SYSTEMATIC VARIABILITY IN PROTON AND COPPER(II) COMPLEXATION BY DISSOLVED ORGANIC MATTER FROM SURFACE FRESHWATERS by Chad D. Luider B.Sc., Okanagan University College, 1999 THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in NATURAL RESOURCES AND ENVIRONMENTAL STUDIES © Chad D. Luider, 2003 THE UNIVERSITY OF NORTHERN BRITISH COLUMBIA June 2003 All rights reserved. This work may not be reproduced in whole or in part, by photocopy or other means, without the permission of the author. National LËxaiy of Canada BkUothèque nationale du Canada Acquisitions and BibliograpWc Services 385 WmNnglnn OewmON K1A0N4 Acquisitions et services bbBograpNques 385. m#W#anglon OUNwmON K1A0N4 Canmdm Our##NosmmUfim^ The author has granted a non­ exclusive Hcence allowing Ac National Libraiy of Canada to rqnodoce, loan, distnbote or sell copies of dns thesis in microhmn, p^)er or electronic formats. L'auteur a accordé une licence non exclusive permettant à la B Aliothèque naticmale du Canada de reproduire, ;nëter, distribuer ou vendre des copies de cette Aèse sous la Anne de microhdie/Glm, de rqnodmAioo sur papier ou sur Armât Aectronique. The anthor retains ownershq) of the c q p y r i^ in this Aesis. Neither Ae thesis nw substantial extracts 6om it may be printed w oAerwise reproduced wiAout Ae auAor's pamission. L'auteur conserve la ^ opiiété du droit d'auteur qui protège cette Aèse. Ni la thèse m des extraits substantiels de celle-ci ne doivent être inqnimés ou autrement reproduits sans smi autmisation. 0-612-84630-X CanadS APPROVAL Name: Chad D. Luider Degree: Master of Science Thesis Title: SYSTEMATIC VARIABILITY IN PROTON AND COPPER (II) COMPLEXATION BY DISSOLVED ORGANIC MATTER FROM SURFACE FRESHWATERS Examining Committee: Chair: Dr. Robert Tait Dean of Graduate Studies UNBC Co-Supervisor: rvisor: Dr. Ellen Petticrew Associate Professor, Geography Program UNBC Dr. Jeff Curtis A(^unct Professor, Geography Program UNBC Committee Member Dr. Todd Whitcombe Associate Ftofessor, Chemistry Program UNBC External Examiner: Dr. David Lean NSERC Chair in Ecotoxicology Professor - Department of Biology University of Ottawa Date Approved: A bstract The reproducibility of potentiometric titrations using a copper (H) ion selective electrode increases directly as a function of dissolved organic matter (DOM) concentration, measured as dissolved organic carbon. Conditional stability parameters (Log(Ki), Log(K2)), calculated 6 om titration curves using a 2-ligand Langmuir Isotherm, are also dependent on DOM concentration explaining 20% to 60% of the variation reported for these parameters in the literature. The effect of DOM concentration on Cu^^ complexation is greater for allochthonous DOM, which also exhibits higher charge density and higher afGnity for Cu^^, than autochthonous DOM (p < 0.05). The Cu^^ complexing properties of DOM collected 6 om surface hreshwaters depend primarily on allochthonous DOM and cumulative hydrologie residence time (p < 0.05) by approximating exposure to transformation and hactionation processes. Comparison to experimental transformation and hactionation processes in micro-reactors suggests that photochemical decomposition, microbial decomposition, and adsorptive hactionation could contribute to the observed pattern of Cu^^ complexation. fuge 2 q/" Table of C ontents ABSTRACT..............................................................................................................................2 TABLE OF CONTENTS........................................................................................................ 3 LIST OF TABLES....................................................................................................................6 LIST OF FIGURES................................................................................................................. 7 ACKNOWLEDGMENTS........................................................................................................ 8 LIST OF ACRONYMS, SYMBOLS AND UNITS................................................................9 PREFACE.............................................................................................................................. 12 1. CHAPTER 1. INTRODUCTION AND LITERATURE REVIEW............................... 13 1.1 Introduction.................................................................................................................... 13 1J2 Hypothesis......................................................................................................................16 1.3 SigniScance.................................................................................................................... 16 1.4 Literature Review...........................................................................................................18 1.4.1 Systematic variability of DOM..............................................................................18 1.4.2 Allochthonous and Autochthonous DOM ............................................................ 19 1.4.2.1 Isolating allochthonous and autochthonous DOM........................................... 22 1.4.3 Transformation and Fractionation of DOM.......................................................... 23 1.4.3.1 Photochemical Decomposition.........................................................................23 1.4.3.2 Microbial Decomposition.................................................................................24 1.4.3.3 Adsorption to Iron Oxyhydroxides and Hydroxides........................................ 26 1.4.4 Copper in Aquatic Systems.................................................................................. 27 1.4.4.1 Sources and effects........................................................................................... 27 1.4.4.2 Complexation of Cu^^ by DOM........................................................................29 1.4.4.3 Measuring Organic Acidity and Cu^^ Complexation....................................... 35 1.4.4.4 Pretreatment requirements................................................................................37 1.5 Conclusions o f Introduction and Literature Review....................................................38 fuge 3 q/"77^ 2. CHAPTER 2. ALLOCHTHONOUS AND AUTOCHTHONOUS DISSOLVED ORGANIC CARBON (DOC) CONCENTRATION DEPENDENCE OF PROTON AND COPPER(II) P0TENTI0IV1ETRIC TITRATION ANALYSES............................. 40 2.1 Abstract......................................................................................................................... 40 2.2 Introduction................................................................................................................... 42 2.3 Methods......................................................................................................................... 46 2.3.1 Sanc^le Collection and Preparation....................................................................... 46 2.3.2 Organic Acidity..................................................................................................... 48 2.3.3 Copper Titrations.................................................................................................. 49 2.4 Results........................................................................................................................... 52 2.4.1 Organic Acidity..................................................................................................... 52 2.4.2 Copper Complexation........................................................................................... 54 2.5 62 Discussion............................................................................................ 2.5.1 Organic Acidity..................................................................................................... 62 2.5.2 Copper Complexation........................................................................................... 63 2.6 Conclusions....................................................................................................................67 3. CHAPTER 3. THE EFFECT OF ENVIRONMENTAL AND EXPERIMENTAL TRANSFORMATION AND FRACTIONATION OF DISSOLVED ORGANIC MATTER ON ORGANIC COPPER (II) COMPLEXATION................................. 68 3.1 Abstract..........................................................................................................................6 8 3.2 Introduction................................................................................................................... 69 3.3 Site Description............................................................................................................. 72 3.4 Methods..........................................................................................................................72 3.4.1 Sample Collection................................................................................................. 72 3.4.2 Experimental Trans&rmation and &actionation..................................................74 3.4.3 Sample Concentration and Exposure to the Cation Exchange Resin.................. 76 3.4.4 Sample Analysis.................................................................................................... 77 3.5 Results............................................................................................................................78 3.5.1 Environmental Transformation and Fractionation of Dissolved Organic Matter (DOM).................................................................................................... ,78 3.5.2 Experimental Transformation and Fractionation of Dissolved Organic Matter (DOM)................................................................................................................. 83 3.5.2.1 Photochemical Decomposition........................................................................ 83 3.5.2.2 Microbial Decomposition................................................................................ 85 3.5.2.3 Adsorption........................................................................................................ 8 8 Page 4 77^ 3.6 Discussion .............................................................................................................. 8 8 3.6 .1 Environmental Trans&rmation and Fractionation of Dissolved Organic Matter (DOM).................... 88 3.6.2 Experimental Transformation and Fractionation of Dissolved Organic Matter (DOM)................................................................................................................. 90 3.6.2.1 Photochemical Decomposition......................................................................... 90 3.6.2.2 Microbial Decomposition................................................................................. 92 93 3.6.2.3 Adsorption........................... 3.7 Conclusion................................................................................................................... 93 4. CHAPTER 4. SUMIVIARY AND CONCLUSIONS................................................ 95 5. LITERATURE CITED............... 98 List of Tables Table 2.1. Literature values of conditional stability parameters (Log(Ki), Log(Kz)) and Cu^^ complexing edacity (CCi, C Q ; mol C-g'^) for whole water and fulvic acid factions of dissolved organic matter (DOM).................................................................. 55 Table 3.1. Dissolved organic carbon (DOC) concentration and parameter values of the 2ligand Langmuir Isotherm 6 )r reference samples &om study lakes spanning a gradient in cumulative hydrologie residence time (CHRT, yr)........................................81 List of Figures Figure 2.1. Acid dissociation (pK*) and charge density (CD) parameter values for allochthonous and autochthonous sources of DO M ........................................................53 Figure 2.2. Parameter values of the Langmuir Isotherm model for allochthonous DOM as a function of DOC concentration.................................................................................... 57 Figure 2.3. Parameter values of the Langmuir Isotherm model for autochthonous DOM as a function of DOC concentration.................................................................................... 58 Figure 2.4. Parameter values of the Langmuir Isotherm model for allochthonous and autochthonous DOM as a function of &ee labile Cu^^ concentrations corresponding to titration end-points.......................................................................................................59 Figure 2.5. DOC concentration dependence of copper (Cu^^ binding density (CuL, pmol C-g'^) calculated at 1.0 pmol L'^ of 6 ee labile Cu^^ for allochthonous and autochthonous DOM .......................................................................................................61 Figure 3.1. M ^ of study a re a ................................................................................................ 73 Figure 3.2. Potentiometric titration curves for allochthonous and autochthonous DOM and for DOM of increasing cumulative hydrologie residence time (CHRT) for reference and treated samples..........................................................................................79 Figure 3.3. Molar concentration of Cu^^ complexing sites (C ul^), dissolved organic carbon (DOC) concentration and speciSc absorbance (SAC350) as a function of increasing cumulative hydrologie residence time (CHRT).............................................82 Figure 3.4. Ratio of treated to reference values for parameters of the 2-ligand Langmuir Isotherm as a function of cumulative hydrologie residence time (CHRT) and for allochthonous and autochthonous sources of D O M ........................................................84 Figure 3.5. Ratio of treated to reference values for DOC concentration as a function of cumulative hydrologie residence time (CHRT) and for allochthonous and autochthonous sources of D O M ...................................................................................... 8 6 Figure 3.6. Net loss of DOC concentration with the cation exchange resin pretreatment step as a function of cumulative hydrologie residence time (CHRT).............................87 P age 7 q/"77^ Acknowledgments I would like the opportunity to thank a number of people without whom this work would not have h^ypened. # To my academic supervisor, Dr. Ellen Petticrew, who provided support that was instrumental in the successful completion of this project # To my research supervisor. Dr. JefT Curtis, who guided me through the practical and theoretical aspects of the this project # To the members of my committee. Dr. David Lean and Dr. Todd Whitcombe, who provided very useful input into this project # To Dr. Richard Playle and Dr. John Crusius, who collaborated on this project and have helped me progress professionally as a scientrGc researcher # To David Aikinstall, who provided technical support in the lab and helped me resolve some of the theoretical aspects of this project # To Kristin Mueller and Robert Bunn, who played an important role in sample preparation and data collection # To Erinn Radomske, who reviewed my work on countless occasions, provided technical expertise in areas ofbiology, and most importantly, provided the emotional support required to complete this endeavor List of Acronyms, Symbols and Units ABS350 Photometric absorbance at an optical wavelength of 350nm CCi.CCz Parameter values of complexing capacity empirically approximated 6 0 m a 2-hgand Langmuir Isotherm that correspond to the density of organically complexed Cu^^ (mol C-g-') CD Charge Density (mol C-g'') calculated as function of pH CDr Total Charge Density (mol C-g'') CER Cation Exchange Resin C-g Grams of carbon measured for DOM CHRT Cumulative Hydrologie Residence Time (yr) Cu^+ Copper (n) CuL Copper binding density (Cu-pmol C-g'') CuLw Molar concentration of Cu^^ binding sites (Cu-pmol L'') calculated 6 0 m empirical titration curves at 1.OE-06 mol L'' of &ee labile Cu^^ and ambient DOC concentrations Cu-mol Moles of Cu^^ CV Coefficient of Variation d Diameter DO Dissolved Oxygen (mg L"') DOC Dissolved Organic Carbon (C-mg L'') DOM Dissolved Organic Matter EDTA Ethylenediaminetetra-acetate g Grams H" Hydrogen ion HO Hydrochloric acid HRT Hydrologie Residence Time (yr) ICP Inductively Coupled Plasma ISE Ion Selective Electrode Kf Formation constant for the complexation of metal ions KHP Potassium Hydrogen phthalate Kpa Kilopascals L Liters Log(Ki), Log(K2) Parameter values empirically approximated 6 0 m a 2-ligand Langmuir Isotherm that correspond to the binding afGnity of organic ligands M Molar concentration mg Milligrams mL Milliliters |imol Micromoles mn Mean or average mol Moles n Sample size nm Nanometre °C Degrees Celsius P Probability that an observation is consistent with the null hypothesis PAR Photosynthetically Active Radiation pH The negative of the base-10 logarithm of hydrogen ion concentration pKa Acid dissociation parameter - relates the concentradon of an acid to the concentration of its ionization products pR-w Water 6 )rmation constant r Correlation - measures the strength and direction of linear relationships Square of the correlation - measures the haction of the variation in the values of y that is explained by the least squares regression of y on x RO Reverse Osmosis SAC350 SpeciGc Absorbance CoefGcient calculated at a photometric ABS wavelength of 350nm fnge 70 q/"77& Standard Deviation UV Ultraviolet UVR Ultraviolet Radiation Preface Research that was conducted for this thesis is part of a larger research initiative that focuses on systematic control over the concentration and composition of dissolved organic matter (DOM) in surface &eshwaters. The following research was conducted but not included in this thesis: scavenging of DOM by iron hydroxide particles as a function of cumulative hydrologie residence time (CHRT); watershed control over DOM in alpine lakes; the relationship between optical properties of DOM and CHRT; and a comparative study of Cu^^ toxicity measured by rainbow trout bioassay, a copper (H) ion selective electrode and diffusion gradient thin film samplers. fngg 72 q/"77^ 1. Chapter 1. Introduction and Literature Review 1.1 Introduction Dissolved organic matter (DOM) is approximately 40% carbon by weight (Perdue 1998), measured as dissolved organic carbon (DOC), and is transformed over time in the environment as a result of biological and chemical degradation processes (BufHe and De Vitre 1994; Brezonik 1994). Three of these processes are photochemical decomposition, microbial decomposition and adsorption (Wetzel 1983; Buffle and De Vitre 1994; Lindell et nZ. 2000). These types of processes will alter the chemical composition of DOM, which may influence its function as a metal complexing ageiit. DOM is composed of organic acids, which can complex metals and play an important role in regulating trace metal spéciation dynamics in surface hreshwaters (Sunda and Lewis 1978; Azenha et oZ. 1995; MacRae et nZ. 1999b). Therefore, a change in organic acidity may have a signiScant impact on aquatic ecosystems by an increase in metal toxicity. The objective of this research is to identify systematic variation in organic acidity and metal complexation as a function of DOM source and exposure to transformation and ûactionation processes. DOM is loaded to aquatic systems 6 om two sources. The 6 rst is &om the terrestrial environment (allochthonous) and the second is from within lakes (autochthonous) (Sposito 1981; Wetzel 1983; Pukushima ez uZ. 1996). Allochthonous DOM is derived from the degradation of plant material, and is transported to lakes by various hydrologie pathways. Autochthonous DOM is derived 6 om primary productivity within lakes (Sposito 1981; Wetzel 1983; Fukushima eZnZ. 1996). These differences in derivation result in fundamental differences in chemical composition (Fukushima eZ nZ. 1996). Thus, the composition of Fnge Z3 q/"ZZ& DOM loaded to surface 6 eshwaters can be investigated using allochthonous and autochthonous sources of DOM for comparison purposes (Luider ef of. in Press). Photochemical decomposition, microbial decomposition and adsorption processes change the properties of DOM over time in the environment because these processes transform and &actionate DOM (Strome and Miller 1978; Geller 1986; Kieber et a/. 1990; Gu et a/. 1995). For example, DOM plays an integral role in metabolism because it is a fundamental food source for microbes (Wetzel 1983; Moran and Hodson 1990; Sun et a/. 1997). Microorganisms, such as bacteria, may consume portions of dissolved organic molecules for energy. Excess carbon is transformed to microbial waste products, such as carbon dioxide, which are lost to surrounding waters (Wetzel 1983). Dissolved organic molecules are &actionated as a result of this decomposition process, and the overall chemical properties of DOM are changed (K ^lan and Newbold 1995; Volk et a/. 1997; Moran et a/. 2000). Transformation and hractionation processes may also be coupled m fztw to collectively change the properties of DOM over time. For example, photolysis may be coupled to microbial decomposition (Lindell et a/. 1995; Miller and Moran 1997; Moran and Zepp 1997a). Natural ultraviolet radiation (UVR) can cleave simple organic molecules &om DOM, which can stimulate microbial metabolism (Strome and Miller 1978; Wetzel et a/. 1995; Moran and Zepp 1997a). Thus, photolysis and microbial decomposition have a combined influence on the concentration and composition of DOM in surface &eshwaters. Due to this combined influence, it is difBcult to measure the individual effects of each process on the overall concentration and composition of DOM. The magnitude and rate of photochemical decomposition, microbial decomposition and adsorption processes is difBcult Page 7^ q/"77<ÿ to investigate in surface 6 eshwaters because of the complex mixture of physical, biological and chemical activity occurring in .y/fw (Wetzel 1983; Frimmel 1990; BufHe and De Vitre 1994). As a result, it is difBcult to investigate the influence of speciBc transformation and fractionation processes on the metal complexing function of DOM. Aspects of the complex nature of aquatic systems can be simplif ed by isolating transformation and fractionation processes in micro-reactors. The effects of each process on the concentration and composition of DOM can then be investigated directly, and compared to the transformation and factionation of DOM in surface freshwaters (Buffle and De Vitre 1994; Gu ef a/. 1995). Furthermore, characteristic changes in DOM can be used to investigate the influence of each process on organic acidity and the metal complexing function of DOM. Metal complexation is one of the most important functions of DOM, and it involves metal binding by organic ligands to form a single integral molecule (Heiing and Morel 1988a; Buffle 1990; Christensen ef oZ. 1999). It was determined in the 1970's that trace metal toxicity is a function of 6 ee ionic and not total metal concentrations (Brown ef a/. 1974; Pagenkopf gf nZ. 1974; Mantoura ef oZ. 1978; O'Shea and Mancy 1978). Thus, the ability of DOM to complex metals contributes to the regulation of trace metals in surface ûeshwaters (McKnight ef nZ. 1983; Buffle 1990; Bruland 1992). For example, Cu^^ is readily complexed by DOM, which can reduce the toxicity of Cu^^ in surface Aeshwaters (Mantoura gf nZ. 1978; McKnight and Morel 1979; Winner 1985; Hering and Morel 1988b; Meador 1991; Azenha gf oZ. 1995; Erickson gf nZ. 1996). The contribution of organic complexation as a regulating process of 6 ee ionic metals, such as Cu^^, has been investigated by many diSerent researchers (Playle and Dixon 1993; fagg ZJ q/"ZZ& MairgfoZ. 1999; MacRae ef a/. 1999a). DOM has been observed to signiûcantly influence &ee ionic concentrations in lake systems (Pagenkopf ef a/. 1974; Sunda and Lewis 1978). However, the protective ejects of DOM are unclear at environmentally signiûcant Cu^^ concentrations, which can be as low ^ 2.0 x 10'^° M (Anderson and Morel 1978). Therefore, organic complexation may not completely reduce the toxicity of Cu^\ and only thermodynamically or kinetically stable Cu^^ complexes may be important h"om an ecotoxicological perspective. The function of DOM as a Cu^^ complexing agent may also change over time as DOM is transformed and 6 actionated in surface ûeshwaters. Transformation and huctionation processes alter the chemical composition of DOM, which may inhibit the formation of stable metal complexes. Subsequently, the susceptibility of surface Aeshwaters to Cu^^ toxicity may vary as a function of exposure to trans&rmation and &actionation processes. Therefore, the influence of trans&rmation and Aactionation processes on the Cu^^ complexing function of DOM needs to be investigated. 1.2 Hypothesis Organic acidity and Cu^^ complexation by DOM vary systematically according to DOM source and exposure to environmental transformation and factional processes. 1.3 Significance This research addresses one of the most important topics in enviromnental toxicity. Freshwater metal contamination is quickly becoming a global problem because of the Page 76 q/"77<9 atmospheric transport of metal contaminants (Nriagru and Pacyna 1988). Dissolved organic matter (DOM) is one of the most signiGcant substances in Geshwater that is available in sufBcient concentrations to mediate metal toxicity (Perdue 1998; Rozan and Benoit 1999). As global populations continue to rise, metal contamination will undoubtedly be a problem of the future as well as the present (Runnels et a/. 1992; Myers 1993; Cohen 1995). In order to sustain the biodiversity of aquatic Gora and fauna, resource management guidelines must better address metal contamination (Savage 1995; Naeem and Li 1997). A better understanding of the role of DOM in metal complexation processes will aid in the development of these resource management guidelines. The sensitivity of lakes to metal contamination will be a function of DOM concentration and composition. However, these properties of DOM vary among systems and may change as a function of exposure to transformation and Gactionation processes. Therefore, it is crucial to determine how changes in DOM inGuence metal complexing processes, especiaUy since changes in global climate may increase trans&rmation and Gactionation processes, such as photolysis (Kerr and McElroy 1993; Schindler et nZ. 1997). DeGning the effects of DOM transfbrmaGon and GacGonaGon on metal complexaGon is cnGcal to the development of site speciGc metal regulaGons. If metal sensiGve lakes can be idenGGed on the basis of DOM concentraGon and composiGon, then more speciGc management guidelines can be developed and implemented. It is not unGl the role of organic metal complexaGon is better understood that the problem of metal contaminaGon can be better regulated. f a g e 7 7 q/"Z7& 1.4 Literature Review f.4. f Sysfemaffc varfab/Z/fy of OOW The concentration and composition of dissolved organic matter (DOM) in aquatic systems is dependent on loading 6 om allochthonous and autochthonous sources and exposure to transformation and ûactionation processes. Exposure to transformation and ûactionation processes is approximated by hydrologie residence time (HRT; Eq 1.1) For example, changes in colour and dissolved organic carbon (DOC) concentration are inversely dependent on HRT for enclosure experiments (Curtis and Schindler 1997). Similarly, HRT was used to ^)proximate losses of DOM due to mineralization and sedimentation in lakes (Engstrom 1987). Therefore, HRT can be used as an independent variable to investigate the eSects of transformation and ûactionation processes on organic acidity and organic Cu^^ complexation. Increasing HRT will be associated with a transition ûom reactive (labile) to less reactive (reûactory) DOM (Luider ef of. in Press). For example, systems that exhibit a long HRT should exhibit an accumulation of photo-reûactory DOM and a decrease in photochemical activity (Lindell ef aZ. 2000). As DOM is exposed to more ultra violet radiation (UVR), photochemically labile components of DOM will be preferentially consumed in photochemical reactions (Brezonik 1994). The reûactory components of DOM will resist photochemical degradation processes and will accumulate over time (Lindell ef a/. f age Z&of ZZ& 2000). Therefore, the rate of photochemical decomposition processes will decrease over time as re&actory DOM accumulates, (Clair and Soyer 1997; Lindell er a/. 2000). The same reasoning can be applied to other trans&rmation and ûactionation processes. For example, as DOM is decomposed by microbial assemblages, biologically reactive components of DOM will be preferentially consumed and biologically unreactive components will accumulate (Brady et al. 2000). As the labile components of DOM are consumed over time, the bioavailability of DOM is reduced (Leff and Meyer 1991; Brady et al. 2000). There have been similar Sndings 6 r adsorption of DOM to Fe(0 H)3(g) (Luider er n/. in Press). This transition ûom labile to reûactory DOM with increasing HRT will vary among aquatic systems because some systems may receive DOM that has been pre-exposed to transformation and ûactionation processes. As a result, DOM entering a system may exhibit a high proportion of reûactory DOM. Therefore, lakes in series may result in a cumulative effect of HRT. As DOM passes ûom one lake to the next, the accumulation of reûactory DOM would successively increase. This effect can be approximated by cumulative hydrologie residence time (CHRT)(Luider ef of. in Press). f.4.2 j4//ocûfûonous and Aufocûfûonous DOW Dissolved organic matter (DOM) is loaded to aquatic systems ûom allochthonous and autochthonous sources. Sources of allochthonous DOM ûom the terrestrial environment primarily depend on the chemical and biological breakdown of plant material (Wetzel 1983; Fukushima et of. 1996; McKnight and Aiken 1998). Leaching processes in soils will mobilize soluble organics, which wiU result in influxes of dissolved organic acids to creeks Page q/"77^ and streams (Singer and Munns 1987). The bulk of these leachates are derived from animal and microbial digestion of plant material (Ertel er oZ. 1984; Singer and Munns 1987; Carter 1996; Fukushima ef uZ. 1996; Hessen eTuZ. 1997; McKnight and Aiken 1998). The concentration and composition of allochthonous DOM produced 6 om soils wiU generally be a function of carbon supply fo m terrestrial plant material and regional climate (Carter 1996; Hessen gZuZ. 1997; Curtis 1998). The type of terrestrial plant material will affect the composition of DOM produced (Ertel gZuZ. 1984; Singer and Munns 1987; Carter 1996; McKnight and Aiken 1998). The concentration of DOM produced will depend on the supply of plant material, temperature and soil moisture (Singer and Munns 1987; Carter 1996; Curtis 1998). Subsequently, the mobilization of organic matter will depend on hydrologie flows through the watershed, which will largely be a function of precipitation and basin morphology (Wetzel 1983; Schindler gZnZ. 1997; Curtis 1998). DOM produced 6 om the terrestrial environment consistently has a few characteristic properties (McKnight and Aiken 1998). Allochthonous DOM is usually richly coloured as a result of high aromaticity and is thought to be chemically polyfunctional, which means that it exhibits a variety of chemical functional groups (Meili 1992; Brezonik 1994; Coveney and Wetzel 1995). Although, carboxyl and hydroxyl groups are thought to constitute the bulk of the chemical functional groups for allochthonous DOM (Bufde gZuZ. 1990; Manahan 1994). In contrast to allochthonous DOM, autochthonous DOM is derived internally within a lake system (Wetzel 1983; Fukushima gZaZ. 1996). The production of autochthonous DOM is primarily derived fn m the by-products of primary production within lakes, which is generally dominated by phytoplankton (Wetzel 1983). Factors controlling primary productivity with lakes include nutrient availability, temperature, dissolved oxygen (DO) fagg 20 ZZO concentrations and photosynthetically active radiation (PAR) (Wetzel 1983; Roberts ef aZ. 1992; Zhang and Prepas 1996). Each of these parameters vary temporally as a result of physical, biological and chemical lake processes and seasonal climate changes (Wetzel 1983). Therefore, the greatest influx of autochthonous DOM will occur when the optimal combination of conditions is achieved, which will generally occur in the spring. Although the exact composition is difBcult to determine, autochthonous DOM is thought to be composed of amino acids, proteins and polysaccharides (BufBe and De Vitre 1994; Brady et al. 2000). These types of molecules exhibit less aromatic character and less colour, as compared to allochthonous DOM (Meili 1992; BufBe and De Vitre 1994; Curtis 1998). These types of molecules are also generally smaller and less complex (BufBe and De Vitre 1994) (BufBe 1990). Carboxyl and hydroxyl functional groups will be present, but in lesser abundance than allochthonous DOM (Zumstein and BufBe 1989; BufBe and De Vitre 1994). Diflerences in chemical composition cause allochthonous and autochthonous DOM to function diSerently. As in any chemical process, the composition of the reactants will determine the outcome of the reaction (Kotz et al. 1994; Radel and Navidi 1994). For example, many components of allochthonous DOM are chemically inactive or refractory and resist transformation and f-actionation processes (Filip 1985; Manahan 1994). In contrast, organic compounds released during autochthonous DOM production may be chemically reactive or labile (McKnight and Morel 1979; Wilhelm and Trick 1994; Meador eT a/. 1998). It is useful to consider allochthonous and autochthonous sources of DOM separately because the bulk properties of DOM in aquatic systems can be simplif ed, even though the bulk of DOM is derived fo m allochthonous sources (Curtis 1998). The characterisfc properfes of Page 27 q/"77& each source of DOM can be compared as end-members to the function of DOM in surface freshwaters (Luider er uZ. in Press). Y.4.2 Y /so/afmg a//oc/?fhonous and aufocAf/tonoas OOM Allochthonous DOM can most easily be isolated within creeks and streams because the DOM in these type of systems is usually primarily 6 om allochthonous sources (Cummins er oZ. 1982; Wetzel 1983; Bilby and Bisson 1992). Phytoplankton and attached algae can produce small amounts of autochthonous DOM, but this source of DOM is minimized in smaller streams because photosynthetically active radiation (PAR) is limited by riparian canopy (Wetzel 1983; Bilby and Bisson 1992). There&re, smaller streams should provide a sufGciently isolated source of allochthonous DOM. Autochthonous DOM is more difScult to isolate because most systems contain a mixture of DOM ûom allochthonous and autochthonous sources. The m^ority of DOM is loaded ûom allochthonous sources and allochthonous DOM is generally ubiquitous in ûeshwaters as a result (Wetzel 1983; Zumstein and Buffle 1989; Curtis 3998). One ^iproach to alleviating this problem is to produce autochthonous DOM within an enclosure (Pukushima et aZ. 1996). Primary producers can be isolated in a mesocosm so that the growth environment can be controlled to maximize productivity and allochthonous sources of DOM can be excluded. Page 22 q/"ZZ^ f.4.3 Trans^rmaÉfom and fracdonadon of OOAf f.4.3. f P/)ofocAem/ca/ OecomposA/on Photochemical decomposition will contribute to changes in the concentration and composition of dissolved organic matter (DOM) over time and will occur directly as a function of UVR (Brezonik 1994; Wetzel er a/. 1995; Morris er a/. 1995). UVR is composed of photons of energy according to Planck's Law (Radel and Navidi 1994; Kotz et al. 1994; Brezonik 1994), and DOM contains chromophores that absorb photons of energy, giving rise to photochemical reactions (Brezonik 1994; Clair and Soyer 1997). Photochemical reactions can result in chemical bond rearrangements and bond dissociation, which can result in the transformation and fractionation of DOM in surface ûeshwaters (Kieber er a/. 1990; Skoog and Leary 1992; Brezonik 1994). These types of photochemical processes will alter the chemical and optical characteristics of DOM. Photolysis can result in net losses of carbon, decreases in colour and changes in the chemical composition of DOM, such as a changes in carboxyl character (Kieber er a/. 1990; Kieber ei n/. 1999; Lindell ef a/. 2000). The degree of these alterations will depend on the initial concentration and composition of DOM in surface ûeshwaters, the duration of exposure and the intensity of UVR (Mopper er aZ. 1991; Brezonik 1994; Wetzel era/. 1995). Latitudes north and south of the equator will experience seasonal variation in UVR. In northern latitudes, UVR will decrease over winter and increase over summer (Ahrens 1998). Thus, photochemical activity will potentially be maximized throughout the spring and summer. However, transmittance can also regulate availability and intensity. For example, ice and snow cover can reduce the amount ofUVR that enters the water column (Wetzel Page 23 o f / / 3 1983; Ahrens 1998). Similarly, cloud cover and atmospheric particles can result in decreases in UVR (Ahrens 1998). The seasonal variation of these parameters will generally be a function of solar elevation and climate (Ahrens 1998). Photochemical decomposition of DOM is also coupled to other processes m .sifw, such as photosynthesis and metabolic activity. The UVR and light attenuating properties of DOM can regulate photosynthesis and metabolic activity by regulating the euphotic zone. Highly concentrated and coloured DOM will effectively protect microorganisms bom harmful UVR, but will also limit photosynthetic activity (Morris er a/. 1995; Sommaruga er of. 1999). In contrast, low concentrations of weakly coloured DOM may result in a high transmittance of light, which may promote photosynthetic activity, but subsequently inhibit metabolism by increased exposure to harmful UVR (Lind and Hongue 1994). There is also the possibility of a positive feedback mechanism. Biodégradation may increase the susceptibility of DOM to photochemical reactions (Miller and Moran 1997). Similarly, photo-transfbrmation and bactionation of DOM can produce simple organic molecules that can stimulate microbial activity (Wetzel ef a/. 1995; Moran and Zepp 1997b; Lindell ef a/. 2000). Therefore, photochemical decomposition and microbial decomposition processes may funcbon interdependently to degrade DOM in surface beshwaters. Y.4.3.2 M/crob/a/ OecomposA/on Microbial decomposition of dissolved organic matter (DOM) will contribute to changes in the concentration and composition of DOM over time, and will occur directly as a function of metabolism. DOM is readily utilized by aquatic microbes, such as bacteria, as a fundamental energy source by processes of assimilation and respiration (Wetzel 1983; Brady et al. 2000). The result of these processes is the complete or partial breakdown of DOM. Complete breakdown results in the total trans&rmation of an organic molecule and typically occurs as a result of carbon assimilation into microbial cell structures (Moran and Hodson 1990). Assimilated carbon is digested by the mechanism of respiration and therefore, some carbon is lost to carbon dioxide. Assimilated carbon can also be released as microbial exudates and cellular debris, which can result in influxes of simple sugars, proteins and simple fatty acids into the water column (Brady et al. 2000). Influxes of these compounds into the water column (Park et nZ. 1997) (Buffle and De Vitre 1994)may result in changes to the chemical and optical characteristics of DOM (Fukushima et o/. 1996). Partial breakdown of DOM occurs by the same process as complete breakdown except only a fraction of an organic molecule is assimilated. The usable fractions of DOM will be preferentially consumed over time (Geller 1986; Sun er aZ. 1997). As portions of an organic molecule are removed and transformed, the original organic molecule is f-actionated. Therefore, f-actionation may result in a selective decrease in the molecular weight of organic molecules over time, which may influence the characteristic properties of DOM (Leff and Meyer 1991; Sun er aZ. 1997). The chemical and optical characteristics of DOM will likely change over time as a result (Sun er aZ. 1997). The degree DOM is decomposed will vary as a function of microbial activity. Microbial activity wül depend on nutrient availability, the concentration and composition of DOM, temperature, dissolved oxygen (DO) and photosynthetically active radiation (PAR) (Wetzel 1983; Brady et al. 2000). Each of these variables will vary seasonally, which results in temporal variation of microbial decomposition processes. Conditions that will maximize fagg ZZ& microbial activity within a lake will generally occur after the spring turnover and the spring melt when nutrient availability, DOM and DO concentrations wül be maximized (Wetzel 1983). As a result, microbial assemblages will usually exhibit maximum productivity &r approximately two months in the spring (Wetzel 1983). Y.4.3.3 Adso/pf/on to /mn Oxy/tydmx/des and Hydmx/des Adsorption to iron (Fe) oxides will contribute to changes in the concentration and composition of dissolved organic matter (DOM) over time, and will occur directly as function of surface complexation processes (Buffle and De Vitre 1994). Functional groups of DOM can react with Fe ox)4iydroxide and hydroxide particle surfaces (Schlautman and Morgan 1994; Buffle and De Vitre 1994; Gu et uZ. 1995). This reaction is known as surface complexation and can result in complete or partial fractionation of DOM, resulting in net losses of DOM from surface f-eshwaters (Buffle and De Vitre 1994; Lambert and Graham 1995; Luider et nZ. in Press). The primary source of Fe in lakes arises from the terrestrial environment (Buffle and De Vitre 1994). Thus, the spring melt should result in an infux ofFe to aquatic systems. The majority of ferrous species (Fe^^ are oxidized to ferric species (Fe^^) in surface feshwaters, typically occurring as iron oxyhydroxide and hydroxide (Fe(OH)s(,)) precipitates, which can then fractionate DOM (Wetzel 1983; Buffle and De Vitre 1994; Stumm and Morgen 1995). As Fe particles enter a lake system they will pass through surface freshwaters and settle to the sediment-water interface (Buffle and De Vitre 1994). During settling, chemical adsorption and desorption processes can result in the redistribution of DOM (Buffle and De Page 2d q/"ZZ3 Vitre 1994; Gu ef a/. 1994). For example, under high redox conditions Fe^^ can complex organic ligands, nutrients and trace metals, resulting in Fe-hydroxide and Fe-phosphate species (BufQe and De Vitre 1994). As the Fe particulates settle through the water column, the ambient redox conditions can fall (Wetzel 1983; BuSle and De Vitre 1994). The subsequent reduction of oxidized iron complexes can result in the release of complexed surface groups (Manahan 1994; Stumm and Morgen 1995). The result is a net loss of DOM 6 om surface freshwaters, and the loss may be specifc to organic functional groups that are susceptible to surface complexation. f.4.4 Copper /n Systems Y.4.4.Y Sources and e/Tecfs More research focus has been given to the topic of Cu^^ contamination in the last 20 to 30 years because it is a growing problem and because it can have serious toxicological effects in aquatic environments. Contamination of Cu^^ has been steadily increasing over the last 400 years (Sposito 1981; Nriagru 1996). Most aquatic systems have been estimated to exhibit up to a seven fold increase in Cu^^ concentrations as compared to background concentrations, which on a global scale has been approximated as 3.9 x 10"^ mol L"^ (Nriagru and Pacyna 1988). Furthermore, Cu^^ contamination is no longer considered a localized problem. Although global estimates are vague, metal contamination is quickly becoming of global signiGcance (Nriagru and Pacyna 1988). Increases in Cu^^ concentrations within aquatic systems are not directly related to increases in toxicity because it was determined in the 1970's that C u^ toxicity is strongly correlated to fe e ionic or labile Cu^^ rather than total concentrations. It was determined that Page 27 q/"77^ &ee ionic within aquatic systems can be regulated by inorganic and organic complexation processes (Brown et a/. 1974; Pagenkopf et a/. 1974; Mantoura et a/. 1978; O'Shea and Mancy 1978). Similar studies followed throughout the 1970's and the early 1980's (Sunda and Guillard 1976; Andrew et oA 1978; Sunda and Lewis 1978; Wood 1983; McKnight et a/. 1983; Borgmann and Ralph 1984). The more recent investigations address site specrGc parameters, such as pH, water hardness and alkalinity, which can influence Cu^^ complexation processes and the subsequent availability of Cu^^ to aquatic organisms (Cusimano et at. 1986; Lauren and McDonald 1986; Welsh et aZ. 1993; Welsh et a/. 1996). Depending on the organism, 6 ee labile Cu^^ can be toxic in very low concentrations. For example, at concentrations of ^proximately 2.0 x 10"^° mol L'^ dinoflagellates were observed to be 100% non-motile (Anderson and Morel 1978). At higher concentrations, such as 3.2 X 10'^mol L '\ whole microbial ecosystems can be killed (Meador et aZ. 1998). In contrast, the 6 eshwater algal species Ooeyftü jwfZZZa was found to be uninhibited by 6 ee labile Cu^^ concentrations ranging &om 8 x 10"^ to 20 x 10"^ mol L'^ (Meador et aZ. 1998). It should also be noted that Cu^^ is a micronutrient (Wood 1983; Wetzel 1983; Buffle and De Vitre 1994; Azenha et aZ. 1995). Therefore, a sensitive balance exists between deûciency and excess of available Cu^^ concentrations. The uptake and resulting toxicological effect of Cu^^ also varies among organisms (Ariza et al. 1999). In general however, the environmentally relevant range, ûom a toxicology perspective, seems to be 6 om Cu^^ concentrations of approximately 1 .0 x 1 0 "^mol L'^ down to the detection limits of most analytical techniques. fuge of ZZ& V.4.4.2 C om p/exaf/onofC u^byO O A f The availability of complexed varies according to the type of complex that is formed (Buffle 1990). Under typical pH and redox conditions the aqua Cu^^ complex (Cu(H2 0 )4^^) will dominate, which is considered as 6 ee ionic or labile Cu^^(Meador 1991; Kotz et al. 1994; Biicker and Jones 1995). However, Cu^^ can also form complexes that are unavailable to aquatic organisms (Wetzel 1983; Buffle 1990; Manahan 1994; Milne er of. 1995) because Cu^^ can be complexed by a variety of different ligands and bonding arrangements (Kettle 1969; Wul6 berg 1987; Radel and Navidi 1994). For example, if a ligand binds to more than one site on a metal cation, then it is referred to as a chelating agent (Kettle 1969; Buffle 1990; Martell and Hancock 1994). These different chemical and structural combinations of ligands will determine the availability of complexed Cu^^ to undergo other reactions, which is usually discussed in terms of thermodynamic and kinetic stability (Buffle 1990; Martell and Hancock 1994). Cu^^ will preferentially reside in the most stable form possible, which is why weakly complexed Cu^^, such as Cu(H2 0 )4^^, can be bioavailable (Buffle 1990; Buffle and De Vitre 1994; Martell and Hancock 1994). Thermodynamic stability will depend inversely on the 6 ee energy associated with a metal complex under equihbrium conditions (Radel and Navidi 1994; Kotz et al. 1994). Complexation reactions that are associated with low free energy will be more stable, and thus more thermodynamically preferential (Kettle 1969; Buffle 1990; Kotz et al. 1994). Kinetic stability involves the mechanism that facilitates the complexation reaction (Radel and Navidi 1994; Kotz et al. 1994). A reaction may be thermodynamically favoured, but is limited by a viable mechanism that will produce thermodynamically stable end products (Kettle 1969; Kotz et al. 1994). Thermodynamic stability depends on two types of interactions between the metal and the ligand. The hrst is the electrostatic and covalent interaction between the metal and ligand (Buffle 1990; Kotz et al. 1994). The second is the coordination geometry between the metal and ligand, which involves a discussion of chelation (Wulfsberg 1987; Radel and Navidi 1994; Kotz et al. 1994). One way of ^proximating the electrostatic and covalent interaction between a metal and ligand is by hard and soft character (Kettle 1969; Cotton and Wilkinson 1976; Wulfsberg 1987; Buffle and De Vitre 1994). Hard cations preferentially participate in electrostatic interactions and will bind with hard donor atoms. Soft cations preferentially participate in covalent interactions and bind with soft donor atoms (Cotton and Wilkinson 1976; Wulfsberg 1987; Buffle and De Vitre 1994). The hardness of cations is determined on the basis of charge and radius (Cotton and Wilkinson 1976; Wulfsberg 1987; Buffle and De Vitre 1994). Small, highly charged cations, such as hydrogen, will be compact and resist electrostatic polarization (Cotton and Wilkinson 1976; Radel and Navidi 1994; Buffle and De Vitre 1994). These metals include the group I and n metals. In contrast, soft cations include the heavier transition metals, such as mercury and lead, which are usually larger, exhibit more complex valence conGgurations and tend to be more polarizable (Cotton and Wilkinson 1976; Radel and Navidi 1994; Buffle and De Vitre 1994). Cu^^ has both hard and soft characteristics and will function as both a hard and soft metal cation (Cotton and Wilkinson 1976; Wulfsberg 1987; BufQe and De Vitre 1994). Hard and soft ligand donors can be separated into two groups; inorganic and organic. Hard inorganic ligands in natural waters will generally include simple molecules, such as carbonates, sulfates and phosphates, which are rich in oxygen (Buffle and De Vitre 1994). fo g e JO q/"77 J Hard organic ligands include carboxyl, hydroxyl and phenolic groups, which are also rich in oxygen (BnSle and De Vitre 1994). In contrast, soA inorganic and organic hgands will include nitrogen and sulAir species, such as sulfhydryl (-SH) groups (Wetzel 1983; Wul6 berg 1987; Bufde and De Vitre 1994). The interaction of hard metals and ligands does not usually produce stable complexes, probably because of thermodynamic or kinetic instability. In contrast, soA metals can Arrm very stable complexes with soA hgands (Cotton and Wilkinson 1976; Wulfsberg 1987; BufAe and De Vitre 1994). The main difference between hard and soA metal interactions is the covalent nature of the bonds involved. SoA metals are larger, more easily polarized, and react with soA hgands that readily exhibit a lone pair of electrons (Kotz et al. 1994; Radel and Navidi 1994; Buffle and De Vitre 1994). These conditions promote strong covalent interactions, which may increase the stabihty of metal complexes (Kotz et al. 1994; Radel and Navidi 1994). Therefore, the covalent nature of soA metal interactions likely contributes to stabihty, and may be a useful consideration when investigating Cu^^ complexation processes. Chelation reactions can also increase thermodynamic stabihty because a hgand binds to a central metal ion in two or more places (Marteh and Hancock 1994). Stabihty is increased when donor groups of the hgand are approximately positioned to coordinate the central metal ion (Kotz et al. 1994; Martell and Hancock 1994). The Aee energy of the complex is lowered "because of a built in rigidity [that] results more or less in the donor groups being Aozen into a position favorable for complex formation" (Martell and Hancock 1994). Thus, chelating agents are usually large molecules with a functional group orientation that may favour metal complexation. Page 37 q/"77& A good example of a chelating agent is ethylenediaminetetraacetate (EDTA). EDTA contains six potential binding sites that can interact with a metal ion (Manahan 1994; Radel and Navidi 1994). Four of these sites are carboxylic groups, which would be considered hard donors that would not preferentially react with soft metals (BufBe and De Vitre 1994; Radel and Navidi 1994). However, EDTA is one of the strongest known complexing agents of heavy metal species, such as Cu^\ Pb^^ and Hg^^ (Morel and Hering 1993; Manahan 1994). Therefore, the stability of Cu^^ complexes will depend on a combination of variables. The electrostatic and covalent interactions between C u^ and ligands of DOM may influence complex stability. However, large molecules with a plethora of potential donor sites, such as organic acids of DOM, may chelate Cu^^, which can increase stabihty. Therefore, chelation processes may also be important to consider when investigating Cu^^ complexation processes. The stabihty of metal complexes is conventionally quantihed by the formation constant of the complexation reaction, which is essenhaUy equal to the concentration of the reaction products [Cu-Ligand,] divided by the concentration of the reactants [Cu^^; Ligand] under equihbrium conditions (Radel and Navidi, 1994; Kotz et al, 1994)(Equation 1.2). Metals will preferentiahy reside in the most stable form possible. Therefore, a strong shift in complexation reactions to the metal complex side of the equihbrium indicates a preferentially stable state, which is quantihed by a higher A good example is the reaction Page 32 q/"77A between free ionic Cu?^ and the chelating agent EDTA. The fbnnation constant for EDTA and is approximately 3.16 x 10^, which means that under equilibrium conditions essentially all 6 ee ionic Cu^^ is complexed by EDTA (Morel and Hering 1993). Since A/values also provide a measure of stability, they approximate the susceptibility o f organically complexed Cu^^ to competing ligands, such as biological surfaces (Buffle and De Vitre 1994). For exaniple, EDTA may complex Cu^^ bound by DOM because the value associated with EDTA-Cu complexes is high. Similarly, if biological surfaces provide higher stability, then Cu^^ will preferentially bind to the biological surfaces (Buffle and De Vitre 1994). These types of substitution reactions will be governed by the dissociation and formation kinetics between each stepwise reaction in the overall process (Kettle 1969; Kotz et al. 1994; Radel and Navidi 1994). However, not all reactions are reversible. Thus, more thermodynamically stable situations may exist, but a mechanism to facilitate the reaction may not. Stable complexes will be unlikely to undergo dissociation because the activation energy required to facilitate the reaction is high (Kotz et al. 1994; Radel and Navidi 1994). In contrast, unstable complexes, such as many simple inorganic complexes, will readily undergo dissociation reactions and more thermodynamically stable organic metal complexes may result (Morel and Hering 1993; Brezonik 1994). The overall interaction between metal cations and ligands will depend on the ambient chemical conditions. Dissociation and formation reactions will be influenced by pH, group I and n metal concentrations and redox potential (Wetzel 1983; Buffle and De Vitre 1994; Brezonik 1994). Hydrogen (H ^ behaves like a hard metal and will not necessarily form strong complexes; however, at low pH values the concentration of H^ will drastically exceed trace metal concentrations, which will result in strong competition for ligands (O'Shea and Mancy 1978; Morel and Hering 1993). Even if only a small proportion of H^ are active, the m^ority of ligands will become protonated. High pH values can influence the interaction between metals and ligands through the formation of inorganic metal hydroxide complexes. Cations in solution act as Lewis acids, which results in aqua metal complexes (Kotz et al. 1994; Radel and Navidi 1994). The hydroxide ion is an excellent Lewis base (Kotz et al. 1994; Radel and Navidi 1994). As hydroxide concentrations increase with increasing pH, metal hydroxides will be preferentially formed. Depending on the Lewis character of the transition metal, hydroxide precipitates will form with increasing pH (Kotz et al. 1994; Radel and Navidi 1994). Groiq) I and H metal cations can also result in competitive effects. Group I and n metals, such as sodium, magnesium and calcium, can be present in moderately high concentrations (Wetzel 1983). These metals can influence trace metal complexation reactions in two ways. Firstly, group I and II metals can compete for complexing sites, either on DOM or biological surfaces (Playle and Dixon 1993; Richards and Playle 1998). The resulting complexes are usually weak and probably do not signihcantly aSect Cu^^ toxicity (Lauren and McDonald 1986; BufEe and De Vitre 1994; Brezonik 1994; Erickson er oZ. 1996). Secondly and most importantly, high concentrations ofhard metals can influence the polyelectric Geld at the surface of organic molecules (Tipping 1993; De Wit er a/. 1993a; Bufde and De Vitre 1994; Milne et a/. 1995). The net negative character and chemical function of DOM as a metal complexing agent may be reduced as a result. Finally, redox conditions can influence complexation processes by affecting the oxidation states of metals and ligands (BufQe 1990; Stumm and Morgen 1995). For Page 34 ax'ZZ& example, interactions between and Fe hydroxide surfaces may change according to redox conditions. Under sufhciently negative redox conditions iron hydroxide can be reduced 6 om ferric to ferrous iron species, which may subsequently react to form reduced species, such as iron sulphides (Wetzel 1983; BufBe and De Vitre 1994). Cu^^ will likely he released when these reduced iron species form and similar reactions may occur with ligands of DOM. f.4.4.3 MeasunngO/pan/c^Cfd/fya/idCf/* Comp/exaf/oa Copper (Cu^^ complexation by DOM can be quantised by a number of diSerent approaches. One ^iproach is organic acidity, which accounts for the concentration and composition of organic acids that comprise DOM (Lydersen 1998). Another approach is to directly quantify organic Cu^^ complexation (Buffle et a/. 1977; Saar and Weber 1980; Hales gf u/. 1999). Potentiometric titration analyses can be used in each approach (Midgely and Torrance 1991), thus providing information on organic pH buffering and Cu^^ complexation. Furthermore, only a fraction of organic acids will realisf cally contribute to Cu^^ complexaf on. This acfve fac fo n of organic acids can be quantifed by comparing measures of organic acidity to measures of Cu^^ complexation. Organic acidity is measured by ftrating sanq)les of acidifed DOM with sodium hydroxide (NaOH) and recording the subsequent changes in pH with a pH electrode. Concentrafons of organically complexed protons (H ^ are then calculated by difference of added hydroxide (OH^ ion concentrafons and measured changes in pH according to the water fbrmafon constant (pK*) to produce a charge density measure (CD). The resulting ftrafon data is modeled to produce total charge density (CD?) and acid dissociafon (pK.) parameters (Milne et aZ. 1995; Lydersen 1998), which provide an approximation of total ionizable binding site density and the pH of binding site ionization respectively. The determination of CDr and pK* parameters is limited by sensitivity to low concentrations of organically bound because the relevant pH range of organic acid dissociation is generally 6 om 2.5 to 4.5 (Leuenberger and Schindler 1986; Tipping and Hurley 1992; De Wit et nZ. 1993b; Driscoll and Lehtinen 1994; Lydersen 1998), and because pH measurements are typically only accurate to ±0.1 pH units (Midgely and Torrance 1991). These accuracy limitations correspond to H^ ion concentrations that range &om 1.5 x 10"^ to 2.0 X 10"^ M respectively 6 om pH 2.5 to 3.5. Concentrations of organically complexed H^ must sufSciently exceed these concentrations to measure organic acidity with conGdence. Organic Cu^^ complexation is measured similarly to organic acidity. Samples of DOM are titrated with copper sulfate (CUSO4) and the subsequent increase in the concentration of Cu^^ ions is measured with a cupric ion selective electrode (ISE). Concentrations of organically complexed Cu^^ are then determined by difference of added and measured Cu^^ according to Nemstian response of the ISE (Midgely and Torrance 1991). The resulting titration data is typically modeled to produce complexing capacity (CC) and conditional stability (Log(K)) parameters (Bresnahan et oZ. 1978; McKnight et nZ. 1983; Xue and Sunda 1997; MacRae er oZ. 1999b), which empirically correspond to the density of complexed Cu^^ (CuL) for a given concentration of Gee labile Cu^^ (BuSle 1990). Potentiometric analyses of organic Cu^^ complexation are sensitive to experimental conditions, such as temperature and most importantly, the ionic composition of the sample matrix (Avdeef eZ aZ. 1983; Belli and Zirino 1993), which is why addition of an ionic strength adjuster is typically recommended (Midgely and Torrance 1991). The ionic composition of Page 3d q/"ZZ3 the sample matrix will influence processes of Cu^ complexing by a change in Cu ^ ion activity. Similarly, cationic species in solution, especially other transition metals, can compete with Cu^^ for binding sites (Lauren and McDonald 1986; Meador 1991; Tipping 1993), and interfere with the Nemstian response of the ISE by changing the activity coefBcient for Cu^^ (Midgely and Torrance 1991). Therefore, variation in the ionic composition of the sample matrix among samples can increase instrumental variation, which is a particularly important consideration when low concentrations of 6 ee ionic Cu^^ are being investigated (Avdeef et a/. 1983). It is important to consider low concentrations of 6 ee ionic Cu^^ because environmentally significant levels of Cu^^ complexation by DOM should be addressed. It is ligands that have a high afGnity for Cu^^ that will efkctively reduce the availabihty of Cu^^ in surface &eshwaters by forming stable complexes (Vasconcelos gf a/. 1997; Marr e/ a/. 1999). Furthermore, Cu^^ toxicity ranges widely 6 om approximately 1.0 x 10"^° to 1.0 x 10"^ M among aquatic biota. Concentrations of high afGnity ligands that buSer Cu^^ to within this range may be below the detection limits of potenGometric Gtration analyses in some surface Greshwaters (BufGe gf a/. 1990). Therefore, potenGometric GtraGon analyses may not provide an accurate approximaGon of organic Cu^^ complexaGon in füw for some surface Goshwaters because of sensiGvity limitaGons to high afGnity Cu^^ binding sites. f.4.4.4 Pmfmafmenf mqu/mmenfs The sensiGvity of potenGometric GtraGon analyses to organically complexed concentraGons and high afGnity Cu^^ sites can be maximized by increasing the total concentraGon of organic acids and exposing samples to a caGon exchange resin (CER). The fa g g 37 q/" total coiicentration of organic acids can be increased by sample concentration in a reverse osmosis (RO) concentrator (Serkiz and Perdue 1990; Clair et aZ. 1991), and will increase the signal to noise ratio for potentiometric analyses. Exposure of samples to the CER replaces inorganic cations in solution, such as Mg^^ and Ca^^, with hydrogen ions, which will m inim ize differences in the ionic composition of sample matrices, thus reducing instrumental variation. Each of the pretreatment processes may increase the sensitivity of potentiometric titration analyses, but may also impose certain limitations that are relevant to an investigation of Cu^^ complexation. Sample concentration by RO wül increase ionic strength, which may change the spatial orientation of potential Cu^^ binding ligands to account for the increased concentration of repulsive and attractive forces. Simüarly, exposure of samples to the CER may alter Cu^^ complexation chemistry by reducing competition with other cations. Tbere&re, results of potentiometric analyses are subject to these experimental conditions. 1.5 Conclusions of Introduction and Literature Review Dissolved organic matter (DOM) is composed of organic acids and is an important regulator of trace metals, such as Cu^^, in surface ûeshwaters. Ligands of DOM can complex Cu^^ and these complexation processes will vary according to the concentration and composition of organic acids in aquatic systems. The overall objective of this thesis is to identify systematic variation in organic acidity and Cu^^ complexation processes as a function of DOM loading and increasing hydrologie residence time (HRT). DOM loading 6 om allochthonous and autochthonous sources will influence the initial concentration and composition of DOM and HRT provides a good ^iproximation of DOM exposure to Page q/"ZZ& transformation and fractionation processes. Three of these processes include photochemical decomposition, microbial decomposition and adsorption, all of which can change the concentration and composition of DOM over time. The Grst step in addressing the overall objective of this thesis is to determine how to measure organic acidity and Cu^^ complexation by DOM. These issues are considered in chapter 2 of this thesis, where the reproducibility and Cu^^ potentiometric titration analyses is measured, and the effects of increasing dissolved organic carbon (DOC) concentration on potentiometric titration analyses is investigated. The second step in addressing the overall objective of this thesis is to test the hypothesis that organic acidity and Cu^^ complexation vary systematically as a function of DOM loading and HRT. This hypothesis is tested in chapter 3 of this thesis, where the Cu^^ complexing properties of DOM are compared among samples of allochthonous and autochthonous DOM and among samples of DOM 6 om surface freshwaters spanning a gradient in cumulative hydrologie residence time (CHRT). In addition, DOM is treated in micro-reactors that maximize rates of photochemical decomposition, microbial decomposition and adsorption processes to better determine how each process contributes to changes in organic Cu^^ complexation in surAce f-eshwaters. Page JP q/" 2. Chapter 2. Allochthonous and autochthonous dissolved organic carbon (DOC) concentration dependence of proton and copper(ll) potentiometric titration analyses 2.1 Abstract The dependence of proton (IT^ and copper (Cn^^ potentiometric titration analyses on dissolved organic carbon (DOC) concentration was investigated for allochthonous and autochthonous sources of dissolved organic matter (DOM) using a Langmuir Isotherm model. Parameter values and their reproducibility 6 om the model were calculated for a gradient of DOC concentrations ranging 6 om approximately 5 to 435 C-mg L '\ Organic acid composition was controlled by concentrating samples in a reverse osmosis (RO) concentrator and then serially diluting samples to produce the DOC concentration gradient. Results indicate that the reproducibility of charge density (CD) and acid dissociation (pK.) parameters is limited by the accuracy of a pH electrode at ambient DOC concentrations in the m^ority of aquatic systems. The reproducibility of Cu^^ complexing edacity (CC%, CC2) and conditional stability (Log(Ki), Log(K2)) parameters is mainly limited by a low density of strong Cu^^ binding sites, particularly in the case of autochthonous DOM. In addition, condition stability parameters (Log(Ki), Log(K2)) are inversely dependent on titration end­ point concentrations of &ee labile Cu^^. The result is higher Log(Ki) and Log(K2) values with lower titration end point concentrations. This effect is similar to the dependence of parameter values on DOC concentration with respect to Cu^^ complexation, which is consistent with a combination of two processes; multidentate complexation of Cu^^ and occlusion of Cu^^ binding ligands. Regardless of the mechanism, DOC concentration could explain 20% to 60% of the reported variation in parameter values of the Langmuir Isotherm model. Therefore, DOC concentration should be considered in the comparison of organic fuge 40 of and complexation processes among diOerent systems or sources of DOM. This research indicates that allochthonous DOM exhibits a higjher total charge density (CDr) and a higher density ofhigher afGnity Cu^^ complexing sites, as compared to autochthonous DOM. 2.2 Introduction Dissolved organic matter (DOM), typically measured analytically as dissolved organic carbon (DOC), is partly composed of organic acids that buSer pH and contribute to the complexation of trace metals in surface heshwaters (Perdue and Lytle 1983b; De Wit et a/. 1993b; Lydersen 1998). The ability of DOM to complex metals has been studied for some time and is one of the most important factors affecting trace metal complexation in aquatic systems (McKnight et o/. 1983; BufHe 1990; Buffle and De Vitre 1994). Organic acidity and metal complexation processes are of particular concern in surface waters of freshwater ecosystems because of acidic rain and anthropogenic metal contamination (Nriagru and Pacyna 1988; Urban gf aZ. 1990; Lydersen 1998). The ability of DOM to control pH and to complex metals, such as copper (Cu^^, will vary according to the concentration and composition of weak organic acids (Morel and Hering 1993; Buffle and De Vitre 1994; Buffle and De Vitre 1994). These properties of DOM are typically investigated directly by potentiometric titration analyses because &ee labile (or non-complexed) H^ and Cu^^ ion activity can be measured under the ^propriate experimental conditions with a pH and cupric ion selective electrode (ISE), respectively (Sjdberg and Ldvgren 1993; Lydersen 1998). Experimental conditions, such as pH, temperature and ionic strength, must be controlled, and measurements should be made within the appropriate sensitivity window for potentiometric titrations to be as reproducible and accurate as possible (Midgely and Torrance 1991; Hales et a/. 1999). Measurements of proton (H^ activity in water are routinely conducted between pH 3 and 10 (Leuenberger and Schindler 1986; Ephraim ef aZ. 1989; Midgely and Torrance 1991), but the literature is less conclusive for measurements of free labile Cu^^. A cupric ISE can be fag^e ^2 q/"f f & calibrated to produce a linear Nemstian response between pCu 3 (1.0 x 10"^ mol L"^) and 19 ( 1 .0 X 1 0 "^^ mol L'^) in buffered solutions because excess ligands counter the dissolution of Cu^^ 6 om the electrode surface that cause drift in response (Avdeef ef u/. 1983; Hales er a/. 1999). Manufacturers however, recommend that the lower end of the linear calibration range is ^rproximately between 1.0 x 10'^ and 1.0 x 10"^ mol L'^ in non-buSered solutions. The latter corresponds to a range that is more typically used for collecting potentiometric titration data because of the relative ease in the calibration procedure and the relevance to natural systems (Cabaniss and Shuman 1988; Xue and Sunda 1997; Hales ct a/. 1999; MacRae et a/. 1999b). Potentiometric data for titration curves of DOM reflect a continuous distribution of organic ligand types that can be discretely modeled to empirically investigate the average bulk properties of organic ligands. Dissociation of organically bound H^ 6 om organic acids is pH dependant and is measured as charge density (CD), which can be modeled with a Langmuir Isotherm model to approximate an acid dissociation (pK.) parameter (Buffle 1990). Similarly, Cu^^ complexation and 6 ee labile Cu^^ activity can be approximated for the determination of complexing capacity (CCi, CCz) and conditional stability (log(Ki), LogfKz)) parameters. A m inim al number of 6 ee parameters are included in the model according to tests of signiGcance using least-squares regression analyses, and therefore, these parameters do not reGect speciGc ligand classes, but rather an empirical means of comparison (Perdue and Lytle 1983b; Town and Filella 2000). The reproducibility of parameter values 6 om the Langmuir Isotherm model depends on the sensitivity of potentiometric titration analyses to organically complexed H^ and Cu^^ ions. This issue is particularly relevant to titrations of H^ complexation because the bulk of Page ^3 q/"77& organic acid dissociation typically occurs below pH 3 (Lydersen 1998). Therefore, most of the acid dissociation occurs throughout a low pH range where small changes in proton activity are difBcult to discern 6 om the high background signal of H^ ions. Cu^^ titrations may have similar sensitivity limitations injveakly buffered samples with low total concentrations of Cu^^ binding sites (Avdeef et a/. 1983), as indicated by a low DOC concentration and a low density of Cu^^ binding ligands (Midgely and Torrance 1991). The upper and lower extreme of organic Cu^ complexation is generally observed in DOM that is derived 6 om the catchment (allochthonous), and DOM that is derived 6 om primary production (autochthonous), respectively (McKnight and Aiken 1998; Richards et a/. 2001). The bulk of DOM in aquatic systems originates hom allochthonous sources (Meili 1992) and is characteristic of a higher complexing capacity (CC^CCz) ofhigher afGnity binding sites (higher Log(Ki) and Log(K2)), as compared to autochthonous DOM (Town and Filella 2000). Consequently, there is greater precision for measurements of Cu^^ complexation in allochthonous DOM. However, current research suggests that Cu^^ binding varies among systems (Hirose 1994; Rozan and Benoit 1999), and ambient concentrations of DOC vary signiGcantly among surface water samples, ranging 6 om less than 1 to over 100 C-mg L'^ in extreme cases (Curtis and Adams 1995; Curtis and Schindler 1997; Curtis 1998). Titration analyses conducted on samples with low Cu^^ binding density and low DOC concentrations may not accurately reflect complexation reactions that are relevant to real world concentrations of Cu^\ which have been approximated at 3.97 x 10"^ (SD = 4.35E-08) mol L'^ (Wetzel 1983; Borg 1995). Sample concentration by reverse osmosis (RO) and exposure to a cation exchange resin (CER) increases the sensitivity of potentiometric titration analyses to organically bound Fogg 44 and ions by increasing the total concentration of organic acids, and particularly high afGnity Cu^^ complexing sites (Serkiz and Perdue 1990; Clair et a/. 1991). However, sample concentration imposes potential limitations to the interpretation of Cu^^ titration data because a 1:1 ligand-metal ratio is typically assumed (BufGe 1990; Morel and Hering 1993). This ratio may increase as a function of DOC concentration because of physicochemical changes to organic ligands of DOM. For example, Cu^^ may undergo processes such as multidentate complexation or binding site occlusion with increasing DOC concentration. Multidentate complexation would be consistent with an increased afGnity for Cu^^ complexation (higher Log(Ki) and LogCKz)) and both processes would be consistent with a decrease in complexing capacity (lower CC] or CC2) (Manahan 1994; Stumm and Morgen 1995). There were a number of objectives to this research. The Grst objective was to assess the reproducibility of H^ and Cu^^ potentiometric titration analyses, with particular reference to organic acid dissociation below 3 and high afGnity Cu^^ complexing sites. The second objective was to determine the sensitivity of the Langmuir Isotherm model to titration end­ points of Aee labile Cu^^ concentration. Titration data were modeled with the Langmuir Isotherm at each successive titration point for samples of DOM. The third objective was to test for effects of increasing DOC concentration on organic C u^ complexation, as indicated by parameter values of the Langmuir Isotherm model. We controUed for organic acid composition by concentrating and then serially diluting samples of DOM to provide a DOC concentration gradient. All analyses were compared among split samples of allochthonous and autochthonous DOM to determine the influence of organic acid composition on each research objective. Page 45 o/"77& 2.3 Methods 2.3. Y Samp/e Co/fectfom and Preparadon Allochthonous dissolved organic matter (DOM) was collected 6 om a Srst order stream in the Okanagan Basin of British Columbia Canada (50° 47' N, 15° 40' W). Autochthonous DOM was produced in a clear acrylic mesocosm (1 m^) with almost 100% light transmittance throughout the visible range. The mesocosm was Elled with dissolved organic carbon (D0C)-6ee deionized water, amended with salts and nutrients, and inoculated with plankton collected onto a glass Gber Glter (Whatman GF/C) 6 om a eutrophic lake. Reagent grade salts were added to simulate 1/10* of the concentrations of global average river water (Wetzel 1983). Concentrations of phosphorus were amended to 0.1 mg L"\ but no nitrogen was added to favor the growth of cyanobacteria, with the dominant taxon being jp. The mesocosm was stored outdoors in a locked compound for 6 )ur months (6 om May to August), where it was exposed to ambient lighting conditions. 7» vivo fluorescence, measured at a wavelength of 660nm with an excitation wavelength of436nm on a Shimadzu RF 1501 spectrofluorophotometer, and DOC concentration were measured weekly to monitor the growth of the plankton community and production of autochthonous DOM. Autochthonous DOM was collected at the beginning of the stationary phase of plankton growth. Approximately 200 L of water 6 om allochthonous and autochthonous sources were collected and Sltered through 142mm-glass Sber ûlters to remove seston, however colloidal and bacterial size factions are generally not removed by this process. The samples were then concentrated using a stainless steel reverse osmosis (RO) concentrating unit (Limnological Research Corp., Kelowna, B.C.). The RO unit consisted of a stainless steel Page 46 Deleon centrifugal pump, a stainless steel holding tank and a Filmtec FT30 US Filters thin composite RO membrane housed in a stainless steel cartridge. Operating pressure was between 700 and 1200 kPa. The permeate was discarded and the retentate was recycled to a final volume of 5 to 10 L. Concentrated sample was drained into acid washed glass bottles and stored at 4°C. Concentrated samples were exposed to an Amberlite IR-118H cation exchange resin (CER) to replace potentially interfering cations with and to fully protonate organic acids. The resin was replenished with 4 mol L'^ HCl and washed ten times with ultra pure water between each sample treatment. Leaching of organics and inorganics from the resin was negligible after the tenth wash with ultra pure water. The resin was added to samples until further additions failed to lower pH (generally about pH 2). Concentrated samples that were exposed to the cation exchange resin were then serially diluted to produce a concentration gradient of DOC for allochthonous and autochthonous DOM. DOC concentrations ranged hrom 3.7 to 437.6 C-mg L'^ for allochthonous DOM and 6 om 3.5 to 435.1 C-mg L"^ for autochthonous DOM. Diluted samples were then split into triplicate samples for each analysis. DOC concentration was measured for all samples with a Shimadzu TOC-5000A Total Organic Carbon Analyzer. Dissolved inorganic carbon was removed by acidifying samples (pH < 2) with select grade hydrochloric acid (2N HCl) and purging with oxygen for seven minutes behare analysis. The coefBcient of variation 6 >r instrumental response was consistently less than 2% and the detection limit of the instrument was 0.05 C-mg L'^ at an iiyection volume of 750 pL. 2.3.2 Ofigan/c^c/d/fy Organic acidity was measured by titrating 20 mL of each diluted sample with 0.05 mol L"' NaOH (standardized to 0.05 mol L"^ potassium hydrogen phthalate (KHP)) 6 om pH 2.0 to 6.0 using a QC Mantech Autotitrator and a glass double junction Ag-Ag/Cl pH electrode. The pH electrode was calibrated using the autotitrator with stabilization criteria of ± 0.05 pH/min according to a 3-point calibration curve (pH 2 ,4 and 7; p < 0.01, r^ = 0.99)(Ephraim ef a/. 1989). Experimental conditions used Air calibration were the same as for sample analyses, including ambient temperature (20.1°C, SD = 0.8"C, n = 2848) and constant lighting conditions with a Anal amended ionic strength of 0.1 mol L'' NaNOs (Midgely and Torrance 1991; Sjôberg and Lôvgren 1993). Charge density (CD) and acid dissociation (pK#) parameters were calculated for each sample Aom corrected titration data between pH 2.0 and 6.0 (Equations 2.1,2.2 and 2.3). Titration data above pH 6 were not included in the model because there was a negligible increase in CD. Corrections to titration data included background inorganic OH" consumption Aom the Airmation ofHzO and the dissociation of HSO4'. Sulfate (S0 4 ^") concentrations required for HSO4' corrections were measured using a Waters 501 HPLC with a Waters 431 Continuity detector and a Waters IC PAC 4.6 x 50mm anion column. C0 = ( 0 H , ) - ( 0 H ^ ) - ( 0 H , )' DOC Eq2.1 where, OHx is equal to OH" added (mol L'^), OHgc is equal to the ar^ustment for inorganic OH" consumpAon (mol L'^), OHm is equal to OH" measured (mol L'^), and DOC is Page 43 q/"773 in units of C-g L '\ Total charge density (CDr) was calculated operationally as CD 6 om pH 2 to 6 and is expressed as mol C-g'\ The acid Armation parameter ( ^ ) was calculated using Systat 8.0 non-linear regression analyses according to equation 2 .2 . where CO is equal to charge density (mol C-g'^), CC is complexing capacity (mol Cg'^), which is equivalent to CDr, and M is equal to OHm (mol L'^). AH non-linear regression analyses were signihcant (p < 0 .0 1 ) with only two 6 ee parameters in the model. Subsequently, the acid dissociation parameter (pK*) is calculated according to equation 2.3 and was found to be consistent with empirical titration curves in each case. = 1 4 - log(.^^ ) Eq 2.3 2.3.3 Copper 77(ra(/ons Copper (Cu^^ complexation was investigated by titrating 20 mL of each sample with l.OE-05 mol L'^ CuSO^ (standardized to l.OE-05 mol L'^ ethylenediamine tetraacetate (EDTA)) using a QC Mantech Autotitrator with an Orion solid state ion selective electrode (ISE) and a Ag-Ag/Cl sure flow double junction reference half electrode. Free labile Cu^^ concentrations were determined 6 om electrode response by cahbrating the electrodes 6 om 5.0 X 10"^ to 1.0 X 10"^ mol L"^ Cu^^ using the autotitrator with stabilization criteria of ± 0.1 Page q/"f 7& mV/min. The electrodes were conditioned with EDTA, rinsed with ultra pure water, and the Cu^^ ISE was polished between each titration to minimize electrode fouling and carry over of organic or inorganic contaminants. Nemstian response was consistently observed (29.5 ± 0.4 mV pC u'\ n = 5) under the same operational experimental conditions used for sample titrations. Samples were prepared for titration by adjusting ionic strength to 0.1 mol with NaNOg to overwhelm small diSerences in ionic strength. In addition, pH was ar^usted to 6 (± O.I pH units) for comparison with previous research, to avoid interference's due to inorganic spéciation (Midgely and Torrance 1991), and because organic acid dissociation was negligible at higher pH values. pH was adjusted under equilibrium conditions with atmospheric CO2, let stand 6 )r 2-3 hours and re-a<^usted prior to running titrations with stabilization criteria of ± 0.05 pH/min. In addition, samples were titrated at room temperature (20.1°C, SD = 0.8°C, n = 2848) and under constant lighting conditions. The titration end-point was operationally set to 1.0 pmol L'^ hree labile Cu^% as determined by the response of the ISE, because higher &ee labile Cu^^ concentrations are not relevant to typical 6 eshwaters (Rozan and Benoit 1999). Free labile Cu^^ concentrations below the lower end of the linear calibration curve were approximated assuming Nemstian response Aom the ISE because Cu^^ was buSered by excess ligands of DOM (Avdeef et aZ. 1983; Hales ef uZ. 1999). This assumption is conservative with respect to the Cu^^ complexing properties of DOM because dissolution of Cu^^ 6 0 m the surface of the ISE results in an underestimate of complexed Cu^^. Complexing edacity (CCi, CC2) and conditional stability (log(Ki), log(Kz)) parameters were calculated 6 0 m titration data by correcting for initial residual Cu^^ that was fo g e JO q/"77 & not exchanged in the CER process, and then modeling titration data using a 2-ligand Langmuir Isotherm discrete binding model (BufDe 1990)(Equation 2.4). where, v is equal to the moles of bound copper (Cu^^ per gram of carbon (Cu-mol Cg '), CC is equal to complexation capacity (mol C-g"^), M is equal to 6 ee labile Cu^^ concentration (mol L''), and Æ/ a n d a r e equal to the average conditional stability parameters for high and low afBnity Cu^^ binding sites respectively. It was assumed that the initial residual Cu^^ in samples was complexed, thereby adding to the total complexing capacity (Town and FileUa 2000). Initial residual Cu^^ was determined by inductively coupled plasma (ICP) analyses on a Leeman Labs PCIOOO. A m inim um number of 6 ee parameters were included in the model according to tests of signihcance using Systat 8.0. A 1-ligand model was not signihcant (p > 0.05) so a 2ligand model was used. All non linear regression analyses using the 2-hgand model were signihcant (p < 0.05), and parameter values were consistent with empirical titration data. Furthermore, CC] and CC2 parameter values were compared to Cu-binding density (CuL, Cu-mol C-g'^) among samples, which was operationally calculated at 1.0 x 10"^ mol L"^ of hee labile Cu^^ 6 0 m each of the titration curves. 2.4 Results 2.4. f Organ/c 4c/d/(y The lower detection limit of charge density (CD) with respect to dissolved organic carbon (DOC) concentration was calculated to be ^proximately 435 C-mg L'^ (p > 0.05) by comparison to titrations of DOC 6 ee water. Error associated with the CD model was approximately 0.01 pH units (SD = 0.009, n = 102), and 0.002 mol L"^ (SD = 8.5E-04, n = 3) of H^ complexation was measured in DOC &ee water. This background complexation was within the reported accuracy limitations (± 0.1 pH units) of the pH electrode (p > 0.1) throughout the lower end of the pH titration range (approximately 2.0 to 2.5). Given these limitations to measuring pH and subsequently calculating CD, base titrations conducted on samples with a DOC concentration below the detection limit were not included in comparisons of allochthonous and autochthonous DOC. Total charge density (CD?) ranged 6 om 0.024 to 0.006 H-mol C-g ' independent of DOC concentration for allochthonous and autochthonous DOM. Similarly, pK, values ranged &om 2.4 to 3.1 (Figure 2.1). The reproducibility (95% conhdence interval) of total charge density (CDr) and acid dissociation parameters (pK») was better for allochthonous dissolved organic matter (DOM) by 15% to 30%, compared to autochthonous DOM, and generally decreased with increasing DOC concentration (Figure 2.1). DiGkrences in CDT and pKa measures for allochthonous and autochthonous DOM were assessed using a 1-factor ANOVA. The CDT of allochthonous DOM was signiScantly higher than CDT 6 r autochthonous DOM by a factor of approximately 2.3 (p < 0.05, n = 3; Figure 2.1), with mean values of 0.018 and 0.008 H-mol C-g-1 respectively. In contrast to CDT measures, the pKa of allochthonous and autochthonous DOM was statistically similar Page 52 3.4 3.2 _ 3.0 4 ^ 2.8 4, A 2.6 Î 2.4 4 0.025 W) I U G 0.020 0.015 4 0.010 A u 0.005 0.000 100 150 200 250 300 350 400 450 DOC, mg L ' Figure 2.1. Parameter values for organic acidity plotted against dissolved organic carbon (DOC) concentration for allochthonous (closed circle) and autochthonous (open circle) sources of DOM. The 95% conGdence interval is shown for all data points (n = 3). A. Acid dissociation constants (pK«). B. Total charge density (CDr). at about 2.6 (p > 0.05, n = 3). Both measures far allochthonous and autochthonous DOM are within the reported range of variation and approximately correspond to the upper and lower extreme respectively for CD r (Table 2 . 1 ). C D r typically ranges 6 om 0.0033 to 0 . 0 2 2 mol Cg'^ (mn = 0.011, SD = 0.0040, n = 22) and pK. values typically range 6 om 2.1 to 4.8 (mn = 3.38, SD = 0.58, n = 48) for comparable titration ranges (Leuenberger and Schindler 1986; Bufde 1990; Tipping et a/. 1990; Tipping and Hurley 1992; De Wit et at. 1993b; Driscoll and Lehtinen 1994; Manuza et of. 1995; Mihie et a/. 1995; Kinniberg et a/. 1996; Lydersen 1998). 2.4.2 Copper Comp/exaftom Initial residual Cu^^ concentrations in each sample prior to Cu^^ titrations were all less than 0.32 pmol L'^ and were below the detection limit of the ICP analysis at DOC concentrations below approximately 109 C-mg L"' (p < 0.01). Measurements of 6 ee labile Cu^^ suggest that more than 99.9% of this initial residual Cu^^ is complexed, even in autochthonous samples. Titration data were corrected for these values and parameter values of the Langmuir Isotherm model were compared among corrected and uncorrected data using a 2-factor ANOVA. Parameter values were signiGcantly influenced by correcting for initial total copper concentrations (p < 0.05), but the absolute change in values was typically less than 5%. The detection limit of Cu^^ titrations was calculated as the DOC concentration at which CCi and CCz values were signiGcantly different &om zero (p < 0.05). All samples of allochthonous DOM were above this detection limit with CC% and CCz values of 38 (SD = 4.3, n = 3) pmol C-g ' and 320 (SD = 52, n = 3) pmol C-g"' respectively at DOC Page 54 q/"f 7& Table 2.1. Literature values of conditional stability parameters (Log(Ki), Log(K2)) and Cu2+ complexing edacity (CCi, CC2; mol C-g'^) for whole water and fulvic acid fractions of dissolved organic matter (DOM). Range Whole log(K:i) ( 1/mole) 7.5-13.1 CCi (pmol C-g'^) 9 -2 9 Whole 4 .6 -9 .0 6 —6100 ----- Whole 7.26 210 5.1 Fulvic 3.6 - 7.3 1900-6900 Fulvic 7 -8 .5 100-650 Fraction : K2 CC2 n Ref. ; 5.3 - 6.5 101 - 299 5 a 10 b 2890 : 5 .4 - 6 .6 440-1900 c 9 b 18 d a.) (Curtis et a/., unpublished data) b.) (Buffle 1990). c.) (MacRae et a/. 1999b). d.) (McKnight et a/. 1983). feg e JJ q/"7/^ concentrations as low as 3.71 C-mg L"' (Figure 2.2). In contrast, samples of autochthonous DOM were below the detection limit for DOC concentrations less than 27.2 C-mg L'^ (Figure 2.3). The highest CCi and CC: values measured at DOC concentrations above the detection limit were 6 . 6 (SD = 0.1, n = 3) pmol C-g"^ and 28 (SD = 38, n = 3) pmol C-g'^ respectively, which is more than a factor of 6 less than allochthonous DOM (Figure 2.2 and 2.3). The reproducibility of parameter values 6 om the Langmuir Isotherm model was calculated as a 95% conGdence interval (shown in Ggures 2.2 and 2.3) and a standard coefBcient of variation (CV). The assumption of normality was investigated Gom ten rephcate titrations of allochthonous DOM with skewness and kurtosis values not signiGcantly different Gom zero (P > 0.1). Reproducibility was better G)r allochthonous DOM, decreasing Gom a maximum CV of 16% to less than 1% for all parameter values with increasing DOC concentration (Figure 2.2). The reproducibility of parameter values for autochthonous DOM decreased Gom 58% to less than 1% with increasing concentrations of DOC above the detection limit (Figure 2.3). Sensitivity of the LangmuG Isotherm model to the Gtration end point of complexed Cu^^ (CuL, Cu-pmol C-g'^) and Gee labile Cu^^ (Cu^^, pmol L"^) was determined by sequendally removing each Gtration data point Gom the model for samples of allochthonous and autochthonous DOM. The resulting parameter values were plotted against the free labile Cu^^ concentradon associated with the tiGation end-point that was modeled (Figure 2.4). Log(Ki) and Log(K2) values decrease logarithmically as a function of increasing concentrations of complexed and Gee labile Cu^^ that were modeled as the titration end point. Similarly, CCi and CC: increased as a Gmction of increasing titration end point. This eGect was typically greater G)r allochthonous DOM and particularly for Log(Ki) values, fa g e o/"77^ 8.5 _____ 4 8.0 7.5 ! 7.0 6.5 6.0 5.5 [ 5.0 u Î 5 É 400 350 300 250 200 150 100 50 0 50 Ù 'q 100 200 300 400 D 100 200 300 400 DOC, C-mg L" Figure 2.2. Parameter values of the Langmuir Isotherm model for allochthonous dissolved organic matter (DOM) calculated at a titration end point of 1.0 pmol L"' of 6 ee labile copper (Cu^^ and plotted as a function of increasing dissolved organic carbon (DOC) concentration. The 95% conGdence limit is shown for all data points (cannot be seen for all data points). Logarithmic regression is signiGcant in each case (p < 0.05). A and B. CondiGonal stability parameters (Log(Ki), Log(Kz)). C and D. Complexing capacity (CCi and CC%). fu g e 57 q f 77^ Lf a B D Û 'o g 100 200 300 400 DOC, C-mg L'^ Figure 2.3. Parameter values of the Langmuir Isotherm model for autochthonous dissolved organic matter (DOM) calculated at a titration end point of 1.0 pmol L"^ of 6 ee labile copper (Cu^^ and plotted as a function of increasing dissolved organic carbon (DOC) concentration. The 95% conûdence limit is shown for all data points (cannot be seen for all data points). A. Conditional stability parameters (Log(Ki)). Linear regression is signiScant (p < 0.05). B. Log(K2). Logarithmic regression is signihcant (p < 0.05). C and D. Complexing capacity (CCi and CCz). 12 B 11 C: & % 8 7 6 60.0 Ü0 U 50.0 U 40.0 § 30.0 -o 20.0 i 8 10.0 0.0 — 0 0.2 0.4 8 —G • 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0 Cu^^, pmol L'^ Figure 2.4. Parameter values of the Langmuir Isotherm model calculated at titration end­ points that correspond to increasing concentrations of complexed copper (Cu^^ and plotted against corresponding concentrations of 6 ee labile Cu^^. Data shown far allochthonous (closed circle) and autochthonous (open circle) sources of dissolved organic matter (DOM) respectively at 438 and 435 C-mg L '\ Logarithmic regression is signiûcant in each case (p < 0.05). A and B. Conditional stability parameters (Log(Ki), Log(K2)). C and D. Complexing edacity (CCi and CCz). which decreased 6 om 11.2 to 8.3 as a function of increasing CuL and 6 ee labile Cu^^ concentrations included in the model (Figure 2.4). The effect of DOC concentration on Cu^^ complexation was determined 6 om titration data with an operational end-point of 1 .0 pmol L'^ of 6 ee labile Cu^^ for all samples. Parameter values vary signiGcantly (p < 0.05) as a function of DOC concentration, but generally only 6 >r allochthonous sources of DOM with Log(Ki) values increasing 6 om 7.1 to 8.4 and Log(Kz) values increasing 6 om 5.6 to 6 . 6 (Figure 2.2 and 2.3). CCi values for allochthonous DOM decreased 6 om 37.8 to 17.7 pmol C-g'^ as a function of DOC concentration and CCz values decreased &om 322 to 64.7 pmol C-g \ as compared to no effect of DOC concentration for autochthonous DOC (Figure 2.2 and 2.3). This trend is consistent with empirical titration curves. Changes in Cu^^ binding density (CuL, Cu-pmol g'^) that were operationally measured 6 om titration data at 1.0 pmol L"^ of 6 ee labile Cu^^ were consistent with the DOC concentration-dependence of CC; and CC2 values for allochthonous and autochthonous DOM. Cu^^ binding density decreases significantly 6 0 m approximately 140 to 69.3 pmol Cg'^ as a function of increasing DOC concentration for allochthonous DOM (p < 0.05, Figure 2.5), whereas diere is no effect for autochthonous DOM (p > 0.05). Copper binding densities are also signiGcantly higher for allochthonous DOM, corresponding to higher CCi and CC2 values. Comparison with literature values (Table 2.1) indicates that DOC concentraGon could potenGally explain a signiGcant percentage of the reported variaGon in parameter values of the Langmuir Isotherm model. For example, the range in Log(Ki) values is 3 to 4 Log(K%) units. ConservaGvely 20 to 30% of this range in Log(Ki) values could result Gom variaGon Page 60 o/"77^ 140 120 on 1 100 U Q G =L 80 60 kJ U 40 20 s s 0 100 200 300 400 -1 DOC, C-mg L Figure 2.5. Density of complexed Cu^^ (CuL, Cu-pmol C-g'^) calculated at 1.0 pmol L'^ of Êee labile copper (Cu^^ for increasing dissolved organic carbon (DOC) concentrations of allochthonous and autochthonous dissolved organic matter (DOM). The 95% conGdence limit is shown for all data points. Linear regression is signiGcant 6 r allochthonous DOM (p < 0.05). in the concentration of allochthonous DOC (1 to 50 C-mg L'^). Similarly for Log(K2), to 60% of the typical range in reported values could result hrom variation in DOC concentration. CCi and CC; values are approximately ten times lower than the m^ority of literature values, probably due to differences in experimental conditions, and are therefore difGcult to compare. 2.5 Discussion 2.5. f Organ/c Measurements of charge density (CD) and acid dissociation (pK,) parameters for dissolved organic matter (DOM) are limited by dissolved organic carbon (DOC) concentration. DOC concentration is the limiting factor because the total charge density (CDr) of DOM is low compared to background proton (H ^ activity throughout the relevant range of organic acid dissociation, which is between 2.1 and 3.1 for allochthonous and autochthonous DOM. Therefore, error in measuring background proton activity, which is generally dictated by the accuracy limitations of a pH electrode, results in poor resolution of organic acid dissociation (Kinniburgh et oA 1999). The overall result is an indistinguishable signal of CD at ambient DOC concentrations for the vast m^ority of aquatic systems (Curtis 1998). Total charge density (CDr) and acid dissociation (pK&) parameters measured at DOC concentrations of approximately 435 mg L'^ are consistent with literature values. To my knowledge, values for CDy and pK« 6 )r autochthonous DOM have not been reported previously in the literature. The reproducibility of these values is better for allochthonous DOM because of higher CDy, which results in a higher signal to noise ratio with respect to Page 62 q/" background proton activity. Therefore, the DOC concentration required to measure organic acidity is inversely dependent on CDp, which means that lower concentrations of DOC are required for DOM of higher CDr. DOM with a higher C D r will tend to occur in aquatic systems that are characteristic of allochthonous DOM and lower C D r will occur for autochthonous-like systems. However, pKa values are similar among sources of DOM and are consistent with the majority of reported values in the carboxyhc group range (Lydersen 1998), which have been estimated to constitute 60 to 90% of organic acids (Morel 1983; Perdue and Lytle 1983b). This consistency among pK* values supports the conclusion that there is very little detectable variation in the bulk composition of organic acids among surface ûcshwaters (Tipping and Hurley 1992). 2.5.2 Copper Comp/exaffon Higher Cu^^complexing capacity (CC%, CCz) for allochthonous DOM is consistent with a higher C D r, which supports two conclusions regarding the composition of DOM in aquatic systems. Firstly, there is a correlation between total organic functional group density and the density of organic Cu^^ binding sites (CuL, pmol C-g'^). Secondly, allochthonous DOM is more efGcient than autochthonous DOM at complexing Cu^^ (Richards et a/. 2001) because of a higher density of higher afBnity complexing sites. Both conclusions are conservative, given that difkrences between allochthonous and autochthonous DOM may be underestimated by CCi, CC2 and Log(Ki), Log(K2) parameter values of the Langmuir Isotherm model. fa g e 63 o f The comparison of parameter values &om the Langmuir Isotherm model among allochthonous and autochthonous sources of DOM is limited by two factors. The 6 rst limiting factor is the reproducibility of titration analyses to high afBnity Cu^^ binding sites because analyses are lim ited by the lower lim it of the linear calibration range B)r the cupric ion selective electrode (ISE). Depending on the composition of DOM that is being analyzed, ambient concentrations of DOC may be too low for comparisons of parameter values among samples. The increase in precision of parameter values as a function of increasing DOC concentration is due to a higher total concentration of Cu^^ binding sites (Avdeef et uZ. 1983; BufBe 1988). By increasing the total concentration of Cu^^ bindiug sites, titration analyses become more sensitive to low concentrations of high afGnity binding sites, which may be the most relevant to m processes (Coale and Bruiand 1988; BufBe ef a/. 1990; Marr ef aZ. 1999). Sensitivity to high afBnity binding sites is particularly relevant to autochthonous-like DOM. Autochthonous DOM may exhibit high afBnity Cu^^ complexing sites (Xue and Sigg 1993; Leal et aZ. 1999), but these sites are difBcult to resolve because of such low densities, approximated by CCi and CCz parameters, and low ambient DOC concentrations (McKnight et aZ. 1994). Therefore, this research suggests that autochthonous DOM makes a negligible contribution to Cu^^ complexation processes in surface B-eshwaters, as compared to allochthonous DOM. The second limiting factor to the comparison of parameter values is the sensitivity of the Langmuir Isotherm model to the operational end-point of experimental titration data. It is well documented that discrete ligand models do not reflect the complexing capacity and stability of Cu^ complexes over the entire distribution of organic functional groups (BufBe ef a/. 1990; Town and FileUa 2000). This issue has been previously investigated as the ratio of metal to ligand concentrations, which indicates that Log(Ki) values increase as the ratio of metal to ligand concentration asymptotically approaches zero (Perdue and Lytle 1983a). High afBnity Cu^^ binding sites are prefermtially consumed by Cu^^ ions (Town and FileUa 2000), but concentrations are smaU by comparison to lower afBnity binding sites when they are evaluated numericaUy (BufUe ef uZ. 1990). This drastic difference in concentration skews the distribution of Ugands modeled by Log(Ki) and Log(Kz) values to lower afBnity binding sites so that real differences in high afBnity sites are underestimated Brom parameter values of the model. The distribution of Ugands modeled by Log(Ki) and Log(K2) values wiU be increasingly skewed towards lower afBnity binding sites, as indicated by lower Log(Ki) and Log(K2) values, as a function of increasing concaitrations of complexed and Bee labile Cu^^ that are modeled. This trend is a Btting artifact of the Langmuir Isotherm model and wiU likely depend on proportional diSerences in high and low afBnity binding sites for DOM. For example, higher proportions of low afBnity binding sites will more drastically skew the distribution of Ugands modeled by Log(K%) values. Therefore, this effect is particularly relevant to aUochthonous DOM because it exhibits higher aŒnity Cu^^ binding sites and a proportionally higher capacity of lower afBnity binding sites, as compared to autochthonous DOM. This conclusion implies that the measured difference in binding afBnity among aUochthonous and autochthonous DOM is underestimated by discrete Ugand modeling. Sample concentration aUeviates problems of reproducibiUty and it is generaUy accepted that the Langmuir Isotherm model is conditional to experimental conditions, including operational titration end-point criteria (BufBe 1990; Town and FileUa 2000). However, sample concentration imposes limitations to the interpretation of titration data at ambient DOC concentrations. Titration data are sufSciently reproducible to identify a statistically signiGcant increase in Cu^^ binding afBnity ((Log(Ki), Log(K2)) and a decrease in binding capacity (CCi, CCz) as a function of increasing DOC concentration (Figure 2.2 and 2.3). These trends are consistent with at least two non-exclusive possibilities. Firstly, the overall trends are consistent with multidentate complexation of a central Cu^^ ion by organic ligands of DOM. Multidentate binding involves more than one ligand binding a central Cu^^ ion, which typically results in higher complex stability. This effect is consistent with the direct dependence of Log(K;) and Log(K2) parameter values on DOC concentration, as well as the inverse dependence of CCi and CC2 parameter values (Morel 1983; Manahan 1994; Martell and Hancock 1994). A second possibility is the occlusion of Cu^^ binding ligands. Potential Cu^^ binding sites may be blocked or effectively deactivated by physicochemical changes in the composition of DOM as a function ofDOC concentration. This eSect could contribute to the observed decrease in complexing capacity for allochthonous DOM. However, more research is required to determine the contribution of either possibility to changes in Cu^^ complexation as a function ofDOC concentration. Regardless of the exact mechanism, increases in DOC concentration could potentially explain some of the variation in parameter values of the Langmuir Isotherm model that is reported in the literature (Table 2.1). This further explanation is because DOC concentration and organic acid composition have a combined efkct on Cu^^ complexation in aquatic systems. Organic acids of the same source and composition complex Cu^^ differently at different DOC concentrations. Page dd q/" This efiect ofDOC concentration suggests that there are concessions to différent analytical t^proaches of quantifying Cu^^ complexation by ligands of DOM. One approach is to concentrate samples by reverse osmosis, thereby m m im i/in g the effect ofDOC concentration on Cu^ complexation and maximizing the precision of the analysis. However, results 6 om this type of analysis poorly reflect Cu^^ complexation at in .yiiw DOC concentrations. The alternative is to analyze samples at in .;iiw DOC concentrations, which may result in poor precision and increased variability due differences in DOC concentration because titration data are typically normalized to carbon. 2.6 Conclusions The precision and accuracy of proton (H ^ and copper (Cu^^ potentiometric titration analyses of dissolved organic matter (DOM) is directly dependent on experimental conditions, including dissolved organic carbon (DOC) concentration, which could explain 20% to 60% of the reported variation in organic Cu^^ complexation. This effect varies for allochthonous and autochthonous DOM because of differences in composition. Allochthonous DOM exhibits a higher total charge density (CDr), which is consistent with a higher density of higher afBnity Cu^^ complexing sites. It is important to recognize that these results are conditional to experimental conditions, but differences between allochthonous and autochthonous DOM are likely underestimated in this research. Therefore, autochthonous DOM likely makes a negligible contribution to organic pH buffering and Cu^^ complexation in surface feshwaters. fo g e 67 of 3. Chapter 3. The Effect of Environmental and Experimental Transformation and Fractionation of Dissolved Organic Matter on Organic Copper (il) Complexation 3.1 Abstract The effect of environmental and experimental trans&rmation and hractionation of dissolved organic matter (DOM) on organic Cn^ complexation was investigated. The efiect of environmental processes was investigated as a function of (1) DOM source, and (2) exposure to transformation and fractiohation processes, as ^iproximated by cumulative hydrologie residence time (CHRT). The effect of experimental processes was investigated by individually treating the same samples across micro-reactors that maximized rates of three commonly identiGed transformation and Gactionation processes; including photochemical decomposition, microbial decomposition and adsorptive Gactionation of DOM. Results show that Cu^^ complexation is primarily dependent on allochthonous sources of DOM and inversely dependent on CHRT (p <0.05). This trend is consistent with an inverse dependence of DOM colour and dissolved organic carbon concentration (DOC) on CHRT (p < 0.05), with values ofDOC ranging Gom 10.1 to 5.2 mg L"\ All three of the experimental transformation and fractionation processes investigated could potentially contribute to this trend, either by inGuencing DOM conqwsiGon or DOC concentraGon. The latter may be the most important variable, as indicated by a proporGonaUy greater dependence on CHRT, which results in a 60% decrease in the molar concentraGon of C u^ binding cites (CuL^i) Gom 1.7 to 0.7 pmol L"^ as a funcGon of CHRT (p < 0.05). 3.2 Introduction Dissolved organic matter (DOM) is a heterogeneous mixture of organic acids that can complex trace metals in surface heshwaters (Langford et of. 1983; BufBe 1990; Perdue 1998). Ligands of DOM contain organic functional groups that complex or bind trace metals, decreasing trace metal toxicity to biota (Sunda and Guillard 1976; Sunda and Lewis 1978; Playle and Dixon 1993). Thus, trace metal availability depends on organic complexing capacity, and complex stability. Complexing capacity is directly proportional to dissolved organic carbon (DOC) concentration by the density of binding sites, and complex stability directly depends on the binding afBnity of Ugands that constitute DOM (BufBe 1990). The concentration and composition of DOM in aquatic systems is controlled by two factors, including (1) DOM source and (2) exposure to transformation and hactionation processes (MeiU 1992; Fukushima et of. 1996; Curtis 1998; Bertilsson and Tranvik 2000). DOM is mainly loaded to aquatic systems 6 om terrestrial (allochthonous) sources, which is characteristically dark in colour (Waiser and Robarts 2000; hnai ef a/. 2001) and high in organic fmctional group density (Chapter 2). However, small contributions of DOM are also derived from sources of primary productivity (autochthonous) (McKnight et of. 1994; McKnight and Aiken 1998). Therefore, DOM in surface freshwaters is an amalgam of DOM from allochthonous and autochthonous sources and DOM that has been transformed and Bactionated by environmental processes. Three environmental processes that can transform or factionate molecules of DOM over time include photochemical decomposition, microbial decomposition and adsorption (Moran and Hodson 1990; McKnight eTa/. 1992; Volk et aZ. 1997; Bertilsson and Tranvik 2000). Page 69 q/"77^ It has been determined for some properties of DOM that environmental transformation and 6 actionation varies as a function of exposure (Curtis and Schindler 1997; Curtis 1998). This relationship was determined by an inverse relationship between hydrologie residence time (HRT), which is calculated as the quotient of lake volume and outflow, and measures ofDOC concentration and colour, as compared by a specihc absorbance coefGcient (SAC). An inverse correlation between colour and the toxicity of trace metals (Richards ef of. 2001) suggests a similar relationship between HRT and trace metal complexation by DOM in surface heshwaters. Organic trace metal complexation is commonly measured for Cu^^ by potentiometric titration analyses, which provides a direct empirical measure of complexing capacity and complex stability (Buffle ef a/. 1977; Cabaniss and Shuman 1988; Buffle 1988; Holm and Curtiss m 1990; MacRae er a/. 1999b). Limitations to this approach include a low resolution of high afBnity binding sites at m ligand concentrations, and instrumental variation due to matrix effects, such as interfering cations and variation in Cu^^ ion activity (Midgely and Torrance 1991; Hales ef a/. 1999). These limitations are minimized by operationally standardizing experimental conditions. For example, increasing DOC concentration by reverse osmosis (RO) increases the sensitivity of potentiometric titrations to high afBnity binding sites (C huter 2). Similarly, exchanging interfering cations with a cation exchange resin (CER) and ac^usting ionic strength minimizes instrumental variation (Avdeef gf a/. 1983; BufBe 1988; Midgely and Torrance 1991). Potentiometric data are conditional to experimental conditions and are typically interpreted by modeling for pairs of parameter values that empirically represent the complexing capacity (CC) and binding afBnity (Log(K)) of organic ligands (BufBe 1990). fu g g 70 Parameter values typically correspond to a low CCi - high Log(Ki) pair, termed high afGnity sites 6 )r discussion pmposes, and a high CCg - low Log(Kz) pair, termed low afGnity sites (BufGe ef a/. 1990; Town and FileUa 2000). However, it is important to recognize that these pairs of parameter values do not represent specihc ligand classes (Town and Filella 2000), but rather the bulk Cu^^ complexing properties of DOM for empirical comparisons of titration curves among samples (Perdue and Lytle 1983a). These pairs of parameter values inevitably predict Cu^^ binding by multiple hgands because DOM is a heterogeneous mixture of organic acids with a continuous distribution of binding afGnities (Perdue and Lytle 1983a; BufGe gr a/. 1990; Benedetti ef aZ. 1995). The rationale for this approach is to avoid statistical over parameterization of Gtration curves that may occur with more complicated models (McKnight et a/. 1983; Coale and Bruiand 1988). The objective of this research is to determine how environmental and experimental transformation and hactionation of DOM affects organic Cu^^ complexation. The eSects of environmental transformation and fractionation are determined by measuring Cu^ complexation in surface freshwater sampled ûom aquatic systems of increasing cumulative hydrologie residence time (CHRT), which accounts for the pre-exposure of DOM to transformation and factionation processes. The efkcts of experimental transformation and factionation are determined by treating sub-samples in micro-reactors that maximize rates of photochemical decomposition, microbial decomposition and adsorption processes. Patterns in sensitivity to experimental transformation and factionation are used to infer whether individual mechanisms are consistent with environmental transformation and factionation. Cu^^ complexation is measured by potentiometric titration analyses and empirically investigated with a two-ligand Langmuir Isotherm for comparisons among reference and Fogg 77 q/"77& treated samples. The results depend on experimental conditions, which are operationally standardized among samples. 3.3 Site Description Samples were collected hom a headwater creek (unnamed) and three downstream lakes (Duck, Wood and Kalamalka Lakes) in the Okanagan Lake watershed of the upper Columbia River (Southern Interior of British Columbia, Canada, Figure 3.1). Cumulative hydrologie residence time (CHRT) increases with flow-path because the hydrology of the lake chain is surface discharge dominated, and the climate at the elevation of the lakes is semi-arid. Thus, the lakes are analogous to batch reactors positioned in series. 3.4 Methods 3.4.1 Samp/e Co//ecOon The headwater creek sample, with a CHRT less than 1 day, is dominated by allochthonous dissolved organic matter (DOM). Autochthonous DOM was produced in an acrylic mesocosm (1 m^) as described in Chapter 2, and was therefore not included in regression analyses according to CHRT with the rest of the samples because the mesocosm is not representative of a natural system. Instead, autochthonous DOM was compared to other DOM samples as the lower extreme of colour and Cu^^ complexation (see chuter 2) Approximately 20 L of surface water were collected into polyethylene containers &om each sample site and 6 om the mesocosm in August 2001. Lake samples were collected 6 om shore, well away &om the entrance of any tributaries to attain a representative sample R age 72 q/" Canada km ^ Site CHRT 1. Creek 2. Duck Lake 3. Wood Lake 4. Kalamalka Lake <1yr ---------- 20 anaBasaiiBaa O n Figure 3.1. Map of the study area showing hydrologie flowpath (1 through 4), drainage basins and cumulative hydrologie residence time (CHRT) for each system that was sampled. of the surface mixing layer for each lake. All samples were ûltered through GF/C Whatman glass ûber hlters to remove seston and the 2 0 L Gltered samples were split into four subsamples of 5 L each; one reference and three experimental subsamples. Each set of experimental subsamples was individually treated in one of the three types of micro-reactors designed to maximize rates of photochemical decomposition, microbial decomposition and adsorption processes. 3.4.2 Expenmenfa/ Tiransfomtaffon and Aacdonadon Photochemical decomposition of DOM was experimentally investigated in the laboratory by exposing subsamples to light emitted from a high-pressure mercury v ^ r lamp. Glass tubes (295 mL) were used as reactor vessels to hlter lower wavelengths of ultraviolet radiation (UVR) emitted from the lamp (0% transmittance below a wavelength of 280nm), which are not common in the natural environment (Wetzel et nZ. 1995). Furthermore, the top of each tube was covered during exposure to minimize contamination. Exposure time was experimentally determined prior to sample treatment to ensure complete reaction of the photo-labile fraction of DOM (24hrs) and the precision of the treatment process was measured by replicate analyses (Coeffcient of variation < 5% for dissolved organic carbon (DOC), n = 11). In addition, the initial sample volume in each tube was conserved by the addition of ultra-pure water at the end of the exposure period for measurements of DOC concentration and colour, as determined by a specif c absorbance coeffcient at a wavelength of 350nm (SAC 350 ). Treated samples were collected in polyethylene sample bottles and stored in the dark at 4°C. Page 74 Microbial decomposition of DOM was experimentally investigated in the laboratory by treating subsamples in plug-flow bioSlm reactors following a similar construction to Kaplan and Newbold (K ^lan and Newbold 1995). Samples were diluted by a factor of 8 to maintain oxic conditions in all reactors, nutrient enriched with N and P (1.5 and 0.1 mg L'^ respectively with NO3 and PO 4 salts) to mimic eutrophic conditions, aerated, and hltered through in-line Whatman GF/C glass ûber ûlteirs to remove seston, however colloids and bacterial size fractions are generally not removed by this process. Reactors were constructed &om 750 mL amber ^ass bottles that were darkened by covering with tin foil and packed with glass beads (d = 1.5 mm), resulting in a potential biofilm surface area of approximately 2.1 m^. Sample residence time was ^proximately 11 hr with a flow rate of ^proximately 21 mL/br, and the treated sample was collected for a period of 12 weeks. Dissolved oxygen (DO) and DOC concentrations were measured once a week to monitor microbial activity within each reactor and to ensure aerobic conditions. Dissolved oxygen was measured in reactor inlet and outlet water. Dissolved oxygen typically decreased between 50% and 60% across reactors, indicating microbial respiration within the reactors. DOC in water from the reactors varied widely. Treated samples were collected daily in polyethylene sample bottles and stored in the dark at 4°C. Adsorptive factionation of DOM was experimentally investigated in the laboratory in batches by adding a slurry of concentrated iron hydroxide particles (Fe(0 H)3(:)) to 5 L subsamples in polyethylene bottles. The concentrated Fe(0 H)3(s) slurry was produced synthetically by ac^usting the pH of an iron (ID) chloride solution (FeCl3; 0.36 M) to pH 7 (Morel 1983). The resulting Fe(0 H)3(g) floe was then repeatedly washed until the conductivity and DOC concentration of the wash water was the same as ultra pure water. P ag e 75 o/" The volume o f concentrated Fe(0 H)3(g) slurry added to each sample was determined experimentally (Fe(0 H)3(a) 3 g L'^) to maximize adsorptive &actionation (Luider ef a/, in Press). Samples treated with the slurry were stirred for ten minutes, filtered through Whatman GF/C glass Sber ûlters, and then stored in clean polyethylene bottles at 4"C in the dark. The precision of the process was measured by replicate analyses of DOC scavenging, with a coefGcient of variation less than 5% for DOC concentrations (n = 3). 3.4.3 Samp/e Concemfraf/on and Exposure (o (he Ca(/on Exchange Res/n Reference and treated samples were concentrated using a bench-top stainless steel reverse osmosis (RO) concentrator (Limnological Research Corp., Kelowna, B.C.) with a stainless steel Deleon centrifugal pump, a stainless steel holding tank, and a Filmtec FT30 US Filters thin composite RO membrane housed in a stainless steel cartridge. Operating pressure was between 700 and 1200 f ^ a and the cartridge was cooled with a water jacket to keep samples at ^proximately room temperature (21°C). Concentrated sample was drained into 500 mL polyethylene bottles and stored at 4°C. Concentrated samples were exposed to an Amberlite IR-118H cation exchange resin (CER) to replace potentially interfering cations with and to fully protonate organic acids. The resin was replenished with 4 mol L'^ HCl and washed ten times with ultra pure water between each sample treatment. Leaching of organics and inorganics 6 om the resin was negligible after the tenth wash with ultra pure water. The resin was added to samples until further additions fmled to lower pH (generally about pH 2). Page q/" 3.4.4 Samp/e ^na/ys/s DOC concentration was measured in all samples with a Shimadzu TOC-5000A Total Organic Carbon Analyzer. Dissolved inorganic carbon was removed by acidifying samples (pH < 2) with select grade hydrochloric acid (2N HCl) and sparging with oxygen for seven minutes before analysis. The coefBcient of variation (CV) for instrumental response was consistently less than 2% and the detection limit of the instrument was 0.05 C-mg L'^ at an injection volume of 750 pL. Organic colour was measured spectrophotometricahy at 350nm with a Shimadzu UV2101PC ultra violet (UV)-Visible scanning spectrophotometer in a 1cm quartz cuvette. A specihc absorbance coefBcient (SAC350) for DOM was calculated according to equation 3.1 and reported in units of cm^ C-mg ' (Curtis and Schindler 1997). (DO C)/(lOO O ) where, pathlength is equal to 1cm and DOC is in units of mg L '\ The coefBcient of variation for this method was less than 9%, as determined by rephcate analyses on split sample (n = 12). Copper (Cu^^ complexation by DOM was measured using the same methodology and experimental conditions described previously in Chapter 2. Complexing edacity (CC], CC2) and conditional stabihty (log(Ki), log(K2)) parameters were calculated from titration data using a 2-hgand Langmuir Isotherm discrete binding model. In addition, the molar concentration of Cu^^ binding sites (CuLw, pmol L"') was operationally calculated at l.OE-06 mol L'^ of &ee labile and at m DOC concentrations. ConEdence limits (95%) for each of the parameter values and for C u l^ were approximated 6 om the results of C huter 2 as a function of DOC concentration. 3.5 Results 3.5. f EnWronmemta/ Transformation and Eracdonatfon of 0/sso/yed Organic Matter (DOikÇ Titration curves for reference samples were empirically ranked 6 om 1 to 5 according to the degree of Cu^^ complexation, which was determined by the density of Cu^^ binding sites (CuL, pmol C-g'^) for a given concentration of &ee labile Cu^^ (pmol L'^). Using these criteria, the empirical rank order 6 )r reference titration curves was the same as the rank order for cumulative hydrologie residence time (CHRT; Figure 3.2a). For example, the titration curve for allochthonous dissolved organic matter (DOM), which corresponds to the lowest CHRT (< 1 day), consistently showed the h ip e st density of Cu^^ binding sites (CuL, pmol C-g'^) for any given concentration of measured 6 ee labile Cu^^ (pmol U^). This ranking indicated that allochthonous DOM exhibited the highest CuL, reaching a maximum of 164 pmol C-g"^ at the operational titration end-point of 1 .0 pmol L"^ of hee labile Cu^^. hi contrast, the titration curve for autochthonous DOM corresponded to the lowest extreme of Cu^^ complexation, with a maximum CuL of 68.5 pmol C-g'\ With respect to the Langmuir Isotherm model, higher CuL (pmol C-g'^) corresponds to higher complexing capacity (CC%, CCz), and lower buSered concentrations of hee labile Cu^^ correspond to ligands of higher binding afBnity (Log(Ki), Log(K2)). The complexing capacity of high afGnity binding sites (CCi) and the binding afBnity of low afGnity sites A Reference Figure 3.2. Potentiometric titration curves for allochthonous (closed circle) and autochthonous (open circle) dissolved organic matter (DOM) and for DOM of increasing cumulative hydrologie residence time (CHRT; square-lyr, triangle- B Photobleached 20yr and diamond-8 lyr respectively). Curves are shown for reference (A) and treated samples (B, C and D). u 200 a 150 C Microbial Decomposition 100 50 200 D Adsorption 150 100 50 cgOgP ^ 0.2 0.4 0.6 0.8 1.0 Cu^^, |imol L'^ Page 79 q/" (Log(Kz)) were inversely dependent on CHRT (i^ = 0.99 and = 0.91 respectively; p < 0.05; Table 3.1), which was consistent with the empirical rank order for reference titration curves. CCi vaines decreased by 19% ranging 6om 39 to 32 pmol C-g'^ and Log(K2) values decreased by 4% ranging 6om 6.3 to 6.1 as a function of CHRT (Table 3.1). Furthermore, the upper and lower extreme in CCi and LogCKz) parameter values corresponded to allochthonous and autochthonous DOM respectively. Log(K;) and CC2 parameter values were independent of CHRT. Dissolved organic carbon (DOC) concentration and DOM colour, measured by a speciGc absorbance coefBcient (SAC 350), were inversely dependent on CHRT (r^ = 0.99 and r^ = 0.98 respectively, p < 0.05), which was consistent with dependence of organic Cu^^ complexation on CHRT (Figure 3.3). DOC concentration decreased by 48% &om 10.1 to 5.2 C-mg L'^ (p < 0.05) as a function of CHRT (Figure 3.3), which is a greater proportional decrease as compared to CCi and LogfKz) qualitative variables. The molar concentration of Cu^^ binding cites (Cul^), which is calculated as product of Cu^^ binding density and DOC concentration, is inversely dependent on CHRT (r^ = 0.99, p < 0.05; Figure 3.3), and is strongly correlated to SAC350 (r = 0.99). Values of CuL^, decreased by 60% ûom 1.7 to 0.7 pmol L'^ as a function of increasing CHRT, which approximately corresponds to the sum of the proportional decrease observed in parameter values (CCi and LogRz) and DOC concentration. Page Table 3.1. Dissolved organic caibon (DOC) concentration and parameter values of the 2ligand Langmuir Isotherm for reference samples horn study lakes spanning a gradient in cumulative hydrologie residence time (CHRT, yr). CHRT,yr DOC (C-mgL^) CCi (pmol C-g'^) Log(Ki) (1/mole) CCz Log(B Alio/ « 1 10.1 39 7.9 180 6.3 1 9.2 35 8.1 200 6.2 20 6.9 32 8.2 180 6.1 81 5.2 32 8.1 190 6.1 Auto 1.4 31 7.8 82 6.0 Mean 6.6 34 8.0 160 6.1 Standard 3.5 Deviation 3.3 0.2 47 0.1 fo g e o/"77^ 2.0 L-l 1.5 Ê 1.0 3 0.5 0.0 12 10 60 8 u u 6 4 o 2 o 0 60 u 15 u 10 o in (L? 5 < (/] 0 0 AUo 20 40 60 CH R T,yr 80 Auto Figure 3.3. Molar concentration of Cu^^ complexing sites (Cul^)(A), dissolved organic carbon (DOC) concentration (B) and speciGc absorbance (SAC350XC) as a function of increasing cumulative hydrologie residence time (CHRT; p < 0.05) for reference samples. The 95% confidence interval is shown for CuL^. Page 3.5.2 Experfmenfa/ Tirans/bnnaf/on and R acdonadon of 0/sso/yed Ofgan/c Afader (DOA() 3.5.2. Y P/)ofoc/?en7/c8/ Oecompos/don Titration curves for photobleached samples were empirically ranked 6om 1 to 5 as done for the reference curves. Using these criteria, the empirical rank order is exactly opposite to reference curves, with the exception of autochthonous DOM, which consistently represents the lower extreme of Cu^^ complexation (Figure 3.2b). For example, the titration curve for allochthonous DOM, which exhibits the lowest CHRT, consistently shows the second lowest density of Cu^^ binding sites (CuL, pmol C-g'^) next to autochthonous DOM for any given concentration of 6ee labile Cu^^. In contrast, the titration curve for the sample of DOM corresponding to the highest CHRT exhibits the upper extreme of Cu^^ complexation, with CuL reaching a maximum of 188 pmol C-g ' . Parameter values of the Langmuir Isotherm for photobleached samples are consistent with the order of empirical titration curves for photobleached samples (Figure 3.2b) and were compared to reference samples by calculating the percent change. Therefore, negative scores correspond to a proportional decrease in parameter values and vice versa for positive scores (Figure 3.4a). The effect of photochemical decomposition on Log(Ki) values ranged 6om 2% to -10% and is independent of CHRT, whereas changes to CCi, Log(K2) and CCz values are all dependent on CHRT (Figure 3.4a). CCi and Log(K2) values decreased by as much as -40% asymptotically as a function of CHRT, which is consistent with the environmental trend measured in reference samples (Figure 3.3). For example, the largest reproducible decrease in CCi and log(K2) values is for allochthonous DOM, which exhibits the lowest CHRT and the least exposure to transformation and huctionation processes. In contrast, CC2 Page ^3 q/"773 B 5 ^ Line of no effect j 0 -5 % ? ? -10 -15 r ! -5 -10 : -15 40 20 8 0 " ^ -20 -20 -40 -60 -60 5 5 0 •5 -10 -10 -15 200 200 150 150 U 100 < 50 ^ n -15 -50 O -50 7/ 0 AUo 20 40 60 CHRT,yr 80 0 Auto AUo 20 40 60 80 CHRT.yr Auto Figure 3.4. Percent change of treated to reference values for parameters of the 2-ligand Langmuir Isotherm as a function of cumulative hydrologie residence time (CHRT). Allochthonous and autochthonous sources of dissolved organic matter (DOM) are presented at extreme ends of CHRT scale. Data shown for photobleached DOM (A) and DOM exposed to high rates of microbial decomposition (B). Line of no effect is shown in each g r ^ h (horizontal grey dash-dot line). Trend lines only shown for signiGcant regression analyses (p < 0.05). The 95% conGdence interval is shown for aG data points. values increased by as much as +84% directly as a function of increasing CHRT. The DOC concentration of photobleached samples was compared to reference samples in the same way as parameter values of the Langmuir Isotherm (Figure 3.5a). Changes in DOC concentration range 6om -60% to -37% as a function of CHRT (p < 0.05). Furthermore, the hnal concentration of DOC in all photobleached samples is similar (4.8 Cmg L '\ SD = 1.3) to the DOC concentration measured in high CHRT systems (5.2 C-mg L'^). 3.5.22 M/cmb/a/ Oecompos/f/on Titration curves for DOM exposed to nutrient enriched microbial decomposition were empirically ranked &om 1 to 5 as done for reference curves. However, the empirical rank order is unclear because the curves tend to overlfgi (Figure 3.2c). In addition, the interpretation of the titration curves is suspect because of signiûcant losses in DOC concentration across the cation exchange resin (CER) pre-treatment step. Losses of DOC with the CER pre-treatment step range 6om -37.1 to -71.1 C-mg L'^ as a function of CHRT (p < 0.05), which corresponds to -28% and -62% respectively (Figure 3.6). This effect is not observed for any other set of reference or treatment samples, where only negligible losses of DOC are observed. Parameter values of the Langmuir Isotherm for samples exposed to nutrient enriched microbial decomposition are suspect because of the DOC losses in the CER pre-treatment step, but are consistent with the order of empirical titration curves (Figure 3.2c). The parameter values were compared to reference samples using the same criteria as photobleached samples. The effect of microbial decomposition on parameter values of the Langmuir Isotherm is independent of CHRT in all cases (Figure 3.4b). However, the fn ge q/" Line of no effect -10 -20 -30 -40 -50 -60 O 120 100 u o -20 20 0 -20 4 -40 -60 -80 € 0 AUo 20 40 60 C H R T ,y r 80 Auto Figure 3.5. Percent change of treated to reference values for dissolved organic carbon (DOC) concentration as a Amction of cumulative hydrologie residence time (CHRT) and for aUochthonous (closed circle) and autochthonous (open circle) sources of dissolved organic matter (DOM). Data shown for photobleached DOM (A), DOM exposed to high rates of microbial decomposition (B), and DOM exposed to high rates of adsorption Aactionation (C). Line of no effect is shown in each graph (horizontal grey dash-dot line). Trend lines only shown for signiUcant regression analyses (p < 0.05). Page Line of no effect u ^ -20 -40 -60 0 AUo 20 40 60 CHRT,yr 80 Auto Figure 3.6. Net loss of dissolved organic caibon (DOC) with the cation exchange resin pretreatment step as a function of cumulative hydrologie residence time (CHRT; p < 0.05) for samples that were exposed to high rates of microbial decomposition. Line of no efkct is shown (horizontal grey dash-dot line). Page eSect of microbial decomposition is positive for Log(Ki) values of allochthonous DOM and negative for Log(Ki) values of autochthonous DOM, with values of +3% and -5% respectively. Furthermore, CCi values decrease far allochthonous DOM by -29% and increase for autochthonous DOM by +3%. This trend is consistent with comparisons of reference and treated empirical titration curves (Figure 3.2). 3.5.2.3 y4dso/pf/on Titration curves for samples treated with FeOH3(,) show a signiGcant decrease in CuL (Figure 3.2d), but are suspect because DOC concentrations were below recommended concentrations (See Chapter 2), between 2 and 9 C-mg L"^ for all samples. Low DOC concentrations are due to signiGcant scavenging of DOM by FeOH3(,) parGcles, which range Gom -90% to -79% directly as a funcGon of CHRT (Luider er a/, in Press). No parameter values were produced for these GtraGon curves because Cu^^ complexaGon was near the lower extreme of detectable limits and because the Gt of the Langmuir Isotherm to experimental data was not signiGcant (p > 0.05). 3.6 Discussion 3.6. f Enwrommenfa/ Tiransfbrmaffon and Eracdonadon of 0/sso/yed Orpan/c Mader (DOA() The rank order of Cu^^ GtraGon curves for reference samples, and the dependence of CCi and Log(Kz) values on cumulaGve hydrologie residence time (CHRT), siqiports the conclusion that Cu^^ complexaGon varies systemaGcally in surface G-eshwaters as a funcGon of two factors. These two factors include (1) dissolved organic maGer (DOM) source, and Page 33 q/"773 (2) exposure to transformation and &actionation processes with increasing CHRT. The dependence of organic Cu^^ complexation on DOM loading is consistent with a critical review of published data on organic Cu^^ complexation (Town and Filella 2000), and appears to depend primarily on DOM 6om allochthonous sources (Richards et a/. 2001). Although high afGnity binding sites are present in autochthonous DOM, the density of these sites is very low. A similar trend is observed for low afGnity binding sites of autochthonous DOM, and autochthonous dissolved organic carbon (DOC) concentrations are comparatively low in surface Greshwaters (Zumstein and BufGe 1989; Meili 1992; Jonsson et a/. 2001). Therefore, the contribution of autochthonous DOM to Cu^^ complexation is hkely negligible compared to allochthonous sources. The dependence of organic Cu^^ complexation on CHRT is conservative because of an inverse dependence of Cu^^ binding density on DOC concentration, as determined in ch u ter 2. Cu^^ binding density decreases 6)r allochthonous DOM as a function of increasing DOC concentration via reverse osmosis (RO) concentration. Therefore Cu^^ binding density may be underestimated for allochthonous DOM because samples were concentrated prior to potentiometric analyses. In addihon, the dependence of organic Cu^^ on CHRT is consistent with other analytical measures of DOM, such as speciGc absorbance (SAC350) and DOC concentraGon. This relationship supports conclusions that opGcal measures of DOM, such as SAC350, can provide good ^proximations of organic Cu^^ complexation (Saar and Weber 1980; Gamble 1980; Ryan and Weber 1982). Comparisons of qualitative measures of organic C u^ complexaGon to systemaGc variability in DOC concentraGon support the conclusion that the effects of increasing CHRT are greater on DOC concentraGon. There is relaGvely litGe change in CCi and Log(Ki) parameter values with increasing CHRT, as compared to the change in DOC concentration. The result is a systematic decrease in the molar concentration of Cu^^ binding sites (C ul^) as a function o f increasing CHRT. Therefore, an inverse relationship between DOC concentration and CHRT augments systematic decreases in organic Cu^^ complexation due to changes in DOM composition. The implication of this trend to Cu^^ complexation in surface h-eshwaters is that DOM composition is of secondary importance to DOC concentration, which is consistent with current Cu^^ spéciation prediction models (Breault ef a/. 1996; Christensen ef a/. 1999; Bryan ef of. 2002). 3.6.2 Expenmenfa/ Tirans^rmaffon and A acdonadon o f 0/sso/yed Organ/c Wader 3.6.2. y P/?ofoc/7em/c8/ OecomposWon Changes in CCi and Log(K2) values for photobleached samples indicate that photochemical decomposition of DOM could contribute to the dependence of organic Cu^^ complexation on CHRT for two reasons. Firstly, the net effect of the photochemical decomposition treatment is a decrease in CC; and Log(K2) values, which is consistent with the effect of CHRT. Similarly, photochemical decomposition reduces DOC concentration, which contributes C u^ complexation by the molar concentration of Cu^^ biuding sites (CuLM). Increasing exposure of DOM to photochemical decomposition processes, such as photolysis, could therefore contribute to the inverse dependence of organic Cu^^ complexation on CHRT, which is consistent with previous research on increased metal bioavailability following DOM exposure to ultraviolet (UV) B radiation (Winch er a/. 2002). fo g e 90 o f 773 Secondly, the net effect of photochemical decomposition on CCi and Log(K2) values is inversely dependent on the pre-exposure of DOM to environmental transformation and Sactionation processes, as approximated by CHRT. For example, the largest decrease in CCi and Log(K2) values for photobleached samples is observed in allochthonous DOM, which exhibits the least pre-exposure to environmental transformation and hactionation processes (CHRT < 1 day). Losses of DOC concentration with photochemical decomposition treatment follow a similar trend. Therefore, DOM that is pre-exposed to transformation and fractionation processes is more re6actory to photochemical decomposition than unexposed DOM, with respect to Cu^^ complexation. In contrast to changes in CCi values, changes in CC2 values for photobleached samples are not consistent with the dependence of organic Cu^^ complexation on CHRT. Increases in CC2 values of the Langmuir Isotherm with increasing CHRT suggest that the net effects of photochemical decomposition processes change as a function of increasing DOM exposure to environmental transformation and 6actionation processes. Photochemical deconqx)sition may therefore result in a higher density of low afGnity Cu^^ complexing sites as DOM becomes increasingly re&actory. This increase in binding site density is not consistent with the inverse dependency of organic Cu^^ complexation on CHRT. This discrepancy suggests that decreases in DOC concentration with increasing CHRT have a greater effect on CuL^ than increases in Cu^^ binding density that may occur as a result of photochemical decomposihon. P ag e q/"77& 3.6.2.2 M/cmb/a/OecomposA/on Decreases in DOC concentration with the cation exchange resin (CER) pretreatment step suggest that the effects of microbial decomposition on DOM are directly dependent on CHRT. For example, the smallest decrease in DOC concentration is observed in allochthonous DOM, which exhibits the shortest CHRT (< 1 day). In addition, the faction ofDOC lost in the CER process likely corresponds to transformed or hactionated DOM because this effect is exclusive to microbially decomposed samples. Therefore, DOM that is pre-exposed to transformation and fractionation processes is more susceptible to microbial decomposition than unexposed DOM, as approximated by CHRT. This rate dependence of microbial decomposition on CHRT is consistent with an identifed coupling mechanism to photochemical decomposition processes because the byproducts of photolysed DOM can stimulate microbial productivity (Bushaw ef oA 1996; Moran and Zepp 1997a). The exposure of DOM to photochemical decomposition processes increases directly as a function of CHRT, and therefore, so may the byproducts of phototransformed and fnctionated DOM. However, the extrapolation of this bend to surface f-eshwaters is limited by the controlled experimental conditions of the microbial decomposition treatment method. Interpreting the effect of microbial decomposition on Cu^^ complexation is obviously limited by the decrease in DOC concentration with the CER pretreatment step because DOM was lost fo m the samples prior to potentiometric titration analysis. Therefore, analyses of Cu^^ complexation for microbially decomposed samples are suspect depending on the Cu^^ complexing properties of the lost DOM fraction. For example, analyses are conservative assuming that the lost f-action of DOM is similar to the remaining fraction with respect to organic complexation. Similarly, complexation may be under or overestimated in microbially decomposed samples if the lost faction of DOC binds more or less eSectively than the remaining faction of DOM respectively. 3 .6 .2 3 ^dso/pf/on The effect of adsorptive &actionation on Cu^^ complexation by DOM could not be measured directly because DOC concentrations remaining after the treatment process were below recommended levels for potentiometric titrations (See C huter 2) due to signiGcant scavenging by FeOH3(g). This decrease in DOC concentraGon suggests that the potenGal aSect of adsorpGve GacGonaGon on Cu^^ complexaGon is high, but is ^parently not fully expressed in surface Geshwaters (Luider ef a/, in Press). AdsorpGve GacGonaGon of DOM could therefore signiGcanGy reduce organic Cu^^ complexaGon by effecGvely reducing DOC concentraGon, and parGcularly if the decrease in DOC concentraGon is proporGonal to a decrease in Cu^^ binding density (CuL, pmol C-g'^). 3.7 Conclusion This research supports the conclusion that the bulk Cu^^ complexing properGes of DOM vary systemaGcally in sur6ce Geshwaters as a funcGon of the following two factors: (1) DOM Gom allochthonous sources, and (2) exposure to transfbrmaGon and GacGonaGon processes, as approximated by cumulaGve hydrologie residence time (CHRT). Photochemical decomposiGon, microbial decomposiGon and adsorpGve GacGonaGon processes ^parenGy contribute to this trend either by inGuencing the composiGon of DOM fn g e q/"7 /3 or by contributing to a decrease in DOC concentration in surface 6eshwaters. The latter appears to be the most important variable in predicting organic Cu^^ complexation because the dependency of DOC concentration on CHRT is proportionally greater than Cu^ complexation, as approximated by parameter values of the 2-ligand Langmuir Isotherm. Page q/" f 7& 4. Chapter 4. Summary and Conclusions Dissolved organic matter (DOM) plays an important role in surface ûeshwaters by buffering pH and trace metals, such as Cu^^. DOM is composed of a heterogeneous mixture of organic acids, and a Êaction of these organic acids act as ligands that can complex Cu^^ ions. The objective of this research was to determine if organic acidity and Cu^^ complexation vary systematically in surface 6eshwaters. This research is relevant to the Geld of ecotoxicology because of environmental problems due to acid rain and increasing metal contamination Gom anthropogenic sources, such as mining and smelting industries. Therefore, better understanding of how organic acidity and Cu^^ complexation vary in surface Geshwaters is important Gom a management perspective. Mv overall hvpothesis was that organic aciditv and Cu^^ complexation varv svstematicallv according to DOM source and increasing hvdroloeic residence time CHRT). DOM is loaded Gom terrestrial (allochthonous) and aquatic (autochthonous) sources which each have characteristic properties. Moreover, it has been determined that the concentration and optical properties of DOM can be predicted Gom HRT by q)proximating exposure to transGrmation and Gactionation processes, such as photochemical decomposition, microbial decomposition and adsorption processes. This approach was ^p lied to organic acidity and the Cu^^ complexing properties of DOM. Testing of this overall research hypothesis was divided into two parts. The Grst part of this research addressed the reproducibility of potentiometric titration analyses, as modeled by a two-ligand LangmuG Isotherm. Moreover, I concentrated samples of DOM in a reverse osmosis (RO) concentrator to test the effects of increasing dissolved organic carbon (DOC) concentration on potentiometric analyses of organic Cu^^ complexation. This analysis was Page PJ q f ^ applied to allochthoqous and autochthonous sources of DOM, which are typically thought to represent the upper and lower extreme of organic acidity and C u^ complexation respectively. I determined that analyses of organic acidity are limited by low concentrations of organic acids and by the accuracy of a pH electrode at m DOC concentrations in the m^ority of aquatic systems. Similarly, a low proportion of high afGnity Cu^^ binding sites limits analyses of Cu^^ complexation. In addition, conditional stability parameters of the Langmuir Isotherm model decreased according to increasing titration end-point concentrations of complexed and 6ee labile Cu^^ because of modeling artifacts. This eSect was similar to the efkct of increasing DOC concentration, which is consistent with a combination of two processes; multidentate complexation of Cu^^ and occlusion of Cu^^ binding ligands. Collectively, these results indicate that the precision and accuracy of potentiometric titration analyses is directly dependent on experimental conditions, such as DOC concentration, which could explain 20% to 60% of the reported variation in measures of organic Cu^^ complexation. The second part of this research tested the overall research hypothesis. Organic acidity and the Cu^^ complexing properties of DOM were measured in samples from lakes spanning a gradient in cumulative hydrologie residence time (CHRT) and in allochthonous and autochthonous DOM end members. In addition, subsamples of DOM were treated across micro-reactors to investigate the contribution of photochemical decomposition, microbial decomposition and adsorption processes to systematic variation in DOM. Reference and treated samples were concentrated in a reverse osmosis (RO) concentrator prior to potentiometric titration analyses. Sample concentration was investigated in the Srst part of this thesis, and therefore, I could measure organic Cu^^ complexation by h i ^ afBnity binding sites and compare parameter values of the Langmuir Isotherm among reference and treated samples of DOM. However, DOC concentrations were too low 6)r analyses of organic acidity. Analyses of organic acidity were there&re excluded 6om the results and the overall hypothesis of this research was applied primarily to Cu^^ complexation by ligands of DOM. I determined that there is a 2.5-fbld decrease in the molar concentration of Cu^^ complexing sites as a function of increasing CHRT (p < 0.05), which reflects a decrease in the Cu^^ binding properties of DOM and a decrease in DOC concentration. Treatment of DOM in the micro-reactors suggests that photochemical decomposition, microbial decomposition and adsorption processes could contribute to this trend. Therefore, the overall hypothesis of this research was accepted, which siq)ports the conclusion that organic Cu^^ complexation will vary among surface heshwaters according to the loading of allochthonous and autochthonous DOM and CHRT. The majority of DOM is loaded 6om allochthonous sources, and there is an overall transition in surface heshwaters &om DOM that is characteristic of allochthonous sources to DOM that is characteristic of autochthonous sources with increasing CHRT. fg g e P7 q/" 77^ 5. Literature Cited Ahrens, D. A. 1998. Essentials of Meteorology: An invitation to the atmosphere. Wadsworth Publishing Company. Anderson, P. M. and F. M. Morel. 1978. Copper sensitivity of GamyoMAzx: towareMsw. Limnology and Oceanography. 23: 283-295. Andrew, R. W., K. E. Biesinger, and G. E. Glass. 1978. Effects of inorganic complexation on the toxicity of copper to Afogna. Water Research. 23: 283-295. Ariza, M. E., G. N. Bÿur, and M. V. Williams. 1999. Environmental metal pollutants, reactive oxygen intermediates and genotoxicology: Molecular ^ ro a c h e s to determine mechanisms of toxicity. Kluwer Academic Publishers. Avdeef^ A., J. Zabronsky, and H. H. Stuting. 1983. Calibration of a copper ion selective electrode response to pCu 19. Analytical Chemistry. 55:298-304. Azenha, M., M. T. Vasconcelos, and J. P. S. Cabral. 1995. Organic ligands reduce copper toxicity in PseWomonns aynMgoe. Environmental Toxicology and Chemistry. 14: 369373. Belli, S. L. and A. Ziiino. 1993. Behavior and calibration of the copper (H) ion selective electrode in high chloride media and marine waters. Analytical Chemistry. 65: 25832589. Benedetti, M. F., C. J. Milne, D. G. Kinniburgh, W. H. van Riemsdijk, and L. K. Koopal. 1995. Metal ion binding to humic substances: application of the non-ideal competitive P age q/" adsorption model. Environmental Science and Technology. 29: 446-457. Bertilsson, S. and L. J. Tranvik. 2000. Photochemical transformation of dissolved organic matter in lakes. Limnology and Oceanography. 45: 753-762. Bilby, R, E. and P. A. Bisson. 1992. AUochthonous versus autochthonous organic matter contributions to the trophic support of Esh populations in clear-cut and old growth forested streams. Canadian Journal of Fisheries and Aquatic Sciences. 49: 540-551. Borg, H. 1995. Trace elements in lakes, p. 177-202. Jh B. Salbu and E. Steinnes [eds.]. Trace elements in natural waters. CRC Press Inc. Borgmann, U. and K. M. Ralph. 1984. Copper complexation and toxicity to 6eshwater zooplankton. Archives of Environmental Contamination and Toxicology. 13: 403-409. Brady, J. E., J. W. Russel, and J. R. Holom. 2000. Chemistry: Matter and it's changes. John Wiley and Sons. Breault, R. F., J. A. Colman, G. R. Aiken, and D. McKnight. 1996. Copper spéciation and binding by organic matter in copper contaminated streamwater. Environmental Science and Technology. 30: 3477-3486. Bresnahan, W. T., C. L. Grant, and J. H. Weber. 1978. Stability constants for the complexation of copper (II) ions with water and soil fulvic acids measured by an ion selective electrode. Analytical Chemistry. 50:1675-1679. Brezonik, P. L. 1994. Chemical kinetics and process dynamics in aquatic systems. Lewis Publishers. fo g e P P q /'J J g Bricker, P. O. and B. F. Jones. 1995. Main factors affecting the composition of natural waters , p. 1-20. Jh B. Salbu and E. Steinnes [eds.], Trace elements in natural waters. CRC Press. Brown, V. M., T. L. Shaw, and D. G. Shuiben. 1974. Aspects of water quality and the toxicity of copper to rainbow tront. Water Research. 8: 797-803. Bruland, K. W. 1992. Complexation of cadmium by natural organic ligands in the Central North PaciSc. Limnology and Oceanography. 37:1008-1017. Bryan, S. E., E. Tipping, and J. Hamilton-Taylor. 2002. Comparison of measured and modeled copper binding by natural organic matter in heshwaters. Comparative Biochemistry and Physiology. 133: 37-49. Buffle, J. 1988. Electroanalytical measurement of trace metals complexation, p. 180-232. Jh T.S. West [ed.]. The determination of trace metals in natural waters. Blackwell Scientific Publications. Buffle, J. 1990. Complexation reactions in aquatic systems: an analytical approach. EUis Harwood. Buffle, J., F. Greter, and W. Haerdi. 1977. Measurement of complexation properties ofhumic and fulvic acids in natural waters with lead and copper ion selective electrodes. Analytical Chemistry. 49: 216-222. Buffle, J., R. S. Altmann, M. Filella, and A. Tessier. 1990. Complexation by natural heterogeneous compounds: Site occupation distribution functions, a normalized description of metal complexation. Geochimica et Cosmochimica Acta. 54: 1535-1553. Page 700 q/'77<5 BufQe, J., R. S. Altmann, and M. Filella. 1990. ESect of physico-chemical heterogeneity of natural complexants. Analytica Chimica Acta. 232: 225-237. BufOe, J. and R. R. De Vitre. 1994. Chemical and biological regulation of aquatic systems. Lewis Publishers. Bushaw, K. L., R. G. Zepp, M. A. Tarr, D. Schulz-Jander, R. A. Bouibonniere, R. E. Hodson, W. L. Miller, D. A. Bronk, and M. A. Moran. 1996. Photochemical release of biologically available nitrogen 6om aquatic dissovled organic matter. Nature. 381:404407. Cabaniss, S. E. and M. S. Shuman. 1988. Copper binding by dissolved organic matter. I. Suwanee river fulvic acid equihbria. Geochimica et Cosmochimica Acta. 52: 185-193. Carter, M. R. 1996. Analysis of soil organic matter storage in agroecosystems, p. 3-14.7» M.R. Carter and B.A. Stewart [eds.]. Structure and organic matter storage in agricultural soils. Stewart Lewis Publishers. Christensen, J. B., J. J. Botman, and T. H. Christensen. 1999. Complexation of Cu and Pb by DOC in polluted groundwater: A comparison of experimental data and predictions by computer spéciation models (Wham and MinteQZA). Water Research. 33: 3231-3238. Clair, T. A., J. R. Kramer, M. Sydor, and D. Eaton. 1991. Concentration of aquatic dissolved organic matter by reverse osmosis. Water Research. 25: 1033-1037. Clair, T. A. and B. G. Soyer. 1997. Environmental variability in the reactivity of hreshwater dissolved organic carbon to UV-B. Biogeochemistry. 36: 89-97. Page 707 q/"770 Coale, K. H. and K. W. Bruland. 1988. Copper complexation in the Northeast PaciGc. Limnology and Oceanography. 33: 1084-1101. Cohen, J. E. 1995. Population growth and the Earth's carrying edacity. Science. 269: 341346. Cotton, F. A. and G. Wilkinson. 1976. Base inorganic chemistry. John Wiley and Sons. Coveney, M. F. and R. G. Wetzel. 1995. Biomass, production, and speciGc growth rate of bacterioplankton and coupling to phytoplankton in an oligotrophic lake. Limnology and Oceanography. 40:1187-1200. Cummins, K. W., J. R. Sedell, F. J. Swanson, G. W. Minshall, S. G. Fisher, C. E. Cushing, R. C. Peterson, and R. L. Vannote. 1982. Organic matter budgets for stream ecosystems: problems in their evaluation, p. 299-353. Jh G.W. Minshall and J.R. Barnes [eds.]. Stream ecology: application and testing of general ecological theory. Plenum Press. Curtis, P. J. 1998. Climatic and hydrologie control of DOM concentration and quality in lakes, p. 93-104. Jh D.O. Hessen and L.J. Tranvik [eds.]. Aquatic humic substances: ecology and biogeochemistry. Springer. Curtis, P. J. and H. E. Adams. 1995. Dissolved organic matter quantity and quahty 6om heshwater and saline lakes in east-central Alberta. Biogeochemistry. 30: 59-76. Curtis, P. J. and D. W. Schindler. 1997. Hydrologie control of dissolved organic matter concentration and quality in low-order Pre-Cambrian Shield Lakes. Biogeochemistry. 36: 125-138. P ag e 702 q/" 77<$ Cusimano, R. F., D. F. Brakke, and G. A. Ch^man. 1986. Effect of pH on the toxicities of cadmium, copper, and zinc to steelhead trout fiSa/mo gaWnenp. Canadian Journal of Fisheries and Aquatic Sciences. 43: 1497-1503. De Wit, J. C. M., W. H. van Riemsdijk, and L. K. Koopal. 1993a. Proton binding to humic substances. Environmental Science and Technology. 27: 2005-2014. De Wit, J. C. M., W. H. van Riemsdijk, and L. K. Koopal. 1993b. Proton Binding to Humic substances. 2. Chemical heterogeneity and adsorption models. Environmental Science and Technology. 27:2015-2022. Driscoll, C. T. and M. D. Lehtinen. 1994. Modelling the acid-base chemistry of organic solutes in the Adirondack, New York, lakes. Water Resources Research. 30: 297-306. Engstrom, D. R. 1987. Influence of vegetation and hydrology on the humus budgets of Labrador lakes. Canadian Journal of Fisheries and Aquatic Sciences. 44:1306-1314. Ephraim, J. H., H. Boren, C. Pettersson, I. Arsenie, and B. Allard. 1989. A novel description of the acid-base properties of an aquatic fulvic acid. Environmental Science and Technology. 23: 356-362. Erickson, R, J., D. A. Benoit, V. R. Mattson, H. P. Nelson, and E. N. Leonard. 1996. The efiects of water chemistry on the toxicity of copper to fathead minnows. Environmental Science and Technology. 15: 181-193. Ertel, J. R., J. I. Hedges, and E. M. Perdue. 1984. Lignin signiture ofhumic substances. Science. 223: 485-487. Filip, Z. 1985. The occurrence of copper, iron, zinc and other elements and the nature of some copper and iron complexes in humic substances 6om municipal refuse disposed in a land SU. The Science of the Total Environment. 44:1-16. Frimmel, F. H. 1990. Characterization of organic acids in Seshwater: A current status and limitations, p. 5 - 2 5 . E.M. Perdue and E.T. Gjessing [eds.]. Organic acids in aquatic ecosystems. John Wiley & Sons. Fukushima, T., J. C. Park, A. hnai, and K. Matsushige. 1996. Dissolved organic carbon in a eutrophic lake: Dynamics, biodegradability and origin. Aquatic Sciences. 58:139-157. Gamble, D. S. 1980. Copper(n) titration of fulvic acid ligand sites with theoretical potentiometric, and spectrophotometric analysis. Analytical Chemistry. 52:1901-1908. Geller, A. 1986. Comparison of mechanisms enhancing biodegradability of reSactory lake water constituents. Linmology and Oceanography. 31: 755-764. Gu, B., J. Schmitt, Z. Chen, L. Liang, and J. F. McCarthy. 1994. Adsorption and desorption of natural organic matter on iron oxide: Mechanisms and models. Environmental Science and Technology. 28: 38-46. Gu, B., J. Schmitt, Z. Chen, L. Liang, and J. F. McCarthy. 1995. Adsorption and desorption of different organic matter factions on iron oxide. Geochimica et Cosmochimica Acta. 59: 219-229. Hales, L. T., C. A. Simon, and E. B. Graeme. 1999. Use of copper ion selective electrodes in ecotoxicological studies: Appropriate validation and quality control. Setae News. July 21- 22. Hering, J. G. and F. M. M. Morel. 1988a. Humic acid conq)lexation of Ca and Cu. Environmental Science and Technology. 22: 1234-1237. Hering, J. G. and F. M. M. Morel. 1988b. Kinetics of trace metal complexation: role of alkaline-earth metals. Environmental Science and Technology. 22: 1469-1478. Hessen, D. O., E. Gjessing, J. Knulst, and E. Fjeld. 1997. TOC fluctuations in a humic lake as related to catchment acidiEcation, season and climate. Biogeochemistry. 36:139-151. Hirose, K. 1994. Conditional stability constants of metal complexes of organic ligands in sea water: past and present, and simple coordination chemistry model. Analytica Chimica Acta. 284: 621-634. Holm, T. R. and C. D. Curtiss m . 1990. Copper complexation by natural organic matter in ground water, p. 5 0 8 - 5 1 8 . D C. Melchior and R.L. Bassett [eds.]. Chemical modelling of aqueous systems H. American Chemical Society. hnai, A., T. Fukushima, K. Matsushige, and Y. H. Kim. 2001. Fractionation and characterization of dissolved organic matter in a shallow eutrophic lake, its inflowing rivers, and other organic matter sources. Water Research. 35: 4019-4028. Jonsson, A., M. Meili, A.-K. BergstrOm, and M. Jansson. 2001. Whole lake mineralization of allochthonous and autochthonous organic carbon in a large humic lake (Ortrâsket, N. Sweden). Limnology and Oceanography. 46: 1691-1700. K ^lan, L. A. and J. D. Newbold. 1995. Measurement of streamwater biodegradable fa g g 705 q/" 77& dissolved organic carbon with a plng-flow bioreactor. Water Research. 29: 2696-2706. Kerr, J. B. and C. T. McElioy. 1993. Evidence for large upward trends of ultraviolet-B radiation linked to ozone depletion. Science. 262: 1032-1034. Kettle, S. F. A. 1969. Coordination compounds. Appleton-Century-Crofts. Kieber, R. J., X. ^ o u , and K, Mopper. 1990. Formation of carbonyl compounds ûom UVinduced photodegradation ofhumic substances in natural waters: Fate of riverine carbon in the sea. Limnology and Oceanography. 35:1503-1515. kieber, R. J., A. Li, and P. J. Seaton. 1999. Production of nitrite 6om the photodegradation of dissolved organic matter in natural waters. Environmental Science and Technology. 30: 993-998. Kinniberg, D. G., C. J. Milne, M. P. Benedetti, J. P. Pinheiro, J. Filius, L. K. Koopal, and W. H. van Riemsdijk. 1996. Metal binding by humic acid: Application of the NICA-Donnan model. Environmental Science and Technology. 30:1687-1698. Kinniburgh, D. G., W. H. van Riemsdijk, L. K. Koopal, M. Boikovec, M. F. Benedetti, and M. J. Avena. 1999. Ion binding to natural organic matter: competition, heterogeneity, stoichiometry and thermodynamic consistency. Colloids and Surfaces A: Physiochemical and Engineering Aspects. 151: 147-166. Kotz, J. C., J. L. Wood, M. D. Joesten, and J. W. Moore. 1994. The chemical world: concepts and applications. Saunders College Publishing. Lambert, S. D. and N. J. D. Graham. 1995. Removal of non-specifc dissolved organic matter P age 706 q/"77& 6 om upland potable water supplies I: Adsorption. Water Research. 29: 2421-2426. Lang&rd, C. H., D. S. Gamble, A. W. Underdown, and S. Lee. 1983. Interaction of metal ions with a well characterized fulvic acid, p. 2 1 9 - 2 3 7 . R.F. Christman and E.T. Gijessing [eds.]. Aquatic and terrestrial humic materials. Ann Arbor Science. Lauren, D. J. and D. G. McDonald. 1986. Influence of water hardness, pH and alkalinity on the mechanisms of copper toxicity in juvenile rainbow trout, gaWwer. Canadian Journal of Fisheries and Aquatic Sciences. 43:1488-1496. Leal, M. F. C., M. T. S. D. Vasconcelos, and C. M. G. van den Berg. 1999. Copper-induced release of complexing ligands similar to thiols byjEmi/ianm Awx/eyz in seawater cultures. Limnology and Oceanography. 44: 1750-1762. Leff^ L. G. and J. L. Meyer. 1991. Biological availabihty of dissolved organic carbon along the Ogeechee River. Limnology and Oceanography. 36: 315-323. Leuenberger, B. and P. W. Schindler. 1986. Explication of integral pK spectrometry to the titration curve of fulvic acid. Analytical Chemistry. 58: 1471-1474. Lind, V. and D. Hongue. 1994. Ultraviolet irradiated water containing humic substances inhibits bacterial metabolism. Water Research. 28:1111-1116. Lindell, M. J., W. Graneli, and L. J. Tranvik. 1995. Enhanced bacterial growth in response to photochemical transformation of dissolved organic matter. Limnology and Oceanography. 40:195-199. Lindell, M. J., W. H. Granih, and S. Bertilsson. 2000. Seasonal photoreactivity of dissolved organic matter 6om lakes with contrasting humic content. Canadian Journal of Fisheries and Aquatic Sciences. 57: 875-885. Luider, C. D., E. Petticrew, and J. Curtis, in Press. Scavenging of dissolved organic matter (DOM) by amorphous iron hydroxide particles Fe(0H)3(s). Hydrobiologia. Lydersen, E. 1998. Humus and AcidiEcation, p. 63-92. D O. Hessen and L.J. Tranvik [eds.], Aquatic humic substances: ecology and biogeochemistry. Springer. MacRae, R. K., D. E. Smith, N. Swoboda-Colberg, J. S. Meyer, and H. L. Bergman. 1999a. Copper binding afBnity of rainbow trout (OncorAyncAuf and brook trout (SaJve/mufyôMtfWü) gills: Implications &r assessing bioavailable metal. Environmental Toxicology and Chemistry. 18: 1180-1189. MacRae, R. K., A. S. Maest, and J. S. Meyer. 1999b. Selection of an organic acid analogue of dissolved organic matter for use in toxicity testing. Canadian Journal of Fisheries and Aquatic Sciences. 56: 1484-1493. Manahan, S. A. 1994. Environmental Chemistry, 6th ed. Lewis Publishers. Mantoura, R. F. C., A. Dickson, and J. P. Riley. 1978. The complexation of metals with humic materials in natural water. Estuarine Coastal Marine Science. 6: 387-408. Manuza, B., S. Deiana, V. Maddua, C. Gessa, and R. Seeber. 1995. Stability constants of metal-humate complexes: titration data analyzed by bimodal Gausian distribution. Soil Science Society of America. Journal. 59: 1570-1574. Marr, J. C. A., J. Lipton, D. Cacela, J. A. Hansen, J. S. Meyer, and H. L. Bergman. 1999. P ag e 70& q /'JJg Bioavailability and acute toxicity of copper to rainbow trout (OncorAyncAMf myAiMrl in the presence of organic acids simulating maternal dissolved organic carbon. Canadian Journal of Fisheries and Aquatic Sciences. 58: 1471-1483. Martell, A. E. and R. D. Hancock. 1994. The chelate, macrocyclic and cryptate effects. Jn G.B. Kauf&nan [ed.]. Coordination chemistry: a century of progress. American Chemical Society. McKnight, D. M. and F. M. M. Morel. 1979. Release of weak and strong copper-complexing agents by algae. Limnology and Oceanogr^hy. 24: 823-837. M cKni^t, D. M., G. L. Feder, E. M. Thurman, R. L. Wershaw, and J. C. Westall. 1983. Complexation of Cu by aquatic humic substances hom different environments. Environmental Science and Technology. 28. McKnight, D. M., K. E. Bencala, G. W. Zellwager, G. R. Aiken, G. L. Feder, and K. A. Thom. 1992. Sorption of dissolved organic carbon by hydrous aluminum and iron oxides occurring at the confluence of Deer Creek with the Snake River, Summit County, Colorado. Environmental Science and Technology. 26:1388-1396. McKnight, D. M., E. D. Andrews, S. A. Spaulding, and G. R. Aiken. 1994. Aquatic fulvic acids in algal-rich Antarctic ponds. Limnology and Oceanogr^hy. 39: 1972-1979. McKnight, D. M. and R. Aiken. 1998. Sources and age of aquatic humis, p. 9-37. Jh D.O. Hessen and L.J. Tranvik [eds.]. Aquatic humic substances: ecology and biogeochemistry. Springer. P age JOP q/"JJ& Meador, J. P. 1991. The interaction of pH, dissolved organic carbon, and total copper in the determination of ionic copper and toxicity. Aquatic Toxicology. 19: 13-32. Meador, J. P., T. H. Sibley, G. L. Swartzman, and F. B. Taub. 1998. Copper tolerance by hreshwater algal species and its abihty to alter &ee ion copper. Aquatic Toxicology. 44: 69-82. Meih, M. 1992. Sources, concentration and characteristics of organic matter in soAwater lakes and streams of the Swedish forest region. Hydrobiologia. 229: 23-41. Midgely, D. and K. Torrance. 1991. Potentiometric water analysis, 2nd ed. John Wiley and Sons. Miller, W. L. and M. A. Moran. 1997. Interaction of photochemical and microbial processes in the degradation of dissolved organic matter hom a coastal marine environment. Limnology and Oceanogr^hy. 42:1317-1324. Milne, C. J., D. G. Kmniburg, J. C. M. De Wit, W. H. Riemsdqk, and L. K. Koopal. 1995. Analyses of proton binding by a peat humic acid using a simple electrostatic model. Geochimica et Cosmochimica Acta. 59: 1101-1112. Mopper, K., X. Zhou, R. J. Kieber, D. J. Kieber, R. J. Sikorski, and R. D. Jones. 1991. Photochemical degradation of dissolved organic carbon and its impact on the oceanic carbon cycle. Nature. 353: 60-62. Moran, M. A. and R. E. Hodson. 1990. Bacterial production ofhumic and non-humic components of dissolved organic carbon. Limnology and Oceanogr^hy. 35: 1744-1756. Page 770 Moran, M. A. and R. G. Zepp. 1997. Role of photoreactions in the formation of biologically labile compounds 6om dissolved organic matter. Limnology and Oceanography. 42: 1307-1316. Moran, M. A., W. M. Sheldon, and R. G. Zepp. 2000. Carbon loss and optical property changes during long-term photochemical and biological degradation of estuarine dissolved organic matter. Limnology and Oceanogr^hy. 45: 1254-1264. Morel, F. M. M. 1983. Principles in aquatic chemistry. John Wiley and Sons. Morel, F. M. M. and J. G. Hering. 1993. Principles and application of aquatic chemistry. John Wiley and Sons. Morris, D. P., H. Zagarese, G. E. Wilhamson, E. G. Balseiro, B. R. Hargreaves, B. Modenutti, R. Moeller, and C. Queimalinos. 1995. The attenuation of solar UV radiation in lakes and the role of dissolved organic carbon. Linmology and Oceanography. 40: 1381-1391. Myers, N. 1993. Population, environment, and development. Environmental Conservation. 20: 205-216. Naeem, S. and S. Li. 1997. Biodiversity enhances reliability. Nature. 390: 507-509. Nriagru, J. 0 . 1996. A history of global metal pollution. Science. 272: 223-224. Nriagru, J. O. and J. M. Pacyna. 1988. Quantitative assesment of worldwide contamination of air, water and soils by trace metals. Nature. 333:134-139. O'Shea, T. A. and K. H. Mancy. 1978. The eSect of pH and hardness metal ions on the Page 777 q/"77g competitive interaction between trace metal ions and inorganic and organic complexing agents found in natural waters. Water Research. 12: 703-711. Pagenkopf^ G. K,, R C. Russo, and R V. Thurston. 1974. Effect of complexation on toxicity of copper to hshes. Journal. Fisheries Research Board of Canada. 31: 462-465. Park, J., M. Aizaki, T. Fukushima, and A. Otsuki. 1997. Production of labile and reAactory DOC by zooplankton excretion: An experimental study using large continuous flowthrough ponds. Canadian Journal of Fisheries and Aquatic Sciences. 54: 434-443. Perdue, E. M. and C. R. Lytle. 1983a. A critical examination of metal-ligand complexation models: application to dehned multiligand mixtures, p. 295-314. Jh R F. Christman and E.T. Gjessing [eds.]. Aquatic and terrestrial humic materials. Ann Arbor Science. Perdue, E. M. and C. R. Lytle. 1983b. Distribution model for the binding of protons and metal ions by humic substances. Environmental Science and Technology. 17: 654-660. Perdue, E. M. 1998. Chemical composition, structure and metal binding properties, p. 41-61. 7» D.O. Hessen and L.J. Tranvik [eds.]. Aquatic humic substances: ecology and biogeochemistry. Springer. Playle, R C. and D. G. Dixon. 1993. Copper and cadmium binding to 6sh gills: estimates of metal-gill stabihty constants and modeling of metal accumulation. Canadian Journal of Fisheries and Aquatic Sciences. 50: 2678. Radel, S. R and M. H. Navidi. 1994. Chemistry, 2nd ed. West Publishing Company. Richards, J. G. and R. C. Playle. 1998. Cobalt binding to gills of rainbow trout (DncorAynwj Page 772 q/" 77^ An equilibrinm model. Comparative Biochemistry and Physiology. 119C: 185197. Richards, J. G., P. J. Curtis, B. K. Bumison, and R. C. Playle. 2001. Eflects of natural organic matter sources on reducing metal toxicity to rainbow trout and on metal binding to their gills. Environmental Toxicology and Chemistry. 20: 11591166. Roberts, R. D., M. S. Evans, and M. T. Arts. 1992. Light, nutrients, and water temperature as determinants of phytoplankton production in two saline, prairie lakes with high sulphate concentrations. Canadian Journal of Fisheries and Aquatic Sciences. 49: 2281-2290. Rozan, T. F. and G. Benoit. 1999. Geochemical factors controlling free Cu ion concentrations in river water. Geochimica et Cosmochimica Acta. 63: 2316-2323. Runnels, D. D., T. A. Shq)herd, and E. E. Angino. 1992. Metals in water: determining natural background concentrations in mineralized areas. Environmental Science and Technology. 26: 2316-2323. Ryan, D. K. and J. H. Weber. 1982. Fluorescence quenching titration for determination of complexing capacities and stabihty constants of fulvic acids. Analytical Chemistry. 54: 986-990. Saar, R. A. and J. H. Weber. 1980. Comparison of spectrofluorometric and ion-slective electrode potentiometry for determination of complexes between fulvic acid and heavy metal ions. Analytical Chemistry. 52: 2095-2100. fo g e J J3 q/"JJ& Savage, J. M. 1995. Systemaücs and the biodiversity crisis. BioScience. 45: 673-679. Schindler, D. W., P. J. Curtis, S. E. Bayley, K. G. Beaty, S. E. Parker, and M. P. Stainton. 1997. Climate induced changes in the dissolved organic carbon budgets of boreal lakes. Biogeochemistry. 36: 9-28. Schlautman, M. A. and J. J. Morgan. 1994. Adsorption of aquatic humic substances on coUoidal-size aluminum particles: influence of solution chemistry. Geochimica et Cosmochimica Acta. 58:4293-4303. Serkiz, S. M. and E. M. Perdue. 1990. Isolation of dissolved organic matter 6om Suwanee River using reverse osmosis. Water Research. 24: 911-916. Singer, J. M. and D. N. Munns. 1987. Soils: an introduction, 2nd ed. Macmillan Publishing. Sjdberg, S. and L. Ldvgren. 1993. The application of potentiometric techniques to study complexation reactions at the mineral/water interface. Aquatic Sciences. 55: 324-335. Skoog, D. A. and J. J. Leary. 1992. Principles of instrumental analysis, 4th ed. Harcourt Brace Publishers. Sommaruga, R., R. Psenner, E. Schaflerer, K. A. Koinig, and S. Sommaruga-Wograth. 1999. Dissolved organic carbon concentration and phytoplankton biomass in high mountain lakes of the Austrian Alps: potential efkct of climate warming on UV under water attenuation. Arctic, Antarctic and Alpine Research. 31:247-253. Sposito, G. 1981. Trace metals in contaminated waters. Environmental Science and Technology. 40: 396-403. P ag e 774 q/" 77^ Strome, D. J. and M. C. Miller. 1978. Photolytic changes in dissolved hmnic substances. Verhandlungen Intemationaler Verein Fuer Limnologie. 20: 1248-1254. Stumm, W. and J. J. Morgen. 1995. Aquatic chemistry: chemical equilibria and rates in natural waters, 3rd ed. John Wiley and Sons. Sun, L., E. M. Perdue, J. L. Meyer, and J. L. Weis. 1997. Use of elemental composition to predict bio-availability of dissolved organic matter in a Georgia river. Limnology and Oceanogr^hy. 42: 714-721. Sunda, W. and R. R. L. Guillard. 1976. The relationship between cupiic ion activity and the toxicity of copper to phytoplankton. Journal of Marine Research. 34: 511-529. Sunda, W. G. and J. M. Lewis. 1978. Effect of complexation by natural organic ligands on the toxicity of copper to a unicellular alga, AJo/zocA/yg» /«rAen. Limnology and Oceanography. 23: 870-876. Tipping, E. 1993. Modeling the competition between alkaline earth cations and trace metal species for binding by humic substances. Environmental Science and Technology. 27: 520-529. Tipping, E., M. M. Reddy, and M. A. Hurley. 1990. Modeling electrostatic and heterogeneity eSects on proton dissociation 6om humic substances. Environmental Science and Technology. 24: 1700-1705. Tipping, E. H. and M. A. Hurley. 1992. A u n if ^ g model of cation binding by humic substances. Geochimica et Cosmochimica Acta. 56: 3627-3641. Pugg J J J Town, R. M. and M. Filella. 2000. Dispelling the myths: Is the existence of LI and L2 ligands necessary to explain ion spéciation in natural waters. Limnology and Oceanography. 45: 1341-1357. Urban, N. R., E. Gorham, J. K. Underwood, F. B. Martin, and J. G. Odgen. 1990. Geochemical processes controlling concentrations of Al, Fe and Mn in Nova Scotia lakes. Limnology and Oceanogr^hy. 35: 1516-1534. Vasconcelos, M. T. S. D., M. A. O. Azenha, and J. P. S. Cabral. 1997. Comparison of availability of copper (II) complexes with organic ligands to bacterial cells and to chiton. Environmental Science and Technology. 16:2029-2039. Volk, C. J., C. B. Volk, and L. A. Kaplan. 1997. Chemical composition of biodegradable dissolved organic matter in streamwater. Linmology and Oceanography. 42: 39-44. Waiser, J. M. and D. R. Robarts. 2000. Changes in composition and reactivity of allochthonous DOM in a prairie saline lake. Limnology and Oceanography. 45: 763-774. Welsh, P. G., J. F. Skidmore, D. J. Spry, D. G. Dixon, P. V. Hodson, N. J. Hutchinson, and B. E. Hickie. 1993. Effect of pH and dissolved organic carbon on the toxicity of copper to larval fathead minnow /^rome/os) in natwal lake waters of low alkalinity. Canadian Journal of Fisheries and Aquatic Sciences. 50: 1356-1362. Welsh, P. G., J. L. Parrot, D. G. Dixon, P. V. Hodson, D. J. Spry, and G. Mierle. 1996. Estimating acute Cu toxicity to larval fathead minnow in soft water &om measurements of dissolved organic carbon, calcium and pH. Canadian Journal of Fisheries and Aquatic Sciences. 53: 1263-1271. Page J J 6 q/" JJ& Wetzel, R. G. 1983. Ümnology, 2nd ed. Saunders College Publishing. Wetzel, R. G., P. G. Hatcher, and T. S. Blanchi. 1995. Natural photolysis by ultraviolet irradiance of recalcitrant dissolved organic matter to simple substrates for rapid bacterial metabolism. Limnology and Oceanography. 40:1369-1380. Wilhelm, S. W. and C. G. Trick. 1994. Iron limited growth of cyanobacteria: Multiple siderophore production is a common response. Limnology and Oceanography. 39: 19791984. Winch, S., J. Ridal, and D. Lean. 2002. Increased metal bioavailability following alteration of ûeshwater dissolved organic carbon by ultraviolet B radiation exposure. Environmetal Toxicology. 17: 267-274. Winner, R. W. 1985. Bioaccumulation and toxicity of copper as affected by interactions between humic acid and water hardness. Water Research. 19: 449-455. Wood, A. M. 1983. Available copper ligands and the apparent bio-availability of copper to natural phytoplankton assemblages. The Science of the Total Environment. 28: 51-64. WuUsberg, G. 1987. Principles of descriptive inorganic chemistry. Brooks/Cole Company. Xue, H. and L. Sigg. 1993. Free cupric ion concentration and Cu(II) spéciation in a eutrophic lake. Limnology and Oceanogr^hy. 38: 1200-1213. Xue, H. and W. G. Sunda. 1997. Comparison of [Cu+2] measurements in lake water determined by ligand exchange and cathode stripping voltometry and by ion selective electrode. Environmental Science and Technology. 31:1902-1909. P ag e 777 Thang, Y. and E. E. Prepas. 1996. Regulation of the dominance of planktonic diatoms and cyanobacteria in four eutrophic hardwater lakes by nutrient, water column stability, and temperature. Canadian Journal of Fisheries and Aquatic Sciences. 53: 621-633. Zumstein, J. and J. Bufde. 1989. Ciiculadbn ofpedogenic and aquagenic organic matter in an eutrophic lake. Water Research. 23: 229-239. Page