Transport and Environmental Applications of Carbon Nanotubes in Porous Media

MISSING IMAGE

Material Information

Title:
Transport and Environmental Applications of Carbon Nanotubes in Porous Media
Physical Description:
1 online resource (226 p.)
Language:
english
Creator:
Tian, Yuan
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Agricultural and Biological Engineering
Committee Chair:
Gao, Bin
Committee Co-Chair:
Kiker, Gregory
Committee Members:
Bonzongo, Jean-Claud
Munoz-Carpena, Rafael
Ziegler, Kirk

Subjects

Subjects / Keywords:
attachment -- breakthrough -- colloids -- dlvo -- metals -- nanomaterials -- sulfonamide -- treatment -- wastewater
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre:
Agricultural and Biological Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
There are increasing concerns over the environmental impact and health risks of carbon nanotubes (CNTs) because they may be released into the environment, such as soil and groundwater systems. This Ph.D. dissertation systematically investigated the fate and transport of CNTs in porous media as well as the applications of CNTs in wastewater treatment to remove heavy metal and antibiotic pollutants.   Laboratory columns packed with quartz sand and glass beads showed that deposition of CNTs in saturated porous media was very sensitive to solution chemistry. More CNTs were found to be trapped in the primary minimum. Under unfavorable conditions, weak associated CNTs in the secondary minimum may be transferred into the primary minimum due to the effect of hydrodynamic force and/or local favorable sites associated with surface heterogeneity.   Dispersion/solubilization methods showed great influence on the stability and mobility of CNTs. Humic acid dispersion granted higher mobility to CNTs compared with other surface modification methods. Reductions in moisture contents showed greater influence on the retention of surface oxidized CNTs. This is possibly due to CNT deposition on the air-water interface and through film straining. Combined mechanisms could be responsible for the retention and transport of CNTs in unsaturated porous media.   Retention and transport of functionalized CNTs in natural sand porous media were mainly controlled by strong surface deposition through the electrostatic and/or hydrogen-bonding attractions between surface function groups of the CNTs and metal oxyhydroxide impurities on the sand surfaces.   Fixed-bed columns packed with functionalized multi-walled CNTs and natural sand were used in laboratory to evaluate and optimize the applications of CNTs in removing heavy metals and antibiotics from water. They significantly improved the fixed-bed’s filtering efficiency of Pb2+ and Cu2+ by 55%-75% and 31%-57%, respectively. The fixed bed column experiments also showed that pH could be a key factor that affects the removal of antibiotics by controlling the protonation of antibiotics and surface charge of the carboxyl and hydroxyl functional groups on CNT surfaces. The column removal efficiency of antibiotics decreased only slightly after regenerations, suggesting the CNT-sand columns can be efficiently used and regenerated to remove contaminants from water.
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.
Statement of Responsibility:
by Yuan Tian.
Thesis:
Thesis (Ph.D.)--University of Florida, 2012.
Local:
Adviser: Gao, Bin.
Local:
Co-adviser: Kiker, Gregory.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044405:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 TRANSPORT AND ENVIRONMENTAL APPLICATION S OF CARBON NANOTUB ES IN POROUS MEDIA By YU AN TIAN 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 201 2

PAGE 2

2 201 2 Yu an Tian

PAGE 3

3 To my family

PAGE 4

4 ACKNOWLEDGMENTS I express my sincerest gratitude to my advisor, Dr. Bin Gao for his support, guidance, inspiration and humor throughout my graduate study at the University of Florida. I also thank the other members of my committee, Dr. Gregory A. Kiker Dr. Rafael Muoz Carpena Dr. Kirk J. Ziegler, and Dr. Jean Claude Bonzongo, for their valuable advice, help, and support in the past several years. I truly appreciate that Prof. Xiaorong Wang at Nanjing University and Prof. Xiaohua Lu at Huazhong University of Science and Technology are encouraging me towards my academic life. I thank Dr Yu Wang, Dr. Lena Q Ma Dr. Shiny Methews and Dr. Carlos Silvera Batista for their valuable advice, help, and support in my research. I extend my gratitude to my colleagues ( Dr. Huimin Sun, Dr. Congrong Yu and Lei Wu ) in the Environmental N anotechnology Research G roup for their help. I would like to acknowledge Steven Feagle, Paul Lane, and Orlando Lanni for their lab support. I am grateful for the facilities and support from the Major Analytical Instrumentation Center and the Particle Engin eering Research Center at the University of Florida. Special thanks go to my wonderful gran dparents and parents for their tremendous love and support in my life. Without them, I could not have achieved this goal.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 11 LIST OF ABBREVIATI ONS ................................ ................................ ........................... 14 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 18 Background and Problem Statement ................................ ................................ ...... 18 Research Objectives ................................ ................................ ............................... 20 Objective 1: Fate and Transport of C NT s in Saturated Clean Porous Media ... 20 Objective 2 : Fate and Transport of C NT s in Unsaturated Clean Porous Media ................................ ................................ ................................ ............ 21 Objective 3 : Fate and Transport of CNTs in Natural Porous Media .................. 21 Objective 4 : Interaction of CNTs with Contaminants ................................ ........ 21 Organization of the Dissertation ................................ ................................ .............. 22 2 TRANSPORT OF ENGINEERED NANOPARTICLES IN SATURATED POROUS MEDIA ................................ ................................ ................................ ... 24 Introduct ory Remarks ................................ ................................ .............................. 24 Methods and Materials ................................ ................................ ............................ 26 CNTs ................................ ................................ ................................ ................ 26 AgNPs ................................ ................................ ................................ .............. 27 Colloids ................................ ................................ ................................ ............. 27 Porous Medium ................................ ................................ ................................ 28 Column Experiments ................................ ................................ ........................ 29 DLVO T heory ................................ ................................ ................................ ... 30 Transport Model ................................ ................................ ............................... 31 Results and Discussion ................................ ................................ ........................... 32 DLVO E nergy ................................ ................................ ................................ ... 32 Transport in Porous Media ................................ ................................ ............... 34 Mo del Simulations ................................ ................................ ............................ 37 Chapter Conclusions ................................ ................................ .............................. 37 3 EFFECT OF SOLUTION CHEMISTRY ON MULTI WALLED CARBON NANOTUBE DEPOSITION AND MOBILIZATION IN CLEAN POROUS MEDIA ... 43 Introduct ory Remarks ................................ ................................ .............................. 43 Material s a nd M ethods ................................ ................................ ............................ 45

PAGE 6

6 MWNTs ................................ ................................ ................................ ............ 45 Characterization of MWNTs ................................ ................................ ............. 45 Po rous Media ................................ ................................ ................................ ... 46 MWNT Stability ................................ ................................ ................................ 46 Column Experiments ................................ ................................ ........................ 47 Mathematical Model ................................ ................................ ......................... 48 Results and D iscussion ................................ ................................ ........................... 48 Characteristics of MWNTs and Por ous Media ................................ .................. 48 Deposition of MWNTs under DI Water Condition ................................ ............. 49 Effect of IS on MWNT Deposition ................................ ................................ ..... 51 Mobilization of MWNTs ................................ ................................ ..................... 52 Chapter Conclusions ................................ ................................ .............................. 54 4 HIGH MOBILITY OF SDBS DISPERSED SINGLE WALLED CARBON NANOTUBES IN SATURATED AND UNSATURATED POROUS MEDIA ............ 63 Introduct ory Remarks ................................ ................................ .............................. 63 Material s and Methods ................................ ................................ ............................ 67 SWNTs ................................ ................................ ................................ ............. 67 Porous Media ................................ ................................ ................................ ... 67 Sand column Expe riments ................................ ................................ ............... 68 Bubble column Experiments ................................ ................................ ............. 70 Results and Discussion ................................ ................................ ........................... 70 DLVO Energy ................................ ................................ ................................ ... 70 SWNT Transport i n Medium sand Columns ................................ ..................... 72 SWNT Transport in Fine sand Columns ................................ ........................... 73 SWNT Transport in Coarse sand Columns ................................ ...................... 74 SWNT Transport in Bubble Columns ................................ ............................... 74 Modeling SWNT Transport in the Sand Columns ................................ ............. 75 Chapter Conclusions ................................ ................................ .............................. 78 5 EFFECT OF SURFACE MODIFICATION ON SINGLE WALLED CARBON NANOTUBE RETENTION AND TRANSPORT IN GRANULAR POROUS MEDIA ................................ ................................ ................................ .................... 83 Introduct ory Remarks ................................ ................................ .............................. 83 Material s and M ethods ................................ ................................ ............................ 85 Surface Modified SWNTs ................................ ................................ ................. 85 Porous Media ................................ ................................ ................................ ... 86 Bubble Column Ex periment ................................ ................................ .............. 86 Sand Column Experiment ................................ ................................ ................. 87 Mathematical Model ................................ ................................ ......................... 88 Results and Discussion ................................ ................................ ........................... 88 Characteristics and Stability of Surface Modified SWNTs ................................ 88 Interaction between Surface Modified SWNTs and Air Bubbles ....................... 90 Transport of Surface Modified SWNTs in Medium Sand ................................ .. 90 Transport of Surface Modified SWNTs in Fine Sand ................................ ........ 91

PAGE 7

7 Transport of Surface Modified SWNTs in Coarse Sand ................................ ... 93 Model Simulations ................................ ................................ ............................ 94 Chapter Conclusions ................................ ................................ .............................. 94 6 DEPOSITION AND TRANSPORT OF FUNCTIONALIZED CARBON NANOTUBES IN WATER SATURATED SAND COLUMNS ............................... 104 Introduct ory Remarks ................................ ................................ ............................ 104 Materials and Methods ................................ ................................ .......................... 107 Functionalized of CNTs ................................ ................................ .................. 107 Porous Media ................................ ................................ ................................ 108 Packed Sand Column Preparation ................................ ................................ 109 CNT Transport in Natural, Baked, and Acid cleaned Sand Columns ............. 109 Distribution of Retained CNT in Natural and Baked Sand Columns ............... 110 Mobilization of Retained CNTs in Natural and Baked Sand Columns ............ 111 Effect of pH on CNT Transport in Natural and Baked Sand Columns ............ 111 Modeling ................................ ................................ ................................ ......... 112 Results and Discussion ................................ ................................ ......................... 112 Surf ace Characteristics ................................ ................................ .................. 112 Retention in the Acid cleaned Sand Columns ................................ ................ 114 Retention in the Natural and Baked Sand Columns ................................ ....... 115 Distribution of Retained CNTs in Natural and Baked Sand Columns ............. 117 Mobilization of Retained CNTs in Natural and Baked Sand Columns ............ 118 pH Effect ................................ ................................ ................................ ......... 118 Chapter Conclusions ................................ ................................ ............................ 120 7 METHODS OF USING CARBON NANOTUBES AS FILTER MEDIA TO REMOVE AQUEOUS HEAVY METALS ................................ ............................... 127 Introduct ory Remarks ................................ ................................ ............................ 127 Materials and Methods ................................ ................................ .......................... 129 CNTs ................................ ................................ ................................ .............. 129 Sand ................................ ................................ ................................ ............... 130 Heavy Metals ................................ ................................ ................................ .. 131 Batch Sorption Experiment ................................ ................................ ............. 131 Fixed bed Column Experiment ................................ ................................ ....... 131 Results and Discussion ................................ ................................ ......................... 133 CNT and Sand Properties ................................ ................................ .............. 133 Sorption Ability of the CNTs ................................ ................................ ........... 134 Single Metal Removal in the CNT Columns ................................ ................... 136 Dual Metal Removal in the CNT Columns ................................ ...................... 138 Chapter Conclusions ................................ ................................ ............................ 139 8 REMOVAL OF SULFAMETHOXAZOLE AND SULFAPYRIDINE BY CARBON NANOTUBES IN THE FIXED BED COLUMN ................................ ...................... 146 Introduct ory Remarks ................................ ................................ ............................ 146

PAGE 8

8 Materi als and Methods ................................ ................................ .......................... 149 Bed Materials and Conditions ................................ ................................ ......... 149 Antibiotics ................................ ................................ ................................ ....... 150 Fixed bed Column Experiments ................................ ................................ ..... 150 Column Regeneration and Recycling ................................ ............................. 152 Results and Discussion ................................ ................................ ......................... 153 Column Analyses ................................ ................................ ........................... 153 Effect of CN T Incorporation Method ................................ ............................... 154 Effect of pH ................................ ................................ ................................ ..... 154 Effect of Bed Depth ................................ ................................ ........................ 156 Effect of Adsorbent Dosage ................................ ................................ ............ 158 Effect of Adsorbate Initial Concentration ................................ ........................ 158 Effect of Flow Rate ................................ ................................ ......................... 159 Column Regeneration and Recycling ................................ ............................. 159 Chapter Conclusions ................................ ................................ ............................ 160 9 CONCLU SION ................................ ................................ ................................ ...... 173 Summary ................................ ................................ ................................ .............. 173 Recommendations for Future Work ................................ ................................ ...... 178 APPENDIX A SU PPORTING INFORMATION FOR CHAPTER 3 ................................ .............. 179 B SU PPORTING INFORMATION FOR CHAPTER 4 ................................ .............. 186 C SU PPORTING INFORMATION FOR CHAPTER 6 ................................ .............. 189 D SU PPORTING INFORMATION FOR CHAPTER 7 ................................ .............. 196 E SU PPORTING INFORMATION FOR CHAPTER 8 ................................ .............. 197 LIST OF REFERENCES ................................ ................................ ............................. 204 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 226

PAGE 9

9 LIST OF TABLES Table page 2 1 Physicochemical properties of nanoparticles, natural colloids, and porous medium ................................ ................................ ................................ ............... 39 2 2 Best fit parameter values and recovery rate calculations for nanoparticle and colloid transport in saturated porous media ................................ ........................ 40 3 1 Summary of conditions and procedures of the column experiments .................. 56 3 2 Physicochemical characteristics of MWNTs and porous media used in the column experiment. ................................ ................................ ............................ 57 3 3 Summary of mass recovery and model results of MWNT transport in saturated porous media with different combinations of solution pH, IS, and porous medium t ype and size. ................................ ................................ ............ 58 5 1 Surface properties of the SWNTs and the sand used in this work. ..................... 96 5 2 Summary of sand column experimental conditions and model parameters ....... 97 6 1 Surface characteristics of the functionalized CNTs and the porous media ....... 121 6 2 Summary of experimental conditions and model results of CNT transport in the sand columns. ................................ ................................ ............................ 122 7 1 Summary of the best fit Langmuir model parameters ................................ ....... 140 7 2 Summary of fixed bed column experimental results ................................ ......... 141 8 1 Chemical structures and properties of sulfamethoxazole (SMX) and sulfapyridine (SPY) ................................ ................................ ........................... 162 8 2 Summary of transport and adsorption parameters of the two antibiotics in CNT/sand columns ................................ ................................ ........................... 163 8 3 Definition and formula of the primary adsorption zone (PAZ) parameters used in this study ................................ ................................ ................................ ....... 164 C 1 PZCs of metal (hydro)oxides. ................................ ................................ ........... 195 E 1 Summary of transport and adsorption parameters of sulfamethoxazole (SMX) and sulfapyridine (SPY) in CNT/sand columns with different CNT incorporation methods ................................ ................................ ...................... 198 E 2 Summary of transport and adsorption parameters of sulfamethoxazole (SMX) and sulfapyridine (SPY) in CNT/sand columns under various pH conditions ... 199

PAGE 10

10 E 3 Summary of transport and adsorption parameters of sulfamethoxazole (SMX) and sulfapyridine (SPY) in CNT/sand columns under various bed depth conditions ................................ ................................ ................................ ......... 200 E 4 Summary of transport and adsorption parameters of sulfamethoxazole (SMX) and sulfapyridine (SPY) in CNT/sand columns under v arious adsorbent dosage conditions ................................ ................................ ............................. 201 E 5 Summary of transport and adsorption parameters of sulfamethoxazole (SMX) and sulfapyridine (SPY) in CNT/sand columns under various adsorbate initial concentration conditions ................................ ................................ ................... 202 E 6 Summary of transport and adsorption parameters of sulfamethoxazole (SMX) and sulfapyridine (SPY) in CNT/sand columns under various flow rate conditions ................................ ................................ ................................ ......... 203

PAGE 11

11 LIST OF FIGURES Figure page 2 1 DLVO energy between the porous medium and (a) silver nanoparticle, (b) carbon nanotube, (c) SDBS dispersed montmorillonite, and (d) water dispersed montmorillonite. ................................ ................................ .................. 41 2 2 Transport of (a) silver nanoparticles, (b) carbon nanotubes, (c) SDBS dispersed montmorillonite, and (d) water dispersed montmorilloni te in water saturated porous media. ................................ ................................ ..................... 42 3 1 Deposition of MWNTs in saturated porous media at stage 1 with different combinations of solution pH, IS, and porous medium type and size. .................. 59 3 2 SEM images of (A) acid cleaned glass beads, (B) acid cleaned sand, (C) glass beads with MWNTs attached on their surface, and (D) sand with MWNTs attached on their surface. ................................ ................................ ..... 60 3 3 Effect of solution I S on MWNT mobilization in saturated porous media at stages 2 and 3 with different combinations of solution pH, IS, and porous medium type and size. ................................ ................................ ........................ 61 3 4 Schematic diagram of migration of a MWNT from secondary minimum to primary minimum due to hydrodynamic force. ................................ .................... 62 4 1 DLVO energy between SDBS dispersed SWNT and (a) medium sand, (b) fine sand, (c) coarse sand, and (d) air water interface. ................................ ...... 80 4 2 Transport of SDBS dispersed SWNTs in sand columns under different volumetric moisture content conditions. ................................ .............................. 81 4 3 SWNT concentrations at the eight sampling ports within the bubble columns measured at different time intervals. ................................ ................................ ... 82 5 1 FTIR spectra of pristine SWNTs and O SWNTs. ................................ ................ 98 5 2 Stability of surface modified SWNTs in aqueous solutions. ................................ 99 5 3 Temporal changes of surface modified SWN T concentrations at the eight sampling ports within the bubble column. ................................ ......................... 100 5 4 Transport of surface modified SWNTs in medium sand columns under saturated (volumetric moisture content, 0.39) and unsaturated (volumetric moisture content, 0.21 and 0.13) conditions. ................................ .................... 101

PAGE 12

12 5 5 Transport of surface modified SWNTs in fine sand columns under saturated (volumetric moisture content, 0.38) and unsaturated (volumetric moisture content, 0.33 and 0.29) conditions. ................................ ................................ ... 102 5 6 Transport of surface modified SWNTs in coarse sand columns under saturated (volumetric moisture content, 0.41) and unsaturated (volumetric moisture content, 0.11 and 0.09) conditions. ................................ .................... 103 6 1 acid cleaned sand, baked sand, and the natural sand at 4000X, respectively. EDS spectra were recorded at the same locations showed in the SEM image. 123 6 2 Transport of CNTs in water saturated columns in DI water system packed with (a) acid cleaned sand, (b) baked sand and (c) natural sand. .................... 124 6 3 Distri bution of retained CNTs in the natural sand column: (a) supernatants from different layers after DI water washing; (b) supernatants from different layers after ultrasonication; (c) sand from different layers dried after ultrasonication. ................................ ................................ ................................ 1 25 6 4 Transport of CNTs in water saturated columns packed with baked and natural sand at different pH conditions. ................................ ............................ 126 7 1 FTIR spectra (average of 32 scans) of CNTs before and after heavy metal adsorption ................................ ................................ ................................ ......... 142 7 2 Sorption isotherms of Pb 2+ and Cu 2+ onto sand, and dispersed and undispersed CNTs in single and dual metal solutions ................................ ...... 143 7 3 Transport of Pb 2+ and Cu 2+ in single metal solutions through different types of fixed bed columns ................................ ................................ ............................ 144 7 4 Transport of Pb 2+ and Cu 2+ in dual metal solutions through different types of fixed bed columns ................................ ................................ ............................ 145 8 1 Breakthrough curves of sulfamethoxazole (SMX) and sulfapyridine (SPY) in the sand/CNT fixed bed columns under various CNT incorporation methods .. 165 8 2 Breakthrough curves of sulfamethoxazole (SMX) and sulfapyridine (SPY) in the sand/CNT fixed bed columns under various pHs ................................ ........ 166 8 3 Breakthrough curves of sulfamethoxazole (SMX) and sulfapyridine (SPY) in the sand/CNT fixed bed columns under various bed depths ............................ 167 8 4 Bed Depth Service Time (BDST) model for sulfamethoxazole (SMX) and sulfapyridine (SPY) at breakpoint (C/C0 = 5%) and exhausted point (C/C0 = 95%) ................................ ................................ ................................ .............. 168

PAGE 13

13 8 5 Breakthrough curves of sulfamethoxazole (SMX) and sulfapyridine (SPY) in the sand/CNT fixed bed columns under various ad sorbent dosages ................ 169 8 6 Breakthrough curves of sulfamethoxazole (SMX) and sulfapyridine (SPY) in the sand/CNT fixed bed columns un der various adsorbate initial concentration s ................................ ................................ ................................ .. 170 8 7 Breakthrough curves of sulfamethoxazole (SMX) and sulfapyridine (SPY) in the sand/CNT fixed bed columns under various flow rates ............................... 171 8 8 Relationship between changes in column capacity (mg/g) of (A) sulfam ethoxazole (SMX) and (B) sulfapyridine (SPY) for 5 fixed bed regeneration cycles ................................ ................................ .......................... 172 A 1 Absorbance spectra (A) and calibration c urves (B) of functionalized MWNTs. 181 A 2 Stability of functionalized MWNTs with DI water, 1 mM and 10 mM ionic strength (IS) at (A) pH 5.6 and (B) pH 10.0. ................................ ..................... 182 A 3 DLVO calculation for MWNTs stability using hydrodynamic diameter ( D h ), physical diameter ( D ) and length ( L ) ................................ ................................ 183 A 4 DLVO energy profiles between MWNT particles and grains calculated using physical diameter ( D ) of MWNTs ................................ ................................ ...... 184 A 5 DLVO energy profiles between MWNT particles and grains calculated using physical Length ( L ) of MWNTs ................................ ................................ ......... 185 B 1 Stability of the SDBS dispersed SWNT suspension ................................ ......... 188 C 1 Stability of functionalized CNT suspensions. ................................ .................... 189 C 2 Absorbance spectra (a) and calibration curves (b) of functionalized SWNTs and MWNTs ................................ ................................ ................................ ..... 190 C 3 DLVO energy between the CNTs and porous media. ................................ ....... 193 C 4 Mass titration curves of the porous media and CNTs. ................................ ...... 194 D 1 Mass titration curves of the porous media and CNTs. Point of Zero Charge (PZC) of the sand and the CNTs were determined using the mass titration metho d. ................................ ................................ ................................ ............ 196 E 1 Release of sulfamethoxazole (SMX) and sulfapyridine (SPY) from the post adsorption CNT/sand fixed bed columns during regeneration. ......................... 197

PAGE 14

14 LIST OF ABBREVIATION S ADE A dvection dispersion equation AgNPs S ilver nanoparticles CNTs Carbon nanotubes CVD C hemical vapor deposition DI D eionized DLVO Derjaguin Landau Verwey Overbeek EDL E lectrical double layer ENMs E ngineered nanomaterials E NPs Engineered nanoparticles EPM Electrophoretic mobility HA H umic acid ICP AES I nductively coupled plasma with atomic emission spectrometry MWNTs M ulti walled nanotubes NOM Natural Organic M atter OC O rganic carbon PAZ P rimary adsorption zone PZC P oint of zero charge PV P ore volume SDBS Sodium dodecylbenzene sulfonate SEM EDS Scanning electron microscope coupled with energy dispersive spectroscopy SRHA Suwannee River Humic A cid SWNTs S ingle walled nanotubes SMX Sulfamethoxazole SPY S ulfapyridine

PAGE 15

15 TOC Total Organic Carbon UVS UV visible spectrophotometer

PAGE 16

16 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy TRANSPORT AND ENVIRONMENTAL APPLICATION S OF CARBON NANOTUBES IN POROUS MEDIA By Y uan T ian August 201 2 Chair: Bin Gao Co Chair: Gregory A. Kiker Major: Agricultural and Biological Engineering There are increasing concerns over the environmental impact and health risks of carbon nanotubes (CNTs) because they may be released into the environment, such as soil and groundwater systems. T h is Ph D diss ertation systematically investigated the fate and transport of CNTs in porous media as well as the application s of CNTs i n wastewater treatment to remove heavy metal and antibiotic pollutants Laboratory columns packed with quartz sand and glass beads sho wed that d eposition of C NTs in saturated porous media was very sensitive to solution chemistry. M ore C NTs were found to be tr apped in the primary minimum U nder unfavorable conditions, w eak associated CNTs in the secondary minimum may be transferred into the primary minimum due to the effect of hydrodynamic force and/or local favorable sites associated with surface heterogeneity. D ispersion/solubilization methods showed great influence on the stability and mobility of C NTs. Humic acid dispersion granted higher mobility to C NTs compared with other surface modification methods. Reductions in moisture contents showed greater influence on the retention of surface oxidiz ed C NTs This is possibly due to CNT

PAGE 17

17 deposi tion on the air water interface and through film straining. C ombined mechanisms coul d be responsible for the retention and transport of C NTs in unsaturated porous media. R etention and transport of functionalized CNTs in natural sand porous media w ere mainly controlled by strong surface deposition through the electrostatic and/or hydrogen bonding attractions between surface function groups of the CNTs and metal oxyhydroxide impurities on the sand surfaces. F ixed bed column s packed with functionalized multi walled CNT s and natural sand w ere used in laboratory to evaluate and optimize the applications of CNTs in removing heavy metals and antibiotics from water T hey s ignificantly improved the fixed filtering efficiency of Pb 2+ and Cu 2+ by 55% 75% a nd 31% 57%, respectively. The fixed bed column experiments also showed that pH could be a key factor that affects the removal of antibiotics by controlling the protonation of antibiotics and surface charge of the carboxyl and hydroxyl functional groups on CNT surface s The column removal efficiency of antibiotics decreased only slightly after regenerations suggesting the CNT sand columns can be efficiently used and r egenerated to remove contaminants from water.

PAGE 18

18 CHAPTER 1 1 INTRODUCTION Background and Problem Statement Nanoparticles (NPs) defined as particles with at least one dimension smaller than 10 0 n m have received much recent attention because of the ir potential toxic effects and the rapid development of n anotechnology [ 1 8 ] Carbon nanotubes (CN Ts) are among the top concerned N Ps in the environment [ 9 11 ] Entirely composed of carbon w ith a significantly large length to diame ter ratio and unique physicochemical properties CNTs are rolled up graphene sheets with exceptional mechanical, electrical, optical, and thermal properties [ 12 15 ] There are mainly two types of CNTs: single and multi walled. Single walled carbon nano tubes (SWNTs) are one layered graphitic cylinders having diameters on the order of a few nanometers, while multi walled carbon nanotubes (MWNTs) comprise of 2 to 30 concentric cylinders having outer diameters often between 30 50 nm. They are largely used i n many novel applications in nanotechnology, electronics, optics, thermal conductors, and other fields in material science and engineering [ 16 19 ] CNTs are considered to pose potential ecological and health risks although their applications greatly en hance and accelerate the industrial renovation and revolution [ 20 21 ] Recent studies of the environmental impacts of CNTs have revealed their toxic effects towards a variety of aquatic and mammalian organisms [ 22 24 ] As a result, there is increased concern among the scientific communit y and regulation agencies over the environmental and ecological risks of CNTs [ 25 26 ] T he exponential growth in production of CNTs and their widespread application in consumer products will inevitably result in their release into the environment [ 27 28 ]

PAGE 19

19 W hen release d into soils they could impose contamination risks on groundwater if the CNTs can move through the vadose zone to reach the water table [ 27 29 31 ] Groundwater is an importan t resource of po t able water and thus t he investigation of fate and transport of CNTs in porous media is essential to the assessment of ecological and health risks of CNTs. M oreover, because of their strong sorption ability, CNTs could also serve as carrier s for other contaminates significantly influenc ing contaminant movement in porous media, such as soil and groundwater systems [ 1 32 35 ] C urrent understanding of fate and t ransport of CNTs in the porous media, however, is still limited. Most of the theories and mechanisms of CNT rete ntion and transport in porous media are mainly based on the well established colloid transport theories [ 36 39 ] Colloids ( i.e., particles with at least one dimension smaller than 10 m ) are ubiquitous in soils and thus are very important to both surface water and groundwater quality due to their small particle size and high surface area [ 40 45 ] Colloids are usually classified as organic colloids and inorganic colloids Organic colloids mainly includ e organ ic matters and biocolloids (e.g. viruses and bacteria) and inorganic colloids are constitute d by colloidal mineral s, clay s engineered nano particles (NPs) etc. [ 46 50 ] Theories and mechanisms of colloid transport in porous media are well dev e lope d and are widely available in the liter ature [ 27 36 44 51 53 ] A ccording to the theories of colloid transport in porous media attachment on grain surfaces has been suggested to be the primary mechanism s c ontr olling colloid retention in saturated porous media [ 54 57 ] P hysical trapping mechanisms, such as wedging and pore straining are also considered as important to the retention of colloid in saturated porous media [ 42 52 53 58 60 ] For colloid transport in unsaturated porous media,

PAGE 20

20 multiple mechanisms, including attachment on grain surfaces, p hysical trapping, air water interface deposition thin water film straining, and attachment on air water solid interfaces [ 61 66 ] However, it is unclear whether the se theor ies /mechanisms ca n applied to describe the deposition, mobilization, and transport of CNTs in po rous media Good understanding of the fate and transport of CNTs in porous media is also very important to the development of CNT based filters for water purifications. As indic ated above, CNTs have very large surface area, which grants them hi gh potential ability to adsorb various chemical poll utants [ 35 67 70 ] P rev ious studie s have demonstrated strong affinities of heavy metals (e.g. Cu 2+ Pb 2+ Cd 2+ ) and antibiotics (e.g. s ulfamethoxazole ) to CNTs under diffe rent conditions [ 71 75 ] A lthou gh many bat ch sorption experiments have been conducted to determine sorption characteristics and mechanisms of these contaminants on CNTs [ 76 77 ] to my knowledge, no stu dy has attempted previously to test the feasibility and effective ness of us ing CNTs as filter media in fixed bed settings to remove contaminants from water. This knowledge gap may impede the applications of CNTs in the environmental field, particularly wit h respect to the application of CNTs in water treatment. Research Objectives The main research objectives of this Ph.D. dissertation are follows: Objective 1: Fate a nd Transport o f C NT s i n Saturate d Clean Porous Media The specific objectives are to ( 1) compare the transport behavior of CNTs with that of engineered NPs and natural colloids in water saturated porous media ( 2) examine the effect of the surfactant (SDBS) on the filtration and transport of CNTs in the media ( 3) investigate the effect s of solution IS and pH on the deposition and mobilization of C NTs in saturated porous media ( 4 ) study the effect of porous media type and size on

PAGE 21

21 the fate and transport of C NTs and (5) determine whether theories developed for colloid transport in porous media can be applied to CNTs. Objective 2 : Fate a nd Transport o f C NT s i n Unsaturated Clean Porous Media The specific objectives are to ( 1) evaluate the effect of dispersion methods on the retention and transport of the CNTs in porous media (2) examine how changes in moisture content affect the retention and transport of the CNTs in porous media (3 ) determine the effects of porous medium sizes on the retention and transport of the CNTs and (4) model the retent ion and transport of C NTs in water saturate d and unsaturated porous media Objective 3 : Fate a nd Transport o f CNTs in Natural Porous Me dia The specific objectives are to ( 1) compare the retention and transport of the CNTs in laboratory column packed with natural, acid cleaned, and baked sand ( 2) determine the effect of perturbations in flow direction, flow rate, and surfactant concentration on remobilization of initially retained CNTs in natural and baked sand and ( 3) examine the retention and transport of the CNTs in natural and baked sand under different pH conditions. Objective 4 : Interaction o f CNTs with Contaminants The objectives are to ( 1) determine the effect of sonication promoted dispersion on sorption capacity of CNTs to heavy metals and antibiotics in the fixed bed column (2) examine the removal of heavy metals and antibiotics from aqueous solutions by trickling the contaminated water through a sand/CNT fixed bed column under different sets of conditions and (3) evaluate the efficiency of regeneration of the fixed bed columns for reuse

PAGE 22

22 Organization of the Dissertation This Ph D dissertation has nine chapters, including the present introductory chapter (Chapter 1). Chapter 2 discusses transport behavior s of two NPs, silver nanoparticles (AgNPs) and carbon nanotubes (CNTs), in saturated porous media. Sodium dodecylbenzene sulfonate (SDBS), an anionic surfactant, was used to disperse the engineered NPs to enhance their stabilities in water. The solubilized NPs were then applied to laboratory columns packed with two types of wa ter saturated quartz sand to obtain their breakthrough curves. Chapter 3 describes the transport, deposition, and mobilization behaviors of multi walled carbon nanotubes (MWNTs) in saturated columns packed with acid cleaned glass beads and quartz sand of t wo different grain sizes. Combined effect s of pH (5.6 and 10) and ionic strength (IS: DI water, 1 mM, and 10 mM) on the fate and transport of the MWNTs in the columns were examined. Chapter 4 explore s the transport mechanisms of sonication shortened, s odiu m dodecylbenzene sulfonate (SDBS) dispersed single walled nanotubes (SWNTs) in both saturated and unsaturated sand columns Laboratory columns packed with quartz sand with different combinations of moisture content and grain size distribution were used to examine the breakthrough behavior of the SDBS dispersed SWNTs. Bubble column experiments were also conducted to study the interactions between the SDBS dispersed SWNTs and the air water interface. Chapter 5 discusses the effect of different surface modific ation methods including oxidization surfactant ( sodium dodecylbenzene sulfonate) coating, and humic acid coating on SWNT stability and their mobility in granular porous media of different grain sizes under both saturated and unsaturated conditions. Chapter 6 mechanistically compare s the retention and transport of two types of functionalized CNTs (i.e., single walled nanotubes and multi walled nanotubes ) in

PAGE 23

23 acid cleaned, baked, and natural sand under unfavorable conditions. Chapter 7 describes novel m etho ds of using carbon nanotubes as filter media to remove aqueous heavy metals B atch sorption and fixed bed experiments were conducted to examine the ability of functionalized multi wa lled CNTs as filter media to remove two heavy metal ions (Pb 2+ and Cu 2+ ) from infiltrating water. Chapter 8 conce ntrates on removal of sulfamethoxazole ( SMX ) and sulfapyridine ( SPY ) by c arbon nanotubes in fixe d bed columns under a broad range of conditions including: CNT incorporation method, solution pH, bed depth, adsorbe nt dosage, adsorbate initial concentration, and flow rate Chap ter 9 summarizes the results of all the previous chapters and makes recommendations on future work R eferences are included at the end of this document.

PAGE 24

24 CHAPTER 2 2 TRANSPORT OF ENGINEERED NANOPARTICLES IN SATURATED POROUS MEDIA 1 Introduct ory Remarks Nanotechnology is often referred to as the foundation of the next industrial revolution, which has immense potential to improve the quality of life. Recent research, however, strongly points t o the health risks associated with many engineered nanoparticles (NPs) [ 21 78 79 ] An increased production of various engineered NPs and their widespread application in consumer products will inevitably result in their release into the environment [ 27 28 ] It is therefore crucial to examine the fate and transport of NPs in the environment. Carbon nanotubes (CNTs) and silver nanoparticles (AgNPs) are two of the most commonly used engineered NPs that have attracted much attention recently [ 33 80 82 ] CNTs are carbon nanomaterials that belong to the fullerene family, i.e., cylindrical fullerenes with a high length to diameter ratio [ 12 ] The novel properties of CNTs prompt their applications in many emerging technologies, particularly in c reating new materials with extraordinary strength, and unique electrical and thermal properties [ 16 83 ] R ecen t studies of environmental impacts of CNTs however, have generally shown toxic effects of CNTs toward the ecosystem, particularly with respect to aquatic organisms [ 22 24 8 4 87 ] AgNPs have wide application in numerous consumer products including clothing, cosmetics, medical care products, food storage, toys, and even food supplements [ 88 ] AgNPs can also be toxic to the ecosystem because silver is well 1 Reprinted with permission from Tian, Y.A., Gao, B., Silvera Batista, C., and Ziegler, K.J., Transport of engineered nanoparticles in saturated porous media. Journal of Nanoparticle Research, 2010. 12 (7): p. 2371 2380.

PAGE 25

25 known for its antibacterial and other destructive behaviors. The toxic effects of AgNPs to microorganisms, plants, fish, animals, and human cells have been observed in several studies [ 78 88 89 ] The toxicity studies of CNTs and AgNPs have also heightened the concern over their env ironmental exposure to humans through groundwater contaminations [ 90 91 ] S everal investigations have been conducted to study the environmental fate and transport of engineered NPs including CNTs [ 92 93 ] Column studies with well defined porous media are often used to explore the governing mechanisms of CNT filtration and transport in groundwater [ 93 ] These studies are valuable and have expa nded the knowledge of the fate and transport of CNTs and other engineered NPs in porous media. Nevertheless, the current understanding of engineered NPs in porous media is far from complete. For instance, there is little/no research effort being spent on s tudying the transport behavior of AgNPs in porous media. Further experimental and theoretical investigations are still in critical need to facilitate the development of regulatory guidelines to manage engineered NPs in soils and groundwater [ 79 ] Fate and transport of colloids, such as clay, metal oxides, and bacteria, are well documented in the literature [ 92 94 95 ] The Derjaguin Landau Verwey Overbeek (DLVO) theory and the colloid filtration theory have been developed and applied to the fate and transport of colloids in well defined porous media [ 95 97 ] Because of the similarities between natural colloids and engineered NPs, these theories may also be applicable to CNT and AgNP transport in well defined porous media. T here are also no t able differences between engineered NPs and natural colloids in many physicochemical properties [ 98 ] The ultrafine size and extremely high surface area of

PAGE 26

26 engineered NPs may introduce new mechani sms affecting their transport behavior in porous media. In this work, laboratory column experiments were conducted to examine the transport of CNTs and AgNPs in water saturated porous media. An anionic surfactant, sodium dodecylbenzene sulfonate (SDBS), w as used to modify the surface of the NPs to enhance their stability in solution. The NP suspensions were then introduced to an acrylic column packed with two types of water saturated quartz sand to obtain the breakthrough curves. Mathematical models based on the DLVO theory and the colloid filtration theory were used to simulate the experimental data. Our objectives were to examine 1) the transport behavior of two engineered NPs (CNTs and AgNPs) in water saturated porous media; 2) the effect of the surfacta nt (SDBS) on the filtration and transport of NPs in the media; and 3) whether theories and models developed for colloid transport in porous media can be applied to the engineered NPs. Methods and Materials CNTs There are mainly two types of CNTs: single a nd multi walled. Single walled carbon nanotubes (SWCNTs) are one layered graphitic cylinders having diameters on the order of a few nanometers, while multi walled carbon nanotubes (MWCNTs) comprise of 2 to 30 concentric cylinders having outer diameters oft en between 30 50 nm. Only HiPco SWCNTs (Rice HPR 162.3) were used in this study to prepare suspensions [ 99 ] As the SWCNTs have a strong tendency to form bundles in water, surfactants are often used to disperse them [ 99 100 ] In this study, the CNT suspensions were prepared by mixing 20 mg of raw SWCNTs with 200 m L of an aqueous SDBS surfactant solution (1% by weight). High shear homogenization (IKA T

PAGE 27

27 25 Ultra Turrax) for 1.5 2.0 h and ultrasonication (Misonix S3000) for 10 min were used to aid dispersion. Subsequently, the suspension was ultracentrifuged at 20,0 00 rpm (Beckman Coulter Optima L 80k) for 5 h to eliminate the aggregates and large SWCNTs from suspension. Previous studies have demonstrated that this method can create a surfactant shell on the sidewall of the nanotubes to stabilize aqueous SWCNT suspen sions [ 99 ] The SWCNTs were tubular shape with average sizes of 400 nm in length and 0.8 nm in diameter measured using atomic force microscope [ 100 ] The zeta potential ( of the surfactant coated SWCNTs were measured by Brookhaven ZetaPlus (Brookhaven Instruments, Holtsville, NY). The zeta potential of the SWCNTs in the suspension was 42.84 mV ( Table 2 1). AgNPs Silver nanoparticles were obtained from the Particle Engin eering Research Center at the University of Florida. They were synthesized by gas phase condensation and comprise a metal core with a thin metal oxide coating [ 101 ] The AgNP suspensions were prepared by mixing 20 mg of raw material with 200 mL o f an aqueous SDBS surfactant solution (1% by weight). The same procedures used in preparing CNT suspensions were used to aid dispersion of AgNPs. The average particle size of AgNPs was 52.4 nm and was 49.49 mV ( Table 2 1). Colloids Montmorillonite powde rs (Southern Clay Products Inc., Gonzales, TX) were used to make colloid suspensions used in the experiments. Two types of colloid suspensions were made, water dispersed and SDBS dispersed. To make the water dispersed montmorillonite suspension, about 1.0 g of kaolinite powder was suspended in 1 L deionized (DI) water. The suspension was shaken vigorously, placed in an ultrasonic

PAGE 28

28 bath for 30 minutes, and then left to stand for 24 h. The fraction of montmorillonite remaining in suspension after 24 h was sip honed into a second flask. The concentration of montmorillonite in an aliquot of this stock suspension was determined gravimetrically before diluting the stock to colloid suspensions used in the experiments. Part of this stock suspension was used to make the SDBS dispersed montmorillonite by adding the same amount of SDBS as used in CNT and AgNP suspensions (1% by weight). The mean sizes of the water and SDBS dispersed montmorillonite were 236.8 49.66 mV and 50.56 mV resp ectively ( Table 2 1). Porous Medium Quartz sand (45/30) was used as the porous medium (Standard Sand & Silica Co.) in the column experiments. The purity of the sand was examined using X ray diffraction (XRD) and contained more than 99.99% of quartz. The sand was sieved to a size range of 0.35 0.60 mm ( d 50 =0.50 mm) and treated with two methods. Sand A was washed sequentially by tap water and deionized water, and then baked in Fisher Isotemp muffle furnace (Fisher Scientific, Rochester, NY) at 550 C to rem ove the trace quantities of organic impurities. Sand B was washed sequentially by tap water, 10% nitric acid and deionized water, and then baked at 550 C to remove the metal oxides and organic impurities. The zeta potentials of the two resulting quartz sa nd were determined according to the procedure of Johnson et al. [ 94 ] Colloidal quartz particles were collected by placing a mixture containing 100 g of cleaned quartz sand and 2 00 mL DI water in an ultrasonic bath for 30 min. After the ultrasonic treatment, aliquots of the quartz colloids were removed and filtered through a 0.45 m filter. The filtrate was analyzed for zeta potential. The for sand A and B were 24.30 mV and 19 .73 mV respectively ( Table 2 1).

PAGE 29

29 Column Experiments The quartz sand (A & B) was wet packed into an acrylic column measuring 2.5 cm in diameter and 15 cm in height. To make a saturated porous media, a small amount of sand was poured gently into 8 mL of DI water standing at the bottom of the column until the sand surface was 0.5 to 1 cm below the water level. A polypropylene stir rod was used to stir the sand in the column. Approximately 8 mL DI water was then added to the column and the sides of the colu mn were gently tapped a uniform number of times. This procedure was repeated several times until the column was packed to a height of 15 cm. Approximately 140 g of sand was used to pack one column with a porosity of 0.42. A peristaltic pump (Masterflex L /S, Cole Parmer Instrument, Vernon Hills, IL) was connected to the inlet at the top of the column to regulate the downward flow at a constant specific discharge of 0.2 cm/min. DI water was first pumped through the saturated column for about 2 h to remove i mpurities followed by a SDBS solution (1% by weight) for another 2 h. The breakthrough experiment was then initiated by switching from SDBS solution to proper suspensions/solutions. CNT, AgNPs, and SDBS dispersed montmorillonite suspensions were introduce d to the sand columns for a two stage breakthrough study. In the first stage (stage 1), the nano/colloid suspension was applied to the column as a 1 h pulse, and then the column was flushed with SDBS water for 2 h. In the second stage (stage 2), the sand c olumn was flushed with DI water for another 2 h. The breakthrough study of water dispersed montmorillonite involved only one stage (stage 1) because no SDBS was used in the experiments. Effluent samples were collected from the bottom of the column with a f raction collector (IS 95 Interval Sampler, Spectrum Chromatography, Houston, TX)

PAGE 30

30 during sample injection and column flushing to analyze SDBS and particle concentrations. Water and SDBS dispersed montmorillonite concentrations were determined by measuring the total absorbance of light at a wavelength of 350 nm with a UV visible spectrophotometer (UVS). CNT, AgNP, and SDBS concentration were also determined by UVS at wavelengths of 320 nm, 395 nm, and 245 nm respectively. Bromide was applied to the column a s a conservative tracer for the breakthrough studies. The experimental procedures were the same as those used for the water dispersed montmorillonite and an ion chromatograph (ICS 90, Dionex Corporation, Sunnyvale, CA) was used to determine bromide concent rations. All the breakthrough experiments were performed in duplicate. DLVO T heory The classic DLVO theory was used to estimate the particle sand grain interaction energy at the experimental conditions by combining the van der Waals attraction and electrical double layer (EDL) repulsion. The Lifshitz van der Waals attraction energy ( ) for a sphere plate system can be written as [ 102 ] : ( 2 1) where A 132 me ( Table 2 1), h is the separation distance, and r is the radius of the particle. The EDL repulsion energy ( )for a sphere plate system can be written as [ 102 ] : ( 2 2)

PAGE 31

31 where is the dielectric constant of the medium (78.4 for water), 0 is the vacuum permittivity (8.85410 12 C 2 N 1 m 2 ), k 23 C 2 J K 1 ), T is the temperature, z is the valence of electrolyte, e is the electron charg e (1.60210 19 C), 1 and 2 are the surface potential of the particle and the sand surface, and is and sand can be determined following van Oss et al. [ 103 ] : ( 2 3) where d is the distance between the surface of the charged particle and the slipping plane and usually taken as 5 angstroms (van Oss et al. 1990). The total DLVO interaction energies between the particles and the sand were determined using equations (1) (3) and were normalized with kT. Transport Model A transport model based on the colloid filtration theory was used to simulate the retention, transport, and re mobilization of engineered NPs in the water saturated sand column. The governing equation can be written as [ 104 ] : ( 2 4) where C is the concentration of suspended particles, D is the dispersion coefficient, v is the av erage linear water velocity, z is the travel distance in the direction of flow, b is the media bulk density, n is the porosity, and S is the deposited/retained particle concentration. The last term is a particle filtration term, which includes both the rate of retention and re mobilization. The colloid filtration theory first developed by Yao et al. [ 96 ] provides an easy way to calculate particle retention rate in porous media using a

PAGE 32

32 single collect model. More recent studies of colloid transport in porous media have shown the re mobilization processes should also be considere d [ 36 104 ] The filtration term can be expressed as: ( 2 5) where k a is the particle retention rate constant and k b is the particle re mobilization rate constant (h 1 ). According to the colloid filtration theory, k b is zero under steady flow condition. When there are perturbations in flow physicochemical conditions, the d eposited particles in porous media will be released (i.e. k b >0). We solved the governing equations of the transport model (equations 4 5) numerically for a zero initial concentrations, a pulse input boundary condition at the column inlet, and a zero conc entration gradient boundary condition at the outlet. The Levenberg Marquardt algorithm was used to estimate the value of the model parameters to minimize the sum of the squared differences between model calculated and measured breakthrough concentrations. This model optimization method was first applied to bromide breakthrough data to estimate D We assumed that the dispersion coefficient ( D ) of particles and SDBS is the same as that of the bromide tracer in the column. The retention and re mobilization rat es of SDBS and particles in water saturated porous media then were determined by identifying the best fit values of k a and k b Results and Discussion DLVO E nergy The Hamaker constant between the AgNPs and the porous media in water were estimated from the Hamaker constants of two silver particles in vacuum without the

PAGE 33

33 consideration of the surfactant influences ( Table 2 1). The calculated Hamaker value of AgNPs in the experimental system was 4.4710 20 J, which was much larger than that of the montmorillonit e colloid (1.8110 21 J). Although there is no reported value of Hamaker constant of two SWCNTs in vacuum, Chen and Elimelech [ 80 86 ] showed that fullerene (C60) nanoparticles in aqueous media have a Hamaker constant around 8.5 10 21 J. Because of the strong similarity between the SWCNT and fullerene nanoparticles, we used their Hamaker constant value to calculate the Hamaker constant between SWCNTs and the porous media in water. The calculated Hamaker value of SWCNTs in the experimental system was 9.810 21 J, which was also larger than that of the montmorillonite colloid (1.8110 21 J). The high values of the Hamaker constant of AgNPs and SWCNTs suggested strong van der Waals attractive forces between the NPs and the grain surfaces under the experimental conditions. The EDL energy between all the particles and the porous media was repulsive because all the measured Table 2 1). As the DLVO theory is for spherical particles, the average sizes of the parti cles were used in this study as the effective diameters (De). The obtained DLVO energy profiles between the AgNPs and the porous media showed secondary energy minima at a separation distance around 30 nm ( Figure 2 1a), suggesting a weak attraction between the AgNPs and the porous media The repulsive energy barrier was high with values around 30 kT (sand A) and 25 kT (sand B) ( Figure 2 a). The high energy barrier between the AgNPs and porous media indicated that the experimental condition s w ere unfavorable for AgNP deposition onto the grain surfaces in the primary energy minima.

PAGE 34

34 Because CNTs are tubular, both the tube length (De = 400 nm) and the tube diameter (De = 0.8 nm) were used to calculate the DLVO energy profiles between the CNTs and the porous med ia ( Figure 2 1b). When the tube length was used, the energy profiles were similar to those of the AgNPs: second minimum and high energy barriers. When the tube diameter was used in the calculations, the energy profiles were much lower without a secondary e nergy minimum ( Figure 2 1b). Although the DLVO energy profiles showed much difference between the two effective diameters, both cases suggested that the experimental condition s w ere unfavorable for CNT deposition onto the grain surfaces in the primary ener gy minima. The energy profiles of SDBS and water dispersed montmorillonite were similar to those reported in the literature [ 105 ] and had repulsive energy barriers around 250kT ( Figure 2 1c and 1d). The presence of SD BS in the solution greatly reduced the repulsive forces between the clay colloids and the two porous media. The DLVO energy of SDBS dispersed montmorillonite showed secondary energy minima at a separation distance around 20 nm ( Figure 2 1c). This is becaus e the presence of SDBS ( 5.74 mM ) in the experimental solution increases the ionic strength, compressing the thickness of the diffusion layer and reducing the repulsive EDL forces ( Table 2 1) [ 102 ] Transport in Porous Media The transport of AgNPs in the two sand media exhibited typical breakthrough behavior s in stage 1 After they were applied to the sand columns, AgNPs were detected in the effluents around 1 pore volume. The breakthrough curves then quickly climbed to a peak value and stayed ther e during the particle injection. The AgNP concentrations decreased quickly to zero after the columns were flushed with particle free SDBS solution. Mass balance calculations showed that about 24% (sand A) and 14%

PAGE 35

35 (sand B) of AgNPs were retained in the two sand columns in stage 1 before the columns were flushed with DI water ( Table 2 2). The retention of AgNPs in the saturated porous media was mainly controlled by deposition in the secondary energy minima ( Figure 2 1a), where the attraction between the parti cle and grain surface is weak and reversible [ 106 ] The secondary minimum deposition was further confirmed by the transport behavior of AgNP i n stage 2, where most of the retained AgNPs in the columns were released back into the pore water and the breakthrough curves showed pulses ( Figure 2 2a). In stage 2, the influents were switched from SDBS solution to DI water, reducing the ionic strength i n pore water and increasing the repulsive EDL forces to eliminate the secondary minima for AgNP attachment. Because there were more AgNPs retained in A than B columns more AgNPs were released from the A column at stage 2 The differences in AgNP depositio n and release in the two porous media may be caused by the different surface properties of the sand resulting from the different treatments (sand A: organic free, sand B: organic and metal oxide free). It has been demonstrated that metal oxides can increas e the deposition of negatively charged colloids to surfaces because of their positive charge and hydrophobicity [ 107 108 ] CNTs and montmorillonite showed similar breakthrough behaviors to AgNPs in the two saturated porous media: a brief period of rapi dly increasing concentrations followed by a plateau and a retreat (Figures 2 2b, 2 2c, and 2 2d ) Mass balance calculations showed that almost all CNTs and montmorillonite colloids passed through the sand columns in stage 1 with a recovery rate around 100% ( Table 2 2). For CNTs and SDBS dispersed montmorillonite, no particles were detected in stage 2 when the influents

PAGE 36

36 were switched to DI water. Our results indicate that the SWCNTs used in this study are highly mobile in saturated porous media. In column st udies of functionalized SWCNTs, Jaisi and his colleagues (2008, 2009) found that the SWCNTs exhibited low or limited transport in well defined porous media and natural sandy soils because of pore straining effects. However, we did not observe any straining effects because the SWCNTs used in this study may be too short (400 nm) to be strained by the pore throats. The transport of water dispersed montmorillonite in the porous media and DLVO predictions showed that the experimental conditions were unfavorabl e for attachment ( Figure 2 1d). The results of CNT and SDBS dispersed montmorillonite, however, were contradictory to the calculations of the DLVO theory (Figures 2 1b and 2 1c), which suggests particle deposition in the secondary minima. These differences could be because of deviations from DLVO theory. First, the effective diameters were used for the tubular and irregular shaped particles, which may not accurately reflect the actual interactions between the particles and grain surfaces. Second, the influe nces of the surfactant (SDBS) on the DLVO interactions were only included in ionic strength calculations. While several previous studies suggested that the presence of surfactants may also alter the surface properties of the particles (both NPs and colloid s) and sand grains by forming coating layers [ 109 110 ] This alteration not only may affect the extent of the van der Waals and EDL forces between the particles and grain surfaces, but also may introduce some non DLVO interactions [ 103 111 ] Although several studies have attempted to quantify these non DLVO interactions and apply them to NP stabilities [ 103 109 111 ] an extended DLVO theory for engineered NP interactions with porous medium is far from complete. Additional theoretical and experiment

PAGE 37

37 investigations are still needed t o modify and extend the DLVO theory to improve its predictions to engineered NP systems. Model Simulations The transport model was first applied to the breakthrough data of bromide to obtain the dispersion coefficient and the simulations matched the experimental data well with r 2 =0.99. Using the obtained dispersion coefficient, the model described the breakthrough curves reasonably well for all experiments with r 2 >0.96 ( Figure 2 2). The best fit values of k a for CNT, water and SDBS dispersed montmorillonite were very small ( Tab le 2 2), indicating the deposition rate of these particles to the sand grain surface is negligible under the experimental conditions. As expected, the best fit k a for AgNPs in sand A (0.43 h 1 ) was higher than that in sand B (0.34 h 1 ) indicating faster de position of AgNPs in sand A. As there were more AgNPs retained in the sand A column, the best fit k b in sand A (6.43 h 1 ) was higher than that in sand B (6.00 h 1 ) indicating faster release in sand A. The good match between the model simulations and experi mental data indicates that colloid transport models can be used to describe the fate and transport of engineered NPs in porous media. Chapter Conclusions The fate and transport of two engineered NPs (CNTs and AgNPs) in water saturated porous media were ex amined. Our results showed that the engineered NPs exhibited similar transport features to natural clay colloids. It is therefore feasible to take advantage of the existing theories and models of colloid transport to describe the fate and transport of engi neered NPs in porous media. Both the classic DLVO theory and filtration theory were applied to the engineered NP systems in this study; however, the presence of the surfactant (SDBS) caused deviations from the theory. Further

PAGE 38

38 investigations are still neede d to modify and improve the DLVO theory and to ensure the accuracy of its predictions in engineered NP systems.

PAGE 39

39 Table 2 1 Physicochemical properties of nanoparticles, natural colloids, and porous medium Particle Effective diameter (nm) Debye length (nm) Zeta potential (mV) Surface potential (mV) Hamaker constant (10 20 J) Silver nanoparticle 52.4 4.01 49.49 57.13 38.5 ref. [ 112 ] Carbon nanotube 90.0/0.8 4.01 42.84 55.87/ 125.39 8.2 calculated from [ 113 114 ] SDBS dispersed montmorillonite 252.7 4.01 50.56 57.50 7.8 ref. [ 105 ] Water dispersed montmorillonite 236.8 241.77 49.66 49.97 7.8 ref. [ 105 ] Sand A 0.510 5 24.30 27.52 8.8 ref. [ 115 ] Sand B 0.510 5 19.73 24.35 8.8 ref. [ 115 ] Water 3.7 ref. [ 115 ]

PAGE 40

40 Table 2 2 Best fit parameter values and recovery rate calculations for nanopart icle and colloid transport in saturated porous media K a (h 1 ) K b (h 1 ) Stage 1 Recovery (%) Stage 2 Recovery (%) r 2 Sand A Silver nanoparticles 0.43 6.43 76.0 14.5 0.97 Carbon nanotubes 3.00E 07 0 101.6 0 0.99 SDBS dispersed montmorillonite 3.00E 07 0 102.6 0 0.99 Water dispersed montmorillonite 3.00E 07 0 99.3 0 0.99 Sand B Silver nanoparticles 0.34 6.00 86.1 10.3 0.97 Carbon nanotubes 3.00E 07 0 103.1 0 0.99 SDBS dispersed montmorillonite 3.00E 07 0 104.3 0 0.96 Water dispersed montmorillonite 3.00E 07 0 101.3 0 0.99

PAGE 41

41 Figure 2 1 DLVO energy between the porous medium and (a) silver nanoparticle, (b) carbon nanotube, (c) SDBS dispersed montmorillonite, and (d) water dispersed montmorillonite.

PAGE 42

42 Figure 2 2 Transport of (a) silver nanoparticles, (b) carbon nanotubes, (c) SDBS dispersed montmorillonite, and (d) water dispersed montmorillonite in water saturated porous media.

PAGE 43

43 CHAPTER 3 3 EFFECT OF SOLUTION CHEMISTRY ON MULTI WALLED CARBON NANOTUBE DEPOSITION AND MOBILIZATION IN CLEAN POROUS MEDIA 1 Introduct ory Remarks Carbon nanotube s (CNT s ) allotropes of carbon with a cylindrical nanostructure are among the most popular engineered nanomaterials. With a significantly large length to diameter ratio and unique physicochemical properties, CNTs are widely used in many applications in nanotechnology, electronics, optics and architectural fields [ 16 ] However, serious environmental concerns about the negative impact s of CNTs have been aroused along with these prevailin g applications [ 27 ] Several risk assessment studies suggested that, after disposal, CNTs may pose risks to the ecosystems and the public health [ 21 78 79 ] Groundwater is an important resource of po t able water a nd thus the investigation of CNTs in the groundwater is the key to effective assessment of ecological and health risks of CNTs. As disposal, they may be potentially released into soil, subsequently transport in the soil profile, and finally enter groundwat er environment. Therefore, studies on transport behaviors of CNTs in the soil profile and related impact factors are very meaningful for the further risk assessment. Transport of CNTs in the porous media was thus increasingly studied recently and enhanced current understanding of their fate in soil and groundwater systems [ 67 93 110 116 117 ] Water saturated porous medium columns have been often used to explore the governing mechanisms of deposition and transport of CNTs. Effect of 1 Reprinted with permission from Tian, Y., Gao, B., Wu, L., Muoz Carpena, R., and Huang, Q., Effect of solution chemistry on multi walled carbon nanotube deposition and mobilization in clean porous media. Journal Of Hazardous Materials. 10.1016/j.jhazmat.2012.06.039(0).

PAGE 44

44 various physicochemical factors, such as solution ionic strength (IS) [ 93 ] pore water velocity [ 27 116 ] dissolved organic matter concentration [ 110 ] surfactant coating [ 118 ] moisture content [ 117 ] and grain size [ 67 117 ] on deposition and t ransport of single walled or multi walled CNTs were determined with laboratory columns packed with sand or glass bead porous media. Theories and models developed for describing the fate and transport of spherical colloids have also been used to analyze the transport behaviors of CNTs in laboratory porous medium columns [ 93 110 116 ] However, it is still unclear whether the well established colloid theories are applicable to CNTs, which are nanosi zed, tubular particles with extremely high aspect ratio [ 118 ] In the literature of colloid transport, pore water chemistry is always considered to play an important role in controlling the deposition and mobilization of colloids in porous media [ 36 119 120 ] It was found that increase in solution chemistry would reduce the double layer thickness and electrostatic repulsion and thus facilitate colloid deposition in porous media, particularly in secondary minimum [ 36 106 121 ] Solution pH was also regarded as one of the influential factors of fate and transport of colloids in porous media [ 36 108 ] Perturbations in solution pH may result in changes of surface potential to affect the deposition and mobilization of colloids in porous media [ 108 ] However, limit ed studies have be en conducted to investigate the effect of solution chemistry on CNT fate and transport in porous media [ 93 116 ] Systematic investigations are therefore needed to determine this effect, especially with respect to understanding the combined effect of solution IS and pH on CNT deposition and re entrainment in different types of porous media.

PAGE 45

45 In this work, laboratory column experiment was conducted to examine deposition, mobilization and transport of multi walled carbon nanotubes (MWNTs) in water satu rated porous media. Acid cleaned quartz sand or glass beads of different sizes were used in the study to pack the columns. The transport behaviors of MWNTs in the columns was examined in details with different combinations of solution IS and pH. Mathematic al models were used to aid in the data interpretations. The objectives were to: (1) determine the effect s of solution IS and pH on the deposition and mobilization of MWNTs in saturated porous media, (2) determine the effect of porous media type and size on the fate and transport of MWNTs, and (3) understand governing mechanisms of MWNT deposition and mobilization in saturated porous media. Materials a nd M ethods MWNTs MWNTs were produced using a chemical vapor deposition method with nickel and magnesium c atalysts and further functionalized by an acid mixture concentrated sulfuric and nitric acids (3:1, v:v) to introduce functional groups (e.g., carboxyl groups) to the nanotube surface [ 122 ] Stock suspension was prepared by dispersing 16 mg MWNTs in 1000 ml deionized (DI) water with Misonix S3000 u ltrasonicat or ( QSonica, Newtown, CT ) for 3 0 min. The stock suspension was used to make experimental MWNT suspensions by adding 1M KCl, KOH, or HCl solutions to obtain desired solution chemistry. Two pHs (5.6 and 10.0) and three ISs (DI water, 1 mM, and 10 m M) were tested in this work ( Table 3 1). Characterization of MWNTs Hydrodynamic diameter of the MWNTs was determined by dynamic light scattering using Brookhaven ZetaPlus (Brookhaven Instruments C orporation, Holtsville,

PAGE 46

46 NY). Actual dimensions of the MWNTs were determined through scanning electron microscope (SEM) (JEOL 6335F, Tokyo, Japan). Electrophoretic mobility (EPM) of the MWNTs was determined using ZetaPlus and Smoluchowski's model was used t o estimate the zeta potential [ 11 0 ] MWNT concentrations were calibrated by measuring the total absorption of light at wavelengths of 255 nm using Evo lution 60 UV Vis Spectrophotometer (Thermo Scientific, Waltham, MA) ( Figure A 1). Porous Media Two types of porous media, glass bead (Cole Parmer Inc., Vernon Hills, IL) and Quartz sand (Standard Sand & Silica Co., Davenport, FL), were sieved into two si ze distributions: fine (0.1 0.2 mm) and medium (0.5 0.6 mm). The porous media were cleaned sequentially by tap water, 10% nitric acid (v:v ), and DI water to remove the metal oxides and organic impurities. The zeta potentials of porous media under various solution chemistries were determined according to the procedure of Johnson et al. [ 94 ] In brief, 100 g of porous media were added into 200 mL background electrolyte solution and the mixture was ultrasonicated for 30 minutes. Aliquots of the quartz colloids in the mixture were then removed and filtered through a 0.4 5 m filter. The filtrate was analyzed for electrophoretic mobility and zeta potential. MWNT Stability Temporal changes of MWNT concentrations under various solution chemistry conditions were measured to evaluate its stability. 1000 ml of MWNT stock suspe nsions were fresh prepared. Every 30 min, 2 ml solution was sampled from the stock suspensions for 24 h. MWNT concentrations were determined by measuring the total absorption of light at wavelengths of 255 nm with UV Vis Spectrophotometer.

PAGE 47

47 Column Experime nts Packed column experiments were used to investigate the deposition, mobilization, and transport of MWNTs in water saturated porous media. The porous media were wet packed into an acrylic column measuring 2.5 cm in diameter and 15 cm in height. M embranes with 50 m pores (Spectra/Mesh, Spectrum Laboratories, Inc.) were used at the column inlet and outlet to seal the column and distribute the flow. The procedures of Tian et al. [ 121 ] were used to packed the columns to avoid l ayering or air entrapment The porosities of the columns were approximately 0.38 (fine sand), 0.40 (medium sand), 0.38 (fine glass bead), and 0.40 (medium glass bead). Transport of MWNTs in the columns was investigated in 1 3 stages depending on experimental working solution chemistry, which varied among three ISs and two pHs ( Table 3 1). A peristaltic pump (Masterflex L/S, Cole Parmer Instrument, Vernon Hills, IL) was connected to the inlet at the top of the column to regulate the downward flow at a specific discharge 2.0 mL/min for all the stages. At stage 1, after the column was flushed with background working solution for 3 PVs, MWNT suspension of the same chemistry was applied to the column for 2 PV s, and then the column was flushed with the working solution for another 2 PVs. For the experiments with relatively higher solution ISs (i.e. IS= 1 and 10 mM), the retained MWNTs in the columns at stage 1 were mobilized by reducing solution ISs in one or t wo stages (i.e., stages 2 and 3, Table 3 1). Effluent samples were collected from the bottom of the column using a fraction collector (IS 95 Interval Sampler, Spectrum Chromatography, Houston, TX) to analyze MWNT concentrations.

PAGE 48

48 Mathematical Model DLVO th eory ( Appendix A 1 ) was used to estimate the interaction energy in MWNT water MWNT and MWNT water medium systems under various solution chemistry conditions. The DLVO theory was developed spherical particles or flat surfaces. Because CNTs are tubular, the calculations can be only used as exploratory estimations. In this work, hydrodynamic diameter ( D h ), physical diameter ( D ), and length ( L ) of the MWNT were used to calculate the DLVO energy profiles for different circumstances ( Appendix A 1 ). Advection di spersion equation coupled with first order kinetics was used to describe MWNT deposition and transport in the porous media [ 104 ] The gover ning equation can be written as: ( 3 1 ) where C is the MWNT concentration, D is the dispersion coefficient, v is the pore water velocity, z is the travel distance in the direction of flow, and k is the first order retention rate This model was applied to the experimental data and solved numerically with a zero initial concentrations, a pulse input boundary condition, and a zero concentration gradient condition at the outlet boundary. Results and D iscussion Characteristics of MWN Ts and Porous Med ia The MWNTs were stable in the suspension and their concentration maintained unchanged for more than 10 hours for different IS and pH conditions ( Figure A 2 ). In

PAGE 49

49 this work, the column experiments were therefore performed within 10 hours t o avoid MWNT aggregation. The hydrodynamic diameter of the MWNTs was around 175 nm for all the tested conditions. SEM measurements indicated that their actual average diameter was about 40 nm and the length was about 400 nm. Zeta potential of the MWNTs inc reased from 38.2 to 13.5 at pH 5.6 and from 39.1 to 17.5 at pH 10 when solution IS increased ( Table 3 2). The estimated DLVO interaction energy profiles of the MWNTs confirmed that the experimental conditions were unfavorable for their aggregation, par ticularly when L or D h was used in the calculation ( Figure A 3 ). The DLVO profiles also demonstrated that the maximum energy barrier for MWNT aggregation decreased when solution IS increased or solution pH decreased. However, the secondary minimum well bec ame deeper when solution IS increased or solution pH decreased. This indicates that low solution pH or high solution IS may promote the aggregation of MWNTs (mainly through second minimum), which is consistent with traditional colloid/nanoparticle stabili ty theories [ 114 123 ] E PM and zeta potential measurements showed that the surfaces of the two porous media (i.e., sand and glass bead) were also negatively charged for all the tested treatment s ( Table 3 2). The zeta potential of the glass bead was more negative than that of sand for every solution chemistry, suggesting glass bead had higher surface charge density. The estimated DLVO interaction energy profiles between the MWNTs and the porous media suggested tha t all the tested experimental conditions were unfavorable for attachment ( Figure A 4 and A 5 ). Deposition of MWNTs under DI Water Co ndition When the experimental working solution was DI water, MWNTs were highly mobile in the two porous media (sand and gla ss bead) at pH 5.6 and 10 .0 ( Figure 3 1), which

PAGE 50

50 seems to be consistent with the prediction of the DLVO theory. Mass recovery calculations showed around 100% of injected MWNTs were recovered from the eluent at the end of stage 1 (only one stage) in most of the columns under DI water condition except three columns packed with fine glass bead at pH 5.6 (94%), fine sand at pH 5.6 (77%), and fine sand at pH 10 .0 (89%) ( Table 3 3). The transport model simulated the transport of MWNTs in DI water through the colum ns well ( Figure 3 1). The best fit k values were close to zero for almost all experiments except for the three fine porous medium columns ( Table 3 3). T he Damkohler number for MWNT retention ( where L is the column length and v is flow velocity) was smaller than 0.1 00 for most of the columns under DI water conditions except one fine sand columns at pH 5.6 (0.26 2 ). This result confirms the high mobility of MWNTs in the saturated columns under DI water conditions because only one fine sand column had a retention time scale comparable to that of the advection process. The retention of MWNTs in the three fine porous medium columns under unfavorable conditions could be caused by several factors, such as surface heterogeneity (i.e., c harge and roughness) and/or physical straining [ 124 125 ] I n this work, however, we did not observe no t able MWNT retention in the fine glass beads (same size as the fine sand) at pH 10 .0 ( Figure 3 1), suggesting physical straining might not be a major MWNT retention mechanism. Although the both the sand and glass bead were acid washed to reduce surface coatings, the clean procedure may not sufficient to completely remove the local charge heterogeneity or surface roughness, particularly for the fine porous media [ 125 126 ] As a result, the remaining local charge hetero geneity on the fine porous medium surfaces

PAGE 51

51 could serve as favorable sites for MWNT deposition, although most of the surfaces were unfavorable based on the overall pore water chemistry [ 127 128 ] This mechanism is consistent with the fact that deposition of MWNTs in both fine sand and fine glass bead columns at pH 10 .0 was low er than at pH 5.6 ( Figure 3 1) because increase of solution pH may result in more negatively charged surfaces and thus reduce the affinity of surface charge heterogene ous sites to MWNTs [ 129 130 ] Surface physical heterogeneity such as surface roughness may also play a vital role on the deposition of MWNTs in the fine porous media, particularly in fine sand columns. SEM image analysis result showed that the glass bea ds have much smoother surfaces than sand grain ( Figure 3 2). Meanwhile, the retention of MWNTs in the glass beads was lower than that of the sand, which is consistent with the findings of previous studies that increase in surface roughness may enhance part icle deposition in porous media [ 131 132 ] Effect of IS on MWNT D eposition When the solution IS increased slightly from DI water to 1 mM or 10 mM, the mobility of the MWNTs in the two types of s porous media reduced dramatically and MWNT transport at pH 10 .0 was higher than at pH 5.6 ( Figure 3 1). Those indicate that MWNT deposition and transport in saturated porous media are very sensitive to solution chemistry, particularly to solution IS. Although the DLVO energy profiles in the Supplementary data showed all experimental conditions w ere unfavorable for the attachment of MWNTs, mass recovery calculations indicated strong MWNT deposition during stage 1 in the porous media with recovery rate of 5 61% and 1 42% for 1 mM and 10 mM solution IS, respectively ( Table 3 3). Enhanced deposition of particles in saturated porous media due to increase in solution IS or reduction in solution pH has

PAGE 52

52 been often observed in the literature of fate and transport of nanoparticles and colloids, such as CNTs [ 93 116 ] and natural colloids [ 108 133 ] The failure of the DLVO theory confirmed that it cannot be applied directly to describe the interaction forces between of MWNTs and porous media [ 117 121 ] A new or m odified DLVO theory therefore is needed to advance current understanding of fate and transport of CNTs in porous media [ 117 ] In addition, new technologies of measuring and calculating zeta potential of tubular particles are also in great need because current method may overestimate the actual surface potential of CNTs [ 110 ] Simulations of the transport model also well matched all the MWNT breakthrough curves under higher IS conditions. The best fit k values were between 0.032 and 0.280 min 1 and the Damkohler number was larger than 0. 494 for all columns under the two IS conditions, confirming the importance of deposition to the fate and transport of MWNTs in the porous media. Similarly, the deposition rate ( k ) was very sensitive to the solution chemistry and increased when IS i ncreased or pH reduced. Mobilization of MWNTs The fate of MWNTs in the two saturated porous media also relied on solution IS. When IS was decreased to a lower level (i.e., from 10 mM to 1 mM and/or from 1 mM to DI water), some of the retained MWNTs were released as pulses from the columns with peak concentrations as high as 0.7C 0 ( Figure 3 3). Re entrainment of colloids and nanoparticles under transient flow chemistry conditions have been also observed in many previous studies [ 93 108 ] Mass recovery calculation s indicated that the release of MWNTs in stage 2 or 3 was relatively low and for most of the cases less than 1 1 % of the total injected MWNTs were recovered ( Table 3 3). In total, less than 2 7 % of the retained MWNTs were mobilized in either the one stage (IS 1 mM, stage 2) or two

PAGE 53

53 stages (IS 10 mM, stages 2 and 3), suggestin g that a large portion (>70%) of MWNT deposition in the two porous media was irreversible under all experimental conditions. SEM image analysis of the sand and glass beads surfaces at the end of stage 3 (IS 10 mM) confirmed the presences of MWNTs attached on the surfaces ( Figure 3 2 C and D). The released portion of MWNTs from the two saturated porous media through IS reductions can be considered as a re entrainment of MWNTs from the secondary minimum well. The deposition and re entrainment of particles, i ncluding MWNTs, in the secondary minimum well are considered as a dynamic capture and release process that is believed to be reversible [ 106 134 ] If the particles in the secondary minimum well can overcome the primary ene rgy barrier to reach the primary minimum well, however, their deposition is considered as an irreversible process and the retained particles will not be released with perturbations in solution IS [ 106 ] Thus, the column experimental results presented in this work indicate that MWNTs deposited in the two porous media under the higher IS conditions (1 or 10 mM) were trapped in both the secondary mi nimum and primary minimum wells, but more were in the primary minimum. Because the electrolyte solutions used in this work had relatively low IS (1 and 10 mM), it would be difficult for the MWNT attachment in the porous media occurring directly in the pri mary minimum [ 121 134 ] Previous studies, however, suggested that particles, particularly small sized particles, in deposited in saturated porous media can transit from the secondary minimum to the primary minimum [ 135 136 ] The transit of MWNT deposition from secondary minimum to primary minimum in the two saturated porous media could be trigged by local favorable conditions resulted from the surface heterogeneity [ 116 128 137 ] In addition, because CNTs have an extremely high

PAGE 54

54 aspect ratio, hydrodynamic force may also increase their kinetic energy to override the energy barrier to transfer from secondary minimum to the primary minimu m [ 116 138 ] For instance, for a MWNT attached on a porous medium surface in the second ary minimum well in an end to end position ( Figure 3 4a), flow around the porous medium grain will impose hydrodynamic force on the tubular MWNT and force it to roll along the grain surface ( Figure 3 4b). Because the long length of the MWNT, the hydrodynam ic force may overcome the repulsive force in the primary barrier and move it into the primary minimum well ( Figure 3 4c). Additional investigations are recommended to further determine the effect of hydrodynamic force on CNT deposition and transport in wat er saturated porous media. Chapter Conclusions Laboratory column experiment results demonstrated that solution IS and pH played important roles on deposition, transport and mobilization of MWNTs in saturated porous media. Although MWNTs were highly mobile in the columns under DI water conditions, slight increase of solution IS to 1 or 10 mM dramatically reduced their mobility in sage 1. Mass recovery calculations indicated only about 5 61% and 1 42% of MWNTs were transported through the columns at IS of 1 and 10 mM, respectively. Subsequent experiment at stages 2 and 3 showed only small portion (<2 7 %) of deposited MWNTs was released when solution IS reduced. Although the experimental conditions are unfavorable, the mobilization experimental result suggest ed that more MWNTs were deposited in the primary minimum than in the secondary minimum. The MWNTs trapped in primary minimum might migrate from the secondary minimum, where weak associated MWNT particles could be driven by hydrodynamic force to roll along the grain surface, override the energy barrier, and further deposit in the primary

PAGE 55

55 minimum. Comparisons of experimental results in sand and glass bead columns indicated that surface heterogeneity (charge and roughness) may also play a vital role in control ling the deposition, transport, and mobilization in saturated porous media.

PAGE 56

56 Table 3 1 Summary of conditions and procedures of the column experiment s Treatments Deposition Mobilization Stage 1 Stage 2 Stage 3 DI water (pH 5.6 or 10.0) CNT with background solution 1 mM IS (pH 5.6 or 10.0) CNT with background solution DI water at the same pH 10 mM IS (pH 5.6 or 10.0) CNT with background solution 1 mM solution at the same pH DI water at the same pH

PAGE 57

57 Table 3 2 Physicochemical characteristics of MWNTs and porous media used in the column experiment. Size (mm) Electrophoretic mobility (10 8 m 2 /(Vs)) Zeta potential (mV) pH=5.6 pH=10.0 pH=5.6 pH=10.0 DI water Glass bead 0.5 0.6 3.3 4.3 44.0 58.0 0.1 0.2 3.1 3.8 41.6 52.1 Sand 0.5 0.6 1.5 3.3 19.7 45.2 0.1 0.2 1.5 3.0 19.6 40.5 MWNT 2.8 2.9 38.2 39.1 1 mM IS Glass bead 0.5 0.6 3.0 3.6 40.0 48.3 0.1 0.2 2.9 3.4 38.8 46.4 Sand 0.5 0.6 1.4 2.8 18.9 38.5 0.1 0.2 1.4 2.7 18.4 35.9 MWNT 2.2 2.3 29.8 31.4 10 mM IS Glass bead 0.5 0.6 2.6 3.4 35.8 45.5 0.1 0.2 2.6 3.3 35.4 45.2 Sand 0.5 0.6 1.2 2.3 15.8 30.5 0.1 0.2 1.2 2.2 15.8 29.2 MWNT 1.0 1.3 13.5 17.5

PAGE 58

58 Table 3 3 Summary of mass recovery and model results of MWNT transport in saturated porous media with different combinations of solution pH, IS, and porous medium type and size. Treatments Mass recovery (%) Breakthrough model ** Size (mm) pH Deposition M 1 Mobilization k (min 1 ) D a R 2 M 2 M 3 M T DI water Glass bead 0.5 0.6 5.6 101.5 0.000 0.000 0.990 0.5 0.6 10.0 100.6 0.000 0.000 0.997 0.1 0.2 5.6 93.8 0.004 0.046 0.993 0.1 0.2 10.0 98.6 0.000 0.000 0.993 Sand 0.5 0.6 5.6 95.7 0.002 0.065 0.997 0.5 0.6 10.0 99.0 0.000 0.000 0.985 0.1 0.2 5.6 76.3 0.017 0.262 0.979 0.1 0.2 10.0 88.9 0.005 0.078 0.945 1 mM IS Glass bead 0.5 0.6 5.6 44.6 4.4 8.0 0.048 0.802 0.986 0.5 0.6 10.0 59.3 10.9 26.8 0.032 0.541 0.985 0.1 0.2 5.6 42.4 3.2 5.6 0.051 0.790 0.964 0.1 0.2 10.0 60.8 6.8 17.4 0.032 0.494 0.955 Sand 0.5 0.6 5.6 37.7 1.8 2.9 0.058 0.969 0.984 0.5 0.6 10.0 55.4 9.0 20.2 0.038 0.638 0.974 0.1 0.2 5.6 5.5 10.1 10.7 0.171 2.693 0.808 0.1 0.2 10.0 48.5 7.5 14.5 0.042 0.659 0.964 10 mM IS Glass bead 0.5 0.6 5.6 25.2 1.6 5.6 9.6 0.080 1.338 0.988 0.5 0.6 10.0 41.6 4.8 10.1 25.6 0.051 0.858 0.976 0.1 0.2 5.6 15.5 1.2 2.1 3.9 0.119 1.844 0.943 0.1 0.2 10.0 35.1 4.5 8.7 20.4 0.059 0.914 0.980 Sand 0.5 0.6 5.6 22.9 0.0 2.1 2.7 0.088 1.439 0.968 0.5 0.6 10.0 35.9 3.6 10.6 22.1 0.065 1.063 0.943 0.1 0.2 5.6 1.3 0.8 17.8 18.9 0.280 4.410 0.831 0.1 0.2 10.0 25.7 2.9 6.1 12.0 0.069 1.083 0.964 M ass recovery of injected of CNTs in porous media for three different stages ( M 1 M 3 ) were given. M TM represents the total mass recovery of mobilization from the deposited CNTs. ** Breakthrough model fitting parameters were given as deposition rate coefficient (k) Damkohler number (D a ) and coefficient of determination ( R 2 )

PAGE 59

59 Figure 3 1 Deposition of MWNTs in saturated porous media at stage 1 with different combinations of solution pH, IS, and porous medium type and size.

PAGE 60

60 Figure 3 2 SEM images of ( A ) acid cleaned glass beads, ( B) acid cleaned sand, ( C ) glass beads with MWNTs attached on their surface, and (D) sand with MWNTs attached on their surface.

PAGE 61

61 Figure 3 3 Effect of solution IS on MWNT mobilization in saturated porous media at stages 2 and 3 with different combinations of solution pH, IS, and porous medium type and size.

PAGE 62

62 Figure 3 4 Schematic diagram of migration of a MWNT from secondary minimum to primary minimum due to hydrodynamic force.

PAGE 63

63 CHAPTER 4 4 HIGH MOBILITY OF SDBS DISPERSED SINGLE WALLED CARBON NANOTUBES IN SATURATED AND UNSATURATED POROUS MEDIA 1 Introduct ory Remarks Carbon nanotubes (CNTs) are among the top concern of engineered nanoparticles (ENPs) in the environment [ 9 ] Entirely composed of carbon, CNTs are rolled up graphene sheets with exceptional mechanical, electrical, optical, and thermal properties [ 12 ] They are largely used in many novel applications in nanotechnology, electronics, optics, thermal conductors, and other fields in material sc ience and engineering [ 16 ] Recent studies of the environmental impacts of CNTs, however, have revealed their toxic effects towards a variety of living organisms and highlighted the potential risks of CNTs to the ecosystem [ 25 26 ] The widespread application and exponential growth in production of CNTs may result in their release into soils. This release imposes contamination risks on groundwater if the CNTs can move through the vadose zone to reach the water Table [ 29 ] As a result, there is increased concern among the scientific communit y and regulation agencies over the fate and tra nsport of CNTs in porous media In addition to their own potential contamination risks, the presence of mobile CNTs in soils and groundwater may also alter the transport behavior of other contaminants. With extremely small sizes and superior surface affiliation abilities, the mobile CNTs in soils can facilitate the transport of a variety of contaminants previously thought to have very limited mobility, such as heavy metals [ 73 ] and organic pollutants [ 139 ] 1 Reprinted with permission from Tian, Y., Gao, B., and Ziegler, K.J., High mobility of SDBS dispersed single walled carbon nanotubes in saturated and unsaturated porous media. Journal Of Hazardous Materials, 2011. 186 (2 3): p. 1766 1772.

PAGE 64

6 4 Facilitated/enhanced transport of reactive contamin ants in soils by natural nano sized and colloidal particles is a well known phenomenon, which has been observed in many previous studies [ 140 141 ] A number of investigations have been conducted to study the environmental fate and transport of ENPs [ 27 118 142 ] Column studies with water saturated porous media have been commonly used to explore the governing mechanisms of CNT fate and transpo rt in groundwater systems [ 93 110 143 144 ] Experimental results from columns packed with well defined and natural porous media suggest that surface retention and physical trapping could play important role s in controlling the fate and transport of CNTs in water saturated porous media [ 30 143 ] Although progress has been made toward a better understanding of the environmental fate and transport of ENPs, the current knowledge of the fate and transport of CNTs in por ous media is far from complete. Reliable data on the transport behavior and interaction dynamics of CNTs in porous media are still limited [ 30 ] This is particularly true for unsaturated porous media; to our knowledge, no research has been conducted to examine the transport behavior of CNTs in unsaturated systems. Due to the addition of an air phase in the unsaturated system s, it is expected that the transport behavior of CNTs could be much more complicated in unsaturated porous media than in saturated porous media. Extensive work has been done in recent years to determine the transport mechanisms of colloidal particles in unsaturated porous media. Unsaturated porous medium columns are often used in the laboratory to examine the transport behavior of different types o f colloids, including viruses, bacteria, clay particles, silicates, and synthetic microspheres [ 65 145 146 ] In addition, microscopic visualization

PAGE 65

65 technologies have been adapted in pore scale investigat ions to determine the retention mechanisms of colloids in unsaturated porous media [ 61 119 ] Findings from these investigations have revealed several important mechanisms that govern the fate and transport of colloidal particles in unsaturated porous media. Fo r a well defined homogeneous porous media, it is recognized that the retention of colloidal particles within the unsaturated system is mainly affected by the following mechanisms: a) porous medium grain surface (solid water interface) deposition, which is controlled by the physicochemical and electrostatic forces between the particle and grain surfaces [ 54 ] ; b) air water interface deposition, which is controlled by the physico chemical and electrostatic forces between the particle surface and the air water interface [ 61 ] ; and c) physical trapping, including wedging, pore straining, and water film straining, which is controlled by the collective contributions from particle size, porous medium grain size, and moisture content [ 147 149 ] T here are no t able differences between CNTs and natural colloids in many physicochemical properties particularly in res pect to aspect ratio and surface properties [ 98 ] It therefore needs to be investigated whether these well documented colloid retention mechanisms are applicable to describe the fate and transport of CNTs in unsaturated porous media. Because CNTs have a strong tendency to form bundles, they are essentially insoluble in either aqueous or organic phase. It is thus often required to disperse or solubilize CNTs for their optimum uses. Stable aqueous CNT disper sions have been prepared primarily through the use of surfactants or amphiphilic polymers to solubilize them or through strong oxidative treatments to introduce hydrophilic functionalities to the CNT surfaces. For instance, SWNT dispersion is now a routine procedure using

PAGE 66

66 surfactant stabilization [ 99 150 ] It is conceivable that surfactant induced changes in SWNT surface properties could bring drastic changes in terms of transport behavior. Well dispersed SWNTs should be more mobile in soil and water systems, therefore, of greater environmental concerns. In this work, laboratory experiments were conducted to investigate the transport of single walled carbon nanotubes (SWN Ts) in unsaturated porous media An ionic surfactant, sodium dodecylbenzene sulfonate (SDBS), was used to stabilize the SWNTs to enhance their dispersions in water. High energy sonication was used to aid in the dispersion, which may shorten the dispersed SWNTs. T he SDBS surfactant has been used in several studies to stabilize SWNTs and examine their environmental fate and transport [ 26 27 118 144 ] L aboratory columns packed with quartz sand with different grain size and moisture content combinations were used to examine the breakthrough behaviors of the SDBS dispersed SWNTs. In addition, bubble column experimen ts were used to examine the interactions between the SDBS dispersed SWNTs and air water interface. Our overarching objective was to explore the mechanisms governing the fate and transport of the SDBS dispersed SWNTs in porous media, particularly under unsa turated conditions. Specific objectives were to: 1) examine how changes in moisture content affect the retention and transport of the SDBS dispersed SWNTs in porous media; 2) determine the effects of porous medium sizes on the retention and transport of th e SDBS dispersed SWNTs; and 3) model the retention and transport of the SDBS dispersed SWNTs in water saturated and unsaturated porous media.

PAGE 67

67 Materials and Methods SWNTs HiPco SWNTs obtained from the Richard Smalley Institute at Rice University (Houston, TX) were used in this study (Rice HPR 162.3). The SWNT suspensions were prepared by mixing 20 mg of raw SWNTs with 200 mL of an aqueous SDBS solution (1% [ 99 ] High shear homogenization (IKA T 25 Ultra Turrax) for 1.5 2.0 h and ultrasonication (Misonix S3000) for 10 min were used to aid dispersion The SWNT suspension was then ultracentrifuged at 20,000 rpm (Be ckman Coulter Optima L 80k) for 5 h to minimize the presence of aggregates Previous studies have demonstrated that this method can create a surfactant shell on the sidewall of the nanotubes to stabilize the SWNT suspensions [ 99 ] In addition, high power sonication could also shorten the SWNTs and create defects to form surface functional groups (e.g., COOH) [ 151 ] The presences of surfactant shell and functional groups introduced negative surface charges, and thus the Z eta potential of the SWNT was 42.84 mV as determined by the ZetaPlus (Brookhaven Instrument Co., Holtsville, NY). The average diameter and length of the dispersed SWNT was about 0.8 and 400 nm mea sured by atomic force microscopy [ 152 ] Porous Media Quartz sand (Standard Sand & Silica Co.) was used in this study and was sieved into three different size distributions: fine (0.1 0.2 mm), medium (0.5 0.6 mm), and coarse (1.4 1.6 mm). The s and was then washed sequentially by tap water, 10% nitric acid (v:v) and deionized water, and th en baked at 550 C to remove the metal oxides and organic impurities. The zeta potential of the acid cleaned quartz sand was

PAGE 68

68 determined according to the procedure of Johnson et al. [ 94 ] In brief, a mixture containing 100 g of acid cleaned quartz sand and 200 mL DI water was ultrasonicated for 30 minutes. Aliquots of the quartz colloids were then removed and filtered through a 0.45 m filter. The filtrate was analyzed for electrop horetic mobility. The zeta potential for the sand was about 1 9.73 mV Sand column E xperiments Laboratory columns packed with acid cleaned sand were used to investigate the transport of SWNT suspensions in porous media with different grain size and moistu re content combinations ( Table 4 1). The quartz sand (i.e., fine, medium, or coarse) was wet packed (i.e., saturated) into an acrylic column measuring 2.5 cm in diameter and 15 cm in height. Stainless steel membranes with 50 m pores (Spectra/Mesh, Spectr um Laboratories, Inc.) were used at the column inlet and outlet to distribute the flow and to maintain the capillary pressure. The membranes were in immediate contact with the porous media and sealed by rubber O rings. Six vent holes were drilled on opposi te sides at 3, 7.5, and 12 cm from the top of the column. The vent holes were sealed with gas permeable porous PTFE membranes (Milliseal Disk, Millipore) to allow air to enter under unsaturated conditions. To pack a saturated sand column, a small amount o f sand was poured gently into 8 ml of DI water standing at the bottom of the column until the sand surface was 0.5 to 1 cm below the water level. A polypropylene stir rod was used to stir the sand in the column. Approximately 8 mL DI water was then added to the column and the sides of the column were gently tapped a uniform number of times. This procedure was repeated several times until the column was packed to a height of 15 cm. The porosities of the columns packed with fine medium, and coarse sized s and were 0.38,

PAGE 69

69 0.40, and 0.40, respectively. For each experiment, a fresh column was packed and was placed on a Digital Bench Scale to monitor the column weight. The moisture content of the entire unsaturated column was then determined by the gravimetric m ethod. Two peristaltic pumps (Masterflex L/S, Cole Parmer) at the column inlet and outlet were used to regulate the sand pack moisture content and to control the downward flow of water and SWNTs through the unsaturated columns. To prepare for the unsatur ated transport experiment, an initially saturated column was drained by elevating the outflow rate 5% higher than the inflow rate. Approximately 4 h were required to drain the column to reach the target moisture content, whereupon a unit hydraulic head gra dient was established throughout the sand pack by equalizing the inflow and outflow rates. A total of 18 columns were used in the experiment with different combinations of moisture content and sand size ( Table 4 1). Once the column moisture stabilized, a n SDBS solution (1% by weight) was first pumped through the column for about three pore volumes (PVs) to stabilize the bulk fluid chemistry. The breakthrough experiment was then initiated by switching from SDBS solution to SWNT suspension. The SWNT suspensi on was applied to the column for more than three PVs. The E ffluent samples were collected from the bottom of the column using a fraction collector (IS 95 Interval Sampler, Spectrum Chromatography, Houston, TX) during the breakthrough experiment to analyze SWNT concentrations with a UV Vis Spectrophotometer (Evolution 60, Thermo Scientific) at a wavelength of 245 nm [ 118 ] Although stability study showed that SDBS dispersed SWN T concentrations remained almost unchanged after 24 hrs ( Figure B 1), the initial and final SWNT

PAGE 70

70 concentrations in the influent reservoirs were monitored for all experiments to ensure the data quality. Only one pump was used at the outflow end of the colu mn to regulate the flow rate in experiments with the water saturated column. The influent concentrations of SWNTs (16 mg/L) used in the saturated experiments matched those of the unsaturated experiments. The procedures of sample collection and analysis fo r saturated experiments were the same as those for the unsaturated experiments. Bubble column Ex periments Bubble column experiments were used to examine the interactions between the SWNT particles and the air water interface. The design and procedure of the bubble column experiments were similar to those of Wan et al. [ 153 ] In brief, an acrylic column measuring 2.5 cm in diameter and 100 cm in height was filled with the SDBS dispersed SWNT suspension. The bottom of the bubble column was carved into a small chamber and sealed with two stainless steel membranes with 50 m pores. Eight sampling ports sealed with stoppers were evenly distributed along the length of the column. Air was pumped into the chamber at the bottom of the column through the stainless steel membranes, resulting in relatively uniform air bubbles rising from the bottom to the top. A flow control system was used to control the air flow rate to be at 5 mL min 1 Water samples were obtained fro m the eight sampling ports at different time intervals to determine the temporal and spatial distribution of SWNTs in the bubble column. Results and Discussion DLVO Ene rgy The Derjaguin Landau Verwey Overbeek ( DLVO) theory was used to calculate the interac tion forces between the SWNTs and porous media, and between the SWNTs and

PAGE 71

71 air water interface (see supporting information). The DLVO theory was originally developed for spherical particles, which might not be sui Table 4 for this study. To our knowledge, ho wever, no theory has been developed to determine the interaction energy between a tubular shaped particle, such as SWNT, and a flat surface. Recent studies suggested that using either the length or diameter as an effective size may provide an alternative w ay to explore the interactions between carbon nanotubes and porous media [ 118 143 ] In this study, the DLVO energy was estimated using either the tube length (400 nm) or the tube diameter (0.8 nm) as an effective size [ 118 143 ] When the tube diameter was used in the calculations, all the energy profiles suggest repulsive inte ractions between the SWNTs and the three types of sand, as well as between the SWNTs and air water interface; however, the energy barriers were very low ( Figure 4 1). When the tube length was used, the repulsive energy barriers increased dramatically for b oth the three porous media and air water interface ( Figure 4 1). The DLVO energy profiles between the SWNT and three types of sand showed secondary energy minima at a separation distance around 20 nm ( Figure 4 1) The secondary energy minimum was not obse rved between the SWNTs and the air water interface. In general, although the DLVO energy profiles of the two effective sizes (i.e., length and diameter) were very different, both cases suggested that the experimental condition s w ere unfavorable for the dep osition of SWNTs in the primary energy minima onto either sand grain surface or air water interface. The energy profiles, however, could not be used to conclusively determine if the SWNTs could be deposited onto sand grain surface in the secondary energy m inima. Further investigation is recommended to

PAGE 72

72 modify the DLVO theory or develop a new theory to calculate the DLVO energy for tubular shaped particles. SWNT Tran sport in Medium s and Co lumn s Transport of SWNTs in columns packed with medium sized quartz sand displayed the typical colloid breakthrough characteristics ( Figure 4 2a). After applied to the column, SWNTs were first detected in the effluent at less than 1 PV, suggesting that dispersion contributes to the transmission of SWNTs through the column. The breakthrough concentrations then increased quickly and plateaued. The plateau concentrations (C/C 0 ) for all the conditions tested were close to one, suggesting low SWNT retention in the medium sized porous media under the three moisture content condit ions. The high mobility of SWNT suspensions in the saturated, medium sand columns, to a certain extend was consistent with the predictions of the DLVO theory, which suggest the deposition of SWNTs onto the porous medium surface is unfavorable, particularl y with respect to primary minimum deposition ( Figure 4 1). In addition to surface deposition, previous studies have suggested that SWNTs could also be retained in saturated porous media through pore straining [ 30 93 ] In this study, however, the breakthrough curves did not display no t able pore straining of the SWNT suspensions. This might be attributed to the grain size (medium sand), which may not be small enough to construct the narrow pores required to retain the nanotubes. The high mobility of both SWNT suspensions in the unsaturated, medium sand columns is contradictory to the prediction of the current theories of colloid deposition in porous media (i.e., air water interface deposition and water film straining). The three breakthrough curves almost overlapped with each other ( Figure 4 2a), suggesting that

PAGE 73

73 the reduction in moisture content had little effect on the transport of the SDBS dispersed SWNT suspensions in the medium sand columns. SWNT Tra nsport in Fi ne sand C olumn s The breakthrough responses of the SWNT suspensions in the fine sand columns under three moisture conditions ( Figure 4 2b) resembled those in the medium sand columns. The pl ateau concentrations for all the experiments were also close to one. In addition, the breakthrough curves for all the experiments were similar to each other. These results also suggest that the retention of the SWNT suspensions in the fine sand columns was low. Similar to their transport in medium sand columns, SWNT transport in the fine sand columns was not sensitive to the moisture perturbations. It seems that neither air water interface deposition nor water film straining were important to SWNT transpor t in unsaturated porous media under these experimental conditions. The low retention of the SWNTs in those column through film straining could probably be attributed to the fact that moisture contents in the fine and medium sand columns (particularly fine sand) were still relatively high ( Table 4 1). The transport experiments in fine sand columns were also designed to examine the pore straining mechanisms of SWNTs in porous media because the pore sizes of the fine porous media were small. If physical trap ping of SWNTs was controlled by their lengths, we might be able to observe their retention in the fine sand columns through pore straining. When SWNTs are transmitted in porous media, however, they might orient parallel to streamlines and thus reduce the p hysical trapping [ 154 ] Additional investigations are still needed to modify and improve the pore straining theory, which were developed for spherical particles, to better predict the straining and transport of SWNTs in porous media.

PAGE 74

74 SWNT Tra nsport in Co arse sand C olumn s SWNT s showed high mobility in saturated, coarse sand columns ( Figure 4 2 c ) and their transport was similar to that in the fine and medium sand columns. In unsaturated columns, however, SWNT suspensions were transmitted to a lesser extent ( Figure 4 2c), indica ting that the air phase in the unsaturated columns increased the retention of SWNTs. This increased SWNT retention could not be attributed to the deposition onto the air water interface because it had little influence on SWNT retention and transport in the medium and fine sand columns. This was further confirmed by the bubble column experimental results, as discussed later. The water film straining mechanism, however, may play an important role in affecting the transport of SWNTs in the unsaturated coarse sand columns. According to the film straining theory, the straining of particles in unsaturated porous media must the connections between pendular rings in the por ous media are broken; and 2) water film thickness is smaller than the particle sizes [ 65 ] For colloidal or nano sized particles in porous media, therefore, water film straining only occurs at low moisture conditions. In this study, the moisture content in the unsaturated, coarse sand were very low (i.e., 0.07 and 0.10), which may be sufficient to cause the formation of thin water film to trap the SWNTs. On the other hand, the moisture content was much higher in the unsaturated, fine (i.e., 0.29 and 0.34) and medium (i.e., 0.14 and 0.21) sand columns ( Table 4 1). SWNT Tran sport in Bubble Co lumn s SWNT concentrations at the eight sampling ports within the bubble column stayed unchanged for more than 20 hours with continuous air bubble injections ( Figure 4 3).

PAGE 75

75 The result was consistent with findings from both the DLVO energy calculations and the sand column studies that showed the experimental con ditions are not favorable for the SDBS dispersed SWNT suspensions to deposit onto the air water interface. Modeling SWNT Tr ansport in the Sand C olumns Transport of SWNTs in porous media can be described by the advection dispersion equation coupled with kinetic expressions to describe particle retention. If we assume that retention of SWNTs in the sand columns was mainly controlled by surface deposition ( i.e., grain surface and air water interface) and physical trapping (i.e., wedging, pore straining, and film straining), the governing equation then can be written as: ( 4 1) where C is the concentration of suspended SWNTs in pore water (M L 3 ) t is time (T), b is the medium bulk density (M L 3 ), is the dimensionless volumetric moisture content, C d and C t are the concentrations of SWNTs retained by surface deposition and physical trapping, respectively (M M 1 ), z is the distance traveled in the direction of flow (L), D is the dispersion coefficient (L 2 T 1 ), and v is the average linear pore water velocity (L T 1 ). T he retention of colloidal and nano sized particles by surface deposition ( C d ) or trapping ( C t ) can b e described by different kinetic expressions. In this study, the surface deposition of SWNTs in the columns was small and it can be approximated as first order kinetics [ 155 ] Examination of the breakthrough curves suggests that physical

PAGE 76

76 trapping (mainly film st r aining) o f SWNTs in the columns could be described by the Elovich equation [ 156 ] As a result, the retained SWNT concentrations can be written as: ( 4 2) ( 4 3) where k 1 and k 2 are the rate constants for SWNT retention at the grain surfaces and in the thin water film (T 1 ), and is a constant reversely proportional to the trapping capacity, which controls the exponential decline of the reaction rate when the amount of SWNTs tra pped in the porous media increase (M 1 M). The Elovich equation (i.e., equation 3) can be approximated as first order kinetics (equation 2) when is very small (i.e., the physical trapping capacity is huge). Equation 1 3 can be solved numerically for a ze ro initial concentrations, a pulse input boundary condition at the column inlet, and a zero concentration gradient boundary condition at the outlet. Simulations of the mathematical model were applied to describe SWNT transport in both water saturated and unsaturated sand columns. For most of the cases, the physical trapping terms were turned off by setting k 2 to zero because film straining was only observed in the unsaturated, coarse sand columns. In general, t he model reproduced all the breakthrough chara cteristics closely ( Figure 4 2), with computation of R 2 exceeding 0.99 ( Table 4 1). The model estimated dispersion coefficients ( D ) ranged from 0.14 to 6.33 cm 2 min 1 ( Table 4 1). Calculations of the Peclet number (

PAGE 77

77 where H is the height of the column) ranged from 17.62 to 95.38, confirming the dependency of the SWNT transport to dispersion within the column. SWNT retention in all the medium and fine sand columns as well as the saturated, coarse sand columns were quantified with the first order kinetics (i.e., equation 2). Best fit estimates of the deposition rate coefficient, k 1 were very small and ranged from 0.0003 to 0.0078 min 1 ( Table 4 1), corresponding to the observed high mobility of SWNTs in those columns. Calcul ations of the Damkohler number for deposition ( ) ranged from 0.011 to 0.045, indicating the time scale was much smaller for surface deposition than for advection through the columns. Because SWNT retention and transport in the unsaturated, coarse sand co lumns were controlled by both surface deposition and water film straining, the coupled equations (1 3) were used to quantify the reaction rate coefficients. The best fit values of the straining rate coefficients ( k 2 ) were several orders of magnitude larger than the deposition rate coefficients ( k 1 ) ( Table 4 1). This result indicates that retention of SWNTs in the unsaturated, coarse sand columns might be dominated by the film straining process. The Damkohler numbers for SWNT physical trapping ( ) were 0.61 to 2.99, implying the film straining rates were comparable to their advective transport in the columns. The model estimated trapping capacity factor, increased from 7.15 kg/mg to 9.46 kg/mg when the moisture content decreased from 0.10 to 0.07, probably because decreases in moisture content increased the fraction of the column exhibiting thin water film behavior.

PAGE 78

78 Chapter Conclusions We found that SDBS dispersed SWNTs were highly mobile in porous media under most of the experimental conditions tested. Deposition of the SWNTs onto the grain surface and air water interface was low because all the interactive surfaces were negatively char ged. Film straining of the SWNTs in the porous media only occurred at very low moisture content at which the thickness of the water film on the grain surfaces are sufficiently thin to trap the nanotubes. Pore straining of SWNTs was not observed in the poro us media across a wide range of size distributions due to the fact that SWNTs may orient parallel to streamlines to reduce their retention. It can be concluded that SDBS dispersed SWNTs, when released into the soil, might have a great impact on the groundw ater quality due to their high mobility. Our results also suggest that mathematical models based on the advection dispersion equation coupled with different reaction rate laws could be used as a monitoring tool to predict the fate and transport of SWNTs in the vadose zone and groundwater.

PAGE 79

79 Table 4 1 Summary of experimental conditions and model results. No. Sand Size Pore Velocity (cm/min) Mass Recovery D (cm 2 min 1 ) Pe k 1 (min 1 ) k 2 (min 1 ) (kg mg 1 ) Da 1 /Da 2 R 2 1 Medium 0.40 0.49 1.01 0.35 23.31 0.0003 0.011 1.00 2 Medium 0.21 3.73 0.99 0.81 29.56 0.0028 0.032 1.00 3 Medium 0.14 1.40 0.99 3.52 17.62 0.0051 0.022 1.00 4 Fine 0.38 0.52 0.98 0.14 59.32 0.0004 0.012 1.00 5 Fine 0.34 2.30 0.98 0.19 59.90 0.0011 0.026 1.00 6 Fine 0.29 0.68 0.97 0.41 95.38 0.0040 0.028 1.00 7 Coarse 0.40 1.96 0.98 0.19 41.21 0.0013 0.045 1.00 8 Coarse 0.10 7.83 0.95 4.36 10.43 0.0021 0.50 7.15 0.013/2.99 1.00 9 Coarse 0.07 2.80 0.93 6.33 18.58 0.0078 0.26 9.46 0.018/0.61 1.00

PAGE 80

80 Figure 4 1 DLVO energy between SDBS dispersed SWNT and (a) medium sand, (b) fine sand, (c) coarse sand, and (d) air water interface.

PAGE 81

81 Figure 4 2 Transport of SDBS dispersed SWNTs in sand columns under different volumetric moisture content conditions.

PAGE 82

82 Figure 4 3 SWNT concentrations at the eight sampling ports within the bubble columns measured at different time intervals.

PAGE 83

83 CHAPTER 5 5 EFFECT OF SURFACE MODIFICATION ON SINGLE WALLED CARBON NANOTUBE RETENTION AND TRANSPORT IN GRANULAR POROUS MEDI A Introduct ory Remarks Because of their exceptional electrical, mechanical and thermal properties, carbon nanotubes (CNTs) have been widely manufactured and used in many applications [ 16 ] This may result in unintended and/or accidental releases of CNTs into the natura l environment. Upon release, CNTs could be transported by water flow through the vadose zone and cross the water table to reach the groundwater and drinking water aquifers. Several studies have found that bioaccumulation and bioaugumentation of CNTs by the food chain may impose great risks to the ecosystems and the public health [ 23 15 7 ] Because CNTs have a strong affiliation to many contaminants, such as heavy metals and organic compounds, their potential risks in soil and groundwater systems could be aggravated [ 73 158 ] Therefore, fate and transport of CNTs in po rous media need thorough investigations. Pristine CNTs have large surface areas and strong van der Waals attraction forces among individual particles, inevitably causing their unwanted self aggregations in aqueous solutions Various surface modification methods thus have been developed to enhance CNTs dispersion in solutions to facilitate their applications. The most common CNT surface modification methods include either strong oxidative treatments or use of amphiphilic polymers or surfactants. Surface ox idization with strong acids can introduce hydrophilic functional groups (e.g. carboxyl and hydroxyl groups) to the CNTs surfaces to help their dispersion [ 9 ] Amphiphilic polymers or surfactants, such as natural organic matter (NOM) treatment and ionic surfactants, can wrap around and solubilize CNTs in aqueous solutions [ 159 ] The surface modification methods in general can introduce

PAGE 84

84 surface charges to the CNTs and thus electrostatic repulsion forces to reduce their aggregation in aqueous solutions. As a result, dispersed or solubilized CNTs may have a higher mobility in soil and groundwater systems than the pristine CNTs. Although several studies have been conducted to examine the fate and transport of various types of CNTs in porous media [ 142 160 163 ] little research effect has been made to determine the effect of different surface modifications on CNT retention and transport in porous media, partic ularly under unsaturated conditions. In this work, laboratory experiments were performed to investigate the retention and transport of surface modified single walled carbon nanotubes (SWNTs) in porous media. Three types of SWNTs with different surface mo difications were tested in this work: acid oxidized, humic acid (HA) dispersed, and sodium dodecylbenzene sulfonate (SDBS, anionic surfactant) dispersed. Stabilities of the three types of SWNTs in aqueous solutions were compared. Columns packed with quartz sand with different combinations of grain sizes and moisture contents were used to compare the breakthrough behaviors of the surface modified SWNTs. In addition, bubble column experiments were conducted to determine the interactions between the modified S WNTs and air water interfaces. The overarching objective of this work was to determine the effect of different surface modification methods on the retention and transport of SWNTs in saturated and unsaturated porous media. Specific objectives were to: 1) evaluate effectiveness of the three surface modification methods in stabilizing SWNTs in aqueous solutions; 2) determine the interactions between the three types of SWNTs and the air water interfaces; 3) compare the retention and transport of the three typ es of SWNTs in

PAGE 85

85 saturated porous media of different grain sizes; and 4) compare the retention and transport of the three types of dispersed SWNTs in unsaturated porous media of different combinations of moistures and grain sizes. Materials and M ethods Su rface Mo dified SWNTs Pristine SWNTs were obtained from Cheap Tubes Inc. (Brattleboro, VT) with diameters of 1.0 2.0 nm and lengths of 0. 5 2 The pristine SWNTs were used in this work to make all the three surface m odified SWNTs. Surface oxidized SWNTs (O SWNTs) were prepared from the pristine SWNTs using a 3 to 1 ratio of sulfuric and nitric acids to add hydrophilic surface functional groups (e.g., hydroxyl and carboxyl). To prepare the suspensions for experimentat ion, 16 mg O SWNTs were suspended in 1000 ml deionized (DI) water, followed by 30 min of ultrasonication (Misonix S3000). Surfactant and HA dispersed SWNTs (S SWNTs and H SWNTS) were prepare d using of SDBS and HA solutions, respectively. SDBS (25155 30 0) and HA (Elliott Soil HA Standard, 1S102H) were obtained from Sigma Aldrich Co. (St. Louis, MO) and International Humic Substances Society (St. Paul, MN, USA), respectively. The SWNT suspensions were prepared by mixing 20 mg of pristine SWNTs with 200 mL o f the SDBS solution (1% by weight) or the HA solution (5 mg/L), followed the method [ 99 ] High shear homogenization (IKA T 25 Ultra Turrax) for 1.5 2.0 h and ultraso nication (Misonix S3000) for 10 min were used to aid dispersion. The resulted SWNT suspensions were then diluted to a concentration of 16 mg/L for future uses.

PAGE 86

86 Electrophoretic mobility (EPM) of the surface modified MWNTs was determined using a ZetaPlus (B rookhaven Instrument Co., Holtsville, NY) and the Smoluchowski's model was used to estimate their zeta potential [ 142 ] Fourier transform infra red (FTIR) analysis was conducted to characterize the surface functional groups of the pristine and oxidized SWNTs. To obtain observable adsorption spectra, each of the SWNT s was mixed with potassium bromide powder at a weight ratio of 0.1% and then the mixtures were pressed into pellets. The spectra were measured using a Bruker Vector 22 FTIR spectrometer (OPUS 2.0 software). Porous Media Quartz sand (Standard Sand & Silica Co.) was used in this study and was sieved into three different size distributions: fine (0.1 0.2 mm), medium (0.5 0.6 mm), and coarse (1.4 1.6 mm). The sand was then washed sequentially by tap water, 10% nitric acid (v:v) and deionized water, and th en baked at 550 C to remove the metal oxides and organic impurities. The zeta potential of the acid cleaned quartz sand was determined according to the procedure as reported by Johnson et al. [ 94 ] Bubble Column Experiment Bubble column experiment was used to examine the interactions between the SWNT particles and the air water interface. In brief, an acrylic column measuring 2.5 cm in diameter and 100 cm in height was filled with each of the different surface modified SWNT suspensions Air was pumped into the chamber at the bottom of the column, resulting in relatively uniform air bubbles rising fro m the bottom to the top. SWNT samples were obtained at different time intervals from the eight sampling ports positioned at different depths to determine the temporal and spatial distribution of

PAGE 87

87 SWNTs in the bubble column. Details of the design and procedu re of the bubble column experiment were reported previously [ 160 ] Sand Column Experiment Column studies were performed to investigate the retention and transport of the surface modified SWN T suspensions in porous media with different grain size and moisture content combinations ( Table 5 2 ). The quartz sand (i.e., fine, medium, or coarse) was wet packed (i.e., saturated) into an acrylic column measuring 2.5 cm in diameter and 15 cm in height. Six vent holes were drilled on opposite sides at 3, 7.5, and 12 cm from the top of the column. The vent holes were sealed with gas permeable porous PTFE membranes (Milliseal Disk, Millipore) to allow air to enter under unsaturated conditions. The moisture content of the unsaturated column was then generated by the gravimetric method. Two level unsaturated conditions were generated for fine, medium, and coarse sand, respectively. Moisture contents in the saturated and unsaturated columns were shown in the T able 5 2 Detailed information about column preparations can be found in previous study [ 160 ] Once the column moisture stabilized, background solution of the tested SWNT suspen sion was first pumped through the column for about three pore volumes (PVs) to stabilize the bulk fluid chemistry. The breakthrough experiment was then initiated by switching from background solution to the SWNT suspension. The SWNT suspension was applied to the column for more than three PVs. The Effluent samples were collected from the bottom of the column using a fraction collector (IS 95 Interval Sampler, Spectrum Chromatography, Houston, TX) during the breakthrough experiment to analyze SWNT concentrat ions with a UV Vis Spectrophotometer (Evolution 60, Thermo Scientific) at a wavelength of 245 nm [ 160 ]

PAGE 88

88 Only one pump was used at the outflow end of the column to regulate the flow rate in ex periments with the water saturated column. The influent concentrations of SWNTs used in the saturated experiments matched those of the unsaturated experiments. The procedures of sample collection and analysis for saturated experiments were the same as thos e for the unsaturated experiments. Mathematical Model The advection dispersion equation (ADE) coupled with kinetic expressions was used to simulate the retention and transport of the SWNTs in the sand columns. The governing equation can be written as [ 164 ] : (1) where C is the concentration of suspended SWNTs in pore water (M L 3 ) t is time (T), z is the travel distance in the direction of flow (L), D is the dispersion coefficient (L 2 T 1 ), v is the average linear pore water velocity (L T 1 ), and k is the deposition rate (T 1 ). This equation can be solved analytically with a zero initial concentrat ions, a pulse input boundary condition, and a zero concentration gradient condition at the outlet boundary [ 165 ] Re sults and Dis cussion Characteristics and Stabili ty of S urface Modi fied SWNTs The FTIR spectra were used to determine the acid oxidation effect on the O SWNTs. Compared to the pristine SWNTs, the O SWNTs showed more FTIR peaks in the spectra ( Figure 5 1 ). As shown in the Figure, peaks at 344 5 1635, 1045 cm 1 of the O SWNTs could be assigned to the O O bond, respectively [ 160 ] This indicates that the

PAGE 89

89 acid oxidation processes introduced carboxyl and hydroxyl functional groups to SWNT surfaces. The EPM and zeta potential values of the tested SWNTs were all negative ( Table 5 1 ), confirming that their surfaces were coated w ith the negatively charged functional groups, SDBS, or HA under the tested experimental conditions. The absolute values of the EPM and zeta potential followed the order of H SWNTs > S SWNTs > O SWNTs, suggesting the same trend for their stability. The stru cture and composition of HA were comprised with numerous functional groups, which tend to bring more negative charges to the surface of H SWNTs, enhancing the repulsive electric double layer forces to separate SWNT bundles into individual ones, and thus fa cilitating the dispersion of H SWNTs in aqueous solutions. Similarly, functional groups and surfactants on the surfaces of the modified SWNTs also introduces negative charges, but less than that of the H SWNTs, to help the dispersion/solubilization. Tempo ral changes of the surface modified SWNT concentrations are shown in Figure 5 2 The relative concentrations of the S SWNTs and H SWNTs remained unchanged for 24 h, indicating strong stability due to the surface modification. The concentrations of the O SWNTs was unchanged for the first 10 h, but decreased slightly (3%) from 10 to 24 h. This suggests that although all the surface modification methods were effective, the HA and SDBS modifications might have slightly better effect than the acid oxidization method to disperse and stabilize SWNTs in solutions. To avoid the potential inter ferences from aggregations of the surface modified SWNTs, all transport experiments used fresh prepared SWNTs suspension. In addition, each experiment, including sample collections and analyses, was controlled to be completed within 10 h.

PAGE 90

90 Interaction betw een Surface Modi fied SWNTs and Air Bub bles The temporal changes of concentrations of the surface modified SWNTs at the eight vertically distributed sampling ports (P1 8 from top to bottom) are shown in Figure 5 3 All the reported concentrations were normalized to the initial concentration. Concentrations of S SWNT and H SWNT from the eight sampling ports within the bubble column stayed unchanged for more than 20 h with continuous air bubble injections. This resul t suggests that the surfactant or the humic acid dispersed SWNTs do not attach to the air water interface under the tested experimental conditions, which is consistent with the results reported previously [ 160 ] The O SWNTs, however, showed a different concentration distribution in the bubble column after 20 h. At the end of the experiment, the O SWNT concentration at P1 (top) increased about 5%, but decreased about 10% at P8. This result indi cates that O SWNTs were attached to the air bubbles and transferred from the bottom of the bubble column to the top. Transport of Surface Modified SWNTs in Medium Sa nd Breakthrough curves of the three types of surface modified SWNTs in saturated and unsat urated columns packed with medium sized sands are shown in the Figure 5 4 The breakthrough points of all the modified SWNTs were less than 1 PV, reflecting the particle dispersion effect and possible enhanced particle velocity caused by the size exclusion effect [ 166 ] Under saturated conditions, the three types of SWNTs showed similar transport behaviors in the sand columns. After applied to the colum ns, the relative concentrations (i.e., C/C 0 ) of the three modified SWNTs quickly climbed to a peak at one, indicating all three surface modification methods are effective in enhancing the mobility of SWNTs in porous media. For S SWNTs and H SWNTs, the moisture contents did not show no t able influence on their breakthrough behaviors and they had

PAGE 91

91 almost the same breakthrough curves for all the tested conditions ( Figure 5 4 ). For O SWNTs, however, the peak breakthrough concentrations under moisture content of 0.21 and 0.13 were about 4 % and 1 2 % lower than that of saturated (0.39) experiment, respectively. This indicates that the O SWNTs were less mobile than the H SWNTs and S SWNTs in unsaturated medium porous media. Those results are consistent with the res ults obtained from the bubble column experiment that the air water interface (AWI) can capture (retain) the O SWNTs but not the S SWNTs or the H SWNTs. Previous studies have demonstrated that colloidal and nanosized particles could be retained in unsatura ted porous media through irreversibly deposited on the AWI under certain conditions [ 119 146 ] Although electrostatic forces between negatively charged particles and the AWI are repulsive and unfavorable for attachment [ 148 167 ] hydrophobic and capillary forces could dominate the interactions between the particles and the AWI [ 148 168 169 ] Transport of Surface Modified SWNTs in Fine Sa nd Breakthrough curves of the surface modified SWNTs in the fine sand columns under differ ent moisture conditions are shown in Figure 5 5 The peak breakthrough concentrations of H SWNTs in all find sand columns were close to one under the three moisture conditions, indicating no/little H SWNTs was retained in the porous media. Small portions o f retention for O SWNTs and S SWNTs were observed in the fine sand under various moisture conditions ( Figure 5 5 ). Similar to the results in the medium sand columns, the O SWNTs had the lowest breakthrough (i.e., lowest mobility) in the fine sand. The H SW NTs again were highly mobile and had a slightly higher mobility than the S SWNTs in the fine sand.

PAGE 92

92 Under statured conditions, the retentions of O SWNTs and S SWNTs in the fine sand columns could be attributed to their attachment on the grain surfaces. The attachment of CNTs on sand grain surfaces is mainly controlled by the DLVO or the extended DLVO forces, which are affected by the solution chemistry and physiochemical characteristics of the particle and grain surfaces [ 103 111 ] Although the experimental conditions were chemically unfavorable, the fine sand had less negatively charged surfaces than the medium or the coarse sand ( Table 5 1 ). In addition, the surface char ge (negative) of the O SWNTs and S SWNTs was lower than that of the H SWNTs ( Table 5 1 ). As a result, it might be easier for the O SWNTs and the S SWNTs to attach to the fine sand surfaces through either second minimum deposition or primary minimum deposit ion after overriding the energy barrier [ 161 ] Reductions in moisture content showed little impact on the transport of H SWNTs and S SWNTs in the fine sand columns ( Figure 5 5 ), which is consistent with the results from the bubble column experiments. While there was no retention of H SWNTs in the two unsaturated columns, the peak breakthrough concentrations of S SWNTs in the two unsaturated fine sand column were around 0.92 C 0 similar to that in the saturated one. This suggests that moisture had no impact on S SWNT retention and transport; theref ore, the retention of the S SWNTs in the unsaturated fine sand columns can also be attributed to the attachment on the grain surfaces. The transport of the O SWNTs in the unsaturated fine sand, particularly in the one with moisture of 0.29, showed lower br eakthrough than that in the saturated fine sand ( Figure 5 5 ), suggesting O SWNTs may be captured by both grain surfaces and the air water interfaces. Compared to the unsaturated medium sand, the moisture contents of the unsaturated fine sand were

PAGE 93

93 higher. A s a result, the air water interfaces showed lower impact on the transport and retention of O SWNTs in the fine sand than that in the medium sand. Transport of Surface Modifie d SWNTs in Coarse Sa nd The transport of the three surface modified SWNTs in satur ated coarse sand was similar to each other with peak concentrations close to one ( Figure 5 6 ), indicating no/little deposition on the grain surfaces. This is consistent with the zeta potential measurements that the coarse sand had the highest negative surf ace charges. The two moisture contents used in the unsaturated coarse sand columns (i.e., 0.11 and 0.09) were lower than those in the unsaturated medium (i.e., 0.21 and 0.13) and fine (i.e., 0.33 and 0.29) sand columns ( Table 5 2 ). The relatively low mois ture contents in the unsaturated coarse sand reduced the mobility of all the surface modified SWNTs ( Figure 5 6 ). Compared to the other two SWNTs, the H SWNTs still had the highest mobility, but their breakthrough under low moisture conditions were lower than that under saturated conditions ( Figure 5 6 ), which were not observed in fine or medium sand columns ( Figures 5 4 and 5 5). Similarly, low moisture content also lowered the breakthrough of S SWNTs and O SWNTs in the unsaturated coarse sand columns. T he retention of S SWNTs and H SWNTs in the unsaturated coarse sand columns could be attributed to the water film straining mechanism [ 160 ] The water film straining mechanism is controlled by multiple factors and must satisfy two conditions: 1) moisture rings in the porous media are broken; and 2) water film thickness is smaller than the particle sizes [ 65 ] For nano sized particles, such as CNTs water film straining only occurs at low moisture conditions [ 160 ] and thus no film straining was observed for the three SWNTs in the medium and the fine sand columns because their moisture was

PAGE 94

94 relatively high. The transport of the O SWNTs in the unsaturated coarse sand was lower than that of the other two SWNTs, probably due to combined retention effects from both film training and air water interface attachment mechanisms. Mod el Sim ulations T he ADE model was used to simulate the transport behavior of SWNTs in the sand packed column with various moisture contents and the simulations matched the experimental breakthrough curves very well with R 2 bigger than 0.99 ( Figure 5 4 5 5, and 5 6). Pecle t number ( where H is the column length ) Damkohler number ( ) and maximum travel distance ( where C/C 0 is 0.001) were then calculated ( Table 5 2 ). The Pe values were relatively large (10 101) indicating flow advection may play more important role in mass transfer of the surface modified SWNTs in the sand columns than dispersion [ 170 ] The Da values were relatively small (0.001 0.250), indicating the time scale for deposition mass transfer was larger than that for advective transport. Compared to the O SWNTs and the S SWNTs, the Da values of the H SWNTs were generally smaller under the same co nditions, confirming that the experimental results that H SWNTs were more mobile than the other two. In addition, the values of L max for H SWNTs were larger than those for C SWNTs and S SWNTs for most of the tested conditions, further confirming that H SWN Ts had the highest mobility in the porous media. Chapter Conclusions In this work, laboratory experiments were conducted to investigate the retention and transport of three types of surface modified SWNTs (i.e., surface oxidization,

PAGE 95

95 surfactant modified, a nd HA modified) in saturated and unsaturated columns packed with fine, medium, or coarse sand. We found that all the three surface modification methods were effective in solubilizing the SWNTs. The absolute values of the EPM and zeta potential followed the order of H SWNTs > S SWNTs > O SWNTs, which is consistent with their stability trend. While all three surface modified SWNTs were highly mobile in the packed sand columns with different combinations of grain sizes and moisture, the H SWNTs showed the high est mobility for most of the tested conditions. NOMs, such as humic acid, are ubiquitous in soil and groundwater environment, as well as other aquatic systems. When CNTs are released in the environment, therefore, they may have high mobility in water flow due to the presence of NOMs. The results from this work also suggest that the retention and transport of SWNTs in porous media are controlled by both dispersion/solubilization methods and environmental conditions. Single or combined retention mechanisms, including gain surface attachment, air water interface attachment, and thin water film straining, may be responsible for the retention of the surface modified SWNTs in the saturated and unsaturated porous media depending on the experimental conditions. In addition, we also found that the advection dispersion model could be used to simulate the retention and transport of surface modified SWNTs in porous media for both saturated and unsaturated conditions.

PAGE 96

96 Table 5 1 Surface properties of the SWNTs and th e sand used in this work Materials Electrophoretic mobility (10 8 m 2 /(Vs)) Zeta potential (mV) O SWNT s 2.69 36.44 S SWNT s 3.30 42.23 H SWNT s 3.34 45.20 Fine Sand 1.53 19.59 Medium Sand 1.54 19.73 Coarse Sand 2.15 27.56

PAGE 97

97 Table 5 2 Summary of sand column experimental conditions and model parameters Treatment Moisture Model parameters v (cm/min) Pe Da L max (m) R 2 Medium sand O SWNTs Saturated 0.39 1.8 33.1 0.025 46 1.00 Unsaturated 0.21 3.1 14.7 0.037 31 0.99 Unsaturated 0.13 1.3 27.7 0.133 9 0.99 S SWNTs Saturated 0.39 2.0 22.5 0.035 33 1.00 Unsaturated 0.21 3.7 24.1 0.051 23 1.00 Unsaturated 0.13 1.4 21.0 0.068 17 1.00 H SWNTs Saturated 0.39 1.8 43.2 0.028 41 1.00 Unsaturated 0.21 3.1 37.5 0.018 64 1.00 Unsaturated 0.13 1.2 36.9 0.026 44 1.00 Fine sand O SWNTs Saturated 0.38 2.1 22.6 0.108 11 1.00 Unsaturated 0.33 2.4 18.6 0.094 12 0.99 Unsaturated 0.29 0.6 24.5 0.157 7 0.99 S SWNTs Saturated 0.38 0.5 100.8 0.073 16 1.00 Unsaturated 0.33 2.4 60.9 0.078 15 1.00 Unsaturated 0.29 0.7 77.2 0.097 12 1.00 H SWNTs Saturated 0.38 2.1 30.3 0.009 123 0.99 Unsaturated 0.33 2.5 33.6 0.016 74 0.99 Unsaturated 0.29 0.7 49.4 0.026 45 1.00 Coarse sand O SWNTs Saturated 0.41 1.6 37.8 0.018 63 1.00 Unsaturated 0.11 4.6 18.4 0.146 8 1.00 Unsaturated 0.09 1.1 18.1 0.250 5 1.00 S SWNTs Saturated 0.41 2.0 41.7 0.023 51 1.00 Unsaturated 0.11 5.7 12.1 0.064 18 1.00 Unsaturated 0.09 2.2 10.1 0.069 17 1.00 H SWNTs Saturated 0.41 1.7 48.4 0.008 139 0.99 Unsaturated 0.11 5.1 49.3 0.001 845 1.00 Unsaturated 0.09 1.4 25.5 0.040 29 1.00 v, Pe, Da, L max and R 2 represent average linear pore water velocity, Peclet number, Damkohler number, maximum travel distance a nd coefficient of determination, respectively.

PAGE 98

98 Figure 5 1 FTIR spectra of pristine SWNTs and O SWNTs

PAGE 99

99 Figure 5 2 Stability of surface modified SWNTs in aqueous solutions

PAGE 100

100 Figure 5 3 Temporal changes of surface modified SWNT concentrations at the eight sampling ports within the bubble column

PAGE 101

101 Figure 5 4 Transport of surface modified SWNTs in medium sand columns under saturated (volumetric moisture content, 0.39 ) and unsaturated (volumetric moisture content, 0.21 and 0.13) conditions

PAGE 102

102 Figure 5 5 Transport of surface modified SWNTs in fine sand columns under saturated (volumetric moisture content, 0.38) and unsaturated (volumetric moisture content, 0.33 and 0.29 ) conditions

PAGE 103

103 Figure 5 6 Transport of surface modified SWNTs in coarse sand column s under saturated (volumetric moisture content, 0.41) and unsaturated (volumetric moisture content, 0.11 and 0.09 ) conditions

PAGE 104

104 CHAPTER 6 6 DEPOSITION AND TRANSPORT OF FUNCTIONALIZED CARBON NANOTUBES IN WATER SATURATED SAND COLUMNS 1 Introduct ory Remarks Advances in nanotechnology make it possible to develop many new and exciting materials to benefit society. Large scale production of engineered nanomaterials (ENMs) however, will inevitably result in their release into the environment and compromise the ecosystem and public health [ 26 28 171 172 ] Carbon nanotube s (CNT s ) are one of the most commonly used ENMs, which have attracted much research attention recently. CNTs can be categorized as single walled nanotubes (SWNTs) and multi walled nanotubes (MWNTs). SWNTs are one layered graphitic cylinders having diameters on the order of a few nanometers. MWNTs comprise of 2 to 30 concentric cylinders having outer diameters often between 10 50 nm Due to their special mechanical, electrical, optical, and thermal properties CNTs have been widely applied in a broad range of commercial and industr ial applications [ 16 ] Concerns about the environmental impact and risk assessment of CNTs have surfaced with the discovery of broader applications. Many studies have shown toxic effects of CNTs to various aquatic and mammalian organisms [ 22 24 86 ] It is therefore important to improve current understanding of the transport, behavior, and fate of CNTs in the environment. In many applications, CNT surfaces are often functionally engineered in order to increase their stability and optimize their performance [ 114 ] Because functionalized 1 Reprinted with p ermission from Tian, Y., Gao, B., Wang, Y., Morales, V.L., Carpena, R.M., Huang, Q., and Yang, L., Deposition and transport of functionalized carbon nanotubes in water saturated sand columns. Journal Of Hazardous Materials, 2012. 213 214 (0): p. 265 272.

PAGE 105

105 CNTs are more stable in solution, when released into soils and groundwater, they are likely to experience higher mobility, which increases the risk for fresh water contamination and aquatic organism toxicity. In addition, several studies have also demonstrated that functionalized CNTs have high affinity for soil and groundwater contaminants, including heavy metals and recalcitrant organic pollutants [ 73 139 173 174 ] Transport of CNTs in soils is consequently suspected to affect the fate of other contaminants thought to have very limited mobility in soils. Such, facilitated transport of reactive contaminants in soils by colloids and nanoparticles is a well known phenomenon that has been observed in several circumstances [ 41 175 176 ] Despite the environmental importance, only a limited amount of studies have been conducted to examine the transport of CNTs in porous media [ 27 93 116 117 142 176 177 ] In most of these studies, laboratory columns packed with acid cleaned sand or glass beads (i.e., clean columns) are used as model porous media to explore the governing mechanisms of CNT transport Results obtained from the model systems suggested that surface pro perties of the CNTs play an important role in controlling their retention and transport in porous media. It was found that CNTs coated with natural organic matter (NOM) a natural and ubiquitous substance in terrestrial and aquatic environments, have much higher mobility in saturated porous media than uncoated CNTs [ 14 2 ] Recent studies have a lso show n that CNTs coated with anionic surfactant are highly mobile in both saturated and unsaturated porous media [ 27 117 176 ] The transport of functionalized CNTs in model porous media has also been investigated and showed only slightly lower transport compared to NOM or surfactant coated CNTs under similar experimental conditions [ 93 116 ] In natural sandy soils, however, the

PAGE 106

106 mobility of functionalized CNTs reduced substantially [ 30 ] ; suggesting that surf ace properties of the porous media are also important to CNT transport Taken together, evidence up to now strongly suggest s that surface deposition (interaction) is an important mechanism of CNT retention in porous media but a mechanistic understanding of this effect is lacking Physical straining has been suggested as an important process driving retention and transport of CNTs in saturated porous media ; particularly under favorable conditions [ 30 93 ] Straining refers to the trapping of particles, such as colloids or nanoparticles, in pores that are too small to allow their passage This process is believed to occur when the ratio of the particle a nd collector grain diameters falls below a critical value [ 60 138 178 ] Under unfavorable conditions, however, there is a debate on whether st r aining should be considered as one of the governing mechanisms of particle retention in porous media [ 124 149 ] Previous studies of CNT transport in saturated porous media also had sp lit opinions on the importance of the straining mechanism On one hand, some studies suggested that physical straining plays an important role in retention and transport of functionalized CNT s, particularly CNT s with a large aspect ratio and their aggregates, in both model and natural porous media [ 30 93 ] On the other h and, straining was considered insignificant in several studies of the retention and transport of humic acid coated, surfactant coated, and functionalized CNTs in model porous media [ 116 117 142 176 ] Additional investigations are thus needed to examine the governing mechanisms of CNT transport in porous media. The overarching goal of this study was to determine the governing retention mechanisms of two types of functionalized CNTs (SWNTs and MWNTs) in natural

PAGE 107

107 porous media. We hypothesized that the ret ention of functionalized CNTs in natural sand are mainly controlled by surface deposition through interactions between the functional groups of the CNTs and surface impurities (metal oxyhydroxides) of the natural sand. To test this hypothesis, retention an d transport of functionalized SW NTs and MWNTs were examined in laboratory columns packed with natural, baked, and acid cleaned sand under different conditions Our specific objectiv es were to : ( 1) compare the retention and transport of the CNTs in laboratory column packed with natural, acid cleaned, and baked sand, ( 2) determine the effect of perturbations in flow direction, flow rate, and surfactant concentration on remobilization of initially r etained CNTs in natural and baked sand, and ( 3) examine the retention and transport of the CNTs in natural and baked sand under different pH conditions. Materials and Methods Functionalized of CNTs Carboxyl functionalized SWNTs (Cheap Tubes Inc., Brattle boro, VT) were used as received from the manufacturer Functionalized M WNTs were synthesized using a chemical vapor deposition (CVD) method with nickel and magnesium catalysts and acid oxidized using a 3 to 1 ratio of sulfuric and nitric acids to add hydro xyl and carboxyl functional groups on its surface. To prepare the suspensions for experimentation, 16 mg CNTs were suspended in 1000 ml deionized (DI) water, followed by 30 min of u ltrasonication (Misonix S3000). A fresh CNT stock was used for each transpo rt experiment and the experimental time was controlled to be within 4 hours. The pH of the CNT suspensions in DI water was around 5. 6 Stability experiments confirmed that two CNT suspensions were stable at least for 24 hours after the sonication ( Figure C 1). The electrophoretic mobility (EPM) and hydrodynamic diameter of the SWNT and

PAGE 108

108 MWNT were determined using a ZetaPlus (Brookhaven Instrument Co., Holtsville, NY). CNTs. B this estimation may overestimate the actual zeta potential by up to 20% [ 142 ] The point of zero charge (PZC) of the CNTs was determined using the mass titration method [ 179 ] A brief description of the mass titration method can be found in the Supporting Information. Porous Media Natural silica sand porous media were obtained from Standard Sand & Silica Co. (Davenport, FL) and were carefully sieved into a size range of 0.5 0.6 mm to ensure uniformity. The sand was washed sequentially by tap water and DI water, oven dried at 70 C, and used in this study as natural sa nd. to determine the organic matter on the surface of sand. Part of the washed natural sand was baked in a furnace at 550 C to remove surface organic matters and used as baked sand [ 176 ] Another batch of the sand was cleaned with 10% nitrate acid (i.e., acid cleaned sand) to remo ve surface impurities using the procedures reported by Tian et al. [ 117 ] The organic carbon (OC) content of the sand was determined using the Walkley Black chromic acid dig estion method [ 180 ] There were about 0.08% of OC on the natural sand, but no detec Table 6 OC on the baked or acid cleaned sand. The zeta potential of the natural, baked, and acid cleaned sand was determined according to the procedure of Johnson et al. [ 94 ] The ir PZC were determined using the mass titration method ( Supporting Information ) Scanning electron microscope coupled with energy dispersive spectroscopy (SEM EDS) was employed to examine th e presence of metal elements on all of the sand surfaces. The EDS has a penetration depth of about 100 nm at the surface of the samples.

PAGE 109

109 Additionally, metal element concentrations of the natural sand were determined by inductively coupled plasma with atomi c emission spectrometry (ICP AES), after digestion with concentrated HNO 3 and HCl, as per the EPA protocol 200.7. Packed Sand Column Preparation The sand was wet packed (i.e., saturated) uniformly into an acrylic column measuring 2.5 cm in diameter and 16 .7 cm in height. M embranes with 50 m pores (Spectra/Mesh, Spectrum Laboratories, Inc.) were used at the column inlet and outlet to distribute the flow and to maintain the capillary pressure. The membranes were in immediate contact with the porous media a nd sealed by rubber O rings. To pack each saturated sand column, a small amount of sand was poured gently into 8 ml of DI water standing at the bottom of the column until the sand surface was 0.5 to 1 cm below the water level. A polypropylene stir rod was used to stir the sand in the column. Approximately 8 mL DI water was then added to the column and the sides of the column were gently tapped a uniform number of times. This procedure was repeated several times until the column was packed to a height of 1 5 cm. The porosity of the natural and acid cleaned sand columns was approximately 0.40. CNT Transport i n Natural, Baked, a nd Acid c leaned Sand Columns A peristaltic pump (Masterflex L/S, Cole Parmer Instrument in Vernon Hills, IL) was connected to the inl et at the top of the sand column to regulate the flow. DI water was first pumped through the column for approximately 2 hr. at velocity of 1.0 cm/min to wash the column and ensure no background interferences in the effluent for CNT measurements. DI water w as used in this work to make the experimental condition be way to explore alternative (non DLVO) retention mechanisms of fine particles in porous

PAGE 110

110 media [ 178 181 ] A pulse of prepared CNT suspension of either MWNTs or SWNTs in the same chemical conditions was then introduced into the column, followed by CNT free DI water to flush out unretained C NTs at the same flow velocity. The suspension pulse volumes were about 15 pore volumes (PVs) for natural sand columns, and 2 PVs for acid cleaned sand columns. E ffluent samples were collected from the bottom outlet of the column using a fraction collector (IS 95 Interval Sampler, Spectrum Chromatography in Houston, TX). CNT concentration s were determined by measuring the total absorption of light at wavelengths of 255 nm with UV Vis Spectrophotometer (Evolution 60, Thermo Scientific) The absorbance spectra and calibration curves of the two CNTs can be found in the supporting information ( Figure C 2). Bromide (NaBr, 40 ppm) was also applied to the column as a conservative tracer. Distribution o f Retained CNT i n Natural a nd Baked Sand Columns The deposition p rofiles of retained CNTs in the natural and baked sand columns were analyzed at the end of the transport experiment s by dividing the columns post transport experiments into nine layers of equivalent mass of sand along the depth gradient. Each sand layer wa s placed into a vial and washed with 25 mL DI water by shaking the mixture by hand. The supernatant was carefully transferred into a new vial to determine the resuspended CNT concentration. An additional 25 mL of DI water was subsequently added to the sand and the vial was ultrasonicated for 30 min. The second supernatant was carefully transferred into a vial to determine the CNT concentration. The residual sand was oven dried at 75 C to determine the presence of CNTs on its surfaces. A fresh batch of natu ral sand was used in the experiment following the same procedure as the control.

PAGE 111

111 Mobilization o f Retained C NT s i n Natural a nd Baked Sand Columns For some of the natural sand and the baked sand columns, after the CNT injection and DI water flushing, the effect of flow direction, flow rate, and surfactant concentration on mobilization of retained CNTs was investigated. The flow direction experiment was initiated by switching downward DI water flow to upward at the same velocity (1.0 cm/min) for four PVs. The effect of flow rate on CNT mobilization was tested by increasing the flow velocity to 2.0 cm/min (downward) for four PVs. Sodium Dodecylbenzenes ulphonate (SDBS), an anionic surfactant, was used in the experiment as the mobilization agent in this study because it is commonly used to disperse CNTs in water [ 117 ] The SDBS (1% by weight) solution was applied to the natural sand column for four PVs at flow velocity of 1.0 cm/min to mobilize previously retained CNTs. E ffluent samples from the mobilization experiment were collected in the same fashion stated previously, b y fraction collector and CNT concentration s were determined with the spectrophotometer. Effect o f pH o n CNT Transport i n Natural a nd Baked Sand Columns To test the effect of solution pH on CNT transport in the natural and baked sand columns, 0.1 mol /L NaO H solution was used to adjust the pH of the experimental solutions (CNT suspensions and DI water) to either 8.0 or 10. A fresh sand column was first flushed with the high pH working solution at a flow velocity of 1.0 cm/min until the effluent pH reached th e same value. The inflow was then switched to the CNT suspension at the same pH and flow velocity for about 2 PVs. The column was then flush ed again with the working solution for about 2 PVs at the same conditions. E ffluent samples were collected to determ ine CNT breakthrough concentration s

PAGE 112

112 Modeling A transport model based on the colloid filtration theory was used to simulate the retention and transport of the CNTs in the sand columns. The governing equation can be written as [ 104 ] : ( 6 1 ) where C is the concentration, D is the hydrodynamic dispersion coefficient, v is the average linear water velocity, z is the travel distance in the direction of flow, and k is the clean bed filtration rate constant This equation can be solved analytically with a zero initial concentrations, a pulse input boundary condition, and a zero concentration gradient condition at the outlet boundary [ 165 ] The Levenberg Marquardt algorithm was used to estimate the value of the model parameters to minimize the sum of the squared differences between model calculated and measured breakthrough concentrations [ 182 183 ] This model optimization method wa s first applied to bromide breakthrough data to estimate D. Triplicated bromide tracer experiments were conducted and the best fit D was 0.062 cm2/min (R2=1.00). Results and Discussion Surface Characteristics Measurements of the EPMs ( Table 6 1) indicated that the two functionalized CNTs were negatively charged under all experimental conditions used in this study, which are consistent with the results reported in the literature [ 93 116 ] MWNTs had slightly lower EPMs than the SWNTs, probably becau se of the presence of both hydroxyl and

PAGE 113

113 carboxyl functional groups on their surfaces. The ionization of functional groups (i.e., deprotonation of hydroxyl and carboxyl groups) introduces strong negative charges and electrostatic repulsion between functiona lized CNTs, and consequently enhances their stability in suspension. The estimated zeta potentials of the two functional CNTs under all experimental conditions were below 30 mV ( Table 6 1), indicating good stability. The Derjaguin Landau Verwey Overbeek ( DLVO) theory was used to estimate the interaction energies (Supporting Information). The DLVO theory was developed for interactions between spherical particles and homogeneous surfaces; therefore, the calculations can be only used as exploratory estimation s here. The estimated DLVO energy profiles between the CNT particles in solution confirmed that the experimental conditions were unfavorable for CNT aggregating ( Figure C 3a b). This is consistent with the results of the stability experiments ( Figure C 1). Mass titration curves ( Figure C 4) indicated that the PZC of the two CNTs were similar with values lower than 3, confirming both SWNT and MWNT were negatively charged for all the tested conditions. The surfaces of the natural, baked, and the acid cleaned sand were also determined to be negatively charged under all experimental conditions tested ( Table 6 1). Both the EPM and zeta potential of the baked and natural sand was slightly higher than the acid cleaned sand, probably due to presence of positively charged impurities on the baked and natural sand surfaces. Without considerations of surface charge heterogeneity, the DLVO energy estimations indicated that the experimental conditions were unfavorable for attachment (i.e., traditional DLVO dep osition) of the CNTs onto either the acid cleaned, baked, or the natural sand ( Figure C 2c h). The PZC values of the t hree types of sand were similar to each other with values close to 3, which agree

PAGE 114

114 with reported values of silica/quartz PZC in the litera ture [ 184 ] The similarities of EPM zeta potential, and PZC among natural, baked, and acid cleaned sand grains suggested the dominance of the silica on the sand surfaces. SEM EDS analyses confirmed the dominance of silica signature on natural, baked, and acid cleaned sand surfaces ( Figure 6 1). While no impurities were identified on the acid cleaned sand surface, the SEM EDS spectrum did show the presence of metallic impurities (i .e., Fe, Al, and Ni) on both natural and baked sand surfaces ( Figure 6 1). ICP AES analyses corroborated the SEM EDS results and showed relatively high levels of Fe ( 167.4 mg/kg) and Al ( 1086 .6 mg/kg), and a very low level of Ni (0.7 mg/kg) on the natural sand. The identification of these elements agrees with the previously reported aluminum and iron oxyhydroxides on natural sand surfaces [ 185 ] Because metal oxyhydroxides are positively charged under most p ractical circumstances, their presence introduces charge heterogeneity on the natural and the baked sand surfaces; a physicochemical factor that will likely affect the transport of functional CNTs. As pointed out by Johnson et al [ 124 ] surface charge heterogeneity is likely a primary factor controlling deposition and transport of nanoparticles in porous media. Retention in the Acid c leaned Sand Columns Both the SWNTs and the MWNTs showed high mobility in the saturated columns packed with acid cleaned sand ( Figure 6 2a). This is consistent with interaction energy calculations from the DLVO theory. SWNTs and MWNTs in their respective experiments were detected in the effluent about one PV after the pulse was applied. The breakthrough response for both CNT suspensions quickly increased and plateaued with additional input of CNTs. The breakthrough concentrations of the two CNTs quickly returned back to zero when the column influent was switched to CNT free D I water. The

PAGE 115

115 plateau of the normalized breakthrough curves (C/C 0 ) for both SWNTs and MWNTs approached unity (i.e., C/C 0 = 0.96 and 1.00, respectively), and the general shape of the breakthrough curves resembled that of bromide ( Figure 6 2). The resemblance of BTCs to the conservative tracer confirms low CNT retention in the acid cleaned sand columns, and lack of interaction with the porous media surfaces. Mass balance calculations showed that only 3% and 1% of the SWNTs and MWNTs were retained, respectively in the acid washed sand columns. Simulations of the transport model matched well the breakthrough data of CNTs in acid cleaned sand columns ( R 2 >0.9 7 ) ( Table 6 2). The best fit, clean bed filtration coefficients ( k ) of the SWNTs and MWNTs in the acid clea ned sand columns were 0.0036 and 0.0012 min 1 respectively. The Damkohler number for SWNT and MWNT deposition ( where L is the column length and v is flow velocity) were 0.0 59 and 0.0 20 respectively. The magnitude of the Da number indicates that the time scale was much longer for retention than for advection through the columns. These results confirmed that experimental conditions where acid washed sand is used as the porous medium do not favor for CNT retention. The high m obility of functionalized CNTs in laboratory columns packed with acid cleaned sand was also observed in other studies under similar experimental conditions [ 93 116 ] Retention in the Natural and Baked Sand Columns The breakthrough responses of t he CNTs in the natural and the baked sand columns were similar and were much lower than that in the acid cleaned sand columns ( Figure 6 2). This is inconsistent with interaction energy calculations from the DLVO theory, suggesting the strong retention of C NTs in the columns may not be controlled by the traditional DLVO deposition. Although the CNTs were applied to both the natural

PAGE 116

116 sand and the bake sand columns for an extended period of time, the normalized breakthrough concentrations reached the plateau st age quickly and remained at low peak C/C 0 concentration of 0.02 0.12. The recovery rates of SWNTs and MWNTs was 2.1% and 8.6% for natural sand columns, and 9.5% and 12% for the baked sand columns ( Table 6 2), indicating that CNT mobility is significantly l ower in chemically heterogeneous sand (i.e., natural and baked sand) than in oxyhydroxides free sand columns (i.e., acid cleaned). The transport model described almost all transport data well but performed weakly for MWNT transport in the natural sand colu mn (R 2 = 0.60) ( Table 6 2). The best fit k values of the SWNTs and MWNTs in the natural and baked sand columns were 0.13 0.24 min 1 which are two orders higher than that in acid cleaned sand columns ( Table 6 2). The Damkohler numbers for the SWNTs and MWNTs in the natural and the baked sand columns were 2.1 3.9 ( Table 6 2), indicating the time scale of retention was comparable to that of the advection process. The enhanced retention of the two CNTs in the natural and the baked sand columns can be attrib uted to the surface deposition mechanism because all types of sand (i.e., natural, baked, and acid cleaned) were similar in most the physicochemical properties, except in the surface characteristics. While the acid cleaned sand has relatively homogeneous s urfaces, surfaces of both natural and baked sand contain metal oxyhydroxides, which introduce surface charge heterogeneities. The CNTs are covered with negatively charged functional groups (of carboxyl dominance for SWNTs and a combination of carboxyl and hydroxyl for MWNTs), and they thus can attach strongly to the metal oxyhydroxides on natural or baked sand surfaces through electrostatic and/or hydrogen bonding interactions [ 142 ] Previous studies have

PAGE 117

117 demonstrated the strong interactions between metal oxyhydroxides and hydrophilic surface functional groups, espec ially carboxyl and hydroxyl groups [ 186 188 ] The results also suggested that physical straining might not be considered as a dominant retention mechanism for CNT tra nsport in the natural sand columns under the tested experimental conditions. This is consistent with previous transport studies of surfactant dispersed SWNTs, NOM coated SWNTs and functionalized MWNT in clean porous media [ 116 117 142 176 ] In a rec ent study, Tian et al. [ 117 ] did not observe any straining of surfactant dispersed SWNTs in water saturated columns packed with acid cleaned quartz sand with different grain size distributions and assumed that CNT might orient parallel to the streamlines in flow to avoid the pore straining in porous media. Distribution o f Retained C NT s i n Natural a nd Baked Sand Columns Measurements of the retained CNTs in the natural and the baked sand columns confirmed their strong deposition on the sand grain surfaces. When the sand w as excavated from the column, it s color was much darker than the original natural sand, especially near the column inlet. Supernatant samples obtained after wa shing the sand with DI water were transparent and appeared to be the same as the control solution ( Figure 6 3a). The CNT concentrations in the supernatants were zero as measured by the UV Vis spectrophotometer at 25 5 nm, confirming that CNTs in the porous media were not retained through physical straining. Subsequent sonication of the sand layers successfully detached a fraction of the retained CNTs from the sand; particularly from the layers closer to the inlet ( Figure 6 3b). Although we were not able to d etermine the CNT concentration of the sonicated supernatants by spectrophotometry because the sonication step generated additional sand colloids, the light absorption of all samples at

PAGE 118

118 25 5 nm was higher than that of the control. Sonication could detach retained CNTs from the sand surface; however, the color of the sonicated sand was still darker than the original sand ( Figure 6 3c). This result further confirmed the strong interaction betwe en the functionalized CNTs and the metal oxyhydroxides on sand surfaces. Mobilization o f Retained C NT s i n Natural a nd Baked Sand Columns The concentration of CNTs in the effluent after the reverse, high velocity, or SDBS flushing was below detection limit ; suggesting that none of the tested physical and chemical transients can effectively remobilize the retained CNTs in the natural sand or the baked sand columns. If pore straining was one of the dominant mechanisms of CNT retention in the columns, change i n flow direction would likely result in the release of the retained CNT s from pore throats of the porous media [ 42 189 ] On the other side, the lack of mobilization of CNTs from either increases in flow velocity or addition of SDBS suggest that interactions between the grain and CNT surfaces are controlled by relatively strong attraction forces, such as electrostatic and/or hydrogen bonding attractions. This result is consistent with the hypothesis that the retention of the functionalized CNTs in the sand is mainly controlled by surface deposition through interactions between the functional groups and metal oxyhydroxides. p H E ffect When higher pH was used, the breakthrough of the two types CNTs in both the natural sand and the baked sand columns increased ( Figure 6 4). For the functionalized SWNT in natural sand, the normalized peak breakthrough concentrations were 0.2 4 and 0.7 4 when the column pH values were at 8.0 and 10, respectively (see breakthrough curves in Figure 6 4a). Similarly, the normalized peak breakthrough concentrations for the functionalized MWNT were 0.2 3 and 0.7 8 with pH of 8.0 and 10.0,

PAGE 119

119 respectively (see breakthrough curves in Figure 6 4b). The recovery rates of the SWNTs and MWNTs from the natural sand columns also increas ed dramatically when the transport conditions were alkaline. Approximately 22% and 78% of the SWNT and 23% and 71% of the MWNT were recovered from the natural sand columns at pH of 8.0 and 10, respectively. Similarly about 23% and 75% of the SWNT and 24% and 75% of the MWNT were recovered from the baked sand columns at pH of 8.0 and 10, respectively. The transport model described all the pH experiment very well with R 2 higher than 0.9 7 ( Table 6 2). The best fit k values of the SWNTs and MWNTs in the natura l and the baked sand columns at two pH values ranged from 0.22 to 0.78 min 1 ( Table 6 2). The Damkohler numbers for the SWNTs and MWNTs in the natural sand and the baked sand columns were 0.21 1.5 ( Table 6 2), suggesting the time scale of retention was co mparable to that of the advection process. Both experimental and modeling results revealed that pH played an important role in enhanc ing the transport of functionalized CNTs in sand porous media. This result concurs with the hypothesis that deposition of functionalized CNTs onto the natural sand surfaces is driven by electrostatic and/or hydrogen bonding attractions between the functional groups and metal oxyhydroxides. Characterization of the natural and the baked sand indicated that the surface charge he terogeneity is mainly introduced by the Fe, Al, and Ni oxyhydroxides. Metal oxyhydroxide impurities on the sandy medium act as favorable deposition sites that are positively charged at pH 5.6 (DI water) and promote deposition of negatively charged CNTs. Pr evious studies have shown that the points of the zero charge of the metal oxyhydroxides identified in our sand fall between 6 .1 and 11.3 ( Table C 1 ) [ 184 190 191 ] When the system pH increased from 5.6 to

PAGE 120

120 8.0 or 10, it reduced or reversed the positive charge of the metal oxyhydroxides and diminished the surface charge heterogeneity. The higher the system pH is, the less positive the surface charge of the metal oxyhydroxides will be. As a result, transport of the functionalized CNTs in the sand columns increased with the pH increments. Chapter Conclusions This study demonstrates the importance of chemical surface interactions between the nanoparticles and the porous media, which stro ngly control transport. Our results suggest that surface charge heterogeneity, is a key factor that controls the transport behavior of nanoparticles moving through porous media. Although it is commonly believed that the nanoparticle filtration efficiency o f sandy soils is low, our study presents evidence that functionalized ENMs have limited mobility in natural sandy porous media that are rich in metal oxyhydroxides. Surface functionalization is a routine step in the preparation of nanoparticle suspensions to improve their performance, but may render additional benefits for the prevention of uncontrolled dispersal of nano litter in the environment as nanoparticles encounter chemically heterogeneous soil environments and become chemically immobilized. Finding s from this study also suggest that natural sand may be used as an efficient filter to remove functionalized ENMs from wastewater.

PAGE 121

121 Table 6 1 Surface characteristics of the functionalized CNTs and the porous media Size (mm) Effective hydro dynamic diam eter (nm) Point of zero charge Electrophoretic mobility (10 8 m 2 /(Vs)) Zeta potential (mV) pH=5.6 pH=8 pH=10 pH=5.6 pH=8 pH=10 SWNT 152.7 2.9 2.71 2.69 2.75 34.6 36.4 37.2 MWNT 179.8 2.4 2.99 2.82 2.89 38.2 38.2 39.1 Acid cleaned sand 0.5 0.6 2.9 1.43 3.04 3.34 19.4 41.1 45.2 baked sand 0.5 0.6 3.1 1.13 2.61 2.87 15.3 35.4 38.8 Natural sand 0.5 0.6 3.1 1.11 2.46 2.69 15.1 33.3 36.4

PAGE 122

122 Table 6 2 Summary of experimental conditions and model results of CNT transport in the sand columns pH Mass recovery Deposition rate, k (min 1 ) Damkohler number, Da R 2 SWNT transport in Acid cleaned sand 5.6 0.97 0.0036 0.059 0.99 Baked sand 5.6 0.095 0.14 2.3 0.97 Natural sand 5.6 0.021 0.24 3.9 0.83 Baked sand 8.0 0.23 0.088 1.5 0.98 Natural sand 8.0 0.23 0.088 1.5 0.97 Baked sand 10 0.75 0.017 0.27 1 Natural sand 10 0.71 0.018 0.30 0.97 MWNT transport in Acid cleaned sand 5.6 0.99 0.0012 0.02 0.97 Baked sand 5.6 0.12 0.13 2.1 0.97 Natural sand 5.6 0.086 0.15 2.5 0.6 Baked sand 8.0 0.24 0.084 1.4 0.99 Natural sand 8.0 0.22 0.091 1.5 0.98 Baked sand 10 0.75 0.013 0.21 0.99 Natural sand 10 0.78 0.015 0.25 0.99

PAGE 123

123 Figure 6 1 SEM images (a, b and c ) and corresponding EDS and c acid cleaned sand baked sand and the natural sand at 4000X respectively EDS spectra were recorded at the same locations showed in the SEM image.

PAGE 124

124 Figure 6 2 T ransport of CNTs in water saturated columns in DI water system packed with (a) acid cleaned sand (b) baked sand and (c) natural sand

PAGE 125

125 Figure 6 3 Distribution of retained CNTs in the natural sand column : (a) supernatants from different layers after DI water washing ; (b) supernatants from different layers after ul trasonication ; (c) sand from different layers dried after ultrasonication.

PAGE 126

126 Figure 6 4 T ransport of CNTs in water saturated columns packed with baked and natural sand at different pH conditions.

PAGE 127

127 CHAPTER 7 7 METHODS OF USING CARBON NANOTUBES AS FILTER MEDIA TO REMOVE AQUEOUS HEAVY METALS Introduct ory Remarks The exceptional properties of nanomaterials, particularly carbon nanotubes (CNT), have revolutionized the electronic and optic industries, energy sectors, and material engineering and manufacturin g. Much research on carbon based nanoparticle in the environment has been devoted to elucidate the physical, chemical, and biological mechanisms that affect their stability, mobility, and toxicity [ 142 ] Relatively less attention has been paid to their potential application for addressing a number of environmental p roblems [ 27 176 ] The hollow and layered nanostructure of CNTs endow them with a characteristically large surface area and a correspondingly high potential sorption capability for chemical pollutants [ 67 ] Several studies have demonstrated that CNTs, particularly those that are functionalized, have a strong affinity to many common water pollutants, including heavy metals [ 171 ] and organic pollu tants [ 121 ] It is well accepted that pristine CNTs are insoluble in water [ 192 ] therefore oxidativ e treatments are often used to introduce hydrophilic functionalities to their surfaces in order to facilitate ability to remove heavy metals in aqueous phase by increasing t heir cation exchange capacity as well as promoting attractive electrostatic interactions [ 173 ] As such, functionalized CNTs have been reported to have a superior sorption ability to heavy metal ions than pristine CNTs and have been recommended as a potential adsorbent f or the removal of heavy metals in contaminated water [ 74 193 194 ]

PAGE 128

128 A vast proportion of published investigations concerned with the removal of contaminants by CNTs have investigated sorption characteristics and mechanisms through batch sorption experiments [ 76 77 171 ] has previously examined the feasibility and effectiveness of using CNTs as filter media in fixed bed settings to remove contaminants from aqueous solutions. A potential deterrent for the novel application of CNTs for water treatment is the concern of nanoparticle elution from the fixed bed reactors, which could result in secondary contamination of the receiving water bodies. A recent study by the authors suggested that natural sand media can be used as a good natural filter to remove functionalized CNTs from water [ 171 ] The deposition of functionalize d CNTs on natural sand surface is mainly controlled by strong surface interactions that are irreversible despite chemical and hydrodynamic disturbances [ 76 ] This makes it possible to envision the development of a new technology to use sand as a safeguard for f unctionalized CNTs in a fixed bed filter to remove contaminants from water. Sand is very a common filter material used in water treatment, but it has relatively poor ability to remove heavy metals from water [ 77 195 ] We hypothesize that the combination of sand media with functionalized CNTs would greatly improve the performance of fixed beds for removing heavy metals from water. The overarching goal of this work is to assess the best procedures and their respective effectiveness of the CNT sand filter me dia for heavy metal removal from water. First, functionalized multi walled CNTs were incorporated to natural quartz sand media in laboratory fixed bed columns via three packing methods: 1) layered, where CNTs and sand were packed as separate layers in the columns; 2) mixed, where CNTs

PAGE 129

129 and sand were mixed together and subsequently packed in the columns; and 3) deposited, where a well dispersed CNT suspension was applied to the columns and allowed to become deposited onto sand surface via filtration. Second, filtration experiments were conducted with the CNT sand columns in which the removal efficiency of two heavy metals in aqueous phase (Pb 2+ and Cu 2+ ) was evaluated. The specific objectives of the work included: 1) determine the effect of sonication promoted dispersion on sorption capacity of the functionalized multi walled CNTs to Pb 2+ and Cu 2+ in single and dual (competing) metal systems; 2) compare removal efficiency of single metal solution (Pb 2+ or Cu 2+ ) across the three packing types of CNT sand media; and 3) compare removal efficiency of dual metal solution (Pb 2+ and Cu 2+ ) across the three packing types of CNT sand media. Materials and M ethods CNTs M ulti walled CNTs were produced using a chemical vapor deposition method with nickel and magnesium catalys ts. Subsequently, these were functionalized by an acid mixture of concentrated sulfuric and nitric acids (3:1, v:v) to introduce carboxyl and hydroxyl functional groups to the nanotube surface [ 195 ] Fourier transform infra red (FTIR) analysis was used to characterize those functional groups. To obtain the observable adsorption spectra, the CNTs were mixed with KBr to 0.1 wt % and then pressed into pellets. The spectra were measured using a Bruker Vector 22 FTIR spectrometer (OPUS 2.0 software). Part of the functionalized CNTs was used directly in some of the sorption and column experiments without any dispers ion promoting treatments and was referred here as undispersed CNTs A separated dispersed CNT suspension was produced by

PAGE 130

130 suspending the functionalized CNTs in deionized (DI) water at a desired concentration and then sonicated with Misonix S3000 u ltrasonica t or ( QSonica, Newtown, CT ) for 3 0 min, and was referred here as dispersed CNTs The physiochemical characteristics of the CNTs were measured for: i) Surface area via the Brunauer Emmett Teller (BET) nitrogen adsorption method at 77 K (NOVA 1200 surface area analyzer, Quantachrome Instruments, Boynton Beach, FL), ii) point of zero charge (PZC) via the mass titration method [ 176 ] iii) h ydrodynamic diameter of dispersed CNTs by dynamic light scattering (ZetaP lus, Brookhaven Instruments Corporation, Holtsville, NY), and iv) electrophoretic mobility (EPM) of dispersed CNTs (ZetaPlus), which was used estimating their zeta potential via Smoluchowski's model. Concentration of the dispersed CNTs was regulated by spe ctroscopic light absorption at a wavelength of 255 nm using Evolution 60 UV Vis Spectrophotometer (Thermo Scientific, Waltham, MA). FTIR analysis was conducted with the post sorption CNTs (i.e., metal laden CNTs) to examine the functional groups using the method described above. Sand Quartz sand of grain size 0.5 0.6 mm (Standard Sand & Silica Co., Davenport, FL) was washed sequentially with tap water and DI water to remove loose impurities. Quartz fragmentation by ultrasonication procedure proposed by Joh nson et al. [ 196 ] was used to cre ate quartz colloids, which were subsequently filtered through a 0.45 m filter. The identification and concentration on the sand surface were determined by i nductively coupled p lasma with optical emission spectrometry (ICP OES, Optima 2100 DV, PerkinElmer Inc., Waltham, MA), as per EPA protocol 200.7.

PAGE 131

131 Heavy Meta ls Lead nitrate and copper nitrate were used to prepare stock solutions. Single metal solutions were prepared at concen trations of 10 mg L 1 of Pb 2+ or Cu 2+ In addition, a dual metal solution containing 10 mg L 1 of Pb 2+ and 10 mg L 1 of Cu 2+ was prepared. ICP OES was used to monitor the metal concentration in the solutions. HCl was used to adjust the pH of the metal sol utions to 5.6. Batch Sorption Experim ent Batches of each sorbent (10 mg of dry CNTs, dispersed CNT suspension containing 10 mg of CNTs, or 1g of sand) were added to 20 mL of each metal solution (Pb 2+ Cu 2+ and Pb 2+ and Cu 2+ ) at seven different concentrati ons to build adsorption isotherms. The mixtures were shaken in a shaker for 12 h at room temperature, which was determined sufficient time to ensure equilibration [ 179 197 ] After equilibrium was equilibrium metal concentrations then measured using the ICP OES. The mass of sorbed metal w as calculated as the difference between initial and equilibrium metal concentration. Fixed bed Column Experi ment Fixed bed column experiments were used to investigate the removal and transport of heavy metals in CNT sand media via three different CNT packi ng methods. Layered, mixed and deposited CNTs were the three ways in which about 10 mg of CNTs were incorporated into 16.8 mg of sand for each CNT sand packed column. Columns packed with only natural sand were used as the control in the study. For the laye red CNT packing method, natural sand was first wet packed into an acrylic column (1.5 cm inside diameter 5 cm height), and un dispersed CNTs evenly

PAGE 132

132 placed at the top of the sand column as a single layer. M embranes with 50 m pores (Spectra/Mesh, Spectrum Laboratories, Inc.) were used at the column inlet and outlet to distribute the flow. For the mixed CNT packing method, natural sand was thoroughly mixed with the un dispersed CNTs and subsequently wet packed into the column. For the deposited CNT packing me thod, the column was first wet packed with the sand, and the dispersed CNT suspension was then applied as a pulse (in downward flow direction) using a peristaltic pump set at a steady flow rate of 1 mL/min. After continuous CNT pulse injection, the influen t was switched to DI water for an additional 2 hrs to flush out unretained CNTs. Effluent concentration of CNTs was monitored in discrete samples via UV Vis Spectrophotometry at a wavelength of 255 nm to determine when 10 mg of CNTs had been deposited in t he sand column. Because small amount of CNTs (~0.06% w/w of CNT/sand) were used in the fixed bed columns, the bed porosity of all three types of column packing was maintained at 0.40. Once packed with sand and CNTs, the columns were subjected to pulses of single and dual metal solutions and the removal efficiency evaluated according to the metal breakthrough from the columns. Prior to metal pulse injection, the columns were flushed with DI water (pH 5.6) for more than 1 hr and no CNT was detected in the el uents of any the columns. Metal removal experiments consisted of two stages. At stage one, the single (Pb 2+ or Cu 2+ ) or dual (Pb 2+ and Cu 2+ ) metal solutions (as described above) were injected into the top of the column for 2 hrs at a steady flow rate of 1 mL/min. At stage two, the influent was switched to metal free DI water for an additional 2 hrs to elute unfiltered metals from the column. Effluent samples were collected discretely with a fraction collector and the metal concentration monitored with the I CP OES.

PAGE 133

133 Results and Di scussion CNT and Sand Pro perties In dispersion, hydrodynamic diameter of the CNTs was around 180 nm, and their EPM was recorded at 2.99 10 8 m 2 V 1 s 1 Mass titration curves ( Figure D 1) indicated that PZC of the CNTs was 2.4, corroborating their negative charge under the tested experimental conditions. Those surface charges may arise from the abundances of acidic functional groups. FTIR spectrum of the CNTs were characterized by four s ignificant bands at wave number 344 5 (O H stretch), 1635 (C=O bond), 1045 (alkene, C O bond), and 138 5 (vibrational band of NO 2 ) cm 1 ( Figure 7 1). The first three bands confirmed the presences of carboxyl and hydroxyl functional groups on the CNT surface BET nitrogen adsorption measurement revealed a surface area of un dispersed CNT powder of 112.7 m 2 g 1 a significantly lower value than previously reported for well dispersed CNTs (>1000 m 2 g 1 ) [ 94 ] This result suggests that un dispersed CNT dispersed in a medium. Thus, it is anticipated that the dispersed CNTs would display higher sorption capacity for metals than the undispersed CNTs based solely on the available surface area for sorption. The PZC of the sand was 3.0 ( Figure D 1), indicating that its charge was negative under the tested conditions. Surface element analysis revealed the presence of meta l impurities (mainly iron oxyhydroxides) on the sand surface [ 77 ] which could potentially serve as adsorption sites for heavy metal ions [ 176 ]

PAGE 134

134 Sorption Abi lity of the CNTs Both dispersed and undispersed CNT s were much more efficient adsorbents of heavy metals than the sand ( Figure 7 2 ), which is consistent with the literature that functionalized CNTs have strong affiliations to heavy metal ions in aqueous solutions [ 195 198 201 ] As expected from available surface area, the dispersed CNTs had greater adsorption ability to bo th aqueous metals in single and dual metal systems than the undispersed CNTs, stressing the importance of CNT surface area on heavy metal sorption. The isotherms also suggest that the sonication promoted dispersion processes could potentially improve the p erformance of the CNTs in water treatment to remove heavy metal contaminants The Langmuir model used to describe the experimental isotherms was: where q e is the mass of metals sorbed per mass of sorbent at equilibrium (mg g 1 ), q m the maximum mass of m etals sorbed per mass of sorbent as the concentration of metal increases (mg g 1 ), K is the Langmuir equilibrium constant for mass of metal (L mg 1 ), and C e the dissolved metal concentration at equilibrium The model fit the experimental data very well, as indicated by R 2 values exceeding 0.92 ( Table 7 1). The q m values of Pb 2+ and Cu 2+ sorption on the sand were low in both single and dual metal systems ( Table 7 1), indicating that natural sand alone is no t an effective sorbent for heavy metals in solution. The single metal maximum sorption capacity of the CNTs was 74.5 and 92.3 mg g 1 (0.36 and 0.45 mmol g 1 ) for Pb 2+ and 51.3 and 67.8 mg g 1 (0.81 and 1.07 mmol g 1 ) for Cu 2+ for undispersed and dispersed CNTs, respectively. The maximum sorption capacity measured for the CNTs in this study is several fold

PAGE 135

135 higher than that of the natural sand and is even better than metal sorption maxima reported for some activated carbons [ 115 200 202 ] The dual metal maximum sorption capacity for undispersed CNTs for Pb 2+ and Cu 2+ was 49.3 and 33.0 mg g 1 (0.24 and 0.52 mmol g 1 ), while that for dispersed CNTs was 65.0 and 43.6 mg g 1 (0.31 and 0.69 mmol g 1 ), respectively. A likely explanation for higher molar sorption capacity of Cu 2+ than for Pb 2+ is the atomic radii of these ions, which is 70 and 112 pm, respectively. Because of a steric over crowding, larger ionic radius of Pb 2+ compared to that of Cu 2+ m ay induce a quick saturation of adsorption sites, resulting in lower molar capacity [ 203 ] Compared to the undispersed CNTs, the sorption capacity of the dispersed CNTs increased for 23.9 32.2% for Pb 2+ and Cu 2+ in both single and dual metal systems. These data further confirm the importance of CNT dispersal in order to maximize its sorption capacity for water treatment purposes. It is suggested that the adsorption of heavy metals onto the CNTs are mainly contro lled by the strong interactions between the metal ions and hydrophilic surface functional groups, especially carboxyl and hydroxyl groups [ 115 204 205 ] Comparison of the FTIR s pectra of CNTs before and after metal sorption confirmed this mechanism ( Figure 7 1). Compared to the original FTIR spectrum, the adsorption of heavy metals on the CNTs resulted in variations of FTIR peaks at wavenumbers of 1635 (C=O) and 1045 (C O) cm 1 which could be attributed to the interactions between metal ions and carboxyl and hydroxyl groups. Previous studies have suggested that positively charged metal ions can form strong complexes with these two functional groups through electrostatic and/or hy drogen bonding interactions [ 73 142 206 207 ]

PAGE 136

136 Single Metal Remov al in the CNT Colu mns The evaluation of three packing methods of CNT sand media was conducted through analyses of effluent breakthrough of dissolved metals in water. Breakthrough curves (BTCs) for single metal pulse injections were analyzed for filtration and transport as they moved through the fixed bed sand y columns with and without CNTs as illustrated in Figure 7 3 The BTCs were constructed as plots of the cumulative pore volume of liquid eluted versus the relative concentration (C/C 0 ) of the injected metal in the effluent. For injections of single metal (P b 2+ and Cu 2+ ) pulses, the BTCs corresponding to fixed beds with deposited CNTs (inverted triangle in Figure 7 3 ) had the lowest mass of metal eluted and the most delayed breakthrough elution, indicating that deposited CNT packing method was the most effective way in improving the fixed f deposited CNT fixed beds agrees with the sorption experiment results that indicated that dispersed CNTs were better sorbents than undispersed CNTs, which were used for layered and mixed CNT fixed beds. Based on the BTC shapes and arrival times, the metal removal efficiency of the fixed bed columns was ranked in decreasing order as: deposited CNT > mixed CNT > layered CNT > CNT free sand. The following three filtration performance parameters were calculated to evaluate the column packing methods against the different combination of metals in solution: removal rate ( r m ), breakpoint ( b p defined as the BV at which the metal concentration in the effluent reaches 5% of the initial metal concentration [ 186 ] ), and recom mended bed capacity ( q r ) ( Table 7 2). The recommended bed capacity represents the capacity in the columns at which the sorbent in the fixed bed should be replaced, and can be

PAGE 137

137 determined as [ 187 ] : where Q is the flow rate (L min 1 ), m is mass of the sorbent in the fixed (g), t is time (min), and t h is time when the metal concentration in the effluent reaches half of that in the influent (i.e., C/C 0 = 0.5). As a conservative measure, the calculated q r of a sorbent should not exceed its Langmuir capacity, as determined from batch sorption experiment. The CNT free sand columns were able to remove 32 .0 % and 12 .5 % of the inputted single metal Pb 2+ and Cu 2+ respectively. In contrast, the CNT sand columns improved the metal removal performance of the columns at r m values of up to 75 .0 % for Pb 2+ and 5 6.9 % for Cu 2+ for the best performing CNT packing method (deposited), and for 55.3% for Pb 2+ and 31.4% for Cu 2+ for the worst performing CNT packing method (lay ered). Whereas the range of removal efficiency varied for the three types of CNT packing, the r m of all CNT sand columns in all cases exceeded that for the CNT free sand columns. The breakpoints ( b p ) of the single metal pulses in the CNT sand columns were also much longer than those in the CNT free sand columns. The most delayed b p was observed for deposited CNT columns, which was recorded at 9.5 min for Pb 2+ and 5.2 min for Cu 2+ indicating that metal elution in these CNT enabled columns was delayed by mo re than twice the time metals were eluted from CNT free sand columns ( Table 7 2). The recommended bed capacities ( q r ) were the lowest for CNT free sand columns ( q r of 0.016 mg g 1 for Pb 2+ and 0.011 mg g 1 for Cu 2+ in Table 7 2), indicating that sand alone is not an effective heavy metal removal system. The q r values for CNTs in all the CNT sand columns were three to four orders of magnitude higher than that for sand, with deposited CNTs at the highest capacity and layered CNTs at the lowest ( Table 7 2 ).

PAGE 138

138 These trends are consistent with the findings of individual material batch sorption experiments. The q r values for deposited CNTs in CNT sand columns ( Table 7 2) corresponded to 87% and 63% of their measured Langmuir maximum sorption capacity q m ( Table 7 1) for Pb 2+ and Cu 2+ respectively. The apparent reduced capacity of the metal sorption for the same quantity of CNTs when added to a fixed bed system is likely due to the shorter residence time between metal ions and the CNTs when flow of the metal solu tion was passed through the column, compared to the longer contact time between the metals and the sorbents in batch tests. Dual Metal Rem oval in the CNT C olumns The three types of CNT sand columns were also superior at removing Pb 2+ and Cu 2+ simultaneous ly from the solution than the CNT free sand columns ( Figure 7 4 ). Similar to the single metal experiments, the metal breakthrough in the CNT sand columns were lower in mass and more delayed than that in the CNT free sand columns, where deposited CNT metho d was ranked with the best filtration performance. These results are in agreement with the sorption and single metal removal experiments for dispersed CNT sorption efficiency of metals. The presence of competing metal ions in the solution for the sorption sites reduced the removal rate, the breakpoint, and the recommended bed capacity in all the columns for the removal of the individual metals ( Table 7 2). It is known that both Pb 2+ and Cu 2+ cations react strongly with the carboxyl and the hydroxyl groups of the CNT surfaces [ 188 208 ] and competition for these sites on the CNTs is expected. As a result, the removal rates of the two heavy metal ions in the dual metal system in the three CNT sand columns ranged from 50.2% to 63.7% for Pb 2+ and 23.6% 39.9% for Cu 2+ Although the molar co ncentration of Cu 2+ in the dual metal solution was much higher

PAGE 139

139 than that of Pb, the competition only reduced about 9 14% of Pb 2+ removal in the three CNT sand columns. On the other hand, the presence of Pb 2+ reduced about 25 30% of the Cu removal. These re sults suggest that the functionalized CNTs have a higher affinity to Pb 2+ than to Cu 2+ ions in aqueous solutions. Similar trend was observed in previous studies of heavy metal sorption onto CNTs and it has demonstrated that Pb 2+ has a much stronger affinit y to the surface function groups on CNTs than Cu 2+ in aqueous solutions [ 73 115 208 ] Chapter Conclusions Laboratory experimental results indicated that functionalized multi walled CNTs had strong sorpt ion ability to aqueous Pb 2+ and Cu 2+ Dispersion of the CNT particles enhanced their sorption ability to the heavy metals. As a result, the CNT sand columns packed with CNT deposited method showed the best effect to remove the heavy metal from aqueous solutions. All the three CNT sand packin g methods, however, were effective and safe ways use the CNTs as filter media to remove heavy metal contaminants from water. These results suggest that, the high metal sorption affinity of CNTs along with the high porosity of natural sand could be exploite d jointly in a single filter to effectively remove multiple types of heavy metals from water. Although CNTs have been referred as potentially promising sorbents for heavy metal removal in the literature; however, little research effort has been made to a pply them in fix bed setting to purify water. As the first of this kind of study, findings from this work may inform the development of innovative and high efficiency CNT based filters for various environmental applications, especially for the treatment of heavy metal contaminated water.

PAGE 140

140 Table 7 1 Summary of the best fit Langmuir model parameters Adsorbent Pb 2+ Cu 2+ q m (mg g 1 ) / (mmol g 1 ) K (L mg 1 ) R 2 q m (mg g 1 ) / (mmol g 1 ) K (L mg 1 ) R 2 Single metal Sand 0.027 / 1.31E 04 0.28 0.98 0.015 / 2.32E 04 0.04 0.92 Undispersed CNT 74.5 / 0.36 0.33 1.00 51.3 / 0.81 0.19 0.99 Dispersed CNT 92.3 / 0.45 0.37 0.99 67.8 / 1.07 0.31 1.00 Dual metal Sand 0.018 / 8.83E 05 0.40 0.99 0.006 / 9.75E 05 0.09 0.98 Undispersed CNT 49.3 / 0.24 0.20 1.00 33.0 / 0.52 0.07 1.00 Dispersed CNT 65.0 / 0.31 0.21 0.98 43.6 / 0.69 0.13 0.99

PAGE 141

141 Table 7 2 Summary of fixed bed column experimental results Sand column Pb 2+ Cu 2+ r m (%) b p (min) q r (mg g 1 ) / (mmol g 1 ) r m (%) b p (min) q r (mg g 1 ) / (mmol g 1 ) Single metal CNT free 32.0 4.4 0.016 / 7.72E 05 12.5 1.2 0.006 / 9.44E 05 Layered CNT 55.3 6.0 25.9 / 0.13 31.4 3.1 12.6 / 0.20 Mixed CNT 65.9 7.3 44.3 / 0.21 41.6 3.3 21.1 / 0.33 Deposited CNT 75.0 9.5 80.1 / 0.39 56.9 5.2 42.6 / 0.67 Dual metal CNT free 23.1 2.4 0.011 / 5.31E 05 5.4 1.2 0.006 / 9.44E 05 Layered CNT 50.2 5.4 25.3 / 0.12 23.6 2.2 15.3 / 0.24 Mixed CNT 56.7 6.7 29.9 / 0.14 30.1 3.1 18.0 / 0.28 Deposited CNT 63.7 9.5 40.9 / 0.20 39.9 4.5 21.2 / 0.33

PAGE 142

142 Figure 7 1 FTIR spectra (average of 32 scans) of CNTs before and after heavy metal adsorption

PAGE 143

143 Figure 7 2 Sorption isotherms of Pb 2+ and Cu 2+ onto sand, and dispersed and undispersed CNTs in single and dual metal solutions

PAGE 144

144 Figure 7 3 Transport of Pb 2+ and Cu 2+ in single metal solutions through different types of fixed bed columns

PAGE 145

145 Figure 7 4 Transport of Pb 2+ and Cu 2+ in dual metal solutions through different types of fixed bed columns

PAGE 146

146 CHAPTER 8 8 REMOVAL OF SULFAMETHOXAZOLE AND SULFAPYRIDINE B Y CARBON NANOTUBES IN THE FIXED BED COLUMN Introduct ory Remarks Sulfonamide antibiotics are widely used as human and veterinary pharmaceuticals to treat disease [ 209 211 ] Sulfamethoxazole (SMX) and sulfapyridine (SPY) are two commonly used sulfonamide antibiotics known to bioaccumulate up the food chain and cause acute as well as chronic adverse effects. [ 21 2 ] Simultaneous exposure to multiple antibiotics could result in the enhanced toxic effects [ 213 ] The use for these antibiotics is mainly for treating human patients. A general concern for public health is the development of antibiotic resistance from chronic exposure and accumulation of antibiotics dissolved in water. Therefore, it is urgent to investigate the pathways through which SMX and SPY disperse in the environment, and develop systems that can efficiently remove these substances from water. Sulfonamide antibiotics could directly enter aquatic environments through aquaculture activity, ph armaceutical manufacturing, and medical waste disposal. These antibiotics could also indirectly gain access to surface and subsurface waters from leached waste of livestock receiving drug treatment [ 214 ] Precipitation events could accelerate the release of antibiotics concentrated in animal manure. Once dissolved in surface waters, antibiotic loading could mix with gr oundwater in the soil profile. The protection of surface and groundwater quality, as two primary sources for drinking water, from contamination of leached antibiotics is of great priority for pubic and environmental health. A variety of physiochemical tec hniques have been developed to remove or destroy antibiotics from water sources, including oxidation [ 215 ] ion exchange [ 216 ] reverse

PAGE 147

147 osmosis [ 217 ] and adsorption [ 218 ] Adsorption is a prevailing method to remove the contaminants from the aquatic environment. Adsorbents, such as clay [ 219 ] zeolite [ 220 ] and activated carbon [ 221 ] have been examined for the removal efficiency of sulfonamide antibiotics in aqueous solutions. Solution chemistry has been shown to strongly affect the removal efficiency of these adsorbents with particular importance placed on solution pH [ 222 ] ionic strength [ 223 ] and presence of competitive sulfonamide antibiotics [ 224 ] Carbon nanotubes (CNTs), as newly introduced adsorbents, are gaining increasing attention because of their exceptional sorbing properties. CNTs have hollow and layered structures with characteristically large surface area of great poten tial for high sorption capability [ 225 ] Previous studies have demonstrated high adsorption ability of CNTs to both heavy metals [ 226 ] and organic pollutants [ 74 ] Investigations on the adsorption of sulfonamide antibiotics on CNTs have found that removal efficiency varies according to the type and pretreatment of CNTs [ 227 228 ] Pretreatment of CNTs (e.g., by surface functionalization) is widely used to disperse these nanoparticles in aqueous solutions for their optimum uses [ 193 ] The introduction of functional groups also increases the ion exchange capacity of the CNTs and augments the number of available sites that can be used for el ectrostatic adsorption [ 194 ] As such, functionalized CNTs have been report ed to exhibit greater potential for removing antibiotics from aqueous environments than pristine CNTs [ 74 ] However, the current application of CNTs for antibiotic removal from water has been restricted to batch sorptio n methods. Two key disadvantages of batch methods include difficulty to collect exhausted/spent adsorbents and interruptions incurred when

PAGE 148

148 integrated to existing continuous processes [ 229 ] The in situ removal of antibiotics in the environment, such as an agricultural field, requires the continuous flow setup to ensure both efficiency and safety, which batch adsorption method might not possess. It is thus important t o develop and optimize CNT enabled water treatment methods to take advantage of the large sorption capacity of CNTs. Unlike batch sorption systems, fixed bed filters permit continuous flow and adsorption of antibiotics from solution, and thus opening a w ealth of opportunities to be used in wastewater in situ treatment. The fixed bed column packed with particular adsorbent could simulate the potential performance of the adsorbent and provide practical operational adsorption efficiency. Despite the above me ntioned benefits of fixed removal of antibiotics from aquatic environment using CNTs in fixed bed columns has not been explored. Moreover, CNTs are expected to maintain their large surface area by staying dispersed, thus obtaining greater adsorption ability than undispersed CNTs. The removal efficiency of antibiotics by dispersed CNTs vs. undispersed CNTs in fixed bed columns is unknown. Similarly, information is needed on the effect of physicochemical conditions on the removal efficiency of antibiotics by CNTs. In this work, we used fixed bed column methods to investigate the removal of SMX and SPY by functionalized CNTs from aqueous solutions under various conditions. Our ov erarching objective was to investigate the removal efficiency of SMX and SPY by CNTs in a fixed bed system under various conditions. The specific objectives of the work were: (1) examine the removal of SMX and SPY from aqueous solutions by trickling the co ntaminated water through a sand/CNT fixed bed column under different

PAGE 149

149 sets of conditions (i.e., pH, CNT incorporation method, adsorbent dosage, bed depth, adsorbate initial concentration and flow rate), and (2) evaluate the efficiency of regeneration of the fixed bed columns for reuse. Materials and Methods Be d Ma terials and Co nditions Functionalized multi walled carbon nanotubes (CNTs) ( Cheap Tubes Inc., Brattleboro, VT) and quartz sand were used as filter materials in the fix bed columns. Three methods of CNT incorporation to the sandy medium were tested (layered, mixed, and deposited), bed depth used ranged from 6 to 15 cm and flow rates ranged from 1 to 2 mL min 1 Chemical conditions tested for fixed bed experiments were 3.0 to 9.0 pH, 10 to 40 mg adsorb ent dosage, and 10 to 40 mg L 1 of antibiotic concentration. CNTs were produced using a chemical vapor deposition method with nickel and magnesium catalysts. Subsequent functionalization was achieved with an acid mixture of concentrated sulfuric and nitric acids (3:1, v:v) to introduce carboxyl and hydroxyl functional groups to the nanotube surface [ 122 ] A batch of the synthesized and functionalized CNTs were used in the dry powder form (undispersed CNTs), while a second batch was dispersed in w ater to create a suspension of CNTs (dispersed CNTs). To make the dispersion, 16 mg of the synthesized CNT powder were dispersed in 1000 ml deionized water and sonicated for 30 minutes in a Misonix S3000 ultrasonicator (QSonica, Newtown, CT). Thorough cha racterization of the CNTs for physiochemical properties was done for the following properties. Surface area of undispersed CNTs was measured using the NOVA 1200 surface area analyzer (Quantachrome Instruments, Boynton Beach, FL), following the Brunauer Emm ett Teller (BET) nitrogen adsorption method at 77 K. Point

PAGE 150

150 of zero charge (PZC) of the undispersed CNTs was determined using the mass titration method [ 179 ] Hydrodynamic diameter of dispersed CNTs was determined by dynamic light scattering with a Brookhaven ZetaPlus (Brookhaven Instruments Corporation, Holtsville, NY). CNT concentration in susp ension was calibrated by measuring the total absorption of light at wavelengths of 255 nm using Evolution 60 UV Vis Spectrophotometer (Thermo Scientific, Waltham, MA). Quartz sand (Standard Sand & Silica Co., Davenport, FL) was sieved into 0.5 0.6 mm g rain size and washed using deionized water (DI water). Basic properties and surface elemental compositions of the sand were reported previously [ 195 ] Antibiotics Two sulfonamide antibiotics, sulfamethoxazole (SMX) and sulfapyridine (SPY) (99%, Sigma Aldrich Co., St. Louis, MO), were used as adsorbates in the fixed bed column experiments. Chemical structure s and properties of SMX and SPY are reported in Table 8 1 Stock solutions of SMX/SPY were prepared (200 mg/L) and subsequently spiked with either 0.1 M KOH or HCl to adjust for desired pH (3.0 9.0) of the final solution. Antibiotic concentrations were d etermined by light absorption at wavelengths of 265 nm with a UV Vis Spectrophotometer [ 221 ] Lack of temporal change of concentration of SMX and SPY confirmed their stability for the duration of our experiments (data not shown). Fixed bed Column Experim ents Fixed bed column experiments were used to investigate the removal of dis solved antibiotics from water in a sand/CNT fixed bed system. A summary of the conditions tested is presented in Table 8 2 which includes CNT incorporation methods, pH, bed depth, adsorbent dosage, and adsorbate initial concentration and flow rate. For CNT

PAGE 151

151 incorporation methods, CNTs were incorporated onto sand by means of: layering, mixing and deposition. An acrylic column of 2.5 cm in diameter and 15 cm height was used to contain the CNT/sand system, to which 10 mg of CNTs were packed with 14 6.8 g of sand for all packing methods. For the CNT layered method, undispersed CNTs were placed as a layer on top of the wet packed sa nd column. For the CNT mixed method, undispersed CNTs were mixed with sand and subsequently wet packed into the column. For the CNT deposited method, the column was first wet packed sand and a dispersion of CNTs injected through the column with a peristalt ic pump (Masterflex L/S, Cole Parmer Instrument, Vernon Hills, IL) in a downward flow direction for more than 10 h. Then, the influent was changed to DI water was for 2 h to elute suspended (not deposited) CNTs from the column. Effluent samples were collec ted from at the outlet of the column with a fraction collector (IS 95 Interval Sampler, Spectrum Chromatography, Houston, TX) to monitor the concentration of eluted CNTs. Dispersed CNTs concentration was determined by total absorption of light at wavelengt hs of 255 nm with UV Vis Spectrophotometer. Mass balance calculations were used to determine the amount of retained CNTs in the column. It is important to note that prior to use for antibiotic sorption experiments, no CNTs were detected in the effluent of the CNT/sand packed columns, therefore CNT incorporation to the sandy medium considered irreversible. A column packed only with natural sand was used as the control. During packing of the column, layering and air entrapment were minimized by means of wet p acking. The porosity of all the CNT/sand columns was approximately 0.40. The peristaltic pump was connected to the inlet at the top of the column to

PAGE 152

152 regulate the flow rate. Membranes with 50 m pores (Spectra/Mesh, Spectrum Laboratories, Inc.) were used at the inlet and outlet of the column to distribute the flow. After packing the column with the methods described above, an antibiotic free background solution was injected to the column for 2 hr until the pore water chemistry reached equilibrium. Subsequen tly, an antibiotic solution was continuously injected to column and effluent samples collected at fixed intervals using a fraction collector. The concentration of antibiotics in the column effluent was measured with UV Vis Spectrophotometer. The experiment was terminated when the effluent concentration matched the initial concentration. Colu mn Regen eration and Rec ycling Post adsorption CNT mixed columns were selected as an example to evaluate the regeneration and recycling efficiency. The experimental con ditions of the selected columns were as follows: pH of 5.6, adsorbent dosage of 10 mg bed depth of 15 cm adsorbate initial concentration of 20 mg/L, and flow rate of 2 ml/min. To regenerate the spent columns, a solution containing 30 g/L NaCl and 1.5 g/L NaOH of pH 12 was injected to the fixed bed columns at the same flow rate as that used for antibiotic sorption experiments until effluent SMX or SPY concentrations fell below the detection limit [ 230 ] Effluent samples were collected and analyzed using the same protocol as that in the column filtration experiment. The column recycling was used to evaluate the viability of reuse of the sand/CNT fixed bed column. After regeneration of the column and equilibration with background solution, column adsorption and sample analysis were performed under the same conditions. The column was recycled up to five times and column capacity at each reuse cycle determined.

PAGE 153

153 Results and Discussion Column Analyses The two antibiot ics showed very high mobility in the CNT free sand columns ( Figure 8 1 ), which can be attributed to their low reactivity to sand [ 231 ] The columns with CNTs, however, showed high retention ability for both SMY and SPY ( Figure 8 1 ), indicating CNTs can be used in fixed bed colu mn to remove antibiotics from water. For all the tested conditions, the CNT/sand fixed bed column s had adsorption capacities of 11.7 92.0 mg/g for SMX and 77.8 123.4 mg/g for SPY ( Table 8 2 ). The p rimary adsorption zone (PAZ) concept was used to anal yze the breakthrough data of the antibiotics in the CNT/sand columns [ 232 233 ] W hen the antibiotics were injected into the fixed bed column the PAZ was initially formed as a narrow band on the top of column. After the adsorbents (CNTs) on the top were saturated with the antibiotics, t he PAZ moved downwards through the fixed bed column and finally reached the bottom of column [ 208 234 ] T he antibiotics started to breakthrough from the column and became detectable in the effluent s In this work, normalized concentration s (C/Co) of antibiotics in the effluent s were plotted against the operation time (min) to obtain the breakthrough curve s of the antibiotics in the columns Parameters controlling the formation and movement of the PAZ can be quantitatively determined form the breakthrough data [ 234 235 ] Definitions of the PAZ parameters ( e.g., t b t e v z et al. ) used in this study are summarized in Table 8 3 These parameters were analyzed to determine the effect of pH (3.0, 5.6, 7.0, or 9.0), CNT incorporation method (layered, mix ed, or deposited), adsorbent dosage (10, 20, 30, or 40 mg), bed depth (6, 9, 12, or 15 cm), adsorbate initial concentration (10, 20, 30, or 40 mg/L), and flow rate (1, 2, 3, or 4 ml/min), on the removal efficiency of CNT/sand

PAGE 154

154 columns to the two antibiotics For all the experimental conditions tested in this study, the calculated PAZ and column adsorption parameters are summarized in Table 8 2 as well as in the supporting information (Table s E 1 E 6). Effect of CNT Incorporation Me thod The effect of CNT incorporation method on the adsorption and transport of SMX and SPY in the columns is illustrated in Figure 8 1 The column maximum adsorption capacities to SMX and SPY, defined as mg of antibiotic per g of CNT/sand media, are summarized in Table 8 2 Experimental results indicate that the column adsorption capacity for SMX was 68.5 mg/g for layered method, 69.3 mg/g for mixed method, and 77.7 mg/g for dispersed method, with the capacity trend in the order of layered < mixed < dispersed. A similar trend was observed for SPY. Table 8 2 also shows the same trend for t e indicat ing dispersed method could retain antibiotics for the longest time which is corresponding to the largest capacity. As showing in the supporting information ( Table E 1 ) the t f and t z also followed the same order, further confirming that the effectiveness/capacity of the three CNT incorporation methods should follow: dispersed> mixed>layered. From a practical perspective, however, the construction of a CNT enabled fixed bed column by the mixed CNT method is recommended with little compromise on the sorption performance of the system. Therefore, the remaining facets of the study are focused on fixed bed systems built with the mixed CNT method. Effect of pH The effect of s olution pH on the adsorption and transport of SMX and SPY in the columns is illustrated in Figure 8 2 Experimental results showed that t he breakpoint time (t b ) decreased from 33.4 to 16.6 with the increase of pH from 3.0 to 9.0 ( Table 8 2 ), indicating tha t pH induced acceleration of SMX and SPY occurred in the columns as

PAGE 155

155 conditions became more alkaline. As shown in the supporting information ( Table E 2 ), decreased t f with pH increase represented that the PAZ formed faster at the higher pH. Decreased t z and increased v z with pH increase indicated faster time to create and move the PAZ Table 8 2 shows that, when pH increase from 3.0 to 9.0, the column adsorption capacity to SMX and SPY decreased from 84.9 to 11.7 mg/g and 93.9 to 77.8 mg/g, respectively. Thi s suggests alkaline pH could decrease of the column ability to removal antibiotics, especially for SMX. Mass titration measurements indicated that the PZC of CNTs was 2.2 [ 176 195 ] signifying that the surface of CNT was negatively charged for all the treatments in the present study. The surface charge of SMX and SPY varied greatly with pH, as represented by the acidity constants, pK a in Table 8 1 SMX and SPY became increasingly negatively charged with increase in solution pH, thus electrostatic repulsion between CNTs and antibiotics is expected. The larger second pK a2 of SPY (8.4), relative to that of SMX (5.6) can explain the observed differenc es in sorption capacity at the highest pH tested. This pH dependent electrostatic phenomenon explains the increased mass recovery of antibiotics from fixed bed column. Additionally, at equivalent pH conditions, the adsorption capacity of SPY by CNTs was la rger than that of SMX. For example, adsorption capacities at pH 7 for SPY were 86.1 mg/g and for SMX 45.8 mg/g. This difference in capacity can be explained by the speciation of the two antibiotics into dominant SMX and neutral SPY main species, which exp erience different degrees of electrostatic repulsion with the CNTs at the given pH. As illustrated in the Figure 8 2 the adsorption capacity of SMX by sand/CNT fixed bed column was greatly affected by changes in pH, while SPY removal did not vary much at the same conditions.

PAGE 156

156 groups could be also an important mechanism for the adsorption of SMX and SPY by CNTs, as has been previously been implicated for sorption of phen olic compounds by CNTs [ 236 ] Amin e functional groups and N heteroaromatic rings in the structure of acceptors; whereas, carboxyl and hydroxyl functional acceptors and donors [ 74 ] Thus the authors propose that the attraction of acceptor acceptor and acceptor donor pairs between the antibiotics and the CNTs is responsible for the observed strong adsorption. Effect of Bed Dep th The effect of bed depth on the ad sorption and transport of SMX and SPY in the columns is illustrated in Figure 8 3 The column adsorption capacities at various bed depths were in general greater for SPY (91.9 to 97.1 mg/g) than for SMX (69.3 to 72.8 mg/g) ( Table 8 2 ). To evaluate the eff ect of bed depth on the breakthrough time, the breakthrough data were interpreted using the Bed Depth Service Time (BDST) model. The BDST model was first introduced by Bohart and Adams [237 ] and further developed by Hutchins [238 ] This model has primarily been u sed in the analysis of fixed bed breakthrough data of heavy metals [ 239 240 ] and organic pollutants [ 159 241 ] The relatio nship between service time and bed depth is as follows: ( 8 1 ) where t is the service time at breakthrough (min); N 0 is the dynamic adsorption capacity of the bed (mg/L); C 0 is the influent antibiotics concentration (mg/L); v is the

PAGE 157

157 linear flow velocity feed to fixed bed (cm/min); h is the sand/CNT bed depth (cm); k is the adsorption rate constant (L/mg min); C t is the effluent solute concentration at duration, t (mg/L). The linear relationship between t and h ( Figure 8 4 ) was described by the s lope and the intercept The values of N 0 and k were thereafter back calculated. The critical bed depth (h 0 ) is the minimum depths of sand/CNT required to achieve the effluent breakthrough (C b ), which can be calculated from the lines best fitting equatio ns by letting t = 0 [ 234 242 ] The BDST model was developed for breakpoint (t b ) and exhausted point (t e ), respectively. The linear regression between service time and bed depth was shown in the Figure 8 4 R egression results suggested the fitted curves can be used for predicting the service time corresponding to the given bed depth both for SMX and SPY (R 2 > 0.9) The relationship between bed depth and service time demonstrate that changes in t b are parallele d by changes in t e ; thus indicating that the time required for the PAZ to move the length of its own height (t z ) is independent of bed depth. The horizontal distance between parallel is the height of the exchange zone [ 234 ] resulting in heights of 4.5 and 5.2 cm for SMX and SPY, respectively. The parameters N 0 k and h 0 were then calculated using the regression result for breakpoint at different bed depth. The value of N 0 k, and h 0 was 16.58 mg/L, 0.038 L/mg min, and 1.81 cm for SMX, and 20.72 mg/L, 0.023 L/mg min, and 2.42 cm for SPY, respectively which suggesting the removal ability of CNTs for SPY is better than that for SMX under the same conditions. T he parameter calculated using BDST model could

PAGE 158

158 also be used to inform the design of fixed bed column s to remove antibiotics from water [ 242 ] Effect of Adsorbent Dosa ge The effect of adsorbent dosage on the adsorption and transport of SMX and SPY in the columns is illustrated in Figure 8 5 As expected, the breakpoint time increased with the increase of CNT dosage. The breakpoint time (t b ) increased from 27.5 to 66.5 min for SMX and 33.5 to 85.5 min for SPY ( Table 8 2 ). The column adsorption capacity slightly decreased from 69.3 to 62.8 mg/g for SMX and 91.9 to 82.4 mg/g for SPY when CNT dosage was increased from 10 to 40 mg. From the fixed bed capacity, it can be concluded that the mass of adsorbed antibiotics per until mass of CNTs decreased with greater amount of CNTs in the fixed bed. A possible explanation could be interference of ant ibiotic adsorption from bonds on the functionalized CNTs can be formed with donors as well as acceptors [ 243 244 ] Effect of Adsorbate Initial Concentrat ion The effect of adsorbate initial concentration on the adsorption and transport of SMX and SPY in the columns is illustrated in Figure 8 6 As shown in the figure, the breakpoint time decreased with the increase of SMX and SPY initial concentration. The breakpoint time decreased from 37.5 to 25.5 min for SMX and 51.5 to 25.5 min for SPY as the initial concentration of antibiotics quadrupled. The column adsorption capacity of antibiotics by CNT was increased from 54.8 to 92.0 mg/g for SMX and 80.9 to 123.4 mg/g for SPY, when antibiotics initial concentratio n was increased from 10 to 40 mg/L ( Table 8 2 ). This large increase in fixed bed capacity is ascribed to the greater concentration gradient generated that promotes greater mass transfer of antibiotics from the liquid phase to the solid (CNT surface) phase [ 245 ]

PAGE 159

159 Effect of Flow Ra te The effect of flow rate on the adsorption and transport of SMX and SPY in the columns is illustrated in Figure 8 7 As shown in t he figure, the breakpoint time decreased with increasing flow rates f rom 59.1 to 14.8 min for SMX and 67.1 to 16.8 min for SPY ( Table 8 2 ). The column removal capacity of antibiotics by CNT was decreased from 78.3 to 61.9 mg/g for SMX and 94.9 to 83.6 mg/g for SPY when flow rate was increased from 1 to 4 ml/min. This reduction in sorption capacity is due to the decrease in contact time betw een the antibiotics and CNTs. The increase of flow rate also accelerates the movement of PAZ downwards in the fixed bed, which contributes to decreased adsorption capacity [ 243 ] Column Regene ration and Recyclin g The release curves of SMX and SPY in the post adsorption columns during the regeneration process are presented in the supporting information ( Figure E 1 ). Results showed that the release of SMX and SPY reached the peak of 1.2 C/C 0 and completed after regeneration process for 22 min. The mass recovery for the release process was 0.97 for SMX and 0.91 for SPY, suggesting that 3% of SMX and 9% of SPY of the previously sorbed antibiotics were strongly bond to the sand/CNT porous media during each regeneration cycle. Thus, although the sorption ca pacity was reduced for each adsorption regeneration cycle, the fixed beds continued to perform well as antibiotic sorbents. The trends of column capacity change after 5 regeneration cycles are plotted in Figure 8 8 From the plots it is evident that column capacity decreased after reuse from 67.9 to 32.6 mg/g for SMX and from 91.9 to 40.0 mg/g for SPY. As proposed previously, the reduced column capacity could be a result from decreased available adsorption

PAGE 160

160 sites (i. e., blocking) and repulsion by the irreversibly sorbed antibiotics from previous cycles. It is important to note that although the absolute sorption capacity of the CNT sand fixed beds is lower for SMX than for SPY on a mg/g basis, regeneration of the colu mns after 5 cycles reduced the sorption capacity of both antibiotics by a similar fraction (52% for SMX and 56% for SPY). Chapter Conclusions The removal of sulfamethoxazole ( SMX) and sulfapyridine (SPY) from aqueous solutions by fixed bed columns was tes ted under various conditions. Experimental results indicate that, from the multiple factors that can be varied in a fixed bed filtration system (CNT incorporation, pH, bed depth, adsorbent dosage, a d sorbate initial concentration, and flow rate), the effect of pH most strongly affects column adsorption capacity. The mechanism driving this phenomenon can be attributed to the level of protonation of both the antibiotics and the functional groups on the CNT surfaces. Of the three CNT incorporation methods teste d, dispersed CNTs rendered the columns with greater adsorption capacity than undispersed CNTs due to the greater surface area. The Bed Depth Service Time (BDST) model well describes the relationship between service time and bed depth and gives the column design parameters. The increase of adsorbent dosage slightly d ecreases the capacity of the filters, but extends the lifetime of the filter. Higher adsorbate initial concentration offers greater mass transfer driving force to move antibiotics towards CNTs and thus enhances the efficiency of antibiotic sorption. The increase of flow rate results in the decrease of contact time between th e antibiotics and CNTs and accelerates the movement of PAZ downwards in the fixed bed, which contributes to decreased adsorption capacity, since a larger fraction of fixed bed material is left behind with sorbing capabilities before the PAZ moves through t he entire

PAGE 161

16 1 column height. For regeneration of the fixed beds, small fractions of SMX and SPY were irreversibly sorbed from the sand/CNT porous medium. Thus, each regeneration cycle reduced the capacity of the bed by 8 26% for both types of antibiotics.

PAGE 162

162 Table 8 1 Chemical structures and properties of sulfamethoxazole (SMX) and sulfapyridine (SPY) Name Structure Molecular weight (g/mol) Acidity constant pK a Speciation pH 3.0 pH 5.6 pH 7.0 pH 9.0 SMX 253.3 pK a1 = 1.8 pK a2 = 5.6 SMX + (6%) & SMX (94%) SMX (50%) & SMX (50%) SMX (4%) & SMX (96%) SMX SPY 249.3 pK a1 = 2.3 pK a2 = 8.4 SPY + (17%) & SPY (83%) SPY SPY (96%) & SPY (4%) SPY (20%) & SPY (80%)

PAGE 163

163 Table 8 2 Summary of transport and adsorption parameters of the two antibiotics in CNT/sand columns Treatment SMX SPY t b (min) t e (min) v z (cm/min) Capacity (mg/g) t b (min) t e (min) v z (cm/min) Capacity (mg/g) Introduction Layered 28.5 43.5 0.42 68.5 34.5 44.5 0.37 89.4 Mixed 27.7 44.5 0.43 69.3 33.6 51.5 0.37 91.9 Dispersed 29.6 47.5 0.41 77.7 35.4 55.4 0.35 97.8 pH 3.0 31.7 49.7 0.38 84.9 33.4 50.4 0.36 93.9 5.6 27.6 44.5 0.43 69.3 33.5 51.5 0.37 91.9 7.0 22.6 32.7 0.53 45.8 31.6 49.3 0.39 86.1 9.0 16.6 23.5 0.76 11.7 29.5 44.6 0.40 77.8 Bed depth 6 cm 8.5 20.6 0.39 72.8 9.5 25.5 0.34 97.1 9 cm 15.5 26.5 0.41 71.6 17.5 32.5 0.36 94.5 12 cm 22.5 33.6 0.41 76.9 25.6 37.7 0.37 94.7 15 cm 27.4 44.5 0.42 69.3 33.6 51.8 0.37 91.9 Adsorbent dosage 10 mg 27.5 44.6 0.42 69.3 33.5 51.3 0.37 91.9 20 mg 45.4 62.3 0.28 70.0 53.3 73.9 0.24 86.8 30 mg 58.3 76.9 0.23 64.9 70.4 101.3 0.18 85.1 40 mg 66.7 95.4 0.19 62.8 85.8 122.8 0.15 82.4 Adsorbate C 0 10 mg/L 37.4 53.5 0.33 54.8 51.8 62.4 0.26 80.9 20 mg/L 27.4 44.5 0.42 69.3 33.7 51.4 0.37 91.9 30 mg/L 26.6 37.5 0.48 84.8 25.5 46.8 0.42 109.6 40 mg/L 25.6 32.5 0.52 92.0 25.5 41.5 0.45 123.4 Flow rate 1 ml/min 59.3 93.5 0.20 78.3 67.0 103.2 0.18 94.9 2 ml/min 27.7 44.7 0.43 69.3 33.6 51.3 0.37 91.9 3 ml/min 19.8 27.7 0.65 67.7 22.3 29.7 0.55 90.2 4 ml/min 14.8 19.9 0.91 61.9 16.8 21.7 0.79 83.6 Th e parameters include breakpoint time (t b ), exhausted time (t e ), adsorption zone velocity (v z ), and column capacity

PAGE 164

164 Table 8 3 Definition and formula of the primary adsorption zone (PAZ) parameters used in this study Parameter Unit Definition Formula h cm B ed depth C 0 mg/L Influent antibiotic concentration C t mg/L effluent antibiotic concentration at time t during the experiment t b min Breakpoint time when the normalized effluent concentration (C/C 0 ) reaches 5% t e min Time when C/C 0 reaches 95%, also defined as the time when the column is exhausted/ spent t z min Time required for the PAZ to move through a fix bed column after its establishment F F raction of the PAZ that still has adsorption ability (i.e., unspent adsorbent fraction in the PAZ) t f min Elapse of time between initial injection and the breakpoint, also defined as the time required for the initial form of the PAZ v z cm/min Velocity of PAZ passing through the bed h z cm Height of the adsorption zone s Saturation percentage of the total column at breakthrough

PAGE 165

165 Figure 8 1 B reakthrough curves of sulfamethoxazole (SMX) and sulfapyridine (SPY) in the sand/CNT fixed bed columns under various CNT incorporation m ethods

PAGE 166

166 Figure 8 2 B reakthrough curves of sulfamethoxazole (SMX) and sulfapyridine (SPY) in the sand/CNT fixed bed columns under various pHs

PAGE 167

167 Figure 8 3 B reakthrough curves of sulfamethoxazole (SMX) and sulfapyridine (SPY) in the sand/CNT fixed bed columns under various bed depths

PAGE 168

168 Figure 8 4 Bed Depth Service Time (BDST) model for sulfamethoxazole (SMX) and sulfapyridine (SPY) at breakpoint (C/C 0 = 5%) and exhausted point (C/C 0 = 95%)

PAGE 169

169 Figure 8 5 B reakthrough curves of sulfamethoxazole (S MX) and sulfapyridine (SPY) in the sand/CNT fixed bed columns under various adsorbent dosages

PAGE 170

170 Figure 8 6 Br eakthrough curves of sulfamethoxazole (SMX) and sulfapyridine (SPY) in the sand/CNT fixed bed columns under various adsorbate initial concentration s

PAGE 171

171 Figure 8 7 B reakthrough curves of sulfamethoxazole (SMX) and sulfapyridine (SPY) in the sand/CNT fixed bed columns under various flow rates

PAGE 172

172 Figure 8 8 R elationship between changes in column capacity (mg/g) of (A) sulfamethoxazole (SMX) and (B) sulfapyridine (S PY) for 5 fixed bed regeneration cycles

PAGE 173

173 CHAPTER 9 9 CONCLUSION S ummary There is increasing concern over the environmental impact and health risks of carbon nanotubes (CNTs) because they may be released into soil and groundwater systems. T he PhD dissertation systematically investigated the fate and transport of CNTs in the porous media as well as environmental application of CNTs in wastewater treatment to remove heavy metals and antibiotics. In Chapter 2, we examined the transport behavio r of two NPs, silver nanoparticles and CNTs, in saturated porous media. The experimental results showed that the surfactant solubilized NPs were highly mobile in the saturated porous media. The transport of CNTs in the column was similar to that of colloid al montmorillonite and their recovery rates were around 100%. Predictions of the transport model matched the experimental breakthrough data of the two engineered NPs well. Our results indicate that theories and models of colloid transport in porous media m ay be applicable to describe the fate and behavior of engineered NPs under certain circumstances. In Chapter 3 we systematically investigated the transport, deposition, and mobilization behaviors of multi walled carbon nanotubes (MWNTs) in saturated colum ns packed with acid cleaned glass beads and quartz sand with two different grain sizes. Combined effect s of pH (5.6 and 10) and ionic strength (IS: DI water, 1 mM, and 10 mM) on the fate and transport of the MWNTs in the columns were examined. MWNTs were relatively mobile in all the tested conditions with DI water as the experimental solution. Their depositi on in the saturated porous media, however, was very sensitive to solution chemistry, particularly IS. Slight increase in solution IS (1 mM) caused strong deposition

PAGE 174

174 of MWNTs in both quartz sand (> 44 %) and glass beads (>39%). Mobilization experimental resul ts indicated that most of the MWNT attachment (> 73 %) to the porous media was irreversible and reduction in solution IS only caused a small portion of re entrainment (<2 7 %) of deposited MWNT for all the tested conditions. This indicates that more MWNTs are trapped in the primary minimum although the deposition of MWNTs in saturated porous media occurs in both primary and secondary minimum. It is suggested that, under unfavorable conditions, w eak associated MWNTs in the secondary minimum may be transferred i nto the primary minimum due to the effect of hydrodynamic force and/or local favorable sites associated with surface heterogeneity. In Chapter 4 a series of laboratory experiments were conducted to explore the transport mechanisms of sonication shortened sodium dodecylbenzene sulfonate (SDBS) dispersed single walled nanotubes (SWNTs) in both saturated and unsaturated sand columns. Laboratory columns packed with quartz sand with different combinations of moisture content and grain size distribution were u sed to examine the breakthrough behavior of the SDBS dispersed SWNTs. Bubble column experiments were also conducted to study the interactions between the SDBS dispersed SWNTs and the air water interface. Packed column experimental results showed that the S DBS dispersed SWNTs were highly mobile for most of the experimental conditions tested. The surface deposition of the SWNTs in the sand columns was low because all the interactive surfaces were negatively charged. Physical trapping was not observed for the SWNTs in the saturated porous media of different grain size distributions because the SWNTs might orient parallel to the streamlines in flow to reduce their retention. Retention of the SWNTs in unsaturated porous media occurred only at a very low moisture content

PAGE 175

175 (<0.10). Otherwise, reduction in moisture content showed little impact on the retention and transport of the SWNTs in unsaturated porous media. Findings from the bubble column experiments confirmed that the SDBS dispersed SWNTs did not attach to t he air water interface. In Chapter 5 the effect of different surface modification methods was investigated including oxidization surfactant ( sodium dodecylbenzene sulfonate) coating, and humic acid coating, on SWNT stability and their mobility in granu lar porous media of different grain sizes under both saturated and unsaturated conditions. Characterization and stability studies demonstrated that the three surface modification methods were all effective in solubilizing and stabilizing the SWNTs in aqueo us solutions. Packed sand column experiments showed that although the three surface medication methods showed different effect on the retention and transport of SWNTs in the columns, all the modified SWNTs were highly mobile. Compared with the other two su rface modification methods, the h umic acid coating method introduced the highest mobility to the SWNTs. While reductions in moisture content in the porous media could promote the retention of the surface modified SWNTs in some sand columns, results from bubble column experiment suggested that only oxidized SWNTs were retention in unsaturated porous media through attachment on air water interface. Other mechanisms such as grain surface attachment and thin water film straining could also be responsible for the retention of the SWNTs in unsaturated porous media. An a dvection dispersion model was used to simulate the experimental data of surface modified SWNT retention and transport in porous media. The model results matched data very well, suggesting that

PAGE 176

176 mat hematical models can be used as a tool to monitor and predict the fate and transport of SWNTs in soil and groundwater systems. In Chapter 6 laboratory column and modeling experiments were conducted to mechanistically compare the retention and transport of two types of functionalized CNTs (i.e., single walled nanotubes and multi walled nanotubes ) in acid cleaned, baked, and natural sand under unfavorable conditions. The CNTs were highly mobile in the acid cleaned sand columns but showed little transport in the both natural and baked sand columns. In addition, the retention of the CNTs in the both baked and natural sand was strong and almost irreversible even after reverse, high velocity, or surfactant flow flushing. Both experimental and modeling results sho wed that pH is one of the factors dominating CNT retention and transport in natural and baked sand. Retention of the functionalized CNTs in the natural and baked sand columns reduced dramatically when the system pH increased. Our results suggest that the retention and transport of the functionalized CNTs in natural sand porous media w ere mainly controlled by strong surface deposition through the electrostatic and/or hydrogen bonding attractions between surface function groups of the CNTs and metal oxyhydro xide impurities on the sand surfaces. In Chapter 7 batch sorption and fixed bed experiments were conducted to examine the ability of functionalized multi walled CNTs as filter media to remove two heavy metal ions (Pb 2+ and Cu 2+ ) from infiltrating water. Batch sorption experiments confirmed the strong sorption affinity of the CNTs for Pb 2+ and Cu 2+ in both single and dual metal solution systems. In addition, sonication promoted dispersion of the CNT particles enhanced their heavy metal sorption capacity by 23.9 32.2%. For column

PAGE 177

177 experiments, laboratory scale fixed bed columns were packed with CNTs and natural quartz sand by three different packing: layered, mixed, and deposited. While all the three packing methods enhanced the fixed bed filtering efficienc y of Pb 2+ and Cu 2+ from single and dual metal systems, the CNT deposited packing method was superior. Although the amount of the CNTs added into the fixed bed columns was only 0.006% (w/w) of the sand, they s ignificantly improved the fixed efficiency of Pb 2+ and Cu 2+ by 55% 75% and 31% 57%, respectively. Findings from this study demonstrate that functionalized multi walled CNTs, together with natural sand, can be used to effectively and safely remove heavy metals from water. In Chapter 8 we systematically investigated the removal of SMX and SPY by Carbon nanotubes (CNTs) in fixed bed columns under a broad range of conditions including: CNT incorporation method, solution pH, bed depth, adsorbent dosage, adsorbate initial concentration, and fl ow rate. Fixed bed experiments showed that pH is a key factor that affects the adsorption capacity of antibiotics to CNTs. The Bed Depth Service Time model describes well the relationship between service time and bed depth and can be used to design appropr iate column parameters. During fixed bed regeneration, small amounts of SMX (3%) and SPY ( 9 %) were irreversibly bonded to the sand/CNT porous media, thus reducing the column capacity for subsequent reuse from 67.9 to 50.4 mg/g for SMX and from 91.9 to 72.9 mg/g for SPY The reduced column capacity resulted from decreased available adsorption sites and repulsion (i.e., blocking) of incoming antibiotics from those previously adsorbed. Findings from this study demonstrate that fixed bed columns packed with CNT s can be efficiently used and regenerated to remove antibiotics from water.

PAGE 178

178 Recommendations for Future Work R esults from this Ph D research not also advance current understand ing of the fate and transport of CNTs in saturated and unsaturated porous media but also inform the development of innovative and high efficiency methods to incorporate CNTs in to wastewater treatment technologies Followings provide recommendations for future work: Develop models to better quantify interaction energies/forces betwe en two CNT particles o r a CNT particle and porous medium (e g. sand grain and glass beads ) ; D etermine whether CNT based filters can be used to treat wastewater containing multiple contaminants, including both organic and inorganic compounds. Develop str ategies or technologies to apply and evaluate the CNT based filters in a pilot plant for wastewater treatment.

PAGE 179

179 APPENDIX A A SU PPORTING INFORMATION FOR CHAPTER 3 A 1 DLVO Calculations The Derjaguin Landau Verwey Overbeek (DLVO) theory was used to estimate the interaction energy between functionalized CNTs and the porous media by combining the van der Waals attraction and electrical double layer (EDL) repulsion. Because CNTs are tubular, effective hydrodynamic diameter was used to calculate the DLVO energy profiles. The Lifshitz van der Waals attraction energy ( ) for a sphere plate system can be written as [ 102 ] : ( A 1) where A 132 h is the separation distance, and r is the radius of the particle. The EDL repulsion energy ( )for a sphere plate system can be written as [ 102 ] : ( A 2) where is the dielectric constant of the medium (78.4 for water), 0 is the vacuum permittivity (8.85410 12 C 2 N 1 m 2 ), k (1.38110 23 C 2 J K 1 ),

PAGE 180

180 T is the temperature, z is the valence of electrolyte, e is the electron charge (1.60210 19 C), 1 and 2 are the surface potential of the particle and the grain surface, and is the reciprocal of the Debye length. From measured and grain can be determined following van Oss et al. [ 103 ] : ( A 3) where d is the distance between the surface of the charged particle and the slipping plane and usually taken as 5 angstroms (van Oss et al. 1990). The total DLVO interaction energies between the particles and the grain were determined using equations (1) (3) an d were normalized with kT

PAGE 181

181 Figure A 1 Absorbance spectra (A) and calibration curves (B) of functionalized MW NTs.

PAGE 182

182 Figure A 2 Stability of functionalized MWNTs with DI water, 1 mM and 10 mM ionic strength (IS) at (A) pH 5.6 and (B) pH 10.0

PAGE 183

183 Figure A 3 DLVO calculation for MWNTs stability using hydrodynamic diameter ( D h ), physical diameter ( D ), and length ( L )

PAGE 184

184 Figure A 4 DLVO energy profiles between MWNT particles and grains calculated using physical diameter ( D ) of MWNTs

PAGE 185

185 Figure A 5 DLV O energy profiles between MWNT particles and grains calculated using physical Length ( L ) of MWNTs

PAGE 186

186 APPENDIX B B SU PPORTING INFORMATION FOR CHAPTER 4 B 1 DLVO energy calculations The classic DLVO theory was used to estimate the interaction energy between the SWNT and porous medium and between the SWNT and air water interface at the experimental conditions by combining the van der Waals attraction and electrical double layer (EDL) r epulsion. Because SWNTs are tubular, the tube length ( D e = 400 nm) and the tube diameter ( D e = 0.8 nm) were used as effective size to calculate the DLVO energy profiles [ 118 ] The Lifshitz van der Waals attraction energy ( ) for a sphere plate system can be written as [ 102 ] : ( B 1) where A 132 h is the separation distance, and r is the radius of the particle. The Hamaker constants of between SWNT and quartz sand and between SWNT and air water interface are 8.5 10 21 and 1.8110 21 J, respectively [ 118 167 ] The EDL repulsion energy ( )for a sphere plate system can be written as [ 102 ] : ( B 2 )

PAGE 187

187 where is the dielectric constant of the medium (78.4 for water), 0 is the vacuum permittivity (8.85410 12 C 2 N 1 m 2 ), k 23 C 2 J K 1 ), T is the temperature, z is the valence of electrolyte, e is the electron charge (1.60210 19 C), 1 and 2 are the surface potential of the particle and the sand surface, and is particles and sand can be determined following van Oss et al. [ 103 ] : ( B 3 ) where d is the distance between the surface of the charged particle and the slipping plane and usually taken as 5 angstroms (van Oss et al. 1990). The total DLVO interaction energies between the particles and the sand were determined using equations (1) (3) and were normalized with kT.

PAGE 188

188 Figure B 1 Stability of the SDBS dispersed SWNT suspension

PAGE 189

189 APPENDIX C C SU PPORTING INFORMATION FOR CHAPTER 6 Figure C 1 Stability of functionalized CNT suspensions.

PAGE 190

190 Figure C 2 Absorbance spectra (a) and calibration curves (b) of functionalized SWNTs and MWNTs a b

PAGE 191

191 C 1 DLVO Calculations The Derjaguin Landau Verwey Overbeek (DLVO) theory was used to estimate the interaction energy between the two functionalized CNTs and the two poro us media by combining the van der Waals attraction and electrical double layer (EDL) repulsion. Because CNTs are tubular, effective hydrodynamic diameter ( D e = 179.8 nm for MWNT; D e = 152.7 nm for SWNT) was used to calculate the DLVO energy profiles. The L ifshitz van der Waals attraction energy ( ) for a sphere plate system can be written as [ 102 ] : ( C 1) where A 132 n be determined from the Hamaker constant of each material h is the separation distance, and r is the radius of the particle. The Hamaker constant of between CNT and quartz sand is 8.5 10 21 J [ 118 ] The EDL repulsion energy ( )for a sphere plate system can be written as [ 102 ] : ( C 2 ) where is the dielectric constant of the medium (78.4 for water), 0 is the vacuum permittivity (8.85410 12 C 2 N 1 m 2 ), k 23 C 2 J K 1 ),

PAGE 192

192 T is the temperature, z is the valence of electrolyte, e is the electron charge (1.60210 19 C), 1 and 2 are the surface potential of the particle and the sand surface, and is and sand can be determined following van Oss et al. [ 103 ] : ( C 3) where d is the distance between the surface of the charged particle and the slipping plane and usually taken as 5 angstroms (van Oss et al. 1990). The total DLVO interaction energies between the particles and the sand were determined using equations (1) (3) and were normalized with kT (Figure C 3).

PAGE 193

193 Figure C 3 DLVO energy between the CNTs and porous media.

PAGE 194

194 C 2 PZC Point of Zero Charge (PZC) of the porous media (i.e., acid clean baked and natural sand) and CNTs (i.e., SWNT and MWNT) were determined using th e mass titration method as described by Noh and Schwarz [ 246 ] Three aqueous NaNO 3 (0.01 M) solutions of different pHs (3, 6 and 11) were prepared using HNO 3 (0.1 M) and NaOH (0.1 M). The solid samples were then added to the solutions (20 mL) at different mass ratios (1%, 5%, 10%, 15%, 20%). The final pH of the mixture was measured after 24 h of shaking at 25 0.1 C. P ZC was determined as the converging pH value from the pH vs. sample mass curves (Figure S4). Figure C 4 Mass titration curves of the porous media and CNTs.

PAGE 195

195 Table C 1 PZCs of metal (hydro)oxides. (Hydro)Oxides PZC [ 184 190 191 ] Aluminium oxide (Al2O3) 7.6 9.2 Gibbsite (Al(OH)3) 9.1 10.1 Magnetite (FeOFe2O3) 6.5 8.2 Fe2O3) 6.1 7.5 Fe2O3) 7.0 9.3 FeO(OH)) 7.2 9.6 8.2 8.7 Nickel oxide (NiO) 9.9 11.3 Nickel oxide (Ni(OH)2) 8.8 11 (Fe0.89Ni0.11)yOz 7.2 8.4

PAGE 196

196 APPENDIX D D SU PPORTING INFORMATION FOR CHAPTER 7 Figure D 1 Mass titration curves of the porous media and CNTs. Point of Zero Charge (PZC) of the sand and the CNTs were determined using the mass titration method

PAGE 197

197 APPENDIX E E SU PPORTING INFORMATION FOR CHAPTER 8 Figure E 1 Release of sulfamethoxazole (SMX) and sulfapyridine (SPY) from the post adsorption CNT/sand fixed bed columns during regeneration.

PAGE 198

198 Table E 1 Summary of transport and adsorption parameters of sulfamethoxazole (SMX) and sulfapyridine (SPY) in CNT/sand columns with different CNT incorporation methods Introduction t b (min) t e (min) t z (min) F t f (min) v z (cm/min) h z (cm) s (%) Capacity (mg/g) SMX Layered 28.5 43.5 15.0 0.48 7.9 0.42 6.3 77.9 68.5 Mixed 27.7 44.5 16.8 0.42 9.7 0.43 7.2 72.1 69.3 Dispersed 29.6 47.5 17.9 0.38 11.1 0.41 7.4 69.6 77.7 SPY Layered 34.5 44.5 10.0 0.59 4.1 0.37 3.7 89.9 89.4 Mixed 33.6 51.5 17.9 0.36 11.5 0.37 6.7 71.4 91.9 Dispersed 35.4 55.4 20.0 0.37 12.6 0.35 7.0 70.6 97.8 Th e parameters include breakpoint time (t b ), exhausted time (t e ), adsorption zone unit movement time (t z ), fraction (F), adsorption zone initial formation time (t f ), adsorption zone velocity (v z ), adsorption zone height (h z ), saturation percentage (s) and column capacity

PAGE 199

199 Table E 2 Summary of transport and adsorption parameters of sulfamethoxazole (SMX) and sulfapyridine (SPY) in CNT/sand columns under various pH conditions pH t b (min) t e (min) t z (min) F t f (min) v z (cm/min) h z (cm) s (%) Capacity (mg/g) SMX 3.0 31.7 49.7 18.0 0.40 10.7 0.38 6.9 72.5 84.9 5.6 27.6 44.5 16.9 0.42 9.8 0.43 7.3 71.8 69.3 7.0 22.6 32.7 10.1 0.58 4.2 0.53 5.3 85.1 45.8 9.0 16.6 23.5 6.9 0.46 3.7 0.76 5.2 81.3 11.7 SPY 3.0 33.4 50.4 17.0 0.45 9.3 0.36 6.2 77.5 93.9 5.6 33.5 51.5 18.0 0.41 10.6 0.37 6.6 74.1 91.9 7.0 31.6 49.3 17.7 0.40 10.6 0.39 6.9 72.5 86.1 9.0 29.5 44.6 15.1 0.53 7.0 0.40 6.0 81.3 77.8 Th e parameters include breakpoint time (t b ), exhausted time (t e ), adsorption zone unit movement time (t z ), fraction (F), adsorption zone initial formation time (t f ), adsorption zone velocity (v z ), adsorption zone height (h z ), saturation percentage (s) and column capacity

PAGE 200

200 Table E 3 Summary of transport and adsorption parameters of sulfamethoxazole (SMX) and sulfapyridine (SPY) in CNT/sand columns under various bed depth conditions Bed depth t b (min) t e (min) t z (min) F t f (min) v z (cm/min) h z (cm) s (%) Capacity (mg/g) SMX 6 cm 8.5 20.6 12.1 0.57 5.2 0.39 4.7 65.9 72.8 9 cm 15.5 26.5 11.0 0.57 4.7 0.41 4.5 78.4 71.6 12 cm 22.5 33.6 11.1 0.63 4.1 0.41 4.5 86.2 76.9 15 cm 27.4 44.5 17.1 0.47 9.0 0.42 7.2 74.5 69.3 SPY 6 cm 9.5 25.5 16.0 0.51 7.9 0.34 5.5 55.0 97.1 9 cm 17.5 32.5 15.0 0.50 7.4 0.36 5.4 70.3 94.5 12 cm 25.6 37.7 12.1 0.53 5.6 0.37 4.5 82.4 94.7 15 cm 33.6 51.8 18.2 0.36 11.7 0.37 6.8 71.0 91.9 Th e parameters include breakpoint time (t b ), exhausted time (t e ), adsorption zone unit movement time (t z ), fraction (F), adsorption zone initial formation time (t f ), adsorption zone velocity (v z ), adsorption zone height (h z ), saturation percentage (s) and column capacity

PAGE 201

201 Table E 4 Summary of transport and adsorption parameters of sulfamethoxazole (SMX) and sulfapyridine (SPY) in CNT/sand columns under various adsorbent dosage conditions Adsorbent dosage t b (min) t e (min) t z (min) F t f (min) v z (cm/min) h z (cm) s (%) Capacity (mg/g) SMX 10 mg 27.5 44.6 17.1 0.47 9.0 0.42 7.2 74.8 69.3 20 mg 45.4 62.3 16.9 0.48 8.8 0.28 4.7 83.4 70.0 30 mg 58.3 76.9 18.6 0.44 10.5 0.23 4.2 84.2 64.9 40 mg 66.7 95.4 28.7 0.48 14.9 0.19 5.4 81.4 62.8 SPY 10 mg 33.5 51.3 17.8 0.42 10.4 0.37 6.5 74.6 91.9 20 mg 53.3 73.9 20.6 0.41 12.2 0.24 5.0 80.3 86.8 30 mg 70.4 101.3 30.9 0.39 19.0 0.18 5.6 76.9 85.1 40 mg 85.8 122.8 37.0 0.40 22.1 0.15 5.5 78.1 82.4 Th e parameters include breakpoint time (t b ), exhausted time (t e ), adsorption zone unit movement time (t z ), fraction (F), adsorption zone initial formation time (t f ), adsorption zone velocity (v z ), adsorption zone height (h z ), saturation percentage (s) and column capacity

PAGE 202

202 Table E 5 Summary of transport and adsorption parameters of sulfamethoxazole (SMX) and sulfapyridine (SPY) in CNT/sand columns under various adsorbate initial concentration conditions Adsorbate C 0 t b (min) t e (min) t z (min) F t f (min) v z (cm/min ) h z (cm) s (%) Capacity (mg/g) SMX 10 mg/L 37.4 53.5 16.1 0.49 8.1 0.33 5.3 82.1 54.8 20 mg/L 27.4 44.5 17.1 0.47 9.0 0.42 7.2 74.5 69.3 30 mg/L 26.6 37.5 10.9 0.44 6.1 0.48 5.2 80.4 84.8 40 mg/L 25.6 32.5 6.9 0.45 3.8 0.52 3.6 86.7 92.0 SPY 10 mg/L 51.8 62.4 10.6 0.56 4.7 0.26 2.8 91.9 80.9 20 mg/L 33.7 51.4 17.7 0.36 11.3 0.37 6.6 72.0 91.9 30 mg/L 25.5 46.8 21.3 0.49 10.9 0.42 8.9 69.7 109.6 40 mg/L 25.5 41.5 16.0 0.50 8.0 0.45 7.2 76.2 123.4 Th e parameters include breakpoint time (t b ), exhausted time (t e ), adsorption zone unit movement time (t z ), fraction (F), adsorption zone initial formation time (t f ), adsorption zone velocity (v z ), adsorption zone height (h z ), saturation percentage (s) and column capacity

PAGE 203

203 Table E 6 Summary of transport and adsorption parameters of sulfamethoxazole (SMX) and sulfapyridine (SPY) in CNT/sand columns under various flow rate conditions Flow rate t b (min) t e (min) t z (min) F t f (min) v z (cm/min) h z (cm) s (%) Capacity (mg/g) SMX 1 ml/min 59.3 93.5 34.2 0.42 19.8 0.20 7.0 73.1 78.3 2 ml/min 27.7 44.7 17.0 0.42 9.8 0.43 7.3 71.7 69.3 3 ml/min 19.8 27.7 7.9 0.40 4.7 0.65 5.2 79.5 67.7 4 ml/min 14.8 19.9 5.1 0.32 3.5 0.91 4.7 78.8 61.9 SPY 1 ml/min 67.0 103.2 36.2 0.46 19.7 0.18 6.5 76.4 94.9 2 ml/min 33.6 51.3 17.7 0.36 11.3 0.37 6.6 71.9 91.9 3 ml/min 22.3 29.7 7.4 0.66 2.5 0.55 4.1 90.6 90.2 4 ml/min 16.8 21.7 4.9 0.46 2.7 0.79 3.9 86.1 83.6 Th e parameters include breakpoint time (t b ), exhausted time (t e ), adsorption zone unit movement time (t z ), fraction (F), adsorption zone initial formation time (t f ), adsorption zone velocity (v z ), adsorption zone height (h z ), saturation percentage (s) and column capacity

PAGE 204

204 LIST OF REFERENCES 1. Hofmann, T. and von der Kammer, F., Estimating the relevance of engineered carbonaceous nanoparticle facilitated transport of hydrophobic organic contaminants in porous media. Environmental Pollution, 2009. 157 (4): p. 1117 1126. 2. Perez, S., Farre, M., an d Barcelo, D., Analysis, behavior and ecotoxicity of carbon based nanomaterials in the aquatic environment. Trac trends In Analytical Chemistry, 2009. 28 (6): p. 820 832. 3. Petersen, E.J., Pinto, R.A., Landrum, P.F., and Weber, W.J., Influence of carbon na notubes on pyrene bioaccumulation from contaminated soils by earthworms. Environmental Science & Technology, 2009. 43 (11): p. 4181 4187. 4. Peralta Videa, J.R., Zhao, L.J., Lopez Moreno, M.L., de la Rosa, G., Hong, J., and Gardea Torresdey, J.L., Nanomater ials and the environment: A review for the biennium 2008 2010. Journal Of Hazardous Materials, 2011. 186 (1): p. 1 15. 5. Hardman, R., A toxicologic review of quantum dots: Toxicity depends on physicochemical and environmental factors. Environmental Health Perspectives, 2006. 114 (2): p. 165 172. 6. Crespy, A., Boleve, A., and Revil, A., Influence of the dukhin and reynolds numbers on the apparent zeta potential of granular porous media. Journal Of Colloid And Interface Science, 2007. 305 (1): p. 188 194. 7. Oberdorster, G., Oberdorster, E., and Oberdorster, J., Nanotoxicology: An emerging discipline evolving from studies of ultrafine particles. Environmental Health Perspectives, 2005. 113 (7): p. 823 839. 8. Zhang, W.X., Nanoscale iron particles for environmen tal remediation: An overview. Journal of Nanoparticle Research, 2003. 5 (3 4): p. 323 332. 9. Chen, K.L., Smith, B.A., Ball, W.P., and Fairbrother, D.H., Assessing the colloidal properties of engineered nanoparticles in water: Case studies from fullerene c 60 nanoparticles and carbon nanotubes. Environmental Chemistry, 2009. 7 (1): p. 10 27. 10. Klaine, S.J., Alvarez, P.J.J., Batley, G.E., Fernandes, T.F., Handy, R.D., Lyon, D.Y., Mahendra, S., McLaughlin, M.J., and Lead, J.R., Nanomaterials in the environmen t: Behavior, fate, bioavailability, and effects. Environmental Toxicology And Chemistry, 2008. 27 (9): p. 1825 1851. 11. Nel, A., Xia, T., Madler, L., and Li, N., Toxic potential of materials at the nanolevel. Science, 2006. 311 (5761): p. 622 627.

PAGE 205

205 12. Zheng L.X., O'Connell, M.J., Doorn, S.K., Liao, X.Z., Zhao, Y.H., Akhadov, E.A., Hoffbauer, M.A., Roop, B.J., Jia, Q.X., Dye, R.C., Peterson, D.E., Huang, S.M., Liu, J., and Zhu, Y.T., Ultralong single wall carbon nanotubes. Nature Materials, 2004. 3 (10): p. 6 73 676. 13. Pan, B. and Xing, B., Adsorption mechanisms of organic chemicals on carbon nanotubes. Environmental Science & Technology, 2008. 42 (24): p. 9005 9013. 14. Nowack, B. and Bucheli, T.D., Occurrence, behavior and effects of nanoparticles in the env ironment. Environmental Pollution, 2007. 150 (1): p. 5 22. 15. Lam, C.W., James, J.T., McCluskey, R., Arepalli, S., and Hunter, R.L., A review of carbon nanotube toxicity and assessment of potential occupational and environmental health risks. Critical Revi ews In Toxicology, 2006. 36 (3): p. 189 217. 16. Mauter, M.S. and Elimelech, M., Environmental applications of carbon based nanomaterials. Environmental Science & Technology, 2008. 42 (16): p. 5843 5859. 17. Avouris, P., Freitag, M., and Perebeinos, V., Carb on nanotube photonics and optoelectronics. Nature Photonics, 2008. 2 (6): p. 341 350. 18. Takagi, S., Eguchi, M., Tryk, D.A., and Inoue, H., Porphyrin photochemistry in inorganic/organic hybrid materials: Clays, layered semiconductors, nanotubes, and mesopo rous materials. Journal of Photochemistry and Photobiology C Photochemistry Reviews, 2006. 7 (2 3): p. 104 126. 19. Gruner, G., Carbon nanotube transistors for biosensing applications. Analytical and Bioanalytical Chemistry, 2006. 384 (2): p. 322 335. 20. Wi esner, M.R., Hotze, E.M., Brant, J.A., and Espinasse, B., Nanomaterials as possible contaminants: The fullerene example. Water Science And Technology, 2008. 57 (3): p. 305 310. 21. Guzman, K.A.D., Taylor, M.R., and Banfield, J.F., Environmental risks of nan otechnology: National nanotechnology initiative funding, 2000 2004. Environmental Science & Technology, 2006. 40 (5): p. 1401 1407. 22. Warheit, D.B., Laurence, B.R., Reed, K.L., Roach, D.H., Reynolds, G.A.M., and Webb, T.R., Comparative pulmonary toxicity assessment of single wall carbon nanotubes in rats. Toxicological Sciences, 2004. 77 (1): p. 117 125. 23. Monteiro Riviere, N.A., Nemanich, R.J., Inman, A.O., Wang, Y.Y.Y., and Riviere, J.E., Multi walled carbon nanotube interactions with human epidermal ke ratinocytes. Toxicology Letters, 2005. 155 (3): p. 377 384.

PAGE 206

206 24. Kennedy, A.J., Hull, M.S., Steevens, J.A., Dontsova, K.M., Chappell, M.A., Gunter, J.C., and Weiss, C.A., Factors influencing the partitioning and toxicity of nanotubes in the aquatic environme nt. Environmental Toxicology And Chemistry, 2008. 27 (9): p. 1932 1941. 25. Alpatova, A.L., Shan, W.Q., Babica, P., Upham, B.L., Rogensues, A.R., Masten, S.J., Drown, E., Mohanty, A.K., Alocilja, E.C., and Tarabara, V.V., Single walled carbon nanotubes disp ersed in aqueous media via non covalent functionalization: Effect of dispersant on the stability, cytotoxicity, and epigenetic toxicity of nanotube suspensions. Water Research, 2010. 44 (2): p. 505 520. 26. Gao, J., Youn, S., Hovsepyan, A., Llaneza, V.L., W ang, Y., Bitton, G., and Bonzongo, J.C.J., Dispersion and toxicity of selected manufactured nanomaterials in natural river water samples: Effects of water chemical composition. Environmental Science & Technology, 2009. 43 (9): p. 3322 3328. 27. Lecoanet, H. F., Bottero, J.Y., and Wiesner, M.R., Laboratory assessment of the mobility of nanomaterials in porous media. Environmental Science & Technology, 2004. 38 (19): p. 5164 5169. 28. Thomas, K. and Sayre, P., Research strategies for safety evaluation of nanomaterials, part i: Evaluating the human health implications of exposure to nanoscale materials. Toxicological Sciences, 2005. 87 (2): p. 316 321. 29. Cullen, E., O'Carroll, D.M., Yan ful, E.K., and Sleep, B., Simulation of the subsurface mobility of carbon nanoparticles at the field scale. Advances In Water Resources, 2010. 33 (4): p. 361 371. 30. Jaisi, D.P. and Elimelech, M., Single walled carbon nanotubes exhibit limited transport in soil columns. Environmental Science & Technology, 2009. 43 (24): p. 9161 9166. 31. Savage, N. and Diallo, M.S., Nanomaterials and water purification: Opportunities an d challenges. Journal of Nanoparticle Research, 2005. 7 (4 5): p. 331 342. 32. Yu, C., Colloid transport in surface runoff through dense vegetation, 2011, University of Florida: United States -Florida. p. 215. 33. Theron, J., Walker, J.A., and Cloete, T.E ., Nanotechnology and water treatment: Applications and emerging opportunities. Critical Reviews In Microbiology, 2008. 34 (1): p. 43 69. 34. Cho, H. H., Smith, B.A., Wnuk, J.D., Fairbrother, D.H., and Ball, W.P., Influence of surface oxides on the adsorpti on of naphthalene onto multiwalled carbon nanotubes. Environmental Science & Technology, 2008. 42 (8): p. 2899 2905.

PAGE 207

207 35. Chen, W., Duan, L., and Zhu, D., Adsorption of polar and nonpolar organic chemicals to carbon nanotubes. Environmental Science & Technol ogy, 2007. 41 (24): p. 8295 8300. 36. Ryan, J.N. and Elimelech, M., Colloid mobilization and transport in groundwater. Colloids And Surfaces A physicochemical And Engineering Aspects, 1996. 107 : p. 1 56. 37. Kjaergaard, C., Poulsen, T.G., Moldrup, P., and d e Jonge, L.W., Colloid mobilization and transport in undisturbed soil columns. I. Pore structure characterization and tritium transport. Vadose Zone Journal, 2004. 3 (2): p. 413 423. 38. Kjaergaard, C., Moldrup, P., de Jonge, L.W., and Jacobsen, O.H., Collo id mobilization and transport in undisturbed soil columns. Ii. The role of colloid dispersibility and preferential flow. Vadose Zone Journal, 2004. 3 (2): p. 424 433. 39. Zhuang, J., Qi, J., and Jin, Y., Retention and transport of amphiphilic colloids under unsaturated flow conditions: Effect of particle size and surface property. Environmental Science & Technology, 2005. 39 (20): p. 7853 7859. 40. Kretzschmar, R., Barmettler, K., Grolimund, D., Yan, Y.D., Borkovec, M., and Sticher, H., Experimental determina tion of colloid deposition rates and collision efficiencies in natural porous media. Water Resources Research, 1997. 33 (5): p. 1129 1137. 41. Sen, T.K. and Khilar, K.C., Review on subsurface colloids and colloid associated contaminant transport in saturate d porous media. Advances In Colloid And Interface Science, 2006. 119 (2 3): p. 71 96. 42. Bradford, S.A., Simunek, J., Bettahar, M., van Genuchten, M.T., and Yates, S.R., Significance of straining in colloid deposition: Evidence and implications. Water Reso urces Research, 2006. 42 (12). 43. Ginn, T.R., Wood, B.D., Nelson, K.E., Scheibe, T.D., Murphy, E.M., and Clement, T.P., Processes in microbial transport in the natural subsurface. Advances In Water Resources, 2002. 25 (8 12): p. 1017 1042. 44. Kretzschmar, R., Borkovec, M., Grolimund, D., and Elimelech, M., Mobile subsurface colloids and their role in contaminant transport. Advances in Agronomy, Vol 66, 1999. 66 : p. 121 193. 45. Ouyang, Y., Shinde, D., Mansell, R.S., and Harris, W., Colloid enhanced transpor t of chemicals in subsurface environments: A review. Critical Reviews In Environmental Science And Technology, 1996. 26 (2): p. 189 204. 46. Hiemenz, P.C. and Rajagopalan, R., Principles of colloid and surface chemistry1997, New York: Marcel Dekker.

PAGE 208

208 47. Bou rgeat Lami, E., Organic inorganic nanostructured colloids. Journal of Nanoscience and Nanotechnology, 2002. 2 (1): p. 1 24. 48. Guo, L.D. and Santschi, P.H., Composition and cycling of colloids in marine environments. Reviews Of Geophysics, 1997. 35 (1): p. 17 40. 49. Antonietti, M. and Goltner, C., Superstructures of functional colloids: Chemistry on the nanometer scale. Angewandte Chemie International Edition in English, 1997. 36 (9): p. 910 928. 50. Buffle, J. and Leppard, G.G., Characterization of aquatic colloids and macromolecules .1. Structure and behavior of colloidal material. Environmental Science & Technology, 1995. 29 (9): p. 2169 2175. 51. Tufenkji, N. and Elimelech, M., Correlation equation for predicting single collector efficiency in physicochemi cal filtration in saturated porous media. Environmental Science & Technology, 2004. 38 (2): p. 529 536. 52. Bradford, S.A., Simunek, J., Bettahar, M., Van Genuchten, M.T., and Yates, S.R., Modeling colloid attachment, straining, and exclusion in saturated p orous media. Environmental Science & Technology, 2003. 37 (10): p. 2242 2250. 53. Bradford, S.A., Yates, S.R., Bettahar, M., and Simunek, J., Physical factors affecting the transport and fate of colloids in saturated porous media. Water Resources Research, 2002. 38 (12). 54. Elimelech, M. and Omelia, C.R., Kinetics of deposition of collo idal particles in porous media. Environmental Science & Technology, 1990. 24 (10): p. 1528 1536. 55. Torkzaban, S., Kim, H.N., Simunek, J., and Bradford, S.A., Hysteresis of colloid retention and release in saturated porous media during transients in soluti on chemistry. Environmental Science & Technology, 2010. 44 (5): p. 1662 1669. 56. Kim, H.N., Walker, S.L., and Bradford, S.A., Coupled factors influencing the transport and retention of cryptosporidium parvum oocysts in saturated porous media. Water Researc h, 2010. 44 (4): p. 1213 1223. 57. Bradford, S.A., Kim, H.N., Haznedaroglu, B.Z., Torkzaban, S., and Walker, S.L., Coupled factors influencing concentration dependent colloid transport and retention in saturated porous media. Environmental Science & Technology, 2009. 43 (18): p. 6996 7002. 58. Torkzaban, S., Bradford, S.A., van Genuchten, M.T., and Walker, S.L., Colloid transport in unsaturated porous media: The role of water content and ionic strength on particle straining. Journal Of Contaminant Hyd rology, 2008. 96 (1 4): p. 113 127.

PAGE 209

209 59. Bradford, S.A., Torkzaban, S., Leij, F., Simunek, J., and van Genuchten, M.T., Modeling the coupled effects of pore space geometry and velocity on colloid transport and retention. Water Resources Research, 2009. 45 6 0. Bradford, S.A., Bettahar, M., Simunek, J., and van Genuchten, M.T., Straining and attachment of colloids in physically heterogeneous porous media. Vadose Zone Journal, 2004. 3 (2): p. 384 394. 61. Wan, J.M. and Wilson, J.L., Visualization of the role of the gas water interface on the fate and transport of colloids in porous media. Water Resources Research, 1994. 30 (1): p. 11 23. 62. DeNovio, N.M., Saiers, J.E., and Ryan, J.N., Colloid movement in unsaturated porous media: Recent advances and future direct ions. Vadose Zone Journal, 2004. 3 (2): p. 338 351. 63. Lenhart, J.J. and Saiers, J.E., Transport of silica colloids through unsaturated porous media: Experimental results and model comparisons. Environmental Science & Technology, 2002. 36 (4): p. 769 777. 6 4. Schafer, A., Ustohal, P., Harms, H., Stauffer, F., Dracos, T., and Zehnder, A.J.B., Transport of bacteria in unsaturated porous media. Journal Of Contaminant Hydrology, 1998. 33 (1 2): p. 149 169. 65. Wan, J.M. and Tokunaga, T.K., Film straining of collo ids in unsaturated porous media: Conceptual model and experimental testing. Environmental Science & Technology, 1997. 31 (8): p. 2413 2420. 66. Wan, J.M. and Wilson, J.L., Colloid transport in unsaturated porous media. Water Resources Research, 1994. 30 (4): p. 857 864. 67. Mattison, N.T., O'Carroll, D.M., Rowe, R.K., and Petersen, E.J., Impact of porous media grain size on the transport of multi walled carbon nanotubes. Environmental Science & Technology, 2011. 45 (22): p. 9765 9775. 68. Tofighy, M.A. and Moh ammadi, T., Adsorption of divalent heavy metal ions from water using carbon nanotube sheets. Journal Of Hazardous Materials, 2011. 185 (1): p. 140 147. 69. Lee, S., Zhang, Z.T., Wang, X.M., Pfefferle, L.D., and Haller, G.L., Characterization of multi walled carbon nanotubes catalyst supports by point of zero charge. Catalysis Today, 2011. 164 (1): p. 68 73. 70. Chen, C.L., Hu, J., Shao, D.D., Li, J.X., and Wang, X.K., Adsorption behavior of multiwall carbon nanotube/iron oxide magnetic composites for ni(ii) a nd sr(ii). Journal Of Hazardous Materials, 2009. 164 (2 3): p. 923 928.

PAGE 210

210 71. Li, Y.H., Ding, J., Luan, Z.K., Di, Z.C., Zhu, Y.F., Xu, C.L., Wu, D.H., and Wei, B.Q., Competitive adsorption of pb2+, cu2+ and cd2+ ions from aqueous solutions by multiwalled carb on nanotubes. Carbon, 2003. 41 (14): p. 2787 2792. 72. Li, Y.H., Luan, Z.K., Xiao, X., Zhou, X.W., Xu, C.L., Wu, D.H., and Wei, B.Q., Removal of cu2+ ions from aqueous solutions by carbon nanotubes. Adsorption Science & Technology, 2003. 21 (5): p. 475 485. 73. Rao, G.P., Lu, C., and Su, F., Sorption of divalent metal ions from aqueous solution by carbon nanotubes: A review. Separation And Purification Technology, 2007. 58 (1): p. 224 231. 74. Zhang, D., Pan, B., Zhang, H., Ning, P., and Xing, B.S., Contributi on of different sulfamethoxazole species to their overall adsorption on functionalized carbon nanotubes. Environmental Science & Technology, 2010. 44 (10): p. 3806 3811. 75. Zhang, X., Pan, B., Yang, K., Zhang, D., and Hou, J.A., Adsorption of sulfamethoxazole on different types of carbon nanotubes in comparison to other natural adsorbents. Journal of Environmental Science and Health Part a Toxic/Hazardous Substances & Environ mental Engineering, 2010. 45 (12): p. 1625 1634. 76. Tawabini, B., Al Khaldi, S., Atieh, M., and Khaled, M., Removal of mercury from water by multi walled carbon nanotubes. Water Science And Technology, 2010. 61 (3): p. 591 598. 77. Rahbari, M. and Goharrizi A.S., Adsorption of lead(ii) from water by carbon nanotubes: Equilibrium, kinetics, and thermodynamics. Water Environment Research, 2009. 81 (6): p. 598 607. 78. Asharani, P.V., Wu, Y.L., Gong, Z.Y., and Valiyaveettil, S., Toxicity of silver nanoparticles in zebrafish models. Nanotechnology, 2008. 19 (25): p. 79. Wiesner, M.R., Lowry, G.V., Alvarez, P., Dionysiou, D., and Biswas, P., Assessing the risks of manufactured nanomaterials. Environmental Science & Technology, 2006. 40 (14): p. 4336 4345. 80. And reescu, S., Njagi, J., Ispas, C., and Ravalli, M.T., Jem spotlight: Applications of advanced nanomaterials for environmental monitoring. Journal Of Environmental Monitoring, 2009. 11 (1): p. 27 40. 81. Narr, J., Viraraghavan, T., and Jin, Y.C., Applications of nanotechnology in water/wastewater treatment: A review. Fresenius Environmental Bulletin, 2007. 16 (4): p. 320 329. 82. Zhao, Y.L., Xing, G.M., and Chai, Z.F., Nanotoxicology: Are carbon nanotubes safe? Nature Nanotechnology, 2008. 3 (4): p. 191 192.

PAGE 211

211 83. Yu, M.F., Lourie, O., Dyer, M.J., Moloni, K., Kelly, T.F., and Ruoff, R.S., Strength and breaking mechanism of multiwalled carbon nanotubes under tensile load. Science, 2000. 287 (5453): p. 637 640. 84. Roberts, A.P., Mount, A.S., Seda, B., Souther, J., Qi ao, R., Lin, S., Ke, P., Rao, A.M., and Klaine, S.J., In vivo biomodification of lipid coated carbon nanotubes by daphnia magna. Environmental Science & Technology, 2007. 41 (8): p. 3025 3029. 85. Shvedova, A.A., Castranova, V., Kisin, E.R., Schwegler Berry D., Murray, A.R., Gandelsman, V.Z., Maynard, A., and Baron, P., Exposure to carbon nanotube material: Assessment of nanotube cytotoxicity using human keratinocyte cells. Journal Of Toxicology And Environmental Health part A, 2003. 66 (20): p. 1909 1926. 86. Bottini, M., Bruckner, S., Nika, K., Bottini, N., Bellucci, S., Magrini, A., Bergamaschi, A., and Mustelin, T., Multi walled carbon nanotubes induce t lymphocyte apoptosis. Toxicology Letters, 2006. 160 (2): p. 121 126. 87. Manna, S.K., Sarkar, S., Barr J., Wise, K., Barrera, E.V., Jejelowo, O., Rice Ficht, A.C., and Ramesh, G.T., Single walled carbon nanotube induces oxidative stress and activates nuclear transcription factor kappa b in human keratinocytes. Nano Letters, 2005. 5 (9): p. 1676 1684. 88. N avarro, E., Piccapietra, F., Wagner, B., Marconi, F., Kaegi, R., Odzak, N., Sigg, L., and Behra, R., Toxicity of silver nanoparticles to chlamydomonas reinhardtii. Environmental Science & Technology, 2008. 42 (23): p. 8959 8964. 89. Kvitek, L., Vanickova, M ., Panacek, A., Soukupova, J., Dittrich, M., Valentova, E., Prucek, R., Bancirova, M., Milde, D., and Zboril, R., Initial study on the toxicity of silver nanoparticles (nps) against paramecium caudatum. Journal of Physical Chemistry C, 2009. 113 (11): p. 42 96 4300. 90. Oberdorster, E., Zhu, S.Q., Blickley, T.M., McClellan Green, P., and Haasch, M.L., Ecotoxicology of carbon based engineered nanoparticles: Effects of fullerene (c 60) on aquatic organisms. Carbon, 2006. 44 (6): p. 1112 1120. 91. Templeton, R.C. Ferguson, P.L., Washburn, K.M., Scrivens, W.A., and Chandler, G.T., Life cycle effects of single walled carbon nanotubes (swnts) on an estuarine meiobenthic copepod. Environmental Science & Technology, 2006. 40 (23): p. 7387 7393. 92. Doshi, R., Braida, W ., Christodoulatos, C., Wazne, M., and O'Connor, G., Nano aluminum: Transport through sand columns and environmental effects on plants and soil communities. Environmental Research, 2008. 106 (3): p. 296 303. 93. Jaisi, D.P., Saleh, N.B., Blake, R.E., and El imelech, M., Transport of single walled carbon nanotubes in porous media: Filtration mechanisms and reversibility. Environmental Science & Technology, 2008. 42 (22): p. 8317 8323.

PAGE 212

212 94. Johnson, P.R., Sun, N., and Elimelech, M., Colloid transport in geochemic ally heterogeneous porous media: Modeling and measurements. Environmental Science & Technology, 1996. 30 (11): p. 3284 3293. 95. Hermansson, M., The dlvo theory in microbial adhesion. Colloids And Surfaces B biointerfaces, 1999. 14 (1 4): p. 105 119. 96. Yao K.M., Habibian, M.M., and Omelia, C.R., Water and waste water filtration concepts and applications. Environmental Science & Technology, 1971. 5 (11): p. 1105 1112. 97. Elimelech, M., Particle deposition on ideal collectors from dilute flowing suspensions: Mathematical formulation, numerical solutions, and simulations. Separations Technol., 1994. 4 : p. 186 212. 98. Christian, P., Von der Kammer, F., Baalousha, M., and Ho fmann, T., Nanoparticles: Structure, properties, preparation and behaviour in environmental media. Ecotoxicology, 2008. 17 (5): p. 326 343. 99. O'Connell, M.J., Bachilo, S.M., Huffman, C.B., Moore, V.C., Strano, M.S., Haroz, E.H., Rialon, K.L., Boul, P.J., Noon, W.H., Kittrell, C., Ma, J.P., Hauge, R.H., Weisman, R.B., and Smalley, R.E., Band gap fluorescence from individual single walled carbon nanotubes. Science, 2002. 297 (5581): p. 593 596. 100. Wang, R.K., Reeves, R.D., and Ziegler, K.J., Interfacial tra pping of single walled carbon nanotube bundles. Journal Of The American Chemical Society, 2007. 129 : p. 15124 15125. 101. Baker, C., Pradhan, A., Pakstis, L., Pochan, D.J., and Shah, S.I., Synthesis and antibacterial properties of silver nanoparticles. Jou rnal of Nanoscience and Nanotechnology, 2005. 5 (2): p. 244 249. 102. Elimelech, M., Gregory, J., Jia, X., and Williams, R.A., Particle deposition and aggregtion: Measurement, modeling and simulation, 1995, Butterworth Heinemann Ltd. 103. van Oss, C.J., Gie se, R.F., and Costanzo, P.M., Dlvo and non dlvo interactions in hectorite. Clays And Clay Minerals, 1990. 38 (2): p. 151 159. 104. Smith, J., Gao, B., Funabashi, H., Tran, T.N., Luo, D., Ahner, B.A., Steenhuis, T.S., Hay, A.G., and Walter, M.T., Pore scale quantification of colloid transport in saturated porous media. Environmental Science & Technology, 2008. 42 (2): p. 517 523. 105. Medout Marere, V., A simple experimental way of measuring the hamaker constant a(11) of divided solids by immersion calorimetry in apolar liquids. Journal Of Colloid And Interface Science, 2000. 228 (2): p. 434 437.

PAGE 213

213 106. Hahn, M.W., Abadzic, D., and O'Melia, C.R., Aquasols: On the role of secondary minima. Environmental Science & Technology, 2004. 38 (22): p. 5915 5924. 107. Vadillo Rodriguez, V. and Logan, B.E., Localized attraction correlates with bacterial adhesion to glass and metal oxide substrata. Environmental Science & Technology, 2006. 40 (9): p. 2983 2988. 108. de Jonge, L.W., Kjaergaard, C., and Moldrup, P., Colloids and co lloid facilitated transport of contaminants in soils: An introduction. Vadose Zone Journal, 2004. 3 (2): p. 321 325. 109. Phenrat, T., Saleh, N., Sirk, K., Kim, H.J., Tilton, R.D., and Lowry, G.V., Stabilization of aqueous nanoscale zerovalent iron dispersi ons by anionic polyelectrolytes: Adsorbed anionic polyelectrolyte layer properties and their effect on aggregation and sedimentation. Journal of Nanoparticle Research, 2008. 10 (5): p. 795 814. 110. Wang, P., Shi, Q.H., Liang, H.J., Steuerman, D.W., Stucky, G.D., and Keller, A.A., Enhanced environmental mobility of carbon nanotubes in the presence of humic acid and their removal from aqueous solution. Small, 2008. 4 (12): p. 2166 2170. 111. Grasso, D., Subramaniam, K., Butkus, M., Strevett, K., and Bergendahl J., A review of non dlvo interactions in environmental colloidal systems. Reviews in Environmental Science and Biotechnology, 2002. 1 : p. 17 38. 112. Butt, H.J., Cappella, B., and Kappl, M., Force measurements with the atomic force microscope: Technique, interpretation and applications. Surface Science Reports, 2005. 59 (1 6): p. 1 152. 113. Chen, K.L. and Elimelech, M., Aggregation and deposition kinetics of fullerene (c 60) nanoparticles. Langmuir, 2006. 22 (26): p. 10994 11001. 114. Chen, K.L. and Elimel ech, M., Relating colloidal stability of fullerene (c 60) nanoparticles to nanoparticle charge and electrokinetic properties. Environmental Science & Technology, 2009. 43 (19): p. 7270 7276. 115. de Mesquita, L.M.S., Lins, F.F., and Torem, M.L., Interaction of a hydrophobic bacterium strain in a hematite quartz flotation system. International Journal Of Mineral Processing, 2003. 71 (1 4): p. 31 44. 116. Liu, Q., Lazouskaya, V., He, Q.X., and Jin, Y. Effect of particle shape on colloid retention and release in saturated porous media. Journal Of Environmental Quality, 2010. 39 (2): p. 500 508. 117. Tian, Y.A., Gao, B., and Ziegler, K.J., High mobility of sdbs dispersed single walled carbon nanotubes in saturated and unsaturated porous media. Journal Of Hazardous Materials, 2011. 186 (2 3): p. 1766 1772.

PAGE 214

214 118. Tian, Y.A., Gao, B., Silvera Batista, C., and Ziegler, K.J., Transport of engineered nanoparticles in saturated porous media. Journal of Nanoparticl e Research, 2010. 12 (7): p. 2371 2380. 119. Gao, B., Saiers, J.E., and Ryan, J., Pore scale mechanisms of colloid deposition and mobilization during steady and transient flow through unsaturated granular media. Water Resources Research, 2006. 42 (1): p. W0 1410, doi:10.1029/2005WR004233. 120. Zevi, Y., Dathe, A., Gao, B., Zhang, W., Richards, B.K., and Steenhuis, T.S., Transport and retention of colloidal particles in partially saturated porous media: Effect of ionic strength. Water Resources Research, 2009. 45 : p. W12403. 121. Bradford, S.A. a nd Kim, H., Implications of cation exchange on clay release and colloid facilitated transport in porous media. Journal Of Environmental Quality, 2010. 39 (6): p. 2040 2046. 122. Zhang, L., Petersen, E.J., and Huang, Q., Phase distribution of (14)c labeled m ultiwalled carbon nanotubes in aqueous systems containing model solids: Peat. Environmental Science & Technology, 2011. 45 (4): p. 1356 1362. 123. Muller, V.M., The electrical double layers and the colloidal stability a historical review of the dlvo theor y. Abstracts Of Papers Of The American Chemical Society, 1994. 208 : p. 210 COLL. 124. Johnson, W.P., Ma, H.L., and Pazmino, E., Straining credibility: A general comment regarding common arguments used to infer straining as the mechanism of colloid retentio n in porous media. Environmental Science & Technology, 2011. 45 (9): p. 3831 3832. 125. Xu, S.P., Gao, B., and Saiers, J.E., Straining of colloidal particles in saturated porous media. Water Resources Research, 2006. 42 (12): p. W12S16, doi:10.1029/2006WR004 948. 126. Chen, J.Y., Ko, C.H., Bhattacharjee, S., and Elimelech, M., Role of spatial distribution of porous medium surface charge heterogeneity in colloid transport. Colloids And Surfaces A physicochemical And Engineering Aspects, 2001. 191 (1 2): p. 3 15. 127. Walker, S.L., Redman, J.A., and Elimelech, M., Influence of growth phase on bacterial deposition: Interaction mechanisms in packed bed column and radial stagnation point flow systems. Environmental Science & Technology, 2005. 39 (17): p. 6405 6411. 12 8. Tufenkji, N. and Elimelech, M., Spatial distributions of cryptosporidium oocysts in porous media: Evidence for dual mode deposition. Environmental Science & Technology, 2005. 39 (10): p. 3620 3629.

PAGE 215

215 129. Lin, S., Cheng, Y., Bobcombe, Y., L. Jones, K., Liu J., and Wiesner, M.R., Deposition of silver nanoparticles in geochemically heterogeneous porous media: Predicting affinity from surface composition analysis. Environmental Science & Technology, 2011. 45 (12): p. 5209 5215. 130. de Kerchove, A.J., Weronski P., and Elimelech, M., Adhesion of nonmotile pseudomonas aeruginosa on "soft" polyelectrolyte layer in a radial stagnation point flow system: Measurements and model predictions. Langmuir, 2007. 23 (24): p. 12301 12308. 131. Morales, V.L., Gao, B., and Ste enhuis, T.S., Grain surface roughness effects on colloidal retention in the vadose zone. Vadose Zone Journal, 2009. 8 (1): p. 11 20. 132. Shen, C.Y., Li, B.G., Wang, C., Huang, Y.F., and Jin, Y., Surface roughness effect on deposition of nano and micro siz ed colloids in saturated columns at different solution ionic strengths. Vadose Zone Journal, 2011. 10 (3): p. 1071 1081. 133. Compere, F., Porel, G., and Delay, F., Transport and retention of clay particles in saturated porous media. Influence of ionic stre ngth and pore velocity. Journal Of Contaminant Hydrology, 2001. 49 (1 2): p. 1 21. 134. Redman, J.A., Walker, S.L., and Elimelech, M., Bacterial adhesion and transport in porous media: Role of the secondary energy minimum. Environmental Science & Technology 2004. 38 (6): p. 1777 1785. 135. Pelley, A.J. and Tufenkji, N., Effect of particle size and natural organic matter on the migration of nano and microscale latex particles in saturated porous media. Journal Of Colloid And Interface Science, 2008. 321 (1): p. 74 83. 136. Shen, C.Y., Li, B.G., Huang, Y.F., and Jin, Y., Kinetics of coupled primary and secondary minimum deposition of colloids under unfavorable chemical conditions. Environmental Science & Technology, 2007. 41 (20): p. 6976 6982. 137. Velegol, D. Anderson, J.L., and Garoff, S., Determining the forces between polystyrene latex spheres using differential electrophoresis. Langmuir, 1996. 12 (17): p. 4103 4110. 138. Bradford, S.A., Torkzaban, S., and Walker, S.L., Resolving the coupled effects of hydr odynamics and dlvo forces on colloid attachment in porous media. Langmuir, 2007. 23 (19): p. 9652 9660. 139. Pan, B. and Xing, B.S., Adsorption mechanisms of organic chemicals on carbon nanotubes. Environmental Science & Technology, 2008. 42 (24): p. 9005 90 13. 140. Flury, M. and Qiu, H.X., Modeling colloid facilitated contaminant transport in the vadose zone. Vadose Zone Journal, 2008. 7 (2): p. 682 697.

PAGE 216

216 141. Walshe, G.E., Pang, L.P., Flury, M., Close, M.E., and Flintoft, M., Effects of ph, ionic strength, di ssolved organic matter, and flow rate on the co transport of ms2 bacteriophages with kaolinite in gravel aquifer media. Water Research, 2010. 44 (4): p. 1255 1269. 142. Li, Y.S., Wang, Y.G., Pennell, K.D., and Abriola, L.M., Investigation of the transport and deposition of fullerene (c60) nanoparticles in quartz sands under varying flow conditions. Environmental Science & Technology, 2008. 42 (19): p. 7174 7180. 143. Liu, X.Y., O'Carroll, D.M., Petersen, E.J., Huang, Q.G., and Anderson, C.L., Mobility of multiwalled carbon nanotubes in porous media. Environmental Science & Technology, 2009. 43 (21): p. 8153 8158. 144. Lecoanet, H.F. and Wiesner, M.R., Velocity effects on fullerene and oxide nanoparticle deposition in porous media. Environmental Science & Technology, 2004. 38 (16): p. 4377 4382. 145. Han, J., Jin, Y., and Willson, C.S., Virus retention and transport in ch emically heterogeneous porous media under saturated and unsaturated flow conditions. Environmental Science & Technology, 2006. 40 (5): p. 1547 1555. 146. Gao, B., Saiers, J.E., and Ryan, J.N., Deposition and mobilization of clay colloids in unsaturated poro us media. Water Resources Research, 2004. 40 (8): p. W08602, doi:10.1029/2004WR003189. 147. Bradford, S.A., Simunek, J., Bettahar, M., Tadassa, Y.F., van Genuchten, M.T., and Yates, S.R., Straining of colloids at textural interfaces. Water Resources Researc h, 2005. 41 (10): p. W10404, doi:10.1029/2004WR003675. 148. Gao, B., Steenhuis, T.S., Zevi, Y., Morales, V.L., Nieber, J.L., Richards, B.K., McCarthy, J.F., and Parlange, J.Y., Capillary retention of colloids in unsaturated porous media. Water Resources Res earch, 2008. 44 (4): p. W04504, doi:10.1029/2006WR005332. 149. Johnson, W.P., Pazmino, E., and Ma, H.L., Direct observations of colloid retention in granular media in the presence of energy barriers, and implications for inferred mechanisms from indirect ob servations. Water Research, 2010. 44 (4): p. 1158 1169. 150. Iijima, S. and Ichihashi, T., Single shell carbon nanotubes of 1 nm diameter. Nature, 1993. 363 (6430): p. 603 15. 151. Barman, S.N., LeMieux, M.C., Baek, J., Rivera, R., and Bao, Z.N., Effects of dispersion conditions of single walled carbon nanotubes on the electrical characteristics of thin film network transistors. Acs Applied Materials & Interfaces, 2010. 2 (9): p. 2672 2678.

PAGE 217

217 152. Ziegler, K.J., Rauwald, U., Gu, Z.N., Liang, F., Billups, W.E., H auge, R.H., and Smalley, R.E., Statistically accurate length measurements of single walled carbon nanotubes. Journal of Nanoscience and Nanotechnology, 2007. 7 (8): p. 2917 2921. 153. Wan, J.M. and Tokunaga, T.K., Measuring partition coefficients of colloid s at air water interfaces. Environmental Science & Technology, 1998. 32 (21): p. 3293 3298. 154. Auset, M. and Keller, A.A., Pore scale visualization of colloid straining and filtration in saturated porous media using micromodels. Water Resources Research, 2006. 42 (12): p. 155. Saiers, J.E., Hornberger, G.M., and Liang, L.Y., First order and second order kinetics approaches for modeling the transport of colloidal particles in porous media. Water Resources Research, 1994. 30 (9): p. 2499 2506. 156. Ho, Y.S. and McKay, G., A comparison of chemisorption kinetic models applied to pollutant removal on various sorbents. Process Safety And Environmental Protection, 1998. 76 (B4): p. 332 340. 157. Chappell, M.A., George, A.J., Dontsova, K.M., Porter, B.E., Price, C. L., Zhou, P., Morikawa, E., Kennedy, A.J., and Steevens, J.A., Surfactive stabilization of multi walled carbon nanotube dispersions with dissolved humic substances. Environmental Pollution, 2009. 157 (4): p. 1081 1087. 158. Kumar, S.A. and Wang, S.F., Adsor ption of ciprofloxacin and its role for stabilizing multi walled carbon nanotubes and characterization. Materials Letters, 2009. 63 (21): p. 1830 1833. 159. Abu Lail, L., Bergendahl, J.A., and Thompson, R.W., Adsorption of methyl tertiary butyl ether on gra nular zeolites: Batch and column studies. Journal Of Hazardous Materials, 2010. 178 (1 3): p. 363 369. 160. Bao, Q.B., Lin, Q., Tian, G.M., Wang, G.H., Yu, J.A., and Peng, G.Q., Copper distribution in water dispersible colloids of swine manure and its trans port through quartz sand. Journal Of Hazardous Materials, 2011. 186 (2 3): p. 1660 1666. 161. Tian, Y., Gao, B., Morales, V.L., Carpena, R.M., Huang, Q., and Yang, L., Deposition and transport of functionalized carbon nanotubes in water saturated sand colum ns. Journal Of Hazardous Materials, 2012(213 214): p. 265 272. 162. Chen, G., Liu, J., Tawfiq, K., Yang, K., and Banks, C., Colloid retention in unsaturated porous media as impacted by colloid size. Particulate Science And Technology, 2009. 27 (1): p. 35 49

PAGE 218

218 163. Bouchard, D., Zhang, W., Powell, T., and Rattanaudompol, U.S., Aggregation kinetics and transport of single walled carbon nanotubes at low surfactant concentrations. Environmental Science & Technology, 2012. 46 (8): p. 4458 4465. 164. Oyanedel Craver V.A. and Smith, J.A., Sustainable colloidal silver impregnated ceramic filter for point of use water treatment. Environmental Science & Technology, 2008. 42 : p. 927 933. 165. van Genuchten, M.T. and Alves, W.J., Analytical solutions of the one dimensional convective dispersive solute transport equation. Technical Bulletion, U.S. Department of Agriculture, 1982. 1661 : p. 151. 166. Grolimund, D., Elimelech, M., Borkovec, M., Barm ettler, K., Kretzschmar, R., and Sticher, H., Transport of in situ mobilized colloidal particles in packed soil columns. Environmental Science & Technology, 1998. 32 (22): p. 3562 3569. 167. Crist, J.T., Zevi, Y., McCarthy, J.F., Throop, J.A., and Steenhuis T.S., Transport and retention mechanisms of colloids in partially saturated porous media. Vadose Zone Journal, 2005. 4 (1): p. 184 195. 168. Chatterjee, N., Lapin, S., and Flury, M., Capillary forces between sediment particles and an air water interface. Environmental Science & Technology, 2012. 46 (8): p. 4411 4418. 169. Keller, A.A. and Sirivithayapakorn, S., Transport of colloids in unsaturated porous media: Explaining large scale behavior based on pore scale mechanisms. Water Resources Research, 2004. 4 0 (12). 170. Akbour, R.A., Douch, J., Hamdani, M., and Schmitz, P., Transport of kaolinite colloids through quartz sand: Influence of humic acid, ca2+, and trace metals. Journal Of Colloid And Interface Science, 2002. 253 (1): p. 1 8. 171. Darnault, C.J.G. a nd Godinez, I.G., Aggregation and transport of nano tio(2) in saturated porous media: Effects of ph, surfactants and flow velocity. Water Research, 2011. 45 (2): p. 839 851. 172. Torkzaban, S., Kim, Y., Mulvihill, M., Wan, J., and Tokunaga, T.K., Transport and deposition of functionalized cdte nanoparticles in saturated porous media. Journal Of Contaminant Hydrology, 2010. 118 (3 4): p. 208 217. 173. Stafiej, A. and Pyrzynska, K., Solid phase extraction of metal ions using carbon nanotubes. Microchemical Journal, 2008. 89 (1): p. 29 33. 174. Yang, K., Wu, W.H., Jing, Q.F., and Zhu, L.Z., Aqueous adsorption of aniline, phenol, and their substitutes by mu lti walled carbon manotubes. Environmental Science & Technology, 2008. 42 (21): p. 7931 7936.

PAGE 219

219 175. Cheng, T. and Saiers, J.E., Colloid facilitated transport of cesium in vadose zone sediments: The importance of flow transients. Environmental Science & Techn ology, 2010. 44 (19): p. 7443 7449. 176. Sun, H.M., Gao, B., Tian, Y.A., Yin, X.Q., Yu, C.R., Wang, Y.Q., and Ma, L.N.Q., Kaolinite and lead in saturated porous media: Facilitated and impeded transport. Journal Of Environmental Engineering asce, 2010. 136 (1 1): p. 1305 1308. 177. subsurface mobility of carbon nanoparticles at the field scale. Advances In Water Resources, 2010. 33 (4): p. 361 371. 178. Porubcan, A.A. and Xu, S.P., Collo id straining within saturated heterogeneous porous media. Water Research, 2011. 45 (4): p. 1796 1806. 179. Noh, J.S. and Schwarz, J.A., Effect of hno3 treatment on the surface acidity of activated carbons. Carbon, 1990. 28 (5): p. 675 682. 180. Walkley, A. a nd Black, I.A., An examination of degtjareff method for determining soil organic matter and a proposed modification of the chromic acid titration method. Soil. Sci., 1934. 37 : p. 29 38. 181. Xu, S.P., Liao, Q., and Saiers, J.E., Straining of nonspherical c olloids in saturated porous media. Environmental Science & Technology, 2008. 42 (3): p. 771 778. 182. Simunek, J., van Genuchten, M.T., and Sejna, M., Development and applications of the hydrus and stanmod software packages and related codes. Vadose Zone Jo urnal, 2008. 7 (2): p. 587 600. 183. Castanheira, N.L., and Pereira, L.S., Field evaluation of a multicomponent solute transport model in soils irrigated with saline waters. Journal Of Hydrology, 2011. 407 (1 4): p. 129 144. 184. Kosmulski, M., The ph dependent surface charging and points of zero charge v. Update. Journal Of Colloid And Interface Science, 2011. 353 (1): p. 1 15. 185. Penn, R.L., Zhu, C., Xu, H., and Veblen, D.R., Iron oxid e coatings on sand grains from the atlantic coastal plain: High resolution transmission electron microscopy characterization. Geology, 2001. 29 (9): p. 843 846. 186. Tredgold, R.H., An introduction to ultrathin organic films from langmuir blodgett to self assembly ulman,a. Nature, 1991. 354 (6349): p. 120 120. 187. Saha, B., Das, S., Saikia, J., and Das, G., Preferential and enhanced adsorption of different dyes on iron oxide nanoparticles: A comparative study. The Journal of Physical Chemistry C, 2011. 115 (16): p. 8024 8033.

PAGE 220

220 188. Yu, S. and Chow, G.M., Carboxyl group ( co2h) functionalized ferrimagnetic iron oxide nanoparticles for potential bio applications. Journal Of Materials Chemistry, 2004. 14 (18): p. 2781 2786. 189. Silliman, S.E., Particle transp ort through two dimensional, saturated porous media: Influence of physical structure of the medium. Journal Of Hydrology, 1995. 167 (1 4): p. 79 98. 190. Kosmulski, M., Chemical properties of material surfaces. First ed2001: Marcel Dekker, New York. 191. Me yer, M., Be, A., Talbot, D., Cabuil, V., Boyer, J.M., Rpetti, B., and Garrigos, R., Synthesis and dispersion of ni(oh)2 platelet like nanoparticles in water. Journal Of Colloid And Interface Science, 2004. 277 (2): p. 309 315. 192. Solovitch, N., Labille, J., Rose, J., Chaurand, P., Borschneck, D., Wiesner, M.R., and Bottero, J.Y., Concurrent aggregation and deposition of tio2 nanoparticles in a sandy porous media. Environmental Science & Technology, 2010. 44 (13): p. 4897 4902. 193. Tian, Y., Gao, B., and Ziegler, K.J., High mobility of sdbs dispersed single walled carbon nanotubes in saturated and unsaturated porous media. Journal Of Hazardous Materials, 2011. 186 (2 3): p. 1766 1772. 194. Atieh, M.A., Bakather, O.Y., Al Tawbini, B., Bukhari, A.A., Abuilaiw i, F.A., and Fettouhi, M.B., Effect of carboxylic functional group functionalized on carbon nanotubes surface on the removal of lead from water. Bioinorganic Chemistry And Applications, 2010: p. 603978. 195. Tian, Y., Gao, B., Wang, Y., Morales, V.L., Carp ena, R.M., Huang, Q., and Yang, L., Deposition and transport of functionalized carbon nanotubes in water saturated sand columns. Journal Of Hazardous Materials, 2012. 213 214 (0): p. 265 272. 196. Diels, L., Spaans, P.H., Van Roy, S., Hooyberghs, L., Ryngae rt, A., Wouters, H., Walter, E., Winters, J., Macaskie, L., Finlay, J., Pernfuss, B., Woebking, H., Pumpel, T., and Tsezos, M., Heavy metals removal by sand filters inoculated with metal sorbing and precipitating bacteria. Hydrometallurgy, 2003. 71 (1 2): p 235 241. 197. Petersen, E.J., Zhang, L.W., Mattison, N.T., O'Carroll, D.M., Whelton, A.J., Uddin, N., Nguyen, T., Huang, Q.G., Henry, T.B., Holbrook, R.D., and Chen, K.L., Potential release pathways, environmental fate, and ecological risks of carbon nan otubes. Environmental Science & Technology, 2011. 45 (23): p. 9837 9856.

PAGE 221

221 198. Hiraoka, T., Izadi Najafabadi, A., Yamada, T., Futaba, D.N., Yasuda, S., Tanaike, O., Hatori, H., Yumura, M., Iijima, S., and Hata, K., Compact and light supercapacitor electrodes from a surface only solid by opened carbon nanotubes with 2 200 m(2) g( 1) surface area. Advanced Functional Materials, 2010. 20 (3): p. 422 428. 199. Mohapatra, M., A comparative study on pb (ii), cd (ii), cu (ii), co (ii) adsorption from single and binar y aqueous solutions on additive assisted nano structured goethite. International Journal Of Engineering Science, 2010. 2 200. Wang, H.J., Zhou, A.L., Peng, F., Yu, H., and Chen, L.F., Adsorption characteristic of acidified carbon nanotubes for heavy metal pb(ii) in aqueous solution. Materials Science and Engineering a Structural Materials Properties Microstructure and Processing, 2007. 466 (1 2): p. 201 206. 201. Peng, X.J., Jia, J.J., and Luan, Z.K., Oxidized carbon nanotubes for simultaneous removal of en drin and cd(ii) from water and their separation from water. Journal Of Chemical Technology And Biotechnology, 2009. 84 (2): p. 275 278. 202. Doong, R.A. and Chiang, L.F., Coupled removal of organic compounds and heavy metals by titanate/carbon nanotube composites. Water Science And Technology, 2008. 58 (10): p. 1985 1992. 203. Cao, X.D., Ma, L., Gao, B., and Harris, W., Dairy manure derived bioc har effectively sorbs lead and atrazine. Environmental Science & Technology, 2009. 43 (9): p. 3285 3291. 204. Inyang, M., Gao, B., Ding, W., Pullammanappallil, P., Zimmerman, A.R., and Cao, X., Enhanced lead sorption by biochar derived from anaerobically di gested sugarcane bagasse. Separation Science And Technology, 2011. 46 (12): p. 1950 1956. 205. Moreno Pirajan, J.C., Gomez Cruz, R., Garcia Cuello, V.S., and Giraldo, L., Binary system cu(ii)/pb(ii) adsorption on activated carbon obtained by pyrolysis of co w bone study. Journal Of Analytical And Applied Pyrolysis, 2010. 89 (1): p. 122 128. 206. Vukovic, G.D., Marinkovic, A.D., Colic, M., Ristic, M.D., Aleksic, R., Peric Grujic, A.A., and Uskokovic, P.S., Removal of cadmium from aqueous solutions by oxidized a nd ethylenediamine functionalized multi walled carbon nanotubes. Chemical Engineering Journal, 2010. 157 (1): p. 238 248. 207. Pillay, K., Cukrowska, E.M., and Coville, N.J., Multi walled carbon nanotubes as adsorbents for the removal of parts per billion l evels of hexavalent chromium from aqueous solution. Journal Of Hazardous Materials, 2009. 166 (2 3): p. 1067 1075.

PAGE 222

222 208. Sousa, F.W., Oliveira, A.G., Ribeiro, J.P., Rosa, M.F., Keukeleire, D., and Nascimento, R.F., Green coconut shells applied as adsorbent f or removal of toxic metal ions using fixed bed column technology. Journal Of Environmental Management, 2010. 91 (8): p. 1634 1640. 209. Thiele Bruhn, S., Pharmaceutical antibiotic compounds in soils a review. Journal of Plant Nutrition and Soil Science, 2 003. 166 (2): p. 145 167. 210. Halling Srensen, B., Nors Nielsen, S., Lanzky, P.F., Ingerslev, F., Holten Ltzhft, H.C., and Jrgensen, S.E., Occurrence, fate and effects of pharmaceutical substances in the environment a review. Chemosphere, 1998. 36 (2): p. 357 393. 211. Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S.D., Barber, L.B., and Buxton, H.T., Pharmaceuticals, hormones, and other organic wastewater Environment al Science & Technology, 2002. 36 (6): p. 1202 1211. 212. Gbel, A., Thomsen, A., McArdell, C.S., Joss, A., and Giger, W., Occurrence and sorption behavior of sulfonamides, macrolides, and trimethoprim in activated sludge treatment. Environmental Science & Technology, 2005. 39 (11): p. 3981 3989. 213. Park, S. and Choi, K., Hazard assessment of commonly used agricultural antibiotics on aquatic ecosystems. Ecotoxicology, 2008. 17 (6): p. 526 538. 214. cts of the presence of sulfonamides in the environment and their influence on human health. Journal Of Hazardous Materials, 2011. 196 (0): p. 1 15. 215. Beltrn, F.J., Aguinaco, A., Garca Araya, J.F., and Oropesa, A., Ozone and photocatalytic processes to remove the antibiotic sulfamethoxazole from water. Water Research, 2008. 42 (14): p. 3799 3808. 216. Choi, K. J., Son, H. J., and Kim, S. H., Ionic treatment for removal of sulfonamide and tetracycline classes of antibiotic. Science Of The Total Environment 2007. 387 (1 3): p. 247 256. 217. Adams, C., Wang, Y., Loftin, K., and Meyer, M., Removal of antibiotics from surface and distilled water in conventional water treatment processes. Journal of Environmental Engineering, 2002. 128 (3): p. 253 260. 218. Terne s, T.A., Meisenheimer, M., McDowell, D., Sacher, F., Brauch, H. J., Haist Gulde, B., Preuss, G., Wilme, U., and Zulei Seibert, N., Removal of pharmaceuticals during drinking water treatment. Environmental Science & Technology, 2002. 36 (17): p. 3855 3863.

PAGE 223

223 2 19. Avisar, D., Primor, O., Gozlan, I., and Mamane, H., Sorption of sulfonamides and tetracyclines to montmorillonite clay. Water Air And Soil Pollution, 2010. 209 (1 4): p. 439 450. 220. Braschi, I., Blasioli, S., Gigli, L., Gessa, C.E., Alberti, A., and M artucci, A., Removal of sulfonamide antibiotics from water: Evidence of adsorption into an organophilic zeolite y by its structural modifications. Journal Of Hazardous Materials, 2010. 178 (1 3): p. 218 225. 221. Caliskan, E. and Gokturk, S., Adsorption cha racteristics of sulfamethoxazole and metronidazole on activated carbon. Separation Science And Technology, 2010. 45 (2): p. 244 255. 222. Lertpaitoonpan, W., Ong, S.K., and Moorman, T.B., Effect of organic carbon and ph on soil sorption of sulfamethazine. C hemosphere, 2009. 76 (4): p. 558 564. 223. Gao, J.A. and Pedersen, J.A., Adsorption of sulfonamide antimicrobial agents to clay minerals. Environmental Science & Technology, 2005. 39 (24): p. 9509 9516. 224. Gao, J. and Pedersen, J.A., Sorption of sulfonamid e antimicrobial agents to humic acid clay complexes. Journal Of Environmental Quality, 2010. 39 (1): p. 228 235. 225. Wang, X., Li, Q., Xie, J., Jin, Z., Wang, J., Li, Y., Jiang, K., and Fan, S., Fabrication of ultralong and electrically uniform single wall ed carbon nanotubes on clean substrates. Nano Letters, 2009. 9 (9): p. 3137 3141. 226. Stafiej, A. and Pyrzynska, K., Adsorption of heavy metal ions with carbon nanotubes. Separation And Purification Technology, 2007. 58 (1): p. 49 52. 227. Ji, L.L., Chen, W ., Zheng, S.R., Xu, Z.Y., and Zhu, D.Q., Adsorption of sulfonamide antibiotics to multiwalled carbon nanotubes. Langmuir, 2009. 25 (19): p. 11608 11613. 228. Ji, L.L., Shao, Y., Xu, Z.Y., Zheng, S.R., and Zhu, D.Q., Adsorption of monoaromatic compounds and pharmaceutical antibiotics on carbon nanotubes activated by koh etching. Environmental Science & Technology, 2010. 44 (16): p. 6429 6436. 229. Eckenfelder, W.W., Industrial water pollution control2000, Boston, Mass. [u.a.]: McGraw Hill. 230. Faki, A.e., Tur an, M., Ozdemir, O., and Turan, A.Z., Analysis of fixed bed column adsorption of reactive yellow 176 onto surfactant modified zeolite. Industrial & Engineering Chemistry Research, 2008. 47 (18): p. 6999 7004. 231. Chen, H., Gao, B., Li, H., and Ma, L.Q., Effects of ph and ionic strength on sulfamethoxazole and ciprofloxacin transport in saturated porous media. Journal Of Contaminant Hydrology, 2011. 126 (1 2): p. 29 36.

PAGE 224

224 232. Gupta, V.K., Srivastava, S.K., Mohan, D., and Sharma, S., Design parameters for fixed bed reactors of activated carbon developed from fertilizer waste for the removal of some heavy metal ions. Waste Management, 1997. 17 (8): p. 517 522. 233. Gupta, V.K., Srivastava, S.K., and Tyagi, R., Design parameters for the treatment of phenolic wastes by carbon columns (obtained from fertilizer waste material). Water Research, 2000. 34 (5): p. 1543 1550. 234. Kundu, S. and Gupta, A.K., Analysis and modeli ng of fixed bed column operations on as(v) removal by adsorption onto iron oxide coated cement (iocc). Journal Of Colloid And Interface Science, 2005. 290 (1): p. 52 60. 235. Benefield, L.D., Judkins, J.F., and Weand, B.L., Process chemistry for water and w astewater treatment1982, Englewood Cliffs, N.J.: Prentice Hall. 236. Lin, D. and Xing, B., Adsorption of phenolic compounds by carbon nanotubes: Role of aromaticity and substitution of hydroxyl groups. Environmental Science & Technology, 2008. 42 (19): p. 7 254 7259. 237. Bohart, G.S. and Adams, E.Q., Some aspects of the behavior of charcoal with respect to chlorine.1. Journal Of The American Chemical Society, 1920. 42 (3): p. 523 544. 238. Hutchins, R.A., New method simplifies design of activated carbon syste ms. Chemical Engineering, 1973. 80 (19): p. 133 138. 239. Cortes Martinez, R., Solache Rios, M., Martinez Miranda, V., and Alfaro Cuevas, R., Removal of cadmium by natural and surfactant modified mexican zeolitic rocks in fixed bed columns. Water Air And So il Pollution, 2009. 196 (1 4): p. 199 210. 240. Ko, D.C.K., Porter, J.F., and McKay, G., Optimised correlations for the fixed bed adsorption of metal ions on bone char. Chemical Engineering Science, 2000. 55 (23): p. 5819 5829. 241. Patel, H. and Vashi, R.T. Fixed bed column adsorption of acid yellow 17 dye onto tamarind seed powder. The Canadian Journal of Chemical Engineering, 2012. 90 (1): p. 180 185. 242. Mohan, S. and Sreelakshmi, G., Fixed bed column study for heavy metal removal using phosphate treated rice husk. Journal Of Hazardous Materials, 2008. 153 (1 2): p. 75 82. 243. Ozdemir, O., Turan, M., Turan, A.Z., Faki, A., and Engin, A.B., Feasibility analysis of color removal from textile dyeing wastewater in a fixed bed column system by surfactant modif ied zeolite (smz). Journal Of Hazardous Materials, 2009. 166 (2 3): p. 647 654.

PAGE 225

225 244. Montazer Rahmati, M.M., Rabbani, P., Abdolali, A., and Keshtkar, A.R., Kinetics and equilibrium studies on biosorption of cadmium, lead, and nickel ions from aqueous soluti ons by intact and chemically modified brown algae. Journal Of Hazardous Materials, 2011. 185 (1): p. 401 407. 245. Nasuha, N., Hameed, B.H., and Din, A.T.M., Rejected tea as a potential low cost adsorbent for the removal of methylene blue. Journal Of Hazard ous Materials, 2010. 175 (1 3): p. 126 132. 246. Noh, J.S. and Schwarz, J.A., Effect of hno3 treatment on the surface acidity of activated carbons. Carbon, 1990. 28 (5): p. 675 682.

PAGE 226

226 BIOGRAPHICAL SKETCH Yu an Tian was born in Wuhan China H e received his e nvironmental e ngineering from School of Environmental Science and Engineering, Huazhong University of Science and Technology, Wuhan, China in 2005 After three year study and research in Nanjing University graduate school, he got a m aster s degree i n environmental scien ce in 200 8 He then started his doctoral research in Department of Agricultural and Biological Engineering, University of Florida, Gainesville, FL under the supervisory of Dr. Bin Gao