Citation
Ultraviolet Activated Mercury Removal for Wastewater with Selected Competing Ligands

Material Information

Title:
Ultraviolet Activated Mercury Removal for Wastewater with Selected Competing Ligands
Creator:
Rogers, Ana Maria
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (200 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Environmental Engineering Sciences
Committee Chair:
MAZYCK,DAVID W
Committee Co-Chair:
CHADIK,PAUL A
Committee Members:
BONZONGO,JEAN-CLAUDE J
BOYER,TREAVOR H
MA,LENA Q
Graduation Date:
8/6/2016

Subjects

Subjects / Keywords:
Alkalies ( jstor )
Chlorides ( jstor )
Electrons ( jstor )
Irradiation ( jstor )
Ligands ( jstor )
Mercury ( jstor )
Oxidation ( jstor )
pH ( jstor )
Sulfides ( jstor )
Wastewater ( jstor )
Environmental Engineering Sciences -- Dissertations, Academic -- UF
dom -- mercury -- photo-oxidation -- photocatalysis -- treatment
Suwannee River, FL ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Environmental Engineering Sciences thesis, Ph.D.

Notes

Abstract:
This work investigates photochemical processes that enable Hg removal from wastewater in which the roles of photo-active sulfide, organic matter, and chloride determine Hg phase distribution. The findings meet the objective to further develop a photochemical treatment method for Hg removal termed ultraviolet activated chelation (UVAC), which uses high intensity ultraviolet irradiation at 254 nm (UV-C) followed by sub-micron filtration. The results from UV bench-scale and pilot UVAC studies showed removal of dissolved ppb Hg down to low ppt Hg levels (0.45 micormeters filterable Hg complexes following greater than or equal to 30 minutes UV-C irradiation. UVAC Hg removal was further enhanced for pH 7 and 11 solutions, relative to pH 3, which produced effluent concentrations below the instrumental detection limit in solutions with thiol-rich HA (<6.0 ppt Hg). The observations indicate that while multiple photo-redox reactions acted competitively, photo-degradative pathways that targeted organic binding ligands altered Hg speciation from its soluble form. In order to further develop UVAC as a viable Hg removal technology for variable wastewater chemistry, it is important to quantify binding ligand action in photochemical valence transfer within soluble Hg complexes. In organic-rich Hg wastewaters with sulfide, this possibly includes 1) complexation with sulfhydryl ligand to produce soluble Hg-DOMRS and Hg-S-DOM; 2) photo-activation of ionized phenolic and carboxylic groups as reducible moieties that oxidize Hg(0), and 3) formation of filterable Hg complexes by interactions with lower molecular weight organic and mineralized carbon. If UVAC photo-chemical processes that form filterable Hg occur when photo-activated DOM functional groups act as oxidizing reactants, then oxidation reduction potential may be a design variable determined by solution pH and UV-C irradiation period. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2016.
Local:
Adviser: MAZYCK,DAVID W.
Local:
Co-adviser: CHADIK,PAUL A.
Statement of Responsibility:
by Ana Maria Rogers.

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright Rogers, Ana Maria. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Classification:
LD1780 2016 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

1 ULTRAVIOLET ACTIVATED MERCURY REMOVAL FOR WASTEWATER WITH SELECTED COMPETING LIGANDS By ANA MARIA ROGERS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2016

PAGE 2

2 © 2016 Ana Maria Rogers

PAGE 3

3 To my parents

PAGE 4

4 ACKNOWLEDGMENTS I would like to first thank my advisor, David Mazyck, for his encouragem ent and guidance throughout this tremendous opportunity to conduct fulfilling work in this area of water treatment research. I a m immensely grateful for all his su pport, trust and patience, which helped me grow into a stronger person as well as a better e ngineer; and I am proud to have been a member of his team. I would like to thank each of my co chairs: Jean Claude Bonzongo, Trevor Boyer, Paul Chadik, and Lena Ma. I am grateful for the time and energy generously devoted to guiding my research efforts th rough sound and learned advice on experiment al and theoretical studies; I am especially grateful for the teachings that have made more a more capable research scientist and for those that have made me a better person. I thank my research group, especially my supportive mentors, Anna CasasÂœs, Heather Byrne, and mentoring colleagues, Katherine Deilz and Akua Oppong Anane; my lab and research assistant and especially Maria Arroyo, whose dedication to experimental research was an important contribution. I than k the ESSIE department staff and the graduate editorial committee for all their support and assistance; I especially thank Berdenia Monroe, Melissa Centurrion and Barbi Jackson from EES. I also would like to t hank the University of Florida l ibraries for th eir tireless efforts to provide access to research generated by the scientific community worldwide. I am grateful for the privilege of receiving grants and scholarships that supported my professional de velopment and degree completion: I thank the Universit y of Florida Alumni Association, the office of Graduate and Minority Program and the Fulbright Fellowship foundation . I thank Anne Donnelly, Samesha Barnes and NSF SEAGEP for providing exciting learning opportunities and research assistant support.

PAGE 5

5 I woul d like to thank each of my friends and family members for all their support and encouragement over the years. I thank Damian Rogers for his un reserved encouragement to meet challenges and grow from them in the process of working towards my goals, and I am thankful for the blessing of spending our lives together. I am especially grateful for the unconditional support of my parents, Robert and Wilma Hagan, who’s actions and advice guide me to seek a purpose that may serve the greater good of our environment and humanity. F inally, I thank God for staying by my side.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ... 4 ! LIST OF TABLES ................................ ................................ ................................ ............. 9 ! LIST OF FIGURES ................................ ................................ ................................ ......... 11 ! LIST OF ABBREVIATIONS ................................ ................................ ............................ 14 ! ABSTRACT ................................ ................................ ................................ .................... 15 CHAPTER 1 PROBLEM STATEMENT ................................ ................................ ........................ 17 ! Photochemical Treatment of Hg Wastewater ................................ .......................... 18 ! Development of UVAC Treatment ................................ ................................ ........... 19 ! Hypotheses ................................ ................................ ................................ .............. 21 ! Objectives ................................ ................................ ................................ ................ 21 ! Organization ................................ ................................ ................................ ............ 22 ! 2 LITERATURE RESEARCH ................................ ................................ ..................... 23 ! Mercury Contamination in the Aqueous Environment ................................ ............. 23 ! Aqueous Mercury Speciation and Solubility ................................ ............................ 26 ! Sulfide in Mercury Speciation ................................ ................................ ............ 2 8 ! Organic Matter in Mercury Speciation ................................ ............................... 30 ! Formation of Solid Phase Hg DOM ................................ ................................ ... 33 ! Formation of Soluble Aqueous Hg DOM ................................ ........................... 34 ! Mercury Methylation ................................ ................................ .......................... 35 ! Aqueous Mercury Photochemistry ................................ ................................ ........... 36 ! Photo transformation Processes ................................ ................................ ....... 38 ! General Mercury Photochemistry ................................ ................................ ...... 40 ! Nitrate Photochemistry ................................ ................................ ...................... 41 ! Sulfide Photochemistry ................................ ................................ ..................... 43 ! Chloride Photochemistry ................................ ................................ ................... 44 ! Organic Matter Photochemistry ................................ ................................ ......... 47 ! Mercury Transformations by Photo Active Organic Matter ............................... 50 ! 3 METHODS AND MATERIALS ................................ ................................ ................. 56 ! Solution Preparation and Chemical Reagents ................................ ......................... 56 ! Batch Reactor Experiments ................................ ................................ ..................... 58 ! Experimental Analysis and Characterization ................................ ........................... 60 ! Detection of Mercury and Analysis ................................ ................................ .... 61 !

PAGE 7

7 Estimation of Experimental Error ................................ ................................ ...... 64!4 DEVELOPMENT OF ULTRAVIOLET ACTIVATED CHELATION TECHNOLOGY ................................ ................................ ................................ ........ 67!Scope ................................ ................................ ................................ ....................... 67!Chlor-alkali Use of Mercury ................................ ................................ ............... 68!Conventional Treatment Effluent Characteristics ................................ .............. 69!Materials and Methods ................................ ................................ ............................ 70!Mercury Removal Definition ................................ ................................ .............. 71!Measuring Redox ................................ ................................ .............................. 72!Chlor-alkali Batch Sample pH ................................ ................................ ........... 72!Investigation of Aqueous Phase Hg Speciation ................................ ....................... 72!UVAC Design Parameters ................................ ................................ ....................... 75!Experimental Results ................................ ................................ ............................... 76!UVAC Mercury Removal ................................ ................................ ................... 76!UVAC Redox Potential ................................ ................................ ...................... 77!Investigation of Possible UVAC Processes ................................ ............................. 80!Effect of Adjusted Solution pH ................................ ................................ .......... 80!Effect of Solution Temperature ................................ ................................ ......... 81!Effect of Sample-Age ................................ ................................ ........................ 84!Investigation of Possible Photo-Reactants ................................ .............................. 85!Hg-Chloride Interactions ................................ ................................ ................... 85!Hg-DOM Interactions ................................ ................................ ........................ 87!Hg-Sulfide Interactions ................................ ................................ ...................... 91!Summary ................................ ................................ ................................ ................. 94!5 DETERMINATION OF UV EFFECT ON AQUEOUS MERCURY REMOVAL BY SULFIDE PRECIPITATION ................................ ................................ ................... 115!Scope ................................ ................................ ................................ ..................... 115!Experimental Parameters ................................ ................................ ................ 115!Mercury Sulfide Solutions ................................ ................................ ............... 115!UV Experimental Results ................................ ................................ ....................... 115!Trace Sulfide ................................ ................................ ................................ ... 116!Residual Sulfide ................................ ................................ .............................. 117!UV Wavelength and Sulfide Photo-Reactivity ................................ ................. 118!Investigation of UV Effects on Hg-S Dissolution ................................ .................... 119!Radical Scavengers ................................ ................................ ........................ 121!Summary ................................ ................................ ................................ ............... 122!6 MIXED LIGAND EFFECTS ON MERCURY PHOTO-TRANSFORMATIONS ....... 132!Scope ................................ ................................ ................................ ..................... 132!Experimental Parameters ................................ ................................ ................ 132!Investigation of Possible Hg Species ................................ .............................. 133!UV Experimental Results ................................ ................................ ....................... 134!

PAGE 8

8 Excess Sulfide in DI Water Solutions ................................ .............................. 134 ! DOM with Trace Sulfide in pH Buffered Solutions ................................ .......... 135 ! Investigation of P hoto Reactive Mercury Species ................................ ................. 136 ! Summary ................................ ................................ ................................ ............... 139 ! 7 SUMMARY OF FINDINGS ................................ ................................ .................... 148 APPENDIX A EXPERIMENTAL PARAMETERS ................................ ................................ ......... 152 ! UV / UVAC Batch Reactor Parameters ................................ ................................ . 152 ! Photometric Output Calculations ................................ ................................ ........... 153 ! Mercury Removal Calculations ................................ ................................ .............. 154 ! Batch Reactor UV Heat Transfer ................................ ................................ ........... 156 ! OR P Probe Temperature Correction ................................ ................................ ..... 157 ! B SUPPORTING DATA ................................ ................................ ............................ 159 ! UVAC treated Chlor alkali Wastewater ................................ ................................ . 159 ! Hg Removal by Heat and Filtration ................................ ................................ . 160 ! Hg Removal by Heated UVAC Treatment ................................ ....................... 161 ! Effect of UV C a nd pH on Oxidation Reduction Potential ............................... 163 ! Effect of UV C on Solution pH ................................ ................................ ......... 165 ! UV with Synthetic Mercury Wastewater ................................ ................................ . 166 ! DOM Photo Transformations ................................ ................................ ................. 168 ! Sulfide Speciation Calculations ................................ ................................ ............. 169 ! C LITERA TURE REFERENCE DATA ................................ ................................ ....... 171 ! DOM Photo reactivity and Activation ................................ ................................ ..... 171 ! UV Spectra Characterization of Sulfur Species ................................ ..................... 174 ! Natural Organic Matter Characterization ................................ ............................... 175 ! UV Wavelength and DOM Photo activation ................................ .................... 175 ! Dissolved Organic Matter Fraction Classification and Possible Components . 176 ! IHSS Organic Matter Characterization ................................ ............................ 177 ! D OM functional group charge density calculations ................................ ... 177 ! Estimations of DOM electron donating and accepting capacities ............. 179 ! DOM elemen tal composition ................................ ................................ ..... 179 ! Carbon distribution in DOM functional groups ................................ .......... 180 ! LIST OF REFERENCES ................................ ................................ .............................. 182 ! BIOGRAPHICAL SKETCH ................................ ................................ ........................... 200 !

PAGE 9

9 LIST OF TABLES Table page 2-1 Equilibrium constants for cinnabar solubility product in the pres ence of polysulfides occurring as rhombic S(0). ................................ ............................... 53!2-2 Octanol-water partitioning coefficient, KOW, for soluble Hg complexes (Hg(II)(aq)) and other environmentally relevant species. ................................ ...... 53!2-3 Constants of formation (Log K) and solubility product (Log Ksp) for Hg-S in the absence of polysulfide for pH and sulfide concentration constraints. ............ 54!3-1 Acid and base reagents used to prepare pH -buffered solutions. ........................ 66!3-2 Operational definitions that characterize Hg during experimentation. ................. 66!4-1 Water quality for chlor-alkali samples A and C ................................ ................... 96!4-2 Visual MINTEQ input criteria used to predict inorganic Hg speciation in untreated chlor-alkali samples. ................................ ................................ ............ 97!4-3 UVAC-treated filtrate Hg concentration for chlor -alkali sample C. ....................... 97!4-4 Standard electrode potentials of relevant Hg and aqueous redox couples in solution at equilibrium. ................................ ................................ ......................... 98!5-1 Solution compositions for sulfide-free (B.0) and trace-sulfide (S.0) samples. ... 123 5-2 Experimental UV driven Hg removal results for B.0 and S.0 samples. ............. 124!5-3 SUVA-290 for IHSS SRHA and SRFA for 5, 30, and 60 min UV irradiation...... 125!5-4 UV-C *Hg losses for pH 7 buffered solution with SHRA and residual sulfide. ... 125 6-1 Hg(NO3)2 and mixed ligand solution composi tion with measured pH over 60 minutes UV-C contact time. ................................ ................................ ............... 140!6-2 HgCl2 and mixed ligand solution composition and pH -values measure with varying UV-C contact time. ................................ ................................ ................ 140!6-3 Molar compositions of prepared solutions of trace -sulfide and DOM (S1-S4)... 141 6-4 Calculated DOM charge density (meq/g-C) for pH-buffered solutions. ............. 142!6-5 Formation stability constants for Hg-DOM complexes in pH~8 solutions .......... 142!6-6 *Hg losses from UV-C in prepared Hg solutions as total Hg loss, *C/C0. .......... 142

PAGE 10

10 6-7 UVAC Hg removal results in variable pH-buffered solutions (S1-S4) ................ 143!A-1 Radial photometric energy per unit volume in UV/UVAC batch reactors. ......... 154!B-1 ANOVA results for heated UVAC re for chlor-alkali sample A. .......... 161! B-2 Eh (mV) values for pH dependent redox potential in chlor -alkali samples. ....... 165!B-3 Reagents used to create pH-buffered synthetic Hg wastewater solutions. ....... 167!B-4 Baseline UVAC Hg removal results (!g/L Hg) using 0.45 !m nylon filters. ....... 167!B-5 Measured Abs-254 (cm1) in UVAC treated chlor-alkali sample C. ................... 168 B-6 Changes in DOC concentration from irradiation by UV -C and UV -B. ................ 168!B-7 S/Hg ratio for C0 = 20 g/L Hg (0.1 M Hg) in pH 7 solutions. .......................... 169!B-8 Summary of S/Hg mass and molar ratios for pH 3, 7, and 11 solutions. ........... 169!B-9 Distribution summary for principle sulfide species for pH 3, 7, and 11. ............. 170!C-1 Photo-reactant activation energies for OH production in water. ....................... 173 C-2 Photo-reactant activation energies for H2O2 in water production. ..................... 173!C-3 Hydration energies and select optical transition energies for comparison of Hg-S complex formation and photo-dissociation. ................................ .............. 175!C-4 Classes of dissolved organic matter fractions. ................................ .................. 176!C-5 Identification of IHSS DOM isolate by name and acronym. ............................... 177!C-6 Acidic functional groups of IHSS DOM. ................................ ............................. 178!C-7 Metal concent and electron donating (EDC) and electron accepting (EAC) capacities of IHSS DOM. ................................ ................................ ................... 179!C-8 Composition of IHSS DOM based on elemental analysis. ................................ 179!C-9 Estimates of carbon distribut ion in IHSS DOM using 13CNMR. ......................... 180

PAGE 11

11 LIST OF FIGURES Figure page 2-1 Sources of anthropogenic Hg emissions into the aqueous environment. ............ 55!3-1 Schematic of UV/UVAC batch reactors. ................................ .............................. 66!4-1 Eh/pH diagram for an oxic system groundwater system. ................................ .... 99!4-2 Effect of pH on pore-size defined DOC in chlor-alkali sample C. ........................ 99!4-3 UVAC Hg removal by varying filter pore size in chlor -alkali sample A. ............. 100!4-4 UVAC effluent filtrate Hg (*CF) in chlor-alkali sample C for pH 3 to 11. ............ 100!4-5 Potentiometric Eh-pH results for chlor-alkali samples A and C. ........................ 101!4-6 Eh-UVC for chlor-alkali sample C with varying pH and contact time. ................ 101!4-7 Eh-UVC for comparison of solution pH in chlor-alkali sample A. ....................... 102!4-8 Eh-UVAC for Sample A1 pH-dependent for tUVC up to 80 minutes. .................. 103!4-9 Eh-UVAC for Sample A2 pH-dependent for tUVC up to 80 minutes. .................. 104!4-10 Regression analysis of Eh-UVAC Hg(s) removal in chlor-alkali sample A. ......... 105!4-11 UVAC Hg losses for pH-variable adjused chlor -alkali sample A. ...................... 106!4-12 Regression analysis of Eh-UVAC Hg(0) losses in chlor-alkali sample A .......... 107!4-13 Comparison of acid pH adjustment for UVAC in chlor -alkali sample C. ............ 108!4-14 Effects of solution temperature on Hg(0) solubility in UVAC batch reactors. .... 109!4-15 Effect of UV-C on pH for DI and chlor-alkali sample A, adjusted pH 3 to 11. .... 109!4-16 Heated UVAC Hg removal in chlor -alkali sample C, variable filter pore size. ... 110 4-17 UVAC Hg removal in chlor-alkali sample C with varying pH adjustment . .......... 111!4-18 Replicate results for Eh-UVAC Hg removal in chlor-alkali sample A. ................ 112 4-19 Replicate regression of Eh-UVAC Hg(0) losses in chlor-alkali sample A. ......... 113 4-20 DOM characterized by EEM fluoro-spectroscopy in chlor-alkali sample C. ...... 114!5-1 UV-C Hg removal in pH-buffered B.0 solutions prepared in DI water. .............. 126!

PAGE 12

12 5 2 UV C Hg removal in pH buffered S.0 solutions with trace sulfide. .................... 127 ! 5 3 UV C Hg removal for pH 7 buffered solutions with residual sulfide. .................. 128 ! 5 4 UV C Hg removal for pH 7 buffered solutions with molar equivalent sulfide. .... 129 ! 5 5 UV B Hg removal in pH 7 buffered solutions with molar equivalent sulfide. ..... 130 ! 5 6 UV C Hg removal in pH 7 buffered solutions with DOM and residual sulfide. ... 131 ! 6 1 UV C Hg removal for DI mi xed ligand solutions of S, Cl and DOM. .................. 144 ! 6 2 UV C Hg removal for pH 3 solutions with trace sulfide and DOM (S1 S4). ....... 145 ! 6 3 UV C Hg removal for pH 7 solutions with trace sulfide and DOM (S1 S4). ....... 146 ! 6 4 UV C Hg removal for pH 11 solutions with trace sulfide and DOM (S1 S4). ..... 147 ! A 1 Schematic of UV bulbs with d imensions (mm). ................................ ................. 152 ! A 2 UV C photometric output of Philips PL S 9W/12 G23 lamps. ............................ 152 ! A 3 UV B photometric output of Philips PL L 9W/01/2P lamps. ............................... 152 ! A 4 Batch reactor dimensions used to calculate photon flux. ................................ .. 153 ! A 5 Empirical derivation of UVAC batch reactor heat transfer from UV C bulbs. .... 156 ! A 6 Linear regression used for temperature correction for ORP electrode. ............. 158 ! B 1 UVAC Hg removal with HCl acid pH adjustment in chlor alkali sample C. ........ 159 ! B 2 Heat only Test (1) Hg removal in chlor alkali sample C by filtration alone. ....... 160 ! B 3 Heat only Test (2) Hg removal in chlor alkali sample C by filtration alone. ....... 160 ! B 4 UVAC Hg removal with variable solution temperature and 60 min UV C. ......... 162 ! B 5 UVAC Hg removal with variable solution temperature and 120 min UV C. ....... 162 ! B 6 Eh UV C for alkaline DI solution vs. chlor alkali sample A for UV C " 20 min. . 163 ! B 7 Unfiltered vs. filtered solution UVAC ORP readings in chlor alkali sample A. . .. 163 ! B 8 ORP results from acid titration with HCl in chlor alkali samples A and C. ......... 164 ! B 9 Changes in solution pH for chlor alkali sample A during UV C irradiation. ....... 165 ! B 10 UVAC Hg removal in DI solutions prepared with variable Hg spikes. ............... 166 !

PAGE 13

13 B 11 Free S 2 Ð ion pC pH diagram for total [S 2 Ð ] !M concentrations. ......................... 170 ! C 1 Estimated anthropogenic contribution to multi phasic gaseous and particulate Hg, represented as fluxes of Hg (Mg/ yr) and Hg reservoirs (Gg). ..................... 171 ! C 2 Schematic energy level diagram for excited state DOM. ................................ ... 172 ! C 3 UV spectra after 273 min photolysis for sulfur 6.25 mM solutions. .................... 174 ! C 4 IHSS DOM charge density vs solution pH from carboxyl and phenol groups. .. 178 ! C 5 Distribution of major elements in IHSS DOM based on elemental analysis. ..... 180 ! C 6 Percent carbon distribution in IHSS DOM functional groups using 13 CNMR. .... 181 !

PAGE 14

14 BDL DGM DOC DOM EAC ECC EDC EH FA HA Hg HMW HS IHSS LMCT LMW NEU/ CHA NOM ORP ROS SHA TOC UVAC VHA LIST OF ABBREVIATIONS Below detection limit Dissolved gaseous mercury Dissolved organic carbon Dissolved organic matter Electron accepting capacity Electron carrying capacity Electron donating capacity Oxidation reduction potential with standard correction Fulvic acid Humic acid Mercury as total Hg unless otherwise stated High molecular weight Humic substances International Humic Substances Society Ligand to metal charge transfer Low molecular weight Neutral hydrophilic acid and base Natural rganic atter Oxidation reduction potential Reactive oxygen species Slightly hydrophobic base and neutral Total organic carbon Ultraviolet activated chelation Very hydrophobic acid

PAGE 15

15 Abstract of Di ssertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ULTRAVIOLET ACTIVATED MERCURY REMOVAL FOR WASTEWATER WITH SELECTED COMPETING LIGANDS By Ana Maria Rogers August 2016 Chair: David W. Mazyck Major: Environmental Engineering Sciences This work investigates photochemical processes that enable Hg removal from wastewater in which the roles of photo active sulfide, organic matter, and chlo ride determine Hg phase distribution. The findings meet th e objective to further develop a photochemical treatment method for Hg removal termed ultraviolet activated chelation (UVAC), which uses high intensity ultraviolet irradiation at 254 nm (UV C) follo wed by sub micron filtration. B ench scale and pilot UVAC studies with secondary process chlor alkali wastewater (i.e., conventionally treated with su lfide) resulted in removal of dissolved ppb Hg down to low ppt Hg levels (<40.0 ppt Hg) . Lab e xperiments investigated UV Hg removal from pH adjusted DI solutions prepared with C 0 ~20 ppb Hg and complexing ligands of sulfide, chloride and dissolved organic matter (DOM). In the absence of UV C , Hg filte rability was enhanced by addition of molar equivalent sulfi de an d impaired by addition of humic acid (HA) alone. For solutions of soluble Hg prepared with trace sulfide and ~10 mg /L DOM, si multaneous photoreactions of sulfide and DOM oxidation removed ! 80% Hg as >0.45 !m filterable complexes following # 30 minute s UV C irradia tion . UVAC Hg removal was further enhanced for pH 7 and 11 solutions , relative to pH 3, which produced effluent

PAGE 16

16 concentrations below the instrumental detection limit (<6.0 ppt Hg) for solutions with # 5 mg C/L thiol rich HA . The observations indicate that wh ile multiple photo redox reactions act ed competitively , photo deg radative pathways that target ed organic bind ing ligands alter ed Hg speciation from its soluble form. In order to further develop UVAC as a viable Hg removal technology for va riable wastewater chemistry, it is important to quantify binding ligand action in photochemical valence transfer within soluble Hg complexes. In organic rich Hg wastewaters with sulfide, this possibly includes 1) complexation with sulfhydryl ligand as sol uble Hg DOM RS and Hg S DOM; 2) photo activation of ionized phenolic and carboxylic groups as reducible moieties that oxidize Hg(0), and 3) formation of filterable Hg comp lexes by interaction s with lower molecular weight organic and mineralized carbon. If UVAC photo chemical processes that form filterable Hg occur when photo activated DOM functional groups act as oxidizing reactants, then oxidation reduction potential may be a design variable determined by solution pH and UV C irradiation period.

PAGE 17

17 CHAPTER 1 PROBLEM STATEMENT Mercury (Hg) pollution of natural systems continues to be an issue of concern on a global scale, especially with respect to the contamination of food supplies. The efforts to curb Hg must take into account that the background chemical content of the waters requiring treatment are present in much higher concentrations than Hg (typically 3 to 6 orders of magnitude). Moreover, industrial waters will be different from site to site thus requiring robust technologies that are Hg selective and non excl usive. The current best available technology for Hg removal from many industrial wastewaters is referred to as secondary process treatment in which sulfide is added to Hg laden brine solution runoff and wash down water to produce HgS (s) precipitates that a re removed via filtration or possibly by settling of sufficiently large particles. While this method is suitable for brine wastewater w ith high chloride content (e.g., # 70 g L 1 Cl Ð ), it is not designed to remove some species of Hg, including elemental Hg(0) or Hg complexed with organic compounds, and produces average effluent Hg concentrations that range from about 5 to 60 µg L 1 Hg [1 3] . This type of effluent has Hg levels higher than U.S. EPA requirements for new Clean Water Act permitting as early as 2018: the Final Rule establishes zero Hg discharge limits for wastewater from steam electric plants (i.e. coal fired power) using method detection limit for total Hg analysis by EPA Method 245.7, ~2.0. ng L 1 [4] . Therefore, development of photo chemical treatment techn ology aims to meet the public safety goal of eliminating Hg waste discharge in the intensified effort to remove aqueous Hg across industrial process wastewaters.

PAGE 18

18 Photochemical Treatment of Hg Wastewater The use of UV light in advanced oxidation processes widely targets organic contaminants [5, 6] and yet, its use is emerging in wastewater treatment of multiple hazardous trace metals including, As(IV), Cr(VI), Hg(II), Pt(II) and U(VI,V) [7 10] . Previous research in our group on photochemical Hg treatment methods includes investigations of Hg photo reduction and photo catalytic a dsorption [11, 12] . Authors Byrne et al. focused on design processes for ppm level aqueous Hg removal using high photon energy from UV 254 nm (UV C) and photocatalysts such as TiO 2 . Its use in heterogeneous aqueous Hg removal has been identified as electron transfer that promotes Hg photo reduction reactions [13 15] , thus producing Hg(0) in the n et reaction in Eq ( 1 1 ) [16] . The Hg(0) aq can then react to adsorb onto the catalyst surface and is removed from solution, e.g., by filtration or settling. Similarly, composite catalysts ar e proven effective, such as silica titania composite photocatalysts (STC) developed for air phase chlor alkali Hg waste were proven effective in aqueous Hg removal [17 20] . ( 1 1 ) The continued research by our group investigated wastewater treatment that uses Hg photo reduction processes to convert aqueous Hg species into purge able Hg(0) aq so that it is both removed from solution by a carrier gas and efficiently captured in the air phase as gaseous Hg(0) g (e.g., adsorbed onto an activated carbon trap). The lab studies measured total Hg removal from prepared ppm level Hg solutio ns following batch experiments that employed artificial UV lamps to promote Hg photo reduction together with a continuous in situ gas purge designed promote Hg(0) volatilization and act as a phase transfer carrier gas. The formation rate of purgeable Hg(0 ) was H g ( a d s ) 2 + + H 2 O h ! T i O 2 " # " " H g ( a d s ) 0 + 2 H + + 1 2 O 2

PAGE 19

19 observed to depend on the experimental variables of UV wavelength and contact time, oxygen content of the purge carrier gas and the addition of dissolved organic matter in variable concentrations as humic acid or fulvic; it was observed that Hg photo reduction is favorable in anoxic solutions that contain photo active organic matter [11, 21] and the analysis on variabl e dependent Hg photo reduction supports the continuation of rese arch on the role of organic matter in natural systems [22] . Development of UVAC Treatment The previous research findings on photochemical treatment of Hg obtained by our group led to a university collaboration with a c hlor alkali plant employing the Hg cell process in order to develop industrial wastewater treatment technologies. This permitted further experimentation by our group using samples of secondary process wastewater effluent that had an average initial concen tration of 20 !g L 1 H g. In one lab study on the use of photocatalysts, the ojective was to optimize Hg photo reduction in order to increase subsequent adsorption of Hg(0) onto the suspended catalyst. The batch experiments used irradiation by UV C with a 15 minute contact time following filtration using 0.45 µm pore size. In order to test this, samples of secondary process effluent were prepared prior to UV using a three step process: 1) Reduce electron competition from dissolved oxygen (DO) by lowering co ncentrations to ~2 mg L 1 DO using nitrogen purge (Airgas , Inc. ); 2) Addition of a photocatalyst as 1000 mg L 1 of either powdered STC or TiO 2 (Degussa P25) (38 to 40 µm particle size); and 3) Addition of a weak organic acid, as citric acid (reagent grade monohydrate, Sigma Aldrich) to act as an electron donor by scavenging unpaired holes on the photo activated catalyst surface [14, 23] . Citric acid is low molecular weight carboxylic acid, an d was selected as the organic agent due to its versatile use in experimental research as a natural soils

PAGE 20

20 system model compound [24] , as an industrial chelating agent [25] as well as a reducing agent in heterogeneous photocatalysis discussed above. Two concentrations were tested, 10 and 20 mg L 1 citric acid (equivalent to 4 or 8 mg L 1 TOC) for solutions with initial concentrations of C 0 = 4.3 µg L 1 Hg. While the experiments resulted in over 95% removal in both photocatalytic systems, an interesting observation was made in control experiments with only citric acid and UV C: Hg removal was comparable to UV irradiated solutions that contained STC or TiO 2 . The findings from ad ditional control experiments on various batch samples determined that addition of citric acid and de oxygenation were not necessary, as UV alone followed by filtration resulted in practically identical removal . Furthermore, any photo produced Hg(0) that m ay have formed did not evolve from the solutions, which suggests that formation of filterable Hg is dependent on the availability of a precipitating ag ent such as residual sulfide. Th e process occurring by UV C irradiation followed by filtration of the sol ution was termed Ultraviolet Activated Chelation (UVAC) and led to the development of a consecutive series of hypotheses based on chlor alkali wastewater treatment results from two manufacturers producing varying effluent water quality: 1) UV oxidizes Hg t o a divalent state, promoting chelation with an organic agent in a process enhanced by a constituent in chlor alkali secondary process effluent and subsequently 2) UV oxidizes organic compounds originally chelated to the Hg, allowing ionic Hg to combine wi th residual sulfide to form HgS, which is insoluble and can be filtered out of solution [26] . In either case, the process is conceptually and theoretically unreported elsewhere; owing

PAGE 21

21 to its novelty, the f indings could not be explained by the current understanding of photochemical Hg treatment that is primarily driven by heterogeneous processes. Hypotheses 1. UVAC treatment removes ppb levels of aqueous Hg from secondary process wastewater to produce ppt level Hg effluent by using high intensity photon energy from UV 254 nm irradiation to convert soluble Hg into a solid phase that is filterable out of solution using 0.45 !m pore size filters. a) It is possible to identify optimal UVAC treatment design parameters of UV solution pH and UV contact time and intensity (mW cm 2 ) by quantifying their respective effects on the electrochemical redox potential of brine wastewaters that contains dissolved organic carbon. b) The relevant Hg transformations processes can be repl icated and investigated in lab studies using synthetic Hg wastewaters. 2. Photochemical oxidation and reduction reactions occurring photo reactive Hg binding ligands can both promote solid phase Hg and inhibit Hg(0) evolution in processes that are determined by photo oxidation reactions; and therefore, UVAC may be best described as an advanced photo oxidation process. a) UV C irradiation of organic and inorganic constituents determines Hg ligand complexation; the rate of reaction is dependent on solution pH, Hg l igand ratio and mixing conditions. b) UV promoted formation and dissolution of Hg S precipitates in sulfidic waters is determined by photo reactive of sulfide species. c) In photo activated chloride waters, Hg(II) reduced to Hg(0) may adsorb onto Hg S, which can be removed as a filterable species. d) In solutions with residual sulfide, longer mixing periods create stable Hg S DOM RS complexes, whereby solid phase Hg occurs by photo degradation (oxidation) of excess sulfide and DOM RS sulfhydryl groups. e) In the absence of residual sulfide, DOM enhance s formation of complexes and may be favorable in solutions that contain chloride. Objectives 1. Develop a current understanding of photochemistry for inorganic and organic constituents that are dominant in natural waters that c ould be also be relevant in Hg wastewater by means of literature research.

PAGE 22

22 a) Examine possible competing effects and governing reactions in photo mediated transformation of Hg into filterable solid phase species. 2. Evaluate UVAC Hg removal in chlor alkali waste water and possible reactants driving Hg photo transformations. a) Characterize Hg speciation in chlor alkali wastewater by investigating thermodynamic equilibrium conditions using measured water quality parameters. b) Monitor total Hg concentration and operatio nally defined fractions: filterable solid phase and filtrate soluble phase Hg species and volatized Hg(0). c) Identify key parameters that control UVAC reaction kinetics including: i) Inorganic and organic Hg ligand photochemistry. ii) Effect of UV C irradiation on wastewater reduction oxidation potential. 3. Identify a measure that can be further developed to predict the potential of UVAC Hg treatment in wastewaters of varying composition. a) Develop UV experiments using synthetic Hg wastewater solutions that support iden tification of possible mechanisms of UVAC Hg removal occurring by sulfide, chloride and dissolved organic matter. b) Identify determinants for photo activated reduction and oxidation potential in and relevant Hg wastewaters. Organization The first objective i s met in the discussion of literature research findings in Chapter 2, which is an overview of aqueous systems photochemistry that is relevant to Hg photo transformations . Objective two is met in Chapter 4 wit h the investigation of UVAC treatment results us ing chlor alkali Hg wastewater . The third objec tive is met in Chapters 5 and 6, which present experimental results on UV activated Hg removal from synthetic Hg wastewater containing inorganic and organic constituents. Chapter 7 is the summary of conclusi ons, which is followed by the Appendix that contains supportive experimental results from this work and literature reference material.

PAGE 23

23 CHAPTER 2 LITERATURE RESEARCH Mercury Contamination in the Aqueous Environment The persistence of mercury (Hg) manifests as a glob ally and continually circulating pollutant by existing in every environmental compartment, albeit in in trace concentrations, in solid, liquid and gaseous forms. Its multiphasic distribution creates challenges for Hg capture and environmental remediation due to its mobility by long range aqueous and atmospheric transport coupled with speciation transformations into methyl Hg (CH 3 Hg) forms that can bioaccumulate in fish and biota and ultimately biomagnify up the human food chain [27, 28] . It is known that methyl Hg toxicity has severe human health impacts that include irreversible neurological and autoimmune impairment human; yet, human health studies have shown that exposure to all physiochemica l forms of Hg can produce these toxic endpoints [29 31] , which is why ongoing research aims to further understand toxicity threshold levels for either acute expos ure to high levels (ppm Hg) or prolonged exposure to low levels (ppb Hg or less) [32, 33] . Ultimately, the total elimination of new anthropogenic Hg and remediation of environmental Hg a re challenges that must be met in order to reduce contamination by its methylated form and protect human health [47] (i.e., fish with methyl Hg). Current point sources of Hg wastewater discharge that contribute to aqueous Hg contamination occur from a rang e of industries including dentistry, oil and gas refinery, cement manufacturing and chlor alkali processing ( Figure 2 1 ). In 2012, these and other anthropogenic sources contributed to the estimated 220 tonnes of Hg discharged int o rivers, streams, lakes and the ocea n according to UNE's Global Mercury Assessment report. For example, up to 5.5% of current aqueous inputs of

PAGE 24

24 anthropogenic Hg is estimated to occur from chlor alkali plants that still employ Hg cell manufacturing to pro duce chemicals such as chlorine, hydrogen, and caustic soda [34] . Despite improvements in Hg removal through conventional processes and other efforts that span over more than a decade to prevent environmental Hg releases, the challenge to eliminate Hg from such complex industrial wastewaters must be met by the development of more robust wastewater treatment methods. An i ncreased awareness that anthropogenic Hg significantly increases human health risks of Hg toxicity has p rompted world wide efforts to develop and implement more stringent Hg standards for industrial releases as well as for standards that reduce Hg in water and wastewater; organizations leading these efforts include the World Health Organization (WHO), the U. S. Environmental Protection Agency (EPA) and the United Nations Environmental Programme (UNEP). An opportunity to successfully implement improved standards exists in the establishment of the Global Mercury Partnership, a multi national organization establ ished in 2013, which completed the creation of an enforceable international treaty called the Minamata Convention on Mercury in 2015 that will phasing out the manufacture, import or export of many Hg products by 2020 (current status of ratification and imp lementation available at www.mercuryconvention.org ). The treaty will also limit Hg contamination by controlling anthropogenic releases to air, water and land throughout its lifecycle, thus requiring developm ent of effective and robust Hg wastewater treatment methods, whose development is a important area of research. Global circulation of mercury . Mercury is poorly water soluble in its reduced form as liquid Hg(0) and has high volatility at ambient temperatu res, which facilitates

PAGE 25

25 both the initial emission from the Earth's surface and its long atmospheric circulation lifetime of approximately 6 to12 months before depositing onto the Earth's surface such as into ocean water s or at the earth's poles [35, 36] . The distribution of sources to the global Hg flux occurring as either natural or anthropogenic Hg has been estimated using analyses of 270 year old arctic glacial ice cores. According to the findings, Schuster et al. estimate that approximately half occurred from anthropogenic inputs (~52%) while the remainder occurred from environmental sources including water or sediment (~42%) and volcanic events (~6%), and that overall the atmospheric Hg deposition rates have decreased since the turn of the 21 st century [37] . These estimates were further refined in a study by Driscoll et al. that modeled anthropogenic contributions over the last 150 years [38] (Appendix Figure B 1 ). The cumulative health and environmental effects of Hg contamination are not yet fully understood, and may change with the gradual temperature rise of the ocean and atmospheric temperatures effects by increasing background diet levels. Its potential effects may already be evident in the findings by Fitzgerald and Sunderland et al. which reported that methyl Hg accumulation in the ocean at depths between 200 and 1000 m is occurring at a higher rate relative to a century ago, and that total Hg concentrations in this range is greater than in surface waters or in the zones below 1000 meters [39, 40] , thereby acting as a source of Hg exposure to marine predators and large fish species foraging in this zone (e.g., swordfish). Regulatory mandates. Stringent surface water Hg discharge limits were developed to mitigate methyl Hg productio n and accumulation in water bodies such as the Great Lake Initiatives Total Mean Daily Load (TMDL) standard of 1.3. in the ng L $ 1

PAGE 26

26 TMDL [41] . The EPA proposed the Clean Air Act's Mercury Rule (CAMR) in 2006 aiming to reduce utility emissions of mercury from 48 tons a year to 15 tons, a reduction of nearly 70% [42] , which was replaced in 2011 by the Final Rule on Mercury and Air Toxics Standards (MATS) [4] . As a result, promising technologies for gas phase Hg removal have become more apparent [19, 43 45] . Aqueous Hg removal technology development is expected to intensify following the EPA final ruling in 2015 to lower effluent Hg standards to non detec table levels for steam generating plants i.e. coal fired power (~2.0 ng L $ 1 ) [4] . This rule applies to Hg wastewater from wet flue gas desulphurization, which is the current maximum achievable control technology ( MACT) for gaseous Hg emissions from coal fired power . Aqueous Mercury Speciation and Solubility Whether occurring from atmospheric deposition or industrial point source contamination, the fate of Hg in aqueous systems determined by speciation transformatio ns that can cycle between its +2, +1 and 0 oxidation states: Hg(I), Hg(II), and Hg(0), respectively, occurring in equilibrium by Eq ( 2 1 ). While monovalent Hg(I) is considered a transition species, it can be an important species in less than alkaline solutions. The forms Hg(0) and Hg(II) are more stable and occur interchangeably according to Eq ( 2 2 ) [46] . The oxidation state of Hg will also determine the natur e of inorganic and organic Hg complexes formed and the extent of their environmental stability and solubility. Environmental conditions drive formation of highly volatile zero valence Hg as dissolved gaseous Hg 0 (aq) (DGM) and the evolved gaseous phase Hg 0 (g) according to the reaction in Eq ( 2 3 ) [46] . H g 2 2 + ! H g 2 + + H g 0 ( 2 1 )

PAGE 27

27 H g ( a q ) 0 + 2 e ! ! H g ( g ) 0 L o g K = 2 2 . 3 ( 2 2 ) H g ( a q ) 0 ! H g ( g ) 0 L o g K = 0 . 9 3 ( 2 3 ) In aqueous systems with natural organic matter (NOM), Hg has a predominant tendency to bind with S containing functional groups with the follow ing order of affinities for common ligands: S >I >Br >OH >Cl >N,O. The overall dissolved vs. insoluble distribution is determined by Hg that is (i) dissolved as a free ion or soluble complex (ii) by means of electrostatic attraction, nonspecifically adsor bed primarily as hydroxo complexes (iii) specifically adsorbed via covalent or coordinative bonding (iv) chelated with bi dentate organic ligan ds and (v) precipitated i.e. as sulfide or hydroxide [47] . The relative hydrophobicity of the Hg complex determines how Hg(II) (aq) partitions into dissolved phase NOM, which is measured using octanol water partitioning coefficients (K OW ) shown in Table 2 2 . Neutral chloro and sulfide Hg species, HgC l 2 0 (aq) and HgS 0 (aq) , are more hydrophobic relative to charged HgCl + (aq) and HgS 2 2 $ (aq) , which do not significantly partition into octanol (K OW ~0) [48] . Redox conditions . The stability of solid phase Hg species in aqueous systems is largely influenced by redox conditions. This is evident i n the findings of Lindsay and Schšndorf et al. who examined Hg solubility and mobility in contaminated groundwater from a former wood impregnation plant that used HgCl 2 . The remediation study used thermodynamic equilibrium calculations together with pH an d redox potential to predict aqueous Hg distribution among chloro , hydroxy and sulfide species. Their findings illustrated how this type of modeling is useful in remediation design when interpreting unexpected field sampling results: Analysis of reduct ion oxidation potential explained why samples predominantly measured inorganic Hg occurring as the neutral species

PAGE 28

28 Hg 0 instead of HgCl 2 [49] . To this end, while HgCl 2 is easily reducible, it can exist within a wide spectrum of electron activity seen in oxic and anoxic environments, e.g., in surface waters, Hg(II) not bonded to organic or sulfide ligands likely exist as hydroxyl and/or chloride complexes [50] . Sulfide in Mercury Speciation The background concentration of naturally occurring Hg in anoxic soil and water systems typically ranges from 0.01 to 0.1 nM Hg, and exists primarily as a sulfide complex that forms in systems containing at least pM sulfide concentration s (10 $ 12 ). The formation of a solid phase molecule of Hg S can produce precipitated polymorphs of metacinnabar, % HgS (s) (black) and cinnabar, & HgS (s) (red). Metacinnabar, whose formation is uncommon in natural systems, has a cubic crystalline structure, hig h temperature stability [61,141], and has a reported band gap energy from the valance band to conductance band of 1 eV [51] . Cinnabar, which more commonly forms in the environment, forms a trigonal crystalline struc ture that contains more S Hg S linkages [52] , and has a reported electron band gap energy of 2 eV [53, 54] . The two polymorphs have the common tendency to f orm covalent bonds by coordination with 2 or 4 sulfur ligands [55] , and both exhibit properties of n type semiconductors [56] . HgS (s) formation. The formation of Hg S comp lexes includes particles that can span both nano and micro size ranges, although solid phase is defined by particulate Hg retain ed on 0.45 µm pore size filter. Any Hg that passes through is labeled as soluble or dissolved Hg, which can include colloidal Hg or nano particulate HgS (s) (<50 nm ) [57] . The molecule HgS (s) is known to be hi ghly insoluble in both water and sediment and has been estimated to solubility product values range from Log sp values of 10 to 53 ; yet, there is ambiguity in phase definition of the aqueous molecule, Hg S (aq) , due to

PAGE 29

29 varying degrees of hydration wheneve r water molecules are present during the formation of molecular Hg S. Dissolved neutral and hydrophobic complexes such as Hg(SH) 2 0 or HgOHSH 0 (aq) form in slightly basic solutions that contain excess sulfide by means of hydration reactions described in Eq ( 2 4 ) [58] or ( 2 5 ) [59] . ( 2 4 ) ( 2 5 ) ( 2 6 ) ( 2 7 ) While HgOHSH 0 (aq) and its isomer, HgS ' H 2 O (aq) , are two soluble species, they differ in theoretical hydration e nergies that determine stability, complexation and photo transformation characteristics [60] . Furthermore, recent findings suggest that nano HgS (s) greatly alters thermodynamic models Hg S species formation . The au thors Dyrssen and Wedborg reported stability of HgOHSH 0 (aq) in equilibrium with HgS (s) in Eq ( 2 6 ) and ( 2 7 ) , has Log K sp from 22 to 10; although experimentally derived value of 10 was validated in abs ence of nano HgS (s) and at very low sulfide concentrations (0.06 to 1 µM S 2 $ ) [61] whe reas the value 22.3 is assumed valid in much more sulfidic waters (>85 µM S 2 $ ), and hence less likely to occur under natural conditio ns [58] . HgS (s) dissolution. Dissolution of HgS (s) can occur by dissolved sulfide in systems such as anoxic soils and sediments. Oxic solutions saturated with S 0 can produce polysulfides S n 2 $ and their protonate d forms (i.e. H x S n x $ 2 ) built of linear sulfur chains of one divalent sulfur atom and one or more S 0 atom s . HgS (s) has increased solubility in sulfidic waters containing S 0 due to the formation of charged Hg polysulfide H g 2 + + O H ! + H S ! " H g O H SH ( a q ) 0 L o g K = 3 0 . 3 H g 2 + + H S ! + H 2 O " H g O H SH ( a q ) 0 + H + L o g K = 2 6 . 7 H g S ( s ) + H 2 O ! H g O H SH ( a q ) 0 L o g K sp = " 2 2 . 3 H g S ( s ) + H 2 O ! H g S " H 2 O ( a q ) L o g K sp = # 1 0

PAGE 30

30 complexes ( Table 2 1 ), which was reported by Jay et al. to be 3 fold higher at neutral pH 7 and up to 200 fold more soluble in alkaline pH range 8 to 11. The authors found that in higher sulfide waters (1 9 mM S 2 $ ), soluble species occur as charged hydrophilic Hg po lysulfide species, primarily Hg(S n ) 2 2 $ followed by HgS n OH $ , and that HgS n HS $ species are only relevant at neutral pH 7. In contrast, at lower sulfide concentrations (10 90 µM S 2 $ ), the hydrated complex HgS n OH $ is more likely to occur with hydrophobic spec ies such as HgS 0 (aq) [62] . Organic Matter in Mercury Speciation In the natural environment Hg can adhere to surfaces and form covalent bonds with solid phase organic matter or NOM P as sediment and mineral oxides (e.g. Fe and Mn oxide) [63, 64] , whereas DOM can determine dissolved phase Hg speciation [65] . Aqueous environments can have variable qu antities of DOM, i.e., within the range of ~0.001 to 100 mg L 1 DOM, and is generally comprised of a heterogeneous mixture of complex molecules containing ligands that act as potential Hg binding sites. This complexation is due to variations in compositio n of the central heteroatom (i.e. , S or N), the structure surrounding the heteroatom (i.e. aliphatic, aromatic, or electron accepting/donating), and ligand stereochemistry [66] . Reduced sulfur groups in DOM . DOM with reduced sulfur groups, RS $ , originates from reducing environments through assimilation and reduction of sulfate ions [58] . Early investigations of Hg Ð ligand binding used spectroscopic evidence to identify RS $ as primary binding sites on DOM using extended X Ray absorption fine structure analysis (EXAFS) , which found that two electron donors exist in the first coordination shell of DOM: 1) a reduced sulfur group (RS $ or thiol ) , and 2) a mixed functional group c onsisting of oxygen, nitrogen, or sulfur atoms [67] . While total sulfur

PAGE 31

31 constitutes up to 3.6% in soil derived HS and up to 1.43% in aquatic HS, Xia et al. identified that RS $ represented 10% and 50% of total su lfur in soil derived and aquatic derived HS, respectively, using near edge structure adsorption spectroscopy (XANES) [68] . Further research by Drexel and Hesterberg et al. found that while the single strongest binding s ite for Hg on DOM is inarguably the RS $ group, at high Hg concentrations relative to DOM Hg ions will first saturate RS $ groups and then subsequently form relatively weaker bonds with liga nds containing available groups. Namely, carboxylic (RCOOH), phenolic (ROH), hydroquinone (RC=O) or amino (RNHR or RNH 2 ) [69, 70] . This suggests that only a small fraction of RS $ (i.e. ~1.6 to 2%) is involved in relatively stronger Hg DOM binding [ 71] . The distribution of these functional groups determines how DOM responds to water treatment processes, such as alum coagulation, leading to the development of system that classifies DOM fractions by their composition of different compounds [72] . A summary of typical classes of each compound found in Appendix Table B 4 shows four fractions: very (VHA) and slightly (SHA) hydrophobic, and charged (CHA) or neutral (NEU) hydrophilic. While a ll four fractions contain compounds that have been identified as relevant in Hg speciation, the distribution of hydrophobic fractions in Hg DOM complexes has been estimated as 50% on average, and can be as much 85 90% in Hg DOM complexes [65, 73] . The VHA and SHA fractions are molecularly composed of phenol groups and both aliphatic and aromatic carboxylic acid, which corresponds with higher reactivity towards oxidants [74] , and preferential sensitization by UV irradiation [75] An important factor that determines formation of Hg DOM complexes is the pH dependent availability of DOM binding sites; Hg c an bind with DOM at deprotonation

PAGE 32

32 sites that occur when the pH of the system is higher than the pKa of the DOM functional group. Carboxylic acids have low pKa values (pKa ~4.8) relative to other major functional groups of and act as important Hg binding li gands in water and soil systems that are classified as acidic in the pH range of ~6.1 to 7 [76] . Complexes could form when adsorbs onto the DOM molecule at the deprotonated site of the carboxyl group, which has be en cited as a possible explanation for reduced Hg mobilization in acidified soil systems [77] . Mercury complexation with dissolved organic m atter . The extent of Hg DOM binding with individual organic ligands is theo retically understood through modeled formation stability constants, Log K, which represent equilibrium conditions at a specific pH in the equation Hg 2+ + L n Ð ! HgL (n 2) Ð , where L n $ represents the fully ionized dissolved ligand with concentration [DOML] calculated at its pKa. Since numerous experimental methods exist for measuring Hg DOM binding strength it is challenging directly compare and validate reported values [78] ; however, plausible alignment on competitive binding for Hg Cl and Hg DOM RS Ð complexes has emerged from experimental research on modeling Hg contamination in natural systems. A study by Benoit et al. reported significant H g DOM complexation in fully aerobic and sulfide free surface waters (i.e. , <0.1 µM S) via ligand exchange reactions, whereas Hg DOM complexes destabilized below pH 7. The authors used DOM isolates from the Florida Everglades to test for competitive binding with HgCl 2 0 (aq) in an Hg Cl DOM system containing 0.01 M Cl $ (aq) and ide ntified the reactivity of the DOM binding site from an acidic functional group with high pK a , which is typical of thiol groups with values ranging 9 to 11. With this information and results from experiments using DOM concentrations

PAGE 33

33 of 0.24 to 0.48 mg L $ 1 DOM (f rom ~20 to 40 µM DOM), the authors developed Log K values of 20.6 and 21.8 for the hydrophobic and hydrophilic fractions, respectively, calculated using a pK a value of 10 for reduced sulfur groups found in both fractions [79] . While the authors initially prefaced the findings by stating that Hg S could outcompete Hg DOM in the Everglades sulfidic sediment pore water, related work identified that Hg S DOM complexes can drive Hg speciation instead of by Hg S al one when neutral Hg S partitions into DOM [80] . Formation of Solid Phase Hg DOM In oxic waters, the distribution of Hg strongly favors sorption onto suspended particulate organic matter (>0.45 µm) (NOM P ), followed by complexation with colloidal (~1 nm to ~1 µm) and then the dissolved phase ( <1 kDA) . Research findings demonstrate tha t significant factors affecting sorption of inorganic Hg onto NOM P include pH and complexing ligands, like chloride and sulfur species. DOM is reported to enhance NOM P metal sorptive properties [81, 82] and furthermore, lower pH conditions appear to create favorable conditions for Hg NOM P complexation, potentially by lowering the electrostatic repulsion of cations by surfaces on NOM P with a low point of zero charge [83] . In the formation of hydrophobic neutral complexes with higher k ow ( Table 2 2 ), Turner et al. found that primarily Hg DOM and the chloro complex, HgCl 2 0 , increased Hg sorption by NOM P [63] . The mechanisms driving Hg NOM P adsorption in saline environments remain poorly characterized and are thought to occur through electrostriction (i.e. "salting out") and/or coagulation; the effects are expected to be limited by th e formation of stable, non sorbing HgCl 2 (aq) and competitive ligand adsorption [83] . This mechanism could be a relevant to Hg phase transformation in

PAGE 34

34 wastewater with high chloride concentrations, like chlor a lkali and acid drainage waste. In the absence of NOM P , however, it is also possible for solid phase Hg DOM species to form as mixed ligand complexes such as Hg sulfide DOM complexes, which can occur via covalent bonding, oxidative complexation through adso rption, or possibly chelation via the mixed ligand bidentate binding configuration, i.e. Hg DOM RS$ DOM RX . Formation of Soluble Aqueous Hg DOM Current research presents a consensus that DOM mostly causes dissolution of particulate HgS (s) resulting in subseq uent mobility of free ionic Hg, thereby acting as a primary mechanism for its transport ; this is not always predictable, however. This is evident in the previous discussion on how Hg complexation with organic ligands varies in strength and occurs with NOM P as well as DOM via functional groups. Furthermore, Hg solubility depends on how the organic matter partitions between the aqueous and solid phase [80] , the molecular weight and aromaticity of the DOM , and on t he effect of competing ligands such as chloride [65, 84] . Size characterization studies do not agree on particle size in the definition of solid phase Hg DOM. In a study that use d a particle size cutoff of 0.45 µm to test for co precipitation of Hg with DOM isola tes from the Florida Everglades, Lu et al. found no effect on the formation of larger size aggregate Hg DOM [85] . On the other ha nd, thiol organic ligands can stabilize small colloids of HgS (s) ( <0.2 µm) and promote precipitation of nano HgS (s) clusters measuring approximately 0.02 µm while formation of larger HgS (s) particles was reportedly inhibited under the same conditions [57, 86] . While HgS (s) is very insoluble and much less volatile and leachable from NOM P in sediment than other forms of Hg (i.e. HgCl 2 (aq) ) it can dissolve with DOM by means of su rface complexation [71] . Hesterberg et al. examined how soluble Hg was distributed

PAGE 35

35 when both DOM and sulfide were present in an anoxic system. The authors determined that competing reactions led to formation o f both Hg S and Hg DOM complexes depending on Hg concentration, the most stable of which is primarily determined by the type of sulfide available [70] , which can include thiol, sulfide and polysulfide species. Whereas Miller et al. reported that neutral and hydrophobic Hg S DOM complexes increase solubility of Hg formed with S 2 $ and S 0 [80] . Dissolved gaseous aqueous mercury. Organic matter is capable of converting Hg into its soluble form as DGM through reacted Hg DOM complexes. The reduction reaction by DOM reportedly occurs via i ntra molecular ligand to metal charge transfer (LMCT), directly through electron withdrawal by Hg(II) at a rate determined by pH, dissolved oxygen, and chloride concentrations [87, 88] . This supports earlier findings by Gu et al. that relatively high levels of DOM can inhibit DGM formation through complexation, particularly under reducing conditions. It was reported that Hg(II) could be effectively reduced to Hg(0) with as little as 0.2 µg L 1 reduced HA, whereas production of Hg(0) is inhibited by complexation as HA concentration increases. Both reduction and complexation by DOM is likely prevalent in anoxic sediments and water. The authors suggested that initial physical sorption takes place followed by the S H bond cleavage or charge transfer from Hg(0) to HA, leading to the formation of Hg(II) HA complexes via H thiolate bonds. This transformation reaction reportedly occurs under strongly reducing conditions in which strong Hg DOM RS complexes form [89] . Mercury Methylation Once Hg is in water and sediments, less toxic elemental and inorganic Hg transforms to t he organic form, methylmercury (methyl Hg), as the result of microbial metabolization. Primary Hg methylators are sulfate reducing bacteria (SRB) that exist in

PAGE 36

36 anoxic aquatic and soil systems, such as sediment pore water. Bioavailability of Hg to the SRB is an important factor in controlling methylation potential. Experimentally derived k ow values ( Table 2 2 ) serve as a benchmark for Hg bioaccumulation potential in aquatic systems, where ionizable and non polar Hg species with k o w < 4 do not biomagnify within aquatic food webs (e.g. Hg 0 , HgCl 2 0 ) . Sulfide can influence the rate of SRB methylation two fold: by determining both the size distribution and the speciation of Hg molecules. This has been shown to occur by d issolved inorg anic Hg S molecules with high k ow such as neutral hydrophobic sulfide, HgS 0 (aq) , which have much higher capacity to partition into biological tissues relative to HgS (s) or charged complexes by passively diffusing through the phospholipid bilayer of SRB cel l membranes [90] . However, methyl Hg production from nano HgS(s) can be as much as 6 times greater than from micro HgS (s) [91] . Toxicity significance . While mercury can have profound and intricate effects on the human neurological and immune system, its speciation is one of the most important factors in determining its toxicity potential: All forms are nephrotoxic, while methylated Hg is a potent neuro toxicant, primarily af fecting the central nervous system by damaging DNA Ð making it especially toxic to infants during all developmental phases. Methylated Hg forms polar and ionizable compounds that are the most bioavailable species in aquatic systems for transport to the li pids of water respiring organisms where it can bio accumulate and subsequent ly biomagnify up wards in the food chain. Aqueous Mercury Photochemistry Photochemical Hg wastewater treatment. The use of UV irradiation is a growing field of interest in treatment technology development even though Hg photochemical transformations are not easily predicted in most industrial waters. While

PAGE 37

37 background literature on Hg photochemical wastewater treatment is still limited, useful findings on Hg photochemistry are drawn from the extensive field of ongoing study in natural systems; for example, how Hg photo transformations occurring in surface waters and top soils impact its biogeochemical cycling and toxicity [92 95] . It is important to consider that experimental research on Hg photo transformations in natural waters commonly uses simulated solar radiation (UV Solar) with wavelength 290 " ( " 400 nm to model the troposphere's sunlight spectrum which has intrinsic variability in intensity (available energy) depending on the hour of the day, time of year, altitude and latitude and air composition and quality . Previous research has also focused on low er intensity radiation found within the solar spectrum using UV A (( = 315 Ð 400 nm; E ave = 3.52 eV), which constitutes about ~94% of terrestrial UV, and UV B wavelengths (( = 280 Ð 315 nm; E ave = 4.18 eV) which account for the remaining ~6% [96] . Noting that the energy of a photon is inversely proportional to its wavelength by the Planck Einstein relation, a photochemical reaction could take place when one Einstein of radiant energy is absorbed per mol of reactant . Hence the knowledge gained from Hg photochemistry research using UV Solar in natural systems is applicable to engineering photochemical treatment for Hg wastewater using higher energy radiation from UV C , which has an average photon emissions energy ~ 4 .8 eV and average wavelength, ( ave = 254 nm (( = 100 to 280 nm; E = 4.43 to 12.4 eV). Previous research on Hg wastewater treatme nt research in our group focused on UV C in the design of photochemical treatment processes for aqueous Hg removal; photochemica l aqueous Hg removal as DGM was studied using continuous purge

PAGE 38

38 reactions with UV irradiation by examining the roles of DOM, dissolved oxygen, and UV wavelength [11, 12, 20, 21, 26] . Photo transformation Processes Aqueous photo reactants . Aqueous reactants produce excited state species upon photon ( hv ) absorption and energy emission with the potential to act as either oxidizing or reducing agents by accepting or donating electrons, respectively. The photo reactivity of a molecule is determined by system variables including irradiation wavelength, water temperature and depth, and ligand or quenching chemistry of the solution. Primary photochemical processes are those in which hv absorption results in direct electron transfer to a higher state within the molecule or transfer out of the molecule while secondary processes involve intermediate photo reactants. The hy droxyl radical • OH is one of the most reactive photochemically produced species with standard electrode oxidation potential of 2.80 V in acid solution and pKa value of 11.9 [97] . Although the mechanism of • OH production varies greatly by photo reactant, it has been proposed as a primary photo oxidant of Hg in natural waters [98] . An irradiated water molecule produces • OH by photo excitation and photo ionization in Eq ( 2 8 ) to ( 2 9 ) leading to reactive oxidative species (ROS) by the proposed mechanisms in Eq ( 2 10 ) to ( 2 14 ); • OH is also a product of photolysis of inorganic components like peroxide, nitrate and chloride, and photo active organic matter [99] . ( 2 8 ) ( 2 9 ) ( 2 10 ) H 2 O h v H 2 O ! " # " " O H $ + H + O H ! h v " # $ • O H + e ! 2 O H ! h v " # $ H 2 O 2 + 2 e !

PAGE 39

39 ( 2 11 ) ( 2 12 ) ( 2 13 ) ( 2 14 ) Measuring photo reactivity. A useful measure of a molecules photo reactivity is known as quantum yield, which is defined as the number of reactant molecules consumed or the number of molecules of an apparent pr oduct for a given flux of absorbed photons in the system (J hv ), defined in Eq ( 2 15 ) in which the hydroxyl radical is measured as the apparent product. The experimental • OH yield for UV Solar DOM photolysis () •OH ~1 to 6 * 10 11 M 1 s 1 ) is two orders of magnitude lower than for NO 3 $ ( ) •OH ~1.4* 10 9 M 1 s 1 ) [100] . A related measure, termed apparent activation energy (E a ), quantifies photo produced ROS by substrates such as DOM (discuss ed further in Appendix Table B 1 and Table B 2 ). ( 2 15 ) Photo reactant lifetime. The effective exposure time (also known as lifetime) for the • OH rad i cal is exceptionally short, with reported values ranging from 2.7 to 11 µs. While a wide discrepancy between theoretical and measured ROS lifetime values is due to differences in methodology, there is agreement that • OH has the longest lifetime relative t o other oxidative species H 2 O 2 and superoxide radical , O 2 • Ð . A recent study by Attri et al. examined ROS generation mechanisms using plasma initiated UV photolysis with photoemissions energy in the same range as UV C (~4.8 eV) and found that • OH H 2 O 2 h v ! " ! 2 • O H H 2 O 2 + • O H ! H O 2 • + H 2 O O 2 • ! + H + " H O 2 • H O 2 • + O 2 • ! H 2 O " # " " H 2 O 2 + O 2 ! a = R • O H J h v [ = ] [ R ] m o l / L u n i t o f t i m e

PAGE 40

40 lifetime in water is inversely related to its density ( + • OH ) , e.g., increasing from 3.15 to 3.92 µs for decreasing + • OH from 4.2 to 0.8 * 10 16 cm 3 [101] . General Mercury Photochemistry Across the wavelength spectrum, the 6s orbital of Hg is assumed to be the photo n acceptor orbital [102] that causes Hg molecular bonds to cleave and transfer electrons between photo activated Hg and another species; this sequence can produce Hg with 0, +1, or +2 oxidation states that depen ds on several parameters. For example, s olution pH and redox potential, photo reactive organic or inorganic bindings ligands, as well as the type and concentration of radical scavengers. Hg photo reduction . Hg photo activation can occur via photo ionizatio n that results in photo reduction of an Hg(II) compound. The photo reaction sequence for aqueous Hg(OH) 2 , developed by Nriagu is shown in Eq ( 2 16 ) through ( 2 18 ). In this process, the excited state Hg( OH) • forms aqueous Hg(OH), which is unstable under reducing conditions; the molecule will either cleave under photolysis to form Hg(0) and • OH , or accept electrons to form the final unreactive product species, depicted as the elemental form Hg(0) [103] . ( 2 16 ) ( 2 17 ) ( 2 18 ) In abiotic DGM production in con taminated lake waters, both UV Solar and visible spectrums contribute a maximum of 73% and 27%, respectively [104] . Previous H g ( O H ) 2 ( a q ) h v ! " ! H g ( O H ) • # " ! H g ( O H ) ( a q ) + • O H H g ( O H ) ( a q ) h v ! " ! H g 0 + • O H H g ( O H ) a q + H + + e ! " # " H g 0 + H 2 O

PAGE 41

41 research on natural systems demonstrates that both UV A and UV B wavelengths facilit ate Hg(0) photo oxidation although there appears to be multiple pathways responsible and therefore , is not likely restricted to lower intensity radiation. Hg photo oxidation . Similarly, photo oxidation of Hg(0), summarized by the net reaction in Eq ( 2 19 ), can be initiated by • OH acting with reactive Hg(OH) • OH as shown by reactions in Eq ( 2 20 ) to ( 2 21 ). The absorption band required for LMCT in naturally occurring Hg species, 2 50 " ( a " 400 nm, coincides with spectrum found in UV C and UV Solar in which the energy produced is enough to convert common ligands into photolytic reactants [103] . In photochemical wastewater treatment, ke y ligands in this process include nitrate and nitrite, sulfide, chloride and DOM. ( 2 19 ) ( 2 20 ) ( 2 21 ) Nitrate Photochemistry Hg(NO 3 ) 2 synthetic wastewater. Experimental methods in this research use Hg(NO 3 ) 2 stock to create artificial Hg wastewater solutions, in which nitrate • OH photo production may contribute to Hg photo transformations. For exam ple in acidic solutions in the range of 4 " pH " 6, Hg 2+ dissociates from NO 3 2 to form primarily fully hydrolyzed Hg(OH) 2 depicted by the reaction in Eq ( 2 22 ). Above pH 6, Hg(OH) 2 remains the dominant species, even with chlorid e ions at low concentrations [Cl $ ] " 10 5 M Cl $ [105] . H g 0 + h v ! " ! H g 2 + + 2 e # H g 0 + • O H ! " ! H g 2 + + O H # o r H g ( O H ) • H g ( O H ) • + • O H ! " ! H g ( O H ) 2 ( a q )

PAGE 42

42 ( 2 22 ) UV C can produce • OH from nitrite at a rate ~2.60 * 10 11 M 1 s 1 • OH and a fraction of that from nitr ate, ~0.015 * 10 11 M 1 s 1 • OH [106, 107] . Although nitrite is present at much lower concentrations, it has a greater effective hv absorption and higher ) a • OH across the UV spectrum [108] . Nitrate ions can undergo photolysis to form nitrite or weakly oxidizing radicals NO 2 • and O • $ according to the reaction in Eq ( 2 23 ). Photolysis of nitrite ions can then produce stro ngly oxidizing radical • OH via intermediates (NO • , and O • $ ); this dichotomous process was first described in reaction sequence proposed by Zepp et al. in Eq ( 2 24 ) to ( 2 25 ) [109] . ( 2 23 ) ( 2 24 ) ( 2 25 ) A recent study by Zhang et al. compared the effect of U V wavelength on Hg photo reduction in artificial waters examined the role of • OH scavengers and producers from ligands sources potassium nitrate, KNO 3 $ , and methanol, CH 3 OH. The findings reported increasing evolved Hg 0 by order of UV C < UV A < UV B; the authors conclude that shorter wavelength of UV C is more effective in NO 3 $ photolysis, which produces excess • OH capable of re oxidizing Hg(0). Howev er, the same study found that under H g ( N O 3 ) 2 H 2 O ! " ! ! 2 N O 3 2 # + H g 2 + H 2 O ! " ! ! H g ( O H ) 2 + H + N O 3 ! h v " # " N O 3 ! $ # N O 2 ! + O ( 3 P) o r N O 2 • + O • ! N O 2 • + O • ! H 2 O " # " " N O 2 • + • O H + O H ! N O 2 ! h v " # " N O • + O • ! H + $ # " " • O H

PAGE 43

43 UV A irradiation the addition of the organic molecule CH 3 OH negated the reported inhibitory effect of added KNO 3 $ , reportedly by scavenging all excess • OH [110] . For illustrativ e purposes, the reported inhibition of Hg(0) formation with nitrate is described by reaction Eq ( 2 26 ), which summarized the simultaneous occurrence of Hg photo activation reaction in Eq ( 2 17 ) and the re actions for NO 3 $ photolysis and OH • product formation in Eq ( 2 23 ) . ( 2 26 ) Sulfide Photochemistry Free sulfide ions, principally bisulfide, HS $ can be homogeneously photo oxidized to produce s ulfate, hydrogen gas, reactive sulfur nucleophiles (i.e. polysulfides, and thiosulfate), and free sulfur S(0). Free sulfur can become complexed as polysulfide anions such as disulfide ion, S 2 2 $ [111] . This occu rs when HS $ ions are subject to one electron oxidation leading to formation of HS • and S • $ shown in Eq ( 2 27 ) to ( 2 29 ) . ( 2 27 ) ( 2 28 ) ( 2 29 ) The stepwise photochemical process in water is described below in Eq ( 2 30 ) to ( 2 31 ) according to Linkous et al. who proposed reactions for adsorbed HS $ on the inner sleeve walls of a batch photo reactor system, HS $ (aq) ! HS $ (ads) . The absorption of photon produces the excited state, i.e. HS $ (ads) ! *HS $ (ads) , which is reactive in water H g ( N O 3 ) 2 h v H 2 O ! " ! ! H g ( O H ) • + N O 3 # $ h v H 2 O ! " ! ! H g 0 + • O H a n d N O 2 • + O • # % H g 2 + H S ! h v " # " H S • + e ! H S • ! S • ! + H + 2 S • ! " S 2 2 !

PAGE 44

44 and subsequently decompose into free sulfur and hydrogen, occurs at the rapid rate of *HS $ (ads) lifetime (~1 to 10 ns). Thereafter, solvation of the free sulfur can form polysulfide ion, leaving the adsorbed H atoms to pair up and evolve gaseous hydrogen, H 2(g) . The authors found that pH 8.5 was at the lower end of the a lkaline range for H 2(g) evolution, wh ereby more acidic conditions would promote formation of HS $ and S 0 according to the equilibrium reaction in Eq ( 2 32 ) [112] . ( 2 30 ) ( 2 31 ) ( 2 32 ) Oxidation of sulfur in dissolved organic matter. The oxidation of reduced sulfur found in DOM (e.g . thiols DOM RS$ ), can produce disulfide through the summary reaction in Eq ( 2 33 ). While contributions from photoreactions have not been studied, Pascal and Tarbell had previously identified the role of H2O2 in the oxidation reaction of merc aptan to disulfide [113] . Therefore, if H 2 O 2 is an intermediate photo reactant it may be possible for DOM RS$ photo oxidation to simultaneously oxidize RS $ groups that could influence overall Hg DOM binding. ( 2 33 ) Chloride Photochemistry Individual chlorine species HOCl and OCl Ð with acid base equilibrium pK a of 7.5 will photolyze at approximately the same rate to produce radicals • OH and Cl • shown by Eq ( 2 34 ) [108] . The authors Watts et al. observed photo generated • OH radi cal production from is more significant at pH >8, and approximated that UV C irradiation of ! H S ( a d s ) " + H 2 O # S ( a d s ) 0 + H 2 + O H " S ( a d s ) 0 + H S ! + O H ! " S 2 2 ! + H 2 O S 2 2 ! + H + " H S ! + S 0 2 R SH o xi d a t i o n ! " ! ! ! R SSR + 2 H +

PAGE 45

45 aqueous solutions with HOCl h as ) •OH of 1.4 mol Es 1 [114] . A proposed mechanism for Cl • production is the direct oxidation of chloride ion by • OH to produce HOCl • Ð according to Eq ( 2 35 ). The radical HOCl • Ð can be further oxidized in chloride rich solutions with pH < 5, which would then form weakly oxidizing radicals like Cl • in Eq ( 2 36 ) and ( 2 37 ), but would remain as the reactant HOCl • $ at higher pH. The radical Cl 2 • Ð is both an oxidant and a weakly chlorinating agent whose oxidizing action continues to be investigated, although it is characterized as an intermediate transient species that is primarily scavenged by DOM [115, 116] . ( 2 34 ) Cl Ð + • OH ! HOCl • Ð ( 2 35 ) HOCl • Ð + H + ! H 2 O + ) Tj ET Q q 0.24 0 0 0.24 207.4745 412.74cm BT 0.0002 Tc 50 0 0 50 0 0 Tm /TT1 1 Tf (Cl ( 2 36 ) Cl • Ð + Cl Ð ! Cl 2 • Ð ( 2 37 ) Effect on Hg photo redox . The photochemical reduction and oxidation of Hg in solutions that contain chloride ion Cl $ has been studied over both the natural sunlight and artific ial UV wavelength spectrum (210 to 400 nm wavelength ) and over a range of pH. The re is extensive discussion on the action of Cl $ as a photo reactant in Hg(0) oxidation [105, 119 121] . In studies by Zhang and Yamamoto, t he reduction of ionic Hg compounds HgNO 3(aq) and HgCl 2(aq) Hg(0) occur red in waters with lower redox potential from dep leted diss olved oxygen . By comparison of nitrate and chloride binding ligands, Cl $ decreased Hg(0) production in dark experiments, while during UV A irradiation, the Hg would readily cleave from its ion instead, thus resulti ng in rapid Hg(II) reduction [110, 117] . On t he other hand, Sun et al. recently reported that Hg Cl species are O C l ! h v H 2 O " # $ $ • O H + C l •

PAGE 46

46 photochemically inert based on observations that Hg(II) photo reduction was inhibited with increasing concentrations of charged aqueous Hg Cl species ( i.e., HgCl n (n 2)$ ) for Hg(OH) 2(aq) sol utions prepared over an NaCl gr adient and irradiated over the range of wavelength ( = 280 700 nm [118] . Authors Lalonde et al. experimentally demonstrate d that Cl $ indirectly mediates reduction at a pseudo first or der rate of 0.01 min 1 [94] . Similarly, Qureshi et al. reported pseudo fi rst order rates ranging (0.0066 to 0.032) min 1 , and postulated that Cl $ stabilizes Hg(II) following Hg(0) photo oxidation [122] . The developments may help explain findings from an early study by Allard et al. showing inhibited photo reduction of Hg(II) in natural waters with organic matter and Cl $ ; thereby supplementing the authors' initial th eory that inhibited Hg(II) photo reduction occurs when Cl $ competes with NOM for Hg binding [88] . Formation of solid phase mercury chloride species . Extensive research by Horvath and colleagues on direct photolysis of aqueous HgCl 2 (aq) provides evidence that its photo generated products include solid phase chloro mercurial species. This is described as reduction of the Hg metal center followed by the formation of insoluble Hg 2 Cl 2 (s) , wherein highly oxidative radical Cl • formed via a primary process photoreaction [123, 124] . The authors reported an increase in the Cl $ /Hg 2+ concentration ratio that shifted the Hg complex equilibria towards the formation of chloromercurials with Cl in higher stoichiometric ratios. However, the reaction appears to be relevant in aqueous solutions containing organic solvents (e.g., ethan ol, cyclohexane, or Eder's reaction), and could be relevant in photochemical treatment of petrochemical process Hg wastewater.

PAGE 47

47 Organic Matter Photochemistry Photochemical reactions affect optical properties of DOM, such as the loss of DOC fluor escence or photo bleaching, thereby creating photo intermediates that can alter Hg redox states [104] . Functional groups found in humic and fulvic acids (HA and FA, respectively) that absorb light over the full UV and vis ible light spectrum are optically defined as chromophoric DOM (CDOM). Principally affected groups are highly aromatic photo activate electron rich phenols [125, 126] . This explains w hy phenol groups found within VHA and SHA moieties ( Appendix Table B 4 ) react preferentially to UV irradiation due to unsaturated bonds [75] as opposed to less aromatic CHA and NEU fractions with lower absorbance at 254 nm. Irradiated DOM is known to be a photochemical source of ROS where early studies on the action of UV C on HA observed that radical mediated degradative oxidation reactions are intrinsically associated with the ability of CDOM to form 3 DOM * by photon absorption (see Appendix Figure B 2 ) and related excited states occurring in LMW DOM fluorescent species, chemi luminescence, and free radicals (e.g. • OH) [127 ] . Postulated pathways leading to • OH production are summarized by ( 2 38 ) where in the absence of iron, H 2 O 2 dependent and independent pathways have been identified. Hydrogen peroxide in UV driven hydroxyl formation . Recent wo rk by Lee et al. used UV Solar to measure H 2 O 2 independent • OH production by variable molecular weight DOM, and the author higher • OH production efficiency from though photo oxidized LMW fractions of CDOM (<1 kDa) in comparison with other forms found in mu nicipal waste [128] . Whereas McKay et al. proposed a mechanism for an H 2 O 2 independent • OH production that involves H atom abstraction from OH Ð or an H 2 O

PAGE 48

48 molecule by photo exci ted DOM, e.g. oxidation by 3DOM * [129] . In oxic systems on the other hand, t he H 2 O 2 dependant pathway has been examined to a greater extent in which dissolved O 2 has a promoting role in photo activating DOM to produce • OH reactive precursors [107, 130] . ( 2 38 ) Th e general understanding is that o nce DOM absorbs a photon, the deactivation of its excited state , 3 DOM • , can oc cur in several ways, such as reacting with dissolved O 2 to form ROS singlet oxygen 1 O 2 , which is commonly reported in natural sunlit waters. Similarly, 3 DOM * can react with O 2 to product ROS anion superoxide radical, O2 • $ , in Eq ( 2 39 ), which is also a precursor to • OH formation (i.e., Eq ( 2 11 ) to ( 2 14 )). ( 2 39 ) While the lifetime of 3 DOM • is longer in de oxygenated solution than with dissolved O 2 , in the absence of oxygen substrate degradation processes could be higher followed by de activation via radical scavenging. The two processes are depicted by proposed mechanisms in Eq ( 2 40 ) and ( 2 41 ) for the generic substrate, S [108] . ( 2 40 ) ( 2 41 ) D O M h v ! " ! p h o t o F e n t o n H 2 O 2 d e p e n d e n t 3 D O M # + O 2 " H 2 O 2 p h o t o l ysi s H 2 O 2 i n d e p e n d e n t 3 D O M # " H 2 O ; O H $ p h o t o l ysi s % • O H 3 D O M ! + O 2 " D O M • + + O 2 • # o r D O M + 1 O 2 3 D O M ! + S " # " D O M • $ + S • + D O M • ! + S • + " # " D O M + S

PAGE 49

49 Ions that readily react with DOM produced • OH radicals include Cl Ð , HCO 3 Ð and SO 4 2 Ð , which are abundant in seawater and chlor alkali and many other industrial wastewaters. Rubio and colleagues recently demonstrated the dampening effect that such compet ing ions can have on •OH and other radical action . The authors used UV C and H 2 O 2 for bacterial disinfection in lake water and artificial seawater spiked with HCO 3 Ð and DOM. Th e authors found that the rate of disinfection (99.9%) declined from 0.15 s 1 in Milli Q water to 0.14 s 1 and 0.13 s 1 for spiked lake water and artificial seawater, respectively. While the roles of Cl Ð and DOM were not fully investigated, the authors concluded that bicarbonate ions scavenge • OH to produces less reactive radicals (e .g. carbonate radical, CO 3 • $ ) thereby decreasing effectiveness by • OH [131] . Photo degradation of organic matter. Just as DOM photo generates • OH that contribute to Hg transformations, • OH scavenging by DOM can account for as much as ~90% in na tural waters [132] . While the redox capacity of DOM by • OH varies by its nature and concentration, the end products are lower molecular weight products that form by DOC mineralization , depicted in Eq ( 2 42 ). During photo fragmentation (i.e. polymerization), the aromatic VHA fract ion of HMW DOM (1 kDa
PAGE 50

50 Mercury T ransformations by P hoto A ctive Organic M atter Indirect photo transformation . The successful design of photochemical treatment for Hg wastewaters that contain photo active DOM is reliant on research developments in the currently active area of research Hg DOM photochemistry, and especially findings that identify individual DOM functional groups in photo transformation pathways in solid phase Hg formation. For example, the authors Si et al. recently reported experimental observations that aqueous Hg was photo transformed into particulate Hg in the form of % HgS (s) following UV A exposure for anoxic and acidic solutions prepared with Hg(II) and thioglycolic acid [134] . The photo chemical conversion of organic sulfur into available sulfide was previously noted by Bonzongo et al. following UV B irradiation of DOM for uptake by methylating bacteria, wherein experimental observations of photo induced methylatio n of dissolved Hg complexes The authors observed that Hg bound to UV B treated humic material (Sigma Aldrich) increased microbial methylation from 0.041 ng ml 1 h 1 compared to 0.33 ng ml 1 h 1 for the Hg fractio n bound to non irradiated HS . The increased methylation was postulated to have occurred by uptake of Hg DOM complexes formed from LMW DOM [92] . Direct photo transformation . The oxidation of Hg(0) by photo activated DOM is thought to occur by one of two ways 1) either competitively with O 2 driven reactions (H 2 O 2 dependent) or 2) in the absence of dissolved O 2 (H 2 O 2 independent). Organic matter bonded to Hg in wastewater can be especially susceptible to photo degradative processes during which the DOM binding ligand of Hg(II) is oxidized and Hg(0) is produced and possibly re oxidized to Hg(II) [65, 135] . Authors Ababneh et al. conducted UV A experiments using solutions prepa red with Hg(II) and weakly chelating organic ligands such as oxalate, citrate, and malonate, which are photo sensitive in aqueous

PAGE 51

51 solution, and reported observations of Hg photo reduction and photo oxidation [136 ] . In order to explain the observations, the authors cited the possible roles of ligand to metal charge transfer (LMCT) and dissolved oxygen according to the proposed photochemical Hg electron transfer mechanism by author Nriagu. As previously discusse d, t he process is postulated to occur when a photochemically excited state DOM molecule transfers electrons to Hg(II) thereby reducing it to form either Hg(I) or Hg(0); the reduced Hg re oxidized by dissolved O 2 [103] . The findings agree with those of Borello et al. who investigated photochemical formation of purgeable Hg(0) in ppm level Hg solutions with Hg DOM by examining the role of UV wavelength, dissolved oxygen , and DOM as either HA or FA with the use of continuous gas purging during UV irradiation with gaseous O 2 , N 2 or Air . The results showed that UV irradiation of oxygenated Hg HA solutions formed less purgeable Hg(0) (either by Air purge or O 2 purge) relative to the N 2 purge results. Based on the expe rimental findings the author predicted that Hg photo reduction can occur when excited state DOM produces excess O 2 • $ that transfers electrons to Hg(II) according to the proposed oxidation reaction: Hg 2+ + O 2 • ! ! Hg 0 + O 2 . This reaction process was predicted to be valid at higher concentrations of DOM (i.e., 10 vs 1 mg L 1 HA for solutions spiked with 1 ppm Hg (NO 3 ) 2 ). A comparison of wavelength UV A vs. UV C in Borello's experiments show tha t oxygenation via air purging affects processes of both Hg photo red u ction and FA photo degradation, which was more effective with UV A compared with UV C; whereas de oxygenation via N 2 purging results show that UV C was more effective compared with UV A. For each comparative experiment using HA or FA, the rate of Hg photo

PAGE 52

52 reduction was faster for FA relative to HA. It was postulated that experiments that resulted in less Hg photo reduction occurred due to fewer DOM molecular weight alterations that produced fewer ROS [21] . As pr eviously discussed, it has been noted that ROS produced from CDOM is wavelength dependent, producing either reducing or oxidizing radicals . The authors Aeschbacher et al. observed that treatment with either UV A or UV C, the same pathway for oxidizing tri plet states produced by 1 O 2 occurred via phenol groups, which constitute a higher concentration in SRHA (1.94 mmol g 1 HS) compared with SRFA (1.48 mmol g 1 HS ) [137] . Comparatively, Lester et al. reported findings that UV C more effectively produced • OH relative to UV A or UV B [138] , which appears to be a determining factor in Hg photo reduction .

PAGE 53

53 Table 2 1 . Equilibri um constants for cinnabar solubility product in the presence of polysulfides occurring as rhombic S (0). Reaction, n = 4 to 6 (1 ) Log K sp Ð 3.8 Ð 5.9 Ð 11.7 Ð 15.4 ! (1) Data sourced as original Paquette & Helz thermodynamic constants refined by Ja y et al . [62] Table 2 2 . Octanol water partitioning coefficient, K OW , for soluble Hg complexes , Hg(II) (aq) , and other environmentally relevant species . Complex a K O W (1) Complex K OW (2) Hg(OH) 2 0 1.2 Hg 0 4.13 HgCl 2 0 3.3 ! *HgSH + , HgS 2 2 $ 0 HgOHCl 0.05 CH 3 HgOH 0.07 Hg(SH) 2 0 , HgS 0 72 CH 3 HgCl 1.7 Hg DOM 1.7 to 3.3 (CH 3 ) 2 Hg 200 a. Charged complexes are assumed not to significantly partition into octanol [48] . Source data: (1) Adapated from original table by author Miller [139] (2) Data from Loux et al. [140] H g S ( s ) + H S ! + ( n ! 1 ) S 0 " H g S n H S ! H g S ( s ) + ( n ! 1 ) S 0 " H g S n 0 H g S ( s ) + H S ! + 2 ( n ! 1 ) S 0 " H + + H g S 2 n 2 ! H g S ( s ) + ( n ! 1 ) S 0 + H 2 O " H + + H g S n O H !

PAGE 54

54 Table 2 3 . Constants of formation (Log K) and solubility product (Log K sp ) for Hg S in the absence of polysulfide for pH and sulfide concentration constraints . Values are a ssumed valid under equilibrium conditions. Source data: (1) Origina l Schwarzenbach & Widmer constants refined by Drott et al . [61] (2) Dyrssen et al. [58] and (3) Barnes et a l. [141] Reaction Constraints Constant pH (1 Ð 11) S 2 $ (0.060 µM Ð 20 mM) Log K 39.1 (1) 32.5 (1) 23.2 (1) 6.19 (2) ! 8.30 (2) 0.57 (2) pH (6 Ð 10) S 2 $ (0.060 Ð 85 µM) Log K sp Ð 37.6 (1) pH (6.13 Ð 6.87) S 2 $ (100 < mM) Log K sp Ð 4.25 (3) Ð 3.50 (3) Ð 3.51 (3) H g 2 + + 2 H S ! " H g ( H S ) 2 0 H g 2 + + 2 H S ! " H + + H g H S 2 ! H g 2 + + 2 H S ! " 2 H + + H g S 2 2 ! H g ( H S ) 2 0 ! H + + H g H S 2 " H g H S 2 ! " H + + H g S 2 2 ! H g S 2 2 ! " H g S ( s ) + S 2 ! H g S ( s ) + H + ! H g 2 + + H S " H g S ( s ) + 2 H 2 S ! H g S ( H 2 S ) 2 H g S ( s ) + H 2 S + H S ! " H g S ( H S ) 3 ! H g S ( s ) + 2 H S ! " H g S ( H S ) 2 2 !

PAGE 55

55 Figure 2 1 . Sources of anthropogenic Hg emissions into the aqueous environment. Global estimate of chlor alkali waste in 2012 ~ 2.8 tonnes; sourced from U NEP Technical Background Report [34] . !"# $ %&''"()*+,#,#-* .*/'"0(12,"# 3* 456 7"#)(+&'******** /'"0(12*89)2& 3* :46 7;<"' $ 9<=9<,***** ,#0()2'> 3* :6 ?,<*'&%,#,#3* 46 @"<0*+,#,#3* :6 A2&9+*&<&12',1*/<9#2) 3* :6 ?2;&'*,#0()2',9<* 89)2& 3* B6

PAGE 56

56 CHAPTER 3 METHODS AND MATERIALS Lab safety. Experiments that required Hg handling were conducted under a ventilated and ducted fume hood using OSHA approved methods for protection and safe ty. Annual lab safety training with the university included the course on hazardous waste management from the Department of Environmental Health and Safety. Experiments were conducted using certiÞed A.C.S. or trace metal grade reagents and Milli Q NanoPure deionized (DI) water that had standard resistivity of 18.1 M, cm $ 1 . Solution Preparation and Chemical Reagen ts Experiments used real secondary process effluent samples obtained in batches from operations chlor alkali plant A and plant C during the period of 2008 2010. In addition, synthetic Hg wastewater solutions were prepared in the lab to represent Hg proces s indus trial wastewater effluent using Nano Pure DI water or by pH buffered Nano Pure DI water according to the reagents summarized in Table 3 1. Mercury . A 20 µM Hg stock solution was prepared by diluting an aliquot of store bought 100 mg L 1 Hg(NO 3 ) 2 solution (Ricca) or by weight using HgCl 2 crystals (Sigma Aldrich), where specified. A typical working solution of 0.1 µM (20 µg L 1 ) Hg was then prepared after appropriate dilutions using either pH buffered or non buffered DI water containing a known amount of selected inorganic and organic ligands. Chloride . The addition of chloride ligand was calculated from pH 3 buffered solutions or by the addition of stock solution prepared with sodium chloride (Fisher Scientific) and de ionized water. Specified experiments used HgCl 2 as Hg stock instead of Hg(NO 3 ) 2 in order to avoid possible interference by nitrate radicals.

PAGE 57

57 Sulfide . Sulfide concentrations were sele cted Solutions of NanoPure DI water and were spiked with sulfide stock to achieve predetermined concentrations that create d a range of Hg to sulfide ratios (Hg/S) , which ranged from t race sulfide to residual sulfide , as follows: • t race sulfide for sub molar equivalent Hg/S ratios , from 6 to 8 * 10 4 µM S 2 $ • m id range sulfide for molar equivalent Hg/S ratios, from 0.2 to 40 µM S 2 $ • and excess sulfide for ratios of Hg/S <1, from 160 to 1200 µM S 2 $ The s ulfide concentrations were experimentally verified using HACH spectrophotome try calibrated using the methylene blue idiometry method with potentiometric titrations [142] . The sulfide stock solutions were freshly prepared for each experiment by adding washed Na 2 S crystals to Nano pure DI water using reagent grade nonahydrate sodium sulfide, Na 2 S ' 9H 2 O (Sigma Aldrich) , and stored at 4 ¡C. The dissolution of crystalline Na 2 S primarily produces bisulfide ion HS $ rather than sulfide ion S 2 $ by hydrolysis reactions in Eq ( 3 1 ) through ( 3 3 ) and subsequently raises solution alkalinity. This is avoided by using pH buffered solution such that there is a known distr ibution of sulfide species; using a pH 7 buffer , for example, S is distributed as ~44% HS $ and ~56% H 2 S and the analysis of pC p H diagrams, free S 2 $ ion is not expected to significantly contribute at pH 7 calculated using pK a1 = 7 and pK a2 = 19 ( Appendix Table B 9 and Figure B 11 ) . ( 3 1 ) ( 3 2 ) ( 3 3 ) N a 2 S H 2 O ! " ! ! H 2 S + O H # H 2 S ! H S " + H + L o g K = " 7 H S ! " S 2 ! + H + L o g K = ! 1 9

PAGE 58

5 8 Natural organic matter. Natural organic matter was from the International Humic Society (IHSS) because of their uniformity, availability and benefits of extensive documented characterization (Appendix C Table B 5 ) . Experiments used Nordic Lake reference material for humic and fulvic acid isolates (NLHA and NLFA, respectively) as well as standard material for humic and fulvic acid from the Suwannee River (SRHA and SRFA, respectively) . In experiments that used non specific DOM , t he source was reverse osmosis aquatic NOM from the Suwannee River (SRNOM) . The Hg DOM solutions were made with NanoPure water and aliquots of DOM stock solutions, prepared by di ssolving solid isolate of DOM in Nano Pure DI water (i.e., 100 to 200 mg DOM ad ded to 60 mL Nano Pure water) and stored at 4¡C in amber borosilicate vials . Batch Reactor Experiments Dark mixing and UV irradiation. The prepared wastewater solutions are mixed in the dark and then illuminated as follows: First the solution is distribute d among identical reactors, each containing 100 mL, proceeded by continuous mixing under dark conditions that varied from one to 24 hours. The solutions were then illuminated under continuously mixed conditions for target contact times, e.g., t UV = 5, 15, 60 minutes (min), while enclosed in a reflective aluminum sleeve for UV blocking and protection. UV batch reactor setup. Batch experiments used custom built borosilicate glass UV reactors with Teflon coated stir bars. Borosilicate glass was chosen as the material for the reactor because it creates a net negative wall charge that can theoretically minimize Hg wall adsorption. The reactors are capped with a Teflon plug equipped with a UV bulb insert and porthole for a sampl ing tube, shown in Figure 3 1. The UV exposure portion of the experiments uses pre heated ballast and a single ended 9 W UV bulb placed in the center of each reactor. The batch reactors were designed to

PAGE 59

59 maximize photon energy using high ratios of volume to photon quantum yield that could be replicated in industrial treatment systems using similar UV radiation intensities. The radial energy was calculated per unit volume (mL) with horizontal cross sectional area A x = 12.10 cm 2 . Reactor dimensions and radial photometric output are provided in Appendix Figure 7 4 and Table A 1. Sampling and Hg filtration. The procedure for s ample collection and filtration was designed to minimize the introduction of oxygen into the solution that could result in non UV related Hg tr ansformations prior to Hg preservation and analysis; and the use of teflon sampling equipment and batch reactor lid minimizes experimental error through procedural Hg mass losses or cross contamination. Sample s needed for Hg analysis were collected from t he test solutions by extraction using sterile polypropylene syringes with 1 m L graduation and 50 mL capacity; the sampling tubes were connected to the batch reactor with a teflon sampling tube inserted through portholes in the teflon cap. The syringe was f illed with 20 mL solution and then connected with a screw top polypropylene encased filter (35 mm diameter) and the solution was filtered the solution directly into 40 mL sampling vials for Hg preservation and digestion. The same syringe and tubing setup was used to collect samples that were not filtered and deposited into the sample glass vials for preservation and analysis. UV wavelength and photomemetric output. UV C wavelength band was selected in order to most effectively study advanced photo oxidativ e processes, which occurs in the shortwave region of light and has the highest production of energy per photon in the UV spectrum, but represent just 0.5 to 3% total UV Solar, ( < 280 nm [143] .

PAGE 60

60 The purpose of using of UV B ha s relevance to naturally occurring processes of DOM and microbial activation related to Hg photochemical transformations [92, 144] . The 9W UV bulbs sourced from Phillips Inc. are G23 base plug in lamps: TUV PL S 9W/2P and PL S 9W/12/2P 1CT (Appendix Figure 7 1 to Figure 7 3 ). The 9W UV C bulbs emit 2.3 W radiation with photometric output of 90% at 253.7 nm wavelength and less than 10% over 300 500 nm, and has an equivalent radial intensity of 198.3 mW cm 2 . The total photon flux from UV C in the solutions for the batch reactor system was determ ined to be 2.46 * 10 5 Einsteins (Es) min $ 1 using ferrioxalate actinometry by Byrne in our research group [145] . The 9W UV B bulbs emit 1.2 W of irradiation within a narrow wavelength band from 310 315 nm, which has an equivalent radial intensity of 99.1 mW cm 2 . While its photon flux was not measured, an approximate value of 1 * 10 5 einsteins min 1 was determined by comparison with 8W UV A bulbs (365 nm) with 1.6W output and produced 1.26 * 10 5 Es min 1 [12] . The equivalent values for photon and energy flux are shown by wavelength comparison in Appendix Table A 1. Experimental Analysis and Characterization Organic matter characterization. Total organic carbon was analyz ed on a Shimadzu TOC 5000 analyzer or Tekmar Dohrmann Apollo 9000HS auto sampler using 30 mL of solution that was collected in organic free glass vials acidified with 10% phosphoric acid (Fisher Scientific). TOC analysis used combustion and air stripping to remove inorganic carbon by zero grade air (Airgas, Inc.) and the sample is oxidized to CO 2 on a platinum/alumina catalyst in a high temperature (680 ¡C) furnace; the CO 2 is then analyzed with an NDIR (non dispersive infrared gas analyzer) detector. The excitation emissions (EEM) spectra of dissolved organic matter was characterized using

PAGE 61

61 a 2500 Hitachi Fluorescence Spectrophotometer coupled with fluorescence regional integration, a quantitative technique that Chen et al. developed to delineate EEM resul ts into five regions based on fluorescence of model compounds, DOM fractions, and marine waters or freshwaters [146] . Spectrometric absorbance was measured for the Hg-DOM solutions at either 254 nm or 290 nm waveleng ths; values obtained at 254 nm are used to compute SUVA254, which reflects structural changes in aromatic and unsaturated groups relative to changes in molecular weight. The response values over the wavelength range of 275-290 nm may be useful in characterizing changes in DOM optical properties such as ECC [147, 148] . Detection of Mercury and Analysis The contents of each reactor were analyzed for total Hg content following bot h dark-mixing and UV irradiation periods of each experiment. After the specified run time elapsed, 20 mL of the solution was collected and preserved for analysis according to the method of Hg detection, which remained consistent per sample batch or experimental trial. The method used to detect Hg concentration is specified in the results as either a) Atomic adsorption (AA) following Method 245.1 [149] using an AA Spectrometer (Teledyne Leeman Labs, Inc ), which has a detection limit of 200 ng L1 [150]b) Atomic fluorescence (AF) following Method 245.7 [151] using an AF Spectrometer (P S Analytical PSA 10.025 Millennium Merlin), which has detection limit of 6.0 ng L1 or c)third party verification for measuring Hg in chlor -alkali samples where stated, which used Method 1631E and detection limit of 0.5 ng L1 [150] . The AA and AF detection limits were calculated as three times the standard deviation of the instrumental blank experiment, found to be between 100 and 200 ng L1 using AA and 6.0 ng L1 using AF;

PAGE 62

62 where AA analysis had more variability in detection limit due to differences in calibration curves obtained for each batch analysis. Differences in Hg method detection by AA and AF analysis . The key differences between AA and AF methods in wastewater analysis are the reagents used for oxidation of Hg, detection limit and effect of interferences in chemically complex solutions. Detection of Hg by AA uses both nitric and hydrochloric acids for preservation and heated oxidation method with potassium permanganate (Fisher Scientific) for digestion. Detection of Hg by AF is more sensitive, and uses only hydrochloric acid for preservation and a stronger oxidant, bromine monochloride, BrCl, for digestion. Both AA and AF methods detect total Hg in the form of gaseous Hg(0) by first converting organic Hg present in solution into ionic Hg(II), which is then reduced to Hg(0) by mixing with a purified 10% w/v stannous chloride solution (Sigma Aldrich) . The mixtures is then purged using argon gas (Airgas, Inc.) so that Hg(0) is transported from the solutions into the air phase can be detected at its absorbance wavelength, 254 nm. In both methods the samples underwent cold acid preservation to facilitate the complete oxidation of Hg. The BrCl oxidizing reagent is prepared using furnace dried (500C) bromamine salts (Fisher Scientific) prepared in solution with Nano-pure DI water and trace-metal grade hydrochloric acid (Fisher Scientific). BrCl has been found to be an excellent oxidant and preservative for total Hg in freshwaters, working faster and more effectively on many organomercurials than the potassium permanganate oxidation used in AA detection [152] . Trace-metal grade BrCl also has the advantage of a typically lower reagent and procedural handling Hg cross contamination compared

PAGE 63

63 with a heated oxidation method that uses more chemicals; therefore, AF detection was the preferred method for Hg analysis in sulfur and organic rich solutions. Mass balance of mercury products. A mass balance approach was used to quantify changes in Hg phase distribution, where (*) denotes Hg products of UV treated solutions. The terms C 0 and *C represent the initial concentration in untreated and UV treated solutions, respectively; C F denotes Hg concentration in filtrate; and *C F is the concentrati on of the filtrate of UV treated solutions. The various fractions are operationally characterized for practical relevance to the treatment industry (see Table 3 2 ). Total Hg in the mass balance, [Hg T ] = [Hg (aq) ] + [ Hg (s) ] , consis ts of aqueous complexes, which are either soluble as Hg (aq) or in particulate form as Hg (s) , together with reactor adsorbed Hg as Hg R ads , and zero valence Hg(0) that could be either dissolved Hg 0 (aq) or in progression to the gaseous phase as Hg 0 (aq) . The equations for the experimental mass balances are defined prior to irradiation by dark reacted Hg T using Eq ( 3 4 ) and for UV irradiated solutions as UV reacted *Hg T using Eq ( 3 5 ). dark reacted Hg ( 3 4 ) UV reacted *Hg H g T h v ! " ! # H g T + # H g ( g ) 0 where, H g T + ! H g ( g ) 0 = * H g ( a q ) + * H g ( s ) + * H g R " a d s + * H g ( a q ) 0 U V ! " # # # # # # # # # # # # # # # # # # # # # # # # + H g ( g ) 0 U V ! " # # ( 3 5 ) With the exception of fugitive gaseous Hg 0 (g) that evolved from solut ions during UV irradiation, Hg concentrations were measured using AA or AF Hg sample analysis. Since the experiments used filtration to remove Hg from solutions, filter retained Hg is H g T = H g ( a q ) + H g ( s ) + H g R a d s + H g ( 0)

PAGE 64

64 considered as solid phase particulate Hg (s) . The Hg remaining in the sol ution filtrate is considered soluble aqueous Hg (aq) , i.e., assumed as either Hg(II) or Hg(0). The solid phase Hg (s) is calculated by [Hg (s) ] = [Hg T ] Ð [Hg (aq) ], and Hg removal in the following equations: ( 3 6 ) ( 3 7 ) Estimation of Experimental Error Experimental methods were designed to reduce Hg cross contamination and achieve reproducible results. A three step process was used for cleaning glassware t o ensure all experiments were carried out with minimal potential for background Hg, consisting of cleaning with a 10% v/v HCl solution to remove residual DOM, and 20% v/v HNO 3 to remove residual Hg, and then rinsed 3 times with DI water. Experimental error was calculated from a set of results repeated in at least duplicate. The level of Hg contamination during procedural and reagent blank experiments is less than 0.02 µg L 1 Hg which is accounted for by mostly AA procedure digestion reagent background Hg. The relative standard deviations were lower wit hin each individual experiment. Measuring experimental mercury losses. A mass balance approximation of Hg losses used the ratio of initial and final unfiltered Hg concentrations (*C/C 0 ). Values *C/C 0 <1 ind icated Hg losses occurred, which were predicted to occur as either Hg(0) evolution from the solution into the batch reactor head space, or Hg adsorbed on the borosilicate reactor walls: [*Hg Loss ] = [*Hg(0) (g) ]+ [*Hg (R ads) ]. Photo reduced Hg in solution a s DGM can enter the atmosphere as Hg(0) g [121, 153] through evolution at a u n t r e a t e d H g r e m o va l = [ H g ( s ) ] [ H g T ] = C 0 ! C F C 0 = 1 ! C F C 0 U VAC H g r e m o va l = ! [ H g ( s ) ] ! [ H g T ] = ! C " ! C F ! C = 1 " ! C F ! C

PAGE 65

65 diffusive rate that is determined by DGM saturation [154] . The rat io (*C/ C 0 ), was calculated for each UV contact period tested and control runs measured the residual Hg adsorbed onto the reactor, stir bar and UV bulb by Hg analysis of a reactor rinse solution that used a heated acid treatment (0.1 M acid as HNO 3 or HCl ). Control dark mixing only experiments using Hg(NO 3 ) 2 in acidic Nano Pure solutions had total Hg recovery of 99 ±1% (i.e. [Hg T ] = [Hg (s) ] + [Hg (aq) ] + [Hg Loss ], which suggests the mass balance approach has a maximum probable error of ±3%, which is also considered the allowable experimental error for the methods of measurement.

PAGE 66

66 Table 3 1 . Acid and base reagents used to prepare pH buffered solutions. pH Salt reagents and acid or base 3 1L = 10.21 g potassi um hydrogen phthalate and 22.3 mL of 1.0 M HCl 7 1L = 6.81 g potassium di hydrogen phosphate and 29.1mL of 1.0 M NaOH 11 1L = 2.10 g sodium bicarbonate and 22.7 mL of 1.0 M NaOH Table 3 2 . Operationa l definitions that characterize Hg during experimentation. Definition Oxidation State(s) Symbol Measurement Total Aqueous 2+,1+,0 Hg T [Hg] analysis of unfiltered sample Soluble 2+,1+,0 Hg (aq) [Hg] analysis of filtrate sample Filterable 2+,1+, 0 Hg(s) M ass balance Residual a 2+,1+,0 Hg R ads [Hg] analysis of reactor rinse sample Fugitive Gas b 0 Hg(0) (g) Mass balance DGM 0 Hg(0) (aq) Not directly measured a. Residual batch reactor Hg was measured in empty reactors immediately following UV irradiatio n, * Hg R ads . b. Fugitive gaseous Hg was observed in some UV irradiated solutions *Hg(0) (g ) . Figure 3 1 . Schematic of UV/UVAC batch reactors.

PAGE 67

67 CHAPTER 4 DEVELOPMENT OF ULTRAVIOLET ACTIVATED CHELATION TECHNOLOGY Scope Chlor alkali facilities that employ mercury (Hg) cells for chemical manufacturing conventionally treat ppm Hg brine water using precipitation, which produces "secondary process" water with ppb Hg effluent levels that do not meet increasingly stringen t standards Ð widely prompting conversion to non Hg processing. However, as Hg cells are still used in several dozen facilities worldwide, this work presents the findings of an alternative treatment method called Ultraviolet Activated Chelation (UVAC). T he UVAC treatment process uses high intensity UV C irradiation to photochemically remove Hg from chlor alkali secondary process effluent, which pr oduced low ppt level effluent (<40 ppt) following alkaline pH adjustment before UV C irradiation. Bench scale UVAC batch experiments using chlor alkali secondary process effluent of varying composition was investigated prior, during and following the 2009 2010 UVAC pilot study in order to obtain physiochemical characteristics that determine Hg removal by this homo geneous process. This included measuring the interchangeable transformations of Hg from its soluble to solid filterable phase and Hg reduction and evolution as gaseous Hg(0). Variables that were examined include the effect of UV C contact time in relatio n to changes in pH and oxidation reduction potential. The findings of this experimental research enable a proposed method that can predict UVAC Hg treatment for secondary process wastewaters in which Hg speciation is governed by sulfide, chloride and orga nic carbon constituents: It requires measuring the co dependence of pH and UV C irradiation with the voltametric response in redox capacity of the Hg wastewater solution. The use of this method aims to provide

PAGE 68

68 a better understanding of UVAC treatment appl ication in suitable wastewaters of interest, e.g., brine purge water, carbon filtered water, and coal combustion Hg waste control, such as flue gas purge water. Chlor alkali Use of Mercury In the 1980's, nearly half of all chlor alkali manufacturing used elemental Hg(0) as a catalyst in the electrochemical conversion of saltwater brine into common chemicals such as caustic soda, potash, hydrochloric acid and chlorine and hydrogen gases [155] . The use of Hg(0) as a cathode for electrolysis and as an anode in the decomposing reaction of the Hg cell process was first developed and commercially used in the 1880's and the industry has undergone drastic change in recent decades as a priority for addressing environm ental health, and human occupati onal health and safety concerns [156 159] . The Chlorine Institute operating in the U.S., Canada, and Mexico reduced its Hg use from 145 to 4 metric tons (97%) from 1995 to 2008 [3] by implementing Hg free alternative technologies such as diaphragm cells or ion exchange membrane cells [155] . However, several dozen Hg cell plants currently operating are challenged to achieve sufficiently low effluent concentrations [34] that a) meet stringent U.S. EPA requirements [4] and b) meet objectives set by the World Chlorine Institute in partnership with the global treaty of the Minamata Convention on Mercury that aim to reduce contamination of this widely toxic and environmentally persistent contaminant [1] . A source of fugitive Hg waste occurs in Hg cell process manufacturing despite recycling the Hg(0), taking the form of either gaseous or aqueous Hg from equipment maintenance and minor operational perturb ations [160, 161] . Most emissions occur from small scale process leaks, which has been found to produce concentrations up to 75 times the background in soils surrounding operational chlor al kali plants [162] ;

PAGE 69

69 contamination can also occur in aquatic and terrestrial systems in the vicinity of impr operly decommissioned facilities [163] . More commonly, the high est concentrations are found in secondary process effluent released into natural water systems following conventional treatment of ppm Hg brine solution [164 166] . C onventional Treatment Effluent Characteristics Hg removal from brine wastewater is a multi stage batch process summarized by the reaction in Eq ( 4 1 ) , in which Hg can have multiple, often unknown, oxidation states before reacting to form HgS (s ) . This is because secondary treatment alone is not designed to removing liquid Hg(0) or Hg(I) occurring in the brine water. Since Hg(I) forms stable complexes of Hg 2 Cl 2 (s) precipitate with low chloride concentrations [167] , primary treatment before sulfide precipitation recommends moving the Hg equilibrium, Hg 2 2+ ! Hg 2+ + Hg 0 , to the right by bri nging the solution to pH 10 11. L iquid Hg(0) and Hg(I) solids can then be filtered or settled out of s olution, which also removes brine sludge and iron compounds [168] . ( 4 1 ) An example of the treatment method using Chlorine Institute Inc. guidelines r equires: 1) Oxidize Hg(0) to its tetra chlorinated form, HgCl 4 2 Ð , by means of acidification to pH 3.5 5.0 and/or by addition of hypochlorite; 2) Remove excess free chlorine by addition of sodium thiosulfate; and 3) Remove excess oxidant and form Hg precipi tate by addition of sulfide, shown in Eq ( 4 2 ) (e.g. NaHS in acid water is ~100% H 2 S with pK a1 ~7 for H 2 S and pK a2 ~19 for HS Ð ). While the size of Hg S particles can vary, the precipitate is operationally defined as micro HgS (s) using 0.45 µm pore size filters. Therefore, the solution must circulate for 1 to 3 hours in order to allow sufficient H g 2 + + H g 2 2 + + H g 0 + 2 S 2 ! ! H g 0 + 2 H g S ( s )

PAGE 70

70 crystal growth before filtration. Longer mixing periods may not be practical for sulfide precipitation treatment due to competing reacti ons that can slow the formation, growth and aggregation of HgS precipitate [52, 62, 169] . ( 4 2 ) Secondary process effluent contains Hg that is filter passing Hg(0) (aq) as well as any Hg(II) that reacted with sulfide to form nano p articles (<0.02 µm) [57] , or colloids (<0.2 µm), or in the case of alkaline solutions, soluble Hg polysulfides [61, 141, 170] . Effluent containing colloidal or dissolved organic matter (DOM) can either promote or inhibit soluble Hg formation under sulfidic conditions depending on both concentration and the nature of its binding characteristics [65, 80, 86] ; furthermore, nano Hg(s) that forms during secondary treatment can be stabilized by DOM thus preventing particle growth for the 1 to 3 hour conventional residence ti me [57, 171] . The fate and transformation of Hg discharged into natural waters would be largely determined by the type of Hg DOM complexes formed in this type of sulfidic effluent, whe reby a growing body of evidence points to its role in photo reduction potential or availability for methylation [87, 91, 172, 173] . The water qu ality summary reported in reflects the range observed in effluent two chlor alkali manufacturing plants included in this study (hereon chlor alkali s amples A and C). Materials and Methods Wastewater samples. Bench scale experiments were designed for UVAC testing on secondary process wastewater efßuent samples received directly from chlor alkali facilities operating in the Southeastern U.S. The samples were collected in different batches from a secondary clarifier following sulfide precipitation treatment by an independent EPA certiÞed analytical laboratory using "clean hands dirty hands" H g C l 4 2 ! + H 2 S ! H g S ( s ) + 2 C l 2 + 2 H +

PAGE 71

71 procedures required by Method 1669 for low level metal analysis [174] . The samples we re stored at 4 ¡ C in polyurethane bottles and a record of storage time (sample age) was maintained. The general experimental design in Chapter 3 describes the methods and materials for batch reactor UVAC treatment, and measuring concentrations of Hg by sample preservation, digestion and analysis. Merc ury Removal Definition The formation of particulate Hg followed by filtration achieves maximum Hg removal in chlor alkali wastewater. Therefore, notwithstanding inherent sample batch heterogeneity, UVAC treatment development is based on experimental find ings that relate the overall performance to solid phase Hg removal by filtration. The total Hg consists of operationally defined solid filterable Hg (S) and aqueous soluble Hg (aq) in the filtrate, which is expressed simply in a mass balance: [Hg (T) ] = [Hg (aq) ] + [Hg (S) ], as discussed in Chapter 3. Batch experiments measure total and filtrate Hg in solutions before and after UVAC, which uses (*) to denote Hg measured after the specified contact time, t UV . Hg removal by UVAC is defined using the normalize d ratio (1 Ð *C F / *C) (multiply by 100 for %Hg removal) in which *C = [Hg T ] UV and *C F = [Hg (aq) ] UV . For each UV C irradiation period (t UVC ) total Hg is measured and used to calculate the ratio (*C / C 0 ), whereby values < 1 are expected to occur due to Hg(0 ) evolution when Hg R ads is less than ~1 to 2% Hg Loss . The treated water was digested using Method 245.1 and analyzed using a Hydra AA Atomic Absorption Spectrometer (Teledyne Leeman Labs, Inc). The detection limit was generally 200 ng L 1 Hg , although c arefully developed standard curves were performed for each batch analysis, and a detection limit as low as 60 ng L 1 was also obtained. Selected samples were verified by third party analyses via Method 1631E which has detection limit of 0.5 ng L 1 Hg.

PAGE 72

72 Me asuring Redox Relative changes in electric potential were measured with varying pH and UV C contact time. Measurements used an oxidation reduction potential (ORP) electrode with platinum pin and glass sensing bulb combined with Ag/AgCl reference half cell (Fisher Scientific accumet Metallic ORP indicating electrode, Model 13620115), filled with 4M KCl with AgCl for storage, which was selected for its stability at higher temperatures and with strongly basic samples. The method of measurement followed manufac ture specifications; the ORP probe was standardized using a single point calibration and the observed values were corrected for temperature using the relationship developed in Appendix Figure 7 6 . The ORP results are shown as SHE corrected values, Eh (V), for comparison with standard redox couples, Eh ¡ (V). Chlor alkali Batch Sample pH The batches of chlor alkali wastewater were received in varied volumetric quantities (between 1 and 5 L), thus permitting a series of bench scale U VAC experiments. The pH will be an important factor because it affects the speciation of mercury, thus affecting the reactions that may be occurring. Its effect on UVAC performance was studied by adjusting pH using 0.1 M NaOH or 0.1 M HCl . The initial pH of each batch sample depended on "pre treatment" before dispatching the chlor alkali plant: batches with pH in ~ 9 to 12 were pH adjusted at the chlor alkali plant with NaOH from an initial pH range of 2 to 4. As a control, samples were tested for filterab le Hg removal upon receipt in the lab before pH adjustment or UVAC testing. Investigation of Aqueous Phase Hg Speciation The water quality analysis indicates the main determinants for Hg speciation before UVAC treatment are chloride, sulfide, organic matt er and solution pH. The

PAGE 73

73 possible dissolution of solid Hg sulfide species may help explain the incomplete formation of filterable HgS (s) during secondary treatment, which can occur through hydration, or as HgS 0 (aq) from the effect of polysulfides in anoxic and neutral to alkaline waters with S 0 saturation. In sulfidic solutions at low pH, & HgS dissolution is much more favorable due to H 2 S which reacts to form Hg(SH) 2 0 [175] . In oxic waters, HgS is unstable and decomposes to produce sulfate by the oxida tive reaction in Eq ( 4 4 ), which is thermodynamically favorable in alkaline pH solution relative to simple dissolution according to Eq ( 4 3 ) or Eq ( 4 5 ) [46] . ( 4 3 ) ( 4 4 ) ( 4 5 ) A comparison of potential pH diagrams developed for Hg S and Hg S Cl systems by Twidwell et al. in Figure 4 1 demonstrate that solvation of Hg(0) is not effective without chloride ligands except under oxidizing/acidic conditions within the region marked as HgSO 4 (aq) [176] . O n the other hand, the addition of chloride has been shown to decrease Hg(0) concentration s due to the strong oxidizing potential of Cl $ ion on Hg(0) [117] . Th e high chloride content in chlor alkali samples is therefore expected to facilitate Hg dissolution over a wide range of solution potentials and pH values. While a dissolution reaction f or HgS (s) is possible with hypochlorite according to Eq ( 4 6 ) and Eq ( 4 7 ) [46] it has been reported that that once HgS (s) forms in oxic and neutral to alkaline solutions, chloride ligands are unable to compete with sulfide and polysulfide Hg species [170] . H g S ( s ) + 2 H 2 O ! H g ( O H ) 2 0 + H 2 S L o g K = " 3 8 . 2 H g S ( s ) + H 2 O + 2 O 2 ! H g ( O H ) 2 0 + S O 4 2 " + 2 H + L o g K = " 9 3 H g S ( s ) + 4 H 2 O ! H g 2 + + S O 4 2 " + 8 H + + 8 e " L o g K = " 7 0

PAGE 74

74 ( 4 6 ) ( 4 7 ) Inorganic Hg speciation was approximated using Visual MINTEQ software, using thermodynamic parameters of the geochemical assessment tool [177] . Speciation calculations used a sulfide concentration gradie nt and average water qua lity values shown ( Table 4 2 ) The model favored insoluble Hg S formation at low sulfide concentrations and low pH, whereas for a system of higher sulfide concentrations, sulfur , chloro , and hydroxy merc urials occurred across the pH spectrum. The highest Hg S activities were reported for solid phase Hg(HS) 2 (s) for pH " 5, and upon pH increase, speciation shifted towards dissolved phase aqueous Hg occurring predominantly as Hg(SH) 2 0 , HgHS 2 Ð and HgS 2 Ð for pH # 9. The high DOC content in batch chlor alkali samples led to characterization by EEM fluorescence spectroscopy analysis for Sample C ( Figure 4 20 ); the resulting peaks were consistent with mostly humic and hydrophilic protein like structures. Solution pH can modify DOM surface area and produce changes in macromolecular structure over the reported size range from 0.1 to 1.2 µm , which Myneni et al, reported based on aromaticity and carboxyl content [178] . Therefore, experimental filtration analysis examined the effect of pH on DOM size distribution chlor alkali solutions by comparing pore sizes 0.22 µm (mixed cellulose ester or MCE) and 0.45 µm (nylon). The results in Figure 4 2 show significantly less DOC in neutral solutions with 0.22 µm MCE filters, whereas 0.45 µm nylon filters produced comparable DOC for solutions with pH 2, 7 and 11. H g S ( s ) + 4 N a O C l + 2 C l ! " H g C l 4 ( a q ) 2 ! + N a 2 S O 4 + 2 N a C l H g 0 + H O C l + H + + 3 C l ! " H g C l 4 ( a q ) 2 ! + H 2 O

PAGE 75

75 UVAC Design Parameters Filtration media . Factors considered in filter sel ection for Hg removal were pore size, filter media and cost. The measurement of Hg (s) was defined from Hg particle size exclusion results using 25 to 55 mm diameter filters for a sample volume of 100 mL using filters that are housed in single us e sterile p olypropylene casings. The Hg filtration results in Figure 4 3 compare filtrate concentrations using pore sizes 0.22 to 6.0 µm before and after UVAC treatment (t UVC = 5 min); the sample tested was chlor alkali sample C that had been adjusted to alkaline pH at the plant (pH in ~10, 11 and 12). The size distribution of filterable Hg particles is similarly characterized befo re and after 5 min UVAC treatment for the pH range tested and a comparison filter pore sizes 0.45 and 0.22 µm produced results that were statistically the same. The types of filter media tested were nylon, nitro cellulose and micro glass fiber, which were comparatively tested for differences in Hg removal from chlor alkali solutions that were prepared of 3 to 12 and solutions temperature in the range of ~ 4 to 50 ¡C), and the reported differences were statistically insignificant. Cost was the final variable considered since the filters used for bench scale experiments are not designed for regeneration, and it was determined that 0.45 µm pore size and either nylon or nitro cellulose is an acceptable filter material for UVAC treatment. UV C redox measurement . A series of experiments was developed with the aim of finding a relationship between Hg photo transformations and changes in solution oxidation potential during UVAC treatment. This was done by measuring redox potential in solutions with varying photomet ric doses using chlor alkali wastewater samples A and C that were prepared to pH levels in the acidic, neutral or alkaline range . The ORP values of the solutions were measured following UV C irradiation periods from

PAGE 76

76 0 to 80 minutes. The corresponding pho ton fluence is calculated from radial photon flux calculated as 20.3 Es cm 2 min 1 , corresponding to energy flux 9.6 * 10 9 J cm 2 min 1 . Experimental Results UVAC Mercury Removal I n order to identify photo chemically driv en mechanisms that optimize UVAC, th e chlor alkali samples were characterized filterable Hg formation, Hg losses, and changes in redox potential, and organic content . The results provide the basis for experimental work using synthetic Hg wastewater presented in Chapters 5 and 6. Chlor alkal i sample C . Chlor alkali secondary process waste water from Sample C was treated in UVAC batch experiments following pH adjustment and mixing in the dark for 1 hour. A comparison of Hg removal results for various pH values in Figure 4 4 shows the lowest filtrate effluent concentration occurred at pH 11 and 30 min contact time. Several different batches were then tested using pH 11 adjustment and 30 minute contact time in order to evaluate the consistency of Hg removal u sing these treatment variables. A summary of the results in Table 4 3 show that UVAC batch experiments produced at least 97% Hg removal for chlor alkali Plant C batch samples adjusted to pH 11 with C 0 ranging 2 to 12 µg L 1 Hg. The variabi lity in the wastewater chemistry of the different batches is thought to explain why some, but not all, produced low ppt levels in the UVAC treated effluent. Further bench scale testing demonstrated that UVAC Hg removal had varying pH and contact time crit eria for optimal Hg removal in process waters from different chlor alkali manufacturers. The results show that S ample C had UVAC Hg removal that correlated positively with U V C irradiation periods up to 50 min. W hile a similar trend was expected for Sampl e A, Hg removal decreased with increasing irradiation periods

PAGE 77

77 and instead, the best Hg removal was obtained between one and five minutes and did not require pH adjustment as there was no significant difference in Hg removal between solutions adjusted to pH values between 4 and 12 [26] . Chlor alkali sample A. The UVAC Hg removal for chlor alkali sample A ( Figure 4 8 ) had experimentally measured Hg losses occurring by Hg(0) g evolution th at varied by pH adjustment ( Figure 4 10 ). The acidified pH 3 solutions produced 94% filterable Hg following the longest UV C contact period tested at t 80 despite u nremarkable Hg(0) losses (~5%), which had treated effluent filtrat e *[Hg (aq) ] = 400 ng L 1 Hg. The spiked solution, sample A2 , that was adjusted to pH 4 had the greatest losses (~20%) and also greater Hg removal of ~99%, which resulted in treated effluent filtrate Hg levels below the AA detection limit, *[Hg (aq) ] <<100 n g L 1 Hg. UVAC Redox Potential The objective of this lab study on UVAC Hg removal processes used ORP measurements to monitor changes in oxidizing potential from the effects of UVAC variables of solution pH, UV contact time and Hg concentration. The study compared aged Sample C with Sample A batch samples labeled as A1 and A2, where A2 was s piked with an additional 50 !g L 1 Hg. The Hg spike used aliquots of diluted Hg prepared with crystalline HgCl 2 in an effort to conserve the speciation o f Hg, which wa s postulated to occur predominately as HgCl 2 (aq) . The results in Figure 4 5 show potential values corresponding to unit pH increments obtained during an acid titration with 0.1M HCl (Appendix Table B$ 2) . The ORP values measured in chlor alkali samples A and C had a range from 0.2 to 0.7 Eh(V). The titration experiment shows that sample A had an acid/base equivalence point at pH ~6 and Eh ~0.45 V by addition of 0.5 and 0.7 mol HCl for

PAGE 78

78 Sample A1 and A2, respectively (shown in Appen dix Figure A 8 ); sample C values were less discernable although they appear similar to sample A. By comparison with Hg redox equilibr ia ( Table 4 4 ), the results appear relevant to chloro mercurials HgC l 2 and HgCl 4 2 Ð that have standard reduction potentials of E ¡ (V) = 0.46 and 0.41, respectively. The effect of UV C produced t he largest gradient change in redox potential occurs during the first 5 minutes of irradiation for all the pH values tested, shown i n Figure 4 7 . The decreasing electrode potential values demonstrate that UV C could initially catalyze reducing conditions depending on the solution pH. It appeared that reducing potentials measured initially in alkaline pH wast ewater samples remained in this range for longer periods of UV C irradiation. The observation may be relevant to identifying reaction mechanisms that drive UVAC Hg removal since alkaline pH, especially pH 11 , was previously identified as an opt imized UVAC design variable during its initial development [26] . The trend was similar for samples prepared to pH as acid and neutral, although a reversal towards oxidizing conditions occurs after approximately 35 mi nutes. The initial rate of redox conversion for Sample C while a comparable trend occurs for pH >10 solutions between 20 and 60 minutes. A comparison of ORP measurements from samples A1 and A2 with added HgCl 2 produced statistically identical results. B ased on this finding it is evident that UV C driven ORP changes better correlate with photo transformations of a non Hg constituent occurring in the chlor alkali wastewater. This could be a constituent that has concentrations in the ppm range or greater , such as organic matter, chloride or sulfide. Hg removal redox measurement. The results for UVAC Hg removal were obtained for Sample A1 and compared with Sample A2 with ~50 !g L 1 added HgCl 2 .

PAGE 79

79 The samples were adjusted from an initial pH i ~11 to pH 9, 7, and 3-4 using HCl and mixed for 1 hour. The reported ORP values reflect UV-treated solutions and thus represent a measure of redox potential that is relative to concurrent UVAC Hg removal results and Hg losses, shown for Sample A1 in Figure 4-8 and sample A2 in Figure 4-9. The regression analysis of Eh vs. Hg removal in Figure 4-10 suggests a trend with increasing irradiation periods: Decreasing Eh (y-axis) is favorable for Hg removal (xaxis) by formation of filterable solid-phase Hg. The apparent outliers are investigated by analysis of Hg losses as Hg(0) and Hg R-a ds . . Hg losses during UV irradiation were measured as evolved Hg(0) and reactor adsorbed Hg after 15 and 80 minutes. The results were used in the comparative analysis of Hg(0) vs. Eh in Figure 4 $12; the findings show that greater photo-reduced Hg(0) occurring in acid pH samples, which after 80 minutes of UV-C was ~7% for Sample A1 (Eh ~0.65 V) whereas sample A2 produced ~21% over the same irradiation period (Eh ~0.36 V). The measured 0.3 V Eh difference in potentials may be explained by redox couples occurring with the added HgCl 2 i n Sample A2 as HgCl 2 /Hg 0 a nd HgCl 4 2– /Hg 0 w ith E(V) 0.41 to 0.46. T he high incidence of reactor adsorbed Hg for pH 7 at t 80 in both solutions reduced overall Hg removal, and as a result appear as outliers in the Eh-Hg (s) regression analysis. The results i show a similar trend compared to the initial series of experiments. The key difference observed in Hg removal and Eh values is that an overall shift towards higher Eh values corresponds to decreased Hg (s) removal and decreased Hg(0) evolution.

PAGE 80

80 In vestigation of Possible UVAC Processes Effect of Adjusted Solution pH The potential effect of added chloride from pH adjustment using HCl acid was tested in UVAC Hg removal by comparison of acids H 2 SO 4 , HNO 3, and HCl. The experiment used chlor alkali S amp le C (pH in ~11) that had been acidified to pH 5 and UV C irradiated for a total of 40 minutes. The results in Figure 4 13 suggest that the presence of added chloride ion from the HCl used for acidification does not appear to detrac t from bench scale studies of the UVAC process, likely due the existing abundance of chloride in chlor alkali wastewater. However, the observation that acid selection should be a design consideration for industrial photochemical of Hg wastewater is based on experimental results that greater Hg(0) evolution occurred over the 40 min UV C test in solutions prepared with HNO 3 . Conversely, an interesting observation is that using H 2 SO 4 (dashed line) had the least amount of Hg(0) evolution and slower Hg removal kinetics. For the HCl adjusted solution, the extension of UV C irradiation to 90 minutes produced a high fraction of evolved Hg(0) (Appendix Figure A 1 ). One possible explanation is that weaker oxidant chloride radicals, Cl • and Cl 2 • $ , form at pH 5 in chloride rich solutions, i.e. Eq ( 2 36 ) and ( 2 37 ), which are merely intermediate transient radicals and can be easily scavenged by DOM [10 8] . In comparison with the HNO 3 adjusted solution, the rat e of Hg(0) evolution appears to be slower. This can be explained by literature research findings that UV C photolysis of NO 3 $ produces excess • OH that are less effective in re oxidizing Hg(0) wh en radical scavengers are present, i.e. by addition of the organic molecule CH 3 OH [110] .

PAGE 81

81 Effect of Solution Temperature The observation that solution temperatures increased during UV -C experiments likely occurred from radiative heat transfer in the UVAC batch reactors; the largest increase observed was -T ~25C following 80 min UV -C irradiation, where T0 ~23C and T80 ~50C. Using a time-series record of solution temperature in 100 mL solutions irradiated by UV-C in the UVAC batch reactors solution, an empirical relationship was developed to approximate the rate of change. The relationship in Eq (4-8) was developed using regression analysis of (T/T0)4 vs. tUVC (2 to 80 min), which produced a linear plot (R2 >0.99) with a defined slope and intercept ( Appendix Figure 7-5). The relationship agrees with Stefan-Boltzmann's Law that radiative heat transfers are proportional to differences in temperature to the fourth power, which is used to predict radiation energy flux emitted per unit of time from a blackbody with known absolute temperature, i.e. qT4 " T0 4. (4-8) Elemental mercury solubility . The effect of this temperature increase is an important consideration in the analysis of experimental Hg losses during UVAC testing since Hg(0) solubility is temperature-dependent. The diffusive process of Hg(0)(g) evolution occurs at a rate that is determined Hg(0)aq saturation levels of dissolved gaseous mercury (DGM) or from reduced solubility of Hg(0) occurring by te mperature dependent equilibrium [154] . Therefore, an approximation of the effects of temperature sought to examine its role as a driving force for progression of Hg(0)g evolution as a displacement from equilibrium as function of UV -C contact time in the batch reactors. The results in Figure 4-14 compare time-dependent solution temperature with values for T = T00 . 2 8 5 ! tU V C+ 0 . 5( )1 4

PAGE 82

82 dimensionless Henry's law constant, k H', which was developed by Andersson et al. using elemental Hg(0) in Milli Q DI water for the environmentally relevant temperature range of 5 to 35¡C, shown by Eq ( 4 9 ) and ( 4 10 ) [179] . ( 4 9 ) ( 4 10 ) Effect on solution pH . While solutions were not pH buffered in order to minimize changes to the overall water quality, a record was kept of pH vs. UV C contact time. The results in Figure 4 15 show decreases can occur within ±2 of the adjusted pH that indicate possible changes i n solution chemistry. This can include the significant reaction of sulfide oxidation to sulfate in Hg S dissolution [180] . Increasing temperatures can have an acid effect although, it cannot readily account for the first 15 minutes of irradiation where decreases in pH were the greatest for some samples irradiated for longer periods. It is more likely the result of radical production and Hg or Hg ligand speciation changes, and is therefore considered relevant in thi s investigation of UVAC Hg photo chemistry. Radical production . The most commonly reported variable in temperature sensitive photo reactivity is radical yield. Nitrate, for example, has • OH quantum yield that is positively correlated with higher solution temperature, but negatively correlated with decreasing pH [181] . Similarly, a positive correlation was found for DOM photo produced • OH and increasing temperature (10 to 40.0 ¡C), quantified by reduced apparent acti vation energy at higher temperatures [182] . Over prolonged UVAC k H ' = [ H g ( g ) ] [ H g ( a q ) ] ( D I w a t e r ) k H ' = e xp ! 2 4 0 4 . 3 T + 6 . 9 2 " # $ % & '

PAGE 83

83 treatment solution temperatures increased up to ~50C after 80 minutes UV-C contact tim e. Therefore, significant effects from higher solution temperatures could occur UV-C contact > 60 min, whereby an increase in photo-produced OH could react with Hg and sulfide, organic and chloride ligands. An important consideration is that increasing temperature can lead to a more acidic environment, which co uld explain the pH decrease observed for some solutions (Figure 4-15). Heat-only v UVAC plus heat . Experiments examined whether temperature alone is a confounding effect in bench-scale study of the UVAC process by comparison o f Hg removal at a controlled temperature of 50C both with and without UV-C contact for pH-adjusted chlor-alkali solutions. An increase in Hg losses was not observed although it was expected since an almost two-fold decrease in Hg(0) solubility is predicted at 50C compared with room temperature solution. Experiments tested for UVAC temperature dependence in pH adjusted solution with Sample A by comparing two groups of prepared solutions: (1) solutions were heated to 50C then pH-adjusted and (2) solutions were pH adjusted then heated to 50C. The results are shown in Appendix F igure A-2 and Figure A-3, which were analyzed for variance using single factor ANOVA (! =0.05). The statistical analysis shows the effect of heat alone does not significantly affect Hg removal by filtration or by Hg losses for solutions with pH 3 to 11. A second series of tests incorporated UV-C in the evaluation of temperature dependence with 60 and 120 minute contact time UVAC treatments using pH 11 Sample C. The results are shown in Appendix Figure A-4 and Figure A-5. Solutions that reached 80C produced significant Hg(0) loss (~40%), but not for solutions

PAGE 84

84 reaching temperatures 4 0, 60 and 70 ¡C; whereas solutions continuously cooled in a ~ 4 .0 ¡C water bath had less effective Hg removal by filtration overall. Effect of Sample Age Batch studies on the UVAC process used fresh and "aged" chlor alkali samples. While characterization resu lts from both types provide insight, it is important to consider the effect of sample aging when interpreting the obtained UVAC treatment results. Since an assumption is that particles formed competitively as Hg S and mixed Hg S DOM complexes, the strengt h of these interactions is known to vary with time and availability of dissolved sulfide. Another time dependent consideration is that clusters of smaller HgS(s) particles can gradually form larger aggregates [57, 86] . In order to examine the potential effect of sample aging on particle size distribution, a UVAC experiment used Sample C (pH 11) after 12 months of its collection. The results from two tests are shown in Figure 4 16 for (a) comparison of filter pore size 0.45 with 0.22 !m that had filter media of 0.45 ! m nylon with high protein binding characteristics and 0.22 ! m pore mixed cellulose ester (MCE), which has low protein binding and (b) c omparison of UVAC with and without added heat using the nylon 0.45 !m filters. There was a significant reduction in solid Hg formation during UV C irradiation for both tests, and instead, there was an increase in Hg losses. In the first test with no added heat, the rate of Hg(0) evolution appeared linear, resulting in ~10% Hg losses after 60 minutes and ~20% after 120 minutes UV C contact time . The heated UVAC test ran for shorter UV C contact times, which produced similar Hg(0) evolution to the unheated test for 40 minutes of contact time. The filter pore size comparison results are counterintuitive since there was less Hg removal by smaller pore size over almost the entire 120 minute test, which was observed as approximately 8% less removal as *Hg (s) .

PAGE 85

85 The difference in nylon and MCE is that nylon exhibits highly effective protein bindings whereas MCE does not. The observation should be considered in context with the understanding that protein like structures can exist as LMW phenol functional groups of DOM that can form Hg complexes, and could also occur from microbial by-products[183]. The heated UVAC test also produced interesting results in the aged wastewater: after approximately 10 minutes of heated UV -C, Hg removal rates started to increase instead. The experimental observations may shed like on a key rate determining reaction occurring in the UVAC process; that is, the necessary photo -reduction of Hg(II) to Hg(0), since the heated experiment with aged sample appeared to effectively drive sufficient Hg into its reducible form, possibly occurring from increased radicals produced by the heat of solution. If there were sufficient oxidation potential and recombination substrate (i.e. residual sulfide) in the aged sample, less Hg(0) would have occurred and possibly increased filterable Hg instead. This makes the most sense in the context of previous studies on formation of DGM occurring by DOM by Miller et al., who found a time-dependent component following l igand exchange Hg-DOM complexation. The authors report that over time, Hg-DOM complexes become less reactive as stannous reducible Hg, possibly by forming stronger bonds or through Hg being sterically protected following ligand exchange and re -orientation withi n the Hg -DOM complex [184] . Investigation of Possible Photo-Reactants Hg-Chloride Interactions The high chloride content in the wastewater likely creates favorable conditions for soluble formation of HgCl2 0, HgCl+ and HgCl4 2 – and has been reported to slow photo reduction (UV-Solar) rates occurring by LWM DOM in anoxic acid waters [185] . In

PAGE 86

86 contrast, low chloride concentrations are favorable for the formation of Hg(I) species such as Hg 2 Cl 2(s) when pH # 7 according to the reaction in Eq ( 4 11 ) [46, 167] . K sp = 1.1 x10 18 ( 4 11 ) A gap energy of 3.2 e V for Hg 2 Cl 2(s) indicates it is relatively photolytically stable, and Tennakone et al. proposed photochemical formation of Hg 2 Cl 2(s) according to the reaction in Eq ( 4 12 ) , developed for pH 7 solutions pr epared as suspensions of HgCl 2 (aq) and Hg 2 Cl 2(s) that were irradiated by U V wavelength < 372 nm. The authors predicted the reaction is reversible according to Eq ( 4 13 ) in the absence of electron acceptors and under photo reducin g conditions, in which case Hg 2 Cl 2 reduce d to Hg(0) due to its relatively low standard reduction potential E¡( V ) = 0.27. The authors noted that HgCl 2 behaved as a self regenerating photocatalyst that could measured by O 2(g) evolution occurring in either E q ( 4 12 ) or ( 4 13 ) , which was attributed to unique semiconductor properties of Hg 2 Cl 2 [186] . While the findings are published, it is unclear whether the ph enomenon has been further investigated. Despite this ambiguity, if Hg 2 Cl 2 is present, the process of its photo degradation (and possible formation) could be significant in limiting UVAC Hg removal if it is a solid phase present in filtered *Hg(s). An in teresting observation from the results of Eh pH model for Hg S Cl by Twidwell et al. ( Figure 4 1 ) is that the solid Hg 2 Cl 2 forms in reducing conditions (Eh ~0.25 to 0.5 V) whereas experimental Eh Hg (s) analysis found that only pH 3 produced filterable Hg (s) i n this range. And yet, the experimental observations of redox potential for solutions prepared to pH 7 or pH 9 had an upper Eh limit of approximately Eh " 0.20 V ( Figure 4 10 ); therefore, Hg 2 Cl 2 that has E¡( V ) = 0.27 is less likely to have occurred as a filterable product of UVAC treatment for the pH 7 and pH 9 solutions . H g 2 2 + + 2 C l ! ! H g 2 C l 2 ( s )

PAGE 87

87 (4-12) H g2C l2+ H2Oh v! " ! 2 H C l + 2 H g0+1 2 O2 (4-13) Hg-DOM Interactions The chlor-alkali samples were characterized for organic content, which led to observations of high levels of humic substances. The emission-excitation results in Figure 4-20 indicate DOM with peaks emissions occurring in in the humic acid region, possibly marine-type and hydrophobic humic acid, as well as hydrophilic proteinaceous structures resembling trypt ophan and soluble microbial by-products [146] . Hydrophilic proteins constitute heterogeneous low molecular weight moieties that contain large functional groups, especially nitrogen ; previous studies by Muresan et al. identified this type of DOM as reactive with Hg in municipal wastewater while others have shown its association with copper [187 189] . It became of interest to better understand Hg-DOM interactions and UV-promoted electron transfer reactions. In particular, the effects of pH and UV -C on the electron carrying capacity of DOM and its potential effect on Hg transformations during UVAC treatment. DOM-HgS(s) photo-reactivity. Previous studies on the UV reactions of HgS(s) have noted both that dissolution of HgS(s) occurs, and that its crystal growth is also possible. Both were reported in the study by Pal et al. on the effect of UV on natural and synthesized crystals of HgS(s) in aqueous solutions of HgCl2, &-HgS(s) and lysine (C6H14N2O2). The authors reported observations of partial decomposition of HgS(s) following UV -C irradiation periods of 20 to 70 hrs by UV >300 nm, which presumably formed H2S and free Hg2+, and measured the formation of pipecollinic acid, or PCA (C6H11NO2). The authors also postulated that oxidation of lysine to PCA in which free 2 H g C l2+ H2Oh v! " ! H g2C l2+ 2 H C l +1 2 O2

PAGE 88

88 Hg 2+ acted as an electron acceptor led to the possible occurrence of HgS (s) growth via in situ reducti ve deposition of Hg 0 [190] . The proposed process of Hg (s) particle growth via reductive deposition may be relevant to UVAC Hg removal since some of the chlor alkali samples had noticeable color change over the c ourse of UV irradiation, changing from a clear solution to a suspension of dark grey particles that were retained on the 0.45 µm filter following filtration. Redox c apacity of DOM . An important consideration is whether it is possible to identify the actio n of DOM as an electron acceptor or electron donor, in which the initial oxidation of DOM appears to be a determining factor for the overall capacity of UVAC to produce filtrate with low level Hg concentrations . The effect of UV on DOM redox activity was recently investigated by Sharpless at al. who identified that UV Solar irradiation irreversibly decreases EDC relative to EAC while producing increased ROS 1 O 2 by LMW DOM species [148] . This makes sense in the context of experiments by Aeschbacher et al. who found that oxidized DOM had less EAC overall compared with DOM pre reduced by bulk electrolysis [191] . It was noted in experiments by authors Lu et al. that DOM exhibits increased EDC and EAC in alkaline solution depending on the other species in the respective redox couple, where by EAC increased 3 fol d and EDC increased 6 fold as the solution pH increased from pH 4 to 10. The authors experimentally dev eloped cyclic voltammograms for DOM at various pH and ionic strength s, and found that ECC (termed ETC for electron transfer capacity) peaked at 0.1M KCl but then decreased for greater ionic strength due to molecular compression (i.e. double electric layer effect). It was observed that with increasing pH, both HA and FA like DOM had shifted EEM

PAGE 89

89 fluorescence peaks towards shorter emissions wavelength by approximately 20 and 15 nm, respectively, which was explained by the increased quinone activity in fluorop horic DOM. Whereas the swelling from repulsive forces between ionized DOM groups was a behavior they noted as a possible explanation for the increased ECC and DOM volume expansion occurring at higher pH and electrolyte concentration [192] . This phenomena is characteristic of phenolic groups that cause coulomb -repulsion forces in the negatively charged DOM and less attributed to carboxylic groups that are still capable of inter and intramolecular hydrogen bonding [193] . This and other developments in electrochemical characterization of DOM redox properties are potentially useful in understanding how organic agents facilitate UVAC Hg removal. In particular, the recent development of a method to identify proton and electron transfer capacity for reducible DOM moieties. The work of authors Aeschbacher et al. used mediated electrochemical reduction and oxidation (Appendix Table B-7) and Eh-pH titrations in order to characterize the corresponding changes in DOM. The authors identified pH-dependent behavior of humic acid isolates in which DOM reduction potential generally decreases with increasing pH; that is, while there was an increase in proton uptake, there was a greater overall increase in the relative concentration of reducible functional groups. This was attributed to increased quinone activity that act as the major electron accepting moieties [191, 194] . These findings could be related to the experimental results on UVAC Hg removal from chlor-alkali wastewater; it appears that both conditions of alkaline pH and high ionic strength resulting from high chloride content would significantly influence DOM redox capacity and can reasonably explain the change in electrode potential values

PAGE 90

90 occurring with varying UV C contact time . Furthermore, the Eh vs. UV C time series results ( Figure 4 5 to Figure 4 6 ) exhibit ed pH dependent trends that are in likeness to the electrochemical redox results reported by Aeschbacher et al. A possible explanation for the apparent similarities and is due to the possible formation of UV C driven photo electric effect that is comparable to current induced electron transfer [195] due to the high photon flux per unit volume in the batch reactors ( 20.3 * 10 3 Es per cm 2 min, or 9.6 * 10 9 Joules per cm 2 min ) , which may result in DOM photo activa tion. Role of pH and DOM . It is helpful to consider what is known about the effect of pH on DOM photo transformation occurring by UV promoted CDOM excitation. The authors Paul et al. concluded that pH effects on carboxylic g roups determines the structural state of NOM (e.g. still capable of H bonding at alkaline pH ) and further determines stabilization of certain type of organic radicals. Carbon centered radicals are produced at low pH whereas alkaline pH shows peak free radi cal content from phenol derived semiquinone type moieties. This increase is thought to mainly occur by electron transfer reactions between quinone and phenolic groups, polymerization reactions, or the generation of semiquinone radicals by autoxidation of hydroquinone at alkaline pH [196] . Thus, pH sensitive DOM radical formation may help explains why pre treatment solution pH had a significant effect on Hg removal in aged samples. One could consider the experimen tal UVAC results for aged chlor alkali sample C that was stored at pH 11 ( Figure 4 17 ) and that the aging process altered the UVAC Hg remov al response of this wastewater. This could indicate that Hg DOM photo transformations decr eased after extended periods under alkaline conditions, which is evident by

PAGE 91

91 comparison of UVAC results using fresh sample with initial pHi ~2 (Figure 4-4). It is possible excess hydroxyl groups caused photo-active components to non-photolytic degradation by, possibly by autoxidation of hydroquinones or by electron transfer from ionized phenol groups. This could have in effect “de-activated” DOM ECC, which could only be re-activated by highly acidic solution, or by heat ( Figure 4-16), as this was the only pH that had UVAC Hg removal in the aged chlor-alkali sample. Abs -254. An estimation of UV-C promoted changes in aromatic DOM in the UVAC treated chlor-alkali may be possible. Abs -254 was measured over a 120 min UV C irradiation for pH 2-11 solutions, summarized in Appendix Table A-5. While the corresponding change in DOC from UVA C experiments was not monitored and so, the calculation of SUVA254 requires further investigation. The most significant loss of absorbance occurred in acidic solutions, which was less pronounced in neutral and alkaline solutions, and is likely an indication of the degree of DOM photo -degradation. Hg-Sulfide Interactions The speciation model developed f or the chlor-alkali wastewater predicted that Hg would occur as sulfur-, chloro-, and hydroxy-mercurials; the highest activity was predicted for negatively charged species HgHS2 $, and HgS2 2 for alkaline pH above 9. Therefore, the enhanced Hg removal at pH 11 is unexpected because it is thermodynamically unfavorable for conventional HgS(s) precipitation. This was evident in the findings of Ravichandran et al. that examined similar S/Hg ratios and reported inhibited Hg precipitation occurs at pH >10 and ins tead, 90% of the Hg was soluble by passing a 0.1 +m filter. On the contrary, for solutions with sub-molar equivalent residual sulfide concentrations and a solution pH range of 4 to 8, it was noted that the Hg -S

PAGE 92

92 system rapidly formed HgS (s) crystals that w ere large enough for filtration and removal from solution [86] . On the other hand, sulfur species are known to form crystalline compounds in alkaline sulfidic solutions under chlor ide saturation concen trations. An early investigation of this solution chemistry by Giggenbach et al. identified increased activity of free sulfide ion in highly alkaline pH (18M NaOH) by comparing UV absorption spectra for HS Ð and S 2 Ð for varying NaOH solutions prepared using crystals of Na 2 S ' 9H 2 O (s) or Na 2 S ' 5H 2 O (s) . The findings led the authors to observe that S 2 Ð absorbed photons within in the UV C band up to 246.3 nm wavelength, compared with HS Ð that had a maximum absorption UV wavelength 229 nm [197] . The observations indicate that once HS Ð forms, its transformation is not exp ected to occur via direct photo lysis, but rather through a secondary reaction. Sulfide photo reactivity . Homogeneous sulfide photochemical reactions that produce disulfide (S 2 2 $ ) include photo oxidation of HS Ð via cause direct reduction of water to produce hydrogen gas H 2 . Brandon et al. noted that w hether or not H 2 evolution occurs, it is likely unreactive with respect to HgS reduction [198] . The UV spectrum for aqueous sulfide species d eveloped by Linkous et al. predicted that S 2 2 $ absorbs UV light shorter than 320 nm and acts as a strong optical filter affecting the rate of H 2 evolution [112] (Appendix Figure B 3 ). Since polysulfide Hg S formation can occur from excess sulfide at pH 11 (Table 2 4), the reactivity of S 2 2 $ could be relevant to the UVAC process by acting as an electron scavenge r. Sulfate photo reactivity . Sulfate is a relevant species in UVAC treatment since the water quality characterization shows high concentrations for most of the bat ches tested ( Table 4 1 ). The production of SO 4 2 $ in alkaline so lution increases oxidative HgS

PAGE 93

93 dissolution, and thus could be a possible explanation for the experimental results obtained using aged chlor-alkali sample C. It is possible that SO4 2 – and other sulfoxy intermediates produced by Hg-S oxidation are significant in the UVAC process. An interesting consideration of how this occurs was postulated by Anaf et al. based on experimental findings of UV -532 nm irradiated solutions of cinnabar that produced more SO4 2$ in a 1M NaCl solution compared with DI water, which was accompanied by a decrease in pH. Of particular interest to the identification of possible UV-promoted solid-phase Hg formation are the author’s proposal of a two -step Hg recombination and adsorption process depicted in Eq (4-14) to (4-16). First the photoreduction of adsorbed Hg2+ by HgS in chloride solutions occurs within the conduction and valence bands of HgS (i.e. 0.02 and 2.02 V vs. NHE, respectively) whereby electrons from the photo-activated Hg-S complex are transferred to adsorbed Hg2+ ions which can then favorably forms HgCl4 2 – in t he absence of residual sulfide. T he subsequent photo-reduction of HgCl4 2 – in Eq (4-16) was postulated to produce Hg(0) that can hen be adsorbed by HgS(s) [199] . (4-14) (4-15) (4-16) The proposed metal to metal charge transfer reaction by Anaf et al. is supported by the findings of a recent quantum mechanics study by Da Pieve et al. which investigated the electrochemical stability of Hg photo -reduction occurring by electron transfer on HgS mineral surfaces. By modeling the valence and conductance band edge positions of &-HgS and &-Hg3S2Cl2 (cordierite) in comparison with relevant H g S + 4H2O ! H g2 ++ SO4 2 !+ 8H+ H g S + 4 C l!+ 4H2O ! H g C l4 ( a d s ) 2 !+ SO4 2 !+ 8H+ H g C l4 ( a d s) 2 !+ 2 e!! H g( a d s ) 0+ 4 C l!

PAGE 94

94 potential of redox couples that could occur in solutions such as halogen acids, the authors concluded that cinnabar is not photodegradable across the pH spectr um by direct electron transfer but tha t instead, the formation of HgCl 2 can lead to Hg(0) [200] . Summary Th e experimental observations point to a possible mechanism of Hg molecule photolysis occurring by the electron transfer to Hg, whereby photo produced Hg 0 is then re oxidized and can react with constituents in the solution to form a solid phase complex. In this case, the depletion of photo oxidants appears more significant in acidic pH whereby Hg photo reduction was observed as the predominant transformation process with increasing UV C contact time . The observations suggest that formation of Hg(0) is a pro cess in UVAC driven solid Hg formation, which could have occurred when Hg 2+ adsorbed by HgS (s) precipitate was reduced, and the subsequent Hg(0) re oxidation occurs by photo chemically produced oxidizing radicals and the presence of with electron scavenger s. It seems reasonable to postulate that reductive Hg(0) deposition onto a mineral surface is a possible mechanism of UVAC treatment in which the mineral surface is possibly mineralized carbon, which would grow to form particles sufficiently larger than 0. 45 !m and filtered out of solution. The findings on ORP measurements were helpful in identifying pH and UV contact as key parameters for Hg removal and provided insight into the occurrence of photo redox couples that may be driving the UVAC Hg removal pr ocesses. In the application of UVAC treatment for wastewaters with unknown solution chemistry, it could be useful to first measure changes in ORP for variable treatment conditions including pH a nd testing UV C contact time. With further refinement that is specific to the Eh UVAC resp onse of the influent waters, it may be possible to identify variables

PAGE 95

95 that can reduce the dependence of p H adjustment; for example, such that UVAC contact time to be sole operational parameter. It is useful to consider that when sulfide is exposed to high intensity UV -C irradiation, the thermodynamic instability of HgS(s) is accelerated by photo-oxidants that can occur as OH from photo-excited DOM and dissolved O2, which would deplete residual sulfide and potentially form phot ochemically unstable Hg-sulfide species. As UVAC Hg removal was observed outside the alkaline pH range of sulfide photo reactivity, Hg photo-transformations are most likely to occur in combination with that of competing ligands sulfide and chloride, whic h is investigated by experimental work in Chapters 5 and 6.

PAGE 96

96 Table 4-1. Water quality for chlor-alkali samples A and C expressed as upper and lower limit values. Parametera (unit) Sample A & C observed (2007-2010) Sample C reported (2008-2009) Detection limit Lower Upper Al !g/L < 0.5 160 < 0.5 < 0.5 Ca – 28.1 < 0.5 Fe < 0.2 3.06 < 0.2 Mg (!g/L) 3.67 1300 3.67 < 0.5 Mn < 0.2 60 < 0.2 < 0.2 Ni < 0.3 57 < 0.3 < 0.3 Si – 3.34 < 0.5 Zn 0.3 110 0.3 < 0.2 DO mg/L 4.3 8 – < 0.1 COD (mg/L) 600 6000 – < 1 TOC 18 4000 – < 0.1 DOM b 10 500 – < 0.1 Chlorid e (g/L) Cl– 0.03 2.47 3000 < 0.5 Free Cl (!g/L) Cl2 20 110 42.5 < 5 Sulfate (g/L) SO4 2 – 0.5 10.5 105 < 1 Sulfide (!g/L) S2 – 2 417 10 < 0.1 ORP mV -120 44 44 Cond. !S 20 1764 – TSS mg/L 70 470 – Hg !g/L 3 330 10 < 0.2 pH 2.6 12 3 a.Detected below 0.5 !g/L: Ba, Cd, Co, Cr, C u. Mo, Pb, Sr, Ti, and V b. DO M HA and FA characteristics in chlor alkali s ample C by EEM Fl ourospectroscopy, measured as ~ 50% TOC . “ – “ not measured.

PAGE 97

97 Table 4-2. Visual MINTEQ input criteria used to predict inorganic Hg speciation in untr eated chlor-alkali samples. Parameter (unit) Input Parameter (unit) Input Al !g/L 0.05 Chloride (g/L) Cl– 7.5 Cu 0.02 Free Cl2 (!g/L) Cl2 30 Fe 0.05 Sulfate (g/L) SO4 2 – 6 Mg (mg/L) 1.3 Sulfide (!g/L) S2 – 4 – 400 Mn 0.06 Ni 0.06 ORP mV -80 Zn 0.01 Mercury (!g/L) Hg 20 DO mg/L 7 pH 2 – 12 COD 569 Table 4-3. UVAC-treated filtrate Hg concentration for chlor -alkali sample C. (*CF) for tUV C = 30 min batches adjusted to pH 11 .5 with pHi =2.6 to 3.9. Batch ID C0 [HgT] (g/L) *CF *[Hgaq] (ng/L)a Hg removal (%) 2 3.7 1.4 99.96 3 8.0 19.0 99.76 4 3.0 5.3 99.83 5 5.6 90.0 98.41 6 3.5 99.0 97.15 7 4.0 1.3 99.97 8 2.0 2.1 99.90 9 2.4 0.5 99.98 10 2.2 2.0 99.91 a.Effluent analysis conducted by third party lab.

PAGE 98

98 T able 4 4 . Standard electrode potentials of relevant Hg and aqueous redox couples in solution at equilibrium . Redox couple E¡(V) Redox couple E¡(V) Hg(II)/Hg(0) Ð Hg(I) Ion redox boundary 2Hg 2+ + 2e Ð ! Hg 2 2+ 0.91 2SO 4 2 $ +4H + +2 e $ ! H 2 SO 3 2 $ + 2H 2 O 0.172 Hg 0 (l) + HgCl 2(aq) ! Hg 2 Cl 2(s) 0.59 {HS Ð }/{SO 4 2 Ð } 0.247 Hg 0 (l) + Hg 2+ ! Hg 2 2+ 0.115 Hg(I)/Hg(0) Hg(II)/Hg(0) Hg 2 2+ + 2e Ð ! 2Hg 0 0.79 Hg 2+ + 2e Ð ! Hg 0 0.86 Hg 2 Cl 2(s) + 2e Ð ! Hg 0 + HgCl 2(s) 0.586 HgCl 2 + 2e Ð ! Hg 0 + 2Cl Ð 0.41 Hg 2 Cl 2(s) + 2e Ð ! 2Hg 0 + 2Cl Ð 0.27 HgCl 4 2 Ð + 2e Ð ! Hg 0 + 4Cl Ð 0.46 Hg 2 2+ + 2e Ð ! 2Hg 0 + Hg 2 Cl 2(s) 0.27 HgO (s) + H 2 O + 2e Ð ! Hg 0 + 2OH Ð 0.0984 Hg 2 2+ ! Hg 0 + Hg 2+ 0.115 HgO (s) + 2H + + 2e Ð ! Hg 0 + 2H 2 O 0.926 HgS (s) + 2H + + 2e Ð ! Hg 0 + H 2 S (g) 0.72 Data source from Lindsay et al. and Stumm and Morgan [46, 201] .

PAGE 99

99 Figure 4-1. Eh/pH diagram for an oxic system groundwater system. (a)Hg-S-H2O and (b) Hg-S-Cl-H2O. Sourced from Twidwell et al. [176]Figure 4-2. Effect of pH on pore-size defined DOC in chlor -alkali sample C. Pore sizes 0.22 m MCE and 0.45 m nylon membranes; C0 = 10 g L1 Hg, pHi ~12. ! "# #! $# %!! %"# %#! %% $ " &'()*+ ,-./ 01,234+25 !6""+7, !68#+7,

PAGE 100

100 Figure 4 3 . UVAC Hg removal by varying filter pore size in chlor alkali sample A. Untreated and 5 min UVAC treated chlor alkali samp le A; C 0 = 11 µg/L Hg and pH i ~11. Figure 4 4 . UVAC effluent filtrate Hg (*C F ) in chlor alkali sample C for pH 3 to 11. C 0 ~5 to 12 µg/L Hg , and pH in . t UVC = 180 for Batch 1, t UVC = 30 min for b atches 2, 3 (error ±5%). ! !"# !"$ !"% !"& ' % '"# !"& !"$( !"# )*+,-./01.-/ 2*3-/4567 8 ! 9/:;,.-<,-= >89/?@A8 B ,.-<,-=/4(/6*;/,:CD7 EF/G-61C<+ ' H 48 I/ J87 !"! !"# $"! $"# %"! %"# & # ' ( $$ )*+ , -+ ./+ 0+ ! -12 3, 4.567+$ 4.567+% 4.567+& ! #! $!! $#! $ % & 3, $$+++) * , . / 0+ n -12

PAGE 101

101 Figure 4-5. Potentiometric Eh-pH results for chlor-alkali samples A and C. Using 0.1M HCl titration; A1 C0 ~6 !g/L Hg, A2 C0 ~50 !g/L Hg, and aged sample C C0 ~20 !g/L Hg. Figure 4-6. Eh-UVC for chlor-alkali sample C with varying pH and contact time. !"# !"$ !"% !"& !"' !"( !") !"* # $ % & ' ( ) * + #! ## #$ ,-./ ! 0 12 3# 3$ 45 ! "#$ "#$ "#% " &" '" (" )*+,-. % $"#/ $$#/ 01+0"2+&"+3456+74 8 9-0 ,:;<.

PAGE 102

102 Figure 4 7 . Eh UVC for comparison of solution pH in chlor alkali sample A . Contact time, t U VC " 80 min where A2 contains added HgCl 2 . ! "#$ " "#$ "#% "#& "#' " $" %" &" '" ()*+,. % / ' 00 +1-*2$3*4 " 5*6"*789:*;8 < =,4 +>?@! "#$ " "#$ "#% "#& "#' " $" %" &" '" ()*+,$#A . / A 00 +B-*203*4 " 5*&*789:*;8

PAGE 103

103 Figure 4-8. Eh-UVAC for Sample A1 pH-dependent for tUVC up to 80 minutes. (a)filtered *Hg(s) and (b) total *Hg Loss with (c) Eh values. !"# !"$ !"% !"& ' ( ) * ! !"' !"# !"( ( ) * + !"# ! !"# !"$ !"% !"& ! #! $! %! &! ( ) * ,-./012345'67/!8% 9:;<7=: 0>4=:7
PAGE 104

104 Figure 4 9 . Eh UVAC for Sample A2 pH dependent for t UVC up to 80 minutes. (a) filtered *Hg (s) and (b) total *Hg Loss with (c) Eh values. !"# !"#$ !"#% !"#& !"#' ( % ) # ! !"( !"$ !"* % ) # + !"$ ! !"$ !"% !"& ! $! %! &! '! % ) # , -./ 01234 5$67/ ! 89!7:;<=7>; 0?47>;7=@AA 67 (B0C/;7EF1@GDH 67 (B0C/ I
PAGE 105

105 Figure 4-10. Regression analysis of Eh-UVAC Hg(s) removal in chlor-alkali sample A. Values plotted from tUV C = (0, 5, 15, 35 and 80) min; (a) A1 C0 ~6 !g/L Hg, and (b) A2 C0 ~50 !g/L Hg (added HgCl2). ! "#$ " "#$ "#% "#& "#' "#$ "#% "#& "#' ( )*+, ! ./ ,0-+ 1234567 ,6-8(9+: ! ;+&++./ ? @ A ! "#$ " "#$ "#% "#& "#' "#($ "#(% "#(& "#(' ) *+,! . /0 -1., 2345678, -9.:$;,< ! =,>",?0@A,/0 % B ( !"# $%&'( !"# !"# $%&'( !"#

PAGE 106

106 Figure 4 11 . UVAC Hg losses for pH adjus ted chlor alkali samp le A. Measured as reactor adsorbed Hg or evolved Hg(0) for 15 and 80 min UV C. ! !"!# !"$ !"$# !"% !"%# !"& '()* '()+ '(), '()* '()+ '(), (-)./0123. 4 056 (-7!8)9:-;2;)2 ?@ A!)=;>)2 ?@ (-) B366 7C8)D% E) F ! G) #!)H-IB) (-) ! !"!# !"$ !"$# !"% !"%# !"& '()& '()+ '(), '()& '()+ '(), (-)./0123. 4 056 (-7!8)9:-;2;)2 ?@ A!)=;>)2 ?@ (-) B366 708)D$ E) F ! G) J)H-IB) (-)

PAGE 107

107 Figure 4-12. Regression analysis for Eh-UVAC Hg(0) losses in chlor-alkali sample A. Values plotted are from UV-C contact period for tUVC = 15 and 80 min. !"#$ " "#$ "#% "#& " "#"' "#( "#(' "#$ "#$')*+,! -./,"-+0122 ,3-4$5+6 ! 7+'"+8/90+./ % : ; !"#$ " "#$ "#% "#& "#' " "#"$ "#"% "#"& "#"'()*+! ,-.+",*/011 +2,345*6 ! 7*&*8.9/*-. : ; <

PAGE 108

108 Figure 4 13 . Comparison of acid pH adjustment for UVAC in chlor alkali sample C. (a) Hg removal by filtration and (b) Hg losses occurring by UV C (12 mo nths aged), C 0 ~20 µg/L Hg, adjusted to pH 5 from pH in = 11. !"# !"#$ !"#% !"#& !"#' ( ! (! $! )! %! ( * +, ./ 0, -1 2 34 * +5671 8-9 8 ! :; " 8<; # +=1/,8> +?1 !"@ !"@A !"' !"'A !"# !"#A ( ! (! $! )! %! +,-0! 1 2 34 * +5671 8-9 8 ! :; " 8<; # + B1/,8>/CD??

PAGE 109

109 Figure 4 14 . Effects of solution temperature on Hg(0) solubility in UVAC batch reactors . Initial T 0 ~23¡C with k H' ~0.30 [Hg (g) ]/[Hg (aq) ], with 1 00 mL solution volume. Figure 4 15 . Effect of UV C on pH for DI and chlor alkali sample A, adjusted pH 3 to 11. Sample A C 0 ~ 6 to 50 µg/L Hg ; statistically insignificant difference between samples A1 and A2. !" !# "! "$ %& %' () ) !) %) ') #) &)) ) !) %) ') #) *+,-. /0123 4 5 6/ ) 7!"3 4 538 93 :.;<0=>03 ?@6) 83 >+,AB:,:CD 6 ! ?-) 7)E") 833 C FG H 5 61:.8 ! " # $ % & ' ( )* )) * $ )* )$ !* !$ "* "$ +, ./0 12345 67 892+:;<=>?) @>?!

PAGE 110

110 Figure 4 16 . Heated UVAC Hg removal in chlor alkali sample C, variable filter pore size. Sample aged for ~12 months at pH in 11; C 0 ~20 µg/L Hg for (a) Filtration only by 0.22 µm MCE and 0.45 µm nylon ( b) Heat + UVAC with 0.45 µm nylon filters. !"# !"$ !"% & ! '! $! &(! ) *+, -./01 2,3, ! !"'456.5 !"((56. !"$ !"% & ! &! (! 7! '! 4! 4! , 85) *+ 9 ,5 -./01 2,3, ! !"'456. -: 155*+,5 ;0<=5>?5@A.;B:< -C155*+,5 85 >A:)5 >?5@A.;B:< ,5-D>5&&1E5, ! F(!556?3G5>? :?AH5&(5.;0)IJ

PAGE 111

111 Figure 4 17 . UVAC Hg re moval in chlor alkali sample C with varying pH adjustment. Sample aged for ~12 months at pH in 11; C 0 = 20 µg L 1 Hg and pH i 11) from untrea ted and UVAC treated (t UV C = 60 min) solutions with varying pH. Total losses as Hg Loss " 8% (error ±5%). 0 0.2 0.4 0.6 0.8 1 2 5 7 9 12 1 C F c C pH C 0 (t UV-C = 60 min) , untreated *C, UVAC-treated

PAGE 112

112 Figure 4-18. Replicate results for Eh-UVAC Hg removal for chlor-alkali sample A. 2 month aged s ample A1 (a) *Hg(aq) filtrate (b) total *Hg losses with (c) Eh (V) values for tUVC = 0 to 70 minutes. !"# !"$ !"% !"& !"' ( ) $ ! !"!# !"( !"(# ) $ ! !"* !"+ !"$ !"& ! *! +! $! ) $ ,-./01234 5(6 07489/!:9$9;<=>9?< 0@4A?<0B497C1DE@F 89(G0A/H=A/4 0I4?<9>DBB 89 (G0A/=/!4 0J4 KL90.4

PAGE 113

113 Figure 4-19. Replicate regression of Eh-UVAC Hg(0) losses for chlor-alkali sample A. 2 month aged sample A1, values plotted are from tUVC = 0 to 70 minutes (a) filtered *Hg(s) and (b) Hg(0) losses. !"#$%&'($)*+$,$-./0$1. ! !"# !"$ !"% !"& !"' !"% !"( !"& !") * +,-. ! / 01.2/--3456789 : % .8/-01.2/-3456789;<=>?@-!A-*'A-:'A(!-5BC ! !"# !"$ !"% !"& ! !"!' !"( !"(' )*+, ! ./,!-+0122 3 % ,4-./,!-+0122 5++6789:+(';+? !"# $%&'( !"# *Hg(s) removal ;

PAGE 114

114 Figure 4-20. DOM characterized by EEM fluoro-spectroscopy in chlor-alkali sample C. Sample adjusted to at p H with NaOH; C0 ~20 !g/L Hg and ~50 mg/L DOC, measured a fresh batch sample. ! "#$% ! "#&% ! "#& ! "#"% "'()*+,+*-. /.01 23"4 ! 25"4 ! 2$"4 ! 2""4 ! $6" ! $3" ! $5"4 ! $$"4 ! 2"" 2$% 2%" 27% 5"" 5$% 5%" '0*88*-. /.01

PAGE 115

115 CHAPTER 5 DETERMINATION OF UV EFFECT ON AQUEOUS MERCURY REMOVAL BY SULFIDE PRECIPITATION Scope An investigation of Hg speciation studies and experimental results on UV irradiated solutions containing aqueous Hg over a gradient of sulfide concentration seeks to understand UVAC-related Hg photochemical transformation mechanisms. Experimental Parameters In order to limit non-photocatalytic oxidation reactions, some UV experiments were conducted under ultrapure nitrogen gas (Airgas) in an atmosphere controlled glove box. The conditions were not included in further experiments as there was no information gained by the setup. The instrumental detection limit by AF was found to be 6.0 ng L1 Hg. Mercury Sulfide Solutions UV experiments were conducted in UVAC batch reactors with artificial Hg wastewater prepared using pH-buffered DI water (Appendix Table A-3) as the canvas for Hg solutions with a target concentration of 0.1 M Hg using Hg(NO3)2. Solutions were prepared according to secondary treatment guidelines that use a mass ratio of ~0.6 S/Hg in order to enable Hg removal by submicron filtration; the inclusion of mass ratios of S/Hg > 1 permits an investigation into the effects of residual sulfide. UV Experimental Results Solutions were prepared with Hg and sulfide to concentrations that herein defined as trace, molar equivalent with Hg and excess residual sulfide. UV -C Hg removal was obtained in pH-buffered solutions for samples with ID “B.0” and “S.0” for sulfide free and trace-sulfide compositions, respectively. The sample compositions and

PAGE 116

116 prec ipitation reactions ( Table 5$ 1 ) illustrate thermodynamically favorable pathways for Hg(s), and Table 5 2 summarizes UV %Hg removal in the prepared solutions. The results for B.0 solutions in Figure 5 $ 1 shows the acidic samp le 3.B0 formed ~10% more filterable Hg compared with 7B.0 and 11B.0 , even in the absence of sulfide. For each solution pH tested using DI Nano pure water , there was little change in Hg removal indicating there was no formation of filterable *Hg(s) during 70 min irradiation . By comparison, of pH adjustment , the greatest filterable Hg occurred at acid pH 3, while UV irradiation produced higher Hg losses while higher Hg solutions prepared with alkaline pH 11 adjustment. Trace Sulfide The results from pH ad justed solutions with added trace sulfide (S.0) each show different effects of UV on both Hg losses and the formation of filterable Hg. A comparison of those results is shown Figure 5 2 along with corresponding measurements of Eh redox potential. The amount of filterable Hg in the pH 3 buffered solution, 3S.0, was considerable before irradiation, ~60% at t 0 , whic h occurred in the solutions that had an equivalent of 0.02 H 2 S/Hg mol ratio. By comparison, the pH 3 sulfide free sample , 3.B0, had with ~23% Hg(s) at t 0 . Given the high chloride content from the pH 3 buffer, adsorption of HgCl 2 0 HgCl + or Hg 0 onto HgS (s) by Eq ( 4 15 ) and ( 4 16 ) may account for the initial filtration. The observed Hg losses occurring at t 60 were significantly higher for the acidic pH solution 3S.0 in comp arison with 7S.0 and 11S.0. A considerable factor in the UV mediated evolution of Hg(0) may be the form of sulfide ; for pH 3, H 2 S sulfide species correspo nded to higher Hg losses compared sulfide species as HS Ð , which corresponded to fewer Hg losses for pH 11 where the predominant Hg species is

PAGE 117

117 predicted to have been Hg(OH) 2 . The amount of filterable Hg remaining in solution at t 60 for 3.S0 also decreased from a value of *Hg (s) = 56±3% at t 0 to *Hg (s) = 41±4% at t 60 , which suggests that photo reduction of free Hg 2+ led to Hg(0) g evolution due to the decreased availability of sulfide to form HgS (s) . H ence , it is predicted that this reduced the possibility fo r Hg 0 deposition onto HgS (s) . Given the high chloride content of the solutions, it is possible that chloro mercurials HgCl 2 and HgCl 4 2 Ð were the predominant species, which underwent photo reduction thus resulting in large Hg losses . If so , the abundance of free chloride ions did not appear to facilitate re oxidation of Hg(0). The oxidation potential increased during UV irradiation for each pH solution where pH 3S.0 and 7S.0 had similar values, which does not elucidate relevant pathways that could explain differences in the Hg photo transformation results. The lowest oxidation potential was measured for 11.S0 that corresponded to lower Hg(0) evolution. Since the buffer used in the pH 11 solution consists of sodium bicarbonate NaOH buffer, it is possible that formation of long lived carbonate radicals CO 3 • Ð capable of oxidizing Hg(0) [202] were effective in decreasing the potential for Hg photo reduction by • OH [131] . Residual Sulf ide Solutions were prepared in pH 7 buffered DI water to produce S/Hg mol ratios of approximately 4 and 40 using 1.0 or 0.1 mg L 1 total sulfide and target C 0 ~20 !g L 1 Hg. The formation of HgS is predicted to occurs by the same reactions as Soluti on 7.S 0, which is provided in Table 5$ 1 . In order to examine how Hg S complexation strength variations that may occur over time affect photo transformations of molecules of Hg S (s) , the solutions had 1 and 24 hours of dark mixing before UV C irradiation ( 90 minute

PAGE 118

118 contact period) . A comparison of results in Figure 5 3 (a) shows that both solutions had decreased *Hg (s) removal for 10, 30 and 90 minutes. The effect of reduced filterable Hg (s) that occurred simultaneously with increasing Hg losses was most sig nificant for solutions with less residual sulfide. The lower S/Hg mol ratio value of ~4 resulted in both higher Hg dissolution and Hg losses following 90 minutes irradiation regardless of mixing period. By comparison, of Hg losses in solution with 1 and 2 4 hours of dark mixing in Figure 5 3 (b), there was an approximately 10% more loss for 1 hour UV C , measuring 37% and 26% (±5%) for 24 hours . The photo reactivity of Hg decreased following longer mixing periods, it is likely tha t Hg molecules with stronger intramolecular Hg S bonds are more from photo dissolution. Previous study on HgS photo dissolution by Hseish et al. identified a conceptual model in which HgS (s) produces both free ionic Hg 2+ and gaseous Hg 0 [203] , although the mechanisms of reactions do not include sulfide. The experimental results suggest that sulfide photo oxidation could act to promote HgS (s) dissolution and Hg photo reduction. U V Wavelength and Sulfide Photo R eactiv ity Irradiation by UV B selectively excludes UV absorption by sulfide species HS Ð or H 2 S (Appendix Figure B 3 ) and so an experiment was designed to compare the effects of sulfide photo reactivity in the absence of chloride ion and molar equivalent sulfide concentrations. The experiment compared UV C and UV B wavelength using pH 7 buffered solutions prepared with Hg(NO 3 ) 2 and sulfide; solutions with S/Hg mol ratios 0.4, 0.6 and 1.2 have sample ID NS(0.4), NS(0.6 ) and NS (1.2), resp ectively. The results in Figure 5 4 shows that UV C dissolution of Hg S is delayed for the first 15 minutes of irradiation, occurring thereafter at comparable rates regardless of S/Hg ratio. In contrast, Hg losses measured as Hg (0) by photo reduction was apparent within 5

PAGE 119

119 minutes of irradiation. Comparison of S/Hg ratios 0.6 and 1.2 show that reduced concentrations of HgS (s) formed during dark mixing result in significantly faster Hg(0) evolution . Since Hg losses occurring duri ng the first 15 minutes of irradiation did not correlate with less *Hg(s) removal, the Hg(0) source is unlikely by photolysis of HgS (s) molecules formed when the S/Hg mol ratio was calculated to be less than 1. While added nitrate from Hg(NO 3 ) 2 could theor etically produce highly oxidizing • OH radical, it does not appear more significant than Hg S speciation and sulfide photo reactivity. This effect may be evident by comparison with UV B irradiated solutions that is predicted to eliminate photo reactions occ urring by H 2 S and HS Ð , and would instead occur by NO 3 2 Ð , OH Ð and Hg. The UV B experimental results shown in Figure 5 5 unexpectedly produced an increase in filterable *Hg (s) for irradiation periods longer than 15 minutes. The in itial rate of Hg photo reduction appears slightly greater for samples NS(0.6) and (1.2) in comparison with UV C, which agrees with findings by Zhang et al. of higher Hg photo reduction by UV B compared with UV C [110 ] . Following the 60 minute irradiation period, the results show that greater *Hg(s) removal occurred for sample NS(1.2) which had the highest ratio of bisulfide ion, approximately 0.5 HS Ð /Hg. This potentially point s to a key experimental finding in that increased formation of filterable Hg(s) occurs when HgS (s) photo dissolution does not; the HgS (s) precipitates can instead adsorb additional Hg. The results may help to identify H 2 S or HS Ð as the primary photo reactive sulfide species in in UV C promoted Hg S dissolution. Investigation of UV Effects on Hg S Dissolution The photo reactivity of Hg S is considered fo r species that can occur by order of increasing hydration energies: HgS < HgS ' H 2 O < HgOHSH 0 < Hg(SH) 2 0 [60 ] ( Appendix

PAGE 120

120 Table B 3 ). During solution preparation the addition of Na 2 S crystals to the Hg(NO 3 ) 2 rapidly forms HgS as HgS ' H 2 O, which has the lowest optical transition energy of all Hg S species, E (S 0 ! S 1 ) of ~2 eV. Under thi s assumption, if HgS ' H 2 O is predominant after 1 hr of dark mixing, it will readily undergo photolysis by UV C and release Hg 2+ into solution. If the isomer HgOHSH 0 formed over time, approximately 1.5 to 2 times more energy ( ~6 eV) would be theoretically required for this molecule to undergo photolysis [60] . It is also important to consider that Hg may also form bi dentate Hg S binding through tetrahedral coordination, resulting in weaker Hg solvation relative to highly soluble polysulfide Hg species such as HgSnSH $ (Appendix Table B 3 ). Based on the experimental evidence and supporting theory, sulfide photo reactivity is a significant factor in the observed increased Hg solubility and pho to reduction on reactions during UV irradiation. Photochemical oxidation of HS $ produces S 0 , which would become complexed mostly as S 2 2 Ð in alkaline pH. While the extent of this reaction is likely determined by limiting photo reactants S 0 and HS Ð , it is not limited to neutral and alkaline pH; in acid solutions with H 2 S, oxidation reactions occurring by H 2 O 2 produce S 0 [204] , whereas oxidation reactions by • OH and NO • 3 can produce HS • radical [205] . By using pH 8 as the upper constraint for photochemically transformations in the experimental solutions, it is possible to postulate relevant photo produced sulfide species according to the findings of pre vious research. Photo reactions of sulfide and HgS. Residual sulfide occurring as HS Ð in neutral and alkaline solutions initially reacts to form HgS (s) and neutral complexes HgOHSH 0 and Hg(HS) 2 0 . As photochemical oxidation of the excess HS Ð forms increa sing levels of S 0 saturation it is possible that polysulfide Hg complexes form.

PAGE 121

121 While polysulfide Hg is predicted to be less photo reactive than HgOHSH 0 , the excess sulfide can produce larger polysulfide anions, which could react with the Hg complex and c ause its destabilization. This can occur by photo oxidation of anion S 2 2 $ to S n • Ð by HS • or S • Ð when the loss of an electron from highest occupied molecular orbital produces larger polysulfide anions, i.e. 2S n • Ð ! S 2n • Ð for n = 2$ 4 [111] . Based on the experimental findings and supporting literature research , solutions at neutral pH would undergo a photo transformation sequence by HS • radicals , developed in Eq ( 5 1 ) and ( 5 2 ) , which describe UV promoted Hg S dissolution for polysulfide complexes HgS n SH $ and Hg(S n ) 2 2 $ . Whereas i n alkaline solutions , if the initial reaction with excess sulfide produces HgOHSH 0 or HgS n OH $ , upon UV irradiation the possible soluble products are Hg polysulfides that can similarly dissociate in secondary processes according to the proposed formation reactions with S 2 2 $ depicted Eq ( 5 1 ) and ( 5 2 ) . ( 5 1 ) ( 5 2 ) Radical Scave ngers The results obtained in UVAC experiments and with synthetic solutions of Hg S Cl convey appear to have a common pathway for photochemically produced Hg(0) losses whereby the reversal can occur by means of Hg 2+ adsorption onto HgS precipitates. This would be possible when Hg(0) re oxidation occurs by photo chemically produced oxidizing radicals and the presence of with electron scavengers. It is known that photo reducible organic matter can promote Hg(0) photo oxidation by acting as an effective elec tron scavenger [94] . On this basis, Hg(NO 3 ) 2 spiked solutions H g S n ( H S ) ! + H S • " H g S 2 n 2 ! + H S ! + H + H g S 2 n 2 ! + H S • + e ! " H g 2 + + S 2 n 2 ! + H S !

PAGE 122

122 were prepared with residual sulfide occurring at 4 and 40 S/Hg and organic matter as ~10 mg L 1 DOC as Suwannee River humic acid standard (IHSS SR HA). The solutions were irradiated for 90 minutes following 1 and 24 hour periods of dark mixing and measured for filterable *Hg (s) ; the measurement of *Hg losses are summarized in Table 5 4 , which are postulated to have occurred by the evolution of gaseous Hg(0) formed during UV treatment. Summary The results in Figure 5 6 indicate that photoreactions did not deplete filterable Hg S species, which is evident by comparing *Hg losses after 10 minutes of U V C irradiation for solutions with 1 hour and 24 hours of dark mixing. The addition of DOM reduced Hg(0) evolution occurring for 90 minute irradiation periods by approximately 50%; while significantly less Hg(0) evolution occurred for the solutions that ha d longer mixing periods, there was no apparent affect on Hg S dissolution and inconclusive evidence of increased filterable Hg. This could be explained by growing evidence that non purgeable dissolved gaseous Hg 0 (aq) adsorbs onto HgS (s) [190] or onto suspended matter thereby acting as a significant form of particulate Hg in the natural environment [206] . By comparison with the UV B results, in which the possible increase filterable Hg formed concurrent with increased Hg(0) evolution, aqueous Hg was more likely dissolved Hg 2+ and stable aqueous HgS 2 2 Ð or polysulfide species that are postulated to form with sulfide photo reactions by Eq ( 5 1 ) and ( 5 2 ). The electron donating and accepting capacity for SHRA was investigated by monitoring changes in SUVA 290 measurements, summarized in Table 5$ 3 , which is predicted to be comparable to SUVA 280 absorbance that has been identified as a useful measure of DOM EDC changes as well as CDOM activation [75] . The decrease

PAGE 123

123 in DOC (see Appendix Table A 6 ) and SUVA 290 suggest that a higher rate of DOM photo oxidation occurs by UV C relative to UV B . Following UV C irradiation the measured changes in absorbance for Hg SRFA solutions appeared comparable to reported changes in absorba nce for IHSS SRFA during a 52 hour UV Solar and oxygenation experiment by Sharpless et al. [148] , which has similar calculated photon energy emissions as 60 minutes UVAC treatment. Furthermore, a comparison of the photon fluence in the batch reactors and their simulated UV Solar irradiation reveals that the qualitative shift in SUVA (C/C 0 ) resembles the UV C system. Therefore, it may be a reasonable assumption that the observed decrease in SUVA 290 values i ndicate decreasing quantum yield for H 2 O 2 (and • OH) from diminished DOM charge transfer reactions. This is predicted to correspond with increasing 3 DOM • electron acceptance and 1 O 2 quantum yield according to the observations of Dalrymple et al. [147] . The main outcome from this set of UV experiments was the observed effects of Hg removal from solutions prepared with DOM and and varying concentrations of sulfide. The experimental results indicate that additio n of DOM promoted re oxidation of photo reduced Hg, which is in agreement with extensive findings reported in the literature of this dual action by UV activated DOM. A significant variable in determining the effective role photo oxidation appeared to have been solution pH, by which comparison of results from pH 11 solutions indicate that alkaline solutions enhance this action. The action by photo reducible DOM moieties such as phenol groups [207] may offer an e xplanation for the increase in DOM oxidation capacity , which would occur by ionized phenol groups solutions with pH # 9 that act as electron acceptors.

PAGE 124

124 Table 5 1 . Solution compositions for sulfide free (B.0) and trace sulfide (S.0) samples. Solutions prepared with pH buffered DI water a ; 2 hrs t DARK mixing. pH/ID a Hg (nM) S (nM) Cl Ð (mM) S/Hg mol Sulfide Hg (s) precipitation reaction G 0 (1) (kJ/mol) 3.B0 127.2 23 Ð 116 7.B0 129.4 137 11.B0 125.6 137 3.S0 128.5 2.3 23 0.02 100% H 2 S Ð 111 Ð 129 7.S0 119.3 1.3 0.01 56% H 2 S 44% HS Ð Ð 22 2 Ð 183 11.S0 132.3 1.3 0.01 100% HS Ð Ð 183 a. Reagents used for to prepare pH buffered solutions prov ided in Appendix Table B 3 (1) G 0 values c alculated for & HgS based on data provided by Stumm and Morgan [ 46] . Table 5 2 . E perimental UV C Hg removal results for B.0 and S.0 solution samples . pH/ID Hg (nM) Cl Ð (mM) S/Hg mol ratio UV C (min) %*Hg(s), 1 (*C F /*C) %*Hg Loss, 1 (*C/C 0 ) 3.B0 127.2 23 t 70 23 ±3 6 ±0.2 7.B0 129.4 t 70 9 ±3 7.4 ±0.2 11.B0 125.6 t 70 13 ±2 10 ±0.3 3.S0 128.5 23 0.02 H 2 S t 60 42 ±4 65 ±3 7.S0 119.3 0.006 H 2 S 0.004 HS Ð t 60 9 ±3 17 ±3 11.S0 132.3 0.01 HS Ð t 60 8 ±3 9 ±2 H g ( O H ) 2 + 2 H C l ! H g C l 2 + 2 H 2 O H g ( O H ) 2 + O H ! ! H g O + H + + H 2 O H g ( O H ) 2 + O H ! ! H g O + H + + H 2 O H g C l 2 + H 2 S ! H g S + 2 H + + 2 C l ! H g C l 4 2 ! + H 2 S ! H g S + 2 H + + 4 C l ! H g ( O H ) 2 + H 2 S ! H g S + 2 H 2 O H g ( O H ) 2 + H S ! ! H g S + H 2 O + O H ! H g ( O H ) 2 + H S ! ! H g S + H 2 O + O H !

PAGE 125

125 Table 5 3 . SUVA 290 for IHSS SRHA and SRFA for 5, 30, and 60 min UV irradiation. UV C (min) SRHA SRFA UV B (min) SRHA SRFA 0 7.67 4.94 0 7.44 5.09 5 7.49 4.92 5 7.13 5.07 30 6.53 4.41 30 6.97 4.92 60 5.31 3.77 60 6.53 4.89 Table 5 4 . UV C *Hg losses for pH 7 buffered solution with SHRA and residual sulfide. Hg as Hg(NO 3 ) 2 (±5% error). S/Hg mol ratio (t DARK ) DI only; C 0 ~15 !g/L Hg 10 mg/L DOC; C 0 ~20 !g/L Hg 4 (1 hr) 37.6% 19.2% 4 (24 hr) 26.0% 10.0% 40 (1 hr) 26.3% 13.8% 40 (24 hr) 21.7% 10.9%

PAGE 126

126 Figure 5 1 . UV C Hg removal in pH buffered B.0 solutions prepared in DI water. (a) *Hg (s) removal and (b) total *Hg Loss ; C 0 ~25 !g/L Hg(NO 3 ) 2 , 2 hr t DRK . ! !"# !"$ !"% ! & #& %& '! ()*+,()*./0,, ! !"# !"$ !"% ! & #& %& '! ()*+,()*./0,, ! !"# !"$ !"% ! & #& %& '! ()*+,()*./0,, +1-..##"2! +3-. '"2! +4 -.. %"2! 5 678. +9:;-

PAGE 127

127 Figure 5 2 . UV C Hg removal in pH buffered S.0 s olutions with trace sulfide. (a) *Hg (s) (b) *Hg Loss compared with (c) Eh; C 0 ~20 !g/L Hg(NO 3 ) 2 , 2 hr t DRK . !"# !"$ !"% !"& !"' ! #! %! '! $"(! )"(! **"(! ! !"* !"# !"$ !"% !"& !"' !") ! #! %! '! $"(! )"(! **"(! ! !"* !"# !"$ !"% !"& !"' !") ! #! %! '! $"(! )"(! **"(! + ,-. /0123 /4 355675/ ! 3 /83 9:; <=>> /? 3 9:; />35 (a) *Hg (s) r emoval

PAGE 128

128 Figure 5 3 . UV C Hg removal for pH 7 buffered solution s with residual sulfide. C 0 = 15.0 µg L 1 Hg(NO 3 ) 2 with variable mixing (t DARK ); (a) *Hg(s) removal and (b) *Hg Loss . ! !"# !"$ !"% !"& ' ! (! %! )! $!*+* !"#$ $*+* !"#$ $!*+* !%$ $*+* !%$ +,-.*/01 2 34560 74 8**9-.7:8*3;/0<41 !"$ !"% !"& ' ! (! %! )! $*+* !"#$ $*+* !%$ +,-.*/01 2 34560 7=8** 9-. >0::

PAGE 129

129 Figure 5 4 . UV C Hg removal for pH 7 buffered solutions with molar equivalen t sulfide. C 0 = 9.5 13.5 µg/L Hg(NO 3 ) 2 with variable S/Hg and 24 hr t DARK ; (a) *Hg(s) removal and (b) *Hg(0) evolution . !"#$ %&'!(#$)*+,-". / 0 !%1 2 3%1# !4#$ %&'!5#$6,(($ %131 5 7 891 !+:;# 89 0 1 !! ! !! <= !5>?#$ =3&'@ 1 5$ A$B>C$D'36$&' E F E <=!5>G#$=3&'@ 1 5$ A$/H>C$D'36$&' E ! E <=!/>I# =3&'@ 1 5$ A$/H>C$D'36$&' 5>C 5>G 5>J 5>K 5>B / 5 I5 ?5 G5 5>C 5>G 5>J 5>K 5>B / 5 I5 ?5 G5 &' L 0 "M( ! 7 G5 #$N$/O$

PAGE 130

130 Figure 5 5 . UV B Hg removal in pH 7 buffered solu tion s with molar equivalent sulfide. C 0 = 9.5 13.5 µg/L Hg(NO 3 ) 2 with variable S/Hg, 24 hr t DARK . ; (a) *Hg(s) removal and (b) *Hg(0) evolution . !" # $ % !"$ &'()* &+*, -./&0*,12'34+5 6 # &-7 8 9-7* &:*, -./&;*,<300, -797 ; !! ! !! => &;?@*, >9./A 7 ;, B,C?D,E/9<,./ F G F =>&;?H*,>9./A 7 ;, B,6I?D,E/9<,./ F ! F =>&6?J* >9./A 7 ;, B,6I?D,E/9<,./ ;?C ;?CJ ;?C@ ;?CH ;?CK 6 ; J; @; H; ;?D ;?H ;?L ;?K ;?C 6 ; J; @; H; ./ M # +N0 & % H; *,O,6P,

PAGE 131

131 Figure 5 6 . UV C Hg removal i n pH 7 bu ffered soluti ons with DOM and residual sulfid e . C 0 = 20.0 µg L 1 Hg(NO 3 ) 2 with variable S/Hg and mixing (t DARK ) ; (a) *Hg(s) removal and (b) *Hg(0) Loss . !"# !"#$ !"#% !"#& !"#' ( ! )! &! #! %!*+* !"#$ %*+* !"#$ %!*+* !%$ %*+* !%$ +,-. /01 2 34560 74 8**9-.7:8* 7;8** 9-. <0:: 5 =>? 7/6@8 !"A !"' !"# ( ! )! &! #! (! /.,<*+B-C (a) *Hg (s) removal

PAGE 132

132 CHAPTER 6 MIXED LIGAND EFFECTS ON MERCURY PHOTO TRANSFORMATIONS Scope In this work experiments on UV irradi ated solutions containing synthetic wastewater with aqueous Hg and its complexing ligands chloride, sulfide and DOM. The experimental evidence is combined with observations Hg speciation studies from are discussed with the aim of improving theoretical unde rstanding of photochemical transformation mechanisms that determine solid phase Hg formation. Water quality analysis from chlor alkali wastewater is the basis for synthetic wastewater composition; Hg solutions were prepared using chemical reagents specifie d in Chapter 3. The methods used to measure the concentration of Hg in solutions according to mass balance approach for Hg product analysis and characterization. In order to compare the effect of controlled system on Hg speciation and photo transformatio ns synthetic Hg wastewater solutions were prepared with and without pH buffers. Nylon filters we re used for Hg removal from solution using a filter pore size of 0.45 µm, thereby enabling comparison of previously reported results in this work. Experimental Parameters Two series of experiments were developed to compare the effects of Hg photo transform ations with sulfide as a precipitating agent in solutions with mixed ligands. The first series compared the eff ects of excess sulfide (~40 mg L 1 S) while the second examined trace sulfide (20 ng L 1 S). The first series was designed for the study of mixed ligands using non buffered Nano Pure DI water solutions prepared with excess sulfide, added chloride and SRNOM and had added Hg as either HgCl 2 or Hg(NO 3 ) 2 , with a target C 0 of 20 ! g L 1 Hg (0.1 µM). Solutions prepared for the study of chloride ion

PAGE 133

133 solely used HgCl 2 stock together with added NaCl while sulfide solutions without chloride were prepared using Hg(NO 3 ) 2 stock. Solution pH in this series varied by type of Hg stock used and whether highly alkaline sulfide was added. The compositions and measured pH values for or mixed ligand solutions are provided in Table 6 -1 and Table 6-2. The effect of UV-C on non -pH buffered solution show a moderate decline. Solutions with added sulfide and produced the most noticeable change from t 0 pH~10.5 to t 60 pH~9.5 wherein the shift in equilibrium increases H 2 S concentrations by a modest amount, approximately ~1.2%, of total sulfide. Trace-sulfide and organic matter . A second experimental series was developed using pH-buffered solutions with trace sulfide and IHSS DOM standard reference humic and fulvic acids. Detailed characterization information is provided in Appendix T able B-5. T he pH-buffered solutions were prepared at pH 3, 7 and 11 according to the method of adding acid or base to a solid salt using 0.1M HCl or NaOH solutions prepared with Nano pure DI water (Appendix Table A-3). The solution compositions are provided in Table 6$3 using measured initial Hg and DOM concentrations. An estimation of the pH-dependence of DOM surface charge was developed using IHSS characterization data on functional groups of the standard a nd reference samples (Appendix Figure B-4). Investigation of Possible Hg Species Mixed ligand omplexes. It is possible for mixed ligand Hg complexes to form at lower DOM:Hg ratios as aqueous Hg species with increasing order of equilibrium constants (Log K): HgOHCl (18.1) [46] < HgClSH (25.7) < HgOHSH 0 (30.3) < HgOHRS (32.2) [58] ; indeed, the hydroxy-sulfur mixed ligands, HgOHSH 0 (aq) and HgOHSR ( aq) have stability constants comparable to Hg polysulfide complexes.

PAGE 134

134 Hg-DOM. Reactive functional groups of DOML (i.e. L = RX) are thiol (RS$), oxygen (RO), and nitrogen (RN). Thiol is the primary binding site in interactions of free ionic Hg(II) and unsaturated Hg:DOM conditions and has reported solubility product Log Ksp of 10, for the H+ abstraction reaction DOMRSH ! DOMRS – . Findings from speciation models predict that Hg-DOM binding configurations occur as either a (1:1) monodendate ratio of Hg to RS$, or a (1:2) bidentate ratio for complexes with either two RS$, or one RS$ and one other RX group in DOM [78, 208] described by a range of reported stability constants in Table 6$5. According to the Hg:DOM composition of the experimental solutions, it is understood that monodendate binding with RS– likely occurs first and that additional Hg-DOM would form as mixed ligand complexes with abundant O-containing functional groups, such as carboxylic or phenolic moieties. It is also possible to form Hg-DOM chelates through a composition of several functional groups, which Haitzer et al postulate as sterically controlled and requires proper orientation of the DOM molecule and its functional groups for strong binding. Furthermore, Haitzer et al. hypothesized that reduced sulfur groups are the dominant Hg binding sites at low Hg/DOM concentration ratios and that Hg is mostly bonded to carboxyl groups at high ratios of Hg to DOM; the authors reported that strong-binding DOM is approximately equal to the molar equivalent of 1 g Hg per mg1-DOC (i.e. 5 *109 mol -Hg per mg1 DOC) [71] . UV Experimental Results Excess Sulfide in DI Water Solutions The experimental results of filterable *Hg(s) removal and the comparative Hg loss and Eh potentials were obtained for the mixed ilgand solutions. A comparison *Hg(s) removal is shown in Figure 6-1 for three different solution compositions: (a) Hg-

PAGE 135

135 Cl DOM (b ) Hg S DOM and (c) Hg S Cl DOM and the results for DOM only solutions are shown for comparison. In contrast with previous results obtained from synthetic Hg waters using only Hg(NO 3 ) 2 and sulfide, the recovery of Hg in solutions following irradiation periods was relatively high, most often 95% or higher, whereas solutions that had greater Hg losses are inclu ded in the summary of results ( Table 6 6 ). The results show that UV C promoted format ion of filterable *Hg (s) in mixed ligand solutions with S Cl, whereas DOM only solutions also had enhanced filtered Hg removal. The addition of chloride ion and DOM with excess sulfide resulted in faster photo catalyzed formation of filterable solid phase complexes, whereas the results in Figure 6 1 (c) show that an increase in Cl Ð /Hg mol ratios from 8 and 66 results in measurably greater *Hg (s) removal. DOM with Trace Sulfide in pH Buffered Solutions A series of experiments sought to compare the effect of varying pH and DOM isolate in solution with trace sulfide (0.65 nM S); a tabulated summary of the findings is given in Table 6 7 followed by graphic descriptions in Figure 6 2 to Figure 6 4 . Acid pH . The results for UV C Hg removal in pH 3 solution with trace sulfide (0.65 nM S) in Figure 6 2 are mostly comparable across the different DOM types used, with the exception of SRHA, w hich had less initial removal and moderately higher Hg removal rate. A comparison of DOM surface charge density appears to correlate well with Hg filtration before irradiation by order of for increasing surface charge densities SRHA < NLHA < SRNOM < NLFA . However, over the entire 60 min irradiation period, solutions with SRHA had the greatest change in filterable Hg, with a total increase of 30%, compared with <20% for the other DOM. The comparative H g loss and Eh potential values are also provided ( Figure 6 2 ).

PAGE 136

136 Neutral and alkaline pH. The results Figure 6 4 for Hg removal from pH 11 solutions show greater overall filterable Hg relative to pH 3 and 7 in Figure 6 4 , which cor responds with higher overall Hg Losses and lower measured Eh potential. By comparison, of results by the DOM isolate variable, a notable difference in UVAC Hg removal occurred for solutions that contained thiol rich SRHA, which is predicted to form strong er Hg DOM RS complexes due to the relatively greater availability of RS Ð binding groups in SRHA. Results from the pH 7 and 11 SHRA solutions had the lowest Hg removal following 30 min UV C whereas after 60 min UV C, Hg removal was equal to or greater than the other solutions. In comparing results from the trend observed pH 7 and pH 11 SHRA solutions, the higher apparent rate of Hg phase transformation occurring from 30 to 60 min UV C was most pronounced for pH 11 solutions, which also corresponds to greater Hg DOM binding reactivity through thiol groups (RS Ð pK a ~9 11). Investigation of Photo R eactive Mercury Species A useful measure of photo reactivity is a molecules optical transition energy, which relates to Hg ligand bonding strength. In solutions with neutral to acidic pH, the predicted species HgS and HgS ' H 2 O have the lowest S 0 to S 1 optical excitation in the visible or near UV [60] . Therefore, it is possible in the system without DOM for HgS ' H 2 O first dissociat es its water molecule and then forms ionic Hg and free sulfide. Furthermore, optical studies for Hg S species show lower transition energies for single bonded Hg than for diatomic chelated Hg [112] , which woul d affect overall Hg photo transformation in terms of reaction rates and Hg products. When at least one sulfhydryl group is present Hg DOM complexes are ! 30 orders of magnitude more stable than complexes involving only oxygen or nitrogen

PAGE 137

137 ligands, whereby an average of 2.4 ±0.2 sulfur atoms from DOM thiol functional groups bond with Hg in solutions void of inorganic sulfide [209] . The molecule HgOHSH 0 forms in neutral to alkaline solutions and, due to its small hydration energy, it can partition into DOM. On the one hand, HgOHSH 0 complexed with DOM would more stable and on the other, it is expected that alkaline pH promotes DOM photo oxidation. This can be explained by a more dispersive DOM structure from libera ted phenolic ligands and favorable steric positioning within Hg DOM complexes. A comparison on the effect of DOM in dark oxidation experiments suggests that alkaline pH results in more Hg(0) oxidized by DOM than acid pH solutions. Oxidizing conditions ar e known to convert thiols to disulfides, reported by Tarbell et al. in which where molecular oxygen and Fe(III) can serve as electron acceptors in the oxidation of thiol to disulfide [210] . This type of sulfide pho to transformation was recently cited in the recent findings by authors Ariya and Si who describe anoxic solution with acidic pH in which photo chemical formation of particulate HgS (s) occurs from the addition of ionic Hg(II) and thioglycolic acid [134] . The findings corroborate the earlier work by Frimmel et al. who first reported photochemical formation of HgS (s) by UV C irradiated solutions with 1 !mol Hg(II) and model thiol cysteine ligands (CYS C 3 H 7 NO 2 S) in de oxyge nated solutions over pH 3 6 whereby the precipitates were removed from s olution via membrane filtration. T he proposed reaction would have occurred by the dissociation of sulfide from the thiol group, (Hg S R)OH + hv ! HgS +R' [211] . Strong chelate complexes are known to form with amino acids sulfhydryl cysteine and glutathione (GSH C 10 H 17 N 3 O 6 S) in acid by forming complexes with high Hg S coordination numbers [212] . For Hg GSH complexes, neutral to alkaline pH

PAGE 138

138 changes the structure to linear S Ð Hg Ð S coordination thereby creating stronger S Hg bonds with deprotonated [GS 3 ] ligands [213] , which may help explain the greater UVAC removal for Hg DOM solutions prepare d at alkaline pH co mpared with neutral solutions. While the CYS content of the IHSS DOM isolates has not yet been characterized, GSH GSH/Hg mol ra tios are provided in Table 6 3 and can be calcul ated for HgLH, HgLH 2 , HgL 2 H and HgL 2 H 2 using average a cid dissociation constants pKa 1 2.06, pKa 2 3.5, pKa 3 8.6, and pKa 4 9.6 [213] . Total thiol content for SRNOM was recently determined by Rao et al. to be surprisin gly low (0.7 !mol g 1 DOC) [214] , and provide s a possible explanation for relatively higher observed *Hg(0) losses at pH 11 in Figure 6 4 . In the experimental solutions with excess sulfide Hg S likely formed as a stable anion, such as Hg(SH) 2 OH $ that has one of the largest hydration energies and is thus confined to aqueous solution, which is also expected for similarly hydrated species, HgHS 2 $ and HgS n OH Ð . The observations agree with those o f Xiao et al. who found that precipitated HgS occurred before its photo reduction in prepared HgS 2 2 $ solutions continuously irradiated with UV 290nm over the 6 hr test [215] . While the kinetics occurring in Xiao's experiments are unique because of differences in reactor setup, UV wavelength and irradiation periods, it is possible that comparable Hg photo transformations would occur in the system studied resulting in formation of both precipitated HgS (s) and dissolve d Hg 0 (aq) . Charged polysulfides Hg S are hydrophilic and previous findings indicate they would not bind directly with DOM [48, 139] . T herefore , in solutions with excess sulfide, it is like ly not possible for HgS (aq) photo reduction reactions by DOM to occur by LMCT reactions. This may explain why higher *Hg losses occurred for DOM solutions with

PAGE 139

139 trace sulfide ( Figure 6 4 ) compared with residual sulfide ( Table 6 6 ). However, it is important to note that for either sulfide concentration neutral Hg sulfide complexes could form complexes with DOM due to the favorable Hg molecular configuration into a linear, coordinated species [55, 60] . Despite the possibility of strong Hg complexation by sulfide and thiols, both types of reduced sulfur containing ligands are unstable under oxidizing conditions, which have been shown to promote the formation of weaker Hg DOM RX complexes occurring through other functional groups [216] . Previous research under oxidizing conditions identified organic matter acts as important electron acceptors in the proces s of thiol induced oxidation of Hg(0) aq [135, 217] and further quantified experimental findings of Gu et al. in which HA ligand induced oxidative complexation was estimated to have binding capacity of ~3.5 µmol Hg/g HA [89] . Summary The postulated mechanisms of DOM oxidation reactions have dual significance in Hg photo transformations leading to filterable Hg(s): reducible moieties act to both scavenge electrons and mediate Hg photo reduction, while photo oxidation of sulfhydryl groups is a possible source of sulfide ligand for Hg S precipitate formation. The formation of available sulfide to react with Hg may be explained by the findings of Jocelyn in which excess sulfide formed disulfides resulting in increased oxidizing action on disulfide by a thiol group [218] . This provides supporting evidence in the identification of possible UVAC products as Hg(II) and Hg(0) molecules forming filterable complexes with stable precipitates of HgS (s) .

PAGE 140

140 Table 6 1 . Hg(NO 3 ) 2 and mixed ligand solution composition with measured pH over 60 minutes UV C contact time. pH of synthetic Hg wastewater with added constituents Hg(NO 3 ) 2 only S DOM S + DOM t UV C (min) 20 µg /L Hg 40 mg /L S ~ 10 mg /L DOC 40 mg /L S ~ 10 mg /L DOM 0 3.96 10.42 3.91 10.9 5 3.95 10.71 3.72 10.76 30 3.85 10.56 3.66 10.32 60 3.7 9.9 3.01 9.36 Table 6 2 . HgCl 2 and mixed ligand solution compos ition and pH values measure with varying UV C contact time. HgCl 2 only Cl Cl + S Cl + DOM Cl + S + DOM 12 µg /L Cl $ t UV C (min) 15 µg /L Hg 12 µg /L Cl $ 12 µg /L Cl $ 12 µg /L Cl $ 40 mg /L S (0.147 µM Cl $ ) 40 mg /L S 10 mg /L DOC 10 mg /L DOC 0 5.52 5.5 2 10.56 4.96 10.59 5 5.71 5.71 10.72 4.71 10.62 30 5.01 5.01 10.48 4.54 10.48 60 5.31 5.31 9.87 4.46 9.68

PAGE 141

141 Table 6-3. Molar compositions of prepared solutions of trace-sulfide and DOM, labeled as S1 to S4. pH/ ID S/Hg mol -ratio Hg (nM) S (nM) Cl– (mM) DOC mg/L (mM) DOM-RS (nM) DOM absent (S0) 3.S0 0.02 128.5 2.3 23 7.S0 0.01 119.3 1.3 11.S0 0.01 132.3 1.3 NLHA (S1) as GSH(1) 3.S1 lo 0.01 108.7 1.3 23 3 (0.23) 46.9 3.S1 hi 0.02 100.3 2.3 23 5 (0.44) 90.7 7.S1 0.02 89.9 1.6 5 (0.44) 90.7 11.S1 0.02 89.3 1.6 5 (0.44) 90.7 NLFA (S2) as GSH(1) 3.S2 0.02 92.2 1.6 23 12 (0.98) 72.0 7.S2 0.02 86.1 1.6 12 (0.98) 72.0 11.S2 0.02 84.1 1.6 12 (0.98) 72.0 SRHA (S3) as GSH(1) 3.S3 0.01 159.0 1.6 23 5 (0.44) 64.1 7.S3 0.02 82.9 1.6 5 (0.44) 64.1 11.S3 0.02 67.9 1.6 5 (0.44) 64.1 SRNOM (S4) as R-SH(2) 3.S4 0.02 99.9 1.6 23 10 (0.85) 7 7.S4 0.02 80.9 1.6 10 (0.85) 7 11.S4 0.02 86.2 1.6 10 (0.85) 7 (1) GSH calc ulated from IHSS characteri zation data in Appendix C (Table B 6) and (2) Total thiol c ontent ( . R SH) calculated from observations by Rao et al., which noted an approximate value of 0.7 !mol/ R SH per g DOM [214] .

PAGE 142

142 Table 6-4. Calculated DOM charge density (meq/g-C) for pH-buffered solutions. pH NLHA (S1) NLFA (S2) SRHA (S3) SRNOM (S4) 3 3.38 4.70 2.73 3.74 7 8.38 10.53 8.46 9.27 11 11.55 13.39 12.69 12.86 Table 6-5. Formation stability constants for Hg-DOM complexes in solutions with pH ~8. Configuration of Hg-DOM(1) Log K range HgL, L=RS$ 21 – 24 HgL2 , L=RS$ 28 – 32 HgL2 , L=RS$, RX 28 – 47 (1) Data source from generally accept ed values adapted from Dong et a l. [78, 208]Table 6-6. *Hg losses from UV-C in prepared Hg solutions as total Hg loss, *C/C0. C0 =15 to 20 !g/L Hg; insignificant Hg losses (%) occurred for the remaining solutions. ID S C CS_lo CS_hi C+DOM UV-C (min) 8 Cl–/Hg 8 Cl–/Hg 66 Cl–/Hg 8 Cl–/Hg 5 96. 0% 91.1% 100.0% 100.4 100.7% 30 93.0% 87.4% 98.8% 99.3 93.8% 60 91.4% 82.5% 80.3% 85.1 88.6%

PAGE 143

143 Table 6 7 . UVAC Hg removal results for p H buffered solutions with Hg DOM and trace sulfi de (S1 to S4 an d baseline S .0 ) as Hg concentration (!g/L Hg) . Solution ID No DOM S1 (NLHA) S2 (NLFA) S3 (SRHA) S4 (SRNOM) pH 3 ( ! g/L Hg ) C 0 25.778 32.501 29.897 31.892 32.377 C F (0 min) 11.281 5.645 1.222 19.936 3.550 *C F 5 min 10.234 5.710 3.900 19.704 9. 561 30 min 4.755 2.096 1.805 10.184 5.452 60 min 5.237 0.949 0.392 5.607 1.656 pH 7 ( ! g/L Hg ) C 0 23.937 29.128 27.918 22.875 26.217 C F (0 min) 19.135 20.914 17.567 20.271 17.035 *C F 5 min 18.675 17.779 10.747 15.012 14.175 30 min 19.3 63 4.734 2.150 2.999 2.538 60 min 18.020 2.107 0.500 0.000 0.566 pH 11 ( ! g/L Hg ) C 0 26.548 28.952 27.254 22.002 27.937 C F (0 min) 16.288 19.652 11.151 18.629 16.998 *C F 5 min 22.366 7.884 5.882 13.350 6.865 30 min 22.705 0.218 0.031 0.1 85 0.038 60 min 20.899 0.043 0.068 0.205 0.034 Hg concentatration for initial untreated solutions ; C 0 . 0.45 ! m filtrate concentration after 2 hrs dark mixing (0 min UV C) as C F ; and variable UV C contact time as treated filtrate concentration, *C F .

PAGE 144

144 Figure 6 1 . UV C Hg removal for DI mixed ligand solutions of S, Cl and DOM. (a) Hg Cl DOM (b) Hg S DOM and (c) Hg S Cl DOM. C 0 ~ 20 !g/L Hg, 40 mg/L S and ~ 10 mg/L DOC and given Cl Ð /Hg mol ratio (2 h r t DARK mixing). ! !"# !"$ !"% !"& ' &()* ! +,-./01 &()* ! +,/01 ! !"# !"$ !"% !"& ' 2./01 /01 2 ! !"# !"$ !"% !"& ' ! '3 4! $3 %! &()* ! +,-.2./01 %%()* ! +,-.2 &()* ! +,-.2 /01 5678( 9:;<= >,-9?= 9@= ,A )* A /01 9B=(,A 2 A /01 98= ,A 2 A )* A /01 y axis: *Hg (s) removal (1 $ *C F / *C)

PAGE 145

145 Figure 6 2 . UV C Hg removal for pH 3 solutions wi th trace sulfide and DOM (S1 S4). C 0 =20 µg/ L Hg , 20 ng/L S and ~ 10 mg/L DOC (2 hr t DARK mixing). !"#$ !"#% !"&! !"&' !"&# ! (& )! #& $! *(+,*(+./ *' *) *# ! !"' !"# !"$ !"% ( *(+,*(+./ *' *) *# !"0 !"% !"1 ( *(+,*(+./ *' *) *# 2345 67/89 65 9::;.:6 ! 9 6<9 =>? @-AA 6B 9 =>? 6A9: *(: C D@>E *' C D@FE *): C *G>E *#: C *GDHI J>:) removal

PAGE 146

146 Figure 6 3 . UV C Hg removal for pH 7 solutions with trace sulfide and DOM ( S1 S4). C 0 =20 µg/ L Hg , 20 ng/L S and ~ 10 mg/L DOC (2 hr t DARK mix ing). !"# !"$ !"% & '& '( ') '* ! !"( !"* !"+ !"$ & '& '( ') '* ,-./ 01234 !")$ !"*( !"*+ !"5! !"5* ! &5 )! *5 +! '& '( ') '* 0/ 4667860 ! 4 094 :;< =>?? 0@ 4 :;< 0?46 '&6 A B=;C '( A B=DC ')6 A 'E;C '*6 A 'EBFG H;6# removal

PAGE 147

147 Figure 6 4 . UV C Hg removal for pH 11 solutions with trace sulfide and DOM (S1 S4). C 0 =20 µg/ L Hg , 20 ng/L S and ~10 mg/L DOC (2 hr t DARK mixing). !"# !"$ !"% & '& '( ') '* ! !"( !"* !"+ !"$ & '& '( ') '* ,-./ 01234 !"(+ !")! !")* !")$ !"*( ! &5 )! *5 +! '& '( ') '* 0/ 4667860 ! 4 094 :;< =>?? 0@ 4 :;< 0?46 '&6 A B=;C '( A B=DC ')6 A 'E;C '*6 A 'EBFG H; && removal

PAGE 148

148 CHAPTER 7 SUMMARY OF FINDINGS Th e experimental findings i dentified UVAC treatment as a method of removal for solvated Hg in industrial secondary process wastewaters by photo mediated electron transfer that is governed by the chemical composition of the solution and the intensity of UV C irradiation . Therein, th e content of competing inorganic and organic matter appeared to be predominant factor in determining whether ppt level Hg concentration are possible in treated effluent filtrate . It is evident that while multiple simultaneous and competing Hg photo reacti ons determine the formation of filterable Hg products, photo degradative pathways that target organic binding ligands alter Hg speciat ion from its soluble form . Th e disc ussion of experimental results postulates photo mediated electron tr ansfer reaction me chan isms to explain observations of photo transformed dissolved Hg. The findings build on the sc ientific knowledge base needed to identify photo reactions and degradative pathways that form filterable Hg products in organic rich waters in which convention al treatment and filtration alone cannot achieve ppt level Hg effluent. The e xperimental UVAC Hg removal results show that in the absence of UV irradiation, the filterability of Hg was enhanced by the addition of molar equivalent concentrations of su lfide and impaired by the addition of HA alone. Solutions prepared with both sulfide and HA and irradiated with UV C for periods between 5 and 60 minutes, the simultaneous photoreactions of dissolved sulfide and DOM oxidation formed > 0.45 !m filterable Hg compl exes , and achieve effluent conc entrations below the instrumental detection limit ( BDL < 6.0 ng L 1 Hg ).

PAGE 149

149 The investigation studied the effects of UV C in both real secondary process industrial Hg wastewater (~5 to 50 µg L 1 Hg ) an d synthetic Hg solutions (~ 20 µg L 1 Hg ) . The formation o f non volatile Hg aggregate was favorable for removal to l evels below detection via 0.45 µ m filtration when lower molecular weight or ionized organic agents were produced by high intensity UV C irradiation . This occurred for contacts periods between 30 and 60 minutes by which d issolved Hg was photochemically transformed into insoluble Hg in pH # 7 solutions prepared with DOM. The observations are consistent with coupling of photo reduction and oxidation reactions of Hg and i ts binding ligands in o rganic rich Hg water systems. On the basis of experimental results from prepared synthetic Hg wastewater solutions, UVAC Hg removal process are postulated to have occurred when ionic Hg was freed from stable aqueous phases of Hg Cl, Hg S, or Hg DOM RS molecules and subsequently formed non volatile Hg aggregate with inorganic precipitation agents. In the case of sulfide precipitation, the photo oxidizing action by UV may have been a limiting reaction . It can be further postulated that the triplet state of photo activated DOM enhanced UVAC treatment by acting as an effective electron scavenger that catalyzed the oxidation of electron donors, and thereby enhanced the formation Hg solid phase formation as oxidized Hg instead of Hg(0). The results from trace sulfide solutions (S1 S4) prepared with different DOM isolates of either N LHA, NLFA, SRHA or SRNOM ( Table 6 7 ) , show there was increased removal following 30 min contact and the rate of removal was higher in al kaline pH 11 relative to pH 3 and pH 7 buffered solutions. The lowest average ppt level Hg effluent concentrations were obtained following 60 minutes o f UV C contact

PAGE 150

150 from pH 11 solutions prepared with ~20 !g L 1 Hg(NO 3 ) 2 and 10 mg L 1 DOM . Furthermore, S3 solutions with thiol rich SHRA, with pH 7 buff er and U V C irradiated for 60 minutes produced treated effluent with BDL Hg levels. It is possible that Hg DOM RS complexation occurred from thiol found in different DOM isolates due to their importance as bindi ng sites for Hg DOM sorption. T he photo degradation of thiol within soluble Hg DOM RS complexes appears conducive to optimal Hg removal that achieves low level ppt H g effluent following filtration; h owever, the experimental observation that thiol rich SRHA solutions required longer UV C irradiation periods is an important consider ation for further UVAC research. The results support the postulation that once thiol is oxidiz ed, Hg is liberated from its binding ligand and favorably occurs as divalent Hg in th e presence of reducible moieties that act as electron scavengers. However, since trac e sulfide Hg DOM solutions with fewer thiol groups (S1, S2 and S4) also produced low level ppt Hg (< 40 ng L 1 Hg), a key consideration is that carboxyl groups in each DO M ty pe may act as the principa lly available complexing ligand s; furthermore, since phenol groups are known to ionize at alkaline pH , they are predicted to be important oxidizing agents that may determine solid phase Hg distribution in UV C wastewater trea tment. Future work . While the experimental findings are unreported elsewh ere in the scientific community, qualitative and characterization studies of UVAC transformation processes are needed in order to further develop its app lication for wastewater with v arying chemistry. In doing so, its use is predicted to be useful in wastewater with halide ligands other than chloride that produce brominated Hg species, such as in the coal and petrochemical industry.

PAGE 151

151 Further UVAC research could focus on photo degradatio n of thiol DOM g roups and possibly emphasize its role along with other functional groups in Hg re oxidation and formation of filterable solid phase Hg. Based on the experimental observations, f u ture work on the role of photo active DOM moieties could consi der whether the formation rate of insoluble Hg increase s when photo reducible DOM phenol groups act as electron acceptors. Therein, it is important to consider that the ionized phenol group may either act to form Hg complexes or more readily photo degrade into LMW and thereby provide an electron source for Hg reduction. The observations of ORP measured in UVAC treated waters suggest that both reducible and oxidizing couples facilitate Hg removal and occurred in bo th chlor alkali wastewater and prepared syn thetic Hg DOM solutions . In order to better quantify its meaning , f urther investigation could consider the role s of both HA and FA functional moieties in photo activated ligand to ligand and ligand to Hg metal interactions , and their respective effects o n the reduction oxidation potential of Hg wastewater.

PAGE 152

152 EXPERIMENTAL PARAMETERS UV / UVAC Batch Reactor Parameters Figure 7 1 . Schematic of UV bulbs with d imensions (mm). A = 128.8, B = 144.5, C = 167.0, D = 28.0 and D1=13.0. Figure 7 2 . UV C photometric output of Philips PL S 9W/12 G23 lamps. Figure 7 3 . UV B photometric output of Philips PL L 9W/01/2P l amps. D D I

PAGE 153

153 Photometric Output Calculations Photon flux is calculated for horizontal cross sectional area (A x ) illuminated per unit volume (1 cm 3 ) using in Figure 7 4 . Figure 7 4 . B atch reactor dimensions used to calculate photon flux. An Einstein is a unit defined as the energy in one mole ( N hv = 6.022* 10 23 ) of photons. The term E J/e is used to convert photon energy flux from Einsteins (j hv ) to Joules (E hv ) calculated by Eq ( 7 1 ) . ( 7 1 ) A x = ! ( r 1 2 " r 2 2 ) E h v ( J ) = j h v ! E ( J / e ) E ( J / e ) Jo u l e s e i n st e i n ! " # $ % & = N h v ' h ' c (

PAGE 154

154 Table 7 1 . Radial photometric energy per unit volume in UV/UVAC batch reactors. H orizontal cross sectional area A x = 12. 10 cm 2 . ( (output) Intensity Photon flux, j hv Energy flux, E hv nm (W) mW/cm 2 Es/(cm 2 min) * 10 3 Es/(cm 2 s) J/(cm 2 min) * 10 9 J/(cm 2 s) * 10 7 E j/e * 10 5 UV C a 254 (2.4) 198.3 20.3 339 9.6 16.0 4.71 UV B c 310 (1.2) 99.1 8.26 138 3.2 5.3 3.86 UV A b 365 (1.6) 132. 2 10.3 172 3.4 5.6 3.28 Ferrioxalate actinometry results in findings by Byrne produced a. j hv = 2.46 * 10 5 Es/min for UV C and b. j hv = 1.26 * 10 5 Es/min for UV A from Bryne et al. [12] c. Estimated as j hv = 1.0 * 10 5 Es/min. Mercury Removal Calculations The distribution of soluble aqueous and fi lterable solid phase Hg was calculated using [Hg (T) ] = [Hg (aq) ] + [Hg (s) ] based on measured [Hg T ] and [Hg (aq) ] in solutions before and after UV treatment, where (*) i ndicates values after UV treatment. T he relevant Hg fractions are defined according to Eq ( 7 2 ) through ( 7 7 ). untreated solution total, [HgT] = C 0 ( 7 2 ) untreated filtrate soluble aqueous, [Hg(aq)] = C F ( 7 3 ) particulate solid phase, [Hg(s)] = C 0 $ C F ( 7 4 ) UVAC treated solution total, *[HgT] = *C ( 7 5 ) UVAC filtrate soluble aqueous, [*Hg(aq)] = *C F ( 7 6 ) UVAC particulate solid phase, [*Hg(s)] = *C $ *C F ( 7 7 ) UVAC reactor adsorbed, [*Hg] = *C R ads ( 7 8 ) [*HgLoss] = [*Hg(0)(g)]+ [*Hg(R a ds)] *C Loss = *C g +*C R ads ( 7 9 )

PAGE 155

155 Approximation of elemental mercury losses. A mass balance approach is developed to account for Hg loss from reactor residual * Hg R ads and formation of gaseous *Hg(0) (g) that volatilized from solution according to Eq ( 7 10 ) or by *Hg(0) (g) alone according to Eq ( 7 11 ) when measured less than approximately ±1% Hg T (untreated total Hg), which is lower than the allowable experim ental error of 3%. The summation of both residual reactor and volatilized Hg are represented in Eq ( 7 12 ) as Hg Loss for experiments in which residual reactor Hg was not measured. ( 7 10 ) ( 7 11 ) ( 7 12 ) Residual reactor adsorbed Hg . In order account for Hg adsorbed to the batch reactor, UV bulb or stir bar during UV irradiatio n (hereon reactor components), which possibly occurs by the forces electrostatic attraction, this fraction of Hg is defined as residual reactor Hg (Hg R ads , or Hg R ). This fraction was measured by de sorbing the Hg into a nitric acid soak of the reactor co mponents after the batch reactor solutions were emptied. The procedure calls for adding 100 mL acid solution as 25% equivalent of HNO 3 or HCl (Fisher Scientific trace metal) to the emptied reactors immediately after each UV experiment, which is then conti nuously stirred (using the same stir bar) at 50¡C for approximately 30 min to 1 hr, or at roo m temperature for 6 to 12 hours, depending on the initial Hg concentration. This process releases any that is Hg bonded to the reactor walls into the acid solutio n, from which a 20 mL sample was collected and the Hg concentration was measured using AF analysis. Note that this fraction was ! H g ( g ) 0 = H g T " ( ! H g T + ! H g R ) ! H g ( g ) 0 = H g T " ! H g T ! ! H g R < 1 % o f H g T ! H g L o ss = H g T " ! H g T # ! H g ( g ) 0 + ! H g R ! ! H g R n o t m e a su r e d

PAGE 156

156 typically low, account for less than 1% Hg T , but that larger concentrations were also observed, up to 8% Hg T , depending on the s olution composition. Hg(0) losses. The occurrence of Hg(0) (g) evolution from UV irradiated solutions is approximated by a mass balance calculation using Eq ( 7 10 ) or ( 7 11 ) from the discussion above. A c ontrol experiment was conducted in order to establish an understanding of the baseline distribution of UV Hg fractions in UV irradiated synthetic Hg wastewater over a series of UV C contact times. The results in Table B 4 report UV Hg products in terms of *Hg(0) (g), *Hg (s) and *Hg (R ads) for Nano Pure DI and Hg(NO 3 ) 2 buffered acid, neutral and alkaline pH using reagents in Table A 3 . Batch Reactor UV Heat Transfer Figure 7 5 . Em pirical derivation of UVAC batch reactor heat transfer from UV C bulbs. (a) UV contact time greater than 2 minutes and initial solution temperature T 0 ~ 23 ¡C and (b) Modeled vs measured s olution temperature . !"#"$%&'()*"+"$%,-&' ./"#" $%000( $ 1 2$ 21 &$ &1 $ &$ ,$ -$ '$ 3456 7383 $ 9 , : ;< = >" 75?@9 7383 ! 9A B?@4CD 7C9 !" !# $" $# %" %# #" " !" %" &" '" ()*+ , . /0 1 2*345 (676( ! 2"8!'#6 9 .:;6<6"8#5=> (6?@.A?BC6 D 2E5

PAGE 157

15 7 ORP Probe Temperature Correction Redox overvie w. In water systems, reducing or oxidizing conditions are limited by the extent to which water can dissociate into H 2(g) or O 2 (g). The limit of reduction is given by: H + + e ! " H 2(g) and the oxidation limit is given by: H + + e Ð + # O 2(g) ! " H 2 O. Electroche mical potential, Eh, is a theoretical representation of potential electron activity, E ¡ = E¡ ox + E¡ red , which uses the Nernst equation below in (7 13) : ( 7 13 ) E ¡ is measured as the voltametric differenc e between an inert platinum electrode and the standard reference hydrogen electrode occurring from E ¡(V) = 2.0 to 2.0, n is the number of moles of electrons transferred in the balanced equation, F is Faraday's constant representing the charge on a mole o f electrons, and R is the ideal gas constant. The Nernst equation is temperature dependent and so electrode measurements require temperature correction. The term p e is used in mole liter 1 units and can be compared to pH [201] . Conversion from Eh to p e is shown in ( 7 14 ) . ( 7 14 ) where: , Eh ( V ) = E o + 2 . 3 R T n F ! l o g O x { } R e d { } p e = F 2 . 3 0 3 R T ! Eh [ =] m o l L p e ( M ) = Eh 0 . 5 9 1 6 a t 2 5 o C F = 9 6 . 4 2 ! 1 0 3 [ =] J m o l R = 8 . 3 1 4 [ =] J m o l K

PAGE 158

158 Figure 7 6 . Linear regression used for temperature corre ction for ORP electrode. (a) the relevant range 20 to 50 ¡C, which was developed from the manufactures reference data for (b) 0 to 50 ¡C using 4M KCl solution with SHE =220 mV at 25¡C. !"#" $ %&'()*+","--.&*/ 01"#" %&''/*' 2)% 2/% 2'% -%% -2% -% .% (% 3% 45" 6 $ 7800 9 :;< =>?@ 6A:B"9 C 7<" 9D<"6"#"-%" E>" 3% F 7 !"#" $ %&//*/+","--2&(( 01"#"%&'')(2 2)% 2/% 2'% -%% -2% --% -.% % 2% -% .% (% 3% 45" 6 $ 7800 9 :;< =>?@ 6A:B" 9 C 7<" 9G<"6"#"%"E>"3% F 7 80H"6A:B" 7>IIAJEK>@L"(M" N7? 45"9:;<"#"4 8O= ,"9 45 04P ," 45 6 $ 7800 <"

PAGE 159

159 APPENDIX A SUPPORTING DATA UVAC T reated Chlor alkali Wastewater The effect of pH a djustment acid is shown in the results in Figure A 1 , which compares Hg removal and Hg losses with HCl, H 2 SO 4 ,and HNO 3 over a 90 minute UVAC contact period. Figure A 1 . UVAC Hg re moval with HCl acid pH adjustment in chlor alkali sample C. Sample was 12 month aged with C 0 ~20 µg/L Hg, adjusted to pH 5 from pH in 11, (a) Hg removal by filtration and (b) Hg losses. !"# !"#$ !"#% !"#& !"#' ( ! () *! %) &! +) #! ( , -. / 01 2. /3 4/5 4 ! 67 " 487 # -931.4: -;3 !"+ !"+) !"' !"') !"# !"#) ( ! () *! %) &! +) #! -./2/ ! 3 < => , /1 -?@A3 4/5 4 ! 67 " 487 # -B31 .4:1 CD;;

PAGE 160

160 Hg R emova l by Heat and F iltration Figure A 2 . Heat only Test (1) Hg removal in chlor alkali sample C by filtrati on alone. Before UV: 1 hr dark mixing, adjusted to pH 3 9 then heated; C 0 ~ 20 µg/L Hg. Figure A 3 . Heat only Test (2) Hg removal in chlor alkali sample C by filtration alone . Prior to UV: dark 15 min mixin g, heated and adjusted to pH 2 12; C 0 ~ 20 µg/L Hg. !"!! !"!# !"$! !"$# !"%! & # ' ( ) *+* ! ,*./* ! 0/%!/ 12+3 $/4567/89:9;2/ ?@ ,-/?AB/ ).8.+,.
PAGE 161

161 A single factor ANOVA (! 0.05) was run on the experimental results to determine if there were difference between the effect of heat on UV-C Hg removal using heat as the variable factor and pH adjustment as the independent variable between the two groups. Two tests were conducted by (1) solutions were heated to 50C then pHadjusted and (2) solutions were pH adjusted then heated to 50C. Two independent tests used different mixing periods following pH adjustment as specified. The summary is: Test (1) F(1,10) = 1.397, p = 0.969; and Test (2) ANOVA F(1,10) = 1.397, p = 0.743. Table A-1. ANOVA results for heated UVAC for chlor-alkali sample A. Test (1) with groups A) no heat and B) pH adj + 50C (1 hr mixing); Test (2) with groups A) no heat and B) heated to 50C then pH adj (15 min mixing). Source of Variation SS df MS F Pvalue F crit Test 1: Between groups 7.65E 06 1 7.65E 06 1.61E 03 0.969 4.965 Wi thin groups 0.0476 10 4.76E 03 Total 0.0476 11 Test 2: Between groups 4.16E 05 1 4.16E 05 1.15E 01 0.743 5.318 Within groups 0.0029 8 3.60E 04 Total 0.0029 9 Hg Removal by Heated UVAC Treatment ample C wastewater with pH in 11 was tested for the effect of temperature at the time of filtration in UVAC treatment with 60 and 120 minute contact time. The results are shown in Figure A-4 and Figure A-5 by using a) UV generated heat only (40C and 60C); b) heated with hot plate plus UV generated heat (70C and 80C); and c) continuously cooled to 4C in a circulating water bath.

PAGE 162

162 Figure A 4 . UVAC Hg removal with variable soluti on temperature and 60 min UV C . Fresh chlor alkali sample C adjusted to pH 11 and C 0 ~ 20 µg/L Hg. Figure A 5 . UVAC Hg removal with variable solution temperature and 120 min UV C. Fresh chlor alkali sampl e C adjusted to pH 11 and C 0 ~ 20 µg/L Hg. !"! !"# !"$ !"% !"& '"! ()*+,-./01*23456347 ()*8.99*23464 ! 7 4:*4 ! ;*#!*<)68*() %!*-=>*? @A4 B! C 4 $! C 4 $ C 4 !"! !"# !"$ !"% !"& '"! ()*+,-./01*23456347 ()*8.99*23464 ! 7 4:*4 ! ;*#!*<)68* () '#!*-=>*? @A4 &! B 4 %! B 4 $ B 4

PAGE 163

163 Effect of UV-C and pH on Oxidation-Reduction Potential Figure A-6. Eh-UV-C for alkaline DI solution vs. chlor-alkali sample A for UV-C " 20 min. Figure A-7. Unfiltered vs. filtered solution UVAC-ORP readings in chlor-alkali sample A. UVAC treatment for pH adjustment; SHE corrected (reference 225 mV). ! "#$ " "#$ "#% "#& "#' " ( $" $( %" )*+,-. /0#1 2$#1 2%#1 /0#$$ 2$#$$ 2%#$$ 3 4-5 ,678. ! "#$ " "#$ "#% "#& "#' "#( "#) "#* " %" '" )" +" ,-./01 2345./6781 : &#: (#9: (#: $$#9: $$#:

PAGE 164

164 (a) Eh vs. HCl (b) pH vs. HCl Figure A 8 . ORP results from acid titration with HCl in chlor alkali samples A and C . Co mparison of (a) Eh and (b) pH using chlor alkali samples A1, A2 and aged sample C. ! !"# !"$ !"% !"& !"' !"( !") !"* ! !"$ !"& !"( !"* # #"$ #"& #"( #"* +,-. ! / 012-342 5# 5$ 4 ! " # $ % & ' ( )* )) )! * *+! *+# *+% *+' ) )+! )+# )+% )+' ,-. /01.-21 3) 3! 2

PAGE 165

165 Table A-2. Eh (m V) values for pH dependent redox potential in chlor-alkali samples. Eh as ORP corrected for temperature and SHE (reference 240 mV). A1: pHi ~11, ~6 g/L Hg A2: pHi ~11, ~50 g/L Hg C+: pHi ~9 ~20 g/L Hg pH Eh (m V) pH Eh (m V) pH Eh (m V) 10.89 125 11.26 196 8.86 194 9.64 185 9.08 285 8.22 272 7.27 260 7.03 314 6.7 308 5.66 291 6.02 332 5.91 351 3.95 571 4.95 411 4.45 438 3.21 652 4.16 556 3.67 540 2.46 666 2.99 680 2.81 594 2 675 2 694 2.07 617 1.41 683 1 703 1.69 638 1 690 1.08 675 Effect of UV-C on Solution pH Figure A-9. Changes in solution pH for chlor -alkali sample A during UV -C irradiation. ! " # $ % & ' ( )* )) )! * )* !* "* #* $* %* &* '* +, ./0 12345 672+89:;<) =;
PAGE 166

166 UV with Synthetic Mercury Wastewater Baseline emoval DI ater. experiment us Hg solutions irradiated with UV C in UVAC batch reactors under an open air atmosphere olutions were filtered using 0.45 m pore size nylo n (Whatman) media filters. The results for baseline Hg removal by comparison of Hg stock type Hg(NO 3 ) 2 and HgCl 2 are found in Figure A-10. Figure A-10. UVAC Hg removal in DI solutions prepared with variable Hg spikes. 0.45 m filtration with (!) 20 (/) 200 g L$1 Hg for Hg(NO3)2 (solid line) and HgCl2 (dotted line). The distribution of Hg in DI water was measured before and after UV irradiation in DI water solutions buffered to pH 3, ~7 and 11 and then spiked with Hg(NO 3 ) 2 (reagents shown below in Table A-3). ! !"# !"$ !"% !"& ! #'%!&'(!)'*!#+ ,-.. ,-012 3-456789:;<=>/?=@A<= 4BC ! 8 ";<=>/?=@A<=-D " $! $! $!! $!! E6F/E05

PAGE 167

167 contact times were varied up to 70 min in total, and Hg was measured before and after each contact time in order to calculate aqueous, filterable, reactor adsorbed and volatile gaseous Hg fraction. The results are shown in Table B-4 and Figure 5-1. Table A-3. Reagents used to create pH-buffered synthetic Hg wastewater solutions. pH Salt reagents and acid or base 3 1L = 10.21 g potassium hydrogen phthalate and 22.3 mL of 1.0 M HCl ~7 1L = 6.81 g potassium di-hydrogen phosphate and 29.1mL of 1.0 M NaOH 11 1L = 2.10 g sodium bicarbonate and 2 2.7 mL of 1.0 M NaOH Table A-4. Baseline UVAC Hg removal results (!g/L Hg) using 0.45 !m nylon filters. Hg concentrations represent fractions occurring as filtered solida , *Hg(s), gaseous, *Hg(0)g , or as residual reactor Hg, *HgR . tUV -C pH 3 %HgT pH ~7 %HgT pH 11 %HgT (min) *Hg(s) *Hg(0)g *HgR *Hg(s) *Hg(0)g *HgR *Hg(s) *Hg(0)g *HgR 0 18.7 12.2 14.7 5 23.0 0.0 0.3 9.9 2.3 0.5 11.2 6.1 1.4 15 19.3 4.9 0.4 11.3 5.7 1.2 12.6 5.2 0.7 35 28.7 2.6 0.7 7.3 5.9 2.1 10.9 5.2 0.6 70 25.6 5.5 0.5 6.1 6.1 1.3 12.4 9.8 0.7 pH 3 %stdev pH ~7 %stdev pH 11 %stdev (min) *Hg (s) *Hg(0) g *Hg R *Hg (s) *Hg(0) g *Hg R *Hg (s) *Hg(0) g *Hg R 0 0.28 0.54 0.93 5 0.49 0.77 0.22 0.22 1.29 0.20 1.32 0.9 5 0.16 15 1.18 2.11 0.31 0.44 0.57 0.28 0.31 0.65 0.02 35 0.33 1.44 0.01 0.55 0.96 0.16 0.46 1.26 0.19 70 0.88 0.40 0.19 0.00 0.65 0.22 0.79 0.93 0.32 a.Non UV treated Hg fractions were identical to the results shown for tUv-C =0 over the 70 min te st. UV irradiated *Hg distribution as % *HgT in solutions of DI water and Hg(NO3)2 with varying pH and tUVC, C0=20 g/L Hg for pH 3, pH ~7 and pH 11. Standard deviation calculated from duplicate experimental results.

PAGE 168

168 DOM Photo-Transformations Table A-5. Measured Abs-254 (cm-1) in UVAC treated chlor-alkali sample C. Sample pHin of 11, C0 ~50 mg/L DOCa. tUV C (min) 0 15 35 55 80 110 pH 3 0.797 0.732 0.481 0.451 0.412 0.405 4 0.842 0.731 0.613 0.593 0.578 0.537 5 0.725 0.673 0.666 0.632 0.597 0.597 6 0.741 0.74 0.72 0.723 0.765 0.638 7 0.759 0.776 0.715 0.719 0.649 0.588 9 0.761 0.748 0.753 0.695 0.686 0.618 11 0.762 0.741 0.722 0.72 0.684 0.682 a.Values reflect base10 absorbance divided by the quartz cell path length.The organic solution DOC concentrations were measured following UV irradiation in the batch reactors using IHSS DOM isolates. Initial DOC is given and C/C0 values are provided in Table A-6. In each experiment, the rate of change was largest in the first 5 minutes of UV-C irradiation. Table A-6. Changes in DOC concentration from irradiation by UV-C and UV -B. Initial DOC C0 and C/C0 for 5, 30, and 60 min. tUv -C SRNOM SRHA SRFA NLHA NLFA NLDOM C0 (mg/L) 22.2 9.4 25.4 21.0 26.5 15.2 C/C0 (UV-C) 5 min 0.97 0.91 0.99 0.78 0.73 0.80 30 min 0.93 0.89 0.95 0.73 0.68 0.79 60 min 0.78 0.69 0.81 0.53 0.55 0.62 C/C0 (UV-B) 5 min 0.98 0.98 0.97 30 min 0.98 0.96 0.96 60 min 0.97 0.93 0.95

PAGE 169

169 Sulfide Speciation Calculations In solutions prepared using Na2S'9H2O salt crystals, total sulfide represents the summation S(II) = [H2S]+[HS–]+[S2–]. The ratio for S/Hg is calculated for mass and molar sulfide equivalents, where for solutions prepared with pH buffers 3, 7 and 11, sulfide principally occurs as H2S and HS– species with negligible contributions from free sulfide ion, <<0.01% total S(II). Table B -7. S/Hg ratio for C0 = 20 g/L Hg (0.1 M Hg) in pH 7 solutions. total S mol eq total S mass eq S/Hg mass ratio S/Hg mo lratio sulfide species S/Hg mol ratio [S2$] M [S2$] mg/L total S total S [H2S] M [HS$] M 56% H2S 44% HS$ 6.7E 04 2.1E 08 0.033 0.21 1.2E 02 9.2E 03 0.12 0.10 1.3E 03 4.2E 08 0.067 0.42 0.023 0.018 0.25 0.19 0.0134 4.2E 07 0.67 4.2 0.232 0.184 2.47 1.91 0.134 4.2E 06 6.68 41.8 2.32 1.84 24.7 19.1 Table B -8. Summary of S/Hg mass and molar ratios for pH 3, 7, and 11 solutions. C0 =20 g/L Hg. total S S/Hg S/Hg mol-ratio mass eq mol eq mass ratio mol ratio pH 3 pH 7 pH 11 S2$ mg/L [S2$] M total S total S 100% H2S 56% H2S 44% HS$ 100% HS$ 2.1E 05 6.5E 10 1.0E 03 0.007 0.007 0.004 0.003 0.007 6.7E 04 2.1E 08 0.033 0.21 0.22 0.12 0.10 0.22 1.3E 03 4.2E 08 0.067 0.42 0.44 0.25 0.19 0.43 2.7E 03 8.3E -08 0.13 0.84 0.89 0.49 0.38 0.86 6.7E 03 2.1E 07 0.33 2.1 2.2 1.2 1.0 2.2 0.013 4.2E 07 0.67 4.2 4.4 2.5 1.9 4.3 0.13 4.2E 06 6.7 41.8 44 25 19 43 0.67 2.1E 05 33 209 222 124 95 215 1.3 4.2E 05 67 418 444 247 191 431 5.3 1.7E 04 267 1670 1 775 989 763 1723 40 1.2E 03 2000 12511 13297 7411 5713 12889

PAGE 170

170 Table B -9. Distribution summary for principle sulfide species for pH 3, 7, and 11. The distribution of sulfide species are shown for Na2S concentrations tested in synthetic wastewater experiments, calculated by pKa1 = 7 and pKa2 = 19. sulfide species total S pH 3 pH 7 pH 11 [S2$] M [H 2 S] M [HS $ ] M [H 2 S] M [HS $ ] M [H 2 S] M [HS $ ] M 6.5E 10 6.5E 04 5E 08 3.6E 04 2.9E 04 8E 08 6.5E 04 2.1E 08 0.02 2E 06 1.2E 02 9.2E 03 3E 06 0.021 4.2E 08 0.04 3E 06 0.023 0.018 5E 06 0.042 8.3E 08 0.08 7E 06 0.046 0.037 1E 05 0.08 2.1E 07 0.21 2E 05 0.116 0.092 3E 05 0.21 4.2E 07 0.42 3E 05 0.23 0.18 5E 05 0.42 4.2E 06 4.16 3E 04 2.3 1.8 5E 04 4.2 2.1E 05 21 2E 03 11.6 9.2 3E 03 21 4.2E 05 42 3E 03 23.2 18.4 5E 03 42 1.7E 04 167 0.013 93 74 0.021 167 1.2E 03 1247 0.10 695 552 1.57 1246 Figure B-11. Free S2– ion pC -pH diagram for total [S2–] !M concentrations. 8 9 10 11 12 13 14 15 16 1 3 5 7 9 11 13 [S!!] -log C (M) pH 1.2E+03 1.7E+02 4.2E+01 2.1E+01 4.2E+00 4.2E-01 2.1E-01 8.3E-02 4.2E-02 2.1E-02 6.5E-04

PAGE 171

171 APPENDIX B LITERATURE REFERENCE DATA Figure B-1. Estimated anthropogenic contribution to multi -phasic gaseous and particulate Hg, represented as fluxes of Hg (Mg yr1) and Hg reservoirs (Gg). The percentages in brackets are estimated increases in pools and fluxes due to anthropogenic activities over the past 150 years; plus a negligible contribution from inert particulate Hg. Sourced from Driscoll et al. [38] . This is an unofficial adaptation of an article that appears in an ACS publication. ACS has not endorsed the content or the context of its use. DOM Photo-reactivity and Activation The stepwise reaction sequence for DOM photo -activation is illustrated in Figure B-2 for absorption of a photon (hvex) by CDOM functional groups with varying photo sensitizing potential (the ability of CDOM to form 3DOM). This causes DOM to undergo electronic excitation from the ground state to the s inglet state (1DOM*), during which electrons are excited from S0 level to the S1 level in a process of leaving the highest occupied molecule orbital (HOMO) and moving to the lowest un -occupied molecular

PAGE 172

172 orbital (LUMO). It can then cross into an uncharged triplet state (3DOM) by intersystem crossing (ISC), or relax to its ground state by fluorescing energy ( hvf). Figure B-2. Schematic energy level diagram for excited state DOM. Adapted from processes proposed for a diatomic molecule by Senesi [219] . Temperatureependent hoto-reactivity. The temperature dependence of a reactants photo-reactivity can measured by activation energy, Ea (kJ mol$ 1), which is calculated from the Arrhenius equation in (B-1) by using experiment al measures of apparent quantum yield ()a) for radicals produced per photon absorbed. (B-1) The following photo -activation energies were measured from heterogeneous and complex DOM isolates and NOM using unspecified wavelength from UV -Solar. A comparison of apparent activation energies in Table B-2 and Table B-1 of Appendix B reveals relative differences among common reactants, including heterogeneous and complex DOM isolates and NOM found in water measured using UV-Solar and UV over a range 290 " ( " 400 nm. Similar values were found for NOM<1 kDa and fulvic acids, which generally have lower average molecular weights, indicating a consistent !"#$%&'()"%'*& $"+,-'% $%.%' / 0123&2456,#78+* $%.%' / 0193& 456 "+%'#+.-&)7+:'#$"7+ !"#!-87#'$)'+)' !"$%.;$7#<%"7+ 0= 23&>456 / %#"<-'% $%.%' $')7+*&'()"%'*& $"+,-'% $%.%' / 0 1? 3 0= ?3 / '()"%'*&%#"<-'% $%.%'& l n (!a) = l n ( A ) " Ea R T

PAGE 173

173 mechanism for OH production, which McKay et al. propose occurs by secondary photoreactions [182] . Table B-1. Photo-reactant activation energies for OH production in water. UV with Reactant/ Reaction in water (1) Ea e (kJ mol$1) UV-Solar with peroxide 5 – 7 UV-Solar with nitrate 19 UV-Solar with DOM in deionized water 21.6a – 34.3b 21.6a – 34.3b 17.8b 26.6b 33.3b a.DOM of Antarctic origin of terrestrial influence b. Suwannee River origin of microbia l influence. (1)Adapted from McKay et al. with error (1.1 7.2) [182] .Table B-2. Photo-reactant activation energies for H2O2 in water production. UV with Reactant/ Reaction in water (1) Ea e (kJ mol$1) UV with DOMa in seawater 16.9 – 19.7 21.1 – 28.7 22.5 – 33.2 33.5 – 52.7 a.DOM of Antarctic seawater origin or oligotrophic influence. (1) Adapted from Kieber et al. error (2.0 6.3) [220] . H2O2 h v! " !O H N O3 ! h v N O3 ! "# $ # # N O2 + O! H2O# $ # #O H + O H! F AD O M h v! " !O H H AD O M h v! " !O H N O MBU L K h v! " !O H N O M< 5 kD a h v! " !O H N O M< 1 kD a h v! " !O H D O Mh v 2 9 0 n m! " ! ! ! H2O2 D O Mh v 3 0 0 n m! " ! ! ! H2O2 D O Mh v 3 2 0 n m! " ! ! ! H2O2 D O Mh v 4 0 0 n m! " ! ! ! H2O2

PAGE 174

174 UV Spectra Characterization of Sulfur Species Figure B 3 . UV spectra after 273 min photolysis for sulfur 6.25 mM solutions . (A) H 2 S (B) HS Ð (as Na 2 S ' 9H 2 O ) and (C) S 2 2 $ (as Na 2 S 2 ); and (D) 0.1 M NaHS. (b) hydrogen evolution by 254 nm wavelength for Na 2 S. b) after 273 minutes of photolysis. Data sourced from Linkous et al. [112] This is an unofficial adaptation of an article that appears in an ACS publication. A CS has not endorsed the content or the context of its use. !"#$%$&'()* +&,./012/"&3$ ! "#$ 4$"3(51&*(5,$*+,5&67821'$&*$#1%9(51&*+,:"%$

PAGE 175

175 Table B-3. Hydration energies and select optical transition energies for comparison of Hg-S complex formation and photo-dissociation. Molecule hydration energy(1) lowest optical transition energy, -E (S0 !T)a (kcal mol$1) (eV) H2S –2.6 HS$ –71.1 HgS –27.6 0.1, 0.1, 1.6 HgS' H2O –37.3 3.1, 3.2, 4.2 HgOHSH0 –17.4 4.9, 5.0, 5.5 Hg(SH)2OH$ –57.3 Hg(SH)2 0 –6.2 HgSH2 $ –54.2 HgSOH$ –61.9 HgCl2 –9.0 6.4, 6.8, 6.8 a.Measured in the gas phase; values for the aqueous phase predicted to be ~1eV greater; T represents the S1 Hg L 0 * orbital. (1) Quantum mechanical calculations by Tossell et al. [60]Natural Organic Matter Characterization UV Wavelength and DOM Photo-activation The absorption characteristics of CDOM vary with the wavelength spectrum such that longer wavelengths, i.e. visible light , are more likely to cause charge transfer between dono rs and acceptors, whereas at shorter wavelengths, i.e. UV -Solar and UVC, direct absorption occurs from acceptor and donor CDOM moieties. Direct absorption from UV-C produces higher OH formation, i.e. ~1.8 *10$ 0 M$ 1 s1 OH at 308 nm compared with ~6.1 *1011 M$ 1 s1 OH at 355 nm [100, 221] , although the mechanism of direct photo-production of OH by CDOM has yet to be conclusively addres sed [222] . The findings support the hypothesis that UV wavelength affects the activation of either electron-accepting or electron-donating CDOM. Namely, that UV -C forms 3DOM* from sensitized electron-donor type CDOM, an on the other hand, UV-Solar produced 3DOM*

PAGE 176

176 from electron-accepting type CDOM, which can instead produce 1O2. A comparison of DOM production of photo-oxidants found no difference in quantum yields for 1O2, 3DOM* and H2O2; the exception was higher OH yields by an order of magnitude for UV-C relative to UV-Solar reactions [138] . The authors attribute increased OH formation to H2O2 produced from electron–donor type CDOM via photo-oxidation of substitut ed phenol groups, which principally absorb in the UV-C range [125] . On the other hand, phenol moieties can produce 1O2 from electron–accepting CDOM that photo-produce oxidizing triplet ketones/aldehyde sensitizers and quinone triplet states [126, 223, 224] . Dissolved Organic Matter Fraction Classification and Possible Components Table B-4. Classes of dissolved organic matter fractions. Fraction Class of organic compounds Reference Hydrophobic fractions (VHA + SHA) hydrophobic acid soil fulvic acids, C5-C9 aliphatic carboxylic acids 1, 2) ring aromatic carboxylic ac ids, ( 1, 2)ring phenols [225 228] hydrophobic base (1, 2)ring aromatics (except pyridine), protein substances [226 228] hydrophobic neutral mixture of hydrocarbons, >C5 aliphatic alcohols, amides, aldehydes, ketones, esters, >C9 aliphatic carboxylic acids and amines, >3 ring aromatic carboxylic acids and amines Hydrophilic fractions (CHA + NEU) hydrophilic acid mixtures of hydroxy acids,
PAGE 177

177 IHSS Organic Matter Characterization The following section describes IHSS DOM isolates listed in Table B-5 that were used in experiments with a summary of available characteristics in literature research. Further information can be obtained on the IHSS website, i ncluding fluorescence spectra (www.humicsubstances.org/spectra). Table B-5. Identification of IHSS DOM isolate by name and acronym. Name of DOM isolate Acronym IHSS ID Suwannee River Aquatic NOM SRNOM 1R101N Suwannee River II Standard HA SRHA 2S101H Suwannee River II Standard FA SRFA 2S101F Nordic Lake Reference HA NLHA 1R105H Nordic Lake Reference FA NLFA 1R105F DOM functional group charge density calculations The overall charge density is humic substances, Qtot in meq g1 C, increases with pH. This value can be calculated using the modified Henderson -Hasselbalch equation in (B-2) for two classes of binding sites: Carboxyl groups are defined by the charge density occurring at pH 8 and p henolic groups are characterized as having two times the change in charge density between pH 8 and pH 1 0 [229] . (B-2) Qt o t= Q11 + K1[ H+]( )! n1 " # $ $ % & ' ' + Q21 + K2[ H+]( )! n2 " # $ $ % & ' '

PAGE 178

178 Table B-6. Acidic functional groups of IHSS DOM. Sample Amino acid (2) (!mol/g) Carboxyl (meq/ C) pH 8(1) Phenolic (meq/C) pH 810(1) Q1 Log K1 n1 Q2 Log K2 n2 Glutathione Cysteine NLHA 17 nd 9.06 3.23 10.32 4.32 4.22 1.64 9.89 1.11 NLFA 6.1 nd 11.16 3.18 12.15 3.79 3.95 1.49 9.67 1 SRHA 12 nd 9.13 3.72 9.74 4.35 3.3 4.48 10.44 1.73 SRNOM nd nd 9.85 3.94 10.57 3.94 3.6 2.61 9.74 1.19 (1) Source data from Ritchie et al . [229] (2 ) Source data from Laboratory of Chemistry of t he Colloids of Soils and Waters, The University of Birmi ngham, Edgbaston, Engl and (ndnot determine d); retriev ed from IHSS w ebsite, www.humicsubstances.org. Figure B-4. IHSS DOM charge density vs solution pH from carboxyl and phenol groups. Calculated for pH 2 to 12 using the modified Henderson-Hasselbalch equation with IHSS characterization data in Table C -6. 0 2 4 6 8 10 12 14 2 4 6 8 10 12 charge density meq/g-C pH SRNOM SRHA SRFA NLHA NLFA

PAGE 179

179 Estimations of DOM electron donating and accepting capacities Table B-7. Metal concent and electron donating (EDC) and electron accepting (EAC) capacities of IHSS DOM. EAC pH 7 EDC pH 7 EDC pH 7 EDC pH 9 Fe Cu Eh 0.49 V Eh 0.61 V Eh 0.73 V Eh 0.61 V DOM mmol e -/(g-HS) !mol/ (g-HS) NLFA 10533 14800 3326125 28684 4.89.02 0.06.02 NLHA 11977 18601 412825 32412 18.15.49 0.2.02 SRHA 9626 182896 36846 33978 22.91.53 0.46.02 SRFA 671 1367 28485 24732 6.81.11 0.33.01 SRNOM 653 1131 24212 2 1559 n/a n/a Data source from Aeschbacher et al. ; retrieved from supporting i nformation of published manuscript [137]. DOM elemental composition Table B-8. Composition of IHSS DOM based on elemental analysis. DOM H2O Ash H N S P C O C/O ratio SRNOM 8.15 7 4.19 1.1 0.65 0.02 52.47 42.69 0.81 SRHA 9.2 0.45 3.98 0.68 0.46 0.01 52.31 45.12 0.86 SRFA 16.90 0.58 4.36 0.67 0.46 0.004 52.34 42.98 0.82 NLHA 9.1 0.31 3.97 1.16 0.58 0.01 53.33 43.09 0.81 NLFA 20.4 1.04 4.28 1.17 0.54 0.013 52.63 42.04 0.80 H2O content is the %(w/w) of H2O in the air equilibrated sample (a function of relative humidity). Ash is the %(w/w) of inorganic residue in a dry sample. C, H, O, N, S, and P are the elemental composition in %(w/w) of a dry, ash free sample. Source data: Elemental analyses by Huffman Laboratories, Wheat Ridge, CO, USA; Isotopic analyses by Soil Biochemistry Laboratory, Dept. of Soil, Water, and Climate, University of Minnesota, St. Paul, MN, USA ; retrieved from IHSS website, www.humicsubstances.org.

PAGE 180

180 Figure B-5. Distribution of major elements in IHSS DOM based on elemental analysis. Figure is based on data in Table C-8. Carbon distribution in DOM functional groups Table B-9. Estimates of carbon distribution in IHSS DOM using 13CNMR. Sample (1) Carbonyl Carboxyl Aromatic Acetal Heteroaliphatic Aliphatic NLHA 10 19 38 7 11 15 NLFA 10 24 31 7 12 18 SRHA 6 15 31 7 13 29 SRNOM 8 20 23 7 15 27 (1)Source data from Thorn et al. [230] ; retrieved from IHSS web site, www.humicsubstances.org. 0 5 10 15 20 25 H!O Ash H N S P SRNOM SRHA SRFA NLHA NLFA % elemental distribution

PAGE 181

181 Figure B 6 . Percent carbon distribution in IHSS DOM functional groups using 13 CNMR . Figure is based on data in Tab le C 9. 0 10 20 30 40 Carbonyl Carboxyl Aromatic Acetal Heteroaliphatic Aliphatic SRNOM SRHA SRFA NLHA NLFA !" C NMR %C distribution

PAGE 182

182 LIST OF REFERENCES [1] World Chlorine Ins t itute and UNEP G l obal Chlor alkali Mercury Partn ership. Reports & Publications: Estimate of Global Mercury Cell Chlorine Capacity. Retrieved from: http://www.worldchlorine.org/publications/unep chlor alkali mercury partnership/ [2] Tonini, D.R.; Gauvin, D.A.; Soffel, R.W.; Freeman, W.P. Achieving low mercury concentrations in chlor ! alkali wastewate rs. Environ. Prog. 2003, 22 (3), 167 173. [3] Chlor Alkali Industry 2008 Mercury Use and Emissions in the United States (12th Annual Report), August 2009, pp. 2 Ð 3. The Chlorine Institute, Inc. Retrieved from U.S. EPA archived reports. [4] Final Effluent L imitations Guidelines and Standards for the Steam Electric Power Gene rating Point Source Category, S 2015 (EPA HQ OW 2009 0819). U.S EPA. Retrieved from http://www.regulations.gov [5] Legrini, O.; Oliv eros, E.; Braun, A. Photochemical processes for water treatment. Chem. Rev. 1993, 93 (2), 671 698. [6] William IV, L.; Kostedt, I.; Drwiega, J.; Mazyck, D.W.; Lee, S.; Sigmund, W.; Wu, C.; Chadik, P. Magnetically agitated photocatalytic reactor for photoca talytic oxidation of aqueous phase organic pollutants. Environ. Sci. Technol. 2005, 39 (20), 8052 8056. [7] Litter, M.I. Treatment of chromium, mercury, lead, uranium, and arsenic in water by heterogeneous photocatalysis. Advances in Chemical Engineering 2 009, 36 , 37 67. [8] Prairie, M.R.; Evans, L.R.; Stange, B.M.; Martinez, S.L. An investigation of titanium dioxide photocatalysis for the treatment of water contaminated with metals and organic chemicals. Environ. Sci. Technol. 1993, 27 (9), 1776 1782. [9] Zheng, S.; Jiang, W.; Rashid, M.; Cai, Y.; Dionysiou, D.D.; O'Shea, K.E. Selective Reduction of Cr (VI) in Chromium, Copper and Arsenic (CCA) Mixed Waste Streams Using UV/TiO2 Photocatalysis. Molecules 2015, 20 (2), 2622 2635. [10] Tanaka, K.; Harada, K.; Murata, S. Photocatalytic de position of metal ions onto TiO 2 po wder. Solar Energy 1986, 36 (2), 159 161. [11] Byrne, H.E.; Borello, A.; Bonzongo, J.; Mazyck, D.W. Investigations of photochemical transformations of aqueous mercury: Implications for water ef fluent treatment technologies. Water Res. 2009, 43 (17), 4278 4284. [12] Byrne, H.E. Adsorption, Photocatalysis, and Photochemistry of T race Level Aqueous Mercury. PhD Dissertation , University of Flo rida, 2009. [13] Botta, S.G.; Rodrõguez, D.J.; Leyva, A.G.; Litter, M.I. Features of the transformation of Hg(II) by heterogeneous photocatalysis over TiO 2 . Catalysis today 2002, 76 (2), 247 258.

PAGE 183

183 [14] Bussi, J.; Ohanian, M.; V‡zquez, M.; Dalchiele, E.A. Photocatalytic removal of Hg from solid wastes of chlor alkali plant. J. Environ. Eng. 2002, 128 (8), 733 739. [15] Wang, X.; Pehkonen, S.; Ray, A.K. Photocatalytic reduction o f Hg (II) on two commercial TiO 2 catalysts. Electrochim. Acta 2004, 49 (9), 1435 1444. [16] Khalil, L.; Ro phael, M.; Mourad, W. The removal of the toxic Hg (II) salts from water by photocatalysis. Applied Catalysis B: Environmental 2002, 36 (2), 125 130. [17] Pitoniak, E.; Wu, C.; Londeree, D.; Mazyck, D.; Bonzongo, J.; Powers, K.; Sigmund, W. Nanostructured s ilica gel doped with TiO2 for mercury vapor control. Journal of nanoparticle research 2003, 5 (3 4), 281 292. [18] Mazyck, D.W. and Casasœs, A.I. System and method for purifying air via low energy, in situ regenerated silica titania composites. United Stat es Patent 8980171 B2. Issued Mar 17, 2015. Primary Class: 422/4, 422/122. 2011, US 12/903,543 (US8980171 B2). [19] Stokke, J.M. and Mazyck, D.W. Development of a regenerable system employing silica titania composites for the recovery of mercury from end bo x exhaust at a chlor alkali facility. J. Air Waste Manage. Assoc. 2008, 58 (4), 530 537. [20] Byrne, H.E. and Mazyck, D.W. Removal of trace level aqueous mercury by adsorption and photocatalysis on silica Ð titania composites. J. Hazard. Mater. 2009, 170 (2) , 915 919. [21] Borello, A.M. Photochemical Reaction Mechanisms Of Aqueous Mercury For Application Of Removal Technologies. PhD Dissertation, University of Florida, 2013. [22] Si, L. and Ariya, P.A. Aqueous photoreduction of oxidized mercury species in pre sence of selected alkanethiols. Chemosphere 2011, 84 (8), 1079 1084. [23] Tennakone, K. and Ketipearachchi, U. Photocatalytic method for removal of mercury from contaminated water. Applied Catalysis B: Environmental 1995, 5 (4), 343 349. [24] Quici, N.; Mo rgada, M.E.; Gettar, R.T.; Bolte, M.; Litter, M.I. Photocatalytic degradation of citric acid under different conditions: TiO 2 heterogeneous photocatalysis against homogeneous photolytic processes promoted by Fe (III) and H 2 O 2 . Applied Catalysis B: Environm ental 2007, 71 (3), 117 124. [25] Sulfide Precipitation Of Heavy Metals: Effect Of Complexing Agents. Industrial Environmental Research Laboratory (EPA/600/S2 84 023). Retrieved from http://nepis.epa.gov [26 ] Hagan Rogers, A.M.; Casasœs, A.I.; Mazyck, D.W. Aqueous mercury removal from chlor alkali wastewater activated by UV 254 nm. Manuscript under revision for re submission to Journal of Water Process Engineering (2016) . 2016 .

PAGE 184

184 [27] Ewers, U.; Krause, C.; Sch ulz, C.; Wilhelm, M. Reference values and human biological monitoring values for environmental toxins. Int. Arch. Occup. Environ. Health 1999, 72 (4), 255 260. [28] Mergler, D.; Anderson, H.A.; Chan, L.H.M.; Mahaffey, K.R.; Murray, M.; Sakamoto, M.; Stern, A.H. Methylmercury exposure and health effects in humans: a worldwide concern. AMBIO: A Journal of the Human Environment 2007, 36 (1), 3 11. [29] Silbergeld, E.K.; Silva, I.A.; Nyland, J.F. Mercury and autoimmunity: implications for occupational and envir onmental health. Toxicol. Appl. Pharmacol. 2005, 207 (2), 282 292. [30] Gardner, R.M.; Nyland, J.F.; Silbergeld, E.K. Differential immunotoxic effects of inorganic and organic mercury species in vitro. Toxicol. Lett. 2010, 198 (2), 182 190. [31] Via, C.S.; Nguyen, P.; Niculescu, F.; Papadimitriou, J.; Hoover, D.; Silbergeld, E.K. Low dose exposure to inorganic mercury accelerates disease and mortality in acquired murine lupus. Environ. Health Perspect. 2003, 111 (10), 1273 1277. [32] Baeuml, J.; Bose O'Reil ly, S.; Gothe, R.M.; Lettmeier, B.; Roider, G.; Drasch, G.; Siebert, U. Human biomonitoring data from mercury exposed miners in six artisanal small scale gold mining areas in Asia and Africa. Minerals 2011, 1 (1), 122 143. [33] Zahir, F.; Rizwi, S.J.; Haq, S.K.; Khan, R.H. Low dose mercury toxicity and human health. Environ. Toxicol. Pharmacol. 2005, 20 (2), 351 360. [34] AMAP Technical Background Report for the Global Mercury Assessment 2013. UNEP Arctic Monitoring and Assessment Programme, Oslo, Norway. 2 013 . [35] Ariya, P.A.; Dastoor, A.P.; Amyot, M.; Schroeder, W.H.; Barrie, L.; Anlauf, K.; Raofie, F.; Ryzhkov, A.; Davignon, D.; Lalonde, J. The Arctic: a sink for mercury. Tellus B 2004, 56 (5), 397 403. [36] Selin, N.E.; Jacob, D.J.; Park, R.J.; Yantosca , R.M.; Strode, S.; JaeglŽ, L.; Jaffe, D. Chemical cycling and deposition of atmospheric mercury: Global constraints from observations. Journal of Geophysical Research: Atmospheres (1984 Ð 2012) 2007, 112 (D2). [37] Schuster, P.F.; Krabbenhoft, D.P.; Naftz, D.L.; Cecil, L.D.; Olson, M.L.; Dewild, J.F.; Susong, D.D.; Green, J.R.; Abbott, M.L. Atmospheric mercury deposition during the last 270 years: a glacial ice core record of natural and anthropogenic sources. Environ. Sci. Technol. 2002, 36 (11), 2303 2310. [38] Driscoll, C.T.; Mason, R.P.; Chan, H.M.; Jacob, D.J.; Pirrone, N. Mercury as a global pollutant: sources, pathways, and effects. Environ. Sci. Technol. 2013, 47 (10), 4967 4983. [39] Fitzgerald, W.F.; Lamborg, C.H.; Hammerschmidt, C.R. Marine biogeoc hemical cycling of mercury. Chem. Rev. 2007, 107 (2), 641 662.

PAGE 185

185 [40] Sunderland, E.M. and Mason, R.P. Human impacts on open ocean mercury concentrations. Global Biogeochem. Cycles 2007, 21 (4). [41] Final Water Quality Guidance for the Great Lakes System: C riteria for the Protection of Wildlife: DDT, Mercury, 2,3,7,8 TCDD, PCBs (60 FR 15366). Retrieved from https://federalregister.gov/a/95 6671 [42] Regulatory Impact Analysis of the Final Clea n Air Mercury Rule (EPA 452/R 05 003). U.S. EPA Office of Air Quality Planning and Standard s , 2005 . [43] Casasœs, A.; Gruss, A.; Baun, D.; Morales, M.; Mazyck, D. Silica Titania Ð Coated Packing: Novel Solution Capable of 90% Hg Capture with Low Operation an d Maintenance Costs. J. Environ. Eng. 2012, 139 (1), 86 94. [44] O'Dowd, W.J.; Hargis, R.A.; Granite, E.J.; Pennline, H.W. Recent advances in mercury removal technology at the National Energy Technology Laboratory. Fuel Process Technol 2004, 85 (6), 533 54 8. [45] Yan, R.; Liang, D.T.; Tay, J.H. Control of mercury vapor emissions from combustion flue gas. Environmental Science and Pollution Research 2003, 10 (6), 399 407. [46] Stumm, W. and Morgan, J.J. Aquatic chemistry: chemical equilibria and rates in nat ural waters. John Wiley & S ons: 2012. [47] Schuster, E. The behavior of mercury in the soil with special emphasis on complexation and adsorption processes a review of the literature. Water Air & Soil Pollution 1991, 56 (1), 667 680. [48] Benoit, J.M.; Maso n, R.P.; Gilmour, C.C. Estimation of mercury ! sulfide speciation in sediment pore waters using octanol Ñ water partitioning and implications for availability to methylating bacteria. Environmental toxicology and chemistry 1999, 18 (10), 2138 2141. [49] Lindsa y, W. and Sadiq, M. Use of pe pH to predict and interpret metal solubility relationships in soils. Sci. Total Environ. 1983, 28 (1), 169 178. [50] Sadiq, M. Toxic metal chemistry in marine environme nts. 1992 . [51] Cardona, M.; Kremer, R.; Lauck, R.; Siegle , G.; Mu–oz, A.; Romero, A. Electronic, vibrational, and thermodynamic properties of metacinnabar % HgS, HgSe, and HgTe. Physical Review B 2009, 80 (19), 195204. [52] Charnock, J.M.; Moyes, L.N.; Pattrick, R.A.; Mosselmans, J.F.W.; Vaughan, D.J.; Livens, F .R. The structural evolution of mercury sulfide precipitate: an XAS and XRD study. American Mineralogist 2003, 88 (8 9), 1197 1203. [53] Pai, M.; Buttrey, D.; Joshi, G.; Honig, J. Electrical and magnetic properties of & HgS (cinnabar). Physical Review B 19 81, 24 (2), 1087.

PAGE 186

186 [54] Xu, Y. and Schoonen, M.A. The absolute energy positions of conduction and valence bands of selected semiconducting minerals. Am. Mineral. 2000, 85 (4), 543 556. [55] Skyllberg, U.; Bloom, P.R.; Qian, J.; Lin, C.; Bleam, W.F. Complexa tion of mercury (II) in soil organic matter: EXAFS evidence for linear two coordination with reduced sulfur groups. Environ. Sci. Technol. 2006, 40 (13), 4174 4180. [56] Liu, J. and Huang, C. Electrokinetic characteristics of some metal sulfide water inter faces. Langmuir 1992, 8 (7), 1851 1856. [57] Deonarine, A. and Hsu Kim, H. Precipitation of mercuric sulfide nanoparticles in NOM containing water: Implications for the natural environment. Environ. Sci. Technol. 2009, 43 (7), 2368 2373. [58] Dyrssen, D. a nd Wedborg, M. The sulphur mercury (II) system in natural waters. Water Air & Soil Pollution 1991, 56 (1), 507 519. [59] Benoit, J.M.; Gilmour, C.C.; Mason, R.P.; Heyes, A. Sulfide controls on mercury speciation and bioavailability to methylating bacteria in sediment pore waters. Environ. Sci. Technol. 1999, 33 (6), 951 957. [60] Tossell, J. Calculation of the structures, stabilities, and properties of mercury sulfide species in aqueous solution. The Journal of Physical Chemistry A 2001, 105 (5), 935 941. [ 61] Drott, A.; Bjorn, E.; Bouchet, S.; Skyllberg, U. Refining Thermodynamic Constants for Mercury(II) Sulfides in Equilibrium with Metacinnabar at Sub Micromolar Aqueous Sulfide Concentrations. Environ. Sci. Technol. 2013, 47 (9), 4197 4203; 10.1021/es3041 324n. [62] Jay, J.A.; Morel, F.M.; Hemond, H.F. Mercury speciation in the presence of polysulfides. Environ. Sci. Technol. 2000, 34 (11), 2196 2200. [63] Turner, A.; Millward, G.E.; Le Roux, S.M. Sediment water partitioning of inorganic mercury in estuarie s. Environ. Sci. Technol. 2001, 35 (23), 4648 4654. [64] Heyes, A.; Miller, C.; Mason, R.P. Mercury and methylmercury in Hudson River sediment: impact of tidal resuspension on partitioning and methylation. Mar. Chem. 2004, 90 (1), 75 89. [65] Ravichandran, M. Interactions between mercury and dissolved organic matter ÐÐ a review. Chemosphere 2004, 55 (3), 319 331. [66] Aiken, G.; Brown, P.; Noyes, T.; Pinckney, D. Molecular Size and Weight of Fulvic and Humic Acids from the Suwanee River, Humic Substances in t he Suwannee River, Georgia: Interactions, Properties, and Proposed Structures. US Geological Survey, Denver, Colorado 1989 . [67] Hesterberg, D.; Hansen, P.; Zhou, W. EXAFS study of metal sulphide stability in a contaminated soil. Goldschmidt Conference Tou louse 1998, 62A .

PAGE 187

187 [68] Xia, K.; Skyllberg, U.; Bleam, W.; Bloom, P.; Nater, E.; Helmke, P. X ray absorption spectroscopic evidence for the complexation of Hg (II) by reduced sulfur in soil humic substances. Environ. Sci. Technol. 1999, 33 (2), 257 261. [69] Drexel, R.T.; Haitzer, M.; Ryan, J.N.; Aiken, G.R.; Nagy, K.L. Mercury (II) sorption to two Florida Everglades peats: Evidence for strong and weak binding and competition by dissolved organic matter released from the peat. Environ. Sci. Technol. 2002, 36 (19), 4058 4064. [70] Hesterberg, D.; Chou, J.W.; Hutchison, K.J.; Sayers, D.E. Bonding of Hg (II) to reduced organic sulfur in humic acid as affected by S/Hg ratio. Environ. Sci. Technol. 2001, 35 (13), 2741 2745. [71] Haitzer, M.; Aiken, G.R.; Ryan, J.N. Binding of mercury (II) to dissolved organic matter: the role of the mercury to DOM concentration ratio. Environ. Sci. Technol. 2002, 36 (16), 3564 3570. [72] Chow, C.; Fabris, R.; Drikas, M. A rapid fractionation technique to characterise natural organic matter for the optimisation of water treatment processes. Aqua 2004, 53 , 85 92. [73] Black, F.J.; Bruland, K.W.; Flegal, A.R. Competing ligand exchange solid phase extraction method for the determination of the complexation of dissolved inorganic mercury (II) in natural waters. Anal. Chim. Acta 2007, 598 (2), 318 333. [74] 1wietlik, J.; D2browska, A.; Raczyk Stanis3awiak, U.; Nawrocki, J. Reactivity of natural organic matter fractions with chlorine dioxide and ozone. Water Res. 2004, 38 (3), 547 558. [75] Buchanan, W.; Roddick, F.; Porter, N.; Drikas, M. Fractionation o f UV and VUV pretreated natural organic matter from drinking water. Environ. Sci. Technol. 2005, 39 (12), 4647 4654. [76] Leenheer, J.; Wershaw, R.; Brown, G.; Reddy, M. Characterization and diagenesis of strong acid carboxyl groups in humic substances. Ap pl. Geochem. 2003, 18 (3), 471 482. [77] Kalbitz, K. and Wennrich, R. Mobilization of heavy metals and arsenic in polluted wetland soils and its dependence on dissolved organic matter. Sci. Total Environ. 1998, 209 (1), 27 39. [78] Dong, W.; Bian, Y.; Lian g, L.; Gu, B. Binding constants of mercury and dissolved organic matter determined by a modified ion exchange technique. Environ. Sci. Technol. 2011, 45 (8), 3576 3583. [79] Benoit, J.; Mason, R.P.; Gilmour, C.C.; Aiken, G. Constants for mercury binding by dissolved organic matter isolates from the Florida Everglades. Geochim. Cosmochim. Acta 2001, 65 (24), 4445 4451.

PAGE 188

188 [80] Miller, C.L.; Mason, R.P.; Gilmour, C.C.; Heyes, A. Influence of dissolved organic matter on the complexation of mercury under sulfidic conditions. Environmental Toxicology and Chemistry 2007, 26 (4), 624 633. [81] Ramamoorthy, S.; Springthorpe, S.; Kushner, D. Competition for mercury between river sediment and bacteria. Bull. Environ. Contam. Toxicol. 1977, 17 (5), 505 511. [82] You, S.; Yin, Y.; Allen, H.E. Partitioning of organic matter in soils: effects of pH and water/soil ratio. Sci. Total Environ. 1999, 227 (2), 155 160. [83] Randall, P.M. and Chattopadhyay, S. Mercury contaminated sediment sites Ñ An evaluation of remedial options. En viron. Res. 2013, 125 , 131 149. [84] Waples, J.S.; Nagy, K.L.; Aiken, G.R.; Ryan, J.N. Dissolution of cinnabar (HgS) in the presence of natural organic matter. Geochim. Cosmochim. Acta 2005, 69 (6), 1575 1588. [85] Lu, X. and Jaffe, R. Interaction between Hg (II) and natural dissolved organic matter: a fluorescence spectroscopy based study. Water Res. 2001, 35 (7), 1793 1803. [86] Ravichandran, M.; Aiken, G.R.; Ryan, J.N.; Reddy, M.M. Inhibition of precipitation and aggregation of metacinnabar (mercuric sul fide) by dissolved organic matter isolated from the Florida Everglades. Environ. Sci. Technol. 1999, 33 (9), 1418 1423. [87] Alberts, J.J.; Schindler, J.E.; Miller, R.W.; Nutter, D.E. Elemental mercury evolution mediated by humic acid. Science 1974, 184 (4 139), 895 897. [88] Allard, B. and Arsenie, I. Abiotic reduction of mercury by hum ic substances in aquatic system an important process for the mercury cycle. Water Air & Soil Pollution 1991, 56 (1), 457 464. [89] Gu, B.; Bian, Y.; Miller, C.L.; Dong, W.; J iang, X.; Liang, L. Mercury reduction and complexation by natural organic matter in anoxic environments. Proceedings of the National Academy of Sciences 2011, 108 (4), 1479 1483. [90] Mason, R.; Reinfelder, J.; Morel, F. Bioaccumulation of mercury and meth ylmercury, In Mercury as a Global Pollutant, Springer: 1995; pp. 915 921. [91] Zhang, T.; Kim, B.; Levard, C.; Reinsch, B.C.; Lowry, G.V.; Deshusses, M.A.; Hsu Kim, H. Methylation of mercury by bacteria exposed to dissolved, nanoparticulate, and microparti culate mercuric sulfides. Environ. Sci. Technol. 2012, 46 (13), 6950 6958. [92] Bonzongo, J.J. and Donkor, A.K. Increasing UV B radiation at the earth's surface and potential effects on aqueous mercury cycling and toxicity. Chemosphere 2003, 52 (8), 1263 1 273. [93] Costa, M. and Liss, P. Photoreduction and evolution of mercury from seawater. Sci. Total Environ. 2000, 261 (1), 125 135.

PAGE 189

189 [94] Lalonde, J.D.; Amyot, M.; Kraepiel, A.M.; Morel, F.M. Photooxidation of Hg (0) in artificial and natural waters. Enviro n. Sci. Technol. 2001, 35 (7), 1367 1372. [95] Moore, C. and Carpi, A. Mechanisms of the emission of mercury from soil: Role of UV radiation. Journal of Geophysical Research: Atmospheres (1984 Ð 2012) 2005, 110 (D24). [96] Diffey, B.L. Sources and measuremen t of ultraviolet radiation. Methods 2002, 28 (1), 4 13. [97] Buxton, G.V.; Greenstock, C.L.; Helman, W.P.; Ross, A.B. Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals ( " OH/ " O $ in aqueous solution. Journal of physical and chemical reference data 1988, 17 (2), 513 886. [98] GŒrdfeldt, K.; Sommar, J.; Stršmberg, D.; Feng, X. Oxidation of atomic mercury by hydroxyl radicals and photoinduced decomposition of methylmercury i n the aqueous phase. Atmos. Environ. 2001, 35 (17), 3039 3047. [99] Mack, J. and Bolton, J.R. Photochemistry of nitrite and nitrate in aqueous solution: a review. J. Photochem. Photobiol. A. 1999, 128 (1), 1 13. [100] Mostofa, K.M.; Liu, C.; Sakugawa, H.; Vione, D.; Minakata, D.; Wu, F. Photoinduced and Microbial Generation of Hydrogen Peroxide and Organic Peroxides in Natural Waters, In Photobio geochemistry of Organic Matter, Springer: 2013; pp. 139 207. [101] Attri, P.; Kim, Y.H.; Park, D.H.; Park, J.H.; Hong, Y.J.; Uhm, H.S.; Kim, K.; Fridman, A.; Choi, E.H. Generation mechanism of hydroxyl radical species and its lifetime prediction during the plasma initiated ultraviolet (UV) photolysis. Scientific reports 2015, 5 . [102] Kunkely, H.; Horv‡th, O.; Vogler , A. Photophysics and photochemistry of mercury complexes. Coord. Chem. Rev. 1997, 159 , 85 93. [103] Nriagu, J.O. Mechanistic steps in the photoreduction of mercury in natural waters. Sci. Total Environ. 1994, 154 (1), 1 8. [104] Garcia, E.; Amyot, M.; Ari ya, P.A. Relationship between DOC photochemistry and mercury redox transformations in temperate lakes and wetlands. Geochim. Cosmochim. Acta 2005, 69 (8), 1917 1924. [105] Hahne, H. and Kroontje, W. The simultaneous effect of pH and chloride concentrations upon mercury (II) as a pollutant. Soil Sci. Soc. Am. J. 1973, 37 (6), 838 843. [106] Takeda, K.; Takedoi, H.; Yamaji, S.; Ohta, K.; Sakugawa, H. Determination of hydroxyl radical photoproduction rates in natural waters. Anal. Sci. 2004, 20 (1), 153 158.

PAGE 190

190 [ 107] Vaughan, P.P. and Blough, N.V. Photochemical formation of hydroxyl radical by constituents of natural waters. Environ. Sci. Technol. 1998, 32 (19), 2947 2953. [108] Vione, D.; Minella, M.; Maurino, V.; Minero, C. Indirect photochemistry in sunlit surf ace waters: Photoinduced production of reactive transient species. Chemistry A European Journal 2014, 20 (34), 10590 10606. [109] Zepp, R.G.; Hoigne, J.; Bader, H. Nitrate induced photooxidation of trace organic chemicals in water. Environ. Sci. Technol. 1 987, 21 (5), 443 450. [110] Zhang, Y.; Sun, R.; Ma, M.; Wang, D. Study of inhibition mechanism of NO3 on photoreduction of Hg (II) in artificial water. Chemosphere 2012, 87 (2), 171 176. [111] Steudel, R. Mechanism for the formation of elemental sulfur fr om aqueous sulfide in chemical and microbiological desulfurization processes. Ind Eng Chem Res 1996, 35 (4), 1417 1423. [112] Linkous, C.A.; Huang, C.; Fowler, J.R. UV photochemical oxidation of aqueous sodium sulfide to produce hydrogen and sulfur. J. Pho tochem. Photobiol. A. 2004, 168 (3), 153 160. [113] Pascal, I. and Tarbell, D.S. The kinetics of the oxidation of a mercaptan to the corresponding disulfide by aqueous hydrogen peroxide. J. Am. Chem. Soc. 1957, 79 (22), 6015 6020. [114] Watts, M.J. and Lin den, K.G. Chlorine photolysis and subsequent OH radical production during UV treatment of chlorinated water. Water Res. 2007, 41 (13), 2871 2878. [115] Feng, Y.; Smith, D.W.; Bolton, J.R. Photolysis of aqueous free chlorine species (HOCl and OCl) with 254 nm ultraviolet light. Journal of Environmental Engineering and Science 2007, 6 (3), 277 284. [116] De, A.K.; Chaudhuri, B.; Bhattacharjee, S.; Dutta, B.K. Estimation of " OH radical reaction rate constants for phenol and chlorinated phenols using UV/H 2 O 2 photo oxidation. J. Hazard. Mater. 1999, 64 (1), 91 104. [117] Yamamoto, M. Stimulation of elemental mercury oxidation in the presence of chloride ion in aquatic environments. Chemosphere 1996, 32 (6), 1217 1224. [118] Sun, R.; Wang, D.; Mao, W.; Zhao, S. ; Zhang, C. Roles of chloride ion in photo reduction/oxidation of mercury. Chinese Science Bulletin 2014, 59 (27), 3390 3397. [119] Nazhat, N. and Asmus, K. Reduction of mercuric chloride by hydrated electrons and reducing radicals in aqueous solutions. Fo rmation and reactions of mercury chloride (HgCl). J. Phys. Chem. 1973, 77 (5), 614 620. [120] Stephens, C.R.; Shepson, P.B.; Steffen, A.; Bottenheim, J.W.; Liao, J.; Huey, L.G.; Apel, E.; Weinheimer, A.; Hall, S.R.; Cantrell, C. The relative importance of chlorine

PAGE 191

191 and bromine radicals in the oxidation of atmospheric mercury at Barrow, Alaska. Journal of Geophysical Research: Atmospheres (1984 Ð 2012) 2012, 117 (D14). [121] Ariya, P.A.; Khalizov, A.; Gidas, A. Reactions of gaseous mercury with atomic and molec ular halogens: kinetics, product studies, and atmospheric implications. The Journal of Physical Chemistry A 2002, 106 (32), 7310 7320. [122] Qureshi, A.; O'Driscoll, N.J.; MacLeod, M.; Neuhold, Y.; HungerbŸhler, K. Photoreactions of mercury in surface oce an water: gross reaction kinetics and possible pathways. Environ. Sci. Technol. 2009, 44 (2), 644 649. [123] Horv‡th, O.; Ford, P.; Vogler, A. Photocatalytic self generation. Mercury (II) reduction via photochemical reactions of the dimercury (I) cation, H g22. Inorg. Chem. 1993, 32 (12), 2614 2615. [124] Horv‡th, O. and Mik—, I. Photoredox chemistry of mercury ions in aqueous ethanol solutions. J. Photochem. Photobiol. A. 1999, 128 (1), 33 38. [125] Canonica, S. and Freiburghaus, M. Electron rich phenols fo r probing the photochemical reactivity of freshwaters. Environ. Sci. Technol. 2001, 35 (4), 690 695. [126] Canonica, S.; Jans, U.; Stemmler, K.; Hoigne, J. Transformation kinetics of phenols in water: photosensitization by dissolved natural organic materia l and aromatic ketones. Environ. Sci. Technol. 1995, 29 (7), 1822 1831. [127] Lipski, M.; S3awi4ski, J.; Zych, D. Changes in the luminescent properties of humic acids induced by UV radiation. J. Fluoresc. 1999, 9 (2), 133 138. [128] Lee, E.; Glover, C.M.; Rosario Ortiz, F.L. Photochemical formation of hydroxyl radical from eff luent organic matter: role of composition. Environ. Sci. Technol. 2013, 47 (21), 12073 12080. [129] McKay, G.; Kleinman, J.L.; Johnston, K.M.; Dong, M.M.; Rosario Ortiz, F.L.; Mezyk, S.P. Kinetics of the reaction between the hydroxyl radical and organic ma tter standards from the International Humic Substance Society. Journal of soils and sediments 2014, 14 (2), 298 304. [130] Mostofa, K.M.; Liu, C.; Sakugawa, H.; Vione, D.; Minakata, D.; Saquib, M.; Mottaleb, M.A. Photoinduced generation of hydroxyl radical in natural waters, In Photobiogeochemistry of Organic Matter, Springer: 2013; pp. 209 272. [131] Rubio, D.; Nebot, E.; Casanueva, J.; Pulgarin, C. Comparative effect of simulated solar light, UV, UV/H 2 O 2 and photo Fenton treatment (UV Ð Vis/H 2 O 2/Fe 2 , 3 ) in the Escherichia coli inactivation in artificial seawater. Water Res. 2013, 47 (16), 6367 6379. [132] Vione, D.; Falletti, G.; Maurino, V.; Minero, C.; Pelizzetti, E.; Malandrino, M.; Ajassa, R.; Olariu, R.; Arsene, C. Sources and sinks of hydroxyl radicals upon irradiation of natural water samples. Environ. Sci. Technol. 2006, 40 (12), 3775 3781.

PAGE 192

192 [133] Sulzberger, B. and Durisch Kaiser, E. Chemical characterization of dissolved organic matter (DOM): a prerequisite for understanding UV induced chang es of DOM absorption properties and bioavailability. Aquat. Sci. 2009, 71 (2), 104 126. [134] Si, L. and Ariya, P.A. Photochemical reactions of divalent mercury with thioglycolic acid: Formation of mercuric sulfide particles. Chemosphere 2015, 119 , 467 472 . [135] Zheng, W.; Liang, L.; Gu, B. Mercury reduction and oxidation by reduced natural organic matter in anoxic environments. Environ. Sci. Technol. 2011, 46 (1), 292 299. [136] Ababneh, F.A.; Scott, S.L.; Al Reasi, H.A.; Lean, D.R. Photochemical reductio n and reoxidation of aqueous mercuric chloride in the presence of ferrioxalate and air. Sci. Total Environ. 2006, 367 (2), 831 839. [137] Aeschbacher, M.; Graf, C.; Schwarzenbach, R.P.; Sander, M. Antioxidant properties of humic substances. Environ. Sci. T echnol. 2012, 46 (9), 4916 4925. [138] Lester, Y.; Sharpless, C.M.; Mamane, H.; Linden, K.G. Production of photo oxidants by dissolved organic matter during UV water treatment. Environ. Sci. Technol. 2013, 47 (20), 11726 11733. [139] Miller, C.L. The role of organic matter in the dissolved phase speciation and solid phase partitioning of mercury. . PhD Dissertation, University of Mary land, 2006. [140] Loux, N.T. An assessment of mercury species dependent binding with natural organic carbon. Chemical Speciat ion & Bioavailability 1998, 10 (4), 127 136. [141] Barnes, H.; Romberger, S.; Stemprok, M. Ore solution chemistry Part 2: Solubility of HgS in sulfide solutions. Economic Geology 1967, 62 (7), 957 982. [142] Cline, J.D. Spectrophomet r ic Determination of Hy drogen Sulfide in Natural Waters. Limnol. Oceanogr. 1969, 14 (3), 454 458. [143] Frederick, J.; Snell, H.; Haywood, E. Solar ultraviolet radiation at the earth's surface. Photochem. Photobiol. 1989, 50 (4), 443 450. [144] Rozema, J.; Bjšrn, L.O.; Bornman, J.; Gaber56 ik, A.; HŠder, D.; Tro5 t, T.; Germ, M.; Klisch, M.; Gršniger, A.; Sinha, R. The role of UV B radiation in aquatic and terrestrial ecosystems Ñ an experimental and functional analysis of the evolution of UV absorbing compounds. Journal of Photochem istry and Photobiology B: Biology 2002, 66 (1), 2 12. [145] Hatchard, C. and Parker, C.A. A new sensitive chemical actinometer. II. Potassium ferrioxalate as a standard chemical actinometer. 1956, 235 (1203), 518 536. [146] Chen, W.; Westerhoff, P.; Leenhe er, J.A.; Booksh, K. Fluorescence excitation emission matrix regional integration to quantify spectra for dissolved organic matter. Environ. Sci. Technol. 2003, 37 (24), 5701 5710.

PAGE 193

193 [147] Dalrymple, R.M.; Carfagno, A.K.; Sharpless, C.M. Correlations between dissolved organic matter optical properties and quantum yields of singlet oxygen and hydrogen peroxide. Environ. Sci. Technol. 2010, 44 (15), 5824 5829. [148] Sharpless, C.M.; Aeschbacher, M.; Page, S.E.; Wenk, J.; Sander, M.; McNeill, K. Photooxidation i nduced changes in optical, electrochemical, and photochemical properties of humic substances. Environ. Sci. Technol. 2014, 48 (5), 2688 2696. [149] Determination of mercury in water by cold vapor atomic absorption spectrometry; Method 245.1, Rev. 3.0. U.S. EPA National Exposure Research Laboratory: Methods for the determination of metals in environmental samples (EPA/600/R 94/111). Retrieved from http://www.epa.gov/nerl/ [150] Mercury in Water by Oxidatio n, Purge and Trap, and Cold Vapor Atomic Fluorescence Spectrometry; EPA Method 1631, Rev. E. U.S. EPA Office of Water (EPA 821 R 02 019). Retrieved from http://www.epa.gov/nerl/ [151] Mercury in Water by Cold Vapor Atomic Fluorescence Spectrometry; Method 245.7, Rev. 2.0. U.S. EPA Office of Research and Development (EPA 821 R 05 001). Retrieved from http://www.nemi.gov/ [152] Bloom, N.S. and Crecelius, E.A . Determination of mercury in seawater at sub nanogram per liter levels. Mar. Chem. 1983, 14 (1), 49 59. [153] Carpi, A. and Lindberg, S.E. Sunlight mediated emission of elemental mercury from soil amended with municipal sewage sludge. Environ. Sci. Techno l. 1997, 31 (7), 2085 2091. [154] Loux, N.T. A critical assessment of elemental mercury air/water exchange parameters. Chemical Speciation & Bioavailability 2004, 16 (4), 127 138. [155] O'Brien, T.F.; Bommaraju, T.V.; Hine, F. Handbook of Chlor Alkali Tech nology: Volume II: Brine Treatment and Cell Operation. Springer Science & Business Media: 2007. [156] Neghab, M.; Amin Norouzi, M.; Choobineh, A.; Reza Kardaniyan, M.; Hassan Zadeh, J. Health effects associated with long term occupational exposure of emplo yees of a chlor alkali plant to mercury. Int. J. Occup. Saf. Ergonomics 2012, 18 (1), 97 106. [157] Piikivi, L. and Tolonen, U. EEG findings in chlor alkali workers subjected to low long term exposure to mercury vapour. Br. J. Ind. Med. 1989, 46 (6), 370 3 75. [158] Bluhm, R.E.; Bobbitt, R.G.; Welch, L.W.; Wood, A.J.; Bonfiglio, J.F.; Sarzen, C.; Heath, A.J.; Branch, R.A. Elemental mercury vapour toxicity, treatment, and prognosis after acute, intensive exposure in chlor alkali plant workers. Part I: History , neuropsychological findings and chelator effects. Hum. Exp. Toxicol. 1992, 11 (3), 201 210.

PAGE 194

194 [159] Smith, R.; Vorwald, A.; Patil, L.; Mooney, T. Effects of exposure to mercury in the manufacture of chlorine. The American Industrial Hygiene Association Jou rnal 1970, 31 (6), 687 700. [160] Lodenius, M. Dry and wet deposition of mercury near a chlor alkali plant. Sci. Total Environ. 1998, 213 (1), 53 56. [161] Southworth, G.; Lindberg, S.; Zhang, H.; Anscombe, F. Fugitive mercury emissions from a chlor alkali factory: sources and fluxes to the atmosphere. Atmos. Environ. 2004, 38 (4), 597 611. [162] Rice, G. E., et al. Mercury study report to Congress. Volume 3. Fate and transport of mercury in the environment . No. PB -98 124753/XAB; EPA -452/R 97/005. U.S.EPA Office of Air Quality Planning and Standards, 1997. [163] Ullrich, S.M.; Ilyushchenko, M.A.; Tanton, T.W.; Uskov, G.A. Mercury contamination in the vicinity of a derelict chlor alkali plant: part II: contamination of the aquatic and terrestrial food chain and potential risks to the local population. Sci. Total Environ. 2007, 381 (1), 290 306. [164] Capsule Report, Aqueous Mercury Treatment. U . S. Environmental Protection Agency, Office of Research and Development; EPA 625 R 97 004. Retrieved from http:/nepi s.epa.gov/ . [165] Otto, M. and Bajpai, S. Treatment technologies for mer cury in soil, waste, and water. Remediation Journal 2007, 18 (1), 21 28. Retrieved from http://clu in.org/542R07003. [166] Br avo, A.G.; Cosio, C.; Amouroux, D.; Zopfi, J.; Chevalley, P.; Spangenberg, J.E.; Ungureanu, V.; Dominik, J. Extremely elevated methyl mercury levels in water, sediment and organisms in a Romanian reservoir affected by release of mercury from a chlor alkali plant. Water Res. 2014, 49 , 391 405. [167] Hildebrand, D. The differentiation of mercury speciation in sub microgram quantities i n aqueous solution . PhD Dissertation, University Ke ntucky, 1972. [168] Chlorine Institute Inc. Guidelines for the Optimization of Mercury Wastewater Treatment (Sulfide Precipitation Process) Systems, Ed. 1, December 2003 (Internal Report). Retrieved from personal communications. 2003 . [169] Pham, A.L.; Morris, A.; Zhang, T.; Ticknor, J.; Levard, C.; Hsu Kim, H. Precipitation of n anoscale mercuric sulfides in the presence of natural organic matter: Structural properties, aggregation, and biotransformation. Geochim. Cosmochim. Acta 2014, 133 , 204 215. [170] Paquette, K.E. and Helz, G.R. Inorganic speciation of mercury in sulfidic wa ters: the importance of zero valent sulfur. Environ. Sci. Technol. 1997, 31 (7), 2148 2153.

PAGE 195

195 [171] Slowey, A.J. Rate of formation and dissolution of mercury sulfide nanoparticles: The dual role of natural organic matter. Geochim. Cosmochim. Acta 2010, 74 (1 6), 4693 4708. [172] Benoit, J.; Gilmour, C.; Heyes, A.; Mason, R.; Miller, C. Geochemical and biological controls over methylmercury production and degradation in aquatic ecosystems. 2003, 835 , 262 297. [173] Whalin, L.; Kim, E.; Mason, R. Factors influen cing the oxidation, reduction, methylation and demethylation of mercury species in coastal waters. Mar. Chem. 2007, 107 (3), 278 294. [174] Sampling Ambient Water for Trace Metals at EPA Water Quality Criteria Levels; Method 1669. U.S.EPA Office of Science and Technology. Retrieved from http://www.epa.gov/nerl/ . [175] Tossell, J. Theoretical studies on the formation of mercury complexes in solution and the dissolution and reactions of cinnabar. Am. Minera l. 1999, 84 , 877 883. [176] Twidwell, L. and Thompson, R. Recovering and recycling Hg from chlor alkali plant wastewater sludge. JOM Journal of the Minerals, Metals and Materials Society 2001, 53 (1), 15 17. [177] MINTEQA2/PRODEFA2, A Geochemical Assessmen t Model for Environmental Systems: Version 3.0 User's Manual, 106 p. ; Environmental Research Laboratory: Athens, GA. [178] Myneni, S.C.; Brown, J.T.; Martinez, G.A.; Meyer Ilse, W. Imaging of humic substance macromolecular structures in water and soils. Sc ience 1999, 286 (5443), 1335 1337; 7989 [pii]. [179] Andersson, M.E.; GŒrdfeldt, K.; WŠngberg, I.; Stršmberg, D. Determination of Henry's law constant for elemental mercury. Chemosphere 2008, 73 (4), 587 592. [180] Okouchi, S. and Sasaki, S. Photochemical behavior of mercury ore in water. Environ. Int. 1983, 9 (2), 103 106. [181] Zellner, R.; Exner, M.; Herrmann, H. Absolute OH quantum yields in the laser photolysis of nitrate, nitrite and dissolved H2O2 at 308 and 351 nm in the temperature range 278 Ð 353 K. J. Atmos. Chem. 1990, 10 (4), 411 425. [182] McKay, G. and Rosario Ortiz, F.L. Temperature Dependence of the Photochemical Formation of Hydroxyl Radical from Dissolved Organic Matter. Environ. Sci. Technol. 2015, 49 (7), 4147 4154. [183] Maie, N.; Scully, N.M.; Pisani, O.; JaffŽ, R. Composition of a protein like fluorophore of dissolved organic matter in coastal wetland and estuarine ecosystems. Water Res. 2007, 41 (3), 563 570.

PAGE 196

196 [184] Miller, C.L.; Liang, L.; Gu, B. Competitive ligand exchange reveals time dependant changes in the reactivity of Hg Ð dissolved organic matter complexes. Environmental Chemistry 2012 . [185] Si, L. and Ariya, P.A. Reduction of Oxidized Mercury Species by Dicarboxylic Acids (C2$ C4): Kinetic and Product Studies. Environ. Sci. Techn ol. 2008, 42 (14), 5150 5155. [186] Tennakone, K. and Wickramanayake, S. Photocatalytic properties of mercury (I) chloride and photogeneration of oxygen from water. Solar energy materials 1987, 15 (1), 59 63. [187] Pernet Coudrier, B.; Clouzot, L.; Varraul t, G.; Tusseau Vuillemin, M.; Verger, A.; Mouchel, J. Dissolved organic matter from treated effluent of a major wastewater treatment plant: characterization and influence on copper toxicity. Chemosphere 2008, 73 (4), 593 599. [188] Muresan, B.; Pernet Coud rier, B.; Cossa, D.; Varrault, G. Measurement and modeling of mercury complexation by dissolved organic matter isolates from freshwater and effluents of a major wastewater treatment plant. Appl. Geochem. 2011, 26 (12), 2057 2063. [189] Sarathy, V. and Alle n, H.E. Copper complexation by dissolved organic matter from surface water and wastewater effluent. Ecotoxicol. Environ. Saf. 2005, 61 (3), 337 344. [190] Pal, B.; Ikeda, S.; Ohtani, B. Photoinduced chemical reactions on natural single crystals and synthes ized crystallites of mercury (II) sulfide in aqueous solution containing naturally occurring amino acids. Inorg. Chem. 2003, 42 (5), 1518 1524. [191] Aeschbacher, M.; Vergari, D.; Schwarzenbach, R.P.; Sander, M. Electrochemical analysis of proton and elect ron transfer equilibria of the reducible moieties in humic acids. Environ. Sci. Technol. 2011, 45 (19), 8385 8394. [192] Lu, Q.; Yuan, Y.; Tao, Y.; Tang, J. Environmental pH and ionic strength influence the electron transfer capacity of dissolved organic m atter. Journal of Soils and Sediments 2015 1 8. [193] Alvarez Puebla, R. and Garrido, J. Effect of pH on the aggregation of a gray humic acid in colloidal and solid states. Chemosphere 2005, 59 (5), 659 667. [194] Aeschbacher, M.; Sander, M.; Schwarzenbach , R.P. Novel electrochemical approach to assess the redox properties of humic substances. Environ. Sci. Technol. 2009, 44 (1), 87 93. [195] Arons, A. and Peppard, M. Einstein's Proposal of the Photon Concept a Translation of the Annalen der Physik Paper of 1905. American Journal of Physics 1965, 33 (5), 367 374.

PAGE 197

197 [196] Paul, A.; Stšsser, R.; Zehl, A.; Zwirnmann, E.; Vogt, R.D.; Steinberg, C.E. Nature and abundance of organic radicals in natural organic matter: Effect of pH and irradiation. Environ. Sci. Tech nol. 2006, 40 (19), 5897 5903. [197] Giggenbach, W. Optical spectra of highly alkaline sulfide solutions and the second dissociation constant of hydrogen sulfide. Inorg. Chem. 1971, 10 (7), 1333 1338. [198] Brandon, N.; Francis, P.; Jeffrey, J.; Kelsall, G .; Yin, Q. Thermodynamics and electrochemical behaviour of Hg S Cl H 2 O systems. J Electroanal Chem 2001, 497 (1), 18 32. [199] Anaf, W.; Janssens, K.; De Wael, K. Formation of Metallic Mercury During Photodegradation/Photodarkening of & ! HgS: Electrochemica l Evidence. Angewandte Chemie 2013, 125 (48), 12800 12803. [200] Da Pieve, F.; Stankowski, M.; Hogan, C. Electronic structure calculations of mercury mobilization from mineral phases and photocatalytic removal from water and the atmosphere. Sci. Total Envi ron. 2014, 493 , 596 605. [201] Lindsay, W.L. Chemical equilibria in soils. John Wiley and Sons Ltd.: 1979. [202] He, F.; Zhao, W.; Liang, L.; Gu, B. Photochemical Oxidation of Dissolved Elemental Mercury by Carbonate Radicals in Water. Environmental Scienc e & Technology Letters 2014, 1 (12), 499 503. [203] Hsieh, Y.; Tokunaga, S.; Huang, C. Some ch emical reactions at the HgS (s) water interface as affected by photoirradiation. Colloids and Surfaces 1991, 53 (2), 257 274. [204] Hoffmann, M.R. Kinetics and me chanism of oxidation of hydrogen sulfide by hydrogen peroxide in acidic solution. Environ. Sci. Technol. 1977, 11 (1), 61 66. [205] Scaldaferri, M.C.L. and Pimentel, A.S. Theoretical study of the reaction of hydrogen sulfide with nitrate radical. Chemical Physics Letters 2009, 470 (4), 203 209. [206] Wang, Y.; Li, Y.; Liu, G.; Wang, D.; Jiang, G.; Cai, Y. Elemental Mercury in Natural Waters: Occurrence and Determination of Particulate Hg (0). Environ. Sci. Technol. 2015, 49 (16), 9742 9749. [207] Matthiesse n, A. Determining the redox capacity of humic substances as a function of pH. Vom Wasser 1995, 84 , 229 235. [208] Dong, W.; Liang, L.; Brooks, S.; Southworth, G.; Gu, B. Roles of dissolved organic matter in the speciation of mercury and methylmercury in a contaminated ecosystem in Oak Ridge, Tennessee. Environmental Chemistry 2010, 7 (1), 94 102. [209] Gerbig, C.A.; Kim, C.S.; Stegemeier, J.P.; Ryan, J.N.; Aiken, G.R. Formation of nanocolloidal metacinnabar in mercury DOM sulfide systems. Environ. Sci. Tech nol. 2011, 45 (21), 9180 9187.

PAGE 198

198 [210] Tarbell, D. The mechanism of oxidation of thiols to disulfides. Organic sulfur compounds 1961, 1 , 97 102. [211] Frimmel, F.; Immerz, A.; Niedermann, H. Complexation capacities of humic substances isolated from freshwater with respect to copper (II), mercury (II), and iron (II, III), In Complexation of trace metals in natural waters, Springer: 1984; pp. 329 343. [21 2] Jalilehvand, F.; Leung, B.O.; Izadifard, M.; Damian, E. Mercury (II) cysteine complexes in alkaline aqueous solution. Inorg. Chem. 2006, 45 (1), 66 73. [213] Mah, V. and Jalilehvand, F. Mercury (II) complex formation with glutathione in alkaline aqueous solution. JBIC Journal of Biological Inorganic Chemistry 2008, 13 (4), 541 553. [214] Rao, B.; Simpson, C.; Lin, H.; Liang, L.; Gu, B. Determination of thiol functional groups on bacteria and natural organic matter in environmental systems. Talanta 2014, 119 , 240 247. [215] Xiao, Z.; Munthe, J.; Stršmberg, D.; Lindqvist, O. Photochemical behaviour of inorganic mercury compounds in aqueous solution. Mercury Pollution; Integration and Synthesis, CJ Watras and JW Huckabee (Eds.) 1994 . [216] Hsu Kim, H. and Se dlak, D.L. Similarities between inorganic sulfide and the strong Hg (II) complexing ligands in municipal wastewater effluent. Environ. Sci. Technol. 2005, 39 (11), 4035 4041. [217] Yamamoto, M. Possible mechanism of elemental mercury oxidation in the prese nce of SH compounds in aqueous solution. Chemosphere 1995, 31 (2), 2791 2798. [218] Jocelyn, P.C. Chemical reduction of disulfides. Methods Enzymol. 1987, 143 , 246 256; 0076 6879(87)43048 6 [pii]. [219] Senesi, N. Molecular and quantitative aspects of the chemistry of fulvic acid and its interactions with metal ions and organic chemicals: Part II. The fluorescence spectroscopy approach. Anal. Chim. Acta 1990, 232 , 77 106. [220] Kieber, D.J.; Miller, G.W.; Neale, P.J.; Mopper, K. Wavelength and temperature d ependent apparent quantum yields for photochemical formation of hydrogen peroxide in seawater. Environmental Science: Processes & Impacts 2014, 16 (4), 777 791. [221] Grannas, A.M.; Martin, C.B.; Chin, Y.; Platz, M. Hydroxyl radical production from irradia ted Arctic dissolved organic matter. Biogeochemistry 2006, 78 (1), 51 66. [222] Sikorski, K.A. Investigating the mechanism of phenol photooxidation by chromopho ric dissolved organic matter. PhD Dissertation, University of Maryland , 2014. [223] Faust, B.C. and Hoigne, J. Sensitized photooxidation of phenols by fulvic acid and in natural waters. Environ. Sci. Technol. 1987, 21 (10), 957 964.

PAGE 199

199 [224]Golanoski, K.S.; Fang, S.; Del Vecchio, R.; Blough, N.V. Investigating the mechanism of phenol photooxidation by humic substances. Environ. Sci. Technol. 2012, 46 (7), 39123920. [225]Aiken, G.; McKnight, D.; Thorn, K.; Thurman, E. Isolation of hydrophilic organic acids from water using nonionic macroporous resins. Org. Geochem. 1992, 18 (4), 567 573. [226]Leenheer , J.A. Comprehensive approach to preparative isolation and fractionation ofdissolved organic carbon from natural waters and wastewaters. Environ. Sci. Technol. 1981, 15 (5), 578 587. [227]Marhaba, T.F.; Van, D.; Lippincott, R.L. Changes in NOM fractionat ion through treatment: A comparison of ozonation and chlorination. 2000. [228]Barber, L.B.; Leenheer, J.A.; Noyes, T.I.; Stiles, E.A. Nature and transformation ofdissolved organic matter in treatment wetlands. Environ. Sci. Technol. 2001, 35 (24), 48054 816. [229]Ritchie, J.D. and Perdue, E.M. Protonbinding study of standard and reference fulvic acids, humic acids, and natural organic matter. Geochim. Cosmochim. Acta 2003, 67 (1), 85 96. [230]Thorn, K.; Folan, D.; MacCarthy, P. Characteri zation of the IHSS standard andreference fulvic and humic acids by solution state 13 C and 1 H NMR spectrometry. US Geological Survey, Water Resources Investigations Report 1989 894196.

PAGE 200

200 BIOGRAPHICAL SKETCH Ana Maria Rogers began her career in water science and engi neering with a Bachelor of Science in civil and environmental engineering at the University of Rhode Island with Spanish and French languages and completed the international engineering program. After spending the summer conducting hydro -geologic research at the University of Puerto Rico in Mayaguez, she obtained a Master of Science in civil and water resources engineering at the Georgia Institute of Technology. During her master’s program, she worked in hydrologic modeling and forecasting as a student res earch fellow at the Southeast River Forecasting Center of the National Oceanic and Atmospheric Agency. Upon completing the master’s degree she continued hydrologic modeling work as a research fellow at the South Florida Water Management District during the summer before beginning doctoral study at the University of Florida under Professor David Mazyck. Following experimental research and lab studies, she sought to use the knowledge gained and apply it to a Fulbright fellowship project in Ulaanbaatar, Mong olia, that aimed to improve human health assessments for environmental and occupational mercury exposure from small scale gold mining. While living there for one year, she worked with a team in the Toxicology Lab at the Public Health Institute on the impl ementation of aqueous mercury environmental analysis and waste monitoring for remediation strategy development; u pon returning to the University of Florida, she completed the Ph.D. degree in environmental engineering sciences. Ana Maria’s career ambitions are to work in industry and to teach in academia with a focus on addressing treatment technology development challenges from heavy metals contamination in water and wastewater, as well as its applications in public health.