. The procedure was based on work by Boyer and Singer (2008), which follows a similar procedure as Ritchie and Perdue (2003) in which potentiometric titration of NOM was shown to be in good agreement with previous approaches documented in the literature. A 50 mL stock solution of NOM in DI water was prepared so as to have a concentration of 300 mg NOM/L and 0. 1 M KCl. The NOM stock solution was covered with parafilm to limit interference by the atmosphere, and purged with N 2 gas (Industrial Grade, Airgas) for 30 min while being continuously stirred using a magnetic stir plate and stir bar. After 30 min, the ini tial pH was recorded and the NOM stock solution was titrated with 0.04 M NaOH in increments of 0.1 mL After each addition of titrant, the solution was allowed 1 min to equilibrate before the pH was recorded and the titration step was repeated. The complet e titration procedure was conducted in duplicate and the titration curves were averaged. The charge density was calculated using Equation 2 1 as follows: NOM charge density (meq/mg C) ( 2 1) The NOM concentration on a charge basis ( in meq/L) was calculated as the product of the charge density (meq/mg C) and the mass concentration (mg C/L).
27 2.2.5 Analytical M ethods The pH was measured with an Ac cumet AP71 pH meter with a pH/automatic temperature compensation probe that was calibrated using pH 4, 7, and 10 buffer solutions. After pH was measured, all samples were filtered through 0.45 m nylon membrane filters (Millipore) to ensure only dissolved species were analyzed and to prevent unsettled MIEX resin from interfering with analysis. Membrane filters were pre rinsed with 500 mL of DI water and ~20 mL of sample. DOC and dissolved inorganic carbon (DIC) were measured on a Shimadzu TOC V CPH total organic carbon analyzer equipped with an ASI V autosampler. DOC and DIC calibration standards were prepared in DI water using potassium hydrogen phthalate (Shimadzu Scientific Instruments, Inc.) and sodium carbonate/sodium bicarbonate (Shimadzu Scientific Instruments, Inc.), respectively. Bicarbonate concentrations were calculated by converting DI C into total carbonate and, if necessary, using alpha values for carbonic acid at the pH of the sample (Benjamin 2002). Chloride, nitrate, and sulfate were measured on a Dionex ICS 3000 ion chromatograph equipped with IonPac AG22 guard column and AS22 anal ytical column using 4.5 mM Na 2 CO 3 /1.4 mM NaHCO 3 eluent (AS22 eluent concentrate, Dionex Corporation) as described in Apell and Boyer (2010). Calibration standards for chloride, nitrate, and sulfate were prepared in DI water using sodium salts that were ACS reagent grade purity. All DOC, DIC, chloride, nitrate, and sulfate measurements were made in duplicate and averaged. Each run on the TOC V CPH and the ICS 3000 was monitored using calibration check standards to ensure that the measured concentration was wi thin 10% of the known value. Ultraviolet absorbance at 254 nm (UV 254 ) was measured on a Hitachi U 2900 spectrophotometer in a 1 cm
28 quartz cuvette. Specific UV 254 absorbance (SUVA 254 ) was calculated by dividing UV 254 by DOC. 2.3 Results and Discussion 2. 3.1 Affinity There are a range of potential problems associated with chloride form anion exchange, as discussed in the Introduction. This is particularly relevant to MIEX treatment, because a majority of the peer reviewed research studying MIEX is based on MIEX resin with chloride as the mobile counter ion (i.e., MIEX Cl) leaving a gap in the knowledge of alternative forms of MIEX resin. Furthermore, knowledge about the affinity of MIEX resin for inorganic anions and NOM is fundamental to modeling and desig ning the MIEX process (Boyer et al. 2010). Therefore, research is needed that investigates the performance of MIEX resin using non chloride mobile counter ions. The following criteria were used to evaluate non chloride anion exchange: (i) affinity for targ et contaminants, (ii) efficacy of regeneration, and (iii) ion exchange mechanism. These criteria were systematically evaluated by studying different forms of MIEX resin and two synthetic model waters. This section focuses on the selective removal of target contaminants. Preliminary jar tests were conducted with 1, 2, and 4 mL /L of MIEX Cl resin using the synthetic inorganic model water (results not shown). A MIEX resin dose of 4 mL /L was chosen for all subsequent jar tests because this dose achieved adequat e removal of anions while still maintaining competition among the various anions. Figure 2 1 shows the results of jar tests conducted with 4 mL /L of different forms of MIEX resin using the synthetic inorganic model water. MIEX HCO 3 (100), MIEX HCO 3 (10),
29 MIEX NO 3 (10), and MIEX 2 SO 4 (10) were produced in their respective sodium salt solution as described in Section 2. 3. 2. 1. MIEX HCO 3 (100) and MIEX HCO 3 (10) showed very similar removal of nitrate and sulfate, with MIEX HCO 3 having a greater affinity f or sulfate over nitrate. The increase in chloride seen for MIEX HCO 3 (10) resin suggests that there was a small fraction of chloride mobile counter ions remaining on the resin; however, this did not appear to affect nitrate or sulfate removal. Both 100 a nd 10 MIEX HCO 3 resins showed similar removal of nitrate and sulfate as MIEX Cl resin. The greater affinity of MIEX Cl for sulfate over nitrate is in agreement with previous selectivity studies using MIEX Cl resin and synthetic inorganic model waters (Boy er et al. 2008). The MIEX NO 3 and MIEX 2 SO 4 resins demonstated the same affinity trends as MIEX HCO 3 and MIEX Cl resins. For example, MIEX NO 3 resin showed substantial removal of sulfate and minimal removal of other anions, and MIEX 2 SO 4 resin showed <25% removal of all anions. Thus, Figure 2 1 illustrates that bicarbonate form anion exchange performs nearly ideatical to chloride form anion exchange, and that all forms of MIEX resin exhibit the same affinity trend with sulfate > nitrate > bicarbonate ~ chlo ride. Figure 2 2 evaluates the effect of NOM on the removal efficiencies of MIEX HCO 3 and MIEX Cl resin. Bicarbonate form MIEX resin was produced using a 10 regeneration solution. MIEX HCO 3 and MIEX Cl correspond to the data in Figure 2 1. MIEX HCO 3 (NOM) and MIEX Cl (NOM) are from jar tests using the synthetic NOM containing model water. The presence of NOM had a negligible effect on the removal of nitrate for MIEX HCO 3 and MIEX Cl resin. Sulfate removal appeared to be slightly improved in the presence of NOM for both resins. This is likely an artifact of conducting
30 single jar tests for the synthetic NOM containing model water, and will be discussed in more detail in Section 188.8.131.52 evaluating regeneration. MIEX Cl resin removed more DOC than MIEX HCO 3 res in (82% vs. 61%, respectively); however, both resins removed >90% of UV 254 absorbing compounds. The preference of MIEX Cl resin for the UV 254 absorbing fraction of NOM, especially in high SUVA 254 waters, has been reported in the literature ( Singer and Bily k 2002, Boyer and Singer 2005 ), and it appears that MIEX HCO 3 resin has a similar preference as MIEX Cl resin for UV 254 absorbing NOM. Figure 2 2 also shows that both MIEX HCO 3 and MIEX Cl resin have a similar affinity for sulfate and DOC. Previous researc h studying conventional chloride form resin and MIEX Cl resin discuss that sulfate and NOM are competitors for anion exchange sites and that NOM properties play a key role (Fu and Symons 1990, Boyer et al. 2008). In summary, the affinity sequence for both MIEX HCO 3 and MIEX Cl resin, in the presence of NOM, is UV 254 absorbing substances > DOC ~ sulfate > nitrate > bicarbonate ~ chloride. Most importantly, the results in Figures 2 1 and 2 2 clearly show that MIEX HCO 3 resin is a suitable alternative to MIEX Cl resin in terms of affinity and removal efficiency. 2.3.2 Regeneration 184.108.40.206 Salt regeneration Many studies have investigated the performance of fresh ion exchange resin ( Li and Sengupta 1998 Jelinek et al. 2004, Hsu and Singer 2010 ), but only a few studies have investigated the regeneration of the same batch of resin to simulate real world conditions of ion exchange treatment (Mergen et al. 2008, Apell and Boyer 2010 ). It was shown in Section 2. 3.1 that MIEX HCO 3 resin has a high affinity for sulfat e. As a result,
31 studies were conducted to evaluate whether sodium bicarbonate could regenerate MIEX resin that was saturated with sulfate. The regeneration procedure is described in Section 2. 2. 2. Briefly, virgin MIEX Cl resin was saturated with a 10 sodi um sulfate solution to produce MIEX 2 SO 4 resin and then regenerated with a 10 sodium bicarbonate solution to produce fresh MIEX (SO 4 ) HCO 3 resin. Similarly, virgin MIEX Cl resin was regenerated with a 10 sodium bicarbonate solution to produce fresh MIEX HCO 3 synthetic model water. The fresh resins were tested following the jar test procedure using the synthetic inorganic model and then regenerated. The sequence of jar test and resin regeneration was repeated two additional times. The results are summarized in Figure 2 3. The fresh MIEX (SO 4 ) HCO 3 resin removed less nitrate and sulfate than the fresh MIEX HCO 3 resin. This suggests that the regeneration procedure did not fully saturate the MIEX 2 SO 4 resin with bicarbonate, i.e., sulfate remained on the resin which resulted in less removal of aqueous nitrate and sulfate. However, MIEX (SO 4 ) HCO 3 and MIEX HCO 3 resins showed nearly identical removal of nitrate and sulfate after one regener ation cycle, and the performance remained similar after two regeneration cycles. These results provide quantitative data which demonstrate that sodium bicarbonate can effectively regenerate exhausted anion exchange resin. The effect of multiple regenera tion cycles on MIEX resin and NOM removal was the next step in evaluating the efficacy of sodium bicarbonate regeneration. Figures 2 4a and b show the final normalized concentrations of inorganic anions and NOM after three regeneration cycles using MIEX HC O 3 and MIEX Cl resin, respectively. Fresh MIEX HCO 3 resin was produced by regenerating virgin MIEX Cl with a 10 sodium
32 bicarbonate solution. The experimental procedure, i.e., jar tests and regeneration of resin, was analogous to the procedure correspondin g to Figure 2 3. Fresh MIEX HCO 3 resin showed greater removal of UV 254 nitrate, and sulfate than regenerated MIEX HCO 3 resin (Figure 2 4a). After three regeneration cycles the order of removal for MIEX HCO 3 was UV 254 > DOC > sulfate > nitrate. MIEX Cl res in followed a similar trend as MIEX HCO 3 resin with respect to fresh resin showing greater removal efficiencies than regenerated resin, and regenerated MIEX Cl showing the same order of NOM and inorganic anion removal. The consistent removal and affinity t rends for NOM and inorganic anions indicate that sodium bicarbonate is as effective as sodium chloride at regenerating MIEX resin. Figures 2 4a and b also show that regenerated MIEX HCO 3 resin has a higher affinity for NOM and a lower affinity for inorgani c anions than regenerated MIEX Cl resin. This result illustrates that regeneration of anion exchange resin is critical for understanding the true behavior of the resin. Other researchers have also reported that regenerated anion exchange resin performs dif ferently than fresh anion exchange resin ( Apell and Boyer 2010 ), thus underscoring the importance of evaluating ion exchange treatment over multiple regeneration cycles. In summary, the regeneration efficiency of MIEX HCO 3 resin was similar to MIEX Cl resi n with respect to removal of inorganic anions and NOM. This is significant because it is the first work to quantify that sodium bicarbonate can be used in place of sodium chloride to regenerate MIEX resin which has been partially exhausted with NOM. 2.3.2 .2 CO2 regeneration Experiments were conducted to investigate the potential of producing MIEX HCO 3 resin by sparging CO 2 gas in DI water that contained MIEX Cl resin. Figure 2 5
33 shows the aqueous chloride concentration versus the amount of time gas (either CO 2 or N 2 ) was sparged in the water/resin slurry. Sparging CO 2 gas resulted in release of chloride from the MIEX Cl resin and an aqueous chloride concentration that was ~6 7 times greater than the initial chloride concentration. This data suggests that bi carbonate species are exchanging with resin phase chloride to produce MIEX HCO 3 resin. In contrast, sparging N 2 gas resulted in no change in the aqueous chloride concentration relative to its initial concentration. The result for N 2 gas was expected becaus e N 2 does not generate anions in water. Overall, while the generation of aqueous bicarbonate and carbonate species from carbonic acid is well known, these results illustrate that sparging CO 2 gas leads to generation of bicarbonate species as a mobile count er ion that can then exchange with resin phase chloride and produce bicarbonate form resin. Future research is needed that optimizes the operating conditions at which CO 2 gas is sparged in water and investigates the regeneration of MIEX resin that is exhau sted with sulfate, NOM, and other anions. 2.3.3 Stoichiometry The stoichiometry of ion exchange was investigated for all forms of MIEX resin using the synthetic inorganic and NOM containing model waters. The approach used in this work to quantify ion exch ange stoichiometry has been previously published and used successfully (Boyer and Singer 2008, Boyer et al. 2008). The data were evaluated by plotting the total uptake of anions by the resin (in meq/L) against the total release of anions by the resin (in m eq/L), where the anion released is the mobile counter ion. The data were expressed in meq/L to account for exchange of monovalent and divalent inorganic anions and polyanionic NOM. The NOM concentration in meq/L was plotted
34 over the pH range of the sample and is shown in Figure 2 6, with calculations performed as described in Section 2 .2 .4. The change in concentrations for organic matter and the anions were then compared to ensure that each component contributed to the stoichiometry of the system. Figure 2 7 shows the stoichiometry of various forms of MIEX resin using the synthetic inorganic model water, i.e., in the absence of NOM. The y = x line shows the equivalent exchange of anions in solution for anions on the resin, and represents ideal ion exchange. The data points were collected from the various affinity and regeneration experiments (see Sections 2. 3.1 and 2. 3.2 ), which in the case of MIEX Cl and MIEX HCO 3 resins generated multiple data points on the figure due to the multiple regeneration steps the resins were subjected to. Data for MIEX Cl, MIEX NO 3 and MIEX 2 SO 4 resins fall along the y = x line, and confirm the expected ion exchange stoichiometry. Data for MIEX HCO 3 resin are generally scattered above the y = x line, which indicates that more chl oride, nitrate, and sulfate ions were removed by the resin than bicarbonate ions were released by the resin. This was surprising because bicarbonate was expected to follow the same ion exchange stoichiometry as the other inorganic anions. A review of prev ious literature suggested that it was possible for multiprotic mobile counter ions, such as carbonic acid and arsenic acid species, to deprotonate within anion exchange resins (Kimura et al. 1982, Horng and Clifford 1997). It is important to note that whil e regenerating the bicarbonate form resin (as described in Section 2. 2.2), there was a decrease of ~1 2 pH units during the DI water rinse step. Although the pH of the rinse water was not measured with every regeneration, there
35 was a consistent decrease in pH. Thus, it was hypothesized that bicarbonate was deprotonating within the MIEX resin, and the exchange process for MIEX HCO 3 was a combination of bicarbonate and carbonate mobile counter ions exchanging with chloride, nitrate, and sulfate in solution. T his hypothesis was evaluated both theoretically and experimentally. First, data for MIEX HCO 3 was recalculated assuming that the change in DIC in solution was due solely to carbonate release by the resin (see calc. R2 CO3 in Figure 2 7). Second, MIEX 2 CO 3 resin (see R2 CO3 in Figure 2 7) was generated in a solution of pH 11.5 under N 2 atmosphere to prevent interference by CO 2 and tested following the jar test procedure described previously. The calculated data for MIEX 2 CO 3 generally fall below the y = x l ine indicating that the equivalent concentration of mobile counter ions released from the resin (i.e., carbonate) is greater than the equivalent concentration of anions removed from solution, which is in agreement with bicarbonate and carbonate acting as t he mobile counter ion. The experimental data point for MIEX 2 CO 3 falls on the y = x line suggesting that carbonate is the mobile counter ion on the resin, and the carbonate does not become protonated in the resin. Thus, experimental and calculated data, in conjunction with the noted decrease in pH during regeneration, suggest that the true mobile counter ion within MIEX HCO 3 resin is a mixture of monovalent bicarbonate and divalent carbonate. Figure 2 8 shows the stoichiometry of MIEX HCO 3 and MIEX Cl resin s in the presence of NOM and as a function of the number of regeneration cycles corresponding to Figures 2 4a and b. Although previous work has confirmed that removal of NOM by virgin MIEX Cl resin is a stoichiometric processes (Boyer and Singer 2008, Boye r et al. 2008), this is the first work to evaluate the stoichiometry of ion exchange in the presence of NOM as a
36 function of the regeneration process. This is a very important consideration because MIEX resin is intended to be regenerated and reused multip le times. Data for fresh MIEX HCO 3 and virign MIEX Cl resin fall along the y = x line, as expected from previous research (Boyer et al. 2008). Importantly, the data for regenerated MIEX Cl and regenerated MIEX HCO 3 resin also fall along the y = x line. Thi s is not a contradiction to Figure 2 7, rather it suggests that the bicarbonate mobile counter ion does not always deprotonate within the resin phase, and that further investigations are needed to fully identify the mechanisms associated with why and to wh at extent the deprotonation occurs. Figure 2 8 does confirm, however, that over the course of multiple regeneration cycles ion exchange remains the operative mechanism of NOM removal, thus confirming that MIEX resin is reuseable without permanent fouling. Although MIEX HCO 3 resin performed very similarly to MIEX Cl resin with respect to affinity and regeneration, the stoichiometry results in Figure 2 7 indicate that there is a difference between the pore scale chemistry of bicarbonate form resin and chlorid e form resin. For example, chloride mobile counter ions are not expected to be affected by variations in pH or the presence of cations, whereas bicarbonate mobile counter ions will be affected by variations in pH and the presence of divalent cations. Furth ermore, the deprotonation of resin phase bicarbonate to carbonate and the reaction of carbonate with calcium could form calcium carbonate minerals within the resin pore structure, which would foul the resin. Thus, additional research is needed before bicar bonate form anion exchange can be implemented at the full scale.
37 2.4 Chapter Summary The goal of this work was to evaluate the extent to which bicarbonate form anion exchange is a viable alternative to chloride form anion exchange. The motivation for this work was the potential adverse effects of chloride form anion exchange, such as increased corrosion and excessive chloride loading to wastewater treatment plants and surface waters. The major conclusions of this work are: Fresh MIEX HCO 3 resin showed very similar removals of nitrate, sulfate, DOC, and UV 254 absorbing substances as virgin MIEX Cl resin. Fresh MIEX HCO 3 resin and virgin MIEX Cl resin showed greater removal of inorganic anions and NOM than corresponding regenerated resins. Nevertheless, sodiu m bicarbonate had approximately the same level of regeneration efficiency as sodium chloride when both anions were used at a concentration 10 times the equivalent capacity of MIEX resin. The affinity sequence for MIEX resin was UV 254 absorbing substances > DOC > sulfate > nitrate > bicarbonate ~ chloride. This affinity sequence was based on using bicarbonate chloride nitrate and sulfate form MIEX resin, synthetic inorganic and NOM containing model waters, and fresh and regenerated resin. Sparging CO 2 gas in a water/MIEX Cl resin slurry resulted in release of chloride from the resin, which is suggested to result from bicarbonate species acting as mobile counter ions. No release of chloride was observed when N 2 gas was sparged in a water/MIEX Cl resin s lurry. Thus, the sparging of CO 2 gas is a potential technique for regeneration of bicarbonate form MIEX resin because it does not require salt. The stoichiometry of chloride nitrate and sulfate form MIEX resin followed ideal ion exchange behavior. In c ontrast, the stoichiometry of bicarbonate form MIEX resin did not always follow predictable ion exchange. It appears that resin phase bicarbonate is deprotonating and resulting in a mixture of bicarbonate and carbonate mobile counter ions. Bicarbonate and chloride form MIEX resin showed clear ion exchange stoichiometry for NOM removal over multiple regeneration cycles.
38 Table 2 1. Average Measured Composition of Synthetic Model Waters used in Ion Exchange Experiments Species Inorganic model water NOM c ontaining model water chloride (mg/L) 15 14 bicarbonate (mg/L) 140 95 nitrate (mg N/L) 28 27 sulfate (mg/L) 46 46 DOC a (mg C/L) 4.1 a SUVA 254 ~ 5.4 L/mg Cm Figure 2 1. Effect of mobile counter ion on affinity of MIEX resin using synthetic inorganic model water (C/C 0 = 4.6 for Cl ). 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 R-HCO3 (100x) R-HCO3 (10x) virgin R-Cl R-NO3 (10x) R2-SO4 (10x) C/C 0 form of MIEX resin HCO3ClNO3SO4--
39 Figure 2 2. Comparison of bicarbonate form and chloride form MIEX resin in the absence and presence of SFR NOM. Figure 2 3. Effect of multiple regeneration cycles on the removal efficiency of bicarbona te form MIEX resin using synthetic inorganic model water. Legend shows anion removed from solution, form of MIEX resin. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 NO3SO4-DOC UV254 C/C 0 anions in solution R-HCO3 R-HCO3 (NOM) R-Cl R-Cl (NOM) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 fresh resin 1 2 C/C 0 number of regenerations NO3-, R-HCO3 NO3-, R-(SO4)-HCO3 SO4--, R-HCO3 SO4--, R-(SO4)-HCO3
40 Figure 2 4. Effect of multiple regeneration cycles on the removal efficiencies of (a) bicarbonate form MIEX resin and (b) chloride form MIEX resin using synthetic NOM containing model water. Figure 2 5. Generation of bicarbonate form MIEX resin by CO 2 gas sparging. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 fresh resin 1 2 3 C/C 0 number of regenerations a. MIEX HCO 3 UV254 DOC SO4-NO30 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 fresh resin 1 2 3 C/C 0 number of regenerations b. MIEX Cl UV254 DOC SO4-NO30 2 4 6 8 10 12 14 0 10 20 30 40 50 60 chloride in solution (mg/L) time of gas sparge (min) carbon dioxide nitrogen
41 Figure 2 6. Charge density of SFR NOM derived from titration. Figure 2 7. Stoichiometry of various forms of MIEX resin using synthetic inorganic model water. All results are experimental data except calc. R2 CO3, which was calculated using R HCO3 data. 0 5 10 15 20 25 30 2 4 6 8 10 12 charge density (meq/g C) pH SFR NOM 0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3 anion uptake (meq/L) anion release (meq/L) y = x R-Cl R-NO3 R-HCO3 R2-SO4 R2-CO3 calc. R2-CO3
42 Figure 2 8. Effect of regeneration on the stoichiometry of bicarbonate form and chloride form MIEX resin u sing synthetic NOM containing model water. Number of regeneration cycles is given by the number in parentheses. 0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3 anion uptake (meq/L) anion release (meq/L) R-HCO3 (0) R-Cl (0) R-HCO3 (1) R-Cl (1) R-HCO3 (2) R-Cl (2) R-HCO3 (3) R-Cl (3)
43 CHAPTER 3 IMPACT OF REGENERATION ON FRACTIONATION OF NOM IN MIEX TREATED SOUTH AUSTRALIAN WATERS 3.1 Background Growing water demands and diminishing sources have led to the increased use of water supplies with high dissolved organic carbon (DOC) contents. In many of these waters, high DOC tends to be primarily composed of natural organic matter (NOM). NOM is a complex mix ture of organic co mpounds that consist of humic and fulvic acids (Collins et al. 1986) These high DOC water are an issue during treatment due to their complex chemical nature and their associated issues, such as the formation of disinfection by products. One such means of removal is through the use of ion exchange (IEX). IEX has been found to be an effective means of removal of DOC from surface waters (Kitis et al. 2007, Boyer and Singer 2005, 2008a, Jarvis et al. 2006, Neale and Schafer 2009, Cornelissen et al. 2009) A ma jor benefit to the use of IEX is regeneration process of these resins is that the resin becomes fouled over time, and loses some of its capacity to remove DOC from source waters, thus decreasing the effectiveness of the process (Walker and Boyer 2011) While IEX processes are well established and well researched this work focuses on the application of magnetic ion exchange (MIEX) resin for the removal of DOC from surface w aters and the characterization of the molecular fractions of NOM in the MIEX treated waters and spent regeneration brine. MIEX has been shown in many instances Reprod uced with permission from Rokicki, C.A., Morran, J., Drikas, M., and Boyer, T.H., 2013. Impact of regeneration on fractionation of NOM in MIEX treated South Australian Waters. Under review at Water Research
44 to be an effective means of removing DOC from water (Boyer and Singer 2005, 2008b, Drikas et al. 2007, Humbert et al. 2005) MIEX is able to remove DOC by interaction with negatively charged functional groups in the structure of the organic compounds, thus effectively removing them from solution (Boyer et al. 2008) One of the advantages of MIEX resi n is its small particle size which allows for easy fluidization, and once agitation is discontinued its magnetic properties causes the resin to aggregate and settle (Singer et al. 2007) The primary means of evaluating the performance of MIEX is through D OC analysis of bulk water composition. One method of characterizing organic matter in solution is through the use of UV 254 and subsequently SUVA. SUVA determination is performed by measuring the UV absorbance of the sample at 254 nm and then normalizing it with the concentration of DOC in solution. SUVA has been found to provide an estimation of the molecular weight and aromaticity of organic matter (Weishaar et al. 2003) Another method for further investigation of organic matter properties is through the use of high performance size exclusion chromatography (HPSEC) (Pelekani et al. 1999) Size fraction analysis of DOC in drinking waters is important to consider because different molecular weight fractions have different properties that impact on downstream drinking water treatment processes. Studies have shown that larger molecular weight fractions are more easily removed by coagulation, while treatment techniques such as activated carbon can better target lower molecular weight fractions (Chow et al. 2008) The hydrophilic nature of the smaller to midsize fractions of organic matter allows for it to be targeted by IEX (Croue et al. 1999) The removal of NOM is driven by electrostatic interactions which are governed by the
45 selectivity of resins for different anions (Boyer et al. 2008) In work done investigating the performance of MIEX in comparison and in conjunction with other water treatment processes, it has been shown that MIEX is extremely effective at reducing the mid range of organic fractions between 500 1500 Da (Humbert et al. 2005) with the majority of the fractions in this size range including humic substances that have both high and low molecular weight (Allpike et al. 2005) Within this specific range, it has also been found that the resin has a further additional targeted fraction. MIEX has been found to target three areas within the mid to low molecular weight organic matter fraction (Huang et al. 2012) One of the concerns of these molecular weight fractions, aside from disinfection by products (DBP) formation, is their ability and tendency to foul membrane systems (Huang et al. 2012, Howe and Clark 2002, Lee et al. 2004) treatment range is ideal for the coupling of IEX with coagulation due to the tendency of coagulation to target the larger fraction (> 1500 Da) of colloidal and biological residues (Chow et al. 2008) The combination of MIEX with treatment systems have been shown to create a more stable treated water that reduces the DBP formation potential (Drikas et al. 2007, Humb ert et al. 2005, Drikas et al. 2003, Drikas et al. 2011, Boyer and Singer 2006, Fearing et al. 2004) The use of MIEX to reduce DOC in treated systems can have broader impacts on the finished water and subsequent treatment streams. Two of the primary issu es associated with MIEX treatment process is the generation of high salt concentration brine after regeneration and potential impacts on finished water corrosivity following treatment (Edwards and Triantafyllidou 2007, Ishii and Boyer 2011, Willison and Bo yer 2012) In lieu of chloride it is possible to utilize the bicarbonate ion as the mobile ion and
46 it has been shown that bicarbonate and chloride have similar selectivities and perform the same when removing DOC from water (Rokicki and Boyer 2011) Previo us work done has shown that in lab scale studies it is sufficient to use stoichiometric excesses when determining concentrations of regeneration agents for MIEX resins (Rokicki and Boyer 2011, Apell and Boyer 2010) It was found that 10 the required stoic hiometric dose was able to regenerate resin such that resin was able to perform as effectively as a resin regenerated at a much stronger dose of 100 (Rokicki and Boyer 2011) .While it has been found that stoichiometric based regeneration techniques for lab oratory scale experiments can yield similar performances, those experiments still observe a decrease in resin performance over multiple regeneration cycles (Walker and Boyer 2011, Rokicki and Boyer 2011) It is not completely clear through these limited mu ltiple regeneration cycle experiments what the impact of regeneration is on the resin processes, due to only a handful of studies focused on multiple regeneration cycles and the majority of experimentations being focused on single batch experiments. It is unclear how different regeneration agents and concentrations factor into the performance and fouling of the resins used in these systems. One way to track potential fouling of resins is through the exploration of organic matter fraction uptake by resin. Th ere have been multiple studies (Humbert et al. 2005, Allpike et al. 2005, Fabris et al. 2007) but a lack of literature into which fractions of NOM cause this build up or fouling on the resin. The lack of literature relating the resin regeneration and the preferred organic matter fractions of the resin do not provide a complete picture of the fouling and regeneration of MIEX resins.
47 The overall goal of this work was to evaluate how different regen eration agents can impact the performance of MIEX resin in real world applications. This was accomplished by taking resin from two different full scale MIEX treatment plants and applying two regeneration agents at different concentrations. After regenerati ng each resin, the resins were used to treat the raw water from the treatment plants. The specific objectives of this work were to: (1) evaluate the impact that the composition and concentration of regeneration agents has on DOC and UV absorbance (UVA 254 ) removal; (2) evaluate the impact of regeneration agent on the molecular weight fraction of NOM in the MIEX treated water and waste regenerant brines; and (3) determine how the bench scale regeneration compared with the performance of full scale regeneratio n systems in the water treatment plants. Multiple loading jar tests were performed on the regenerated resins from the treatment plants to simulate the treatment conditions in the full scale system. The resins were then regenerated to simulate continuous re use of resin. 3.2 Materials and Methods 3.2.1 Treatment P lants and R aw W aters All waters and MIEX resins used in this work were collected from treatment plants located in South Australia, Australia. The two treatment plants sampled were the Mount Pleasa nt Water Treatment Plant (Mount Pleasant WTP), located in Mount Pleasant, South Australia and the Middle River Water Treatment Plant (Middle River WTP) on Kangaroo Island, South Australia. A summary of water quality parameters from on site measurements fro m these two sites can be found in Table 3 1.
48 Mount Pleasant WTP draws water from the River Murray, which is a high turbidity and high DOC body of water. The WTP utilizes two separate water treatment trains, both utilizing the MIEX as a pre treatment proce ss for DOC removal, with one train using conventional treatment with coagulation sedimentation and filtration, and the second stream utilizing a microfiltration system. The MIEX system at Mount Pleasant WTP utilizes a mixing tank for the contactor and coll ects resin in an adjacent settling tank for recycling (Drikas et al. 2011) Middle River WTP draws from the Middle River Reservoir which is a water source characterized with lower turbidity and high DOC. The treatment train at Middle River utilizes MIEX as a pre treatment process for DOC removal, followed by coagulation, sedimentation, and filtration. The MIEX system at Middle River WTP utilizes a fluidized bed, high rate contactor, with tube settlers to prevent MIEX loss. Resin is pumped directly out of th e contactor in the Middle River WTP and sent into the regeneration system. 3.2.2 Anion E xchange R esins All experiments in this work were performed using MIEX resin from Orica Watercare. Resins used with each water sample were collected from their respecti ve water treatment plant at either Middle River or Mount Pleasant, and were sampled from the resin collection systems, after use but prior to regeneration. Resins were collected prior to regeneration specifically so that the resins were partially saturated with DOC and anions from the WTP to allow more realistic starting conditions of the experiment.
49 3.2.3 Multiple L oading I on E xchange J ar T ests Multiple loading jar tests were performed so that bench scale experiments would more realistically mimic perfor mance conditions in the full scale water treatment processes (Kitis et al. 2007, Mergen et al. 2008) In full scale operations only a fraction of spent resin is regenerated, while water is continuously being treated. In this system it is only a small fract ion of the regenerated resin that is in contact with the water being treated. Full scale MIEX systems are designed for the resin to treat a certain number of bed volumes (BV) of water prior to resin regeneration. Experiments were performed on a Phipps & B ird PB 900 jar tester with 2L B KER 2 square acrylic jars. All experiments were performed in triplicate and all data presented represents the mean of the triplicate experiments with error bars signifying one standard deviation. 2 L of raw water taken from the same WTP as the resin was added to 10mL of the regenerated resin. The water was mixed at 100 rpm for 20 min. After 20 min the water was decanted with care to prevent resin loss, and an additional 2L of raw water was added then mixed at 100 rpm for an additional 20 min. This process was repeated until there was a contact of 600 BV for the Middle River water/resin and 800 BV for the Mount Pleasant water/resin. The BV contact difference is due to matching operating conditions at the respective WTPs. All s amples were analyzed for pH, UVA 254 DOC, and molecular weight fractionation as described in Section 3 2.5. 3.2.4 Resin Regeneration Spent resins were collected from the WTPs following the contactor and were regenerated in the laboratory, subjected to be nch scale experiments, regenerated
50 again, and subjected to another round of bench scale experimentation. The two regeneration agents that were investigated were (1) sodium chloride (NaCl) and (2) sodium bicarbonate (NaHCO 3 ). The goal of these two regenerat ion agents were to generate a resin with either a (1) chloride mobile counterion or (2) a bicarbonate mobile counterion. The two regeneration agent concentrations were determined using (a) a stoichiometric excess based on ion exchange capacity of MIEX resi n or (b) To regenerate the resin, resin was measured in a graduated cylinder and added to a B KER 2 jar with 1L of regeneration agent solution. Regeneration agents were made using either a solution of NaCl o r NaHCO 3 The stoichiometric based dosage was based on the total equivalents capacity of the dose of resin and multiplied by 10 to ensure adequate presence of mobile ions In the regeneration scheme used in this experiment, this is approximately 0.3% and 0. 4% by weight of the chloride and bicarbonate salt, respectively. The alternative treatment scheme was based on the manufacturer suggestions that a chloride solution should be a 10% solution, which is approx. 100 g/L of NaCl and that the bicarbonate shoul d be a 50 g/L of NaHCO 3 solution (5% by weight), which is equivalent to approximately 340 and 115 stoichiometric excess, respectively, with the different concentrations being due to the less soluble nature of the NaHCO 3 salt. Resins were mixed in the sol ution of regeneration agent at 100 rpm for 10 min. Resins were allowed to settle and were decanted, during decanting the waste regenerant brine was sampled and analyzed by HPSEC. 1 L of deionized (DI) water was added to the resin, was mixed for 10 min at 1 00 rpm, allowed to settle and decanted, this DI water rinse was repeated a second time.
51 3.2.5 Analytical M ethods All samples were filtered through 0.45 m mixed cellulose ester membrane filters, with pH and conductivity being measured prior to filtration. Filtered samples were analyzed for DOC by conductometric methods using a Sievers 900 Total Organic Carbon Analyzer, with each analysis including check standards in both the low and high end of the instruments range, these standards were changed regularly and varied slightly by the Australian Water Quality Centre quality assurance/quality control department. Samples were filtered through 0.2 m mixed cellulose ester membrane filters for apparent molecular weight (AMW) determination using HPSEC analysis. HPS EC was undertaken using a Waters Alliance 2690 separations module and 996 photodiode array detector at 260 nm. 0.1 M phosphate buffer in 1.0 M NaCl was flowed through a Shodex KW802.5 packed silica column at a rate of 1.0 mL /min, which provided effective s eparation range from approximately 1000 Da to 50000 Da. The AMW calibrations were performed with 35000, 18000, 8000, and 4600 Da polystyrene sulphonate standards as described by (Fabris et al. 2008) using methods described in (Chin et al. 1994) 3.3 Resu lts and Discussion 3.3.1 Impact of regeneration agent on DOC removal The bulk of experimentation was performed in multiple loading jar tests. Figures 3 1 and 3 2 represent the effluent DOC normalized by the influent DOC for each set of resins used to gau ge the performance of the resins in treatment systems. In this work, the regeneration cycle refers to how many times a certain resin has been regenerated after being collected from the water treatment plant. The first regeneration cycle
52 represents resin th at has been collected, then regenerated, and then subsequently used for experimental purposes. The second regeneration cycle represents resin that has been regenerated following the first cycle and then subsequently used in experimentation. Figure 3 1 show s the results from resins taken from the Mount Pleasant water treatment plant and plots the effluent DOC concentration normalized by the influent DOC concentration for the different resins over the span of the multiple treatment bed volumes. Figure 3 1 sho ws effluent DOC concentration normalized by the influent DOC for the resins regenerated using the different regenerants for the first and second regenerations. Figure 3 1a shows the resin regenerated by Cl at the operators recommended conditions (10% NaCl ), Figure 3 1b shows the resin regenerated by HCO3 at the operators recommended conditions (5% NaHCO3) and finally Figures 3 1c,d show the resins regenerated at the stoichiometrically determined dose, (0.3% NaCl and 0.4% NaHCO3, respectively.) These resul ts clearly show that with continuous loading, the resin performance decreases with increasing bed volumes treated. This trend was apparent under all regeneration conditions used. DOC removal after the second regeneration was generally similar to the first regeneration Comparison of the chloride and bicarbonate regeneration (Figure 3 1a,b and Figure 3 1c,d) show that the regenerated resins perform similarly to each other at removing DOC. The primary difference in resin performance comes from the difference between the concentration of regeneration agents as shown in Figures 3 1a vs Figure 3 1c and Figure 3 1b vs Figure 3 1d. It is important to take into account that at the 10% NaCl or 5% NaHCO 3 regeneration conditions that the molar concentration of HCO 3 i s
53 due to the fact that HCO 3 has a lower solubility, which limits the strength of the HCO 3 solution, leading to a difference between the concentration of the Cl and HCO 3 regeneration solutions. Figure 3 1 shows that the overall trend for resin performance from highest DOC removal to lowest is as follows: NaCl at 10% > NaHCO 3 at 5% > NaCl at 0.3% > NaHCO 3 at 0.4%. The decreased performance, however slight, implies that the resins are not being completel y regenerated at the lower concentrations of regeneration agents. In situations where waste regenerant brine is an issue, a lower salt concentration may be easier to dispose of and the benefit of such may be able to offset a slight decrease in DOC removal. Depending on the influent water quality, and necessary effluent quality, regeneration brine strength is another means of optimization the efficiency of a water treatment system. Figure 3 2 is identical to Figure 3 1, with the exception being that the res ins and waters used were taken from the Middle River WTP. Figure 3 2a confirms the trend apparent in Figure 3 1 of decrease in resin performance with increasing bed volume treated. Figure 3 2b shows a decrease in DOC removal with subsequent bed volumes, wi th the second regeneration showing the same performance as the first regeneration. There is a decrease in performance between the Cl resin and the HCO 3 resin regenerated at 10% NaCl or 5% NaHCO 3 which is most significant at 200 bed volumes and is likely due to the difference in the regenerant concentration strength between the two counterions. Figure 3 2c shows that the DOC removal appears to improve with multiple loadings and the best DOC removal is seen at 600 bed volumes for the first regeneration whe reas in the second regeneration the DOC removal decreases with
54 increasing bed volumes treated. Figure 3 2d matche those in Figure 3 2c. The two resins in Figures 3 2c,d behave similarly to each other which suggest that the difference in performance seen in Figures 3 2a,b is in fact likely due to the difference in concentrations with the 10% NaCl and 5% NaHCO 3 regeneration rather than the affinity of the resin. The similarity in the performance of the resins in Figures 3 2c,d imply that for the Middle River water the regeneration solution strength has more of an impact on the regeneration efficiency than the regeneration agent. Figure 3 2 shows that with the 10% NaCl and 5% NaHCO 3 regenerant (Figure 3 2a,b) the resins are more completely regenerated and are a ble to behave as expected in addition to removing more DOC than the resins regenerated using the 0.3% NaCl or 0.4% NaHCO 3 regenerant (Figure 3 2c,d.) as was observed with the resin sourced from the Mt Pleasant WTP. Furthermore, the even stronger 10% NaCl regenerant solution yields better performing resin compared to the 5% NaHCO 3 resin due to the higher molar concentration. The resins represented in Figures 3 2c,d are most likely incompletely regenerated, which explains the performance of the first regene ration of the resin. The resin may be saturated with an ion that competes with DOC for surface sites and over time the DOC concentration is able to shift so that it is more preferred by the resin due to concentration gradients. The subsequent regenerations cause the resin to behave as expected, which implies that there was fouling on the resin that the initial regeneration was unable to reverse. In comparing the performance of both resins, the primary similarity is the trend of 200 BV treatment having achi eved the best DOC removal with subsequent treatment bed volumes suffering a decrease in performance, which is consistent with trends seen
55 in other multiple loading experiments in which MIEX undergoes an initial saturation phase and then sees a decline in p erformance (Kitis et al. 2007) However there are also noticeable differences between the performances of the resins between the two water sources. Figure 3 1 showed the best performance at just over 60% DOC removal at 200 BV, with the median removal for a ll bed volumes at approximately 40%. The data in Figure 3 2 showed that the best removal was approximately 45% at 200 BV, with the median removal across all the bed volumes closer to approximately 25% for all the resins (both regeneration agents at both co ncentrations) treating Middle River source water. The difference between the performances of the resins in these two different waters is likely due to both the much higher level of DOC, sulfate, and bromide in the Middle River water, which utilizes a great er proportion of the ion exchange capacity of the resin with the sulfate and bromide competing with the DOC. It also appears that the resin may be more heavily loaded and/or the type or organics present in these two water sources differs to the extent that the Middle River DOC is more closely (and irreversibly) bound to the resin and hence is not as readily removed during regeneration. 3.3.2 Impact of R egeneration A gent on NOM F ractionation Another aspect to consider is whether or not the different regene ration agents impact the molecular weight fraction of the DOC that is removed during treatment. Figure 3 3 shows chromatograms representing the performance of the resins and water taken from Mount Pleasant for the different combination of regeneration agen ts and concentrations. The x axis shows the apparent molecular weight in Da on a logarithmic scale and plotted against that is the absorbance of the DOC. In these figures, the larger
56 the absorbance corresponds to a greater amount of UV absorbing organic ma tter at that specific molecular weight. Each part of Figure 3 3 shows the raw water compared to the sampled from Mount Pleasant. Figure 3 3a,b show the use of the resin s regenerated with 10% NaCl or 5% NaHCO 3 for the first and second regenerations, respectively. These figures show similar performance for all the resins between the first and second regeneration. Figures 3 3c,d show the resins that were regenerated using w ith 0.3% NaCl or 0.4% NaHCO 3 for the first and second regeneration, respectively, and confirm the previous observations that less organics were removed. Figure 3 4 is arranged identically to Figure 3 3 and represents the data from the Middle River resins a nd water. Figures 3 3 and 3 4 demonstrate the same overall trend of decreasing performance with increasing bed volumes with the HCO 3 trailing the Cl at each bed volume. The one deviation from the performance trend being in Figure 3 3d, where the performa nce trend showed that the Cl resin at the 400 and 600 bed out performed the HCO 3 resins at 400 and 600 bed volumes, with the rest of the resins following the same trend of HCO 3 following Cl As shown in Figures 3 1 and 3 2 the the Mount Pleasant resins more effectively reduces the peak in the 1000 2000 Da range than the Middle River resin. While comparing these two water sources, it is important to consider the difference in the source water quality; Middle River water, in addition to having an overall greater UV absorbance, a greater proportion of the UV absorbing material is greater than 1000 Da when compared to the Mount Pleasant water. The disparity in scale of the two figures is due to the higher DOC level in the Middle River water supply, coupled w ith the tendency of the carboxyl and aromatic fraction of organic matter to be in this range (Abbt Braun et
57 al. 2004, Lankes et al. 2008) with this high aromatic and UV absorbing fraction being the most rapidly removed by IEX. The difference in performance between Mount Pleasant and Middle River is likely due to the resin targeting this fraction and then becoming exhausted. The 1000 this fraction, and the drop off in removal after the 200 bed volumes treated as shown in Figure 3 3. Figure 3 4 shows a decreased peak in the 1000 2000 Da range, which supports the argument that the resin favors this fraction, followed by the smaller fract ions. Based upon Figures 3 3 and 3 4, it appears that the 1000 by the 500 700 Da range, and lastly the peak at 200 300 Da range, which represents an agglomeration of organic matter associated with the salt peak that precedes HPSEC (Huber et al. 2011, Swift and Posner 1971) The removal of the large peak primarily is likely to be due to the larger overall concentration of the organic matter in this fraction, with the overall preference of MIEX for organic mat ter fractions in the 500 1500 Da range being consistent with previous findings (Humbert et al. 2007) and the mid range preference being supported elsewhere in the literature (Allpike et al. 2005, Huang et al. 2012) Previous work found that the MIEX resin prefers the organic fractions in the 5 to 2 kDa size range (Mergen et al. 2009) was in good agreement with the findings of this preference over multiple use cycles HPSEC has been used in previous studies to investigate the performance of MIEX for the removal of DOC (Humbert et al. 2005, Allpike et al. 2005, Fearing et al. 2004, Fabris et al. 2008) but there is a limited investigation into the fractionation of
58 orga nic matter in the regenerant brine following treatment (Mergen et al. 2009) After the laboratory experiments, the resin was regenerated and the brine analyzed to investigate how the different regeneration solutions impact the fractionation of organic matt er in the spent brine. Figure 3 5 shows the chromatogram from the analysis of the brine, with the y axis showing the absorbance and x axis showing the AMW on a logarithmic scale. These figures represent the organic matter fraction that is coming off the re sin after use, with the peak height relating to concentration of UV absorbing organic matter in solution, with the location of the peaks representing the size and properties of the organic matter fractions. Figure 3 5a shows the waste regenerant brine afte r the resin treated the Mount Pleasant water following the first regeneration experiment. It is clear that the 10% NaCl or 5% NaHCO 3 brine releases significantly more organic matter, with the primary peak seen in the regeneration at 1000 Da. Figure 3 5b sh ows the same information as Figure 3 5a, but represents the second regeneration experiment of the Mount Pleasant water. The second regeneration resulted in a wider range of AMW compounds coming off the resin, but overall a decreased intensity; the 0.3% NaC l and 0.4% NaHCO 3 regenerants more closely matched the 10% NaCl and 5% NaHCO 3 regenerants. A major difference seen in Figures 3 5 is the clear difference in the concentration of UV absorbing compounds coming off the resins between the different regeneratio n strengths. Figures 3 5 a,c, and d show a very significant decrease in the release of UV absorbing material when using the 0.3% NaCl and 0.4% NaHCO 3 regenerant solution, with figure 3 5b showing a less pronounced decrease. The likely cause for these dif ferences between the regeneration performance is due to insufficient regeneration of the resin, as these resins had been
59 subjected to multiple loadings, it was likely that in the WTP the resins were fully saturated and the lower strength of the regenerant solution was inadequate to fully recharge the resins. In comparing Figures 3 5a and c, there is a difference in the total UV absorbing materials coming off the resins, with the Mount Pleasant resin releasing much more UV absorbing material which implies t hat the Middle River resin was either more The difference in intensity seen in Figure 3 5a between first re generation and second regenerations is likely a result of the resin directly taken from the treatment plant being more fully saturated, whereas the second regeneration having a lower intensity is due to the limitations of one jar test to fully saturate the resin. Figures 3 5c and 3 5d show the same patterns with the Middle River resins and water. The results in Figures 3and 4show that overall during treatment the fraction of organic matter being removed, in descending order of peak intensity reduction, is t he 1000 1500 Da, the 500 700 Da, and the 200 300 Da range. This same trend is not what is seen during regeneration, during regeneration there are primarily two peaks, the 500 1500 Da peak and the 500 700 Da peak. The shift of peaks in Figure 3 5 implies th at either there is potentially a breakdown or compression of the lower molecular weight fraction into smaller fractions, or that the Ion exchange resins do show changes in affinity for diff erent ions depending on ionic strength of solution, so it is possible that the resins have very different affinities when the resin is being regenerated. Figure 3 5 also shows that only Cl at the 10% NaCl regenerant
60 concentration is able to generate the p eak at 100 Da, which is associated with the HPSEC salt peak, that elutes a mixture of smaller and mid sized organic fractions that cannot be categorized particularly by size (Swift and Posner 1971) however the elution of this mixed fraction implies that t his is the only regenerant that is able to fully treat the resin and reverse fouling, as none of the other regenerations are able to generate significant quantities of these organics. It is possible that the breakdown, or shift of organic matter in this hi gher ionic strength solution is only found with the 10% NaCl regeneration and is the cause for the improved regeneration of the resin. In addition to comparing the fractionation the number averaged ( M n ) and weight averaged ( M w ) molecular weights for the DO C was determined from the HPSEC data. M n and M w were derived by using the following equations as described by (Chin et al. 1994) : ( 3 1) ( 3 2) where h represents absorbance at an eluted volume and M represe nts the molecular weight of the analyte. Using these equations it was possible to normalize the M w with the M n organic matter fractions in solution. As polydispersity approac hes 1 the mass of fractions remaining in solution becomes more uniform. Table 3 2 shows a summary of the M w values for all first regeneration laboratory experiments, with additional data representing the brine from the second regeneration experiments All data for the second regeneration is available in Appendix A and shows similar trends to the first regeneration. M n values are not shown for brevity and can be inferred based on M w and
61 ctiveness of the treatment process to remove certain factions of organic matter. A drop in polydispersity is a good indicator of a treatment techniques ability to target a specific fraction of organic matter, as the polydispersity is used as a means to con vey the overall variation in the size range of the UV absorbing organic matter. Table 3 2 shows that there tends to be a slight increase in the polydispersity from the raw water to the treated waters. This confirms that the MIEX is targeting, not just one fraction of organic matter, but multiple fractions. Additionally, the brine has a significantly lower polydispersity than the raw and treated waters, implying that fractions of organic matter being removed by the resin are not coming off the resin during r egeneration but that regeneration is targeting more specific fractions. It is likely that certain fractions of organic matter are remaining on the resin it appears that these are the larger AMW (greater than 1500) compounds as the MW of the brine is of t he order between 500 1000, depending on the raw water source. In T able 3 2 it is shown that overall there are only small differences in the performance of the resin during the multiple loading jar tests, but there is a clear difference in the performance while comparing the brine. The Middle River brine shows no difference between the first and second regenerations, while all the Mount Pleasant brine shows a decrease in M w regeneration, with this decrease sho wing that after the second regeneration there are fewer large AMW fractions coming off the resin. The chloride brine generated by the Mount Pleasant resins tended to have a similar or greater M w value to the treated waters, suggesting that the higher AMW m aterial is more readily removed from the resin during regeneration. It can be seen in the second regenerations of the Mount Pleasant
62 chloride resin that there is less effective removal, due to the decrease in the M w of the brine, the same trends are not ap parent in the lower strength regenerant solutions. The lower strength regenerant solutions have M w lower than that of the treated water, which implies that the regeneration process is leaving the largest of the AMW fractions on the resin. The same trend of higher M w in the brine than the treated water no longer holds in the Middle River water. The Middle River water has a much higher M w than the Mount Pleasant water, which indicates a much greater portion of the larger AMW fractions, due to these greater po rtions of larger AMW fractions there is no significant decrease in M w values of treated water. What is seen in the Middle River brine is that the difference M w is much more pronounced between the weaker and the stronger regenerant concentrations. The lower M w seen in the regenerants is due to the much higher concentration of large AMW fractions in the raw water, with the resin being saturated by these large portions, once again the high strength regenerant is better able to remove the large AMW fractions fr om the resin, while the weaker regenerant solution is not. In fractions dominating the distribution, with the exception of the chloride brine at the The overall trend showed that weaker regenerant solutions had a slight tendency to remove fewer of the larger AMW compounds as exhibited by the decrease in the Mw. Similarly the second regenerations of the resins tend to remove fewer larger molecular weight fractions, overall implying that weaker regenerant solutions and multiple regenerations may not be fully regenerating the resins when they are in the presence of larger AMW fractions. These results are consisten t with findings elsewhere in the literature investigating the
63 fractionation of organic matter in brine, in that the larger AMW fractions tend to block access to resin sites (Mergen et al. 2008) 3.3.3 Comparison of B ench scale to F ull scale S ystems An imp ortant concern in water treatment is the limiting of resources used, while maintaining a production of quality potable water. This work investigated the impact that the composition and concentration of regeneration agent and the impact it had on performanc e of the resin. Table 3 3 shows the averages of water quality monitoring parameters at the two water treatment plants used in this study. The data was collected by plant operators and provided to the authors. The data from the experimental values was deriv ed by averaging the entirety of the data generated for a specific experiment type including each bed volume treated and all the regenerations. This allows for the direct comparison of lab scale performance with full scale performance. Table 3 3 shows that the data from the Mount Pleasant WTP agrees with the experimental data, in that the resins that were regenerated using the same regenerant as the plant achieved a DOC removal within approximately 5% of the full scale plants removal however there was a decr ease in DOC removal by 10 15% with the 0.3% NaCl and 0.4% NaHCO 3 regenerants. The data for Middle River does not agree as well as the Mount Pleasant data. This difference in performance is likely due to the small resin dosages used in the laboratory experi ments and the high concentration of DOC in the source water; in the full scale WTP resin dosages are much higher approximately 100 mL /L, through a continuous upflow clarifier. Table 3 3 shows that a reduction from a 10% NaCl solution to a 0.4% NaCl solut ion can achieve comparable performance which only yields an approximate 7% decrease in DOC removal in the higher DOC Middle River water. In the
64 lower DOC Mount Pleasant water the decrease in performance was closer to 15% when using the weaker brine. In sit uations where high DOC waters are being treated, it stands to reason that a decreased salt demand can benefit treatment systems greatly in cost and waste production. In situations where generating too much brine may be a concern, this can potentially allev iate some concern and allow for the use of ion exchange in situations where it was previously not a viable option. 3.4 Chapter Summary The conclusions of this work are: MIEX targets a wide range of organic matter fractions in natural waters regardless of regeneration agent, regeneration agent concentration, or DOC level. In waters with higher DOC levels, the largest peaks will have the largest removal due to favorability of that fraction and competition for charged sites. The MIEX resin tends to target th ree particular size ranges of organic matter, with the 1000 1500 Da and the 500 700 Da range the preferred regions, and the 200 300 Da range being the least preferred peak being reduced. This is likely due to the mobility of the size ranges, the aromatic s tructure of the larger range and the likely presence of carboxylic acid groups in the 500 700 Da range. The difference in performance between Cl or HCO 3 as regeneration agents are slight in regards to DOC removal. When averaging the performance of each f orm of the Cl form resin and the HCO 3 form of the resin there is a notic eable 3 5% decrease in performance with the HCO 3 resin. In some instances lower concentrations of regeneration agents may provide adequate levels of performance, thus decreasing bri ne treatment requirements. In this work it was shown that a reduction of regeneration agent concentration by almost 30 fold only yielded a 17% decrease in DOC removal, thus creating much less brine waste while removing significant amounts of DOC. The frac tions of UV absorbing organic matter in the brine following regeneration do not match the fractions removed during treatment. This implies that certain fractions 1000 1500 and 500 700 Da, of organic matter are remaining on the resin, which is likely the c ause of fouling. The concentration of Cl or HCO 3 as regeneration agents may also impact the efficiency of removing the organics during regeneration. It appears that the lower concentrations are even less effective at removing the larger MW material whic h
65 may then result in more rapid fouling of the resin. This would be an issue in source waters with higher MW organics. 3.5 A cknowledgments This material is based upon work supported by the National Science Foundation under Grant No.1209962, with addition al support provided by the Australian Academy of Science. The authors would like to thank SA Water operations staff for providing water, resin samples, and access to operational data. Additionally, the authors would like to thank Edith Kozlik, Paul Colby, and Martin Harris for their analytical support; and Rolando Fabris for additional support in analysis.
66 Table 3 1 Characteristics of raw water sources Characteristic River Murray (at MtP WTP) Middle River Reservoir (at MR WTP) DOC 1 6.8 mg/L 13.3 mg/L Turbidity 1 41 NTU 8 NTU Color 1 42 HU 150 HU UV 254 2 0.266 1/cm 0.594 1/cm Sulfate 1 12 mg/L 23 mg/L Bromide 1 0.11 mg/L 0.63 mg/L 1 Averaged values from water monitoring at treatment plants 2 Derived from laboratory experiments
67 Table 3 2. experiments Sample 1 M w MtP Cl 10 1 raw 1005 1.42 MtP Cl 10 1 200BV 860 1.63 MtP Cl 10 1 400BV 846 1.54 MtP Cl 10 1 600BV 830 1.53 MtP Cl 10 1 800BV 835 1.50 MtP Cl 10 1 Brine 738 1.20 MtP Cl 10 2 Brine 756 1.12 MtP Cl op 1 raw 1007 1.37 MtP Cl op 1 200BV 897 1.64 MtP Cl op 1 400BV 879 1.59 MtP Cl op 1 600BV 863 1.54 MtP Cl op 1 800BV 840 1.50 MtP Cl op 1 Brine 1160 1.50 MtP Cl op 2 Brine 958 2.13 Sample 1 M w MR Cl 10 1 raw 1607 1.74 MR Cl 10 1 200BV 1732 1.80 MR Cl 10 1 400BV 1698 1.81 MR Cl 10 1 600BV 1684 1.80 MR Cl 10 1 Brine 536 1.22 MR Cl 10 2 Brine 570 1.22 MR Cl op 1 raw 1527 1.74 MR Cl op 1 200BV 1722 1.84 MR Cl op 1 400 BV 1701 1.82 MR Cl op 1 600 BV 1673 1.82 MR Cl op 1 Brine 100 3 2.59 MR Cl op 2 Brine 1014 2.58 1 stoichiometrically determined and regenerated using 0.3% NaCl, and sample names containing Sample 2 M w MtP HCO3 10 1 raw 1005 1.42 MtP HCO3 10 1 200 BV 867 1.59 MtP HCO3 10 1 400 BV 846 1.53 MtP HCO3 10 1 600 BV 840 1.50 MtP HCO3 10 1 800 BV 841 1.49 MtP HCO3 10 1 Brine 699 1.19 MtP HCO3 10 2 Brine 647 1.23 MtP HCO3 op 1 raw 1007 1.37 MtP HCO3 op 1 200 BV 882 1.61 MtP HCO3 op 1 400BV 869 1.54 MtP HCO3 op 1 600BV 852 1.51 MtP HCO3 op 1 800BV 848 1.49 MtP HCO3 op 1 Brine 991 1.22 MtP HCO3 op 2 Brine 784 1.24 Sample 2 M w MR HCO3 10 1 raw 1607 1.74 MR HCO3 10 1 200 BV 1709 1.78 MR HCO3 1 0 1 400BV 1707 1.81 MR HCO3 10 1 600BV 1685 1.81 MR HCO310 1 Brine 549 1.21 MR HCO310 2 Brine 565 1.20 MR HCO3 op 1 raw 1527 1.74 MR HCO3 op 1 200BV 1688 1.79 MR HCO3 op 1 400 BV 1688 1.82 MR HCO3 op 1 600 BV 1667 1.82 MR HCO3 op 1 Brine 881 1.45 MR HCO3 op 2 Brine 902 1.44 2 stoichiometrically determined and regenerated using 0.4% NaHCO 3 and sample names NaHCO 3
68 Table 3 3. Change in DOC for full scale and laboratory scale treatments Avg DOC in Avg DOC out Avg % remaining MtP WTP 1 6.8 3.3 52 MtP M C 10x 8.75 5.80 66 MtP M C op 8.72 4.24 49 MtP M H 10x 8.75 6.04 69 MtP M H op 8.72 4.66 53 MR WTP 2 13.3 7.3 45 MR M C 10x 14.02 11.15 80 MR M C op 14.02 10.26 73 M R M H 10x 14.02 11.65 83 MR M H op 14.02 10.91 78 1 The n for the samples of DOC in is 24 and the n of samples for DOC out is 48 2 The n for the samples of DOC in is 14 and the n of samples for DOC out is 52
69 A) B) C) D) Figure 3 1 Effect of in creasing bed volume loading on DOC removal in Mount Pleasant water by resin regenerated in a) 10% NaCl; b) 5% NaHCO 3 ; c) 0.3% NaCl; and d) 0.4% NaHCO 3 for the first and second regeneration cycles. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 MtP 10% NaCl r1 MtP 10% NaCl r2 DOC C/C 0 Jar test batch 200 BV 400 BV 600 BV 800 BV 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 MtP 5% NaHCO3 r1 MtP 5% NaHCO3 r2 DOC C/C 0 Jar test batch 200 BV 400 BV 600 BV 800 BV 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 MtP 0.3% NaCl r1 MtP 0.3% NaCl r2 DOC C/C 0 Jar test batch 200 BV 400 BV 600 BV 800 BV 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 MtP 0.4% NaHCO3 r1 MtP 0.4% NaHCO3 r2 DOC C/C 0 Jar test batch 200 BV 400 BV 600 BV 800 BV
70 A) B) C) D) Figure 3 2 Effect of increasin g bed volume loading on DOC removal in Middle River water by resin regenerated in a) 10% NaCl; b) 5% NaHCO3; c) 0.3% NaCl; and d) 0.4% NaHCO3 for the first and second regeneration cycles. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 MR 10% NaCl r1 MR 10% NaCl r2 C/C 0 Jar test batch 200 BV 400 BV 600 BV 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 MR 5% NaHCO3 r1 MR 5% NaHCO3 r2 C/C 0 Jar test batch 200 BV 400 BV 600 BV 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 MR 0.3% NaCl r1 MR 0.3% NaCl r2 C/C 0 Jar test batch 200 BV 400 BV 600 BV 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 MR 0.4% NaHCO3 r1 MR 0.4% NaHCO3 r2 C/C 0 Jar test batch 200 BV 400 BV 600 BV
71 A) B) C) D) Figure 3 3 Molecular weight distribution b y HPSEC for resin and waters from Mount Pleasant WTP using 10% NaCl or 5% NaHCO 3 regeneration solution, A) first regeneration, and B) second regenerations; resin and water from Mount Pleasant WTP using 0.3% NaCl or 0.4% NaHCO 3 regeneration solution at the c) first regeneration, and d) second regenerations. Legend caption listed in order of appearance at 1000 Da from top to bottom. 0 0.005 0.01 0.015 0.02 10 100 1000 10000 Absorbance @ 260 nm (1/cm) Apparent Molecular Weight (Da) Mount Pleasant: 10% NaCl and 5% NaHCO 3 ; first regeneration RAW H-800 BV Cl-800 BV H-600 BV Cl-600 BV H-400 BV Cl-400 BV H-200 BV Cl-200 BV 0 0.005 0.01 0.015 0.02 10 100 1000 10000 Absorbance @ 260 nm (1/cm) Apparent Molecular Weight (Da) Mount Pleasant: 10% NaCl and 5% NaHCO 3 ; second regeneration RAW H-800 BV Cl-800 BV H-600 BV Cl-600 BV H-400 BV Cl-400 BV H-200 BV Cl-200 BV 0 0.005 0.01 0.015 0.02 10 100 1000 10000 Absorbance @ 260 nm (1/cm) Apparent Molecular Weight (Da) Mount Pleasant: 0.3% NaCl and 0.4% NaHCO 3 ; first regeneration RAW H-800BV Cl-800BV H-600BV Cl-600BV H-400BV Cl-400BV H-200BV Cl-200BV 0 0.005 0.01 0.015 0.02 10 100 1000 10000 Absorbance @ 260 nm (1/cm) Apparent Molecular Weight (Da) Mount Pleasant: 0.3% NaCl and 0.4% NaHCO 3 ; first regeneration RAW H-800 BV Cl-800 BV H-600 BV H-400 BV Cl-600 BV Cl-400 BV H-200 BV Cl-200 BV
72 A) B) C) D) Figure 3 4 Molecular weight distribution by HPSEC for resin and waters from Middle River WTP using 10% N aCl or 5% NaHCO3 regeneration solution at the A) First regeneration, and B) second regeneration; resin and water from Middle River WTP using 0.3% NaCl or 0.4% NaHCO3 regeneration solution at the C) first regeneration, and D) second regenerations. Legend ca ption listed in order of appearance at 1000 Da from top to bottom. 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 10 100 1000 10000 Absorbance @ 260 nm (1/cm) Apparent Molecular Weight (Da) Middle River: 10% NaCl and 5% NaHCO 3 ; first regeneration RAW H-600 BV Cl-600 BV H-400 BV Cl-400 BV H-200 BV Cl-200 BV 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 10 100 1000 10000 Absorbance @ 260 nm (1/cm) Apparent Molecular Weight (Da) Middle River: 10% NaCl and 5% NaHCO 3 ; second regeneration RAW H-600 BV Cl-600 BV H-400 BV Cl-400 BV H-200 BV Cl-200 BV 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 10 100 1000 10000 Absorbance @ 260 nm (1/cm) Apparent Molecular Weight (Da) Middle River: 0.3% NaCl and 0.4% NaHCO 3 ; first regeneration RAW Cl-600 BV H-600 BV H-400 BV Cl-400 BV H-200 BV Cl-200 BV 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 10 100 1000 10000 Absorbance @ 260 nm (1/cm) Apparent Molecular Weight (Da) Middle River: 0.3% NaCl and 0.4% NaHCO 3 ; second regeneration RAW Cl-600 BV H-600 BV H-400 BV Cl-400 BV H-200 BV Cl-200 BV
73 A) B) C) D) Figure 3 5 Molecular weight distribution by HPSEC for A) Mount Pleasant regeneration brine from the first regeneration, B) Mount Pleasant regeneration brine from the second regeneration, C) Middle River regeneration brine from the first regeneration, and D) Middle River regeneration brine from the second regeneration. Legend caption listed in order of appearance at 1000 Da from top to bottom. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 10 100 1000 10000 Absorbance @ 260 nm (1/cm) Apparent Molecular Weight (Da) Mount Pleasant brine, first regeneration 10% NaCl 5% NaHCO3 0.3% NaCl 0.4% NaHCO3 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 10 100 1000 10000 Absorbance @ 260 nm (1/cm) Apparent Molecular Weight (Da) Mount Pleasant brine, second regeneration 10% NaCl 5% NaHCO3 0.3% NaCl 0.4% NaHCO3 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 10 100 1000 10000 Absorbance @ 260 nm (1/cm) Apparent Molecular Weight (Da) Middle River brine, first regeneration 10% NaCl 5% NaHCO3 0.3% NaCl 0.4% NaHCO3 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 10 100 1000 10000 Absorbance @ 260 nm (1/cm) Apparent Molecular Weight (Da) Middle River brine, second regeneration 10% NaCl 5% NaHCO3 0.3% NaCl 0.4% NaHCO3
74 CHAPTER 4 THE ROL E OF DI VALENT CATIONS IN THE FOULING OF AN ION EXCHANGE RESINS 4.1 Background The removal of natural organic matter (NOM) from drinking water is a growing area of concern as NOM impacts many aspects of the water treatment processes, and by reacting with disinfect ants, forms disinfection by products (Krasner et al. 2006). NOM in natural waters behaves as an anionic species due to the deprotonation of carboxylic acid functional groups (Boyer and Singer 2008, Collins et al. 1986) which makes the removal of NOM by ani on exchange possible. The removal of NOM by anion exchange is well documented (Bolto et al. 2002, Boyer and Singer 2006, Boyer et al. 2008, Drikas et al. 2007, 2011, Karpinska et al. 2013) and this work focus es on the use of magnetic ion exchange (MIEX) re sin, which has been shown to be very effective at removing dissolved organic carbon (DOC) from drinking water. Unfortunately when using anion exchange, the regeneration of the resin produces concentrated salt brines which are difficult to regenerate which has led to the investigation of the use of NaHCO 3 as a regenerant rather than NaCl (Rokicki and Boyer 2011). While the use of bicarbonate as the mobile ion appears to be effective, it seems that the bicarbonate form tends to lose its capacity more rapidly than the chloride form over multiple resin life cycles (Walker and Boyer 2011). It appears that the increased concentration of bicarbonate may be contributing to the fouling of the resin in natural water systems In laboratory based anion exchange reaction s, it has been shown that ion exchange reactions will undergo fouling with multiple regeneration cycles (Mergen et al. 2008, Walker and Boyer 2011), which is described as the loss of capacity of the resin.
75 Of particular interest in regards to bicarbonate f orm resins is the potential for inorganic fouling to occur by means of scale formation of the resin. Inorganic scale formation is a well investigated phenomenon in membrane processes (Jung and Son 2009, Kimura et al. 2004, Mitsoyannis and Saravacos 1977), but is not readily investigated in ion exchange processes. It is believed that the precipitation of minerals on membranes increase the extent of fouling (Bhattacharjee and Johnston 2002, Mitsoyannis and Saravacos 1977). The precipitation of minerals from a queous solution is a process that is governed by the solubility product of the mineral (Benjamin 2002). This work investigated the impact that magnesium, calcium, and cadmium had on the fouling of the bicarbonate form anion exchange process. Table 4 1 list s the solubility products of the likely carbonate mineral to form using the selected cations, with magnesite being the most soluble and otavite being the least soluble The selected divalent cations not only have the potential to form precipitates, but i n teraction with NOM is likely to occur (Dzombak et al. 1986, Kinniburgh et al. 1999). Calcium is understood to readily form complex es with NOM, while magnesium and cadmium are also known to complex with NOM, they do so to a lesser extent (Kalinichev and Kir kpatrick 2007, Leenheer et al. 1998). The ability of these cations to complex with NOM can play greatly into the mobility of the cations. The more mobile the cation, the greater the potential of the divalent cation to enter the por ous resin structure and f oul the resin. It is expected that cations would be excluded from the resin structure due to Donnan exclusion principles (Cumbal and Sengupta 2005, Sarkar and SenGupta 2010 ): however while complexed with NOM, or in the hydroxide form, the cations would be able to enter into the resin structure.
76 While research does show that the fouling of resins do occur (Rokicki and Boyer 2013, Walker and Boyer 2011), it is not entirely clear what the mechanism is that causes the fouling of bicarbonate form resins. The precipitation of carbonate minerals is what would appear to be the most likely cause of fouling of the resin. It is unclear whether fouling occurs as precipitation on the resin, or if it is caused by steric interferences with the resin structure. Any steri c interference would need to be able to overcome Donnan exclusion to enter into the resin structure to remove access to charged sites. Overall, it is unknown how divalent cations contribute to the fouling of bicarbonate form anion exchange, and by which me chanisms the fouling occurs. The hypothesis of this work was that the solubility of carbonate minerals would govern the inorganic fouling that occurs on bicarbonate form anion exchange resin, with less soluble species fouling the resin most. The specific objectives of this research were to: (1) d etermine the effect of divalent metal cations on the ability of bicarbonate form anion exchange resin to remove DOC over multiple regeneration cycles; (2) evaluate the impact of divalent cations on bicarbonate form anion exchange in the presence of nitrate in lieu of NOM; and (3) determine whether the anion exchange process is likely to alter water chemistry enough to trigger precipitation. Anion exchange resins in the bicarbonate or chloride form were used to trea t synthetic waters that contained Suwannee River NOM isolate and either: Mg 2+ Ca 2+ or Cd 2+ or nitrate with one of the divalent cations of interest. After treatment, the resin was regenerated using either NaCl or NaHCO 3 and used to treat another batch o f the same synthetic water, with this process being repeated for a total of four regeneration cycles.
77 4.2 Materials and Methods 4.2.1 Materials Synthetic waters were developed specifically to keep competition between the different ions and resin consiste nt in the different waters using different cations. Three waters were designed each using either: Mg 2+ Ca 2+ or Cd 2+ as the divalent metal cation. Waters were designed so that each ion in water had the same equivalence charge as the other constituents. Us ing a same equivalence for the constituents of the waters minimized the extent to which concentration difference of constituents drove anion One deviation from the con sistent equivalence charge was in the Cd 2+ derived water, in that the chloride concentration was higher, as listed in Table 4 2, due to solubility of cadmium salts available. Waters were created by weighing out dosages of American Chemical Society (ACS) ce rtified salts of sodium bicarbonate and sodium chloride, Suwannee River NOM isolate (Catalog number: 1R101N ) from the International Humic Substances Society, and either ACS grade MgO, CaO, or CaCl 2 Sodium chloride was not added to the Cd 2+ synthetic water because chloride was added from the cadmium salt. Magnetic anion exchange resin (MIEX, Orica Watercare) was used in this work. MIEX resin was used in two different presaturant ion forms: chloride or bicarbonate. Prior to the first experiment, virgin MIEX resin was regenerated, as described in 4. 2.3 to convert to the desired presaturant ion.
78 4.2.2 Experimental methods All experiments were performed on a Phipps & Bird PB 700 jar tester, using stainless steel paddles, 2 L glass beakers containing 2000 mL of test water, with all conditions tested in triplicate. 4. 2.3 Ion E xchange R egeneration Two regeneration agents were used in this investigation, a (1) sodium chloride (NaCl) and (2) a sodium bicarbonate (NaHCO 3 ) based solution. These two agents were used to generate a resin with either the chloride or bicarbonate as a mobile counterion. The concentration of the regenerant was determined by calculating a stoichiometric excess based on the ion exchange capacity of the MIEX resin this concentration was calc ulated by determining the equivalents capacity of the dose of resin in each sample, and multiplied by 10 to drive equilibrium. Regenerants were made using either a solution of NaCl or NaHCO 3 each 1 mL dose of resin had regenerant dosed with 1 L of 3.04 mg /L NaCl or 1 L of 4.37 mg/L of NaHCO 3 Resins were mixed in the regenerant solution at 100 rpm for 30 min, once mixing was complete the resins were allowed to settle, and the used regenerant was decanted. 1 L of DI water was added to the resin and mixed fo r 10 min at 100 rpm allowed to settle and decanted, with this DI water rinse step being completed one more time. 4. 2.4 Ion E xchange T reatment Treatment experiments were completed following regeneration of the resin. One milliliter of MIEX resin was added to 2 L of synthetic water and mixed at 100 rpm. Mixing was stopped at 5, 15, 30, and 120 min; the resin was allowed to settle and 50 mL of supernatant was drawn using a syringe before resuming the mixing. All samples were
79 syringe filtered using millex PVD F 0.45 m filter tips and analyzed for pH. Depending on the water characteristics, samples were also analyzed for DOC, DIC, inorganic anions, or inorganic cations. Bench scale experiments with data shown as percent removal compared to the raw water had s t atistical errors calculated using E quation 4 1, where s represents the standard deviation, N represents the sample size, c o represents the initial concentration, and c represents the concentration at that sample point. Error associated with samples present ed as absolute concentrations were represented with standard deviation. ( 4 1) 4. 2 5 Analytical M ethods All samples were filtered using 0.45 m sterile PVDF membranes. Immediately following filtration, the pH of all samples were measured using an Accumet AB15 pH meter with probe that was calibrated using pH 4, 7, and 10 buffer solutions. DOC and DIC were analyzed using a Shimadzu TOC V CPH total organic carbon analyzer as described by Rokicki and Boyer ( 2011 ) Chloride, nitrate, magnesium, and calcium values were measured on a Dionex ICS 3000 ion chromatograph. During chloride and nitrate analysis the ICS 3000 was equipped with an IonPac AG22 guard column and AS22 analytical column running Na 2 C O 3 /1.4 mM NaHCO 3 eluent solution. During calcium and magnesium analysis the ICS 3000 was equipped with an Ion P ac CG12A guard column and and IonPac CS12A analytical column running 0.02 M methanesulfonic acid eluent solution. All analysis was compared to cal ibration
80 standards and checks using ACS reagent grade purity salts in DI water. Cadmium samples were sent to the Analytical Services Laboratory located at the University of Florida, for inductively coupled plasma analysis. All TOC and IC samples were analy zed in duplicate for precision, with all duplicates being found acceptable if they were within 5% difference between the duplicates Check standards were used at the end of each run to confirm the accuracy of calibration curves, for both the TOC and IC, an d were always within 10% variability. In addition to check standards the TOC analyzer utilized outside check standards from Ricca Chemical were confirmed to be within 10% 4. 2.6 Visual MINTEQ Generated model waters were analyzed using Visual MINTEQ versi on 2.61 to compare to the actual experiments and determine whether or not precipitation would occur in the waters following the anion exchange process. To prevent difficulties associated with modeling the complex nature of NOM and its interactions with div alent cations, analysis was only carried out using the waters containing nitrate. For all runs the system was set to open atmospheric CO 2 pressure (0.00038 atm), and bicarbonate was entered as equal composition of CO 3 2 and pH was fixed at the measured pH values of the waters used in experiments, with the 120 min time sample value being used for post ion exchange pH values. Remaining ion concentrations were entered in as their representative ionic concentration as millimola r units, with concentrations bein g listed in Table 4 3. The baseline composition was set by the theoretical composition of the water, as briefly described, with the post anion exchange water composition being calculated based on adding the total chloride or bicarbonate capacity of the res in to the solution. With MIEX resin having approximate capacity of 0.5 meq/m L and each dose
81 of resin was 1m L 0.5 meq of chloride was added to model the post chloride resin waters, and 0.5 meq of bicarbonate was added to model the post bicarbonate resin w aters. Each post anion exchange model water was run twice, once with no precipitation allowed, and once again with precipitation to quantify the amount of potential precipitates On initial runs, precipitation was restricted to determine the saturation ind ex of each species in solution, to identify those species most likely to form, once these species were identified, the models were set to allow them to precipitate. 4.3 Results and Discussion 4. 3.1 Effect of D ivalent M etal C ations on T arget A nion R emova l 4. 3.1.1 DOC removal Figure 4 1 shows the percent DOC removal in each of the model waters containing Suwannee River NOM isolate, plotted against the mixing time of the resin for each of the resin regeneration cycles being labeled as R1, R2, R3, or R4 wit h the numeral integer representing the total number of regeneration cycles the resin underwent. Figure 4 1a) shows the DOC removal for all chloride form resins in the m agnesium water as described in Table 4 2. Figure 4 1a shows that at the longest time int erval, the resin is able to achieve 48% DOC removal, and gradually decreases over multiple regeneration cycles to approximately 33% DOC removal. This gradual loss of capacity of the resin at our sample point is likely due to a combination of the incomplete regeneration of the resin that was described by Rokicki and Boyer ( 2013 ) and fouling of the resin Figure 4 1b, shows the same model water as Figure 4 1a and how it behaves when treated by bicarbonate form resin. The first regeneration of the bicarbonate form
82 has a reduction in DOC removal when compared to the first regeneration in the chloride form resin with bicarbonate form achieving 33% removal and chloride form achieving 48% It appears that the interaction of magnesium and bicarbonate have an immedi ate negative impact on the performance of the removal of DOC by anion exchange, while previous work has shown that bicarbonate form resin has the same affinity for DOC as the chloride form (Rokicki and Boyer 2011) meaning that the decrease in performance i s caused by the interactions of magnesium with the resin. Figure 4 1b also shows that over multiple uses the resin performance continues to drop and approach 23% DOC removal. Figures 4 1c,d) show the performance of the chloride and bicarbonate form resins respectively, in the calcium water. In Figure 4 1c the first regeneration of the resin is able to achieve 47% DOC removal at 120 min but all the other subsequent regenerations fall to approximately 30% removal. Figure 4 1d shows that the bicarbonate for m resin achieves 41% DOC removal with a more gradual decrease in performance over multiple regenerations to 25% removal. Figures 4 1c,d do show a slight difference in performance between the chloride and bicarbonate forms, it is not as significant as the m impact of calcium and magnesium is not as expected, as magnesite is the most soluble of the likely minerals to precipitate of the selected cations. The most likely cause of fouling of the resin may be the formation of Mg(OH) 2 which would allow for the magnesium to interact directly with the pores of the resin. In comparison to the calcium and magnesium waters, the cadmium appears to have the least impact on the fouling of the resin. Figure 4 1d,e shows th e chloride and bicarbonate form in the cadmium containing synthetic water. The performance of the resins in Figure 4 1d,e are
83 essentially the same with the chloride form resin achieving the highest DOC removal at 120 min of 46% and the lowest of 28% while the bicarbonate form achieves 49% DOC removal at its first regeneration and 31% DOC removal at the last regeneration. The hypothesis that the cadmium water would cause the most fouling of the bicarbonate water, followed by calcium, with magnesium fouling the water the least does not appear to be holding, instead the opposite is occurring and magnesium is causing the most resin fouling. 4. 3.1.2 Divalent metal cation removal Figure 4 1 showed that the presence of divalent cations do have an impact on the p erformance of the bicarbonate form resin. It stands to reason that with any decrease in performance from the chloride form resin when compared to the bicarbonate form resin, there should be a noticeable difference in the divalent cation concentration. An i ncrease in the cation removal should occur if the fouling of the resin is occurring due to inorganic scale. Figure 4 2 shows the residual concentration of magnesium for each sample time, plotted against the number of regeneration cycles the resin underwent for both the chloride and bicarbonate form resin. Figure 4 2a shows the chloride form of the resin, which shows minimal difference in magnesium concentration for the first regeneration, a slight removal in the second regeneration, with slight leaching of magnesium off the resin in the third and fourth regenerations at the 120 min sampling point. Figure 4 2b shows the magnesium removal by the bicarbonate form resin. Figure 4 2b shows that there is little removal overall in the first, second or third regene ration series, but in the fourth regeneration the resin leaches magnesium into the bulk solution. It is likely that the magnesium is interacting with the resin or
84 organic matter initially, with the later regenerations eventually desorbing magnesium f rom th e resin and releasing it into the bulk water solution. It is clear that the magnesium is interacting with either the resin, NOM, or the bicarbonate, but the exact mechanism is unclear as it does not appear to be precipitation due to the magnesium concentra tion not decreasing. The performance of the resins in the presence of calcium is more straight forward. Figure 4 3a,b is configured the same as Figure 4 2, and shows that the percent removal of calcium is 10%, with the exception of the 15 min sample for the first regeneration of the bicarbonate form resin. The mobile ion of the resin in the presence of calcium ultimately does not drastically impact the removal of calcium, with no calcium removal there can be no precipitation of calcium minerals. Figure 4 4a,b, which is arranged as Figures 4 2 and 4 3, show the cadmium removal achieved in the waters treated by the chloride and bicarbonate form resins respectively. In Figure 4 4a, it can be seen that the first three regenerations of the chloride form resin have a slight increase in the cadmium concentration at 120 min while, but subsequently in the fourth regeneration there is a removal of cadmium during the 5, 15, and 30 min sampling points, but by the 120 min sampling point the solution equilibrates and th ere is no net removal noticed. Compared to the bicarbonate form resin shown in Figure 4 4b, in which the first three regeneration cycles achieve minimal change in cadmium removal, and the last regeneration releases cadmium. It appears that the cadmium conc entration increases to a value that is greater than the initial concentration during treatment, while this may be reasonable in F igure 4 4b, assuming that the resin is able to accumulate cadmium in the pore structure, and release it after several treatment s, this accumulation
85 is not seen as decreased DOC removal in Figure 4 1f. It is also unlikely that the resin would be able to accumulate free cations due to Donnan exclusion processes (Cumbal and Sengupta 2005) This change in divalent cation concentration is likely due to initial precipitation of the cadmium into the bulk solution, and as the concentration of chloride and bicarbonate ions change through the anion exchange process, the equilibrium shifts and more cadmium dissolves. The presence of NOM and divalent cations definitely appear to have an impact on the fouling of bicarbonate form anion exchange resins, with magnesium having the greatest impact, as evaluated by decrease in DOC removal. In the presence of all divalent cations there is some extent of decreased DOC removal experienced, but in none of the waters were there significant decreases in metal cations to be able to conclude that the fouling was caused by precipitation. pH was recorded for all samples and is re port ed in Appendix B. Figure 4 1 and it shows that the pH for the bicarbonate form and the chloride form waters were similar, but the divalent cation impacted the pH. Waters containing magnesium were found to have a pH between 7 and 8, while the calcium and cadmium waters had pH values p rimarily in between 6 and 7. 4. 3.2 Effect of D ivalent M etal C ations on T arget A nion R emoval in the A bsence of NOM 4. 3.2.1 Nitrate removal It is well documented that organic matter has a tendency to interact with cations in waters (Kinniburgh et al. 1996 Kinniburgh et al. 1999) and these interactions may somehow be impacting the performance of bicarbonate form resin. To eliminate any concerns associated with the interaction of the divalent cations with natural organic
86 matter, synthetic waters using nitra te as the targeted anion for removal were used. To determine the cause of resin fouling, it was important to consider anion exchange in which the only source of potential fouling was from the divalent cation interacting with the other anions and the resin. To achieve this nitrate was used in lieu of NOM as the targeted anion, so that organic interactions with the resin or divalent cations would not interfere with the performance. Figure 4 5 shows the nitrate removal from the various model waters that used n itrate in lieu of NOM as the targeted anion with each of the divalent cations for the multiple regenerations of the resin. Nitrate was selected because previous work showed that MIEX has a higher affinity for the nitrate than the chloride or bicarbonate an d will selectively remove the nitrate from waters (Rokicki and Boyer 2011) Any decrease in nitrate removal between chloride and bicarbonate form resin can be attributed to inorganic fouling of the resin. Figure 4 5a shows the chloride form resin in the s ynthetic water containing nitrate and magnesium for each regeneration cycle. In Figure 4 5a, each regeneration cycle peaks at 15 min, with the first regeneration cycle having the highest nitrate removal with a peak of 17% removal, and a decline to 12% remo val at the 120 min time interval. Each subsequent regeneration cycle follows the same pattern, with a consistent decline in performance with the final regeneration achieving only 8% nitrate removal at the 120 min sampling point. This peak in nitrate remova l signifies that the resin is removing the nitrate initially, but eventually desorbing the nitrate as the experiment progresses. The nitrate peak removal and subsequent desorbing is likely related to shift in ion exchange equilibria and is discussed subseq uently with the cation removal performance. A similar trend is seen in Figure 4 5b in which the bicarbonate form of the resin is used to treat a water
87 containing nitrate and magnesium. Figure 4 5b shows that the bicarbonate form of the resin achieves much lower nitrate removal almost immediately when compared to the chloride form. The bicarbonate form of the resin follows the same removal trend, in which the highest removal is seen at 30 min, with the nitrate desorbing after this point. The highest removal achieved by the bicarbonate form resin in the presence of magnesium was approximately 8% nitrate removal, which was at the first regeneration step, while the chloride form resin achieved 8% nitrate removal at its worst performance on the final regeneratio n cycle showing that the bicarbonate in the magnesium water has a clear impact on the resin and is fouling the resin in the form of inorganic scale. The trend of desorbed nitrate was not a continued observation and was only noticed in the magnesium waters In Figure 4 5c,d chloride form and bicarbonate form resin are shown in the presence of the calcium ion for all regeneration cycles, and both resins behave as expected with percent nitrate removal increasing until 30 min and performance leveling off, rath er than decreasing. The increase of removal until a maximum removal is obtained is standard for most anion exchange processes and is expected (Helfferich 1962) Figure 4 5c shows that the final nitrate removal for all regeneration cycles of the chloride fo rm resin falls between 23 % and 31%, while Figure 4 5d, which shows the bicarbonate form of the resin in the calcium water, only achieves 21 % to 24% nitrate removal. This slight decrease in performance between the resin forms, shows that there is some sligh t inorganic fouling that is occurring, but not to the same extent that the magnesium was fouling the resin. The final portions of Figure 4 5, which includes Figure 4 5e,f show the chloride and bicarbonate form of the resin in the nitrate water that conta ins cadmium. This figure
88 closely resembles the performance of the resin in Figures 4 5c,d, in that the resin achieves its max performance after 30 min and levels off after that, with each regeneration cycle decreasing the overall max performance slightly. In Figure 4 5e, the chloride form achieves a maximum nitrate removal of 36%, which decreases to 29% nitrate removal at the end of the regeneration cycles. In Figure 4 5f, the bicarbonate form resin experienced a slight decrease in nitrate removal compared to the chloride form resin. The bicarbonate form resin in the presence of cadmium was only able to achieve a maximum nitrate removal of 30%, while over the course of the regeneration cycles it decreased to 27%. The decrease of the nitrate removal in the ca dmium was a clear indicator that there was some inorganic fouling that was occurring. All of the synthetic waters tested experienced some extent of inorganic fouling, with magnesium being the most drastically fouled by inorganic precipitation. 4. 3.2.2 Div alent metal cation removal Figure 4 6 shows the concentration of magnesium for each sample time in the waters containing magnesium and nitrate plotted against the number of regeneration cycles that the resin underwent for both the chloride and bicarbonat e form resin. Figure 4 6a shows the change in magnesium in the presence of the chloride form resin which in all regenerations shows a steady increase in magnesium concentration as the contact time increases. Figure 4 6b shows the same trend with the bicar bonate form of the resin. This data appears to be rather anomalous and is not believed to be related to instrumental error due to calibration checks being within 10% of their value, it is reasonable to assume that due to the low concentration of the sample s that instrument signal interfered with analysis. Figure 4 6 is showing a net increase of magnesium,
89 rather than a decrease further supports that precipitation is not occur r ing on the resin, but the precipitation may have occurred in the bulk solution, wi th ion exchange allowing the magnesium to become more soluble Figure 4 7 is configured as Figure 4 6, showing the calcium concentrations in the nitrate and calcium water. Both Figure 4 7a, b show that the chloride and bicarbonate form respectively show t hat the concentration of calcium remains relatively consistent across all regenerations. Figure 4 8 a, b show the concentration of cadmium for the nitrate model waters both Figure 4 8 a, b show that over the course of the ion exchange experiments, the cad mium concentration is increasing. As with Figure 4 6 this is likely due cadmium and magnesium becoming more soluble during ion exchange. The divalent cation concentration data further supports the idea that precipitation is not necessarily occur r ing on the resin, but the divalent cations have a clear impact on the performance of the resin, which can be seen when comparing the first regeneration of the chloride form resins to the bicarbonate form resins. It appears that the cadmium and magnesium may be preci pitating in the bulk water solution, and overtime as the ion exchange process occurs, the equilibrium of the bulk water solution shifts and more of these divalent cations are dissolving in the bulk water solution. This is likely caused by different speciat ion occurring with the addition of chloride and bicarbonate to the bulk solution. When looking at the pH values in Appendix B it is shown that the magnesium waters have a pH around 10, while the calcium water starts at 9 and decreases to 7, and the cadmiu m water stays in between 6 and 7. The magnesium interactions with the water and the resin are clearly the driving force in the fouling of the resin, and is likely related to the pH shift.
90 4. 3.2.3 Chemical speciation modeling Visual MINTEQ modeling shown in Table 4 3, show that in the theoretical post ion exchange waters some species approach saturation. In the waters containing calcium, both the post chloride and post bicarbonate waters would be likely to precipitate a ragonite, c alcite, and v aterite, but there is a clearly increase in the saturation index in the bicarbonate waters, showing that these minerals are more likely to precipitate. In the magnesium model waters, the most likely precipitate is magnesite, for both the post chloride and post bicarbon ate treatments, with the bicarbonate water once again having the higher saturation index. These model results conflict with the data seen in the bench scale experiments, both the NOM containing water and the nitrate containing water experienced increased f ouling in the presence of bicarbonate, while theoretically the magnesium waters should be the least likely to foul. The most likely reason that fouling is occurring is that the magnesium is likely forming magnesium hydroxide that is interacting with the re sin structure. The only water that ha d a positive saturation index, and therefor was likely to form a precipitate wa s the cadmium water following bicarbonate form anion exchange which shows that otavite is likely to precipitate in both the post chloride a nd pos t bicarbonate waters. The model w as run again, with precipitation allowed and otavite precipitated in both waters showing that the increased concentration of bicarbonate would increase the amount of precipitate. This data also conflicts with the exp erimental data that was seen, the cadmium waters tended to have the least amount of impact on performance, which implies that this precipitation is not occurring on the resin surface. The precipitation occurring in the bulk water is not impacting the resin and is not fouling the resin, this is likely due to otavite being readily able to precipitate in the bulk water solution, where the ionic potential is
91 much lower than the resin phase aqueous solution which has a much higher ionic strength. It is possible that a majority of the cadmium is precipitating out of the bulk water solution, rather than interacting with the resin, thus decreasing the likelihood of cadmium interacting with the resin, as opposed to the magnesium waters, where it is likely the magnesi um is forming other complexes and interfering with the resin. In the absence of NOM, it does appear that the bicarbonate form of the resin does experience more fouling than the chloride form. In the presence of all divalent cations the performance decrease d on the first regeneration cycle and often to a greater extent than the NOM containing samples. It is likely that without the NOM interactions that the divalent cations are directly interacting with the resin. As supporting T able 4 2 shows, most pH fell w ithin typical raw water range of 6 9, with the exception of the magnesium water which had a higher pH. 4.4 Chapter Summary In most waters containing NOM, there was an almost immediate decline in performance in regards to DOC removal in the presence of div alent cations with the exception of cadmium. There was no decrease in divalent cation concentration to support the hypothesis that the divalent cation precipitation was the cause of resin fouling. Magnesium appeared to have the greatest impact on the perf ormance of anion exchange, and this was likely due to magnesium speciation in water. Mg(OH)2 was likely to form and would be able to enter the pore structure of the resin, thus blocking access to resin surface sites. Cadmium, while models state is likely foul the resin. Precipitation is likely occurring in the bulk solution and not impacting the resin.
92 A) B) C) D) E) F) Figure 4 1 Kinetic plots showing percent DOC removal over time for e ach resin in either the chloride or bicarbonate form in the presence of either: Mg, Ca, or Cd in the NOM containing water. Arranged in order of decreasing pK sp 0% 10% 20% 30% 40% 50% 0 50 100 % DOC removal time (min) R1 R2 R3 R4 chloride form Mg 0% 10% 20% 30% 40% 50% 0 50 100 % DOC removal time (min) R1 R2 R3 R4 bicarbonate form Mg 0% 10% 20% 30% 40% 50% 0 50 100 % DOC removal time (min) R1 R2 R3 R4 chloride form chloride form Ca 0% 10% 20% 30% 40% 50% 0 50 100 % DOC removal time (min) R1 R2 R3 R4 bicarbonate form Ca 0% 10% 20% 30% 40% 50% 0 50 100 % DOC removal time (min) R1 R2 R3 R4 chloride form Cd 0% 10% 20% 30% 40% 50% 0 50 100 % DOC removal time (min) R1 R2 R3 R4 bicarbonate form
93 A) B) Figure 4 2 Absolute m agnesium concentration for each resin in either the A) chlo ride or B) bicarbonate form in the synthetic water containing NOM and magnesium sampled at 0, 5, 15, 30, 120 min. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 R1 R2 R3 R4 Magnesium (mg/L) regeneration series chloride form Magnesium 0 min 5 min 15 min 30 min 120 min 0 0.2 0.4 0.6 0.8 1 1.2 1.4 R1 R2 R3 R4 Magnesium (mg/L) regeneration series bicarbonate form Magnesium 0 min 5 min 15 min 30 min 120 min
94 A) B) Figure 4 3 Absolute c alcium concentration for each resin in either the A) chloride or B) bicarbonate form in the synthetic water containing NOM and calcium sampled at 0, 5, 15, 30, 120 min. 0 0.5 1 1.5 2 2.5 R1 R2 R3 R4 Calcium (mg/L) regeneration series chloride form Calcium 0 min 5 min 15 min 30 min 120 min 0 0.5 1 1.5 2 2.5 3 3.5 R1 R2 R3 R4 calcium (mg/L) regeneration series bicarbonate form Calcium 0 min 5 min 15 min 30 min 120 min
95 A) B) Figure 4 4 Absolute c admium concentration for each resin in either the A) chloride or B) bicarbonate form in the synthetic water containing NOM and Cadmium sampled at 0, 5, 1 5, 30, 120 min. 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 R1 R2 R3 R4 Cadmium (mg/L) regeneration series chloride form Cadmium 0 min 5 min 15 min 30 min 120 min 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 R1 R2 R3 R4 Cadmium (mg/L) regeneration series bicarbonate form Cadmium 0 min 5 min 15 min 30 min 120 min
96 A) B) C) D) E) F) Figure 4 5 Kinetic plots showing percent nitrate removal over time for each resin in either the chloride or bicarbonate form in the presence of either: Mg, Ca, or Cd in the nitrate containing water. Arra nged in order of decreasing pK sp 0% 10% 20% 30% 40% 0 50 100 % NO3 removal time (min) R1 R2 R3 R4 chloride form chloride form 0% 10% 20% 30% 40% 0 50 100 % NO3 removal time (min) R1 R2 R3 R4 bicarbonate form Mg 0% 10% 20% 30% 40% 0 50 100 % NO3 removal time (min) R1 R2 R3 R4 chloride form Ca 0% 10% 20% 30% 40% 0 50 100 % NO3 removal time (min) R1 R2 R3 R4 bicarbonate form Ca 0% 10% 20% 30% 40% 0 50 100 % NO3 removal time (min) R1 R2 R3 R4 chloride form chloride form Cd 0% 10% 20% 30% 40% 0 50 100 % NO3 removal time (min) R1 R2 R3 R4 bicarbonate form Cd
97 A) B) Figure 4 6 Absolute m agnesium concentration for each resin in either the A) chloride or B) bicarbonate form in the synthetic water containing nitrate and magnesium sampled at 0, 5, 15, 30, 120 min. 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 R1 R2 R3 R4 Magnesium (mg/L) regeneration series chloride form NO3/Mg 0 min 5 min 15 min 30 min 120 min 0 1 2 3 4 5 6 R1 R2 R3 R4 Magnesium (mg/L) regeneration series bicarbonate form NO3/Mg 0 min 5 min 15 min 30 min 120 min
98 A) B) Figure 4 7 Absolute c alcium concentration for each resin in either the A) chloride or B) bicarbonate form in the synthetic water containing nitrate and calcium sampled at 0, 5, 15, 30, 120 min. 1.5 1.6 1.7 1.8 1.9 2 2.1 R1 R2 R3 R4 Calcium (mg/L) regeneration series chloride form NO3/Ca 0 min 5 min 15 min 30 min 120 min 1.55 1.6 1.65 1.7 1.75 1.8 1.85 1.9 1.95 2 R1 R2 R3 R4 Calcium (mg/L) regeneration series bicarbonate form NO3/Ca 0 min 5 min 15 min 30 min 120 min
99 A) B) Figure 4 8 Absolute c admium concentrati on for each resin in either the A) chloride or B) bicarbonate form in the synthetic water containing nitrate and cadmium sampled at 0, 5, 15, 30, 120 min. 0 1 2 3 4 5 6 R1 R2 R3 R4 Cadmium (mg/L) regeneration series chloride form NO3/Cd 0 min 5 min 15 min 30 min 120 min 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 R1 R2 R3 R4 Cadmium (mg/L) regeneration series bicarbonate form NO3/Cd 0 min 5 min 15 min 30 min 120 min
100 Table 4 1 Solubility products of likely precipitates. Mineral Chemical composition pKsp 1 Magnes ium carbonate (Magnesite) MgCO 3 3.68 Magnesium hydroxide Mg(OH)2 9.22 Calcium carbonate (Calcite) CaCO 3 8.01 Calcium hydroxide Ca(OH)2 5.26 Cadmium carbonate (Otavite) CdCO 3 11.3 Cadmium hydroxide Cd(OH)2 13.64 1 Values obtained from Knovel Critical Tables (2 nd Edition)
101 Table 4 2 Theoretical model water compositions Modeled synthetic water Cl (mmol/L) NO 3 (mmol/L) SR NOM (mmol/L) H CO 3 2 (mmol/L) Ca 2+ (mmol/L) Mg 2+ (mmol/L) Cd 2+ (mmol/L) NOM Ca 0.06895 0.06895 0.06895 0.13790 NOM Mg 0.06895 0.06895 0.06895 0.13790 NOM Cd 0.13790 0.06895 0.06895 0.13790 Nitrate Ca 0.06895 0.06895 0.06895 0.13790 Nitrate Mg 0.06895 0.06895 0.06895 0.13790 Nitrate Cd 0.13790 0.06895 0.06895 0.13790 Table 4 3 Synthetic model water compos itions for waters using nitrate in lieu of NOM, with varying divalent cations. Used for Visual MINTEQ input. Modeled synthetic water pH Cl (mmol/L) NO3 (mmol/L) CO32 (mmol/L) Na (mmol/L) Ca (mmol/L) Mg (mmol/L) Cd (mmol/L) Nitrate Ca 7.20 0.06895 0.06895 0.06895 0.20685 0.13790 Nitrate Ca post Cl 7.13 0.56895 0.06895 0.06895 0.20685 0.13790 Nitrate Ca post HCO3 7.01 0.06895 0.06895 0.56895 0.20685 0.13790 Nitrate Mg 9.75 0.06895 0.06895 0.06895 0.20685 0.13790 Nitrate Mg post Cl 9.95 0.56895 0 .06895 0.06895 0.20685 0.13790 Nitrate Mg post HCO3 10.10 0.06895 0.06895 0.56895 0.20685 0.13790 Nitrate Cd 6.37 0.13790 0.06895 0.06895 0.13790 0.13790 Nitrate Cd post Cl 5.78 0.63790 0.06895 0.06895 0.13790 0.13790 Nitrate Cd post HCO3 6.35 0.13790 0.06895 0.56895 0.13790 0.13790
102 Table 4 4 Results of Visual MINTEQ Simulations Mineral most likely to precipitate a saturation index a,b equilibrium solid amount (mmol/L) c removal of cation by precipitation c Nitrate Ca post Cl Aragonite 3.024 Calcite 2.88 Vaterite 3.447 Nitrate Ca post HCO3 Aragonite 2.24 Calcite 2.096 Vaterite 2.663 Nitrate Mg post Cl Brucite 1.136 Magnesite 1.207 Mg(OH) 2 2.830 Nitrate Mg post HCO3 Brucite 1. 067 Mag nesite 0.305 Mg(OH) 2 2.761 Nitrate Cd post Cl Cd(OH)2 6.019 Otavite 1.314 Nitrate Cd post HCO3 Cd(OH)2 4.861 Otavite 0.56 9.36E 0 5 68 % a No precipitation allowed. b log(IAP/Ks0); where IAP is the ion activity product and K s0 i s the stability constant of the solid. c Precipitation allowed.
103 CHAPTER 5 DETERMINATION OF BIOLOGICAL FOULING ON ANION EXCHANGE USED IN WASTEWATER TREATMENT 5.1 Background Magnetic ion exchange resin (MIEX) has been shown to be effective at removing di ssolved organic carbon in many potable water treatment systems (Boyer and Singer 2006, 2008, Drikas et al. 2007, Kingsbury and Singer 2013, Singer et al. 2007) and is increasingly being used as a treatment technology for pre and post treatment needs (Boye r et al. 2011, O'Shea et al. 2009) Wastewater treatment often requires the use of tertiary treatment to meet governmental requirements to meet water reuse or land application standards. In instances with high dissolved organic carbon concentrations it has been found that anion exchange is an effective means tertiary treatment technology for wastewater treatment (MIEX) (Gan et al. 2013, Lu et al. 2012, Zhang et al. 2006) One concern associated with the use of anion exchange for tertiary wastewater treatmen t is that the highly active microbial populations may attach to the resin surface and any biofilm on the surfaces of anion exchange resins will interfere with resin surface sites and cause the loss of performance. The formation of a biofilm would be detri mental to the use of anion exchange as a tertiary wastewater treatment technology, with any means to quantify and/or mollify the extent of biological fouling being of interest. Anion exchange is a process that utilizes particles that have fixed charged sur face sites. In anion exchange th ese fixed surface sites are positively charged and anionic mobile ions attach until a more favorable ion is present and exchanges places with the mobile ion (Helfferich 1962)
104 particle size so it is easily fluidized in the contact reactor, and is able to rapidly aggregate and settle while not being agitated (Drikas et al. 2003) The standard mobile ion used in anion exchange is the chloride ion, but research has been done to inv estigate alternative ions that generate less harmful waste regenerant products (Rokicki and Boyer 2011) The use of different mobile ions has been shown to not alter the affinity of the resin for other mobile ions, and should not greatly impact the overall performance of the resin when targeting DOC removal (Rokicki and Boyer 2011) To more accurately mimic full scale anion exchange treatment processes, multiple loading jar test experiments were used due to their more accurate representation of resin loadin g conditions (Kitis et al. 2007, Mergen et al. 2009) In the full scale systems only a small percentage of the resin is drawn off and regenerated and these systems are designed such that each resin dose is in contact with a certain number of total bed volu mes prior to regeneration. I t is expected that DOC polishing c ould be accomplished with the additional benefits of additional nitrate removal (Kitis et al. 2007) t hrough the use of ion exchange for wastewater tertiary treatment at times this approach diff icult even with denitrification processes (Zhang et al. 2012) Previous literature has shown that bacteria, particularly gram positive bacteria, due to the negatively charged nature of the peptidoglycan layer (van der Wal et al. 1997) can adhere to resin s tructure. The adherence of microorganisms to the resin structure would require that bacteria have a negative charge to enter into the resin phase; otherwise the positive fixed mobile ions would exclude any adhesion to the resin (Cumbal and Sengupta 2005, S arkar and SenGupta 2010) It is feasible that any organisms that attach to the resin structure would have the necessary carbon source
105 provided by DOC within the resin In some instances, microbial populations may be able to reduce anions in addition to org anic carbon as a potential energy source, including sulfates and nitrates (Bitton 2005) In wastewater treatment, denitrification is the process in which microorganisms reduce oxidized nitrogen as a means of energy production. As such it is important to c onsider how different mobile ions would effect the biological fouling of the anion exchange. The method for quantifying biofilm growth is EPA method 1684, which is a way to quantify the total, fixed, and volatile solids. EPA method 1684 is based on the rep eat drying and weighing of samples. In the use of ion exchange the continued baking of the ion exchange resin may denature the resin and provide inaccurate results. Previous work has shown that an alternative to the use of EPA method 1684 is the use of cou lter counter analysis to estimate counts of suspended cells, polymers, and aggregates ( Casey et al. 2012, Huang et al. 2012, Zhang et al. 2013) Coulter counter analysis measures changes in conductance at an aperture opening as liquid sample travels throug h the orifice. C hange s in condunctance are then quantified and algorithms provide data as particle size/count. Coulter counters have been used previously in water treatment to evaluate flocculation (Treweek and Morgan 1977) and can further be used to quant ify the estimated amounts of microbial counts. Currently ion exchange is a well understood process, but it is still not readily investigated or used in wastewater treatment. The impact of wastewater on the resin is not thoroughly understood, and biologic al fouling is potentially of major concern, due to the lack of published research regarding the biological fouling of ion exchange resins in standard water treatment processes. There is a lack of knowledge pertaining to how to
106 quantify and estimate biologi cal fouling on ion exchange resins and how to differentiate between fouling associated with use or inorganic scale. The challenge is relat ed to collecting and accurately quantifying biomass on from the surface of small fluidized particles. Also an area tha ion impacts the extent of biological growth that occurs on the resin. The overall goal of this work was to determine whether or not different mobile ions impact the biological fouling that occurs in ion exchange used for wastewater treatment. To accomplish this goal resins were regenerated with different mobile ions and used in multiple loading experiments to treat clarifier effluent from a wastewater treatment plant. Biofilms were separated from the re sins and analyzed by particle size distribution as described in Huang (2012) and Zhang (2013) The objectives of this work w ere to: (1) develop a method to analyze biological growth on the surface of the ion exchange resin; and (2) determine the impact of mobile counter ion on the extent to which biofouling occurs. 5.2 Materials and Methods 5.2.1 Resin The resin used in this study was virgin MIEX resin provided by Orica Watercare. Prior to use resin was rinsed and regenerated into the desired mobile ion. Regeneration occurred in bulk batches of resin and was dosed stoichiometrically at 10 the capacity of the MIEX resin. 20 mL of resin was measured out in graduated cylinders, and added to 1 L of water containing 104 meq/L of the desired mobile ion of eithe r: chloride, bicarbonate, nitrate, or sulfate. Resins were mixed in the regenerant solution for 30 min and rinsed twice, as described in previous work (Rokicki and Boyer 2011)
107 5.2.2 Water S ource Water for experimentation was collected from the Water Recl amation Facility located at the University of Florida in Gainesville, F L The Water Reclamation Facility services the entirety of the University of Florida campus, which utilizes the four stage Bardenpho treatment process This process offers effective nit rogen and phosphorus removal with minimal chemical addition (Burdick et al. 1982) The Bardenpho process involves an anoxic denitrification zone, followed by aerobic nitrification, with another anoxic dentrification stage, followed by an aerobic phase (Bur dick et al. 1982) Samples were collected on the day of experimentation following the clarifier, but prior to sand filtration. Samples were not filtered prior to use and were not treated prior to contact with the resin. 5.2.3 Multiple L oading R esin E xhaus tion To more closely mimic the conditions of full scale water treatment processes, multiple loading jar tests were conducted on the resins (Kitis et al. 2007, Mergen et al. 2008) These multiple loading jar tests recreate conditions in which resin will be in contact with multiple bed volumes of water before it is regenerated. 1 mL of regenerated resin was dosed into beakers and 250 mL of clarifier water was added to the resin. The water was then mixed by magnetic stir bar on a stir plate at 250 rpm for 30 min. After mixing the water was decanted with care to prevent loss of resin and an additional 250 mL of clarifier water was added and mixed for an additional 30 min. This process was repeated an additional two times which totals a total contact of 1000 BV. After multiple loading experiments, resins were allowed to sit for varying periods of time to mimic resins sitting in plate settlers prior to regeneration.
108 All experiments were performed in triplicate, with each data point representing the mean of the tri plicate data. 5.2.4 Biofilm A nalysis To quantify the extent of biofilm formation on the resin, the supernatant fluid was decanted from the resin. To effectively count microorganisms as particles they must be cleaved from the resin. Sodium dodecyl sulfate is a commonly used surfactant and is often used to break extracellular polysaccharide binding to surfaces (Simoes et al. 2008) Resin was added to 25 mL of a 0.1 mg/l sodium dodecyl sulfate wash solution and then vortexed for approximately 1 min. Resin was allowed to settle, and 0.25 mL of the supernatant was then added to 20 mL of Beckman Isoton solution. Samples were which provides particle counts in the size range of 0.6 m to 18 m. Statistical error associated with particle counts is represented as standard error of total particle counts. 5.3 Results and Discussion 5.3.1 Evaluation of biofilm growth Figure 5 1 shows the total absolute particle count numbers after allowing biofilm to develop for 5 d following the multiple loading regeneration sequence. The figure shows the control sample which just represents a chloride form resin following the multiple loading jar test procedure, but using DI water in lieu of the clarifier effluent. This control was established to ensure that the mixing and decanting the resin multiple times did not cause the resin to fracture and increase particle counts in the aqueous phase. In addition to the control sample, resins were regenerated into chloride bicarbonate
109 nitrate and sulfate form. After 5 d of contact, Figure 5 1 shows that the control has minimal total particle count, while the chloride bicarbonate and nitrate form having approximately 55000 total particle count and sulfat e form having approximately 35000 particle count. It is clear that the control sample and the treated samples have drastically different particle counts, which is to be expected as DI water has no particulates, microorganisms or carbon source for microbes Figure 5 2 is configured identically to Figure 5 1, with the exception being that it represents the extent of biological growth that occurred after 7 days of contact. In Figure 5 2 the control value is approximately 7000 particles, which is in line with what was seen in Figure 5 1 This means that the expected baseline for background particle counts is in the range of 5000 7000 particles. In Figure 5 2 the chloride form resin has a total count of 100000 particles, the bicarbonate form having 77000, nitra te 87000, and sulfate 67000. There appears to be a greater total particle count than in Figure 5 1, in addition to a slightly larger amount of standard error. Based on these preliminary results, it does appear that there is biofilm developing on the resin and it does increase with time. However, there is no clear trend in how the mobile ion impacts the biofouling, with the exception of the sulfate resin, which appears to have a slightly decreased total particle count. To ensure that particles coming off of the resin were not just turbidity from the clarifier process and actually related to microorganisms, some alternative waters were used. Two alternative waters were generated, both based on the clarifier effluent used in the previous figures. The first wate r was unfiltered clarifier water that was treated with 10 mL /L of 6% sodium hypochlorite solution. This water was allowed to mix for 1 hour
110 prior to contact with the resin. The second water was the clarifier effluent, but was filtered through 0.45 m Milli pore filters. The goal of using these alternative waters was to act as further confirmation that biological growth on the resin was what was being detached, rather than other particulates in solution, or the swelling and potentially breaking of the resin f ollowing the treatment process. Once these waters were generated, they were added to the resin and mixed for a total contact time of 120 min, which was the same total contact time as the multiple loading experiments described previously Due to a shortage of water only 250 bed volumes were achieved. Figure 5 3 shows the results of these alternative water loading experiments. The chloride form resin in the presence of the inactivated water did have a relatively high particle count of 30000, which implies tha t the inactivated microorganisms, fragments of microorganisms and turbidity are contributing to the fouling of the resin. It may also be on the resin. The filtered water samples for chloride bicarbonate nitrate and sulfate form resin all have very low particle counts all of which are less than 1600 total particles. 5.3.2 F ollow up Work This work serves as the foundations for future research projects with the bulk of the effort spent on developing a system for quantifying the biofilm on the resin. Further efforts to optimize the coulter counter quantification will be developed by investigating the necessary concentration of surfactants needed to separate microbial matter from the resin. To optimize the removal of microbial matter, it will be necessary to test multiple surfactants at different concentrations. It is important to limit the concentration
111 of surfactant added so that the surfactant does not form aggregate s on their own which would show up in particle counts, thus requiring several series of optimization experiments to find an ideal concentration of surfactant to remove microbial matter, without forming surfactant colloids. Subsequent work to fully develop the understanding of biological fouling of anion exchange resins will need to be pursued. Several additional multiple loading experiments will need to be performed, with multiple contact times. It is necessary to evaluate the resin almost immediately foll owing multiple loading to see if it is biological material that accumulates on the resin, or the biological growth that is being quantified. Additional time steps would also further allow to evaluate whether there is growth on the resin. Due to the destruc tive process of sample analysis, organization of the experiment will need to be altered to track the change over time. To track the growth over time, one form of the resin will be chosen (likely the chloride form) and used to treat clarifier water from the wastewater treatment plant. Multiple samples will be generated and designated sample times will be selected, likely 0 d, 1 d, 2 d, 3 d, 7 d. Through this method, it will be possible to track the impact that time has on biological growth with consistent so urce water. An additional experimental plan is to confirm that coulter counter method with another method of analysis t o insure that the particle counts represent the biological matter. Crystal violet staining will be utilized (Zhang et al. 2013) The ben efit of crystal violet staining is that it is a well documented and recognized method for detecting and quantifying microbial populations (Shanks et al. 2005) Crystal violet stains the peptidoglycan layer of the bacteria and fluoresces when excited at 590 nm the
112 fluorescence signal can be measured and quantified using a plate reader (Zhang et al. 2013) The crystal violet staining would provide a reliable confirmation of the coulter counter methods. 5. 4 Chapter Summary The coulter counter analysis appea rs to be able to demonstrate the extent of biofilm formation on anion exchange resins. There does not appear to be a significant difference between mobile ion and the sulfat e may cause the sulfate to remaining on the resin surface, preventing access to the resin by microorganisms.
113 Figure 5 1 Total particle counts for resins with allowed biofilm growth for 5 days. Figure 5 2 Total particle counts for resins with allo wed biofilm growth for 7 days. 0 20000 40000 60000 80000 100000 120000 control MIEX-Cl MIEX-HCO3 MIEX-NO3 MIEX-SO4 total particle count (#) Resin form 5 day contact 0 20000 40000 60000 80000 100000 120000 control MIEX-Cl MIEX-HCO3 MIEX-NO3 MIEX-SO4 total particle count (#) Resin Form 7 day contact
114 Figure 5 3 Total particle counts for resins with allowed biofilm growth for 5 days in disinfected water, or filtered water. 0 5000 10000 15000 20000 25000 30000 35000 40000 M-Cl, hypochlorous MIEX-Cl, filtered MIEX-HCO3, filtered MIEX-NO3, filtered MIEX-SO4, filtered Total particle counts # Resin form 5 day contact, alternative water
115 CHAPTER 6 CONCLUS IONS 6.1 Conclusions of D issertation The goal of this work was to investigate the regeneration and fouling of bicarbonate form anion exchange resin. Some key findings of this work are: The affinity of MIEX resin is not altered by the bicarbonate form of the resin and the resin maintained the ideal ion exchange stoichiometry in prel iminary experiments showing that using the bicarbonate form of the resin is not inherently detrimental to the ion exchange process. MIEX resin will target a wide range of organic matter regardless of whether the resin was in the bicarbonate or chloride fo rm. The resin maintains the same organic matter fraction preference regardless of mobile ion and is an extension of the affinity. The organic matter fractions best removed by the resin are those fractions in the 1000 1500 Da size range, which is the range that contains carboxylic acid functional groups. This fraction also tends to foul resin the most. As shown previously the regeneration strength of the solution matters much more than the mobile ion. Differences between the performance of chloride and bica rbonate form resins are slight, but the biggest impacts come after multiple regeneration cycles. Rather than increased fouling in some instances, the decrease in performance may be an additive effect of weaker regeneration. Chloride concentrations can be m uch higher than bicarbonate concentrations, thus increasing the ionic strength and driving ion exchange equilibrium for more complete regeneration of the resin in the chloride form. Divalent cations do impact the resin, but the impact of performance is no t likely directly related to precipitation on the resin, but caused by the metal cations interfering with the organic matter or structure of the resin. Biological fouling will occur on the resin surface regardless of mobile ion if there is a carbon source for the microorganisms. 6.2 Implications of Findings The majority of the implications of these findings are related to a deeper understanding of the mechanisms related to the ion exchange process. Previously it
116 appeared that bicarbonate form resin was an undesirable alternative to the use of chloride form resin but was something to be used when disposal of brine was not feasible. Now with further understanding of the relationship between bicarbonate and the resin, organic matter, divalent cations and micro organisms make it so that a more complete picture can be presented. The bicarbonate form of the resin is able to behave to a similar extent as the chloride form and this was demonstrated when the resin performance was compared when regenerated using simil ar milliequivalnce based regenerant solutions. Knowing when the bicarbonate form of the resin may be fouled or lose its capacity allow for the more targeted use of this resin, with the specific option to regenerate it as needed. This work also provides add itional information about the use of the chloride form resin and how different regeneration agents impact it. Ultimately this work clarifies how ion exchange works, and how the regeneration process impacts ion exchange. This work serves as a foundation for those hoping to investigate the use of bicarbonate as a mobile ion, or for anyone that is curious about how regenerants impact resin performance or how resin is fouled by either cationic metals or biological growth.
117 APPENDIX A SUPPORTING INFORMATION FO R CHAPTER 3 Table A 1 experiments Sample M w MtP Cl 10 2 raw 1005 1.41 MtP Cl 10 2 200BV 863 1.60 MtP Cl 10 2 400BV 847 1.52 MtP Cl 10 2 600BV 830 1.50 MtP Cl 10 2 800BV 837 1.49 MtP Cl op 2 raw 1016 1.42 MtP Cl op 2 200BV 806 1.78 MtP Cl op 2 400BV 795 1.68 MtP Cl op 2 600BV 771 1.64 MtP Cl op 2 800BV 788 1.59 Sample M w MR Cl 10 2 raw 1612 1.76 MR Cl 10 2 200BV 1698 1.80 MR Cl 10 2 400BV 1673 1.80 MR Cl 10 2 600BV 1693 1.80 MR Cl op 2 raw 1537 1. 74 MR Cl op 2 200BV 1745 1.85 MR Cl op 2 400 BV 1695 1.83 MR Cl op 2 600 BV 1677 1.82 Sample M w MtP HCO3 10 2 raw 1005 1.41 MtP HCO3 10 2 200 BV 851 1.56 MtP HCO3 10 2 400 BV 844 1.50 MtP HCO3 10 2 600 BV 831 1.48 MtP HCO 3 10 2 800 BV 839 1.47 MtP HCO3 op 2 raw 1016 1.42 MtP HCO3 op 2 200 BV 829 1.74 MtP HCO3 op 2 400BV 824 1.66 MtP HCO3 op 2 600BV 796 1.60 MtP HCO3 op 2 800BV 811 1.55 Sample M w MR HCO3 10 2 raw 1612 1.76 MR HCO3 10 2 200 BV 1677 1.77 MR HCO3 1 0 2 400BV 1685 1.80 MR HCO3 10 2 600BV 1680 1.81 MR HCO3 op 2 raw 1537 1.74 MR HCO3 op 2 200BV 1669 1.78 MR HCO3 op 2 400 BV 1674 1.81 MR HCO3 op 2 600 BV 1677 1.82
118 APPENDIX B SUPPORTING INFORMATION FOR CHAPTER 4 Figure B 1 Kine tic isotherm showing pH change for each resin in either the chloride or bicarbonate form in the presence of either: Mg, Ca, or Cd for waters using NOM as the targeted anion for removal. Arranged in order of decreasing pK sp 5 6 7 8 9 10 11 0 50 100 150 pH time (min) chloride form NOM Mg R1 R2 R3 R4 5 6 7 8 9 10 11 0 50 100 150 pH time (min) bicarbonate form NOM Mg R 1 R 2 R 3 5 6 7 8 9 10 11 0 50 100 150 pH time (min) chloride form NOM Ca R1 R2 R3 R4 5 6 7 8 9 10 11 0 50 100 150 pH time (min) bicarbonate form NOM Ca R 1 R 2 R 3 5 6 7 8 9 10 11 0 50 100 150 pH time (min) chloride form NOM Cd R1 R2 R3 R4 5 6 7 8 9 10 11 0 50 100 150 pH time (min) bicarbonate form NOM Cd R 1 R 2 R 3
119 Figure B 2 Kine tic isotherm showing pH change for each resin in either the chloride or bicarbonate form in the presence of either: Mg, Ca, or Cd for waters using nitrate as the targeted anion for removal. Arranged in order of decreasing pK sp 5 6 7 8 9 10 11 0 50 100 150 pH time (min) chloride form NO3 Mg R1 R2 R3 R4 5 6 7 8 9 10 11 0 50 100 150 pH time (min) bicarbonate form NO3 Mg R1 R2 R3 R4 5 6 7 8 9 10 11 0 50 100 150 pH time (min) chloride form NO3 Ca R1 R2 R3 R4 5 6 7 8 9 10 11 0 50 100 150 pH time (min) bicarbonate form NO3 Ca R1 R2 R3 R4 5 6 7 8 9 10 11 0 50 100 150 pH time (min) chloride form NO3 Cd R1 R2 R3 R4 5 6 7 8 9 10 11 0 50 100 150 pH time (min) bicarbonate form NO3 Cd R1 R2 R3 R4
120 LIST OF REFERENCES Abbt Bra un, G., Lankes, U. and Frimmel, F.H. (2004) Structural characterization of aquatic humic substances The need for a multiple method approach. Aquatic Sciences 66(2). Ahn, W.Y., Kalinichev, A.G., Clark, M.M., 2008. Effects of background cations on the foul ing of polyethersulfone membranes by natural organic matter: Experimental and molecular modeling study. Journal of Membrane Science 309, 128 140. Allpike, B.P., Heitz, A., Joll, C.A., Kagi, R.I., Abbt Braun, G., Frimmel, F.H., Brinkmann, T., Her, N. and Am y, G. (2005) Size exclusion chromatography to characterize DOC removal in drinking water treatment. Environmental Science & Technology 39(7). Apell, J.N., Boyer, T.H., 2010. Combined ion exchange treatment for removal of dissolved organic matter and hardne ss. Water Res 44, 2419 2430. Benjamin, M.M., 2002. Water Chemistry. McGraw Hill, New York, NY. Bhattacharjee, S. and Johnston, G.M. (2002) A model of membrane fouling by salt precipitation from multicomponent ionic mixtures in crossflow nanofiltration. Env ironmental Engineering Science 19(6), 399 412. Bitton, G. (2005), Wastewater microbiology, 3 rd ed. John Wiley & Sons, Inc. Blackhurst, M., Hendrickson, C. and Vidal, J.S.I. (2010) Direct and indirect water withdrawals for US industrial sectors. Environmen tal Science & Technology 44(6), 2126 2130. Bolto, B., Dixon, D., Eldridge, R., King, S. and Linge, K. (2002) Removal of natural organic matter by ion exchange. Water Research 36(20), 5057 5065. Boyer, T.H. and Singer, P.C. (2005) Bench scale testing of a m agnetic ion exchange resin for removal of disinfection by product precursors. Water Research 39(7), 1265 1276. Boyer, T.H., Singer, P.C., 2006. A pilot scale evaluation of magnetic ion exchange treatment for removal of natural organic material and inorgani c anions. Water Research 40, 2865 2876. Boyer, T.H., Singer, P.C., 2008. Stoichiometry of removal of natural organic matter by ion exchange. Environmental Science & Technology 42, 608 613. Boyer, T.H. and Singer, P.C. (2008b) Removal of natural organic mat erial by a magnetic ion exchange resin. Water Science and Technology: Water Supply 8(2), 167 172.
121 Boyer, T.H., Singer, P.C., Aiken, G.R., 2008. Removal of dissolved organic matter by anion exchange: Effect of dissolved organic matter properties. Environmen tal Science & Technology 42, 7431 7437. Boyer, T. H., Miller, C. T. and Singer, P. C. (2010) Advances in modeling completely mixed flow reactors for ion exchange. Journal of Environmental Engineering ASCE, doi: 10.1061/(ASCE)EE.1943 7870.0000241. Boyer, T.H., Graf, K.C., Comstock, S.E.H. and Townsend, T.G. (2011) Magnetic ion exchange treatment of stabilized landfill leachate. Chemosphere 83(9), 1220 1227. Burdick, C.R., Refling, D.R. and Stensel, H.D. (1982) Advanced Biological Treat ment to Achieve Nutrient Removal. Journal Water Pollution Control Federation 54(7), 1078 1086. Chin, Y.P., Aiken, G. and Oloughlin, E. (1994) Molecular weight, polydispersity, and properties of aquatic humic substances Environmental Science & Technology 2 8(11). Chow, C.W.K., Fabris, R., van Leeuwen, J., Wang, D. and Drikas, M. (2008) Assessing natural organic matter treatability using high performance size exclusion chromatography. Environmental Science & Technology 42(17), 6683 6689. Collins, M.R., Amy, G .L. and Steelink, C. (1986) Molecular Weight Distribution, Carboxylic Acidity, and Humic Substances Content of Aquatic Organic Matter Implications for Removal during Water Treatment. Environmental Science & Technology 20(10), 1028 1032. Cornelisse n, E.R. Beerendonk, E.F., Nederlof, M.N., van der Hoek, J.P. and Wessels, Collins, M.R., Amy, G.L. and Steelink, C. (1986) Molecular Weight Distribution, Carboxylic Acidity, and Humic Substances Content of Aquatic Organic Matter Implications for Removal during Water Treatment. Environmental Science & Technology 20(10), 1028 1032. Cornelissen, E.R., Beerendonk, E.F., Nederlof, M.N., van der Hoek, J.P. and Wessels, L.P. (2009) Fluidized ion exchange (FIX) to control NOM fouling in ultrafiltration. Desalination 23 6(1 3), 334 341. Croue, J.P., Violleau, D., Bodaire, C. and Legube, B. (1999) Removal of hydrophobic and hydrophilic constituents by anion exchange resin. Water Science and Technology 40(9), 207 214. Cumbal, L. and Sengupta, A.K. (2005) Arsenic removal usi ng polymer supported hydrated iron(III) oxide nanoparticles: Role of Donnan membrane effect. Environmental Science & Technology 39(17), 6508 6515.
122 Dahlke, T., Mathes, P.A. and Adams, B. (2007) Treatment of highly polluted paper and pulp effluent using comb ined treatment processes including a continuous ion exchange process. Proceedings Water Reuse and Recycling, pp 122 129. Davis, W.M., Erickson, C.L., Johnston, C.T., Delfino, J.J. and Porter, J.E. (1999) Quantitative Fourier Transform Infrared spectroscopi c investigation of humic substance functional group composition. Chemosphere 38(12), 2913 2928. Dean, J.A., 1999. Lange's Handbook of Chemistry, 15th ed. McGraw Hill. Drikas, M., Chow, C.W.K. and Cook, D. (2003) The impact of recalcitrant organic character on disinfection stability, trihalomethane formation and bacterial regrowth: An evaluation of magnetic ion exchange resin (MIEX (R)) and alum coagulation. Journal of Water Supply Research and Technology Aqua 52(7), 475 487. Drikas, M., Dixon, M. and Morran J. (2007) Removal of DBP precursors using the MIEX process in combination with granular activated carbon. Abstracts of Papers of the American Chemical Society 233, 410 410. Drikas, M., Dixon, M., Morran, J., 2009. Removal of MIB and geosmin using granula r activated carbon with and without MIEX pre treatment. Water Research 43, 5151 5159 Drikas, M., Dixon, M. and Morran, J. (2011) Long term case study of MIEX pre treatment in drinking water; understanding NOM removal. Water Research 45(4), 1539 1548. Dudal, Y., Gerard, F., 2004. Accounting for natural organic matter in aqueous chemical equilibrium models: a review of the theories and applications. Earth Science Reviews 66, 199 216. Dzombak, D.A., Fish, W. and Morel, F.M.M. (1986) Metal Humate Interactions .1 Discrete Ligand and Continuous Distribution Models. Environmental Science & Technology 20(7), 669 675. Edwards, M. and Triantafyllidou, S. (2007) Chloride to sulfate mass ratio and lead leaching to water. Journal American Water Works Association 99(7), 9 6 109. Eisman, P.C., Kull, F.C., Mayer, R.L., 1949. THE BACTERIOLOGICAL ASPECTS OF DEIONIZED WATER. Journal of the American Pharmaceutical Association Scientific Edition 38, 88 91. Fabris, R., Chow C.W.K., Drikas, M. and Eikebrokk, B. (2008) Comparison of NOM character in selected Australian and Norwegian drinking waters. Water Research 42(15), 4188 4196 Fabris, R., Lee, E.K., Chow, C.W.K., Chen, V. and Drikas, M. (2007) Pre treatments to reduce fouling of low pressure micro filtration (MF) membranes. Jour nal of Membrane Science 289(1 2), 231 240.
123 Fearing, D.A., Banks, J., Guyetand, S., Eroles, C.M., Jefferson, B., Wilson, D., Hillis, P., Campbell, A.T. and Parsons, S.A. (2004) Combination of ferric and MIEX (R) for the treatment of a humic rich water. Wa ter Research 38(10). Flemming, H.C., 1987. Microbial growth on ion exchangers. Water Research 21, 745 756. Fu, P.L.K. and Symons, J.M. (1990) Removing aquatic organic substances by anion exchange resins. Journal American Water Works Association 82(10), 70 77. Gan, X.J., Kim, D. and Karanfil, T. (2013) MIEX (R) treatment of an effluent impacted stream. Journal American Water Works Association 105(4), 53 54. Garland, J.L., 1997. Analysis and interpretation of community level physiological profiles in microbia l ecology. Fems Microbiology Ecology 24, 289 300. Greenleaf, J.E. and SenGupta, A.K. (2009) Flue gas carbon dioxide sequestration during water softening with ion exchange fibers. Journal of Environmental Engineering ASCE 135(6), 386 396. Hammann, D., Bour ke, M., Topham, C., 2004. Evaluation of a magnetic ion exchange resin to meet DBP regs at the Village of Palm Springs. Journal American Water Works Association 96, 46 50. Helfferich, F.G., 1962. Ion exchange. McGraw Hill, New York. H ll, W. and Kiehling, B (1981) Regeneration of anion exchange resins by calcium carbonate and carbon dioxide. Water Research 15(8), 1027 1034. Horng, L.L. and Clifford, D. (1997) The behavior of polyprotic anions in ion exchange resins. Reactive & Functional Polymers 35(1 2), 41 54. Howe, K.J. and Clark, M.M. (2002) Fouling of microfiltration and ultrafiltration membranes by natural waters. Environmental Science & Technology 36(16). Hsu, S., Singer, P.C., 2010. Removal of bromide and natural organic matter by anion exchange. Wa ter Research 44, 2133 2140. Huang, H., Cho, H. H., Schwab, K.J. and Jacangelo, J.G. (2012) Effects of magnetic ion exchange pretreatment on low pressure membrane filtration of natural surface water. Water Research 46(17). Huber, S.A., Balz, A., Abert, M. a nd Pronk, W. (2011) Characterisation of aquatic humic and non humic matter with size exclusion chromatography organic carbon detection organic nitrogen detection (LC OCD OND). Water Research 45(2), 879 885.
124 Humbert, H., Gallard, H., Jacquemet, V. and Croue, J.P. (2007) Combination of coagulation and ion exchange fo r the reduction of UF fouling properties of a high DOC content surface water. Water Res 41(17), 3803 3811. Humbert, H., Gallard, H., Suty, H. and Croue, J.P. (2005) Performance of selected anion exchange resins for the treatment of a high DOC content surface water. Water Research 39(9), 1699 1708. Ishii, S.K.L. and Boyer, T.H. (2011) Evaluating the secondary effects of magnetic ion exchange: Focus on corrosion potential in the distribution syst em. Desalination 274(1 3), 31 38. Iskrenova Tchoukova, E., Kalinichev, A.G., Kirkpatrick, R.J., 2010. Metal cation complexation with natural organic matter in aqueous solutions: molecular dynamics simulations and potentials of mean force. Langmuir 26, 1590 9 15919. Jarvis, P., Jefferson, B. and Parsons, S.A. (2006) Floc structural characteristics using conventional coagulation for a high doc, low alkalinity surface water source. Water Research 40(14), 2727 2737. Jelinek, L., Parschova, H., Matejka, Z., Paida r, M. and Bouzek, K. (2004) A combination of ion exchange and electrochemical reduction for nitrate removal from drinking water Part I: Nitrate removal using a selective anion exchanger in the bicarbonate form with reuse of the regenerant solution. Water Environment Research 76(7), 2686 2690. Jung, C.W. and Son, H.J. (2009) Evaluation of membrane fouling mechanism in various membrane pretreatment processes. Desalination and Water Treatment 2(1 3), 195 202. Kalinichev, A.G. and Kirkpatrick, R.J. (2007) Mol ecular dynamics simulation of cationic complexation with natural organic matter. European Journal of Soil Science 58(4), 909 917. Karpinska, A.M., Boaventura, R.A.R., Vilar, V.J.P., Bilyk, A. and Molczan, M. (2013) Applicability of (MIEXDOC) D A (R) proces s for organics removal from NOM laden water. Environmental Science and Pollution Research 20(6), 3890 3899. Kimura, K., Hane, Y., Watanabe, Y., Amy, G. and Ohkuma, N. (2004) Irreversible membrane fouling during ultrafiltration of surface water. Water Resea rch 38(14 15), 3431 3441. Kimura, E., Sakonaka, A. and Kodama, M. (1982) A carbonate receptor model by acromonocyclic polyamines and its physiological implications. Journal of the American Chemical Society 104(18), 4984 4985. Kingsbury, R.S. and Singer, P. C. (2013) Effect of magnetic ion exchange and ozonation on disinfection by product formation. Water Research 47(3), 1060 1072.
125 Kinniburgh, D.G., vanRiemsdijk, W.H. and Koopal, L.K. (1996) Concepts of ion binding to natural organic matter. Abstracts of Pape rs of the American Chemical Society 211, 103 GEOC. Kinniburgh, D.G., van Riemsdijk, W.H., Koopal, L.K., Borkovec, M., Benedetti, M.F. and Avena, M.J. (1999) Ion binding to natural organic matter: competition, heterogeneity, stoichiometry and thermodynamic consistency. Colloids and Surfaces a Physicochemical and Engineering Aspects 151(1 2), 147 166. Kitis, M., Harman, B.I., Yigit, N.O., Beyhan, M., Nguyen, H. and Adams, B. (2007) The removal of natural organic matter from selected Turkish source waters usin g magnetic ion exchange resin (MIEX (R)). Reactive & Functional Polymers 67(12), 1495 1504. Krasner, S.W., Weinberg, H.S., Richardson, S.D., Pastor, S.J., Chinn, R., Sclimenti, M.J., Onstad, G.D. and Thruston, A.D. (2006) Occurrence of a new generation of disinfection byproducts. Environmental Science & Technology 40(23), 7175 7185. Lankes, U., Luedemann, H. D. and Frimmel, F.H. (2008) Search for basic relationships between "molecular size" and "chemical structure" of aquatic natural organic matter Answer s from C 13 and N 15 CPMAS NMR spectroscopy. Water Research 42(4 5). Larson, T.E. (1975) Corrosion of domestic waters. Illinois State Water Survey Bulletin, Champaing, Ill. Lee, N.H., Amy, G., Croue, J.P. and Buisson, H. (2004) Identification and understan ding of fouling in low pressure membrane (MF/UF) filtration by natural organic matter (NOM). Water Research 38(20). Leenheer, J.A., Brown, G.K., MacCarthy, P. and Cabaniss, S.E. (1998) Models of metal binding structures in fulvic acid from the Suwannee Riv er, Georgia. Environmental Science & Technology 32(16), 2410 2416. Letterman, R.D., 1999. Water Quality and Treatment A Handbook of Community Water Supplies, 5th edition ed. McGraw Hill. Li, P. and Sengupta, A.K. (1998) Genesis of selectivity and reversi bility for sorption of synthetic aromatic anions onto polymeric sorbents. Environmental Science & Technology 32(23), 3756 3766. Lu, Z.Y., Jiang, B.C., Li, A.M., Guo, H.Y., Sun, S.G. and Chu, L.M. (2012) Toxicity reduction in an industrial nitro aromatic w astewater plant: an assessment and a proposed improvement. Water Science and Technology 66(2), 415 421.
126 Matosic, M., Mijatovic, I. and Hodzic, E. (2000) Nitrate removal from drinking water using ion exchange Comparison of chloride and bicarbonate form of the resins. Chemical and Biochemical Engineering Quarterly 14(4), 141 146. McLamore, E.S., Jaroch, D., Chatni, M.R., Porterfield, D.M., 2010. Self referencing optrodes for measuring spatially resolved, real time metabolic oxygen flux in plant systems. Pla nta 232, 1087 1099. Mergen, M.R.D., Adams, B.J., Vero, G.M., Price, T.A., Parsons, S.A. Jefferson, B. and Jarvis, P. (2009) Characterisation of natural organic matter (NOM) removed by magnetic ion exchange resin (MIEX Resin). Water Science and Technology: Water Supply 9(2), 199 205. Mergen, M.R.D., Jefferson, B., Parsons, S.A., Jarvis, P., 2008. Magnetic ion exchange resin treatment: Impact of water type and resin use. Water Research 42, 1977 1988. Mills, G.A. and Urey, H.C. (1940) The kinetics of isotopic exchange between carbon dioxide, bicarbonate ion, carbonate ion and water. Journal of the American Chemical Society 62, 1019 1026. Mitsoyannis, E. and Saravacos, G.D. (1977) Precipitation of Calcium Carbonate on Reverse Osmosis Membranes. Desalination 21( 3), 235 240. Neale, P.A. and Schafer, A.I. (2009) Magnetic ion exchange: Is there potential for international development? Desalination 248, 160 168. O'Shea, M.G., Dahlke, T., Staunton, S.P., Pope, G.M. and Holmquist, A. (2009) Evaluation of a magnetic ion exchange (MIEX (R)) resin for decolorisation applications within a raw sugar factory. International Sugar Journal 111(1324), 220 +. Panswad, T. and Anan, C. ( 1999) Impact of high chloride wastewater on an anaerobic/anoxic/aerobic process with and without inoculation of chloride acclimated seeds. Water Research 33(5), 1165 1172.Rey Castro, C., Mongin, S., Huidobro, C., David, C., Salvador, J., Garces, J.L., Galc eran, J., Mas, F., Puy, J., 2009. Effective Affinity Distribution for the Binding of Metal Ions to a Generic Fulvic Acid in Natural Waters. Environmental Science & Technology 43, 7184 7191. Pelekani, C., Newcombe, G., Snoeyink, V.L., Hepplewhite, C., Assem i, S. and Beckett, R. (1999) Characterization of natural organic matter using high performance size exclusion chromatography. Environmental Science & Technology 33(16).
127 Ritchie, J.D. and Perdue, E.M. (2003) Proton binding study of standard and reference fulvic acids, humic acids, and natural organic matter. Geochimica Et Cosmochimica Acta 67(1), 85 96.Singer, P.C. and Bilyk, K. (2002) Enhanced coagulation using a magnetic ion exchange resin. Water Research 36(16), 4009 4022. Rokicki, C.A., Boyer, T.H., 20 11. Bicarbonate form anion exchange: Affinity, regeneration, and stoichiometry. Water Research 45, 1329 1337. Rokicki, C.A. and Boyer, T.H. (2013) Impact of regeneration on fractionation of NOM in MIEX treated South Australian waters. Submitted to Water R esearch. Sarkar, S. and SenGupta, A.K. (2010) The Donnan Membrane Principle: Opportunities for Sustainable Engineering Processes and Materials. Environ. Sci. Technol 44, 1161 1166. Shanks, R.M.Q., Donegan, N.P., Graber, M.L., Buckingham, S.E., Zegans, M.E. Cheung, A.L. and O'Toole, G.A. (2005) Heparin stimulates Staphylococcus aureus biofilm formation. Infection and Immunity 73(8), 4596 4606. Sharma, P., Ofner, J., Kappler, A., 2010. Formation of Binary and Ternary Colloids and Dissolved Complexes of Organ ic Matter, Fe and As. Environmental Science & Technology 44, 4479 4485. Simoes, M., Simoes, L.C., Pereira, M.O. and Vieira, M.J. (2008) Sodium dodecyl sulfate allows the persistence and recovery of biofilms of Pseudomonas fluorescens formed under different hydrodynamic conditions. Biofouling 24(1), 35 44. Singer, P.C., Boyer, T., Holmquist, A., Morran, J., Bourke, M., 2009. Integrated analysis of NOM removal by magnetic ion exchange. Journal American Water Works Association 101, 65 +. Singer, P.C., Schneider, M., Edwards Brandt, J. and Budd, G.C. (2007) MIEX for removal of DBP precursors: Pilot plant findings. Journal American Water Works Association 99(4), 128 139. Swift, R.S. and Posner, A.M. (1971) GEL CHROMATOGRAPHY OF HUMIC ACID. Journal of Soil Science 22(2), 237 &. Takasaki, S. and Yamada, Y. (2007) Effects of temperature and aggressive anions on corrosion of carbon steel in potable water. Corrosion Science 49(1), 240 247. Tan, Y.R. and Kilduff, J.E. (2007) Factors affecting selectivity during dissolve d organic matter removal by anion exchange resins. Water Research 41(18), 4211 4221. Thurman, E.M., Malcolm, R.L., 1985. Organic Geochemistry of Natural Waters. Nijhoff/Junk, Boston, MA.
128 Treweek, G.P. and Morgan, J.J. (1977) Size Distributions of Floccula ted Particles Application of Electronic Particle Counters. Environmental Science & Technology 11(7), 707 714. Tripp, A.R. and Clifford, D.A. (2006) Ion exchange for the remediation of perchlorate contaminated drinking water. Journal American Water Works Association 98(4), 105 114. van der Wal, A., Norde, W., Zehnder, A.J.B. and Lyklema, J. (1997) Determination of the total charge in the cell walls of Gram positive bacteria. Colloids and Surfaces B: Biointerfaces 9(1 2), 81 100. Walker, K.M. and Boyer, T.H (2011) Long term performance of bicarbonate form anion exchange: Removal of dissolved organic matter and bromide from the St. Johns River, FL, USA. Water Research 45(9), 2875 2886. Warton, B., Heitz, A., Zappia, L.R., Franzmann, P.D., Masters, D., Joll, C.A., Alessandrino, M., Allpike, B., O'Leary, B., Kagi, R.I., 2007. Magnetic ion exchange drinking water treatment in a large scale facility. Journal American Water Works Association 99, 89 101. Weishaar, J.L., Aiken, G.R., Bergamaschi, B.A., Fram, M.S., F ujii, R. and Mopper, K. (2003) Evaluation of specific ultraviolet absorbance as an indicator of the chemical composition and reactivity of dissolved organic carbon. Environmental Science & Technology 37(20). Willison, H. and Boyer, T.H. (2012) Secondary ef fects of anion exchange on chloride, sulfate, and lead release: Systems approach to corrosion control. Water Research 46(7), 2385 2394. Yoon, S.H., Lee, C.H., Kim, K.J., Fane, A.G., 1998. Effect of calcium ion on the fouling of nanofilter by humic acid in drinking water production. Water Research 32, 2180 2186 Zhang, X., Li, F.Z. and Zhao, X. (2012) Application of a Magnetic Resin (MIEXA (R)) in Wastewater Reclamation for Managed Aquifer Recharge. Water Air and Soil Pollution 223(8), 4687 4694. Zhang, W., McLamore, E.S., Garland, N.T., Leon, J.L.C. and Katherine Banks, M. (2013) A Simple Method for Quantifying Biomass Cell and Polymer Distribution in Biofilms. Submitted to Water Research. Zhang, R., Vigneswaran, S., Ngo, H.H. and Nguyen, H. (2006) Magnetic ion exchange (MIEX (R)) resin as a pre treatment to a submerged membrane system in the treatment of biologically treated wastewater Desalination 192(1 3), 296 302.
129 BIOGRAPHICAL SKETCH Chris grew up in Miami, F L and from a very young age was fascinated wi th the sciences. Having been a Florida native, Chris decided to pursue his education at the University of Florida, and completed all his degrees there and is a loyal gator. He s in 2010, and his Ph.D. in 2013.