|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 ABIOTIC NATURAL DISSOLVED ORGANIC MATTER-CARBONATE INTERACTION IN KARSTIC AQUIFERS By JIN JIN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2009
2 2009 Jin Jin
3 To my family
4 ACKNOWLEDGEMENTS I would like to sincerely thank my advisor, Dr. Andrew Zimmerman, for his continuous advice and guidance throughout my graduate study in the Department of Geological Sciences. His patience and encouragement helped me overcome lots of difficulties and eventually accomplish this research project. I am also very grateful to my committee members, Dr. Jonathan Martin and Dr. Jean-Claude Bonzongo, for all th e help and support they provided during my experiments and thesis development. I thank all my friends in the Department of Geological Sciences for their friendship and laughter. Special thanks go to my lab mates, Atanu Mukherjee, Dr. Mi-Youn Ahn, Ann Laffey and Dr. Gabriel Kasozi, for their assistance in my research as well as the good time we shared together. Lastly, I would like to thank my beloved parents, who raise my up to more than I can be.
5 TABLE OF CONTENTS Page ACKNOWLEDGEMENTS .............................................................................................................4LIST OF TABLES ................................................................................................................ ...........7LIST OF FIGURES ............................................................................................................... ..........8ABSTRACT ...................................................................................................................... ...............9 CHAPTER 1INTRODUCTION .................................................................................................................. 112BACKGROUND .................................................................................................................... 14NDOM in Groundwater ........ ................ ................ ................ ................ ................ ...........14NDOM-Carbonate Interactions .......................................................................................173MATERIALS AND METHODS ...........................................................................................244RESULTS ....................................................................................................................... ........30Quantity of NDOM-Carbonate Adso rption ............ ................ ................ ................ .........30Composition of NDOM-Carbonate Adsorption ..............................................................325DISCUSSION .................................................................................................................... .....45Factors Controlling Karst Aquifer NDOM Adsorpti on .................. .............. ............. ......45NDOM Transformation due to Abiotic Carbonate Interaction .......................................52NDOM Effect on Carbonate Dissolution ........................................................................556AQUIFER ENVIRONMENTAL IMPLICATION AND CONCLUSION ............................59 APPENDIX ADETAILS OF ANALYTICAL METHODS USED ...............................................................62BFLORIDAN AQUIFER GEOLOGIC FRAMEWORK .........................................................64CALTERNATIVE ADSORPTION MODEL PARAMETERS ................................................65DSORBENT X-RAY DIFFRACTION RESULTS ...................................................................66 EHPLC PEAK RETENTION TIME OF SORBENT INDIGENOUS ORGANIC MATTER ........................................................................................................................ ........68
6 FEEM GRAPHS FOR NDOM-SORBENT PAIRS .................................................................69LIST OF REFERENCES ............................................................................................................ ...80BIOGRAPHICAL SKETCH .........................................................................................................91
7 LIST OF TABLES Table Page 3-1Sorbent background information ...... ................ ................ ................ ................ ............. .........293-2Sorbent chemical composition ................................................................................................294-1Linear model parameters (weight-normalized) ......................................................................364-2Linear model parameters (surface area-normalized) ..............................................................364-3Amount NDOM adsorbed at two NDOM solution conc entrations ........................................37A-1Calibration of the HPLC-SEC column ...................................................................................63C-1Langmuir model parameters (weight-normalized) .................................................................65C-2Freundlich model parameters (weight-normalized) ................................................................65E-1HPLC peak retention time of sorbent indigenous or ganic matter ............... ................ ...........68F-1EEM graphs for NDOM-sorbent pairs ...................................................................................70
8 LIST OF FIGURES Figure Page 4-1Adsorption data for soil NDOM on R-O and R-S carbonate rocks of two grain size fractions (<0.15 and 0.15-0.5 mm) and for 2 and 4 day interaction periods ..........................384-2Adsorption data for stream NDOM on six carbonate sorbents (C-S, C-H, C-A, C-O, RO and R-S) .................................................................................................................... ..........394-3Adsorption and desorption data for soil NDOM on R-S of two grain sizes (<0.15 and 0.15-0.5 mm), 4-days .......................................................................................................... ....404-4Representative EEM and change in EEM during interaction .................................................414-5Excitation-emission matrices of indigenous organic matter released from aquifer core materials (C-S, C-H, C-A and C-O, all <0.15 mm) in to water ...............................................424-6Liquid size exclusion chromatograph (HPLC-SEC) of soil NDOM (14.47 mg C/L) sorbate before and after interaction with four aqui fer core materials .....................................434-7Ca2+ and Mg2+ released during 2-day in teraction of distilled water and three concentrations of stream NDOM with aquife r core materials (C-S, C-H, C-A and C-O, all 0.15-0.5 mm) .............................................................................................................. .......445-1Plot of sorbent surface area (m2/g sorbent) versus adsorption affinity (L/g sorbent) for six Floridan Aquifer materials ................................................................................................ 575-2Plot of sorbent indigenous organic matter content (g C/g sorbent) versus adsorption affinity (L/g sorbent) for six Floridan Aquifer materials ........................................................58A-1HPLC-SEC column calibration function ................................................................................63B-1Floridan Aquifer geologic framework ....................................................................................64D-1Sorbent X-ray diffraction results ............................................................................................6 6
9 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science ABIOTIC NATURAL DISSOLVED ORGANIC MATTER-CARBONATE INTERACTION IN KARSTIC AQUIFERS By Jin Jin May 2009 Chair: Andrew R. Zimmerman Major: Geology Natural dissolved organic matter (NDOM)-m ineral interaction (e.g. adsorption, desorption, mineral precipitation/dissolution) may be a significant factor controlling geochemical and microbial processes, and contaminant fate in karst groundwater systems. Despite its importance, the abiotic interactions of NDOM with subsurface carbonate remains poorly understood. Our study examines the abiotic NDOMcarbonate interactions that take place in a karst aquifer through laboratory adsorption-desorption experiments using NDOM from soils and stream water collected in north Florida and six representative carbonate sorbents from each of the four major formations making up the Floridan aquifer. While significant adsorption was observed at higher NDOM concentrations, in the lower natural range of NDOM concentrations, most carbonate materials released indigenous organic matter (OM). No difference was found in the amount of organic carbon adsorbed from the two different NDOM types onto carbonate sorbents, however, the core materials generally sorbed more NDOM than aquifer rocks. Longer interaction periods for NDOM-carbonate led to more NDOM adsorption indicating that adsorption equi librium was not achieved. Sorbent surface area was found to be a controlling factor of adsorption for relatively pure carbonate samples, but not
10 when sorbent indigenous OM and other impurities were present. The NDOM-carbonate adsorption system was well described by a modified linear isotherm, and was mostly reversible, though a small amount of hysteresis was observed. Preferential adsorption of a high over low molecular weight NDOM, and a humic-like over fulvic-like fraction onto carbonate sorbents were detected. The presence of NDOM inhibited mineral dissolution, but the dissolution inhibition observed was not always proportional to NDOM solution concentration. Surface area, and more importantly, the mineralogy of carbonate sorbents were found to be the controlling factors in determining susceptibility to dissolution and degree of NDOM dissolution inhibition. Though the NDOM-carbonate adsorption mechanism could not be completely determined by these experiments because of the heterogeneity and complexity of NDOM and adsorbent surfaces, it is clear that both rapid and weak outer-sphere bonding and stronger but slower hydrophobic interaction are at play. The results pr esented here provide a starting point for further studies examining subsurface NDOM-carbonate inter action, as well as some insight into the possible environmental consequences of human introduction of NDOM-rich solutions into aquifers that may occur during aquifer recharge or storage and recovery projects.
11 CHAPTER 1 INTRODUCTION While groundwater aquifers hold 97% of the Earths unfrozen fresh water (Hancock et al. 2005), our understanding of the biogeochemical processes occurring there remains very limited. Although it has long been known that water and its solutes, microbes and rock interact in these subsurface systems, the importance and rates of the various biotic and abiotic geochemical reactions that occur there remain poorly defined. One example, the interaction of natural dissolved organic matter (NDOM) with solid surfaces in aquifers, has not received the scientific attention that it warrants, as it likely plays a major role in controlling groundwater quality. Karstic aquifers, provide 40% of the groundwater used for drinking and more than 25% of the worlds human population lives above karst. With their relatively high surface connectivity and groundwater flow rates, these aquifers are particularly vulnerable to contamination from above (Lau & Mink, 1987). While NDOM, itself, is considered a contaminant in drinking water (Bob & Walker, 2001), it may contain organic contaminant compounds and also plays a role in aquifer contaminant fate and transport as a carrier of contaminants in the dissolved or colloidal state (Frimmel, 1998). Many contaminants that are deemed as virtually immobile in aqueous systems can interact with NDOM or colloidal NDOM, resulting in migration of hydrophobic chemicals far beyond predicted distances. In addition, groundwater NDOM plays important roles in controlling biogeochem ical processes by acting as proton donors and acceptors and as pH buffers, and by participating in mineral precipitation and dissolution, and by affecting the transport and degradation of pollutants (Findlay et al. 2003; Frimmel, 1998; Gu et al. 1994; Schlautman & Morgan, 1994). For example, NDOM is a food source, fueling microbial activity and thus, contaminant degradation, and consumes oxygen, thus controls the oxidation state of the subsurface (Lovley & Chapelle, 1995). Groundwater NDOM is
12 likely to encourage the growth of potentially harmful microbial populations (Boyes & Elliott, 2006; Fisher et al. 2006). The presence of NDOM considerably influences the mobilization and fixation of heavy metals (e.g., Zn, Cu, Pb, Cd, Hg, Cr, Co, U, Pu) (Lee et al. 2005; Petrovic et al. 1999). Due to its reactive nature, a thorough understanding of the processes which govern groundwater NDOM quantity and quality deserves investigation. Interactions between NDOM and carbonate minerals such as calcite and dolomite, the major constituent of a karstic aquifer, can affect the subsurface environment. Sorption of NDOM onto aquifer minerals changes the physicochemical properties of the solid to a large extent, including the complexation, dispersity and surface electrical property (Davis, 1982; Dunnivant et al. 1992; Gu et al. 1994; Jardine et al. 1989). Enhanced carbonate dissolution is another environmental concern of NDOM-carbonate intera ction in a karstic aquifer. Organic acids occurring in NDOM and the production of CO2 via heterotrophic microbial activity can both enhance carbonate mineral dissolution, leading to geological hazards such as land surface subsidence (Wu, 2003). Understanding of subsurface NDOM-carbonate interaction is also important to assess the biogeochemical consequences of hydrogeologic projects such as aquifer storage and recovery (ASR), a process by which groundwater is recharged through wells to an aquifer and extracted for human use at some later time, and aquifer recharge (AR). Lastly, NDOM-carbonate interaction, which occurs during primary and secondary petroleum migration, could play an important role in the final composition of subsurface petroleum reserves (Mazzullo & Harris, 1992). In spite of its relevance to environmental and human health, very few studies have specifically examined NDOM-carbonate dynamics in karst. Because these systems may be hydrogeological complex and ill-defined, our research employs laboratory examinations of
13 carbonate-NDOM interaction as a starting point toward this goal. A range of karst materials from the Floridan Aquifer and NDOM types were combined in vitro and relevant geochemical measurements made such that the important abiotic processes that transform NDOM in the subsurface could be identified. An additional future goal would be to use both abiotic and biotic signature of NDOM interaction in groundwater sy stem, if identifiable, to trace the sources and travel paths of groundwater in the subsurface.
14 CHAPTER 2 BACKGROUND NDOM in Groundwater NDOM is a ubiquitous component of natural waters and is derived from microbial, plant and animal matter and the products of their decomposition in various environmental systems. As the major form in which terrestrial organic matter is made available to subsurface microbial systems and transferred to the fr eshwater aquatic and, ultimately, marine realms, much research on NDOM chemistry and reactivity was carried out. Although approximately 25% of NDOM can be characterized as amino acids, nucleic acids, carbohydrates, hydrocarbons, fatty acids or phenolic compounds, the vast majority of NDOM remains uncharacterizable on a molecular level. It is of complex and heterogeneous nature with a wide range of molar masses and chem ical structures that give it a multifunctional role in the natural environment (Frimmel, 1998). However, it is commonly operationally defined as humic substances, including fulvic, humic and hydrophilic acid fractions, based upon its solubility at various pH conditions. There are three main sources of NDOM to groundwater: organic matter (OM) deposits (e.g., buried peat, kerogen or coal), soil and sediment, and other water bodies (i.e. OM infiltrating into the subsurface from rivers, la kes, wetlands or marine systems) (Aiken, 2002). NDOM derived from different source materials should have distinctive chemical characteristics associated with those source materials. For instance, NDOM derived from higher plants have relatively large amounts of aromatic-C and phenolic content, and low N content. Microbially or algal-derived NDOM, on the other hand, has greater N content, and low aromatic-C and phenolic content (Aiken, 2002). The relative contributions of these sources of groundwater NDOM vary
15 between aquifer to aquifer and would be quantifi able were it not for the microbial and abiotic processes can slowly alter the structure and chemical reactivity of the NDOM. A major factor responsible for NDOM attenuation is biodegradation by indigenous soil and aquifer microorganisms. The respiration rate, which is an index of subsurface microbial activity, varies along the flow path as groundwa ter transitions from oxic to anoxic conditions (Murphy & Schramke, 1998). In many pristine aqui fers, where microbial activity is limited by the availability of electron donors (usually NDOM), dissolved inorganic carbon (DIC) accumulates slowly along aquifer flow paths and available electron acceptors are consumed sequentially in the order dissolved o xygen > nitrate > Fe (III) > sulfate > CO2 (Chapelle, 2000). Thus, biodegradation rates in the subsurface are not only controlled by the physiological capacity of the microbes, but also by subsurface nutrients available to the microbes (Simoni et al. 2001). The transformation of NDOM fractions occurs during its travel through the subsurface. For example, high molecular weight (Mw) components of NDOM was found preferentially adsorbed to surfaces, and have lower mobility than smaller Mw components during groundwater travel (Rauch & Drewes, 2005). Unretarded transport of the hydrophobic and macromolecular NDOM also found after adsorption indicating NDOM can exhibit considerable mobility in an aquifer and NDOM could alter the transport of contaminant in groundwater (McCarthy et al. 1993). Groundwater OM is one of the concerns during the implementation and evaluation of hydrogeologic projects such as soil-aquifer treatment (SAT) and ASR for treatment of both reclaimed water (Pavelic et al. 2005) and wastewater (Cha et al. 2004; Rauch & Drewes, 2004), which indirectly influences the trace elements st atus in an aquifer, such as As (Arthur, 2002; Price & Pichler, 2006). It was reported that approximately 20-25% of the organic carbon (OC)
16 injected into a carbonate aquifer was removed through aerobic respiration and denitrification (VanderzalM et al. 2006). During SAT, transphilic and hydrophilic OM can be preferentially removed during SAT, including non-aromatic OM such as large molecule polysaccharides and protein (Drewes et al. 2006). Biological processes are found to be the main contributor for the removal of OM, which interestingly appears to be a limiting factor for biomass growth in the system (Rauch & Drewes, 2005). Changes in OM chemical properties and Mw distribution were also found in a artificial groundwater recharge in Finland (Lindroos et al. 2002). Microorganisms, including bacteria, fungi, algae and protozoa, serve as geochemical agents in the subsurface by controlling a series of reactions that influence the subsurface environment. Microbes are capable of affectin g the fate of groundwater NDOM, for example, catalyzing the oxidation of OM (Lovley & Chapelle, 1995). In addition, they also influence the chemical composition of groundwater, such as accumulation of dissolved iron, methane and H2S (Lovley & Chapelle, 1995). The hydraulic property of an aquifer can also be affected by microbial processes. The production of CO2 and organic acids can lead to increased mineral solubility, developing secondary porosity thus increasing permeability in aquifer (Chapelle, 2000). On the other hand, microbes are found to cause active or passive mineral formation by precipitation and subsequent nucleation of crys tal formation (Ehrlich, 1998). A decrease in hydraulic conductivity also indicates clogging of a limestone fracture by groundwater microbes (Ross et al. 2001). In addition to biodegradation and NDOM-mineral interaction (discussed in following section), other major processes that are considered to control NDOM attenuation in the subsurface are co-precipitation and dilution (Leenheer, 2002). NDOM co-precipitation (with groundwater cations such as Al3+ and Ca2+) is likely to be controlled by pH and Eh changes in
17 the subsurface. Lastly, factors such as rainfall, snowmelt, recharge, and evapotranspiration can affect dilution and precipitation of NDOM in an aquifer indirectly. NDOM-Carbonate Interactions Although a variety of minerals may exist in carbonate aquifers, we shall assume a predominantly calcite composition (Randazzo & Jones, 1997) for the purposes of discussion here. There are many possible abiotic interactions between NDOM and calcite that can be examined including adsorption, desorption, dissolution and precipitation. These abiotic reactions may also, in turn, influence the microbially-media ted fate or transformation of NDOM. Adsorption: Adsorption of NDOM onto aquifer carbonate is a process that occurs when groundwater NDOM molecules accumulate on mine ral surface (sorbents), whereas desorption is a phenomenon whereby NDOM is released from carbonate mineral surfaces or matrices. Adsorption of organic compounds onto inorganic solid surfaces can significantly alter the physicochemical properties of the underlying solid, whose behavior (e.g., electrophoretic mobility, colloidal stability, transport, and inte raction with environmental contaminants) may be dominated by the adsorbed NDOM (Dunnivant et al. 1992; Gu et al. 1995; Gu et al. 1994; Jardine et al. 1989). For example, most solid surfaces in natural groundwater systems are negatively charged due to adsorbed NDOM (Gu et al. 1994; Liang et al. 1993; McCarthy et al. 1993; McKnight et al. 1992). The adsorbed NDOM coatings may also render hydrophilic surfaces hydrophobic and more capable of sorbi ng organic contaminants (McCarthy & Zachara, 1989; Murphy et al. 1992; Murphy et al. 1990). This is especially important to many waste disposal sites where organic ligands were found to compete with NDOM for adsorption onto sorbent surfaces. Besides sorption onto inorganic solids, organic compounds can bind with NDOM in the soil or groundwater environment, and thus have a considerable effect on the
18 dissipation of organic pollutants during soil and aquifer remediation (Amiri et al. 2005; Chen et al. 2005; Chen & Xing, 2005; Raber et al. 1998). The adsorption and desorption mechanisms of NDOM on carbonate minerals are not yet completely understood due to the heterogeneity and complexity of NDOM and adsorbent surfaces. The major mechanisms proposed by which NDOM may adsorb onto sorbent surfaces include: (i) anion exchange (electrostatic interaction), (ii) ligand exchange-surface complexation, (iii) hydrophobic interaction, (iv) entropic effect s, (v) hydrogen bonding, and (vi) cation bridging (Sposito, 2004). However, there are few direct measurements with which these mechanisms can be distinguished. Ligand exchange between surface-coordinated hydroxyl and H2O and organic compounds is most often invoked to explain mineral oxide-NDOM adsorption. Adsorption of a fulvic acid on goethite, for example, involves the complexation between COOof fulvic acid and -OH of the goethite and is usually accompanie d by an increase of pH, indicating that NDOM replaces hydroxyls on iron oxide surfaces (Tipping, 1981). However, Davis (1982) reported very little change in pH with the ad sorption of NDOM onto aluminum oxide and speculated that the stereo-chemical arrangement functional groups (e.g ., -COOH, -OH) in NDOM may play a role in determining which NOM molecules are adsorbed and which are not. A physical adsorption mechanism driven by favorable entropy changes has also been suggested as the dominant interaction mechanism in NDOM-soil adsorption system. For instance, Gu et al. (1994) showed that larger and more hydrophobic NDOM fractions were preferentially adsorbed by iron oxide over smaller size hydrophilic NDOM fractions. Due to their high surface area homogeneous mineral surface, metal oxides are widely used in adsorption investigations, particularly iron and aluminum oxides (Gu et al. 1994; Johnson et al. 2004; Saito et al. 2004; Yoon et al. 2005). Preferential adsorption of higher Mw,
19 more aromatic, and more hydrophobic fractions of OM was reported (Hur & Schlautman, 2003; Meier et al. 2004; Meier et al. 1999), and the cause for which was ascribed to presence of aromatic structures and acidic groups, as revealed by the close relationship between maximum sorption and total acidity and the content of ca rboxyl-C (Kaiser, 2003). Tombacz et al. (2004) reported that humic acids can be bound to th e most reactive surface sites on surface via complexation reaction. Some researchers pointed that sorbents mineralogy dominates the relation between OM adsorption and surface area (Kaiser & Guggenberger, 2003), while others believe that the number (and probably the position) of acidic groups attached to aromatics is the primary control of adsorption (Kaiser, 2003). Clay minerals are also commonly studied geosorbents (Baham & Sposito, 1994; Kahle et al. 2003; Kahle et al. 2004; Petrovic et al. 1999). Similar preferential adsorption phenomenon as observed in OM-metal oxides systems was also found when use clay minerals as sorbents (Hur & Schlautman, 2003; Kaiser, 2003; Wang & Xing, 2005). The adsorption affinity, capacity and desorption hysteresis were found correlated with structural features of the humic substances during their interaction w ith Na-kaolin clay (Balcke et al. 2002). The surface of clay mineral is likely to be modified during OM adsorption (Tombacz et al. 2004; Wang & Xing, 2005). Unlike metal oxide and clay minerals, adsorp tion of organic compounds onto calcite has not been extensively investigated, but a few st udies have reported significant adsorption. One study (Lee et al. 2005) reported adsorption of Suwannee River humic acid (SRHA) onto calcite to be rapid and mostly irreversible, with corresponding changes in electrostatic properties. The sorption of soil humic acids onto calcite was investigated for quite a long time (Carter & Mitterer, 1978; Muller et al. 1986; Petrovic et al. 1999) and even a computer model (de Leeuw, 2004) was employed to elucidate the adsorption mechanism between organic molecules and polymorph
20 calcite. Depletion in organic acidic compo unds in NDOM following ca lcite adsorption was ascribed to be preferential adsorption of these compounds to calcite (Carter, 1978; Muller et al. 1986). Another study of the adsorption of benzoate, citrate, tartrate and glutamate on calcite concluded that adsorption of organic compound with carboxylic groups resulted in a negative surface charge imparted to calcite (Plank & Bassioni, 2007). However, all these previous used finely ground pure and highly crystalline carbonate minerals such as calcite and dolomite do not necessarily represent the adsorption behavior of N DOM that will be observed in karst aquifer. Heterogeneous geosorbents such as aquifer rock, like soil and sediments, are complicated mixture of minerals that are expected to have a range of surficial properties and, therefore, behaviors with respect to NDOM adsorption. This is evident in the previous studies on OM adsorption to aquifer materials which have mainly been carried out in sedimentary, mainly sandy, aquifers. For example, it is reported that the behavior of NDOM sorption onto a sand and gravel aquifer is dominated by solid size fractions, as well as the ferrous minerals, which comprise only a small part of the aquifer material (Barber et al. 1992). Studies examining the adsorption of phenols onto a sandy aquifer material revealed that sorption decreases with increasing DOC concentration and pH in the solution (Amiri et al. 2005; Amiri et al. 2004; Ball & Roberts, 1991a). An intraparticle diffusion model was utilized to explain the adsorption of halogenated organic chemicals to aquifer materials (Ball & Roberts, 1991b). The origin and long-term transport of high and low Mw DOC fractions were investigated in a clay-rich till aquitard system (Hendry & Wassenaar, 2005). Retardation of NOM in a field-scale test (in a sandy aquifer) proved the NOM adsorption phenomena that observed in laboratory studies (McCarthy et al. 1993).
21 Adsorption mechanism has also been investigated on aquifer materials. For example, ligand exchange was regarded as the dominant reaction in sorption of fulvic and humic acids onto both synthetic (iron-oxide-coated quartz) and natural aquifer materials (soil and sediments) (Chi & Amy, 2004). In contracts, preferential adsorption of hydrophobic versus hydrophilic OM was also observed (Ussiri & Johnson, 2004), indicating the importance of hydrophobic interaction. Others have emphasized the importance of mineral texture on sorption processes. For example, a pore-filling mechanism was suggested by a study on low-polarity organic compounds (such as PAHs) sorption within a microporous aquitard (Xia & Ball, 1999). Similarly, entrapment of hydrophobic organic compounds in an aquifer kerogen isolate was explained by hole-filling mechanism (Ran et al. 2004). A conceptual model was proposed to describe microscale, intergranular variability in sorption properties of the aquifer sediments (Barber et al. 1992). However, no studies have examined the adsorption of bulk or NDOM fractions on natural heterogeneous carbonate materials. Desorption: Desorption of NDOM in an aquifer syst em, whether partial or complete, can significantly affect the OM and aquifer solid, and thus controls the fate and transport of subsurface OM and other organic or inorganic contaminants. For example, slow desorption kinetics limits the efficiency of soil and groundwater remediation as well as the bioavailability of contaminants (Carmichael et al. 1997; Karapanagioti et al. 2001). When adsorption is irreversible, that is, the whole of the adsorb ed NDOM cannot be completely desorbed under similar conditions, the system is termed hys teretic (Ditoro & Horzempa, 1982; Vaccari & Kaouris, 1988). Hysteresis was observed for a number of organic compounds adsorbed on soil sediments, pure clay minerals, and metal oxides (Barriuso et al. 1992; Ditoro & Horzempa, 1982; Pu & Cutright, 2006; Vaccari & Kaouris, 1988). However, strictly speaking, completely
22 irreversible reactions are thermodynamically impossible. Thus, the hysteresis of an adsorptiondesorption reaction is often considered as an experimental artifact due to either kinetic limitation or other factors such as chemical and biological activities (Gu et al. 1994). In any case, the degree to which adsorption of organic compounds may be reversible can be taken as an indication of the strength of the sorben t-sorptive interaction (Essington, 2004c). Dissolution/precipitation: A unique feature of karst, versus other types of aquifers, is that dissolution/precipitations may occur on short (experimentally observable) timescales. An issue that is still debated is whether organic compounds enhance or inhibit the dissolution of carbonate minerals. Some organic compounds were shown to enhance dissolution (Fredd & Fogler, 1998; McMahon et al. 1995; Perry et al. 2003; Wu & Grant, 2002) or at least inhibit precipitation (Inskeep & Bloom, 1986; Lin & Singer, 2005; Lin et al. 2004; Luttge & Conrad, 2004). Natural organic moieties such as acids may enhance calcite dissolution by simply releasing proton to the aqueous system (Equation 2-1): 2H+ + CaCO3 = Ca2+ + H2O + CO2 (2-1) Organic moieties may also fuel microbial respiration, producing CO2, which leads to calcite dissolution, following the overall reaction (Equation 2-2): CO2 + CaCO3 + H2O = Ca2+ + 2HCO3 (2-2) The calcite dissolution rate has also been found to increase significantly in the presence of chelating agents which bind calcium. A surface chelation mechanism was introduced to describe the dissolution and the rate of dissolution is found enhanced by the influence of proton attack at low pH (Fredd & Fogler, 1998; McMahon et al. 1995; Perry et al. 2003; Wu & Grant, 2002). On the other hand, some compounds were shown to inhibit calcite dissolution as well as enhance precipitation (Frye & Thomas, 1993; Hoch et al. 2000). The finding that compounds of
23 greater Mw and hydrophobicity (e.g., humic versus fu lvic acids) inhibit calcite growth most (Hoch et al. 2000; Lin & Singer, 2005; Lin et al. 2004), indicates that surface coverage plays an important role. Thermodynamic modeling suggest that carboxylic acids, hydroxyl aldehydes, and amides will be effective crystal growth inhibitors through their strong adsorption to the growth steps, thereby blocking these sites to further attachment by calcium (de Leeuw, 2004). Again, most previous research on the influence of OM on carbonate mineral dissolution or precipitation was carried out using pure mineral samples. Our study investigates the abiotic NDOM-carbonate rock interactions that may be pr edicted to take place in the Floridan aquifer through laboratory adsorption-desorption experiments using NDOM (collected from North Florida creeks and soil extracts) and aquifer mate rials. The present study is designed to be a starting point toward an understanding the range and complexity of behaviors of karst OMcarbonate subsurface system.
24 CHAPTER 3 MATERIALS AND METHODS Since the main purpose of our study is to model and investigate subsurface abiotic NDOM-carbonate interaction, experimental conditions and materials were chosen to as to mimic, as closely as possible, the subsurface aquifer environment. Natural forest soil and stream water NDOM sources were used as the sorbate and various aquifer rocks and core material from a range of depths were used as sorbent. Two sorbent size fractions, and NDOM concentrations spanning the natural NDOM concentration range, were tested in batch adsorption experiments and chemical analyses of the solute composition before and after mineral interaction were used to understand the role of abiotic processes on NDOM transformation. Sorbents: Adsorption experiments were carried out using two groups of aquifer materials: core material provided by the Florida Geological Survey (FGS) from regions of south-central Florida where ASR projects are underway, and relatively pure Floridan Aquifer rock, quarried and provided by the Florida Department of Transportation (FDOT) (designated C and R, respectively: Table 3-1). The original (before granulating) FDOT aquifer rocks are dense and relatively light-color carbonate comparing to FGS core material which were friable and brownish in color. The samples include representatives from each of these four major formations of the Upper Floridan Aquifer formations; Hawthorn, Suwannee, Ocala, and Avon Park (designated H, S, O and A, respectively: see Appendix II for Florida hydrogeologic framework). The major elemental composition of each FGS sample is shown in Table 3-2, as is the Brunauer-EmmettTeller (BET) surface area determined by N2 adsorption using a Quantachrome A1 Autosorb (Quantachrome Corp., Boynton Beach, FL). FGS core sorbent mineralogy was identified (by Dr. W. Harris, UF Soil and Water Science Department) using X-ray diffraction (XRD: see Appendix IV for figures). The C-A sample is nearly pure dolomite, C-O is nearly pure calcite, C-S is
25 mainly calcite and quartz, and the C-H sample has calcite, dolomite, quartz, apatite, and palygorskite. All samples were granulated and then sieved into <0.15 mm and 0.15-0.5 mm size fractions prior to surface area, XRD analysis, and adsorption experiments. The OM content of both FGS and FDOT samples was measured by lo ss-on-ignition (LOI, 4 50C for 4.5 hours) and converted to organic carbon (OC) content assuming 0.5 g OC/g OM (Drever, 2002a; Essington, 2004a). Only C-H had a surface area and OC content significantly greater than the others, while the OC content of both rock samples were clearly lower than that of the core samples. Sorbates: We have measured local groundwater samples ranging 0.5-13 mg OC/L and surface water concentrations ranging 1-32 mg OC/L, though OC concentrations in swamp waters can be significantly greater. The NDOM in most groundwater, however, is likely to be soilderived (Tipping et al. 1997). In our study, we use a north Florida mixed forest soil extract and stream water collected from Santa Fe River wa tershed, FL as the primary NDOM sources for adsorption experiments. For comparison, two other types of NDOM (two lakes near Gainesville, FL) were also tested as NDOM sorbates. Soil sa mple were mixed with water (soil/water volume ratio ~1:2) in a polycarbonate container and placed on an end-over-end shaker for 4 to 5 days. The soil-water mixture was then centrifuged at 4000 rpm for 5 min and the supernatants were filtered (0.45 m membrane, Millipore) and then concentrat ed by freeze-drying (not to dryness). Extracts were stored in the dark at 4C, prior to dilution (with deionized, photo-oxidized OC-free water) to prepare sorbate NDOM solutions of various concentrations. The stream and lake NDOM samples were filtered, concentrated, stored and diluted in the same manner as the soil NDOM sample. Adsorption and desorption experiments: Samples for batch adsorption experiments were prepared by mixing 40 ml background NDOM solutions (concentration ranges from 0 to
26 153.49 mg C/L) with amount of granulated aquifer materials that have 20 m2 surface area in 50 ml polypropylene centrifuge tubes. Tubes were then place in horizontal position on a shaker table (200 rpm) for 2-4 days at room temperature (22 2C). At the completion of the adsorption period, the suspensions were centrifuged at 4000 rpm for 5 min, and the supernatants were immediately pipetted into glass vials and stored in dark at 4C prior to chemical analyses. Duplicate reaction tubes along with controls with no sorbent were prepared for each NDOM concentration. Desorption studies were also conducted on selected samples. After removal of the supernatant of the adsorption experiment, tubes were weighed so that entrained solution amount could be calculated, and 40 ml OC-free water was added to each tube. Dissolved organic carbon (DOC) was then measured in the supernatant after another 4-day incubation period to determine the amount of OM desorbed. The amount of DOC adsorbed onto or desorbed from carbonate surfaces was assumed to be the difference between DOC in starting and ending solution. Adsorption and desorption isotherms were constructe d by plotting amount DOC ad sorbed or desorbed (qe, mg C/g sorbent) versus final equilibrium concentration (Ce, mg C/L) for a range of solution concentrations. The adsorption data were fit using three empirical adsorption models (linear, Freundlich and Langmuir). Model fits were quantified using least squares techniques. Although these three models can fit the data equally well (i.e. high R2 values for each model, complete fitting parameters listed in Appendix III), as the simplest model with the fewest assumptions, the linear model was chosen as best in keeping with the principle of parsimony. In addition, a simple model is more easily applied to estimations of NDOM-carbonate adsorption in natural hydrogeologic system where, in many instances, the amount of data available is insufficient to
27 justify a more sophisticated a pproach. Lastly, many more complex adsorption models, including Freundlich and Langmuir, may reduce to linear m odel under certain restrictive assumption or at lower sorptive concentration ranges. The equation for linear isotherm is given as follows (Equation 3-1): e D eC k q (3-1) where qe is the amount of adsorbed sorptive, normaliz ed to sorbent weight (in mg C/g sorbent), and Ce is the equilibrium organic carbon concentration in the background solution (in mg C/L). The distribution coefficient, kD (in L/g sorbent), is a function of the properties of the sorbent and sorptive interaction and is often referred as an indication of adsorption affinity (Drever, 2002b; Essington, 2004d; Schwarzenbach et al., 2003a). A slightly modified linear equation, which includes a y-axis intercept term, was employed in our study (Equation 3-2): b C k qe D e (3 -2) This was done to express the amount of indigenous OM released from the aquifer materials. Analytical Methods: Dissolved organic carbon was measured on a Shimadzu total organic carbon analyzer (TOC-5000A) after acidifying to pH 1 to 2 with 1M HCl and sparging for 2 minutes with OC-free air to remove inorganic carbon from samples prior to the measurement. The mean of three to five injections of 60 l is reported for every sample and precision, the coefficient of variance, was <5% for the replicate injections. The composition of NDOM in sorbate solution before and after adsorption experiments was examined by a variety of methods in order to determine the chemical nature of the adsorbed OM. Spectroscopic characteristics were examined using a Hitachi F-7000 fluorescence spectrometer to obtain three dimensional excitation-emission matrix (EEM) spectra of the samples. This instrument is equipped with a xenon lamp that energizes the sample with nearly
28 constant intensity in the range from 200 to 800 nm. Excitation wavelengths from 200 to 450 nm incremented at 10 nm intervals were collected, chosen as a compromise between data resolution and data collection time. For each excitation wa velength, emission was measured from 200 to 700 nm at 3 nm intervals. The molecular weight (Mw) distribution of NDOM before and after adsorption was determined by high performance liquid chromatography-size exclusion chromatography (HPLCSEC) according to the methods of Chin (1994), modified by Zhou (2000). The liquid chromatography system consisted of a solvent pump (Shimadzu LC-20AT5) and a UV-VIS detector (Shimadzu SPD-20A). The SEC column used was a TSK-GEL G3000SW (30 cm 7.8 mm diameter), purchased from Tosoh Bioscien ce LLC (Montgomeryville, PA) and compound detection was by UV-absorbance at 254 nm. A phosphate buffer was used as a mobile phase (0.1M NaCl, 0.002 M KH2PO4, and 0.002 M K2HPO4 buffered to pH 6.8) and flow rate was 1 ml/min. The HPLC-SEC was calibrated using four protein standards (albumin bovine, Mw=67 kD, ovalbumin Mw=45 kD, chymotrysinogen A Mw=24 kD, and bacitracin Mw=1450 D; all purchased from MP Biomedicals) at concentrations of 110-260 ppm. Retention time was converted to molecular weight using the calibration curve (see Appendix I). Major cations (Ca2+ and Mg2+) concentration and pH of the background solution, both before and after adsorption, were measured in order to detect possible mineral dissolution/precipitation reactions. Supernatant solutions were acidified to pH=1~1.5 and stored at 4C before analysis using an automated Dionex DX500 ion chromatograph (Dionex Corp., Sunnyvale, CA). Internal standards reflect a precision of better than 3% of the value of the measurement.
29Table 3-1. Sorbent background information Sample code used Geologic formation Location Sample information/core depth Core samplesa C-H Hawthorn Palm Beach County, FL W-18728; 797-798 ft C-S Suwannee Collier County, FL W-17801; 974-975 ft C-O Ocala Hendry County, FL W-18720; 1190-1191 ft C-A Avon Park Osceola County, FL W-18725; 980-985 ft Rock samplesb R-S Suwannee Brooksville, FL quarried R-O Ocala Newberry, FL quarried a. Core samples provided by the Florida Geological Survey (FGS), Tallahassee, FL. b. Floridan Aquifer rock samples provided by Florida Department of Transportation (FDOT), Tallahassee, FL. Table 3-2. Sorbents chemical composition Sorbent Sorbent surface area (m2/g)a Organic carbon contentb Major elemental composition (weight %)c <0.15 mm size fraction 0.15-0.5 mm size fraction SiO2 Al2O3 Fe2O3 MgOCaO Na2OK2O TiO2 P2O5 C-H 9.6 13 1.29 17.7 1.9 0.3 3.2 38.9 0.7 0.5 0.1 6.4 C-S 3.9 4.1 0.38 15.3 0.6 0.1 0.5 45.6 0.2 0.2 bdl 0.3 C-O 0.7 1.7 0.74 0.5 0.2 bdld 0.9 53.1 0.1 0.1 bdl bdl C-A 0.6 0.5 0.66 0.7 0.2 0.1 18.5 32.8 0.1 0.1 bdl bdl R-S 3.7 2.8 0.02 Not determined R-O 1.5 0.8 0.11 a. Sorbents surface area (m2/g sorbent) determined using multipoint N2 adsorption and calculated using BET theory. b. Sorbent organic carbon content (in weight %) determined by loss-on-ig nition (LOI) method (assuming organic matter composition of 50% organic carbon). c. Major elemental composition data (in weight %) of sorbents provided by Florida Geological Survey, Tallahassee. d. below detection limits
30 CHAPTER 4 RESULTS Quantity of NDOM-Carbonate Adsorption Four different types of NDOM (soil, stream and two lake NDOM) were tested using the same experimental treatment, but only soil and stream NDOM were selected for further investigation since the other two NDOM have similar chemical composition (i.e., EEM and HPLC results) as stream NDOM. Significant adsorption of both soil and stream-derived NDOM samples onto all carbonate rock sorbents was measured. Generally, there was no difference in the amount of DOC adsorbed from the two different NDOM samples onto carbonate sorbents, as all the adsorption data were within the confidence level of our isotherms. However, the FGS core materials generally sorbed more NDOM than FDOT aquifer rocks at the same background NDOM concentrations, indicating a greater adsorption affinity for the core materials. The FGS core materials showed release of indigenous rock OM when the concentration of background solution is low (especially in controls when ba ckground solution is OC-free DI water), while there was no such release for the FDOT aquifer materials. There was no significant pH change over the course of the 4-day adsorption period. Furthermore, the fine particle size fraction (<0.15 mm) for both aquifer rocks and core materials, although not very significant, tended to adsorb more NDOM than the coarse particle size fraction (0.15-0.5 mm). Adsorption isotherms for soil NDOM on granulated aquifer rocks of two grain sizes (<0.15 and 0.15-0.5 mm) and two time periods (2 and 4 days) are linear or near-linear (average R2 value of 0.995). This linearity extends to an equilibrium concentration of at least 153 mg C/L (Figure 4-1, model parameters in Table 4-1). In all cases, Suwannee formation rock (R-S), exhibited greater adsorption affinity (higher kD) than Ocala formation rock (R-O). Within rocks of the same geologic formation, the sorbents of the finer size fraction (<0.15 mm) had a larger kD
31 than their 0.15-0.5 mm counterp art when expressed on a weight-normalized basis. There was no significant difference in kD values, however, when surface-a rea normalized isotherms were plotted (Table 4-2). Longer equilibration periods for NDOM-carbonate interaction (i.e. 4 versus 2 days) led to greater kD values, though individual data points for 2 versus 4 day adsorption were not always significantly different. Therefore, our soil NDOM-carbonate adsorption system did not, technically, reach equilibrium in two or likely ev en four days and the full adsorption capacities of the sorbents remain unknown. Longer time period adsorption experiments were not carried out, however, due to the occurrence of dissolution (discussed later in text). Two-day adsorption isotherms for stream NDOM (0-20 mg C/L) sorb ed on each of the aquifer sorbents (C-S, C-H, C-A, C-O, R-O and R-S, all of 0.15-0.5 mm grain sizes) are plotted in Figure 4-2 (isotherm parameters in Table 4-1) In general, the isotherm linearity for FDOT aquifer rocks (average R2=0.9954) is better than that of FGS core material (average R2=0.9701, excluding C-A which has R2=0.4322). The unexpectedly low R2 for C-A (poor isotherm linearity), it is not likely caused by experimental error since experimental duplicate treatments are in close agreement (further discussion later in text). It is also clear from both plotted isotherms (Figure 4-2) and isotherm parameters (Table 4-1), that, unlike the aquifer rock isotherms with y-intercepts close to zero, those of aquifer core materials have negative values, implying that indigenous OM is released into solution. The amount of OM released from the C-O sample is greatest of all the aquifer sorbents tested, and sorption of NDOM only occurs at background solution concentrations greater than 13 mg C/L (Figure 4-2). To make comparisons easier, the percent of NDOM adsorbed by sorbent at two background concentrations for each NDOM-carbonate pair was calculated using the linear
32 adsorption isotherms constructed (Table 4-3). At low NDOM concentration (i.e. 5 mg C/L), the adsorption percentage is relativel y low, some of the treatments even show desorption (negative value). The adsorption percentage increases with the increasing of NDOM concentration. At a relatively high concentration (i.e. 20 mg C/L), the percentage for most the treatments fall in 3070% range and 72.3% was observed as the highest, indicating there were still some free NDOM in the solution that are ready to be adsorbed, and sorbents had been adsorbing maximum amount of NDOM in the given experimental time (2 or 4 days). Therefore, our adsorption systems were not restricted by the lack of enough NDOM in background solution, and our experimental design is fairly grounded. Immediately after the 4-day adsorption, a 4-day desorption experiment for Suwannee and Ocala rocks was also carried out using OC-free DI water to replace the NDOM solution. Since both rocks have similar adsorption-desorption patterns, only Suwannee (both <0.15 mm and 0.15-0.5 mm) is shown (Figure 4-3). At the low solution concentrations at which desorption was carried out, only a small portion of the NDOM remained sorbed to the aquifer rocks. That is, a small amount of hysteresis is observed, indicating that adsorption was not perfectly reversible. Composition of NDOM-Carbonate Adsorption Chemical analyses of soil and stream NDOM (sorbates) both before and after mineral interaction were carried out to investigate the possibility of pr eferential adsorption of specific NDOM components. Fluorometry: Fluorescence spectrometer is useful as a tool to compare the bulk fluorophore composition of heterogeneous sample s such as NDOM. A representative EEM plot of soil NDOM (Figure 4-4) shows four ma in fluorophores types detected and their corresponding location in EEM space which were id entified in the literature previously. They
33 include a fulvic-like peak (C : Excitation/Emission wavelengths 295-305 nm/410-420 nm), a humic-like peak (A: 210260/410450 nm), and two tryptophan-like peaks (T1: 275280/340 360 nm and T2: 215220/310340 nm) (Coble, 1996). EEM graphs of stream NDOM are similar to that of soil extract but lack the tryptophan-like fluorophores (Peaks T1 and T2), which are believed to be related to microbial activity (Parlanti et al., 2000). In addition, EEM identified the indigenous OM released from each of the core materials (into deionized OC-free water) as spectraphotometrically similar to each other as well as to stream water NDOM (Figure 4-5). That is, they lacked the tryptophan like fluorophores present in soil NDOM extract. While EEM data for each NDOM-sorbent pair are presented in Appendix VI, a single representative EEM is shown here (Figure 4-4) The difference EEM (fluorescent intensity matrix of NDOM before carbonate rock interaction, on top, minus, background solution intensity EEM after interaction, on bottom), shows the fluorophores that preferentially adsorbed as negative value intensities. Humiclike fluorophores were found pr eferentially adsorbed over the other three fluorophores by core materials during two days interaction period as it exhibits negative peak on the difference EEM graph (Figure 4-4). HPLC-SEC: Both NDOM sorbate types tested in our study displayed same Gaussian molecular weight distributions and retention times in size exclusion chromatograms. The average peak retention time of 11.7 0.1 minutes (Figure 4-6), corresponds to a Mw of 30.6 1.1 kD (calculated using the retention time-Mw linear relationship of protein standards, see Appendix I). Groundwater well sample collected from Santa Fe River Basin in north Florida were also found to have similar Mw distributions. These findings justify the use of soil and stream NDOM interchangeable and as repres entative of groundwater NDOM in the region.
34 Further, because freeze dried and non-fre eze dried NDOM displayed the same Mw distribution, indicating little chemical change occurred during laboratory processing. Two peaks were found in the chromatograms of indigenous OM via aquifer rock-DI water blank controls, with retention time of 11.9 (corresponding to 28.5 kD in Mw) and 12.7 minutes (19.7 kD), respectively. In addition to these two peaks, two other peaks were found in solution of core material-DI water controls: 5.9 (94.1 kD) and 16.2 minutes (~ 1000 D), indicating the difference in Mw distribution among indigenous OM of different sorbents and NDOM used as sorbate (Appendix V). Mw distributions of both NDOM types decreased (longer peak retention time) following sorbent interaction (FGS core materials, 2-day interaction: example in Figure 4-6). Thus, there was preferential adsorption of larger Mw over smaller Mw compounds which was slightly less for C-A and C-O (a 1 minute retention time increase corresponding to about 10.9 kD), and slightly larger for C-S and C-H (a 2 minutes retention time increase, corresponding to about 21.9 kD). No HPLC evidence shows that the sorbent size has any effect on NDOM adsorption character. Ion Chromatography: The concentrations of Ca2+ and Mg2+ ions in supernatant solutions following NDOM-carbonate interaction were measured to gauge the influence of NDOM on mineral dissolution af ter 2-day incubations of stream NDOM with FGS core material. In experimental controls (DI water-sorbent pairs), mineral dissolution occurred (as gauged by Ca2+ release) that was greatest for C-H and least for C-A samples on a weightnormalized basis. On a surface area-normalized basis, however, these trends were just the opposite (Figure 4-7). In addition, samples of fine particle size (< 0.15 mm) released more cations to solution th an those of the larger particle size fraction (0.15-0.5 mm).
35 Mineral dissolution was also detected after interaction with NDOM; however, the amount of cations released was decreased relative to that in DI water alone (Figure 4-7). Further, the decrease in mineral dissoluti on was enhanced with increasing concentrations of NDOM sorbent. This was observed for sorbents in both size fractions (<0.15 mm and 0.15-0.5 mm) and for both soil and stream NDOM. For example, after core materials (0.15-0.5 mm) two-day interaction with stream NDOM (concentration up to 18.19 mg C/L), on a weight-normalized basis, strongest Mg2+ dissolution inhibition was found in C-O (82.3% inhibition), followed by CS (73.1%), while C-H and C-A experienced simi lar relatively lower inhibition (38.7% and 44.5%, respectively). The extent of Ca2+ dissolution inhibition does not synchronize with that of Mg2+, with strongest inhibition found in C-A (66.6%) and weakest in C-H (16.2%). The same dissolution inhibition effect of NDOM also was found on a surface area-normalized basis (Figure 4-7).
36 Table 4-1. Linear model parameters (weight-normalized) Experimental treatmenta kD b bc R2 R-S, <0.15 mm, soil NDOM, 4-day, adsorption 0.005 -0.011 0.997 R-S, <0.15 mm, soil NDOM, 2-day, adsorption 0.005 -0.012 0.994 R-S, 0.15-0.5 mm, soil NDOM, 4-day, adsorption 0.004 -0.015 0.994 R-S, 0.15-0.5 mm, soil NDOM, 2-day, adsorption 0.003 -0.009 0.995 R-O, <0.15 mm, soil NDOM, 4-day, adsorption 0.002 -0.003 0.995 R-O, <0.15 mm, soil NDOM, 2-day, adsorption 0.002 -0.016 0.999 R-O, 0.15-0.5 mm, soil NDOM, 4-day, adsorption 0.001 -0.005 0.996 R-O, 0.15-0.5 mm, soil NDOM, 2-day, adsorption 0.001 -0.008 0.993 C-H, 0.15-0.5 mm, stream NDOM, 2-day, adsorption 0.012 -0.033 0.991 C-S, 0.15-0.5 mm, stream NDOM, 2-day, adsorption 0.011 -0.061 0.942 C-O, 0.15-0.5 mm, stream NDOM, 2-day, adsorption 0.013 -0.153 0.977 C-A, 0.15-0.5 mm, stream NDOM, 2-day, adsorption 0.005 -0.050 0.432 R-S, <0.15 mm, soil NDOM, 4-day, desorption 0.005 0.004 1.000 R-S, 0.15-0.5 mm, soil NDOM, 4-day, desorption 0.003 0.004 1.000 a. Treatment listed as sorbent type, particle size sorbate type, interaction time period, and experimental stage. b. kD is the slope (distribution coefficient) for linear model (L/g sorbent). c. b is the y-axis intercept of the linear model (mg C/g sorbent). Table 4-2. Linear model parameters (surface area-normalized) Experimental treatmenta kD b bc R2 R-S, <0.15 mm, soil NDOM, 4-day 0.002 -0.003 0.997 R-S, <0.15 mm, soil NDOM, 2-day 0.001 -0.003 0.994 R-S, 0.15-0.5 mm, soil NDOM, 4-day 0.002 -0.006 0.994 R-S, 0.15-0.5 mm, soil NDOM, 2-day 0.001 -0.003 0.995 R-O, <0.15 mm, soil NDOM, 4-day 0.001 -0.002 0.995 R-O, <0.15 mm, soil NDOM, 2-day 0.001 -0.011 0.999 R-O, 0.15-0.5 mm, soil NDOM, 4-day 0.001 -0.010 0.993 R-O, 0.15-0.5 mm, soil NDOM, 2-day 0.001 -0.007 0.996 C-H, 0.15-0.5 mm, stream NDOM, 2-day 0.001 -0.003 0.991 C-S, 0.15-0.5 mm, stream NDOM, 2-day 0.003 -0.015 0.942 C-O, 0.15-0.5 mm, stream NDOM, 2-day 0.008 -0.092 0.977 C-A, 0.15-0.5 mm, stream NDOM, 2-day 0.011 -0.102 0.432 a. Treatment listed as sorbent type, particle size, sorbate type, and contact time period. b. kD is the slope (distribution coefficient) for linear model (L/m2 sorbent). c. b is the y-axis intercept of the linear model (mg C/m2 sorbent).
37 Table 4-3. Amount NDOM adsorbed at two NDOM solution concentrations Experimental treatment 5.00 mg C/L 20.00 mg C/L NDOM adsorbeda(mg C/g sorbent) % NDOM adsorbed NDOM adsorbed (mg C/g sorbent) % NDOM adsorbed R-S, <0.15 mm, 4-day 0.014 44.1 0.089 67.2 R-S, <0.15 mm, 2-day 0.013 35.2 0.088 59.7 R-S, 0.15-0.5 mm, 4-day 0.005 18.6 0.065 59.7 R-S, 0.15-0.5 mm, 2-day 0.006 28.7 0.051 52.8 R-O, <0.15 mm, 4-day 0.007 51.0 0.037 64.4 R-O, <0.15 mm, 2-day -0.006 -47.6 0.024 34.6 R-O, 0.15-0.5 mm, 4-day 0.000 -1.3 0.015 50.0 R-O, 0.15-0.5 mm, 2-day -0.003 -40.3 0.012 35.3 C-H, 0.15-0.5 mm, 2-day 0.027 38.6 0.207 72.3 C-S, 0.15-0.5 mm, 2-day -0.006 -5.2 0.159 57.3 C-O, 0.15-0.5 mm, 2-day -0.088 -122.9 0.107 34.0 C-A, 0.15-0.5 mm, 2-day -0.025 -31.2 0.500 19.9 a. Calcualted using linear isotherms listed in Table 4-1.
38 Figure 4-1. Adsorption data for soil NDOM on RO and R-S carbonate rocks of two grain size fractions (<0.15 and 0.15-0.5 mm) and for 2 and 4 day interaction periods. Error bars represent the standard deviation of duplicate analyses. The solid lines are the linear model adsorption isotherms (details in text).
39 Figure 4-2. Adsorption data for stream NDOM on six carbonate sorbents (C-S, C-H, C-A, C-O, R-O and R-S). All the sorbents are of 0.15-0.5 mm grain size and adsorption time period is 2 days. Error bars represent the standard deviation of duplicate analyses. The solid lines are the linear model adso rption isotherms (details in text).
40 Figure 4-3. Adsorption and desorption data for soil NDOM on R-S of two grain sizes (<0.15 and 0.15-0.5 mm), 4-days. Error bars represent the standard deviation of duplicate analyses. The solid lines are the linear m odel fit of the adsorption and desorption isotherms (details in text).
41 Figure 4-4. Excitation-emission matrices (EEM) of A) soil NDOM extracts illustrating position of fulvic-like (Peak C), humic-like (Peak A), and tryptophan-like (Peak T1 and T2) fluorophores, B) stream NDOM, C) preferential adsorption of humic-like fluorophores: supernatant EEM following 2-da y interaction of carbonate rock (C-H, 0.15-0.5 mm) with soil NDOM (12.47 mg C/L) subtracted by starting supernatant EEM, D) preferential adsorption of humic-like fluorophores: following 2-day interaction of carbonate rock (C-H, 0.15 -0.5 mm) with stream NDOM (18.19 mg C/L).
42 Figure 4-5. Excitation-emission matrices of indigenous organic matter released from aquifer core materials (C-S, C-H, C-A and C-O, all <0.15 mm) into water.
43 Figure 4-6. Liquid size exclusion chromatograph (HPLC-SEC) of soil NDOM (14.47 mg C/L) sorbate before and after interaction with f our aquifer core materials (C-S, C-H, C-A and C-O, all 0.15-0.5 mm). Also shown are three protein molecular weight (Mw) standards (chymotrysino A, Mw=24 kD; albumin bovine, 67 kD; bacitracin, 1450 D).
44 Figure 4-7. Ca2+ and Mg2+ released during 2-day interact ion of distilled water and three concentrations of stream NDOM with aquifer core materials (C-S, C-H, C-A and CO, all 0.15-0.5 mm). Left: Cations released normalized to sorbent weight (mg/g sorbent), right: Cations re leased normalized to sorbent surface area (mg/m2 sorbent).
45 CHAPTER 5 DISCUSSION The results of NDOM-carbonate interaction experiments revealed significant adsorption of all NDOM types to all karst aquifer materials examined. While this adsorption occurred at all NDOM concentrations for the Floridan Aquifer rock samples tested, release (desorption) of indigenous OM from the aquifer core sample te sted occurred within the lower portion of the natural range of NDOM concentrations (<10 mg C/L). However, at higher NDOM concentrations, NDOM adsorption to the core materials occurred with an even greater affinity than that of the aquifer rock materials. For example, based on our linear adsorption model, at a commonly measured groundwater NDOM concentration (5 mg C/L), most of core materials as well as R-O showed negative values in both amount and percentage of NDOM adsorbed indicating the release of indigenous OM from the sorbents (Table 4-3). However, at a higher groundwater NDOM concentration level (e.g., 20 mg C/L) such as might be measured after a major precipitation event, all sorbents adsorbed NDOM (20-72% of that originally present in solution). A better understanding of the factors producing these trends can be obtained through an examination of the relationships between adsorption behavior and the characteristics of the aquifer sorbents tested. Factors Controlling Karst Aquifer NDOM Adsorption Unlike most previous research that has carried out adsorption experiments using pure mineral sorbents such as metal oxides, calcite, clays etc., and purified or fractionated organic sorbates such as individual compounds or compound classes, our study is distinguished by its use of both heterogeneous sorbents and sorbates. This was done in an effort to obtain empirical data that could be used to predict or explain temporally and spatially distributed field observations of aquifer geochemistry. Elemental and XRD analyses indicate that the carbonate
46 sorbents used in our study contain, in addition to calcite and/or dolomite (and quartz in the case of Hawthorn and Suwannee Fm. materials), variable amounts of apatite, metal oxides, clay, and even OM. It is unsurprising, therefore, that the more mineralogically homogeneous aquifer rocks produce highly linear isotherms, while those of the aquifer core materials contain significant scatter. To the contrary, it is in some ways surprising that the combination of heterogeneous sorbents and heterogeneous sorbates produce isotherms with any significant linearity at all. One explanation is that, although isotherms can be thought of as composed of a superposition of several individual isotherms that are characteristic for each specific type of sorbent and sorbate, they mostly reflect the adsorption character of the prevalent sorbate and sorbent pair (Schwarzenbach et al., 2003a). In our experiments, that appears to be humic acids and calcite. Adsorption data for NDO M interaction with carbonate are well-fit using a linear model, which is commonly associated with the adsorption of nonionic, non-polar, and hydrophobic organic compounds via weak, physical forces such as van der Waals attraction (partition theory, Essington, 2004b). Bulk, as opposed to fractioned, NDOM may vary widely in its hydrophobicity, as it commonly contains ionic or polar moieties (e.g., carboxyl and amine groups) coexisting with non-polar ones (e.g., alky l chains and aromatic rings). However, there is no dividing line separating hydrophobic from hy drophilic compounds in NDOM, nor fast rule to dictate which isotherm model should apply to each. However, accurate as partition theory may be to describe the sorption behavior of hydrophobic NDOM, it has its limitation when dealing with our NDOM-carbonate adsorption system. On the one hand, it fails to elucidate the sorption mechanism for polar, ionic and hydrophilic fraction of NDOM, whose isotherms are not linear over broad concentration ranges. On the other hand, the partition theory generally refers to an
47 absorption rather than an adsorption mechanism. This may be applicable if OM is the primary sorbent in aquifer materials but not, generally, for minerals without significant microporosity. Sorbent characteristics: Certainly, the physical and chemical property of carbonate aquifer sorbent materials is another factor likely to control adsorption. The contrasting character of the aquifer rock versus core materials make an interesting comparison in regards to their sorbent behavior. Though not determined, we can assume the aquifer rocks are of the same mineralogical composition as their core counterparts (i.e. from same geologic formation). Yet the rock materials are massive, unfractured and relati vely light-color relative to the more friable and tan-colored core materials. The latter is likely to have seen greater groundwater, NDOM and microbial interaction resulting in alteration and addition of minor impurities such as secondary minerals, indigenous OM and metal oxides. While each may have considerable macroporosity, the core materials are likely much more permeabl e, and so, may be better representatives of the aquifer materials likely have the greatest chance of chemically interacting with groundwater, and thus, affect their chemistry. One might expect mineral surface to exert a do minant control on surficial reactions such as adsorption. However, having very little significant microporosity, none of the aquifer materials displayed large surface areas. No significant correlation was found between NDOM adsorption affinity (kD) and the surface area of th e six carbonate sorbents (R2=0.774, p value of the slope=0.163). However, an excellent linear correlation is found between kD and surface area for the aquifer rock samples alone (Figure 5-1, R2=0.990, p=0.023). No such correlation was found among the core materials, alone (Figure 5-1, R2=0.933, p=0.543), though all but the Avon Park Fm. core sample displayed ro ughly twice the adsorption affinity of their rock counterparts.
48 One can conclude, therefore, that surface area may be a controlling factor of adsorption for relatively pure carbonate samples, but no t when other impurities are present. Another possible controlling factor of adsorption is the presence of indigenous OM in the aquifer sorbents. The pola r nature of the surface of most mineral sorbents including calcite (commonly hydroxyland oxy-moieties) will preferentially attract pol ar substances (e.g., water) over non-polar and nonionic NDOM via hydrogen bonding. Adsorption of organic molecules onto carbonate surfaces requires displacing the water molecules at such a surface, which is quite thermodynamically unfavorable. However, OM a dhering to, or incorporated within, a mineral renders its surface more hydrophobic and will not require the displacement of tightly bound water molecules prior to adsorption of additional OM (Schwarzenbach et al., 2003b). Previous studies found OM can bind with other organic compounds (Amiri et al., 2005; Chen et al., 2005; Chen & Xing, 2005), and the organic coatings may also render hydrophilic surfaces hydrophobic and more capable of sorbing organic contaminants (McCarthy & Zachara, 1989; Murphy et al., 1990). We should expect, therefore, that NDOM shows greater adsorption affinity for sorbents with more indigenous OM. A moderately strong linear correlation was found between the amount of indigenous OM in each sorbents and kD (Figure 5-2, R2=0.868, p=0.101). Though only significant to the 90% level of confidence, considering the other known or unknown impurities in the sorbents that could affect this relationship (and surface area as discussed previously), it is still reasonable to conclude that the presence of indigenous OM strongly influences NDOM adsorption. Adsorption kinetics: Long-term adsorption of humic acid to calcite (Lee et al., 2005), as well as to metal oxide (Gu et al., 1994), was described as largely irreversible, and was attributed to ligand exchange and inner-sphere-like interactions (Gu et al., 1994; Murphy et al., 1990).
49 However, for environmental processes, it is, perhaps, inappropriate to model adsorption reactions as either completely reversible or completely irreversible (discussed here,Ditoro & Horzempa, 1982) and not all are in complete agreement as to the predominant OM-mineral adsorption mechanism (discussed in next sections). It was suggested that the adsorption process may be made up of a rapid and reversible initial stage followed by a much slower nonreversible stage (i.e. kinetically biphasic,Selim et al., 1992; Vaccari & Kaouris, 1988). The formation of inner-sphere surface complexes and bonds of a covalent character are commonly believed to occur during the slow phase (Gu et al., 1994), whereas, we posit that the bonds formed during the 2 and 4-day experiments conducted in our study are mainly weaker outer-sphere interactions of the rapid initial phase. Because the carbonate sorbents exhibited greater NDOM adsorption after 4versus 2days, it can be said that adsorption equilibrium was not achieved. For instance, at a 5 mg C/L background NDOM concentration, R-S (<0.15 mm) showed an increase of 8.9% in NDOM adsorption after 4relative to 2-days, while there was an 14.7% increase for R-O (0.15-0.5 mm) at 20 mg C/L (Table 4-3). However, the increase was most apparent at higher NDOM concentration (>60 mg C/L), much higher than likely to be found in aquifers. By comparison, a previous study reported that SRHA adsorption to calcite was rapid, reaching equilibrium within in one day (Lee et al., 2005). In contrast, the non-equilibrium found here in our study may reflect the heterogeneity of both sorbents and sorbates. That is, carbonate rocks with a variety of minerals and surface functional groups may bind by a variety of mechanisms (rapid and slow forming) to a complex NDOM mixture. OM-mineral sorption systems were shown to reach equilibrium in timescales ranging from a few minutes to years (Avena & Koopal, 1999; Ball & Roberts, 1991a; Day et al., 1994;
50 Gu et al., 1995; Gu et al., 1994; Kleineidam et al., 2004; Ochs et al., 1994; Rugner et al., 1999; Zhou et al., 2001). For example, Zhou et al. (2001) reported that a fulvic-goethite adsorption system reached steady-state in about six hours. In contrast, no sorption eq uilibrium was found in an adsorption system of phenanthrene and aquifer materials (mineral separates from sandy aquifer sediments and fresh rock fragments) during adsorption experiments lasting 1010 days (Rugner et al., 1999). The transport of organic components in groundwater was shown to display non-equilibrium features due to very slow sorption kinetics onto aquifer materials or soils (Curtis et al., 1986; Pignatello et al., 1993; Roberts et al., 1986). In some cases, the slow sorption was attributed to intraparticle pore diffusion (Ball & Roberts, 1991b; Grathwohl & Reinhard, 1993; Werth et al., 2000), which is not applicable to our study (discussed further below). More recently, slowly re ached equilibrium was attributed to a high sorption capacity (Kleineidam et al., 2004). The carbonate system may be similar in that the linear isotherm displayed no evidence of reaching maximal adsorption capacity at the study conditions. However, from a thermodynamic prospective, all sorption system must, eventually, reach maximum adsorption capacity given either high enough NDOM concentration or long enough interaction time, or both. Adsorption/desorption mechanism: Assuming the studied carbonate system has a maximal adsorption capacity, though not observed, and considering the fact that adsorption of hydrophilic components can produce nonlinear adsorp tion behavior, it is likely that the linear isotherms observed in our study are only the low-concentration portion of a Freundlich or Langmuir isotherm with an adsorption ca pacity much higher th an the natural NDOM concentration range. In this case, adsorption sites on carbonate minerals are far from being saturated at low NDOM concentrations. However, at much higher NDOM concentration,
51 adsorption would become more and more difficu lt until all the sites are occupied (when kD=0). Others have also reported that different isotherm models apply to different sorbate concentration ranges (Lee et al., 2005; Murphy et al., 1992; Murphy et al., 1990). For example, Lee et al. (2005) found that a SRHA-calcite adsorption system showed Langmuir behavior at low concentrations (0-15 mg C/L) and non-Langmuir behavior at higher concentrations. The reason for such change in adsorption behavior was attributed to (1) a change of NDOM molecular orientation at the sorbent surfac e (Lu & Miller, 2002), (2) mutual interactions between sorbate molecules which enhances the adsorption affinity at higher loadings (Sposito, 1984), and (3) aggregation of NDOM on the sorbent surface as background concentration increases (Namjesnik-Dejanovic et al., 2000). During the desorption experimental stage, when background NDOM solution concentration was lowered, most, but not all of the adsorbed NDOM was released (e.g., 73-88% desorption for C-S), indicating that the NDOM-carbonate rock system, though close to reversible, had some hysteresis. However, factors other than true irreversible adsorption, can cause hysteresis, such as experimental artifacts, inappropriate experimental design, and nonequilibrium of either the adsorption or desorption process. Given that the adsorption leg is apparently not at equilibrium and that the appearan ce of hysteresis might be created by releasing addition OM into the low-concen tration solution during desorp tion stage, it may be that the hysteresis is a false one. However, a signal that desorption equilibrium wa s reached is the near equivalence of the desorption and adsorption kD (isotherm slope) (Essington, 2004c). For the 4day adsorption and desorption stage of R-S (<0.15 mm) with soil NDOM, both kD values were found to be 0.005, while for R-S (0.15-0.5 mm), adsorption kD (0.004) and desorption kD (0.003) were very similar (Table 4-1). In addition, considering the hydrophobic interaction between our
52 sorbates and sorbents (e.g., preferential adsorption of hydrophobic fraction of NDOM, discussed later), which would cause hysteresis, would lead to the conclusion that the NDOM-carbonate system is not completely reversible. Observations of the preferential adsorption of acidic organic compounds and an accompanying pH increase in background solution has led to ligand exchange to be a proposed adsorption mechanism OM adsorption onto calcite (Carter & Mitterer, 1978) and iron oxide (Tipping, 1981). Ligand exchange is not likely the primary adsorption mechanism occurring in our system, as no pH change was detected. However, considering that the carbonate mineral provide a strong pH buffer, the change in pH (reflecting the extent of ligand exchange) might be too small to be detected. Conformational changes in certain NDOM functional groups (e.g., COOH and OH) during hydrophobic interaction between NDOM and sorbents, is also able to give a good explanation for little acidity cha nge during adsorption (Avena & Koopal, 1999; Davis, 1982; Geffroy et al., 2000). On the other hand, the transformation of carbonate surface from hydrophilic to hydrophobic, (e.g., incorporating indigenous OM or forming organic coating via prior sorption), also facilitates the hydrophobic interaction. We propose that the NDOM-carbonate adsorption process consisted of a rapid and reversible initial stage, followed by a slower ir reversible stage. During the rapid initial stage, weak and outer-sphere bonding is the main association between carbonate sorbents and sorbed organic molecules. However, at longer time inte rvals, the stronger hydrophobic interactions occur, creates irreversibility and the observed desorption hysteresis. NDOM Transformation due to Abiotic Carbonate Interaction Our observation of preferential adsorption of a high over low Mw NDOM, and a humiclike over fulvic-like fraction onto carbonate sorbents is in agreement with the findings of other
53 published studies. For example, the higher Mw and hydrophobic fraction of Suwannee River fulvic acid (SRFA) standard were prefer entially adsorbed, re lative to lower Mw and hydrophilic fractions, to both kaolinite and hematite (Hur & Schlautman, 2003). Similar results were reported in studies of bulk NDOM adsorption to various clay and metal oxide sorbents (Gu et al., 1994; Meier et al., 1999). Kaiser (2003) also found such preferential adsorption in a NDOM-synthetic goethite adsorption system and concluded that strong sorption of the hydrophobic fractions may be caused by the presence of aromatic structures It was hypothesized th at hydrophobic effects, and to a lesser extent, ligand exchange, were the dominant mechanisms contributing to the preferential adsorption of SRFA and SRHA onto both synthetic and natural aquifer materials (Chi & Amy, 2004). Preferential adsorption of humic-like over fulvic-like compounds shown in our study (Figure 4-4c. and d.) also agrees with observations that post-mineral-interacted NDOM is likely to contain more fulvic acid than humic acid (Hedges & Oades, 1997). The preferential adsorption patterns observed in our systems were similar to those of other mineral oxides systems. Therefore, in aquifer systems, the aromaticity of NDOM is likely to be an important determinant for preferential adsorption of hydr ophobic and humic-like fraction of NDOM. This can well explained by adsorption mechanism that we proposed earlier. During the slow hydrophobic interaction stages, high over low Mw NDOM are preferentially adsorbed, and thus lead to sorption hysteresis. The indigenous OM detected in the aquifer materials may have both altered NDOM composition and sorbent surface chemistry, and thus influenced the adsorption process. Indigenous OM released from the carbonate sorbents was most apparent at low NDOM solution concentrations and among the core sorbent materi als. Little OM release from aquifer rocks was observed, which was not surprising given their very low OC contents (Table 3-2). However, no
54 significant linear correlation was found betw een the amount of indigenous OM released and measured sorbent OC content in each sample. For example, C-H, with the highest OC content among the four core materials, had the weakest tendency to release OM into solution, while C-O released the most OC. This indicates that the type and OM-mineral association of indigenous OM in each aquifer material is not likely the same. Other sorbent chemical parameters (such as surface area, elemental composition, adsorption affinity, or cations released to solution via dissolution) were also not found to be clearly related to OM release. We conclude that indigenous OM in these samples may be heterogeneously distributed in aquifer material and may be present as inter-particle inclusions as well as crystal surface coatings. The possibility of exchange of sorbate NDOM for indigenous OM must also be considered, and for this, OM compositional indicators can be useful. EEM results showed that indigenous OM released from carbonate samples generally contained humic-like substances (peak A), except for C-O, which contains both humic-like and fulvic-like substances (Figure 45). In addition, HPLC-SEC Mw distributions showed indigenous OM of both core materials and aquifer rocks to be composed of two common peaks (average Mw=28.5 and 19.7 kD), and two additional peaks in core materials (average Mw=94.1 and about 1 kD, see Appendix V), distinct from the NDOM peaks (average Mw=30.6 kD). Considering both EEM and HPLC results, we conclude that, although the main composition of indigenous OM may be categorized as humiclike substances, similar to that found in soil and stream NDOM, the indigenous OM also contains both larger and smaller Mw compounds that may be easily released when background NDOM concentrations are low, should fresh surfaces be exposed due to, for example, dissolution. The direct observation of these indigenous OM peaks (HPLC data) in the postinteracted background solution further confirms the exchange of OM between the aqueous and
55 solid phases. The origin of this indigenous OM is not known, but a likely source is microbiallyderived degradation products. It follows that some of this material is relatively more labile than the infiltrating NDOM for which it is exchanged on mineral surfaces and, may, therefore, fuel subsurface microbial activity. NDOM Effect on Carbonate Dissolution Though microbial conversion of OC to inorgani c C should increase the rate of carbonate dissolution (Perry et al., 2003; Perry et al., 2004), it is likely this effect was small to absent in our experiment because samples were treated aseptically and the 2-day interaction period allowed little time for microbial acclimatizati on to the OC source. Thus, release of Ca2+ and Mg2+ cations into solution (final starting cati on concentration in supernatant), which we attribute to mineral dissolution, is likely due wholly to abiotic hydrolyzation of mineral surfaces. On a weight-normalized basis, C-H had the greatest dissolution tendency and C-A least. However, on a surface area-normalized basis th e trend was just the opposite. This is likely because the higher surface area of C-H is due to clays which do not release cations. Although surface area may have been a contributing factor in determining dissolution extent, mineralogy played a more important role. Dolomite, the main constituent of C-A, has a much smaller solubility product than calcite, the dominant mineral of the other three core materials (at 25C, Kcalcite=10-8.48, Kdolomite=10-17.2 Drever, 2002c). This is reflected in our results that C-A released least amount of Ca2+ in DI water on a weight normalized basis. C-S showed smaller Mg2+ and Ca2+ release than C-O, likely due to the presence of quartz in C-S which co rrespondingly lowered its calcite content. For the relatively high cation release from C-H, though it contains significant quartz, may be due to the high solubility of apatite.
56 In general, the presence of NDOM inhibited mineral dissolution, but the dissolution inhibition observed was not always proportional to NDOM solution concentration. For example, no or only a small amount of inhibition occurred at 6.6 mg C/L NDOM (the concentration at which little NDOM adsorption occurs). There is only a small difference in extent of inhibition (and NDOM adsorption) at 11.8 versus 18.2 mg C/L NDOM. These observations suggest that the inhibition of mineral dissolution is due to protection of mineral surface by organic coating, and not to organic complexation of cations in solution. Though the issue of whether the present of OM inhibit or enhance carbonate dissolution is still being debated, our conclusion is supported by many previous studies. Thomas et al (Thomas et al., 1993) reported that organic compounds that adsorb strongly onto calcite and dolomite (fatty acids and carboxylated polymers) inhibited dissolution, while weak adso rbates and non-adsorbat es showed little or no effect on the dissolution rates. Phillips et al (200 5) reported that citrate molecules in a solution of precipitating calcite are incorporated structurally into the calcite crystal, and suggested that water and hydrogen-bonding interactions plays a ro le in the organic molecules-calcite interface. Lastly, it is noticed that, for each of the four core materials, the dissolution indicated by Ca2+ release is not the same as Mg2+. For example, the order of the extent of dissolution inhibition indicated by of Ca+2 among four sorbents at the highest NDOM concentration (18.19 mg C/L) is: C-A (66%) > C-S (58%) > C-O (41%) > C-H (16%), while for Mg2+, it is: C-O (82%) > C-S (73%) > C-A (45%) > C-H (39%). This asynchronous dissolution inhibition suggests that NDOM does not inhibit the dissolution of all minerals equally. For example, because Mg2+ release was more inhibited by the presence of NDOM than Ca2+ release, in all cases, Mg2+-containing minerals (such as dolomite) must have been relatively more protected by NDOM coverage.
57 Figure 5-1. Plot of sorbent surface area (m2/g sorbent) versus adsorption affinity (L/g sorbent) for six Floridan Aquifer materials.
58 Figure 5-2. Plot of sorbent indigenous organic matter content (g C/g sorbent) versus adsorption affinity (L/g sorbent) for six Floridan Aquifer materials.
59 CHAPTER 6 AQUFIER ENVIRONMENTAL IMPLICATION AND CONCLUSION The hydrogeology of an aquifer plays a role in the subsurface fate and transport of organic pollutants and natural OM. Factors such as depth to water table, sediment porosity and permeability, and groundwater flow velocity all co ntrol the pace and extent of OM transference both spatially and temporally. While it is well known that microbial utilization will alter the amount and character of NDOM along its subsurface flow path, the results presented here make clear the important role that abiotic factors can potentially play as well. Additionally, given its relatively greater surface connectivity and perm eability, and, thus, variability in NDOM concentration and contact time, the effects of ab iotic mineral interaction may be temporally and spatially variable. The results of our study show that, due to slow adsorption kinetics, longer groundwater retention or slower flow rates will lead to greater NDOM adsorption in a karst aquifer. At lower NDOM concentrations, such as during dry periods or at locations distant from NDOM sources, OM may be released from carbonate rock. In contrast, at greater NDOM concentrations, such as following rain even ts or at locations close to NDOM sources, considerable NDOM sequestration may occur via carbonate interaction. These processes will be enhanced by the presence of indigenous OM which will have the tendency to be released during periods of low NDOM concentration and enhance NDOM adsorption when greater concentrations of NDOM are present. Preferential adsorptions of some NDOM co mponents onto carbonate minerals may modify the properties (e.g. electrostatic) of aquifer mineral surfaces and, thus, have an important effect on their subsequent sorption behavior. Al most all solid surfaces in natural groundwater systems are negatively charged due to organic coating (Gu et al., 1994; Liang et al., 1993; McCarthy et al., 1993; McKnight et al., 1992). Hence, positively ch arged organic compounds in
60 groundwater (e.g., amino acids and proteins) ar e readily removed by cation exchange, while hydrophilic neutral (e.g. carbohydr ates and alcohols) and low Mw anionic OM (e.g. organic acids) are poorly retained by aquifer solids (Aiken, 2002). Preferential adsorption of humic acid-type to carbonate materials that we observed is also expect to affect groundwater quality as it can render the groundwater NDOM to be more hydrophilic. These OM components that are not adsorbed may be considered the more likely fuel for micr obial metabolism and are, potentially, pollutants of greater concern in regards to drinking water qual ity. Further, the fate and transport of nutrients such as nitrates and phosphates, too, may be tied to that of NDOM. The adsorption of NDOM to aquifer minerals removes organic N and P forms and prevents their possible remineralized by microbes. In contrast, NDOM released by carbonate can serve as an electron acceptor and stimulate the conversion of N and P into microbial biomass. The finding that the dissolution of aquifer material of all types examined is inhibited by the presence of NDOM at all concentrations has environmental implications as well. The presence of NDOM in aquifers might be expected to prevent the mobilization of heavy metals, radio-nuclides and other constituents from the aquifer matrix into groundwater. However, biotic processes must also be considered such as the release of inorganic carbon by microbial mineralization of NDOM, which will enhance carbonate dissolution, and the coupling of NDOM reduction with the oxidation of reduced minerals forms, thus altering the solubility of metals and radio-nuclides. As enhanced car bonate dissolution in karstic areas leads to geological hazards such as land surface subsidence, it will be import ant to determine the relative influence of abiotic and biotic factors on carbonate dissolution when considering projects that may change to groundwater chemistry.
61 An improved understanding of the abiotic NDOM-carbonate interaction processes which may occur in karst aquifer systems may be of benefit when conducting hydrogeologic projects such as ASR and AR. For example, a project manager may want to consider whether a pretreatment removal of NDOM prior to subsurface injection may ultimately be of benefit or harm, both from an environmental and from a gr oundwater quality point of view. While it is understood that caution is required when conducting projects such as ASR using water that contains NDOM, our study has shown that water with all NDOM removed may have implications for groundwater chemistry as well. Clearly additional work should be carried out to examine both abiotic and biotic NDOM-mineral inte ractions simultaneously, such as incubations using native groundwater microbial populations. These experiments should be carried out in both batch and column modes, and, lastly field scale ground-truthing studies should be carried out in well-monitored systems. The initiation of both ASR and AR projects in Florida afford just such an opportunity to examine the effects of NDOM interaction on subsurface biogeochemical processes.
62 APPENDIX A DETAILS OF ANALYTICAL METHODS USED HPLC: High-performance size exclusion chromatography (HPLC-SEC), also known as gel permeation chromatography or gel filtration ch romatography, is often used to characterize the molecular weight distribution of NDOM due to its sensitivity and reproducibility (Wu, 2003). In addition, HPLC requires relatively small samp le volumes, in many cases, without sample preconcentration. With the careful use of standards, it reveals not only the molecular size distribution, but also the weight (Mw) and number (Mn) averaged molecular weight of a sample. HPLC-SEC coupled with UV detection was commonly used to examine NDOM composition (e.g., Her et al., 2003; Lepane, 2001; Lepane & Kudrjashova, 2001), reactivity (e.g., Namjesnik-Dejanovic et al., 2000; Zhou et al., 2001), as well as changes in molecular size distribution following adsorption reactions (e.g., Hur & Schlautman, 2003; Specht et al., 2000). In our study, the Mw distribution of NDOM before and after adsorption was determined by HPLC-SEC according to the methods of Chin (1994), modified by Zhou (2000). The SEC column was calibrated using four protein standards. Name, concentration, Mw of the four proteins used and their corresponding retention time are shown in Table A-1. Calibration function obtained by linear regression of the Mw versus Retention time of protein standards (Figure A-1). EEM: Fluorescence is one possible outcome from molecular absorption of a quantum of light energy equal to the different between two electron orbital levels. The native fluorescence characteristics of NDOM can be utilized as a sensitive and non-destructive means of investigating the composition, source and transformation of NDOM (Baker, 2005). Recent instrument advances in fluorescence spectrometer permit the rapid measurement of NDOM fluorescence properties at low concentration over a continuous range of both excitation and emission wavelengths (Coble, 1996) and requires only a small sample size (<5 ml) (Baker & Lamont-Black, 2001). In our study, excitation-emission matrixes (EEM) of each NDOM-carbonate sorbent pair, both before and after 2-day interaction, were examined using a Hitachi F-7000 fluorescence spectrometer. Excitation wavelengths from 200 to 450 nm incremented at 10 nm intervals were collected. For each excitation wavelength, emission was measured from 200 to 700 nm at 3 nm intervals.
63 Table A-1. Calibration of the HPLC-SEC column Standard Name Concentration (ppm) Mw (kD) Retention Time (minute) albumin bovine 220 67 9 ovalbumin 260 45 9.7 chymotrysinogeo A 200 24 12.2 bacitracin 110 1.45 14.5 Figure A-1. HPLC-SEC colu mn calibration function
64 APPENDIX B FLORIDAN AQUIFER GEOLOGIC FRAMEWORK SERIES STRATIGRAPHIC UNIT HYDROGEOLOGIC UNIT Miocene Hawthorn Group Intermediate Confining Unit Florida Aquifer System Upper Florida Aquifer Oligocene Suwannee Limestone Eocene Ocala Limestone Avon Park Formation Middle Confining Unit Oldsmar & Cedar Keys Formations Lower Florida Aquifer Paleocene Figure B-1. Floridan Aquifer geologic fra mework (after Randazzo & Jones, 1997)
65 APPENDIX C ALTERNATIVE ADSORPTION MODEL PARAMETERS Table C-1. Langmuir Model ( e L e L eC K C K q q 1max) Parameters (weight-normalized) Experimental treatment KL (L/mg) qmax (mg/g) R2 R-S <0.15 mm, soil NDOM, 4-day 1.041E-08 5.126E+05 0.999 R-S <0.15 mm, soil NDOM, 2-day 1.661E-08 2.948E+05 0.997 R-S 0.15-0.5 mm, soil NDOM, 4-day 1.159E-08 3.446E+05 0.995 R-S 0.15-0.5 mm, soil NDOM, 2-day 1.702E-08 1.968E+05 0.997 R-O <0.15 mm, soil NDOM, 4-day 4.703E-08 4.295E+04 0.994 R-O <0.15 mm, soil NDOM, 2-day 3.446E-08 5.007E+04 0.994 R-O 0.15-0.5 mm, soil NDOM, 4-day 4.994E-08 1.955E+04 0.996 R-O 0.15-0.5 mm, soil NDOM, 2-day 6.960E-08 1.183E+04 0.989 C-H 0.15-0.5 mm, stream NDOM, 2-day 1.779E-03 6.958E+00 0.994 C-S 0.15-0.5 mm, stream NDOM, 2-day 9.843E-03 1.323E+00 0.973 C-O 0.15-0.5 mm, stream NDOM, 2-day 1.227E-02 1.236E+00 0.993 C-A 0.15-0.5 mm, stream NDOM, 2-day 1.769E-07 2.498E+04 0.583 Table C-2. Freundlich Model (n e F eC K q ) Parameters (weight-normalized) Experimental treatment KF (L/g) n R2 R-S <0.15 mm, soil NDOM, 4-day 0.004 1.063 0.999 R-S <0.15 mm, soil NDOM, 2-day 0.003 1.098 0.999 R-S 0.15-0.5 mm, soil NDOM, 4-day 0.002 1.150 0.999 R-S 0.15-0.5 mm, soil NDOM, 2-day 0.002 1.095 0.999 R-O <0.15 mm, soil NDOM, 4-day 0.001 1.077 0.995 R-O <0.15 mm, soil NDOM, 2-day 0.001 1.141 0.998 R-O 0.15-0.5 mm, soil NDOM, 4-day 0.001 1.082 0.997 R-O 0.15-0.5 mm, soil NDOM, 2-day 0.000 1.170 0.994 C-H 0.15-0.5 mm, stream NDOM, 2-day 0.012 1.013 0.994 C-S 0.15-0.5 mm, stream NDOM, 2-day 0.013 0.950 0.972 C-O 0.15-0.5 mm, stream NDOM, 2-day 0.017 0.900 0.992 C-A 0.15-0.5 mm, stream NDOM, 2-day 0.002 1.313 0.592
66 APPENDIX D SORBENT X-RAY DIFFRACTION RESULTS XRD measurements were conducted in Soil and Water Science Department, University of Florida by Dr. W. Harris. Results show that the C-A is nearly pure dolomite, the C-O is nearly pure calcite, C-H has calcite, dolomite, quartz, apatite, and palygorskite, and C-S is calcite and quartz. Figure D-1. Sorbent X-ray diffraction results
67 Figure D-1. Continued
68 APPENDIX E HPLC PEAK RETENTION TIME OF SORBENT INDIGENOUS ORGANIC MATTER Table E-1. HPLC peak retention time of sorbent indigenous organic matter Sorbent name Peak 1 Peak 2 Peak 3 Peak 4 R-O 11.8 12.8 R-S 11.9 12.6 C-H 5.9 11.9 12.8 16.2 C-S 5.9 11.9 12.8 16.2 C-O 5.8 11.9 12.6 16.2 C-A 5.9 11.9 12.6 16.2
69 APPENDIX F EEM GRAPHS FOR NDOM-SORBENT PAIRS EEM data for each NDOM-sorbent pair are presented in the form of a table consisted of three main columns: before adsorption (left column), after adsorption (middle), and EEM intensity change (right). The EEM intensity change was generated by deducting the intensity of before adsorption from that of after adsorption Therefore, change in intensity greater than zero shows desorption occurred, while less than zero means adsorption occurred. All the EEM graphs have the same scales: the horizontal axis of the EEM graph represents the emission wavelength, ranges from 200 to 700 nm, measured at 3 nm intervals, and the vertical axis represents the excitation wavelength, ranges from 200 to 450 nm, measured at 10 nm intervals. The water used in controls as background solution was deionized, photo-oxidized OC-free water.
70Table F-1. EEM graphs for NDOM-sorbent pairs Sorbent Sorbate Before adsorption After adsorption EEM intensity change C-S <0.15 mm Stream NDOM C=24.53 mg C/L C-S <0.15 mm Soil NDOM C=11.19 mg C/L C-H <0.15 mm Stream NDOM C=24.53 mg C/L
71Table F-1. Continued Sorbent Sorbate Before adsorption After adsorption EEM intensity change C-H <0.15 mm Soil NDOM C=11.19 mg C/L C-A <0.15 mm Stream NDOM C=24.53 mg C/L C-A <0.15 mm Soil NDOM C=11.19 mg C/L
72Table F-1. Continued Sorbent Sorbate Before adsorption After adsorption EEM intensity change C-O <0.15 mm Stream NDOM C=24.53 mg C/L C-O <0.15 mm Soil NDOM C=11.19 mg C/L C-S 0.15-0.5 mm Stream NDOM C=6.56 mg C/L
73Table F-1. Continued Sorbent Sorbate Before adsorption After adsorption EEM intensity change C-S 0.15-0.5 mm Stream NDOM C=11.81 mg C/L C-S 0.15-0.5 mm Stream NDOM C=18.19 mg C/L C-H 0.15-0.5 mm Stream NDOM C=6.56 mg C/L
74Table F-1. Continued Sorbent Sorbate Before adsorption After adsorption EEM intensity change C-H 0.15-0.5 mm Stream NDOM C=11.81 mg C/L C-H 0.15-0.5 mm Stream NDOM C=18.19 mg C/L C-A 0.15-0.5 mm Stream NDOM C=6.56 mg C/L
75Table F-1. Continued Sorbent Sorbate Before adsorption After adsorption EEM intensity change C-A 0.15-0.5 mm Stream NDOM C=11.81 mg C/L C-A 0.15-0.5 mm Stream NDOM C=18.19 mg C/L C-O 0.15-0.5 mm Stream NDOM C=6.56 mg C/L
76Table F-1. Continued Sorbent Sorbate Before adsorption After adsorption EEM intensity change C-O 0.15-0.5 mm Stream NDOM C=11.81 mg C/L C-O 0.15-0.5 mm Stream NDOM C=18.19 mg C/L C-S 0.15-0.5 mm Stream NDOM C=14.73 mg C/L
77Table F-1. Continued Sorbent Sorbate Before adsorption After adsorption EEM intensity change C-H 0.15-0.5 mm Stream NDOM C=14.73 mg C/L C-A 0.15-0.5 mm Stream NDOM C=14.73 mg C/L C-O 0.15-0.5 mm Stream NDOM C=14.73 mg C/L
78Table F-1. Continued Sorbent Sorbate Before adsorption After adsorption EEM intensity change C-S 0.15-0.5 mm Soil NDOM C=12.47 mg C/L C-H 0.15-0.5 mm Soil NDOM C=12.47 mg C/L C-A 0.15-0.5 mm Soil NDOM C=12.47 mg C/L
79Table F-1. Continued Sorbent Sorbate Before adsorption After adsorption EEM intensity change C-O 0.15-0.5 mm Soil NDOM C=12.47 mg C/L
80 LIST OF REFERENCES Aiken, G., (2002) Organic matter in ground water. In: U.S. Geological Survey Artificial Recharge Wo rkshop Proceedings Open-File Report 02-89 U.S. Geological Survey, Sacr amento, California. Amiri, F., Bornick, H., Wo rch, E., (2005) Sorption of phenols onto sandy aquifer material: the effect of dissolved organic matter (DOM). Water Research 39(5), 933-941. Amiri, F., Rahman, M.M., Bornick, H., Wo rch, E., (2004) Sorp tion behaviour of phenols on natura l sandy aquifer material du ring flow-thr ough column experiments: The effect of pH. Acta Hydrochimica Et Hydrobiologica 32(3), 214-224. Arthur, J.D., Dabous, A. A., Cowart, J. B., (2002) Mobili zation of arsenic and other trace elements during aq uifer storage and recovery, southwest Florida. In: U.S. Geological Survey Artificial Re charge Works hop Proceedings Open-File Report 02-89 U.S. Geological Survey Sacramento, California. Avena, M.J., Koopal, L.K., ( 1999) Kinetics of humic acid adsorption at solid-water interfaces. Environmental Science & Technology 33(16), 2739-2744. Baham, J., Sposito, G., ( 1994) Adsorption of Dissolved Organic-Carbon Extracted From Sewage-Sludge on Montmorillonite and Kaolinite in the Presence of Metal-Ions. Journal of Environmental Quality 23(1), 147-153. Baker, A., (2005) Thermal fl uorescence quenching properti es of dissolved organic matter. Water Research 39(18), 4405-4412. Baker, A., Lamont-Black, J ., (2001) Fluorescence of diss olved organic matter as a natural tracer of ground water. Ground Water 39(5), 745-750. Balcke, G.U., Kulikova, N.A., Hesse, S., Kopinke, F.D., Perminova, I.V., Frimmel, F.H., (2002) Adsorption of humic substan ces onto kaolin clay related to their structural features. Soil Science Society of America Journal 66(6), 1805-1812. Ball, W.P., Roberts, P.V., (1991a) Long-term Sorption of Halo genated OrganicChemicals by Aquifer Mate rial. 1. Equilibrium. Environmental Science & Technology 25(7), 1223-1237. Ball, W.P., Roberts, P.V ., (1991b) Long-term sorption of halogenated organic-chemicals by aquifer material 2. intraparti cle diffusion. Environmental Science & Technology 25(7), 1237-1249. Barber, L.B., Thurman, E.M ., Runnells, D.D., ( 1992) Geochemical He terogeneity in a sand and gravel aquifer effect of se diment mineralogy an d particle-size on
81 the sorption of chlorobenzenes. Journal of Cont aminant Hydrology 9(1-2), 35-54. Barriuso, E., Baer, U ., Calvet, R., (1992) Organic-chemicals in the environment dissolved organic-matter and adsorption-desorption of dimefuron, atrazine, and carbetamide by soils. Journal of Enviro nmental Quality 21(3), 359-367. Bob, M., Walker, H.W., (2001) Enhanced adsorp tion of natural or ganic matter on calcium carbonate particles through surf ace charge modifica tion, pp. 17-25. Boyes, S., Elliott, M., ( 2006) Organic matter and nutrie nt inputs to the Humber Estuary, England. Marine Pollution Bulletin 53(1-4), 136-143. Carmichael, L.M., Christman, R.F., Pf aender, F.K., (1997) Desorption and mineralization kinetics of phenanthrene and chrysene in contaminat ed soils. Environmental Science & Technology 31(1), 126-132. Carter, P.W., (1978) Adsorp tion of Amino Acid-Conta ining Organic-Matter by Calcite and Quartz. Geochimica Et Co smochimica Acta 42(8), 1239-1242. Carter, P.W., Mitterer, R.M., (1978) Amin o-Acid Composition of Organic-Matter Associated with Ca rbonate and Non-Car bonate Sediments. Geochimica Et Cosmochimica Acta 42(8), 1231-1238. Cha, W., Choi, H., Kim, J., Kim, I.S., (2004) Evaluation of wastewater effluents for soil aquifer treatment in South Korea. Water Science and Technology 50(2), 315-322. Chapelle, F.H., (2000) The si gnificance of microbial pro cesses in hydrogeology and geochemistry. Hydrogeology Journal 8(1), 41-46. Chen, B.L., Johnson, E. J., Chefetz, B., Zhu, L.Z., Xing B.S., (2005) Sorp tion of polar and nonpolar aromatic organic contaminants by plant cuticular materials: Role of polarity and accessibility. Environmental Scie nce & Technology 39(16), 6138-6146. Chen, B.L., Xing, B.S., (2005 ) Sorption and conformati onal characteristics of reconstituted plant cuticular waxes on montmorillonite. Environmental Science & Technology 39(21), 8315-8323. Chi, F.H., Amy, G.L., (2004 ) Kinetic study on the sorption of dissolved natural organic matter onto differe nt aquifer materials: th e effects of hydrophobicity and functional groups. Journal of Colloid and Interface Science 274(2), 380391.
82 Chin, Y.P., Aiken, G., Olough lin, E., (1994) Molecular-Wei ght, Polydisp ersity, and Spectroscopic Proper ties of Aquatic Humic Substances. Environmental Science & Technology 28(11), 1853-1858. Coble, P.G., (1996) Characterization of mari ne and terrestrial DOM in seawater using excitation emi ssion matrix spectroscopy. Marine Chemistry 51(4), 325-346. Curtis, G.P., Roberts, P.V., Reinhard, M., (1986) A natura l gradient experiment on solute transport in a sand aquifer. 4. sorption of organic solutes and its influence on mobility. Water Resources Research 22(13), 2059-2067. Davis, J.A., (1982) Adsorpti on of Natural Diss olved Organic-Matt er at the Oxide Water Interface. Geochimica Et Cosmochimica Acta 46(11), 2381-2393. Day, G.M., Hart, B.T., McKe lvie, I.D., Beckett, R., (1994) Adsorp tion of Natural Organic-Matter Onto Goethite. Colloids and Surfaces a-Physicochemical and Engineering Aspects 89(1), 1-13. de Leeuw, N.H., (2004 ) A computer modelling study of the uptake and segregation of fluoride ions at the hy drated hydroxyapatite (0001 ) surface: intr oducing Ca10(PO4)(6)(OH)(2) potential mode l, pp. 1860-1866. Ditoro, D.M., Horzempa, L.M ., (1982) Reversible and re sistant compon ent of PCB adsorption desorption isotherms. Environmental Scie nce & Technology 16(9), 594-602. Drever, J.I., (2002a) The Geochemistry of Natural Waters: Surface and Groundwater Environments, pp. 114116. Prentice Hall, Uppe r Saddle River, NJ. Drever, J.I., (2002b) The Ge ochemistry of Natural Wat ers: Surface and Groundwater Environments, pp. 88-90. Prentice Hall, Upper Saddle River, NJ. Drever, J.I., (2002c) The Geochemistry of Natural Waters: Surface and Groundwater Environments, pp. 419442. Prentice Hall, Uppe r Saddle River, NJ. Drewes, J.E., Quanrud, D.M ., Amy, G.L., Westerhoff, P.K., (2006) Character of organic matter in soil-aqu ifer treatment systems. Journal of Environmental Engineering-Asce 132(11), 1447-1458. Dunnivant, F.M., Schwarze nbach, R.P., Macalady, D. L., (1992) Reduction of Substituted Nitrobenzenes in Aqueous-Solutions Cont aining Natural OrganicMatter. Environmental Science & Technology 26(11), 2133-2141. Ehrlich, H.L., (1998) Geomicrobiology: its si gnificance for geology. Earth-Science Reviews 45(1-2), 45-60.
83 Essington, M.E., (2004a) Soil and Water Ch emistry: An Integr ative Approach, pp. 163-167. CRC Pre ss, Boca Raton. Essington, M.E., (2004b) Soil and Water Ch emistry: An Integr ative Approach, pp. 338-339. CRC Pre ss, Boca Raton. Essington, M.E., (2004c) Soil and Water Ch emistry: An Integr ative Approach, pp. 358-364. CRC Pre ss, Boca Raton. Essington, M.E., (2004d) Soil and Water Ch emistry: An Integr ative Approach, pp. 335. CRC Press, Boca Raton. Findlay, S.E.G., Sinsab augh, R.L., Sobczak, W.V., Hoos tal, M., (2003) Metabolic and structural response of hyporheic microbial communities to variations in supply of dissolved organic matter. Limnology and Oceanography 48(4), 1608-1617. Fisher, T.R., Hagy, J.D., Boynton, W.R., Williams, M. R., (2006) Cultural eutrophication in the Choptank and Patuxent estu aries of Chesapeake Bay. Limnology and Oceanography 51(1), 435-447. Fredd, C.N., Fogler, H.S., (1998) The influen ce of chelating agents on the kinetics of calcite dissolution. Journal of Colloid and Interface Science 204(1), 187-197. Frimmel, F.H., (1998) Ch aracterization of na tural organic matter as major constituents in aquatic systems. Journal of Cont aminant Hydrology 35(1-3), 201-216. Frye, G.C., Thomas, M.M., (1993) Adsorption of Organic-Compound s on Carbonate Minerals .2. Extraction of CarboxylicAcids from Recent and Ancient Carbonates. Chemical Geology 109(1-4), 215-226. Geffroy, C., Stuart, M.A.C., Wong, K., Cabane, B., Berger on, V., (2000) Adsorption of nonionic surfactants onto polystyrene: Kine tics and reversibility. Langmuir 16(16), 6422-6430. Grathwohl, P., Reinhard, M., (1993) Desorption of tric horoethylene in aquifer material rate limitati on at the grain scale. Environmental Science & Technology 27(12), 2360-2366. Gu, B.H., Schmitt, J., Chen, Z., Liang, L. Y., McCarthy, J.F., ( 1995) Adsorption and Desorption of Differe nt Organic-Matter Frac tions on Iron-Oxide. Geochimica Et Cosmochimica Acta 59(2), 219-229. Gu, B.H., Schmitt, J., Chen, Z.H., Liang, L. Y., McCarthy, J.F., ( 1994) Adsorption and Desorption of Natural Organic-Matt er on Iron-Oxide Mechanisms and Models. Environmental Scie nce & Technology 28(1), 38-46.
84 Hancock, P.J., B oulton, A.J., Humphreys, W.F., ( 2005) Aquifers and hyporheic zones: Towards an ecologi cal understanding of groundwater. Hydrogeology Journal 13(1), 98-111. Hedges, J.I., Oades, J.M., (1997) Comparativ e organic geochemist ries of soils and marine sediments. Organic Geochemistry 27(7-8), 319-361. Hendry, M.J., Wassena ar, L.I., (2005) Origin and mi gration of diss olved organic carbon fractions in a clay-rich aquitard: C-14 and delta C-13 evidence. Water Resources Research 41(2). Her, N., Amy, G., McKnight D., Sohn, J., Yoon, Y.M., (2003) Characterization of DOM as a function of MW by fluorescence EEM an d HPLC-SEC using UVA, DOC, and fluorescence detection. Water Research 37(17), 4295-4303. Hoch, A.R., Reddy, M.M ., Aiken, G.R., (2000) Calcite crystal growth inhibition by humic substances with emphasis on hydrophobic acids from the Florida Everglades. Geochimica Et Cosmochimica Acta 64(1), 61-72. Hur, J., Schlautman, M.A ., (2003) Molecular weight fractionation of humic substances by adsorp tion onto minerals. Journal of Colloid and Interface Science 264(2), 313-321. Inskeep, W.P., Bloom, P.R., (1986) Kinetics of Calcite Precipitatio n in the Presence of Water-Soluble Or ganic-Ligands. Soil Science Society of America Journal 50(5), 1167-1172. Jardine, P.M., Weber, N.L ., McCarthy, J.F., (1989) Me chanisms of Dissolved Organic-Carbon Adso rption on Soil. Soil Science Society of America Journal 53(5), 1378-1385. Johnson, S.B., Yoon, T.H., Ko car, B.D., Brown, G.E., ( 2004) Adsorption of organic matter at mineral/water interfaces. 2. Outer-sphere adsorpti on of maleate and implications for dissolution processes. Langmuir 20(12), 4996-5006. Kahle, M., Kleber, M., Jahn R., (2003) Reten tion of diss olved organi c matter by illitic soils and clay fractions: Infl uence of mineral phase properties. Journal of Plant Nutrition and Soil Science-Zeitschrift Fur Pflanzenernahrung Und Bodenkunde 166(6), 737-741. Kahle, M., Kleber, M., Jahn R., (2004) Reten tion of diss olved organi c matter by phyllosilicate and soil clay fractions in relation to mi neral properties, pp. 269276. Kaiser, K., (2003) Sorp tion of natural organic matter fractions to goethite (alphaFeOOH): effect of chemical compositio n as revealed by liquid-state C-13 NMR and wet-chemical analysis. Organic Geochemistry 34(11), 1569-1579.
85 Kaiser, K., Guggenberger, G., (2003) Mineral surfaces and soil organic matter. European Journal of Soil Science 54(2), 219-236. Karapanagioti, H. K., Gossard, C.M., Stre vett, K.A., Kolar, R.L ., Sabatini, D.A., (2001) Model coupling intraparti cle diffusion/sorption, nonlinear sorption, and biodegradation processes. Journal of Contam inant Hydrology 48(1-2), 1-21. Kleineidam, S., Rugner, H., Grathwohl, P., (2004) Desorption kinetics of phenanthrene in aquifer material lacks hysteresis. Environmental Science & Technology 38(15), 4169-4175. Lau, L.S., Mink, J.F., ( 1987) Organic Contamination of Groundwater a LearningExperience. Journal American Wate r Works Association 79(8), 37-42. Lee, Y.J., Elzinga, E.J., Reeder, R.J., ( 2005) Cu(II) adsorption at the calcite-water interface in the presence of natural organic matte r: kinetic studies and molecular-scale characterization. Geochimica Et Cosmochimica Acta 69(1), 49-61. Leenheer, J.A., (2002) Processes Controll ing Attenuation of Dissolved Organic Matter in the Subsurface. In: U.S. Geological Survey Artificial Recharge Workshop Proceedings Open-File Re port 02-89 (Ed. by G.R.A.a.E.L. Kuniansky). U.S. Geological Su rvey, Sacramento California. Lepane, V., (2001) Characteriz ation of Baltic Se a dissolved organic matter as oil shale precursor by separation and frac tionation by adsorp tion chromatographic XAD method, and size excl usion chromatography. Oil Shale 18(3), 239-257. Lepane, V., Kudrjashova, M., (2001 ) High-performance size exclusion chromatographic characteriz ation of humic substances and dissolved organic matter from Baltic aquatic environment. Oil Shale 18(4), 350-372. Liang, L.Y., McCarthy, J.F., Jolley, L.W ., McNabb, J.A., Mehlhorn, T.L., (1993) Iron Dynamics Transformation of Fe (II)/Fe (III) Duri ng Injection of Natural Organic-Matter in a Sandy Aquifer. Geochimica Et Cosmochimica Acta 57(9), 1987-1999. Lin, Y.P., Singer, P.C., (2005) Inhibition of calcite crysta l growth by po lyphosphates. Water Research 39(19), 4835-4843. Lin, Y.P., Singer, P.C., Aiken, G.R., ( 2004) Inhibition of he terogeneous calcite crystallization by Suwannee River fu lvic acid: Kinetics, mechanism and thermodynamics. Abstracts of Papers of the American Chemical Society, 228, U611-U611.
86 Lindroos, A.J., Kitunen, V., Derome, J., Helmisaari H.S., (2002) Changes in dissolved or ganic carbon during artificial recharge of groundwater in a forested esker in Southern Finland. Water Research 36(20), 4951-4958. Lovley, D.R., Chapelle, F. H., (1995) Deep subsurf ace microbial processes. Reviews of Geophysics 33(3), 365-381. Lu, Y.Q., Miller, J.D., ( 2002) Carboxyl stretc hing vibrations of spontaneously adsorbed and LB-transfe rred calcium carboxylates as determined by FTIR internal reflection spectroscopy. Journal of Co lloid and Interface Science 256(1), 41-52. Luttge, A., Conrad P.G., (2004) Direct ob servation of microbial inhibition of calcite dissolution. Applied and Enviro nmental Microbiology 70(3), 1627-1632. Mazzullo, S.J., Harris, P.M., (1992) Mesogenetic Dissolutio n Its Role in Poro sity Development in Carbonate Reservoirs. Aapg Bulletin-Americ an Association of Petroleum Geologists 76(5), 607-620. McCarthy, J.F., Williams, T.M., Liang, L.Y., Jardine, P.M., Jolley L.W., Taylor, D.L., Palumbo, A.V., Cooper, L.W ., (1993) Mobility of Natu ral Organic-Matter in a Sandy Aquifer. Environmental Scie nce & Technology 27(4), 667-676. McCarthy, J.F., Zachara, J.M ., (1989) Subsurface Transport of Contaminants Mobile Colloids in the Subsurface Envir onment may Alter the Transport of Contaminants. Environmental Scie nce & Technology 23(5), 496-502. McKnight, D.M., Bencala, K. E., Zellweger, G.W ., Aiken, G.R., Fe der, G.L., Thorn, K.A., (1992) Sorption of Dissolved Or ganic-Carbon by Hydrous Aluminum and Iron-Oxides Occurring at the Confluence of D eer Creek With the Snake River, Summit C ounty, Co lorado. Environmental Science & Technology 26(7), 1388-1396. McMahon, P.B., Vroblesk y, D.A., Bradley, P.M., Chapel le, F.H., Gullett, C.D., (1995) Evidence for Enhanced Mineral Dissol ution in Organic Acid-Rich Shallow Ground-Water. Ground Water 33(2), 207-216. Meier, M., Chin, Y.P., Maur ice, P., (2004) Variations in the composition and adsorption behavior of dissolved organic matter at a small, forested watershed. Biogeochemistry 67(1), 39-56. Meier, M., Namjesnik-Dejanovi c, K., Maurice, P.A., Chin Y.P., Aiken, G.R., (1999) Fractionation of aquatic na tural organic ma tter upon sorption to goethite and kaolinite. Chemical Geology 157(3-4), 275-284.
87 Muller, P.J., Suess, E., Ungerer, C.A., (1986) Amino-Acids and Amino-Sugars of Surface Particulate and Sediment Trap Material from Waters of the Scotia Sea. Deep-Sea Research Part aOceanographic Research Papers 33(6), 819-838. Murphy, E.M., Schramke, J.A ., (1998) Estimation of microb ial respiration rates in groundwater by geochemical modeling co nstrained with st able isotopes. Geochimica Et Cosmochimica Acta 62(21-22), 3395-3406. Murphy, E.M., Schramke, J.A ., Fredrickson, J. K., Bledsoe, H.W ., Francis, A.J., Sklarew, D.S., Line han, J.C., (1992) Th e Influence of Micr obial Activity and Sedimentary Organic-Carbon on the Isotope Geochemi stry of the Middendorf Aquifer. Water Resources Research 28(3), 723-740. Murphy, E.M., Zachara, J.M., Smith, S.C., (1990) Influence of Mineral Bound Humic Acids on the Sorption of Hydrophob ic Organic-Compounds. Abstracts of Papers of the Americ an Chemical Society 199, 98-ENVR. Namjesnik-Dejanovic, K., Ma urice, P.A., Aiken, G.R., Cabaniss, S., Chin, Y.P., Pullin, M.J., (2000) Adsorption and fractionation of a muck fulvic acid on kaolinite and goethite at pH 3.7, 6, and 8. Soil Science 165(7), 545-559. Ochs, M., Cosovic, B., Stumm, W., (1994 ) Coordinative and Hy drophobic Interaction of Humic Substances With Hydrophilic Al2O3 and Hydrophobic Mercury Surfaces. Geochimica Et Co smochimica Acta 58(2), 639-650. Parlanti, E., Worz, K., Geo ffroy, L., Lamotte, M., (2000) Dissolved or ganic matter fluorescence spectroscopy as a tool to estimate biologi cal activity in a coastal zone submitted to an thropogenic inputs. Organic Geochemistry 31(12), 17651781. Pavelic, P., Nicholson, B.C., Dillon, P.J., Barry, K.E., (2005) Fate of disinfection byproducts in groundwater during aquifer storage an d recovery with reclaimed water. Journal of Cont aminant Hydrology 77(1-2), 119-141. Perry, T.D., Duckworth, O. W., McNamara, C.J., Martin, S.T., Mitchell, R., (2003) Accelerated calcite dissolutio n caused by alginic acid. Abstracts of Papers of the American Ch emical Society 226, U591-U591. Perry, T.D., Duckworth, O. W., McNamara, C.J., Martin, S.T., Mitchell, R., (2004) Effects of the biologically produced polymer alginic acid on macroscopic and microscopic calcite dissolution rates. Environmental Science & Technology 38(11), 3040-3046. Petrovic, M., Kastelan-Macan, M., Horvat, A.J. M., (1999) Interact ive sorption of metal ions and humic acids onto mineral particles. Water Air and Soil Pollution 111(1-4), 41-56.
88 Phillips, B.L., Lee, Y.J., Re eder, R.J., (2005) Organic c oprecipitates with calcite: NMR spectroscopic evidence. Environmental Science & Technology 39(12), 4533-4539. Pignatello, J.J., Ferra ndino, F.J., Huang, L. Q., (1993) Elution of aged and freshly added herbicides from a soil. Environmental Scie nce & Technology 27(8), 1563-1571. Plank, J., Bassioni, G., (2007 ) Adsorption of carboxy late anions on a CaCO3 surface. Zeitschrift Fur Naturfor schung Section B-a Journa l of Chemical Sciences 62(10), 1277-1284. Price, R.E., Pichler, T., ( 2006) Abundance and mi neralogical association of arsenic in the Suwannee Limestone (Florida): Imp lications for arseni c release during water-rock interact ion, pp. 44-56. Pu, X.C., Cutright, T.J., (2006 ) Sorption-desorptio n behavior of PCP on soil organic matter and clay minerals. Chemosphere 64(6), 972-983. Raber, B., Kogel-Knabner, I ., Stein, C., Klem, D., (1998) Partitioning of polycyclic aromatic hydrocarbons to dissolved organic matter from different soils. Chemosphere 36(1), 79-97. Ran, Y., Xing, B.S., Rao, P.S.C., Fu, J.M., (2004) Im portance of adso rption (holefilling) mechanism for hydrophobic orga nic contaminants on an aquifer kerogen isolate. Environmental Scie nce & Technology 38(16), 4340-4348. Randazzo, A.F., Jo nes, D.S., (1997) The Geology of Florida University Press of Florida. Rauch, T., Drewes, J.E., ( 2005) Quantifying biological organic carbon removal in groundwater recharge systems. Journal of Environmen tal Engineering-Asce 131(6), 909-923. Rauch, T., Drewes, L., ( 2004) Assessing the removal potential of soil-aquifer treatment systems for bulk organic matter. Water Science and Technology 50(2), 245-253. Roberts, P.V., Goltz, M.N., Mackay, D.M., (1986) A natural gradie nt experiment on solute transport in a sand aquifer. 3. retardation estimates and mass balances for organic solutes. Water Resources Research 22(13), 2047-2058. Ross, N., Villemur, R., Desc henes, L., Samson, R., (2001 ) Clogging of a limestone fracture by stimulating groundwater microbes. Water Research 35(8), 20292037.
89 Rugner, H., Kleineidam, S., Grathwohl, P., (1999) Long term sorption kinetics of phenanthrene in aquifer materials. Environmental Science & Technology 33(10), 1645-1651. Saito, T., Koopal, L.K., va n Riemsdijk, W.H., Nagasa ki, S., Tanaka, S., (2004) Adsorption of humic acid on goethite: Isotherms, charge adjustments, and potential profiles. Langmuir 20(3), 689-700. Schlautman, M.A., Morg an, J.J., (1994) Adsorp tion of Aquatic Hu mic Substances on Colloidal-Size Aluminum-Oxide Particles Influence of Solution Chemistry. Geochimica Et Cosmochimica Acta 58(20), 4293-4303. Schwarzenbach, R.P., Gschwend, P.M., Im boden, D.M., (2003 a) Environmental Organic Chemistry (Second Ed ition). In: S. Edition (Ed. ) (Ed. by S. Edition), pp. 280-283. Wile y-Interscience. Schwarzenbach, R.P., Gschwend, P.M., Imboden, D.M., ( 2003b) Environmental Organic Chemistry (Second Ed ition). In: S. Edition (Ed. ) (Ed. by S. Edition), pp. 291. WileyInterscience. Selim, H.M., Buchter, B., Hinz, C., Ma, L., (1992) Mo deling the transport and retention of cadmium in soils multi reaction and multicomponent approaches. Soil Science Society of America Journal 56(4), 1004-1015. Simoni, S.F., Schafer, A., Harms, H., Ze hnder, A.J.B., (2001) F actors affecting mass transfer limited biodegradation in saturated porous media. Journal of Contaminant Hydrology 50(1-2), 99-120. Specht, C.H., Kumke, M.U., Frimmel, F.H., (2000) Characterization of NOM adsorption to clay minerals by size exclusion chromatography. Water Research 34(16), 4063-4069. Sposito, G., (1984) The Surface Chemistry of Soils Oxford University Press. Sposito, G., (2004) The Surface of Natural Particles Oxford University Press. Thomas, M.M., Clouse, J.A., Longo, J.M ., (1993) Adsorption of organic-compounds on carbonate minerals. 3. infl uence on dissolution rates. Chemical Geology 109(1-4), 227-237. Tipping, E., (1981) The Adsorption of Aqua tic Humic Substances by Iron-Oxides. Geochimica Et Cosmochimica Acta 45(2), 191-199. Tipping, E., Marker, A.F.H., Butterwick, C., Collett, G.D ., Cranwell, P.A., Ingram, J.K.G., Leach, D.V., Lishman, J.P., Pinder, A.C., Rigg, E., Simon, B.M., (1997) Organic carbon in the Humber rivers. Science of the Total Environment 194, 345-355.
90 Tombacz, E., Libor, Z., Illes, E., Majzik, A ., Klumpp, E., (2004) Th e role of reactive surface sites and complexati on by humic acids in th e interaction of clay mineral and iron oxide particles, pp. 257-267. Ussiri, D.A.N., John son, C.E., (2004) Sorption of organic carbon fractions by Spodosol mineral horizons. Soil Science Society of America Journal 68(1), 253-262. Vaccari, D.A., Kaouris, M., (1988) A model for irreversible adso rption hysteresis. Journal of Environmental Science and Health Part a-Environmental Science and Engineering & Toxic and Hazardous Subst ance Control 23(8), 797-822. VanderzalM, J.L., La Salle, C.L.G., Dillon, P.J., (2006) Fa te of organic matter during aquifer storage and recovery (ASR) of reclaimed water in a carbonate aquifer. Applied Geochemistry 21(7), 1204-1215. Wang, K.J., Xing, B.S., (2005 ) Structural and sorption ch aracteristics of adsorbed humic acid on clay minerals. Journal of Envir onmental Quality 34(1), 342349. Werth, C.J., McMillan, S.J., Castilla, H.J., (2000) Struct ural evaluation of slow desorbing sites in model and natura l solids using temperature stepped desorption profiles. 1. Model development. Environmental Science & Technology 34(14), 2959-2965. Wu, Y., (2003) Mech anism analysis of hazards caus ed by the inte raction between groundwater and ge o-environment. Environmental Geology 44(7), 811-819. Wu, Y.T., Grant, C., ( 2002) Effect of chelation chemistr y of sodium polyaspartate on the dissolution of calcite. Langmuir 18(18), 6813-6820. Xia, G.S., Ball, W.P., (1999 ) Adsorption-partitioning upt ake of nine low-polarity organic chemicals on a natural sorbent. Environmental Scie nce & Technology 33(2), 262-269. Yoon, T.H., Johnson, S.B., Brown, G.E., (2005) Adsorpti on of organi c matter at mineral/water interfaces. IV. Adsorp tion of humic substances at boehmite/water interfac es and impact on bo ehmite dissolution. Langmuir 21(11), 5002-5012. Zhou, Q.H., Cabaniss, S.E., Maurice, P.A., (2000) Cons iderations in the use of high-pressure size exclusion chromatogra phy (HPSEC) for determining molecular weights of aquatic humic substances. Water Research 34(14), 3505-3514. Zhou, Q.H., Maurice, P.A., Cabaniss, S.E., (2001) Size fr actionation upon adsorption of fulvic acid on goethite: Eq uilibrium and kinetic studies. Geochimica Et Cosmochimica Acta 65(5), 803-812.
91 BIOGRAPHICAL SKETCH Jin Jin was born in Shangh ai, China, in Ju ly 1984. He attended University of Science and Technology of China (USTC) from 2002 th rough 2006. Upon earning B.S. in environmental scie nces from USTC, he entered University of Florida and began his gradua te study in the Depart ment of Geological Sciences under the guidance of Dr. Andrew R. Zimmerman. Jin served as both teaching and research assistant during his graduate study and received M.S. in geology in May 2009. He will continue to purs ue his interests in environmenta l studies by enro lling in a PhD program.