Iv Geochemical and Isotopic Characterization of Coal

Geochemical and Isotopic Characterization of Coal Combustion Residuals:

Implications for Potential Environmental Impacts

Laura Suzanne Ruhl

Earth and Ocean Sciences Duke University

Date:______Approved:

______Avner Vengosh, Supervisor

______Gary Dwyer

______Heileen Hsu-Kim

______Paul Baker

______James Hower

Dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Earth and Ocean Sciences in the Graduate School of Duke University

2012

ABSTRACT

Geochemical and Isotopic Characterization of Coal Combustion Residuals:

Implications for Potential Environmental Impacts

Laura Suzanne Ruhl

Earth and Ocean Sciences Duke University

Date:______Approved:

______Avner Vengosh, Supervisor

______Gary Dwyer

______Heileen Hsu-Kim

______Paul Baker

______James Hower

An abstract of a dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Earth and Ocean Sciences in the Graduate School of Duke University

2012

Abstract

Coal fired power plants are widespread in the United States and most developed countries around the world, providing affordable electricity to consumers. In the US, approximately 600 power plants generate 136 million tons of Coal Combustion Residuals (CCRs) annually, encompassing fly ash, bottom ash, and flue gas desulfurization materials. The range and blends of

CCRs vary substantially across coal-fired plants and depend on a unique set of circumstances for each plant and coal source. Current U.S. regulations mandate the installation of advanced capture technologies to reduce atmospheric pollution, but do not address the transfer and storage, or the potential impacts to water resources. Thus, improved air quality is traded for significant enrichments of contaminants in the solid waste and effluent discharged from power plants.

This work examines the geochemical and isotopic characteristics of CCRs, as well as potential environmental impacts from CCRs. This investigation looks at several different aspects of CCR and environmental interactions from 1) the immediate impacts of the 2008 TVA coal ash spill in Kingston, TN, 2) the long-term (18-month) exposure of the spilled ash in the Emory and

Clinch rivers, 3) impacts on waterways in North Carolina that receive CCR effluent from coal fired power plants, and 4) examination of boron and strontium isotopes of CCRs from leaching experiments and their application as tracers in the environment of the TVA spill and NC waterways. These investigations have lead to the conclusion that 1) contact of surface water with

CCRs brings about leaching of high concentrations of certain CCR contaminants, such as As, Se,

B, Sr, Mo, and V in the surface waters; 2) the dilution effect is critical in determining the concentration of contaminants from the CCRs in surface water (both at the spill and in waterways receiving CCR effluent); 3) recycling of trace elements (such as As) through adsorption/desorption can impact water quality; and 4) elevated boron and strontium concentrations, in addition to their distinctive isotopic ratios, can trace CCR effluent in the iv

environment. Combining the geochemical behavior and isotopic characteristics provides a novel tool for the identification of CCR effluents in the environment.

Dedication

I dedicate this work in honor of my grandfather, Dan H. Ruhl, a great man and one of my favorite gators.

Contents

Abstract ...... iv

List of Tables...... x

List of Figures ...... xi

Acknowledgements ...... xiv

1. Introduction ...... 1

1.1 Background...... 1

1.2 Geochemistry and Isotopic Signature...... 3

1.3 Dissertation Research and Objectives ...... 6

1.4.1 Chapter 2 Synopsis: A Survey of the Potential Environmental and Health Impacts in the Immediate Aftermath of the Coal Ash Spill in Kingston, Tennessee ...... 8

1.4.2 Chapter 3 Synopsis: The Environmental Impacts of the Coal Ash Spill in Kingston, Tennessee: An Eighteen-Month Survey ...... 9

1.4.3 Chapter 4 Synopsis: The Impact of Coal Combustion Residual Effluent on Water Resources: A North Carolina Case Study...... 9

1.4.4 Chapter 5 Synopsis: Boron and Strontium Isotopic Characterization of Coal Combustion Residuals ...... 10

2. A Survey of the Potential Environmental and Health Impacts in the Immediate Aftermath of the Coal Ash Spill in Kingston, Tennessee ...... 12

2.1 Introduction ...... 12

2.2 Analytical Methods...... 13

2.3 Results and Discussion ...... 14

2.3.1 Coal ash and sediments ...... 14

2.3.2 Water Contamination...... 17

2.3.3 Potential Environmental Impacts ...... 19

2.3.4 Potential Health Impacts...... 22

vii

3. The Environmental Impacts of the Coal Ash Spill in Kingston, Tennessee: An Eighteen-Month Survey ...... 26

3.1 Introduction ...... 26

3.2 Analytical Methods...... 28

3.3 Results and Discussion ...... 29

3.3.1 Laboratory TVA coal ash leaching...... 29

3.3.2 Surface Water ...... 32

3.3.3 Sediment Acid Volatile Sulfide [101] ...... 33

3.3.4 Pore Water ...... 33

3.3.5 The Control of pH, redox state, and ash/water ratio on contaminant mobilization...... 38

3.3.6 Implications for tracing and prediction of CCR contaminants...... 41

4. The Impact of Coal Combustion Residual Effluent on Water Resources: A North Carolina Case Study ...... 43

4.1 Introduction ...... 43

4.2 Analytical Methods...... 47

4.3 Results and Discussion ...... 48

5. Boron and Strontium Isotopic Characterization of Coal Combustion Residuals: Validation of New Environmental Tracers ...... 61

5.1 Introduction ...... 61

5.2 Analytical Methods...... 64

5.2.1 Field Sampling and Leaching Experiments...... 64

5.2.2 Boron Isotopes...... 65

5.2.3 Strontium Isotopes...... 65

5.3 Results and Discussion ...... 66

5.3.1 Boron and Strontium in Coal Combustion Residuals...... 66

5.3.2 CCR Isotopic Signatures ...... 70 viii

5.3.4 Environmental Applications...... 74

5.3.4.1 The 2008 TVA Coal Ash Spill...... 74

5.3.4.2 CCR Effluent Discharges and Water Resources...... 76

6. Conclusions ...... 84

Appendix A- Supplementary Material for Chapter 2...... 87

Appendix B- Supplementary Material for Chapter 3...... 90

Appendix C- Supplementary Information Chapter 4 ...... 98

References ...... 103

Biography...... 119

List of Tables

Table 1: Average metals concentrations (mg/kg) in TVA coal ash and background soil in Kingston, TN. Samples were collected and analyzed by Tennessee Department of Environment and Conservation and the Tennessee Department of Health...... 15

Table 2: Hg results (mg/kg) in coal ash and river sediments associated with the spill area in Kingston, TN. Background data of Hg in Tennessee soil from the USGS Open Geochemical Database...... 16

Table 3: Radioactivity data (pCi/g unit) and activity ratios of coal ash and background data from the spill area in Kingston, TN. TDEC data are from The Tennessee Department of Environment and Conservation and the Tennessee Department of Health. Kentucky coal ash data are from Zielinski and Budahn [57] ...... 17

Table 4: Chemical composition of major (mg/L) and trace (mg/L) elements of water samples in the area of the TVA coal ash spill in Kingston, TN. “U” refers to unfiltered samples. EPA-MCL is the Maximum Contaminant Level for drinking water and EPA-CCC is the Criterion Continuous Concentration, which is an estimate of the highest concentration of a material in surface fresh water to which an aquatic community can be exposed indefinitely without resulting in an unacceptable effect [66, 67] . For Site location see Figure 3 and Supporting Information (Appendix A)...... 19

Table 5: Summary of concentration range, mean ± standard deviation (in brackets) of contaminants measured in different water sources in this study (n – number of samples). DL is the detection limit of the analytical method...... 33

Table 6: Listing of all of the water bodies sampled during this investigation. Also listed are the size of the power plants, the occurrence of a wet FGD system, and the amount of water discharged via the ash pond (CCR effluent) and the cooling water, which does not come in contact with CCRs...... 47

Table 7: The concentrations of ions in the effluent and the EPA regulatory levels are listed...... 51

Table 8: R squared and P values for the regression of the listed elements versus chloride. Group 1 showed a linear correlation. Group 2 had some correlation to chloride, while the third group had none...... 55

Table 9: List of the plants and CCRs that were sampled and then had leaching experiments performed on them...... 68

Table 10: Table of Plants and Coal Basins sampled. The δ11B and 87Sr/86Sr of the leachates from the 0.1 solid to liquid ratio leaching experiment as well as the concentration of the trace and major elements (in ppm normalized to weight) are listed...... 69

List of Figures

Figure 1: Variations of δ11B in the B isotopic system. Coal and coal ash have a much more depleted δ11B value than meteoric water [42] ...... 5

Figure 2: Strontium (87Sr/86Sr) isotopic ratio of a variety of geologic materials...... 6

Figure 3: Map of the sampling sites of the TVA coal ash spill in Kingston, Tennessee. Site descriptions are reported in Supporting Information (Google maps provided the base map)...... 13

Figure 4: Variations of Na+ and Cl- (A), Ca2+ and As (B), Sr2+ and B (D), and Sr2+ and As (D) in water samples from the Cove (triangles), Emory River upstream (open circles) and downstream (closed circles), and Emory - Clinch upstream (open squares) and downstream (closed squares). Mixing of Emory and Clinch Rivers (Line M1) is identified by the major ion composition (A). Elements that are enriched in coal ash, as reflected by the high concentrations in the Cove area (dashed line), show higher concentrations in the downstream Emory-Clinch river samples relative to the expected Emory-Clinch river mixing composition (Line M1). Line M2 reflects possible mixing of contaminants derived from coal ash leaching near the spill area (dashed line) and the uncontaminated Clinch River composition (Line M1)...... 21

Figure 5: Map of spill site of the TVA Kingston Fossil Plant. Emory and Clinch River mile markers (ERM and CRM) are labeled. The possible extent of the spilled ash upstream ...... 27

Figure 6: Variations of flow rate (cubic meters per second; red line, left-hand side axis) in the upstream Emory Rive USGS gauge at Oakdale, TN[116] and arsenic concentrations in downstream river water (black circles, right-hand side axis). Note that the low typical river discharges in the summer were associated with higher As concentrations in the downstream river water. Also note the timing of the spill event (black square) and dredging activities (arrow)...... 29

Figure 7: Results of leaching experiments of TVA coal ash under different ash/water ratios (upper panel) and pH conditions (lower panel). Variations of solid to water ratios, which are also expressed as total suspended solids[119] , show different pH effects and As concentrations in the deionized water (checkered symbols, upper panel) and upstream Emory River (blue circles) leaching experiments. Slopes for the linear correlations up to TSS=1000 mg/L are marked. pH variations show differential mobilization of LCACs in the experimental lechates (checkered symbols, lower panel), as compared to field measurements of pore water (black squares) and Cove water (red circles)...... 31

Figure 8: Spatial and temporal variations of As concentrations in filtered river water (top) and filtered pore water (bottom) as a function of distance from the spill site (defined as zero). Note that summer sampling yielded higher As contents in downstream river and that As in pore water increases with time in further downstream sites...... 36

Figure 9: Variations of B, Sr, As, Ba, and Se concentrations (mg/L unit, logarithmic scale) in pore water (black squares), Cove water (red circles), upstream river (open squares), and downstream river (blue circles). Note the relative enrichment of As and depletion of Se in pore water relative to the Cove water...... 37

Figure 10: AsIII versus total As measurements in pore water samples from the spill site in Kingston, TN (right) and As/B versus SO4/Cl ratios in pore water (black squares) and Cove water (red circles; left). Note that As species in most of the pore water is dominated by the reduced AsIII species and that As enrichment is associated with sulfate depletion in the pore water relative to the Cove water...... 38

Figure 11: Map of NC waterways sampled. The waterways have a coal-fired power plant utilizing the water for steam power generation or CCR associated processes. Also included is the reference lake, Jordan Lake...... 46

Figure 12: This plot shows the average outfall concentrations of CCR effluent from plants with an FGD system [154] and without an FGD system. The FGD outfalls tend to have higher concentrations of most constituents, except Li and V...... 50

Figure 13: Concentration Range found at the NPDES outfalls at the sampled waterways. Red symbols correspond to plants with and FGD system. Blue symbols are ash only, green are the reference lake (Jordan Lake) and black are cooling water lakes with no CCR ...... 52

Figure 14: Boron vs Chloride Concentration in Hyco and Mayo Lakes. The average outfall concentration (with standard deviation) is the red circle, the lake is blue squares, and the porewater is black diamonds. The concentration of boron in the lake is linear indicating its conservative behavior in the lake...... 56

Figure 15: Boron and arsenic concentration in porewater collected from Hyco Lake, Mayo Lake, Mountain Island Lake (MIL), and Jordan Lake at upstream (up), outfall, and downstream (down) locations relative to the NPDES outfall. Red symbols correspond to plants with and FGD system, blue symbols are ash only, and green are the reference lake (Jordan Lake). The EPA boron health advisory level is indicated, as well as the EPA CCC freshwater aquatic regulatory level...... 58

Figure 16: Leaching experiments on the spilled TVA CCRs [154] and the Duke Steam Plant (blue) fly ash. The Duke Steam Plant ash had lime addition post combustion. Low pH values leached off more strontium and boron than the neutral and basic pH values...... 70

Figure 17: Range of δ11B and 87Sr/86Sr found within major US coal basins, as well as the δ11B and 87Sr/86Sr of some common B and Sr source materials. The Spilled CCR is from an unknown coal source and the App/PRB mixture is a mixture of Appalachian and Powder River Basin source coals...... 73

Figure 18: The δ11B and 87Sr/86Sr from the variable pH leaching tests on the Spilled TVA ash [154] and the Duke Steam Plant Ash (Blue). The Spilled TVA ash has a more radiogenic 87Sr/86Sr value than the Duke Steam Plant Ash (that was treated with CaO)...... 73

Figure 19: The δ11B values of the CCRs produced in different units in the same plant from the same source coal (Table 9). This variation seen is a function of the heterogeneity of coal and the physical processes within the plant (mixing and pulverizer efficiency).. 74

xii

Figure 20: The delta B values versus B concentration and the 87Sr/86Sr values vs Sr concentration of the TVA Spill area waters (surface and pore water), as well as the leaching experiments on the spilled ash...... 76

Figure 21: The delta B and87Sr/86Sr values versus concentration of the CCR effluents from Tennessee and North Carolina coal fired power plants. They are divided up according to state and if they had and FGD system. The green point on the strontium graph represents an actual FGD sample...... 80

Figure 22: The delta B values of North Carolina water bodies and the CCR effluent discharged into them. Hyco and Mayo lakes show evidence of CCRs (negative delta B) throughout the lakes, while the outfalls (circles) of the other water bodies have different delta B values at the outfall compared to upstream...... 81

Figure 23: The 87Sr/86Sr values of the CCR effluent discharged into North Carolina waters. Each outfall (circles) has a distinctly different Sr isotopic value relative to the upstream 87Sr/86Sr values. This indicates that locally, 87Sr/86Sr values may be used as a tracer...... 81

Figure 24: Delta B and B concentration in Hyco and Mayo lakes. The porke water has a negative delta B value like the lake water (Figure 24). The pore water has slightly less boron than the outfall or surface waters (indicating possible adsorption)...... 82

Figure 25: Long Core pore water results. Delta B increases with depth, while both B and Cl concentration decrease with depth. The B/Cl ratio increases with depth...... 83

xiii

Acknowledgements

I embarked on this journey four years ago, with much road ahead of me. After countless hours in the field and lab, I am done with this chapter of my life. Getting my PhD has been challenging, yet rewarding. I’ve learned a lot about coal ash. I’ve learned a lot about boats and field sampling. I’ve learned patience and hard work can pay off.

None of this would have been possible without my advisor or my committee. Avner took me on as a student, and throughout this journey he has been a wonderful mentor. While challenging me, he has taught me the value of hard work, asking questions, looking at data in other ways, and getting that extra sample. He has given me the opportunity to travel, do lots of field sampling, and present at many conferences; things that most other students do not get to do during their PhDs. Gary has been a tremendous help throughout my PhD, and in my opinion rivals superman. Not only coming to the rescue at the TIMS and running my ICPMS samples, but also taking the time to address my questions about lab or field techniques; he has always been there to help. Helen has been a great mentor and teacher as well. I enjoyed learning geochemistry from her and feel that she has been a great influence in my success. Braving the extreme conditions (below freezing temps, where ice was forming on the deck) to join me in the field. She has also provided me with advice at times that I’ve found invaluable. Paul has been a great teacher and fun person to interact with throughout my PhD. Giving me advice on field sampling techniques and providing comments when needed, Paul has been a great help throughout. And last but certainly not least, Jim has been a wonderful person to get to know, and I am very grateful that we happened upon him after my first paper. Jim has been great reference for everything coal related, in addition to providing samples to complete this dissertation. He is always a bright spot at every meeting where I encounter him. I have been very lucky with such a

xiv

great and involved committee. I thank you for your time and the countless hours of help I’ve gotten from you all.

None of this could have been possibly without the help and support of my labmates: Nat

Warner, Brittany Merola, Hadas Raanan-Kiperwas, and David Vinson. Not only have they become great friends, they have always been around to answer questions, change a sample, and provide support and encouragement when needed. Also, I cannot forget the field-help and friendship I have received from others at Duke: Amrika Deonarine, Grace Schwartz, Alissa

White, Katie Barzee, and Andy Nunnery. These folks have been through heat, freezing temperatures, rain and many other hazards to help me gather samples, and they always do it with a smile and good attitude. Much of my North Carolina fieldwork would not have been possible without the help of Autumn Romanski and Danny Smith. I also appreciate their time and effort to provide edits/comments on my paper. I would also like to thank Neil Carriker and R. L. Pope for their assistance in collecting samples from the TVA plants.

My friends have provided me with many great memories from my time here at Duke, and been there to keep my spirits high. In addition to dealing with my hectic sampling schedule, they ensured that I’ve enjoyed my time here. Marianne Combs, Amy Morsch, Katie Grogan, Hannah

Aird, Brittany Merola, Nat and Casey Warner, Julie Ruhl, Amrika Deonarine, Alana Belcon, and

Kyle Bradbury have been great friends throughout my PhD, and I thank them for all of their kindness, fun times, and support. I want to give a special thank you to Kiril Kolev. Without him, my time here wouldn’t have been as sweet, enjoyable, or memorable. Thank you for your constant support, creative distractions, and your kind words of advice. I would lastly like to thank my family for their unending support and love. They have always been by my side to encourage, comfort, and strengthen. My parents, Leslie and Dan Ruhl, brother, Daniel, and sister, Julie have been lighthouses in the storm for me.

1. Introduction

1.1 Background

Coal is a major source of energy, providing over 42% of the world’s electricity [1] . Coal fired power plants are ubiquitous in the United States and most developed countries around the world, providing affordable electricity to consumers. A recent survey of the amount of coal ash generation in the USA revealed that 600 power plants nationwide generate approximately 136 million tons of Coal Combustion Residues (CCRs) each year, 43% of which is recycled into other materials. The remaining 70 million tons is stored in 194 landfills and 161 ponds in 47 states [2] .

CCRs encompass fly ash, bottom ash and boiler slag, and flue gas desulfurization (FGD) byproducts and are generally characterized as being enriched in trace elements such as As, B, Se,

Hg, and Sr. The prevention and isolation of CCR leachates from getting into the environment and water sources are of utmost importance to our health and natural resources. Developing integrated geochemical and isotopic fingerprints of the CCRs and their source coals, provides a novel approach that supplies the necessary diagnostic tools to combat one of the increasing sources of anthropogenic pollution in the United States and the rest of the world. The results of the laboratory experiments combined with field sampling are paramount in providing new insights into the true environmental fate of CCRs.

On December 22, 2008, the retaining wall on a waste retention pond at the Kingston

Fossil Plant, Tennessee, ruptured and an estimated 4.1 million m3 of coal ash slurry was spilled onto the surrounding land surface and into the adjacent Emory and Clinch Rivers [3] . This was the largest coal ash spill in US history and brought CCRs storage to the nation’s attention [4] .

One of the major potential hazards of coal ash storage in ponds is the continuous leaching of contaminants and their transport to the hydrological system. An EPA study [5] identified 63 coal ash landfills and ponds in 23 states where the coal combustion waste is associated with

contaminating groundwater and the local ecosystem. As such, the TVA coal ash spill and NC water bodies receiving CCR effluent have provided an “opportunity” to investigate these processes in a larger scale and to define geochemical tools for investigating the impact of CCRs on the environment.

The range and blends of CCRs varies substantially across coal-fired plants and depends on a unique set of circumstances for each plant and coal source [6] . Most coal-fired power plants burn a blend of coals. These chemical variations induce several implications with respect to the compositions of leachates that could be generated from CCRs in coal fired power plants and alternative migration paths in the environment. For example, the literature shows that the acidity of CCRs has important control on the quality and the trace metal abundances of CCR leachates

[7-10] and, moreover, the long-term evolution of contaminants in leachates [11] , which has important implications for the impact on the environment.

This dissertation establishes geochemical and isotopic tracers of CCR contaminants, providing the tools to identify and predict the fate of CCR contaminants in the environment. This dissertation is based on research of the TVA spill and CCR effluent in North Carolina, where the migration of metals from CCRs and the water quality impacts were monitored. The data generated in this project provides an in-depth evaluation how different coal sources control the levels and compositions of CCR contaminants, the environmental conditions in which contaminants are mobilized from CCRs, and their possible fate in aquatic systems. In addition, this novel approach for geochemical and isotope fingerprinting CCR contaminants provides a new monitoring tool that can be used to identify CCRs contaminants in aquatic systems and to distinguish them from other pollution sources that may also contaminate the environment.

Understanding the leaching systematics of CCRs is also important for developing adequate CCR management and beneficial use of recycling products of CCRs.

1.2 Geochemistry and Isotopic Signature

Coal is a sedimentary rock that consists of complex mixture of organic and inorganic constituents of detrital and authigenic origins [12, 13] . Once coal is burned in coal-fired power plants, the composition can change in a variety of ways. Some portions undergo no transition throughout the combustion stage, while some minerals are formed during combustion or during the transport and storage of CCRs [14] . Much of the organic material in coal is burned and released as CO2 during coal combustion, releasing some of the elements associated with the organic fraction (B, Ge, V, Be, Cr, Ni, etc) [15-17] . Most modern fluidized-bed combustion units maintain temperatures at or above 800-900ºC (with furnace temps up to 1500 ºC), which causes the decomposition of carbonates and sulfides, and releases some elements through sublimation

[15] . The elements likely to be in the flue gas are C, O, H, S, N, P, Na, K, As, Zn, Cd, Pb, Hg, and Se [15] . As the flue gas is vented away from the furnace, it begins to cool and the volatilized elements condense and adsorb to the surface of the fly ash particles [18-22] . The resulting coal ash is enriched in toxic elements like As, B, Be, Cd, Cr, Mn, Ni, Pb, Zn [23-28] . Elements that are adsorbed to the fly ash particle surface are more easily leached into solution during fly-ash water interaction [29] . Major minerals found in fly ash are quartz, mullite, magnetite, and hematite, which tend to be unreactive in the weathering environment [30] . The glass silicates associated with fly ash are prone to weathering at rates somewhere between the mineral phases and the adsorbed phases [30] . Previous leaching experiments have shown that there are several factors that control the mobilization of the coal ash contaminants such as pH, solid-liquid ratio, leaching time, combustion characteristics, and phase mineralogy [28, 31-34] .

While there is extensive literature on the chemistry of coal, CCRs, and their leachates [35-37] , only a few studies have addressed the isotopic compositions of CCRs. For example, boron has two stable isotopes (11B/10B ratio ~4), which has a large variation (total δ11B

range ~ 90 ‰; Venogsh 1998) in nature due to the large fractionation of relatively light isotopes.

The boron isotope ratios have been used to trace groundwater contaminants such as agricultural runoff, wastewater, and petroleum waste [38-41] . Williams and Hervig [42] measured B isotopic values of different coals in the USA. They determined that 10B was highly enriched in coal compared to most terrestrial rocks, and concluded that coal is characterized by negative δ11B

11 11 10 values (as low as -70 ‰) relative to the international standard (δ B is defined as [( B/ B)sample/

11 10 ( B/ B)NIST-SRM951 -1]x 1000 and the data is reported in permil). Williams and Hervig [43] suggested that this negative value could provide a tool for elucidating organic contaminants compared to relatively positive δ11B values found for most uncontaminated water (Figure 1). The limited data available on boron isotope ratios in CCRs suggest that coal ash has a similar δ11B value to the coal, measured at -4 to -19‰ [38, 44] , although the direct relationship between the composition of co-existing coal and CCRs has not yet been established. High temperatures associated with coal combustion volatilize most of the elements associated with the organic phases (including B) and volatile elements associated with minerals [15] . Spivak-Birndorf’s [44] analysis of CCRs indicate that there is no isotopic fractionation due to coal combustion, and that is retains its 10B enrichment. Boron is associated with the easily leachable fraction of elements

[29] that adsorbed onto the fly ash particles during cooling of the exhaust gas. Since there is no species specific (11B-enriched boric acid species relative to the 11B-depleted tetrahedral boron species) preferential leaching in water, there should be no isotopic fractionation when leached from CCRs. In this paper, the CCR boron isotopic composition is shown to be “universal” and is applied to different ash sources (based on coal origin) and plant configuration.

Figure 1: Variations of δ 11B in the B isotopic system. Coal and coal ash have a much more depleted δ 11B value than meteoric water [42] .

An additional isotope tracer proposed in this study is strontium (87Sr/86Sr) that is unique due to the lack of fractionation in the hydrological system [45] . Coal beds in Wyoming were analyzed for 87Sr/86Sr isotopes as a tracer for hydrologic basin movement, which revealed heterogeneity in Sr ratios ranging from 0.712 to 0.714 [46] . Hurst et al [47] analyzed coal, fly ash, and bottom ash for 87Sr/86Sr and found a range of 0.70883 to 0.70972. Each of these analyses was performed on specific coal and coal ashes, therefore, Sr isotopes could be different for different basins, which could be used as a tool to detect the link between CCRs and the original coal source. Given that the Sr isotopic composition of coal and CCRs are expected to reflect the local rocks and hydrology of the original depositional environment in which coal was generated

(Figure 2), isotopic variations among CCRs could be linked to their original coal source. In fly ash, a minor portion of strontium is potentially leachable due to surface associations, but the remainder is associated with silicate phases, therefore, the leaching is quite slower [30] . There

could be some different Sr isotopic ratios due to different compositions of different phases in the ash. Another factor to consider with Sr isotopes would be the addition of lime (CaO) or CaCO3 with a different Sr isotopic ratio in a SO2 scrubbers and its impact on the Sr isotope fingerprints of leachates originated from such CCRs (e.g., CaCO3 contains 500-1500 ppm Sr).

Figure 2: Strontium (87Sr/86Sr) isotopic ratio of a variety of geologic materials.

1.3 Dissertation Research and Objectives

This in-depth research investigation was driven by the following hypotheses:

1) CCRs are highly enriched in contaminants that are highly mobilized from CCRs that could affect associated water.

2) Specific environmental conditions (pH, redox state) control the species distribution of oxyanions (arsenic, selenium, boron) that are important components of the CCRs contaminants.

Therefore, the distribution of oxyanions derived from CCRs would control the reactivity of these contaminants with the host sediments (lake/river) and/or aquifer rocks (groundwater).

3) Boron and strontium isotopes of coal and CCRs have unique isotopic values; therefore, 6

they are good indicators of CCRs and or source coals, which could have significant environmental applications.

4) CCRs and their leachates are released to the environment through the National

Pollutant Discharge Elimination System outfalls and groundwater leaching; therefore, the contaminants in the leachate can remain and be traced in the hydrologic system (i.e. surface and pore water).

5) The large quantities of water utilized in NC for coal-fired power generation and discharge to the environment could have major impacts on the quality of surface water, ground water, and pore water in the bodies of water, which receive discharge from ash disposal facilities.

6) Pore water in river and lake sediments could be an important representative monitoring site for evaluating the long-term impact of CCRs on the water resources and predicting the fate and bioaccumulation of contaminants in the biological food chain.

Based on these hypotheses, this investigation evaluated the environmental impacts, the storage, and fate of leachates originating from CCRs, as well as the characterization of the geochemical and isotopic properties on CCRs. The specific objectives of this study were:

• Objective 1 –Establish novel “geochemical/isotopic fingerprints” for identification of CCR

contaminants in the environment that will include trace metal distribution, boron

(11B/10B), and strontium (87Sr/86Sr) isotopes.

• Objective 2 – Determine the factors that control the reactivity of CCRs and the

mobilization of trace metals, as well as the possible links to the coal origin and CCRs

generation.

• Objective 3- Explore the range of Sr and B isotopes for CCRs to determine whether there is

a “universal” isotopic signature or whether they can be used to determine the source coal.

• Objective 4 – Determine the environmental conditions that affect the reactivity of CCRs

such as pH, redox, salinity, and properties of the CCRs.

• Objective 5 – Evaluate the overall impact of CCRs on water resources in NC due to the

immense amount of water used in the coal fired power plant process.

• Objective 6- Apply the laboratory based geochemical and isotopic results to lakes and river

field sites (impacted by CCRs) in North Carolina by sampling and analyzing the water,

sediment, and pore water.

• Objective 7- Determine the fate of CCR leachates in the aquatic realm by measuring

contaminants in pore water.

1.4.1 Chapter 2 Synopsis: A Survey of the Potential Environmental and Health Impacts in the Immediate Aftermath of the Coal Ash Spill in Kingston, Tennessee

This chapter reports the results of water and sediment immediately following the TVA coal ash spill. An investigation of the potential environmental and health impacts in the immediate aftermath of one of the largest coal ash spills in US history at the Tennessee Valley

Authority (TVA) Kingston coal-burning power plant has revealed three major findings. First, the surface release of coal ash with high levels of toxic elements (As=75 mg/kg; Hg=150 ug/kg) and radioactivity (226Ra+228Ra=8 pCi/g) to the environment has the potential to generate resuspended ambient fine particles (<10 um) containing these toxics into the atmosphere that may pose a health risk to local communities. Second, leaching of contaminants from the coal ash caused contamination of surface waters in areas of restricted water exchange, but only trace levels were found in the downstream Emory and Clinch Rivers due to river dilution. Third, the accumulation of Hg- and As-rich coal ash in river sediments has the potential to have an impact on the ecological system in the downstream rivers by fish poisoning and methylmercury formation in anaerobic river sediments.

1.4.2 Chapter 3 Synopsis: The Environmental Impacts of the Coal Ash Spill in Kingston, Tennessee: An Eighteen-Month Survey

This paper reports on the longer term environmental impacts from the Tennessee Valley

Authority (TVA) coal ash spill in Kingston, Tennessee. An eighteen month investigation of the environmental impacts of the spill combined with leaching experiments on the spilled TVA coal ash have revealed that leachable coal ash contaminants (LCACs), particularly arsenic, selenium, boron, strontium, and barium, have different effects on the quality of impacted environments.

While LCACs levels in the downstream river water are relatively low and below the EPA drinking water and ecological thresholds, elevated levels were found in surface water with restricted water exchange and in pore water extracted from the river sediments downstream from the spill. The high concentration of arsenic (up to 2000 ug/L) is associated with some degree of anoxic conditions and predominance of the reduced arsenic species (arsenite) in the pore waters.

Laboratory leaching simulations show that the pH and ash/water ratio control the LCACs’ abundance and geochemical composition of the impacted water. These results have important implications for the prediction of the fate and migration of LCACs in the environment, particularly for the storage of CCRs in holding ponds and landfills, and any potential CCRs effluents leakage into lakes, rivers, and other aquatic systems.

1.4.3 Chapter 4 Synopsis: The Impact of Coal Combustion Residual Effluent on Water Resources: A North Carolina Case Study

Chapter 4 presents the findings from the investigation of the quality of CCR effluents and their receiving waters. This study systematically documents the quality of effluents discharged from CCR settling ponds at nine sites and the impact of associated waterways (lakes, rivers) in

North Carolina as compared to a reference lake. We measured the concentrations of major and trace elements in over 300 samples from CCR effluents, surface water from lakes and rivers at different downstream and upstream (background) points, and pore water extracted from lake 9

sediments. The study is based on a one-year detailed investigation of two lakes (Hyco and Mayo) and a single sampling event for nine other sites, including a reference site. The data show that

CCR effluents contain high levels of contaminants that in several cases exceed the US

Environmental Protection Agency guidelines for both drinking water and ecological effects. This investigation demonstrates the quality of receiving waters in North Carolina depends on 1) the relationship between levels of contaminants in the CCR effluent and the ratio between effluent flux and freshwater resource volumes (i.e., the dilution effect); and 2) recycling of trace elements through adsorption on suspended particles and release to deep surface water or pore water in bottom sediments during periods of thermal water stratification and induced anoxic conditions.

The impact of CCRs is therefore accumulative, which influences metal accumulation and the health of aquatic life in water associated with coal fired power plants in North Carolina.

1.4.4 Chapter 5: Boron and Strontium Isotopic Characterization of Coal Combustion Residuals

The fifth chapter combines research from the laboratory with field investigations to investigate the use of boron and strontium isotopes as CCR tracers. Boron and strontium have been identified as good indicators of CCRs in leaching tests and in the environment, but only a few studies have addressed the isotopic compositions of CCRs. The objectives of this investigation were to integrate the geochemical and isotopic signatures of CCRs from a variety of coal sources, and to understand their environmental applicability. Eight coal-fired power or steam plants were sampled for CCRs and each was subjected to leaching experiments under varying conditions. The original coal sources have distinctive strontium isotopic (87Sr/86Sr) ratios in the

Appalachian (0.7098-0.7111), Illinois (0.7114-0.7125), and Powder River (0.7122) Basins (with slight overlap of the Powder River Basin and the Illinois Basin), while the δ11B values were nearly all depleted (-18 ‰ to -4 ‰)(with one exception the, Illinois Basin coals ranged from -7

‰ to +8 ‰). Post-combustion treatment (FGD) appeared to change the strontium isotopic composition of CCRs. Leaching experiments with variable pH values resulted in varying 87Sr/86Sr ratios, and a slight variation in δ11B values (but all were negative). Environmental sampling took place at the 2008 TVA coal ash spill site in Kingston, TN, as well as at ten CCR effluent discharges. Boron and strontium isotopic analyses were performed on surface water as well as the interstitial porewater. The environmental sampling revealed elevated concentrations of Sr and B from CCR effluents and receiving waters, as well as the distinctive boron and strontium isotopic signatures from the CCRs were traceable in the environment. The combination of the geochemical indicators from the laboratory experiments and field sampling (surface and pore water) provided a unique and practical identification method for evaluating the impact of CCRs on the environment.

2. A Survey of the Potential Environmental and Health Impacts in the Immediate Aftermath of the Coal Ash Spill in Kingston, Tennessee

(As published in Environmental Science & Technology; 2009; Volume 43, 6326–6333)

(Supplemental Material in Appendix A)

2.1 Introduction

On Monday, December 22, 2008, the containment structure surrounding the storage of coal ash at the Kingston coal-burning power plant of Tennessee Valley Authority (TVA) collapsed, which resulted in massive release of coal combustion products (CCP) ash to the environment near Harriman, Tennessee. The CCP material, consisting of fly ash and bottom ash, spilled into tributaries of the Emory River and directly into the Emory River (Figure 3), which joins the Clinch River that flows to the Tennessee River, a major drinking water source for downstream users. The Kingston coal ash spill released over 4.1 million cubic meters of ash, which is one of the largest spills in U.S. history. Some previous coal ash spills in the USA include the 2000 Martin County spill in Kentucky, which released over 1.1 million cubic meters of coal slurry into abandoned mines and nearby creeks [48] , and the 1972 Buffalo Creek incident in

West Virginia, which released almost a half million cubic meters of coal mining residue slurry into a nearby town [49] . Numerous studies have shown that coal ash contains high levels of toxic metals that can harm the environment [6, 9, 10, 34, 36, 50-55] and some of these elements are soluble in water and easily leached in aquatic systems [29, 36, 55, 56] .

This paper aims to provide on initial assessment of the potential environmental impacts and health risks associated with the Kingston TVA coal ash spill. In particular, the paper examines the reactivity of trace metals known to be enriched in CCP ash [36, 52] in surface water and the potential ecological effects associated with the accumulation of CCP ash in river sediments. Furthermore, the paper emphasizes the relatively high content of radionuclides in CCP

ash and the potential health impact of their resuspension in the atmosphere. While most studies have investigated the potential for radon emanation from cement and fly ash used as building materials [57-60] , here we examine possible health risks associated with elevated radium activity in CCP ash. The study includes measurements of trace metals in solid ash, sediments from the river, and water samples that were collected in the vicinity of the coal ash spill. Given the limited data collection since the accident, this paper provides only an initial evaluation, and does not provide a comprehensive assessment of the overall environmental impacts of the TVA coal ash spill.

Figure 3: Map of the sampling sites of the TVA coal ash spill in Kingston, Tennessee. Site descriptions are reported in Supporting Information (Google maps provided the base map).

2.2 Analytical Methods

Coal ash, sediments from the rivers, and water samples from tributaries, the Emory and

Clinch Rivers, and springs near the spill area in Kingston and Harriman, TN (Table S1; Figure 3)

were collected in three fieldtrips on January 9-10, February 6-7, 2009, and March 27-28, 2009.

The surface water samples were collected near the river shoreline from sites located upstream and downstream (at different distances) of the spill. Each location was determined by availability of public access and/or allowance by property owners. Water sampling strictly followed USGS protocol [61] ; trace metal and cation samples were filtered in the field (0.45µm syringe filters) into new and acid-washed polyethylene bottles containing high-purity HNO3. Trace metals in water were measured by inductively coupled plasma mass spectrometry (ICP-MS); mercury in sediments and coal ash by thermal decomposition, amalgamation, atomic absorption spectroscopy

(Milestone DMA-80) [62] ; and radium isotopes by g-spectrometry (see supporting text S1).

2.3 Results and Discussion

2.3.1 Coal ash and sediments

A comparison of the chemical composition of the TVA coal ash and local soil in

Kingston, Tennessee (Table 1) shows marginal enrichments of major elements of calcium (by a factor of 2), magnesium (1.3), and aluminum (1.5) and large enrichments of trace elements such as strontium (30), arsenic (21), barium (5), nickel (5), lithium (5), vanadium (4), copper (3), and chromium (2). The high arsenic concentration in the TVA coal ash (mean=75 mg/kg) is consistent with previously reported As in ash residue of both hard coal (anthracites, bituminous, and subbituminous A and B; As=50 mg/kg) and brown coal (lignites and subbituminous C;

As=49 mg/kg) [9] . In addition, the mercury level of the TVA coal ash (an average of 151.3±15.9 mg/kg; Table 2) is higher than background soil in Tennessee (45 mg/kg). These concentrations are consistent with the range of values reported in fly ash (100 to 1500 mg/kg) [63] . Likewise, the 226Ra (a mean of 4.4±1.0 pCi/g) and 228Ra (3.1±0.4 pCi/g) activities of the coal ash are higher than those in local soil in Kingston (1.1±0.2 and 1.4±0.4 pCi/g, respectively; Table 3). The Ra activity of the TVA coal ash is similar to the levels reported previously for fly and bottom ash 14

from a Kentucky utility (Table 3) with a consistent activity 228Ra/226Ra ratio of ~0.7 [57] . The potential impact of Ra on the environment and human health is an important consideration in remediation of the spill and is discussed below.

The Hg concentration increases from 16-54 mg/kg in upstream sediments, collected at the shoreline of the Emory and Clinch Rivers, to a level of 53 mg/kg directly across from the spill site (Site 8; Figure 3), and up to 92-130 mg/kg in sediments from the downstream Clinch River at

Sites 9 and 10 (Tables 2 and S1). The Hg concentrations of the upstream sediments are consistent with previously reported Hg data for the overall Tennessee River (Table 2) [61] . A historical massive release of Hg from Oak Ridge Y-12 plant into East Fork Poplar Creek has resulted in accumulation of Hg in the sediments from the downstream Clinch River [64, 65] , which could have provided a Hg legacy source for the Clinch River sediments. Our direct sampling of two sites in the upstream Clinch River (relative to the coal ash spill; Table 2) and downstream of the

Y-12 source in Oak Ridge resulted however low Hg contents of the river sediments (16 ± 5 upstream and 54 ± 11 mg/kg downstream to Poplar Creek), which are similar to background values we report for the Emory River (Table 2). In contrast, sediments from the downstream

Clinch River (sites 9 and 10) have higher Hg content (>100 mg/kg), which suggests a significant contribution of Hg from the coal ash to the river sediments. We therefore conclude that ash transport and deposition in the Clinch River has increased the Hg content in the river sediments.

Table 2: Hg results (mg/kg) in coal ash and river sediments associated with the spill area in Kingston, TN. Background data of Hg in Tennessee soil from the USGS Open Geochemical Database.

2.3.2 Water Contamination

Results show that the tributary that was dammed by the coal ash spill and turned into a standing pond (“the Cove” in the area of Swan Pond Circle Road; Figure 3) has relatively high levels of leachable coal ash contaminants (LCAC), including arsenic, calcium, magnesium, aluminum, strontium, manganese, lithium, and boron (Table 4; Figure 4). Some of these elements are highly enriched in coal ash [34, 51] (Table 1), and are known to be highly soluble in aquatic systems [51] . Among the LCAC’s, arsenic stands out with concentrations of up to 86 mg/L in the

Cove area. Groundwater data from the other tributary (Figure 3) show negligible LCAC levels, thus indicating that the shallow groundwater is not contaminated. In this hydrological setting, non-contaminated groundwater discharges into the dammed tributary and causes leaching of

LCAC from the coal ash. Under restricted water exchange, the formation of standing water in the

Cove resulted in contaminated surface water. In contrast, surface waters from the Emory River and Emory- Clinch River downstream from the breached dam show only slight LCAC levels, and all river inorganic dissolved constituents concentrations are below the EPA-Maximum

Contaminant Levels (MCL) and EPA-Criterion Continuous Concentration (CCC) for aquatic life

(Table 4) [66, 67] .

The upstream Clinch River has a distinct chemical composition (higher Na+, Ca2+, Mg2+,

2+ 2- Sr , and SO4 ) relative to the upstream Emory (Table 4), and thus the major inorganic constituents show mixing relationships between these two water sources (Line M1 in Figure 4) downstream from the confluence of the Emory and Clinch rivers (Figure 3). The concentrations of arsenic, strontium, and boron in the downstream river samples deviate towards higher values, however, relative to these river mixing relationships (Figure 4). These geochemical shifts reflect a small but traceable indication of leaching of contaminants from coal ash that was spilled into the river and further mixing with the uncontaminated river water (Line M2 in Figure 4). The data show that the river flow is effective in reducing the LCAC’s contents by an order of magnitude relative to directly contaminated water measured in the Cove.

The water samples were collected near the shoreline of the river, which may be an underestimate of the concentration of dissolved elements throughout the river vertical profile. The spatial distribution of contaminants (dissolved and suspended particulate fractions) in a river water column depends on a number of factors including particulate size, turbulent flow conditions, seasonal flow changes, and channel morphology [68] . The dissolved phase, which was measured in this study, is typically homogeneously distributed in a river, but can vary with the proximity to a point source of pollution [68] . Assuming that metal mobilization in the river derives from both suspended ash and bottom sediments, the distribution of dissolved constituents in the water column would depend on numerous factors such as differential river velocity, rate of mobilization, and water depth [69] . Further investigation is therefore required to determine the vertical distribution of metals in the river water, and whether sampling of the upper river section represents the most diluted segment of the river flow.

2.3.3 Potential Environmental Impacts

While the downstream river water shows only trace levels of LCAC’s (at the surface), the downstream river sediments show high Hg concentrations similar to the coal ash levels (115-130 mg/kg; Table 2). The ecological effects of Hg in the coal ash and sediments depend on the chemical ability of Hg in the solids and the potential for mercury methylation in the impacted area. While previous studies have demonstrated that Hg in CCP ash is not readily soluble through acid-leaching protocols [70] , Hg has a high affinity for natural organic matter [71, 72] , which can promote desorption if the Hg is associated with weaker binding sites on metal-oxide minerals in the ash material [73] . Furthermore, the transformation of Hg to methylmercury by anaerobic

bacteria in river sediments is a concern because of bioaccumulation of methylmercury in food webs. Previous studies have shown that sulfate addition can promote methylation in freshwater ecosystems by stimulating sulfate-reducing bacteria [74] , the primary organisms responsible for producing methylmercury in the environment [75] . In coal ash-containing waters, a 10- to 20-

2- fold increase in SO4 concentrations was observed in the Cove area relative to unaffected upstream sites (Table 4). Therefore, the methylation potential of mercury from this material could

2- be high because the coal ash also provides an essential nutrient (SO4 ) that encourages microbial methylation. In addition, accumulation of arsenic-rich fly ash in bottom sediment in an aquatic system could cause fish poisoning via both feed chains and decrease of benthic fauna that is a vital food source [9, 76] .

2.3.4 Potential Health Impacts

Of particular concern to human health is the wind-blown re-suspension of fly ash into the atmosphere. It is well known that wind-blown dust can travel long distances, as exemplified by

Asian dust storms that result in transport to locations as far away as the U.S. [77] . It is possible that coal ash exposed to the atmosphere can be resuspended and transported to populated areas where human exposure may occur. Fly ash–airborne particles with diameters less than10 µm

(PM10) are regarded as respirable and may affect the human lung and bronchus [78-80] . The process of particulate re-suspension will depend on a variety of factors, including the fly ash particulate size and related chemical and physical properties, wind speed and atmospheric turbulence, and likely the relative humidity and surface moisture [81, 82] . The particles that are of most importance for human health are in the fine particulate (PM2.5) mode, which readily deposit deep in the lung [83] . Past work has shown that CCP ash have particulate sizes ranging from less than 1 mm to 100’s of microns in size [84, 85] . In addition, there is a compositional relationship as a function of fly ash particle size [84] . Several studies have also measured ambient fine particulate matter associated with elevated concentrations of toxic metals in the vicinity of coal-fired power plants [84, 86-88] . In some cases, fly ash–airborne particles were also found in remote areas (up to 30 km from power stations) [89] . Overall, past work indicates that coal ash contains inhalable particulate matter, and that fly ash emitted from the burning of coal is readily transported in the atmosphere.

The high concentrations of trace metals (Tables 1 and 2) and radioactivity (Table 3) reported in this study for the bulk TVA coal ash are expected to magnify, as fine fractions of fly ash, which may be resuspended and deposited in the human respiratory system, are typically 4–10 times enriched in metals relative to the bulk ash and the coarse size fraction [84] . The toxic metal

content in coal ash, the sizes of fly ash particulates, and the ionizing radiation (IR) exposure

(both, incorporated and external) may act synergistically or, less frequent, antagonistically, affecting human health directly (predominantly through inhalation of contaminated air) and indirectly through the food chains (consuming contaminated agricultural products) [55] . Coal ash was recognized as a Group I human carcinogen (based on occupational exposure studies) associated with increased risks of skin, lung, and bladder cancers [90] . Arsenic and radium exposures in humans are associated with increased risks of skin, lung, liver, leukemia, breast, bladder, and bone cancers [91] for exposure predominantly due to chronic ingestion or chronic inhalation, with the dose-response curve dependent on location, sources, and population susceptibility and/or tolerance.

Health impacts of CCP ash have been predominantly studied on animal models and human cell lines, with few short-term epidemiological follow-ups. CCP ash particulates affect lung epithelial and red blood cells in animal studies and human in vitro models, causing inflammation, changing the sensitivity of epithelia, altering immunological mechanisms and lymphocyte blastogenesis, and increasing the risk of cardiopulmonary disease (e.g., pulmonary vasculitis/hypertention) [92-95] . Individuals with pre-existing chronic obstructive pulmonary disease, lung infection, or asthma are more susceptible to the coal ash affliction [96] . Several epidemiological studies have proved the significant health hazards (such as enhanced risk for adverse cardiovascular events) of fine-particulate air pollution for individuals with type II diabetes mellitus and people with genetic and/or disease-related susceptibility to vascular dysfunction, who are a large part of the population [43] .

Radium-226 and 228Ra, which are the main sources of low-dose IR exposure in coal ash, can remain in the human lung for several months after their inhalation, gradually entering the blood circulation and depositing in bones and teeth with this portion remaining for the lifetime of

the individual. When inhaled, the radionuclides can affect the respiratory system even without the presence of the other coal ash components. Thus, the airborne particles containing radioactive elements inhaled by clean-up workers of the nuclear accident at the Chernobyl nuclear power plant caused bronchial mucosa lesions, in some cases preneoplastic, with an increased susceptibility to the invasion of microorganisms in bronchial mucosa [97, 98] . Consequently, the combined radioactivity of coal ash at the TVA spill, together with other enriched trace metals such as Ni, Pb, and As, may increase the overall health impact in exposed populations, depending on duration of exposure, and particularly for susceptible groups of the population. It is important to underscore the fact that at this time it is not possible to estimate the health impacts of CCP ash resuspended particulates due the a lack of information on the rate at which they are entrained into the atmosphere, as well as their chemical, physical, and synergistic properties linked to morbidity and mortality. Clearly, future studies are needed linking ambient element and radionuclide concentrations with ground level CCP ash characteristics, ambient meteorological characteristics, and human population exposure.

This study has provided an initial assessment of the environmental impacts and potential health effects associated with the TVA coal ash spill in Kingston, Tennessee. The study shows that the high metals contents of coal ash and their high solubility resulted in contamination (e.g.,

As) of surface water associated with the coal ash spill in areas of restricted water exchange. In the downstream Emory and Clinch Rivers the leaching of trace metals is significantly diluted by the river flows. While the levels of contaminants in the downstream Emory and Clinch Rivers are below the MCL levels, high concentrations of Hg found in the river sediments pose a serious long-term threat for the ecological system of these rivers. This study also highlights the high probability of atmospheric re-suspension of fine fly ash particulates, which are enriched in toxic metals and radioactivity, and could have a severe health impact on local communities and

workers. Based on these initial results, this study provides a framework for future and long-term monitoring of the TVA coal ash spill and remediation efforts. Future studies should focus on evaluating the ecological ramifications, such as methylmercury formation in the sediments in the downstream Emory and Clinch Rivers, and the composition of particulate matter in the air in the vicinity of the spill area. Finally, future prognoses of the health impacts of residents exposed to coal ash requires long-term follow-ups of various population groups, including children and adolescents, pregnant women, persons exposed in utero, individuals with pre-existing broncho- pulmonary diseases and diabetes mellitus. All these factors must be included in remediation efforts for the TVA Kingston coal ash spill.

3. The Environmental Impacts of the Coal Ash Spill in Kingston, Tennessee: An Eighteen-Month Survey

(As Published in Environmental Science & Technology; 2010; Volume 44, Issue 24, 9272-9278)

(Supplementary material is in Appendix B)

3.1 Introduction

On December 22, 2008, the Kingston coal-fired power plant of Tennessee Valley

Authority (TVA) had a containment structure rupture, spilling over 3.7 million cubic meters of wet coal ash (fly ash with intermixed bottom ash), inundating the Emory River, its tributaries, and the adjacent landscape near Harriman, TN [3] . The wet coal ash spilled into the Emory River, which joins the Clinch River, and then converges with the Tennessee River (Figure 5), a major drinking water source for populations downstream [4] . Previous investigation has shown that the concentration of some leachable coal ash contaminants (LCACs) such as boron, arsenic, strontium, and barium increased slightly downstream of the spill relative to the upstream river concentrations, but remained below the EPA’s maximum contaminant level (MCL)[66] and the

Continuous Criterion Concentration (CCC)[67] for aquatic life. However, areas of restricted water exchange (i.e., the “Cove”; Figure 5), where water remained stagnant and in direct contact with coal combustion residues (CCRs), had high levels of LCACs [4, 99] .

The high concentrations of LCACs in CCRs [6, 9, 34, 36, 50-54, 100, 101] and their enhanced mobility in aquatic systems [29, 55, 56] are the key factors for evaluating potential risks of CCRs to the environment. The TVA spill has provided a unique opportunity to examine these effects on a regional field scale, beyond laboratory leaching tests [3, 102-106] . The objectives of this study are (1) to provide a systematic eighteen-month monitoring survey of the environmental impacts following the TVA coal ash spill in Kingston, Tennessee; (2) to examine the composition of major constituents and trace metals in water samples from different sites 26

associated with the ash spill, including upstream and downstream river waters, tributary waters, and porewater extracted from the river sediments; (3) to conduct laboratory leaching experiments on the TVA coal ash for evaluating the factors that control LCACs composition and mobilization; (4) to examine the possible river quality impacts upon dredging of the ash from the rivers, which was part of the major remediation plan by TVA [3] ; and (5) to evaluate the impact of redox conditions that prevail in river sediments on the reactivity of coal ash in the environment, particularly for arsenic mobilization and species distribution.

Figure 5: Map of spill site of the TVA Kingston Fossil Plant. Emory and Clinch River mile markers (ERM and CRM) are labeled. The possible extent of the spilled ash upstream continued to CRM 5 and ERM 6 due to the magnitude of the spill’s initial force and altered

hydrodynamics of the power plant water intake, and continued further downstream than shown on this map (due to subsequent storm events) (base map provided by Google maps).

3.2 Analytical Methods

Twelve field trips were made to the spill site between January 2009 and June 2010 with over 220 surface and porewater samples collected during high (winter) and low (summer-fall) river flow regimes (Figure 6). Water sampling strictly followed USGS protocols [107] ; after filtration of samples in the field (0.45 µm syringe filters), trace elements were measured by inductively coupled plasma mass spectrometry (ICP-MS), major elements by direct current plasma optical emission spectrometry (DCP-OES), and anions by ion chromatography (IC). Pore waters were extracted from river bottom sediments obtained using a Wildco box core (up to 25 cm depth) and VibeCoreTM (up to 182 cm depth). Inorganic arsenic species were measured using the Bednar method [108] in which the uncharged arsenic species AsIII was separated from porewater and preserved in the field. Acid volatile sulfide [97] [109, 110] was quantified in whole sediment samples from the box core. Laboratory-leaching experiments were conducted on a bulk coal ash sample that was collected from the Cove area (Figure 5) soon after the spill.

Leaching experiments were simulated reactions of the spilled TVA ash collected in the Cove area

(Figure 5) with solutions exhibiting a wide range of acidity (pH of 0.4 to 12) and ash/water ratios

(1x10-5 to 3.5x10-3).

Figure 6: Variations of flow rate (cubic meters per second; red line, left-hand side axis) in the upstream Emory Rive USGS gauge at Oakdale, TN[111] and arsenic concentrations in downstream river water (black circles, right-hand side axis). Note that the low typical river discharges in the summer were associated with higher As concentrations in the downstream river water. Also note the timing of the spill event (black square) and dredging activities (arrow).

3.3 Results and Discussion

3.3.1 Laboratory TVA coal ash leaching

Two types of leaching experiments on the TVA coal ash were conducted: (1) leaching under different pH conditions; and (2) reaction with deionized water and an upstream Emory

River water under different ash/water ratios (referred to as total suspended solids, TSS). Leaching 29

simulations under a wide range of pH conditions resulted in differential LCAC leaching. As, B,

Sr, and Se showed maximum dissolved concentrations in leachates under extreme acidic (pH=0.4 to 1.7) conditions. As and Se, and to much lesser extent B, reached minimum concentrations at pH values between 4.5 and 7, and then increased in leachate concentration at pH 12. Sr and Ca showed progressively lower concentrations in leachates with increasing pH. These results are consistent with previous laboratory leaching experiments [102-106, 112, 113] . The differential

LCACs mobilization generated large variations in the contaminants’ ratios (e.g., As/B, Se/As;

Figure 7).

The second type of experiments showed that the TVA ash had low buffering capacity.

The reaction of the ash with an upstream Emory river sample did not change the pH of the effluents, even at high ash/water ratios (Figure 7). This non-alkaline characteristic of the TVA coal ash is different from other CCRs reported elsewhere [102-106] . Sensitivity tests for tracing

LCACs upon reacting TVA ash with deionized water and upstream river water showed that As,

Sr, Ba, and to a lesser extent B were traceable even at low solid/water conditions. This result also reflects the sensitivity of the analytical method. The concentrations of these elements were linearly correlated with ash/water ratios up to 1x10-3 (TSS =1000 mg/L). For As, the empirical linear slopes for the variations of As concentrations with TSS (at range <1000 mg/L) were 4x10-3 and 1x10-2 mg As per mg TSS for river water and deionized water, respectively (Figure 7). The results also show that As concentrations above the MCL of 10 mg/L could be obtained above TSS values of 3000 and 1000 mg/L for river water and deionized water, respectively (Figure 7).

Figure 7: Results of leaching experiments of TVA coal ash under different ash/water ratios (upper panel) and pH conditions (lower panel). Variations of solid to water ratios, which are also expressed as total suspended solids[114] , show different pH effects and As concentrations in the deionized water (checkered symbols, upper panel) and upstream Emory River (blue circles) leaching experiments. Slopes for the linear correlations up to TSS=1000 mg/L are marked. pH variations show differential mobilization of LCACs in the experimental lechates (checkered symbols, lower panel), as compared to field measurements of pore water (black squares) and Cove water (red circles).

3.3.2 Surface Water

Surface water. A small tributary that was dammed by the ash spill, creating a standing pond (the “Cove”; Figure 5), initially had relatively high concentrations of LCACs above the background levels measured in the upstream Emory and Clinch Rivers (Table 5) [4] . During the physical removal of the ash, while inflowing water was temporally diverted around the Cove area, the concentration of LCACs decreased (e.g., As 86 µg/L in Feb 2009 to 5µg/L in Nov 2009). The concentrations of LCACs in the downstream river were significantly lower than those found in the Cove (Table 5), but nonetheless higher than levels in the upstream Emory and Clinch Rivers for each sampling date. During the 18-month survey, the As concentrations measured in the downstream river sites were all below the EPA’s MCL and CCC thresholds (Figure 6).

The spatial and temporal distributions of As concentrations in the river surface water

(Figure 8) showed some distinctive peaks during June, August, and September 2009 and again in

April and June 2010 at sites located at and downstream of the spill region (ERM 3-ERM 1). In contrast, winter sampling periods resulted in low As concentrations, likely due to greater dilution by higher winter flows (Figure 6). Between March 20, 2009 and May 2010 over two million cubic meters of ash were dredged out from the river as part of the TVA remediation activities

[115] . The seasonal dependence of As concentrations in the downstream river water (Figure 8), combined with relatively constant and low As contents during the major dredging activities (fall and winter 2010), indicates that ash dredging had minimal, if any impacts on the river surface water quality. This observation is consistent with our lab experiments (Figure 7), in which TVA ash leaching under TSS conditions similar to those measured in the Clinch and Emory Rivers (20 to 100 mg/L; [116, 117] caused only a minor As contribution (0.5-1 mg/L).

3.3.3 Sediment Acid Volatile Sulfide [97]

AVS was detected in the surface bottom river sediments (composite of the top 20-25 cm) in all sites, with concentrations ranging from 0.02 mmol g-1 to 1.5 mmol g-1 (dry weight basis).

No seasonal or spatial trends were observed over the course of sampling.

3.3.4 Pore Water

The huge quantity of released ash inundated many parts of the river and was found at significant distances upstream and downstream of the spill. The spill occurred at approximately

Emory River mile 2.5 (ERM 2.5), and following redistribution by subsequent rain events, TVA investigations indicated the extent of ash from ERM 0.0 up to about ERM 6 (9.6 km), from

Clinch River mile (CRM) 0.0 to about CRM 5 (8 km), and from Tennessee River Mile (TRM)

561.8 to TRM 568.7 with traces extending about 3.2 km downstream beyond TRM 561.8 [118]

(Figure 5). The ash was distributed in the river either by the large force of the initial spill flow, the hydrodynamics of the power plant’s water intake changing direction of river flow near the confluence of the Emory and Clinch Rivers, and subsequent storm events [119] . The spill region

(ERM 3.0 to ERM 1.5) was characterized by deep deposits of ash, which formed the matrix for bottom sediment in this stretch of the river. Since the initial spill event, the ash has been washed further downstream, mixed with native sediment, and buried by fresh deposits of native sediments. Pore waters extracted from buried ash and native river sediment mixtures were found to have neutral-pH (a range of 6.6 to 7.4) and high concentrations of B, Sr, Ba, and As relative to the upstream surface and pore waters (Table 5). The As concentrations (up to 2011 µg/L; mean=323.8±494 µg/L) far exceeded the EPA’s arsenic MCL (10 µg/L) and CCC (150 µg/L) for aquatic life. In contrast, concentrations of Se, Pb, Cr, and U in the pore waters were low (Table

5).

The data show that pore water extracted from deep coring (sediment depth of 0-100 cm and 100-175 cm) had significantly higher LCACs concentrations relative to shallow pore water (depth of 25 cm) of the same sites (ERM 2 and CRM 4), suggesting that the deep pore water was a better representation of the internal sediment process, while the shallow pore water was likely affected by dilution from the overlying river water and/or oxidation during sampling.

Nonetheless, high LCAC levels were also observed in the shallow pore waters (Table 5). Boron and strontium have been identified as good indicators for CCR leachates [4] [56, 120, 121] , which is also confirmed by the leaching experiments performed in this study. A similar range of

B/Sr ratios were also observed in the artificial leaching experiments (0.55 to 0.85 at neutral pH), the surface water region of limited water exchange (the Cove; 0.46-1:00), and pore water (0.5-

1.5). In contrast, the Se/As ratios show large variations, particularly for porewater with Se/As ratios (2x10-5 to 0.10) lower than those of the Cove (0.01-0.53) and experimental leachates at neutral pH (~0.4) (Figures 7 and 9). Furthermore, As was enriched in the pore waters (As/B ratios

=0.1 to 2.1) relative to the Cove waters (0.02-0.22) and neutral-pH leachates (~0.6) (Figure 7). In summation, we show that surface water from the Cove area had higher Se concentrations and

Se/As ratios relative to the pore water, while the pore water had significantly higher As concentrations for similar B or Sr concentrations (i.e., higher As/B ratios; Figure 9).

3.3.5 The Control of pH, redox state, and ash/water ratio on contaminant mobilization

The sampling procedure of the pore water involved an exposure of the sediments and pore waters to the atmosphere. Thus, direct measurements of the redox state (e.g., redox potential and dissolved oxygen) of the pore water were not possible. However, we use the following observations to postulate that the river bottom sediments had some degree of a reducing state: (1) sulfate contents in the pore water (normalized to chloride; Figure 10) were significantly lower than that of the Cove water, suggesting the occurrence of sulfate reduction processes in the pore water; (2) the presence of acid volatile sulfide in the river sediments from which pore waters were extracted (0.01- to 0.9 µmol/g); (3) high concentrations of manganese and iron in pore waters

relative to downstream river water (Table 1); and (4) direct measurement of the reduced form of arsenic (AsIII) in the pore water (Figure 10).

Numerous studies have determined that the predominant species of arsenic and selenium in CCRs are the oxidized form arsenate [80] and the reduced form selenite (SeIV), respectively [122-125] . Based on the neutral-pH (6.6 to 7.4) and arsenic speciation

0 measurements (Figure 10), we deduced that the uncharged H3AsO3 species was the predominant species in the pore water. While the overall higher concentrations of As, B, and Sr in the pore water relative to the Cove water could reflect higher ash/water ratios, the differential enrichment of As in the pore water relative to the Cove water (Figures 9 and 10) could be due to the combination of (1) reductive dissolution of hydrous ferric oxides (HFO) and Mn-oxides (as evidenced by the high Mn and Fe contents in the pore water; Table 1) that could further release

0 V adsorbed As; and (2) decreased sorption of the uncharged H3AsO3 species (compared to As ) to the host sediments, including oxides and sulfides [126, 127] . Furthermore, we show that the relative As enrichment (i.e., an increase in As/B ratios) is associated with sulfate depletion in the pore water (lower SO4/Cl ratios; Figure 10), which suggests that the reducing conditions were sufficient to mobilize As but not as advanced to the stage for precipitation of arsenic sulfides

[127, 128] .

Conversely, the reduced Se species (SeIV) that was leached from CCRs was apparently preserved in the reducing conditions of the bottom sediments. Given the near neutral-

- pH of the pore water, we hypothesize that the predominant selenium species was HSeO3 , which is known to have a strong sorption affinity for both HFO [126, 129] and clay minerals [130] that would result in lower Se concentrations in the pore waters. In oxidizing conditions, similar to the conditions in the restricted tributary (the Cove), SeIV that is leached from CCRs could be oxidized to SeVI, which is less reactive towards sorption [126, 129] . Therefore oxidation of selenium could

result in relatively higher Se concentrations in surface waters; however, the kinetics of this transformation is rather slow [131] . The apparent reduction of sulfate (Figure 10) is also associated with higher Ba/B ratios in the pore water relative to the Cove surface water (Figure 9), thus suggesting higher Ba mobility in the pore water. One explanation for this enrichment could be the saturation state (SI) of the mineral barite (BaSO4), which would be significantly lower in the low-sulfate pore waters (SIbarite=2.9) relative to high-sulfate surface waters of the Cove

(SIbarite=3.5).

The low Se/As and high As/B and Ba/B ratios of the pore waters were distinctively different from those of the background upstream river water, while the downstream river samples had intermediate values (Figure 9). This apparent mixing relationship may indicate that the concentration of these elements in the downstream river water was not derived primarily from in- situ interaction with suspended ash in the river but rather from subsurface LCAC-rich solutions that emerged from the river bottom. The level of LCAC concentrations in the river water depended therefore on the mass-balance between upstream river water and subsurface solutions rather than leaching from suspended ash in the river. A mass-balance calculation indicates that an increase from a background As concentration of 1 mg/L to a “contaminated” value of 5 mg/L in the downstream river would result from a contribution of only 0.2% of pore water with As concentration of 2000 mg/L into the overlying river water column. The association of high As concentrations in the downstream river during summer sampling (Figures 8) could be the result of lower flow rates in the summer (Figure 6) and a higher contribution of the subsurface and

LCACs-rich solutions. This conclusion is consistent with our leaching experiments (Figure 7) that show negligible As concentrations in effluents containing TVA ash in an amount that is equivalent to the typical TSS values in the Clinch and Emory Rivers [116, 117] .

3.3.6 Implications for tracing and prediction of CCR contaminants

The data presented in this study indicate that the massive remediation efforts of TVA by dredging and removing over two million cubic meters of coal ash from the Emory and Clinch

Rivers had only a minor effect on the river surface water quality because of the large dilution and the low ash/water ratios, as demonstrated in our leaching experiments (Figure 7). Our data show that buried TVA ash that accumulated within the river bottom sediments with limited exchange with the surface river flow is highly reactive and generated high levels of dissolved As, B, Sr, and

Ba in the associated pore waters. While the presence of these elements in high concentrations in the pore water presents a potential direct threat to infaunal species that live in the subsurface, particularly due to the high toxicity of the AsIII species [9, 132] , the abundance of these tracers in pore waters might be used to assess the distribution and extent of buried ash in different locations downstream of the spill area in Kingston, TN.

The results of this study indicate that CCR leaching would be expected to vary for different conditions, particularly ash/water ratio, pH, and redox conditions. Oxygenated surface waters in contact with coal ash are expected to be enriched in many LCACs, including As and Se, whereas reducing subsurface waters (groundwater and pore waters in landfills and river sediments) are expected to be more enriched in As and depleted in Se. In contrast, the redox conditions have no apparent impact on the differential mobilization of B and Sr, which makes those elements good tracers for CCR leaching to aquatic systems. Effects of pH vary depending on whether the predominant form of an element in natural waters is an oxyanion or hydrated cation; elements such as B, As, and Se that occur as oxyanions exhibit higher solubility, and thus greater leachability under low and high pH conditions, while the solubility of most cationic species (e.g., Ca and Sr) decreases with increasing pH. The EPA Toxicity Characteristic

Leaching Procedure (TCLP) [133] , which is used to determine whether a material must be

regulated as a hazardous waste, only considers leaching in weakly acidic conditions (pH ~4), and does not consider leaching of contaminants under a wide range of pH conditions [112, 113] , nor possible anaerobic conditions. In the case of coal ash waste, our results indicate that the TCLP test would greatly underestimate leachate concentrations of As for anaerobic disposal conditions, thus would under-estimate the potential impact of coal ash leachate in many situations. Future studies should focus on evaluating the potential ecological ramifications, particularly for infaunal species that would be exposed to AsIII in the shallow pore waters. Finally, future prognoses of the potential environmental hazards of CCRs and possible migration and fate of ash contaminants in water resources should take into account the results of investigations of the TVA coal ash spill in

Kingston, TN.

4. The Impact of Coal Combustion Residual Effluent on Water Resources: A North Carolina Case Study

(Supplemental Material is in Appendix C)

4.1 Introduction

Numerous studies have shown that effluents generated from leaching of Coal Combustion

Residuals (CCRs) typically have high concentrations of toxic elements [120, 134] . Yet, the overall impact of disposed CCR wastes on the quality of water resources in the U.S. is largely unexplored. In the US, approximately 600 power plants [135] generate 136 million tons of

CCRs annually, of which 56% is stored in surface impoundments and landfills, while the remaining are reused for concrete, cement, and construction industries [136] . CCRs, encompassing fly ash, bottom ash, and flue gas desulfurization (FGD) material, represent one the largest industrial waste streams in the U.S., and are not classified as hazardous waste [137] .

Despite the large volume of CCR effluents generated annually and their disposal into hundreds of surface water bodies, the environmental risks associated with these disposal practices are not well known. Moreover, due to the lack of CCR waste data [138] , the effluents that are discharged from coal fired power plants and permitted by the national and state regulatory bodies lack consistent monitoring and limit requirements that are relevant to CCR effluents.

Water in coal-fired power plants is used in steam production and cooling, as well as the transport of CCRs from the plant to CCR holding ponds. In spite of some losses [139] , the residual effluent water is discharged to the environment and is permitted through the National

Pollution Discharge Elimination System (NPDES) Program. The NPDES Program as established by the Clean Water Act requires the control and permitting of point source discharges of wastewater [140] . In principle, the permits issued for NPDES outfalls consist of requirements for discharge, monitoring of the effluent, and limits on contaminants discharged to receiving waters 43

[140] . Despite significant changes in coal combustion practices and technologies, the Federal guidelines concerning effluents discharged from power plants have not been revised since 1982

[138, 141] . Although the NPDES regulations for CCR effluents disposal vary between states, in most cases they consist of only limited factors. For example, in the federal regulations guiding the permit limits on effluent discharge include only total suspended solids and oil and grease [141] , but do not include other constituent limits that could be relevant to CCR effluents, unless written in specifically by the permitting body.

In recent years, air regulations have become more stringent (e.g., Interstate Clean Air Rule and Clean Air Act), requiring the capture of potential atmospheric pollutants, like sulfur oxides

(SOx). The FGD process effectively removes many of the volatile elements associated with the

SOx rather than releasing them to the atmosphere. Most coal-fired power plants (up to 88% in

2010) use a wet FGD disposal system [134] where the volatile constituents are captured by injecting lime or CaCO3 and are then entrained into the wastewater either as dissolved constituents or in particulate form [134] . This process results in cleaner air emissions, but the trade-off is significant enrichments of contaminants in solid wastes and wastewater discharged from power plants. Data from FGD wastewater reveal extremely high levels of certain constituents, specifically boron, manganese, and selenium, which tend to be in dissolved forms in

FGD wastewater with concentrations up to 250,000 µg/L, 5,460 µg/L, and 21,700 µg/L respectively [134] . Other contaminants are also found in high concentrations in the FGD wastewater, like arsenic (5,070 µg/L), antimony (23 µg/L), cadmium (302 µg/L), chromium (350

µg/L), lead (252 µg/L), and mercury (872 µg/L) [134] . Several studies have shown that groundwater near these CCRs disposal facilities was contaminated by CCR leachates [5] . Yet, the long-term exposure and accumulation of contaminants, as well as the impact of a significantly

more enriched waste stream, like FGD, are poorly studied in surface water surrounding coal-fired power plants.

This study aims to investigate the impact of CCR disposal on surface water surrounding coal fired power plants in North Carolina. We systematically document the quality of discharged effluents from ten CCR discharge sites and the impact on associated waterways (lakes, rivers) and from one reference (control) lake in North Carolina (Figure 11; Table 6). We measured the concentrations of major and trace elements in 76 CCR effluent samples, 129 surface water samples from lakes and rivers from different downstream and upstream (background) points, and

98 pore water samples extracted from the lake sediments. The study is based on an investigation of monthly sampling over one year at two lakes (Hyco and Mayo) and a single sampling for nine other waterways (Table 6, Figure 11). The data show that CCR effluents contain high levels of contaminants that in several cases exceed the US Environmental Protection Agency (EPA) guidelines for both drinking water and ecological effects. We also demonstrate that the impact of

CCRs is accumulative, which increases the risks to the health of aquatic life in waterways associated with coal fired power plants in North Carolina and similar sites across the U.S.

4.2 Analytical Methods

A total of thirty-six field trips were made to bodies of water in North Carolina between

August 2010 and February 2012 with over 300 surface and pore water samples collected.

Samples were collected monthly from Hyco and Mayo Lakes from August 2010 through August

2011. The other water bodies sampled include Lake Norman, Mountain Island Lake, Lake Wylie,

High Rock Lake, Belews Lake, Dan River, French Broad River, Lake Julian, and Jordan Lake as a reference lake (Figure 11). The other lakes and rivers were sampled during the summer of 2011,

with the exception of Mountain Island Lake, which was sampled both during the summers of

2010 and 2011. Water sampling strictly followed USGS protocols [107] . Water samples were taken at a variety of depths with a Wildco Niskin Bottle (Type A for trace metals) in order to capture the epilimnion and hypolimnion during stratification. Cations and trace metals were measured in both dissolved and total samples. After filtration of samples in the field (0.45µm syringe filters), trace elements were measured by inductively coupled plasma mass spectrometry

(ICP-MS), major elements by direct current plasma optical emission spectrometry (DCP-OES), and anions by ion chromatography (IC) at Duke University. Pore water was extracted from lake bottom sediments obtained using a Wildco box core (up to 25 cm depth), then vacuum filtration or centrifugation to extract the pore water. The long core was extracted using a peat corer (up to

1m depth). Inorganic arsenic species were measured using the Bednar method [108] in which the uncharged arsenic species AsIII was separated from pore water and preserved in the field. The method was verified by laboratory experiments of known mixed combinations of As species that were processed through an anion exchange resin cartridge.

4.3 Results and Discussion

Several of the NPDES outfall locations discharging CCR effluent from coal-fired power plants in NC (Figures 12 and S1; Table S1) in this investigation exceeded EPA’s water quality standards. Despite the existence of NPDES permits and Effluent Guidelines Monitoring, the limits of contaminants from CCRs in the effluent are currently not set by Federal Regulation. This study documented elevated pollutant concentrations in effluents discharged into receiving waters

(Figures 12 and S1). Dissolved and total samples were collected at each location. They were comparable in concentration of trace elements; therefore the concentrations reported in this paper are dissolved concentrations. For example, during the Summer 2011 sampling, the Roxboro,

Asheville, and Riverbend Plant outfalls contained arsenic concentrations above the NC water

quality standard of 10 µg/L with concentrations reaching 11 µg/L, 44.5 µg/L, and 92 µg/L, respectively. Mayo NPDES discharge yearlong average selenium concentration exceeded the 5

µg/L EPA Chronic Criterion Concentration (CCC) for aquatic life. Some of the individual monthly sampling events at the Mayo NPDES outfall showed Se concentrations almost four times the CCC limit, as high as 19 µg/L. The summer sampling event at the Asheville Plant revealed selenium concentrations over 17 times the CCC (87.2 µg/L). The NPDES outfall for the

Asheville plant also exceeded other human and aquatic life benchmarks, including antimony above the EPA’s MCL (6 µg/L) at 10.9 µg/L, cadmium exceeded the fresh water aquatic life

(EPA CCC) standard (0.25µg/L) at 0.8 µg/L, and thallium concentrations were greater than the 2

µg/L EPA MCL at 2.9 µg/L (Table 7). The outfalls were sampled directly from the pipe or at some sites from water near the outfall where direct sampling was not accessible (Table 7). Thus, the data at some of the outfall sites (Roxboro, Mayo, Marshall, and Allen) underestimate the full extent of the CCR waste stream contaminant level of the discharge. In spite of efforts to reduce the levels of contaminants discharged through the NPDES outfall by using a settling pond, clarifier, bioreactor, or wetland at some sites [142-144] , our data clearly show high contaminant levels that suggest the need for enhanced removal/wastewater treatment.

Figure 12: This plot shows the average outfall concentrations of CCR effluent from plants with an FGD system [145] and without an FGD system. The FGD outfalls tend to have higher concentrations of most constituents, except Li and V.

Table 7: The concentrations of ions in the effluent and the EPA regulatory levels are listed.

Many of the outfalls sampled consisted of wastewater from the FGD process that was subsequently diluted with the ash pond water (and/or other process water) and at some locations mixed with the cooling water (e.g., Roxboro plant at Hyco Lake; Table S2) prior to discharge at the outfall. Therefore for plants with an FGD system, the effluent concentrations represent some dilution of the original FGD wastewaters [142] . The data show that outfalls sampled from coal fired power plants with an FGD system (n=69) have significantly higher concentrations of major ions (Ca, Mg, and Cl; p<0.01) and minor constituents such as B (p<0.01), Br (p<0.01), and Cr

(p<0.05) relative to outfalls with only ash pond water or cooling water disposal (n=5 and n=7 respectively; Figure 12). The plants with no FGD system, but with wet ash disposal systems and subsequent discharges (n=5), had higher concentrations of several constituents including As, V, 51

Sb, Li, Tl and Mo (p<0.01) relative to effluents from FGD systems (Table S1). Selenium concentrations were also higher at FGD outfalls with several plants exceeding the EPA’s CCC of

5 µg/L (Mayo at 19 µg/L and Asheville at 82 µg/L compared to Riverbend at <3.5 µg/L and Dan

River at <1 µg/L selenium). The CCR outfalls were enriched in many constituents compared to the upstream waters that feed them (Figure 13). The FGD had larger enrichments in many ions compared to the ash discharge only outfalls (Figure 12).

outfalls. The EPA water quality criteria for boron, arsenic, and selenium are referenced in the graphs, and several outfalls exceed the limits.

In contrast to the CCR outfalls, separated cooling water effluents that were sampled in this study had much lower metal concentrations, did not exceed any of the human or aquatic life benchmarks (Table S1), nor were enriched in any constituents compared to their respective upstream waters and reference lake (Jordan Lake) (Table S1). Consequently, in outfall sites where CCR effluents and cooling water were blended the contaminant level was significantly reduced. For example, in Hyco lake, where the cooling water constitutes >98% of the effluent volume (Table S2) the contaminant levels of the NPDES outfall would be significantly higher if the cooling component was reduced or restricted (e.g. recirculating cooling water at Mayo Lake).

The direct effluents from the FGD process and ash ponds at Roxboro were reported to have concentrations of As ranging between 1.6-394 µg/L and Se ranging 4.3-238µg/L during the yearlong sampling (Table S2) [144] . Therefore, cooling water has an important mitigating effect on the quality of NPDES outfalls in NC.

We further analyzed the impact of CCR effluents on the quality of receiving waters by systematically comparing the chemical composition in waters downstream of the disposal sites relative to upstream waters from the same river/lake and a reference lake that has no connection to coal plant discharge (B. Everett Jordan Lake; Figure 11). Jordan Lake was selected as a control lake because it is representative of most lakes found in North Carolina. Also, it is rain and drainage fed, a drinking water source, and used extensively for recreation. The data show

2- elevated concentrations, particularly for Ca, Mg, Sr, Li, B, V, Cr, Se, Mo, F, Cl, Br, SO4

(p<0.01) as well as for As and Tl (p<0.05), in downstream water relative to upstream water.

Likewise, the concentrations of Ca, Sr, Li, B (p<0.01), as well as V, Se, and Mo (p<0.05) were

elevated in sites downstream of the outfalls relative to concentrations in the reference lake (Table

S1).

In spite of the large dilution of effluent discharge, which plays a key role in diluting the dissolved constituents released to surface waters, we observed significant variations and differential impacts of various constituents after release into the receiving waters. We grouped the major and minor elements according to their chemical behavior in Hyco and Mayo Lakes (Table

8). In group 1, the concentrations of boron (R2=0.88; Figure 3), calcium (R2=0.96), strontium

(R2=0.95), bromide (R2=0.91), and sulfate (R2=0.86) in filtered water (0.45 um) show linear correlations with chloride during the yearlong sampling in Hyco and Mayo Lakes (Figure 13,

Table S3), reflecting their conservative (i.e., non-reactive) behavior in the lake system. Thus dilution seems to be the key factor determining their concentrations in the affected rivers/lakes.

The concentration of other elements (Se, Mg, Cr, V, and Ba in Group 2) in filtered water show a non-linear R2 with Cl (0.3 < R2 > 0.6) that suggests some attenuation in the lakes (e.g. sorption to particles).

In contrast, As, Fe, and Mn show low or no correlation with chloride (R2= 0.01, 0.07, and

0.001, respectively; Table 8), indicating strong association with suspended particles in the water column. Higher dissolved concentrations of these constituents were observed at the bottom of the lake during periods of thermal stratification in the summer and low dissolved oxygen content

(Figure S2). Seasonal stratification leads to the depletion of oxygen in bottom water during summer months and an overturning of the water column during the fall [146] . We hypothesize that under oxygenated water conditions As oxyanions would be adsorbed to Fe oxyhydroxides particles in the water column and sediment [55, 147] . During the stratification periods, when the bottom waters become anoxic, reductive dissolution of Fe (and Mn) oxyhydroxides results in release of dissolved As, Fe, and Mn to the bottom water. The reducing conditions would also 55

convert arsenate (As(V)) into arsenite [148] , a neutrally-charged form of arsenic at pH 7 (i.e.,

0 H3AsO3 ) that is less reactive towards sorption on oxyhydroxides [Raven, 1997

#44;Wolthers, 2005 #42] and also more toxic to wildlife [132] . The co-variance of As with other redox sensitive elements like Fe and Mn during thermal stratification in Hyco Lake

(Figure 14) supports this model.

In contrast, selenium does not increase with decreasing dissolved oxygen (and depth) in

Hyco and Mayo Lakes, but rather shows a linear relationship with chloride, although with a relatively weak correlation (R2= 0.65; Figure 3, Table 8), reflecting both dilution and retention effects. This is consistent with the selenium species geochemistry: under oxic conditions the oxidized species selenate (Se(VI)) would be less reactive toward sorption with oxyhydroxides and thus behave conservatively in the water column. In contrast, under anoxic conditions the partially-reduced Se species selenite (Se(IV)) would have a strong sorption affinity for both oxyhydroxides [126] and clay minerals [129, 130] . The most reduced forms of selenium (e.g. elemental Se0 and FeSe) tend to persist as sparingly soluble minerals. Overall, a transition to

anoxic conditions in the lake hypoliminion would result in lower dissolved Se concentrations

[131, 149] .

In addition to differential distribution of contaminants in the surface waters, this study revealed elevated levels of CCR contaminants (Fe, Mn, Sr, As, Mo, Sb, Ni, V and Br (p<0.01), as well as Mg and F (p<0.05)) in shallow pore water extracted from the lake bottom sediments that were significantly higher than those of the overlying bottom water. For example, As concentrations in pore water from Hyco, Mayo, and Mtn. Island lakes were as high as 83, 297, and 240 µg/L, respectively, exceeding the EPA’s MCL (10 µg/L) and CCC (150 µg/L) standards

(Figure 15). For comparison, the concentrations of trace elements (B, p<0.1 and As, p<0.1) in pore water from the reference lake were significantly lower (Figure 15). We hypothesize that retention of CCR contaminants from the lake water via adsorption onto suspended matter in the water column results in accumulation of these contaminants in the sediments that are deposited on the lake bottom. Recent reports of higher concentrations of As and Se in lake bottom sediments at the outfall at Hyco and Mayo relative to the upstream branch of the lake (Table S4) (PEC2008;

PEC 2009; PEC 2010) confirm that both As and Se are recycled through adsorption and desorption due to changes in the lake water chemistry induced from thermal stratification during the summer. Changes in the ambient conditions (pH, redox state) in the lake sediments would release these metals to the pore water [150] . We documented high levels of As in pore water and other redox-sensitive elements (Mn, Fe) that confirm this model (Figure 15). Additionally, over

82% of arsenic in the pore water collected at Hyco and Mayo Lakes were composed of the reduced and more mobile species arsenite, as detected by direct arsenic speciation (Table S4). In contrast to As, the Se concentrations were significantly higher in the CCR effluents and lake water relative to the pore water (p<0.01) (Figures 13 and 15). This indicates that Se from CCR effluent can become associated with the sediment, but that Se species become immobilized by

forming elemental selenium and metal-selenium complexes in the sediment and therefore are not incorporated into the pore water.

The accumulative nature of arsenic, selenium, and other CCR contaminants in lake systems could have ecological implications, particularly for benthic organisms and therefore the rest of the food chain. Indeed, elevated As and Se levels were reported in fish tissues from Hyco and Mayo Lakes especially near the NPDES outfall (Table S4) [144, 151-153] . Furthermore, the

2010 Mayo Lake Environmental Report [152] showed deformities in some fish, including an extended lower jaw and spinal curvature, both of which are indicators of ingestion of high levels of Se [147] . If the base of the food chain is exposed to high levels of contaminants through the sediment and/or pore water, other organisms could be at risk if they feed on those organisms that live in the contaminated sediments and pore water [55, 154] .

The impact of the effluent discharged from the NPDES outfalls on water quality in the downstream waterways is dependent on the flow rate to the river/lake (i.e., dilution effect), residence time in the water body, as well as the mobilization (e.g., adsorption/desorption to sediment) properties of specific contaminants in the water. For example, the outfall on the French

Broad River from the Asheville power plant had effluents with high contaminant concentrations

(Table 7), but because of high river discharge flow, the downstream water was significantly diluted (although still detectable). These hydrologic systems, however, can be greatly affected by droughts. During the severe drought of 2007-2008 in North Carolina, the discharge of the French

Broad River decreased drastically to just over 5 m3 s-1 , approximately five times lower than the river flow rate during the time of our sampling (25 m3 s-1) [155] . With CCR effluent discharge of

0.1 cubic meters/sec and assuming that this rate would not be changed during drought periods, such effluent discharge would make up almost 2% of the river flow downstream of the Asheville plant’s outfall. Our data show CCR discharge into smaller lakes appears to have a greater impact relative to the larger lakes (e.g., Mayo Lake versus Lake Norman). This impact is therefore a combination of the volume of released CCR effluents, the lake inflows, plant water usage

(removing from the lake system), and residence time. All of these factors can play a major role in the lake’s water quality. For instance, Hyco and Mayo Lakes have B concentrations of 958µg/L and 703µg/L, respectively compared to upstream creek B concentrations of <3µg/L and <7µg/L, respectively (Tables 7 and S1). This is a 300 and 100 fold enrichment in the B content in the lakes. Conversely, Lake Norman, the largest lake in NC, 13-14 times the size of Hyco and Mayo

Lakes, had little difference (12 µg/L) between its upstream and the downstream boron concentrations. We conclude the smaller lakes and hydrological systems are more sensitive to

CCR effluent contamination, particularly during drought periods when the dilution factor in the receiving water is expected to be reduced. Moreover, as water regulatory agencies encourage

power plants to install recycled cooling water systems rather than once-through cooling water as a way to conserve water resources, a potential unintended consequence of this policy is the discharge of CCR effluents with greater concentrations of contaminants.

This study shows that coal-fired power plants that discharge their ash pond and FGD wastewater had a significant effect on water quality of receiving waters of North Carolina. We show that even low concentrations of some contaminants, such as As, below health benchmarks at the NPDES outfall could become problematic as As is retained in suspended sediments and remobilized with environmental changes in reduced bottom and pore waters. The results of this study have significant implications for hundreds of similar sites across the U.S. given that CCR storage facilities continuously generate contaminants via leaching and transport to nearby hydrological systems. While this study focused on surface waters near CCR facilities, groundwater may have similar issues. Many CCR disposal ponds and landfills are not lined and, in many instances, are not adequately monitored, nor regulated with respect to their effects on groundwater and surface waters. This study highlights the need for rigorous monitoring and clear regulations for limiting the CCR contaminants that are being discharged into U.S. waterways.

5. Boron and Strontium Isotopic Characterization of Coal Combustion Residuals: Validation of New Environmental Tracers

5.1 Introduction

Coal fired power plants, ubiquitous in the United States and most developed countries around the world, provide affordable electricity to consumers. In the US, approximately six hundred power plants [135] generate 136 million tons of Coal Combustion Residuals (CCRs) annually, of which 56% is stored in surface impoundments and landfills. The remaining CCRs are reused for concrete, cement, and in the construction industries [136] . CCRs encompass fly ash, bottom ash and boiler slag, and flue gas desulfurization (FGD) products and are generally characterized as being enriched in trace elements such as As, B, Se, Hg, and Sr [120, 134] . The range and blends of CCRs vary substantially across coal-fired plants and depends on a unique set of circumstances for each plant and coal source [6] . Most coal-fired power plants burn a blend of coals, depending on sulfur content and the plant scrubber technology, and price of coal. These variations influence the compositions of leachates generated from CCRs in coal plants and complicate identification and tracking of the migration of CCR contaminants in the environment.

CCRs can have a significant impact on the water quality [156] , therefore distinguishing CCR effluents from other anthropogenic pollution sources is of utmost importance.

The objectives of this study are to integrate the geochemical and isotopic signatures of CCRs from a variety of coal sources and to understand their environmental applicability. While there is extensive literature on the chemistry of coal, CCRs, and their leachates [35, 36, 52, 157] , only a few studies have addressed the isotopic compositions of CCRs. Boron has two stable isotopes

(11B/10B ratio ~4), which have a large variation in nature (total δ11B range ~ 90 ‰) [39] due to the large fractionation of relatively light isotopes. The boron isotope ratios have been used to

trace groundwater contaminants such as agricultural runoff, wastewater, and petroleum waste

[38-41] . Williams and Hervig [42] measured B isotopic values of different coals in the USA

(finding a range from -1 ‰ to -70 ‰). They determined that 11B is highly depleted in coal compared to most terrestrial rocks, and concluded that coal is characterized by negative δ11B

11 11 10 values (as low as -70 ‰) relative to the international standard (δ B is defined as [( B/ B)sample/

11 10 11 ( B/ B)NIST-SRM951 -1]x 1000 and the data are reported in permil). This negative δ B value could provide a tool for elucidating CCR effluents compared to the relatively positive δ11B values found for most uncontaminated water [42] . The limited data available on boron isotope ratios in CCRs suggest that coal ash has a similar δ11B value to the coal, measured at -4 to -19‰ [42] [38, 44] , although the direct relationship between the composition of co-existing coal and CCRs has not yet been established. Coal combustion at high temperature volatilizes most of the elements associated with the organic phases (including B) and volatile elements associated with silicate phases in coal [15] . Analysis of CCRs [42, 44] indicates that there is no isotopic fractionation associated with coal combustion, and CCRs retain their 11B depleted signature. Boron is associated with the easily leachable fraction of elements [29] that adsorbed onto the fly ash particles during cooling of the exhaust gas. Since there is no species specific preferential leaching to water (i.e., 11B-enriched boric acid relative to the 11B-depleted tetrahedral boron), one would assume that there is no isotopic fractionation during the leaching of boron from CCRs.

In addition to boron, we introduce strontium isotopes (87Sr/86Sr) as an additional isotope tracer that could be utilized for tracing CCR effluents. Strontium isotopes are largely unfractionated during the mobilization of Sr in the hydrological system [45] . Coal beds in

Wyoming were analyzed for 87Sr/86Sr isotopes as a tracer for hydrologic basin movement, which revealed heterogeneity in the Sr ratios, ranging from 0.712 to 0.714 [46] . Hurst et al [47] analyzed coal, fly ash, and bottom ash for 87Sr/86Sr and found a range of 0.70883 to 0.70972.

Each of these analyses was performed on specific coal and CCRs, therefore it seems that Sr isotope ratios could vary in different geological basins, which could be used as a tool to detect the link between CCRs and the original coal source and/or environmental effects. Given that the Sr isotopic composition of coal and CCRs is expected to reflect the local geological rocks and hydrology of the original depositional environment in which coal was deposited, we hypothesize that Sr isotopic variations of CCRs would reflect their original coal source. In fly ash, a minor portion of strontium is potentially leachable due to surface adsorption, but the bulk is associated with the silicate phases and therefore the leaching is relative slower [30] . Different compositions and phases in the ash could lead to variations in Sr isotopic ratios. Another factor to consider with

Sr isotopes would be the addition of lime (CaO) or CaCO3 with a different Sr isotopic ratio in

SO2 scrubbers (Flue Gas Desulfurization- FGD) and its impact on the Sr isotopic fingerprints of leachates originated from such CCRs (e.g., CaCO3 contains 500-1500 ppm Sr).

By developing integrated geochemical and isotopic fingerprints of the CCRs and their source coals, this investigation provides the necessary diagnostic tools to combat one of the increasing sources of anthropogenic pollution. The study is based on (1) systematic laboratory experiments of eight coal fired power and steam plants burning an assortment of coals from the

Appalachian, Illinois, and Powder River Basins in the United States (Table 8) and (2) field environmental sampling from surface and porewater in a river system affected by the 2008 TVA coal ash spill in Kingston, TN and ten CCR effluent discharges in NC and TN, as well as affected surface and pore waters in NC. The combination of the geochemical indicators from the laboratory experiments and environmental sampling provides a unique and practical identification method for evaluating the impact of CCRs on the environment.

5.2 Analytical Methods

5.2.1 Field Sampling and Leaching Experiments

Coal combustion residual samples were collected from eight coal fired power and steam plants burning an assortment of coals from the Appalachian, Illinois, and Powder River Basins in the United States. All of the fly ash samples from the TVA Fossil Plants (John Sevier, Bull Run, and Kingston), as well as the samples collected by Jim Hower (from plants coded PRB, E, M, and

R) were leached according to the potential EPA Method 1316 (varying solid to liquid ratios). The de-ionized water leaching test results reported in this paper are from the 0.1 solid to de-ionized water ratio portion of the Method 1316 leaching procedure. Leaching experiments were simulated with the spilled TVA spilled ash collected from the 2008 TVA coal ash spill (the Cove) [4] and the Duke University Steam plant ash with solutions exhibiting a wide range of acidity (pH range of 0.4 to 12).

Field trips were made to the TVA spill site between January 2009 and June 2011 with over 270 surface and porewater samples collected during high (winter) and low (summer-fall) river flow regimes [4, 156] . Field trips were made to lakes and rivers in North Carolina associated with disposal of CCRs wastewater between August 2010 and February 2012. Surface and porewater samples were collected monthly from Hyco and Mayo Lakes from August 2010 through August 2011, and the other water bodies sampled include Lake Norman, Mountain Island

Lake, Lake Wylie, Dan River, French Broad River, and Jordan Lake as a reference lake [158] .

Water sampling strictly followed USGS protocols [107] ; after filtration of samples in the field

(0.45 µm syringe filters), trace elements were measured by inductively coupled plasma mass spectrometry (ICP-MS), major elements by direct current plasma optical emission spectrometry

(DCP-OES), and anions by ion chromatography (IC). Pore waters were extracted from river 64

bottom sediments obtained using a Wildco box core (up to 25 cm depth), VibeCoreTM (up to 182 cm depth), or peat core (up to 100 cm depth).

5.2.2 Boron Isotopes

Boron from surface water, effluents and leachates was processed through cation- exchange resin (AG® 50W-x8 Resin) to remove all cations (particular interference of Ca with the

BO2- ions), treated with peroxide to remove organic matter and CNO complexes, loaded on the

Triton (Thermo) thermal ionization mass spectrometer at Duke University and measured as BO2- ions on low-temperature negative ion method developed recently by Dywer and Vengosh [159] .

All sample loading was carried out in a vertical laminar flow clean hood equipped with boron- free PTFE HEPA filtration. Data on standards (NIST951, OISL Atlantic seawater, and IAEA

Groundwater B-3) loaded using this method yield external precision of approximately 0.5‰ δ11B.

The average boron ratio received for NIST951 during these analyses was 4.00281. The variability within replicates was ±1.5‰. Total loading blank is <15pg B as determined by isotope dilution

(NIST951). The load solution delivers ionization efficiency similar to seawater and has negligible

CNO- (mass 42) interference, based on negligible signal at proxy mass 26 (CN-).

5.2.3 Strontium Isotopes

Strontium from surface water, effluents and leachates was evaporatively preconcentrated in HEPA filtered clean hood and re-digested in 0.6mL of 3.5N HNO3 from which strontium was separated using Eichrom Sr-specific ion exchange resin. Approximately 1 to 10µg Sr was loaded onto out-gassed single rhenium filaments along with TaCl activator solution and loaded onto the

Triton TIMS at Duke University. Samples and standards were gradually heated to obtain an 88Sr beam intensity of ~3V, after which 220 cycles of data were collected, yielding a typical internal

precision of ~ 0.000004 for 87Sr/86Sr ratios (1 sd). External reproducibility on standard NIST987 yielded a mean 87Sr/86Sr ratio of 0.710265 ± 0.000009 (1 sd).

5.3 Results and Discussion

5.3.1 Boron and Strontium in Coal Combustion Residuals

Coal Combustion Residuals (CCRs) can vary widely in their chemical and mineralogical make-up based on their source coals, combustion temperature and process, as well as post-combustion treatment [6, 120] . Boron and strontium have been shown to be high in abundance and easily leached off of CCRs [4, 120, 156] , therefore can be good indicators of

CCR leachate. The results from the 0.1 solid to de-ionized water leaching test are shown in Table

9. The results showed that there is a wide range of concentrations of leachable constituents from the fly ash (e.g., B ranged from 1 ppm to >140ppm). In addition to Ca, Mg, V, Mo, Ba, and Li, boron and strontium had elevated concentrations in the leachates of the tested CCRs samples

(Table 9). The concentration variations could be accounted for with different combustion and treatment processes, as well as different coal sources (Table 8). In this study we investigated

CCRs originated from source coals of a variety of coal basins (Appalachian, Illinois, and Powder

River Basin). Large variations in boron leachate concentration were seen based on the coal source burned at the various plants (Table 9). For example, the Illinois basin coal ash had the highest boron concentration of all of the other sources sampled, with concentrations ranging from 60 ppm to 142ppm (normalized to weight). The Appalachian and Powder River Basin (PRB) (and mixtures) had much lower boron concentrations (below 32 ppm B). The Powder River Basin had the greatest concentration of strontium (273 ppm) and the mixture of Appalachian and PRB had the second greatest concentration (95 ppm). The Appalachian and Illinois Basins had Sr

concentrations less than 60 ppm. In addition to the coal sources, we observed variations in the

CCRs leachable concentrations in different units within the same plant (Table 9). Fly ash was sampled from seven units within one plant and individually tested with the same 0.1 solid to deionized water ratio (Table 9). The different units represent variations seen in the plant due to actual process differences in coal homogenization, pulverizer efficiency, and particle size differences. The units showed a range in Sr (50-85ppm) and B concentration (4-35ppm), yet they are from the same source coal (App/PRB Mixture). These differences demonstrate the potential heterogeneity of the concentrations of leachable constituents from CCRs from the same plant.

These variations could be due to the heterogeneous nature of coal, differing combustion conditions (temperature), or capturing different size fractions at each unit [18, 120, 160, 161] ).

This chemical heterogeneity also complicates the prediction of the environmental effects of

CCRs, and thus additional tools, such as isotopes are required to have a better assessment of the environmental fate of CCRs leachates.

Some of the CCR samples were also leached under a range of pH conditions in order to mimic possible variety of environmental conditions that they could encounter (e.g., acid mine drainage with a low pH). The data show that the concentrations of B and Sr removed from the

CCRs during leaching significantly vary with pH (Figure 16). At low pH, B concentration in leachates was the highest, indicating that boron mobilization is enhanced under acidic conditions.

Likewise, the Sr concentrations in leachates under low pH were the highest (Figure 16).

Table 9: List of the plants and CCRs that were sampled and then had leaching experiments performed on them.

Figure 16: Leaching experiments on the spilled TVA CCRs [145] and the Duke Steam Plant (blue) fly ash. The Duke Steam Plant ash had lime addition post combustion. Low pH values leached off more strontium and boron than the neutral and basic pH values.

5.3.2 CCR Isotopic Signatures

Previous studies have shown that U.S. coal are characterized by δ11B values of -1 to -

70‰ [38, 42] , while CCRs; a bituminous coal ash [38] , two unknown sourced fly ashes [44] , and NIST1633a [42] ; are also characterized by negative δ11B values, as low as -19‰. This preliminary data suggest that the boron isotopic characteristics of coal and CCRs [38, 42, 44] are relatively unique. In this study we expand the CCRs collection and include CCRs originated from the most of the major coal basins in the USA. The boron isotopes ratios from the 0.1 solid to deionized water leaching-tests had a range of δ11B values from -18 ‰ to +7 ‰ (Figure 17). These results are similar to previously published values for CCRs, ranging from -19 ‰ to +16 ‰ [38,

44, 162] . Previous studies have considered the isotopic variations based on the coal rank, which revealed no pattern of δ11B values (bituminous and lignite fly ashes with negative δ11B and the sub-bituminous coal fly ash with a positive δ11B)[38] . They have yet to be discriminated based on their geographical (therefore depositional and historical) locations. Figure 17 shows that CCRs 70

derived from the Appalachian basin coals had δ11B range of -18 ‰ to -14 ‰, Powder River Basin has a δ11B of -4 ‰, and a blend of Appalachian and Powder River Basin had δ11B value of -8 ‰.

The CCRs originated from coal from Illinois basin had the largest δ11B range from -8 ‰ to +7 ‰.

Overall, the boron isotopic signatures of CCRs leachates indicate that CCRs have some different

δ11B values that are related to different coal sources. This variation could reflect different boron sources during coal formation and/or subsequent metamorphism during coalification.

The results from the variable pH leaching experiments (Figure 18) of the TVA coal ash show a range of δ11B values from -18‰ to -14‰. The data show a progressive decrease in

δ11B values in the leachates through the transition from acid to neutral conditions for the TVA ash

(Figure 18), and then a slight increase under leaching in basic conditions (Figure 18). Given that we assume that no preferential leaching of boron species occurs, these isotope variations could imply different boron sources in the CCRs with different boron species distribution; under low pH an intensive mobilization of boron would release boron also from the mineral phase with higher δ11B values (and apparently higher trigonal boron) relative to leaching under basic conditions. Nonetheless, the δ11B variation seen across the whole pH scale was only 4‰, which is negligible relative to the large boron isotope variation in nature.

In addition, we observed isotopic variations in leachates from individual units sampled from the same plant, with δ11B values from -9 ‰ to -4 ‰ (Figure 19). This, again, adds an internal variation among units of the same plant but with only 5 ‰ variations in boron isotopic ratio. This difference could be due to heterogeneity of the B isotopic ratio in the source coal [42] .

Strontium isotope ratios were also recognized as a possible CCR tracer [163, 164] . It was shown that Sr isotopic ratios for CCRs can be characterized by coal rank [164] , in which the

87Sr/86Sr ratio increases with ranking; lignites of 0.70767, subbituminous of 0.70874, and bituminous coal of 0.71022 [164] , due to the lower rank coals containing more Sr bearing

carbonate and sulfate minerals [164] . Yet differentiating 87Sr/86Sr based on coal rank could be misleading, due to the fact that the rank of coal doesn’t dictate its geographic origin nor age, which can have major impact on its isotopic composition. The ashes in the above study were collected from plants spanning the United States, but it is not stated if the coal source was also from that region. The 87Sr/86Sr isotopic ratio of the CCR leachates in this study ranged from

0.7110 to 0.7125, the most radiogenic 87Sr/86Sr ratio (Figure 17). The data show that the 87Sr/86Sr isotopic ratios of the Appalachian Basin coals had the least radiogenic ratio (0.70975-0.71108), while CCR leachates originated from coals from Powder River Basin had a more radiogenic ratio

(0.71221), and CCR leachates from Illinois Basin had a broader range of 87Sr/86Sr ratios (0.71137 to 0.7125) (Figure 17). These ratios are inconsistent with the Sr ratios subdivided by coal rank.

Appalachian and Illinois coals are generally characterized as bituminous, while the Powder River

Basin range from subbituminous to lignite (less common).

The Sr isotope results of experimental TVA ash leaching show a progressive decrease in Sr concentrations and 87Sr/86Sr ratios with pH (Figure 18). This indicates that higher extraction of Sr from CCRs under acid condition mobilizes more radiogenic 87Sr/86Sr ratio. The

87Sr/86Sr ratio change could be a result of the dissolution or break-up of a less soluble phase of the

CCR material that has a more radiogenic signature, compared to the adsorbed (i.e. easily leachable) strontium. In contrast, leaching experiments of lime-treated ash generated in Duke

University steam plant show significantly lower 87Sr/86Sr ratios (Figure 18), possibly as a result of lime addition to the Duke Ash post-combustion.

Figure 18: The δ11B and 87Sr/86Sr from the variable pH leaching tests on the Spilled TVA ash [145] and the Duke Steam Plant Ash (Blue). The Spilled TVA ash has a more radiogenic 87Sr/86Sr value than the Duke Steam Plant Ash (that was treated with CaO).

5.3.4 Environmental Applications

5.3.4.1 The 2008 TVA Coal Ash Spill

One of the world’s largest CCR spills occurred at the Tennessee Valley Authority’s

Kingston Fossil Plant in Kingston, TN in December 2008. The retention pond breach resulted in over 4.1 million cubic meters of wet CCRs spilled into the Emory and Clinch Rivers as well as the surrounding land surface. The results from the TVA spill [4, 156] confirmed that boron is one of the sensitive indicators for metals leaching from CCRs, with boron content up to 1600 µg/L in downstream pore water under ash covered sediments, relative to the upstream and

uncontaminated river water with boron of 6 to 9 µg/L (Figure 20). Results from surface water at a region covered with spill ash of limited water exchange (known henceforth as the Cove) [4, 156] from the TVA spill area yielded conspicuously low δ11B values of -12‰, which is consistent with previous reported δ11B values of -4 to-19‰ in coal ash [38] and our leaching experiments of the same TVA coal ash (-18‰ to -14‰). The depleted 11B composition is significantly different from that of common uncontaminated boron, as demonstrated by the δ11B values of the upstream

Emory (+9.6‰) and Clinch (+10.2‰) rivers (Figure 20). The conspicuously depleted isotopic signal of boron isotopes in CCRs is also different from other anthropogenic source such as domestic wastewater (δ11B =0 to +10‰) [39] , and thus could be used as a sensitive tracer for quantification of CCRs leachates migration to the environment.

Similar to boron, water in the ash spill area (Cove) is characterized by high Sr content (up to 1240 µg/L) relative to the pristine upstream Emory River (25 µg/L) (Figure 20). The 87Sr/86Sr ratios found in the ash spill area (Cove) (0.7110 to 0.7128) are consistent with the TVA spilled

CCRs leachates (0.712) that were generated through the experimental leaching experiment

(Figure 20). The range of 87Sr/86Sr values found in the upstream Emory was 0.7117-0.7124 and the Clinch River water was 0.7124, while the downstream river had a much narrower range from

0.7119 to 0.7127 (a result of the mixing relationships between the two rivers). Thus, the Sr isotopic ratio of spilled CCRs overlap, in this case study with the 87Sr/86Sr variability of the upstream source waters. This demonstrates that, in spite of the high concentration of Sr in CCR leachates, there could be circumstances that Sr isotopes could not be used as distinguishing tracers for elucidating the CCR leachates in a water system due to similar isotopic fingerprints.

Nonetheless, the downstream pore water and cove water had distinctly higher Sr concentrations, which could be helpful in identifying CCRs in combinations with the depleted (and distinctive)

δ11B values.

5.3.4.2 CCR Effluent Discharges and Water Resources

The CCRs effluents were sampled from ten coal fired steam and power plants in North

Carolina [158] and Tennessee from August 2010 until August 2011. The outfalls (i.e., discharge of CCRs effluents to the environment) show a range of leached chemical constituents (B, Sr, Mo,

V, etc) with high enrichment factors relative to the upstream source water [158] . The boron and strontium isotopes were measured in CCR effluent from some of the coal fired power plants in

Tennessee and North Carolina [158] . The data show that all CCR effluents are characterized by elevated boron concentrations and relative low δ11B values ranging from -12‰ to -0.2 ‰ (Figure

21). The range of boron isotope ratios measured in the effluents could be indicative of the various processes undertaken in the plant as well as the coal sources. For instance, one of the TN plants disposes of their fly ash in a dry manner and therefore only transports the bottom ash via water

(hence the discharge is only representative of bottom ash). Also, many of the NC plants (and one 76

TN) have a Flue Gas Desulfurization component, which could also alter the boron and strontium isotopic ratios (Figure 21).

The depleted 11B composition of the CCR effluents is different from that of common boron in water resources, as demonstrated by the reference lake (Lake Jordan) in North Carolina with lower B concentrations (<100 ug/L) and δ11B values ranging from 0 to +7‰ in surface water. The boron concentration variations found in the CCR effluents indicate that the FGD effluent has higher concentrations of boron than those effluents without FGD. There is no significant difference in the δ11B values however between the FGD and non-FGD effluents, which indicates that the FGD process does not add a significant source of boron from the injected carbonate or lime, but rather captures more of the volatile boron from the combusted coal before it escapes the smokestacks. The North Carolina and Tennessee plants were burning Appalachian basin coals during the sampling period, except for one TN plant that was burning a 60/40 mixture of Appalachian and Powder River basin coals (Table 8). The δ11B values of the effluents from these plants are consistent with those of the leaching experiments performed for the selected

CCRs sorted by coal basin origin (Figure 17).

In the North Carolina investigation samples were collected upstream and downstream of the CCR effluent outfalls to determine the effect on water resources receiving CCR effluent. The boron concentrations of the CCR effluents were significantly high relative to the upstream

(reference) waters (the upstream waters were used to transport the CCR materials) [158] . The

δ11B values of the rivers and lakes at the outfall sites and downstream from CCRs effluent disposal were significantly lower relative to the upstream (non-impacted) sites (Figure 22). The only exceptions were enclosed lakes with high CCR discharge volume and long residence time

(Hyco and Mayo Lakes), in which the lake water is recycled as a source for the coal plant. In these cases, the lakes had an even distribution of low δ11B values (-8 ‰ to +2 ‰), without

showing the isotopic transition from non-impacted to impacted lakes as shown in most of other lakes and rivers in NC. In systems where water is not recycled for cooling the coal plants, we observed a sharp increase in B concentration and a drop in δ11B at the outfall, relative to the upstream waters (with the exception of Lake Wylie). The downstream waters had a typical composition that reflects a mixture of the upstream waters and CCR effluents. For instance, the

French Broad River upstream water had a boron concentration of 8 ug/L and δ11B of +8 ‰, while the CCR effluents had B concentration of 2379 ug/L and δ11B of -7.7 ‰. The downstream mixture of these two end-members resulted in a boron concentration of 115 ug/L and δ11B of -7

‰, revealing the influence of the CCR effluent discharge into the river (Figure 22). A simple mass-balance calculation shows that in this case the contribution of the CCRs input to the downstream river is about 5% (i.e., 5% of the dissolved constituents of the river is derived from

CCRs effluents discharge). Similarly, we observed similar patters in the chain of lakes (Lake

Norman, Mountain Island Lake, and Lake Wylie) near Charlotte, NC (Figure 22).

The 87Sr/86Sr ratios of the in the North Carolina CCR effluents ranged from 0.7075 to

0.7147 while the Tennessee CCR effluents ranged from 0.7107 to 0.7117 (Figure 21). The only available FGD effluent sample (from the Bull Run Fossil Plant) had a very high Sr concentration

(>10,000 ug/L) and a lower 87Sr/86Sr ratio of 0.708 relative to the other CCRs effluents. The bulk

CCRs effluents of the same plant in which we collected the separated FGD effluents had lower Sr concentration (500 ug/L) and higher 87Sr/86Sr (0.71081; Figure 21; green dot and circle). At the same time, experimental leaching of coal ash from the same plant resulted yielded a higher

87Sr/86Sr (0.711076). This indicates that the Sr isotope composition of the bulk CCR effluents discharging from this plant represent a mixture of Sr derived from leaching of coal ash and FGD effluents with high and low 87Sr/86Sr ratios, respectively.

As mentioned above, the North Carolina and Tennessee plants were burning Appalachian basin coals during the sampling period, except for one TN plant that was burning a 60/40 mixture of Appalachian/Powder River basin coals. Some of the North Carolina CCR effluents sampled fit within the Appalachian range (Figure 17) determined with the leaching experiments, but several are less or more radiogenic than the Appalachian range. The low 87Sr/86Sr shift could be a result of the contribution of the FGD effluents in which lime or limestone with typical lower 87Sr/86Sr ratio (~0.7068 0.7095) [165] are added to the flue gas, introducing an additional Sr component to the effluent. Another explanation is that the source water for the plant (river or lake water) may have sufficient Sr and an isotopic signature that is mixed with CCR leachates and therefore could change the 87Sr/86Sr range of the Appalachian coals determined by the leaching experiments. The

Tennessee effluents are consistent with their appropriate ranges, whether they are Appalachian coals or the mixture of Appalachian and Powder River basin. Overall, these results indicate that the CCRs without the FGD system may be more representative of the ranges determined by the leaching experiments (Figures 17 and 21), while FGD could introduce a significant amount of Sr with a lower 87Sr/86Sr signature.

Samples were collected upstream and downstream of the CCR effluent outfalls to determine the effect on the water quality and isotopic composition of water resources receiving

CCR effluent. The Sr isotopes in the North Carolina river and lake downstream samples showed mixing relationships between the upstream (non-impacted) water and the CCR effluents. Every water body sampled had higher Sr concentrations in the outfall and downstream samples and typically (with the exception of Hyco Lake) lower 87Sr/86Sr ratios relative to the upstream samples

(Figure 23). In all of these cases, the downstream Sr composition represents a mixture of the upstream and outfall (Figure 8). In contrast, results from Hyco Lake showed an opposite phenomenon, where the 87Sr/86Sr of the outfall was higher than the uncontaminated lake water

(Figure 23). The relationships between Sr concentration and 87Sr/86Sr in the downstream samples show a clear mixture of the CCR effluents and background waters demonstrating that Sr isotopes can reveal the influence of CCRs on the water resource. Likewise, we see similar patterns in

Mayo Lake (Figure 23), where the 87Sr/86Sr ratio of the CCR effluents is higher, yet still lower than the uncontaminated water as shown in other NC lakes. We used the Sr isotope variation to calculate the mixing fraction of CCRs effluents in the two lakes, showing a range of 88% to 98% in Hyco lake and 46% to 69% in Mayo Lake. Hyco and Mayo Lakes show more of a gradational relationship with Sr compared to B, due to the higher input of Sr (21-82 ug/L Sr compared to <8 ug/L B) from the streams feeding the lakes. The 87Sr/86Sr signatures from the CCR effluents in

North Carolina are distinct enough, unlike the TVA spill that they allow for the Sr isotopic fingerprinting of the CCR effluent into the environment.

Pore water, or the interstitial water extracted from the lake bottom sediment, was sampled at Hyco and Mayo Lakes in the North Carolina. The top 25 cm (or less) of the sediment column was sampled for porewater, and revealed slightly lower concentrations of boron in the porewater relative to the overlying surface water (Figure 24). The porewater showed a range of δ11B that were similar to those found in the overlying surface water at Hyco and Mayo lakes with relatively low δ11B values ranging -8‰ to -0.5‰ (Figure 24). A long core was also taken at Hyco Lake

near the outfall site, up to 52 cm deep within the bottom lake sediment column. The boron concentration decreased with depth in the column (from 817 ug/L to 81 ug/L), while the δ11B increased from -5.5‰ to +5.5 ‰ with depth (Figure 25). Chloride, which behaves conservatively in the lake [158] , also decreases with depth in the column from 58 mg/L to 26 mg/L, yet by only two-fold while boron decreases by ten-fold. The B/Cl ratio, therefore, decreases with depth and parallel to the 10‰ increase in δ11B values (Figure 25). This suggests that boron in the pore water is being adsorbed by the sediments, in addition to a general shift in the pore water chemistry along the 52 cm core. Numerous studies have shown that the association of boron depletion and

11B enrichment is the result of preferential retention of the light boron isotope (boron-10) from the water, resulting in 11B enrichment in the residual waters [166-169] .

Figure 24: Delta B and B concentration in Hyco and Mayo lakes. The pore water has a negative delta B value like the lake water (Figure 24). The pore water has slightly less boron than the outfall or surface waters (indicating possible adsorption).

Figure 25: Long Core pore water results. Delta B increases with depth, while both B and Cl concentration decrease with depth. The B/Cl ratio increases with depth.

Overall, this study shows that in spite of some isotopic variations in the coal sources

(boron and strontium isotopes), combustion procedures, and post treatment such as FGD

(strontium isotopes), the boron and strontium isotopic fingerprints of leachates generated from

CCRs are unique and could be different from uncontaminated water and/or other anthropogenic contamination sources. The application of boron and strontium isotopes, combined with the geochemical characteristics, therefore provides a novel tool for identifying CCR effluents and quantifying their impact on the environment. The application of strontium isotopes alone may not be able to identify CCR effluent in some cases where the Sr isotopic composition of the uncontaminated water overlaps with that of CCRs, but when combined with geochemistry and boron isotopes, can still provide sufficient evidence of CCR impact. Future studies should also use these tracers for other application, such as monitoring groundwater contamination underlying the many unlined CCR ponds across the United States.

6. Conclusions

Coal combustion residual disposal presents an interesting environmental challenge. CCRs can be used in concrete, cement, and many other applications, providing an alternative to mining a new source material, or they are held in storage ponds indefinitely. Both situations bring up concerns of leaching elements adsorbed to the CCRs into nearby waterways. In addition, recent air regulations have become more stringent, resulting in better capture technology, which removes the sulfur oxides along with more of the volatile elements from the flue gas, preventing them from being released to the atmosphere. If these enriched CCRs are disposed of in a wet manner, the resulting effluent can be high in many potentially hazardous elements; therefore, trading clean air for dirtier water. This dissertation presented the findings from an investigation of an accidental spill as well as “business as usual” scenarios in the disposal of coal combustion residuals. This investigation established the geochemical and isotopic tracers of CCR contaminants, providing the tools to identify and predict the fate of CCR contaminants in the environment. This dissertation is based on research of the TVA spill and CCR effluent in North

Carolina and Tennessee, where the migration of metals from CCRs and the water quality impacts were monitored. The data generated in this project provide an in-depth evaluation how different coal sources control the levels and compositions of CCR contaminants, the environmental conditions in which contaminants are mobilized from CCRs, and their possible fate in aquatic systems. In addition, this novel approach for geochemical and isotope fingerprinting CCR contaminants provides a new monitoring tool that can be used to identify CCRs contaminants in aquatic systems and to distinguish them from other pollution sources that may also contaminate the environment. Understanding the leaching systematics of CCRs is also important for developing adequate CCR management and beneficial use of recycling products of CCRs.

Together these investigations have illuminated several conclusions. First, contact of 84

surface water with CCRs can leach high concentrations of leachable CCR contaminants, such as

As, Se, B, Sr, Mo, and V in the surface waters. High concentrations were measured at the Cove area at the TVA spill, as well as from some of the CCR effluent discharges sampled from the NC and TN plants. This conclusion can be applied to many situations where water comes into contact with CCRs, for instance, within the holding ponds, the groundwater below unlined holding ponds, and re-use situations such as structural filling for overpasses or golf courses. Secondly, the relationship between levels of contaminants in the CCR effluent and the ratio between effluent flux and freshwater resource volumes (i.e., the dilution effect) controlled the concentration within the body of water. This was illustrated in North Carolina where the larger lakes, like Lake

Norman, saw little impact from the discharge of CCR effluent, whereas the smaller lakes (Hyco and Mayo) with relatively larger effluent discharge showed much higher impacts (Ex. Boron concentrations 753-1000ppb in the lake, whereas the upstream was less than 7ppb). A further illustration of the dilution effect is the mixing of the concentrated FGD discharge with cooling water at Hyco Lake (Table S2 in Appendix C). Third, recycling of trace elements occurs through adsorption on suspended particles and release to deep surface water or pore water in bottom sediments during periods of thermal water stratification and induced anoxic conditions. This phenomenon was seen at several lakes with porewater as well as the deep surface waters with arsenic concentrations well above the average surface water concentration. CCRs can have an accumulative impact in certain areas, as evidenced by Hyco and Mayo Lakes. These lakes have relatively long residence times and don’t flush out the CCR effluent; therefore the CCR contaminants accumulate in the lake. Lastly, elevated boron and strontium concentrations, in addition to their isotopes, can be used as tracers of CCR effluent in the environment. The surface water and porewater revealed high concentrations of both boron and strontium, with a negative

δ11B value. The 87Sr/86Sr values showed distinction from the surrounding water and rocks

87Sr/86Sr values in some situations, while in others there was no distinction. Strontium is therefore a tracer that can be used in situations where the CCR 87Sr/86Sr values are significantly different from those of the surrounding geology and water sources. Overall, this investigation has proven that the combination of the geochemical indicators from the laboratory experiments and field sampling (surface and pore water) provides a unique and practical identification method for evaluating the impact of CCRs on the environment.

This investigation was novel due to the combination of leaching experiments with environmental applications. Previous investigations have performed numerous leaching experiments, but field component truly sets this investigation apart. CCRs are ubiquitous throughout the world, and therefore the field application of the geochemical and isotopic characteristics allows for a true environmental understanding. The ability to use the elevated concentrations (B and Sr) as well as the unique isotopic signatures makes a distinctive indicator of CCRs.

Appendix A- Supplementary Material for Chapter 2

Supplementary Text S1: Analytical techniques

Inorganic elemental concentrations (Li, B, Mg, Ca, Cr, Cd, Mn, Co, Ni, Cu, Zn, As, Se,

Rb, Sr, Mo, Ba, Pb, Th, U, Al) were determined by inductively coupled plasma mass spectrometry (ICP-MS) on a VG Plasmaquad 3 at Division of Earth and Ocean Sciences in Duke

University. Instrument calibration standards were prepared from serial dilution of certified water standard NIST1643e. Samples and standards were diluted to the same proportions with an internal standard solution in 1% HNO3 containing 10 ppb In, Tm, and Bi to monitor and correct for instrument drift. All dilutions were carried out with solutions prepared from deionized/quartz- distilled H2O and quartz-distilled HNO3. Instrument drift was also monitored and corrected for each element by analysis of one of the calibration standards at regular intervals between analysis of samples and standards. To allow for determination of U concentrations,

1643e, which does not contain U, was spiked with plasma-grade single-element solutions of U.

External precision for most elements is typically 4% or less, based on replicate analysis of samples and standards.

Sediment samples for Hg analysis were collected in polypropylene centrifuge tubes and stored at

4°C until analysis. Hg content was quantified in triplicate analyses with thermal decomposition, amalgamation, atomic absorption spectroscopy (Milestone DMA-80) (1).

Instrument calibrations were conducted with a Hg(NO3)2 stock acidified with 0.1 M HNO3.

SRM 2709 San Joaquin Soil [170] was used as a reference standard to validate instrument calibration during the analysis. We quantified a value of 1.44 ±0.01 mg/kg in triplicate analyses of the reference (NIST reported value was 1.40 ± 0.08 mg/kg).

Radium measurements of solid ash samples were performed on a Canberra (DSA2000) broad-energy Ge gamma-ray spectrometer at Duke University. 228Ra was quantified from a 87

weighted average of counts from the 338 keV and 911 keV peaks of 228Ac. Counting error of individual samples depends on the activity of the Mn-fiber sample (total 228Ra range of 1 to 60

Bq) and run time. The precision of our measurements is typically less than 10%. Activities of

226Ra were obtained through the 352 keV energy line of its radioactive granddaughter, 214Pb after incubation for three weeks. 210Pb activities were measured by 46keV energy line. The activities of all these nuclides were calibrated using CCRMP U-Th ore standard DL-1a and

Canberra Multi-Gamma ray standard MGS-5C, measured under physical conditions identical to the samples (2).

Table S1: Table of sampling locations at the TVA spill site. The site numbers correspond to Figure 3 in the paper.

Appendix B- Supplementary Material for Chapter 3

Supplementary Text S1: Analytical techniques

A fieldtrip to the site included field measurements (pH, dissolved oxygen, conductivity) and collecting of water and solid ash samples from the site. Water sampling strictly followed

USGS protocol [107] ; trace metal and cation samples were filtered directly into new, high-purity acid-washed polyethylene bottles containing high-purity HNO3 in the field for preservation using syringe-tip 0.45 µm filters. Sediment was collected from the river bottom with a Wildco box core

(with complete recovery reaching 25 cm depth). A vibracore was also used to extract deeper sediments (up to 182 cm deep) from the river bottom. The river sediments collected from the box core were homogenized and porewaters were extracted in the field with a Nalgene vacuum filtration unit with a 0.45 µm filter. Inorganic elemental concentrations (Li, B, Mg, Ca, Cr, Cd,

Mn, Co, Ni, Cu, Zn, As, Se, Rb, Sr, Mo, Ba, Pb, Th, U, Al) were determined by inductively coupled plasma mass spectrometry (ICP-MS) on a VG Plasmaquad 3 at the Division of Earth and

Ocean Sciences in Duke University. Ca, Mg, Na, Sr, Ba, Fe, Mn, and Si concentrations were determined by direct current plasma optical emission spectrometry (DCP-OES) on an ARL-

Fisons Spectraspan 7 that was calibrated using solutions prepared from plasma-grade single-

− − 2− element standards. Major anion (Cl , NO3 , SO4 ) concentrations were determined by ion chromatography on a Dionex DX-500 ion chromatograph (IC). Bicarbonate concentrations were determined by titration to pH 4.5.

For acid volatile sulfides [97] , AVS in whole sediment samples from the box core were homogenized and collected in borosilicate glass jars and frozen until analysis. AVS is generally

- comprised of acid-soluble sulfide phases such as dissolved inorganic sulfide (H2S, HS ),

0 amorphous FeS(s), and ZnS(s) and does not include more recalcitrant phases such as pyrite, S , polysulfides, and other metal (Cu, Pb, Hg, Ag) sulfides[109] . AVS separation was performed in

the laboratory by adding HCl to thawed sediment slurries, purging the sample with ultra high purity N2, and trapping volatile H2S in a 1 N NaOH solution [110] . Dissolved sulfide in this solution was quantified immediately by colorimetry [110] and was sampled and measured using the Allen method [109] . Dry-wet ratios were determined in separate aliquots of the samples and were used to report AVS concentration normalized to sediment dry weight.

Inorganic arsenic species were measured using the Bednar method [108] for field preservation of arsenic species in water. Each sample underwent 0.45µm filtration and preservation with 1.25mM EDTA to chelate metal cations, buffer the sample pH, and reduce microbial activity. The arsenic species are then separated using a syringe to elute the EDTA- preserved sample through the cartridge containing a strong anion-exchange resin (Supelco, Belle- fonte, PA; LC-SAX, 3-mL cartridge barrel) into a new, high-purity acid-washed polyethylene bottle containing high-purity HNO3. The charged arsenic species (primarily As (V)) are removed in the cartridge, and the remaining uncharged arsenic in solution is As[148] .

In order to test this methodology, primary standards of 100mg-As/L of the two arsenic species were prepared from sodium m-arsenite (NaAsO2) for As[148] and potassium dihydrogen arsenate (KH2AsO4) for As(V). Five standard solutions were prepared from different mixing ratios of As[148] to As(V). Each mixed solution was filtered with a 0.45 µm filter into a 30 mL new, high-purity acid-washed polyethylene bottle containing 50µL of 1.25 mM EDTA. The sample was then processed through the anion exchange resin cartridge. The residual As (primarily as As(III)) was determined by ICP-MS. The results from this experiment are shown in Figure S1.

Figure S1: Results of arsenic species separation experiment (based on Bednar Method[108] ). Known mixtures of arsenite and arsenate were processed through anion exchange resin and measured for the residual uncharged arsenite concentrations. The measured arsenic concentrations are consistent with the expected values (R2 =0 .998), indicating the robust nature of this field procedure.

Supplementary Text S2: Coal ash leaching experiments

We have conducted laboratory-leaching experiments on TVA bulk coal ash from the

Kingston spill area in Kingston and Harriman, TN. Leaching experiments were simulated reactions of CCPs with strong acid (HCl in 0.5N), weak acid (HCl in 0.02N), strong-basic (NaOH in 0.2N), weak-basic (NaOH in 0.00063N), deionized water under two S/L ratios, and Toxicity

Characteristic Leaching Procedure (TCLP; EPA method 1311; [133] ). Each simulation was run for ~24 hours then the liquid was decanted off, filtered, and then analyzed in the method designated above. Leaching conditions varied over a wide range of acidity (pH of 0.4 to 12). The leachates were analyzed at Duke in a manner consistent with the samples collected.

Experiments were also performed to simulate Total Suspended Solid values measured in the Emory and Clinch Rivers near the spill site. TSS values were simulated by mixing varying ratios of ash from the TVA spill site (collected near the Cove shortly after the spill) with DI water and upstream Emory River Water. The TSS values replicated were 10, 100, 500, 1000, 3000, and

3500 mg/L each in DI and river water, which equates to solid/liquid ratios of .00001, .0001,

.0005, .001, .003, and .0035, respectively. The ash was mixed with the water then shaken for 24 hours then the liquid was decanted off, filtered, and analyzed in the method designated above.

Figure S2: Results of leaching experiments of TVA coal ash from the Kingston spill site in TN. Concentrations of arsenic, selenium, boron, and strontium are in ppb normalized to the ash weight in the experiments. The pH values were determined in the leachate solutions after 24 hours of reactions.

Table S1: Major ion composition data (mg/L) of pore water and Cove surface water investigated in this study.

Table S2: Trace element composition (mg/L) of pore water and Cove surface water investigated in this study.

Table S3: Acid volatile sulfide [97] data in units µmol/g (dw) of river sediments obtained at Emory River and Clinch River sampling sites. Values represent the average ± 1 standard deviation (n = 2-3). N/A indicates sample was not available. Limit of quantification was 0.01 µmol/g (dw).

Appendix C- Supplementary Information Chapter 4

Figure S1: Arsenic concentration plotted against chloride concentration in Hyco and Mayo Lakes. The average outfall concentration (with standard deviation) is the red circle, the lake is blue squares, and the porewater is black diamonds. The concentration of arsenic is lower in the lake and outfall relative to the porewater indicating its removal and subsequent remobilization to the porewater during anoxic conditions.

Figure S2: Hyco Lake arsenic, iron, and manganese concentration compared to dissolved oxygen. Low dissolved oxygen conditions were found in deeper water near the bottom of the lake during thermal stratification.

Table S1: Table of ions in North Carolina bodies and the CCR effluent discharged into it.

Table S2: Table of the internal outfalls that contribute to the total outfall sampled at Hyco Lake.

100

Table S3: Table of Total As concentration and percentage as As(III) in the porewater at Hyco, Mayo, and Jordan lakes.

101

Table S4: Table of As and Se concentrations in sediment and fish tissue from Hyco and Mayo Lakes. This data was compiled from the Progress Energy Annual Reports (Hyco: 2008, 2009, 2010 and Mayo 2010).

102

References

[1] International Energy Association. Power generation from Coal: Measuring and reporting efficiency performance and CO2 emissions. 2008 .

[2] Lombardi, K., Coal Ash: The hidden story; How Industry and the EPA Failed To Stop a Growing Environmental Disaster 2009, http://www.publicintegrity.org/articles/entry/1144.

[3] Tennessee Valley Authority, Corrective action plan for the TVA Kingston fossil plant ash release, 2009, http://www.tva.gov/kingston/admin_record/pdf/G/G4.pdf.

[4] Ruhl, L., Vengosh, A., Dwyer, G. S., Hsu-Kim, H., Deonarine, A., Bergin, M. and Kravchenko, J. Survey of the potential environmental and health impacts in the immediate aftermath of the coal ash spill in Kingston, Tennessee. Environmental Science & Technology 2009, 43. 6326–6333.

[5] Environmental Protection Agency. Coal Combustion Waste Damage Case Assessments. 2007, Office of Solid Waste, EPA-HQ-RCRA-2006-0796.

[6] Mardon, S. M. and Hower, J. C. Impact of coal properties on coal combustion by-product quality: examples from a Kentucky power plant. International Journal of Coal Geology 2004, 59. 153-169.

[7] Wang, T., Wang, J., Tang, Y., Shi, H. and Ladwig, K. Leaching Characteristics of Arsenic and Selenium from Coal Fly Ash: Role of Calcium Energy & Fuels 2009, 23. 2959-2966.

[8] Punshon, T., Seaman, J. C. and Sajwan, K. S. The Production and Use of Coal Combustion Products. Kluwer Academic Pub2003, 1-12.

[9] Yudovich, Y. E. and Ketris, M. P. Arsenic in coal: a review. International Journal of Coal Geology 2005, 61. 141-196.

[10] Swaine, D. J. and Goodarzi, F., Environmental aspects of trace elements in coal, 1995, Kluwer Academic Publishers. Dordrecht, The Netherlands.

103

[11] Ziemkiewicz, P. F., Simmons, J. S. and Knox, A. S. The Mine Water Leaching Procedure: Evaluating the Environmental Risk of Backfilling Mines with Coal Ash. Kluwer Academic Pub2003, 75-90.

[12] Vassilev, S. V. and Vassileva, C. G. Occurrence, abundance, and origin of minerals in coals and coal ashes. Fuel Proccessing Technology 1996 48. 85-106.

[13] Finkelman, R. B. Modes of occurrence of trace elements in coal. 1981.

[14] Vassilev, S. V. and Vassileva, C. G. Mineralogy of combustion wastes from coal-fired power stations. Fuel Proccessing Tech 1996 47. 261-280.

[15] Tishmak, J. K. and Burns, P. E. The chemistry and mineralogy of coal and coal combustion products. Geological Society2004, 236. 223-246.

[16] Valkovic, V., Trace elements in coal:volumes I and II, 1983, CRC Press, Inc. Boca raton, FL.

[17] Zubovic, P., Stadnichenko, T. and Sheffey, N. B. The association of minor elements with organic and inorganic phases of coal. 1960, paper, U. p., Washington, DC.

[18] Kaakinen, J. W., Jorden, R. M., Lawasani, M. H. and West, R. E. Trace element behavior in coal-fired power plant. Environ. Sci. Technol., 1975, 9. 862–869.

[19] Smith, R. D., Campbell, J. A. and Neilson, K. K. Characterization and formation of sub- micron particles in coal-fired power plants. . Atm. Env. 1979. , 13. 607-617.

[20] Smith, R. D. The Trace-Element Chemistry of Coal During Combustion and the Emissions from Coal-Fired Plants. Prog. Energy Combust. Sci. 1980, 6. 53-119.

[21] Summers, D. V., Rupp, G. L. and Gherini, S. A. Physical-Chemical Characteristics of utility solid wastes. 1983, EA-3236.

[22] Van Der Sloot, H. A. N., B.J.T. Release of Trace elements from surface-enriched fly ash in seawater. Wastes Ocean 1985, 4. 449-465.

104

[23] Coles, D. G., Ragaini, R. C., Ondov, J. M., Fisher, G. L., Silberman, D. and Prentice, B. A. Chemical Studies of Stack Fly-Ash from a Coal-Fired Power Plant. Environmental Science & Technology 1979, 13. 455-459.

[24] Smith, R. D., Campbell, J. A. and Nielson, K. K. Concentration-Dependence upon Particle- Size of Volatilized Elements in Fly-Ash. Environmental Science & Technology 1979, 13. 553- 558.

[25] Hansen, L. D., Silberman, D., Fisher, G. L. and Eatough, D. J. Chemical Speciation of Elements in Stack-Collected, Respirable-Size, Coal Fly-Ash. Environmental Science & Technology 1984, 18. 181-186.

[26] Tazaki, K., Fyfe, W. S., Sahu, K. C. and Powell, M. Observations on the Nature of Fly-Ash Particles. Fuel 1989, 68. 727-734.

[27] Vassilev, S. V. Trace-Elements in Solid-Waste Products from Coal Burning at some Bulgarian Thermoelectric-Power Stations. Fuel 1994, 73. 367-374.

[28] Khanra, S., Mallick, D., Dutta, S. N. and Chaudhuri, S. K. Studies on the phase mineralogy and leaching characteristics of coal fly ash. Water Air Soil Pollut. 1998, 107. 251-275.

[29] Jankowski, J., Ward, C. R., French, D. and Groves, S. Mobility of trace elements from selected Australian fly ashes and its potential impact on aquatic ecosystems. Fuel 2006, 85. 243- 256.

[30] Spears, D. A. and Lee, S. Geochemistry of leachates from coal ash. Geological Society of London2004, 236. 619-639.

[31] Ondov, J. M. and Heft, R. E. Element compostion of submicrometer particles from coal combustion. Trans. Of the American Nuclear Society 1982, 41. 195-196.

[32] Haynes, B. S., Neville, M., Quann, R. J. and Sarofim, A. F. Factors Governing the Surface Enrichment of Fly-Ash in Volatile Trace Species. J. Colloid Interface Sci. 1982, 87. 266-278.

[33] Spears, D. A., Tarazona, M. R. M. and Lee, S. Pyrite in UK Coals- Its Environmental Significance. Fuel 1994, 73. 1051-1055.

105

[34] Querol, X., Fernandez-Turiel, J. L. and Lopez-Soler, A. Trace elements in coal and their behavior during combustion in a large power station. Fuel 1995, 71. 331-343.

[35] Sajwan, K. S., Coal combustion byproducts and environmental issues, 2006, Springer. New York.

[36] Hower, J. C., Robl, T. L., Anderson, C., Thomas, G. A., Sakulpitakphon, T., Mardon, S. M. and Clark, W. L. Characteristics of coal combustion products (CCP's) from Kentucky power plants, with emphasis on mercury content. Fuel 2005, 84. 1338-1350.

[37] Hower, J. C., Robl, T. L. and Thomas, G. A. Changes in the quality of coal combustion byproducts produced by Kentucky Power Plants, 1978 to 1997: consequences of Clean Air Act directives. Fuel 1999, 78. 701-712.

[38] Davidson, G. R. and Bassett, R. L. Application of boron isotopes for identifying contaminants such as fly ash leachate in groundwater. Environmental Science & Technology 1993, 27. 172-176.

[39] Vengosh, A., Heumann, K. G., Juraske, S. and Kasher, R. Boron isotope application for tracing sources of contamination in groundwater. Environmental Science & Technology 1994, 28. 1968-1974.

[40] Bassett, R. L., Buszka, P. M., Davidson, G. R. and Damaris, C. D. Identiﬁcation of groundwater solute sources using boron isotopic composition. Environmental Science & Technology 1995, 29. 2915-2922.

[41] Hogan, J. F. and Blum, J. D. Boron and lithium isotopes as groundwater tracers: a study at the Fresh Kills Landfill, Staten Island, New York, USA. Appl. Geochem. 1999, 18. 615-627.

[42] Williams, L. B. and Hervig, R. L. Boron isotope composition of coals: a potential tracer of organic contaminated fluids. Appl. Geochem. 2004, 19. 1625-1636.

[43] O'Neill, M. S., Veves, A., Zanobetti, A., Sarnat, J. A., Gold, D. R., Economides, P. A., Horton, E. S. and Schwartz, J. Diabetes enhances vulnerability to particulate air pollution - Associated impairment in vascular reactivity and endothelial function. Circulation 2005, 111. 2913-2920.

[44] Spivak-Birndorf, L. J. and Stewart, B. W. Use of Boron Isotopes to Track the Interaction of Coal Utilization Byproducts with Water in the Environment. 2006 . 106

[45] Faure, G., The principles of Isotope Geology, 1986, Wiley. New York, NY.

[46] Campbell, C. E., Pearson, B.N., and Frost, C.D. Strontium isotopes as indicators of aquifer communication in an area of coal-bed natural gas production, Powder River Basin, Wyoming and Montana. Rocky Mountain Geology 2008, 43 171–197.

[47] Hurst, R. W., Davis, T. E. and Elseewi, A. A. Strontium isotopes as a tracer of coal combustion residue in the environment. Engineering Geology 1991, 30.

[48] WV DIvision of Culture and History, Buffalo Creek Disaster, 2009, 2/10/2009. http://www.wvculture.org/history/buffcreek/bctitle.html.

[49] Smith, S. MSHA Assesses Maximum Fines for Martin County Sludge Spill.

[50] Meij, R. Trace elements behavior in coal fired power plants. Fuel 1994, 39. 199-217.

[51] Cornelis, G., Johnson, C. A., Gerven, T. V. and Vandecasteele, C. Leaching mechanisms of oxyanionic metalloid and metal species in alkaline solid wastes: A review. Appl. Geochem. 2008, 23. 955-976.

[52] Hower, J. C., Sakulpitakphon, T., Trimble, A. S., Thomas, G. A. and Schram, W. H. Major and minor element distribution in a coal-fired utility boiler in KY. Energy Sources 2006, 28. 79- 95.

[53] Sakulpitakphon, T., Hower, J. C., Trimble, A. S., Schram, W. H. and Thomas, G. A. Arsenic and mercury partitioning in fly ash at a Kentucky power plant. Energy & Fuels 2003, 17. 1028-1033.

[54] Hower, J. C., Graham, U. M., Dozier, A., Tseng, M. T. and Khatri, R. A. Association of the sites of heavy metals with nanoscale carbon in a Kentucky electrostatic precipitator fly ash. Environmental Science & Technology 2008, 42. 8471-8477.

[55] Rowe, C. L., Hopkins, W. A. and Congdon, J. D. Ecotoxicological Implications of Aquatic Disposal of Coal Combustion Residues In The United States: A Review. Environmental Monitoring and Assessment 2002, 80. 207-276.

[56] Elseewi, A. A., Page, A. L. and Grimm, S. R. Chemical characterization of fly-ash aqueous systems. Journal of Environmenal Quality 1980, 9. 424-428. 107

[57] Zielinski, R. A. and Budhan, R. B. Radionuclides in fly ash and bottom ash: improved characterization based on radiography and low energy gamma-ray spectrometry. Fuel 1998, 77. 259-267.

[58] Kovler, K., Perevalov, A., Steiner, V. and Metzger, L. A. Radon exhalation of cementitious materials made with coal fly ash: Part 1- scientific background and testing of the cement and fly ash emanation. Journal of Environmental Radioactivity 2005, 82. 321-334.

[59] Mahur, A. K., Kumar, R., Mishra, M., Sengupta, D. and Prasad, R. An Investigation of Radon Exhalation rate and estimation of radiation doses in coal and fly ash samples. Applied Radiation Isotopes 2008, 66. 401-406.

[60] Cevik, U., Damla, N., Koz, B. and Kaya, S. Radiological characterization around the Afsin- Elbistan coal-fired power plant in Turkey. Energy & Fuels 2008, 22. 428-432.

[61] Wilde, F. D. R., D.B.; Gibs, J.; Iwatsubo, R.T. Processing Water Samples (Vesion 2.1). 2004, Book 9, Chapter A5.

[62] Environmental Protection Agency. Mercury in Solids and Solutions by Thermal Decomposition, Amalgamation, and Atomic Absorption Spectrophotometry; Method 7473. 1998, USEPA, Washington, DC.

[63] Sanchez, F. K., R.; Kosson, D.: Delapp, R. Characterization of mercury-enriched coal combustion residues from electric utilities using enhanced sorbents for mercury control. 2006 .

[64] Campbell, K. R., Ford, C. J. and Levine, D. A. Mercury distribution in Poplar Creek, Oak Ridge, Tennessee, USA. Environmental Toxicology and Chemistry 1998, 17. 1191-1198.

[65] Burger, J. and Campbell, K. R. Species differences in contaminants in fish on and adjacent to the Oak Ridge Reservation, Tennessee. Environmental Research 2004, 96. 145-155.

[66] Environmental Protection Agency, National Primary Drinking Water Regulations: Maximum Contaminant Levels 2009, January. http://www.epa.gov/safewater/contaminants/index.html.

[67] Environmental Protection Agency, National Recommended Water Quality Criteria, 2009, July 15. http://www.epa.gov/ost/criteria/wqctable/.

108

[68] Droppo, I. G. and Jaskot, C. Impact of River transport on Contaminant Sampling Error and Design. Environmental Science & Technology 1995, 29. 161-170.

[69] Fischer, H. B. Longitudinal dispersion and turbulent mixing in open-channel flow. Annual Reviews of Fluid Mechanics 1973, 59-78.

[70] Noel, J. D., Biswas, P. and Giammar, D. E. Evaluation of a sequential extraction process used for determining mercury binding mechanisms to coal combustion byproducts. Journal of the Air & Waste Management Association 2007, 57. 856-867.

[71] Haitzer, M., Aiken, G. R. and Ryan, J. N. Binding of mercury(II) to dissolved organic matter: The role of the mercury-to-DOM concentration ratio. Environmental Science & Technology 2002, 36. 3564-3570.

[72] Khwaja, A. R., Bloom, P. R. and Brezonik, P. L. Binding constants of divalent mercury (Hg2+) in soil humic acids and soil organic matter. Environmental Science & Technology 2006, 40. 844-849.

[73] Cruz-Guzman, M., Celis, R., Hermosin, M. C., Leone, P., Negre, M. and Cornejo, J. Sorption-desorption of lead (II) and mercury (II) by model associations of soil colloids. Soil Science Society of America Journal 2003, 67. 1378-1387.

[74] Gilmour, C. G., Henry, E. A. and Mitchell, R. Sulfate stimulation of mercury methylation in freshwater sediments. Environmental Science & Technology 1992, 26. 2281-2287.

[75] Compeau, G. C. and Bartha, R. Sulfate-reducing bacteria: Principal methylators of mercury in anoxic estuarine sediment. Applied and Environmental Microbiology 1985, 50. 498-502.

[76] Hopkins, W. A. Effects of coal combustion wastes on survival, physiology, and performance of the benthic-feeding ﬁsh, Erimyzon sucetta. University of South Carolina2001, PhD Dissertation.

[77] Darmenova, K., Sokolik, I. N. and Darmenov, A. Characterization of east Asian dust outbreaks in the spring of 2001 using ground-based and satellite data. Journal of Geophysical Research-Atmospheres 2005, 110.

[78] Reynolds, L., Jones, T. P., BeruBe, K. A. and Richards, R. Toxicity of airborne dust generated by opencast coal mining. Mineralogical Magazine 2003, 67. 141-152.

109

[79] Linak, W. P., Yoo, J. I., Wasson, S. J., Zhu, W., Wendt, J. O. L., Huggins, F. E., Chen, Y., Shah, N., Huffman, G. P. and Gilmour, M. I. Ultrafine ash aerosols from coal combustion: Characterization and health effects. Proceedings of the Combustion Institute 2007, 31. 1929- 1937.

[80] Iordanidis, A., Buckman, J., Triantafyllou, A. G. and Asvesta, A. Fly ash - airborne particles from Ptolemais-Kozani area, northern Greece, as determined by ESEM-EDX. International Journal of Coal Geology 2008, 73. 63-73.

[81] Nicholson, K. W. A Review of Particle Resuspension. Atmospheric Environment 1988, 22. 2639-2651.

[82] Harris, A. R. and Davidson, C. I. A Monte Carlo Model for Soil Particle Resuspension Including Saltation and Turbulent Fluctuations. Aerosol Science and Technology 2009, 43. 161- 173.

[83] Wilson, W. E. and Suh, H. H. Fine particles and coarse particles: Concentration relationships relevant to epidemiologic studies. Journal of the Air & Waste Management Association 1997, 47. 1238-1249.

[84] Teixeira, E. C., Samama, J. C. and Brun, A. Study of the Concnetration of Trace-Elements in Fly-Ash Resulting from Coal Combustion. Environmental Technology 1992, 13. 995-1000.

[85] Blaha, U., Sapkota, B., Appel, E., Stanjek, H. and Rosler, W. Micro-scale grain-size analysis and magnetic properties of coal-fired power plant fly ash and its relevance for environmental magnetic pollution studies. Atmospheric Environment 2008, 42. 8359-8370.

[86] Rose, N. L. Inorganic fly-ash spheres as pollution tracers. Environmental Pollution 1996, 91. 245-252.

[87] Sui, J. C., Xu, M. H., Du, Y. G., Liu, Y., Yu, D. X. and Yi, G. Z. Emission characteristics and chemical composition of PM10 from two coal fired power plants in China. Journal of the Energy Institute 2007, 80. 192-198.

[88] Bhanarkar, A. D., Gavane, A. G., Tajne, D. S., Tamhane, S. M. and Nema, P. Composition and size distribution of particules emissions from a coal-fired power plant in India. Fuel 2008, 87. 2095-2101.

110

[89] Jones, T., Blackmore, P., Leach, M., Berube, K., Sexton, K. and Richards, R. Characterisation of airborne particles collected within and proximal to an opencast coalmine: South Wales, UK. Environmental Monitoring and Assessment 2002, 75. 293-312.

[90] IARC. Overall evaluation of carcinogenicity: an updating of IARC Monographs Volumes 1 to 42. 1987.

[91] Lyman, G. H., Lyman, C. G. and Johnson, W. Association of Leukemia with Radium Groundwater Contaminantion. Jama-Journal of the American Medical Association 1985, 254. 621-626.

[92] Shifrine, M., Fisher, G. L. and Taylor, N. J. Effect of Trace-Elements Found in Coal Fly- Ash, on Lymphocyte Blastogenesis. Journal of Environmental Pathology Toxicology and Oncology 1984, 5. 15-24.

[93] Costa, D. L. and Dreher, K. L. Bioavailable transition metals in particulate matter mediate cardiopulmonary injury in healthy and compromised animal models. Environmental Health Perspectives 1997, 105. 1053-1060.

[94] Goldsmith, C. A. W., Hamada, K., Ning, Y. Y., Qin, G. Z., Catalano, P., Murthy, G. G. K., Lawrence, J. and Kobzik, L. Effects of environmental aerosols on airway hyperresponsiveness in a murine model of asthma. Inhalation Toxicology 1999, 11. 981-998.

[95] Proctor, S. D., Dreher, K. L., Kelly, S. E. and Russell, J. C. Hypersensitivity of prediabetic JCR : LA-cp rats to fine airborne combustion particle-induced direct and noradrenergic-mediated vascular contraction. Toxicological Sciences 2006, 90. 385-391.

[96] Becker, S., Soukup, J. M. and Gallagher, J. E. Differential particulate air pollution induced oxidant stress in human granulocytes, monocytes and alveolar macrophages. Toxicology in Vitro 2002, 16. 209-218.

[97] Poliakova, V. A. S., V.A.; Tereshchenko, V.P.; Bazyka, D.A.; Golovnia, O.M.; Rudavskaia, G.A. Invasion of Micro-organisms in Bronchial Mucosa of liquidators of the Chernobyl accident consequences Mikrobiology 2001, 63. 41-50.

[98] Chizhikov, V., Chikina, S., Gasparian, A., Zborovskaya, I., Steshina, E., Ungiadze, G., Samsonova, M., Chernyaev, A., Chuchalin, A. and Tatosyan, A. Molecular follow-up of preneoplastic lesions in bronchial epithelium of former Chernobyl clean-up workers. Oncogene 2002, 21. 2398-2405.

111

[99] Vengosh, A. R., L; Dwyer, G.S. Possible environmental effects of the coal ash spill at Kingston, Tennessee. Phase I: Preliminary results Duke University2009.

[100] Swaine, D. J. Trace-elements in coal and their dispersal during combustion. Fuel Processing Technology 1994, 39. 121-137.

[101] Swaine, D. J. Environmental aspects of trace-elements in coal. Environmental Geochemistry and Health 1992, 14. 2-2.

[102] Kashiwakura, S., Kubo, H., Kumagai, Y., Matsubae-Yokoyama, K., Nakajima, K. and Nagasaka, T. Removal of boron from coal fly ash by washing with HCl solution. Fuel 2009, 88. 1245-1250.

[103] Dutta, B. K., Khanra, S. and Mallick, D. Leaching of elements from coal fly ash: Assessment of its potential for use in filling abandoned coal mines. Fuel 2009, 88. 1314-1323.

[104] Skodras, G., Grammelis, P., Prokopidou, M., Kakaras, E. and Sakellaropoulos, G. Chemical, leaching and toxicity characteristics of CFB combustion residues. Fuel 2009, 88. 1201- 1209.

[105] Fulekar, M. H. and Dave, J. M. Heavy-metals release from ash ponds to soil-water environment - A simulated technique Environment International 1992, 18. 283-295.

[106] Warren, C. J. and Dudas, M. J. Leaching behaviour of elected trace-elements in chemically weathered alkaline fly-ash. Science of the Total Environment 1988, 76. 229-246.

[107] US Geological Survey, National Field Manual for the Collection of Water-Quality Data, 2008, Jan 1. http://water.usgs.gov/owq/FieldManual/index.html.

[108] Bednar, A. J., Garbarino, J. R., Ranville, J. F. and Wildeman, T. R. Preserving the distribution of inorganic arsenic species in groundwater and acid mine drainage samples. Environmental Science & Technology 2002, 36. 2213-2218.

[109] Allen, H. E., Fu, G. and Deng, B. Analysis of acid volatile sulfide (AVS) and simultaneously extracted metals (SEM) for the estimation of potential toxicity in aquatic sediments. Environmental Toxicology and Chemistry 1993, 12. 1441-1453.

112

[110] Bowles, K. C., Bell, R. A., Ernste, M. J., Kramer, J. R., Manolopoulos, H. and Ogden, N. Synthesis and characterization of metal sulfide clusters for toxicological studies. Environmental Toxicology and Chemistry 2002, 21. 693-699.

[111] US Gelogical Survey, USGS Real- Time Data for 03540500 Emory River at Oakdale, TN, 2010, http://waterdata.usgs.gov/usa/nwis/uv?site_no=03540500.

[112] Thorneloe, S. A., Kosson, D. S., Sanchez, F., Garrabrants, A. C. and Helms, G. Evaluating the fate of metals in air pollution control residues from coal-fired power plants. Environ. Sci. Technol. 2010, 44. 7351-7356.

[113] Hesbach, P. A., Kim, A. G., Abel, A. S. P. and Lamey, S. C. Serial batch leaching procedure for characterizaton of coal fly ash. Environ Monit Assess 2010, 168. 523-545.

[114] Drott, A., Lambertsson, L., Bjorn, E. and Skyllberg, U. Importance of dissolved neutral mercury sulfides for methyl mercury production in contaminated sediments. Environmental Science & Technology 2007, 41. 2270-2276.

[115] Francendese, L. Memorandum: Dredging Determination 3/5/10. 2010, .

[116] Tennessee Valley Authority. TVA Water Testing Results. 2009, http://www.tva.gov/kingston/water/results2.pdf .

[117] Tennessee Valley Authority. Maximum Value Sampling Results. 2010, http://www.tva.gov/kingston/water/max_results.pdf..

[118] Tennessee Valley Authority, Continuing Investigation of the Nature and Extent of Ash in the Emory, Clinch and Tennessee River Bottoms, 2009, http://www.tva.com/kingston/ash_distribution.pdf. .

[119] ERDCWES, Summary of ERDCWES Model Simulations, 2010, http://www.epakingstontva.com/Nature%20and%20Extent/Summary%20of%20Storm%20Events .doc.pdf. .

[120] Adriano, D. C., Page, A. L., Elseewi, A. A., Chang, A. C. and Straughan, I. Utilization and disposal of fly-ash and other coal residues in terrestrial ecosystems - A review. Journal of Environmental Quality 1980, 9. 333-344.

113

[121] Landsberger, S., Cerbus, J. F. and Larson, S. Elemental characterization of coal ash and its leachates using sequential extraction techniques. Journal of Radioanalytical and Nuclear Chemistry-Articles 1995, 192. 265-274.

[122] Huggins, F. E., Helble, J. J., Shah, N., Zhao, J., Srinivasachar, S., Morency, J. R., Lu, F. and Huffman, G. P. Forms of occurrence of arsenic in coal and their behavior during coal combustion Abstr. Pap. Am. Chem. Soc. 1993, 205. 12.

[123] Huggins, F. E., Huffman, G. P., Miller, C. A. and Linak, W. A. Leaching and XAFS chracterization pf PM 2.5 from combustion of U.S. coals. 2003 .

[124] Huggins, F. E., Senior, C. L., Chu, P., Ladwig, K. and Huffman, G. P. Selenium and Arsenic Speciation in Fly Ash from Full-Scale Coal-Burning Utility Plants. Environmental Science & Technology 2007, 41. 3284-3289.

[125] Shoji, T., Huggins, F. E., Huffman, G. P., Linak, W. P. and Miller, C. A. XAFS spectroscopy analysis of selected elements in fine particulate matter derived from coal combustion. Energy & Fuels 2002, 16. 325-329.

[126] Plant, J. A., Kinniburgh, D. G., Smedley, P. L., Fordyce, F. M. and Klinck, B. A. Arsenic and Selenium. Elsevier-Pergamon2003, 9 17-66.

[127] Wolthers, M., Charlet, L., van Der Weijden, C. H., van der Linde, P. R. and Rickard, D. Arsenic mobility in the ambient sulfidic environment: Sorption of arsenic(V) and arsenic(III) onto disordered mackinawite. . Geochim. Cosmochim. Acta 2005, 69 3483-3492.

[128] Belzile, N. and Tessier , A. Interactions between arsenic and iron oxyhydroxides in lacustrine sediments. Geochimica Cosmochimica Acta 1990, 54. 103-109.

[129] Dzombak , D. A. and Morel, F. M. M., Surface complexation modeling – Hydrous Ferric Oxide 1990 Wiley New York .

[130] Bar Yossef, B. and Meek, D. Selenium sorption by kaolinite and montmorillonite Soil Sci. 1987, 144 12-19.

[131] White, A. and Dubrovsky, N. Chemical oxidation-reduction controls on selenium mobility in groundwater systems. 1994, Chp 8. 185-221.

114

[132] Duker, A. A., Carranza, E. J. M. and Hale, M. Arsenic geochemistry and health. Environment International 2005, 31. 631-641.

[133] Environmental Protection Agency. Method 1311: Toxicity Characteristic Leaching Procedure. 1992, Agency, E. P .

[134] Environmental Protection Agency. Steam Electric Power Generating Point Source Category: Final Detailed Study Report. 2009, EPA 821-R-09-008.

[135] Energy Information Administration. Count of Electric Power Industry Power Plants, by Sector, by Predominant Energy Sources within Plant, 2002 through 2010. Energy Information Administration2011, 2010.

[136] American Coal Ash Association, 2008 Coal Combustion Product (CCP) Production and Use Survey Report, 2009, http://www.acaa-usa.org.

[137] Environmental Protection Agency, Wastes- Non-Hazardous Waste- Industrial Waste, 2012, Agency, U. E. P. April 15. http://www.epa.gov/osw/nonhaz/industrial/special/fossil/coalashletter.htm.

[138] Hanlon, J. Memorandum: National Pollutant Discharge Elimination System (NPDES) Pennitting of Wastewater Discharges from Flue Gas Desulfurization (FGD) and Coal Combustion Residuals (CCR) Impoundments at Steam Electric Power Plants . 2010, Office of Waste Water Management, Washington, DC .

[139] Department of Energy. Power Plant Water Usage and Loss Study. 2007.

[140] Environmental Protection Agency, NPDES Permit Program Basics: Fequently Asked Questions, 2012, Agency, E. P. April15. http://cfpub.epa.gov/npdes/faqs.cfm#107.

[141] Codes of Federal Regulations. Steam Electric Power Generating Point Source Category. 1982 .

[142] North Carolina Department of Environment and Natural Resources. Permit to discharge wastewater under the National Pollutant Discharge Elimination System (NPDES). NC; DENR; DWQvarious .

115

[143] Progress Energy, Air and Water Resources- Progress Energy Carolinas, 2012, April 15. https://www.progress-energy.com/commitment/environment/what-we-are- doing/airandwater.page.

[144] Progress Energy. Montly Data Monitoring Report (DMR) for Progress Energy Carolinas' Roxboro Steam Plant NPDES NC0003425 2009-2011 .

[145] Badireddy, A. R., Hotze, E. M., Chellam, S. and Wiesner, M. R. Toxicity of Fullerol Nanoparticles to Bacteriophages. Envirionmental Science and Technology 2007, 41. 6627-6632.

[146] Balistrieri, L. S., Murray, J. W. and Paul, B. The Cycling of Iron and Managnese in the Water Column of Lake Sammamish, Washington Limnol. Oceanogr. 1992, 37. 510-528.

[147] Lemly, A. D. Symptoms and implications of selenium toxicity in fish: the Belews Lake case example. Aquat. Toxicol. 2002, 57. 39-49.

[148] Bool III, L., E. and Helble, J. J. A laboratory study of the partitioning of trace elements during pulverized coal combustion. Energy & Fuels 1995, 9. 880-887.

[149] Weres, O., Jaouni, A.-R. and Tsao, L. The distribution, speciation, and gochemical cycling of selenium in a sedimentary environment, Kesterson Reservoir, California, U.S.A. . Appl. Geochem. 1989, 4. 543-563.

[150] Yan, X., Kerrich, R. and Hendry, M. J. Distribution of arsenic(III), arsenic(V) and total inorganic arsenic in porewaters from a thick till and clay-rich aquitard sequence, Saskatchewan, Canada. Geochim. Cosmochim. Acta 2000, 64. 2637-2648.

[151] Progress Energy. 2009 Hyco Lake Annual Report (reported to NC DENR). 2009.

[152] Progress Energy. 2010 Mayo Lake Annual Report (reported to NC DENR). 2010 .

[153] Progress Energy. 2010 Hyco Lake Annual Report (reported to NC DENR). 2010 .

[154] Hopkins, W. A., Roe, J. H., Snodgrass, J. W., Staub, B. P., Jackson, B. P. and Congdon, J. D. Effects of chronic dietary exposure to trace elements on banded water snakes (Nerodia fasciata). Environmental Toxicology and Chemistry 2002, 21. 906-913.

116

[155] US Gelogical Survey. USGS Gage 03447687 French Broad River Near Fletcher, NC 2012.

[156] Ruhl, L., Vengosh, A., Dwyer, G. S., Hsu-Kim, H. and Deonarine, A. Environmental Impacts of the Coal Ash Spill in Kingston, Tennessee: An 18-Month Survey. Environmental Science & Technology 2010, 44. 9272-9278..

[157] Hower, J. C., Trimble, A. S., Eble, C. F., Palmer, C. and Kolker, A. Characterization of fly ash from low-sulfur and high-sulfur coal sources: partitioning of carbon and trace elements with particle size. Energy Sources 1999, 21. 511-525.

[158] Ruhl, L., Vengosh, A., Dwyer, G. S., Hsu-Kim, H. and Schwartz, G. E. The Impact of Coal Combustion Residual Effluent on Water Resources: A North Carolina Case Study. (submitted to journal) 2012 .

[159] Dwyer, G. and Vengosh, A. Alternative filament loading solution for accurate analysis of boron isotopes by thermal ionization mass spectrometry. 2008 .

[160] Klein, D. H., Andren, A. W., Carter, J. A., Emery, J. F., Feldman, C., Fulkerson, W., Lyon, W. S., Ogle, J. C., Talmi, Y. A., Vanhook, R. I. and Bolton, N. E. Trace-Element Measurements at Coal-Fired Allen Steam Plant Mass Balance and Concentrations in Fly Ash. Abstracts of Papers of the American Chemical Society 1975, 50-51.

[161] Natusch, D. F. S. Physiochemical Associations of Trace Contaminants in Coal Fly Ash. Abstracts of Papers of the American Chemical Society 1975, 52-52.

[162] Buszka, P. M., Fitzpatrick, J., Watson, L. R. and Kay, R. T. Evaluation of Ground-Water and Boron Sources by Use of Boron Stable-Isotope Ratios, Tritium, and Selected Water Chemistry Constituents near Beverly Shores, Northwwestern Indiana, 2004. 2004 .

[163] Spivak-Birndorf, L. J., Stewart, B. W., Capo, R. C., Chapman, E. C., Schroeder, K. T. and Brubaker, T. M. Strontium Isotope Study of Coal Utilization By-Products Interacting with Environmental Waters. Journal of Environmental Quality 2012, 41. 144-154.

[164] Mattigod, S. V., Rai, D. and Fruchter, J. S. Strontium Isotopic Characterization of Soils and Coal Ashes. Appl. Geochem. 1990, 5. 361-365.

117

[165] Mc Arthur, J., Howarth, R. and Bailey, T. Strontium Isotope Stratigraphy: LOWESS version 3: Best Fit Marine Sr-Isotope curve for 0-509 Ma and accompanying look-up table for deriving numerical age. Journal of Geology 2001, 109. 155-170.

[166] Schwarcz, H. P., Agyei, E. K. and McCmulle.Cc. Boron Isotopic Fractionation During Clay Adsorption from Sea-Water. Earth Planet. Sci. Lett. 1969, 6. 1-&.

[167] Spivack, A. J. and Edmond, J. M. Boron Isotope Exchange between Seawater and the Ocean-Crust. Geochim. Cosmochim. Acta 1987, 51. 1033-1043.

[168] Palmer, M. R., Spivack, A. J. and Edmond, J. M. Temperature and pH Controls Over Isotopic Fractionation During Adsorption of Boron on Marine Clay. Geochim. Cosmochim. Acta 1987, 51. 2319-2323.

[169] Spivack, A. J., Palmer, M. R. and Edmond, J. M. The Sedimentary Cycle of Boron Isotopes. Geochim. Cosmochim. Acta 1987, 51. 1939-1949.

[170] National Institute of Standards and Technology, Critical Stability Constants of Metal Complexes Database, 1993 .

118

Biography

Laura, a third generation gator, received her BS and MS in Geological Sciences from the

University of Florida in 2006 and 2008, respectively. In 2008, Laura began her PhD at Duke

University. While working toward her PhD at Duke, she advised the Beta Rho chapter of Kappa

Alpha Theta and volunteered at several elementary, middle, and high schools presenting on geology topics. She was invited to present her research at the EPA’s Toxics Release Inventory

Annual Meeting in 2010. She was awarded honorable mention in 2011 from the Nicholas School

Dean’s Award for Outstanding Student Manuscript. She was also awarded the best student presentation at the 2011 MedGeo conference in Bari, Italy. After completing her PhD, Laura will begin an assistant professor position at University of Arkansas at Little Rock.

119